Comprehensive Coordination Chemistry: Applications

Citation preview

COMPREHENSIVE COORDINATION CHEMISTRY

The Synthesis, Reactions, Properties & Applications of Coordination Compounds

Editor-in-Chief: Sir Geoffrey Wilkinson, FRS Executive Editors: Robert D. Gillard & Jon A. McCleverty

Pergamon An imprint of Elsevier Science

COMPREHENSIVE COORDINATION CHEMISTRY IN 7 VOLUMES

PERGAMON MAJOR REFERENCE WORKS Comprehensive Inorganic Chemistry (1973) Comprehensive Organic Chemistry (1979) comprehensive Organometallic Chemistry (1982) Comprehensive Heterocyclic Chemistry (1984) International Encyclopedia of Education (1985) Comprehensive Insect Physiology, Biochemistry & Pharmacology (1985) Comprehensive Biotechnology (1985) Physics in Medicine & Biology Encyclopedia (1986) Encyclopedia of Materials Science & Engineering (1986) World Encyclopedia of Peace (1986) Systems & Control Encyclopedia (1987) Comprehensive Coordination Chemistry (1987) Comprehensive Polymer Science (1988) Comprehensive Medicinal Chemistry (1989)

COMPREHENSIVE COORDINATION CHEMISTRY The Synthesis, Reactions, Properties & Applications of Coordination Compounds Volume 6 Applications

EDITOR-IN-CHIEF

SIR GEOFFREY WILKINSON,

FRS

Imperial College of Science & Technology, University of London, UK EXECUTIVE EDITORS

ROBERT D. GILLARD

JON A. McCLEVERTY

Universiiy College, Cardig UK

University of Birmingham, UK

PERGAMON PRESS OXFORD * NEW YORK * BEIJING FRANKFURT S i 0 PAUL0 * SYDNEY * TOKYO * TORONTO

U.K.

Pergamon Press, Headington Hill Hall, Oxford OX3 OBW, En g1and

U.S.A.

Pergamon Press, Maxwell House, Fairview Park, Elmsford, New York 10523, U.S.A.

PEOPLES REPUBLIC OF CHINA

Pergamon Press, Room 4037, Qianmen Hotel, Beijing, People's Republic of China

FEDERAL REPUBLIC OF GERMANY

Pergamon Press, Hammenveg 6, D-6242 Kronberg, Federal Republic of Germany

BRAZIL

Pergamon Editora, Rue Ega de Queiros, 346, CEP 04011, Si0 Paulo, Brazil

AUSTRALIA

Pergamon Press Australia, P.O. Box 544, Potts Point, NSW 2011, Australia

JAPAN

Pergamon Press, 8th Floor, Matsuoka Central Building, 1-7-1 Nishishinjuku, Shinjuku-ku, Tokyo 160, Japan

CANADA

Pergamon Press Canada, Suite No. 271, 253 College Street, Toronto, Ontario M5T 1R5, Canada Copyright @ 1987 Pergamon Books Ltd.

All Rights Reserved. No parr of rhis publicarion may be reproduced, stored in a retriewl system or transmirred in any form or by any means: electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without permission in wriring from rhe publishers. First edition 1987 Library of Congress Cataloging-in-Publication Data Comprehensive coordination chemistry. Includes bibliographies. Contents: v. 1. Theory and background - v. 2. Ligands Main group and early transition elements - [etc.] 1. Coordination compounds. I. WiYkinson, Geoffrey, Sir, 192111. Gillard, Robert D. 111. McCleverty, Jon A. QD474.C65 1987 541.2'242 86-12319

-

v. 3.

British Library Cataloguing in Publication Data Comprehensive coordination chemistry: the synthesis, reactions, properties and applications of coordination compounds. 1. Coordination compounds. I. Wilkinson, Geoffrey, 192111. Gillard, Robert D. 111. McCleverty, Jon A. 541.2'242 QD474

ISBN 0-08-035949-3 (vol. 6 ) ISBN 0-08-026232-5 (set)

Printed in Great Britain by A. Wheaton & Co. Ltd., Exeter

Contents vii

Preface Contributors to Volume 6

ix

Contents of All Volumes

xi 1

57

Electrochemical Applications W. P. MCWHINNIE,University of Aston in Birmingham, UK

58

Dyes and Figments R. P R I C E , ICI PLC, Manchester, UK

35

59

Photographic Applications and E. R. SCHMITTOU, Easrman Kodak Company, Rochester, NY, D. D. CHAPMAN USA

95

60

Compounds Exhibiting Unusual Electrical Properties A. E. UNDERHILL, University College of North Wales, Bangor, UK

Uses in Synthesis and Catalysis 61.1 Stoichiometric Reactions of Coordinated Ligands D. St. C . BLACK,University of New South Wales, Kensington, NSW, Australia

133

155

61.2

Catalytic Activation o f Small Molecules A. S P E N C E Rformerly , of Ciba-Geigy AG, Basel, Switzerland

229

61.3

Metal Complexes in Oxidation H . MIMOUN,Institur Frangais du Petrole, Rueil-Malmaison, France

317

61.4

Lewis Acid Catalysis and the Reactions of Coordinated Ligands R. W . HAY, University of Stirling, UK

41 1

61.5

Decomposition of Water into its Elements D. J . COLE-HAMILTON, University o f s t Andrews, U K and D. W. BRUCE, University of Liverpool, UK

487

Biological and Medical Aspects 62.1 Coordination Compounds in Biology M. N. HUGHES,King’s College London, UK

541

62.2

Uses in Therapy H. E. HOWARD-LOCK and C. J . L. LOCK,McMaster University, Hamilton, Ontario, Canada

755

63

Application to Extractive Metallurgy M. J. N r c o ~ C. , A. FLEMING and J. S. PRESTON, Council for Mineral Technology, Randburg, Transvaal, South Africa

779

64

Geochemical and Prebiotic Systems P. A. WILLIAMS, University College, Cardix UK

843

65

Applications in the Nuclear Fuel Cycle and Radiopharmacy C. J . J O N E S , University of Birmingham, UK

881

V

vi 66

Contents Other Uses of Coordination Compounds A. SPENCER, formerly of Ciba-Geigy AG, Basel, Switzerland

1011

Subject Index

1031

Formula Index

1079

Preface Since the appearance of water on the Earth, aqua complex ions of metals must have existed. The subsequent appearance of life depended on, and may even have resulted from, interaction of metal ions with organic molecules. Attempts to use consciously and to understand the metalbinding properties of what are now recognized as electron-donating molecules or anions (iigands) date from the development of analytical procedures for metals by Berzelius and his contemporaries. Typically, by 1897, Ostwald could point out, in his ‘Scientific Foundations of Analytical Chemistry’, the high stability of cyanomercurate(I1) species and that ‘notwithstanding the extremely poisonous character of its constituents, it exerts no appreciable poison effect’. By the late 19th century there were numerous examples of the complexing of metal ions, and the synthesis of the great variety of metal complexes that could be isolated and crystallized was being rapidly developed by chemists such as S. M. Jorgensen in Copenhagen. Attempts to understand the ‘residual affinity’ of metal ions for other molecules and anions culminated in the theories of Alfred Werner, although it is salutary to remember that his views were by no means universally accepted until the mid-1920s. The progress in studies of metal complex chemistry was rapid, perhaps partly because of the utility and economic importance of metal chemistry, but also because of the intrinsic interest of many of the compounds and the intellectual challenge of the structural problems to be solved. If we define a coordination compound as the product of association of a Bronsted base with a Lewis acid, then there is an infinite variety of complexing systems. In this treatise we have made an arbitrary distinction between coordination compounds and organometallic compounds that have metal-carbon bonds. This division roughly corresponds to the distinction - which most but not all chemists would acknowledge as a real one -between the cobalt(II1) ions [Co(NH3)J3’ and [Co( V ~ - C ~ H ~ ) Any ~ ] ’ .species where the number of metal-carbon bonds is at least half the coordination number of the metal is deemed to be ‘organometallic’ and is outside the scope of our coverage; such compounds have been treated in detail in the companion work, ‘Comprehensive Organometallic Chemistry’. It is a measure of the arbitrariness and overlap between the two areas that several chapters in the present work are by authors who also contributed to the organometallic volumes. We have attempted to give a contemporary overview of the whole field which we hope will provide not only a convenient source of information but also ideas for further advances on the solid research base that has come from so much dedicated effort in laboratories all over the world. The first volume describes general aspects of the field from history, through nomenclature, to a discussion of the current position of mechanistic and related studies. The binding of ligands according to donor atoms is then considered (Volume 2 ) and the coordination chemistry of the elements is treated (Volumes 3, 4 and 5) in the common order based on the Periodic Table. The sequence of treatment of complexes of particular ligands for each metal follows the order given in the discussion of parent ligands. Volume 6 considers the applications and importance of coordination chemistry in several areas (from industrial catalysis to photography, from geochemistry to medicine). Volume 7 contains cumulative indexes which will render the mass of information in these volumes even more accessible to users. The chapters have been written by industrial and academic research workers from many countries, all actively engaged in the relevant areas, and we are exceedingly grateful for the arduous efforts that have made this treatise possible. They have our most sincere thanks and appreciation. We wish to pay tribute to the memories of Professor Martin Nelson and Dr Tony Stephenson who died after completion of their manuscripts, and we wish to convey our deepest sympathies to their families. We are grateful to their collaborators for finalizing their contributions for publication. Because o f ill health and other factors beyond the editors’ control, the manuscripts for the chapters on Phosphorus Ligands and Technetium were not available in time for publication. However, it is anticipated that the material for these chapters will appear in the journal Polyhedron in due course as Polyhedron Reports. We should like to acknowledge the way in which the staff at the publisher, particulariy Dr Colin Drayton and his dedicated editorial team, have supported the editors and authors in our vii

...

Vlll

Preface

However, it is anticipated that the material for these chapters will appear in the journal Polyhedron in due course as Polyhedron Reports. We should also like to acknowledge the way in which the staff at the publisher, particularly Dr Colin Drayton and his dedicated editorial team, have supported the editors and authors in our endeavour to produce a work which correctly portrays the relevance and achievements of modern coordination chemistry. We hope that users of these volumes will find them as full of novel information and as great a stimulus to new work as we believe them to be. JON A. McCLEVERTY Birmingham

ROBERT D. GILLARD

Cardif GEOFFREY WILKINSON London

Contributors to Volume 6 Professor D. St. C. BIack School of Chemistry, University of New South Wales, PO Box 1, Kensington, NSW 2033, Australia Dr D. W. Bruce Department of Chemistry, University of Sheffield, Sheffield S3 7HF, UK Dr D. D. Chapman Photoscience Division, Kodak Research Laboratories, Eastman Kodak Co., 1669 Lake Avenue, Rochester, NY 14650, USA Professor D. J. Cole-Hamilton Department of Chemistry, University of St Andrews, Purdie Building, St Andrews, Fife KY16 9ST, UK Dr C. A. Fleming Mineral and Process Chemistry Division, Mintek, Council for Mineral Technology, Private Bag X3015, Randburg 2125, Transvaal, South Africa Dr R. W. Hay Department of Chemistry, University of Stirling, Stirling FK9 4LA, UK Dr H. E. Howard-Lock Institute for Materials Research and Laboratories for Inorganic Medicine, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M1, Canada Dr M. N. Hughes Department of Chemistry, King’s College London, Strand, London WC2R 2LS, U K Dr C. J. Jones Department of Chemistry, University of Birmingham, PO Box 363, Birmingham B15 2TT, UK Professor C. J. L. Lock Institute for Materials Research and Laboratories for Inorganic Medicine, McMaster University, 1280 Main Street West, Hamilton, Ontario LXS 4M1, Canada Professor W. R. McWhinnie Department of Molecular Sciences, University of Aston in Birmingham, Gosta Green, Birmingham B4 7ET, U K Dr W. Mirnoun Institut Franpis du Petrole, BP 311, 92506 Rueil-Malmaison Cedex, France Dr M. J. Nicol Mineral and Process Chemistry Division, Mintek, Council for Mineral Technology, Private Bag X3015, Randburg 2125, Transvaal, South Africa CCC6-A*

ix

X

Contributors to Volume 6

Dr J. S. Preston Mineral and Process Chemistry Division, Mintek, Council for Mineral Technology, Private Bag X3015, Randburg 2125, Transvaal, South Africa Dr R. Price IC1 PLC (Organics Division), Hexagon House, PO Box 42, Blackley, Manchester M9 3DA, UK Dr E. R. Schmittou Color Instant Photography Division, Research Laboratories, Eastman Kodak Co., 1999 Lake Avenue, Rochester, NY 14650, USA Dr A. Spencer 111 Killigrew St, Falrnouth, Cornwall TR11 3PU, UK Professor A. E. Underhill Department of Chemistry, University College of North Wales, Bangor, Gwynedd LL57 2UW, UK Dr P. A. Williams Department of Chemistry, University College, PO Box 78, Cardiff CF1 IXL, UK

Contents of All Volumes Volume 1 Theory & Background 1.1 1.2 2 3 4 5 6 7.1

7.2 7.3 7.4 7.5 8.1 8.2 8.3

9 10

General Historical Survey to 1930 Development of Coordination Chemistry Since 1930 Coordination Numbers and Geometries Nomenclature of Coordination Compounds Cages and Clusters Isomerism in Coordination Chemistry Ligand Field Theory Reaction Mechanisms Substitution Reactions Electron Transfer Reactions Photochemical Processes Reactions of Coordinated Ligands Reactions in the Solid State Complexes in Aqueous and Non-aqueous Media Electrochemistry and Coordination Chemistry Electrochemical Properties in Aqueous Solutions Electrochemical Properties in Non-aqueous Solutions Quantitative Aspects of Solvent Effects Applications in Analysis Subject lndex Formula Index

Volume 2 Ligands 11 12.1 12.2 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 14 15.1 15.2 15.3 15.4 15.5 15.6 15.7

Mercury as a Ligand Group IV Ligands Cyanides and Fulminates Silicon, Germanium, Tin and Lead Nitrogen Ligands Ammonia and Amines Heterocyclic Nitrogen-donor Ligands Miscellaneous Nitrogen-containing Ligands Amido and Imido Metal Complexes Sulfurdiimine, Triazenido, Azabutadiene and Triatomic Hetero Anion Ligands Polypyrazolylborates and Related Ligands Nitriles Oximes, Guanidines and Related Species Phosphorus, Arsenic, Antimony and Bismuth Ligands Oxygen Ligands Water, Hydroxide and Oxide Dioxygen, Superoxide and Peroxide Alkoxides and Aryloxides Diketones and Related Ligands Oxyanions Carboxylates, Squarates and Related Species Hydroxy Acids xi

Contents qf all Volumes

xii

Sulfoxides, Amides, Amine Oxides and Related Ligands Hydroxamates, Cupferron and Related Ligands Sulfur Ligands 16.1 Sulfides 16.2 Thioethers 16.3 Metallothio Anions 16.4 Dithiocarbamates and Related Ligands 16.5 Dithiolenes and Related Species 16.6 Other Sulfur-containing Ligands 17 Selenium and Tellurium Ligands Halogens as Ligands 18 Hydrogen and Hydrides as Ligands 19

15.8 15.9

20.1 20.2 20.3 20.4

Schiff Bases as Acyclic Polydentate Ligands Amino Acids, Peptides and Proteins Complexones Bidentate Ligands

21.1

Porphyrins, Hydroporphyrins, Azaporphyrins, Phthalocyanines, Corroles, Corrins and Kelated Macrocycles Other Polyaza Macrocycles Multidentate Macrocyclic and Macropolycyclic Ligands Ligands of Biological Importance Subject Index

21.2 21.3 22

Volume 3 Main Group & Early Transition Elements

23 24 25.1 25.2 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

Alkali Metals and Group IIA metals Boron Aluminum and Gallium lndium and Thallium Silicon, Germanium, Tin and Lead Phosphorus Arsenic, Antimony and Bismuth Sulfur, Selenium, Tellurium and Polonium Halogenium Species and Noble Gases Titanium Zirconium and Hafnium Vanadium Niobium and Tantalum Chromium Molybdenum Tungsten Isopolyanions and Heteropolyanions Scandium, Yttrium and the Lanthanides The Actinides Subject Index Formula Index

Volume 4 Middle Transition Elements 41 42 43

Manganese Technetium Rhenium

Contents of All Volumes

44.1 44.2 45 46 47 48 49

Iron Iron(1I) and Lower States Iron(II1) and Higher States Ruthenium Osmium Cobalt Rhodium Iridium Suhject Index Formula Index

Volume 5 Late Transition Elements 50 51

52 53 54 55

56.1 56.2

Nickel Palladium Platinum Copper Silver Gold Zinc and Cadmium Mercury Subject Index Formula Index

Volume 6 Applications 57 58 59 60

61.1 61.2 61.3 61.4 61.5 62.1 62.2 63 64 65 66

Electrochemical Applications Dyes and Pigments Photographic Applications Compounds Exhibiting Unusual Electrical Properties Uses in Synthesis and Catalysis Stoichiometric Reactions of Coordinated Ligands Catalytic Activation of Small Molecules Metal Complexes in Oxidation Lewis Acid Catalysis and the Reactions of Coordinated Ligands Decomposition of Water into its Elements Biological and Medical Aspects Coordination Compounds in Biology Uses in Therapy Application to Extractive Metallurgy Geochemical and Prebiotic Systems Applications in the Nuclear Fuel Cycle and Radiopharmacy Other Uses of Coordination Compounds Subject Index Formula Index

Volume 7 67

Indexes

Index of Review Articles and Specialist Texts Cumulative Subject Index Cumulative Formula Index

...

Xlll

57 Electrochemical Applications WILLIAM R. McWHlNNIE University of Astun in Birmingham, UK 57.1 INTRODUCTION

i

57.2 ELECTRODEPOSITION OF METALS 57.2.1 Electrochemicul Cells 57.2.2 Overvoltage m d Polarization 57.2.3 Role of Coordination Compounds 57.2.4 Addition Reagents 57.2.5 Electrodeposition of Specific Metals 57.2.5.1 Chromium 57.2.5.2 Copper 57.2.5.3 Nick1 57.2.5.4 Precious metals 57.2.5.5 Tin 57.2.5.6 Zinc

1 1 3 4 5

7 7 9 10 11 11 12

13 14

57.2.5.7 Ah.Yphthtg

57.2.5.8 Miscellaneous ligands 57.2.6 Concludirrg Remtvks

14

57.3 COORDINATION COMPOUNDS A N D ELECTRODE PHENOMENA SURFACE MODIFIED

15

ELECTRODES 15 15 15 16

57.3.1 Introduction 57.3.2 Surface M o d f i d Electrodes 57.3.2.1 Nafwn mod@d electrodes 57.3.2.2 Poly-4-vinylpyridine and related polymers 57.3.2.3 Polyvinrrferrocene and related systems 57.3.2.4 Prussian Blue modi3ed electrodes 57.3.2.5 Miscellaneous surface modified electrodes 57.3.2.6 Applications

19 21 23 26

57.4 GENERAL CONCLUSION

31

57.5 REFERENCES

31

57.1 INTRODUCTION This chapter will carefully differentiate situations in which coordination of metal ions assists in the achievement of specific electrochemical aims from experiments designed to study the electrochemistry of coordination compounds. (For information on the latter topic, see, particularIy, Chapters 8.1-8.3). Two major areas have been selected €or consideration: one almost classical, namely the electrodeposition of metals, the other of more recent origin, namely the modification of electrode surfaces. The two areas specified will require the coverage of a wide range of chemistry. Although the basic principles of electroplating are well understood, and the role of coordination compounds in some plating baths well established, new work and new directions are discernible, particularly in relation to bath additives and their mode of action. Work on the modification of electrode surfaces is of more recent origin, but the research has already gathered considerable momentum. Coordination compounds have not played a unique role in this development, but they have on occasion made important advances possible. Some authors believe that this work will influence the direction of electrochemistry for many years to come. It certainly has implications for electrocatalysis, electronic devices, visual display units and photoelectricity to mention but a few topical objectives which currently drive the research. 57.2 ELECTRODEPOSITION OF METALS 57.2.1 Electrochemical Cells Within an electroplating bath the deposition of a metal from solution on to a solid substrate clearly involves the transport of charge across the solution-substrate interface. The solution will 1

Elec trochernical Applications

2

be an electrolyte, i.e. a phase in which charge is carried by ions, and the solid substrate will generally be a metal electrode in which charge is carried by electronic conduction. The focus of interest is generally the e!zctrode at which electrodeposition occurs; however, it is not possible to study that electrode-electrolyte interface in isolation since it must be linked to at least one other such interface to form an electrochemical cell. Electrochemical cells may be one of two types. Should a current spontaneously flow on connecting the electrodes via a conductor, the cell is a gaivanic cell. An electrolytic cell is one in which reactions occur when an external voltage greater than the reversible potential of the cell is applied. Simple examples involving copper are given in Figure 1. It is the electrolyticcell which is of interest in the electrodeposition of metals. c -

P

Electron f l o w

Electron f l o w

(b)

(a)

Figure 1 Examples of (a) a galvanic cell and (b) an electrolytic oell. Copper metal i s the cathode in (a), but the anode in (b)

When a metal is immersed in a solution of an electrolyte, a potential difference is set up at the mqtal-solution interface; this is the electrode potential. When a metal dips into a solution of its own ions, some ions may leave the metal and enter the solution, while others will deposit on the metal from sohtion. Since the ions are charged, an electrical double layer is created at the metal-solution interface. The equilibrium potential difference between metal and solution is the Galvani potential. When ions are transferred from solution to deposit on the metal, the metal consititutes the positive side of the double layer and vice versa. In practice, in order to assign a relative numerical value for an electrode potential, two electrodes are combined in an electrochemical cell and the potential of one electrode, the reference electrode, is assigned an arbitrary value. The standard hydrogen electrode (SHE; platinum black electrode contacting dihydrogen gas at 1 atm and hydrogen ions at unit activity) is assumed to have a potential of zero. Experimentally more convenient is the standard calomel electrode (SCE; HglHg2ClZIKC1saturated in water), which has a potential of 0.242 V vs. the standard hydrogen electrode. It is clear that the passage of charge across the electrode-solution interface may cause oxidation or reduction. Since the amount of chemical change is governed by Faraday's laws, such processes are termed faradaic. A link exists between the electrode potential, E, and the concentrations (properly activities) of compounds of the electrode process. In general Ox

+ ne-

F==

Red

(1)

where Ox and Red are the oxidized and reduced forms, respectively; hence using the Nernst equation E

=

E" $. RTInF In(q,x/aRcd)

(2)

where R is the gas constant, F the faraday and u represents activity. At unit activities E-E"

(3)

where E" is the standard potential. In the case where the reduced form is a metal M Mn++ ne

F==M

(4)

-+ RTlmF In Wni

(5)

Equation (2) may be rewritten as E = E"

Thus it follows that a change in the concentrationof M d will influence E. To take a simple example Co(s) + Ni2 '(as)

=

Cozt(aq)

+ Ni(s)

(6)

If both aquo ions are present in MIM2+(aq)half cells at 1 M concentration,nickel metalis deposited on connecting the half cells. However, if the concentration of Co*+(aq)were reduced to 0.01 M in the cobalt half cell, deposition of cobalt would occur. Thus a change in concentration of a factor

3

Electrochemical Applications

of ten has caused E to change by the order of the difference of the E" values for the two half cells ( i e . 0.03 V). It is not appropriate in this chapter to tabulate quantities of electrochemical data since that required may be obtained from texts on electrodeposition.lJ However, a brief mention of sign conventions must be made since, particularly in the early literature, confusion can arise. Two conventionshave been used: the 'European' and the 'Ameri~an'.~ It is sometimes erroneously stated that the conventions differ only in sign; however, the real difference lies in the distinction between the potential of an actual electrode and the EMF of a half cell reaction. The European convention assigns the sign (positive or negative) which the electrode would assume when immersed in a solution of its own ions. Thus a zinc electrode in a solution of Zn2+(aq) would assume a negative potential and P is -0.763 V. From a thermodynamic viewpoint the entry of zinc ions from the electrode to the solution is a spontaneous process in the presence of acid, hence AGO = -nFE"

(7)

The American convention would assign a positive value to E" for the ZnlZn2+(aq)half cell written as an oxidation, but a negative sign if written as a reduction. It is seen that the European convention refers to the invariant electrostatic potential of the electrode with respect to the SHE, whereas the American convention relates to the thermodynamic Gibbs' free energy which is sensitive to the direction of the cell reaction. IUPAC recommends that 'electrode potential' be reserved for the European convention, whereas the EMF of a half cell is dealt with via the American convention. The predictive value of standard electrode potentials is nicely illustrated following Bard and Faulkner (an excellent source of fundamental theory) in Figure L4As the potential of the Pt or Hg electrode is moved from its equilibrium (zero current) value to more negative potentials, the species of more positive E" should be reduced first. This prediction is verified for the Pt electrode, but Cr3+is reduced before H+ for the Hg electrode. -

-

-0.76

ZnZ+t2e-

Zn

-0.41

Cr3+ + e -

Cr'+

2H'

H2

-0.25 0

0

+2e-

0 . 15 0.77

+

+ io 1 P t electrode 0 01 M ~ ~ l u t i o nofs Fe3', Sn"+. k ? + In I M Y C t

Hg electrode. Zrt7+.CrZt in

0 01 M solutions of IM H C I

Cr3',

Figure 2

As is often the case in chemistry, it must be recognized that for real processes kinetic as well as thermodynamic factors must be considered.

57.2.2 Overvoltage and Polarization The equilibrium at an electrode is dynamic, the potential determining ions moving in opposite directions at equal rates. If the process is reversible, the ions move smoothly from one phase to the other via the phase boundary and perform no work. The electrode potential remains constant and independent of current. Kinetic factors may induce a variation of electrodepotential with current; the difference between this potential and the thermodynamic equilibrium potential is known as the overvodtage and the electrode is said to be polarized. In a plating bath this change: of potential can be attributed to the reduced concentration of depositing ions in the double layer which reduces the rate of transfer to the electrode, but the dissdution rate from the metal increases. Since the balance of these rates determines the electrode potential, a negative shift in the value occurs: the concentration polarization (qco,). Anodic effects are similar but in the opposite direction.

Electrochemical Applications

4

Other contributions to the polarization are activation polarizations (qact)caused by inhibition of the passage of ions through the phase boundary which may arise in the discharge mechanism. Films on the electrode may also contribute (e.g. oxide, metal already deposited, impurity) by offering a resistance to current flow differing from the bath resistance (ohmic polarization, qOhm).Hence the observed overvoltage is given by E - E eqwhbnum .. . =

('lconc

+ 'lact + 9 0 h ) ANODE +

('lconc

+ 9act + 'lohm) CATHODE +

A

(8)

where A represents the ohmic term for the bulk bath. Quite commonly there is an overvoltage for hydrogen discharge, as in the example given in Figure 2. Current-potential curves obtained under steady state conditions are called polarization curves. If two or more faradaic processes occur at the electrode, the fraction of current (ir)driving the rth process is the instantaneous current efficiency. Over a period of operating time the fraction of the total number of coulombs used in the rth process (QR) is related to the overall current efficiency of that process (OCE), i.e. OCE

= &/Qtotd

(9)

Should this approach loo%, only one electrode process occurs. 57.2.3

Role of Coordination Compounds

The majority of commercial electroplating involves an aqueous medium; hence, depending on the pH of the solution, the metal ions will be present in the bath as aquo or hydroxo complexes. In many cases another complexing agent, usually cyanide, is added. Such baths are operated under alkaline conditions to prevent the release of HCN. The choice of a particular bath composition is determined by technical factors.Thus, for example, zinc-alkaline cyanide baths have an excellent macroscopic throwing power, that is the ability to produce a uniform plated surface on a substrate that has macroscopic irregularities. The immediate effect of adding a strong ligand such as cyanide to a plating bath will be to alter the concentration of free metal ion. If we consider M"+ + mCN-

-

[M(CN),](m-")-

(10)

The activity of the free metal ions aM"+ is given by %+

= ucomph /%iab(wN)m

where Kstabis the stability constant of the cyano complex. Substituting in equation ( 5 ) gives E

=

E" - R T / n F In Kstab+ RTf nF In acomp,ex/(aCN)m

(12)

Since for stable cyano complexes &tab will be large, the hKstab term becomes dominant and causes a large negative shift in potential. This can have some interesting effects. For example, in a mixed copper-zinc bath containing cyanide, since Kstab(Cu)> Kstab(Zn),it is possible for the two metals to be deposited at almost the same potential and, indeed, the alloy brass m a y be plated from such solutions (vide infra). Although metallurgists have done much work on the morphology of the electrodeposited metal and the chemistry of the bulk bath solution can be understood, ideas covering the actual deposition act remain qualitative for the most part. If deposition occurs from an aquo ion, one view2 of the process is that the ion becomes polarized in a diffusion layer adjacent to the Helmholtz double layer, the water ligands being repelled from the side of the complex approaching the electrode. Removal of coordinated water and discharging then occur in the double layer and the neutral atom is adsorbed on to the electrode surface. Migration of the atom over the surface to a crystal growth point then occurs. Although there is support for the contention that the discharged metal may move to a site of crystal growth, the prior concept of an uncharged metal atom transferring from the Helmholtz layer to the electrode surface seems less attractive. A large number of commercially important plating processes occur from complex ion baths in which the metal is a constituent of an anionic complex, e.g. copper, zinc, cadmium, silver and gold are all commonly plated from cyanide baths, and tin plates from a stannate bath in which [Snrv(OH)6]2-is present. Chromium is commonly plated from a chromate bath although in this case the background medium is acid rather than alkaline. Thus the mechanism of deposition of metals from anionic complexes is of particular interest. It will be instructive to comment on two situations, one occurring in alkaline baths, the other in acidic baths.

Electrochemical Applications

5

Although, as explained above, zinc is commonly deposited from cyano baths, it may also be deposited from strongly alkatine baths (‘alkaline non-cyanide bath’). Typically a zinc : hydroxide ratio of 1 : 10 would be employed and the major solution species would be the tetrahedral zincate anion [Zn(OH),]2- as evidenced from Raman studies.5 Careful studies have been carried out to investigate the electrode reaction.6 An electrokinetic study gave data consistent with a four-step mechanism for the overall process [ih(OH)4]2-

+

2e- = Zn

+ 40H-

(13)

The rate determining step was identified as equation (1 5). [Zn(OH)4]2-

[ZU(OH)~]-

F==+

[ZII(OH)~]- f e-

=

+

[ZI-I(OH)~]-

OH-

+

(14)

(15)

OH-

There was reason to believe that the dissociation step in equation (14) was in fact a proton transfer [ZII(OH)~]’-

+

HOH

__

[(HO)3Zn(H20)]-

+

OH-

(16)

and that, therefore, the complex species present may be written as [Zn(OH)z{H20)4-J2-i.A range of possibilities exists for the step following the rate determining step. Since mechanisms requiring the adsorption of negatively charged species on an electrode at a high negative potential may be rejected as unlikely, choice fell upon a mechanism requiring deposition of zinc from an uncharged species, the data favouring the following overall process [Zn(OH),]’[ZU(OH)~]- + e[Zn(OH)2][Zn(OH)] + e-

__ F==+

+

[Zn(OH)3][Zr1(0€€)~]-

CZn(UH)] + Zn + OH-

+ +

OHOHOH-

The rate determining step need not always be, as in this case, one of the reduction steps. Thus at low overpotential, slow surface diffusion was rate deterrnining for the deposition of ~ o p p e r . ~ Plating from chromate(V1) baths involves an altogether more complex set of processes which are not yet fully understood.s The dominant ions in the baths are [HCr04]- when the solutions are dilute, and [HCr207]- when more concentrated. A quantity of chromium(II1) is generally present. In the absence of foreign ions a film forms on the cathode (‘basic chromium chromate’) which, whilst porous to hydrogen, prevents reduction of the chromate(V1) species. The presenoe of foreign anions, particularly fluoride or sulfate, will loosen the film and allow reduction of chromate to proceed. Complexes of sulfate with chromiumfII1) have been incriminated. It is believed that their solubility can prevent the formation of the film and can help to break down the film initially formed. Certainly one complex containing Cr:S04 = 2: 1 has been identified. This stoichiometry suggests a bridging sulfato complex. It has been established that most cathode metals are to some extent soluble in chromic acid solutions, and ions will enter the solution in the highest available oxidation state [e.g. copper(II), gold(III)]. Polarization of the cathode will then cause reduction to lower oxidation states [kinetic factors will prevent the prior raduction of chromate(VI)j, then new low-valent species may then initiate a chemical reduction of the chromium(V1). Chromium deposition occurs within the potential range for the evolution of dihydrogen and, indeed, the latter is the dominant cathode process with the result that typically cathode current efficiencies of only 10-20% are achieved (see equation 9). Although some chromium(II1) is initially present in the bath, early radiochemical data1*suggest that the deposited chromium does originate exclusively from chromate(V1). 57.2.4

Addition Reagents

The success of a commercial process depends very much on addition reagents. Two important classes are the levelling agents and the brightening agents. Levelling agents will encourage deposits to grow more rapidly on microindentations on the surface and less rapidly on peaks, thus achieving a finished surface with a high degree of smoothness. Marketability of a product is usually increased if the electrodeposit is bright. Brighteners can achieve lustrous coatings without the need for subsequent buffing. It is interesting that even coatings of zinc or cadmium, which are unlikely to retain their brightness in use are, none the less, more acceptable to customers when bright.

Electrochemical Applications

6

The role, if any, of coordination chemistry in the processes is obscure. Some of the additives are potential ligands, although they may not be effective as such in the bulk solution; however if, as seems likely, they are active at the electrode-solution interface," some possibility of coordination compound formation must be allowed. Examination of deposits has revealed that material which could only have originated from the addition reagent has been incorporated into the electrodeposit." This lends support to the view that the agents do indeed adsorb on to the electrode surface. It will be neither profitable nor relevant to study each additive to each plating process, but it should be recognized that the additives themselves need not be chemically or electrochemically inert. Some examples follow. Coumarin (1) is a leveller and partial brightener in nickel plating." It can be electrochemically reduced to (2): at pH 4,90% is reduced to (2) and only 10% incorporated in the deposit; at pH 2 the figures are 99% and 1%. Melilotic acid (2) is poorly adsorbed at the electrode.

(1)

(2)

Some recent work on brighteners in zinc plating is also of interest. The first report of successful plating from an alkaline cyanide zinc bath was in 190712but it was not until 1935 that the first bright alkaline cyanide zinc baths appeared in the plating industry. Numerous patents have been issued since 1935 claiming novel brighteners and many commercial processes use mixtures of additives which may act synergistically. A vast range of substances has been employed including metal ions (e.g. Ni", CrI" (but not C P ) , SeIV,TeIV);animal proteins such as glue and gelatine, and aromatic aldehydes such as p-anisaldehyde (3)or o-vanillin (4) have more recently proved effective. An important patentI3 introduced a new famiiy of brighteners of which compounds (5)-(8) are examples. The compounds are effective for alkaline non-cyanide baths as well as for the cyanide baths. Another patent14 added a range of aromatic betaines to the list of effective compounds.

oco2Et 0""" Q""'

C1- CH2C02Me

CI- C H 2 C q M e

C1- CH2Ph

(7,

(6)

(8)

Probably the most widely used compound is 1-benzylpyridinium-3-carboxylate(9). It is noted that baths containing this material require a 'working in' period before deposits reach maximum brightness. This correlates with the fact that a new absorption maximum at 360 nm grows when (9) is added to alkali. The rate of growth of the maximum is independent of zinc concentration and the rate law isI5 rate =

GbS[ OH-]2

[(9)12

(18)

Change of the quaternizing group may lead to a slower rate of growth of the 360 nm maximum, and also in practice to a poorer performance as a brightener. Thus it is clear that (9) is the precursor of the true brightening agent. There is good evidence to suggest that the mechanism shown in equations (19) and (20) is operative. Whereas (9) was unlikely to be an effective ligand for zinc(I1) in the plating medium, the so-called bimolecular ether (11) is in fact a potential terdentate ligand (N,N,O or O,O,O).Compound (11) is diMicult to isolate from solution in a pure form but model compounds (O,O,O) or (O,S,O) such as (12) and (13) have been shown to function as ligands to zinc(I1) and other transition metal ions.16However, complexes of (12) and (13), which in the solid state are Zn(L-2H) 1.5H20and Zn(L-H)(C104) (L = 12 and L = 13), decompose on contact with strong alkali. Thus it is improbable that a significant quantity of brightening agent is complexed in the bulk solution (see also Section 57.3.2.6(ii)).

r-

7

Electrochemical Applications

( J c o H pH02C(CH&E(CH&CO,H

CHzPh

0

(9)

(12)

aiH +-

(9)

+

OH-

',

(13) E=O, S

+

__+

CH2Ph

H20

(20)

CH2Ph (11)

The brightening agents may also undergo electrochemical reaction as well as chemical reaction. Thus compounds related to (7) may be electrochemicallyreduced as shown in equation (21). Thus, dihydropyridine species such as (14) may also be present in the plating bath. -e-

Me

J

\-e-

(14)

In view of the fact that species such as (11) and (14) will in any event be derived from (91, it was interesting to note a patent claimI4 for (11) as a new brightener for zinc plating. The claim is doubtless correct, but it is superfluous. The above examples, which relate to but few of the many possible addition reagents in use, indicate that chemically the systems may be quite complex. The additives are certainly active at the electrode-solution interface and indeed many that are used effectively are known to be surface active. The role of coordination chemistry in the detailed mechanism of action remains somewhat obscure, not so much because the appropriate questionshave not been answered, but rather because they have rarely been posed. 57.2.5

Electrodeposition of Specific Metals

It is not relevant within the context of this chapter to detail the technical aspects of the electrodeposition of all possibIe metals and alloys. Such information is available in a number of specialist or handbook^.'^ However, it is pertinent to illustrate more specifically than in the above sections the role of coordination compounds in particular cases. 57.2.5.1

Chromium

The point of major interest is that for the most part chromium plating occurs from a chromium(V1) bath (Section 57.2.3).20 This bath has a low current efficiency, gives off both hydrogen (cathode) and oxygen (anode) and this vigorous evolution of gas causes a mist which creates a health hazard requiring appropriate precautions. Generally a lead anode is used since this is insoluble in the acid medium and will allow the anodic oxidation of chromium(II1) to chromate(V1). This obviously requires that the chromium must be replenished as GO3. The mechanism of chromium plating has been touched upon earlier (Section 57.2.3) and the role of chromium(II1) sulfato complexes was mentioned. Interestingly,in addition to the disadvantageous

8

Electrochemical Applications

role of chromium(II1) in creating a cathode film, the presence of some chromium(I1I) is required to give the bath an adequate throwing power. The previous remarks can be expanded to include three developments which are designed to overcome some of the unattractive features of the process. (i) Tetrachromate baths It is known that trichromates [Cr3010I2- and tetrachromates [Cr4013]*- m a y crystallize from strongly acid solutions as their alkali metal salts. Tetrachromates are believed to be major constituents of a plating bath for which the following composition is typical: Cr03 300 g I-’ NaOH 6 0 g 1-1

H2S04 0.75 g 1-I Ethanol I cm3 1-1

This ‘tetrachromate’ bath is characterized by a low operating temperature of 16-22 “C and by an excellent throwing power. Although a high current density requires the bath be cooled, the current efficiency (30-37%) is high by chromium plating standards. The deposit tends to be matt and softish and hence polishing is required for a bright plate and the process wiIl not be suitable if a hard plate is required. (ii) Triualent chromium plating Since the reduction of chromate(V1) to metal is a six-electron process, it would seem logical to plate the metal from an electrolyte based on a lower oxidation state of chromium. A good deal of work has been done on the development of trivalent chromium plating,* but it is only recently that a commercial process has been available in the United Kingdom. The process works with carbon anodes and at a cathode current density of 10 A dm-’. The electrolyte is contained in a plastic or rubber lined bath and air agitation is required. Although a chromium ‘complex’ is involved, full details of the electrolyte composition are not available in the journal paper reporting the successful development of the new process.20 Chromium is consumed as the basic sulfate and the ligand is described as inorganic. Conductivity salts are required to improve the conductivity of the solution and some boric acid is present as a buffer. An earlier American process21is equally vague as to details, but chromium(II1) chloride complexed with an ‘organic’ ligand in the presence of a ‘secondary organic’ (1% v/v> is said to give good results under laboratory conditions. Some chlorine evolution at the graphite anode was observed but one respect in which the process was attractive was that pollution problems seemed generally less severe than with a chrornate(V1) bath. One reason is the lower toxicity of chromiurn( 111). Although the process never translated successfully to industry, more chemical details are available for a system based on aqueous DMF.u A typical composition, with quantities in g 1-’ is: Crrrr(as CrCly6H20) NH4+ (as NH4Cl) Na+ (as NaCl)

Boric acid DMF

43 10 20

2 400

The cathode efficiency is 10% (evolution of dihydrogen), 60% ( C P --+ Cr“) and 30% for plating chromium metal. The pH remains constant at 1.4; this is attributed by the authors to the processes shown in equations (22) and (23). However, given an apparent overall mole ratio of Cr:C1 of 1:4.7 and the electrochemical generation of CrlI on which substitution reactions are rapid, it seems that chloro complexes of chromium(II1) must be present, in which case equations (22) and(23) may represent an over simplification. 2[Cr(H2O)d3+

e

2[Cr(0H)(H2O)J2+

e

/O\ [(H20)4Cr

Cr(H2014j”

(22)

3H20

(23)

‘O/

2[cr(H20)b]2+

+

2H@+

+ [o]

__f

2[Cr(H20),]3+

+

One disadvantage of the chromium(II1) process now in operation is that only a limited thickness of plate is possible even at high current density. Indeed, a maximum deposit of just under 0.5 pm is achieved in 3 minutes at 10 A dm-2. Higher current densities actually result in a thinner deposit and only densities in excess of 40 Adm-’ restore the thickness achieved at 10 A drn-2. The contrast with the chromate(VI) process is shown in Figure 3. (iii) Brush plating

Some special instances arise when deposition in an electrodeposition bath would be inappropriate. Small areas c m be plated by making a cathodic connection and touching the area with a pad soaked in an electrolyteand in which is located an insoluble anode (often graphite).*A popular

~

Electrochemical Appkations

9 Chromium(V1)

/ C h ro m ium

(m1

Current density (A drn-?)

Figure 3 Comparison of thickness of plate deposited in three minutes as a function of current density

choice of electrolyte is the ammonium salt of tris{oxalato)chromate(III),(NH4)3[Crrr1(C,04)3], dissolved in methanol or in hydroxylamine. 57.2.5.2

Copper

Three major electrodeposition media for copper may be identified. All involve coordination compounds. (i) Cyanide copper baths

Cyanide copper baths are of importance for the copper plating of ferrous metals and zinc die castings, often to give pratection prior to the deposition of a second metal such as nickel or chromium from a medium (acid) in which the base metal would react. Three baths may be identified: ‘Strike’, ‘Rochelle’ and ‘High Efficiency’. The Strike bath is valuable for the deposition of thin coatings. The Rochelle bath is robust and generally free from interference by foreign substances and is used for coatings of intermediate thickness. The High Efficiency bath, for which current efficiency can approach loo%, is used for thick coatings but is sensitive to impurities. Typical compositions are given in Table 1. Table 1 Cyanide Copper Bath Compositionsa

cum NaCN or KCN Na2C03

15 (0.168 M) 23 (0.469 M) -

26 (0.290 M) 35 (0.714 M) -

75 (0.838 M) 93 (1.898 M) 115 (1.769 M)

15

30

a

NaOH or KOH

30 42

Rochelle salt KNaCfi&4H20 Cu:CN a

-

45

Optional

3.19

(0.160 M) 3.46

3.26 (Na)

These baths contain adventitious carbonate v i a absorption of C 0 2 from the air and via the anodic oxidative decomposition of cyanide.

Cyanide above that required for a copper:cyanide ratio of 1:3 is termed free cyanide and it is generally assumed that the major cyano complex in solution is [Cu(CN),I2-. However, this ion was first crystallographically characterized in Na2[Cu(CCN)3].3H20only quite recently.23 Whilst it is certainly possibie that the tricyanocuprate(1)ion is present at lower cyanide concentrations, when a significant excess of the ligand is present it is likely that the tetracyanocuprate(1)is present. However, as seen in Table 1 most baths typically operate with a mole ratio of copper to cyanide between 1:3 and 1:4. The sodium carbonate provides a buffering effect @H 10.3) in the Strike and Rochelle baths. The mechanism of deposition is not well understood. The ion [Cu(CN)d- has been implicated but this may be subject to the same objections discussed in Section 57.2.3 when the deposition of zinc was considered. More free cyanide will polarize the cathode with the result that hydrogen

10

Electrochemical Applications

is evolved at the cathode in the Strike and Rochelle baths, although this provides extra cleaning. The free cyanide is also required for the corrosion of the copper anode. Potassium sodium tartrate (Rochelle salt) is believed to play a complex forming role at both anode and cathode. Thus, the tartrate may complex with the initial products of electrolysis in the anode film and it may also form short-lived complexes in the Cathode film.

(ii) Acid copper baths The history of plating copper from acid baths may be traced back to 1810.' The relatively low running costs of the baths make them attractive for electroforming of copper articles and for the electrorefining of copper metal, as well as in standard plating applications. The solutions may be based on copper(I1) sulfate/sulfuric acid or, of more recent introduction, copper(I1) tetrafluoroborate/fluoroboric acid. The latter solutions may be operated at extremely high current densities and hence are very suitable when heavy deposits are required in a short time; consequently they tend to contain a greater concentration (range 0.94-1.89 M) of copper than do the sulfate baths (range 0.60- 1.00 M). For many applications, addition reagents are required. A bewildering range has been employed for both levelling and brightening; some of the many that are effective include a combination of poly(viny1 alcohol) and glue; amino acids; thiourea; and benzotriazole. As discussed in Section 57.2.4, it is likely that, in many cases, complex formation in the cathode region has some role to play in the mode of action of these reagents. In this context it is interesting that some of the addition agents might be expected to be effective ligands for copper(1). Indeed in one study," a copper(1)-lxnzotriazole complex was actually identified as an inclusion in the electrodeposit obtained from a benzotriazole-containingcopper(I1) sulfate acid bath.

(iii) eropkosphate copper baths The advantages of this system are that it presents a much lesser effluent problem than the cyanide bath and, operating around neutral pH, is less corrosive than the acid bath. It has a good throwing power but some metals, e.g. steel and zinc, need to be flash plated in a cyanide bath before plating in a pyrophosphate solution is possible. The bath in commercial use is based on potassium bis(pyrophosphato)cuprate(II), which may in fact be isolated as the hexahydrate K,[CU(P,O~)~].~H~O. Ammonia is generally present, as is a source of nitrate, The ammonia is said to assist in anode corrosion and to contribute to brightening the cathode deposit; nitrate inhibits the reduction of hydrogen ions. 5 7.2.5.3

Nickel

Nickel is the most widely used electrodeposited metal. It is usually used in connection with chromium when the nickel deposit is followed by one of chromium, the finish being known as 'bright chromium plate'. The basic bath has altered little since Watts introduced high-speed nickel plating in 1916. The bath is based on nickel@) sulfate, nickel(II) chloride and boric acid. Whilst this offers little scope for novel coordination chemistry, the importance of nickel plating has ensured that more work than usual has been performed on addition agents25and some aspects of this might be profitably considered. The reduction of coumarin (1)mentioned in Section 57.2.4 is typical of the fate of a group of levellers used in nickel plating which have in common the groups (15). It has been recognized that the depositing nickel may function as a hydrogenation catalyst and that this would certainly imply adsorption of the alkene on the nickel surface. The vapour phase hydrogenation of ethylene catalyzed by nickel is said to be most rapid when nickel films have the (1 10) planes preferentially exposed,26which may imply that the alkene function shows preferential adsorption on some planes. 0

II (1s)

The levelling action is also dependent on pH. At lower pH (1 S),greater reduction of (15) occurs, e.g. to alcohol, and levelling power falls off. At pH 4-5 the alkene function is reduced and a thin film of a hydroxonickel species is seen on the cathode. The leveller and the inorganic film seem to function synergistically to control adsorption of the leveller which, in turn, may influence the development of the nickel surface. Thus again coordination, on this occasion to afford an organometallic species, is implicated but the details of mechanism remain vague and speculative.

Electrochemical Applications

11

Sulfur compounds, e.g. arenesulfonicacids, used in conjunction with the unsaturated compounds mentioned above can provide level, bright films. After several hours of operation of the bath, the odour of the aromatic hydrocarbon can be detected and it is well known that the bright deposits contain sulfur. Thus reduction of organic species at the developing nickel surface is a general phenomenon.

57.2.5.4 Precious metals Those aspects of the plating of silver, gold and the platinum metals which specifically involve coordination compounds are now briefly considered. (i) Silver and gold Silver and gold were probably the first metals to be e l e ~ t r o p l a t e dand ~ ~ were the subject of a patent which could be said to have started the electroplating industry.28Both metals are normally plated from cyanide baths and the processes have enjoyed some prominence since World War I1 as electronics technology'has developed and the good electrical conductivity and chemical resistance of the metals has been attractive. The silver bath normally operates with a mole ratio of AgCN between 1:3 and 1:3.27. Potassium carbonate and hydroxide are also present. The excess or 'free' cyanide ensures formation of the dicyanoargenate(1) ion [Ag(CN)J and prevents precipitation of the sparingly soluble silver(1) cyanide. Generally, a commercial preference exists to add the 'free' cyanide as KCN rather than NaCN. Many addition reagents have been used to obtain bright silver plate. Most of the successful compounds contain sulfur; indeed, the use of carbon disulfide in 1847 was probably the first recorded instance of an addition reagent improving a deposit. Chemically the requirements for gold plating are similar. For baths containing a greater concentration of gold the more soluble potassium dicyanoaurate(I), K[Au(CN)~],is preferred over the sodium salt. The [Au(CN);I- ion is stable over a wide range of pH; below pH 3, however, AuCN precipitates.The pH is normally maintained at less than 11.8 with carbonate and phosphate buffers, which also improve the solution conductivity. Some baths for flash plating of gold are prepared from solutions of the metal in aqua regia and contain a low concentration of free cyanide. The possibility that tetracyanoaurate(III),[Au"'(CN)J, is present in such solutions has been discussed.29 (ii) Platinum metals

Only three of the metals, rhodium, palladium and platinum, need be considered. Of these only rhodium plating is of significant commercial importance, but the relatively low cost of palladium has made it attractive for contacts and printed circuits. Electrodeposited platinum is harder than the annealed bulk metal and finds applications in jewellery, the plating of scientific instruments, standard weights and parts of electrical apparatus. Rhodium is normally plated from a sulfate or phosphate bath. It is palladium and platinum that give more scope for some classic coordination compounds. Palladium has been successfully plated from baths after introduction as the tetraamminepalladium(I1) cation l?d(NH3)J2+, as the nitrite, nitrate or chloride salts. In the latter case, dichlorodiamminepalldium(I1) is added to a bath containing ammonium salts (chloride, sulfate and carbonate) and free ammonia. Baths based on [Pd(NO2),I2-/NaCl and [pdC14]2-/NH&1 have also been used. Platinum is generally introduced as dinitrodiammineplatinum(II), [Pt(NH&(N02)d; the solution also contains ammonium nitrate and sodium nitrite. Alternatively, H2PtC16or Na2Ft(oH)6], both derivatives of platinum(IV), have been used. 57.2.5.5

TiR

Although relatively soft and expensive, tin has a number of properties such as low toxicity and good resistance to corrosion that make it an attractive plating metal. Since the melting point of the metal is 232 "C, dipping in molten tin is an alternative which has been employed over the centuries, particularly for the tinning of copper cooking utensils. However, little control exists over the thickness of the coating. The formation of steel strip by continuous cold rolling required a process capable of tinning the strip. This is now achieved by electrodepositingtin at speeds as high as 170-600 rn min-l using current densities of 20- 100 A drrP2.

Electrochemical Applications

12

Two common alternatives are available for electrodepositing tin:30 alkaline and acidic baths. The alkaline bath has good throwing power but consumes more power than the acid bath. Tin is the bath being approximately 0.25 M present in the alkaline bath as stannate(IV), in ‘free’ hydroxide ion (PH 13.4). The hydroxide ion is the principal charge carrier. Potassium is superior to sodium as the counter ion (greater ionic mobility) but economic factors lead to the continued use of sodium in many plants. The hydroxide ions, acting as a sink for dissolved C 0 2 , also prevent two undesirable reactions (equations 24 and 25).

-

SnO,

[Sn(OH)6]*[SII(OH)~]~- + CO,

-+

20HSnOz

+ 2HzO + C0:- +

(24

3H@

(25)

Hydrogen peroxide is present to oxidize any tin(I1) to tin(1V). The presence of stannate(II), [Sn(OH)3]-, causes a spongy non-adherent plate, possibly partly as a result of a disproportionation (equation 26). 2[Sn(OH)3]-

Sn

+

+

[SIl(OH)6]2-

(26)

Most interesting is the behaviour of the tin anode. If the anode current density is too low, the tin will dissolve in the alkaline medium as stannate(I1) with disastrous results for the plating. If the current density is high, the anode is rendered passive and dioxygen is evolved on electrolysis and the tin must be replenished as stannatew). At intermediate current densities of 1-2 A dm-*, the anode assumes a greenish yellow film and the, tin enters solution as stannate(1V). The deposits from the alkaline bath are fine grained, pure, but dull and somewhat porous. If bright non-porous films are required, the plated article is treated in a process known as ‘flow brightening’ in which it is taken through a bath of palm oil operating at temperatures in excess of 232 “C, the melting point of tin. The problem of obtaining bright tin deposits direct from the plating bath proved difficult to solve, but it is now possible using acid baths in which organic brightening agents act synergistically with surface active agents. The acid of choice is sulfuric acid (-1.33 M) and the metal is present as tin(I1) sulfate (-0- 19 M), although the fluoroborates [HBF4/Sn(BF4)dprovide an alternative choice. Of more interest to coordination chemists is the Du Pont halogen tin process which is based on halo complexes of tin(I1). The presence of fluoride is considered to form [SnFJ, but much chloride is present to enhance the conductivity. Inhibitors are required to control the oxidation of tin(I1) to tinw)and the consequent separation of the sparingly soluble Na*[SnFd] as a sludge. The process was originally introduced in 1942 for the.continuous electrotinning of steel strip. 57.2.5.6

Zinc

Examples in earlier sections have been taken from zinc plating (Sections 57.2.3, 57.2.4); hence it is appropriate that Zinc should be the last of the metals selected for individual consideration in this chapter. Four different processes are available for electrodepositing zinc3l: (a) high cyanide-high metal (HCHM); (b) low cyanide-low metal (LCLM); (c) alkaline non-cyanide (ANC); and (d) acid (pH 5 ) (ACID). Typical compositions are given in Table 2. Table 2 Composition of Zinc Plating Baths (g 1-I) .- -

.

Zinc ~~

HCHM

LCLM ANC ACID

. NaCN ~~

30-50 7-15 12-18 35-55

75-150 10-30 -

NaOH

Chloride

75-150 80-120 115-145

-

~~

-

140-200

(i) High cyanide-high metal The electrolytein these baths is robust and the throwing power of the bath is excellent; however, current efficiency falls with increasing current density as hydrogen evolution increases. The bath also presents significant effluent disposal problems since cyanide must be destroyed by chlorine or hypochlorite oxidation, thus adding to the capital costs of the plant. The Zn:CN ratio may vary between 1:3.3 and 1:4; however, the Zn:OH ratio is in the range 1:4 to 1:5 and thus the electrolyte will contain both cyano and hydroxo complexes. Indeed, one estimate32suggests that as much as 75-90% of the zinc is present as [Zn(OHJ4]’ and the rest as

13

Electrochemical Applications

[Zn(CN>$. It is believed that the mechanism of deposition involves the hydroxo c o m p I e ~ in ,~~ which case Section 57.2.3 should be consulted. (ii) Low cyanide-low metal

The main advantage offered by this bath is the reduction of the cost of effluent treatment;34 however, the low metal content means that stricter day to day analytical control of the bath contents is required. Thus the process is probably viable only if laboratory facilities are available in the same plant. To some extent the disadvantages may be overcome by a compromise move to a medium cyanide-medium metal bath. The 2n:CN ratio is 1:2, but the Zn:OH ratio is 1:20; hence, the hydroxo- rather than the cyano-zincate will be the major species in the electrolyte.

(iii) Alkaline non-cyanide A natural deveiopment was to remove cyanide altogether from the bath. However, until recently, the bright ductile coatings which were characteristic of the cyanide processes could not be reproduced in the alkaline cyanide-free bath. The development of brighteners such as those discussed in Section 57.2.4 has now made this a viable process which avoids major effluent treatment problems. The current efficiency falls off with current density more rapidly than for the high cyanide-high metal bath, but in this respect it is no worse than the low cyanide bath. (iv) Acid zinc plating Acid baths, so called because they operate at a pH of about 5, offer a virtually 100% current efficiency because of the high hydrogen overvoltage; however, until recently they offered poor throwing power and a matt deposit. The development of new brighteners, e.g. a mixture of a poly(ethy1ene glycol) with a ketone35or 3-substituted pyridinium now make this a competitive means of producing decorative zinc plate. Zinc is present in the electrolyte as the aquated cation and chloride is present as a charge carrier. It is interesting to note37an early process which was based on ammonia/ammonium chloride in which the electrolyte contained the tetraammine [Zn(NH3),J2+.This solution never found general acceptance, although it has been used for wire plating. 57.2.5.7

Alloy pkting

The values of the standard electrode potentials of copper (Cu+ + e- -+ Cu, 0.522 V) and zinc (Zn2++ 2e- + Zn, -0.76 V) make it appear unlikely that the metals could be codeposited as the alloy brass. However, if, for example, solutions which are 0.025 M in [Zn(CN)4]2-and 0.05 M in [Cu(CN),I2- are mixed, simple calculation can show that the static electrode potentials of the two ions have values which approach each other more closely and codeposition becomes much more probable. This can be understood with the help of equation ( 5 ) and the knowledge that the dissociation constants of [Zn(CN),I2- (1.3 x lo-'') and [CU(CN)~]-(5.6 x will greatly modify the activity term. Typical bath constituents for alloy plating are given in Table 3, which also gives some impression of the variety of alloys which may be electrodeposited. Table 3 Electrodeposition of Alloys: Bath Composition

AIioys

Ni-Co, Cu-Zn, Cu-Sn, Pb-Sn,

Ni-Zn, Ni-Fe Ag-Pb, Ag-Cd, gold alloys Cd-Sn

Bath constituents

AIioys

Baih cunsiiiuenrs

Sulfate, chloride Cyanide

Ag, Au, Pt alloys Cu-Sn Ni-cu In-Pb

Complex halides Cyanide, pyrophosphate Sulfate, citrate Sulfamate

BF4-

The use of mixed complex baths is interesting since the concentration of one free metal ion may be altered by varying the amount of one ligand. Thus, in a copper-tin bath, cyanide content may be varied to alter the activity of copper ions, with little or no effect on tin which is present as stannate or as a pyrophosphate complex. It is evident that some knowledge of tbe coordination chemistry will reduce the degree of empiricism in developing alloy plating baths. The electrodeposited alloys are alloys in the true metallurgical sense, showing a phase structure indicated by the appropriate phase diagram to be stable at the temperature of electrodeposition; sometimes some interesting differences in behaviour are noted. Thus, for example, cast alloys of

14

Electrochemical Applications

nickel and tin (Ni:Sn = 1:l) contain equal mole fractions of Ni3Sn2and Ni3Snq whereas the electrodeposited alloy is the single intermetallic compound NiSn. The NiSn is deposited from a nickel(I1) chloride-tin(I1) chloride bath containing NH4HF2;since the metals are deposited in an accurate 1:1 ratio the involvement of a species such as NiSnF4 has been suggested. It has also been proposed that fluoro-bridged complexes are present and the possibility of trichloro- or trifluoro-stannate(1I) complexes has been considered, X3Sn-+Ni1l.A laser Raman study of the solution failed to reveal evidence of species other than aquonickel(I1) species and fluorotin(X1)species3* 57.2.5.8

Miscellaneous ligands

( i ) Ammonia

Deposits fmm solutions of ammine complexes are often good. Two specific examples have been mentioned, namely platinum (Section 57.2.5.4(ii)) and zinc (Section 57.2.5.6). However, such processes have found no large-scale applications. (is) Halides Halides, particularly the less expensive chloride, are often found in plating baths. When a concentrated solution of metal is required, the chloride salt often proves more soluble than the sulfate. The free chloride ions often carry much of the current. Deposition from true halo complexes can occur; thus good deposits have been reported from [AuC14]- and [AgId-. (iii) EDTA and related ligands Ethylenediaminetetraacetic acid (EDTA) has found a variety of uses; not all have been successful. EDTA can be a useful additive to an acid c o b r bath since it can offer some control over dissolution of the anode. In high-efficiency cyanide copper plating, the addition of EDTA can counter the harmful effects of chromate(V1). Et is speculated that the chelating agent facilitates the reduction of Crvl to CrlI1 by CUI. Iron, which has not been specifically dealt with in t h s chapter, may be deposited in thin films from alkaline media containing complexes of the metal with EDTA or with triethanolamine. The effluent disposal problems presented by cyanide baths, together with stricter environmental legislation, have prompted much work to remove cyanide from such systems whilst retaining the good throwing power of cyanide baths and the brightness of the deposits. One approach has been to substitute other ligands for cyanide and EDTA has been a popular choice. Thus, for example, gold will deposit from a gold-EDTA, cyanide-free bath. A variation of this theme has been the use of related ligands such as ethylenediaminetetraphosphoric acid (16).39Thus (16) and related compounds have been claimed to give satisfactory deposits from cyanide-free, neutral (pH = 71, solutions of copper, nickel, cadmium, iron, cobalt and Zinc. Unfortunately, in removing the cyanide problem, ligands of this type create another in that it becomes dificult to free eMuent from excess metal. This is well illustrated by a simple example. A solution containing only [Zn(OH),]Z- is effectively cleared of zinc, as insoluble Zn(OH)?, by adjusting the pH to between 8.5 and 9.0. In the presence of EDTA, no zinc precipitates. Thus chelating agents of this type are unsuitable substitutes for cyanide. By contrast, triethanolamine IS very satisfactory from the viewpoint of effluent technology. 0

57.2.6

Concluding Remarks

The coverage of electrodeposition of metals in this chapter has attempted to identify those aspects of the topic in which coordination compounds have a role. The coverage is not comprehensive in terms of metals that can be electrodeposited, but the examples chosen give a fair

Electrochemical Applications

15

indication of the variety of conditions encountered in-terms of pH range, addition reagents and variety of complex species involved. Inevitably, from the viewpoint of the coordination chemist, there is an impression of qualitativeness. In one sense t h s feeling is understandable since most of the interest in the topic has been metallurgical. After all, commercially speaking, it is the quality and properties of the electrodeposited film that matter rather more than the chemical details of its origin. Some fundamental chemical studies have been carried out but much more work could be done. With a strong resurgence of interest in electrode surface phenomena, more studies of a quantitative nature will doubtless be available in the near future.

57.3 COORDINATION COMPOUNDS AND ELECTRODE PHENOMENA: SURFACE MODIFIED ELECTRODES 57.3.1

Introduction

The second topic of this chapter is the role of coordination compounds in advancing electrochemical objectives, particularly in the sphere of chemically modified electrodes. This involves the modification of the surface of a metallic or semiconductor electrode, sometimes by chemical reaction with surface groups and sometimes by adsorption. The attached substrate m a y be able to ligate, or it may be able to accept by exchange some electroactive species. Possibly some poetic licence will be allowed in defining such species since many interesting data have been obtained with ferrocene derivatives; thus these organometallic compounds will be considered coordination compounds for the purpose of this chapter. A number of review articles have recently appeared which give a good summary of the present state of progress. No attempt is made to duplicate the coverage of those articles, nor is the coverage given claimed to be comprehensive in the sense that one would expect of a review article; rather an overview of the role of coordination compounds in a topical and exciting area of research is offered. Since most work relates to surface modification, the section is conveniently subdivided into sections dealing with different surface materials. Intigrated into these subsections will be details of the various methods by means of which the surface is modified. 57.3.2

Surface Modified Electrodes

Nafion modifid eiectrories Nafion (17) is a perfluorinated polymer related to teflan (polytetrafluoroethylene). An electrode is conveniently coated by allowing an ethanolic solution of the polymer to evaporate. The film produced is stable, rather more so in fact than other polymer films, e.g. polyvinylpyridine (see Section 57.3.2.2). At the microscopic level the polymer separates into two phases, the bulk polymer and the lower density ionic cluster phase. Diffusion of ions can occur quite freely; for example, the diffusion coefficient of Na' in Nafion (MW 1200) is only slightly less than in water.44

57.3.2.1

-(CF&Fdx(CFCF&-

I

O--(C~P~)-O-(CF~CF2)SO~- Na+ (17)

The polymer coated electrode may be doped with an electroactive species by exposing it to a dilute solution of the chosen material. A good example is [Ru(bipy)J2+(bipy = 2,2'-bipyridyl), which can be exchanged from a solution of the ruthenium complex in sulfuric acid. It is observed that the value of E" for the [R~(bipy)~]%+ couple is the same as the aqueous solution value. Also the loading of the polymer can be such that the local surface concentration of the electroactive complex is greater than that in the solution from which it is exchanged; thus larger currents are observed than with the bare electrode under the same conditions. The diffusion of the electroactive ions is both physical and due to electron transfer reactions.45 The occurrence of either or both mechanisms is a function of the electroactive species present. It has been observed that the detailed electrochemica1 behaviour of the electroactive species often deviates from the ideal thin film khaviour. For example, for an ideal nernstian reaction under Langmuir isotherm conditions there should be no splitting between the anodic and cathodic peaks in the cyclic voltammogram; further, for a one-electroncharge at 25 "Cthe width at half peak height should be 90.6 mV.4 In practice a difference between anodic and cathodic potentials may be finite even at slow scan rates. This arises from kinetic effects of phase formation and of interconversion between different forms of the polymer-confined electroactive molecules with different standard potentials.&

16

Electrochemical Applications

An interesting example of the type of chemistry that can occur with a Nation modified electrode is provided by a pyrolytic graphite-Nation electrode exchanged with [ R u ( b i ~ y ) ~which ] ~ ' is found to catalyze some oxidation reactions. An orange emission is noted at the electrode when the electrolyte is aqueous oxalate at p p 6 (electrogenerated chemiluminescence - ec1).47,48Oxalate penetrates the polymer layer and is oxidized by the electrochemically generated [R~(bipy)3]~+, the following mechanism being proposed (equation 27; L = bipy).

-

RUL~,+ RUL33+$.

c,o,2c,o4-

R u L ~ ~ + C02? +

+

R u L , ~ + + e-

RUL;+

+ c204-

co, + c0,co, + RUL,*~* c q + RUL?'

R u L ~ ~ * +CO,' + RUL,'~+ RuL?' + hv RUL?++ coz7 RuL3+ + CO, RuL: + RuL,~' --+ RuL302+ + RUL?'

The ecl arises from the reaction of the radical anion COZT with [R~(bipy)~]~'. It was noted that some quenching of [Ru(bip~)3]*~+ by the ruthenium(II1)complex was possible. Interestingly, water promotes a non-radiative decay of [R~(bipy)~]~'. It seems reasonable that which helps within the Nafion film the ruthenium(I1) species minimizes its contact with to explain the abiIity of the film to retain relatively hydrophobic counter ions. The more hydrophilic [ R ~ ( b i p y ) ~is] ~in+fact more labile from the film than the bipositive complex.44

57.3.2.2 Poly-kvinylpyridine and related pofymers Nafion was rendered electroactive by cation exchange of a redox centre. Other polymers such as poly-4-vinylpyridine (PVP) may be protonated or quaternized to provide the possibility of introducing a redox anion such as [IrCl6I3- or [Fe(CN>613-.50?51 Alternatively, the pendant pyridyl groups on the neutral polymer may function as ligands and species such as [Ru(bipy)2I2' may be polymer bound. ( i ) Attachment of polymers to electrode surfaces

The polymer may be preformed or, alternatively, the monomer may be polymerized on the electrode surface.In the case of PVP, preformation of the polymer by radical initiated polymerization of 4-vinylpyridine leads to (18), but reaction with dry hydrogen chloride gives the ionene form (19) which isalso an anion exchanger.

Q (19) Ionene polymer

A process commonly used to apply PVP to an electrode surface is to allow a solution of the polymer in a solvent such as methanol to evaporate on the surface. The redox active species can be applied by exposing the modified electrode to a solution of the appropriate complex. Alternatively the PVP-redox complex may be formed prior to coating the electrode. A more sophisticated variation of the above technique is spin casting. A small quantity of polymer solution is dropped on to an electrode, e.g. platinum, which is then rotated at thousands of rotations per minute. Electropolymerization is useful and has been successfully applied to 4-vinylpyridine complexes and to 4-methyl-4'-~inyl-2,2'-bipyridyl.~~ Vinylferrocene (vide infra) has been polymerized on to platinum, glassy carbon and titanium dioxide electrodes by introduction to a radiofrequency argon plasma discharge. Electropolymerization and plasma polymerization are likely to be of value to produce copolymers on electrode surfaces.

17

Electrochemical AppIica tions

The techniques described may provide polymer matrices of many monolayers which will generally provide large electrochemical responses since many layers of redox sites can react. The polymer adheres to the electrode surface by adsorption or by covalent, possibly in the case of PVP coordinate, bonding. Little work has been done with PVP copolymers (that with PVP-lysine is an exception), but it is of interest to note that polymerization with 2% &vinylbenzene will produce an insoluble polymer with no impairment of coordinating ability. (ii) Protonated and quaternized PVP electrode coatings Protonated or quaternized poly-4-vinylpyridine will function as an anion exchanger; hence ion exchange with anions of interest in redox processes is an easy and obvious way of rendering the polymer redox active. Thus, for example, a graphite disc electrode coated with protonated PVP and exchanged-withhexachloroiridate(Iv), [IrCl6I2-,has been used to effect the catalytic oxidation of aqueous iron(I11).50 Some leaching of the redox active anions may occur unless the electrode is in contact with a solution containing the same redox species. The protonated and quaternized PVP will generally swell in contact with solvents, implying ingress of solvent to the polymer matrix. The matrix is a multilayer film, yet redox processes occur with great kinetic facility. This raises the question of how redox centres within the polymer but remote from the electrode-polymer interface can be involved in the necessary oxidation-reduction cycle. One possibility is the migration of the centre through the matrix to the electrode surface. Another possibility is a process of 'electron hopping' between adjacent redox centres. This is illustrated schematically in Figure 4. A centre adjacent to the electrode may be oxidized at an appropriate potential; this will then be reduced by an adjacent centre, and in this way the process will influence chemical events at the polymer/solution interface. In some cases a mixture of diffusion and electron hopping may be involved.43

+,----, e-

Electrode

-OX

REO-OX

RED-

OX

Figure4 Schematic illustration of 'electron hopping' mode of charge transport in a redox polymer

When an exchanged film is in contact with a solution of the Same redox species, the selectivity of the ion exchange medium may render the concentratim ratios of reduced to oxidized forms in the polymer film different from corresponding solution values. This may be illustrated by consideration of glassy carbon electrodes covered with films of PVP quaternized with alkyl bromides of varying carbon chain length and exchanged with [Fe(CN)6]x-. Accumulation of the iron(II1) for transfer complex within the polymer matrix is favoured and estimates of the molar work (WT) of the the redox ion from the solution to the membrane phase have been made in the presence of potassium trifluoroacetate." The value of WTKdequates to the free energy of dissociation of the ion pair K'[Fe(CN),I4- (AG = -13 kJ mol-I). The Ionger the alkyl chain, the greater the preference for hexacyanoferrate(III), negative values of WTox implying ion pair formation, Q+.[Fe(CN)&, within the polymer matrix. For a 1Zcarbon chain the positive value of Vd (13 kJ mol-') implies that total dissociation of the aqueous phase ion pair is required before the ironflI) species can enter the matrix. However, for short carbon chains, which give much better solvated membranes, values of WTrd can become negative, implying that the iron(I1) enters these membranes as K+[Fe(CN),I4-. (iii) Redox centres ligated to PVP electrode coatings

Poly-4-vinylpyridine is a versatile polymer since in the non-protonated form it may function as a polymeric nitrogen-donor ligand. Thus, redox centres may be anchored by coordination to pendant 4-pyridyl groups on the polymer chains. PVP may react with dichIorobis(2,2'-bipyridyi)ruthenium(II) and the precompiexed polymer may then be used to dip coat various electrode^.^^ EDTA complexes of ruthenium(II1) will react with PVP and may thereby be immobilized on electrodes. The use of transparent graphite electrodes facilitates the spectroscopic monitoring of both the quantity of PVP and ruthenium on the electrode. Also the Ru"' Rdlreduction may be followed as it proceeds. The UV spectra of the immobilized EDTA complexes are similar to those in solution. It has been possible to use a band --.)

Electrochemical Applications

18

at 290 nm to monitor the rate of loss of the Ru"'/EDTA complex from the polymer and a rate constant of 4 x lo4 s-l for the breaking of the 4-pyridyl-ruthenium(III) coordinate bond.53 Further variations on the theme have been achieved% by anchoring species such as [RuIVO(terpy)(py)]2+or complexes of Iron complexes have also been studied; for example, evaporation of a solution containing [Fe(CN),(H20)l3- and PVP on to an electrode will immobilize the pentacyanoferrate as a pyridyl complex, one in three available pyridyl groups being used to avoid precipitation prior to evaporation of solvent. The transport of charge by electron hopping is an attractive model for these systems. In the case mentioned above, the electrode response is better from the precomplexed polymer film than from one prepared by first coating with PVP, then dpping into a solution containing a source of [Fe(CN),(H2O)l3-;thus the spatial distribution of redox centres is important as well as their number in determining eIectrode response. Data for the pentacyanoferrate system support charge transport via adjacent redox sites and the rate of this transport falls off rapidly below a critical concentration of centres.56 The polymer must be able to accommodate geometric changes demanded by the metal ion as the oxidation state changes. Actually, since much work is done with Fe"/Fe"', Rd1/Ru1'' and O S ~ ~ / Owhere S ~ ~in' ,all cases octahedral geometry is dominant, this presents little problem for PVP. Mobility of polymer chains could, however, be an important factor in bringing two redox sites into juxtaposition. Some sophistication is often seen in biological redox polymers where the redox sites are held in favourable spatial relationship and the metal is located in sites which represent a compromise between the geometrical requirements of the oxidized and reduced forms. The electron transport obeys diffusion laws where studied by transient methods. An experimental system which has provided recent information on conduction in redox polymers consists of a platinum electrode on which the redox polymer is deposited. Porous gold is deposited at the other interface, porous so that charge compensating ions may migrate in and out to the polymer film as the oxidation state of the redox centres changes. In one example the electrode may be represented The experiments were then carried out in acetonitrile 0.1 M in as Pt]Os(bipy)z(PVP)22+IAu. Et4N+C104-. If the gold electrode is held at 0 V (vs. SCE) and the platinum electrode made increasingly positive, the current at each electrode can be monitored as a function of the potential of the platinum electrode. The E x for half maximum current was 0.74 V, close to E"' (0.73 V) for the osmium polymer-bound complex (both figures vs. SCE). When the limiting value of current is reached, a11 sites in contact with the gold eiectrode are OsII and those contacting the platinum electrode are OsJr1and steady state concentration gradients exist across the polymer film. Current flows only when mixed oxidation sites are present.55 Using the same apparatus it is possible to explore electron mobility involving 0s1I/Os1and even Osl/Oso(Figure 5). This is achieved by scanning the platinum electrode to negative potential. It was observed that the relative current carrying capacity of the three couples Osll~/Os~~:Osll/Osl:Os'/ was in the ratio 1:3:25, with the most highly conducting state passing currents of 0.6 A cmP2.

-I Pt

+

Pf

Os ( bi py I2 (PVP:)

-

+ Os ( b Ipy)2 (PVP)'+

205(b,pyl2 (PVP)+

Au

Electrochemical Applications

19

Having debated the mechanism of charge transport within the polymer film, it is now useful to consider a few examples of chemical applications of polymer modified electrodes. Electrodes coated with [R~(bipy)~Ci(PVP)lCl or [R~(bipy)~(py)(PVP)]Cl~ show strong catalytic effects for the reduction of cerium(1V) and the oxidation of i r ~ n ( I I ) . ~ ~ Another interesting application is the study of the kinetics of thermodynamically unfavourable oxidations of a series of iron, ruthenium and osmium complexes with 2,2'-bipyridyl or 1,lOphenanthroline (MLJ2+)(equation 28). If Eso' for MLz+i3+is more positive than E"' for the electrocatalyst, currents will be controlled by the slower rate of equation (28) and k12becomes measurable, thus providing a measure of electron transfer rate in a thermodynamically unfavourable back direction. It was found that the Marcus theorys8 holds in this thermodynamically unfavourable regime. The reaction is driven by the electrode regeneration of 0s1I1in the polymer film. Detailed analysis of kinetic data is consistent with a model in which electron transfer to or from ML32+/3+ occurs exclusively with the outmost monolayer of the osmium sites on the polymer.

-

Ptl[Os(bipy),(PVP)#+

+

ML32+

k12

Pt][Os(bipy)2(PVP)d2+

+

ML33+

(28)

-e

A number of polymer films related to PVP are known. An obvious variation is polybipyridyl. Thus the complex cation tris(4-vinyl-4'-methyl-2,2'-bipyridyl)ruthenium(II) may be electroreduced in acetonitrile solution to give a polymer film on a platinum electrode. Pulsing of the electrode at 0.5 Hz between + i . 5 and -1.5 V (vs. SCE) gives an orange emission (ecl) arising from annihilation between RulI1and RuI centres.59 Compounds (20) and (21) undergo electrochemical oxidation to give films of poly(2-methyl8-quinolinol) and poly(8-quinolinol) on a variety of metal substrates.60Copper(I1) can ke complexed from aqueous media, but cobalt(I1) requires an organic medium. X-ray photoelectron spectroscopy shows the copper(I1) complex of films derived from (20) to be complexed to both N and 0 and also shows that water is absent from the primary coordination sphere. However, for cobalt(I1) on the polymer derived from (21),water is present in the primary coordination shell.

(20)

(21)

Tetracoordinated copper(I1) enhances electron hopping but of the cobalt(I1) doped polymers it appears that only those cationic sites adjacent to the metal surface are active.

57.3.2.3 Polyvinylferrocene and related systems The previous section has illustrated the variety of redox centres that can be used to modify electrode behaviour with the help of polymers containing heterocyclic nitrogen atoms. Thus the redox centre has been retained close to the electrode surface both by ion exchange and by coordination to a pendant group of the polymers. Further variations would be provided either by makmg the redox centre a pendant group on an appropriate polymer chain, or by seeking some means of binding the atoms covalently to the actual electrode surface. Both possibilities are usefully introduced via ferrocene derivatives. ( i ) Attachment to the electrode surface

Vinylferrocene (22) may be polymerized (Section 57.3.2.2.i) to give a polymer in which the iron(II/III) redox centres are pendant from a carbon backbone. Copolymers have also been formed with styrene61and acrylonitrile.62Another approach using a different polymer is illustrated by the covalent binding of poly(methacry1chloride) to Sn02electrodes followed by attachment of pendant ferrocene centres by reaction of hydro~ymethylferrocene.6~

CCC6-0

Electrochemical Applications

20

Anodic oxidation of a nickel electrode will give a surface structure that can be represented as Ni-surface-0, H, (A). Reaction with, say, a dichlorosilane will then follow (equation 29): (A)

+ y/2C12SiR2

--+

Ni-surfa~-OxSiR2

+ yHCl

(29)

If (23) is selected as the dihalosilane, a convenient way of modifying the nickel surface is available.@The electrochemical properties of the treated nickel electrode are very similar to those of a similarly derivatized platinum electrode; for example, both are equally effective in the electrocatalytic oxidation-reduction of solution ferrocene. Normally oxidation of the nickel surface would be a competing process ultimately rendering the electrode passive. The surface modification clearly eliminates this prablem and opens up the possibility of using surface modified inexpensive metals as electrodes. The above example of using silanes to surface modify an electrode is but one of many. The method was pioneered by who was able to demonstrate the successful reaction depicted in equation (30) by use of X-ray photoelectron spectroscopy. The presence of a good donor group can then be used to coordinate metal based redox centres, or alternatively further organic chemistry may be carried out, e.g. following equation (31). The method is then quite general, but is well suited to forming monolayers rather than a multilayer coverage, which characterizes most redox polymer derivatized electrodes.

A further example of the use of this technique to introduce a ferrocene redox centre to a platinum surface is given in equation (3.2). A comparative survey was made of the rates of heterogeneous charge transfer between the platinum electrode and ferrocene both in solution and immobilized on the surface. Both processes show an Arrhenius temperature dependence but AGAcT(S0h) # AhCAcT(swface bound). Absolute rate theory was unsatisfactory for the surface reaction and the need to involve electron tunnelling and a specific model for the conformation of the surface was indicated.66

Fe

(ii) Same properties of ferrocene mod$ed electrodes

Whilst it is agreed that electron transport within polymer supported ferrocene involves ferrocene-ferrocenium electron hopping and requires no participation from the organic framework, recent studies of electron transfer to a substrate in solution indicate that t h i s need not be mediated by iron(II/III) hopping. Thus X-ray photoelectron spectroscopy gave no evidence for unexposed platinum on an electrode coated with polyvinylferrocene, but scanning electron microscopy revealed channels in the polymer layer. It was concluded that the reactant could either diffuse through such channels or pinholes to the electrode surface, or indeed diffuse through the polymer matrix.68These possibilities are illustrated in equation (33).

21

Electrochemicnl Applications

M+ ~~~

A-

(33) ,,

h

Attempts to develop a model for the digital simulation of the cyclic voltametric behaviour of PVF films on platinum62electrodes required inclusion of the following features: (a) environmentally distinct oxidized and reduced sites within the film; (b) interconversion of the above sites and interaction between them; (c) rate of electrochemical reactions to depend on the rate of interconversion of redox sites, the rate of heterogeneous electron transfer between film and substrate, intrafilm electron transfer and the rate of diffusion of counter ions; and (d) dependence on the nature of the supporting electrolyte and the spacing of electroactive groups within the film. Although the detailed mechanism of electron transport and transfer involving PVF electrodes may be complex, they show remarkable stability and rapidly reversible redox behaviour in nonaqueous solvents such as acetonitrile. This has led to the suggestiod9 that they might function as standard electrodes for non-aqueous solvents. Such standards are required since the SCE is unsatisfactory in a number of respects. Particularly, the liquid junction potential between aqueous and non-aqueous solutions is unknown and irreproducible; also there is a danger that the test solution will become contaminated with water and with potassium and sodium ions. PVF films may alter the electrochemical response of semiconductor electrodes in useful ways. For example, for potentials more positive than -0.8 V (SCE) only reductions are possible with single crystal or polycrystalline TiOz in contact with a ferrocene solution (acetonitrile). However, with plasma coated with PVF,ferrocene sites are oxidized in a potential region where n-Ti02 is considered blocked to electron transfer.70Here two factors are involved: obviously the coating of the electrode with an electroactive species, but also alteration of the surface energetics of the semiconductor. This section can be concluded by reference to an example of a chemical application of a PVF modified e l e ~ t r o d e .Halocarbons ~~ may be electrocatalytically reduced using PVF on carbon as the electrode with photoassistance. The overall scheme is given in equation (34):

+

F ~ ( C P ) ~ RX

K e

hv

Fe(Cp)z-XR

I < 400nm

Fe(Cp),

+

X-

+

R'

(34)

If the electrode is held at the iron(II/III) potential, equation(35) applies. carbonlFe(Cp)zFe(CphI

+

K

RX

e

CIFe(Cp)zFe(Cp)z.XRI

hv

S+

CI Fe(Cp)zFe(Cp)z.XRI C C IFe(CphFe(Cp),%- RI

57.3.2.4

Prussian 3lue motlified electrodes

It has already been noted that hexacyanoferrate ions may be exchanged on to protonated and quaternized PVP (Section 57.3.2.2(ii)), and indeed other polymeric ion exchange materials may be used, e.g. poly(viny1 sulfate) and poly(styrene Also, similar redox species may be anchored using chlorosilanes, e.g.

I

C12Si[(CH33CNl2 [F$'(CN)5VW)I3(24)

22

Electrochemical Applications

In this section, electrodes with relatively pure films of hexacyanoferrate(II/III) salts will be considered. They can be produced by a variety of means. In one series of experiments, graphite electrodes were treated with Fe(CO)5 in a glow discharge, after which the electrode surface contained iron(II1) oxides and carboxylates (from oxidation of carbon monoxide). When the electrode is placed in aqueous K4fFe(CN),] the [Fe(CN)#- couple is attached. The film is stable over many thousand electrochemical cycles and colour changes corresponding to those shown in equation (36) are noted. IMFelllFe(CN)dl~5 IFe'n{Fe'1'(CN)~}lo.5 Berlin Green

-%e-, - %M f

MFei"[Fe(CN)d Prussian Blue

e- M

[M2Fen{Fe(CN)6)] (36) Everitt's salt

Films of Prussi& Blue may be electrodeposited on materials such as platinum, glassy carbon, Sn02 or gold. The electrodes are cathodically polarized in solutions of ~ ~ O ~ ( I I I ) / [ F ~ ~ ~ ~ ( C N which are 0.01 M in HC1. Deposits of Prussian Blue are obtained most rapidly on gold and rather less rapidly on SnOz. Some 57FeMossbauer data have been obtained for films on transparent Sn02 (which may have some value as electrochromic displays). Thus at 0.6 V the parameters are 6 = 0.37 mm s-I, A = 0.41 mm s-'; at -0.2 V (SCE), 6 = 1.14 mm s-l, A = 1.13 mm s-l. These data, together with ESCA measurements, serve to indicate that the film is indeed water insoluble Prussian Blue, Fe4[Fe(CN)&. The optical spectrum of the electrode was studied as a function of potential with the results given io Table 4.The voltammetric wave at -0.2 V was found to be dependent on the nature of the supporting cation in solution; the cations may therefore migrate into the open 'zeolitic' structure.74 Table 4

Optical Maxima of Prussian Blue on SnO, as a Function of Potential .-..

Potential of electrod@ (V)

Optical maxima (nm)

Potential of electrode" (V)

Optical maxima (nm)

0.6 -0.2

700 (Prussian Blue) No distinct visible bands

1.4

420 (full oxidized) 420, 770

a

1.1

2:

SCE. 770 nm band is intervalence charge transfer and varies with the degree of oxidation.

Some further information on the migration of supporting cations through zeolitic channels in The the Prussian Blue type of structure comes from a study of a related system, CU[F~(CN),].~~ copper(I1) hexacyanoferrate(I1)may be electrodepositedon glassy carbon when cyclic voltammetry reveals two. redox couples at 0.69 V (reversible, Fe) and 0.00 V (cathodic), 0.35 V (anodic) (Cu). Potassium ions are transported into the film during reduction preferentially to sodium. This observation correlates with effective ionic size (K+,2.4 A; Na+(aq), 3.6 A) in comparison with the zeolitic channel size (3.2 A), although it is not in fact possible to return the fih to the potassium only form after exposure to aquated sodium ions. However, potassium will displace totally the larger aquated lithium ion (4.7 A). The results of the study indicated a very low barrier to electron transfer in Cu[Fe(CN),J films and, in contrast to the immobilized redox polymer mediations considered earlier, the behaviour of the film is similar to a conductor in that solution species undergo reactions governed by their redox potentials and independently of the film redox reaction. Deposition of C4Fe(CN)6] on to transparent SnOz gives an optically transparent electrode which undergoes a marked change in the visible spectrum on redox reaction, thus making this material a further candidate for electrochromic devices. The possibility of using surface modification of cheap metals to make them effective electrode materials has been mentioned (Section 57.3.2.3(i)). A further example employs cyanoferrates and cyanoruthenates as the redox centres.76Complexes such as [M(CN),L]"- (M = Fe, Ru; L = CN, HzO, NO, L-histidine) may be immobilized on a partially corroded nickel surface. The surfaces have good stability and diffuse reflectance IR spectroscopy shows the presence of bridging cyano groups, implying the presence of a binuclear mi, M) species in the surface. A general equation for the redox reaction is: R, [Ni"(NC) M"'(CN),]y

+e-

-e

R, [Ni"(NC) M"(CN),IY

(37)

The redox potential is found to be a function of the ionic size of R, varying over 500 mV for the series Li-Cs, thus raising the possibility of developing ion sensitive electrodes.

Electrochemical Applications 57.3.2.5

23

Miscellmeous sarface modified electrodes

In the foregoing sections, surface modifications by materials which have received significant attention have been gathered under separate headings. In this section, other modifications to electrode surfaces which have involved coordination compounds are considered. (i) Clay modified electrodes

Montmorillonites (smectite clays) have structures resembling that of pyrophyllite but the structure is not electrically neutral. Exchangeable cations are located in interlamellar regions of the clay and, furthermore, the clay can be flocculated such that the plate-like crystals compact with parallel c-axes to give coherent layers. The smectites are then attractive materials with which to modify electrodes. Some experiments have been carried out with a sodium montmorillonite dispersion on an Sn02 e l e ~ t r o d eThe . ~ ~ layer of clay adhered well to the surface and [R~(bipy)~]~' was successfully exchanged on to the clay. The film was electroactive but cracked readily. The addition of powdered platinum gave a more coherent layer. Other species exchanged on to the clay included [Fe(bipy)J2+ and a trimethylammonium derivatized ferrocene. An interesting variation is to exchange an optically pure tris chelate complex on to montmorillonite, e.g. A-[Ru(phen)J2+. If this clay is used to modify the surface of an Sn02 electrode, it is found that the electrochemical oxidation of [ C ~ ( p h e n ) ~at ] ~30 + "C produced A-[C~(phen)~]~+ with 7% optical purity.78 (ii) Novel polymers

The electrochemical oxidation of tyramine in NaOHlMeOH media gives films of polytyramine (25). The film, on a platinum electrode, can complex copper(I1) ions from aqueous media and

cobalt(II), iron(II), manganese(I1) and zinc(I1) from organic media. X-ray photoelectron spectroscopy established that coordination of the metal ions had occurred. For cobalt, evidence of coordination to both ether and amine functions is obtained, bur for the other metal ions evidence of ether coordination is less definitive.

(25) R=CHzCHzNHz

Cyclic voltametric studies show a strong pH dependence for the electrode response when doped with copper. Free amino groups protonate at low pH, causing swelling of the polymer and a more open structure. The detailed redox behaviour of the copper electrode is determined by the rate of charge diffusion through the film. By contrast, the cobalt and iron doped electrodes form a perfectly reversible redox system showing voltammograms which are independent of film thickness. It was suggested that differences arose from more facile electron hopping in the case of 'four coordinate' copper(I1) than for five or six coordinate M1lI1ll(M = Co, Fe). Electrodes doped with manganese(I1) or Zinc(1I) gave no response. Polymeric coatings considered thus far have been based on organic polymers. An example of an electrode (Sn02on glass) modified with a coordination polymer is now ~ o n s i d e r e dComplexes .~~ of chromium(II1) with carboxylic acids have a variety of uses. For example, 'volan', a chromium(II1)-methacrylate complex, is used as a coupling agent for reinforced plastic laminates. Stearic acid complexes are used to waterproof paper and cotton. The structure of such complexes appears uncertain but reasonable speculation suggests that equation (38) may be valid. Thus, if a surface containing OH groups (e.g. SnO2) was treated with a pH adjusted solution of the complex, condensation polymerization of the complex can occur (26). If R is chosen as ferrocenyl, cyclic voltammograms in acetonitrile (NBq'C104-) are the same as those for surface attached ferrocene (Section 57.3.2.3). A novel approach to electrode design using polymeric ligands was recently illustrated using the nickel-triphenylphosphine system.*O An effective homogeneous electroactive catalyst (Ni/PPh3) is solubilized with 2% crosslinked polystyrene phenylphosphine. The complexed polymer is then

Electrochemical Applications

24

H

R

R

(26)

mixed with graphite paste to form a ‘physical mixture’ electrode which, it is claimed, has a longer life than a surface modified electrode and is effective in electrocatalysis without the need for coordinating ligands in solution. Charge is readily transferred in and out of the electrode and the cathodic electrochemistry is similar to the solution electrochemistry, e.g. equation (39). -

~

NiL;’ L

=

5

NiLz+

5

NiLz

-

NiL3

(39)

phosphine; NiL, is most stable for Nio

(The bis-phosphine complex must live long enough to capture a further phosphine ligand or else nickel metal precipitates.) It is predicted that other metals should be effective in such physical mixtures, in which case the idea has greater potential. Much work has been undertaken to modify electrode surfaces with films which are themselves conducting. The most promising approaches involve organic charge transfer and radical ion polymers. Coordination chemistry has, to date, played little part in this work (a good recent review is a ~ a i l a b l e )but , ~ ~one example relating to ferrocene chemistry can be quoted. In th~sexample a well known electron acceptor, 7,7’,8,8’-tetracyanoquinodimethane(TCNQ; 27), is modified and incorporated into polymer (28) in which the iron(I1) of the ferrocene unit is the electron donor. The electrical conductivity of such a film will depend on partial electron transfer between ion and TCNQ centres as well as on the stacking of the polymer chains. The chemistry opother materials, based on coordination compounds, which have enhanced electrical conductivity is covered in Chapter 61.

(iii) Phthalocyanine

Dye molecules covalently attached or adsorbed on to semiconductor electrodes may sensitize the electrode to visible light. Thus it may be possible to produce photocurrents and photovoltages corresponding to longer wavelengths of light than would correlate with the semiconductor band gap. Metal phthalocyanines (MPc) are a very a t k t i v e choice of dye since they are stable and the chromophores absorb strongly in the solar spectrum.Also a wide variety of metal phthalocyanines are known; thus a wide range of redox potentials is available, and indeed some are semiconductors in their own right. The semiconductors n-Ti02 and n-W03 have been sensitized with thin films (100-250 A) of MPc (M = Mg, Zn,TiO, Co, Fe, ClAl,H2)*l and have been shown to effect the photo-oxidation of I-, [Fe(CN)#-, hydroquinone and [FeI1(EDTA)l2-. Photocurrents were proportional to the light

Electrochemical Applications

25

intensity and the spectral response was characteristic of the MPc film. If the electrode was held at negative potentials, the photosensitized reduction of benzoquinone, [Fe"'(EDTA)]- and dioxygen occurred. This is consistent with p-type semiconductor behaviour of the MPc films. Iron phthalocyanine derivatives are amongst the most powerful catalysts for dioxygen reduction, but thin films of FePc on glassy carbon show deactivation on repeated electrochemical cycling.82 The cyclic voltammogram of FePc (3000 A) on gold in helium-saturated 0.5 M H2S04shows three anodic and three cathodic waves, each corresponding to a one-electron change. It was concluded that the waves corresponded to dihydrogen adsorption and desorption at different surface sites. The electrode will promote the reduction of dioxygen, the reduction wave shifting from -0.5 V on bare gold to -0.05 V on the FePc film,but the intensity of the wave decays on repeated scanning indicating some deactivation process. The deactivation is caused by a combination of blocking by peroxide intermediates and actual loss of iron centres from the film. (iv) Bipyridyl, phenanthroline and related systems Complex cations containing 2,2'-bipyridyl or 1,lO-phenanthrolineas ligands, particularly those of ruthenium, have been encountered in Section 57.3.2.2(iii). However, the bis-chelate complex was generally anchored to a polymer chain by coordination to pendant pyridyl groups although the possibility of electrochemical polymerization of a 4-vinyl-4-methyl-2,2'-bipyridylwas considered.59In this section, some miscellaneous examples of the role of bipyridyl and phenanthroline complexes are considered. The theme of photosensitizing semiconductor electrodes introduced in Section 57.3.2.5(iii) may be developed with an example from ruthenium-bipyridyl chemistry. The sequence (40) is well known. The effectiveness of the photosensitization should be increased by the covalent attachment of the tris(bipyridyl>ruthenium(II) entity to the semiconductor surface, for example to Sn02. This has been achieved using the versatile halosilane chemistry shown in equation (41). The counter anion was PFs-. Cyclic voltammetry showed that the behaviour of the syitems Sn02/aqueous [ R ~ ( b i p y ) ~and ] ~ +Sn02(film)/electrolytewere very similar but with a +0.05 V shift in E". The coated electrode gives a photocurrent with a red shift of 10 nm which is twice as large as for the non-coated electrode. Unfortunately the current falls off due to promotion of the hydrolysis of the film. [Ru(bipy)$!+

hv

[Ru(bipyb?*+

1=455 nm

+e-

[Ru(bipy)#

(RED)

(40)

1

-~.

( ~ = 1 . 4 x104M-Icm-*)

I

CHlSiC13

0 0 0

L SnOz substrate A neat exploitation of the ideas of Section 57.3.2.2(ii) is achieved via the ligand bathophenanthrolinedisulfonic acid (29), which can be ion exchanged on to a quaternized polymer such as poly(N-methylethylenimine) (30). Subsequent treatment with iron(I1) will give the orange-red bathophenanthroline complex.84An interesting potential application as an electrochromic device is illustrated by sputtering A1203on to an Sn02 surface, except for seven segments defining the number '8' as in a digital display. Treatment of the surface with the bathophenanthroline polymer gives an orange-red colour. If the potential of the electrode is held at +1.2 V (vs. SCE; Fe" + FeIT1),within 0.2 s a negative '8' appears. Bis(1,lo-phenanthroline)copper(I) has an absorption spectrum which is a function of concentration. This has been attributed to oligomerization, possibly via K-E intermolecular stacking interactions. The tendency to oligomerize has a marked effect on the electrochemistry of the complex. For example, the complex exhibits extensive adsorption on the surface of a graphite electrode. The multilayers exhibit good electron mobility and the layers probably grow by reduction of surface copper(I1). Rotating disc voltammetric measurements of the reduction of

26

Electrochemical Applications

(30)

(29)

[ C ~ " ( p h e n ) ~ ]on ~ +electrodes coated with [Cu1(phen)2]+allowed estimation of the rate constant for electron transfer between the two complexes, a value of lo5 M-'s-' being obtained.*' 1,lO-Phenanthroline inhibits the anodic dissolution of iron and the new technique of surface] ~the + electrode enhanced Raman spectroscopy (SERS) has identified the presence of [ F e ( ~ h e n ) ~on surface during dissolution. 4,4'-Bipyridyl has been used to modify silver electrodes to overcome the sluggish response of proteins at the electrodes. Strong SERS is observed both at the oxidation (+0.45 V) and reduction (-0.6 V) potentials, indicating that the ligand remains adsorbed. The spectra observed as a function of potential show the presence of Agl-bipy complexes, silver metal with adsorbed bipyridyl and, at -1.4 V, bipy + b i ~ y - . ~ ~

57.3.2.6

Applications

In this section, many of the objectives driving current research in the field of modified electrode surfaces are identified. The constraint that the application should involve coordination compounds is applied; hence a somewhat constricted view is obtained which is nonetheless, fortuitously, reasonably comprehensive. (i) Ion selective electrodes The development of electrodes to measure rapidly the concentration of ions to which they are uniquely sensitive has important applications in many fields. Two recent examples are quoted of surface modification of an electrode with a ligand showing some specificity for particular ions. In one approach crown ethers (31) or (32)are mixed with PVC and a plasticizer such as 04trophenyl octyl ether in THF to give a membrane on evaporation of the solvent. Discs may then be cut and incorporated into electrodes. The resulting electrodes show good selectivity for thallium(1) and are useful in the measurement of thallium concentrations.s8 0

(

>c-(cw"/2+ 0

I

Me(CH2)&02CH, I

CH2 I

(31)

(32)

Enhanced binding of sodium ions by electrochemically reduced nitrobenzene substituted lariat ethers (e.g. 33) has been observed.89This is attributed to interaction between ring bound Na+ and the reduced side chain. Log Ks (33.Na) has been estimated to be 6.33, where K, is the binding constant for Na+ and the anion of (33). (ii) Reactions ut electrodes Nafion modified electrodes (Section 57.3.2.1) have been shown to promote electrogenerated chemiluminescence from the [R~(bipy)~]~+/oxalate system.44The problem of the quenching of the ecl by water was overcome in the Nafion case since sections of the electrode coating were hydrophobic. However, aqueous systems continue to be of interest in the development of practical ecl devices such as digital display units (see also Section 57.3.2.5(iv)). It has been shown that ecl is

Electrochemical Applications

27

(33)

obtained at the surface of a platinum or glassy carbon electrode immersed in an aqueous solution which is M in [ R ~ ( b i p y ) ~and ] ~ +5 x M in oxalate when a potential more positive than that required for oxidation of the ruthenium(T1) complex is applied. The pH should be 3 2 . Removal of dioxygen is not necessary, but the intensity of the orange luminescence is enhanced by flushing with dinitrogen. Other organic acids such as pyremic, malonic and lactic acids are effective.90 A different application of ecl is illustrated by the alternating current electrolysis of transition metal carbonyl complexes. Figure 6 shows the cycling of a complex A between oxidized and reduced forms. Use of alternating current electrolysis enables chemical use to be made of ecl. For example, reaction (42) is known to be achievable by conventional photochemical means and can also be carried out by applying an alternating potential in an electrochemical cell.91

'Photochem ica I' reactions

Figure6 Cycling between E1 (RED)and E2 (OX)to produce ecl

The phenomenon of ligand bridging between redox centres in homogeneous electron transfer reactions is well established. It has now been shown that such ligands may adsorb on to the electrode surface and mediate electron transfer to or from the electrode from some oxidizable or reducible species in solution. Sulfur atoms in thiolate or thioether ligands are well known for their ability to mediate electron transfer in homogeneous redox reactions; indeed, in reactions involving chromium(II), a CrIII-S bond is often found in the product, indicating an inner sphere mechanism. If chromium(I1) is oxidized at a mercury electrode in the presence of S(CH2C02H)2,there is extensive incorporation of the thioether in the chromium(I1) product when the oxidation is carried out between -0.3 and -0.4 V (vs. SCE), but incorporation diminishes at more positive potentials. It is argued that the initial chromium(II1) product may be (34)in which the chromium-sulfur bond is labile.92 (H,O)&r LOoc\ -S

(3) Kfomation

121

7 C02H 4x

The thioether (13) considered above has previously been mentioned in Section 57.2.4, when it was concluded that substantial complexing of zinc ions in the bulk solution of a plating bath did not occur. This work now establishes that such agents can be active at the electrode surface and helps to establish that adsorption on to the electrode must be an important prelude to the effective action of addition reagents in electroplating. CCC6-0*

28

Electrochemical Applications

Both RuIII(EDTA) and Ru"'(HEDTA) (HEDTA = N-2-hydroxyethylenediaminetnacetato-) are adsorbed at mercury electrodes, but if a ligand is present which can bind the labile sixth coordination position, adsorption is reduced. Thiocyanate is effective in reducing the quantity of complex adsorbed, possibly due to the formation of a thocyanato complex.93 (iii) Electrocatalysis

Equations such as (1) in many cases have a rate constant that has the Arrhenius form even when they occur at the electrode surface (equation 43): k

=

A exp( -EJRT)

(43)

(k = rate constant, A = frequency factor, E, = activation energy). An important difference between a reaction occurring at the electrode surface and that in homogeneous solution is that the rate of the process depends on the potential of the electrode. Since the electrons can be considered reactants, a more negative potential will effectively stabilize the reactants and, hence, lower E,. It is seen that rate constants for oxidation and reduction processes vary in opposite senses with potential. Some reactions are kinetically facile. A good example is the reduction of [RU(NH~)~]~' since this involves a one-electron change and no major stereochemical change between oxidized and reduced species. Furthermore, n6 adsorbed intermediates are involved. Where the reaction is more complex involving several steps and adsorbed intermediates, it is often necessary to apply a more extreme potential than that implied by thermodynamic considerations in order to see an enhancement of the rate constant. Many electrode reactions are potentially attractive but may be too slow for practical application. An electrocatalyst may increase the rate to practicable values and often takes the form of a modification to the electrode surface. A popular reaction for study has been that shown in equation (44),which is actually a rather complex process which proceeds through high-energy intermediates and has a tendency to stop after the addition of only two electrons (equation 85). 0, + 4Hf f 4e02

f

2H'

t 2e-

2HzO

(44)

H202

(45)

-

----t

One of the economic incentives for the study of reaction (44) is that the reaction relates to the successful development of methanol-oxygen and hydrogen-oxygen fuel cells which are capable of extremely efficient conversion of the energy contained in a fuel to electricity. One of the most elegant approaches to the problem to date involves a cofacial cobait-porphyrin system adsorbed on to graphite electrodes. The active porphyrin is represented in (35). The mechanistic scheme which has been proposed is represented schematically in equation (46), where the symbol 'CoIL1l\Co"'' represents one bifacial porphyrin entity in which both metal atoms are cobalt(I1I). The spacing of the two porphyrin rings appears critical for the effective catalytic action.94

1

c=o

-

-

I

Electrochemical Applications co"'

I

I

co"'

c o"

I

I

CO'"

CO"

+

CO"

+ o2

co"

c o"

0 2

+o-0

+o-0

+e-

-e-

+e-

CO"

I

l

CO"'

CO"

I

I

CO"

-e-

I

29

CO"'

I

co"

I

1

+e-

CO'U

Coli

+o-0

CO"

-*

+o-0

I

CO"

CO"

-e-

'

(46)

co"

i fast 3H+,2e-

Another approach which has been effective in achieving the four-electron reduction of dioxygen to water is based on the irreversible adsorption of iron(II1) tetrakis(N,N,N-trimethylani1inium)porphyrin (FeTMAP) on glassy carbon electrode^.^^ Data suggest that the dioxygen reduction is determined by a surface redox couple in a surface eiectrochemical catalysis mechanism (equation 47). Unfortunately, repeated scanning of the potential range results in loss of activity (10% per cycle). This is attributed to loss of porphyrin from the surface and to degradation by the intermediate H202. F$TMAP(adsorbed)

+

e-

t

-

Fe"TMAP(ads0rbed) I

(47)

02

Even reactions which are thermodynamically unfavoured, e.g. with Kaqas low as 5 x lop6,may be electrocatalytically mediated. For example, rotating platinum electrodes covered with polymerized [Ru(4-vinyl-4'-methyl-2,2'-bipyridyl)3]2+Wif1 electrocatalytically mediate the oxidation96 of such species as [Ru(bipy)3I2+,[R~(bipy)2(4,4'-bipy)~]~+, [Ru(bipy)2(py)(MeCN)l2+,[Ru(bipy)z(MeCN)2I2+and [R~(bipy)2(pyrazine)2]~+.

(iv) Electronic devices Considerable potential exists to design surface modified electrodes which can mimic the behaviour of electronic components. For example, a rectifying interface can be produced by using two-layer polymer films on electrodes. The electroactive species in the layers have different redox potentials. Thus electron transfer between the electrode (e.g. platinum) and the outer electroactive layer is forced to occur catalytically by electron transfer mediation through the inner electroactive layer. Such bilayers can conveniently be built up by successive electropolymerization of complexes containing ligands with vinyl substituents, e.g. 4-vinylpyridine or 4-vinyl-4'-methyl-2,2'-bipyridyl. The films may be deposited on metallic or semiconductor electrodes (e.g. Pt, glassy carbon, Sn02, TiOz). More efficient metallation of the films is obtained by polymerization of coordinated ligand than by subsequent metaliation of a preformed polymer film. An alternative to discrete films would be a copolymer with distinct redox sites, or a combination of a single polymer film with a copolymer film in a bilayer device. The concept can be illustrated with the help of Figure 7.97 The inner layer (redox system A) insulates the outer layer (copolymer of redox system A and B). The outer layer must be permeable

30

Electrochemicd Applicutions

to counter ion flow as required by redox reactions of the inner layer. As the potential is scanned from negative to more positive potentials, it is seen that current can flow only at certain potentials; hence maintaining the potential at value (i) or (ii) will give a device which permits current flow in one direction only. Suitable chains of A and B are [R~~~(bipy),(4-VP)~]~+ and [R~(bipy)~(4~iny1-4‘-methyl-2,2‘-bipy)]~+ where 4-VP is 4-vinylpyridine.

Solution A

B A

0 A

Figure 7

A variation of the conceptgghas been provided by studying electrodes coa :d by elec rochemical polymerization of ruthenium(II)-4-viny1-4’-methyl-2,2’-bipyridylcomplexes such that the thickness of the coating can be controlled. If such electrodes are placed in contact with a solution containing both Fe(Cp), and [R~(bipy)~fpy)Cl]PF~ (methyl cyanide solvent), it is found that Fe(Cp), may be selectively oxidized with a covering of 10 monolayers. The selectivity depends on the relative permeability and diffusional characteristics of the redox reagents within the polymer film. As the film thckness builds up towards 60 monolayers, charge transfer through the film becomes the only mechanism for oxidation of the ruthenium(I1) complex, but this cannot occur until the film couple is approached, when partiat oxidation Rurl -+ RulI1in the film will create an electron hopping pathway through the film. Thus with thick films the oxidation-reduction behaviour at the electrode is determined by the characteristics of the film rather than by the inherent redox couple. The film may then have properties more usually associated with devices such as diodes and transistors. A novel approach to data storage applications involves a glassy carbon electrode with a coating of [ R u ( ~ ~ ~ ~ ) ~ ( P V Pwhere ) C ~ ] PVP C ~ , is poly-4-vinylpyridine. The rotating eIectrode oxidizes iron(I1) at 640 mV [the polymer coating will inhrbit iron(I1) oxidation until ruthenium(I1) centres are oxidzed to ruthenium(III)]. If the electrode is irradiated, a photochemically mediated substitution of the chloro ligand by an aquo ligand will occur and, after this radiation, a more positive potential (740 mV) is required to oxidize iron(I1). Thus if the electrode potential is held between 640 and 740 mV a ‘yes’rno’ device is possible since irradiation will eliminate the flow of current.99 A rather general problem is that the devices indicated above are usually too slow to be of immediate practical application but clearly the opportunity to develop more research is enormous.

(v) Solar energy conversion The harnessing of solar energy as a seemingly inexhaustible energy source is attractive but many problems have yet to be solved before efficient conversions of solar energy are available. Two thrusts in the research are relevant to this chapter. One possibility is to generate photoelectricity by illumination of a semiconductor electrode in contact with an appropriate solution redox couple which lies within the band gap of the semiconductor. Schematically, this is represented in Figure 8 following W r i g h t ~ n . ~ ~ Alternatively, the photoelectrode may be used to generate fuel, the photoelectrolysis of water being an obvious target. In this case the excited electron-hole pairs must have sufficient reducing and oxidizing powers to drive the 0 2 / H 2 0and HzO/Hzhalf cells. A practical problem with n-type semiconductors is that photoanodic oxidation can quickly render the system useless; surface derivatization was seen as a way round the problem. A range of silanated ferrocenes was used to derivatize silicon electrodes with a remarkable increase in the stability of the photocurrent as a function of time.’** Illumination causes the oxidation of surface bound ferrocene which, in turn, oxidizes a redox species such as [Fe(CN)6]4-in solution. Another approach has been to use a photogalvanic cell with a rotating optically transparent SnOz electrode. Light is used to drive a redox reaction in solution, e.g. equation (48):”

31

Electrochemical Applications

1

Lwd

Volence band

Photogenerated voltage, E, = Et

- Et,

Figure 8 Photovoltaic cell based on an n-lype semiconductor [R~"(bipy)~]'+

kv __f

[R~"(bipy)~*]'+

The above system is in fact not attractive in practice, mainly because rate constants are significantly greater than is desirable. (vi) Metallization

It is interesting to conclude this section with an example that, in a sense, brings the chapter full circle. The metallization of plastic materials used as metal substitutes is a process with actual and future commercial potential. Usually, plastics are plated after appropriate sensitization by an electroless process which involves reduction of metal ions (e.g.Ni2+,Cu2+)by chemical rather than electrical means.19 There seems no reason why the reducing agent should not be incorporated in the polymer and Murray and his collaborators'ol have demonstrated that copper, silver, cobalt coated and nickel may each be electrodepositedon to films of [p01y-Ru(bipy)~(4-vinylpyridine)~~+ on to platinum electrodes. The metal reductions are mediated by the Ru' and Ruo states of the polymer. 57.4

GENERAL CONCLUSION

The unifying theme of this chapter has been chemistry at the electrode surface, an approach which it is hoped has introduced some cohesion to the material selected for coverage. The constraint that coordination chemistry should be involved has been observed but, even within this constraint, no attempt has been made to be totally comprehensive in the coverage of the literature in the sense that would be expected of a review article. Rather a small scale map has been drawn and those now requiring a larger scale map of particular areas are referred in the first instance to those general t e ~ t s ~ , ~and , ~ review , ~ J ~ a r t i ~ l e s ~ Ocited. -~~,~~ It is clear that the field of chemically modified electrodes has a great future. Commercial success is already evident within the chlor-alkali industry where dimensionally stable Ti02/Ru02 derivatized electrodes have replaced graphite and produced significant savings in the consumption of electric power. Dispersed electrodes will attract much attention in the future. One goal might be particles of a semiconductor modified in different ways on opposite sides to drive the H20/Hz and 02/H20 half cell reactions. The coordination chemist does not have a unique role to play since this is very much a multidisciplinaryarea of research however, it would be surprising if no major breakthrough of the future were based on the redox chemistry of coordinated metal ions. 57.5

REFERENCES

1. F. A. Lowenheim (ed.), 'Modern Electroplating', 3rd. edn., Wiley, New York, 1974. 2. E. Raub and K. Muller, 'Fundamentals of Metal Deposition', Elsevier, Amsterdam, 1967. 3. G. Lewis and M. Randall, 'Thermodynamics', 2nd edn., McGraw-Hill, New York, 1961. 4. A. J. Bard and L. R.Faulkner, 'Electrochemical Methods - Fundamentals and Applications', Wiley, New York, 1980. 5. J. S . Fordyce and R. L. Baum, J. Chem. Phys., 1965,43,843. 6. J. O M . Bockris, Z. N a g and A. Damjanovic, J. Electrochem. Sac., 1972,119,285 7 . E. Mattson and J. O'M.Bockris, Trans. Faraday Soc., 1959,55,1586. 8 . R. Weiner and A. Walmsley, 'Chromium Plating', Finishing Publications Ltd., Teddin@on, England, 1980. 9. M. Frey and C. A. Knorr, 2. Elektrochem., 1956,60, 1089. 10. F. Ogbum and A. Brenner, J. Electrochem. SOC.,1949, %, 347. 11. .IEdwards, . Trans. Inst. Met. Finish., 1964,41, 169. 12. R. C. Snowdon, Tmns. Electrochem. Soc., 1907,11, 131.

-

32

Electrochemical Applications

Udylite Corp., Br. Put. 1049 132 (1965). E. I. Du Pout De Nemours and Co., Br. Put. 1 170058 (1966). B. S. James and W. R. McWhinnie, Truns. Inst. Met. Finish., 1980,58,72. B. S. James and W. R. McWhinnie. Transition Met. Chem.. 1981.6. 151. C. O.~Schmakel,K. S. V. Santhanah and P. J. Elving, J . Electrochem. SOC.,1974,121, 1033. Stanffer Chemical Company, US Put. 3755097 (1973). ‘Canning Handbook on Electroplating’, W. Canning and Co.Ltd., Birmingham, England, 1970. C. Barres, J. J. B. Ward and J. R. House, Trans. h t . Met. Finish., 1977,55, 73. J. B. Bride, H u f h g , 1972, 59, 1027. J. J. B. Ward and I. R A. Christie, Trans. Inst. Met. Finish., 1971, 49, 148. C. Kappenstein and R. P. Hugel, Inorg. Chem., 1978, 17, 1945. J. K. h a l l and L. L. Schreir, Trans. Inst. Mer. Finish., 1964,41, 29. H.Brown, Plating, 1968,55, 1047. 0. Beck, A. E. Smith and A. Wheeler, Proc. R. Sot. London, Ser. A, 1940, 177,62. L. Brignatelli, Ann. Chim (Pnuia), 1800, 18, 152. G. Elkington and H. Elkington, Br. Put. 8447 (1840). A. F. Uohmheim, Plating, 1961,48, 1104. M. E. Roper, Metal Finish. J., 1966,255. H. G. Todt, Trans. Inst. Met. Finish., 1973, 51,91. R. 0.Hull and C . J. Wernlund, Modern Electroplating, 1942,5.362 (issued by the Electrochemical Society of America). H. Gerischer, 2.Phys. Chem., 1952,202,302. W. Fairweather, Electroplut. Met. Finish., 1973,29. Schering Aktiengesellschaft, Br. Put. 1309946 (1970). E. I. Du Pout De Nernours and Co., Br. Put. 1381939 (1972). W. Fairweather, Electroplut. Met. Finish., 1973,2?. B. E. Wynne, .I. M. Sykes and G. P. Rothwell, J. Electrochem. Soc., 1975, 122,646. Lea-Ronal, Inc., Br. Pat. 1314268 (1970). 40. M. S. Wrighton, Acc. Chem. Res., 1979, 12, 304. 41. R. W. Murray, Acc. Chem. Res., 1980,13, 135. 42. W. J. Albery, Acc. Chem. Res., 1982,15, 142. 43. L. R . Faulkner, Chem. Eng. News, 1984,28. 44. C. R. Martin, I. Rubenstein and A. J. Bard, J. Am. Chem. SOC.,1982,104,4817. 45. H. S. White, J. M d y and A. J. Bard, J. Am. Chem. SOC.,1982, 104,4811. 46. T. P. Hemming and A. J. Bard, J. Electrochem. SOC.,1983,130,613. 47. I. Rubenstein and A. J. Bard, J. Am. Chem. SOC.,1980, 102,6641. 48. I. Rubenstein and A. J. Bard, J . Am. Chem. SOC.,1981, 103, 5007. 49.. N. E. Prieto and C. R. Martin, J. Electrochem. SOC.,1984, 131,751. 50. N. Oyama and F. C. Anson, A n d . Chem., 1980,52, 1192. 51. H. Brauo, W.Storck and K. Doblhofer, J. Electrochem. SOL,1983,130,807. 52. 0. Haas and J. G. Vos, J. Electrounul. Chem., 1980,113, 139. 53. N.S. Scott, N. Oyama and F. C. Anson, J. Electramal. Chem., 1980, 110, 303. 54. G. I. Sammuels and T. J. Mayer, J. Am. Chem. SOC.,198L,103,307. 55. P.,G. pickup, W.Kutner, C. R. Leidner and R.W. Murray,J . Am. Chem. SOC.,1984,106, 1991. 56. K. Shigehara, N. Oyama and F. C. Anson, J. Am. Chem. Soc., 1981,103,2552. 57. C. R. Leidner and W. R. Murray, J. Am. Chem. Soc., 1984,106,1606. 58. R. A. Marcus, Annu. Rev. Phys. Chem., 1964, 15, 155. 59. H. 0. Abruiia and A. J. Bard, J. Am. Chem. Soc., 1982, 104, 2641. 60. M.-C. Pham,J.-E. Dubois and P.-C. Lacaze, J. Electrochem. Soc., 1983,130,346. 61. A. H. Schroeder, F. 8. Kaufman, V. Patel and E. M. Engler, J. Electrounal. Chem., 1980,113, 193. 62. P. J. Preece and A. J. Bard, J. Electrounal. Chem., 1980, 114, 89. 63. K. Itaya and A. J. Bard, Anal. Chem., 1978, SO, 1487. 64. A. B. Bocarsly, S.A. Galvin and S . Sinha, J. Electrochem. SOC.,i983, 130, 1319. 65. P. R. Moses, L. Wier and R. W. Murray, A d . Chem., 1975,47, 1882. 66. H. Sharp, M. Peterson and K. Edstom, J. Electroanal. Chem., 1980,109,271. 67. A. Skorobogaty and T.D. Smith, Coord. Chern. Rev., 1984,53,55. 68. P. J. Peerce and A. J. Bard, J. Elecfrounul. Chem., 1980,112, 97. 69. P. J. Peerce and A. J. Bard, J . Electrounul. Chem., 1980, 108, 121. 70. D. R. Rolison and R. W. Murray, J. Electrochem. Soc., 1984,131,337. 71. M. F. Dantartas, K. R.Manu and J. F. Evans, J . Elecfrounai. Chem., 1980,110,379. 72. N. Oyama, T. Shimomura, K. Shigehara and F. C. Anson, J. Electroanal. Chem., 1980,112,271. 73. A. L. Crumbliss, P. S. Lugg,D. L. Patel and N. Morosoff,Znorg. Chem., 1983,22,3541. 74. K. Itaya, T. Ataka and S. Toshima, J . Am. Chem. SOC.,1982,104,4767. 75. L. M. Siperko and T. Kuwana, J. Electrochem. Sac., 1983,130, 396. 76. S. Sinha, B. D. Humphrey and A. B. Bocarsly, Inorg. Chem., 1984,23,203. 77. P. K. Ghosh and A. J. Bard, J . Am. Chem. SOC.,1983,105, 5691. 78. A. Yamagishi and A. h a m a t a , J. Chem. SOC.,Chem.Commun., 1984,452. 79. R. E. Malpas, 1.Electrounnl. Chem., 1981, 117, 347. 80. R,Jasinski, J . Elecrrochem. SOC.,1983,130, 834. 81. A. Giraudeau, Fu-Ren F. Fau and A. J. Bard, J. Am. Chem. Sac., 1980,102, 5737. 82. C. A. Melendres and X . Feng, J. Electrochem. Soc., 1983, 130, 811. 83. P.K. Ghosh and T. G. Spiro, J. Am. Chem. Soc., 1980, 102, 5543. 84. K. Itaya, H. Akahoshi and S . Toshima, J. Electrochem. SOC., 1982,129,762. 85. Chi-Woo Lee and F. C. Anson, Inorg. Chem., 1984,23,837. 86. R. P. Van Duyne and M. Janik-Czachor, J. Electrochem. SOC.,1983,130,2320. 87. T. M. Cotton, D. Kaddi and D. Iogra, J . Am. Chem. SOC.,1983,105,7462. 88. H. Tamura, K. Kimura and T. Shono, J. Electrounul. Chem., 1980,115, 115.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

Electrochemical Applications

33

A. Kaifer, L. Echegoyen, D. A. Gustowski, D. M. Goli and G. W. Gokel. J . Am. Chem. Soc., 1983, 105,7168. I. Rubenstein and A. J. Bard, J. Am, Chem. SOC.,1981,103, 512. H. Kunkley, A. Mmtz and A. Vogler, J. Am. Chem. SOC.,1983,105,7241. P. K. Wrona and F. C. Anson, J. Electroanal. Chem., 1981,127,87. S . Gonzalez and F.C. Anson, J. Electroanal. Chem., 1981,129,243. J. P. Collman, M. Marrocco, P. Denisevich, C. Koval and F. C. Anson,J . Electroanal. Chem., 1979, 101, 117. A. Bettelheim. R. Parash and D. Ozer, J . Electrochem. Soc., 1982, 129, 2247. T. Ikeda, C. R. Leidner and R. W. Murray, J. Am. Chem. Suc., 1981,103,7422. H. D. Abruila, P. Denisevich, M. Umana, T. J. Meyer and R.W. Murray, J . Am. Chem. SOC.,1981, 103, 1 . C. D. Ellis, W. R. Murphy, Jr. and T. J. Meyer, J . Am. Chem. Soc., 1981, 103, 7480. 0.Haas, M. W e n s and 5. G . Vos, J. Am. Chem. Soc., 1981, 103, 1318. J. M. Botts, A. B. Bocarsly, M.C. Palazzotto, E. G. Walton, N. S.Lewis and M. S . Wrighton, J . Am. Chem. Soc., 1979,101, 1378. 101. P. G. Pickup, K . PJ.Kuo and R. W. Murray, J. Electrochem. Soc., 1983, 130,2205. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100.

58 Dyes and Pigments RAYMOND PRICE IC1 plc, Organics Division, Blackley, Manchester, UK 58.1 INTRODUCTION

35

58.2 METAL COMPLEXES OF AZO COMPOUNDS 58.2.1 Nature of Bonding by ik Azo Group 58.2.2 Bidentate Azo Compowrds 58.2.2.1 Medially metallized fypes 58.2.2.2 Terminully meralkrd types 58.2.3 Tridentate Azo Compounds 58.2.3.1 Methods of preparation 58.2.3.2 Tridentate diarylazo compounds containing o-amino substituents 58.2.3.3 Stereochemistry and iromerism 58.2.3.4 Tridentate azo compounds containing hetero donor atoms 58.2.4 Tetradentate Azo Compounds 58.2.5 Pentadentate Azo Compounds 58.2.6 Hexadentate Azo Compounds

40

58.3 METAL COMPLEX FQRMAZANS 58.3.1 Bidentatv Formazans 58.3.2 Tridentare Formazans 58.3.3 Tetradentate Formzans 58.3.4 Formazan Analogues

77 78 79 81 83

58.4 METAL COMPLEX AZOMETHINES

83

41 42 42 44 46 46

57 62 73 75 76 76

58.5 METAL COMPLEXES OF O-HYDROXYNITROSO COMPOUNDS

84

58.6 METAL COMPLEXES OF HYDROXYANTHRAQUINONES

86

58.7 MACROCYCLIC METAL COMPLEXES 58.7.1 Phthalocyanines SX.7.2 Others

87 87

58.8 REFERENCES

91

91

58.1 INTRODUCTION

Man has been fascinated by colour since time immemorial, as witnessed by the coloured representations of animals with which the early cavemen adorned the walls of their dwellings. This fascination persisted through the Egyptian, Greek and Roman eras and remains undiminished right up to the present day. From the very earliest times man desired to colour the fabrics he possessed and initially he turned to Nature for the materials with which to accomplish this. The origin of organic dyes and pigments is lost in antiquity but both sprang from the use of natural products such as Brazilwood, logwood, kermes, lac-dye, cochineal, Persian berries and madder. Many of these emerged as useful dyestuffs because they could be applied by one of the earliest dyeing processes. This, to become known as the mordant method, depended upon the formation of insoluble metal salts or complexes within the fibre. In the early days, dyeing was not a science but an art or a craft and could hardly be otherwise. Frequently dyeing processes involved treating the fibres with mixtures of 'dyes' of animal or vegetable origin and varying proportions of clay, earth, ashes, slate dust, powdered brick and various gums, all with the objective of achieving the best and most consistent result. Gradually it came to be recognized that the results obtained with certain dyestuffs and clay materials could be improved by replacing the latter by alum and other metallic salts and mordant dyeing was born. In this the cloth to be dyed was impregnated with a solution of a soluble salt of a metal such as aluminum, iron, chromium or tin and the insoluble metal hydroxide was precipitated within the fibre. This 'mordanted' fibre was then treated with a solution of a naturally occurring colouring matter capable of forming an insoluble complex with the mordant. Various natural colouring matters were used which differed widely in chemical structure. All, however, contained a chelating system. The most important was alizarin (l),obtained from madder root, which gave a beautiful bluish red shade with an aluminum mordant and had excellent fastness properties. In 1868 some

35

Dyes and Pigments

36

400000 acres of land were given up to the cultivation of the madder root and world production was 70000 tons a year. It was an industry in itself, but was destroyed almost overnight when Graebe and Lieberman identified the principal colouring matter as 1,2-dihydroxyanthraquinoneand were successful in making it artificially. Other naturally occurring dyes containing the 1-hydroxyanthraquinone chelating system are cochineal and kermes, the active constituents of which are carminic acid (2) and kermesic acid (3), respectively. Both are obtained from insects, the former from the female of the species Coccus cacti, which lives on cacti in Mexico, and the latter from Coccus ilicis, which infests the kermes oak. About 200000 insects are required to produce 1 kg of cochmeal but it was used in the UK at least until 1934with a tin mordant for dyeing the Guards' tunics and the piping on postmen's trousers! n

HO I

0

HO

I

I 0

(3)

(4)

The success of logwood, from the heartwood of the tree Haematoxylan campechimum L found in Central America, as a black dyestuff also depended on complex formation. The active constituent of logwood is the chroman hematein (4), which is itself red. When applied with a chromium mordant, however, logwood gave black shades and held the premier position for blacks and blues until the late 1890s, by which time synthetic dyes accounted for some 90% of the dyes used. The principal characteristic of a pigment which distinguishes it from a dyestuff is that it is substantiallyinsoluble in the medium in which it is used. There are many other distinctions of vital technical importance but, for the present purpose, this very simplified differentiationis adequate. The development of organic pigments was slower than that of organic dyestuffs but sprang from the use of crude inorganic materials to 'fix' dyestuffs of natural origin within fibres. It was recognized that natural dyes such as alizarin formed highly coloured insoluble substances with clays. These could be incorporated into oily or resinous materials and used for decorative purposes; hence pigment from the Latin pingere, to paint. As methods of dyeing became better understood, so the clays were replaced by metal salts such as alum and the foundation for the development of modem metal complex organic pigments was firmly established. The advent of synthetic dyes, particularly the all important azo series, did nothing to diminish the importance of metal complex formation in dyeing. It continued to provide, through the mordant method, solutions to problems such as the colouring of fabrics with dyestuffs which had no natural affinity for the fibre in question and the achievement of good fastness to washing and light. With the wide variety of azo dyes available it became unnecessary to use different metals to obtain different colours and salts of chromium rapidly displaced those of other metals since 'chromed' azo dyes were found to have the k s t fastness to light and washing, particularly on wool. Three principal dyeing methods were employed, the chrome mordant, the afterchrome, and the metachrome processes. In the first, the wool was treated with sodium dichromate and an agent such as oxalic acid to reduce chromium to the trivalent state. The wool was then dyed with an azo dyestuff, e.g. (5), which formed a chromium complex on the fibre. The process was effectively reversed in the afterchrome method in which the wool was first dyed with a suitable azo dyestuff, then treated with sodium dichromate in the same or a fresh dyebath. In the metachrome process the dyeing was carried out in a single bath which contained the dyestuff, sodium dichromate and

1 H8s03Na /po$ -

-_

H20

HZO\

/OH2

I

Na'

S03Na

\ NO2

0,s

Na+

(7) Perlon Fast Violet BT (IG)

Na'

(8) Irgalan Brown Violet DL (Geigy)

The next milestone in the history of chromium dyes for wool arose from the observation that 1:l chromium complex dyestuffs react with a second molecule of a metal-free, tridentate azo compound to give the 2: 1 chromium complex. If the second dyestuff is the same as that from which

38

Dyes and Pigments

the 1 :1 chromium complex was derived, a symmetrical 2:1 chromium complex (9) is produced. If, however, a 1:l chromium complex reacts with a molecule of a different tridentate azo compound, an unsymmetrical 2: 1 complex (10) results. The range of available shades can be extended by using such mixed structures. Further, it is possible to prepare unsymmetrical 2:l chromium complex dyestuffs from two azo compounds only one of which centains a sulfonic acid group. In this way, relatively cheap products having good solubility have been obtained which give level dyeings on wool from neutral dyebaths with negligible damage to the fibre. Fastness properties are good and dyes of this type occupy a leading position for wool.

2-

(10)

Both 2: 1 chromium and cobalt complex azo dyestuffs have little or no afinity for cellulosic fibres and until the early 1960s their use was restricted to wool and nylon. With the introduction by IC1 of their Procion range of fibre-reactive dyes, however, their use was extended to cellulosic fibres on which they give prints having excellent fastness to light and wet treatments. Before that time the development of metal complex dyes for cellulose had followed a similar pattern to that of the development of such dyes for wool but, in this case, the most important metal was copper. Early work in this field has been reviewed by several authors.' The after-treatment of dyeings on cotton obtained from dyestuffs such as (11) with copper salts was used for many years to improve fastness

Dyes and Pigments

39

to light and washing. On repeated washing, however, the effect was lost: it was gradually recognized that this was because of the low stability of the copper complex formed by (11) and that improvements could be obtained by using dyes containing the o,o’-dihydroxydiarylazo system rather than the o-hydroxy-0’-methoxydiarylazosystem in (1 1). Water-soluble copper complexes of o,o’-dihydroxydiarylazo dyes were first prepared by SCI (now Ciba-Geigy) in 1915 and they could be applied2 to cotton in the same way as ordinary direct dyes. Following that observation, a large number of water-soluble copper complexes of tridentate azo dyestuffs (e.g. 12) were marketed. Such dyes were usually very fast to light but had only moderate wash fastness. This was overcome with copper complex dyestuffs containing fibre-reactive groups, such as some memlxrs of ICI’s Procion ranges, which give dyeings on cellulose having excellent fastness to light and to wet treatments.

(11)

”‘r

H2O

N

~

O

’

q

= 0-cu = p

-N S03Na

J‘y&JN--p CU-0 / \

/ \ -

-

/ \ -

/

NHPh

Na03S

op\$ (12)

N-

&;% \rc

I

1

(13)

So far as metal complex pigments are concerned, the phthalocyanines must occupy the prime position. Phthalocyanine itself was dwovered by von Braunin 1907and copper phthalocyanine some 20 years later by de Diesbach. It was, however, left to Dandridge, Drescher, Dunsworth and Thomas of Scottish Dyes Ltd. (later ICI) to realize the potential of the copper complex as a pigment. They first noticed the presence of a coloured impurity in phthalic anhydride prepared by the reaction of ammonia with molten phthalimide in an enamelled iron vessel in which the enamel was damaged. The coloured impurity proved to be iron phthalocyanine, the iron coming from the reaction vessel. Subsequently copper phthalocyanine (13) was shown to possess even better properties as a pigment and today copper phthalocyanine and its halogenated derivatives are probably the most important pigments employed in high quality blue and green automotive finishes. Dyes and pigments are effect chemicals and it is vital that their technical properties are appropriate for the market for which they are intended. It is immediately obvious that features such as shade, strength and brightness of hue are of prime importance in dyes and pigments but many other requirements must be met by successful products. In the case of dyestuffs, for example, vastly different properties are required for the dyeing of cotton, wool and the three main man-made fibres - nylon, polyester and polyacrylonitrile. Further, the colouration produced should be permanent throughout the useful life of the dyed material, requiring that the dyestuff be fast to light, washing, dry cleaning, heat and various other agencies. Similarly, pigments must be easily dispersed in the medium in which they are employed and must satisfy a whole range of technological criteria. The colour of dyes and pigments stems from the chromogens upon which they are based but many of the other properties are achieved through the introduction of appropriate substituents into the parent chromogen. For example, high wet-fastness on cotton is obtained by the intro-

Dyes and Pigments

40

duction of the dichlorotriazinyl group into a dyestuff which then reacts with cellulose under alkaline conditions to form a covalent bond between the dyestuff and the cellulose. So far as chromogens are concerned, most of those presently employed in dyes and pigments had been discovered by the end of the 19th century and since then very few new chromogens have been added to the range of available dyestuffs, a notable exception being copper phthalocyanine. Similarly, most major advances in metal complex dyestuffs in which the metal may be regarded as part of the chromogen, for example the 2:1 chromium and cobalt complexes of tridentate azo compounds, had been made when this subject was reviewed in the late 1 9 6 0 ~Thus, 3 ~ whilst a large number of publications, particularly patents, have appeared since that time, most have related to the achievement of desirable technical effects by modification of dyestuff structures rather than to coordination chemistry. For example, in the field of 2: 1 chromium complexes of tridentate azo compounds, many publications relate merely to building up new complexes from previously known dye ligands using known methods for their preparation. It is not the purpose of this chapter to present a detailed account of the history and technology of metal complex dyes and pigments since many comprehensive reviews are a ~ a i l a b l e . ~ ,The ~~” discussion has, therefore, been limited to those aspects of the subject which centre on coordination chemistry rather than dyestuff and pigment technology, with special reference to the effects of coordination on ligand reactivity. This brief introduction will serve to demonstrate that coordination chemistry has figured prominently in the development of the dyes and pigments industries from the earliest times. Further, it is at least in part due to the large amount of research in these industries that many of the advances in coordination chemistry since its recognition by Werner have been made. This chapter is concerned with dyes and pigments which are complexes of azo, formazan, azomethine, nitroso, anthraquinone and phthalocyanine ligands. Many of these compounds find important applications in other fields, particularly colour photography and reprography, analysis, catalysis, biology, and some modem high technology industries such as electronics. These applications are described in other chapters of this volume. 58.2

METAL COMPLEXES OF AZO COMPOUNDS

Metal complex azo dyes and pigments may be conveniently divided into two classes: those in which the azo group is involved in coordination to the metal and those in which it is not. The former are by far the more important and are derived from chelating diarylazo compounds (14) having at least one donor function ortho to the azo group. For obvious reasons, only those metal complex dyes stable under dyebath conditions and towards agencies to which the dyed material may be subsequently exposed are used commercially. The stability of the complexes depends upon factors including the size of the chelate ring, the formation of annelated chelate rings, the basicity of the ligand, and the nature of the metal. X

Y

H O H } bidentate NH2

Ho+

?H

A.

A ..

(15)

A decisive factor in the formation of a chelate ring is the number of atoms involved in ring formation: chelates with five- or six-membered rings are the most stable; this criterion is met by

Dyes and Pigments

41

the compounds (14). Similarly, annelation of chelate rings results in enhancement of the stability of complexes. Not surprisingly, therefore, the metal complexes formed by the tridentate azo compounds are more stable than those formed by the bidentate ligands. As a rule the most basic ligand forms the most stable complex and the most acidic the least stable complex. In accord with these principles the relative stabilities of metal complexes of bi- and tri-dentate azo compounds follow the general sequence (15). In practice, only metal complexes of tridentate azo compounds find wide commercia1 application. Chromium and cobalt are the metals most commonly used in dyestuffs for polyamide fibres and leather because of their kinetic inertness and the stability of their complexes towards acid. Since the advent of fibrereactive dyestuffs, chromium and cobalt complexes have also found application as dyestuffs for celldosic fibres, particularly as black shades of high light-fastness. Copper complexes are of more importance as dyes for cellulosic fibres and are unsuitable for polyamide fibres because of their rather low stability towards acid treatments. The importance of such metal complexes stems principally from their very high light-fastness, attributed4 to protection of the azo group by the metal against attack by, for example, singlet oxygen. This explanation appears eminently reasonable since one result of a ligand becoming coordinated to a metal is that it becomes less susceptible to electrophilic attack. This is also manifested in the generally good fastness of such dyestuffs to oxidizing agents such as sodium hypochlorite. The penalty paid for the improvement in these and other technically important properties is that such metal complexes are generally duller in shade than the parent azo compounds. This dulling effect and the associated bathochromic shift does not arise from the forbidden d + d transitions of the metal itself. These are so weak (E-,< IO3) that they are completely swamped by the permitted, intense T + n* transitions of the chromogen (E, ea. 25 000). The colour change is the result of chelation of the major donor groups, azo, X and Y (14), to the metal, which produces a major perturbation of the n-electron distribution. Metal complex azo dyestuffs in which the azo group is not involved in coordination at the metal are derived from azo compounds (e.g. 16) based on coupling components such as salicylic acid, 8-hydroxyquinoline and salicylaldehyde. In general, metal complexes of this type are brighter in hue than those in which the azo group is involved in coordination to the metal, little shade change occurs on complex formation and there is little or no enhancement of the light fastness of the dyestuff. The principal reason for the introduction of dyestuffs of this type was their relatively good fastness to washing. This is now achieved by other means and dyes of this type are of little or no commercial importance.

ON" J J

1C02H

(16)

58.2.1

Nature of Bonding by the Azo Group

The donor properties of the azo group are weak and its involvement in coordination to a metal ion was originally inferred from the observation that whereas those azobenzenes having a hydroxy or amino group ortho to the azo group form metal complexes, those having such groups in meta or pura positions do not. However, it was not clear whether the bonding between the azo group and the metal involved the sp2 lone pair of electrons of one of the nitrogen atoms or the x-electrons of the azo group, and as recently as 1952 this uncertainty was indicated5 by a bracket enclosing the entire azo group (17). In 1947, Werner6 proposed a a-bonded structure (18) for a compound obtained7 by the interaction of azobenzene and platinum(IV) chloride but this was subsequently shown* to be a salt derived from hydrazobenzene and hexachloroplatinate(1V). Later workers succeeded in isolating silver and palladium complexes of azobenzene9 and 5,6-benzocinnoline,lo which may be compared with cis-azobenzene. No information on their structures is available but their existence was taken to demonstrate coordination of the azo group in the absence of other donor groups in the molecule. The mode of coordination of the azo group to a metal ion was not fully resolved until X-ray data on several metal complex azo compounds established that only one nitrogen atom of the azo group is involved in bonding to the metal. In the case of the copper cornplexll (19) of 1-phenylazo-2-naphthol the P-nitrogen atom of the azo group is bonded to the metal which forms part of a six-membered chelate ring.

Dyes and Pigments

42

58.2.2 58.2.2.1

Bidentate Azo Compounds

MeaYaiiy metallized types

In general, the stability towards acids and alkalies of metal complexes of bidentate diarylazo compounds (20) is inadequate for them to find application as dyestuffs and only a limited number are used as pigments. They are, however, of some historical interest and their relative simplicity compared with metal complexes of tridentate azo compounds makes them useful models for the study of the latter, technologically important class. They are, therefore, considered here in some detail.

(20) X=OH, NH2

The first substantially pure complex of a bidentate diarylazo compound appears to have been prepared as long ago as 1893, when CabertiI2 obtained a brown pigment by the interaction of 1-(4-nitrophenylazo)-2-naphtholand copper salts. Copper, nickel and cobalt complexes of o-hydroxy- and o-amino-diarylazo compounds were subsequently prepared by a number of work of Drew and his coworkers14that the structures ~ ~ r k e rbut~it~was~ not J until ~ ~ the , classical ~ of these complexes were established with any degree of certainty. These workers confirmed that both classes of azo compound reacted with copper, nickel, and cobalt(I1) salts to give neutral complexes having 2:l stoichiometry in whch a proton was lost from each molecule of the azo compound. Comparison of the copper complex of o-hydroxyazobenzene with those of the copper complexes of N-phenylsalicylaldimine (21j and benzylidene-o-aminophenol (22) led them to propose structure (23). The structure was subsequently confirmed by X-ray crystal structure determinations on the copper l 1 and nickelI5complexes of 1 -phenylazo-2-naphthoL The copper complex consists of two molecules of 1-phenylazo-2-naphthol centrosymmetrically disposed about the central copper ion, the two oxygen atoms and the p-nitrogen atoms of the azo groups forming a square (24). The molecule as a whole is not planar, the azo group lying out of the plane of the P-naphthol nucleus and that of the benzene ring. This is indicative that the ligand is coordinated to the metal in the azo rather than the hydrazone form (Section 58.2.3.2). Despite the incontrovertible evidence regarding the structure of copper and nickel complexes of o-hydroxydiarylazo compounds, confusion remained with regard to their cobalt complexes. Thus some ~ 0 r k e r ~ ~ 3 ~reported , ~ J ~ J 6the isolation of complexes having 2:1 stoichiometry whilst others5J7 reported 3:l stoichiometry. The oxidation state of the cobalt was also in dispute. The situation was clarified18when more modern techniques were employed to study the reaction of 1-phenylazo-2-naphthol and related compounds with various cobalt salts and complexes. The results of this work are summarized in Scheme 1. Several points are noteworthy: first, the conversion of the 2: 1 cobalt(I1) complex into the 3: 1 cobalt(II1) complex in the presence of mineral acid and atmospheric oxygen, which no doubt accounts for the variable results obtained by earlier workers; and second, the isolation of the complexes of formula [(LH)ZC~llXZ] and their facile conversion into the complexes of formula [L2C011].The former were considered to be octahedrally coordinated cobalt(I1) complexes on the basis of magnetic and conductance measurements and represented the first examples of metal

Dyes and Pigments

[(LH)2COIIX3

43

C0"XZ

LH = 1-phenylazo-2-naphthobX = Cl, Br

Scheme 1

complexes of o-hydroxydiarylazo compounds in which no proton loss from the ligand had occurred. Implicit in this conclusion is the fact that o-hydroxydiarylazo compounds are capable of functioning as bidentate ligands per se. These results led to an acceptable mechanism (Scheme 2)19 for the formation of neutral 2:l cobalt(I1) complexes of o-hydroxydiarylazocompounds by the interaction of the latter and divalent and it is unlikely cobalt salts. o-Hydroxydiarylazo compounds are very weak acids (pKA x that they are ionized appreciably under the conditions typically employed for their conversion to metal complexes. The first step in the reaction is, therefore, seen as coordination of the un-ionized ligand to the metal ion. o-Hydroxydiarylazocompounds exist in solution as an equilibrium mixture of strongly hydrogen-bonded azo and hydrazone forms (Section 58.2.3.2), the position of the equilibrium being determined by several factors, including the structure of the azo compound, solvent, etc. It is not clear which tautomer is involved in the first step in the reaction sequence but this is seen as involving the oxygen atom of the ligand (25). Chelation then leads to the hexacoordinate, neutral complex (26). Treatment of the latter with potassium acetate results in the loss of the elements of hydrogen halide to give the neutral, tetrahedral 2: 1 cobalt(I1) complex (27). It is apparent from thls that the chelated ligand is a stronger acid than is acetic acid (pKA4.75). Not surprisingly, therefore, (27)is the product when 1-phenylazo-2-naphthol reacts with coblt(I1) acetate. Unfortunately the PKA value of the coordinated 1-phenylazo-2-naphthoIwas not reported but, clearly, it is markedly lower than that of the free azo compound. This is a good example of the remarkable enhancement of the acidity of a ligand as a result of its becoming coordinated to a metal ion, which is of considerable importance in problems encountered in the manufacture of chromium complexes of tridentate metallizable azo compounds (Section 58.2.3.1(i)(a)).

Dyes and Pigments

44

O N N / # /

H-0

(27)

(26)

+ 2KX + 2MeCOZH Partial structures; X = C1, Br

Scheme 2

Drew14 proposed structures (28) for the copper, nickel and cobalt complexes of 1-phenylazo2-naphthylamine which were analogous to those of the corresponding complexes of 1-phenylszo-Znaphthol. Doubts regarding this formulation have, however, been raised by the observations18that whereas the 2: 1 cobalt(1I) complex of 1-phenylazo-2-naphthol has tetrahedral has square planar geometry. geometry, the 2: 1cobalt(I1) complex of 1-phenylazo-2-naphthylamine These assignments were made on the basis of magnetic measurements and no X-ray structural information is available. They have, however, been used to support the suggestion that the anitrogen atom of the azo group is involved in the coordination of 1-phenylazo-2-naphthylamines (29).Complexes of this type have no practical application as dyes or pigments but considerable theoretical significance, since their formation apparently involves the loss of a proton from a primary aromatic amino group. This is discussed in more detail in Section 58.2.3.2.

(28) M = CuII, Ni", Co"

58.2.2.2

TerminrrUy metallized types

Terminally metallizable dyes (30) are obtained by the interaction of a diazonium salt and a coupling component containing a chelating system, for example salicylic acid, catechol, salicylaldoxime or 8-hydroxyquinoline, and their coordination chemistry is typical of these compounds. Such dyes were rarely used as preformed metal complexes but were usually applied to cotton and then converted to their copper complexes on the fibre to improve their fastness to wet treatments. A typical example is the blue dyestuff (31).

Dyes and Pigments

45

U N / N \ O C O z H \

OH

(31)

Dyes of this type have been superseded by fibre-reactivedyestuffs which have excellent fastness to washing as a result of the formation of a covalent bond between the dyestuff and the cellulosic substrate and are considerably brighter in hue. Despite the fact that terminally rnetallizable dyes are no longer of commercial interest, one phase in their development affords a further example of the modification of the properties of a ligand as a result of its becoming coordinated. Thus it was discovered2' that bisazo dyes such as (32)gave dyeings having excellent wash-fastness when applied on cellulosic fibres in conjunction with an aftertreatment with a copper salt. The dyestuffs derived their water solubility from the sulfate ester groups which were rapidly hydrolyzed in the presence of copper salts. The hydrolyzed dyestuff was not only insoluble in water but also formed a polymeric copper complex (33) within the structure of the fibre which endowed it with remarkable fastness to washing. In the copper(11) ion-catalyzed hydrolysis of 8-hydroxyquinolinesulfate, Hay and Edmonds22concluded that the intermediate formation of a copper(I1) complex such as (34) promoted cleavage of the sulfur-oxygen bond.

Dyes and Pigments

46

58.2.3

Tridentate Azo Compounds

Complexes derived from medially metallizable, tridentate azo compounds represent the single most important class of metal complex dyestuffs in commercial terms. Some indication of this is afforded by the fact that in 1983 the world consumption of premetallized dyes of this type was of the order of 20000 tons. The most important azo compounds employed in the manufacture of dyes of this type are those containing the o,o’-dihydroxyazo-, the o-hydroxy-0‘-carboxyazo-and the o-hydroxy-0’-aminodiarylazo systems. It is well e s t a b l i ~ h e d ’ that . ~ ~ ~these form four-coordinate copper and nickel complexes (35) in which the coordination sphere of the metal can be completed by a variety of neutral ligands. In both cases the light-fastness of the parent azo compound is improved as a result of complex formation but the nickel complexes are insufficiently stable towards acid to be of commercial interest as dyestuffs. The history of copper complexes has already been discussed (Section 58.1) and will not be considered further here, although it is worthy of mention that currently the most important copper complex dyestuffs are those containing fibre-reactive systems, e.g. (36),for application on cellulosic fibres. L

A

(35) X = CO2, 0; Y = 0,NH; M =CUI*, Nil1; L = H20, NH3, pyridine, etc.

S03Na (36) Procion Rubine MXB

Much more important commercially are the 2: 1 chromium(II1) and 2: 1 cobalt(II1) complexes of tridentate azo compounds, which find a wider application, particularly as dyestuffs for wool, polyamide fibres and leather. These have been the subject of review^^^,^^ which discuss their dyeing properties in detail. The patent literature on metal complex dyes of these types is vast but since this relates principally to the achievement of specific, desirable technical effects by appropriate substitution of the azo compounds it will not be considered in detail here. Rather wil1 the emphasis be placed upon those aspects of dyestuffs of this type which are of general interest in the context of their coordination chemistry and, more particularly, on those areas where uncertainties exist or conflicting results have been reported.

58.2.3.1 Methods of preparation ( i ) Chromium Complexes

It has already been stated that chromium complexes of tridentate metallizable azo compounds occupy their position as the single most important class of metal complex dyestuffs because of their high stability. It should be noted, however, that in this context the term stability is not used in the thermodynamic Sense but relates to the kinetic inertness of the c o r n p l e ~ e s .Octahedral ~~ chromium(II1) complexes have a 8 electronic configuration and the ligand field stabilization energy associated with this is high.26Ligand replacement reactions involve either a dissociative

Dyes and Pigments

47

(SN1) or an associative (SN2)mechanism requiring the formation of five-coordinate or sevencoordinate intermediates, respectively, both of which lead to a considerable loss in stabilization energy.27Hexacoordinate d3 chromium(II1) complexes are, therefore, very resistant to distortions of these types with the consequence that rates of ligand replacement reactions are slow. This is a highly desirable situation so far as the chromium complex dyestuffs are concerned but leads to problems in their synthesis since the starting point is tradtionally a chromium salt containing the hexaaquachromium ion which is itself kinetically inert.*' The consequence of this is that tridentate azo compounds react very slowly with aquated chromium salts in aqueous medium and in many cases the reaction fails to go to completion unless large excesses of chromium are used. This is commercially very undesirable since it results in long plant occupation times, high energy requirements and, in some cases, the use of pressure equipment. Consequently the dyestuffs industry has devoted a considerable amount of research to devising improved methods for the preparation of chromium complex dyestuffs. Before going on to discuss these, however, it is worth considering the preparation of chromium complex dyestuffs from tridentate azo compounds and aquated chromium salts in more detail. ( a ) Syntheses employing aquated chromium salts. Typically, the metal-free azo compound is treated with an excess of a chromium(II1) salt such as the acetate, chloride or sulfate in aqueous medium at the boil, or at higher temperature under pressure. In the case of azo compounds devoid of sulfonic acid groups it is common to add organic solvents such as alcohol, ethylene glycol or formamide to overcome solubility problems. Increasing acidity of the reaction mixture favours the formation of 1: 1 chromium complexes and at a pH of 1.9 or below the 1:1 complex is usually the sole product. At hgher pH values of the order of 7 or above the 2:l chromium complex is usually obtained. The use of a particular chromium salt is, therefore, determined by the product required. Thus chromium(II1) salts of strong acids, such as the chloride or sulfate, which allow the pH of the reaction mixture to fall to low values, are usually employed in the preparation of 1:l chromium complex dyestuffs. Chromium salts of weak acids, such as the acetate or formate, which exert a marked buffering effect, are used in the preparation of 2: 1 complexes. DH,

+

2DH2 DH,

[Cr(H20),]3+3X-

-

+ [Cr(H20)6]3+3X+ [DCr(H20)3]tX-

[DCr(H,O),]+X-+

2HX

+ 3H20

--+ [D2Cr]-Hf

+

3HX -k 6H2O

[D2Cr]-H+

f

HX

f

(1)

3H20

DH, = tridentate diarylazo compound

Tabie 1 pKaLValues of o,a-Dihydroxydiarylazo Compounds Azo compound

PK,'

Azo compound

PK,'

In simple terms the overail reactions involved in the formation of the 1:l and 2:1 chromium complexes may be represented by equations (1) and (2) respectively. In fact the formation of the 2: 1 chromium complex must involve the intermediacy of the 1: 1 chromium complex (equation 3). It is, therefore, significant that when the reaction between a tridentate azo compound and a chromium salt is carried out under conditions which result in the formation of the 2: 1 chromium complex, Le. at relatively high pH, no 1: I complex is detected in the reaction mixture. The obvious conclusion to be drawn from this is that the second reaction (equation 3) occurs very much more rapidly than does the first (equation 1) and that the rate-determining step is involved in the formation of the 1:l complex. Further, the overall reaction is very much more rapid at high than

48

Dyes and Pigments

at low ?Hvalues and, indeed, this has been utilized in the preparation of 1:1 chromium complexes. The metallizable azo compound is first converted to its 2 1 chromium complex at relatively high pH and the latter is treated with a mineral acid to obtain the 1:l complex. In the light of the reported29 p&' values of a series of o,o'-dihydroxydiarylazo compounds (Table l), the clear implication of this is that the hexaaquachromium ion reacts more rapidly with the ionized azo compound than with the un-ionized form. However, it is equally apparent that the hexaaquachromium ion can react with the latter to form a complex in which deprotonation of the ligand has occurred. Unfortunately, although several authors30 have studied the kinetics of the overall reaction, no conclusive results have been obtained since the system is not really suitable for study by the usual titration methods because of the very slow rate of reaction.31 It is, however, possible to make reasonable speculations regarding the rate-determining step on the basis of the reported results (Scheme 3). - 3+

H-0

7Cr(H20)5

\ & NNN%

r

-

\

-

B

1+

H 2 I0

(37) srbeme 3

The first step in the formation of a 1:l chromium complex by the interaction of an opdihydroxydiarylazo compound and the hexaaquachromium ion at low pH must involve replacement of a coordinated water molecule in the hexaaquachromium( 111) ion by one of the donor atoms in the azo compound. The next step in the reaction sequence is formation of a chelate complex by replacement of a further molecule of coordinated water. Polarization of the ligand as a result of chelate formation markedly enhances its acidity and proton loss ensues (cf. Section 58.2.2.1). Replacement of a further coordinated water molecule by the remaining phenolic oxygen atom in the azo compound gives a complex in which the metal atom is a member of two annelated chelate rings. Proton loss from the chelated azo compound, again promoted by the polarization associated with coordination to the positively charged metal ion, then leads to the 1:l chromium complex

(37). Thus five steps are involved in the reaction between a hexaaquachrornium(II1) ion and an o,o'-dihydroxydiarylazo compound at low pH to form the 1:1 chromium complex dyestuff: three ligand replacement reactions and two ionization stages. Studies with other ligands iead to the conclusion that the probable rate-determining step is replacement of the first of the six coordinated water molecules in the highly symmetrical hexaaquachromium(II1) ion by one of the donor functions of the ligand. For example, it has been shown32that the rate of reaction between the hexaaquachromium(II1) ion and the bidentate oxalate or malonate ions to form monochelate

49

Dyes and Pigments

complexes (equation 4), although slow, is effectively the same as that between the hexaaquachromium(II1) ion and acetate ions to form the monoacetate complex (equation 5), i.e. the second, chelation step is relatively fast. Similarly, it has been rep0rted3~that the first step in the reaction between the hexaaquachromium(II1) ion and glycine to form a chelate complex is slow and determines the overall rate. [Cr(H20)d3+

+

-OC0CO2-

slow __*

[(H20)5CrOCOC02-C

[CrtH20)6]3+ f MeCO;

slow

fast

P-co

[(H20)4C


0015

i/r (K-’) Figure 6 Variation of electrical conductivity with inverse temperature for a series of bis(oxa1ato)platinate salts

NiOP already exists in the Pccn group at room temperature and weak diffuse X-ray scattering and kF and 2kF satellite reflections are ob~erved.’~ The 2kF spots disappear above 305 K and the conductivity is similar to that observed for CoOP and ZnOP?9*72

142

Compounds Exhibititq Unusual Electrical Properties

Although MgOP and CoOP are isostructural above 290 K, their behaviours below this temperature are very different. MgOP undergoes an orthorhombic to twinned monoclinic phase transition at about 283 K, in which the Pt atom chain tilts by 1" towards the b axis.69This does not result in any significant change in either the interchain Pt-Pt separation or in the interchain distance. In fully hydrated crystals the phase transition is also accompanied by the development of satellite spots, indicating the formation of a superstructure. The position of the spots indicates a three-dimensional modulation of the platinum atom chain giving rise to a superlattice with unit-cell dimensions (a, 6 S b , 5.3~).A diffuse X-ray line due to the period of the Peierls modulation of the platinum atom chain and corresponding to a repeat distance along c of 3 . 2 has ~ also been observed in MgOP at room temperature. Thus the periods of the Peierls modulation and the 3D modulation appear to be incommensurate in MgOP, although they are commensurate in CoOP, ZnOP and NiOP. The electrical conduction and thermopower properties of MgOP are quite different from those At- ~the ~ phase transition there is an increase in conductivity of about of the Co, Zn and Ni ~ a l t s . ~ O 10% on going from the high-temperature orthorhombic to the low-temperature monoclinic phase in MgOP, but changes by a factor of up to ten were observed for the other salts. Whereas the changes in conductivity were spread over 30-50 degrees for CoOP, NiOP and ZnOP, the transition occurs over a temperature range of only a few degrees in MgOP. Time-dependent changes in conductivity and thermopower at a constant temperature were observed during the phase change in MgOP at timescales of several hours, whereas no such observations were made with the other salts.70J' The ironGI) salt at room temperature is isomorphous with CoOP but undergoes a transition to the same twinned monoclinic structure with accompanying superstructure exhibited by MgOP.78 The solid state properties of this salt have not been studied in detail. MnOP also exists in the twinned monoclinic form at room temperature. However, the conductivity and thermopower properties as a function of temperature are very different from those of MgOP.72*71 The electrical conduction studies of MnOP show a transition at 225 K, which corresponds to a change in the thermopower from metallic to semiconducting proper tie^.^^ A second transition can be observed at 120 K from the conductivity data but this has only a minor effect on the t h e r m ~ p o w e rStudies .~~ on compressed pellets of Pbo.8[Pt(C204)&zH20 and Cdo.8[Pt(C204)2].nH20have been reported.81 Two models have been developed to describe the superstructure found in these salts. In the Kobayash er al. model, cation ordering in the channels adjacent to the Pt(C204),] chain is responsible for the development of the superstructure and the 3D modulation of the Pt atom chain.79On the other hand, Bertinotti and coworkers have proposed that the chains are fragmented into micro-domains by periodic intrinsic defects associated with polaron^.^^.^^ Three orthogonal deformation modes, one longitudinal and two transverse, are present in each chain and a fourth mode corresponds to a global sliding of the molecular column. As can be seen from-the above discussions, the electrical conduction properties of these compounds are strongly affected by the phase changes and superstructure development observed from X-ray studies. The anomolous behaviour of MgOP has been ascribed to chaos arising from competition between in~tabilities.~~ The interdependence of these, however, is not simple, and a full understanding of these systems is not available. Partially oxidized bis(squarat0)platinum and bis(croconato)platinum salts have been reported but detailed studies have been restricted by the lack of suitable single crystals.84

60.2.3

DihalodicarbonyIiridate~~~

The reaction of K21rCl6with CO under pressure at 170 "C produces a material of gold appearance, which can be recrystallized from acetone solution to yield very fine needle-like golden crystals of variable stoichiometry K[Ir(C0)2Cl,] (x = 2.3-2.7). Cleare and Griffith showed that a series of compounds Y[Ir,(CO),X,], where Y = K+, Cs+, Ph,As+ or Bu,N+ and X = Cl- or Br-, could be readily prepared from the reaction of [IrX6j3- with formic and hydrohalogenic acids.86 However, these materials are unstable and do not readily form single crystals and this has severely restricted both the chemical and physical characterization of these materials. In particular, the Stoichiometry of many of these compounds is controversial and non-reproducible and as yet no complete three-dimensional structure determination has been completed. The chemical stability of the compounds decreases in the order chloro > bromo > iodo. Krogmann and Geserich showed these compounds to be linear metal atom chain compounds with short intrachain Ir-Ir separations of 2.86 A, and that they have related structures to the cation-deficient bis(o~a1ato)platinates.~~ A detailed study of these compounds showed that the

Compounds Exhibiting Unusual Electrical Properties

143

needle crystals are strongly dichroic, exhibiting a metallic lustre in reflected light.88The oxidation state of the iridium ranges from 1.39 to 1.44 and MSssbauer studies show that all the Ir atoms are equivalent. The variation of stoichiometry observed by different workers probably arises from the presence of alkali halides as interstitial precipitates in the lattice and, therefore, material such as Nao.931r(CO)2C12.32 should be formulated as Nao.6r[Ir(C0)2C12]-0.32NaC1.88 The electrical conductivity of several of these compounds has been reported.**Compressed pellet four-probe measurements at room temperature are usually in the ranie 10-*-5 K' cm-' although a value of 150-500 has been reported for orientated polycrystals. Most of the compounds show semiconductor behaviour with activation energies of about 35 meV. K0.60[Ir(C0)2C12]~0.5H20 shows evidence of a transition to more metal-like behaviour near room temperature.88Et seems very likely that if good quality single crystals of these compounds could be obtained then they would exhibit conduction properties similar to those of the cation-deficient tetracyanoplatinates or bis(oxa1ato)platinates. 60.3 NEUTRAL COMPLEXES AND DERIVATIVES Neutral square coplanar complexes of divalent transition metal ions and monoanionic chelate or dianionic tetrachelate ligands have been widely studied. Columnar stack structures are common but electrical conductivities in the metal atom chain direction are very low and the temperature dependence is that of a semiconductor or insulator. However, many of these compounds have been shown to undergo partial oxidation when heated with iodine or sometimes bromine. The resulting crystals exhibit high conductivities occasionally with a metallic-type temperature dependence. The electron transport mechanism may be located either on predominantly metal orbitals, predominantly ligand n-orbitals and occasionally on both metal and ligand orbitals. Recent review articles deal with the structures and properties of this class of compound in detai1.89>90J2 60.3.1 a$-Dionedioximates Square coplanar complexes of the nickel triad with dimethylglyoxime(DMG), diphenylglyoxime (DPG) and benzoquinone dioxime (BQD) have been extensively studied (see Figure 7). There are two basic structural motifs. The Type A structure as typified by a-Pd(BQD);! contains the planar units stacked in columns with the M-M vector parallel to the stacking axis and usually with a repeat distance of 2 units.91The M-M distance is short (3.15-3.40 A)and the packing inefficient, leading to the existence of channels in the stacking direction. Some evidence of anisotropic properties has been observed.92In the Type B structure as shown by Ni(BQD)293the planar units are in a slipped stack arrangement so that the M-M vector is not parallel to the stacking direction. This results in a large M-M distance but more efficient packing and often a herringbone arrangement. All these compounds exhibit semiconductor behaviour with very low conductivities at room temperatures in the metal atom chain direction.89 Under high pressure, Pt(DMG)l has a conductivity of 10 L2-l cm-' (40 kbar).94

(4

@)

Figure 7 a, B-Dionedioximates: (a) R = R = Me, dimethylglyoxime (DMG); R = R = Ph, diphenylglyoxime @PG); (b) benzoquinone dioxime (BQD)

60.3.1.I

Non-integral oxidation state compounds

Conducting halogenated linear metal atom chain systems were first reported in 195095and are now known for Ni and/or Pd complexes of glyoxime (GLY), DPG and BQD. The compounds are usually produced by treating a hot solution of the neutral complex in o-C6H4C12with X2 and lustrous coloured crystals of the NIOS compound are obtained on cooling the solution. The limited range of compounds studied is shown in Table 3 and all have metal-over-metal stacked structures. A complete 3D structure of Ni(DPG)21 has been published and this shows a tetragonal space group with stacks of Ni@PG)2 units staggered by 90". The iodine species are located in channels parallel to the Ni(DPG);! stacks.97The presence of 15- ions has been identified from resonance

144

Compounds Exhibiting Unusual Electrical Properties Tsble 3 Non-integral Oxidation State Metal Glyoximates and Benzoquinone Dioximates Compoulpd

d ~ - M

(A) Pd(GLY)-J Ni(DPG)21 Ni(DPG12Br Pd(DPG)$ Pd(DPG)2Bri.i Ni(BDQ)2I0.018 Ni( BDQ)*1,.,,.0.32CsH,CH, Pd(BDQ),ln ,'0.520ChH,CH C

DC conditions (SI-'

3.244(1) 3.223(2) 3.27 1( 1) 3.26(1) 3.27(1) 3.180(2) 3.153(3) 3.184(3)

(2.3-1) x 10-3 1.8 x 10-3 (7.7-47) x 10-5 (8.9-13) x lo-' dcation.121,122 Slow aerial oxidation of a 50% aqueous acetone solution of H,[Pt{ S$2,(CN),),] and LiCl yielded a black microcrystalline product of small shining black needles and black platelets. Four-probe DC conduction studies on the needle-shaped crystals showed the room temperature conductivity along the needle axis to be -100 Q-' cm-'. The needle-shaped crystals were a cation-deficient compound Lio.~[Pt(S&(CN)2}2]-2H20 (LiPt(mnt)). X-Ray studies show that the room temperature structure of LiPt(mnt) consists of stacks of nearIy eclipsed [Pt(mnt)2]anions along the c axis, with c = 3.639 8, (Figure The unit cell is triclinic and the stacks form sheets along b separated along a by Lif and € 3 2 0 . There are short sulfur contacts between the chains within the sheets as well as within the chain, suggesting a relatively two-dimensional n e t ~ 0 r k . IAlthough ~~ the compound is cation deficient, the long Pt-Pt interchain separation shows that the bonding within the chain cannot be of the simple metal orbital overlap type found for the Krogmann salts (Section 60.2). The high conductivity must therefore result from interaction of the whole anion in the chain direction. The temperature dependence of oll for freshly prepared crystals slowly increases with decreasing temperature, with mIl passing through a maximum at about 250 K.12, The conductivity falls to its room temperature value at about 200 K and then falls rapidly with decreasing temperature. Below 100 K the behaviour is that of a semiconductor with activation energy of about 34 meV. Crystallographic studies have shown a superstructure below T, = 215 K which is preceded by one-dimensional diffuse scattering typical of a Peierls t r a n ~ i t i 0 n .The I ~ ~position of the diffuse line along c indicates that the extent of band filling is 0.41 or 0.59. Studies have shown that the thermopower above T, is approximately constant and positive, implying hole characteristics for the carriers in the metallic region.125Further chemical studies have shown that LiPt(mnt) contains approx. 0.3 protons per platinum in the lattice and, therefore, the extent of band filling is 0.59, in agreement with the thermopower results.126Below T, the thermopower reflects the intrinsic properties of the charge-density wave semiconductor. Magnetic susceptibility studies have revealed the P a d paramagnetism of the conduction electrons above T, and a sharp drop in xp at 220 K coFesponding to the metal-semiconductor transiti~n.'~'Optical reflectivity measurements in the detal atom chain direction show a plasma edge characteristic of a metal with a plasma frequency of -6-7000 Assuming a one-dimensional tight binding approximation to model the electronic structure of LiPt(mnt), this leads to a band width of -0.4 eV, somewhat lower than that derived from the Pauli susceptibility results.125 MO calculations based on the neutral Pt(mnt)2 molecule show that, as expected, the overlap integrals along the face-to-face stacking direction are much larger than the interstack overlap integrals.124The overlap integral of the dz2 orbital in the platinum chain is one-tenth of those found in the partially oxidized tetracyanoplatinates because of the large intrachain platinum separation in LiPt(mnt). On the other hand, the intrastack overlap integrals of the S atoms are of the same order as those of the d,Z orbitals of the tetracyanoplatinates. The estimated band-gap is about 0.5 eV. Preliminary studies on the nickel analogue of LiPt(mnt) indicates that it displays similar electrical conduction properties. 126 On the basis of these extensive studies, LiPt(mnt) behaves as a simple quasi-one-dimensional conductor with a mean-field-like metal-semiconductor transition due to the PeierIs instability.

Compounds Exhibiting Unusual Electrical Properties

149

The short S-S distances between adjacent stacks produces sufficient coupling between the chains to give rise to the almost mean field like behaviour at the t r a n ~ i t i 0 n . l ~ ~ Studies have been made of the effect of changing the central metal, the cation and the substituent on the ligand, on the conduction properties of this class of Unfortunately, although LiPt(mnt) can be obtained as small single crystals, the other systems studied have only been obtained as microcrystalline powders. This has meant that no detailed correlations can be made of structure/chemical composition on the conductivities of these compounds. Attempts to produce analogues of LiPt(mnt) with ligands in which the terminal cyano group was replaced by phenyl, methyl or hydrogen were not successful and only integral oxidation state complexes could be obtained. The electron distributionin the monoanion HOMO is very dependent on the nature of the substituent. Electron withdrawing groups, such as CN, increase the electron density on the metal and reduce it on the S compared with the phenyl substituted ligand. T h i s may aid the intermolecular overlap necessary for the formation of a 1D metal. The black plates which are formed along with LiPt(mnt) in the acid oxidation of H2[Pt{S2C2(CN)2}d/LiCl mixtures in acetone-water have been shown to be Lio,[Pt{ S2C2(CN)2}d.2H20.129 X-Ray studies reveal that the [Pt(mnt)2]0.5-anions are stacked face-to-face along the a axis of the unit cell to form a fourfold distorted linear chain. This can be regarded as the PeierIs distorted state analogous to that found in Rbl 67[Pt(C204)&H20. The room temperature conductivity (all) is -1 ! X I cm-l and the temperature dependence is that of a semiconductor.129 A highly conductingcomplex, ( N B ~ ~ ) ~ . ~ ~ [ N i ( dhas m i been t ) ~ ] prepared , by electrocrystallization. The conductivity in the stacking direction is 10 SZ-' cm-l at 300 K but the temperature dependence indicates a thermally-activated conduction mechanism. A conducting iron complex has also been reported. Compounds of the type Z,[Ni(dimt)~],where x ,(PP~,),~.~ A selective hydrogenation of 1,4-androstadiene-3,17-dioneto 4-androstene-3,17-dione can be achieved with complex (2). The rate of hydrogenation is dependent on both the base added to promote equation (11) and on para substituents in the phosphorus ligand. Thus for the bases, the sequence was found to be Et3N= Et2NH> PhNH?> BuNH2, with pyridine inhibiting the reaction. The effect of substituents in the phosphine was Me0 > Me > H > F9.The further hydrogenation of (4) to the saturated diketone was suppressed relative to the formation of(4) by increasing the hydrogen pressure." With temperature and hydrogen pressure both elevated, complex (2) efficiently hydrogenates both aldehydes" and ketones12to alcohols. Under similar conditions, the reduction of anhydrides to lactones has been reported (equation 13).13

23 3

Catalytic Activation of Small Molecules

(13)

The hydrogenation of both aliphatic and aromatic nitro compounds to the corresponding amines ~ -the ' ~ hydrogenation of sugars, the presence of HCl was found with (2) is also k n ~ w n . ~In necessary to prevent the decarbonylation of aldehyde groups to give [RuH(Cl)( CO)(PPh&], which is inactive." In almost all the above studies, complex (2) was used as catalyst precursor. It possesses the advantage of being air-stable, whereas (3) is Dinuclear complexes of the type [RuH(X)L2I2(X = Cl, Br; L = PP$, AsPh,) are reported to hydrogenate the C=C bond of acrylamide.'* The stoichiometric reaction of (3) with an alkene leads to the ortho-metallated complex (5).19

I PPh3 (5)

This reaction is much too SIOW to play any role in catalytic hydrogenation. In the presence of D2, the ortho positions of the phenyl rings are d e ~ t e r a t e dH-D . ~ exchange between HZ and D2 and between D2 and solvent ethanol is also catalyzed by (2) and (3):' Although certain points relating to the hydrogenation remain unclear, Scheme Z gives the generally accepted me~hanisrn.~,~' [RuH(C1)(PPh3),]+ RCH=CH,

[

PRuH(Cl)(PPh,),

CHR

1

+

.-t

[

1

7 ~ R u H ( C L ) ( P P h 3 ) , +PPh3 CHR

[Ru(CH,CH,R)Cl(PPh,),]

[Ru(CH,CH,R)CI(PP~,)~]+ H,+ PPh,

--L

[RuH(Cl)(PPh,),]

-tREt

Scheme 1

The cis-dihydride complex [Ru(H)~(PP~,),]( 6 ) reacts with one mole of alkene to give a stoichiometric hydrogenation and it catalyzes the hydrogenation of terminal alkenes. The ruthenium(0) species [Ru(PPh,),] is thought to be formed. This then reacts with alkene or hydrogen depending on the conditions. The hydrogen activation step involves oxidative addition to [Ru(PPh,),] or its alkene complex instead of the heterolytic cleavage seen with ruthenium(1I) species.22The nitrosyl complexes [RuH(NO)L,] (L = PPh3, PPh,Pr', PPh&H11) are reported to catalyze the hydrogenation of styrene.23 [RuH(Cl)( v6-C6Me,)PPh3] hydrogenates monoenes, dienes, phenylacetylene and ben~ene.'~

PPh3 (6)

F

low temperature leads to Treatment of [RuH(Cl)(PPh,),] (3) with potassium naphth 1' the isolation of complexes formulated as K"[(PPh,)2Ph2 C6H4Ru(H)2]-C,oH8-Et20 and K2+[(PPh3)3(PPh2)Ru2(€i)4]2-*2C,H,403. These species hydrogenate ketones, aldehydes and esters to alcohols, and nitriles to amines. Acrolein gave either propionaldehyde alone or a mixture with allyl alcohol.25 The same complexes catalyze the partial hydrogenation of some polynuclear aromatic hydrocarbons, but isolated aromatic rings were not reduced.26 Wilkinson and his co-workers introduced another class of highly active hydrogenation catalysts in the form of the hydridocarboxylate complexes [RuH(O,CR)(PPh,),] ( R = Me, Et,

Uses in Synthesis and Catalysis

234

Pr', CF3, C6H,,, 2-OHC6H4). These complexes possess the advantage of being air-stable, unlike the hydridochloro analogue (3), and no base is needed, as with the dichloro complex (2).27In benzene, [RuH(OAc)(PPh,),] (7)shows the same marked selectivity for terminal alkenes as does (3). The mechanism of the hydrogenation is believed to be similar to that for complex (3). In methanolic solution, complex (7)is also active as a hydrogenation catalyst for terminal alkenes, and its activity is enhanced by the addition of a strong acid having a weakly coordinating anion, such as HBF, ?' Species showing similar catalytic properties in acidic methanol were obtained using [Ru(H),( PPh3)J or [Ru(OAc),(PPh,),] directly or by reduction of the trinuclear complex [Ru,O(OAc),(PPh,),] (8)29in the presence of PPh3 or by treating [RuCl,(PPh,),] with AgC10,. It appears that in all cases cationic methanol-solvated ruthenium(11)-triphenylphosphine complexes are formed. The catalytic activity was found to be dependent on the acid concentration.28

'04

I '0'

Me (81

In a more detailed study of the hydrogenation using [Ru(OAc),(PPh,),] (9) in benzene, methanol and methanol containing p-toluenesulfonic acid, the rate passed through a maximum at low acid concentration. This was attributed to the stepwise protonation of the acetate ligands, leading to cationic, methanol-solvated complexes (equation 17). [WOAc?APPh,)J

n+ M~OH. [Ru(OAc)(PP~~~*(MeOH)~]+ & [Ru(PPh,),(MeOH),]Z+

(17)

Benzene or methanol without acid gave much poorer rates. Similar behavior was observed with both conjugated and non-conjugated dienes and these could generally be hydrogenated to the monoenes with very high selectivity. The superior coordinating ability of the diene appears to be the cause of this.30 Using the suffonated phosphine PPh2(m-C6H,S0,H) ( mSPPh2) the complexes [RuC12(mSPPh2),], [RuH(Cl)( mSPPh,),] and [RuH(OAc)(mSPPh,),] have been prepared. These are water-soluble analogues of complexes (2), (3) and (7).In aqueous solution at normal pressure they hydrogenate both C=C and C=O bonds?l A kinetic investigation led to a mechanism analogous to that of Scheme 1being postulated for the hydrogenation of crotonic and pyruvic acids. Carbonyl complexes of ruthenium have been mentioned briefly above. They are dealt with in detail here. [RuCl2(C0),(PPh,),] (10) in dimethylacetamide has been shown to reduce the C=C bond of methyl vinyl ketone at 80 "C and normal pressure of hydrogen. AIthough 1-octene was At higher temperature and not hydrogenated it was efficiently isomerized to 2-, 3- and 4-0ctene.~~ pressure, I. ,5,9-cyclododecatriene was hydrogenated with very good selectivity to cyclododecene with a turnover number (moles product-fomedjmoles catalyst used) of 32 Other ruthenium complexes were investigated but were less efficient. Rates of hydrogenation of a series of alkenes

co

PPh,

PPh3

PPh,

(10)

(11)

with (10) were found to vary in the order conjugated dienes > non-conjugated dienes> terminal The mechanism probably involves a hydrogenolysis step, as for alkenes > internal [RuCl,(PPh,),] (equation 18). [RuCI~(CO),(PP~,)J+H,

--c

[RuH(CL)(CO)z(PPh3)2] fHCl

(18)

Catalytic Activation of Small Molecules

235

At 160-200 "C and 1.5 MPa hydrogen pressure, (10) is an efficient catalyst for the hydrogenation of simple aliphatic and aromatic aldehydes to alcohols, giving turnover numbers of 10 000-95 [RuHCl(CO)( PPh3)3](1 1) has been shown to reduce propionaldehyde, heptaldehyde, benzaldehyde and acetone to alcohols. The hydrogen pressure used here was significantly higher than in the above studies. [RuCl,(PPh,),] and [RuH(Cl)(PPh,),] were also about as effective as (ll), but [Ru(CO),(PPh,),] was less In a further study, acetone and other ketones were reduced to alcohols by the four complexes above and by [RuH(NO)(PPh,),], [RuCI,(CO),(PFhJ,], [Ru(H),(CO)(PPh,)J (12), [Ru(H)~(PP~,),] and [Ru(H),(PPh,),] (6). The best results were obtained with complex (ll),the nitrosyl complex and the dichloride (10). The addition of water to these reactions was found to be beneficial in most cases?' Use of the same complexes to catalyze the hydrogenation of propionaldehyde showed that [RuH(Cl)(PPh,),] (3) was the most effective. However, complex (11) was preferred for general use because of its stability in air.38Turnover numbers up to 32 000 were achieved. Complex (11) decomposed during the reactions to give a variety of species which could not be characterized. The hydrogenation of heptanal and cyclohexanone is also catalyzed by the complexes [Ru(O2CCF,),(CO)(PPh,),1 (13), [RU(O~CCF~)~(CO)(PP~,)(P~,PCH~CH~ASP~~)] (14) and [Ru(O2CCF3),(C0)(Ph2PCH,CH,CH2PPh2)] (15). Complex (13) formed a significant amount of the dicarbonyl [Ru(O,CCF,),(CO),( PPh,),] during the reaction, whereas with (14) no other complex could be detected. Complex (15) was the most active catalyst and, here again, a number of uncharacterized complexes were formed.39The proposed mechanism for the hydrogenation is shown in Scheme 2. These complexes also catalyze the dehydrogenation of alcohols. CF3

1 P%# Ph3P'

I w. ao R "** I l C 40 PPh3

F3C \ ,"-0

CF3

/ cI '0

PPh3 '04@// I@CO

' 0

I

04&

Rrt,,hz O

Ph,Aq,)

CF3

CF3

I

I qco

O4 RrbPPhz y P h z P d

CF3

I 13)

(15)

(14)

Ru( O,CCF,),CO( PPh,)(Arphos)

CF,CO,H RuH(O,CCF,)CO(PPh,)( Arphos)

IPPh3

R' \

/

R

c=o

OCWRR' I

7

Ru(H),(O,CCF,)CO(Arphos)

c It

0

RuH(O,CCF,)CO( Arphos) RR'CHOH

R \ /

l!uH(O2CCF3)CO(Arphos)

H2

A

!

OCHRR 1 Ru(O,CCF,)CO(Arphos)

Scheme 2

In addition to reducing C=O bonds, the dihydride (12)also hydrogenates 1-hexene., At normal pressure and 45 "C (12) acts principally as an isomerization catalyst. At 1.7 MPa, hexane is formed but 2-hexene remains the major product. The ruthenium hydrogenation catalysts described above all contain phosphorus ligands. Some phosphorus-free catalytic systems have been reported. Thus the oxygen-centred trinuclear acetate [Ru,O(O,CMe),(H,O),j+ (16), whose structure is analogous to (81, hydrogenates alkenes in DMF solution at normal pressure of hydrogen and 80 0C.40The catalysis shows a number of unusual features. Thus internal alkenes are hydrogenated more rapidly than terminal ones. The rate order is cyclic alkene > internal alkene > terminal alkene, Terminal alkynes are reduced even more

236

Uses in Synthesis and Catalysis

slowly. The phosphine complexes of ruthenium generally show a marked preference for terminal alkenes. In hydrogenations with (16) it was also noted that no isomerization products could be detected. The reactions are generally autocatalytic. [HRu,O(O,CMe),]+ is thought to be first formed, followed by [Ru,O(O,CMe),]+, both of which are solvated by DMF. The latter complex is responsible for the bulk of the hydrogenation, which involves only one ruthenium atom in the complex. Activation of hydrogenation precedes alkene coordination. Chloro complexes of ruthenium(11) were found to hydrogenate maleic and fumaric acids to succinic acid slowly at 60-80 "C and normal pressure of hydrogen. Non-activated alkenes lead to the production of ruthenium metal. The structures of the species involved are unknown. The mechanism involves coordination of the alkene followed by heterolytic cleavage of hydrogen, giving a ruthenium(I1) hydride as the second step.41 In N, N-dimethylacetamide solution the reduction by hydrogen of ruthenium(TI1) chloride is claimed to produce ruthenium( I) complexes which hydrogenate ethylene, maleic and fumaric acids. The complexes are thought to be dimeric but their precise structures are unknown. Interestingly, these d ruthenium(I) complexes are believed to activate hydrogen by oxidative addition whereas heterolytic cleavage of hydrogen occurs with most ruthenium catalysts (equation 19). Ru1+H2 + Ru"'(H),

(19)

Hydrogen activation precedes the coordination of the alkene, Reviews on hydrogen activation ~ ~ * ~ reviews on homogeneous hydrogein relation to homogeneous catalysis are a ~ a i l a b I e .General nation are listed at the end of Section 61.2.2. 61.2.2.2

Cobalt

The pentacyanocobaltate(11) ion has long been known to catalyze alkene hydrogenation, mainly The catalyst system shows negligible of conjugated dienes. A review of the early work is a~ailable.4~ activity for the hydrogenation of non-activated monoenes. A major disadvantage is that the system is inhibited by excess substrate, and the turnover numbers obtained are generally less than 2. The activity and selectivity observed in these systems are markedly solvent dependent. In water alone, deactivation eventually occurs. This has led to the use of mixtures of water and organic solvents, OF to polar organic solvents alone. For hydrogenation studies the catalyst is generally prepared in situ by mixing the appropriate quantities of a cobalt salt and a Group I cyanide. The resulting green solution is generally considered to contain the ion [Co(CN),I3-, but it has been pointed out that the colour is more probably due to the presence of [ C O ( C N ) ~ H ~ O ] ~ - . ~ ~ In the hydrogenation of cyclopentadiene to cyclopentene it was found that ethanolic solutions retained their catalytic activity for months, whereas aqueous solutions rapidly lost their activity. Fresh solutions in either solvent gave similar initial rates.47The product of the hydrogenation is dependent on both the solvent and the CN/Co ratio. In the hydrogenation of butadiene in aqueous solution the major product was trans-2-butene at CN/Co= 5.1 and l-butene at CN/Co> 5.1. The total concentration also affects the product distribution. In glycerol-methanol solution similar results were observed, the change occurring at CN/Co = 5.4. In ethylene glycol-methanol solution the production of cis-2-butene was favoured.48 The hydrogenation of trans- 1 -phenyl-l,3-butadiene has been discussed49in terms of the mechanism originally proposed>5 and v3-allylspecies were invoked. NMR studies on the hydrogenation of butadiene have shown that q1-but-2-enyl and q3-l-methylallyl complexes are present in the By using a reaction mixture. With isoprene, an v1-methylbuf-2-enyl complex was ob~erved.~' two-phase system and a phase-transfer catalyst, the turnover numbers attainable with conjugated dienes can be increased to about 30.51 In addition to conjugated dienes, cyanocobalt catalysts also hydrogenate the C=C bond of activated alkenes.'2 Carvene, mesityl oxide, 2-cyclohexenone, benzalacetone and an androstenone derivative were reduced in this way.53 Hydrogenations with [Co(CN),I3- as catalyst have been long thought to involve radical reactions. Hydrogen is activated prior to reaction with the alkene, the mechanism involving homolytic fission by two [CO(CN)~],-ions (Scheme 3).54 The transfer of hydrogen to the alkene also occurs by a radical and not by the insertion reaction which is much more common in homogeneous hydrogenation. These kinetic studies were carried out using cinnamic acid as alkene. The observation of alkenyl and allyl species in diene hydrogenation suggests that a different mechanism of hydrogen transfer may

Catalytic Activation of Small Molecules [Co(CN)J3-+H,

4

[CoH(CN)J3-+H.

[Co(CN),]’-+H*

+

[CoH(CN),I3-

[CoH(CN),I3-+ X

+

[Co(CN)J3-+ HX.

-+

[Co(CN),]’-+H,X

[CoH(CN),]’-+ HX.

237

(X = conjugated diene, activated alkene)

Scheme 3

operate there, as these reactions appear slow enough for intermediates to be observed directly by NMR. Modification of the [CO(CN),]~-catalyst by addition of diamines has been reported to lead to higher catalytic activity in diene hydrogenation and a predominant 1,2-addition of hydrogen. Such systems are thought to contain the species [CoH(CN),(diamine)]-. Ethylenediamine, 2,2‘bipyridyl and 1,lO-phenanthroline were the diamines sed.^^*'^ Bis(dirnethylglyoxirnato)cobalt(II) (17), generally in the presence of a base such as pyridine, has been studied as a potential analogue of the cobalt-containing vitamin BIZ.The base coordinates in the axial position, giving a square pyramidal structure. These complexes reduce the C=C bond of activated alkenes at normal temperature and hydrogen pressure.’’ As with the [Co(CN)J3system, the attainable turnover numbers are very limited. With methyl methacrylate the pyridine adduct of (17) gave a turnover number of 10. The hydrogenations are generally carried out in alkaline media and there appears to be an optimum basicity for a given alkene!’ The active species is thought to be [CoH(DMG),(base)] or [Co(DMGH),(base)]-, where [Co(DMGH),] represents complex (17).

/

0, (B = morpholine)

9

6 (18)

In presence of morpholine, (17) catalyzes the hydrogenation of nitrobenzene to aniline.61 A near twofold excess of the amine relative to (17) was used. At a higher excess, catalysis ceased. This was attributed to formation of a bis-morpholine complex. The hydrogenation proceeds best in solvents such as acetone, ethyl acetate or THF.6’ The transfer of the first hydrogen atom to nitrobenzene was found to be rate determining and a dinuclear nitrobenzene complex (18) was postulated. The base (B) is morpholine. In presence of pyridine, (17) catalyzes the reduction of activated alkenes by both hydrogen and NaBH,. Vitamin BIZalso catalyzed these reactions.63 Reduction of a cobalt( 11) halide in presence of 2,2’-bipyridyl with zinc in THF-ethanol leads to cobalt(1)-bipyridyl complexes which hydrogenate butadiene to cis-2-butene at 25 “C and normal pressure of hydrogen. For different halides the rate decreases in the order I>Br>Cl. 1,lOPhenanthroline complexes were also active.64Here again, the catalyst does not tolerate an excess of diene. The proposed mechanism for the hydrogenation is given in Scheme 4.

Scheme 4

The triphenylphosphine complexes [CoX(PPh,),] (X = C1, Br, I) exhibit fairly high selectivity in the hydrogenation of dienes. Internal double bonds were reduced preferentially. The addition

238

Uses in Synihesis and Catalysis

of a Lewis acid to activate the complexes was necessary. The activation is thought to involve the removal of the halide ligand and formation o f a diene complex (equation 24).65 [CoBr(PPh,),] +diene+BF,.OEt,

+

[Co(diene)(PPh,),]+ [BF,Br]-+ Et,O

(24)

The complex [CoH(N,)(PPh,),] (19) in presence of sodium naphthalide gave a paramagnetic species which hydrogenates styrene at normal temperature and hydrogen The kinetics of the hydrogenation of cyclohexene by (19) were studied and the mechanism given in Scheme 5 was deduced, in which two alternative routes were considered. It was not possible to distinguish between them on the basis of the kinetic e~idence.6~ N

CoH(N),(PPh,),

alkene

CoH(alkene)(PPh,),+ N,

I I

5 CoH(alkene)(PPh,),+

Co(H),(alkene)(PPh,),

I

alkanc

PPh,

alkene

CoH(alkene)(PPh,),+alkane

R

d Co(alkyl)(alkene)(FPh,)2

Scheme 5

In a further kinetic study on the hydrogenation of cyclohexene, [Co(H),(PPh,)J was used as catalyst:' The mechanism is thought to involve dissociation of a PPh, ligand to permit alkene coordination. The hydrogenation leads to a cobalt( I) species formulated as [CoH( PPhJ2], probably solvated, which then oxidatively adds hydrogen to regenerate the trihydride. [CoH( BH,)( PCy3)J in benzene solution hydrogenates terminal alkenes and styrene faster than internal alkenes or dienes. Isomerization of terminal alkenes caused a rapid deterioration of the reaction rate.6g An ortho-metallated triphenyl phosphite complex, for which the structure (20) was proposed, gave a turnover number of more than 300 in the hydrogenation of 1-butene. Isomerization was not observed in this case.7o Phosphinecarbonyl complexes of cobalt have long been known to act as hydrogenation catalysts. In a recent study involving cyclohexene the kinetics of its hydrogenation by the complex [CoH(CO),( PBun3)J were studied. Unlike the systems described above, the carbonyl complexes generally require elevated temperature and pressure. The proposed mechanism is given in Scheme 6. CoH(CO),(PBu",),

I/

CoH( CO),PBu",

alkane+CoH(CO),PBu", y,

+ PBu",

A CoH(alkene)(CO),PBu",

\

,p% /;,,

\

'CoH( H),(alkene)(CO), PBu", Scheme 6

In many studies, [Co,(CO),] was treated in situ with phosphorus ligands to obtain the catalytic species. Such systems lie outside the scope of this work and the companion volumes 'Comprehensive Organometallic Chemistry' should be consulted. Cobalt catalysts have been shown to hydrogenate arenes to saturated hydrocarbons. The Co(acac),-AlHBui,-PBu"3 system hydrogenates benzene to cyclohexene, but the presence of styrene was necessary, otherwise the reaction ceased. The styrene was also hydrogenated." Muetterties and co-workers have reported the hydrogenation of arenes catalyzed by [Co{P(OMe),},( q2-C,H,)] (21) at room temperature and normal pressure of l1ydrogen.7~The

Catalytic Activation of Small Molecules

239

product of benzene hydrogenation was cyclohexane, no cyclohexadienes or cyclohexene being detected. Using deuterium, it was shown that cis addition of all three molecules of deuterium occurred. The low activity of the triphenyl phosphite complex could be raised considerably by using more bulky phosphites or phosphines, but the lifetime of the catalyst is reduced. Surprisingly, arenes were reduced at rates similar to those of alkenes, whereas most hydrogenation catalysts react far more rapidly with alkenes and benzene can often be used as a solvent for alkene hydrogenation.

“i“ (MeO),p/‘Pp(OMe), P(OMe), (21)

The attainable turnover numbers lie in the range 20-40. Complex (21) hydrogenates many arenes and alkenes though the conversions are often The crystal structure of the q 3 cyclooctenyl analogue [Co{P(OMe),},( v3-CSHI3)](22) has been determined. Considering the v3-unit as a single ligand, the molecule is pseudotetrahedral. This complex also catalyzes arene hydrogenation.” Alkyl-substituted arenes undergo H/ D exchange in the alkyl group during arene hydrogenation. The v3-cyclooctenyl group of (22) was removed by hydrogenolysis early in the reactions and the coordinatively highly unsaturated hydride [CoH{P(OMe)&] is thought to be involved.76Scheme 7 shows the proposed mechanism. /

D

co

co

?

D. V”

I

Y

Scheme 7 Phosphite ligands omitted

61.2.2.3

Rhodium

Homogeneous hydrogenation by rhodium complexes is perhaps the classic example occurring in catalysis of activation of the hydrogen molecule by oxidative addition to d s rhodiurn(1) complexes (equation 3). The relatively high H-H bond energy (450 kJ mol-’) is believed to be overcome by the donation of electron density from a filled metal d-orbital to the empty u* antibonding orbital of the hydrogen molecule. Very often, this oxidative addition is highly reversible at normal temperature and hydrogen pressure and the equilibrium can be displaced to either side almost instantaneously by introducing or removing hydrogen. The discovery of the complex [RhCl(PPh,),] (231, now known as Wilkinson’s catalyst, represenIt is a highly active catalyst ted a tremendous step forward in homogeneous for the hydrogenation of alkenes and, unlike the cobalt complexes, it remains active at high

Uses in Synthesis and Catalysis

240

alkene/catalyst ratios." Its introduction led to greatly increased interest in both hydrogenation and homogeneous catalysis generally. The complex may be obtained as either red or orange crystals, the red form being the more common one. Both forms have a near square planar structure which is slightly distorted towards the tetrahedral, The distortion is somewhat more pronounced in the red form, There are close interactions between ortho hydrogen atoms of the phenyl groups and the metal, which are different in the two forms." In their original studies, Wilkinson and co-workers postulated a mechanism involving oxidative addition of hydrogen to rhodium( I) to give a rhodium( 111) dihydride, followed by coordination of alkene. The precise nature of the hydrogen transfer to the alkene was not clear at that time, but the general outlines of the mechanism were correct. There have been may inconclusive kinetic studies of the hydrogenation, but the detailed mechanism has now been elucidated by Halpern and his ~ o - w o r k e r s , and ~ ~ - is ~ ~summarized in Scheme 8.

p%Rh@

'

P

c1."".Rh@p

C 'l

*IjI

P' I

I

\ I ,c=c \ / ,

Catalytic cycle Scheme 8

The actual catalytic cycle is enclosed in the rectangle. The species outside this can also lead to hydrogenation, but at a much slower rate. The dissociation of one PPh3 ligand forming [RhCl(PPh,),] leads into the catalytic cycle. Further ligand dissociation, not shown in Scheme 8, may occur. When the more strongly coordinating styrene is hydrogenated (Scheme 8 relates to cyclohexene), additional steps occurred in the reaction, involving the formation after hydrogenation of a bis-styrene complex [Rh(H),CI( PPh3) (~tyrene),].~'Halpern has discussed the difficulties involved in obtaining a reliable mechanistic understanding of such a complex system?' In the 18 years since its discovery, Wilkinson's catalyst has been used to hydrogenate all kinds of unsaturated compounds and great efforts have been made to develop other phosphorus ligand-containing catalysts of both rhodium and other metals. This in turn has led to catalytically active complexes which do not contain phosphorus. In Table 1 the hydrogenation of various unsaturated compounds catalysed by [RhCl( PPh3),] is summarized. In some cases the addition of deuterium and tritium to the substrate was i n ~ e s t i g a t e d . ~ ~ ~ ~ ~ ~ ' ~ ~ ~ ~ ~ ~ Osborn and co-workers introduced a series of cationic rhodium-diene complexes of the general type [Rh(diene),(ph~sphine)~]+ X- (X- = non-coordinating anion).lo3-IMNorbornadiene was the preferred diene, although 1,5-cyclooctadienewas also used. PPh3,PPh,Me, PPhMe, or a chelating diphosphine were generally the phosphorus ligands and the anions were typically C104- or PF6-. These complexes were shown to hydrogenate alkenes to alkanes,Io4dienes to monoeneslo6 and alkynes to ci~-alkenes.'~~ Complexes of this type in which chiral diphosphines are present have become most important in asymmetric catalytic hydrogenation (see Section 6 1.2.3). Under hydrogenation conditions the diene is reduced to give cationic solvated rhodium(1)-phosphine complexes which initiate the

Catalytic Activation of Smdl Molecules

24 1

Table 1 Applications of [RhCl(PPh,),] in the Catalytic

Hydrogenation of Unsaturated Compounds Substrates

Ref:

Ethylene, cyclohexene Cycloalkenes Alkenes, cycloalkenes Methylenecyclohexene derivatives Allene Terpenes Triterpenoids Perhydroaromatics, butadiene rubbers Cyck~hexenolderivatives Cycloalkenes, activated alkenes Alkenes, cycloalkenes, dienes, conjugated dienes, activated alkenes Cyclohexene, unsaturated acids Monoepoxybutadiene, phenylcyclopropane, styrene Unsaturated thiophene derivatives Nitroaromatics Activated alkenes, nitrostyrene derivatives Lipids

5 84 85 86,87 88

89 90 91 92 93 94 95 96 91 98 99 100

catalytic cycle. The triphenylphosphine complexes then activate hydrogen by oxidative addition whereas with diphos as ligand, no reaction with hydrogen to give [Rh(H)2(PPh,),(soivent),]', occurs in absence of a1ker1e.l'~From solutions of the solvated species [Rh(diphos)]:+ a dinuclear complex has been isolated, which reverts to the mononuclear complex on redissolving. Some other hydrogenations catalyzed by cationic rhodium complexes of this type are summarized in Table 2. Table 2 Some Other Hydrogenations Catalyzed by Cationic Rhodium-Phosphine Complexes Complexes

Subsrrates

Re$

3-Phenylbut-3-en-2-01, 4-phenylpent-4-en-2-01 Unsaturated esters

108

1-Hexene, cyclohexene, benzophenone Alkenes, cycloalkenes, ketones, cyclic ketones Ketones Ketones, aldehydes

110

109

111 112 113

Ketones, aldehydes Aldehydes Styrene oxide

114 115 116

3,4-Epoxybut- 1 -ene

117

In addition to the hydrogenation of C-C multiple bonds, such cationic species have now been shown to reduce ketones, aldehydes and epoxides. The mechanism of hydrogenation of a C=O bond is believed to be analogous to that for alkenes, with the insertion step leading to an alkoxide complex (Scheme 9).lt3 The implication of metal hydrides in homogeneous hydrogenation led to their direct use as catalyst precursors. [RhH(PPh,),] (24) was shown to hydrogenate 1-hexene with accompanying isomerization to 2-hexene. Dissociation of PPh, occurred to give coordinatively unsaturated species."* These underwent oxidative addition to give rhodium(II1) trihydrides, this being the rate determining step. Alkene insertion and reductive elimination of hexane complete the catalytic cycle, regenerating a rhodium( I) monohydride. In the hydrogenation of ethylene with complex (24), however, it was believed that a rhodium(1)-ethyl complex was first formed, which then reacted with hydrogen."' In the presence of an equimolar amount of PEt,, complex (24) effected

242

Uses in Synthesis and Catalysis

RzcH0H7 I

I -CHR,

(P = phosphorus ligand, S = solvent) Scheme 9

the hydrogenation of 1,3-dienes to 1-alkenes.'" Another mixed phosphine catalyst which reduces terminal alkenes is [€UIH(PF,)(PP~~),].~~~

(24)

(25) DBP

The catalyst system [RhCl(nbd)],/ PPh,/ Et,N has been shown to hydrogenate ketones under mild conditions.'*' Many examples were given. The catalytic activity was again attributed to rhodium( I)-hydrido-phosphine complexes generated according to equation (25) after hydrogenation of the coordinated diene. [Rh(H),CI(PPh,),]+Et~N S [RhH(PPh,),]

+ Et3NH+C1-

(25)

The 5-phenyldibenzophosphole (25) complex [ RhH(DBP),] catalyzes the hydrogenation of terminal alkenes, lJ-hexadiene and some substituted alkenes. With 1-hexene it was far more active than [RhCl(PPh,),] or [RhH(PPh,),]. The mechanism is again thought to involve dissociation of DBP followed by alkene insertion into the rhodium hydride bond and subsequent reaction with h ~ d r 0 g e n . l ~ ~ The hydrido complex [RhH(PPr',),] and [Rh,(N,){P(C,H,,),},] have been shown to hydrogenate nitriles to primary amines at normal temperature and pressure, with turnover numbers of up to 100 being attained.'24 This is a most interesting observation in view of the coordinating ability of primary amines, which might have been expected to saturate the coordination sphere of the rhodium, putting an end to catalysis. In the homogeneous hydrogenation of alkynes it has generally been found that cis-alkenes are formed exclusivelyin the hydrogenation step, any trans-alkene occurring being due to a subsequent isomerization. It has now been reported that [Rh,(H),{P(OPr'),},] (26) converts alkynes directly to trans-alkene~."~ The dinuclear unit is thought to remain intact during the catalytic cycle. From the reaction of (26) with di(p-tolyl)acetylene, a vinyl complex (27) was isolated in which the trans-alkene was present. P P

PPr',

(27)

The intermediates formed in these reactions are unstable in presence of excess of the alkene, and catalysis rapidly ceases. The carbonate complex [Rh(H)2(PPri3)2( q2-0,COH)] (28) catalyzes

Catalytic Activation of Small Molecules

243

the hydrogenation of diphenylacetylene to trans-stilbene.'26Here again, the trans isomer is believed to be formed directly. Complex (28) appears to be a more efficient catalyst for such reactions than (26), but details of the turnover numbers that can be achieved were not given.

L

PPh3

I

(29)

By addition of carboxylate anions to solutions of cationic rhodium-triphenylphosphine complexes under hydrogen, complexes of the type [Rh( H)2(02CR)(PPh,)J have been isolated which have the structure (29). In benzene-methanol solution these complexes reduce the C=C double bond of activated alkenes.'27Carboxylate complexes of the type [Rh(O,CR)(PPh,),], which may be regarded as precursors of (29), have been prepared by addition of PPh, and a carboxylate anion to an acidic solution of [ R ~ , ( O A C ) ~ ]They . ~ ~ *hydrogenate 1-hexeneand 1-hexyne in benzene solution but at lower rates than [RhCl(PPh,),]. Several carboxylate anions were used, the best results being obtained with the acetate. [Rh(OAc)(PPh,),] (30) in methanol solution gave rates superior to those obtained in benzene, and in presence of p-toluenesulfonic acid even better rates were obtained.,' This was true for terminal, internal and cyclic alkenes and for both conjugated and non-conjugated dienes. By varying the acid/acetate ratio, a wide range of dienes could be hydrogenated to the monoene with high selectivity. The acid dependence was thought to be due to conversion of complex (30)to cationic species by protonation of the acetate group. Complex (30) is more easily prepared from [RhC13.3H,0] directly30 than by the original method."* Another variation on the structure of [RhCl(PPh,),] is the complex [Rh(PPh,),( s3-C3H3>], which has been shown to reduce 1-octene without isomerization occurring."' Water-soluble hydrogenation catalysts have been obtained using ligands such as (31)-(33) which contain strongly polar groups. 0

S03Na

II

N(CH2CH2PPh2)2

(31)

P~,PCH,CH,&M~I~~NO,-

(33)

(32)

[RhC1,.3H20] in presence of (31) was shown to hydrogenate cyclohexene in 1: 1 mixtures of water and a polar organic solvent.130Similarly, rhodium complexes of (32) and ligands of related structure reduced activated a1l~enes.I~' [RhCl( mSPPh,),] ( mSPPh, = 31) hydrogenated 1-hexene, 2-hexene and cyclohexene in a two-phase system, the complex remaining in the aqueous phase.I3* [Rh(nbd)(amphos)J3* (amphos = 33), also in a two-phase system, reduced 1-hexene and styrene.',, Although carbonyl groups generally tend to deactivate rhodium hydrogenation catalysts, a number of complexes have been studied. [RhH(CO)(PPh,),] (34), which is better known as a

(35)

(34)

(36)

hydroformylation catalyst (see Section 61.2.4.4), also hydrogenates alkenes at normal temperature and hydrogen pressure.'34 The complex shows a marked preference for terminal double bonds. Alkynes, conjugated dienes and internal and cyclic alkenes were not hydrogenated under these conditions. Following a kinetic investigation of the hydrogenation of 1-hexene and 1-decene, the mechanism shown in Scheme 10 was proposed. The catalyst slowly loses its activity owing to dimerization according to equation (26). 2[RhH(CO)(PPh,),]

CCC6-I

+

[Rh(CO)(PPh,),],+ZPPh3+ H,

(26)

Uses in Synthesis and Catalysis

244

PPh, OC-Rh--H

4 %

PhoP

PPh3

It PPh3

I

OC-Rh-H I

RCH,CH,

PPh3

w=CH,,,,

PPh3 I

CH2CH2R

PPh3

I 1

OC-Rh-CHZCH2R PPh3 Scheme 10

In the hydrogenation of ethyl acrylate, it was shown that deactivation also occurred, but irradiation with a weak source of UV light led to reactivation and complete hydrogenation to ethyl Using the ligands PhzPCH2PPhz(dpm) and Ph2AsCH2PPh2(dam) a series of complexes of the type [Rh,X(CO),L,]' (X = C1, Br; L = dpm, dam) were introduced. These complexes have the so-called 'A-frame' structure illustrated in (35) for [Rh,Cl( CO),(dpm),]+. Terminal, internal and cyclic alkenes and terminal and internal alkynes were hydrogenated. Complex (35) appears to be the better catalyst, followed by its dam analogue. However, for the reduction of phenylacetylene to styrene, [FUI~(CN),(CO),(~~~),] was far more active than the other c~mplexes.'~' The nitrosyl complex [Rh(NO)(PPh,),] (36)catalyzes the hydrogenation of both 1-hexene and cyclohexene in dichloromethane as solvent and was also found to add deuterium to cyclohexene without H/D ~crambling.'~'A further study extended the range of substrates hydrogenated to internal alkenes, to conjugated and non-conjugated dienes, activated alkenes and terminal and internal a 1 k ~ n e s . l ~ ~ Nitrogen ligands have played a minor role in homogeneous hydrogenation compared with those of phosphorus, but some such catalysts are known. The 1,lO-phenanthroline complex (37)has been shown to hydrogenate cycl~hexene.'~~ H

(37)

1,lO-Phenanthroline and 2,2'-bipyridylcomplexes have been studied in greater detail by Mestroni and c o - w o r k e r ~A. ~series ~ ~ ~of~ ~ ~ complexes of the type [Rh(chel)(HD)]X (chel=bipy, phen, 4,7-Me2-phen, 5,6-Me,-phen, 3,4,7,8-Me4-phen;HD = 1,5-hexadiene; X = PF6-, BPh,-) were prepared. These complexes react with hydrogen to give solvated species, the hexadiene undergoing hydrogenation (equation 27; S = solvent). [Rh(chel)(HD)]++H,

s

[Rh(chel)S,]++ hexane

(27)

In contrast to the phosphine analogues already mentioned, reaction (27) does not occur with the cod and nbd complexes. Further, reaction (28) occurs quantitatively on addition of cod to the hexadiene c o m p l e x e ~ . ' ~ ~

Catalytic Activation of Small Molecules [Rh(chel)(HD)]++cod

--t

245

[Rh(chel)(cod)]++ H D

(281

In alkaline solution, the rhodium-hexadiene complexes hydrogenate ketones at norma1 temperature and hydrogen pressure. Carbon-carbon double bonds are also reduced. An excess of the nitrogen ligand was found to be advantageous in some cases and the ion [Rh(bipy)J+, which is presumably solvated, reduced a range of ketones almost quantitatively. When equimolar mixtures of cycloalkenes and cyclic ketones were used as substrates in alkaline methanol, [lU~(bipy),Cl~]+ gave almost exclusive hydrogenation of the ketones. On the other hand, the C=C bond of ~ ~ ~ most unsaturated esters and ketones was preferentially reduced by [R h ( b i ~ y ) S , ] + .ClearIy interesting selectivities can be achieved with these catalysts, ~ H 2-aminopyridine ~,)~] led to a species believed to be a cationic The reaction of [ ~ U I ~ C ~ ~ ( Cwith solvated rhodium(1)-arninopyridine complex. This was more active than [RhCl(PPh,),] or [RuH(Cl)(PPh,),] for the hydrogenation of cyclohexene. The mechanism was thought to involve oxidative addition of hydrogen to rhodium(I) prior to alkene c ~ o r d i n a t i o n . ' ~ ~ James et al. have shown that dialkyl sulfide complexes of rhodium can act as homogeneous hydrogenation catalysts. The complex [RhCl,(SEt,),] in DMA solution is believed to react with hydrogen according to equations (29)-(31). [RhC13(SEtZ)3]-+ [RhCI,(SEt,),]+

(29)

EtzS

[RhCl,(SEt,)J+ H,

-+

[Rh(H)CI3(SEt,),]-+ H+

(30)

[RhCI,(SEt,),]+

--t

[Rh(H)CI,(SEt,)J+ Hf+C1-

(31)

Hz

The rhodium(II1) hydride then loses a proton to give a rhodium(1) complex which after coordination of the alkene oxidatively adds hydrogen. Ethylene, cinnamic acid and maleic acid ~ ) ~also ] an active catalyst for maleic acid hydr0genati0n.l~~ By were reduced.'" [ I U I C ~ ~ ( S B Zwas addition in situ of Et$ to the cyclooctene complex [RhCl(C,H,,),],, species were obtained which exhibited the same kinetics as the rhodium(II1) complexes in maleic acid h~dr0genation.I~~ The complex [Rh2(pSBu'),{P(OMe),),1 and related complexes with mixed thiol bridges and with other phosphites have been shown to hydrogenate cy~lohexene.'~' OAc),] reduces terminal and cyclic alkenes, activated alkenes The rhodium(11) complex [a2( and alkynes.I4* Various polar solvents could be used, but DMF was preferred. Following a kinetic investigation of the hydrogenation of 1-decene, the mechanism shown in equations (32)-(35) was proposed. [Rh,(OAc)4]+ H, [Rh,H(OAc),]+alkene

-3

[Rh,H(alkene)(OAc),] [Rh,(dkyl)(OA~)~] + H++AcO-

-+

[Rh,H(OAc),]fH++ AGO-

(32)

[Rh,H(alkene)(OAc),]

(33)

[Rh,(alkyl)(OAc),j

(341

[Rh,(OAc),]+ alkane

(35)

Unfortunately it is not known whether hydride and alkene coordinate to the same or different rhodium atoms in the dimer, but the former appears more probable in terms of our general mechanistic knowledge of hydrogenation. The obsemation of alkene isomerization during the reaction indicates that equation (34) is reversible. Some other rhodium homogeneous hydrogenation catalysts and the substrates they reduce are collected in Table 3. Two reviews which concentrate on rhodium-catalyzed hydrogenation have Table 3 Some Other Rhodium Hydrogenation Catalysts Complex

[RhH(C,B,H,,)(PPh,)2] ~RhZCI,(CO)41 [Rh(cod)L,]C1O4 L = MeCN, PhCN [Rh(cod)L']ClO, L' = succinonitrile Rh amino acid complexes CRhCI(HD)I, HD = 1,s-hexadiene

Substrate

ReJ

1-Hexene 1,3-Pentadiene 1-Hexene, cyclohexene, dienes

149 150 151

1-Hexyne Arenes Arenes, substituted Arenes, aromatic heterocycles

152 153

246

Uses in Synthesis and Catalysis

61.2.2.4 Iridium In acetic acid solution, [Ir(H),(PPh,),] (38) catalyzes the hydrogenation of 1-butanal to n-butanol at normal pressure of hydrogen and slightly above room temperature. Simple alkenes were not reduced under these conditions, but the C=C bonds of acrylic acid and methyl acrylate were. Acetate complexes were believed to be responsible for the ~ a t a l y s i s . In ' ~ ~dichloromethane, (38) hydrogenated both ethylene and I-hexene. Dissociation of one PPh, ligand to give the coordinatively unsaturated species [Ir(H),( PPh3)J was a necessary preliminary to cata1y~is.l~~ Another bis(tripheny1 hosphine) species, [Ir( H),(OCMe2)2(PPh3)2]+, has been shown to hydrogenate n~rbornadiene.'~'When (38) was used under a hydrogen pressure of l MPa, turnover numbers of more than 8000 could be achieved in aldehyde hyd~ogenation.'~~ Cationic Complexes of the type [Ir(cod)L,]+(L = PPh3, PMePh2), having a non-coordinating anion, efficiently hydrogenate alkenes and dienes. lSy These complexes oxidatively add hydrogen to give the corresponding dihydrides, [Ir(H)2(cod)L2]+.The diene is then hydrogenated to cyclooctene. The proposed mechanism for L = PMePh, is shown in Scheme 11. [ Ir(cod)( PMePh2)J'

--i

HZ [Ir(H)2(cod)(PMePh2),l+

alkene

^.i.--.

cyclooctene

alkwianr [ Ir(alkene),( PMePh?)?]'

