Advances in Chromatography: Volume 45 (Advances in Chromatography)

Citation preview

DK4478_C000.fm Page i Friday, September 22, 2006 8:48 AM

Half Title Page VOLUME 45

Advances in CHROMATOGRAPHY

DK4478_C000.fm Page ii Friday, September 22, 2006 8:48 AM

DK4478_C000.fm Page iii Friday, September 22, 2006 8:48 AM

Title Page VOLUME 45

Advances in CHROMATOGRAPHY EDITORS:

ELI GRUSHK A Hebrew University of Jerusalem Jerusalem, Israel

NELU GRINBERG Boehringer-Ingelheim Pharmaceutical, Inc. Ridgefield, Connecticut, U.S.A.

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

DK4478_C000.fm Page iv Friday, September 22, 2006 8:48 AM

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2007 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 1-57444-735-1 (Hardcover) International Standard Book Number-13: 978-1-57444-735-4 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www. copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

DK4478_C000.fm Page v Friday, September 22, 2006 8:48 AM

Contributors Abby Brown Department of Chemistry and Biochemistry California State University Los Angeles, California Violet Calderon Department of Chemistry and Biochemistry California State University Los Angeles, California Dinora B. Chinchilla Department of Chemistry and Biochemistry California State University Los Angeles, California Frank A. Gomez Department of Chemistry and Biochemistry California State University Los Angeles, California Amber M. Hupp Department of Chemistry Michigan State University East Lansing, Michigan Daniela Hutanu Department of Chemistry Oregon State University Corvallis, Oregon Xiaoping Li Department of Chemistry Michigan State University East Lansing, Michigan

Victoria L. McGuffin Department of Chemistry Michigan State University East Lansing, Michigan Ruth Montes Department of Chemistry and Biochemistry California State University Los Angeles, California Colin F. Poole Department of Chemistry Wayne State University Detroit, Michigan Alejandra Ramirez Department of Chemistry and Biochemistry California State University Los Angeles, California Vincent T. Remcho Department of Chemistry Oregon State University Corvallis, Oregon Sabine Schefzick Pfizer Global Research and Development Discovery Technologies Ann Arbor Laboratories Ann Arbor, Michigan Nicholas H. Snow Department of Chemistry and Biochemistry Seton Hall University South Orange, New Jersey

DK4478_C000.fm Page vi Friday, September 22, 2006 8:48 AM

Taguhi Sogomonyan Department of Chemistry and Biochemistry California State University Los Angeles, California James D. Stuart Department of Chemistry University of Connecticut Storrs, Connecticut

Jose Zavaleta Department of Chemistry and Biochemistry California State University Los Angeles, California

DK4478_bookTOC.fm Page vii Friday, September 22, 2006 8:49 AM

Contents Chapter 1

The Thermodynamic and Kinetic Basis of Liquid Chromatography...................................................................................1

Xiaoping Li, Amber M. Hupp, and Victoria L. McGuffin Chapter 2

Applications of Ionic Liquids in Extraction, Chromatography, and Electrophoresis ............................................................................89

Colin F. Poole Chapter 3

Applications of Capillary Electrophoresis to Molecular Recognition and Analysis of In-Capillary Enzyme-Mediated Transformations................................................................................125

Jose Zavaleta, Dinora B. Chinchilla, Abby Brown, Alejandra Ramirez, Violet Calderon, Taguhi Sogomonyan, Ruth Montes, and Frank A. Gomez Chapter 4

Aptamers as Molecular Recognition Elements in Chromatographic Separations ..........................................................173

Daniela Hutanu and Vincent T. Remcho Chapter 5

Advances in High-Throughput High-Performance Liquid Chromatography and “Purification-Friendly” Combinatorial Library Design Strategies ................................................................197

Sabine Schefzick Chapter 6

Biological, Clinical, and Forensic Analysis Using Comprehensive Two-Dimensional Gas Chromatography ...............215

Nicholas H. Snow Chapter 7

Determination of Endocrine Disrupting Chemicals Found in Environmental Samples by Gas Chromatography/ Mass Spectrometry...........................................................................245

James D. Stuart Index ......................................................................................................................275

DK4478_bookTOC.fm Page viii Friday, September 22, 2006 8:49 AM

DK4478_book.fm Page 1 Thursday, September 14, 2006 1:35 PM

1

The Thermodynamic and Kinetic Basis of Liquid Chromatography Xiaoping Li, Amber M. Hupp, and Victoria L. McGuffin

CONTENTS I. Introduction....................................................................................................2 II. Theory ............................................................................................................2 A. Thermodynamics ...................................................................................4 B. Kinetics..................................................................................................8 III. Methods........................................................................................................10 A. Perturbation Methods ..........................................................................10 1. Temperature Jump Method ...........................................................11 2. Pressure Jump Method ..................................................................11 3. Dipole Jump Method.....................................................................12 B. Shallow Bed Method...........................................................................12 C. Chromatographic Methods..................................................................13 1. Frontal Method..............................................................................13 2. Impulse Method.............................................................................13 IV. Applications of Thermodynamic and Kinetic Methods..............................19 A. Partition ...............................................................................................19 1. Fundamental Studies of the Retention Mechanism......................20 2. Effect of Alkyl Chain Length and Bonding Density....................22 3. Effect of Support ...........................................................................33 4. Effect of Mobile Phase Composition............................................ 34 5. Effect of Solute Structure and Concentration............................... 37 B. Adsorption ...........................................................................................45 1. Poly(styrene-divinylbenzene) ........................................................45 2. Porous Graphitic Carbon...............................................................47 3. Other Organic Adsorbents............................................................. 52 4. Silica ..............................................................................................53 C. Ion Exchange.......................................................................................57 D. Size Exclusion .....................................................................................60

1

DK4478_book.fm Page 2 Thursday, September 14, 2006 1:35 PM

2

Advances in Chromatography, Volume 45

E.

Chiral ...................................................................................................62 1. Brush CSPs....................................................................................62 2. Inclusion CSPs ..............................................................................64 3. Polysaccharide CSPs .....................................................................68 4. Affinity CSPs.................................................................................73 V. Conclusion ...................................................................................................76 References................................................................................................................77

I. INTRODUCTION In liquid chromatography, the retention mechanisms broadly consist of partition, adsorption, complexation, ion exchange, and size exclusion [1]. These mechanisms may occur individually or in various combinations. When a combination of mechanisms occurs intentionally, that is, by specific and deliberate design, it is generally beneficial to the resulting separation. However, when a combination occurs unintentionally or by accident, it is generally detrimental to the separation. A deep understanding of the retention mechanism is necessary in order to maximize the beneficial contributions and minimize the detrimental contributions to the separation. This understanding is also helpful to guide the development of commercially available stationary phases with better performance. Thermodynamic information provides an understanding of the energetic interactions between the solute and the mobile and stationary phases, which are intrinsically related to the aspects of retention and selectivity. Kinetic information provides an understanding of the rates of mass transport processes, which are intrinsically related to the efficiency or plate height. Both thermodynamic and kinetic information are necessary in order to optimize resolution [1,2]. In this chapter, the fundamental theories of thermodynamics and kinetics are briefly summarized and the experimental methods of determining thermodynamic and kinetic parameters are described. The application of these methods to the most common retention mechanisms in liquid chromatography is reviewed. Because of the great depth and breadth of information available, particularly in thermodynamics, this review is not intended to be exhaustive, but rather illustrative of the characteristic features that distinguish each retention mechanism.

II. THEORY The thermodynamics and kinetics of solute transfer between the mobile and stationary phases in liquid chromatography can be treated in several ways. Most commonly, the retention process is treated as if it were analogous to a chemical reaction. For a partition mechanism, this may be represented as a simple first-order reaction, ⎯⎯ ⎯ → Xs Xm ← ⎯

(1.1)

DK4478_book.fm Page 3 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

3

for a solute (X) distributed between the mobile (m) and stationary (s) phases. For an adsorption, complexation, or ion exchange mechanism, where the number of stationary phase sites (S) is limited, a second-order reaction is more appropriate: ⎯⎯ ⎯ → XS Xm + S ← ⎯

(1.2)

⎯⎯ ⎯ → XS + Dm , Xm + DS ← ⎯

(1.3)

or

where D is a displacing agent that is added to the mobile phase. If the concentrations of stationary phase sites and the displacing agent are sufficiently large, these mechanisms may be treated as pseudo-first-order reactions. The treatment of the retention process as a chemical reaction involves a rather simplistic view. First, chemical reactions occur in homogeneous solution and the reactants have relatively well-defined energies and diffusion distances between them. The chromatographic system is intrinsically heterogeneous at the macroscopic level, as it involves two immiscible phases. Moreover, the phases themselves are also heterogeneous at the microscopic and molecular levels. The bulk mobile phase composition may be substantially different from that near the surface of the stationary phase or in the pores. The stationary phase may have several different types of retention sites or even different retention mechanisms. Because of these heterogeneities, the energy of the solute in the mobile and stationary phases is not as well defined as in a chemical reaction. Second, given sufficient time, the chemical reaction can achieve a state of equilibrium in which the rates of the forward and reverse reactions are equal and the activities (or concentrations) of the products and reactants remain constant. The chromatographic system, by virtue of dynamic flow, can never achieve this macroscopic state of equilibrium, but may, under suitable conditions, achieve a microscopic condition of steady state [2]. Despite this simplistic view, the treatment of liquid chromatography as a chemical reaction is prevalent, as it provides considerable insight into the rank and magnitude of the various contributions to the retention mechanism. If the retention process is considered as a chemical reaction, then an energy coordinate diagram (Figure 1.1) can be used to explain the analogous thermodynamic and kinetic contributions. In this diagram, the solute transfers from the mobile phase (Xm) to the stationary phase (Xs) through a high-energy transition state (X‡). The thermodynamic parameters, such as the change in molar enthalpy (ΔH) and molar volume (ΔV), are characterized by the difference between the final and initial states. These thermodynamic quantities represent the weighted average of all of the available states in the heterogeneous stationary and mobile phases, respectively. The kinetic aspects of the retention event can be described by using transition-state theory. The transfer from the mobile to stationary phase is characterized by a fast equilibrium between the mobile phase and transition state with preequilibrium

DK4478_book.fm Page 4 Thursday, September 14, 2006 1:35 PM

4

Advances in Chromatography, Volume 45

X‡

Energy

Km‡

ΔVm‡ ΔHm‡

kms

Xm ΔV ΔH Xs Reaction coordinate

FIGURE 1.1 Energy coordinate diagram depicting the transfer of solute X from the mobile phase (m) to the stationary phase (s). Associated thermodynamic and kinetic parameters are described in the text.

constant Km‡, followed by a rate-limiting step between the transition state and the stationary phase with rate constant kms. The kinetic parameters, such as the activation enthalpy (ΔHm‡) and activation volume (ΔVm‡), can be used to elucidate the kinetic description of the retention event. For the transfer from the stationary to mobile phase, the preequilibrium constant Ks‡ and rate constant ksm are used to determine the corresponding activation enthalpy (ΔHs‡) and activation volume (ΔVs‡). These kinetic quantities represent the weighted average of all available paths between the final and initial states. The derivation of these thermodynamic and kinetic quantities is described in the following sections.

A. THERMODYNAMICS As noted above, equilibrium is not attained due to the dynamic nature of the chromatographic system. However, at the microscopic level, steady-state conditions may be achieved in which the activity (or concentration) of the solute in the mobile and stationary phases (am and as, respectively) achieves a constant value: K=

as = kβ , am

(1.4)

where K is the equilibrium constant, k is the retention factor, and β is the phase ratio. The phase ratio is defined as the volume of the stationary phase divided by the volume of the stationary phase. The ratio of the equilibrium constants or retention factors for two solutes defines the selectivity (α): α=

K 2 k2 , = K1 k1

(1.5)

DK4478_book.fm Page 5 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

5

where the subscripts 1 and 2 denote the less and more retained solutes, respectively. This ratio may reflect the differences between members of a homologous series (methylene group selectivity), between solutes with the same aliphatic or aromatic structure but varying in functional group (functional group selectivity), or between chiral solutes (enantiomeric selectivity). The equilibrium constant is related to the change in molar Gibbs free energy (ΔG) by ln K =

−ΔG , RT

(1.6)

where R is the gas constant and T is the temperature. The selectivity for two solutes is correspondingly related as ln α = ln

K 2 − ΔΔG , = K1 RT

(1.7)

where ΔΔG represents the difference in molar Gibbs free energy for methylene selectivity, functional group selectivity, or enantiomeric selectivity. The Gibbs free energy is related to the changes in molar enthalpy (ΔH) and molar entropy (ΔS) by the Gibbs-Helmholtz equation [3]: ΔG = ΔH − T ΔS .

(1.8)

By substitution, the retention factor can be related to the molar enthalpy and molar entropy: ln k = ln K − ln β =

− ΔH ΔS + − ln β . RT R

(1.9)

The change in molar enthalpy can be determined by graphing the natural logarithm of the retention factor versus inverse temperature at constant pressure. This will yield a slope that is related to the change in molar enthalpy and an intercept that contains information about the change in molar entropy and the phase ratio. When the phase ratio is known, the molar entropy can be determined from the intercept. However, the phase ratio is often unknown and accurate determination of the molar entropy is not possible. The phase ratio and the molar entropy must be independent of temperature for accurate determination of the molar enthalpy. The change in molar enthalpy will be positive for an endothermic transfer and negative for an exothermic transfer of the solute from the mobile to stationary phase. Although there are some noteworthy exceptions (vide infra), most successful separations in liquid chromatography involve an energetically favorable exothermic process. The change in molar entropy will be positive when there is a favorable dissipation of energy through an increase in the number or distribution of microstates. This most commonly occurs

DK4478_book.fm Page 6 Thursday, September 14, 2006 1:35 PM

6

Advances in Chromatography, Volume 45

through a change in volume or concentration, leading to a change in configurational entropy as the solute is transferred from the mobile to the stationary phase. The selectivity for two solutes is correspondingly related as ln α =

− ΔΔH ΔΔS + . RT R

(1.10)

Thus, a graph of the natural logarithm of the selectivity vs. inverse temperature will yield from the slope and intercept the differential change in molar enthalpy (ΔΔH) and molar entropy (ΔΔS), respectively, associated with methylene selectivity, functional group selectivity, or enantiomeric selectivity. From the definition of molar enthalpy [3], ΔH = ΔE + P ΔV ,

(1.11)

where ΔE is the change in molar internal energy, V is the change in molar volume, and P is the pressure. Substitution of Equation 1.11 into Equation 1.9 yields ln k =

− ΔE P ΔV ΔS − + − ln β . RT RT R

(1.12)

The change in molar volume can be determined by graphing the natural logarithm of the retention factor vs. pressure at constant temperature. This will yield a slope that is related to the change in molar volume and an intercept that contains information about the molar internal energy, molar entropy, and phase ratio. These parameters must be independent of pressure for accurate determination of the molar volume. The change in molar volume will be positive when the solute occupies a larger volume and negative when the solute occupies a smaller volume in the stationary phase than in the mobile phase. The selectivity for two solutes is correspondingly related as ln α =

− ΔΔE P ΔΔV ΔΔS − + . RT RT R

(1.13)

Thus, a graph of the natural logarithm of the selectivity vs. pressure will yield from the slope the differential change in molar volume (ΔΔV) associated with the methylene selectivity, functional group selectivity, or enantiomeric selectivity. To gain a greater understanding of the balance of thermodynamic contributions to solute retention, enthalpy-entropy compensation is very useful. For a pair or series of solutes that obeys a linear free energy relationship [4], there exists a hypothetical temperature at which the relative changes in enthalpy and entropy are balanced and the net change in free energy is zero. By rearrangement of Equation 1.8, this compensation temperature (Tc) can be expressed as

DK4478_book.fm Page 7 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

Tc =

ΔΔH . ΔΔS

7

(1.14)

At this temperature, the pair or series of solutes would coelute and no separation would be achieved (α = 1.0). Hence, it is also called the isoselective or isoenantioselective temperature (Tiso) in chiral separations. A compensation temperature that is substantially greater than the ambient temperature suggests that the retention mechanism is enthalpy dominated, whereas a compensation temperature that is less than the ambient temperature is entropy dominated. The compensation temperature can be determined from the slope of a graph of the change in molar enthalpy versus the change in molar entropy, both derived from Equation 1.9. However, for statistical reasons, a linear relationship may be observed between ΔH and ΔS, even in the absence of enthalpy-entropy compensation [5]. Krug et al. [6] investigated different methods to identify true enthalpy-entropy compensation that is not influenced by statistical artifacts. By means of Equation 1.8, the free energy at a specific temperature (T) can be related to the free energy at the compensation temperature (Tc) by T ⎞ T ΔGTc ⎛ . ΔG = ΔH ⎜ 1 − ⎟ + ⎝ Tc ⎠ Tc

(1.15)

The retention factor can then be related to the free energy at the compensation temperature by ln k = −

ΔH ⎛ 1 1 ⎞ ΔGTC − + + ln β . R ⎜⎝ T Tc ⎟⎠ RTc

(1.16)

Thus, a graph of the natural logarithm of the retention factor vs. the change in molar enthalpy can be used to evaluate enthalpy-entropy compensation. If compensation occurs, this graph will be linear and the slope can be used to calculate the compensation temperature. The compensation temperature can be used to compare the retention mechanisms for different solutes, different mobile phases, or different stationary phases. As discussed by Ranatunga et al. [7], if the compensation temperatures are identical, then the relative contributions of enthalpy and entropy to the overall free energy are the same for the two systems. However, if compensation temperatures are different, then the underlying retention mechanism must be distinctly different. This concept of compensation is readily extensible [8]. In the same way that the changes in molar enthalpy and entropy are related through free energy in Equation 1.8, the changes in molar enthalpy and volume are related through the internal energy in Equation 1.11. Hence, for a pair or series of solutes that obey a linear internal energy relationship, there exists a hypothetical pressure at which the relative changes in enthalpy and volume are balanced and the net change in internal

DK4478_book.fm Page 8 Thursday, September 14, 2006 1:35 PM

8

Advances in Chromatography, Volume 45

energy is zero. By rearrangement of Equation 1.11, this compensation pressure (Pc) can be expressed as ΔΔH . ΔΔV

Pc =

(1.17)

A compensation pressure that is substantially greater than the operating pressure suggests that the retention mechanism is enthalpy dominated, whereas a compensation pressure that is less than the operating pressure is volume dominated. The compensation pressure can be determined from the slope of a graph of the change in molar enthalpy versus the change in molar volume, derived from Equations 1.9 and 1.12, respectively. Because these thermodynamic quantities are not derived by linear regression from the same experimental data, there are no difficulties with statistical artifacts. The compensation pressure can be used, in combination with the compensation temperature, to compare the retention mechanisms for different solutes, different mobile phases, or different stationary phases.

B. KINETICS Whereas thermodynamic information is vital, kinetic information provides a deeper understanding of the retention mechanism. There are two common approaches to describe kinetic behavior. The first approach, consistent with the thermodynamic treatment, is to consider the retention process as a chemical reaction. The general principles of chemical kinetics are then applicable [3]. The second approach, commonly used in chemical engineering, is to characterize the rates at which the individual mass transfer processes occur. Each of these approaches provides a valuable and complementary description of the retention process and will be described herein. If the retention process is treated as a first-order or pseudo-first-order reaction according to Equation 1.1, kms

⎯⎯⎯ ⎯⎯ →X , Xm ← k ⎯ s

(1.18)

sm

then kms and ksm are the corresponding rate constants for solute transfer from the mobile to stationary phase and from the stationary to mobile phase, respectively. These designations for the rate constants will be used throughout this chapter, regardless of the retention mechanism (partition, adsorption, ion exchange, size exclusion, etc.). These are “lumped” rate constants, that is, they comprise all contributions to the kinetic behavior, including the sorption/desorption event, diffusion in the mobile and stationary phases, diffusion in pores and stagnant layers, interfacial resistance to mass transfer, etc. The rate constants, which reflect the kinetic behavior, are related to the retention factor, which reflects the thermodynamic behavior, by k=

kms . ksm

(1.19)

DK4478_book.fm Page 9 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

9

By analogy to the thermodynamic treatment, the ratio of the rate constants for two solutes may be used to define the kinetic selectivity (αms and αsm). For example,

α ms =

(k ) (k )

ms 2

,

(1.20)

ms 1

where the subscripts 1 and 2 denote the less and more retained solutes, respectively. This ratio may reflect the kinetic differences associated with methylene selectivity, functional group selectivity, or enantiomeric selectivity. The detailed kinetic parameters (Figure 1.1) can be elucidated by applying a combination of thermodynamic and transition-state theories. The rate constant kms is expressed by the Arrhenius equation [3] as ⎛ − ΔEm ‡ ⎞ , kms = Am ‡ exp ⎜ ⎝ RT ⎟⎠

(1.21)

where Am‡ is the preexponential factor and ΔEm‡ is the activation energy for the transfer from the mobile to stationary phase. By rearrangement,

ln kms = ln Am‡ −

ΔEm‡ . RT

(1.22)

The activation energy can be determined by graphing the natural logarithm of the rate constant versus inverse temperature at constant pressure. This will yield a slope that is related to the activation energy and an intercept that contains information about the preexponential factor. The activation energy represents the energetic barrier that must be overcome for solute transfer from the mobile to stationary phase. The kinetic selectivity is correspondingly related as

ln α ms = ln

(k ) (k )

ms 2 ms 1

=

− ΔΔEm‡ , RT

(1.23)

where ΔΔEm‡ represents the difference in activation energy for methylene selectivity, functional group selectivity, or enantiomeric selectivity. By using classical thermodynamic relationships, the activation energy is given by ΔEm ‡ = ΔH m ‡ + RT − P ΔVm ‡ . By substitution into Equation 1.22,

(1.24)

DK4478_book.fm Page 10 Thursday, September 14, 2006 1:35 PM

10

Advances in Chromatography, Volume 45

ln kms = ln Am ‡ −

ΔH m ‡ + RT − P ΔVm ‡ . RT

(1.25)

The activation volume can be determined by graphing the natural logarithm of the rate constant vs. pressure at constant temperature. This will yield a slope that is related to the activation volume and an intercept that contains information about the activation enthalpy and preexponential factor. The activation enthalpy and preexponential factor must be independent of pressure for determination of the activation volume. The activation volume represents the volumetric barrier that must be overcome for solute transfer from the mobile to stationary phase. The activation enthalpy can then be calculated from the activation energy and activation volume via Equation 1.24. In a similar manner, the rate constant ksm can be used to determine the activation energy (ΔEs‡), activation volume (ΔVs‡), and activation enthalpy (ΔHs‡) for solute transfer from the stationary to mobile phase. These kinetic quantities provide a detailed description of the energetic and volumetric requirements of the retention process that complement the thermodynamic description. Another approach, which is more common in the engineering literature, is to evaluate the individual mass transfer processes that contribute to band broadening. Miyabe and Guiochon [9] identified four primary mass transfer processes: axial dispersion, external mass transfer, intraparticle diffusion, and sorption/desorption kinetics. Axial dispersion is broadening that occurs in the bulk mobile phase, including axial diffusion and multiple paths through the packed bed. External mass transfer occurs as the solute moves from the bulk mobile phase through the thin stagnant film at or near the particle surface. Intraparticle diffusion occurs as the solute migrates into the pore of the particle. This process consists not only of solute diffusion through the mobile phase in the pore, but also surface diffusion on the wall of the particle. Finally, as the solute sorbs to and from the surface, slow mass transfer arises from the sorption/desorption event. These mass transfer processes are considered as individual and separable contributions to the overall broadening and rate constant. The mathematical treatment of this approach will be described in Section III.C.2.a.

III. METHODS A. PERTURBATION METHODS Among the available methods for evaluation of thermodynamic and kinetic parameters, perturbation methods play an important role [10,11]. Perturbation methods represent the simplest approach, wherein the solute, mobile phase, and stationary phase of interest are contained in a small, static cell. A fast perturbation is applied to the equilibrium in order to alter the activity or concentration of the solute in the mobile and stationary phases (Equation 1.4). The perturbation may be achieved by rapidly changing conditions such as temperature, pressure, or dipole moment. After perturbation, the rate of relaxation of the system to the new conditions is monitored, followed by mathematical extraction of the equilibrium and rate constants. The

DK4478_book.fm Page 11 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

11

advantage of the perturbation methods is their simplicity, where thermodynamic and kinetic behavior can be measured without interference from flow contributions. The measured rate constants consist of the sorption/desorption process and diffusional contributions to mass transfer. One of the limitations of these methods is that they require perturbation, during which the actual thermodynamic and kinetic behavior of the system is altered. Hence, the determined values of equilibrium and rate constants are neither those of the initial state nor those of the final state after perturbation. In order to minimize the required perturbation, the equilibrium constant should be as close to unity as possible. This ensures the greatest change in activity or concentration for a given perturbation of the system. In addition, the perturbation must be applied uniformly to the entire system and with sufficient speed that it does not contribute to the observed kinetic behavior. Although these methods are very promising, their application to chromatographic systems has been relatively limited. 1. Temperature Jump Method The temperature jump method is the most versatile and useful of the perturbation methods. This method is based on the concept that chemical equilibria, including retention processes, are usually associated with a nonzero enthalpy change and thus are temperature dependent. The shift in the equilibrium constant depends on the change in molar enthalpy according to Equation 1.9: ⎛ ∂ ln K ⎞ ΔH ⎜⎝ ∂T ⎟⎠ = RT 2 . P

(1.26)

The temperature pulse can be induced by Joule heating [12–15], dielectric heating [16], or optical heating [17]. The instrument is relatively simple to construct and the speed of heating is sufficient for a wide range of inorganic, organic, and biochemical equilibrium reactions. 2. Pressure Jump Method The pressure jump method is based on the concept that chemical equilibria, including retention processes, display a more or less marked dependence on pressure. The shift in the equilibrium constant depends on the change in molar volume according to Equation 1.12: ⎛ ∂ ln K ⎞ − ΔV ⎜⎝ ∂P ⎟⎠ = RT T

(1.27)

The pressure pulse can be produced either by a sudden application or release of pressure in the system. The pressure jump method does not have as wide a range of applications as the temperature jump method because of the limited number of reactions that exhibit a sufficiently large pressure dependence. However, this method

DK4478_book.fm Page 12 Thursday, September 14, 2006 1:35 PM

12

Advances in Chromatography, Volume 45

has been beneficial for the study of sorption/desorption of charged species, due to their relatively large changes in molar volume [18]. 3. Dipole Jump Method The dipole jump method is based on the change in dipole moment induced by photoexcitation. Many chromophores exhibit different dipole moments in the ground and excited states as a consequence of electronic redistribution after photoexcitation. This change in the dipole moment can influence the relative affinity of the solute for the mobile and stationary phases [19]. The net change in solubility depends upon the magnitude of the change in the dipole moment as well as the absorption cross section and fraction of molecules in the excited state. Because high photon flux may be required to ensure large absorption cross sections, due care must be taken to avoid thermal heating. In addition, excited-state reactions of the solute must be avoided. In the absence of these effects, the shift in the equilibrium constant depends on the change in solute activity (or concentration) according to Equation 1.4.

B. SHALLOW BED METHOD The shallow bed method can be considered as an intermediate between perturbation methods and chromatographic methods. In this method, the stationary phase particles are packed in an extremely short (“zero length”) column. The experiments may be performed in two ways, by sorption (uptake) or desorption (release) methods. In the sorption method, a mobile phase solution containing the solute flows through the shallow bed at a high linear velocity. The concentration of the solute in the effluent solution is nearly identical to that in the influent solution. If the linear velocity is sufficiently high, the stagnant diffusion layer surrounding the particle will be thin. Hence, the time required for diffusion through this layer will be small compared to the time required for the slow intraparticle processes. As a result, the intraparticle processes will determine the observed sorption rate. After a certain time, the flow is stopped and the amount of solute that has been sorbed is measured by eluting it out of the shallow bed. This measurement provides a single point on the sorption curve. This experiment is then repeated for varied sorption times to construct the complete sorption rate curve. An alternative approach that is less time consuming and labor intensive is the desorption method. In this method, the shallow bed is preequilibrated by flow of a mobile phase solution containing the solute. To initiate the experiment, the flow is abruptly shifted to a solute-free mobile phase, which causes the solute to be desorbed from the bed. By detecting the eluted solute concentration as a function of time, the desorption curve can be constructed from a single experiment. Specifically, the concentration, C(t), is a function of the instantaneous molecular desorption rate of the solute, dni(t)/dt, and the flow rate, F, as follows: C (t ) =

( dn (t ) / dt ) . i

F

(1.28)

DK4478_book.fm Page 13 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

13

After integration, the desorption rate curve is constructed as t

∫

ni ( t ) = F C ( t ) dt .

(1.29)

0

The desorption rate curve is then fit by nonlinear regression to different theoretical models, such as the linear driving force model or spherical diffusion model, to extract the rate constants [20–23].