[IrH(alkyl)(alkene)(PMePh,),]'

t

[Ir(H),(alkene)z(PMePh2)2]+

Scheme 11

Other phosphorus ligands used include PCy, and diphos. The complex [Ir(cod)py(PPr'),]+ was also active. After the cessation of hydrogenation, deactivation of the catalyst occurs rapidly.16' This was caused by the formation of the dinuclear hydride [II-,(H)~LJ+.The PPh, complex has

(39)

(40)

The pyridine complexes also undergo deactivation, but the nature of the resulting complex is unknown. A series of para-substituted triarylphosphines was used to give the complexes [Ir(cod){P(p-XC,H,),),]+ (X=MeO, Me, H, F, C1).16' The rate of hydrogenation of 1-heptene and 1,Ccyclohexadiene increased with the electron releasing ability of the substituent. It was thought that the more basic phosphines favoured both the oxidative addition of hydrogen and alkene coordination. Deactivation of the catalyst again occurred after complete hydrogenation of the alkene. [IrCI( PPh3)Jrthe iridium analogue of Wilkinson's catalyst (23),does not function as a hydrogenation catalyst under mild conditions, as the necessary dissociation of one PPh3 ligand does not occur. However, bis(tripheny1phosphine) species prepared directly from [TrCI(C,H,,),] (40) and the ligand were found to be about 10 times as active as (23).There was a marked preference for terminal over internal alkenes and the isomerization activity was much greater than with (23).16' In a further study using complexes of the type [IrCl(cod)(PPh,-,Cy,)], it was found that for all four phosphines the highest activity for 1-hexene hydrogenation was exhibited at P/Ir = 1. The PCy, and PPhCy, complexes were most eff e c t i ~ e . ' ~ ~ The complex [Ir(cod)py(PCy,)]+ (41) as the PF6- salt has been found to be superior to complex (23)in the selective hydrogenation of steroids to the 5a p r 0 d ~ c t . lIndenones ~~ and a series of substituted cyclohexenols could also be hydrogenated in good yield and with generally very high selectivity with (41).'66A brief summary of some of the above chemistry is a~ailab1e.I~'

Catalytic Activation of Small Molecules

247

[ IrH(CO)(PPh,),] (42) catalyzesthe hydrogenation of ethylene, butadiene and dimethyl maleate. The rate of its reaction with hydrogen is far greater than that of hydrogenation. Oxidative addition and dissociation of PPh, give the species [Ir(H),(CO)(PPh,),], which then enters the catalytic c ~ c l e . ' ~A~ ,precise ' ~ ~ mechanism could not be defined and it appears that this is in any case dependent on the alkene."' For the hydrogenation of 1,3-butadiene, equations (36)-( 44) were postulated. 7 3- C,H, represents the trihapto-1-methylally1 group. [IrH(CO)(PPh,),]

+ H,

[IrH(CO)(PPh,),]

-+

[Ir(H),(CO)(PPh,),]f

PPh,

(36)

[IrH(CO)(PPh,),]+ PPh,

(43)

s [Ir(H),(CO)(PPh,M

(44)

[IrH(CO)(PPh,),l+ H2

In the reduction of dimethyl maleate, isomerization to fumarate was much faster than hydrogenation."' In addition to a mechanism broadly similar to that for butadiene, a further pathway was thought to operate which involved an ortho-metallation of one PPh3ligand, as shown in equations (45) and (46). [Ir(alkyl)(CO)(PPh,),1

-+

[I~6H,PF'h,)(CO)PPh3]+alkane

[Iz6H4PPhZ)(CO)(PPh3)]+Ht + [IrH(CO)(PPh,),l

(43)

The iridium( I) complex [IrCl( CO)(PPh3)2](43),generarly known as Vaska's compound, has played a most important role in the study of oxidative addition and Strohmeier and co-workers have shown that under appropriate conditions it is an extremely efficient hydrogenation catalyst. With activated alkenes, turnover numbers of several thousands could be achieved at raised temperature and pressure. [IrCl(CO)(PCy,),] gave even better results, and with ethyl crotonate Table 4 Some Other Homogeneous Iridium-catalyzed Hydrogenations Substrate

Complex

[IrCl(cod)L] L = PPh,, PCy, [Ir(OAc)(cod)I, [Ir(OAc)(cod)PPh,] [IrH(C1)(BF4)N,(PPh,),1 [ Ir,Cl, L] L = C,H,(CH,P(CH,CH,PPh,),}, [Tr( u-carb)CO(PhCN)PPh,] u-a r b = 7-C,HS-1,2-C, R , H I,,

Re$

1-Hexene

178

Cyclohexene Hexene, cyclohexene

179 180

Alkenes, alkynes

181

Alkenes, dienes, trienes

182

Alkenes. activated alkenes

183, 184

Activated alkenes

185

Alkenes, activated alkenes, ketones

186

248

Uses in Synthesis and Catalysis

turnover numbers over 150 000 were attained.”’ The precise mechanism of homogeneous hydrogenation with (43) is still ~ n c e r t a i n . ’In~ ~diene hydrogenation, which occurs with high selectivity towards monoene formation, v3-allyl complexes are believed to be formed.174 1,4-Cyclohexadiene,”’ but-2-yne-l,4-di01’~~ and dimethyl maleate’77 have also been hydrogenated using complex (43) as catalyst. Some other reports of homogeneous iridium hydrogenation catalysts are given in Table 4.

61.2.2.5

Nickel, Palladium and Platinum

[NiJ,( PPh,),] (44)hydrogenates linear and cyclic dienes, both conjugated and non-conjugated. With the former, 1,6addition of hydrogen predominates, the reaction involving $-allyl intermediates. The complex also reduces 1,5,9-~yclododecatriene.’~~ In a further study of this reaction, (44) was believed to activate hydrogen by heterolytic cleavage (equation 47). The reaction is then believed to proceed by alkylene complexes such as (45) to give the monoene.ls8 These reactions require hydrogen pressures well above normal. PPh

[NII,(PPh,),] +H, A [Ni(H)I(PPh,),] +Ph,P.HI

(44)

(47)

(45)

Treatment of [ N i ( a c a ~ ) ~ with ] Al2Et3Cl3in presence of PPh, gave solutions which selectively hydrogenated cyclic dienes to m o n o e n e ~ at ’ ~normal ~ hydrogen pressure. [Ni(acac),] alone was shown to hydrogenate cyclohexene, but again only at raised hydrogen pressure. From a kinetic study of the reaction, alkene coordination was concluded to precede hydrogen acti~ation.”~ Palladium complexes of the general kind [PdX,L,] (X = halogen, L = phosphine), either alone or in presence of SnCI,, hydrogenate linear and cyclic dienes including sterically hindered ones [PdC1(PPh3)(rl3-CsHI3)],was isolated from at raised hydrogen p r e ~ s u r e .An ~ ~allyl ~ , ~complex, ~~ these solutions and was shown to be a more active catalyst than [PdCl,(PPh,),]. An excess of PPh, poisoned the reaction but neither catalyst was affectedby an excess of chloride ion. The complex [PdC1( PPh3)(v3-C4H7)]and related complexes have independently been shown to hydrogenate linear and cyclic o~tadienes.‘~’ The atlyl group is first reduced, creating coordinative unsaturation. Complexes of diphosphines such as [PdCl,(diphos)] and [Pd(diphos)J hydrogenate acetylene, monoenes, dienes and trienes.’93 These reactions were accelerated by exposure of the catalyst solution to oxygen, possibly due to the generation of coordinative unsaturation by partial oxidation of the phosphorus ligand to the phosphine oxide. In DMF as solvent, PdC1, or [PdCl,( DMF)J will catalyze the hydrogenation of conjugated dienes and alkynes to alkenes.lg4The above systems quite generally give selective reduction of dienes to monoenes. In presence of pyridine, solutions of [Pd(acac),] catalyzed the reduction of nitrobenzene to aniline at normal pressure of hydrogen and the complex [PdH(acac)( PhNO,)py] was isolated. The dinuclear complex [PdZCl4(PPh,),] (46) also catalyzes the reduction of nitrobenzene. At normal hydrogen pressure, complete reduction to aniline occurred.’96A nitrobenzene complex, believed to be an intermediate in the reaction, was isolated (equation 48).

(46)

Platinum compiexes of the general type [PtX2L2](X = halogen, pseudohalogen; L = phosphine, arsine, sulfide, selenide) in presence of SnCI, hydrogenate terminal alkene^,'^^-'^^ linear and cyclic dienes and triene^,'"^'^*^,'^^ unsaturated nitriles lg8 and long-chain alkenes such as methyl In general, dienes are reduced to monoenes selectively. These linoleatezo0and soybean reactions require raised hydrogen pressure (3.4-5MPa). The complex [PtC12(PPh3),] has been most generally used. 198~199*201 With S K I , , this gives [PtCl(SnCl,)(PPh,),] (47). Complex (47) then gives the hydride [PtH(SnC13)(PPh3),] (48) wirl-

Catalytic Activation

of Small Molecules

249

hydrogen. 187~1993200Other Lewis acids than SnCl, have been tried but were far less effective.200 For the hydrogenation of isoprene the mechanism shown in Scheme 12 has been proposed.lS7 This mechanism requires the heterolytic cleavage of the hydrogen molecule but does not show the formation of a hydridoalkene-hydride complex, which is generally believed to be an intermediate stage in hydrogenation2’ CH3 [PtH(SnCL)(PPh&

J

-+

I

CH2=C-CH=CH2

\

SnC13

SnC13

I

!

PhgP-Pt-PPh3 I

Ph3P-Pt-PPh3

I

CH~--~H-CH=CH~ -t

[PtH(SnC13)(PPh3)2] Scheme 12

[PtCl(SnCl,)(PPh,),] and related complexes have been shown to be useful for the partial hydrogenation of soybean oiL201Complexes of the type [PtCl,LL’] ( L = PPh,; L’ = sulfide, amine) are better catalysts for the hydrogenation of styrene than the corresponding [PtC12L2]or [RC1,L’2] complexes. Here again, SnC1, was present.’02 The ligands L’ were shown to be more labile than PPh, and their value appears to lie in facilitating the production of coordinative unsaturation. This, in turn, may lead to the alkene complex lacking in Scheme 12. Further evidence for monophosphine species is provided by the complexes [PtHL( v3-C3H5)](L = PCy,, PBu’,), which catalyze the hydrogenation of dienes to monoenes even at -78 0C.203The mechanism shown in yl was shown to cause immediate Scheme 13 was proposed. Addition of diene to the ~ j ~ - a l lcomplex displacement of propylene. No reaction with hydrogen occurred in absence of diene.

\-

‘H

II

Scheme 13

Uses in Synthesis and Catalysis

250

[PtH( NO,)(PEt,)J at raised temperature and pressure hydrogenates monoenes and dienes to alkanes. The reduction of dienes was found to be more difficult than that of monoenes, in contrast to the behaviour of other platinum catalysts.*” Methanolic solutions of H2PtC16.2H20and SnC12.2H20catalyze the hydrogenation of ethylene and acetylene under mild ~onditions.~’~ Hydrogen activation was attributed to equation (49). This step was believed to precede coordination of ethylene, after which alkene insertion and protonation of the platinum-ethyl bond complete the cycle. In isopropanol in the presence of HBr this system has been shown to hydrogenate a range of simple and activated alkenes at normal temperature and hydrogen pressure.*06The hydrogenation of methyl cinnamate by [PtC12(SnC1&I2- was shown by deuteration studies to involve a cis or suprafacial transfer of the first hydrogen atom to the double bond. The second hydrogen transfer was non-~pecific.~” [Pt(SnC13)5]3-+H,

-B

[F’tH(SnCI,),]3--t H+f SnCI,-

(49)

61.2.2.6 Other Metals [Os(H),(PEtPh,),] catalyzes the hydrogenation of I-octene at normal pressure and raised temperature, but considerable isomerization occurs. The complex can be recovered unchanged after the reaction. The cis complex [Os(H),(PEtPh,),] catalyzes isomerization more rapidly than hydrogenation.’08 [CuBr(PPh,),] hydrogenates nitrobenzene, giving nitrosobenzene, azobenzene and aniline at high temperature. The dinuclear complexes [ C U ~ ( H ) ~ B ~ ~ ( Pand P~~)~ [Cu,Br,( PPh,),(PhN=NPh)] were isolated from the reaction and are believed to be intermediates.z09 There are a number of reviews available on homogeneous hydrogenation. Speciaiized reviews are available on hydrogenation with phosphine complexes of rhodium,154on hydrogenation of arenes,2*0on hydrogenation with water-soluble catalysts,211and on mechanistic aspects.81There are general reviews covering the older212-216 and more r e ~ e n t ~ ” ~literature. ’”~~*

61.2.3 ASYMMETRIC HYDROGENATION 61.2.3.1 C=C Bonds

The introduction of the highly efficient homogeneous hydrogenation catalyst [RhCl( PPh?)J (23)7’ led to attempts to adapt this system to catalytic asymmetric homogeneous hydrogenation by the introduction of chiral phosphorus ligands in place of PPh3. In 1968 the groups of K n o w l e ~ ~ ~ and Homer’” both demonstrated the use of methyl-n-propylphenylphosphineand a rhodium compound in asymmetric hydrogenation. Knowles et ai. used a sample of (-)-MePr”PhP which was of only 60% optical purity together with RhC1,.3H20 to hydrogenate a-phenylacrylic acid to hydratropic acid, which had an optical purity of 15%. Horner et al. used (+)-MePr”PhP with [RhC1(1,5-hexadiene)1, to hydrogenate a-ethylstyrene to 2-phenylbutane having an optical purity of 8%. The method of preparing the catalyst in situ from a rhodium(1) complex of this type and the phosphorus ligand has now found general application in asymmetric catalysis. For this very moderate start, asymmetric catalytic hydrogenation has now made dramatic progress. Optical yields of over 90% are now commonplace, and 97-99% is not unusual. From the start, there was great interest in the production of amino acids by asymmetric hydrogenation and a-acetamidocinnamic acid, its derivatives and related compounds rapidly became the principal alkenes studied (equation 50). H\

/

NHCOMe

c=c,

0

‘OZH

+

H2

-

NHCOMe * / O C H , - C H \COPH

(50)

Initial efforts concentrated on monodentate ligands in which the phosphorus atom itself was the asymmetric centre. Morrison and co-workers introduced the ligand neomenthyldiphenylphosphine, in which the phosphorus atom itself is not chira1.22’Such a phosphine is easily prepared from a natural, chiral substance and does not require the resolution step necessary for those chiral

Catalytic Activation of Small Molecules

25 1

at phosphorus. Using this ligand and a rhodium(1)-ethylene or -diene complex, (E)-P-rnethylcinnamic acid could be hydrogenated with an enantiomer excess of 61% ,far surpassing the previous results. A further important step was taken by Kagan, who introduced the first chelating diphosphine ligand to asymmetric h y d r o g e n a t i ~ n .His ~ ~ ligand, ~ ’ ~ ~ ~2,3- O-isopropylidene-2,3-dihydroxyl,Cbis(diphenylphosphino)butane, generally abbreviated to DIOP (49; Figure l ) , could be used with [ R ~ - I C ~ ( C to ~ hydrogenate H ~ ~ ) ~ ~ ~a-acetamidocinnamic acid in an optical yield of 72%. Other amino acid precursors could be reduced in up to 80% optical yield. DIOP has the same advantage as neomenthyldiphenylphosphine of being prepared from a chiral natural substance, in this case L-tartaric acid. The value of the a-acetamido group in these reactions was soon re~ognized.2’~ The complexity of the full chemical names of most ligands used in asymmetric homogeneous catalysis has led to the general introduction of abbreviated names such as DIOP and this practice will be followed here. Knowles next reported optical yields of up to 90% in the hydrogenation of further aacetamidocinnamic acid derivatives, still using monodentate phosphines chiral at phosphorus. Again the catalysts were prepared in situ from the ligand and [ RhC1( 1,5-hexadiene),] or a related complex. The best optical yield of 90% was obtained using cyclohexyl(o-anisy1)methyIphosphine (CAMP).224 The above reports laid the basis for catalytic asymmetric hydrogenations giving optical yields of close to 100%. Quite generally rhodium was the metal used with a chelating diphosphine as ligand. It seems to matter but little from the point of view of the enantiomer excess whether the asymmetric centre(s) in the ligand are at phosphorus or elsewhere. a-Acylamino-cinnamic and -acrylic acid derivatives have generally been the substrates used. The value of an alkene which can act as a bidentate chelate has long been recognized. Reviews of the early literature are systems which have led to optical yields of 90% or a ~ a i l a b l e . ~Those ~ ~ - ~rhodium-diphosphine ~’ better are summarized below. The reviews given at the end of this section should be consulted for other more recent reports.

(49) DIOP

(50) DIPAMP

(51) CHIRAPHOS

pc

(52) SKEWPHOS

PPh,

PhzP PhZP (53) PROPHOS PPh,

PPhz

(54) CYCPHOS

PPhz

(55)

(56) NORPHOS

Figure 1 Some diphosphine ligands used in asymmetric hydrogenation

[ R . ~ - I C I ( C ~ Hin ~ ~presence )~] of DIOP (49) gave an optical yield of 92% in the hydrogenation of a-acetamidocrotonic acid.’*’ The complex [Rh(cod)(R&DIPAMP)]+ reduced a number of a-acylamino-cinnamic and -acrylic acid derivatives in optical yields of 93-96% .230 DIPAMP is the ligand used in the Monsanto synthesis of L-dihydroxyphenylalanine(L-DOPA) by rhodiumcatalyzed asymmetric hydrogenation. L-DOPA is important in the treatment of Parkinson’s disease. The complexes [Rh(diene)(S,S-CHIRAPHOS)]+(diene = cod, nbd) were used to prepare Rphenylalanine, R-leucine and R-tyrosine in optical yields of 99, 100 and 92% respectively as the N-benzyl derivatives by hydrogenation of the appropriate dehydro compounds.231The N-acetyl derivatives of R-alanine, R-phenylalanine, R-tyrosine and R-DOPA were obtained in 90-9870 optical yields by the hydrogenation of the corresponding dehydroamino acid derivatives using [Rh(nbd)(S,S-SKEWPH0s)lf as catalyst. When the related ligand (S)-1,3-bis(diphenylphosphino)butane, which possesses only one asymmetric centre, was used in place of S,S-SKEWPHOS, the optical yields dropped to 1-20%, This was attributed to the phenyl groups being in an achirdl array in this Both of these ligands form six-membered chelate rings. When (R)-1,2bis(diphenylphosphino)propane, or R-PROPHOS (53), was used in similar hydrogenations, the optical yields were again 91-93%, indicating that in the less mobile five-membered chelate ring

CCC6-I*

252

Uses in Synthesis and Catalysis

one asymmetric centre is sufficient.233The use of ( R ) -1-cyclohexyl-l,2-bis(diphenylphosphino)ethane, (R)-CYCPHOS(54), in which the methyl group of PROPHOS is replaced by the more bulky cyclohexyl ring, led to optical yields of 90-98% in such hydrogenation^.'^^ The cyclobutane derivative (55) gave an optical yield of 91% in the hydrogenation of aacetamidocinnamic acid. The catalyst was here prepared in situ from [RhCI( 1,5-hexadiene)12. The corresponding 1,2-derivative of cyclopentane gave an optical yield of only 73% and the cyclopropane and cyclohexane derivatives gave 15% and 36% respectively.235[ RhCl(cod),] in presence of the norbornadiene-based ligand NORPHOS (56) gave up to 96% optical yields with a-acetamidocinnamic acid as The above reports indicate that those diphosphines which may be regarded as derivatives of 172-bis(diphenylphosphino)ethaneand which form five-membered chelate rings are easily adapted to use in asymmetric synthesis, the introduction of a single asymmetric centre in the carbon backbone being sufficient to give high optical yields (PROPHOS, CYCPHOS, NORPHOS). Two such centres (CHIRAPHOS), or asymmetry at both phosphorus atoms (DIPAMP), also leads to very high optical yields. For those diphosphines which form six- or seven-membered rings, the necessary asymmetric induction is less readily achieved. A number of comparative studies of phosphorus li ands in asymmetric hydrogenation catalysed by rhodium complexes have appeared.237-244 Considerable ingenuity has been shown in designing ligands for asymmetric hydrogenation and some other rhodium catalysts which lead to very high optical yields are collected in Table 5. Chiral ligands will be found there and in Figure 1. A number of groups have shown interest in the mechanism of asymmetric hydrogenation, principally of { a)-2-acetamidocinnamic acid and its derivatives. Kagan and co-workers have shown that cis addition of deuterium occurs to the 2 isomer. Rhodium complexes of DIOP were used here?5s Koenig and Knowles obtained similar results with the ligands DIPAMP, cyclohexyl(oanisy1)methylphosphine and others.ZShBoth groups showed that E - 2 isomerization occurred during the hydrogenations. NMR spectroscopy has been used to study the species formed in solution by interaction of cinnamic acid derivatives with asymmetric hydrogenation catalysts.2s7725s Such studies are necessarily limited to those species which accumulate in adequate concentration and have sufficientlylong lifetimes for observation by NMR. In catalytic reactions as rapid as those described here, such complexes appear likely to be outside rather than in the operating catalytic cycle?' Halpern and co-workers have carried out a detailed investigation of the mechanism of the asymmetric hydrogenation of methyl (MAC) and ethyl (EAC) (2) -a-acetamidocinnamate by rhodium complexes of the ligands DIPAMP (So)and CHIRAPHOS (51)?59Coordination of alkene precedes the oxidative addition of hydrogen. For both ligands, one of the two possible diastereoisomers of the rhodium-diphosphine-alkene complex predominates in solution to a large extent. From the reaction of EAC with the S,S-CHIRAPHOS complex, this diastereoisomer has been isolated. Its structure is represented in (57).260

\Me (57)

A feature of this structure is the arrangement of the four phenyl groups of the CHIRAPHOS ligand. Each PPhz unit has one phenyl group presenting an edge and one a face towards the

Catalytic Activation of Small Molecules

253

TabIe 5 Some Other Highly Selective Asymmetric Hydrogenation Catalysts Catalyst

% e.ea

[Rh(cod)L]+

a-Acylaminoacrylic acids

Re$ 241, 242

Me

91-94 Me

phxNHpPhz

L = Ph

93

NHPPhz

a-Acylaminocinnamic acids

[Rh(cod)L]+

a

243

Me

I

N-PPhz

L=

92

N-PPh,

I

Me

NHPPhz L=

93-94 NHPPhz a-Acetamidocinnamic acid

[Rh(codjL]+

94

244

a-Acetamidocinnamic acid

91-93

245

a-Acetamidocinnamic acid

91-98

246

91

241

phfpph2

L = Ph Me PPh2 [Rh(cod)L]+ (menthyl)

I

N-PPhz

L=

(

N-PPh2

I

(menthyl) [Rh(diene)CAPP]+ diene = cod, nbd

CAPP =

"'h PPhz

N CONHR

R = aryl, aikyl [RhCI(1,5-hexadiene)12+ L

X X = H (PPM) X = C0,Ru" (3PPM)

a-Acetamidocinnamic acid

Uses in Synthesis and Catalysis

254

Table 5 (continued) Catalyst

YO e.e.

Ref:

e-Benzoylarninocinnamic acid

95

248

a-Benzoylaminocinnamic acid

98

249

a-Acetamidocinnamic acid

90

250

90-91

251

a -Acetarnidocinnami c acid

92

252

a-Acetamidocinnarnic acid

93

253

91-92

254

Substrate

a

NHPPh,

PH,P

PPh,

(DIOXOP) [Rh(nbd)L]+

a-Acetamido-acrylic and -cinnamic acids

.o1

NC

0

PhzP

[RhC1(1,5-hexadiene)I2+

CHMeNMe2 [Rh(nbd)Lj+

N-Acetamidoacrylic acid, N- benzoylaminocinnamic acid, 4-acetoxy-3-methoxya - acetami do cinnamic acid

Bu'OCH~ L=

PPhz

1 , PPh,

a

PhCH~OCH~xPPh~

e.e. = enantiomer excess.

PPh,

Catalytic Activation of SmaII Molecules

255

alkene. This is shown more clearly in (58). This arrangement, which appears to be common to many rhodium-diphosphine complexes used in asymmetric hydrogenation, accounts for the chiral recognition by the rhodium atom of the alkene and explains the preponderance of one diastereoisomer of the [€#(EAC)( S,S-CHIRAPHOS)]+ complex. Halpern has shown that this predominant isomer exhibits negligible activity towards the oxidative addition of hydrogen. The minor isomer, which could be detected in solution for DIPAMP but not for CHIRAPHOS, reacts far more rapidly with hydrogen and is responsible for producing the major enantiomer of the hydrogenation product. The optical selectivity is thus due to this difference in reaction rates and not simply to the preferred manner of coordination of the alkene to the rhodium-diphosphine specie^.*^^*^^^ The precise reasons for this large difference in the rates of reaction of the two diastereoisomers with hydrogen are not yet known. The full mechanism is shown in Scheme 14.

I

I J

J

L

I

I Me

Me

I

Ay

'0

/J

Ph

\Ph

I

I

[Rh(P-P)S',]'

N H yHC 0 2 M e

Mev

H Me02C =

IRh(P--P)S',]+

NH

0

0 Ph

Ph (S )

(R) (S' =solvent) Scheme 14

The mechanism requires that the initial alkene coordination step be reversible, otherwise all of the rhodium would soon be present as the major diastereoisomer of the alkene complex and the reaction could then only proceed by this route, reversing the optical selectivity and drastically

Uses in Synthesis rand Catalysis

256

reducing the rate. If the hydrogen pressure were to be increased, the degree of reversibility of this step would be reduced through increased competition by the oxidative addition of hydrogen. The selectivity would then become increasingly dependent on the mode of coordination of the alkene, leading to a reduction in tbe optical yield, with a possible reversal of the optical selectivity at sufficiendy high pressure. Such an inverse dependence af optical yield on hydrogen pressure has been observed in asymmetric hydrogenation^.^^^,^^^ It is noteworthy that, at normal temperature and hydrogen pressure, these systems not only give optical selectivitiesvery close to 100°/~,but also show activities well up in the range previously thought to be attainable only by enzymes. That this is possible with such comparatively simple transition metal complexes indicates how prodigal Nature has been in constructing its own catalysts. Following the success in producing amino acids by asymmetric hydrogenation, this research has been extended to dipeptides. Using rhodium complexes of the same or similar ligands to those above, the hydrogenation of dehydrodipeptides is also possible in opticaI yields in the range of 90-98% (equation 51).261-264 R'

\

/

H /c=c\

NHCOR3

NHCOR3

+ Hz -+

I

R'CHz-*CH-CONH-*CHCOzR4

I

CONH-*CHC0zR4

(51)

R2

I

RZ

Poulin and Kagan have shown that a double asymmetric hydrogenation can be carried out on the same dipeptide precursor with excellent optical selectivity (equation 52). [Rh(cod)(R,RDIPAMP)]+ was used as catalyst. The major diastereoisomer of the product, which had the S,S-configuration, showed an optical purity of better than 95% .265

MeCONH

CONH

phkphk

C02Me

MeCONH

CONH

(52)

C0,Me

Efficient asymmetric hydrogenation of alkenes other than the amino acid and dipeptide precursors described above has met with only limited success. This appears to be at least in part due to the inability of many alkenes to function as bidentate chelates. Ethyl 2-acetoxyacrylate was liydrogenated with an enantiomer excess of 89% using [Rh(cod)( R,R-DlPAMP)]', giving the S-enantiomer (equation 53). The ligands CHIRAPHOS, PROPHOS, DIOP, BPPM and CAMP were less effective.266 H

\

/

H /c=c\

C02Et

+H, 4 CH,-*CH 0-C-Me

/ \

COZEt

(53) 0-C-Me

II

II

0

0

Itaconic acid and its derivatives represent the only other alkene type which has so far been reduced with optical yields of over 90%. Using the usual cationic rhodium-diene (cod) precursor, a best optical yield of 94% was obtained with the diphosphine FPPM (59). The presence of triethylamine was necessary.267With the related BPPM (Table 5), again in presence of triethyl[Fth(cpd)(DIPAMP)]+ gave an optical yield amine, optical yields up to 92% were of 88% for the dimethyl ester of itaconic acid, and 90% with a monoamide of a-methyleneglutaric PhzP

I QCH2PPhz CHO (59) FPPM

Several recent reviews on catalytic asymmetric hydrogenation are available. The mechanistic studies have been summarized by H a l p e ~ n . 8 ' ~The ~ ' ~synthesis and use of phosphorus( 111) ligands have been reviewed by H ~ r n e r . ~ Ferrocene-based ~' and nitrogen-containing phosphorus ligands have also been covered.271General reviews of the subject have a p ~ e a r e d . ~ ' ~ - ~ ' ~

Catalytic Activation of Small Molecules

257

61.2.3.2 C=O Bonds The asymmetric hydrogenation of C=O bonds have now been achieved in optical yields up to 95%, rivalling the performance of alkenes. Here also, rhodium complexes have been used almost exclusively, but some success has been obtained with cobalt catalysts. Using [Co(HDMG),] in presence of optically active bases, benzil could be reduced to benzoin (equation 54) in an optical yield of 78%. Quinine or quinidine were the chiral bases employed. The best optical yields were obtained with quinine (60). It was found that when benzylamine was also present, the rate of hydrogenation was greatly enhanced without any decrease in the optical yield.276 OH 0

0 0

I1

II

II

I

PhC-CPh+H,

+ Ph-*CH-CPh

(54)

Using a quadridentate ligand derived from 1,3-propanediamineand 2,3-butanedione the cobaltoxime complex (61) was prepared, and again in presence of the chiral base quinine (60)this was used to hydrogenate benzil according to equation (54) in an optical yield of 79% .*" These cobalt systems permitted only low turnover numbers to be achieved.

&

PPh2

I

Me0

Fe

I

*' 1-

CH(Me)O Ac

i

(60)

(62) BPPFOH

(61)

Kumada and co-workers have introduced the ferrocene-based ligand BPPFOH (62) for the rhodium-catalyzed asymmetric hydrogenation of carbonyl compounds. In presence of a stoichiometric quantity of triethylamine, pyruvic acid was hydrogenated by the complex [Rh(cod)(BPPFOH)J+in an optical yield of 83% to lactic acid (equation 55).278 The same ligand has been used to hydrogenate aryl aminoalkyl ketones with optical yields of up to 89% (equation 56).279 OH

0

I1

MeCCO,H+H,

-+

I

Me-*CHC02H

(55)

The a-ketolactone (63) can be hydrogenated to give pantolactone (64; equation 57). Using [RhCl(cod)J2and the ligand BPPM (Table 5 ) an enantiomer excess of 81% was obtained. Related ligands possessing different substituents on the nitrogen atom of BPPM were less ef€ective.2'* When a neutral chloro complex of BPPM was employed as catalyst, the optical yield was raised to 87% .**l With DIOP (49) the best result was only 40%.

Q +

0

0

0

(63) 0

(57)

H2

-

(64)

OH

mccH2NMe2 II

\

/

+ H?

QJyHCHzNMe2

(58)

[RhCl(nbd)12and DIOP gave an optical yield of 95% in the hydrogenation of a-diethylamino-2acetonaphthalene (equation 58).282This is the highest optical yield yet reported for the asymmetric

258

Uses in Synthesis and Catalysis

hydrogenation of a carbonyl group. There has so far been no really detailed study of the mechanism of these asymmetric hydrogenations and the origin of the optical selectivity remains unknown. 61.2.4

HYDROFORMYLATION

61.2.4.1

Introduction The hydroformylation reaction is of particular interest here as it involves the activation of two or even three small molecules (equation 59). Isomer formation is quite general where the structure of the alkene permits this and much effort has been invested in attempts to improve the ratio of normal to branched products. CH,-CH=CHI+

H,+CO

-+

CH,-CH,-CH,-CHO

or CH3-CH-CH,

I

(59)

CHO

Since its discovery some 50 years ago by Roelen, a great deal of research has been carried out on the reaction and its industrial importance is great. The initial work used as catalyst precursor [Co2(CO),] or simple cobalt salts which were carbonylated under the reaction conditions. Subsequently, phosphine-modified cobalt catalysts were introduced and, more recently, rhodium and platinum catalysts. Only the cobalt and rhodium catalysts have found industrial use to date. Catalysis of hydroformylation by [Co,(CO),] and other binary metal carbonyls is outside the scope of this work. Reviews are a ~ a i l a b l e . ~ ~ ~ - ~ * ~ Iron and Ruthenium The complexes of these metals show only limited activity as hydroformylation catalysts. [Fe(CO),( PPh,),] and related complexes were found to hydroformylate 1-pentene,but the conversion remained Its ruthenium analogue [Ru(CO),(PPh,),] (65) was significantly more efficient.287-289 [RuCl,(PPh,)MeOH] was also active, but [RuC12(PPhJ3] (2) gave very poor conversion and precipitation of the insoluble dicarbonyl [RuC~,(CO),(PP~,)~] (10). In a more detailed study of ruthenium complexes in hydroformylation, conversions to aldehydes of more than 80% were obtained using 1-hexene as substrate with the catalyst precursors [Ru(CO),( PPh3)2] (65), [ R U ( H ) ~ ( C O ) ~ ( P (661, P ~ ~ [Ru(H)KO(PPh,),I )~I (12), [RuH(NO)(PPh,),l, [Ru(H),(PPh,),I (6), [ R u ( H ) ~ ( P P ~ , )[ ~ R]U, ( O ~ C M ~ ) ~ ( P P(9) ~ ,and ) J [Ru(O,CBU'),(PP~,)~].In a11 cases, complex (65) was recovered from the reaction r n i ~ t u r e s . 2 Variation ~~ of the phosphorus ligand showed that PPh, and P(OPh), gave the best results. A mechanism was proposed for the hydroformylation using [Ru(CO),(PPh,),] as catalyst (Scheme 15).

61.2.4.2

PPh3

PPh,

PPh3 (65)

PPh3 OC-RU

PPh3

I @ co

I

bco

PPhi RCH,CH,CHO

PPh?

Scheme 15

Catalytic Activation of Small Molecules 61.2.4.3

259

Cobalt

[Co2(CO)4itself as a hydroformylation catalyst lies outside the scope of this article but reviews are available. R3-2R6 One of the technical problems with this catalyst is the high reaction pressure needed when using it. It was shown by Slaugh et ai. that hydroformylation could be carried out at very much lower pressure if a phosphine ligand was added to the system. Trialkyl- and mixed alkylaryl-phosphines were used, the latter including the diphosphines Ph2P(CH2),,PPH2( n = 2,4,5); AsBu, and AsPhEt, were also effective. The best ligand was PBu"~'?~' [Co,(CO),] reacts with PBu", to give the complex [ C O ( C O ) ~ ( P B U ~ ~[Co(CO),])~]+ (67). Under hydrogen, [CoH(CO),PBu",] (68)is formed. This system is more strongly reducing than [Co,(CO),] alone and it facilitates the direct production of alcohols by hydrogenation of the aldehydes formed. Some loss of alkene by hydrogenation occurs. The PBu",-containing catalyst system also exhibits a stronger preference for terminal alkenes than does [ co~(Co)~]."'

(68)

(69)

The reactions are solvent dependent. Non-polar solvents favour the formation of the neutral hydride (68), whereas in polar media, ionic species predominate. Hydroformylation activity for propylene was observed only under conditions where (68) was formed.291The dinuclear species ~ * ~ ~is ~added in excess [Co,(CO),(PBu",),] (69) also appears to be catalytically i n a ~ t i v e ?If~ PBun3 over cobalt, only complexes (68) and (69) were present. The rate of hydroformylation then became first order in hydrogen, owing to the equilibrium shown in equation (60).292

The kinetics of the interconversion of complexes (67) and (68) have also been studied.293It seems that the mechanism of hydroformylation in the presence of PBu", is similar to that for [co2(co)8] alone.285,294 The use of diphosphine ligands, especially diphos itself, together with [CO,(CO)~]can lead to catalysts which show significantlyimproved hydroformylation activity, but as yet nothing is known of the complexes which are formed.295Extremely stable and highly selectivecatalysts were obtained from [Co,(CO),] and derivatives of the ligand 1,Z-diphos hacyclopent-5-en-4-one (70). Again, nothing is known of the coordination chemistry involved.'

F

Ro\

R'P-PR

v o C02R' z R 0 (70) R = aryl

R'= alkyl

61.2.4.4

Rhodium and Iridium

Wilkinson and co-workers showed that the hydrogenation catalyst [RhCl( PPh,),] (23)was also an effective catalyst for the hydroformylation reaction.297The complex reacts rapidly with CO to give [RhCl(CO)(PPh,),] (71). Both complex (71) and a range of its derivatives having other phosphine ligands were shown to catalyze the hydroformylation of 1-pentene under conditions far milder than those needed when [co2(co)8] was In the presence of excess PPh, the complex [RhH(CO)(PPh3)3](34)was isolated, and this was found to be a better catalyst precursor Ph,P

h '

c1/

h

co l

,

Uses in Synthesis and Catalysis

260

than (23) or (71). With this complex, alkenes could be hydroformylated rapidly at room temperature and pressure. Terminal alkenes again reacted more rapidly than internal alkenes.298Complex (34) in the presence of CO gives in solution the species [RhH(CO),(PPh,),] (72).294The mechanism of the hydroformylation is given in Scheme 16?w RhH(CH,=CHR)CO(PPh3)2

J[

RCH=CH,

RhH(?P)(PPh,),

e

Rh(CH,CH,R)CO(PPh,),

co e Rh(CH2CH2R)(C0),(PPh3),

RCH,CH,CHO

a

n Rh(H)2(COCH2CH,R)CO(PPh3)2 & Rh(COCH2CH2R)CO(PPhJ2

Scheme 16

A most important discovery was that the addition of a large excess of PPh, relative to rhodium leads to a very high normal/branched aldehyde ratio, especially when PPh3 itself is used as solvent.29sThis was attributed to steric effects on the alkene insertion, the excess of PPh3preventing its dissociation in this step. When P(OPh), was used as ligand, the effect of an excess of it on the isomer ratio was far less significant."' These studies have led to the introduction of an industrial process for the rhodiumcatalyzed hydroformylation of propylene to n-butyraldehyde which is rapidly gaining in importance relative to the older, cobalt-catalyzed r ~ ~ t eThe ? ~relative ~ , ~ merits ~ ~of the two processes have been d i s c ~ s s e d . ~ ~ ~ * ~ ~ ~ It has been found that in the rhodium-catalyzed process a slow loss of catalytic activity occurs and this is partly due to the presence of the more basic PPh,Pr", formed according to equation (61) and involving ortho-metallated rhodium intermediates such as (73).305 PPh, + CO +H,

+ C,H6

PPh,Pr"

+ PhCHO

Ph2po

(61)

(Ph3P)?Rh \ (73)

A kinetic study of the hydroformylation has been carried and the mechanism proposed by Wilkinson (Scheme 16) was extended to express both the associative and dissociative modes of alkene and the formation of n- and iso-butyraldehydes. High-pressure IR spectroscopy usin CO and either H2 or D, has confirmed the formation of [EUIH(CO),(PPh,),] in these reactions.807 The hydroformylation of ( E ) - and (2)-3-methyl-2-pentene with [RhH(CO)(PPh,),] (34) was investigated to clarify the stereochemistry of the reaction. By use of D2 it was shown that the addition of the H and CHO groups to the alkene occurred with high selectivity in a cis manner.," Some further reports of hydroformylation catalyzed by complex (34) are collected in Table 6. Table 6 Alkene Hydroformylations Catalyzed by [RhHICO)(P%)J

Alkenes 1-Hexene n-Dodecenes Cycloheptatriene Conjugated dienes 1,6-Dienes 3,3,3-Trifluoropropene Pentafluorostyrene para-Substituted styrenes Methyl methacrylate Unsaturated acids and esters Diallylarnines Bis(a1kenyl)carbamates

Re$ 309, 310 311 312 313 314 315 315 316 317 318, 319 320 320

Analogues of [ RhH(CO)(PPh,),] with other phosphorus ligands also act as hydroformylation catalysts. Using complex (34) in the presence of 1,l'-bis(dipheny1phosphino)ferrocene as ligand,

Catalytic Activation of Small Molecules

26 1

the mechanism of the hydroformylation of I-hexene was believed to differ from that when only PPh3 was used, and dinuclear species were postuiated?21 The effect of the diphosphines Ph2P(CH2).PPh, (n = 2-4) on the hydroformylation of terminal alkenes has been s t ~ d i e d . ~ ~ ~ - ~ ~ The effects on rate Various other 1,4-bis(diphenylphosphino)butanederivatives were also and on selectivity to the various aldehyde isomers were complex, and little is known of the coordination chemistry involved other than the obvious implication of chelation. Following the work of Wilkinson, the complex [RhCl(CO)(PPh,),] (71) and its derivatives with other phosphorus ligands have been studied as catalysts for hydroformylation. These reactions quite generally require significantly higher temperature and pressure than those catalyzed by [RhH(CO)(PPh,),]. Complex (71) has found use in connection with the synthesis of a-hydroxyaldehydes and succinnaldehyde monoacetals.32xIn high-pressure IR work on the hydroformylation of 1-hexene and cyclohexene, it was shown that a small amount of cyclohexenyl peroxide greatly accelerated the reaction. This was due to the conversion of (71) to the dicarbonyl [RhCl(CO)2PPh3].329 The effect of replacing PPh, in complex (71) by phosphorus ligands having long alkyl chains has been studied using the phosphines PBu”, , P( n-C,H,,), , P( n-CI6H3,),, P( p-C6H4Et)3, P( pC6H4Bu), and P( p-C6H4C5H11)3. The trialkylphosphines promote alkene isomerization relative to PPh, . The substituted arylphosphines, however, gave better ratios of linear/branched aldehydes, with. P( P - C ~ H ~ B giving U ) ~ the most satisfactory combination of yield and selectivity.330With 3-phenylpropyleneas substrate, rhodium-catalyzed hydroformylation in the presence of the Group V ligands XPh, ( X = N, P, As, Sb, Bi) was studied at high pressure of H2 and CO. NPh, and PPh3gave complete conversions to aldehydes, with the distribution of the three possible aldehydes being the same in each case. With AsPh, and SbPh, the conversion began to decrease and 2-phenyl-n-butyraldehyde was hardly formed, though the other two aldehydes were again formed in similar proportions. With BiPh3, hydroformylation did not occur.331 Complexes containing aminophosphine ligands of the type [RhCl(CO)LJ ( L = PPh2NEt2,PPh(NEtJ2, P(NEt,),, PPh,NMe,, PPh(NMe2)2, PPh,N(Me)Cy, PPh{N(Me)Cy}, and { PPh2N(Me)CH,},) have been studied as catalysts for the hydroformylation of l - h e ~ e n e . , ~ ~ These ligands all have basicities lying between those of PPh, and PBu3. IR and ESCA studies showed the aminophosphines in these complexes must be considered as electron acceptor ligands, the order of electron affinity being N(Me)Cy > NEt,> NMe2. The electron density on phosphorus and nitrogen increased as the number of phenyl groups in the ligand decreased. In the free ligands, electron donation from nitrogen to phosphorus occurred, whereas in the rhodium complexes it was in the reverse direction. Both the conversion to aldehydes and the selectivity to normal aldehydes observed in the hydroformylation of 1-hexene by these complexes were markedly ligand dependent. A linear relationship between the electron density on the nitrogen atom and the normal/branched aldehyde ratio was found, indicating that for these aminophosphines the ratio is largely controlled by electronic factors.332 By ligand exchange in [RhCI(CO)(PPh,),] (71), or by direct reaction of the phosphine ligand with [Rh2C1,(C0),], a series of silicon-containing phosphine complexes were prepared.333The latter reaction in presence of LiI gave the corresponding iodides. The complexes obtained in this and way were [RhX(CO)(PPh2CH2SiMe,),] (X = Cl, I), [IUII(CO)(PP~~CH~CH~C€I~S~M~~)~] [RhX(CO)(PPh,CH(Me)Si(OSiMe,),O),l (X = C1, I). [RhC1(CO)(PPh2CH2SiMe3)2] gave very similar behaviour to complex (71) in the hydroformylation of 1-hexene, but the iodo complexes gave both lower conversions to aldehydes and much poorer selectivities to n-heptanal. [Rhl(CO)(PPh,CH,SiMe,),] gave branched aldehyde as the major The hydroformylation of methyl acrylate, methyl methacrylate, methyl crotonate and methyl tiglate has been studied with reference to the selectivity for introduction of the formyl group at the a- or @-carbon atom. The catalysts used were prepared in situ from [Rh,CI,(CO),] and Ph2P(CH2).PPh2 ( n = 2-5), Cy2P(CH2).PCy2 ( n = 2-4) and DBP’(CH2)2DBP’ (DBP’= SH-dibenz~phospholyl).~~~ Good selectivity to a-formyl derivatives was obtained only with those ligands having n = 2-4. The dibenzophosphole group was also shown to cause significant rate enhancement in alkene hydroformylation.”’ In addition to C=C bonds, [RhCl(CO)(PPh,),] (71)has been shown to catalyze the hydroformylation of formaldehyde to give glycol aldehyde (equation 62).”‘ Small amounts of methanol were obtained as a by-product. The reaction requires N,N-dialkylamides as solvents, otherwise methanol is the major product and no glycol aldehyde is formed. Coordinated amide solvent was thought H,C=O+H,+CO

+

HOCH,CHO

(62)

Uses in Synfhesh and Catalysis

262

to direct the formaldehyde insertion step towards formation of a hydroxymethyl rather than a methoxy intermediate (equation 63).

A large number of rhodium complexes having different anions and phosphorus ligands were studied, but complex (71) was the best catalyst, being under these conditions somewhat surprisingly better than [RhH(CO)(PPh,),]. The mechanism proposed for the reaction is shown in Scheme 17. Deuteration studies showed that the distribution of deuterium in the glycol aldehyde was consistent with this mechanism.336It has been shown that, in presence of small amounts of an . ~ ~ ~amines also amine, the reaction can be carried out in solvents other than d i a l k y l a r n i d e ~The significantly increased the reaction rate.

MeOH+(71)

T PPh3

PPh,

PPh7

MeOH+(71)

s (S =solvent)

Scheme 17

Complexes of carbonic or carboxylic acid anions have been used as hydroformylation catalysts for various alkenes. The bicarbonate compkx [ Rh( H),(02COH)( PPri3)J as catalyst enabled 1-hexene to be converted to aldehydes using paraformaldehyde as source of hydrogen and carbon monoxide in place of the more usual gas mixture.338The acetate complex [Rh(OAc)CO(PPh,),] (74) has been shown to effect the selective hydroformylation of cyclic dienes. The cyclohexadienes gave predominantly dialdehydes, whereas 1,3- and 1,5-cyclooctadiene gave the saturated monoaldeh~des.3~' PPh3)J, which is the Propylene hydroformylation was catalyzed by [ Rh(02C-n-C5HII)CO( caproyl analogue of (74).340In the hydroformylation of alkenes with [Rh,(CO),,], complexes of the type [Rh(02CR)(CO)2]2were isolated from the reaction mixture. As the R groups always corresponded to the alkene used, their presence was attributed to oxidation of rhodium-acyl intermediates by extraneous oxygen.341With PPh, these complexes gave analogues of (74).

Me (74)

Catalytic Activation of Small hblecules

263

Diene complexes have been used as precursors for hydroformylation catalysts. [RhCl(nbd)], in presence of PPh, was used in a kinetic study of the hydroformylation of l-he~tene.~" Cationic diene complexes of the type [Rh(cod)(XPh,),]+ (X = N, P, As, Sb, Bi) having a non-coordinating anion (ClO,-) were employed in the hydroformylation of 1-heptene. The performances of the different ligands were surprisingly similar as regards conversion and selectivity even when large (75) hydroformylates l-hexene at normal excesses of the ligands were used.,,, [Rh(~od)(PPh,)~]+ temperature and pressure in presence of Et3N but isomerization of the alkene is extensive. In presence of PPh,, [RhH(CO)(PPh,),] (34) was isolated in good yield from these reactions. [Rh(cod)py(PPh,)]+ was much less effective than (75). [Rh(cod),]' was inactive unless PPh, was added.344 The complex [RhH(CO)(mSPPh,),] containing the water-soluble phosphine rnSPPh, (31) has also been prepared, and it was shown to hydroformylate l-hexene in the pH range 5.2-9.2.345 [RhH(CO)(NBz3),1 (76) also hydroformylates terminal aikenes at normal temperature and pressure. In solution under an atmosphere of H2 and CO, dissociation of NBz, and formation of a dinuclear species (77) occur (Scheme 18).346Using l-octene as substrate it was found that n-nonanal was the predominant aldehyde formed, but extensive alkene isomerization BZ3N

H

% I

BziN ( I

Rh-NB7-3

co

I e Bz3N-Rh-NB7-3 I

CO

The complex [Rh(nbd)(arnpho~)~]~+, which contains the water-soluble ligand amphos (33), has been prepared. In a two-phase system this complex acted as a hydroformylation catalyst for of the alkene occurred to a limited extent. In the presence of H, and 1 - h e ~ e n e .Isomerization l~~ CO the dicarbonyl [Rh(C0)2(amphos),]3+was thought to be formed. The tris(2-pyridy1)phosphine complex [RhH(CO)(PPh,)(Ppy,),] in the presence of an excess of the pyridylphosphine ligand has been shown to hydroformylate 1-hexene. A high selectivity towards l-hexene was achieved.347 Iridium complexes show very limited activity as hydroformylation catalysts compared with those of rhodium. [IrCI(CO)(PPh,),] (43) has been shown to hydroformylate terminal alkenes but much higher temperature and pressure are required than with Reviews on rhodium-catalyzed hydroformylation are available.350335'

61.2.4.5

Platinum

In presence of SnCl,, platinum complexes are effective hydroformylation catalysts. For terminal alkenes they generally give much higher selectivity to normal aldehydes than do cobalt or rhodium catalysts. [PtH( SnCl,)CO( PPh,),] (78) at high temperature and pressure converted l-pentene to n-hexenal. [PtH(SnC1,)(PPh3),] (48) and [PtH(CI)(PPh,),] (79) in presence of SnCl, were also active catalysts. In all three cases, (78) was recovered from the reaction mixtures after hydroformylation. Using (78) as catalyst the rate of the reaction was five times higher than with [Co,(CO)J under the same conditions and the selectivity to normal aldehydes was also higher.352Complex (78) may have the four-coordinate structure [PtH(CO)(PPh,),J+ SnC13-. The use of [PtC12(PPh,)2](80) in presence in SnC1, and other main group metal halides as catalysts for hydroformylation has been studied in detail.353In absence of SnC12, (80) gave no significant conversion of 1-heptene to aldehydes. Both SnC1, and GeCl, with complex (80) led to hydroformylation but they were much less effective than SnC12; PbCl,, SiC1, and SbC13 were also ineffective. Other tin halides and various phosphine, arsine and stibine ligands were investigated, but the best yield of n-octanal was obtained with the [PtClz(PPh,),]-SnClz combination. The preferred S n j R ratio was about five. The order of reactivity of different alkenes was similar

Uses in Synthesis and Catalysis

264

to that observed with rhodium and cobalt catalysts, terminal straight-chain alkenes giving the best results. It was believed that [RH( SnCl,)CO(PPh,)] was the active species formed in these reactions. The basics of the proposed mechanism are given in Scheme 19; SnC1,- and PPh, ligands are not included. Chloride was thought not to be present in view of the excess of SnC12. Isomerization can occur by reversal of the alkene insertion, as usual. The probable oxidation states of the various platinum complexes involved were not discussed, nor was the mode of hydrogen activation.353 RCH2CHZCHO

H -P t -CO

I

RCH,CH,C-Pt-CO II I O H

\\

H

H-Pt-CO

I

I

RCH,CH,-Pt-CO

Hzl I

RCH=CH,

I

H

Scheme 19

The use of an appropriate chelating diphosphine ligand greatly enhances the catalytic activity observed with the [PtClz(PPh3)2]-SnC12system.354For diphosphines of the type Ph,P(CH2),PPh2 ( n = 1-6,lO) the activity reached a maximum at n = 4, this system being roughly four times as active as those with n = 3, 5 or 6 . For n = 1, 2 or 10 the activity was minimal in comparison; the alkene used was 1-pentene. By restricting the mobility of the alkene chain of the phosphine, even better selectivity to normal aldehyde could be achieved. Using the cyclobutane derivative (55) the selectivity to n-hexenal was 99%. This required only one mole of (55) per mole of platinum, whereas in the [RhH(CO)(PPh,),]-catalyzed reactions very large excesses of PPh, are needed to attain a selectivity of cu. 94% (see Section 61.2.4.4). The strong dependence on the size of the chelate ring was interpreted as being due to the need for the phosphirle to function as both a bidentate and a monodentate ligand at differing points in the catalytic cy~le.3'~ The carbonyl complexes [PtCl,(CO)(XR,)] (X = P, As; R = aryl, alkyl) in presence of SnC1, all catalyzed the hydroformylation of 1-hexene with good selectivity to n - h e ~ t a n a l . ~It' ~ has been shown that the high selectivity towards normal aldehydes exhibited in platinum-catalyzed hydroformylation can be used to produce terminal (Le. normal) aldehydes from internal alkenes. With the cationic complex [PtCl(CO)(P(OPh),],]', again in presence of SnCI,, trans-5-decene gave up to 17% n-undecanal; with [RCl(CO)(PPh,),]+ (81) in presence of ZnBr, a linearity of 62% was achieved, though at reduced conversion.356 From the reaction of ~is-[PtCl,(PPh,)~] with 1-hexene and CO under pressure in ethanol, the complex [PtCI(CO-n-C6H13)(PPh3)21(82) has been isolated. In the presence of SnCll it catalyzes the hydroformylation of 1-hexene. The crystal structure of (82) was determined: the complex is approximately square planar, and is the tmns isomer; no unusual features were present in the structure. It is interesting that the free chloro complex was obtained although more than a threefold excess of SnClzwas present in the preparation of (82).357The complex [Pt(SnCI,)(COPr")(PPh,),] has also been isolated; it is believed that a platinum-tin bond is present in the complexes taking part in the catalytic cy~le.3~' Ph3P

c1

oEr=NCH(Me)Ph

'Pt/

'

WC H C 1311

'PPh,

0 (82)

(83)

The high selectivity shown by the platinum-tin catalysts and the importance of the industrial synthesis of normal aldehydes suggests that further study will be most rewarding. A review on hydroformylation by the platinum metals is available.359

265

Catulytic Activation of Small Molecules 61.2.4.6

Asymmetric Hydroformylation

Using [Co,(CO),] in the presence of the (S)-form of the ligand N-a-methylbenzylsalicylaldimine (83) as catalyst in ethanolic solution, styrene gave 2-phenylpropanal as its diethyl acetal with an optical purity of 15°h?60 More successful attempts at asymmetric hydroformylation have involved rhodium and platinum complexes. As in asymmetric hydrogenation, best results have been obtained with opticaily active chelating diphosphines as ligands, but some studies of monophosphines have been made. Using

/ \

Pr"

Me

(S)-(+)-methyl(n-propy1)phenylphosphine and [RhCl(C,H,,)],) styrene gave ( R ) - ( - ) - 2 phenylpropanal in an optical purity of 20-30% (equation 64). The mechanism shown in Scheme 20 was proposed for the reaction.361

co

H p*

I

%-co

'*P

co

co

lH2

Ph

H

*

I

co

I

I

co

cI \\\\\I\ Me

+

P*-Rh-P*

Me)

H'

CO Scheme 20

With [Rh,CI,( CO),] as catalyst precursor, together with ( R )-benzyl(methyl)phenylphosphine or neomenthyldiphenylphosphine, styrene was hydroformylated to 2- and 3-phenylpropanal (equation 65); 2-phenylpropanal was the major product. The optical yields were not given, but PhCHZCH,

+ H, + CO

-D

Ph-CH-CH,

I

+ PhCHZCHzCHO

CHO

the observed optical rotations showed that benzyl(methy1)phenylphosphine was several times more effective than the neomenthyl In a related study of the asymmetric hydroformylation of styrene, [Rh,Cl,(CO),] gave 2-phenylpropanal with optical purities of 17.5 and 21% using as ligands (+)-benzyl(methy1)phenylphosphine and (-)-methyl( n-propy1)phenylphosphine respectively. Neomenthyldiphenylphosphine gave an optical yield of less than 1 % Using [RhH(CO)(PPh,),] (34)in the presence of DIOP (49) the optical purity was 23%. Simple aliphatic alkenes were also studied using (34) and (49) and cis-2-butene gave 2-methylbutanal in an optical yield of 27% as the best result.360With [Rh,(CO),,] and the chelating bis(dibenzophosphoIy1) ligand (84) at a phosphorus/rhodium ratio of 4,styrene gave 2-phenylpropanal in an optical yield of 40%. Both 1- and 2-butenes and 2-phenylpropylene lead to significantly poorer r e ~ u I t s ? ~ ~ N-Vinylsuccinnimide and N-vinylphthalimide were hydroformylated using complex (34) and the ligands DIOP (49), DIPAMP (50) and DIPHOL (85). Optical yields for the resulting 24midopropanals of up to 27% for the succinnimide and 38% for the phthalimide derivatives could be obtained, in both cases with (85) as ligand.364 I

266

Uses in Synthesis and Catalysis

An attempt to use the 3-trifluoro-1R-camphorate complex (86) as a catalyst for the asymmetric hydroformylation of styrene failed as the ligand dissociated rapidly under the reaction condition~.~~~

(85) DIPHOL

(84)

(86)

The majority of studies of asymmetric hydroformylation with rhodium and platinum complexes have made use of DIOP (49) as a ligand. With either the complex [RhCl(CO)(DIOP)] or [RhC1(C2H,)2]2 plus DIOP, styrene was hydroformylated to 2-phenylpropanal with optical yields of only 16%.366When a-monodeuterostyrene was used as substrate, with DIOP and complex (34) as catalyst, essentially the same optical yield was obtained.367The same catalyst with non-deuterated styrene under different conditions gave an optical yield of 25% ?hS In diene hydroformylation with complex (34) and DIOP, the attainable optical yields were critically dependent on the diene. Thus isoprene gave, after oxidation of the aldehyde with Ag20, 3-methylpentanoic acid as major product having an optical purity of 32%. Butadiene, 2,3dimethylbutadiene and 2-methyl-1-butene all gave optical yields of less than 6% ?69 The hydroformylation by (34)and DIOP of a number of alkenes having C,, symmetry gave best results with cis-2-butene, with an optical yield of 27% .370 [Rh2Clz(CO),] together with DIOP, benzyl(methy1)phenylphosphine or neomenthyldiphenylphosphine has been investigated in the hydroformylation of vinyl acetate (equation 66). 0

1

OCMe

0

II

CH,=CHOCMefH,+CO

I

+

Me-C-H I

(66)

CHO

The monophosphines gave optical yields of 2.1 and 0.5% respectively, but with DIOP a best result of 23% was obtained.37’The product of this hydroformylation is a potential precursor for the amino acid threonine. When [Rh(acac)(cod)] was used with DIOP as ligand, a best optical yield of 40% was achieved. With the related ligand DIPHOL (85) the optical yield was raised to 51%. The chemical yields of the 2-acetoxypropanal were in the range 75-95%.372 The complex [PtCI,(DIOP)] in the presence of SnCl, has been used as catalyst for the asymmetric hydroformylation of various alkenes. With a series of butene and styrene derivatives very low optical yields were obtained. The best results were achieved with 2,3-dimethyl-l-butene which gave 15% optically pure 3,4-dimethylpentanal, and with a-ethylstyrene which gave 15% optically pure 3-phenylpentanoic acid after oxidation of the aldehyde.373 Using D2and CO, with [PtCl(SnCl,)(DIOP)] as catalyst, ( Z ) -and (E)-2-butene were converted to eryfhro- and threo- 1,3-[2H]z-2-methylbutanal respectively with high selectivity. This indicates that the H and CO groups are added in a cis manner to the double bond, as is the case with [RhH(CO)(PPh,),] (34).374Using [PtCl2(DIOP)]-SnClz or the DIPHOL (85) analogue, it was shown that a given enantiomer of the ligand gave different enantiomers of the resulting 2~’ to a previous report.373Here methylbutanal when starting with (2)-or ( E ) - 2 - b ~ t e n e , ~contrary again, the behaviour is the same as for the rhodium catalyst. That very high optical yields can be obtained in asymmetric hydroformylation has now been demonstrated. Using [PtCl,(-)-(DTOP)] and SnC12 in benzene solution, careful choice of the reaction conditions permitted the conversion of styrene to (+)-2-phenylpropanal in an optical ~ DIOP the best result was 35%. Even slight variation of the reaction conditions yield of 9 5 ’ / 0 . ~ ’With caused the optical yield obtained with DIPHOL to fall significantly, indicating clearly that to obtain the best possible results in asymmetric catalysis, all reaction parameters must be optimized.

Catalytic Activation of Small Molecules 61.2.5 61.2.5.1

267

CARBONYLATION Introduction

Whilst the number of carbonyl ligand-containing complexes known today is vast, only relatively few have been shown to take part in homogeneously catalyzed reactions. Coordination and activation are therefore not synonymous. The bonding involves donation from the carbon lone pair to an empty metal orbital, and back donation from a filled metal d-orbital to an empty antibonding .rr*-orbital of the carbonyl ligand (87).

(87)

The relative extent of this u-donation and wacceptor behaviour determines the electron density in the neighbourhood of the carbon atom and consequently its reactivity. This electron density is further very dependent on the transition metal, its oxidation state and the other ligands in its coordination sphere. The frequency of the CO stretching vibration in the IR spectrum gives an indication of the resulting electron distrib~tion.~’~ The higher this frequenc the more positive is the carbon atom and the greater is its susceptibility to nucleophilic attackJ8 Unfortunately, the carbonyl intermediates occurring in catalytic carbonylations are rarely amenable to IR study. The so-called insertion reaction (equation 4) is the most common way in which a new bond to the carbonyl carbon atom is formed. It should be noted that the reaction is misnamed as a more correct picture involves the migration of the group X to the carbonyl carbon, rather than an insertion of CO into the M-X bond.379This may therefore be regarded as a nucleophilic attack on the carbonyl carbon atom and the relevance of it bearing a partial positive charge is apparent. In many carbonyls no significant change in the electron density at the carbon atom occurs, and they are thus not activated by coordination. A review is available on this Carbonylation by binary metal carbonyls is outside the scope of this work and the companion volumes ‘Comprehensive Organometallic Chemistry’ should be consulted.

61.2.5.2

Ruthenium

The carbonylation of methanol, dimethyl ether and methyl acetate using as catalyst precursors [RuI,(CO),] and [Ru(acac),] (88) with NaI, HI or Me1 as promoter has been reported.381The major products were ethanol from methanol, methyl acetate from dimethyl ether, and acetic anhydride plus acetic acid from methyl acetate (equations 67-69). MeOH+ CO + 2H2

-+

EtOH + H,O

MeOMe+CO

+

MeCOMe

(67)

tl

0 MeCOMe+CO

II

0

+

MeCOCMe I1 )I 0 0

Further carbonylation can occur, adversely affecting the selectivities obtained. Where it is used, the effect of hydrogen pressure on selectivity is also significant, as not all of the reactions require hydrogen. The production of water and the presence, of alcohols lead to esterification-hydrolysis equilibria and water can affect the hydrogen pressure via the reversible water-gas shift reaction (equation 70). H,O+CO

-+

H,+CO,

(70)

Relatively high temperatures and pressures are required for these carbonylations. When complex (88) is used as catalyst precursor in presence of an iodide source, two ruthenium(I1) complexes ~ ] and fac-[RuI,(CO),]- (90). have been shown to be formed. They are u i ~ - [ R u I ~ ( C 0 )(89) Use of (88) requires the presence of hydrogen to reduce it to ruthenium(I1). Compounds (89) and (90) are believed to be in equilibrium with one another under the reaction conditions, and

Uses in Synthesis and Catalysis

268

co

I

1

CO

co (90)

(89)

complex (90)is thought to initiate the catalytic cycle. Hydrogen or a proton source (H20, MeOH, HI) is essential. Complex (90) or one similar reacts to give a ruthenium methyl species, which then undergoes CO insertion to give an acyl. Reaction of this with hydrogen, water, methanol or iodide could then lead to the various products observed, but the full details of the mechanism remain unknown. A further reaction of the above type is ester homologation. The production of ethyl acetate from methyl acetate (equation 71) using similar catalytic systems to the above has been s t ~ d i e d . ~ ” 0

0

II

MeCOMe+C0+2H2

11

MeCOEt+H20

-+

(71)

Complex (88) with an iodide source was again used. The selectivity was surprisingly dependent on the cation of the iodide. It has been found that cobalt(I1) acetate improves the conversion and selectivity achieved in equation (71) when used together with complex (88) as catalyst.383In a related reaction, acetic acid was carbonylated to propionic acid (equation 72). MeC0,H

+ 2H2+ CO

-+

EtC0,H

+ HZO

(72)

The ruthenium source was typically Ru02, RuC13, R ~ ( a c a c ) [~R, U ~ ( H ) ~ ( C Oor) ~a ~binary ] MeI, Et1 or HI was employed. carbonyl, again with a source of iodide present as a ) ~ ] observed in the reaction mixtures. The reaction gave Both complex (90) and [ R U ~ I ~ ( C Owere carboxylic acids of longer chain length than propionic acid. A mechanism was proposed in which the precise nature of the iodocarbonyl complexes was not specified (Scheme 21). MeC0,H

+ HI

MeCOI + H 2 0

+

0

0

I1

/I

CH,CI+RUI,_~(CO), --c Ru(-CCH,)I,(CO),

1

0

II

Ru(-CCH2CH3)I,(CO),

Ru(CH2CH,)I,(CO),+

h20

RuI,-,(CO),

+ CH,CH,CO*H+

HI

Scheme 21

The polynuclear complex [Ru(OAc),(CO),], (91) has been shown to catalyze the carbonylation of secondary amines to formamides at normal pressure and 75 “C (equation 73). Some primary amines also r e a ~ t . 3 ~ ~ ~ ~ ’ ~ RR”H+CO

---c

RR’NCHO

(73)

In amine solution, (91) gives dinuclear species [Ru(OAc),(CO),( RR’NH)J (92). The carbonylation is autocatalytic, suggesting that these complexes must undergo further reactions before Me

I

I

Me (92)

Catalytic Activation of Small Molecules

269

becoming catalytically active. The polynuclear hydridocarbonyl complex [RuH(CO),], when used as catalyst precursor showed no autocatalytic behaviour, and possibly reacts with the amine to enter the catalytic cycle directly. [Ru(acac),] (88) has also been used as catalyst precursor for the reduction of CO with H2.388 These reactions require very high temperature and pressure. The only products formed by CO reduction were methanol and methyl formate. Complex (88) is carbonylated under the reaction conditions and some carbonyl complexes can be used in its place. [Ru(CO),j was shown to be formed in all cases in the reaction mixture. The use of acetic acid as solvent promoted the formation of ethylene glycol esters in these reactions.389Complex (91) was also used and again [Ru(CO),] was observed. When iodide was added to the system, the activity and selectivity to the two carbon products ethanol and ethylene glycol increased ~ignificantly.~~' Using KI, no [Ru(CO),] was detected, but both complex (90) and [Ru,H(CO),,]- were present. Neither of these complexes when used alone gave more than very limited activity. Optimum performance required a 2 : 3 ratio of the hydride to (90).Using catalyst precursors such as RuC13.3H20, [Ru(acac),], [Ru(CO),(PPh,),] (65) and others, the yields and selectivities to C, products have been found to be markedly affected by the presence of bulky cations, possibly owing to their effect on the polarity of the medium.391Attempts to hydrogenate CO to C2products in n-propanol ~ ] or RuCl,.3H20 led only to the formation of solution using [RuC12(PPh3),] (2), [ R ~ ( a c a c ) (88) propyl formate and propyl acetate.392 The water-gas shift reaction (70) is of great importance in adjusting the H,/CO ratio in synthesis gas mixtures, and attempts have been made to find homogeneous catalysts capable of better performance than the heterogeneous catalysts currently used in industry. In aqueous KOH, RuCI3.3H20acts as a catalyst for the reaction at 90 "C and near normal pressure.393The porphyrin complex [Ru(TPPS)C0I4- (TPSS = meso- tetrakis(4-sulfonatopheny1)porphyrinate) behaves similarly and both gave better results than did [Ru3(CO),J, The complex [R~Cl(CO)(bipy)~]+ has been shown to catalyze the water-gas shift reaction in both photochemical and thermal reactions.396The details of the mechanism are not yet clear, but attack of hydroxide on a carbonyl ligand to give a formate complex is involved. The complexes [Ru(cod)pyJ2+, [RuCl2(C0),py2] and [RuCl,py,] catalyze the reduction of nitrobenzene by carbon monoxide and water (equation 74).397The cationic complex gave the best results, and a number of substituted nitrobenzenes were reduced in good yields. Unlike some other catalysts for this reaction, it did not catalyze the water-gas shift (equation 70). PhN0,+3COfH20

2NO+CO

--c

+

PhNB,+3C02

N,O+CO,

(74)

(75)

[Ru( NO),(Pph,),] catalyzes the reduction of nitric oxide by carbon monoxide (equation 75).3y8 61.2.5.3

Cobalt

The carbonylation of methanol is an important route to acetic acid and is the basis for two industrial processes (cobalt- and rhodium-catalyzed).The cobalt-catalyzed route uses [CO,(CO)~] as catalyst with an iodide promoter, and lies outside the scope of this work. The companion volumes 'Comprehensive Organometallic Chemistry' should be consulted. The conversion of methanol to ethanol with carbon monoxide and hydrogen has attracted considerable attention. Further carbonylation to higher alcohols occurs much more slowly, but acetic acid formation i s a competing reaction and this leads to ester formation. Using CoI, in presence of PBu", as catalyst, the selectivity to ethanol was improved by addition of the borate ion B4072-.3w This was attributed to an enhanced carbene-like nature of an intermediate cobaltacyl complex by formation of a borate ester (equation 76). This would favour hydrogenolysis to

M e k o

+ (HO)*BO-

Me&-&

ethanol rather than hydrolysis to acetic acid. Using Co12 and the diphosphines Ph,P(CH,).PPh, ( n = 1 4 ) , selectivities to ethanol of up to 89% were achieved.400The reaction sequence given in Scheme 22 was proposed. No further details regarding the nature of the cobalt complexes involved were available. The diphosphines having n = 2 or 3 gave the lowest selectivityto ethanol (35-65%). These ligands should form the most stable chelate rings with the metal. For n = 1, 4 or 5, the selectivities were of roughly the same order (81-89Oh). The extent of formation of acetic acid,

Uses in Synthesis and Catalysis

270

methyl acetate and propanol was negligible, but significant quantities of methane were obtained. The high selectivity to ethanol was attributed to the ease of hydrogenation of the acylcobalt intermediate. MeOH

EtOH

HI

Me1

-.

H

CO

MeCoI -A CH,

H

+&MeCHO + Co

fl

0

11

MeCCol

-

ROH

MeCOR

(R= Me, H) Scheme 22

Using Co(OAc), in presence of PPh, and HI as catalyst precursor the mechanism of the reaction has been studied.” A number of cobalt complexes having I-, PPh, and CO as ligands were also investigated and [CoI,(PPh,),] gave the best results. When CD30D was used, no H/D exchange occurred in the methyl group, ruling out the possibility of carbene formation according to equation MeOH+Co

-+

H,C=Co+H20

(77)

(77). The acyl route (Scheme 22) was accepted and the detailed mechanism given in Scheme 23 was evolved. MeOH + H I

MeCHO

-

MeI

+ H20

(78)

!

7 Co(COMe)(CO)(PPhy), Co(COMe)(H),(CO)(PPh& -

HZ MeCHO

f

H2 % EtOH

Scheme 23

There remains the question of whether the acyl group can be further reduced without elimination of acetaldehyde, giving ethanol directly. The reductive elimination of iodine is perhaps in need of further substantiation. Hydrogenation of the acyl complex [ C O ( C O M ~ ) ( C O ) ~ P P ~ gave , M ~ ]a product distribution very similar to that obtained when [ C O ( C O ) ~ P P ~ ~was M ~used ] as a catalyst for methanol homologation.m2 This was regarded as further evidence for the presence of an acyl intermediate in the catalytic reaction. [CoNO(CO),] (93) catalyzes the carbonylation of benzyl bromides and chlorides to give carboxylic The reactions, carried out in a two-phase system, occurred at normal temperature and pressure and their behaviour was quite different to that observed when [Co,(CO),] is used as catalyst. The anion [CoH(NO)(CO),]- is believed to be a key intermediate, formed by attack of hydroxide on (93) followed by loss of CO,. Ketones, bibenzyls and other by-products were detected in low yields. The mechanism of the carbonylation is given in Scheme 24. Co(OAc), in the presence of sodium hydride and a sodium alkoxide has been used to catalyze the carbon lation of aryl bromides, giving mixtures of carboxylic acids and esters, again at normal pressure.40: When amines were present, amides were formed. Unfortunately, nothing is known about the nature of the cobalt complexes involved.

Catalytic Activation of Small Molecules

27 1

NO

I

C O1111111 CO

oc'

5[Co(CO2H)(C0),NO]

'co (93)

kco, [CoH(CO),NO] -

co CoH(COBz)(C0)2NO

CoH(Bz)(CO),NO

Scheme 24

From the hydroesterification of 1-decene the cluster [Co,( H),py5(CO),] has been is0lated.4~~ In the presence of 2-vinylpyridine, intermediates having the partial structure (94) were detected by NMR spectroscopy. This was believed to be an intermediate in the catalytic hydroesterification of 2-vinylpyridine to methyl 3 4 2-pyridy1)propionate (equation 79).

In presence of diphos, [Co,(CO),] catalyzes the reaction of propylene with CO and H 2 0 to give aldehydes (Reppe hydroformylation, equation &0):06 The reaction requires high temperature

CH,CH=CH,+ 2CO+ H,O

+

CHO I CH,CH,CH,CHO+CH,CHCH,

-t CO,

(80)

and pressure. Induction periods were observed and were attributed to formation of hydrido complexes of uncertain composition, [CoH(CO),(diphos),], which lead to catalysis. In the presence of a large excess of propylene, di-n-prapyl ketone became the major product. Under these conditions the ligands PBu, , PPh, , P(OPh)3, Fh,PCrCPPh,, PhzAs(CHz)2PPh2and PhzP(CH2).PPhz ( n = 1-4) were all investigated. The best ligands for ketone formation were diphos and Ph2PC=CPPh2?'Y7Using [Co,(CO),] and diphos, ethylene gave diethyl ketone with a selectivity of 99%.408 The suggested general reaction mechanism is given in Scheme 25. C,H, CoH(CO),(diphos),

Co(Et)(CO),(diphos),

CO, + CoH(CO),(diphos),,

1

Co(COEt)(CO),-,(diphos),

2CO+H,O

I

Co,(CO),,-,(diphos),,

EtzCO

tf-

Co(CH,CH,COEt)(CO),-,(diphos),

CoH(CO), (diphos), Scheme 25

The small amount of propanal formed in the reaction (less than 1%) arises from the reaction of the hydride with the acyl. The same catalyst system converted methyl acrylate to dimethyl

272

Uses in Synthesis and Catalysis

4-oxopimelate (equation 81) in a yield of 94%, and a mechanism analogous to Scheme 25 was proposed.409 CH,=CHCO,Me+

2CO + H 2 0 'c02(co'81 ~i~~~ w OC(CH,CH,C02Me),

(81)

The synthesis of bifurandiones from acetyylenes and CO using [Co,(CO),] in presence of PBun3 or P(OMe), has been briefly r e p ~ r t e d . ~ "

61.2.5.4

Rhodium

As mentioned in the previous section, the carbonylation of methanol to acetic acid is an important industrid process. Whereas the [C~~(CO)~]-catalyzed, iodide-promoted reaction developed by BASF requires pressures of the order of 50 MPa, the Monsanto rhodium-catalyzed synthesis, which is also iodide promoted and which was discovered by Roth and co-workers, can be operated even at normal pressure, though somewhat higher pressures are used in the production units.41 1-413 The rhodium-catalyzed process gives a methanol conversion to acetic acid of 99%, against 90% for the cobalt reaction. The mechanism of the Monsanto process has been studied by F o r ~ t e r . ~ The ' ~ anionic complex cis-[RhI,(CO),]- (95) initiates the catalytic cycle, which is shown in Scheme 26. Me

Ij

1'

'co

(95)

I

MeCOzH

bo hHZ0 r

co

1

Scheme 26

Reaction (78) regenerates Me1 from methanol and HI. Using a high-pressure IR cell at 0.6 MPa, complex (95) was found to be the main species present under catalytic conditions, and the oxidative addition of Me1 was therefore assumed to be the rate determining step. The water-gas shift reaction (equation 70) also occurs during the process, causing a limited loss of carbon monoxide. A review of the cobalt-, rhodium- and indium-catalyzed carbonylation of methanol to acetic acid is a~ailable.4'~ Kinetic studies of the acetic acid synthesis catalyzed b RhCl,.3H2O have confirmed that the The effectof nitrogen and phosphorus oxidative addition of Me1 is the rate determining ligands on the reaction has been studied, and bidentate ligands were shown to inhibit the reaction almost ~ o m p l e t e I yA . ~ kinetic ~~ study of the solvent dependence of the reaction showed that

'

Catalytic Activation of Small Molecules

273

solvents of medium polarity such as ketones led to both high selectivity and high activity when RhC13.3H20or [ Rh,Cl,(CO),] (96) was used as catalyst precursor!20

Ethanol421is carbonylated to propionic acid, and is carbonylated to n- and iso-butyric acids. In the latter case it is not known whether the isomerization occurs in the alcohol, the iodide or the rhodium-alkyl complex. In presence of hydrogen, RhCl, and [RhH(CO)(PPh,),J (34)catalyze the homologation of methanol to ethanol, again with methyl iodide as promoter. The mechanism was believed to involve the oxidative addition of methyl iodide followed by CO insertion to give an acyl. This then either underwent direct hydrogenation to ethanol or gave acetaldehyde which was subsequently reduced. The sequence generates HI which reacts with methanol according to equation (78), regenerating methyl iodide.423 Dimethyl ether has been carbonylated in two steps to give acetic anhydride (equation 82). The first step is the faster of the two. The catalyst was RhC13.3H20in presence of PPh3 and [Cr(CO>,]. Nothing appears to be known about the rnechani~rn.~" 0

MeOMe+CO

+

II

MeCOMe

0

%

0 0 II II MeCOCMe

0

II

MeCCH2CH20R+C0 -D

0

II

I1

MeCCH,CH,COR

(83)

Esters of levulinic acid have been synthesized by carbonylation of 4-alkoxy-2-butanones (equation 83). Methyl iodide was used as promoter and the best results were obtained with [RhCl(PPh,),] (23),[Rhl(CO)( PPh,),] or RhCI3-3H20and PBu", as catalyst precursors. Cleavage of the ether linkage with HI to give 4-iodo-2-butanone was proposed as the initial step in the reaction.425[RhCl(CO)(PPh,),] (71) has been shown to catalyze the carbonylation of epoxides to p-lactones, but uncertainty exists regarding the mechanistic details.426The homologation of acetic acid is catalyzed by complexes such as [Rh(acac),], RhC13and Rh203.3x5 The main product is propionic acid but butyric and valeric acids were also formed. The carbonylation of methyl acetate to acetic anhydride is likely to become an industrial process in the near RhC13.3H20is typically used as catalyst precursor and an iodide promoter is used. A mechanistic study indicated that methyl iodide formed from the ester and HI is carbonylated as in acetic acid synthesis (Scheme 26). The resulting acyl, perhaps via reductive elimination of acetyl iodide, converts the acetic acid formed in the ester cleavage to acetic a n h ~ d r i d e . ~ [RhI(CO)( ~ ~ - ~ ~ ' PPh,),] also catalyzes the reaction though apparently more slowly The mechanism of this reaction is given in Scheme 27. than complex (95).430*431 Me1 Rhl(CO)(PPh3)2

Rh(Me)l2(CO)(PPh>)z

The five-coordinate acyl complex may not be mononuclear. Using a rhodium complex in presence of a diamine, alkenes were converted to alcohols according to reaction (84). RCH=CHz+ 3CO + H20

+

RCH?CH,CH20H

OT

RCHCH,

I

CH,OH

+ 2C0,

274

Uses in Synthesis and Catalysis

[Rh,Cl,(CO),] (96),[RhCI(PPh,),] (23) and RhC1,.3Hz0 were all employed as catalysts though none was as good as [Rh,(C0),6]. The most suitable diamines were Me2N(CH2),NMe2( n = 2,3) and 4-dimethylaminopyridine. The linear alcohols predominated.432If amines are present in reactions such as (84), aminomethylation takes place. This provides a ready synthesis of secondary and tertiary amines containing a -CH,+ N E group. A typical example is given in equation (85).433

3

+ 3CO + H20 4- HN

+

o - C H 2 - N a

Although the reaction was first discovered more than 30 years ago, the mechanism is still not certain. An acyl complex was believed to be an intermediate. This then reacted with the amine to give an imine or enarnir1e.4~~ The complexes [RhCl,py,], [RhC1(NHJ)5]2' and [Rh,(CO),,J all catalyzed the reaction with very similar conversion and selectivity, and in view of the high temperature and CO pressure it seems likely that similar carbonyl complexes are involved in all three cases. [RhCl(PPh,),j (23) and [RhCl(CO)(PPh,)J (71) catalyze the formation of enol silyl ethers and derivatives according to equation (86). Both the (E) and ( 2 ) isomers of the product are ~btained.~"" The dinuclear complex [Rh2C1z(CO)4](96) catalyzes the reaction of phenylacetylene with carbon monoxide to give 3,6-diphenyl-l-oxabicyclo-[3.3.0]-octa-3,6-diene-2,4-dione (97) (equation 87).435b CO + HSiEt,Me

n-C,H,CH=CH,+

2PhC=CH+3CO

+

n-C,H,,CH=CHOSiEt,Me

-

(86)

0

Ph (97)

In presence of an alkene, furanone derivatives were obtained (equation 88).436The hydrogen needed for the reaction was derived from the solvent, ethanol. [RhCI(CO)(PPh,),] (71), [RhCl(PPh,)J (23) and RhC13.3H20 as well as binary carbonyls could all be used as catalysts. The mechanism shown in Scheme 28 was proposed.

PhC-CPh+CO+C2H,(+H1)

P h e 0 Ph Et

co

GH4

Rh-H

4

Rh-Et

A Rh-COEt

I

PhClCPh

Et

I 0

Et

Et

Scheme 28

In presence of a mild base such as Na2C03 or NaOAc, a hydroesterification to furanones i! possible (equation 89).437RhC13-3H20or a binary carbonyl was used as catalyst. lndanom derivatives were obtained as by-products. The mechanism probably involves the formation of at alkoxycarbonyl intermediate (Scheme 29).

Catalytic Activation of Small Molecules

PhC=CPh+2CO+EtOH

---L

EtORh-CO

275

P h e o Ph OEt

PhCGCPh Rh--CO,Et 0~~~~~ I

0

OEt

:'"

CO, H+

0

0

OEt

Scheme 29

[ Rh(acac)(CO),] in presence of 2-hydroxypyridine has been shown to catalyze the hydrogenation of CO at high temperature and pressure. The reaction is of poor selectivity, more than a dozen C , , C, and C3 products being formed.438Ethylene glycol was the major product and formaldehyde was believed to be a key intermediate. Although its formation from H, and CO is thermodynamically very unfavourable, it would only need to be present in low concentration. Many rhodium complexes catalyze the water-gas shift reaction (equation 70). [RhCl(PPh,),f (23), [RhC1(cod)],, [Rh(nbd)(diphos)]+, [Rh(cod)diphos)]+, [Rh( cod)(PMePh2)J+ and [ Rh(cod)(mSPPh,),]- (mSPPh, = 31) all catalyzed the hydrogenation of benzalacetone using CO and H 2 0 as the hydrogen s0urce.4~~ A similar hydrogenation of methyl crotonate has been described, in which trans-[Rh(OH)(CO)(PPr',),J (98) is the catalyst.440This complex was formed in situ from [Rh(H),(02COH)(PPr'3)2]or from [RhCl(C,H,)], and PPri3 and Bu"Li. In these reactions, hydroformylation also occurred.

(98)

[Rh,Cl,(CO),] (96) in acidic solution also catalyzes the water-gas shift. In presence of excess iodide, [RhI,(CO),]- (95) is believed to be the active rhodium species and the reaction required only mild conditions."'.442 The proposed mechanism is given in Scheme 30. HI

[RhL(CO)21COZ

+ HI

[RhH(I),(CO),I-

i co Scheme 30

Other iodocarbonyl complexes might be present, depending on the CO pressure. In a further study of the water-gas shift reaction a large number of phosphine-containing rhodium complexes were investigated as catalysts. They were mainly either neutral or cationic carbonyl derivatives

CCCK-J

Uses in Synthesis and Catalysis

276

and are too numerous to mention here.M' The mechanism was studied using complex (98) as precursor, and differs from that with the phosphine-free complex (95) in that carbonate intermediates are involved. The chemistry is summarized in Scheme 31. As no kinetic study was undertaken, the relative importance of the different pathways is not known. The solvent (S) was typically acetone or pyridine. RhfOHMCO)

wio#

% RCo2 e

co,

H2, S

RhH(OCOzH)(CO)LZ

d

(L = PPr', , S = solvent)

H,

co2 C O Y H 2 0

RhH(CO)L2

co

H2kO3

Scheme 31

The complexes [ R h ( b i p ~ ) ~ ][RhC12(bipy)J+ ~+, and [Rh(OH2),(bipy)2]3' also catalyze reaction (70), with the latter two being most effective. Using sulfonated bipyridyl or phenanthroline

derivatives, water-soluble catalysts were prepared.444 If [RhCl(CO)(PPh,),] (71) is treated with an aryl azide such as p-tolyl azide in presence of CO, reactions (90) and (91) occur. RhCI(CO)(PPh,)2+ MeC6H,N, RhCl(N)2(PPh,)Z+CO

--t

MICI(N),(PP~,)~+ MeC6H4NC0

-+

RhCl(CO)(PPh3)2+N*

The nitrogen complex has only transient existence. These reactions permit a catalytic carbonyla. ~ ~complexes [RhX(CO)LJ ( X = C1, Br, F; tion of azides to isocyanates at low CO p r e s ~ u r eThe L = P(OPh), , PPh,, P(4-CIC6H4>,, P(4-MeOC6H&, P(C6H1,),, PBu",) were investigated and [RhCl(CO)(PBu",),] gave the highest rate for the carbonylation of p-tolyl azide. In a separate study, complex (71) and [Rh(diphos),]* were shown to catalyze the same azide carbonylation and also to give ureas in presence of aniline derivatives (equation 92) and carbamates in presence of alcohols (equation 93).446,447 p-RC6H,N,

+ CO +p-R'C,H4NH2

P - R C ~ H ~+CO N ~ +EtOH

4

p-RC,H,NHCONHC6H,R'-p

(92)

+

P-RC~H~NHCO~E~

(93)

In a reaction related to azide carbonylation, aromatic nitroso compounds also give isocyanates (equation 94). [Rh,Cl,(CO),] (96), [RhCI(CO),pyj, [RhCl(CO)(PPh,),] (71) and [RhCl(CO){P(OPh),},] all catalyzed the reaction. Deoxygenation of the nitrosobenzene derivative was the rate determining step.M8Nitrosomethane can be similarly ~ a r b o n y l a t e d . ~ ~ - ~ ~ ~ R N 0 4 2 C O + RNCO+CO2

(94)

The more usual way of producing isocyanates by catalytic carbonylation employs the more readily available nitro compounds as starting materials (equation 95). There are a number of PhN02+ 3CO

4

PhNCO + 2CO2

(95)

reports on rhodium catalysts for this reaction. The complexes used include RhCl3.3H20 and [Rh,Cl,(CO),] (96),452[RhH(CO)(PPh,),] (34)453and complex (96) in presence of pyridine454or in presence of a metal halide such as MoCls.455-457 The mechanism proposed in the latter case is given in Scheme 32. MoCl, is believed to act as a Lewis acid which coordinates to one atom of the nitro group. Reaction of the resulting RN(O)OMoCl, entity with complex (96) leads to cleavage of the chloride bridges to give (99). The principal advantage of the presence of MoC15 is that it permits the reaction to be carried out at normal pressure, though raised temperature is still needed.457 A further reaction in which CO acts as a reducing agent is that with nitric oxide (equation 96). C0+2NO

4

COz+N,O

(96)

Catalytic Activation of Small Molecules

277

RNO,, MoCIS

I1

[ R h C I(CO)z]

%. /"-P 7 4 R N C O , MoCl5,

OdRh\

/

MoCls

C-N,

co2

R

Scheme 32

It has become of increased importance because of environmental problems relating to oxides of nitrogen. With RhC13.3H20 as catalyst the reaction proceeds slowly at normal temperature and pressure.398The anionic complex [RhCl,(CO),]- has also been used as a catalyst and this leads to the formation of [RhCl,(CO)(NO),]- (100) which is believed to take part in the catalytic Complex (100) contains rhodium(III), as judged by the frequency of the CO stretching vibration and the nitrosyl ligands were regarded as coordinating as NO-. At the end of the reaction, [RhCI,(CO)2]- was recovered quantitatively. T h e reaction requires aqueous acid. Two possible mechanisms were discussed, both of which involved complex ( In a more recent study, [Rh(N0),(PPh3),]+ has been used as catalyst for reaction (96).460This system also shows an increased rate in presence of water, but this is far less marked than with complex (100). Dinuclear complexes form during the reaction but do not appear to be catalyticallf active. The details of the mechanism are not known at present. CH2=CHCH2NH2

+

CO

--+

a.

(97)

N

H

Some other rhodium-catalyzed reactions of CO are collected here. The cyclization of allyl amine to y-butyrolactone (equation 97) is catalyzed by the complexes RhCl,, [Rh(acac),], [Rh2C12(C2H4)4]r [RhCl(PPh,),] (23)and [RhCl(CO)(PPh,),] (71). If allyl chloride or iodide is used together with ammonia or a primary amine, analogous cyclizations occur?61 2-Arylazirines are carbonylated by [Rh,Cl,(CO),] (96) to give isocyanates under mild conditions?62 The reaction can be carried out at normal pressure and at room temperature or below. Vinyl isocyanates are produced (equation 98). By addition of alcohols or amines, carbamates or ureas were obtained.

In presence of RhCI3*3H20,[Rh,Cl,(cod),], [RhCl(PPh,),] (23)or [RhCl(CO)(PPh,)2] (71), alkylation of amines could be achieved using an aldehyde, CO and water?63 RhC1,.3H20 effected the a-methylation of ketones in the presence of aqueous formaldehyde.464RhC13.3Hz0 or [RhCl(nbd),] catalyzes the formation of quinoline derivatives from aromatic nitrocompounds and aliphatic aldehydes again in presence of CO and ~ a t e r . 4 The ~ ' aromatic compound is partly converted to the dialkylaniline derivative (equation 99). In all of these reactions the water-gas shift (equation 70) or a related reaction provides the reducing power.

The carbonylation of cyclopropane catalyzed by [Rh2C12(CO)4]has been The selectivity of the reaction was poor, with up to eight products being formed. Cyclobutanone was the major carbonylation product.

27 8

Uses in Synthesis and Catalysis

A number of general reviews on carbonylation are available, which include rhodium-catalyzed reactions. They are listed at the end of Section 61.2.5.8.

61.2.5.5

Iridium

Iridium complexes i n the presence of iodide promoters are, like rhodium complexes, efficient catalysts for the carbonylation of methanol to acetic a ~ i d . ~Complexes ~ ~ , ~ of~ the ~ , type ~ ~ ~ [IrCl(cod)L] (L = C5H5N,3-C1CsH,N, 2-MeC5H,N) and [Ir(cod)L’]+(L’= phen, bipy) have been used as precursors with various iodide promoters-467The cationic complexes gave the highest activities. It was shown that, under reaction conditions, anionic iodocarbonyl complexes of iridium were formed. IrCl, can be used as catalyst precursor and methyl iodide appears to be the best promoter.416The iridium-catalyzed carbonylation of methanol is mechanistically more complicated The reaction can involve either neutral or anionic complexes of iridium than that of depending on the reaction conditions, particularly on the acidity and the iodide concentration. The complexes [Ir12(CO)2]- and [1rI3(CO)J were isolated from the reaction mixture by addition of iodide and iodine respectively to it. This suggests that the complex [IrI(CO),] is present in solution. The water-gas shift reaction (equation 70) can occur to a significant extent during iridium-catalyzed methanol carbonylation. The proposed mechanism is shown in Scheme 33.468 CO, HI

lrH(MC0)dOHd

HZ

A

HI Irll(CO)3 W k i ( I ) d C O ) J

Hz [ I ~ L( CO ) J

lrI(CO)2

4 I

Me1

Ir(COMe)12(C0)3

[Ir(COMe)13(CO)Z1-

I

Scheme 33

The acetyl iodide formed reacts with water or methanol to give acetic acid or methyl acetate. The latter is subsequently hydrolyzed to acetic acid. Methyl iodide is regenerated according to equation (78). The synthesis of acetic acid from methanol catalyzed by complexes of cobalt, rhodium and iridium has been reviewed.469 Iridium complexes in the presence of iodide also catalyze the carbonylation of methyl acetate to acetic anhydride (equation 69). The reaction mechanism is similar to that of Scheme 33. The ester reacts with HI to give methyl iodide which is carbonylated as in Scheme 33 to acetyl iodide. This reacts with acetic acid to give the anhydride.4295430 If ethylene is present during the carbonylation of methanol catalyzed by IrC14, once again with Me1 as promoter, methyl propionate is formed.416The reaction depends on the presence of iridium hydride species in solution, and a rhodium analogue of the reaction exists. The full details of the mechanism are not known but the basic steps are shown in Scheme 34. The intermediates are all believed to be complexes of iridium(II1). As mentioned above in connection with the acetic acid synthesis, iridium complexes catalyze the water-gas shift reaction (equation 70). From IrClB.3H20and sulfonated derivatives of bipy and phen, water-soluble catalysts were Using dioxane as solvent, complexes of the type [Ir(cod)LZ]~ ( L = PMePhz, PPh3), [Ir(cod)L’]+ (L’=diphos, phen, 4,7-Me2-phen, 4,7-Ph2phen, 3,4,7,8-Me4-phen) and [Ir(cod)X]- (X = 4,7-diphenylphenanthroline disulfonate) also catalyzed the reaction, with the anionic species being most The mechanism was thought

Catalytic Activation of Small Molecules

279

C2H4

1r-Et

Ir-H EtC02Me

H

-i

kC0

0

0

'O---Ir-C--Et

II

II

Ir-C-Et

/

MeOH

Me

Scheme 34

to involve a hydroxycarbonyl intermediate which lost CO, to give a dihydride. Reductive elimination o f hydrogen, coordination of CO and nucleophilic attack by water then close the cycle. The complexes [IrCl(CO)L,] (L = PPh, ,P(OPh), both catalyze the carbonylation of aryl nitroso compounds to give isocyanates (equation 94)!48

61.2.5.6 Nickel

Many of the carbonylation reactions of nickel involve [Ni(CO),] as catalyst. These are outside the scope of the present work and the companion volumes 'Comprehensive Organometallic Chemistry' should be consulted. The carbonylation of dimethyl ether to methyl acetate (equation 6 8 ) is catalyzed by [Ni(H,O),]CI, in the presence of PPh, and [Cr(CO),]."4 This reaction was much faster than the subsequent carbonylation of the ester to acetic anhydride. The nature of the complexes formed under reaction conditions is unknown. The reaction requires both high temperature and CO pressure. [Ni(CO),py] catalyzes the carbonylation of 2-mercaptoethanol with ring closure (equation The product is 0,Sethylenethiocarbonate. A thiol complex, [Ni(SCH2CH20H),], was isolated from the reaction mixture and was also shown to catalyze the reaction, which occurred at a temperature and pressure only slightly above normal. [NiI(CO),]- catalyzed the carbonylation o f acetylene in the presence of methanol to give methyl Unsaturated esters of higher molecular weight were also obtained. 3-vinylacrylate (equation 101).472 Methyl ketones can be prepared by the carbonylation of aryl, benzyl and thienyl bromides and iodides in the presence of tetramethyltin (equation 102). This reaction also requires raised temperature and pressure. (Ni(CO),(PPh,),] (101) was used as catalyst.473 HOCH2CH,SH

b

+C O

0 2 H C r C H + C O + MeOH

CH,=CHCH=CHCO,Me

+

(101)

0 RI

+ CO t SnMe,

H

RCMe

+

+ SnTMe,

(102)

:Phi

I

Ni IIIIIIICO

Ph,/

'co (101)

The complexes [Ni(CO),PPh,] and [NiCl,(PPh,),] were Iess active, and [NiCl,(diphos)l gave only traces of product. The basic steps of the mechanism are given in Scheme 35.

RCNiI

Scheme 35

Uses in Synthesis and Catalysis

280 61.2.5.1

Palladium

Under oxidizing conditions, generally achieved by admission of oxygen itself, alcohols can be carbonylated to give oxalate or carbonate esters in the presence of palladium catalysts (equation 103). ROH+CO

-+

R0,CC0,R+ROC02R

(103)

PdC& in the presence of another metal chloride as co-catalyst and oxygen gave diethyl oxalate The reaction requires both high temperature and and diethyl carbonate from ethanol and C0.474 pressure. CuCI, as co-catalyst gave the best yields of the oxalate ester. The details of the mechanism are not yet clear, but it is thought that attack of alcohol on coordinated CO gives an alkoxycarbonyl complex, which coordinates a second CO molecule cis to this group. Further nucleophilic attack of the alcohol on this second CO ligand could then promote formation of the oxalate ester. Phenol has been carbonylated to diphenyl carbonate in a two-phase system in presence of oxygen and NaOH as base (equation 104). A quaternary ammonium or phosphonium salt was used as a phase-transfer catalyst, and [Mn(acac),] as co-catalyst. PdBr, was the palladium source and diphenyl carbonate was the only carbonylation 2PhOH+CO

-+

(PhO)&O

(104)

The phase-transfer method has also been used for the synthesis of aryl, benzyl, vinyl and heterocyclic carboxylic acids by carbonylation of the corresponding halides (equation 105).476 RX f CO f 2NaOH

RC0,Na t H,O+ NaX

(105)

The second mole of base is needed to effect reaction (10) and complete the catalytic cycle. [PdCl,(PPh,),] (102) was used as catalyst in the presence of excess PPh3. The basic steps of the mechanism are given in Scheme 36. Pd L

2NaOH

RX R

P

/Leo d

X

RCPdX

Scheme 36

The phase-transfer method has also been employed for the carbonylation of benzylic halides to Carboxylic a ~ i d s . 4The ~ ~ palladiurn(0) complexes [Pd(PPh,),] (103), [ Pd(diphos)J (104) and [Pd(DBA),] (105; DBA = dibenzylideneacetone)were used as catalysts. With (103) and (104) the carboxylic acid was the major product. Complex (105) gave little or none of the acid, the toluene and bibenzyl derivatives corresponding to the benzyl halide used being formed. Benzyl esters of the carboxylic acid were sometimes present as minor products. The reaction has been adapted to provide a new synthesis of anthranilic acid derivatives (equation 106).478Tri- n-butylamine was used to neutralize the HBr formed. 0

Ph3P

,

,PPhs Pd Ph3P / ‘PPh3

Catalytic Activation of Small Molecules

281

Heck and co-workers have reported the catalytic carbonylation of aryl and vinyl bromides and iodides and of benzyl chlorides in the presence of alcohols to give esters.479The general reaction is summarized in equation (107) in which RX represents the above organic halides.

+ +

RX CO R O H + R”,N

4

RCOOR’+ R”,NH+ X-

(107)

The reactions occur at normal pressure of CO but require raised temperatures. The complexes employed as catalyst precursors were [Pd(OAc),] (106), [PdI(Ph)(PPh,),] (107), [PdBr(Ph)( PPh,),] (108), [Pd2Br4(PPh3),](log), [ PdBr,( PPh,),] (110) and complex (102). Organic iodides reacted with (106) alone, otherwise the presence of PPh3 was necessary. Electron withdrawing substituents in the aryl halides accelerated the reaction and electron releasing substituents hindered it. The palladium(I1) complexes are reduced under the reaction conditions by alcohol and CO to give the palladium(0) species [Pd(CO)(PPh,),]. This undergoes oxidative addition of the organic halide to initiate the catalytic cycle. The proposed mechanism is given in Scheme 37.

RX

PPhi

Pd(CO)(PPh,),

.1.

HX

co

PdR(X)CO(PPh,)

7~ R” INH’X K”N

Pd H ( X )( PP h3) 2 \J*

IPPh3

€‘d(COR)X(PPh3),

RCOOR R O H Scheme 37

The product of the oxidative addition may be [PdX(R)(PPh,),] and not the carbonyl complex shown. Inorganic bases such as NaOAc can also be used in these reactions.480Complex (102) was used here as catalyst precursor. In a further study of the reaction the complexes [Pd(CO)(PPh3)J (ill), [Pd3(C0)3(PPh3)4](112), [Pd,(CO),(PPh,),] (113) and complex (102) were used as catalysts with diethylamine as base?8’ The presence of an acyl intermediate was confirmed. Some of these reactions could be carried out at normal temperature and pressure. n

/

Ph3P

,

0

\

PPh3

FPh3

With complex (102) as catalyst, reaction (107) has been used to prepare a precursor of the natural product curvularin.482If cyclic ethers are used instead of alcohols, o-haloesters are

Uses in Synthesis and Catalysis

282

Epoxides can be used in a similar manner (equation 109). Complex obtained (equation (107) was used as catalyst. No base is needed in this variant of the reaction.

If the halide and the hydroxy group are present in the same molecule, reaction (107) leads to the synthesis of lactones.4s4With complex (102) as catalyst a series of butenolides were prepared in good yields from vinyl iodides (equation 110). Four- and six-membered ring lactones and The mechanism proposed was analogous to that of a-methylene lactones were Scheme 37. This cyclization has been used in the synthesis of the natural product eara ale none.^^^ PdCl, was the catalyst.

Aryl, vinyl and heterocyclic bromides and iodides in the presence of H2, CO and a base react to give aldehydes (equation 111). [PdI,(PPh,),] (114) or its analogues (102) and (110) were used as catalysts. The reaction requires both high temperature and pressure.48xIn some cases, aroyl chlorides are also converted to aldehydes under these conditions. RX+H,+CO+R,N

-

RCHO+R3NHCX-

(Ill)

Reaction of the catalyst precursor [PdX2(PPh,),] ( X =halide) with HZ,CO and two moles of base is believed to give [Pd(CO)(PPh,),] (equation 112). This then undergoes oxidative addition and dissociation of a PPh3 ligand. The proposed mechanism is given in Scheme 38. Pf‘h

PdX,(PPh,),+ H,+CO+2R3N 2 Pd(CO)(PPh,), +2R3NH+ XRX

(112)

PPh

Pd(CO)(PPh,),

Pd(R)X(CO)(PPh,)

R’,NH’X-

R’;N, CO

1

PdH(X)(PPh,), \J* RCHO

Pd(COR)X(PPhp), H,

Scheme 38

As with other palladium-catalyzed reactions of organic halides, if the group R is such as to permit a B-hydride elimination, reaction (7) may occur preferentially. Unsymmetrical ketones can be obtained by the palladium-catalyzed carbonylation of aryl, alkyl and vinyl bromides and iodides in the presence of tetraalkyl- or tetraaryl-tin compounds (equation 1 13).4893490 The catalyst precursors used were complexes (102), (107) and [ PdCl,(AsPh,),]. 0

RX+CO+R’,Sn

!I

+

RCR+R’,SnX

(113)

When alkyl halides were used, some elimination occurred to give the corresponding alkene.490 [Pd(CU)(PPh,),] was again believed to be formed in situ when PPh, complexes were used as catalysts. The mechanism of ketone formation is shown in Scheme 39. For those cases where R is an alkyl group, the extended reaction sequence shown in Scheme 40 was invoked to explain the production of the alkene and the occurrence of i ~ o m e r i z a t i o n ~ ~ ~ AsPh, as Ligand was found to favour the carbonylation reaction relative to PPh3.

Cntalytic Actiuation of Small Molecules RX

CO

Pd(CO)(PPh,),

co

A

283

Pd(R)X(CO)(PPh3)2

1

I

Pd(R)(COR)(PPh,),

'J\ Pd(COR)X(PPh3)2 R',SnX R'&n Scheme 39

qR+ X Pd

I

R'4sy;

Pd-X

+

R

R

-1 R

$P,d-~X H

I Pd-X

Pd

R ,SnX R

'5 +

% *

R'H

+

Pd

R',SnX

-

isomers

R

Scheme 40

At raised temperature and CO pressure, terminal acetylenes can replace the organotin derivatives, giving acetylenic ketones (equation 114).'"' Some of these reactions occur at normal pressure. Aryl, heterocyclic and vinyl bromides and iodides could be used. These reactions again required a stoichiometric quantity of base. The complexes (102), [PdCl,L,] ( L = 1,l'-bis(dipheny1phosphino)ferrocene) and [ PdI(Ph)( AsPh,),] were used as catalysts. The mechanism was not discussed. 0

RX + CO + H C E C R + R",N

II

+

RCC=CR+ R'I3NH+X-

(114)

The reaction of an aryl and an alkyl iodide in presence of complex (103) and the Zn-Cu couple leads to alkyl aryl ketones in good yields (equation 115).492By-products such as ArR, ArH and Ar, were found in some cases. Replacement of aikyl iodides by benzyl chlorides gave benzyl ketones, but the formation of by-products due to coupling reactions was significant. Dialkylzinc complexes were formed here and the proposed mechanism is given in Scheme 41. 0 I1

ArX+CO+RX

ArCR

ArI

Pd

ArPdl

..IRj

;znfR2-

ArCPdR

+ 0

II

ArPdCR Scheme 41

CCC6-J'

(115)

Uses in Synikesis and Catalysis

284

The introduction of a primary or secondary amine into reaction (107) in place of the alcohol leads to the production of amides (equation l16).493The complexes used as catalyst precursors were (102), (110), (114) and (108). Complex (110) was most often used. RX+ CO+ R"H,+R",N

+

RCONHR'+ R';NHf X-

(116)

The reaction can be carried out with aryl, heterocyclic and vinyl bromides and iodides and with some vinyl chlorides. It requires raised temperature but only normal CO pressure. If the primary or secondary amine is a strong enough base, it can be used in excess and the tertiary amine is not needed, The mechanism given in Scheme 42 was proposed for the reaction. It was not known whether the primary amine enters the reaction sequence at the point shown or at an earlier stage.

-.r.

CO

RX

Pd(CO)(PPh,),

r

PdR(X)(PPh,),

7R"3NH'X-

HX

R;N

co

co

PdH(X)(PPh,), ' J \

Pd(COR)X(PPh3)2

RCONHR' R ' N H ? Scheme 42

The initial reduction of the palladium(I1) precursors to [Pd(CO)(PPh,),] occurs according to equation (1 17). Pd[CO)(PPh3),+(R'NH),CO+2HX

PdX,(PPh3),+2CO+R'NH,

(117)

Using [PdI(Ph)(PPh,),] (107) as catalyst, it has been shown that at higher temperature and CO pressure amides may be formed directly from tertiary amines according to equation (1 18).494 This reaction appears to be limited to R'= Et if reasonable yields are to be attained. FS+CO+R',N

+

RCONR',+R'X

(118)

In a further variant of this type of carbonylation, aryl, vinyl and heterocyclic bromides and iodides underwent a double carbonylation to give a-ketoamides (equation 1 19). RX -t2CO f 2R',NH

4

RCOCONR', + R',NH,+ X-

(119)

This reaction appears to be limited to secondary amines, which were used in excess.4953496 Simple amides were usually obtained as by-products in these reactions, though the selectivity to the a-ketoamides was generally good and often better than 90%. It is at present not completely RX, CO

0

\

R'NH~

R'NH2 RCOCONHR'

I H X

1

0

HPdX

II

RC-PdCNHR

1-

Pd (*)

0

II

Pd Scheme 43

(R)

RCOCONHR'

Catalytic Activation of Small Molecules

285

clear which reaction parameters affect the course of the reaction as regards the occurrence of single or double carbonylation, but it is clear that alkyl- or dialkyl-phosphines favour the latter course. Many complexes have been used as catalyst precursors for the reaction. The best complexes for double carbonylation were PdClz , [PdCl,(NCPh),], [PdC12(PMePh2),], [PdCl2(PMezPh),I, [PdBr(Ph)(PMePh,),], [PdC12(PEtPh2)21, CPdC12(PEtzPh)zl, [PdClZ(DIOP)], [ PdClz{PhzP(CH2),PPh,}] and [PdI(Ph)(PPh,),] (107).The precise mechanism is uncertain. An acyl complex is believed to be formed by the oxidative addition of the organic halide to palladium(0) followed by insertion of CO. There are then two possibilities for the production of a-ketoamides and they are given in Scheme 43. The above synthesis of simple amides has been adapted to intramolecular reactions leading to lac tarn^.^^^ Five-, six- and seven-membered ring benzolactams were prepared (equation 120).

[Pd(OAc),] in the presence of PPh, was used as catalyst. The use of vinyl bromides in lactam formation has also been rep0rted.4~'Imides were obtained from aryl bromides. The method has further been applied to the synthesis of diazepam and 1,4-ben~odiazepines~~~ and to a-methylene lactams and In connection with the synthesis of natural products, the reaction has been employed in the preparation of hexadehydr~himbane,~~' anthram~cin~ and ' ~ berbine derivative~.~''~ The catalyst was prepared in situ from [Pd(OAc),] (106) and PPh3 in all cases. The mechanism of lactam formation is analogous to that for amides (Scheme 42). A brief review on the palladium-catalyzed carbonylation of organic halides has ap~eared.5'~ PdCI, in presence of PPh, catalyzes the hydrocarboxyiation of alkenes at raised temperature and CO pressure (equation 121).506Both the normal and branched isomers of the acid are formed. The normal isomer is generally the major product if R is an alkyl group. Various other phosphorus ligands were tried and tri-o-tolylphosphine gave the highest ratio of linear to branched acids. Use of an alcohol in place of water in equation (121) leads to the formation of esters (hydroesterification). Thus cyclohexene and methanol gave methyl cyclohexanecarboxylate in high yield using PdC12 and PPh, as catalyst precursor. This reaction does not have the problem of isomer formation (equation 1 2 2 1 . ~ ~ ' RCH=CH,+CO+ H,O

-.

RCH,CH,CO,H+RCH(Me)CO,H

fCO+MeOH

-G

C02Me

(121)

(1221

In a study of the hydroesterification of 1-heptene the complexes [PdCl,L,] ( L = PPh3, AsPh,, P(OPh), , PMe,Ph, P( 1-MeOC,H,), , P(4-C1C6H4)3,P(4-MeOC,H4), or P( o-tol),) in the presence of Lewis acids such as SnC12, SnIz, GeClZand PbCl, were used as cataly~t.~"The combination [PdCl,{P( p - t ~ l ) ~plus } ~ ] IOSnCl, gave the best yield and seiectivity to the linear ester. Terminal alkenes generally reacted well, as did cyclooctene, but internal linear alkenes gave poor results. Primary alcohols performed far better than secondary ones. The catalyst precursor in the case of PPh, and SnCl, above is believed to give rise to a hydrido complex of the type [PdH(SnCI,)(PPh,)L] ( L = SnCl,-, PPh3, CO or Cl). The expected reaction mechanism is given in Scheme 44. RCH=CH2 PdH(SnCI3)(PPh3)L

I PdH(SnC13)(PPh3)L I RCH=CH2

RCH2CH2COOR' R'OH Pd(COCH&H,R)(SnCI,)(PPh,)L

*----i-Pd(CH2CH,R)(SnC13)(PPh,)l, co

(L= SnCI,-, C1-, PPh3 or CO) Scheme 44

286

Uses in Synthesis and Catalyst$

Isomer formation can occur at the alkene insertion step: This is probably reversible and the rates of the CO insertion for the alkyl complexes formed will therefore also affect the ultimate product distribution. Higher temperatures favour the formation of the branched Often, the alcohol is used as solvent for these reactions. The addition of another solvent of lesser polarity has been shown to reverse the usual selectivity, giving the branched product as the major isomer. Again, complex (102) was used as ~atalyst.”~ In the hydroesterification of styrene (equation 123) the isomer distribution was found to be very dependent on the phosphorus ligand used. PhCH=CH2+ CO+ROH

-+

PhCH,CH,CO,R+ PhCHMe

I

COzR

Complex (102) gave only the branched isomer, whereas [PdCl,(DIOP)] favoured the linear product. The same was found with a-methylstyrene, this time with almost complete regioselectivity in both cases.”’ The effect of chelate ring size on the selectivitywas studied using the diphosphines ~ ~ six-, ~ ~ ~seven’~ PhzP(CHz).PPhz ( n = 1-6,lO). Complex (102) was included for c o m p a r i s ~ n .The and eight-membered chelate rings ( n = 3,4,5) favoured the linear isomer slightly, the sevenmembered ring giving the best result. For n = 1,6 or 10 the branched isomer predominated, to a large extent in the latter two cases. This was also the case with the monodentate ligands PPh,, PPhzBu, PPh,(CH,Ph), PCy, and PBu, , which gave almost complete specificity to the branched product. The effect was attributed to the stability of the chelate rings. Bisphosphine complexes favour the formation of the linear isomer on steric grounds, whereas in monophosphine species the branched alkyl complex is preferred. The larger, less stable chelate rings therefore show an increasing tendency to give rise to monophosphine complexes through chelate ring opening, favouring the branched Interestingly, no catalytic reaction occurred when diphos was used as ligand. This would form the most stable, five-membered ring. The hydroesterification of allyl acetate461and N-~inylphthalimide~’~ has been studied using complex (102) in presence of excess PPh3 or SnC1,. Significant dependence of the regioselectivity of the solvent was again observed. 3,3,3-Trifluoropropene and pentafluorostyrene were subjected to hydroesterification and hydrocarbonylation using complex (102) or the complexes [ PdClZL] (L = 1,l‘-bis(diphenyIphosphino)ferrocene, 1,4-bis(dipheny1phosphino)butane)as catalysts.515By suitable choice of ligand and reaction conditions, either isomer of the respective products could be obtained in good yield. The selectivity changed markedly on changing the nucleophile from water to alcohol. This may be due to a change in the mechanism of the reaction to the alkoxycarbony1 route shown in Scheme 45, instead of the acyl route (Scheme 44).

Pd-CO

H2c fH+

0

,

RCH2CH2COzH

RCH=CHz

Pd- CHCH2C02H

R

Scheme 45

When the cluster [Pd4(C0)4(02CMe)4].2MeC02H was dissolved in styrene and treated with methanol under an atmosphere of carbon monoxide, methyl cinnamate was formed.516This reaction was also believed to occur by an alkoxycarbonyl route. The reaction became catalytic when [Pd(OAc),] (106)was used in presence of NaOAc and a stoichiometric amount of copper(I1) as reoxidant for the palladium(0) formed. Stille and co-workers have investigated this reaction, sometimes called carboalkoxylation, in detail. A basic difference between this reaction and the hydroesterification described above is that the oxidative nature of carboalkoxylation permits double functionalization of a double bond. Thus ( E ) - and (Z)-2-butene react readily with CO and methanol in the presence of a catalytic amount of PdClz and a stoichiometric amount of CuCl, to give methyl 3-methoxy-2-methylbutanoate(equation 124).517,518 MeCH=CHMe+COS MeOW

4

MeCM(OMe)CH(CO,Me)Me

(124)

Catalytic Activation of Small Molecules

287

The addition to the double bond was, at least in the early stages of the reaction, exclusively trans. In presence of NaOAc this changed completely to give cis addition and the product under these conditions was dimethyl 2,3-dimethylsuccinate (equation 125). MeOH

MeCH=CHMe+CO

MeCH(CO,Me)CH(CO,Me)Me

(125)

The former reaction involves attack of methanol on the double bond as the first step. In presence of NaOAc however, carbomethoxylation occurs (Scheme 46). CO, MeOH

A

]-Pd

>-Pd-COzMe

I

HE H llllli iiiiiiH ""OwMe Me Pd

MeOzC

Pd

CO, MeOH

C O , MeOH

4

J

Me0 Hll$+-(M; Me

\CO2Me

MeOzC

COZMe

Scheme 46

A number of linear and cyclic alkenes have been studied in this reaction. The presence of a base favoured the formation of diesters from linear alkenes, whereas cyclic alkenes gave diesters In a further study using dienes and unsaturated alcohols, ketones and even in absence of esters, the dicarboxylation reaction could quite generally be achieved in good yields.5207521 In presence of CCI4 and a base, alkenes react with CO and alcohols with [Pd(OAc),] (106) and PPh3 as catalyst precursor to give trichloromethyl derivatives (equation 126).'22 The ester was generally the major product. This reaction requires higher pressure than the above carboalkoxylations. Diphos and P(o-tol), were as effective as PPh3 as ligands but P(OPh), and PBu3 gave inferior results. The mechanism of the reaction is at present unknown. RCH=CH2

+ CCI4 + CO

EtOH

R

R (126)

Using ( E )-alkenylpentafluorosilicates, unsaturated esters can be readily obtained by reaction with CO and methano1 (equation 127)."' PdClz or PdBrz was generally effective as catalyst. Complex (106) was less satisfactory and no carbonylation occurred with [PdCl,(PPh,),] (102). R K2[H;C=C

L .

(65)

A related epoxidation reaction was found to occur catalytically when an alkene was reacted with azobenzil in the presence of 0, and catalytic amounts of palladium acetate (equation 47).194 PhC-CPh

II II

N, 0

+

>=(

+O,

& PhC-CPh Pd(0Ac)

II II

0

+

(47)

0

Either a palladadioxetane species (67) formed upon addition of 0, to the ketocarbene metal has been suggested to be the active intermediate in this complex (66),or a dioxirane form (a), reaction (equation 48).

Metal Complexes in Oxidation Ph

\

0-0

I

/Pd=C-CPh

I1

-.. 1 -.L+ ,Pd-C-CPh 0

I

\ --+

I/

/

Ph 0

0

N2

II

Pd +

1

o,

Ph-.. lC ,, Ph!

331

H -

PhC-CPh

/I

0

II

+

(48)

0

0

II

PhC-CPh

-N*

It is noteworthy that Pt02(PPh3)2can oxidize benzyl to benzoic anhydride with 55% yield by inserting one oxygen atom between the two carbonyl The palladium peroxometallic adducts were found to be less stable than the platinum analogs and decompose mainly by C-C bond cleavage to carbonyl compounds (equation 49).'43 A similar decomposition may occur in the oxidation of ketoximes to ketones by Pd02(Ph,P), (equation 50f.'Y5 0 (Ph3P)*Pd' ' 0 &&-Me 'Me NC// CN

heat

Me\ /C=O Me

+

(49)

NC

CN

The addition of trifluoroacetic acid to the palladium or platinum peroxide ahducts with electrophilic alkenes results in the formation of epoxide in high yield and with high stereoselectivity.143~148 The mechanism shown in equation (51) has been suggested for this reaction.'48

(51)

y e OH.40

CF,CO,H ___,

OHC v

\

,,OCOCF, PZR, OH

M 0

e

,OCOCF, P,Pt\ OCOCFj

f

+H 2 0

A large number of Group VI11 metal-dioxygen complexes catalyze the oxidation of phosphine However, to phosphine oxide or isocyanides to isocyanates by molecular oxygen.6~'2~56~'40~'41~146,'84 their use as catalysts for the oxidation of alkenes enerally leads to the same products as those obtained from free radical chain a u t o x i d a t i ~ n s .B'96 ~-~198~ With the notable exception of rhodium, Group VI11 metal-peroxo complexes are generally reluctant to react with simple alkenes by nonradical pathways. However, such an oxygen transfer has been shown to occur in the reaction of '80-labeled [(A~Ph,)~Rh0,]'C10,- with terminal alkenes under 0,-free, anhydrous conditions, producing '80-labeled methyl ketone (equation 52).13' ClO, (69)

+ RCH=CH2

R-C-Me

II

I8O 85%

This oxygen-transfer reaction was based on an important observation by Read et d., who found that rhodium complexes such as RhCI(PPhJ3 were able to promote the cooxygenation of terminal

338

Uses in Synthesis and Catalysis

alkenes to methyl ketones and PPh3 to Ph3P0 using conditions under which a Wacker-type oxidation (hydroxymetalation of the coordinated alkene by water) was disfavored (equation 53).199 RCH=CH2+PPh3+O2

RhCI!PPhjl, b

C,H,

R-C-Me+Ph,PO I1

(53)

Q

In this reaction, one oxygen atom of the dioxygen molecule is incorporated into the alkene, while the phosphine acts as a coreducing agent with the second oxygen atom to produce phosphine oxide. Tn the absence of phosphines and under more forced conditions, the use of Rh' complexes as catalysts for the autoxidation of alkenes results in the formation of the expected ketone together with aliylic or cleavage oxygenated products, presumably coming from parallel radicalchain reactions. 198.200-203 The use of rhodium trichloride associated with copper(I1) perchlorate or nitrate in an alcoholic solvent resulted in a major improvement in the catalytic oxidation of terminal alkenes by 0, at room temperature, without the need for a coreducing agent (equation 54).'04 2RCHzCHZ + 0,

MCl,

+ CU(CIO,), b

2R-C-Me

II

EtOH, ?O-80°C

(54)

0 298%

In this reaction, both oxygen atoms are incorporated into two moles of alkene to give two moles of ketone with a selectivity, based on consumed alkene and 0 2 ,of up to 98%. The characteristics of the reaction are different from conventional Wacker chemistry. (1) In contrast to the corresponding palladium system, water is not involved as the oxygen source. In fact it is an inhibitor, and the presence of a dehydrating agent such as 2,2-dimethoxypropane speeds up the reaction. (2) Although the alcohol solvent does not act as a coreducing agent, it is necessary for the reaction and probably intervenes in the formation of the active specie^.^^^.^'^ (3) The presence of copper(I1) salts, in the optimum ratio Cu: Rh = 1 or 2, considerably increases the rate and the selectivity of the reaction. Rhodium perchlorate alone acts as a catalyst, but cooxidizes the solvent alcohol to a carbonyl compound. With Cu :Rh = 1, 85% of the copper precipitates from the reaction mixture as CuCl. (4) The presence of chloride anions in an optimum ratio C1: Rh = 2 or 3 and an acidic medium (H+: Rh = 3) are necessary for the ( 5 ) The nature of the oxidation products is traceable to the nature of the rhodium-alkene interaction. Terminal alkenes and internal ones (e.g. cycloheptene), which form ii-complexes of rhodium( I), e.g. [RhC1(alkene)2]2, are selectively converted into methyl ketones, whereas alkenes which form rr-allylic complexes of rhodium( 111) ( e.g. cyclopentene) give alkenyl ethers via oxidative substitution of the alkene by the solvent a I c o h 0 1 . ~ ~ ~ OR

(6) The RhC13-Cu(C104)2 catalyzed oxidation of 1-hexene is first-order dependent on rhodium and alkene concentration, and independent of dioxygen pressure. Jn fact, higher 0, partial pressure inhibits the reaction. Strongly complexing alkenes, e.g. cyclooctene, were found to be less reactive than terminal ones. Further, the reaction is strongly inhibited by ligands, e.g. PPh,, or strongly compleiing dialkenes, e.g. 1,5-cyclooctadiene. The overall catalytic oxidation of terminal alkenes to methyl ketones by 0, has been tentatively interpreted as resulting from the consecutive consumption of each oxygen atom by two moles of alkene in two complementary reactions as shown in Scheme 3. The first peroxide oxygen atom is transferred to the alkene by a peroxymetalation pathway involving coordination of both 0, and alkene to the metal in (71a) or (71b), followed by the formation of a peroxometallacycle (72a) or (72b) which decomposes to give the methyl ketone and the rhodium-oxo (73a) or -hydroxo (73b) species. The second hydroxo oxygen atom is transferred via an internal cis hydroxymetalation mechanism (Wacker type), which produces a second mole of ketone and regenerates the initiai Rh"' hydride or the MI'species by reductive elimination of the latter. Two plausible alternative versions of the peroxymetalation pathway can be envisaged133and these are described below.

Metal Complexes in Oxidation

I

Hydroxy metalation

I I

Peroxy metalation

kh

trisubstituted > disubstituted > monosubstituted. This is illustrated by the regiospecific monoepoxidation of conjugated dienes (equation 66)

Electron withdrawing groups on the alkene considerably retard the epoxidation. For example, acrylic esters and acrylonitrile are not reactive, while allyl chloride is about one tenth as reactive ~ epoxidation . ~ ~ ~ of alkenes is completely stereoselective: as propylene towards TBHP/ M o ~ The cis-alkenes are exclusively transformed into cis-epoxides and trans-alkenes into trans-epoxides. Oxygen addition to the double bond preferentially occurs from the less shielded face of the substrate, e.g. in the selective epoxidation of terpenes by t-pentyl hydroperoxide (t-amyl hydroperoxide, TAHP) (equations 67 and 68).225

The presence of the substrate of functional groups capable of interacting with the metal directs the stereoselectivity of the epoxidation, as shown by the comparative reactivity of (82) towards TAHP/Mo(CO)~and peroxybenzoic acid (PBA).241

(82)

TAHP/Mo(CO), PBA

100% 50%

0% 50%

Allylic and homoallylic alcohols are particularly good substrates for epoxidation by TBHP/VV and TBHP/ Mow catalysts, with the former being superior in activity and selectivity (equations 70-72). 57,226,242 Allylic alcohols have also been shown to be particularly good substrates for enantioselective epoxidation. Good results were observed in some cases with TBHP/VO(acac),/chiral hydroxamates (equation 73);' but a major breakthrough was obtained

Uses in Synthesis and Catalysis

344

using TBHP/Ti(OR), in the presence of dialkyl tartrate as the chiral ligand (equation 74). This method has proved to be very useful for the asymmetric epoxidation of a wide range of prochiral allylic alcohols with good yields and enantiometric excesses exceeding 90%.2437244

TBHP/Mo(CO), TBH P/ VO( acac)

TBHP/ Mo( CO), TB HP/VO( acac),

2% 2%

98% 98%

98 % 98%

2% 2%

I X P.' 2m01 TBHP PhMe -20 "C 4 d

(73)

Ph 90% (80% e.e.1

Ph

D( -)-diethyl

tartrate

\

R.2 R'

R1

TBHP!Ti(OiR),

OH

CH2CI,,-20'C

or

R2 L(+)-diethyl tartrate

R3 0

70-87% yield 290% e.e.

( c ) Eflect of solvent, ligands and coproduct nlcohol. The d" metal-catalyzed epoxidation of alkenes requires anhydrous conditions, with the inhibiting effect of water being more pronounced for V and Ti than for Mo and W. Chlorinated (e.g. CH&, C2H4C12)or aromatic solvents (e.g. benzene, toluene) are suitable for good catalyst activity, but alcohols or basic solvents, e.g. DMF, THF and dioxane, strongly retard or completely inhibit the oxidation. The coproduct alcohol derived from the hydroperoxide also exerts an inhibitory effect in the order W < Mo < Ti < V245 and competes with the alkyl hydroperoxide by forming metal alkoxides, preventing the formation of metal-alkyl peroxides. ( d ) Mechanism. It is now generally admitted that this reaction, owing to its very high selectivity and stereospecificity, proceeds according to a heterolytic rather than a homolytic mechanism. Two alternative mechanisms (A and €3; Scheme 5 ) emerge from the numerous interpretations proposed.Sb (A) Nucleophilic attack of the alkene on the 'electrophilic' oxygen atom covalently bound to the metal, which is reminiscent of Bartlett's butterfly mechanism for epoxidation of alkenes by percarboxylic acids.229

Metal Complexes in Oxidation

345

\

ROH

ROOH Scheme 5

(B) Complexation of the alkene to the metal followed by its insertion between the metaloxygen bond according to an intramolecular 1,3-dipolar mechanism, forming a fivemembered pseudocyclic peroxometallacycle which decomposes to give the epoxide and the metal alkoxide. Although there is no definitive proof establishing one or the other proposal, mechanism (B) appears to explain most of the characteristics of the reaction. This mechanism is consistent with the general T - u rearrangement procedure occurring in most heterolytic metal-catalyzed transformations of alkenes,I5*and is close to that proposed for the epoxidation of alkenes by Mo-peroxo complexes (equation 261, which has similar characteristics. As it involves a Lewis base-Lewis acid metal-alkene bond, it is consistent with the reactivity of alkenes increasing with their nucleophilic nature, and with the strong inhibition by basic ligands and solvents which compete with the alkene for vacant sites on the metal. Further, the absence of heterolytic reactivity of the dipicolinatovanadium(V)-alkyl peroxide complex (22), which has no available coordination sites, disfavors noncomplexation mechanism (A). Also, mechanism B, involving rigid metallacyclic intermediates, is better suited to explain the very high stereoselectivity of the reaction, e.g. in the case of allylic alcohols, as shown by equation (75).121,'62 1213162,193m4

If we consider the d o metal-N,N-dialkylhydroxylamino complexes (79), (80) and (81) as valid models for the reactive but unstable alkyl peroxide species Mo02(00R),, VO(OOR)3 or V203(OOR)4,and Ti(OOR), presumably involved in catalytic oxidations, the low activity of vanadium and titanium for the epoxidation of simple alkenes could be interpreted by the fact that these alkenes cannot displace the 0,O-bonded alkyl peroxide groups in the coordinatively saturated Vv- and TiIv-alkyl peroxide species, whereas allylic alcohols can displace the alkyl peroxide groups by forming bidentate allylic alkoxides as in equation (75).16' Since the metal-alkene association preceding the peroxymetalation reaction in mechanism ( 3 ) is a pure Lewis acid/Lewis base interaction, it would be expected that compounds having alkylperoxy groups bonded to a Lewis acid center should be active for the epoxidation o f alkenes. This is indeed found for boron cornpounds, which are active as catalysts for the epoxidation of alkenes by alkyl hydroperoxide^.'^^**^^ Isolated boron tris(alky1 peroxides), B(OOR)3, have been shown to epoxidize alkenes stoichiometrically, presumably through alkylperoxyboration of the double bond (equation 76).248

Uses in Synthesis and Catalysis

346

Aluminum compounds such as Al(OBu'), have also been shown to catalyze the selective transformation of allylic alcohols into s y n - e p o ~ i d e sRelated . ~ ~ to the mechanism of this reaction is the decomposition of the aluminum compound (83) to give ethylene oxide?" P'CH, Et,AI)\ I O-CH, I OBu'

--+

H?C--CH, \ / 0

+ Et,AIOOBu'

(77)

(83)

Epoxides can also be formed from the oxidation of alkenes by molecular oxygen via in situ generation of hydroperoxides by An interesting example is the direct stereoselective oxidation of cyclohexene by 0, to syn- 1,2-epoxycyclohexan-3-01catalyzed by CpV(CO), with a 65% yield and -99% stereoselectivity (equation 78).253 OH 1

(iii)

Oxidation of other substrates

Alkyl hydroperoxides can oxidize a variety of other nucleophilic substrates in the presence of d o metal catalysts. Thus molybdenum and vanadium catalysts have been used for the selective oxidation of tertiary amines to the corresponding N-oxides (equations 79 and 80).225,254 R,N+Bu'OOH

VO(acaci,

R3NO+BuLOH

(79)

Primary amines such as cyclohexylamine are selectively converted to oximes by TBHP in the ~ , found presence of Ti, V, Mo and W compounds. Titanium compounds, such as T ~ ( O B U )were to be the best catalysts (equation 81).255

The oxidation of sulfides to sulfoxides by TBHP in the presence of Mo and V catalysts has been extensively s t ~ d i e d . ~A~modified ' , ~ ~ ~ Sharpless reagent,243i.e. Ti(0Pri),/2 diethyl tartrate/ 1 H20, was used for the asymmetric oxidation of prochiral sulfides to sulfoxides with enantiomeric excess greater than 90% (equation 82). 160,257 Me

S/ Ti(OPr'j,, 2 diethyl tanrate+ IH,O CH,CI,, -20 "C:. 4 h

Me

Me

e.e.-91% yield = 90%

Metal Complexes in Oxidation

347

61.3.2.2.2 Group VI11 metal alkylperoxo and hydroperoxo complexes

(i)

Palladium catalysts

Palladium ?-butyl peroxide carboxylates [Pd( OOBu') ( RCO2)I4,easily prepared by reaction (9), and characterized by X-ray crystallography (formula 23),are very effectivereagents for the selective oxidation of terminal alkenes to methyl ketones under anhydrous and anaerobic conditions?2 Their reactivity decreases in the order R = CF, > CC13>> CH3. The reaction has been shown to proceed in two steps: (a) oxygen transfer from the ?-butyl peroxide complex to the alkene, producing the methyl ketone and the palladium- r-butoxy complex (equation 83), followed by (b) a rapid substitution of the OBu' group by the alkene on the metal, resulting in the formation of t-butyl alcohol and the rr-allylic complex (equation 84). Only terminal alkenes are selectively converted to methyl ketones. Internal alkenes directly produce the .rr-allylic complex upon exchange with the OOBd groups. 0

+ 7CFSCO,Pd-O-Bu'

CF3CO2Pd-00But

CF3C02Pd-OBu'

+

(83)

+ /",/

(84) \

kz x k,

Et

The genera1 trends of this oxidation are consistent with the mechanism depicted in Scheme 6. This involves the complexation of the alkene to the metal followed by its insertion into the palladium-oxygen bond, forming the five-membered pseudocyclic intermediate which decomposes to give the methyl ketone and the palladium-t-butoxy complex. The decomposition of (Ma) is similar to that of the rhodium dioxametallacycles previously shown in Scheme 3.42 (CF3COJ2Pd

i;

BU'OOH

CF3C02H

)

Bu'OOH

RCH=CH2

CF3C02Pd-:

,

/

RCH=CH,

I

CF3CO,Pd-0But

CF3CO;Pd'OOBut

JH

CH2 CF3C02Pdq/ \

0

P-d

Scheme 6

/

C- R

But

(84s)

It can be seen that the pseudocyclic intermediate (Ma) strongly resembles the stable alkylperoxymercury compound (Ma) prepared from the reaction of TBHP with an alkene in the presence of The X-ray structure of the similar BrHg{CH(Ph)CH( Ph)(OOBuf)} mercury(I1) compound has clearly shown the pseudocyclic nature of this adduct by the interaction existing between mercury and the OBu' group.259The transmetalation of mercury by palladium in (84b) produces acetophenone in 9 5 % yield, presumably via the formation of the pseudocyclic intermediate (85; equation 85).42 H

CCC6-L*

Uses in Synthesis and Catalysis

348

Methyl ketones can be catalytically produced when an excess of TBHP is used for regenerating the initial t-butyl peroxide species from the resulting alkoxy complex in Scheme 6. To prevent the formation of a m-allylic complex from causing lower selectivities, a large excess of TBHP with respect to the alkene is required (equation 86).260 WIRCO,),

+ Bu'OOH

R

+ Bu'OH

R-C-Me

!I

0

70-8 5 %

Among palladium complexes, [Pd(CF,CO,)(OOBuf)], was found to be the best catalyst for the oxidative cleavage of cyclic acetals to monoesters of diols by TBHP (equation 87).261Palladium carboxylates are also effective catalysts for the selective oxidation of terminal alkenes to methyl ketones by hydrogen peroxide (equation 88)?62 This reaction occurs at room to moderate (80 "C) temperature, in a two-phase (benzene or AcOEt) or single-phase (Bu'OH or AcOH) solution. Only terminal alkenes, e.g. 1-hexene, 1-octene, allyl acetate, are selectively converted to methyl ketone. l 8 0 Labeling studies showed that ketonic oxygen is derived from hydrogen peroxide, and not water as in the Wacker process. The suggested mechanism is similar to that of the ketonization of alkenes by PdOOBu' complexes, and involves palladium hydroperoxide as the reactive intermediate (equation 89).262

RCH=CHz+H,Oz

Pd(RCO,l,

83%

R-C-Me+H,O

(88)

II

0

70-90% 0

R R

RCO2Pd-00H

t

=/

d I

RC0,Pd-OOH

-

RC02Pdq "QR

0 ' 0 I

-

!I

R-C-Me

+

RC02Pd-OH

(89)

I

Evidence for the pseudocyclic intermediate (86) was obtained by the formation of acetophenone in 85% yield upon reaction of Li2PdC1, with CF3C02HgCH2CH(Ph)OOH,as in reaction (85).263 Treatment of nonreactive Pd02(PPh3)2with one equivalent of MeS03H results in the formation of a coordinatively unsaturated hydroperoxide complex which reacts with terminal alkenes to give the corresponding methyl ketone. Labeling studies showed that the ketonic oxygen atom is derived from molecular oxygen, not water.263Using CF3C02Hinstead of MeS0,H results in the formation of an inactive hydroperoxo complex due to the occupation of the vacant site S in (87) by the trifluoroaceto group, impeding coordination of the alkene (equation 90).263alp-

349

Metal Complexes in Oxidation

Unsaturated esters or a#-unsaturated ketones are efficiently oxidized to P-keto esters and p-diketones by TBHP or H,O, in the presence of Na2PdCl, (equation 91). The oxidation of terminal alkenes under these conditions results in extensive double bond migration and lower

60-80%

(ii)

Platinum catalysts

Platinum-alkylperoxo and -hydroperox0 complexes are much less effective ketonization reagents than their palladium analogs. The platinum-hydroperoxide complex generated by protonation of Pt( PPh3)202as in equation (90) was found to be inactive,'33as well as Pt(CF,)(OOH)(depe) obtained from the reaction of W 2 0 2 with the corresponding hydroxo However, the platinum-t-butyl peroxide complex [Pt(CF,CO,),(OOBu')( Bu'OH)], selectively transforms terminal alkenes (e.g. 1-octene) into the corresponding methyl ketone?' but much less efficiently than the palladium complex (23)." This reaction becomes slightly catalytic in the presence of excess TBHP. Strukul and coworkers recently prepared a series of cis and trans isomers of the platinum- t-butyl peroxide complexes Pt(R)(OOBu')L, (R=CF3,Ph, etc; L = tertiary phosphine). The X-ray crystal structure of the trans-[ R(Ph)(OOBu')(PPh,),] (88) revealed a square-planar arrangement with end-bonded OOBut group.219Interestingly, only the trans isomers were found to be capable of oxidizing terminal alkenes to methyl ketones, whereas the cis isomers were inactive. The suggested mechanism involves the coordination of the alkene leading to a five-coordinate intermediate, followed by a pseudocyclic peroxymetalation, as in Scheme 6 (equation 92).'lY

R

4

I /L

R-Pi-OORu'

L/

R-Pt

-

?qR

/ '0-0

L I

/

L

R-Pt--OBui /

+ R-C---Me II

L

0

Bu'

(iii) Rhodium and iridium catalysts

Although no direct oxygen transfer reactions from well-characterized rhodium-hydroperoxo or -alkylperoxo complexes to alkenes have yet been reported, these species are probably involved in the rhodium-copper catalyzed ketonization of terminal alkenes by 02,as previously shown in Section 61.3.2.1.3. Rhodium trichloride has been used to catalyze the ketonization of terminal alkenes by H,O, or TBHP in alcoholic solvents, but these reactions occur less efficiently than with the Rh/Cu/Oz system.*07 Anthraquinone was produced in 96% yield from the oxidation of anthracene with THHP in the presence of RhCI(PPh,), catalyst (equation 93).*66Oxidation of cylic alkenes, e.g. cyclopentene, cyclohexene, cycloheptene, by TBHP in the presence of Rh2(OAc)4results in the formation of a$-unsaturated ketones and allylic acetates as the major products (equation 94).267

96 %

Uses in Synthesis and Catalysis

350

In contrast to inactive iridium(TT1)-Peroxo complexes, 1r"'-hydroperoxo species have been shown to transfer oxygen to a coordinated alkene, for example in the slightly catalytic oxidation of cyclooctene to cyclooctanone by 02+H, mixtures in the presence of IrHCI2(C8H,,) (equation 95).26sOxygen transfer presumably occurs as for palladium hydroperoxides in equations (89) and (90). IrHC12(C8H,,)

0 &+ Ir(00H)CI2(CsHI2)

4

-

Ir(OH)CI2 + CsH120

(95)

I

A similar ketonization of coordinated cyclooctene was observed in the reaction of [ I ~ ( C ~ H , Z ) ~ ] 'with B F ~O2 - in the presence of one equivalent of fluoroboric acid, and probably occurs cia Ir-OOH species according to a similar r n e c h a n i ~ m . ' ~ ~ As previously shown for platinum versus palladium complexes, iridium complexes are generally much less effective catalysts than rhodium ones for the ketonization of alkenes by 02,H 2 0 2 or ROOH.

61.3.2.3 Conclusion

The catalytic properties of transition metal complexes in the reactions considered above are related to the existence and reactivity of their peroxide derivatives. For d n metals the general characteristics of the oxidations and the resulting oxygenated products are similar using hydrogen peroxide or alkyl hydroperoxides, although the nature of the active species, peroxo or alkyrperoxo, is different. The same is true for Group VI11 metals for which the ketonization of alkenes occurs using O,, H,Oz or ROOH as the oxygen source, and involves different peroxide species. One likely rationale is to consider that the active peroxo, alkylperoxo or hydroperoxo species reacts in the triangular form (89). This has been shown to be the case for d o metal peroxides, but not for Group VI11 metal alkyl- or hydro-peroxides which, in the solid state, have an end-on linear or bridged structure. However, the recent preparation of v2-SSR complexes, e.g. [ Ir( $-SzMe)(dppe),12+268a,b and [Os( q2-S2Me)(CO),(PPh,),lf,268cwhich are the sulfur analogues of ~ ~ - 0 0 R complexes, could suggest that such 0,O-bonded alkyl peroxide species might also exist in solution. A further rationale is that the heterolytic or homolytic nature of the catalytic oxidation seems to be strongly governed by the heterolytic or homolytic dissociation mode of the peroxide moiety, which depends on the nature of the metal (Scheme 7).

I-' v, C'r

(89)

I

Ti.V.Mo.W

Rhl Ir, Pd, I f

M,:

[

;O

Homolytic cleavage

;o

M;

0

Heterolytic cleavage

R Scheme 7

Heterolytic oxidations require releasable coordination sites on the metal, involve strained metallocyclic intermediates and are highly selective. By contrast, homolytic oxidations involve bimolecular radical processes with no metal-substrate association and are less selective.

61.3.3 METAGOXO COMPLEXES IN OXIDATION The use of high-valent metal-oxo compounds as stoichiometric oxidizing reagents of organic substrates has been known for over a century. Table 5 lists some of the most representative examples of such compounds, together with their structural characteristics and their reactivity. A majority of metal-oxo reagents, viz. CrO,CII, MnO, , Fe042-, R u 0 4 and OsO,, involve the metal in its higher oxidation state with a do configuration and a tetrahedral str~cture.~' In contrast to catalytic oxidations involving metal peroxide intermediates, catalysis by metal-oxo complexes does not in theory require the presence of a coreducing agent. Indeed, the reduction

Metal Complexes in Oxidation

351

Table 5 Reactivity of Metal-Oxo Complexes

Reoxidizing agent for catalysis

d(M=O) v(M=O) Oxo complex

V 0 2(pic)(HMPA) Crv'02C12 CrViOZ(OH),

Smtcrure"

(A)

TB TH TH

1.6 1.57

-

crV103(PY)z crv'03ct TH Crv'O, (NO,)? SP Cr"O(salen)+ M O ~ ~ O ~ C ~ ~ ( D M F ) ~ OH OH MoV'02(Et2dtc), MnV1'04TH ReV",O7 TH FeV'O:TH RuV"'O, OSV"'0,

SeIVO, a

-

(m-l)

935,948 984

-

-

-

-

1.53

907,954

-

-

-

1.49 1.68 1.63 1.62 1.65, 1.74 1.70

TH

1.697

-

1.78

997 905,940 908,877 838,921b 1009 778,800b 883, 9Hb

PhIO O,(for PR,) H2O2 CIO-, PhIO, IO4-, S20:-, R 3 N 0 964, 953b H,Oz, TBHP, C10-

c103

924

TBHP

Reactivity andlor refs.= N R ~ ~ ~ ~ 2 6~ ~9 2 7 0 , 2 7 OA27a.272 1

~

~

2 07~ 20 7 0

0~273,274

0~275,276

SE277 SE278 NR2'9 p02xn,2x~

282, HA, OC2s".z84 E, KZa5 HA, 540~286.287 OC, OA288,289 DO57s290 ~0291-293

Abbreviations: Et,dtc= diethyl dithiocarbamate; TB = trigonal-hipyramidal; TH =tetrahedral; SP = square-pyramidal; OH = octahedral. v l ( A , ) and v3(F3) absorptions, data taken from ref. 50. N R = not reactive toward hydrocarbons; SE = stereoselective epoxidation; E = epoxidation; HA = hydroxylation of alkanes; OA = oxidation of alcohols to carbonyl compounds; PO=oxidation of phosphines to phosphine oxides; OC = oxidative cleavage of alkenes; K = ketonization of alkenes; DO =hydroxylation of alkenes to diols; A 0 =allylic oxidation of alkenes.

of the oxo metal by a substrate could directly produce the reduced complex, which can be reoxidized by dioxygen to the initial oxo complex. In fact, with one notable exception (the catalytic oxidation of phosphines by O2 in the presence of dioxodithiocarbamatemolybdenurn(VI) complexes), the reduced species of metal-oxo reagents cannot be reoxidized by air, but require stronger oxidants such as C10-, PhIO, IO4-, S202-,etc., and in some cases H202or TBHP (Os, Se). 61.3.3.1 Chromium

The oxidative properties of chromium-oxo complexes towards organic substrates have been thoroughly investigated, and several reviews have appeared in recent years.270-276 We will only briefly consider the oxidation of alcohols to carbonyl compounds, the epoxidation of alkenes and the hydroxylation of hydrocarbons. 61.3.3.1.1 Oxidation of alcohols to carbonyl compounds

A variety of chromium(V1)-oxo reagents, e.g. chromic acid, dichromate ion, Cr03(py)2, Cr03(Bipy), Cr02C12,CrO,(OAc), , Cr02(OBu'), , CrO,Cl-pyH+, have been extensively used by organic chemists in both soluble and supported forms for the selective oxidation of alcoho\s.270-276.294 Under optimal conditions, 1.5 equivalents of carbonyl compound? are produced per mole of chromium(V1)-oxo reagent (equation 96). 3R2CHOH+Cr0,+6H+

+

3R2C=0+2Cr"'+6H,0

(96)

It is now generally admitted that this reaction involves both one-electron and two-electron transfer reactions. Carbony1 compounds are directly produced from the two-electron oxidation of alcohols by both Crm- and Crv-oxo species, respectively transformed into Cr'" and Cr"' species. Chromium(IV) species generate radicals by one-electron oxidation of alcohols and are responsible for the formation of cleavage by-products, e.g. benzyl alcohol and benzaldehyde from the oxidation of f , Z d i p h e n y l e t h a n ~ l . *The ~ ~ *key ~ ~ step ~ for carbonyl compound formation is the decomposition of the chromate ester resulting from the reaction of the alcohol with the Crvi-oxo reagent (equation 9 ~ ) : ' ~

352

Uses in Synthesis and Catalysis

The most widely used chromium(V1)-oxo reagents for synthetic use in organic chemistry are the Collins reagent, Cr03(py)2:73 and the Corey reagent, Cr03C1-pyH+,2’7 with the latter being superior for the oxidation of primary alcohols to aldehyde^.'^'

61.3.3.1.2 Epoxidatwn of alkenes

Chromium(V1)-oxo compounds are generally poor reagents for the selective transformation of alkenes to epoxides and give a complex mixture of However, the reaction of CrO,Cl, with cis-alkenes at low temperature was found to give mainly cis-epoxides, as well as chlorohydrin and vicinal dichloride arising from cis addition processes. These results were interpreted by Sharpless by a mechanism involving an initial alkene-chromium( VI) r-complex which rearranges to an oxametallacyclobutane intermediate which decomposes to give the epoxide with high retention of configuration (equation 98).269

Ab initio theoretical methods showed that oxametallacyclobutane formation was indeed the most favored route for the reaction of chromyl chloride with ethylene.298Whereas the chromium metallacycle preferentially decomposes to epoxide by reductive elimination, calculations predict that the molybdenum and tungsten analogs wouId preferentially undergo carbon-carbon bond cleavage leading to metal-carbene species and carbonyl products arising from oxidative cleavage of the double bond (equation 99). However, MoO,Cl, and W02C12complexes have been found to date to be quite reluctant to react with alkenes at moderate temperature^.'^^ M=Cr

0

Chromyl nitrate, Cr02(NO3),, was found by Miyaura and Kochi to be a much better epoxidizing reagent than Cr02C1,.277This reaction is solvent dependent. In basic solvents such as DMF or pyridine, epoxide is the major product (equation loo), whereas in an acetone solution alkene ketal, resulting from the addition of acetone to epoxide, is predominantly produced (equation 101).

The epoxidation of alkenes in DMF occurs with a high retention of configuration. Both cisand trans-P-methylstyrene are stereospecifically converted to the corresponding cis- and transepoxides, respectively, in high yield. The reactivity of alkenes follows the order norbornene > a-methylstyrene > P-methylstyrene > styrene > cyclohexene >> 1-octene. ESR experiments showed that the actual oxidizing species are oxochromium(V) compounds formed by one-electron oxidation of solvent. In further support of this finding, Kochi and coworkers isolated, and characterized by X-ray crystallography, the chromium(V)-oxo complex CrO(salen)+PF,- (90a) obtained by . ~ ~ ~complex (90a) is an effective treatment of Cr(SaIen)(H20)2tPF6- with i o d o s y l b e n ~ e n e This stoichiometric reagent for the selective epoxidation of alkenes, and can be used to catalyze the epoxidation of alkenes by iodosylbenzene (equation 102). The reactivity of various alkenes with

Metal Complexes in Oxidation

3 53

this catalytic system is similar to that previously observed in stoichiometric epoxidation by chromyl nitrate. PhIO +

)=(

- +A CrO(salm)+

PhI

(102)

The catalytic epoxidation of equation (102) is similar to that previous!y found by Groves and coworkers using chromium(II1) tetraphenylporphyrin as catalyst (equation 103).299These authors convincingly demonstrated, by means of IR, ESR and "0-labeling experiments, that the active species of reaction (103) is an oxoporphinatochromium(V) complex, CrO(TPP)C1(90b), presumably with a structure similar to (90a). One question raised by the reactivity of (90a) and (90b), which have no suitable coordination site, is whether the alkene is coordinated to the metal before oxametallacyclobutane formation as in the previous Sharpless proposal of equation (98). Groves suggested that the attack of the double bond occurs from the side and parallel to the plane of the porphyrin, which corresponds to a maximum overlap between the .rr-orbitalsof the approaching alkene and those of the singly occupied .rr-antibonding orbitals of the metal-oxo group?'" PhIO

+)=(

Cr(TPP)CI

PhI

+

61.3.3.1.3 Hydroxylaation of hydrocarbons The oxidation of toluene to benzoic acid by CrOzClzin acetic acid was described by Carstanjen more than 100 years Shortly thereafter, Etard isolated an organometallic complex with the formula PhMe.2Cr02C12 from the reaction of CrOZCl, with toluene."' This complex gave benzaldehyde in high yield upon treatment with water. A similar complex from triphenylmethane gave triphenyl methanol on hydrolysis.271Unfortunately, X-ray structural analyses of these complexes are still lacking and would greatly help in understanding the mechanism of hydrocarbon hydroxylation by chromium-oxo reagents. The main characteristics of the oxidation of hydrocarbons by chromic acid and chromyl compounds are as follow^.^^^*^^ (1) Alkanes are preferentially hydroxylated at the more nucleophilic C--H bonds, with This reaction relative reactivities tertiary :secondary :primary hydrogens = 7000 : 110: l.303 occurs with a high retention of configuration at the hydroxylated carbon atom, as shown by the selective formation of cis-9-decal01from the oxidation of cis-decalin with chromyl acetate in an acidic r n e d i ~ m "and ~ the hydroxylation of chiral (+)-3-methylheptane (91) to chiral alcohol (92) with 72 to 85% retention of c ~ n f i g u r a t i o n . ~ ~ ~

(92)

(91)

(2) Toluene and its substituted derivatives are oxidized by sodium dichromate or chromyl chloride mainly into the corresponding benzoic acids or benzaldehydes in good yield^.^" No or very little ring hydroxylation is observed. (3) Polynuclear aromatic compounds such as naphthalene or anthracene are oxidized by chromyl reagents mainly into quinones, with a significant NIH shift, providing evidence for the intermediacy of arene oxide Among the various proposals for the hydroxylation of C-H bonds by chromyl reagents, the mechanism involving a hydrogen atom abstraction by the oxo reagent giving a chromium(V)carbon free radical cage structure (93), followed by a rapid collapse to give the alkyl-chrornium(V1) 0

on

I

(94)

X,Cr-O

\ + IIII~IIC-

4

3 54

Uses in Synthesis and Catalysis

intermediate ( 9 4 ) or the Cr'" ester (95), which decomposes to alcohol and Cr'", satisfactorily accounts for most of the data concerning this reaction.

61.3.3.2 Molybdenum and Tungsten Trioxo and cis-dioxo complexes of molybdenum(V1) and tungsten(V1) are more stable and therefore much less reactive than their chromium analogs. However, the use of strongly donating sulfur chelate ligands such as dialkyl dithiocarbamate (R2dtc)280results in the labilization of the Mo=O bonds in the complexes MoO,(R,dtc),, as shown by the decrease in the IR frequencies of the Mo=O stretching vibrations in comparison with other MeV'-cis-dioxo complexes.3o7 Rarral and Bocard showed that the complexes Mo02(R2dtc), (R=Et, Pr", Bu') catalyze the oxidation of phosphines to phosphine oxides at room temperature and atmospheric oxygen pressures.280The steps of this catalytic reaction have been studied individually. The molybdenumdioxo complex (96) reacts with PPh, in an inert atmosphere to give Ph,PO and the isolated Mo"-monooxo complex (97), and this reduced complex (97) reacts readily with dioxygen to regenerate the active dioxo complex (96). This catalytic oxidation is complicated by a disproportionation between (96) and (97) giving a dinuclear Mo' species (98; equation 106).280 PR;

PR;O

u \

0

0

(R,dtc),Mo'V=O

II

II

e (R2dtc)2Mov-Q-Mov(R,dtc),

(106)

1

io,

(96)

(97)

(98)

Since this reaction is inhibited by the presence of strongly complexing diphosphines, it seems likely that oxidation proceeds through the prior complexation of the phosphine to the metal followed by the transfer of an oxo oxygen atom to the coordinated phosphine (equation 107).12 These complexes (96) also oxidize hydrazobenzene io a~obenzene,~'~ but are weak oxidants and do not react with alkenes.s6

c;:

Fh,Pr*MoV'=O

+

Ph

:3

i Mo=O y

6

4

Ph,PO-Mo=O

+

Mo'"=O+Ph,PO

(107)

Molybdenum(V1)-oxo complexes intervene as reactive species in the selective allylic oxidation of propene to acrolein in the gas phase over bismuth molybdate catalysts at high temperatures (>300 0 ~ ) . 5 6 , 3 0 8 - 3 1 0 In industrial processes, selectivities in acrolein reaching 90% can be obtained using multicomponent catalysts based on Bi203-Mo03.308 It is now understood that this reaction proceeds via the initial formation of a symmetric allylic intermediate. Grasselli has proposed that the role of bismuth oxide is to perform the rate-determining a-hydrogen abstraction step in a Bi-Mo dinuclear species (100) to give a n-allyl complex (101). Oxygen insertion into the metal-carbon bond followed by decornpostion of (102) results in the formation of acrolein and reduced dinuclear species ( 103).308~309

(102)

(103)

A somewhat similar mechanism involving Mo"'-imino species (Mo=NH) resulting from the reaction of ammonia with Mo=O bonds has been suggested for the industrially important ammoxidation of propene to acrylonitrile (equation 109).308

Metal Complexes in Oxidstion CH,=CHMe+ NH,+$O,

RI,O,/MOO,

CH2=CHCN+3H2O

355 (109)

Bismuth molybdate catalysts can also cause other allylic oxidations such as the conversion of 2-methylpropene to methacrolein and a-methylstyrene to atropaldehyde, and ammoxidations such as the conversion of 2-methylpropene to methacrylonitrile and a-methylstyrene to atr~ponitrile.~'~

61.3.3.3 Manganese and Rhenium 61.3.3.3.1 Oxidations by permanganate

The permanganate ion MnO; has been widely used by organic chemists for its strong oxidizing properties towards organic substrates.284"''-313The low solubility of Permanganate in organic solvents requires the use of oxidation-resistant cosolvents such as acetic acid, acetone or t-butyl alcohol. Recent improvements consist of the use of crown or tetraalkylammonium and arsonium salts,315*3l 6 for solubilizing permanganate in inert solvents such as CH2C12or benzene. Impregnation on inorganic supports such as molecular sieves or silica gel also activates permanganate for reaction in organic ~olvents.~"The reactivity of permanganate is pH dependent and is generally enhanced in basic conditions. Manganese dioxide is usually obtained as a brown precipitate at the end of the oxidations. The main characteristics of permanganate oxidation are described below. (i)

Alcohols

Primary and secondary alcohols are readily oxidized by permanganate while tertiary alcohols are stable.2849311 Primary alcohols are transformed into carboxylic acid via the formation of aldehydes. Secondary alcohols are cleanly oxidized to ketones.294 (ii)

Alkenes

Decoloration of a permanganate solution (Baeyer test) is a standard procedure far detecting unsaturation in organic compounds. Oxidation of alkenes by aqueous permanganate generally results in the formation of cis- 1,2-glycols, a-hydroxycarbonyl compounds and oxidative cleavage products. It is generally accepted that cyclic manganese(V) esters (104)are formed in the initial step of the reaction (Scheme 8)18**318as well as in the oxidation of alkenes by Ru04 and Os04 (vide infra). cis-Glycols result from hydrolysis of (104), while cleavage products and ketols are presumably derived from a cyclic Mn"' ester (105) resulting from the oxidation of (104) by the permanganate. H

Y +

AH

Scheme 8

Some synthetically useful oxidations of alkenes by permanganate can be performed under controlled conditions. For example, 1-decene could be oxidized to nonanoic acid in 91% yield by permanganate in the presence of the phase-transfer agent Aliquat 336.319In a benzene solution with crown ether and permanganate, a-pinene is oxidized to cis-pinonic acid in 90% yield (equation 1 1 0 ) . ~ ' ~

Uses in Synthesis and Catalysis

356

(iii)

Alkanes

Permanganate is a strong oxidant capable of hydroxylating alkanes under relatively mild conditions at the more nucleophilic C--H bonds in the order tertiary> secondary> primary. Hydroxylation of tertiary C-H bonds occurs with retention of configuration, as shown by the selective formation of cis-9-decalol from the oxidation of cis-decalinwith benzyltriethylammonium permanganate.316Oxidation of secondary methylene groups by permanganate gives the corresponding ketone, as in the oxidation of 2-benzylpyridine to 2-benzoylpyridine (80% yield).”” Methyl-substituted aromatic hydrocarbons are readily oxidized by permanganate into benzoic acid derivatives.311Toluene, p-xylene and mesitylene are respectively converted to benzoic, terephthalic and trimesic acids in ca. 90% yield by potassium permanganate in the presence of cetyltrimethylammonium ~hloride.~” 61.3.3.3.2 Oxidation by manganese dioxide

Freshly prepared Mn02 is a useful reagent in organic chemistry and has been used in a large variety of oxidative transformation^.^" These reactions involve the allylic oxidation of alkene to a,P-unsaturated carbonyl compounds, the transformation of methylarenes to benzaldehyde and benzoic acid derivatives,the oxidation of secondary methylene groups to ketones, and the oxidation of alcohols to carbonyl c o m p o ~ n d s . ~ The ’ ~ yields are generally fair to good. Manganese(V)-oxo-porphyrinato complexes have been presumed to be the active species in the epoxidation of alkenes by iodosylbenzene or hypochlorite catalyzed by manganese( 111) porphyrins. This and related oxidations will be examined in more detail in Section 61.3.6.2.2. 61.3.3.3.3 Oxidation by rhenium-oxo species

Although high-valent rhenium-oxo complexes such as perrhenate, Re,O,, and ReO,X (X = F, C1, Br) are widely known,322their use as oxidant$ is rather rare. The stoichiometric oxidation of alkenes to a mixture of ketone and epoxide by Re,O, has been reported in the patent literat~re.”~ For example, 2-butene is transformed to 2,3-epoxybutane and 2-butanone by Re207at 100 “C. A catalytic oxidation was observed in the presence of excess hydrogen peroxidezg5but precise data on this reaction are lacking.

61.3.3.4 Iron, Ruthenium and Osmium 61.3.3.4.I

Oxidation by ferrate ions

The only relatively stable high-valent iron-oxo complexes reported to date are ferrate(V1, V, Fe033- and Fe0,4-?23 There is considerable interest in knowing the reactivity of high-valent iron-oxo compounds, owing to the current opinion favoring their involvement as active species in biological oxygenases. Unfortunately, rather little about the chemistry of such compounds has been reported. Potassium ferrate, K2Fe0,, is isomorphous with chromates and readily oxidizes alcohols to carbonyl compounds in the presence of alkali.286*2g7 The purple color of ferrate ion fades with the progress of the reaction. Yields as high as 96% have been reported for the oxidation of cinnamyl alcohoI to cinnamaldehyde (equation 11 l).287 IV) ions, i.e. Fe0:-,

Barium ferrate suspended in glacial acetic acid oxidizes squalene to monoepoxysqualene in low yield.306Potassium ferrate is unreactive towards alkanes, but when reduced by oxalic acid it

Metal Complexes in Oxidation

357

produces a presumed FeLv-oxospecies capable of hydroxylating adamantane to I-adamantanol and 2-adamantan0ne.~~

61.3.3.4.2 Oxidation by rutlrenium-oxo compounds

Ruthenium tetroxide, Ru04, is a powerful oxidant, soluble in organic solvents, which was introduced by Djerassi in 1953 for applications in organic synthesis.324It is easily prepared by the oxidation of low-valent ruthenium compounds, e.g. RuC& or Ru02, by strong oxidants such as periodate, hypochlorite, chromic acid, etc., and can be conveniently extracted from aqueous solutions into carbon tetrachloride,288Although Ru04 is a strong stoichiometric oxidant by itseIf, modern improved procedures use only catalytic amounts of ruthenium precursors along with a suitable oxygen source such as hypochlorite, periodate, persulfate, iodosylbenzene or tertiary amine N - ~ x i d e . ' ~ ~ (i) Oxidation of alcohols to carbonyl compounds

Ruthenium tetroxide readily converts secondary alcohols to the corresponding ketone, and primary alcohols to aldehydes and It is particularly recommended for converting alcohols which are difficult to oxidize with other reagents, for example, the hydroxylactone (105a) in equation (1 1 2 p 5

RuOJ NalO,

'0 (105.)

80%

Highly selective ruthenium-catalyzed oxidation of alcohols can be achieved using low-valent complexes and milder cooxidants. For example, oxidation of primary alcohols can be cleanly stopped at the aldehyde stage by using RuC12(PPh3),as a catalyst and iodosylbenzene d i a ~ e t a t e ~ ' ~ or N-methylmorpholine N-oxide (NMO)327as the oxygen source. The latter system oxidizes only hydroxyl groups and preserves double bonds, as shown by the oxidation of (+)-carved (equation 113).327

OH

'0

94%

Ruthenium complexes such as RuC13,328RuC1, + C U ( C ~ Oand ~ ) RuC12(PPh3), ~~~~ 330 have been shown to catalyze the direct oxidation of alcohols to carbonyl compounds by molecular oxygen. Allylic alcohols are cleanly oxidized to the corresponding a,p-unsaturated carbonyl compounds. For example, (+)-cameo1 is selectively oxidized by air to (+)-carvone (92% yield) at room ~ . ' ~ reaction ' temperature in C2H,C12 in the presence of catalytic amounts of R u C ~ ~ ( P P ~ , ) This presumably involves the formation of a Ru'" alkoxide, which undergoes @-eliminationto produce the carbonyl compound and a ruthenium hydride which could be reoxidized by molecular oxygen.330 (ii) Oxidative cleavage of alkenes

Ruthenium tetroxide is a four-electron oxidant which directly transforms alkenic compounds into oxidative cleavage products, i.e. carbonyl compounds and carboxylic acids.288The reaction can be visualized as proceeding according to a [4+ 23 cycloaddition of the cis-dioxo moiety with the alkene, resulting in the formation of a Ru"' cyclic diester which decomposes to ruthenium(1V) dioxide and oxidative cleavage products (equation 114).**' This reaction can be made catalytic

358

Uses in Synthesis and Catalysis

using small amounts of RuO,, RuC13 or RuOz in conjunction with NaOCl (equation 115),331 NaIO, (equation 116j332or peracetic acid (MeCHO+ O,, equation 117j333as cooxidant. Me(CH2),CH=CH(CH2),C02H

NaOCI:RuO,

Me(CH,),CO,H+ H0,C(CH,),C02H 94%

u

(115)

94%

U

94?0

(iii) Miscellaneous oxidations

One of the most remarkable reactions performed by ruthenium-oxo complexes is the selective oxidation of aliphatic and cyclic ethers to the corresponding esters and lactones, respectively, with high yields under mild condition^.^^'.^^^ For example, tetrahydrofuran is oxidized by RuO, to y-butyrolactone with nearly quantitative ~ i e l d s . 3 ~ A ~catalytic reaction can be carried out in the presence of excess hypochlorite or p e r i ~ d a t e . ~ ~ ~ ~ ~ ~ '

-

Ph

,Me 0

0 N ~ I O/ R ~ C I

(118)

A 89%

Carbon-carbon bonds can be selectively cleaved by ruthenium-oxo compounds, such as in the transformation of phenylcyclohexane to cyclohexanecarboxylic acid (equation 119j.332

OPh

NaIOJRuCI,

94%

Polynuclear aromatic hydrocarbons are oxidatively cleaved by RuO, under mild conditions, and this reaction can be performed catalytically in the presence of NaOCl (equation 120).336

65%

Dialkyl- and diaryl-acetylenes are selectively converted to a-diketones (equation 121 1, whereas terminal alkynes give the corresponding carboxylic acid with loss of one carbon atom.337 PhCECPh

liaC€l/RuO,

CC1,-H20

PhC-CPh

II II

0 0 83%

61.3.3.4.3 Oxidation by osmium-oxo compounds

Osmium tetroxide is a mild, reliable reagent for the selective oxidation of alkenes to cis-1,2glycols. Discovered a long time ago,338,341 this reaction has been widely used for synthetic purposes by organic ~ h e m i s t s . ~ ~ ~ . ~ ~ ~ The Criegee mechanism originally proposed involves the formation of an osmium(V1)-ester complex (106) from the [4+2] cycloaddition of the Osv1" cis-dioxo moiety with an alkene, followed by the hydrolysis or reduction of (106) to cis-glycol and reduced osmium species. In support of this mechanism a variety of OS"' cyclic esters such as (107) or (108) (L = quinuclidine) have recently been synthesized from Os04 and the alkene, and characterized by an X-ray crystal structure.290343 In solution the dimeric complex (108) dissociates to give the monomeric dioxo trigonal-bipyramidal complex (109), which is simiIar to ( 106).344

Metal Complexes in Oxidation

359

(10s)

Sharpless and Hentges proposed that the cyclic ester (112) is formed via an oxametallacyclobutane (111) resulting from the [2 21 intramolecular addition of an oxo bond to the coordinated alkene in ( l10)?45They utilized the high stereoselectivity of this reaction to induce chirality with good optical yields (up to 83%) during the hydroxylation of various alkenes by OsO, in the presence of chiral pyridine ligands.345

+

The same researchers found that imino derivatives of Os04 react with alkenes to produce This oxyamination reaction can be made catalytic in cis-p-amino alcohols (equation 124).346,347 the presence of chloramine salts, e.g. ArS0,NClNa or ROCONC1Ag.3487349

Since OsO, is volatile, toxic and expensive, considerable effort has been devoted to the catalytic application of the cis dihydroxylation of alkenes in the presence of excess c o o ~ i d a n tPrevious .~~~~~~ procedures used metal chlorate (Hoffman reagent),339or hydrogen peroxide (Milas reagent)350 as cooxidant, usually in BdOH or acetone.290Recent procedures utilize t-butyl hydroperoxide or N-methylmorpholine N in conjunction with ammonium salts ( Et4NOH or Et4NOA~)573351 o ~ i d e , "and ~ are generally more selective.

*

Me

TBHP/ Et,NOAc/OsO, acelone

CO,Et

HO

OH

Meu1-H H 72%

COzEt

NMOIOsO, THF-H,O-Bu'OH

61.3.3.5 Selenium and Tellurium

61.3.3.5.1 Allylic oxidations by selenium dioxide

Selenium dioxide, SeO,, is a widely used reagent for the allylic oxidation of alkenes and ketones. The subject has been extensively covered by recent review articles56*291-293 and only a short summary will be given in this chapter.

360

Uses in Synthesis and Catalysis

The oxidation of alkenes by SeOz (or its hydrated form SeO(0H)J results in the formation of allyl alcohols and a,&unsaturated compounds, or allylic acetate when the reaction is carried out in acetic acid. (a) The oxidation generally occurs without double bond rearrangement. (b) Oxidation of 2-methyl-2-alkenes occurs at the (E) methyl group. (c) 1-Substituted cyclohexene is oxidized at the 6-position; this rules out the formation of allylic free radicals or carbonium ions. The widely accepted mechanism of this reaction has been proposed by Sharpless: 353-355 it involves the initial ene addition (113) of the Se=O group to the alkene to form the allylseleninic acid (114), which undergoes a [2,3] sigmatropic rearrangement to give the allyl selenenate (115). The allyl alcohol (116) is formed by hydrolysis of (115), whereas the @-unsaturated carbonyl compound (117) results from the reductive elimination of (115), coproducing seIenium metal in its colloidal form.

(117) 0

Allyl alcohols can be produced catalytically by oxidation of alkenes with TBHP in the presence of small amounts of SeO, (equation 128). In contrast to the stoichiometric reaction, this catalytic oxidation can be performed under mild conditions (r.t., CH2C12solvent).356Alkynic compounds under o a predominant allylic dihydroxylation upon reaction with TBHP/CH2Cl2 (equation 129).3

P

C7H15-

TBHPJScO,

CH,CI,, 25 T, 2d

’ OH 60%

- -

-

0 55%

Selenium dioxide is a most useful reagent for the oxidation of ketones or aldehydes to a-dicarbonyl compounds along with a$-unsaturated carbonyl compounds as b y - p r o d u ~ t s . ~ ~ ’ , ’ ~ The carbonyl compound probably reacts in its enol form in a way similar to that of alkene oxidation (equation 130).358

61.3.3.5.2 Oxidation by tellurium-oxo compounds The use of tellurium dioxide as an oxidation reagent has been seriously hampered by its very low solubility in almost all organic solvents. However, TeO, appears to be a much milder oxidant than SeOz. Catalytic systems containing TeO,, HBr and AcOH have been used industrially by Oxirane to convert ethylene to ethylene glycol via the formation of mono- and di-acetate (equations 131 and 132).359-361 The overall yield from ethylene to ethylene glycol is more than 90%, making this reaction competitive with respect to the older silver-catalyzed ethylene epoxidation process.

Metal Complexes in Oxidafion

361

However, the development of this process has been suspended, largely because of corrosion problems.361 CH,=CH,+2AcOH+0.5O2

TeO /HBr

AcOCH,CH,OAc+ H 2 0

AcOCH2CH20Ac+2H20 + HOCH,CH,OH +2AcOH

(131)

(132)

A similar TeO,/HBr catalyst has been used in acetic acid for the oxidation of propene at ca. 120 "C to propene oxide and propene glycol via 1,2-dia~etoxypropane.~~~*~~~ However, this reaction gives lower yields because of side allylic oxidation reactions. Oxidation of toluene under similar conditions (TeOJLiBr, 160 "C) results in the formation of methylbenzyl acetate mixtures rather than benzyl acetate as observed with SeO,/LiBr catalyst (equation 133).363 120 ' C , AcOH

aMe 1 2 0 T , AcOH

CHH,OAc

Bis(pmethoxypheny1) telluroxide (118) and the corresponding tellurone (119) have recently been shown to exhibit mild oxidizing properties towards easily oxidizable substrates. Thus the telluroxide (118) oxidizes thiocarbonyl compounds RR'C=S to the corresponding ketone ~ ~ ! ~ = 0 3 6 4 , 3 6 5and 3,4-di-t-butylpyrocatecholto 3,4-di-t-b~tyl-o-quinone.~~ The tellurone (119) oxidizes benzyl alcohols to the corresponding carbonyl compounds, a reaction ,which is not observed with (1 18)?66 0

0

I1

)I

p - MeOC,H,-Te-C,H,OMe-p

p-MeOC,H,-Te-C,H,OMe-p

II

0 (118)

(119)

61.3.4 OXIDATION 3Y NUCLEOPHILIC ATTACK AT HYDROCARBON SUBSTRATES COORDINATED TO PALLADIUM 61.3.4.1

Introduction

Discovered by Phillips in 1894,3'' the oxidation of ethylene to acetaldehyde by palladium(l1) salts in an aqueous solution was developed into a commercial process about 60 years later by Smidt and coworkers at Wacker Chemie.3837384 These researchers succeeded in transforming this stoichiometric oxidation by a precious metal (equation 150) into a catalytic reaction through the reoxidation of the resulting Pdo by molecular oxygen in the presence of copper salts (equations 151-1 52). CH2=CH2+ PdC12+H 2 0

Cu,CI,

+

MeCHO+Pdo+2HC1

(150)

Pdo f 2CuC1,

+

PdCl,

+ Cu2C12

(151)

+ 2HCl+ 0.50,

4

2CuC1,+

H20

(152)

The commercial success of this reaction was a major step leading to the withdrawal of acetylene-based petrochemical technologies, and provided a considerable stimulus in both academic and industrial circles for research on homogeneous catalysis by noble-metal complexes. This is reflected in the appearance of several books and many review articles devoted to the Applications of palladium catalysts in organic synorganic chemistry of palladium.367-370,384-394 thesis, including oxidation reactions, have been extensively covered by Trost and Verhoeven in 'Comprehensive Organometallic Chemistry'.395This section will deal only with catalytic reactions in which molecular oxygen is involved as the palladium reoxidant. Table 6 lists some of these catalytic oxidations which involve ketonization, glycolate formation, acetoxylation, oxycarbonylation, oxycyanation, oxychlorination, oxidative coupling, oxidation of alcohols and oxidation by coordinated nitro ligands. Some of these reactions, e.g. the oxidation of ethylene to acetaldehyde and of propene to acetone, the acetoxylation of ethylene to vinyl acetate and of butadiene to 1,4-diacetoxy-2-butene,and the oxycarbonylation of methanol to dimethyl oxalate, have been the object of large-scale industrial development.

Uses in Synthesis and Catalysis

362

Table 6 Palladium-catalyzed Oxidation of Hydrocarbons Caialyfic reaction ~~~~

Complex

NucIeophiIe

Refs.

~

Oxidation of alkenes io carbonyl compounds ( Wacker)

RCH=CH2+0.50,+

RCOMe

(134)

Pd--k

H,O

56,367-370

AcOH

371

R

Oxidation ofalkenes to glycolares

RCH=CH,+AcOH+O.502

4

RCH-CH2

I

(135)

1

R

OH OAc Acetoxylations OAc

CHZ=CH2+AcOH+O.5O2

/'-+AcOH+O:502 ,

-

+ HzO

+ H20

-0Ac

4

+2AcOH+0.502

PhMe+AcOH+0.50,

-+

4

+ H,O

AcO nOAc

PhCH,OAc+H,O

(136)

Pd- II

AcOH

367-370

(137)

Pd-lj

AcOH

367

AcOH

372

AcOH

373

Pd- II

-

374

(145)

Pd- It

HCN

379

(146)

Pd-II

HCI

370

(138) Pd)/

(139)

;", \

Oxycarbonylations

CH,=CHz+C0+0.5O2

C02H

-+

=/

PhH+CO+0.50,+ PhCOzH RCH=CH,+CO+H,O -D RCH(C0,H)Me 2ROH+2CO+ 0.502- R02C-COZR Oxycyanation

CH2=CH2+HCN+0.502+

CN

4

+

H20

Oxychlorination

CH,=CH2+HCI+0.50,+

4

+ H20

Oxidatiue coupling 2ArH+0.50, -+ A r c A r + H 2 0

3 80

Oxidation of alcohols to carbonyl compounds RR'CHOH + 0 . 5 0 , 4 RR'C=O + H,O

Most of the reactions listed in Table 6 involve prior activation of the substrate by coordination to palladium in the form of a T-, a n-allylic, a T-benzylic, or an alkyl or aryl complex. Once coordinated to the metal, the substrate becomes an electron acceptor and can react with a variety of different nucleophiles. The addition of nucleophiles (Nu) to the coordinated substrate may ~ ' -external ~~~ occur in two different ways, as shown by Scheme 9 for T-alkene c o m p l e ~ e s : ~(a) attack leading to trans addition of palladium and nucleophile across the r-system (path A); or (b) internal addition of the coordinated nucleophile to the complexed alkene resulting in cis addition of palladium and nucleophile to the double bond. The cis and trans adducts (120) and (121) may then undergo p-hydride elimination (p-H), producing the vinylic oxidation product

Metal Complexes in Oxidation

363

(122), or palladium-carbon oxidative cleavage (O.C.)to give the cis and trans addition products (124) and (123). During this reaction the paIladiurn(I1) complex is reduced to metallic pal-

ladium(O), probably uia reductive elimination from the palladium hydride species. -INu

I

\

Pd"

PdO

+ H"

/

+ Pd'lH.

Scheme 9

Nucleophilic attack at a-allylic complexes may also occur from (a) external trans addition or (b) internal cis addition, resulting in the formation of allylic oxidation products (equation 153).398,399

Backwall and coworkers have extensively studied the stereochemistry of nucleophilic additions on .rr-alkenic and r-allylic palladium(11) complexes. They concluded that nucleophiles which preferentially undergo a frans external attack are 'hard' bases such as amines, water, alcohols, acetate and stabilized carbanions such as P-diketonates. In contrast, 'soft' bases are nonstabilized carbanions such as methyl or phenyl groups and undergo a cis internal nucleophilic attack at the coordinated substrate.398,399 The pseudocyclic alkylperoxypalladation procedure occurring in the ketonization of terminal alkenes by [RCO,PdOOBu'], complexes (see Section 61.3.2.2.2)42belongs to internal cis addition processes, as well as the oxidation of complexed alkenes by coordinated nitro ligands (vide

61.3.4.2

Oxidation of Alkenes to Carbonyl Compounds

The stoichiometric oxidation of ethylene to acetaldehyde by PdC1, in an aqueous solution (equation 150) has the following characteristic^.^^^*^^^ (i) At low Pd" concentration the rate expression is: -d(CZHJ / d t

=k[

PdCl2-] [C2H4]/ [Cl- ]'[ H+].

(ii) No H/D exchange occurs when C2H, is oxidized in DzO, or when C2D4is oxidized in H20. (iii) The kinetic isotope effect kH/kD is 1.07 when C2D4is oxidized in H 2 0 , and 4.05 when C2H4is oxidized in D 2 0 . The mechanism depicted in Scheme 10 satisfactorily accounts for most of the kinetic and stereochemical data concerning this r e a ~ t i o n , although ~ ~ ~ ~ ~some ~ " important ~ ~ details are still subject to debate.367 The key feature of this mechanism is the a-u rearrangement of the coordinated alkene (127) + (128) and the ensuing P-hydride elimination (128) + (129). Although kinetic measurements favor an internal cis migration of the coordinated hydroxyl group to the alkene (or ethylene insertion into a Pd-OH bond),"' recent stereochemical studies of oxidation400and oxycarbonylation,'' of cis- and tram-CHD=CHD indicate that the hydroxypalladation is trans, and probably results from the external nucleophilic attack of the coordinated ethylene by water. As for acetoxylation (vide infra), it is possible that both mechanisms operate, depending on the reaction conditions. The hydroxypalladation procedure has been convincingly illustrated by Moiseev who

Uses in Synthesis and Catalysis

364

Scheme 10

found that transmetalation of mercury by palladium in HOCH,CH,HgCl results in the exclusive formation of acetaldehyde (equation 154)."02This procedure has been used to obtain high yields of methyl ketones from higher terminal alkene^:'^

-

C1-Hg-CH2CHR+PdC1,

-HgCI,

I

OH

C1-Pd-CH,CHR

I

+

R-C-Me

I1

+ Pdo+HC1

(154)

0

OH

The commercial Wacker process for the manufacture of acetaldehyde operates in a single-stage or a two-stage pr0cedure.4'~In the single-stage process, ethylene and oxygen are fed into a reactor containing an aqueous hydrochloric solution of PdC1, (-4gl-') and a large excess of CuC12 (-200 g 1-') at 120 "C and 4 atm pressure. Ethylene conversion is ca. 40%, and oxygen conversion is almost complete. Acetaldehyde is distilled out of the reaction mixture to remove the reaction heat [AH = -242 kJ mol-'). In the two-stage process, ethylene reacts with the same catalyst solution at 105-110 "C and 10 atm until most of the Cu" has been reduced to CUI.After distillation of the acetaldehyde, the aqueous catalyst solution is reoxidized by air in a second reactor, and then recycled to the first reactor. Both processes give 95% yields of acetaldehyde, together with acetic acid and chlorinated by-products. A similar single- or two-stage process has been developed for the manufacture of acetone from propene (1 10-120 "C, 10 atm, PdCl2-CuC1,-HC1 catalyst) but on a smaller scale (80 000 t/year) The * ~yield is ca. 93%. than for acetaldehyde (2 500 000 t / ~ e a r ) . ~ Higher alkenes can also be converted to methyl ketones with the Wacker catalyst, but the rates and selectivities are lower. Improved procedures use b a s i ~ ~ ' ~ ,or " ~alcoholic ' solvents:" Tsuji and coworkers used the PdCl,/CuCl catalyst in DMF for the synthesis of a variety of natural products and fine cherni~als.~'~ Only terminal alkenes are ketonized under these conditions, even when the substrate contains other functional gr0ups.3~~ The Wacker PdC12-CuC12catalyst is a highly corrosive one, and i t s use requires special vessels and apparatus, such as titanium or tantalum alloys. Heteropolyacids have been used as alternative noncorrosive reoxidants of palladium for a variety of organic transformations?' €or example in reaction (155):"

0

+ 0.502

PdSOJ H,PMo,W,O,,

100%

Good yields of carbonyl compounds have also been obtained from the vapor-phase oxidation In this case, no of alkenes by steam and air over palladium catalysts supported on ~harcoal.~" copper cocatalyst is needed, presumably because palladium( 11) is not reduced to palladium(O), but remains in the form of a stabilized palladiurn(I1) hydride which can react with 0, to give the hydroperoxidic species.

Metal Complexes in Oxidation

3 65

In fact, the role of copper and oxygen in the Wacker Process is certainly more complicated than indicated in equations (151) and (152) and in Scheme 10, and could be similar to that previously discussed for the rhodium/copper-catalyzed ketonization of terminal alkenes. Hosokawa and coworkers have recently studied the Wacker-type asymmetric intramolecular oxidative cyclization of trans-2-(2-butenyl)phenol(132) by O2 in the ,presence of (+)-(3,2,10-7pinene)palladium(II) acetate (133) and Cu(OAc), (equation 156).413It has been shown that the chiral pinanyl ligand is retained by palladium throughout the reaction, and therefore it is suggested that the active catalyst consists of copper and palladium linked by an acetate bridge. The role of copper would be to act as an oxygen carrier capable of rapidly reoxidizing palladium hydride into a hydroperoxide species (equation 157).4'3 Such a process is also likely to occur in the palladium-catalyzed acetoxylation of alkenes (see Section 61.3.4.3). &Me

U p ')2 l O A C (133)

QTL

O*,c~~OAc)2,MeoH'

70%

(132)

+

o\

(156)

12%0

H

I

Acetalization of alkenes can be achieved in good yields when the oxidation is carried out in the presence of alcohols or diols. Acetalization of terminal alkenes such as 1-butene occurs preferentially at the 2-position in the presence of PdC1,-CuCl2 (equation lSS),"'" whereas terminal alkenes bearing electron-withdrawing substituents are acetalized at the terminal position in the presence of PdC12-CuC1 in 1,2-dimethoxyethane (equation 159)?15

68 %

30%

86%

61.3.4.3 Acetoxylation of Alkenes, Dienes and Aromatic Hydrocarbons 61.3.4.3.1

Vinyl acetute from ethylene

First discovered by Moiseev et ~ 1 , "the ' ~ palladium-catalyzed acetoxylation of ethylene to vinyl acetate has been the subject of very active investigations, particularly in industry, as shown by the considerable number of patents existing in this area. Vinyl acetate is an extremely important petrochemical product which is used for the synthesis of polymers such as poly(viny1 acetate) and poly(viny1 alcohol). Most of its annual production (-2.6 Mt) results from the acetoxylation of ethylene (equation 160). CH2=CH,+AcOH+0.50,

3 CH,=CHOAc+H,O

(160)

The vinyl acetate process exists in both homogeneous and heterogeneous versions. The liquidphase process developed by IC1 is essentially a Wacker reaction performed in acetic acid: ethylene, O2 and AcOH are reacted at 110 "C in the presence of PdC12, Cu(OAc)*and HCI. Overall yields are greater than 90%. Acetaldehyde is formed as a coproduct in the reaction, owing to the presence

Uses in Synthesis and Catalysis

366

of water, and is oxidized to acetic acid which is used in the process. However, this process encountered severe corrosion problems and was aband~ned.'~ The gas-phase process, successfully commercialized independently by Bayer and USI,417involves passing a mixture of ethylene, acetic acid and oxygen over a supported palladium catalyst contained in a multitubular reactor at 150 "C and about 5-10 atm pressure. The overall yield in vinyl acetate is about 92%, and the major by-product is CO,. The catalyst consists of a palladium salt (e.g. Na,PdCl,) deposited on silica (or alumina) in the presence of a cocatalyst (e.g. HAuCl,), reduced and impregnated with potassium acetate before use.3843418 The lifetime of the catalyst is about 2 years.4" A widely accepted mechanism for acetoxylation of ethylene is shown in equation (161) and consists of the nucleophilic attack of the acetate anion on the coordinated ethylene, followed by acetoxypalladation and P-hydride elimination, giving vinyl acetate and palladium hydride.367 AcO;

[XPd-OOH] (137)

It should be noted that heterogeneous palladium acetoxylation catalysts do not contain copper cooxidants, presumably because the support stabilizes the resulting palladium(11) hydride such as (136) and prevents the formation of metallic palladium. The stabilized palladium hydride (136) may react with O2to give the hydroperoxide (137), which is probably an important intermediate for the regeneration of the initial Pd" catalyst. Such a stabilization of the palladium catalyst can also be achieved in homogeneous liquid phase by the use of appropriate ligands. Thus, it has recently been shown that palladium(I1) hydroxamates are effective catalysts for the acetoxylation of ethylene with high selectivity and a high turnover (>200) (equation (162), whereas Pd(OAc), rapidly becomes deactivated and precipitates in the form of metallic palladium.419It is probable that the bidentate hydroxamate ligand stabilizes the hydride Pd-H species and prevents palladium from precipitating.

)::!I:(

CH,=CH,

r d , YaOAc

+ AcOH + 0.502

IoO"C,3h

* CH,=CHOAC+H~O

(162)

96%

61.3.4.3.2 Oxidation of ethylene to glycol acetate

When palladium-catalyzed acetoxylation is camed out in the presence of nitrate or nitrite ions, ethylene glycol monoacetate (EGMA) results as the major product of the reaction (equation 163).420 3C2H,+2LiN0,+SAcOH

4

CH2CH,f2LiOAc+2NO+ H 2 0

I

I

OH OAc

This reaction can be made catalytic by reoxidizing the resulting nitric oxide to nitric acid by air, and has been developed into a commercial process by Kuraray for the manufacture of ethylene glycol from the hydrolysis of EGMA?2i The overall yield of ethylene glycol in this process is ca. 95%. The mechanism of EGMA formation (equation 163) has not been studied intensively, but it is generally thought that EGMA results from the oxidative cleavage of the acetoxypalladation adduct (135; equation 161) by nitrite or nitrate ions which prevent P-hydride elimination.361 Yermakov et d.have recently shown, by 1 7 0 NMR studies, that the oxygen carbonyl atom of EGMA is derived from LiN03, and have therefore suggested that palladium( 11)-nitro species are involved as the actual oxidizing spe~ies.~" 61.3.4.3.3 Allylic acetoxylation of alkenes

In contrast to ethylene, which gives only vinylic or oxidative addition products, the acetoxylation of higher alkenes results in the formation of a mixture of allylic and vinylic acetates.367The

Metal Complexes in Oxidation

367

selectivity of this reaction can be controlled by adjusting the oxidation conditions. Thus the oxidation of propylene by PdC12 in acetic acid without added acetate gives 96% 2-propenyl acetate, whereas allylic acetate is the major product (94% ) when the same reaction is carried out in the presence of excess sodium Although allylic complexes are remarkably stable, the allylic oxidation of alkenes seems to proceed via a classical acetoxypalladation-palladium hydride elimination rather than from r-allyl &Hydride elimination preferentially occurs at the carbon center not containing the acetate group (equation 164).

-

+ Pd(0Ac)i

+ AcOPdH AcO

PdOAc

(164)

OAc

Both homogeneous and heterogeneous procedures have been developed for the synthesis of allyl acetate from propene (equation 137 and Table 6). Allyl acetate is an interesting intermediate which can be hydrolyzed to allyl alcoh01,4~~ or hydroformylated to a lP-butanediol or THF Acetoxylation of .propene to allyl acetate can be performed in the liquid phase with high selectivity(98% ) in acetic acid in the presence of catalytic amounts of palladium t r i f l ~ o r o a c e t a t e . ~ ~ ~ The stability and activity of this catalyst can be considerably increased by adding copper (11) trifluoroacetate and sodium acetate as cocatalysts (100 "C, 15 bar, reaction time = 4 h, conversion = TO%, selectivity = 97% >.Gas-phase procedures for the manufacture of allyl acetate are described in several patents and use conventional palladium catalysts deposited on alumina or silica, together with cocatalysts (Au, Fe, Bi, etc.) and sodium acetate. The activity and selectivity reported for these catalysts are very high (100-1000 g I-' h-', ~electivity=90-95%).~~~ A similar procedure has been used for the synthesis of methallyl acetate from 2-methylpr0pene.~~~

61.3.4.3.4

Acetoxylation of dienes

Palladium-catalyzed 1,4-diacetoxylation of butadiene is a useful reaction of commercial interest which provides an interesting alternative for the synthesis of butanediol and tetrahydrofuran, previously based on acetylene feedstocks (equation 165).

+ 2AcOH + 0.5OZ

-n o

(i) H,/cat

2

/--LOH

HO -

AcO-.-,

(165)

The selectivity in the formation of 1,4-diacetoxy-2-butene(1,CDAB) is considerably enhanced when tellurium compounds are used as cocatalysts. Thus a heterogeneous catalyst, prepared by impregnation of Pd(N03):! and Te02 dissolved in HNO, over active charcoal (Pd/Te = lo), can be used for the oxidation of butadiene (by O2 in AcOH at 90OC) to 63% trans-1,4-DAB, 25% cis-1,4-DAB and 12% 3,4-diacetoxy-l-butene. Conventional soluble catalysts such as Pd(OAc),/Li(OAc) are much less selective in the formation of 1,4-DAB.429The gas-phase 1,4diacetoxylation of butadiene in the presence of Pd-Te catalysts is currently being industrially developed by Mitsubishi and BASF.430 The mechanism of 1 ,4-diacetoxylation of dienes and its stereochemical aspects have been studied in great detail by Backwall and coworker^.^^^-^^^ Thus the 1,4-diacetoxylation of 1,3cyclohexadiene mainly occurs trans in the presence of Pd(OAc),/LiOAc/benzoquinone (BQ) (equation 167), whereas the use of an Li,PdCI,/LiOAc/BQ system results in the major formation

(a) Li,PdCI,, LiOAc

q -Pd-

1

AcO-PdI

I (138)

I--I

(b) W(OAc),,LiOAc

BQ

,

u

(166)

Uses in Synthesis rand Catalysis

368

of cis- 1,4-diacetoxy-2-cyclohexene(equation 166). The mechanism involves 1,2-acetoxypalladation of the diene followed by nucleophilic attack of the resulting .rr-allylic complex (138) by the coordinated acetate, giving the trans product (140; path b). In the presence of chloride ligands, external acetate addition on (138) occurs and gives rise mainly to formation of the cis isomer (139)."'

61.3.4.3.5 Aceioxyladon of aromatic compounds The reaction o f palladium(I1) salts with benzene in acetic acid gives a mixture of biphenyl and phenyl Nuclear acetoxylation is favored by the presence of sodium acetate, whereas oxidative arene coupling occurs at high oxygen pressure. The acetoxylation of benzene to phenyl acetate can be made catalytic in the presence of strong oxidants such as K2Cr207?32Supported palladium catalysts appear to have greater stability for the gas-phase acetoxylation of benzene. High yields of phenyl acetate can be obtained from the reaction of benzene, acetic acid and This reaction is of industrial interest oxygen over Pd-Au/Si02 catalysts at 150 "C (equation 168).433 as a more direct route to phenol than the conventional cumene oxidation process.

80%

Acetoxylation of toluene using a Pd(OAc),-Sn(OAc)2-charcoal catalyst selectively produces The active catalyst presumably contains benzyl acetate with high turnover numbers (Pd-Sn bonds. Tin ligands are known to increase the r-acceptor ability of palladium, and may favor the coordination of the toluene in the form of a benzylic r-allyl complex (141) which is nucleophilically attacked by the acetate ani0n.4~~

6

0

CH~OAC

+ AcOH + OSO2

W(OAc),/Sn(O,4c),

+

/OAC

Me

I

AcO-\

+ (AcO)&-PdH

+ AcO-Pd-Sn(OAc),

(170)

61.3.4.4 Oxycarbonylations 61.3.4.4.1 Alkenes and dienes

The reaction of carbon monoxide with alkenes in the presence of palladium catalysts under oxidative conditions has been extensively studied owing to its industrial importance (equation 171).361,436-438 The oxidative carbonylation of ethylene to acrylic acid has been developed into a commercial process by Union The reaction is performed at 150 "C, 77 atm, CzH4/C0= 1, in an acetic acid-acetic anhydride solvent mixture, and in the presence of a Wacker catalyst ( PdC12/CuC1,/LiOAc). Acrylic acid is produced as the main product, together with pacetoxypropionic acid (overall yield = 80% ). The suggested mechanism (equation 172) involves intramolecular insertion of coordinated ethylene into a palladium-carboxylic group obtained by water hydrolysis of the Pd-CO species. Acrylic acid results from &hydride elimination of the carboxypalladation intermediate (142), whereas oxidative cleavage by AcOH results in the formation of P-acetoxypropionic acid.440 CH,=CH,

+ CO + 0.50,

WCl,jCuCI,/LiOAc AcOHjAc,O

C02H

* A

Metal Complexes in Oxidation

369

Bj 1 4 2 \ c O H

=/

COiH

l-7 COIH

AcO

The oxycarbonylation of propene under the same conditions results in the formation of crotonic acid as the major product instead of the more valuable methacrylic When the oxycarbonylation of ethylene or terminal alkenes is carried out in anhydrous alcoholic solvents instead of acetic acid, dialkyl succinates and @alkoxy esters are the major products (equation 173).44'.442 8'CH=CHz

oz/Co/RoH +R'CHCHz PdCI,/CuCI, ' R'CHCH2 I I I I ROZC COZR RO C02R

(173)

Non-oxidative hydrocarboxylation of alkenes to carboxylic acids with CO and H,O is catalyzed by palladium complexes such as PdC12(PhCN)2or PdCl,(PPh,), ,and a-methyl acids predominate in the presence of H C ~ .A ~recent ~ ~improvement , ~ ~ of this reaction consisted of the use of a PdC12/CuC12/HC1 catalyst under oxidative conditions.377Almost quantitative yields of a-methyl carboxylic acids and dicarboxylic acids were obtained from terminal alkenes and terminal dialkenes respectively, at room temperature and atmospheric pressure (equation 174).377 RCH=CHZ+CO+HzO

PdCIJCuClJ HCI

O,,rt,Iatm

RCHMe

I

(174)

COZH

The synthesis of dialkyl hex-3-ene-lY6-dioatefrom the dioxycarboxylation of butadiene in an alcoholic solvent, and in the presence of a dehydrating agent such as trimethyl orthoformatew or 1 ,l-dimethoxy~yclohexane,~~ provides an economically attractive route for the synthesis of adipic acid (equation 175).

61.3.4.4.2

Aromatic compounds

The oxidative carbonylation of arenes to aromatic acids is a useful reaction which can be performed in the presence of Wacker-type palladium catalysts (equation 176). The stoichiometric reaction of Pd(OAc), with various aromatic compounds such as benzene, toluene or anisole at 100 "C in the presence of CO gives aromatic acids in low to fair yields.446This reaction is thought to proceed via CO insertion between a palladium-carbon (arene) v-bond. It can be made catalytic in the presence of excess TBHP and allyl chloride, but substantial amounts of phenol and coupling by-products are f0rrned.4~~ ArH + CO+ 0.50,

ArC0,H

(176)

The catalytic oxycarbonylation of benzene and naphthalene to benzoic or naphthoic acid in but difficulties the presence of Wacker-type catalysts has been reported in several in reoxidizing the reduced palladium have inhibited industrial use of this chemistry.

61.3.4.4.3

Alcohols and phenols

The synthesis of dialkyl oxalates by oxycarbonylation of alcohols in the presence of a dehydrating As for the agent and a Wacker catalyst was first reported by Fenton in 1968 (equation 177).378*449 previous oxycarbonylations, the presence of water is a strong inhibitor of the reaction and favors the side-formation of CO, (equation 178). Dehydrating agents such as triethyl orthoformate or boric anhydride are necessary to prevent water formation and subsequent deactivation of the

Uses in Synthesis and Catalysis

370

catalyst. Improved procedures use Pd(OAc),/Co(OAc),/PPhJbenzoquinone as a catalyst system and enable the 'oxycarbonylation of methanol by CO and dioxygen (yield = 83%, turnover number = 140).4'0 WCI,/CuCI, wc(oEt),

2ROH+2C0+0.502

PdC1, + CO + H,O

RO,CCO,R+ H,O

Pd + CO, + 2HC1

4

(177) (178)

A likely mechanism for this reaction involves the nucleophilic attack of alcohol on two molecules of coordinated CO, followed by the coupling of the two palladium-bonded alkoxycarbonyl moieties In support of this mechanism, the reaction of the bis( methoxycarbonyl) (equation 179).438,451 complex (143) with CO and PPh3 produces dimethyl oxalate and the reduced palladium(0) complex.45'

co

c1 'Pd'

c1/

c1

\co

CO,R]'-

\Pd'

+ 2 R O ~ -+

COzR

I

-+

CO,R

c1

(PPh,),Pd(CO,Me),

+Pdo+2C1-

CO PPh

W(CO)(PPh,),+(CO,Me),

(143)

The synthesis of dimethyl oxalate and ethylene glycol from methanol, CO and H, is currently being industrially developed by UBE and Union Carbide.452Methyl nitrite is readily obtained from the reaction of methanol with nitric oxide in the presence of oxygen (equation 181). The carbonylation of methyl nitrite to dimethyl oxalate is achieved over a palladium/charcoal heterogeneous catalyst at 110-140 "C and low pressure (equation 182), and the resulting nitric oxide is recycled to the formation of methyl nitrite. Hydrogenolysis of dimethyl oxalate over a copper chromite catalyst results in the formation of ethylene glycol (equation 183). The overall yield of ethylene glycol in the combined steps is 90%. This syn-gas route to ethylene glycol might compete in the near future with older ethylene-based oxidation processes. 2MeOH+ 2NO + 0.50,

50 "C

2MeONO + H,O

(181)

90%

90%

The oxycarbonylation of phenol in the presence of a palladium catalyst, a tertiary amine and a manganese cocatalyst at room temperature and atmospheric pressure results in the formation of diphenyl carbonate in good yield (equation 184).453 ZPhOH + co

+

*,

PdBr,i>Wc4 b

KEr,

PhO-C-OPh+H,O

II

(184)

0

61.3.4.5 Miscellaneous Oxidations 61.3.4.5.1 Oxycyanation and oxychlorination

Ethylene stoichiometrically reacts with Pd(CN), in benzonitrile at 150 "C to give acrylonitrile in 55% yield. This reaction can be performed catalytically by reacting ethylene, hydrogen cyanide and oxygen at 350 "Cover PdCl,/V,O,/CsCI supported on alumina. The selectivity in acrylonitrile is about 90%, and the productivity about 20 g/g Pd/hour (equation 185).379 P d i ViLs/A120,

CH,=CH,+HCN

+0.50,

CH,=CHCN+ H 2 0

(185)

The catalytic oxychlorination of ethylene with HC1 and 0, to vinyl chloride can be catalyzed in the liquid phase by a PdCl,-CuC12 mixture, or in the vapor phase over supported palladium

Metal Complexes in Oxidation

37 1

catalysts (equation 146).37u*454 However, these processes are not productive enough to compete with the conventional dehydrochlorination of 1,2-dichloroethane.

61.3.4.5.2

Oxidative coupling of arenes and alkenes

This reaction has been thoroughly and will be considered only briefly here. A variety of aromatic compounds can be oxidatively coupled by reaction with palladium( 11) salts, e.g. PdC12, Pd(OAc),, Pd(TFA), or Pd(CIO,),, at a temperature generally not exceeding 120 "C (equation 186). In the general case, a mixture of all the theoretically possible biaryl isomers is formed, with their proportions depending on the nature of the substituent R. The reactivity of arenes increases with increasing nucleophilic substitution of the aromatic ring. The reaction is thought to proceed via n--arene (144), v-arylpalladium (145) and arylpalladation of a second arene molecule to give (146), followed by P-hydride elimination.455

+ PdX2

2

-

R

+Pdo + 2 H X /

R

R

PdX2 (145)

(144)

PdX (146)

A catalytic arene-arene coupling can occur in the presence of cooxidants such as Cu", Fe"' and heteropolyacids along with molecular oxygen, but this reaction is still not selective enough for industrial use. Oxidative coupling of specific alkenes such as styrene derivatives4" and vinyl acetatee0 to This 1,3-diene derivatives can also be achieved in the presence of palladium coupling essentially occurs 'head to head', i.e. the C-C bond formation involves the least substituted carbon atoms of the double bonds (equation 188).461 R

R

\

R

PdC1,JNaOAcJAcOH

\

IS0 "C

/

Ph/C=CHz

Ph

C=CHCH=CH

/ 'Ph

The alkene arylation reaction has been extensively studied by Moritani and coworkers462and by Heck.46' An interesting application of this chemistry is the synthesis of styrene from the oxidative coupling of benzene and ethylene (equation 189).464 PhH+CH,=CH2+0.5O2

Pd(OAc),/O,(lO atm) 100°C

PhCH=CH,

+H 2 0

(189)

650%/Pd

61.3.4.5.3 Oxidation of alcohols to carbonyl compounds

Palladium compounds are effectivecatalysts for the selective oxidative transformation of primary and secondary alcohols into carbonyl compounds. Although conventional PdC12/CuC1, catalysts can operate at 70-120°C and under O2 pressure (3 atrn)F5 milder procedures using a PdCl,/NaOAc mixture are effective at room temperature and atmospheric 0, pressure (equation

CCCB-M

Uses in Synthesis and Catalysis

372

190).466The mechanism of this transformation presumably involves palladium alkoxide formation followed by P-hydride elimination (equation 191). PdCIJHaOAc 38"c

R1RZCHOH+0.50, R'RZCHOH+PdXz

R'RZC=O+HzO

R'

XPd-0

+

(190)

R'RZC=O+HPdX

'C/

H'

'R2

61.3.5 OXIDATION BY COORDINATED NITRO LIGANDS

It has been shown in the previous sections that only a very few catalytic systems are able to transfer both oxygen atoms of the dioxygen molecule onto two molecules of substrate. Representative examples are, in peroxide chemistry, the rhodium-catalyzed ketonization of alkenes by O,, and, in oxo chemistry, the oxidation of phosphines to phosphine oxides in the presence of molybdenum dithiocarbamate. A recent interesting approach uses the oxidizing properties of coordinated nitro groups towards organic substrates S (equation 192) and the ensuing reoxidation of the reduced nitrosyl group by molecular oxygen to the initial nitro species (equation 193) to form a catalytic system.467 MNO,+S

+

MNO+O.SO,

MNO+SO

(192) (193)

MNO,

3

61.3.5.1 Oxidation by Cobalt-Nitro Complexes

Five-coordinate cobalt(111)-nitrosyl complexes possessing bent geometries react, in the presence .~~~,~ of bases (e.g. L = pyridine), with molecular oxygen to give cobalt(II1)-nitro c o m p l e x e ~ This reaction is generally thought to proceed via the formation of the bimetallic peroxy-nitrate intermediate ( 147).47"The related pyCo(saloph)NO, complex [saloph = N,N'-bis(salicy1idene)o-phenylenediaminato] has been used as a stoichiometric oxidant of PPh3 to PPh30 and MOO(S,CNRz), to MOO,(S,CNR,), . The oxidation of phosphines to phosphine oxides becomes catalytic in the presence of dioxygen (equation 195)!7' Z(sa1en)Co-NO+O,

L(sa1en)Co-N

/

0 0

\ /

\

0.50,.

N-Co(sa1en)L

+

ZL(sa1en)Co-NO,

(194)

PY

Schiff base-cobalt-nitro complexes are too mild as oxidants to react as such with alkenes. However, the addition of Lewis acids (e.g. BF, * Et20, LiPF6) to these complexes activates the nitro ligand and produces a variety of both stoichiometric and catalytic oxidations. Stoichiometric transformations involve the oxidation of sulfides to sulfoxides and 1,3-~yclohexadieneto benzene.&' Alcohols such as benzyl alcohol and cycloheptanol are catalytically transformed into the corresponding carbonyl 2R'R'CHOH

+ 0.502

pyCo(TPP)NO,/BF, 60 T

2R'R2C=0

+ H,O

400-800%/Co

Oxygen-transfer reactions have been shown to occur from cobalt(II1)-nitro complexes to alkenes coordinated to palladium.472Thus ethylene and propene have been oxidized stoichiometrically in quantitative yields to acetaldehyde and acetone respectively, with the concomitant reduction of the nitro- to the nitrosyl-cobalt analog. A catalytic transformation with turnover numbers of 4-12 can be achieved at 70 "C in diglyme. The mechanism shown in Scheme 11 has been suggested.

Metal Complexes in Oxidation

373

X = C1, L = TPP Scheme 11

A somewhat similar catalytic acetoxylation of ethylene to vinyl acetate by 0, has been carried out in acetic acid in the presence of a Pd(OAc),-pyCo(TPP)NO, system.472A stoichiometric epoxidation of alkenes such as 1-octene or propene by cobalt-nitro complexes has been shown to occur in the presence of thallium(II1) benzoate. Oxygen labeling studies showed that the epoxide oxygen atom comes only from the nitro ligand (equation pyCo(TPP)NO,+ RCN=CH,

TI"'

Co(TPP)NO+RCH-CH+py

(197)

'o/

61.3.5.2 Oxidation of Alkenes by Palladium(I1)- and Rhodium(II1)-Nitro Complexes Andrews and coworkers have recently shown that bis(acetonitrile)(chloro)(nitro)palladium(II) (148) stoichiometrically reacts with terminal alkenes to give the corresponding methyl ketone in almost quantitative yields and a reduced [PdCl(NO)], complex. When the same reaction is carried out in the presence o f dioxygen, a slightly catalytic production of 2-alkanone occurs (2-7 turnovers).396Labeling experiments with "0-enriched complex (148) clearly showed that the ketonic oxygen comes from the coordinated nitro group. More interestingly, a palladium- rr-alkene complex (149) and a relatively stable five-membered metallacyclic intermediate (150) were detected during the reaction of (148) with propene.396All these data are consistent with the following mechanism (equation 198; L = MeCN).

It is noteworthy that the metallacycle (150) strongly resembles the pseudocyclic intermediate (84a) postulated in the mechanism of ketonization of terminal alkenes by palladium(I1)-&but 1

peroxide complexes (Scheme 6). On the basis of decomposition studies of platinametallacycles,475 Andrews favored the cleavage of (150) between the Pd-N bond, in order to obtain a 0" Pd-C-C-H dihedral angle which is necessary for the ensuing P-hydride elimination to However, owing to the high strength of the Pd--N bond, which is illustrated by the formation D f the resulting Pd-NO complex (151), plausible alternative mechanisms involving breakdown of the metallacycle (150) by a [C-p, C-cy] hydride shift as in equation ( 5 8 ) , or a [ C - p , O - p ] hydride shift as in equation ( 5 9 ) , should also be considered. The reaction of (148) with norbornene produces a stable, well-characterized bright yellow netallacyclic compound (152) which, upon standing at 25 "C in toluene, slowly decomposes to Tive exo-epoxynorbornene in almost quantitative yield (equation 199).'" The particular decompo;ition, which is related in some aspects to the mechanism of alkene epoxidation by d" metal-peroxo

374

Uses in Synthesis and Catalysis

and -alkylperoxo species, can occur because norbornene has no possibilities for p- hydride elimination, a-allyl complex formation or hydride shift. Norbornene and some of its derivatives can be catalytically epoxidized by O2 in the presence of (148) with 4-6 turnover number^."'^

-

(152)

90%

The cationic rhodium( 111)-nitro complex [ ( M~CN),~UINO,]~+( BF,), also oxidizes ethylene and 1-octene to acetaldehyde and 2-octanone, respectively. However, rro catalytic oxidation takes place in the presence of d i o ~ y g e n . ~ ~ '

61.3.6 CATALYSIS BY FIRST-ROW TRANSITION METAL COMPLEXES (Mn, Fe, Co, Cu) 61.3.6.1 Introduction The oxidation chemistry of first-row transition metals has attracted considerable interest in recent decades owing to its close relationship with important biological processes. However, the mechanism of these oxidative processes still remains poorly understood, mainly because the characterization of reactive intermediates is difficult. Thus, neither peroxide examples of manganese, iron and copper nor their OXO derivatives of medium oxidation state have as yet been isolated and characterized by X-ray crystallography. Oxidation catalysis by Mn, Fe, Co and Cu is mainly homolytic in nature. These metals, which are characterized by one-electron oxidationreduction steps, act in their oxidized form as one-electron oxidants and generally produce radical species from hydrocarbons. The direct interaction of cobalt( 111) acetate or manganese( 111) acetate with hydrocarbons is generally thought to proceed in two ways: (a) electron-transfer reaction (equations 200 and 201) and (b) electrophilic substitution at the metal center (equations 202 and 203).56Bath processes involve one-electron reduction of the metal and concomitant carbon radical formation from the hydrocarbon and depend on both the oxidation-reduction potential of the ~ +1.51 V for Mn'+/Mn2i] and the ionizmetal [ E o(aqueous solution) = 1.82 V for C O ~ + / C Oand ation potential of the hydrocarbon." Electron transfer M(OAc);+ RH'

M(OAc),+ RH

RH' -+ R-+H+

(200

(201

Electrophilic substitution M(OAc),+RH RM(OAc),

+

RM(OAc),+AcOH

(202

--+

R*+M(OAc),

(203

In the presence of dioxygen, the carbon radical R. produced by reactions (201) and (202) ar transformed into alkylperoxy radicals ROO., reacts with Co" or Mn" species to regenerate th Co"' or Mn"' oxidants, and produce primary oxygenated products (alcohol, carbonyl compounds which can be further oxidized to carboxylic acids. This constitutes the basis of several industril processes such as the manganese-catalyzed oxidation of n-alkenes to fatty acids, and the cobal catalyzed oxidation of butane (or naphtha) to acetic acid, cyclohexane to cyclohexanol-on mixture, and methyl aromatic compounds (toluene, xylene) to the corresponding aromatic mom or di-carboxylic The relevance of catalytic oxidations mediated by first-row transition metals to enzymic oxy en; tions has stimulated considerable interest in creating biomimetic catalytic oxidation systems.go,,, A list of the most representative systems is given in Table 7 together with their oxidizing propertie Some of these systems, particularly those using first-row metal porphyrins, are particularly effecti! hydrocarbon oxidants, with a high selectivity, sometimes approaching that observed in enzym oxygenases.

Metal Complexes in Oxidation

37 5

Table 7 Oxidation of Hydrocarbons by Model Systems"

Alkene epoxidation S NS

System fn(TPP)CI/PhIO ln(TPP)CI/O,+ H,/Pt 4n(TPP)CI/ Bu, NBH,/O, tn(TPP)OAc/ NaOCl/%NCI In(TPP)C1/ascorbate/02 'e"/ H,O,(Fenton) 'e(acac),/ H,O,/ MeCN 'e"/ EDTA/ ascorbate/ 0, (Undenfriend) 'e''/ o-mercaptobenzoatel O2 'e''/ hydrazobenzene/H+/O, 'eo/AcOH/py/H,S 'e(TPP)Cl/ PhIO 'e(TPP)Cl/p-CN-DMANO 'e(TPP)CI/CHP :o(TPP)/CHP Abbreviations: p-CN-DMANO nonstereoselective.

+

-

+

-

+ -

Alkane hydroxylation S NS

No

-

t

+

-

-

+

+

+

-

= p-cyano-N,N-dimethylanilineN-oxide;

-

478-480 48 1 482 483 484 485-488 48Y

+

-

-

490-492

No

-

488,493,494

-

49 5 496

-

+

-

-

+

Refs.

-

LOW

i-

Alkene ketonization

NO

-

Low

-

No

+

+

-

-

+ -

+ -

+

+

-

-

-

-

Low

No No

+

-

-

No

-

Low -

+

-

+

+

+

-

Arene hydroxylation NIH sh$t

-

488,497-499

500 501 502

CHP = cumyl hydroperoxide; S = stereoselective; NS =

Manganese Catalysts

i1.3.6.2 i1.3.6.2.1

Oxidation by manganese(1II) salts

Owing to their lower oxidation-reduction potential, manganese(111) salts are weaker oxidants han cobalt(111) compounds. The reactivity of Mn(OAc>, towards hydrocarbons is strongly :nhanced by the presence of strong acids or bromides. Alkenes The stoichiometric oxidation of alkenes by Mn(OAc), in acetic acid at 120 "C affords y-lactones n good yield, via the homolytic addition of carboxymethyl radical to the double bond (equations !04 and 205).504-506 'i)

-09 Me

Ph

Ph

+ AcOH

Mn(OAc),

5

Me

R

heat

Mn(0AcL v .CH,COOH

=

/

a

RCH-CH,

Mn"'

f

RCH-CH,

-Mn"

COiH

0

2 RCH-CH, I /

(205)

0 9 0

Under appropriate conditions, Mn(OAc)3can be used as a free-radical initiator for the homolytic iddition of acetic anhydride to terminal alkenes. Linear or a-branched carboxylic acids can be xoduced in 70-80% yields based on a - a l k e n e ~ . ~ ' ~ In the presence of bromide, the oxidation of alkenes by Mn(OAc), in acetic acid readily occurs it 70-80 "C and produces allylic acetates in good yields. Thus cyclohexene is oxidized to cycloand a-methylstyrene to P-phenylallyl acetate in 70% yield,'09 kexenyl acetate in 83% yield:" vith a mechanism involving allylic hydrogen abstraction by bromine atoms coming from the ixidation of bromide by Mn"' (equation 206).

Uses in Synthesis and Catalysis

376

(ii) Aromatic compounds The oxidation of substituted aromatic hydrocarbons by Mn(OAc), in refluxing acetic acid proceeds by two competing mechanisms: (a) a free radical addition of carboxymethyl ,radical to the nuclear ring, forming 2-arylacetic acids; and (b) a benzylic hydrogen abstraction by Mn"' resulting in the formation of benzyl In the presence of strong acids such as sulfuric, trichloro- or trifluoro-acetic acids, only benzylic oxidation products are readily formed at room temperature (equation 207).512,513

12%

81%

7%

(iii) Alkanes Manganese( HI) acetate is poorly reactive with saturated hydrocarbon^."^ However, oxidation of adamantane by Mn(OAc), in trifluoroacetic acid gives relatively high yields of 1-adamantyl trifluoroacetate, showing a preferential attack at tertiary C-H bonds.515Oxidation of n-alkanes by air in the presence of manganese catalysts constitutes the basis for an industrial process for the manufacture of synthetic fatty acids from n-alkanes of petroleum origin, which has been commercially developed in the Soviet Union.516

61.3.6.2.2 Catalysis by manganese porphyrins

The ability of heme-containing cytochrome P-450enzymes to catalyze the hydroxylation of hydrocarbons and the epoxidation of alkenes by 0, in the presence of hydrogen-donor NADPH cofactor has stimulated the screening of synthetic metalloporphyrins for their oxidizing properties in the presence of various oxygen sources such as O,+coreducing agents, iodosylbenzene or hypochlorites. Manganese-porphyrin complexes have recently been shown to possess many interesting properties as epoxidation and hydroxylation catalysts. (i) Oxidation by dioxygen and coreducing ngents In 1979, Tabushi and coworkers reported that oxygenation of cyclohexene in the presence of Mn(TPP)CI and sodium borohydride unusually produced cyclohexanol as the major product, while the same reaction carried out in the absence of NaBH, afforded the conventional radical autoxidation products. Since cyclohexene oxide was readily transformed into cyclohexanol under the reaction conditions, it was reasonably suggested that the epoxide is the primary product of the reaction (equation 208).518Later, the same authors were able to stop the reaction at the epoxide stage by using Mn(TPP)Cl, 1-methylimidazole (NMI), O2 and (H2+ colloidal platinum) as the coreducing agent (equation 209).,'l Although a substantial amount of water was directly produced from the reaction of 0 2 + H 2 , epoxide was catalytically formed with respect to manganese (65 turnovers) and platinum (300 turnovers). The reactivity order of alkenes, i.e. increasing with their nucleophilic nature, was found to be similar to that observed with a related Mn(TPP)Cl/PhJO catalytic epoxidation system (see below). However, in contrast to this PhIO system, the epoxidation of alkenes with O2 H2 proceeds with retention of configuration (cis-alkenes gave cis-epoxides only). The same catalytic system was also found to hydroxylate alkanes, but at a lower rate. The hydroxylation preferentially occurs at the tertiary C--H bonds, as shown by the formation of 1-adamantanol as the main product from adamantane.481

+

O,/Mn(TPP)Cl/ NaBH,

(208

PhH/ElOH

0 Mn(TPP)CI-hMI +

O2 +

HZ

cc4loidalR

(201)'

' 92%

6%

Metal Complexes in Oxidation

377

A surprisingly different oxidation of alkenes was found to occur in the presence of Mn(TPP)Cl catalyst and [NBuJ BHJ coreducing agent.482Except for cyclohexene (transformed into cyclohexen-3-01 and cyclohexanol), the other alkenes, viz. cyclooctene, styrene and 1-octene, were transformed into the corresponding ketone and the ensuing alcohol hydrogenation product (equation 210).