C. CHROMATOGRAPHIC METHODS Chromatographic methods are the most widely used methods for determining the thermodynamic and kinetic behavior of a system. In contrast to perturbation methods, chromatographic methods are dynamic, with the mobile phase flowing through the column at all times. In this way, the contribution of flow phenomena to the zone profile cannot be neglected. The two most common chromatographic methods are the frontal and impulse methods. 1. Frontal Method In frontal analysis, the column is preequilibrated with the mobile phase. Then a solution containing a known concentration of solute in the mobile phase is introduced continuously. As the solute sorbs onto the stationary phase, the column becomes saturated and the concentration of solute eluting from the column gradually increases, forming a characteristic breakthrough curve. The mean position of the breakthrough curve can be related to the concentration and equilibrium constant of the solute as well as the sorption capacity of the column. The amount of solute sorbed is calculated from the breakthrough curve in order to derive a single point on the equilibrium isotherm. This process is repeated for progressively increasing concentrations of the solute to construct the complete isotherm. To obtain kinetic information, theoretical breakthrough curves are numerically calculated for each concentration step. These theoretical curves use different values of the rate constant in the transport model, together with the equilibrium isotherm determined from the breakthrough data. The theoretical breakthrough curves are then compared to the experimental ones. The rate coefficient is determined in such a way that the best agreement is observed between the theoretical and experimental breakthrough curves. In this approach, it is assumed that the rate constant remains invariant during each concentration step in the frontal analysis measurement and that this rate constant corresponds to the average concentration of each step [9]. 2. Impulse Method The impulse method is generally performed by injecting a small volume of the solute onto the column. The elution of the solute zone is monitored by on-column or

DK4478_book.fm Page 14 Thursday, September 14, 2006 1:35 PM

14

Advances in Chromatography, Volume 45

postcolumn detection. The retention factor and rate constants can be extracted by any of the following models. a. Plate Height Model In this model, the thermodynamic and kinetic parameters are determined from the mean and the broadening of the zone profile, respectively. The broadening or plate height may be determined in a variety of ways, but the most accurate is the statistical moments, as they make no assumptions about the shape of the zone profile. The first and second statistical moments are calculated from the zone profile as

M1 =

∫ C(t ) t dt ∫ C(t ) dt

(1.30)

and

∫ C(t )(t − M ) dt . ∫ C(t ) dt 2

M2 =

1

(1.31)

The first moment (M1) represents the mean retention time. It can be directly related to the retention factor by

k=

M1 − t0 , t0

(1.32)

where t0 is the elution time for a nonretained compound. The method of calculating kinetic rate constants is derived by extrapolation of Giddings’ work [2]. The plate height in the length domain (HL) can be related to the second moment (M2) in the time domain by 2

HL =

σ 2L M 2 ⎛ u0 ⎞ , = L L ⎜⎝ 1 + k ⎟⎠

(1.33)

where u0 is the mobile phase linear velocity. The mass transfer term (Cs) for slow kinetics in a system that exhibits a partition or adsorption mechanism is given by

Cs =

2k

(1 + k )

2

. ksm

(1.34)

DK4478_book.fm Page 15 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

15

Thus, ksm =

2 k u0 (1 + k )2 dH L

kms =

and

2 k 2 u0 , (1 + k )2 dH L

(1.35)

where dHL is the contribution to plate height arising from slow mass transfer. As discussed in Section II.B, Miyabe and Guiochon [9] evaluated the plate height contributions to various mass transfer processes. In their approach, the statistical moments can be further elucidated as M1 =

L δ0 u0

(1.36)

and M2 =

(

)

2L δ ax + δ f + δ d , u0

(1.37)

where δ0 is a dimensionless retention parameter, equal to 1 + k, and is given by δ 0 = ε + (1 − ε)(ε p + ρ p K ) .

(1.38)

The contributions to M2 from axial dispersion (δax), external mass transfer (δf), and intraparticle diffusion (δd) are given by ⎛D ⎞ δ ax = ⎜ 2L ⎟ δ 20 , ⎝ u0 ⎠ ⎛ Rp ⎞ δ f = (1 − ε) ⎜ ⎟ ε p + ρp K ⎝ 3k f ⎠

(

(1.39)

)

2

,

(1.40)

and ⎛ R p2 ⎞ δd = 1 − ε ⎜ ⎟ ε p + ρp K ⎝ 15 De ⎠

(

)

(

)

2

,

(1.41)

respectively, where ε is the interparticle void fraction, εp and ρp are the porosity and density of the packing material, Rp is the particle radius, DL is the axial dispersion

DK4478_book.fm Page 16 Thursday, September 14, 2006 1:35 PM

16

Advances in Chromatography, Volume 45

coefficient, kf is the external mass transfer coefficient, De is the intraparticle diffusivity, and K is the equilibrium constant. Under conditions where the rate of sorption/desorption is negligibly small, the plate height can be written as ⎛ M ⎞ ⎛ L ⎞ DL δ f δ d . H = ⎜ 22 ⎟ ⎜ = + + ⎝ M1 ⎠ ⎝ 2 u0 ⎟⎠ u02 δ 20 δ 20

(1.42)

The different parameters that are related to mass transfer processes can be estimated or calculated based on the above equations. For example, the external mass transfer coefficient, diffusion coefficient in the mobile phase, and pore diffusion coefficient can all be estimated, and the diffusion coefficient in the stationary phase can thus be calculated. More specific information on the foundations and applications of this model can be found in the excellent review by Miyabe and Guiochon [9]. The plate height model assumes that all contributions to variance, both symmetric and asymmetric, that are not directly attributable to fast processes such as axial dispersion arise from slow kinetics. There are several potential sources of error in this method. First, this method requires that the solute concentration be within the linear region of the isotherm, such that it does not contribute to the variance. Second, this method relies on the accurate calculation and subtraction of all fast mass transfer terms, which is difficult for packed columns in liquid chromatography. Empirical estimations of these parameters will introduce errors. Moreover, any extra-column contributions to variance, including those from the injector, detector, connectors, etc., must be accurately calculated and subtracted. Again, there are no theoretical methods for the calculation of these parameters, so empirical estimations are necessary. b. Exponentially Modified Gaussian Model Consider an incremental length of the chromatographic column. Zone broadening can arise from multiple paths, diffusion, and mass transfer processes that are fast relative to the time spent in the incremental length. These processes contribute to the symmetric broadening, which is described by a Gaussian function,

()

C t =

2 ⎛ ⎛ t − tg ⎞ ⎞ exp ⎜ −0.5 ⎜ ⎟ ⎟ , ⎜⎝ ⎝ σ g ⎠ ⎟⎠ 2 πσ g

A

(1.43)

where A is the area, tg is the retention time of the Gaussian component, and σg is the standard deviation of the Gaussian component. In addition, zone broadening can arise from mass transfer processes that are slow relative to the time spent in the incremental length. These processes contribute to the asymmetric broadening. For a partition or adsorption mechanism that can be considered as a first-order or pseudo-first-order reaction, this contribution is given by an exponential function,

DK4478_book.fm Page 17 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

(

⎛ − t − tg C t = A exp ⎜ ⎜⎝ τ

()

) ⎞⎟ , ⎟⎠

17

(1.44)

where τ is the standard deviation of the exponential component. The zone profile observed at the end of the total column length is the convolution of the Gaussian and exponential contributions, that is, the multiplication of the functions and integration within each incremental length. This convolution is given by the exponentially modified Gaussian (EMG) equation,

C (t ) =

⎛ σ 2g tg − t ⎞ A exp ⎜ 2 + ⎟ τ ⎠ 2τ ⎝ 2τ

⎛ ⎛ t−t σg ⎞ ⎞ g ⎜ erf ⎜ − ⎟ + 1⎟ . ⎜⎝ ⎜⎝ 2 σ g 2 τ ⎟⎠ ⎟⎠

(1.45)

The zone profile is fit to the EMG equation by nonlinear regression to extract the regression parameters (A, tg, σg, τ). From these parameters, the retention time, tr, is calculated as tr = tg + τ

(1.46)

and the corresponding retention factor is calculated as

k=

tr − t0 . t0

(1.47)

The method of calculating kinetic rate constants from the EMG model is derived by extrapolation of Giddings’ work [2,24]. The mass transfer term for slow kinetics in the partition or adsorption mechanism is given by Equation 1.34. By rearrangement, the rate constants are given by

ksm =

2 kt0 τ2

and

kms =

2 k 2 t0 . τ2

(1.48)

The EMG model assumes that all contributions to asymmetric broadening (τ) arise from slow kinetics. There are several potential sources of error in this method. First, this method requires that the solute concentration be within the linear region of the isotherm, such that it does not contribute to the asymmetry. Second, any extra-column contributions to asymmetry must be minimized or eliminated. In practice, this is achieved by detection at several points along the chromatographic column and subtraction of the parameters determined at each detector [24]. It is noteworthy that the EMG method does not require a priori estimation, whether by

DK4478_book.fm Page 18 Thursday, September 14, 2006 1:35 PM

18

Advances in Chromatography, Volume 45

theoretical or empirical means, of symmetric column or extra-column contributions to broadening. c. Giddings Model By means of stochastic theory [2], Giddings derived a model suitable for theoretical treatment of the adsorption mechanism under first-order or pseudo-first-order conditions,

C ( x ) = Aγ

k

x

(

)

(

)

I 1 2 γ kx exp − γx − γk ,

(1.49)

where γ is a dimensionless constant, equal to the product of the desorption rate constant (ksm) and the elution time of a nonretained compound (t0). After conversion from the time domain (t) to the retention factor domain (x), the zone profile is fit by nonlinear regression to Equation 1.49. From the regression parameters (A, k, γ), the corresponding retention factor and rate constants are calculated. The Giddings model assumes that all contributions to symmetric and asymmetric broadening arise from slow kinetics. There are several potential sources of error in this method. First, this method requires that the solute concentration be within the linear region of the isotherm, such that it does not contribute to the broadening. Second, this method requires that column contributions from multiple paths and diffusion in the mobile and stationary phases be negligible. Moreover, any extracolumn contributions, including those from the injector, detector, connectors, etc., must also be negligible. Unlike the plate height and EMG methods, there is no a posteriori method to correct for these contributions. d. Nonlinear Chromatography Models All of the impulse methods described above are suitable for thermodynamic and kinetic measurements within the linear region of the isotherm. However, some mechanisms such as adsorption, ion exchange, and complexation have stationary phases with a limited number of sites that may be easily overloaded. Hence, it is desirable to be able to evaluate their behavior under nonlinear conditions. A convenient method to extract thermodynamic and kinetic information from frontal profiles was reported by Thomas [25] and later modified for elution zone profiles by Wade et al. [26]. This model was derived for mechanisms that can be considered as second-order sorption and first-order desorption reactions (Equation 1.2) under linear and nonlinear conditions. This theoretical model is given by

⎛ Aγ ⎞ C ( x) = ⎜ 1 − exp − γKC 0 ⎝ KC 0 ⎟⎠

(

where

(

))

(

(

))

(

⎛ k x I 2 γ kx exp − γx − γk 1 ⎜ ⎜ − ⎜⎝ 1 T γk, γx 1 − exp − γKC 0

(

)(

(

))

) ⎞⎟

⎟ , (1.50) ⎟⎠

DK4478_book.fm Page 19 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

u

T (u, v ) = e

−v

∫e I ( −t

0

0

)

2 vt d t .

19

(1.51)

In this equation, γ is a dimensionless constant equal to the product of the desorption rate constant (ksm) and the elution time of a nonretained compound (t0), k is the retention factor, K is the equilibrium constant, and C0 is the initial concentration. After conversion from the time domain (t) to the retention factor domain (x), the zone profile is fit by nonlinear regression to Equation 1.50. From the regression parameters (A, k, γ, KC0), the corresponding retention factor and rate constants are obtained. The Thomas model assumes that all contributions to symmetric and asymmetric broadening arise from nonlinear isotherms and slow kinetics, which is a combination of mass transfer and sorption/desorption processes. There are several potential sources of error in this method. First, this method assumes the kinetics and the isotherm to be Langmuirian. Second, this method requires that column contributions from multiple paths and diffusion in the mobile and stationary phases be negligible. Moreover, any extracolumn contributions, including those from the injector, detector, connectors, etc., must also be negligible. Unlike the statistical moment and EMG methods, there is no a posteriori method to correct for these contributions. Other nonlinear chromatography models have been reviewed by Golshan-Shirazi and Guiochon [27,28]. Among those that provide both thermodynamic and kinetic information are the reaction-dispersive and transport-dispersive models. The reaction-dispersive model presumes that sorption/desorption kinetics are slow relative to the fast kinetics of mass transfer. The transport-dispersive model presumes that mass transfer kinetics are slow relative to the fast kinetics of sorption/desorption. These simplifications allow a numerical solution of the general rate model for comparison of theoretical and experimental zone profiles. Further information, including direct comparison of these models with the Thomas model, can be found in the review by Golshan-Shirazi and Guiochon [27].

IV. APPLICATIONS OF THERMODYNAMIC AND KINETIC METHODS A. PARTITION Although a wide range of stationary phases are available, the vast majority of reversed-phase separations are performed with alkyl-bonded silica phases. These phases are based on dispersion (London) forces [1], the most universal of all interactions, and hence are broadly applicable. Although many studies have added insight and attempted to definitively characterize the retention mechanism, many questions and controversies still remain. It is the purpose of this section to review the thermodynamic and kinetic studies that have provided a better understanding of the underlying processes of retention. In addition, the following parameters will be examined with regard to their contribution to the retention mechanism: properties

DK4478_book.fm Page 20 Thursday, September 14, 2006 1:35 PM

20

Advances in Chromatography, Volume 45

of the alkyl-bonded phase and support, properties of the mobile phase, solute structure and concentration, as well as temperature and pressure. 1. Fundamental Studies of the Retention Mechanism For many years, there has been a lively debate concerning the retention mechanism in reversed-phase liquid chromatography. Many models have been developed to describe the retention mechanism, some of which are directly and others indirectly based on thermodynamic concepts. Examples of models that are directly based on thermodynamics include the general solution or solubility parameter model reviewed by Tijssen et al. [29] and the solvophobic model developed by Horvath et al. [30,31]. Models that are indirectly based on thermodynamics include the lattice models developed by Martire and Boehm [32] and Dill [33]. The first important issue to be addressed is whether the mechanism of reversedphase separation with alkylsilica phases is predominantly partition-like or adsorption-like in character. A theoretical perspective was provided by Dill [33], who developed lattice models for both the partition and adsorption mechanisms. In both mechanisms, the surface consisted only of alkyl chains and the underlying silica support was not considered. Solute retention was predicted by using simple, thermodynamically related parameters: the entropy of mixing of the solute, the configurational entropy of the bonded chain, and the contact interactions of the solute with the solvent and bonded chain. This model predicted that the extent of partitioning becomes more important with increasing chain length. Solute molecules have the lowest concentration near the proximal or bound end and the highest concentration near the distal or free end of the chain. As the bonding density increases and space between the chains becomes more restricted, adsorption becomes more important. In fact, partition is predicted to decrease to zero as the chains reach their maximum density of approximately 8.1 μmol/m2. However, under most typical conditions, the predominant mechanism for long-chain alkylsilicas, particularly octadecylsilica, is likely to be partition or partition-like in nature. Experimental studies have also been used to examine the partition and adsorption mechanisms in reversed-phase liquid chromatography. In a study by Tan and Carr [34], liquid-liquid extraction was used as a model of a true partition system and compared with monomeric octylsilica and octadecylsilica phases. The alkylbenzenes were used as model solutes in order to determine the change in free energy of a methylene group (ΔΔG; Equation 1.7) in each system. The ratio of the change in free energy of a methylene group for transfer from the mobile phase to bulk hexadecane to that for transfer from the mobile phase to the stationary phase was calculated. Values of this ratio that are close to unity denote a partition mechanism, where the methylene group can be fully embedded in the bonded phase chain. For both octylsilica and octadecylsilica, the observed ratio was close to unity when the concentration of methanol in the mobile phase was less than 70%, indicating a partition-like mechanism. Tan and Carr [34] further demonstrated that shorter chains have larger ratios ranging from 2.2 to 1.5 for methylsilica to hexylsilica, indicative of a more adsorption-like mechanism. Moreover, phases with low bonding density of 0.6 μmol/m2 have significantly higher ratios of 1.8 to 2.0, while those with higher

DK4478_book.fm Page 21 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

21

bonding densities of 1.4 and 2.3 μmol/m2 have smaller and approximately equal ratios of 1.3. These results confirmed the predictions of the lattice model of Dill [33]. Later studies by Park et al. [35] examined polymeric phases ranging from methylsilica to octadecylsilica. The ratios were closer to unity for the polymeric phases than for the corresponding monomeric phases. This implies that polymeric phases act more like bulk hexadecane and exhibit a more partition-like mechanism than do monomeric phases. For all stationary phases examined in these studies, larger ratios were observed when the concentration of methanol in the mobile phase was greater than 70%. The authors suggested that this did not necessarily imply that adsorption was dominant under these conditions, but instead that the structure of the bonded phase may be different or that methanol may be more soluble in the stationary phase at higher concentrations. These studies demonstrate that the retention of a nonpolar solute with both monomeric and polymeric alkylsilica phases is more similar to a partition-like mechanism than an adsorption-like mechanism. The second important issue is whether the driving force for the separation arises from interactions of the solute with the mobile phase or with the stationary phase. The solvophobic model developed by Horvath et al. [30,31] is based on the general concept that the predominant contribution to the net change in free energy of retention arises from processes taking place in the mobile phase. The model includes changes in free energy arising from the formation of a cavity in the solvent, the reduction in free volume of the solute, and the van der Waals and electrostatic interactions between the solute and solvent [36]. This model has been very successful in predicting trends in retention, particularly those upon a change in mobile phase composition and solute structure. However, this view was challenged in a critical study by Carr et al. [37], which demonstrated that interactions with the stationary phase contribute more to the free energy of retention than was previously thought. In order to separate the mobile and stationary phase contributions, partition coefficients for solute transfer were measured from the gas phase to water (as a model of the mobile phase) and from the gas phase to hexadecane (as a partition model of the stationary phase). These partition coefficients are summarized for the alkylbenzenes in Table 1.1. The partition coefficients to water are relatively small and decrease with increasing chain length of the alkylbenzene. In contrast, the partition coefficients to hexadecane are substantially larger and increase with increasing chain length. The corresponding changes in molar free energy, calculated from Equation 1.6, exhibit similar trends. The average free energy for transfer of a methylene group is 0.67 kJ/mol from the gas phase to water, –2.53 kJ/mol from the gas phase to hexadecane, and –3.20 kJ/mol from water to hexadecane. It is evident from these data that the attractive interactions of the solute with the stationary phase are significantly greater than the repulsive interactions with the mobile phase, and hence constitute the predominant contribution to their retention. This work was extended by Ranatunga and Carr [38], who determined the change in free energy for transfer of a methylene group as a function of the mobile phase composition. For all mobile phases examined (40 to 100% methanol-water and acetonitrile-water), the changes in molar free energy were positive. In other words, the transfer of alkylbenzenes from the organic mobile phase to the gas phase was an energetically unfavorable event. The contributions of molar enthalpy were

DK4478_book.fm Page 22 Thursday, September 14, 2006 1:35 PM

22

Advances in Chromatography, Volume 45

TABLE 1.1 Partition Coefficient (K) and the Change in Molar Free Energy (ΔG) of Transfer for Alkylbenzenes to Water and n-Hexadecane as Models of the Mobile (m) and Stationary (s) Phasesa Solute Benzene Toluene Ethylbenzene n-Propylbenzene n-Butylbenzene CH2b

Km 4.26 3.94 3.22 2.27 1.75

Ks 633 2,210 5,830 16,600 48,500

ΔGm (kJ/mol)

ΔGs (kJ/mol)

ΔΔG (kJ/mol)

–3.59 –3.39 –2.90 –2.03 –1.39 0.67

–16.0 –19.1 –21.5 –24.1 –26.7 –2.53

–12.4 –15.7 –18.6 –22.1 –25.3 –3.20

a K and ΔG for transfer from gas to water at 25°C. K and ΔG for transfer from m m s s gas to n-hexadecane at 25°C. ΔΔG for transfer from water to n-hexadecane at 25°C. b Average change in molar free energy of methylene group calculated from toluene to n-butylbenzene.

Source: Adapted from Carr, P.W., Li, J., Dallas, A.J., Eikens, D.I., and Tan, L.C., Revisionist look at solvophobic driving forces in reversed-phase liquid chromatography, J. Chromatogr. A, 656, 113, 1993.

unfavorable and slightly greater than the favorable changes in molar entropy. In all cases, however, the mobile phase contributions were significantly less than the stationary phase contributions. The net changes in molar free energy, enthalpy, and entropy for a methylene group are summarized in Table 1.2. These studies conclusively demonstrate the predominant role of the stationary phase in the retention mechanism of reversed-phase liquid chromatography. 2. Effect of Alkyl Chain Length and Bonding Density Many studies have been performed to elucidate the role of the bonded phase in the retention process. Some of these studies have been focused on discerning the effect of alkyl chain length and bonding density, whereas others have examined the effect of the synthetic method. Krstulovic et al. [39] examined a series of silica bonded phases with alkyl chain lengths varying from 1 to 18. With increasing chain length, a small but systematic increase in the logarithm of selectivity for a homologous series of alkylbenzenes was observed in both methanol-water and acetonitrile-water mobile phases. The selectivity is related to the differential change in free energy of the methylene group, according to Equation 1.7. This effect, which was not observed in similar studies of liquid-liquid extraction (partition), was attributed to the more ordered structure of the alkyl chain after it was immobilized on the silica support. Tchapla et al. [40] provided more detailed measurements of retention factor and selectivity as a function of the alkyl chain length. They found that the methylene selectivity for several

DK4478_book.fm Page 23 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

23

TABLE 1.2 Thermodynamics of Transfer for a Methylene Group from Methanol-Water and Acetonitrile-Water Mobile Phases to n-Hexadecane Methanol-Water to n-Hexadecane

Acetonitrile-Water to n-Hexadecane

Percent Organic Modifier

ΔG (kJ/mol)

ΔH (kJ/mol)

T ΔS (kJ/mol)

ΔG (kJ/mol)

ΔH (kJ/mol)

T ΔS (kJ/mol)

40 50 60 70 80 90 100

–2.44 –2.06 –1.79 –1.51 –1.20 –0.94 –0.72

–2.78 –2.49 –2.11 –1.72 –1.09 –0.92 –0.89

–0.34 –0.44 –0.32 –0.21 0.11 0.03 –0.12

–1.82 –1.57 –1.39 –1.28 –1.10 –0.97 –0.77

–0.96 –0.44 –0.36 –0.75 –0.77 –0.91 –1.04

0.79 1.06 0.97 0.46 0.26 — –0.33

Source: Adapted from Ranatunga, R.P.J. and Carr, P.W., A study of the enthalpy and entropy contributions of the stationary phase in reversed-phase liquid chromatography, Anal. Chem., 72, 5679, 2000.

homologous series decreased in a stepwise manner near a critical value of carbon number for each particular stationary phase. As shown in Figure 1.2, the decrease was observed between solutes with 4 and 8 carbons for a hexylsilica phase, yet was observed between 12 and 18 carbons for an octadecylsilica phase. The authors noted that this decrease in selectivity occurred when the length of the alkyl chain of the solute was equal to or exceeded that of the stationary phase. This seemed to suggest that the partition mechanism involves a vertical penetration of the solute into the stationary phase. When the solute carbon number is less than that of the stationary phase, all carbon atoms are able to intercalate and interact directly with the corresponding carbon atoms of the stationary phase. However, when the solute carbon number is greater, the remaining carbon atoms cannot enter the stationary phase and thus undergo weaker dispersive interactions. Hence, solutes with a carbon number greater than the chain length of the stationary phase are less retained and exhibit lower selectivity. Tchapla et al. [41] further examined the temperature dependence of butadecylsilica and octadecylsilica phases. The van’t Hoff plots (log k versus 1/T; Equation 1.9) showed linear behavior for all homologous series under all conditions in the temperature range of 25° to 60°C. Extrapolation of the data from the van’t Hoff plot showed two convergent points where the critical value for all homologous series was observed at carbon number 12–13 for the butadecylsilica phase and carbon number 14–15 for the octadecylsilica phase. From these plots, the change in molar enthalpy decreased while the change in molar entropy increased with increasing solute carbon number in the homologous series. However, a discontinuity was observed in each plot at the same critical carbon numbers cited above. All of these trends were viewed to be consistent with the previous explanation of vertical insertion of the solute into the stationary phase, where changes in both enthalpy and entropy were altered at the critical carbon number. Tchapla et al. [41] also performed

DK4478_book.fm Page 24 Thursday, September 14, 2006 1:35 PM

24

Advances in Chromatography, Volume 45

1.23

Selectivity

1.21 1.19 1.17 1.15 1.13 0

5

10

15 20 Carbon number (a)

25

30

0

5

10

15 20 Carbon number (b)

25

30

1.27

Selectivity

1.25 1.23 1.21 1.19 1.17

FIGURE 1.2 Selectivity versus carbon number plots for monomeric alkylsilica phases: (a) methylsilica, (b) hexylsilica, (c) octylsilica, (d) octadecylsilica. Mobile phase: 90% methanolwater, 25°C. Selectivity is the mean value for five homologous series: n-alkanes, n-alkyl chlorides, n-methyl esters of carboxylic acids, n-alcohols, and 2-n-alkanones. (Adapted from Tchapla, A., Colin, H., and Guiochon, G., Linearity of homologous series retention plots in reversed-phase liquid chromatography, Anal. Chem., 56, 621, 1984.) Continued.

enthalpy-entropy compensation studies and determined compensation temperatures below and above the critical carbon number. In general, the solutes with shorter alkyl chains had larger compensation temperatures than those with longer chains. For example, Tc for alkanes below the critical carbon number was 542°C, while that above the critical carbon number was 477°C. Also, homologous series with more polar head groups had compensation temperatures that were consistently higher than those of series with nonpolar head groups. For example, Tc for methoxyalkanes

DK4478_book.fm Page 25 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

25

1.32

Selectivity

1.30 1.28 1.26 1.24

1.22 0

5

10

15 20 Carbon number

25

30

25

30

(c) 1.41 1.40

Selectivity

1.39 1.38 1.37 1.36 1.35 0

5

10

15 20 Carbon number (d)

FIGURE 1.2 Continued.

below the critical carbon number was 836°C, while that above the critical number was 594°C. These studies suggest a consistency in the retention mechanism of alkylsilica stationary phases. Similarly, Jinno and Ozaki [42] evaluated the enthalpy-entropy compensation of alkylbenzenes on octylsilica and octadecylsilica stationary phases. The change in molar enthalpy for the alkylbenzenes was constant and negative over the temperature range from –1° to 50°C. Enthalpy-entropy compensation was found to exist in all cases and the compensation temperatures were calculated. As the percent of water in the acetonitrile mobile phase was increased from 0.1 to 20%, the compensation temperature for octylsilica increased from 360 to 586 K, respectively, and that for octadecylsilica increased from 381 to 639 K, respectively. From the similarity in

DK4478_book.fm Page 26 Thursday, September 14, 2006 1:35 PM

26

Advances in Chromatography, Volume 45

these compensation temperatures, the retention mechanism of octylsilica and octadecylsilica was concluded to be the same. Miyabe and Guiochon [43] studied the kinetic properties of silica bonded phases with alkyl chain lengths ranging from 1 to 18 carbons. By using the plate height model, they determined the intraparticle diffusion (Equation 1.41) and the components of pore diffusion and surface diffusion for homologous alkylbenzenes. From the measured values of the surface diffusion coefficient, they calculated a hypothetical surface diffusion coefficient normalized for the carbon content of the stationary phase. By using an Arrhenius relationship (Equation 1.21), a graph of the preexponential (frequency) factor versus the activation energy of surface diffusion revealed enthalpy-entropy compensation behavior with a compensation temperature of 420 K. To validate this temperature, the authors adopted the suggestion of Krug et al. [5,6] and examined the compensation behavior from several perspectives to ensure that it was not a consequence of statistical error. A linear correlation was observed in a plot of ΔH versus ΔG for surface diffusion, with a compensation temperature of 390 K. Statistical tests confirmed the validity of this enthalpy-entropy compensation behavior. Based on these results, a simple model was derived that permitted estimation of surface diffusion coefficients under different experimental conditions with an accuracy of approximately ± 20%. These results confirm that the kinetic behavior arising from surface diffusion was similar for alkylsilica materials with varying alkyl chain length. The bonding density of alkylsilica stationary phases is another parameter that is of critical importance to the retention mechanism. Sentell and Dorsey [44] examined the partition coefficient for a series of monomeric octadecylsilica phases with bonding densities of 1.6 to 4.1 μmol/m2. They found that the partition coefficient increased linearly with bonding density up to 3.1 μmol/m2, after which it began to decrease. This trend was observed for a variety of aromatic solutes (ethylbenzene, biphenyl, p-terphenyl, naphthalene), mobile phases (methanol-water and acetonitrile-water), and temperatures (20° to 35°C). In the low-density region (< 3.1 μmol/m2), the alkyl chain order or packing structure had little effect on solute retention, whereas in the high-density region (> 3.1 μmol/m2), the chain packing was constrained and interfered with solute partitioning into the phase. These results confirmed the prediction of the lattice model of Dill [33], as discussed previously. Sentell and Dorsey [45] subsequently compared the selectivity of monomeric octadecylsilica phases as a function of bonding density. When the mobile phase composition was held constant (55% methanol-water), all changes in selectivity were attributed to free energy contributions arising from solute interactions with the stationary phase. For a homologous series of alkylbenzenes, the methylene selectivity was nearly constant at 1.96 ± 0.03 for all bonding densities from 1.74 to 4.07 μmol/m2. However, for a homologous series of polycyclic aromatic hydrocarbons (PAHs), the phenyl selectivity increased in an approximately linear manner from 7.05 to 8.18 with increasing bonding density. This increase was attributed to an increase in shape selectivity for the solutes (benzene, biphenyl, and p-terphenyl) as the bonding density increased and the alkyl chains became more ordered. Based on these observations, selectivity was determined for a series of four-ring PAHs with different annelation structures. A distinct correlation was observed between selec-

DK4478_book.fm Page 27 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

27

tivity and bonding density, which was related to the length-to-breadth ratio of the PAHs. The selectivity was also determined for the series of planar and nonplanar PAHs previously identified for stationary phase characterization by Sander and Wise [46]. For bonding densities from 1.74 to 3.56 μmol/m2, the selectivity of tetrabenzonaphthalene and benzo[a]pyrene was approximately 1.7. However, for bonding densities of 3.60 and 4.07 μmol/m2, the selectivity decreased to 1.56 and 1.63, respectively. Concurrently, a change in elution order was observed for benzo[a]pyrene and phenanthro[3,4-c]phenanthrene, indicative of a significant change in the stationary phase structure. The authors suggested that the shape selectivity observed for these monomeric octadecylsilica phases could be due to ordering of the alkyl chains rather than the degree of polymerization or thickness of the stationary phase. More detailed thermodynamic studies of the effect of bonding density were performed by Cole and Dorsey [47] using alkylbenzenes, nitroalkanes, and small PAHs as model solutes. As shown in Figure 1.3, van’t Hoff plots were linear for all solutes on monomeric octadecylsilica phases with bonding densities of 1.60 and 2.84 μmol/m2. On these phases, the change in molar enthalpy for benzene was –3.07 and –2.19 kcal/mol, respectively, and the change in molar entropy for benzene was –6.36 and –2.61 cal/mol K, respectively. Phases with bonding densities of 3.06 to 4.07 μmol/m2 were also examined, however, linear van’t Hoff plots were not observed over the temperature range from –5.0° to 80.0°C. As the bonding density increased, both ΔH and ΔS became less negative. This suggested that the entropic contribution became more significant than the enthalpic contribution as the bonding density increased. The chromatographic values for ΔH and ΔS were compared to thermodynamic values for dissolution of liquid benzene in water reported by Gill and Wadso (ΔH = 0.497 kcal/mol, ΔS = –13.8 cal/mol K at 298 K) [48]. Here, ΔH and ΔS represent the transfer from the nonpolar environment (benzene) to the polar environment (water), so the sign should be opposite from that of the chromatographic system. When this was considered, Cole and Dorsey [47] concluded that the thermodynamic values obtained for the chromatographic system were similar to those for dissolution. Cole and Dorsey [47] also examined the effect of enthalpy-entropy compensation and calculated compensation temperatures for each solute. They found that the difference in Tc values for a particular solute (e.g., toluene) on the phases with 1.60 and 2.84 μmol/m2 bonding density was greater than the difference in Tc for different solutes on the same phase, as shown in Table 1.3. This indicated that the difference in bonding density of the stationary phase contributed more significantly to the retention process than did the difference in solute structure. Other investigators have examined the effect of the synthetic method by which the alkylsilica stationary phase is prepared. Rimmer et al. [49] synthesized a series of bonded phases, tridecylsilica to octadecylsilica, by using monomeric, solution polymerized, and surface polymerized methods on the same silica support. Each alkyl chain length was synthesized by each method to yield 20 different phases. The monomeric phases had similar bonding densities of 1.19 to 3.64 μmol/m2. The solution polymerized phases had slightly higher bonding densities of 4.21 to 5.92 μmol/m2. The surface polymerized phases had the greatest variation in bonding density from 1.90 to 7.97 μmol/m2. The phases were compared with respect to