,?

OH

800%/Mn

2000%/Mn

In the presence of ascorbate coreducing agent and under phase-transfer conditions, manganese porphyrins have recently been shown to activate oxygen by catalyzing epoxidation of alkenes and hydroxylation of alkanes to alkanol-one (ii) Oxidation by iodosylbenzene Iodosylbenzene has been extensively utilized over the last few years for its ability to cleanly transfer 'oxene' oxygen atoms to metals and possibly generate high-valent reactive metal-oxo species. PhIO has been successfullyused instead of O,+ NADPH in conjunction with cytochrome P-450 to hydroxylate alkanes,s1and has found a variety of interesting applications as a two-electron oxidant in the presence of first-row transition metal porphyrins. The oxidative properties of iodosylbenzene in the presence of Mn(TPP)Cl towards hydroxylation of alkanes47*~479 and epoxidation of alkenes47s were simultaneously reported by Hill479and Gr0ves.4~'These oxidations have the characteristics of a stepwise radical reaction. Alkenes are epoxidized with loss of stereochemistry at the double bond, as shown by the formation of a 1.6: 1 mixture of trans- and cis-epoxides with an 88% overall yield (based on PhIO) from the oxidation of cis-stilbene (equation 211).478A major amount of cis-stilbene oxide was found in the presence of Mn(TTP)Cl (TTP'= tetra-o-tolylporphyrin), showing that the product distribution is sensitive to the nature of the porphyrin ring. Epoxidation of norbornene in the presence of [''OJH20 results in the formation of 80% labeled norbornene oxide, and suggests the intermediacy of a reactive high-valent manganese-oxo species.47R

phvph+

Phwph + PhIO CH,CI, Mn(TPP)CI " b,, + 0

Phl

(211)

Oxidation of alkanes (R-H) by PhIO in the presence of Mn(TPP)Cl results in the formation of alcohol and alkyl chloride along with minor amounts of coupling products R-R (equation 212).479,519 Hydroxylation of alkanes preferentially occurs at the more nucleophilic tertiary C-H bonds. Oxidation of norcarane has all the characteristics of a radical reaction.478

% yields/PhlO

26

5

2

0.5

Attempts to isolate the reactive intermediates from the reaction of PhIO with manganeseporphyrin precursors led to the characterization of MnTV-PhIOadducts with the presumed formula (153) [from Mn(TPP)OAc] and (154) [from Mn(TPP)CI]. Reaction of Mn(TPP)N, with PhIO yielded the dinuclear p o x 0 complex (155), which has been characterized by an X-ray crystal structure.52nComplexes (153) and (154) were found to epoxidize alkenes and to hydroxylate alkanes in the same way as in the catalytic reacti0ns.5'~The p-oxo compound (155) oxidizes PPh3to PPh,O but is much less reactive than (153) and (154) toward alkanes.

Uses in Synthesis and Catalysis

378

cL.I-Ph

Ph I/ '0

'0

&

\OAc

,,

1

0

(d) A likely mechanism which is consistent with the experimental data involves MnV-oxo species (156) generated from (153), (154) or (155) as the active intermediate. This intermediate (156) can either homolytically add to double bonds to give a free-rotating free-radical intermediate (157), which decomposes into a cis- + trans-epoxide mixture or abstract hydrogen atoms from alkanes to give a free radical Re with subsequent formation of alcohol in the cage species (158) (Scheme 12).478-480,519 R

X PhI

PhIO

Scheme 12

Epoxidation of alkenes by sodium hypochlorite An interesting epoxidation reaction of synthetic interest has recently been reported by Meunier and The reaction of sodium hypochlorite with styrene catalyzed by Mn(TPP)X (X = C1, OAc) under phase-transfer conditions affords styrene oxide in high yield (equation 213). (iii)

Addition of pyridine bases to the catalytic system caused a considerable increase in the rate and selectivity of the reaction, reaching 80% yield of styrene oxide for 100% styrene conversion (r.t., 30 mia). In the presence of this pyridine-modified system, the reactivity o f alkenes is in the order styrene > trisubstituted > cis-disubstituted > trans-disubstituted > monosubstituted. The epoxidation of alkenes is not stereoselective. In the absence of pyridine, cis-stilbene was converted into a 1.8: 1 fruns:cis epoxide mixture, whereas the cis isomer prevails in the presence of excess pyridine ligands. Neither chlorohydrins nor pyridine N-oxides are involved in this catalytic system. Attempts to isolate the reactive intermediate led to the characterization of a relatively stable

(159a) Oxo radical cation

(159b) Oxo-like

Metal Complexes in Oxidation

379

manganese(1V) complex with the formula Mn(TMP)OCl (159) (TMP = tetramesitylpor hyrin) This for which the oxo radical cation (159a) or the oxo-type structure (159b) was suggested!' complex (159) was found to epoxidize styrene both stoichiometrically and catalytically in the presence of NaOCI. Anchoring Mn(TPP)OAc to a rigid polyisocyanide support has resulted in a considerable increase in the epoxidation rate of cyclohexene by NaOC1. These solid catalysts can be re.covered and retain their activity during several consecutive experiments.522 61.3.6.2.3 Miscellaneous manganese-catalyzed oxidations Catalytic oxidations using manganese complexes and dioxygen or peroxide oxidants are rather rare. Although reversible dioxygen adducts of the superoxo or peroxo type have been detected in the interaction of manganese(I1)-pophyrin or -Schiff base complexes with 0,at low temperature," or from the reaction of MnX2(PR3)(X = C1, Br, I) with oxygen,s23no reactivity towards hydrocarbons has been reported. p-Isophorone can be cleanly oxidized by air to 1,5,5-trimethylcyclohexene-3,6-dionein the presence of Mn(sa1en) and triethylamine (equation 214).524The mechanism suggested involves carbanion (160) formation from the reaction of NEt, with p-isophorone, followed by the formation of an alkyl peroxide-manganese(II1) complex (161) which decomposes to the dienone (162).524756 0

0

aa 6

90%

0

+ LMn"'0OMn"'

0

-LMn"'O,

0

2

H

Mn"'

H

OOMn"'

61.3.6.3 Iron Catalysts 61.3.6.3.1 Oxidation by peroxides

The hydroxylating properties of hydrogen peroxide in the presence of iron(1I) salts have been known since 1894.485This system, known as Fenton's reagent, has been widely used as an effective oxidant of a great variety of organic substrates.486-488,525'-526 A widely accepted mechanism, proposed by Haber and Weiss, involves the formation of hydroxyl radicals from the cleavage of the peroxide bond by iron(I1) ions (equation 216)?27 FeZ++H,Oz

-+

Fe3'+OH-+H0.

(216)

Fenton's reagent hydroxylates aromatic rings, generally with low yields, at the more nucleophilic carbon positions, and gives substantial amounts of biaryl by-products resulting from the dimerization of intermediate radical species (equation 217).49'

CCCS-M*

Uses in Synthesis and Catalysis

380

The use of Fenton's reagent in nonprotic solvents (MeCN, CH2C1,) results in substantial changes in the reactivity and the selectivity of the transformations. Thus alkenes are stoichiometrically epoxidized in a nonstereoselective fashion by H,O,/Fe(acac), in acetonitrile (equation 218).489,528 ph

ph

'

Fe:)=(:H,O,=10:1:180

Ph

Ph

H,O,/Fe(acac),/MeCN

w

+

0

Ph b 0p

:

4

h

96

68%

However, the same catalytic system did not hydroxylate aromatic compounds in a way significantly different from usual aqueous Fenton's reagents?29 Oxidation of cyclohexanol by H,O, in the presence of Fe(C104)? and perchloric acid in acetonitrile was found to give cyclohexanone and the six possible cyclohexanediols in which This unusual cis- 1,3-cyclohexanediol represented 72% of the total amount of diols regioselectivity was explained by the formation of a ferry1 ion species FeOz+ or Fe'"=O by heterolysis of the hydroperoxo compound capable of abstracting the cis hydrogen atom at carbon 3 (equations 219 and 220).530Several Fenton's reagents modified by the addition of ~ a t e ~ h o 1 ~ or quinoness3*have been shown to give increased yields and selectivities in the hydroxylation of aromatic hydrocarbons.

72% regiospecificity H

H'

cy '

Fe"

H-0

9

I c-c

H

I

Hp,

T

H-0

7

H

I

1

c-c

61.3.6.3.2 Oxidation by dioxygen and

II

-

-H,O

H-0

I

/FeV=O H

c-c

I -

H-0

9

I

c-c

Fe:

,H

0

I

(220)

cowducing ugmt

Several biomimetic systems containing iron catalysts and a coreducing agent have been found to hydroxylate hydrocarbons by dioxygen under mild conditions (equation 221).'2,s03,'26,535 The first of these hydroxylating systems was described by Undenfrie~~d~~O and involves Fe"( EDTA) as the catalyst and ascorbic acid as the coreducing agent. This reagent was found to hydroxylate arenes to phenols, and alkanes to Several related systems using other coreducing agents such as diamin~purine,'~~ 2-mercaptobenzoic a c i d y hydrazoben~ene?~'2-mercaptoethan01'~~ and metallic iron p o ~ d e r ' ~have * * ~also ~ ~ been used, and these oxidizing properties are listed in Table 7. The yields in oxidation products are generally proportional to the concentration of the coreducing agent, but are much lower than those expected from reaction (221), since parallel oxidation of the coreducing agent occurs. An accurate comparison of the oxidizing properties of these model systems is difficult to make. However, these systems differ both from cytochrome P-450and from PhIO/iron porphyrin reagents, and have more or less electrophilic properties depending on the nature of the coreducing agent. RH+DH,+O,

% ROH+D+H,O

(221)

(a) Hydroxylation of alkanes preferentially occurs at the more nucleophiIic tertiary C-H positions, but some of these systems using 2-mercaptoethan01~~~ or metallic iron hydroxylate adamantane with preference for the secondary position. However, none of these s stems hydroxylates cis-decalin with retention of configuration at the hydroxylated atom.4 (b) Hydroxylation of aromatic hydrocarbons by these model systems generally occurs at the more nucleophilic ring carbon atoms, but none of these systems gives any significant NIH shift during the hydroxylation. (c) Epoxidation of alkenes does not occur to a significant extent with these model systems, and allylic hydroxylation is the main reaction observed.

P

Metal Complexes in Oxidation

38 1

(d) Oxidative demethylation of anisole occurs with a low kinetic isotope effect, in contrast to cytochrome P-450and PhIO/Fe(TPP)Cl.539 No simple picture has emerged concerning the mechanism of these oxidations. In one rationale, dioxygen is reduced to hydrogen peroxide by the coreducing agent, and hydroxylation is performed by hydroxyl radicals as with Fenton's reagent.491 Further proposals involve superoxide Fe111-O-O',494 hydroperoxide Fe111-00H:95 oxenoid Fe"I-0-071*540 or ferry1 ion Fe02+ (or Fe1V=O)54as the active oxidant. Recent studies by Barton and coworkers concern the hydroxylating properties of the Feo/RCOOH/py-H,0/U2 system which exhibits particular selecSince sulfides are not oxidized to sulfoxides tivity towards secondary C-H bonds of by this catalytic system, it has been assumed that the active species is neither a radical nor 'oxenoid' reagent, but may be a coordinatively unsaturated iron complex capable of activating alkanes by insertion between C-H b0nds.4~~ 61.3.6.3.3 Catalysis by iron porphyrins

Oxidation by iodosylbenzene Epoxidation of alkenes and hydroxylation of alkanes can be achieved under mild conditions with iron porphyrin catalysts and iodosylbenzene as the o ~ i d a n t . ~ ~ ~ - ~ ~ ~ . ~ ~ ~ (1) Epoxidation of alkenes has the following characteristic^.^^^*^^^*^^^ The reactivity of alkenes increases with increasing alkyl substitution at the double bonds in the following order: tetramethylethylene :2-methylbut-2-ene:cyclohexene :oct-1-ene = 204 :76 :20 :I -488,498 High yields of epoxide based on the iodosylbenzene consumed are obtained with electron-rich alkenes. 1,3-Cyclohexadieneis oxidized to the corresponding monoepoxide in 93% yield (equation 222).498 (i)

93%

cis-Alkenes are much more reactive than trans i s ~ r n e r s . ~ Thus ' ~ ~epoxidation ~~~ of cisstilbene or cis-Cmethylpentene in the presence of Fe(TPP)Cl is at least 14 times faster than that of the trans analogs, in contrast to epoxidation by peracids. The relative reactivity of the cis- and trans-alkenes is sensitive to the substitution on the porphyrin periphery, and the reaction becomes more selective for cis-alkenes as the steric bulk of the porphyrin increases. In contrast to the PhIO/Mn(TPP)Cl system, epoxidation of alkenes occurs with high retention of configuration. cis-Alkenes are almost exclusively converted to cis-epoxides. Norbornene is transformed into a 67 :3 mixture of em- and endo-norbornene oxide (e.g. equation 223).498 Ph

Ph

PhlO/Fe(TPP)CI

w

0 71Yn

Asymmetric cpoxidation of prochiral alkenes can be performed by PhIO in the presence of chiral iron porphyrins."' Thus styrene was epoxidized to (R)-(+)-styrene oxide in 67% chemical yield and 48% e.e. in the presence of iodosylmesitylene and FeT(a,P,a,P[a,P,a,P-bindp = 5a,lOP,15 ru,20P-tetrakis(0-[(S)-2'-carboxymethyl- 1,l'binap)PPCl binaphthyl-2-carboxamido]phenyl)porphyrin] (equation 224).541

0" '0 +

FeT(u,B,u,P-binapiPPC~

@ /

+

9 /

(224)

67% yield 48% e.e.

The proposed epoxidation mechanism depicted in Scheme 13 involves oxygen transfer from a reactive iron(V)-oxo intermediate to the alkene. The approach of the alkene to the iron-bound oxygen presumably occurs 'side-on' and parallel to the porphyrin plane, and involves the interaction of partially filled oxygen-iron p v - d v antibonding orbitals.

382

Uses in Synthesis and Catalysis

0-I-Ph

-'i

/

I

1

I-Fh

?'

Scheme 13

(2) Oxidation of alkanes by PhIO in the presence of iron(II1) porphyrins at room temperature results in the formation of alcohol as the major product along with consecutive carbonyl oxidation Hydroxylation of cis-decalin occurs with retention of configurproducts (equation 225).498,488,502 ation at C-9 and mainly produces ~ i s - 9 - d e c a l o l .Oxidation ~ ~ ~ , ~ ~ ~of n-heptane results in the formation of 2-, 3- and 4-heptanols with an overall yield of ca. 20% based on PhIO. The heptanol distribution depends on the nature of the porphyrin ligand, and w - 1-hydroxylation increases as the iron accessibility of the catalyst decreases.502Carbon radical intermediates were revealed by the formation of larger amounts of bromocycloheptane when the hydroxylation of cycloheptane was carried out in the presence of CC1,Br.498 The hydrogen isotope effect for cyclohexane hydroxylation by Fe(TTP)Cl was found to be 12.9* 1. The proposed hydroxylation mechanism involves hydrogen atom abstraction from the alkane by the reactive FeV=O intermediate, and rapid collapse of this radical to give the product alcohol and to regenerate the Fe"' catalyst (equation 2 2 6 1 . ~ ~ ~ OEl

31%

0 II

6 '/o

H 0'

I

X

I

X

I

X

I

X

l+BrCC'3 RBr

(3) In contrast to cytochrome P-450 and iron-based models using O2 and a coreducing agent, iron porphyrins are poor catalysts for the hydroxylation by PhIO of aromatic hydrocarbons at Benzene is not oxidized, and toluene gives mainly benzylic oxidation the ring products, Le. benzyl alcohol and benzaldehyde. However, high NIH shift values (60-70%) were observed in the hydroxylation of [4-2H]anisoleand [ 1,4-2H2]naphthaleneby PhIO in the presence of Fe(TPP)CI and Fe(TFPP)CI [TFPP = tetra(pentafluorophenyl)porphyrin].488~4Y9

(ii) Oxidation by miscellaneous oxidants Tertiary amine oxides can be used instead of iodosylbenzene for the epoxidation of alkenes in the presence of Fe(TPP)CLS0"p-Cyano-N,N-dimethylaniline N-oxide (p-CN-DMANO) has

Metal Complexes in Oxidation

383

oxidizing properties similar to PhlO and provides the same products. The major difference between these two oxidants is that while PhlO is both stereospecific and kinetically selectivefor cis-alkenes, p-CN-DMANO is only stereospecific and also reacts with rrans-alkenes (equation 227).

(227)

Hydroxylation of alkanes can be performed by TBHP or cumyl hydroperoxide in the presence of iron porphyrin catalysts.s0'*s02The characteristics of this reaction are very different from PhlO/Fe(TPP) systems. No epoxidation of alkenes occurs, and the alcohol distribution is completely independent of the nature of the iron porphyrin used (equation 228). A 'Fenton-type' mechanism involving active species not including the metal has been suggested (equation 229).

Cum-OOH

Fe(TPP)CI

+

0Fe(TPP)CI

(TPP)ClFe'"OH+RO.

Cum-OH

+ 20%

40YO

RH -KOW. R.+(TPP)CIFelV-OH

d

R'OH+Fe(TPP)Cl

(229)

The porphyrin-iron( 111)-peroxo complex [Fe(TPP)O,] (163)was prepared by the reaction of KO, with Fe"(TPP) in the presence of a crown ether, and characterized by spectroscopic methods [ v ( 0 - 0 ) = 806 cm-1]542.This peroxo complex (163) was found to be inactive toward hydrocarbons. However, addition of excess acetic anhydride to (163) dissolved in a benzene-cyclohexane mixture results in the formation of cyclohexanol and cyclohexanone. This reaction is thought to which decomposes proceed via acylation of the peroxo group, giving iron percarboxylate (la), to an FeV-oxo compound (165) capable of hydroxylating alkanes.543Such a mechanism has been suggested for the hydroxylation of camphor by Pseudumonns cytochrome P-450.544

/ O=C-Me

(163)

(165)

(164)

Low-spin iron'"-oxo-porphyrin complexes such as (NMeIm)(TPP)Fe'"=O can be obtained from the homolytic dissociation of the dinuclear Fe"'-p-peroxo complex in the presence of bases (B) (equation 231).545However, this iron(1V)-oxo complex was only found to oxidize phosphines to phosphine Hence, although there is strong evidence for the involvement of high-valent reactive iron-oxo species in oxidations by iodosylbenzene and possibly in oxidations by cytochrome P-450, the lack of structural data for these reactive intermediates raises some doubt concerning the nature and the metal valency of these reactive species. The difference in reactivity of iodosylbenzene, on the one hand, and of peroxide or O2+ coreductant, on the other, makes it difficult to propose single reactive intermediates. Furthermore, the hydroxylating and epoxidizing properties of vanadium and chromium peroxides also give some consistency to the involvement of iron peroxide species, or related iron-iodosylbenzene or -hypochlorite adducts as reactive intermediates. There is no doubt that the intensive efforts devoted to this subject will shed some light in the near future on the mechanism of these fascinating oxidation reactions. 2PFe"+0,

61.3.6.3.4

-+

PFe"'-O-O-Fe"'P

23

2PFeIV=0

(231)

MiscelIaaneous oxidations

p,-Oxo-triiron complexes [ Fe,O(RCO,),L,]X [ R = Me, But(piv); L = MeOH, Py;X = Cl, OAc] have recently been shown to catalyze the epoxidation of alkenyl acetate by molecular oxygen

384

Uses in Synthesis and Catalysis

under mild conditions (60 "C, 1 atm) (equation 232).546 The regiospecificity of this reaction resembles that of peracid epoxidation but is opposite to that of Bu'OOH/VO(acac),. Alkenyl acetates were found to be the best substrates for this reagent. Simple alkenes are much less efficiently epoxidized. The reaction is not stereospecific, as shown by the formation of a 5: 1 mixture of trans- and cis-epoxides from the oxidation of cis-p-methylstyrene.One mole of dioxygen was consumed per mole of epoxide produced. This is twice the amount required for the stoichiometry of the reaction, and the detected oxidative degradation products account for one It is noteworthy that a similar p3-oxo-iron oxygen atom of 0,not found in the epoxide cluster Fe,O(AcO),(py), has been isolated from the reaction of iron powder with acetic acid in aqueous pyridine in air, a system which had previously been shown to hydroxylate alkanes.496

(232)

Qo*c 87%

61.3.6.4

Cobalt Catalysts

61.3.6.4.1 Oxidation by cobalt(III) salts

The oxidation of hydrocarbons by cobalt(Il1) acetate has been thoroughly investigated, due to its relevance to industrial homolytic oxidation p r o c e ~ s e s . ~Radical ~ , ~ ~ intermediates '~~~~ are produced from one-electron oxidation of hydrocarbons according to an electron transfer or an electrophilic substitution mechanism previously described in equations (200)-(203). These oxidations are dramatically accelerated by the presence of strong acids or halide salts. Oxidation of n-alkanes by Co"' acetate in acetic acid occurs with a remarkable regioselectivity (rs) at the w-1 position, giving 2-alkyl acetate as the major product in anaerobic conditions, and 2-alkanone in the presence of oxygen (equations 233 and 234).548Cyclohexane is readily oxidized in nitrogen by Co(OAc), in acetic acid to mainly cyclohexyl acetate and Z-acetoxycyclohexanone. s49,550 In the presence of oxy en and a high cobalt concentration, adipic acid is the major product formed (equation 235).'!? Oxidation of adamantane by Co(OAc), and TFA in AcOH preferentially occurs at the tertiary positions, producing 1-adamantyl acetate as the major product

'

I

81% rs

83% rs

Oxidation of alkenes by Co(OAc), preferentially occurs at the allylic positions, yielding 2-aikenyl acetate. Oxidation of cyclohexene by Co(OAc), and TFA in AcOH results in the formation of cyclohexenyl acetate along with minor amounts of the corresponding allyl alcohol.5s0The oxidation of ethylene by Co(TFA), in TFA affordsethylene glycol bis(triflu~roacetate).~~~ Oxidation of alkylbenzene by Co(OAc), mainly occurs at the side-chain benzylic positions, without formation of nuclear adducts. In nitro en, benzylic acetates predominate, whereas aromatic acids are favored when oxygen is pre~ent."~ Thus toluene is transformed into benzyl acetate and benzaldehyde by Co(OAc), in AcOH under anerobic condition^.^'^ When this reaction is

Metal Complexes in Oxidation

385

carried out in the presence of O2 and trichloracetic acid, benzoic acid is almost exclusively formed with relatively high turnover numbers (equation 236).512

85% 1628%/CO

Halide salts also have a pronounced synergistic effect on the cobalt-catalyzed autoxidation of alkyl aromatic compounds to acids, which is of considerable industrial importance (see For example, p-xylene is selectively converted to terephthalic acid by O2 in the presence of Co(OAc), and NaBr at 60 "C and atmospheric pressure. Oxidation of bromide by Co"' produces bromine atoms, which act as chain-transfer agents (equation 237 and 238).56 When cobalt(I1) acetate is used, long induction periods corresponding to the oxidation of Co" to CoI'' are observed. These induction periods can be eliminated by initiators such as TBHP or ozone. The oxidation of Co(OAc), by ozone produces a p,-oxo-bridged trimer CO,O(OAC),(ACOH)~, which is itself a good initiator.554This trimeric complex forms spectrophotometrically characterized charge-transfer complexes with a variety of aromatic hydrocarbons.554This may be of significant importance for the initiation mechanism of CoIII-catalyzed oxidations.554 Co(OAc), t BrBr.+ArMe

ArCH,.

eCo(OAc),Br ArCH,O,-

Co(OA4, b

-

Co(OAc), t Bra

ArCHO+Co(OAc),OH

(237)

% ArCOzH

(238)

61.3.6.4.2 Industrially important cobalt-catalyzed oxidation of hydrocarbons (i) Adipic acid from cyclohexane Adipic acid is a most important petrochemical product which is mostly used for the synthesis of nylon 6.6 from its condensation with hexamethylenediamine. Cyclohexane is transformed to adipic acid in two steps: (a) oxidation of cyclohexane to a cyclohexanol-cyclohexanone mixture (ol-one) via the formation o f cyclohexyl hydroperoxide followed by (b) oxidation of the ol-one mixture to adipic acid by nitric acid (equation 239). OOH

OH

0 II

In the Dupont process, cyclohexane is reacted with air at 150°C and 10 atm pressure in the presence of a soluble cobalt(I1) salt (naphthenate or stearate). The conversion i s limited to 8-10% in order to prevent consecutive oxidation of the ol-one mixture. Nonconverted cyclohexane is recycled to the oxidation reactor. Combined yields of ol-one mixture are 7 0 4 0 % .83,84,555 The ol-one mixture is sent to another oxidation reactor where oxidation by nitric acid is performed at 70-80 "C by nitric acid (45-50% 3 in the presence of a mixture of Cu(NO3>, and NH,VO, catalysts, which increase the selectivity of the reaction. The reaction is complete in a few minutes and adipic acid precipitates from the reaction medium. The adipic acid yield is about 90%. Nitric acid oxidation produces gaseous products, mainly nitric oxides, which are recycled to a nitric acid synthesis unit. Some nitric acid is lost to products such as N2and N20 which are not recovered. In a modified version developed by Scientific Design, the autoxidation of cyclohexane is carried out in the presence of boric acid which forms boric esters with cyclohexanol and cyclohexyl hydroperoxide. Boric esters such as B(OC6H11)3are hydrolyzed to cyclohexanol and cyclohexanone. The resulting boric acid is dehydrated and recycled. This method produces an approximately 10 : 1 01 :one mixture, with an overall yield of 90% for a cyclohexane conversion of 10%. Although this process gives higher yields of ol-one mixture, it requires added investments and operating costs for boric acid recycling."v84 A similar procedure has been developed by Chemische Werke Hulls for the oxidation of cyclododecane to a cyclododecanol-one mixture by air in the presence of boric acid with trace amounts of cobalt(I1) carboxylate at 160-180°C and 1-3 atm.556The ol-one mixture (5: 1 ) is

386

Uses in Synthesis and Catalysis

produced in ca. 80% yield at 33% conversion. Further oxidation of cyclododecanol-one by nitric acid affords dodecanedioic acid, a most important intermediate for the synthesis of polyamides (nylon 12) for several specialty applications. Acetic acid from butane Butane from natural gas is cheap and abundant in the United States, where it is used as an important feedstock for the synthesis of acetic acid. Since acetic acid is the most stable oxidation product from butane, the transformation is carried out at high butane conversions. In the industrial processes (Celanese, Hiils), butane is oxidized by air in an acetic acid solution containing a cobalt catalyst (stearate, naphthenate) at 180-190 "C and 50-70 atm.3613ss7 The AcOH yield is about 4045% for ca. 30% butane conversion, By-products include CO, and formic, propionic and succinic acids, which are vaporized. The other by-products are recycled for acetic acid synthesis. Light naphthas can be used instead of butane as acetic adic feedstock, and are oxidized under similar conditions in Europe where natural gas is less abundant (Distillers and BP processes). Acetic acid can also be obtained with much higher selectivity (95-97%) from the oxidation of acetaldehyde by air at 60°C and atmospheric pressure in an acetic acid solution and in the presence of cobalt acetate.3617s58

(ii)

(iii) Benzoic acid from toluene Benzoic acid is an important chemical intermediate which can also be used as a phenol precursor by decarbonylation in the presence of copper catalysts (Lummus process). It is produced industrially by oxidation of toluene by air in the presence of cobalt catalysts (Dow and Amoco processes; equation 240). The reaction can be carried out without solvent, or in an acetic acid solvent. The oxidation of toluene without solvent uses a cobalt octoate catalyst and operates at higher temperature (180-200 "C).Yields of benzoic acid are about 80% for ca. 50% toluene c o n v e r ~ i o n . ~ ~ ' In an acetic acid solution and in the presence of cobalt acetate, the reaction occurs at lower temperature conditions (110-120 "C) and gives higher yields in benzoic acid (90%).83384

Terephthalic acid from p-xylene This is one of the most important industrial oxidation processes. Terephthalic acid (TPA) is mostly used for the manufacture of polyester fibers, films and plastics, and its world production capacity reaches 8 Mt/year. Two major processes have been developed. The Amoco-Mid Century process produces terephthalic acid by the one-step oxidation of p-xylene in acetic acid, whereas the Dynamit Nobel process yields dimethyl terephthalate in several steps and in the absence of solvent.83~84*86 In the Ammo process, p-xylene is oxidized at 200 "C under 15-20 atm in acetic acid and in the presence of a catalyst consisting of a mixture of cobalt acetate (So/, weight of the solution), manganese acetate (1%) and ammonium bromide. Owing to the highly corrosive nature of the reaction mixture, special titanium reactor vessels are required. One of the main difficulties of this process is to remove the intermediate oxidation products such as p-toluic acid or p-carboxybenzaldehyde which contaminate TPA obtained by precipitation from the reaction medium. A series of recrystallization and solvent extraction apparatus is required to obtain fiber grade TPA with 99.95% purity. The overall yield in TPA is CA. 90% for a 95% conversion of p-xylene, The Dynamit Nobel process produces dimethyl terephthalate (DMT) by a complicated series In the oxidation section, p-xylene is of oxidation and esterification stages (equation 241).83*84,86 oxidized at 150 "C and 6 atm without solvent and in the presence of cobalt octoate to TPA and p-toluic acid. These oxidation products are sent to another reactor for esterification by methanol at 250 "C and 30 atm. Fiber grade DMT is purified by several recrystallizations, and monoesters are recycled to the oxidation reactor. The overall yield in DMT is about 80%, which is lower than in the h o c 0 process. However, this process is competitive because it is not corrosive and requires lower investments. It provides high-quality fiber-grade dimethyl terephthalate. (iv)

Me

Me

C02H

Metal Complexes in Oxidation

387

61.3.6.4.3 Catalysis by cobalt porphyrins and Scbifl base complexes Oxidation of terminal alkenes by dioxygen and a coreducing agent Styrene derivatives can be selectively converted to the corresponding benzyl alcohols by molecular oxygen in the presence of bis(dimethylglyoximato)chloro(pyridine)cobalt( 111) and sodium tetrahydroborate (equation 242).559A likely mechanism for this reaction involves insertion of the alkene into the cobalt-hydride bond, followed by 0, insertion into the cobalt-carbon bond, as in equation (ll), and decomposition of the peroxide adduct (168) to the ketone, which is reduced to alcohol by NaBH4 (equation 243).

(i)

-0 -

O,/Co(DMGH),Clpy/Ns EH,

R

4

NaBH,

L ~ c o ~ + - c7 ~ LnCo-H

2

LnCo-CHMe

I

R

-

H I I R

-LnCo-OH

LnCo-0-0-CMe

OH

R-C-Me !I 0

NaBH I 2 R-C-Me I

H

(168)

A somewhat similar oxidation of terminal alkenes to methyl ketone and alcohol by O2 in the presence ofCo(sa1MDPT) [salMDPT = bis(salicylideneiminopropyl)methylamine] and in ethanol solvent has recently been reported by Drago and coworkers (equation 244).560Only terminal alkenes were found to be reactive with this catalytic system. The reaction is alcohol dependent and occurs in ethanol and methanol but not in f-butyl or isopropyl alcohols. The alcohol is concomitantly oxidized during the reaction, and may act as a coreducing agent and/or favor the formation of cobalt hydride. This oxidation might occur according to the mechanism of equation (243). Co(wlMUPT) +

O2 EtOH.50-70T

'

V + OH Y 0

- 1ooo%/co

(244)

-750%/C0

(ii) Oxidation of phenols Cobalt-Schiff base complexes catalyze the selective oxidation of phenols by dioxygen into quinols (equation 245j61) or quinones (equations 246562*563 and 247561)under mild conditions.

(245)

R

BuQBu' R OH (169)> 90%

?H

Uses in Synthesis and Catalysis

388

Nishinaga and co-workers isolated a series of stable cobalt(111)-alkyl peroxide Complexes such as (170) and (171) in high yields from the reaction of the entacoordinated Cot'-Schiff base complex with the corresponding phenol and O2 in CHzCl2:7,s Complex (170; R=Bu') has been characterized by an X-ray structure. These alkyl peroxide complexes presumably result from the to the phenoxide radical obtained by homolytic addition of the superoxo complex C O " ' - O ~ ~ hydrogen abstraction from the phenolic substrate by the Co"'-superoxo complex. The quinone The product results from p- hydride elimination from the alkyl peroxide complex (172)56'*56*56535M quinol (169) produced by equation (245) has been shown to result from the reduction of the Co"'-alkyl peroxide complex (170) by the solvent alcohol which is transformed into the corresponding carbonyl compound (equation 248).5bi

(iii) Oxidative cleavage reactions Cobalt"-Schiff base complexes, e.g. C~(salen),'~'C o ( a c a ~ e n and ) ~ ~cobalt( ~ 11) porphyrins,569 e.g. Co(TPP), are effective catalysts for the selective oxygenation of 3-substituted indoles to keto amides (equation 249), a reaction which can be considered as a model for the heme-containing enzyme tryptophan-2,3-dioxygenase(equation 21)." This reaction has been shown to proceed via a ternary complex, Co-02-indole, with probable structure (175), which is converted into indolenyl hydroperoxide (176). Decomposition of (176) to the keto amide (174) readily occurs in the Catalytic presence of Co(TPP), presumably via formation of a dioxetane intermediate (177).569956 oxygenolysis of flavonols readily occurs in the presence of Co(sa1en) and involves a loss of one mole of CO (equation 251).57"'. 0

0

Hydroxylation of alkanes Cobalt porphyrins catalyze the hydroxylation of cyclohexane by cumyl or &butyl hydroperoxide under mild conditions (equation 252).'" This reaction presumably occurs outside the coordination (iv)

Metal Complexes in Oxidation

389

sphere of the metal, with the active species being alkoxy radicals resulting from the decomposition of the hydroperoxide by the metal, as previously shown for iron catalysts in equation (229).502

+Cum-OOH

61.3.6.5

66

-

0

Co(TPP)

+ Cum-OH

+

052)

23 %

45 Yo

Copper Catalysts

61.3.6.5.1 Copper dioxygen complexes

The chemistry of copper superoxides, peroxides and oxo compounds is still poorly understood owing to the lack of structural data concerning these unstable species. Dioxygen is generally completely reduced by copper(1) salts. Reaction of copper(1) halides with O2 in the presence of bases affords an active catalyst for the oxidative coupling of phenols; acetylenes and amines by 02.571s572 Complexes with the general formula Cu4XqOZLn [ n = 3 or 4; X = C1, Br; L = py, MeCN, N-methylpyrrolidone (NMP),DMA, DMSO, etc.] have been isolated from the reaction of 4 mol copper(1) halide with I mol O2 in the presence of a base.s72*575,582*s83 Although no X-ray data are available as yet, it is generally understood that these tetrameric complexes involving two p*-oxo oxygen atoms are formed via homolytic dissociation of the p-peroxo dioxygen adduct (179), followed by reaction of the Cu"'=O or Cu"-O- species (180) with the initial copper(1) complex (178).14,572,573

-.

-.

LXCU'

~ L X C U '0, + L X C U ~ ~ ~ - O - O - C ~ ~ ~ X~ L ( L X C+. ~L~ XC ~~L ~~O - O___c .) LXC~~~--O--C~~~XL (178)

( 180)

(179)

(181)

(253)

In the presence of trace amounts of water, the tetrameric p2-oxo complex (182) in 1,2dimethoxyethane is transformed into a OXO tetrameric complex (183; equation 254), characteris inactive towards the oxidation of phenols. ized by an X-ray In contrast, (182)572*575 The reaction of N,NYN',N'-tetramethyl-l,3-propanediamine (TMP) with CuC1, C 0 2and dioxygen results in the quantitative formation of the p-carbonato complex (184; e uation 255).576 This compound acts as an initiator for the oxidative coupling of phenols by 02.' Such p-carbonato complexes, also prepared from the reaction of Cu(3PI)CO with O2 [BPI= 1,3 bis(2-(4-methylpyridyl)imin~)isoindoline],~'~are presumably involved as reactive intermediates in the oxidative carbonylation of methanol to dimethyl carbonate (see Upon reaction with methanol, the tetrameric complex (182; L = Py;X = C1) produces the bis(p-methoxo) complex (185; equation 256), which has been characterized by an X-ray structure,s79and is reactive for the oxidatiye cleavage of pyrocatechol to muconic acid derivatives.580,s81

s,

(CL2-0)2Cu,X4L, (182) CUCl

H2O L=NMP

(fi4-O)L,Cu4Cl,(OHJL (183)

(254

co*

Lxcu-o-c-o-cuxL

(255)

L =TMP,X = C1

I1

0 (184)

R I

R (185)

A polymeric p- hydroperoxocopper(I1) compound has been prepared from the reaction of hydrogen peroxide with an aqueous solution of copper(1I) acetate and has been characterized by spectrophotometric methods and peroxide oxygen determination (equation 257).222 Such copper( 11) peroxide compounds have recently been claimed to be active catalysts €orthe epoxidation of light alkenes by dioxygen.223

Uses in Synthesis and Catalysis

390

61.3.6.5.2

Oxidation of nonfunctionalized hydrocarbons

Catalytic oxidations by copper compounds are mainly homolytic in nature. Copper salts have been extensively used in conjunction with molecular oxygen, peroxides and persulfate for the oxidation of a variety of alkenic and aromatic hydrocarbon^.^^'^^^ Oxidarion of alkenes and alkynes Copper(I) sulfate generated from the reduction of aqueous copper(11) sulfate with metallic copper promotes the epoxidation of alkenes such as allyl alcohol or propene in low yields.585The reaction of t-butyl hydroperoxide or t-butyl peracetate with alkenes in the presence of copper(1) chloride results in the homolytic formation of allylic t-butyl peroxides or allylic esters (equations 258-260). (i)

90YO

+ Bu'OOH + AcOH

4-J--.-

+ BU'OOAC

-6

+ Bu'OH + H20

CUCl

'"'

(259)

89%

+B ~ ~ O H OAc 88%

Oxidation of cyclohexene by peroxydisulfate in the presence of copper(I1) salts results in the formation of cyclopentanecarboxaldehyde as the main product in an aqueous acetonitrile solution (equation 26f), and 2-cyclohexenyl acetate in an acetic acid solution (equation 262).5s8,5*9 Reaction (261) has been interpreted as the formation of a radical cation (186) by oxidation of cyclohexene with S,Oi-, followed by hydrolysis of (186) to the ,f3-hydroxyalkyl radical (1871, which is oxidized by copper( 11) salts to the rearranged aldehydic product (188; equation 263).s89

a yo

(186)

(187)

(188)

Monosubstituted acetylenic compounds can be oxidatively dimerized by air at room temperature in the presence of copper salts in a pyridine-methanol solution. This method has been applied to a wide variety of acetylenic compounds and gives high yields in disubstituted diacetylenic compounds (equation 264)."'

Metal Complexes in Oxidation

391

Oxidation of arenes The reaction of hydrogen peroxide with copper(1) salts produces a Fenton-like hydroxylating system involving reactive hydroxyl radical intermediates (equation 265).486,49’Hydroxylation of benzene to phenol can be achieved by air in the presence of copper(1) salts in an acidic aqueous s o l ~ t i o n . ~This ~ * -reaction ~ ~ ~ is not catalytic (phenol yields are ca. 8% based on copper(1) salts) and stops when all copper(1) has been oxidized to copper(I1). A catalytic transformation of benzene to phenol can occur when copper(I1) is electrolytically reduced to copper(1) (equation 266).594,595 (ii)

Cu++HzOz

Cu2++OH-+HO’

(265)

‘e. -

Anthracene can be selectively oxidized to anthraquinone by molecular oxygen in the presence of copper(I1) bromide in an ethylene glycol solution. Ethylene glycol (EG) acts as a bidendate ligand for copper and prevents the formation of bromoanthracene as a by-product (equation 267).596

97%

Nuclears9’ or s i d e - ~ h a i nacetoxylation ~ ~ ~ , ~ ~ ~ of arenes can be performed with good yields by persulfate and copper(I1) salts in acetic acid (equations 268 and 269). As previously shown for cyclohexene (equation 263), persulfate oxidizes the aromatic ring to a radical cation which loses a proton to give a carbon radical, which is further oxidized by copper(I1) acetate to the final acetoxylated product.

(268)

61.3.6.5.3 Oxidation of phenols Copper-catalyzed oxidations of phenols by dioxygen have attracted considerable interest owing to their relevance to enzymic tyrosinases (which transform phenols into o-quinones; equation 24) and laccases (which dimerize or polymerize diphenol~),~’ and owing to their importance for the synthesis of specialty polymers [poly(phenylene oxides)15* and fine chemicals (p-benzoquinones, muconic acid). A wide variety of oxidative transformations of phenols can be accomplished in the presence of copper complexes, depending on the reaction conditions, the phenol substituents and the copper catalysLS6 (i) Oxidation of monophends Copper(1) salts in the presence of a tertiary amine catalyze the oxidation of 2,6-disubstituted phenols by O2 into poly(pheny1ene oxides) (high arnine/copper ratio) or biphenyl-4,4’-quinanes (low amine/copper ratio). When R is a bulky substituent (e-g. Bu‘), biphenyl-4,4’-quinone is the sole product (equations 270 and 271).599

Uses in Synthesis and Catalysis

392

r n

G

O

H

+ 50,

R

Under different reaction conditions, phenols can be oxidized to p-quinones (equations 272600-602 and 273603),but in the case of phenol itself, insufficient selectivity has prevented, as yet, the commercial application of this potentially important synthesis of p-benzoquinone and hydroquinone. The selectivity of p-benzoquinone, or p-quinol formation can be increased at the expense of oxidative coupling products by using a large excess of the copper reagent [Cu4C1402(MeCN), The suggested mechanism or CuCl+Oo, in MeCN] with respect to the phenolic sub~trate.6'~ involves the oxidation of the phenoxide radical (189) by a copper( 11)-hydroxo species to p-quinol (190) which can rearran e (for R 2 = H ) to hydroquinone (191; Scheme 14), which is readily oxidizable to p-quinone. 6%

CuCl/MeCN/0,(70 bar)60G CuCI/MeOH/O,( 100 bar)"'

C = 93%, S = 80% C = 89%, S = 60%

Ortho hydroxylation of phenols can be accomplished by molecular oxygen in the presence of copper( I) chloride and metallic copper in acetonitrile. This particular ortho selectivity, which is similar to that of enzymic tyrosinases, has been attributed to the formation of a stable copper(I1) catecholate resulting from the reaction of copper(I) phenates with dioxygen (equation 274).605

Metal Complexes in Oxidation

393

70-90%

(ii) Oxidative cleavage of catechols Tsuji and coworkers reported that copper(1) chloride in the presence of pyridine, methanol and dioxygen promotes the stoichiometric oxidation of pyrocatechol to methyl muconate.6°6 Labeling 1 8 0 2 studies have shown that only one atom of the dioxygen molecule is incorporated in the substrate, while the other one is transformed into water as in enzymic monooxygenases (equation 275)607(and not as in dioxygenases, viz. pyrocatechase). This reaction has been shown by Rogic et al. to proceed via two steps (equation 276).5R0,5s' (a) The first step is the oxidation of pyrocatechol to o-quinone by the 'cupric reagent', which has the presumed formula [Cu"Cl(OMe)py], related to complex (185; equation 256).579 This reaction presumably involves formation of a dicopper( 11) catecholate intermediate in which electron transfer from the aromatic ring to two Cu" centers provides obenzoquinone and two copper(1) species, reoxidized by O2 to copper( (b) The second step is the oxidative cleavage of the o-benzoquinone to methyl muconate induced by the 'cupric reagent'.5s' Such an oxidative cleavage reaction of 9,lO-phenanthrenequinone by Cu4C14py302(182) has been shown to occur stoichiometrically in argon (equation 277).'"

-

(Cu2+ClpyOMe),

go

+ CU4C1402PY3

P

(192)

C02CuClpy, Y

1 8CO,CuClpy2

f

2CUClPY3

/

61.3.6.5.4 Oxidation of alcohols

Copper(1) chloride in the presence of pyridine derivatives catalyzes the oxidation of alcohols to the corresponding carbonyl compounds by molecular oxygen.60s The best catalytic system involves 1,lO-phenanthroline as the basic ligand and is well suited for the oxidation of benzylic or allylic alcohols (equation 278 and 279). Cycloalkanofs are poorly reactive, and linear primary alcohols give a large amount of aldehydic oxidation products having at least one less carbon atom than the initial substrate. Diacetone alcohol can be oxidatively dehydrogenated to the corresponding a-keto epoxide by dioxygen in the presence of CuCl and pyridine (equation 280).'09 PhCH,OH+&

PhCH=CHCH,OH+fO,

CuCl/phen

PhCHO+H,O

CuCl/phen

86%

PhCH=CHCHO+ H,O 83%

Me2C--CH-CMe

I

I

II

HO H

0

+

CuCl/py

* Me2C--CH-CMe \0/

8

C=28%,S=95%

394

Uses in Synthesis and Catalysis

61.3.6.55 Oxidation of nitrogen compounds

(i) Oxidation of methylene to carbonyl groups Copper salts catalyze the selective transformation of activated methylene groups into carbonyl groups by dioxygen. These reactions (281)-(283) can be considered as valid models for enzymic internal monooxygenases (equation 22), since " 0 labeling studies have shown that the carbonyl oxygen atom of (194) comes from molecular oxygen.61° Oxidation of bis( 1-methylbenzimidazol-2y1)methane (193; equation 281)610and bis(2-pyridy1)methane (195; equation 282)61'results in the formation of the carbonyl compounds (194) and (196), respectively, in almost quantitative yields. Copper(I1) salts also catalyze the selective transformation of trimethylamine into dimethylformamide (equation 283).612

(282) (1%) > 90%

(195)

MezNMe

CuCIJDMF f 01

llo'c

+ H20

Me2N-C-H

II

C

0 40%, S

(283)

99%

(ii) Oxidative cleavage of activated double bonds Copper complexes are particularly effective catalysts for the oxidative cleavage of enamines under extremely mild condi(equation 284) 613,615 and 3-substituted indoles (equation 285)616*617 tions (-0 "C). Me

\

,c=c

/

H

CUCI,/MecCN

+0

\

2

osc

+ H-C-N A

' Me-C-Me II

0 87%

(284)

0 87%

..

H

70%

(iii) Miscellaneous oxidations Several specific oxidative transformations of nitrogen compounds can be carried out in the presence of copper salts. Oxidation of o-phenylenediamine with molecular oxygen in the presence of a twofold excess of CuCl in pyridine results in the formation of ciqcis-mucononitrile in high yield (equation 286).6'8-619The bis-pox0 tetranuclear complex Cu4C1402( py)., was found to be the active species in this transformation.'18 A similar procedure can be used for the selective oxidative coupling of diphenylamine to tetraphenyihydrazine by CuCI/py/ 0, or Cu4Cl,02py4 (equation 287).619

96% Ph Ph

-' Ph

\

2

/

NH+:O,

CUCl/P>

Ph

'N-N'

Ph

+H,O

'Ph 83 %

Metal Complexes in Oxidation

395

Copper(I1) acetate in methanol catalyzes the oxidation in air of acid hydrazides to carboxylic acids or esters (equation 288),"2' and that of 1,2-dihydrazones to acetylenes (equation 289).618 Ph-C-NHNHZ

II

+0 2

Cu(OAc),/MdH

PhCO,H+N2+HZO

0

95 %

PhC -CPh

II

H,NN

II

+0 2

Cu(OAc),/MeOH

NNH,

PhCZCPh +2N2+2H20 78%

61.3.6.5.6 Oxycarbotiylution of alcohols and amines

It has long been known that copper(1) salts in complexing solvents absorb carbon monoxide under mild conditions,62' and several copper(1)-carbonyl complexes having terminal CO (e-g. [Cu(dien)CO]BPh, , dien = diethylenetriamines22; CuHB(C3N2H3)C0, HB(C3N,H3), = hydrotris( 1-pyra~olyl)borato~~~) or bridging CO (e.g. [Cu,(tmen),( p CO)(p PhCO,)]BPh,624) have been characterized by an X-ray crystal structure. As previously shown in equation (255), p-carbonato-copper(I1) complexes can be generated from the reaction of copper(l1)-p-oxo complexes with C 0 2 ,or from the reaction of copper(1) CO complexes with 0 2These . pcarbonato complexes are probably the precursors for the synthesis of dimethyl carbonate by oxycarboxylation of methanol. Dimethyl carbonate is an interesting material which can be used instead of toxic dimethyl sulfate as a multipurpose alkylating reagent.4"B.578 Its synthesis can be performed in one step from cheap methanol, CO and oxygen materials in the presence of copper salts (e.g. copper(1l) methoxychloride or CuCl/py) at ca. 100 "C and 15-70 atm (equation 290).578*625 This reaction is thought to proceed in two steps:"' (a) formation of copper(I1) methoxychloride from the reaction of copper(1) chloride, O2 and methanol (equation 291) and (b) reduction of copper(I1) methoxychloride with CO to form dimethyl carbonate and regenerzte copper( I) chloride (equation 292).626 2MeOH+C0+40,

MeO-C-OMe+H,O

(I

0

C-10-20%, S>90"/0

~ C U C I + ~ M ~ O H ++$ ~ZCuClOMe+H,O 2CuClOMe + CO

--*

(MeO),C=O+ 2CuC1

(291) (292)

A related oxycarbonylation of secondary amines to N,N-disubstituted ureas in the presence of copper(1) salts at room temperature and atmospheric pressure has been disclosed by B r a ~ k r n a n . ~ , ~ ~ The reaction is particularly effective with cyclic secondary amines such as piperidine and morpholine (equation 2931, with primary and aliphatic secondary amines being much less reactive.626a (293)

61.3.7 CONCLUDING REMARKS In general, the catalytic properties of transition metals for the oxidation of hydrocarbons are strongly governed by the existence and nature of the metal-oxygen intermediates. Among the various possible reactive intermediates, two families of metal-oxygen species emerge as the most important ones, ie. metal peroxides and metal-oxo compounds. In metal peroxide chemistry, the heterolytic or homolytic nature of catalytic oxidation seems to be strongly dependent on the heterolytic or homolytic dissociation mode of the peroxide intermediate, for which the triangular coordination mode of the peroxide moiety of the metal appears to be a key feature. Heterolytic oxidations require attainable coordination sites on the metal, involve strained metallacyclic reaction intermediates, and are highly selective. In contrast, homolytic oxidations involve bimolecular radical processes with no metal-substrate interactions and are less selective. In the important field of palladiurn oxidation chemistry, hydroperoxo

396

Uses in Synthesis and Catalysis

species probably play an important role in the reoxidation of palladium hydride by air, which is a key step for achieving the catalytic oxidation cycle. Although stoichiometric oxidations by high-valent d o oxo compounds have been firmly established, a recent approach using first-row transition metal porphyrins and various oxidants (0,+ coreductant, peroxides, PhIO, NaOC1) has raised the possibility that high-valent metal-oxo species (e.g. Crv=O, Mn"=O, Fev=O) could act as the key intermediate oxidants for the epoxidation of alkenes and the hydroxylation of alkanes. Although these oxidations are homolytic in nature, they can be controlled for better stereoselectivity and enantioselectivity by changing the substituents on the porphyrin periphery. The oxidizing properties of iron- and cobaIt-p3-oxo species M,O(RCO,),L, and of copper(I1)-p'-oxo complexes Cu4X40JL3also seem to be of particular interest and can be considered as possible precursors for reactive first-row metal-oxo species. Except for a few cases (e.g. the rhodium-catalyzed ketonization of terminal alkenes), the direct and selective oxidation of hydrocarbons by molecular oxygen without the need for a coreducing agent has proved to be difficult to achieve owing to the different reactivities of the first peroxide oxygen atom and of the second oxo or hydroxo oxygen atom of the coordinated dioxygen molecule. New approaches using metal-oxo oxidants which can be regenerated by air, or nitro-metal complexes, are interesting alternatives which deserve to be pursued. Intensive efforts by many groups of researchers, including organic and inorganic chemists and biochemists, in the area of oxidation chemistry should ensure that this goal will probably be achieved in the near future so as to provide a larger number of well-characterized oxidizing reagents and new selective catalytic transformations.

61.3.8 ADDENDUM

and a number of significant new Since this chapter was written, several review papers have appeared. They are briefly noted here in relation to the most relevant sections of the text.

Section 613.2 Metal Peroxides in Oxidation Novel vanadium(V)-alkyl peroxide complexes with the general formula VO(OOBut)(ROPhsalR') ( 197; ROPhsalR' = tridendate N-(2-oxidophenyl)salicylidenaminatoSchiff base ligand) were found to epoxidize alkenes stoichiometrically with high selectivity.h'"

The reactivity of complexes (197) is highly relevant to the Halcon epoxidation process (see Section 61.3.2.2.1). The reaction proceeds at room temperature with high yields, and i s completely stereoselective (&-alkenes are transformed into &-epoxides only). The reactivity of alkenes increases with their nucleophilic nature, as for MoO,(HMPA) complexes. The reaction is strongly inhibited by water, alcohols and basic ligands, and is accelerated in polar nondonor solvents (C2H,CI, , PhNO,). The Michaelis-Menten epoxidatim rate V = k,K [alkene][ complex]/ (1 + K [alkene]) shows that the alkene reversibly coordinates to the metal, forming a metal-alkene adduct which decomposes in the rate determining step !n the epoxide and the metal t-butoxy complex. The proposed mechanism (Scheme 15) involves the coordination of the alkene to vanadium(V), displacing the weakly bonded OBu' oxygen atom, followed by the insertion of the coordinated alkene into the V-0 bond, forming the pseudocyclic metalladioxetane (199) which decomposes by a [1,3] dipolar cycloreversion mechanism to the epoxide and the vanadium(VI- a-butoxy complex (ZOO). The reaction becomes catalytic in the presence of excess Bu~OOH."~~

Meta7 Complexes in Oxidation

397

I

k2

But

Scheme 15

(199)

The heterolytic reactivity of complex (197) contrasts with the homolytic reactivity of the vanadium(V) (dipicolinato)(alkyl peroxide) (22), and is presumably due to the lower binding ability of the OBu' oxygen atom in (197), shown by NMR, which allows the coordination of alkene to the metal. In the Halcon epoxidation process, the reaction of the zirconium(1V) methyl trialkoxide (201) with O2 yields the epoxy alkoxide (203), via intramolecular epoxidation of the coordinated allyl alcohol by the incipient methyl peroxide complex (202).630

t- Butylperoxy complexes of bis(pentamethylcyclopentadieny1) hafnium( IV) [(C5Me5)2Hf(OOBu')R;204; R = Me, Et, Ph] have been prepared from the reaction of Bu'OOH with (C,Me,)Hf(H)( R), and thermally decompose to give the mixed alkoxide (C,Me,)Hf(OBu*)(OR). The X-ray crystal structure of (204; R = Ph) indicates amonodendate t-butylperoxy Titanium(1V)-porphyrin [TiO(TPP)] as well as molybdenum(V)-porphyrin [MoO(OMe)(TPP)] complexes were found to be active for the catalytic epoxidation of alkenes by alkyl hydroperoxides, whereas their peroxo derivatives are inactive.632Iron(II1)-porphyrin-peroxo complexes such as Fe(TPP)O;NMe: did not react with hydrocarbons, but form sulfato complexes upon reaction with S02.632 Manganese(II1)-porphyrin-peroxo complexes Mn(02)(TPP)-K' were recently characterized by X-ray ~rystallography.~~~ Cobalt( 111)-alkyl peroxide complexes with the formula Co(BPI)(OOR)(OCOR') (205; BPI = 1,3-bis(2'-pyridylimino)isoindoline;R = But, CMe2Ph;R' = Me, Ph, But) have been prepared from the oxidation of Co"(BPI)( OCOR') complexes by alkyl hydroperoxides. The X-ray crystal structure of complex (205); (R=Bu', R'=Ph) revealed a distorted octahedral environment, with a monodendate OOBut group and a bidendate carboxylate.635

(W Saturated hydrocarbons are homolytically oxidized by complexes (205) into alcohols, ketones and t-butyl peroxide products. The hydroxylation reaction occurs at the more nucleophilic C-H

398

Uses in Synthesis and Catalysis

bonds, with extensive epimerization at the hydroxylated carbon atoms. The following mechanism has been suggested (equation 295).

*

LnCo"'-O

LnCo"'-OOBu'

% L~CO"'-OH+R

LnCo"+ROH

+

(295)

Ln= (BPI)(OCOR)

These complexes (205) were found to be good models for the reactive intermediates involved in the catalytic decomposition of alkyl hydroperoxides (Haber Weiss mechanism), and in the catalytic hydroxylation of hydrocarbons by ROOH. The platinum complex (diphoe)R(CF,)(OH) is an effective catalyst for the selective epoxidation of terminal alkenes by dilute H,O, under mild conditions (20°C). The reaction is thought to proceed via external attack of the HOO- anion on the coordinated alkene (equation 296).636

+

(diphoe)Pt(CF3)(OHJ

H20z

20T,N,.THF

+ H20

'- o

(296)

99 %

Section 61.3.3 Metal-Oxo Complexes in Oxidation In the presence of excess bipyridyl, ruthenium trichloride is an effective catalyst for the selective epoxidation of alkenes by sodium periodate (equation 297). The epoxidation is syn and stereospecific for cis and trans alkenes.637 Ph

%Ph

RuCI,/bqy/HalO, CH,Cl,-H,O, 5 "C,15 h

,

(297)

Ph 83%

Ruthenium-oxo species (RuO,)'- and (RuO,)- oxidize primary alcohols to carboxylic acids and secondary alcohols to ketones. New species (RuO,Cl,)( PPh,) and RuO,(bipy)Cl, cleanly oxidize a wide range of alcohols to aldehydes and ketones without attack of double bonds.638

Section 61.3.4 Palladium-catalyzed Oxidations

The synthetic applications of the palladium-catalyzed oxidation of alkenes to ketones have in the Wacker palladium-catalyzed ketonization of recently been r e ~ i e w e d . 6Improvements ~~ terminal alkenes have been obtained using phase-transfer catalysis,64' polyethylene or phosphomolybdovanadic Allylic acetoxylation of cyclohexene can be selectively effected by palladium acetate in the presence of MnO, and p-benzoquinone (bq) (equation 298).640

9S%

Palladium-catalyzed oxidation of 1,3-dienes in the presence of LiCl and LiOAc produces l-acetoxy-4-chloro-2-alkenes with high selectivity. The reaction is stereospecific and cyclic dienes give an overall cis [1,4] addition (equation 299).644

e

Pd(OAc),iLiCI/ LiOAc

AcOHfBQ.25'C

c1m

O

A

(299)

c

81% ( E / Z = 9 / 1 )

Carbamates have been prepared in good yields from the reaction of amines, alcohols, CO and oxygen in the presence of palladium or rhodium catalysts and iodide (equation 300).645 0 NH2

+ EtOH + CO + 0.502

It

Pd black/KI.

,60"C,80atm

+

PhNH-C-OEt 93%

+ HzO

Metal Complexes in Oxidation

399

The oxidative carbonylation of aromatic hydrocarbons can be effected under mild conditions (rat., 1 atm CO) in CF3C02H in the presence of an equimolecular mixture of Pd(OAc), and Hg(OAc), (equatibn 301).646

1 33% / Pd

Section 61.3.5 Oxidation by Coordinated Nitro Ligands This subject has recently been reviewed.647Several additional papers have appeared on the catalytic oxidation of alkenes by O2 in the presence of PdCl(MeCN),NO,( 148).648Terminal alkenes and trans- cyclooctene yield the corresponding ketones, cyclopentene and cyclohexene the corresponding allyl alcohol, and bicyclic alkenes the corresponding epoxide. Heterometallacyclopentanes such as (152) have been isolated from the reaction of (148) with norbornene (dicyclopentadiene), and characterized by X-ray Glycol monoacetates were obtained from the reaction of (148) with terminal alkenes in acetic acid.649

Section 61.3.6 Catalysis by First-row Transition Metal Complexes (Mn, Fe, Co, Cu) The epoxidation of alkenes by sodium hypochlorite in the presence of manganese porphyrins under phase-transfer conditions has been thoroughly Kinetic studies of this reaction revealed a Michaelis-Mentera rate equation.652As in Scheme 12, the active oxidant is thought to be a high-valent manganese(V)-oxo-porphyrin complex which reversibly interacts with the alkene to form a metal oxo-alkene intermediate which decomposes in the rate determining step to the epoxide and the reduced Mn"' porphyrin. Shape selective epoxidation is achieved when the sterically hindered complex Mn(TMP)Cl is used as the catalyst in the hypochlorite oxidation.652 Other oxygen sources, such as potassium hydrogen sulfate,653ROOH,65402+ascorbatefi5'or NaBH, ,656 and iod~sylbenzene,'~~ have also been used in cpnjunction with manganese porphyrins for the epoxidation of alkenes and the hydroxylation of alkanes. In the presence of imidazole, manganese porphyrins catalyze the epoxidation of alkenes, including the terminal ones, by hydrogen peroxide.658Nonene epoxide was obtained in 90% yield using 5 equivalents of H202 at 20°C in MeCN in the presence of catalytic amounts of Mn(TDCPP)Cl and imidazole The catalyst was not destroyed at the end of the (TDCPP = tetra-2,6-di~hlorophenylporphyrin).6~~ reaction. In relation to enzymic cytochrome P-450oxidations, catalysis by iron porphyrins has inspired many recent studie~."~ The use of C6F510as oxidant and Fe(TDCPP)Cl as catalyst has resulted in a major improvement in both the yields and the turnover numbers of the epoxidation of alkenes,659The Michaelis-Menten kinetic rate, the higher reactivity of alkyl-substituted alkenes compared to that of aryl-substituted alkenes, and the strong inhibition by norbornene in competitive epoxidations suggested that the mechanism shown in Scheme 13 is heterolytic and presumably involves the reversible formation of a four-rnembered FeV-oxametallacyclobutaneintermediate."* on Picket-fence porphyrin (TPiVPP)FeCl-imidazole, 0, and [H2+ colloidal F't supp;c$d poly(vinylpyrroIidone)] act as an artificial P-450 system in the epoxidation of alkenes. Cu2+ions, e.g. Cu(NO&, catalyze the epoxication of alkenes by iod~sylbenzene.&~ Oxidation of alcohols to aldehydes can be effected by 0, in the presence of CUI' ions and Tempo (2,2,6,6-tetramethylpiperidinyl l - ~ x i d e )Arene . ~ ~ ~hydroxylation of the binucleating ligand (206)

1

I+

400

Uses in Synthesis and Catalysis

by 0, is catalyzed by copper(1) salts, presumably via the intermediate formation of a p-peroxo CU~~-OO--CU” species. Complex (208) has been characterized by means of an X-ray crystal structure:& 61.3.9 REFERENCES 1. 2. 3. 4. 5. 6.

0. Hayaishi, in ‘Oxygenases’, ed. 0. Hayaishi, Academic, New York, 1976, p. 1. L. Vaska, Acc. Chem. Res,, 1976, 9, 175. J. S. Valentine, Chem Rev., 1973, 73, 235. E. Bayer and P. Stretzman, Struct. Bonding (Berlin), 1967, 2, 181. J. A. McGinetti, MTP, Int. Rev. Sci., Znorg. Chem., 1972, 5, 229. V. Y. Choy and C. J. OConnor, Coord. Chem. Rev., 1972, 9, 145. 7. G. Wenrici Olive and S . Olive, Angew. Chem., Int. Ed. Engl, 1974, 13, 29. 8. A. V. Savitskii and V. I. Nelyobin, Russ. Chem. Kev. (Engl. Transl.), 1975, 44,110. 9. R. W. Erskine and B. 0. Field, Struct. Bonding (Berlin), 1976, 28, 1. 10. R. D. Jones, D. A. Sommerville and F. Basolo, Chem Reu., 1979, 79, 139. 11. R. Boca, Coord Chem. Rev., 1983, 50, 1. 12. H. Mimoun, in ‘Chemical and Physical Aspects of Catalytic Oxidation’, ed. CNRS, Paris, 1980, p. 19. 13. H. Mimoun, Reu. Inst. Fr. Pet., 1978, 33, 259. 14. E. I. Ochiai, Inorg. NucL Chem. Lett., 1974, 10, 45. 15. G. A. Rodley and W. T. Robinson, Nature (London), 1972, 235,438. 16. J. P. Collman, R. R Gagne, C. A. Reed, W. T. Robinson and G. A. Rodley, hoc. Natl. Acad. Sci. USA, 1974,71,1326. 17. J. P. Collman, T. R. Halpert and K. S. Suslick, in ‘Metal Ion Activation of Dioxygen’, ed. T. G. Spiro, Wiley, New York, 1980, p. 1. 18. M. 1. Coon and R. E. White, in ref. 17, p. 73. 19. J. A. OConnor and E. A. V. Ebsworth. Ado. Inora. Chem. Radiochem.. 1964.6. , . 279. ed. S. Patai, Wiley, New York, 1983, p. 463. 20. H. Mimoun, in ‘The Chemistry of Functional Gro;ps--‘Peroxides’, 21. M. Zehnder and U. Thewalt, Z.Anorg. Allg. Chem., 1980, 461, 53. Fischer and R. Weiss, J. A m Chem. Soc., 1983, 105, 3101. 22. H. Mimoun, L. Saussine, E. Daire, M. Postel, .I. 23. A. Bkouche-Waksman, C. Bois, J. Sala-Pala and J. E. Guerchais, J. Organornet. Chem., 1980, 195, 307. 24. R. Stomberg, A r k Kemi, 1964, 22, 29. 25. J. M. Le Carpentier, A. Mitschler and R. Weiss, Acta Crysrallogr., Sect. B, 1972, 28, 1288. 26. J. Chatt, J. Chem. SOC.,1953, 2939. 27. L. Vaska, Acc. Chem. Res., 1968, 1, 335. 28. L. Vaska, L. C. Chen and W. H. Miller, J. Am. Chem. SOC.,1971,93, 667. 29. M. Laing, M. J. Nolte and E. Singleton, J. Chem Soc., Chem Commun., 1975, 660. 30. M. S. Weininger, I. F. Taylor and E. L. Amma, Chem. Commun., 1971, 1172. 31. M. Matsumoto and N. Nakatsu, Acta Crystullogr. Secc B, 1975,31, 2711. 32. T. Yoshida, K. Tatsumi, M. Matsumoto, K. Nakatsu, A. Nakamura, T. Fueno and S. Otsuka, Nouv. J. Chim., 1979, 3, 761. 33. T. Kashigawi, Id. Yasuoka, N. Kasai, M. Kakudo, S. Takahashi and N. Hagihara, Chem. Cornmun, 1969,743. 34. B. Bosnich, W. G. Jackson, S. T. D. Lo and J. W. McLaren, Inorg. Chem., 1974, 13, 2605. 35. R. Guilard, J. M. Latour, C. Lecomte, J. C. Marchon, J. Protas and 0. Ripoll, Inorg. Chem., 1978, 17, 1228. 36. M. Nakajima, J. M. Latour and J. C. Marchon, J. Chem. Soc., Chem Commun., 1977, 763. 37. M. Bennett and R. Donaldson, J. Am. Chem. SOC, 1971, 93, 3307. 38. L. H. Vogt, H. M. Faigenbaum and S. E. Wiberley, Chem. Rev., 1963, 63, 269. 39. S. Bhaduri, P. R. Raithby, C. I. Zuccaro, M. B. Hursthouse, L. Casella and R. Ugo, J. Chem SOC.,Chem. Commun., 1978, 991. 40. F. Sakurai, H. Suzuki, Y. Moro-Oka and T. Ikawa, J. Am. Chem. SOC.,1980, 102, 1749. 41. H. Suzuki, K. Mizutani, Y. Moro-Oka and T. Ikawa, J. Am Chem. SOC.,1979, 101, 748. 42. H. Mimoun, R. Charpentier, A. Mitschler, J. Fischer and R. Weiss, J. Am. Chem. SOC.,1980, 102, i047. 43. C. Gianotti, C. Fontaine, A. Chiaroni and C. Riche, J. Organornet. Chem., 1976, 113, 57; 1972.38, 167. 44. Y. Tatsuno and S . Otsuka, J. Am. Chem. Soc., 1981, 103, 5832, 45. A. Nishinaga, H. Tomita, K. Nishizawa, T. Matsuura, S. Ooi and K.Hirotsu, J. Chem, Soc., Dfllron Trans., 198I, 1504. 46. H. Mimoun, P. Chaurnette, M. Mignard, L. Saussine, J. Fischer and R. Weiss, Nouv. J. Chim., 1983, 7,467. 47. G . Strukul, R. Ros and R. A. Michelin, Inorg. Chem., 1982, 21, 495. 48. J. H. Bayston and M. E. Winfield, J. Catal., 1964, 3, 123. 49. S. Muto, H. Ogara and Y. Kamiya, Chem. Lett., 1975, 809. 50. W. P. Griffith, Coord. Chem. Reo., 1970,5,459. 51. B. Spivack and Z. Dori, Coord. Chem. Reo., 1975, 17, 99. 52. K. Saito and V. Sasaki, Ado. Inorg. Bioinorg. Mech., 1, 1982, 179. 53. D. B. Dadybujor, S. S. Jewur and E. Ruckenstein, Cafal. Reo. Sci. Eng., 1979, 19, 293. 54. J. T. Groves, in ref. 17, p. 125. 55. J. T. Groves, A d a Inorg. Biochem., 1, 1979. 56. R. A. Sheldon and J. K. Kochi, ‘Metal-Catalyzed Oxidations of Organic compounds’, Academic, New York, 1981. 57. K. B. Sharpless and T. R. Verhoeven, Aldrichimica Acta, 1979, 12, 63. 58. K. Bloch and 0. Hoyaishi (eds.), ‘Biological and Chemical aspects of Oxygenases’, Maruzen, Tokyo, 1966. 59. 0. Hayaishi, Annv. Rev. Biochem., 1969, 38, 21. 60. 0. Hayaishi (ed.), ‘Molecular Mechanisms of Oxygen Activation’, Academic, New York, 1974. 61. T. E. King, H. S . Mason and M. Morrisson (eds.) ‘Oxidases and Related Redox Systems’, Wiley, New York, 1965, vols. I and 11.

Metal Complexes in Oxidation 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75.

40 1

P. Hartter, Methoden Org. Chem. (Houben-Weyl), 1981,964. L. Que, Jr., Stmct. Bonding (Berlin), 1980, 40, 39. M. Nozaki, in ref: 60, p. 135. J. M. Wood, in ref. 17, p. 163. P. Feigelson and F. 0. Brady, in ref. 60, p. 87. W. H. Vanneste and A. Zuberbiihler, in ref. 60, p. 371. M. E. Winkler, K. Lerch and E. I. Salomon, J. Am. Chem Sot., 1981, 103, 7001. I. C. Gunsalus, J. R. M e e k , J. D. Lipscomb, P. Debrunner and E. Munck, in ref. 60, p. 559. S. Onenius and L. Emster, in ref. 60, p. 215. V. Ullrich, Angew. Chem., Int. Ed. Engt., 1972, 11, 701. V. Ullrich, Top. Curr. Chem., 1979, 83, 68. R. E. White and M. J. Coon, Annu. Reu. Biochem., 1980,49, 315. D. M. Jerina, Chem. Technot., 1973, 120 and refs. therein. S. D. Nelson, M. R. Boyd and J. R. Mitchell, in ‘Drug Metaboiism Concepts’, ed. D. M. Jerina, American Chemical Society, Washington DC,1977, p. 155. 76. W. Levin, A. W. Wood, A. Y.H.Lu, D. Ryan, S. West, D. R. Thakker, H. Yagi and D. M.Jerina, in ref. 75, p. 99. 77. D. Mansuy, Rev. Biol. ToxicoL, 1981, 3, 283. 78. J. T. Groves, S. Krishnan, G. E. Avaria and T. J. Nemo, Adv. Chem. Ser., 1980, 191, 277. 79. J. T. Groves, G. A. McClusky, R. E. White and M. J. Coon, Biochern. Biophys. Res. Commun., 1978, 81, 154. 80. E. Daire, Ph.D. Thesis, Universitd Paris VI, 1983; E. Daire, H. Mimoun and L. Sapssine, N o w . J. Chim., 1984, 8,271. 81. F. Lichtenberger, W. Nastainczyk and V. Ullrich, Biochem. Biophys. Res. Commun., 1976, 70, 939. 82. R. E. White, S. G. Sligar and M. J. Coon, J. Biol. Chem., 1980, 255, 11 108. 83. P. Leprince, A. Chauvel and J. P. Catry, ‘Proctdts de P&rochimie’, Technip, Paris 1971. 84. K. Weissermel and H. J. Arpe, ‘Chimie Organique Industrielle’. Masson, Paris, 1981. 85. ‘Riegel’s Handbook of Industrial Chemistry’, 7th edn., ed. J. A. Kent, Van Nostrand Reinhold, New York, 1974. 86. L. F. Hatch and S. Matar, ‘From Hydrocarbons to Petrochemicals Hydrocarbon Processing’, Gulf Publishing Co, Houston, TX, 1981. 87. Catalytica Associates Inc., ‘Selective Oxidation of Hydrocarbons: a Critical Analysis’. Multiclient Study no. 1077, 1979, Santa Clara, CA. 88. Hucknall, D. J., ‘Selective Oxidation of Hydrocarbons’, Academic, New York, 1974. 89. G. T. Austin, Chem Eng. (NY), 1974,81 (1) 127; (2) 125; (3) 87; (4) 86, 143; (5) 101; (6) 149; (7) 107; (8) 96. 90. ‘Ulmanns Encyklopadie der Technischen Chemie’, Verlag. Cbemie, Weinheim, 1972-1983. 91. ‘Kirk-Othmer Encyclopedia of Chemical Technology’, Wiley, New York, 1978-1983. 92. J. E. Lyons, Hydrocarbon Process., 1980 (11) 107. 93. H. W. Prengle and N. Barona, Hydrocarbon Process., 1970, 49 (3), 106. 94. W. R. Patterson, in ‘Catalysis in Chemical Processes’, ed. R. Pearce and W. R. Patterson, Wiley, New York, 1981, p. 251. 95. R. Landau and A. Saffer, Chem. Eng. Prog, 1968, 64 (10) 20. 96. Y. Kamiya, Adv. Chem Ser., 1968, 76, 192. 97. J. W. Parshall, J. MOL Catal., 1968, 4, 243. 98. R. Jira, W. Blay and G. Grimm, Hydrocarbon Process., 1976, 55 (3), 97. 99. R. Landau, G. A. Sullivan and D. Brown, Chemtech, 1979, 602. 100. R. A. Sheldon, J. MOL Cutal., 1980, 7, 107. 101. C. F. Cullis and D. J. Hucknall, Catalysis, 1982, 5, 273. 102. P. A. Kilty and W. M. H. Sachtler, Catal. Rev., 1974, 10, 1. 103. D. L. King, K. K. Ushiba and T. E. White, Hydrocarbon Process, 1982, 61 (Il), 131. 104. A. Aquilo, J. S. Alder, D. N. Freeman and R. J. H. Voorhove, Hydrocarbon Process., 1983, 62 (3), 57. 105. G. Luft and M. Schrod, J. Mol. Catal., 1983, 20, 175. 106. H. Mimoun, M. Postel, F. Casabianca, A. Mischler and J. Fischer, Inorg. Chem., 1982, 21, 1303. 107. J. C. Marchon, J. M. Latour and C. J. Boreham, J. Mol. CataL, 1980, 7 , 227. 108. J. Sala-Pala, J. Roue and J. E. Guerchais, J. Mol. CataL, I980,7, 141. 109. H. Mimoun, Isr. L Chem., 1983, 23,451. 110. G. Mathern and R.Weiss, Acta Crystaltogr., Sect. B, 1971, 27, 1572. 111. G. W. F. Fleet and W.Little, Tetrahedron Lett., 1977, 42, 3749. 112. R. Stomberg, Acta Chem. Stand., 1963, 17, 1563. 113. D. Westlake, R. Kergoat and J. E. Guerchais, C.R. Hebd. Seances Aead. Sci., Ser. C, 1975, 280, 113. 114. S. E. Jacobson, R Tang and F. Mares, Inorg. Chem., 1978, 17, 3055. 115. H. Mimoun, I. Seree De Roch and L. Sajus, Bull. Soc Chim. Fr., 1969, 1481. 116. H. Mimoun, I. Seree De Roch and L. Sajus, Tetrahedron, 1970, 26, 37. 117. K. B. Sharpless, J. M. Townsend and D. R. Williams, J. Am Chem. SOC.,1972,94, 295. 118. H. Arakawa, Y . Moro-Oka and A. Nozaki, Bull. Chem Soc Jpn., 1974,47, 2958. 118a. F. Di Furia and G. Modena, f i r e Appt. Chem., 1982, 54, 1853. 119. W. Winter, C. Mark and V. Schurig, Inorg. Chem., 1980, 19, 2045. 120. H. B. Kagan, H. Mimoun, C. Mark and V. Schurig, Angew. Chem, In$. Ed. Engl., 1979,6, 485. 121. P. Chaumette, H. Mimoun, L. Saussine, J. Fischer and A. Mitschler, J. Organomet. Chem., 1983, 250, 291. 122. B. Chewier, Th. Diebold and R. Weiss, Inorg. Chim. Act4 1976, 19, L57. 123. (a) H. Mimoun, unpublished results; (b) G. Amato, 0. Bortolini, F. Di Furia et al., J. Mol. Catal, 1986, 37, 165. 124. F. W. B. Einstein and P. R.Penfold, Acta Crystallogr., 1964, 98, 5473. 125. S. E. Jacobson, D. A. Muccigrosso and F. Mares, J. Org. Chem, 1979, 44,921. 126. B. W. Graham, K. R. Laing, C . J. O’Connor and W. R. Roper, Chem. Commun.,1970, 1272; 1968, 1556, 1558. 127. B. W. Graham, K. R L i n g , C. J. O’Connor and W. R. Roper, J. Chem Sac., Dalton Trans., 1972, 1237. 128. B. E. Cavit, K. R. Grundry and W. R. Roper, J. Chem. Soc., Chem Commun., 1972, 60. 129. V. L. Johnson and J. F. Geldard, Inorg. Chem., 1978, 17, 1675.

402

Uses in Synthesis and Catalysis

130. L. M.Haines, Inorg, Cbem., 1971, 10, 1685. 131. F. Igersheim and H. Mimoun, Noun J. Chim., 1980, 4, 161; L Cbem. SOC.,Cbem. Commun., 1978, 559. 132. A. Nakamura, Y. Tatsuno and S. Otsuka, Inorg. Cbem, 1972, 11, 2058. 133. H. Mimoun, Pure Appl. Cbem., 1981, 53, 2389. 134. M. J. Nolte, E. Singleton and M. Laing, J. Am. Chem. SOC., 1975, 97, 6396. 135. J. A. McGinetti, R J. Doedens and J. A. Ibers, Inorg. Cbem, 1967, 6, 2243. 136. J. Valentine, D. Valentine and J. P. Collman, Inorg. Cbem., 1971, 10, 219. 137. R. W. Horn, E. Weissberger and J. P. Collman, Inorg. Cbem, 1970, 9, 2367. 138. W. B. Beaulieu, G. D. Mercer and D. M. Roundhill, 1. Am. Cbem. SOC.,1978, 100, 1147. 139. M. Matsumoto and N. Nakatsu, Acta Crystallogr., Sect. B, 1975, 31, 2711. 140. S. Otsuka, A. Nakamura, Y. Tatsuno and M. Niki, J. Am. Chem. Soc., 1972, 94, 3761. 141. G. Wilke, H. Schott and P. Heirnhach, Angew Chem., 1967,79, 62. 142. C. J. Nyman, C. E. Wymore and G. Wilkinson, J. Chem SOC.( A ) , 1968, 561. 143. R. A. Sheldon and J. A. Van Doom, J. Organornet. Cbern., 1975,94, 115. 144. F. Igersheim and H. Mimoun, Nouu. J. Cbim., 1980,4, 711. 145. T. Kashigawi N. Yasuoka, N. Kasai, M. Kakudo, S. Takahashi and N. Hagihara, Cbern. Commun., 1969, 743. 146. A. Sen and J. Halpern, J. Am. Chem. SOC.,1977,99, 8337 and refs. therein. 147. R. Ugo, G. M. Zanderighi, A. Fusi and D. Carreri, J. Am. Chem. SOC.,1980, 102, 3745 and refs. therein. 148. M. J. Broadhurst, J. M. Brown and R. A. John, Angew. Cbem, Int. Ed. Engl., 1983, 22, 47. 149. G. Read, M. Urgelles, A. M. R. Galas and M. B. Hursthouse, J. Chem. SOC.,Dalton Trans., 1983, 911. 150. C. D. Cook and G. S. Jauhal, J. Am. Chem. Soc., 1967,89,3066. 151. J. J. Levison and S. D. Robinson, J. Am. Cbem. Soc., 1971, 93, 617. 152. G. Gordon and H. Taube, J. Inorg. Nucl. Chem, 1961, 16, 268. 153. G. A. Olah and J. Welch, J. Org. Cbem., 1978, 43, 2830. 154. A. D. Westland and M. T. H. Tarafder, Inorg. Chem, 1981, 20, 3992. 155. A. D. Westland and N. T. H. Tarafder, Inorg. Cbem., 1982, 21, 3228. 156. H.Tomioka, K. Takai, K. Oshima and H. Nozdki, Terrahedron Lett., 1980, 21, 4843. 157. (a) K. F. hrcell, 1. Organomet. Cbem., 1983, 252, 181; Organometalks, 1985, 4, 509; (b) K. A. Jorgensen and R. Hoffman, Acta Cbem. Scand., Ser. B, 1986, 40, 411. 158. M. Tsutsui and A. Courtney, Adu. Organornet. Cbem., 1977, 16,241. 159. R. H.Grubbs and A. Miyashita, in ‘Fundamental Research in Homogeneous Catalysis’, ed. M. Tsutsui, Plenum, New York, 1979, vol. 3, p. 151. 160. P. Pitchen, Ph.D. Thesis, Universitt Paris-Sud, Centre d’Orsay, France, 1983. 161. K. Hintzer, Ph.D. Thesis, Tubingen University, FRG, 1983. 162. H. Mimoun, J. Mol. Catal., 1980, 7, 1. 163. E. Vedejs, D. A. Engle and J. E. Tolschow, J. Org.Cbem, 1978,43, 188. 164. S. E. Jacobson, R. Tang and F. Mares, J. Cbern. SOC.,Cbem. Commun., 1978, 888. 165. S. L. Regen and G. M. Whitesides, J. Organomet. Chem., 1973, 59, 293. 166. G. Schmitt and 3. Olbertz, J. Organomet. Cbem., 1978, 152, 271. 167. M. M. Midland and S. B. Preston, J. Org. Chem., 1980, 45, 4514. 168. S. A. M a t h , P. G. Sammes and R. M. Upton, J. Chem SOC.,Perkin Trans 1, 1979, 2481. 169. G . A. Brewer and E. Sinn, Inorg. Cbem., 1981, 20, 1x23. 170. M. R. Galobardes and H. W. Pinnick, Tetrahedron Lett., 1981, 52, 5235. 171. (a) G. B. Payne and P. H. Williams, J. Org. Chem., 1959, 24, 54. (b) K. S. Kirshenbaum and K. B. Sharpless, I. 0%.Chem., 1985, 50, 1979. 172. (a) H. C. Stevens and A. J. Kaman, J. Am. Cbem. SOC., 1965, 87, 734. (b) B. M. Trost and Y. Masuyama, Isr. J. Chem., 1984, 24, 134. 173. J. P. Schinnann and S. Y. Delavarenne, ‘Hydrogen Peroxide in Organic Chemistry’, Edition et documentation industrielle, Paris, 1979. 174. Z. Raciezewski, J. Am. Cbem. SOC.,1960,82, 1267. 175. G. J. Carlson, J. R. Skinner, C. W. Smith and C. H. Wilcoxen (Shell), US Pat. 2 833 787 (1958) (Chem. Abstr., 1958, 52, 16367). 176. M. Pralus, J. C. Lecoq and J. P. Shirman, in ref. 159, p. 327. 177. C. Venturello, E. Alneri and M. Ricci, J. Org. Cbem, 1983, 411, 3831. 178. P. A. Grieco, Y. Yokogama, S. Gilman and M, Nishizawa, J. Org. Cbem., 1977,42,2034. 179. H. J. Reich, F. Chow and S. L. Peake, Synthesis, 1978, 299. 180. C. H. McMuHen (Union Carbide), US Par. 3 993 673 (1976). 181. S. E. Jacobson, F. Mares and P. M. Zambri, . I Am Chem Soc., 1979, 101, 6946. 182. W. Treibs, Angew. Cbem., 1964, 76, 990. 183. N. A. Milas, J. Am. Cbem. SOC.,1937, SO, 2342. 184. J. E. Lyons, in ‘Aspects of Homogeneous Catalysis’, ed. R. Ugo, D. Reidel, Dordrecht, 1977, vol. 3, chap. 1. 185. R. Ugo, F. Conti, R Mason and G. Robertson, Cbem. Commun., 1968, 1498. 186.. P. J. Hayward and C. J. Nyman, J. Am. Chem S N , 1971, 93, 617. 187. G. Read, M. Urgelles, A. M. R. Galas and M. B. Hursthouse, J. Chem. SOC.,Dalton Trans., 1983, 911. 188. R. Ugo, Engelbard I d . , Tech. Bull., 1971, 11, (2), 45. 189. H. C. Clark, A. B. Goel and C. S. Wong, J. Organornet. Cbem., 1978, 152, C45. 190. H.C. Clark, A. B. Goel and C. S. Wong, J. Am Cbern. SN, 1975, 100, 6241. 191. M. J. Y.Chen and J. K. Kochi, J. Cbem. Soc,, Cbem Commun., 1977, 204. 192. H. Kwart and D. M. Hoffman, J. Org. Cbem., 1966, 31, 419. 193. H. Mimoun, Angew Chem., Tnt. Ed. Engl., 1982, 2, 734. 194. H. S. Ryang and C. S. Foote, J. Am. Chem. SOC.,1980, 102, 2119. 195. K. Maeda, I. Montani, T. Hosokawa and S. I. Murahashi, Tetrahedron Lett., 1974, 797. 186. A. Fusi, R. Ugo. A. Pasini and S. Cenini, J. Organornet. Cbem, 1971, 26, 417.

~

~

Metal Complexes in Oxidation

403

197. J. E. Lyons and J. 0. Turner, J. Organomet. Chem., 1972,37, 2881; Tetrahedron Lett., 1972, 2903. 198. K. Takao, Y. Fujiwara, T. Imanaka and S. Teranishi, Bull. Chem SOC.Jpn, 1970,43, 1153; 1973,46, 3810; 1970,43, 3898. 199. C. Dudley and G. Read, Tetrahedron Lett., 1972, 5273; J. Chem SOC.,Dalton Trans., 1977, 883; J. Mol. Caial, 1980, 7 , 31. 200. D. Holland and D. J. Milner, J. Chem. SOC.,Dalton Trans., 1975, 2440. 201. J. Farrar, D. Holland and D. J. Milner, J. Chem. Soc, Dalton Trans., 1975, 815. 202. B. R. James, Ado. Chem Ser., 1980, 191, 253 and refs. therein. 203. H. Bonnemann, W. Nunez and D. M. Rohe, Helu. Chim. Acta, 1983,66, 177. 204. H. Mimoun, M. M. Perez-Machirant and I. Seree de Roch, J. Am. Chem. SOC.,1978, 100, 5437. 205. (a) E. D. Nyberg, D. C. Pribich and R. S. Drago, J. Am. Chem SOL, 1983,105,3538; 1985,107,2598; (b) J. Dehlmann and E. Huft, Oxid. Commun., 1983, 5, 391, 405; (c) M. Faraj, 3. Martin and C. Martin, J. Organornet. Chem., 1984, 276, C23; J. Mol. Catal., 1985, 31, 57. 206. M. M. Perez-Machirant, Ph.D. Thesis, Universiti Paris VI, 197'7. 207. E. D. Nyberg, Ph.D. Thesis, University of Illinois, 1982. 208. J. Jeffery, R. J. Maowby, M.B. Hursthouse and N. P. C. Walker, J. Chem. SOC.,Chem. Commun., 1982, 1411. 209, J. E. Backvall, B. Akermark and S. 0. Ljungreen, J. Am Chem. SOC., 1979, 101, 2411. 210. M. A. Andrews and C. W. F.Cheng, J. Am. Chem. SOC,1982, 104, 4268. 211. 0. Bortolini, F. Di Furia, G. Modena and R. Seraglia, J. Mol Cutul, 1984, 22, 313. 212. L. Carlton, G. Read and M. Urgelles, J. Chem. SOC., Chem. Commun., 1983, 586. 213. J. Martin, C. Martin, M. Faraj and J. M. Bregeault, Nouu. J. Chim., 1984,8, 141. 214. (a) H. Van Gaal, H. G. A. M. Cuppers and A. Van der Ent, Chem. Commun., 1970,1594. (b) J. M. Brown, R. A. John and A. R. Lucy, J. Organornet. Chem, 1985, 279, 245. 215, J. M. Le Carpentier, A. Mitschler and R. Weiss, Acta Crysfallogr., Sect. B,1972, 28, 1288. 216. H. L. Roberts and W. R. Symes, J. Chem. SOC.( A ) , 1968, 1450. 217. H. Suzuki, S. Matsuura, Y. MOrO-Okd and T. Ikawa, Chem Lett., 1982, 1011; J. Organomet. Chem., 1985, 286, 247. Chem. Commun., 1972, 1074; 1 Chem. SOC.,Dalton 218. B. L. Booth, R. N. Hazeldine and G. R. H. Neuss, J. Chem. SOC., Trans., 1982, 37, 511. 219. G. Strukul. R. A. Michelin. .I.D. Orbell and L. Randaccio. Inow. Chrm.. 1983., 22., 3706. 220. J. M. Bregeault and H. Mimoun, Nouu. J. Chim., 1981, 5, 287. -' 221. G. Ferguson, M. Parvez, P. K. Monaghan and R. J. Puddephatt, J. Chem Soc., Chem. Commun., 1983, 267. 222. E. I. Ochiai, Inorg. NucL Chem. Lett., 1973, 9, 987. 223. D. Y. Waddan and D. Williams (ICI), US Pat. 4 390 738 (1983). 224. R. Hiatt, in 'Oxidation', ed. R. L. Augustine and D. J. Trekker, Dekker, New York, 1971, vol. 2, chap. 3. 225. G. A. Tolstilov, V.'P. Yur'ev and U. M. Dzhemilev, Russ. Chem Rev. (Engl. Transl.), 1975, 44,319. 226. R. A. Sheldon, in ref. 184, 1981, vol. 4, p. 3. 227. R. A. Sheldon and J. K. Kochi, Adu. Catal., 1976, 25, 272. 228. J. Sobczak and J. J. Ziolkovski, J. Mol. Catal., 1981, 13, 11. 229. A. 0.Chong and K. 8. Sharpless, J. Org. Chem., 1977, 42, 1587. 230. R. Curci, F. Di Furia and C. Modena, in ref. 159, 1978, vol. 2, p. 255. 231. R. E. Drew and F. W. B. Einstein, lnorg. Chem., 1972, 11, 1079. 232. B. Nuber and J. Weiss, Acta Crystallogr., Sect. B, 1981, 37, 947. 233. I. V. Spirina, V. N. Alyasov, V. N. Glushakova, N. A. Skorodumova, V. P. Sergeeva, N. N. Balakshina, V. P. Malennikov, Y.A. Aleksandrov and G. A. Razuvaev, Rum J. Org. Chem (Engl. Trans.), 1982, 18, 1570. 234. N. Indictor and W. F. Brill, J. Org. Chem, 1965, 30,2074. 235. C. C. Su, J. W. Reed and E. S. Gould, Inorg. Chem., 1973, 12, 337. 236. F. Trifiro, P. Forzatti, S. Preite and I. Pasquon, J. Less-Common Met, 1974, 36,319. 237. L. Saussine, H. Mimoun, k MitschIer and J. Fischer, Nouu. I. Chim, 1980, 4, 235. 238. K. Wieghardt, W. Holzbach and J. Weiss, Inorg. Chem., 1981, 20, 3436. 239. K. Wieghardt, M. Hahn, I. Weiss and W. S. Swiridoff, Z. Anorg. Allg. Chem., 1982,492, 164. 240. R. A. Sheldon, Recl. Trav. Chim. Pays-Bas, 1973,92, 253. 241. V. P. Yurev, 1. A. Gailyunas, L. V. Spirikhin and G. A. Tolstikov, J. G e n Chem. USSR (Engl. Transl.), 1975,45,2269. 242. T. Itoh, K. Jitsukawa, K. Kaneda and S. Teranishi, J. A n Chem. SOC, 1979, 101, 159. 243. T. Katsuki and K. 3. Sharpless, J. Am. Chem. Soc., 1980, 102, 5974. 244. (a) K. B. Sharpless, S. S. Woodard and M. C.Finn, Pure Appl. Chem., 1983,55, 1823; J. Am. Chem. SOC.,1984, 106, 6430; (b) A. F'fenninger, Synthesis, 1986, 89. 245. R. A. Sheldon and J. A. Van Doorn, J. CutaL, 1973, 31, 427, 246. R, A. Sheldon and J. A. Van Doorn, J, Catal,, 1974, 34, 242. 247. P. F. Wolf, J. E. McReon and D. W. Connel, J. Org. Chem., 1975, 40, 1875. 248. J. C. Brunie and N. Grenne, Fr. Pat. 1 477 267 (1966) (Rhone-Poulenc). 249. K. Takai, K. Oshima and H. Nozaki, Tetrahedron Lett., 1980,21,1657. 250. G . A. Razuvaev, L. P. Stepovik and V. A. Dodonov, Zh. Obshch. Khim., 1969,39, 1595. 251. A. Fusi, R. Ugo and G. M. Zanderighi, J. Catal., 1974, 34, 175. 252. H. Arzoumanian, A. Blanc, V. Hartig and J. Metzger, Tetrahedron Lett, 1974, 1011. 253. J. E. Lyons, Tetrahedron Lett., 1974, 2737; Adv. Chem. Ser., 1974, 132, 64. 254. M. N. Sheng and J. G. Zajacek, J. Org. Chem., 1968,33, 588; Org. Synth., 1970, 50, 56. 255. G. N. Koshel, M. I. Farberov, L. L. Zalygin and G. A. Krushinskaya, J. Appl. Chem. USSR (Engl. Transl.), 1971, 44,885. 256. 0. Bortolini, F. Di Fnria and G. Modena, J. Mol. Catal., 1982, 14, 53, 63. 257. P. Pitchen and H. Kagan, Tetrahedron Lett, 1984, 25, 1049; J. Am. Chem. Soc., 1984, 106, 8188. 258. A. J. Bloodworth and I. M. Griffin, 1. Chem. SOC.,Perkin Trans. 1, 1975, 195. 259. J. Halfpenny and R. W. H. Small, J. Chem. SOC.,Chem. Commun, 1979, 879. 260. H. Mimoun and R. Charpentier (Institut Eransais du Wtrole), Pr. Pal. 79 008 828 (1979).

CCC6-N

404

Uses in Synthesis and Catalysis

T. Hosokawa, Y.lrnada and S. Murahashi, J. Chem. SOC, Chem. Commun.,1983, 1245. M. Roussel and H. Mimoun, J. Org. Chem., 1980,45, 5381. F. Igersheim and H. Mimoun, Nouu. J. Chim., 1980, 4, 711. J. Tsuji, H. Nagashima and K. Jori, Chem. Lett., 1980, 257. G, Strukul, R. Ros and R. A. Michelin, Inorg. Chem., 1982, 21, 495. P. Mulier and C. Bobillier, Tetrahedron Lett., 1981, 22, 5157. S. Uemura and S. R. Patil, Chem. Lett., 1982, 1743. M. T. Atlay, M. Preece, G. Strukul and B. R. James, J. Chem Soc., Chem. Commun., 1982,406; Can. J. Chem., 1983, 61, 1332. 268. (a) K L. K. Plute, R. C. Haltiwanger and M. Rakowski-Du Bois, Inorg. Chem., 1979, 18, 3246. (b) J. E. Hoots and T. B. Rauchfuss, lnorg. Cbem, 1983, 22, 2806. (c) G. R. Clark and D. R. Russel, J. Organornet. Chem, 1979, 173, 377. 269. K. B. Sharpless, A. Y.Teranishi and J. E. Backvall, J. Am. Chem. SOC.,1977, 99, 3120. 270. K. B. Wiberg, in ‘Oxidation in Organic Chemistry’, Academic, New York, 1965, part A, p. 69. 271. C. U. Nenitzescu, Bull. SOC. Chim. Fr., 1968, 1349. 272. S. Sundaram, N. Venkatasubramanian and S. V.Anantakrishnan, J. Sci. Ind. Res., 1976, 35, 518. 273. J, C. Collins, W. W. Hess and F. J. Frank, Tetrahedron Lett., 1968, 30, 3363. 274. D. H. G. Bosche, Methoden Org. Chem. (Houhen-Wqvl), 1975, 425. 275. G. Piancatelli, A. Scettri and M. DAuria, Synthesis, 1982, 245. 276. De Hasish, J. Sci. Ind. Res., 1982, 44, 484. 277. N. Miyaura and J. K. Kochi, J. Am. Chem. SOC.,1983, 105, 2368. 278. T. L. SiddaI, N. Miyaura, J. C. Huffman and J. K. Kochi, 1 Chem. Soc., Chem. Commun, 1983, 1185. 279. L. R. Horian and E. R. Corey, Inorg. Chem, 1968, 7 , 722. 280. R. Barral, C. Bocard, I. Seree De Roch and L. Sajus, Tetrahedron Lett., 1972, 1693. 281. R. Durant, C. D. Gamer, M. R. Hyde and F. E. Mabbs, J. Chem. Soc., Dalton Trans., 1977, 955. 282. G m e h Handbuch der Anorganischen Chemie, Mangan, Part C,, Springer-Verlag, Berlin, 1975. 283. D. Amdt, in ref. 274, p. 466. 284. R. Stewart, in ref. 270, p. 2. 285. D. M. Fenton {Union Oil), US Pat. 3 316 279 (1967) (Chem. Abstr., 1967,67,64 236). D. M. Fenton and L. G. Wolgemuth (Union Oil), US Put. 3 518 285 (1970) (Chem. Abstr., 1970, 73, 55 610). 286. R. J. Audette, J. W. Quail and P. J. Smith, J. Chem. Soc., Chem. Commun., 1972, 38. 287. Y.Tsuda and S . Nakajima, Chem. Lett., 1978, 1397. 288. D. G. Lee and M. Van Den Engh, in ref. 270, 1973, part B, p. 177. 289. E. S. Gore, Platinum Met. Rev., 1983, 111. 290. M. Schroder, Chem Rev., 1980, 80, 187. 291. N. Rabjohn, Org. React. ( N Y ) , 1976,24, 261. 292. H. J. Reich, in ref. 270, ed. W. S. Traknovsky, part C, p. 1. 293. Gmelin Handbuch der Anorganischen Chemie, Se Suppl uol. B l , Springer-Verlag, Berlin, 1981. 294. P. Miiller, in ‘The Chemistry of the Functional Groups: Ethers, Grown Ethers and Hydroxyl Groups’, ed. S. Patai, Wiley, Chichester, 1980, Vol. 1, p. 469. 295. J. RoEek and A. E. Radkowsky, J. Am. Chem. Soc, 1973, 95, 7123. 296. R. Durand, P. Ceneste, G. Lamaty, S. Moreau, 0. Pomares and J. P. Roque. Red. Trau. Chim.Fays-Bas, 1978,97,42. 297. E. J. Corey and J. W. Suggs, Tetrahedron Lett., 1975, 2647. 298. A. K. Rappe and W. A. Goddard, J. Am. Chem. Soc, 1982, 104, 3287. 299. J. T. Groves and W. J. Kruper, J. Am. Chem. Soc., 1979, 101,7613. 300. J. T. Groves and T.E. Nemo, J. Am. Chem. Soc., 1983, 105, 5786. 301. E. Carstanjen, J. Prakt. Chem., 1870, 110, 51. 302. A. Etard, C.R Hebd. Seances Acad. Sci., 1877, 84, 127. 303. F. Mares and J. RoEek, Collect. Czech. Chem Cummun., 1961,26, 2370. 304. D. P. Archer and W. I. Hickinbottom, J. Chem. Sm,1954, 4197. 305. K. B. Wiberg and G. Foster, J. Am. Chem. Soc., 1961,83, 423. 306. K. B. Sharpless and T. C. Flood, J. Am. Chem. Soc., 1971,93, 2316. 307. A. Nakamura, M. Nakayama, K. Sugihashi and S. Otsuka, Inorg. Chem., 1979, 18,394. 308. R. K. Grasselli and J. D. Burrington, Adv. Catal., 1981, 30, 133. 309. R. K. Grasselli, Adv. Chem. Ser., 1983, 222, 317. 310. F. Weiss, in ref. 12, p. 157. 311. D. Anrdt, in ref. 274, p. 465. 312. D. G. Lee, in ‘Oxidations: Techniques and Applications in Organic Synthesis’, ed. R. L. Augustine, Dekker, New York, 1969, p. 55. 313. D. Benson, ‘Mechanisms of Oxidation by Metal Ions’, Elsevier, New York, 1976, p. 149. 314. D. J. Sam and H. F. Simmons, J. Am. Chem. Soc., 1972, 94, 4024. 315. N. A. Gibson and J. W. Hosking, Aust. J. Chem., 1965, 18, 123. 316. H. J. Schmidt and H. J. Schafer, Angew. Chem, Int. Ed. Engl., 1979, 18,68. 317. S. L. Regen and C. Koteel, J. Am. Chem. SOC.,1977,99, 3827. 318. F. Freeman, C. 0. Fuselier, C. R. Armstead, C. E. Dalton, P. A. Davidson, E. M. Marchesfski, D. E. Krochman, M. N. Johnson and N. IC. Jones, J. Am. Chem. Sor,1981, 103, 1154. 319. C. M. Stark, 1 Am. Chem. SOC.,1971,93, 195. 320. E. H Huntress and H. C. Walter, J. Am. Chem. Soc., 1948,70, 3702. 321. G. A. Artamkina, A. A. Grinfeld and 1. P. Beletskaya, Russ. J. Org. Chem. (Engl. TmnsL), 1980, 16, 612. 322. G. Rouschias, Chem. Rev., 1974, 5, 531. 323. I. G. Kokarovtseva, I. N. Belyaev and L. V. Semenyakova, Russ. Chem. Reu. (Engl. Transl), 1972, 41, 929. 324. C. Djerassi and R. R. Engle, J. Am. Chem. SOL,J953, 75, 3838. 325. H. Copal, T.Adams and R. M. Moriarty, Tetrabedrun, 1972,UI. 4259.

261. 262. 263. 264. 265. 266. 267. 268.

Metal Complexes in Oxidation

405

P. Muller and J. Godoy, Heh. Chim. Acta, 1983, 66, 1791; Tetrahedron Lett, 1981, 22, 2361. K. B. Sharpless, K. Akashi and K. Oshima, Tetrahedron Lett., 1976, 29, 2503. R. Tang, S. E. Diamond, N. Neary and F. Mares, J. Cham Soc., Chem. Commun., 1978, 562. H. Mimoun and R. Charpentier (Institute Franpais du Pbtrole), Ger. Pat. 2920678 (1978) (Chem. Abstr., 1980,92, 128 393). 330. M. Matsumoto and S. Ito, J. Chern. Soc., Chem. Cornmun., 1981,907. 331. T. A. Foglia, P. A. Barr, A. T. Mdloy and M. J. Costanzo, J. Am. Oil Chem Soc., 1977,54, 858A, 870A. 332. P. H. J. Carlsen, T. Katsuki, V. S. Martin and K. B. Sharpless, J. Org. Chern., 1981, 46, 3936. 333. K. Kawamoto and T. Woshioka (Mitsui Petrochemical Industries), Eur. Pat. 21 118 (1981) (Chem. Abstr., 1981,94, 174 356). 334. L. M. Berkowitz and P. N. Rylander, J. A m Chem. Soc., 1958, 80, 6682. 335. (a) A. B. Smith and R. M. Scarborough, Synth. Commun., 1980, 10, 205. (b) G. Balavoine, C. Eskenazi and F. Meunier, J. Mol. C a r d , 1985,30, 125. 336. U. A. Spitzer and D. G . Lee, J. Org. Chem., 1974,39,2468. 337. H. Gopal and A. J. Gordon, Tetrahedron Lett., 1971, 31, 2941. 338. 0. Makowka, Bar., 1908, 41,943. 339. K. A. Kofmann, Ber., 1912, 45, 3329. 340. R. Criegee, Justus Liebigs Ann. Chem., 1936, 522, 75. 341. R. Criegee, B. Marchand and H. Wannowius, Justus Liebigs A n n Chem, 1942, 550,99. 342. L. F. Fieser and M. Fieser, ‘Reagents for Organic Synthesis’, Wiley, New York, vol. 1-6. 343. D. J. Gulliver and W. Lewason, Coord. Chem Rev., 1982, 46, 1 and refs. therein. 344. M. Schroder, A. J. Nietson and W: P. Griffith, J. Chem SOC., Dalton Trans., 1979, 1607. 345. S. G. Hentges and K. B. Sharpless, J. Am. Chem. Soc., 1980, 102, 4263. 346. A. 0. Chong, K. Oshima and K. B. Sharpless, J. Am. Chem. Soc., 1977, 99, 3420. 347. D. W. Patrick, L. K. Truesdale, S. A. Biller and K. B. Sharpless, J. Org. Chem., 1978,43, 2628. 348. K. Akashi, R E. Palermo and K. B. Sharpless, J. Org. Chem., 1978,43, 2063. 349. E. Herranz, S. A. Biller and K. B. Sharpless, J. Am. Chem SOC.,1978, 100, 3596 350. N. A. Milas and S. Sussman, J. Am. Chem. SOC.,1936,58, 1302; 1937, 59, 2345; 1939,61, 1844. 351. K, Akashi, R. E. Palermo and K. B. Sharpless, J. Org. Chern., 1978,43, 2063. 352. V. Van Rhenen, R. C. Kelly and D. Y. Cha, Tetrahedron Left., 1976, 1973; Org. Synth., 1978, 58, 43. 353. K. B. Sharpless and B. F. Lauer, J, A m Chem. Soc., 1972, 94, 7154. 354. H. P. Jensen and K. B. Sharpless, J. Org. Chem., 1975,40, 264. 355. D. Arigoni, A. Vasella, K. B. Sharpless and H. P. Jensen, J. Am. Chem. Sac, 1973,95, 7917. 356. M. A. Umbreit and K.B. Sharpless, J. Am. Chem. Sac., 1977, 99, 5526. 357. B. Chabaud and K. B. Sharpless, J. Org. Chem., 1979, 44,4202. 358. K. B. Sharpless and K. M. Gordon, J. Am. Chem. SOC, 1976, 98, 300. 359. A. M. Brownstein, Hydrocarbon Process., 1974, 53 (6), 129. 360. J. Kollar (Oxirane) US Put. 3 985 795 (1976) (Chem. Abstr., 1976, 85, 123 335). 361. G. W. Parshall, ‘Homogeneous Catalysis. The application of Catalysis by Soluble Transition Metal Complexes’, Wiley, New York, 1980. 362. R. J. Harvey, J. Kollar and J. P. Schmidt (Oxirane), US Pat 3 907 874 (1975) (Chem. Abstr., 1974, 80, 120 324). 363. J. Bergman and L. Engman,J. Org. Chem., 1982, 47, 5191; Tetrahedron Lelt., 1979, 35, 3279. 364. D. H. R. Barton, S. V. Ley and C. A. Meerholz, J. Chem. SOL,Chem. Commun., 1979, 755. 365. S. V. Ley, C. A. Meerholz and D. H. R. Barton, Tetrahedron Lett., 1980, 21, 1785. 366. L. Engman and M. P. Cava, J. Chem. Soc., Chem. Commun., 1982, 164. 367. P. M. Henry, ‘Palladium Catalyzed Oxidation of Hydrocarbons’, D. Reidel, Dordrecht, 1980. 368. P. M. Maitlis, ‘The Organic Chemistry of Palladium’, Academic, New York, 1971, vol. 1 and 2. 369. P. N. Rylander, ‘Organic Syntheses with Noble Metal Catalysts’, Organic Chemistry Monographs, Academic New York, 1973, vol. 28. 370. F. R. Hartley, Chem Rev., 1969,69, 799; J. Chem. Ed., 1973, 50, 263. 371. A. Mitsutani, Chem. Econ. Eng. Rev., 1973, 5 (3), 32. 372. K. Takehira, H. Mimoun and I. Seree De Roch, J. CataL, 1979, 58, 155. 373. D. R. Bryant, J. E. McKeon and B. C. Ream, J. Org. Chem, 1968, 33,4123. 374. D. M. Fenton and K. L. Olivier, Chem. Technology, 1972, 2, 220. 375. H. S. Kesling and L. R. Zehner, (Atlantic Richfield), US Pat. 4 171 450 (1979) (Chem. Abstr., 1979, 90, I21 035). 376. Y. Ichikawa and T. Yamagi (Teijin Limited), Fr. Pal. 2 195 612 (1973) (Chon. Ahsrr., 1974, 81, 25 389). 377. H. Alper, J. B. Woeli, N. Despeyroux and D. J. H.Smith, J. Chem. Soc., Chem. Commun., 1983, 1270. 378. D. M. Fenton and P. Steinwand, J. Org. Chem., 1974, 39, 701. 379. N. Kominami, Bull. Jpn. Pet. Insi., 1971, 13, 109. 380. H. Itatami and H. Yoshimoto, J. Org. Chem., 1973, 38,76; Bull Chem. SOC.Jpn., 1973,46, 2490; J. Catal, 1973, 29, 92. 381. T. F. Blackbum and J. Schwartz, 1. Chem. Sac., Chem. Commun, 1977, 157. 382. F. C. Phillips, 2. Anorg. AIlg. Chem., 1894, 6, 213; Am. Chem J., 1894, 16, 225. 383. J. Smidt, W. Hafner, R. Jira, J. Seldmeier, R. Sieber, R. Ruttinger and H. Kojer, Angew. Chem., 1959,71, 176; Angew. Chem., In?. Ed. Engl, 1962, 1, 80. 384. R. Jira and W. Freisleben, Organomet. React., 1972, 3, 1. 385. J. Tsuji, ‘Organic Synthesis with Palladium Compounds’, Springer-Verlag, Berlin, 1980. 386. M. Herberhold, ‘Metal n-Complexes’, Elsevier, Amsterdam, 1972. 387. R. F. Heck, ‘Organotransition Metal Chemistry’, Academic, New York, 1974. 388. B. M. Trost, Tetrahedron, 1977, 33, 2615; Pure Appl. Chem, 1979, 51, 787; Acc. Chem. Res., 1980, 13, 385. 389. J. Tsuji, Adv. Org. Chem., 1969, 6, 109; Acc. Chem. Res., t 9 6 9 , 2 , 144; 1973, 6, 8; Adv. Organornet. Chem., 1979, 17. 141; Pure AppL Chem., 1979, 51, 1235. 390. R. F. Heck, Top. Cum Chem., 1971, 16, 221; Adv. Catal., 1977, 26, 323; Ace. Chem. Res., 1979, 12, 146.

326. 327. 328. 329.

406 391. 392. 393. 394. 395.

U s ~ in s Synfhesvis and Catu1,ysis P. M. Maitlis, Acc. Chem. Res., 1979, 12, 146. P. M. Maitlis. Acc. Chem. Res., 1976, 9, 93. P. M. Henry, Acc. Uhem. Res., 1973, 6, 16; Adv. Organornet. Chem., 1975, 13, 363. L. S. Hegedus, ‘New Applications of Organometaltic Reagents in Organic Synthesis’, Elsevier, .4msterdam, 1976. B. M. Trost and T. R Verhoeven, in ‘Comprehensive Organometallic Chemistry’ ed. G. Wilkinson, Fergamon, Oxford,

1982, vol. 8, chap. 57, p. 799. M. A. Andrews and K. P. Kelly, J. Am. Chem. Sac., 2981: 103, 2894. J. E,. Backvall, Pure Appl. Chem., 1983, 55, 1669. E. Akermark, J. E. Backvall and K. Zetterbcrg, Acta Chim. Stand., Ser. B, 1982, 36, 577. J. E. Backvall, Acc. Chem. Res., 1983, 16, 335. J. E. Bickr,all, B. Akermark and S. 0. Ljunggrecn; J. Am. Chrm. Soc., 1979, 101, 2411. J. R. Stille and R. Divakaruni, J. Orgununzer. Chem., 1979, 169, 239. I. I. Moiseev and M. N. Vargaftik, Dokl. A h d . Naui, SSSR, 1966, 166, 370. G. T. Rodeheawr and D. F. Hunt, Chem. Cornmun., l V l ?818. R. Jim, W. Blau and D. Gnmm, Hydrocarbon Process.. 1976, 55 (3i, 97. Hydrocarbon Procesr., 1983, 62 ( I l), 69, 72. W. H. Clement and C. M. Selwitz, J. Org. Chem., 1964, 29, 241. D. R. Fahey and E. A. Zuech, J. Org. Chem., 1974, 39, 22. A. A. Grigorv? I. I. Moiseev, M. J. A. Klimenko and U. N. Lipina, Khim. Prom. (Moscow), 1972, 48, 14 (Chem. Absrr., 1972, 76, 85 317). 409. J. Tsuji, Pure Appl. Chem., 1981, 53, 2371 and refs. therein. 410. I. V. Kozhevnikov and K. I . Matveev, Appl. C a r d , 1983, 5 , 135; Kinet. Katal., 1977, 18, 862. 411. H . Ogawa, H. Fujinami, K. Taya and S. Teratani, J. Chem. Sac., Chem. Commun., 1981, 1274. 412. K. Fujirnoto, H. Takeda and T. Kunugi, Ind. Eng. Chem., Prod. Res. Dev., 1974, 13, 237; 1976+15, 259. 413. T. Hosokawa, T. IJIIO, S. lnui and S. I . Murahashi, J. Am. Chem. Sot., 1981, 103, 2318; J. Org. Chew., 1985, SO, 1282. 414. W. G. Lloyd and B. J. LuberoB, .I. Org. Chrm., 1969, 34, 1949. 415. ‘T.Hosoknwa, T. Ohta and S. Murahashi, J. Chem. SOC..C k m . C‘ommun., 1983, 848. 416. 1. 1. Moiseev, M. N. Vargaftik and Y. Scrkin, Dokl. Acud. Nuuk SSSR, 1960, 133, 377. 417. Hydrocarbon Process., 1983. 62 (11). 153, 154. 418. W. Schwerdtel, Hydrocarbon Process., 1968, 47 (1 I),187. 419. 1. P. Laloz,H . Mimono, J. J. Rouxel and L. Saussine (Institut FranGais du Pttrole), Fr. Pur. 2 185 52.5 jlYbl), (Chem. Abstr., 1982, 96, 199 116). 420. M. Tamura and T. Yasui, Chem. Commun., 1968, 1209. 421. Kuraray. Br. Par. 1 124 862 (1967) (Chem. Absfr., 1968, 69: 66391). 422. N. I. Kuznetsova, V. A. Likholobov, M. A. Fedotov and Y.I. Yermakov, J. Chem. Soc., Chem. Commun., 1982,973. 422. (a) S. Winstein, J. McCoskie, H. B. Lee and P. M. Henry, L Chem. Soc., 1976, 98, 6913. 423. W. Kitching, 2 . Rappoport, S. Winstein and G. C. Young, J. Am. Chem. Sac., 1966, 88, 2054. 424. P. Hayden, W. Featherstone and J. E. 1,loyd (IC:]), Ger. Par. 2037 179 (1971) (Chem. Abstr., 1971, 74, 87 363). 425. W. E. Smith, Br. Pat. 1461 831 (1977) (Chern. Abstr., 1975,82, 139 309). 426. I-. Saussine, J . P. Laloz and H. Mimoun (Institut Frdnyais d u Pktrole), Fr. Paf. 2 450 802 (1979) (Them. Abstr., 1981, 94, 208 348’). 427. W. Swodenk and G. Scharfe (Bayer), US Pot. 3 325 452 (19751 (Chem. Abrrr., 1971, 74, 3323). 428. H . Fernholz, F. Wunder and H. J. Schmidt (Hoechst), Ger. Put. 2 057 087 (1977), (Chem. Ahrr., 1972, 77, 87 902). 429. K. Takchira, H . Mimoun and I. Seree de Roch, J. CataL, 1979. 58, 155 and refs. therein. 430. A. M. Brownstein and H. L. List, Hydrocarbon Process., 1977, 5 6 ( 9 ) , 1.54 and refs. therein. 431. J. M. Davidson and C. Triggs, J. Cliem. Sac. ( 4 ) , 1968, 1331. 432. P. M. HenI?;, J. Org. Chem., 1971, 36, 1886. 433. H. J. Arpe and L. Hornig, Erdol Kohle, Erdgas Petrochem. Brennst. Chem., 1970, 23, 79. 434. J. E. McKeon and D.R. Bryant (Union Carbide!, US Pat. 3 547 982 (1970) (Chem. Abstr., 1971, 74, 141 294). 435. F. R. S. CIark, R. 0. C. Norman, C. B. Thomas and J. S. Willson, J. Chem. Soc., Perkin Trans. I , 1974, 1289. 436. J. K. Stille and 0. E. James, in ‘The Chemistry of Double-Bonded Functional Groups’, ed. S. Patai, Wiley, London, 1977, pari 2, Chap. 12. 437. J. Falbe, ‘New Synrheais with Carbon Monoxidc’, Springer-Verlag, Berlin, 1980. 438. R. Ugo, in ‘Catalysis in C , Chemistry’, ed. W. Keirn, D. Reidel, Dordrrcht, 1983, p. 135. 439. K. L. Olivier, D, M. Fcnton and J. Bide, Hydrocarbon Proccw., 1972, 51, 9s. 440. 1).M. Fenton, K. L. Olivier and 3. Biale, Prcp. Diu. 1’t.t. Chum., Am. Chem. Sot:., 1469, 14, C 7 7 . 441. D. M. Fenton and P. J . Steinwand, J . Org. C’hern., i972. 37, 2034. 442. D. E. James and J. K. Stille, J. Am. Chern. %IC., 1976, 98, 1810. 443. N. Yon Kutepow, K. Rittler and D. Neuhauer (RASFI, US Par. 3 437 676 (1969) (Chem. A b m , 1967, 66, 37 510). 444. J. K. StiHe, (Polymer Sciences Corporation), US Far. 4259 519 (1981) (Chem. Abstr., 1980, 92, 214907). 445. H. S. Kesling and L. R. Zehner (Atlantic Richfieid C,o.), US Par. 4 171 450 (1979) (Chem. ilbstr., 1979, 90, 121 035). 446. Y. Fujiwara, T. Kawanchi and H. Taniguchi, J. Chem. SOC., Chem. Commun., 1980, 220. 447. Y. Fujiwara, I. Kawata, H. Sugimoto and H. Taniguchi, J. Orgunomet. Chem., 1983, 256, C35. 448. D. M. Fenton (Union Oil), US Put. 3 700 729 (1972), Chern. Abstr., 1973, 78, 29 459). 449 D. M. Fenton and P. J. Steinwand (Union Oil), US Pur. 3 393 136 (1972) (Chem. Abstr., 1968, 69,95 982). 450. S. P . Current, J. Org. Chem., 1983, 48, 1779. 451. F. Riverti and U. Romano, J. Organorner. Chem., 1978, 154, 323. 452. K. Fuji et al. ( U B E Industries), US Pur. 4 229 591 (1980) (Chem. Absrr., 1979, 91, 4958). 453. J. E. Hallgreen, G. M. Lucas and R. 0. Matthcws, J. Organornet. Chem., 1981, 204, 135. 454. Ci. D. Cooper and D. E . Floryan, Ger. Par. 2 754 887 (1978). (Chem. Ahsfr., 1978, 89, 75 647) 455. I. V. Kozhenikov, Ruw. Chem. Rev. ( E n g . Trrrnsli, 1983, 52, 138. 456. H. W. Krause, R. Selke and H. Pracejus, Z. Chem., 1976, 16, 465.

396. 397. 398. 399. 400. 401. 402. 403. 404. 405. 406. 407. 408.

Metal Complexes in Oxidation

407

I. V. Kozhevnikov and K. I. Matveev, Rusr. Chem. Rev. (EngL T r a m . ) , 1978,47, 649. M. Sainsbury, Tetrahedron, 1980, 36, 3327. R. Hiittel, J. Kratzer and M. Bechter, Chem. Ber., 1961, 94, 766. C. F. Kohl1 and R. Van Helden, Red. Trav. Chim. Pays-Bas, 1967, 86, 193. H. C. Volger, Red. Trav. Chim. Pays-Bas, 1967,86, 677. I. Moritani and J. Fujiwara, Synthesis, 1973, 524. R. F. Heck, J. Am. Chem SOC.,1968, 90, 5518, 5526, 5531, 5535, 5542. R. S. Shue, J. Am. Chem SOC.,1971,93, 7116; Chem. Commun., 1971, 1510. W. G. Lloyd, J. Org. Chem., 1967,32, 2816. T. F. Blackbum and J. Schwartz, J. Chem. SOC.,Chem Commun., 1977, 157. B. S. Tovrog, F. Mares and S. E. Diamond, in ‘Catalysis of Organic Reactions’, ed. W. R. Moser, Dekker, New York, 1981, p. 139. 468. S. G. Clarkson and F. Basolo, Inorg. Chem., 1973, 12, 1528. 469. J. A. McCleverty, Chhem. Rm., 1979,19, 53. 470. R. Ugo, S. Bhaduri, B. F. G. Johnson, A. Khair, A. Pickard and Y. Ben Taarit, J. Chem. SOC.,Chem. Comrnun.,1976, 694. 471. B. S. Tovrog, S. E. Diamond and F. Mares, J. Am. Chem SOC.,1979, 101, 270. 472. B. S. Tovrog, F. Mares and S. E. Diamond, J. Am. Chem SOC.,1980, 102, 6616. 473. S. E. Diamond, F. Mares, A. Szalkiewicz, D. A. Muccigrosso and J. P. Solar, J. Am. Chem. SOC.,1982, 104,4266. 474. B. S. Tovrog, S. E. Diamond, F. Mares and A. Szalkiewicz, J. Am. C h e m Soc., 1981, 103, 3522. 475. J. X. McDermott, J. F. White and G. M. Whitesides, J. Am. Chem. Soc., 1976, 98, 6521. 476. A. Heumann, F. Chauvet and B. Waegell, Tetrahedron Lett., 1982, 23, 2767. 477. D. A. Muccigrosso, F. Mares, S. E. Diamond and J. P. Solar, Inorg. Chem., 1983, 22, 960. 478. J. T. Groves, W. J. Kruper and R. C. Haushalter, J. Am. Chem. Soc., 1980, 102, 6375. 479. C. L. Hill and E. C. Schardt, J. Am. Chem. SOC.,1980, 102, 6374. 480. J. A. Smegal and C. L. Hill, J. Am. Chem. SOC.,1983, 105, 3515. 481. 1. Tabushi and A. Yazaki, J. Am. Chem. SOC.,1981, 103, 7371. 482. M. Perree-Fauvct and A. Gaudemer, J. Chem. Soc., Chem. Commun.,1981, 874. 483. E. Guilmet and B. Meunier, Tetrahedron Lett., 1980, 21, 4449; 1982, 23, 2449; Noun J. Chim., 1982;6, 51t; J. Am. Chem. SOC.,1984, 106,6668. 484. D. Mansuy, M. Fontecave and J. F. Bartoli, J. Chem. Soc., Chem Commun.,1983, 253. 485. H. J. H. Fenton, J. Chem. SOC.,1894,65, 899. 486. C. Walling, Acc. Chem. Res., 1975, 8, 125. 487. D. M. Jerina, D. R. Boyd and J. W. Daly, Tetrahedron Lett., 1970, 457. 488. J. R. Lindsay-Smith and P. R. Sleath, J. Chem. SOC.,Perkin Trans. 2, 1983, 1165; 1982, 1009. 489. T. Yamamoto and M. Kimura, J. Chem. Soc., Chem. Commun., 1977, 948. 490. S. Undenfriend, C . T. Clark, J. Axelrod and B. Brodie, J. B i d Chem., 1954, 208, 731. 491. J. R. Lindsay-Smith and R 0. C. Norman, in ‘Oxidases and Related Redox Systems’, Wiley, New York, 1965, p. 131. 492. V. Ullrich and H. J. Staudinger, ‘Biochemie des Sauerstoffs’, Springer-Verlag, Berlin, 1968, p. 229. 493. V. Ullrich 2. Naiurforseh., Ted €3, 1969, 24, 699. 494. J. R. Lindsay-Smith, B. A. J. Shaw, c). M. Foulkes, A. M. Jeffrey and D. M. Jerina, J. Chem. Soc., Perkin Trans. 2, 1977, 1583. 495. H. Mimoun and I. Serree de Roch, Tetrahedron, 1975, 31, 777. 496. D. H. R. Barton, M. J. Gastiger and W. E. Motherwell, J. Chem. SOC.,Chem. Commun., 1983, 41, 731; Tetrahedron Lett., 1983, 24, 1979. 497. J. T. Groves, T. E. Nemo and R. S. Myers, J. Am. Chem. Soc., 1979, 101, 1032. 498. (a) J. T. Groves and T. E. Nemo, J. Am. Chem. SOC.,1983, 105, 5786, 6243. (b) J. T. Groves and D. V. Subramanian, J. Am. Chem. SOC.,1984, 106, 2177. 499. C. K. Chang and F. Ebina, J. Chem. Soc., Chem. Commun., 1981, 778. 500. (a) M. W. Nee and T. C. Bruice, J. Am. Chem. SOC.,1982, 104, 6123. (b) M. F. Powell, E. F. Pai and T. C. Bruce, J. Am. Chem. Soc, 1984, 106, 3277. 501. D. Mansuy, J . F. Bartoli, J. C. Chottard and M. Lange, Angew. Chem., Int. Ed. Engl., 1980, 19, 909. 502. D. Mansuy, J. F. Bartoli and M. Momenteau, Terrahedron Leit., 1982, 23, 2781. 503. T. Matsuura, Tetrahedron, 1977, 33, 2869. 504. E. I. Heiba, R. M. Dessau and W.J. Koehl, J. Am. Chem SOL, 1968,90, 2706, 5905; 1974, 96, 7977. 505. M. Okano, Chem. I n d (London), 1972, 423. 506. J. E. Bush and H. J. Finkbeiner, J. Am. Chem. SOC.,1969,91, 5903. 507. W. J. De Klein, R e d Tmv. Chim. Pays-Bas, 1975, 94, 48. 508. J. R. Gilmore and J. M. Mellor, J. Chem. Soc. (C), 1971, 2355. 509. A. Kasahdra, R. Saito and T. Izurni, Bull. Chem. SOC.Jpn., 1973,46, 2610. 510. E. I. Heiba, R. M. Dessau and W. J. Koehl, J. Am. Chem SOC.,1969, 91, 138. 511. J. R. Gilmore and J. M. Mellor, Chem. Commun., 1970, 507. 512. J. Hanotier, M. Hanotier-Bridoux and P. De Radzitzky, J. Chem. SOC.,Perkin Trans. 2, 1973, 381. 513. E. Baciocchi, L. Mandolini and C. Rol, J. Org. Chem, 1980, 45, 3906. 514. A. Onopchenko and J. G. D. Shulz, J. Org. Chem., 1978,38, 909. 515. S. R. Jones and J. M. Mellor, J. Chem. SOC.,Perkin Trans. 2, 1977, 511. 516. F. Asinger, ‘Paraffins Chemistry and Technology’, Pergamon, New York, 1968. 517. N. S. Enikolopian, K. A. Bogdanova and K. A. Askarov, Russ. Chem. Rev. (EngI. Transl.), 1983, 52, 13. 518. 1. Tabushi and N. Koga, J. Am. Chem. SOC.,1979, 101, 6456. 519. J. A. Srnegal, B. C. Schardt and C. L. Hill, J. Am. Chem. Soc., 1983, 105, 2920, 3277, 3510, 3515. 520. B. C. Schardt, F. J. Hollander and C. L. Hill, J. Am. Chem. Soc., 1982,104, 3964. 521. 0. Bortolini and B. Meunier, J. Chem. SOC.,Chem. Commun., 10r13, 1364.

457. 458. 459. 460. 461. 462. 463. 464. 465. 466. 467.

408

Uses in Synthesis and Catalysis

522. A. W. Van Der Made, J. W. H. Smeets, R. J. M. Nolte and W. Dreuth, J. Chem. Soc, Chem. Commun.,1983, 1204. 523. C. A. McAuliffe, H.F. AI Khateeb, 0. S. Barratt, J. C. Briggs, A. Challita, A. Hosseini, M. G. Little, A. G. Mackie and K. Minten, J. Chem. Soc., Dalton Trans., 1983, 2147 and refs. therein. 524. M. Constantini, A. Dromard, M. Jouffret, B. Brossard and J. Varagnat, J. Mol. Cafal., 1980, 7 , 89. 525. L. S. Boguslavskaya, Russ. Chem. Rev. (Engl. Transl), 1965, 34, 503. 526. D. I. Metelitsa, Russ Chem. Rev. (Engl. Transl.), 1971,40, 563. 527. F. Haber and J. Weiss, Proc. R Soc. London, Ser. A, 1934, 147,332. 528. M. Tohma, T. Tomita and M. Kimura, Tetrahedron Lett., 1973, 4359. 529. A. Brook, L. Castle, J. R. Lindsay-Smith, R. Higgins and K. P. Moms, J. Chem. Soc., Perkin Trans. 2, 1982, 687. 530. J. T. Groves and M. Van Der Puy, J. Am. Chem. Sac, 1974, %, 5274; 1976,98, 5290. 531. J. M. Maissant, J. M. Bodroux, C. Bouchoule and M. Blanchard, J. Mol. Catal., 1980, 9, 237. 532. S. Tamagaki, K. Suzuki, H. Okamoto and W. Tagaki, Tetrahedron Lett., 1983, 24, 4847. 533. S. Tamagaki, K. Hotta and W. Tagaki, C h m Lett., 1982,651. 534. G. A. Hamilton and J. P. Friedman, J. Am. Chem. Soc, 1963, 85, 1008; 1966, 88, 5266, 5269. 535. A. A. Kumar, C. S. Vaidyanathan and N. A. Rao, J. Sci. Ind. Res., 1978, 37, 698. 536. G.A. Hamilton, R. J. Workman and L. Woo, 3: Am. Chem. Soc., 1964,86, 3390. 537. 1. Tabushi, T. Nakajima and K. Seto, Tetrahedron Lett., 1980, 21, 2565. 538. J. M. Maissant, C. Bouchoule and M. Blanchard, J. Mol Catal., 1983, 14, 333. 539. J. R. Lindsay and P. R. Sleath, J. Chem. Soc, Perkin Trans. 2, 1983, 621. 540. G. A. Hamilton, in ref. 60, p. 405. 541. (a) J. T. Groves and R. S. Myers, J. Am. Chem Soc., 1983,105, 5791. (b) D. Mansuy, P. Batlioni, J. P. Renaud and P. Guerin, J. Chem. Soc., Chem. Commun., 1985, 155. 542. E. McCandlish, A. R. Miksztal, M. Nappa, A. Q. Sprenger, J. S. Valentine, J. D. S o n g and T. G. Spiro, J. Am. Chem. Soc, 1980, 102, 4268; Inorg. Chem., 1984, 23, 3548. 543. A. M. Khenkin and A. A. Shtejman, J. Chem. Soc, Chem. Commun., 1984, 1219. 544. S. G. Sligar, K A. Kennedy and D. C. Pierson, Proc. Nurl. A m d . Sci. USA, 1980, 77, 1240. 545, D. €3. Chin, G. N. La Mar and A. Balch, J. Am. Chem SOL, 1980, 102, 5947; 1983, 105,782. 546. S. Ito, K. Inoue and M. Matsumoto, J. Am. Chem SOC,1982, 104, 6451. 547. J. Hanotier and M. Hanotier-Bridoux, J. Mol. CataL, 1981, 12, 133. 548. J. Hanotier, Ph. Carnerman, M. Hanotier-Bridoux and P. De Radzitzky, J. Chem SOL, Perkin Trans. 2, 1972,2247. 549. A. Onopchenko and J. G. D. Schulz, J. Org. Chem., 1978,38,3729. 550. L. Verstraelen M. Lalrnand, A. J. Hubert and P. Teyssie, J. Bhem. Soc., Perkin Trans. 2, 1976, 1285. 551. K. Tanaka, Chemtech, 1974, 555. 552. R. T. Tang and J. K. Kochi, J. Znorg. NucL Chem., 1973, 35, 3845. 553. E. I. Heiba, R. M. Dessau and W. J. Koehl, J. Am. Chem. Soc., 1969, 6830. 554. T. Szymanska-Bozar and J. J. Ziolkowski, Koord. Khirn, 1976, 2, 1172; J. MOL CataL, 1979, 5, 341. 555. V. D. Luedecke, ‘Adipic Acid‘, in ‘Encyclopedia’ of Chemical Processing and Design’, ed. J. J. McKetta and W. A. Cunningham, Dekker, New York, 1977, vol. 2, p. 128. 556. F. Broich and H. Grasermann, Erdol Kohle, Erdgas, Petrochem., 1965, 18, 360. 557. F. Broich, H. Hofermann, W. Hunsmann and H. Simmrock, Erdol Kohle, Erdgas, Petrochem., 1963, 16, 284. 558. R. P. Lawry and A. Aguilo, Hydrocarbon Procesx., 1974, 53 ( l l ) , 103. 559, T. Okamoto and S. Oka, Tetrahedron Lett., 1981, 22, 2191. 560. A. Zornbek, D. E. Hamilton and R. S. Drago, J. Am. Chem SOC.,1982, 104, 6782. 561. A. Nishinaga and H.Tomita, J. Mol. CataL, 1980, 7 , 179 and refs. therein. 562. H.M. Van Dort and H. J. Geursen, Recl. Trav. Chim. Pays-Bas, 1967, 85, 520; J. Org. Chem., 1969, 34, 273. 563. V. M. Kothari and T.J. Tazuma, J. CataL, 1976,41, 180. 564. A. Nishinaga, H. Tomita, K. Nishizawa, T. Matsuura, S. Ooi and K. Hirotsu, J. Chem. SOC.,Dalton Trans., 1981, 1504 and refs. therein. 565. A. Zombeck, R. S. Drago, B. B. Corden and J. H. Gaul, J. Am. Chem. Soc., 1981, 103, 7580. 566. A. M. Martell, Pure Appl. Chem., 1983,55, 125. 567. A. Nishinaga, Chem. Lett., 1975, 273. 568. K. Nomiya, N. Wako, T. Komiyama and M. Miwa, 2.Naturforsch., Teil B, 1979, 34,442. 569. M. N.Dufour-Ricroh and A. Gaudemer, Tetrahedron Leu., 1976,45, 4079; J. Mol. Catal., 1980, 7 , 277. 570. A. Nishinaga, T. Tojo and T. Matsuura, J. Chem. Soc., Chem. Commun., 1974, 896. 571. A. D. Zuberbuhler, in ‘Metal Ions in Biological Systems’,ed. H. Sigel, Dekker, New York, 1975, vol. 5 , chap 7. 512. G. Davies, M. F. El Shazly, D. R. Kozlowski, C. E. Kramer, M. W. Rupich and R. W. Slaven, Ads. Chem. Ser., 1979, 173, 178. 573. G. Davies and M. A. El-Sayed, Znorg. Chem., 1983,9, 1257. 574. G. Davies, M. F. El Shazly, M. W. Rupich, M. R. Churchill and F. J. Rotella, J. Chem. SOL,Chem Commun., 1978, 1045. 575. P.Capdevielle and M. Maumy, Tetrahedron Lett., 1983, 5611. 576. M. R. Churchilk, G. Davies, M. A. El Sayed, M. F. El Shazly, J. P. Hutchinson, M. W. Rupich and K, 0. Watkins, Inorg. Chem., 1979, 18, 2296. 577. R. R. Gagne, R. S. Gall, G. C. Lisensky, R. E. Marsh and L. M. Speltz, Znorg. Chem., 1979, 18, 771. 578. U. Romano, R. Tesel, M. Mauri and P. Rebora, Ind Eng. Chem., Prod. Res. Dev., 1980, 19, 396. 579. R. D. Willett and G. L. Breneman, Inorg. Chern., 1983,22,326. 580. M. M. Rogie and T. R. Demmin, J. Am. Chem. Soc,, 1978, 100, 5472. 581. T. R. Demmin, M. D. Swerdloff and M. M. Rogic, J. Am Chem. Soc., 1981, 103, 5795. 582. G. Speier and 2.Tyeklar, J. Chem. SOC.,Dalton Trans., 1983, 1995. 583. G , Spier, 2. Tyeklar and A. Rockenbauer, Inorg. Chim. Acta, 1982, 66, L69. 584. D. Manegold, in ref. 284, p. 55. 585. ICI, Belg. Fat. 775 998 (1970).

Metal Complexes in Oxidation

409

586. G. Sosnovsky and D. J. Rawlinson, in ‘Organic Peroxides’, ed. D. S.Wem, Wiley, New York 1971, vol. 11, chapter 11, p. 151. 587. D. G. Rawlinson and G. Sosnovsky, Synthesis, 1972, 1. 588. F. Minisci, A. Citterio and C. Giordano, Acc. Chem. Res., 1983, 16, 27. 589. C. Amoldi, A. Citterio and F. Minisci, J. Chem. SOC.,Perkin Trans. 2, 1983, 531. 590. G. Eglinton and W. McCrae, Adv. Org. Chem., 1963, 4, 225. 591. I. D. Campbell and G. Eglinton, Org. Synth., 1965, 45, 39. 592. J. R. Lindsay-Smith and R. 0. C. Norman, J. Chem. Soc., 1963, 2897. 593. K. Sasaki, S. Ito, Y. Saheki, T. Kinoshita, T. Yamasaki and J. Harada, Chem. Lett., 1983, 37. 594. E. Steckham and J. Wellmann, Angew. Chem., 1976,88, 306. 595. K. Sasaki, S. Ito, T. Kinoshita and J. Harada, Chem. Lett., 1983,445. 596. B. Brossard, R. Janin, L. Krumenacker and J. Varagnat, Tetrahedron Letf, 1977, 26, 2273. 597. C. Giordano, A. Belli, A. Citerio and F. Minisci, J. 0%.Chem., 1979, 44,2314. 598. C. Walling, C. Zharo and G. M. El Taliawi, J. Org. Chem., 1983, 48,4910. 599. A. S. Hay, Polym. Eng. Sei., 1976, 16, 1; H. L. Finkbeiner, A. S.Hay and D. M. White, in ‘Polymerization Processes’, ed. C. E. Schildknecht and I. Skeist, Wiley-Interscience, New York, 1977, p. 537. 600. E. L. Reilly (Du Pont) Br. Pat. 1511 813 (1978) (Chern. Absrr., 1981,94, 174647). 601. P. Beltrame, P. L. Beltrame and P. Carniti, Ind. Eng. Chem., Prod. Res. Deu., 1979, 18, 208. 602. M. Constantini and M. Jouffret (Rhone Poulenc), Eur. Pat. 1199 (1979) (Chem. Abstr., 1979,91, 39 140). 603. W. Brenner (Hoffman-La-Roche), Ger. Pat. 2 221 624 (1972) (Chem. Abstr., 1973,78, 58 057). 604. P. Capdevielle and M. Maumy, Tetrahedron Lett., 1983, 24, 5611. 605. P. Capdevielle and M. Maumy, Tetrahedron Lett., 1982, 23, 1573, 1577. 606. J. Tsuji and H. Takayanagi, J. Am Chem. SOC.,1974, 96, 7349. 607. J. Tsuji, H. Takayanagi and I. Sakai, Tetrahedron Lett., 1975, 14, 1245. 608. C. Jallabert and H. %viere, Tetrahedron Lett, 1977, 1215; 1980, 36. 1191. 609. C. Neri and E. Perrotti (Snarn Progetti), US Pat. 4064 146 (1977). 610. C. A. Sprecher and A D. Zuberbuhler, Angew. Chem., Int. Ed. Engl, 1977, 16, 189. 611. F. L. Urbach, U. Knopp and A. D. Zuberbuhler, Helu. Chim. Acta, 1978, 61, 1097. 612. J. C. Sauer (Du Pont), US Pat. 4042621 (1977) (Chem. Abstr., 1977, 87, 22445). 613. V. Van Rheenen, Chem Commun..1969, 314. 614. T. Ho, Synth. Commun., 1974,4, 135. 615. K. Kaneda, T. Itoh, N. Kii, K. Jitsukawa and S. Terani’shi, J. Mol. Catal, 1982, 15, 349. 616. J. Tsuji and H. Takayanagi, Chem Lett., 1980, 65. 617. J. Tsuji, H. Kezuka, H. Takayanagi and K. Yamamoto, Bull. Chem Soc. Jpn., 1981,54, 2369. 618. T. Kajimoto, H. Takahashi and J. Tsuji, J. Org. Chem., 1976,41, 1389. 619. J. Tsuji, H. Takayanagi and Y. Toshida, Chem. Lett., 1976, 147. 620. T. Kajimoto, H. Takahashi and J. Tsuji, Bull. Chem. SOC.Jpn., 1982, 55, 3673. 621. M. I. Bruce, J. Organornet. Chem., 1972,44, 209. 622. M. Pasqueli, F. Marchetti and C. Floriani, Inorg. Chem., 1978, 17, 1684. 623. M. R. Churchill, B. G . De Boer, F. J. Rotella, 0.M.Abu Salah and M.I. Bruce, Inorg. Chem., 1975, 14, 2051. 624. M. Pasquali, C. Floriani, A. Gaetani-Manfredotti and C. Guastini, 1. Am. Chem. Soc,, 1981, 103, 185. 625. G. Stamman, R. Becker, J. Grolig and H. Waldmann (Bayer), US Pat. 4 370 275 (Chem. Abstr., 1982, 96, 19 658). 626. T. Saegusa, T. Tsudaand K. Isayama, J. Org. Chem., 1970, 35, 2976. 626. (a) W. Brackman, Discuss. Faraday SOC.,1968, 46, 122. 627. J, E. Lyons, in ‘Applied Industrial Catalysis’, Academic, New York, 1984, vol. 3, p. 131, 627. (a) D. E. Roundhill, in ‘Homogeneous Catalysis with Metal Phosphine Complexes’, ed. L. H. Pignolet, Plenum, New York, 1983, p. 377. 628. B. Meunier, Bull. SOC.Chim Fr., Part 2, 1983, 345. 629. H. Mimoun, M. Mignard, P. Brechot and L. Saussine, J. Am Chem. SOC., 1986, 108, 8711. 630. T. V. Lubben and P. T. Wolczanski, J. Am. Chem. SOC.,1985, 107, 701. 631. A. Van Asselt and I. E. Bercaw, Abstracts of the 1984 International Chemical Congress of Pacijic Basin Societies, Honolulu, Haway, Abstract 07H40; J. Am. Chem. Soc., 1986, 108, 8291. 632. H. J. Ledon and F. Varescon, Inorg. Chem., 1984, 23, 2735; J. Am. Chem. Soc., 1981, 103, 3601. 633. A. R. Miksztd and J. S. Valentine, Inorg. Chem., 1984, 23, 3548. 634. J. S. Valentine, J. N. Buntyn rf al, ref. 633, Abstract 07H07. 635. L. Saussine, E. Brazi, A. Robine, H. Mimoun, J. Fischer and R. Weiss, J. Am. Chem. Soc., 1985, 107, 3534. 636. G . Strukul and R. Michelin, J. Chem. SOC.,Chem. Commun., 1984, 1538; J. Am. Chem. Soc., 1985, 107, 7563. 637. G. Balavoine, C. Eskenazi, F. Meunier and H. Riviere, Tetrahedron Lett, 1984, 25, 3187; J. Chem. SOL, Chem Commun., 1985, 1111. 638. G. Green, W. P. Griffith, D. M. Hollinstead, S. V. Ley and M. Schroder, J. Chem. Soc., Perkin Trans. I, 1984, 681. 639. J. Tsuji, Synthesis, 1984, 369. 640. A. Heumann and B. Akermark, Angew. Chem., Inr. Ed. Engl, 1984, 23, 453. 641. K. Januszkiewicz and H. Alper, Tetrahedron Lett., 1983, 24, 5159. 642. H. Alper, K. Januszkiewicz and D. J. H. Smith, Tetrahedron Lett, 1985, 26, 2263. 643. S. F. Davison, B. E. Mann and P. M. Maitlis, J. Chem. SOC.,Dalton Trans., 1984, 1223. 644. J. E. Backvall, J. E. Nystrom and R. E. Nordberg, J. Am. Chem Soc., 1985, 107, 3676. 645. S. Fukuoka, M. Chono and M. Kohno, J. Org. Chem., 1984,49, 1460. 646. A. Chiesa and R. Ugo, J. Organornet Chem., 1985, 279, 215. 647. J. P. Solar, F. Mares and S. E. Diamond, Catal. Rev. Sci. Eng., 1985, 22, 1. 648. M.A. Andrews, T. C. T. Cheng et al., Organometallics, 1984, 3, 1479, 1777; 1985, 4, 268; J. Am. Chcm. SOL,1984, 106, 5913. 649. F. Mares, S. E. Diamond, F. J. Regina and J. P. Solar, J. Am. Chem. SOC, 1985, 107, 3545.

410

Uses in Synthesis and Catalysis

650. B. Meunier, E . Guilmet et al., J. Am. Chern. SOC.,1984, 106, 6668; J. Mol. Caral., 1984, 23, 115; 1985, 31, 221. 651. A. W. Van den Made and R. J. M. Nolte, J. Mol. C u d , 1984, 26, 333; 1985, 31, 271. 652. J. P. Collman, I. L. Braumann et al., J. Am. Chem. Soc, 1985, 107,2000; Roc. Natl. Acad. Sci USA, 1984,81, 3245; 1983, 80, 7039. 653. B. De Poorter and B. Meunier, Noun J. Chim, 1985, 9, 393. 654. D. Mansuy, P. Battioni and J. P. Renaud, J. Chern. SOL, Chem. Commun., 1984, 1255. 655. M. Fontecave and D. Mansuy, Tetrahedron, 1984,40, 4297. 656. H. Sakurai, Y.Hataya, T. Goromaru and H. Matsuura, J. Mol. Catal., 1985, 29, 153. 657. K. Suslick, B. Cook and M. Fox, J. Chem. SOC., Chem. Commun., 1985, 580. 658. J. P. Renaud, P. Battioni, J. F. Bortoli and D. Mansuy, J. Chem. SOC.,Chem. Commun., 1985, 888. 659. (a) P. C. Traylor, D. Dolphin and T. S. Traylor, J. Chem SOC.,Chem. Commun., 1984, 279. (b) T. Mashiko, D. Dolphin, T. Nakano and T. G. Traylor, J. Am. Chern. SOC.,1985, 107, 3735. 660. J. P. Collman, T. Kodadek et a[., J. Am. Chern. Sac., 1985, 107, 4343. 661. J. R. Lindsay-Smith and D. N. Mortirner, J. Chern. Soc., Chem Commun., 1985, 64,410. 662. M. Fontecave and D. Mansuy, J. Chem. SOC.,Chem. Commun., 1984, 879. 663. I. Tabushi, M. Kodera and M. Yokoyama, J. Am. C k m . Soc., 1985, 107, 4466. 664. C. C. Franklin, R. B. Van Atta, A. F. Tai and J. S. Valentine, J. Am. Chem. SOC.,1984, 106, 814. 665. M. F. Semmelhack, C. R. Schmid, D. A. Cortes and C. S. Chou, J. Am. Chem. SOC.,1984, 106, 3374. 666. K. D. Karlin el al., J. Am. Chem. SOC.,1984, 106, 2121, 3372.

51.4 Lewis Acid Catalysis and the Reactions of Coordinated Ligands tOBERT W. HAY Jniversity of Stirling, Stirling, UK 1.4.1 INTRODUCTION 61.4.1.1 The Hydrolysis of Cations 61.4.1.2 The Kinetic Investigation of Metal-promoted Reactions Involving Labile Metal Ions

412 413 413

1.4.2 THE HYDROLYSIS O F AMINO ACID ESTERS, AMIDES AND PEPTIDES 61.4.2.1 Labile Metal Systems 61.4.2.1.1 Base hydrolysis of amino acid esters 61.4.2.1.2 Monoamino esters 61.4.2.1.3 Mixed ligand complexes of monoamino esters 61.4.21.4 Bidentate and polydentate esters 61.4.2.1.5 Palladium(II) complexes 61.4.2.1.6 Kinetic stereoselectivity 61.4.2.1.7 Hydrolysis of amino acid amides ond peptides 61.4.2.1.8 Peptide bond formation 61.4.2.2 Non-labile Cobalt(III) Complexes 61.4.2.2.1 Ester hydrolysis 61.4.2.2.2 Peptide and amide hydrolysis 61.4.2.23 Cobalt hydroxide-promoted hydrolysis and lactonization 61.4.2.24 Peptide bond formation

414 414 416 416 417 419 423 424 425 426 427 427 43 1 434 436

1.4.3 HYDROLYSIS OF CARBOXYLIC ESTERS AND AMIDES 61.4.3. I Reactions Involving Coordinated Hydroxide in Labile Metal Systems

437 442

1.4.4 HYDROLYSIS OF PHOSPHATE ESTERS 61.4.4.1 Lobile Metal Ions 61.4.4.2 C:obalt(III) Complexes 61.4.4.3 Metal Hydroxide Gels

443 443 446 448

1.4.5 REACTIONS OF COORDINATED NITRILES

449

1.4.6

DECARBOXYLATION O F p-OXO ACIDS

453

1.4.7

ORGANOSULFUR COMPOUNDS

456

1.4.8 FORMATION O F IMINES

458

1.4.9 IMINE HYDROLYSIS

460

1.4.10

46 1

LACTAM HYDROLYSIS: PENICILLINS, CEPHALOSPORINS

463

1.4.11 HYDROLYSIS OF ANHYURIDES 1.4.12 HYDROLYSIS OF GLYCOSIDES, ACETALS AND THIOACETALS

464

1.4.13 HYDROLYSIS O F SULFATE ESTERS

465

1.4.14 REACTIONS OF COORDINATED AMINO ACIDS 61.4.14.1 Aldol Condensations 61.4.14.2 Schiff Bases 61.4.14.3 Isotopic Exchange and Racemization

466 466 467 467

1.4.15

AMINO ACID SYNTHESIS AND ALDOL CONDENSATIONS

468

1.4.16 ESTER EXCHANGE REACTIONS

469

1.4.17 HYDROLYSIS O F EPOXIDES

470

81.4.18 UREA HYDROLYSIS

470

4.4.19 ACYL TRANSFER REACTIONS

47 1

411

.CCb-N*

LJses in Synthesis and Cutnl-vsis

412 61.4.20 ENOLlZAT'lON

473

61.4.21 CARBONYL GROUP HYDRA'TION

474

61.422 ME.TAL 10s;-PROMOTEDCARBONYL REDUCTIONS

415

61.4.23 METAL. IOK-PROMOTED REACTIONS O F ALKENES

415

61.4.24

ISOMERIZATION AND CHELATE RING OPENING

477

61.4.25 REFERENCES

478 ~

61.4.1

.. ..

INTRODUCTION

Metal ions are Lewis acids and as such catalyze many reactions which are also subject to specific acid catalysis by the proton. Reactions in which metal ions are involved are often best described as metal ion-promoted reactions as the products of the reaction often remain bound to the metal ion. Although scattered references to metal ion-promoted reactions are to be found in the early literature it was not until the late 1950s that such reactions began to be studied in detail. A strong driving force has been the realization that some 30% of enzymes are metalloenzymes or require metal ions for activity. Many of the reactions dealt with in this article have been studied in an attempt to delineate possible mechanisms for enzymic processes. Before considering the consequences of the coordination of substrates to metal centres, it is instructive to compare some of the properties of metal ions with those of the proton. ( 1 ) The proton, although limited to a single positive charge, has an extremely high charge density and therefore great polarizing power. The charge density on a metal ion due to its larger size is generally much less. The charge density on PtIv may approach that of H+ and i.ts effect can be seen in the properties of complexes of this metal ion ( c g . pK, values and reactivity of coordinated amines and water).' (2) The coordination number of the proton is usually one, although it can be greater if hydrogen bonding is considered. The larger coordination numbers of the transition metal ions allow them more flexibility in grouping re.agents for intramolecular reactions. (3) The concentration of protons (and of protonated substrates) is limited at pH values near neutrality. Analogous metal complexes are often stable over a wide pH range including physiological pH, and may exist in the presence of appreciable concentrations of OH- or other nucleophiles. Westheimer' coined the term 'superacid catalysis' to describe the a.bility of metal ions to catalyze reactions at pH values where substantial concentrations of the protonated substrates cannot exist. If a substrate is subject to specific acid catalysis by a proton then it is fikely that an analogous Lewis acid-catalyzed reaction will occur. If the substrate becomes less reactive on protonation, then similar effects will be observed with the metal ion complex. The activating or protective effect of coordination arises from withdrawal of electron density from the ligand by the metal ion. In complexes where there is substantial n--character in the metal-ligand bond, back donation of electron density from filled d-orbitals on the metal ion into, for example, m*-antibonding ligand orbitals may result in little or no net electron withdrawal, or in some cases a net gain.' In these systems the effect of coordination on the reactivity of the ligand will be minimal.3 Table 1 Rate Constants for the Bromination of Aniline, the Anilinium Ion and a Metal Complex"

Substrate

PhNH, cis-[Co(en),CI(NHzPh)] PhNH3+

kiM-l

s-1)

3 x 1o'O 1 . 4 lo-' ~ 3 x io-'

Data taken from M . M. Jones, 'Ligand Reactivity and Catalysis', Academic, Yew York, 1968.

The Lewis acidity of a particular metal ion will be primarily a function of the charge density on the metal centre. Complexes of Cu", Zn" and C O " ~should therefore display properties intermediate between those of the free ligand and its conjugate acid. Early measurements on the bromination of aniline, the anilinium ion and coordinated aniline provide a good illustration of this point (Table I ) . Since this reaction is an electrophilic substitution, protonation and metal

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

413

:omplex formation inhibit the reaction. Many of the examples to be discussed in the chapter deal vith accelerated rates of nucleophilic attack due to the coordination of a carbonyl group uia a one pair on oxygen, or a nitrile group through its nitrogen atom which leads to an increased usceptibility of the carbon atom to nucleophilic attack. Although the metal complex is less eactive than the protonated substrate, it can exist at pH values where the concentration of the irotonated substrate can be negligible. For this reason many hydrolytic enzymes have metals such is zinc" at the active site. Enzymes use general acid catalysis and metal ion catalysis rather than he specific acid catalysis normally employed by the synthetic chemist. Several books and review articles are available dealing with various aspects of metal ion ;atalysis, 4-2429-32 and the role of metal ions in enzymic ~ a t a l y s i s . ~ ~ - ~ *

51.4.1.1 The Hydrolysis of Cations Metal ions in aqueous solution generally hydrolyze to form a series of mononuclear md polynuclear hydroxy complexe~.~ ,34 Iron( 111) for example forms [FeOHI2+, [Fe(OH),]+, Fe(OH),(aq), [Fe(OH)J, [Fe2(OH),I4+ and probably other polynuclear species. The formation 3f a monohydroxy complex (equation 1) can be used as a guide to the degree of electron withdrawal by the cation. Typical pK, values for various cations are listed in Table 2. (The pK, values of H30+and HzO are -1.7 and 15.7 respectively.) The tetrahedrally coordinated [Be(H,O),]*+ ion is more acidic than expected when compared with octahedrally coordinated aqua ions of divalent metals. Each hydrogen must carry a somewhat greater portion of the positive charge than it would in an octahedral complex, and this makes removal of a proton easier. Similar considerations apply to square planar Pd2'(aq). MercurylII) and tin(I1) are also more acidic than would be expected. The high acidity of rnercury(I1) is due to the unusual stability of 'linear' HzO-Hg--OH+, while Sn"(aq) has between two and three coordinated water molecules in the inner coordination sphere.35

'H Table 2 The pK, Values of Aqua Cations"

Ion ~

"

Be"

Bin1

Cd" co" CUI' E P HP" Fe"'

a

61.4.1.2

1

PK,

10A

PK,

4.97 5.40 4.0 10.08 9.65 7.96 7.9 0.25 2.19

Pd"

-2.3 7.71 10.59 9.86 0.5 0.62 3.2 0.65 8.96 -0.3

m'1

Mn" Ni" Pur" 1111

WV UW

Zn" ZrN

Data from D.W. Barnum, Inorg. G e m . 1983,22, 2297.

The Kinetic Investigation of Metal-promoted Reactions Involving Labile Metal Ions

Most metal ion-promoted reactions in which labile metal ions are employed involve a reaction scheme of the type shown in equations (2)-(5). There is a rapid pre-equilibrium formation of MS2+ (M2+= metal ion; S = substrate) (equation 2), which can be followed by attack of hydroxide ion (equation 3) of water (equation 4) to give products. In some cases unimolecular decomposition of MS2+ occurs, as for example in decarboxylation reactions (Section 61.4.6). Most reactions are studied with a constant substrate concentration and a varying excess of the metal ion under constant conditions of pH. Plots of kObs(the observed first order rate constant) versus the metal ion concentration can be of two types (Figures 1 and 2). In Figure 1, there is a linear dependence 3n the [M2*], indicating limited complex formation. The slope of the line kobs/[M2+] gives the :atalytic rate constant kea,. The behavior shown in Figure 2 illustrates 'saturation' conditions. At iigh metal ion concentrations the reaction rate becomes independent of the metal ion concentration h e to complete formation of the complex MS2+. Such plots can be used to obtain both rate

414

Uses in Synthesis and Catalysis

constants and eauilibrium constants. In this case it can readily be shown that kobo is given by equation ( 6 ) which can be rearranged to give equation (7). M2++S

& MSZ+

MS2++OH-

5 products

MS*++H,O

k.rs products

MS2' k

A products = -

oh

kX[M2+] 1+K[M2+]

CM2'1

CM'+]

Figure 1

Figure 2

A plot of l/kobs versus 1/[M2+]is linear of slope l/kK and intercept l / k The pH dependence of the plateau rate constant klimcan be used to determine if the reaction shows a dependence on the hydroxide ion concentration. A typical situation is shown in Figure 3. In this case there is a positive intercept so that kli, = k,+ k,,,[OH-]. The ko term normally indicates a solvolytic reaction (nucleophilic attack by water) while koH, the slope of the line, relates to base hydrolysis.

[OH-: Figure 3

Many additional complexities can occur in reactions of this type due to the presence of more than one complex in solution and to metal-buffer interactions which can lead to additional metal equilibria involving the buffer anion and even to mixed ligand complex formation.

61.4.2

61.4.2.1

THE HYDROLYSIS OF AMINO ESTERS, AMIDES AND PEPTIDES

Labile Metal Systems

The metal ion-catalyzed hydrolysis of amino acid esters and peptides has been a subject of continuing interest over the past three decades. In the early 1930s the hydrolysis of peptides was

Lewis Acid Cafalysis and the Reactions of Coordinated Ligands

415

round to be subject to metal ion catalysis, but the discovery by Kroll in 1952 that the hydrolysis 3f a-amino acid esters was catalyzed by metal ions stimulated considerable interest in the area. Many of these reactions can be considered as simple model systems for such metalloenzymes as sarboxypeptidase A, leucine aminopeptidase and glycylglycine dipeptida~e.~' A variety of reviews on various aspects of the topic are available.22,23,24,28,30 The work carried 3ut has dealt with catalysis by labile metal ions such as Cu" and Ni", with non-labile Co"' systems and with catalysis by Pd" and some rare earth ions. A major problem in studying the hydrolysis of a-amino acid esters in the presence of labile metal ions has been the determination 3f the ligand binding sites, and the identification of the catalytically active species in solution. Many of the early kinetic studies were carried out with monoamino esters, where the formation constants are quite low. Solutions containing, for example, copper(I1) and the ester ligand (E) will contain a variety of complexes such as [CUE]*', [CUE,]'+, [CUE,]'+ and [CuEJ2+. The most abundant species will depend on the copper(I1) : ester ratio, and inevitably there is more than one complex present. In addition, there will be mixed species [CuEA]+, [CuE2A]+,etc. formed as the hydrolysis of E to A- (the anion of the amino acid) occurs. As a result, it can be difficult to establish the nature of the hydrolytically most active species. A number of these difficulties have been overcome by the use of polydentate amino acid derivatives containing additional donor atoms capable of interacting with a metal ion or by using mixed ligand complexes. There has been considerable success in interpreting the kinetic data for such systems. Many of the initial investigations (as with all studies of metal ion-catalyzed reactions) were complicated by the use of complexing buffers which led to additional problems due to metal ion-buffer interactions. The use of pH-states and the advent of non-complexing buffers such as Hepes and Pipes3' and the new Elias buffers38have essent.ially resoIved this problem. Amino acid esters can be monodentate (1) or bidentate (2), and examples of both types of complex have been isolated and their solid state IR spectra ~ t u d i e d . ~Many ~ - ~ linvestigations have shown that formation of the monodentate ester species has similar e f f e p to protonation of the a-amino group. Thus the pK, values of MNH2CH(R)C02H and NH3CH2C02H (carboxyl ionization) are usually quite similar.42

Essentially three different routes can be considered for the base hydrolysis of an amino acid ester in the presence of ametal ion (equations 8-10). In general terms hydrolysis of the monodentate N-coordinated ester (equation 8) would be expected to be somewhat similar to base hydrolysis M"*-NH,CH,CO,R+OH-

+

M"+-NH2CH2C02-+ROH

,NH2\

M"+

I

HO:

CH2

I c=o I

/NH"CH2 - M

~

I

1Ln-i'+

(8)

+ ROH

\o

OR

of &H3CH2C02R(EHC), and this is generally observed to be the case (see later). Formation of - ~ 'carbonyl stretching frequency, the chelated ester species leads to a considerable r e d u ~ t i o n ~in~the e.g. from 1740 to 1600 cm-' in the case of copper(1I) indicating significant polarization of the initially suggested that such chelated ester species were carbonyl group by the metal ion. Kr01l~~ involved in the metal ion-catalyzed hydrolysis of glycine methyl ester and this view is now supported by much experimental evidence. Such carbonyl-bonded ester species have been characterized in the solid state (Section 61.4.2.2.4), a typical example being the cobalt(I1I) complex (3).13'Polarization of the carbonyl group by the metal ion assists nucleophilic attack by reagents

416

Uses in Synthesis and Catalysis

such as OH-, OH, and RNHz (Scheme 1).These reactions involve the formation of the tetrahedral in the hydrolysis intermediate { 4 ) and such an intermediate has been detected spectros~opically~~ of some cobalt(II1) com lexes of the type shown in (3). Oxygen exqhange data are also fully consistent with this view? Presumably in these reactions formation of the tetrahedral intermediate is the rate-determining step. OH

OR

0

Scheme 1 Metal ion-catalyzed hydrolysis of glycine esters

The intramolecular hydrolysis involving coordinated hydroxide (equation 10) was first detected in kinetically inert cobalt(II1) complexes and these reactions are considered in detail in Section 61.4.2.2.3.

61.4.2.1.1 Base hydrolysis of amino acid esters

In order to assess the magnitudes of the catalytic effects of metal ions on the hydrolysis of a-amino acid esters, it is necessary to have kinetic data on the base hydrolysis of such compounds NH,CHRCO,R'+OH-

k,. NH,CHRCO,-+

R'OH

~~H,CHRCO,R'+OH- k,,i NH,CHRCO,-+ R'OH

The reaction scheme can be summarized by equations (11) and (12). The pX, for the ionizatior of the a-amino group falls in the range 7 to 7.6.', Values of k, and kEH- have been determined for a series of a-amino acid esters and typical kinetic data are given in Table 3. Values of kEH+/k, are ea. 100 for a-amino acid esters. Withdrawal of the amino group to the @-positionlower: k,,+/k, to about 50 (e.g. methyl p-alaninate) and further withdrawal to the €-position (methy lysinate) reduces the ratio to 2.7.

61.4.2.1.2 Monoamino esters

Initially studies of metal ion-promoted hydrolysis were centred on simple monoamin1 ester^.^^,^,^' However, many of the initial investigations led to rather conflicting results. Th reactions are difficult to study due to the low formation constants of the active complexes. Mor recent have provided rate constants (Table 4) which show only order c magnitude agreement; however, it has been possible to establish that hydroxide ion is th predominant nucleophile at pH values of ca. 5. Higher pH values lead to precipitation of met: hydroxides. Evidence for nucleophilic attack by water has also been ~ b t a i n e d . ~ ~ - ~ ~ The rate enhancements in the copper(I1)-glycine ester systems are large (cu. 105-106-fold These rate accelerations are similar to the rate accelerations of cu. lo6 observed in the inecobalt(II1) systems where direct metal-ester carbonyl bonding occurs. It is thus likely that suc hydrolyses occur by the reactions outlined in Scheme 1. However, attack by coordinated hydroxid (equation 10) cannot be excluded and hydrolysis could occur by a combination of both reactia pathways.

417

Lewis Acid Catalysis and the Reactions of Coordinated Ligands Table 3

Rate Constants for the Base Hydrolysis of Amino Acid Esters at 25 "C and 1 = 0.1 M (KCI)'

Amino esfer

Methyl glicinate Ethyl glycinate Methyl a-alaninate Methyl a-amino-nbutyrate Methyl norvalinate Methyl norleucinate Methyl valinate Methyl leucinate Ethyl leucinate Methyl isoleucinate Methyl p-alaninate Methyl serinate Methyi @-phenylalaninate Ethyl @-phenylalaninate Methyl methioninate Methyl typtophanate Methyl 2,3diaminopropionate Methyl histidinate Methyl cystein'are Methyl lysinate Ethyl cysteinate a

kE (M-' s-')

ICEH+

(M-' s-')

1.28 0.64 1.11 0.39

28.3 22.9 80.3 44.5

0.40 0.37 0.076 0.455 0.187 0.067 0.136 0.99 0.55

40.2

kH:+ (M-'

s-')

40.8

0.235 0.77 0.29

57.3

0.73 0.62 0.07 (E-) 0.46 0.04 (E-)

-

46.5

I .07 (EH)

3.83 (EH')

1.26 0.60 (EH)

73.5 6.67 (EH*)

Data taken from ref. 22. Rate constants are not listed for amino acid esters such as methyl 4-amino-n-butyratewhere lactamization occurs (see ref. 38). In the case of the cysteine esters, E- = -SCH2CH(NH2)C0,R EH = HSCH,CH(NH,)CO,R EH" = -SCH,CH&H,)CO,R

Table 4 Rate Constants for the Metal Ion-promoted Hydrolysis of Methyl Glycinate and Ethyl Glycinate Reactiona

CuE2++OHCuE2++H,O ME** + OH-b NiE*'+ H,Ob CUE" + O K b CuEA+ + OH-b CUE'+ + H,Ob a

61.4.2.1.3

Rate constani (ionic strength, temperature)

ReJ

7.6 x lo4 M-' s'-' (0.16 M,25 "C) 4.3 x lod5 s-' (0.16 M, 25 "C) 3.98 x 10' M - I S - ~ (l.OM, 30°C) 1 . 9 ~ 1 0 - ~ s(l.OM, -~ 30°C) 1 . 4 105 ~ M-1 s-I (?, 25 "C) 8.41 ~0' M-' s-' (?, 25 "C) 2.5 x s-' (?, 25 "C)

46 46 47 47 48 48 48

'

E is H2NCH,C02Me or H2NCH2C02Et;A- is H,NCH,CO,-. These values refer to glycine ethyl ester.

Mixed ligand compl~xesof monoamino esters

The measurements made with mixed ligand complexes of palladium(11) illustrate the useful data which can be obtained from mixed ligand complexes containing simple monoamino acid ligands. Angelici and coworkers have studied the hydrolysis of a-amino acid esters in mixed ligand complexes of the type [CUL{NH,CH(R)CO,R'}]~+using ligands (L) such as iminodiacetate (imda):9 nitrilotriacetate (nta)," 2,2',2"-tris(aminoethyl)amine (tren)," terpyridyl (terpy)," . ~ ~ work by Hay and diethylenetriamine (dien)" and bis(2-pyridylmethylamine) ( d ~ a ) Further Banerjee discusses mixed ligand complexes of nta,54 ethylenediaminemonoacetate (edma),'5 glycylglycine dianionS6and iminodiacetate (imda).57To various degrees these studies can be said to mimic metalloenzyme-substrate complexes. The chemistry involved can be illustrated by a specific example. Using a ligand such as ethylenediaminemonoacetate;' ternary complexes with amino acid esters can be formulated as either (5) or (6). The ester ligands in the ternary complexes of edma- undergo base hydrolysis some lo3 times faster than the free unprotonated esters, although considerably lower effects occur with methyl L-histidinate (Table 5 ) . The magnitude of these effects suggests that the amino acid

418

IJses in Synthesis and Catalysis

I'

01.4,

Table 5 Catalytic Effects for the Hydrolysis of the Mixed Ligand Complexes [Cu(edma)(NH,CH(R)C02R')]+ at 25 "C and I = 0.1 M (KNO,!"

IO-' k,, Ester

(complex)

(M-1 s-l

>

iester)

k,,

(M-1

1.71

GlyOMe L-p-PheOMe

0.99

GlyOEr L-n- AlaOEt r-p-PheOEt L-HisOMe

0.63 0.59 0.38 0.025

s-1)

1.28 0.55 0.64 0.55 0.24 0.62

Ratio

1 . 3 4 10' ~ 1.18 x io3 io3 1.1 x 10:

LOX

1 . 6 10' ~ 4u

Data tnkcn from R. W. Hay and P. K. Baneries. J, Chew. S o r , Dnlton 7 r a n s , 19x0, 2452.

ester ligand is probably bidentate with a rather weak alkoxycarbonyl interaction (6). A weak interaction is expected for the Jahn-Teller-distorted d9 configuration of copper([I). Intramolecular attack by coordinated hydroxide (7) appears unlikely in this system in view of the relatively low rate accelerations observed, and the fact that the reaction shows a first order dependence on the hydroxide ion concentration up to pH 8. 0 1 l2

(7 1

Actibation parameters have been determined5' for the hydrolysis of a-amino acid esters in mixed ligand complexes (Table 6). For base hydrolysis of the complcx [Cu(nta)(NH,CH,CO,Et)]-, AH* = 20.5 kJ mol-' and AS' = -138 J K-' mol-'. Catalysis in this system is primarily due to a substantial decrease in AH' (by cu. 21 kJ mol-') compared with the free ligand. A detailed discussion of the activation parameters i s available.58 Table 6 Activation Parameters for the Hydrolysis of Methyl Glycinate in the Presence of Metal Yitriloacetates and Tetradentate Nickel(I1) Chelates at r = 0.1 Ma k,, , 2 5 "C Complex

log K ,

[ Ni(tren)( GlyOMe)]'+ [~i(trien)(G~y~~~e)]~+ [Ni(edda)(GlyOMc)1 [Ni(nta)(GlyOMe)]

14.8 14.0 13.5 11.47 10.81

LCo(nta)(tilyDMe)] [Cu(nta)jCiyOMe)]

[Zninta)(GlyOMe)] Cu (npa)(ClyOEt)]

13.05

10.44 13.05

(M-1

s-1)

67.1 53.1

41.2 52.3

18.6 460

34.6 78.2

AH+ (kJ mol-') 14.2i-3.3 30.1 + 2.9 20.1 t 3 . 3 3.8 f 2.9 6.3 + 3.3 14.2+ 5.0 16.7 2.9 20.5

*

AS *

(J K-' mol-') -163i8 -109k4 -14618 -197+8

-3.1+8 -159* 13 -159~8

-138

'The data are taken from the compilation given by D. E. Yewlin, M. A. Pellack and R. Nakon, J. .4m.Chem. Soc. 19?7,99, 1078

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

419

An interesting correlation has been observeds3 between the formation constant XcuLof the metal complex and its catalytic activity in a mixed ligand with an amino acid ester. Large values of KCuL(equation 13) lead to lower base hydrolysis rates in the ternary complex. The Lewis CUZ++L

L

L

CUL2+

(13)

acidities of the CuL"' complexes decrease as the formation constant for CuL"' increases (Table 7). Strong a-donors such as tren greatly decrease the Lewis acidity of copper(I1). In addition the effects of charge on these reactions do not appear to be of great importance. Thus [Cu(edma) (GlyOMe)]+ undergoes slower base hydrolysis than [Cu(irnda)(GlyOMe)], while the hydrolysis of [Cu(dpa)(GlyOMe)]'+ is some 10 times slower than the edma derivative. A n y electrostatic facilitation of the reaction seems to be overshadowed by changes in the Lewis acidity of copper(11). However, it should be appreciated that these conclusions assume that complexes all react by a similar mechanism. Table 7 Rate Constants (koH) and Equilibrium Constants (log K,,,) Associated with the CuL-promoted Hydrolysis of Methyl Glycinate at 25 "C

K,

[CuL(GlyOMe)]

(M-I s-')

[Cu(irnda)(GlyOMe)]

7.6 x lo3 a

[Cu (edma)(GlyOMe)]+

1.7~10'" 4.6x IO2

[Cu(nta)(GIyOMe)][Cu(terpy)(GlyOMe)12' [C~(dpa)(GlyOMe)]~' [C~(dien)(GlyOMe)]~+ [Cu(tren)(GlyOMe)lZf a

2.2X1OZL 1.7x le

1.4x 1.3

Id a

log Kcur

Ref:

10.63 13.29 13.10 13.40

51

14.4 15.9 18.8

55 50

51 53 52 51

I =0.1 M. I =0.07 M. I =0.05 M.

The results are of some significance as far as metalloenzymes are concerned as the Lewis acidity of the metal ion can presumably be 'tailored' to a degree by the appropriate degree of u-donation and +acceptance. For the bonding of Zn" to apocarboxypeptidase a value of log K of 10.5 has been estimateds9 and a similar value has been reported6' for Zn" with apocarbonic anhydrase. Such considerations may be of relevance in determining metal ion specificities of metalloenzymes. Metal ions with high binding constants for apoenzymes may be poor Lewis acid catalysts. Walker and Nakon6' have determined values of koH and the appropriate activation parameters for the base hydrolysis of a series of mixed ligand complexes [Cu(L)(GlyOMe)]"+ where L is a tetradentate or tridentate ligand. Tridentate and tetradentate chelates fall on different isokinetic lines suggesting different hydrolytic mechanisms. Both the strengths of the donor atoms of the auxiliary ligands and the charge carried by the complex are of importance in determining the hydrolysis rates. Some related studies have also been published.62 Correlations have been values for tridentate copper(I1) chelates and both log koH for base observed63 between A,, hydrolysis of [Cu(L)(GlyOMe)]"+ and log KOHfor the equilibrium (14). The formation constants KCuL(equation 13) were used as a measure of the Lewis acidity of the various complexes [CuL]"+. Some representative data are given in Table 8. KOH

[CuL]"++OH'

===[CUL(OH)]'"~''+

(14)

61.4.2.1.4 Bidentate and polydentate esters

The metal ion-promoted hydrolysis of a number of bidentate esters such as methyl 2,3diaminopropionate (81, methyl histidinate (9), methyl cysteinate (10) and the ethyl ester of ethylenediaminemonoacetate (11) have been studied. The first three esters give very thermodynamiTypical kinetic data cally stable metal complexes in solution with pendant ester f~nctions.6~-'~ for these systems are given in Tables 9 and 10, Base hydrofysis of the L-(+)-histidine methyl ester complexes72follows the reactivity order [CuEZ12+>[NiE,l2+> [CUE]"> [NiE]'+ > EH> CuEA> [CUEOH]+>[NiEAj+> E at 25 "C ( I =0.1 M) where E is the ester and A- is L-histidinate. The bis complex of methyl histidinate

Uses in Synthesis and Catalysis

420

Table 8 Formation Constants for Hydroxo Complex Formation by [CuL]*+ Complexes ( K O , ) , Second Order Rate Constants ( kH) for Base Hydrolysis of [CuL(GlyOMe)jx+ and A, Values for [CuLIx+a log

CUL

[Cu(tren)]*+ [Cu(dtma)]' [Cu(paa)l CCu(nta)l [Cu(dpa)lZ* [cu(dien)l2' CCu(tev~)l~+ [Cu(lrnda)]

log

kOH

0.95 0.83 2.25 2.50 2.06 1.39 2.28

4.44 4.25 4.34 4.20

860 710 670 715 637 611 680 730

5.05

4.67 5.71 6.26

4.50

(nm)

A,,,

I(OH

Data from J. K. Walker and R. Nakon, Inorg. Chem. Acta, 1981, 55, 135.

CH,-CH-C02Me

I

1

NH2 NH2

03)

CH2-CH-C02Me

CHZ-CH,

SH

NH2 NHCH2C0,Et

I

!

I

NHZ

1

(11)

(10)

Table 9 Rate Constants k,

for the Metal Ion-promoted Base Hydrolysis of some Bidentate Amino Acid Ester Species at 25 "C

bH for

complexes" (M-' s-')

System

ME?+ Cu"-L-methyl histjdinate NiII-L-methyl hislidinate CU"-Q, L methyl 2,3-diaminopropionate Hg"-D, L-methyl 2,3-diaminopropionate a

MEZ+

MEA+

-

198.gb 328 118 37Sb 188.3 305 115

42.7 16.5

-

28.3 88.7 24.3

ME(OH)+

-

-

175 78.3 322.5 87.7 618 82

33.7 I

135 50

All rate constants refer to base hydrolysis at 25 "C and I =0.1 M unless otherwise stated,

r=0.16~.

Table 10 Rate Constants for the Metat Ion-promoted Base Hydrolysis of some Bidentate Species of Cysteinate Esters at 25 "C System

Ni"-L-methyl cysteinate Pb"-L-methyl cysteinate Zn'l-L-methyl cysteinate Cd"-cmethyl cysteinate Mg"-L-methyl cysteinate Nil1-L-ethyl cysteinate Zn"-L-ethyl cystejnate CdII-r-ethyl cysteinate

10.8 5.7 4.7

3.5 2.3 2.6(b1 1.9'b' LOP'

All rate constants refer to base hydrolysis at 25 'C and I

r =0.16 M.

= 0.1

3.6 0.97 1.38 0.7 0.77

M unless otherwise stated.

71 71 71 71 71 67 67 67

R6$ 67 72 84 67 72 73 73

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

42 1

with copper( IJ) (CuE2") hydrolyzes some 265-fold more rapidly than the unprotonated ester (E) when the appropriate statistical corrections are made. The protonated ester (EH') hydrolyzes some 75 times more rapidly than E. The rate acceierations are relatively low as the metal ion does not interact with the alkoxycarbonyl group of the ester. In these complexes, electrostatic effects due to the positive charge(s) carried by the species appear to be of prime importance in determining hydrolysis rates. Similar considerations apply to complexes of methyl 2,3-diaminopr0pionate.7~Thermodynamic parameters AH* and AS&, for the base hydrolysis of the [CuE212+and [CuEA]+ complexes of methyl histi ti din ate^^ and methyl ~~-2,3-diaminopropionate~~ indicate that AH' values are ca. 4-8 kJ mol-' higher for the metal complexes compared with the free ligands. The rate increases are due to more positive values of AS,+,,for the complexes. Angelici and his coworkers have overcome some of the practical problems arising from the low formation constants of simple monoamino esters, by incorporating the ester moiety into a larger polydentate ligand. A typical example is the use of ethyl glycinate-N,N-diacetic acid (12) which hydrolyzes to give nitdottiacetic acid (13). CH,CO,H

CH,CO,H

/

N -CHzCOzEt 'CH2C02H (121

(13)

Kinetic measurements have been r e ~ o r t e d ' ~ -for ' ~ ligands involving ethyl glycinate (14a), ethyl alaninate (14b), ethyl valinate (14c), ethyl leudnate (14d), butyl glycinate, ethyl p-alaninate (15a) and ethyl 4-aminobutyrate (15b). Metal ion-promoted hydrolysis has been studied using a variety of metal ions (Cd", Ni", Mn",Fe", Co", Zn", Pb", La'", Nd"', Gd'", Dy'", Er"', Yb'", Lu"' and Sm"') for a few oE these ligands. The 1 : 1 complexes have large formation constants and display little tendency to add a second ligand. However, hydroxo complexes are formed at higher pH due to deprotonation of water molecules.

COzEt

COzEt

(14) a; R = H b; R = M e c; R=CHMe, d; R = CH,CHMe,

(15) a; n = 2 b; n = 3

The ethyl glycinate-N,N-diacetic acid derivative (12) may be used to illustrate the general features of these reactions. Hydrolysis of the copper(1I) complex in the pH range 5-7, studied by pH-stat, corresponds to equation (15) where EGDA2- = ethyl glycinate-N,N-diacetate and [Cu(EGDA)I+OH-

-

[Cu(NTA)]-+ EtOH

(15)

NTA3- = nitrilotriacetate. The rate expression takes the form, rate = koHICu(EGDA)][OH-] with koH= 2.18 X lo4 M-' s-' at 25 0C.74The kinetic measurements do not differentiate between bimolecular hydroxide ion attack on the 1: 1 aqua chelate (14a) and intramolecular attack by coordinated hydroxide (ca. 0.01 to 20% of [Cu(EGDA)(OH)]- exists in solution in the pH range 5.0-7.0). Attack by coordinated hydroxide does not appear to be favoured in this system due to the relative dispositions of the hydroxide ion and the alkoxycarbonyl group. The hydrolytic reactions are very rapid, with rate accelerations similar to those observed with copper(I1) and methyl glycinate, suggesting that in aqueous solution there is a direct interaction between the metal ion and the alkoxycarbonyl group as shown in (16). Rodgers and Jacobson" have published the crystal structure of DL-(ethyl valinate-N,N-diacetato)diaquacopper( H), the copper(IT) complex of EVDA (14c). The stereochemistry around copper is distorted octahedral, the equatorial coordination sphere consisting of the tertiary

Uses in Synthesis and Catalysis

422

1 OEt (16)

nitrogen, the two acetate oxygens and a water molecule. The second water molecule occupies an axial position at a greater distance from the metal, while the ether oxygen of the ester group occupies the opposite axial position lying at 2.34 8, from the metal. The ester carbonyl group is not coordinated in the solid state. This system provides a good example of some of the difficulties involved in defining the nature of catalytically active species of labile complexes in solution. Kinetic data for the base hydrolysis of a variety of metal complexes of EGDA are summarized in Table 11. Rare earth ions such as Yb"' and Lu'" have very marked catalytic effects, and these systems provide one of the few examples of detailed studies of rare earth catalysis. Activation parameters have been obtained (Table 12) and these show some interesting variations. In complexes of nickel(II), copper(I1) and lead(II), the rate accelerations are due primarily to a lowered AH* with little change in AS* compared with the free ligand. For rare earth ions quite the reverse is observed with more positive (i.e. less negative) AS* values leading to the rate enhancement with no change in AH*. Different mechanisms appear to operate in the two systems. Nucleophilic catalysis (by acetate, €€PO:'-,pyridine, 4-picoline and nitrite) of some of these reactions has been ~ t u d i e d using '~ esters derived from ethyl valinate and ethyl leucinate which are sterically hindered to base hydrolysis. Table 11 Rate Constants for the Base Hydrolysis of Metal Complexes of Ethyl Glycinate-N,N-Diacetate at 25 "C" ~ _ _ _ _ _ _ _ ~

Metal ion

1O-'k (M-I s-')

Metal ion

2.14 3.89 4.18 10.1 38.6 66.2 218 283

La(II1) Nd(II1) Sm(ll1) Gd(I1I) DY(W

Cd" Ni" Mn" CO"

Fe" Zn" CU"

Pb" a

Er(TI1)

Yb(II1) Lu( 111)

lO-'k (M-'

s-l)

115

347 447 376 877 1920 3460 3450

Data from B. E. Leach and R. J. Angelici, J. Am Cbem. Soc., 1968, 90, 2504. I = 0.05 M (KNO,), [MI= [EGDAI = 6.7x M.

Table 12 Activation Parameters for the Base Hydrolysis of Ethyl Glycinate and Metal Complexes of Ethyl Glycinate-N,N-Diacetate a Ester

Ethyl glycinate Betaine ethyl ester [Cu(EGDA)] [Ni(EGDA)] [WEGDA)I Ism( EGDA)]+ [Lu(EGDA)]+ a

AH' (kJ mol-') 44.4

* 2.1

40.6* 1.3 20.9 k2.1

23.8 * 1.7 18.4k2.5 41.8zk1.7

42.7 f2.5

AS* (J K-' mol-') -90.8 f4.2 -77.4 f 4.6 -91.6 f6.3 -101.3~5.0 -98.3 f8.8 -16.7f5.0 +3.3 f8.4

Data for the metal complexes taken from B. E. Leach and R. J. Angelici, J. Am Chem. SOC.,1968,90. 2504.

The copper(11)-catalyzed hydrolysis of the methyl esters of glycylglycine and glycylsarcosine has been as has the copper(I1)-promoted hydrolysis of ethyl glycylglycinate, ethyl Under the conditions of the experiments (pH > glycyl-a-alaninate and ethyl glycyl-~-leucinate.~~ 7.0) the peptide esters act as tridentate ligands, donation occurring from the terminal amino group and the deprotonated amide nitrogen atom with a weak interaction between the metal ion and the carbonyl group of the ester (17). Rate accelerations of the order of lo3 (compared with the

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

H2NL ,N>

1'K.

d

o

HaN,

,N\

423

+ H+

(17)

unprotonated ligands) were observed (Table 13). Analogous studies have been carried out with Pd" complexes" but in this case much higher rate accelerations (ca. lo') are observed. Nakon and Angelicis' have carried out a potentiometric and IR study of copper(I1) complexes of glycylglycine methyl ester and glycylsarcosine methyl ester in an attempt to define the nature of the active complexes in aqueous solution. Table 13 Rate Constants and pK, Values for the Base Hydrolysis of some

Copper(I1) Complexes of Dipeptide Esters at 25 "C and I =0.01 Ma Ligand

Ethyl glycylglycinate Ethyl glycyl-0-alaninate Ethyl glycyl-L-leucinate a

8.0 9.1

2.87 x io3 1.36 x 10'

1.98 x 10' 1.51 x 10' 1.68 x Id

Data taken from R. W. Hay and K. B. Nolan, 1; Chem. Sac., Dalton Trnns., 1974, 2542. Base hydrolysis or the aqua complex. Base hydrolysis of the hydroxo complex.

Base (and water) hydrolysis of the 1: 1 complex ofthe tetramethyl ester of EDTA with copper(I1) ([Cu(Me4EDTA)]") has been investigated.82In the complex probably two ester groups of the ligand are bonded to the metal ion in conjunction with the two nitrogen donors. Rate accelerations of the order of lo5 were observed in base hydrolysis compared with the unprotonated ligand (Table 14). A variety of metal complexes of the ligand have also been isolated and characteri~ed.~~ Table 14 Rate Constants for the Base Hydrolysis of EDTA Methyl Esters and their Copper(I1) Complexes at 25 "C and I = 0.1 Ma koli

Ligand

Me,edta Me,edtaMe,edta2Meedta3-

kOH

( W s - 1 )

2.10 0.86

0.35 0.06 ~~~~~~

a

Cornpiex

(M-' s-')

Cu(Me,edta)*+ Cu(Me3edta)+ Cu(Me,edta) Cu(Meedta)-

3.2 x 10' 8.0~10~ 1.5 x lo3 7.5 x 10'

Rare enhancemen! 1.6~10~ 8.Ox 104 4.4 x 103 1 . 2 io3 ~

~~~~

Data taken from R W. Hay and K. B. Nolan, J. Chem Soc, Dallon Trans., 1975, 1348

61.4.2.1.5 Palladium(I1) complexes

A number of recent studies have been carried out with palladium(I1) c o m p l e x e ~ , ~ of~the -~~ types (18) and (19). These mixed ligand complexes are formed in solution by the addition of the a-amino acid ester to the corresponding diaqua species, for example [Pd(en)(OH,),]'+. The kinetics of hydrolysis of the ester ligand in such mixed ligand complexes have been studied in detail. Nucleophilic attack by both OH- and OW, occurs. Typical kinetic data are summarized in Table 15. Substantial rate accelerations are observed in these systems for base hydrolysis. Thus for the ethylenediamine complex (18)rate increases of 4 x lo" for GlyOEt to 1.4 x lo7 for ethyl picolinate are observed.*' These rate accelerations are consistent with the formation of carbonyl-bonded species (18). The effects with methyl L-cysteinate and methyl L-histidinate are much Iess marked as such ligands can give mixed ligand complexes which do not involve alkoxycarbonyl donors. Thus in the case of methyl L-histidinate the complex (20) is formed. For these latter two esters only relatively small rate accelerations of 20-100 occur. For the chelate ester complexes, the ratios of kOH/kHzOfall within the range 3.8 x lo9 to 3.2 x 10". Such values for the relative nucleophilicity of water and hydroxide ion are comparable with those previously noted for copper(I1) complexes.82

Uses in Synthesis and Catalysis

424

Table 15 Hydrolysis Data for [Pd(en)(NH2CH(R)C02R']2* Complexes at 25 "C and I = 0.1 M

(KNW Esier

GlyOEt GlyOMe a-AlaOEt PheOEt CysOEt HisOMe picOEtb a

2.45 x io4 6.25 x io4 6.15 x io4 11.75X lo4 4.20 12.76 6.47 x lo6

5.3 x 4.9 x 1 . 6 io+ ~

I .04 x io+ 5.17X 5.48 X 2.05 x 1 0 - ~

0.64 1.28 0.55 0.24 0.04

0.62 0.46

Data taken from R. W. Hay and P. Banerjee, J. Cham. SOL,Dalton Trans., 1981, 362.

b .

picOEt = ethyl picolinate (2-carbethoxypycidine).

Although intramolecular attack by coordinated hydroxide ion (21) could occur in these systems, this pathway in not supported by the experimental evidence.86 The palladium(11)-promoted hydrolysis of methyl glycylglycinateand isopropyl glycylglycinate has been investigated over a temperature range." Complexes of type (22) are formed in which the amino, deprotonated amide and alkoxycarbonyl groups act as donors. Hydrolysis by both H 2 0 and OH- ion is observed. Base hydrolysis of the coordinated peptide esters is ca. 105-fold faster than the unprotonated peptide esters.

61.4.2.1.6 Kinetic stereoselectivity

The general topic of stereoselectivity in metal complexes of amino acids and their derivatives has been reviewed.B8Stereoselective effects have been observed in the base hydrolysis of some mixed ligand complexes involving amino acid esters. The complex [Ni( D-HisOMe)( his)]+ undergoes base hydr~lysis*~-~' some 40% faster than [ N~(D-HisOMe)(n-His)]+. Stereoselectivity is not observed for the analogous copper(I1) c o m p l e x e ~ . ~The ~ . ~formation ~ constants of the diastereoisomeric nickel complexes are identical? so that a kinetic stereoselective effect must operate. The [MEA]' complex (A- =the amino acid anion, E = the amino acid ester) probably involves tridentate A- and primarily bidentate E. The octahedral coordination of nickel(11) results in the possibility of a direct interaction between metal ion and ester carbonyl (23).Such an interaction can only occur when the two ligands are of opposite configurations as a trans arrangement of the bulky imidazole donors is referr red.^'-^^ The interaction must be very weak in view of the small rate differences between the diastereoisomers. Such transient coordination is less likely for copper(II), which favours a tetragonal geometry. In the complex [&EA]+, both E and A- are probably bidentate with possibly weak apical carboxylate coordination with A-. More detailed studies of reactions of this type have been reported.9697Nickel(I1) complexes of histidine and tryptophan provide stereoselectivity in the hydrolysis of histidine methyl ester, but stereoselectivity is not observed with nickel( II) complexes of aspartic acid or methionine. Only tridentate ligands with a minimum steric bulk appear to be capable of exhibiting stereoselectivity in reactions of this type.

Lewis Acid Catalysis and the Reactions of Coordinated Ligands OMe

425

1'

\

'0 (23)

The coordination of optically active amino acids and their methyl esters to nickel(I1) complexes of lY2-bis(2-(S)-aminomethyl-1-pyrrolidiny1)ethane(24; R = H) and lY2-bis(2-( SI-N-methylaminomethyl-I-pyrrolidiny1)ethane(24; R = Me) has been studied?* Some amino acidate ions coordinate stereoselectivity, as do their methyl esters, so that base hydrolysis of the esters proceeds stereoselectively. C N c H 2 c H 2 N a CHzNHR

CHzNHR

(24)

Hatano and Nozawa9' have observed stereoselective effects in the hydrolysis of D- and Lphenylalanine esters using poly-L-lysinecopper(11) complexes as catalysts and some aspects of the topic have been reviewed?'

61.4.2.1.7 Hydrolysis of amino acid amides and peptides

A variety of metal ions, i.e. nickel(II)y'OO~'O' copper(II),'" ~ o b a l t ( I 1 )and ' ~ ~palladium(II),'04~'05 promote the ionization of peptide (or amide) hydrogens and the to ic has been reviewed.'06 Promotion of peptide hydrogen ionization increases in the series Co < Ni" < CUI'< Pd11.104A number of bisdipeptide-cobalt( 111) complexes containing deprotonated amide groups have been investigated and their spectral properties studied.107-109 Protonation in acidic media occurs on oxygen rather than nitrogen to give the imino1 tautomer of the dipeptide ligand."*' '' Many studies have shown that the oxygen atom is the site of both protonation and metal ion coordination to the neutral amide function, while the nitrogen atom becomes the metal bonding site when ionization of an amide hydrogen occurs. A number of discussions are available on the general and the subject is currently reviewed.2B topic of metal peptide cornple~es'~~~''~~'~~-~~~ Early work1l6established that CUI', Ni" and Co" promote the hydrolysis of glycinamide in the pH range 9.35 to 10.35 at temperatures of 6.5 to 75 "C. Bamann and his collaborators carried out an extensive series of studies on the metal ion-promoted hydrolysis of peptides and related c~mpounds"~ and a review of this early work is available."' Highly charged ions such as thorium(1V) were found to promote the hydrolysis of leucylglycylglycine at pH values as low as 5. The thorium(IV) species is very extensivelyhydrolyzed at this pH and the reaction is presumably heterogeneous. Gel hydrolysis is effective at relatively low temperatures (37 "C),whereas observable effects were only obtained with such ions as copper(l1) at temperatures of ca. 70 "C. Bamann et a1'19also found that the hydrolysis of di- and tri-peptides is catalyzed by rare earth ions at pH 8.6. Cerium(1V) and Cell1 were particularly effective even at a temperature of 37 "C. The same reactions with La"' as a catalyst were much slower and only occurred at an appreciable rate at 70 "C. Many of these reactions merit further study. The copper(11)-promoted hydrolysis of glycylglycine has been studied in some detail.**' Copper(I1) ions catalyze the hydrolysis of glycylglycine in the pH range 3.5 to 6 at 85 0C.120The pH rate profile has a maximum at pH 4.2, consistent with the view that the catalytically active species in the reaction is the carbonyl-bonded complex. The decrease in rate at higher pH is associated with the formation of a catalytically inactive complex produced by ionization of the peptide hydrogen atom. This view has subsequently been confirmed by other workers,12' in conjunction with an IR investigation of the structures of the copper(I1) and zinc(I1) complexes l ~ ~zinc(II), nickel(I1) and manganese(l1) has also in D20 solution.Iz2Catalysis by ~ o b a l t ( I I ) ,and been studied. t24-126

R

426

Uses in Synthesis and Catalysis

Copper(I1) has been found to inhibit the hydrolysis of glycylglycine in basic solution (pH> Conley and MartinI2*have also found that, at pH values in excess of 11, copper(I1) inhibits the hydrolysis of glycinamide due to amide hydrogen ionization. Similar results were obtained with picolinamide, and a bis-picolinamide complex of nickel(11) containing deprotonated amide groups was isolated.”’ The above studies indicate that metal ions catalyze the hydrolysis of amides and peptides at pH values where the carbonyl-bonded species (25) is present. At higher pH values where deprotonated complexes (26) can be formed the hydrolysis is inhibited. These conclusions have been amply confirmed in subsequent studies involving inert cobalt(II1) complexes (Section 61.4.2.2.2). Zinc(I1)-promoted amide ionization is uncommon, and the first example of such a reaction was only reported in 1981.’03 Zinc(I1) does not inhibit the hydrolysis of glycylglycine at high pH, and amide deprotonation does not appear to occur at quite high pH values. Presumably this is one important reason for the widespread occurrence of zinc(11) in metallopeptidases. Other metal ions such as copper(I1) would induce amide deprotonation at relatively low pH values leading to catalytically inactive complexes. R

(25) active

(26) inactive

Polarization of the carbonyl group by the metal ion in the carbonyl-bonded complex (25) promotes nucleophilic attack at the carbonyi carbon by nucleophiles such as hydroxide ion and water. An interesting metal-promoted peptide hydrolysis has been observed during an investigation of Schiff base complexes of copper(11) with glycine, diglycine and trigly~ine.”~ Diglycine reacts with bis(saIicylaldehydato)copper(II) and bis(benzoylacetaldehydato)copper(II) to give the glycylglycinato Schiff base copper( 11) complexes. A similar reaction with triglycine at pH 4.5 gave, however, the copper(I1) complex of the diglycine Schiff base (27; Scheme 2) and free glycine. Subsequent has shown that hydrolytic cleavage occurred at the carboxyl end of the tripeptide, although not with 100% specificity.

cu

Scheme 2

The reactions of Cu(sal), (sal = salicylaldehyde monoanion) with glycinamide, glycine esters and the crystal structure of and the amide and esters of glycylglycine have been N-salicylideneglycinatoaquacopper(I1) tetrahydrate has been determined.”’ Reactions of this type merit detailed kinetic study. It should be possible to develop useful peptide-cleaving reagents using reactions of this general type.

61.4.2.1.8 Peptide bond formation

Peptide bond formation using non-labile cobalt(II1) complexes has now been developed to a useful synthetic level (Section 61.4.2.2.4), but few attempts have been made to use other metal centres. The formation of glycine peptide esters in the presence of copper(I1) has been noted.’33 Treatment of glycine esters with copper(I1) (other metal ions can also be used) in a non-aqueous solvent at room temperature gave di-, tri- and tetra-glycine peptide esters. After carbobenzyloxylation, the peptide esters were separated by column chromatography, and no evidence was obtained

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

42 7

for diketopiperazine formation. The polymerization of the methyl esters of amino acids uia their copper(11) complexes has also been r e ~ 0 r t e d . I ~ ~ Presumably these reactions involve intermediates of type (28) which undergo nucIeophiIic attack by a molecule of the ester leading initially to dipeptide ester complexes. These complexes then undergo further reactions of a similar type.

61.4.2.2

Non-labile Cobalt(II1) Complexes

The use of kinetically inert cobalt(II1) complexes has led to important developments in our understanding of the metal ion-promoted hydrolysis of esters, amides and peptides. These complexes have been particularly useful in helping to define the mechanistic pathways available in reactions of this type. Work in this area has been the subject of a number of review^.^'-^" Although most of the initial work was connected with cobalt(III), investigations are now being extended to other kinetically inert metal centres such as Rh"', lr'" and Ru"'. It is convenient to discuss these reactions under the following headings: ester hydrolysis, peptide (amide) hydrolysis and peptide bond formation; but it should be remembered that similar intermediates occur in each case and the discussion overlaps the three sections.

61.4.2.2.1

Ester hydrolysis

Alexander and B ~ s c h " ~first described the preparation of complexes of the type cis[CO(~~)~(CH,CH~CO~R)C (29) I ] CbyI ~reacting the appropriate amino acid ester hydrochloride with trans-[Co(en),Cl,]Cl in aqueous solution, the free amino acid ester being generated in situ by the presence of a weakly coordinating base such as Me2NH. CI

1*

f

&jljl,CH,CO,R

c1

NH2

Me,NH.H,O

H~N,

12+ I, NH1CHzCOzR

NH,

ACO\

W.J,W I

CI (29)

Hydrolysis of the N-coordinated amino acid ester complexes with 4 M HCl leads to the corresponding complexes containin the N-coordinated amino in acidic solution (equation Mercury(I1)-promoted h y d r ~ l y s iof j ~cis-[Co(en),X(GlyOR)j2* ~ 16) gives [Co(en),(Gly)]", and it was suggested that a chelated ester species [Co(en),(GtyOR)]'* (30) was the reactive intermediate in the Hg"-promoted reaction (Scheme 3). cis-[C0(en),X(GlyOR)]*++Hg*~~+H~0 -t [C~(en),(Gly)]~'+ROH+H++HgXt (X = C1, Br; R = Me, Et, hi)

(16)

Evidence in support of this mechanism was provided by changes in the C=O stretching frequency as the monodentate ester (1735 cm-') was first chelated (1610 cm-I) and then hydrolyzed to [ C ~ ( e n ) ~ ( G l y )(1640 ] ~ + cm-I). The reaction presumably proceeds via the five-coordinate species (31), and the ester carbonyl oxygen competes so effectively with solvent water for the vacant coordination site that the chelate ester is formed exclusively. It was subsequentIy found that treatment of ~is-[Co(en),(GlyOR)Cl]~+with AgClO, in acetone'3' gave [Co(en),(GlyOR)](C10,), (30). This complex can be used for the synthesis of O ~ ) ~amino acid or peptide esters peptide esters.''%Thus treatment of [ C ~ ( e n ) ~ ( G l y O M e ) l ( C l with in anhydrous sulfolane, dimethyl sulfoxide or acetone solution gave the [Co(en),(peptide-OR)J3+

Uses in Synthesis and Catalysis

428

ion (32).Kinetic data for the hydrolysis of [Co(en),(GlyOPr')](C10,), have been reported.'39 For the pH-independent attack by water in acidic solution k = 1.15 x s-' at 25 "C, and evidence for base hydrolysis at pH 8.5 was obtained. The pH-independent rate constant was identical to that obtained in the second step of the Hg"-promoted hydrolysis of ~is-[Co(en)~Cl(GlyOPr')]~+. These kinetic results thus confirm that a chelated ester intermediate is formed following the Hg"-promoted removal of halide ion from the chloropentamine complex (29). Isotope studies using I8O establish that hydrolysis of the chelated ester occurs without rupture of the chelate ring. A similar chelated ester intermediate is generated following HOC1 oxidation of the coordinated bromide in [Co(en),Br(GlyOR)I2+ but some of the aqua complex [Co(en),(H,O)( GlyOR)lfCis also produced in this reaction. /NH2

HxN,

1 ,NH2,

,Co,

H 2 N LI N H z

o=c

13+

CH,

I/ \

+ :NH,CH,CO,R

-

,.fNHz HzN, ,NH2,

iH2

HzN*NH2 I

O=

OMe

13+

I

c,

NHCHzCOzR

(32)

The rates of hydrolysis of chelated ester species are independent of pH in the range 0-4.'",'39,'40 This result is expected as coordination to Co"' prevents protonation of the carbonyl oxygen and consequent acid hydrolysis. Base hydrolysis of the chelated isopropylglycine in [Co(en),(GlyOPri)J3'( koH = 1.5 x lo6 M-' s-' at 25 "C) is about 106-fold faster than hydrolysis of the free unprotonated ester.'" The rate enhancement is entirely due to a more positive AS* in the chelated ester species.'4oz141 Detailed kinetic investigations have subsequently been made142of the hydrolysis and aminolysis of [C~(en),(GlyOPr')]~+ (Scheme 4). It was found that over the pH range 0-7 in buffer solutions kobs= kH2,+ koHIOH-] + k,[B], where B is the buffer base. Both oxygen bases (e.g. acetate) and nitrogen bases acted as nucleophiles rather than as general bases; the nitrogen nucleophiles formed stable chelated amides or peptides. Somewhat surprisingly it was concluded that kl is rate determining in all such substitution reactions even though (33) is highly activated towards nucleophilic attack and contains a poor leaving group. The reaction of (33) with glycine ethyl ester in DMSO solution has also been studied kinetically and the tetrahedral intermediate (34)detected spectroph~tometrically.'~~ It has been suggested'& that high concentrations of C10,- alter the mechanism of the Hg"(X = C1, Br), with high Clod- concentrations promoted hydrolysis of ci~-[Co(en)~X(GlyOEt)]~'

k,lfast

Scheme 4

429

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

(>4 M) giving a significant amount of [C0(en)~(H,O)(Gly0Et)]~'which only slowly ( t i = 23 min)

reverted to the chelated ester intermediate with loss of bound water. As such high Clodconcentrations would be expected to significantly reduce the free water concentration these observations seem questionable. Buckingham and WeinI4' have subsequently shown that C104does not alter the mechanism, but alters the rate, and in the presence of Hg" the reaction proceeds exclusively via the chelated ester intermediate. ions led to interesting developments. Studies of the base hydrolysis of cis-[C~(en),(GlyOR)X]~* Base hydrolysis of halopentarnine complexes of cobalt( 111) occurs by an SNICB mechanism leading to a five-coordinate intermediate. In the N-bonded ester complexes of the [Co(en)2(GlyOR)C1]2+ type, the ester carbonyi group and solvent water can compete for the vacant site in the five-coordinate intermediate (Scheme 5) giving rise to the chelated glycine ester species and the hydroxypentamine respectively.

12+

1+

fast

+OH-

I

I

c1

CI

NHzCHnCOzR

I

J

H,O(fast)

I

+HZ0

1z+

\

Hf(fast)

im

Tracer s t ~ d i e s ' ~using ' **Ohave established that the two main pathways to the chelated glycine complex (which is not the sole product) are (a) internal nucleophilic attack by bound hydroxide ion on the N-coordinated ester and (b) attack of 'external' hydroxide on the chelated ester. Further investigation^"^ have dealt with the Hg"-promoted hydrolysis and base hydrolysis of the P,-[C~(trien)Cl(GlyoEt)]~+ in (35). In this case, base hydrolysis gives only chelated glycine products, and ''0tracer studies confirm that some 84% of the ~,-[Co(trien)Gly]*+arises via the chelated ester and the remainder by intramolecular nucleophilic attack by coordinated hydroxide.

Uses in Synthesis and Catalysis

430

' HNANH~ ( HN,

I ,,NH1CH2CO2Et

YC0\

HjN,

NH3 ,,NH,CH,CO,Et

I

,.p,

H,N

1

3+

+OHNH3

NH3

H3N,

e

12+

NH, ,NH2,

I

CH2

+ H2O

c-0 1

H,N./'PN= NH3

I

OEt

(36)

An interesting example of intramolecuiar attack is seen147 in the base hydrolysis of [Co(NH,),NH,CH,CO,R]". Base hydrolysis in the pH range 9-14 gives both [Co(NH-,),( NH2CH2C02-)]*+, containing the monodentate glycinate anion, and [Co(NH3)J NH2CH2CONH)I2+, with glycinamide bonded uia both nitrogen atoms. The formation of chelated glycinamide can be rationalized in terms of intramolecular attack within the amido complex (36). Similar intramolecular effects are seen in the base hydrolysis of [c~(en)~Br(NH,cH,Br)]~' which leads to N-coordinated a z i r i d i n e ~ . ' ~ ~ Intramolecular effects are only observed with N-coordinated esters of glycine and p-alanine which can give rise to five- and six-membered chelate rings. Base hydrolysis of cis[COC~(~~)~(NH,(CH,)~CO~M~)]~+'~~ and cis-[CoCl(en),( NH2(CH,),C0,Me)l2* 150 has been studied. For the first complex, initial chloride hydrolysis is followed by a slower base hydrolysis of the pendant ester function to give the hydroxypentamine, and similar results were obtained with the second complex. Base hydrolysis of the N-coordinated NH2(CH2)3C02Meis only ca. seven times faster than for NH2(CH2)3C02Me.Coordination to cobalt(II1) via the nitrogen has somewhat similar effects to protonation of the amino group. (X = C1 or Br) in which the ester Base hydrolysis of c~~-[CO(~~),X(NH,CH,CH,OCOM~)]~+ of an amino-alcohol is employed as the ligand has also been studied.'51 A two-step hydrolysis is observed, the first invoiving C1- or Br- loss and the second ester hydrolysis. It is noteworthy that N-coordination to cobalt prevents the rapid base-catalyzed isomerization of 2-aminoethyl acetate to 2-acetylaminoethanol. The various pathways by which esters, amides and peptides can undergo hydrolysis can be summarized as shown in Scheme 6.lS2Reaction 1 (the 'metal carbonyl' mechanism) leads to rate enhancements of 104-106for all substrates independent of the leaving group Y. The rate enhancement observed in reaction 1 is due solely to a more favourable entropy term, Le. a more positive AS' for the promoted reaction. Reaction 2 is only effective with more reactive species (CO,, anhydrides, aldehydes and esters with good leaving groups such as 2,4-dinitrophenyl acetate'', the hydrolysis of alkyl esters of amino acids, amino and 4-nitrophenyl a ~ e t a t e ' ~ ~ As ' ' ~a~result ). acid amides and simple peptides has not been observed in a bimolecular 'metal hydroxide' process. Both AH' and AS' contribute to the rate acceleration in reaction 2 . Reaction 1 also leads to nucleophilic attack by nucleophiles other than OH- (e.g. NH,R, ROH, H20) and general acid or genera1 base catalysis can occur.13x,140~142 Only hydrolysis is observed with metal hydroxides and general acid or general base catalysis has not been 0 b ~ e r v e d . l ~ ~ Reaction 3, the intramolecular counterpart of reaction 2, has been observed for amino acid esters,'56 amides'" and nitrile^,'^' where five- and six-membered chelate rings can be formed. In the case of the aminoacetonitrile complex (37) a rate enhancement of ca. 10" occurs at pH 7, and For reaction this may be compared with an acceleration of ca. lo6 for the reaction 1 analog~e."~ 3, AH* values become of considerable significance. In the case of amino acid ester and amide complexes, the intramolecular hydrolysis reaction was not observed directly, but was deduced from the results of "0 tracer studies. However, recently the cis-hydroxo and cis-aqua complexes derived from the bis(ethylenediamine)cobalt( 111) system, containing glycinamide, glycylglycine and isopropylglycylflycinate, have been isolated and their subsequent cyclization studied over the pH range 0-14.'6 ,16' Other related studies have dealt with the base hydrolysis'62 of cis-[Co(en),X(GlyO)] ' ions (X = Cl, Br) and oxygen exchange and glycinato ring opening in [ C ~ ( e n ) , ( G l y O ) ] ~ ~ . ' ~

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

43 1

Reaction I

\I

-co-o=c 1 1

/Rl-+ ?*H \

Y

--+

or OH2

\ I / I

-Co-Or-C

'\

+ YH

'0

Reaction 2

Reaction 3

OR Scheme 6

61.4.2.2.2 Peptide and amide hydrolysis A number of complexes of the general type (CoN,(OH)(OH,)]*+ (N4= a system of four nitrogen donors) stoichiometrically cleave the N-terminal amino acid from di-or tri-peptides. Reactions have been described for N, = en2r164-165 trien1W167 and tren. 168~186In the case of trien complexes, the proposed mechanism for peptide hydrolysis is shown in Scheme 7. Hydrolysis can occur by two pathways: (a) attack by external hydroxide on the 0-bonded chelated peptide, and (b) intramolecular attack of coordinated hydroxide on the N-bonded peptide.

[Co(trien)(H,O)(OH)]*+

[Co( trien)( Gly)]'+

+ NH2CH,CONHCH,CO-Pep

+ NH,CH,CO-Pep

A o=c, NH2CH2CO- Pep Scheme 7 Proposed mechanisms for peptide hydrolysis

A variety of N-0-chelated glycine amide and peptide complexes of the type [ C O N , ( G I ~ N R * R ~have ) ~ ~been + prepared and their rates of base hydrolysis The kinetics are consistent with Scheme 8. Attack of solvent hydroxide occurs at the carbonyl carbon of the chelated amide or peptide. Amide deprotonation gives an unreactive complex. Rate constants kOH are summarized in Table 16. Direct activation of the carbonyl group by cobalt(TI1) leads to rate accelerations of ea. 104-IOb-fold. More recent investigations160*161 have dealt with

432

Uses in Synthesis and Catalysis Table 16 Rate Constants for the Base Hydrolysis of Ester, Amide and Peptide Bonds in Various Cobak(II1) Complexes (25 “C, I = 1.0 M)”

1.5 x lo6 25 1.6 1.1 2.6 2

5 3 3 from D. A. Buckingham, C. E. Davis, D. M. Foster and A. M. Sargeson, J. Am. Chem. SOC, 1970.92.5571.

a Data

intramolecular hydrolysis of glycinamide and glycine dipeptides coordinated to cobalt( 111). The intramolecuIar addition of cobalt( 111)-bound H,O and OH- to glycinamide, glycylglycine isopropyl ester and glycylglycine in complexes of type (38) has been investigated in detail.I6’ times more rapidly than the free dipeptide (pH 5 ) The aqua dipeptide complex hydrolyzes l o L 1 and a factor of lo7 is involved for the hydroxo dipeptide complex at pH 8. Coordinated water is more reactive than coordinated hydroxide mainly due to a more positive value of AS”. An extensive review of this topic is available.23Also see reference 21. rNH2 H2N, ,NH,CH,CONHR

1’+I*+

(38) R = H, CH,CO,W, CHZCO,‘

Some of the difficulties in the use of p-[C~(trien)(OH)(OH~)]’~ for the sequential analysis of peptides have been d i s ~ u s s e d . ’ An ~ ~ -in~estigation”~ ~~~ of the relative effectiveness of the complexes cis$- [Co(trien) (OH)(OH2)I2+, cis-a-[Co(trien)( OH)(OHZ)l2’, [Co(en)z(OH)(0H2)I2”, cis-[Co(tren)(OH)(OH,) J and [CO(NH&,(OH)(OHJ~+ in the hydrolysis of glycyl dipeptides and dipeptide esters has shown that the cis-~-[Co(trien)(OH)(OH,)]’+ species is at least 50 times more effective than any of the other complexes. The [Co(edda)(OH)(OHz)] ion has been r e p ~ r t e d ”to~ be as effective as cis-p-[Co(trien)(OH)(edda = ethylenediaminediacetate) in the hydrolysis of L-Ala-L-Phe, and a variety of other ligand systems have been eva1~ated.I~~ Gillard and P h i p p ~ ”established ~ that [Co(dien)X,] complexes reacted with triglycine to give [Co(dien)(GlyGly)lf and free glycine. In this case, the two N-terminal amino acids can be subsequently described the preparation of a number removed in a single step. Wu and B~sch’’~ of complexes of the type [CoX(dien)(GlyGlyOR)]’+ and [CoX(dien)(GlyO)]*+ (X = Cl, NOz). Peptide bond formation (Section 61.4.2.2.4) was shown to occur using the three-site moiety [Co(dien)13+.Girgis and Legg’79established that [Co(dien)(OH)(OH,)] cleaves the peptide bond of a-L-aspartylglycine, and also studied the [C~(trien)(OH)(OH~)]~+-prornoted hydrolysis of the diethyl ester of L-aspartic acid and dipeptides containing aspartic acid, glutamic acid and methionine. The [C~(dien)(GlyGlyOR)X]~+ complexes prepared by Wu and Bus~h’~’ have the configuration (39), in which the dien ligand adopts a mer configuration and the ligand X lies trans to oxygen. A series of rrans-(0,X)-[CoX(dien)(AA)]+ complexes (AA- =amino acid anion, X = C1, NO2, CN) have been prepared. Complexes such as (39) are likely intermediates in the reaction of [Co(dien)(OH)(0Hz)J2+ with dipeptides. Hay and Piplani“’ have studied the kinetics of base hydrolysis of a variety of complexes of type (39)and have determined values of k, for hydrolysis of the peptide bond and the NOz and C1 ligands (Table 17). The rate constant koH for peptide bond hydrolysis falls within the range 0.7 to 7 M-’ s-’ and is some lo4to 10’ times greater than that for the uncomplexed peptide.

Lewis Acid Catalysis and the Reactims of Coordinated Ligands

433

~HCH~CO~R (39)

Table 17 Base Hydrolysis of the Peptide Bond and the NOz and C1 Ligands in Co(dien) Complexes at I = 0.1 M and 25 "C" ~

Initial complex

kpeptide on

[Co(dien) (GlyElyO)NOz]+ [Co( dien)(GlyGIyGlyO)NOz]+

0.88 0.67 0.68 7.0

[Co(dien)(GlyGlyOEt)N02]'+ [CoCl(dien) (GlyGlyOEt)]'+ [CoCI(dien)(GlyO)]+ [CoCl(dien)( GlyNH,)]" [Co(dien)(GlyO)NO$ a

~~~

kggz (M-'s-')

(W's-l)

k&

(M-'s-')

2.5 x lo-' 1.1 x 106 1.3 x IO4 5.85 x io5

-

-

2.45 x lo-'

For base hydrolysis of glycylglycine, k , = = 4 x lO-'M-'s-' at 26°C. Data from R. W. Hay and D. P. Piplani, Kern. K o d 1977, 48, 47. For additional data see R. W. Hay, V. M. C. Reid and D. P. Piplani. Trunsition Met. Chern., 1986, 11, 302.