DK4478_book.fm Page 28 Thursday, September 14, 2006 1:35 PM

28

Advances in Chromatography, Volume 45

Retention factor

10

1

0.1 0.0028

0.0030

0.0032 0.0034 1/temperature (K−1)

0.0036

0.0038

0.0036

0.0038

(a)

Retention factor

10

1

0.1 0.0028

0.0030

0.0032

0.0034

1/temperature (K−1) (b)

FIGURE 1.3 Van’t Hoff plots for benzene on octadecylsilica columns (15 cm × 4.6 mm inner diameter) with varying bonding density: (a) 1.60 μmol/m2, (b) 2.84 μmol/m2, (c) 3.06 μmol/m2, (d) 3.56 μmol/m2. Mobile phase: 60% acetonitrile-water. (Adapted from Cole, L.A. and Dorsey, J.G., Temperature dependence of retention in reversed-phase liquid chromatography. 1. Stationary-phase considerations, Anal. Chem., 64, 1317, 1992.) Continued.

methylene selectivity, shape selectivity, and band broadening. No significant chromatographic differences were found to result from the synthetic routes, other than the changes in bonding density mentioned above. The methylene selectivity, measured with homologous alkylbenzenes, was observed to increase slightly from 1.28 to 1.71 with bonding density. In contrast, selectivity for tetrabenzonaphthalene and benzo[a]pyrene decreased from 1.67 for a monomeric heptadecylsilica phase (3.56 μmol/m2), to 0.64 for a solution polymerized heptadecylsilica phase (5.50 μmol/m2),

DK4478_book.fm Page 29 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

29

Retention factor

10

1

0.1 0.0028

0.0030

0.0032 0.0034 1/temperature (K−1)

0.0036

0.0038

0.0036

0.0038

(c)

Retention factor

10

1

0.1 0.0028

0.0030

0.0032

0.0034

1/temperature (K−1) (d)

FIGURE 1.3 Continued.

to 0.19 for a surface polymerized heptadecylsilica phase (7.28 μmol/m2). The concomitant inversion in retention order indicated a significant change in the stationary phase structure. To further understand the influence of the synthetic method, Jinno et al. [50] examined the temperature dependence of monomeric octadecylsilica, monomeric octadecylsilica with endcapping, polymeric octadecylsilica, and diphenylsilica phases. Van’t Hoff plots for PAHs on the monomeric octadecylsilica phases and the diphenylsilica phase were linear in the temperature range from 20° to 200°C, yet showed very high curvature on the polymeric phases. The selectivity was calculated for each solute pair at each temperature on each of the stationary phases. The authors concluded that polymeric phases were more capable of discriminating the planar/nonplanar character of PAHs than monomeric phases. This distinction was more

DK4478_book.fm Page 30 Thursday, September 14, 2006 1:35 PM

30

Advances in Chromatography, Volume 45

TABLE 1.3 Comparison of Compensation Temperatures for Toluene on Octadecylsilica Columns with Different Bonding Densitiesa Bonding Density (μmol/m2)

Tc for All Solutes (°C)

Tc for Toluene (°C)

Difference (°C)

1.60 2.84

400–800 500–1800

549 1517

968

a

Experimental conditions: 60% acetonitrile-water mobile phase.

Source: Adapted from Cole, L.A. and Dorsey, J.G., Temperature dependence of retention in reversed-phase liquid chromatography. 1. Stationary-phase considerations, Anal. Chem., 64, 1317, 1992.

marked with polymeric phases synthesized from trichlorosilanes than those synthesized from dichlorosilanes. This selectivity decreased with temperature, with a critical point between 40° and 60°C, which coincided with the temperature range where a phase transition was proposed to occur. The occurrence of a phase transition has frequently been proposed for alkylsilica phases, particularly those with long alkyl chains, high bonding density, and polymeric synthesis. Wheeler et al. [51] reviewed the published literature in this area derived from calorimetric, spectroscopic, and chromatographic studies. The primary chromatographic evidence of a phase transition has been the nonlinear nature of van’t Hoff plots (Figure 1.3). This nonlinearity suggests that the change in molar enthalpy is not constant with temperature and, hence, that the heat capacity at constant pressure (ΔCP = (∂ΔH/∂T)P) is also not constant [8]. A change in heat capacity is the clearest indication of a phase transition. Morel and Serpinet [52,53] and Claudy et al. [54] directly confirmed this conclusion by using differential scanning calorimetry and nuclear magnetic resonance (NMR) spectroscopy. The transition temperature was observed to increase with the length and density of the alkyl chains bonded to the silica surface. By using adiabatic calorimetry, Van Miltenburg and Hammers [55] demonstrated that the phase transition is actually second order and extends over a broad temperature range from –203° to 37°C for octylsilica and from –123° to 32°C for octadecylsilica phases. This second-order phase transition is viewed as an order-disorder transition rather than a traditional solid-liquid or melting transition. An ordered alkyl chain consists of all trans carbon-carbon bonds and the degree of disorder is indicated by the number and position of gauche bonds. The ordered state, in which the contact area between adjacent alkyl chains is maximized, is enthalpically preferred but entropically unfavored. In contrast, the highly disordered state, in which the chains have a great conformational diversity with many gauche bonds, is entropically preferred but enthalpically unfavored. The phase transition is thought to occur progressively, beginning at the distal or free end of the alkyl chain then gradually prevailing to the proximal or bound end. Sander et al. [56] confirmed by Fourier transform infrared (FTIR) spectroscopy that the bonded

DK4478_book.fm Page 31 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

31

TABLE 1.4 Retention Factor (k), Rate Constants (kms and ksm), and Characteristic Time (t = 1/(kms + ksm)) for Homologous Fatty Acids Separated on Polymeric Octadecylsilicaa Solute

k

kms (s–1)

ksm (s–1)

t (s)

C10 C12 C14 C16 C18 C20 C22

0.56 1.29 3.12 7.88 18.8 38.5 67.3

5.05 2.68 2.26 1.62 0.98 0.83 0.90

9.08 2.08 0.72 0.21 0.052 0.021 0.013

0.071 0.21 0.34 0.55 0.97 1.18 1.09

a

Experimental conditions: methanol mobile phase, 298 K, 4570 psi. Source: Adapted from McGuffin, V.L. and Lee, C., Thermodynamics and kinetics of solute transfer in reversed-phase liquid chromatography, J. Chromatogr. A, 987, 3, 2003.

alkyl chains contain several gauche bonds at 44°C, with a high probability of occurrence at the distal end and very low probability at the proximal end. The degree of disorder is comparable to that of the corresponding n-alkane coated in a thin film on the silica surface, but less than that of the n-alkane in bulk solution. McGuffin and Lee [57] compared the thermodynamic and kinetic behavior of a polymeric octadecylsilica phase (5.4 μmol/m2) below and above the transition temperature of 45°C. The retention factors and rate constants for a homologous series of fatty acids (C10 to C22) are summarized in Table 1.4. It is apparent that the retention factors increased systematically with carbon number by approximately two orders of magnitude. The retention factors for C10 to C22 decreased by 84.4 to 99.6%, respectively, as temperature was increased from 20° to 60°C. Hence, the phase transition had a significant effect on thermodynamic behavior. In Table 1.4, the increase in retention factor with carbon number was accompanied by changes in the rate constant for transfer from the mobile to stationary phase by approximately one order of magnitude and the rate constant for transfer from the stationary to mobile phase by approximately three orders of magnitude. Hence, the transition from the stationary to mobile phase seemed to have the greatest effect on kinetic behavior. The characteristic time (t = 1/(kms + ksm)) increased with carbon number for C10 to C18 fatty acids, but remained relatively constant thereafter. This is consistent with the mechanism described by Tchapla et al. [40,41], wherein the solutes insert vertically in the stationary phase. Hence, solutes with carbon number greater than that of the stationary phase (C18) inserted the same distance into the stationary phase and had comparable kinetic behavior. Above the transition temperature, the rate constants increased significantly and were comparable to those for monomeric octadecylsilica phases.

DK4478_book.fm Page 32 Thursday, September 14, 2006 1:35 PM

32

Advances in Chromatography, Volume 45

TABLE 1.5 Retention Factor (k) and Rate Constants (kms and ksm) for PAHs Separated on Monomeric and Polymeric Octadecylsilicaa Monomeric Octadecylsilica Solute

k

kms (s–1)

Phenanthrene Chrysene Picene Benzo[a]pyrene Phenanthro[3,4-c]phenanthrene Tetrabenzonaphthalene

0.25 0.41 0.71 0.65 0.51 1.10

≥ 400 ≥ 400 36 57 ≥ 400 ≥ 400

a

Polymeric Octadecylsilica

ksm (s–1)

k

kms (s–1)

ksm (s–1)

≥ 400 ≥ 400 43 79 ≥ 400 ≥ 400

0.44 1.15 6.14 2.36 0.70 1.29

≥ 400 221 118 207 ≥ 400 ≥ 400

≥ 400 132 9 57 ≥ 400 ≥ 400

Experimental conditions: methanol mobile phase, 293–303 K.

Source: Adapted from Howerton, S.B. and McGuffin, V.L., Thermodynamic and kinetic characterization of polycyclic aromatic hydrocarbons in reversed-phase liquid chromatography, Anal. Chem., 75, 3539, 2003.

In subsequent studies, Howerton and McGuffin [58] compared the kinetic behavior of PAHs on monomeric (2.7 μmol/m2) and polymeric (5.4 μmol/m2) octadecylsilica phases. As shown in Table 1.5, the rate constants for transfer from the mobile to stationary phase were generally larger for the polymeric phase than for the monomeric phase. However, the rate constants for transfer from the stationary to mobile phase were generally smaller for the polymeric phase than for the monomeric phase. In other words, it was faster and more facile for solutes to enter, but slower and more difficult for them to exit the polymeric stationary phases with higher bonding density. All solutes on both monomeric and polymeric phases showed a decrease in rate constants as the retention factor was increased. Finally, a number of studies have examined the effect of other synthetic methods that modify the stationary phase structure. Vervoort et al. [59] compared six different octadecylsilica stationary phases, including polymeric phases, sterically protected phases, phases with bidentate alkyl chains and embedded polar groups, as well as those synthesized with high-purity silica. Based on linear van’t Hoff plots, it was concluded that basic pharmaceutical compounds showed no change in the retention mechanism over the temperature range from 280 to 360 K. The degree to which each compound interacted with the stationary phase depended on the stationary phase properties. For example, neutral compounds (e.g., benzene, phenol) showed similar ΔH values on each stationary phase, whereas the more basic compounds (e.g., nortriptyline) showed more negative ΔH and ΔS values on the polar embedded phases and those synthesized with high-purity silica. Neue et al. [60,61] developed a method to classify these various octadecylsilica materials that illustrates the utility of graphs of ln k versus ln k and graphs of ln α vs. ln α for different solutes, previously introduced by Vailaya and Horvath [62]. Layne [63] has similarly compared conventional, polar-embedded, and polar-endcapped octadecylsilica phases by

DK4478_book.fm Page 33 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

33

using various selectivity values for nonpolar, polar, and hydrogen-bonding solutes. These studies illustrate the importance of synthetic methods and the incorporation of polar groups, particularly for the separation of acidic and basic solutes. With a few noteworthy exceptions [64], detailed thermodynamic and kinetic evaluations of these phases have yet to be reported. 3. Effect of Support The underlying solid support also has a very significant influence on retention in reversed-phase separations. As noted in the studies discussed above, the most common support material is silica. Sander and Wise [65] compared bonded phases prepared on 22 silica materials with differing particle shape, diameter, density, surface area, and pore diameter. They observed large variations in bonding density on the silica materials when performed under the same synthetic conditions, 1.26 to 2.78 μmol/m2 for monomeric octadecylsilica phases and 2.57 to 5.35 μmol/m2 for polymeric octadecylsilica phases. Two parameters were found to have the greatest influence: pore diameter and silica pretreatment with acid or base. Differences in retention factor and selectivity of PAHs were most pronounced for these parameters on the polymeric phases. For example, the retention factor of benzo[a]pyrene varied from 1.56 to 13.60 and the selectivity of tetrabenzonaphthalene and benzo[a]pyrene varied from 0.62 to 1.33. The authors concluded that the underlying silica substrate has a great effect, often unpredictable and uncontrollable, on the retention and selectivity of alkylsilica phases. In addition, much attention has been paid to the adsorptive nature of silica and its detrimental effect on the partition mechanism of alkylsilica phases. In a seminal study, Nahum and Horvath [66] examined the effect of the silica surface on the retention factor and change in molar enthalpy for dibenzo-18-crown-6 in a methanol mobile phase. The retention factor was greatest on silica and decreased progressively on octadecylsilica with 5% and 12% carbon loading. The corresponding change in molar enthalpy was determined to be –8.797, –8.433, and –7.356 kcal/mol, respectively. If partition was the predominant mechanism here, then retention and molar enthalpy should increase with carbon loading. However, if adsorption was dominant, then the phase with the highest carbon loading should have the lowest concentration of residual adsorptive sites and thus the smallest effect of these sites on retention. Hence, the retention of hydrogen bonding solutes such as the crown ethers was shown to be largely controlled by the underlying silica support, even with high carbon loading of the bonded phase. In a two-part review by Nawrocki [67,68], the nature of the silica adsorption sites was discussed and several methods for their blockage and removal were reviewed and compared. The use of pure silica as an adsorption phase will be discussed further in Section IV.B.4. Although the most common underlying support is silica, many other supports such as zirconia, titania, and synthetic polymers have been used. Zirconia supports are more stable than silica supports and can handle extremes of pH and temperature that silica cannot. However, complete coverage of the surface is essential to avoid deleterious adsorption. Li and Carr [69,70] have performed a number of studies on zirconia-based reversed-phase systems. Changes in molar enthalpy were

DK4478_book.fm Page 34 Thursday, September 14, 2006 1:35 PM

34

Advances in Chromatography, Volume 45

determined for alkylbenzenes on polybutadiene (PBD)-coated zirconia in a temperature range of 40° to 100°C. The values ranged from –2.07 to –4.00 kcal/mol for benzene to n-pentylbenzene, respectively, and are similar to those reported for octadecylsilica phases. The enthalpy of transfer of a methylene group for the PBDzirconia phase (–0.39 kcal/mol) was comparable to those for octadecylsilica phases (–0.27 and –0.41 kcal/mol) as well. In fact, the hydrophobic selectivity of PBDzirconia was found to be comparable to conventional bonded phases [69]. A temperature study by Guillarme et al. [71] compared the thermodynamic behavior of octadecyl- and PBD-zirconia with octadecylsilica phases. Over a limited temperature range of 25° to 80°C, the van’t Hoff plots were linear for alkylbenzenes and other solutes on all phases. Over a wider temperature range of 25° to 200°C, however, distinct differences were observed. Typical silica-based phases showed linear van’t Hoff plots for both methanol-water and acetonitrile-water mobile phases. However, slight curvature in the van’t Hoff plots was observed for zirconiabased phases with pure water and acetonitrile-water mobile phases, while a linear relationship was observed with methanol-water mobile phases. The authors demonstrated that the curved van’t Hoff plots were not due to changes in the system pressure with temperature, which could obscure the true temperature-dependence of retention. A confirming study with more detailed thermodynamic measurements was performed by Coym and Dorsey [72]. They observed curvature in the van’t Hoff plots for nonpolar aromatic solutes on PBD-zirconia using a pure water mobile phase. From these plots, the changes in molar enthalpy were found to be significantly more negative (exothermic) at high temperature than at low temperature. For example, the change in molar enthalpy for toluene was reported to be –5.7 kJ/mol in the temperature range of 15° to 55°C and –31.3 kJ/mol in the temperature range of 125° to 175°C. The authors attributed the differences in molar enthalpy to a change in the hydrogen-bonding network in water that influences the solvation of the solute. In spite of this large difference in molar enthalpy, the retention factors were significantly smaller in the high temperature range than in the low temperature range. The authors suggested that water is strongly hydrogen bonded at low temperature, creating a large and highly favorable change in molar entropy (hydrophobic effect). However, water undergoes little or no hydrogen bonding at higher temperatures, which results in a smaller entropy change. At higher temperatures, the entropic contribution was speculated to be larger than the enthalpic contribution, which would result in a net decrease in retention. However, we note here that a temperature-dependent change in molar entropy cannot be separated from a temperature-dependent change in molar enthalpy by means of a van’t Hoff plot. Hence, the molar enthalpies reported by Coym and Dorsey [72] may not be accurate if their proposed mechanism is correct. Clearly, further studies of this phenomenon and its mechanism are warranted [64]. 4. Effect of Mobile Phase Composition Whereas the contributions of the alkylsilica stationary phase clearly predominate [37,38], the contributions of the mobile phase also greatly influence the retention and selectivity of reversed-phase separations. To elucidate these contributions, many

DK4478_book.fm Page 35 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

35

investigators have varied the type and composition of the mobile phase to examine changes in the thermodynamic and kinetic behavior. The most common mobile phases are aqueous mixtures of methanol, tetrahydrofuran, or acetonitrile. These solvents are chosen to enhance acid, base, or dipole interactions, respectively, and hence can influence the selectivity. To influence the retention, the proportion of water in the mobile phase is varied. The effect of water on retention has been investigated extensively. Grushka et al. [73] studied the effect of water on retention factor, selectivity, and the concomitant changes in molar enthalpy and entropy. As the water content was increased, the retention factors for a series of alkylbenzenes increased and the methylene selectivity increased. The selectivity was found to be more strongly dependent on the mobile phase composition than on the temperature. Values for ΔH became more negative as the water content increased, ranging from –1.82 to –4.39 kcal/mol in 90% methanol-water and ranging from –2.06 to –6.21 kcal/mol in 80% methanol-water for ethylbenzene to dodecylbenzene, respectively. Values for ΔΔH of the methylene group were also determined as –0.267 and –0.405 kcal/mol in 90% and 80% methanol-water, respectively. Therefore, it is energetically more favorable for the solute to be in the stationary phase as the water content of the mobile phase is increased. Cole et al. [74] investigated the role of the mobile phase by means of van’t Hoff plots for benzene. When water-rich mobile phases of protic solvents (e.g., methanolwater and propanol-water) were used, the van’t Hoff plots were highly curved over the temperature range from –5° to 80°C. The values of ΔH were positive for temperatures from –5° to 20°C and negative for temperatures from 30° to 80°C. When mobile phases with aprotic solvents (e.g., acetonitrile-water) were used, the van’t Hoff plots were more nearly linear. The authors concluded that the hydrophobic (solvophobic) mechanism may be reasonable for protic solvents, but is not adequate to explain the retention mechanism in other situations. A related mobile phase effect was observed by Sentell et al. [75] through the examination of retention and selectivity for several homologous series consisting of alkylbenzenes, phenylenes, and PAHs. The selectivity changed more with temperature for methanol-water than for acetonitrile-water mobile phases. This was attributed to the hydrogen-bonding capability of the methanol-water mobile phases, which led to a more structured mobile phase. It is noteworthy that the graphs of the logarithm of selectivity vs. inverse temperature were linear and the reported values for ΔΔH and ΔΔS for the methylene and phenyl groups were negative over all temperature ranges. Sander and Field [76] examined the retention behavior of N,N-diethylaniline and 2-propylbenzene on an octadecylsilica phase over a wide range of methanol-water compositions. The change in molar enthalpy values for both solutes, although all negative, increased with an increase in the methanol concentration. That is, the transfer became less favorable for higher concentrations of methanol. The change in molar entropy for both solutes, although all negative, slightly increased with an increase in the methanol concentration. The enthalpy term was found to be the predominant contribution to retention. Barman and Martire [77] examined the effect of methanol-water and ethanol-water composition on the retention of a series of alkylbenzenes by using a semiempirical relationship between the retention volume and the temperature and mobile phase composition. Through graphs of the logarithm of retention volume vs. the volume

DK4478_book.fm Page 36 Thursday, September 14, 2006 1:35 PM

36

Advances in Chromatography, Volume 45

fraction of water in the mobile phase, the enthalpic contribution was found to be greater and opposite the entropic contribution. That is, as the volume fraction of water was increased, the enthalpic contribution to the retention volume systematically increased, while the entropic contribution systematically decreased for xylenes, ethyltoluenes, and diethylbenzenes. In all cases, the enthalpic term was the predominant contribution to retention. Wu et al. [78] examined the thermodynamic behavior of methanol-water and acetonitrile-water mobile phases as a function of pH. The mobile phase type and composition greatly affected both the retention and selectivity of piperazine diastereomers. The neutral form of the diastereomers is predominant at pH 6.4, whereas the protonated form of the diastereomers is predominant at pH 3.0. At both pH values, the logarithm of the retention factor decreased monotonically with an increase in methanol-water or acetonitrile-water composition. However, the logarithm of the diastereomeric selectivity showed interesting behavior as a function of the mobile phase composition. For methanol-water systems at pH 6.4, the selectivity decreased linearly with composition. For acetonitrile-water systems at pH 3.0 and 6.4 and for methanol-water systems at pH 3.0, the selectivity initially increased, reached a maximum value, and then decreased with composition. The authors suggested that the retention mechanism is complex, passing from a region that is dominated by selective solvation of the piperazine moiety at a low concentration of the organic modifier to a region that is dominated by more general solvent strength effects at a high concentration of the organic modifier. Through examination of the temperature behavior, the thermodynamic behavior of the mobile phases was also found to differ. At pH 3.0, the slope of the van’t Hoff plot was negative for both methanol-water and acetonitrile-water mobile phases. This was attributed to a change in the pKa of the piperazine diastereomers with increasing temperature, which favored the neutral form that was more highly retained. At pH 6.4, where the neutral form was already predominant, the behavior of the two mobile phases was distinctly different. The logarithm of the retention factor decreased with increasing temperature in methanolwater, but slightly increased in acetonitrile-water. The slope of the logarithm of the selectivity factor versus inverse temperature was positive in the methanol-water system, suggesting that the retention process was enthalpy dominated according to Equation 1.10. In contrast, the slope was slightly negative in the acetonitrile-water system, suggesting that the retention process was entropy dominated. Another common type of mobile phase involves the use of a buffer or a mixture of buffer and organic solvent. Melander et al. [79] examined aromatic carboxylic acids, substituted hydantoins, and allantoin separated on octadecylsilica using aqueous 50 mM Na2HPO4 buffer (pH 2.0) or a mixture of this buffer with 6% acetonitrile. An enthalpy-entropy compensation plot (ln k vs. –ΔH; Equation 1.16) for the aromatic carboxylic acids showed indistinguishable slopes and intercepts for the two mobile phases. Thus, the retention mechanism for the solutes did not rely on the presence of an organic modifier in the mobile phase. Similar compensation behavior was observed for the hydantoin solutes for the two mobile phases. When two other solutes (not used in the original compensation plot) were separated using a higher concentration of 30% acetonitrile, it was found that they shared the same compensation behavior. It appeared that the balance of enthalpic and entropic forces was

DK4478_book.fm Page 37 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

37

indifferent to the presence of the organic modifier and that the mechanism of solute interaction with octadecylsilica was the same for a variety of solutes. The compensation temperature for the aromatic carboxylic acids was 647 K, whereas that for the substituted hydantoins was 596 K. These temperatures were compared to substituted benzene derivatives analyzed by Knox and Vasvari [80], where the compensation temperature was 639 K. At the 95% confidence level, the range of compensation temperatures was found to be between 554 and 755 K. Both sets of solutes examined by Melander et al. [79] fell in this range, suggesting an identical retention mechanism for the solutes investigated. Compared with the thermodynamic effects, relatively little work has been done to elucidate the kinetic effects of the mobile phase. Miyabe et al. [81] studied the effect of the organic modifier on retention and mass transfer kinetics for an octadecylsilica phase. The equilibrium constant (K) increased with the surface area of the alkylbenzene solutes for mobile phases of 50% tetrahydrofuran-water, 70% methanol-water, and 70% acetonitrile-water. The isosteric heat of sorption (Qst) was calculated from van’t Hoff data and related to K0, which is the value of K at 1/T = 0 or Qst = 0. The former parameter is related to the change in molar enthalpy, while the latter is related to the change in molar entropy. From the slope of a graph of K0 versus Qst, a compensation temperature of 400 K was determined for the 50% tetrahydrofuran-water mobile phase. A kinetic characterization was also performed. The relative importance of the individual mass transfer processes was evaluated for the alkylbenzene solutes. Axial dispersion (δax) contributed about 30%, fluid to particle mass transfer (δf) contributed from 30% to 40%, and intraparticle diffusion (δd) contributed from 30 to 35%. Since each of these mass transfer processes contributed approximately equally to the second moment, none could be neglected. Among the contributions to intraparticle diffusion, surface diffusion was shown to be more significant than pore diffusion. The surface diffusion coefficient was correlated to the molar volume of the solute through the use of an empirical equation. As the solute volume increased, the surface diffusion coefficient increased. The effect of the organic modifier was also observed, where the surface diffusion coefficient increased in the order methanol < tetrahydrofuran < acetonitrile. As this order is different from the order of retention strength, this suggests that the mobile phase has a different effect on surface diffusion than on retention. Based on these studies, the type and concentration of organic modifier as well as the type and concentration of buffer and pH have a significant effect on thermodynamic behavior. Presumably, these parameters also have a significant effect on kinetic behavior, however much work remains to be done in this area. 5. Effect of Solute Structure and Concentration The retention mechanism depends not only on the characteristics of the mobile and stationary phases, but also on the solute structure and properties. Many solutes are helpful for characterization of the retention process, particularly homologous series of aliphatic and aromatic solutes. These series provide information about the contributions to retention and selectivity from an individual methylene or benzene group. McGuffin and Chen [8] determined the changes in molar enthalpy and molar volume

DK4478_book.fm Page 38 Thursday, September 14, 2006 1:35 PM

38

Advances in Chromatography, Volume 45

TABLE 1.6 Changes in Molar Enthalpy (ΔH) and Molar Volume (ΔV) for Homologous Fatty Acids Separated on Monomeric and Polymeric Octadecylsilicaa Monomeric Octadecylsilica

Polymeric Octadecylsilica

Solute

ΔH (kcal/mol)

ΔV (ml/mol)

ΔH (kcal/mol)

ΔV (ml/mol)

C10 C12 C14 C16 C18 C20 C22

–1.7 –2.2 –2.6 –3.0 –3.4 –3.8 –4.2

1.9 0.1 –0.7 –1.4 –2.3 –3.2 –4.3

–10.6 –14.2 –17.8 –21.4 –25.2 –28.1 –30.5

–27.1 –37.4 –51.7 –68.6 –83.4 –92.0 –104.0

a

Experimental conditions: methanol mobile phase, 303 K.

Source: Adapted from McGuffin, V.L. and Chen, S.H., Molar enthalpy and molar volume of methylene and benzene homologues in reversed-phase liquid chromatography, J. Chromatogr. A 762, 35, 1998.

for a homologous series of saturated fatty acids, as summarized in Table 1.6. On a monomeric octadecylsilica phase (2.7 μmol/m2), the change in molar enthalpy ranged from –1.7 to –4.2 kcal/mol for C10 to C22, respectively. The differential change in molar enthalpy (ΔΔH) per ethylene group was relatively constant at –0.41 ± 0.02 kcal/mol. This suggests that each ethylene group contributes equally to the retention process from an energetic perspective. On a polymeric octadecylsilica phase (5.4 μmol/m2), the change in molar enthalpy was much greater and ranged from –10.6 to –30.5 kcal/mol for C10 to C22, respectively. The differential change in molar enthalpy per ethylene group was relatively constant at –3.65 ± 0.13 kcal/mol, approximately an order of magnitude greater than that on the monomeric phase. The changes in molar volume were similarly distinctive and informative. For the monomeric phase, the changes in molar volume ranged from 1.9 to –4.3 ml/mol for C10 to C22, respectively. A positive change indicates that the solute occupies a greater volume in the mobile phase than in the stationary phase, whereas a negative change indicates the reverse. The differential change in molar volume (ΔΔV) per ethylene group was –1.0 ± 0.4 ml/mol, which confirmed that each ethylene group contributes equally to the retention process from a volumetric perspective. On the polymeric phase, the changes in molar volume ranged from –27.1 to –104 ml/mol for C10 to C22, respectively. This change in molar volume is a significant proportion of the total molar volume, which is approximately 201 to 441 ml/mol for C10 to C22, respectively. The differential change in molar volume per ethylene group was –14.1 ± 2.8 ml/mol, approximately an order of magnitude greater than that on the monomeric phase. These studies provide a clear indication of the thermodynamic contributions of the methylene group to retention.

DK4478_book.fm Page 39 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

39

TABLE 1.7 Activation Enthalpy (ΔH) and Activation Volume (ΔV) for Homologous Fatty Acids Separated on Polymeric Octadecylsilicaa Solute

ΔHm‡ (kcal/mol)

ΔHs‡ (kcal/mol)

ΔVm‡ (ml/mol)

C10 C12 C14 C16 C18 C20 C22

26.1 30.0 28.3 25.8 24.4 22.1 19.5

41.2 48.7 49.7 51.8 54.2 55.4 56.7

7.2 55.3 64.8 79.3 87.1 93.4 110.0

a

ΔVs‡ (ml/mol) 31.7 91.4 116.0 147.0 171.0 179.0 211.0

Experimental conditions: methanol mobile phase, 303 K.