The synthesis and base hydrolysis of [M(NH,),DMFJ3+ containing 0-bonded DMF (M = 184 and Irl[1 184 ) has been studied in detail. Two reaction pathways are observed with the cobalt(II1) complex. The major pathway corresponds to hydroxide ion attack at the carbonyl centre giving [Co(NH,),OOCH]*+ and HNMez. The minor path leads to [Co(NH3),0H]'+ and DMF, and is the dissociative conjugate base process. Amide cleavage is accelerated at least lo" times due solely to a more favourable entropy term. Hydrolysis of the Rh"' and Ir"' complexes corresponds to equation (17). In this case there is no evidence for any ligand loss by a conjugate base mechanism to give free DMF and the pentaamminehydroxo species. The kinetics fit the rate equation kobs= k,[OH] k2[OHI2. For the k, pathway the rate S'. Values of AH' are enhancement (ca. 106-fold) occurs due to a more positive value of A essentially the same as for the free ligand.

~0111,183

+

[(NH,),MOCHNMe2]3++OH-

+

[(NH,),MOOCH]'++HNMe,

(17)

The work described above suggeststhat carbonyl-bonded amides and peptides when coordinated to Co"', Rh"' and Ir"' will undergo base hydrolysis e a lo6 times faster than the free ligand, the rate acceleration arising primarily from a more positive value of AS*. At high pH, amide deprotonation occurs (Scheme 8) leading to catalytically inactive complexes. Much higher rate accelerations can be obtained if intramolecular attack by coordinated water or hydroxide ion can take place (ca. lo'-10"-fold). Recent work'6o also clarifies a previous report on the hydrolysis of ci~- [Co( en)~Br( GlyNH~) l~+. In the initial it was concluded that the hydroxo grycinamide ion reacted rapidly under the conditions of base hydrolysis and the subsequent process observed spectrophotometrically

1%

?HI.

CHz I (enIzCo '04, NHR

/NHZ-CH2

I*&

I

'C) , ( O H ) ( G I ~ N H in , ) ]addition ~+ to the chelated amide. Base hydroIysis of 0-coordinated N,N-dimethylformamide in the complex [Co([ 15]aneN5)DMFl3+(40)18' is accelerated some 327-fold compared with the free ligand. The effect is considerably less marked in the macrocyclic complex when comparisons are made with the [Co(NH,),(DMF)J2+ complex (rate acceleration CQ. lo4). The Lewis acidity of Co"' is presumably reduced by the stronger r-donors of the macrocyclic ligand.

q+

1

3'

N\ ( N&NH H P , I

I

0 >C-N
IO4 compared to the non-catalyzed reaction. Thl 0

II

cx,, -0-F- 0I

I

N H

Lewis Acid Catalysis and the Reactions of Coordinated Ligands 0

445

q;

It

0% ,O-

-0-p-o-

N H (74)

(75)

transition state in the copper(I1)-promoted reaction has been formulated as (74) or (75). In (74) the copper(I1) ion acts as a more effective acid catalyst than a proton, lowering the pK, of the leaving group so that facile hydrolysis of the dianion (generally observed only with leaving groups of p K , < 7) may be observed. In (75) the copper(I1) is expected to induce strain in the P-0 bond and/or partially neutralize charge on the phosphate, leading to nucleophilic displacement by solvent on phosphorus. Copper(I1) ions have been observed to catalyze the hydrolysis of a number of other phosphate monoesters including salicyl phosphate (76),2683269,274 2-pyridylmethyl phosphate (77)270and 8-quinolyl phosphate (78).27*,272,274 The catalytic effect observed with these esters was apparently quite small (ca. 10-fold). However, the uncatalyzed reactions of esters such as salicyl phosphate are subject to intramolecular general acid catalysis by the neighbouring carboxyl group.273When allowances are made for this catalysis it can be shown that copper(I1) ions exert a very marked catalytic effect on the hydrolysis of the normally unreactive phosphate monoester dianion of salicyl phosphate (ca. IO* rate a~celeration).'~~ Intramolecular general acid catalysis occurs via the transition state (79)?73Cleavage of the P-0 bond which is well advanced is assisted by general acid catalysis by the neighbouring carboxyl group, although the proton transfer has scarcely begun. The formation of a complex such as (80) should lead to a similar type of transition state.

0

CH,-0-P-OH I

0

I

O=P-OH I

OH (77)

II

0 (79)

The phosphate esters (81) and (82) are also subject to catalysis by metal ions275,27h and possible reactive complexes are illustrated. Metal complexes of adenosinediphosphoric acid and adenosinemonophosphoric acids have been and the effect of divalent metal ions on the hydrolysis of ADP and ATP has been investigated.2787279

446

Uses in Synthesis and Catalysis

As much of biological phosphate chemistry appears to be subject to metal ion catalysis there is a great incentive to clarify and define much of the chemistry described above.

61.4.4.2

Cobalt(II1) Complexes

In recent years there has been considerable interest in the hydrolytic activity of cobalt(II1) complexes of phosphate esters. This approach has been adopted in order to avoid some of the mechanistic complexities which can arise with kinetically labile metal centres, Some developments in this field have been recently discussed by Dixon and Sargeson.zl Farrell et ~ l . " appreciated ~ that chelation of ROP032'"to cobalt(II1) ta give a four-membered chelate ring should activate the phosphorus atom towards nucleophilic attack by strain induction. Metal ion binding as in (83) could compress the OPO angle (Y in the ground state and so reduce the activation energy necessary for attainment of the trigonal bipyramidal intermediate (or transition state) in phosphoryl transfer. Similar arguments have been used to account for the high reactivity of five-membered cyclic phosphates such as ethylene phosphate. 241,281 Earlier work by Lincoln and Stranks282had shown that the phosphate complex [Co(en),P04] 'existed in rapid reversible equilibrium with the monodentate complex [Co(en),OH(HPO,)]. An X-ray studyzs3of the crystal structure of [Co(en),PO,] establishes that the ring OPO angle (84) has been deformed from the unstrained tetrahedral angle of 109.5' to 98.7", similar in magnitude to that found in methylethylene phosphate2s4(99.1") and cytidine 2,3-cyclic phosphate (95.8°).285

(83)

(84)

Sargeson and ~ o w o r k e r initially s ~ ~ ~ reported that 4-nitrophenyl phosphate in the complex (85) underwent base hydrolysis some lo9 times faster than the uncoordinated ester. Subsequent X-ray work286has now shown that the complex (85) and the analogous ethylenediamine derivative do not contain chelated phosphate esters but are dimeric species (86) with a surprisingly stable eight-membered chelate ring. The reactivity of such bridged complexes in basic solution has been (87) has recently discussed in a review.21The synthesis of cis-[Co(en),(OH2)0,P~c6H4No~]+ been described2" and its hydrolysis studied over the pH range 7-14 by 31PNMR spectroscopy and by monitoring nitrophenol release at 400 nm. Intramolecular attack by '*O-labelled coordinated hydroxide gives initially a five-coordinate phosphorane which decays to [Co(en),P04] and nitrophenol (kbs=7.8xlOP4s-' at 25°C in the pH range 9-11.8). The hydrolysis is lo5 times faster than that of the uncoordinated ester under the same conditions. At pH 10 there is evidence for " 0 exchange between solvent and the five-coordinate phosphorane and therefore Co-OH addition and ester hydrolysis are not concerted.

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

447

The base hydrolysis of [CO(NH~)~O~POC,H,NO,J+ has also been studied.'** In this case a coordinated aminato ion is the intramolecular nucleophile (88). The hydrolysis of ~ , ~ - [ C O ( N H ~ ) ~ H(89) , P ~isO greatly ~ ~ ] accelerated in the presence of cis[Co(cy~len)(OH~)~ (Cyclen ] ~ + = 1,4,7,10-tetratazacyclododecane)and the intermediate (900)has been suggested as the active complex.28931PNMR studies have now provided evidence in support of such an hydrolysis occurring via intramolecular attack by coordinated hydroxide.

H

9

0

The preparation of cis-[Co(cyclen)P04] has been described and its reactivity in acidic and basic solution studied.z91This complex undergoes rapid ring opening in acidic or basic solution to give the monodentate phosphato species. Loss of monodentate phosphate in acidic solution follows a rate expression of the form kobr= k,+ kHE,EH'] where ko = 5 x lop4s-' and kH = 4.05x lo-' M-' s-' at 25 "C ( I = 0.49 M). Loss of monodentate phosphate from the complex [Co(cyclen)(OH)(OPO,)]- also takes place in basic solution, and the reaction shows a first order dependence on the hydroxide ion concentration with bH = 2.7 x lo-' M-' s-' at 25 "C and I = 0.49 M. Similar studies with [Co(en),PO,] are also described in the paper. This work provides supporting evidence for the participation of the intermediate (90) in the [Co(cy~len)(OH~)~]~*promoted hydrolysis of P - ~ [ C O ( N H ~ ) ~ H ~ The P ~ Omonodentate ~~]. intermediate will be sufficiently long lived in solutions of pH 8-11 to participate in the reaction. The [(NH3)5CoOP(OMe)3]3+ion has recently been shown292to react with SCN-, I- or SzO;to produce (NH3)5C002P(OMe)z]z+ and respectively MeSCN, Me1 or MeS203-. The reactions are believed to occur by SN2substitution at carbon. This establishes a new mode of reaction for a coordinated phos hate ester. The rate enhancement on coordination is ca. 150. The Cr"' and Co8 1 complexes of ATP (adenosine 5'-triphosphate) have been shown to be very good analogues for MgATP. The hydrolysis and decomposition of these complexes have now been studied in some ~letai1.2'~ The breakdown of the tridentate [Co(NH,),ATP] is rapid, producing high levels of free ATP and lesser amounts of ADP. Rate constants for hydrolysis are 100-5000 times greater than those for uncomplexed ATP. Polyphosphates such as pyrophosphate and triphosphate are interesting ligands. Recently the unidentate and bidentate pyrophosphate complexes [Co(NH3)5HPzQ7] H20294and [Co(NH3)4HP207)-2H20 have been characterized and their crystal structures determined. The cu,P,y-tridentate [Co( NH3)3(H2P3010)]complex contains two fused six-membered chelate rings formed by the facial coordination of one 0 from each of the three phosphate residues.295The 0 the a,y-bidentate triphosphate ligand has an complex [Co(NH& H z P 3 0 1 0 ) ] ~ H zcontaining eight-membered chelate ring in a boat conformation stabilized by two interligand H bonds from the axial ammines above and below the chelate ring to the p- and y-ph~sphates.'~~ The rate of hydrolysis of bidentate triphosphate in [Co(NW,),H,P3010] has been studied by phosphomolybdate analysis and 31PNMR.z97,z9sBoth the X-ray crystal structurezw and the 31PNMR spectrum of ~ C O ( N H ~ ) ~ Hare ~ Pconsistent ~ O ~ ~ with ] structure (91) in which one terminal phosphate residue is not bonded to the metal centre. Kinetic studies establish that hydrolysis of the chelated ligand occurs at some two thirds of the rate for the free ligand. Hubner and M i l b ~ r n ~ ~ have noted large rate increases of ca. lo5 for cobalt(II1) complexes with a 3 :1 metal: pyrophosphate stoichiometry. 0

0

II It HO-P-0-P-0I I 0-

0,

0

I/

P-OH

p1

COWHA (91)

448

Uses in Synthesis and Catalysis

The hydrolysis of adenosine 5'-triphosphate (ATP) in the presence of various cobalt( 111) complexes has been ~tudied.~''Complexes such as [Co(en),13' which have no available sites for coordination o l the substrate display no catalytic activity. Complexes having one site or two sites in a trans configuration such as tetraethylenepentaminecobalt(II1) or bis(dimethylg1yoximato)cobalt(III) slightly enhance ATP hydrolysis. However, complexes with two available exhibit considerable activity. Both sites in a cis configuration such as cis-a- or ~is-p-Co(trien)~' the reactions ATP+ H 2 0+.ADP+ PI and ATP+ H 2 0+ AMP+ PPI occur with these systems. The complex [Co(dien)13+ effectively enhances the hydrolysis of ATP to ADP+Pi. At pH 4.0 the uncatalyzed hydrolysis rate constant for ATP hydrolysis is 1.18 x lop6s-' at 50 "C. For ATP (1 x lo-, M) and [Co"' (dien)13' (2 x lop3M) at pH 4.0, kobs= 1.75 x s-' at 50 "C, a rate enhancement of 1SO-fold. The substitution-inert complexes of chromium( 111) and ATP prepared by DePamphilis and Cleland302have been used with considerable success to elucidate the kinetic mechanism of yeast hexokinase303and other enzymes.304Cornelius and have recently isolated and characterized the complexes [Co(NH,),HATP] ( n = 2, 3 or 4) and [Co(NH,).ADP] ( n = 4 or 5) in addition to the complexes [Co(en),HP207], [Co(NH3).HP207] ( n = 4 or 5 ) and [Co(NH3),H2P3Ol0] ( n = 3 or 4). The 31PNMR spectra of the simpler phosphato complexes provide definitive evidence of pyrophosphate both as a monodentate and as a bidentate ligand, and of tripolyphosphate both as a bidentate and as a tridentate ligand. A correlation between OPO bond angles and the 31PNMR chemical shift for a number of phosphate esters has been noted.306If such a correlation could be extended to cobalt(II1) complexes, "P NMR would be a powerful tool for determining the solution geometry of phosphates bound to cobalt. Clearly important results and developments are to be expected in this experimentally difficult area.

61.4.4.3 Metal Hydroxide Gels

A number of studies have been reported on the hydrolysis of phosphate esters by metal hydroxide In view of the heterogeneous nature of these reactions it is difficult to come to firm gels.307-309 mechanistic conclusions; however, it seems likely that metal-bound nucleophiles are probably involved. Bamann3" first noted that the hydroxides of lanthanum, cerium and thorium promoted the hydrolysis of a-glycerophosphate in the pH range 7-10 and suggested that the reaction could be regarded as a model for the metal-containing alkaline phosphatases which cleave phosphate esters around pH 9. When the ester is adsorbed on the hydroxide gel, the cation neutralizes the negative charges on the phosphate dianion ROP032- and allows more facile attack by hydroxide ion. Butcher and Westheimer3'" have investigated similar reactions of this type and have found that the process resembles the enzymatic reaction in that cleavage occurs exclusively at the phosphorus-oxygen bond. The reaction does not occur readily unless the phosphate ester is substituted in the p-position, so that the hydrolysis of ethyl phosphate is not greatly promoted by lanthanum hydroxide gel, but the hydrolysis of p-methoxyethyl, p- hydroxyethyi and paminoethyl phosphates is strongly catalyzed at pH 8.5 and 78 "C. The hydrolysis of P-methoxyethyl phosphate is regarded as occurring by the steps shown in Scheme 16. The ester 1-methoxy-2-propyl phosphate is hydrolyzed at pH4, or by La(OH)3 at pH 8.5 with complete retention of stereochemical configuration and P-0 bond cleavage. Catalysis by rare earth ions has been discussed by Trapn~ann.~"Scheme 14 was suggested prior to the recognition of the high reactivity of metal-bonded OH species in hydrolysis. Hydrolysis via a complex of type (92) could provide a very reactive pathway.

HzC-

9

HL.,

,La,

I

0" I

Me

P03-+H,0

-

I

Me

H,PO,'~

Scheme 16

+

+ Po,OH

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

449

61.4.5 REACTIONS OF COORDINATED NITRILES

In recent years there has been considerable interest in the reactions of nitriles in the coordination first reported that the hydrolysis of 2-cyano-1,lO-phenanthrosphere of metal ions. Breslow et line to the corresponding carboxamide is strongly promoted by metal ions such as copper(II), nickel(I1) and zinc(1I). Base hydrolysis of the 1 : 1 nickel complex is lo7 times faster than that of the uncomplexed substrate. The entire rate acceleration arises from a more positive value of A S '. Somewhat similar effects have been observed for base hydrolysis of 2-cyanopyridine to the corresponding carboxamide. In this case rate accelerations of lo9 occurred with the nickel(11) complex.313 A number of studies have also been made of the hydrolysis of nitriles in the coordination sphere of cobalt( 111). Pinnell et aL314found that benzonitrile and 3- and 4-cyanophenol coordinated to pentaamminecobalt(111) are hydrolyzed in basic solution to the corresponding N-bonded carboxamide (equation 22). The reaction is first order in hydroxide ion and first order in the = 18.8 M-' sC1 at 25.6 "C for the benzonitrile derivative. As kOH for the base complex with kIl hydrolysis of benzonitrile is 8.2 x lop6M-' s-' at 25.6 "C, the rate acceleration is ca. 2.3 x 106-fold. The product of hydrolysis is converted to [(NH,),CoNH,C0Phl3+ in acidic solution and the pK of the protonated complex is 1.65 at 25 "C. Similar effects have been observed with aliphatic nitrile^.^" Thus, base hydrolysis of acetonitrile to acetamide is promoted by a factor of 2 x lo6 on coordination to [Co(NH3),I3+. 0

II

[(NH3),CoN=CPhl3++ OH

3 [(NH3),CoNHCPh]'+

(22)

Particular emphasis has been directed towards comparison of ligand properties and reactions of different organonitrile complexes of the type [M(NH,),(NCR)]"+ where M = Ru", Ru'", Rh"' or Ir"' 337 Specific rates of base hydrolysis of the pentaammineruthenium(111) complexes of

.

acetonitrile ( koH = 2.2 X lo2 M-'s-' at 25 "c) and of benzonitrile ( koH = 2.0 x lo3 M-' s - I ) are some lo8 times faster than the free ligand, and about 10'times faster than the analogous cobalt(1II) complexes.316The reaction products are the corresponding amido complexes which reversibly protonate in acidic solution to give the amide complexes. Base hydrolysis of the pentaammineoccurs at a rate comparable to that rhodium(II1) complex of acetonitrile (koH= 1.0 M-'s-1)316*337 of the analogous cobalt( 111) complex, while the ruthenium( 11) complex is at least lo6 times less reactive than the corresponding ruthenium(111) complex. Representative kinetic data for these reactions are summarized in Tables 20 and 21. Table 20 Rate Constants ( koH) and Activation Parameters for the Base Hydrolysis of Nitriles to Carboxamides Substrate ~

1.6 x 8.2 x 3.4 1.o 2.2 x 102 0.23 HgC13- (0. I) >> HgCI4'- (10.001).The rate constant for reaction via Hg" is cu. lO'-fold larger than for hydrolysis in the presence of H 3 0 + alone. Addition of HCl leads to a reduction in rate, rather than an increase, presumably due to the formation of less reactive chloro complexes. Hydrolysis of (107)

Uses in Synthesis and Catalysis

458

p-XC&I,Cl o ]

RI ' S

+ H20

HgCl

p-XC,H,C-R+

II

HO(CH&SH

0-

( 106)

p-X C A d 3

p-XC6H4SH+ HO(CH,),CHO

+ H20

(107)

appears to involve rapid pre-equilibrium formation of a mercury complex with the S-atom of the acetal followed by a slow unimolecular breakdown of the acetal-mercury complex, while hydrolysis of (106) proceeds by slow attack of solvent water on the acetal-mercury complex. Studies of metal ion promotion of the reactions of thiocarboxylic acids and anhydrides,363 thio amide^^^'-^'^ and a variety of other sulfur compounds have been carried out. The interested reader should consult the comprehensive review by Sat~hel1.I~

61.4.8

FORMATION OF IMINES

Recent studies have shown that coordinated ammonia and amine ligands under basic conditions ~'-~~ reactions ~ may effect nucleophilic attack at carbonyl centres in organic c o m p o ~ n d s . ~These occur due to formation of deprotonated amido species which can act as nucleophiles. For example, reaction of cobalt( 111) and platinum( 1V)ammines with ketones gives the corresponding Co"' and Pt'" imine complexes. A similar reaction between [Ru( NH3)6]3f and diones produces the corresponding Ru" diimine ( 108).377It has also been found3" that nitrilepentaammineruthenium(11) [Ru(NH&I3'+ MeCOCOMe

(NH,),R lo2 occur for water attack.

OH2 (1344)

Buckingham and EngelhardtzW have studied the hydrolysis of propionic anhydride in the presence of kinetically inert complexes of the type [M(NH3),0H]"+. These reactions occur by nucleophilic attack of coordinated hydroxide on the anhydride (Scheme 32). For reactions of M-oH'"-L'+ with propionic anhydride, the Bronsted plot of log kMOH versus the pK, of M-0H2"+ is a smooth curve if values for reaction with H,O and OH- are included. Although kMOHfor [(NH3),CoOHl2+ (3 M-'s-') is about 103-foldless than koH, its reaction will compete favourably at neutral pH with base hydrolysis. Such effects are considered in more detail in Section 61.4.2.2.3.

0 Scheme 32

61.4.12

HYDROLYSIS OF GLYCOSIDES, ACETALS AND THIOACETALS

T h e hydrolysis of glycosides is susceptible to both general and specific acid ~ a t a l y s i s ,and ~~~, thus cleavage of the glycosidic bond would also be expected to be subject to metal ion catalysis. Simple 0-acetals are weak bases and even highly charged ions such as iron(II1) have limited effects on their rates of hydrolysis?25The use of chelating ligands leads to quite significant effects. Thus Clark and Hay426have found that the hydrolysis of 8-quinolyl p-D-glucopyranoside is subject to pronounced catalysis by copper(I1) (8-hydroxyquinoline was chosen as the aglycone in order to provide a binding site for the metal ion close to the glucosidic bond). Copper(II), nickel(I1) and cobalt(I1) catalyze the reaction but saturation kinetics were not observed. Copper(I1) is a particularly effective catalyst and it is estimated that the copper(I1) complex is hydrolyzed ca. 105-106times faster than the uncomplexed glycoside in the pH range 5.5-6.2.The reaction is presumed to occur as shown in Scheme 33. Przystas and Fife4*' have studied the hydrolysis of substituted benzaldehyde methyl 8-quinolyl acetals such as (135) in 50% dioxane-water at 30 "C.These acetals are subject to both general and specific acid catalysis. A variety of divalent metal ions (Cu", Co'', Ni", MnT1and Zn") exert

,

I

i

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

465

I

Scheme 33 Metal ion-catalyzed hydrolysis of 8-quinolyl P-D-glucopyranoside

large catalytic effect even though complexation is quite weak. For example, a 0.02 M concentraion of Ni" (1000-fold excess over the substrate) at pH 7.2 leads to a 2 x 105-fold increase in kobs b r the hydrolysis of (135). Metal ion catalysis is attributed to a transition state effect in which he leaving group is stabilized. L

SR Ph,C

/

R S '

S-Acetals are very readily hydrolyzed in the presence of soft metal ions and large rate accelerations (e.g. ca. 106-foldcompared with the proton) are found. In the case of S-acetals such as (136)only a small extent of pre-equilibrium complex formation occurs42*even with Hg", but for substrates such as (1371, 1 :1 complex formation readily occurs with weaker binding metal ions such as Ag'. Kinetic terms reflecting binding of two Ag' ions to the substrate (137) are detectable.429 Both A1 and A2-like schemes have been proposed for these S-acetal reactions and the topic was reviewed17 in 1977. Use can also be made of some of the effects described above for the synthesis of glycosides. in the presence of mercury( 11) salts are readily solvolyzed Thus phenyl 1-thio-o-glucopyranosides to give alkyl D-glucopyranosides with inverted anomeric c~nfiguration.~~' Methanolysis of the pand a-anomers gave the methyl a- and &glycosides which were isolated in yields of 74 and 87% respectively. The approach was extended to the synthesis of complex glycosides (the a-anomers of which are of special interest) by the preparation of cholestanyl and 1-naphthyl a-D-glucopyranosides and a disaccharide derivative. A number of other reactions in this general area such as the mutarotation of ~ - ( + ) - g l u c o s e ~ ~ ' . ~ ~ ~ and aldose-ketose i s o m e r i ~ a t i o n sare ~ ~also ~ subject to catalysis by metal ions.

61.4.13 HYDROLYSIS OF SULFATE ESTERS

Although metal-promoted hydrolysis of phosphate esters is a topic of very current interest (Section 61.4.4), little work has been published dealing with the effects of metal ions on the hydrolysis of sulfate esters. The acid-catalyzed hydrolysis of aryl sulfates has been shown to occur by an A-1 mechanism (Scheme 34).434Nucleophilic catalysis by amines has been observed in the hydrolysis of p-nitrophenyl sulfate435and intramolecular carboxyl group catalysis occurs with salicyl sulfate436as with salicyl phosphate. The hydrolysis of 8-quinolyl sulfate (138) is strongly promoted by copper(I1) in the pH range 5.4-5.8 at 39.8 0C?37The 1: 1 copper(I1) complex which is believed to have the structure (139)

Uses in Synthesis and Catalysis

466

0 ArOSO,-+H* 2 ArOS0,H

e

Ar-&?k!-%-

I

4

II

ArOH+SO,

H O Scheme 34

hydrolyzes 105-106times more rapidly than the free ligand in the pH range 5-6. Further investigations of the effect of metal ions on sulfate ester hydrolysis are required. Most aryl sulfatases have pH optima in the acidic range (pH4-6), but currently there appears to be no evidence for any metal-activated aryl sulfatase enzyme.

p cu-

0

l

I

o=s=o

o=s=o

I

l

0-

61.4.14

REACTIONS OF COORDINATED AMINO ACIDS

The various reactions undergone by coordinated amino acids have been the subject of several review^^^^^^^^^,^^^,^ and only a brief discussion will be given here. The reactions which occur can be roughly classified under three headings: (a) aldol condensations, (b) reactions of complexes of amino acid Schiff bases, and (c) isotopic exchange and racemization at the a-carbon of the amino acid.

6 1A.14.1 Aldol Condensations The synthetic applications of these reactions are considered in Section 61.4.15. Reaction of copper(I1) glycinate with formaldehyde in aqueous solution at pH 12 gives serine, while reaction with acetaldehyde gives a mixture of threonine and allothreonine (Scheme 35). This reaction has been extended to a variety of aldehydes to obtain longer chain p-hydroxyamino H

Scheme 35

Similar reactions are undergone by Schiffbase derivatives. Thus the Schiff base formed between pyruvic acid and glycine undergoes a similar reaction under weakly basic conditions (Scheme 36).M0

Scheme 36

Schiff base complexes such as N-salicylideneglycinatoaquacopper(I1) will also react with alkyl halides under basic conditions giving the alanine derivative (with MeI) and phenylalanine with benzyl bromide.&' The synthesis of threonine can be made stereospecific using optically active complexes of the type ~-LCo(en)~Gly]~+ but with low asymmetric yield.442In the case of dipeptide complexes only

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

467

the C-terminal residue is activated,443thus in the reaction between [Co(dien)(GlyGly)]' and acetaldehyde at pH ca. 11, the product is [Co(dien)(GlyThr)]+.

61.4.14.2

Schiff Bases

Schiff base formation can have a considerable effect on both the position and degree of activation of the coordinated amino acid. The Schiff bases derived from amino acids and pyridoxal have attracted considerable attention due to the biochemical significance of vitamin B6 and the realization that many of the enzymic reactions involving B6 could be brought about in the absence of enzyme by using pyridoxal and various metal ions.44d*~4s~461;662,542 Schiff base formation between pyridoxal (140) and amino acids leads to complexes of type (141) which are in tautomeric equilibrium with (142). This tautomeric equilibrium leads to transamination, thus the same metal complexes can be obtained when either pyridoxal and alanine or pyridoxamine and pyruvic acid are allowed to react together in the presence of a metal ion. H ~ p g o o has d ~ studied ~~ the rates of transamination of 15 amino acids in the presence of zinc(I1) and pyridoxal 5-phosphate (143). On mixing the reagents zinc(11)-aldimine complexes are rapidly formed (ca. 5 min) and these species subsequently transaminate in a slow second step. AI"' and Zn" systems have been particularly well ~ t u d i e d . ~ ~The - ~ role " of the metal ion seems to involve both stabilization or 'trapping' of the Schiff base, and in addition it also ensures the planarity of the conjugated wsystem. In the case of the aldimine tautomer, extensive 'H NMR s t ~ d i e s ~ ~ ' * ~ ~ * have shown that formation of the ternary complex results in activation at the amino acid 2-carbon. At room temperature the reaction occurs without incorporation of 2H into the aldehyde methine position indicating that the primary mechanism is carbanion formation rather than tautomerism. RCHCO; ,

RCC0,-

OH

61.4.14.3

Isotopic Exchange and Racemization

The activation of the normally inert methylene group of glycine by coordination to a transition ion has been recognized for many years. The substantial increase in the acidity of the methylene hydrogens has been directly confirmed by deuterium labelling experiments:' and has also been ~ ' -p~l a~ t~i n ~ m ( I I ) . ~ ~ ~ shown to occur in other a-aminocarboxylate complexes o f ~ o b a l t ( I I I ) ~and Norman and P h i p p ~ carried ~ ~ ~out a systematic investigation of the effect of various metal ions on the activation of the methylene protons of EDTA using 'H NMR techniques. Of the 30 metal ions studied only seven, all transition metal ions VO", Cr'", Fe", Co", Co"', Rh'I', Ni", Cu", enhanced the activation of the methylene groups of the ligand. A true catalytic effect was observed with Ni" which induced complete exchange even in the presence of a large excess of the ligand. Stadherr and A n g e l i ~ i "studied ~ ~ the effects of various metal ions on the base-catalyzed racemization and deuterium exchange at the methine hydrogen of N,N-bis( carboxymethyl)-D-phenylglycine (144). The metal complexes underwent racemization at rates up to 5 x 10' times greater than for the free ligand. -O2CCH2 -02CCH,

'N-CH(Ph)C02 / (14)

A recent paper4mhas questioned some of the above results and reports that Ni" when complexed to L-alanine and Co"' when complexed to Gly-Ala or Ala-GIy retard the rate of racemization of the L-alanine. A detailed study of the pH rate profile on free and Ni"-complexed L-alanine indicated that above pH 4 the free amino acid racemized faster than the complex. Further studies of these reactions are required.

CCC6-P

Uses in Synthesis and Catalysis

468

61.4.15

AMINO ACID SYNTHESIS AND ALDOL CONDENSATIONS

Coordinated ligand reactions have been used to synthesize a variety of amino acids, and some of the approaches employed are discussed below. Intramolecular imine formation between a coordinated aminate ion and a 2-oxo acid has been utilized to synthesize the two racemic amino acids 2-cyclopropylglycine and proline.463Thus anation of [ C O ( N H ~ ) ~ O Hby ~ ] Br(CH2)3COC02H ~+ at pH 5 gives two major products (145) and (146). Both are converted to tetraammineiminocarboxylato chelates by attack of an adjacent deprotonated ammonia. The cyclopropylimine complex can, for example, be reduced by alkaline BH4- to give the ( R S )-2-cyclopropylglycine complex.

( 146)

( 145)

The cobalt(II1)-promoted synthesis of P-carboxyaspartic acid has also been reported464by the intramolecular addition of coordinated aminate ion to the alkenic centre in the (3,3-bis(ethoxycarbonyl)-2-propenoato)pentaamminecobalt(llI) ion [(H3N)5CoOOCCH==C(C02Et)z]2+. In aqueous solution, an intramolecular addition of the cis-aminate ion at the alkenic centre gives the N,O-chelated diester of P-carboxyaspartic acid (Scheme 37).

Scheme 37

Coordination of a-amino acids to metal ions has long been known to increase the acidity of the C-H bond at the a-carbon atom and various reviews of the topic are a~ailabIe.~'*~' The original observation465that copper( 11) ions promote the base-catalyzed condensation of acetaldehyde with glycine to give threonine (147) was rapidly followed by a number of reports of similar Since this early work a great deal of literature has appeared concerning the reaction of metal complexes of amino or small with various aldehydes. The reaction has also been carried out using chiral octahedral cobalt(II1) complexes containing glycine with relatively high asymmetric yields but low overall yield^.^^'-^^' It was later discovered that the use of an amino acid Schiff base complex instead of an amino acid metal complex increased the acidity of the C--H bond at the a-carbon atom of the amino acid and also prevented the occurrence of N - a l k y l a t i ~ n . ~ ~Thus ' - ~ ' the glycine residue in N-salicylideneglycyl-L-valinatocopper(IT) reacts with formaldehyde in aqueous solution at pH 8.5. Decomposition of the complex with HZSat pH 2 gives seryl-L-valine containing optically active

( 147)

In the case of copper(I1) complexes, the reaction with acetaldehyde is believed to proceed by the steps shown in Scheme 38. The bis(oxazolidine)copper(II) complex (148; R = Me) has been characterized by X-ray analysis?86 Treatment of this complex with H2S in acid solution gives threonine. Synthetic procedures have been developed giving threonine in 95% yields.4s3 The condensation of formaldehyde with glycine or glycine Schiff bases coordinated to CoIII, CUI' and Nil' has been shown to give a-hydr~xymethylserine."~~ With [Cu(GlyO),] the reaction of formaldehyde at the a-carbon is preceded by condensation on the amino group and is followed by cyclization to give an oxazolidine-type ring.

Lewis Acid Catalysis and the Reactions of Coordinated Ligands RCHOH I

7 ,

/ N, CHI

,c.

I

’ I

II

I

RCHOH

RCH CHR

RCH \

469

RCHO

,

/NH-CH ,cU

\*

I /c.,

(148)

Scheme 38

Condensation of formaldehyde with [Co(en),(GlyO)j2+ gives as the initial product a-hydroxymethylserinebis(ethylenediamine)cobalt( 111) which is subsequently converted into the a-hydroxymethylserine-l,4,8,1l-tetraaza-6,13-dioxacyclotetradecanecobalt( 111) ion containin the macrocyclic ligand (149) which is coordinated via the nitrogen donors as in (lW).48g Cooke and Dabr~wiak have ~ ~ studied ~ the reaction of acetaldehyde with optically active [Co(en),(GlyO)]’+. The isolated threonine and allothreonine contain an excess of the S-enantiomer indicating that the aldehyde prefers to attack the S ‘side’ of the coordinated glycine.

61.4.16 ESTER EXCHANGE REACTIONS

A number of metal complexes of carboxylic esters undergo transesterification on refluxing with a l ~ o h o l s .Thus ~ ~ ~copper( , ~ ~ ~11) complexes of ethyl icolinate (151) on refluxing in methanol give the analogous complexes of methyl picolinate: i? In some cases methanolysis rather than transesterification occurs, as with bis(ethy1acetoacetato)copper(II) which gives methoxy-bridged derivatives in refluxing methan01.~~ Sigman and J~rgenson~~’ have found that zinc( 11) catalyzes the transesterification reaction between N- (p-hydroxyethy1)ethylenediamine and 4-nitrophenyl picolinate. This reaction involves a reactive mixed ligand complex (152) in which the zinc(1I) ion perturbs the pK, of the hydroxyethyl group of N-(P-hydroxyethy1)ethylenediamineto provide a high effective concentration of the nucleophile. Intramolecular nucleophlic attack then occurs at the carbonyl group of p-nitrophenyl picolinate. This system provides a somewhat unique example of intramolecular

0

I 0-cu-0 II I

EtO’ C

II

470

Uses in Synthesis and Catalysis

attack by a coordinated nucleophile in a labile complex, the reaction pathway being confirmed by the nature of the products obtained. has been found to be a The complex bis(pentane-2,4-dionato)-~,~'-dimethoxy-dicopper(II) very effective and selective catalyst for ester exchange of both 2-ethoxycarbonylpyridine and ethyl 2-pyridylacetate to give the corresponding methyl esters.491 A number of studies have been carried out on the ligand reactivities of the metal complexes of Schiff bases derived from amino acid esters and carbonyl compounds. Ester exchange reactions were first reported by Pfeiffer et and extended by other investigator^.^'^-^^^ Copper(I1) complexes of the Schiff bases derived for salicylaldehyde and dibenzyl asparate or glutamate have recently been characteri~ed:'~ Refluxing these complexes in methanol for 30 min gives bis( a-methyl-P-benzyl-N-salicylideneaspartato)copper(I~) in which selective ester and bis( cr-methyl-y-benzyl-N-sali~lideneglutamato)copper(II)exchange occurs. A similar selective ester exchange reaction has also been observed with the copper(I1) complex of the Schiff base derived from salicylaldehyde and dibenzyl as as par ate.^"

61.4.17

HYDROLYSIS OF EPOXIDES

Some interesting work has been published dealing with the metal ion-promoted hydrolysis of epoxides. Hanzlik and Michaely4" first observed that in the presence of copper( 11) the hydration of 2-pyridyloxirane (153) is accelerated by a factor of 1.8 x lo4 and its reaction with Cl-, Br- and MeO- becomes 100% regios ecific for p-attack. The magnitude of the catalytic effect decreases in the order Cur'> CO"> ZnR >> Mn". The pH rate profile for the copper( TI)-catalyzed reaction is a bell-shaped curve with a maximum at ca. pH 5. Further investigations of the reaction suggest that the bidentate complex (154) is the reactive species which undergoes external attack by water or other nucleophiles present in solution. The enzymic hydration of epoxides and the possible role of metal ions has been the results obtained suggest that a metal ion is not involved at the active site of epoxide hydrase.

X = O H , OMe, C1, Rr

61.4.18

(154)

UREA HYDROLYSIS

The enzyme urease catalyzes the hydrolysis of urea to form carbamate ion (equation 32). At pH 7.0 and 38 "C, the urease-catalyzed hydrolysis of urea is at least 1014 times as fast as the spontaneous hydrolysis of urea. Jack bean urease is a nickel(I1) metalloenzyme'02 with each of its six identical subunits containing one active site and two metal ions, and at least one of these nickel ions is implicated in the hydrolysis. It has been suggested503that all substrates for urease (urea, N-hydroxyurea, N-methylurea, semicarbazide formamide and acetamide) are activated towards nucleophilic attack on carbon as a result of 0-coordination to the active nickel(I1) site as in (155). Nickel(I1) ions have been foundsw to promote the ethanolysis and hydrolysis of N-(2-pyridylmethyl)urea (Scheme 39) and this system is considered to be a useful model for the enzyme. 0

I1

H,NCNH,

+ H 2 0 + H,NC02-+

NH4+

(155)

R = NH,, NHOH, NHMe, NHNH2, H, Me

(32)

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

47 1

Scheme 39

The synthesis of both 0- and N-bonded linkage isomers of the [Rl~(NH~)~(urea)]~' complexes has been described505and their reactions studied in detail. The formation of [Rh(NH,),I3+ from [Rh(NH3)5(NH,CONH,)]3+ represents the first observation of non-enzymatic enhancement of the decomposition of urea at low pW. Comparison of the rate constant ( k = 2.0 x IO-' s - ' ) with that observed for the slow decomposition of urea at low pH (equation 33; k = 6 x 10-los-l at 25 "C)indicates that coordination to the metal centre increases the reactivity by some 3 x 104-fold. NH,CONH,

4

NH,+HNCO

e

NH,'+NCO-

(33)

The reactions of the pentaarnminecobalt( 111) complex of urea (0-coordinated) have also been Under basic conditions [Co( NH,),OH]*+ is the only cobalt( 111) product. The main reaction pathway (ca, 97%) is S,lCB displacement of coordinated urea (Scheme 40) with koH = 15.3 M-' s-' at 25 "C.A limiting rate was approached at high pH as the complex dissociated to its inactive conjugate base. Hydrolysis of coordinated urea was not observed. /

OH-

urea+ (NH,),Co-OH

'S,ICB

(NH,),Co-O=C

NHzlJ+

\ NH2

Kn

YNH1 I+

(NH,),Co-0-C

L 1

\ NH,

Scheme 40

The complexes [Cr(NH,),LI3+ ( L = OCHNH2,OC(NH2)2and OC(NHMe2)]have been characterized as the 0-bound isomers.507At 25 "C in aqueous base they isomerize to the deprotonated N-bound isomers without competitive hydrolysis. Reacidification regenerates the 0-bound isomer rapidly and completely. The urea pentaammine complexes of CoI", Rh"' and Cr"' do not provide any evidence for detectable production of [(NH,),M-OOCNH,] required for reactions to model the chemistry of urease. This is not altogether surprising as there is little apparent activation upon coordination. The pK, of the free ligand (ca. 14) is not lowered substantially on Co"' (13.19), Rh"' (13.67) or Cr"' (13.50), and free urea is hydrolyzed slowly in base. Studies with the inert metal centres suggest that coordination through oxygen alone does not activate urea towards hydrolysis. This result suggests that coordination of urea to a single Ni" at the active site of urease is insufficient for catalysis and that a mechanism involving the two metal ions at the active site may be required.

61.4.19

ACYL TRANSFER REACTIONS

One particularly interesting area which has not been subjected to detailed study is metalpromoted acyl transfer. Studies on esterases and peptidases have shown that acyl transfer occurs to a nucleophilic group of the enzyme within an enzyme-substrate complex and the acyl group is then hydrolyzed in the second step. Many of these enzymes, e.g. carboxypeptidase, contain zinc(I1). Breslow and Chipmanso*have shown that zinc( 11) pyridinecarboxaldimine anion (156) provides a strong free nucleophile and is a particularly effective nucleophilic catalyst in the hydrolysis of 8-acetoxyquinoline 5-sulfonate (157; Scheme 41). The reaction involves the catalyst-substrate complex (158). Molecular models show that in the mixed ligand compIex (158), the N-0- group is in a position to attack the acetyl group of (157). The zinc complex (156) is also an excellent catalyst for the hydrolysis of p-nitrophenyl acetate; in fact it is comparable in reactivity to hydroxide ion, although its pK, is only 6.5, (Table 29). The hydrolysis of 8-acetoxyquinoline (159) is subject to catalysis by metal ions and detailed kinetic studies of the reaction have been reported.5w The metal ion could be bound to the carbonyl oxygen (160) or the ether oxygen (161) and the actual structure of the catalytically active complex was not unequivocally defined.

47 2

Uses in Synthesis and Catalysis

0 I

CH II

Zn-N

\

+

@ 0 t

0-

Table 29 Rate Constants for Nucleophilic Attackasb

Nucleophile

OH-

15.7

PCA anion

10.04

HZO PCA-Zn"

2.01

14.8

77.2

-1.7

6.98 7.3 x

6.5

10

1.ox 10-8 10

from R. Breslow and D.Chipman, 3. A m Chem Soc., 1965,87,4195 PCA = pyridine-2-carboxaldoxime,AQS = 8-acetoxyquinoline-5-sulfonate, PNA = p-nitrophenyl acetate.

a Data

At 25 "C.

Sakan and Mori510 have shown that 8-acetoxyquinoline reacts with copper(I1) chloride in dry tetrahydrofuran to give the green complex (162). The IR spectrum of the complex has v(C0) at 1790 cm-' consistent with the ether oxygen acting as a donor. Chelation via the carbonyl oxygen would lead to a substantial decrease in the carbonyl stretching frequency (the value of v ( C 0 ) for the free ligand is 1750cm-'). In (162) the carbonyl oxygen is very reactive to nucleophiles owing to polarization of the C-0 bond and the substantial thermodynamic stability of the copper( 11) complex of 8-hydroxyquinoline ( i.e. complexation of the leaving group).

I

Me\

I

P-cu-c' I C c1 II

0 (162)

The complex (162) can be used as an acylating agent for alcohols, phenols and amines in THF or benzene solvent (Scheme 42). Acetanilide can be isolated in 83% yield after 2 days at room temperature. In addition 8-acetoxyquinoline will react with Grignard reagents to give high yields (80% ) of ketones. The reaction is believed to proceed via the intermediate magnesium(I1) complex (163)(Scheme 43).

Lewis Acid Catalysis and the Reactions of Cvordinated Ligands

473

MeCONHPh

,O-Cu-cl

Me, C

It

I

c1

0

Scheme 42

- Q” R”MgBr

R\

o , C II

0

K ,

,0-Mg-Br C

II

1

R’’

*

+R’CR”

0-Mg

II

0

I Br

Some kinetic work on the reaction of Bu”NH2with the copper(I1) complex of 8-acetoxyquinoline at 25°C in DMF as solvent has been carried out. The second order rate constant is 9 . 5 ~ 10-2

M-l

61.4.20

s-1*511

ENOLIZATION

The rates of enolization or ionization of certain keto acids in which the carboxylic group is suitably placed to assist the ionization by interaction with the carbonyl oxygen have been found to be considerably higher (factors of up to 2 x lo3)than those of the corresponding keto If the carbonyl oxygen can be coordinated to a metal ion, a considerable enhancement of the rate of ionization of the active C-H bond would be expected. Pedersen first observed such a rate enhancement for the acetate-catalyzed halogenation of both ethyl acetate514 and ethyl 2oxocyclopentanecarboxylate5‘5in the presence of Cu”. The effects were not large presumably due to only weak complexation between Cu” and the ketones. Complex formation between 2-acetylpyridine (164 = I) and Zn”, Ni” and Cu” leads to large increases in the rates of enolization or ionization of the C-H bond adjacent to the carbonyl group,516the catalytic constants for the acetate-catalyzed iodination being 5 x lo3 (ZnL2+) and 2 x lo5 (CuL2+)times larger than that of the uncomplexed substrate. This effect, which is considerably larger than that observed on protonation of 2-acetylpyridine (ca. 200-fold rate increase), is consistent with stabilization of the negative charge developing on the carbonyl oxygen during ionization of the C-H bond as shown for the transition state (165). -0Ac

Somewhat similar effects have been observed with acetonyl phosphonate (MeCOCH2Po,’-).”’~’’8 The rate of deuteration at the 2-position is some 2000 times faster in the presence of magnesium( 11) compared with the free substrate. ~ ~ studied keto-enol tautomerism rates for oxaloacetate ( oxac2-) Covey and L e u s ~ i n g ” ’ . ~have and zinc(I1) oxaloacetate complexes in acetate buffer solutions at 25 “C. Thus for the equilibrium the value of k,, the rate constant for the forward reaction, for the reaction oxac,,,, 2- 4oxace,,:oxacketoZP+ OAc- is 6.6 x lo-’ M-’ sC1 while for Zn(oxac)k,,,) + OAc-, kf is 25 M-’ s-I. The rate

-

Uses in Synthesis and Catalysis

474

acceleration at 25 "C is ca. 4 x lo3, very similar to that of 5 x lo3 which occurs in the zinc(1I)promoted acetate-catalyzed enolization of 2-a~etylpyridine.~'~

61.4.21 CARBONYL GROUP HYDRATION Carbonic anhydrase is a zinc( 11)metalloenzyme which catalyzes the hydration and dehydration of carbon dioxide, CO,+ H 2 0 e H++ H C 0 3.525 As a result there has been considerable interest in the metal ion-promoted hydration of carbonyl substrates as potential mode1 systems for the enzyme. For example, Pocker and Meany519 studied the reversible hydration of 2- and 4pyridinecarbaldehyde by carbonic anhydrase, zinc( II), cobalt(11), H 2 0 and OH'^'.The catalytic efficiency of bovine carbonic anhydrase is ca. 10' times greater than that of water for hydration of both 2- and 4-pyridinecarbaldehydes. Zinc(I1) and cobalt(I1) are ca. lo7 times more effective than water for the hydration of 2-pyridinecarbaldehyde, but are much less effective with 4pyridinecarbaldehyde. Presumably in the case of 2-pyridinecarbaldehyde complexes of type (166) are formed in solution. Polarization of the carbonyl group by the metal ion assists nucleophilic attack by water or hydroxide ion. Further studies of this reaction have been but the mechanistic details of the catalysis are unclear. Metal-bound nucleophiles (M-OH or M-OH2) could, for example, be involved in the catalysis.

( 166)

(167)

The interaction of 2-pyridinecarbaldehyde with Cu( 11) has been studied by potentiometric and spectrophotometric methods,522 and formation constants have been determined. The neutral deprotonated complex (167) was characterized in the solid state. The hydration of acetaldehyde is catalyzed by zinc(I1) and the catalysis is enhanced by the Various processes were considered (Scheme presence of anions such as acetate and 44) to account for the hydroxide-dependent and -independent pathways involving zinc( 11). Zinc( IT)

Scheme 44

catalysis of the hydration of acetaldehyde has been shown to be markedly increased by a ligand carrying a remote general base.524This result provides evidence that hydration when cooperatively 'atalyzed by zinc(l1) and a base requires independent attack by base and metal ion, that is the base must not be coordinated to the metal ion. This system involved the use of the zinc complex of the pyridine-2-aldoximate ion (pyox- = 168). Catalysis is believed to involve general base-

H

'N

p

N

' 0 (168) = pyox-

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

47 5

catalyzed attack at the carbonyl group which is already polarized and thus activated by the zinc ion (169). The remote general base increases the activity of zinc(I1) by a factor of ca. 2 x 10'.

61.4.22 METAL ION-PROMOTED CARBONYL REDUCTIONS The NAD+-dependent alcohol dehydrogenase from horse liver contains one catalytically essential zinc ion at each of its two active sites. An essential feature of the enzymic catalysis appears to involve direct coordination of the enzyme-bound zinc by the carbonyl and hydroxyl groups of the aldehyde and alcohol substrates. Polarization of the carbonyl group by the metal ion should assist nucleophilic attack by hydride ion. A number of studies have confirmed this view. Zinc(I1) catalyzes the reduction of 1,lO-phenanthroline-2-carbaldehydeby N-propyl- 1,Cdihydronicotinamide in acetonitrile,"' and provides an interesting model reaction for alcohol dehydrogenase (Scheme 45). The model reaction proceeds by direct hydrogen transfer and is absolutely dependent on the presence of +zinc(II). The zinc(I1) ion also catalyzes the reduction of 2- and 4-pyridinecarbaldehyde by Et4N BH4-.526The zinc complex of the 2-aldehyde is reduced at least 7 x lo5 times faster than the free aldehyde, whereas the zinc complex of the 4-aldehyde is reduced only lo2 times faster than the free aldehyde. A direct interaction of zinc(I1) with the carbonyl function is clearly required for marked catalytic effects to be observed.

Scheme 45

The reduction of (E)-2-, ( E ) - 3 - and (E)-4-cinnamoylpyridinesby 1,Cdihydropyridines to give dihydro ketones has also been shown to be catalyzed by zinc(I1) and magnesiurn(II)."7 Kinetic measurements show that the rate of reduction is fastest in the case of the 2-isomer where the metal is simultaneously complexed with the nitrogen and oxygen donors. A very fast zinccatalyzed reduction of pyridine-2-carbaldehyde by the alcohol dehydrogenase coenzyme model N,N'-diethyl-N-benzyl-1,4-dihydronicotinamide (170) has also been described."*

61.4.23 METAL ION-PROMOTED REACTIONS OF ALKENES Nucleophilic attack at an alkene is not a favoured process in organic chemistry. Forcing conditions are required, such as high concentrations of a strong base and/or high temperatures.529 Biological reactions are, however, known where alkenes are rapidly hydrated or aminated in near neutral conditions and ambient temperatures. Some of these enzymic systems involve metal ions and as a result there has been an interest in developing suitable model systems using metal complexes. The enzyme aconitase catalyzes the isomerization of citric acid to isocitric acid via the interrnediate cis-aconitic acid (Scheme 46),530 and various attempts have been made to model this reaction.2' The cobalt(ll1) complexes derived from methyl maleate (171) and methyl fumarate (172) have been prepareds3' to study intramolecular attack by coordinated hydroxide on the alkene. Generation of the hydroxo species of the maleic acid complex leads to rapid cyclization to give the

CCCB-P'

476

Uses in Syntkesi.r and Catalysis

-0

CHZCOZH I HO-C-CO,H

I

CHZCOZH

citrate

HC-C02H !I C-C02H I CH,CO,H cis- aconitate

l=

HO-CH-COZH

I

H-C-COZH I CH2COZH

(243S)-isocitrate

Scheme 46

chelated malate half ester in two diastereoisomeric forms ( A S and AR). The crystal structure of the AR isomer has been determined,532and the reactions leading to its formation are shown in Scheme 47. In 0.2 M imidazole buffer at pH 8 the rate constant for cyclization is ca. 0.1 s-l at 25 "C, some 107-fold greater than that for hydration of the uncoordinated half ester under the same conditions. The cyclization gives exclusively the five-membered chelate, and the rate is independent of pH in the range 8-11 as would be predicted for such an intramolecular reaction. Tracer studies have also shown that bound ''OH- is retained in the chelated malate.

O==T

AR

OMe Scheme 47

Superficially, it might be expected that coordinated acrylic acid would undergo a similar type of hydration reaction with nucleophilic addition taking place at the P-carbon (the acid is ester-like when coordinated). Experimentally" hydration is not observed (Scheme 48), but both the 2- and

Scheme 48

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

477

3-chloroacrylato complexes are hydrated by this intramolecular process. The latter gives a fivemembered chelate ring while the former gives a six-membered ring. Interestingly the five-membered ring appears to be formed more slowly than the six-membered ring. A consequence of the addition of coordinated OH- to alkenes is that other nucleophilies, for example a coordinated aminate ion, should also be active. This type of reaction is seen with the chloropentaammine complex [CO(NH~),OOCCH=CHCO~B~']~+ in aqueous base533(Scheme 49). The reaction of the t-butyl maleate complex occurs to the extent of ca. 50% and is complicated by hydrolysis of the maleate ester and some decomposition of the cobalt(II1) complexes. Reactions of this type have recently been ex loited in the synthesis of P-carboxyaspartic acid in the coordination sphere of ~ o b a l t ( I I I )!? .~~

CH

II CH

0

I

Scheme 49

61.4.24

ISOMERIZATION AND CHELATE RING OPENING

The acid-catalyzed aquation of cis-[Cr(mal),(OH,),]- to give [Cr(mal)(OH,),]+ has been studied in The CUI'-, Ni"-, Co"- and 2n"-catalyzed dissociation of the cisdiaquabis(malonato)chromate( 111) ion has also been recently investigated using perchloric acid solutions.536Under these conditions kobs = kH[H"] + kM[M2*]where kobsis the observed first order rate constant and k , and k , are the appropriate rate constants for the proton- and metal ion-catalyzed pathways respectively. The metal ions exert marked catalytic effects with kMvalues following the sequence Cu" > Ni" > Co" > Zn". Saturation kinetics were not observed. A possible mechanism for the reaction is outlined in Scheme 50. Metal ion-catalyzed aquation of tris(malonato)chromate(III) has also been in~estigated.'~~ In perchloric acid solutions ([ H+] = 0.05 M), k& = kH[H*]-t- kw[M2+], with values of kM following the sequence Cu"> Nil1> Co"> Zn" > Mn". The k, values parallel the formation constants for the mono complexes of the metal ions with malonate,. A similar mechanism to that shown in Scheme 50 is predicted for this reaction. A number of other studies on metal ion-catalyzed aquation of Cr"' complexes have also been publi~hed.~~~,~~~ The trans e cis isomerization of [Cr(Cz04)(H20)2]-is also catalyzed by metal ions and a detailed study of catalysis by Mg" has been published.s40 Dissociation of oxalato complexes of Cr"' of the type [C~(OX),(OH,),_,,]'~-~"'+( n = 1,3) is also promoted by a series of transition metal ions.541

Scheme 50 Possible mechanism for the metal ion-catalyzed aquation of cis-[Cr(mal),(H,O),]-

Uses in Synthesis and Catalysis

47 8 41.4.25

REFERENCES

1. J. Sen and H. Taube, Acta Chem. Scand., Ser A, 1979, 33, 125. 2. F. H. Westheimer, Trans. N . Y.Acad. Sci., 1955, 18, 15. 3. I. P. Evans, G. W. Everett and A. M. Sargeson, J. Chem. SOL,Chem. Commun., 1975, 139; J. Am. Chem. Soc., 1976, 98, 8041. 4. Q. Fernando, Adv. Inorg. Chem. Radiochem., 1965, 7 , 185. 5. R. W. Hay, Reo. Pure Appl. Chem., 1963, 13, 157. 6. R. W. Hay, J. Chem. Educ., 1965,42,413. 7. M. M. Jones, ‘Ligand Reactivity and Catalysis’, Academic, New York, 1968. 8. E. Kimura, Yuki Gosei Kagaku Kyckuishi, 1971,29, 12. 9. M. M. Jones and W. A. C.onnor, Ind. Eng. Chem., 1963, 55, 15. 10. M. L. Bender, in ‘Reactions of Coordinated Ligands’ (Advances in Chemistry Series, No. 37), American Chemical Society, Washington, DC, 1963. 11. T. C. Bruice and S. J. Benkovic, ‘Bio-organic Mechanisms’, Benjamin, New York, 1966, vols. I and 2. 12. M. L. Bender, ‘Mechanisms of Homogeneous Catalysis from Protons to Proteins’, Wiley-Interscience, New York, 1971. 13. H. Sigel (ea.), ‘Metal Tons in Biological Systems, VOI. 5, Reactivity of Coordination Compounds, Dekker, New York, 1976. 14. R. P. Hanzlik, ‘Inorganic Aspects of Biological and Organic Chemistry, Academic, New York, 1976. 15. R. P. Houghton, ‘Metal Complexes in Organic Chemistry, Cambridge University Press, London, 1979. 16. A. E. Martell (ed.), ‘Xth International Conference on Coordination Chemistry’, Butterworths, London, 1968. 17. D. P. N. Satchell, Chem. SOC.Reo., 1977, 6, 345. 18. J. P. Candlin, K. A. Taylor and D. T. Thompson, ‘Reactions of Transition Metal Complexes, Elsevier, Amsterdam, 1968. 19. A. E. Martell, in ‘Metal Ions in Biological Systems, ed. H. Sigel, Dekker, New York, 1976, vol. 2, pp. 208-262. 20. M. M. Taqui Kahn and A. E. Martell, ‘Homogeneous Catalysis by Metal Complexes’, Academic, New York, 1974. 21. N. E. Dixon and A. M.Sargeson, in ‘Zinc Enzymes’, ed. T. G. Spiro, Wiley, New York, 1983, chap. 7. 22. R. W. Hay and P.J. Morris, in ‘Metal Ions in Biological Systems’, ed. H. Sigel, Dekker, New York, 1976, vol. 5. 23. D. A. Buckingham, in ‘Biological Aspects of Inorganic Chemistry, ed. A. W. Addison, W. R. Cullen, D. Dolphin and B. R. James, Wiley, New York, 1976. 24. R. W. Hay, in ‘Reactions of Coordinated Ligands, ed. P. S. Braterman, Plenum, New York, vol. 2 (in press). 25. J. E. Coleman, Prog. Bioorg. Chem., 1971, 1, 159. 26. G. L. Eichhom (ed.), ‘Inorganic Biochemistry’, Elsevier, London, 1973, vols. 1 and 2. 27. ‘Inorganic Biochemistry’ (Specialist Periodical Reports), Royal Society of Chemistry, London, vols. 1-3. 28. R. W. Hay and D. R. Williams, ‘Metal Complexes of Amino-acids, Peptides and Proteins’, in each odd issue of ‘Amino-acids, Peptides and Proteins’ (Specialist Periodical Reports), ed. R. C. Sheppard, Royal Society of Chemistry, London. 29. E.-I. Ochiai, Coord. Chem. Rev., 1968, 3, 49. 30. D. P. N. Satchell and R. S. Satchell, Annu. Rep. Chem. SOC.A., 1978, 25. 31. A. Pasini and L. Casella, J. lnorg. Nucl. Chem, 1974, 36, 2133. 32. D. A. Phipps, J. Mol. Catal., 1979, 5, 81. 33. C. F. Baes and R. E. Mesiner, ‘The Hydrolysis of Cations’, Wiley-Interscience, New York, 1976. 34. D. W. Barnum, Inorg. Chem., 1983, 22, 2297. 35. G. Johansson and H.Ohtaki, Acta. Chem. Scand, 19?3,27,643. 36. H. Kroll, J. Am. Chem. SOC.,1952, 74, 2036. 37. See for example D. D. Pemn and B. Dempsey, ‘Buffers for pH and Metal Ion Control’, Chapman HaIl, London, 1974. 38. U. Bips, H. Elias, M. Hauroder, G. Kleinhans, S. Heifer and K. J. Wannowius, Inorg. Chem., 1983,22, 3862. 39. M. P. Springer and C. Curran, Inorg. Chem., 1963, 2, 1270. 40, R. W. Hay and L. J. Porter, Aust. J. Chem., 1967, 20, 675. 41. H. Shindo and T. L. Brown, J. Am. Chem. Soc., 1965, 87, 1904. 42. R. W. Hay, R. Bennett and D. J. Barnes, J. Chem Soc, Dalton Trans., 1972, 1524. 43. D. k Buckingham, J. Dekkers and A. M. Sargeson, J. Am Chem Soc., 1973, 95, 4173. 44. M. L. Bender and B. W. Turnquest, J . Am. Chem. Soc., 1957.79, 1889. 45. J. M. White, R. A. Manning and N. C. Li, J. Am. Chem. SOC.,1956, 78, 2367. 46. H. L. Conley and R. B. Martin, J. Phys. Chem., 1965, 69, 2914. 47. J. E. Hix and M. M. Jones, Inorg. Chem., 1966, 5 , 1863. 48. C. Regarth, Acia Pharm. Suec., 1966, 3, 101. 49. B. E. Leach and R I. Angelici, Inorg. Chem., 1969, 8, 907. SO. R. J. Angelici and D. Hopgood, J. Am. Chem. SOC.,1968, 70, 2514. 51. R. D. Wood, R. Nakon and R.J. Angelici, horg. Chem, 1978, 17, 1088. 52. J. W. Allison and R. J. Angelici, Inorg. Chem., 1971, 10, 2338. 53. R. Nakon, P. R Rechani and R. J. Angelici, J. Am. Chem. SOC.,1974,96, 2117. 54. R. W. Hay and P. K. Banerjee, unpublished results. 55. R. W. Hay and P. K. Banerjee, J. Chem. SOC, Dalton Trans., 1980, 2452. 56. R. W. Hay and P. K. Banerjee, J. Chem. SOC., Dalton Trans., 1980, 2385. 57. R. W. Hay and P. K. Banerjee, J. Inorg. Biochem., 1981, 14, 147. 58. D. E. Newlin, M. A. Pellack and R. Nakon, J. Am Chem. Soc, 1977,99, 1078. 59. J. E. Coleman and B. L. Valee, J. Bid. Chem., 1961, 236, 2244. 60. S. Lindskog and P. 0. Nyman, Biochem. Biophys. Acta, 1964, 85, 462. 61, J. K. Walker and R. Nakon, Znorg. Chem., 1978, 17, 1151. 62. S, A. Bedell and R. Nakon, Inorg. Chem., 1977, 16, 3055. 63. J. K. Walker and R. Nakon, Inorg. Chim. Acta, 1981, 55, 135. 64. R W. Hay and P. J. Morris, J. Chem. Soc ( A ) , 1971, 3562.

Lewis Acid Caialysis and the Reactions of Coordinated Ligands

479

R. W. Hay and P.J. Morris, 3. Chem. Soc. ( A ) , 1971, 1518. N. C. Li, E. Doody and J. M. White, J. Am. Chem. Soc., 1957, 79, 5859. H. L. Conley and R. B. Martin, J. Phys. Chem., 1965,69, 2923. L. J. Porter, D. D. Perrin and R. W. Hay, J. Chem. SOC.( A ) , 1969, 118. J. M. White, R. A. Manning and N. C. Li, J. Am. Chem. SOC.,1955, 77, 5225. R. Mathur and N. C. Li, J. Am. Chem. SOC.,1964, 86, 1289. R. W. Hay and L. J. Porter, J. Chem. SOC.( A ) , 1969, 127. R. W. Hay and P. J. Moms, J. Chem. SOC.( A ) , 1971, 1524. R. W. Hay and P. J. M o m s , J. Chem. SOC.,Dalton Trans., 1973, 56. R. J. Angelici and B. E. leach, J. Am. Chem. SOC.,1967, 89, 4605. R. J. Angelici and B. E. Leach, J. Am. Chem. Soc., 1968,90, 2499. B. E. Leach and R. J. Angeiici, 1.Am. Chem. SOC.,1968, 90, 2504. J. Rodgers and R. A. Jacobson, hm-g, Chim. Acra, 1975, 13, 163. R. Nakon and R. J. Angelici, 1. Am. Chem. Soc., 1973,95, 3170. R. W. Hay and K. B. Nolan, J. Chem. SOC.,Dalton Trans., 1974, 2542. 80. R. W. Hay and M. P.Pujari, J. Chem. SOC.,Dalton Trans., 1984, 1083; R W. Hay and M. P. h j a r i , Inorg. Chim. Acta, 1986, 123, 175. 81. R. Nakon and R. J. Angelici, Inorg. Chem., 1973, 12, 1269. 82. R. W. Hay and K. B. Nolan, J. Chem. Soc., Dalton Trans., 1975, 1348. 83. R. W. Hay, K. B. Nolan and M. Shuaib, Transition Met. Chem., 1980, 5, 230. 84. T. R. Kelly, Ph.D. Thesis, University of Glasgow, 1962. 85. R. W. Hay and P. Banerjee, J. Chem SOC.,Dalton Trans., 1981, 362. 86. R. W. Hay and A. K. Basak, J. Chem. SOC.,Dalton Trans., 1982, 1819. 87. M.-C. Lim, J. Chem. SOC., Dalton Trans., 1983, 1675. 88. R. D. Gillard, Inorg. Chem. Acta Rev., 1967, 1, 69. 89. J. E. Hix, Jr. and M. M. Jones, J. Am. Chem. Soc., 1968, 90, 1723. 90. R. W. Hay and P. J. Morris, J. Chem. Soc. ( A ) , 1971, 1524. 91. R. W. Hay and P. J. Morris, Chem. Common., 1969, 18. 92. P. J. Moms and R. B. Martin, J. Inorg. Nucl. Chem., 1970, 32, 2891. 93. R. W. Ketsinger, F. A. Cotton and R. F. Bryan, Acta Cryslallogr., 1963, 16, 651. 94. M. M. Harding and H.A. Long, J. Chem. Soc ( A ) , 1968, 2554. 95. R. J. Sundberg and R. B. Martin, Chem. Rev., 1974, 74, 471. 96. J. R. Blackburn and M. M.Jones, J. Inorg. Nucl. Chem., 1973, 35, 1597. 97. J. R. Blackburn and M. M. Jones, J. Inorg. Nucl. Chem., 1973, 35, 1605. 98. S. Kitagawa, T. Murakami and M. Hatano, Inorg. Chem., 1976, 15, 1378. 99. M. Hatano and T. Nozawa, in ‘Metal Ions in Biological Systems’, ed. H. Sigel, Dekker, New York, 1976, vol. 5. 100. R. B. Martin, M. Chamberlin and J. T. Edsall, J. Am. Chem Soc,, 1960,82,495. 101. H. C. Freeman, J. M. Guss and R. L. Sinclair, Chem. Commun., 1968, 485. 102. See for example D. W. Margerum, L. F. Wong, F. P. Bossu, K. L. Chellappa, J. J. Czarnecki, S. T. Kirksey, Jr. and T. A. Nuebecker, in ‘Bioinorganic Chemistry 11’ (Advances in Chemistry Series, No. 162), ed. K. N. Raymond, American Chemical Society, Washington, DC, 1977. 103. See for example E. A. Lance and R. Nakon, Inorg. Chem Acta, 1981, 55, L1. 104. E. W. Wilson and R. B. Martin, Inorg. Chem., 1970,9, 528. 105. H. A. 0. Hi11 and K. A. Raspin, J. Chem. Soc. ( A ) , 1969,619. 106. R, B. Martin and H. Sigel, Chem Rev., 1982, 82, 385. 107. R. D. Gillard, E.D. McKenzie, R. Mason and G. B. Robertson, Coord. Chem. Rev., 1966, 1, 263. 108. E. D. McKenzie, 1 Chem. SOC.( A ) , 1969, 1655. 109. M. T. Barnet, H. C. Freeman, D. A. Buckingham, I-Nan Hsu and D. van der Helm, Chem Commun., 1970,367. 110. D. L. Rabenstein, Can. 1.Chem., 1971,49, 3767. 111. R. D. Gillard, Inorg. Chem Acta Rev., 1967, 1, 69. 112. H. C. Freeman, Adv. Protein Chem., 1967, 22, 257. 113. A. S. Brill, R. B. Martin and R. J. P. Williams, in ‘Electronic Aspects of Biochemistry’, ed. B. Pullman, Academic, New York, 1964. 114. H. C. Freeman, in ‘The Biochemistry of Copper’, ed. J. Peisach, P. Aisen and W. E. Blumberg, Academic, New York, 1966. 115. C. A. McAuliffe, in ‘Inorganic Biochemistry’ (Specialist Periodical Reports), ed. H. A. 0.Hill, Royal Society of Chemistry, London, 1979, vol. 1. 116. L. Meriwether and F. H. Westheimer, J. Am. Chem. Soc., 1956,78, 5119. 117. E. Bamann, J. G. Haas and H.Trapmann, Arch. Pharm., 1961,294, 569. 118. E. Bamann and H. Trapman, Adv. Enzymol., 1959, 21, 169. 119. E. Bamann, H. Trapman and A. Rother. Chem. Ber., 1958, 91, 1744; Natunuissenschaften, 1956,43, 326. 120. I. J. Grant and R. W. Hay, Aust. J. Chem., 1965, 19, 1189. 121. T. Nakata, M. Tasumi and T. Miyazawa, Bull. Chem. SOC.Jpn., 1975, 48, 1599. 122. M. Tasumi, S. Takahashi, T. Nakata and T. Miyazawa, BuIL Chem. SOC.Jpn, 1975,48, 1595. 123. L. Lawrence and W. J. Moore, J. Am. Chem. SOC.,1951,73, 151. 124. D. A. Long, T. G. Truscott, J. R. Cronin and R. G. Lee, Trans. Faraday SOC,1971, 67, 1094. 125. J. R. Cronin, D. A. Long and T. G. Truscott, Trans. Faraday Soc., 1971, 67, 2096. 126. K. Ohkubo and H. Sakamoto, Bull. Chem. SOC.Jpn., 1973,46, 2579. 127. M. M. Jones, T. J. Cook and S. Brammer, J. Inorg. Nucl. Chem., 1966, 28, 1265. 128. H. L. Conley and R. B. Martin, J. Phys. Chem., 1965, 69, 2914. 129. A. Nakahara, K. Hamada, I. Miyachi and K. Sakurai, Bull Chens. SOC.Jpn., 1967,40, 2826. 130. A. Nakahara, K. Hamada, Y. Nakao and T. Higashiyama, Coord. Chern. Rev., 1968, 3, 207. 131. K. Hamada, H. Ueda, Y. Nakdo and A. Nakahara, Bull. Chem. SOL Jpn., 1969,42, 1297. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

480

Uses in Synthesis and Catalysis

132. T. Ueki, T. Ashida, Y. Sasada and M. Kakudo, Acta CrystaNogr., Sect. B., 1969, 25, 328. 133. S. Yamada, M. Wagatsuma, Y. Takeuchi and S.Terachima, Chem Pharm. Bull., 1971,19,2380. See also Telrahedron Lett., 1970, 1501. 134. A. Brack, D. Louembe and G. Spach, Origins Lge, 1975, 6,407. 135. M. D. Aiexander and D. H. Busch, Inorg. Chem, 1966,5, 602. 136. M. D. Alexander and D. H. Busch, Inorg. Chem, 1966, 5, 1590. 137. M. D. Alexander and D. H. Busch, J. Am. Chem SOC., 1966, 88, 1130; for work with the t-butyl ester see Y. Wu and D. H. Busch, J. Am. Chem. SOC.,1970,92, 3326. 138. D. A. Buckingham, L. G. Marzilli and A. M. Sargeson, J Am. Chem. Soc., 1967,89,4539. 139. D. A. Buckingham, D. M. Foster and A. M. Sargeson, J. Am. Chem. SOC.,1968,90,6032. 140. D. A. Buckingham, D. M. Foster and A. M. Sargeson, J; Am. Chem. SOC.,1970,92, 5701. 141. R. W. Way and L. J . Porter, J. Chem. SOC.( B ) , 1967, 1261. 142. D. A. Buckingham, J. Dekkers, A. M. Sargeson and M. Wein, J. Am. Cbem. Soc., 1972,94, 4032. 143. D. A. Buckingham, J. Dekkers and A. M. Sargeson, 3. Am. Cbem. Soc., 1973,95,4175. 144. K. Nomiya and H. Kobayashi, Inorg. Chem., 1974, 13, 409. 145. D. A. Buckingham and M. Wein, Inorg. Chem., 1974,13, 3027. 146. D. A. Buckingham, D. M. Foster, L. G. Manilli and A. M. Sargeson, Inorg. Chem., 1970, 9, 11. 147. D. A. Buckingham, D. M. Foster and A. M. Sargeson, J. Am. Chem SOC, 1969,91, 3451. 148. D. A. Buckingham, C. E. Davis and A. M. Sargeson, I. Am. Chem. Soc., 1970,92, 6159. 149. R. W. Hay, R. Bennett and D. J. Barnes, J. Chem. Soc., Dalton Trans., 1972, 1524. 150. R. W. Hay, R. Bennett and D. P. Piplani, 1 Chem. Soc., Dalton Trans., 1978, 1046. 151. K. B. Nolan, B. R Coles and R. W. Hay, J. Chem. SOC,Dalton Trans., 1973, 2503. 152. C. J. Boreham, D. A. Buckingham and F. R. Keene, Inorg. Chem, 1979, 18, 28. 153. D. A. Buckingham and C. R. Clark, Aust. J. Chem, 1982, 35, 431. 154. R. W. Hay and R. Bembi, Inorg. Chem. Acta, 1982,64, t179. 155. See for example D. A. Buckingham and L. M. Engelhardt, J. Am Chem. Soc., 1975,97, 5915; R.3.Martin, J. Inorg. Nucl. Chem, 19?6,38, 511. 156. D. A. Buckingham, D. M. Foster and A. M. Sargeson, J. Am. Chem. Soc., 1969,91,4102. 157. D. A. Buckingham, D. M. Foster and A. M. Sargeson. J Am. Chem. Soc., 1970,92, 6151. 158. D. A. Buckingham. P. J. Moms, A. M. Sargeson and A. Zenella, Inorg. Chem., 1977, 16, 1910. 159. D. A. Buckingham, F. R. Keene and A. M. Sargeson, J. A m Chem. SOC.,1973,95, 5649. 160. C. J. Boreham, D. A. Buckingham and F. R. Keene, Inorg. Chem, 1979, 18, 28. 161. C. J. Boreham, D. A. Buckingham and F. R. Keene, L Am. Chem. SOC.,1979, 101, 1409. 162. C. J. Boreham, D. A. Buckingham and C. R. Clark, Ifiorg. Chem, 1979, 18, 1990. 163. C. J. Boreham and D. A. Buckingham, Aust. 3. Chem., 1980,33, 27. 164. J. P. Collman and D. A. Buckingham, J. Am. Chem. Soc., 1963,85, 3039. 165. D. A. Buckingham and J. P. Collman, Inorg. Chem., 1967, 6, 1803. 166. D. A. Buckingham. J. P. Collman, D. A. R. Happer and L. G. Manilli, J. Am. Chem. SOL,1967, 89, 1082. 167. L. G. Manilli and D. A. Buckingham, Inorg. Chem, 1967,7, 1042. 168- E. Kimura, S. Young and J. P. Collman, Inorg. Chem, 1970,9, 1183. 169. D. A. Buckingbam, C. E. Davis, D. M. Foster and A. M. Sargeson, J. Am. Chem. SOC.,1970,92, 5571. 170. M. D. Fenn and J. H. Bradbury, Anal. Biochem., 1972,49,498. 171. K. W. Bentley and E. H. Creaser, Inorg. Chem., 1974, 13, 1115. 172. E. Kimura, Inorg. Chem., 1974, 13, 951. 173. K. W. Bentley and E. H. Creaser, Biochem. J., 1973, 135, 507. 174. M.-J. Rhee and C.B. Storm, J. Inorg. Biochem., 1979, 11, 17. 175. S. K. Oh and C. B. Storm, Bioinorg. Chem., 1973, 3, 89. 176. M.-J. Rhee and C. B. Storm, J. Inorg. Biochem., 1979, 11, 17. 177. R. D. Gillard and D. A. Phipps, Chem. Commun., 1970, 800. 178. Y. Wu and D. H. Busch, J. Am. Chem SOC, 1972, 94, 4115. 179. A. Y. Girgis and J. 1. Legg, J. Am. Chem. SOC., 1972, 94, 8420. 180. K. Ohkawa, J. Fujita and Y. Shimura, Bull. Chem. Soc. Jpn., 1972,45, 161. 181. L. F. Was-Boas, Ph.D. Thesis, University of Kent, UK, 1974. 182. R. W. Hay and D. P. Piplani, Kern. Kozl., 1977,48,47. 183. D. A. Buckingham, J. MacB. Harrowfield and A. M. Sargeson, J. Am. Chem. SOC.,1973,96, 1726. 184. N. J. Curtis and A. M. Sargeson, J. Am. Chem SOC,1984, 106, 625. 185. S. C. Chan and F. K. Chan, Aust 1. Chem., 1970, 23, 1175. 186. Y. Mitsui, 1. Watanabe, Y. Iitaka and E, Kimura, J. Chem. Soc,, Chem. Commun., 1975, 280. 187. R. W. Hay and R. Bembi, Inorg. Chim. Acta, 19R2. 64, L199. 188. R. K. Murmann and H . Taube, J. Am. Chem. Soc., 1956,78, 4886; D. E. Klimek, B. Grossman and A. Haim, Inorg. Chem., 1972, 11, 2382. 189. A. M Sargeson and H. Taube, Inorg. Chem., 1966, 5, 1094. 190. F. A. Posey and H. Taube, J. Am. Chem. SOC.,1953, 75, 4099; E. Chaffee, T. P. Dasgupta and G. M. Hams, J. Am. Chem. SOC, 1973,95,4169. For a review see D. A. Palmer and R. van Eldik, Chem. Rev., 1983, 83, 651. 191. R. van Eldik and G. M. Hams, Inorg. Chem., 1980, 19, 880. 192. J. MacB. Harrowfield, V. Noms and A. M. Sargeson, J Am. Chem. SOC.,1976,98, 7282. 193. D. A. Buckingham, J. MacB. Harrowfield and A. M. Sargeson, J. Am. Chem. SOC.,1973, 95, 7281. 194. A. U. Fowless and D. R. Stranks, Inorg. Chem., 1977, 16, 1271, 1276, 1282. 195. H. Gamsjager, G. A. K. Thompson, W. Sagrnutler and A. G . Sykes, Inorg. Chem., 1980, 19, 997. 196. R. S. Taylor, Inorg. Chem., 1977, 16, 116. 197. T, G Beech, N. C. Lawrence and S . F. Lincoln, A w t . J. Chern., 1973, 26, 1877. 198. R.K. Wharton, R. S. Taylor and A. G. Sykes, Inorg. Cbem., 1975, 14, 33. 199. D. A. Buckingham and C. R. Clark, Aust. J. Chern., 1982,35, 431.

Lewis Acid Catalysis and the Reacfions of Coordinated Ligands

48 1

D. A. Ruckingharn and L. M. Engelhardt, J. Am. Chem. Sac+, 1975, 97, 5915. R. W. Hay and R Bembi, Inorg. Chim. Acta, 1982,64, L179. D. A. Buckingham and C. R Clark, J. Chem. SOC.,Dalton Trans., 1979, 1757. D. A. Buckingham, L. G. Manilli and A. M. Sargeson, J. A m Chem Soc., 1967,89, 2772. J. P. Collman and E. Kimura, J. Am. Chem. Sac., 1967,89, 6096. D. A. Buckingham, J. Dekkers, A. M. Sargeson and M. Wein, 1 Am. Chem. SOC.,1972,94, 4032. D. A. Buckingham, P. A. ManilIi, I. E. Maxwell and A. M. Sargeson, Chem. Commun., 1968, 488. See for example H. Wautier, V. Datte, M.-N. Smets and J. Fastrez, J. Chem. Sac., Dalton Trans., 1981, 2479, and H. Wautier, D. Marchal and J. Fastrez, J. Chem. SOC.,Dalton Trans., 1981, 2484. 208. C. R. Clark, R. F. Tasker and D. A. Buckingham, J. Am. Chem. SOC., 1981, 103, 7023. 209. D. R. Knighton, D. R K. Harding, M. J. Firar and W. S. Hancock, J. Am. Chem. Soc., 1981, 103, 7025. 210. S. S. Isied, A. Vassilian and J. M. Lyon, J. A m Chem SOC.,1982, 104, 3910. 211. E. J. Corey and R. L. Dawson, J. Am. Chem. SOC,1962, 84, 4899. 212. C. R. Wasmuth and H. Freiser, Talanta, 1962, 9, 1059. 213. R. H. Barca and H. Freiser, 1 Am. Chem. SOC.,1966,88, 3744. 214. R. W. Hay and C. R. Clark, J. Chem. SOC.,Dalton Trans., 1977, 1993. 215. J. Suh, M. Cheong and M. P. Suh, J. Am. Chem. SOC.,1982, 104, 1654; J. Suh, E. Lee and E. S. Jang, Inorg. Chem., 1981, 20, 1932. 216. R. P. Houghton and R. R Puttner, Chem. Commun., 1970, 1270. 217. R. M. Propst, 111 and L.S. Trzupek, J. Am. Chem. SOC.,1981, 103, 3233. 218. R. W. Hay and C. R. Clark, J. Chem. SOC.,Dalton Trans., 1977, 1866. 219. K. H. Gerber and R G. Wilkins, ACS Meeting, Dallas, Texas, 1973, Abstract 173. 220. T. H. Fife and T. J. Przystas, 1. Am. Chem. SOC.,1980, 102, 7297. 221. T. J. Przystas and T. H. Fife, J. Am. Chem. SOC., 1980, 102,4391. 222. T. H. Fife and T. J. Przystas, J. Am. Chem. SOC.,1982, 104, 2251. 223. T. H. Fife and V. L. Squillacote, J. Am. Chem. SOC.,1978, 100, 4787. 224. T. H. Fife, T. J. Przystas and V. L. Squillacote, J. Am. Chem. SOL, 1979, 101, 3017. 225. T. H. Fife and V. L. Squillacote, J. Am. Chem SOC,1977,99, 3762. 226. K. Ogino, K. Shindo, T. Minami, W. Tagaki and T. Eiki, Bull. Chem SOC.Jpn., 1983, 56, 1101. 227. R. F. Boyer, J. Znorg. Nucl. Chem., 1980,42, 155. 228. P. Domiano, P. L. Messori, C. Pelizzi and G. Predieri, Inorg. Chem. Acta, 1983, 70, 21. 229. M. A. Wells, G. A. Rogers and T. C. Bruice, J. Am. Chem. Sw,1976, 98, 4336. 230. P. R. Woolley, Nature (London),1975, 258, 677. 231. R. H. Prince, D. A. Stotter and P. R. Woolley, Inorg. Chem. Acta, 1974, 9, 51. 232. M. R. Caira, L. R. Nassimbeni and P. R. Woolley, Acta Crysrullogr., Sect. B, 1975, 31, 1334. 233. E. Breslow, in ‘The Biochemistry of Copper’, ed. J. Peisach, P.Aisen and W. E. Blumberg, Academic, New York, 1966, pp. 149-156. 234. R. Breslow, D. E. McClure, R. S. Brown and J. Eisenach, J. Am. Chem. Soc., 1975, 97, 194. 235. J. T. Groves and R. R. Chambers, J. Am. Chem. Soc., 1984,106, 630. 236. E. Bamann and H. Trapmann, Adv. EnzymoL, 1959, 21, 169. 237. T. C. Bruice and S. J. Benkovic, ‘Bioorganic Mechanisms’, Benjamin, New York, 1966, vol. 2, chaps. 5, 6 and 7. 238. J. R. Cox and 0. B. Ramsay, Chem. Rev., 1964,64, 317. 239. C. A. Bunton, J. Chem. Educ., 1968, 45, 21. 240. J. Emsley and D. Hall, ‘The Chemistry of Phosphorus’, Harper Row, London, 1975. 241. F. H. Westheimer, AEC. Chem. Res., 1968, 1, 70. 242. P. Gillespie, F. Ramirez, I. Ugi and D. Marquading, Angew. Chem, Int. Ed. Engl, 1973, 12, 91. 243. R. F. Hudson, ‘Structure and Mechanism in Organophosphorus Chemistry’, Academic, New York, 1965. 244. B. J. Walker, ‘Organophosphorus Chemistry’, Penguin, London, 1972. 245. R. D. Cook, C. E. Diebert, W. Schwarz, P. C. Turley and P. Haake, J. Am. Chem. Sac., 1973, 95, 8688. 246. J. F. Momson and E. Heyde, Annu. Rev. Biochem., 1972,41, 29. 247. A. S. Mildvan, Enzymes, 1970,2, 446. 248. J. Imsande and P. Handler, Enzymes, 1961, 5, 281. 249. B. S. Cooperman, in ‘Metal Ions in Biological Systems’, ed. H. Sigel, Dekker, New York, 1976, vol. 5 . 250. A. S. Mildvan and G. M. Grisham, Struct. Bonding (Berlin), 1974, 1, 20. 251. T. G. Spiro, in ‘Inorganic Biochemistry’, ed. G. L. Eichhorn, Scientific Publishing, New York, 1973, VOI. 1, chap. 17. 252. P. J. Briggs, D. P. N.Satchel1 and G. F. White, J. Chem. Soc. ( E ) , 1970, 1008. 253. C. H. Oestriech and M. M. Jones, Biochemistry, 1967, 6, 1515. 254. C. H. Oestreich and M. M. Jones, Biochemistry, 1966, 5, 2926. 255. D. E. Koshland, J. A m Chem. Soc., 1952,74, 2286. 256. C. H. Oestreich and M. M. Jones, Biochemistry, 1966,5, 3151. 257. J. P. Klinman and D. Samuel, Biochemistry, 1971, 10, 2126. 258. G. D. Sabato and W. P. Jencks, J. Am. Chem. SOC.,1961,83, 4393, 4400. 259. D. A. Buckingham and C. R Clark, Aust. J. Chem., 1981,34, 1769. 260. M. Watanabe, Bull. Chem. SOC.Jpn., 1982, 55, 3766. 261. M. Watanabe, M. Matsuura and T. Yamada, Bull. Chem SOC.Jppn., 1981, 54, 738. 262. E. Thilo and W. Wieker, J. Polym. Sci., 1961, 53, 55. 263. W. Wieker and E. Thilo, 2.Anorg. Allg. Chem., 1960, 306, 48. 264. W. Wieker and E. Thilo, Z. Anorg. Allg. Chem., 1961, 313, 296. 265. T.Eiki and W. Tagaki, Chem Lett., 1981, 1465. 266. T.Eiki, T. Horiguchi, M. Ono, S. Kawada and W. Tagaki, J. Am. Chem. Soc., 1982, 104, 1986, 267. S. J. Benkovic and L. K. Dunikoski, Jr., J. Am. Chem. Soc., 1971, 93, 1526. 268. Y. Murakami and A. E. Martell, J. Phys. Chem., 1963, 67, 582. 269. R. Hofstetter, Y. Murakami, G . Mont and A. E. Martell, J. Am. Chem. SOC.,1962, 84, 3041.

200. 201. 202. 203. 204. 205. 206. 207.

Uses in Synthesis and Catalysis

482

Murakami and M.Takagi, J. Am. Chem. Soc., 1969, 91, 5130. Murakami, J. Sunamoto and Y. Sadamori, Chem. Comrnun, 1969, 983. Murakarni and J. Sunamoto, Bull. Chem. SOC.Jpn., 1971, 44,1827. H. Bromilow and A. J. Kirby, J. Chem. SOC.,Perkin Trans. 2, 1972, 149. R. W. Hay and A. K. Basak, Inorg. Chim. Acta, 1983, 79, (B7), 255; R. W. Hay, A. K. Basak, M. P. Pujari and A. Perotti, J. Chem. Soc., Dalton Trans., 1986, 2029. 275. J. E. Loran and P. A. Naylor, J. Chem. Soc., Perkin Trans. 2, 1977, 418. 276. C.-M. Hsu and B. S. Cooperman, J. Am. Chem SOC,1976,98, 5652, 5659. 277. M. M. Taqui Khan and A. E. Martell, J. Am. Chem. Sac., 1962,84, 3037. 278. M. Tetas and J. M. Lowenstein, Biochemistry, 1963, 2 , 350. 279. N. Nelson and E. Racker, Biochemistry, 1973, 12, 563. 280. F. J, Farrell, W. A. Kjellstrom and T. G. Spiro, Science, 1969, 164, 320. 281. E. A. Dennis and F. H. Westheimer, J. Am. Chern. Soc., 1966, 88, 3422. 282. S. F. Lincoln and D. R. Stranks, Aust. J. Chem., 1968, 21, 57, 283. B. Anderson, R. M. Milburn, J. MacB. Harrowfield, G. 8. Robertson and A. M. Sargeson, J. Am. Chem. Soc., 1977, 99,2652. 284. A. Steitz and W. N. Lipscomb, J. Am. Chem. SOC.,1965,87, 2488. 285. C. L. Coulter, J. Am. Chem. SOC.,1963, 95, 570. 286. D. R Jones, L. F. Lindoy, A. M. Sargeson and M. R Snow, Inorg. Chem., 1982,21, 4155. 287. D. R. Jones, L. F. Lindoy and A. M. Sargeson, J. Am Chem. Soc., 1983, 105, 7327. 288. J. MacB. Harrowfield. D. R. Jones, L. F. Lindoy and A. M. Sargeson, J. Am. Chem. SOC, 1980, 102, 7733. 289. P. R. Norman and R. D. Cornelius, J. Am. Chem Sac., 1982, 104, 2356. 290. R. D. Cornelius and P. R. Norman, Inorg. Chim. Acza, 1982, 65, L193. 291. R. W. Hay and R. Bembi, Inorg. Chim. Acta, 1983, 78, 143. 292. W. G. Jackson and B. C. McGregor, Inorg. Chem. Acta, 1984,83, 115. 293. S. H. McClaugherty and C. M. Grisharn, Inorg. Chem., 1982,21,4133. 294. E. A. Merritt and M. Sundaralingam, Acta Crystullogr., Sect. B, 1980, 36, 2576. 295. E. A. Merrio and M. Sundaralingam, Acta Crysrallogr., Seck B, 1981, 37, 1505. 296. E. A. Merritt, M. Sundaralingam and R. D. Cornelius, Acra Crystnhgr., Sect. B, 1981, 37,657. 297. R. D. Cornelius. I.nnrg. C h ~ m1980, . ~ 19;1286. 298. R. D. Cornelius, Inorg. Chim. Acta, 1980, 46, L109. 299. E. A. Merritt, M. Sundaralingarn, R. D. Cornelius and W. W. Cleland, Biochemistry, 1978, 17,3274. 300. P. W. A. Hubner and R. M. Milburn, Inorg. Chem., 1980, 19, 1267. 301. S. Suzuki, S. Kimura, T. Higashiyama and A. Nakahara, Bioinorg. Chem., 1974, 3, 183. 302. M. L. De Pamphilis and W. W. Cleland, Biochemistry, 1973, 12, 3714. 303. K. D. Danenberg and W. W. Cleland, Biochemistry, 1975, 14, 28. 304. C. A. Janson and W. W. Cleland, J. Bid. Chem, 1974, 249, 2562, 2567, 2572; M. I. Schimerlik and W. W. Cleland, J. Bid. Chem. 1973,248,8418; D. A. Armbruster and F. B. Rudolph, J. Bid. Chem., 1976,251,320; J. Bar-Tana and W. W. Cleland, 3. Bid. Chem., 1974, 249, 1271. 305. R. D. Cornelius, P. A. Hart and W. W. Cleland, Inorg. Chem., 1977, 16, 2799. 306. D. G. Gorensrein, J. Am. Chem. Soc., 1975, 97, 898. 307. E. Bamann, Angew. Chem., 1939, 52, 186. 308. E. Barnann and M. Meisenheimer, Chem. Ber., 1938,71, 1711, 1980, 2086, 2233. 309. E. Barnann, F. Fischer and H. Trapmann, Biochern. Z., 1951, 32$, 413. 310. W. W. Butcher and F. H. Westheirner, J. Am. Chem, Snc, 1955, 77, 2420, 311. H. Trapmann, Arzneim-Forsch., 1959, 9,341, 403. 312. R. Bresiow, R. Fairweather and J. Keana, J. Am. Chem. Soc., 1967, 89, 2135. 313. R. Breslow and M. Schmir, J. Am. Chem. Soc., 1971, 93, 4960. 314. D. Pinnell, G. B. Wright and R. B. Jordan, J. Am. Chem. SOL, 1972, 94, 6104. 315. D. A. Buckingham, F. R. Keene and A. M. Sargeson, J. Am. Chem. Soc., 1973, 95,5649. 316. A. W. Zannell and P. C. Ford, Inorg. Chem., 1975,14, 42. 317. R. E. Clarke and P. C. Ford, Inorg. Chem:, 1970,8, 227, 495; P. C. Ford, Chem. Comrnun., 1971, 7. 318. I. I. Creaser and A. M. Sargeson, J. Chem SOC.Chem Commun., 1975, 974. 319. I. I. Creaser, S. F. Dyke, A. M. Sargeson and P. A. Tucker, J. Chem. Soc., Chem. Cornmum, 1978, 289. 320. D. G. Butler, I. I. Creaser, S. F. Dyke and A. M. Sargeson, Acta Chem. Scand., Ser. (A), 1978, 32, 789. 321. W. R. Ellis and W. L. Purcell, Inorg. Chem., 1982, 21, 834. 322. 1. 1. Creaser, J. MacB. Harrowfieldd, F. R. Keene and A. M. Sargeson, J. Am. Chem. Soc, 1981, 103, 3559. 323. W. P. Norris, J. Org. Chem., 1962, 27, 3248. 324. W. Fleming, J. W. Fronabarger, M. L. Leiberman and V. M. Loyola, Second Chemical Conference of the North American Continent, Las Vegas, NV, August 1980, American Chemical Society, Washington, DC, Abstract INOR 13. 325. R. J. Balahura and W. L. Purcell, Inorg. Chem., 1981, 20, 1459. 326. R. Lopez de la Vega, W. R. Ellis, Jr. and W. L. Purcell, Inorg. Chim. Acta, 1983, 68, 97. 327. D. A. Buckingham, B. M. Foxman, A. M. Sargeson and A. Zannella, J. Am. Chem. Soc., 1972,94, 1007. 328. R W. Hay and K. B. Nolan, J. Chem. Soc., Dalton Trans., 1974, 914. 329. D. A. Buckingham, P. Morns, A. M. Sargeson and A. Zanella, Inorg. Chem., 1977, 16, 1910. 330. B. Anderes and D. K. Lavallee, Inorg. Chem., 1983, 22, 3724. 331. C. R. Ciark and R. W. Hay, J. Chem. Sac, Dalfon Trans., 1974, 2148. 332. B. N. Storhoff and H. Lewis, Coord. Chem. Rev., 1977, 23, 1. 333. R. Ros, M. Lenarda, T. Boschi and R. Roulet, Inorg. Chim. Acta, 1977, 25, 61. 334. R. Ros, R. A. Michelin, T. Boschi and R. Rouiet, Inorg. Chim. Acta, 1979, 35, 43. 335. L. Calligaro, Polyhedron, 1984, 3, 117. 336. D. A. Buckingham, C. R. Clark, 8. M. Foxman, G. J. Gainsford, A. M.Sargeson, M. Wein and A. Zanella, Inorg. Chem., 1982,21, 1986.

270. 271. 272. 273. 274.

Y. Y. Y. R.

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

483

N. J. Curtis and A. M. Sargeson, J. Am. Chem. SOC.,1984, 106, 625. R. W. Hay, in ‘Metal Ions in Biological Systems’, ed. H.Sigel, Marcel Dekker, New York, 1976, vol. 5 . See for example W. D. Covey and D. L. Leussing, J. Am. Chem. SOC.,1974,96, 3860 and papers listed. See for example G Kubala and A. E. Martell, J. Am. Chem. Soc., 1982, 104, 6602 and papers listed. R. Steinberger and F. H.Westheimer, J. Am. Chem. SOC.,1951, 73,429. R. J. Dummel, M. N. Berry and E. Kun, Mol. Pharmacol., 197t, 7 , 367. J. E. Prue, J. Chem. SOC., 1952, 2331. R. W. Hay and K. N. Leong, J. Chem. SOC.( A ) , 1971, 3639. D. W. Larsen and M. W. Lister, Can. J. Chem., 1968, 46, 823. R. W. Hay and S. J. Harvie, Aust. J. Chem., 1965, 18, 1197. K. J. Pedersen, Acta Chem. Scand., 1955,9, 1640. K. J. Pedersen, Acra Chem. Scand., 1958, 12,919. A. Kornberg, S. Ochoa and A. Mehler, J. Biol. Chem., 1948, 174, 159. K. 3. Pedersen, J. Am. Chem SOC., 1929, 51, 2098. K. J. Pedersen, J. Am. Chem. SOC.,1936, 58, 240. E. Gelles and R. W. Hay, J. Chem. Soc., 1958, 3673. B. R. Brown, Q. Rev., Chem. Soc., 1951, 5, 131. C. Reyes-Zamora and C. S. Tsai, Chem. Commun., 1971, 1047. G. Kubala and A. E. MartelI, Inorg. Chem., 1982, 21, 3007. N. V. Raghaven and D. L. Leussing, J. Am. Chem. Soc., 1974,%, 7147. W. D. Covey and D. L. Leussing, J. Chem. SOC.,Chem. Commun., 1974, 31. N. V. Raghaven and D. L. Leussing, J. Am. Chem. Sac., 1977,99,2188. J. V. Rund and R. A. Plane, J. Am. Chem. SOC., 1964,86, 367. K. N. Leong and M. W. Lister, Can. J. Chem., 1972,50, 1461. D. P. N. Satchell and I. I. Secemski, J. Chem. SOC.( B ) , 1970, 1306. D. P. N. Satchell, M. N. White and T . J. Weil, Chem. Ind. (London), 1975, 791. A. J. Hall and D. P.N. Satchell, J. Chem. SOC., Perkin Trans. 2, 1976, 1274, 1278. See for example D. Gravel, c. Vazin and S. Rahal, J. Chem. Soc., Chem. Commun., 1972, 1323; T.-L. Ho and C. M. Wong, Can. J. Chem., 1972, 50, 3740. 365. L. R. Fedor, J. Am. Chem. SOC,1968,90, 7266. 366. L. R. Fedor and B. S. R. Murty, J. Am. Chem. Soc., 1973,95, 8407. 367. D. P. N. Satchell and L. 2. Zdunek, unpublished results quoted in ref. 361. 368. A. J. Hall and D. P. N. Satchel], J. Chem. SOC., Perkin Trans. 2, 1975, 778, 953, 1273, 1351. 369. A. J. Hall and D. P. N. Satchel!, J. Chem. Soc., Perkin Trans. 2, 1977, 1366. 370. 0. M. Peters, N. M. Blaton and C. J. DeRanter, J. Chem. Soc., Perkin Trans. 2, 1978, 23. 371. J. M. Harrowfield and A. M. Sargeson, J. Am. Chem. Soc., 1974,%, 2634. 372. B. T. Golding, J. M. Harrowfield and A. M. Sargeson, J. Am. Chem. Soc., 1974,96, 3003. 373. B. T. Golding, J. M. Harrowfield, G. B. Robertson, A. M. Sargeson and P. 0. Whimp, J. Am. Chem. SOC, 1974,96, 3691. 374. J. D. Bell, A. R. Gainsford, B. T. Golding, A. J. Herlt and A. M. Satgeson, J. Chem. SOC.,Chem. Commun., 1974,980. 375. A. R. Gainsford and A. M. Sargeson, Ausf. J. Chem., 1978, 31, 1479. 376. D. A. Buckingham, J. M. Harrowfield and A. M. Sargeson, J. Am Chem Soc., 1973,95, 7281. 377. I. P. Evans, G. W. Everett, Jr. and A. M. Sargeson, J. Chem. Snc., Chem. Commun., 1975, 139. 378. K. Schug and C. P. Guengerich, J. Am. Chem. Soc., 1975,97,4135. 379. D. Wayshort and G. Navon, Chem. Commun., 1971, 1410. This value may be erroneous; see J. N. h e r . J. Inorg. Nucl. Chem., 1973, 35, 2067. 380. A. R. Gainsford, R. D. Pizer, A. M. Sargeson and P. 0. Whimp, J. Am Chem. Soc., 1981, 103, 792. 381. P. D. Ford, IL B. Nolan and D. C. Povey, Inorg. Chim. Acra, 1982, 61, 189. 382. A. M. Sargeson, Chem. Br., 1979, 15, 23; see also J. Am. Chem SOL, 1977,99, 3181; 1982, 104, 6016. 383. For a discussion see ‘Coordination Chemistry of Macrocyclic Compounds’, ed. G. A. Melson, Plenum, New York, 1979. 384. J. MacB. Harrowfield and A. M. Sargeson, J. Am. Chem. Soc., 1979, 101, 1514. 385. E. Rotondo, R. Pietropaola, G. Tresoldi, F. Faraone and F. Cusmano, Inorg. Chem. Acfa, 1976, 17, 181. 386. E. Rotondo, R. Pietropaola and F. Cusmano, Inorg. Chem. Acfa, 1978, 26, 189. 387. E. Rotondo and F. Priolo, J. Chem. SOC.,Dalton Trans., 1982, 1825. 388. J. MacB. Harrowfield, A. M. Sargeson, J. Springborg, M. R. Snow and D. Taylor, Inorg. Chem., 1983, 22, 186. 389. D. L. Leussing, in ‘Metal Ions in Biological Systems’, ed. H. Sigel, Dekker, New York, 1976, vol. 5 , p. 1. 390. G. L. Eichhorn and J. C. Bailar, J. Am. Chem. Soc., 1953, 75, 2905. 391. G. L. Eichhorn and i. M. Trachtenberg, J. Am. Chem. Soc., 1954, 76, 5183. 392. G. L. Eichhorn and N. D. Marchand, L Am. Chem. Soc., 195&78, 2688. 393. D. H. Busch and J. C. Bailar, J. Am. Chem. SOC.,1956, 78, 1137. 394. D. F. Martin and F. F. Cantwell, J. Inorg. Nucl. Chem., 1964,26, 2219. 395. L. J. Numez and G. L. Eichhorn, J. Am. Chem SOC.,1962, 84, 901. 396. D. L. Leussing and C. K. Stanfield, J. Am. Chem. Soc., 1966, 88, 5726. 397. A. C. Dash and R K. Nanda, J. Am. Chem. Soc., 1969, 91, 6944. 398. L. F. Lindoy, Q. Reu., Chem. SOC.,1971, 25, 379. 399. C. M. Harris, S. L. Lenzer and R. L. Martin, Aust. J. Chem., 1961, 14, 420. 400. C. V. McDonnell, Dim. Abstr., 1968, 28B, 3242. 401. R. W. Hay and P. R. Norman, unpublished observations. 402. A. C. Braithwaite, C. E. F. Rickard and T. N. Waters, 1,Chem. Soc., Dalton Trans., 1975, 2149. 403. E. Hoyer and B. Lorenz, Z. Anorg. Allg. Chem., 1965, 336, 192. 404. V. W. Skopenko and E. Hoyer, 2. Anorg. Allg. Chem., 1965,339,214. 405. E. Hoyer, Nafurwissenachafren, 1959, 1.

337. 338. 339. 340. 341. 342. 343. 344. 345. 346. 347. 348. 349. 350. 351. 352. 353. 354. 355. 356. 357. 358. 359. 360. 361. 362. 363. 364.

Uses in Synthesis and Catalysis

484

Hoyer, 2. Chem, 1965, 231. W. Hay and K.B. Nolan, J. Chem. Soc., Dalton Trans., 1976, 548. C. Dash, B. Dash and S. Praharaj, J. Chem. SOL,Dalton Trans., 1981, 2063. C. Dash, B. Dash, P. K. Mahapatra and M. Patra, J. Chem. SOC.,Dalton Trans., 1983, 1503. N. P. Gemmantel, E. W. Gowling and M. 1. Page, J. Chem. SOC.,Perkin Trans. 2, 1978, 375. N. P. Gensmantel, P. Proctor and M. I. Page, J. Chem SK, Perkin Trans. 2, 1980, 1725. R. W. Hay and A. K. Basak, unpublished results. W. A. Cressman, E. T. Sugita, J. T. Doluisio and P. J. Niebergall, J. Pharm. Sci., 1969, 58, 1471. N. P.Gemmantel, D. McLellan, J. J. Morris, M. I. Page, P. Proctor and G. S. Randahawa, in ’Recent Advances in the Chemistry of P-Lactam Antibiotics’, ed. G . Gregory, Royal Society of Chemistry, London, 1980. 415. J. Fisher and J. R Knowles, Annu. Rep. Med. Chem., 1978, 13, 239. 416. S. Kuwahara and E. P. Abraham, Biochem. 1,1967, 103, 27C. 417. J. T. Groves and R. M. Dias, J. Am. Chem. Soc, 1979, 101, 1033. 418. For reviews see E. T. Kaiser and B. L. Kaiser, Acc. Chem Res., 1972, 5, 219; W. N. Lipscomb, Tetrahedron, 1974, 30, 1725. 419. M. W. Makinen, L. C. Kuo, J . J. Dymowski and S. Jaffer, J. Biol. Chem., 1979,254, 356. 420. R. Breslow, D. E. McClure, R. S. Brown and J. Eisenach, J. Am. Chem. SOC.,1975, 97, 194. 421. R. Breslow and D. Chipman, J. Am. Chem. Soc., 1965, 87, 4195. 422. T. H. Fife and T. J. Przystas, J. Am. Chem. SOC.,1983, 105, 1638. 423. J. N. BeMiHar, Adv. Curbohydr. Chem., 1967, 22, 25. 424. B. Capon, Chem. Rev., 1969, 69, 407. 425. G. Wada and M. Sakamoto, Bull. Chem. Soc. Jpn., 1973,46, 3378. 426. C. R. Clark and R. W. Hay, J. Chem. Soc., Perkin Trans. 2, 1973, 1943. 427, T. J. Pnystas and T. H. Fife, J. Am. Chem. Soc., 1980, 102, 4391. 428. L. R Fedor and B. S. R. Murty, J. Am. Chem. SOC.,1973, 95, 8407. 429. D. P.N. Satchel1 and T. J. Weil, Inorg. Chem. Acra Lett., 1978, 29, L239. 430. R. J. Ferrier, R. W. Hay and N. Vethaviyasar, Carbohydr. Res., 1973, 27, S5. 431. C. J. OConnor, A. L. Odell and A. A. T.Bailey, Ausr J. Chern., 1982,35, 951. 432. C. J. OConnor, A. L. Odell and A. A. T. Bailey, Ausl. J. Chem., 1983,36, 279. 433. B. E. Tilley, D. W. Porter and R. W. Gracey, Carbohydr. Res., 1973, 27, 289. 434. J. L. Kice and J. M. Anderson, 1.Am. Chem SOC.,1966, 88, 5242. 435. S. I. Benkovic and P. A. Benkovic, J. Am. Chem. Sm., 1966,88, 5504. 436. S. 5. Benkovic, J. Am. Chem. SOC.,1966,88, 5511. 437. R. W. Hay, C. R. Clark and J. A. G. Edmonds, 1. Chem. SOL,Dalton, Trans, 1974, 9. 438. R. D. Gillard, Irrorg. Chim. Acta Rev., 1967, 1, 69. 439. S. Akabori, T. T. Otani, R. Marshall, M. Winitz and J. P. Greenstein, Arch. Biochem. Biophys., 1959, 83, 1. 440. T. Ichikawa, S. Maeda, Y. Araki and Y. Ishido, J. Am. Chem. SOC.,1970,92, 5514. 441. A. Nakahara, S. Nishikawa and J. Mitani, Bull. Chem. SOC.Jpn., 1967,40, 2212. 442. M. Murakami and K. Takahashi, Bull. Chem. Soc Jpn., 1959,32, 308. 443. I. G. Browning, R. D. Gillard, J. R. Lyons, P. R. Mitchell and D. A. Phipps, J. Chem. Soc., Dalton Trans., 1972, 1815. 444. E. E. Snell, P. M. Fasella, A. Braunstein and A. Rossi-Fanelli, ‘Chemical and Biological Aspects of Pyridoxal Catalysis‘, Macklillan, New York, 1963. 445. T. C. Bruice and S. J. Benkovic, ‘Bio-Organic Mechanisms’, Benjamin, New York, 1966, Vol. 2. 446. A. E. Martell, in ‘Metal Ions in Biological Systems’, ed. H. Sigel, Dekker, New York, 1973, vol. 2, pp. 207-293. 447. D. Hopgood, J. Chem. SOC.,Dalton Trans., 1972, 482. 448. E. M. Abbott and A. E. Martell, J. Am. Chem. SOL, 1970, 92, 5845. 449. 0. A. Gansow and R. H. Holm, J. Am. Chem S m , 1969, 91, 573. 450. 0. A. Gansow and R. H. Holm, J. Am. Chem Soc, 1969,91, 5984. 451. R D. Gillard and D. A. Phipps, Chem. Commun., 1970, 800. 452. D. H. Williams and D. H. Busch, J. Am. Chem. SOC.,1965, 87, 4644. 453. D. A. Buckingham, L. G. Marzilli and A. M. Sargeson, J. Am. Chem. SOC.,1967, 89, 5133. 454. J. D. Sudmeirer and G. Occupati, Inorg. Chem., 1968, 7,2524. 455. R. D. Gillard, P. R. Mitchell and N. C. Payne, Chem. Commun., 1968, 1150. 456. R. D. Gillard, S. H. Laurie, D. C. Price, D. A. Phipps and C.F. Weick, J. Chem Sot., Dalron Trans., 1974, 1385. 457. L. E. Erickson, A. J. Dapper and J. C. Uhlenhopp, J. Am. Chem. SOC.,1969, 91, 2510. 458. P. R. Norman and D. A. Phipps, Inorg. Chem Acta, 1977, 24, L35. 459. L. G. Stddtherr and.R. J. Angelici, Inorg. Chern., 1975, 14, 925. 460. G. G. Smith, A. Khatib and G. S. Reddy, J. Am. Chem Soc,, 1983, 105, 293. 461. E. E. Snell, A. E. Braunstein, E. S. Severin and Y. M. Torchinsky, ‘Pyridoxal Catalysis: Enzymes and Model Systems’, Wiley, New York, 1968. 462. R. H. Holm, in ‘Inorganic Biochemistry’, ed. G. L. Eichhorn, Elsevier, Amsterdam, 1973, vok. 2, p. 1137. 463. P. J. Lawson, M. G. McCarthy and A. M. Sargeson, J. Am. Chem. SOC.,1982, 104, 6710. 464. N. E. Dixon and A. M. Sargeson, J. Am. Chem. SOC.,1982, 104, 6716. 465. M. Sato, K. Okawa and S. Akabori, Bull. Chem. SOC.Jpn., 1957, 30, 937. 466. S. Akabori, T. T. Otani, R. Marshall, M. Winitz and J. P. Greenstein, Arch. Biochem. Biophys., 1959, 83, 1. 467. L. Benoiton, M. Winitz, R. F. Coleman, S. M. Birnbaum and J. P. Greenstein, J. Am. Chern. SOC.,1959, 81, 1726. 468. Y. Ikutani, T. Okuda and S. Akabori, Bull. Chem. Soc. Jpn., 1960, 33, 582. 469. R. D. Gillard and P. M. Harrison, J. Chem. Sac. ( A ) , 1967, 1957. 470. J. R. Brush, R. J. Magee, M. J. O’Connor, S. 8. Teo, R J. Geue and M. R. Snow, J. Am. Ckem. SOL,1973,95,2034. 471. R. D. Gillard, S. H. Laurie, D. C. Price, D. A. Phipps and C. F. Weick, J. Chem. SOL, Dulton Trans., 1974, 1385. 472. M. J. O’Connor, J. F. Smith and S. B. Teo, Atrsr. J. Chem., 1976, 29, 375. 473. K. Noda, M. Bessho, T. Kato and N. Izumiya, Bull. Chem. SOL,Jpn., 1970, 43, 1834. 474. M. Fujioka, Y. Nakao and A. Nakahara, J. Inorg. NucI. Chem., 1977,39, 1805.

406. 407. 408. 409. 410. 411. 412. 413. 414.

E. R. A. A.

Lewis Acid Catalysis and the Reactions of Coordinated Ligands

485

M. Murakami and K. Takahashi, Bull. Chem. SOC.Jpn., 1959, 32, 308. J. C. Dabrowiak and D. W . Cooke, Inorg. Chem., 1975, 14, 1305. D. A. Phipps, Xnorg. Chim. Acta, 1978, 27, L103. K. Harada and J. Oh-hashi, J. Org. Chem., 1967, 32, 1103. T. Ichikawa, S. Maeda, Y. Arakai and Y. Ishido, J. Am. Chem S o c , 1970,92, 5514. T. Ichikawa, S. Maeda, T. Okamoto, Y. Araki and Y. Ishido, BulL Chem. SOC.Jpn., 1971, 44, 2779. S. Ohdan, T. Ichikawa, Y.Araki and Y. Ishido, Bull. Chem. SOC.Jpn, 1974, 47, 1295. S. Suzuki, H. Narita and K. Harada, J. Chem. SOC.,Chem Commun, 1979, 29. P. Sharrock and C. H. Eon, J. Inorg. Nucl. Chem., 1979,41, 1087. L. Cassella, A. Pasini, R. Ugo and M. Visca, J. Chem. Soc., Dalton Trans., 1980, 1655. R. J. Geue, M. R. Snow, J. Springborg, A. J. Hert, A. M. Sargeson and D. Taylor, J. Chem. Sw, Chem Commun., 1976, 285. 486. J. P. Aune, P. Maldonado, G. Larcheres and M. Peirrot, G e m . Commun., 1970, 1351. 487. R. W. Hay, Ausr. J. Chem, 1964, 17, 759. 488. R. W. Hay and C. R Clark, Transition Mer. Chem., 1979, 4, 28. 489. J. A. Bertrand and D. Caine, J. Am. Chem. Soc., 1964, 86, 2298. 490. D. S. Sigman and C. T. Jorgenson, J. Am. Chem. SOC.,1972,94, 1724. 491. R. P. Houghton and C. S. Williams, Tetrahedron Lett., 1967, 5091. 492. P. Pfeiffer, W. Offerman and H. Werner, J. Prakt. Chem., 1942, 159, 313. 493. H. S. Verter and A. E. Frost, J. Am. Chem. SOC.,1960, 82, 85. 494. R. P. Houghton and D. J. Pointer, J. Chem. SOC.,1964, 3302. 495. G. A. Auld and A. Davidson, Inorg. Chem., 1968, 7, 306. 496. R. D. Gillard and R. Wootton, J. Chem. SOC.( E ) , 1970, 364. 497. Y. Nakao, M. Demichi and A. Nakahara, Bull Chem. SOC Jpn., 1980, 53, 1564. 498. Y. Nakao and A. Nakahara, Chem Lett., 1977, 145. 499. R. P. Hanzlik and W. J. Michaely, J. Chem. SOC.,Chem. Commun., 1975, 113. 500. R. P. Hanzlik and A. Hamburg, J. Am. Chem. Sac., 1978, 100, 1745. 501. R. P. Hanzlik, M.Edelman, W. J. Michaely and G. Scott, J. Am. Chem Soc., 1976,98, 1952. 502. N. E. Dixon, C. Gazzola, R. L. Blakeley and B. Zerner, .I. Am Chem. SOL, 1975, 97,4131. 503. N. E. Dixon, P. W.Riddles, C. Gazzola, R. L. Blakely and 8. Zerner, Can J. Biochem., 1980, 58, 1335. 504. R. L. Blakeley, A. Treston, R. K. h d r e w s and B. Zerner, J. Am. Chem. Soc., 1982, 104, 612. 505. N. J. Curtis, N. E. Dixon and A. M. Sargeson, J. Am. Chem. Soc., 1983, 105, 5347. 506. N. E. Dixon, W. G. Jackson, W. Marty and A. M. Sargeson, Inorg. Chem., 1982,21, 688. 507. N. J. Curtis, G. A. Lawrance and A. M. Sargeson, Aust. J. Chem, 1983, 36, 1495. 508. R. Breslow and D. Chipman, J. Am. Chem. SOC.,1965,87, 4195. 509. R. H. Barca and H. Freiser, L Am. Chem. SOC.,1966, 88, 3744. 510. T. Sakan and Y. Moni, Chem Le??., 1972, 793. 511. R. W. Hay and J. F. Ridlington, unpublished results. 512. R. P. Bell and M. I. Page, J. Chem. SOC.,Perkin Trans. 2, 1973, 1681. 513. B. G. Cox and R. E. J. Hutchinson, J. Chem. SOC.,Perkin Trans. 2, 1974, 613. 514. K. J. Pedenen, Acta Chem Scand., 1948,2, 385. 515. K. J. Pedersen, Acta Chem Scand., 1948, 2, 252. 516. B. G. Cox, J. Am. Ckem. Snc., 1974, 96, 6823. 517. R. Kluger and P. Wasserstein, J. Am. Chem. SOC.,1973, 95, 1071. 518. R. Kluger and A. Wayda, Can.J. Chem., 1975,53, 2354. 519. Y. Pocker and J. E. Meany, J. Am Chem Soc., 1967, 89, 631. 520. Y. Pocker and J. E. Meany, J. Phys. Chem, 1968,72,655. 521. S . Cabani, P. Gianni and E. Matteoli, J. Phys. Chem., 1972, 76,2959. 522. A. E. El-Hilaly and M. S. El-Ezaby, J. Inorg. Nucl. Chem., 1976, 38, 1533. 523. R. H. Prince and P. R Wooliey, J. Chem. SOC.,Dalton Trans., 1972, 1548. 524. P. R. Woolley, J. Chem Soc., Chem. Commun., 1975, 579. 525. For a recent review of the mechanism of action of carbonic anhydrase see R. D. Brown, I11 and S. H. Koenig, Inorg. Chim. Acta, 1984, 91, 183. 526. D.J. Creighton, J. Hajdu and D. S. Sigman, J. Am. Chem SOC.,1976,98, 4619. 527. R. A. Case and U . K. Pandit, J. Am. Chem. Soc., 1979, 101, 7059. 52R. R. A. Hood and R. H. Prince, J. Chem. SOC.,Chem. Commun, 1979, 163. 529. J. March, ‘Advanced Organic Chemistry’, 2nd edn., McGraw-HiI1, New York, 1977, p. 693. 530. J. P. Glusker, Enzymes, 1971, 5, 413. 531. L. R. Gahan, J. MacB. Harrowfield, A. J. Herlt, L. F. Lindoy, A. M. Sargeson and P. L. Whip, to be published. 532. P. 0. Whimp, to be published. 533. A. J. Herlt, to be published. 534. N. E. Dixon and A. M. Sargeson, J. Am. Chem. SOC,1982, 104,6716. 535. D. Banerjea and C . Chatterjee, J. Inorg. Nucl. Chem., 1967, 29, 2387. 536. A. K. Basak, D. Banerjea, R. W. Hay and C. Chatterjee, J. Coord. Chem., 1981, 11, 195. 537. A. K. Basak and C. Chatterjee, Bull. Chem. SOC.Jpn., 1983, 56, 318. 538. D. Banerjea, J. Indian Chem SW, 1977, 54, 44. 539. L. Goswami, S. Sarker and D. Banerjea, Z.Anorg. Allg. Chem, 1977,435 301. 540. D. Huchital and X. Yang, J. Conrd. Chem., 1981, 11, 57. 541. S. Das, R. N.Banerjee and I).Banerjee, J. Cnord. Cbcm., 1984. J3, t23. 542. D. Dolphin, R. Poulson and A. Avrarnovic (eds.), ‘Vitamin B, Pyridoxal Phosphate: Chemical, Biochemical and Medical Aspects’, Parts A and B, 1986, Wiley, New York.

475. 476. 477. 478. 479. 480. 481. 482. 483. 484. 485.

61.5 Decomposition of Water into its Elements DAVID J. COLE-HAMILTON" and DUNCAN W. BRUCE$ University of Liverpool, UK INTRODUCTION 61.5.1.1 General Considerations 61.5.1.2 The Thermodynamics of Hydrogen Production from Water

488

WATER CLEAVAGE USING SIMPLE METAL SALTS AS CATALYSTS 615 2 . 1 Hydrogen Production 61.5.2.2 Oxygen Production 61.5.2.3 Hydrogen and Oxygrn Production 61.5.2.3.1 The Fe"/Fe"' system 61.5.2.3.2 The Cer"/CeiV system

493

61.5.1

61.5.2

61.5.3

WATER CLEAVAGE USING METAL COMPLEXES AS CATALYSTS

INTERMOLECULAR ELECTRON TRANSFER REACTIONS IN THE PHOTOCHEMICAL DECOMPOSITION OF WATER 61.5.4.1 General Considerations 615 4 . 2 Photochemical Hydrogen Production Mediated by [Ru(bipy)J2' 6134.2.1 Photochemical properties of [ R ~ ( b i p y ) ~ ] and ~ ' its excited states 61.5.4.2.2 Photochemical hydrogen production via oxidative quenching of [Ru(bipy)J2+ * by methylviologen (1, I '-dimethyL4,4'-bipyridinium dication) 61.5.4.2.3 Photochemical hydrogen production via oxidatiue quenching of[R~u(bipy)~]~'* by chemically modified viologens 61.5.4.2.4 Hydrogen production via oxidative quenching of ruthenium(I1) complexes containing chemically modixed bipyridyl Iigands 61.5.4.2.5 Hydrogen production via oxidative quenching o f [ R u ( b ~ $ y ) ~ ] ~ by ' * non-viologens 61.5.4.2.6 Photochemical hydrogen production via reductiue quenching of [ R ~ ( b i p y ) ~ *] ~ " 61.5.4.3 Photoproduction of Hydrogen from Water Using Metailoporphyrin and Metallophthalocyanine Sensitizers 61.5.4.3.1 Photochemical properties 61.5.4.3.2 Metalloporphyrin sensitizers 61.5.4.3.3 Metallophihaloc~~anine sensitizers 61.5.4.4 Redox Catalysts 61.5.4.5 Oxygen Production by intermolecular Electron Transfer Processes 61.5.4.5.I Uncatalyzed reactions 6154.5.2 Homogeneous catalysts .for oxygen production 615 4 . 5 . 3 Heterogeneous catalysts for oxygen production 61.5.4.6 Cyclic Water Cleavage - Hydrogen and Oxygen Production uia Electron Transfer Catalysis 61.5.4.7 The Use of Molecular Assemblies in Water Splitting and Related Reactions 61.5.4.7.1 Introduction 61.5.4.7.2 Micelles 61.5.4.7.3 Reverse miceNes and microemulsions 61.5.4.7.4 Vesicles

488 489

493 493 494 494 49 5 495

61.5.4

61.5.5

PHOTOELECTROCHEMICAL DECOMPOSITION OF WATER

61.5.6 COORDINATION COMPLEXES AS CATALYSTS FOR THE ELECTROCHEMICAL PRODUCTION OF HYDROGEN OR OXYGEN FROM WATER 61S.6.1 Hydrogen Production 61.5.6.2 Oxygen Production 61.5.7

REFERENCES

498 498 499 499 500 502 506 506 508

510 5 10 511

513 513 515 515 517 519 523 525

523 526 527

528 530 532 532 534 534

* Now

at the University of St. Andrews. f Now at the University of Sheffield.

487

Uses in Synthesis and Catalysis

488

61.5.1 INTRODUCTION

61.5.1.1

General Considerations

Ever since Nehemiah demonstrated, albeit with divine assistance, that naphtha would burn,' man has increasingly relied on oil and oil based products for fuel as well as for feedstocks for the chemical industry. However, it was not until the invention of the internal combustion engine in the latter part of the 19th century that total exploitation of this resource became a reality. Oil can therefore best be seen as the raw material of the 20th century and its exploitation increased almost exponentially until 1973 when the price was suddenly raised and people began to consider the likely future availability of this commodity. Predictions based on consumption at that time suggested that available or exploitable resources were probably only sufficient for a maximum of 50 years. Ten years have passed since then and substantial ecohomies in the use of oil have been made but it is still clear that oil has a limited lifetime as our major source of energy. Indeed the most pessimistic current forecasts suggest that alternatives will have to make the major contribution in about 60 years. The major alternative is coal but this is also a non-renewable energy source with a predicted useful lifetime of about 200 years. Whatever the timescale, both of these resources are non-renewable and the realization that their lifetime may be comparatively short, together with the economics of exploiting, for example, shale oil and sincere worries about the use of nuclear power have led scientists to search for renewable sources of energy which might be exploitable on a commercial basis. Most renewable energy options must rely on a net input of energy into the earth and since the sun is our only external energy source, harnessing its energy is the main objective of almost all alternative energy strategies. Exceptions are tidal power, which relies on the gravitational attraction of the sun and particularly the moon, and geothermal energy, which cannot really be described as renewable. The sun can be considered to emit radiation as a black body at about 5900 "C hence the solar emission spectrum has a maximum at 500 nm and extends into the IR and the UV regions of the electromagnetic spectrum, although most of the energy is in the form of visible light. Absorption processes in the atmosphere, mainly by H 2 0 and C 0 2 ,remove a large proportion of the IR region of the sun's energy before it reaches the surface of the earth so that the main energy available at sea level is in the visible region (see Figure 1). 0.25

I

fi

Solar irradiotion curve outside otmosphere

lor irradiotion curve a t s e a level

Wavelength ( p m )

Figure 1. Spectral distribution curves related to the sun. Shaded areas indicate absorption at sea level due to thi atmospheric constituents shown (reproduced from ref. 2)

The main drawbacks to the use of solar energy are its rather low density (ca. 1 kW m-' on i clear day with the sun directly overhead), its wide wavelength spread and its somewhat unpredict able nature caused by different weather patterns. Nevertheless, calculations have shown that it i' in theory possible to provide the energy needs of the whole world provided that the energy cai

Decomposition of Water into its Elements

489

be stored in a usable form and that this storage process has an efficiency of -10%. Under these circumstances, it would be necessary to cover 1% of the land area of the earth with solar collectors or an area of about the size of Quebec province in Canada.2 Despite this apparently optimistic assessment, it is probably more realistic to think in terms of solar energy making a contribution, along with many other processes, to our major energy needs. In particular, isolated communities in areas where sunshine is prevalent may depend on solar energy in the near future (indeed satellites already do), whilst more general exploitations will increase once the prices of coal and oil start to rise again. There are very many options available for solar energy storage but this chapter is concerned solely with the use of solar energy to decompose water into hydrogen and oxygen, the former of which will be used as a fuel (see equation 1). H20 -+ H,+0.50,

AG0=238kJ mol-'

(1)

Solar hydrogen production from water has a number of extremely attractive properties. These include the fact that water is highly abundant and hence cheap and is regenerated on burning hydrogen so that no raw material is consumed. Hydrogen has a very high fuel value -the reverse of equation (1) produces 287 kJ of heat for every 2 g of hydrogen burnt. Hydrogen is easily stored, transported and used. Indeed feasibility studies suggest that the natural gas pipelines currently in use in the United States could be converted for transport of hydrogen gas with little or no modification.' Hydrogen could be burnt in stoves and a number of manufacturers have demonstrated the feasibility of hydrogen powered aircraft and motor cars? It is considered that underground storage in large caves is also practical for hydrogen.* When burnt, hydrogen is essentially non-polluting: the oxides of nitrogen that can be produced when hydrogen burns in air can be reduced to a minimum by efficient carburetion. Finally, hydrogen is indefinitely kinetically stable at room temperature so that long term storage is straightforward. The main disadvantage to storage of solar energy in the form of hydrogen is that consumers believe that hydrogen is a highiy dangerous commodity. Most people's only knowledge of hydrogen is associated with explosions, e.g. of the RlOl airship. The actual dangers are vastly overstated since massive explosions only occur if hydrogen is mixed with oxygen in a 2: 1 ratio. Even then, the very low density of hydrogen means that the explosion tends to go upwards and causes little damage on the ground. In contrast, fires and explosions arising from petroleum fumes are much more dangerous since they tend to cling to the ground and incinerate everything in the vicinity. Hydrogen derived from water can, thus, be considered as a viable fuel or method fur storing solar energy, provided that it can be produced cheaply. Some excellent books are available on the background to a hydrogen based fuel economy.2-6

61.5.1.2

The Thermodynamics of Hydrogen Production from Water

Hydrogen can easily be produced from water since all electropositive metals give hydrogen on treatment with water or dilute acid. Indeed, in 1785, during the course of experiments on large scale hydrogen production, Lavoisier used the observation that hydrogen was produced when water was contacted with hot iron as a crucial piece of evidence that water was a compound and not an element as had previously been supposedh7These and other experiments led to the development of many of the current theories of chemistry and to the overturning of the phlogiston theory. However, reactions between metals and water lead to metal oxides as the other product and release of oxygen from these is very difficult, particularly since the thermodynamics are unfavourable. For large scale commercial hydrogen production from water it is thus essential that the water is decomposed to produce hydrogen and oxygen (equation 1). The thermodynamics of this reaction are highly unfavourable, as is essential if hydrogen is to be a useful fuel, and it can be shown that temperatures in excess of 2800°C are required for the direct thermolysis. However, it is possible to design cyclic thermal systems involving simple chemical transformations that operate at much lower temperatures (700-1000°C) and research into the use of solar heat for these reactions is advancing. These reactions will not be considered further since, in general, they do not involve coordination compounds on account of the low stability of these compounds at high temperatures.

490

Uses in Synthesis and Catulysis

Alternative methods for surmounting thermodynamic barriers such as that of equation (1) are electrolysis and photolysis. Electrolysis is particularly attractive since hydrogen and oxygen are produced in different places and hence no separation process is required. However, electricity input is necessary and this must be generated. For solar energy conversion, direct photovoltaic conversion of sunlight into electricity using, for example, silicon cells is possible and subsequent electrolysis of water has been demonstrated. Alternatively, photoelectrochemical cells can be used and some of these are discussed later in this chapter, as are catalysts which lower the overpotential for hydrogen production at a mercury electrode. Finally, direct photolysis of water is a possible and highly attractive means of storing solar energy, particularly since the major portion of the solar energy arriving on earth is of a n energy comparable to electronic energy level differences within molecules. On account of the quantized nature of light, any photochemical reaction will have a threshold wavelength (Ag) such that light of longer wavelength will not drive the reaction.' If we assume that all light with wavelengths shorter than this threshold wavelength is absorbed and converted into chemical potential, it would appear at first sight that the amount of energy stored or the light energy to chemical energy conversion efficiency should increase as the threshold wavelength increases and, provided all the solar radiation were absorbed, 100% conversion efficiencies would be theoretically possible. However, this naive approach ignores the fact that quanta of light which have shorter wavelengths than the threshold wavelength contain more energy ( h c / h ) than is necessary to drive the required reaction ( h c l h , ) and hence for each quantum an energy of hc[(l/A)-(l/AJ] is not used to drive the chemical reactions and must be dissipated as heat. The amount of this loss increases as the threshold wavelength increases (Figure 2 ) and the consequence of this is that a plot of conversion efficiency against threshold wavelength rises steeply to a maximum at ca. 840 nm and then falls slowly. Allowing for other unavoidable energy losses, the graph is as in Figure 3. The ripples on the high energy side of the maximum arise from absorption in the atmosphere. It should be noted that the absolute maximum conversion efficiency for solar radiation is ca. 31% and since this is a thermodynamic limit, efficiencies above this limit cannot be achieved.* In practice, losses caused during chemistry subsequent to the light absorption step will almost certainly cause a further reduction in the conversion efficiency. It is probably unrealistic to expect efficiencies in chemical storage systems to be higher than 20%.

500

I

Wavelength (nrn)

Plot o f energy loss per mole of photons due to the quantized nature of light (ELoss)against wavelength for A, = 800 or 400 nm. Energy from all light of wavelength longer than the threshold wavelength is also lost

Figure 2

Turning to water photolysis specifically, it can easily be shown that, ignoring losses, the threshold wavelength for the breakdown of water into hydrogen and oxygen is 500, 1000 or 2000nm depending upon whether one, two or four photons are used to decompose each molecule of water. These threshold limits are shown in Figure 3 by full vertical lines. Calculations which aIlow for unavoidable losses shift these values for two or four photons to 778 and 1352nm (dotted lines in Figure 3 ) and further blue shifting will occur if the mechanism of decomposition involves, for example, energy wasting back reactions (discussed in detail below).8 These threshold values are thermodynamic values and assume that hydrogen and oxygen are the primary products. If free

*

If more than one reaction is used for photons of different energy, e.g. in stacked cells of different band gap semiconductors, higher efficiencies C Q ~be achieved but this is unlikely to be practical for direct chemical storage.

Decomposition of Wafer into its Elements

49 I

4c

-5 r c a,

I

840 nm

I

33%

I

3c

11

+ 'c

c P

2c

R

L

I

W L

I

V

I

I

I I

I

10

I

I

I

I I

/ 2 photons

I photon,

/ 400

photons

7 7 8 n y A I

2

I I

600

I

I .

803

352 nrn

IC

1200

\I

I

, 1400

.I, (nm)

Figure 3. Plot showing the variation of solar energy conversion efficiency with threshold wavelengths at sea level. The solid vertical lines represent the energies required for decomposition of water using one or two photons per molecule of water respectively (4 photons would be at 2000nm). The dashed vertical lines represent computed values for these threshold wavelengths taking into account unavoidable thermodynamic losses (adapted from ref. 8)

radical intermediates are involved, the threshold wavelengths are substantially shorter ((367 nm) and close examination of Figures 1 and 3 shows that systems in which free radicals are generated need not be considered for the practical storage of solar energy, where minimum efficiencies of ca. 12% will be required. Nevertheless, some systems of this kind are described in the subsequent sections of this chapter. As discussed above, and considering Figure 3, it is clear that only systems which involve the use of two or four photons to decompose one water molecule need to be considered as practical candidates for efficient solar energy storage. However, the complexity of designing a four-photon system means that most attention has been given to the two-photon system. One other important criterion for successful water cleavage that must be considered is the solution pH. Although the potential difference between the two half reactions for water decomposition is fixed at 1.23 V and is independent of pH, the half-cell reactions are dependent upon pH (Figure 4). Thus, by altering the pH of a solution it is sometimes possible to alter the half-cell potentials to be compatible with the redox properties of a photosensitizing catalyst. The oxidant must have a redox potentia1 above the oxygen line, whilst the reductant must have a redox potential below the hydrogen line. The effect of pH is ilhstrated in subsequent sections of this chapter. In the following sections of this chapter, we consider water decomposing systems based on simple inorganic ions or complexes, and then those which involve multistep intermolecular electron

PH Figure 4. Variation of electrode potentials (a)with pH for hydrogen (-)

and oxygen (- - -) production from water

49 2

Uses in Synthesis nnd Catalysis

.n 0

1 I I

c? 0

1 14 A l 0

0

Decomposition of Water into its Elements

493

transfers in attempts to model photosynthesis with synthetic reagents. We do not consider system% where the main components are naturally occurring nor do we consider unsensitized semiconductor systems.

61.5.2 WATER CLEAVAGE USING SIMPLE METAL SALTS AS CATALYSTS 61.5.2.1 Hydrogen Production It is well known that a large number of metal ions, when placed in water, are capable of oxidizing or reducing water and that others can carry out the oxidation or reduction photochemically. Simple ions capable of hydrogen production have been reviewed' and some of the results are listed in Table 1 . In some cases, the reactions occur thermally but are accelerated by light, whilst in others the incident light provides the necessary thermodynamic driving force. In general, the reactions proceed via metal to solvent (or metal to ligand, since the ligand and the solvent are the same) charge transfer, which leads to the production of H.or possibly solvated electrons, the latter of which subsequently react with H+ to give He. In many cases catalysts have been used to enhance the rate of the reaction, possibly by allowing it to proceed without the formation of free H.- either by forming adsorbed radicals (heterogeneous catalysts) or metal hydrides (homogeneous catalysts). In all photochemical cases, UV light is required for hydrogen production and in most cases the reactions occur in strongly acidic solutions. Sulfuric, phosphoric, perchloric and nitric acids are the preferred media, although halo acids have occasionally been employed. Usually, the metal containing products of these reactions are rather stable and do not produce oxygen thermally or photochemically, although isolated examples of hydrogen production on photolysis of higher oxidation state ions, e.g. E u " ' , ~U"' ~ 25 and suggest that cyclic behaviour is occurring and this may also be the case for Mn1'.26 Nevertheless, the requirement for UV light and the generally stoichiometric nature of the reactions renders them unsuitable for solar energy storage. Cyclic systems based on Fe"'"' and Cei'"'" are discussed in detail below.

61.5.2.2

Oxygen Production

Table 2 lists a number of simple metal ions that have been used in the thermal or photochemical oxidation of water. Most of these contain metals in high oxidation states and are strong oxidizing agents. Recycling of these ions by thermal or photochemical means is, then, highly unlikely. The exceptions to this generalization are Co"' and Mn'", the former of which, although strongly oxidizing in acid solutibn, is easily formed from Co" in basic solution or if coordinated with high-field ligands. As yet catalytic cycles based on Coil'll' have not been demonstrated but it is possible that some may be accessible. A cyclic water splitting reaction based on Mn"'"' has reported.26 Table 2

Production of Oxygen from Water and Simple Metal Ions

Reaction

Ion

TI'

+

Mn'"

co'+ NilV cu3+ NpV" pUV"

Amlv [MnO,I-

WPe hU

A A A h A h

h h

hv [FeO,]'-

A

Producr

Conditiom

H2S0,(6-15 M) H2S0,(10-'-3 M), HCIO,, HNO, pH 1.0 pH 1.0 HCIO,(0.1-4 M) pH>7 H3 PO, HzS0,(7.5 M ) pH 7-14 pH 0.5-8.8 HCIO, or phosphate

TI ' Mn"'

co2+ NiZ+ cu2+ Npv'

Puvl Am"' Mn'", Mn"' MnO, b*e3+

Refs. 20 27 28-3 1 31 31 32 33.34 35 36 37 38

In general, the mechanism of 0, production in these systems is believed to involve formation of OH. or oxygen formation by coupling of hydroxo bridges as shown in Scheme 1 for cobalt.'

Uses in Synthesis and Catalysis

494 H 0

2C01"+20H-

-*

[(H20)Jo

/ \ Co(H20),I4" \ / 0 H

-

H 0 ,/ j , ,, H O [(H,O),Co, : Co(H20),I4+H202+2[Co(H,0)6]2+ j ,,' 0

.\

H

Scheme 1 Proposed mechanism for the thermal oxidation of water by Co3+

61.5.2.3

Hydrogen and Oxygen Production

61.5.2.3.1 me Fe1'/Fe'''

system

Redox potentials for various different I;e"/Fe"' couples are shown in Table 3 and although [Fe(OH),] is thermodynamically capable of reducing water to hydrogen and does so at elevated temperatures? the rate of the reaction is low. It can, however, be increased by the addition of suitable catalysts, e.g. Ni excess [FeS04j,40metals such as Ni?' CUTPd,18341 or colloidal Pt.40 Nevertheless, cyclic water cleavage based on this system is unlikely since [Fe(OH),] is a. highly stable material. Table 3 Half-wave Potentials (TO) for Various ~e'+''+ Couples coupr;

[Fe( H20)6]3+'2+ Fe(OH),/Fe(OH), [Fe(CN),I3-l4[Fe(bipy),]'+/*+

7P

+0.77 -0.88 +0.36 +0.96

For catalytic water decomposition, it is therefore necessary to work in acidic solution where, although Fe" is not thermodynamically capable of reducing water and Fe"' is not thermodynamically capable of oxidizing water, both reactions can be driven photochemically. Because of this, two photons are used for the reaction and a large portion of the solar spectrum can in theory be collected. In practice this is not the case as both Fell and Fe"' absorb mainly in the UV region of the spectrum. Irradiation of aqueous solutions of Fe" salts with UV light gives hydrogen and Fe'" with initial quantum yields in the range 0.1-0.7,* depending upon solution pH, although these decrease with time to 0.01 over prolonged periods.42 It appears that it is essential for the iron to be present at least partly as [Fe(S04)n]2n-2complexes since higher quantum yields are obtained at very low The quantum yield for H2 production is generally rather pH where these species insensitive to [Fe"] but at low pH (0.5) some positive dependence is Studies carried out using deuterium labelling," trapping reagents4 or in glass matrices at low temperatures' confirm that hydrogen radicals are the primary product and are produced with quantum yields of ca. O . l . I 7 There is some suggestion that these react with Fe" to form H- from which hydrogen is obtained by protonation." The rate of hydrogen production is not increased by the addition of other simple ions but it can be sensitized to visible light by addition of [Rh2(bridge),12+(bridge = 1,3-diisocyanopropane). l n this case, the hydrogen is produced directly from the.;hodium species with formation of [ FCh2(bridge),C12]2' and Fer' simply regenerates this starting Rh catalyst.45 Oxygen production can also be achieved by irradiation of iron(II1) salts with 254 nm light in aqueous solution. The greatest successes appear to occur using the perchlorate although the chloride has been empl0yed.4~It has been shown by detailed kinetic studies that the mechanism of this reaction involves primary formation of hydroxyl radicals, two of which give an oxygen atom. Dimerization of these atoms produces oxygen?6 The success of the reaction probably arises because back electron transfer between OH. and Fez+ is slow ( k < 3 x lo7 dm3 mol-' 46 The O2 producing reaction can be catalyzed by W 0 3 , ZnO or TiOz but not by SiO, or A& RuO: adsorbed on W 0 3 seems to be an even more efficient catalyst4' and it has been suggested that 0: is not formed on photolysis of ammonium iron(II1) sulfate unless these catalysts are present

ji

* All quantum yields are quoted in mol einstein-'.

Decomposition of Water into its Elements

49 5

Similar results were obtained with [Fe(CN),I3-. The initial electron transfer event in these cases occurs between catalyst adsorbed Fe3+(ads) and 0 H - ( a d ~ ) . ~ * Combining these two half reactions leads to catalytic production of hydrogen and o x ~ g e n , ~ ~ , ~ ~ although quantum yields are very low ( E o > -0.50V. Working on viologens of related structures, Sasse and confirmed these obscrvations and went on to explain them in terms of k, and the rate of hydrogen production via the reaction of the viologen radical cation with protons over the Pt catalyst. Thus, whilst k, decreases with increasingly negative redox potentials of the viologen, the rate of hydrogcn production increases, and it is the balance of these two rates which leads to the observed upper and lower limits of E'(R/R~).

503

Decomposition of Water into i t s Elements

0

9 0-

2

I

d

-1

z

x

2

2

I

I

wl

2

0

L

0

b,

b,

tn

Ln

m

wl

0 N *

,..

8

2

W

c

0

8

0

N

3

0

0

t I

d

2I

N

c: 0 I

Uses in Synthesis and Catalysis

504

r

a,E t

:

x

I I I I

8

c

r

0 I

f

Li

x

4

I

t:

I

I

m m o m

Y\q?r:

0 0 0 0

I

I

I

I

m

- 0

?T

30

I

I

I

r0

505

Decomposition of Water into its Elements

Using the zwitterionic viologen (17) Willner and Degani'413'45demonstrated the use of colloidal suspensions as an aid to charge separation. When a solution containing [Ru(bipy),]'+, TEOA and (17; DQS) was irradiated at pH 9.6, no DQS' formation was observed, but on the introduction of SiOz colloid into the system, DQS- was produced with a quantum efficiency of 2.4%. Before excitation, the positively charged sensitizer is held in close proximity to the surface of the colloidal particle, which is negatively charged due to the ionization of the surface silanol groups at pH > 6. On electron transfer, DQST possesses a net negative charge and is therefore repelled by the colloidal surface. Not only does this facilitate charge separation, but it also results in a decrease in kb to a value of about lo7mol-' dm3 s-l. Addition of colloidal platinum to solutions containing [Ru(bipy),12+, TEOA, DQS and Si02 yielded hydrogen with a quantum efficiency of 2.2 x lop3 at pH 8.5. (ii) Polymeric systems A logical extension of the work described above on long chain viologens, is to attach the viologen to a polymeric support and some examples of this are now described. Work by Lee et used N-(benzylpolyvinyl)-N'-methyl-4,4'-bipyridinium(PVBV*+) with 41YO of the benzyl groups having viologen functions, and polyvinyl-N-acetate-Nt-methyl-4,4'bipyridinium ( PVAV2+)with 40% of the acetate groups having viologen functions. PVBV2' and PVAV" quenched [Ru(bipy)J2+* with rate constants kq = 2.7 x 10' and 3.1 x 10' mol-' dm3 s-' respectively, with kb = 1.8 X 10" mol-' dm3 s-' in both cases. In the presence of EDTA, the yield of PVBV' on photolysis was 6.3% and in the presence of colloidal platinum (protected by the in situ polymer), the hydrogen yield was only 5% of that observed for an equivalent system containing MV2+in place of PVBVz+ and with the Pt supported by carbowax 20M. Matsuo and ~oworkersl~'examined poly( N-vinyIbenzyl-N'-n-propyl-4,4'-bipyridinium) (PV1002+)in conjunction with [Ru(bipy),]'+ and the results are collected in Table 9. The quenching of [Ru(bipy),12+* by PVIOOz+ proceeded at a rate which was less than 20% of that for MVZC, whilst the rate of hydrogen production in a sacrificial system was found to be 35% of that obtained for MV". However, when k , and the overall rate of hydrogen production are compared, it is found that PV1002+ produces hydrogen twice as fast as MV" on reaction with H+ over a Pt catalyst (equation 26). The same study also showed that the electron transferred to the viologen polymer is delocalized over the polymer chain, revealing the possibility of trapping the electron in an appropriate sink built into the polymer backbone. Table 9

Comparison of Monomeric and Polymeric Viologens

Relative rate Ru complex [Ru(bip~)~l'+ [Ru(biPY)'l*+ a

Viologen

Environment

MV" PV1Oo2+

Aqueous solution Aqueous solution

kJmol-' dm' s-') 1.7 x lo9" 8.7 x 10'

of reduction

loob 19

Under the conditions employed. Arbitrary units-reference.

In a slightly different system, Matsuo '48~'49 also examined electron transfer to a polymer PV"] via a zwitterionic supported viologen [ poly( N-vinylbenzyl-N'-n-butyl-4,4'-bipyridinium; PSV). viologen (N,N'-bis(3-sulfonatopropyl)-4,4'-bipyridini~m; PSV quenched 31% of the luminescence from [Ru(bipy),]'+* and gave a charge separation yield of 57% with respect to a control system using MV' as quencher. In the presence of PV2+, ' however, whilst the quenching efficiency did not change (PV' does not quench [R~(bipy)~]'+* to an appreciable extent), the charge separation yield increased to 151% of that obtained in a system containing MVZ+.Similar effects were observed when [Ru(phen),]'+* was employed as sensitizer (phen = 1,lO-phenanthroline). Furthermore, studies on the lifetime of the reduced form of PSV implied that the electron was further transferred to the polymer which was then acting as an electron sink. In the presence of colloidal Pt, the decay of the reduced viologen was enhanced, presumably because of electron transfer to t'F and this decay became extremely rapid when the solution was acidified with 0.5 mol dm-' H,S04. However, when EDTA was added to the system, hydrogen production was not observed, as EDTA is not an efficient electron donor (to [R~(bipy)~]")at this pH. Nevertheless, hydrogen was evolved on irradiation (in the presence of EDTA) at neutral pH, although much more slowly than in a system containing MV2+,and in fact PV+ accumulated. This points to proton reduction as the rate determining step and the effect is ascribed to a shift

506

Uses in Synthesis and Catalysis

in the redox potential of Ht/H2 on Pt, which in this system is supported by PV2+; similar shifts have been reported for Pt electrodes covered with ammonium salts.'50

61.5.4.2.4 Hydrogen production via oxidative quenching of ruthenium(II) complexes containing

chemically modified bipyridyl ligands

Much of the work in this area has centred around efforts to optimize the photochemical and redox properties of the Ru" complexes which are related to water cleavage reactions, e.g. lifetime of excited state, absorption maxima, etc. A detailed account of these properties is found in Chapter 8.3 and hence it is only intended here to present the results of these studies on hydrogen producing systems. Using [ R ~ (b ip y )~ (DECb ip y )]~ (DECbipy + = diethyl-2,2'-bipyridine-4,4'-dicarboxylate),Sasse and coworkers'51 found that hydrogen production yields in a system comprising [ R ~ ( b i p y )DEC~( bipy)]2+/MV2+/EDTA/pt were only 10% of those obtained in an identical system employig [Ru(bipy)J2+. This decrease was attributed to two effects. Firstly, Eo(Ru3+"+") for the DECbipy substituted complex was some 100mV less negative than that of the unsubstituted complex, rendering the former a weaker reducing agent. This is reflected in the values of k,(MV2') obtained, which were 9.6 x 10' and 1.3 x lo8mol-' dm3 s-' for the unsubstituted and substituted complex respectively. This lower value of k,, coupled with the observation that water would quench [Ru(bipy)3(DECbipy)]2+*,albeit with k,< IO5, is thought to be responsible for the low yields of hydrogen. compounds of the formula [R~(bipy),,(thpy)~-,]~+(thpy = In a related 2-(thiazol-2'-yl-pyridine)were examined as chromophores. It was found that the rate of hydrogen production steadily decreased with increasing n, but the stability of MV2' in the system was markedly enhanced. Usually, MV*' in these systems is gradually consumed via catalytic hydrogenation over the pt catalyst (see Section 61.5.4.2.2), although sulfur containing compounds have ] ~ +chromophore, the been shown to inhibit this r e a c t i ~ n . ' ~ Thus, ~ , ' ~ ~using [ R ~ ( b i p y ) ~ ( t h p y ) as but + , the percentage of rate of hydrogen production is 67% of that obtained for [ R ~ ( b i p y ) ~ ] ~ unhydrogenated MV2+ increases from 53 to 93.15*Inhibition of the hydrogenation reaction is due ] ~ + the Pt catalyst, presumably via the sulfur atoms to adsorption of some [ R ~ ( b i p y ) ~ ( t h p y )onto of the thpy ligand. Various phenanthroline complexes of Ru" have been examined as potential chromophores including 1,lO-phenanthroline (phen), 5-chlorophen (5-Clphen) and 4,7-dimethylphen ( d r n ~ h e n ) . All ' ~ ~of these complexes have much longer lived excited states than [Ru(bipy)J*+, notably [Ru(dmphen)J2+ where the excited state is three times longer lived. In a system using hydrogenase as catalyst and a thiol as electron donor, the rate of production of MV' increased according to bipy < phen < 5-Clphen < dmphen. The rate of hydrogen production, however, followed a different series (phen < bipy = 5-Clphen < dmphen) and the lack of correlation between d[MV']/dt and d[H,]/dt is attrib~ted''~to the known inhibition of hydrogenase activity by phen complexes of ruthenium.'54 Highly efficient quenching of ( N ,N ' - didodecyl-2,2'-bipyridine-4,4'-dicarboxamide)bis(bipyridy1)Ru" ([RuC,,B]) has been reported using long chain or polymer bound vio10gens.'~~ These systems are approximately twice as efficient as the [Ru(bipy),12+/ MV2+ combination, although hydrogen production has not been reported. pendant on polystyrene, photoreduction of MV2+ was achievedlS5 at Using [ Ru(bipy)$' ca. 25% of the rate attained in a [Ru(bipy)312+/MV2+system. Addition o f HCI and Pt black to the former system led to hydrogen evolution, although efficiencies were not given. Work by Lever and later Tazuke on bipyrazinyl and bipyrimidyl complexes of Ru" is described in Section 61.5.4.2.6.

61.5.4.2.5 Hydrogen production via oxidative quenching of [Ru(bipy)#+

* by non-viologens

Whilst methylviologen and its derivatives have attracted much attention as electron relays in multicomponent hydrogen producing systems (see Sections 61.5.4.2.2and 61.5.4.2.3), other compounds have been studied and it is these that are described here. Lehn and coworkers 156~157investigated the rhodium complex [Rh(bipy)J3+, which on photolysis in solutions containing [ Ru(bipy),]'+, TEOA and K2ptC14 produces hydrogen after a short induction period during which [ptCl,]*- is reduced to R0 (see below). They found that an

Decomposition qf Water into its Elements

507

oxidative cycle (in Ru) operates with [Rh(bipy)J’+ quenching [R ~ ( b i p y ) with ~ ] ~a~rate constant kq = 3.5 x 10’ mol-’ dm3 s-’ (Eo(Ru3+l2+*)= -0.84 V; E0(Rh3+/2+) = -0.67 V). Reactions (32) to (38) were proposed for hydrogen production at pH 7 in the presence of excess bipy and involve the further reduction of [Rh(bipy)Jz+ to [ R h ( b i ~ y ) ~ and ] + free bipy (reaction 36): TEOA- (produced after proton loss from TEOA’ formed in reaction 34) was proposed as the source of the electron for the reduction to Rh’.This Rh’species is,%henprotonated to a Rh”’ hydride species (reaction 37), which reacts with H+ to form hydrogen (reaction 38). Recoordination of a bipy ligand regenerates [ R h ( b i ~ y ) ~ ]and ~ + completes the cycle. Reaction (38) can occur in the absence of platinum, although the rate of hydrogen production increases 20 fold in its presence. [R~(bipy)~]’+ % fRu(bipy)J2+* [Ru(bipy),J3’ f[lU~(bipy)~]”

[Ru(bipy),]’+*+[Rh(bipy),]”‘ [ R ~ ( b i p y ) ~ ] ’ +TEOA +

TEOA?

4

[Rh(bipy),]’++e-

[Ru(bipy)J2* +TEOA’

decomposes

-+

+

[Rh(bipy)2]++bipy

!% [RhH(H20)(bipy)z]2’ [W~(bipy)~]++H+ [RhH(H,O)(bipy),]’++ Hf

[Rh(bipy),13++ H 2 0 + H2

At pH < 6 , in the absence of excess bipy and with EDTA as eIectron donor, hydrogen production takes place by a different mechanism (reactions 32 to 37 and 39 to 42: for TEOA read EDTA). Presumably, the [Ru(bipy)J3+ generated in reaction (40) is reduced by EDTA. However, the source of the electron in reaction (41) is unclear, although presumably it could arise from EDTA. formed by deprotonation of EDTA’ in an analogous reaction to that described above for TEOA. Further studies on the same system by Sutin and ~oworkers’~’ implied some variation on the mechanisms proposed by Lehn.ls7 Sutin found that whilst the quantum yield of [Ru(bipyM3+ formation with [F2h(bipy)J3+ as quencher was 0.15, the quantum yield of [Rh(bipy)J*+ was 0.3. This was interpreted in terms of reactions (43) to (46), where the TEOA, radical reduces the Rh”’species rather than the Rh” species. Furthermore, the fact that the quantum yield of the final Rh’ photoproduct is 0.13 suggests that it is formed via disproportionation of [ R h ( b i ~ y ) ~ ] ~ + and not by reduction of [Rh(bipy)J2+ by TEOA.

-

[RhH(HzO)(bipy)2]2++H+& -!

[Rh(bipy),(H20)2]3+-kHZ

(39)

[ W b i p ~ ) ~ ( H ~ O ) ~[lR’ +u + ( b i ~ y ) ~ l ~ * * [R~(~~PY),(~,O)~I~++[R~(~~PY)~~~+ (40) [ R h ( b i p ~ ) ~ ( H ~ O ) , ] ~ +4 + &[Rh(bipy),]++ 2H,O [Rh(bipy),]++ H+

[RhH(H10)(bipy)2]2’

[R~(bipy)~]~+*+[Rh(bipy)~]~+ + [Ru(bipy),13++[Rh(bipy),]*+ [Ru(bipy),I3++TEOA TEOA’

4

+

[Rh(bipy),13++TEOA.

[Ru(bipy)J2++TEOAt

TEOA.+H+

+

[Rh(bipy),l’++products

(41) (42)

(43) (44)

(45) (46)

With regard to the hydrogen producing reactions, it was proposed’58 that [ Rh(bipy),12+ was the active species and this was supported by the following observations. Solutions of the bis(bipyridine)rhodium( I) species produced no detectable hydrogen when left in contact with Pt for extended periods. Secondly, under catalytic conditions with high [Pt], hydrogen is evohed and no Rh’ is observed, whilst at low [pt], the yield of hydrogen drops and Rh’ is formed ai its expense. ~ period in That the reduction of [PtC14]2pto Pto is the cause of the o b ~ e r v e d ”induction hydrogen production was shown by the following experiments. Replacement of K2PtC14 by colloidal Pto in a catalytic system resulted in hydrogen evolution with no induction period.1s8 Secondly, the induction time for a catalytic solution containing K2PtC14(3 x lop4mol dm-3) was equivalent to the time required to form 3 x 10-4moldm-3 of Rh’ (in the absence of K2PtC14). Coupled with the observation that free bipy accumulates at early photolysis times, this latter observation points to [Rh(bipy),]+ as the reductant.Is8 Other group^'^^-'^' have investigated various [Co(cage)]” complexes as electron relays with varying degrees of success. Many of these species seem to quench by energy transfer or do not charge separate in the encounter complex160but a notable exception is [Co(sepulchrate) J3+ ( [ C ~ ( s e p ) ] ~ + ) . ~This ~ ~ - complex ’~~ oxidatively quenches [ Ru(bipy)J*+* with a rate k = 3-6 x 10’ mol-’ dm3 s-l, but the quantum yield for cage escape is 0.9 (cf; = 0.25 for MV4+) CCCS-Q*

508

Uses in Synthesis and Catalysis

and this is a t t r i b ~ t e d ’to ~ ~the extra charge on the complex with respect to MVZ+.However, [ C ~ ( s e p ) ] ~is’ not as good as MV2+in hydrogen producing systems as it is orange in colour and acts as an inner filter. Furthermore, the thermodynamic driving force for proton reduction is only 25 mV compared to 160 mV for MV2’. Photolysis of buffered aqueous solutions of [Ru(bipy),I2+ and [Fe,S,(SCH2Ph),l2- (t2-; a synthetic analogue of the iron-sulfur proteins found in photosystem I) generates hydrogen in the presence and absence of EDTA.162This latter result is particularly interesting as it implies that water itself is acting as both electron donor and acceptor, although the authors failed to observe simultaneous oxygen evolution, even on addition of unhydrated R u 0 2 . Furthermore, that the proposed mechanism (reactions 47 to 51) requires four photons to produce one mole of hydrogen necessarily limits the potential efficiency of the system. 2[Ru(bipy),JZi [ Ru(bipy),]’+*

a 2[Ru(bipy),12‘*

+ tZ- + [ Ru(bipy),]’+

[Ru(bipy),I3++ OH[Ru(bipy)J’+*+t3[Ru(bipy),]++ H+

(47)

-t t3-

(48)

+

(49)

--t

[Ru(bipy)J2+ 0.2502+0.5H20

-+

[Ru(bipy),]++t’-

-+

[Ru(bipy),12++0.5H2

Ti3+ was reported to produce hydrogen uia oxidative quenching of [R~(bipy)~]’+* with a quantum efficiency of cu. 0.1, with the Ru”’ so produced oxidizing a second mole of Ti3+to Ti’” (as Ti02+) in preference to oxidizing water.’63However, the observation163that the solutions of [ R ~ ( b i p y ) ~used ] ~ + in the experiments produced hydrogen on photolysis (ie. in the absence of Ti3+) prompted other workers to reappraise the initial claims. They that the quantum efficiency (in the presence of Ti3+) was and whilst agreeing with the previously measured value of kq = 6 x lo6 mol-’ dm3 s-l, interpreted the quenching mechanism as one of energy transfer. This was supported by the observation that the rate of reaction (52) was lo5dm3 mal ' cm-') which is blue shifted on complexation with third row transition elements. They can absorb a large fraction of the solar spectrum, but their potential application to hydrogen production is limited by their tendency to aggregate in solution"R (although where the system allows, this aggregation can be prevented by the use of pyridine ( 5 % w/w) or micelles). Almost all photoreactions of porphyrins and phthalocyanines occur via the triplet state whose lifetime (ca. 1 ms at T = 300 K) and yields of formation are greatest for porphyrins. The redox potentials of porphyrins are also more amenable to water photoreduction via oxidative or reductive cycles and some selected values are found in Table 11. Table 11 Redox Potentials" for Some Ground and Excited State Porphyrins and PhthaIocyanines'

*'

a

Compound

TO(P+/P)

HzTPP ZnTPP CdTPP PdTPP PC

+1.19 +0.78 +OX7 f 1.26 +1.34

ZnPc

+0.92

PO!

P/ P-)

-0.81 -1.11 -1.01 -0.76 -0.42 -0.65

T O ( P+/P;-*)

TO(

Ps.*/P-)

-0.24 -0.70 -0.67 -0.53

+OH

f0.10

f0.82

-0.21

+0.48

+0.37 +0.53 $.1.03

Measured in eV cs. NHE. :P denotes the triplet excited state.

The coupling of these photochemical and redox properties to water reduction reactions is now described.

61.5.4.3.2 Metalloporphyrin sensitizers

Hydrogen producing systems employing metalloporphyrin sensitizers have usually consisted of the same components as systems employing [Ru(bipy),]'+ as sensitizer, that is MV", EDTA and colloidal Pt or hydrogenase. Whilst some reports have appeared133'1'6-18'of hydrogen evolution employing water insoluble zinc meso-tetraphenylporphyrin (ZnTPP; Figure 6; R = Ph) in micelles or mixed organic/water systems, the bulk of the literature has centred on the water soluble analogues, tetra( N-methylpyridinium)porphyrinatozinc(II)( [ZnTMPyPI4+; Figure 6; R = py-Me+) and zinc(II)tetra(4-sulfonatophenyl)porphy~~a~ozin~(I~)([ZnTSPP]~~; Figure 6; R = p-C6H4S03-). Hydrogen production from these latter compounds was demonstrated independently by Gratzel'89 and McLendon and Miller'9o in 1980. On photolysis, [ZnTMPyPI4+ is excited and rapidly intersystem crosses (with an efficiency mol dma3 of ca. 0.9) to the triplet state which has a lifetime of 1.3 ms. At [ZnTMPyPI4+= and [MV2+]= 5 x lop3mol dm-3 this triplet state is oxidatively quenched with k, = 2 x 10' mol-' dm3 s-l. However, at low [MV2+]and in the presence of EDTA (2 x mol dm-3) the triplet state can be reductively quenched with kq= 4 x 10' mol-' dm3 s-'. Thus photolysis of solutions containing [ZnTMPyPI4+, MV2+, EDTA and Pt leads to hydrogen production and, in " the light of the two possible quenching mechanisms, equations (75) to (82) were p r ~ p o s e d . ~At low [MVZ+],a substantial amount of hydrogen is produced on photolysis, but reaction (77) occurs faster than reaction (76). This implies that reaction (81) occurs very much faster than reaction (80). In fact, reaction (76) only predominates at [MV2+J> IOp3 mol dm-3. In the absence of EDTA and at [Pt] = 8 x mol drnp3,photolysis yields no hydrogen and leads to extensive destruction of the sensitizer, presumably by reduction to the dihydro- or tetrahydro-porphyrin. When [Pt]is raised to about 6 x lo-' mol dmP3 however, hydrogen production is observed, presumably uia reaction (83).lS9 Thus, it is possible for hydrogen production to occur by parallel oxidative and reductive cycles. 189~190Interestingly, McLendon and Miller'" interpreted their findings in terms

Uses in Synthesis and Catalysis

512

MPc (R = H) MPcTS4- (R = SO,-)

0

Metallophtbalocyanines R

R\

/ \

MTPP (R = Ph)

R

-/ R Metalloporphyrins Figure 6. Structures and nomenclature! for metalloporphyrins and metallophthalocyanines

of a singlet state mechanism, but recent reports''8 indicate that as a result of spin selection rules, singlet quenching does not normally lead to redox products. Furthermore the short lifetime of the single excited state (ca. s ) would seem to preclude collisional encounter. ZnTMPyp+

5 ZnTMF'yp+*

(75)

ZnTMPyp+*+MV2+ + ZnTMI'yP'++ MV? ZnTMPyp+* + EDTA ZnTMPyp+++DTA EDTA' 2ZnTMPyP"

+

MV'+H+

ZnTMF'@++ EDTA?

-D

ZnTMF'yP"++EDTAt

+

H++EDTA*

(79)

ZnTMPyF'++ ZnTMPyp+

ZnTMPyP++MV'+

ZnTMPyp++ H+

--*

-D

ZnTMPyp++ MV'

% MV2++0.5H, ZnTMF'yp++0.5H2

(83)

Replacing [ZnTMPyPI4+ by [ZnTSPPI4- also led to hydrogen e v o l ~ t i o n , ' ~although ~ " ~ ~ the yields were substantially lower; in fact, the turnover number in the porphyrin dropped from 350 h-' [ZnTMPyPI4+to 50 h-' [ZnTSPPI4-. Interestingly, these compare well with [R~(bipy)~]" systems where the turnover under similar conditions is only 4 h-'.I9O The triplet excited state of [ZnTSPPI4- is oxidatively quenched by MV2' with kq= 1.4 x 10" mol-' dm3 s-l, and this high value is almost certainly due to static quenching arising from ground state cornplexati~n.'~~ Despite the high rate of quenching, hydrogen production remains low and this is due to a low quantum yield of MV' caused by electrostatic attraction between the redox products. ' defined the optimum conditions for These findings were confirmed by Harriman et ~ 1 . ' ~ who the reaction. In a solution buffered to pH 5 containing [ZnTMPyPI4+ (2 x lo-' mol dmP3), MV" (8 x lop3mol dm-3) EDTA, (1.5 x (MV+) was 0.75 mol dm-3) and Pt (ca. lo-' mol dm-3), and the quantum yield for hydrogen production was 0.3. This in fact represents 60% of the maximum possibIe efficiency for such a system. Irradiation of the same solution but omitting MV2' gave a quantum yield for hydrogen production of 2 x lo-' for the reductive cycle. However, whilst prolonged photolysis led to only minor bleaching of the chromophore, this bleaching was greater for the oxidative cycle. Thus exhaustive photolysis of the two systems gave approximately the same amount of hydr~gen.'~' More recent work by R i c h o u ~ showed '~~ that these efficiencies could be further increased, as the quantum yield for product ion formation on quenching of [ZnTMQPI4+*by MV2' increased

Decomposition of Water into

its

Elements

513

with temperature, whilst the rate of back electron transfer decreased. Interestingly, replacement of MV2+ by C,,MV2+, gives cage escape yields of nearly 100%.'38 Employing [ZnTSPPI4- as sensitizer and ascorbic acid as electron donor, Tabushilg4 also demonstrated photo eneration of hydrogen. Thus, irradiation of 8 cm3 of a solution containing [ZnTSPPI4- (7 x 10- mol dm-3),ascorbicacid (0.1 mol dm-3), MV" (lo-' mol dm-3) and Pt (1.1 x mol dm-3) at 0 "C yielded 100 ~1 (H2) h-', with more than 99% of the chromophore and relay being unchanged after 9 h. On extended photolysis, the rate of hydrogen production decreased, and this was attributed to electron transfer from MV2+ to the dehydroascorbic acid produced. If MV2' was left out of the reaction solution, then irradiation produced only 130 p,1 (H,) in 3 h, by which time the chromophore was completely degraded. This was attributed to a demetallation of the porphyrin, which was confirmed by the observation that irradiation of a solution containing [HzTSPPj4-, ascorbic acid and EV yielded the same amount of hydrogen. Replacement of MV2+ by hexylmethylviologen generated 7150 ~1 (H2) in 10 h.'95 [ZnTSPPI4- was shown to be an efficient photosensitizer in a system which did not require an electron relay.'96 Thus irradiation of 5 cm3 of a solution containing [ZnTSPPI4- (4x mol dmW3),TEOA (0.2 mol dm-3) and platinized zinc oxide (Pt/ZnO 0.1-10 wt% Pt) gave hydrogen at a rate of 39.2 p.1 h-' and a quantum yield of 0.03 (0.5 HJ. This system produced hydrogen via an oxidative cycle and was shown to be stable with no denaturation of the Pt/ZnO and negligible denaturation of the chromophore in 152 h.'% Whilst most of the above examples have been based on Zn chromophores, other metals have been investigated"' and thus Sn1V,'973'98 pdII,199 RuII 200,201 and Cd"202 porphyrins have been shown to be capable of acting as effective chromophores and of these, the last has been found to be potentially superior to Zn?"

F

61.5.4.3.3

Metallopkthalocyanine sensitizers

Despite their superior photoproperties, metallophthalocyanines"8~203 have not found use as photosensitizers for hydrogen production from water due to poor cage escape yields204and extensive ground state comprexation by MV2+,205indeed MV2+ enhances the formation of phthalocyanine dimers. Thus, irradiation of solutions containing [ZnPcTSI4- (see Figure 6) EDTA and Pt produces hydrogen with quantum yields of

61.5.4.4

Redox Catalysts

Crucial to the success of reactions designed to produce hydrogen via intermolecular electron transfer reactions is the addition of an efficient redox catalyst which allows reduction of protons by the reduced form of the relay (e.g. MV+) formed by the initial photochemical electron transfer. although other metals, e.g. Rh and Pd,206 Early studies relied on colloidal platinum or RO, have been employed. Ru, Au, Ir and Ag206are reported to be ineffective, although some authors have claimed HZproduction using a colloidal gold redox catalyst.207Ru02 is active 1363208 for hydrogen production from MV2+ and has the added advantage that it does not catalyze the hydrogenation of MV2+ (see Section 61.5.4.2.2). In general, the cataiysts have been prepared by methods developed in the 1 9 4 0 ~ , *which ~~ involve reduction of H2PtC16or K2PtC14 with, for example, citrate,'" basezo9or NaBH4 138211 and can then be used directly, or, more usually, protected with a polymer, typically poly(viny1 a l c o h 0 1 ) ~carbowax ~~ 20M13' or poly(vinylpyrrolidone),212 a colloidal semiconductor such as T ~ ,213-217 O ~ SrTi03,218,186 ZnOig6 etc., in zeolites or in cyclodextrins.21gWhen unprotected, the catalysts tend to agglomerate giving them rather short active lifetimes.220,221 An alternative synthesis involves radiolysis of solutions containing appropriate metal salts, e.g. H2PtC16, and a water soluble acrylic monomer, e.g. acrylamide.222Radiolysis causes reduction of the salt with concomitant formation of polymer in which the colloidal metal particles (average dimeter ca. 17 A) are trapped. Supported catalysts prepared in this way are indefinitely stable towards agglomeration. is The efficiency of the catalyst in the archetypal system, EDTA/[RU(~~~~),]~+/MV~+/P~ dependent upon its mode of preparation, its concentration, possibly the particle size, and the Thus, unprotected colloids give lower rates of hydrogen production nature of the support.2193223-226 than those protected with PVA and most catalysts reach their maximum efficiency at [&I= 1-5 105 mol dm-3,129.224,226.227 although for particles with diameters of ca. 30 A (probably the most active particles), maximum rates are achieved at 2 x mol dm-3.'38 Under these conditions

514

Uses in Synthesis and Catalysis

it is probable that all of the MV' escapes back electron transfer and is intercepted by the pt particles. In systems where semiconductors are used to support the colloidal platinum, the electron transfer catalyst is often omitted and the primary electron transfer event involves injection of an electron from the excited state of the chromophore into the conduction band of the semiconductor^ 196216,217,228-230 Hydrogen production then occurs at the supported platinum particles. These systems are discussed in more detail in Section 61.5.4.6 but they can be highly efficient (quantum yields for hydrogen production up to 0.1)2293230 and problems associated with relay hydrogenation are removed. There is some dispute as to the importance of the size of the platinum particles in determining the rate of hydrogen production in the system [ Ru(bipy),J2 '/EDTA/MV2+/Pt but most researchers suggest213*219.22' that with PVA supported catalyst, the rate increases with decreasing catalyst size (at a given concentration) and then falls below ca. 20-30 A.211*212+21y The data available are rather limited and it has been shown'*' that particles of average diameter 16A are as effective as semi-aggregates of diameter 1000 A. This may indeed by the case, but to assume that this implies If indeed there were no size no size dependence in between seems an over-~implification.~~' dependence, since measurements are made at constant catalyst concentration, it would be necessary to infer that each surface platinum atom were more active in the aggregates than in the smaller particles. It is more logical that the activity of the catalyst should be inversely proportional to the particle radius and, from the limited data available, this appears to be approximately the case (see Figure 7).226

Reciprocal radius o f P t particles

nrn.')

Figure 7. Variation of rate of hydrogen production with the reciprocal of the particle radius in the system, [R~(bipy)~]'+ (4x mol dm-3), MV2+ mol dm-3), EDTA ( 3 x lO-'moI ~ I I - ~ t)'F, protected with poIy(v-inyl alcohol) (1.4 x lo-' mg dm-3) (data taken from ref. 226).

The fall off in efficiency of the catalysts below 20-30A can probably be attributed either to problems associated with electron transfer to the particle or, more likely, to the inability of the small particIes to accumulate sufficient electrons (or hydrogen radicals) to allow hydrogen production. The reaction specificity of the catalysts towards MV' oxidation may also be impaired. Qualitatively, the role of the redox catalyst can be seen as a way of accumulating reducing equivalents. For platinum, uptake of an electron from, for example, MV : is immediatdy followed by protonation so that hydrogen radicals are effectively stored:31 whereas for gold, protonation is much slower and the particle acts more as an electron store.231The surface of the particle may also allow specific orientation of the reducing agent so the hydrogen production at the remaining sites is easier.232 Quantitatively, the rate of hydrogen production is second order in [Pt] or [Au]232-234and, although this has been attributed to particle-particle interaction^,'^^ it is found that, if the particles are considered as microelectrodes. sDherica1 electrode kinetics will explain this behaviour. (at least for particles above 100 in Adiarneter), whereas homogeneous reaction kinetics 'will not.21 1,226,233,235-237 At sufficiently high [ Pt J the hydrogen-producing step has been r e p ~ r t e d ~ 'to' ,be ~~ 100% ~~~~~ efficient and not rate determining and it has been suggested that further work to improve the

Decomposition of Water into its Elements

515

efficiency of the catalyst is unnecessary.2L7However, this is not the case since ideally a platinum catalyst which was sufficiently active to compete with back electron transfer (so that all MV' underwent reaction 26 rather than 23), would allow hydrogen production and the accumulation of [Ru(bipy)13+ in the absence of a sacrificial electron donor and, since oxygen production from [Ru(bipy),I3+ is possible (see Section 61.5.4.5), this would allow cyclic water cleavage. For this objective to be realized, it is important that the platinum catalyst should react selectively with MV' and that it should not then transfer the electron back to the oxidized form of the sensitizer. Some rather elegant experiments towards this goal have been carried out. Thus, using colloidal platinum protected by the positively charged polysoap poly( N-hexadecylvinylpyridinium bromide) and C,,MV2+ as an electron relay, hydrogen can be produced'38 from either [Ru(bipy),12+ or [Zn(TMPyP)I4+ in the absence o f added reducing agent. The success of this reaction arises because C,,MVi interacts with the Pt particle in a manner similar to that in which it is incorporated into CTAC micelles (see Section 61.5.4.7.2) so that electron transfer to t'F is highly efficient, and because the oxidized form of the sensitizer is repelled from the platinum particle by the positively charged surface, hence preventing reduction of the oxidized form of the sensitizer by the reduced platinum particle. Using a slightly different but related approach, it has been shown203that colloidal platinum supported by CTAC can selectively interact with reduced diheptylviologen (HV') and not allow back electron transfer to [ZnTMPyP]'+ since this is electrostatically repelled from the reduced platinum particle by the surrounding CTAC molecules. Using the more hydrophilic MV2+, the platinum particles cannot interact with MV' since it too is repelled by the positively charged surface. Unfortunately, in neither case i s hydrogen produced on continuous photolysis. This may be because each pIatinum atom only accumulates a small number of electrons or because the surface coating increases the overpotential for hydrogen evolution. Attempts to support colloidal platinum particles with surfactant zinc porphyrin micelles allow hydrogen production at lower platinum concentrations than when the porphyrin and the particle are separate but an irreversible electron donor is still necessary for hydrogen production.116 Finally, as an alternative to colloidal metal particles it is possible to use hydrogenase, 133,153,186,187.200,202,238-248 normally from Desulphovibrio vulgaris, or a synthetic analogue derived from Fe,S, clusters248in the presence of bovine serum albumin, as the redox catalyst for production of hydrogen from MV' and protons. Hydrogenase has comparable activity to colloidal platinum (although the direct comparison is difficult since the concentration of hydrogenase is usually unknown) and has the advantages of longer lifetime2" (since agglomeration is not possible) and total homogeneity removing problems associated with light scattering. Its main disadvantages are its extraction, which requires skilled techniques and is time consuming, and its instability meaning that it must be stared at -196 "C.

61.5.4.5

Oxygen Production by Intermolecular Electron Transfer Processes

61.5.4.5.1 Uncatalyzed reactions

One of the major difficulties associated with catalytic photochemical water decomposition reactions is the requirement that four electrons be provided for each molecule of oxygen that is formed and there are very few compounds which allow this reaction to take place without the intermediacy of high energy species such as hydroxyl radicals. We therefore treat this subject in some detail. Examination of equation (49) shows that this kind of reaction is in theory possible with [ R ~ ( b i p y ) ~ ,] and ~ ' indeed it was the demonstration of oxygen p t o d u c t i ~ n from ' ~ ~ basic solutions of [Ru(bipy),13+ that spawned the enormous interest in the possible use of [Ru(bipy),]'+ as a sensitizer for water decomposition. It had been known249325o for some time that [M(bipy),l2+", which is chemiluminescent, was generated on production of [ M ( b i ~ y ) ~ ] ,(M + = Fe, Ru, os) in basic s o l ~ t i o n ~and ~ ~ that - ' ~ at ~ least for [FeLJ3+ (L=bipy or phen), oxygen was p r o d u ~ e d . 2 ~ ~ " There was also some evidence that this reaction was accelerated by light.251 Initially it was be1ieved2j2that hydroxyl radicals were the first formed products in these reactions but lack of any effect on the rate of the reaction by adding bromide, alcohols or benzene demonstrated that this was not the case. Furthermore, hydroxyl radicals react rapidly with [Ru(bipy),12' to give a complex with A,,, = 750-800 nrn which is similar to an intermediate in

* In acidic solution, different products can he ~ h t a i n e d . ' ~ ~

Uses in Synthesis and Catalysis

516

the reaction between [Ru(bipy)J3+ and OH-. This suggests that OH- bound to the metal or bipyridyl ligand might be being Highly complicated pH dependencies are observed for these reactions such that maximum 0, yields are obtained at p H 9 for M = R u (80%)’05 and 13 for M = F e (70’/0).~~*One report259 suggests that 0, can be produced at pH 1 from [Ru(bipy)J3+ in an uncatalyzed reaction. For M = Os, however, oxygen is evidently not produced, and more recently it has been suggested by a number of ~ o r k e r s , ~ that ~ O in - ~the ~ ~absence of catalysts 0, is not produced from any of these species. Reduction of M3+ does occur, however, the reducing agent being bipyridyl, which can be oxidized to C0,.262 There is some evidence to suggest that photochemically excited [Ru(bipy),13+ does oxidize water to oxygen in the absence of added catalysts.26’ Recent studies26’of the [Fe(bipy)J3+ system have shown that, although O2is produced from [Fe(bipy),13+ in basic aqueous solution, it is only formed if dissociation of bipyridyl occurs. The ~ + then be attributed to the lower lability much lower yields of O2 obtained from [ F e ( ~ h e n ) ~ ]can of this complex. The complex formed on dissociation of bipyridyl is believed to be (19) or possibly the oxo-bridged dimer [Fe(bipy),O];+, which can be formed [Fe(bipy)2(0H)2]Z+ from [Fe(bipy)J”+(n = 2 or 3), and which is known to catalyze 0, production from water.266 Although these FeIVcomplexes do not themselves produce oxygen from water, it is believed261,266 that they catalyze the production of oxygen from the bipyridyl N-oxide (bipyO) complex [Fe(bipy0),12+ (18), which in turn is formed along with [Fe(bi~y)~]’+ from [ F e ( b i ~ y ) ~ ]OH~+, and [Fe(bipy),OH]” (see Scheme 3). The exact nature of this last mentioned hydroxo intermediate is unclear. It may be an ion pair or it may be formed by attack of OH- on the metal, a nitrogen atom or a carbon atom of the bipy group, Attack on C would appear to be eliminated by the observation that bipye) is a reaction product. It has been suggested267that attack on the metal leads to 0, production, whilst attack on the ligand, which is favoured by electrophilic substituents, leads to ligand hydroxylation. Perhaps electron donating substituents would lead to higher yields of 02.A discussion of the reactions of metal bipyridyl complexes with water and hydroxide ions appears elsewhere in these volumes. bipy+‘[Fe(bipy)(OHt2]’

e

[Fe(bipy)J3’

[Fe(bipy),OH]’+

/

(19)

5[Fe(b1py)~]~++ 50H-

[bipyOl

6

[Fe(bipy0),l2’+ 5[Fe(bipy),l2++3H2O

uy Scheme 3

Mechanism of ‘uncatalyzed’ formation of O2 from [Fe(bipy)J*’

The release of oxygen from [Fe(bipy),13+ is believed to occuf as shown in Schemes 3 and 4 and does not involve high energy radical intermediates.261The first step, attack of hydroxide ion on [Fe(bipy)J3+, is generally accepted as10532589261 being rate determining but other workers have concluded250that FeIVintermediates of the form [Fe(OH)(bipy)J3+ produce hydrogen peroxide by reaction with OH-.and that [Fe(bipy)J3+ oxidizes H 2 0 2to 0,. Evidence for these intermediates is lacking, although it has been reported that H202 can sometimes be detected as a This result has recently been questioned.26*

,---.

(19~+[Fe~biavi3)2~f[bipy01+0, N N=bipy Scheme 4

Mechanism of oxygen formation from [Fe(bipy0)J2+ 261

As far as ruthenium is concerned, a number of unidentified ruthenium containing products are ~ b t a i n e d ’ ”in ~ small amounts during the reaction of [Ru(bipy)J3+ with OH-. These could be compounds such as [Ru(bipy),O]’+ (with monodentate bi~yridyl,~’),[ { R ~ ( b i p y ) ~ ( H ~ 0 ) } , 0 ] ~ +

Decomposition of Wuter into

its

Elements

517

or even ~is-[Ru(bipy),(H,Q)~]~+~~~ all of which (or close analogues) have been shown to be active catalysts for water oxidation by oxidizing agents such as Ce4+ or even [ R ~ ( b i p y ) J ~ + ? ~ " The available evidence clearly indicates that little or no oxygen is produced in the uncatalyzed reaction of [M(bipy)$+ with OH- and that small amounts of decomposition lead to adventitious catalysts for oxygen evolution. It is hardly surprising therefore that considerably more efficient 0, generation is observed if catalysts are added.

61.5.4.5.2 Homogeneous catalysts for oxygen production

Hexaaqua ions of a large number of transition metals, e-g. Mn"'' CUI', Co", Fe", Fe"' and catalyze the production of ox gen from [M(bip j3I3+ (M = Ru, Os or Fe) and one 9 electron oxidizing agents such as [hcl,] and [Co(H,O),] Y+ over a wide pH range. Yields up to 75% of 0, are obtained from [Ru(bipy)J3+ catalyzed by Co" at pH6-12 and yields of decomposition products containing Ru are much lower than for the uncatalyzed in terms of The kinetics of this reaction at pH 7 are highly complex but can be interpreted2633274 rate determining formation of a cobalt(1V) intermediate which oxidizes water to H202 followed by oxidation of Hz02by [Ru(bipy),I3+ (see equations 84-88). Unfortunately, reaction (89) causes loss of cobalt as an insoluble oxide and this accounts for the observed decrease of catalytic activity with time. ~ i 1 262,272 1

z

CO(OH)~+ZH+

Co2++2Hz0

Co(OH),+ [R~(bipy)~]'+," [Co(OH)J++ [Ru(bipy)J'+

[Co(OH),]++[Ru(bipy),13+

5 [CoO]Z++[Ru(bipy)3]2+ t HzO

[C00]2*

4

Z [ R ~ ( b i p y ) ~ ] ~HzO, ++

C02++HZO2

+

[R~(bipy)~j~++0~+2H+

[coo]2++co2+ -+ [COOC0]4+

4

ppt

For manganese, a slightly different mechanism involving formation of [MnO,]- and MnO,, presumably via oxidation to [Mn04]*- and disproportionation, has been and oxygen is then formed by the known oxidation of water by [MnOJ catalyzed by MnO,. A similar mechanism accounts for the catalysis of O2 production from PbO, and water by Mn'r,275as well as for 0, production from Ce4+ catalyzed by MnO, or C O ~ O ~ . ~ ' ' Complexes of similar metals with other ligands such as bipy, ammonia, and 1,2-diamin0ethane'~~ are also catalytically active for oxygen production from [ Ru(bipy)J3" but it is believed that partially hydrolyzed forms are the active species. Many of these complexes are more active oxygen evolving catalysts than the parent hydrated species and it has been concluded that cis-diaqua complexes which can dimerize to di-p-hydroxo forms are the active species.277Since considerable activity is observed at pH(5, where on account of the redox potentials of [R~(bipy),]~+'~+ (1.26 V) and Hz02/H20 (1.4V), equilibrium concentrations of H20, can only be mol dmP3,it is concluded that genuine four-electron oxidation of H 2 0 to O2 occurs, perhaps by a reaction such as (90) involving a dioxo bridged intermediate. The most active species appear to be those containing Co and Fe. 2[C00]2+

-L

2c0~+'00,

(90)

Similarly, using phthaloc anine or porphyrin complexes of a range of transition elements, cobalt and iron again a ~ p e a toJ be ~ ~the best metal ions. Although the mechanisms of the reactions are not fully understood, it is believed that two-electron oxidation is again important and some correlation between oxygen yield and redox potential (M3+/M2+) for the phthalocyanine complexes is observed. The anomalously low efficiency of zinc compounds compared with those of cobalt, which have similar first oxidation potentials, suggests that the second oxidation potentials are also important.278 The catalytic photochemical oxidation of water can also be achieved272s279 by similar methods, for example photolysis of [Ru(bipy)J'+ using [Co( NH3)5C1]2+as a sacrificial electron acceptor. The quantum yield for O2 production is ~ a . 0 . 0 2 5in~ the ~ ~ absence of added redox catalyst, although this result has been questioned.280,281 The mechanism of this photochemical reaction is as shown in equations (91)-(94) and relies upon oxidative quenching of [Ru(bipy)J2+* by the cobalt complex to give [Ru(bipy),13* and highly labile cobaIt(l1) amine complexes which decompose at a rate which competes with back

Uses in Synthesis and Catalysis

518

electron transfer to yield much more weakly reducing aquacobalt( TI) complexes. The [ Ru(bipy)J3+ then oxidizes water either in the presence of an added Coz* (Fe”, Co“, Ni“ or preferably insoluble Cot” or Fe“’ hydroxides if [SZO8]’- is used as acceptor)273or relying upon the cobalt(II) complexes formed from [Co(NH3),C1l2+ in the photochemical r e a c t i ~ n . ” ~ [Ru(bipy),]’+

[R~(bipy)~]’+*

[Ru(bipy),]’+*+ [Co(NH,),C1I2+ [Co(NH3),Cl]+f6H,0

+

4

[Ru(bipy),13*+ [Co(NH,),CI]+

[CO(H~O)~]~++~NH~+CI-

[Ru(bipy),l3++ OH- % [Ru(bipy),]2++0.250,f0.5HZ0

(94)

Efficient O2 production is observed over a narrow pH range (4-6) and it is probable that both the rate of oxygen production and the yield maximize at pH 5.280 It is also found that the dependence on added [Co2’] is complicated, the rate maximizing at mol dm -3, whilst the 0, yield decreases as [Co”] increases?727280 This photochemical reaction, as well as O2 production from Pb02, can also be ~ a t a l y z e d ~ ~ ~ by [R~(bipy),(H,O),]~+in free solution or with all the components adsorbed onto hectorite at pH values down to 1. The independen~e,’~of the rate of 0, production upon pH (Figure 8) is remarkable and suggests that catalytic oxygen generation itself is not rate determining but rather that the rate limiting step is some physical process such as ion transport, cage recombination, etc.

1

2

3

4

5

6

7

8

9

PH

Figure 8. pH dependenw of the initial oxygen evolution rate in the following photochemical test system: ( W ) SO ml or [Co(NH3),C1]*’ saturated H,O, 10Omg of hectorite 10% of the cations of the clay exchanged with [R~(bipy)~]‘+ and 10% exchanged with [R~(bipy),(H,O),]~+;(e) same total concentrations, but in a purely homogeneous medium (reproduced from ref. 283)

The active isomer of the Ru complex is believed to be cis despite the fact that it isomerizes to trans on adsorption on the clay mi11erals.2~~ It is believed that the isomerization does not occur with 100% yield and that the small amount of unisomerized ci~-[Ru(bipy),(H,O),]~+is responsible for the observed catalytic activity. This is supported by the observation that reactions starting from tran~-[Ru(bipy),(H~O)~]~+ do not produce 0,. Once again, the overall reaction is believed to involve two-electron oxidations to a ruthenium( IV) oxo species but in this case oxidation to a monomeric dioxoruthenium( IV) intermediate apparently takes place. Loss of 0, occurs via a peroxo intermediate (equations 95-98, equation 96 occurs s t e p ~ i s e ) ?The ~ ~ need for the cis isomer is then obvious. Unfortunately, catalyst instability (only about 30 electrons can be transferred per molecule) means that this interesting system is not suitable for commercial exploitation.283

+

[Ru(bipy),I3+ cis-[Ru( bipy)2(H20)2]2+ --c [Ru(bipy),]’+

+ H+ + ~is-[Ru(bipy)~(OH)(OH,)12t

~is-[Ru(bipy),(OH)(OH,)]~++3[Ru(bipy),]’+ -+ c i s - [ R ~ ( O ) ~ ( b i p y ) ~ 1 ~ ’ + [Ru(O,)(bipy),]” ~is-[Ru(O)~(bipy),]~+

2H20 [Ru(Oz)(bi~~)z12++

+

[Ru(bi~~),(H~0)~1’++0~

(95) (96) (97) (98)

One other interesting aspect of and related279systems is that dinitrogen is a product. It is apparently formed via catalytic oxidation of NH3 or NH4+ which are released by the rapid solvolysis of [CO(NW,),CI]~. The crucial importance of the use of an irreversible electron acceptor for oxygen production in these photochemical systems is demonstated by the observation that replacing [CO(NH,),C~]~+ by Fe3+ causes the yield of 0, to drop to presumably because back electron transfer

Decomposition of Water into its Elements

519

competes with OH- oxidation, even if the most active redox catalysts are employed. The one exception to this general observation is that at high pH (7-13.5) photolysis of solutions containing only [R~(bipy)~]’+ and MV2+leads284-286 to the catalytic formation of MV’ with an initial quantum yield of 0.029.286Since MV” is an efficient quencher of [Ru(bipy),]‘+*, it is probable that the mechanism of this interesting reaction is as in equations (21), (22) and (49) and the reaction of [ R ~ ( b i p y ) ~ with ] ~ + OH- must compete with back electron transfer to MV’. The exact fate of the [Ru(bipy),13+ is unknown except that [Ru(bipy),]’+ is regenerated. It has been suggested that O2 is produced but in view of the discussion above, it is perhaps more likely that a small amount of ligand oxidation occurs. Et would seem that attempts to carry out this reaction in the presence of O2 producing catalysts would be highly beneficial. The importance of this reaction lies in the fact that MV’ is efficient for hydrogen evolution in the presence of suitable catalysts but only at pH~*’” or PbO, yay Colloidal dispersions were for O2 production from C e 4 + Y X X ca. 100 times as active as powders.’” The original results concerning the reaction with Ce4+were questioned290and it was suggested that RuO, was oxidized to Ru04, which did not decompose to RuO, and O2 under the reaction conditions. However, it was subsequently pointed out”’ that the ratio of Ce4+/Ru02is crucial to the success of the reaction. At high catalyst concentrations and surface areas, the electrochemical potential of the RuO, particle is quite low and oxygen production is favoured over RuO, corrosion, whereas if only small amounts of Ru02 are employed, RuO, production will be the major reaction. Similar electrochemical arguments have been used to rationalize291 the observed decrease in oxygen yield with [Ce“’] at high [RuO,]. Using [Ru(bipy)J3* as oxidizing agent in thermal or photochemical reactions, adsorbed Ruv”’ species have also been proposed’” to be formed from Ru02. These can lead to Ru04, water oxidation or ligand oxidation to CO?. That RuV”’ intermediates are probably important has been elegantly demonstrated292by the observation that the pH at which O2 production begins for [M(bipy),13+ (M = Fe, Os or Ru) correlates well with the pH at which Ru04 becomes capable of being generated from Ru02 by [M(bi~y)~],+. Furthermore, water is then oxidized directly by Ru04 in a tetra-electronic step rather than via HzO,, which is thermodynamically inaccessible ]~+ (see Figure 9). These results also hold if [Ru(bipy)J3+ is photogenerated from [ R ~ ( b i p y ) ~ with [S2O,]’- or [Co(NH3),C1]*+as sacrificial electron donors, and higher rates of oxygen production are apparently observed.’92 This may be because photoexcited [ Ru(bipy),13+ is the water oxidant.265 Ru02 supported by polybrene increases the rate of reduction of [Ru(bipy)J3+ by two orders of magnitude and the rate increases with increasing pH from 1 to 10 and increasing RuO, c o n c e n t r a t i ~ n .Care ~ ~ ~ should, however, be taken in interpreting these results since the rate of formation of [ R~(bipy)~]’+ rather than 0, was measured. Other workers have shown262that O2 production in a similar system maximizes at pH 4 and does not occur above pH 7. Similar problems have been noted for [Fe(bipy),]” in the oxidation of water catalyzed by RuO, in that at p H 4 no 0, is observed but a product which has a similar visible spectrum to (i)

520

Uses in Synrhesis and Catalysis

PH

Figure 9. (a) OEyields as a function of pH in water oxidation with [M(bipy),]3+(10-3moldm-3) complexes in the presence of RuO, catalyst ( mol dm-’). (b) E-pH diagram for water oxidation, RuO, reduction and redox potentials of tris(bipyridine) complexes (0, [ R ~ ( b i p y ) ~ ] ~ + ”[7, + ;[Fe(bipy),]3+’2+; a, [ O ~ ( b i p y ) ~ ] ~ ” ”(reproduced } kom ref. 292)

[Fe(bi~y)~]’+is obtained. Interestingly, neutralization of this solution produces O2 and gives authentic [Fe(bipy)3]2+.294 The onset of oxygen production in these systems is at ca. pH 42893292,294 with both O2 yield289,294 and the rate of 02295 production maximizing at ca. pH 9 (pH 7t for powdered R ~ 0 ~ 1 The . 2 ~rate ~ of oxygen production is approximately first order in catalyst concentration. 289f95 Of the oxidants described above, only [Ru(bipy),I3+ can be efficiently prepared in a photochemical reaction and hence has potential for use in a cyclic water cleavage system. Many studies have, thus, been carried out on the RuO, catalysis of O2production from [ R ~ ( b i p y ) ~ generated ]~+ photochemically from [Ru(bipy),]’+ and a sacrifkial electron acceptor such as [Co(NH3)5C1]2+ or [s~oJ-. Initial studies suggested2’l that oxygen production occurred at pH 1.4 but that it was considerably faster at pH 4.9. Other colloidal metal oxides also catalyzed the reaction. Labelling studies confirmed that O2 was derived from wate?60*28‘but examination of the relative yields of Co2+ and O2when using [ C O ( N H ~ ) ~ C ~as] *acceptor, + suggest that only about 20% of the [Ru(bipy)J3+ generated produces 0, oxidatively.279Similar results have been suggested’” by quantum yield studies using Ru02/Ti02as catalyst, which show that although the quantum yield for [ Ru(bipy),]’+ formation is 0.3, that for 0, formation is only 0.03 (for a stoichiometric reaction, a value of ca. 0.075 would be expected). This discrepancy almost certainly arises279from the concomitant oxidation of NH, or NH4+ in the system to N2 as well as from a small amount of deep ligand oxidation of bipyridyl. This, perhaps, explains why better O2 yields can be obtained in the dark from [Ru(bipy)J3+ (see above) or using other irreversible donors such as PbOz, although in the last reaction, significant uncatalyzed O2 production also occurs.26o although In many of the reactions catalyzed by Ru02, reproducibility has been a it is apparently improved by using colloidal RuO,. (Non-hydrated RuO, is not catalytically active.294)Early experiments were carried out with RuO, (or Ru) derived from RuC132a8but colloidal Ru02 is preferably prepared from Ru04 by stirring with a polymer support, e.g. styrene-maleic anhydride copolymer (SMA), at pH 9 when spontaneous RuO, production occurs.294Alternatively KRuO;’, (which is more easily purified than RuO,) can be reduced with H202292or in boiling base289to an RuO, colloid which can be stabilized by sodium dodecylsulfate (SDS) or poly(viny1 al~ohol)?’~ Such stabilized RuO, colloids have radii of -90 nm and shelf lives of over six months289compared with 30 nm and less than one month for more conventional colloids.295Typical first order rate constants for the oxidation of water by [Fe(bip~)~],+ are -0.4 s-* at pH 7 for the SDS protected and -0.04 s-’ at pH 9 for the SMA protected

Somewhat similar results to those obtained with Ru02 have been obtained using Mn02296in the photochemical system employing [Co(NH3)5Cl]Ztas sacrificial electron acceptor. Thus, the

Decomposition of Water info its Elements

52 1

rate of 0, production increases to [MnO,] of 8.6 mol dm-3 and then drops slightly, presumably because of light absorption and scattering. The maximum 0,yield is observed at pH 3.6. This appears to be due to the fact that, although the initial rate of O2 production increases from pH 3.6, at higher pH, 0, is only generated for a short time (30min at pH4.9 and 6min at pH6.0). Apparently only activated p-MnO, catalyzes 0, production efficiently.296

(ii)

Supported metal oxide and related catalysts

The low reproducibility of certain of the systems involving heterogeneous catalysts, described above, as well as the potential for charge separation using semiconductor catalysts has led to the use of heterogeneous catalysts supported on semiconducting materials, e.g. TiO, . These supported catalysts are generally from commercial or specially prepared Ti02particles by stirring with RuCI3 at pH 4-6, filtering and drying at 100 "C. Catalysts prepared in this way are at least an order of magnitude more efficient for 0, production than powdered RuOZ2" or than those where RuO, is incorporated into the body of the Ti02 particle at the time of its preparation from, e.g. TiC1,.298Alternatively, the RuO, can be deposited from [ R u , ( C O ) , , ] ~ ~ ~ or RuO, 300,301 and may be polymer pr~tected.~" Typical particle diameters are of the order of 20nm3" and, as with unsupported Ru02, the hydrate, Ru02.3.6H20is apparently the active species.303 ~ eas thed sacrificial ~ electron ~ ~ donor~ in photochemical ~ ~ ~ , In general, [S208]'- has been ~ oxidation of water sensitized by [R~(bipy)~]'+.This has certain advantages since [Ru(bipy)J3+ is formed with a quantum yield of 2 according to equations (99)-( 101). However, this system has certain disadvantages since substantial loss of chromophore occurs on prolonged photolysis, and there is some suspicion2" that 0,may be formed without the intermediacy of [Ru(bipy)J3'-. Although it is generally accepted that in the absence of sensitizer [S208]2- and Ru02/Ti02 do not produce oxygen with appreciable rates in the dark or on irradiation, no data are available on the reaction of [S0J5 formed in equation (100) with water in the presence of RuOJTiO,, and it is at least possible that 0,may be produced from this reaction.

Nevertheless, [S2OSl2-provides a useful electron acceptor which, unlike [Co( NH3)5C1]2f,does not cause such large changes in the pH of the solution consequent upon its rapid decomposition after reduction. This has allowed combined studies of conductance and absorbance of solutions obtained3" in flash photolysis experiments to prove that 0,is liberated at the same rate as that with which [Ru(bipy),f3+ decays and hence to show that Ru02 does not 'collect holes' but rather passes them directly to a water molecule or, more likely, to an OH-. The rate determining step at low catalyst loadings or concentrations using [Co(NH3)5C112+280 or [s208]'- 298'302'303 is probably diffusional encounter of [ Ru(bipy),13+ with Ti02/Ru02 particles and this is thought3*, to have a second order rate constant as high as 5 x 10'' mol-' dm3 s-l, which is very close to the diffusion controlled limit. At higher catalyst loadings (>2.S% RuOz 3023303 or 10% if RuO, is incorporated into the body of the particlesz9'), this step is no longer rate determining and the rate can even of rate at higher catalyst loadings, which is also usually drop ~ l i g h t l y . ~ This ~ ~ ~ reduction "~ at higher concentrations of catalysts with a given loading, has been attributed to light absorption by the Ru02 leading to a lower overall light intensity for absorption by the chromophore. Consistent with this explanation are the observations that: (a) the rate of oxygen evolution is proportional to incident light and (b) in non-photochemical systems such a fall off is not observed at higher catalyst loadings.303In these cases O2 evolution increases3032304 with increasing RuO, coverage up to ca. 7% and then levels off. This levelling off is generally assumed to arise because at these high coverages, the catalyst is so efficient for electron transfer that this electron transfer step is no longer rate determining.28073003304 However, very recently it has been shown304that the total yield of 0, from Ce4+ catalyzed by RuO,/TiO, (under irradiation?) increases up to an RuOt loading of ca. 6% and then falls slightly. Interestingly, the magnetic susceptibility per gram of Ru02/Ti02 for the particles shows exactly analogous behaviour (see Figure 10). This suggests that paramagnetic particles are required for the O2 producing reaction and that at higher coverages, antiferromagnetic coupling between particles occurs to give lower magnetic moments.304These coupled particles may then be more susceptible to corrosion to RuO,, accounting for the lower 0,yields.

~

522

Uses in Synthesis and Catalysis

0

25

50

75

IO

RuOn in Ti02

Figure 10. (a) Evolution as percentage of stoichiometric O 2 From Ce4+ and H,O as a function of RuOz loading on TiO,. (b) Variation with composition of the magnetic susceptibility per gram (x,)of RuO,/TiO, as a function of RuO, loading on TiOz (reproduced from ref. 304)

The photochemical oxygen producing reactions are sensitive to pH, the rate constant for O2 production increasing by two orders of magnitude from pH 4-7.””’ Broadly similar conclusions have been reached using non-photochemically generated [ R ~ ( b i p y ) , ] ~ +or~ Ce4+ ~ ’ as electron acceptors, although no rate reduction is obs5rved at higher catalyst loadings or catalyst concentrations. With [Ru(bipy),I3+, the rate of reaction can be by 2-3 orders of magnitude compared with the uncatalyzed reaction and particular advantages are observed at pH 2-6, although the maximum absolute rates are obtained at pH 8-10. It should be noted, however, that these results refer to [Ru(bipy),13+ decay - 0, production was not measured.293 Because of their relationship to chlorophyll, metalloporphyrins have received considerable attention in water splitting systems. Success has been achieved in hydrogen production using sacrificial systems (see Section 61.5.4.3.2) but there is only one report of 0, p r o d u ~ t i o nUsing .~~ Fe3+or Ag’ as reversible electron donors, [ Z ~ T M P Y P ]produces ~+ 0, on photolysis in the presence of Ru02/Ti02. Maximum rates o f O2 production are observed at 1% Ru02 loading and above, and at pH 1.6 (Fe”) or 4.5 (Ag+). The rate using Fe3’ as acceptor drops by a factor of 5 at pH 2.5 or 0.8, the latter because of substantial demetatlation of the porphyrin. The mechanism of this reaction may involve quenching of [ZnTMPyPI4+ by Fe3+ and catalysis of the production of O3 by the RuOz/Ti02 from the subsequently formed [ZnTMPyP]’+. At first sight, it would appear that the efficiency of the catalyst must be exceptional since Fe3+/Fe2+is a reversible couple and back electron transfer is extremely rapid,280at least from [Ru(bipy),I3+. However, using the more highly charged [ZnTMPyPI4+, flash photolysis studies suggest that Fe2+ can be formed in high quantum yields from Fe3+ at low pH and in the absence of a catalyst. Thus, R u 0 2 / T i 0 2should be able to intercept [ZnTMPyPI5+before it can back react with Fe2+.The alternative, and perhaps more likely, explanation involves injection o f an electron from the excited state of the sensitizer into the conduction band of the TiO,. Charge separation then occurs at the semiconductor surface and [ZnTMPyP]’+ oxidizes H,O as before. The back electron transfer from Fe2+ is less likely in this case since Fez+ is generated by the injected electron at a site remote from the formation of [ZnTMPyP]’ ‘.94 In the absence of a redox catalyst, 0, is not produced even if [S,O,j’- is employed as electron donor?’’ More seriously, other workers have been unable to observe3060, production from identical systems using RuO,/TiO, as redox catalyst. A wide range of donors, pH and Ru02/Ti03 preparations was employed but oxygen could not be detected as a product on photolysis even though: (a) flash photolysis studies showed that [ZnTMPyPI5+ was formed, although its lifetime was short and it probably disproportionated or was further oxidized to [ZnTMPyPI6+, and (b) oxygen production was observed using the same catalysts and [Co( NH3)$1I2+ but [Ru(bipy)J2+ as c h r o r n o p h ~ r e . ~Furthermore, ’~ oxygen production using [ZnTMPyPj4+ was not observed even using Ru02/Ti02 provided by the group that had previously been successful in observing 0. p r o d ~ c t i o n . ’The ~ initial observations of oxygen production using this catalytic system must therefore, be viewed with some suspicion.

Decomposition of Water into its Elements

523

Other porphyrins which undergo"' photooxidation but which do not produce oxygen from their oxidized forms include [MTSPPI4-, M = Zn", Pd or SnC1, and [MTMPyPI4+, M = Pd or SnC1,. In view of the presence and undoubted significance of manganese in photosynthetic pigments it might be expected that manganese porphyrins and related complexes should be capable of taking part in oxygen producing systems. This is particularly the case since zinc porphyrins may be active (see above) and yet replacement of manganese by zinc in, for example, thylkaloid membranes causes a decrease in the activity towards oxygen A number of studies of phthalocyanine and porph rin complexes of manganese have been carried out and it is clear that both Mn"' and Mn' Y complexes can be photoreduced in the presence of water or OH- and in the absence of an electron a ~ c e p t o r . ~These ~ - ~ 'complexes ~ can be produced from Mn" or Mn"' complexes in the presence of an electron acceptor e.g. quinones,314 and the reaction can be sensitized by, for example, zinc porphyrins in suitable model membrane

environment^.^'^ However, all the available evidence suggests that these photoreduction reactions do not lead to oxygen f ~ r m a t i o n ~ (although ~ ' - ~ ~ ~ small amounts of H202 may be p r o d u ~ e d ) , ~ ' ~but , ~ rather '~ hydroxyl radicals, which subsequently attack the somewhat sensitive skeletons of the porphyrin or phthalocyanine Similar results have been obtained from porphyrin complexes of other metals.315Few attempts appear to have been made to catalyze 0, production from these higher oxidation state manganese derivatives and this would appear to be an area warranting further study. The supported R u 0 2 catalysts described in this section, and which are superior299to colloidal or powdered RuO,,~" are probably amongst the most active for water oxidation although they sufferfrom the serious disadvantages that Ti0, is an oxygen sponge and adsorbs O2 very strongly until it is saturated, whereupon further OH- reduction does not occur. This effect is enhanced by deposited redox catalysts and UV irradiation and varies depending upon the type of TiOl employed.300330' Oxygen absorption can be reduced by carrying out reactions in the presence of, for example, nitrate ion, but even then 20% of the 0, formed remains ad~orbed.~" Ru02 supported on solids other than TiO,, such as zeolite^^^',^'^ and clay mineral^,^" can also be effective in the production of oxygen from water. Once again, [Ru(bipy),13+generated thermally or photochemically using [Co(NH3),C1I2+ as electron acceptor produces up to 95% of the stoichiometric amount of 0,at pH 9 and this only drops to 75% at pH 1. For [Fe(bipy),I3+ or [Os(bipy),13+, the pH sensitivity is more dramatic but ca. 100% of 0, can be generated at pH 7 and 12 respectively using zeolite supported RuOz and the rate of reaction is much higher than is observed with unsupported RuO, .26873'6 With clay minerals, the three dimensional structure of the mineral plays an important part in determining the success of the experiment. Thus, for the normally structured hectorite, no oxygen is evolved3" on photolysis of supported RuOz and [Ru(bipy),12+ in the presence of aqueous [Co( NH3)&1]". This is because the cationic sensitizer is strongly adsorbed within the interlayer spaces whilst the RuOz aggregates on the outer surface of the mineral. Thus, RuO, and [Ru(bipy),I3+ cannot come into mutual contact. However, if the clay is collapsed by pretreatment in air or hydrogen at 600 "C, both [Ru(bipy),12+ and RuOa must be on the exterior surface of the clay and O2evolution does occur317when this system is irradiated with aqueous [Co(NH,),Cl]*+. Sepiolite has smaller interfibre distances and so enetration by [Ru(bipy)J'' does not occur and oxygen is readily evolved using this support.3 2 Finally, R u 0 2 supported on WO, has proved mote effective49for photochemical O2 production from Fe3+ than WO, The reaction requires UV light and is inhibited by Fez' and by higher concentrations of Fe3+.Quantum efficiencies are low and other metals, e.g. Pt, Rh or Ru on W 0 3 , reduce the rate of 0, production. 61.5.4.6 Cyclic Water Cleavage - Hydrogen and Oxygen Production uia Electron Transfer Catalysis

Examination of the various half reactions for hydrogen and oxygen production described in Sections 61 54.2-61 -5.4.5 suggests that some of these might be coupled to give cyclic water cleavage. For example, hydrogen can be produced from a sacrificial electron donor, a chromophore, an electron transfer catalyst and a redox catalyst. The role of the sacrificial electron donor is to * More recent studies suggest that oxygen can be produced from [S2ORJ2-and [ZnTSPP]4-.307

524

Uses in Synthesis and Catalysis

remove the oxidized chromophore formed during the first electron transfer event and intercept the energy wasting back electron transfer reaction. Since the oxidized form of the chromophore, e.g. [Ru(bipy)J3+ or [ZnTMPyP]'+, can, in the presence of suitable catalysts, oxidize water, hydrogen and oxygen production should be possible. The only criteria for successful water cleavage are that the rate of oxidation of water by the oxidized sensitizer (e.g. equation 105) must be sufficient to compete with back electron transfer (equation 104) and that the pH values at which the two half reactions occur must be compatible. The first reports of successful water cleavage followed the development of RuO, as a catalyst for O2 production from [Ru(bipy),13+. Simply mixing colloidal RuO, and colloidal platinum protected by styrene-maleic anhydride copolymer with [Ru(bipy)J2+ and MV2' leads"' to H2 and O2 production on photolysis according to equations (102)-( 106) with quantum yields of ca. 1.5 x lo-'. This value is 100 times smaller than that for hydrogen production in a similar system using EDTA as a sacrificial electron donor,225partly because back electron transfer from MV' to [Ru(bipy),j3" occurs at a rate comparable to that of O2 production from [Ru(bipy),13+ and water, and partly because oxygen produced during the reaction reacts with MV' regenerating MV2+and giving H202but preventing H2 production. This reaction has proved somewhat fickle and is highly dependent upon the quality of the RuO, catalyst. Indeed, it has recently been claimed3I9that this reaction does not produce H2 and O2.$ Similar processes have been reported using [ZnTMFyPI4+ as sensiti~er.9~ [Ru(bipyj312' -!% [R~(bipy)~]'+* [Ru(bipy)$+* +MY2'

-

[Ru(bipyj,l3+f M P [Ru(bipy),]'++OH-

4

KuO,

-P

[Ru(bipy),I3'

+ MV'

[R~(bipyj~]~'+MV~+

[R~(bipy)~]~++O.250~+0.5H~O

MV'+H+ -% MV2++0.5H,

In principle, since RuO, supported on Ti02 is more active for O2 production than Ru02 alone, this catalyst, too, should be capable of interacting with [Ru(bipy)J3+ at a rate competitive with that of back electron transfer. Such reactions leading to cyclic water cleavage with efficiencies up to 33% those of systems involving sacrificial electron donors and hydrogen production rates of up to 4 5 ~ m ~ d m - ~ h have - ' , been reported?2' The amount of oxygen produced is less than stoichiometric, probably because of absorption onto TiO?, which is known to be an oxygen ~ p o n g e . ~ ~ *However, -~" the rate of H2 production is undiminished after 48 hours photolysis. Originally, it was thought that TiO, was simply an innocent support for Ru02 and FV but subsequent study has shown that this is not the case and that cyclic water cleavage occurs, albeit with ca. 25% efficiency, if MV2+ is omitted from the mixture.297A diagrammatic representation of what is believed to occur in the absence of MV2* is shown in Figure 11. As in other systems, light absorption forms [Ru(bipy),12+*which injects an electron into the conduction band of TiO, . Electron energy

Figure 11. Schematic representation of processes involved in the photochemical decomposition of water catalyzed by [R~(bipy)~]'+(S) and colloidal TiO, bearing R and RuOZ

Charge separation occurs immediately and the electron passes to an island of platinum on the surface of the particle, where it effects proton reduction. Meanwhile, [Ru(bipy)J3+ efficiently oxidizes OH- to 0,at the adsorbed islands of RuO,. Thorough studies of this reaction have shown quantum yields up to 0.002 at pH 4 and hydrogen evolution rates of 0.38 cm3 h-', whereas with MV" present 4 = 0.06 at pH 5.5 and hydrogen is produced at 1.2 cm3h-' from 25 cm3 of $It has been suggested that H, production in this system may arise from the Pt support acting as an electron donor.""

Decomposition of Warer into its Elements

525

solution. All attempts to reproduce these experiments in other laboratories have failed and it is reported that even the original investigators have been unable to repeat these observation~.~’~ In the absence of a photosensitizer, Ti02/pt/Ru02 is reported to produce H, and O2 on UV photolysis in water, by direct band gap excitation, with a quantum yield of 0.3.2’’ Another system for which both H2 and O2 production has been observed323involves the use of Ru02 and [Ru(bipy)J2+ cosupported on sepiolite combined with platinum deposited on Eu3+ embedded in alumina. The two solids, alumina and the clay, associate in solution to give a system as shown in Figure 12. Photolysis of this system produces hydrogen and oxygen in an oscillatory reaction as shown in equations (102)-( 106) but with Eu3+ replacing MV2+,323[Ru(bipy)J2+* reduces Eu3+ for €I2 production whilst [Ru(bipy),13’ oxidizes OH- in the presence of RuOz. Turnover numbers in this system are low, perhaps because of poisoning of the Ru02 catalyst.

Figure 12. Schematic representation of water decomposition using sepioljte and europium(II1) (A) exchanged alumina as support for [Ru(bipy),]’+ (D) (reproduced from ref. 323)

Finally, it has recently been r e p ~ r t e d ~that ~ ” ’irradiation ~~~ of a solution of [R~(bipy)~]’+ containing colloidal K[Fe( FeCN)J (Prussian blue, PB) produces hydrogen and oxygen continuously. It is necessary for both PB (hm,,=700nrn) and [Ru(bipy)J2+ (hm,,=452) to be irradiated and for potassium or rubidium ions to be present in the solution (the reaction is inhibited by Naf, Li+ or Ca2+). Evidently, PB performs a variety of different functions in this reaction. Control experiments show that it can act as an electron transfer catalyst from [Ru(bipy)J2’*, as a chromophore for H, production from strong electron donors, and as a catalyst for O2 production from [Ru(bipy)J3+ at pH as low as 2. Quantum yields in this system are -0.1 x loe3 but it does appear to be a genuine example of simultaneous H2 and O2 production. It must be concluded that, although intermolecular electron transfer reactions can show very high quantum yields (up to ca. 0.6) for both hydrogen and oxygen formation, attempts to couple production of the two gases together have been disappointing. No generally reproducible system with anything approaching commercially viable efficiency has been reported. Undoubtedly there is promise in these systems but other systems in which reversible electron donors or acceptors are employed, but for which back electron transfer is severely inhibited, must be urgently sought. This will undoubtedly require much more active redox catalysts, in particular for oxygen evohtion since platinum cataIysts of very high efficiency for hydrogen production are known. However, the major contribution in this field is likely to come from the use of organized assemblies (micelles, bilayers, membranes, rnicroemulsions, etc.) to promote charge separation and to aliow the use of reversible electron donors and acceptors with back electron transfer reactions being slowed down by intervention of the submicroscopic two phase systems.