Source: Adapted from McGuffin, V.L. and Lee, C., Thermodynamics and kinetics of solute transfer in reversed-phase liquid chromatography, J. Chromatogr. A, 987, 3, 2003.

The kinetic contributions to retention for the homologous fatty acids on the polymeric octadecylsilica phase were examined by McGuffin and Lee [57]. As shown in Table 1.4, the retention factor increased logarithmically and the rate constants decreased logarithmically with increasing carbon number. The rate constants from the mobile to stationary phase decreased by approximately one order of magnitude, while the rate constants from the stationary to mobile phase decreased by approximately three orders of magnitude for C10 to C22. Hence, under most conditions, the transfer from the stationary to mobile phase was the rate-limiting step in the retention mechanism. The concomitant changes in activation enthalpy and activation volume are summarized in Table 1.7. The activation enthalpy from the mobile phase to transition state showed small variations, but was roughly constant with increasing carbon number. In contrast, the activation enthalpy from the stationary phase to transition state increased monotonically with carbon number and was approximately twofold larger than the activation enthalpy from the mobile to stationary phase. These activation enthalpies were also substantially greater than the net change in molar enthalpy, given in Table 1.6. Hence, it is likely that the solutes do not enter and leave the stationary phase in a single step (as shown in Figure 1.1), but rather in a stepwise or progressive manner. When considered as a progressive process, the average increase in activation enthalpy in Table 1.7 was 1.6 ± 0.6 kcal/mol per ethylene group, which is energetically feasible within the chromatographic system. The activation volumes are also summarized in Table 1.7. The activation volume from the mobile phase to transition state increased monotonically by 11 ± 4 ml/mol per ethylene group. The activation volume from the stationary phase to transition state was approximately twofold larger and increased by 24 ± 9 ml/mol per ethylene group. These activation volumes were also substantially greater than the net change in molar volume, given in Table 1.6. Again, this suggests that the solutes do not

DK4478_book.fm Page 40 Thursday, September 14, 2006 1:35 PM

40

Advances in Chromatography, Volume 45

enter and leave the stationary phase in a single step, but rather in a stepwise or progressive manner. Zhao and Carr [82] examined the thermodynamic contributions of the methylene group for several nonpolar homologous series. The differential change in molar free energy (ΔΔG) per methylene group was determined for alkylbenzenes, alkylacetates, alkylphenones, alkylanilines, and nitroalkanes. The ΔΔG values were very similar, ranging from 0.303 to 0.322 kcal/mol, with an average deviation of ± 0.007 kcal/mol. Hence, the authors concluded that the terminal functional group of the homologous series has little effect on the net retention of the methylene group for alkylsilica phases. More detailed studies by Chen et al. [83] evaluated the changes in molar enthalpy and entropy for an extensive set of solutes. For homologous alkylbenzenes, p-alkylphenols, p-alkyliodobenzenes, p-alkylacetophenones, and p-alkylanilines, ΔH became systematically more negative as the alkyl chain length was increased. Although there were small variations between these homologous series, the differential change in molar enthalpy for the methylene group ranged from 1.69 to 1.91 kJ/mol, with an average deviation of ±0.09 kJ/mol. In addition, the intercept (ΔS/R – ln β) of the van’t Hoff plot became more negative with increasing alkyl chain length. These values were also relatively similar, ranging from 0.224 to 0.314. Hence, the terminal functional group of the homologous series appeared to have little effect on the enthalpic and entropic contributions to retention of the methylene group for alkylsilica phases. Carr et al. [84] examined the retention behavior of more polar functional groups, such as methoxyl, methyl ester, aldehyde, nitro, and cyano. Unlike methylene groups (Tables 1.1 and 1.2), these polar groups exhibited different values for the change in molar free energy in the chromatographic system and in a liquid-liquid extraction system using hexadecane as a model of the stationary phase. This suggests that a simple partition mechanism may not be sufficient to describe the retention of polar functional groups. Moreover, unlike methylene groups, the interactions of the polar groups with the mobile phase were energetically favorable and much greater than their interactions with hexadecane. This observation may suggest that the retention process is, in fact, driven by the mobile phase for polar groups. However, these interactions are not solvophobic, as suggested in the model of Horvath and Melander [30,31], but solvophilic in nature. Chen et al. [83] examined the changes in molar enthalpy and entropy for substituted benzenes, as summarized in Table 1.8. These values provide insight into the enthalpic and entropic contributions for individual functional groups. For solutes with a more negative change in molar enthalpy than benzene, their functional groups have a more energetically favorable interaction with the stationary phase or, equivalently, a less favorable interaction with the mobile phase. The most negative values were observed for chloro, bromo, iodo, and phenylketone functional groups. Conversely, for solutes with a less negative change in molar enthalpy than benzene, their functional groups have a more energetically favorable interaction with the mobile phase or, equivalently, a less favorable interaction with the stationary phase. The least negative values were observed for hydroxyl, methylenehydroxyl, and methylketone functional groups. Similar insight can be gained from the entropy term, provided that the accessible phase ratio remains constant for these solutes. For solutes with a more negative change in this term than

DK4478_book.fm Page 41 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

41

TABLE 1.8 Changes in Molar Enthalpy (ΔH) and Molar Entropy (ΔS/R – ln β) for Substituted Benzenes on Monomeric Octadecylsilicaa Solute

ΔH (kJ/mol)

ΔS/R – ln β

Benzene Benzyl alcohol Phenol Benzamide Benzonitrile Nitrobenzene Nitrosobenzene Fluorobenzene Chlorobenzene Bromobenzene Iodobenzene Acetophenone Benzophenone

–6.720 –5.293 –6.646 –7.513 –7.590 –6.946 –6.875 –6.798 –8.067 –8.343 –9.110 –5.408 –8.716

–2.975 –3.547 –4.048 –4.908 –3.906 –3.491 –3.261 –3.093 –3.209 –3.208 –3.332 –3.039 –3.596

a

Experimental conditions: 80% acetonitrile-water mobile phase. Source: Adapted from Chen, Z., Nakayama, T., Nakagama, T., Uchiyama, K., and Hobo, T., Thermodynamic approaches to intermolecular interaction and retention behavior in liquid chromatography, J. Liq. Chromatogr. Relat. Technol., 26, 2809, 2003.

benzene, their functional groups have a less favorable entropic contribution to retention. The most negative values were observed for amide, hydroxyl, nitrile, methylenehydroxyl, and phenylketone functional groups. Conversely, solutes with a less negative change in this term than benzene will have a more favorable entropic contribution. For these experimental conditions, no solutes had a less negative value than benzene. The balance of enthalpic and entropic contributions to retention has also been characterized. Vailaya and Horvath [62] analyzed thermodynamic data obtained by Alvarez-Zepeda et al. [85] for alkylbenzenes separated on an octadecylsilica phase. A graph of the change in molar enthalpy versus carbon number revealed a common intersection point for mobile phase compositions of 55 to 100% methanol-water. At this point, the isoenthalpic carbon number was approximately –4.5. Similarly, a graph of the entropy term (ΔS + R ln Vs) vs. carbon number revealed an isotropic carbon number of approximately –3. Because of the inequality of these values, the enthalpyentropy compensation temperature determined for varying mobile phase compositions was not constant, ranging from 1086 to 880 K. In contrast, graphs of the change in molar enthalpy and entropy versus mobile phase composition for carbon numbers ranging from two to six revealed a common isoenthalpic and isoentropic composition

DK4478_book.fm Page 42 Thursday, September 14, 2006 1:35 PM

42

Advances in Chromatography, Volume 45

of 117% methanol-water. Hence the enthalpy-entropy compensation temperature determined for varying carbon number was constant at 755 ± 28 K. Opperhuizen et al. [86] examined the enthalpy-entropy compensation behavior for an extensive series of alkylbenzenes, chlorobenzenes, chlorotoluenes, chloronaphthalenes, and chlorobiphenyls on an octadecylsilica phase. The slopes of the compensation plot (ln k vs. –ΔH) were almost identical for these solutes, which indicated that compensation temperatures were comparable at a fixed mobile phase composition. The compensation temperatures were 609 ± 14 K in 85% methanolwater and 552 ± 7 K in 90% methanol-water. The contribution of methylene and chloro groups to retention was also examined. Values for ΔΔH per methylene group increased from –1.71 to –1.56 kJ/mol, while values for ΔΔS increased from –2.81 to –2.79 J/mol K as the composition of the mobile phase increased from 85 to 90% methanol-water. Substitution of a chloro group gave a contribution to retention that was mechanistically comparable to the methylene group. These extrathermodynamic contributions were independent of the structure of the parent compound. In other studies, homologous series that vary in the number and position of benzene groups have been examined. McGuffin and Chen [8] determined the changes in molar enthalpy and molar volume for a series of PAHs, as summarized in Table 1.9. On a monomeric octadecylsilica phase (2.7 μmol/m2), the change in molar enthalpy for the cata-annelated homologous series of phenanthrene, chrysene, and picene ranged from –0.8 to –2.9 kcal/mol. The change in molar enthalpy per benzene group was not constant, which suggests that each additional benzene group did not contribute equally to the retention process from an energetic perspective. Benzo[a]pyrene, with the same number of aromatic rings (five) but a more condensed structure than picene, had a slightly less negative change in molar enthalpy of –2.2 kcal/mol. The nonplanar PAHs phenanthro[3,4-c]phenanthrene and tetrabenzonaph-

TABLE 1.9 Changes in Molar Enthalpy (ΔH) and Molar Volume (ΔV) for Polycyclic Aromatic Hydrocarbons Separated on Monomeric and Polymeric Octadecylsilicaa Monomeric Octadecylsilica

Polymeric Octadecylsilica

Solute

ΔH (kcal/mol)

ΔV (ml/mol)

ΔH (kcal/mol)

ΔV (ml/mol)

Phenanthrene Chrysene Picene Benzo[a]pyrene Phenanthro[3,4-c]phenanthrene Tetrabenzonaphthalene

–0.8 –1.6 –2.9 –2.2 –1.0 –1.5

–1.9 –1.8 –3.1 –2.1 1.9 1.6

–3.6 –7.6 –15.8 –9.1 –4.5 –3.2

–1.7 –8.2 –28.4 –10.1 1.7 6.8

a

Experimental conditions: methanol mobile phase, 303 K.

Source: Adapted from McGuffin, V.L. and Chen, S.H., Molar enthalpy and molar volume of methylene and benzene homologues in reversed-phase liquid chromatography, J. Chromatogr. A 762, 35, 1998.

DK4478_book.fm Page 43 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

43

thalene have significantly fewer negative changes in molar enthalpy than would be expected based on their six-ring structure. In fact, they are more comparable to the three- and four-ring homologous PAHs. On a polymeric octadecylsilica phase (5.4 μmol/m2), the changes in molar enthalpy were significantly greater, but exhibited the same trends as the monomeric octadecylsilica phase. The changes in molar volume were similarly distinctive and informative. On the monomeric phase, the change in molar volume for the cata-annelated homologous series of phenanthrene, chrysene, and picene ranged from –1.9 to –3.1 ml/mol. The change in molar volume per benzene group was not constant, which suggests that each additional benzene group did not contribute equally to the retention process from a volumetric perspective. Benzo[a]pyrene, with the same number of aromatic rings (five) but a more condensed annelation structure than picene, had a slightly less negative change in molar volume of –2.1 ml/mol. In contrast, the nonplanar PAHs phenanthro[3,4c]phenanthrene and tetrabenzonaphthalene exhibit positive changes in molar volume. Hence, these solutes occupy a greater volume in the stationary phase than in the mobile phase. On the polymeric phase, the changes in molar volume were greater, but exhibited the same trends as the monomeric octadecylsilica phase. These studies, as well as others [75,83,87], provide a clear indication of the thermodynamic contributions of the benzene group to retention. The kinetic contributions to retention for the PAHs on the polymeric octadecylsilica phase were examined by Howerton and McGuffin [58]. The retention factors and rate constants are summarized in Table 1.5. The rate of transfer for the threering PAH phenanthrene and the six-ring nonplanar PAHs phenanthro[3,4-c]phenanthrene and tetrabenzonaphthalene were very fast and, in fact, were greater than the capabilities of the chromatographic measurement method (≥ 400 s–1). For the remaining PAHs, the rate constants decreased with increasing ring number and increased for more condensed annelation structure. Transfer from the mobile to stationary phase was the rate-limiting step for the monomeric octadecylsilica phase, whereas transfer from the stationary to mobile phase was the rate-limiting step for the polymeric octadecylsilica phase. The concomitant changes in activation enthalpy and activation volume for these PAHs are summarized in Table 1.10. The activation enthalpies increased with ring number and decreased for more condensed annelation structure. More detailed examination of the effect of annelation structure was reported by Howerton and McGuffin [87]. The activation enthalpies and volumes were substantially greater than the net changes in molar enthalpy and volume, given in Table 1.9. As noted previously, this suggests that the solutes do not enter and leave the stationary phase in a single step, but rather in a stepwise or progressive manner. The thermodynamic and kinetic behavior of nitrogen-containing PAHs (NPAHs) has also been elucidated on a polymeric octadecylsilica phase by McGuffin et al. [88]. The retention factors for the NPAHs were smaller than those for the parent PAHs in the methanol mobile phase, while the reverse was true in acetonitrile. In fact, retention factors for aza-PAHs increased by approximately an order of magnitude in acetonitrile compared to methanol. This was attributed to the ability of the NPAHs to interact only weakly with residual silanol groups on the silica surface in the protic solvent, but much more strongly in the aprotic solvent. In general, the NPAHs had more negative changes in molar enthalpy than their parent PAHs because

DK4478_book.fm Page 44 Thursday, September 14, 2006 1:35 PM

44

Advances in Chromatography, Volume 45

TABLE 1.10 Activation Enthalpy (ΔH) and Activation Volume (ΔV) for Polycyclic Aromatic Hydrocarbons Separated on Polymeric Octadecylsilicaa Solute

ΔHm‡ (kcal/mol)

ΔHs‡ (kcal/mol)

ΔVm‡ (ml/mol)

ΔVs‡ (ml/mol)

Chrysene Picene Benzo[a]pyrene

19.9 55.7 18.0

27.5 71.9 27.3

77.9 153.0 54.0

84.6 181.0 64.2

a

Experimental conditions: methanol mobile phase, 303 K.

Source: Adapted from Howerton, S.B. and McGuffin, V.L., Thermodynamic and kinetic characterization of polycyclic aromatic hydrocarbons in reversedphase liquid chromatography, Anal. Chem., 75, 3539, 2003.

of this adsorption. It is noteworthy, however, that the changes in molar enthalpy were relatively comparable in the two mobile phases. For example, benz[a]acridine had ΔH values of –6.6 ± 0.1 kcal/mol in methanol and –8.3 ± 2.1 kcal/mol in acetonitrile. Hence, the differences in thermodynamic behavior were attributed to changes in the intercept (e.g., molar entropy or phase ratio) rather than the slope (e.g., molar enthalpy) of the van’t Hoff plot. The kinetic behavior was equally dramatic, with rate constants differing by two to four orders of magnitude in the two mobile phases. For example, the rate constants for transfer of benz[a]acridine from the mobile to stationary phase were 47 and 1.3 × 10–2 s–1 in methanol and acetonitrile, respectively, at 303 K. The corresponding rate constants for transfer from the stationary to mobile phase were 80 and 2.9 × 10–3/s–1, respectively. In nearly all cases, the latter process was the rate-limiting step of the retention mechanism. These results clearly indicate the importance of adsorption on the underlying support in alkylsilica phases. Further studies with aqueous mobile phases and additives are desirable for more thorough characterization. The effect of solute concentration on kinetic behavior was examined by Miyabe and Guiochon [89] on an octadecylsilica phase. From the estimation of plate height, the axial dispersion coefficient was found to increase linearly with increasing concentration in the range of 0 to 102 g/ml of 4-(2-methyl-2-propyl)phenol. The mass transfer rate constant was also found to increase linearly with concentration. Among the kinetic contributions to the rate constant, the sorption/desorption and external mass transfer processes were negligibly small and intraparticle diffusion was dominant. Moreover, among the contributions to intraparticle diffusion, surface diffusion was shown to be much more significant than pore diffusion. Hence, the authors concluded that the concentration dependence of the rate constant arises primarily from the surface diffusion coefficient. This concentration dependence of the surface diffusion coefficient was explained by the chemical potential driving force model. These results are significant because they clearly indicate the important role of surface diffusion processes in the kinetic behavior of alkylsilica phases.