61.5.4.7

The Use of Molecular Assemblies in Water Splitting and Related Reactions

61.5.4.7.1 Introduction

The high surface potentials and differential polarities of molecular assemblies such as micelles, vesicles and microemulsions suggest that they may be of use in effecting charge separation after a photochemical redox event either by preferential electrostatic repulsion of one of the products or by differential solubilities of the two products in the different phases. This area of research has been extensively r e v i e ~ e d ~ ~ and ~ -we ” ~give a brief overview of the use of these systems.

Uses in Synthesis and Catalysis

526 61.5.4.7.2 Micelles

Molecules which have hydrophilic head groups attached to hydrocarbon chains are surface active but form micelles in water above a certain concentration (the critical micelle concentration, CMC). These micelles are generally spherical or cylindrical and have a positive, neutral or negative surface with a hydrocarbon like interior. Although a large number of different possible ways of using micelles to promote electron transfer reactions has been discussed,329we consider only those in which hydrogen, or a hydrogen precursor, is generated using a metal complex as sensitizer. The simplest use of micelles is in the solubilization of hydrophobic species and early experiments showed that [ZnTPP] in TritonX-100 micelles could act as a chromophore in the photochemical reduction of MV2+by ethanethiol or TEOA. Hydrogen was produced if hydrogenase or colloidal Pt were added?31 Zinc porphyrins in neutral micelles (TritonX-100) hydrogen on photolysis in the presence of TEOA, bipy and K2[RC1,]. Secondly, micelles can be used to retard back electron transfer reactions between products produced in the primary photochemical redox reaction. Perhaps the most impressive reaction of this kind is the 1000 fold reduction in the rate of back electron transfer between methyltetradecylviologen cation radical (C,,MV') and [ R ~ ( b i p y ) ~ ]formed ~+ on photolysis of C,,MV2+ and [Ru(bipy),]'+ in the presence of CTAC micelles rather than in free solution.i39 At the optimum C14MVZfconcentration, it is present in the free solution, does not incorporate into the micelles and does not self micellize. However, on electron transfer, the cation radical becomes amphiphilic and immediately incorporates into the micelle whilst [Ru(bipy),13 ' is repelled by the positive surface of the micelle. The rate of geminate recombination (equation 104) is < 2 x lo7 mol-.' dm3 s-l. Evidently CI4is the optimum chainlength, the C , , derivative being hydrophilic before and after electron transfer and the CI6and C18derivatives being incorporated into the micelles as d i ~ a t i 0 n s . l ~ ~ Similar effects are observed using [ZnTMPyPI4+ as sensitizer, or in the absence of CTAC provided the concentration of C,,MV2+ is above its CMC. Under these conditions, the molecules in free solution are reduced and then incorporated into the micelles.139 Replacing CTAC micelles with a cationic polysoap protecting colloidal P t also leads to successful charge separation and in this case H2 production is observed in the absence of a sacrificial electron donor.138The use of a positively charged polysoap is highly important since it allows reaction of C,,MV' with the R catalyst but protects [ R ~ ( b i p y ) ~from ] ~ + the reduced t'F particle by electrostatic It has been claimed that, in the presence of R u 0 2 , H2 and O2 are produced but details have not been published.'38 A related example, which simply utilizes the surface potential for charge separation, involves quenching of [ R ~ ( b i p y ) ~ ] ~by+ PSV * (see Section 61.5.4.2.3). Positively charged micelles or bilayers enhance the quantum yield for PSVT formation over that in free solution by retaining it in the near surface layers of the micelle whilst expelling the positively charged [ R ~ ( b i p y ) , ] ~ + . ~ ~ ~ The importance of surface charge is amply demonstrated by the reductive quenching of surfactant [R~(bipy)~]'+*derivatives by pyrene-N,N'-dimethylaniline (DMA). Figure 13 shows the decay of photogenerated [Ru(bipy),]+ in the millisecond time domain for various different micellar environments and for free MeCN solution. Apparently the crucial step is the repulsion of DMA' by the micelle and hence CTAC provides considerably more efficient charge separation than does SDS.328

T r i t o n X - 100 v)

n

a

0 .I

i

-

I

0

SDS r

.

r

I

t

2

m

-

I

I

3

4

Time lmsl

Figure 13. Decay curves of surfactant [Ru(bipy),lt (515 nm), formed from photolysis of surfactant [Ru(bipy)J'+ and pyrene-N,N'-dimethylaniline,in various different environments (reproduced from ref. 328)

Decomposition of Water into its Elements

521

Very high quenching rate constants are obtained if the nitrogen derivative is attached directly to [ R ~ ( b i p y ) ~ making ]~+ it amphiphilic, however, very low yields of redox products are achieved even in the presence of EDTA, since back electron transfer is also extremely efficient.’49Addition of CTAC above the CMC makes little difference but if didodecyhiologen is also added, the rate of cation radical formation increases five-fold. Still higher rates of cation, radical production ( 5 5 ~ ) are observed if a polysoap with pendant viologen units is added.’49 Attachment of the ruthenium complex to the polysoap leads to even greater rate enhancements and it is believed that sequential passage of the electron from the ruthenium attached viologen after the photochemical electron transfer to neighbouring viologen units and then on down the chain of local viologens leads to efficient charge ~ e p a r a t i o n . ’ ~ ~ Cooperative effects have been observed when surface active zinc porphyrin derivatives and amphiphilic naphthoxyvioIogens are irradiated in the presence of EDTA.”’ Evidence for this comes from the observation that quantum yields of radical cations of the viologens on irradiation into both the absorption band of the zinc porphyrin derivatives and of the viologen are higher than the sum of the quantum yields obtained on irradiation of the individual absorption maxima. ]’~ the zinc porphyrin, but Similar results probably occur with surface active [ R ~ ( b i p y ) ~ replacing selective excitation is not possible in this case.333 Finally, it is possible to incorporate platinum particles into micelles of [ZnT0PyPI4+, which has a structure similar to that of [ZnTMPyPI4+, except that one of the pyridyl nitrogen atoms is quaternized with a CI8HT7chain, making it amphiphilic, either alone or diluted with CTAC.* Irradiation in the presence of EDTA produces hydrogen (4 = 0.004 f 0.002) by reductive quenching of the zinc porphyrin by EDTA. Addition of MVZ+increases the quantum yield to 0.01, either because of reduction of MV2+by [ZnTOPyPI4+*or because it is reduced by [ZnTOPyPI3+ formed by the reductive quenching by EDTA.

61.5.4.7.3 Reverse micelles and microemulsions When surfactant molecules are dissolved in organic solvents, the head groups cluster together and the hydrocarbon chains point into the bulk phase, forming reverse micelles. If water is trapped in the area of the head groups, the system is termed a microemulsion. Sometimes, a short chain alcohol (e.g. hexanol) is added to aid dispersion. These systems have been used in attempts to allow charge separation in electron transfer events. For example, with microemulsions of toluene containing dodecylammonium propanoate with EDTA and [Ru(bipy),12’ in the trapped water pools, it is possible to okserve reduction of bis(hexadecy1)viologen ((C,,),V”) in the near surface region (4 = 0.013). Interestingly, neither MV2+ nor diheptylviologen are reduced in a similar system made up of dodecane, CTAB and hexan01.~~~ It is also possible to showg8complete charge transfer across the oil-water interface, made up of dodecylammonium propanoate, to benzylnicotinamide in the surface region. This renders benzylnicotinamide neutral and it then extracts into the toluene phase as shown by further electron transfer to 4-dimethylazobenzene (see Figure 14). Again quantum yields are ca. 0.013?’

Figure 14. The separation of reduced acceptor (A, = 4-methylazobenzene) from oxidized donor ([Ru(bip~)~]~’) by an acceptor ( A = benzylnicotinamide) located at the interface (reproduced from ref. 98)

The only apparent example of hydrogen production in such a system involves electron transfer from benzenethiol in CHCIJoctane to water pools containing MVz* and hydrogenase. The chromophore is an amphiphilic [Ru(bipy)J’+ derivative embedded in the CTAC interface region.245The amphiphilic chromophore allows electron transfer ( 4 = 0.13) cu. three times more ~ ~ ~as, [Ru(bipy)J2+ ~~~ and about five times more efficiently than water soluble ~ h r o m o p h o r e ssuch efficiently than when the sensitizer (e.g. [ZnTPP]) is present in the organic phase.

528

61.5.4.7.4

Uses in Synthesis and Catalysis

Vesicles

In addition to the properties of micelles described above, vesicles, which are bilayer structures and can be considered to be model membranes, separate two distinct aqueous phases: an entrapped or inner water pool and the bulk aqueous phase. In principle, therefore, electron transfer may be possible across the bilayer and the sites of hydrogen and oxygen production in a water splitting system can be separated spatially. Although complete water splitting has not been observed in such a system, the principle can be tested since it is possible to form the vesicles in a solution containing one or more of the redox active components of an electron transfer system and then remove these components from the bulk solution, but not from the inner pools, by gel filtration or ion exchange. Other components of the electron transfer systems can then be added to the bulk water phase. It has been reported336that in systems of the kind shown in Figure 15, with a sensitizer and an electron donor in the bulk phase separated from methylviologen in the inner aqueous pools, electron transfer can occur directly across the membrane. Further studies, however, showed that MV2+ was escaping photochemically (perhaps because of localized heating) to the bulk phase where it was reduced.337

Figure 15. Originally proposed model for transmembrane electron phototransfer from EDTA to MV2+ oia [ Ru(bipy)J2+, It was subsequentlyshown that MV2' was escaping photochemically into the bulk aqueous phase337(adapted from ref. 341)

It is, therefore, necessary to have at least one component embedded in the vesicle bilayer. In general, this has been the chromophore, which can be hydrophobic, such as tetraphenylporphyrinatozinc(lI), or amphiphilic so that it aligns in the bilayer structure with the water soluble metal containing head group in the water. For example, in the system shown in Figure 16 with EDTA or NADH as electron donor trapped in the inner pools, [ZnTPP] in the bilayer and MV2' in the bulk solution, quantum yields for MV' formation of u to 0.1% have been observed on illumination.'86"8' This is enhanced to 0.57% if [Ru(bipy),] is also included in the inner pools and hydrogen is generated if either hydrogenase or polymer supported rhodium metal is added to the bulk phase. At pH9, with CoClz, a known oxygen producing catalyst, replacing EDTA, MV' can still be generated photochemically. It has been suggested that this system (similar to Figure 16) provides the basis for cyclic water cleavage, but oxygen has not been detected as a p r o d u ~ t . ' ~ ~ , ' ~ ~ Using surfactant chromophores, trans-bilayer electron transfer can also be observed, although in the absence of a mediator, the chromophore must be embedded in both surfaces of the bilayer structure. Thus, MV2' in the bulk aqueous phase can be reduced by EDTA or sodium ascorbate in the inner water pools if they are separated by a bilayer consisting of 5,10,15-tris(l-methylpyridine)-4-yl-20-(4-stearoxyph~nyl)porphinatozinccolyophilized with dipalmitoyl-L-phosphatidylcholine (DPPC).338 Similar results are obtained using surfactant derivatives of [Ru(bipy)J2+ and increased rates of MV' formation are observed if mediators such as hexadecylviologen;' 2-methyl-1,Cnaphthoquinone (vitamin K3)338,339 didodecyl- or dibutyl-alloxazine (DBS)328are employed. The rate of electron transfer and its dependence upon chromophore concentration suggest that the electron tunnels from a chromophore on the inner surface of the vesicle to one on the outer surface rather than being carried by flipping of the surfactant chromophore from one surface to the other?* In the system involving amphiphilic zinc porphyrin

R

Decomposition of Water into its Elements

5 29

MVZ Water

MVZ+

Figure 16. Scheme for successful electron phototransfer from EDTA to MV" through the bilayer of lipid vesicles containing (ZnTPP)(reproduced from ref. 186)

mediated by DBA, where the electron acceptor is disodium 9,10-anthraquinone-2,5-sulfonate (replacing MV*+), there is evidence that two photons may be involved in transport of the electron across the vesicle wall (Figure 17). Unfortunately, the actual rates of geminate recombination of the oxidized form of the sensitizer and, for example, MV' appear often to be higher in vesicles than in free solution, perhaps because of their higher viscosities and, although quenching rate constants can be very high, this is generally not reflected in the quantum yield for formation of MV' -328 HP Inner

H20 Outer

EDTA, RH 7

PH 7

AcONH~

ACONH4

buffer

buffer

Figure 17. Schematic representation for two-step activation of electron transport across a vesicle wall. ZnP is octadecylpyridiniumyltris(4-pyridyl)porphyrinatozinc(II) cation. DBA is 1,3-dibutylalloxazine,AQDS is 9,10-anthroquinone-2,6-

disulfonate (reproduced from ref. 328)

An interesting variation of this system involves embedding surface active zinc and manganese( 111) porphyrins in lecithin membranes. Provided that the porphyrin is negatively charged, and with zwitterionic PSV in the bulk aqueous phase, the Mn"' can be photochemically oxidized to MnrV.No electron transfer from PSVT to MnJVoccurs, perhaps because the MnIVis deeply embedded in the membrane.310 Using vesicles where the amphiphilic chromophore is embedded only in the outer wall, electron transfer across the bilayer is not observed in the absence of mediators whether the EDTA is in the inner pools with MV2' in the bulk phase or vice ver3u.338*339 With EDTA in the inner pools, electron transfer can, however, be effected on addition of, for example, ubiquinone, ZnTPP or H2TPP to the bilayer. In the last two cases, electron transfer is much more efficient than when the amphiphilic zinc porphyrin is omitted.339 Finally, a number of studies have employed viologens in the bilayer, either added as an In the amphiphilic reagent to DPPC,340or as the sole component of the vesicle wa11s.3373341,342 latter cases, double bonds in the hydrocarbon chain of the viologen have been polymerized to give added stability and rigidity to the bilayer structure. For added viologens, photochemical electron transfer from inner pools containing EDTA or K2C204 and [Ru(bipy),]*+ or [ZnTMPyPI4+ to methylene blue, [Fe(CN),]"-, [{Ru(bipy)2(H,0)}20]"f or [ S ~ M O , ~ O ~was ~ ] 'observed and in the absence of an electron acceptor in the bulk, viologen cation radical was formed. Again, tunnelling of an electron from a reduced viologen on the inner surface to one on the other surface is believed to occur. Quantum yields up to 0.1 have been observed.3A0

530

Uses in Synthesis and Catalysis

Using the polymerized viologen bilayers, the viologen units were removed from the outer surface to which they were attached by ester linkages by reaction with iminoethanol. Photolysis with [Ru(bipy),]*+ in the bulk phase then led to viologen cation radical formation on the inner surface. Presumably the iminoethanol groups acted as electron donors.342 Finally, vesicles have been used to store reducing equivalents by incorporating C,,MV’, formed from Cl,MV2+ on irradiation in the presence of EDTA and [Ru(bipy),12+,as multimers (monomers are obtained in the presence of CTAC micelles). These multimers are more stable to air than the monomers but generate hydrogen from water on the addition of carbowax protected colloidal platinum.343 From the foregoing discussion, it would appear that ordered assemblies have enormous potential for increasing the efficiency of water splitting systems since they can, under favourable circumstances, lead to efficient charge separation, although considerably more research in this area is required. In particular, the development of more efficient neutral electron transfer catalysts will assist electron transfer across bilayers. Bis(2,3‘-diphenylditholato)nickel has been successfully employed344as such a neutral electron transfer catalyst by dissolving it in diphenyl ether containing dicyclohexano-18-crown-6 supported in a millipore filter. Electron transfer occurs from EDTA via [Ru(bipy),]” and MV2+ on one side of the membrane to [Fe(CN),I3- on the opposite side. Direct transfer from [R~(bipy)~]’+* via the nickel complex is also possible if SDS is added to the aqueous solution containing [Ru(bipy)J2+. The more efficient electron transfer quenching by nickel in this case is believed to arise because the SDS molecules take [Ru(bipy)J2+ to the H20/Ph,0 interface. The electron is carried across the membrane by the [ Ni(S2C2Ph2)J0’couple.3q4Other neutral electron transfer catalysts include zwitterionic sulfonated viojogens (see Section 61.5.4.2.3). The development of highly selective redox catalysts for H2 praduction (see above) and investigations of oxygen producing reactions are also goals worthy of the highest priority. Other types of ordered assembly that would appear worthy of study in water-splitting systems include polyelectrolytes, particularly since their charged surfaces can enhance certain quenching rate constants by up to three orders of magnitude,345and solid supports bearing a surface charge. Some work on solid supports has been described in Section 61.5.4.5.3(ii), but another interesting series of reports concerns the use of colloidal SiOz. The negatively charged surface adsorbs [Ru(bipy),12+ and evidently holds it in such a way that oxygen does not quench the emission whereas the positively charged MV2* does. Very efficient formation of MV‘ is observed in the presence of EDTA or TEOA.346 Furthermore, whereas back electron transfer from the radical anion derived from the zwitterionic N,N’-bis(3-sulfonatopropyl)-2,2‘-bipyridiniumion (DQS) to [ R ~ ( b i p y ) , j ’ ~in alkaline homogeneous solution is so rapid that redox products cannot be observed on flash photolysis, in the presence of SiOz colloids the rate constant for this reaction, kb,is lo7moi-’ dm3 s-1.145 The role of the S O 2 , which has a negatively charged surface at this pH, is to repek DQST whilst strongly attracting [Ru(bipy),13+. In homogeneous solution these two products attract one another on account of their opposite charges and cage escape does not compete with back electron transfer. In addition, SiOz prevents DQST, once it has escaped from the cage, from returning into the environment of IRu(bipy),l3+and hence further retards the back electron transfer. In the presence of TEOA and colloidal platinum, hydrogen is evolved with C$ = 2.2 x lop3at pH 8.5. Similar results are obtained3473346 with PSV except that no effect on cage escape is observed, just a retardation in the subsequent back electron transfer by a factor of 100 compared with homogeneous solution. The less negative reduction potential of PSV means that hydogen cannot be generated at the high pH required for S i 0 2to have a negatively charged surface. Anthraquinone sulfonates can, however, be reduced by PSV’ generated in this way.349735o

61.5.5 PHOTOELECTROCHEMICALDECOMPOSITION OF WATER Although it is apparent from the foregoing discussion that the simultaneous catalytic production of hydrogen and oxygen from water has only been demonstrated in a very few cases and even most of these have proved irreproducible, all of the systems produce hydrogen and oxygen in very close proximity to one another. A possible way to surmount this problem is to use photoelectrochemistry and generate hydrogen at one electrode and oxygen at another. These electrodes can be in the same or different compartments but if they are in different compartments, changes in pH will occur and replenishment of H’ in the cathode compartment and of OH- in the anode

Decomposition of Water into its Elements

53 1

compartment will be necessary.* Using the same compartment can lead to short circuiting of the desired reactions. A large amount of research has been carried out into the photoelectrochemical production of hydrogen and/or oxygen from [R ~ ( b i p y ) ~ ] ~ + . It is possible to generate hydrogen from a solution containing [Ru(bipy)J2+, MV2+ and a sacrificial electron donor such as TEOA, EDTA or cysteine at a pH (8-10) where reduction of H+ by MV' is not thermodynamically possible. This is achieved3'l by attaching the solution via platinum anode and cathode to water at a lower pH and joining the two solutions by an agar/KCl bridge. Hydrogen is then produced from the cathode in the solution which does not contain photosensitizer etc. At lower pH values in the anode compartment, some hydrogen may be produced by direct Pt catalyzed reduction of protons by MV'. This problem can be overcome by use of a mercury anode"' at which the overpotential for H2 formation is higher and then both cathode and anode compartment can contain solutions at pH 4.7 with EDTA as the irreversible electron donor.352 Alternatively, hydrogen can be generated at a remote cathode from [R ~ ( b i p y ) ~generated ]+ by reductive quenching of [ Ru(bipy),12+ by Et3N.353 More effort has been devoted to studies of O2production. In this case, the cathode compartment contains [Ru(bipy),12+orand an electron acceptor such as [S208]2-,354-356 [CO(CzO.AI'-,"' [Co(NH3)5C1]2+,353.357 f l 3 + ,353 and a platinum or carbon cathode. The pH of the anode compartment depends not only on the oxidation potential of the [Ru(bipy)J3+, which can be varied by use of substituted bipyridyls, but also on the overpotential of the anode material for 0, production. For example, systems based on [Ru(bipy),]'+ itself can produce oxygen at p H > 7 at a Pt anode,351whereas smooth oxygen production is observed at pH 4.7 from an anode prepared from RuOz deposited on titanium Still lower pH values for the anode compartment are possible if [Ru(bipy),LJZf (L = dipropyl 2,2'-bipyridyl-4,4'-dicarboxylate)for which E 0 ( 3 + / 2 + ) is 1.59V (us. SCE)?56 There is, however, a suggestion that using RuO, and [Ru(bipy),12+, O2 production can occur from an anode at pH as low as 1.0.3s8 In the case of the substituted b i ~ y r i d y l ,it~is~found that the rate of O2production is dramatically enhanced by addition of Ce3+to the cathode compartment. Apparently, [ Ru(bipy),LI3+ decomposes rapidly once formed so that not all of it arrives at the cathode. Ce3+reduces [Ru(bipy),LI3+ in bulk solution and Ce4* accepts the electron from the cathode. The chemistry of these systems is identical to that already described for O2 and H2 production in homogeneous systems except that MV' or [ Ru(bipy)J3' generated during the photochemistry of the bulk solution then regenerate MV2' or [Ru(bipy),]'+ by donation or acceptance of an electron at the electrode. Photopotentials of -1.0 V and currents of -1 mA can be generated for long periods in this way. Similar photoelectrical cells using metal phthalocyanine or porphyrin sensitizers have also been d e v e I o ~ e d . ~ ~ ~ - ~ ~ * There have been two reports of cyclic water cleavage in a photoelectrochemical system. The first simply involves358addition of protected colloidai Pt to a solution containing MV2' and [ R ~ ( b i p y ) ~in] ~0.1 + M H2S04. Instead of MV" injecting an electron into the electrode for remote H2production, hydrogen production is achieved by reduction of H+ by MV' catalyzed by colloidal Pt. The [ R ~ ( b i p y ) ~ which ] ~ + accumulates then receives an electron from the t'F electrode and oxygen formation occurs at an Ru02 electrode in H2S04(0.5 mol dm-3) separated from the cathode Compartment by nafion. Apparently H, and 0, are produced photochemically with a photocurrent of 0.2mA. It is difficult to see how this reaction occurs with such efficiency since the rate of diffusion of [Ru(bipy)J3+ to the electrode and the rate of H2 production from MV' and H t are expected to be much slower than back electron transfer between [Ru(bipy),13' and MVt.358 Chinese workers have reported that cells based on n-TiO,/NaOH//[Rh(bipy),]CI,/Ptor Pt/[R~(bipy),]Cl//[Rh(bipy)~]Cl~/Pt also produce H2 and 0, on photolysis.363 One other approach to the photoelectrolysis of water that has been adopted involves the photosensitization of semiconductor electrodes such as Ti02,362,364,365 SrTiO, 3653366 or sno2265,365,367 by, for example, [R~(bipy)~]". The photochemically excited state of the chromophore injects an electron into the conduction bond of a semi-conductor; this is then passed uia an external circuit to a platinum electrode for H2 production. The oxidized form of the quencher then forms O2 apparently in an uncatalyzed reaction.? Unfortunately, all such systems * Using different compartments may, however, overcome reactions of oxygen with, for example, MV',

which can drastically reduce rates of water decomposition. t There is some evidence that photoexcited [Ru(bipy),I3+ oxidizes water much more rapidly than does the ground state.26s

CCC6-R

532

Uses in Synthesis and Catalysis

that have been reported require an external bias (usually ca. 1.0V) and it has been shown that O2 adsorbed on the semiconductor leads to low rates of electron transport364and even adsorbing the chromophore on the electrode physically26s or chemically365does not lead to substantial increases in efficiency. The only related system which allows H2 production with lower bias potentials is that derived from adsorption of [Ru(bipy),( MeOCH2CH20Me)]’+ onto a polysulfurnitride electrode. Hydrogen production is observed on irradiation of the face perpendicular to the fibre bundles with a bias of only -0.1 V relative to SCE.368 Finally, p ~ I y r n o l y b d a t e and ~ ~ ~tungstate371 ,~~~ ions can be used for photogalvanic hydrogen production on irradiation with UV light. The absorption of light produces OH- identified by spin trapping and a hydtoxomolybdenum(V) species, which injects an electron into a platinum electrode. Hydrogen is produced at a remote platinum cathode in H2S04(2.5 mol dm-3). Unfortunately, oxygen does not appear to be produced from OH. but rather oxidation of the molybdate occurs to give peroxomolybdate species.37oUsing tetrathiomolybdate and a carbon hydrogen can again be produced photoelectrochemically and the hydrolyzed catalyst can be recycled by addition of H2S.The overall reaction is then the photochemical production of hydrogen and sulfur from H2S.373

61.5.6 COORDINATION COMPLEXES AS CATALYSTS FOR THE ELECTROCHEMICAL PRODUCTION OF HYDROGEN OR OXYGEN FROM WATER 61.5.6.1 Hydrogen Production In view of its high overpotential for the production of hydrogen from water, the mercury electrode allows electrochemical evaluation of possible catalysts for proton reduction since redox potentials and efficiencies for hydrogen production can be measured. In general, a species that is catalytically active for hydrogen production will give an amplified wave in both polarography and cyclic voltammetry. The first waves of this kind were observed by B r d i ~ k ain~solutions ~~ containing proteins and Co2+ at cu. -1.3 V (vs,SCE). Subsequently, numerous studies have shown that these waves are produced from solutions of Co2+ or Co3+ containing a wide range of sulfur containing compounds. 3757376 Cysteine complexes have especially been studied. It seems that complex formation is very i r n p ~ r t a n t ~ ~and ~ - ”at~least two functional groups on the ligand must coordinate378 to the metal atom. One of these must be sulfur, whilst the other can be -C02H, -NH2, et^.^^* Analogous complexes with oxygen replacing sulfur do not produce the eff e~t.~’‘ Nitrohydr~xylaminate’~~-~’~ anion or NCS-387*388 will also produce a catalytic wave and some investigators have observed catalytic hydrogen waves in the absence of added ligand.389-391 Similar complexes of Ni2+ also give catalytic hydrogen waves in p~larography.~~’ It is known that a number of metals which do not amalgamate with mercury are capable of lowering the averpotential of hydrogen production at mercury electrodes and indeed catalytic hydro en waves attributable to metal deposition can be observed in polarograms o f simple salts of Ng !i* Ru,393-3Ys w , 3 9 4 Ir,3Y4,395 and ptm344,395For Au, Ag, Cu and Pd, which amalgamate with mercury, catalytic hydrogen waves are not observed.394 Interestingly, a catalytic hydrogen wave is obtained from bis( dithiocarbamato)zinc( 11) at a This is probably because the zero carbon (paste OT graphite) electrode although not at valent complex i s unstable with respect to loss of zinc metal to form an amalgam, whilst it is sufficiently stable on carbon to promote ~atalysis.”~ Initially, it was believed397that the Brdicka waves were attributable to cobalt deposition on the mercury surface leading to a lowering of hydrogen overpotential, and indeed this may be the case when no ligand is present or in the presence of, for example, h i ~ t i d i n e . 3However, ~~ attempts379 to measure directly the effect of cobalt deposition by coating the mercury surface with cobalt suggest that it does not lower the overpotential for hydrogen significantly and it has been suggested that even in the absence of added ligands, a zerovalent complex may be the active species in the catalytic cycle.389 Further evidence that cobalt metal is not responsible for the catalytic hydrogen waves comes from the observation that the higher the stability of the complex formed between the sulfurcontaining ligand and Co”, the more pronounced is the catalytic wave.37’ This is confirmed since

Decomposition of Wuter into its Elements

533

the wave is depressed in acid s ~ l u t i o n ”where ~ certain functional groups can be protonated off the cobalt leading to a reduction in the stability of the complex. The generally accepted,376although not fully proven, mechanism for the catalytic wave is as shown in equations (107)-(109) and involves reduction to a zerovalent complex which is the catalytically active species. Whether the catalysis involves two or one electron steps (Co2+ or Co’) intermediates is not known but it is probable that the neutral zerovalent complex is adsorbed on the electrode surface.380J96 M”L2+2e

+

MOL,+ nH+

MOL,+^^

MOLz MoL,H,

-. --L

MO~+OSI~H,

An alternative mechanism (equations 110-112) has been reported to operate in certain cases, for example for M L , M = Co2+or Ni2+, L = [ROCS,]‘, [R2NCS2]- or [RHNCS2]-.376A related mechanism is also believed3’* to be responsible for the extremely efficient (ca. 2 x lo5 catalyst turnovers h-’) catalytic reduction of water at a mercury electrode using [MH(PEt3)J+, M = pt or Pd. In these cases the system is further complicated since cyclic voltammagrams give inverted and amplified waves (Figure 18). In the first negative sweep, it is believed that a two-electron reduction (equation 113) occurs to give a monolayer of adsorbed [MH(PEt,),]- which is not catalytically active. The reductive catalytic cycle is then initiated (equation 114) on the first subsequent positive sweeps by a one electron oxidation to [MH(PEt3),Iads*.The catalytic cycle (equations 115-117)then continues until the potential is insufficiently reducing to drive equation (115).On the second and subsequent negative scans, the catalytic cycle restarts once the potential is sufficiently reducing to drive equation (1 15) and continues until the potential allows reduction of [ MH(PEt3)Jads. to the catalytically inactive [MH(PEt3),ladS- (equation 118). At concentrations higher than 5 x lop4mol dm-’ the cyclic voltammograms are complicated by desorption of [M( PEt,),]. Interestingiy, solutions containing platinum(1V) complexes and PH, also give catalytic hydrogen waves in their polarograms but these have been attributed to reduction of [pH,]+ catalyzed heterogeneously by the platinum ML,+H+ [ML,H]++e ML.+HA

[ML,H]+ -D

ML,+0.5H2

e

[ML,H)++A-

-E

(VI

Figure 18. Cyclic voltammograms of [PtH(PEt,),]+ (1.5 x l o 4 mol dm-’) in phosphate buffer (pH 6.88) at a mercury drop electrode: (a) first scan; (b) second and subsequent scans

61.5.6.2

Oxygen Production

There are comparatively few coordination complexes which lower the overpotential for oxygen production. Hydroxides of both iron( 11)400 and ~ o b a l t ( I I ) catalyze ~~~*~ oxygen ~ ~ production at a rotating platinum disc electrode at high pH. The mechanism of the reaction is analogous to that proposed for the catalytic production of oxygen from, for example, [Ru(bipy)J3+ (see section 61.5.4.5). Related to this is the use of [{(NH3)sCo)20]5+for the catalytic production of oxygen or hydrogen peroxide402at platinurn4O3or mercurym2 electrodes. Ruthenium trichloride catalyzes the production of oxygen at a platinum anode4” via higher oxidation state species, whilst one of the most effective catalysts for oxygen or chlorine evolution is formed by reduction of a solution containing [Fe2(S04),] and K,[Ru(CN)J~*~This leads to a film on the electrode of ruthenium purple, Fe,[Ru(CN),],, with a structure analogous to that of Prussian blue. Oxygen can then be produced at 0.2 V us. SCE.404 Finally, both manganese( 11) gluconate405and alkaline solutions of cobalt phthalocyanine^^^^ also promote catalytic oxygen evolution.

61.5.7 REFERENCES 2 Macabees, chap. 1, v. 21. L. B. McGown and J. O’M. Bockris, ‘How to Obtain Abundant Clean Energy’, Plenum, New York, 1980. J. O M . Bockris, ‘Energy Options’, Taylor and Francis, London, 1980. C. A. McAuliBe, ‘Hydrogen and Energy’, Macmillan, London, 1980. J. R. Bolton (ed.), ‘Solar Power and Fuels’, Academic, New York, 1976. J. S. Connolly (ed.), ‘Photochemical Conversion and Storage of Solar Energy’, Academic, New York, 1981. M. Dumas and D. Duveen, Chymia, 1959, 5 , 113. J. R. Bolton, A. F. Haught and R. T. Ross, ref. 6, p. 297. K. Kalymasundaram and M. Gratzel, in ‘Photovoltaic and Photoelectrochemical Solar Energy Conversion’, ed. F. Cardon, W. P. Gomes and W. Dekeyser, Naro Advanced Studies Institute Series, Series B, 1981, p. 349 (Chem. Abstr., 1981, 96, 22 352a). 10. L. Kruczynski and H.D. Gesser, Inorg. Chim. Acta, 1983, 72, 161. 11. Y.Doi and M. Tanaki, Inorg. Chim. Acta, 1982, 64, L145. 12. 0.I. Micic and M. T. Nenadovic, in ‘Energy Storage. Transactions of the 1st International Assembly’, ed. J. Silverman, Pergamon, Oxford, 1980, p. 445 (Chem. Abstr., 1982, 94, 194 89831). 13. I. A. Potapov, M. 8. Rozenkevich and Yu. A. Sakharovskii, Koord. Khim., 1981, 71, 229. 14. E. R. Buyanova, L. G. Matvienko, A. I. Kokorin, G. L. Elizarova, V. N. Parmon and K. I. Zamaraev, React. Kinet. Catal. Lett., 1981, 16, 309. 15. C. Shao, J. Li, L. Pan, X. Yang, M. Bei and H. Guo, in ‘Fundamental Research in Organometallic Chemistry. Proceedings of the 1st China-Japan-US Trilateral Seminar on Organometallic Chemistry’, ed. M. Tsutsui, Y. lshii and Y. Huang, Van Nostrand Rheingold, New York, 1982, p. 715 (Chem. Abstr., 1982, 97, 136 537r). 16. D. V. Sokoluukii, Ya. A. Dorfman and Yu. M. Schindler, Zh. Obshch. Khim., 1974,44, 20. 17. N. Bubnov, V. V. Voevodskii, N. V. Fok and B. N . Shelimov, Opt. Spektrosk, 1961, 11, 7 8 (Chem. Abstr., 1961, 55, 25 479h). 18. W. Traube and W. Lange, Ber. Dtsch. Chem Ges. B, 1925, 58, 2773. 19. F. S. Dainton, E. Collinson and A. Malati, Trans. Faraday SOC., 1959, 55, 2096. 20. R. E. West, H. Mahmoud, D. G. Burkhard, H. Ito and R. S. Kirk, NASA Doc N63-19875, Sci. Tech. Aerosp. Rep., 1963, 1, 1459 (Chem. Abstr., 1964, 60,7607b). 21. D. R. Eaton and W. R Stewart, J. Phys. Chem., 1968,72, 400. 22. D. L. Douglas and D. M. Yost, J. Chem. Phys., 1949, 17, 1345. 23. P. R. Ryason, Sol. Energy, 1977, 19, 445. 24. T. Arakawa, T. Takata, M. Takakuwa, G. Y. Adachi and J. Shiokawa, Mater. Res. Bull., 1982, 17, 171. 25. I. S. Shchegoleva, Khim. Vys. Energ., 1982, 16, 556 (Chem. Abstr., 1983, 98, 44059d). 26. R. E. West, US Dept. Comm. Ofice Tech. Serv. PB. Rep., 1960, 150 (Chem. Abstr., 1963, 58, 7537~). 27. T. S. Dtabiev, V. Ya. Shafirovich and A. E. Shilov, React. Kinet. Catal. Lett., 1976, 4, 11. 28. H. A. Bhakare and C. V. N. Rao, J. Indian Chem. Soc., 1974, 51, 543. 29. K. Jijee and M. Santappa, Roc. Indian Acad. Sci, S e d . A , 1967, 65, 155 (Chem. Abstr., 1967, 67, 103 199j). 30. M. Anbar and I. Pecht, J. Am. Chem. SOC.,1967,89, 2553. 31. M. Anbar and 1. Pecht, Trans. Faraday Soc., 196X, 64, 744. 1. 2. 3. 4. 5. 6. 7. 8. 9.

Decomposition of Water into its Elements

535

J. C. Sullivan and A. J. Zielen, Inorg. Nucl. Chem. Lett, 1969, 5, 927. Yu. A. Komkov and N. N. Krot, Radiokhimiya, 1970, 12, 227 (Chem. Abstr., 1970,73,39 1070). Yu. A. Komkov and N. N. Krot, Radiokhimiya, 1970, 12,692 (Chem. Absrr., 1971,74,91696~). V. Ya. Frenkel, I. A. Lebedev, I. A. Kulikov, A. A. Ryabova, B. F. Myasoedov and M. V. Vladimirova, Radiokhimiya, 1978, 20, 109 (Chern Abstr., 1978, 89, 15 9602). 36. V. Y. Shafirovich and A. E. Shilov, Kinet. Katal., 1978, 19, 877 and refs. therein. 37. G. Zimmerman, J. Chem Phys., 1955, 23, 825. 38. H. Goff and R. K. Murmann, 1. Am. Chem. SOC.,1971,93, 6058. 39. F. J. Shipko and D. L. Douglas, J. Phys. Chem., 1956, 60, 1519. 40. V. R. Evans and J. N. Wanklyn, Nature (London), 1948, 162, 27. 41. C. F. V. Mason, M. G. Bowman and M. A. David, J. Inorg. Nucl. Chern., 1978,40, 1739. 42. L. J. Heidt, M. G. Mullin, W. B. Martin and A. M. Johnson Beany, J. Phys. Chem., 1962,66, 336. 43. E. Hayon and J. Weiss, J. Chem. Soc., 1960, 3866. 44. J. Jortner and G. Stein, J. Phys. Chem., 1962, 66, 2200. 45. I. S. Sigal, K. R. Mann and H. B. Gray, J. Am. Chem. SOC.,1980, 102, 7252. 46. G. V. Buxton, S. P. Wilford and R. J. Williams, J. Chem. Suc., 1962, 4057. 47. W. Good and W. A. B. Purdon, Chem. Ind. (London), 1955, 1594. 48. A. A. Krasnovskii and G. P. Brin, Dokl. Akad. Nauk SSSR, 1962, 147, 656 (Chem. Abstr., 1963,58,9782d). 49. J. R. Darwent and A. Mills, J. Chem. SOC.,Faraday Trans. 2, 1982,78, 359. 50. L. J. Heidt, J. Chem. Educ., 1966, 43, 623. 51. L. J. Heidt, in ‘Proceedings of the World Symposium on Applied Solar Energy, Pheonix, Arizona, 1955’, p. 275 (Chem. Abstr., 1956,50, 11 095e). 52. L. J. Heidt and M. E. Smith, J. A m Chem. SOC.,1948, 70, 2476. 53. T. J. Sworski, J. Am. Chem Soc., 1955, 77, 1074. 54. M. G. Evans and N. Uri, Nature (London),1950, 166, 602. 55. F. R. Duke and J. A. Anderegg, Iowa State Coll. J. Scl, 1953,27, 491 (L‘hem. Abstr., 1953,47, 10974d). 56. J. Weiss and D. Porret, Nature (London), 1937, 139, 1019. 57. L. J. Heidt, Proc. Am. Acad. Arts Sci., 1951, 79, 228 (Chcrn Abstr., 1952, 46, 2915). 58. L. J. Heidt and A. F. McMillan, Science, 1953, 117, 75. 59. R. J. Marcus and H. C. Wohlers, Ind. Eng. Chem., 1959, 51, 1335. 60. R. J. Marcus and H. C. Wohlers, Ind. Eng. Chem., 1960, 52, 825. 61. S. Pragasan, G. A. S. Raj and X . R. Rajkumar, Acta. Cienc. Indica [Ser.] Phys., 1979, 5,28 (Chem. Abstr., 1980,92, 25 447p). 62. G. Kalatzis, J. Konstantinos, E. Vrachnou-Astra and D. Katakis, 1. A m Chem. Soc., 1983, 105, 2897. 63. H. Enomoto, Ph.D. Thesis, University of California, Irvine, 1979 (Chem. Absrr., 1979, 91, 157 OOOb). 64. M. Shirom and G. Stein, Nature (London), 1964, 204, 778. 65. A. L. Poznyak and G. A. Shagisultanova, Dokl. Akad. Nauk SSSR, 1967,173,612 (Chem. Abstr., 1967,67,16 713d). 66. N. Kelso King and M. E. Winfield, J. Am. Chem. SOC.,1958, 80, 2060. 67. A. V. Gogolev, I. A. Potapov, M. E. Rozenkevich and Yu. A. Sakharovskii, Kznel. Catal. (Engl. TransL), 1980,21,296. 68. A. V. Gogolev, I. A. Potapov, M. B. Rozenkevich and Yu. A. Sakharovskii, Kine!. Karol., 1981, 22, 237. 69. R. E. Hintze and P. C. Ford, J. Am. Chem. SOC.,1975,97, 2664. 70. P. K. Eidem, A. W. Maverick and H. B. Gray, Inorg. Chim. Acta, 1981, 50, 59. 71. D. D. Davis, G. K. King, K. L. Stevenson, E. R. Birnham and J. H. Hageman, J. Solid State Chem., 1977, 22, 63. 72. K. L. Stevenson, D. M. Kaehr, D. D. Davis and C. R. Davis, Inorg. Uhem., 1980, 19, 781. 73. G. Ferraudi, Inorg. Chem., 1978, 17, 1370. 74. R. F. Jones and D. J. Cole-Hamilton, J. Chem. SOC.,Chem Commun, 1981, 58. 75. R. F. Jones and D. J. Cole-Hamilton, J. Chem. SOC.,Chem. Commun., 1981, 1245. 76. D. J. Cole-Hamilton, K. F. Jones, 3. R. Fisher and D. W. Bruce, in ’Photogeneration of Hydrogen’, ed. A. Harriman and M. A. West, Academic, London, 1982, p. 105. 77. A. Bino and M. Ardon, J. Am. Chem. SOC.,1977, 99, 6446. 78. W. C. Trogler, G. L. Geoffroy, D. K. Erwin and H. B. Gray, J. A m Chem. SOC.,1978, 100, 1160. 79. H. B. Gray and A. W. Maverick, Science, 1981, 214, 1201. 80. S: Miller and A. Hain, J. Am. Chem. SOC.,1983, 105, 5624. 81. D. K. Erwin, G. L. Geoffroy, H. B. Gray, G. S. Hammond, E. I. Solomon, W. C. Trogler and A. A. Zagan, J. Am. Chem. SOC.,1977, 99, 3620. 82. K. R. Mann, N. S. Lewis, V. M. Miskowski, D. K. Erwin, G. S. Hammond and H. B. Gray, J. Am. Chem. SOL, 1977, 99, 5525. 83. Y.Ohtani, Y. Yamamoto and H. Yamazaki, Inorg. Chim. Acta, 1981, 53, 481. 84. H. B. Gray, V. M.Miskowski, S. J. Milder, T. P. Smith, A. W. Maverick, J . D. Buhr, W. L. Gladfelter, I. S. Sigal and K. R. Mann, in ‘Fundamental Research in Homogeneous Catalysis 3’, ed. M. Tsutsui, Plenum, New York, 1979, 819. 85. W. M. Miskowski, I. S. Sigal, K. R. Mann, H. B. Gray, S. J. Milder, G. S. Hammond and P. R. Ryason, J. Am. Chem. SOC.,1979, 101, 4383. 86. E. N. Savinov, S. S. Saidkhanov, V. N. Parmon and K. Zamaraev, React. Kinet. Catul. Lett., 1981, 17, 407. 87. T. Yamase, Inorg. Chim Acta, 1982, 64, L155. 88. T. Yamase, Inorg. Chim. Acta, 1983,76, L25. 89. R. Battaglia, R. Henning, B. Dinh-Ngoc, W. Schlamdnn and H. Kisch, J Mol. Catal., 1983, 21, 239. 90. S. Ydmagida, T. Azuma and H. Sakurai, Chem. Lett., 1982, 1069. 91. J. Buchelev, N. Zeug and H. Kisch, Angew. Chem., I n f . Ed. Engl. 1982, 21, 783. 92. D. W. Bruce and D. J. Cole-Hamilton, unpublished observations. 93. H.B. Gray, V. M. Miskowski and A. Gupta, Roc. DOE ChemlHydrogen Energy Syst. Contract Rev., 1978 (Publ. 1979) (CONF 7&1142),171 (Chem. Abstr., 1980, 93, 242 533p). 94. E. Borgarelfo, K. Kalyanasundaram, Y. Okuno and M. Gratzel, Helu. Chim.Acta, 1981, 64, 1937.

32. 33. 34. 35.

536

Uses in Synthesis and Catalysis

95. A. Harriman, private communication. 96. E. N. Savinov, S . S. Saidkhanov and V. N. Parmon, Kinel, catal. (Engl. Transl.), 1983,24, 55. 97. S. S. Saidkhanov, A. I. Kokorin, E. N. Savinov, A. I. Vokov and V. N. Parmon, J. Mol. CataZ., 1983, 21, 365. 98. I. Willner. W. E. Ford. J. W. Otvos and M. Calvin. Bioelectrochemistw. .. 1979, 55. 99. R. Mitzner and W. Depkat, 2.Phys. Chem. (Leipzig), 1974, 255, 861. 100. D. A. House and R. P. Moon, J. Inorg. NucL Chem., 1981,43, 2572. ;. 101. Y. Otsuji, K. Sawada, I. Morishita, Y. Taniguchi and K. Mitzuno, Chem. Lett., 1977, 983. 102. F. Ashmawy, C. A. McAuliffe, R. V. Parish and J. Tames, J. Chem. SOC.,Chem. Commun. 1984, 14. 103. J. E. Earley and H. Razavi, Inorg. Nucl. Chem. Let!., 1973, 9, 331. 104. E. Baur and A. Rebmann, Helv. Chim. Acta, 1921, 4, 256. 105. C . Creutz and N. Sutin, Roc. Natl. Acad. Sci. USA, 197572, 2858. 106. G. Sprintschnik, H. W. Sprintschnik, P. P. Kirsch and D. G . Whitten, J. Am. Chem. SOC., 976, 98, 2337. 107. G. Sprintschnik, H. W. Sprintschnik, P. P. Kirsch and D. G.Whitten, 1.Am. Chem Soc., 977, 99, 4947. 108. A. Harriman, J. Chem. Soc., Chem. Commun., 1977,778. 109. G. L. Gaines and S. J. Valenty, J, Am. Chern. SOL, 1477,99, 1285. 110. L. J. Yellowlees, R. G. Dickinson, C. S. Halliday, J. S. Bonham and L. E. Lyons, Aust. J. Chem, 1978,31, 431. 111. J. M. Kelly and J. G. Vos, Angew. Chem., Inr. Ed. Engl., 1982, 21, 628. 112. D. J. Cole-Hamilton, J. Chem. Soc., Chem Commun, 1980, 1213. 113. D. Choudhury and D. J. Cole-Hamilton, J. Chem. SOC., Dalton Trans., 1982, 1885. 114. R. F. Jones and D. J. Cole-Hamilton, Inorg. Chim. Acta, 1981, 53, L3. 115. J. M.Clear, J. M. Kelly, C. M. O’Connell, J. G. Vos, C. J. Cardin, S. R. Costa and A. J. Edwards, J. Chem. SOC., Chem. Commun, 1980, 750. 116. K. Kalyanasundaram, Coord. Chem. Rev., 1982, 46, 159 and refs. therein. 117. M.-C. Richoux, in ‘Photogeneration of Hydrogen’, ed. A. Harriman and M. A. West, Academic, London, 1982, p. 39. 118. J. R. Darwent, P. Douglas, A. Harriman, G. Porter and M.-C. Richoux, Coord. Cbem. Rev., 1982,44,83 and refs. therein. 119. A. Harriman and A. Mills, J. Chem. SOC,,Faraday Trans. 2, 1981,77, 2111. 120. E. Amouyal and B. Zidler, Isr. J. Cbem., 1982, 22, 117. 121. G . L. Gaines, J. Phys. Chem., 1979,83, 3088. 122. M. A. J. Rogers and 3 . C . Becker, J. Phys. Chern., 1980, 84,2762. 123. K. Kalyanasundaram and M. Neumann-Spallart, Chem. Pbys. Leff.,1982, 88, 7. 124. K. Kalyanasundaram, J. Kiwi and M. Gratzel, Helv. Chim Acta, 1978, 61, 2720. 125. A. Moradpour, E. Amouyal, P. Keller and H. Kagan, Nouv. J. Chim., 1978, 2, 547. 126. P. Keller, A. Moradpour, E. Amouyal and H. Kagan, Nouv. J. Chim., 1980,4, 377. 127. M. Gohn and N.Getoff, 2. Naturforsch. Teil A, 1979, 34, 1135. 128. P. Keller, A. Moradpour, E. Amouyal and H. Kagan, J. MOL Catal., 1980, 7, 539. 129. P. Keller and A. Moradpour, J. Am. Chem. SOC.,1980, 102, 7193. 130. 0. Johansen, A. Launikonis, J. W. Loder, A. W. H.Mau, W. H. F. Sasse, J. D. Swift and D. Wells, Aust. J. Cbem., 1981, 34, 981. 131. 0. Johansen, A. Launikonis, J. W. Loder, A. W.H. Mau, W. H. F. Sasse, J. D. Swift and D. Wells, Ausr. J. Chem., 1981,34, 2347, 132. See, for example, P. N. Rylander, ‘Catalytic Hydrogenation over Platinum Metals’, Academic, New York, 1967. 133. M. W. W. Adams, K. K. Rao and D. 0. Hall, Photochem. Photobiophys., 1979, 1, 33. 134. 1. Okura and N. Kim-Thuan, J. Chem. Res. ( S ) , 1979, 334. 135. A. I. Krasna, Photochem., Photobiol., 1980,31, 75. 136. E. Amouyal, P. Keller and A. Moradpour, J. Chem Soc., Chem. Commun.,1980, 1019. 137. A. Demorfier, M. DeBacher and G. Lepoutre, Nouv. J. Chim., 1983,421. 138. P.-A. Brugger, P. Cuendet and M. Gratzel, J. Am. Chem SOC.,1981, 103, 2923. 139. P.-A. Brugger and M. Gratzel, J. Am. Chem. SOC.,1980, 102, 2461. 140. A. Launikonis, J. W. Loder, A. M. H. Mau, W. H. F. Sasse and D. Wells, Isr. J. Chem., 1982, 22, 158. 141. Y. Degani and I. Willner, J. Am. Chem. S K , 1983, 105, 6228. 142. E. Amouyal, B. Zidler, P. Keller and A. Moradpour, Chem. a y s . Lett., 1980, 74, 314. 143. P. Keller, A. Moradpour, E. Amouyal and B. Zidler, J. MOL Catal., 1981, 12, 261. 144. A. Launikonis, J. W. Loder, A. W. H. Mau, W. J-I. F. Sasse, L. A. Summers and D. Wells, Ausr. J. Cbem., 1982, 35, 1341. 145. Y. Degani and I. Willner, J. Chem. SOC., Chem. Commun., 1983, 710. 146. P. C. Lee, M. S . Matheson and D. Meisel, Isr. J. Chem, 1982, 22, 133. 147. T. Nishijima, T. Nagamura and T. Matsuo, J. Polym. ScL, Polym. Lett. Ed., 1981, 19, 65. 148. T. Ohsako, T. Sakamoto and T. Matsuo, Chem. Lett., 1983, 1675. 149. T. Matsuo, T. Sakamoto, K. Takuma, K. Sakura and T. Ohsako, J. Phys. Chem, 1981,85, 1277. 150. J. L. Stickney, M. S. Soriaga and A. T. Hubbard, J. Electronnnl. Chem. Interfacial Electrochem., 1981, 85, 179. 151. 0. Johansen, A. Launikonis, A. W. H. Mau and W. H. F. Sasse, Aust. J. Chem., 1980, 33, 1643. 152. L. J. Fitzpatrick, H. A. Goodwin, A. Launikonis, k W. H. Mau and W. H. F. Sasse, Ausr. J. Cbem., 1983, 36, 2169. 153. I. Okura and N. Kim-Thuan, J. Chem. SOC.,Faraday Trans. 1, 1981,77, 1411. 154. I. Okura, K. Nakamura and S. Nakamura, J. Mol Catal., 1979,6, 311. 155. M. Kaneko, A. Yamada and Y. Kurimura, Inoi-g, Chim Acta, 1980,45, L73. 156. J.-M. Lehn and J. P. Sauvage, Nouv. J. Chim., 1977, 1, 449. 157. M. Kirch, J.-M. Lehn and J. P. Sauvage, Helv. Chim. Acta, 1979, 62, 1345. 158. S. F. Chan, M. Chou, C. Creutz, T. Matsubara and N. Sutin,J. Am. Chem. Soc,, 1981, 103, 369. 159. V. Houlding, T. Geiger, U. Koelle artd M. Gratzel, J. Chern. Sac., Chem. Commun., 1982, 681. 160. P. A. Lay, A. W. H. Mau, W. H. F. Sasse, I. I. Creaser, L. R. Gahan and A. M. Sargeson, Inorg. Chem., 1983, 22, 2347. 161. M. A. Rampi Scandola, F. Scandola, A. Indelli and V. Balzani, Inorg. Chim. Acta, 1983, 76, L67. 162. T. Okuno and U. Yonemitsu, Chem. Lett., 1980, 959.

537

Decomposition of Water into its Elements

163. G. Giro, G. Casalbare and P. G . DiMarco, Chem. Phys. Lett., 1980, 71, 476. 164. B. S. Brunschwig and N. Sutin, Chem. Phys. Lett., 1981, 77, 63. 165. J. S. Clayton, D. W. Bruce, D. J. Cole-Hamilton and P. Camillen, Inorg. Chim. Acta, 1985, 96, L11. 166. C. Dainty, D. W. Bruce, D. J. Cole-Hamilton and P. Camilleri, 1.Chem. SOC.,Chem. Commun., 1984, 1324. 167. C. Creutz, N. Sutin and B. Brunschwig, J. Am. Chem Soc., 1979, 101, 1297. 168. C. V. Krishnan, C. Creutz, D. Mahajan, H. A. Schwan and N. Sutin, Isr. J. Chem., 1982,22, 98. 169. G. M. Brown, B. S. Brunschwig, C. Creutz, J. F. Endicott and N. Sutin, J. Am. Chem. SOC.,1979, 101, 1298. 170. C. Creutz, Inorg. Chem., 1981,20, 4449. 171. C. V. Krishnan and N. Sutin, J. Am. Chem. SOC.,1981, 103, 2141. 172. J. R. Fisher and D. J. Cole-Hamilton, J. Chem. SOC.,Dalton Trans., 1984, 809. 173. A. Deronzier and T, J. Meyer, Inorg. Chem., 1980, 19, 2912. 174. J. Hawecker, J.-M. Lehn and R. Ziessel, Nouv. J. Chim., 1983, 7, 271. 175. J.-M. Lehn and R. Ziessel, Roc. Natl. Acad. Sci. U S 4 1982, 79, 701. 176. P. J. DeLaive, B. P. Sullivan. T. J. Meyer and D. G . Whitten, J. Am. Chem. Soc., 1979, 101,4007. 177. K. Monserrat, T. K. Foreman, M. Gratzel and D. G. Whitten, J. Am. Chem. Soc., 1981, 103, 6667. 178. R. J. Crutchley and A. B. P. Lever, J. Am. Chem. Soc., 1980, 102, 7128. 179. R. J. Crutchley and A. B. P. Lever, Znorg. Chem., 1982,21, 2276. 180. R. J. Crutchley, N. Kress and A. B. P. Lever, J. Am. Chem. SOC.,1983, 105, 1170. 181. N. Kitamura, Y. Kawanishi and S. Tazuke, Chem. Lett, 1983, 1185. 182. A. B. P. Lever, Adv. Inorg. Chem. Radiochem., 1965, 7, 27. 183. J. W. Buchler, in ‘The Porphyrins’, ed. D. Dolphin, Academic, New York, 1977, vol. 1, chap. 10. 184. K. M. Smith (ed.), ‘Porphyrins and Metalloporphyrins’, Elsevier, New York, 1976. 185. A. Harriman, J. Chem SOC.,Faraday Trans. 2, 1980, 76, 1978. 186. V. N. Parmon. S. V. Lvmar. I. M.Tsvetkov and K. I. Zamaraev. J. Mol. CataL. 1983.21. 353. 187. I. M. Tsvetkov, E. R. Buyanova, S. V. Lymar and V. N. Parmon, React. Kinet. Catal. Lett., 1983, 22, 159. 188. X.-R.Xiao, C.-3.Wang and H. T. Tien, L Mol. Catal, 1984, 23, 9. 189. K. Kalyanasundaram and M. Gratzel, Helu. Chim. Acta, 1980, 63, 478. 190. G. McLendon and D. S . Miller, J. Chem. SOC.,Chem. Commun., 1980, 533. 191. A. Hamman, G. Porter and M.-C. Richoux, J. Chem. SOC., Faraday Trans. 2, 1981,77, 833. 192. A. Harriman and M.-C. Richoux, J. Photochem., 1981, 15, 335. 193. M.-C. Richoux, J. Photochem., 1983, 22, 1. 194. I. Tabushi and A. Yazaki, J. Org. Chem, 1981,46, 1899. 195. I. Tabushi and A. Yazaki, Tetrahedron, 1981, 37, 4185. 196. T. Shimidzu, I. Iyoda, Y. Koide and N. Kanda, Noun J. Chim, 1983,7,21. 197. W. Kruger and J. H. Fuhrhop, Angew. Chem., Int. Ed. Engl., 1982,21, 131. 198. J. H. Fuhrhop, W. Kruger and H. H. David, Liebigs Ann. Chem, 1983, 204. 199. I. Okura, S. Aono, M. Takeuchi and S. Kusunoki, Bulk Chem Soc Jpn., 1982, 55, 3637. 200. I. Okura and N. Kim-Thuan, J. Chem. Soc., Chem. Commun., 1980, 84. 201. D. P. Rillema, J. K. Nagle, L. F. Barringer and T. J. Meyer, J. A m Chem. SOC.,1981, 103, 56. 202. 1. Okura, S, Aono, M. Takeuchi and S. Kusunoki, Noun J. Chim., 1982, 6, 221. 203. A. Harriman, G. Porter and M.-C. Richoux, in ‘Photogeneration of Hydrogen’, ed. A. Harriman and M.A, West, Academic, London, 1982, p. 67. 204. A. Harriman and M.-C. Richoux, J. Photochem., 1980, 14, 253. 205. A. Harriman and M.-C.Richoux, J. Chem. Soc., Faraday Trans. 2, 1980, 76, 1618. 206. T. R. Webb, Report 1980, W82-02527, OWRT-A-OSO-ALA(1); Order No. PB82-157801 (Chem. Abstr., 1982, 97, 45 054q). 207. T. Sakata, in ‘Proceedings of the Symposium on Water and Metal Cations in Biological Systems, 1978’, ed. B. Pullman and K. Yagi (Chem Abstr., 1981, 94, 159 701s). 208. P. Keller, A. Moradpour and E. Amouyal, J. Chem. Soc., Faraday Trans. 1, 1982, 78, 3331. 209. L. D. Rampino and F. F. Nord, J. Am, Chem. SOC.,1941,63, 2765. 210. G. C. Bond, Trans. Faraday Soc., 1956, 52, 1235. 211. A. J. Frank and K. L. Stevenson, Chem Soc., Chem Cornmum. 1981, 593. 212. N. N. Toshima, M. Kuriyarna, Y.Yamada and H. Hirai, Chem. Lcrr., 1981, 793. 213. Y. Chen, Z. Wei, H. Liu and Y. Chen, Cuihua Xuebao, 1981, 2, 194 (Chem. Abstr., 1982,96, 41 540b). 214. M.T. Nenadovic, 0. I. Micic, T. Rajh and D. Savic, J. Photochem., 1983, 21, 35. 215. J.-M. Lehn, J. P. Sauvage and R Zeissel, Noun J. Chim., 1981,5, 291. 216. T. Li, H. Tang, K. Qi, W. Gu, Cuihua Xuebao, 1983,4, 159 (Chem Abrtr., 1983,99, 77 561k). 217. K. Hauffe, Chm.-Zig., 1983, 107, 190 (Chem. Abstr., 1983,99, 73 746b). 218. D. H. W. M. Thewissen, M. Eeuwhorst-Reinten, K. Timner, A. H. A. Tinnernans and A. Mackor, Sol Energy R and D Eur. Cornrnunify Ser. D, 1982, 56 (Chem. Abstr., 1982, 97, 166010f). 219. Y. Okuno, Y. Chiba and 0. Yonemitsu, Chem. Lett., 1983, 893. 220. Z. Shi, H. Tang, X . Zhu, Cuihua Xuebao, 1981,2, 252 (Chem. Abstr., 1982,96, 172011~). 221. I. Okura and S. Kusunoki, Inorg. Chim. Acta, 1981, 54, L249. 222. R. Rafaeloff, Y. Haruvy, J. Binenboym, G. Baruch and L. A. Rajbenbach, J. Mol. Catal., 1983 , 22, 219. 223. A. Harriman, G. Porter and M.-C. Richoux, J. Chem. Soc,Faraday Trans. 2, 1981, 77, 1939. 224. M. Maestri and D. Sandrini, Nouv. J. Chim., 1981, 5, 637. 225. J. Kiwi, in ‘Proceedings of the International Symposium on Polymer Dispersion Properties, 1981’, ed. T. F. Tadvos, Academic, London, p. 245 (Chem. Abstr., 1982,97, 185 3212). 226. J. Kiwi and M. Gratzel, J. Am. Chem. Soc., 1979, 101, 7214. 227. E. Amouyal, D. Grand, A Moradpour and P. Keller, Nuuv. J. Chim., 1982, 6, 241. 228. J. Kiwi, Chern. Phys. Lett., 1981, 83, 594. 229. E. Pelizzetti, M. Visca, E. Borgarello, E. Pramauro and A. Palmas, Chim. Jnd. (Milan), 1981, 63, 805 (Chem. Absrr., 1982, 96, 110 875d). .

I

I

.

538

Uses in Synthesis and Catalysis

K. Hashimoto, T.Kawai and T. Sakata, Nouv. 1. Chim., 1983, 7 , 249. M. S. Matheson, P. C. Lee, D. Meisel and E. Pelizzetti, J. Fhys. Chem., 1983, 87, 394. 0. Evea, Noun J. Chim., 1982, 6, 423. D. S. Miller and G. McLendon, J. Am. Chem. SOC.,1981, 103, 6791. E. Sutcliffe and M. Neumann-Spallart, Helv. Chim. Acta, 1981, 44,2148. M. Gratzel, Faraday Discuss. Chem. SOC.,1980, 70, 359. D. S. Miller and G. McLendon, Znorg. Chem, 1981, 20, 950 and refs. therein. W. J. Albery, P. N. Bartlett and A. J. McMahon, in ‘Photogeneration of Hydrogen’, ed. A. Harriman and M. A. West, Academic, London, 1982, p. 85. 238. 1. Okura and N. Kim-Thuan, J. Mol. Catal., 1979, 5 , 311. 239. I. Okura and N. Kim-Thuan, Chem. Lett., 1980, 1511. 240. I. Okura, S. Nakamura, N. Kim-Thuan and K. I. Nakamura, J. MOL Catal, 1979, 6, 261. 241. S. Oishi and K. Nozaki, J. Mol. CataZ., 1980,9, 231. 242. I. Okura, S. Kusunoki, N. Kim-Thuan and M. Kobayashi, J. Chem. SOC.,Chem. Commun., 1981,56. 243. I. Okura, M. Takeuchi and N. Kim-Thuan, Chem Lett., 1980, 765. 244. I. Okura, N. Kim-Thuan and M. Takeuchi, Angew. Chem., Znt. Ed. Engl., 1982, 21, 434. 245. R. Hilhorst, C. Laane and C. Veeger, Roc. NatL Acad Sci. USA, 1982,79, 3927. 246. I. Okura, S. Aono and S. Kusunoki, Znorg. Chim. Acta, 1982, 65, L215. 247. I. Okura, S. Kusunoki and S. Aono, Znorg. Chem., 1983, 22, 3828. 248. I. Okura and S. Nakamura, J. Mol. Catal., 1980, 9, 125. 249. D. M. Hercules and F. E. Lytle, J. Am. Chem. SOL, 1966, 88, 4744. 250. F. E. Lytle and D. M. Hercules, Photochem. Photobiol., 1971, 13, 123. 251. T. S. Glikman and M.E. Podlinyaeva, Ukr. Khim. Zh. (Russ. Ed.), 1955, 21, 211 (Chem. Abstr., 1956, 50, 1472f). 252. G. Nord and 0. Wernberg, J. Chem. SOC.,Dalton Trans., 1972, 866. 253. F. Blan, Monatsfi. Chem., 1898, 19, 647. 254. W. W. Brandt, F. P. Dwyer and E. C. Gyarfas, Chem. Reu., 1454, 54,959. 255. B. Z. Shakhashiri and G. Gordon, J. Am. Chem. Soc., 1969, 8, 1103. 256. G. Nord and 0. Wernberg, J. Chem. Soc., Dalton TTUnS., 1975, 845. 257., A. E. Harvey and D. L. Manning, J. Am. Chem. Sw., 1952,14, 4744. 258. V. Ya. Shafirovich, A. P. Moravskii, T. S. Dzhabiev and A. E. Shilov, Kinet. Katal., 1977, 18, 509. 259. X. Wong, Y. Song and C. Gu, Taiyangneng Xuebao, 1981,2,337 (Chem. Abstr., 1982,96,78 852f). 260. M. Kaneko, N. Awaya and A. Yamada, Chem. L e f t , 1982, 619. 261. G. Nord, B. Peterson and E. Bjergbakke, J. Am. Chem Soc., 1983, 105, 1913. 262. N. K. Khannanov, A. V. Khramov, A. P. Moravskii and V. A. Shafirovich, Kinet. KataL, 1983, 24, 858. 263. N. K. Khannanov and V. Ya. Shafirovich, Kinet. Xatal, 1981, 22, 248. 264. G. L. Elizarova, L. G. Matvienko, N. V. Lozhkina and V. N. Parmon, React. Kinet. CataZ. Lett., 1983, 22, 49. 265. H. Daifuku, K. Aoki, K. Tokuda and H. Matsuda, J, Electroanal. Chem. Interfacial Electrochem, 1982, 140, 179. 266. G. L. Elizarova, L. G. Matvienko, V. N. Parmon and K. I. Zamaraev, Dokl. Akad. Nauk SSSR, 1979, 249, 863. 267. V. Ya. Shafirovich and V. V. Strelets, lzu. A k d . Nauk SSSR, Ser. Khim., 1980, 7 (Chem. Abstr., 1980, 92, 118 502t). 268. J. P. Collin, J.-M. Lehn and R. Zeissel, Noun J. Chim., 1982, 6, 405. 269. S. Goswami, A. R Chakravarty and A. Chakravorty, 1 Chem. SOC., Chem. Commun., 1982, 1288. 270. S. W. Gersten, G. J. Samuels and T. J. Meyer, J. Am. Chem Sus, 1982, 104, 4029. 271. V. Ya. Shafirovich and A. E. Shilov, Kinet. Cafal. (Engl. TransL), 1979, 20, 950. 272. V. Ya. Shafirovich, N. K. Khannanov and V. V. Strelets, Nouv. J. Chim., 1980,4, 81. 273. N. K. Khannanov and V. Ya. Shafirovich, Dokl. Akad Nauk SSSR, 1981,260, 1418. 274. B. S. Brunschwig, M. H. Chou, C. Creutz, P. Ghosh and N. Sutin, J. Am. Chem. SOC.,1983, 105, 4832. 275. V. Ya. Shafirovich, N. K. Khannanov and A. E. Shilov, Kinet. Katal., 1978, 19, 1591. 276. T. S. Glikman and I. S. Shchegoleva, Kinef. Kat& 1968,9, 461. 277. G. L. Elizarova, L. G. Matvienko, N. V. Lozhkina, V. N. Parmon and K. I. Zamaraev, React Kinet. Catal. Lett., 1981, 16, 191. 278. G. L. Elizarova, L. C. Matvienko, N. V. Lozhkina, V. E. Maizhish and V. N. Parmon, React. Kinet. Catal. Left., 1981, 16, 285. 279. A. Juris and L. Moggi, Int. J. Sol. Energy, 1983, 1, 273. 280. A. Harriman, G. Porter and P. Walters, J. Chem SOC.,Faraday Trans. 2, 1981, 77, 2373. 281. J.-M. Lehn, J.-P. Sauvage and R. Zeissel, Nouo. J. Chim., 1979,3, 423. 282. H. Nijs, M. 1. Cruz, J. J. Fripiat and H. van Damme, J. Chem. Sac., Chem. Commun., 1981, 1026. 283. H. Nijs, M. I. Cmz, J. J . Fripiat and H. van Damme, Nouv. J. Chim., 1982, 6, 551. 284. I. Okura, N. Kim-Tnuan and M. Takeuchi, Inorg. Chim. Acta, 1981, 53, L149. 285. I. Okura, N. Kim-Thuan and M. Takeuchi, Inorg. Chim. Acta, 1982,57, 257. 286. J. G. Xu and G. B. Porter, Can. J. Chem., 1982, 60,2856. 287. J. Kiwi and M. Gr%tzel,Angew. Chem, Znt. Ed. Engl, 1978, 17, 860. 288. J. Kiwi and M. Gratzel, Angew. Chem., Int. Ed. Engl., 1979, 18, 659. 289. J. Kiwi, J. Chem. Sac., Faraday Trans. 2, 1982, 78, 339. 290. A. Mills and M. L. Zeeman, J. Chem. SOC.,Chem. Commun, 1981,948. 291. J. Kiwi, M. Gratzel and G. Blondeel, J. Chem. Sac., DaZton Trans., 1983, 2215. 292. V. Ya. Shafirovich and V. V. Strelets, Noun 1.Chirn, 1982, 6, 183. 293. C. Minero, E. Lorenzi and E. Pramauro, Inorg. Chim. Acta, 1984, 91, 30. 294. K. Kalyanasundaram, 0. 1. Micic, E. Pramauro and M. Gritzel, Heh. Chim. Acta, 1979,62, 2432. 295. E. Prarnauro and E. Pelizzetti, Inorg. Chim. Acta, 1980, 45, L131. 296. Y.Okuno, 0. Yonemitsu and Y . Chiba, Chem. Lett., 1983, 815. 297. E. Borgdrello, J. Kiwi,. E. Pelizzetti, M. Visca and M. Gratzel, J. Am. Chem. Soc., 1981, 103, 6324. 298. D. H. M. W. Thewissen, M. Eeuwhorst-Reinten, K. Timmer, A. H, A. Tinnemans and A. Mackor, Recl. Trau. Chim. Pays-Bas, 1982, 101, 79.

230. 231. 232. 233. 234. 235. 236. 237.

Decomposition of Water into its Elements

539

299. E. Yesodharan and M. Gratzel, Helu. Chim. Acta, 1983, 66, 2145. 300. E. Borgarello and E. Pelizzetti, Inorg. Chim. Acta, 1984, 91, 295. 301. Nederlandse Centrake Organisatie voor Toegepart - Natuurwetenshappelijk Onderzoek, Neth. Pat. 80 03 898 (1980) (Chem. Abstr., 1982,96, 224018g). 302. R. Humphrey-Baker, J. Lilie and M. Gratzel, J. Am. Chem Soc,1982, 104, 422. 303. G. Blondeel, A. Hamman, G. Porter, D. Urwin and J. Kiwi, J. Phys. Cbem., 1983,87, 2629. 304. P. Baltzer, R. S. Davidson, A. C. Tseung, M. Gratzel and J. Kiwi, J. Am. Chem. SOC.,1984, 106, 1504. 305. A. Harriman and G. Porter, J. Phorochem., 1982, 19, 183. 306. A. Harriman, G. Porter and P. Walters, J. Chem. SOC.,Faraday Trans. I , 1983, 79, 1335. 307. A. Harriman, 2nd international Conference on the Chemistry of the Platinum Group Metals, Edinburgh, July 1984. 308. M. Miller and R. P. Cox, FEBS Lett, 1983, 155, 331. 309. A. Harriman and G. Porter, J. Chern. SOC.,Faraday Trans. 2, 1979,1S, 1543. 310. R. Wohlgemuth, J. W. Otvos and M. Calvin, Proc. NatL Acud. Sci. USA, 1982, 79, 5111. 311. G . Engelsma, A. Y. E. Markham and M. Calvin, J. Phys. Chem, 1962, 66, 2517. 312. L. N. Zavgorodnaya and T. S. Glikman, J. Gen. Chem. USSR (Eng.? Trawl), 1969,39, 1414. 313. T. S, Glikman and 0. V. Zabroda, Biokhimiya, 1969, 34, 302. 314. A. Harriman, G. Porter and A. Wilowska, J. Chem. SDC.,f=anradny Trans. 2, 1983, 79, 807 and refs. therein. 315. A. Harriman and G. Porter, Comm. Eur. Communities [Rep.], EUR 1982, EUR 768 (Chem. Abstr., 1982,97,227 307g). 316. J.-M. Lehn, J. P. Sauvage and R. Zeissel, Nouu. J. Chim., 1980,4, 355. 317. H. Nijs, H. van Damme, F. Bergaya, A. Habti and J. J. Fripiat, J. Mol. Catal., 1983, 21, 223. 318. K. Kalyanasundaram and M. Gratzel, Angew. Chem., Inf. Ed. Engl., 1979, 18, 701. 319. A. Hamman, J. Phorochem., 1984,25, 33. 320. M. Kaneko, N. Takabayashi and A. Yamada, Chem. Lett., 1982, 1647. 321. J. Kiwi, E. Borgarello, E. Pelizzetti, M. Visca and M. Gratzel, Angew. Chem., Ini. Ed. Engl., 1980, 19, 646. 322. L. Milgrom, New Sei., 1984, 26. 323. H. Nijs, J. J. Fripiat and H. van Damme, J. Phys. Chem., 1983,87, 1279. 324. M. Kaneko, N. Takabayashi, Y. Yamauchi and A. Yamada, BuU. Chem SOC.Jpn., 1984,57, 156. 325. I. Willner, C. Laane, J. W. Otvos and M. Calvin, ACS Symp. Ser., 1982, 177, 71. 326. J. Kiwi, K. Kalyanasundaram and M. Gratzel, Siruct. Bonding ( B e d i n ) , 1981, 49, 37. 327. M. Gratzel, Pure Appl. Chem., 1982, 54, 2369. 328. T. Matsuo, Pure Appl. Chem., 1982, 54, 1693. 329. J. K. Thomas, Chem Rev., 1980,80,283. 330. A. J. Frank, in ‘Micellisation, Solubilization and Microemulsions’, ed. K. L. Mittal, Plenum, New York, 1977, vol. 2, p. 549. 331. I. Okura and N. Kim-Thuan, J. Mol. Caral., 1979,6, 227. 332. F. Gao, S. Li and Y. Chen, Taiyangneng Xuebao, 1981, 2, 396 (Chem. Abstr., 1982,96, 107 058j). 333. T. Nagamura, N. Takeyama and T. Matsuo, Chem. Lett., 1983, 1341. 334. S. S. Atik and J. K Thomas, 1.A m Chem. SOC.,1981, 103,4367. 335. I. Willner, J. W. Otvos and M. Calvin, in ‘Proceedings of the International Symposium on Solution Behaviour of Surfactants: Theoretical and Applied Aspects, 1980’, ed. K. L. Mittal and E. J. Fendler, Plenum, New York, 1982, vol. 2, p. 1237 (Chem. Ahsn., 1983,98, 75 306m). 336. J. H. Fendler, J. Phnwchem., 1981, 17, 303. 337. L. Y. Lee, J. K. Hurst, M. Politi, K. Kurihara and 1. H. Fendler, J. A m Chern. SOC.,1983, 105, 370. 338. T. Katagi, T. Yamamura, T. Saito and Y. Sasaki, Chem. Lett., 1981, 503. 339. T. Katagi, T. Yamamura, T. Saito and Y. Sasaki, Chem. Left., 1981, 1451. 340. E. Yablonska and V. Ya. Shafirovich, Nouu. J. Chim., 1984, 8, 117. 341. M. Tunuli and J. H. Fendler, 1. Am. Chem. SOC, 1981, 103, 2507. 342. P. Tundo, K. Kurihara, D. J. Kippenberger, M. Poiliti and J. H. Fendler, Angew. Chem., In?. Ed. Engl., 1982, 21, 81. 343. K. Monserrat and M. Gratzel, J. Chem. SOC.,Chem. Commun., 1981, 183. 344. R. Frank and H. Rau, 2 Nuturforshch. Teil A., 1982, 37, 1253. 345. D. Meisel and M. S. Matheson, J. Am. Chem. SOC.,1977,99,6577. 346. J. Wheeler and J. K. Thomas, J. Phys. Chem., 1982, 86, 4540. 347. W. I. Willner, J. W. Otvos and M. Calvin, J. Am. Chem. SOC.,1981, 103, 3203. 348. W. I. Willner, J.-M. Yang, J. W. Otvos and M. Calvin, J. Fhys. Chem., 1981, 85, 3277. 349. W. 1. Willner and Y.Degani, J. Chem. SOC.,Chem. Comrnun., 1982, 761. 350. W. I. Willner and Y. Degani, Isr. J. Chem., 1982, 22, 163. 351. W. J. Uressick, T.J. Meyer, B. Durham and D. P. Rillema, Inorg. Chem, 1982, 21, 3451. 352. P. Liu, H. Wang, C. Xie and C. S. Cha, Gaodeng, Xuexiao Hueme Xuebao, 1983,4, 131. 353. K. Chandrasekaran and D. G. Whitten, J. Am. Chem. Sac, 1980, 102, 5120. 354. M. Neumann-Spallart and K. Kalyanasundaram, Ber. Bunsenges. Phys. Chem., 1981, 85, 704. 355. M. Neumann-Spallart, K. Kalyanasundaram, C. Gratzel and M.Gratzel, Helu. Chim. Acta, 1980, 63, 1111. 356. M. Neumann-Spallart and K. Kalyanasundaram, J. Chem SOC.,Chem. Commun., 1981, 437. 357. W. J. Dressick, T. J. Meyer and B. Durham, Isr. J. Chem, 1982, 22, 153. 358. M. Gratzel and M. Neumann-Spallart, Fr. Pat. 2 486 705 (Chem Abstr., 1982,97, 9196d). 359. W.-J. Xu and Y.-J. Wu, Taiyangneng Xuebao, 1981, 2, 115 (Chem. Abstr., 1981,95, 153 780r). 360. M. de Backer, M.-C. Richoux, F. Leclercq and G. Lepoutre, Rev. Phys. AppL, 1980, 15, SZY. 361. M. Neumann-Spatlart and K. Kalyanasundaram, J. Phys. Chem., 1982, 86, 2681. 362. G. Lepoutre, M.-C. Richoux and M.de Backer, Comm. Eur. Communities [Rep.], EUR 1981, EUR 7608 (Chem. Abstr., 1982, 97, 26 300a). 363. H.-Y. Tang, W.-2. Li and C.-Y. Yu, TaiyangnengXuebau, 1981, 2, 115 (Chem. Abstr., 1981,95, 153 780r). 364. A. Hamnett, M. P. Dare-Edwards, R. D.Wright, K. R Seddon and I. B. Goodenough, J. Phys. Chem., 1979,83,3280. 365. M. P. Dare-Edwards, J. B. Goodenough, A. Hamnett, K. R Seddon and R. D. Wright, Faraday DISSCUS. Chem. Soc., 1980, 70, 285. CCC6-R’

Uses in Synthesis and Catalysis

540

H.A. Tinnemans and A. Mackor, Recl. Trav. Chim. Pays-Bas, 1981, 100, 295. R. Mernming, E. Schroeppel and U. Bringmann, J. Electroanal. Chem. Interfacial Electrochem, 1979, 100, 307. H. B. Mark, A. Voulgaropoulos and C. A. Meyer, J. Chem Soc., Chem. Commun., 1981, 1021. T. Yamase and T. Ikawa, Inorg. Chim. Acta, 1979,3?, L529. T. Yamase and T. Ikawa, Inorg. Chim. Acta, 1980,45, L55. T. Yamase, Inorg. Chim. Acta, 1981, 54, L165. A. Bhattacharya, J. Basu, K. Das, A. B. Chatterjee, R. G. Bhattacharyya and K. K. Rohatgi-Mukherjee, Adu. Hydrogen. Energy, 1982, 3,771 (Chem. Abstr., 1982,97, 171290b). 373. A. Bhattacharya, J. Basu, R. G. Bhattacharyya, A. B. Chatterjee, and K. K. Rohatgl-Mukhejee, Bull Chem. Soc. Jpn., 1983,57, 939. 374. R. Brdicka, Collect. Czech. Chem. Commun.,1933, 5, 112. 375. A. Calusaru, 3. Electroanal. Chem. Interfacial Electrochem., 1967, 15, 269 and refs. therein. 376. V. F. Toropova, G. K. Budnikov, N. A. Ulakhovich and E. P. Medyantseva, J. Electrnnd Chem Inrefacial Electrochem., 1983, 144, 1 and refs. therein. 377. M. Shinagawa, H. Nezu, H. Sunahara, F. Nakashima, H. Okashita and T. Yarnada, ‘Proceedings of the 2nd International Congress on Advances in Polarography, Cambridge, England, 1959’, 1960, vol. 3, p. 1142 (Chem. Abstr., 1963, 58, 237f and refs. therein). 378. M. Brezina and V. Gultjaj, Collect. Czech. Chem. Commun, 1963, 28, 181. 379. M. Brezina, Collect. Czech. Chem. Commun. 1959, 24, 3046. 380. M. Shinagawa, H. Nezu, A. Muromatsu, and S . Oka, Rev. Polarogr., 1963, 11, 183. 381. P. Zuman, Chem. Listy, 1958, 52, 1349. 382. W. Lambrecht, S . Gudbjarson and H. Katzlmeier, Hoppe-Seyler’s Z. Physiol. Chem., 1960, 322, 52 (Chem Abstr., 1961, 55, 10418a). 383. R. G. Neville, J. Am Chem. Soc., 1957, 79, 518 and refs. therein. 384. A. Calusaru and J. Kuta, Nature (London), 1965, 2M,750. 385. A. Calusaru, Natuwissenschaften, 1966, 53, 475 (Chem Absrr., 1966, 65, 18 158b). 386. A. Calusaru and F. Banica, Elektrokhimiya, 1977, 13, 1417 (Chem. Abstr., 1978, 88,29 569~). 387. J. Pasciak and I. Zjaviony, Rocz. Chem., 1975,49, 315 (Chem. Abstr., 1975,82, 161 791s). 388. G. N. Babkin, L. N. Lebedeva and A. Kh. Iskakova, EZektrokhimiya, 1982,18,287 (Chem. Abstr., l982,96,131984d). 389. M. K. Boreiko, V. V. Shapovalov and D. M. Palade, Ek?ktrokhimiya, 1974,10,1208 (Chem. Absrr., 1975,82,23 666d). 390. C. Feneau and R. Brekpot, Metallurgie (Mons, Belg.), 1969, 9, 115 (Chem. Abstr., 1970,72, 23 727th 391. Ya. I. Tur’yan, Elektrokhirnzya, 1980, 16, 1728 (Chem. Absfr., 1981,94, 56 868y). 392. M. K. Boreiko, 0.L. Zhakova, A. T. Mal’tev and I. S. Kevlich, J. Gen. Chem. USSR (Engl. Trans[), 1980,50,774. 393. M . A. El Guebeley, ‘Proceedings of the 6th Meeting of the International Committee on Electrochemistry, Thermodynamics and Kinetics’, 1955, p. 550 (Chem Absrr., 1956, 50, 10 56%). 394. J. Heyrovsky, Prirodo (Moscow), 1935, 28,212 (Chem. Absrr., 1937,, 31,6116). 395. P. Herasymenko and I. Slendyk, Collect. Czech. Chem Commun., 1933, 5, 479. 396. V. F. Toropova, H. C. Budnikov, V. P. Frolova and N. A. Ulakhovich, J. Anal. Chem. USSR (Engl. Transl.), 1975, 30, 1435. 397. J. Kadlecek, P. Auzenbacher and V. Kalous, J. EleclroanaL Chem. Interfacial Electrochem, 1975, MI, 89. 398. J. R. Fisher, R. G. Compton and D. J. Cole-Hamilton, J. Chern. Soc., Chem. Commun, 1983, 555. 399. V. Vojiv, Collecf. Czech. Chem,. Commun., 1961, 26, 289. 400. V. V. Strelets and V. Ya Shafirovich, Tezisy DokL-Vxs. Soveshch. Polarogr. 7th, 1978, 256 (Chem. Abszr., 1980, 92, 206 009t). 401. V. Ya. Shalirovich and V. V. Strelets, Nouu. J. Chim, 1978, 2. 199. 402. I. M. Reibel, Tr. Kishzneu. Skh. Inst. im. M. V. Fmnze, 1962, 35 (Chem. Abstr., 1963, 59, 9594~). 403. S . Barnett and R G. Charles, J. Electrochem. Soc., 1962, 109, 333. 404. K. Itaya, T. Ataka and S. Toshima, J. Am. Chem SOL, 1982, 104, 3751. 405. D. T. Sawyer and M. E. Bondini, J. Am. Chem. Soc, 1975,97, 6588. 406. S. N. Pobedinskii, A. N. Aleksandrova, A. Trofimenko, K. N. Belonogov and M. I. Al’yanov, Tr. Iwmou. Khim.-Tekhnol. Inst., 1973,16, 31; (Chem. Abstr., 1975,82, 177 119q).

366. 367. 368. 369. 370. 371. 372.

A.

62.1 Coordination Compounds in Biology MARTIN N. HUGHES King's College London, UK 62.1.1 INTRODUCTION 62.1.1.1 Binding Groupsfor Metals in Biology 62.1.1.2 Roles for Metals 62.1.1.3 Analysis 62.1.1.4 Some Geneml Comments

545 5 46 548 549 550

62.1.2 SODIUM A N D POTASSIUM 62.1.2.1 Selective Binding of Sodium and Potassium Cations 62.1.2.2 Transport of Cations 62.1.2.2.1 Some ionophoric antibiotics 62.1.2.2.2 The sodium pump: (Na', K+),-ATPase 62.1.2.2.3 Transport in microbes and other systems 62.1.2.3 Biological Roles for Sodium and Potassium Ions 62.1.2.3.1 AIhIi metal ions as enzyme activators 62.1.2.3.2 Alkali metal ions as structure stabilizers and other functiom

551 551 552 553

62.1.3 MAGNESIUM, MANGANESE A N D CALCIUM 62.1.3.1 Probes for Mg" and Ca2+ 62.1.3.2 Coordination Chemistry and Biological Roles for Mg2+ and Ca2+ 62.1.3.2.1 Binding of Ca'j to proteins 62.1.3.2.2 Ca2' and Mgz+ as structure stabilizers 62.1.3.2.3 Calcium as a regulator of cellular activities 62.1.3.2.4 Magnesium (manganese) and calcium ions as enzyme activators 62.1.3.3 Transport of the IIA Cations 62.1.3.3.1 The (Ca2+, Mg2+)-ATPase of sarcoplasmic reticulum 62.1.3.3.2 Miscellaneous mammalian transport systems for Ca 62.1.3.3.3 Calcium transport in mitochondria 62.1.3.3.4 Transport 01the IIA cations in microbes. Sporulation 62.1.3.3.5 Transport of the IIA cations in plants 62.1.3.4 Calcium-binding Proteins 62.1.3.4.1 Calmodulin (CaM) 62.1.3.4.2 Troponin C 62.1.3.4.3 Parvalbumin 62.1.3.4.4 Myosin light chains 62.1.3.4.5 Calcium-binding intestinal protein 62.1.3.4.6 SlOO proteins 62.1.3.4.7 Proteins with y-carboxyglutamate residues 62.1.3.4.8 Miscellaneous calcium-bindingproteins 62.1.3.S Mg2+ (Mnz+) as Enzyme Activators 62.1.3.5.1 Kinases -some general comments 62.1.3.5.2 Qruvate kinase 62.1.3.5.3 Miscellaneous kinases 62.1.3.5.4 Phosphatases 62.1.3.5.5 Phosphate transfer and biosynthesis 62.1.3.5.6 Nucleotide cyclases 62.1.3.5.7 D N A and R N A polymerases 62.1.3.6 Proteins and Enzymes with a Specific Requirement for Manganese 62.1.3.6.1 Tyrosine-bound manganese: purple manganese acid phosphatase and ribonucleotide reductase 62.1.3.6.2 Concanavalin A and other lectins 62.1.3.6.3 Phosphoenolpyruvatecarboxykinase 62.1.3.6.4 Phosphoglycerate phosphomutase 62.1.3.7 Magnesium and Manganese in Photosynthesis 62.1.3.7.1 Chlorophyll 62.1.3.7.2 n e manganese protein of photosystem II 62.1.3.7.3 Reaction centres of photosynthetic bacteria 62.1.3.8 Calcium as an Enzyme Activator and in Control Ihcesses 62.1.3.8.1 The role of calcium in blood coagulation and in protein C, an anticoagulant 62.1.3.8.2 Calcium as an activator of extracellular enzymes

562 563 563 563 564 565 565 565 5 66 568 568 569 572 572 574 575 576 576 576 577 571 577 578 519 580 580 580 582 583 584 586 587 587 588 588 588 590 590 59 1 592 592 593

'+

541

554 558

559 559 562

542

Biological and Medical Aspects

62.1.3.8.3 Cakium as a trigger 62.1.3.9 Calcifiation and Mobilization 62.1.3.9.1 Uprake of calcium from the intestine 62.1.3.9.2 Deposition of calcium 62.1.4 THE BIOCHEMISTRY OF ZINC

62.1.4.1 Transport of Zinc 62.1.4.2 Carbonic Anhydrase 62.1.4.2.J Structural studies 62.1.4.2.2 Kinetics of the carbonic anhydrase-cutulyzed reversible hydration of CO, 62.1.4.2.3 Meiallocarbonic anhydrases 62.1.4.2.4 Inhibitors for carbonic anhydrase 62.1.4.2.5 Mechanism of action of carbonic anhydrase 62.1.4.3 Carboxypeptidase 62.1.4.3.1 Structural studies 62.1.4.3.2 Metallocorboxypeptidases 62.1.4.3.3 Chemical modification of amino acid residues 62.1.4.3.4 Kinetic studies and the mechanism of action 62.1.4.4 Themolysin 62.1.4.5 Aminopepiidases, Particularly Leucine Aminopeptidase 62.1.4.6 Aspartate Transcarbamylase 62.1.4.7 a-Amylase 62.1.4.8 Gipxdase 1 62.1.4.9 Alcohol Dehydrogenase 62.1.4.9.1 Binding o j coenzyme 62.1.4.9.2 Metal for zinc substitution 62.1.4.9.3 Binding of substrate 62.1.4.9.4 Kinetic studies and mechanism of action 62.1.4.10 A h l i n e Phosphatase 62.1.4JO.l Structural aspects 62.1.4.10.2 Metal substitution in alkaline phosphatase 62.1.4.10.3 Mechanism of action 62.1.4.11 The &Lactamuses 62.1.4.12 Phospholipase C 62.1.4.13 Fmctose-l,&bisphosphatase 62.1.4.14 Zinc and Snake Venoms 62.1.4.15 Some General Conclusions

594 596 596 597 598 599 600 600 601 60 1 602 602 603 603 604 605 605 606 606 606 607 607 608 608 609 609 610 610 611 611 612 612 613 613 613 613

62.1.5 THE BIOCHEMISTRY OF IRON - A SURVEY OF IRON-CONTAINING ACTIVE SITES

614

62.1.5.1 The Iron Porphyrins 62.1.5.1.1 Ihe coordination chemistry of porphyrins 62.1.5.1.2 Solution studies: complex formation 62.1.5.1.3 ?he redox chemistry of iron porphyrins 62.1.5.1.4 Electronic effects in iron prophyrins 62.1.5.2 The Cytochromes 62.1.5.2.1 Classification of cytochromes 62.1.5.2.2 Cytochromes c 62.1.5.2.3 Cytochromes b 62.1.5.2.4 Cytochromes a and d 62.1.5.2.5 Cybchrome interactions in the respiratory chain 62.1.5.3 Chlorins and Isobncteriochlorins 62.1.5.4 The Iron-Sulfur Proteins 62.1.5.4.1 Rubredoxin 62.134.2 [ZFe-ZS] proteins 62.1.5.4.3 [4Fe-4S] proteins 62.1.5.4.4 Three-ironproteins 62.1.5.4.5 Characterization of iron-sulfur clusters 62.1.5.4.6 Enzymes having iron-sulfur clusters 62.1.5.5 Binuclear, Oxo-bridged Iron Centres and the Iron-tyrosinate Proteins 62.1.5.5.1 Ribonucleotide reductase 62.1.5.5.2 Purple acid phosphatase 62.1.5.5.3 Other binuclear oxo-bridged iron centres

614 614 616 616 616 618 619 619 623 624 624 625 626 626 627 629 63 1 633 634 634 634 636 636

62.1.6 THE BIOCHEMISTRY OF COBALT

62.1.6.1 Some General Considerations on B,, Coenzymes and Model Compounds 62.1.6.1.1 Organocobalt compounds:formation and cleuuage of the Co-C bond 62.1.6.1.2 Some reactions of cobalamins and cobinamides 62.1.6.2 Adenosylcobalamin as a Coenzyme 62.1.6.2.1 Ribonucleotide reductase 62.1.6.3 Methylcobnlamin as a Cofactor

637 637 638 639 640 642 642

Coordination Compounds in Biology

543

62.1.7 THE BIOCHEMISTRY OF NICKEL 62.1.7.1 Nickel in Urease 62.1.7.2 Nickel in Merhonogenic Bucteria. Factor F430 62.1.7.3 Nickel in Carbon Monoxide Dehydrogenase from Acefogenic Bacteria 62.1.7.4 Nickel in Hydrogenase 62.1.7.5 Uptake of Nickel by Pianis

643 643 644 645 646 648

62.1.8 THE BIOCHEMISTRY OF COPPER 62.1.8.1 Types of Copper Centres 62.1.8.1.1 Type 1 Copper 62.1.8.1.2 Type 2 copper 62.1.8.1.3 Type 3 capper 62.1.8.1.4 Blue copper proteins 62.1.8.2 The Blue Electron-fransferProteins 62.1.8.2.1 Plasrocyanins 62.1.8.2.2 Azurin, stellacyanin and other type 1 sites 62.1.8.2.3 Specircrscopic studies on blue electron-transfer proteins 62.1.8.2.4 Electron-transfer reactions with inorganic complexes 62.1.8.2.5 Electron-transfer reactions with proteins 62.1.8.2.6 Model studiesjor the type 1 site 62.1.8.3 The Type 3 Binuclear Copper Centre 62.1.8.3.1 Models for binuclear copper centres 62.1.8.4 The Type 2 Copper Centre 62.1.8.5 Ceruloplasmin 62.1.8.5.1 Oxidase actiuity of ceruloplasmin 62,1.8.5.2 Structural nspects of ceruloplasmin 62.1.8.5.3 Mechanism of action of ceruloplasmin 62.1.8.6 Copper and the Eihylene E$ea in Plants

648 648 648 649 649 649 649 649 651 651 652 653 653 654 654 655 656 656 656 656 656

62.1.9 THE BTOCHEMTSTRY OF MOLYBDENUM 62.1.9.1 Molybdenum in Clusters-Copper Antagonism and Niirogenase 62.1.9.2 Molybdopterin -the Molybdenum Cofactor 62.1.9.3 ESR Spectra of Molybdenum Species 62.1.9.4 The Molybdenum Hydroxylases 62.1.9.4.1 Xanthine oxidase, xanthine dehydrogenase and aldehyde oxidase 62.1.9.4.2 Carbon monoxide oxidase, nicotinic acid hydroxylase, formate dehydrogenase and other hydroxylases 62.1.9.5 Sulfite Oxidase 62.1.9.6 Assimilatory and Dissimilatory Nitrate Reductases 62.1.9.7 Some General Commenls on the Role of Molybdenum

656

657 657 658 658 658 662 663 663 664

62.1.10 VANADIUM, CHROMIUM AND OTHER ELEMENTS 62.1.10.1 Vanadium 62.1.10.1.1 Physiologicnl effects of oanadate(V) -competition with phosphaie 62.1.10.1.2 Inhibition of enzymes by vanadyl ion 62.1.10.1.3 Vanadium as an irrsulin mimic 62.1.10.1.4 Accumulaiion of uanadium by sea squirts 62.1.10.1.5 Activation of enzymes by vanadium 62.1.10.2 Chromium 62.1.10.2.1 The glucose-tolerance factor 62.1.10.3 Tungsten

665

62.1.11 THE TRANSPORT AND STORAGE OF TRANSITION METALS AND ZINC 62.1.11.1 Transport and Storage of Iron in Mammalian System 62.1.1 1.1.1 Ferritin 62.1.1 1.1.2 Transferrin 62.1.1 1.1.3 Phosvitin 62.1.11.1.4 The transport of iron 62.1.11.2 Transport and Storage of Other Transition Metals and Zinc in Mommafian Sysiems 62.1.11.2.1 Transport of copper 62.1.11.2.2 Transport of zinc 62.1.11.2.3 Transport of nickel, manganese, cobalt and vanadium 62.1.1 1.2.4 Storage of tmnsirwn metals and zinc. Metallothionein 62.1.1 1.3 Transport of Iron in Microorganisms 62.1.1 1.3.1 Multiple pathways for iron transport 62.1.11.3.2 The catecholate siderophores 62.1.11.3.3 The hydroxamate siderophores 62.1.11.3.4 Pseudobactin 62.1.11.3.5 f i e binding of Fe"-siderophore complexes to receptors and the release of iron into the cyfoplasm 62.1.11.3.6 Microbial iron fransport and infections in animols and p r a m 62.1.11.4 The Storage Of Iron in Microorganisms 62.1.11.4.1 Magnerotactic microorganisms

667 667 667 669 670 67 1 67 1 67 1 672 612 612 67 3 615 67 5 676 678 678

665 665 665 665 666

666 666 666 667

619 619 680

544

BiologicaI and Medical Aspects

62.1.11.5 Transport and Storage of iron in Plants 62.1.1 1.6 Transport and Storage of Transition Metals other than Iron in Microorganisms 62.1.11.6.1 The uptake of molybdenum 62.1.1 1.6.2 The storage of copper. Microbial metallothioneins 62.1.12 DIOXYGEN IN BIOLOGY 62.1.12.1 The Chemistry of Dioxygen 62.1.12.2 Multimetal Centres and Concerted Electron Transfer 62.1.12.3 Transport and Storage of Dioxygen 62.1.12.3.1 Model compounds for heme dioxygen corricrs 62.1.12.3.2 Hemoglobin 62.1.12.3.3 The cooperative effect in hemoglobin 62.1.12.3.4 Monomeric and dimeric hemoglobins 62.1.12.3.5 Giant hemoglobins 62.1.123.6 The reversible binding of dioxygen by hemoproteins 62.1.12.3.7 Hemerythrin 62.1 .I 2.3.8 Hemocyanin 62.1.12.4 Cytochrome Oxidase 62.1.12.4.1 Structure and organization 62.1 J2.4.2 Reaction mechanism 62.1.12.4.3 Electron-transferpathways in cytochrome oxidase 62.1.12.5 Bacterial Cytochrome Oxidases 62.1.12.5.1 Cytochrome oxidases of the aa3 type 62.1.12.5.2 Cytochrome o 62.1.12.5.3 Cytochrome d 62.1.12.5.4 Cytochrome cd, 62.1.12.5.5 Terminal oxidases in the aerobic carboxydobacteria 62.1.12.6 Blue Copper Oxidases 62.1.12.7 Non-blue Copper Oxidases 62.1.12.8 The Superoxide Dismutases 62.1.12.8.1 The copper-zinc superoxide dismutase 62.1.12.8.2 The superoxide dismutases with manganese or iron 62.1.12.9 Peroxidases and Catalases 62.1.12.9.1 Peroxidase 62.1.12.9.2 Cytochrome c peroxidase 62.1.12.9.3 Chloroperoxidase 62.1.12.9.4 Myeloperoxidase 62.1.12.9.5 Glutnthione peroxidase 62.1.1 2.9.6 Catalase 62.1.12.11) Dioxygenases 62,1.IZ.10.1 Nan-heme dioxygenases 62.1.12.10.2 Heme dioxygenases 62.1.12.1 1 Monooxygenases 62.1.12.11.1 Cytochrome P-450 62.1.12.1 1.2 Tyrosinase 62.1.12.11.3 Dopamine P-hydroxylase 62.1.12.1 1.4 Putidamonoxin

680 680 68 1 681 68 1 682 683 683 684 685 687 688 689 689 689 69 1 692 693 694 696 696 697 697 697 698 698 699 700 700 701 703 703 703 705 705 705 706 706 706 706 708 709 709 711 711 711

62.1.13 ELECTRON-TRANSFER REACTIONS 62.1.13.1 Types of Electron-transfer Reaction 62.1.13.2 Requirements for Fast Elecfron Transfer 62.I. 13.3 Intramolecular, Long-distance Electron Transfer 62.1.13.4 Electron-transfer Chains 62.1.13.4.1 The coupling of electron transfer to the Jynthesis of ATP 62.1.13.4.2 Electron transfer in mitochondria 62.1.13.4.3 Respiratory chains in Escherichia coli

711 712 712 713 713 714 714 715

62.1.14 THE NITROGEN CYCLE 62.1.14. 1 Nitrogen Fixation 62.1.14.1.1 The binding and reactivity of dinitrogen in transition metal complexe 62.1.14.1.2 Nitrogenase 62.1.14.1.3 The iron protein 62.1.14.1.4 f i e molybdenum-iron protein. The P clusters 62.1.14.1.5 The iron-molybdenum cofactor 62.1.14.1.6 Strucfure of the iron-molybdenum cofactor 62.1,14.I . 7 The reactivity and mechanism of nitrogenase 62.1.14.1.8 The sensitivity of nitrogenase to dioxygen 62.1.14.2 The Reduction of Nitrate 62.1.14.3 3% Reduction of Nitrite 62.1.14.3.1 Assimilatory nitrite reductases-reduction to ammonia 62.1.14.3.2 The chemical pathway in denitrification 62.1.14.3.3 The biochemistry of denitrification

717 718 718 719 720 720 721 721 722 725 725 725 725 126 727

Coordination Compounds in Biology

545

62.1.14.4 Nitrijcation 62.1.14.4.1 Oxidation of ammonia by Nitrosomonas europaea

727 727

62.1.15 TRANSITION METAL IONS AND ANTIBIOTICS 62.1.15.1 Bleomycin 62.1.15.1.1 Structural siudies on bleomycin and its transition metal complexes 62.1.15.1.2 Mechanistic aspects 62.1.15.2 Adriamycin 62.1.15.3 Bacitrucin

728 728 728 729 729 729

62.1.16 REFERENCES

730

62.1.1 INTRODUCTION

In this chapter it is hoped to review in general terms the major subject of metals in biology, and to draw attention to appropriate further reading for the expansion of particular topics. The essential role of metal ions in a wide range of biological processes is now well known and hardly requires elaboration! It should be noted that important advances have been made in the last two decades, and current studies are serving to demonstrate the breadth of the subject and the exciting prospect that lies ahead. Li

Be

;1 ,“ I Rb Cs

Sr Ba

Sc

Y La

Ti Hf Zr

1

Cr

V Nb Ta

I

Mo

Fe Ru

Mn

I

Tc Re

W

Os

Co

Ni

Rh Ir

Pd Pt

Cu Ag Au

Zn Cd Hg

Vanadium and chromium appear to be essential elements, while this claim has also been made recently for tin (N. F. Cardarelli, ‘Tin as a Vital Nutrient: Implications in Cancer Prophylaxis and other Physiological Processes’, CRC Press, Fort Lauderdale, 1985). For a remnt account of the biochemistry of the trace elements see ‘Biochemistry of the Essential Ultratrace Elements’, ed. E. Frieden, Plenum, New York, 1984

Figure 1 Metal ions of biological significance

The metal ions of major biological significance are indicated in Figure 1, which shows part of the Periodic Table. Some information on the distribution and concentration levels of these metals in living systems is shown in Table 1. The transition metals and zinc are usually regarded as trace elements, as they are present in very small amounts. Of the transition elements, iron is the most abundant metal, and probably the most well studied. Iron is essential for all living systems with the exception of certain members of the lactic acid bacteria, which grow in environments notoriously low in iron, such as milk. Lactic acid bacteria are devoid of cytochromes, peroxidases Table 1 Metal Levels in Adult Humans Total ( 70 kg a d d )

Metal Na K

Mg Ca V Cr Mn Fe

co Ni cu Zn Mo Sn

(mgf

100 140 20 1000 15 2 12-20 4000b 1.5 10 50- 120 1400-3000 9-10

Whole blood (pg dm-3)

Red cells (bg dm-3)

1960 1700 24 61 Na+> NH,+> Rb+= Cs' > Li' , again through the promotion of the loss of phosphate from the phosphorylated intermediate.14' It has been postulated that IC+ may be involved in antiport with Ca2+ to maintain charge balance, but it is currently thought that this involves protons or a flux of anions. Two moles of Ca2+ appear to bind to the protein per phosphorylation site. Binding of one Ca2+ causes a conformational change followed by binding of Ca2+ to a second site. The use of Ca2+ electrodes or arsenazo I11 indicator shows the ratio Ca2'/ATP = 1.82f 0.13 in the presence of saturating Ca2+. The ratio decreased with increasing pH (with an apparent pK = 7.9), and at temperatures over 30 "C. It was also dependent upon Ca2+. Thus the stoichiometry appears to vary with the physiological conditions due to 'pump lipp page'.'^^ Most mechanisms for the (Ca2+,Mg2+)-ATPaseare modifications of the proposal of deMeis el U L ' ~ *An example is shown in Figure 9. The ATPase can utilize CaATP and MgATP as substrate^.'^^ Two moles of Ca2+are bound per mole of the phosphorylation site with high affinity, followed by phosphorylation of the enzyme by ATP. Conformational changes in the phosphoprotein lead to the calcium sites being accessible to the intravesicular space, with decrease in their affinity for

Coordination Compounds in Biology

567

E

E2P M

2Ca2+E,P M

MZf = Mg2+ or Ca2+ E,P= ADP-sensitive phosphoenzyme E,P = ADP-insensitive phosphoenzyme Figure 9 Schematic representation of the (Ca", Mg2+)-ATPase

Ca2+.148 Calcium is released after conformational transition of the phosphoenzyme from EIP (the ADP-sensitive phosphoenzyme) to E2P (the ADP-insensitive phosphoen~yme).'~' An important contribution has been made by Pickart and Jencks,lS0 who have set up a thermodynamic cycle for catalysis of calcium transport based upon equilibrium constants for individual steps determined under a single set of experimental conditions. This cycle provides the necessary information to allow the evaluation of the way in which the binding energies associated with particular stages allow rapid turnover of the system under physiological conditions. Binding sites for Ca2+ and ATP have been explored by the use of metal probes and nucleotide analogues. The Mn2+ion substitutes for Mg2+but also binds at the Ca2+sites. Such complications have led to the use of I a n t h a n i d e ~ as ' ~ ~probes for the Ca2+ sites. Thus Gd3+ and Tb3+compete with Ca2+for the high affinity site. Luminescence studies with laser-excited Tb3+at the Ca" sites show that two water molecules are present in the first coordination hell.'^' Earlier work'34 with Gd3+ shows that the Ca2+sites are a maximum of 16.1 A apart, and that both sites involve a low level of hydration, consistent with a hydrophobic site. Gd3+has also been used as an ESR probe, and, under certain conditions, evidence has been produced for two forms of an E-Gd3' complex, in accord with current mechanistic views. Several classes of mcleotide analogue have been used recently: (1) thiophosphate derivatives of ATP; ( 2 ) 1 7 0 and '*Oderivatives of ATP; (3)Cr"' or Co"' complexes of ATP; (4)the fluorescent analogue 2,3-0-(2,4,6-trinitrocyclohexyldienylidine)adenosine diph~sphate;'~'and ( 5 ) photoaffinity ana10gues.l~~ Vanadate inhibits hydrolysis of ATP by (Ca2+, Mg2+)-ATPase.As in the case of the (Na+, K')-ATPase, it is thought that vanadate inhibits by binding at the active site and mimicking the trigonal bipyramidal transition state of an S,2 reaction. Occupancy of the phosphate site by vanadate inhibits the binding of Ca2+ at the high affinity sites, even as binding of Ca2+inhibits the binding of P,. This demonstrates the interdependence of the Ca2+ and vanadate/phosphate

site^.'^" Efflux of calcium from the SR is very rapid, the concentration of Ca2+in the sarcoplasm rising 100-fold in milliseconds. This is too fast to be accounted for by a simple reversal of the calcium pump. It has been postuiated that rapid efflux occurs through hydrophilic channels in the SR membrane formed by aggregation of the (Ca2+,Mg2+)-ATPase.As noted earlier, the enzyme exists as oligomers in the membrane. Thus the enzyme is postulated to act as a pump in one direction and as a rapid efflux channel in the other. While this idea is speculative, there is some evidence for passive permeability mechanisms for calcium in SR. Thus, in SR vesicles, accumulated Ca2+ can be relea~ed'~'by caffeine, ATP, increased external Ca2+, low Mg2+, alteration in pH or alteration in the ionic environment that could result in changes of surface membrane charge or a depolarized membrane. In addition it appears that a certain proportion of SR vesicles have enhanced 'calcium-release' properties.'56J57 It is generally thought that physiological release of Ca2+from SR is induced by Ca2+,depolarization of the SR membrane, changes in surface charge and/or a pH gradient.

568

Biological and Medical Aspects

62.1.3.3.2 Miscellaneous mammalian trassporl systems for Ca2+ A variety of other calcium transport systems are associated with Ca2'-activated ATPases. The extraembryonic structure, the chorioallantoic membrane, of the chick embryo is responsible for the translocation of over 120 mg of eggshell calcium into the embryo during development. The enzyme responsible for this is a (Ca2+,MgZC)-ATPasewith K , values for Ca2+ of 30 Fmol dm-3 and 0.3 mmol dm-3, and a molecular weight of 170 000. The enzyme can be crosslinked and co-isolated with a calcium-binding pr~tein.'~' Transport of Ca2+ is also associated with (Ca2+,Mg*+)-ATPases in neutrophil plasma membranes,1s9transverse tubule membranes from rabbit skeletal rnuscle,l6' rabbit myocardial endoplasmic reticuIum,162 sar~ o l e r n m a , brain ' ~ ~ microsomes,164the Golgi apparatus16s and rat liver plasma mernbranes.lh6 The Caz+-pumping ATPase of the red cell membrane differs from the SR calcium pump. It involves a phosphoprotein intermediate, but is a larger protein (MW= 140000) and requires activation by a Ca'+-calmodulin complex in the cytosol. It is probable that similar calcium pumps exist in other plasma membranes including some described above. It seems sensible that Ca2+-calmodulin complexes that mediate intracellular responses to changes in the cytosol calcium concentration should also control activation of the calcium pump that expels excess calcium from the cell. The calcium pump of the red cell membrane appears to resemble other ion pumps, having a site for the hydrolysis of ATP and a site for the binding and translocation of calcium all on one protein. However, there must be some additional feature that results in inhibition, so that it is only active when binding the Ca2+-calmodulin complex. A hint comes from the observation that isolated red cell membranes, which usually require calmodulin for activity of the calcium pump, show full activity after gentle hydrolysis with trypsin. This implies that the trypsin hydrolysis has resulted in the loss of a peptide fragment which is responsible for the inhibition and which also contains the receptor site for the ~ a l r n o d u l i n . ' ~The ~ - ' ~properties ~ of the fragments produced on controlled hydrolysis of the ATPase with trypsin have been examined, and it appears that the fragment postulated to contain the calmodulin binding site does not bind calmodulin after it is split from the main r n ~ l e c u l e . This ' ~ ~ does not invalidate the hypothesis but more work needs to be done. Calcium may also be taken up by Na'-Ca" exchange. Thus sperm plasma membranes take up Ca2+ by an ATP-independent Na+/Ca2+ antiporter. Ejaculated sperm cannot do this, due to the presence of a protein' that binds strongly to the plasma membrane."' A highly active, electrogenic Na+/Ca2+ exchange occurs in sarcolemmal vesicles from cardiac muscle.171

62.1.3.3.3

Calcium transport in mitochondria

Calcium levels are believed to be controlled in part at least by the uptake and release of Ca*+ from mitochondria. 172-174 The capacity of mitochondria for Ca2+seems to be more than sufficient to allow the buffering of Ca2+ at low cytosol levels. Mitochondria take up Ca2+ by an energydependent process either by respiration or ATP hydrolysis. It is now agreed that Ca2+ enters in response to the negative-inside membrane potential developed across the inner membrane of the mitochondrion during respiration. The uptake of Ca2+ is compensated for by extrusion of two H+ from the matrix, and is mediated by a transport protein. Previous suggestions for a Ca2+phosphate syrnport are now discounted. The possible alkalization of the mitochondrial matrix is normally prevented by the influx of H+ coupled to the influx of phosphate on the Hf-H2P04symporter (Figure 10).This explains why uptake of Ca2' is stimulated by phosphate. Other cations can also be taken up by the same mechanism. If respiration is inhibited, mitochondria can still take up Ca2+as long as ATP is available, since hydrolysis of ATP can also generate a transmembrane potential. The presence of phosphate increases the ability of the mitochondria to accumulate Ca2+,partly because of the buffering effect of the phosphate on the pH of the matrix, and partly because of the precipitation of insoluble calcium phosphate within the matrix. The formation of insoluble calcium phosphate lowers the internal free [Ca2+] and favours further influx. The amorphous nature of the precipitate is of interest as the formation of crystalline hydroxyapatite would bc expected. Mitochondria must contain a factor that inhibits this process. The proteins that are responsible for mitochondrial Ca2+ uptake have not been identified with certainty. Various glycoproteins isolated from mitochondria are probably Ca2' receptor proteins rather than Ca2+ transporters. However, a 3000 molecular weight Ca2+-binding protein has been isolated from the inner membrane of calf heart mitochondria which has properties appropriate

Coordination Compounds in Biology

569

OUT

Figure

10 Respiration-dependent uptake of CaZ+by mitochondria

for the Ca2+ carrier. This protein, calciphorin, carries Ca2+ through an organic phase, and is inhibited by La3+ and by ruthenium red, the classical inhibitors of calcium uptake in mit~chondria.’~’ Efflux of calcium from mitochondria is less well understood. It can be induced by uncouplers or by inhibition of electron transport, which results in the reversal of the Ca2+ uptake porter as Ca2+efflux runs down the Ca2+ concentration gradient. However, there are additional pathways, possibly via a Ca2+-2H+ antiport carrier. A chemically diverse range of compounds act as Ca2+-releasing agents, while some intracellular regulatory factors may be physiological regulators (e.g. CAMP, Na+, pyrophosphate, phosphoenolpyruvate, redox state of pyridine nucle~tides).”~ It is possible that these agents act indirectly by affecting the mitochondrial permeability to other substances, for example by the activation of hydrophilic channels in the inner membrane. Calcium can produce a reversible increase in permeability of the mitochondrial inner membrane (the Ca2+-inducedmembrane transition). The rate of induction of this transition is also dependent on a number of factors listed above. It should be noted that Ca2+ efflux differs in mitochondria isolated from various sources.

62.1.3.3.4 Transport of the IIA cations in microbes. Sporulation Some illustrative examples of transport of Mg2’, Mn2+ and Ca2+ in microbes will be given in this section, and particular attention paid to cation uptake during sporulation. A number of transport processes will be described, which illustrate well the value of cation-selective ionophores in the study of cellular physiology. Excellent reviews of earlier work are available. 177-1HU In view of the substantial requirement of bacteria for Mg2+, it is not surprising that selective, active transport processes are available for its uptake.’” The bacterium E. coli has two kinetically

570

Biological and Medical Aspects

distinct transport systems for Mg2", as shown by studies with mutants defective or altered in these transport systems.18' The first, system I, has high affinity for Mg2+, with K,,, = 30 pmol dm-3. It also takes up Mn2+, Ni2+ and Co2+ with lowered activity, although it should be stressed that these cations may well have separate transport processes of high affinity for uptake at lower concentrations. The second, system 11, also has a K , value of about 30 kmol dm-3, and is probably specific for Mg2'. It is repressed by growth in media of high magnesium concentration. Some bacteria require citrate as a charge-compensating cotransported anion. The uptake of Mn2+ or Co2+by system I is inhibitory or lethal respectively, and leads to efflux of Mg2+from the cell. It is likely that accumulated Mn2* or Co2+displaces Mg2+from ribosomes, thus increasing intracellular 'free' magnesium ions, which leave the cell via the Mg2+ transport systems. Both systems for Mg2+ uptake are temperature dependent and are inhibited by uncouplers and respiratory chain inhibitors, showing them to be active processes. Energy-dependent transport systems of high affinity for magnesium have also been identified in fungi and yeasts. Mg2+ moderates the mutagenic and growth-inhibiting actions of Mn2+ on Saccharomyces cereuisiae, probably by competing for the same uptake system. Energy-dependent uptake of Mg2+ in this case requires phosphate for cotransport. In eukaryotic cells there are sometimes compartmentalized deposits of magnesium phosphate. Highly specific, active transport systems for manganese have been found for a wide range of organism^."^ Their high affinity for Mn2+ allows organisms to concentrate Mn2+, even in the presence of much higher levels of calcium and magnesium. Thus the E. coli manganese transport system has a IC, value of 0.2 pmol dm-3 and is unaffected by a 105-fold excess of Mg2+or Ca2+. Both Co2' and Fen+appear to compete for the Mn2+-uptake system, but their K , values are over 100 times greater than the K , value for Mn". The system is effectively specific for Mn2+. In addition, at higher concentrations, cells can acquire Mn2+ because it is a low affinity substrate for the Mgz+transport system. Bacillus subtilis has a citrate-Mn2+ cotransport system, while yeasts have a phosphate-Mn2+ system, as observed for Mg2+. Certain marine bacteria take up Mn2+ and oxidize it to Mn'", which is precipitated as MnO,. These processes are thought to be involved in the formation of ferromanganese nodules, which may contain up to 63% of their weight as Mn02.179Relatively little is known about the mechanism of formation of these nodules, but their potential as a reserve of manganese (and other metals) is enormous. A recent suggestion is that the oxidation of Mn2+ is carried out by the spores of a marine Bacillus rather than the cells. Oxidation of MnZt did not occur in the absence of binding of Mn2 ' to the spores. Oxidation and deposition of manganese only took 'place after Mn2+ was bound to the pore.'",''^ B. subtilis has a highly specific requirement (1 pmol dm-3) for Mn2+ for sporulation. No alternative cation can replace Mn2+, although other cations are also required. Uptake of Mn2+ by B. subtilis is an active, highly specific process, similar to that described for E. coli. Thus, uptake of Mn2+ is unaffected by a 105-fold molar excess of Mg2+ or Ca2+,but has a K , value of about 1 pmol dm -3. The manganese accumulated by non-sporulating cells is exchangeable with extracellular Mn2+. In sporulating cells this was not the case, and after 9 h into sporulation 100% of the Mn2+from media containing low levels of Mn2+was found in the mature spores. Mn2' is essential for the synthesis of certain antibiotics by Bacillus, including gramicidin A, bacitracin and my~obacillin."~ These were once thought to be involved in the transport of Mn2+during sporulation, but these ideas are now discounted. The role of cations in sporulation will be discussed at the end of this section. Microbial cells maintain low internal [Ca2+] levels through the operation of several types of highly active and specific efflux p r ~ c e s s e s . Efflux ' ~ ~ ~of ~ ~Ca2+ ~ can only be studied with difficulty in whole cells, partly because of complications resulting from the binding of Ca2+ to the cell walls. Nevertheless, the presence of active mechanisms for the efflux of calcium are shown by the fact that extreme temperatures or inhibitors result in increased levels of intracellular calcium. Efflux of Ca2+ has been demonstrated for B. megaterium, while ATP-linked efflux of Ca2+ has been demonstrated for Streptococcus f a e c a l i ~ . ' ~ ~ Most resuits on calcium transport have been obtained using cytoplasmic membrane vesicles, which may be prepared in 'inside-out' or 'right-side-out' configurations. Inside-out vesicles may be obtained by the disruption of E. coli cells in a French press, These then accumulate Ca2+ in an energy-dependent fashion, provided ATP or an oxidizable substrate is available. Addition of phosphate enhances the uptake of calcium as calcium phosphate is precipitated inside the 'cell, thus accounting for the lack of exchangeability of the calcium.ls6 The calcium uptake process appears to involve exchange with H+. Thus the ionophore nigericin, which catalyzes an H+/K+ exchange, inhibits uptake of calcium in the presence of potassium.

Coordination Compounds in Biobgy

571

This results in the loss of the H+gradient. The ionophore A23187 stimulates calcium uptake. The Ca2’/H+ ratio appears to be electrogenic, but published values of the ratio are 1:1’87 or 1:>2.IS8 These ion movements are shown schematically in Figure 11. It should be noted that the orientation of the protonmotive force is reversed from that in vivo. The proton-pumping Mg2+-ATPase will be described in a later section. This calcium transport system is not inhibited significantly by ruthenium red, the classical inhibitor for calcium uptake in mitochondria. However, uptake of Ca2+by these inside-out vesicles of E. coli is inhibited drarnati~ally’~~ by a dimeric, mixed-valence complex of RUI’~’I’, [( NH3)3RuC13Ru( NH3)3]2+.The mode of action remains to be established. ADP+ Pi

H+

\ I

A - + +

ATPase

NADH

H+ nig = nigericin

I

Ca”

Ca”

Val =valinomycin

Figure 11 Uptake of Ca’+ by inside-out membrane vesicles of Escherichia coli

Recently, a second transport system for efflux of Ca2+from E. coli has been identified.’” This is a calcium-phosphate symporter, which catalyzes a 1 : 1 cotransport of calcium and phosphate, probably in exchange for protons. Active transport of Ca2+across the plasma membrane of Neurospora crassa involves a CaZ+/H+ exchange,”’ although in Halobacterium halobium movement of Ca2+ is linked to Na+.19’ The primary gradient generated is a proton gradient, but an H+/Na+ antiporter develops an Na+ gradient which is used to extrude Ca2+. Efflux of Ca2+ is not inhibited by proton-conducting uncouplers but is sensitive to monensin, an ionophore that catalyzes an H’/Na+ exchange. Thus, at least four different mechanisms have been established for the exclusion of Ca” by bacterial cells, namely, active efflux of Ca2+uia the Ca2+/Ht antiport (E. coli, Azotobacter uinelmdii and Mycobacterium phki), the Ca2+/Na+antiport ( H . halobium), via the hydrolysis of ATP (S. faecalis), and calcium-phosphate symport (E. coli). Massive amounts of Ca”, up to 4% dry weight, are accumulated during sporulation in members of the genus Bacillus. Other divalent cations, Mn2+ and Zn2+, are also accumulated. The Ca2+ is found associated with dipicolinic acid (pyridine-2,6-dicarboxylicacid). In general, bacterial spores show remarkable physiological properties, including a resistance to heat and a wide range of deleterious agents. They have no detectable metabolism when dormant but are able to germinate within minutes in response to agents such as L-alanine, heat and manganese. Jt is speculated that the calcium dipicolinate complex confers heat resistance to bacterial spores and helps to maintain dormancy. However, certain mutants which have normal heat resistance do not contain dipicolinic acid.”? The precipitation of calcium dipicolinate in the spore C O U ! ~ contribute to the passive uptake of calcium. The uptake of Calf by sporulating cells is of interest as it involves the reversal of the normal distribution of calcium between the intracellular compartment and the external medium. It is not CCC6-S*

Biological and Medical Aspects

572

known whether the uptake of Ca2+involves a reversal of the normal calcium extrusion system or whether it is an independent transport pathway. Calcium must first be translocated across the membrane of the mother cell, and then be accumulated by the forespore. The'outer membrane of the forespore arises from invagination of the mother cell membrane, and therefore has inside-out orientation. It will be able to accumulate Ca2+into the forespore. However, the inner membrane of the forespore will have normal orientation. Transport into the mother cell probably involves a Ca2+/H+antiport working in the reverse direction from normal, to give a concentration of Ca2+in the range 3-9 mmol dmP3in the mother cell cytoplasm. Transport into the forespore can then follow the concentration gradient, being facilitated by the precipitation of the calcium in the forespore as the dipicolinate. Added dipicolinate does not stimulate Ca2+uptake into isolated f ~ r e s p o r e s . ' ~ ~ The distribution of the cations in the spores has been investi ated by the use of high resolution electron probe microanalysis for spores from B. mega~riurn'~'and B. cereus.'96 T h i s shows that nearly all the Ca2+, Mg2+ and Mn2+ is concentrated in the core region, while there is a high concentration of silicon in the cortex/coat layer. ESR spectral9' shows the Mn2+ signal in the former case to be virtually identical at 77 and 298 K, suggesting it is present in a crystalline lattice. Such a lattice would be present in a Ca2+/Mn2+complex of dipicolinic acid. Dormancy of the spore has been attrib~ted'~'to the inactivation of phosphoglycerate phosphomutase by binding of Mn2+. Triggering of germination is brought about by several agents. The first event detected in the triggering of B. megaterium spore germination is the release of Zn2+. This may be liberated from a regulatory site on a protein.'99 Some 25% of the total spore Zn2+pool is released from non-heat-activated spores within four minutes of triggering germination. It appears that the sequence of events overallzouis Zn'+ release, commitment, cortex hydrolysis, pyridine-2,6-dicarboxylicacid and Ca2' release and the onset of metabolism. Electron probe microanalysis also shows that Ca2+ is a prominent constituent of the y-particle in the zoospore of the fungus Blastocladiella emersonii.201Ca2+ also induced zoospores of the fungus Phytophthora cinnamomi to encyst, and subsequently to germinate, while other cations induced encystment only.202There is a high Ca" content in Streptomyces spores, and release of Ca2+ is an early event in spore germinati~n.~'~

62.1.3.3.5 Transport of the IIA cations in plants

'

Most studies in this area have emphasized nutritional aspects, and the effects of nutritional deficiencies. A ran e of studies on uptake of Mn2+ by plant tissues show evidence for active transport systems.' *'05 The enzymes of photosystem 11 have an absolute requirement for MnZ+, as does the NAD-dependent malic enzyme found in the mitochondria of certain plants. Some Mn*+-containing proteins have been suggested to be involved in transport and storage of manganese. Manganin is a peanut seed globulin that contains one Mn2+per protein molecule,2o6 while jack bean concanavalin A also contains manganese. Most growth and developmental processes in plants are associated with movements of calcium.207 Calcium enters the plant through the root, mainly at the tip of the root. Large, mainly immobile concentrations of Ca2+are found in the apoplastic compartment, but the cytosolic concentration of Ca2+ is low, and is similar to that found in animal cells. In general it appears that entry of Ca2+into plant cells is passive, although its efflux is driven by active transport mechanisms. There is much evidence chat the low levels of Ca2+are controlled by plasma membrane Ca"-transIocating ATPases and by Ca2+uptake into mitochondria, vacuoles and possibly other organelles. Uptake of Ca2+by plant mitochondria also involves Ca'+-translocating ATPases, and appears to be similar to that found €or animal mitochondria. It is stimulated by phosphate and inhibited by ruthenium red and lanthanide ions.2Q*Isolated plant mitochondria contain large amounts of calcium, up to 700 nmol mg-' mitochondrial protein. Some microsomal plant membranes have Ca2+-translocating ATPases which are moduiated by caimodulin, as described in Section 62.1.3.3.2 for animal cells.

62.1.3.4

Calcium-binding Proteins

A wide range of calcium-binding proteins have been isolated from intracellular and extracellular I 15 119,21)9,210 They are involved in the storage of calcium, the buffering of czlcium concentrations and the triggering of a wide range of processes involving membrane transport, secretion ~

Coordination Compounds in Biology

573

and enzyme action, and so are associated, for example, with nerve and muscle function. Structural data for some calcium-binding proteins were included in Table 6, while Table 7 presents a more comprehensive list. These proteins have high binding constants for calcium ( 10"-107) and much lower constants for magnesium (about lo3), and show sequence homology in certain cases. There are variations in the calcium-binding sites in accord with the different functions of the proteins. They also bind Ca" at rates close to the diffusion-controlled limit. The trigger proteins have four sites for Ca2", which occur as two pairs of 'hands'. Other proteins have two linked sites, i.e. one pair of hands. These highly structural sites2" are not seen in the calcium-storage proteins.

Table 7 Some Calcium Proteins ?'rotein

Prothrombin Calmodulin Troponin C Myosin light chains SI00 Calcineurin Intestinal CaBP Parvalbumin Calsequestrin Calelectrin Bone proteins Saliva proteins Statherin

Function

Extracellular trigger Intracellular trigger of enzymes and pumps Trigger of contractile systems Trigger in some muscle cells Trigger in nerve cells Phosphoprotein phosphatase Calcium transport Calcium buffer Intracellular store Promotion of membrane aggregation Limits crystal growth (Gla proteins) Protection of teeth (proline rich)

Many of these proteins, notably those involved in trigger functions, undergo substantial conformational changes on binding calcium. In this way the presence of Ca2+ can be used to transmit information to enzyme sites to be activated in the cell. The use of the trigger protein to interact with the enzyme allows this information to be transmitted in a more controlled way than would be the case if Ca2+had interacted directly with the enzyme. The use of triggered confonnational changes to control reactivity is known in biology, and is illustrated well by the cooperative effect in the dioxygen carrier hemoglobin. In order to understand these effects in a detailed way for the calcium-binding proteins, it is necessary to compare the structure of the calcium-free and calcium-bound forms of each protein. The structures of parvalbumin and the intestinal calcium-binding protein (Wasserman protein) are known (see Table 6), but at present there are no crystal structures available for calcium-free proteins. Some structural information has been deduced from changes in the NMR spectrum as calcium is removed from the calcium-bound proteins.2w3210 In the light of the available crystallographic studies, the amino acid sequences and NMR studies, certain general points have been notedzw for parvalbumin, Wasserman protein, calmodulin, troponin C and 5100. (1) Each calcium site is formed from residues in a hand which includes a p-strand. The sites contain backbone carbonyl and side-chain carboxylate. (2) Each binding site is linked to two helices. (3) Each &strand backs on to another strand to form a Ca-P-sheet-Ca unit, which then involves four helices. (4) The helices interact with each other through largely hydrophobic surfaces. ( 5 ) The connections between remote ends of helices are relatively mobile strands, and differ from protein to protein. (6) The four-hand proteins, the calmodulins and troponins, are similar to the sum of their two-handed fragments. Removal of calcium leaves most of the secondary structure unchanged and results in a small change in the twist of the 8-sheet, with associated changes in the relative positions of the four helices and the exposure of new surface residues, together with differences in the link regions between helices. Loss of calcium gives a mobile, less well-structured protein, that is continually affected by change in temperature, and melts at low temperature. Binding of calcium in the trigger reaction results in an adjustment of the &sheet, which is transmitted via rotation of the helices (with the exposure of new hydrophobic residues) to the link region of the protein, which undergoes major conformational change and then binds to the enzymes to be activated. The inhibitor trifluoperazine2I2 functions by binding to the surface of the helices, so hindering the trigger movement. The extent to which these hydrophobic and charged

Biological and Medical Aspects

574

domains are exposed may well explain the selective affinity of bovine brain calmodulin for brain phosphodiesterase, troponin C for troponin I, and Tetrahymena calmodulin for guanylate cyclase. Several examples of calcium-binding proteins are discussed in the following sections. Some will be discussed further in later sections.

62.1.3.4.1

Calmoddin (CaM)

This widespread protein, of molecular weight about 17 000, regulates the activity of a number of calcium-dependent enzymes. 119*213-216 Calmodulin binds four Ca2', apparently with the induction of sequential conformational changes, as shown by studies involving NMR,2'7+"'8 fluorescence,21Y*22* CD22'*222 and Raman spectroscopy.223The Raman studies confirmed that the calcium-induced conformational changes involved the a-helix and P-sheet, while the UV-CD work222showed that the main structural event occurring in the step CaMCa, + CaM-Ca, involves a gain in the a-helical content in the 83-92 sequence to expose a hydrophobic region which would interact with a target enzyme.224It is probable that only CaM.Ca3 and CaM-Ca, interact with targets. There is some uncertainty about whether or not the four sites for calcium are filled sequentially. There seems to be evidence from 43CaNMR for two high affinity sites (constants: 4 x lo6 dm3 mol-') and two low affinity sites (constants: 9 x lo4 dm3 mola view supported by 'I3Cd NMR.227 These results were obtained from solutions of low ionic strength. Other data222suggest that calmodulin has four nearly identical intrinsic binding constants. It may be that the binding of Ca2+ depends upon the presence of counter ions to shield charged groups on the protein. Terbium(II1) has been used as a luminescence probe for the calcium sites. Tb3+is luminescent when bound to a protein near to a tyrosine residue, upon excitation of the latter. No significant increase in Tb3+luminescence could be seen up to ratios of Tb3+/calmodulin=2/1, followed by This suggests that the a 400-fold enhancement of Tb3+luminescence on further addition of Tb3+. higher affinity sites under these conditions lack tyrosine, which identifies them as domains I and 1 1 . ~ ~ ~ , ~ ~ ~ Deactivation of calmodulin-dependent enzymes may involve calmodulin-binding proteins such as c a l c i n e ~ r i n , 2although ~~ calcineurin has been shown to possess phosphatase activity.z3o T h e binding of calmodulin to target enzymes is an important topic. Several calmodnlin-regulated enzymes have been purified: these include phosphorylase kinase, phosphodiesterase, calcineurin, (Ca2+,Mg2+)-ATPase, troponin, fodrin and myosin kinases. Their calmodulin-binding subunits are all different from each other. Nevertheless, these subunits could share a common calmodulinbinding domain. However, while some enzymes bind calmodulin in the absence of CaZf (e.g. phosphorylase kinase, troponin I and bacterial adenylate cyclase), others cannot. This suggests that there is not a common interaction mode. The interaction sites have been investigated by several techniques. Cyclic nucleotide phosphodiesterase and the 3-(2-pyridyldithio)propionylsubstituted calmodulin crosslink23' through the formation of -S-Slinks to give an enzymatically active species in the absence of added Ca.'+ Such crosslinking techniques are useful for identifying specific sites on the enzyme which react with calmodulin. Limited cleavage of calmodulin with trypsin to give large fragments allows these fragments to be tested for their ability to activate calmodulin-regulated enzymes and to bind anticalmodulin drugs. Trypsin digests in EGTA yielded peptides 1-106, 1-90 and 107-148. The fragment 1-106 contains sites I, I1 and 111. Digests carried out in the presence of Ca2' gave peptides 1-77 and 78-148, each of which contains two sites for calcium. The purified fragments contained 9. These are attributed to the ionization of Tyr-248 and the bound water molecule. The role of Glu-270 has been much discussed. As noted earlier, this could involve general base catalysis of attack of water, or direct nucleophilic attack on the carbon of the hydrolyzable peptide link. Some studies on the interaction of anions with cobalt carboxypeptidase A have thrown light on this que~tion."~ At pH> 7, the d-d and MCD spectra of Co" CPA are insensitive to anions, but at pH azurin > plastocyanin. This is based upon the assumption that the larger the spread of the self-exchange rate constant k,, values, the less kinetically accessible is the copper centre. There are many instances for azurin and plastocyanin where limiting kinetic behaviour is observed, and attributed to the formation of an adduct between the protein and the inorganic complex folowed by electron transfer. Values of the association constants and of the electrontransfer rate constants may then be calculated. This situation has not been observed in the case of stellacyanin, which differs from azurin and plastocyanin in that it has an overall positive charge at pH 7 (of +7 in the case of the reduced protein). The electron-transfer rate constants are often associated with fairly large negative values for the entropy of activation (in the range -84 to -210 J K-’mol-’), which are not expected for electron transfer within a compact assembly. 9

Coordination Compounds in Siology

653

The observed second order rate constants, which will be a product of the association constant and the electron-transfer rate constant, will be pH dependent if a protonated form of the protein has a different reactivity from the unprotonated form. The oxidation of plastocyanin by a range of negatively charged complexes decreases with decrease in pH, the dependence of rate constant upon pH leading to a pK, value of about 6 . This may be interpreted simply in terms of the protonation of a residue close to the copper which normally binds the complex, but which is unable to do so when protonated. This residue could well be His-87. Oxidation of plastocyanin is independent of pH in this pH region, suggesting that this oxidizing by [C0(4,7-(0-SO~)~phen)~]~agent binds to a different site on the surface of the protein. This has been confirmed by blocking experiments with non-redox active complexes.932Thus, the oxidation by hexacyanoferrate( 111) is blocked by [Co(C,O,),J'- at a site close to His-87. The oxidation of stellacyanin, surprisingly, is independent of pH over the pH range Quenching of excited-state [Ru(bipy)J3+ by reduced blue proteins involves electron transfer from the Cu' with rate constants close to the diffusion limit for electron-transfer reactions in aqueous solution. It is suggested that the excited Ru" complex binds close to the copper-histidine centre, and that outer-sphere electron transfer occurs from CUI through the imidazole groups to Ru". Estimated electron-transfer distances are about 3.3 h; for plastocyanin and 3.8 8, for azurin, suggesting that the hydrophobic bipy ligands of Ru2* penetrate the residues that isolate the Cu-His unit from the solvent?36 Reduction of rusticyanin by iron(I1) and by chromium(I1) has been The former reaction shows effects due to anions, with limiting zero order kinetics observed in the presence of sulfate.

62.1.8.2.5 Electron-transfer reactions with proteins Many of the reactions of the plastocyanins and azurins with other redox proteins follow Marcus b e h a ~ i o u r . 9These ~ ~ reactions all show a single mechanism of electron transfer, with no kinetic selectivity and no specific interactions between the proteins. The notable exception to this behaviour is cytochrome f (&, where a specific interaction O C C U T S , ~appropriate ~~ for its natural redox partner. Equation (48) represents a probable sequence of electron carriers, although it is difficult to extrapolate conclusions to the membrane-bound proteins. plastoquinone

-+

cytochromef

4

plastocyanin 4 chlorophyll a, (WOO)

(48)

Azurin is involved in the electron-transfer chain to cytochrome oxidase in bacterial respiration. Reduction of oxidized azurin by cytochrome cSs1 is very but the reverse process is complicated through the existence of a slow conformational equilibrium.

62.1.8.2.6 Model studies for the type 1 site Before the elucidation of the structural details of the type 1 site, a range of studies on the blue proteinsY21,927,948-950 and on model corn pound^^^^-^^^ had led to proposals that methionine, cysteine and two histidine residues were ligands for copper in azurin and plastocyanin. The question of the Franck-Condon barrier to electron transfer in small copper complexes has been explored. Electron transfer is slow in copper complexes of some N and 0 donor ligands:56 and the rate is not accelerated by the presence of one or two thioether ligands:57 showing that geometric factors are important in this context. The redox potentials of a range of complexes with varying numbers of sulfur donor atoms imply957that a model complex with two sulfur and two heterocyclic nitrogen donors should have a redox potential of about 600mV, which will be increased by a further SO to 200mV by the low dielectric constant at the active site of a protein, and by a further amount (upwards of 300 mV) by imposition of tetrahedral geometry on the Cu" centre. This leads to the suggestion that the type 1 site should be expected to have redox potentials of about 900mV. The values given in Table 22 show that such high values are found in some cases. It is necessary therefore to explain the low value of stellocyanin (184 mV) rather than the high value of fungal laccase (785mV). Suitable explanations could involve detailed geometric considerations.

654

Biological and Medical Aspects

62.1.8.3 The Type 3 Binuclear Copper Centre The type 3 centre is present in the blue oxidases, and in hemocyanin, tyrosine and dopamine @hydroxylase. Cytochrome oxidase, like laccase, contains four metal centres including an antiferromagnetically linked Fe.. .Cu pair. It appears that there are considerable similarities between the type 3 site and the cytochrome a , . . C u B site in cytochrome oxidase. While this section is concerned with structural aspects of the type 3 site, there will be some overlap with the discussion of the blue oxidases, notably laccase, presented in Section 62.1.12.6. Laccase contains one type I Cu and one type 2 Cu in addition to the type 3 pair. The copper can be reversibly removed from the type 2 site (to give T2D-laccase). Reconstitution may be accomplished by adding CuS04 or Cu' under anaerobic conditions.958Loss of type 2 copper has little effect on the redox potentials of the iype 2 and type 3 copper, or on the electron-transfer reactivity of the type 1 copper. It appears that type 2 Cu is a substrate-binding site in the reduction pathway for the blue copper?59 As noted above, the type 3 site has similarities to the cyt a3...Cu, pair in cytochrome oxidase. Both pairs function as 0,-reducing centres and are ESR-siient due to antiferromagnetic interactions in the oxidized and fully reduced enzymes. ESR signals have now been detected from the type 3 Cu" in laccase intermediates in which the second Cu has been reduced to C U ' . ~ ' "Similarly, ~~~ the ESR-silent CUEin cytochrome oxidase has been detected in an intermediate species. The ESR signals for the type 3 Cu in laccase and the Cug in cytochrome oxidase are very similar. The Cu" in the type 3 pair that becomes detectable on reduction of the other Cur' has been studied by ENDOR techniques? which show that the copper is coordinated by at least three nitrogenous ligands, at least one of which is an imidazole. All three ligands could be imidazoles. The fourth ligand appears to be an aqua or a hydroxo group. This could be a replacement for the ligand that normally bridges the two CU" centres of the type 3 site, and which remains with the Cu' in the uncoupled state. Alternatively, the aqua group could be an exchangeable ligand for O2or 02-, and the bridge is an imidazole group, as known for the Cu/Zn superoxide dismutase, which has remained with the CUI' centre. ENDOR measurements on cytochrome oxidase Cu; confirm the similarity with the type 3 Cu in laccase.

62.1.8.3.1

MudeLF fur binuclear copper centres

Many binuclear copper complexes have been synthesized as models for binuclear sites in enzymes?63 The strong antiferromagnetic interactions present in the natural binuclear copper centres probably result from a superexchange mechanism via bridging ligands. Imidazolate bridges are unable to give strong antiferromagnetic interactions in model ~ y s t e m s . 9Models ~~ involving thioethers have been proposed frequently to account for the high redox potentials of laccase, and also for the characteristic 330 nm band of type 3 c~1.9~~ These models seem less likely to be valid now in view of the results described in the previous section. Macrocyclic ligands or dinucleating ligands have also been used to bring two copper centres into close proximity (3-4 A). Other requirements seem to be the presence of perhaps three nitrogen donors and effective bridging for antiferromagnetic interactions. Some of these ligands are shown in Figure 37. Ligand (1) is an amide ligand that undergoes ionization of the amide group on coordination.966Compound (2)967forms a dicopper complex in which each CUI' ion is bonded to two imino nitrogens of the macrocycle, a thiocyanate group and to two bridging ethoxide groups in an approximately trigonal bipyramidal geometry. The Cu...Cu distance is 3.003 A, and the pairs of Cu" are strongly antiferromagnetically coupled (with 2J in the range -600 to -700 cm-'). The dicopper(I1) complexes undergo reduction by heating their solutions in acetonitrile. These complexes catalyzed the oxidation of substrates such as catechols, hydroquinone and ascorbic acid. Imidazolate-bridged dicopper(11) complexes of ligand (3) involve9'* the CUI' centres being strongly bound in a planar geometry to three nitrogen atoms of the macrocycle and to one nitrogen atom of the bridging irnidazolate group, and weakly bound to axial oxygen ligands (C10,or H 2 0 ) . Magnetic susceptibility measurements confirm antiferromagnetic interaction between the two copper centres with 23 about -50 cm-I. Ligand (4) binds two Cu ions with a bridging phenoxide oxygen, with a pyramidal geometry around each C U I , ' ~and ~ a Cu-.-Cu separation of 3.6-3.7A. This ligand mimics the dioxygen reactivity found in the monooxygenases. The binuclear complex of the triketonate ligand (5) is

Coordination Compounds in Biology

655

I"" "") NH2

H2N

/-Y

Q

h

-,c> ,

(YU-PY PY

PY

(8) Figure 37

Model compounds for the type 3 copper site

diamagnetic at room temperature with -25 = 740 crn-'. The Cu" centres can be reduced sequentially, but the redox potentials (-220 to -410 mV) are too low to mimic type 3 sites.970 The binuclear Cu" complex of ligand (6) shows strong antiferromagnetic interactions with -25 = 588 ~ r n - ' . ' ' ~ , ~Similar '~ ligands have been r e ~ i e w e d . 4Finally, ~~ the binuclear complex of ligand (7), with two bridging azide groups, is completely diamagnetic:73 and ligand (8) acts as a carrier of o ~ . ~ ~ ~

62.1.8.4 The Type 2 Copper Centre

The paramagnetic copper present in the non-blue oxidases such as galactose oxidase and amine oxidases, and also in the blue oxidases, has d - d and ESR spectra typical of coordination complexes of copper(I1). ESR spectra of type 2 Cu suggest the presence of three to four nitrogen ligands, while boiind water has also been implicated by proton relaxation rate measurements. A number of anionic inhibitors bind to type 2 Cu. These results suggest that substrates may bind to the type 2 copper centres in oxidases.

Biological and Medical Aspects

656 62.1.8.5 Ceruloplasmin

This blue oxidase, present in the plasma of vertebrates, appears to be multifunctional.90'~975 It accounts for some 95% of the circulating copper in a normal mammal, and its concentration fluctuates considerably in diseased states. It appears that ceruloplasmin has a major role in copper transport (as discussed in Section 62.1.11). In addition it has oxidase activity towards three groups of substrates, although its physiological role is not known with certainty. 62.1.8.5.1 Oxidase activity of ceruloplasmin

This is directed particularly towards Fe"? the substrate with the highest V,,,and the lowest K,, values. Cerulaplasmin-catalyzed oxidation of FeTT is 10- 100 times faster than the non-enzymatic reaction, and ceruloplasmin appears to be the only effective ferrooxidase in human serum. This is linked to the control of iron mobilization by ceruloplasmin. Iron is released as Fe" from ferritin, and the rate at which it is converted to circulating Fe"',-transferrin is dependent upon ceruloplasmin. This explains why Cu-deficient animals develop anaemia. In addition, ceruloplasmin shows oxidase activity towards bifunctional aromatic amines and phenols, but with much higher K , values, and also to a group of pseudosubstrates that can rapidly oxidize Fe". Ceruloplasmin may also have a role as a serum antioxidant. 62.1~ 5 . 2Structural aspects of ceruloplasmin

There has been much uncertainty over the molecular weight and copper content. It appears to involve a single polypeptide chain of molecular weight about 130 000 with six or seven copper ions. On balance, there are two type 1, one type 2 and four type 3 copper centres. This is in accord with the fact that 44% of the total copper is paramagnetic and ESR-dete~table.~" Similarities with respect to the location of cysteine, histidine and methionine residues in the proteins of azurin, plastocyanin and ceruloplasmin indicate that the type 1 centres in ceruloplasmin are similar to those in the other two proteins. The nine-line superhyperfine splitting in the ESR spectrum of the type 2 Cu has been interpreted in terms of four equivalent nitrogen ligands.978 This was observed in a protein from which the type 1 copper was depleted by dialysis against ascorbate.

62.1.8.5.3

Mechanism of action of cemloplasmin

The several functions of ceruloplasmin cannot be explained at present. It seems reasonable that this diversity is related to the activity of the copper centres. The general pattern of oxidase activity is probably similar to that of the other blue oxidases, with a type 3 binuclear site serving to bind and reduce dioxygen, with electrons transferred from the type 1 site. The type 2 copper may represent a substrate-binding site. 62.1.8.6 Copper and the Ethylene Effect in Plants

Ethylene plays an essential role in the development of plants, for example in germination, growth and fruit ripening. Carbons 3 and 4 of methionine appear to be the most important source of the gas in vivo, but the effect is also induced by externally applied gas. Acetylene and carbon monoxide are competitive with ethylene. This suggests that a metal ion is present at the ethylene receptor site, a view confirmed by the inhibition of alkene binding by dithi~carbarnate.~~' The possibility that this metal is copper9*' is supported by the preparation of copper( I)-monoalkene complexes that show the tight binding of monoalkenes characteristic of the ethylene receptor sites of plants.981

62.1.9 THE BIOCHEMISTRY OF MOLYBDENUM Molybdenum is the only transition element in the second row of the Periodic Table that is known to have a biological role. It is present in at least 15 enzymes, which are listed in Table 23,

657

Coordination Compounds in Biology Table 23 Molybdenum in Enzymes Source

MW

Xanthine oxidase Xanthine dehydrogenase Aldehyde oxidase Sulfite oxidase Nitrate reductase (dissimilatory) Nitrate reductase (assimilatory) Formate dehydrogenase Carbon monoxide oxidase Nicotinic acid hydroxylase Chlorate reductase Biotin sulfoxide reductase Pyridoxal oxidase Tryptophan side-chain oxygenase Nitrogenase

Bovine milk Animals, microbes Hog milk Animals, Thiobacillus nouellus Escherichia coli Chlorella vulgaris E. coli Pseudomonas carboxydovorans Clostridium barkerii h i e u s mirabilis

300 000 250 000 270 OOO 110000 200 000 360 000

Trimethylamine oxidase

E. coli

Enzyme

300 000 300 000

Mo

Other cofactors

2

2 2 2 1 4 2-4 2

"

2

4Fe2S, ,4FAD 4Fe, S,, 4FAD 4Fe, S,, 4FAD 2 heme 4Fe, S, 4 heme, 4FAD FenS,, heme, Se 4Fe2S2,2FAD, Se Fe/S, FAD, Se

Drosophila spp. Pseudomonas spp.

2

(4Fe2S,, 4FAD) Heme

Range of microbes

2

Mo, Fe, S cluster, ironsulfur cluster

and in several uncharacterized proteins. A number of the enzymes in Table 23 are hydroxylases. In many cases the substrate coordinates to the molybdenum, and electron transfer follows. This is in accord with the presence of other electron-transfer centres in the enzymes, which may be iron-sulfur, FAD or heme centres as noted in Table 23. A number of reviews are a ~ a i l a b l e . ~ ' ~ - ~ ~ ' It is interesting to speculate on the chemical properties of molybdenum which make it suitable for its biological function. Obvious features in the chemistry of molybdenum are: (a) a range of oxidation states which can be stabilized in aqueous solution by the common ligands of biology; (b) the formation of oxo compounds and the sulfur analogue; (c) the ability to participate in atom-transfer reactions; and (d) the possibility of higher coordination numbers. 62.1.9.1

Molybdenum in Clusters - Copper Antagonism and Nitrogenase

One of the enzymes given in Table 23 is nitrogenase, which is responsible for the fixation of dinitrogen to give ammonia. Molybdenum probably serves as the binding site for N2, and is present in the iron-molybdenum cofactor, which is a molybdenum-iron sulfide cluster. Nitrogenase will be considered in Section 63.1.14, which deals with the nitrogen cycle. A second example of cluster formation involves the antagonism between molybdenum and copper. The presence of high concentrations of molybdenum in pasture soils is known to lead to symptoms of copper deficiency in animals. This has been attributed to the formation of thiomolybdates in the rumen of grazing animals, which then interfere with the metabolism of copper through the formation of cluster compounds of molybdenum and copper. It has been shown that thiomolybdate, M O S ~readily ~ , forms such clusters on reaction with hosphines and copper(I1) salts under appropriate conditions. Structures are shown in Figure 38.g2,993 Reaction between thiornolybdate and copper compounds in aqueous solution have also been reported.994

\

P

P = phosphine

Figure 38 Some copper-molybdenum dusters

62.1.9.2 Molybdopterin - the Molybdenum Cofactor The enzymes listed in Table 23 (with the exception of nitrogenase) appear to contain a molybdenum-binding cofactor?95 These are interchangeable from enzyme to enzyme, where studied, and are probably identical. The factor is usually assayed by its ability to activate the

Biological and Medical Aspects

458

nitrate reductase from the Neurospora crassa mutant nit-I, which lacks the cofactor. This provides striking evidence in support of a common molybdenum cofactor in these enzymes. The cofactor appears to include a novel ~ t e r i n . ~ ~The - ~ ' properties * of the pterin depend upon the nature of the side-chain in the 4-position. The structure shown in Figure 39 has been proposed997 on the basis that molybdopterin is related to urothione, oxidized to pterin-6-carboxylic acid, and contains in the side-chain two sulfur groups, a double bond, a hydroxyl function and a terminal phosphate group. Two stable fluorescent derivatives of molybdopterin have been chara~terized:~~ which may be of value in view of the extreme instability of the native molybdoprotein when released from the enzyme.

"D)

OH

I H

" M o ( a S

o4

4%

,f

'

SCH2CH2S

2

Ir

I

Figure 39 Molybdopterin and a possible model compound

The apocofactor is synthesized in the absence of molybdenum in E. coli and N. crassa. The cofactor from E. coli is soluble, but is isolated bound to a carrier protein from which it has to be Tungstate competes with molybdate for the molybdenum site in the cofactor, leading either to the formation of a demolybdo species or to an inactive form containing tungsten. The tungsten protein has been characterized in the case of sulfite oxidase,1000which, while inactive, is immunologically identical to the molybdenum cofactor. Figure 39 shows a model ligand for the pterin. This provides an S,02 environment for the molybdenum. In addition the species MoIVOL and Mo203(L)2 have been prepared.'"'

62.1.9.3 ESR Spectra of Molybdenum Species This technique has been of especial value in studying molybdoenzyrnes. Mo"', MoV and Mo'" species have do, d' and d 2 configurations. MoV species are therefore ESR-detectable and are usually easily recognizable, with a slightly anisotropic set of g values in the region of g = 1.96-1.99. For 95M0 and 9 7 M I~=,2 but for other isotopes 1 = 0. The I = 1 isotopes represent about 25% of natural molybdenum, and so the six-fold hyperfine splitting may be difficult to observe. Enrichment of the specimen in molybdenum isotopes with I =; may be necessary. One complication may arise from coupling of the MoV centres, as is well known for dimeric complexes of MoV, with resulting loss of signal. One important feature of MoV intermediate species during enzymatic reaction is the presence of superhyperfine splitting from a proton. MoV model compounds with coordinated N-€3 groups display superhyperfine coupling of the same order of magnitude as that found in molybdenum enzymes.1oo2 62.1.9.4

The Molybdenum Hydroxylases

These enzymes catalyze the two-electron oxidation of purines, aldehydes and pyrimidines, sulfite, formate and nicotinic acid in the genera1 reaction shown in equation (49). These enzymes show some differences in properties. Xanthine oxidase, xanthine dehydrogenase and aldehyde oxidase all have relatively low redox potentials and a unique 'cyanolyzable' sulfur atom, and so will be discussed together. RH+H,O

-+

ROH+2H++2e-

(49)

62.1..9.4.1 Xanthine oxidase, xanthine dehydrogenase and aldehyde oxidase

Xanthine oxidase and xanthine dehydrogenase are two closely related enzymes which differ in the nature of their terminal electron acceptor, the oxidase using O2 and the dehydrogenase using

Coordination Compounds in Biology

659

NADC. They catalyze the hydroxylation of purines. Equation (50) shows the xanthine oxidasecatalyzed oxidation of xanthine to uric acid. Xanthine oxidase is probably the most well-studied molybdenum enzyme. There is good evidence that the molybdenum is the site for binding and reduction of xanthine. The enzyme contains MeV' in the resting form, while MoIV and Mo" are implicated during catalysis.

Aldehyde oxidase catalyzes the oxidation of aldehydes to carboxylic acids by dioxygen, but also catalyzes the hydroxylation of pyrimidines. Despite its rather broad specificity for substrates, it may well be that aldehyde oxidase should be regarded primarily as a pyrimidine hydroxylase. Thus, xanthine oxidase and aldehyde oxidase catalyze the hydroxylation of purines and pyrimidines respectively. The oxygen incorporated into the product comes from water, not 0 2 . The dioxygen serves as the electron acceptor and other oxidizing agents may be used. As normally isolated, xanthine oxidase is a mixture of the active enzyme and a catalytically inactive degradation product, namely the desulfo xanthine oxidase. The deactivation of the enzyme by added cyanide results from the abstraction of this sulfur atom to give thiocyanate. Loss of the sulfur atom to give desulfo xanthine oxidase results in marked changes in the ESR signalstoo3of the molybdenum(V), and in changes in the absorption spectra and redox potential of the Changes in these essential properties of the enzyme on loss of the sulfur atom have led to much interest in the identity of the group which loses the sulfur, and in possible structural and mechanistic implications. Early candidates for the source of the sulfur atom were cysteine1005,1006 and pers~lfide.'~~' Molybdenum complexes with persulfido groups have been prepared, 1008~1009and these will undergo sulfur abstraction reactions with cyanide."" The present view is that an Mo=S group is involved. Xanthine hydrogenase and aldehyde dehydrogenase also give desulfo proteins which have no catalytic activity and lower redox potentials. A molybdenum protein from Desulfouibrio gigas has been isolated, which appears to have similar properties to the desulfo proteins."" Desulfo enzymes can be regenerated by anaerobic treatment with sulfide plus dithionite. Cyanide-inactivated turkey liver xanthine dehydrogenase is also reactivated by selenide, suggesting that selenium may replace sulfur in the active centre. Xanthine oxidase has a molecular weight of 300 000 and is dimeric. Each subunit contains one Mo, one FAD and one each of two types of Fez& cluster. The Mo" ESR signals are split by the iron-sulfur centre I, so allowing the calculation of distances between the two centre^.'^'^-'^'^ The most recent estimate is close to 15 A, with a lower limit of 12 A, a value that holds also for aldehyde X-ray absorption edge and EXAFS measurements are now available for a number of molybdenum enzymes, as shown in Table 24. The coordination environment around molybdenum in the desulfo xanthine oxidase involves two terminal oxo groups (Mo=O = 1.75 A), two sulfur atoms at 2.5 8,and one at about 2.9 A. In the active enzyme, a sulfur atom at about 2.25 A (M=S) replaces one of the terminal oxygen atoms of the desulfo en~yrne.'~'".'~" An independent study'"' has given similar but not identical results for xanthine oxidase. Oxidized xanthine dehydrogenase appears to be similar to xanthine oxidase; oxidized desulfo xanthine dehydrogenase shows the loss of the M=S group and the formation of an additional M=O group."" However, both desulfo and active xanthine dehydrogenase in the reduced form show one M=O group These results support strongly the that a terminal sulfur group is lost from the molybdenum on formation of the desulfo enzyme and replaced by a terminal oxo group. This will account for the difference in pK, value observed for xanthine oxidase and the desulfo enzyme, which has been attributed to the protonation of the sulfur and oxygen atoms respectively (equations 51 and 52). Evidence for such protonation arises from the observation that the ESR spectrum of the Mo" species shows ligand hyperfine splitting, which can be attributed to a proton. The doublets coalesce in DzO showing that the proton is exchangeable,*021while the pH dependence of this phenomenon gives an apparent pK, value of about 8. It appears, however, that the molybdenum(V) interacts with two protons during the catalytic reaction.Io2' Studies with 8-deuteroxanthine show that a proton from the 8-position in xanthine transfers to the enzyme, and is the proton more strongly coupled to the molybdenum. This proton could be transferred to the Mo=S group, as CCC6-v

Biological and Medical Aspects

660

Table 24 X-Ray Absorption Edge and EXAFS Data on Molybdenum Enzymes' Enzyme

Xanthine dehydrogenase Desulfo xanthine dehydrogenase

Donor atoms (bond length,

Redox state

A)

Ref:

Oxidized Reduced Oxidized

1 0 (1.70), IS (2.15), 25 (2.47) 1 0 (1.68), 3 s (2.38) 2 0 (1.67), 25 (2.46)

Reduced

10 (1,661, 2/3S (2.33) 10 (1.73,1 s (2.25), 2s (2.50), 1s (2.90) 2 0 (1.75), 2 s (2.50), IS (2.90) 1 / 2 0 (1.71), 2s (2.54), IS (2.84) 2 0 (1.68), 2/3S (2.41) 1 0 (1.69), 3s (2.38) 2 0 (1.71), 2/3S (2.44)

Xanthine oxidase Desulfo xanthine oxidase Xanthine oxidase Sulfite oxidase

Oxidized Reduced Nitrate reduaase (assimilatory, Oxidized C. vulgaris) Reduced Nitrate reductase (dissimilatory, Oxidized E. coli) Reduced

1 0 (1.67), 3s (2.37) O?

4

1 / 2 0 / N (2.10), 2/3S (2.36), X(2.8)

4

'There may be additional ligands not detected by these techniques. 1. S. P. Cramer, R. Wahl and K. V. Rajagopalan, 1. Am Chem. SOC.,1981, 103,7721. 2. J. Bordas, R. C. Bray, C. D. Garner, S. Gutteridge and S. S . Hasnain, Biochem. J., 1980, 191,499. 3. T. D. Tullius, D. M. Kurtz, S. D. Conradson and K. 0. Hodgson, J. Am. Chem Soc., 1979, 101, 2776. 4. S . P. Cramer, L. P. Solomonson, M. W. W. A d a m and L. E. Mortenson, 3. Am. Chem. SOC.,1984, 106, 1467.

shown in equation (53), which also suggests a possible coordination environment for the rnolybdenurn.

E-Mo=O+H+

= E-Mo-OH IO pKa=

Xanthine oxidase also has a site that will bind anions such as nitrate. This appears to be the substrate-binding site on molybdenum. The action of subtilisin on xanthine dehydrogenase monomer gives a fragment of molecular weight 65 000, which contains the molybdenum and iron-sulfur centres, and from which the flavin is absent. This fragment will not catalyze xanthine-dependent reduction of NADf. The action of various proteases suggests that xanthine dehydrogenase is made up of three domains represented schematically in Figure 40.1022

20 000

I

65 000

Cleavage points are shown by

65 000

+

Figure 40 Proteolytic cleavage of xanthine dehydrogenase

66 1

Coordination Compounds in Biology

The technique of ESR spectroscopy has been of especial value in the study of molybdenumcontaining enzymes. Oxidized xanthine oxidase contains MoV1but a number of ESR signals due to Mo" are developed under various reducing conditions.991These have been classified as 'very rapid' (for reduction with xanthine at high pH), 'rapid' (1 and 2) (for reduction with any substrate), 'inhibited' (for the enzyme treated with aldehydes), 'slow' (for the desulfo enzyme reduced for long time periods) and 'resting 11' (a signal resistant to redox changes, found for the desulfo enzyme treated with ethylene glycol). Only the very rapid and rapid signals are observed during the turnover time of the enzyme and are therefore of catalytic relevance. The interaction with aldehydes gives an MoV signal attributed to the reaction product shown in equation (54). The conversion to this species is not stoichiometric, reflecting the distribution of molybdenum in the oxidation states (VI), (V) and (IV). It appears that there is also an M o ' ~aldehyde species, which is not in rapid equilibrium with the Mov-inhibited species.1o23

It is probable that during enzyme turnover, there is a two-electron abstraction from the substrate to give MoIV and of a proton by a molybdenum ligand; that is an effective abstraction of hydride. The two electrons will then be distributed amongst the reducible centres in the protein to extents appropriate to their relative redox potentials, prior to transfer to O2 or other added oxidizing agent. There will be a transient reoxidation of MoIVto give the observed MoVspecies. The redox potentials for the molybdenum and other redox centres are known.1024-1028 Those for molybdenum are given in Table 25. It may be seen that a wide range is covered for the various enzymes, showing the versatility of the molybdenum centre in enzymes. Table 25 Redox Potentials (E,,,, mV) of Molybdenum Centres in Enzymes Enzyme

MO

Xanthine oxidase Xanthine dehydrogenase Aldehyde oxidase Desulfo xanthine oxidase Desulfifo xanthine dehydrogenase Mo protein from D. gigas Sulfite oxidase Nitrate reductase (C.vulgaris) Nitrate reductase ( E . coli) Formate dehydrogenase

M O ~

-373 -357 -359 -354 -397 -415 t38 -34 +180 -330

M O ~ I M O ~ ~

-374 -397 -351

-386 -433 -530 -163 -54 +220 -470

The evidence for the involvement of MoIV seems clear. Upon reduction of the enzyme with less than stoichiometric amounts of reagent or substrate, an ESR signal appears which is lost when further reductant is added. The tun sten-containing sulfite oxidase1ooocan be reduced to ESR-active Wv with dithionite, but the W$signal is stable in the presence of additional reducing agent. It appears that the tungsten cannot be reduced to W'", so accounting for the inactivity of the enzyme. Excellent evidence for the importance of MolV comes from studies with inhibitors. Allopurinol is a substrate for the enzyme, but its oxidation product alloxanthine (9)is strongly bound at the active site. Removal of excess reagents allow quantitative reoxidation with [Fe(CN),I3-. After allowance for oxidation of the flavin and iron-sulfur centres, it may be shown ~ ~ ~ of this that the molybdenum requires two electrons to be oxidized back to M o ~ ' . ' Because property, allopurinol is used in the control of hyperuricemia in gout.1030More potent inhibitors are folate compounds and 2-amin0-4-hydroxypteridine.'~~~ 8-Bromoxanthine binds at the Mo centre and again seems to involve M O ' ~ . ' O ~ ~ 0

662

Biological and Medical Aspects

The very rapid and rapid signals are suggested to be due to species resulting from the breakdown of an intermediate at high and low pH value~.’’~~ The rapid signal shows the interaction of a proton at the active site. The types 1 and 2 rapid signals are thought to involve different coordination geometries around the molybdenum. Any discussion of the mechanism of xanthine oxidase should attempt to incorporate the special features of xanthine oxidase (and xanthine dehydrogenase and aldehyde oxidase) which are not present, for example, in sulfite oxidase. There are two such features at least: (a) the involvement of two protons rather than the one found for sulfite oxidase, and (b) the presence of the cyanolyzable sulfur atom, The mechanistic features discussed so far involve the abstraction of two electrons and a proton. This means that a carbonium ion is generated, which could undergo attack by a nucleophile. Thus, the presence of a nucleophile at the active site could lead to the formation of a covalent intermediate that will break down to give the products.’032The nucleophile could either be the cyanolyzable sulfur atom or a group associated with the second proton. A possible scheme is shown in Figure 41.

MoV ‘very rapid’

H

11

11

Mo”

MOW

‘rapid’

‘rapid’

Figure 41 The mechanism of action of xanthine oxidase

62.1.9.4.2 Carbon monoxide oxidase, nicotinic acid hydroxylase, formate dehydrogenase and other hydroxylases

Nicotinic acid hydroxylase from Clostridium barkerii catalyzes reaction ( 5 5 ) , the hydroxylation of a pyridine group, and bas similarities to xanthine dehydrogenase. Nicotinic acid hydroxylase is a 300 000 molecular weight flavoprotein containing iron-sulfur and FAD centres, and a molybdopterin cofactor.’035Formate dehydrogenase contains selenium as ~elenocysteine,’~~~ but this does not appear to be the case for nicotinic acid hydroxylase. The possibility that the selenium is incorporated into the molybdopterin cannot be excluded at present. nicotinic acid+ H,O+ NADP+

6-hydroxynicotinic acid+ NADPH+ H+

(55)

Carbon monoxide serves as the sole carbon and energy source for the carboxydo bacteria under aerobic conditions. Using water as the oxygen donor, carbon monoxide oxidase catalyzes the hydroxylation of carbon monoxide, giving carbon dioxide or bicarbonate for assimilation. Most work has been carried out on the enzyme from Pseudomonas carbo~ydovorans.‘~”~’~~~ The activity of carbon monoxide oxidase is considerably stimulated upon anaerobic treatment with sulfide and dithionite, or by aerobic treatment with selenite. The binding of selenite to the oxidase specifically activates the CO --* methylene blue reaction.Io3’ The molybdenum cofactor liberated from selenium-activated carbon monoxide oxidase does not contain selenium. Here, then, the

Coordination Compounds in Biology

663

selenium may be present as a selenocysteine residue, in which case this would be expected to be the case also for nicotinic acid hydroxylase. The biochemistry of the aerobic carbon monoxide-oxidizing bacteria is of considerable interest in that the respiratory chain is not sensitive to inhibition by CO, due, it is suggested, to the presence of a CO-insensitive o-type cytochrome functioning as a terminal oxidase.'037 Carbon monoxide oxidase is highly specific for CO and does not catalyze the hydroxylation of xanthine, pyrimidines or aldehydes. It has no nitrate reductase or sulfite oxidase activity. One unusual feature is that the active enzyme shows a stable MoV ESR signal in the absence of substrate or other reducing agent. The substrate, CO, did not affect the different ESR signals (from Mo or Fe/S centres) within the turnover time of the enzyme. The use of "CO showed no isotope effect upon the resting Mo" signal of carbon monoxide oxidase, suggesting that CO is not bound directly or through oxygen bridges to the molybdenum.1039aAt this stage it is not possible to speculate upon reaction mechanism. Formate dehydrogenases have been isolated from aerobic and anaerobic bacteria, yeasts and plants, and comprise a rather heterogeneous group of enzymes. E. coli contains two forms of formate dehydrogenase which can be either soluble or membrane bound.'040 Two forms have been isolated from the methanogen Methanococcus vannielii, namely a 105 000 molecular weight protein containing molybdenum and an iron-sulfur centre, and a high molecular weight complex which contains selenium in addition.1041The formate dehydrogenase from Methanobacterium formicicum'028contains a molybdoprotein. Reduction of the enzyme generated ESR signals, confirmed to be due to molybdenum by the use of 95Mo-enriched formate dehydrogenase and the observation of hyperfine splitting. A second component of the ESR spectrum showed no splitting, and was assigned to a reduced iron-sulfur centre. Cell culture in a medium containin high tungstate concentration gave an enzyme without activity or an ESR spectrum. The Mo5 ESR signal is complicated by ligand hyperfine splitting due to two protons, a situation similar to that found for xanthine oxidase. A unique feature lies in the appearance of the ESR signals at ga,> 2.0. This may suggest that the molybdenum in this formate hydrogenase has a different coordination geometry from the octahedral or square pyramidal structures suggested by ESR data on other enzymes.lW2 '~~~ has a molecular weight A molybdenum cofactor has been isolated from Proteus r n i r a b i l i ~ and greater than 1000. The molybdoenzymes of E. coli, in addition to the formate dehydrogenases described above and the nitrate reductase (Section 62.3.9.6), also include the membrane-bound trimethylamine oxidaselm and the soluble biotin sulfoxide r e d ~ c t a s e . " ~ ~

62.1.9.5

Sulfite Oxidase

Sulfite oxidase catalyzes the oxidation of sulfite to sulfate, with cytochrome c as the ultimate physiological electron acceptor. Sulfite is oxidized at the molybdenum site, which is in turn reoxidized by the heme. The enzyme is a dimer, of molecular weight 5 5 000 per subunit. The molybdenum and heme centres seem to be in different domains, as shown by proteolysis. 1046-1048 Sulfite oxidase (and nitrate reductase) has a higher redox potential for Mo than other rnolybdenum enzymes. Although both are inhibited by cyanide in their reduced states, this process is reversed by oxidation and thiocyanate is not released. There appears, therefore, to be no terminal S atom to be removed with cyanide. Conclusions from EXAFS measurements on sulfite oxidase are shown in Table 24.'"' EXAFS data on Chlorella assimilatory nitrate reductase suggest it is similar to sulfite oxidase.'04YThe important feature in the EXAFS results for sulfite oxidase is that there is a loss of an oxo group in going from the oxidized to reduced form. This leads to a simple mechanism in which the molybdenum site acts as an oxo-transfer reagent, converting sulfite to sulfate. The Mo'" generated in this way can then accept a new oxygen from water on being reoxidized to Mom'.

62.1.9.6 Assimilatory and Dissimilatory Nitrate Reductase

Nitrate reductases have been isolated from bacteria, plants and fungi and always contain molybdenum. Two types may be distinguished: (a) the assimilatory nitrate reductases which catalyze the reduction of nitrate to nitrite, which ultimately is reduced to ammonia and used by

Biological and Medical Aspects

664

the cell for growth, and (b) the dissimilatory nitrate reductases which use nitrate as a terminal electron acceptor in the absence of dioxygen. The structures of nitrate reductases vary considerably, and various electron-transfer centres are used. Thus the electron transfer pathway in the assimilatory nitrate reductase of Neurospora crassa is given in equation (56).

Pseudomonas aeruginosa can synthesize both types of nitrate reductase, depending upon the environmental conditions, the dissimilatory enzyme being repressed by d i ~ x y g e n . " ~ ~ Nitrate reductase from Chlorella, an assimilatory enzyme, is a homotetramer of molecular weight 360 000 and contains one each of Mo, heme and FAD per subunit. The nitrate reductase from E. coli is a dissimilatory enzyme. EXAFS data are available on the molybdenum sites in both enzymes (Table 24).'050 The environment of the molybdenum in the assimilatory enzyme is similar to that found for sulfite oxidase, with at least two sulfur ligands near the molybdenum and a shuttle between monoxo and dioxo forms with redox change in the enzyme. This allows a similar mechanism to be put forward for the assimilatory nitrate red~ctase,"~'shown in equation (57), where an oxo group is transferred from nitrate to MotVwith production of nitrite and MeV'. -Z

\ 0-

The structure of the reduced dissimilatory reductase from E. coli shows noticeable differences. In particular, there is a long-range interaction with a n unidentified neighbour. It is tempting to suggest that this is an iron atom from an iron-sulfur centre, and to propose that there is a concerted two-electron transfer to the nitrate.lo5* The different function of the two enzymes can then be related to the different mechanisms. The assimilatory nitrate reductase from Chlorella contains the molybdenum cofactor, as evidenced by the ability of the enzyme to donate the cofactor to the nitrate reductase of the mutant nit-l of N. crassa. Reduction of the enzyme with NADH gives the Mo" ESR signal, which is abolished on reoxidation with nitrate. Line shape and g values of the signal show a pH dependence similar to those observed previously for sutfite oxidase. The signal observed at pH7.0 shows evidence for interaction with a single exchangeable ESR studies on the dissimilatory nitrate reductase are also available. Two signals arising from a high pH and a low pH species of MoV have been found in the enzyme from E. ~ o f i , the " ~ low ~ pH species having a strongly coupled proton.'0s5 Studies on the dissimilatory nitrate reductase from Ps. a e r ~ g i n o s a 'show ~ ~ ~ evidence for three MoV species, a high pH species, and nitrate and nitrite complexes of a low pH species. The iron-sulfur clusters were classified as a [3Fe-4S] cluster in the oxidized enzyme (arising from degradation of a [4Fe-4S] core), and an unusual low potential [4Fe-4SIf cluster in the reduced enzyme. The nitrate complex with the nitrate reductase involves a strongly coupled proton, which is in accord with reduction of nitrate involving electron transfer and protonation.

62.1.9.7 Some General Comments on the Role of Molybdenum The results described in the previous sections have demonstrated the versatility of molybdenum as a reaction centre in biology. The molybdenum cofactor stands at the centre of an important network of cellular functions that are all catalyzed by molybdoenzymes. The similarities and differences in these reactions are of great interest, and relate clearly to the detailed arrangement of terminal sulfur and oxygen atoms. However, the unexpected results found for carbon monoxide oxidase and the requirement for selenium in some cases indicate that other factors are also important. An organism such as E. coli has at least five molybdoenzymes. It will be of great interest to look at the synthesis, availability and cellular distribution of the molybdenum cofactor and its relationship to the function of these molybdoenzymes at different stages of the cell cycle. The study of ch lo rate-re~ istan t'~ and ~ ~nitrate reducta~e-deficient'~~~ mutants of E. coli, which are

Coordination Compounds in Biology

665

mutations involving molybdoenzymes, should throw light on the way in which synthesis and use of the molybdenum cofactors are controlled. It is clear, then, that while significant advances have been made in the biochemistry of molybdenum, there is much that remains to be done before these processes can be understood at the molecular level.

62.1.10

VANADIUM, CHROMIUM AND OTHER ELEMENTS

Vanadium and chromium have been shown to be essential by nutritional studies. Of the remaining metals, it appears possible that tin and tun sten may be essential. The biological properties of the essential metalloids such as seleni~m'~'' and arsenic'060will not be considered. 62.1.10.1

Vanadium

Vanadium is widely distributed in the biosphere, and is present in mammalian tissues at concentrations of about 10 pmol dmP3 or less. Its role as an essential element was shown in the early 1970s when it was established that the growth of small animals was retarded when they were fed synthetic diets containing less than 10 p.p.b. vanadium. It is now recognized that vanadium has numerous physiological eff ects,1062although the nature of its essential physiological role is not known with certainty.1D63Vanadium-containing nitrogenase and b r o r n o p e r ~ x i d a s e 'have ~~~ been isolated. 62.1.10.1.1

Physiological effects of uanadate(V&comptition with phosphate

In many cases the physiological effects of vanadium are due to the close similarity between vanadate and phosphate. Vanadate seems to inhibit specifically those ATPases that form a covalent phosphoryl enzyme intermediate during the enzymatic reaction cycle. Such behaviour was first noted for the (Na+, K+)-ATPase of the sodium pump, as described in Section 62.1.2.2.2, and attributed57 to the binding of H2V04- to the site of enzyme phosphorylation. This led to the suggestion that the physiological role of vanadium lay in the moderation of these processes, and this naturally occurring inhibitor has been used extensively in the study of the mechanism of the sodium pump since then. Vanadate has been found to inhibit, for example, the plasma membrane ~ ~ Saccharomyces cerevi~iae,"~~ which are related to the ATPase of Neurospora C M S S U ~and (Na+, K+)-ATPase, the proton-translocating ATPase of Mycobacterium phleiIM6 and the Ca2+ATPase of red cell rnernbrane~."~~ Other examples are given in ref. 1062. Acid and alkaline phosphatases with phosphorylated intermediates are inhibited by vanadate. This has been exploited in the study of the role of alkaline phosphatase in m i n e r a l i ~ a t i o n . ' ~ ~ ~ Vanadate also inhibits the ATP-dependent degradation of proteins in reticulo~ytes.'~~~ Vanadate is taken up by N. crassa when the cells are phosphate limited and growing in higher pH values. It takes place by the phosphate transport system 11, which has high affinity for phosphate. Phosphate is a competitive inhibitor of vanadate uptake.'071This confirms the competitive relationship between phosphate and vanadate.

62.1.10.1.2 Inhibition of enzymes by vanadyl ion Vanadate is reduced in red cells to vanadyl ion by intracellular glutathione following uptake through a phosphate transport system.1072~1073Free vanadyl is normally unstable with respect to oxidation, but appears to be stable when complexed with intracellular proteins or smaller molecules.'074 Vanadyl ion is a relatively powerful inhibitor of (Na+, K+)-ATPase. For pure fractions of the enzyme, inhibition was nearly complete .at less than 5 pmol dmT3vanadyl ion.107s The state of vanadyl ion at pH 7 is somewhat uncertain, but may involve a hydroxylated species. Vanadyl ion also inhibits alkaline phosphatase more effectively than does ana ad ate.'"^ 62.1.10.1.3 Vanadium as an insulin mimic

Vanadium appears to mimic the action of insulin. The incubation of rat adipocytes with 10-100 Fmol dm-3 V02+ or vanadate stimulates the transport of glucose and 2-deoxyglucose into

666

Biological and Medical Aspects

the Ce111077.'078 to an extent proportional to the vanadium content of the cells. The magnitude of this effect is equivalent to that produced by insulin. It appears that this may be due to V02-e.'077 Externally applied vanadyl ion is less effective than vanadate, possibly because the vanadate enters the cell more rapidly, and is then reduced to vanadyl ion which causes the insulin response. The mode of action of the vanadyl ion in mimicking insulin is not clearly understood. It may function by regulating phosphatase activity in cells. 62.1.10.1.4 Accumulation of vanadium by sea squirts

The ascidians or tunicates (sea squirts) accumulate vanadium from seawater (about 5 x lo-' mol dm->) to a level of about 1 mol dm-3 and store it in a dilute solution of sulfuric acid (pH < 2) in blood cells called vanadocytes. The tunicates thus concentrate vanadium several NMR, ESR and EXAFS determinations on whole vanadocyte cells o f Ascidia milli~n-fold.'"~~ ceratodes and Ascidia nigra indicate that the vanadium is present mainly as aquated V"' probably complexed with sulfate. Some vanadyl ion (5-10%) is also present.'080,'08' A chromogen has been isolated from several species (tunichrome), but its role is not understo~d.~~~~J~~~ It was once thought that the vanadium in vanadocytes acted as dioxygen carriers but this is now known to be incorrect. Many questions remain to be answered about the role of vanadium in vanadocytes.

62.1.10.1.5 Activation of enzymes by vanadium While in general the biological activity of compounds of vanadium is directed towards the inhibition of enzymes, there are a few cases where vanadate activates enzymes. However, it is probable that this is an indirect effect resulting from the inhibition by vanadate of regulatory processes. This may be illustrated with reference to adenylate cyclase.'084 Adenylate cyclase is regulated by a guanylnucleotide-binding protein. In the presence of GTP plus hormone this protein activates adenylate cyclase. Hydrolysis of GTP by an associated GTPase terminates the activation. Vanadate may thus cause activation of adenylate cyclase by inhibiting the GTPase. Decrease in the rate of hydrolysis of GTP at the regulatory protein will maintain adenylate cyclase in the active state. The recent discovery of a vanadium-containing nitrogenase is described in Section 62.1.14.

62.1.10.2

Chromium

Chromium was proposed as an essential element after the observation that rats fed with Torula yeast developed impaired glucose tolerance.1085The active ingredient in other yeasts was suggested to be a chromium complex, named the 'glucose-toIerance factor'. Brewers' yeast contains chromium in the range 0.35-5.4 pg g-1.1086 There is now much interest in the biology of which has been suggested to be involved not only in the action of insulin, but also in the activation of certain enzymes, and, possibly, the stabilization of nucleic acids. Chromium is also known to be carcinogenic and mutagenic at high concentrations, particularly as chromate. Chromate is reduced in rat liver microsomes to Cr"' and Crv. Cr" is labile and may well be the carcinogenic f01-m."~~

62.1.10.21

The glucose-tolerance factortBg0

It has been proposed that GTF is related to a dinicotinatotris(amino acid)chromium(III) complex, which forms a ternary complex at the membrane receptor with insulin. Porcine insulin binds a Cr(nic),(glutathione) complex very tightly. It has been suggested that GTF is not a chromium complex.'"91This arises from the failure to isolate a biologically active Cr"' complex from extracts of brewers' yeast grown in a medium containing added Cr"'. Two fractions showed biological activity, but further purification resulted in the loss of chromium and the isolation of biologically active chromium-free compounds. One of these was largely tyramine (formed from tyrosine residues), but pure tyramine does not show GTF-type properties. The activity of this fraction must therefore be due to the minor component

Coordination Compounds in Biology

667

( < l o % ) which could not be isolated. At this stage it appears probable that GTF is not a chromium complex, but it is possible that chromium may be involved in the in vivo synthesis of GTF, or that the active chromium complex has not been isolated.

62.1.10.3 Tungsten As noted in Section 62.1.9, tungsten usually acts antagonistically against molybdenum. When tungsten is substituted for molybdate, particularly when administered with selenite, there is a substantial increase in formate dehydrogenase activity of Clostridium thermoaceticum cells and for C. formicometicum. Tungsten may be the preferred metal for formate dehydrogenase in these two cases. A tungsten-requiring, thermophilic methanogen has been rep~rted.'"~

62.1.11 THE TRANSPORT AND STORAGE OF TRANSITION METALS AND ZINC The previous sections have indicated the essential role of these metals in living organisms. It is important, therefore, that transport processes are available for the uptake of these metals from the environment and their delivery to sites where they are incorporated into biological molecules. There may also be mechanisms for the storage of excess metal until it is required by the cell, and for the reprocessing of metals liberated by the breakdown of the molecules in which they are bound. Transport and storage processes involving iron are by far the best understood, both for mammalian systems (transferrin and ferritin) and for microbes (the siderophores in iron transport). In addition, knowledge of the transport and storage of copper and zinc in mammalian systems is advancing steadily, although awareness of microbial transport systems in general is poor. Mammalian control systems for iron transport are more complex than those found in microorganisms. In both cases, there is the problem that, at physiological pH values, iron will be present as highly insoluble Fe"' polymeric species of composition Fe(0) (OH). Organisms need to solubilize iron and to prevent the iron forming insoluble species during storage.

62.1.11.1 Transport and Storage of Iron in Mammalian Systems In an adult human some 65% of the total iron is found in hemoglobin and myoglobin, and the bulk of the remainder is found in the storage proteins ferritin and hemosiderin. A small amount is utilized in iron enzymes at any one time. An account will be given of ferritin and the transport protein transferrin, prior to a general discussion of iron transport and storage.

62.1.11.1.1 Ferritin

This iron storage protein is widely distributed in many mammalian cells, and also in invertebrates, plants, fungi and a number of bacteria, where it is associated with a b-type cytochrome. There have been considerable advances recently in the understanding of its structure and physiological function. 1093-1098 Apoferritin consists of a hollow, spherical shell of external diameter 125 A, which provides an inner cavity of maximum diameter about 80 A for the storage of iron. Apoferritin is made up of 24 subunits, arranged in 432 symmetry (Figure 42). The structure while amino acid sequences are known for has been studied by X-ray crystallography,'098~''00 horse"" and human spleeni102ferritins and for rat liver ferritin.1103The crystal structure shows the presence of two types of channel into the cavity, nameIy six hydrophobic channeh along the three-fold symmetry axes and eight hydro hilic channels along the three-fold symmetry axes. These channels are no more than about 4 wide at their narrowest point, but the protein may have some flexibility and so allow the entry of larger ions. The crystal structure shows that the hydrophilic channel is blocked by two adventitious Cd2+ ions, one bound by three glutamates and the other by three aspartates. The inner cavity is filled with small crystalline particles of composition [Fe(O)(OH)J,[Fe(O)(OPO,H,)], which is 57% iron and contains up to about 4500 iron atoms per ferritin molecule. While the core stores iron as Fe'", it is generaily accepted that deposition and mobilization , Fe"' involve Fe". Apoferritin rapidly accumulates iron in a solution containing Fe" and 0 2 but

1

CCCB-v*

Biological and Medical Aspects

668

Figure 42 The structure of apoferritin. N = N-terminus; protein helices E form hydrophobic channels (reproduced with permission from Adv. Inorg. Biochem., 1984, 5, 39, Elsevier, Amsterdam)

is taken up slowly and to a smaller extent."04 The simplest view of this process is shown in equation ( S S ) , in which apoferritin catalyzes the oxidation of Fe" with a ratio of four Fe" per 0 2 as , found by some worker^."^^ 4Fezf+O2+6H,O

+

4Fe(O)(OH)+SH+

(58)

The deposition of Fe" into apoferritin using '*02 as the oxidant shows that only 3-4% of the oxygen atoms in the core were derived from 0 2 ,irrespective of the amount of iron added. Use of W2"0 as solvent confirms that nearly all the oxygen atoms in the core were derived from the solvent. This work gave a stoichiometry of two Fe" per 0 2 . " 0 6 It is known that this stoichiometry can vary from 2 : 1 to 4: 1 under different conditions. In the case of the lower stoichiometry, superoxide or peroxide must be produced, although they have not been detected. EXAFS measurements on the ferritin core and iron-glycine complex lead to a structure for the core in which the iron centres are surrounded by 6.4* 0.6 oxygens at 1.95 0.02 A distance in a distorted octahedral geometry. Each iron has 7* 1 iron neighbours at 3.29 f 0.05 A. These results, together with data on density and stoichiometry, lead to the structure shown in Figure 43.This involves a layer arrangement with iron in the interstices between two nearly close-packed layers of oxygen atoms, with only weak interaction between adjacent layers. The sheet is terminated by phosphate groups, appropriate for the stoichiometry, so that each strip will be about 608* in width and folded upon itself to fill the core."*'

*

W

--70

d

-

(b)

Figure 43 The structure of the iron-containing core in ferritin. (a) One way of using the phosphorus atoms to terminate the two-dimensional sheets into strips. From the stoichiometry there are nine Fe atoms per P, giving a width of -60 A. The length of the strip depends on the amount of iron in the micelle. (b) Schematic drawing of the folding of the strip into a 70 A diameter micelle (reproduced with permission from J. Am. Chem. Soc., 1979, 101, 67, American Chemical Society, Washington, DC)

Ferritin has also been compared with iron-dextran by the EXAFS technique.'"' The apoferritin controls the deposition of the core. Reconstitution of ferritin under a range of conditions always gives the same structure, which is not the case in the absence.of apoferritin. There are metal-binding sites on the protein shell. There is evidence for the binding of iron to apoferritin, probably by carboxyl groups, but there is little detailed information on these sites.109' On the other hand, other metal ions inhibit the formation of ferritin and may do this by binding at or close to the iron sites. Of most significance appear to be results on Tb3+,Zn2+and V02+,

Coordination Compounds in Biology

669

although several other metals have been studied.'lW*'''o ESR studies suggest that V02+ binds at a stoichiometry of 0.5 V02+ per subunit, indicating that the sites may involve ligands on two chains.'*'' Znz* has little effect on the binding of V02+, although it strongly inhibits ferritin formation, and so may bind at different sites. Crystallographic studies on the Zn2+- and Tb3+substituted ferritin are in progress.1098Zn2+ and Tb3+appear to bind with a stoichiometry of 24 cations per ferritin molecule. The mechanism of deposition and mobilization of iron in ferritin has been much d i s c ~ s s e d . " ~ ~ ~ ' ~Ions ' ~ ~ "and ' ~ neutral molecules can all pass through the protein shell, due probably to the presence of entry channels of different character. The Cd2+ions found in the hydrophilic channel in the crystal structure could be metal ions entering the shell. Two proposals have been made for the mechanism of oxidation of apoferritin-bound Fe2+.In the crystal growth modei, the bulk of the Fe" is oxidized to Fe(O)(OH) on the surface of the It is initiated at catalytic sites on the interior surface of the protein. There growing ~rystallite."~~ is evidence that the mechanism of iron uptake by ferritin changes after the initial uptake. Initial uptake occurs more effectively with O2 as the oxidizing agent, but in the latter stages O2 and KI03 are equally effective. The alternative view is that the protein is involved in catalysis of oxidation at all stages."I3 It is suggested that two Fe" ions bind at each of the pairs of metal sites found in the crystallographic studies, and that each pair interacts with dioxygen to give a p-peroxo intermediate, (Fe-0-0-Fe), which then undergoes reaction to give Fe(O)(OH). The p-peroxo pathway should lead to the incorporation of l80 into the ferritin core, which was not Nevertheless, the "0 evidence is not unambiguous enough to rule out the protein catalysis model. Iron may be released rapidly from ferritin in v i m by the action of reducing agents in the presence of iron chelators. Reducing agents are much more effective than chelators alone. The most effective reducing agent under physiological conditions appears to be reduced flavins, although the rate of reduction b dihydroflavins depends on many factors and seems to be controlled by the protein shell.1 0 9 7 Apoferritin is synthesized in response to several factors, including the presence of iron. Added iron is taken up by the iron transport protein transferrin, which then binds to receptor sites on several types of cells which store i r ~ n . " ' ~ * " ' ~

62.1.3 1.1.2

Transferrin

The transferrins are proteins that bind and transport iron as Fe"'.1*09*11'6-1118 They include lactoferrin from milk, ovotransferrin from egg white, and serum transferrin from a range of organisms. Uteroferrin, considered in Section 62.1.5.5.2 on the purple acid phosphatases, is an iron-binding protein with phosphatase activity, that has been proposed to transport iron from maternal to foetal c i r c u ~ a t i o n . ~There * ~ - ~ are ~ ~ distinct differences between the iron-binding sites in uteroferrin and transferrin, and so uteroferrin will not be discussed in this section. Each of the transferrins is a glycoprotein of molecular weight about 80 000, made up of a single polypeptide divided into two similar domains. Both domains have a specific binding site for one iron(II1) ion and an obligate anion which under physiological conditions will be carbonate. Iron is not bound in the absence of the anion. Other anions such as oxalate, malonate, citrate or EDTA will also activate the metal-binding site. The binding of Fe"' in the two domains of transferrin suggests that there is no interaction between them, in accord with distance measurements. A recent value is 35 A, determined by consideration of fluorescence energy transfer between the two sites."" Amino acid sequences have been reported for human serum transferrin and ovotransferrin,"20-"22 and partially for lact~ferrin."~~ The two halves show a high degree of homology."22 Partial proteolysis of transferrin gives half transferrins with one Fell' site and Thus, a fragment of molecular weight 35 600 has been molecular weights close to 40 oO0.1124~1125 obtained from human transferrin by proteolysis with thermolysin. This is derived from the N-terminal domain and is obtained by cleavage of a Pro-Val bond (residues 345 and 346).1124 The two carbohydrate components of human serum transferrin are attached to residues 415 and 608 in the C-terminal domain of the protein, and appear to have little effect on the metal sites. Loss of the carbohydrate groups has no effecton the ESR spectra of the V02+ complex of human transferrin." l 6 The transferrins belong to the iron-tyrosinate group of proteins discussed in Section 42.1.5.5.2. Charge transfer from phenolate ligands to FelI1accounts for the salmon-pink colour of transferrin. The detailed coordination environment of the iron in transferrin is not known with certainty, as

670

Biological and Medical Aspects

X-ray structural analysis has not yet been carried out at sufficient resolution. However, traditional chemical and spectroscopic methods have given a great deal of information. The involvement of tyrosine and histidine residues as ligands has been suggested by a comparison of potentiometric titration data for transferrin and the apoprotein, and by the use of chemical modification techniques and NMR and Raman spectroscopy. This has led to the conclusion that each iron is coordinated by two histidine and two or three tyrosine residues. The two homologous regions in the primary structure of transfenin which contain the metal-binding site only have two tyrosine residues each. The results of UV difference spectra have also suggested that there were two tyrosine ligands for each iron.1126The involvement of histidine as a ligand is supported by the observation of appropriate superhyperfine splitting in the ESR spectrum of copper transferrin. The coordination is completed by an aqua or hydroxo group and the obligate anion. The involvement of the anion as a ligand has been disputed but now seems established beyond question by spectroscopic studies on a range of metallotransferrins and various anions.1127The anion binds weakly to the apoprotein as a prerequisite for the binding of iron. It is reasonable that the anion binds to a positively charged group, which otherwise would hinder the binding of iron. The identity of the anion-binding group has been proposed to be a protonated imidazole of a histidine residue112' or a guanidinium group of an arginine residue.1129The former suggestion is supported by the fact that the conserved residues in both domains include three histidines. Two of these are ligands to iron, while the third could be the anion-binding site. The iron centres in transferrin, in summary, are of approximately octahedral geometry with two tyrosinate, two imidazole, a carbonate (or bicarbonate) and a water molecule (or hydroxide ion) as ligands. This is consistent with the loss of three protons on binding of iron to apotransferrin, namely from two tyrosines and the aqua group. The ESR spectra of the two monoiron(IT1) half fragments differ slightly from each other, suggesting there are some differences between the two sites. The ESR spectrum is an unusual one which has not yet been reproduced in model compounds. A typical model1130for the iron site in transferrin is the complex of Fe'" with N,N'-ethylenebis( 0hydroxypheny1)glycine. As hydrolysis of this compound occurs, so there is an increasing similarity in the zero-field splitting parameters in the ESR spectra of model and transferrin. These compounds appear to be s i x - ~ o o r d i n a t e . " ~ ~ , " ~ ~ The significance of the two sites in transferrin has been much discussed. The sites are distinguishable spectroscopically and have different affinities for iron, which may be dependent on the anion used. The two sites release iron at different rates in a pH-dependent manner. The site in the C-terminal half of human serum transferrin (once designated the A site) retains its iron at pH 6.0 and so is the acid-stable site. The site on the N-terminal half is the acid-labile site. It is now possible to separate the four principal species of transferrin [diiron(III), N-terminal monoiron(lII), C-terminal monoiron(II1) and the a otransferrin] by the techniques of polyacrylamide gel electrophoresis in the presence of urea."'PThis allowed detailed studies of iron binding to transferrin, so that equations are now available for calculation of the distribution of iron between the various transferrin species.1134 The removal of iron from transferrin by ligands in vitro is difficult. Nevertheless, transferrin must be able to release iron readily when required. Reductive removal of iron does not occur, as the iron is released as iron(II1). Studies on the delivery of iron to reticulocytes have failed to In uitro studies on show that iron is taken up preferentially from either site in transferrin.113531136 the use of chelating agents to remove iron suggest that Fe"' transferrin undergoes a conformational change to give an open conformation. This species can form a complex with the chelating agent, which is eventually released as a complex with the Fe"'.''37 Such a process could occur in vivo. Pyrophosphate mediates the removal of iron from transferrin to isolated respiring rat liver mitochondria. It appears that phosphates may be chelators of iron in removal of iron from transferrin in cells. Lactoferrin occurs in high concentrations in human milk. It has high affinity for iron, and it may have a bacteriostatic function in depriving microbes of essential iron required for their growth. It supplies iron to intestinal tissue and probably has a nutritional role in supplying iron to newly born infants.

62.1.11.1.3 Phosvitin

Phosvitin is a highly phosphorylated, iron-containing protein found in egg yolk and serves as an important source of dietary iron.113R It could be viewed as a source of phosphate, as the protein contains about 130 residues of phosphoserine which appear to serve as ligand groups for iron.

Coordination Compounds in Biology

67 1

It is also possible that this phosphoprotein could be involved in the transport of other metals.1138 Isolated phosvitin usually has two to three atoms of iron, with Fe/phosphate ratio about 1: 50. Considerably larger amounts of Ca2+ and Mg2+ are bound [(Ca2++Mg2+)/phosphate ~ 0 . 7 1It. appears to have no known enzymatic or regulatory role. 62.1.11.1.4 The transport of iron Iron entering the bloodstream from the gastrointestinal tract is thought to be present as Fe", and must be oxidized to Fe"' before binding to transferrin, which then delivers iron to many different types of cell. The non-enzymic route for oxidation of Fe" in serum appears to be too slow for the formation of iron(II1) transferrin. As noted in Section 62.1.8.5.1, the copper protein c e r u l o p l a ~ m i nhas ~ ~ ferroxidase ~-~~~ activity, being responsible for the oxidation of Fe" to Fe"'. It is well known that deficiency of copper influences iron metabolism in animals, in accord with this role for ceruloplasmin. Transferrin also has to cope with the iron liberated on breakdown of iron enzymes. It carries iron from the breakdown sites of hemoglobin in the reticuloendothelial cells of spleen and liver back to the sites in bone marrow for the synthesis of hemoglobin. Serum transferrin accounts for about 4 mg of body iron in humans, but delivers some 40 mg of iron daily to the bone marrow. Serum transferrin is normally only about 30% saturated with iron. This explains why it has a relatively high capacity for binding other metals and so is implicated in the transport of 67Ga,1'39 which is used as an imaging agent for various soft tissue tumours and inflammatory abscesses.1140 Transferrin also facilitates movement of gallium across tumour cell membra ne^."^^ The uptake of iron from transferrin into cells has been best studied for reticulocytes, where there is heavy demand for iron for the synthesis of heme. Reticulocytes and, by analogy, other mammalian cells have receptor molecules on their plasma surface which bind transferrin. Characterization of the transferrin receptor from rabbit reticulocytes suggests that it is probably a glycoprotein of molecular weight around 200 000. The number of receptor sites in cells varies in response to the conditions. For example, treatment of K562 cells with the efficient iron chelator desferrioxamine results' in an increase in the total number of receptors for tran~ferrin."~' It appears that both halves of the transferrin molecule contain a recognition site for the receptor, and that both are necessary for binding. Thus, the two halves of ovotransferrin are much less effective than the whole molecule in binding to the receptor and donating iron into the chick embryo red blood cell.1143One fragment of serum transfenin is ineffe~tive."~~ The uptake of iron into the cell could follow several pathways. The iron could be released from the transferrin at the receptor site and be carried into the cell. Alternatively, the whole transferrinreceptor complex could be taken into the cell via endocytosis, and passed into an acidic compartment, where the iron is released, passed out of the compartment, and stored in ferritin. The final step in the biosynthesis of heme takes place within the inner membrane of the mitochondria where the ferrochelatase catalyzes the insertion of Fell into protoporphyrin IX. Thus, the major part of the iron entering the cell is used in the mitochondria. Little is known about the way in which the iron enters the mitochondria. It seems unlikely that transferrin could deliver iron directly to mitochondria. Probably it is transferred to ferritin, which can then transfer it to the mitochondria. The mitochondria reduce exogenous flavin, which in turn reduces iron in ferritin,"44 causing its release prior to uptake. Experiments suggest that the ferrochelatase reaction is the rate-limiting step in the biosynthesis of heme, and that the rate of mobilization of iron from ferritin by mitochondria with intact energy coupling and energy conservation properties is about three times the rate of formation of protoheme.

62.1.11.2 Transport and Storage of Other Transition Metals and Zinc in Mammalian Systems Most of the transition elements can form complexes with transferrin and other proteins in serum, and can be stored in ferritin. This accounts for the effect of iron metabolism when the concentrations of these metals rise. Limited information which shows the presence of specific transport and storage systems is available on some metals. 62.1.11.2.1 Tmnsport of copper Copper is bound by albumin or histidine after uptake in the gut, and transported in this form to the liver, where it is transferred to ceruloplasmin. Some 95% of copper in human serum is

672

Biological and Medical Aspects

bound to ceruloplasmin, while the remainder is bound to albumin or histidine. It appears, therefore, that ceruloplasmin is the major transporter of coppery75 although definitive confirmation is not available. It is reasonable to assume the existence of receptor sites in the various target tissues, and the possibility of copper release by a reductive method. If this occurs at the cell membrane then there may be an intracellular receptor for CUI.

62.1.11.2.2

Transport of zinc

About one third of the zinc in venous plasma is bound to a,-macroglobulin, and the remainder to albumin, with the exception of trace amounts bound to histidine and cysteine.1145However, transferrin has been implicated in the uptake of zinc from the intestinal membrane, while albumin is involved in the removal of zinc from intestinal mucosal cells and its transport to the liver. Other ligands proposed for various transport processes for zinc are citric acid and picolinic acid.'146

62.1.11.2.3

Transport of nickel, manganese, cobalt and vanadium

Very little is known about the transport of nickel, manganese and cobalt. Plasma transferrin, conalbumin and citrate have been suggested to serve as carrier ligands. Transferrin is the main transport protein for vanadium in humans, and will transfer vanadium to ferritin.

62.1.11.2.4 Stomge of copper and zinc. Tke d e of metallothionein

The addition of copper, zinc, cadmium or mercury to animals results in the synthesis of a cysteine-rich protein called metallothionein. 1147-1149 These proteins have been isolated from a number of sources, and have molecular weights in the range 6000 to 12 000 with a cysteine content of about 30-35'/0 of the total amino acid content. They have also been found in microorganisms and plants. These proteins are thought to play an important role in the storage of zinc and copper, and as a result of their storage capacity, are able to bind and detoxify cadmium and mercury. Foetal and neonatal livers contain exceptionally high levels of copper compared to the adult organ. Thus, the livers of new-born rats contain as much as 20 times the level of copper and zinc metallothionein as that found in 70-day-old rats.'150 Again, a metallothionein from foetal bovine liver contained eight copper and two zinc atoms per molecule of protein.i151These proteins can only be isolated with difficulty under oxygen-free conditions. It appears then that large amounts of copper (and zinc) are stored in the liver bound to metallothionein, and are mobilized as required for enzyme synthesis after birth. 1

50

Figure 44 The amino acid sequence in equine metallothionein

10

60

61

Coordination Compounds in Biology

673

The amino acid sequences of many mammalian metallothioneins are known.i1s2They all contain 61 residues with the cysteine residues distributed along the full length of the polypeptide chain, and with N-acetylmethionine and alanine at the N- and C-termini, respectively. All 20 cysteine residues are involved in the binding of up to seven metal ions via thiolate linkages. A typical amino acid sequence is shown in Figure 44. The involvement of thiolate groups as ligands was first demonstrated by a comparison of the charge-transfer bands in metallothionein with those of thiolate complexes of zinc and cadmium. Tetrahedral environments around the metal were confirmed from a study of the d - d spectra of the cobalt-substituted protein,1153although the use of perturbed angular correlation spectroscopy of y-rays suggests the presence of two types of environment.1154EXAFS measurements on ZnsCuz sheep metallothionein suggest a tetrahedral environment of four sulfur atoms with a Zn-S distance of about 2.29 Tetrathiolate coordination of the metals in metallothionein implies that the metals must be present in clusters with some of the cysteine residues serving as bridging ligands. The classification of the thiolate groups as bridging or terminal is confirmed by ESCA spectros~opy."~~ The existence of clusters is in accord with 'I3Cd NMR spectra of "3Cd-labelled metallothioneins, which show '13Cd-l 13Cd spin 15' Decoupling experiments allow the suggestion of two metalThe A cluster thiolate clusters containing four and three metals termed A and B re~pective1y.l'~~ is formed by the C-terminal half of the protein and contains four metal ions bound to 11 cysteine residues, five of which are bridging. The B cluster binds three metal ions through nine cysteines, three of which are bridging. The two clusters differ in their metal-binding properties. Mammalian Cd, Zn metallothionein contains four Cd ions in cluster A and two Cd ions and one Zn ion in cluster B.1157,1'59 Calf liver metallothionein contains three Cu ions in cluster B with Cd-displaceable zinc ions in cluster A.116o Studies on the binding of zinc and cadmium to human liver metallothionein show that binding occurs first in cooperative fashion to cluster A, followed by cooperative binding to cluster B. Incubation of the 7Cd metallothionein resulted in loss of Cd ions from cluster B initially.'16' The addition of Co" to rabbit liver metallothionein appears to take place through the selective build-up of clusters, as shown by ESR, magnetic susceptibility and electronic On filling up the metal sites on the apoprotein, the ESR signal from the high-spin Co" increases in proportion to the binding of four metal ions, but declines on incorporation of the remaining three. The stepwise incor oration of Cd2+ into the apoprotein has been followed by electronic, CD and MCD spectra.'14'In these studies it is suggested that, for both Co2+ and Cd2+,the first four metal ions bind mainly to four thiolate groups each, and that cluster formation does not take place until the binding of the last three metal ions, where bridging thiolates become necessary as ligands. These two views of the binding of metals to metallothionein require some reconciliation and a decision on whether the first four metals form a cluster before further metal is bound. The magnetic data suggest that they do not. Several studies on models have been reported, both on the binding of a range of metals to apometallothionein and the design of ligands, particularly with Cys-X-Cys and Cys-X-Y-Cys arrangements as found for metallothionein. 1163~1164All metal derivatives appear to bind in metalthiolate clusters.1149Platinum has also been found bound to metallothionein in rat tissues following treatment with cisplatin and the trans isomer.'165 Metallothioneins play an important role in the metabolism of copper, both in absorption and handling by the liver. Dietary supplementation with copper leads to a large increase in liver copper-metallothionein. The biological life of metallothionein can vary considerably. The half-life of Cu-metallothionein in zinc-deficient rats (12.3 h) was lower than that in zinc-sufficient rats (16.9 h)."66 Usually zinc is found in Cu-metallothionein, so zinc may stabilize the copper metall~thionein."~' Metallothioneins probably participate in other cellular functions, for example in nucleic acid metabolism.'168

62.1.11.3 Transport of Iron in Microorganisms

Microorganisms, with the exception of certain of the l a c t ~ b a c i l l i , ~require ' ~ ~ iron, and have efficient mechanisms for its controlled uptake. Many organisms have several transport systems for iron. These have different affinities for iron, but those of high affinity are of especial interest. The extreme insolubility of iron in aerobic environments at physiological pH values presents a remarkable challenge to the microorganism. Furthermore, microorganisms living within higher

Biological and Medical Aspects

674

organisms have to compete for iron with animal transferrins which themselves have high affinity for iron. Pathogenic microorganisms have to scavenge iron in situ, for example, and so the availability of iron is closely linked to infections such as coliform invasions, septicaemia, gangrene, etc. All of this emphasizes the need for microorganisms to have highly efficient means for gaining iron. Microorganisms accomplish this by synthesizing special ligands, the siderophores, which are low molecular weight compounds of remarkable affinity for Fe"', having log K values greater than 30 and sometimes over 50. These ligands are synthesized in the cell in response to low levels of iron, and are then secreted. They bind and solubilize Fe"' to give siderophore-iron complexes, which are taken back into the cell via specific receptor sites on the cell surface. Siderophores are synthesized by a range of organisms and a good many have now been characterized. Table 26 lists some of these. They may be isolated quite easily in many cases by growing organisms under low iron or iron-deficient conditions, when their syntheses are derepressed. The iron has to be removed selectively from the growth medium. This may be accomplished by the use of ion exchange resins such as Chelex-100, or immobilized 8h y d r o x y q ~ i n o l i n e , ~although '~~ a variety of methods have been used in the past.'"' Growth of a number of organisms under such iron-limited conditions will lead to secretion of the siderophore into the culture medium. In some cases this will be shown by the accumulation of coloured degradation products of the siderophore in the culture supernatants. Procedures for isolation of the siderophores either as the FelI1 complex or as the iron-free form have been d e ~ c r i b e d . ' ' ~ ~

Table 26

Type

Example ~~~~

Cyclic tricatecholates Linear catecholates Cyclic trihydroxarnates Cyclic/linear hydroxamates

Hydroxarnatc/citrate Dihydroxamate/phenolate Hydroxamate/catecholate/hydroxy acid

Classes of Siderophores

~

Source

log K

~

Enterobactin Agrobactin Parabactin Vibriobactin Fernchrome Fusarinines Rhodotorulic acid Neurosporin Ferrioxamine Schizokinen Arthrobactin Aerobactin My cobact ins Pseudobactin Exochelin

Enteric bacteria Agrobacterium tumefaciens Paracoccus denitrificans Vibrio cholerae Fungi (used by bacteria) Various fungi Rhvdvtorula spp. Neurospora crassa Streptomyces spp. Bacilli Arthrohacter spp. Salmonella spp. Mycobacteria Pseudomonas spp. Mycobacteria

52

31.2 30.6 22.9

The siderophores fall into two main chemical types, with either catecholate-phenolate or hydroxamate donor groups. Several subdivisions of the hydroxamate siderophores are known, and additional examples continue to be identified. The siderophores, with a small number of possible exceptions, bind Fe"' with six hard, often negatively charged oxygen donor atoms. The exceptions may provide five such donor atoms and one nitrogen donor. These are groups well known to be good ligands for Fe"'. The siderophores bind Fe" and other metal ions to a much lesser extent. Cr'" siderophores have been used as kinetically inert compounds, while Ga"' forms stable complexes, as does Al"' to a lesser extent. The presence of siderophores in a medium may be shown by adding an iron( 111) salt, as complex formation will be demonstrated by the colour of the Fe"'-siderophore complex, due to chargetransfer bands in the visible region. Chemical and spectroscopic tests allow ready classification into catecholate and hydroxamate types, for example the use of the Arnow and Czaky colorimetric reactions, re~pectively.''~~ The siderophores have been discussed in Chapter 22 (Volume 2) and so will only be considered briefly here. All aspects of iron transport in microor anisrns and clinical uses of siderophores have been reviewed on many occasions. 1169~117031172-1

"'

Coordination Compounds in Biology 62.1.11.3.1

67 S

Multiple pathways for iron transport

Many microorganisms have several pathways for uptake of iron, which are induced under particular conditions. This ensures that there is a continuous supply of iron at the correct concentration level. A low affinity system, which only functions in iron-sufficient conditions, seems to be widely distributed. This may involve the adsorption of Fe(O)(OH) polymer on the cell wall and transport of iron into the cell. Escherichia coli has at least five independent transport systems, one of which is the low affinity pathway described above. In addition, it synthesizes enterobactin as a siderophore; it can take up the iron(II1) complex of ferrichrome, a siderophore synthesized by certain fungi; there is a citrate-induced system, and a less common process involving aerobactin. While E. coli only synthesizes one siderophore, other organisms synthesize a range of siderophores. Thus, the mycobacteria produce: several cIasses of the siderophore exochelin, which differ in the solubility of the iron( 111) complex.1179Mycobacterium smegmatis may produce up to eight exochelins.'180The structures of these exochelins are not known, but appear to be peptides with molecular weights less than 1000. Mycobacteria also synthesize mycobactins, which are intracellular iron-binding compounds, but which appear to be involved in iron transport in the early stages of growth. Under other conditions, mycobactins may serve as a store of iron. Mycobacteria also secrete salicylic acid and 6-methylsalicylic acid. Shigella flexneri synthesizes the hydroxamate siderophore aerobactin (and a single outer membrane protein), but certain strains also synthesize enterobactin, a phenolate siderophore, in which case three additional outer membrane proteins are synthesized, similar to those produced b y E. coli K-12 strains."'*

62.1.11.3.2 The catecholate siderophores

The best known example is enterobactin (otherwise called enterochelin), which is produced apparently by all enteric bacteria. It has three 2,3-dihydroxybenzoyl groups attached to a macrocyclic lactone derived from three residues of L-serine condensed head-to-tail. The structures of enterobactin and its iron complex are shown in Figure 45, which shows that the iron is bound by six phenolate oxygen atoms in an octahedral environment. Enterobactin has the highest known affinityfor FeI'I, with Log K = 52 at pH 7.4."" The iron(II1) complex can exist as isomeric forms, which may be associated with selectivity in binding to the receptor site.

0

0

=6

HCH

Enterobact in

IronUU) enterobactin

Figure 45 Enterobactin and its Fe"' complex

Biological and Medical Aspects

676

Linear catecholate siderophores are also known. The structures of parabactin and agrobactin are shown in Figure 46. They contain an oxazoline ring, which is subject to acid hydrolysis. The open-ring derivatives are agrobactin A and parabactin A. In these siderophores there is the possibility of coordination of the ring N atom, although agrobactin can provide six phenolate groups. The conformations of parabactin and its Gar” complex have been examined”83 by high resolution NMR. While the ligand exists in three separate conformers, only the A cis chelate was formed in the complex. This may be of significance as a recognition feature. A related example is vibriobactin, from Vibrio cholera, which contains three 2,3-dihydroxybenzoic acid functions, like agrobactin, but has two oxazoline

R = H Parabactin

Parabactin A

R = Me Agrobactin

co

CO

NH

NH

I

I

I

I

CH -(CHz),

-CHz

I

co NH

I

NH

.

I

t

CHCHzOH

CHz

I

COiH

C02H

CO

I

2,3-Dihydroxybenzoylserine

I

COzH Itoic acid

Figure 46 Parabactin, agrobactin and some small siderophoric phenolates

A number of catechols of lower molecular weight, which have siderophoric properties, have been isolated, Examples are shown in Figure 46. All organisms synthesizing enterobactin also form 2,3-dihydroxybenzoylserine,while a number of related species are also present. One of these, itoic acid, from Bacillus subtilis was the first phenolate siderophoric ligand to be i~olated.”’~ A great deal of work has been carried out on model compounds with phenolate and catecholate 187 and to some extent with modified siderophores.”** The important question of the ligands,1186.~ mechanism of release of iron(II1) from the siderophore has also been a central issue in these model studies.

62.1.11.3.3

The hydroxamate siderophores

Ferrichrome was the first ligand of this type to be isolated (Figure 47). Ferrichrome is a cyclic peptide with three hydroxamic acid side-chains. It gives a neutral complex with Fe”’. A number of variations having substituents on the hexapeptide or on the acyl group are also found. The ferrichromes are synthesized by fungi, but they are also used by many bacteria as a source of iron, even though they do not synthesize the siderophore themselves. The fundamental structural unit of the ferrichromes, N’-hydroxyornithine, is also present in rhodotorulic acid (and related compounds), the fusarinines ( N*-acyl-Ns-hydroxyornithine) and in neurosporin. The diversity of structure and biological activity of these siderophores results from the various substituents present in the peptide and acyl groups. The hydroxamate siderophores are better characterized from a structural viewpoint. X-ray diffraction data are available for the

Coordination Compounds in Biology

677

Fe"' complexes of ferri~hrome,"~~ ferrichrome A,"90 ferri~hrysin"~'and ferricrocin."gz These ferrichromes all show similar conformations of the peptide backbone and similar stereochemistry. ~ ~ ~all~contain ~ ~ ~ ornithine in the L configuration. Together with two fusarinine s i d e r o p h o r e ~ , "they Surprisingly the D isomer of Ns-hydroxyornithine is present in neuro~porin."~~

\ O

F'

" ' I c /

/H'

Femoxamines

Ferrichrornes

pH

leC-C-N

wo 0 I

HO 0

J : r N - T - i M e

II

\

,CH2CH20

HC=C,

H

Me Fusarinines

Me

Me

I

I

c=o

c=o

NOH

NOH

I

1

I

(CH214

I

I

C o d

I

(CH2)4

I

CHNHCOCHzCCHzCONHCH

I

C02H

I

I

OH

C02H

Aerobactin Figwe 47 Ferrichrome and other hydroxamate siderophores

0

It

R

I

/C~NH ,CH,

CH2

(CHt).-N-CMe

I

II

\

HO

0

/

HO

0

HO-C-COZH

I

CH* \C/NH,

I1 0

I A n

L

Rhodotomlic acid

H

,(CH2).-N-CMe CH

I

R

Schizokinen: R = 8, n = 2 Aerobactin: R = CO,H, n =4 Arthrobactin: R = H, n = 4 Figure 48 The citrate hydroxamates

II

Biological and Medical Aspects

678

The ferrioxamines are a family of linear and cyclic siderophores built up from l-amino-5(hydroxyamino)alkanes and succinic and acetic acids. Figure 47 shows ferrioxamine B, a linear trihydroxamic acid, which is probably the best-known example. The mesylate salt of the deferri form of ferrioxamine B is used in the clinical treatment of iion-overloaded patients. Exchange of iron between ferrioxamine 3 and ferrichrome A at pH 7.4 has a half-life of about 220 h.' '96 Siderophores with citrate-hydroxamate structures (schizokinen, arthrobactin and aerobactin) are synthesized by enteric bacilli and Arthrobacter spp. Their structures are shown in Figure 48.

62.I . 11.3.4 Pseudobactin The pseudomonads produce a variety of siderophores, including pyochelin (a phenolic compound from Pseudomonas aeruginosa), the trihydroxamate nocardamine from Ps. stutzeri and pseudobactin from Ps. jluorescens. Pseudobactin i s an unusual siderophore in that it carries the three functional groups found generally in siderophores, namely hydroxamic acid, catechol and a-hydroxyacid, which are substituents on a linear hexapeptide, as shown in Figure 49.'197

HN

HC-CHZ

Me

I I HC-C=O Me HO-CH I I I I NH2CH-CONH-CH-CONH-CH-CONH-CH-CONH-CHCO (CHA

Me

I

I

NH

1

Figure 49 Schematic structure of pseudobactin

62.1.11.35 The binding of Fer"-siderophore complexes to receptors and the release of irun into the

cytoplasm

The iron-siderophore complex is too large to permeate the water-filled channels in the outer membranes of Gram-negative bacteria. Accordingly, a transport mechanism is necessary, involving proteins and expenditure of energy, although these are not well understood at present. The first stage in transport involves the recognition of the iron( 111)-siderophore by receptor proteins in the cell membranes. These proteins are synthesized at low levels of iron. The receptor sites have been identified, partly because they also bind phages and antibiotics (which results in the death of the cell). Specific receptor proteins have been detected in E. coli K12 for the binding of iron(I11) complexes of enterobactin, citrate, ferrichrome and aerobactin, although other proteins may also be necessary. The receptor for Fe"'-enterobactin retains its affinity for the complex and for colicin B in vitro after extraction. The receptor has a dissociation constant of about 10 nmol dm-' for Fe"'enterobactin. The gene for synthesis of the receptor is FepA. A range of modified enterobactin molecules have been synthesized in which the three 2,3'dihydroxybenzamide groups are substituted into a benzene ring rather than the macrocyclic lactone. In addition various substitutions were made."88 The ability of these compounds to support the growth of E. coli K12 under iron-limited conditions and also to protect against colicin B was studied. Growth was observed with certain of these modified siderophores, notably the models with the unsubstituted 2,3-dihydroxybenzene ring, which also protected against colicin B. This was not the case with the model with a sulfonated benzene ring. These results suggest that the tricatechol-iron( 111) centre is specifically recognized by the receptor site on the protein, in accord with the transport activity of both cyclic and linear enterobactin analogues. Growth was also

Coordination Compounds in Biology

679

observed with FepA mutants lacking the receptor protein in the outer membrane. This may mean that amounts of the Fe"' analogues sufficient to support growth passed through the outer membrane. However, FepB mutants deficient in the iron complex-translocating system did not grow. This shows that, for E. coli, active transport of the siderophore-iron complex into or across the cytoplasmic membrane takes place. The transport or release of iron has been much discussed. One view is that the iron is transferred from the siderophore complex at the outer membrane to another membrane-bound protein, for example in the uptake of iron by rhodotorulic acid in Rhodotomla. The other view is that the intact Fe"'-siderophore complex is taken up into the cell. This is supported by studies with inert chromium(II1) complexes in E. coli and also by labelling studies. Particular discussion has centred around the mechanism of release of the iron, in view of its high affinity for the siderophore. A very attractive proposal is that this simply involves the reduction of the Fe"'-siderophore complex. Iron(I1) has much lower affinity for the siderophore, and so can be released: values of log K for Fe" are about 8. Good evidence has been found for siderophore reductase activity in bacteria and fungi.1198Reduction of Fe"'-hydroxamate siderophores can occur readily under physiological conditions. The release of iron from catechol siderophores is less well understood, as reduction of Fell'enterobactin (-750 mV) by physiological reducing agents seems unlikely. Accordingly, it has been suggested that the Fe"'-enterobactin must undergo hydrolysis before reduction, which means that the ligand cannot be recycled. A similar argument has been put forward for Fe"'-parabactin. The reduction potential is -673 mV at pH 7, while that of Fe"'-parabactin A is around -400 mV. It is suggested, therefore, that intracellular release of iron from parabactin may occur by hydrolysis The need for the of the Fell'-parabactin to the more easily reducible Fe'l'-parabactin A.1199 degradation of enterobactin has been challenged. Mossbauer, ESR and titration experiments have shown that iron complexes of enterobactin and synthetic analogues remain as Ferrrbetween pH 2 and 10. But in methanol solution at low pH, there appears to be a quasi-reversible redox reaction, possibly reduction of the protonated Fe"'-enterobactin complex by the catechol group to give a semiquinone radical."86 This offers a mechanism for cellular removal of iron from catechol siderophores without having to degrade the siderophore structure. It should be noted that while the inability to recycle the siderophore seems wasteful, there is good evidence that the hydroxamate siderophore ferrochrome is also modified into an N-acetyl form of low affinity for iron,'200and so cannot be reused. The transport process is unclear. In some cases it is inhibited by uncoupling agents, suggesting that uptake is linked to the electrochemical proton gradient, while evidence has also been found for a symport mechanism involving ferrichrome and Mg2f.'20'

62.1.11.3.6 Microbial iron transport and infections in animals and plants

Microorganisms in higher organisms compete for iron with transferrin and various ironcontaining molecules. There is a good correlation between the virulence of pathogenic infections and the extent of siderophore synthesis, often aerobactin.'202Thus, one role of transferrin is to protect against infection by denying iron to invading organisms. Lactoferrin in milk and conalbumin in eggs have a similar role. The siderophore pseudobactin synthesized by soil organisms is able to stimulate the growth of the plants they invade by strongly complexing iron, so inhibiting the growth of the plant pathogen, Erwinia. The competition between transferrin and infectious organisms is a fascinating topic. During infection, the mammalian host will attempt to diminish the iron available to the bacterium. The iron levels in the blood decrease and the iron stores in the liver increase. The observation that siderophore synthesis is sharply reduced at temperatures just over 37°C may mean that the physiological function of fever is to depress bacterial siderophore synthesis, as shown for a rabbit pathogen, Pseudomonas m u l t o ~ i d a . ' ~ ~ ~

62.1.11.4

The Storage of Iron in Microorganisms

The possible role of the mycobactins in the storage of iron in the mycobacteria has been noted. Of greater significance is the identification of ferritin-like molecules in some bacteria. The cytochrome b557.5from Azotobacter uinelandii, known to be associated with large amounts of iron, is now known to be a ferritin, with an iron content of 13-20% and an electron-dense core of

680

Biological and Medical Aspects

55 b, diameter. This is suggested to be a store of iron for the synthesis of nitrogenase.'204 Cytochrome bl from E. coli is also recognized to be a bacterioferritin, and has 24

62.1.11.4.1

Magnetotactic microorganisms

Certain aquatic bacteria are influenced by the magnetic field of the earth.'206These magnetotactic organisms synthesize small crystals of magnetite, Fe,04 .lZo7The swimming direction of the bacteria is thus controlled by the magnetic field. The intracellular iron grains in Aquaspirillum magnetotacticurn are observable by electron microscopy. They are enclosed by a membrane that is close to the cytoplasmic membrane, and may be contiguous with it. The grains are single crystals of hexagonal prismatic shape, with lattice spacings appropriate for magnetite.'20s Each ceIl contains about 20 cubic-octahedral magnetosomes averaging 420 A each side and arranged in a The particles were within the single magnetic domain range for magnetite. A. magnetotacticum synthesizes magnetite in the presence of an iron source and when O2 is growth-limiting. The cells are sensitive to oxygen. The addition of catalase allowed aerobic growth of the cells without the formation of magnetosomes. It is possible that formation of magnetite results from a protection mechanism against H z 0 2 through the removal of iron which would catalyze the decomposition of peroxide. An important result comes from Mossbauer spectroscopy, which has shown the presence of ferritin and the magnetite in magnetosomes. There is increasing evidence for the occurrence of magnetotactic bacteria and algae.lZo6This has considerable biochemical significance, and suggests that fine-grain magnetic iron compounds in nature may be of microbial origin. It is difficult to be certain of the physioiogical function of this phenomenon, other than, perhaps, to direct the swimming direction of the organisms down towards sediments.

62.1.11.5 Transport and Storage of Iron in Plants A number of low molecular weight chelating agents occur in plants, and have been termed 'phytosiderophores'. These compounds are suggested to be involved in the stimulation of iron uptake, and, for example, the biosynthesis of chlorophyll. An example is mugineic acid (Figure 50), which is an amino acid possessing an iron-chelating activity, from root washings of Herdeum uulgare. It is capable of solubilizing Fe(O)(OH) in the pH range 4-9. However, the iron-solubilizing action is strongly inhibited by divalent transition metal ions, suggesting that mu ineic acid does not show selectivity towards Fe"'. Structures have been reported for the Fe"', CoI,fi,,, and CU"1211 complexes. The ligand is hexadentate, with the in-plane donor atoms being the two nitrogen atoms and terminal carboxylate oxygens, with the OH and intermediate carboxylate providing the axial ligands. It is noteworthy that this compound does not resemble the microbial siderophores. This, coupled with the lack of selectivity towards iron, makes a physiological role for mugineic acid seem unlikely. Storage of iron in plants involves phytoferritin.'211a

OH Figure 50 Structure of mugineic acid

62.1.11.6 Transport and Storage of Transition Metals other than Iron in Microorganisms The uptake of nickel by Clostridium pasteurianum and Methanobacterium bryantii has been considered in Section 62.1.7. There is a great deal of information available on the surface binding of transition metals to cell walls. Such cell walls possess many charged groups, such as peptidoglycan, which can contribute carboxylate groups, and teichoic acids, which provide anionic clusters due to the

Coordination Compounds in Biology

681

presence of phosphodiester residues. There is ood evidence that some cations bind much more strongly to isolated cell walls than others. 1212-1'15 This selectivity is being exploited in the use of microorganisms to recover metal ions from solution. Furthermore, it appears likely that binding of metal ions to Gram-negative organisms may well have been involved in the laying down of mineral deposits, serving as a matrix for d e p ~ s i t i o n . ' ~ ' ~ ~ ' ~ * ~ The binding of metal ions to cell surfaces will be the first stage in uptake. There is little known about the mechanisms by which metal ions are then translocated into cells. Highly specific pathways have been found for manganese in all organisms examined, so that organisms are able to concentrate Mn2+even in the presence of much higher levels of Ca2+ and Mg2+,as noted for Bacillus subtilis sporulation (Section 62.1.3.3.4). The E. coli uptake system for MnZ+has a K , value of 0.02 wmol dm-3 and so is readily saturated. It is unaffected by a 100 000-fold excess of Mg2+and Ca2+, and, while Co2+is a competitive inhibitor, other transition metals are not taken up by this pathway."" Mn2+may also be cotransported with citrate, while low affinity mechanisms are also known. Despite the demonstration of these pathways it is difficult to demonstrate manganese requirements for growth in many cases. The metals copper, chromium and cobalt also appear to be essential for growth for some, if not all, microorganisms.'218 Cobalt is an essential requirement for cobalamin-utilizing bacteria, but apart from being an alternative substrate for an Mg2+ transport system, there appears to be no highly specific transport system for Co2+. E. coli has a high affinity uptake system for cyanocobalamin, even though it does not require B12. Cobalamin may thus serve as a source of cobalt.

62.1.11.6.1 Uptake of molybdenum

Molybdate is taken up by nitrogen-fixing organisms, which require it for synthesis of nitrogenase. Uptake is inhibited by sulfate and tungstate. Bacillus thuringiensis,when growing on iron-deficient medium, secretes a yellow, low molecular weight compound that reacts with sodium molybdate to give a molybdenum-peptide complex. This appears to be analogous to a sider~phore.'~'~ Similar results have been found to occur in molybdenum-limited Azotobacter vinelandii.1220The accumulated molybdenum is bound in a storage protein. An equivalent situation holds for tungsten when the organism is grown in the presence of tungstate. The molybdenum storage protein from A. vinelandii is a tetramer of two pairs of different subunits binding a minimum of 15 Mo per tetramer.'220a

62.1.11.6.2 The storage of copper. Microbial metallothioneins

It is now well established that microorganisms synthesize metallothionein or metallothionein-like proteins in response to high concentrations of copper, zinc or cadmium. Fungal metallothioneins seem to contain copper exclusively. A metallothionein from yeast binds 10 Cu+ per molecule of protein of molecular weight 10 OO0.'221 The protein from Neurospora crassa1222contains only 25 amino acid residues, of which seven are cysteine residues. The protein corresponds to the seven cysteine residues of the N-terminal region of mammalian metallothioneins. It is unlike the mammalian metallothioneins in that it binds six copper atoms per molecule, probably in a single The [Cu]/[cysteine] value is very high. A copper metallothionein which contains four Cu' and eight cysteine residues per molecule of protein (molecular weight 4800) has been isolated from Pseudomonas putida growing on 3 mM cadmium synthesizes three cysteine-rich proteins of molecular weight 4000 to 7000, containing four to seven cadmium, zinc and copper atoms per molecule. The use of '13Cd NMR on the major cadmium protein shows it to be related to cadmium metallothionein, but with some significant diff e r e n c e ~ . ' ~ ' ~ ~

62.1.12 DIOXYGEN IN BIOLOGY The oxygen molecule is a substrate for over 200 enzymes.'2zAThese enzymes fall into two main categories, oxygenases and oxidases. Oxygenuses catalyze reactions in which the atoms of the oxygen molecule are incorporated into oxidized organic substrates. In dioxygenases, both oxygen atoms are inserted into the

682

Biological and Medical Aspects

substrate, while in monooxygenases only one oxygen atom is inserted, the other being reduced to water. Oxidases catalyze electron transfer to dioxygen, which is reduced to superoxide, peroxide or water. 1225-1226The best characterized example of an oxidase is cytochrome oxidase, the last member of the respiratory chain, which catalyzes the four-electron reduction of dioxygen to water. Dioxygen, although a powerful oxidizing agent, is kinetically inert, and in general requires activation by binding to a reduced metal centre, which will be either Fe" or Cu'. The requirement for activation in this way allows the possibility of controlling the reactivity of dioxygen in its biological reactions. Activation involves the transfer of electron density from the reduced metal to the oxygen molecule. Control of this process, i.e. the extent to which such electron transfer occurs, will be one factor that dictates the biological function of the hemoprotein or other enzyme, and so determines whether it will function as an oxidase or an oxygenase, or indeed as a reversible transporter of dioxygen. The products of partial reduction of dioxygen (superoxide, hydrogen peroxide and hydroxyl radical) are not subject to kinetic control in their reactions and are, therefore, vigorous oxidizing agents. If they remain bound to the metal, then their reactivity is controlled. However, these species, particularly hydroxyl radical, may be liberated in solution or may be generated free by other reactions of dioxygen. Such reactive oxidizing agents may have a number of harmful effects. In nature, there are defences against these compounds, namely the superoxide dismutases, catalase and the selenium-dependent glutathione peroxidase, 122671228~1229 which all catalyze their disproportionation or reduction to dioxygen. The requirement for dioxygen as the final acceptor of electrons in cellular respiration in aerobic organisms is usually met by diffusion of dioxygen from the surface. In larger organisms this is not feasible, and so systems for the transport and storage of dioxygen have been developed. These include hemoglobin, which has been remarkably well studied. It has been pointed however, that biologically speaking, hemoglobin is only an auxiliary in the process of cell respiration, and that the general significance of cytochrome oxidase greatly exceeds that of hemoglobin. In the following sections, examples of all the classes of reaction given above will be discussed, with emphasis on those that have been well studied. Many of the enzymes involved in these processes are hemoproteins, but non-heme prosthetic groups are important in oxygenases and are also present in some oxidases. Copper is an important metal in this context, and is present in the oxygen transport protein hemocyanin, and in oxidases such as cytochrome oxidase and laccase. Some flavoenzymes are important too, but will not be covered in this discussion.

62.1.12.1

The Chemistry of Dioxygen

This is discussed by Hill and Tew in Chapter 15.2 (Volume 2). The oxygen molecule is a paramagnetic diradical, with the unpaired electrons occupying the degenerate antibonding T orbitals. Addition of electrons to the antibonding orbitals is shown in equation ( 5 9 ) , and results in the formation of superoxide (paramagnetic, bond order If) and peroxide (diamagnetic, bond order 1). Addition of a third electron results in cleavage of the 0-0 bond to give oxide and hydroxyl radical, while complete reduction to water occurs on addition of a fourth electron.

As noted earlier, there is a kinetic restriction on the reactivity of dioxygen, as the one-electron reduction to superoxide is unfavoured. This is demonstrated in the redox potentials shown in Figure 5 1.The triplet ground state of the molecule may also impose spin restrictions on its reactions. Partially reduced dioxygen species are utilized in the reactions of various enzymes. Thus, monooxygenases may use bound superoxide, while cytochrome P-450 utilizes bound peroxide in its hydroxylation r e a ~ t i 0 n s . l ~ ~ ~ The extent of reduction of dioxygen and the use of reactive intermediate species in enzymatic reactions in the hemoproteins are determined largely by the nature of the axial Iigand provided by the protein, and by other more subtle effects. The nature of the axial ligand can determine the extent to which charge is transferred to the dioxygen bound in the second axial position, through the nature of its a-donor and T-acceptor properties. In cytochrome P-450, the axial ligand is a thiol group, which will supply electron density to the iron, and enhance electron transfer to the dioxygen. In the case of hemoglobin the axial group is imidazole, which is a good u-donor and .rr-acceptor, and so can prevent excessive, irreversible transfer of electron density to the dioxygen.

Coordination Compounds in Biology

683

0.270

0.820

Figure 51 Reduction potentials (V) in the reduction of dioxygen at 25 'C and pH 7

62.1.12.2 Multimetal Centres and Concerted Electron Transfer A notable feature in certain oxidases is the presence of multimetal centres. The 0,-reducing site in the oxidases that catalyze the reduction of dioxygen to water always involves two metal ions. This allows the concerted transfer of two electrons to dioxygen, giving peroxide as the first intermediate in the reduction, and avoiding the thermodynamically unfavoured production of superoxide. This has the added advantage of bypassing a toxic species. The reduction of peroxide to water then follows. Examples of such oxidases are cytochrome oxidase, and the copper proteins laccase, ascorbate oxidase and ceruloplasmin, which contain four or eight metal centres. Coupied binuclear copper sites for binding dioxygen are also found in hemocyanin, the dioxygen-binding protein of molluscs and arthropods, and tyrosinase, an enzyme which shows monooxygenase activity in the hydroxylation of monophenols, and oxidase activity in the oxidation of orthophenols. In both cases dioxygen is present as a bound peroxide group. A different type of concerted reaction involves the bacterial cytochrome c peroxide, where two hemes are coupled together, so that hydrogen peroxide undergoes a two-electron reduction to water without the formation of radical species. In a number of dioxygenases, dioxygen is reduced to peroxide by concerted electron transfer from [2Fe-2S] and non-heme Fe'I centres.

62.1.12.3

Transport and Storage of Dioxygen

A number of well-known and much-studied iron- or copper-containing molecules are involved in these processes. Hemoglobin and myoglobin are tetrameric and monomeric hemoproteins, respectively, which bind one molecule of dioxygen at each heme iron( 11) centre. Leghemoglobin is another monomeric hemoprotein found in Rhizobium-infected, nitrogen-fixing root nodules, while erythrocruorin is an extracellular giant hemoglobin found in various invertebrates. Hemerythrin is a non-heme iron protein from certain marine worms, which binds O2 reversibly to give violet oxyhemerythrin with Fe/O, in a 2/1 ratio. Hemocyanin is the copper-containing pigment of crab blood which also binds O2 to give blue oxyhemocyanin, with copper present as Cui' and Cu/02 in a 2/1 ratio. Both hemerythrin and hemocyanin are colourless in the deoxy form. All these carriers or storers of dioxygen can be oxidized to the Fe"' or CUI' 'met' forms, in which they are not active. These carriers provide sites at the transition metal centre for the reversible attachment of dioxygen. In simpler metal complexes these metals would be irreversibly oxidized by 0, to the thermodynamically favoured Fe"' or Cu" state. In some way, therefore, the protein prevents the irreversible oxidation of the metal centre. Myoglobin has a greater affinity for dioxygen than does hemoglobin, which is necessary to allow the transfer of dioxygen from the carrier hemoglobin to myoglobin for use in respiration in muscle cells. The uptake of dioxygen by hemoglobin is complex, due to interactions between its four subunits. Each subunit in hemoglobin contains a heme group and binds dioxygen reversibly. The subunits interact cooperatively, so that uptake of dioxygen by one of them enhances the affinity of the others for dioxygen. The ratio of successive binding constants is 1 :4: 24 :9. This phenomenon is essential for hemoglobin to function efficiently, to ensure that its binding capacity

Biological and Medical Aspects

684

is fully utilized, and that all the bound dioxygen is released under conditions of low dioxygen concentration. Thus hemoglobin will be fully saturated in the lungs and completely deoxygenated in the capillaries. The extent of cooperativity is measured by the Hill coefficient, n. For myoglobin and isolated subunits of hemoglobin, n = 1, while for hemoglobin, n = 2.8. The value for hemerythrin is about 1.2, suggesting that it is a storage system. The origin of the cooperative effect in hemoglobin is a subject of much interest.

62.1.12.3.1 Model compounds for heme d i q g e n carriers

A wide range of transition metal complexes will carry dioxygen reversib1y.I In all cases, electron transfer takes place from the metal centre to the antibonding .rr-orbitals of the bound dioxygen. In some cases the bound dioxygen appears to be present as superoxide and in others as peroxide ) around 1100 and 800 cm-‘ respectively]. [as shown by ~ ( 0 ,values Of greatest interest are those compounds that attempt to model hemoglobin directly. Simple iron(I1) porphyrins are readily autoxidized first to superoxo species, then to p-peroxo dimers and finally to p o x 0 dimers, as represented in equation (60). Bridge formation must be prevented if carrier properties are to be observed. This has been achieved by the use of low temperature and sterically hindered or immobilized iron(I1) porphyrins. Irreversible oxidation is also hindered by the use of hydrophobic environments. In addition, model porphyrins should be five-coordinate to allow the ready binding of 02; this requires that one side should be protected with a hydrophobic structure. Attempts have also been made to investigate the cooperative effect by studying models in which different degrees of strain have been introduced.

I

I

I 1

I

-

L-Fe-0-Fe-L

(60)

Wang‘232developed the first example of an Fe” porphyrin having dioxygen-carrying-properties, n a ~ e l ya heme-imidazole mixture in a polymer matrix. He attributed this largely to the hydrophobic environment produced by the polymer, but another important feature was the immobilization of the heme group. A good number of iron(I1) porphyrins have been synthesized which have special structural features designed to facilitate reversible binding of O2 without competing oxidation. Some of these are represented schematically in Figure 52, namely the ‘capped’ porphyrin,1233‘picket-fence’

Picw- fera

Capped

smppsd / bridged

Tail

- bo=

0 . But

BU’

I c=o

But

I c=o I

YII

\63

L /O OI

O

,

But

Me

“0

\

de

I

I

Et

Figure 52 Some synthetic iron(I1) porphyrins which function as models for dioxygen carriers

Coordination Compounds in Biology

68 5

‘strapped’ porphyrin,1235in which opposite mesa positions are linked, and ‘tail-base’ ~ 0 r p h y r i n . lOther ~ ~ ~ examples are cyclophane hemes,lZ3’ ‘tailed picket fence’,1238‘basket handle’,1239‘gable’,1240‘pocket”241and ‘ f a ~ e - t o - f a c e ”porphyrins. ~~~ Over this range of synthetic analogues and hemoproteins, there is a dramatic variation in O2 affinities (over a range of lo6 in equilibrium constant). This reflects the nature of the binding site pockets, the effects of solvation, local polarity, hydrogen bonding and the effect of the protein. 1237~1243~1244Only a few of these, notably the ‘picket-fence’ and ‘capped’ porphyrins, bind 0, reversibly at room temperature. Several solid complexes with picket-fence porphyrins have been isolated, and the structure has been reported’245 for Fe(TpivPP)(NMeimid)(02) where TpivPP=a,a,a,a-tetrapivaloylamidophenylporphyrin and NMeimid = N-methylimidazole. This shows end-on binding of 0 2 with , an FeOO angle of about 130”.The ~ ( 0frequency ~ ) has been assigned at 1159 cm-’, confirming the Fe3+(02-)formalism. Resonance Raman spectra show the Fe-O2 stretching frequency at 568 cm-’, close to the value in oxyhemoglobin (567 cm-I).

62.1.12.3.2 Hemoglobin

X-Ray diffraction data of high resolution are available for a number of hemoglobins and myoglobins, due to the work of Perutz and Kendrew and their colleagues. Hemoglobin is a tetramer of molecular weight 64450, and is made up of two identical pairs of units (the a- and /3-units) arranged roughly in a tetrahedron. In each subunit the iron is complexed to protoporphyrin IX, with one axial ligand, an imidazole of a histidine residue, giving a five-coordinate species, although there is a loosely bound water molecule near to the sixth position. A large fraction of the protein has the a-helical structure. All the polar groups of the protein are on the outside of the molecule, with a resulting hydrophobic interior that will contribute to the stability of the Fe” with respect to oxidation to the met form. In addition to the hemeimidazole link, there are other heme-protein interactions that will stabilize the structure. Thus, each heme group of oxyhemoglobin is in van der Waals contact with about 60 atoms of the protein. Further stabilization results from the interaction of the carboxyl groups of the porphyrin with basic groups of the protein, and from the interaction between the vinyl groups of the heme and the aromatic amino acids of the hydrophobic interior. The oxygen affinity of hemoglobin varies with the pH of the medium. This is the Bohr effect, and arises from the effect of pH on the interaction between the heme and the ionizable groups of the protein. It appears that the Bohr effect is linked to the presence of a second imidazole group (the distal imidazole group), which is involved in uptake of dioxygen in some indirect way. Uptake of O2 is associated with substantial conformational changes, which are central to an understanding of the cooperative effect. These conformational changes are associated with the reversible interconversion of a tensed (T), low affinity form of hemoglobin into a relaxed (R) high affinity form. The addition of dioxygen to hemoglobin is also associated with a change in the magnetic properties from paramagnetism to diamagnetism. P e r u t ~ ’ ~has ~ ’ an explanation of long standing for the cooperative effect in hemoglobin, which accommodates the magnetic data and is based on an earlier suggestion by Hoard.1248This is represented in Figure 53.

protein

protein

JL

2.07 8.

I

4 0

T quaternary state Figure 53

R quaternary state

The trigger for the cooperative effect in hemoglobin

586

Biological and Medical Aspects

The deoxy form of hemoglobin is high-spin and paramagnetic, with the iron(11) lying above the plane of the heme, due it was suggested to its large size, which prevented its moving into the cavity in the porphyrin. Binding of 0, results in the formation of low-spin iron(II), which is smaller and can move into the plane of the heme. The bound dioxygen (presemt as 0, in this hypothesis) must be in an excited singlet state to give a, diamagnetic species. The resulting movement of the iron atom relative to the plane of the porphyrin is magnified through the interactions in hemoglobin, so that substantial conformational changes occur. This view of the trigger effect is now thought to be at least partly incorrect. There have been several demonstrations of the occurrence in solution and in the solid state of high-spin iron porphyrins which are six coordinate, and which have the Fe" lying in the plane of the porphyrin ring.1249Thus the movement of the iron(I1) in hemoglobin cannot be attributed directly to the high-spin-low-spin electronic rearrangement. Rather, it is to be expected that a five-coordinate complex would have the metal ion displaced out of the basal plane towards the axial ligand. Formation of a six-coordinate complex will remove this effect. There is also some uncertainty on the extent of movement of the iron with respect to the porphyrin plane. The original X-ray data indicated that in deoxyhemoglobin the iron lies some 0.56 A above the porphyrin plane and moves to a position in or close to the plane in oxyhemoglobin. Earlier EXAFS data on deoxyhemoglobin have been interpreted to show that the iron atom lies ~' Perutz et have published a further only 0.2A above the porphyrin ~ 1 a n e . l ~However, assessment of EXAFS data on deoxyhemoglobin, and have compared this with data on a picket-fence porphyrin, where the displacement of the iron from the porphyrin plane was known to be about 0.4 A. It is claimed that there is a similar displacement of the iron in deoxyhemoglobin. It appears, therefore, that there is a substantial movement of the iron with respect to the porphyrin plane on oxygenation of hemoglobin, but its relationship to the cooperative effect awaits elucidation. A further problem centres on the nature of the Fe(0,) group. It is known that dioxygen is bound end-on (in oxymyoglobin) with an FeOO angle of 121". The Pauling model accounts'252 for the diamagnetism of oxyhemoglobin in terms of low-spin Fe" with '0,. However, we is^'^^^ has suggested that oxyhemoglobin involves a superoxide complex of Fe"', Fe"'( 023.The diamagnetism results from strong antiferromagnetic interaction between low-spin iron( 111) and superoxide. In support of Weiss's view, it should be noted that ~ ( 0 ,has ) been assigned at 1107 cm in oxyhemoglobin (clearly in the range expected for superoxide); that superoxide may be displaced from oxyhemoglobin by weak nucleophiles such as chloride; that addition of superoxide ion to met(Fe"') hemoglobin gives oxyhemoglobin; and that there are similarities between oxyhemoglobin and a range of other oxyhemoproteins that do involve Fe1"(02-).These arguments are not conclusive, and support has been expressed for the Fer'(0,) It is, however, of interest to consider coboglobin, formed by recombination of the globin from hemoglobin with the cobalt-substituted protoporphyrin.1256The coboglobins take up 0, reversibly, showing cooperative and Bohr effects. The deoxycoboglobin involves low-spin cobalt(II), so allowing the application of ESR techniques. The hyperfine splitting from the Co" is coupled with that from the nitrogen donor atom from the axial imidazole group, showing that the unpaired electron occupies the dz2 orbital. In the oxy derivative, the nitrogen hyperfine splitting disappears and the separation of the hyperfine lines due to cobalt is reduced to about a third. This shows that the unpaired electron has been largely transferred from the cobalt, giving Co"'(OIJ. Further confirmation comes from ESR studies on CoHb1'02, which show hyperfine splitting due to the nuclear spin of "0 ( I =$), confirming therefore that the unpaired electron density is transferred from the Co" to 0,. It appears that some 60% of the electron density is ) for oxycoboglobin (1106 cm-') is very similar to that for oxyt r a n~f e r red .'~The '~ ~ ( 0 ,mode hemoglobin (1107 cm-'),showing that the argument based on the cobalt derivative may be applied to hemoglobin. ESR studies also show the similarity between oxycoboglobin and oxygenated cobalt porphyrin, and suggest that the protein does not measurably influence the electronic structure of the heme group, although the protein does control the extent to which dioxygen is bound. The magnetic properties of hemoglobin have been reinvestigated, and the claim made1258that it is paramagnetic at room temperature, although such paramagnetism is not shown in the NMR spectrum. These conclusions have now been challenged,'2sy and evidence presented for the complete diamagnetism of both oxy- and carbonmonoxy-hemoglobin. Thus, at present, while it is not possible to give a definitive assessment of the electronic distribution in ox hemoglobin, the balance of current evidence seems to support strongly its formulation as an Fe" superoxide species.

687

Coordination Compounds in Biology

62.1.12.3.3

The coopamtiile effect in hemoglobin

This effect results from the transition between two alternative structures, the tensed (T) low affinity form and the relaxed (R) high affinity form. Binding of dioxygen at one heme must trigger a stereochemical change, which then sets in motion the transition from the T to the R form. The constraints in the protein which oppose this change result in the low affinity of the T structure. This conformation change induces the T-P R conformation change in the remaining vacant 0,-binding sites in the tetramer, so that 0, affinity increases as more 0,is bound and the R conformation is stabilized to a greater extent. The movement of the iron atom into the plane of the porphyrin on binding of dioxygen means that the proximal (axial) histidine has to move possibly up to 0.4 A towards the porphyrin plane. Perutz has suggested that the low affinity for dioxygen showed by the T structure is due to constraints provided by the protein which oppose this movement of the proximal histidine. The stereochemical nature of these constraints has been established'260 by a detailed comparison of the structures of human deoxyhemoglobin, human carbonmonoxyhemoglobin and horse methemoglobin. Transition from the T to the R structure does not involve significant changes in the contact between the components of the aP dimer. The change in quaternary structure results rather from large changes in the association of the a,P1 dimer with the a#, dimer. A number of differences in tertiary structure between T and R forms have been found. The hemes lie in pockets in the protein, into which they are wedged tightly by about 16 side-chains. These allow no movement of the iron atom and its axial histidine ligand relative to the plane of the porphyrin without a change in the tertiary structure of the protein. Furthermore, in the @-subunitsof the T structure, the site for dioxygen is obstructed by the distal valine residue, and this too can only be cleared by a change in tertiary structure. The detailed differences in tertiary structure are shown in the paper of Baldwin and Chothia,1260and are beyond the scope of this account. Particularly important effects are manifested in the orientation of the proximal histidines. In deoxyhemoglobin the imidazole rings are in asymmetric positions with respect to the hemes. The imidazole plane is approximately normal to the heme, but is tilted in its own plane. This feature is intrinsic to the-T structure, and prevents the imidazole group moving towards the porphyrin plane. The iron atom can only move into the plane of the porphyrin if the tilt can be removed, which is accomplished on switch to the R form. Theoretical investigations on the most likely change in conformation of the heme complex on oxygenation and the pathway of its transmission to the subunit boundaries have attributed the restraint to oxygenation in the T structure as the tilt of the proximal histidines.1261 Thus it appears that the affinity for dioxygen i s controlled by restraint of the movement of the heme iron and by steric hindrance of the ligand site. The chemical viability of the first proposal is confirmed by studies on model porphyrins, where the axial imidazole group is hindered to prevent movement towards the plane of the porphyrin. Thus, binding of dioxygen to iron(11) picket-fence porphyrins with N-methylimidazole as an

LA

T state

A

L

L

II R state

Figure 54 Cobalt(I1) gable porphyrins4 model for the cooperative effect

688

Biological and Medical Aspects

axial ligand is similar to that with myoglobin and isolated subunits of hemoglobin. But the use of 1,2-dimethylimidazole as the axial ligand causes a decrease in affinity for dioxygen, down to the level of hemoglobin (T).The ratio of oxygen affinities for N-methyl- and 1,2-dimethyl-imidazole as axial ligands is 80: 1, quite similar to the ratio of 100: 1 observed for R and T hemoglobin. Close contact between the 2-methyl group and the porphyrin nitrogen hinders the approach of the imidazole and its attached iron to the porphyrin plane.'262 Similar conclusions have been found using 'tail-base' hemes, where strain can be built into the chelation arm.1257By shortening the arm linking the imidazole to the porphyrin (see Figure 52), it is possible to introduce a tilt into the axial imidazole group, as found in R state hemoglobin. The presence of such strain lowers ~ . ~ ~ ~ ~ affinity for o Cooperative binding of dioxygen has been demonstrated'264 in cobalt(I1) 'gable' porphyrin model compounds, where structural change at one porphyrin, produced by binding of O,, is manifested in enhanced affinity at the other porphyrin. The gable porphyrin (Figure 54) involves two triphenylporphyrin groups bridged by a common phenyl group, with the axial positions filled by a bridging N,N'-diimidazolylmethane ligand. The deoxy complex has reduced affinity for 0 2 , due to the lowered ability of the cobalt to move into the plane of the porphyrin. However, the resulting in-plane coordination of the cobalt on oxygenation is manifested in a weakened Co-ligand bond in the second porphyrin, so that the bridging N,N'-diimidazolylmethane is lost and free ligand binds from the other side. Dioxygen then binds with higher affinity. The Hill coefficient is estimated to be 1.5. Such experiments as these confirm the feasibility of dioxygen-binding affinity being reduced by steric constraints, and then being increased on removal of the constraint. The R + Ttransition in hemoglobin may be brought about by a number of allosteric effectors,'265 such as , ' H C1-, CO, and inositol hexaphosphate, which function either by strengthening salt bridges between subunits or by forming additional ones. Such salt bridges are broken on oxygenation. The control of the structure of hemoglobin by the use of such effectors has been of value in the comparison of structural features in Rand T hemoglobin derivatives by various instrumental techniques. Criteria have been established for distinguishing between the two quaternary forms by techniques such as Raman, NMR, CD, MCD and UV spectroscopy.'246*1266*'267

C~2.1.12~3.4Monomeric and dimeric hemoglobins

These include mammalian myoglobin, the hemoglobins of the lamprey, sea-snails and clams, and leghemoglobin. The tertiary structure of all these species is similar to that of sperm whale myoglobin, the first protein structure to be solved, and which has now been determined as various liganded forms at resolutions down to 1.5 A.1268 Data on oxymyoglobin show that the 0, is bound end-on with a bent configuration, as originally suggested by Pauling, with an FeOO angle of 121", and that the iron is out of the plane by 0.33 A. The position of the 0, in the heme pocket is fixed, being constrained by steric hindrance from phenylalanine, valine and distal histidine residues. Neutron diffraction data show that the bound dioxygen is hydrogen bonded to the distal histidine.'269The myoglobin from Aplysia Zimacina provides an example of a myoglobin with some different features. Thus, the distal histidine is missing, and the metal site is more accessible to the solvent. Leghernogl~bin,'~~' from nitrogen-fixing nodules in Rhizobium-infected legumes, is a monomeric hemoprotein which shows high affinity for 02.It functions in parr as a buffer to control dioxygen levels in the nodule and so prevent damage to the nitrogenase enzyme. This high affinity results from a combination of an extremely fast 0, on-rate constant ( k ' = 118 dm3 pmol-' s-l) and a moderately slow O2 off-rate constant ( k = 4.4 s-'), giving an equilibrium dissociation constant K = k/ k'= 37 nmol dm-3. The affinity of myoglobin for dioxygen is much lower than this with K = 530 nmol dmP3.In myoglobin, as noted above, there seems to be a tight compartment around the dioxygen-binding site, while in leghemoglobin there appears to an unusually large and accessible distal cavity in which the distal histidine group can move freely. The cavity appears to collapse partly as O2 is bound, so that inward movement of distal histidine towards the bound 0, might, by steric hindrance, account for the relatively slow 0, off-rate and the stability of the Fe-0, bond. The rate of dissociation of dioxygen is dependent upon pH, with a pK value of 5.46.'271 Furthermore, ESR studies'272on cobalt leghemoglobin, in addition to confirming the presence of the proximal histidine group as a show a pH-dependent spectrum with apparent pK about 5.7. In both cases the pH-sensitive group is suggested to be the distal histidine (His-61).

Coordination Compounds in Biology

689

The pH dependence of the rate of dissociation of dioxygen is attributed to a hydrogen bonding nteraction between the bound dioxygen and the protonated distal histidine group. As the pH ncreases, so the rate of dissociation increases as the extent of protonation of the histidine falls. f i e rate varies by a factor of 5 from acid to alkaline limits. In contrast, the rate of dissociation If dioxygen from oxymyoglobin is independent of pH, and there appears to be a hydrogen bond iom the distal histidine at all pH values. The formation of the hydrogen bond in oxyleghemoglobin jecreases the rate of dioxygen dissociation, and so gives rise to a pH dependence of the dioxygen iffinity, which is very unusual for a monomeric hemoglobin. This may well be of some biological iignificance. Some dimeric myoglobins have been isolated from the muscles of the sea-snail Nussa rnutabilis md from the clams Anadora broughtonii and A. senilis. These have no Bohr effect but show zooperative effects for the uptake of 02.i274

62.1.12.3.5 Giant hemoglobins Extracellular hemoglobins are widely distributed among annelids, molluscs and arthropods. They are large proteins, made up in some cases of two superimposed hexagonal discs, some 250 A n diameter, each containing six subunits. The molecular weights of the subunits lie in the range 20 000 to 29 000 for erythrocruorins and from 25 000 to 35 000 for chlorocruorins. Dissociation into subunits occurs mainly at pH values above pH 8. The extent of association of the subunits has effects upon the behaviour of e r y t h r o c r u ~ r i n . ' ~ ~ ~ The structure of erythrocruorin has been refined to 1.4A resolution.'276 The oxy form binds O2 with a bent geometry and an enhanced FeOO angle of 150". Of particular interest is the fact that the Fe atom is displaced slightly more from the plane of the heme in oxy- (-0.3 A) than in deoxy-erythrocruorin (-0.2A). This may be due to the presence of a water molecule that is hydrogen bonded to the bound 0,. Thus spin-state changes are not necessarily linked to pronounced movements of the metal atom.

62.1.12.3.6 The reversible binding of dioxygen by hemoproteins It is now possible to summarize the various contributions made by protein and heme in the binding of dioxygen. The protein provides a five-coordinate deoxy state and a hydrophobic environment which enhances the binding of 0, and prevents irreversible oxidation to the met form. It also provides the constraints on the movement of the axial histidine group relative to the plane of the porphyrin that are linked to the origin of the cooperative effect in hemoglobin. The axial histidine group, in addition to providing heme-globin interaction, also controls electron distribution in the Fe(02) grouping, through its strong rr-donor and wacceptor properties. The role of the distal imidazole group is being clarified. The interaction between this group and the bound 0, species is enhanced by the charge transfer to the dioxygen to give, effectively, bound superoxide ion. This specific interaction probably contributes to the selectivity of the binding site for dioxygen. These observations help to explain why the same heme is used in nature for so many different functions by being attached to different proteins.

62.1.12.3.7 Hemerythrin Hemerythrin is a non-heme iron protein used in the transport and storage of dioxygen by several phyla of marine invertebrates, notably the sipunculid Deoxyhemerythrin is colourless or pale pink, and becomes violet on oxygenation. Hemerythrin from erythrocytes of the worm GolJngiu gouldii has a molecular weight of about 108 000, and is made up of eight identical subunits, each of which contains two iron atoms and binds one oxygen molecule. Examples are known in which hemerythrin is trimeric, dimeric and tetrameric. Myohemerythrin is found in muscle tissue, and is a monomeric species analogous to the hemerythrin subunit. Oxymyohemerythrin is more stable than oxyhemerythrin, as expected from the analogy with hemoglobin and myoglobin. Deoxyhemerythrin contains two FeII centres which react with dioxygen to give two Fer'' centres with bound peroxide. This process occurs with only little cooperativity, the value of the Hill coefficient lying in the range 1.0 to 1.4. Oxyhemerythrin undergoes autoxidation in solution to

Biological and Medical Aspects

690

give the iron(II1) 'met' form, which does not react with 0,. However, the met form combines with a range of anions, where the anion probably binds at the O2site. The oxy + met transformation is accelerated by anions. These reactions are summarized in equation (61). It should also be noted that semi-met forms of hemerythrin can be prepared in which the iron centres are formally Fe"' and Fe". Two relatively stable species can be prepared either by one-electron reduction of methemerythrin or by one-electron oxidation of deoxyhemerythrin. Both species undergo disproportionation, as shown in equation (62).'27* deoxy +o, Fe", Fe"

* Fe"', Fe"', 022---. Fe"', met + Fe"', LL

OXY

o22-

Deoxyhemerythrin is paramagnetic, having two high-spin ( S = 2) Fe'' centres. The presence of iron( 111) centres in oxyhemerythrin is shown by the similarity between the electronic spectra of oxyhemerythrin and the met form. Resonance Raman spectra show a band at 848 cm-' assigned to v(O,), as shown by studies with 1802,which is appropriate for a peroxo group. Use of 160-'80 shows that the two atoms of the peroxide have different environment^,'^^' while the iron(II1) centres have different environments also, as indicated by the Mossbauer spectra of oxyhemerythrin. These observations are consistent with a structure for oxyhemerythrin in which the peroxide is bound end-on to one iron centre only. Earlier X-ray structures have given conflicting data on the nature of the iron binuclear site'28"-'282 for metazido- and methydroxo-hemerythrin, but the situation has been clarified by a determination resolution of the structure of metazidohemerythrin from 7'hemiste d y s c r i t ~ r n . ' All ~~~ at 2.2 three structures are given in Figure 55, while the subunit arrangement is shown schematically in Figure 56.

,/ ,

,

His-73

His-106

His-77

His-73 ,His-77 Fe

His-101

\

//

H20

Asp-106 H20 Glu-58 \\ /

\

His-25

Metazidomyohemetythrin from T.zostericoln

His-25

/

/7\

'\

Fe

'His-54 Tyr-109

Methydroxohemerythrin from 7: dyscn'curn

His-73 ,His77

AsP-LO~,,?, ,G1~-58

Fe

fN3

Fe

/ L \His-54

Tyr- 1 14

His-101,

y '

\ 'His-54

His-25

Metazidohemerythrin from T. dyscriturn

Figure 55 Structures for the binuclear iron centre in hemerythrin

Figure 56 The arrangement of the subunits in hemerythrin

The most recent structure of the metazidohemerythrin (Figure 55c) shows that each iron is octahedrally coordinated, with the iron centres bridged by carboxylate groups from Asp-106 and Glu-58 and by an oxo group. The remaining coordination positions are filled by three histidine side-chains in one case and two histidines and the azide ion in the other. The azide group probably fills the 0,-binding site in the oxyhemerythrin, again suggesting that O2 is bound end-on. Methydroxohemerythrin has an empty coordination site, so that one iron is five-coordinate. The structure of the azide complex was predicted previously on spectroscopic gro~nds.''*~There is some uncertainty over the nature of the bridging group, but spectroscopic and magnetic evidence, which shows antiferromagnetic interactions between the two iron( 111) species, is consistent with an oxo bridge.'2R5It should also be noted that similar Fe, sites are found in other proteins, where

Coordination Compounds in Biology

69 1

there is evidence for oxo bridges. These include ribonucleotide reductase and purple acid phosphatase, discussed in Section 62.1.5.5. The postulate of an oxo bridge is also supported by EXAFS data,'286*'287 which show that the oxo bridge is present in oxyhemerythrin and metazidohemerythrin, but possibly absent in hemerythrin. This work also shows that the peroxide group in oxyhemerythrin is bound at the azide site in metazidohemerythrin, and that one of the iron centres in deoxyhemerythrin may be five-coordinate. Prior to these structural studies, work using chemically modified proteins had shown the requirement for at least four histidine ligands per subunit together with a glutamic acid residue and Tyr-109 (Tyr-114 in met). The most recent X-ray work has eliminated the possibility that Tyr-109 is a ligand, a possibiiity proposed in earlier X-ray work. Other work on chemical modification of residues has been reviewed.'277 The quaternary structure (Figure 56) has been described as a square doughnut, 75 X 75 X 40 A3, with an open central cavity 20A in diameter. Each iron site is about 28-30A from its two neighbours in the same layer. The amino acid residues involved in the hydrogen bonds and salt bridges between subunits have been identified in T. dyscritum. These residues are largely absent from myohemerythrin, which is in accord with its monomeric character. The arrangements of subunits from other oligomeric forms of hemerythrin have been established. Thus the trimer subunits from Siphonosoma funafuti are arranged in a triangle.'288

62.1.12.3.8 Hemocyanin

The hemocyanins bind and transport dioxygen in many species of the Mollusca and Arthropoda phyla. They are very large molecules, and are often made up of more than 100 subunits with molecular weights between 25 000 and 75 000. The hemocyanin of Loligo pealei has a molecular weight of 3 750 000. Some smaller hemocyanins are known. Thus the horse shoe crab Limulus polyphemus has a unique 48-subunit *hemocyanin made up of eight immunochemically distinct subunits.'289 Two related Chelicerata, the scorpion Androctonus and the spider EurypeEma have 24-subunit hemocyanins that crosslink immunochemically with Limulus hemocyanin, and yet do not give the 48-subunit oligomer. The 48-subunit oligomer in Limulus hemocyanin results from specific stabilization by calcium ions, which must crosslink negative groups to prevent dissociation. If Ca" is removed, then the hemocyanin undergoes rapid dissociation to monomers uia 24- and 12-subunit intermediates. The factors that govern the assembly and stabilization of such large molecules as the hemocyanins and the erythrocruorins are of general relevance to the function of biopolymers. Each subunit in hemocyanins contains a pair of copper atoms which bind one oxygen molecule. The deoxy protein is colourless while the oxy form is blue, showing the oxidation of Cu' to Cu" on oxygenation. In addition to bands at 440, 570 and 700nm, an intense band is observed at 341 nm in the spectrum of oxyhemocyanin, which is assigned to charge transfer between the two copper centres or between Cu" and peroxide. Resonance Raman spectra of several oxyhemocyanins show v ( O z ) at about 745cm-' indicative of a bound peroxo group. The oxyhemocyanin is completely diamagnetic and ESR-silent, showing antiferromagnetic interaction between the copper centres (-2J = 550 cm-I). EXAFS data on Busycon canuliculutum oxyhemocyanin have been interpreted'290 in terms of a model having two copper atoms, separated by 3.67 A, each bound to three histidine ligands, and bridged by the peroxo group and an atom from a protein ligand, possibly a phenolate from tyrosine. Analysis of EXAFS data on oxyhemocyanin of Megathura crenulata and the a- and only two histidine ligands and two low Z atoms (N or /3-components of Helix pornaria 0)per copper, with Cu-Cu separation of 3.55 A. The copper is present in approximately square planar geometry, and one of the low Z atoms will be the peroxide oxygen. The model for the oxyhemocyanin site is shown in Figure 57. The possibility of bridging peroxide bound through one oxygen atom only is eliminated by resonance Raman measurements which show the peroxide atoms to be equivalent. The EXAFS results for d e o ~ y h e m o c y a n i n 'confirm ~~~ that copper i s coordinated to two histidines each, and provide no evidence for C u . . C u interaction within 4 A and some evidence for a Cu...Cu distance of 5.4 A. The coordination number of copper (two) is low for a metalloprotein, but this does allow facile binding of dioxygen. On oxygenation, the two copper atoms become oxidized to Cu", move towards each other (to 3.55 A), and become bridged by the low 2 ligand atom, with the bridging peroxide bound at the binuclear site. Thus oxygenation must involve some reorganization of the protein. CCC6-W

692

Biological and Medical Aspects

n - 3.55 A

1

Figure 57 Model for the active site in oxyhemocyanin

The results of other studies have contributed to the elucidation of the active site of hemocyanin. Thus, imidazole groups have been suggested to be ligands by the use of UV-vis and Raman spectroscopy, acid-base titrations and photooxidation studies. Methemocyanin is formed on oxidation of hemocyanin, in which case an exogenous bridging ligand will replace peroxide. These bridging ligands mediate the antiferromagnetic interaction. The nature of the endogenous bridging ligand has been the subject of much discussion. The possibility that it is an oxygen atom would be appropriate for strong antiferromagnetic interaction, but a tyrosine group'293now seems unlikely on a number of grounds. An alternative group is alkoxide from serine or threonine residues,'294 although their pK, values (>16) are quite high. The binuclear site in hemocyanin seems to be very similar to that in tyrosinase (Section 62.1.12.11.2). A detailed analysis of their electronic structures has been prepared;O6 particular attention being paid to the two intense spectral features which dominate the UV-vis region, namely the bands at 347 nm ( E = 20 000 dm3mol-' cm-') and 570 nm ( E L- 1000). Eiucidation of these unique spectral features is complicated considerably by the binuclear nature of the site. One approach has been to prepare a series of hemocyanin (and tyrosinase) active site derivatives, beginning with the simplest derivative, met-apo, where one copper has been removed and the other oxidized to Cu", and then preparing others of increasing complexity such as the mixed valence 'half met' forms. The two dominant spectral features are not present in methemocyanin, suggesting they can be assigned as 0;- + Cu" charge transfer. Resonance Raman measurements on oxyhemocyanin with laser excitation into either of the two intense bands produce an enhanced the peak shifts to -710 cm-', Raman peak at -750 cm-'. When '*O2is substituted for 1602, confirming the vibration to be due to ~ ( 0for~peroxide. ) This confirms that the absorption bands are 0;- + Cu" charge-transfer transitions. In the CD spectrum of oxyhemocyanin an additional 0;- -D Cu" charge-transfer transition is seen. The observation of at least three 0;- + Cu" CT bands requires that peroxide bridges the two coppers. The band at 440 rim may well be (endogenous bridge) 3 Cu" charge transfer. The binding of CO to hemocyanin is of interest. The two-coordinate copper centres in deoxyhemocyanin might each be expected to bind CO. However, some two-coordinate complexes of CUI with heterocyclic donor ligands show lack of reactivity towards CO in the absence of additionai ligands, probably because the metal does not receive enough electron density to 'back bond' to C0.1295

62.1.12.4 Cytochrome Oxidase Cytochrome oxidase is a membrane-bound enzyme that accepts electrons from cytochrome cy and transfers them to dioxygen in the respiratory chain. The dioxygen undergoes a four-electron reduction to water. The importance of cytochrome oxidase amongst 02-utilizing enzymes is emphasized by the fact that it is probably responsible for more than 90% of the total consumption of dioxygen by living organisms. Cytochrome oxidase also serves as a proton pump, so that the process of electron transfer is associated with the vectorial transfer of protons across the membrane, and thus contributes to the establishment of the proton gradient which is used to drive the synthesis of ATP. Cytochrome oxidase is located in the inner mitochondrial membrane of animal, plant and yeast cells (the eukaryotes) and in the cell membrane of prokaryotes. The arrangement is represented schematically in Figure 58. The complexity of cytochrome oxidase and the problems associated with its solubilization from the membrane have presented great obstacles to the elucidation of the structure and mechanism of the enzyme, but its importance has resulted in an enormous literature, which has been reviewed f r e q ~ e n t l y . ' ~ ~ ~ - ' ~ ~ ~ Mammalian cytochrome oxidase has attracted particular attention. It contains four redox-active transition metal centres, two type e cytochromes ( Q and u3) and two copper ions. Other oxidases

Coordination Compounds in Biology

4H+

Figure 58

693

2H.O

Schematic representation of the arrangement of the redox centres in cytochrome oxidase

that reduce dioxygen to water also have four redox metal centres; thus laccase contains four copper centres, appropriate for a four-electron reduction reaction. It has been suggested that cytochrome a3and one of the copper centres are antiferromagnetically coupled as they are both undetectable by ESR techniques in the oxidized enzyme. It is thought that the cytochrome a3and the copper (CU,) share a common ligand, which has been suggested by various authors to be oxide, hydroride, sulfide and imidazolate. The other centres, cytochrome a and CuA, are isolated and may be detected by ESR in the Fe'" and Cull states. Cytochrome a is usually low-spin and six-coordinate, and will therefore be involved in electron transfer. Cytochrome a3 is high-spin and five-coordinate, and is therefore able to bind small molecules such as O2and CO in the sixth, axial position when present as Fe". It appears therefore that cytochrome a3is the component of cytochrome oxidase that binds dioxygen, and it is possible that at some stage the dioxygen species would bridge the u3 iron and CU,. The close proximity of the u3 and CU, centres has been demonstrated for the low-spin oxidized cyanide complex, where the cyanide acts as a bridging ligand between a3 and CU,. The temperature dependence of the paramagnetic susceptibility of oxidized cytochrome oxidase over the ranges 200-7 and 4.2-1.5 has been interpreted in terms of a model with two isolated S = f centres (low-spin Fe"' and CUI') and a spin-coupled S = 2 centre (high-spin Fe"'. CUI'). Additional information has been provided by MCD on the oxidized low-spin complex with cyanide bound to the a3 heme. This work shows that the cyanide is a bridging ligand to copper. Saturation curves have been observed for the u3 Soret MCD signal, showing that the a,-CN-Cu, centre cannot have a diamagnetic ground state. The data suggest an S = 1 ground state, which has been interpreted1303in terms of ferromagnetic coupling between CU, (S= f)and a3-CN- (S = i).The alternative of Fe" and Cu' seems unlikely. It is difficult to reconcile these two sets of results, but the MCD work seems to be clear cut in its essential conclusion of an S = 1 ground state. Near-IR CD and MCD experiments also suggest that it is unlikely that a blue copper centre is present in cytochrome oxidase. Some considerations that support the opposite view will be given later.

62.1.12.4.3

Structure and organization

As noted, cytochrome oxidase spans the inner mitochondrial membrane. One side is exposed to the outer surface and accepts electrons from ferrocytochrome c. The inside face contains the cytochrome a3-CuB pair. Mitochondrial cytochrome oxidase is reported to contain from six to twelve different subunits. It is generally accepted that the three largest polypeptides are true subunits of the enzyme, but it is still not certain how many of the other subunits fall into this class. The two largest subunits are suggested to be associated with the redox centres, and the third (111) to have a specific role in the translocation of protons. Subunit 111 is not involved in the catalysis of electron transfer. Amino acid sequencing has been carried out for a number of these subunits. Subunit 11of beef heart Cytochrome oxidase has a region that considerable homology with the 'blue' electron-transfer copper protein plastocyanin, and contains two cysteine, one histidine and one methionine residues which could represent a binding site for copper. An interesting comparison may be made at this stage with certain bacterial which closely resemble mitochondrial cytochrome oxidase. Many of these appear to have a much simpler subunit

694

Biological and Medical Aspects

composition. The cytochrome oxidase from Paracoccus denitrificuns appears to have only two subunits (molecular weights 45 000 and 28 000), heme a, heme a3 and two copper ions.'306These subunits resemble the subunits I and I1 of the mitochondrial enzyme, showing similar properties with respect to activity and spectra. This suggests strongly that subunits I and I1 of the mitochondrial enzyme contain all four redox centres. Subunit I1 provides the binding site for cytochrome c, and probably contains heme a in addition to a copper site. There is some evidence that heme a3 is present in subunit The interaction of cytochrome oxidase with lipids is essential for its biological function, and has therefore been widely studied in order to explore the topological arrangement of the subunits in membra ne^.'^^ Subunit TIT is suggested to span the membrane, as required for the role of Ill in proton translocation, subunit T is buried in the membrane and 11 extends into the cytosol. The orientation of the heme groups with respect to the membrane has been studied using orientated multilayers, and measuring the angular dependence of the ESR spectrum. It appears that the heme planes are perpendicular to the membrane Some progress has been made in the characterization of the redox centres. The presence of a potential type 1 blue copper site in subunit I is in accord with EXAFS data that have demonstrated Cu-S interactions, in particular the possibility of two sulfur atoms bound to It seems possible therefore that CuA is a type 1 copper, typical of copper electron-transfer proteins. The nature of CuB is less certain: ESR parameters indicate that CuBis similar to type 3 copper, which occurs pairwise in copper oxidases as the 0,-binding site. An imidazole group of histidine has been identified conclusively as an axial ligand in Cytochrome a3. The nature of the aS-CuBbridge is still unclear. EXAFS data have been interpreted"" in terms of a bridging thiolate li and, but this is not yet generally accepted. Bridging imidazolate has now been eliminated, 1311*1'12 partly because information from model compounds shows that such a bridge would not mediate the strong antiferromagnetic coupling ( - J > 200 crn-') found for cytochrome a3 and CuB. Instead an oxo bridge has been ~ u g g e s te d ,'~formed ~' possibly from the dioxygen. It should be noted that these difficulties in characterizing the nature of the bridging group between u3 and CuBhave led to the suggestion that the two centres are separate from each other, and that the magnetic properties of the oxidized enzyme result from the presence of Fe'" and CUI rather than coupled Fe"' and Cu". However, ESR studies'313 on the reaction of nitric oxide with cytochrome oxidase show that the oxidized enzyme contains cytochrome u3 (Fe3+) and CuBZcat an estimated distance of 3.4 A from each other. EXAFS data on the carbonmonoxy cytochrome u3 indicateI3l0 that the first shell structure is identical to that of hemoglobin, a point of some mechanistic interest.

62.1.12.4.2 Reaction mechanism In general terms, it is assumed that the dioxygen binds to the reduced enzyme ai the a3 CuB site (with K , values in the range 0.3-3 pmol dmP3),and is then reduced to a bound peroxo group, and subsequently to water. The details of the mechanism are the subject of much uncertainty. The reaction of dioxygen with fully reduced cytochrome oxidase is very fast, but low temperature techniques involving flash photolysis of carbonmonoxy derivatives have allowed these reactions to be studied, and individual steps to be identified. In these experiments, the fully reduced or 'mixed valence' enzyme is saturated with carbon monoxide to give the Fe" a3-C0 complex. The suspension is then saturated with dioxygen, which now cannot react with the cytochrome oxidase, and further cooled to an appropriate temperature (between -130 and -60 "C). The reaction with O2 is initiated by an intense flash of light which photolyzes the Fe-CO bond. The O2 may then react at the empty coordination site on the cytochrome u 3 ,and the reaction may be followed by various techniques. This procedure was developed by Chance and his coworkers. The intermediates may represent species with increasing oxidation levels for the metal centres, or they may represent species where erectronic redistribution has occurred within a particular intermediate. Useful parallel experiments involve studies of the reaction of O2 with the 'mixed valence' oxidase where the cytochrome a and Cu, are oxidized, and the other two are reduced. Again a number of intermediate species are seen in the reaction with dioxygen. It is clearly of interest to compare these two situations. The overall picture is further complicated in that various forms of the enzyme may be prepared. The resting or fully oxidized form of the enzyme obtained in normal preparations appears to differ from the form obtained by reducing the resting enzyme and reoxidizing with molecular

Coordination Compounds in Biology

695

oxygen. The latter species is more active, and is termed the 'pulsed oxidase'. It has been suggested that the pulsed form contains a bridging oxo group. It is beyond the scope of this discussion to consider the complex interrelationships between these species in detail, but a clear account is given by Wikstrom et al.1299 Instead, a summary of the intermediates formed in the reduction of dioxygen will be presented. It is probable that not all of the many intermediates observed are of mechanistic significance. On photolysis of the reduced or mixed valence enzyme, the 590 nm band of the a,-CO species disappears and a new band appears at 612 nm, due to the generation of five-coordinate cytochrome a3'+. If dioxygen is present, and at a temperature higher than about -100 "C, then a new s ecies (A) is formed, which cannot be photolyzed, and which is spectrally similar to the a$-CO compound. It is probable that this involves O2bound to the cytochrome a 3 , and does not involve bridging a3 and Cu,. This species would approximate to oxyhemoglobin, and would involve some charge transfer to bound dioxygen. The subsequent reactions differ in the cases of the reduced and mixed valence enzyme. The reaction of the mixed valence compound may well be expected to be simpler than that of the fully reduced enzyme, as electron transfer from cytochrome a and CuA cannot be involved. As the spectrum due to compound A from the mixed valence enzyme disappears, so a new band forms with maximum intensity at 605-610nm. This is compound C, which appears to involve peroxo-bridged Fe'" and Cu". At one time it was thought that cytochrome u3 in compound C was Fe", but this view seems unlikely. Another possibility involves peroxide bridging FeIV and CUI. These suggestions have all been assessed in the light of UV-vis absorption and ESR spectroscopy. They are shown in Figure 59.

The reaction of the fully reduced enzyme with O2 does not follow the pathway described above. Instead, loss of compound Ais followed by the production of compound B through the intermediate formation of another compound (11), which may be equivalent to the peroxo cornpound C, but having reduced cytochrome a and CuA. The transient nature of this species I1 compared to compound C would result from the possibility of electron transfer from cytochrome II and CuA, to give compound B. The fact that cytochrome a and Cu, are at least partially oxidized in compound B is shown by the appearance of the ESR signal (in 30-50% intensity) characteristic of the fully oxidized enzyme. Intermediates beyond compound B in the oxidation sequence have also been described. In one of these, the CuB2+is ESR-detectable.

Fe"'-02--Cu" Figure 60 A scheme for the reduction of O2 at the a,/CuB Centre of cytochrome oxidase

696

Biological and Medical Aspects

Some insight into the understanding of these later intermediates comes from the observation that the fully oxidized enzyme may undergo a one-electron oxidation reaction, in which the electron donor is probably water and the acceptor ferricytochrome c. The overall product would be a one-electron oxidation product of the fully oxidized centre plus water. Presumably, oneelectron reactions in the opposite direction can occur. The transfer of one electron from cytochrome a to the a,/CuB centre in compound c , plus one electron from a3 or CU, will allow a second concerted two-electron reaction with the formation of (Fe'"=O2-Cu~-0H-), and ESR visibility of the copper. In the next stage of the reaction antiferromagnetic coupling would be reintroduced. A mechanism has been put forward by Wikstrom et ul.12gyin which reduction of O2 occurs in two concerted two-electron steps, but with electron transfer to the a,/CuB centre occurring by discrete one-electron steps. The latter feature is suggested to allow the coupling of the redox reactions of cytochrome a to proton translocation. They suggest that the two possible forms of the fully oxidized enzyme correspond to the 'pulsed' (ferryl iron) and resting (p-oxo) states of the enzyme (Figure 60). The addition of the third electron to the bridging peroxide species will result in a breaking of the 0-0 bond and the formation of hydroxide. Rapid electron transfer from the Fe"' to the remaining fragment gives oxide rather than 0-, and effects a two-electron reduction of peroxide. The formation of a ferryl ion is consistent with the ESR signal for the CuB2+ion, which shows interaction with a nearby paramagnetic ion.1313The formulation Fe"'(O-) would be diamagnetic, while the ferryl ion would be paramagnetic.

62.1.12.4.3

Electron-transfer pathways in cytochrome oxidase

The present view is that cytochrome a is the acceptor of electrons from cytochrome c, but that a simple linear electron-transfer sequence from cytochrome a to CuAand then to the cytochrome a,/Cu, centre is unlikely. Instead the sequence shown in equation (63) holds, where cytochrome a is in rapid equilibrium with CuA.These views depend largely upon pre-steady-state kinetics of . the redox half reactions of the enzyme with its two substrates, ferrocytochrome c and 0 2 However, these conclusions are not in accord with kinetic studies1314under conditions when both substrates are bound to the enzyme, and which show maximal rates of electron transfer from cytochrome c to 0 2 .In particular some of the cytochrome c is oxidized at a faster rate than a metal centre in the oxidase. In contrast, at high ionic strength conditions, where the cytochrome c and the cytochrome oxidase are mainly dissociated, oxidation of cytochrome c occurs only slowly following the complete oxidation of the oxidase. These results for the fast oxidation of cytochrome c have been interpreted in terms of direct electron transfer from cytochrome c to the bridged peroxo intermediate involving u33+and CuB2+,or to a two-electron transfer to O2 from cytochromes a and a3 during the initial phase of the reaction. c

+

(a)

-

(CUB,%)

0 2

(63)

lb

cu,

The interaction of cytochrome oxidase with cytochrome c has been considered in Section 62.1.5.2.5.

62.1.12.5

Bacterial Cytochrome Oxidases

Bacterial cytochrome oxidases offer an interesting area of study that complements well that of mitochondrial cytochrome oxidase. In some cases they are readily solubilized from the membrane; they tend to have fewer subunits than the mitochondrial oxidase; and they form stable intermediates during the reaction with dioxygen. They often do not have a role in energy transduction, a fact reflected in their simpler structures. P00le'~'~has reviewed the bacterial cytochrome oxidases, and has drawn attention to features which are not present in the mitochondrial enzyme, and which reflect the metabolic diversity and adaptability of bacteria. These are: ( I ) the synthesis of the oxidases is controlled dramatically by the prevailing environmental conditions; (2) some oxidases are multifunctional, and may use electron acceptors other than dioxygen; (3) more than one type of oxidase may be present, each terminating a branched electron-transfer pathway.

Coordination Compounds in Biology

697

62.1.12.5.1 Otochrome oxidases of the aa, type These are similar to the mitochondrial enzyme, apart from the simpler subunit composition. EXAFS studies1310on the oxidases from Paracoccus denitrijicans and Thermus thermophilus have confirmed the metal centres to be similar to those in the mitochondrial enzyme. In the latter case, the ESR silent cytochrome a3 in the reduced enzyme has been shown by the use of Mossbauer spectroscopy to involve five-coordinate, high-spin Fe”.1315On addition of cyanide to the oxidized enzyme, there is evidence for a low-spin FelI1 complex with cyanide, and for coupling to the CuB centre. One of the advantages of microbial systems is that 57Fe-enriched enzymes may readily be obtained by growing the organism on a medium supplemented with 57Fe. A non-covalent complex that contains cytochromes c, and uaj and has a molecular weight of about 93 000 has been isolated from T. thermophilus. This contains two proteins only, cytochrome c1and a protein of molecular weight about 55 000 which is though to be a single subunit cytochrome oxidase containing the two cytochromes and two copper The use of a thermophilic organism often allows membrane-bound enzymes to be extracted, as their stability is greater. It is of great interest that a single subunit cytochrome oxidase has been isolated, although the two peptides in this complex cannot be separated from each other. Preliminary studies on the cytochrome c1aa3 complex have been reported.13” Cytochrome caa3 has also been purified from the thermophile PS3, and appears to be a three-subunit ~ o r n p l e x . ’ ~ ’ ” , ’ ~ ~ ~ 62.1.22.5.2 Cytochrome o This cytochrome oxidase (‘0’is an abbreviation for oxidase) is widely distributed and may completely or partly replace cytochrome oxidase aa3 under conditions where the supply of dioxygen is limited. It is a membrane-bound enzyme which has proved difficult to purify, and whose spectral characteristics are those of b cytochromes. Its identity is usually confirmed by observation of the distinctive spectral features of its complex with carbon monoxide. The cytochrome o from Azotobacter vinelandii is reported to consist of one polypeptide of molecular weight 2s 000 with two identical heme components. It has also been isolated from the thermophile PS3,I3l9Escherichia coli, Vitreoscilla, Pseudomonas aeruginosa, and Rhodopseudornonas spp. The enzyme from Vitreoscilla consists of two identical polypeptides of molecular weight 13 000 and two moles of protoheme IX. A cytochrome b562-0 complex from E. coli contains two peptides and, strangely, copper.’320 Cytochrome o binds dioxygen with K , values in the range 1.8-6.5 pmol dm-3. Addition of dioxygen to the reduced cytochrome o from Acetobacter suboxydans gave a species assumed to be a stable oxygenated intermediate, which on further oxidation gave the oxidized form. Stable intermediates are also given by cytochrome o from VitreosciZla The IR spectrum of the oxygenated form absorptions at 1134 cm-’ due to bound dioxygen. The assignment was confirmed by the use of ‘*02, and the observation of a shift in frequency to 1078cm-I. This value is appropriate for a bound superoxide group, i.e. Fe’” (02-), and is similar to v ( 0 J for oxyhemoglobin. Formation of the oxidized enzyme from the reduced form requires 0.5 mol O2 per mol cytochrome, and there is spectral evidence for the formation of an intermediate prior to the oxygenated form. The tight binding of superoxide in the oxygenated form limits any toxic effects. The final product is hydrogen peroxide, as shown in equation (64). bZ+.b2+

b2+(02)b*+ + bZ+(02-)b3+ + b3+.b3++02-

(64)

62.1.12.5.3 Cytochrome d This is found in a range of bacteria,I3O5particularly Gram-negative heterotrophic bacteria, and often coexists with other oxidases. The production of cytochrome d occurs under OJimited or sulfate-limited conditions, and is found in obligate anaerobes. Cytochrome d has a chlorin The complex prosthetic group. It occurs in E. coli as a complex with cytochromes b,,, and a1.1322 with cytochrome b5>*has been Cytochrome d has a remarkable affinity for dioxygen, with K,,, values in the range 0.0180.35 Fmol drnp3. Thus, it could be synthesized by anaerobes as a defence mechanism against 0 2 , which would be scavenged by cytochrome d. It will also be synthesized at low dioxygen levels

698

Biological and Medical Aspects

by aerobes, as it will be able to maintain respiration under low-0, conditions. Cytochrome d also appears to confer lack of sensitivity to cyanide. Normally cyanide binds at the open coordination position of five-coordinate hemes. In this case, the affinity for O2 is so high that cyanide cannot compete for the site. Thus Achromobacter in the presence of 1 mmol dm-’ KCN shows adaptive growth and respiration that may be correlated to the increased appearance of cytochromes a , and d, particularly the latter oxidase. Low temperature trapping techniques involving photolysis of the CO-liganded cytochrome d in the presence of dioxygen have produced evidence for an intermediate species which has been suggested to be an oxygenated intermediate analogous to o ~ y h e m o g l o b i n . ’This ~ ~ ~is confirmed by resonance Raman experiments on the solubilized and aerated enzyme from E. coli, which show features at 1078-1105 cm-’ that may be assigned to v(0,) of the oxygenated ~ x i d a s e . ’ ’ ~ ~ 62.1.12.5.4 Cytochrome cd,

The cd, oxidase from Pseudomonas aeruginosa has been most extensively studied. Both hemes are low-spin in the oxidized protein, while heme c is diamagnetic and d , is high-spin, probably five-coordinate, in the reduced state.1326The enzyme is associated with the inner surface of the cytoplasmic membrane, but is readily solubilized, an unusual feature. Molecular weight determinations suggest that the enzyme is a dimer, each subunit having c and d , hemes. It is suggested that all four hemes are at one end of the dimer, a suggestion that is of important mechanistic significance.’”’ The cd, enzyme may operate physiologically with compounds other than dioxygen as the terminai electron acceptor. Thus its dominant role is that of a nitrite reductase, nitrite being reduced to NO. Binding of O2 is suggested to follow two pathways, depending upon O2 tension. In the high oxygen tension pathway, two oxygen molecules bind to the enzyme and are reduced to peroxide, which then follows a catalase-like decomposition reaction. In the low-oxygen-tension pathway, a direct reduction of the peroxide to water occurs.1328 62.1.12.5.5 Terminal oxidases in the aerobic carboxydobacteria

Terminal oxidases are inhibited by small molecules and ions such as carbon monoxide, nitric oxide, nitrite and cyanide which bind tightly at the dioxygen site. It is noteworthy, therefore, that the carboxydobacteria are able to use carbon monoxide as the sole carbon and energy source, with dioxygen as the final electron acceptor. As noted earlier, the carbon monoxide oxidase from these bacteria is a molybdoprotein in which the molybdenum is bound by a pterin group. The remarkable lack of toxicity of carbon monoxide to these aerobic organisms is attributed to the presence of novel cytochrome o (cyt b,,,) as a terminal oxidase, which is suggested to be insensitive to carbon rn~noxide.’’~~ Nevertheless, it is difficult to accept that carbon monoxide cannot compete with dioxygen for the dioxygen-binding site. The question of selective binding of CO and 0, to heme proteins and various model porphyrins has been of some interest recently, in an endeavour to pinpoint factors which may contribute to selectivity. Normally carbon monoxide gives a linear Fe-CO group, while dioxygen gives a bent Fe-0, structure. It is well known that the Fe-CO group is forced to adopt a non-linear geometry in hemoproteins through interaction with various distal groups. This may lower the affinity for CO and so provide some protection for the site against CO. Attempts have been made’33u+’331 to synthesize model porphyrin compounds with pockets to accommodate bent Fe-02 groups, and which provide steric hindrance against the preferred linear geometry of Fe-CO. This work has led to the conclusion that there is substantially lowered affinity for CO in the sterically hindered porphyrins. Kinetic studies on the binding of 0, and CO by cyclophane and cofacial d i p ~ r p h y r i n s have ’ ~ ~ ~shown that distal steric hindrance affects the ligand association rate constant and has no effect on the dissociation rates. The association rates showed a differentiation factor against binding of CO in the range 3 to 8. However, these are not very large factors in view of the substantial differences between these models and the native oxidases. It has also been that changes in the electronic properties of the heme group had little effect on affinity for CO. At this stage, therefore, it is reasonable to say that distal steric effects could provide some protection against carbon monoxide toxicity towards terminal heme oxidases. It is certainly not possible to account for the remarkable resistance of the carboxydobacteria to the toxic effects of carbon monoxide. One conclusion is that the terminal oxidase in the carboxydobacterja is not a hemoprotein.

Coordination Compounds in Biology 62.1.12.6

699

Blue Copper Oxidases

In the discussion of the biochemistry of copper in Section 62.1.8 it was noted that three types of copper exist in copper enzymes. These are: type 1 ('blue' copper centres); type 2 ('normal' copper centres); and type 3 (which occur as coupled pairs). All three classes are present in the blue copper oxidases laccase, ascorbate oxidase and ceruloplasmin. Laccase contains four copper ions per molecule, and the other two contain eight copper ions per molecule. In all cases oxidation of substrate is linked to the four-electron reduction of dioxygen to water. Unlike cytochrome oxidase, these are water-soluble enzymes, and so are convenient systems for studying the problems of multielectron redox reactions. The type 3 pair of copper centres constitutes the 0,-reducing sites in these enzymes, and provides a two-electron pathway to peroxide, bypassing the formation of superoxide. Laccase also contains one type 1 and one type 2 centre. While ascorbate oxidase contains eight copper ions per molecule, so far ESR and analysis data have led to the identification of type 1 (two), type 2 (two) and type 3 (four) copper centres. At present, understanding of these blue oxidases is limited, but progress is being made906 towards the characterization of the metal sites, and certain features in the mechanism seem wet1 established. This is particularly true for the laccase from Rhus vernicifera. The use of optical, CD, MCD and ESR spectroscopy by various has led to a description of the type 1 site as having a flattened tetrahedral geometry, with two histidines and one cysteine ligand. Histidine has also been proposed as a ligand for the type 2 site, and indeed it has been suggested1336that the type 2 Cu in ceruloplasmin contains three nitrogen donor ligands. EXAFS measurements exclude sulfur as a donor to the type 3 site, and lead indirectly therefore to the conclusion that in Section 62.1.8.3. a histidine group is a ligand. The ligands at the type 3 site are It is possible to prepare R. uernicgera laccase which is reversibly depleted in the type 2 copper. This is a great aid in understanding the role of type 2 copper in the mechanism. Type 2-depleted laccase may exist with the type 3 site in the oxidized or reduced form. The spectral features of the blue type 1 copper change with change in the oxidation state of the type 3 copper. This intersite structural interaction may relate to the electron-transfer pathway between type 1 and type 3 The type 2 site has been implicated in the binding of polyphenolic sub~trates.9~~ Solomon et a19*' have suggested that the peroxide is bound to a single copper at the binuclear type 3 site, and so differs from the bridging peroxide found for hemocyanin and tyrosinase (Figure 57). Comparison with a laccase from which the type 2 copper has been removed has led to the suggestion that the type 2 copper is necessary to stabilize the type 3 Cu-hydroperoxide complex. The type 2 and type 3 centres appear to be close together, as rapid-quench ESR experiments with reduced laccase and 1 7 0 2 have shown that the H2170thus formed is bound equatorially at the type 2 cu Kinetic studies with laccase have shown that the enzyme must be reduced by the organic substrate before reaction with dioxygen occurs. The first elstron from the substrate is accepted by the type 1 Cu2+, and the second by the type 2 Cuz+. The electrons from these reduced sites are then transferred to the type 3 copper pair, which then binds dioxygen with reduction to peroxide. It is possible that the type 2 and type 3 centres are in the same cavity, which only becomes accessible to the solvent when the type 1 Cu+ is oxidized. An intermediate in the reaction has been identified. Its optical spectrum is very similar to that of the complex formed by adding hydrogen peroxide to the oxidized enzyme. However, it appears unlikely that the peroxo complex is the observed intermediate, as the optical intermediate has been identified with a paramagnetic intermediate, shown by ESR.'339The use of 170-enriched O2

HzY

\

T, =type of copper; o = oxidized, r = reduced Figure 61 The mechanism of laccase CCC6-W*

700

Biological and Medical Aspects

has shown unambiguously that this is an oxygen radical. It appears that after the formation of peroxide, there is a rapid transfer of an electron from the type 1 copper, so that the first observable intermediate is a three-electron reduced species. The ESR data suggest that the radical is 0-, which would be expected if three electrons are transferred to 0,. The type 1 Cu2+ will immediately be re-reduced and then transfer this electron to the type 3 site, so completing the reduction of the hydroxyl radical to water. A scheme is shown in Figure 61.'340 Ceruloplasmin is the major copper-containing protein in mammalian blood plasma (see Section 62.1.8.5), and appears to consist of a single polypeptide chain of molecular weight 130 OO0.'341 The nine-line superhyperfine splitting in the ESR spectrum of the type 2 copper has been interpreted in terms of four equivalent nitrogen ligand^.^" 62.1.12.7

Non-blue Copper Oxidases

A number of copper-containing proteins show spectral features like those of normal copper complexes, and therefore do not appear to contain 'blue' copper centres. Amongst these are galactase oxidase and the amine oxidases. It is noteworthy that it appears unlikely that the copper is involved in the activation of dioxygen. Galactose oxidase catalyzes the oxidation of many primary alcohols to the corresponding aldehyde, with reduction of dioxygen to hydrogen peroxide.'342 It contains one type 2 copper centre per molecule (molecular weight 68 000) and no other metal centres, and thus represents the simplest known copper protein. The Cu" appears to be coordinated to two imidazole groups, one solvent water and a fourth ligand, which is non-nitrogenous, while NMR studies show that exogenous ligands can bind at two sites (axial and equatorial) with displacement of water to give a five-coordinate complex.'343 The mechanism of reduction of dioxygen is controversial, as a two-electron reduction is linked to one redox centre. One possibility is that the copper is reduced to CUI by the substrate, and then oxidized to Cu"' by dioxygen, with production of hydrogen peroxide. However, galactose cannot reduce Cu", and the Cu" is redox inactive in the absence of mediators. As noted above, ligand-binding sites are available on the copper centre, and galactose binds to Cu" in the enzyme. Studies on the binding of fluoride in the presence of dioxygen failed to show evidence for competition between O2 and bound f l ~ o r i d e . This ' ~ ~ suggests that the role of copper in galactose oxidase may only be one of binding the substrate, which then reacts directly with 02. The amine oxidases catalyze the oxidative deamination of biological amines as shown in equation (65). They have molecular weights greater than 100 000 and are generally composed of two subunits, each containing one copper ion. ESR and NMR studies indicate that the copper is Cu" bound by at least two imidazole groups. Alternatively one ligand may be a pyridine group where pyridoxal phosphate is a prosthetic group of the enzyme. There is no convincing evidence for a role for copper in the activation of dioxygen. On the other hand copper-free amine oxidases are not enzymatically active. Bovine plasma amine oxidase reconstituted with only one Cu" per molecule of protein retained full suggesting only one of the two Cu sites is necessary for catalysis. RCH,NHZ+O,+ H,O

-m

RCHO+ H,Oz+NH3

(65)

In benzylamine oxidase there is evidence that the amine undergoes transamination with the pyridoxal prosthetic group to give a pyridoxamine, which is then oxidized by dioxygen to give H202 and N H 3 . The role for the copper is one of activation of the substrate.1346 62.1.12.8 The Superoxide Dismutases Superoxide dismutase is widespread in nature, and is suggested'347 to have the important function of catalyzing the disproportionation of the superoxide ion (equation 66). Aerobic organisms have high levels of superoxide dismutase in accord with this view. Protection for the cell against these toxic byproducts of oxygen respiration is completed by the enzyme catalase, which catalyzes the decomposition of the hydrogen peroxide. Superoxide undergoes spontaneous disproportionation, but the catalytic effect of the superoxide dismutase is remarkable, possibly as high as a factor of 109.'348,'349 20,-

+

0,+0,2-

(66)

Coordination Compounds in Biology

701

The enzyme isolated from eukaryotic sources contains copper and zinc, while those obtained from other sources may contain iron or manganese. The latter two species are closely related, as shown by extensive homologies in their amino acid sequences.1349

62.1.12.8.1 The copper-zinc superoxide dismutase The best studied dismutase is that isolated from bovine erythrocytes. It has a molecular weight of 31 400 and is made up of two identical subunits each containing one copper and one zinc. X-Ray diffraction have shown the copper to be bound by four imidazole ligands (His-44, His-46, His-61 and His-118) plus water, in a distorted five-coordinate geometry, while the zinc has a tetrahedral environment with three imidazole ligands (His-61, His-69 and His-78) and one oxygen donor (Asp-81). The residue His-61 is a bridging imidazolate group. The overall structure is represented schematically in Figure 62.

Figore 62 The structure of the Cu-Zn superoxide dismutase

Each subunit has a deep channel on the outside surface, with the Cu" slightly exposed on the floor of the channel, and close to the completely buried Zn". The amino acid residues making up this channel are highly conserved in enzymes from different species, suggesting that the channel is important. The site for superoxide binding has been identified as a small pit over the Cu". This region of the enzyme is highly stabilized through an interlocking network of hydrogen bonds. The active site pit contains highly ordered water molecules. Two of them form a 'ghost' of a superoxide ion at the Cu site; one binds to the Cu, and the other to a guanidinium N of Arg-141. These two water molecules form an OOCu angle of 120°, which is the expected geometry of a bound superoxide ion. During catalysis of superoxide disproportionation, the copper centre is reversibly oxidized and reduced by successive encounters with superoxide, giving 0, and H 2 0 2(equations 67 and 68). The zinc(I1) almost certainly has a structural role in the formation and stabilization of the active site,1352and indirectly in enhancing the reactivity of the copper. ECU"+ 0,+ H+ E'Cu1+02-+H+

E'Cu'

e

+ 0,

ECu"fHzOn

(67) (68)

The detailed X-ray data have allowed an elegant elaboration of this mechanism. The protein environment with the tetrahedral distortion results in a high redox potential for the copper (0.42 V). The first catalytic reaction is facilitated by relaxation of distortions in the Zn ligand geometry found for the Cu" enzyme through the bridging Hisdl being allowed to rock away from the copper with copper-bridge bond breaking, when the latter is in its Cu' state. The high rate enhancement brought about by the enzyme over an already fast reaction is of interest. The rate constant for superoxide dismutase is within an order of magnitude of the diffusion-controlled limit. It is evident therefore that the number of non-productive collisions between SOD and 02is low. This rate enhancement has been explained1351by a consideration of the electrostatic potential in and around the active site channel. The presence of positively charged residues at critical positions means that the incoming negatively charged superoxide ion is guided towards the copper in the active site, with a minimization of non-productive collisions. It will be interesting to see if a similar system operates in the manganese and iron superoxide dismutases.

Biological and Medical Aspects

702

Both copper and zinc may be removed readily by dialysis, and replaced by four copper ions. Alternatively the zinc may be replaced by cobalt. In both cases antiferromagnetic interactions between Cu-Cu and Cu-Co have been observed. This has allowed the testing of the hypothesis that the mechanism of superoxide dismutase involves the breaking. and reformation of the Cu-imidazolate bond. Lippard and coworker^'^^^ have monitored the ESR spectrum of the four-copper form of the enzyme, and have detected breaking of the bridge on lyophilization by the elimination of the antiferromagnetic interaction between the copper centres. They suggest that bridge breaking results from conformational changes brought about in the protein by the loss of water of hydration. Addition of water resulted in the restoration of the ESR spectrum characteristic of the Cu-Cu coupling. Table 27 Manganese and Iron Superoxide Dismutases

Metall mol

Ref:

MW

Subunits

Bacteria Bacillus subtilis B. slearothermophilus Streptococcus faecaiis S. mutans Thennus aquaticus T. thermophilus Escherichia coli Mycobuclerium lepruemurium M. tuberculosis Chromatiurn vinomrn Anmystis niddans Desuvovibrium desuvuricans E. coli Pseudomonas ovalis Bacteroides fragilis

45 000 40 000 45 000 42 000 80 000 80 000 40 000 45 000 88 000 82 000 37 000 43 000 39 000 44 000 42 000

2 x 22 500 2x20000 2x22500 2 x 19 000 4 x 2 1 000 4 x 20 000 2 x 20 000 2 x 22 000 4x22000 2 x 4 1 000 2x18500 2 x 21 500 2x18000 2 x 21 000 2x20500

1.1 Mn 4.0 Mn 1.3 Mn 1.2 Mn 2.1 Mn

1.0-1.8 Fe 2.0 Fe 1.8-1.9 Fe

8 9 10 11 12 13" 14" 15

Prokaryotic algae Sprirulina platensis Plecronema boryanum

37 000 37 000

2x18400 2 x 19 000

1.0 Fe 0.94 Fe

16 17

Eukaryotic algae Porphyridium cruentum

40 000

2x20000

2.0 Mn

18

100 000 48 000

4 x 25 000 2 x 24 000

3.0 Mn 1.5 Mn

20

86 000 85 000 80 000

4 x 2 1 500 4 x 2 1 000 4 x 20 000

2.0 Mn 3.9 Mn 2.3 Mn

21 22 4

Source

Yeast Saccharomyces cerevisiae Serratia marcescens Animal Bovine heart mitochondria Human liver Chicken livcer mitochondria

2.0 Mn 1.2 Mn 1.3 Mn 4.0 Fe

2.0 Fe 1.0 Fe 1.6 Fe

1

2 3 4 5

6 437

19

Structure determined. 1. K. Tsukuda, T. Kido, Y. Shimasue and K. Soda, A g k BioL Chem., 1983,47, 2865. 2. J. Bridgen, J. I. Hams and E. Kolb, J. MOL B i d , 1976, 105, 333; J. D. G. Smit, J. Pulver-Sladek and J. N. Jansonius, J. Mol. BioL. 1977, 112, 491. 3. L. Britton, D. P. Malinowski and 1. Fridovich, J. BncrerioL, 1978, 134, 22Y. 4. R.A. Weisiger and I. Fridovich, J. B i d daily intake

Iron

3-5 g

10 mg (males) 18 mg (females)

Zinc

2-3 g

15 mg

Fluorine

3g

1.5-4 mg

2-3 g

-

Dietary imbdance Functions

Deficiency

Electron transport (cytochromes), oxygen carrier (haemoglobin), storage/transport (ferritin), enzymes, immune systern > 160 enzymes in main metabolic pathways, nucleic acid and protein synthesis, immune system

Widespread geographically; fatigue, anemia

Danger in hemochromatosis, CooIey’s anemia, acute poisoning, Bantu siderosis

Occurs in Iran, Egypt; TPN, genetic disease, traumatic stress; growth depression, delayed sexual maturation, skin lesions, depression of immunocompetence, change of taste acuity Dental caries, possibly osteoporosis

Unlikely except from prolonged therapeutic use; can interfere with Fe and Cu mciaboiism

Structure of teeth, bones

Calcification, structure of connective tissues Manganese 2.5-5 mg Enzymes in protein and 1g energy metabolism, superoxide dismutase, mucopolysaccharides synthesis 80 mg 2-3 mg Metal storageltransport Copper (ceruloplasmin); enzymes in synthesis of cartilage, bone, myelin; interaction with iron Lipid metabolism, 15 mg Vanadium regulation of cholesterol synthesis, ATPases Iodine Synthesis of thyroid 11 mg 150 hormones Molybdenum 10 mg I 50- 500 pg In metalloenzymes; xanthine, aldehyde, sulfite oxidases Selenium 6- 12 mg 50-200 pg Glutathione peroxidase. interaction with heavy metals; inhibitory effect on cancer Interaction with iron Nickel 10mg absorption, nucleic acid, lipid metabolism Chromium 2 mg 50-200 pg Potentiation of insulin, maintenance of normal glucose tolerance Silicon

Cobalt

1-5mg

3pg Part of Vitamin BIZ, Vitamin BI2 erythropoiesis

Arsenic

1-2mg

-

a

Iron metabolism

Excess

Unknown

Dental fluorides; severe fluorosis in parts of India and South Africa Unknown

Unknown

Toxicity by inhalation

Occurs in malnutrition, TPN; anemia, neutropenia, skeletal and neurological defects

Danger in Wilson’s disease

Unknown

Unknown

Goiter (‘Derbyshire neck’ in the UK) Faulty metabolism of xanthine, sulfur

May lead to thyrotoxicosis Gout-like syndrome in parts of Soviet Union

Endemic cardiomyopathy, Known in parts of China conditioned by Se deficiency; muscle weakness Unknown Unknownb In malnutrition, aging, TPN; impaired glucose tolerance, relative insulin resistance, elevated serum lipids, peripheral neuropathy Vitamin deficiency only, because of diet or failure to absorb Unknown

Unknownb

Unknown Unknownb

Compiled from refs. 31, 32 and 33. Except that dermatitis hypersensitivity and carcinogeneses because of excess exposure are well-known in environmental toxicolo#.

62.2.2.4.2

Iron

An extensive survey of the historical uses of iron with many original references has been given by Fairbanks et Ancient civilizations calied iron ‘the metal of heaven’, probably because of the relatively high amount in meteorites. Remedies containing iron have been known since 1500 B.C., two being mentioned in the Ebers Papyrus, an Egyptian pharmacopoeia and the oldest surviving manuscript of mankind. One was a recipe for a paste to be applied for pterygium, and the second, quoted here, was a relief for baldness:

Uses in Therapy

764

‘“ 0 shining One, thou who hoverest above! 0 Xare! 0 disc of the sun! 0 protector of the divine Neb-Apt!” to be spoken over Iron Red lead Onions Alabaster Honey make into one and give against.’ Another early example of iron usage is entirely legendary. Prince Iphyclus of Thessaly was unable to beget a child, and the seer-physician Melampus cured him.As a child, the prince had been frightened by his father wielding a bloody knife to geld rams. The father had driven the knife into an oak tree, where over the years it became compietely buried. Melampus had the prince cut down the tree, recover the knife, scrape the rust from the blade into wine, and drink the iron-fortified liquor. The prince soon regained his potency. This appears to be the first reference in Western literature to both castration complex and medicinal use of iron. Pliny the Elder wrote extensively on iron. It seems to have been a theraputic panacea for the Roman era. 1000 years later ibn-Sina catalogued the known 10th century anatomy, physiology and pharamacology in his ‘Canon of Medicine’. The description of the uses of iron for a variety of problems ends with ‘it restores sexual potency and strength to men’, echoing the legend of Iphyclus. The ancient physicians were remarkable in guessing that iron is the source of the colour of blood. In the 17th century, chlorosis was recognized. Sydenham in 1681 is credited with a remedy, ‘...a syrup made by steeping iron or steel filings in cold Rhenish wine...’. Not until the first part of the 18th century was the ash of blood shown to contain iron. In the 19th century the association of chlorosis with iron deficiency and the value of large doses of iron(I1) salts were shown.47The last 100 years have shown rapid advancement in our understanding of iron absorption and metabolism. The incorporation of inorganic iron into hemoglobin was demonstrated in 1932. In the 1920s it was established that iron is important in the oxidative mechanisms of all living cells, Iron is present in simple iron proteins (ferredoxins), hemoproteins (hernoglobin and myoglobin), the oxygen carriers; cytochromes, the electron-transfer proteins, which function as reversible acceptorsdonors of electrons; iron-sulfur proteins such as xanthine oxidase; iron-binding proteins invoived in transport of iron such as transferrin and lactoferrin; the iron-storage proteins, ferritin, found in all cells; and hemosiderin, which seems to be deposits of ferritin. Simple iron-deficiency anemia is our single most important nutritional disease, and probably the most common organic disorder seen in clinical medicine. Deficiency is accompanied by fatigue. palpitation on exertion, sore tongue, tinnitus, headaches, cold hands and feet, ‘spoon nails’, decreased resistance to infection, dysphagia, GI disturbances, angina and other symptoms. Even in patients with classic migraine headaches, treatment for iron deficiency may decrease the frequency of attacks. More rarely, iron deficiency may give rise to neurological symptoms simulating intracranial tumors, accompanied by increased cerebrospinal fluid pressure. Deficiency of zinc often occurs at the same time as iron deficiency. Causes include chronic blood loss, such as from menorrhagia, ulcers, malignancy or infections. Competition exists for growth-essential iron between bacterial, fungal and protozoan pathogens and their vertebrate hosts. The more virulent microorganisms can overcome the host’s ability to withhold iron. Human requirement €or iron is highest in the first 2 years of life, during the rapid growth of adolescence, and through the childbearing years for women, when absorption of approximately 2.0 mg d-‘ is needed. The concepts of iron deficiency and anemia, implying varying degrees of depletion of iron stores and of enzymes, and reduction of levels of cixulating hemoglobin, are well understood. Clinical manifestations and diagnostic studies are thoroughly reviewed by Fairbanks et In summary, in severe iron-deficiency anemia, in addition to morphologic changes in erythrocytes, some laboratory findings are abnormal: plasma iron concentration is low, total iron-binding capacity (TIBC) of plasma is often increased; transferrin saturation is usually < 18% and often < 10%; and serum copper concentration is high, amongst others. For mild anemia, the determination of red cell indices has some value in diagnosis; however, it should be realized that these indices may appear to be entirely normal and that they are subject to errors. Patients with ‘statistically normal’ hemoglobin values may show a positive hemoglobin response to iron therapy. A woman with hemoglobin concentration of 13g (100 ml)-] may be iron deficient through excessive menstrual flow; for her a hemoglobin concentration of 15g (100 rnl)-’ CCCS-Y’

Biological and Medical Aspects

7 64

may be ‘normal’ in the physiological sense. Thus, the diagnosis of anemia based on arbitrarily selected criteria such as hemoglobin levels is unsatisfactory. A better criterion is the change in hematocrit value acheved by administration of oral iron(I1) fumarate (60 mg Fe daily for 3 months). Studies using a correlation between this change and initial hematocrit value showed deficiency in 10-24% of the female population of Sweden. Put another way, the probability of iron deficiency is 100% for a hematocrit of < 31%, falling to 28% and 3.2% at hematocrits of 37% and 40% respectively. A hematocrit value of >37.5% suggests a healthy iron level. There is still a real need for a better, simple, reliable method for detection of mild iron-deficiency anemia, a deficiency so common that it can almost be assumed to exist. The evidence of iron deficiency is often disarmingly vague. Anemia may result from other complications even when iron supply is sufficient. A decrease in hemoglobin synthesis, a fault in transport mechanisms or destruction of erythrocytes have all been noted. Sideroblastic or iron-loading anemias are characterized by a fault in iron metabolism (see Section 62.2.3-2). There are also several other syndromes of iron deficiency known clinically, including pica and Goodpasture’s syndrome (an immune-related lung and kidney disease). The different categories of anemia have been discussed in detail by P r a ~ a d . ~ ~ Oral contraceptive drugs have been shown to have a significant effect on iron metabolism. Serum iron and total iron binding capacity (TIBC) are consistently elevated. Hemoglobin is either unchanged or slightly altered. Both increases and decreases have been reported, but the number of subjects is small. (In pregnancy, a decrease is generally observed.) Serum copper and ceruloplasmin are also consistently elevated in these drug users, probably because of the estrogen component, whereas serum and plasma zinc are decreased. These copper and zinc levels are often inversely related, so that depressed serum zinc concentration may be a result of increased copper. The richest dietary sources of total iron are organ meats (liver and kidney), egg yolk, dried legumes, corn, molasses and parsley. Liver is particularly valuable because of the high absorbability of its iron. However, only about 10% of dietary iron is absorbed. Iron deficiency anemia can be treated with soluble iron(I1) compounds providing 200 mg in three or four daily divided doses. Oral iron(I1) sulfate is the least expensive and is in wide use. Ascorbic acid increases the absorption efficiency of iron@) sulfate. Parenteral administration of iron is used when oral iron is ineffective. Iron-dextran, a colloid formed from iron(II1) chloride and an alkali-modified dextran, is one of several preparations available which has found extensive clinical use. It contains up to 28% Fe by weight and has a structural similarity to ferritin. Transfusion therapy may also be used in severe chronic anemia or acute hemorrhage. Acute iron poisoning is also relevant to the discussion of therapy. Toxic effects of iron were recognized in 14th century Italy: ‘He to whom has been given the magnetic stone ... will become lunatic, melancholy, or furiously mad. Now the treatment is... fragments of emeralds and gold filings... with wine, and he is to be clystered with the milk of ewes and oil of sweet almonds.’47 Nowadays, it is the combination of curious toddlers and bottles of attractive tablets that leads to most cases of iron poisoning. Two tablets (0.6g) caused severe symptoms and 2 g have led to death. Symptoms of bleeding, gastoenteritis, shock and possibly coma may occur. Treatment consists of gastric lavage with 1% sodium bicarbonate solution followed by stomach instiilation of 5 g desferrioxamine. Tn addition, an intramuscular dose of desferrioxamine (1 g, children; 2 g, adults) should be g i ~ e n . ~ 9 62.2.2.4.3

Zinc

Zinc is recognized as an essential element for normal function in all forms of Iife. Its deficiency can lead to various disorders including impairment of the following processes: protein and carbohydrate metabolism, capacity for learning, reproductive maturation and function, bone development and normal growth. Low zinc levels may result in mental lethargy and depression. Zinc is believed to be effective therapeutically in wound healing (e.g. following surgery) and in treatment of atherosclerosis. Zinc appears to have a role in DNA replication and transcription, as well as in protein synthesis. In 1940 carbonic anhydrase was isolated from mammalian erythrocytes; the protein was shown to contain 0.33% zinc. In 1955 carboxypeptidase became the second zinc enzyme to be reported. About 20 zinc metalloenzymes have since been studied in great detail, and about 60-70 await complete characterization. The numerous causes of human zinc deficiency, and conditions in which it exists, have been summarized by P r a ~ a d .They ~ ~ , include ~~ nutrition (inadequate diet), excessive alcohol ingestion, liver disease, gastrointestinal disorders (such as Crohn’s disease), neoplastic diseases (conditioned

Uses in Therapy

765

deficiency), burns and skin disorders (excessive losses of zinc), chronic disease and chronic renal failure. Acute zinc deficiency in Middle Eastern countries is probably a result of high phytic acid content of cereals. Additional zinc is required during normal pregnancy. Iatrogenic zinc deficiency may follow use of chelating agents, antimetabolites, antianabolic drugs and diuretics; use of total parenteral nutrition fluids omitting zinc; and use of penicillamine for Wilson’s disease. Two genetic disorders are helped by zinc therapy. (i) Acrodermatitis enteropathica is a lethal, recessive trait disorder which develops in the first few months of life, in which dermatological, ophthalmic, gastrointestinal and neuropsychiatric problems are exhibited. Zinc supplements can result in complete cure. (ii) Sickle cell anemia patients exhibit many of the same clinical features as zincdeficient patients, including delayed growth and delayed onset of puberty. This disease and its therapy are areas for furiher work. In addition to the classical symptoms of zinc deficiency mentioned above, the following unusual conditions have been reported: liver and spleen enlargement, abnormal dark adaptation and abnormalities of taste. Several laboratory procedures for diagnosing zinc deficiency are available. Measurement of zinc levels in plasma is useful in certain cases. Levels of zinc in the red cells and hair may be used for assessment of body zinc status. More accurate and useful parameters are neutrophil zinc determination and quantitative assay of alkaline phosphatase activity in neutrophils. Determination of zinc in 24 h urine may help diagnose deficiency if sickle cell disease, chronic renal disease and liver cirrhosis are ruled out. A metabolic balance study may clearly distinguish zinc-deficient subjects. The pharmacological use of zinc (use other than for zinc deficiency) is, with one notable exception, very recent. The use of topical zinc to promote healing of wounds (zinc oxide in calamine) has been known since 1550 B.C.53Recent studies report enhanced healing from oral zinc therapy for pilonidal cysts, venous stasis leg ulcers, major burns, leg ulcers in sickle cell anemia, and a variety of types of chronic wounds. The standard dose is 220 mg zinc sulfate (50 mg elemental Zn) given three times a day with meals. While this is a pharmacological dose of zinc (150mg d-l), enhanced healing only occurs until the zinc level is raised to the normal physiological level, so that this is really an example of therapy for a deficiency. This effect is consistent with the fact that zinc is required for collagen synthesis. Zinc therapy has enhanced graft acceptance in children with major burns. Measurements of skin zinc levels in chronic steroid-treated patients are amongst the lowest observed. The effect of various stressors, including steroids, other drugs and radiation, needs to be studied. Anesthesia as well as surgical injury causes zinc levels to fall. Studies to determine if supplemental zinc will increase myocardial concentrations or will have a positive effect on the healing of infarctions have been suggested.j4 Other diseases which respond to zinc therapy include sickle cell anemia and rheumatoid arthritis. Sickle cell anemia was the first condition to be described as a molecular disease. An in vivo antisickling effect with zinc therapy was demonstrated in 1977. The proposed mechanism is one of inhibition of caImodulin function by zinc. Only about 10% of administered zinc is absorbed and the therapy would be more effective if plasma zinc levels could be increased. The observations that rheumatoid arthritis (RA) patients (i) may often be zinc deficient, (ii) have low serum histidine levels, and (iii) often benefit from the drug D-penicillamine (a chelating agent which may enhance zinc absorption) and also that zinc has an antiinflammatory effect, have led to a study of zinc sulfate therapy for RA. Significant improvement in joint swelling, tenderness and morning stiffness was found. Brewer suggests that further clinical studies in this area are warranted, and, moreover, that a variety of inflammatory diseases (e.g, ulcerative colitis) might be successfully treated with zinc.53~-PeniciIlaminetreatment for Wilson’s disease has led to zinc deficiencymanifested as severe skin lesions, hair loss and other problems persisting for several years. The lesions were reversible by zinc acetate therapy within one month.50 Zinc absorption is inhibited by most food and the elevated plasma level lasts only 5 h after a dose; thus it is better given five or six times a day, 1 h before or after meals. Zinc acetate may be better tolerated than the sulfate salt. Brewer prefers 25 mg elemental zinc, administered as the acetate, five or six times a day.53 Zinc is relatively non-toxic as a drug. If more than I g zinc is ingested in a single dose, toxic symptoms such as abdominal pain, nausea, vomiting, fever, drowsiness and lethargy may occur. A more significant toxicity problem is zinc-induced copper deficiency which can be corrected with a supplement of 0.5 mg of copper as copper sulfate per day.53 62.2.2.4.4

Copper

Copper is the third most abundant heavy metal in the human body, the concentration being 1.4-2.1 mg per kg body weight. All tissues of the body require copper for normal metabolism.

Biological and Medical Aspects

7 66

Copper may be stored in the liver and released, as required, in the form of ceruloplasmin, for example. Like other trace elements, trace amounts of copper are essential for life, and excess amounts are toxic. Copper is an essential component of numerous key metalloenzymeswhich are critical in melanin formation, myelin formation and crosslinking of collagen and elastin. Copper plays a vital role in hemopoiesis, maintenance of vascular and skeletal integrity, and structure and function of the nervous system. Thus a deficiency of copper can lead to a variety of adverse effects such as increased fragility in bones, aneurysm formation in arteries and a loss of lysyl oxidase activity in ~ a r t i l a g e . ~Articles " ~ ~ on copper also appear in Sigeil, volumes 3 and 5, all of volumes 12 and 13, and volume 14. The causes of human copper deficiency include: ( I ) low intake - malnutrition, total parenteral nutrition (TPN); (2) high loss I.I cystic fibrosis, nephrotic syndromes; and (3) genetic factors Menkes' dsease. Copper deficiency may also be associated with chronic malabsorption, a situation which is made much worse in cases of gastric and bowel resection. Several special diets, including powdered milk, liquid protein and standard hospital diets are a means of inducing copper deficiency. The amount of copper in US food has decreased steadily since 1942, and may be related to the rising incidence of coronary artery disease. A copper deficiency may also occur as the result of the use of chelators for other purposes: for example, diethyl dithiocarbamate is an in vivo metabolite of ANTABUSETM(disulfiram). For the most part, adequate copper is received in diet and widespread human deficiencies do not occur, but deficiencies may arise because of antagonists. The metals Cd, Hg, Ag and Zn interfere with copper metabolism, probably by competing for copper-binding sites in proteins. Ascorbic acid depresses intestinal absorption of copper56 (in contrast to iron). Some proteins in the diet adversely affect utilization of copper. The sulfide ion is a well known inhibitor of copper absorption, since it forms copper(I1) sulfide which is insoluble.56 Like zinc, copper and its compounds have been used since ancient times, with copper dust, acetate, sulfate and carbonate reported in Egyptian and Hindu prescriptions, and also used by Hippocrates and Galen. Copper arsenite was used in 1892 for anemia and debility. Copper sulfate was recommended to strengthen man, to stimulate the heart and blood vessels, to increase deposition of fat and to treat anemia. The adult requirement is 1.25 mg Cu d-l, about one third of which is absorbed. TPN should be supplemented with 0.5-1.5 mg d-' (adults) and 20 pg (kg weight)-l d-' (children). Two genetic disorders of copper metabolism, Wilson's disease (see Section 62.2.3.3)and Menkes' disease, are known. The latter involves impaired intestinal absorption of ~ o p p e r as ~ well ~ i ~as~ probably subcellular metabolic defects wluch result in copper deficiency with respect to metalloenzyme activity. The characteristic 'steely' hair in Menkes' disease results from free SH bonds in hair protein because of failure of lysyl oxidase to produce the disulfide links. Depigmentation of hair and skin, hypothermia, cerebral degeneration, central nervous system retardation, skeletal demineralization and arterial degeneration are all seen. Copper supplements may benefit hypothermia and increase pigmentation but the disease is not generally cured. A Cu1%-His2 complex has been used in treating Menkes' disease, in a dose of 200 pg Cu (10 kg body weight)-', injected subcutaneously in I ml saline, pH 7.35. This complex has the advantage of resembling the form in which copper is transported in the blood serum and may prove to be more effective than simple copper salts.56 Serum copper is increased in leukemia, Hodgkin's disease, various anemias, 'collagen' disorders, hemochromatosis, and myocardial infarction. Measurement of serum copper level together with other clinical parameters appears to be a useful index for evaluating disease activity and response to therapy in malignant lyrnphorna~.~~ 62.2.2.4.5

Cobalt

Cobalt is rare on the earth's surface, but is found in fertile soil and in plants. It is required in the human body in the form of vitamin B12(cyanocobalamin). Vitamin Bi2 was the first metallo complex in living systems to be studied in great depth.33,34,55 It is needed in the formation of hemoglobin, and the adult human body contains 2-5 mg of vitamin BI2and its derivatives. The stomach secretes a glycoprotein called intrinsic factor, which binds cobalamin in the intestinal lumen. Pernicious anemia is caused by a deficiency of intrinsic factor, leading to impaired absorption of cobalamin. A significant amount of research has centred on the preparation and investigation of model complexes containing cobalt. Co2+can be substituted for Zn2+in a number of enzymes without a gross change in activity, and has therefore been used as a probe of active enzyme sites.

Uses in Therapy

767

Cobalt must be supplied in the diet in its physiologically active form, vitamin B12.GI absorption of cobalt is about 25%, with wide individual variation; excretion occurs mainly via the urine. The major part is excreted within days and the rest has a biological half-life of about two years. Originally, the therapy for pernicious anemia was to have patients eat large amounts of liver. The most reliable treatment now is monthly injections of cobalamin. Toxic levels of C o have occurred when addition of cobalt to beer caused cardiomyopathy in heavy beer drinkers. Industrial exposure has lead to pneumoconiosis and other pulmonary problems. Model reactions have contributed significantly to our understanding of biological processes. Both pyridoxal phosphate (vitamin B6) and BI2-coenzymeshave proved useful in mechanism studies. Methyl transfer reactions to various metals are of environmental significance. in 1968 it was shown that methylcobalamin could transfer a methy1carbanion to mercury(I1) salts in aqueous solutions. Recent research on interaction between B i2-coenzyrnesand platinum salts has shown that charged Pt*I salts Iabilize the Co-C bond. Secondly, the B12-coenzymesare unstable in the presence of platinum salts; this observation correlates with the fact that patients who have received cis-platin develop pernicious anemia.

62.2.3 THERAPEUTIC USES OF LIGANDS WHICH FORM COORDINATION COMPLEXES 62.2.3.1 Chelation Therapy in Heavy Metal Poisoning The mechanisms of metal ion poisoning may be grouped into two classifications: (i) electrolyte imbalance, absorption of negative sites of enzymes and osmotic imbalance, which occur in Li+, Na', K' and Ba2+(type a metal ions) poisoning; and (ii) complex formation with toxicity related to the strength of the complex bond, which occurs for all other metal ions. The subject of electrolyte imbalance will be referred to again in the last section of this chapter. Type b metal ions (those which form stronger coordinate bonds to sulfur than to oxygen) tend to accumulate in the kidneys and liver. Nephrotic syndrome, or kidney failure, is the common toxic action. Therapeutic chelating agents are compounds used as antidotes for metal poisoning. Several reviews on the subject have been published in a work edited by Martell,I8 as well as in book^.^*.^^ There is rapid reaction of an agent with a metal bound to a biological site; the resulting binary or ternary complex involves the metal ion, the agent and possibly another ligand, and it is water soluble. The usual route of excretion is from the bloodstream to the kidney, to the urine, where elevated metal concentration is confirmed by analyses. Excretion via the liver, bile and feces is also an important route for some metals such as copper and iron.58-60 Because time is critical in cases of acute metal poisoning, there may be a high rate of transport of heavy metal complexes leading to kidney damage, so that dialysis may be necessary. When treatment is begun too late, damage may be irreversible. The ideal agent would be specific for a given toxic metal, since it would likely have less tendency to alter the balance of other essential metals. Other requirements of chelating agents are that they should: possess an LDso > 400 mg kg-'; form highly stable complexes with the metal ion to be removed - five- or six-membered chelate rings are usually the strongest, and a ligand that can form two or more rings will produce an even more stable complex; have a high water solubility; form complexes that are water soluble; form complexes that are not more poisonous than the metal ion; permit resulting complexes to be excreted via the kidneys without causing further damage; and not undergo significant metabolism. For chelating agents which can be metabolized, a lugh toxicity is often associated with a high lipid solubility, which permits the metal complexes to enter the brain. Steric aspects should be considered Ag(I) and Au(I) prefer two linear bonds, which are best achieved by eight-membered rings or by two monodentate ligands, and these ions will not form five-membered chelate rings. The first chelating agent to be used effectively in humans as a metal poisoning antidote was 2,3-dimercaptopropanol-l or BAL (British Anti-Lewisite). See Table 3 for abbreviations and chemical names of chelating agents. BAL was originally developed to protect -SH enzymes in the body from arsenicals in war gases. It protects such enzymes from several metals in various oxidation states, including Aul/III, Cd", Hg", Pb", Sbl" and Bi"' as well as As"' ion poisoning. The chelates formed are sufficiently water soluble to be excreted in the urine. BAL is lipid soluble and it has the ability to remove metal deposits not available to attack by other chelators. More effective and less toxic compounds have since become available; these include vicinal dithiols, other dithiols, aminocarboxylic acids, cysteine derivatives, and others. Unithiol (sodium 2,3-dimercaptopropane- 1-sulfonate; DMPS) forms very stable, water soluble complexes with Hg2+,Pb2+,CdZL,Zn2+,Bi3+,As3+,Sb2+and Ni2+.The complexes are nearly all less toxic than

BimlogicaE and Medical Aspects

768

Table 3 Chelating Agents: Abbreviations and Chemical Namesm BAL DCBM DFOA DMPS DMSA DPA DTPA EDDHA EDTA TETA

2.3-Dimercaotourooancll-1 British anti-Lewisiteor dirnercaurol Dithiocarbamaie ' Desferrioxamine-B.DESFERALrM Sodium 2.3-dirnercaptopropane-l-sulfonate,Unithiol 2,3-Dirnercaptosuccinicacid D-( -)-Penicillamine Diethylenetriaminepentaacetic acid Ethylenebis-N,N'-(2-0-hydroxypheny1)glycine; also EHPG Ethylenediamineletraacetic acid Triethyleneletramine

the metal ions alone (except Zn and Cd with BAL), and are excreted via the urine with a half-life of approximately 2 h. 2,3-Dimercaptosuccinic acid (DMSA) is much more soluble in water and much less toxic than BAL, can be given orally and is less expensive than DMPS, and it is an effective antidote for the same range of metals. It has also been used, as the antimony complex, in treating schistosomiasis. For Pb poisoning it is as effective as Ca EDTA. It is capable of enhancing methyl mercury excretion from the brain. Other dithiols. including glucosides of BAL and structurally related sulfonate-containing compounds, have been studied. Hundreds of aminopolycarboxylic acids have been prepared. They all contain the =N-C-C02H entity which gives very stable five-membered ring chelates when N and 0 are bonded to the same metal ion. Na2Ca EDTA was first used as an antidote for lead poisoning and has since been studied for other metals such as the radioactive lanthanides and the actinides. Diethylenetriaminepentaacetic acid (DTPA) is frequently more effective than EDTA because its complexes have higher stability constants. The zinc complex Na3Zn DTPA is the preferred agent for removing deposits of transuranium elements resulting from industrial exposure. In D-penicillamine and cysteine derivatives, the three donor groups -SH, -NH2 and -C02H exist together (see Section 62.2.3.4 and also Chapter 20.2). This entity has wide general coordinating ability, and it chelates with many toxic metals. Cysteine itself is a naturally occurring amino acid and the human body enzyme systems can transform it rapidly. It may give temporary protection against heavy metals and then lose its antidotal action. Its derivative, D-penicillamine (DPA), with two methyl groups on the sulfhydryl carbon, has been found to be a valuable therapeutic chelating agent. It was first introduced by Walshe in 1956 for enhanced copper excretion in patients with Wilson's disease (see Section 62.2.3.3). It has been used for removal of Pb, Hg, Au, Pt compounds and Sb, as well as in treatment of rheumatoid arthritis. It can be admmstered orally, has little inherent toxicity, and is readily available. It seems to be replacing BAL in the treatment of heavy metal poisoning. Acetylation of penicillamine removes the donor properties of the N atom, but the resulting compound penetrates cell membranes more easily. Thus N-acetyl-D-penicillamine can remove Hg in methyl mercury from brain tissue more effectively than DPA and is used for chronic mercury intoxication. A group of ligands known as dithiocarbamates are finding use in treatment of nickel poisoning from exposure to Ni(C0)4 in industrial settings.60

62.2.3.2

Iron Overload

Cooley's anemia is a lethal hereditary disease in which there is a fault in hemoglobin synthesis. This condition is reviewed in four articles in a book edited by Martell.'* The patients must receive blood transfusions starting at age six months and continuing every 3-4 weeks for the rest of their lives. The increased iron input from these transfusions exceeds the capacity of ferritin and transferrin, and results in severe iron overload with iron deposits in the heart, liver, endocrine glands and other organs, usually leading to death before age 20. Iron chelation therapy has been used since the 1960s to retard the condition. Desferrioxamine or DESFERALTMhas proved relatively successful and is now widely used for iron chelation therapy. Desferrioxamines are naturally occurring, selective iron(II1) chelators isolated from actinomycetes. The drug form is subject to a combination of metabolic destruction and rapid excretion. It has moderate toxicity and can also be used as an antidote for iron poisoning. Use of ascorbic acid with the drug enhances the amount of iron removed. It is not an ideal drug because it must be injected, is very expensive ($1000 patient-' month-') and is only effective when the iron load is ten times the normal level. Research has concentrated on development of more effective iron chelators. The ideal chelator should be inexpensive, capable of being administered orally and free of side effects. It should have a selective ability to bind Fe"' rapidly, relative to ferritin and transferrin,

Uses in Therapy

769

and should not interfere with hemoglobin, myoglobin and the cytochromes, or with the normal biochemical processes. With a knowledge of bacterial iron chelators and the techniques of modern chemistry, it is becoming possible to synthesize molecules that are more effective than the natural substance in DESFERALTMin removing iron from tissues. In general, ligands whch have a high selectivity affinity for Fe"' (high effective stability constant in vivo) should be polydentate and relatively acidic oxy-donors, such as derivatives of catechol, %hydroxyquinoline, hydroxamic acid and carboxylic acids. Over 100 such agents have been evaluated for clinical use. Extensive tests with ethylene-N,N'-bis-2-(O-hydroxyphenyl)glycine (EHPG) suggest potential improvement over DESFERALTM.The design of new hexadentate ligands, especially cryptates, is important in finding a source of more effective chelators. Many of these chelates resemble analogous compounds used by bacteria to chelate iron. The coordination chemistry and distribution of plutonium in the body is similar to that of iron and desferrioxamine can mobilize 239Pufrom most of the organs. This fact makes the drug important for safety in the nuclear industry. An intake of 25-75 mg Fe d-I may be safe. Long term iron overload has been noted in South Africa (Bantu siderosis). Ingestion of up to 200 mg Fe d-' may be the result of eating food cooked in iron pots or drinking Kafir beer containing 15- 120mg Fe 1-I. Ethiopians have a high iron intake (up to 500 mg d-l), but as siderosis is not common, much of this iron must be ~ n a v a i l a b l eThis .~~ type of hemosiderosis should be treated by preventative measures to restrict the amount of iron in the diet.48

62.2.3.3

Copper Removal in Wilson's Disease

Wilson's disease is the result of a genetic fault in metabolism, and is characterized by a reduced ceruloplasmin level, an excess of copper in the liver, progressive liver disease, neurological symptoms and the presence of Kayser-Fletcher (KF) rings in the eye. The disease is fatal without treatment. Removal of excess copper by treatment with the drug D-penicillamine, or with triethylenetetramine (TETA) for those who do not respond to D-penicillamine, has been successful. KF rings are always present in patients over 12 years old having Wilson's disease, and may be observed before other clinical abnormalities. The damage to various organs is caused by deposition of excessive amounts of copper. It is desirable to diagnose patients in the asymptomatic state, but t h s usually only happens when they are relatives of known patients. Any person having a ceruloplasmin level of less than 20 mg dl-l, or more than 250 pg Cu per gram of dry liver, must be considered to have Wilson's disease. Chelation therapy was proposed by Cumings in 1948. BAL was used initially. In 1957, Walsh started D-penicillamine as an oral chelation therapy and it is still the treatment of choice. Treatment requires life-long administration of about 1 g daily, and such patients now have normal life expectancies. Only the pure D isomer should be used, as the L isomer is more toxic. Some patients become intolerant to D-penicillamine and may be treated effectively by TETA.24On treatment with D-penicillamine, neurological improvement may precede the disappearance of the KF ring, unless irreversible brain damage has occurred.54There is now evidence that D-penicillamine therapy should be supplemented with pyridoxine (vitamin B6) and zinc acetate, and probably also with vitamin A, in order to minimize toxic side effects.

62.2.3.4 D-Penicillamine for Rheumatoid Arthritis The mechanism of action of D-penicillamine as an antiarthritic drug is still unknown, but since we know penicillamine can form a variety of metal chelates and complexes, it seems appropriate to include a further note on it here. D-Penicillamine has become established through a number of double-blind controlled clinical trials, as an efficacious agent for treatment of rheumatoid arthritis (RA). Amongst available second line agents it is being used as the drug of first choice about as frequently as MYOCNRISINTM in Canada, Benefit is achieved in nearly 70% of patients, and about 30% of patients must discontinue because of adverse side effects. There is no increase in toxicity amongst the elderly. Our results indicate an acceptable efficacy at lower doses (500-600 mg d-l). Only the pure D isomer is used in medicine, since the L and DL forms are toxic. We advocate a strict toxicity monitoring program, including full blood and urine analysis monthly.61 Again, supplements of pyridoxine, zinc acetate and vitamin A may prove useful in reducing side effects.

770

Biological and Medical Aspects

62.2.3.5 Aluminum Removal in Alzheimer’s Disease Aluminum is the most common metal in the earth’s crust, and it is perhaps surprising that its concentration in the human body is so low. About 50 mg A1 d-’ is ingested in foods and dusts. The concentration of Al is relatively high in processed foods such as cheese and pickles, and in plants that thrive in acid soils such as cranberries and onions. Cooking of acidic foods in aluminum pots, addition of alum as a food preservative and the use of antiperspirants may contribute to increased Al levels in the human body. Antacids containing aluminum are common therapy for peptic ulcers. Normally the A1 is eliminated and there is no resulting toxicity, but aluminum has a neurotoxic effect and its absorption from orally administered salts is a potential hazard. Patients undergoing dialysis because of renal failure are usually given about 4 g aluminum hydroxide gel daily. Such patients may develop dialysis dementia and may have a brain level of A1 which is twelve times greater than controls. Alzheimer’s disease (AD) is the most prevalent form of senile dementia. Up to two million people in North America suffer from it, and it is the fourth cause of death in the elderly. The cause and treatment of this disease are therefore extremely important. Although the role of aluminum in AD and its in vivo chemistry is not known in detail, patients with Alzheimer’s disease have been shown to have elevated aluminum concentrations in certain parts of the brain. Aluminum appears to concentrate in the nucleus. Crosslinks with DNA strands have been found. Crosslink formation can be reversed by sequestering the aluminum with EDTA. Three associations between Down’s syndrome and Alzheimer’s disease have been noted. For Alzheimer’s patients there is a higher frequency of Down’s syndrome in the family tree. People having Down’s syndrome who survive to age 30 or 40 often develop Alzheimer’s disease at that time. In both diseases there are fingerprint abnormalities which may serve as markers. In Down’s syndrome an abnormality in chromosome 21 has been identified, and the evidence suggests that there may be a different abnormality in the same chromosome for Alzheimer’s disease. A possible link between AD and the infectious prion (a particle smaller than the smallest virus) is being studied by some researchers. Aluminum may be a compounding factor in AD; it may accelerate or alter the course of the disease via deregulation of calcium. Calmodulin (a protein with four binding sites) activates a number of enzymes and is an important messenger in brain cells. A1 binds to this protein with ten times the affinity of Ca, and changes its conformation. If it can be shown conclusively that A1 removal aiters the course of AD, then A1 exposure should be regulated, and removal may be important. Desferrioxarnine has been found to remove Fe and A1 together in the ratio 5 Fe to 3 Al. The drug has the disadvantage of being very expensive (see Section 62.2.3.2). In a pilot study patients receiving desferrioxamine deteriorated less than the placebo group. The study is currently being continued with a larger group.62 AD is a primary degenerative dementia affecting humans as young as in their forties. The German physician Alois Alzheimer first described the disease in 1906. It is characterized by senile plaques and paired helical filaments (PHFs), and the severity of the condition directly parallels their number.63The involvement of aluminum in this and related dementias (dialysis encephalopathy and amyotrophic lateral sclerosis - Parkinson dementia in Guam) is currently a highly contested issue in neurological research. Elevated levels of A1 have been reported in some regions of the brains of subjects dying with AD.64 The intracellular locus of A1 accumulation is the nuclear chromatin in AD,65,66and the cytoplasm in dialysis brains.67 The association of A1 with the nuclei of brain cells has also been demonstrated in experimental animals with Al-induced e n ~ e p h a l o p a t h y .Researchers ~ ~ * ~ ~ agree that A1 levels increase with human age; however, whether the-levels are pathologically elevated in AD patients, or are just age associated, remains c o n t r o ~ e r s i a lIt. ~seems ~ ~ ~ ~likely that there is a regulatory mechanism for A1 in the human body, and that A1 may be an essential element whose function has not yet been identified. The main research problem is that Alzheimer-type neurofibrillary degeneration has not yet been reproduced in the laboratory. The A1 ligands in the brain remain to be identified and the chemical steps in the formation of Al-contaiaing tissue must be defined. A complicated concentration- and the formation of at pH-dependent interaction between A1 and DNA occurs in vitr0~~~~~4involving least two A1-DNA complexes. The solution interactions of Al”’ with calf thymus DNA74 and ATP75 have been studied with the resulting proposals that three distinguishable AI-DNA complexes and four AI-ATP complexes are formed. In the latter study, it was possible by 27Al NMR to distinguish the varying degrees of binding of Al”’to the bases and phosphates of the nucleotide.

Uses in Therapy

77 1

Clinically, the administration of desferrioxamine (DFO)76 has been reported to increase urinary A1 excretion in humans and to slow the progression of AD symptoms in a small number of patients.77Very recently, a major step in the diagnosis of AD was reported in the preparation of antibodies that specifically recognize PHFs in human brain tissue.78This may lead to a diagnostic test, which is a very pressing need because other conditions that cause dementia, such as depression and drug toxicity, are reversible. Also therapy for AD may be more effective if it is diagnosed early. Presently AD can only be confirmed on autopsy. 62.2.3.6 Antiviral Chemotherapy

At the present time, there is no accepted chelating agent which can be used against common influenza viruses in humans. A virus has a core of either DNA or RNA and a protective coat of many identical protein units. All viruses are either rods or spheres, that is the protein coats are cylindrical shells having helical symmetry or spherical shells having icosahedral symmetry. Viruses reproduce inside living cells, where each viral nucleic acid directs the synthesis of about 1000 fresh viruses. These are then released and the host cell may die. Certain metal chelates have been found to act in an antiviral capacity by interfering with the viral cycle. The therapeutic substance may (i) destroy the virus outside the cell; (ii) occupy sites on the cell surface and prevent penetration by the virus; or (iii) prevent replication processes inside the cell. Current research is centred on understanding the molecular biology of viral infections and on finding substances with selective toxicity. Substances which are effective in vitro may not be so in vivo, where they must be soluble in body fluids, be transported to the site of the viral infection and be capable of blocking the viral cycle. Then too, they may exhibit a curare-like effect and therefore not be usable in therapy. To be effective, antiviral agents would have to be given before the peak of virus multiplication, and not be toxic. Symptoms caused by overreaction of the host immune system will not be relieved except by general supportive treatment (bed rest and acetaminophen). In spite of these limitations some antiviral drugs are known, and the relationsbps of their activity to steps in the viral cycle are discussed by Ferrin and S t ~ n z iPhosphonoacetic .~~ acid has shown activity against herpes viruses by inhibition of the virus DNA synthesis. Acetylacetone (2,4-pentanedione) derivatives are active against herpes simplex and some RNA viruses, including influenza. The role of metal ions in these mechanisms is being studied, for example by computer models of Mood plasma. Cu, Zn, Mg, and Ca ions are believed to be important. Metal complexation in the biological activity of bleomycins is the subject of a review. 8-Hydroxyquinoline has been found to have antifungal and antibacterial properties in the presence of metal ions. Many other examples are known. 62.2.4 COMPOUNDS WHOSE INTERACTIONS IN VIVO ARE UNKNOWN 62.2.4.1 Electrolyte Balance 62.2.4.1.1

The role of Na+, K+, Mg2+ and Ca2+ions

There are several labile elements for which the nature of the coordination sphere around the metal ion in vivo is somewhat speculative, and for which the amount of well-defined coordination chemisty in medical therapy is still rather small. These elements, Li, Na, K, Ca and Mg, are critical in several biological processes: they control osmotic balance in cels and the stability of intracellular structures such as nucleic acid^.^^,^^ Their release, transport and exchange trigger and regulate numerous biologica1 actions. Na', K+, Mg2+and Ca2+are abundant in living systems, reflecting the high natural abundance of these elements in the earth's crust and their ease of incorporation. Their distribution is selective, with K+ and Mg2+concentrated in cytoplasm, and Na+ and Ca2+ concentrated outside cells. Intracellular K+ concentration is 150 mM, 30 times higher than extracellular, whereas intracellular Na' is about 15 mM, 10 times lower than in extracellular fluids. Energy is required to maintain this cation gradient across the cell membrane, and this energy comes from the hydrolysis of ATP, which requires the enzyme adenosine triphosphatase (ATPase) and the various ions. For a cell at rest, K+ is pumped in and Na+ is pumped out. (Perhaps up to one half of the basal metabolic rate in man [17OO kcal d-'; 7100 kJ d-]) is expended in pumping Na+.) Localized reversals of this process coincide with control and trigger of electncal impulses: intracellular NaS level is particularly low in nerve and muscle so that entry of a small amount of Na+ causes a large increase in concentration. The shock observed after severe burning is because K' ions are lost from within

Biological and Medical Aspects

772

cells. The use of these ions in therapy is concentrated on maintaining appropriate balances, for example the lowering of Nat to reduce hypertension. Magnesium ions are found complexed with nucleic acids inside cells and are necessary for nerve impulse transmissions, for muscle contractions and for metabolism of carbohydrates. The physical integrity of the DNA helix is dependent on Mg2+, and the physical size of RNA aggregates is controlled by Mg2+concentrations. Magnesium deficiency may be asymptomatic; some symptoms include muscular twitching and tremors, vertigo, weakness, numbness; sweating and tachycardia; apathy, depression, poor memory and confusion. Increased Mg2+in the diet has been claimed to reduce the extent and frequency of myocardial infarctions. Milk of magnesia and epsom salts (MgS04.7H20) are used as laxatives. Overdoses of magnesium can cause depression and anasthesia. The use of magnesium-containing antacids such as magnesium oxide or trkilicate in patients with renal failure causes magnesium i n t ~ x i c a t i o n . ~ ~ Calcium, phosphorus and vitamin D are needed for formation of bones and teeth. Supplements containing calcium, fluoride and vitamin D are commonly used to treat osteoporosis, and some American physicians recommend ‘Turns’ as an inexpensive source of calcium. Calcium is needed also for formation of milk, maintaining correct heart rhythm, and conversion of fibrinogen into fibrin to form blood clots. Calcium salts are sometimes administered to promote blood clotting. Calcium deficiency in blood plasma causes muscle cramps, twitching and eventually convulsions. If the blood calcium level falls, calcium is leached from the bones, causing osteomalacia. A group of cardioactive drugs known as the ‘Ca antagonists’ block influx of Ca2+ into infarcted, dying heart muscle cells following myocardial infar~tion.~~JO

62.2.1.1.2

Lithium in psychiatry

The use of lithium in medicine has been the subject of recent reviews by Birch,*l Birch and S a d l e ~and - ~ ~references therein, and Tosteson.p2Historically, the use of lithium in medicine began with the treatment of ‘gout and rheumatics’ in 1859. For the following 90 years, lithium was proposed for a variety of disorders and then discarded; for example, lithium bromide was considered to be an effective sedative. In 1949 lithium was introduced into psychatric practice and lithium carbonate, Li2C03, became the first of the modern psychotropic drugs. In a review of double-blind trials Schou and Thomsen (1975) support the prophylactic use of this drug in bipolar (manic-depressive) illness. About 0.1-0.2Oh of the general population are receiving lithium. About 50-60% of these patients are benefitting, if appropriately diagnosed. Daily dose depends on age, body weight and kidney efficiency; it is usually between 1 and 2 g Li2C03per day in divided doses. ‘Slow release’ preparations have been tried but not found to produce an advantage. The objective is to maintain plasma lithium level in the range 0.6-1.2 m o l 1-’. At levels greater than 2 mmol 1-’ toxic symptoms such as skin rash, tremors, confusion and GI disturbances are likely to be seen. Li treatment is by nature a long-term process, and considerable supervision and monitoring are required. A Li saliva test which would correlate with the amount of Li circulating in the cerebrospinal fluid would be an advantage over the continual blood tests presently used for monitoring. Long-term side effects of lithium treatment include weight gain. The treatment is associated with development of hypothyroidism in about 10-15% of cases. There is an association with kidney disease. Birch has expressed the general view that Li may interact with magnesium-dependent processes, and theoretical chemistry supports this view. Despite the widespread clinical significance of Li, there is presently no consensus on its mode of action. One postulate for the mechanism is termed ‘hyperpolarization’. Li affects the conductivity in cell transport channels. Other explanations include modulation of neurotransmitter concentrations and inhibition of Na+/K’ /Mg2+lCa2+ATPases. Lithium has the smallest crystal radius but the largest hydrated radius of the alkaii metals. Its h g h polarizing power gives it a high affinity for oxygen and nitrogen binding sites on ligands; in this respect it is most similar to magnesium. Since a great many biological processes are magnesium dependent, competition between Li and Mg could disturb a number of balances in biochemical reactions. Selective complexation on the basis of size of specific metal ions to facilitate the transport across membranes in vivo has been shown by such naturally occurring antibiotics as valinomycin (for potassium), and actinomycin (for sodium). A series of such ligands, called ionophores, can be produced, including crown ethers and cryptands, both of which are able to transport highly charged ions through lipid membranes. Such ligands do not occur naturally in man, but channels for transport of ions through membranes have specific diameters for specific ions. A study of

Uses in Therapy

773

ionophores and their reactions m a y provide more information on lithium binding and action, which in turn may provide more knowledge of the mechanisms of psychoses. The possibility of a relationship between depressive illness and cancer has been suggested. It should be realized, however, that the attributing of certain diseases to certain personality types has been common in the history of medicine (for example tuberculosis) and has proved to be proportional to ignorance of the true causes. There is a case review showing that patients having depressive illness have greater than average incidence of cancer; there is one prospective study in Iowa and one in Europe. A question which should be posed to and studied by psychiatrists in general is: ‘What percentage of recurrent manic-depressive patients have cancer? A related question which should be posed to oncologists in general is: ‘What percentage of cancer patients are exhbiting manic-depressive illness? How do we know if a patient has been appropriately diagnosed? It seems likely that cancer could produce sufficient metabolic changes in a person that depression or psychotic symptoms would be manifested. Also, a patient realizing that he may be terminally ill is very likely to feel hostility, repression and depression. If more simple screening tests for various carcinomas were available (e.g. blood tests, trace element concentrations), cancer patients could avoid being diagnosed as psychiatric patients, and physicians could avoid mistaken diagnoses. 62.2.4.2 Vitamins, Hormones and Miscellaneous Drugs In addition to the well-defined inorganic complexes and potent chelating ligands already discussed, there are numerous substances - vitamins, hormones and drugs - which contain (i) heteroatoms capable of forming chelates, or chelating structures, in addition to other pharmacophoric groups; or (ii) donor sites which can form complexes but not chelates. For many of these substances, particularly those of group (ii), there is no clear correlation between effect or activity and metal affinity. In other words, the literature does not hold a body of well-defined coordination chemistry for all these substances. Nevertheless, some research is in progress. No chelating agent can be expected to be active physiologicallyunless its stability constants are at least as high as those of the common amino acids. Yet if the stability constants are too high, the substance will be saturated before it reaches the desired site of action. If cell penetration is desired, lipophilic groups are added to the The properties and modes of action of vitamins and hormones are the subjects of many volumes, including refs. 83-85, and will be mentioned only briefly here. Examples of such substances which have strong chelating properties are thyroxine, histamine, noradrenaline and adrenaline. In mammals, both a thyroid and a parathyroid hormone exist to regdate the level of circulating calcium. Calcium controls the permeability of semipermeable membranes and the excitability of the nerve cell membrane. Calcium ions play a role in release of neurotransmitters. In the catecholamines, all the metal binding occurs between the two oxygen atoms attached to the benzene ring. Few attempts have been made to study formation of metal complexes by hallucinogenic drugs, but available data suggest that under physiological conditions there are no strong tendencies. Diazepam (a tranquilizer) binds to metals via nitrogen and forms well-defined complexes ML2X2, M = Cuz+,or ML,X,, M = Co2+,Ni2+,X = C1-, Br-, L = diazepam. Barbiturates form complexes with Co, Ag and Hg ions.2’ 62.2.4.3 Aging, Depression, Senility and Trace Elements Much apparent senility in the elderly is really undiagnosed depression -at least 15% and likely as much as 35% or more. Obviously, investigations for depression should be made by physicians before assuming Alzheimer’s disease or other conditions. We have seen in reviewing the essential trace elements and their deficiencies and excesses how important the levels of these elements are to general health and a feeling of well-being. There have been studies on trace element distribution which show considerable change with aging -accumulation in one organ and reduction in another. This whole area opens an exciting avenue of research into the relationship of the relative distributions of trace elements in various diseases. 62.2.4.4 InteractionsBetween Elements, Vitamins and Drugs There is a danger of creating new imbalances by careless use of supplements of vitamins and trace minerals, hormones and drugs. Examples of interaction between trace elements and vitamins include the following: vitamin C (ascorbic acid) enhances the biological availability of iron, de-

774

Biological and Medical Aspects

presses that of copper, and renders selenium almost totally unavailable. Zinc is required for the function and rnetaboiism of vitamin A.31There is a vitamin E-dependent role for selenium. Glutathione peroxidase (GSH-Px) was identified as a selenoprotein in 1973. Both vitamin E and Se protect membranes from oxidative degradation. It is believed that GSH-Px acts to destroy peroxides before they can attack cellular membranes, while vitamin E acts withm the membrane itself preventing the chain reactive autoxidation of membrane lipid^.^',^^,^^

62.2.4.5

Effect of Metals on Drug Absorption

Drug interactions can occur at many sites at receptor or plasma protein binding sites, or in the liver and Iudney. An important area for interaction is the site of drug absorption in the GI tract. Here a variety of factors, including changes in pH, various foods, antacids, and the formation of insoluble chelates and complexes, contribute to drug interactions. Considerable research has been done on the effects of metal salts or metal-containing antacids. A large number of drugs show decreased absorption because of magnesium or aluminum hydoxide gels (common antacid preparations). As an example, decreased tetracycline aborption results from interaction with Ca2+,Fe2+, Mg2+ and AI3’ to form poorly absorbed chelates: thus if iron salts and tetracycline are both prescribed they shoutd be taken separately at 3 h intervals. A drug absorption interaction could cause unexpected therapeutic failure - non-prescribed medication, especially antacids, could be interacting with the prescribed drugs.Sh -”

62.2.5 STEREOSELECTIVE ACTIONS OF DRUGS There is currently a growing awareness amongst pharmacologists of the importance of stereochemistry, particularly of the chirality of drug molecules. These drugs may be coordination complexes, ligand molecules with potential for in vivo coordination, or molecules whose in vivo interactions are unknown. Stereoselectivity in the metal complexes of amino acids and dipeptides has been reviewed by Pettit and H e f f ~ r d , ~and ’ will not be discussed further here. Most clinicians will know only the generic of trade names of a drug. Few will know the chemical name or will be aware of the two-dimensional representation of the structure, and even fewer will understand the three-dimensional implications of that structure. A classic case of this was the early work by Hill et on the anticancer activity of the platinum complexes &chloro(l,2-diaminocyclohexane)platinum(II). It was not realized that the 1,2-diaminocyclohexaneligand existed in three different forms and that the three derived platinum complexes might have different biological activities. Kidani and his colleagues89were the first to synthesize the three different isomers and to show there was a distinct difference in biological activity and therapeutic index. Investigation of the implications of chirality with respect to gold drugs is being carried out in our own research group. Many of the drugs used in clinical practice are chiral, that is they exist in two structurally different forms, called enantiomers, which are mirror images and are related to one another in the way that the left hand is related to the right. Most of these drugs (over 80%) are administered as racemates (mixtures of the left- and right-handed enantiomers) because (i) the method of preparation produces a racemate and separation may be too costly, or (ii) in the case of a newly discovered drug the techniques for separation of the two forms may not yet be developed, or (iii) as quoted to us by a drug company scientist, ‘one form is active, the other is not, and since we have to use a filler anyway there is no point in separating the two chiral forms’. That the different chiral forms of a drug may have different biological and medical properties is hardly surprising. Many biological molecules are chiral and exist in living systems as one enantiomer onIy. Proteins are made up of only the L forms of amino acids. The normal forms of DNA (A and 3)are right-handed helices. That a particular ‘handed’ enantiomer of one compound interacts differently with the two enantiomers of another compound has been known since the days of Pasteur.90This effect is often used as the basis of separation of enantiomers and can explain why the different enantiomers of a chiral drug have different biological and physiological activity. More importantly, however, the two isomers of a drug may follow different metabolic pathways, and pharmacological activity is often conferred by one isomer only. It has been suggested by Ariens9*that the terms ‘eutomer’ and ‘distomer’ be used for the enantiomers with the higher and lower pharmacological activity. Obvious reasons for using the eutomer alone are (i) the distomer

Uses in Therapy

775

is metabolized to toxic products, (ii) the distomer contributes to adverse side effects, (iii) the distomer is metabolized to products with unfavorable ( e g . too potent) pharmacological action, or (iv) the distomer counteracts the pharmacological action of the eutomer. The drug thalidomide serves to illustrate the first of these points: with respect to hypnotic potency, the (5‘)-(-), (A)-(+) and (RS)-(+) racemate of thalidomide are all equally active, as tested but not (R)-(+)-ihaliby prolonging of sleep induced by phenobarbitol. (,!+(-)-Thalidomide, domide, is transformed in vivo to L-N-phthaloylglutamine and L-N-phthaloylglutamic acid, products that are both embryotoxic and teratogenic in SWS mice and Natal rats. The D isomers of these products, formed by the (R)-(+) drug, are In the area of rheumatology, the importance of using optically pure drugs is well illustrated by penicillamine: D-penicillamine has been used for many years in treatment of Wilson’s disease and cystinuria and is now widely used in rheumatoid arthritis. It is well established now that this drug, which is chiral, should only be given in the pure D (or S) form, because the toxicity of the L (or R ) , or the DL ( R S ) racemic forms, is much greater. This fact was found by trial and error, with some earlier patients on DL-penicillamine experiencing severe adverse side reactions such as optic neuritis. Now only the pure D form is available for prescription (see also 62.2.3.4).61 Even more subtle effects arise for drug interactions of a non-chiral drug with a chiral one. Phenylbutazone is not chiral in itself but it can interact with a chiral drug, warfarin, to change the activity of the latter. Phenylbutazone inhibits the oxidative metabolism of the (S)-(-)form of warfarin, (which is five times more potent than the (R)-(+)form) and thereby decreases its clearance. Conversely, phenylbutazone induces the enzymatic reduction of the ( R ) form thus increasing the clearance.93 Analysis of total warfarin may indicate little departure from normal pharmacokinetics, but the distribution of eutomer and distomer will have changed markedly. Another important example of the importance of chirality, related to rheumatology, is that shown by some members of the NSAID family of aryl acids. Certain molecules which inhibit prostaglandin synthetase in vitro show in vivo antiinflammatory actions. Such compounds include several families of acidic NSAIDs, particularly the aryl acids: salicylates, indomethacin analogs, phenylacetic acids, fenamic acids and enolic compounds. With respect to the indomethacin analogs, an important characteristic has emerged, namely, the requirement for a sinister absolute configuration (the S form, which happens to be +) at the chiral centre. For these drugs the Is>-(+) enantiomers show dominant, if not exclusive, activity.94 Various arylpropionic acids show similar specificity. For most, if not all, the (5‘)enantiomer is the pharmacologically active one, whereas the ( R ) enantiomer i s usually much less active, although the ratio of ( S ) / ( R ) activity varies from drug to drug (and species to species). Only one of these drugs, however, is administered as the separated (S)enantiomer (mproxen, NaprosynTM).Normally these drugs are considered safe, and one cannot readily differentiate between the relative activities of the (5‘) and (R)forms because the in vivo half-life is very short, typically one or two hours. In patients with impaired renal function, where clearance is much slower, however, problems can arise. From in vivo studies of ibuprofen, it was established that the (S)-(+) isomer was responsible for antiinflammatory activity. In vivo, however, the ( R ) - ( - ) isomer may become active because there is stereoselectiveinversion from ( R ) to (S) (but not from S to R)in vivo with a half-life of about two hours.95 This inversion apparently proceeds by stereoselective formation of the coenzyme A (CoA) ester of the (R)-(-)-arylpropionic acid, followed by epimerization and release of the (5‘)-(+)-enantiomer. This epimerization is observed in vivo before the oxidative Such inversion from ( R ) to (S)in vivo is also known for f e n ~ p r o f e nand ~ ~ benoxarnetaboli~m.~~ profen, and is expected to occur for most of the drugs of this series.98 Thus the ( R ) form cannot correctly be described as inert. Statements such as ‘the ( R ) form is inert’ and the ‘(S) form is pharmacologically active and is therefore also responsible for adverse side effects’ must be examined carefully and tested by experiments which are capable of measuring stereospecificdifferences. In the monitoring of drug levers and metabolites in body fluids, methods which do not distinguish between the (R)and (S)forms of chiral drugs and their metabolites are likely to yield pharmacokinetic nonsense. One further point which should be considered is the importance of dose size. Because of the (R) (s)conversion, the dosage of the (S)form administered may be as much as two or three times the anticipated dose. One can visualize an elderly 90 Ib lady, with decreased renal function, who is administered a racemic drug. She receives the normal dose calculated for a 150 lb person (because of the way the tablets are made up). Because of decreased renal function and increased retention there is time for all the ( R ) enantiomer to be converted to the (S) enantiomer. Effectively, she will receive three times the needed dose of the active drug and the area under the dose-time curve will be much greater. It is hardly surprising that adverse side effects sometimes occur.w

-

Biological and Medical Aspects

776

In summary, clinicians and pharmacologists need to be aware of the potential problems arising from the administration of drugs as racemates. The known differences in pharmacological activity and metabolism exhibited by optical isomers can be applied advantageously, and even exploited by manufacturers, and moreover it can be assumed that such important distinctions will eventually be shown for drugs which have not yet been so extensively studied. 62.2.6 CONCLUSIONS

While the use and study of platinum and gold complexes in cancer and arthritis have resulted in the development of a large research area of current interest, these drugs should be seen in the perspective of other discoverieswhich have made an enormous impact on medicine: the development of anesthetics, first nitrous oxide and ether, and then chlorofonn in 1847; the introduction of phenol as an antiseptic by Lister; the discovery of salicylic acid and its analgesic properties in 1876; the discoveries of insulin in 1922 and ascorbic acid in 1927. The use of anesthetics, antiseptics and antibiotics plus modem public health measures providing safe drinking water, sewage and immunization programs have brought about dramatic increases in life expectancy and survival rates, and have made heart disease, cancer and arthritis the modern epidemics. Moreover, a perception of ‘a molecular dsease’ (a term first used by L. Pauhng and H. A. Hano in describing sickle cell anemia), has evolved for atherosclerosis and diabetes, and more recently for cancer and arthntis. Despite the many contributions that chemists and physicists have made to medicine, there remain unanswered problems. Some of them, of course, cannot be resolved by chemists alone, but require political and legislative action, for example the known hazards of smolung to both smokers and non-smokers. The high incidence of lung cancer is well recognized in the 1980s; perhaps less well-known is the fact that lung cancer deaths are expected to exceed those from breast cancer in women in 1985. Occupational health and safety measures in the work setting require input from several professions, but chemists should have a vested interest because chemists as a profession are at a higher risk to certain cancers than are the general population. Chemical research provides data regarding the safe levels and the measurement of levels of toxic substances, as well as the esoteric area of development of new anticancer drugs. The science of epidemiology with its emphasis on statistics has opened up many areas for research. It should be realized, however, that mortality rates -health statistics dealing with causes of death-are not useful because there are inaccuracies in about 42% of death certificates. Statistics relating to the incidence of disease are therefore more useful. There is still much to be learned in medical research from autopsies: a startling fact for the layman is that the major diagnosis has turned out to be wrong in up to 40% of people who are autopsied. Chemists and other scientists can play a vital role in the more legal and political aspects of medicine, such as publicizing the necessity for autopsies, as well as in the development of chemical tests which will aid diagnoses and in the development of new complexes useful in therapy. 62.2.7 GLOSSARY Angina Antiseborrhoeic DNA Diuretic Dysphagia Erythrocyte GI Iatrogenic Hematocrit LD50

Lysis Pica Pterygium RA RNA SOD Splenomegaly TPN Thrombocyte

ischemic chest pain against acne and dandruff deoxyribonucleic acid promoting excretion of urine difficulty in swallowing red blood cell gastrointestinal physician-caused volume of erythrocytes lethal dose, or amount of compound which results in death of 50% of the animals being tested (in a large population) any tissue or cell breakdown eating of mud and clay disease of the conjunctiva of the eye rheumatoid arthritis ribonucleic acid superoxide dismutase enlargement of spleen total parenteral nutrition platelet

Uses in Therapy

777

62.2.8 REFERENCES 1. H. Sigel, (ed.), ‘Metal lons in Biological Systems’, Dekker, New York, 1979-1983, vol. 9-16. 2. B. Rosenberg, L. VanCamp, J. E. Trosko and V. H. Mansour, Narure (London), 1969,222, 385. 3. R. D. Gillard, ‘Heavy Metals in Medicine’, in ‘New Trends in Bio-inorganic Chemistry’, ed. R. J. P. Williams and J. J. R. F. daSilva, Academic, London, 1978, pp. 355-387. 4. Proceedings of the Third International Symposium on Platinum Coordination Complexes in Cancer chemotherapy, Parts I and 11, ed. A. Khan, Wadley Institutes of Molecular Medicine, Dallas, Texas, October, 1976. J. Clin. Hematol. Oncol., 1977, 1-827. 5. A. W. Prestayko, S. T. Crooke and S. K. Carter (ed.), ‘Cisplatin. Current Status and New Developments’, Academic, New York, 1980. 6. S. J. Lippard (ed.), ‘Platinum, Gold and Other Metal Chemotherapeutic Agents’, ACS Symposium Series 209, American Chemicai Society, Washington, DC, 1983. 7. M. P. Hacker, E. B. Douple and I. H. Krakoff (eds.), ‘Platinum Coordination Complexes in Cancer Chemotherapy’, Martinus Nijhoff, Boston, MA, 1984. 8. J. J. Roberts and A. J. Thornson, h o g . Nucleic Acid Res. Mol. B i d , 1979, 22, 71-133. 9. R. E. Dickerson and H. R. Drew, J. Mol. Biol., 1981, 149,761-785. IO. J. P. Caradonna, S. J. Lippard, M. J. Gait and M. Singh, J . Am. Chem. Suc., 1982,104, 5793-5795. 11. A. J. M. Marcelis, J. H. J. den Hartog and J. Reedijk, J. Am. Chem. Suc., 1982, 104,2664-2665. 12. J. H. J. den Hartog, C. Altona, J. C. Chotard, J. P. Girault, J. Y. Lallemand, F. A. A. M. de Leeuw, A. T. M. Marcelis and J. Reedijk, Nucleic Acids Res., 1982, 10,4715. 13. M. J. Cleare and P. C. Hydes, ‘Metal Complexes as Anti-Cancer Agents’, in ref. I , vol. 1 I, pp. 1-62. 14. R. D. Gillard, in ‘Recent Results in Cancer Research’, ed. T. A. Connors and J. J. Roberts, Springer-Verlag, New . , York, 1974, vol. 48, p.29. 15. R. J. Puddephatt, ‘The Chemistry of Gold’, Elsevier, Amsterdam, 1978. 16. K. C. Dash and H. Schmidbauer, ‘Gold Complexes as Metallo-Drug’, in ref. 1, vol. 14, pp. 179-205. 17. B. M. Sutton, ‘Overviewand Current Status of Gold-Containing Anti-Arthritis Drugs’, in ref. 6, pp. 355-366. 18. A. E. Martell, (ed.),‘Inorganic Chemistry in Biology Medicine’, ACS Symposium Series 140, American Chemical Society, Washington, DC 1980. 19. C. F. Shaw, 111, ‘Gold Binding to Metallothioneins and Possible Biomedical Implications’, in ref. 18, pp. 349-372. 20. M. T. Razi, G. Otiko and P. J. Sadler, ‘Ligand Exchange Reactions of Gold Drugs in Model Systems and in Red Cells’, in ref. 6, pp. 371-384. 21. R. C. Elder, M. K. Eidsness, M. J. Heeg, K. G. Tepperman, C. F. Shaw, 111 and N. Schaeffer, in ref. 20, pp. 3855400. 22. D. H. Brown and W. E. Smith, in ref. 20, pp. 401-418. 23. P. M. Ford, J.RhewnatoI., 1984, 11,259-261. 24. J. 0. Nriagu (ed.),‘Copper in the Environment, Part 11, Health Effects’, Wiley, New York, 1979, pp. 83-162. 25. J. R. J. Sorenson, in ref. I, vol. 14, pp. 77-124. 26. K. B. Menander-Huber and W, Huber, in ‘Superoxide and Superoxide Dismutases’, ed. A. M. Michelson, I. M. McCord and I. Fridovich, Academic, London, 1977, pp. 537-549. 27. R. Lontie (ed.), ‘Copper Proteins and Copper Enzymes’, CRC Press, Boca Raton, FL, 1984, vols. I, I1 and 111. 28. Ref. 1, vol. 16, pp. 1-26. 29. N. J. Birch and P. J. Sadler, ‘Inorganic Elements in Biology and Medicine’, Specialist Periodical Report, ‘Inorganic Biochemistry’, Chemical Society, London, 1979, vol. 1, pp. 357-421. 30. Norman Kharasch (ed.), ‘Trace Metals in Health and Disease’, Raven, New York, 1979. 31. W. Mertz, Science, 1981,213, 1332. 32. C. F. Mills, Chem. Br., 1979, 15, 512. 33. E. J. Underwood, ‘Trace Elements in Human and Animal Nutrition’, 4th edn., Academic, New York, 1977. 34. W. G. Hoekstra, J. W.Suttie, H. E. Ganther and W. Mertz (eds.), ‘Trace Element Metabolism in Animals-2’, University Park, Baltimore, MD, 1974. 35. M. Kirchgessner (ad.), ’Trace Elements Metabolism in Man and Animals’, Arbeitkreis fur Tierernahrungsforschung Weihenstephan, 1978. vol. 3. 36. K. Schwarz, in ‘Clinical Chemistry and Chemical Toxicology of Metals’, ed. S. S. Brown, Elsevier/North Holland, New York, IY77, pp. 3-22. 37. C. A. McAuiiffe, ‘Techniques and Topics in Bioinorganic Chemistry’, Wiley, New York, 1975. 38. M. N. Hughes, ‘The Inorganic Chemistry of Biological Processes’, 2nd edn., Wiley, New York, 1981. 39. K. N. Raymond, ‘Bio-inorganic Chemistry 11’, ACS Advances in Chemistry, Series No. 162, American Chemical Society, Washington, DC, 1977. 40. F. H. Nielson, in ‘Inorganic Chemistry in Biology and Medicine’, ed.A. E. Martell, American Chemical Society, Washington, DC, 1980, pp. 23-42. 41. J. K. Aikawa, ‘Magnesium: Its Biologic Significance’, CRC Press, Boca Raton, FL, 1981. 42. W. X. Cheung (ed.),‘Calcium and Cell Function I Calmodulin’, Academic, New York, 1980. 43. D. Burrows (ed.), ‘Chromium, Metabolism and Toxicity’, CRC Press, Boca Raton, FL, 1983. 44. D. Shapcott and J. Hubert (eds.), ‘Chromium Nutrition and Metabolism’, Elsevier/North-Holland Biomedical Press, Amsterdam, 1979. 45. J. E. Spallholz, J. L. Martin and H. E. Ganther (eds.), ‘Selenium in Biology and Medicine’, AVI, Westport, CT, i981. 46. J. M. Bowen, ‘Environmental Chemistry of the Elements’, Academic, New York, 1979. 47. V. F. Fairbanks, J. L. Fahey and E. Beutler, ‘Clinical Disorders of Iron Metabolism’, 2nd edn., Grune and Stratton, New York, 1971. 48. A. S. Prasad, ‘Trace Elements and Iron in Human Metabolism’, Plenum, New York, 1978. 49. A. Jacobs and M. Worwood (eds.), ‘Iron in Biochemistry and Medicine’, Academic, London, 1974. SO. A. S. Prasad (ed.), ’Trace Elements in Human Health and Disease’, Academic, New York, 1976, vol. 1, ‘Copper and Zinc’. 51. A. S. Prasad (ed.), ‘Trace Elements in Human Health and Disease’, Academic, New York, 1976, vol. 2, ‘Essential and Toxic Elements’.

778 52. 53. 54. 55. 46. 57. 58.

Biological and Medical Aspects

A. S. Prasad, ‘Zinc Deficiency and its Therapy’, in ref. 1, vol. 14, pp. 37-55. G . J. Brewer, ‘The Pharmacological Use of Zinc’, in ref. 1, vol. 14, pp. 57-75. Z. A. Karciovglu and R. M. Sarper (eds.), ‘Zinc and Copper in Medicine’, Charles C. Thomas, Springlield, IL, 1980. D. R. Williams, ‘The Metals of Life’, van Nostrand Reinhold, London, 1971. B. Sarkar, in ref. 1, vol. 12, pp. 233-281. C. A. Owen, Jr., ‘Copper Deficiency and Toxicity’, Noyes Publications, New Jersey, 1981. W. G. Levine (ed.), ‘The Chelation of Heavy Metals’, in ‘International Encyclopedia of Pharmacology and Therapeutics’, Pergamon, Oxford, 1979, Section 70. 53. F. P.Dwyer and D. P.Mellor (eds.), ‘Chelating Agents and Metal Chelates’, Academic, New York, 1964. 60. M. M. Jones, ‘Therapeutic Chelating Agents’, in ref. 1, vol. 16, pp. 47-83. 61. H. E. Howard-Lock, A. Mewa, W. F. Kean and C. J. L. Lock, ‘D-Pencillamine: Chemistry and Clinical Use in Rheumatic Disease’, Seminars in Arthritis and Rheumatism, 1986, 15 (4), 261. 62. D. R. McLachlan, personal communication, 1984. 63. P. M. Farmer, A. Peck and R. D. Terry, J. Neuropathol. Exp. Neurol., 1976, 35, 367. 64. D. R. Crapper, S. S. Krishnan and S. Quittkat, Braitz, 1976,99, 67-80. 65. D. P. Ped and A. R. Brody, Science, 1980,208,297-299. 66. D. R. Crapper, S. S. Krishnan and A. J. Dalton, Science, 1973, 180, 511-513. 67. D. R. Crapper and U. DeBoni, Can. Psychiatr. Assoc. J., 1970, 23,229-233. 68. D. R. Crapper, S. Quittkat, S. S. Krishnan, A. J. Dalton and U. DeBoni, Acta Neuropathol. (Berlinj, 1980,50,19-24. 69. U. DeBoni, J. W. Scott and D. R. Crapper, Histochemisiry, 1974,40,31-37. 70. J. R. McDermott, A. I. Smith, K. Iqbal and H. M. Wisniewski, Neurology, 1979,29, 809-814. 71. J. R. McDermott, A. I. Smith, K. Iqbal and H. M. Wisniewski, The Lancet ii, 1977, 710-711. 72. D. R. Crapper, S. Karlik and U. DeBoni, Aging, 1978, 7 , 471-485. 73. D. R. Crapper and U. DeBoni, in ‘The Biochemistry ofDementia’, ed. P.J. Roberts, Wiley, London, 1980, pp. 53-68. 74. S. J. Karlik, G. L. Eichhorn, P. N. Lewis and D. R. Crapper, Biochemistry, 1980,19, 5991-5998. 75. S. J. Karlik, G. A. Elgavish and G. L. Eichhorn, J. Am. Chem. Soc., 1983,105,h02-609. 76. I. 3 . Neilands, S t r u t . Bonding (Berlin), 1966, 1, 59-108. 77. D. R.McLachlan, A. J. Dalton and H. Galin, Presentation at Canadian and American Gerontological Association Meetings, Toronto, Ontario, November 1981. 78. Y. Ihara, C. Abraham and D. J. Selkoe Narure (Londun), 1983, 304,727-730. 79. D. D. Perrin and H. Stunzi, ‘Metal Ions and Chelating Agents in Antiviral Chemotherapy’, in ref. 1, vol. 14, pp. 207-241. 80. M. N.Hughes and N. J. Birch, Chem. Br., 1982,lS (3),9. 81. N. J. Birch, ‘Lithium in Psychiatry’, in ref. I, vol. 14, pp. 257-313. 82. D. C. Tosteson, Sci. Am., 1981, 244 (4), 164. 83. A. Albert, ‘Selective Toxicity’, 4th edn., Methuen, London, 1968. 84. J. B. Taylor and P. D. Kennewell, ‘Introductory Medicinal Chemistry’, Ellis Honvood, Chichester, England, 1981. 85. G. deStevens (ed.), ‘Medicinal Chemistry’, vol. Ill, ‘Drug Design’, ed. E. J. Ariens, Academic, New York, 1972. 86. P. F. D’Arcy and J. C. McElnay, ‘Drug Metal Interaction in the Gut’ in ref. 1, vol. 14, pp. 1-35. 87. L. D. Petfit and R. J. W. Weffort, in ref. 1, vol. 9, pp. 173-212. 88. J. M. Hill, E. Loeb, A. S. Pardue, A. Khan, N.D. Hill, J. J. King and R,W. Hill, in ref. 4, Part 11, J. Clin. Hematol. Oncol., 1977, 681-700. 89. Y. Kidani, K. Inagaki, R. Saito and S. Tsukagoshi, in ref. 4, Part 4,J. Clin. Hemaml. Oncol., 1977, 197-209. 90. L. Pasteur, ‘Researches on Molecular Asymmetry’ (1860), Alembic Club Reprints No. 14, Edinburgh, Scotland, 1905. 91. E. J. Ariens, W. Soudijov and P. B. M. W. W. Timmermans (eds.), ‘Stereochemistry and Biological Activity of Drugs’, Blackwell, Oxford, 1983. 92. G. Blaschke, H. P. Kraft, K. Fickentscher and F. Kohler, Arzneim.-Forsch., 1979.29 (11) (IO), 1640-1641. 93. R. A. OReilly, W. F. Trager, C. H. Motley and W. Howald, J. Clin. Invest., 1980,65,746. 94. J.’ R.Vane and M.R. Robinson (eds.), ‘Prostaglandin Synthetase Inhibitors’, Raven, New York, 1974. 95. E. D. J. Lee, K. M. Williams, G. G. Graham, R. 0. Day and G. D. Champion, J. Pharm. Sci., 1984,73, 1542. 96. D. G. Kaiser, G. J. Vangeissen, R. J. Reisner and W. J. Wechter, J. Pharm. Sci., 1976, 65, 269. 97. A. Rubin, M.P. Knader, P. P. K. Ho, L. D. Bechtol and R.L. Wolen, J. Pharm. Sci., 1985,74 (I), 82. 98. A. J. Hull and J. Caldwell, J. Pharm. Pharmacoi., 1985,35, 693. 99. H. E. Howard-Lock, C. J. L. Lock and W. F. Kean, ‘Doesthe left hand know what the right hand is doing? or Clinical pharmacology: the use of optically pure ( R or S ) forms of chiral drugs rather than racemic mixtures’. Submitted to J, Rheumaid, 1985.

63 Application to Extractive Metallurgy M. J. NICOL, C. A. FLEMING and J. S. PRESTON Council for Mineral Technology, Randburg, Transvaal, South Africa 63.1 INTRODUCTION

779

63.2 MlNERAL PROCESSlNG 63.2.I Collectors 63.2.2 Activators 63.2.3 Depressants

780 781 782 783

63.3 HYDROMETALLURGY 63.3.I Leaching Processes 63.3.1.1 Gold and silver 63.3.1.2 Copper 63.3.1.3 Nickel and cobalt 63.3.1.4 Lead 63.3.1.5 Aluminum 63.3.1.6 Uranium 53.3.2 Solvent-extraction Processes 63.3.2.1 Solvent extraction of base metals by carboxylic acidr 63.3.2.2 Solvent extraction of base metals by organophosphorus acids 63.3.2.3 Solvent extraction of copper and nickel by ortho-hydroxyoximes 63.3.2.4 Solvent extraction of base metals by amine salts 63.3.2.5 Solvent extraction of platinum-group metals and gold 63.3.2.6 Solvent extraction of base metals by solvating extractants 63.3.3 Ion-exchange Processes 63.3.3.1 Cation-exchpnge resins 63.3.3.2 Anion-exchange resins 63.3.3.3 Chelating resim 63.3.3.4 Future trends 63.3.4 Precipitation 63.3.4.1 Oxides 63.3.4.2 Sulfides 63.3.4.3 Metals 63.3.5 Electrochemicai Processes

783 783 784 785 786 787 787 788 788 789 792 799 802 806 810 814 815 817 823 826 827

827 828 828 83 1

63.4 PYROMETALLURGY

832

63.5 REFERENCES

835

63.1 INTRODUCTION The extraction of metals from ores is a complex procedure involving a number of physical and chemical separation and purification steps that are designed to produce a pure metal or metal compound with maximum yield at the lowest cost. A simplified diagram illustrating the major processing steps and their interrelationships is shown in Figure 1. The complexity of such processes can be appreciated if one considers the large number of metals that must be economically recovered from ores of widely differing composition and concentration, such as aluminum from bauxite with an A1203content of 50-55%, or gold from an ore in which the gold content in the metallic state is as low as 1O-3%. Extractive metallurgy can be conveniently classified into the following three main activities: (i) mineral processing or ore-dressing, which involves the separation and concentration of the minerals from the unwanted or gangue material by a variety of physical and physicochemical techniques that exploit characteristics such as colour, specific gravity, shape and size, charge, magnetic susceptibility and surface properties; (ii) hydrometallurgy, which, as the name implies, involves the processing of an ore or concentrate by the dissolution, separation, purification and precipitation of the desired metal by the use of aqueous solutions; and (iii) pyrometallurgy, which produces metals or intermedate products directly from the ore or concentrate by the use of high-temperature oxidative or reductive processes. 779

780

Application to Extractive Metallurgy Ore

Leach

Crush or grind

Gangue

Gangue

Leach

Physi@ separation

angue

oncentrate Gangue

Leach

IPyro-pretreatment

Pyro-pr-ing

Treated concentrate Gangue Leach

-

Matte or metal

Solution

Eurification5

refining of

lpure solution

I

I

Ppt. of metal Byproduct

Intermediate product to market

'Metal 10 marker

Metal to market

Figure 1 Metallurgical process routes

While it is certainly true that coordination chemistry has its greatest application in hydrometallurgy, several aspects of the ore-dressing and pyrometallurgical operations involve the application of coordination chemical principles, and the indications are that these areas can be expected to develop in the future as our understanding of the chemistry of surfaces and of high-temperature systems develops. For convenience, the following sequence and breakdown will be adopted (1) Mineral processing with the emphasis on flotation, (2) Hydrometallurgy - Leachmg - Separation Processes - Precipitation, and (3) Pyrometallurgy with the emphasis on the chemistry of slag systems.

63.2 MINERAL PROCESSING

Materials mined from a mineral deposit usually consist of a heterogeneous mixture of solid phases that are generally crystalline and contain various minerals, Crushng and grinding operations are used to liberate the mineral species from one another and to reduce the size of the solids to a range suitable for subsequent processing. Of the various separation techniques, those of froth flotation and agglomeration exploit the chemical and physical properties of the surfaces of minerals, which can be controlled by various chemical interactions with species in an aqueous phase. The flotation process is applied on a Iarge scale in the concentration of a wide variety of the ores of copper, lead, zinc, cobalt, nickel, tin, moIybdenum, antimony, etc., which can be in the form of oxides, silicates, sulfides, or carbonates. It is also used to concentrate the so-called non-metallic minerals that are required in the chemical industry, such as CaF2, BaS04, sulfur, Ca3(P03)2,coal, etc. Flotation relies upon the selective conversion of water-wetted (hydrophilic) solids to nonwetted (hydrophobic) ones. This enables the latter to be separated if they are allowed to contact air bubbles in a flotation froth. If the surface of the solids to be floated does not possess the requisite hydrophobic characteristic, it must be made to acquire the required hydrophobicity by the interaction with, and adsorption of, specific chemical compounds known as collectors. In separations from complex mineral mixtures, adltions of various modifying agents may be required, such as depressants, which help to keep selected minerals hydrophilic, or activators, which are used to reinforce the action of the collector. Each of these functions will be discussed in relation to the coordination chemistry involved in the interactions between the mineral surface and the chemical compound.

Application to Extractive Metallurgy

781

63.2.1 Collectors

Flotation collectors are generally surface active amphipatic molecules, R-X, consisting of a polar group X attached to a non-polar hydrocarbon group R.The interaction of the collector with the mineral surface occurs through the polar group, whereas the non-polar group confers the hydrophobicity to the surface. The polar groups utilized in flotation collectors generally have the ability to coordinate to the metal ion constituting the lattice of the mineral. Typical groups are -OH, -C02H, -SH, -0CS2- (xanthate), -02PS2- (dithiophosphate), -NH2 and -SO3-, of which the most widely used are the fatty acids for the flotation of oxides and the thio compounds for the recovery of sulfide minerals. The most important raw materials for the production of non-ferrous metals, such as copper, lead, zinc, nickel, cobalt, molybdenum, antimony and cadmium, are the sulfide minerals. The use of collectors containing various thio-type functional groups has proved to be the most successful in the flotation of these minerals. Some of these compounds are shown in Table 1. Table 1 Common Thio Collectors

Thio ukrivative

Parent compound

Alcohols ROH

Thiols

RSH

Carbonic acid Alkyl thiocarbonates

/p

HOC

/p

ROC

\

OH

\S-Mt Alkyl dithiocarbonates (xanthates)

HS ROC 'S-M+

Carbamic acid H

\

//"

H/N-C\oH

R

Alkyl thioncarbamates

\

H/N-C\

Phosphoric acid HO

\yo

RO

H

b 'OH

RO

Xanathate derivatives

NS OR2

Dialkyl dithiophosphates

\/ / \

S-M+ Mercaptobenzothiazoles

These collectors are effective only under oxidizingconditions, and it is generally accepted'J that the species that confers hydrophobicity on the mineral surface is either a chemisorbed metal thio compound or the oxidized form of the collector, dithiolate. The amounts of each species formed will depend on the relative stabilities of the metal-sulfur and sulfur-sulfur bonds. The formation of four-membered chelate rings is also possible with soft metal ions such as copper(1) because the largely covalent character of the bond in this instance is able to overcome the strain within the ring by extensive electron delocalization. This could account for the partial selectivity of some of these reagents for the copper minerals,2which has been put to good use in the sequential flotation of copper, lead and zinc from complex sulfide ores.2 Chelating agents that can form insoluble, hydrophobic chelates on the surface of minerals are potential collectors for the selective flotation of mineral^.^^^ As early as 1927, Vivians reported the use of cupferron, a well-known analytical reagent, as a collector for the flotation of cassiterite @noz).Since then, there have been a number of reports on the use of chelating agents in flotation.

782

Application to Exrractive Metallurgy

The selectivity potentially offered by these reagents has attracted the attention of many workers who have proceeded on the assumption that these reagents will adsorb onto the surface of minerals with a selectivity for the metal ions of the lattice similar to that found in aqueous solution. Examples of this are dimethyglyoxime for nickel,6 cupferron for tin,5 salicylaldoxime for copper and 9-hydroxyquinoline for lead and zinc mineral^.^ The use of chelating solvent extractants such as the hydroxyoximes for the flotation of oxidized copper ores has been reported;8,9 unfortunately, although laboratory tests have shown that the use of chelating collectors has potential, they are not yet in widespread commercial use. A number of attempts have been made to understand the mechanism of the adsorption of chelates on oxide minerals. For instance, IR spectroscopic studies'o have indicated the presence of a basic monosalicylaldoximate copper complex as well as the bis-salicylaldoximate complex on the surface of malachite (basic copper carbonate) treated with salicylaldoxime.However, other workers4have shown that the copper chelate is partitioned between the surface and dispersed within the solution, and that a dissolution-precipitation process is responsible for the formation of the chelate. Research into the chemistry of the interaction of chelating collectors with mineral surfaces is still in its infancy, and it can be expected that future developments will depend on a better understanding of the surface coordination chemistry involved. 63.2.2 Activators

The most widely applied activation procedure is that involving the use of copper(I1) ions to enhance the floatability of some sulfide minerals, notably the common zinc sulfide mineral sphalerite.2 Sphalerite does not react readily with the common thiol collectors, but after being treated with small amounts of copper it floats readily owing to the formation of a surface layer of CuS." A similar procedure is often adopted in the flotation of pyrrhotite (FeS), pyrite (FeS2), galena (PbS) and stibnite (Sb2S3).In the context of coordination chemistry, the major contribution has been to the understanding of the chemistry involved in the deactivation of these minerals, a procedure often adopted in the sequential flotation of several minerals from a complex ore. A deactivating agent for copper-activated sphalerite is any species that has sufficient affinity for copper(1) or (11) to compete for it with sulfide ions in the surface lattice of the mineral, thus removing it from the surface. Ligands such as cyanide or ethylenediamine, which coordinate strongly to copper, have therefore been found to be the most effective. A knowledge of the stability of the species present in a system composed of H+,Zn2+,Cu2+,S2- and CN- ions ha5 enabled12 the extent of deactivation by cyanide ion to be predicted; the results of these predictions are compared with experimental observations in Figure 2. This approach has been successfully extended to the effects of pH and the presence of other ions such as carbonate on the activation and deactivation processes, and is a pertinent example of the quantitative application of coordination chemistry to complex systems. Obwvcd

-Calculated

Figure 2 Variation of the extent of deactivation of copper-activated sphalerite as function of the cyanide concentration (after Marsicano et al.,ref. 12)

Application to Extractive Metallurgy

783

The inadvertent activation of a mineral can often result in an undesirable response to flotation.

A typical example is quartz, a highly unreactive mineral whose recovery by flotation is therefore difficult. However, in the presence of metal cations, it can be activated to respond to anionic collectors, as shown by the datal3 in Figure 3. It is of interest that the pH value at which flotation is effective can be related to that at which the monohydroxy complexes of the metal ions are predominantly stable, suggesting a specific adsorption of these species onto the quartz surface. In some cases this can be used for the selective depression of the quartz minerals in, for example, the flotation of pyrite in the presence of cationic collectors such as the long-chain amines. pH o f hydroxy complex formation FeOH”

PbOH‘

MgOH+

PH

Figure3 Minimum flotation edges of quartz as a function of pH; 1 x Fuerstenau and Palmer, ref. 13)

M sulfonate, 1 x

M metal ion (after

63.2.3 Depressants The mechanisms of the action of the inorganic depressants are reasonably well understood. The following are examples. (i) If cyanide is used to suppress the flotation of the sulfide minerals, the greater stability of the cyano complexes over the corresponding thiolates results in the preferential formation of hydrophilic metal cyanide layers for metals such as zinc, copper, iron, nickel, cobalt and even antimony. (ii) Galena (PbS) is depressed with chromate ions by a mechanism similar to that for cyanide. (iii) Soluble silicates are used to disperse and to depress the siliceous gangue minerals. The organic depressants are usually natural products or modified natural products of high molecular mass, and contain numbers of strongly hydrated polar groups that are the basis of their depressant action. Although some of these compounds contain coordinating centres such as hydroxyl, carboxyl and amino groups, the mechanisms whereby these reagents operate is obscure and, as in the case of the chelating collectors, further research into the coordination chemistry of the interaction of these compounds with mineral surfaces is required.

63.3 HYDROMETALLURGY 63.3.1 Leaching Processes The primary step in a hydrometallurgical process is the dissolution (selectively if possible) of the desired component{s) of the ore, concentrate, or pyrometallurgically produced intermediate. The coordination chemistry of the metal ions produced during leaching is of paramount importance in our understanding of the leaching reactions and in the optimization of the extent of dissolution and of the costs involved. The following list (Table 2) gives an indication of the metals recovered by leaching and of the types of ligands that are employed and those that are potentially useful. These generally occur in nature as minerals of the oxide or sulfide types, the former including complex oxides such as silicates, carbonates or phosphates. In some cases, such as the precious metals (gold, silver and the platinum group), the elemental form is naturally occurring. While the oxide-type minerals are mostly ionic solids with well-defined stoichiometries and structures, the sulfide minerals exhibit a variety of different structures that are attributable to the essentially covalent character of t h e bonding in most transition metal sulfides. However, the semimetallic properties of many sulfides show that the degree of covalency vanes from a small fraction for zinc sulfide to almost 100% for the copper and iron sulfides.l 4 In complex sulfides such as chalcopyrite (CuFeS2), an additional complication is that metal atoms of more than one kind and more than one oxidation state are present. A further characteristic of sulfide minerals is that the composition

Application to Extructive Metallurgy

784

Table 2 Metals That Can Be Recovered by Leaching and Types of Ligands Employed Metal

Metal

Ligand

Zr, Nb, Mo, Hf, Ta, W Ru, Rh, Pd, Os, Ir, Pt Aa

C1-, OH-

Ligand ~~

Ai Ti Cr Mn Fe co Ni Cu Zn, Cd Se, Te

OH-, C1-

c1so42-,c1so42so42-,c1-

AU

Sn Sb Pb Lanthanides

S042-, C1-, NH3 Sod2-, Cl-, NH3 S042-, C1-, NH3, MeCN

CINH3, CN-, MeCN, C1CN-, (NH&CS, S203'-, SCN-, C1OH-, C1-

u

so42-

c1-

c1so420 ~ , 2co32-

OH-, Cl-, MeC02~

rarely conforms to the ideal formula, partial substitution of one kind of atom for another (e.g. silver for copper or lead) being common. The chemistry of leaching processes can be broadly classified into two main reaction types: non-oxidative and oxidative (or reductive) processes. In general terms, non-oxidative dissolution occurs when the formal oxidation states of the elements of which the mineral is composed are unchanged during leaching. Examples of such reactions are the leaching of zinc oxides in acid solutions, the dissolution of aluminum oxides in concentrated alkaline solutions, and the dissolution of iron sulfides in acidic solutions to produce hydrogen sulfide. Oxidative (or reductive) dissolution occurs when there is a change in the formal oxidation state of one of the elements during the leaching process. Examples of such reactions include the leaching of uranium dioxide in acidic solutions by the oxidation of uranium(1V) to uranium(VI), of sulfide minerals by the oxidation of sulfur(-11) to the elemental or higher oxidation states, and of gold(0) to gold(1) in alkaline cyanide solutions. It should be clear that the coordination chemistry of both the ions constituting the mineral (or metal) and of the oxidizing (or reducing) agent is a crucial component in the successful dissolution of a mineral and in our understanding of the mechanisms of these reactions. Attention in the following discussion will be focused on the coordination chemistry of several of these leaclung processes as applied in the metallurgical industry and on others that show some potential for future application. 63.3.1.1 Gold and silver Gold is the most noble of metals (the standard reduction potential for the Au/Au+ couple is 1.73 VI5). This property is, of course, the main reason for its durability and for its occurrence in nature usually as the native metal. Despite its generally high resistance to corrosion, gold can be dissolved freely in solutions containing appropriate complexing species. The most stable complexes of gold(1) and gold(II1) are those formed with the cyanide ion (Bz(Au(CN)*-) = 2 x and p4 (Au(CN),-) z 1056),16 and it is therefore not surprising that the majority of industrially important processes involving gold in solution are based on the use of the relatively inexpensive cyanide ligand. In particular, the cyanidation of gold and silver ores has been the most widely employed process for the recovery of these metals for over 100 years. In the dissolution of gold (or silver) by cyanidation, the milled ore is agitated with dilute alkaline cyanide solution, and the air is introduced (equation 1). 4Au

+ 8CN- +

02

-

+ ZHzO

4Au(CN),-

+

40K

(1)

This reaction has been shown17to occur by a typical electrochemicalcorrosionmechanism in which the anodic dissolution of the gold, i.e. Au

+ 2CN-

Au(CN),-

+

e-

(2)

is coupled to the cathodic reduction of oxygen (with peroxide as an intermediate product of reduction). The exceptional stability of the aurocyanide complex is such that dissolution can be achieved even in the presence of considerable amounts of other metals such as copper, zinc and nickel in the ore. The above anodic reaction (equation 2) has been extensively studiedlg and, in terms of its application, the most significant observation is the passivation of the gold surface, as shown by the peak at -0.4 V in the linear-sweep voltammogram, and of a gold electrode in cyanide solution (Figure 4). It is generally agreed that passivation is due to the formation of a strongly adsorbed layer with the stoichiometry AuCN on the surface of the gold. This is consistent with the known19linear polymeric structure of gold(1) cyanide in the solid state, in which cyanide acts as a bridging ligand. It is of interest that silverv), which does not form such a polymeric

Application to Extractive MetalIurgy

185

structure, does not exhibit passivation, whereas palladium(II), which forms a two-dimensional polymeric structure, exhibits a degree of passivation so severe that it cannot be effectively leached in cyanide solutions. Fortunately, the presence of heavy-metal ions such as those of lead, thallium or mercury in the solution prevents the passivation of gold and, in practice, it is probably seldom encountered as a processing problem. Mechanisms for the action of these ions have been suggested.

_..... -.O.I -0

M KOH I M KOH and

0.1 M KCN

Potential (V)

Figure4 Comparison of the ancddic khaviour of a rotating disc gold electrode in 0.1 M KOH with that in a solution of 0.1 M KOH and 0.1 M KCN (after Nicol, ref. 18)

As a result of increasingly stringent environmental regulations, much attention has been devoted in recent years to the possibility that gold and silver can be leached by ligands other than cyanide. Other considerations, for example the presence of excessive quantities of soluble copper in the ore or the occurrence of gold in refractory (to cyanidation) minerals such as arsenopyrite or stibnite, have also prompted the search for alternatives to cyanide. Of the various possibilities, the one most studied2' and applied is the use of thiourea in acidic solutions with iron(II1) as the oxidant. The principal leaching reaction is Au

+ 2NH2CSNH2 + Fe3+ +

Au(NH,CSNH&+

+ Fe2+

(3)

Extraction is rapid compared to that by cyanidation, and the gold surface is not subject to the passivation observed in alkaline cyanide solutions, At present, the main disadvantage of the process is the htgh cost of the reagents, which is compounded by the instability of thiourea with respect to hydrolysis and to oxidation under the conditions of the leach. Other ligands that have been suggested and tested, at least in the laboratory, for the leachng of gold are chloride22,thiocyanatG3 and polysulfide^^^. Of these, thiocyanate can be expected to play an important role in the future in possible processes for the combined leaching of gold and uranium. Silver chloride is a common source and intermediate product in many extractive metaliurgkal processes, for example it occurs in the anode slimes from copper refineries, the residues of leaching processes for base metals, as a product of the chlorination of impure gold-silver bullion, and in photographic waste. A novel process for the leachng and purification of silver chloride, which was devised by Parker et a1.,26 is based on the observation that silver chloride is very soluble in some dipolar aprotic solvents containing chloride ion but is much less soluble when water is present. The very different khaviour oE the equilibrium AgCl(s)

+ C1-

+ AgCl2-

(4)

in water versus dipolar solvents is a function of the slightly stronger solvation of AgC12- and the very much weaker solvation of the chloride ion by the aprotic solvents. These factors can account for an increase of up to 9 log units in the equilibrium constant for the above reaction, therefore dimethyl sulfoxide (DMSO) saturated with calcium chloride (1.4 M) will dissolve almost 200 g of silver per litre. The addition of water to produce a solution with a DMSO content of less than 70% (v/v) results in the precipitation of more than 99% of the silver as a pure product. 63.3.1.2 Copper The development of hydrometallurgical processes for the extraction of copper from ores and concentrates has been a widespread major research activity for the past 20 years. Coordination chemistry has been applied in many ingenious ways in attempts to devise novel and energy-efficient processes to replace the conventional smelting-electrorefining approach. Several of these processes

786

Application bo Extractive Metallurgy

have achieved commercial application, while others offer prospects for future development. Ammoniacal leaching processes will be discussed in a later section. Processes based on chloride have received much attention, and several plants currently employ leachng reactions based on the use of chloride.27The high stability of copper(1) ion in strong chloride solutions results in a significant change in the chemistry of the copper in such solutions. Copper(I1) is therefore a relatively powerful oxidizing agent in concentrated chloride solutions and, in 5 M chloride, is comparable thermodynamically to the iron(II1) ion.28In addition, the oxidation of copper(1) to copper(l1) by oxygen is rapid,29and t h s results in considerably enhanced rates in the oxidative leachng of, for example, sulfide minerals by oxygen in chloride solutions containing copper ions. Furthermore, in the oxidative leaching of copper-containing sulfide minerals, the use of copper(I1) as the oxidant in chloride solutions avoids the use of iron(lI1) with all its attendant problems of iron removal in subsequent processing. The most common copper sulfide mineral is chalcopyrite, and the predominant reaction with copper(I1) in chloride solutions is30 CuFeSz

+

3Cu"

-

4Cu'

+ Fe" +

2s

(5)

As will be discussed later, the recovery of copper from a solution containing copper(1) provides an attractive alternative in terms of energy consumption, and this has led to the search for other ligands that stabilize copper(1). Of these, the nitriles, RCN, provide some interesting and potentially useful processing options, many of which have been explored by Parker and Muir.31The nitriles, being rather weak bases, interact much less strongly with a limited number of soft metal ions capable of donating &electrons into an antibonding x-orbital of the nitrile group. Therefore acetonitrile solvates copper(1) and silver more strongly than does water, and this provides the basis for a leaching process for copper metal (impure blister copper produced by smelting, scrap, or from other sources of crude copper). cuo

+ CUI1 + 2cu'

(6)

The variation of the equilibrium constant for this reaction in acetonitrile-water mixtures containing dilute sulfuric acid is shown in Figure 5. At pH values below about 5, the curve is more or less independent of acidity because the proton has no significant interaction with acetonitrile in the presence of water. This is an important factor, and distinguishes the nitriles from other ligands such as ammonia and pyridine, which are stronger bases. It should be apparent from the data in Figure 5 that the copper can be effectively leached with copper(T1) solutions containing, say, 50% acetonitrile, whereas removal of the acetonitrile by distillation will reverse the equilibrium and result in the precipitation of pure copper powder and regeneration of the copper(I1) solution for recycle to the leaching step. This process has undergone extensive pilot-plant evaluation, but has not been implemented on a commercial scale owing to the depressed state of the copper market over the past five to ten years. In theory, copper sulfide and oxide minerals can be leached by the use of a suitably modified version of the above chemistry.

H:O

('70

by mass)

Figure 5 Variation of the equilibrium constant for the reaction Cuo + CUI' (after Muir and Parker, ref. 31)

2Cu1in mixtures of water and acetonitrile

63.3.1.3 Nickel and cobalt

Since ammonia forms stable, water-soluble complexes with many metals, leaching can be carried out under alkaline conditions to give these metals in solution. Of particular interest are the metals copper, nickel and cobalt, which form particularly stable amines that have been well characterized as having the following approximate stability constants16 (at high ionic strength): Cu1,j2 = 11; = 5; Fel1,P2 < 2. Cu1I,p4 = 13; Nil1,& = 9; It is of particular significance from a processing point of view that iron(I1) forms weak complexes and that complexes of iron(II1) with ammonia have not been observed. As a result, iron is rejected

Application to Extractive Metallurgy

787

from the process as the oxide at the leaching stage, which gives this process a considerable advantage over other acid-based processes for these metals. This chemistry forms the basis for the successful Sheritt-Gordon ammoniacal leaching process,32 which is widely applied to the processing of nickel-cobalt-copper sulfide concentrates. A finely ground concentrate of sulfide minerals is leached in ammonia solution in the presence of oxygen to produce a solution containing the amines of copper, nickel and cobalt. As in the leaching of sulfide minerals in chloride solutions, the relatively high stability of copper(1) and its rapid oxidation to copper(I1) by oxygen results in considerable catalysis of the leaching process by copper ions. Purification of the leach liquor and its reduction to the metallic state will be reviewed in a later section.

63.3.1.4 Lead The non-oxidative dissolution of galena (PbS) in acidic solutions is an attractive first step in a hydrometallurgical process for the production of lead metal. The formation of relatively stable chloro complexes with lead ions (b3 = 2) results in a displacement of the equilibrium of the reaction PbS

+ 2H' + iC1-

F=+

PbCli(2-i)+

+ H,S

(7)

to the right, so that, at elevated temperatures, concentrations of lead of up to 100 g dmP3can be achieved. A knowledge of the formation constants of the various lead chloro complexeshas enabled on the basis of a kinetic model in which it is assumed the kinetics of reaction (7) to be predi~ted3~ that reaction (7) is at equilibrium at the sulfide surface and that the rate-determining step is the diffusion of the various products away from the surface. The results of the calculated and observed rates of dissolution are shown in Figure 6, and the relatively good agreement is further evidence for the invaluable role of coordination chemistry in the quantitative description of applied metallurgical processes.

'"1

. -0

Experimental 6 M H' Experimental 3 M H * Calculated 6 M H *

Figure6 Experimental and calculated rates of dissolution of galena as a function of the chloride concentration at 298 K (after Scott and Niml, ref. 33)

63.3.1.5

Aluminum

The dissolution of hydrated alumina in concentrated aqueous solutions of sodium hydroxide at elevated temperatures and pressures is the basic step in the extraction of aluminum from bauxite via the Bayer process.34This dissolution reaction is described by the equation AI2O3.xH20 + 2NaOH

-

ZNaAl(OH),

The high stability of the aluminate ion allows the production of concentrated solutions of aluminum with the virtual exclusion of the main metallic impurity, viz. iron as an oxide residue. The resultant impure aluminate solution is clarified and its temperature reduced when the reverse of the above reaction occurs with the formation of Al2O3.3H20by a slow crystallization procedure. The high-purity alumina trihydrate product is calcined and then reduced electrochemically in a molten fluoride bath by the well-known Hall-Heroult process. The major problems in the Bayer process have their origin in the coordination chemistry of aluminum in alkaline solutions. The CCCB-7.

788

Application to Extractive Metallurgy

most generally accepted structures are A1(OH)4pand Al(OH)d(H20)2-, since the aluminate ion is monomeric, monovalent and has a coordination number of either 4 or 6 . Extensive hydrogen bonding between the aluminate ion and the solvent is responsible for the high viscosity of these solutions, which creates major problems in materials handling and heat exchange. In addition, the well-known kinetic inertness of aluminum(II1) results in slow rates of crystallization (more than 50 h), requiring large vessels and large volumes of circulating solution and seed material.

63.3.1.6 Uranium

As the most common uranium-bearing minerals such as uraninite (UO?) are insoluble or dissolve extremely slowly in non-oxidizing aqueous solutions, industrial processes for the leaching of these minerals depend on a reaction that involves oxidation of the uranium from the tetravalent to the hexavalent state with the formation of soluble uranyl species.35Depending on the nature of the other constituents of the ore, dissolution is accomplished in acidic sulfate solutions with iron(II1) or chlorate ions as the oxidant or in alkaline carbonate solutions with oxygen or peroxide as the most commonly employed oxidants. It is currently generally accepted35that the mechanism of the dissolution reaction is electrochemical in nature, anodic oxidation of the semiconducting mineral being coupled to cathodic reduction of the oxidant on the mineral surface. The rate of leaching of U 0 2 by iron(1II) is significantly higher in sulfate than in perchlorate solutions35 and, at a pH value of 1, it passes through a maximum as the sulfate concentration is increased, as shown in Figure 7. Also shown is the effect of the sulfate concentration on the cathodic reduction of iron(II1) at a U 0 2 surface. The close similarity of the two curves demonstrates that the influence of the sulfate ion on the leaching rate is a cathodic phenomenon. Furthermore, the maximum, at a sulfate:iron(III) ratio of approximately 1 : l suggests that the species FeS04+(aq) is the most electroactive of the various iron(II1)-sulfato complexes under these conditions.

4

7

Cathodic reduction of Fe(lf11 potential 0.35 V

11 z'

Medium: O.02M Fe(lll1

0.05

(SO:

r40

2

I : 0.10

cr

I (MI

Figure7 Influence of sulfate on cathodic reduction of Fe"' and leaching of UOz by FelI1 (after Nicoi, Needes and Finkelstein, ref. 35)

Complex uranium ores are often associated with phosphate-bearing minerals, and the presence of soluble phosphate has been found to adversely affect the recovery of the uranium. Studies such as those outlined above have established that, although the iron(II1)-phosphonato complexes are more reactive, the decreased leaching rate in the presence of phosphate is due to the formation of insoluble, non-conducting layers of uranyl phosphate on the surface of the mineral. The uranyl ion forms unusually strong soluble complexes with the carbonate anion16(fl3 > 20), and this has formed the basis for the use of alkaline carbonate solutions in the leaching of uranium, particularly where the ore is leached directly from the ore body (leaching in situ). The rate of dissolution increases with increasing concentration of carbonate but reaches a limiting value at concentrations above about 1 M. It has been shown36that this is due to the rate-determining formation of a layer of hydrated U 0 3 on the surface, which subsequently dissolves by forming complexes with the carbonate ions. 63.3.2 Solvent-extraction Processes

Solvent extraction is a well-established method of separation in the field of extractive metallurgy. First applied to the extraction of uranium for nuclear purposes in the early 194Os, the technique currently now finds widespread use in the recovery of uranium, copper, zinc, cobalt, nickel,

Application to Extractive Metallurgy

789

vanadium, tungsten and molybdenum, as well as in separations amongst gold and the platinum-group metals, in separation amongst yttrium and the rare-earth metals, in the separation of niobium and tantalum, and in the separation of zirconium and hafnium. Although the origin of many of the commercial processes is undoubtedly empirical, subsequent studies of a fundamental nature have often enabled the underlying chemical principles to be clearly identified. Solvent-extraction systems can be classified conveniently according to the nature of the reaction involved in the extraction process, of which the following types can be identified. ( a ) Extraction by physical distribution involves a simple molecular distribution of the solute between the aqueous phase and an inert solvent (Le. one containing no atoms with electron-donor properties). Such a situation arises only for solutes that are scarcely solvated in the aqueous phase. Few, if any, such systems are found in the field of extractive metallurgy, but they are typified by, for example, the extraction of ruthenium tetroxide into a solvent such as carbon tetrachloride. (b) Extraction by solvation involves the replacement by the extractant of some or all of the coordinated water molecules from a metal cation (or hydrogen ion) to form a species that is soluble in the organic phase. A typical example is the extraction of uranyl nitrate by tri-n-butyl phosphate (TBP) UO,(H,O)F

+ 2TBP + 2N03-

-

U0,(N03)2(TBP)2 + 6H20

(8)

where the bars denote the presence of the species in the organic phase. Solvating extractants are most characteristically oxygen-containing organic molecules such as ethers, ketones and neutral organophosphorus compounds. (c) Extraction by anion exchange is characterized by the extraction of an anionic metal species in the form of an ion pair with a suitable cation, following the exchange with a simple (non-metallic) anion between the organic and aqueous phases. A typical example is the extraction of cobalt(l1) from chloride solutions by an organic solution of an amine salt: [CoCI:-]

+ 2R3NH+C1-

-

(R3NH+),[CoCI4*-]

+ 2C1-

(9)

Often, the cationic component of the ion pair is formed by the protonation of a compound of appreciable basicity, most typically a nitrogen-containing base. Alternatively, the cation may be of the quaternary ammonium, phosphonium or arsonium type. (d) Extraction by cation exchange involves the formation, by the extracted metal cation, of an electrically neutral coordination complex with the extractant by displacing another cation (most commonly a hydrogen ion). Frequently, the formation of the extractable complex involves chelation as, for example, in the extraction of copper(I1) by the enol form of a P-diketone:

When the cation displaced from the extractant is a hydrogen ion, as in reaction (IO), the extractability of the metal shows a distinct dependence on the pH value of the aqueous phase. Such systems are conveniently characterized by the pH value at which the concentrations of metal in the organic and aqueous phases are equal at a stated concentration of extractant, this being denoted the pHo.5value for the given metal. In the following sections, the most important commercial applications of solvent-extraction technology in the field of extractive metallurgy are discussed in terms of the coordination chemistry relevant to the extraction process. The sole criterion for the inclusion of a given process is that it should have been utilized on an industrial scale. It may be noted that several interesting extraction systems in the early stages of development have been excluded on this basis, despite the pertinence of the chemical principles involved. 63.3.2.1 Solvent extraction of base metals by carboxylic acids Carboxylic acids represent a group of readily available and relatively inexpensive extractants. They have found rather limited application in commercial processes, however, probably on account of their generally low selectivity and poor pH functionality. Nevertheless, they have been used for the removal of iron from the rare-earthmetal~,~~ separations the separation of copper from among yttrium and the rare earths,39 the recovery of indiumm and gallium,4* the removal of

Application to Extractive Metallurgy

790

impurities from cobalt electrolyte^,^^.^^ and the coextraction of cobalt and nickel before further processing44 The extraction of some metal ions by Versatic 10 acid (a commercial product consisting mainly of 2-ethyl-2-methyiheptanoicacid and similar highly branched isomers) is shown in Figure 8. The order of extraction of the metals investigated in a recent is T1"' > Pd" > Hg" z FelI1 > In111 > ThIV > U02" z A1111 > CUI1 > Pb" > YbIII > AgI > ZnII > NdllI > Ce"1 > La111 > CdlI1 > Ni" > Co" > Fe" x Mn" > Ca" > SrlI > Mg" z Ba". This series can be largely rationalized on the basis of the magnitude of the stability constants for complexation of the respective metal ions by a typical carboxylate anion such as acetate.45Earlier comparisons between such selectivity series and the order of precipitation of the metal hydroxide^^"^^ have a fundamental basis only insofar as RCOy' and OH- both belong to the class of ligands that are anions of a weak acid.48Clearly, such a comparison cannot account for discrepancies such as the relative positions of calcium, strontium and magnesium. In contrast, the poor extraction of magnesium can be understood readily in terms of the expected stability constants for formation of the carboxylate complexes, such a displacement of magnesium from the usual order (e.g. K , for the formation of the acetate Complexes Mg > Ca > Sr > Ba) commonly occurring with ligands of increased steric complexity, as a result of the small size of the Mg2+ion (0.64 A, as compared with 0.99 A for Ca2+ and 1.13 A for Sr2+).4*

4

5

6

7

8

liqoilibrium pH

Figure 8 Extraction of some divalent metals (0.05 M as nitrates) from 1.0 M sodium nitrate solutions by 0.50 M solutions of Versatic 10 acid in xylene at 20 "C

For the divalent metals of the first transition series, the order of extraction follows that expected on the basis of the Irving-Williams series49for the stability constants of the octahedral complexes, viz. Mn < Fe < Co < Ni < Cu > Zn. However, although copper(I1) can be readily separated from the remaining metals, mutual separations in the group nickel, cobalt, iron and manganese are not readily accomplished. For example, the difference between the pHo.5values for the extraction of nickel and cobalt was found to lie between 0.0 and 0.2 pH units for a range of different aliphatic and aromatic carboxylic acids." The removal of nickel from cobalt-containing solutions by extraction with carboxylic acids therefore requires a process consisting of many separate stages if a high eficiency is to be acheved, and is believed to have been applied on a commercial basis only in the Soviet Uni0n.~2743 Although it is evident from electronic spectra that nickel(I1) and cobalt(I1) form complexes of octahedral s y m e t r y with carboxylic acids,50the exact stoichiometry of the extracted complexes has not been fully established. However, it is apparent that, in addition to the coordination of the two carboxylate anions (A-) required to give an electrically neutral complex, further coordination of undissociated carboxylic acid ligands (HA) occurs so that the metal ion can achieve3ts preferred octahedral coordination, complexes of the type N ~ A z ( H Aand ) ~ C O A ~ ( H Abeing )~ f ~ r m e d . ~ ' ~ ~ Oligomerization of the complexes presents an alternative means for the achievement of an octahedral coordmation, and the presence, in organic extracts, of dimeric compounds of the type Ni2A4(HA)4and Co2A4(HA),has been suggested on the basis of metal distribution data. However, no consideration appears to have been given to the structure of such c o r n p l e x e ~ . ~ ~ - ~ ~ The Sumitorno Metal Mining Co. in Japan uses a carboxylic acid extractant to recover nickel and cobalt from a solution obtained from the pressure leaching of metal sulfide concentrates with sulfuric acid.44Iron, copper, zinc and manganese are removed from the leach liquor by conventional precipitation methods prior to the extraction of nickel and cobalt into a 40-60% solution

79 1

Application to Extractive Metallurgy

of Versatic acid in kerosene. Ammonia solution is used t o adjust the equilibrium pH value to 6.9: (Ni,Co)S04 + nHZA2 + 2NH3

+

(Ni.Co)A*(HA)b-2

+ (NH$,S04

(1 1)

The ammonium sulfate contained in the raffinate is treated with lime to produce ammonia for recycle, and the extracted nickel and cobalt are transferred to a chloride medium (for subsequent separation by amine extractants, as described in Section 63.3.2.4) by stripping with 25% hydrochloric acid solution: (Ni,Co)A2(HG)2,-2

+ 2HC1

+

(Ni,Co)C12

+ nH2A2

(12)

The high affinity shown by carboxylic acids for copper (11) compared with the remaining divalent metals of the first transition series appears to be due in part to the stabilization of the extracted complexes by the formation of the well-known dimeric structure (1) in which copper(1I) carboxylates exist in the solid state and in non-donor s0lvents.5~The axial ligands, L, consist of undissociated carboxylic acid moleculesS5or, in the absence of an excess amount of extractant, they may consist of water or other solvent molecules.56 Copper was successfully removed from nickel sulfate solutions on the base-metal plant at Matthey Rustenburg Refiners in South Africa by being extracted into Versatic 10 acid at a controlled pH value. The process is believed to have been discontinued only because improvements in the selective leaching of copper and nickel rendered it unnecessary. R

The use of carboxylic acids for the removal of iron(II1) from solutions of the rareearth metals has been reported,38 but has not been described in detail. The stoichiometries of the extracted complexes of iron(II1) have not been clearly established. The n-decanoic acid complex has bsen variously described as (FeA3)3and Fe3A9-x(OH),(HA)x5 or [Fe(OH)A& and [Fe(OH)2AHA]2,57 the n-octanoic acid complex as (FeA3.H20)3,58the naphthenic acid complex as FeA3,47and that of Versatic 10 acid as ~eA3(HA),], or [Fe(OH)A2]3.59 Versatic acid has also been used by Thorium Ltd. in England for the production of pure europium(III), lanthanum(II1) and yttrium(II1) oxides.39The extracted carboxylate complexes of the lanthanides(IT1) have k e n showdo to have the composition MA3(HA)3.xH20,where x = 1 or 2. Extraction of the lanthanides(II1) by Versatic acid has been reported to increase progressively with atomic number39 whereas, for C9 carboxylic acids of similar structure, the extraction was shown to increase from lanthanum to samarium, and then to decrease from samarium to lutetium.6o A similar pattern, in which extraction increases from lanthanum to europium and decreases from gadolinium to ytterbium, was found for acids of the naphthenic type61 (alkylcyclopentanecarboxylic acids derived from the refining of petroleum). It is of interest that a plot of the stability constants (K1) for the formation of the acetate complexes of the lanthanides also reveals a maximum at samarium.62This effect can probably be attributed to opposing contributions from the enthalpy and entropy terms to the free-energy change on complex formation. Versatic acid has been used in Japan to recover indium and gallium from solutions obtained from Extraction is carried out at the leaching of bauxites, zinc minerals, coal ash and flue dusts.40~41.63 a pH value of 2.5 to 4.0;some coextraction of tin(II), iron(II1) and aluminum(II1) occurs if these metals are present. In the extraction of indium(II1) by n-hexanoic acid,64the predominant species in the organic phase was found to be InA3(HA)3whereas in the extraction by n-decanoic acid65 the existence of trimeric (InA3*HA)3and hexameric [InA2(OH)j6species was also postulated. The literature pertaining to the structures of extracted metal carboxylate complexes has been reviewed.66

792

Application to Extractive Metallurgy

63.3.2.2 Solvent extraction of base metals by ovgiinophosplrorus acids The use of organophosphorus acids, such as di(2-ethylhexy1)phosphoric acid (D2EHPA; di(2-ethylhexyl) monohydrogen phosphate; 2; R = C4H9CH(Et)CH2),is now well established in the recovery of base metals. This reagent has found commercial application in the separation of cobalt from n i ~ k e I , ~the ~ , ~separation * of zinc from impurities such as copper and cadmium,69the recovery of beryllium70and vanadium,71and in separations involving yttrium and the rare-earth

RO'

'OH (2)

It is a well-known and important feature of the chemistry of organophosphorus acids that they exist in the form of dimeric molecules when dissolved in organic solvents of low polarity, such as the aliphatic and aromatic hydrocarbons that are typically used as diluents in commercial extraction processes.74The formation of these dimers represents a self-solvation of the monomeric reagent via the formation of intermolecular hydrogen bonds, a stable eight-membered ring (3) being formed, which can conveniently be represented in the abbreviated form (4). Moreover, it is apparent that, when a sufficient amount of excess extractant is available, these cyclic structures also persist in the extracted metal complexes, only one proton being displaced from each dimer, which then acts as a bidentate ligand.75For divalent metal ions that show a preference for tetrahedral coordination and trivalent metals that prefer octahedral symmetry, coordination-saturated complexes such as (3) and (6) are formed respectively. However, for those divalent metals that show a preference for octahedral coordination, it is clear that additional ligands (B) must be bound to the metal ion so that the preferred coordination number can be achieved. Complexes such as (7) are then formed. If a sufficient amount of excess organophosphorus acid is available, these additional ligands consist of undissociated molecules of the extractant (B = HA or HzA2in structure 7). In the absence of an excess amount of extractant, water or other solvent molecules occupy the remaining sites in the octahedral complexes.

/-\0 - E

RO

"\I

/*

H :-

/A\

H

B

/".\/H

'.. A

A

M

H

...A/ \,+A

/*

.--

1\,/

H

B

A..,

H (5)

(6)

(7)

When large amounts of metal ions are loaded into organic phases containing organophosphorus acids, the Iimiting ratio of metal to extractant of 1:2 is approached, and polymerized complexes such as (8) are then formed, in which the PO2- group acts as a bridging ligand between adjacent metal i0ns.~6Solid products of composition such as CoA2 and NiA2(H20)2have been obtained by the removal of the solvent from such organic phases.77

(8)

The extraction of some metal ions by di(2-ethylhexyl)phosphoric acid is shown in Figure 9.' The order of extraction of the divalent metals of the first transition series, viz. Zn > Cr > M > Cu > Fe Co > Ni > V, shows important differences from that observed with carboxyli acid extractants, particularly with respect to the relative positions of zinc and copper and of coba

Application to Extractive Metallurgy

793

and nickel. The above order clearly shows no correlation with the Irving-Williams series for the stability constants of octahedral complexes, but appears instead to reflect the ease with which the respective metal ions adopt the tetrahedral configuration (or, for copper(I1) and chromium(II), a tetragonally distorted octahedral configuration) that is favoured by the bulky dimeric ligands. Therefore, an important contribution to the selectivity series is made by the magnitude of the differences in the ligand-field stabilization energies of the octahedral and tetrahedral complexes of the respective metals, which lie in the order Zn, Mn < Fe < Co < Cu, Cr < Ni, V. Evidence ,~~ as,from ~ ~ a study of the from spectral, thermogravirnetric, magnetic and ESR s t ~ d i e s , ~as~well effect of temperature upon metal distribution,81 suggests that the zinc(I1) and manganese(I1) complexes have tetrahedral structures, whereas the copper(1I) complex has a square-planar or distorted octahedral structure. The nickel complex has the octahedral structure (7), the axial ligands B consisting of undissociated extractant molecules {HA or H2A2) or water molecules, depending upon the extraction conditions. The complex has been variously formulated as Ni(HA2)2(HA)2,82~s3 as Ni(HA2)2(H2A2)284 and, more generally, as NifHA2)2(H2A2)x(H20)2-x.85 For cobalt(II), tetrahedral (blue) and octahedral (pink) forms of the complex exist in equilibrium,81*86 the tetrahedral form being increasingly favoured as the temperature is raised, owing to the entropy contribution of reactions such as

This effect has been utilized to enhance the separation of cobalt from nickel, since commercial extraction processes usually operate at temperatures of 50 to 70 "C. Thus, the difference between the pHo,5values for the extraction of cobalt and nickel from 1.0 M ammonium nitrate solutions by 0.50 M D2EHPA in xylene increases from 0.16 pH units at 0 "C to 0.43 pH units at 20 "C and 0.78 pH units at 50 "C.85At low metal loadings, the tetrahedral cobalt(I1) complex has been unequivocally formulated as C O ( H A ~ )whereas, ~ , ~ at ~ high ~ ~ metal ~ ~ ~loadings, ~ a polymeric species . ~ structure ~ is formed with a composition approaching the limiting stoichiometry of C O A ~The of the pink octahedral complex is undoubtedly analogous to that of the corresponding nickel(I1) complex. Im 1

0

I

I

I

I

I

2

3

4

5

Equilibrium p H

Figure9 Extraction of some metal ions (divalent unless shown otherwise) by 0.50 M solutions of di(2-ethy1hexyl)phosphoric acid in xylene at 20 "C

Recently new reagents were introduced that extract cobalt entirely in the tetrahedral f0rm,85~879*8 thus enhancing the separation of cobalt and nickel, since the complexes of the latter metal retain an octahedral structure. In these reagents of the phosphonic and phosphinic acid types (structures 9 and 10 respectively), the closer proximity of one or both of the alkyl groups to the metal-coordinating site brings about considerable destabilization of the octahedral nickel complex relative to the tetrahedral cobalt complex, owing to steric crowding, particularly if the alkyl group is branched at the a-carbon atom. So, for example, when R = 2-ethylhexyl, the separation between the curves for cobalt and nickel extraction for a 0.2 M solution of extractant in toluene at 20 "C increases from 0.53 pH units for the phosphoric acid to 1.42 for the phosphonic, and 1.93 pH units for the phosphinic acid.

194

Application to Exrractive Metallurgy

The Nippon Mining Co. refinery in Japan, whch had previously used D2EHPA to extract cobalt from solutions obtained in the leaching of a mixed cobalt-nickel sulfide in sulfuric acid, changed in 1978 to the new extractant 2-ethylhexyl 2-ethylhexylphosphonate enabling a much-improved selectivity for cobalt over nickel to be ~ b t a i n e d .The ~ ~ ,process ~ is carried out in three countercurrent stages with an organic solution of phosphonic acid that has been converted to the ammonium salt by contact with aqueous ammonia: __

H2A2

-

+ NH4OH +

2NH,HA2

CoZc

+ H2O Co(HAd2 + 2NH4' NH4HA2

(14) (15)

Any nickel that is coextracted with the cobalt is displaced from the organic phase by contact with a concentrated solution of cobalt(I1) sulfate: NI(HA,),(H,A,),,

+ CoZf

- +

Co(HA2)2

+ 4 . 4 2 + NiZ'

(16)

Cobalt is stripped from the organic phase with the spent electrolyte from the cobalt electrowinning circuit Co(HA2),

+ H2S04

2H2A2 + C0S04

--j

(17)

and the neutral solution of cobalt(I1) sulfate so formed provides the feed to electrowinning. The h g h affinity of organophosphorus acid extractants for zinc(I1) has been utilized in several commercial processes. The Nippon Mining Co. uses D2EHPA to extract zinc selectively from cobalt and nickel pnor to separation of the latter metals, using the phosphonic acid extractant described above. Zinc is extracted in a single contact at an equilibrium pH value of 2 to 3, where negligible coextraction of cobalt and nickel occurs (Figure 9). Zinc is also recovered with D2EHPA at plants in Spaid9 and Portugal9' after a preliminary extraction from chloride solutions with an amine salt (Section 63.3.2.4). The zinc(I1) chloride strip liquor obtained from the first extraction cycle is extracted with a solution of D2EHPA, milk of lime being used to adjust pH: ZnClz

___

+ 2H2A2 + Ca(OH)2

Zn(HA2)2

+ CaC12 + 2H2O

(18)

The extracted Linc is transferred to a sulfate medium, which is preferred for the electrowinning of i n c , by stripping with a solution containing sulfuric acid (being spent electrolyte from the electrowinning plant): Z U S O ~+ 2H2AZ ~

Zn(HA2)2 -I HlSO,

+

(19)

The concentrated solution of zinc(I1) sulfate produced in this way forms the feed to the zinc electrowinning stage, and the regenerated D2EHPA is recycled to extraction. Zinc(I1) is extracted selectively over copper(I1) and cadmium(II), but some iron(II1) is coextracted into the organic phase. (Iron(II1) is, in fact, more strongly extracted than zinc(I1) under equilibrium conditions, but the rate of extraction of the former metal is slow.) Any iron(II1) that is extracted remains in the organic phase during the stripping of the zinc and is removed subsequently in a separate step with a strong solution of hydrochloric acid. Organophosphorus acid extractants have also been used in the commercial recovery and separation of the rare-earth metals. The largest operation of this type is located at the Mountain Pass mine in California, which is owned by the Molybdenum Corporation of America and is the world's principal source of rare-earth compounds. The plant was inaugurated in 1965 to produce europium(II1) oxide of greater than 99.9% purity for use in the manufacture of red phosphors for colour t e l e ~ i s i o nSubsequently, .~~ facilities were installed for the production of sarnarium(II1) oxide for use in the manufacture of samarium-cobalt magnets for high-efficiency electric motors, and of gadoliniurn(II1) oxide for use in the image enhancement of X-ray films and in the control of nuclear reactors.93Other technical-grade rare-earth products are also produced. In the various solvent-extraction circuits employed in this process, use is made of a solution of D2EHPA in kerosene as the extractant. The selective recovery of the various metals is achieved by careful control of the equilibrium pH value of the aqueous phases in the multistage extraction and stripping operations. After the leach liquor has passed through two separate circuits, each of which comprises five stages of extraction and four of stripping, the europium product is obtained initially as a solution of europium(II1) chloride. Further purification of the product is accomplished by reduction with amalgamated zinc to EuN, which is by far the most stable of the divalent lanthanide ions with respect to the reduction of water (cf. the redox potentials of the Eu3+/Eu2+ and Sm3+/Sm2+couples, which are -0.43 and -1.55 V re~pectively~~). Sulfuric acid is added to the

Application to Extractive Metallurgy

795

europium(I1) solution to precipitate the insoluble sulfate, whch is subsequently calcined to europium(II1) oxide of h g h purity. Other commercial operations known to have used organophosphorus acids for the recovery of rare earths are those of Thorium Ltd. in E ~ ~ g l a n d , ~Megon ~ t h e Company in N o r ~ a y , ’ ~and ,~~ Denison Mines in Canada.96,97 The extracted complexes have been shown 98-100 to have the composition M(HA2)3.Although the lanthanides commonly form complexes in which the metal ion is eight- or nine-coordinate, octahedral coordination appears to be favoured by the complexes with organophosphorus acids. presumably because of the steric demands of the bulky dimeric ligands. Under conditions of h g h metal loading, the extracted complexes form insoluble gels, which are polymeric compounds of stoichometry (MA3),, where n is approximately 600 for lanthanum and ytterbium, and approximately 4000 for yttrium and holmium.101The extractability of the rare-earth metals by DZEHPA increases monotonically with increasing atomic a plot of log D (where D is the distribution ratio of the metal between the organic and aqueous phases) against atomic number ( Z ) being approximately linear, with an average separation factor (/3 = Di+l/DI) of about 2.5. Yttrium(II1) lies on the same straight line if assigned an artificial atomic number of 67.5. 1.e. between those of holmium and erbium. For the trivalent l a n t h d n i d e ~ ~and ~ . ’ actinides,99 ~~ as well as for yttrium and scandium.’j the equilibrium constant for the extraction reaction has been shown to vary inversely with the ionic radius of the metal ion. It has therefore been concluded that the extracted complexes are all of the M(HA2)3 type, involving predominantly ionic metal-ligand bonds.75 The similarity of the IR spectra of the scandium(II1) and thorium(1V) complexes of D2EHPA to those of the alkali metals is also indicative of the importance of ionic bonding.Io2 Scandium(I1lf is extraordinarily strongly extracted by D2EHPA, because its ionic radius is small (Pauling radius 0.81 A) compared to those of the ianthanides (0.93 to 1.15 A), and this property has been used in the recovery of scandium as a byproduct of uranium processing. Thus. at the Vitro Chemical Company in Salt Lake City, Utah, scandium was coextracted with uranium into an organic phase containing dodecyl dihydrogen phosphate. lo3 The uranium was selectivelj stripped into 10 M hydrochloric acid, while the scandium remained in the organic phase and was removed subsequently by stripping with hydrofluoric acid to produce a precipitate of scandmm(II1) fluoride. On a plant installed to recover scandium, thorium, yttrium and the heavy lanthanides from waste sulfate liquor from the Port Pirie uranium refinery in South Australia, the metals were coextracted into 0.1 M D2EHPA in kerosene.Io4All the products except scandium were stripped with 4.5 M sulfuric acid, and the scandium was recovered from the organic phase, whlch was contacted with a 2.5 M solution of sodium hydroxide to precipitate scandium(II1) hydroxide. The precipitate was then separated by filtration. Thermodynamic measurements on the partition of the lanthanides between 2-ethylhexyl phenylphosphonate in heptane and aqueous nitric acid have shown that changes in entropy are largely responsible for the increased stability of the extracted complexes of the metals from lanthanum to promethium, whereas the decreasing enthalpy of the system is the major factor contributing to the continued increase in the stability of the complexes of the remaining metals. lo5 Thus. for the reaction in the interval from lanthanum to promethium, the enthalpy of partition becomes decreasingly exothermic, whereas a very pronounced increase occurs in the entropy change (about 17 entropy units). In contrast, in the interval from promethium to lutetium, the partition process becomes increasingly exothermic, whereas the entropy changes remain approximately constant. Further scrutiny of the thermodynamic data for this system revealed that plots of the change in the free energy of reaction against atomic n u m h r show a distinct ‘tetrad’ effect,lo6as had been ~~~~~The observed previously in plots of logD against the atomic number of the l a n t h a n i d e ~ . ~ ~ effect is also clearly discernible in earlier distribution data for D2EHPA. lo9 Peppardloghas proposed the general hypothesis that, in systems involving all fifteen lanthanides(III), the points on a plot of the atomic number (Z)against the logarithm of a suitable numerical measure of a given property of these elements may be grouped into four tetrads by the use of four smooth curves without inflections, the gadolinium point being common to the second and third tetrads, and the extended smooth curves intersecting additionally in the regions where Z = 60 to 61 and 67 to 68. The origin of the effect has been attributed to the quantum mechanical interelectronic repulsion energy of the 4f-electrons.’ In practical terms, the tetrad effect makes the separation of those adjacent pairs of elements that lie on the zone of the shallowest curvature of each tetrad, v i t . Pr/Nd, EulGd, Dy/Ho and Yb/Lu, even more d i f f i ~ u l t . ~ ~ J l ~ Few details have been revealed concerning the commercial extraction of beryllium by organophosphorus acids. However, alternative processes involve the stripping of the loaded organic phase CCC6-Z*

796

Application to Extractive Metallurgy

with mineral acid followed by the precipitation of beryIlium(I1) hydroxide with alkali, or the direct precipitation of beryllium(I1) hydroxide from the loaded organic phase.70The selective extraction of beryllium(I1) over aluminum(II1) presents a major diffculty,ll2 although the separation of these metals can be achieved in a multistage process with careful control of the equilibrium pH values. The extracted beryllium(I1) complex is of the type Be(HA2)2,and, as would be expected from the small ionic radius of the cation (0.34 A), the steric complexity of the organophosphoric acid exerts a marked effect on its extractive power.' l 3 Although beryllium is extracted more strongly than the oxher alkaline earth metals, it is extracted less strongly than would be expected on the basis of the linear extrapolation of a plot of the logarithm of the extraction constants against the reciprocal of the ionic radii for barium, strontium, calcium and magnesium.75This is also a consequence of the very small ionic radius of beryllium(II), which is less than the limiting value that would result in oxygen-oxygen contact of the ligands, even for tetrahedral c ~ o r d i n a t i o n . ~ ~ Organophosphorus acids were among the first extractants to be used in the commercial recovery of uranium from solutions obtained by the Ieachmg of low-grade ores with sulfuric acid. In the so-called Dapex p r o c e ~ s ,l 4~the ~ , ~leach liquor is extracted with a solution of about 0.1 M D2EHPA in kerosene, and the pH value of the aqueous phase is adjusted to close to 1.O in order to prevent the coextraction of vanadium impurities. Since iron(II1) also extracts under these conditions, the leach liquor is reduced with metallic iron prior to extraction to convert any iron(II1) present to the iron(I1) state. The extraction of uranium(V1) by D2EHPA has been shown115to proceed according to the stoichiometry UO;'

+ 2H,A,

-

UO,(HA,),

+

2H'

(20)

Various structures have been proposed for the extracted complex, the most acceptable being (ll),which was proposed originally by B a e ~ . ~ ~

(11)

The dimeric cornpiex (U02)2A2(HA2)2, reported to be present at low acidities,lI7 can be formulated similarly (structure 12). At high concentrations of uranium, polymers of increasingly long chains and of composition (U02)nA2n-2(HA2)2are formed.

'*

(12)

In the Dapex process, uranium is usually stripped from the loaded organic phase with a solution of sodium carbonate, which converts the extractant to the sodium salt form, while retaining uranium in solution as sodium uranyl tricarbonate: UO,(HA2)2

+ 3Na2C03

2NaHA2

+ Na4U02(C03)3

(21)

Sufficient tri-n-butyl phosphate (TBP) or isodecanol is incorporated into the organic phase to prevent separation of the sodium salt of D2EHPA as a third phase. Uranium is precipitated from the strip solution by the addition of magnesium oxide or ammonia gas to give a product known as 'yellow cake'. On some p l a n t ~ , ~ lvanadium Jl~ is recovered from the barren aqueous phase emanating from the uranium-extraction circuit by the extraction of V02+into a solution of D2EHPA (usually a more concentrated solution than that used for the extraction of uranium) at an equilibrium pH value of about 2.2. The extracted complex has been shown to possess the stoichiometry VO(HA2)? when an excess amount of the extractant is present.s09ll9 ESR data are consistent with a fivecoordinate structure of C,, symmetry, and indicate that polymeric complexes, presumably of the type (V0),A2,-2(HA2>2, are formed at high metal loadings.80Vanadium is recovered commercially from loaded organic phases by stripping with sulfuric acid. The strip solution is heated with sodium chlorate, and ammonia is added to precipitate the vanadium as 'acid red cake' (probably HV03), which is then treated in a fusion furnace to give the final product, 'black oxide',

Application to Extractive Metallurgy

797

Organophosphorus acid extractants have found considerable use in recent years for the recovery of uranium as a byproduct in the manufacture of 'wet-process' phosphoric acid. This acid is obtained by the digestion of phosphate rock with sulfuric acid, and typically contains 0.1 to 0.2 g of uranium per litre.I2*It has been estimated that, in 1976, the wet-process acid produced in the USA alone contained some 2500 t of dissolved uranium;'21this therefore represents a valuable potential source of this strategic metal. Although solvent-extraction processes were operated at two phosphoric acid plants in Florida as long ago as the early 1950s, these were shut down a few years later when uranium became available at low cost from the sandstone deposits of the western USA. The process operated by the International Minerals and Chemical Corporation plant, for example, used a 5 7 % solution of diisodecyl pyrophosphate (13; R = i-CIoHzl)in kerosene to extract uranium(1V) from phosphoric acid solutions that had been reduced with metaltic iron.120,122 The loaded organic phase was treated with 25% sulfuric acid to remove coextracted calcium (as CaS04), and the uranium was then stripped with a mixture of 20-25% sulfuric acid and 15-20% hydrofluoric acid to precipitate UF4.xHF cgreen salt'). Although the product was not of high purity and serious losses of the extractant occurred due to hydrolysis, one of the new-generation plants in the USA is reported to be operating an essentially similar process based on the use of dioctyl p y r o p h o ~ p h a t e . ~ ~ Some recent research has been directed towards improvement of the basic process. 123

(13)

Conflicting information has been published on the relative extents of extraction of U4+ and U022+by dioctyl p y r o p h o ~ p h a t e . ~The ~ ~ reduction J~~ of the feed acid, which is done in the commercial process, may serve mainly to minimize the coextraction of iron(III), which is extracted much more strongly than i r ~ n ( I I ) . 'The ~ ~ uranium(IV) complexes extracted by dialkyl pyrophosphate have been reported to be of the type U(R2HP207)4,125 and those of uranium(V1) to be However, some doubt was subsequently cast on these conclusions of the type U02(R2P207).124 when it was found that the dioctyl pyrophosphate used, which had been prepared by the action of octanol upon phosphorus pentoxide, was a complex mixture of p r o d ~ c t s . ~ Indeed, ~ ~ J ~the ~ strong extractive power of this reagent towards uranium(1V) appears to be the result of a fortuitous synergistic interaction between the components of the mixture.'26 The Oak Ridge National Laboratory has developed two improved processes for the recovery of uranium from wet-process phosphoric acid and, since the dramatic increase in uranium prices in the mid-l970s, these are currently used wideIy on a commercial scale!' Both processes consist of two distinct cycies of extraction that are used to maximize the purification and concentration a mixture of the mono- and di-octylphenyl of the uranium present in the feed acid. In one process,12B esters of orthophosphoric acid (as a 0.3 M solution in kerosene) is used in the first cycle to extract U4+after the reduction of any U022+present in the feed with metallic iron. Uranium is recovered from the loaded organic phase by an oxidative stripping process using sodium chlorate in 12 M phosphoric acid. Uranium(VI), which is much less strongly extracted than U4+,returns to the aqueous phase and, by the use of suitably high 0rganic:aqueous ratios in the extraction and stripping circuits, can be concentrated to about 12 g 1-'. In the second process,129a mixture of D2EHPA (0.5 M) and tri-n-occtylphosphineoxide (TOPO, 0.125 M) in kerosene is used in the first cycle to extract UOzz+after oxidation of the feed solution with sodium chlorate, hydrogen peroxide, or air. The mixed extractant shows a specific synergistic enhancement of extraction for UOZ2+, and the loaded organic phase can be stripped by being contacted with a portion of the raffinate acid, the natural iron content of which has been partly reduced to Fe2+ with scrap iron. The equilibrium UO?

+

2Fe2'

+

4Hf

+

U&

+

2Fe3+ + 2H20

(22)

is displaced to the right by virtue of the strong complexation of Fe3+ by phosphoric acid. Uranium(1V) is less strongly extracted by the mixed extractant than uranium(V1) and returns to the aqueous phase. Both process alternatives use a second cycle of extraction of UOZ2+by 0.3 M D2EHPA plus 0.075 M TOPO in kerosene. Uranium is stripped from the loaded organic phase with a solution of ammonium carbonate (2 to 3 M), which converts the D2EHPA to the ammonium salt form

Application to Extractive Metallurgy

798

-

and precipitates uranium in the aqueous phase as ammonium uranyl tricarbonate: UO,(HA2)2(TOPO),

+ 3(NH&C03

2NH4(HA2)

+ (NH4),UO,(CO& + nTOPO

(23)

The precipitate is filtered and then calcined at 600 "C to yield high-grade U308. Alternatively, stripping with a more dilute solution of ammonium carbonate allows the uranium to remain in solution in the strip liquor and to be precipitated subsequently as U04.2H20by the addition of hydrogen peroxide. I3O The chemistry of the extraction of UO;' by mixtures of organophosphoric acids and neutral organophosphorus compounds has been the subject of many studies, and the available information has been widely discussed.75~1'R,127,131 134 There is no general agreement on the mechanism of the synergistic effect, although an 'addition process' such as UO,(HA2)2

+ TOP0

+ UOl(HA2)2.TOPO

(24)

appears to be favoured more than a 'substitution process' such as UO,(HA&

+ 2TOPO

UO,A,(TOPO),

----+

+ H2A2

In the addition mechanism, the coordination number of uranium is increased from six to seven, whereas in the substitution mechanism, there is no increase in the coordination number. The absence of any synergistic effect for uranium(1V) and t h ~ r i u m ( I V ) which , ~ ~ ~ give coordinationsaturated complexes M(HA2), with organophosphoric acids in which the metal ions are eight coordinate, is in accord with the addition mechanism. The magnitude of the synergistic effect for U022+increases with increasing basicity of the phosphoryl oxygen of the neutral additive,I**viz. (R0)3P0 < (RO)&PO < (RO)R2P0 < R3P0. Other aspects of the extraction of uranium(V1) by mixtures of organophosphorus acids and neutral organophosphorus compounds have been discussed by B a e and ~ ~by~ Blake et d 1 l S In contrast to the chemistry of extraction of uranium(V1) by organophosphorus acids, that of uranium(1V) has not been widely studied. It is known,k35however, that the extraction of U4+by mono(2-ethylhexy1)phosphoric acid (2-ethylhexyl dihydrogen phosphate), for example, exceeds that of UO,2+ by a factor of 105. This clearly provides the basis for the strong extraction of uranium(1V) from phosphoric acid solutions by extractants containing monoalkyl phosphates. The existence of a synergistic effect in the extraction of uranium(1V) by mixtures of mono(2ethylhexy1)phosphoric acid and neutral organophosphorus compounds was also reported recently.136 Organophosphorus acid extractants have also been used commercially to recover thorium from barren solutions obtained from uranium ion-exchange plants. For instance, f i o Tinto Dow Ltd of Canada installed a plant at Elliot Lake in 1959 to extract thorium(1V) from solutions containing only 0.15 g of thorium per litre.137The loaded organic phase is stripped with 5 M sulfuric acid, from which the product subsequently crystallizes as an acid thorium(1V) sulfate. It has been shown that the extraction of thorium(1V) by organophosphoric acids involves the mechanisms Th&

+ 3H,A,

-

and ThX3+ .t

3H,A,

ThA,(HA,),

-

t- 4H'

(26)

+

(27)

'ThX(HA2)?

3H+

where X- is the anion present in the aqueous phase. The extent of extraction of the anion-containing species depends upon the nature of the anion, the nature and concentration of the extractant, and the acidity of the aqueous phase. 138 A solid compound of the limiting stoichiometry ThA4 was obtained by the contacting of organic solutions of D2EHPA with an excess amount of an aqueous solution of thorium(1V) chloride, whereas, in corresponding experiments with thorium(1V) nitrate, a solid of composition Th(N03)A3 was obtained.139IR spectroscopy confirmed that the nitrate ion is covalently bonded to the metal. Among other metals extracted commercially by D2EHPA, indium and thallium may be mentioned. Indium is recovered as a byproduct of copper smelting at Cerro del Pasm in Peru,140while indium and thallium are both recovered by solvent extraction on zinc plants in the USSR.I4' The extraction of indium(1II) is reported to follow the stoichometry In3+ +

3H,A,

-

In(HA2)3

+

3H+

(28)

although oligomeric complexes such as [ I I A ( H A ~ ) are ~ ] ~formed at high metal 10adings.I~~ The extraction of thallium(II1) is reported to proceed according to the reaction143 T10H2+

+ %&

+ TlA,.HA

+

2H'

+

2H2O

(29)

Application to Extractive Metallurgy

79Y

63.3.2.3 Solvent extraction of copper and nickel by ortho-hydroxyoximes Since the mid-l960s, solvent extraction has been appiied on a commercial scale to the recovery of copper from aqueous solutions obtained from the leaching of acid-soluble copper ores with dilute sulfuric acid. At present, several plants producing up to 200 t of copper per day by an integrated leach-solvent extraction-electrowinning route are in operation throughout the world.37,144-146 All the plants in operation use extractants of the ortho-hydroxyoxime type, the active component being the antiisomer (14), in which R = alkyl, phenyl or hydrogen, and R = alkyl. These reagents (HA, where H denotes the replaceable, phenolic proton) extract copper by a pH-dependent cation-exchange mechanism: cu2+

-

+ 2HA

CuA2 +

2H+

(30)

This allows the efficient extraction of copper at low acidities (usually at a pH value between 1.5 and 2.0) and its subsequent back-extraction (stripping) into an aqueous phase of higher acidity with simultaneous regeneration of the free hydroxyoxime for recycle to the extraction stage. Aqueous feeds to the extraction process typically contain I to 3 g of copper per litre and 5 to 25 g of iron per litre, as well as other impurities such as aluminum, magnesium and smaller amounts of nickel, cobalt, manganese, zinc and molybdenum. Copper is selectively extracted into a solution of an ortho-hydroxyoxime in a hydrocarbon diluent, and the barren aqueous phase (raffinate) containing the iron and other impurities is recycled to the leaching operation so that the acid liberated in the sohent-extraction step can be utilized. The copper contained in the loaded organic phase is then stripped into an acidic solution comprising the spent electrolyte from the copperelectrowinning cells. Thus the acid generated during electrowinning is used to reverse the reaction shown in equation (30). In addition to the purification of the copper-containing leach liquor, considerable concentration of the original copper content is achieved during the solvent-extraction process by the use of a high 0rganic:aqueous ratio during stripping, thus producing a sufficiently eoncentrated aqueous solution of copper (about 50 g 1-l) for efficient electrowinning. N /OH

The extracti n of s~me common metal ions by the model compound anti-1-(2-hydroxy-5methylpheny1)octan-1-one oxime (14; R = n-C7HI5,R' = Me) is shown in Figure 10. The order of extraction of some common metals (pHo.5values in parentheses) is147Cu2+(1.0) > Fe3+(2.4) > Ni2+(4.0) > Coz+(6.1) > Zn2+(7.4) > Mn2+(8.8). 100-

.w * G 52x 75 2550 :

0

co

/j+//jj I

1

1

I

1

For the divalent metals, the order is largely as expected from the Irving-Williams series of stability constants. A similar extraction order has been reported for solutions of salicylaldoxime (14; R = R = H) in benzene, although the extraction of zinc(I1) was found to be anomalously

800

Application to Extractive Metallurgy

low.l4*More recently149,the order of extraction by solutions of the commercial reagent LlX 64N (14; R = Ph, R = C9H19) was given as Cu > FeITr> Ni > Zn > Co. The selectivity of these reagents for copper@) over iron(III), which is an important aspect of their successful commercial acceptance, is attributable to the fact that the former metal gives a specifically structured, essentially square-planar complex of the type (15). Is* H-0

(15)

Tn addition to the two six-membered rings formed by the chelation of the ortho-hydroxyoxime group to the central metal ion, this complex contains two five-membered rings formed by hydrogen bonding of the oximic hydrogens to the phenolic oxygens of the other ligand. This results in the unusual situation that the formation constant of the 2:l complex (K2) greatly exceeds that of the 1:l complex by virtue of the additional stabilization of the former by the formation of the fivemembered rings, which can occur only when both ligands are coordinated to the metal ion.'51 In contrast, the tris(salicylaldoximato)iron(III) complex is octahedral in structure, and no such extended planar ring formation is possible between the three oxime ligands. As a result, it appears that the stability of the iron(II1) complex is actually less than that of the copper(I1) complex, which enables the extraction of copper to be carried out at pH values lower than those required for the efficient extraction of iron. It should be noted that, since the extraction of iron(II1) by ortho-hydroxyoximes shows a third-order dependence on the concentration of the extractant152 Fe3+ +

3HA

-

m3+

3H+

(31)

in contrast to the second-order dependence for copper(I1) (equation 30), decreasing selectivities for copper over iron are to be expected as the concentration of the extractant is increased. However, I )selectivity , ~ ~ ~ Jfor~ copper ~ owing to the different rates of extraction of copper(I1) and ~ ~ O ~ ( I I the can be enhanced by the use of short phase-contact times. In the solid state, the metal atoms in bis(salicylaldoximato)copper(II) show two additional contacts with the oxime oxygens of adjacent molecules, resulting in a distorted octahedral structure. However, the axial Cu-0 distance (2.66 A) is much longer than the metal-ligand distances in the square-planar array (Cu-0, 1.92 8, and Cu-N, 1.94 A).154 Studies by ESR of copper(I1) extracts of the commercial reagent SME 529 (14; R = Me, R = C9H19)have shown that the copper complex exists as a square-planar species in hydrocarbon solutions, but that five-coordinate adducts are formed in the presence of ammonia or pyridine. 155 Spectrophotometric evidence suggests that the nickel(I1) and palladium(I1) complexes of salicylaldoxime are analogous to that of copper(I1) and that they contain a trans square-planar N202 donor set, whereas the complexes of cobalt(II), zinc(II), iron(I1) and manganese(I1) contain a cis square-planar N 2 0 2 donor set.150 The anomalous stability of the i r ~ n ( I I ) ' ~complex o in the series (logj& in 75% (vlv) dioxane at 25 "C,shown in par en these^),'^' Cu(21.5) > Fe(I6.7) > Ni(14.3) > Co(13.5), Zn(13.5) > Mn(11.9), cannot be explained in terms of spin pairing in the iron(I1) complex, which shows a paramagnetism equivalent to four unpaired e1ectr0ns.I~~ It has been suggested, on the basis of Mossbauer spectroscopic measurements, that metal-to-ligand backcoordination by K-bond formation occurs, the third ionization potential of iron being lower than that of any of the other metals studied.156 Bis(salicylaldoximato)nickel(II)is diamagnetic in the solid state, but becomes partially paramagnetic in solution in chloroform157and aqueous d i 0 ~ a n e . This l ~ ~ is likely to be due to molecular association or solvation,15*rather than to the adoption of a distorted tetrahedral structure in solution.151The nickel(I1) complex readily forms pseudo-octahedral bis-adducts in the presence of large amounts of amines, 159~160although analogous complexes of commercial ortho-hydroxyoximes have been reported to retain the square-planar structure in hydrocarbon solution in the presence of moderate amounts of ammonia or aliphatic amines.161 A relatively small number of commercial operations, such as those of the SEC Corporation at El Paso, Texas,16' and of the Nippon Mining Company at Hitachi, Japan,163use artho-hydroxyoximes for the recovery of nickel. In the SEC Corporation process, the aqueous feed solution consists of a crystallizer discharge stream from a local copper refinery, and contains about 70 g of copper

Application to Extractive Metallurgy

801

and 20 g of nickel per litre. After the solvent extraction of copper at a pH value of 1.6 to 2.2, the pH value of the solution is raised to between 8.5 and 10 with ammonia gas, and nickel is extracted into a 9% solution of L K 64N. A complete description of the process has been given. the aqueous feed consists of a solution of cobalt, In the Nippon Mining Company nickel and other metal sulfates obtained by the high-pressure oxidative leaching of a mixed metal sulfide ore. Following the remova1 of iron and copper by chemical precipitation methods, and the successive extraction of zinc and cobalt by organophosphorus acid reagents (Section 63.3.2.2) under weakly acidic conditions, the pH value of the solution is adjusted to between 9 and 9.5 with ammonia, and nickel is extracted into a 25% solution of LIX 64N in an alkane solvent. Although it is clear from Figure 3 that ortho-hydroxyoximes are able to extract nickel under mildly acidic conditions (a pW value of 4 to 5), the attainment of equilibrium is slow, at least 1 h of phase contacting being required at moderate stirring speeds. The extraction rate shows an inverse first-order dependence on the hydrogen ion concentration,lh5however, hence much shorter mixing times are adequate under the ammoniacal conditions used in commercial processes. The marked tendency for cobalt(I1) to become oxidized to cobalt(II1) after its extraction by hydroxyoximes is well known, and results from the large decrease in redox potentia1 of the Co2+/Co3+couple in the presence of nitrogen donor ligands

-

[CO(H~O)~]~+

[ C O ( H ~ O ) ~ ] ~+' e-

E" 1.84V

(32)

[CO(NH,),]~~ + e-

E" 0 1 V

(33)

-

and

[CO(NH~)~]~+

Stripping of the extracted cobalt(II1) complexes from the organic phase is very difficult, even with concentrated solutions of mineral acids, presumably because of the slow kinetics of ligand exchange of the d6 ion. The slow ligand exchange of cobalt(I1I) complexes can be used to advantage in the separation of cobalt and nickel, however. If an ammoniacal solution of nickel(I1) and cobalt(I1) is aerated sufficiently for the latter to be oxidized to cobalt(II1) before being contacted with an organic solution of an orlho-hydroxyoxime,166nickel can be extracted with high selectivity owing to its much faster reaction kinetics. Indeed, with aliphatic hydroxyoximes, no appreciable extraction of the species [Co(NH&l3+, [CO(NH,),(H~O)]~+ and [CO(NH~).+(H~O)~]~' was found during contact periods of several days.167 The commercial extraction of palladium(I1) by ortlna-hydroxyoximes is discussed in Section 63.3.2.5; here, it need be noted only that palladium(1Ig is extracted by these reagents at pH values considerably lower than those required for the extraction of copper(I1). The kinetics of the extraction of metals, especially copper(II), by ortho-hydroxyoximes (HA) has been widely studied in recent years.146Of particular interest is the observation that the rate of extraction can be considerably increased by the incorporation of small amounts of aliphatic a-hydroxyoximes (H2B) into the organic phase.168An even more marked effect has been noted Accelerator compounds of the former type are used in Henkel's commercial with a-dioximes. extractant LIX 64N, whereas those of the latter type are used in the Shell reagent SME 530. Although the accelerator compounds can form copper complexes at sufficiently h g h pH it is unlikely that the catalytic action operates by a simple exchange mechanism such as 169917*

Cuzt

+ 2H,B

Cu(HB),

+ C U ( H B ) ~ + 2H'

+T

A

-

~~

CuA,

+ 2H,B

(34) (35)

since the catalysis is stiIl operative at pH values below which reaction (34) takes place to any significant extent, particularly for the a-hydroxyoximes.l70 It is more likely that the catalysis involves the transient formation of mixed complexes of the type CuA(HB), of stability intermediate between those of the complexes CU(HB)~ and CuA2: Cu2'

+ HA

CuA+ -t

H,B

CuA+

+ HA

By this means, the reaction

-

+ CuA+

+

H'

(36)

CuA(H3) +. H'

(37)

m2-!- H'

(39)

802

Application to Extractive Metallurgy

which is the rate-determining step in the absence of H2B172 (probably owing to the need for specific orientation requirements at the organic-aqueous interface), is largely replaced by reaction (37), in which the relatively fast formation of a five-membered ring is involved, and reaction (38), in which the second six-membered ring and hydrogen-bonded bridged structure is formed under homogeneous conditions by ligand exchange in the organic phase. It is probably also relevant that the hydrogen-bonded bridged structure is absent in the mixed complex CuA(HB). The kinetics of the homogeneous reaction between copper(I1) and ortho-hydroxyoximes were also studied in detail re~ent1y.I~~ 63.3.2.4

Solvent extraction of base metab by amine salts

Amine salts represent the only commercially important class of extractants of the anionexchange type. Their most widespread use is in the extraction of uranium from sulfate leach liquors, but they have found application in the recovery of cobalt, zinc and copper from chloride solutions, as well as in the extraction of metals that readily form oxyanions, such as vanadium, molybdenum and tungsten. The order of preference of alkylammonium cations for simple inorganic anions in exchange reactions of the type R3NHX

+ Y-

+ R,NHY

+ X-

(40)

lies in the Clod- > NCS- > I- > NO3- > Br- > C1- > SO4*- > F-.Ths sequence is largely a reflection of the readiness with w h c h the anions Ieave the aqueous phase. Small, strongly polarizing anions such as F- are strongly hydrated, and it is much less energetically favourable for them to leave the aqueous phase than it is for the larger anions of more diffuse charge, such as C10, . According to the same principles, complex metallic anions such as FeCl,-, [UO,( SO4),]'and Co(NCS)2- should be readily extracted in the absence of competition from strongly extracted simple inorganic anions. The formation of the extractable ion-pair is not dependent solely on electrostatic interactions, but may also involve hydrogen bonds of the type R3N+H....C1-, for e ~ a r n p 1 e . Therefore I~~ the extraction of a given anion by different amines follows the primary > secondary > tertiary > quaternary, owing to the decreasing availability of interacting hydrogens bound to the nitrogen atom, as well as the increasing steric hindrance of the substituent alkyl groups to the approach of the extracted anion. In contrast, complex anions, such as halometallates, are incapable of forming hydrogen bonds with the alkylammonium cations, and therefore compete most effectively with the corresponding free halide ions in the order175quaternary > tertiary > secondary > primary, which is the observed order of extraction efficiency for most metals from chloride solutions. 179 The order of extractability of different metals from chloride solutions by a given amine is usually that expected from the tendency of each metal to form anionic chloro complexes.180This is evident in Figure 11, which shows the extraction of several metals from lithium chloride solutions by 0.50 M solutions of the tertiary amine Alamine 336 in xylene, The order of extraction,l8I Pd" > Cdrr > ZnIr > Fe"' > Cu" > Co" > Mnll > Fe" % Nil1,is in qualitative agreement with the order of the stability constants of the corresponding chlorometallate a n i o n ~ , ' 8although ~ , ~ ~ ~ it is difficult to predict the position of some of the metals that form rather weak complexes, such as cobalt, manganese and iron{II). It appears, therefore, that the extractability of a given metal is more a reflection of the stability of the anionic metal species Mz++ 4C1-

-

MC1:-

(41)

than of the relative affinity of the complex anion for the alkylammonium cation of the extractant, e.g. 2R,NHX

+ MC1:-

--+

(R,NH),MCl, -t 2X-

(42)

Spectral investigations of the organic extracts of amine salts from metal chloride solutions are consistent with the extraction of tetrahedral tetrachlorometallate anions for metals such as Fell1, CuIX,Zn", Cdrl, CoITand Mn1r.184-186 Analysis of the metal-saturated organic phases reveals the expected ratio of amine:metal:chloride of 2: 1:4 for the divalent metals and 1: 1:4for the trivalent metals. The extractability of the tetravalent actinides also follows the order of the tendency of those metals to form anionic chloro complexes, and the spectra of the organic extracts are those expected for the octahedral MCls2- species.187

Application to Extractive M e t d l w g y

803

The analysis of the distribution data is also in agreement with extraction stoichiometries such as 2R,NHC1

+

R3NHCI

MCl2-

+

+

MC12-

+ 2CI+ C1-

(43)

+ 2C1-

(45)

(R3"H)2MC14

MC14- + RJNHMCI,

and 2R3NHC1

-

(R3NH),MC1,

(44)

Deviations from such stoichiometries may indicate the formation of mixed associates of the '$8 such as [R3NH+MCl4-R3NH'C1-], and evidence has been given for the quadrupole ty~e,'2~7 formation of species of this stoichiometry for z ~ ~ c ( I Iand ) ~ for ~ ~copper(I1) J ~ ~ and cobalt(I1).lW1 It can be seen from Figure 11 that, by the appropriate choice of the equilibrium concentration of chloride ion in the aqueous phase, separations between certain pairs of metals can be made, for example between copper(I1) and manganese(I1) at a chloride concentration of 3.0 M, and between cobalt(I1) and nickel(1T) at a chloride concentration of 6 to 8 M. Furthermore, the metals can be stripped from the loaded organic phase by being contacted with an appropriate volume of water so that the equilibrium concentration of chloride ion in the strip liquor lies on the lower portion of the extraction curve, where substantial aquation of the extracted chlorometallate occurs (R3NH),MC14

+

6H20

+

M(H,0),2f

+

2CI-

+

2R3NHCI

(46)

and the metal cation returns to the aqueous phase. Thus, for example, copper(I1) can be stripped from 0.50 M solutions of Alamine 336 in xylene at equilibrium concentrations of chloride below about 2 M, and cobalt(I1) at below about 4 M.

I 0 , 0

/

Pd" z Pt" + Rh"I x IrT". The extraction of species such as H21rC16was suggested244without the extraction mechanism being considered, and the distribution data were recently discussed in terms of the generalized species [H'(H20),(TBP)b]n[MC1,n-], in which the extractant is assigned the role of solvating the hydrated proton.246It has been reported that the number of solvating TBP molecules depends upon the identity of the solvent used to dilute the extractant phase.245 The extraction of platinum is carried out commercially under strongly acidic conditions (5-6 M HCl). Although the distribution coefficient for the extraction of platinum into TBP is not very high (typically 5), reasonably efficent recovery of the metal can be obtained in a four-stage countercurrent process. The stripping of platinum is achieved by contacting the loaded organic phase with water, under which conditions the weakly basic extractant is readily deprotonated. Iridium(1V) can be separated from rhodium(XI1) by extraction into TBP, and this method is reported to be used at the Hanau refinery of D e g ~ s s a , ~ as~well ' as at one or more of the South African PGM refineries. Amine salts also find commercial application in the extraction of iridium(1V) in the presence of rhodium(II1) and base metals, stripping being achieved in this case by contacting the loaded organic phase with a solution of a reducing agent such as sodium hydrogen sulfite to reduce the metal complex to the weakly extracted IrC163- s p e ~ i e s . ~ ~ ~ , ~ ~ ~ Ruthenium and osmium are traditionally removed from solutions containing other PGM by distillation of their volatile tetroxides under oxidizing conditions. Osmium can be efficiently removed in this way even under quite strongly acidic conditions, and it would appear that little can be gained by the application of solvent-extraction technology to the recovery of tlm metal. For ruthenium, however, the oxidation process is more difficult, and the pH value of the solution must be maintained at a relatively high value, since the complete removal of Ru04 is difficult to attain. The extraction of the molecular species by carbon tetrachloride has therefore been suggested as an alternative,247as has the extraction of the ruthenium(I1) nitrosyl complex [Ru(NO)C1,I2- by amine salts,248the latter method having been used on a commercial scale in South Africa.249 Although rhodium has probably not been recovered commercially by solvent extraction, it can be extracted by quaternary ammonium salts under certain conditions,250the anion [Rh2C19]3having been identified in extracts from aqueous phases containing the species [Rh(H20)C1,J2-.

810

Application to Extractive Metallurgy

63.3.2.6 Solvent extraction of base metals by solvating extractants

The use of solvating extractants in the recovery of gold and platinum-group metals (PGM) was described in the previous section. These extractants have also found some specialized applications in the extractive metallurgy of base metals. For example, they have been used in the recovery of uranium, the separation of zirconium and hafnium, the separation of niobium and tantalum, the removaI of iron from solutions of cobalt and nickel chlorides, and in the separation of the rare-earth metals from one another. Although solvating extractants such as TBP are still widely employed in the purification of uranium for nuclear purposes and in the reprocessing of spent nuclear fuels (neither of which applications lies within the scope of t h s review), their use in the extractive metallurgy of uranium is largely only of hstorical significance. In the USA during the period 1942 to 1953, solutions obtained by the dissolution of rich pitchblende ores in nitric acid were extracted with diethyl ether, the process owing its origins to the early observations of P e l i g ~ tThe . ~ ~first ~ cycle of extraction was carried out on an aqueous feed solution containing 200 g of uranium per litre and 1 M free nitric acid. The loaded organic phase was stripped with water, and the strip liquor was evaporated to a uranium content of 1000 g 1-I. This removed excess nitric acid and destroyed any coextracted heteropoly acids of molybdenum and vanadium. The strip liquor was treated in a second extraction cycle at low acidity to give a uranyl product of h g h purity.252*253 Uranyl nitrate has been extracted into solvents such as diethyl ether as the entity [U02(N03)2(H20>4j solvated by two to six molecules of the organic s o l ~ e n t . ~The ~ " solvation ~~~ is usually considered to be of the secondary type, in which the solvent molecules are attached by hydrogen bonding to the water molecules in the primary hydration shell. However, IR data have been p r e ~ e n t e d ~ to ~ ~support J ~ 8 Muller's earlier hypothesis259that only two of the water molecules are directly bound to the uranyl cation, the remaining coordination sites being occupied by solvent molecules. So that the much stronger extraction of uranyl nitrate compared with the nitrates of the transition metals (e.g. the partition coefficients for Mn2+,Co2+,Cu2- and UOz2+lie in the ratio 0.1:1.O:2.O: lo7) can be rationalized, it has been suggested that the water and nitrate in the extracted species are The direct both covalently bonded to UOz2+by hybrid orbitals involving electrons of thef-~hell.'~~ contact of the metal and the nitrate has been confirmed by IR spectroscopy.258Solid compounds (where S represents the organic solvent of the types UO2(NO3)y3H20*Sand U02(N03)2.2H20.2S molecule), in which the metal ion retains a coordination number of eight, have been isolated from organic extracts of uranyl nitrate+260 The commercial extraction of uranium by diethyl ether was replaced in the early 1950s by a single-cycle process using TBP (20-30% in kerosene or hexane).252The feed solution was treated with phosphate to form a complex with thorium(IV), and with aluminum(II1) to form a complex with fluoride ions, and adjusted to 200-400 g of uranium per litre and 1-3 M free nitric acid. Uranium was stripped from the loaded organic phase into water at 65 "C to give a strip liquor containing 100 g of uranium per litre and less than 0.1 M free nitric acid. In the UKAEA process, uranium was precipitated as ammonium diuranate, calcined to U03 or U308, and reduced with hydrogen to UU2. In the USAEC process, uranyl nitrate was converted by 'thermal denitration' to U03,which was subsequently reduced with hydrogen to U02. It is reported that TBP still finds use in the primary extraction of uranium on plants at Palabora (South Africa),261 Forez (France)262and Mounana (Gabon).z62At the Palabora plant, a uranothorianite concentrate is leached in hot nitric acid to produce a liquor containing 25-35 g of uranium and 105-125 g of thorium per litre. Uranium, together with some thorium, is extracted into a 10% solution of TBP in kerosene. The coextracted thorium, which is rather less strongly extracted than uranium,263is displaced from the organic phase by being contacted with a concentrated solution of uranyl nitrate, after which the uranium is stripped from the organic phase by water at 40-50 "C. The strip liquor is subsequently treated with ammonia gas to precipitate an ammonium diuranate product containing less than 0.6% thorium. in which, It is generally agreed that the extracted uranium species is U02(N03)2(TBP)2127J64,265 in contrast to the species extracted into ether, the extractant molecules are directly coordinated to the metal ion. The extraction of uranium(V1) by neutral organophosphorus compounds inbonds) and with the increased creases with the number of C-P bonds (as opposed to C-0-P branching of the substituent alkyl groups, as would be expected from the effect of these changes on the electron-donor properties of the phosphoryl group. The extracted thorium(1V) complex has been formulated as Th(N03)4.2S266,267 and as Th(N03)4.3S,268,269 and its composition may depend to some extent on the conditions of extraction. The extraction of thorium (like that of

Application to Extractive Metallurgy

811

uranium(V1)) is also enhanced by the replacement of C-0-P bonds in the extractant with C-P linkages, but the extraction of thorium (unlike that of uranium) is depressed when branched alkyl groups are introduced into the extractant molecule. This result has been explained in terms of the increased importance of steric effects in the Th(N03)4.3S complex as compared with those in uo2(No3)2’2s-270 Processes have been developed271for the recovery of thorium by solvent extraction into solutions of TBP, but their present status is uncertain owing to the limited market for this metal (although some countries are reportedly stockpiling thorium for future use in nuclear applications). The solvent extraction of rare-earth nitrates into solutions of TBP has been used commercially for the production of high-purity oxides of yttrium, lanthanum, praseodymium and neodymium Crom various mineral concentrate^,^^ as well as for the recovery of mixed rare-earth oxides as a byproduct in the manufacture of phosphoric acid from apatite 0res.272?273 In both instances, extraction is carried out from concentrated nitrate solutions, and the loaded organic phases are stripped with water. The rare-earth metals are precipitated from the strip liquors in the form of hydroxides or oxalates, both for adjacent rare earths of which can be calcined to the oxides. Since the distribution coefficients (0) are closely similar, mixer-settler assemblies with 50 or more stages operated under conditions of total reflux are necessary to yield products of adequate purity.39 Initial fundamental s t ~ d i e s prompted ~ ~ ~ , ~the ~ ~suggestion that the extractability of the lanthanides by TBP increased regularly through the series, with a linear dependence of logD upon atomic number. However, subsequent work276,277 revealed that the linear dependence was not observed over the entire series, and that discontinuities in slope, or maxima, were always obtained in the region from europium to dysprosium. The changes in the entropy and enthalpy of the extraction reaction also vary in a discontinuous manner, while the changes in the free energy show that the extracted complexes evidence of a tetrad effect.27sFurthermore, the initial were hydrated species of the type [M(TBP),(H20), J(N03)3, where x is probably 6, was later rejected in favour of the formation of anhydrous trisolvates M(N03)3-3TBP,at as well as high concentrations280of metal in the organic phase. In view of the existence of water-saturated TBP as the monohydrate TBP-H20,the extraction of the lanthanides can be formulated as M3++ 3NO3-

+

3TBP.HZO + M(NO&.3TBP

+

3H2O

(86)

It has been confirmed that the water content of the organic phase varies inversely with the concentration of extracted metat, and falls to zero for metal-saturated solutions.281 The discontinuous variation in the extractability of the knthanides with atomic number has been interpreted as signifying a change in the coordination number of the metal ions in the extracted complex across the series.282Therefore, from a study of the electronic spectra of the extracted complexes of neodymium and erbium, it was concluded that the metal ions are eight- and sixcoordinate respectively. The difference in coordination was proposed to result from the existence of two bidentate nitrate ligands and one monodentate nitrate ligand in Nd(N03)3.3TBP, whereas all three nitrate ligands in E ~ I J V O ~ ) ~ - ~are T Bmonodentate.282 P NMR studies of the proton and 31Pchemical shifts for the lanthanide complexes of tri-n-pentyl phosphate also showed a difference between the structures of the complexes of the light and heavy lanthanides.283However, a more recent study of the lanthanide-induced chemical shifts of the cr-CRz protons in the TBP complexes provided no evidence for a change in coordination across the series, although the remoteness of these protons from the metal coordination centre may preclude their sensitivity to such changesh2W A review of the solvent extraction of the rare-earth metals has been pub1ished.l” Solvent extraction has proved to be the most effective method for the separation of zirconium and hafnium, which invariably occur in nature in close association, owing to their almost identical chemical properties. These metals have found considerable use in the nuclear-power industry on account of their unusually high (hafnium) and low (zirconium) neutron-capture cross-sections. It is evident that the mutual separation of the two metals must be of a high degree to make them suitable for such applications. Two different solvent-extraction processes are known to be used on a commercial scale; in one process, zirconium is selectively extracted from nitrate media into TBP; in the second process, hafnium is selectively extracted from thiocyanate solutions into methyl isobutyl ketone (MJBK). In the TBP process, wbich was developed in the USA28Sand in England,286zirconium(1V) hydroxide (produced, for example, by the hydrolysis of the material obtained from the caustic fusion of zircon sand) is dissolved in nitric acid to give a solution containing 30-100 g of zirconium (plus hafnium) per litre and 5-8 M free nitric acid. The zirconium is extracted into a 5 0 4 0 % solution of TBP in a suitable hydrocarbon diluent, the loaded organic phase is washed in 5 M nitric

Application to Extractive Metallurgy

8 I2

acid to remove coextracted hafnium, and the zirconium is stripped into water to produce a pure solution of the nitrate. The process has been described in detail in several publication^.^*^-^^^ Fundamental studies of the distribution of zirconium(1V) between dilute solutions of TBP in kerosene and aqueous nitric acid indicate that the disolvate Zr(N03)4*2TBPis present in the organic p h a ~ e , * ~although ' , ~ ~ ~ the presence of the monosolvate Zr(N03)4.TBP at low concentrations of extractant has also been s u g g e ~ t e d .Hafnium ~ ~ ~ , ~ has ~ ~ been shown to behave in an analogous manner.294The second-order dependence of metal distribution upon the concentration of hydrogen ion in the aqueous phase has been interpreted in terms of the extraction reaction M02+

+

2Hf

+

4NOJ-

+ %@

-

M(NO,)d.ZTBP

f

H,O

(87)

distribution data were adequately represented in terms of the

However, in a recent reactions (88) and (89). Zr4+

+

4N0,-

ZrOH3+

+

3NO3-

+ +

-

Zr(N03)4.2TBP

Zr(0H)(NO3),.2TBP

(88) (89)

Inextractable and other polymeric specieswere considered to be present in the aqueous phase.295 The selectivity of neutral organophosphorus compounds for the extraction of zirconium over hafnium in nitrate media appears to result from the combination of a number of factors. The stability constants of the nitrate complexes of zirconium are slightly higher than those of the corresponding hafnium complexes.296 Furthermore, the difference between the P=O stretching frequencies in the free extractant and the metal complex is greater for zirconium (60 crn-l) than for hafnium (45 cm-I), indicating the greater strength of the solvating interaction for the former rneta1.2g7 Differences for the two metals in the extent of formation of hydrolyzed298and polymerized species299in the aqueous phase are also undoubtedly important. The selective extraction of hafnium from thiocyanate media into diethyl ether was first reported in 1947.30°730tThe technique was subsequently investigated extensively in the USA with a view to the development of a suitable industrial-scale process, it being found advantageous for MIBK to be used as the solvent in place of diethyl ether.302The first commercial plant was completed in 1952,303and the process has since been used widely in the USA, England, France, Germany and Japan. Several descriptions of the process have been given in the literature.2R7-289~303-30s In a typical operation, a solution of zirconium and hafnium chlorides (1.3 M total metal with a hafnium content of 2-3% by mass) containing 1 M hydrochloric acid and 3 M ammonium thiocyanate is extracted in column contactors with a solution of thiocyanic acid (0.5 M) in MIBK, to produce a raffinate containing about 50 mg of hafnium per litre. The loaded organic phase is scrubbed with 3.5 M hydrochloric acid to remove coextracted zirconium (the resulting aqueous phase being recycled to the extraction circuit), and the hafnium is subsequently stripped from the organic phase with 2.5 M sulfuric acid to give a solution of hafnium(1V) sulfate containing about 2% zirconium. The reasons for the selective extraction of hafnium over zirconium from thiocyanate solutions by solvating extractants are not well understood. Hence, a recent review of the chemistry of these metals described the separation process but offered no explanation for the observed selectivity.306 There is no evidence that differences in the stabilities of the thiocyanate complexes of hafnium(1V) and zirconium(IV) are responsible for the selective extraction of the former, since the formation constants of the respective complexes are essentially identical for both metals.307However, there is some indication that the hafnium thiocyanates are more readily solvated by the extractant than are the corresponding complexes of zirconium. For tri-n-octylphosphine oxide, for example, the shift in the P=O stretching frequency from its value for the free extractant is greater for the hafnium complex (85 cm-l) than for the zirconium complex (70 ~ m - 9 . ~ ~ ' Several studies have been carried out in attempts to determine the stoichiometry of the extracted complexes, and it is generally agreed that hydroxo complexes of the type [Zr(OH),(NCS),_,] solvated by one or two molecules of extractant are the predominant species in the organic phase.308-3 14 Most frequently, the number of thiocyanate moieties in the extracted complex is equal to two, although some authors consider that the extracted species changes progressively to the trithiocyanato complex (x = 1) and tetrathiocyanato complex ( x = 0) on the successive extraction of a given aqueous phase.312IR data show that the thiocyanate groups are bound to the metal via the nitrogen atom,315as would be expected from the distinct A-character of zirconium(1V).

Application to Extractive Metallurgy

813

In a recent study in which MIBK was used as the extractant,316the distribution data were found to be consistent with the reaction ZrOZf

+

2NCS-

+ 2MIBK

-

ZrO(NCS)2.2MIBK

although, in the view of the doubtful integrity of the Zr02+ion, the extracted complex could equally well be formulated as Zr(OH)2(NCS)2*2MIBK.Furthermore, it has been concluded from IR studies of the complexes of hafnium(1V) extracted from thiocyanate solutions by cyclohexanone that no metal-oxygen double bonds are present.314 As was observed in the case of the extraction of zirconium and hafnium from nitrate media, it is probable that the different tendencies of the metals towards hydrolysis has some effect on the selectivity 0 b s e r v e d , 2 ~expecially * ~ ~ ~ ~ in view of the proved extraction of hydroxo complexes. The extraction of both metab decreases markedly in the presence of sulfate ions in the aqueous phase (a feature that is utilized in the stripping of the loaded hafnium with sulfuric acid), although the selectivityfor hafnium over zirconium is simultaneously increased on account of the higher stability constants of the inextractable sulfato complexes of The separation of niobium and tantalum by solvent extraction has been carried out on a commercial scale for about 30 years, and has totally superseded the previously used Marignac process based on the fractional crystallization of K2TaF7 and K2NbOF5.H20from dilute hydrofluoric acid.317,3 18 Early work showed that tantalum could be separated from niobium by being extracted from mixtures of hydrofluoric and mineral acids into diisopropyl ket0ne.~'9Niobium was also found to be extracted at higher acidities and high concentrations of hydrofluoric acid.320Subsequently, a solvent-extraction process using methyl isobutyl ketone (MIBK) was developed at the US Bureau of Mines Laboratories in Albany, Oregon, and was put into commercial operation by the Wah Chang Corp. in Albany in 1956.321In that process, columbite and tantalite ores are dissolved in 70% hydrofluoric acid, and the resulting solution, containing 10-15 M free hydrofluoric acid, is contacted with MIBK in the presence of sulfuric acid in polyethylene pulsed-plate columns. The coextracted niobium and tantalum are separatedin a second column by the stripping of the niobium into dilute acid, and the tantalum is subsequently stripped from the organic phase by being contacted with a large volume of water. Niobiumw) and tantalum(V) oxides are precipitated from the respective strip liquors with ammonia. An essentially similar process is used by Fansteel Metallurgical Corp. at Muskogee, Oklah0ma.3~There, however, contactors of the mixer-settler type are used in the solvent-extraction operation, and tantalum is recovered from the strip liquor by the addition of potassium fluoride solution to precipitate K2TaF7, which is used for the recovery of tantalum metal by fused-salt electrolysis. The use of TBP in place of ketone extractants has also been rep0rted;3~3this reagent offers the advantages of higher separation factors and lower losses of extractant to the aqueous phase.324*325 The preferential extraction of tantalum over niobium from weakly acidic fluoride solutions appears to result from the predominance of inextractable hydrolyzed niobium species such as H2NbOF5under these conditions; tantalum, in contrast, is present largely in the form of extractable HTaFs and H2TaF7.At high concentrations of hydrofluoric acid, the extraction of niobium is enhanced owing to the formation of significant amounts of HNbF6. The increased extraction of niobium in the presence of mineral acids can be related to the suppressionof the hydrolysis reaction

-

+ H,O H,NbOF, + HF (91) Distribution data suggest that when TBP is used the extracted niobium compound is associated HNbF,

with three molecules of extractant, whereas the tantalum compound is associated with two molecules of extractant except at very low metal loadings.326Nevertheless, it is likely that the role of the extractant is to solvate the proton component of the extracted fluorometallic acid rather than to solvate the metal ion directly. Commercial applications of solvating extractants in the extractive metallurgy of the morecommon base metals are very limited in number, the only well-known example being the use of TBP for the removal of iron from solutions of cobalt and nickel chloride^.^^' Nevertheless, it is probable that the use of solvating extractants (particularly neutral organophosphoruscompounds) will become more widespread with the increasing acceptance of chloride-based routes for the processing of certain base-metal ores, especially complex sulfides of zinc327,328 and copper.329,330 At the Falconbridge Nikkelverk in Norway, iron(LI1) is extracted from a solution of 4.5 M hydrochloric acid containing nickel (120 g 1-l) and iron, copper and cobalt (2 g 1-l each). The organic phase consists of a 4% solution of TBP in an aromatic solvent, and two stages of extraction suffice to reduce the iron content of the aqueous phase to 0.005 g 1-l. The loaded organic phase

814

Application to Extractive Metallurgy

is stripped with water in three stages to give a solution containing 33 g of iron per litre in 0.7 M hydrochloric acid. 191 It was concluded from a superficial study of the dependence of the distribution of iron(II1) on the concentration of TBP in chloroform that a complex of stoichiometry FeC13.3TBPwas extracted from 2 M hydrochloric acid, whereas a complex of stoichiometry H[FeCl4(TBP)4 was extracted from 6 M hydrochloric acid.331The coordination of iron(II1) was assumed to be octahedral in both instances. A similar conclusion was reached for solutions of TBP in benzene, the regions of extraction of the different complexes being given as 4 M and 7-9 M hydrochloric acid respectively.332However, it was subsequently shown that in all instances the electronic spectra of the extracts are those of the tetrahedral FeCL- species,333 and that the mechanism of extraction is therefore apparently analogous to that for the extraction of iron(II1) from chloride solutions by reagents such as ethers, ketones and esters where the extracted complex is of the type [H+.nS.mH20][FeCl4-].It has been shown that the e l e ~ t r o n i c ,IR,336,339 ~ ~ ~ , ~and ~ ~ proton magnetic resonance spectra340of such complexes are consistent with the solvation of the proton rather than the metal ion by the extractant, the metal ion remaining in the form of the tetrahedral FeC14- species with no primary solvation or hydration. Water also plays an important role in the formation of the extractable complex, anhydrous HFeC14-2S(where S = diisopropyl ether), for example, becoming soluble in excess ether only after the addition of 5 mol of water per mol of complex and the formation of the species [H30+.4H20mS][FeC14-].335

63.3.3 Ion-exchange Processes As early as the middle of the nineteenth century it was observed that certain naturally occurring substances could extract and concentrate ions from their e n ~ i r o n m e n t . However, ~ ~ ~ , ~ ~50 ~ years elapsed before ion exchange, as the process became known, was used on an industrial scale for water-~oftening,~~~ and another 50 years before ion exchange came of age as a metallurgical process with the advent of the atomic age and the worldwide demand for uranium that accompanied it. The term ion exchange, as used by extractive metallurgists, refers to processes in which a solution containing a mixture of metal ions is contacted with a resin that is insoluble in the metal ion solution. A modern ion-exchange resin consists of a chemically inert organic matrix to which labile functional groups are chemically bonded. These groups interact with and extract metal ions from aqueous solutions. The reaction is generally reversible so that, by alteration of the chemical environment, the metal ions can be subsequently reextracted back into the solution phase. The twin objectives in this operation are always the production of a more concentrated metal solution and the separation of the desired metal ion from other species that may be present in the original solution. The early development of ion exchange as a unit operation in hydrometallurgy was slow, mainly because of the lack of selectivity of the resins under operating conditions, and the limited capacities of the earlier commercial resins. Consequently, ion exchange found applications only in processes where the concentration of metal ions in solution was very low, and where the resin could be used to upgrade the solution prior to some final purification step. Recent developments, however, particularly the introduction of chelating resins, have considerably broadened the scope of resins in hydrometallurgy. Ion-exchange resins have developed from naturally occurring alurnin~silicates~'~+~~~ to synthetic silicates, and finally (as a result of the discovery in 1935344 that polar groups could be fixed to a polymeric matrix) to organic ion-exchange resins. The inert organic backbone in such resins usually consists of polystyrene crosslinked with divinylbenzene, while the functional group bonded to the backbone generally falls into one of three groups, viz. anionic, cationic or chelating. The chemistry involved in the interactions between these various functional groups and metal ions in solution is generally the same in ion exchange as in solvent extraction; therefore, for a range of metal ions, the order of selectivity and the relative distribution between the solution and the extractant are also similar. As a broad generalization it can be stated that ion exchange would probably only be preferred to solvent extraction in the hydrometallurgical processing of low-grade solutions. In these situations solvent losses in solvent extraction become relatively important, whereas the polymeric ion-exchange resins are effectively insoluble in water and do not suffer from this drawback. However, perhaps the greatest advantage of resins is their ability to treat metal solutions containing appreciable quantities of suspended solids, since this eliminates the necessity for the costly solution-clarification steps that are indispensable in solvent-extraction prwesses. Resins with cation-exchange functionality are available either as strongly or weakly acidic ion exchangers. The strong-acid resins have sulfonic acid (-S03H) functional groups, whereas the

Application to Extractive Metallurgy

815

weak-acid resins have carboxylic acid (-CO*H) functional groups. The mechanism of extraction is by the exchange of hydrated metal cations in solution with hydrogen ions on the functional groups, the pH value of the aqueous solution being the most important factor in the determination of metal extractability, particularly for the weak-acid resins. The metal ions remain in the outer Helmholtz layer in the hydrated state, and electrostatic interaction without dehydration accounts for the reversibility of absorption. Cation-exchange resins are not particularly selective, but some separation of the metals can be achieved on the basis of their ionic charge and ~ i z e . ~ ? ~ Two types of anion-exchange resins are available: strong- and weak-base resins. The resins with strong-base functionality have quaternary amine functional groups, whereas the active groups on weak-base resins are primary, secondary or tertiary amines, or a mixture of the three. The rnechanism of extraction is by the exchange of complexed metal anions in solution with simple inorganic anions, such as chloride, nitrate, sulfate, or bisulfate, which are associated with the amine in the resin phase. Anion-exchange resins have better selectivity than cation-exchange resins, owing to the ability of certain metals to form anionic complexes in solution, whereas other metals in the same solution do not.346-34s In anion and cation exchange, the metal ion is generally held in the resin phase by coulombic forces, although in some instances (the extraction of copper by weak-acid resins, for example) there is evidence of complex formation. Strong-acid and strong-base resins are completely ionized in aqueous solutions in the pH range 2-12, and exhibit maximum ion-exchange capacity under these conditions. Weak-acid and weak-base resins are ionized only at pH values higher than 5 (for -COIH) and lower than 9 (for -NR2). Their capacities tend to be lower than those of the strong-acid and strong-base resins, but they often exhibit better selectivities for metals. The chelating resins are generally acidic, and the metal cations are extracted by exchange with protons on the functional groups of the resin. The reactions are therefore of the cation-exchange type but, in contrast to simple cation exchange, the metal cation coordinates with the functional group of the resin, forming strong covalent bonds. The chelating resins are therefore more specific for certain metals than the cation-exchange resins, which allows greater selectivity. Some examples of the functional groups in chelating resins are the iminodiacetate, iminophosphonate, picolylamine and pyridylimidazole groups. The use of ion-exchange resins in extractive metallurgy will be discussed in terms of these three basic types of resins and the chemistry of metal complexation in aqueous solutions, with particular reference to current applications in the hydrometallurgical industry. While the electrostatic interactions in anion- and cation-exchange resins are not strictly within the domain of coordination chemistry, the extractability of metal ions is nonetheless significantly influenced by the extent of their reaction with their aqueous environment. For example, the capacity and selectivity of cation-exchange resins for metal cations depends largely on the degree of coordination of water molecules to the metal ion, whereas the metal-extraction properties of anion-exchange resins depend on the extent of coordination between metal cations and anionic ligands in solution. 63.3.3.1 Cation-exchange resins (i) Principles

Cation-exchange resins extract metal ions from aqueous solutions by forming the metal salt of the acidic functional groups on the resin. This type of reaction can be represented by the following simple equation M(H20),*

+

n (-RH

( 1 -R),M(H,O),

+

nHf

(92)

where R represents the organic acid anion and the symbol 1- represents the inert polymeric matrix. The equilibrium in equation (92) is usually defined as a mass-distribution ratio such as D =

metal ion (mol)/dry resin (g) metal ion (mol)/solution (ml)

(93)

and the equilibrium value is determined by the pK, of the functional groups (Le. the p H value at which 50% of the groups are ionized) and by the strength of the interaction between the metal cation and the organic acid. The ion-exchange behaviour of cation-exchange resin is determined predominantly by the functional groups; the number of fixed ionic groups determines the resin capacity, while the acid strength of the groups determines the effective operating pH range of the resin. Elution of the metals from the resin is achieved by shifting of the equilibrium jn equation (92) to the left through mass action, i.e. by an increase in the concentration of hydrogen ions or of some other competing cationic species. Weak-acid resins can be eluted with essentially a stoi-

816

Application to Extractive Metallurgy

chiometric amount of acid, whereas strong-acid resins require a substantial excess of acid or of some other cation. The predominant influence that determines the selectivity of a cation-exchange resin is the strength of the ion-pair that is formed between the hydrated metal cation and the ionized acidic functional group. As the bond formed is essentially electrostatic in nature, the selectivity is determined by the ionic charge on the metal cation and the distance of closest approach to the functional group, i.e. by the size of the hydrated metal ion. Since the degree of hydration of metal cations decreases with increasing atomic number, the selectivity, as a general rule, increases with increasing valence and atomic number. This is illustrated by the results in Table 3, whlch show the affinities of the Duolite strong-acid resin for various monovalent and divalent cations (at concentrations in the range 0.01 to 0.1 M) relative to the hydrogen ion. The striking feature of these results is the poor selectivity of the cation-exchange resins; in particular, separations within a group of metals (e.g. the first transition series: iron, zinc, copper, cobalt, nickel) are so small as to be of no prsctical value. The other interesting feature of these results is the role of the matrix in determining the selectivity of the resin. Divinylbenzene (DVB) is introduced as a crosslinking agent in polystyrene resins, an increase in crosslinking leading to an increase in the mechanical strength of the resin. However, an increase in DVB also ieads to a decrease in the extent to which the resin will swell in solution and, consequently, to a decrease in the average pore size. The resin therefore becomes more selective for smaller cations with increasing DVB, and this effect is also illustrated in Table 3. Table 3 Relative Afinitiw of Various Ions for the Strona-acid Resin Duolite C-20a 4% 8% 12% 16% DVB DVB DVB DVB Monovalent cations

H Li Na NH4 K Rb

cs

cu Ag

1.0 1.0 1.0 0.90 0.85 0.81 1.3 1.5 1.7 1.6 1.95 2.3 1.75 2.5 3.05 1.9 2.6 3.1 3.2 2.0 2.7 3.2 5.3 9.5 6.0 7.6 12.0

1.0 0.74 1.9 2.5 3.35 3.4 3.45 14.5 17.0

2.2 2.4 2.4 2.6 2.65 2.7

2.7 2.8 2.9 3.0 3.05 3.6 3.95 3.25 5.8 8.1 14.0 14.5 16.5

Divalent cations

Mn Mg Fe Zn

co cu

Cd Ni Ca Sr Hg Pb Ba a

2,35 2.5 2.55 2.7 2.8 2.9 2.8 2.95 2.85 3.0 3.4 3.9 3.85 4.95 5.1 7.2 5.4 7.5 6.15 8.7

2.5 2.6 2.7 2.8 2.9 3.1 3.3 3.1 4.6 6.25 9.7 10.1 11.6

From Warshawsky, ref. 348.

In considering the extraction of metal cations by strong-acid resins in the presence of competing hydrogen ions, some approximations can be made. Monovalent cations absorb below hydrogen ion concentrations of 0.5-1 M, divalent cations below 1-2 M, and trivalent cations below 2-3 M, while a tetravalent cation, such as the Th4+ion, is strongly extracted at all acidities. In those instances where the ionic charge of the metal ion is less than the oxidation number (e.g. the uranyl ion UOZ2+or the vanadyl ion V02+), the cation-exchange behaviour is determined by the ionic charge and not the valency of the metal. Weak-acid resins generally extract metal ions only at a pH value of 3 or more. At high pH values, where the carboxylic acid functional group is predominantly ionized, the metal-extraction behaviour of a weak-acid resin parallels that of a strong-acid resin. The weak-acid resins tend to be more selective than strong-acid resins, however, and practical separations of metals are possible. In certain interactions between metal ions and weak-acid resins it is difiicult to determine whether the interaction is purely electrostatic or whether chemical bonds are formed. In particuiar, there is strong evidence349that Ag+ and Cu2+form complexes with carboxylic acid groups. Copper, in

Application to Extractive Metallurgy

817

fact, has such an affinity for weak-acid resins that, in the pH range 4-5, it can be extracted quite selectively from a solution containing cobalt and nicke1. A secondary, more subtle, effect that can be utilized in the achievement of selectivity in cation exchange is the selective complexation of certain metal ions with anionic ligands. This reduces the net positive charge of those ions and decreases their extraction by the resin. In certain instances, where stable anionic complexes form, extraction is suppressed completely. This technique has been utilized in the separation of cobalt and nickel from iron, by masking of the iron as a neutral or anionic complex with citrate35oor tartrate.351 Similarly, a high chloride concentration would complex the cobalt and the iron as anionic complexes and allow nickel, which does not form anionic chloro complexes, to be extracted selectively by a cation-exchange resin.

(ii) Applicutions Most of the applications of cation exchange in the recovery of metals from aqueous solutions are several decades old. Early attempts to employ cation exchange for the recovery of metals from primary sources failed because of the lack of selectivity. For example, the extraction of uranium from sulfuric acid leach liquors failed owing to the extenive coextraction of iron, aluminum and manganese. Over the years a number of studies have been carried out on the recovery of cobalt and nickel from the tailings of acid-leach hydrometallurgical operations. Because of their value and relatively low concentration in these solutions (100-1000 mg I-'), they are ideal for recovery by ion exchange. The studies have shown that nickel and cobalt (also copper or zinc) present in solutions containing a large excess of sodium, potassium, calcium and magnesium can be selectively extracted with weak-acid resins. No large-scale commercial installations have been introduced, however, and the chelating resins would probably be preferred in this application nowadays. Cation-exchange resins have therefore found their widest application in the removal of heavy metals from industrial liquors, but the objective in these applications, because of the lack of selectivity, has generally been water purification rather than metal recovery. One exception has k e n the treatment of the rare-earth elements. The separation of the rare earths by cation-exchange c h r o m a t ~ g r a p h ywas ~ ~ developed ~ originally for analytical purposes, but the technique was extended to a commercial scale for the preparation of individual rare-earth metals in a very pure state. The method involves loading of all the rare-earth metals non-selectively from a chloride leach liquor onto a strong-acid resin, followed by selective elution with a 0.1% solution of citrate buffered at a value of pH 8. The citrate ion forms complexes most strongly with the heaviest metals, thus reducing their effective positive charge and enhancing their rate of elution relative to the lighter metals in the series. The rare-earth metals emerge from the resin column in the order Er > Ho > Dy > Tb > Gd > Eu, and individual rare earths with a purity greater than 99.9% 353 can be produced. Another important industrial application of cation-exchange resins is in the regeneration of electroplating baths. Chromium has many applications in metal-finishing and aluminum-anodizing operations and, since it is highly toxic on the one hand and relatively valuable on the other, several ion-exchange processes have been developed for the regeneration of spent electroplating baths. Impurities such as iron, aluminum and manganese or valuable metals such as copper or zinc355are removed from the solutions of chromic acid and sulfuric acid with a cation-exchange resin. The anionic chromic acid can be recycled to a plating bath or can be concentrated on an anion-exchange resin.3s6 Finally, one of the first continuous ion-exchange plants installed used a weak-acid resin to recover copper from rayon-fibre spinning solutions. In the Bemberg or copper(I1) ammonium process,357the spinning takes place in an acidic copper sulfate solution, and the fibre is then washed in ammonia solution. The wash water contains as much as 30% of the copper required for the spinning operation and its recovery is important in economic and environmental terms. The copper is extracted as the cationic amine complex by the weak-acid resin, and is then stripped from the resin with the acidic spinning solution. Zinc is recovered in a similar manner from vicrose rayon-spinning operations. 63.3.3.2 Anion-exchange resins

(i) Principles Strong- and weak-base resins exchange anions with their aqueous environment and therefore extract only metals that form anionic complexes in solution. Strong- and weak-base resins display a similar affinity for anionic species, which increases with the charge and the polarizability of the anion. Strong-base resins have quaternary amine functional groups that possess a permanent

Application to Extructive Metallurgy

818

positive charge and extract metal ions of general formula ML,"- (where L is an anionic ligand), according to the following equation: + ML,"- + n I -NR,Xe I -&R~),ML,"- -+ nx(94) The equilibrium in equation (94) is generally defined as a mass-distribution ratio such as that shown above for cation-exchange resins (equation 93), and the position of the equilibrium is determined by the relative concentration of the counter-ion MLxn- and the co-ion X-. The nature of the quaternary amine has little effect on the equilibrium properties of the resin, and the chemistry of metal complex formation in aqueous solution is the dominant factor. Weak-base resins have primary, secondary or tertiary amine functional groups and, in neutral or alkaline solution, the amines are in the free-base (un-ionized) form and have no ion-exchange properties. In acidic solution, the amine groups are protonated and extract anions according to the following equations: I-NR,

+

+ HX

I-NR2HX-

(95)

and n 1 - h ~ ~ +~ ML=~ -

( I --E;R~H),ML,~-

+

nx-

(96)

The pK, of the weak-base resins is determined by the basicity of the amine functional groups and by the nature of the anion in solution. Since the strength of the ion-pair increases with the charge and polarizabiliiy of the anion, the driving force for ion-pair formation permits proton uptake from far more alkaline solutions when strong ion-pairs are formed. For example, the strength of the ion-pair formed between the large polymeric amine cation of a weak-base resin and the large, very stable and highly polarizable anion Au(CN)~-yields a pK, of about 9-10,358 whereas the pK, of the same resin in chloride solution is some 2 log units lower. Anion-exchange resins are eluted either by reversal of the equilibria shown in equations (94) to (96) or, in some cases, by chemical destruction of the anionic complex MLXn-.The elution of strong-base resins requires a large excess of the co-ion, X-, whereas the elution of weak-base resins can be achieved most effectively by treatment of the resin with a stoichiometric amount of hydroxide ions, which restores the resin to the free-base form: II--~~R~H),ML + , ~OH-

n l - ~ ~ , + ML;-

(97)

An example of chemical destruction of an anionic species is the elution of aurocyanide from an anion-exchange resin with thiourea, a neutral ligand that reacts with gold cyanide in the presence of acid to form a cationic complex. The elution reaction can be described by the following equation: I -&R3Au(CN);

+

2CS(NH2),

+ 2H2S04

-

2HCN

+ 1 -&R3HSOi +

[Au{CS(NH32J2]+HS0i

(98)

C

0

s

4

1'2

[HCll ( W

Figure 12 Volume-distribution coefficients in hydrochloric acid for various elements on the strong-base anion-exchange resin Dowex 1 X-10 (after Krause and Moore, ref. 359)

Application to Extractive Metallurgy

819

The greatest potential use of anion-exchange resins in hydrometallurgy obviously lies in the treatment of sulfate, chloride and cyanide leach liquors in which the anionic ligands form strong metal compIexes. An example of the selectivity that can be achieved with a strong-base resin in chloride solution is shown in Figure 12. Tin, zinc, iron and molybdenum all form strong anionic chloro complexes, even in fairly dilute chloride solution, and copper and cobalt form anionic complexes of intermediate strength. Nickel does not form anionic chloro complexes, while trivalent chromium is kinetically inert to chloride attack. It can be seen that by careful control of the chloride concentration in the range 0-10 M excellent separations of these metals, and many others (Figure 13), can be achieved. Str. ad!

lo"

No ads.

-

A s II1)

L

Zr. Feiiri, Pb

Li-Fr. Be-Ra, S c - h c . La-Lu. Th. Ni. AI

I

I

!HCII (M)

Figure 13 The distribution of various ions between aqueous hydrochloric acid solution and a strong-base resin, Dowex 1 X-10, as a function of hydrochloric acid concentration (after Marcus and Kertes, ref. 132)

Far less selectivity can be acheved in cyanide solution because many metals, both precious and base, form anionic complexes in even very dilute cyanide solution. This is illustrated by the equilibrium absorption isotherms for various metal ions in a leach liquor from a gold cyanidation plant on a strong-base resin (Figure 14).358 Owing to the wide variation in the strength of the metal cyanide complexes that occur commonly in leach liquors from gold plants, some degree of selectivity can be achieved by adjustment of the pH value of the solution. The aurocyanide anion (logs2 = 47) is far more stable than the ferrocyanide (logss = 36), nickel cyanide (logs4 = 30), copper(1)cyanide (logs, = 28) and zinc cyanide (log/& = 19) complexes, and remains in solution as an anion at pH values at which the other anions decompose. This is illustrated by the curves shown in Figure 15;358the effect of pH on the selectivity of a strong-base resin for metal cyanide complexes is shown in Figure 16.358 It has also been shown3a that the selectivity of anion-exchange resins for monovalent anions over polyvalent anions is a function of the distance by which the fixed charges on the resin matrix are separated. The uptake of a divalent anion requires the presence of two closely spaced positive CCCS-AA

Application to Exrractive Metallurgy

820

Conditions: pH value 11.5

CN-

O.OO2M

Y E:

e

/ 0

0.1

0.2

0.3

0.4

.

.

0.5

0.6

Metal in solution ( m g l - ' )

Figure14 The equilibrium loading of various metal cyanide complexes from a gold plant pregnant solution onto a strong-base resin (AIOIDU) (after Fleming and Cromberge, ref. 358)

Conditions: Toial ~ ~ i i ~ e ~ i i r a t t o i i of each metal IW4M

cu

0

2

6

4

R

10

PH

Figure 15 'The effect of acidity on the stability of various metal-cyanide complexes in aqueous solution (after Fleming and Cromberge, ref. 358)

I0

9

8

7

6

5

4

3

2

PJI

Figure 16 The effect of pH on the selectivity of a strongbase resin (AlOlDU) for metal cyanide complexes (after Fleming and Cromberge, ref. 358)

charges, whereas no such constraints exist in the extraction of monovalent anions. Results presented recently35*have shown that the application of this theory to the extraction of metal anions from cyanide solution by weak-base resins allows for greatly improved selectivity for the monovalent gold and silver cyanide complexes. With increasing pH of the solution, the number of ionized functional groups in a weak-base resin decreases, and the average distance between adjacent ionized groups increases, improving the selectivity of the resin for monovalent anions. This is illustrated by the results in Table 4 in which the absorption properties of two commercial weak-base resins (A7, made by Duolite International and IRA 93, made by Rohm and Haas) are compared with those of a strong-base resin (AlOlDU, made by Duolite International), at two pH values. Most weak-base resins have a low and variable concentration of permanently charged functional groups in the resin matrix. These pseudo strong-base groups arise as a result of the crosslinking of adjacent amine groups, and have all the properties of the quaternary amine groups of strong-base resins. Russian workers361have observed that certain weak-base resins with a low strong-base content extract gold very selectively from cyanide solutions of high pH, but that the selectivity disappears if the strong-base content exceeds about 20% of the total resin capacity. The mechanism of rejection of the multivalent complex cyanide anions in this instance is presumably also based on the charge separation theory.

(ii) Applications ( a ) Uranium. By far the most important contribution of ion-exchange resins in hydrometallurgy has been in the extraction of uranium. By the late 1940s most of the world's uranium was extracted

82 1

Application to Extractive Metallurgy Table 4 The Distribution of Anionic Metal Cyanide Complexes Between Anion-exchange Resins and an Aqueous Cyanide Solution as a Function of Solution pHa Metal cyanide anion Gold Silver Cobalt

Copper Nickel Iron Zinc a

I R A 93

A7 D,,

D6

71 90 120 20 10 115 13370

1140 150 600 650 1460 1160 8130

D,,

3230 210 170 100 510 240 1080

AlOlDU

D6 3440 280 4870 1040 3920 4250 520

D,,

D6

18950 2210 21170 3880 24400 4260 46000

16750 190 15530 2050 10180 8520 3080

D , , and D,=distribution coefficients at p H = 11 and 6 respectively. Concentration of metal in solution = M (from Fleming and Cromberge, ref. 358).

from low-grade ores (0.05-0.5% uranium), which were leached with sulfuric acid. These solutions contained high concentrations of free sulfuric acid, iron, aluminum, vanadium, magnesium, calcium, titanium and silica, and uranium-recovery techniques such as precipitation with ammonia or extraction with cation-exchange resins were so unselective that the purity of the final uranium product was often not much better than the original ore. Only when it was observed for the first time in the early 1950s that the uranyl ion undergoes stepwise formation of anionic Complexes with sulfate or bisulfate u o p + ns0,2uo,(so4),2-*" (99)

uo:+

-

+

~ H S O ~w UO,(SO~),"~" + n ~ +

(n = 1, 2,3)

(100)

were anion-exchange resins used for the first time, allowing uranium to be concentrated and separated from the major impurities in sulfuric acid leach liquors. Fixed-bed anion exchange became the universally accepted means for the recovery of uranium in the 1950s, and it also represented the only large-scale hydrometallurgical application of ion exchange. In the 1960s and early 1970s a move was made towards solvent extraction with tertiary amines because this technique offered improved selectivity and because it could be operated as a sophisticated continuous process, whereas fixed-bed ion exchange was traditionally a batch operation. With the development in the 1970s of several elegant continuous ion-exchange processes, the pendulum swung back to ion exchange to a certain extent, and the modern trend is generally for very low-grade solutions (with a uranium content of less than 200 mg 1-') to be treated by ion exchange, and for higher-grade solutions to be treated by solvent extraction. On a number of plants in South Africa the so-called Bufflex process employs ion exchange and solvent extraction back-to-back. In this process a low-grade, unclarified, leach liquor is first concentrated by a strong-base resin, and the concentrated eluate from the ion-exchange process is then purified in a small solvent-extraction plant. As the demand for products of higher purity and the need for the processing of lower-grade ores increases, this type of process is likely to become more popular. There are several uranium ore-bodies in the world that cannot be leached economically with sulfuric acid because of the high limestone content of the ore. Such ore-bodies are generally leached with an alkaline solution of sodium carbonate and bicarbonate. Carbonate also forms anionic complexes with the uranyl ion the predominant species being uo2(co3)34-and may therefore also be treated with anion-exchange resins. The leaching reactions and the distribution of the various anionic uranyl species are very dependent on the pH value of the leach liquor and on the sulfate or carbonate concentration. Nominally, only the anionic di- and tri-sulfate or carbonate species will exchange with the functional groups of an anion-exchange resin, but the resin itself can facilitate the formation of complex anions in the resin phase because of the high concentration (approximately 0.5 M) of the co-ion on the functional group. Therefore, a complex equilibrium is established in which the resin is a participant; the following reactions describe these equilibria for sulfuric acid leach liquors: ~

( t -hR,),SOZ-

2( I -b&SO:-

+

2( I - - & R ~ ) ~ S O ~+-

+

U0,S04 F==+

U02(S04)2-

( I -&R,),UO,(S0,),2-

( 1 -&R3)4U02(S04)2-

uo2(so4)?-==== ( I -&R3),UO2(SO4),"

(101)

+

SO$-

(102)

+

2S06

(103)

Application to Extracrive Metallurgy

822

The affinity of a strong-base resin for various anions commonly present in uranium process solutions can be ranked as follows: U02(S04)34- > U02(S04)22- > NO3- > Cl- > HS04- > Fe(S04)?- > SO4-. In the 1950s, urany1 sulfate was stripped from the strong-base resins with a concentrated solution of chloride or nitrate, these being the two most efficient anions in the above sequence. Uraniurn was recovered from the strip solution by precipitation of the diuranate species with ammonia: 2H2[UOz(S04)3]

+

I4NH4OH

---+

(NH4)2UzO7

+ 6(NH&S04 +

1 IHzO

On the modern plants employing the Bufflex flowsheet, uranyl sulfate is stripped from the resins with the bisulfate anion. The eluate contains about 1 M sulfuric acid since this is the optimum concentration for the subsequent solvent-extraction process; at lower acid concentrations, the tertiary amine in solvent extraction is only partially ionized, which reduces its capacity for uranyl sulfate, whereas at higher acid concentrations the bisulfate anion begins to compete with the uranyl sulfate anion for tertiary amine functional groups. In the alkaline process, uranyl carbonate is eluted from the resin by concentrated (approximately 1 M) solutions of chloride, carbonate or bicarbonate, and is recovered from the eluate by treatment with an acid to destroy the carbonate complex, followed by precipitation of the diuranate, e.g. (NH4)4[U02(C03)3]

+ 6HC1

-

4NH4Cl

+ U02C12 + 3H20 +

2UO,CI, -t- 6NHdOH + (NHJIU20,

+ 4NH4C1 + 3H2O

3C02

(105)

(106)

Weak-base resins have been tested from time to time but have not found wide acceptance in the uranium industry. The main reason for this is that the major advantage of weak- over strong-base resins, viz. elution by neutralization, cannot be utilized in uranium processing since it is not possible for the weak-base resin to be converted to the free-base form without diuranate precipitating within the pores of the resin (unless a complexing agent such as carbonate is added to the eluate). In the presence of carbonate, uranium remains in solution as the uranyl carbonate anion, even in very alkaline solution, so is readily eluted from a weak-base resin in the free-base form. This eluate would then be treated as depicted in equations (105) and (106) for the recovery of uranium. Alternatively, weak-base resins can be eluted by ion-exchange mass action. (b) Gold. The conventional process for the recovery of gold involves oxidative leaching with oxygen in calcium cyanide solution to produce a very dilute (usually less than 10 mg 1-I) solution of gold as the anionic complex Au(CN)~-.The leach liquor is then filtered and treated with zinc powder to reduce the gold back to the metallic state. Although the zinc-precipitation process is simple and cheap, and the technology is well-developed, there are certain instances when an anion-exchange concentration and purification step may be profitable. For example, the chemical composition of the leach liquor occasionally does not lend itself to the quantitative or economic recovery of gold by zinc cementation (a very low gold concentration or very high impurity concentration, for example), but the major potential application of anion-exchange resins in gold processing is for the recovery of aurocyanide direct from the pulp in the resin-in-pulp (RIP) process. The filtration and clarification steps are a very costly component of the conventional gold process, especially with certain ores that are difficult to filter, and it is likely that the RIP process would offer significant cost advantages in most applications. RIP for the recovery of gold and silver from cyanide pulp was proposed in South Africa in 1960,362and pilot-plant tests were carried out at that time. More recent progress in the development of RIP for gold extraction has come from the Soviet Union,3633364 and a 1978 report365indicates that one of the largest gold mines in the Soviet Union is using anion-exchange resins to extract aurocyanide in an RIP plant. This is believed to have been the world's first apptication of RIP to gold extraction, although RIP is now used widely for gold extraction in the Soviet Union.365The whole question of the recovery of gold from solutions or pulps by resins recently came under close scrutiny in South A f r i ~ a ~but, ~ ~ although . ~ ~ the ~ -process ~ ~ ~looks promising, no gold mines outside the Soviet Union have implemented this technology. The close association of gold and uranium in many of South Africa's gold-bearing reefs, and the success of the continuous ion-exchange process for the recovery of uranium from these ores, has also led to research and pilot-plant studies on an integrated anion-exchange RIP process for the recovery of both gold and uranium.369 In the extraction process, the aurocyanide complex plus other metal cyanide complexes that are common in cyanide leach liquors (e.g. Fe(CN)64-, Fe(CN)63-, Zn(CN)42-, Ni(CN)42-, C U ( C N ) ~ ~ and Co(CN)& load onto the resin by simple ion exchange: nl-6R3X-

+

M(CN),"-

( I -AR~),M(cN),~-

+ nx-

( 107)

Application to Extractive Metallurgy

823

Because of their symmetry, high polarizability and multiple anionic charge, these complexes are generally very strongly bound to the resin, and eiution, particularly from strong-base resins, is difficult. Anions such as chloride, bisulfate, nitrate or cyanide can be used to reverse the equilibrium shown in equation (107) but, only when the activity of these anions is increased by the addition of a polar organic solvent such as acetone or acetonitrile to the eluate, are reasonable rates of elution achieved. This forms the basis of the organic type of elution procedures that were tried in the past.370Two anions that absorb very strongly onto anion-exchange resins that have to be very effective eluants for aurocyanide, even in aqueous solution, are the been thiocyanate and zinc cyanide anions. Because they absorb so strongly onto the resin, the resin must be regenerated before recycling to extraction. For the zinc cyanide complex, this is achieved effectively and cheaply by destruction of the complex with acid: ( 1 -hR3)2Zn(CN)42-

+

ZH2S04

+

(I-&R3)2S042-

+

ZnSO,

+

4HCN

(108)

The removal of thiocyanate from a strong-base resin can be achieved effectively by treatment with an iron(II1) salt.366In the presence of a slight excess of Fe3+ ions, thiocyanate forms the cationic complexes FeSCN2+and Fe(SCN)2+: 2 1 -&R,SCN-

+

Fe2(S04)3 + ( I - h 3 ) 2 S 0 , 2 -

+

2[FeSCN]S04

t 109)

Thiocyanate can be recovered and recycled from this solution by the precipitation of iron as iron(II1) Aurocyanide can also be eluted from a strong-base resin by chemical coaversion of the gold to a cationic complex with the thiourea ligand, as shown above (equation 98). This method of gold elution is favoured in the Soviet Union, but suffers from the drawback that elution of the other metd cyanide complexes is generally poor, and multi-elution procedures are necessary. The elution of aurocyanide and other anions from a weak-base resin is achieved by neutralization of the amine functional groups (equation 96). The method is cheap and very effective;3b6this is a major factor favouring the use of weak-base resins, particularly since recent research has shown that modern weak-base resins extract5urocyanide efficiently from cyanide leach liquors that have undergone minimal adjustment of pH.358,368 (c) Silver. Until relatively recently, a considerable amount of silver from secondary sources such as plating or photographic wastes was discharged in waste streams. In plating waste, silver is generally present as the anionic cyanide complex Ag(CN)?-, and in photographic waste jt is present as the anionic thiosulfate complex AgS20 Zn2+,and most chelating resins will extract cations in this order. Therefore, copper can be extracted selectively from a solution containing manganese, iron, cobalt, nickel and zinc with most chelating resins, or nickel can usually be extracted fairly selectively provided no copper is present. To find a ligand that reverses this general trend at any stage and to incorporate that ligand into a polymeric matrix is far more difficult, and is a task that has challenged and intrigued chemists for several decades. For example, over the years, a number of research worker^^'^*^'' have attempted to incorporate the well-known nickel-specific ligand dimethylglyoxime into a polymeric matrix, and although nickel-selective resins based on this ligand have been made, they have suffered from other problems, such as the hydrolytic stability of the ligand functional group or low capacity. The history of the early development of chelating resins was reviewed in 1957,378and a recent describes the important developments from 1957 to the present time; only those developments that are likely to make a contribution in large-scale hydrometallurgical processing will be discussed here, however. The first chelating resins that were found to be really suitable for application in the field of selective cation absorption were those based on the aminodiacetate functional g r o ~ p . The 3 ~ ~first to have an affinity for commercial resin based on this functional group, Dowex A l , was a range of metals which was similar to the order of dissociation constants of the metal complexes with ethylmediaminetetraacetic acid (EDTA), Le. Resin selectivity

KEDTA

Sr < Ca < Fe < Co < Zn < Ni < Pb < Cu 8.6 10.7 14.3 16.3 16.5 18.6 18.0 18.7

Moreover, the range of stability constants was some three orders of magnitude greater than that of an ordinary sulfonic acid. Since then, many research groups have made resins based on the

Application to Extractive Metallurgy

825

aminodiacetate functional group, some of which have interesting metallurgical properties. In general, the order of affinity between polymer and transition metal ions is similar for all the polymers based on t h s functional group, and improved selectivities can only be achieved at the expense of the large binding constants typical of aminodiacetate or diaminotetraacetate resins.379 Other types of resin, incorporating oxygen as well as nitrogen functional groups, have been developed in an attempt to reproduce the sort of metallurgical performance that has been achieved with analogous solvent-extraction reagents, such as the LIX reagents based on hydroxyoxime groups. For the most part, the performance of these resins has been disappointing and they have suffered from problems with capacity or selectivity, or both.379 One of the most obvious applications of chelating resins is in the selective extraction of copper from low-grade leach liquors. The hydrometallurgical processing of low-grade copper ores is confined almost exclusively to leaching with sulfuric acid and, although these solutions contain copper at levels from a few grams per litre down to 0.1 g l-l, the application of conventional prccesses (such as cementation or solvent extraction) has been restricted to solutions containing at least 1 g 1-'. Ion-exchange processes would open the way to the treatment of lower-grade ores and solutions, and will inevitably become more attractive as the richer copper ore-bodies become depleted. Another potential application for the future is in the treatment of barren solutions on uranium plants. These dilute sulfuric acid solutions contain copper, nickel and cobalt at concentrations ranging from 0.05 to 1 g 1-' and the extraction of cobalt and nickel from such solutions could become an attractive proposition. These copper and uranium leach liquors generally contain high Concentrations of Fe3+ion, and a major limitation of all the earlier chelating resins in these applications was their poor selectivity for copper, nickel or cobalt over Fe3+ions, and their low exchange capacity at the pH values typical of these leach liquors. However, several resins have been introduced recently that show good selectivity for copper over iron and absorb copper well at pH values around 2. These are the Dow picolylamine derivative resins, XFS 4195, XFS 4196 and XFS 43084.382Their good seiectivity for copper is attributed to the fact that copper forms a covalently bonded five-membered ring complex with the picolylamine group, whereas iron is loaded by an ion-exchange mechanism, forming a weak electrostatic bond between the iron(II1) sulfate anion Fe(S04)2-, and the protonated amine, analogous to that in a weak-base anion-exchange process.383The selectivity of these resins for copper over iron therefore deteriorates with increasing acidity. A similar selectivity for copper over Fe3+ions was observed for the recently described pyridylimidazole derivative resins, PI1 and There, too, iron is believed to be loaded onto the resin as the Fe(S04)2-anion, and it was s h o ~ n that ~ ~ a3six-fold improvement in the selectivity for copper over iron resulted when the aqueous-solution anion was changed from sulfate to nitrate, which does not form anionic complexes with Fe3+ions. Table 6 Metal Ion Absorption by XFS 43084 Resina, Species

cu Felll Ni

co Cd

zn A1

Mg

Ca Ferr a

Equilibrium conc. (g I-') Rem K (1 mol-')

pH

Solution

2.1 2.0 2.0 3.0 2.2 3.5 2.2 5.0 2.2 3.6 3.7 6.3 6.3 2.2 4.0

0.68 0.87 0.80 0.75 0.93 0.85 0.89 0.87 1.06 0.98 0.71 0.W 0.19 1.25 1.13

30 8.6 10.1 16 1.7 8.4 0.3 3.3 1 .o 7.1 0.3 0.33 0.38 0.14 2.1

500 18 29 64 3.2 22 0.6 7.2 1.5 14 0.7 0.6 3.2 0.18 3.1

Resin solution volume ratio = 1.2/100, initial aqueous concentrationsof metal ion 1 g lLi, except Ca, which was 0.2 g I-', K AM]/{[M"t][m]) where barb denote the resin phase and RH represenh the free functional-group concentration at equilibrium. From Grinstead, ref. 382.

At higher pH values, the picolylamine and pyridylimidazole resins also absorb other base metals in the first transition series, and useful separations can be achieved. This is illustrated by the results

Application to Extractive Metallurgy

826

in Table 6,382which refer to the Dow resin XFS 43084. In particular, nickel can be extracted selectively from cobalt solutions with both types of resin, and this property could be applied usefully in the cobalt electrowinning industry, for example, where it is important for the nickel concentration in the advance electrolyte to be kept very low. Most transition metals form their strongest complexes with ligands carrying oxygen or nitrogen as donor atoms. There is a small group of ions, however, including Pd2+,Pt2+,Au+, Ag+,Cu+ and Hg2+,which form their most stable complexes with donor atoms such as phosphorus, sulfur and arsenic. Several resins based on sulfur donor atoms have been prepared; these include m e r ~ a p t o t,h~i ~ g~l y c o l a t eand ~~~ d i t h i o c a r b ~ n a t efunctional ~~~ groups. The interest in polymers of this type has been mainly in pollution-abatement and analytical applications. Two resins that extract the noble metals (the PGM plus gold) very selectively388and that are based on the polyisothiouronium group have been shown379,389 to extract the metals by anion exchange with the weak-base amine groups and by coordination with the sulfur donor atoms. The covalent bond formed between precious metals (PGM, gold and silver) and functional groups with a sulfur donor atom are often so strong that elution without destruction of the resin is generally difficult. In such cases, because of the very high capacity of these resins and because of the value of the metal loaded, incineration of the resin is usually the most economically viable method for recovery of the precious metals.

63.3.3.4

Future trends

( i ) Solveni-impregnated resins The solvent impregnation of resins is a relatively new concept that has developed from a growing awareness of the need for ion-selective resins, and the slow development of such resins caused by fundamental difficulties in the attachment of the selective functional groups to polymers. The concept of a solvent-impregnated resin (SIR) is a compromise which seeks to eliminate the disadvantages inherent in solvent-extraction and ion-exchange processes, and consolidate their respective advantages. In an SIR, the organic ligand is physically absorbed into the pores of an inert polymer without undergoing chemical bonding to the polymer. It therefore combines the advantageous chemical properties of the solvent-extraction reagent, such as high mass-transfer rates, high capacity and good selectivity, with the physical advantages of ion-exchange resins, such as low losses of solvent and simpler organic-aqueous engagement and disengagement. Two approaches for the incorporation of organic reagents into inert polymers have been developed. The first approach, which is based on the physical absorption of the extractant into a high-surface macroporous polymer bead,390can be applied universally to many reagents and polymers.3Y1An SIR prepared in that way by the impregnation of hydroxyoxime into a neutral polymeric support has been shown392to retain the distinctive properties of the reagent with very good selectivity for copper over other divalent ions and iron(II1). Another series of SIRs, which incorporated phosphoric acid or phosphonic ester compounds, retained the distinctive features of the reagents and were able to extract trace quantities of zinc from cobalt sulfate solutions.393The second approach in the development of SIRs, which was pioneered by the Bayer AG Company, employs the incorpo;ation in situ of the extractant during the copolymerization process.394These SIRs are termed Levextrel resins and ligands that have been incorporated into this type of resin include phosphoric acids and esters for the removal of heavy metals from solution.395 An interesting property of resins impregnated with oximes or oxines is that the selectivity of, for example. copper over iron(III), approaches that of a pure solvent-extraction process only when ar, inert solvent is present in the p o r q of the resin.396Thus, in a b-hydroxyoxime SIR, the selectivity for copper over iron(II1) improved by a factor of 20 when the solvent perchloroethylene was introduced into the SIR, and by a factor of 700 in a similar resin impregnated with 8-hydroxyquinoline.396This is believed to be due to kinetic and thermodynamic restrictions in the extraction of iron(III), but not of copper, at an aqueous-organic b o ~ n d a r y . ~ ~ ~ ? ~ ~ ~ The only major impediment to the development of SIRs on an industrial hydrometallurgical scale is the problem of leakage of the reagent from the resin phase into the aqueous feed solution, a problem that is more severe in the first type of SIRs than in the Levextrel resins. If this problem can be solved, a bright future is predicted for this concept in resin technology. (ii) Ion exchange in non-aqueous media

Ton exchange from organic solvents and mixed organic aqueous solvents offers interesting possibilities for the extraction and separation of metals because of the different nature of the solvation processes in these systems. Only cation solvation is significant in dipolar, aprotic, organic

Application to Extractive Mtlrallurgy

827

solvents,398and the anions are relatively more reactive than in water, allowing for enhanced distribution coefficients with anion-exchange resins. As the ratio of organic to aqueous in a mixed solvent is increased, the concentration of water molecules around the cation is reduced, decreasing the size of the hydrated cation. One consequence of this is that metal complex anions will be formed at lower anion concentrations than in pure water. Moreover, because the forces that bind the hydration cloud depend on the charge density of the cation, selective destruction of the hydration cloud starts at lower organic : aqueous ratios for larger cations.399As a result, large cations will penetrate a cation exchanger more easily, and differences in the distribution coefficients of elements of different size will be enhanced. Selectivities of potentia! hydrometallurgical interest that were demonstrated recently400are the very good extractions of nickel and cobalt from acetonitrile, propylene carbonate, sulfolane or dimethylformamide by a cation exchanger; copper, iron(ll), iron(II1) and zinc, present in the same solution, are either weakly extracted or are not extracted at all. It is also possible for copper to be extracted selectively from Fe3+ ions with an anion exchanger in dimethyl sulfoxide, dimethylformamide or dimethylacetamide.400 63.3.4

Precipitation

The recovery or removal of metals from solutions derived from the leaching of minerals is an important step in any hydrometallurgical process. Precipitation by reduction to the metallic state in electrochemical cells will be discussed in Section 63.3.5; this section will cover the use of chemical reagents to control the precipitation process. Therefore, although the production of metallic powders by the reduction of metal ions with hydrogen or iess-noble metals (cementation) is electrochemical in nature, it will be discussed under this heading. 63.3.4.1

Oxides

The selective hydrolysis of metal ions to produce various forms of hydrated oxides is the most widely used form of precipitation. In particular, the removal of iron from hydrometallurgical process streams is a continuing problem. Iron enters the circuit as a constituent of a valuable mineral, such as chalcopyrite (CuFe2), or an impurity mineral, such as the ubiquitous pyrite or pyrrhotite. So far, effective removal of the iron has been achieved by the precipitation of iron(II1) as jarosite ( M FeI(S04)210H)6):01 goethite ( FeOOH)402or hematite ( Fe203).403 The fundamental chemistry associated with the thermodynamics of formation of the alkali jarosites (M = Na, K or NHJ has been studied,404and it has been found that substitution of the alkali ions by Ag+ or 112 Pb2+produces jarosites that are often formed during the processing of lead or zinc ores, which generally contain silver as a valuable byproduct. In addition, partial or complete substitution of iron(II1) by divalent metal ions can occur, resulting in the formation of ( O order H ) ~ .of an extensive range of substituted jarosites such as beaverite, P ~ C U F ~ ~ ( S O ~ ) ~The incorporation into the lead jarosite structure has been found to be Fe3+ B Cu2+ > Zn2+ > Co2+ > Ni2+ > Mn2*.This order appears to be related roughly to the ease of hydrolysis of the various cations, and this knowledge has been of considerable use in the rationalization of the loss of valuable metals such as copper, zinc, cobalt and nickel in the jarosites that are formed during the iron-removal steps. This is particularly severe for galliwn(U1) and indium(III), which can replace all the iron and thus be lost as byproducts. The partition coefficient for gallium is significantly higher than that for aluminum, presumably because gallium is more similar in size to iron(II1) than aluminum. A knowledge of the thermodynamics of the formation of the various hydroxy and sulfato complexes of iron(III), especially at elevated temperatures, has been invaluable in the establishment of the optimum conditions for the precipitation of goethite and hematite.405The results of calculations of the solubility of the various oxides and basic sulfates are often demonstrated by the use of predominance-area (or volume) diagrams, an example of which is given in Figure 17. The dotted lines represent equal concentrations of neighbouring complexes and hence define areas of predominance for each species. Solutions with compositions above the solid line are unstable with respect to the precipitation of goethite. The effect of complexing by sulfate on the equilibria can be calculated, and a third dimension, viz. sulfate concentration, can be incorporated in the diagram. Another process that has benefited from procedures such as that outlined above is the removal of cobalt from leach liquors derived from the extraction of nickel-bearing concentrates or mattes.406Traditionally, the removal of cobalt from such acidic sulfate solutions has been based on the afinity of the trivalent metal ions (such as those of iron and cobalt) for hydroxide, which CCCfi-AA*

828

Application to Extractive Metallurgy

0

1

2

3

5

4

6

7

8

9

1 0 1 1 I Z 1 3 1 4

PH

Figure 17 Solubility o f FeOOH at 25 "C (after McAndrew, Wang and Brown, ref. 405)

is well known to be greater than that of the divalent metal ions. Hence, by the appropriate choice of pH and the oxidizing agent, cobalt(I1) can be effectively removed as the cobaltic oxide trihydrate Co2O3-3H20.At present, most commercial operations use either chlorine or nickel(I11) compounds, the latter being prepared by the electrolytic oxidation of alkaline nickel(I1) hydroxide:

+ OH+ Co2* + 3H20 Ni(OH),

NiOOM

-

+ H20 + eC O ~ O H )+~ Ni(0H)z +

NiOOH

(111)

2H'

(1 12)

It is likely that developments in cobalt-selectivereagents, such as those outlined in Section 63.3.2.2, will result in this approach being replaced by solvent extraction in the future.

63.3.4.2

Sulfides

Most base-metal sulfides are very insoluble compounds and, in principle, processes based on the differences in their solubilities can be used in the selective precipitation of sulfides. For example, the strong affinity of copper ions for sulfide ion is used to good effect in the removal of traces of copper from leach liquors in the recovery of nicke1,406cobalt407and manganese.40sFor manganese, other impurity metal ions such as cobalt, nickel and zinc are also selectively precipitated by the use of sulfide ions. An interesting application409of the use of sulfide ion as a selective precipitant for metals is the removal of molybdenum in the commercial recovery of tungsten from low-grade scheelite (calcium tungstate) concentrates. In this process, dissolution of the tungsten as soluble tungstate and of the molybdenum as molybdate ions is achieved by a pressure-leaching procedure using sodium carbonate. The molybdate ions are then selectively converted to the thiomolybdate species by the addition of excess sulfide ions: MOO:-

f

4HS-

IVIOS~~f 40H-

(1 13)

Although thermodynamically possible, the slower rate of substitution of the tungsten(V1) species results in the formation of only a small amount of the corresponding thiotungstate ion, provided that the conditioning period with sulfide ion is controlled. Molybdenum trisulfide is then quantitatively precipitated by the decomposition of the thio salt by the slow addition of sulfuric acid to a pH value of between 2 and 3: MoS2-

+

2H-

-

MoS,

+

H2S

(114)

Under carefully controlled conditions, the filtrate contains less than 10 mg of molybdenum per litre, and the precipitate less than 1% of the tungsten. Osseo-Assare,41° with a knowledge of the stabilities of the various species involved, recently subjected this process to a thermodynamic analysis.

63.3.4.3 Metals The possibility that reducing gases such as sulfur dioxide, carbon monoxide or hydrogen can be used under pressure to effect the final precipitation of metals from their solutions is an obvious and, in some instances, an attractive alternative to electrowinning or ~ e m e n t a t i o n . ~Of ' ~ the commercial plants built to produce metals in powder form by gaseous reduction of the metals from

Application to Extractive Metallurgy

829

aqueous solutions, those based on the Sherritt Gordon ammoniacal pressure leach process32have been the most successful, and continue to be built and operated. These plants produce nickel powders and small quantities of special copper powders by the reduction with hydrogen of the metals from their ammine sulfate solutions. The overall reaction for the reduction of nickel from ammoniacal.solutions with hydrogen under pressure can be written as Ni(NH3),2f

+

H,

Ni

+

2NH4++ (n-2)NH3

(115)

In practice, x is maintained at about 2.0, and therefore the molar ratio of NH3 to Ni does not change appreciably during reduction. Various catalysts are used to initiate the reaction, and the nickel nuclei formed thus continue to catalyze the reaction. The rate of reduction is first order with respect to the concentration of nickel ion and independent of the hydrogen pressure.412The effect of a change in the ammonia:nickel ratio on the rate of precipitation at 213 "C is shown in Figure 18, together with a line showing the change in the fraction of the total nickel present as the diammine with the ratio. This line was calculated by the use of stability-constant data extrapolated to 213 "C. The similarity of the two lines suggests that the reduction of the diammine species is more rapid than that of other species present; this information is of obvious value in the control and optimization of the reduction process, since it permits the optimum nicke1:ammonia ratio under various conditions to be calculated from thermodynamic data. The results for cobalt4I3are similar, the maximum rate being at a coba1t:ammonia ratio of about 2.8: 1.

.. (.?Ah

Figure 18 Effect of ammonia to nickel ratio (RA)Ni on the apparent flst-order reaction rate constant k Also drawn i s the line ( n = 2) showing the fraction of total nickel initially present as Ni(NH,),'+ ions (after Needes and Burkin, ref, 412)

Cementation, the process by which a metal is reduced from solution by the dissolution of a less-noble metal, has been used for centuries as a means €or extraction of metals from solution, and is probably the oldest of the hydrometallurgical processes. It is also known by other terms such as metal displacement or contract reduction, and is widely used in the recovery of metals such as silver, gold, selenium, cadmium, copper and thallium from solution and the purification of solutions such as those used in the electrowinning of zinc. The electrochemical basis for these reactions has been well established414and, as in leaching reactions, comprises the anodic dissolution of the less-noble metal coupled to the cathodic reduction of the more-noble metal on the surface of the corroding metals. Therefore, in the well-known and commercially exploitedg cementation of copper from sulfate solution by metallic iron, the reactions are

- +

cu2++ Fe

2e-

Fez+

cu

(116)

2e-

(117)

The coordination chemistry of both metal ions involved plays a vital role in governing the thermodynamics and the kinetics of such reactions. This will be illustrated by reference to the chemistry involved in the precipitation of gold from cyanide leach liquors by metallic zinc powder -the most widely applied method for the recovery of gold from The overali reaction is 2Au(CN)2-

+

Zn

2Au

-+

Zn(CN):-"

+

(4-x)CN-

(ll8)

where xis dependent on the concentration of free cyanide and, under normal operating condtions, is between 3 and 4. As a result of the exceptional stability of the dicyanoaurate ion, variations in the solution conditions have little influence on the cathodic reduction reaction. On the other hand, hydrolysis of the zinc ion to produce insoluble hydrated oxides is possible under certain conditions of pH and cyanide concentration, as shown in Figure 19. For efficient cementation, which requires the removal of gold to concentration levels below 5 x lops M, it is vitally important for conditions to be maintained that do not allow oxides to form, since passivation of the zinc surface inhibits

830

Application to Extractive Metallurgy

and, in some instances, prevents further reaction. For this reason, the concentration of free cyanide is generally maintained above about 4 x M, and the solutions are deoxygenated before cementation. The latter operation is essential because the reduction of the dissolved oxygen by the reaction 0,+ 2H20

+ 4e-

(119)

40H-

---+

can result in (a) excessive consumption of the zinc, and (b) high concentrations of zinc and hydroxyl ions at the surface, both of which are conducive to the formation of oxides on the surface. Table 7 illustrates the importance of the concentration of free cyanide in the maintenance of efficient extraction even in the absence of oxygen, particularly at higher concentrations of gold.

-I;

Soh bilirs

-6 4

I

1

1

6

8

10

12

I4

16

PH

Figure 19 Domains of solubility and insolubility of Zn" in cyanide solutions at 25 "C Table 7 Effect of Cyanide Concentration on the Cementation of Gold

0.00 1 0.002

0.05 0.05 0.05 5.0

1.03

1.34 1.32 41.3 99.7 107

0.005 0.00 1 0.005

5.0 5 .o

0.010

Although gold forms particularly stable cyanide complexes,16 it is fortunate that the main impurities in the cyanide leach liquors, such as iron, copper, nickel, cobalt and zinc, also form strong complexes with the cyanide ion. Therefore, although the concentration of iron, in the form of ferrocyanide, can be some orders of magnitude greater than that of the gold, the iron does not interfere, owing to the stability of the ferrocyanide ion, which prevents reduction to the metal at the potentials at which the cementation of gold occurs. The same is true of the other major impurity ions. A knowledge of the thermodynamics of the formation of the cyano complexes is therefore of considerable assistance in the prediction and interpretation of the roles of the various ions in the cementation process. O n the other hand, the recovery of some metals (notably those of the platinum group) by cementation is particularly difficult. This applies particularly to their removal from the waste streams of the refineries. In a typical process for platinum-metal refining,416the metals are dissolved as their chloro complexes and, although reactions such as PtC12-

+

2Fe

F====+

Pt

+

2Ee2+(2C1-)

( 120)

are thermodynamically favourable, the kinetic inertness of these complexes results in unacceptably low rates of reduction of all the platinum-group metals (except palladium, which forms complexes that are considerably more labile) by active metals such as iron, aluminum or zinc. This characteristic of the coordination chemistry of these metals also accounts for the difficulty experienced in the devising of suitable methods for their production in the pure metallic form by electrowinning from aqueous solutions.

Application to Extractive Metallurgy

831

63.3.5 Electrochemical Processes The production of pure metals by electrowinning or electrorefining is often the preferred final step in an extractive metallurgical process. A large number of metals are recovered by electrochemical reduction, the most important of which are (in order of quantity produced) aluminum, copper, zinc, magnesium, nickel, manganese, cobalt, chromium, titanium, the precious metals (gold, silver and the platinum-group metals) and the alkali metals. With the exception of aluminum, magnesium, titanium and the alkalis, which are eIectrowon from fused-salt electrolytes, aqueous solutions are employed for the electrowinning of these metals. On the basis of their electrochemical behaviour in aqueous solutions, the metals can be divided roughly418 into three main classes, as shown by the portion of the periodic table of the elements reproduced in Table 8. (a) Normal metals show reversible behaviour; are deposited with low overpotentials at high current efficiencies to give deposits with well-developed, relatively large grain sizes that are not readily affected by addition agents; and have relatively high overpotentials for hydrogen evolution. (b) Inert metals exhibit irreversible behaviour: are deposited with high overpotentials at current efficiencies that are generally Iower than 90% (in some cases, lower than 50%) to give fine-grained, stressed deposits that are readily modified by the addition of smoothing agents; and have low overpotentials for hydrogen evolution. Table 8 Electrochemical Characttrisiics of the Elements

/Lil

No deposits

No deposits

B

C

N

AI

Si

P

S

Ge

As

Se

Pb

Bi

Po

K

Ca

Sc

Ti

V

Cr

Mn

Fe

Co

Ni

Cu

Zn

Ga

Rb

Sr

Y

Zr

Nb

Mo

Tc

Ru

Rh

Pd

Ag

Cd

In

W

Re

Os

Ir

Pt

Au

Hg

T1

Cs .. . . ..

Ba -

RE. Hf .

.

.

Ta .

Normal as amalgams Inert; no deposits

....

..

Inert

Normal

Mermediate

In general, the practical applications of this classification are apparent when one considers that the electrowinning or refining of the normal metals presents few problems. In contrast, the recovery of the inert metals by electrodeposition is more difficult, the degree of difficulty decreasing from left to right in any row of Table 8, e.g. the electrowinning of chromium419metal is significantly more difficult than that of which, in turn, is more difficult than that of copper. These trends are due to thermodynamic as well as kinetic factors, the nobility of the metals tending to increase from left to right, whereas the kinetic factors cannot be correlated as easily with the positions of the metals in the table. However, it has been suggested4*' that the rate of electrochemical reduction of simple aqua ions (as reflected in the exchange current density) can be correlated with the rate constant for the substitution of water in the inner coordination sphere of the ions, as shown by the data in Figure 20. The rate of loss of coordinated water therefore appears to be an important step in the overall reduction and electrocrystallization of metal aqua ions. This is one of the main reasons why the electrowinning of chromium is difficult, in contrast to the relative ease with which zinc is deposited422even from acidic solutions, despite the similarity of the thermodynamics of the two reactions. Hydrolysis of the metal ions to form non-conducting layers of hydrated oxide on the surface of cathodes is one of the most important characteristics of electrowinning of the inert metals, and is one of the most serious problems. The simultaneous evolution of hydrogen during the deposition of these metals results in an increase in pH close t o the cathode surface that can (if the solution is not adequately buffered) result in the incorporation of oxide into the deposit and even, in some instances, in the cessation of metal deposition. Partly because or ths, sulfate solutions are preferred to, for example, chloride electrolytes, because the buffering action of the sulfate-bisulfate equilibrium reduces the possibility of local high pH values. Therefore chromium of high ( > 99%) purity can be electrowon from sulfate solutions,419but oxide inclusions in deposits from chloride solutions limit the purity to less than 95%. Furthermore, for manganese the formation of relatively weak

832

Application to Extractive Metallurgy

, .. .... .....

ioJ io4

,

10'

..... ,..... ,

io6

IO'

.,.l

1000 "C)temperatures impose severe limitations on the range of techniques suitable for the study of the coordination chemistry involved. As a result, metallurgists have traditionally studied only the macroscopic properties of silicate slags, such as the phase equilibria, thermodynamic activities and transport properties, and only in recent years have attempts been made to adapt spectroscopic techniques, which have proved so useful in studies of solutions at ambient temperature, to high-temperature systems. The coordination chemistry of molten slags is therefore not understood in detail at this stage. Ceramicists and physicists have studied the structures of silicate glasses and have interpreted the results in terns of characteristics such as the size, polarizability, and coordination number of the ions, whereas metallurgists have employed a less quantitative description in terms of the acidity or basicity of the components constituting the slag.424 The many forms of crystalline silica are arrangements of the Si04 tetrahedron linked to four others by bridging (-0-) bonds. During melting, the crystalline network is partially destroyed and, as the temperature is increased further, the degree of polymerization (n) in the structure (Sin03n+1)2(n+1)decreases gradually. Therefore, the transformation from solid silica to a viscous liquid takes place in several stages in which no well-defined fusion point can be discerned, and the physical properties, such as viscosity and electrical conductivity, also show gradual changes. In slags consisting mainly of oxides, an acid can be regarded as an acceptor of oxide (02-) anions, whereas a base is a source of these ions. In the presence of an acid, amphoteric oxides behave as bases, whereas in the presence of a base, they behave as acids. Examples of these are: Acid

sio2+ lo2- F==+

Base

CaO

Amphoteric

AIz03

Ca2+

+

0 ' -

SiO,4-

+ 0"

e

A12042-

(121) ( 122)

023)

Application to Extractive Metallurgy A1,0,

+

2A13+

+

833

302-

(124)

As can be seen, most slags are silicates and, if a basic oxide is added to the hexagonal network of silica, it introduces 02- ions into the network, partially destroying it:

I

I

0

I

I

I 0 I

1 0 I

-O-Si-O--Si-O-

I

+ CaO

0

- ~ - ~ i - ~ -

+

Ca’+

I

+ -0-si-0-

1 0 I

I 0 I

1125)

A neutral slag is a slag that contains enough 02,‘ ions to ensure that each tetrahedron of the acid oxide is independent of the rest. In the binary system CaO-Si02, neutrality will be reached at the composition 2CaOSi02. Therefore, according to whether the mole percentage of Si02 is lower than, equal to, or higher than 33%, the slag will be basic, neutral or acidic respectively. The oxides can be classified424in terms of their acid-base character, as shown in Table 9. Table 9 Acid-Base Character of Common Oxides Oxide

Z / ( R c+ R J 2

Coordination number of liquid

Na20

0.18

CaO

0.15

MnO FeO

0.42 0.44 0.44 0.48

6-8 6-8 6- 8 6 6 6

0.72 0.75 0.83 0.93

4-6 4-6 4-6 4

1.22 1.66

4 4

ZnO

MgO c?03

Fe-0,

A1;O; Ti02 Si02

Character of oxide

Basic - network breakers

Amphoteric

Acidic -network formers

It is of interest that the order parallels that of increasing degree of covalent character of the oxide where R, and R, are the ionic radii. The as approximated by the coulombic term Z/(R, + Q2, assignment of the terms ‘network formers’ and ‘breakers’ is obvious from the discussion above. This classification of the oxides in terms of their coordinating properties for oxide anions has been particularly useful in the interpretation of the transport properties of the slags encountered in metallurgical processes, and has considerably simplified the design of suitable slag compositions. In particular, the addition of fluxes to reduce the viscosity of a slag is an important component in the optimization of a slag system. From the above, it should be clear that the addition oEa basic (or network-breaking) oxide to an acidic slag such as silica should result in decreased viscosity. In the same way, silica acts as a flux for basic slags. Examples of such effects are given in Figure 21. As could be predicted, the addition of alumina increases the viscosity of a basic slag but decreases that of an acidic dag. The correct choice of the material for the refractory lining of furnaces has also been simplified by the application of the concepts outlined above. Recent X-ray and Mossbauer studies425of silicate slags with Compositions similar to those used in the making of iron and steel have revealed that the simplified interpretation of the coordination chemistry of slag systems given above must needs be modified and extended when transition metal ions are involved. This is shown by the Mossbauer spectrum in Figure 22, from which it has been deduced that iron@) is tetrahedrally and octahedrally coordinated in a typical silicate slag. When this type of information can be determined at high temperatures and for the other important metals (such as chromium, manganese, nickel and cobalt) that are present in pyrometallurgical slags, it will be invaluable as an aid to the advancement of the present understanding of the mechanisms of redox reactions wm-ring during pyrometallurgical processes. There are several interesting parallels between the chemistry of aqueous solutions and that of high-temperature systems. This can best be illustrated by a discussion of the chemistry involved in the smelting of base-metal sulfide concentrates, which is the preferred route for the recovery of copper, nickel, cobalt and lead. The first step in such a process is the melting of the concentrate in the presence of additional slag-forming constituents to produce a two-phase system consisting of a matte (or molten sulfide) and a slag (or molten silicate). In principle, this can be likened to a high-temperature solvent-extraction process, the objective of which is the removal of as much

Application to Exlractive Metallurgy

834

I

I

I

I

I600

1500

Temperatrue ("C)

I I500

1

I

I I600

Temperature ("C)

Figure 21 Viscosity of (A) acidic and (B) basic blast furnace slags (after Goudurier, Hopkins and Wilkomirsky, ref. 424)

- 0.05 -2.60

I

-1.95

-

1.30

-0.65

I

0.00

I

0.65

I

1.30

I

I

I

1.95

2.60

3.25

3.90

Velocity (rnrn s - ' )

Figure22 MBssbauer spectrum of a slag containing 25% Si02, 31% CaO, 20% A1203 and 24% Fe203. The pair A corresponds to Fete:+, B to Fe2+ and C to Fe,,?+ (after Ebwker, Lupis and Flinn, ref. 425)

as possible of the impurity metals (such as iron and magnesium) from the sulfide by their extraction into the silicate. On the basis of the aqueous coordination chemistry of the metal ions involved, the following order of affinity for acceptor ligands such as sulfide ions would be predicted, CUI' > Ni" > CO" > Fe" > FeI", and roughly the reverse order would be expected for ligands such as oxide or hydroxide. This order is retained to a large extent in matte-slag systems, as shown in Table 10, which summarizes some data426on the distribution coefficient Diof species i between the sulfide and silicate phases under oxidizing conditions, calculated on a mole fraction basis. For copper concentrates, virtually all the iron can be partitioned into the slag by 'blowing', i.e. by the introduction of oxygen into the system whereas, for concentrates containing nickel and particularly cobalt, a compromise must be reached between the removal of the iron and the loss of the valuable metals into the slag phase. This is shown by the results426in the last two columns of Table 10, which were obtained from a detailed analysis of plant data. An extension of the correlation between the chemistry of aqueous solutions and that of high-temperature melts can be applied to several other important elements that often occur with the base metals. Thus, one would predict a favourable distribution of gold and the platinum-group metals into the sulfide phase and, indeed, the sulfide-rich matte that is formed is a very efficient collector for the precious

Application to Extractive Metallurgy

835

Table 10 Distribution of Metals in Smelting Processes Temperature (“C) Element

1255

1305

Recovery Dj

Matte (%)

Slug (%)

Cu Ni co Fe

245 274 80 6

-

-

cu Ni

180 257 61 7

94 87 41 0.6

Co

Fe

-

-

3 2 44

91

metals, which have an average distribution coefficient of about 170.427On the other hand, as zompared with iron, chromium, which occurs predominantly as chromite, would be expected to favour the slag phase; t h s is borne out by plant data, which suggest that chromium has a lower distribution coefficient than iron. 63.5

REFERENCES

1. R. Woods, in ‘Comprehensive Treatise of Electrochemistry’, ed. J. O M . Bockris, B. E. Conway, E. Yeager and R. E. White, Plenum, New York, 1981, vol. 2, p. 571. 2. J. Leja, ‘Surface Chemistry of Froth Flotation’, Plenum, New York, 1982, p. 498. 3. G. Barbery, J. L. C e d e and V. M. Plichon, in ‘Proceedings of the Twelfth International Mineral Processing Congress’, Institution of Mining and Metallurgy, London, 1977, p. 19. 4. D. R. Nagaraj and P. Somasundaran, Trans. AIME, 1980,226, 1891. 5. A. C. Vivian, Min. Mag.. 1927,36, 348. 6. B. W. Holman, Bull., Inst. Min. Metall., 1930, 315, 53. 7. G. Rinelli and A. M. Marabini, in ‘Proceedings of the Tenth International Mineral Processing Congress’, Institution of Mining and Metallurgy, London, 1973, p. 493. 8. D. R. Nagaraj and P. Somasundaran, in ‘Recent Developments in Separation Science’, CRC Press, Boca Raron, FL, 1929, vol. 5, p. 81. 9. M. R. Atademir, J. A. Kitchener and H. L. Shergold, Int. J . Miner. Process., 1981,8, 9. 10. J. L. Cecile, M. 1. Cruz,G. Barbery and J. J. Fripiat, J. Colioidlnterface Sci., 1981,80, 589. 1 1 . N. P. Finkelstein and S. A. Allison, in ‘A. M. Gaudin Memorial Volume’, ed. M. C. Fuerstenau, AIME, New York, 1976, vol. 1 , p. 414 12. F. Marsicano, P. J. Harris, G. M. McDougall and N. P. Finkelstein, Report 1785, National Institute for Metallurgy, Johannesburg, 1971, p. 11. 13. M. C. Fuerstenau and B. R. Palmer, in ‘Flotation’, ed. M. C. Fuerstenau, AIME, New York, 1976, p. 148. 14. C. T. hewitt and V. Rajamani, in ’Sulfide Mineralogy’, Min. Soc. Am., Washington, DC, 1974, vol. 1, p. PR-1. 15. G. Milazzo and S. Caroli, ‘Tables of Standard Electrode Potentials’, Wiley, New York, 1978, p. 54. 16. A. E. Martell and R. M. Smith, ‘Critical Stability Constants’, Plenum, New York, 1974. 17. V. Kudryk and H. Kellogg, J. Met., 1954, 6, 541. 18. M. J. Nicol, Gold Bull., 1980, 13, 105. 19. M. H. Ford-Smith, ’The Chemistry of Complex Cyanides’, HM Stationery Office, London, 1964, p. 64. 20. J. D. E. McIntyre and W . F. Peck,J. Electrochem. SOC.,1976,123, 1800. 21. T. Groenewald, J. S. Afr. Inst. Min. Metall., 1977,77,217. 22. N. P. Finkelstein, R. M. Hoare, G. S. James and D. D. Howat, J. S . Afr. Imt. Min. Met&/., 1966,66, 196. 23. H. A. White, J. Chern. Merail. Min. SOC.S. Afr., 1905, 109. 24. R. M. Berezowsky, V. B. Sefton and L. Gormely, US Pat. 4070 182 (1978). 25. N. J. Louw, A. M. Edwards and H. W. Gussman, Chemsa, 1977,3, 135. 26. A. J. Parker, B. W. Clare and R. P. Smith, Hydrometallurgy, 1979,4,233. 27. E. Peters, in ‘Chloride Hydrometallurgy’, Sociitt Benelux Metallurgie, Brussels, 1977, p. 1 . 28. M. Bonan, J. M. Dernarthe, H.Renon and F. Baratin, Merail. Trans., 3,1981, 12, 269. 29. M. J. Nicol, S. dJr. J . Chem., 1982,35, 77. 30. E. Peters, Merail. Tranr.,B, 1976, 7, 505. 31. D. M. Muir and A. J. Parker, in ‘Advances in Extractive Metallurgy 1977, Institution of Mining and Metallurgy, London, 1977, p. 191. 32. F. A. Forward and V. N. Mackiw, J. Met., 1955 7,457. 33. P. D. Scott and M. J. Nicol, Truns.-Inst. Min. Metall.. Sect. C, 1976,85,40. 34. A. N. Adamson, Chem. Eng. (London), 1970, CE 156. 35. M. J. Nicol, C. R. S. Needes and N. P. Finkelstein, in ‘Leaching and Reduction in Hydrometallurgy’, ed. A. R. Burkin, Institution of Mining and Metallurgy, London, 1975, p. 1. 36. Ref. 35, p. 16. 37. D. S. Flett, Chem. h d . (London), 1977, 706. 38. A. Gabriel, in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed. D. Dyrssen, J. 0. Liljenzin and J. Rydberg, North-Holland, Amsterdam, 1967, p. 654. 39. C. G. Brown and L. G.Sherrington, J. Chem. Techno[. 3iorechnol., 1979,29, 193. 40. H. Abe (Dowa Mining Co.),Jpn. Put. 74 43048 (1974) (Chem. Abstr., 1975,82, 143358b). 41. T. Katsura and H, Abe (Dowa Mining Co.), Jpn. Put. 74 42765 (1974) (Cham. Abstr., 1975,82, 143354~). 42. L. M. Gindin, P. 1. Bobikov, G. M. Patyukov, V. A. Dar’yalskii, K. P. Brodnitskii and I. A. Kasavin, Tsvern. Met., 1961,34,22 (Chem. Absrr., 1962,57, 3118b).

836

Application to Extractive Metallurgy

43. P. I. Bobikov and L. M. Gindin, Izv. Sib. Old. Akad. Nauk SSSR, 1962,6,46 (Chem. Abstr., 1962,57, 14699g). 44. N. b o , S. ltasako and I. Fukui, in ‘Proceedings of the International Conference on Cobalt: Metallurgy and Uses’, SocittP Benelux Metallurgie, Brussels, 1981, vol. 1, p. 63. 45. J. S. Preston, Hydrometallurgy, 1985, 14, 171. 46. A. W. Fletcher, D. S.Flett and J. C. Wilson, Bull., Ins!. Min. Metall., 1964,13, 765. 47. A. W. Fletcher and D. S. Flett, in ‘Solvent Extraction Chemistry of Metals’, ed. H. A. C. McKay, T. V. Healy, I. L. Jenkins and A. Naylor, Macmillan, London, 1965, p. 359. 48. C. S. G. Phillips and R. J. P. Williams, ‘Inorganic Chemistry’, Oxford University Press. Oxford, 1966, vol. 2, pp. 82, 116,268. 49. H. Irving and R. J. P. Williams, Nature (London), 1948,162, 746. 50. J. S. Preston, Hydrometallurgy, 1983, 11, 105. 51. M. Tanaka, N. Nakasuka and S. Goto, in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed.D. Dyrssen, J. 0. Liljenzin and J. Rydberg, North-Holland, Amsterdam, 1967, p. 154. 52. M. Tanaka, N. Nakasuka and S. Sasane, J. Inorg. Nucl. Chem., 1969,31,2591. 53. A. W.Fletcher and D. S.Flett, J. Appl. Chem., 1964, 14, 250. 54. D . P. Graddon, J. Inorg. Nucl. Chem., 1959, 11,337. 55. D. P. Graddon, Nature (London), 1960, 186,715. 56. D. P. Graddon, J. Inorg. Nucl. Chem., 1961, 17,222. 57. A. Bartecki and W. Apostoluk, Hydrometallurgy, 1980,5, 367. 58. P. Muhl, L. M. Gindin, A. I. Kholkin, K. Gloe and K . S. Luboshnikova, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 74)’, Society of Chemical Industry, London, 1974, vol. 2, p. 1717. 59. A. J. van der Zeeuw, Hydrometallurgy, 1979,421. 60. G. V.Korpusov, N. A. Danilov, Yu. S. Krylov, R. D. Korpusova, A. I. Drygin and V. Ya. Shvartsrnao, in ‘Proceedings of the International Solvent Extraction Conference {ISEC 74)’, Society of Chemical Industry, London, 1974, vol. 2, p. 1109. 61. R. A. Alekperov and S. A. Geibalova, Dokl. Aknd. Nauk SSSR, 1968, 178, 349 (Chem. Absrr., 1968, 68, 10x514~). 62. Ref. 48, p. 116. 63. T. Katsura and H. Abe (Dowa Mining Co.), US Pat. 3 920 450 (1975) (Chem. Absrr., 1976, 84, 77 381s). 64. G. K. Schweitzer and M. M. Anderson, Anal. Chim. Acta, IY68,41,23. 65. M . Tanaka, N. Nakasuka and H. Yamada, J. Inorg. Nuel. Chem., 1970,32,2759. 66. D. S.Flett and M. J. Jaycock, in ‘Ion Exchange and Solvent Extraction’, ed. J. A. Marinsky and Y.Marcus, Dekker, New York, 1973, vol. 3, p. 1. 67. G. M. Ritcey and A. W. Ashbrook, ‘Solvent Extraction: Principles and Applications to Process Metallurgy’, Elsevier, Amsterdam, 1972, part 11, p. 282. 68. D. S. Flett, Chem. Eng. (London), July 1981, 321. 69. E. D . Nogueira, J. M. Regife and P. M. Blythe, Chem. Ind. (London), 1980, 63. 70. Ref. 67, p. 177. 71. L. White, Eng. Min. J., 1976, 177 (I), 87. 72. Ref. 67, p. 386. 73. S. Amer, Rev. Metal. (Madrid), 1981, 17, 245 (Chem. Abstr., 1982, 96, 187969~). 74. Z. Kolarik, in ‘Solvent Extraction Reviews’, ed. Y. Marcus, Dekker, New York, 1971, vol. I,p. I . 75. C. F. Baes, Jr., J. h r g i Nucl. Chem., 1962, 24, 707. 76. Z. Kolarik and R. Grimm, J. Inorg. Nucl. Chem., I976,38, 1721. 77. T. Sato and T. Nakamura, J. Inorg. Nucl. Chem., 1972, 34,3721. 78. J. S. Preston, Hydrometallurgy, 1983, 10, 187. 79. T. Sato and M. Ueda, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 74)’, Society of Chemical Industry, London, 1974, vol. 1, p. 871. 80. T. Sato, T. Nakamura and M. Kawamura, in ‘Proceedingsof the International Solvent Extraction Conference (ISEC 77)’, Canadian Institute of Mining and Metallurgy, Montreal, 1979, vol. 1, p. 159. 81. J. E. Barnes, J. H. Setchfield and G. 0. R. Williams, J. Inorg. Nucl. Chem., 1976, 38, 1065. 82. R. Grimm and Z. Kolarik, J. Inorg. Nucl. Chem., 1974,36, 189. 83. 1. Komasawa, T. Otake and Y.Higaki, J. Inorg. Nucl. Chem., 1981,43,3351. 84. C . Cianetti and P. R. Danesi, Solvent Extr. Ion Exch., 1983, 1, 9. 85. J. S. Preston, Hydrometallurgy, 1982,9, 115. 86. D. S. Flett and D. W. West, in ‘Complex Metallurgy ’78’,cd. M. J. Jones, Institution of Mining and Metallurgy, London, 1979, p. 49. 87. J. S. Preston, J. S.Afr. Inst. Min. Metall., 1983, 83, 126. 88. D. S. Flett, in ‘Hydrometallurgy: Research, Development and Plant Practice’, ed. K. Osseo-Assare and J. D. Miller, The Metallurgical Society of AIME, New York,1982, p. 39. 89. T. Kasai, H. Nakayama, K. Motoba and E. Itoh, in ‘Preprints for Technical Session D-4, Fourth Joint Meeting MMIJ-AIME’, MMIJ-AIME, Tokyo, 1980, p. 41. 90. M. Ando, M. Takahashi and T. 0gata;in ‘Hydrometallurgy: Research, Development and Plant Practice’, ed. K. Osseo-Assare and J. D. Miller, The Metallurgical Society of AIME, New York, 1982, p. 463. 91. E. D. Nogueira, J. M. Regife and M. P. Viegas, in ‘Chloride Electrometallurgy’, ed. P. D. Parker, The Metallurgical Society of AIME, New York, 1982, p. 59. 92. Anon., World Min., March 1966,40. 93. Anon., WorldMin., May 1982, 100. 94. C. S. G. Phillips and R. J. P. Williams, ‘Inorganic Chemistry’, Oxford University Press, Oxford, 1966, vol. 2, p. 118. 95. Ref. 67, p, 399. 96. Ref. 67, p. 407. 97. 8. H. Lucas and G. M. Ritcey, CIA4 Bull., 1975,68 (7531, 124. 98. D. F. Peppard, G. W. Mason, J. L. Maier and W. J. Driscoll, J . Inorg. Nucl. Chem., 1957,4: 334. 99. D. Dyrssen and L. D. Hay, Acta Chem. Scand., 1960,14, 1100. 100. 2. Kolarik and H. Pankova, J. Inorg. Nucl. Chem., 1966,28, 2325. 101. T. Harada, M. Smutz and R. G. Bautista, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 71)’, Society of Chemical Industry, London, 1971, VOI. 2, p. 950. ”

Applicution to Extractive Metallurgy 102. 103. 104. 10s. 106. 107.

D. F. Peppard and J. R. Ferraro, 1. Inorg. Nucl. Chew., 1959, 10,275.

L. D. Lash and J. R. Ross, J . Mei., 1961, 13, 555. R. G . Canning. Proc. rtusrrulu. Inst. Min. Metall., 1961. No. 198, 113. I. Fidelis and S. Siekierski, J. Inorg. Nucl. Chem., 1967, 29, 2629. S. Siekierski, J. Jnorg. Nml. Chem., 1970, 32, 519. D. F. Peppard, G. W.Mason and S. Lewey, J . Inorg. Nucl. Chem., 1969,31, 2271. 108. D. F. Peppard, C. A. A. Bloomqnist, E. P. Horwitz, S. Lewey and G . W. Mason, J . Inorg. Nucl. Chem., 1970, 32, 339. 109. T. B. Pierce and P. F. Peck, Analyst, 1963,88, 217. 1 IO. L. J. Nugent, J. Inorg. Xucl. Chem., 1970, 32, 348.5. 111. B. Weaver, in ‘Ion Exchange and Solvent Extraction’, ed.J. A. Marinsky and Y. Marcus, Dekker, New York, 1974, vol. 6, p. 189. 112. R. W. Cattrall, Ami. J . Chcm., 1961, 14, 163. 113. C. J. Hardy, B . F. Greenfield and D. Scargill, J . Chem. Soc., 1961, 174. tl4. R. C. Merritt, ‘The Extractive Metallurgy of Uranium’, Colorado School of Mines Research Inslitutc, Golden, CO, 1971, pp. 204,422, 447. 1 I S . C. F. Baes, Jr., R. A. Zingaro and C. F. Coleman, J . Phys. Chem., 1958.62, 129. 116. J. Kennedy and A. M. Deane, J . Inorg. Nucl. Chem., 1961,19. 142. 117. T. Sato, J . Inorg. Nucl. Chem., 1962, 24, 699; 1963, 25, 109. 118. C. A. Blake, Jr., C. F. Baes, Jr. and K. B. Brown, Ind. Eng. Chem., 1958,50, 1763. 119. T. Rigg and J. 0.Garner, J . Inorg. Nucl. Chem., 1967,29,2019. 120. R. C. Ross, Eng. Min. J., 1975, 176 (12), 80. 121. F. J. Hurst, W. D. Arnold and A. D. Ryon, Chem. Eng. ( N Y ) , 1977, 84 (I), 56. 122. B. F. Greek, 0. W. Allen and D. E. Tynan, Ind. Eng. Chem., 1957, 49,628. 123. J. M. Demarthe and S. Solar, in ‘Extraction Metallurgy 81’: Institution of Mining and Metallurgy, London, 1981, p. 42. 124. F. Habashi, J . Inorg. hrucl. Chem., 1960, 13, 125. 125. D. Grdcnic and B. Kurpar, J~ Inorg. Nucl. Chem., 1959,12. 149. 126. M. Zangen, Y. Marcus and E. D. Bergman, Sep. Sci., 1967.2, 1137. 127. Y . Marcus, Chem. Rev.,1963.63, 139. 128. F. J. Hurst and D. J. Crouse, Ind. Eng. Chem., Process Des. Drv., 1974, 13, 286. 129. F. J. Hurst, D. J. Crouse and K. B. Brown, Ind. Eng. Chem., Process Des. Dev., 1972.11, 122. 130. W. W. Berry and A. Dubreucq, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 80)’. Association des Ingenieurs sortis de I’Universite de Liege, Liege, 1980, vol. 3, paper 80-235. 131. M. Zangen, J . Inorg. Nucl. Chem., 1963,25, 581. 132. Y. Marcus and A. S. Kertes, ‘Ion Exchange and Solvent Extraction of Metal Complexes’. Wiley-Interscience, New York, 1969, p. 836. 133 A. S. Kertes, in ’Recent Advances in Liquid-Liquid Extraction’, ed. C. Hanson, Pergamon, Oxford, 1971, p. 15. 134. G . Duyckaerts and J. F. Desreux, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 77)’, Canadian Institute of Mining and Metallurgy, Montreal, 1979, vol. I , p. 73. 13s. D. F. Peppard, G . W. Mason and R. J. Sironen, J. Inorg. Nucl. Chem., 1959, IO, 117. 136. T. W. Lee and G. Ting, Sep. Scz. Technol., 1981, 16,943. 137. I. W. Vermeulen, J. Met., 1966, 18, 22. 138. D. F. Peppard, C.W.Mason and S. McCarty, J . Inorg. Nucl. Chem., 1960, 13, 138. 139. D. F. Peppard and J. R. Ferraro, J . Inorg. Nucl. Chem., 1959,10, 275. 140. S. Amer, Rev. &feral.( M a d r i d ) , 1981, 17, 109 (Chem. Absrr., 1982, 96, 107684k). 141. I. P. Tulenkov, E. I. Kazantsev and I. F. Khudyakov, Sov. J. Non-ferrous Met. (Engl. Transl.), 1976, I7 (3), 6. 142. T. Sato, J . Inorg. ;IrucI. Chem., 1975, 37, 1485. 143, I. S. Levin, T. F. Rodina and G. A. Marinkina, Sov. J. Non-ferrolrs Met. (Eng. Transl.), 1970, 11 (3), 51. 144. D. S. Flett, Hydromefalhrgy, 1976, I , 207. 145 D. S. Flett, in ‘Hydrometallurgy Research, Development and Plant Practice’, ed. K. Osseo-Assare and J. D . Miller, The Metallurgical Society of AIME, New York. 1982, p. 39. I46 K.J. Whewell and C. Hanson, in ‘Ion Exchange and Solvent Extraction’, ed. J. A. Marinsky and Y. Marcus, Dekker. New York, 1981, vol. 8, p. I . 147, K. Soldenhoff and J. S . Preston, Council for Mineral Technology, unpublished data, 1984. 148 I. Dahl, A n d . Chim. Acto, 1968, 41, 9. 149 L.Calligaro, A. Mantovani, U. Belluco and M. Acampora, Polyhedron, 1983,2, 1189. 150. K. Burger, F. Ruff, I. Ruff and I. Egyed, Acta Chim. Acad. Sci. Hung., 1965,46, I (Chem. Abstr., 1966,64,4569d). 151. K. Burger and I. Egyed, J . Inorg. Nucl. Chem., 1965,27, 2361. 152. R. Chiarizia and P.R. Danesi, Sep. Sei. Technol., 1981, 16, 1181. 153. C. A. Fleming, B. R. Green and K. G . Ashurst, in ‘Proceedings of the International Solvent Extraction Conference (ISEC go)’, Association des Ingenieurs sortis de I’Universitk de Liege, Liege, 1980, vol. 2, paper 80-224. 154. M. A. Jarski and E. C. Lingafelter. Acta Crystallogr., 1964, 17, 1109, 155. P. OBrien and J. R. Thornback, Hydrometallurgy, 1982, 8, 331. 156. K. Burger, t.Korecz, 1. B. A. Manuaba and P. Mag, J . Inorg.Xucl. Chem., 1966, 28, 1673. 157. J. B. Willis and D. P. Mellor, J . Am. Chem. Soc., 1947, 69, 1237. 158. A. Chakravorty, Coord. Chem. Rev., 1974, 13, I . 159. V. Romano, F. Maggio and T. Pizzino, J . Inorg. Nucl. Chem., 1971,33,261 I . 160. T. S. Kannan and A. Chakravorty, Inorg. Chem., 1970,9, 1153. 161. J. W. Hosking and N. M. Rice, Hydrometallurgy, 1978,3, 217. 162. R . D. Eliasen and E. Edmunds, Jr., CIM Bull., 1974, 67 (742),82. 163. S. Nishimura, in ‘Extraction Metallurgy 81’, Institution of Mining and Metallurgy, London, 1981, p. 404. 164. Ref. 67, p- 334. 165. K. Akiba and H . Freiser, Sep. Sci. Technol., 1982, 17,745. 166. J. L. Drobnick and W. A. MillsaD (General Mills. Inc.). US Pat. 3276863 (1966) (Chem. Abstr.. 1966.66. 13 119hl. 167. J. S. Preston, J. Inorg. Nucl. Chem.; 1975, 37, 1235. ,

I .

.

I

.

1

,

838

Application io Exbruc five Metallurgy

168. D. S. Flett, D. N. Okuhara and D. R. Spink, J. Inorg. J%’ucl.Chenr., 1973, 35,2471. 169. A. J. van der Zeeuw and R. Kok, in ‘Proceedings of the International Solvent Extrdclion Conference (ISEC 77)’, Canadian Institute of Mining and Metallurgy, Montreal, 1979, vol. I , p- 17. 170. J. S. Preston, J. Inorg. Nucl. Chem., 1980, 42,441. 171. A. R. Burkin and J. S. Preston, J . Inorg. Nucl. Chem., 1975, 37. 2187. 172. J. S. Preston and Z . B. Luklinska, J . Inorg. Nucl. Chem., 1980.42, 431. 173. M. J . Nicol, J. S. Preston, J . A. Ramsden and M. Mooiman, Hydrometallurgy, 1985, 14, 83. 174. C. F. Coleman. K.B. Brown, J. G. Moore and D. J. Crouse. Ind. Eng. Chem.. 1958,50.1756. 175. R. M. Diamond, in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed. D. Dyrssen, J. 0. Liljenzin and J . Rydberg, North-Holland. Amsterdam, 1967, p. 349. 176. J . S . Preston, Sep. Sci. Technol., 1982. 17, lb97. 177. C. F. Coleman, Nucl. Sei. Eng., 1963, 17, 274. 178. R. R. Grinstead, in ‘Proceedings of the International Confcrence on Solvent Extraction Chemistry’. ed. D. Dyrssen, J. 0 .Liljenzin and J. Rydberg, North-Holland, Amsterdam, 1967, p. 426. 179. F. G . Sccley and D. J. Crouse, J . Chem. EnK. Datu, 1966,11,424. 180. Y . Marcus and A. S. Kertes, ‘Ion Exchange and Solvent Extraction of Metal Complexes’, Wiley-Interscience, New York, 1969. p. 790. 181. W. Aspeling and J. S. Preston, Council for Mineral Technology, unpublished data, 1984. 182. R. M. Smith and A. E. Martell, ‘Critical Stability Constants’. Plenum, New York, 1976, vol. 4, p. 104. 183. L. G . Sillen and A. E. Martell, ‘Stability Constants of Metal-ion Complexes’, Special Publication No. 17. The Chemical Society, London, 1964, p. 271. 184. M . L. Good and S. E. Bryan, J . Inorg. Nucl. Chem., 1961,ZO. 140. 185. M . L. Good and S. C. Srivastava, J . Inorg. Nucl. Chem., 1965, 27, 2429. 186. T. Sato, K. Adachi, T. Kat0 and T. Nakamura, Sep. Sci. Technol., 1982, 17, 1565. 187. W. E. Keder, J . Inorg. Nucl. Chem., 1962, 24, 561. 188. 0. Levy, G. Markovits and A. S . Kertes. J . InorK. ,Vul:uci. Chem., 1971, 33, 551. 189. M. Aguilar and M. Muhamrncd, J . Inorg. Nucl~Chem., 1976,38, 1193. 190. C. Fischer, H. Wagner and V. V. Bagreev, Polyhedron. 1983, 2. 1141. 191. P. G . Thornhill, E. Wigstol and G. van Weert, J . Met., 1971,23 (7), 13. 192. Ref. 67, p. 307. 193. C. Bozec. J . M. Demarlhe and L. Gandon, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 74)’. Society of Chemical Industry. London, 1974. vol. 2. p. 1201. 194. A. Suetsuna, N. Ono, T. Iio and K. Yamada, The Metallurgical Society of AIME, TMS Paper Selection A 80-2 (1980). 195. C. F. Coleman, K. B. Brown, J . G . Moore and D. J . Crouse, Ind. Eng. Chem., 1958,50. 1756. 196. R.C. Merritt, ‘The Exlractive Metallurgy of Uranium’, Colorado School of Mines Research Institute, Golden, CO, 1971, p. 379. 197. Ref. 67, p. 452. 198. A. I. Bellingharn, Proc. Australas. Inst. Min. Mefall., 1961, No. 198, 85. 199. S . A. Finney, in ’Proceedings of the International Solvent Extraction Conference (ISEC 77)’, Canadian Institute of Mining and Metallurgy, Montreal, 1979, vol. 2, p. S67. 100. D. W. Boydell and E. B. Viljoen, in ‘Proceedings o f the 12th Council of Mining and Metallurgical Institutions Congress’, ed.H . W. Glen, South African Institute OF Mining and Metallurgy, Johannesburg, 1982, p. 575. 201. I. E. Lewis and S . Kesler, in ‘Proceedings of the lnlernational Solvent Extraction Confcrcnce (ISEC KO)’, Association des Ingenieurs sortis de l’Universit6 de LPge, Li&ge$1980, vol. 3, paper 80-146. 202. C. A. Fleming, in ’Vacation School on Uranium Ore Processing’, South African Institute of Mining and Metallurgy, Johannesburg, 1981, p.8/1. 203. W. J . McDowell and C. F. Coleman, in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed. D. Dyrssen, J. 0. Liljenzin and J . Rydberg, North-Holland, Amsterdam, 1967, p. 540. 204. T. Sato, J. Inorg. Nucl. Chem.. 1963, 25, 441. 205. R. W. Cattrall and S. J. E. Slater, Coord. Chem. Rev., 1973, 11, 227. 206. F. G. Seeley and D. J. Crouse, J . Chem. Eng. Dafa, 1971, 16, 393. 207. J, E. House, in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed. D. Dyrssen, J . 0. Liljenzin and J. Rydberg, North-Holland, Amsterdam. 1967, p. 641. 208. J . E. Lik, in ‘Exlraction Metallurgy of Refractory Metals’, ec. H. Y . Sohn, 0.N . Carlson and J . T. Smith, The Metdhrgical Society of AIME, New York, 1980, p. 69. 209. B. N. Laskorin, V. S. Ul’yanov and R. A. Sviridova, Zh. Przkl. Khim., 1965,38, 1133 (Chum. Abslr.. 1965,63, 5001~). 210. R. R. Swanson. H. N . Dunning and J. E. House, Eng. Min. J., 1961, 162 (lo), 110. 211. S. W. H. Yih and C. T. Wang, ‘Tungsten: Sources, Metallurgy, Properties and Applications’l Plenum, New York, 1979, pp. 104, 119. 212. T. K. Kim, R. W. Mooney and V. Chiola, Sep. Sci., 1968,3,467. 21 3. D. N. Gibson, in ‘International Molybdenum Encyclopaedia’, ed. A. Sutulov, Intermet Publications, Santiago de Chile, 1979, vol. 2, p. 260. 214. A. S. Vieux, Kabwe-wa-Bibombe and M. Nsele, Hydrometallurgy, 1980, 6, 35. 215. F. A. Cotton and G. Wilkinson, ‘Advanced Inorganic Chemistry’, 4th edn., Wiley, New York. 1980. p. 853. 216. M. B. MacInnis, T. K. Kim and J . M. Laferty, J. Less-Common Met., 1974, 36, 111. 217. M. B. MacInnis and T. K. Kim, J . Chem. Technol. Biotechnol., 1979, 29, 225. 218. C. J. Lewis and J. E. House, Trans. Min. Snc. A I M E , 1961, 220, 359. 219. R. J. Dowsing, Met. Mater., July 1980, 32. 220. R. I. Edwards, J . M e t . , August IY76, 4. 221. F. S. Clemwts, Ind. Chem., 1962,.38, 345. 222. R. I. Edwards. in ‘Proceedings of the International Solvcnt Extraction Conference (ISEC 77)’. Canadian Institute o f Metallurgy, Montreal, Special Volume 21. 1979, vol. 1 , p. 24. 223. M. J . Cleare, P. Charlesworth and D. J . Bryson, J . Chem. Trchnol. Biotechnol., 1979, 29, 210. 224. L. R.P. Reavill and P. Charlesworth, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 80)’, Association des Ingenieurs sortis de I’Universitt: de Liege, Liege, 1980, vol. 3, paper 80-93. 225. P. Charlesworth, Platinum Met. Rev., 1981, 25, 106.

Application to Extractive Metallurgy

839

226. J. E. Barnes and J. D. Edwards, Chem. Ind. (London), 6 March 1982, 151. 227. R. I. Edwards and W. A. M. te Riele, in ‘Handbook of Solvent Extraction’. ed. T. C. Lo, M. H. I. Baird and C. Hanson, Wiley, New York. 1983, p. 725. 228. V. A. Mikhailov, V.G .Torgov, E. N. Gil’bert, L. N. Mazalov and A. V . Nikolaev, in ‘Proceedings ofthe International Solvent Extraction Conference (ISEC 71)’, Society of Chemical Industry, London, 1971, p. 11 12. 229. R. I. Edwards (National Institute for Metallurgy), S. Afr. Pal. 74 05 109 (1974) (Chem. Abstr., 1977. 87, 121008p). 230. K. J. Shanton and R. A. Grant (Matthey Rustenburg Refiners Ltd.), Ger. Pat. 2901 733 (1979) (Chem. Absrr., 1980. 92, 150688~). 231. W. J. Holland, R. A. Dimenna and R. J. Walker, Mikrochim. Acra, 1977,483. 232. D. F. C. Morris and M. A. Khan. Tulunta, 1968, 15, 1301. 233. M. Fieberg, C. Conncll and W. A. M. Te Riele, Natl. Inst. Merall. Rep. No. 1996. (1978). 234. H. C. Sergeant and N . M. Rice, in ‘Proceedings of the rnternational Symposium on Chloride Hydrometallurgy’+ Societe Benelux Metallurgie, Brussels, 1977, p. 385. 235. M. Zicgler and 0. Glemser, Angew. Chum, 1956, 68, 620 (Chem. Abstr., 1958, 52. 142870. 236. 1’. I. Bobikov, L. M . Findin, E. F. Kouba, A. P. Sokolov and V. 1. Dolgikh. USSR Pur. 142768 (1961) (C‘hPm. Abstr., 1962, 56, 13876f). 237. V. I. Dolgikh, P. I. Bobikov: V. E. Rorbat and M. R. Ferberg. Tr, Vses..Nuuchno-Tekh. Sovushch. Protsrss,v Zhidk. Ekstr. Khamosorbtsii. h d , Lmingrud. 1964, 307 (Chem. Abstr., 1966, 65, 1423d). 238. V. I. Dolgikh, P. I. Bobikov, V. F. Borbat and M. B. Ferberg, Tsvetn. Met., 1967, 40,27 (Chem. Abstr., 1967, 67, 14035k). 239. L. M. Gindin, in ‘Ion Exchange and Solvent Extraction’, ed J. A. Marinsky and Y. Marcus, Dekker, New York, 1981, vol. 8. p . 311. 240. V. F. Borbat and E. F. Kouba, Tr. Vses. Nauchno-Tekh. Soveshch. Prorsessy Zhidk. Ekstr. Khemosorbrsii, 2nd. Leningrad, 1964, 299 (Chem. Abstr., 1966, 65. 1 1 4 2 3 ~ ) . 241. R. A. Grant (Matthey Rustenburg Refiners Ltd.), Fr. Pat. 2472025 (1981) (Chem. Abstr., 1982,96,9963f). 242. R. A . Grant (Matthey Rustenburg Refiners Ltd.), Br. f o r . 2065092 (1981) (Chem. Ahstr., 1Y82. 96. 3K80Yr). 243. R. I. Edwards and M. J. Nattras (National Institute for Metallurgy), S. Afr. Pat. 7603679 (1977) (Chem. Abslr.. 1978, 88, 92954g). 244. E. W. Berg and W. L.Senn, Jr.. Anal. Chim. Acta, 1958, 19, 12, 109. 245. A. T. Casey. E. Davies. T. L. Meek and E. S. Wagner. in ‘Proceedings of the International Conference on Solvent Extraction Chemistry’, ed. D. Dryssen, J. 0. Liljenzin and J. Rydberg, North-Holland, Amsterdam, 1967. p. 327. 246. M. Knothe, Z.Anorg. Aflg. Chem., 1980, 470, 216 (Chem. Absrr., 1981, 94, 21 250n). 247. M. A. Khan and D. F. C. Morris, Sep. Sci., 1967, 2, 635. 248. N. M. Sinitsyn, F. Ya. Rovinskii and V . V . Ispravnikova, Dokl. Akad. Nauk SSSR, 1967, 176, 1320 (Chem. Abstr., 1968,69, 133870. 249. M. M. Fieberg and R. I. Edwards (National Institute for Metallurgy). S. Afr. Pat. 76 03680 (1977) (Chem. Abstr., 1978,89, 46980k). 250. 1.. M. Gindin, S.N. Ivanova, A . A. Mazurova, A. A. Vasil’eva, L. Ya. Mironova, A. P. Sokolov and P. P. Srnirnov, Izv. Sib. Old. Akad. ,Nauk SSSR. Ser. Khim. Nauk, 1967, 89 ( C h m . Absrr., 1967, 67, 68 138q). 251. E. Peligol, Ann. Chim.Phys., 1842, 5, 5. 252. R. G . Bellamy and N. A. Hill, ‘Extraction and Metallurgy ofuranium. Thorium and Beryllium’, PerEamon, Oxford, 1963, p. 32. 253. C. E. Winters (US Government). C’S Pat. 2859091 (1958) (Chem.Abstr., 1959,53, 5612~). 254. E. Glueckauf and H . A. C. McKay, Nature (Londonj, 1950, 165, 594. 255. H. A. C. McKay and A . R.Mathieson, Trans. Faraday Sac., 1951. 41,428. 256. E. Glueckauf, H. A . C. McKay and A. R. Mathieson, Trans. Fmaduy Sac., 1951,47.437. 257. Ya. I. Ryskin, V. N. Zemlyanukhin, A. A. Solov’eva and N. A. Derbeneva. Russ. J. Inorg. Chem. (Engl. Transl.), 1959, 4, 174. 258. Y. I. Ryskin, V. P. Shvedov and A. A. Solov’eva, Russ. J. Inorg. Chem. (Engl. Transl.), 1959,4, 1033. 259. A. Muller, Ber. Drsch. Chern. Ces. B, 1940.73, 1353 (Chem. Abstr., 1941, 35, 2435). 260. A. R. Mathieson, J. Chem. Soc., 1949, S 294. 261. T. H. Tunley and V . W. Nel, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 74)‘, Society of Chemical Industry, London, 1974, vol. 2, p. 1519. 262. I. L. Jenkins, Hyrdrometailur~,1979,4, I . 263. T. Sato, J. Appl. Uhem., 1965, 489. 264. R. L. Moore, US Atomic Energ-v Commission Report, AECD 3196 (1951). 265. H. A. C. McKay, Chem. Ind. (J,ondon), 1954, 1549. 266. H. Hesford, H. A. C. McKay and D. Scargill, J . Inorg. Nucl. Chem., 1957,4, 321. 267. E. P. Maiorova and V. V. Fomin, Russ. J . Inorg. Chem. (Engl. Transl.), 1959,4, 1156. 268. L. I. Katzin, J . Inorg. Nuel. Chem., 1957,4. 187. 269. T. H. Siddall, 111, Ind. Eng. Chem., 1959, 51, 41. 270. T. H. SiddalI, 111, J. Inorg. Nucl. Chem., 1960, 13, 151. 271. W. D. Jamrack. ‘Rare Metal Extraction by Chemical Engineering Techniques’, Pergamon, Oxford, 1963, p. 176. 272. G. Werner, H. Giseke and H. Holzappel, J . Less-Common Met., 1966, 11, 209. 273. V. P. Judin and H. E. Sund, in ‘Hydrometallurgy SI’, ed. G. A. Davies, Society of Chemical Industry, London, 1981, p. F4/1. 274. D. F. Peppard, J. P. Faris, P. R. Gray and G . W . Mason, J. Phys. Chem., iY53,57, 294. 275. D. F. Peppard, I . P. Faris, P. R . Gray and M . M. Markus, J. Am. Chem. Soc., 1953,75, 6063. 276. D. F. Peppard, W. J . Driscoll, R. J. Sironen and S . McCarty, J . Inorg. Nucl. Chem., 1957,4, 326. 277. E. Hesford, E. E. Jackson and H. A. C. McKay, J . Inorg. NucI. Chem., t959, 9, 279. 278. I. Fidelis, J. Inorg. Nucl. Chem., 1970, 32, 997. 279. E. Hesford and H. A. C. McKay, Trans. Faraday SOC.,1958,54, 473. 280. H. Bostian and M. Smutz, J . h o g . Nucl. Chem., 1964, 26, 825. 281. F. V. Robinson and N . E. Topp, J. Inorg. Nucl. Chem., 1964,26. 473. 282. D. Karraker, J . Chem. Educ., 1970,47,424. 283. T. H. Siddall, 111, W. E. Stewart and D. G. Karraker, Inorg. Nucl. Chem. Lett., 1967,3, 479.

840

Application to Extractive Me taIlurgy

284. S. Z. Ndzuta, E. W. Giesekke and K. G. R. Pachler, J. Inorg. Nucl. Chem., 1980,42, 1067. 285. R. P. Cox, H. C. Peterson and G. H. Beyer, Ind. Eng. Chem., 1958, 50, 141. 286. F. Hudswell and J. M. Hutcheon, in ‘Extraction and Refining of the Rarer Metals’, lnstitution of Mining and Metallurgy, London, 1957, p. 402. 287. R. H.Nielsen, in ‘The Metallurgy of Hafnium’, ed. D. E. Thomas and E. T. Hayes, United States Atomic Energy Commission, Washington, DC, 1960, p. 49. 288. W. C. Jamrack, ‘Rare Metal Extraction by Chemical Engineering Techniques’, Pergamon, Oxford. 1963, p. 180. 289. D. Royston and P. G. Alfredson, Australian Atomic Energy Commission Report AAEC/TM538, 1970 (Chem. Abstr., 1971,74, 56354~). 290. Ref. 67, p. 542. 291. K. Alcock, F. C. Bedford, W. H. Hardwick and H. A. C. McKay, J . Inorg. Nucl. Chem., 1957,4, 100. 292. D. F. Peppard, G. W. Mason and J. L. Maier, J . Inorg. N u d Chem., 1956,3, 215. 293. A. S. Solovkin, Zh. Neorg. Khim., 1957,2, 611 (Chem. Abstr., 1957,51, 1735Lg). 294. G. F. Egorov, V. V. Fomin, Yu. G . Frolov and G. A. Yagodin, Russ. J. Inorg. Chem. (Engl. Transl.), 1960,5,503. 295. J. C. Mailen. D. E. Homer, S. E. Dorris, N. Pih, S. M. Robinson and R. G . Yates, Sep. Sci. Technol.. 1980, 15, 9.59. 296. A. S. Solovkin and Z. N. Tsvetkova, Usp. Khim., 1962, 31, 1394 (Chem. Abstr., 1963, 58, 5069g). 297. 0. A. Sinegribova and G. A. Yagodin, Russ. J. Inorg. Chrm. (Engl. Transl.), 1971, 16, 1194. 298. G . A. Yagodin, 0. A. Sinegribova and A. M. Chekmarev, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 71)’, Society of Chemical Industry, London, 1971, vol. 2, p. 1124. 299. G . A. Yagodin, 0. A. Sinegribova and A. M. Chekmarev, in ‘Proceedings of the International Solvent Extraction Conference (lSEC 74)’, Society of Chemical Industry, London, 1974, vol. 3, p. 2209. 300. W. Fischer and W. Chalybaeus, Z, Anorg. Allg. Chem., 1947,255,79 (Chem. Abstr., 1949,43, 2883~). 301. W. Fischer, W. Chalybaeus and M. Zumbusch, 2 Anorg. Allg. Chem., 1947,255,277 (Chem. Abstr., l949,43,2883e). 302. L. G. Overholser, C. J. Barton and W. R. Grimes, US Atomic Energy Commission Report Y-477 (1949) (Chem. Absir., 1956, 50, 116264). 303. Anon., Eng. Min. J., 1964, 161 (I), 78. 304. E. Foley, In ‘Extractive Metallurgy of Refractory Metals’, ed. H. Y. Sohn, 0. N . Carlson and J. T. Smith, The Metallurgical Society of AIME, New York, 1980, p. 341. 305. J. H. McClain and S. M. Shelton, in ‘The Reactor Handbook’, ed. C. R. Tipton, Interscience, New York, 1960, vol. 1, p. 64. 306. E. M. Larsen, in ‘Advances in Inorganic Chemistry and Radiochemistry’, ed. H. J. Emeleus and A. G. Sharpe, Academic, New York. 1970, vol. 13, p. 1 . 307. A. E. Laubscber and K. F. Fouche, J. Inorg. Nucl. Chem., 1971,33,3521. 308. 0. A. Sinegribova and G. A. Yagodin, Rues. J. Inorg. Chem. (Engl. Transl.), 1965,10, 675. 309. N. F. Rusin and I. V. Vinarov, Rum. J. Inorg. Chem. (Engl. Trawl.), 1969,14, 559. 310. 0. A. Sinegribova and G . A. Yagodin, Russ. J. Inorg. Chem. (Engl. Trawl.), 1971,16, 1194. 311. 0. A. Sinegribova and G. A. Yagodin, Russ. J. Inorg. Chem. (Engl. Transl.), 1971, 16, 1471. 312. N. A. Shostenko, L. G. Nekhamkin and G. A. Yagodin, Russ.J. Inorg. Chem. (Engl. Transl.), 1973, I S , 706. 313. Yu. A. Tsylov, N. A. Shostenko and L. I. Lyubushkina, R w s . J. Inorg. Chem. (Engl. Trunsl.), 1977,22, 1361. 314. I . V. Vinarov, A. I. Orlova, L. I. Il’chenko and L. P. Grigor’eva, Russ. J. Inorg. Chem. (Engl. Transl.), 1.977,22, 570. 315. Yu.Ya. Kharitonov, I. A. Rozanov and I. V. Tananaev, Izv. Akad. Nauk SSSR, Otd. Khim. M a d , 1963,596 (Chem. Absir., 1963, 59, 46713g). 316. D. 0.Voit, in ‘Proceedings of the International Solvent Extraction Conference (ISEC go)’, Association des Ingenieurs sortis de 1’Universite de Liege, Liege, 1980, vol. 2, paper 80-59. 317. F. Fairbrother, ‘TheChemistry of Niobium and Tantalum’, Elsevier, Amsterdam, 1967, p. 7. 318. C. Placek and D. F. Taylor, Ind. Eng. Chem., 1956,48,686. 319. P. C. Stevenson and H. G. Hicks, Anal. Chem., 1953,25, 1517. 320. J. R. Werning, K. €3. Higbie, J. T. Grace, B. F. Speece and H. L.Gilbert, Ind. Eng. Chem., 1954,46,644. 321. C . W. Carlson and R. H. Nielsen, J. Met., 1960, 12 (6), 472. 322. D. J. Soisson, J. J. McLafferty and J. A. Pierret, Ind. Eng. Chem., 1961, 53, 861. 323. G. P. Giganov and T. I. Yarinova, Sov.J. Non-ferrcius Met. (Engl. Transl.), 1975, 16 (6), 17. 324. J. M. Fletcher, D. F. C. Morris and A. G. Wain, Trww-Inst. Min. Metall., 1955-56, 65, 487. 325. S. Nishimura, J. Moriyama and I. Kushima, Trans. Jpn. Inst. Met., 1963,4, 259. 326. I. I. Baram, G . E. Kaplan and B. N. Laskorin, Russ. J. Inorx. Chem. (Engl. Trawl.), 1965, 10,272. 327. G. M.Ritcey, B. H. Lucas and K. T. Price, Hydroma#dZurgy. 1982,8, 197. 328. E. D. Nogueira and P. Cosmen, Hydrometallurgy, 1983,9,333. 329. R. F. Dallon, R.Price and P.M. Quan, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 83)’, American Institute of Chemical Engineers, New York, 1983, p. 189. 330. T.Danielssen, G. H. Boe, P. M. Finne and B. Townson, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 83)’, American Institute of Chemical Engineers, New York, 1983, p. 536. 331. S. K. Majumdar and A. K. De, Tdanta, 1960,7, 1. 332. H. Specker and M. Cremer, Z. Anal. Chem., 1959,167, 110 (Chem. Abstr., 1959,53, 18731a). 333. I. V. Seryakova, Yu. A. Zolotov, A. V. Karyakin, L. A. Gribov and M. E, Zubrilina, R w s . 1.Inorg. Chem.(Engl. Transl.), 1962,7, 1039. 334. A. H. Laurene, D. E. Campbell, S. E. Wiberley and H. M. Clark, J. Phys. Chem., 1956,60, 901. 335. H. L. Friedman, J. Am. Chem. Soc., 1952,74, 5. 336. Yu. A. Zolotov, 1. V. Seryakova, 1. 1. Antipova-Karataeva, Yu.1. Kutsenko and A. V. Karyakin, Russ. J. Inorg. Chem. (Engl. Transl.), 1962,7, 615. 337. V. V. Fomtn and A. F. Morgunov, Rues. J. Inorg. Chern. (Engl. Transl.), 1960,5,670. 338. I. V. Seryakova, Yn.A. Zolotov, A. V. Karyakin and L. A. Gribov, Russ. J. Inorg. Chem. ( h g l . Trans!.), 1963, 8, 244. 339. I. V. Seryakova, Yu. A. Zolotov, A. V. Karyakin, L. A. Gribov and M. E. Zubrilina, Russ.J. Inorg. Chem. (Engl. TrMsl.), 1962,7, 1039. 340. G. R. H e a l and H. M. Clark, J. Phys. Chem., 1961,66, 1930. 341. J. T. Way, J. R . Agric. SOC.Engl., 1850,.11, 313; 1852,13, 123. 342. H. S. Thompson, J . Agric. SOC.Engl., 1850, 11, 68.

Application to Ex true tive Metallurgy

84 1

R. Gans, Jahrb. P r m s . Geol. Lmdesanst., 1905,26, 179. B. A. Adams and E. L. WoImes, J. Soc. Chem. Ind., London, Trans. Coinmun., 1935,54, 1. F. Helfferich, ‘Ion Exchange’, McGraw-Hill, New York, 1962, p. 168. H. W. Kauczor, Inst. Chem. Eng. Symp. Ser., 1975,42, 20.1. D. Herve, Ind. Miner. [Ser.] Mineralurgie, 1977, 3, 173. A. Warshawsky, Hydrometallurgy, 1976177, 2, 197. R. H. Stockes and H. F. Walton, J. Am. Chem. Soc., 1954,76, 3327. I. K. Tsitovich and N. G. Nikitina, Izv. Vyssh. Uchebn. Zaved. Khim. a i m . Tekhnol., 1963,6, 567. R. K. Drachevskaya, N. F. Il’uasova, N. V. Dolgashova and N. V. Demenev, Khim. Tekhnol. Sur’my Aknd. Nauk Kirg. SSR Inst. Neorg. Fiz. Khim., 1965, 9. 352. J. J. Katz and G. T. Seaborg, ‘The Chemistry of the Actinide Elements’, Methuen, London, 1957. 353. R. Kunin, ‘loa Exchange Resins’, Robert E. Kreiger, New York, 1972, p. 207. 354. T. J. Hsu et al., Huan Ching Pa0 Hu (Taipei), 1982, 5 (3), 17. 355. T. 1. Scott, V. H . Westlake and M. K. Bridle, US Par. 4376706 (1983). 356. S. Sussman, F. C. Nachod and W. Wood, Ind. Eng. Chem., 1945,37,618. 357. F. Gerstner, Chem.-hg.-Tech.,1954, 26, 264. 358. C. A. Fleming and G. Cromberge, J . S. Afr. Inst. Min. Metall., 1984. 84 (S), 125. 359. K. Krause and G. E. Moore, J . Am. Chem. Soc., 1953,75, 1457. 360. D. Clifford, React. Polym., 1983, 1, 77. 361. B. N. Laskorin, G. I. Sadovnikova, L.N. Petrova and V. D. Fedorov, J. Appl. Chem. USSR (Engl. Transl.). 1977. 47, 1794. 362. J. Davidson, F. 0. Read, F. D. L. Noakes and T. V. Arden, Trans.-Inst. Min. Metall., 1960-61, 70,247. 363. V. I. Demidov, R. Z. Kreines, M. I. Ivanova and A. S. Kartasheva, Sov. J. Non-Ferrous Met. (Engl. TransLj, 1967, 8(8), 53. 364. I. D. Fridman, E. P. Zdorova, L. E. Pochkina, R. Yu. Bek, A. I. Maslii and S. M. Il’ichev, Sov. J . Nan-Fermw Met. (Engl. Traml.), 1971,12 (IO), 76. 365. Anon., ‘Gold from Russia’s Muruntau Deposit’, Consolidated Gold Fields Report Coal, Goldand Base Metals, 1978, 26, 75. 366. C. A. Fleming and G. Cromberge, J. S. Afr. Inst. Min. Metall., 1984,84,269. 367. C. A. Fleming and G . Cromberge, J . S. Afr. Inst. Min. Metall., 1984,84, 369. 368. A. Mehmet and W. A. M.Te Riele, in ‘Ion Exchange Technology’, ad. D. Naden and M. Stread, Society of Chemical Industry, London, 1984, p. 637. 369. C. A. Fleming and G. Cromberge, in ‘MINTEK 50, Proceedings of the International Conference on Mineral Science and Technology’, ed. L. F. Haughton, Council for Mineral Technology, Randburg, 1984, vol. 2, p. 663. 370. F. H. Burstall, N. F. Kember, P. J. Forrest and R. A. Wells, Ind. Eng. Chem., 1953,45, 1648. 371. D. G. Marsh, J. AppI. Phofogr. Eng., 1978,4, 17. 372. H. W. Chou, J. Appl. Photogr. Eng., 1980,6, 14. 373. W. H. Waitz, Plaring Surf. Finish., April, 1982, 56. 374. 0. Samuelsen, ‘Ion Exchange Separations in Analytical Chemistry’, Wiley, New York, 1963, p. 400. 375. J. A. Kitchener, ‘Ion Exchange Resins’, Methuen’s Monographs on Chemical Subjects, Methuen, London, I967. 376. J. Stamberg, J. Seidl and J. Rahm, J . Polym. Sei., 1958,31, 15. 377. R. L. Degeiso, L. G. Donaruma and E. A. Tomic, J. Appl. Polym. Sci., 1963.7, 1523. 378. J. R. Millar, Chem. Ind. (London), 1957,606. 379. A. Warshawsky, Angew. Makromol. Chem., 1982, 109/110, 171. 380. A. Gabert and B . Voiget, Ger. (East) Pat. 44 124 (1965). 381. R. Rosset, Bull. Soc. Chim. Fr., 1966, 59. 382. R. R. Grinstead, J . Met., 1979,31 (3), 13. 383. B. R. Green and R. D. Hancock, in ‘Hydrometallurgy 81’, ed. G. A. Davies, Society of Chemical Industry, London, 1981, p. E6(1). 384. B. R. Green and R. D. Hancock, Hydrometallurgy, 1981,6, 353. 385. H. Egawa, Y. Jog0 and H. Maeda, Nippon Kagaku Kaishi, 1979, 12, 1760. 386. R. J. Philips and 3. S. Fritz, Anal. Chem., 1978,50, 1504. 387. C. 0. Giwa and M. J. Hudson, Hydrometallurgy, 1982,8,65. 388. G . Koster and G. Schmuckler, A n d Chim. Acta, 1967,38, 179. 389. S. Goldstein, M. Silberg and G. Schmuckler, Ion Exch. Memk., 1974, 1,225. 390. A. Warshawsky, Tmns.-Insr. Min. Metall., Sect. C,1974,83, 101. 391. D. S. Flett, Chem. Ind. (London), 1977,641. 392. A. Warshawsky and H. Eerkowitz, Trans.-Inst. Min. Metal!., Sect. C , 1979,88, 36. 393. A. Warshawsky, R. Kalir and H.Berkowitz, Trans.-Inst. Min. Metall., Sect. C , 1979,88, 31. 394. H. W. Kauzcor and A. Meyer, Hydrometallurgy, 1978,3,65. 395. R. Blumberg and L. Logan, Hy&ometallurgy, 1979,4, 389. 396. B. R. Green and R. D. Hancock, Sep. Sei. Technol., 1980,15, 1229. 397. C. A. Fleming, 3.R. Green and K. G. Ashurst, in ‘Proceedings of the International Solvent Extraction Conference (ISEC 80)’, Association des Ingenieurs sortis de 1’Universite de Likge, Likge, 1980, vol. 2, p. 224(1). 398. A. J. Parker, Chem. Rev., 1969,69, 1. 399. F. W. E. Strelow, G. R. van ZyI and C. J. C. Bothma, Anal. Chim. Acta, 1969,45,81. 400. C. A. Fleming and A. J. Monhemius, Hydrometallurgy, 1979,4, 159. 401. R. V. Pammenter and C. J. Haigh, in ‘Extraction Metallurgy ’Sl’, Institution of Mining and Metallurgy, London, 1981, p. 379. 402. P. T. Davey and T. R.Scott, Hydrometallurgy, 1976,2,25. 403. J. E. Dutrizac in ‘Lead-Zinc-Tin ’80’, ed. J. M. Cigan, T. S. Mackey and T. J. O’Keefe, AIME, New York, 1979, p. 532. 404. J. E. Dutrizac, in ‘Hydrometallurgy - Research, Development and Plant Practice’, AIME, New York, 1982, p. 531. 405. R. T. McAadrew, S. S. Wang and W. R. Brown, CIM Bull., 1975, 68, 101. 406. J. R. Bolt and P. Queneau, ‘The Winning of Nickel’, Methuen, London, 1967, p. 336. 407. M. A. Bouchat and J. J. Saquet, J. Met., 1960, 802.

343, 344. 345. 346. 347. 348. 349. 350. 351.

842

Application to Extractive Metallurgy

408. R. S. Dean, ‘Electrolytic Manganese and its Alloys’, Ronald Press, New York, 1957, p. 206. 409. A. I. Belhgham, R. J. Murray and D. E. Collier, in ‘Hydrometallurgy - Research, Development and Plant Practice’, AIME, New York, 1982, p. 87. 410. K. Osseo-Assare, M e f d l . Trans., B, 1982, 13, 555. 41 1. D. J. Evans, in ‘Advances in Extractive Metallurgy’, Institution of Mining and Metallurgy, London, 1968, p. 831. 412. C. R. S. Needes and R. R. Burkin, in ‘Leaching and Reduction in Hydrometallurgy’, Institution of Mining and Metallurgy, London, 1975, p. 91. 413. M. E. Wadsworth, Trans. AIME, 1969,245, 1381. 414. R. M. Nadkarni and M. E. Wadsworth, in ‘Advances in Extractive Metallurgy’, Institution ofMining and Metallurgy, London, 1968, p. 918. 415. M.J. Nicol, E. Schalch and P. Balestra, J. S.Afr. Inst. Min. Metall., 1979, 79, 191. 416. J. E. Barnes and J. I).Edwards, Chem. Ind. (London), 1982, 151. 417. F. Basolo and R. G.. Pearson, ‘Mechanisms of Inorganic Reactions’, Wiley, New York. 1967, p. 351. 418. E. N.Lyons, J. Elecrmchern. Soc., 1954, 101, 363. 419. R. R. Lloyd, I. B. Rosenbaum, V. E. Homme and L. P. Davis, J. Electrochem. Soc., 1948.94, 122. 420. J. R. Bolt and P. Queneau, ‘The Winning of Nickel’, Methuen, London, 1967, p. 251. 421. A. K. Vijh and J. P. Randin, J. Phys. Chem., 1975,79, 1252. 422. D. J. Mackinnon and J. M. Brannen, J. Appl. Electrochem., 1977, 7,451. 423. B. V. Tilak, S. R. Rajagopalan and A. K. Reddy, Tmms. Faraday SOC.,1962,58,795. 424. L. Goudurier, D. W. Hopkins and I. Wilkomirsky, ‘Fundamentals of Metallurgical Processes’, Pergamon, Oxford, 1978. p. 183. 425. J. C. Bowker, C. H. P. Lupis and P. A. Flinn, Can. Metall. Q.,1981, 20,69. 426. V. Rajamani and A. J. Naldrett, Econ. Geol., 1978, 73, 82. 427. S. A. De Waal, Council for Mineral Technology, personal communication.

Geochemical and Prebiotic Systems PETER A. WILLIAMS University College, Cardiff, UK 64.1 INTRODUCTION

843

64.2 MINERALS 64.2.1 Minerals as Coordination Complexes 64.2.2 Specijic Mineralogical Examples of Coordination Compounds

844 844

846

64.3 INORGANIC COMPLEXES IN SOLUTION 64.3.1 Aqueous Solutions at Low Temperatures 64.3.2 Coordination Chemistry of Brines 64.3.3 Inorganic Compnplcxe.~in High-temperature Aqueous Solutions

850

64.4 GASEOUS TRANSPORT AND SILICATE MELTS

854

64.5 COORDINATION GEOCHEMISTRY OF COMPLEXES OF ORGANIC LIGANDS 64.5.1 Complexes of Humic and Fulvic Acids 64.5.2 Complexes of Porphyrins and Related Macrocyclic Ligands 64.5.3 Complexes with Other Organic Ligands 64.5.4 The Possible Involvement of Organic Coordination Compounds in Ore Formation

856 857 86 1 866 868

64.6 PREBIOTIC SYSTEMS

870

64.7 REFERENCES

873

850 852 854

64.1 INTRODUCTION

‘Here certainly we are justified in thinking that chemical principles should give us an insight into the ways of nature.’ K. B. Krauskopf’ Many diverse processes are involved in the transformation of the elements and their compounds in the Earth. Some o f the pathways observed are shown in Figure 1, a version of the so-called ‘geochemical cycle’. This cycle i s very much simplified and i s not a closed one. It may also be ‘short-circuited’ and indicated processes may be very fast on the geological time scale, or, more often as not, occupy very lengthy periods, amounting in some cases to billions of years. Although few authors have said so in as many words, coordination compounds and complex ions are at the heart of every stage of the cycle. These include minerals, species in fluids ranging from aqueous solutions to silicate melts, complexes in sediments and asphalts, and chelates intimately involved in life processes and their decomposition products. This review addresses itself to a description of coordination geochemistry in the geochemical cycle and to related processes in the primitive Earth. We do not pretend that such an immense subject can be exhaustively covered in a short chapter. Rather, attention is given to general concepts, principles and their possible applications. Nor is the list of references by any means exhaustive. However, it does cover the main points and provides an entry to the voluminous literature of the subject. It should be added that many of the possible roles that coordination complexes play in nature are based on speculative interpretations of present day geochemical evidence; such is the nature of many geochemical problems, as many of the reactions are so slow that they do not lend themselves easily to rigorous experimental investigation. No doubt many discoveries will be made in the field of coordination geochemistry in the near future. This is only to be expected in the light of advances in techniques and instrumentation. 843

Geochemical and Prebiotic Systems

844 Remelting

I

METAMORPHIC

t

Metaror Dhos IS

-1

t

ORGANIC MATERIAL

ROCKS

I

Diagenesis

FROM BTOSPNERE

1

Figure 1 The geochemical cycle

Such findings are to be especially welcome, because without a thorough understanding of the coordination chemistry of inorganic and organic ligands, we cannot begin to understand any of the transformations occurring about us in the natural environment.

64.2 MINERALS 64.2.1 Minerals as Coordination Complexes Nearly 4000 discrete minerals have been characterized. Except for a few such as the native elements, purely organic species, and simple inorganic substances such as ice, water and carbon dioxide, they are all coordination compounds of one sort or another. It is inappropriate in a review of this size to examine many individual minerals in detail; rather the aim here is to outline the scope of coordination in naturally occurring solid phases. Several excellent compilations of minerals, their compositions, properties and occurrences have been published?-7 New and discredited species also appear in lists in Mineralogicul Abstracts' (quarterly) and in the Mineralogical Magazine' (biannually). The Commission on New Minerals and Mineral Names of the International Mineralogical Association gives approval for each new species gazetted. In these natural coordination compounds many different donor atoms are involved in bonding to metal ions. Some of these are gathered together in Table 1. Given this ligand set together with all of the combinations afforded by the elements of the periodic table, it is perhaps none too surprising that the number of known minerals is large and that a sizeable number of new ones, perhaps amounting to 100-200, are currently reported each year. The observed diversity is compounded by the many different kinds of structural arrangements possible for the theoretical combinations and by the diversity of new ligands generated by the polymerization of such simple precursor ions as borate and silicate, Fortunately, systematic schemes for the description of polyborates have been proposed"." and the range of ligands constituting the complex silicates has been exhaustively catalogued12 in the light of the fact that silicate minerals make up the bulk of the Earth's crust. Over the last 50 years or so the advance of X-ray crystallography has allowed a detailed insight into the structures of many minerals. The number of known packing arrangements, the general characteristics of which have been outlined elsewhere,"-16 is multiplied in the mineral kingdom by the number of isomorphous analogues of species for which single-crystal X-ray studies have been carried out. Many simple minerals, especially simple salts like halite, NaCl, sulfides, sulfosalts and oxides, have structures based upon cubic or hexagonal closest-packed arrays of either cations or anions. Coordination geometries of metal ions in many of these kinds of minerals are thus confined to more or less regular octahedra and tetrahedra. The occupancy of the two types of sites is dictated by the stoichiometry of the mineral, the radius of the ions involved and their preferred coordination geometries. Coordination of cations in mineral species in terms of bonding and crystal field effects has been extensively Comprehensive lists of ionic radii relevant to cation coordination geometries in minerals have also been compiled.16321

Geochemical and Prebiotic Systems

845

Table 1 Some Donor Atoms and Ligands in Minerals Group

Donor

Ligund types

IV

VI

C Si N P As Sb Bi 0

VI I

S Se Te F

Carbide, CNSilicide Nitride, organics, SCNPhosphide Arsenide, diarsenide and mixed anions such as As;-, Ass2Antimonide and arsenic analogues Bismuthinide and arsenic analogues 02-,OH-, H,O, borates and polyborates, silicate and polysilicates, borosilicates and AsO:-, antimonates, antimonites, arsenites, aluminosilicates, columbates, tantalates, PO:-, CO?-, NO,-, selenate, germanates, germanites, VO;-, CrO:-, Moo,'-, WOd2-, SO:-, selenite, tellurate, tellurite, iodate, organic hydroxyl and carboxylate Sulfide, disulfide and sulfoanions of group V and of Se and Te Selenide analogues of S ligands Telluride analogues of S ligands, Te2F- and complex fluorides such as BF,-, SiF,'-, A1F:CI- and complex chlorides such as FeCk4Br-

V

c1 Br I

I-

In non-close-packed structures, which indeed account for the bulk of known minerals, higher coordination numbers for cations are frequently encountered. Some specific examples are mentioned in the next section, but it is appropriate here to focus attention briefly on silicate minerals in view of their widespread occurrence and importance as rock-forming minerals. For example, in zircon, ZrSi04, the ZrIVion is eight-coordinate as is the Mg" ion in pyrope garnet, Mg3A12Si30,2. A number of cations are present in the amphibole class of silicates with the formula A,_,B2Y5Z8022(OH,F, Cl), ( A = Ca, Na, K; B = Ca, Fe", Mg, Mn", Li, Na; Y = Al, Cr, Fe, Mg, Mn", Ti; Z = Al, Si, Ti). In the micas and in leucite, KAlSi2O6, twelve-coordinate cations are present (K+ in the latter), and the K+ cation is nine-coordinate in the common feldspar orthoclase, KA1Si308. While the general formula of the amphiboles given above seems at first somewhat unwieldy, it is simple to rationalize this structure, and those of the other complex silicates, as consisting of various polymeric groups of Si0;- tetrahedra linked in different ways via bridging apical oxygen atoms. In a number of silicates these tetrahedra are replaced by those of AlO;-, which are generally larger and cause a distortion of the basic framework, and so compositional variations arise because of the extra cations needed in the lattice to ensure electrical neutrality. In the amphiboles, the SO,"- tetrahedra are linked by two or three bridging oxygen atoms to give double chains as shown in (1) with the composition (Si,O,,),6"-. The pyroxenes are based on the simple single-chain structure (Si03)?"- (2), although the chains are kinked as a result of slight rotations of the monomeric units. If each tetrahedral unit shares three corners to form infinite twodimensional sheets, the basic framework of the mica group results, with the formula (Si2O5)?-. When all four oxygen atoms in the monomer are involved in bridging, a three-dimensional framework is formed. Several polymorphs with the basic stoichiometry SiOz are known. Replacement of the Si ions by At"' and occupation of interstices thus formed by other cations yields the basic structure of the feldspars. Thus the variety of coordination numbers and geometries in an important class of minerals, the silicates, can be viewed as arising from the disposition of the ligand network formed simply by the regular crosslinking of SiO2- and A102- tetrahedra in a restricted number of ways. Examination of the garnet structure indicates that it is derived from a related linking of silicate tetrahedra and MO, octahedra where M is Al"', Fe"' or Cr"'.

Geochemical and Prebiotic Systems

846

While such considerations allow the elaboration of many mineral structures, these ideas are comprehensively treated in the main references given above. Particular aspects illustrating the diversity of coordination modes and complexes in minerals of a more specialized kind are set out in the next section.

64.2.2 Specific Mineralogical Examples of Coordination Compounds The variety of coordination chemistry in minerals is to some extent exemplified by naturally occurring halide species. Indeed, compounds identical to and related to several classic synthetic substances are known. To illustrate this point, attention is first drawn to fluorides and hydroxyfluorides of A1r",2-5 some of which are listed in Table 2 together with related Mn", Fe", Fe"' and Larr1minerals of interest. The occurrences of the AI"' fluorides in several pegmatites, the most famous being that at Ivigtut, Greenland, have been compiled by Raade and Hang.31 Table 2 Selected Complex Halide Minerals of AI"', Fe", Fe"', La"' and Mn" Mineral

Cryolite Elpasoli te Pachnolite Thomsenolite Weberite Cryolithionite Usovite Gearksutite Prosopite Chiolite Tikhonenkovite Ralstonite Gagarinite Rinneite Douglasi te Erythrasiderite Chlorrnanganokalite

Rep

&omposition

22,23 23 24

24 25 26 I

27 28

NaK,FeCi6 K2FeC1,.2H,O K,FeCI,.H,O K,MnCl,

29 30

-

In general see refs. 2-5. Selected crystallographic and analytical references are given where appropriate. Ln = lanthanide(lI1) or Y"' ion.

All of these AI"'-containing minerals have that ion hexacoordinated, but a variety of structures are known. Cryolite, elpasolite, pachnolite and thomsenolite all contain isolated A1F;- octahedra. The other aluminum fluorides and hydroxyfluorides shown in Table 2 have these units, variously substituted by hydroxide ions, linked together by the sharing of corners or edges in a fashion analogous to the linking of SO, and BO3 units to form the basic frameworks of the silicates and borates. Cryolithionite thus has the garnet structure and in prosopite the octahedra share two edges, with bridging being effected by hydroxyl ions and the free vertices being taken by fluorides. It is worth noting here that the alkali and alkaline earth cations in these minerals sometimes have unusual coordination sphere geometries. Two crystallographically distinct Na+ sites occur in cryolite, one octahedral and the other a somewhat distorted cubic antiprism. In weberite the two sodium atoms are also found in crystallographically independent sites, but are both octacoordinated. One lies at the centre of a flattened hexagonal bipyramid and the other in a rectangular prismatic polyhedron. Phase equilibria and conditions of formation of many of these species are especially in the light of their relevance to the refining of aluminum. Gagarinite, NaCaLnF,, apparently contains isolated LnF2- octahedra, but its crystal structure is unknown. Hexachloromanganate(I1) and ferrate( 11) ions are present in the minerals chlormanganokalite and rinneite, respectively, and iron(1I) or iron(II1) ions with four and five chloride ions in the coordination sphere are represented by the species douglasite and erythrosiderite. The coordination spheres are completed by water molecules, and these chloro complexes correspond to well-known compounds prepared in the laboratory? Erythrosiderite, its ammonium analogue, kremersite, and chlorrnanganokalite are all very deliquescent, a fact in accord with their occurrence around fumaroles and in ejecta at the volcanoes of Vesuvius and Mt. Etna. The former mineral was found associated with molysite, FeCl,, which

Geochemical and Prebiotic Systems

847

is naturally unstable out of this kind of environment. Rinneite has also been reported from Ve~uvius.~ These somewhat bizarre occurrences have also yielded' several salts of hexafluorosilicate(V1) and tetrafluoroborate(II1). The known minerals of this cIass are listed in Table 3. Heiratite was first found in a similar volcanic sublimate at Vulcano, and the dimorphs of (NH4)'SiF6 have also been reported as sublimates above a burning coal seam at the Bararee colliery at Barari, India. Finally, in this connection, two other simple coordination complexes are known from similar environments at Vesuvius. These are mitscherlichite, K2CuC14.2H20,identical with the artificial salt,37 and pseudocotunnite, K2PbC14, although this latter species might be KPbC13*1/ 3 H20.3 Table 3 Naturally Occurring Hexafiuorosilicates and Tetrafluoroborates Mineral

Composition

Re$

Avogadrite Ferruccite Hieratite Cryptohalite Bararite Malladrite

(K,Cs)BF4

2,34,35 2,35 2,35,36 2 2 2,34,35

NaBF, KzSiF, (NH,),SiF, (NH,),SiF, Na,SiF,

Three other chlorides which illustrate the range of complexity of this class of minerals, varying from the simplest to the most exotic of stoichiometries, should be mentioned. Albritt~nite,~' CoCI2.6H20,contains the trans-dichlorotetraaquacobaIt(I1) species (3). The same coordination geometry obtains in the analogous unnamed Ni" mineral? Metal cluster complexes are unusual in mineralogy but a notable example is found3' in the rare oxychloride boliite, Pb26C~24AggC162(0H)47.H20. In this cubic azure-blue species, six of the silver atoms are arranged in a regular octahedron at the centre of the unit cell. Eight chlorine atoms lie above the faces of this octahedron and six others are positioned normally to the apices as shown in (4). This configuration is close1 related to those found in such simple synthetic metal cluster compounds as MoC1, and WCI,. 4 J

Cl

.:Ag o:CI

Two classic pseudohalide coordination complexes are also known to occur naturally. Julienite, Na,[Co( NCS)4].8H20, is identical to the well-known synthetic compound4' and has the thiocyanate ligands bound to the cobalt(I1) ion tetrahedrally via the nitrogen atoms. The blue complex was originally reported as occurring as needle-shaped crystals in a cobaltian wad from Kantanga, Zaire.' The familiar salt &[Fe(CN)6].3H20 is polymorphous and crystallizes in the monoclinic, orthorhombic and tetragonal systems. Square and rectangular plates together with crystals of a monoclinic habit, with a similar formula, have been reported from a number of Soviet gold mines, and have been given the name kefehydrocyanite.42However, the single analysis reported indicates that the mineral is a monohydrate. The cyanide ligands were presumably derived from decomposing organic matter and not from leach liquors used in the beneficiation of the ore. Thus, the occurrence is of some significance with respect to the mobilization and enrichment of gold in oxidized ore deposits through the formation of cyano complexes (vide infra). Simple nitrogen-containing ligands are uncommon in mineralogy. A series of species which are derived from or analogous to Millon's base are however known.A3In these minerals the coordination backbone polymer [Hg,N],"+ has large cavities which can be occupied by a variety of anions, and end members can be obtained by ion exchange techniques. Kleinite contains both sulfate

848

Geochemical and Prebiotic Systems Table 4 Selected Carbonate Minerals Mineral

Liebigite Andersonite Grimselite Schroekingerite Zellerite Metszellerite Lanthanite Calkinsite Lokkaite Ancylite Burbankite Ewaldite Weloganite Chalconatrite Azurite Hell yerite

Composition

Ca2UO2(CO,),.1 1H,O Na,CaUO,( CO3),5-6H,O K3NaU02(C03),.H,0 NaCa,UO,(CO,),(SO,)F- 10H,O CaUO2(CO,),-5H,O CaU02(C0,),+3H,0 LaZ(CO3),.9H2O (La, Ce)ACOd3.4H2O (Yb, Ca)2(C0,),. 1.58H20 (La, Ce) x(Ca, Sr)z-l(C0,)2(OH),.(2-x)Hz0 (Na, Mz+, Ln3c)h(C03)s Ba(Ca,M,-,)(CO,), (Sr, Ca)ZrNa,(C03),.3H,0 NazCu(CO3),~3H,O CU~(CO~AOH)Z NiC0,.6Hz0

Ret 49 50 51 52 3 3 53 3 54 55

3 56 57 59 2 60

and chloride ions and has a structure similar to that of high-temperature t r i d ~ m i t e .Some ~~ non-essential water of crystallization is also incorporated in the channels. Mosesite, prepared by reacting HgClz with dilute aqueous ammonia, has the cubic cristobalite structure, space group F33m, but natural material contains sulfate, molybdate and carbonate ions together with the essential chloride4' and is a monohydrate. An anhydrous sulfate analogue, gianellaite, has also ~ doubt has been cast been recently as has the simple Hg" amidonitrate ~ a l t . 4Some upon the provenance of this latter mineral, found in an abandoned mine in Colorado, USA, and it may have been formed as a result of the kind of explosives using ammonium nitrate used in the workings. It has also been suggested that the classic coordination compound copper(I1) acetate monohydrate, reported48 as an alteration product of native copper from the Onganja mine, Namibia, may be artificial as well, produced as a result of specimen cleaning. This is reminiscent of the chloroacetate calclactite,' CaCI(CzH,02)-5H,0, found as an efflorescence on calcareous rocks, fossils and pottery stored in oak museum cases. With respect to unusual coordination geometries and the preservation of commonly characterized solution species in mineral lattices, together with the occurrence of coordination compounds frequently encountered in the laboratory, several minerals containing oxygen donors ought to be mentioned. A few carbonates are especially noteworthy. These are collected together in Table 4. The four uranyl carbonates liebigite, andersonite, grimselite and schroekingerite have been shown to contain the U02(C03):- anion, which adopts the expected hexagonal bipyramid geometry. This is also most probably the case4' in the related minerals bayleyite, Mg2U02(C03)3.18Hz0, swartzite, CaMgUOz(C03)~12Hz0, and several others. The tris(carbonat0) complex is a wellknown aqueous species and is thought to be important in the dispersion of uranium in alkaline groundwaters. Similarly, the bis(carbonat0) complex, U02(C03)22-,is probably preserved in the structures of zellerite and metazellerite. High coordination numbers are also frequently encountered among complexes of the rare earths, zirconium and hafnium. A few relevant minerals are also listed in Table 4. In lanthanite, each of the two independent rare earth ions i6 ten-coordinate. Both coordination spheres may be viewed as distorted Archimedean antiprisms whost square faces are replaced by prisms. Only one coordination site for the metal ions is present in ancylite, due to the fact that all the cations are disordered. The geometry about the metals might be described as a distorted hexagonal bipyramid, one of whose apices is replaced by a triangle of oxygen atoms from two of the carbonate ions and one hydroxyl ion. Ewaldite has been described as a barium calcium carbonate, but the calcium site is occupied by a variety of cations, including the rare earths. Here, the cation is six-coordinated with three oxygen atoms from two carbonate ions lying below it. The coordination sphere is completed by three other oxygen atoms from distinct carbonate groups forming an equilateral triangle above the metal ion. A review of hydroxycarbonates of tha lanthanide elementss8 covering much o f the earlier literature has appeared. It has been shown that weloganite crystallizes as discrete enantiomorphs in space group P1 and exhibits a remarkable variety of coordination geometries in its structure. Carbonate ions and water molecules link the six cations such that the ZrIV ion is nine-coordinate with three oxygen donors each lying above, below, and in the plane with the metal ion. The

Geochemical and Prebiotic Systems

849

arrangement has but slight distortions from D3,,symmetry. The three Sr" ions have independent coordination spheres, but are closely related in spatial terms. The ten oxygen atoms which surround are contained approximately within a hexagonal bipyramid, one of whose apices is replaced by a triangular arrangement of donors. The two sodium ions lie in quite dissimilar environments. Both are six-coordinated, but one has slightly distorted octahedral symmetry and the other lies in the centre of a trigonal prism. This structural extravagance is echoed by many other minerals in the general literature. isolated as a corrosion product from ancient bronze objects has been A copper( 11) reported. It is chalconatrite, Na2Cu(C03)2.3H20,and is identical to the known artificial compound. The origin o f its greenish blue colour is due to the solid-state analogue of the very stable coordination complex ion Cu(CO&-(aq), which also imparts the characteristic and beautiful colour to the common secondary copper mineral azurite, Cu,(CO,),(OH),. Other carbonates of the transition elements are somewhat simpler and hellyerite, for example, contains the ordinary hexaaquanickel( 11) ionfi0Such regular or near-regular octahedral coordination is common in oxides and hydroxides as well as simple aquated metal salts. These include sulfates such as the hexahydrites, MS04-6Hz0(M = Mg, Zn, Fe, C o ) and related species,6l and somewhat surprisingly, mohrite:' the naturally occurring analogue of Mohr's salt, (NH4)2Fe(S04)2.6H20. Hexahydroxo complexes of SnIVand Ge'v?363*64 such as wickmanite and tetrawickmanite, MnSn( OH), ,vismirnovite, ZnSn(OH),, stottite, &%%(OH),, and the unnamed complex Zno.5Feo.5Ge(OH)6 and its Mn" analogue, might be thought to belong to this class of minerals as well, although it is unlikely that these species contain the isolated (Sn, Ge)(OH)t- ion. Coordination complexes of organic ligands are widespread in nature, and these are discussed in some detail below. However, but few isolated compounds have been described definitively as mineral species. The same must be said for the many potential organic ligands known to be present in a wide variety of Earth materials. Indeed many of the organic compounds and salts listed in standard reference are ill defined, and a number of them may well be discredited as minerals in the future. Some quite exotic potential ligands are known, including guanine and uric acid (uricite)? and phthalimide (kladnoite),66 but their natural coordination chemistry is largely unknown. A number of poorly characterized complexes of humic and fulvic acids (vide infra) have been reported, such as crenite,' the yellow-coloured materia1 in stalactitic calcite crystals from the Harz mountains in Germany. It is probably the calcium salt of a fulvic acid. Similarly, pigotite2 is an aluminum salt of a related ligand, and volzite is an analogous Zn" complex, but type material was later to be admixed with wurzite, the high-temperature dimorph of ZnS. Humates in peat, lignites, shales and other sedimentary rocks are invariably complex assemblages of many related ligands. The name sogrenite has been applied68to one such mixture containing a considerable quantity of complexed uranyl ion, and a mixed Ca", Mg", Al"' and Fe" (together with minor amounts of many other elements identified spectrographically) salt of a humic acid, called dopplerite, has been described from a number of 10calities.6~'~~ Other kinds of coordination compounds involving organic ligands are much better understood. One such class is that of metal complexes of oxalic acid.* Calcium oxalate mono- and di-hydrate, whewellite and weddellite respectively, are generally of organic and indeed biogenic origin, but the former has also been noted as a primary mineral in hydrothermal ore deposits. These salts are the main phases present in oxalate-containing human urinary c a l c ~ l i . ~Another ' natural coordination backbone polymer is humboltine, the Fe" oxalate Fe(C20,).2H20.72Bridging oxalate groups link the Fe" ions in single crystals along the b axis to form infinite chains. Other complex oxalates are instantly recognized as coordination compounds. Minguzzite i s the potassium salt of the three-bladed propeller tris( oxalato)ferrate(lII), K3[Fe(C204)3].3H20.73 This complex anion is preserved in the sodium magnesium octa- or nona-hydrate mineral ~tepanovite'~ and the corresponding Al"' complex in the same lattice has been given the name zhemchuznikovite. It is appropriate to include two other complexes of oxyacids here, since both are fully characterized. The first is mellite,75A12(C12012). 18Hz0,the Al"' complex ofmellitic acid (5). This honey-coloured mineral has been known for nearly two centuries' and occurs in a number of European and Soviet identical to the artificial salt lignite deposits. The second is a calcium complex of citric Ca3(C6H,0,),~4H,0. This complex, earlandite, was found in sediments dredged from the bottom of the Weddell Sea, Antarctica, together with weddellite (vide supra). No doubt many other naturally occurring salts of organic acids abound and some of these are mentioned in a later section. However, no others have ever been described as minerals. This is in the most part due to the fact that they are difficult to isolate intact from the bitumens etc. in which they occur and

850

Geochemical and Prebiotic Syslems

(6)

that they have not yet been observed as distinct crystals, always being found as complex mixtures with a host of other organic (and perhaps organometallic, in its widest sense) species. Porphyrins and their complexes are ubiquitous in nature, and are discussed in detail in a later section. Only one, abelsonite, has been accorded a separate mineral name. Found as aggregates of purple plates in drill core from the Green River formation, Utah, USA, it proved to be a Ni" complex" with a formula closely approaching C3,H32N4Ni.Spectral studies indicated that the mineral was a deoxophylloerythroetioporphyrin based on (6), and was presumably derived from chlorophyll. However, the natural material probably contained water molecules completing the octahedral Ni" Coordination sphere. The colours and red fluorescence of quincite, a variety of sepiolite, Mg4Si,01,(OH),~6H,0, from Quincy, France, and of red calcites from Austria, have been as~ribed'"'~to porphyrins or their Gar'' and Fe"' complexes. More recent work" on calcite from Deutsch-Altenburg, Austria, has shown however that the colour in this calcite is due to finely dispersed iron(II1) oxides and oxyhydroxides, and the fluorescence to the presence of icosyl alcohol. Other workers" have also found that of the several organic components which can be extracted from quincite, none is a porphyrin. The origin of the colour remains unknown. Limitations of space preclude a comprehensive survey of the coordination chemistry of minerals. The reader is referred to the general references in this section for more information. Naturally, many other examples might have been chosen to exemplify various coordination numbers, geometries and complexes found in minerals. Nevertheless, the various species mentioned above do indicate the wide range of coordination chemistry in the solid state found in the mineral kingdom and the complexity of coordination compounds in the natural environment.

64.3 INORGANIC COMPLEXES IN SOLUTION that the formation of coordination complexes plays an important It has long been role in the geochemistry of natural aqueous solutions. Application of such reactions to problems concerning the deposition of certain kinds of ores was taken up by K r a u ~ k o p f . Many ' ~ ~ ~of ~ ~the ~ applications of complex formation in solution geochemistry and mineral formation were considered in a now classic text by Garrels and Christ.86More recently this material has been expanded and built uponby Stumm and Morgan? who also havecollected the basicbibliographyofthe subject. The application of coordination chemistry is central to the understanding of aqueous geochemical cycles and is more far-reaching than might initially be thought. In this section, the coordination complexes of simple inorganic ligands are discussed. However, it should be remembered that species containing organic ligands can be at least as important or even more important in articular instances. This area is dealt with in a later section. Moreover, adsorption phenomena' may be viewed in part as coordination processes. Chemical modelling of the distribution of various metals in natural waters as well as analysis of metals associated with particulate matter in such systems indicates that adsorption is sometimes the most significant control with respect to species distributions. Effects have been examined for a number of transition metal ions, lead", zinc" and c a d m i ~ r n " . ~This ~ - ~area ~ of geochemistry is, however, outside the scope of this review and the interested reader is directed to the above-mentioned work and other references therein. Coordination complexes in silicate melts and gases of natural origin are examined in Section 64.4.

64.3.1

Aqueous Solutions at Low Temperatures

Aqueous coordination complexes in natural systems which involve most of the ligands listed in Table 1, as well as many others, are known. For solutions at around room temperature, the

Geochemical and Prebiotic Systems

851

general chemistry of complex formation is well u n d e r ~ t o o d ~ and ~ - ' ~comprehensive compilations of stability constants for hydrolyzed species and simple inorganic species are a ~ a i l a b l e . ~ ~ - ' ~ Accordingly, many workers have used this data to examine species distributions of metal ions and complexes in freshwater and seawater. A recent review" has tabulated many of the results in terms of periodic trends and lists most of the relevant earlier work in this field. Table 5 Complexes of Major Dissolved Species in Seawater

Ion

Significant aqueous complexes

CaSO?, CaHCO,*, C a C O t MgSO,', MgHC03+,MgCO: NaS0,-, NaHCO," K.30,-

Ca2+

M g2'Na k+ +

Table 5 lists the important complex species, aside from free aquated metal ions, of the major ions in seawater, and Table 6 gives similar information for first-row transition and other selected elements which are present in much lower concentrations.' Speciation patterns for other elements are listed in an earlier referenceY8 It is clear that only a few types of complexes are formed under these conditions. This is a natural consequence of the concentration and availability ofthe common OH-, SO4'-, C03'-) and the low concentrations of most metal ions in natural ligands (Cl-, H20, aqueous solutions. Even for these simple species, however, complex geochemical interactions are indicated. For instance, it has been that MOH+ (M= Cu, Zn) complexes are those species adsorbed on geothite, a-FeOOH, and that MCl+ (M = Pb, Cu, Zn) species are preferentially adsorbed on the same phase from chloride-containing solutions. Implications arising from such findings for the accumulation of elements in deep-sea manganese nodules and related media are obvious. Table 6 Complexes of Selected Trace Elements in Freshwater and Seawater

Metal ion

Fresh water"

Significant aqueous complexes Seawnlerb

TiO(OH)', TiO(0H): HZVO,-, H3V0,' Cr(Oy),+ CrO, ,HCrO,MnSO4' FeS0,O Fe(0H); cos0,o NiSO? CuCP CuOH+, CuSO,', CuHCO,+, cuco,o, CU(OH),'

TiO(OH)+,TiO(OH)20 HV04'-, H2V04-, NaHV0,Cr(OH),+ CrO?-, NaCrO, ,KCr0,MnCI+, MnCI;, MnHCO,+, MnSO? FeOH+, FeCl+, FeSO,O Fe(0H): CoOH+, Co(OH);, COCO,^, CoCI+, CoSO? As for co" CUCI", cuc1,As for freshwater pIus CuCl', Cu(C0,);-

98

Zn"

ZnSO,O

98,104

Pb"

PbOH+, PbC03", PbHC03+, PbSO,' MgOH+, Hg(OH):, HgCI+, HgCl,", HgCI,-, HgC1,'-

ZnOH+, Zn(OH):, ZnHCO,+, ZnCO3O, ZnSO,", ZnCI+, ZnCI,O, ZnCI, -,ZnC1,'As for freshwater plus Pb{S0,)2z-, PbCl;, PbCI,-, PbCI,'Chloride complexes as for freshwater

Ti'" VV

Ct"' Cr"' Mn" Fe" Fe"' co" Ni" CUI cu"

Hg"

98 98,99

98 98,100,101

98,102 9s 98,103 98 98 98,104

98,104 98

" p H 6 . b p H 8.2.

The role of the simple inorganic species listed in Table 6 in the formation of orebodies from aqueous solution has been reviewed.'06 Bauxite, copper ores and iron and manganese deposits were examined. Particular attention was paid to coordination complexes in slightly alkaline groundwaters. The coordination chemistry of gold in the natural environment, especially with respect to its migration and enrichment in certain oxidized sutfide orebodies, provides another excellent example of the variety of ligands which may be involved in complexes of geochemical interest. This has been the subject of an excellent recent reviewlo7and has attracted the attention of geochemists for several decades.'" Gold has been observed to migrate in groundwaters in the oxide and gossan zones of auriferous sulfide deposits, and to reprecipitate nearer the water table. CCC6-BB

852

Geochemical and Prebiotic Systems

Similar behaviour is also displayed by copper and silver, the former being responsible for large economic concentrations o f that element particularly associated with weathered porphyry and other ores. Since gold is the most inert of elements, its passage into and transport in aqueous solution must be due to oxidation in the presence of suitable ligands. For many years the chloride complex AuC14-, formed in chloride solutions by reaction of gold with, say, MnO,(s), was thought to be the most likely complex. Recent work however has indicatedlo7that AuBr,-, Au(CNS)~-, Au12, Au(CN)?- and Au(S,O,),3- could all be implicated. The mixed ligand complex AuClBr-, which is one of the main forms of gold in seawaterlogtogether with AuCI,-, may also be important. Thiocyanate has a negligible concentration in soils, but cyanide from, for example, decomposing plant glycosides could account for many of the observations. The bis(thiosu1fato) complex may prove to be the most important since thiosulfate can be generated as an intermediate in the oxidation of sulfides such as pyrite, FeS,, themselves in close association with native gold particles. Excluding brines of various kinds, which are dealt with in the next section, groundwaters may have somewhat different concentrations of inorganic ligands than surface waters and processes involving complex formation and speciation are thus altered. In general, the concentration of dissolved silicic acid, H4Si0:, is higher in these environments and several workers have examined complex formation with it. Major silicate species and complexes present in natural waters are H4Si0:, H3Si04-, H2Si0:-, MgH,SiO:, MgH,SiO,+ and their Ca" analogues.98 Other minor species which may have some geochemical significance especially in the formation of deep-sea nodules are FeH,SiO;' and MH,SiO,- (M = Co"', Fe"').110311'Other complexes involving polymerized silicate and aluminosilicate112species may also be important in groundwaters, but little is known of their true significance. The geochemistry of uranium in natural waters has been thoroughly reviewed.1133114 The species UO,HSiO,- may influence the mobility of the uranyl ion under some circumstances, but the most significant complexes of this cation under most geochemical conditions are U02CO:, U02(C03)$-, U03(C03)34-,an extremely stable complex indeed, and one mentioned extensively in the literature, and U 0 2 (HP04);-. At low pH and under suitable redox conditions, U'" fluorides can account for significant amounts of uranium present in solution, and Uvspecies may also be involved. A review of thorium data115indicates a number of similarities to uranyl complexes with respect to geochemical behaviour. Inorganic complexes of geochemical interest are Th(S04);, ThF:,' Th(HP04);, ? ~ I ( H P O ~ ) ~and , - Th(0H):. A final example concerns the formation o f heteropolynuclear hydroxide complexes.116The complexes [(OH)Fe(0H),Crl3+, [(OH)Fe(OH),Cr(OH)]" and [(Ow),Fe(OH),Cr(OH)]', or polymers such as Fe(OH),M,"+ (M = V, Cr, Mn, Co, Ni, Cu; n = 2-4) have been studied with a view to an understanding of the inclusion of transition metals in iron ores as mixed oxides rather than their occurrence as discrete mineral phases. Many other examples might have been chosen in this section. Reference should be made to the general reviews given above. However it should be clear that simple inorganic coordination Complexes play a major role in the chemistry of natural aqueous systems at low temperatures.

64.3.2 Coordination Chemistry of Brines Inorganic coordination chemistry of very saline fluids has received considerable attention with respect to the transport of metals in groundwaters and hydrothermal systems and the emplacement of a number of different kinds o f orebodies. Two lines of evidence are of particular importance, and have both been set out in considerable detail in a series of articles by several authors in a recent work.''' The first of these concerns the nature of fluid inclusions in various ore and gangue minerals from numerous sulfide and oxide ore deposit^."^*"^ In general the contents of these inclusions are very saline with chloride molalities up to the order o f ca. 5 molal and the major balancing cations are Na+, Ca2+,Mg2+and K+. Many fluid inclusions also contain high concentrations of CO, liquid or gas and varying amounts of other ligating components. These observations provide compelling evidence that metals in many deposits were transported as inorganic complexes. Some inclusions contain significant amounts of base r n e t a l ~ . " ~ -Equally '~~ as striking as these observations is the occurrence of hot metalliferous brines associated with continental and ocean deeps, and which are comparable in composition to many fluid inclusion^."^ Particular examples occur in the Red Sea area, the Caspian Sea, the Salton Sea, along mid-ocean ridges and in association with continental oil field salt domes and deposits. The classification and role of systems such as this in metal transport have been outlined by White.',' Other ligands which can also be important in these brines are F-,Br-, I-, NH3 and sulfur-containing species. A recent provocative review'*' suggests that organic ligands may also be involved in certain brines in the emplacement

Geochemical and Prebiotic Systems

853

of Mississippi-type deposits, but their significance has not been fully assessed. Bubble-filling temperatures of fluid inclusions indicate that many sulfides were deposited at temperatures ranging from about 50 to a few hundred degrees. Sulfides associated with geothermal brines such as pyrite, FeS, , chalcopyrite, CuFeS,, galena, PbS, and blende, ZnS, provide indisputable evidence that circulating saline waters are intimately involved in mineralizing processes. A more spectacular example in this respect is the SaIton Sea geothermal brine from which the somewhat rarer species chalcocite, Cu2S, bornite, CuSFeS4, tetrahedrite, (Cu, Fe),,Sb,S,, , stromeyerite, AgCuS, and native silver among others, have been observed to ~ r y s ta lliz e .'~Indeed ~ primary chloride mineral formation has been noted from a few l ~ c a l i t i e s . ' ~ ~ J ~ ~ The solubilities of a number of sulfides in aqueous brines have been the subject of much research and a considerable amount of experimental data has a c c u m ~ l a t e d . ~ Such l ~ ~ ~ data ~" coupled with modelling experiments have led to an appreciation of the inorganic coordination chemistry of many elements in metalliferous brines. Not surprisingly, a number of species of no significance in surface waters are now thought to be important in these conditions. Some of these involving halide, carbonate and ammonia ligands are given in Table 7. Not included are those species listed for seawater in Table 6, to avoid duplication. No doubt many others could be equally as important. The list is not intended to be comprehensive, but to indicate the range of complexes of geochemical interest in this connection. It is also difficult to distinguish between the significance of these species with respect to other complexes known to exist at higher temperatures, and which are discussed in the next section. Table 7 Additional Selected Inorganic Complexes of Possible Significance in Metalliferous Brines -

Metal ion Pb"

Ag' CUI cu" Be" Hg"

Complexes

Ref:

Pb(C03):-, Pb(HCO,):, Pb(HCO,),-, Pb2(C03)&-, Pb(OH)BrCl;-, Pb(OH)Br,C12-, Pb,(OH),Cl:, Pb8(0H)&1~, NaPbCI,', NaPbCl,-, PbBr+, PbBr;, PbNH32+,Pb(NHJ2'+ AgCl', AgCI,-, AgCI,'-, AgC12cuc1,zCU(NH~),'+ BeF.,Hg1,'-

122,127,128

117,129 130 131 132 133

Finally, a separate discussion of sulfide complexes is warranted because of their suggested role"' in metal transport similar to that outlined above. Attention was drawn to such species not only because most economic ore deposits of transition and related metals are of the sulfide type, but because sulfide and its congeners occur in brines as potential ligands and therefore sulfide is transported together with metals in some hydrothermal solutions, and because of the striking association between elements, sulfides and sulfosalts of group V in many ores. Such complexes are naturally of great significance in solutions at higher temperatures. Many of the species implicated are given in TabIe 8, although recent examinations of several lines of geochemical and geological evidence tend to suggest that Pb" complexes of this sort probably do not reach ap reciable concentrations in hydrothermal s y ~ t e r n s . ' ~ ' , ' ~ Mixed ~ , ~ ~ complexes ~ of sulfide and As or Sb"' may contribute to the transport of gold in alkaline sulfur-rich brines. Other sulfur

E'

Table 8 Some Geologically Important Complexes of Aqueous Sulfide

Metal ion

Complexes

Additional ref:

134 135 136 -

854

Geochemical and Prebiotic Systems

ligands could provide alternative modes of complexation for many metals. Thiosulfate and polysulfide (Sn2 ; n = 2-6) ions have been implicated in coordination complexes of Cull, Cd" and The geochemical Pb" in a number of sulfide-rich waters. Polyselenide is also pre~ent.'~''~~' application of polysulfide complex formation has been as~essed'~'with a view to an explanation of pyritization, gangue mineral formation and wall-rock alteration in base-metal orebodies. Recently, the presence of Cu" and Fe" complexes of a number of reduced sulfur species in a tidal marsh has been rep~rted.'~'Under suitable reducing conditions in such an environment organic thiol and polysulfide ligands may also be formed. These are anticipated to be strongly involved in CUI' speciation in pore waters under such a regime.

64.3.3 Inorganic Complexes in High-temperature Aqueous Solutions High-temperature aqueous complex formation is of central importance to metal transport and deposition in hydrothermal systems and the like. Not surprisingly, much less data are available for complex formation in comparison to solutions involving complex systems near room temperature, since such mixtures of interest are experimentally much less accessible. Early work in this field was exhaustively reviewed by H e l g e ~ o n , 'and ~ ~ more recent summaries of the state of the art have appeared."7,'29 General approaches to prediction of coordination complexes formed and their behaviour have been outlined by these and other The work of Helgeson in this field is particularly n~teworthy.'~'In general, all of the complexes mentioned in the last two sections and their congeners might be important, although most recent studies'29 indicate that at high temperatures, of the order of 300-500 "C, the chemistry of complex equilibria may be much simplified compared with analogous multicomponent mixtures at low temperatures. It is only fair to indicate, however, that a much larger body of thermodynamic data needs to be accumulated before an adequate model for high-temperature speciation can be advanced. At very high temperatures, complex formation must only involve inorganic ligands, and only a few simple ones involving H2S, H20, COz and halides are reckoned to have any significance, although sulfates could play a part as well. Accordingly, attention has been focussed upon the coordination chemistry of these donors, particularly on halide and carbonate complexes in view of the fluid inclusion evidence. Besides species tabulated earlier, a number of other complexes have been claimed to be involved in metal transport under these conditions, and are listed for reference in Table 9. It seems clear that FeCl', FeCl: and FeOH+ are involved in the formation of high-temperature iron deposits. Carbonates, fluorides and chlorides are almost certainly associated with ore-forming fluids in high-temperature environments connected with pegmatites, greisens and skarns. Nevertheless, it must be said that all of these complexes, and no doubt many others, require considerably more research work to be carried out before their proper functions are completely understood uoder high-temperature geochemical conditions, and they can be incorporated into the speciation model^'^^^'^^ necessary for a rigorous description of hydrothermal ore deposition. Table 9 Possible Complexes of Importancea in High-temperature Mineralizing Solutions 117~120 Cnmpkxes

Ligand

c1FCO,z-

so:OHa

FeCl ', Au,CI,(HCI),,, LnCI'+, CuCI4'-, GaC1,Ln&, SnF,'-, W&-, AsF,'-, ZnF,'-, SeF:-, ZrF:-, NbF2Ln(C03)4s-, Th(CO3);-, UOz(C03)$- and related species, W"' and Be" carbonates, Sn(OH)$03h(S15;)2%(OH):, %(OH),-, Sn(0H);-

Additional ref

146- 149 149-1 $1 149,152-155 -

-

In addition to complexes listed in previous tables. Ln = rare earths.

64.4 GASEOUS TRANSPORT AND SILICATE MELTS By comparison with studies of complex formation in aqueous solution, the gaseous transport of coordination compounds in geochemical systems has received scant attention despite many pointers to the possibility, especially concerning volcanic emanations, arising from metallurgical processes. In Section 64.2 reference was made to a number of halides associated with fumarolic

Geochemical and Prebiotic Systems

855

activity at Vesuvius, Mt. Etna and Vulcano, Italy. Molysite, the very hygroscopic Fe"' chloride, was presumably transported as the dimer Fe2C16in the gas phase and a number of other halides, aside from the hexafluorosilicates and tetrafluoroborates transported as gaseous SiF4 and BF3, have been noted from the Italian deposits.2 These include the copper( 11) minerals eriochalcite, CuC12.2H20, melanothallite, CuCl(0H) (?), hydrocyanite, CuSO,, and dolerophane, Cu,OSO,. Gaseous SO3 is thought to contribute to the latter two species. Other simple halides known to have been transported in the gas phase are chloraluminite, AlCl,-6H20 (the Al,Cl, dimer possibly being important at higher temperatures), lawrencite, FeClz, scacchite, MnC1, , chlormagnesite, MgC12 and cotunnite, PbCI,. The transport of copper and lead in the vapour phase under such volcanic conditions is well documented. While the Cu3C1, trimer might be important for the former, abundant evidence for CuCl(g) has accrued. It has been observed spectrographically in volcanic flames at K i l a ~ e , and '~~ in high-temperature volcanic gases by other w ~ r k e r s . ' ~ ' The ~ ' ~ ~sublimation of CuCl from lavas leading to the crystallization of primary atacamite, Cu3(OH)3C1,upon condensation has also been reported.I6' Zinc may also be transported as the volatile chloride under the same conditions and has been noted in fumarolic emissions from the Showashinzan volcano, Japan.162 Several workers have speculated on the role that gaseous complexes might play in hydrothermal ore disposition under supercritical conditions or in high-temperature environments where water is absent. Early suggestions were summarized by Shcherbina,"' who considered that SiF,, BF3, TiC14, FeC13,ZrCl,, BeC12, TaClS,TaF, and NbF,, among simple halides, could all be involved in mineral formation from the gas phase at high temperatures. Other volatile oxyhalides such as NbOCl,, MoOCI, and ReOC1, were also considered. More recent work has focussed on the transport of tin'63 as volatile SnC14, rather than SnF,, in the formation of cassiterite (SnO,) deposits. General studies of high-temperature gas-phase metal transport with applications to geochemical processes of ore formation have been reported+'64Particular emphasis was placed on reactions of base metals and their common ore minerals but it is fair to say that the implications for geochemistry of gaseous complex formation are as yet imperfectly explored. The coordination chemistry of silicate and aluminosilicate melts is a particularly important subject in view of magmatic processes involved not only with ore genesis, but the much wider subject of the geochemistry of rock-forming processes in magmas of various sorts. The structures of mch melts are reasonably well understood at least qualitatively in terms of linking of silicate and aluminate tetrahedra but the general approach has been of a theoretical rather than an experimental nature. Generally, it is thought that melts have structures analogous to those found in silicate minerals and much is owed to studies involving glasses, particularly those using vibrational spectroscopic techniques. It is a fact that silicate melts are highly polymerized and this in turn is reflected by the high viscosity of such systems. A recent series of review^'^' has placed mony studies concerned with this phenomenon, and associated coordination chemistry, in perspective. Application of certain aspects of polymer theory indicates that species such as SO:-, Si207,--, Si3010p- and Si40,3'"- are important, together with others. Among these are the metasilicate rings Si3096- and Si4O1;-, which can polymerize in very siliceous melts to give cylindrical structures. This fact is reminiscent of the beryl structure. The silipon atom is always tetrahedrally coordinated. Aluminosilicate melts have much more complex structures, and the aluminum ion may be four-, five-, or six-coordinated. A considerable controversy has accompanied studies on these systems and a recent excellent review'66 has gathered much of the evidence and results concerning them. Central to our present understanding of melts are the concepts of 'network-forming' and 'network-modifying' constituent^.'^^,^^^ Aluminum and silicon, together with accessory phosphorus, belong to the former category and produce melts and glasses with a high degree of polymerization. On the other hand, the 'network modifiers', metallic cations usually, disrupt the basic structures and lower the viscosity of the melts. Concerning cation coordination in silicate melts, most work in the past has concentrated on partitioning of elements between the melt and various numbers of silicates in equilibrium with it. Burns" has analyzed these phenomena in detail in terms of crystal field stabilization energies pertaining to the solid phases. However, only a few studies on coordination sites in the melt have been reported and much of our knowledge of these systems is based on theoretical consideration of potential geometries developed in random rigid-sphere assemblies. The most important contribution in this area arises from the work of Bernai,I7' which was extended in a general way to mineral structures by Whittaker.141 It can be shown that in a random assemblage of spheres the average coordination number i s lower (typically eight or nine) than that in a close-packed array (12). Of relevance to the considerations of coordination geometry here is the kind of holes generated in these assemblages and their geometry. It was shown that

856

Geochemical land Prebiotic Systems

tetrahedral and half-octahedral holes amounted to some 90% of the resulting sites. Smaller numbers of trigonal prisms, Archimedean antiprisms and tetragonally distorted dodecahedra are also present. These results, as mentioned below, d o not fit exactly the experimental picture, as far as it is known, of cation coordination geometries in melts, but they do point to the fact that lower coordination numbers are preferred. This simple notion does fit with a number of experimental observations. In silicate melts approximating the olivine, (Fe, Mg)2Si04,stoichiometry, Ni" is both tetrahedrally and octahedrally ~ o o r d i n a t e d . ~The ' ~ disposition of coordination geometry is dependent upon the temperature of the melt as well as on its composition. Higher temperatures favour tetrahedral coordination as does, for example, changing the composition of nickel-containing glasses from those containing low molecular weight alkali metals to those containing potassium or rubidium. It should be pointed out here, however, that reservations need to be placed on the drawing of analogies from glasses to melts, Recent spectral studies show that appreciable coordination geometry changes accompany glass solidification even if quenching is quickly carried It is difficult, as well, to attempt to rationalize melt coordination chemistry in terms of only one kind of coordination site for a given element and for a given geometry. Studies involving Ni" in melts of the composition of the pyroxene diopside, CaMgSi,O,, indicate that at least two octahedral complexes must be present in order to explain electrochemical b e h a ~ i 0 u r . Similar l~~ results were found for Co" and Zn" under the same conditions. The coordination chemistry of iron in such melts is a little better understood. Fer'' and Fe" can act as network formers and modifiers in both tetrahedral and octahedral coordination g e ~ r n e t r i e s . ' ~It~is - ' also ~ ~ of interest to note in the light of the above results on Ni'l that multiple site occupancy is also indicated for Fe1".'77 In a number of CaO-A1203-Si02-Fe0 glasses two different tetrahedral complexes of this ion can be observed using Mossbauer spectroscopy. in Na,O-SiO,-FeO melts indicate that specific Fe"' anionic units are Speciation formed which involve Si2052pligands. This result, implying an alternative view of metal speciation in silicate melts, is a significant one. It implies that discrete species, which may lend themselves to modelling studies, are important, as opposed to the view of the coordination chemistry of melts being concerned with holes or sites in silicate liquids varying with changing composition. A particularly striking observation in this respect concerns a study of Mn"' in sodium silicate melt^."^ Analyses of electronic absorption and luminescence spectra indicated that Mn"' was present as a discrete complex and thus that the ion could not be viewed strictly as a network modifier. T h e consequences of this finding for crystal field interpretations of the crystallization of such ions from silicate melts were outlined. However, the paucity of detailed experimental data in these kinds of systems at the present time makes it difficult to assess the true role of individual coordination complexes in siliceous melts. This is an active field, nonetheless, and it is anticipated that a considerable understanding of this branch of coordination chemistry might be achieved in the next few years.

64.5 COORDINATION GEOCHEMISTRY OF COMPLEXES OF ORGANIC LIGANDS

As has been noted in a previous section, only a few complexes of organic ligands have been characterized as discrete mineral species. This fact is, however, quite out of proportion with respect to the importance of metal-organic interactions in the natural environment. Of course, many complexes formed are quite ephemeral in nature, and metabolic and ultimately diagenetic processes lead to their destruction as the material proceeds further in the geochemical cycle. Nevertheless, many species are robust enough to survive quite drastic treatment and they persist in geological units of great age. It is quite reasonable also to suggest that many more coordination complexes of organic ligands exist in waters, sediments and rocks, including fossil fuel materials, than have been isolated and characterized. This is in no small part due to the fact that they are frequently denatured during the experimental procedures adopted to isolate the organic molecular constituents. The association of metals with organic fractions of sediments and rocks of various types has long been appreciated.' Data have been tabulated and reviewed for shales and other sedimentary rocks,17971ao coa15,'81asphalts,'82 petroleum'"+184and humus oils.''^ An excellent review of much of this work and its setting in terms of organic coordination geochemistry is available,ls6 and covers most of the early work in this field. Some elemental associations are striking, particularly the occurrence of high concentrations of Ni and V in petroleum and Ga, G e and U in coals and peats. A major fraction of the metal contained in these materials is complexed to organic ligands.

Geochemical and Prebiotic Systems

857

Oxygen and nitrogen donors are probably the most important in this respect, although S , P and As ligands might also play a part. Naturally occurring organometallic compounds, including carbonyls, have received little attention and their significance is barely understood. As well as in these kinds of environments, metal complexation by organic ligands, particularly hi h molecular weight ones, in natural waters is of crucial importance in the geochemical cycle.'"J' In the next few sections the coordination geochemistry of such complexes and its significance is discussed. Divisions have been drawn between the various classes of important ligands known to be involved in such processes, although some overlap is found to occur, especially for the so-called 'humic' and 'fulvic' acids. Finally the potential importance of organic coordination complexes in ore deposition is assessed. Environmental coordination chemistry of ligands derived from mankind's activities is not considered here, although the interested reader is directed towards a number of papers presented at a recent conference on this topic.'89

64.5.1

Complexes of Humic and Fulvic Acids

Humic materials of one sort or another represent the vast bulk of the organic material containing reduced carbon in the Earth's crust and in natural waters. They occur in rather more condensed forms in coals, lignites, etc., but much material in sediments and waters, including interstitial pore waters, can be extracted into solution and has a molecular weight range of perhaps 2000 to 300000. The nature and importance of humic substances has been the subject of a recent authoritative review,'" The role and importance of coordination complex formation with these materials in the natural environment cannot be overstated, and has been discussed in detail for ~ ' ' ~various sedimentary rocks and coals.196In spite of natural water^,'^^"^^-'^^ s e d i r n e n t ~ ' ~ ~and this impressive body of knowledge, however, and the fundamental importance of these complexes in the transport and deposition of many metallic species, our understanding at the molecular level of the kinds of complexes formed is at best sketchy. This rather unsatisfactory state is partly due to the way in which soluble humic materials are isolated and, somewhat arbitrarily, classified. After extraction of low molecular weight components such as amino acids, fats and so on from soils or sediments or the collection of high molecular weight polyaromatic materials on ion exchange resins, the resulting brown-pigmented polymers are treated with dilute acids and alkali. Extraction of these humic substances with alkali leaves a residue called humin, which itself displays many of the coordinating properties discussed below. The supernatant is acidified, with HCl usually, and the isolated precipitate is humic acid. The yellow and light brown compounds remaining in the filtrate are fulvic acids. Obviously these materials are really complex mixtures of many macromolecules and their structural complexity is compounded by the fact that the origin of these acids is the woody and grassy tissues of plants. Given that there are many thousands of different species alive at present, and indeed many more in the past, this finding is not surprising. Nevertheless, a number of studies have pointed out the main kinds of ligands and chelating groups in humic and fulvic acids and the general kinds of complexes formed are fairly well understood. Classically, humic and fulvic acids have been thought of as modified lignins,'" a view that is strengthened by the fact that they display similar properties to oxidized lignins. Lignin itself has a complex polyphenol-type structure made up of phenylpropane derivatives, including coumaryl (7), coniferyl(8) and sinapyl alcohols (9) and their congeners. These are linked through condensations and dehydrogenations and are substituted by other aliphatic, hydroxyl and alkoxy1 groups linked together via ether or carbon-carbon bonds. Oxidation of terminal chains would produce carboxylic acid groups, and dealkylation and oxidation would result in phenolic and catecholic (10) compounds together with 0- and p-quinone groups; these are generally known to be the major ligating functions in humic and fulvic acids, although, in such complex compounds, other donors are no doubt present, some of which are discussed below.'" Further degradation in soils and sediments of lignin probably gives rise to the mixtures of benzoic acid, hydroxybenzoic acids, vanillic acid ( l l ) ,quinones, etc. which may be extracted from such materials. CH=CHCH20H

CH=CHCHzOH

CH=CHCHzOH

COZH

OH

OH

OH

Geochemical and Prebiotic Systems

858

As a general model of humic and fulvic acid structures the above model explains, in simple terms, the complexation of metal ions in many environments. Functionally, localization of metal ions at carboxylate sites of similar acid strength indicates that these ligands might be viewed in the same light as synthetic crosslinked polyaromatic carboxylic acids.'993200 Investigations of the role played by quinone groups also indicate that they make major contributions to the metal chelation capacities of humic and fulvic acids.201This conclusion was borne out by IR studies and synthetic work on model systems involving benzoquinone. The accumulation of uranium, germanium and vanadium in peats, lignites and other related substances has been suggested to particularly involve the phenylpropane units of lignin.2MThese metals are believed to be incorporated most intensively during early stages of decomposition. Later, as the material reacts and condenses to form humic and fulvic acids, complex stabilities are further enhanced. Furthermore, degradation of humic and fulvic materials from natural waters gives mixtures of polyphenols and phenolic acids203and this finding emphasizes that the macromolecular ligands consist of esterand ether-linked aromatic acids. Similar results are found for soil humic substance^.'^^ More recent work on fairly well-characterised fulvic acids indicate that these again are polymers of carboxylic-phenolic acids containing both benzene and naphthalene (12) nuclei. More revealingly, a high resolution 13CNMR studyZo5of freshwater humic substances has apparently identified resonances arising from carboxyl, aromatic, alkyl and alkoxy groups. In the particular humic acid isolated, 40% of the carbon atoms were incorporated in aromatic rings and 24% were present in carboxylic groups, Notwithstanding the consistency of all of these findings, the nature and variety of humic and fulvic acids is in fact even more complicated. It is known'93 that humic and fulvic acids contain nitrogen, sulfur and phosphorus ligands, and the origin of the substances is now thought to arise from the decomposition of a11 plant components into simple molecules, followed by polymerization with amino acids, sugars and tannin^.'^',^^^ This is in accord with the fact that humic acids can be derived from algae, lichens and mosses. These species do not contain lignin. Flavones, stilbenes, tannins and other phenolic metabolites of microorganisms are all known to be incorporated in some humic acids. Up to 10% by weight of humic and fulvic acids has been shown to be made up of amino acids. The most abundant of these are glycine (13), aianine (14) and aspartic acid (15), although many others have been reported to be present. Naturally, these units can participate in metal complexation in the normal way via the formation of stable chelates containing five- and six-membered rings. Monomeric silicic acid or perhaps organically bound silicic acid groups are also frequently found in humic and fulvic acids but their role in bonding is not fully understood, A general reaction schemeZo6for the polymerization of amino acids with phenolic acids to form humic substances is shown in Scheme 1. Condensations of monosaccharides and amino acids via the Maillard reaction, followed by dehydration or polymerization, gives rise to brown nitrogenous polymers called melanoidins with very similar composition and spectral properties to other soil humic material^.^" Other groups known to be incorporated in humic and fulvic acids are various nitrogenous bases and fatty acids.'" In summary, an entirely satisfactory scheme for the synthesis of humic and fulvic acids is not yet available. Detailed structures are poorly understood. However, the general ligating groups are known, and these are principally carboxyl and phenolic oxygen and amino acid nitrogen atoms. The mixture of molecular structures is complex and may involve other kinds of ligands, including porphyrin^.^^^,^^^ In addition the physical form of the humic and fulvic acid can be of importance in the natural environment. Attention has been to the presence of composite clay-humic acid particles in the natural environment. Other workersY'*211*212 have studied their effects on trace metal distributions in the marine environment. The organic coatings substantially control the complexation of Cd, Cu and Pb on particles. This work has been extended to a study of Cu" and Ag' uptake by amorphous iron oxides. It was concluded that rather than reactions with simple metal oxide sites, trace element uptake in the natural environment by such substances is probably controlled by their coatings of humic and/or fulvic acids followed by incorporation in the mineral lattice or on its surface. However, the chemistry of complex formation is for all intents and purposes the same as that for 'free' humic and fulvic acids, and we make no distinction between these systems here.

co2H I

HZN-C-H I (12)

COzH

I I

H,N-C-H

H

Me

(13)

t 14)

CO,H I

H,N-C-H

I

CHZCO, H (15)

Geochemical and Prebiotic Systems

I

I

Humic acids

859

OH

1 ~

similar intermediates

Scheme 1

Many studies have been carried out concerning the stability constants of humic and fulvic acid c~rnp~exes.'~ Stability ~~'~~ constants ~ ' ~ ~ vary considerably with pH and ionic strength213and this, together with the variable nature of the ligands involved, accounts for the range of values reported for individual metal ions in the literature. However, the stabilities of divalent metal complexes generally follow the well-known Irving-Williams order Mg < Ca < Mn < Co < Zn = Ni < Cu < Hg. M2+l-HA

MA+2H'

(1)

Table 10 shows some representative ranges of log K values for reaction (1). It should also be noted that humic materials of differing origins have different stability complex behaviour. Soil fulvic acid in general forms the weakest complexes and marine sedimentary humic acid the strongest. Some of the variations can be ascribed to different modes of binding to the humic and fulvic acids. For example Mn", which does not bind strongly, has been observed to retain a high degree of i0nicity.2'~ Other workers have s h o ~ n ~ ' that ~ , ~the " Mn(H:O):+ ion appears to retain its inner hydration sphere to a significant degree in Mn"-humic and -fulvic acid aggregates of soil and aqueous origin. Adsorption and binding of this ion, particularly by electrostatic forces, is probably important in complexes of this type. Table 10 Ranges of Stability Constants" for Humic and Fulvic Acids of Various Origins with Selected Metal Ions1'*

Mg"

Zn" Ni" cu"

3.2-4.1 3.2-4.7 4.6-5.0 4.3-4.9 4.3-5.0 4.8-5.9 5.0-5.5 7.8-10.4

Hg"

18.3-21.1

Ca" Cd" Mn"

co"

a

lonic strength = 0.02 molal, pH = 8.0.

More detailed studies on the nature of complex formation between humic materials and certain individual metal ions have been carried out and it is appropriate to consider these more fully. Fe"' and Al"' form very strong complexes with fulvic and humic acids and compete for the ligating phenolic and carboxylic groups.216This was anticipated by the earlier work of J ~ s t e . ~ " Similar changes in the IR spectrum were observed upon treatment of a humic acid with Fe"' or Al"'. Carboxylic groups were considered to be involved in coordination because substantially the same spectral pattern was obtained on treatment of the sample with strong base. On this basis the strong correlations between Fe"', AI"' and dissolved organic matter in natural waters is readily u n d e r s t ~ o d . ~A ' ~series ~ ~ ' ~of elegant experiments220using ESR and Mossbauer spectroscopy has also shown that more than one complexing site exists for Fe"' in humic materials. T h e first contains the metal ion in a site with tetrahedra1 or octahedral coordination geometry and binds the metal strongly. The metal ion is resistant to removal by complexation, reduction and substitution

Geochemical and Prebiotic Systems

860

by other metal ions. The second site was associated with the surface of the humic acid ligands. The iron is octahedrally coordinated, more weakly bound and can be reduced and complexed relatively easily. This observation parallels those of other who studied the reduction of Fe"' to Felt by humic substances and in model complexes. Fe" is found in considerable amounts as humates even in well-oxygenated waters. This is due to the fact that Fe"-humic acid complexes are quite stable as well as to the ability of humic materials to reduce Fe"' to Fe". Other ligating groups may also be important for Fe"'. For example, Mossbauer studies have also indicated223 that Fe"' is associated in part with silicate groups in some humic and fulvic acids. Copper, lead, cadmium, zinc and probably other divalent transition metal ions appear to compete for similar coordination sites in humic and fulvic acids. However, in a fashion analogous to the case with Fe"' some workers have identified more than one site for a particular ion. At least three have been suggested as being available in humic acids for Zn11.224 Pb" forms complexes with humic acids which are about as stable as Cu'I species, but Pb" humates and fuivates are less soluble and quickly coagulate and p r e ~ i p i t a t e . This ~ ~ ~simple , ~ ~ ~ fact is of some significance with respect to Pb" distributions in surface water and sediments. The extremely low concentrations of lead in natural waters may partly be explained by the scavenging of lead from solution by such precipitated complexes, via adsorption processes or complexation at sites still unoccupied by other metal ions. Several workers have proposed models for the coordination sites responsible for complexation. It has been widely held227*228 that the interaction of Cu", Cd", Zn" and Pb" with fulvic and humic acids involves a salicylic acid group (17), but later workers229believe that a phthalic acid group (18) is involved. No doubt, in actual fact, both kinds of interaction exist in the natural environment together with others mentioned above.

17 II

0

II

The enrichment of uranium in organic geological units is a well-known phenomenon, and may be sufficient under certain conditions to concentrate enough of the element to make its recovery economic. Szalay"' has summarized much of the early work in this field and he considers that the formation of polycarboxylate complexes involving humic acids is the overwhelmingly important chemical process in the enrichment. Quite spectacular enrichments are observed, of the order of 10 000: 1 in natural water at pH 5-6. Uranium is thought to be present in these complexes as the uranyl, UO;+, ion, and uranyl fulvates have been suggested to play a significant role in uranium mobilization in near-surface ground water^.^^' Subsequent diagenesis of uranium fulvates, humates and humins is expected to lead under reducing conditions to the crystallization of uraninite and pitchblende, Ul-,02. Many deposits of UO, are known in which the uranium minerals are associated with high molecular weight carbonaceous matter, graphite and other highly altered organic material. It is generally assumed that this material is probably the remnants of organic substances originally responsible for uranium scavenging from aqueous solution by complex formation. Considerable attention has also been paid to humic and fulvic acid complexes of vanadium. This is due to the observation of the unusual concentration of this element in a number of bi01iths.~~' Enrichment factors of at least 50 000: 1 are typical. Vanadium is thought to migrate in water as V02+, V 0 3 - and other Vv species,"' and to be reduced by the humic or fulvic acid by to V02+ prior to incorporation into very stable coordination complexes. It has been the application of ESR measurements that the semiquinone groups present in humates are not responsible for the reduction. ESR measurements also lend themselves to the characterization of the V'" humic complexes, and have been employed by a number of w ~ r k e r ~ All . ~these ~ ~ ~ ~ workers agree that the V02+ ion enters rigid surface sites involving oxygen donors and that it loses rotational mobility, although some nitrogen ligands may also be involved.234As opposed to this view, Templeton and C h a ~ t e e have n ~ ~ shown ~ that model complexes of V02+ with phthalic and salicylic acids give ESR spectra apparently identical to those of V02+-fulvic acid complexes. There thus seems no real doubt that the usual aromatic carboxyl and phenolic oxygen donors are responsible for complex formation in this case. Szalay'% also suggested that a similar reduction to Mov species resulted in molybdenum complexes of humic acids and the accumulation of Moo:of this element. Other works seems to bear this but such processes do not appear to be

Geochemical and Prebiotic Sysiems

86 1

involved in the incorporation of tungsten into coals.23bIn this latter instance, it is suggested that complex formation involves the oxygen donor groups of humic and fulvic materials interacting with various cationic hydroxytungstate(V1) species in aqueous solution. A few other metal-humic acid associations are worth noting. It is possible that Pd" occurs as organic complexes of this type:37 and Ag' complexes are well knownFXbeing formed by linkage to the humic material over the same pH range as the dissociation of the carboxyl and phenol groups occurs. K+, Ca2+ and A13+ all decrease the extent of complexation with Ag', and, as expected, this effect is more pronounced with the more highly charged cations. Gold may also be mobilized in soil litter horizons and transported as humate c o m p l e x e ~ . However, ~ ~ ~ * ~ ~it ~is possibly more significant that gold sols are protected by humic and fulvic acids, thus allowing some transport under the right conditions. Inorganic gold complexes with oxidation states 1 and 111, some involving ligands like CN- and SCN- derived from decomposing plant material, and which are important in gold coordination geochemistry, are discussed in an earlier section. Some intriguing variations in the nature of the humate ligands have also recently come to light. Studies of metal fixation in sedimentary organic matter recently deposited in the form of algal mats have been carried The material was originally derived from cyanobacteria and the humic substances are characterized by a very low aromatic content balanced by the incorporation of large amounts of carbohydrates and amino acids. Similarities between cation coordination versus pH curves in titration experiments for UO;+, Ni2+, Mn2+, Zn2+, Co2+, Pb2' and Cu" pointed to a common first step in the complexation of all cationic species. This was suggested to involve bond formation between the metal ion and anionic carboxylate groups. A marked selectivity was also observed. Molybdate, vanadate and germanate were only weakly bound. Mixed ligand complexes may also exert a significant influence on metal speciation and concentration in natural waters. Fuhate-phosphate complexes of Cu", Pb", Cd" and Zn" are more stable than the metal-fulvic acid c o m p l e x e ~ . Incorporation ~ ~ ~ ' ~ ~ ~ of these species into models in order to estimate free metal ion concentrations gave estimates which were in excellent agreement with concentrations determined experimentally using ion-specific electrodes. However, mixed ligand complex formation with humic and fulvic acids has not received widespread attention, despite these results and their potential importance. Finally, humic and fulvic acids can act as chemical weathering agents by complexing reactions. Early work was carried out on metal oxides244and the differential leaching of cations leading to the decomposition of rocks and minerals is frequently cited.' Many early experiments on the role of humic and fulvic dcids in the breakdown of rocks and minerals are summarized by Ong et aL245There is no doubt that chelation of rnetal ions b y humic materials is of great importance in this respect. Furthermore, the process might be responsible for the formation of lateritic aluminum deposits, which are restricted to tropical climates which have an abundance of humic acids in soils and waters. Humic acids possibly are involved in the dispersion of metals from ore deposits, by reaction with both primary and secondary minerals. Experiments show that galena and blende are rapidly attacked by a podzolic soil humic acid extract;246the study was extended to ore metal oxides, secondary minerals and various components of gangues. Typical extraction results for the sulfides were 0.12 mg per day of Zn from blende and 3 mg per day of Pb from galena using a 0.1% (w/v) aqueous soiution of the humic acid preparation. These values are much higher than for rain and other near-surface waters and it was suggested that the concept of relative metal mobilities near ore bodies might need to be reassessed in environments which contain large concentrations of humic and fulvic acids.

64.5.2 Complexes of Porphyrins and Related Macrocyclic Ligands The origins of organic geochemistry could be traced to the pioneering studies of T r e i b ~ ~ ~ ~ , " ' who discovered that a number of petroleums and bitumens contained porphyrins. Later studies confirmed metal complexes of porphyrins in such organic media as well as in shales and the conclusion that the ligand was in fact deoxophylloerythroetioporphyrin (DPEP; 6 ) ,a degradation product of the chlorophylls, was taken as proof that chlorophyll-bearing plants were primarily responsible for the formation of petroleums and other asphalts. This monumental discovery has prompted numerous researches in this field, signifying as it does the continuity of life forms similar to those present today on the earth from early times. Vanadyl porphyrins have been isolated from shales at least one billion years old.24Y Metal complexes of porphyrins involving Mg", Nil', VIv, Fell', Ga"' and Cull have been identified from a number of sediments. Early work in this area has been summarized by Saxby,'86 and a

862

Geochemical and Prebiotic Systems

recent reviewzsuof the many types of tetrapyrroles found in the geosphere has placed much of the published organic and coordination chemistry of these compounds in perspective. Nevertheless, it has only been over the last year or so that the definitive structural evidence proving the link between chlorophyll and many petroporphyrins has been forthcoming. The metal complexes found in geological media are based on four classes of ligand: the porphyrins, the simplest being porphin (19), dihydroporphyrins (chlorins) (20),tetrahydroporphyrins (21) and the corrins (22). These in turn are generally thought to arise from the breakdown products of the chlorophylls, haem-type centres, especially the cytochromes, and related complexes such as vitamin B,* and nirrins, associated perhaps with metal enzymes in methanogenic and allied bacteria.

While the detailed molecular arrangement of most geoporphyrins is not known, much chemical evidence has accumulated concerning the basic structures of the ligands and complexes involved. Their general geochemistry and distribution in roils, sediments and rocks ranging in age from the Precambrian to the present has been e l ~ c i d a t e d . ~In~ sedimentary ' rocks and asphalts, porphyrins are predominantly present as Ni" or VOz+ complexes. This has been attributed to the great stability of these coordination complexes and it is thought that the formation of the coordination compounds themselves confers the overall resistance of the ligands towards degradation during the usual later diagenetic processes. However, early also noted the presence of Fe"'- and CUI'containing species. As is only to be expected, the concentrations of metal porphyrins in geological media are highly variable. Ni" species amount to some 10 p.p.m. in sediments and Fe"' complexes are present at roughly the same levels. Vanadyl compounds are comparatively more concentrated in marine horizons. In asphalts, complexes can be present at about the 0.1% level, but are most frequently represented at much lower concentrations. These comparatively low concentrations are no doubt responsible in part for the lack of detailed molecular structures in the literature. Coupled with this is the fact that complex mixtures are invariably isolated in the first parts of most procedures; the number of isolated single species is rare indeed. Chromatographic results, volatility studies and powder X-ray data all pointed at an early date ta the presence of groups of isomers in isolate fractions rather than the presence of unique compounds.254Finally, the application of mass spectroscopy to the study of petroporphyrins clearly established that this was the case, revolving about classic studies involving material isolated from a Swiss oil shale and units from the Green River, Colorado, oil shale.25532S6 One important fact emerged from these investigations. This was that porphyrins and their complexes (the compounds being demetallated prior to mass spectrographic analysis) occur in homologous series which arise from variations in the number and nature of alkyl substituents at peripheral sites of the basic macrocycles (6), (19)-(21). Observations of peaks in the series 310+14n and 308+ 14n ( n 2 2) attest to increasing degrees of alkylation in the ligands based on porphin and a DPEP, respectively. This compelling evidence was presented some time ago and is summarized in a short review of early work in this field.257Most frequently, masses of from 450-500 are found for fossil porphyrins in sediments and asphalts. Generally, DPEP itself is the largest component of most mixtures although etioporphyrins related to the haem family, but also possibly arising from chlorophylls by the opening of the exocyclic ring, can be present in comparable or greater amounts. Furthermore, significant amounts of either series containing as few as eight or as many as 20 or so methylene groups can be present, although concentrations usually drop off with either decreasing or increasing molecular weights. Inhomogeneity is quite naturally predicted to occur in this class of complexes. This is not only due to the different possible breakdown pathways in soils, sediments, etc., but is also due to the multiplicities of naturally occurring precursors in living organisms. For example, a number o f known chlorophyll structures are shown in (23). The structural complexity is even more elaborate for the haems. Mixtures in geological media thus are to be expected. However, it should be pointed out that the amount of porphyrins from plant sources is overwhelming compared to that

Geochemical and Prebiotic Systems

863

R' = CH=CH2, CHO, CH(0H)Me RZ= CHO, Me, R3 = Et, Pr,Bu' R4= Me, Et R5= H, C 0 , M e R6 = phytyl, famesyl R7= H, Me, Et

from animal sources (by the order of about lo5 by weight) and it is probable that the majority of the haem geocomplexes also arise from plant pigments other than, and via particular degradations of, chlorophylls rather than from animal remains. In the original work mentioned above, Treibs outlines the general mode of decomposition of the chlorophylls in sediments and so on. Magnesium is lost first: this reaction proceeds with ease and it is possible that other metals are coordinated at an early stage. In agreement with this notion is the fact that Ni" and V02' stabilize the overall porphyrin structure, and that Ni" chlorins have been isolated from recent freshwater These reactions are followed by ester hydrolyses, vinyl group hydrogenation, oxidation of the chlorin to the porphyrin, reduction of the carboxyl group(s) and decarboxylation. It is of interest to note that isoprenoids in sediments and asphalts are almost certainly derived from farnesol(24) or phytol (25). The same can be said of numerous other fatty acids and related molecules also present in such and related media.2s9Such a scheme of degradation leads naturally from chlorophyll b or bacteriochIorophyl1 to DPEP, the product advanced by Treibs. Variations of stereochemistry and isomeric distributions since observed have been ascribed to transalkylation reactions or radical combinations with alkenes of different molecular weight. It should also be remembered that ligands based on (20) and their metal complexes may also be generated from those based on (6) via the opening of the exocyclic ring if the ester hydrolyses occur under basic conditions.

Aside from mass spectroscopy, the most frequently applied technique to the study of petroporphyrin and other geoporphyrin complexes is electronic spectrophotometry, and this formed the basis of the original identifications. Distinctive visible and electronic spectra, especially concerning the so-called Soret bands,260have been extensively utilized for gross structural determinations. Details can be found in the general reference given above. Four main bands are evident in the visible region of the electronic spectra of the free ligands, which correspond generally to the four main structural types. These are DPEP, and the etio (alkyl substituted), phyllo ( -CH2CH2C02H) and rhodo ( -CH2CH2C02H and -C02H) series based on the nucleus of (19). The visible region of the electronic spectrum is fairly sensitive to structural alterations of the ligand, and the organic molecules are usuaIly red in colour. This fact distinguishes, in most cases, ligands based on (19) from those based on (20), which are green. A somewhat different picture emerges concerning the visible electronic spectra of metalloporphyrin complexes, which are dominated by two bands. In non-coordinating solvents, Nil' etioporphyrins show a distinctive spectrum characteristic of a square planar complex with maxima around 550 and 510 nm.The corresponding square pyramidal V 0 2 +complexes have maxima around 570 and 530 nm. Extensive discussion of spectral interpretations has appeared.2503251~257 Indeed, the postulated existence of some porphyrins observed in mass spectral studies relies on such work. One example is a series based on (26) which has been noted from a number of geological sources.261Other related complexes, reminiscent of the phthalocyanines, which contain more than one condensed benzene ring, may also be present in sediments and asphalts. No chlorophylls with related structures are known, however, and the origin of such complexes remains rather enigmatic. It is possible that they derive from reactions of other complexes with aromatic fragments in asphalts. Several workers have presented evidence of high molecular weight porphyrin-containing complexes in oil shales and other material^.*^^'^^^ The porphyrin complex nuclei may be bonded to kerogen during diagenesis. Peptides and proteins may also be linked to the central coordination complex in some case^.^^^,^^" Gel-permeation studies have indicated some very high molecular weights.

864

Geochemical and Prebiotic Systems

(26)

and showed that Treibs isolated the first porphyrin coordination complex from a Swiss it (or at least a structurally related derivative) occurred in a wide variety of sedimentary rocks and asphalts. He showed it to contain vanadium and subsequently the involvement of the vanadyl ion was provedFM Treibs also described a second chelate from the same original source, which he proposed to be an iron complex. Somewhat later it was shown that the complex contained the Ni" ion. It is thus appropriate, and in view of much of the foregoing discussion, to focus attention on the detailed structure of the few fully characterized geoporphyrin complexes recently reported, since their elucidation crowns some five decades of intense activity in this area of coordination geochemistry. High resolution 'H NMR techniques have proved to be of immense help in this Very recently, Fookes reported2'* the separation of Ni" porphyrins from the Julia Creek oil shale from Queensland, Australia. The original material contained some 11 p.p.m. of Ni complexes in all and these were separated by a combination of TLC and reverse-phase HPLC procedures. The mass spectra of two of the fractions displayed patterns indicating that they were composed of single complexes, one corresponding to Ni"-DPEP and the other to one of its C,,-homologues. In a series of elegant experiments involving the exhaustive use of the nuclear Overhauser effect (NOE) it was unequivocally established that the first complex was indeed that shown in (6)and that its congener was that shown in (27a). Previous workers2" had reported a partial structure determination. These outstanding results finally proved the long-assumed link between chlorophylls and petroporphyrins and their complexes. A third Ni" complex was isolated and it corresponded to a C,,-homologue as evidenced by its mass spectrum. A full characterization of this species by NOE spectroscopy established this species as that shown in (27b). In addition, X-ray crystallographic evidence has been forthcoming. The major vanadium-containing complex from the Julia Creek oil shale was crystallized and a structure determination has shown it to be the V 0 2 + analogue of (6).272The final structural proof concerning these complexes so long sought by so many workers has thus now been provided.

(27r) R = H (2%) R = M e

the structural characterization of a homologous series of 15,17Fookes has also butanoporphyrins from the Julia Creek material. The three homologues are shown in (28a-c). Other workers274have also established the presence of the free ligand of (28b) in the Serpiano oil shale. In this study, the ligand was derived from the vanadyl complex. The variation in both series of complexes characterized with respect to substitution at C-3 could be accounted for by the original presence of a vinyl group at that position but the origin of the seven-membered ring in (28) is obscure in that it is not known in any present-day chlorophyll, although it may have arisen from an extinct species. Alternatively it could be derived via the condensation of the pmpionic side chain with the exocyclic ring as shown in (29).273 Subsequent hydrolyses, reduction and decarbonylations could lead to the observed species. It is anticipated that further related characterizations of individual species will yield valuable clues as to the origin of such complexes and their bearing upon evolution trends concerning extinct species.

Geochemical and Prebiotic Systems

865

(ZSa) R = M e

(ab) R=Et

(28c) R = H

A few unambiguous characterizations of metal porphyrins not containing an exocyclic ring have also been recently reported. Gilsonite is a generic name given to a complex bitumen from the Unita Basin, Utah, USA. It is known to contain a reasonably simple distribution of etio- and DPEP-based nickel porphyrins on the basis of a series of mass spectrographic and chrornotographic s t ~ d i e s .Up ~ ~to~about , ~ ~100 ~ p.p.m. of Ni" species are present in the gilsonite. Two etioporphyrins (30a,b) have been demonstrated to be present. Their characterization was achieved, however, after removal of Ni" from the c o m p l e x e ~ .The ~ ~ identity * ~ ~ ~ ~of the complex involving etioporphyrin I11 (30b) was proved using a novel NMR technique involving the preparation of a double sandwich complex of the type bis(porphyrinato)mer~ury(II)acetatomercury(II)."' This kind of adduct might prove to be of importance in the future in the identification of other porphyrin complexes of geological origin. A particularly noteworthy feature of this Ni" etioporphyrin species is the fact ,that it is clearly derived from chlorophyll a, the most abundant of the naturally occurring chlorophylls. The exact point in the degradation of chlorophyll a at which the exocyclic ring is opened is not known, but it has been shown2'' that etioporphyrins are transformed from members of the DPEP series if they are heated in clay mixtures. This process is quite analogous to those occurring during diagenesis of the sedimentary column.

C02H

(3Oa) R' = Et; R2-R8 = Me

COiH

(31)

(30b) R', Rs= Me; R2, R4,R6, R7 = Et (3oc) R', R7 = Me; R2, R4, R6, R8 = Et

Recent evidence has come to light that geoporphyrins and their complexes may indeed have a mixed origin. This condusion was drawn on the basis of the isolation of mesohaem (31) from an Australian lignite.239The concentration of the porphyrins present was less than 1 p.p.m. The natural material was demetallated, converted into its dimethyl ester, and the 'H NMR spectrum of its Zn" complex was found to be identical to that of an authentic sample. The Fer" atom present in the original complex might have been that originally incorporated into the macrocycle during biosynthesis. The elucidation of the mesohaem geoporphyrin structure indicates therefore that it and other related complexes might arise in sediments and other media uia the degradation of various cytochromes. The same group has also reported the isolation of a number of gallium(lI1) porphyrins from a bituminous coal.'" One of the components of the mixture of gallium complexes had the required molecular weight for, and the same retention time in HPLC experiments as, the synthetic hydroxogallium(II1) etioporphyrin I chelate (30~). Of the corrin (22) series of Complexes comparatively little is known. Vitamin BIZ,the Co"' complex based on this ligand, and its congeners are known to be present in ~eawater.'~'It is however quickly metabolized and migration of Co"' in muds and related sediments is primarily controlled by biological activity involving the more or less intact vitamin or its derivatives.*"

866

Geochemical and Prebiotic Systems

The complete structures of but few geoporphyrins have been established. This state of affairs does not bring into question the enormous volume of work on such naturally occurring coordination complexes which can be gleaned from the general references given. Quite the contrary! The most recent work provides the final confirmation of the scheme of Treibs and thoroughly establishes the link between living organisms and tetrapyrroles and their complexes observed to be widely if not abundantry distributed in rocks and asphalts. It is anticipated that in the near future many more such detailed studies will be reported as a result of advances in instrumentation, especially those in the field of high resolution NMR spectroscopy. Results of this nature are essential for a complete understanding of the origin of the complex mixtures of porphyrin complexes found in nature and indeed to shed light on evolutionary processes which have taken place over the last few billion years.

64.5.3

Complexes with Other Organic Ligands

from natural Organic geochemists have now isolated many hundreds of organic waters, soils, sediments and rocks dating from the present time back to the Precambrian era.’” Indeed, a considerable number of these species are potential ligands containing the usual donor groups based on elements in groups V and VI of the periodic table, and perhaps others. However it is only fair to conclude that the coordination geochemistry of these ligands has not been deeply studied in the sense of the isolation of individual complexes. The main reason for this is probably concerned with experimental procedures adopted for the separation of the ligands themselves; extraction of sediments and so on with strong mineral acids is nat particularly conducive to preservation of many of the potential complex molecules. Nevertheless, the presence of the ligands in environments rich in metal ions available for complexing, taken together with our knowledge of the formational stabilities of certain chelates reported elsewhere in this compilation of reviews, makes it inconceivable that numerous kinds of coordination complexes of ligands other than humic acids and porphyrins do not exist in a wide variety of geological environments. A number of the probable interactions, some of which are potentially of great importance, are mentioned below. Amino acids are widespread in nature and have been isolated from all waters and rocks, except igneous and highly metamorphosed units, together with peptides and protein^.*^^,^*^ Complex formation is overwhelmingly probable in these circumstances and a few examples put forward in the literature are worth enumerating. The rates of racemization and epimerization of amino acids in sediments and fossil remains have received increasing attention with respect to application as a possible dating r n e t h ~ d . ~ ~ ~ The most extensively studied materials are the calcareous tests of pelagic formanifera, which are abundant in marine sediments. These are biomineralized substances consisting of polypeptides and proteins embedded in calcium carbonate and related mineral matrices.287Amino acids are released slowly by hydrolysis and undergo a series of degradations as well as being racemized. Racemization rates vary from amino acid to amino acid as might be expected; equilibrium has been observed in Miocene samples.’88 Recently it has been shown that metal ion catalysis of the racemization of amino acids may be important in geological environments and such reactions may give rise to misleading dates. The generally accepted scheme for the reaction is shown in Scheme 2. Of course, diastereoisomeric amino acids may give rise to equilibrium constants not

equal to unity. Early evidence for this kind of metal ion catalysis involving a coordination complex was presented by Bada and S~hroeder.~’~ They observed that the rate of epirnerization of (S)-isoleucine (35) is about an order of magnitude greater than would be expected on the basis of experimentally determined reaction rates. In addition, it was noted that (R)-alloisoleucine/ ( S ) isoleucine ratios, (R)-alloisoleucine (36)being the product of epimerization in this case, were

Geochemical and hebiotic Systems

867

higher in the free amino acid pool compared with the protein fractions. It was suggested that divalent alkaline earth complexes were probably responsible for the catalysis, and that these derived from the calcareous mineral of the samples. Other metal ions, notably Cu", were also suggested as being important. More recently, analogous catalysis via coordination complexes has been proposed for the (S)-threonine (37)-( R)-allothreonine (38) epimerization in fossil formanifera.290Other dating discrepancies found for fossil corals291may also derive from related reactions. Many metal ions, particularly of the transition series, could be involved, and are known to participate in these epimerizations. However, their mode of action is and much more work is required to evaluate their significance in a geological context.

Et

Me (37)

(35)

Me (38)

Threonine and allothreonine are thermally unstable amino acids. The same is true for the simplest hydroxy amino acid, (S)-serine (39). High levels of this amino acid have however been reported from certain petroleum brine waters294and its preservation may be due to the formation of stable chelates. A h r e n ~ 'has ~ ~ suggested that metal-amino acid chelates may be important in the sedimentary cycle, and the uptake of certain metals in sediments, particularly in binding of metal-amino acid chelates to clays, has been proposed as a possible mechanism for the incorporation of metal ions in carbonaceous materials.296Similar complexes are thought to contribute to the stabilization of amino acids in coals, lignites and peat^,^'^ but, as has been outlined in a previous section, many other ligands are also present in these media, and their relative importance is difficult to assess. Much of this work is unfortunately of a speculative nature since no well-characterized complex of an amino acid has been isolated from a geological source, as far as the author is aware.

CHsOH

(39)

Many other nitrogen-containing species which are perfectly capable of forming complexes with metal ions have been isolated from sediments, rocks and asphalts. Among them are pyridines, pyrimidines, quinolines, indoles, purines and the like, as well as pyrroles, imidazoles, nucleic acids and amino sugars.294~298-300 In addition many sulfur compounds have been reported, especially from petroleum?'' and these contain potential ligating sulfur atoms as thiols and thioalkane fragments in bulky thioalkanes and heterocyclic molecules. However, while the potential exists for complex formation with these ligands, little is known at the molecular level of their participation in geochemical coordination cycles. Somewhat more detailed information is available concerning complexes of carboxylic acids. Hydenlx4was among the first to suggest that trace metals in crude oil were at least partially present as salts or complexes of organic acids. Similarly, Soviet workers have suggested that these kinds of complexes are important in sediments and Carboxylic acids are the major oxygencontaining components of crude oils303*304 and short chain aliphatic acids including acetic, propionic, butyric, isobutyric, valeric and isovaleric acids have been isolated from oil field brines.30s Phenol and phthalic acids were also detected in these media. The occurrence of hydroxy acids in sediments has been recently reviewed306and low molecular weight acids including the above, together with oxalic, malic, 2-ketoglutaric, tartaric, lactic and citric acids have been isolated from Interactions with metal ions of the aromatic members mentioned above soils and (and their congeners) have been covered in a previous section. Attention is also drawn to the fact that some of these ligands occur as coordination complexes in isolated mineral species as listed in Section 64.2.2. Thus there is no doubt that the coordination geochemistry of these substances is ubiquitous even though few complexes have been isolated intact. Tightly bound a-hydroxy acids observed in recent sediments might owe their resistance to extraction to participation in complexing on mineral surfaces.3n8

868

Geochemical and Prebiotic Systems CO2 H

1

c=o I

y

2

y

2

CO, H (40)

Mineral weathering, or at least that part due to the exudation of chelating agents by microorganisms, is probably the best-studied area of coordination geochemistry involving organic oxygencontaining ligands. This subject has been recently reviewed in an authoritative manner3'' and many examples have been reported. Simple chelating acids including those mentioned above as well as salicylic and 2,3-dihydroxybenzoic acids are formed by microorganisms. The simple species 2-ketoglutaric acid (40) has been shown to be important during the microbial weathering of silicates and forms complexes with Ca" during the solubilization of calcareous rocks, apatite and montmorillonite. Citric and oxalic acids also contribute to the destruction of silicate and other rocks by coupling with Fe"', Al"' and Ca". Several workers have conducted experiments along these lines and results indicate that chelate formation with all of these kinds of ligands is responsible for a wide variety of mineral transformation^.^^*-^'^ Other workers have identified an NiII-stearic acid complex spectroscopically in a hectorite and other kinds of ligands derived from microorganisms might also be important in the geochemical cycle. A number of gallic acid complexes of aluminum have been studied by EMF titration techniques and stability constants have been computed.315The possible role of these complexes in the geochemistry of natural waters was assessed and it was shown that kaolinite solubility can be increased about five-fold at neutral pH. Hydroxamic acids may also be important in related systems.309 A comment on possible interactions of metal ions with sugars and polysaccharides in the neutral environment is warranted. Metal ion incorporation into algal mats was mentioned in Section 64.5.1, in which bonding to sugar residues as well as to amino acidate groups was probably involved. Many sugars have been reported from seawater and sediments316and metal incorporation into detrital polysaccharides and alginates has been suggested to be important in seawater.317 Experiments describing the extraction of uranium from solution by chitin and its derivatives have been rep~rted.~''Related processes may control in part the distribution of uranium in natural waters. Again, however, it must be said that the true role of such ligands in the geochemical cycle is of a speculative nature. The challenge remains for geochemists and coordination chemists to evaluate fully the occurrence and importance of such complexes. Of paramount concern, it would seem, would be the isolation and characterization of some of the many chelates which undoubtedly exist in natural waters, soils and rocks. Finally, attention is drawn to the possibility of organometallic complexes in the natural environment. We do not wish to examine alkyl metal species such as methylmercury and dimethylmercury compounds, important though these might be in terms of pollution studies. Their general occurrence is covered in a recent review.'*' Many alkenes are present in geological media but no evidence is available to suggest that .rr-complexes are formed. Mention has already been made of vitamin B,2 and cyanide as a ligand in minerals, and more importantly in the potential mobilization of gold in its ores under near-surface weathering conditions. Conceivably, metal carbonyls could also contribute to metal ion transport at high temperatures under reducing conditions. A notable recent article, however, points out that organometallic complexes may be much more widespread in nature than i s currently supposed. The report concerns the isolation of a number of organoarsenic compounds from the Green River oil shale and represents the first report of such complexes in fossil fuel Both methyl- and phenyl-arsonic acids were shown to be present together with the arsenate ion. The phenyl-substituted compound is of some significance since no examples of biophenylation have been reported to date. Biomethylation of arsenic compounds, on the other hand, is well known. In this connection it should be noted that AsMe3 is known to be produced by the action of certain microorganisms in One might anticipate that organogeochemicals might receive increasing attention in the future as a result of advances in instrumentation and activity in the field of pollution studies, especially those relating to the fate of metal-containing waste in the environment. 64.5.4 The Possible Involvement of Organic Coordination Compounds in Ore Formation This is a field of potentially great significance, although it has received little attention to date, relying as it does on geochemical inference rather than a solid foundation of experimental fact.

Geochemical and Prebiotic Sysiems

869

A recent article”* has highlighted the possibility of the transport of Pb” as organic complexes in Mississippi-type ore solutions. Development of a purely inorganic model for metal complexing and transport gave rise to estimates of Pb” concentrations which are probably too low to account for the volume of mineralization found in these deposits. Evaluation of stability constants for metal acetate and related species indicated that such complexes were unlikely to enhance the concentration of Pb” in the ore-bearing fluids to the levels thought to be involved. However, if salicylic acid had been present, a considerable increase of Pb” in solution as a complex would result. While salicylic acid has not yet been detected in oil field brines, from which the transporting solutions are thought to originate, either it or a structural analogue could have given rise to the observed metal orebody distribution. Mention has already been made of the accumutation of uranium by organic material, predominantly in the form of humic acid-type complexes, and the involvement of such interactions in uranium ore formation, Other workers have suggested that the oxidation of organic compounds by hydrothermal solutions may give rise to a lowering of the solution redox potential.320This was suggested to play a part in the crystallization of uranium in certain deposits of the Bohemian massif. Isoprenoid hydrocarbons, amino acids and fatty acids are all associated with copper, uranium and hydrothermal fluorite deposits of the region, and may all have been involved in complex formation. Related studies by earlier Soviet workers3” have commented on amino acids present in fluorites from the Azov region. In mineralized limestones, glycine, alanine and valine were detected and the distribution of Cu, Pb, Ga, Sb, Co, Mo and other metals was explained by complex formation between the ligands and metal ions. Methane is a common component of fluid inclusions in minerals from many hydrothermal deposits. Low molecular weight hydrocarbons of various sorts are also present, Kranz”’ has summarized the early work in this field. Alkanes, alkenes and alkylbenzenes were all present in a number of inclusions in fluorite, together with fluorinated hydrocarbons, that he studied from several European localities. In addition the donors methylamine, dimethylamine and ethylamine, together with glycine, alanine and aminobutyric acid, could be isolated from the fluorites and feldspar inclusions. It was concluded that these substances probably derived from the degradation of humic acids and the like and could have been important in the transport of metals as complex ions. In a later a number of nitriles together with amines and amino acids were reported from uraniferous fluorites. A high correlation between U and Th contents in the samples and the organic material was taken to indicate that the organic material was intimately linked with metal transport. It was also suggested that the amino acids themselves might not be of biological origin but could have been synthesized as radiochemical products of reactions of hydrocarbons, ammonia and water in the ore-forming fluids. On the other hand, some model studies indicate that in certain sulfur-rich marine waters the concentrations of amino acids and simple organic acids are too low to much effect metal speciation in solution. The possible importance of humic complexes was again emphasized in these environments. In this connection, such complexes have been proposed to be important in the generation of certain stratiform base-metal ~rebodies.~” Sulfur could be contributed to the system from components of the humic acids themselves. Metal sulfides were precipitated in an experimental investigation from deoxygenated suspensions of metal humates. Framboidal textures, frequently observed in sulfide assemblages from such deposits, may also result from the decomposition of metal humate complexes. In a related study of the Carlin gold humic substances associated with the mineralization were suggested to have been responsible for the gold deposition, Gold Complexes such as AuCl, or AU(CN)~-are thought to be taken up as complexes involving chelation by N, S or 0 donors prior to reduction to metallic gold. The concentration of copper in the Kupferschiefer of Northern Europe has also been attributed to organic complex transport of the element.327While humates are most probably involved, experiments involving tannic, salicylic and citric acids indicated that these or related ligands could also be implicated in copper mobilization. Finally, the transport of manganese by organic acids has been proposed as a mechanism for the formation of a number of ores of this element.328The geochemical cycle of manganese is intimately associated with microbial activity,”’ and some of the ligands mentioned earlier might be involved in its complexation and mobilization. Humic acids and their breakdown products may also be important. Brockamp’s experiments328with these species, salicylic, citric and tannic acids give rise to a reasonable explanation for the origin of various marine sedimentary manganese deposits. There seems little doubt that coordination geochemistry does play a role in ore formation in a number of environments. Much more work in this field remains to be done however. It is possible that sedimentary iron formations also derive from similar cycles to those mentioned above for manganese.330The uptake of iron by bacteria which are capable of synthesizing magnetite, Fe304,

870

Geochemical and Prebiotic Systems

in viva could lead to the deposition of sediments containing significant proportions of that mineral. A proper assessment of complex formation on ore depositions, while offering a tempting and plausible explanation for some geological settings, must however await more rigorous experimental investigation.

64.6

PREBIOTIC SYSTEMS

It is widely accepted that the Earth condensed from a gas cloud about 4.5 billion years ago. The age of life on Earth is not precisely known but the earliest microfossils and stromatolites found in the Warrawoona formation are 3.5 billion years old. Naturally, living organisms must be older than this but the time for their development must be of the order of around 0.5 billion years if the probability that the Earth was molten early in its history is taken into account.331 While some authors have argued for panspermia as the origin of life on earth, this view is not widely held and it is generally acknowledged that life arose via a series of reactions in aqueous solutions containing various organic compounds, the ‘prebiotic soup’. This possibility was first advanced by the Russian biochemist Oparin, who outlined the notion of spontaneous generation of life on the primitive earth. Independently, Haldane333reached much the same conclusions and demonstrated that it was possible that significant concentrations of organic chemicals could have been built up in the early oceans. These ideas have been expanded upon by other Of course, the prevailing chemistry of the atmosphere and the oceans of the primitive earth is a highly contentious subject. The general inorganic coordination Chemistry outlined in previous sections of this review is still applicable, albeit that certain simple ligands such as cyanide or sulfide may have been present at higher concentrations, and other species, particularly more oxidized ones, were probably present at much lower levels. In fhe light of the coordination chemistry covered above, little more needs to be added on these subjects here. Rather, the role of coordination processes and complexes in the chemistry of formation of prebiotic organic compounds will be examined. Again it must be emphasized, however, that the significance of such reactions is speculative in the extreme, though it seems inescapable that coordination complexes must have played a prominent part. The reasons for thinking this are the experimental results amassed on the synthesis of organic compounds in the presence of metal ions under simulated prebiotic conditions, the synthesis of many classical ligands in related experiments, and the fact that it appears that living organisms incorporated coordination compounds in their metabolic cycles at an early date. This latter chemistry has persisted throughout evolution and coordination complexes are indispensable for all living organisms. The atmosphere of the primitive earth was probably highly reducing with carbon, nitrogen, hydrogen and oxygen being overwhelmingly present as methane, nitrogen, hydrogen and water, although not all authors have agreed with this view,336.337 because little carbon is found in the oldest known sedimentary rocks. On the other hand, it must be remembered that changes in the atmosphere to its being one in which carbon was present as C 0 2 could have occurred by the time that these rocks were laid down. Other compounds containing the above elements in the primitive earth were perhaps mineral carbides, water of crystallization in various mineral species and ammonium ion or ammonia. Much of the argument concerning the composition of the primitive atmosphere has been put in perspective in two excellent r e v i e w ~ . ~The ~ ~second * ~ ~ *draws upon our knowledge of other planetary atmospheres, the cornposition of meteorites and the known early deposition of large amounts of iron in Precambrian sediments. This, and particularly the first, points to the general conclusions recorded above, coupled with the fact that prebiotic generation of organics places reasonably tight restrictions on the atmospheric composition in any case. The prime energy sources for the reactions of interest are generally Of overwhelming importance is radiation from the sun. Of less significance is energy from electric discharges. Radioactivity, volcanic activity and shock waves together with cosmic rays and the solar wind probably made negligible contributions to the reactions in the prebiotic soup. Thus it is understandable that most experiments on synthesis in prebiotic systems have used UV or sparking sources in various mixtures. With these kinds of experiments in mind, Beck33yhas elegantly deduced some of the important simple coordination chemistry of the prebiotic soup. He suggests a primary ligand set of H20, NH3, CO, CN-, (CN)zC:-, S2-, H-, N2and CO,. Some of these species are of importance in reactions to form complex organic molecules, as discussed below. Cyanide complexes were possibly important in the primitive oceans, and experiments on extraction of elements from powdered rock samples have shown that significant concentrations of Fe, C o , Cu, Mn and Mo

Geochemical and Prebiotic Systems

87 1

can be generated by complexing and dissolution.340Earlier workers have also commented on cyanide complex f0rrnation,3~'.~~' and such species may be involved as catalysts in other reactions of prebiotic interest. Beck339also pointed out that 'it is inevitable to realise that exotic transition metal complexes, which can be prepared using specid laboratory techniques, could have been a common species on the primordial Earth.' This statement is undoubtedly true and unopen to challenge. However, the nature of such complexes is impossible to discern fully in the sense that as many complexes might be suggested as the imagination permits. The general chemistry of the simple ligand set mentioned above is known, and has been expounded upon at length elsewhere in this compendium of reviews. Typical, simple inorganic complexes, present in the primitive oceans, are easy to imagine, and their chemistry is accepted in this connection to be well established. One particular set of species, that involving sulfide ligands S:-, is perhaps worth mentioning in some greater detail. Given a reducing environment, these kinds of complexes may have been widespread. The connection between them, minerals such as pyrite, FeS,, and its congeners, and the iron-sulfur cluster enzyrne~,'~'now known to be basic to all present life-forms, is easily brought to mind. The evolution of iron-binding biomolecules has been reviewed.3M The first experimental evidence concerning the possible prebiotic synthesis of biorogical molecules was presented by Miller in 1953.345This study was in turn basad upon that of previous workers346who found that C 0 2 reacts to give aldehydes, among other products, in the presence of ionizing radiation. It was found that in a reducing atmosphere comprising CH4, NH3, H 2 0 and HZ,glycine, a- and &alanine, aspartic acid and a-aminobutyric acid were produced by the passage of electrical discharges through the reaction mixture. Aldehydes and HCN were also produced, as were hydroxy acids, short chain aliphatic acids and urea. This remarkable experiment has prompted an enormous amount of experimentation in the field, almost all of it patterned after the original work. It is intriguing that glycine in particular should be formed in such high yields as were found (about 2%, based on carbon) and that amino acids are formed as easily and abundantly in such simple systems. We must draw the inescapable conclusion that amino acids form the basis of life as we know it by virtue of their ease of formation under primitive Earth conditions. MiIler has recently reviewed the work in this field stretching over three de~ades.3~' Four mechanisms have been advanced for the prebiotic formation of amino acids. The first involves a cyanohydrin (reaction 2) and a related route (reaction 3) can be invoked to account for the presence of hydroxy acids. These particular reactions have been studied in considerable detail both kinetically and in terms of thermodynamic q ~ a n t i t i e s . ~An ~ ' alternative route (4) involves the hydrolysis of (Y -aminonitriles, which are themselves formed directly in anhydrous CH$ NH3 mixtures.348 Cyanoacetylene, formed in CH,/N2 irradiations,349 yields significant have amounts of asparagine and aspartic acids (reaction 5). Finally, a number of workers336~350-354 proposed that HCN oligomers, especially the trimer aminoacetonitrile and the tetramer diaminomaleonitrile, could have been important precursors for amino acid synthesis. Reaction mixtures involving such species have yielded up to 12 amino acids. Table 11 indicates the range of amino acids produced in these kinds of sparking syntheses. Of some interest is the fact that close parallels between these kinds of experiments and amino acid contents of carbonaceous chondrite meteorites e ~ i s t . ~ ~ ' , ~ ~ ~ , ~ ~ ~ RCHO+HCN+NH3 RCHO+HCN

+ RCH(NH,)CN % RCH(NH2)CONH2 !%

e RCW(0H)CN

RCH(OH)CONH,

CH,+ NH3 -+ H, NCH2CN + MeCH( NH&N NCCGCH+NH3

CN

-+

H O

NCCH=CHNH, + NCCH,CH(NH,)CN

(2)

RCH(NH,)CO,H

2 RCH(OH)C02H

(3)

H2NCH2C02H + MeCH( NH2)C02H

(4)

5

H2NOCCH,CH(NHJCO,H asparagine

2 H0,CCH2CH(NH,)C0,

H

aspartic acid

(5)

The role played by coordination complexes as templates and catalysts in these reactions is somewhat more difficult to discern, although several important results have been reported. Some early work by Bahadu? showed that amino acids were formed in the reaction of potassium nitrate with paraformaldehyde in the presence of Iight. Many amino acids were produced, but the highest yields were found when the reaction was carried out in the presence of iron(lI1) chloride as a catalyst. Later extended analogous findings to colloidal molybdenum oxide as the catalyst. Molecular nitrogen was fixed in this case, a reaction in a sense comparable to the

XI2

Geochemical and Prebiotic Systems Table 11 Amino Acids Formed in Sparking Experiments with Mixtures of CH,, N,, H,O and NH, 331

G1y cine Alanine a-Amino-n-butyric acid a-Aminoisobutyric acid Valine Norvaline Isovaline Leucine Isoleucine Alloisoleucine Norleucine r-Leucine Pro1i ne Aspartic acid Glutamic acid Serine Threonine Allothreonine

a,y-Diaminobutyric acid a-Hydroxy-y-aminobutyricacid

a$-Diaminopropionic acid Isoserhe Sarcosine N-Ethylglycine N-Propylglycine N- Isopropylglycine N-Methylakdnine N - Ethylalanine ,%Alanine @- Amino-n-butyric acid P- Arninoisobutyric acid y-Aminobutyric acid N- Methyl-P-alanine N-Ethyl-@-alanine Pipecolic acid

well-known role of molybdenum-containing enzymes in the reduction of molecular nitrogen to ammonia in leguminous plants. Other found similar results in the reaction of formaldehyde and hydroxylamine with added molybdate and vanadate. The yield of amino acids increased exponentially with added molybdate. Egami has most elegantly outlined the reasoning behind the accepted role of transition metal complexes in prebiotic syntheses involving FischerTropsch-type reaction^.^^'"^^ The trace metal composition of primitive seawater, especially concerning elements such as iron, copper, molybdenum and zinc, would be expected to influence syntheses and result in the incorporation of such elements into protoenzymes. Although some experiments involving Ni, A1203 and clay catalysts with CO, H2 and NH3 gave inconclusive has shown that the original suggestions were about the role of the metals, later most probably correct. Formaldehyde and hydroxylamine were reacted in a synthetic seawater in the presence of Zn2+, Moo.,-, Fe3+,Co2+,Cu2+ and Mn2+. Some 40 amino acids were detected in the products with glycine, alanine, p -alanine and norvaline being among the most concentrated. Of particular interest was the fact that peptides were also produced which also had some hydrolytic activity. The authors stated in view of this truly remarkable finding that 'this protoenzyme-like activity is in good correlation with the consideration by Egami that the zinc complex in the early stage of evolution may be regarded as a precursor of hydrolytic and transferring enzymes, including enzymes participating in the metabolism of macromolecules and information transfer'. Reactions of coordinated amino acids and their precursors undoubtedly took place in the prebiotic soup. While it is impossible to assess the importance of such processes, we should allow ourselves a little speculation here, and a few examples are included, all of which are based on sound laboratory findings. Beck339observed the formation of glycine by the acid hydrolysis of the complex H2Cu2(CN)4(C2N2). It is possible that similar species might be involved in the prebiotic production of other amino acids. Serine and threonine together with allothreonine can be easily by the reactions of coordinated glycine with formaldehyde or acetaldehyde (Scheme 3). Similar reactions to form ornithine and arginine are known.3w3367 Finally, reactions on clay surfaces, involving coordination complexes of amino acids and other ligands, may have elaborated the speciation of the prebiotic soup. Early work in this field has been gathered together by Weiss in an excellent review.368Under suitable conditions amino acids can polymerize in clays (and on SiO?), among other important reactions. A recent report369has described the alteration of various aromatic amino acids, especially tyrosine, on kaolinite, bentonite and montmorillonite. Oxidative degradation occurs at pH greater than 3 and the possible role of clays in the prebiotic synthesis of polypeptides was assessed. The workers also found that (S)-tyrosine reacted faster than (R)-tyrosine, a result which contradicts the laws of thermodynamics, and which may have arisen as a result of bacterial contamination. A related result3" concerning the supposed difference in stoichiometry of Fe"' and Fe" complexes of, separately, ( R ) - and (S)-alanine, with a view towards the generation of optical activity in prebiotic systems is absolutely in error. This particular field of the origin of optical activity is the subject of much controversy"'~"* and a convincing explanation for the origin of handedness in biological systems is still awaited. Perhaps one of the most important roles of clays and coordination complexes in the synthesis of prebiotic compounds, apart from their catalytic influences, is one of protection. Mention has

Geochemical and Prebioiic Systems

873

Scheme 3

been made earlier of the preservation of porphyrins by complexation, and it is possible that amino acids, among other species, were protected from other degradative reactions by the formation of complexes, and incorporation into clay minerals at the bottom of the primitive oceans. Much work remains to be carried out on these amino acid/complex systems, however, before any more definite conclusions can be reached. Considerable attention has been focussed upon possible amino acid prebiotic syntheses in this section, because of their biological importance. It should be pointed out that many other compounds of biological significance have also been made under simulated primitive conditions and it is proposed to conclude this review with a brief examination of such reactions, particularly with respect to the possible role of coordination compounds in them. Porphyrin complexes can be commonly produced in sparking mixtures. Thus, Hodgson and Baker373showed that pyrrole and paraformaldehyde in the presence of copper(II), nickel(I1) and vanadyl salts give rise to both free and metal complexes of porphyrins. The most effective ion for the template reaction is Ni”, a result parallelling the observations outlined earlier on the preservation of porphyrin complexes in asphalts, sediments and rocks. It was pointed out that coloured species noticed in earlier experiments374were also probably porphyrin coordination that Ni(CN):increases the yield of complexes. More recently it has been porphyrins in such reactions, as does Fe(CN)Q-. Since these ions were probably present in reasonable concentrations in the primitive oceans, their significance is obvious. Nucleobase formation in a Fischer-Tropsch synthesis was first reported by H a y a t ~ u ’using ~~ a mixture of CO, NH3 and H2. This discovery was foreshadowed and expanded upon by a number of other workers and involves reactions of HCN or cyano c o m p l e ~ e s . The ~ ~ ~later - ~ ~work ~ by Hayatsu and coworkers381using Ni-Fe and M203catalysts is noteworthy. Alkyl cyanides, pyrroles, porphyrins, guanidines, hydantoin, uracil and derivatives, thymine, adenine, guanine, xanthine, melamine, alkanes, alkenes and aromatic hydrocarbons were all produced from the simple mixture originally used. It has also been shown that magnesium complexes in reactions involving urea and imidazole activate nucleotides and amino acids to form Intriguingly, it has recenty been that Zn2+ and Pb2+ catalyze the polymerization o f nucleotides with the production of oligomers some 30-40 units in length. The significance of these findings with respect to the prebiotic evolution of RNA polymerase was discussed. It is notable that all DNA and RNA are Zn2+-containing enzymes. This research provides a vital clue to the early development of replicating systems. Finally, the synthesis of sugars under prebiotic conditions has been achieved. This was really foreshadowed by the classic work of B ~ t l e r o v ~who ’ ~ demonstrated that pentoses and hexoses are formed by condensation of formaldehyde in basic aqueous solution. The process depends upon the presence of a suitable catalyst, usually Ca2+ derived from Ca(OH), or CaCO,, and alkaline earth complexes are most probably involved. Later workers, turning their attention towards prebiotic chemistry in particular, have shown that clays and apatite catalyze the formation of monosaccharides, especially ribose, with good yields being Thus all of the main molecules of life, or their simple precursors and analogues, have been synthesized under conditions thought to approximate the primitive Earth. The role coordination complexes plays in these processes must be highly significant as evidenced by a number of experimental findings. Much more work remains to be done in this field, and on the involvement of mineral-organic interactions in the prebiotic soup. The role of minerals in these evolutionary pathways might be much more important than has previously been t h o ~ g h t . ” ~ 64.7 REFERENCES 1. K. B. Krauskopf, ‘Introduction to Geochemistry’, 2nd edn., McGraw-Hill, Tokyo, 1979, 2. C. Palache, H. Berman and C . Frondel, ‘The System of Mineralogy’, 7th edn., Wiley, New York, 1951. 3. M. H. Hey, ‘An Index of‘ Mineral Species and Varieties‘, 2nd edn., British Museum (Natural History), London, 1962; 1st Appendix: Trustees of the British Museum (Natural History), London, 1963; 2nd Appendix (with P. G. Embrey): Trustees of the British Museum (Natural History), London, 1974.

Geochemical and Prebioric Systems

874

H. Strunz, ‘Mineralogische Tabellen’, 5th edn., Akademische Verlagsgesellschaft Geest and Portig K.-G.,Leipzig, 1970. M. Fleischer, ‘Glossary of Mineral Species 1980’, 3rd edn., Mineralogical Record, Tucson, 1980. W. L. Roberts, G. R. Rapp, Jr. and J. Weber, ‘Encyclopedia of Minerals’, Van Nostrand Reinhold, New York, 1974. P. G. Embrey and J. P. Fuller, ‘A Manual of New Mineral Names 1892-1978’, British Museum (Natural History) and Oxford University Press, London and Oxford, 1980. 8. ‘Mineralogical Abstracts’, ed. R. A. Howie, The Mineralogical Society of Great Britain and the Mineralogical Society of America, London, Quarterly. 9. ‘Mineralogical Magazine’. ed. A. M. Clark, The Mineralogical Society of Great Britain, London, Quarterly. 10. C. Tennyson, Fortschr. Mineral., 1963, 41, 64. 11. Hsieh Hsien-te, Sci Abstr. China, Earth Sci., 1966, 4, 6 . 12. W. A. Deer, R. A. Howie, and J. Zussman, ‘Rock Forming Minerals’, 2nd edn., Longman, London, 1982-1985, vols. 4. 5. 6. 7.

1-5.

13. D. McKie and C. McKie, ‘Crystalline Solids’, NeIson, London, 1974. 14. A. F. Wells, ‘Structural Inorganic Chemistry’, 4th edn., Clarendon, Oxford, 1975. 15. J. J. Papike and M.Cameron, Rev. Geophys. Space phys., 1976, 14, 37. 16. D. J. Vaughan and J. R.Craig, ‘Mineral Chemistry of Metal Sulphides’, Cambridge University Press, Cambridge, 1978. 17. H.-J. Bautsch, Wiss. Z. Hurnboldt-Univ. Berlin, Math.-Nahtrwiss. Reihe, 1967, 16, 783. 18. R. G. Burns, ‘Mineralogical Applications of Crystal Field Theory’, Cambridge University Press, Cambridge, 1970. 19. R. G. Burns, Geochim. Cosmochim. Acta, 1973, 37, 2395. 20. R. G. Burns, Geochim. Cosrnochim. Acta, 1975, 39, 857. 21. P. Henderson, ‘Inorganic Geochemistry’, Pergamon, Oxford, 1982. 22. F. C. Hawthorne and R B. Ferguson, Can Mineral., 1975, 13, 377. 23. E. G. Steward and H. P. Rooksby, Acta Clystallogr., 1953, 6, 49. 24. G. Cocco, P. C. Castiglione and G. Vagliasindi, Acta Crystallogr., 1967, 23, 162. 25. G. Giuseppetti and C. Tddini, Tschermah Mineral. Petrugr. Mitt., 1978, 25, 57. 26. S. Geller, Am. Mineral, 1971, 56, 18. 27. Z. V. Pudovkina and Yu. A. Pyatenko, Dokl. Acad. Sci. USSR, Earth Sci. Sect. (Engl. Transl.), 1970, 190, 131. 28, 2,V. Pudovkina and Yu. A. Pyatenko, Dokl. Acad. Sci. USSR, Earth Sei. Sect. (Engl. Trans%), 1967, 174, 193. 29. M. Fleischer, Am. Mineral., 1962,47, 805. 30. F. H. Stewart, Mia Mag., 1951, 29, 557. 31. G. Raade and J. Haug, Mineral. Rec., 1980, 11, 83. 32. P. A. Foster, Jr., J. Am. Ceram. Soc., 1975, 58, 288. 33. D. P. Stinton and J. J. Brown, Jr., J. Am. Ceram. Soc.. 1976, 59, 264. 34. F. Zambonini, BOIL SOC.Geol. Ital., 1931, 49, 179. 35. G. Carobbi, Period. Mineral., 1933, 4, 410. 36. F. Zambonini and G. Carobbi, Atti Accad. Naz. Lincei, Cl. Sci. Fis., Mat. Nat., Rend., 1926, 4, 172. 37. L. Chrobak, 2. Kristallogr., 1934, 88, 35. 38. W. W. Crook, 111 and L.-A. Marcotty, A n Mineral., 1978, 63, 410. 39. R. C. Rouse, J. Solid State Chem., 1973, 6, 86. 40. M. C. Baird, Prog. Inorg. Chem., 1968, 9, 1. 41. A. Preisinger, Tschermaks Mineral. Petrogr. Mitt., 1953, 3, 376. 42. A. S. Povarennykh and L. D. Rusakova, Geol Zh. (RUSS. Ed.), 1973, 33, 24 (Min. Abstr., 1974, 25, 74-508). 43. E. Hayek and P. Lnama, Monatsh. Chem., 1965, 96, 1454. 44. H. Heritsch, Anz. Oesterr. Akad. Wiss., Math.-Natunvirm KL, 1954, 1. 45. G. Switzer, W. F. Foshag, K. J. Murata and J. J. Fahey, A m . Mineral., 1953,38, 1225. 46. G. Tunell, J. J. Fahey, F. W. Daugherty and G. V. Gibbs, Neues Jahrb. Mineral. Monarsh., 1977, 119. 47. R. C. Rouse, P. J. Dunn and D. R. Peacor, Mineral. Rec., 1982, 13, 233. 48. .P. J. Dunn, Mineral. Rec., 1981, 12, 49. 49. K. Mereiter, Tschennaks Mineral. Petrogr. Mitt., 1982, 30, 277. 50. A. Coda, A. Della Giusta and V. Tazzoli, Acta Crystallogr., Secr. B, 1981, 37, 1496. 51. F. M a n i and F. Rinaldi, Period. Mineral., 1961, 30, 1. 52. K. Mereiter, T s c h e m a h Mineral. Petrogr. Mitt., in press. 53. A. Dal Negro, G. Rossi and V. Tazzoli, A m Mineral., 1977, 62, 142. 54. V. Perttunen, BulL Geol. SOC.Finl., 1970, 43, 67. 55. A. Dal Negro, G. Rossi and V. Tazzoli, Am Mineral, 1975, 60,280. 56. G. Donnay and I. D. H. Donnay, Tschermaks Mineral. Petrogr. Mitt., 1971, 15, 201. 57. J. D. Grice and G. Perrault, Can. Mineral., 1975, 13, 209; T. T. Chen and G. Y. Chao, ibid., 1975, 13, 22. 58. J. Sawyer, P. Caro and L. Eyring, Rev. Chim. Miner., 1973, LO, 93. 59. C. Frondel and R. J. Gettens, Science, 1955, 122, 75. 60. T. Isaacs, Min. Mag., 1963, 33, 663. 61. J. L. Jambor and R W.Boyle, Can. Mineral., 1965, 8, 166. 62. C. L. Garavelli, Atti Accad. Naz. Lincei, Cl. Sci. Fis., Mat. Nat., Rend., 1964, 36, 1. 63. B. Geier and J. Otteman, Neues Jahrb. Mineral Abh., 1970, 114, 89. 64. N. K. Marshukova, A. P. Palovskii, G. A. Sidorenko and N. I. Chistyakova, Zap. Vses. Mineral Om., 1981, 110, 492 (Min. Abstr., 1982, 33, 82M-182). 65. P. J. Bridge, Min. Mag., 1974, 39, 889. 66. E. Matzat, Acra Crystallogr., Sect. E, 1972, 28, 415. 67. C. Frondel, Am. Mineral., 1967, 52, 617. 68. M. Fleischer, Am Mineral., 1958, 43, 382. 69. M. Dee and G. F. Mitchell, Sci. F’roc. R. Dublin Soc., 1946, 24, 85. 70. J. D. Vine, V. E. Swanson and K. C. Bell, Proc, U N Int, Con$ Peaceful Uses At. Energy, 2nd, 1958, 2, 187. 71. E, L. Prien and C. Frondel, J. Urol, 1947, 57, 949. 72. S. Carie, Bull. Soc. Fr. Mineral. Cristallogr., 1959, 82, 50.

Geochemical and Prebiotic Systems

875

73. C. L. Craravelli, Atti Accad. Naz. Lincei, C1. Sci. Fix, Mat. Naf., Rend., 1955, 18, 392. 74. Yu. N. Knipovich, A. I. Komkov and E. I. Nefedov, Tr. Vses. Nauchno-Issled. GeoL hst., 1963, new ser., no. 96, 131 (Min. Abstr., 1963, 4, 16, 551). 75. E. A. Jobbins, G. A. Sergeant and B. R. Young, Min. Mug., 1965, 35, 542. 76. F. A. Bannister, Discovery Rep., Cambridge, 1936, 13, 67. 77. C. Milton, E. J. Dwornik, P. A. Estep-Barnes, R. B. Finkleman, A. Pabst and S. Palmer, Am. Mineral., 1978, 63, 930. 78. M. DtribBri, Bull. Soc. Fr. Mineral. Cristallogr., 1961,84, 94. 79. H. Haberlandt, Wien. Chem-Ztg., 1944, 47, 80. 80. C. J. Kelly, Jr., Am. Mineral, 1970, 55, 2118. 81. M. Louis, J.-C. Guillemin, J.-C. Coni and J.-P. Ragot, Proc. Inr. Meet. 0%.Geochem., 4th, 1968, 533. 82. F. W. Clarke, Geol. Surv. BUZZ.(LIS), 1924, 770. 83. V. M. Goldschmidt, ‘Geochemistry’, Clarendon, Oxford, 1954. 84. K. B. Krauskopf, Econ. Geol., 1951,46, 858. 85. K. B. Krauskopf, Eeon. GeoZ., 1951,46, 498. 86. R. M. Carrels and C. L. Christ, ‘Solutions, Minerals, and Equilibria’, Harper and Row, New York, 1965. 87. W. Stumm and J. J. Morgan, ‘Aquatic Chemistry’, 2nd edn., Wiley, New York, 1981. 88. J. D. Hem, Wafer Resour. Res., 1972, 8, 661. 89. J. D. Hem, Geochim Cosmochim. Acta, 1976,40, 599. 90. J. Vuceta and J. J. Morgan, Environ. Sci. Technol., 1978, 12, 1302. 91. J. A. Davis and J. 0. Leckie, Environ. Sci. Technol., 1978, 12, 1309. 92. F. A. Cotton and G. Wilkinson, ‘Advanced Inorganic Chemistry’, 4th edn., Wiley-Interscience, New York, 1980. 93. S. Ahrland, Struci. Bonding (Berlin), 1968, 5, 118; 1973, 15, 167. 94. J. Burgess, ‘Metal Ions in Solution’, Ellis Horwood, Chichester, 1978. 95. L. G. SillBn and A. E. Martell, ‘Stability Constants of Metal-ion Complexes’, The Chemical Society, London, 1964; Supplement No. 1, The Chemical Society, London, 1971. 96, R. M. Smith and A. E. Martell, ‘Critical Stability Constants’, Plenum, New York, 1976. 97. C. F. Baes and R. E. Mesmer, ‘The Hydrolysis of Cations’, Wiley, Mew York, 1976. 98. D. R. Turner, M. Whittield and A. G. Dickson, Geochim. Chosmochirn. Acta, 1981, 45, 855, 99. R. E. Cranston and J. W. Murray, A n d Chirn. Acta, 1978, 99, 275. 100. J. J. Morgan, in ‘Principles and Applications of Water Chemistry’, ed. S. D. Faust and J. V. Hunter, Wiley, New York, 1967, p. 561. 101. D. Lesht and J. E. Bauman, Irrorg. Chem., 1978, 17, 3332. 102. W. Davison, Geochim Cosmochim. Acta, 1979, 43, 1693. 103. B. Cosovi6, D. Degobbis, H. Bilinski and M. Branica, Geochim. Cosmochim. Acta, 1982, 46, 151. 104. D. T. Long and E. E. Angino, Geochim. Cosrnochim. Acta, 1977, 41, 1183. 105. A. M. Posner and J. P. Quirk, Aust. J. Soil Res., 1981, 19, 309. 106. L. R. Pitwell, Chem. Geol., 1973, 12, 39. 107. H. W. Lakin, G. C. Curtin, A. E. Hubert, H. T. Shacklette and K. G. Doxtader, Geol. Surv. BuH. ( U S ) , 1974, 1330. 108. W. H. Emmons, Geoi. Suw. Bull, (US), 1917, 625. 109. B. I. Peshchevitskiy, G. N. Anoshin and A. M. Yerenburg, Dokl. Acad Sci. USSR, Earth Sci. Sect. (EngZ. Trans!.), 1965, 162, 205. 110. E. J. Reardon, Chem. Geol., 1979, 25, 339. 111. P. Halbach, R. Giovanoli and D. von Borstel, Earth Plunef. Sci Lett., 1982, 60, 226. 112. S.-I. Wada and K. Wada, Soil Sci., 1981, 132, 267. 113. D. Langmuir, Geochim. Cosmochim. Acta, 1978, 42, 547. 114. G. Dongarra and D. Langmuir, Geochim. Cosmochim. Acta, 1980,44, 1747. 115. D. Langmuir and J. S. Herman, Geochim. Cosmochim. Acta, 1980, 44,1753. 116. L. Navolic and G. Pedro, C.R. Hebd. Seances Acad. Sci., Ser. D, 1976, 282, 1765, 1917. 117. H. L. Barnes (ed.), ‘Geochemistry of Hydrothermal Ore Deposits’, 2nd edn., Wiley, New York, 1979. 118. E. Roedder, ‘Composition of Fluid Inclusions’, Geol. Sum, Prof: Pap. ( U S ) ,1972,44055. 119. D. M. Pinckney and J. HaiTty, Econ. Geol., 1970, 65, 451. 120. F. J. Sawkins and D. A. Scherkcnbach, Geology, 1981, 9, 37. 121. D. E. White, Econ. Geol, 1968, 63, 301. 122. T. H. Giordano and H. L. Barnes, Econ. Geol., 1981, 76, 2200. 123. R. J. Skinner, D. E. White, H. J. Rose and R. E. Mays, Econ. Geul, 1967, 62, 316. 124. I. N. Goryainov and T. B. Andreyeva, Dokl. Acud. Sci. U S S e Earth Sci. Sect. (Engl. Trund), 1972, 204, 224. 125. D. J. Mossman and K. J. Heffernan, Chem. Geol., 1978, 21, 151. 126. T. J. Barrett and G. M. Anderson, Econ. Gebl., 1982, 77, 1923. 127. L. M. Lebedev and I. B. Mikitina, Dokl. Acad. Sci. USSR, Earth Sci. Sect. (Engl. Transl.), 1971, 197, 229. 128. P. Tsai and R. P. Cooney, Chern. Geol., 1976, 18, 187. 129. T. M. Seward, in ‘Chemistry and Geochemistry at High Temperatures and Pressures’, ed. D. T. Rickard and F. E. Wickman, Pergamon, Oxford, 1981. 130. A. W. Rose, Econ. Geol., 1976,71, 1036. 131. D. T. Rickard, Stockholm Coptrib. Geol., 1970, 23, 1. 132. N. E. Cohen, R. R. Brooks and R. D. Reeves, NZ J. Geol. Geaphys., 1967, 10, 732. 133. M. A. Korasik, Yu. T. Goncharov and A. E. Vasilevskaya, Geochem. Int., 1965, 2, 82. 134. V. V. Shcherbina, Geochem. Int., 1967, 4, 1104. 135. S. B. Romberger and H. L. Barnes, Econ. Geol., 1970, 65, 901. 136. T. M. Seward, Geochirn. Cosmochim. Acta, 1973, 37, 379. 137. 1. Ya. Nekrisov and A. A. Konyushok, Minerul. Zh., 1982, 4, 33 (Min. Absrr., 1983,34, 83M/0378). 138. J. BoulBgue, Geochim. Cosrnochirn. Acta, 1977, 41, 1751. 139. J. Bouligue, P. Genest and G. Michard, C.R. Hebd. Seances Acad Sci., Ser. D, 1976, 282, 145. 140. P. L. Cloke, Cieochirn. Cosrnochim. Acta, 1963, 27, 1299.

876

Geochemical and Prebiotic Systems

141. J. Boulegue, C. J. Lord, 111 and T. M. Church, Geochim. Cosmochim. Acta, 1982, 46, 453. 142. H. C. Helgeson, ‘Complexing and Hydrothermal Ore Deposition’, Pergamon, Oxford, 1964. 143. E. Froese, GeoL Sum.Pap. (Geol. Sum. Can.), 1981, 80-28. 144. M. H. Reed, Geochim. Cosmochim. Acta, 1982,46, 513. 145. H. C. Helgeson, D. H. Kirkham and G. C. Flowers, Am J. Sci, 1981, 281, 1249. 146. R. W. Henley, ?%hi. - Nat. Environ. Res. Counc., Ser. D ( U K ) , 1972,2, 53. 147. I.-M. Chou and H. P. Eugster, Am. J. Sci., 1977, 277, 1296. 148. N. Z. Boctor, R. K. Popp and J. D. Frantz, Geochim. Cosmochim Acta, 1980,44, 1509. 149. F. Guichard, T. M. Church, M. Treuil and H. Jaffrezic, Geochim. Cosmochim. Acta, 1979,43, 983. 150. I. G. Khel’vas and G. G. Grushkin, Zap. Vses. Mineral. Ova., 1974, 103, 670 (Min. Abstr., 1976, 27, 76-323). 151. V. V. Shcherbina, Geochemistry, 1962, 1069. 152. 1. N. Govorov and A. A. Stunzhas, Geochemistry, 1963, 402. 153. V. V. Shcherbina and Sh. A. Abakirov, Geochem Int, 1967, 4, 165. 154. S. M. McLennan and S. R. Taylor, Nature (London), 1979,282,247. 155. N. C. Higgins, Can. J. Earth Sei., 1980, 17, 823. 156. A. H. Truesdell and B. F. Jones, J. Res. US Geol. Sum., 1974, 2, 233. 157. A. H. Truesdell and W. Singers, J. Res. US GeoL Sun?, 1974, 2, 271. 158. K. J. Murata, Am. 1 Sci., 1960, 258, 769. 159. A. H. Delsemme, BulL Acad. R. Sci. d’Outre-Mer, 1960, 6 , 507. 160. H. Tazieff, BUN VoZcanoL, 1960, 23, 69. 161. S. I. Naboko, Bull. VolcanoZ., Ser. 2, 1959, 20, 121. 162. Y. Mizutani, Geochem. .I 1970,4, . , 87. 163. W. R. Hesp and D. Rigby, Pac. Geol., 1972, 4, 135. 164. A. L. Walker and A. S. Buchanan, Econ. Geol., 1969, 64, 923. 165. S. A. McGough, Year Book - Carnegie Inst. Washington, 1982, 81. 166. B. H.W. S. de Jong, C. M. Schramm and V. E. Parziale, Geochim. Cosmochim. Acta, 1983,47, 1223. 167. D. R. Gaskoll, in ‘Advances in Physical Geochemistry’, ed. S. K. Saxena, vol. 2, Springer-Verlag, Berlin, 1982. 168. B. 0. Mysen and D. Virgo, Geochim. Cosmochirn. Acta, 1980, 44, 1917. 169. M. 1. Wood and P. C. ness, Contrib. Mineral. Petrol., 1980, 72, 319. 170. J. D. Bernal, Nature (London), 1960, 185,68. 171. E. J. W. Whittaker, J. Non-Cryst. Solids, 1978, 28, 293. 172. R. G. Bums and W. S . Fyfe, Science, 1964, 144, 1001. 173. J. A. Boon and W. S. Fyfe, Chem. Geol., 1972,10, 287. 174. K. W. Sernkow, R. A. Rizzo, L. A. Haskin and D. J. Lindstrom, Geochim. Cosmochim. Acta, 1982,46, 1879. 175. M. P. Dickenson and P. C. Hess, Contrib. Mineral. Petrol., 1981, 78, 352. 176. D. Virgo, B. 0. Mysen, P. Danckwerth and F. Seifert, Year Book - Carnegie Inst. Washington, 1982, 81, 349. 177. E.-R. Neumdnn, B. 0. Mysen, D. Virgo and F. A. Seifert, Year Book - Carnegie Inst. Washington, 1982, 81, 353. 178. C. Nelson and W. B. White, Geochim. Cosmochim. Acta, 1980,44, 887. 179. H. J. M. Bowen, ‘Trace Elements in Biochemistry’, Academic, London, 1966. 180. J. A. Kornfeid, in ‘Advances in Organic Geochemistry’, ed. U. Colombo and 0.D. Hobson, Pergamon, Oxford, 1964, p. 261. 181. K. 8. Krauskopf, Econ. GeoZ., 1955, 50,411. 182. K. G. Re11 and J. M. Hunt, in ‘Organic Geochemistry’, ed. I. A. Breger, Pergamon, Oxford, 1963, p. 333. 183. W. L. Whitehead and I. A. Breger, in ref. 182, p. 248. 184. H. J. Hyden, Geol. Surv. Bull. ( U S ) , 1961, 1100. 185. F. M. Swain, in ref. 182, p. 87. 186. J. D. Saxby, Rev. Pure Appl. Chem., 1969, 19, 131. 187. E. K. Duursma, in ‘Chemical Oceanography’, ed. J. P. Riley and G. Skirrow, Academic, London, 1965, vol. 1, chap. 11. 188. R. F. C. Mantoura, A. G. Dickson and J. P. Riley, Estuarine Coastal Mar. Sci., 1978, 6 , 387. 189. K. Irgolec and A. E. Martell (eds.), ‘Ehvironmental Inorganic Chemistry’, VCH Publishers, Deerfield Beach, FL, 1985. 190. R. F. Christman and E. Gjessing (eds.), ‘Aquatic and Terrestrial Humic Materials’, Ann Arbor Science, Ann Arbor, MI, 1983. 191. J. H. Reuter and E. M. Perdue, Geochim. Cosmochim Acta, 1977, 41, 325. 192. K. S. Jackson, I. R. Jonasson and G. B. Skippen, Earth-Sci. Rev., 1978, 14, 97. 193. V. S. h t i i i n and 1. M. Varentsov, Chem. Erde, 1980, 39, 298. 194. M. A. Rashid and 3. D. Leonard, Chem. Geol., 1973, 11, 89. 195. A. Nissenbaum and D. J. Swaine, Geochim. Cosrnochim. Acta, 1976,40, 809. 196. A. Szalay, Ark Mineral. GeoZ., 1974, 5, 23. 197. I. A. Breger, Fuel, 1951, 30,204. 198. M. Schnitzer and S. U. Khan, ‘Humic Substances in the Environment’, Dekker, New York, 1972. 199. P. R. Bloom and M. B. McBnde, J. Soil SOC.A m , 1979,43, 687. 200. E. M.Perdue, J. H. Reuter and M. Ghosal, Geochim. Cosmochim. Acta, 1980, 44.1841. 201. M. A. Rashid, Chem Geol., 1972,9, 241. 202. S. M. Manskaya and L. A. Kodina, Geochemistv, 1963, 389. 203. R. F. Christman and R. A. Minear, in ‘Organic Compounds in Aquatic Environments’,.ed. S. J. Faust and J. V. Hunter, Dekker, New York, 1971, p. 119. 204. C. H. Langford, D. S. Gamble, 4. W. Underdown and S. Lee, in ref. 190, p. 219. 205. M. A. Wilson, P. F. Barron and A. H. Gillam, Geochim. Cosmochim.Acta, 1981,45, 1743. 206. F. J. Stephenson and J. H. A. Butler, in ‘Organic Geochemistry’, ed. G. Eglinton and M. T. Murphy, Longman, London, 1969. 207. T.C. Hoering, Year Book - Carnegie Inst. Washington, 1973,72, 682. 208. M. V. Cheshire, M. L. Berrow, B. A. Goodman and C. M. Mundie, Geochim. Cosmochim. Acta, 1977,41, 1131.

Geochemical and Prebiotic Systems

87 7

209. A. L. Abdul-Halim, J. C. Evans, C. C. Rowlands and J . H. Thomas, Geochim. Cosmochim. Acra, 1981,45, 481. 210. E. Tipping and D. Cooke, Geochim. Cosmochim. Acta, 1982,46,75. 211. R. D. Guy, C. L. Chakrabarti and L. L. Schrdmrn, Can. J. Chem, 1975,53, 661. 212. L. W. Lion, R. S. Altman and J. 0. Leckie, Enoiron. Sci. TechnoL, 1982, 16, 660. 213. V. Cheam, Can. L SoiISci, 1973,53, 377. 214. J. J. Alberts, J. E. Schindler, D. E. Nutter, Jr. and E. Davis, Geochim. Cosmochim. Acta, 1976, 40, 369. 215. M. B. McBride, Soil Sd,1978, 126, 200. 216. H. Ker n d o d and M. Schnitzer, Geochim. Cosmochim. Acta, 1980, 44, 1701. 217. C. Juste, C.R. Hebd. Seances Acad. Sci., Ser. D, 1966,262, 2692. 218. K. C. Beck, J. H. Reuter and E. M. Perdue, Geochim. Cosmochim Acta, 1974, 38, 341. 219. E. M. Perdue, K. C. Beck and J. H. Reuter, Nature (London), 1976, 260,418. 220. N. Senesi, S. M. GriEth, M. Schnitzer and M.G. Townsend, Geochirn Cosmochim. Acta, 1977,41, 969. 221. M. SzilLgyi, Soil Sci, 1973, 115, 434. 222. T. L. Tbeis and P. C. Singer, Enoiron. Sei. Technol., 1974, 8, 569. 223. D. P. E. Dickson, L. Beller-Kallai and 1. Rozenson, Geochim. Cosmochim Acta, 1979,43, 1449, 224. N. S. Randhawa and F. E. Broadbent, Soil Sci., 1965, 99, 295. 225. K. S. Jackson and G. 3. Skippen, J. Geochem. Explor., 1978, 10, 117. 226. R. A. Saar and J. H. Weber, Geochim. Cosmochim. Acta, 1980, 44,1381. 227. M. Schnitzer and E. H. Hansen, Soil Sci., 1970, 109, 333. 228. D. S. Gamble, M. Schnitzer and I. Hoffman, Can. J. Chem., 1970, 48, 3197. 229. P. G. Manning and S. Ramamoorthy, J. Inorg. Nucl. Chem., 1973, 35, 1577. 230. P. Halbach, D. von Borstel and K.-D. Gundermann, Chem Geol, 1980,29, 117. 231. A. Szalay and M. Sziligyi, Geochim Cosmochim. Acta, 1967, 31, 1. 232. S. A. Wilson and J. H. Weber, Chem. Geol., 1979, 26, 345. 233. G. D. Templeton, III and N. D. Chasteen, Geochim. Cosmochim. Acta, 1980, 44, 741. 234. B. A. Goodman and M. V. Cheshire, Geochim. Cosmochim. Acta, 1975, 39, 1711. 235. M. Sziligyi, Geochem. hi., 1967,4, 1165. 236. G. Eskenazy, Chem. Geol., 1977,19, 153. 237. W. A. Fuchs and A. W. Rose, Econ. Geol. 1974,69, 332. 238. G. Eskenazy, Yearb. Univ. Sojia, Geol., 1974, 66, 279. 239. H. L. Ong and V. E. Swanson, Colo. Sch. Mines Q., 1969, 64, 395. 240. W. E. Baker, Geochim Cosmochim. Acta, 1978,42, 645. 241. J.-R. Disnar and J. Trichet, Geochim. Cosmochim. Acta, 1981,45, 353. 242. J.-R. Disnar, Geochim. Cosmochim. Acta, 1981,45, 363. 243. S. Ramamoorthy and P.C.Manning, J. Inorg. Nucl. Chem., 1974, 36,695. 244. Ng Siew Kee and C. Bloomfield, Geochim. Cosmochim Acta, 1961, 24, 206. 245. H. L. Ong, V. E. Swanson and R. E. Bisque, Geol. Sum. Pro$ Pap. (US), 1970,700-C, 130. 246. W. E. Baker, Geochim. Cosmochim. Acta, 1973, 37, 269. 247. A. Treibs, Justus Liebigs Ann. Chem., 1934,510, 42. 248. A. Treibs, Angew. Chem., 1936,49, 682. 249. E. S. Barghoorn, W. G. Meinschein and J. W. Schopf, Science, 1965, 148,461. 250. A. J. G. Barwise and E. V. Whitehead, in ‘Advances in Organic Geochemistry’, ed. A. G. Douglas and J. R. Maxwell, Pergamon, Oxford, 1979, p. 181. 251. G. W. Hodgson, B. Hitchon, K. Taguchi, B. L. Baker and E. Feake, Geochim. Cosmochim. Acra, 1968,32, 737. 252. H. N. Dunning and J. W. Moore, Am. Assoc. Pet. Geol. SulL, 1957, 41, 2403. 253. G. W. Hodgson, N. Ushijima, K. Taguchi and I. Shimada, Sci Rep., Tohoku Unio., Ser. 3, 1963,8,483. 254. E. W. Baker, J. Am. Chem Soc., 1966, 88, 2311. 255. D. W. Thomas and M. Blumer, Geochim. Cosmochim. Acta, 1964, 28, 1147. 256. J. R. Morandi and H. B. Jensen, J. Chem. Eng. Data, 1966, 11, 81. 257. E. W. Baker, in ref. 206, p. 464. 258. G. W. Hodgson and E. Peake, Nature (London), 1961, 191, 766. 259. G. Eglinton, in ref. 206, p. 20. 260. J. L. Soret, C.R. Hebd. Seances Acad. Sei., 1883, 97, 1269. 261. E. W. Baker, T. F. Yen, J. P. Dickie, R. E. Rhodes and L. F. Clark, .l Am. Chem. Sot, 1967, 89, 3631. 262. M. Blumer and W. n. Snyder, Chem. Geol., 1967, 2 , 35. 263. G. W. Hodgson, J. Flores and R. L. Baker, Geochim. Cosmochim Acta, 1969,33, 532. 264. G. W. Hodgson, M. A. Holmes and B. Halpern, Geochim Cosmochim Acta, 1970, 34, 1107. 265. A. Treibs, Justlts Liebigs, Ann. Chem., 1935, 517, 172. 266. J. G. Erdman, V. G. Ramsey, N. W. Kalendar and W. E. Hanson, J. Am. Chem. SOC.,1956, 78, 5844. 267. E. A. Glebovskaia and M. V. Vol’kenshtein, Zh. Obshch. Xhim., 1948,18,1440 (Chem. Abstr., 1950,44, 2209). 268. K. N. Ganesh, J. K. M. Sanders and J. C. Waterton, J. Chem. Soc., Perkin Trans. 1, 1982, 1617. 269. A. K. Battersby, C. Edington, C. J. R. Fookes and J. M. Hook, J. Chem SOC.,Perkin Trans. 1, 1982, 2265. 270. C. J. R. Fookes, J. Chem. SOC.,Chem. Commun., 1983, 1472. 271. J. M. E. Quirke, J. R. Maxwell, G. Eglinton and J. K. M.Sanders, Tetrahedron Lett., 1980, 21, 2987. 272. A. Ekstrom, C. J. R. Fookes, T. Hambley, H. J. Loeh, S. A. Miller and J. C. Taylor, Nature (London), 1983,306,173. 273. C. J. R. Fookes, J. Chem. Sw, Chem. Commun., 1983, 1474. 274. G. A. Wolff, M. Murray, J. R. Maxwell, B. K. Hunter and J. K. M. Sanders, J. Chem. SOC.,Chem. Commun, 1983, 922. 275. J. M. E. Quirke, G. Eglinton and J. R. Maxwell, J. Am. Chem Soc., 1979, 101, 7693. 276. J. M. E. Quirke and J. R. Maxwell, Tetrahedron, 1980,36, 3453. 277. M. F. Hudson and K. M. Smith, Tetrahedron, 1975,31, 3077. 278. B. M. Didyk, Y. I. A. Alturki, C. T. Pillinger and C. Eglinton, Nature (London), 1975, 256, 563. 279. R. Bonnett, P. J. Burke and A. Reszka, J. Chem. Soc., Chem. Commun.,1983, 1085.

878

Geochemical and Prebiotic Systems

R. Bonnett and F. Czechowski, Nature (London), 1980, 283, 465. V. V. Koval’skiy and S . V. Letunova, Geochem. Int., 1966, 3, 1176. K. A. Kvenvolden, E. Peterson and G. E. Pollock, Narure (London), 1969,221, 141. P. E. Hare, in ref. 206, p. 438. K. A. Kvenvolden, Annu. Rev. Earth Planet. Sci, 1975, 3, 183. G. Dungworth, Chem Geol., 1976, 17, 135. R. A. Schroeder and J. L. Bada, Earth-Sci. Rev., 1976, 12, 347. P. Westbroek and E.W. de Jong (eds.), ‘Biomineralization and Biological Metal Accumulation’, D. Reidel, Dordrecht, 1983. 288. K. King, Jr. nd P. E. Hare, Micropaleontology, 1972, 18, 285. 289. J. L. Bada and R. A. Schroeder, Earth Planet. Sci. Lett., 1972, 15, 1. 290. R. A. Schroeder and J. L. Bada, Geochim. Cosmochim. Acta, 1977,41, 1087. 291. J. F.Wehmiller, P. E. Hare and G. A. Kujala, Geochim. Cosmochim. Acta, 1976,40, 763. 292. R. D. Gillard, P. O’Brien, P. R. Norman and D. A. Phipps, J. Chem Soc., Dalton Trans., 1977, 1988. 293. G . G. Smith, A. Khatih and G. S. Reddy, J. Am. Chem Soc., 1983, 105. 293. 294. E. T. Degens, ‘Geochemistry of Sediments’, Prentice-Hall, New Jersey, 1965. 295. L. H. Ahrens, Geochim. Cosmochim. Acta, 1966,30, 1111. 296. A. Siegel, Geochim. Cosmochim. Acta, 1966, 30,757. 297. P. Zubovic, A d n Chem. Ser., 1966, 55, 221. 298. G. Anderson, Soif Sci., 1961, 91, 156. 299. W. Van der Velden and A. W. Schwartz, in ‘Environmental Biogeochemistry’, ed. J. 0. Nriagu, Ann Arbor Science, Ann Arbor, MI, 1976, p. 175. 300. G. Dungworth, M. Thijssen, J. Zuurveld, W. Van der Velden and A. W. Schwartz, Chem. Geol., 1977, 19, 295. 301. G. C. Speen and E. V. Whitehead, in ref. 206, p. 638. 302, W. Bergmann, in ref. 182, p. 503. 303. H. M. Smith, Bull. US Bur. Mines, 1968, 642. 304. B. P. Tissot and D. H.Welte, ‘Petroleum Formation and Occurrence’, Springer-Verlag, New York, 1978. 305. L. M. Willey, Y . K. Kharaka, T. S, Presser, J. B. Rapp and I. Barnes, Geochim. Cosmochim. Acra, 1975, 39, 1707. 306. J. N. Cardoso and G. Eglinton, Geochim. Cosmochim Acm, 1983, 47, 723. 307. M. M. Kononova, ‘Soil Organic Matter’, 2nd edn., Pergamon, Oxford, 1966. 308. K. Kawarnura and R Ishiwartari, Nature (London), 1982,297, 144. 309. J. Berthelin, in ‘Microbial Geochemistry’, ed. W. E. Krumbein, Blackwell, Oxford, 1983, p. 223. 310. W. Huang and W. D. Keller, Am. Mineral., 1970, 55, 2076. 311. J. R. Boyle, G. IC. Voigt and B. L. Sawhney, Soil Sci, 1974, 117, 42. 312. M. H. Razzaghe-Karimi and M. Robert, C.R. Hebd Seance3 Acad. Sci., 1975, 280, 2173. 313. M. J. Wilson, D. Jones and W. J. McHardy, Lichenologist, 1981, 13, 167. 314. 0. Sieskind and B. Siffert, C . R Hebd. Seances Acad Sci., 1972, 274, 973. 315. L. 0. Ohman and S. Sjoberg, Acta Chem. Scand, Ser. A, 1981,35,201; 1982, 36,47. 316. K. Mopper and K. Larsson, Geochim. Cosmochim Acra, 1978, 42, 153. 317. R. Chesselet and C. Lalou, C.R. Hebd. Seances Acud. Sci. 1965, 260, 1225. 318. P. F. Andreyev, E. A. Plisko and E. M. Rogozina, Geochemirtry, 1962, 624. 319. R. H. Fish,R. J. Tannous, W. Walker, C. S. Weiss and F. E. Brinkman, J, Chem. SOC.,Chem. Commun., 1983, 490. 320. B. Kfibek, GeoI. Carpathica. 1981, 32, 605. 321. T. V. Drozdova, K. I. Yakubovichand and E. F. Konstantinov, Geochem. Int., 1964, 529. 322. R. L. Kranz, in ref. 206, p. 521. 323. R. L. Kranz, Trans. - Inst. Min. Metall., Sect. B, 1968, 77, 26. 324. L. R. Gardner, Geochim Cosmochim. Acta, 1974, 38, 1297. 325. F. W. Pauli, Soil Sci, 1975, 119, 98. 326. A. S. Radtke and B. J. Scheiner, Econ. Geol., 1970, 65, 87. 327. 0. Brockamp, Contrib. Mineral. Petrol., 1974,43, 213. 328. 0. Brockamp, Sedimentology, 1976,23, 579. 329. K. H. Nealson, in ref. 309, p. 191. 330. K. H.Nealson, in ref. 309, p. 159. 331. S. L. Miller, in ‘Mineral Deposits and the Evolution of the Biosphere’, ed. H. D. Holland and M. Schidlowski, Springer-Verlag, Heidelberg, 1982, p. 155. 332. A. I. Oparin, ‘The Origin of Life’, Macmillan, New York, 1938. 333. J. B. S. Haldane, Rarionalisi Annu., 1928, 148, 3. 334. H. C. Urey, Roc. Nail. Acad. Sci. USA, 1952,38, 363. 335. J. D. Bernal, ‘The Physical Basis of Life’, Routledge and Kegan Paul, London, 1951. 336. P. H. Abelson, Prw NatL Acad. Sei USA, 1966, 55, 1365. 337. P. E. Cloud, Jr., Science, 1968, 160, 729. 338. R. M. Lemmon, Chem. Rev., 1970, 70,95. 339. M. T. Beck, Met. Ions Biol. Syst., 1978, 7, 1. 340. M. T. Beck and J. Ling, Natunvissenschaften, 1977, 64, 91. 341. L. E. Orgel, in ‘TheOrigin of Life and Evolutionary Biochemistry’, ed. K. Dose, S. W. Fox, G. A. Deborin and T. W. Pavlovskaya, Plenum, New York, 1974, p. 369. 342. V. J. Landis, J. Chern. Educ., 1974, 51, 565. 343. B. A. Averill and W. H. Orme-Johnson, Met. loons Bioi. Syst., 1978, 7, 128. 344. J. B. Neilands, Stmcr Bonding (Berlin), 1972, 11, 145. 345. S. L. Miller, SLience, 1953, 117, 528. 346. W. M. Garrison, D. C. Morrison, J. G. Hamilton, A. A. Benson and M. Calvin, Science, 1951, 114,416. 347. J. E. Van Trump and S . L. Miller, Origin Life, Proc ISSOL Meet., 3r4 Jerusalem, 1980. 348. C. Ponnamperuma and F. Woeller, Cum Mod B i d , 1967, 1, 156.

280. 281. 282. 283. 284. 285. 286. 287.

Geochemical and Prebiotic Systems

879

R. A. Sanchez, J . P. Ferris and L. E. Orgel, Science, 1966, 154, 784. C. N. Matthews and R. E. Moser, Proc. Natl. Acad. Sci USA, 1966, 56, 1087. C. N. Matthews and R. E. Moser, Nature (London), 1967,215, 1230. R. E. Moser, A. R. Clagget and. C. N. Matthews, Tetrahedron LetL, 1968, 13, 1599. R. E. Moser, A. R. Claggett and C. N. Matthews, Tetrahedron Lett., 1968, 13, 1605. K. Harada, Nature (London), 1967,214,479. K. A. Kvenvolden, J. Lawless and C. Ponnamperuma, Roc. NutL Acad. Sci. USA, 1971,68,486. R. Hutchison, ‘The Search for Our Beginning’, Oxford University Press and British Museum (Natural History), Oxford and London, 1983. 357. K. Bahadur, Nature (London), 1954, 173, 1141. 358. K. Bahadur, S. Ranganayaki and L. Santarnaria, Nuture (London), 1958, 182, 1668. 359. J. Or6, A. Kimball, R Fritz and F. Master, Arch. Biochem. Biophys., 1959,85, 115. 360. F. Egami, J. Mol. Errol., 1974,4, 113. 361. F. Egami, J. Biochem (Tokyo), 1975,77, 1165. 362. D. Yoshino, R Hayatsu and E. Anders, Geochim. Cosmochim Acta, 1971, 35,927. 363. H. Hatanaka and F. Egami, Bull Chem. Soc Jpn., 1977,50, 1147. 364. M. Sato, K. Okawa and S. Akabori, Bull. Chem. Soc J p n , 1957,30, 937. 365. D. A. Phipps, 1. Mol CatuZ., 1979, 5 , 81. 366. A. C. Kurtz, J. Biol. Chem., 1938, 122, 477. 367. F. Turba and K. H. Schuster, Hoppe-Seyler’s Z. Physiol. Chem., 1948, 283, 27. 368. A. Weiss, in ref. 206, p. 737. 369. T. D. Thompson and A. Tsunashina, Cfays Clay Miner., 1973, 21, 351. 370. A. N. Astanina, A. P. Rudenko, M. A. Ismailova, E. Y. Offengenden and H. M. Yakubov, Inorg. Chim. Acta, 1983, 79(B7), 284. 371. N. H. Horowitz and S . L. Miller, Prog. Chem. Org. Nat. M.,1962, 20, 453. 372. G. Balavoine, A. Moradpour and H. B. Kagan, J. Am. Chem Soc, 1974, 96, 5152. 313. G. W. Hodgson and B. L. Baker, Nature (London), 1967,216, 29. 374. S. L. Miller, J. Am. Chcm. S m , 1955,77, 2351. 375. R. Hayatsu, Science, 1964, 146, 1291, 376. J. Orb, Biochem. Biophys. Res. Commun., 1960,2, 407. 377. J. Or6 and A. P. Kimball, Arch. Biochem. Biophys., 1961, 94, 221. 378. J. Or6 and A. P. Kimball, Arch. Biochem. Biophys., 1962, 96, 293. 379. R. A. Sanchez, J. P. Ferris and L. E. Orgel, J. Mol. Biol., 1967, 30,223. 380. A. W. Schwartz and G. J. F. Chittenden, Biosystems, 1977, 9, 87. 381. R. Hayatsu, M. H. Studier, S. Matsuoka and E. Anders, Geochim Cosmochim. Acta, 1972,36, 555. 382. R. Lohrmann and L. E. Orgel, Nature (London), 1973, 244, 418. 383. R. Lohrmann, P. K. Bridson and L. E. Orgel, Science, 1980, 208, 1464. 384. A. Butlerov, Justus Liebigs Ann. Chem., 1861, 120, 296. 385. N. W. Gabel and C. Ponnamperuma, Nature (London), 1967, 216, 453. 386. C. Reid and L. E. Orgel, Nature (London), 1967, 216,455. 387. A. G. Cairns-Smith, ‘Genetic Takeover and the Mineral Origins of Life’, Cambridge University Press, Cambridge, 1981.

349. 350. 351. 352. 353. 354. 355. 356.

Applications in the Nuclear Fuel Cycle and Radiopharmacy CHRISTOPHER J. JONES University of Birmingham, UK 65.1 INTRODUCTION

88 1

65.2 THE NUCLEAR FUEL CYCLE 65.2.1 The Geochemical Segregation of Uranium from Thorium 65.2.2 The Recouery of Uranium and Thoriumfrom Ores 65.2.2.1 Uranium 65.2.2.2 Thorium 65.2.3 Nuclear Fuel Preparation 65.2.3.1 Final puri$ication of uranium 65.2.3.2 Final puri$curion of thorium 65.2.3.3 Thermal reactor fuel preparation 65.2.3.4 Fast reactor fuel preparation 65.2.4 Reprocessing of Irradiaied Fuel 65.2.4.1 Fuel decanning and dissolution 65.2.4.2 Solvent extraction fundamentals 65.2.4.3 Solvent extraction processes 65.2.4.4 Waste treatment and the recovery of radionuclides

885 886 895 89 5 911 919 919 922 923 924 925 927 928 936 959

65.3 PHARMACEUTICAL APPLICATIONS OF RADIONUCLIDES 65.3.1 Chemical Aspects of Radiopharmaceutical Formulation 65.3.1.1 Gallium and indium 65.3.1.2 Technetium 65.3.2 Diagnostic Imaging Applications of Radiopharmuceuricals 65.3.2.1 Bone 65.3.2.2 Kidney 65.3.2.3 Hepatobiliary system 65.3.2.4 Heart and brain 65.3.2.5 Tumors and abscesses 65.3.2.6 Labelled colloids, cells and proteins

963 968 969 912 98 5 985 987 989

990 992 994

65.4 RECENT DEVELOPMENTS

996

65.5 REFERENCES

997

65.1 INTRODUCTION

Coordination chemistry has played a crucial, though not always well defined, role in the isolation and utilization of metallic radionuclides. Some two billion years ago it was the differences in the coordination chemistries of Uv' and Th'" which led to their geochemical segregation, and the subsequent formation of the richest uranium ore deposits available today. More recently, in 1898, the Curies exploited subtle differences in the chemistries of the metallic elements during their painstaking isolation of radium chloride by fractional crystallization of BaCl,-RaCl, mixtures. In the same year, Marie Curie identified polonium and in 1899 Debierne discovered actinium. It was work such as this, carried out at the close of the 19th century, that laid the foundations for the modern radiochemical industry. Nowadays, tonne quantities of radioactive materials are routinely processed by the nuclear power industry and again coordination chemistry plays a central role in the separation processes used. As an adjunct to this exploitation of the actinide elements as sources of nuclear fission energy, the use of other radionuclides in a variety of applications has been developing apace. Nuclides such as 23*Puand 90Sr can provide reliable sealed heat sources which can operate continuously, without attention, for many years. These find applications in such diverse devices as heart pacemakers and spacecraft. Gamma ray emitting nuclides such as 6oCo may be used as radiation sources in, for example, the sterilization of 881

882

Applications in the Nuclear Fuel Cycle and Radiopharmacy

disposable medical equipment or the radiographic inspection of engineering components. The radiography of human patients is also possible using radiopharmaceutical agents containing short half-life y r a y emitters such as 99mT~. Three basic approaches may be used to generate ‘man-made’ radionuclides, two of which involve neutron irradiation of suitable target materials in a nuclear reactor. In one case this is to induce fission of a heavy nucleus such as 235Uin order to produce fission products such as 90Sr, 99M0 or 137Cs,In the other case, neutron capture reactions are utilized in, for example, the conversion of 98M0to 99M0 or the conversion of 235Uto 238Pu.The third approach is only suitable in cases where small quantities of the product radionuclide are required since it involves the use of a particle accelerator to irradiate a target with ions derived from the nuclei of light elements. In all three approaches the isolation of the desired radionuclide will normally invoIve a separation process. Usually this process will include a wet chemical stage which exploits differences in the coordination chemistries of the metallic elements present. In the case of radiopharmaceutical applications, not only is coordination chemistry involved in the isolation of metallic radionuclides but also in their incorporation into the active ‘drug’. In fact the development of technetium radiopharmaceuticals represents a particularly timely and sophisticated application of coordination chemistry. In some ways this chemistry complements that of the nuclear fuel cycle. A large part of the basic chemistry involved in the reprocessing of nuclear fuels was developed in the post-war years between 1946 and 1970. Although many points of detail are still under investigation, the Purex process, which is most widely used today, employs the same basic chemistry as that developed in the late 1940s.’ In contrast, major developments in the chemistry of technetium, especially as applied to radiopharmaceuticals, have occurred since 1970. This situation is to some extent reflected in the type of ligands used in these two different areas. Actinide separation generally involves hard ‘inorganic’ ligands such as water, nitrate, sulfate or organophosphate. Radiopharmaceutical chemists, on the other hand, are increasingly investigating the use of softer organic ligands containing groups such as amine, thiol or phosphine. Thus the historical development of coordination chemistry is, in some ways, reflected in the chemical systems used in the nuclear fuel cycle and, subsequently, in the formulation of radiopharmaceuticals. It is appropriate, therefore, to consider the nuclear fuel cycle, and the nuclide production processes which have developed in its wake, before turning to the use of radionuclides in pharmacy. The coordination chemistries of the elements considered in this chapter have already been the subject of detailed discussion in earlier sections of these volumes. Consequently, it is the purpose of this chapter to review, in general terms, the nuclear fuel cycle, the production of metal radionuclides and subsequently their incorporation into radiopharmaceutical formulations. Within this framework, specific aspects of coordination chemistry which are relevant to the application in question will be considered. Comprehensive accounts of the nuclear fuel cycle, including its chemical aspects, have been given, by Flagg’ in a book published in 1967, and more recently by Wymer and V ~ n d r a More .~ general texts on solvent extraction chemistry and separation methods for the rarer metals are also along with compilations of papers relating specifically to actinide separations.8 The uses of solvent extraction in the nuclear industry have been reviewed by Jenkins9 and reviews of fuel reprocessing chemistry are given by Bond” and by Wilkinson.” Papers on solvent extraction processes may also be found in the proceedings of the International Solvent Extraction Conferences12 (ISEC) held in Liege (1980), Toronto (1977), Lyon (1974), The Hague (19711, Jerusalem ( 1968),12aGothenburg ( 1966),12bHarwcll ( 1965),’2cBrussels ( 1963)’2dand Gatlinburg ( l962).lZ2 At least one section in these usually relates to the nuclear fuel cycle. The chemistry of nuclear fuels themselves has been described in reviews by C a r n ~ b e l land ’ ~ by Marples et aLI4 Fast reactors have now been developed to the prototype stage and the results of work on the reprocessing of fast reactor fuels were described at an international symposium held at Dounreay in 1979.15 Turning to the subject of radiopharmaceuticals, a basic text on radiopharmacy is provided by Tubis and Wolf.16 The chemistry of radiopharmaceuticals is described in a compilation of papers edited by Heindel et al.,” while the preparation of labelled compounds containing short-lived radionuclides has been reviewed by Silvestor.” Also Meares et al. have described the attachment of chelating groups to biological molecules to produce metal ion carriers which are selectively bound to sites within the test organism.” The metals most widely used in radiopharmaceutical applications are technetium, indium and gallium. In this context, recent reviews of technetium chemistry have been provided by Deutsch et al.:’ by Davison and Jones” and by Schwockau.22 The chemistry of gallium and indium which relates to radiopharmaceuticals has been reviewed by Moerlein and Welch.23

Applications in the Nuclear Fuel Cycle and Radiopharmacy

883

65.2 THE NUCLEAR FUEL CYCLE

The term 'nuclear fuel cycle' refers to the processes involved in the generation of power by the fission of actinide nuclei in a nuclear reactor. These processes comprise the manufacture of nuclear fuel elements, their irradiation in a reactor and the subsequent reprocessing of the irradiated fuel to recover fissile and fertile material." There are also two closely associated processes which are not part of the cycle itself. The first is the extraction of fissile and fertile material from ore deposits and the second is the treatment and disposal of the wastes generated by the cycle, especially by the reprocessing stage. In order to place these processes in perspective it is useful to consider the complete lifespan of the actinide elements as summarized in Figure 1. Since their formation during stellar element synthesis and nuclear reactions in supernovae, most actinide isotopes have had time to decay to insignificant levels by radioactive processes. Only those isotopes with half lives approaching the age of the Universe are found in significant quantities on Earth. These are 238 U ( tl/? = 4.47 x lo' y), 235U (t, , l = 7.04 x lo*y) and 232Th( tlIz = 1.405 x 10" y), of which only -'U is fissile. Small amounts of other actinides, particularly neptunium, actinium, protactinium and plutonium, also occur naturally, either as a result of neutron capture in actinide bearing rocks or from radioactive decay processes. These elements are not of primordial origin and, although gram quantities of protactinium have been obtained by reprocessing uranium ore residues, they are usually best obtained by neutron irradiation of suitable target materials in a nuclear reactor. During the Earth's history, geochemical processes have led to the segregation of the uranium and thorium present in the crustal rocks and to the formation of uranium ore deposits containing as much as a few percent uranium in some cases. If left in the ground this uranium and thorium would eventually decay to stable lead isotopes. However, if instead they are extracted and fed into a nuclear fuel cycle, their destruction may be accelerated and fission energy produced. Since each fission event libertates 208 MeV of energy, the potential energy stored in a fissile actinide represents some three million times that available from the combustion of a similar mass of fossil fuel. Natural uranium contains only 0.7% of fissile 235Uand in the form of metal or oxide cannot sustain a fission chain reaction. This results from the small fission cross section offered by the 235Unucleus to the energetic 'fast' neutrons released during fission. A much higher 235Uconcentration is necessary to maintain a fission chain under such circumstances. However, the fission cross section of 235Uis larger with respect to slow or 'thermal' neutrons. Thus if natural uranium is placed in a medium such as graphite, water or D20, which can act as a moderator and slow down, or thermalize, the fission generated neutrons, a chain reaction may be sustained. It is the application of this strategy which underlies the thermal reactor nuclear fuel cycle set out on the right hand side of Figure 1. Thus natural uranium, in the form of metal in the case of the older graphite moderated British Magnox ractor, or oxide in the D20moderated Canadian CANDU reactor, may sustain a fission chain reaction in a thermal reactor. Although it is possible to utilize natural uranium in this way, greater efficiencies in energy production may be achieved if a higher 235U concentration is used in conjunction with a moderator. Consequently the newer British Advanced Gas Cooled Reactor (AGR) and the American Pressurized Water Reactor (PWR) systems, which are graphite and water moderated respectively, utilize fuels containing uranium enriched to 2.3% and 3.2% 235U.These reactor systems thus require an additional enrichment stage between the uranium extraction process and the fuel fabrication stage. During the irradiation of the fuel in the neutron flux of the reactor, two important processes are occurring. Firstly, the fissile 235Unuclei are splitting to generate some 40 or so fission product elements (FPs). These all have different half-lives and some will have largely decayed at the time the fuel is discharged from the reactor. After three years the more important FPs remaining in irradiated PWR fuel are, in order of decreasing radioactive contribution to total fuel acti~ity,'~ 1 3 7 1~3 ' 7 r n ~ ~144 , Pr, ' W e , 1 0 6 ~ 1~0 ,6 a , 90sr, 90y, 147Pm, 134Cs,1 5 4 E ~%r, , I2%b, I5'Eu, IZSrnTe, 3H, 151Sm, "lmAg, '13Td, 95Nb, 123Sn,95Zr, 9 9 T ~"',Te and 119mSn.Secondly, neutron capture reactions are occurring. One of these results in the formation of fissile 239h from fertile 238U according to Scheme 1. Some of the 239Puproduced will also undergo fission, but 239Pumay represent some 0.5% of the actinide content of the fuel discharged from the reactor. Neutron capture reactions also result in the formation of other plutonium isotopes along with some americium and curium.

Scheme 1 CCC6-cc

Applications in the Nuclear Fuel Cycle and Radiopharmacy

884

I

I--Searewtion Planetary ptocesses Separate U from Th

To increase 23s~ Content Natural U as metal or oxide

Fertile isotopes

Enriched U as oxide

Enriched UOn: AGR,PWR Natural U02:CANDU

23aPu/2?J orZWU/'?h as oxides

Natural U metal : Magnox

Irradiation 2uU

Thermal neutrons

Fast neutrons Pu breeder 239Pu-+FP~:3sU--+ 239Pu

259pu

or

235U-FPs

Thermol neutrons Th breeder 232TP 2uU

233 U -

23e(J__r2=pu

Fps:32Th--+ 233U

t Reprocessing Essential for breeder cycle Remove FPS Reclaim fertile and fissile mater ial

,

CANDU

A net consumption of fissile material occurs

More fissile matariol is formed than consumed

-

-

+

Optional for thermal cycle Remove FPs Recover separative work in enriched U 230Pu Reclaim 23sPu

-

~

Highly Active Waste Recovery Potentially useful rodionuclidas m y be separated from waste a.g. iss~u. '"Cs ~

~

~

FPs t Contains i ~ ~ ~ l 8,g. Cs.Sr. Ru.Zr.Nb.Tc Contuins actinides 8.g. Np.U.Pu. Am.Cm and possibly Po. Th

CAND

Applications in the Nuclear Fuel Cycle and Radiopharmacy

885

It is processes of this second type which make possible the alternative ‘breeder reactor’ fuel :ycle shown on the left side of Figure 1. If the excess neutrons emitted from the care of an )perating fission reactor are absorbed in a surrounding ‘blanket’ of fertile 23’U, capture reactions vi11 lead to the formation of 239Pu.Thus it is in principle possible to generate more fissile material n the blanket than is consumed in the core. The use of natural uranium to produce plutonium n a breeder cycle of this type could increase the available fission energy by a factor of 50-60. 4owever, to realize this goal conditions are needed which differ substantially from those found n a thermal reactor. The neutron capture, or ‘breeding’, process is more efficient when unmoderated fast’ neutrons are used. Thus in order to optimize the ‘breeding gain’ of the reactor an unrnoderated :ore is required. This in turn requires a much higher concentration of fissile material in the reactor :ore if a chain reaction involving fast neutrons is to be sustained. A fuel containing about 20% issile material is necessary. A further feature of such a reactor system is the hi h energy density n its core during normal operation. This may be of the order of 0.5 MW dm-$! and the C 0 2 or water coolants used in the AGR and PWR are not sufficiently good conductors of heat to transfer mergy out of the core at the required rate. More effective coolants, such as liquid metals, are iecessary, sodium being the metal of choice. It is this need for a liquid metal coolant and fast ieutrons in an effective breeder system that gives reactors of this type the name Liquid Metal 2ast Breeder Reactor (LMFBR). The breeder cycle thus involves the fabrication of both core and ,lanket fuel elements. The core contains 239Puas its fissile component while the blanket contains ‘38 U as its fertile component. After irradiation, both the bIanket and core must be reprocessed to .emove the fission products and recover the fissile 239Puand fertile 238Ufor recycling through the ‘eactor. In this way the 238Uin natural uranium can also be used as a source of fission energy. f i e 23’U/239Pubreeder reactor cycle is under development in Britain, France and the USSR with xototype reactors operating in all these countries. An alternative breeder cycle involves 233Uand !32711as its fissile and fertile components as summarized in Scheme 2. This offers the advantage if a superior neutron yield of 233Uin a thermal reactor system. However, the 1 . 9 ~half-life z-emitter 228This also formed and presents a major problem in the reprocessing stage.24Such a :ycle would allow the world‘s thorium reserves to be added to its uranium reserves as a potential iource of fission power. However, the 232Th/233Ucycle is unlikely to be developed in the near ’uture owing to the more advanced state of the 23’U/239Pucycle and the current availability of iranium. Z32n

233m

22 3 min

2 3 3 ~ ~ 2 3 3 ~ 270d

Scheme 2

In the case of the thermal cycle, fuel reprocessing is not essential. In the CANDU system, for example, the fuel is placed in secure long-term storage without reprocessing. However, this spent a substantial amount of 238Uand some fissile 239Pualong fuel still contains some unused 235U, with other actinides and the FPs. In the case of reactors using enriched fuels the final 235Ucontent may still be greater than in natural uranium. Since enrichment is an expensive energy intensive process, recovery of the separative work embodied in any increased 235U content may provide an incentive for reprocessing. Also the plutonium recovered from the reprocessing of thermal reactor fuel may be incorporated into the breeder cycle. The final composition of the irradiated fuel will depend upon its initial fissite material content and the energy spectrum of the neutrons to which it has been exposed. It will also depend upon the extent to which the fuel has been irradiated, that is its ‘burn-up’. The irradiation history of the fuel may then have a bearing on the process chemistry of any subsequent reprocessing stage. The primary aim of the reprocessing stage is to remove neutron absorbing fission products which, if allowed to accumulate in the fuel, could impair reactor performance. However, since the actinide product streams from reprocessing must be accepted back into the enrichment or fuei fabrication processes to complete the cycle, extremely effective separation of the fuel from all FPs and higher actinides is necessary. The concentrations of plutonium and lo6Ruin the uranyl nitrate product from British Nuclear Fuels ‘Thermal Oxide Reprocessing Plant’ (THORP), for example, are required” to be respectively 5 x 10’ and 1 x 10’ less than in the irradiated fuel fed into the plant. These figures are known as decontamination factors, DF, and DF,, in this case, and are normally defined in terms of the radioactivity of a fuel contaminant X measured in nuclear disintegrations per unit time as shown in equation (1). The magnitudes of these DF values illustrate the need for a process which can separate uranium, and plutonium, selectively from a mixture containing up to about 40 elements. The radioactive fission products and hi her actinides separated from the fuel provide a potential source of useful radionuclides such as 4 ’Cs and 23sPu.Further separation processes may thus be included to extract specific elements from the highly active

886

Applications in the Nuclear Fuel Cycle and Radiopharmacy

waste (HAW} produced by fuel reprocessing. Another important aim of the reprocessing stage is to convert the unwanted materials in this HAW to a form suitable for long-term storage. To this end it is desirable to combine all the radioactive wastes in one HAW stream. The FP activity in this stream will have decayed by a factor of IO' before levelling off after almost lo3y. However, it will be over lo6 years before the actinides have decayed to a level where their radiotoxicity falls below that of the original ore had it been left in the ground. The secure storage of radioactive materials over such a timescale presents a major challenge to the nuclear industry. At present the most highly developed process for long-term HAW storage is vitrification. This involves the incorporation of the HAW into borosilicate glass blocks which are expected to provide a thermally stable leach-resistant matrix in which the FPs and actinides will remain immobilized over geological time scales. DF, =

Activity of X per unit mass of U in feed Activity of X per unit mass of U in product

Viewed in the context of the actinide lifespan, the nuclear fuel cycle involves the diversion of actinides from their natural decay sequence into an accelerated fission decay sequence. The radioactive by-products of this energy producing process will themselves ultimately decay but along quite different pathways. Coordination chemistry plays a role at various stages in this diversionary process, the most prominent being in the extraction of actinides from ore concentrate and the reprocessing of irradiated fuel. However, before considering these topics in detail it is appropriate to consider briefly the vital role played by coordination chemistry in the formation of uranium ore deposits.

65.2.1 The Geochemical Segregation of Uranium from Thorium Uranium and thorium are not particularly rare metals. They occur to the extent of about 4 and 12 p.p.m. respectively in continental rocks as a whole.24The average Th :U ratio in the earth's crust is thus about 3, and not greatly different from the cosmological value" of 2.23. This suggests that these elements were not segregated to any major extent during primary differentiation and the formation of the molten planetary core. The electropositive nature and high oxygen affinities of uranium and thorium led to their appearance in the oxide and silicate components of the early rock forming system. However, the ionic charges and radii of U'" and ThIVare sufficiently different from those of the major metallic components of crustal rocks that they would Iargely be excluded from the rock lattice during crystallization. This led to their being mainly associated with minerals forming after the bulk of the rock had crystallized. In support of this general view, autoradiographic studies of granites have shown that most of their radioactivity is associated with microcrystalline inclusions and materials residing along grain boundaries within the rock mass.26 U4++2Hz0 --t U0;++4H++2euIll

-1.95v _ _

L,I\

1

"V

0.07 V

(2)

p I

Scheme 3

The release of uranium and thorium from rnjnerals into natural waters will depend upon the formation of stable soluble complexes. In aqueous media only Th'" is known but uranium may exist in one of several oxidation states. The standard potential for the oxidation of U4+ in water according to equation (2) has been re-eval~ated~' as E" = 0.273 f 0.005 V and a potential diagram for uranium in water at pH 8 is given in Scheme 3." This indicates that U"' will reduce water, while U" is unstable with respect to disproportionation to UIV and U"'. Since the Earth's atmosphere prior to about 2 x lo9 y ago was anoxic, and mildly reducing, UJvwould remain the preferred oxidation state in natural waters at this time. A consequence of this was that uranium and thorium would have exhibited similar chemistry in natural waters, and have been subject to broadly similar redistribution processes early in the Earth's history. Both UIVand Thrv are readily hydrolyzed in aqueous solutions of low acidity. A semiquantitative summary of the equilibrium constants for the hydrolysis of actinide ions in dilute solutions of zero ionic strength has been provided by Liljenzin and coworkers.2BThe results obtained for Th4+, W4' and UO;+ are summarized in Table 1. The tetracations of the oxidation state (IV) metals are the dominant species only below pH 1. Above pH 8 the monoanionic M(OH)5- complexes are expected to be the dominant species in solution, although the bulk of the metal will have precipitated at this

Applications in the Nuclear Fuel Cycle and Radiopharmacy

887

Table 1 Equilibrium Constants for the Formation of Hydroxy Complexes at Zero Ionic Strength and 25 "C

Reaction

Metal Ion M'

{'+OH-= M(OH)'-' d L + 2 0 H - e M(OH)2z-2 /I= OH~(0~1~2-3 / I Z + ~ O He - M(oH),=-~ +SOW' 2 ~ ( 0 ~ 1 ~ 1 MZ+OWM,(OH)2Z~'

AL

5

M'+20H- e M2(0H)22Z-2 M= OH- M , ( o H ) , ~ z - ~ M Z + ~ O H - M,(oH),~-~ M z + 5 0 H - S M2(OH)52'-5 M Z OH- e M ~ ( O H ) ~ Z - ~ M z + 5 0 H - $ M,(OH),3'-' . M Z +SOHM~(oH),~z-* d(OH),'-4(solid) S M z + 4 0 H A(OH),'-*(solid) Mz+20HAO,(solid) +2H,O e M4++40H-

+

=

Estimated log K ualue" 1K4+ u4+

*

10.8 0.3 22.2

40.1

-

I

21.9 f 0.6 I

I

90.9+0.1 -44.7

-

-49.7

UO?

13.4i0.4 (25.5k0.8) (36.6* 1.2) (46.3 1.6) 54.0 f 2.0

8.1 f 0.2 (17.2k0.8) (21 2) -

-

9.6 k 5 22.4*0.2 -

*

21k3 40i4 52*5 62*6

-

( - 5 4 i 2)

-

-57.8 *2

*

-

-

-

44.2* 0.5 55*1

-22.4*0.2 -22.4k0.2 -

Values in parentheses were obtained by extrapolation.

)H.Below p H 6 a mixture of several hydroxy species would be formed. An investigation2' of rh4+hydrolysis in 0.1 M K N 0 3 solution has indicated that, under these conditions, Th(OHI3+, Th,(OH),,4* and Th,(OH),,9* were formed" with overall stability constants log p = 2.98(0.O07), 10.55(0.03) and 34.41(0.03) respectively, the estimated standard deviations being given in parenheses. No evidence for the formation of dinuclear species such as Th2(0H):+ was found. The

hermodynamic data available for uranium complexes have been critically evaluated by L a n g r n ~ i ? ~ n order to predict the speciation of uranium in natural waters. In a model aqueous phase :ontaining fluoride (2 p.p,m.), chloride (10 p.p.m.), suIfate (100 p.p.m.) and phosphate (1 p.p.m.), J(OH)*- was the dominant species above pH 6 and below pH 3 only the fluoride complexes JF3+,UF?+, UF,* and UF, were present. Between pH 3 and 6, U(OH);+, U(OH),+, U(OH), tnd U(OH),- were formed to varying extents, with U(OH)4 predominating at pH 4. At low pH, luoride ion plays a prominent role in the solubility of U 0 2 in water at 25 "C. In the absence of 7- at pH 2 the total dissolved uranium concentration is lo-'* M, while the presence of 2 p.p.m. otal fluoride increases this value to ca. M. However, at pH 4-5 the effectiveness of F- ceases md the dissolved uranium concentration reaches a minimum of M. The uranium concentraM at pH 9. ion then increases linearly with pH to about The low solubilities of UIv and ThIV in natural waters of pW 4-7 would lead to their being ransported mainly in the solid phase. That is either as components of the mineral lattice of 'esistate grains released during rock weathering, or as complexes adsorbed on soils or sediments md, in the case of thorium, on clay minerals.26730 The importance of the organic fractions of soils ind sediments is illustrated by a study of the binding of ThTV to humic and fulvic acids obtained iom lake sediment^.^' The binding to humic acid was thought to occur at two sites, the first nvolving one carboxylate moiety and the second two carboxylates. The thermodynamic data for .hese systems are given in Table 2 and include positive entropy values two to three times the nagnitude of those found for acetate or sulfate complexing of Th4*. This was thought to indicate I higher degree of desolvation in the complexing of Th4" by humic or fulvic acid. This may arise .hrough the organic macromolecule providing regions of comparatively low dielectric constant iround the binding sites, Approximately 2 x lo9 years ago the activities of photosynthetic algae wrought a dramatic :hange in the composition of the Earth's atmosphere through the release of large quantities of 3xygen. Under these new conditions the widespread oxidation of UIVto Uv' became possible in rocks and sediments exposed to the atmosphere. The formation of water soluble uranyl complexes :odd provide a powerful mechanism for the dissolution and aquatic transport of uranium and ;hus would lead to its geochemical segregation from thorium, for which no such oxidized species :odd form. In the U-02-Hz0 soluble uranium species are only formed at rather high 3r low pH values. However, in the presence of C02, soluble carbonate complexes are formed *These species are solvated but to an unknown extent. In this chapter, metal complex formulae will only be written in square brackets when they represent discrete molecular species in which the metal coordination sphere is complete.

888

Applications in the Nuclear Fuel Cycle and Radiopharmacy Table 2 Thermodynamic Dataa for the Binding of Th'" to Humic and Fulvic Acids3'

Th(humate)

Th(humate),

Th(fu1vate)

Th(fulvate),

log P

- AG

AH

PH

(kl eq-'1

(kl eq-')

AS (J eq-' K-')

4.0 3.95 4.60 5.03

-

63.562 0.06

32.6 f3.2

323 5 10

11.140*0.013 12.027 f0.023 13.181 i0.038

4.0 3.95 4.60 5.03

92.23 *0.12

42.1 f3.3

453 * 12

16.168k0.023 17.289+0.043 18.434f0.173

4.0 4.01 5.00

55.905 0.18

18.9 4.2

*

251 *44

9.798 f 0.029 10.824* 0.051

4.0 4.01 5.00

76.97 f0.33

46.4 f8.4

414*30

13.495f 0.056 15.073 0.084

-

-

-

*

'Data obtained at 25k0.5"C at ionic strength 0.1 M as NaCIO,+NaO,CMe.

which stabilize UO;+ in solution over a wide range of pH values.32 Thus although hydrated UO: ' is only found in solution below pH 4 in the absence of CO,, under a C O , partial pressure of 10-3-10-' bar at 25 "C soluble U02C03forms in the region pH 4-5 with [U0,(C03)z(H20)z]2predominating between pH 5 and pH 11. Above pH 11, [UO,(C0,),l4- becomes the dominant species in solution. The rate of dissolution of uraninite, UO,, in groundwater containing 0, and C 0 2has been studied by Grandstaff and the empirical model given by equation (3) was proposed.33 The rate of dissolution (mol d-') is given in terms of the specific surface area ( S A cm2gm-') of the UOz, the total dissolved carbonate concentration (aco2mol dmP3),the dissolved O2concentration (a,, p.p.m.), the mole fraction of non-uranium cations (mainly Th4+ and Pb2+)present ( N ) , the temperature ( T K), the hydrogen ion activity ( a H +mol dm-3) and a retardation factor ( R ) to take account of the poisoning of UO, surfaces by organic matter. The proposed mechanism of dissolution involved the slow adsorption of 02,followed by its splitting and surface migration to form hydrated U 0 3 according to equations (4)and ( 5 ) . Subsequent reactions of U0,.H20 with bicarbonate according to equations (6) and (7) then afford [UO,(CO,),( H20)$. One fast and one slow reaction were proposed to account for the first order dependence on ace,. In order to account for the first order dependence of rate on uH+ a hydration reaction of U 0 2 was proposed as shown in equation (8). Subsequent ionization could then lead to the species given in equations (9) and (10). If only the protonated form of the hydrate could be involved in 0, binding, a dependence on aH+would be expected.

-d[U1 - 102025 S A R - I

--

(10-3.38-EOSN

dt UO2+0,

+

UOztO+H,O

1 aco2 aH+ao2exp (-7045/

UO2.02 (slow)

+

UO3.HzO (fast)

T)

(3) (4)

(5)

The speciation of dissolved uranium in groundwater has been modeled in a study of uranium oxide fuel d i s s o l ~ t i o n The . ~ ~ calculated E,-pH diagram for a typical groundwater composition is shown in Figure 2.34,35This again shows the importance of carbonate complexes but also indicates that halide complexes of UO?+ and, under certain conditions, UO,+ may also be important contributors to the aqueous phase at low pH. The stability constants or uranyl carbonate and hydroxocarbonate complexes have been determined36and are summarized in Table 3. These results were thought to indicate that (UO,),(CO,)(OH),- was also important in the aquatic speciation of uranium. However, natural waters also contain silicate and phosphate ions. Although

Applications in the Nuclear Fuel Cycle and Radiopharmacy

0

2

4

6

e

IO

889

12

pH ( a t 25'Cc)

M sulfate, Figure 2 The E,!pH diagram for uranium in a model groundwater containing IOm3M total carbonate, 8.6 x M chlonde and M fluoride. (Reproduced from L. H. Johnson, D. W. Shoesmith, G. E. Lunansky, M. R. Bailey and P. R. Tremaine, Nud. Technol., 1982, 56, 238, with the permission of the American Nuclear Society)

a,b,x from:

aUO:++

bC02+ c H 2 0 e [(UO,),,(

(OH)c-h]2'--h--C + ( b + c)H+

where c - b = x . b + c = v a n d 2 a - b - c = z . b

(for aqueous solutions containing 0.1 M C10, and up to Z x

p"., for UO:++

M U).

rcoz- e I J O ~ ( C O ~ ) ~ ( " - ~ ~(-u = l , t = 1 - 3 )

& r f o r 3 U 0 ~ + + 6 C 0 3 Ze - (U0,)3(C03),"-

(r=6,y=3)

(for aqueous solutions).

silicate complexation of U0;+ appears relatively unimp~rtant?~ the high formation constant3? for UO2(HPO4);- makes it a major component at intermediate pH, as shown in Figure 3. Thus at intermediate and high pH it is phosphate and carbonate complexes respectively which dominate the speciation of U O:+ in natural waters. At lower pH values, fluoride and, to a lesser extent, sulfate complexes may be present in addition to hydrated UO;+. This complexation of uranyl ion results in typical total dissolved uranium concentrations of between 0.1 and 10 p . p h in continental surface waters compared with a figure of 10-3-10-2 jxpb. for thorium. In aerated groundwaters, higher uranium concentrations of 10-100 p.p.b. are generally found. This difference in net water solubilities had led to major differences in the nature of commercial thorium and uranium reserves. Thorium is generally obtained from alluvial or sedimentary deposits formed in regions where heavy refractory minerals, weathered from a source rock, have been concentrated by mechanical processes. In contrast, the richest uranium deposits have formed in regions where chemical processes have immobilized and concentrated uranium from surface- or ground-water flowing through them. Furthermore, the less mobile thorium is found to be an essential compound of only six minerals, compared with over 70 for uranium. The major source mineral of thorium is monazite, a rare earth phosphate containing up to 10% Th as LnP04, where Ln represents a variety of metal ions but especially Ce"', La'" and Y1ll in and thorite, combination with (Ca", ThIV). Other important thorium minerals are thorianite, Tho2, ThSi04, both of which may also contain significant proportions of U'".'

Applications in the Nuclear Fuel Cycle and Radiopharmacy

890

OH

Figure 3 The distribution of uranyl complexes2' as a function of pH at 25 "C with water containing 0.3 p.p.m. F-, bar 10 p.p.m. C1-, 100 p.p.m. SO:-, 0.1 p.p.m. PO:-, 30 p.p.m. SiOz and with a partial CO, pressure of

The aquatic transport of uranium as carbonate complexes is reflected in the formation of the uranyl carbonate minerals Rutherfordine, U02(C03),38Leibigite, Ca2[U 0 2 (C03)3]-10-llH20,39 and Andersonite, Na2Ca[U02(C03)3]-6H20?D Both natural and synthetic41UO,CO, have similar structures which contain sheets of planar C03'- ions. The uranyl groups have the normal trans dioxo configuration and occupy planar distorted hexagonal sites within these sheets. Each U"' thus has a distorted hexagonal bipyramidal coordination geometry which involves two bidentate and two monodentate carbonate ligands. The [UO,(CO,),]"- ions in Leibigite, Andersonite and synthetically prepared (NH4)4[U02(C03)3]42or &[U02(C03)3]43 also contain distorted hexagonal bipyramidal U"' but in this case there are three bidentate carbonate ligands giving rise to discrete ions of the complex, as shown in Figure 4. These structures are similar to those found

@ Uranium 0 oxygen 0

Carbon

b Figure 4 The structure of [U0,(C0,)3]4- showing selected average bond lengths and angles43

in Rb[U0,(N03),]44 and Na[U02(N03)2(H20)2]45 but contrast with the pentagonal bipyramidal coordination geometries found in [U02(H20)5]C104.2H2046 and [UO,( OAC),(H,O)].H,O~~ illustrated in Figure 5. The molecular structure of a U02(C03)22-complex has not been reported but

@

Uranium

2.

7

(0)

3

(b)

Figure 5 The structures of (a) [UOz(HzO),]2' and (b) [UOZ(H,O)(MeCO,),] showing selected average bond lengths and angles. (Adapted from N. W. Alcock and S . Esperas, J. Chem. Soc., Dalton Trans., 1977, 893, and R. Graziani, G. Bombien and E. Forsellini, J. Chem. Soc., Dalton Trans., 1972,2059, with permission from the Royal Society of Chemistry)

Applications in the Nuclear Fuel Cycle and Radiopharmacy

89 1

studies in NaHC03 media4' indicate that this complex will only be present in solutions which also contain substantial amounts of [U02(CO,), I"-. Despite the importance of carbonate complexes in the aquatic transport of uranium, these carbonate minerah do not represent a major source of recoverable uranium. The chemical compositions of some of the more important uraniferous minerals are presented in Table 4 and their occurrence has been reviewed in several texts relating to the production of nuclear f ~ e l s ? ~ ~ ~ - ' * The 1979 International Fuel Cycle Evaluation Conference defined the lowest production cost range for uranium as $80 kg-' or below including capital, direct mining and processing costs. Reasonably assured world reserves at this cost were estimated at 1.8 x lo6t. Out of this total some 84% was contributed by reserves in Australia, Canada, Namibia, the USA, South Africa and Niger. High-grade resources containing up to 0.2% uranium may be found in the sandstone deposits of the USA, Gabon and Niger. Overall, such sandstones account for about 35% of the uranium recoverable at low cost. A further 20% of low-cost reserves are in the form of veins, some being of hydrothermal origin, while an additional 15% are in the form of conglomerate deposits.24 Table 4 The Chemical Composition of Some Uranium Mineralszs49 ~

Anion

Mineral

Composition

Oxides

Uraninite" Pitchblende" Gummite Schoepite Brannente Davidite Coffinite Uranophane Autunite Torbernite Monazite Xenotine Zippeite Uranopilite Schroeklingerite Andersonite

uo2 %os (u'y,-,

Silicates Phosphates

Sulfates Carbonates

Liebigite

Organics Vanadates

a

Rutherfordine Thucholite Carnotite Tyuyamunite

)

uv12)02+x

U03.nH20 UO, (OH),.H,O (U,Ca,Fe,Th,Y),Ti,O,,

(Fe,Ce,U)(Ti,Fe,V,Cr),(O,OH), U(SiO,),-, (OH),,

Ca(U0,),(Si0,),(OH),.5H20 Ca( U02)2(PO,),. 10- 12H,O C~(UO~)~(PO,),*~~HZO (Ce,Y,La)PO, with Th and U YPO, with Th and U UO2 ( S O A (H,O) '8H20 NaCa,( U02)(C03), SO,F. 10H20 Na,CaU0,(C03),~6H,0~10H,O Ca2U02(C03)3.10-11H,O

uo2co3

Hydrocarbon matrix containing U, Th and rare earths KZ(UO,)Z(V~J~~~H,O Ca(U02),(V0,),-7-10H,0

Although uraninite is traditionally described as UO, and pitchblende as U,O,, both appear to represent a single mineral phase ranging from UO, in c o m p ~ s i t i o n . ~ ~

The leaching of uranium from source rocks, and its subsequent mineralization, have been considered in detail by Hostetler and G a r r e l ~and ~ ~more recently by Langm~ir.~' The formation of uranyl minerals most commonly occurs in regions where evaporation can concentrate groundwater containing uranyl ions and cornplexing ligands such as vanadate, phosphate, silicate or arsenate, and where COz concentrations are relatively low. The precipitation of U022+by vandate to give carnotite or tyuyamunite is an important fixation mechanism and leads to notable similarities between the distributions of uranium and vanadium. In the absence of vanadium, phosphate minerals such as autunite may form but, being more soluble, these are less abundant. The equilibrium constant K,, given by equation (11) for the conversion of potassium autunite to at 25 "C. The formation of carnotite according to equation (12) has been evaluated2' as soluble U02(HP04),*- means that, when phosphate and vanadate are both present, the total phosphate concentration must exceed that of vanadate by a factor of about 500 before autunite will precipitate in preference to carnotite. Similarly, very high silicate concentrations are required to precipitate uranophane in preference to autunite. The equilibrium constant for the conversion of uranophane to autunite according to equation (13) has been estimated" as 1022.6.In general the autunites and uranophanes are rarely precipitated from groundwater, except near to source rocks which have a low vanadium content but give rise to high groundwater phosphate or silicate concentrations. The optimum pH range for the precipitation of uranyl minerals is pH 5 4 . 5 . This pH range also corresponds with the maximum sorption or uranyl on to organic materials, clays ccc6-cc*

I 892

Applications in the Nuclear Fuel Cycle and Radiopharmncy

and metal oxohydroxides. At higher pH the formation of carbonate complexes or species such as (U02)3(OH)5+leads to dissolution. At lower pH, UO;+ or its complexes with fluoride or sulfate dissolve. Uranium minerals, including carnotite, are most stable, and thus least soluble, under low partial pressures of C 0 2 . This in regions where carbonated groundwaters approach the land surface, and equilibrate with a lower atmospheric C 0 2 content, mineral deposition will be favoured. Kpy= [H2W4-]/[ H2V04-]

KJ(UO,)z(PO4)J + 2HZVO4-

4

= lo2

K2[(UOJ2(VO4)J

(11)

+ 2HzP04-

(12)

(13)

Ca[(UOZ)z(Si0,0H),]+2H,P04+2H+ + Ca[(U0,)Z(P0,)Z]f2H,Si0,

The structures of several uranyl minerals have been investigated using X-ray techniques. Carnotite was f o ~ n d ' to ~ ,have ~ ~ a similar structure to C S ~ [ ( U O ~ ) ~ ( V ~which O ~ ) ]contained , infinite sheets of formula [ ( U O , ) , ( V , O , ) ] ~ " ~separated by Cs+ ions. The uranium had a distorted pentagonal bipyramidal coordination geometry in which a monodentate and two bidentate [V20J6- ligands contributed to an equatorial belt of five oxygens, as shown schematically in 6. The Figure

Vanadium

@ Uranium

0

Oxygen

0 Cesium

Figure 6 The uranium environment in carnotite (Reproduced from M. Ross, H. T. Evans and D. E. Appleman, Am. M i n e d . , 1964, 49, 1603, with permission from the Mineralogical Society of America)

[V,0s]6- ions were formed from two distorted square pyramidal [VO,]'- units sharing an edge. The structure of meta-autunite, Ca[ (U02),( P04)2]-SH20,was relateds4to that of a meta-torbernite, Cu[ (U02),(P04),].6H20, which contained sheets of formula [ U02(P04)],"-. These were separated by planar [CU(OH,),]~+units which were also involved in a network of hydrogen bonds. In this case the uranium was in an octahedral coordination environment resulting from the trans oxo ligands and a square of oxygen atoms derived from four monodentate PO:ligands. The silicate like carnotite contained'5756uranium in a minerai uranophane, Ca[H2(U0,),(Si0,),f~5H,0, distorted pentagonal bipyramidal geometry, the equatorial belt of five oxygens being derived from one bidentate and three monodentate Si0,"- ligands as shown in Figure 7 . The mineral structure contained sheets of formula [ (UO,),(SiO,),],""- linked through a network of hydrogen bonds and C a - 0 interactions. Another major processs involved in uranium mineral formation is the reduction of uranyl complexes to UIVspecies of low solubility. This process may occur in regions where uraniferous groundwater passes into a zone of low E,. The &-pH diagrams27for the U-O2-CO7-H20 and K-U-V-02-H,0-C02 systems at 25 "C are presented in Figures 8 and 9. These indicate that, in the presence of vanadium, carnotite will precipitate between pH 5 and 8 even under oxidizing Et, conditions. However, in the absence of vanadium, insoluble U 0 2 or U 4 0 9 are only formed when the E,, falls below about 200 mV. Reducing conditions of this type may result from mixing with anoxic groundwater containing reducing agents such as Fez+ or dissolved H2 S. Equations (14)

Applications in the Nuclear Fuel Cycle and Radiopharmacy

893

Figure 7 The uranium environment in uranophane. (Reproduced from D. K. Smith, J. W. Gamer and W. N. Lipscomb, Am. Mineral., 1957, 42, 594, with permission from the Mineralogical Society of America) 1.0,

.

!

PH

Figure 8 The EJpH diagram for the U/O,/COz/H,O system at 25 "C with a partial COz pressure of lo-* bar?' M dissolved uranium species; U0,am = amorphous UO, Solid-solution boundaries are drawn at

and (15) exemplify the formation of uraninite at pH 8 by such a process. The evolution of COz and the formation of stable Fe(OH), (solubility product = or SO:would serve to drive these reactions to completion. Another mechanism for the formation of U'" minerals is provided [U02(C03),]4-+ 2k2'+3H20 4[UO,(CO3),I4-+HS-+

-

U02+2Fe(OH),+ 3C02

15H* -+ 4U02+S04'-+ 12CO2+8H2O

(14) (15)

894

Applications in the Nuclear Fuel Cycle and Radiopharmacy

PH

Figure 9 The EJpH diagram for the K/U/V/O,/H,O/CO, systemz7at 25 "C with bar at solid sofution boundaries and at a CO, partial pressure of

M' K and

M total uranium

Table 5 Thermodynamic Data for the Binding of Uranyl to Humic Acid" Aqueous medium

Freshb Freshb Sea" Sea"

Complex" UO,(humate) i = 1 UO,(humate), i = 2 UO,(hurnate) i = 1 UO,(humate), i = 2

5.11 f0.02 8.94 f0.03

7.6 11.5

-29.2

f0.1

-51 Zk0.2

-

-2.7

* 0.4

+8*4

-

89 200 I

-

Humic acid of 4.2 meq g-' ionizable acid content. 8 . 4 lo-' ~ eq dm-' humic acid at pH 4.04 with 0.09 M NaC10,+0.01 M glycollic acid Fresh water containing ,"-", where x = 1-6, and hydroxy nitrato complexes. However, species containing Zr02+ are not expected to be present since this ion is unstable in aqueous media and is rapidly hydrated3" to Zr(OH):+. The extraction chemistry is further ~ o m p l i c a t e d ~ by' ~the formation of inextractable polymeric species when the ZrIVconcentration exceeds ca. lop2M. An example of such oligomerization is afforded by the [ Z T ( O H ) ~ ( H ~ O )ion ~ ] ~which ~ + contains four Zr"' ions in a square arrangement linked by two p-OH ligands on each square edge.313Four water molecules complete the Zr"' coordination sphere in an approximately DZddodecahedral geometry. The extraction of ZrIVby pure TBP/OK phases increases rapidly with aqueous phase acidity so that Dzrfor 19% TBP/OK increases from ca. at 0.5 M WN03 to ca. lo2 at 13 M HN03.314 Satisfactory decontamination from zirconium can thus be achieved using aqueous phase acidities in the range 2-3 M where DZrwill be ca. 0.1. T h e extraction reaction was described315by equation (158) but more recent work3I6also indicates the presence of Zr(OH)2(N03)2*2TBP and Zr(0H)Zr(OHL2+(,,) + 2H+,,, +4NO,-(,,)

+ZTBP,,,,)

* Zr(N03),-2TBP~,,~+2H20(,,)

(158)

944

Applications in the Nuclear Fuel Cycle and Radiopharmacy

( N03)3-2TBPwhen TBP concentrations below 0.3 M are in contact with nitrate concentrations below 0.6 M. There is also evidence for a monosolvate Zirconium can give rise to separation problems if the TBP radiolysis products HDBP and H2MBP are allowed to accumulate. At low Zr4+ concentrations, Zr(N03)2(DBP2H)2 is extracted from nitric acid by HDBP.,19 At higher concentrations, Zr(N03)2(DBP)2may form and precipitate as a polymeric solid. Such precipitates may act as emulsion stabilizers, preventing phase disengagement and impairing process performance. The addition of H2MBP to extracts of Zr4+ into TBP containing HDBP led to the formation320of a solid of composition Zr(OH),(NO,),.DBP.MBP, where x + y = 1. The extraction behaviour of Zr"' appeared to be determined by the HDBP present but its retention and precipitation were determined by H2MBP even where an excess of WDBP over H2MBP was present. H,MBP alone gave rise to gelatinous solids containing hydrated Zr(MBP), . TR studies321have also confirmed the presence of MBP and DBP in solids precipitated by aged organic phase Zr"' extracts. Other solvent degradation products derived from the diluent as well as from TBP may arise. Butyl lauryl phosphoric acid (HBLP) has been as a model for diluent derived diesters of M enhances Zr"' extraction by TBP. and at concentrations above 2 x phosphoric HBLP is not removed by alkali washing of the solvent but BLP- is transferred into water at low ionic strength. Dialkylphosphorane complexes of Zr"' may also arise from irradiation and can stabilize emulsions.323Diluent derived hydroxamic acids are present at too low a concentration (10-8-10-9 M) to completely account for the extent of zirconium retention which occurs in irradiated TBP/OK extraction^,^^" although Zr"' does form hydroxamate c~rnplexes.'~~ The relative order of extractability of metal cations by dialkyl phosphoric acids was found305to be Zr4' > Pu4+ > UO:+ > {Ru(NO)},+ > Nb". However, the mass action effect of the UO;+ present is to complex virtually all the HDBP and some of the H2MBP present. The residue will then A single pass of solvent through a pulsed column contactor largely be complexed by Zr"" and h4+. in the first decontamination cycle might give rise to ca. lop5total alkyl phosphoric acids.306Under such conditions, FP decontamination is not significantly affected. Niobium occurs in oxidation state (V) in the dissolver solution and does not extract into pure TBP/OK phases to a significant extent. Values of ca. 5 x lo-" have been for DNb between 20% TBP/OK and 2 M HNO, . Inextractable polymeric hydroxy complexes of NbV are formed but the presence of silicates329or HDBP3' can lead to some extraction of niobium species. A model for the extraction by HDBP is provided by HBLP, which was thought to extract NbV HDBP might similarly lead to some according to equation (159) on the basis of IR synergic extraction of Nb" in TBP/OK but, as mentioned above, free HDBP will not be available in the first process and in practice problems do not usually arise with niobium. Nb(OH),,,,,+HBLP,(,,,)

Nb(OH)4BLP.HBLP(,,,,+HzOjas)

(159)

Molybdenum provides an example of a fission product which would appear in the dissolver solution in oxidation state (VI). However, the half lives of the molybdenum isotopes produced are sufficiently short for decay processes largely to eliminate this element before fuel reprocessing is undertaken. The extraction of MeV' by TBP from aqueous nitrate solutions of pH 0.5-3.0 has been investigated.332It was found that at Mow concentrations above lop4M, slow aqueous phase polymerization reactions gave rise to time dependent extraction behaviour. Thus DMov] between TBP and a solution containing 1.0M KNO, and 1 . 5 ~ 1 0 - ~ MoV1 M at p H 3 rose to 0.1 over a period of 12 hours. At Mo"' concentrations below M7 equilibrium was rapidly attained and in this case DMo"was at a maximum of 0.08 between pH 0.5 and pH 1.0, falling off steeply at higher pH values. The extracted species was found to be Mo(OH),.HNO,. A far more important fission product is 99Tc,which appears in oxidation state (VII) as Tc04in the dissolver solution. In the past, 99Tcwas not a major concern because of its low specific activity and the low energy of its &emission. However, much larger quantities of this isotope are now arising from high burn-up power reactor fuel. Furthermore it has a long half-life (2.1 x lo5y) and the dissolved Tc04- ion is highly mobile in the environment. These factors combine to make 99 Tc extremely important as a component of HAW in long-term storage because of its potential Reactions of Tc04- with reductants used for plutonium valence control environmental eff e~ts.3'~ may also have implications in process design." The extraction of Tc0,- from nitric acid media by TBP/OK has been reported to follow equation (160) in the absence334of UO+ ; and equation (161) in the of UO;.' The values of DTpbetween nitric acid and TBP/OK were to increase with H N 0 3 concentration to a maximum value between 0.6 and 1.0 M H N 0 3 before falling of? because of competition

Applications in the Nuclear Fuel Cycle and Radiopharmacy

945

between H N 0 3 and HTc04 for the available TBP. At 25 "C the maximum DTCv111value with 30% TBP/n-dodecane was 0.9, falling to 0.18 at 60°C. The corresponding values for 20% TBP/ndodecane were some four times smaller. Values of 75 J mol-' for AG and -197 J mol-' K-' for A S were obtained for the extraction reaction.336In the absence of nitric acid the extracted species was found to be HTCO~.TBP.(TBP.H~O)~, a similar species having been proposed to account for the extraction of HReO, by TBP.338-340 H+(,q,+TC04-(aq)+3TBP~,,g)e HTc0,.3TBP,,,) UOz2++N0,-~,,)+TcO,-~,)-1-2TBP~,,,

e UOZ(N03)(Tc0,).2TBP(.,,

( 160)

. (161)

The effect of added nitrite, nitrate and uranyl ions on Tc04- extraction has been M had little effect on DTcvil but the Aqueous NaNO, concentrations between and presence of NH4N03 concentrations from 1 to 4 M was found to depress DTCv11 because of the salting-out effect on aqueous phase TBP. Added UOz2+ ions gave rise to DTCv11 values which decreased with increasing €€NO3concentration and did not show the maximum at 0.6 M HN03 observed in the absence of UOp+. Studies of the extraction of UO,(TCO,)~itself have shown337 that Tcv" extraction according to equation (162) could occur. Thus in the UO;+/TcO,-/ NO3-/ H20-TBP/OK system, different species were present in the organic phase depending upon the relative concentrations of the aqueous phase components. At aqueous UOZ2+ concentrations above M, U02(Tc04)2.2TBP was extracted, while at concentrations below 3X M, U02(N03)2.2TBPwas extracted along with HTc04-3TBP. The mixed anion complex UO2(NO3)(Tc0,)*2TBP was only observed in the organic phase under a limited range of conditions, viz. for UO2+ concentrations between 3 x loL4and 3 x lop3M at acidities between 0 and 0.1 M or UO?+ concentrations between and 10' M at acidities between 0.1 and 1.0 M. This species was not found at acidities in excess of 1 M. However, studies using 6 M HN03 gave results which were consistent with UO,( N03)(Tc04),*3TBP as the stoichiometry of the extracted species. The presence of hydrazine nitrate or iron(I1) sulfamate at concentrations between and M was not found34' to affect DTCvl1 significantly while DBP had only a small effect. Empirical numerical equations have been d e ~ e l o p e d ~ ~to~ describe .~" the extraction behaviour : present together under conditions appropriate to the Purex process. of Tc0,- and UO+

* U02(T~0,)2~2TBP(,,g)

U022+,,,+2T~04-(aq)+2TBP~,,,I

(162)

The extractability of Tc0,- in the presence of UOf may result in a few percent of the 99Tc present following the uranium stream. lo~ll Furthermore, the conditions which favour the separation of NpVfrom Uv' may promote Tcv" extraction to give low DFrcvalues. An additional complication arises from the reactions of Tc04- with reductants and stabilizers used to produce Pdl' in the U/Pu separation section of the process.343These reactions may have a detrimental effect on U/Pu separation. They also give rise to lower oxidation state technetium species which are poorly extracted and may compromise the purity of the plutonium product stream." Recent studies have also that zirconium and technetium may coextract in the first Purex cycle, leading to lower DFT, values than previously expected. The extracted species is a 1: 1 TBP solvate possibly of formula Zr( N0J3(Tc04).TBP. Careful process design and control are thus necessary to achieve acceptable DFTcvalues for high burn-up fuels and the control of technetium behaviour in the Purex process still presents a major challenge to fuel reprocessors. Ruthenium provides t he only example of a fission product which might arise in oxidation state (VIII) as RuO,. The formation of this compound presents special problems because of its organic phase solubility, volatility and reactivity. Fortunately, conditions in Purex process streams do not normally give rise to RuO, since temperatures above 70 "C are necessary for RuO, formation at nitric acid concentrations below 8 M. Furthermore, the presence of nitrous acid or nitrogen oxides would lead to the rapid reduction of Ru""'. However, Ru04 may arise when nitric acid solutions are evaporated or thermally denitrated, for example when HAW is vitrefied to produce solid waste." Where the formation of RuO, is possible it will be necessary to design process off-gas facilities which will absorb it. In fact the thermal oxidation of irradiated oxide fuel in the 'Voloxidation' may be used as a means of removing volatile fission products, including ruthenium, from spent fuel.

(iii) Chemistry of the actinides in the Purex process Several general accounts of actinide chemistry are available,207*242*345*346 including those in specific volumes of the Gmelin Handbook on chemistry in s o l ~ t i o n , ~the ~ ~solvent * ' ~ ~ extraction

946

Applications in the Nuclear Fuel Cycle and Radiopharmacy

of uranium272and irradiated fuel reprocessing.349The solution chemistry of the actinides has also been reviewed by Ahrland et aL90 and, more recently, with the emphasis on thermodynamic considerations, by C h ~ p p i n . ~In " addition to uranium and plutonium, spent thermal reactor fuel will also contain significant amounts of neptunium along with some americium and curium.I3 These will be present in the dissolver solution in oxidation states which depend upon the element and whether there has been any chemical conditioning of the nitric acid solution, for example with nitrogen oxides or fluoride. After dissolution the solution will contain little nitrous acid." However, conditioning with nitrogen oxides is often carried out to convert the plutonium to p U I v so that a nitrous acid concentration of 0.05-0.1 M may arise at this stage. After air sparging this M HNO,. may fall to ca. The increasing stability of oxidation state (111) with increasing atomic number in the transuranium elements results in the higher actinides, americium and curium, appearing in oxidation state ( I l l ) in the dissolver solution. The potential for the oxidation of Am'+ to Am0;+ in 1 M HCIO4 is reported3s' to be 1.69 V so that AmV' is only formed under strongly oxidizing conditions. In aqueous bicarbonate solutions, AmO:+ is reduced rapidly3" by nitrite to b o 2 + . This ion in turn rapidly d i s p r o p o r t i ~ n a t e s ' ~by ~ ,a~ complicated ~~ mechanism, the overall effect of which is represented by equation (163). Rapid disproportionation of Am4+ also occurs353 according to equation (164). Consequently the americium present in the dissolver solution will essentially all be converted to oxidation state (111). The stability constants for the complexation of Am3+in media of unit ionic stren th have been reported as p1= 1.4, p2= 0.04 for nitrate354at 26 "C and PI = 9.1 1 i0.56 for nitrite3' at 25 "C.The extraction of Am3+from nitric acid solutions by TBP was shown"' to involve the trisolvate Am(N03),.3TBP, as for other trivalent actinides. The value of D , between 35% TBP in a hydrocarbon diluent and nitric acid was found358to increase with acidity to a maximum of ca. 0.03 between 2 and 3 M HNO, before falling off and then rising steeply again above 10 M HNO, . The thermodynamics of Am3+extraction and values of from 2 M LiNO, at pH 2 by 0.25 M TBP in dodecane have been AH = -41.8 kJ mol-' and AS = -147.1 J mol-' K-' obtained. Under Purex conditions the americium and curium, which behaves similarly, will thus pass with the FPs into the HAW stream from the decontamination cycle. Much higher DAmvalues are obtained with acidic phosphorus ester extractant~~'~so that the presence of HDBP from solvent degradation could impair decontamination from americium. However, the mass action effect of the UO;+ present is to remove any free HDBP so that americium retention in the solvent does not normally arise.

= 2Arn022++h3++2FIz0 2Am4++ZHz0 = Arn"+AAm02++4H+

3Am0,++4H+

P U O ~ ~ + + H N O ~ +3 ~ HP' u 4 + t H N O j + H 2 O 3Pu4++2H20

2Pu3++Pu03++4H+

(163) (164) (165) (166)

Plutonium is unusual among the actinides in that all four oxidation states from Pu"' to Puv' can coexist in aqueous solution. However, it is Pu4+which predominates in the dissolver solution. The PuO;+ also present initially will be reduced by any nitrous acid3'l in the solution according to equation (165). Further conditioning of the dissoher solution with nitrogen oxides is sometimes used to ensure that all the plutonium is converted to Pu4+ before the first extraction cycle of the h r e x process. Any Puv present would also be slowly reduced by H N 0 2 , but at acidities above lop2M, rapid d i s p r o p o r t i o n a t i ~ n of ~ ~hO,* ~ . ~ ~ ~occurs according to equation (155). At low acidities, disproportionation of Pu4"'may also occur, The overall process is described by equation (166), which implies a dependence upon [H+I4 for the equilibrium constant. A dependence on [H+]s.3 was reported for nitric acid media362and attributed in part to the effects of nitrate complexing. More recent have indicated an equilibrium constant dependence on [Hc]3.2 and [NO3I2-'. The thermal oxidation of Pu4+to PuO;+ in 0.3 M HN03 has also been in~estigated~'~ and appears to involve disproportionation followed by reoxidation of the h3+ formed to h4+, the overall effect being a conversion of Pu4+to PuO:+ which becomes significant at temperatures above 40 "C. The attainment of equilibrium in the disproportionation reaction is accelerated in the presence of UO," and a cation-cation complexation model was based on a stability constant of 0.8 f 0.016 mol-' for the formation of the h r v * U Vcomplex. ' The disproporat ambient temperature is unlikely to require consideration in the Purex process tionation of h4+ unless acidities below cu. 0.5 M are anticipated. At lower acidities still, below ca. 0.1 M, the The rate of polymer formation hydrolysis of Pu4+to give a polymeric colloid may also occur.366,367

Applications in the Nuclear Fuel Cycle and Radiopharmacy

947

was found to be third order in [Pu'"] and was retarded by UOz( The structure of the polymer remains uncertain although bridging hydroxy groups may be involved, as proposed"' for the { P U ( ~ - O H ) ~ P Ucore ) ~ +in Pu(OH),(S0,)-4H20 formed by thermal hydrolysis of Fu(SO,)~. Reactions such as these place a lower limit on the acidity acceptable for aqueous plutonium carrying streams in the Purex process. At nitric acid concentrations in excess of 10 M the PU'" is present369as [PLL(NO,),]~-,while at ionic strength 1.9 the stability constants for the formation of the lower nitrate complexes of Pu4+ have been reported242 as p1=4.0, p2=7.5, p3=4.0 and p4=1.2. Typical aqueous phase acidities in the first extraction cycle of the Purex process might be between 1 and 3 M so that essentially all the plutonium would be present as Pur".

0

2

4

6

8

1 0 1 2 1 4

HNOi IM1

Figure 43 T h e extraction of actinides from HN03 by 19% TBP/OK2",37"

The extraction of Pu4+from nitric acid by TBp272involves the solvate P u ( N O ~ ) ~ * ~ T formed BP in a reaction stoichiometry analogous to that proposedzo0for Th4+and shown in equation (167). The variation in Dhlv with HN03 concentration is shown in Figure 43 for extraction by 19% TBP/OK. This shows that PUlVis more extractable than UtVor Np'" but that Pun is substantially less extractable than Uvr.200~37* The thermodynamics of Purvextraction have been i n ~ e s t i g a t e d ~ ~ ' and it was found that the unfavourable enthalpy term was off set by a positive entropy term arising when forming Pu(N03),.2TBP. In the presence of Uv' the extraction from the dehydration of hi4+ of P u I V is as illustrated in Figure 44.However, D h r v remains sufficiently high that essentially quantitative extraction is possible in the first cycle. ~ ~ ~ a ~ ) + 4 N 0 3 - ( , ~ ) + 2 T B P+~ ,MNOJ,.2TBP(,,,, ,p)

(167)

In order to effect a separation of uranium and plutonium it is necessary to adjust the plutonium oxidation state to (111), which is poorly extracted.373Provided the uranium remains as Uvl a partition is then possible with Uvl remaining in the organic phase while hrr' passes into the aqueous phase. Contacting the hi'" and U"' loaded solvent with an aqueous phase containing a suitable reducing agent will thus allow selective back extraction of the plutonium. The potential diagram3" for plutonium at pH 0 is shown in Scheme 6. In the past, iron(I1) sulfamate has been successfully used to effect this reduction. Unfortunately, this reagent is consumed in large excess over its stoichiometric requirement and contributes nonradioactive salts to the active waste streams. The reason for the nonstoichiometric behaviour and the choice of sulfamate as the anion lies in the fast autocatalytic reaction between PuIII and nitrous a c i d y which may be by equations (168)-(170). Nitrite oxidation of Fe2+ to Fe3+ may also occur according to equation (171). These reactions can result in the consumption of the reductant with little net change in plutonium oxidation state. To prevent this it is necessary to add a stabilizer which can remove CCC6-EE

948

Applications in the Nuclear Fuel Cycle and Radiopharmacy

L

30.0

I

\

14.5M 0

I so

I

I

I 150

100

( U )ora

la 1-7

Figure 44 The variation in D,lv between HNO, and 30X TBP/dodecane at 22 "C with increasing organic phase uranium concentration3"

nitrous acid at a rate which competes with its autocatalytic generation in the Pu3+ oxidation reaction. The most suitable reagents for this are sulfamate, which reacts according to equation (172), and hydrazine, which reacts with excess nitrous acid according to equations (173) and ( Hydrazine has the clear advantage that its reaction products are gaseous and do not contribute to the process waste streams. However, the possibility that hydrazoic acid or metal azides may appear in the process streams must also be considered. Fortunately the azide cornplexing of actinide or lanthanide metals is weak compared with that of d-block metals?79Hydrazoic acid is also readily destroyed in nitric acid media by heating, the products of the reaction being N20, N2 and HN02

h3++ NO2

+ HNO, e 2 N 0 2 + H2O

e h4++ NO,-

N02-+H+ Fe2++NOz-+2HC

(rate controlling)

(169)

* HN02 +

Fe3++NO+H20

NH,SO,- -k NO>--+ N2 + SO:N,H,+ HNO2

(168)

-L

+ H20

HN, + 2H2O

Even in the presence of stabilizers, more than the stoichiometric quantities of reductant will be consumed. This largely results from the inextractability of NH2S0,- or hydrazine, present as N2H5+.Nitric acid present in the organic phase reacts with traces of H N 0 2 , which is also readily Since the e~tracted,~"to produce NO2 which can autocatalytically react with traces of h3+. stabilizer is not extracted the P u 1 I 1 - P u I V oxidation cycle can proceed unchecked in the organic phase, regenerating Pur". This process is rapid at organic phase nitric acid concentrations of 0.2 M and above?" Consequently, an excess of reductant is necessary to maintain process control in the partition cycle. Stabilizers which are soluble in the organic phase might be used to reduce reductant consumption but those considered so far contribute organic products to the process pu~r~

pu~v U 7 V puv

I

1.04V

Scheme 6

puv~

Applications in the Nuclear Fuel Cycle and Radiopharmacy

949

streams and offer an unacceptable alternative. Where hydrazine is used the effect on the distribution of PuIVand Uv' of salting by both the N2H5*N03 and the Pu( N03)3present in the aqueous phase may need to be con~idered.~'~ Alternatives to Fe" as a reducing agent for P u I V are provided by UIVand by hydroxyammonium nitrate (HAN).38473s5 In the case of HAN present in excess the dominant reduction reaction is represented by equation (175). However, where the Pu'" concentration equals or exceeds the HAN concentration the reaction shown in equation (176) becomes important.385The rate of reduction was found to be inversely related to [H+I4 as shown by equation (177), where k'= 0.029~0.008mol' s-' and Kd,the dissociation constant f0r{Pu(N0,)}~+, = 0.33 f 0.15 mol at 30 "C. This rate becomes too slow to be acceptable at acidities above ea. 1 M so that HAN is not a suitable reductant for the first Purex cycle. However, its use in later cycles where lower acidities prevail may offer advantages since HAN is decomposed to N 2 0 by heating in nitric a ~ i d ~ * ' , ~ * * and does not contribute to the HAW. This reaction involves both the oxidation of HAN by H N 0 3 and its reaction with HN02, as shown in equations (178) and (179). Hydroxylamine also reduces Puv' in a mechanism involving reduction to Puv, which disproportionates to give Puvr and Pu'''. The PulI1 is then an effective reductant for Puvl and is itself converted to PuIVwhich is reduced by NHzOH to regenerate h I r l in an autocatalytic reaction sequence.386Other oximes have been found more effective than HAN for PuIV reducti0n,3~' but these contribute organic by-products to the process streams. A combination of HAN and iron(l1) sulfamate is now used in the second cycle of the h r e x process at the Savannah River Plant in the USA.388 2NH30H++2Pu4+ + 2Pu3++ N2+2H,O+4H+

(175)

2NH30H++4Pu4+ 4 4 h 3 + +NzO+H 2 0 + 6 H +

(176)

-d[ PU'~] ___-

[Pu'"]~[NH~OH+]~

I

dt

k'[Pu"']2~H']4(Ka+~N0,-])2

NH,0H++2HN03

+

3HN02 + H30+

NH30H++ HNO,

+

N,O+ H,O

(178)

+ HJO+

(179)

Like Fe", UIVreacts rapidly with Pu1v.389-391 The rate was found to vary inversely with [H+]* and the reaction proceeded according to equation (180). A stabilizer is necessary to prevent the oxidation of U" by nitric or nitrous acid according to equations (181)-(183).392 The oxidation 2PuA++ U4++2Hz0

U4*+ NO3- + H,O

e

2 h 3 +t UOzz+t 4HC

4

U02'*

f

HNO, -b H+ (slow)

U4++2HN02 4 U02**+2H++2N0

2NO+ H N 0 3 + HzO

3HN0,

(180)

(rapid)

(rapid)

(181)

(182) (183)

by nitrate is favoured at low acidity while Pulll oxidation is favoured at high acidity. The potential diagram for uranium oxidation states at pH 0 is given in Scheme 7. The UIV may be generated U"'

U'V

1

061V ,

0.063 V u"

,,v 0.30 V

i

Scheme 7

externally and then fed into the process stream or it may be produced in situ, by electrochemical reduction. Photolytic production of U"' has also been investigated393but, in the presence of hydrazine, photoexcited UO;+* reacts to produce ammonia, which is an unacceptable byThe reactions occurring during the in situ electrolysis are s u m m a r i ~ e d ' ~ in, equations ~~~ (184)-(191). Both hrv and Uvl are reduced at the cathode but Uv' reduction proceeds via Uv, which disproportionates according to equation (187). A polarographic of U" in TBP solution has shown that a one-electron reduction process may also occur in the organic phase. The reduction of PulVby UIVthen proceeds according to equation (180). Nitric acid may also be reduced to nitrous acid according to equation (188), while any hydrazine present may be oxidized at the anode according to equation (189). Any neptunium present may also undergo reduction at the cathode according to equations (190) and (191). Atmospheric oxygen has been

950

Applications in the Nuclear Fuel Cycle and Radiopharmacy

found to oxidize U'" in TBP/dodecane-HNO, mixtures.397In the organic phase the reaction was slow but could be accelerated by autocatalysis. A more rapid reaction occurred in the aqueous phase, which was zero order in [U"'], but both Purr' and Fer' were found to retard UIV oxidation in mixed phase dispersions. Pilot scale trials have indicated that in situ electrolysis can produce U/Pu separation factors an order of magnitude higher than when reductive back extraction by solutions of UIVis used." In the latter case a five- to ten-fold excess of U" is required to obtain a separation factor of PU from u of ca. io4. pu4++,-

e

pus+.,

Eo=0.92V

UOZ2++4H++2e- e U 4 + t 2 H 2 0 ; 2U0$++2e-

E0=0.33 V

e 2U02+ (at the electrode);

2U02++4H++4H20 N03-+3H++2eN2Hs+ -+ N,+ 5H'

E0=0.06V

(185) (186)

e U0;++U4++6H20

(187)

e HN02+H,0;

(188)

+ 4e-;

NpOz2++e- G NpOZf; NpOzC-t 4H+ + e-

( 184)

E,

Eo=0.94V

= 0.23 V

E0=1.15V

e Np4++ 2H,O;

E,, = 0.75 V

(189) (190) (191)

The extraction of F d + by TBP/OK resembles that of Am3+in that the trisolvate F'U(NO,)~+~TBP is extracted357g398 in a stoichiometry analogous to that shown for Ln3+ in equation (157). The value of D P pfor 19% TBP/OK-HNO, reaches a maximum of ca. 2 x lo-' at 2 M HN03373SO that effective back extraction of Pu"' from the organic phase is possible. Where UIVis used as the reducing agent the uranium will be retained in the organic phase as U(N03)4*2TBp99 and UOz(N03)2.2TBP,2003383 although Dulv values348are lower than D h l v , as illustrated by Figure 43. The P u l I 1 may readily be reoxidized to P u I V , prior to extraction in subsequent purification cycles, by conditioning the solution with nitrogen oxides. The general coordination chemistry of plutonium in aqueous media has been reviewed by Cleveland4" and the equilibrium constants for the plutonium oxidation states PuIV/Puv, P U ' ~ / P ~and ' pUv/Puvl have recently been re-evaluated,4°1 good agreement with existing data being found. Short-lived 239Npis formed during the irradiation of 238U containing fuels as shown in Scheme 1. The amount of 239Nppresent will decline substantially during the cooling period between the discharge of the fuel from the reactor and its admission for reprocessing. However, the long-lived isotope 237Np( tli2 = 1.7 x lo6 y) is also formed and does not significantly decay during the cooling period. The presence of this isotope makes it desirable to divert neptunium into the HAW stream from the first Purex cycle, along with the FPs and americium. If uranium is to be recycled for enrichment it will be necessary to achieve DFN, values of ca. lo4. In nitric acid, neptunium is commonly encountered in oxidation states (IV), (V) and (VI). The aqueous coordination chemistry of neptunium has been reviewed in genera1402 and, more specifically, in relation to the Purex The redox potential diagram350for neptunium in 1 M HClO, is shown in Scheme 8 and the potential for the NpV'/NpV couple in 1.02 M HN03 was reportedz4' to be 1.138 V. After fuel dissolution the neptunium is predominantly present in oxidation state (VI)." However, in the presence of nitrous acid, NpV1will be reduced to NpV according to equation (192), but at high acidity and low (less than lop3M) nitrous acid concentration, reoxidation of the Np02' by NO3- can occur.1oDisproportionation of the NpO,+ is also possible according to equation (193), 2Np022++HN02+H20

2Np02++3H++N03-

2Np02++4H+ e Np4++ Np0;++2H2O

(192) (193)

although this is slow at low acidities. Thus Np02+ is essentially stable in the range 0.1-2 M HNO, but above 6 M HN 0 3 will be entirely converted to NpO;+ in a reaction catalyzed by traces of

NpIII

ENp'V

NpV

0.94 V

Scheme 8

3Npv' J

Applications in the Nuclear Fuel Cycle and Radiopharmacy

95 1

HN02.'0,404In h r e x feed solutions of less than 3.5 M FINO3, over 89% of the neptunium was found to be NpV,falling to CQ. 85% in 4.8-5.3 M HN03. No Npv' was found in the feed s o l ~ t i o n ~ ~ ~ The reduction of Npv to NpIVby a variety of reducing agents, including HAN,406hydrazine407 and U1v,408has been investigated. At ambient temperatures the reduction by HAN or N2H4 was found to be slow but was accelerated by increased acidity or temperature.407At 61 "C and an ionic strength of 4 the rate of NpV reduction was given by equation (194) where k = (4.54k0.09)x lop2mol-2-65min-'. The reaction mechanism was complex but could be represented by equations (195)-(199). Rapid reduction of NpV' by HAN was also observed accordin to $ equation (200) or, in the presence of excess NpV, equation (201). The reduction of Np by hydrazine was found406to proceed according to equation (202) with the diimide N2H2decomposing to give either +NT+iN2H4 or $NH,+;HN,. The reaction was catalyzed by Fe"', although Fe"' will itself oxidize Np" to NpV. Oxidation of Np'" by nitrate also occurs according to equation (203).409A five-step mechanism has been proposedw8 to describe the reduction of NpV by UIV, the overall reaction being represented by equation (204). The thermodynamic activation parameters for this reaction were evaluated as AG* = 84k0.84 kJ mol-', AH* = 38* 1.3 kJ mol-' and AS* = 37 f 1.4 J mol-' K-'. Weak cation-cation complex formation between NpV and Uv' has been observed and is of an outer sphere n a t ~ r e . ~ " -d[NpV1 = k[NpV][NH30H+][H+]'.65

dt

N p 0 , H 2 + t NH30H+ -+ 2NHOH

-B

2NHOH+2NpOzH2+ -D

NP(OH)~++NHOH+H,O N2+2W20 (fast) H,O+N2O+2Np(OH);+

NP(OH)~++H++ Np4++H,0 Np(OH),'*+2H+ 2Np0;++

-P

(slow)

Np4++2H,0

(195)

(196)

(fast)

(fast)

(fast)

2NH30H+ + 2NpOz++N2+2Hz0+4H+

(197) (198) (1%)

(200)

4Np02++ N 2 0 + H,O+ 6H'

(201)

2Np0,HZ++NzH5++3H++ 2Np4++ N2H,+4H,0

(202)

4Np0;++2NH30H+

2Np4'+N0,

-+

+3H,O

-+

v'+NpO?+ e

2Np0,++HNOz+5Hf

(203)

U 0 2'+ Np4+

(2041

The extraction of neptunium by TBP is highly oxidation state dependent. Thus Np" and Npv' are and intermediate in behaviour between the corresponding uranium and plutonium species of the same oxidation state, as shown in Figure 43. The solvates Np(N03)4.2TBp11and Np02(N03)2.2TBp12are extracted in reactions analogous to those given for P u I V and Uv' in equations (167) and (1 10).*OoIn contrast, NpVis poorly extracted with DNpv for 30% TBP/OK and 1 M HNO, being 1.7 x lov2for a process which involves the monosolvate Np02(N03)-TBP.413 Thus careful control of the neptunium oxidation state is necessary to ensure an effective separation from P u I V and Uv'. Provided it is maintained in oxidation state (V), neptunium will pass into the aqueous raffinate of the first extraction cycle. However, control of the course of neptunium through the Purex process can sometimes present problems and in some cases conditions are arranged to allow the neptunium to pass through with the uranium before being removed in the uranium purification c y ~ l e s . ' ' ' ~ ~ ~ The chemistry of uranium in the Purex process is simpler than that of plutonium or neptunium. In the dissolver solution, Uvl is present exclusively. The lower oxidation states (IV) and (V) are both unstable in nitric acid media, Uv disproportionating rapidly and U" being formed only in the presence of medium strength reducing agents. In the absence of a stabilizer, U" is prone to oxidation by air and nitric or nitrous acid as discussed above. The extraction of U" by TBP4' has been described in Section 65.2.3.1 and Duvl values are larger than those of D h v l or D N p w under comparable conditions, as shown in Figure 43. The presence of nitrous acid at concentrations in excess of 5 x M has been found to suppress Uv' extraction4I5 because of its competition for the TBP, as shown in equation (205). In its turn the Uv' present in the organic phase exerts a mass action effect on the FPs, present at much lower concentration, enhancing their retention in the aqueous raffinate of the first cycle." H+(,,,+No,-(,,,+TBP(,,,,

*

HNo,'TBP(O,g,

(205)

952

Applications in the Nuclear Fuel Cycle and Radiopharmacy Table 12 Thermodynamic Data for Actinide Extraction from 3 M Nitric Acid by 30% T B P / x ~ l e n e ~ ~ ~

Actinide ion

(kJ mol-')

AG

AH (kJ mol-')

ThIV

-8.11

-10.75

UtV Np'" pur"

-

1.57

10.50 6.02 -11.92

-11.71 -18.70 -11.50

UV'

-

Np"' Puvl

-20.12

-13.01

-

AS (J mol-' K-') -8.83

-

74.5 82.8 -1.44

-

I

The enthalpy change for actinide extraction by TBPl6 is small and exothermic for Anv1 but, with the exception of ThIV, endothermic for A d v as shown in Table 12. In the case of Uv' extraction from HN03 by TBP/dodecane the net enthalpy change was found4" to be 20.9* 0.21 kJ mol-'. The value of D h l v was found to increase with temperature, reaching a maximum at 60 O C ? l 8 Measurements of the enthalpy of extraction of P u I V at low acidities are hampered by hydrolysis reactions but, generally, the enthalpies for actinide extraction by TBP appear to be rather insensitive to aqueous phase HN03 concentrations. The rates of transfer of Purv and Uv' between phases were found to be rapid213*348,4'9 and, in centrifugal contacters, mass transfer may be complete in 5-10 seconds.420The reductive back-extraction of PuIII using Fe" is also rapid in well-mixed two-phase systems, although the net mass transfer rate may be substantially reduced in the absence of effective stabilization to prevent reoxidation and reextraction of the plutonium.421 Actinide retention in the organic phase during back extraction operations may result from the presence of acidic products from radiolytic solvent degradation. The distribution coefficients of H2MBP and HDBP between nitric acid and TBP/OK are such that the bulk of the HDBP will be present in the organic phase while the H2MBP will be largely in the aqueous phase." The M while solubility of Pu(N0J2.(DBP), in 10% TBP/hydrocarbon solutions is4,, ca. U02(DBP), is quite soluble." In fact, U02(DBP)2itself is containing bridging DBP

0

Uranium

0

Oaygsn

Phosphorus

Figure 45 The structure of the UO,(DBP), polymer chain?23 (Adapted from H. Bums, Inorg, Chem., 1983, 22, 1174, with permission from the American Chemical Society)

ligands and octahedrally coordinated uranium as shown in Figure 45. However, in the presence of tri-n-butylphosphine oxide (TBPO) a dimer of formula [UO,(NO,)(DBP)(TBPO)], is formed,423as shown in Figure 46. This may represent the species which is soluble in TBP but having TBPO in place of TBP. The suppression of polymer formation by adduct formation of this type may also account for the synergistic effect of certain neutral ligands on UO:+ extraction by dialkyl phosphates referred to in Section 65.2.2.1(iv). High organic phase uranium loadings can thus remove DBP and suppress Zr(OH)(NO,)( DBP)2 precipitation. However, the formation of ZrtVprecipitates is, in any case, slow and retarded by high HN03 concentrations. Phosphoric acid itself forms highly insoluble precipitates with Zr" but only modestly suppresses P u I V e x t r a ~ t i o n . ~The ~ ' solubility roducts of UO,(HPO,) and (U02)3(P04)2have been evaluated as 1OL-I2 and lop5*respe~tively:'~ Any hydroxamic acids which might result from radiolytic degradation of the diluent in the presence" of HN03 would be effective complexing agents for U"' or Zr1V?25However, they would also be readily destroyed by any nitrous acid present and have not

Applications in the Nuclear Fuel Cycle and Radiopharmacy

BU ~

B

u

953

B Bu

I

Bu

0

Uranium

Nitrogen

Phosphorus

Figure 46 The structure of [U0,(DBP)(TBPO)(N03)]2?23 (Adapted from H. Burns, Inorg. Chem., 1983, 22, 1174, with

permission from the American Chemical Society)

in fact been detected in Purex process Scrubbing the recycled solvent with solutions of sodium carbonate is effective in removing HDBP and its complexes so that DFRuand DFzr values of ca. 100 may be obtained after a 5 minute contact.42s Hydrazine carbonate has been proposed as a 'salt free' reductive scrub for TBP.42sAllowing the solvent to age prior to scrubbing is highly detrimental to the effectiveness of carbonate treatments and greatly increases the contact times needed.426 Many different h r e x flowsheets have been proposed and investigated. Those which are used in commercial plants will be tailored to suit the particular process requirements of the plant?'427 However, some general features relating to the choice of flowsheet conditions are noteworthy." A high uranium loading in the first cycle is desirable to optimize FP decontamination. In practice, 80% saturation of 30% TBP/OK by uranium is the maximum which can be tolerated before the density difference between the phases becomes unacceptably small." High acidity, 2-4 M HN03, suppresses the formation of interfacial precipitates of Zriv-alkyl phosphoric acid complexes and suppresses ruthenium extraction, while lower acidities, 1.5 M HNO,, offer better zirconium decontamination. Consequently a dual scrub with 4 M HNO, followed by 1.5 M HNO, has been proposed for the first cycle.2s6Dual temperature scrubs have also been recommended, with ZrIV being removed at 30 "C after which a 70 "C scrub more effectively removes ruthenium. The decontamination factors achieved will also depend on the type of equipment used. Mixer settlers values than pulsed give longer contact times, greater solvent radiolysis, and thus lower DFZrINb columns, which lead to an order of magnitude less HBBP production. Centrifugal contactors offer a further improvement but at the expense of greater mechanical complexity in the highly active section of the plant." Some typical DF values attained by the Purex process are given in Table 13. After the U/Pu partition stage the uranium loaded organic phase may be back extracted with water or dilute nitric acid to give an aqueous uranium stream which passes to further purification

Table 13 Typical Decontamination Factors Attained by the Purex Process285

P h ionium product DF for

One cycle Two cycles

Total y

Zr-thJb

RU

Total y

Uranium product z r + Nb

RU

2x1d

3 x lo3 6 X IO6

io3 io5

5x103 4 x 10'

3 x lo3 5 x lo6

2 x io3 3 x 10'

2x lo6

954

Applications in the Nuclear Fuel Cycle and Radiopharmacy

v

cycles of solvent extraction. A reductive scrub of the or anic uranium stream ma then be used to remove residual plutonium while scrubbing with > 10- M HN02 will remove Np .The aqueous plutonium stream from the partition cycle may be reoxidized using nitrogen oxides prior to passing to additional purification cycles. In the Pu(N03),-HN03-H20-TBP-OK system a third phase containing 185 g dmP3 Pu'" can form when PuiVloadings exceed 40 g dmP3in 30% TBP/OK?28 This limits the acceptable organic phase actinide content and gives lower FP decontamination factors than are achieved in the first cycle with Uw present. Back extraction of h I V with dilute nitric acid gives rise to a rather dilute aqueous product stream and electroreduction or back extraction with HAN429 may be used to obtain product streams containing ca. 60gdrnp3 plutonium.'" Alternatively, the complexation of Pu by aqueous may be used. The solution produced will be suitable for oxalate precipitation processes but will give rise to a sulfate bearing waste stream. Because of the security issues associated with plutonium production, Purex flowsheets are now also subject to design parameters which relate to the control of fissile material inventories to avoid diversion into weapons Operational experience of the application of the Purex process to oxide fuel reprocessing has been reviewed by Hine,432and technologies for reprocessing nuclear fuels have been surveyed.433 The use of higher levels of power reactor fuel irradiation and the introduction of breeder reactors will lead to increased solvent radiation doses in the first cycle of the Purex process. Improvements in the process which limit solvent radiolysis effects are thus of particular importance for the future. One strategy, described by M ~ K i b b e n is , ~ the ~ ~ so called 'permanganate strike'. This involves the addition of externally produced MnOz to the dissolver solution, or its in situ production from the addition of KMnO, and Mn(N03)2.The MnOz acts as an adsorbent for zirconium and niobium but not for Ce3+, Am3+, UO? or Pu4+ ions. Separation af the solid Mn02 from the dissolver solution prior to the first cycle of solvent extraction thus substantially reduces the radiation dose to the solvent. The carry over of zirconium and niobium into subsequent process cycles is also substantially reduced so that shorter cooling periods may be used, even for highly irradiated fuel. The h r e x process is now so well tried and established that it is difficult to conceive of its being replaced by any alternative method of reprocessing reactor fueIs during the foreseeable future. However, improvements and alternatives continue to be proposed and investigated. McKay has considered2" the possible application of other solvent systems to LMFBR fuel reprocessing. He concluded that TBP would not be superseded as the extractant of choice in the short term, but that reagents such as dibutyl phosphonate (DBBP) might have potential for use in flowsheets using low extractant concentrations. These give high FP rejection in the first cycle at the expense of a reduced solvent capacity for carrying actinides. Extractants of higher capacity, such as DBBP, would thus be better suited for use in such flowsheets. Ether diluents, such as dibutyl ether, might also have advantages over OK since their radiolysis products would be less detrimental to process performance.

P

(iv) Breeder reactor fuels There are two breeder reactor fuel cycles. One involves the irradiation of 238U/239Pu oxide fuel with fast neutrons and is at the prototype stage of development. The other involves the irradiation of 232Th/233Uoxide fuel with thermal neutrons and is at the experimental stage. Fuel from the 238 U/239Pu cycle may be reprocessed using Purex technology adapted to accommodate the significant proportion of plutonium present in the fuel. Increased americium and neptunium levels will also arise compared with thermal reactor fuel. The 232Th/233U fuel may ais0 be reprocessed using solvent extraction with TBP in the Thorex (Thorium Recovery by Extraction) process. In this case the extraction chemistry must also take account of the presence of '33Pa arising as shown in Scheme 2. Plants in the UK, USSR and France are now reprocessing irradiated UOz/Pu02 fuels and LMFBR fuel reprocessing has been the subject of international conferences.'53434The plant at Cap la Hague, France, employs a 30% TBP solution and no U/Pu separation is undertaken, so that a mixed U / h product is obtained. Fluoride is added to the process feed to complex zirconium and suppress its e~traction?~'The Dounreay plant in the UK employs a 20% TBP/OK solution and uses sulfuric acid to effect the U/Pu separation.436TBP poorly extracts Purv or Uv' from sulfuric acid solutions, but in mixed HN03/H2S04the equilibria shown in equations (206) and (207) must be considered. The equilibrium constants for these reactions, KP and Kri, are given

Applications in the Nuclear Fuel Cycle land Radiopharmacy

955

by equations (208) and (209) where [A] = [HNO,]'/[H,SO,] and Kp>>KU. The formation constants for Pu(S04),+, Pu(NO,)(SO,)+ and Pu(SO& have been reported as j3 =6.9 x lo2, 1.08 x lo3 and 5.1 x lo4 respectively on the basis of solvent extraction experiment^.^^' Thermodynamic data for the formation of uranyl sulfate complexes are given in Table 7. Neptunium438and americium439

also form sulfate complexes but in the cases of NpVand Amrrrthese are charged and inextractable. Adjusting the value of [A] provides a means of separating PuIV and U". Because Kp>>KU and K P is related to [A]', conditions can be defined under which Pur" passes to the aqueous phase while leaving Uv' in the organic p h a ~ e . The 4 ~ ~PuIV may be re-extracted by increasing [A]. Although this means of effecting the U/Pu split contributes sulfate to the process waste streams, it avoids the need for a reducing agent. The higher plutonium content of LMFBR fuel would require the use of unacceptably large volumes of iTon(I1) sulfamate in a reductive back-extraction of plutonium. HAN is also unsuitable because of the slow kinetics of its reaction with P u ' . An outline flowsheet for Prototype Fast Reactor (PFR) fuel reprocessing at Dounreay is illustrated in Figure 47. In the first cycle a partial separation of Uv' from Purv is made. In practice, Uv' is found to prevent third-phase formation so that higher solvent loadings are possible than with Pu'" alone. Although a pure P u ( N O ~product )~ stream is ultimately produced by this flowsheet, it would be possible to produce a mixed U/Pu product stream. After blending to adjust the U/Pu ratio this might be used directly in the Sol-Gel process to produce U02/Pu02 fuel. Such an approach has the advantage, in terms of security, that at no stage in the process is a stream containing entirely fissile actinides produced. Irradiated ThO,/UO2 breeder fuels may also be reprocessed using solvent extraction with TBP in the Thorex p r o ~ e s s . ~The ' ~ , 232Th/233U ~~~ fuel cycle presents several additional chalienges to the fuel reprocessor. The 233Paprecursor of 233Uhas a substantially longer half-life than 239Np in the 238U/239Pu cycle. Thus, unless very long fuel cooling times were allowed prior to reprocessing, there would be substantial quantities of highly active 233Pain the process feed. Most of this would pass to the HAW in the first Thorex c cle, and would lead to loss of 23315. from the fuel cycle unless a separate process for 233Paor '3U recovery from the HAW were used. A more serious problem would arise from the formation of 232Uby neutron capture and its decay to 228Th, as shown in Scheme 9. The short half-lives of these isotopes, and of some "'Th daughters, make them intensely radioactive. No matter what process strategy were used they would accumulate in the 232Th/233U fuel cycle and necessitate remote handling at all stages. Also, in combination with the 233Papresent, they would give rise to high solvent radiation doses and increased TBP and diluent degradation compared with Purex experience. Chemical difficulties might then arise from the increased formation of HDBP, which can form insoluble complexes with Th'v?s5 Other chemical problems arise from the possibility of third-phase formation during TBP extraction of ThIVand from the low solubility of Tho, in nitric acid. Addition of fluoride ion to the dissolver solution allows Tho2 dissolution but contributes an additional, and potentially complexing, anion to the first cycle aqueous feed. The extraction chemistry of PaV also differs from that of Np' so that new complications arise from the presence of 233Pa.

a

2 3 3 ~

Z'U

(n)

Z32u

'"Th

,sy . ..

212Pb

74y

Scheme 9

Irradiated ThO,/UO, fuel can be dissolved in the Thorex reagent consisting of 13 M H N 0 3 , 0.05 M HF and 0.1 M AI(N03)3.This gives rise to a solution containing ThlV,Uvr and Pa". The solution chemistry of protactinium has been r e ~ i e w e d ~ ~and, ' , ~ ' although PalVis stable in aqueous media, it is readily oxidized by air so that only Pa" would arise in the dissolver solution. In nitric acid media, Pav forms complexes of the genera1 formula Pa(OH),(N03)yZ where (1,y, z) = (3,1,+ l ) , (2, 1, +2) and ( 2 , 2 ,+1) in 1-2 M HN03, (2,3,0), (2,4, -1) and (1,3, +l) in 2-6 M HN03 and (2,3,0) and (2,4, -1) in 6-12 M HN03.441Some,equilibrium for the formation of such complexes in media of ionic strength 5 M are shown in Table 14. Only the CCC6-EE'

Applications in the Nuclear Fuel Cycle and Radiopharmacy

956

L-l Aqueous

Solvent 20%TBP/OK Aqueous Raffinote FPs,NPP

-

t

First Extraction Cycle I Fission Product Deconlamination

-

Scrub HN03

=

H W @ 4

1

"

-

Orponic U + Pu

Am Cmm

Uranium/Plutonium Partial Split

Aqueous Product

,,

~

Feed Conditioning

U + Pu in HWH2S04

-

Uranium Loaded Solvent

50% of U

to back - uranium extract ion and

I

u

Aqueous Producl Pu (So4l2

solvent recycle

Pu

Solvent

,I

to - uranium back

Feed Conditioning

extraction and solvent recycle Scrub

Solvent 20%TBP/ OK

Third Extraction Cycle

-.-

H NO,

7 Rotf inate

Plutonium Product

I

Plutonium Back €%traction

-

HNO3

25 g d m 3 Pu (NO,),

Figure 47 An outline flowsheet for the reprocessing of Prototype Fast Reactor fuel at D o ~ n r e a y ~ ~ ~

Table 14 Equilibrium Constants4'" for Nitrate Complexes of Pa" Species

Equilibrium expression

log Ka 1.23 2.11 2.73

a

I n 5 M (Li+/Na+/H+CIO,-/NO,-)

medium.

Applications in the Nuclear Fuel Cycle and Radiopharmacy

957

(2,3,0) complex was found to be extracted by TBP, forming the disolvate Pa(OH),(N03),.2TBP. The value of Dr,v between 30% TBP/OK and HNO, was found to increase rapidly from ~ ]may '+ be prepared directly from TcO,- and the Tc' phosphite complex may thus be prepared from Tc04- under mild conditions using a two-step reaction sequence. The Tc' isocyanide complex, [Tc(ButNC),]+, may also be conveniently prepared from wmTc04-by reacting a standard preparation of 99mTcglucoheptonate with the Bu'NC adduct of ZnBr,. Thus even Tc' species are obtainable from TcO,- under conditions which are not precluded by the requirements of a clinical environment. Recently the Tc"' species [TcNXJ (X = C1, Br) have been prepared from the reaction of N3and TcO,- in concentrated HX solution. They were shown to have a square pyramidal structure with an apical nitride.633Substitution reactions of [99mT~NC14]preparations with MDP, DTPA and cysteine have been used to prepare new agents for clinical evaluation. Although radiochemically pure products could be obtained, these radiopharmaceuticals did not prove to be of clinical use. However, their behaviour was different from that of analogous preparations containing Tc03+ instead of TcN'+ and funher studies with this system would seem to be warranted. This development demonstrates that, in principle, all seven technetium oxidation states, from Tc' to Tcvl', are available for medical use. Although the clinical utility ofthe oxidation state (11) and (VI) complexes has yet to be established, it may be presumed that, when suitable complexing agents are identified, these two will find a role and further extend the range of clinical applications of 9 9 m T ~ . Radiopharmaceutical formulations based on 99mTcmay be divided into two major classes.635 The Class I formulations include what may be described as technetium tagged materials. These may be particles such as cells or colloidal substances, or they may be molecular species such as proteins or other large biological molecules. The essential feature of these agents is that their biodistribution behaviour is determined by the nature of the substrate material. Thus the attachment of a 99mTclabel should have little or no effect on their biodistribution. However, it cannot necessarily be assumed that this will be the case, especially when relatively low molecular weight substrates are involved. The Class I1 formulations are the so called 'technetium essential' radiopharmaceuticals whose biodistribution depends not only on the nature of the substrate to be labelled, but also on the presence of 9 Y m Titself. ~ Some examples of the major types of w Y " Tradiopharmaceutical ~ agents in use, or under investigation, are summarized in Table 22. The nature of the 9 9 m Tbinding ~ site

Applications in the Nuclear Fuel Cycle and Radiopharmacy

984

Table 22 Examples of ssmTq Radiopharmaceuticals

Class"

Application

Principle

Blood pool imaging

I

Distribution of' labelled blood cells

Lymph system imaging

I

Stomach imaging

1

Liposome uptake in lymph nodes affected by tumor Suspension of particles in stomach

Liver and spleen structure studies Spleen imaging

I

Lung scanning

I

Thyroid scanning Bone scanning

I1 I1

Liver and bile duct function studies Kidney function studies

I1

Kidney structure studies

I1

Trapping in reticuloendothelial system Damaged blood cell uptake by spleen Trapping of microspheres in capilliaries Reversible anion uptake Binding of phosphonate to bone surface or Tc exchange with hydroxyapatite Transport through hepatocytes to bile duct Glomerular filtration and distribution in extracellular fluid Binding in kidney cortex

Myocardial imaging under development

I1

Cation uptake in myocardium

Brain scanning

I1

Brain scanning experimental studies

I1

Accumulation at tumor disrupted sites Diffusion through blood/brain barrier

a

I1

I1

Formuhtionb

Re$

Citrate + cells Oxine + cells Sn2+ only + cells Tc0,-/Sn2+ incorporated in phospholipid vesicle Reaction of Tc0,'- with polymer beads treated with triethylenetetramine* Tc0,- adsorbed on ion exchange resin beads* Colloidal Tc, S7 preparation from Tc04-+H2S Denatured 9 9 m Tlabelled ~ red blood cells Albumin microspheres

636 637 63 8 639

Tc0,-* MDP or hydroxyethane-1,idiphosphonate (HEDP)

551 20, 553, 644, 645

Carbamoyliminodiacetate derivatives (21) DTPA

20, 551, 553

Dimercaptosuccinic acid or glucoheptonate Chelating ditertiary phosphine or arsine (22) Alkyl isonitrite Tc0,-* Glucoheptonate Bis-N,N'-(2thioethy1)ethylenediamine derivatives (35)

20, 553, 641

640 641 551,642 505 643

553, 646,647

20, 613, 614 607 551 505 648 649

Class 1 are 'tagged materials', Class I1 are 'technetium essential' systems. Formulations normally incorporate a reducing agent such as S K I z unless marked*. Other additives may also be present: ascorbic acid, for example, is sometimes incorporated as an antioxidant.

in most Class I radiopharmaceuticals remains unknown. However, the more recently developed Class II agents, such as those derived from dmpe, Bu'NC and (35), are based on known 99Tc chemistry so that the nature of the 9 9 m Tspecies ~ involved is well established. On the basis of known 99Tcchemistry it might also be presumed that the most stable 9 9 mT ~complex V of dimercaptosuccinic acid (DMSA) would be structurally related to [TcO(edt),]-. The nature of the complexes formed between 9 9 m Tand ~ MDP or hydroxyethane-1,l-diphosphonate(HEDP) is less certain, but a 99Tc model again exists to illustrate one possible binding mode, as shown in Figure 52. Although several 99Tccom lexes of amine-carboxylate ligands have now been structurally characterized, the nature of the '"Tc complexes formed with DTPA or the hepatobiliary agents (21) has not been conclusively determined. Structure (36) has been for the complex of

R