DK4478_book.fm Page 45 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

45

In summary, alkylsilica stationary phases operate predominantly by a partition mechanism, where the interaction of the solute with the stationary phase contributes more greatly than that with the mobile phase to the net change in free energy. The underlying support plays an important role in both the thermodynamic and kinetic behavior, often leading to a combined mechanism of partition and adsorption. As the alkyl chain length and bonding density increase, the changes in molar enthalpy increase and kinetic rate constants decrease. The mobile phase composition, likewise, influences changes in molar enthalpy by means of the water content and the type of organic modifier. The kinetic behavior is influenced by the organic modifier and its effect on the surface diffusion coefficient. Finally, the solute structure provides a broad view of the retention mechanism from both a thermodynamic and kinetic perspective.

B. ADSORPTION The adsorption mechanism has several features that distinguish it from the partition (absorption) mechanism. First, the stationary phase consists of an impermeable, solid surface. Common stationary phases include organic materials such as poly(styrenedivinylbenzene) (PS-DVB) and porous graphitic carbon (PGC), as well as inorganic materials such as silica and alumina. Most of these adsorbents are heterogeneous, having surface sites with different functional groups. These functional groups generally involve strong and selective interactions, including proton donor-acceptor (Brönsted acid-base) and electron donor-acceptor (Lewis acid-base) interactions, among others [1]. As a consequence of these strong interactions, there is often a fixed stoichiometry between the solute and the stationary phase, most frequently 1:1, but occasionally 2:1 or higher. Also, because the stationary phase is a solid with a fixed surface area, there are a limited number of surface sites. Consequently, the solute molecules are in competition for the surface sites and may be displaced by more strongly interacting species. These features of the adsorption mechanism make it distinctly different from partition. In this section, the thermodynamic and kinetic characterization of the most common adsorption phases will be reviewed. 1. Poly(styrene-divinylbenzene) The copolymer PS-DVB has long been used as a stationary phase for size exclusion chromatography and, more recently, for reversed-phase liquid chromatography [90,91]. The stationary phase is prepared by polymerization of styrene and divinylbenzene monomers in aqueous suspension in the presence of an initiator and porogen. By controlling such factors as the concentration of monomers, type and concentration of porogen, temperature, and time duration of the polymerization, the properties of the resulting particles can be regulated. The particles are spherical with well-defined diameter and pore size and with a semirigid to rigid structure, depending upon the degree of cross-linking. PS-DVB is a promising alternative to octadecylsilica for reversed-phase separations, as it is stable at pH 1 to 13, at temperatures up to 130°C, and can be used with a wide range of mobile phase solvents [92]. It does have some disadvantages, however, as polymer-based particles typically exhibit lower efficiencies and lower pressure capabilities than silica-based particles. Although PS-DVB

DK4478_book.fm Page 46 Thursday, September 14, 2006 1:35 PM

46

Advances in Chromatography, Volume 45

TABLE 1.11 Isomeric Selectivity for Phenyltoluenes, Terphenyls, Stilbenes, and Terpinenes on Poly(styrene-divinylbenzene) and Octadecylsilica Phasesa Phenyltoluenes Stationary Phase PS-DVB Octadecylsilica a

Terphenyls

Stilbenes

Terpinenes

m-/o-

p-/m-

m-/o-

p-/m-

trans-/cis-

γ-/α-

1.26 1.05

1.00 1.00

2.13 1.23

1.09 1.09

1.13 1.03

1.08 1.13

Experimental conditions: 50% acetonitrile-water mobile phase, 303 K.

Source: Adapted from Zhao, J. and Carr, P.W., A comparative study of the chromatographic selectivity of polystyrene-coated zirconia and related reversed-phase materials, Anal. Chem., 72, 302, 2000.

has been well characterized in terms of column permeability, reproducibility, porosity, and loading capacity [93], many of the fundamental thermodynamic and kinetic properties have not been characterized. The following studies have proven to be invaluable in understanding the underlying retention characteristics of PS-DVB. A comparison of PS-DVB with other reversed-phase materials, such as octadecylsilica and PBD-coated zirconia and alumina, was performed by McNeff et al. [94]. The retention factors for a homologous series of alkylbenzenes were evaluated on each stationary phase using 40% acetonitrile-water with 50 mM phosphate buffer at pH 3.2 as the mobile phase. A linear relationship was observed between the natural logarithm of the retention factor and the number of methylene groups for all phases. The retention was greatest on PS-DVB, followed closely by octadecylsilica, and much less on PBD-zirconia and PBD-alumina. Values for the change in molar free energy for a methylene group were –0.583 kJ/mol for PS-DVB, –0.880 kJ/mol for octadecylsilica, –0.826 kJ/mol for PBD-zirconia, and –0.906 kJ/mol for PBD-alumina. The PS-DVB phase had the smallest ΔΔG value, which accounted for the difference in hydrophobicity of this phase compared to the other reversed-phase materials. The selectivity factors for several phenyl-substituted isomers were calculated on a PS-DVB phase (PRP-1; Hamilton Company, Reno, NV) and compared to octadecylsilica by Zhao and Carr [82]. Values for the isomeric selectivity are summarized in Table 1.11. The PS-DVB phase showed increased selectivity over octadecylsilica for separation of the ortho- and meta- isomers of phenyltoluene and terphenyl, whereas the meta- and para- isomers were not well separated on either phase. The PS-DVB phase also showed increased selectivity for the trans- and cis-stilbenes. This study demonstrated an inherent difference in the molar free energy of retention for isomeric compounds on PS-DVB compared to octadecylsilica, which is often a characteristic feature of the adsorption mechanism compared to partition. The effect of the mobile phase has been examined by several investigators. Guillarme et al. [71] observed the mobile phase dependence of van’t Hoff plots for small organic solutes on a PS-DVB phase (PLRP-S; Polymer Laboratories, Shropshire, UK). Linear van’t Hoff plots were observed for methanol-water mobile phases,

DK4478_book.fm Page 47 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

47

while curvilinear van’t Hoff plots were observed for acetonitrile-water mobile phases in the temperature range from 20° to 180°C. Possible origins of the nonlinear behavior were discussed, including the ability of PS-DVB to swell as a function of temperature and mobile phase composition. Yarita et al. [95] used superheated water as a mobile phase for the separation of phenols on a PS-DVB phase. Van’t Hoff plots showed linear behavior with an increase in temperature from 100° to 150°C. It was suggested that ΔH and ΔS were independent of temperature, although values for these parameters were not given or discussed in any further detail. The kinetic behavior of PS-DVB has been investigated extensively by Cantwell and coworkers [22,23,96]. Li and Cantwell [22] examined the mass transfer kinetics of PS-DVB phases (PRP-1 and PRP-∞; Hamilton Company, Reno, NV) with 85% methanol-water mobile phase by using the shallow-bed technique in the sorption mode. The sorption rate of naphthalene on PRP-1, a macroporous PS-DVB phase, was monitored as a function of time and the data were fit to a biporous diffusion model. From this fit, it was calculated that 91 ± 3% of the total sorption was due to adsorption at the surface of the pore walls, while 9 ± 3% was due to partition in the matrix. The diffusion coefficient for pore diffusion was estimated to be 3 × 10–6 cm2/s, while that for matrix diffusion was no greater than 10–12 cm2/s. Thus, the slow intraparticle mass transfer rate was attributed to slow diffusion in the polymer matrix. To confirm this result, a nominally nonporous PS-DVB phase, PRP-∞, was examined by using a monoporous diffusion model. The effective diffusion coefficient was 4 × 10–9 cm2/s, which was consistent with the slow matrix diffusion observed in PRP-1, but the magnitude was not in good agreement. Bujalski and Cantwell [23] later used the shallow-bed technique in the desorption mode with correction for experimental artifacts by means of an impulse response function marker (phloroglucinol). This allowed more accurate estimation of the kinetic processes, resulting in a diffusion coefficient of 5.0 × 10–11 cm2/s for PRP-∞, which is in much better agreement with that for PRP-1. Ells et al. [96] examined the effect of swelling on mass transfer rates in PRP-∞. For this purpose, a good wetting solvent for PS-DVB, tetrahydrofuran, was added to the 70% methanol-water mobile phase. As the concentration of tetrahydrofuran was increased from 0 to 10%, the diffusion coefficient of naphthalene increased from 7 × 10–11 to 15 × 10–11 cm2/s. The tetrahydrofuran cosolvent caused two effects: increased swelling of the polymer matrix and decreased sorption capacity for naphthalene. The authors suggested that the latter effect was most likely correlated with the improvement in chromatographic peak shape. They proposed that tetrahydrofuran filled the small micropores between polymer chains, which prevented naphthalene from entering these micropores and thereby enhanced the rate of mass transfer. The use of PS-DVB for reversed-phase separations is growing as more fundamental knowledge of the retention mechanism is elucidated. While the preceding work has proven to be important, many more thermodynamic and kinetic studies are needed to fully elucidate the combined adsorption/partition mechanism. 2. Porous Graphitic Carbon Porous graphitic carbon is another common stationary phase for reversed-phase liquid chromatography that operates by an adsorption mechanism. Detailed studies

DK4478_book.fm Page 48 Thursday, September 14, 2006 1:35 PM

48

Advances in Chromatography, Volume 45

TABLE 1.12 Retention Factor (k) and Kinetic Rate Constants (kms, ksm) for Alkylbenzenes on Porous Graphitic Carbona Solute

k

kms (s–1)

ksm (s–1)

Benzene Toluene Ethylbenzene n-Propylbenzene n-Butylbenzene n-Pentylbenzene n-Hexylbenzene

0.16 0.36 0.37 0.53 0.80 1.43 2.34

1.01 4.86 4.53 8.39 14.0 24.5 37.4

5.87 13.1 11.8 15.0 16.7 16.3 15.4

a

Experimental conditions: methanol mobile phase, 298 K.

Source: Adapted from Zhang, Y. and McGuffin, V.L., Thermodynamic and kinetic characterization of porous graphitic carbon in reversed-phase liquid chromatography, J. Chromatogr. A, submitted for publication.

of the synthesis, structure, and performance of PGC have been performed by Knox et al. [97] and Knox and Ross [98]. PGC is synthesized by injection and subsequent polymerization of a phenol-formaldehyde mixture inside the pores of silica particles. The material is heated to 1000°C in nitrogen, after which the silica is dissolved with potassium hydroxide. The remaining carbonaceous material is then heated at 2000°C in argon to form the PGC. The resulting particles are spherical with a diameter and pore size similar to the original silica template. The surface consists of intertwined ribbons of carbon, which impart mechanical strength and the ability to withstand high pressure. Two common forms are the Bernal structure of three-dimensional graphite layers ordered ABABAB and the Warren structure of two-dimensional graphite layers that are randomly ordered. The homogeneous graphitic surface is hydrophobic and highly polarizable, hence retention is largely controlled by solute polarity and size. A review by Hanai [99] summarizes the retention characteristics of PGC. In many studies, the retention and selectivity of homologous series have been used to characterize PGC and compare it with octadecylsilica and other reversedphase materials. Zhang and McGuffin [100] examined the thermodynamic and kinetic behavior of alkylbenzenes and methylbenzenes on Hypercarb PGC (Thermo Electron Corp., Waltham, MA) with a methanol mobile phase. Representative data for the retention factors are summarized in Table 1.12. Van’t Hoff plots were linear over the temperature range from 23° to 53°C and pressure range from 1000 to 5000 psi (6.9 to 34.5 MPa). For all solutes examined under all conditions, negative values for ΔH indicated that transfer from the mobile to stationary phase was energetically favorable. The molar enthalpy became more negative as the carbon number increased. The differential change in molar enthalpy (ΔΔH) was determined to be

DK4478_book.fm Page 49 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

49

–2 kJ/mol for the alkylbenzenes and –2.5 kJ/mol for the methylbenzenes, which suggested that each methylene or methyl group produced a similar effect on retention. The change in molar volume was determined to be nearly zero, which was consistent with the adsorption mechanism of PGC. Representative data for the kinetic rate constants are also summarized in Table 1.12. The rate constants from the mobile to stationary phase increased as the number of methylene or methyl groups increased. However, the rate constants from the stationary to mobile phase remained relatively constant. The activation enthalpy from the mobile phase to transition state was small and relatively constant, whereas that from the stationary phase to transition state became progressively larger as the carbon number increased. Hence, the energetic barrier for desorption was most significant. In contrast, the activation volume, which represents the volumetric barrier, for both adsorption and desorption was nearly zero. These thermodynamic and kinetic properties were significantly different from those for octadecylsilica, as discussed previously [57]. Viron et al. [101] reported the behavior of fatty acid methyl esters with varying carbon number and degree of unsaturation. In general, the retention of all fatty acids was significantly greater on PGC than on representative octadecylsilica phases. The methylene selectivity for the PGC phase was 1.58 with a 50% methanol-dichloromethane mobile phase, while that for the octadecylsilica phases was 1.14 to 1.20 with a methanol mobile phase. Thus, both phases had a constant contribution to retention for each methylene group, but the interaction was stronger with PGC than with octadecylsilica. The addition of one to three double bonds caused a decrease in retention on both phases. However, the contribution to retention was not constant with each successive double bond. For PGC, the selectivity was 2.77 for addition of the first double bond, 1.59 for the second, and 1.04 for the third. For octadecylsilica, the selectivity was 1.38 to 1.66 for addition of the first double bond, 1.17 to 1.20 for the second, and 1.10 to 1.17 for the third. Thus, double bonds had greater interaction with PGC than with octadecylsilica. Interestingly, retention on the PGC phase increased with the degree of unsaturation for fatty acids with five or six double bonds. This behavior, which is inconsistent with conventional reversed-phase separations, was attributed to the stereochemistry of the double bonds and their ability to align well with the planar surface of the graphite. The effect of the mobile phase composition was evaluated by Gaudin et al. [102]. They examined a variety of mobile phases for the separation of fatty acid methyl esters on Hypercarb PGC. Methanol was selected as the reference solvent, as it is typically a weak eluent for PGC. By taking the logarithmic ratio of the methylene selectivity for methanol and the solvent of interest, the eluotropic solvent strength (ε°) was established. The values for methylene selectivity and solvent strength are summarized in Table 1.13. In general, the weakest solvents were more polar (methanol and acetonitrile), while the strongest solvents were less polar (tetrahydrofuran, dichloromethane, chloroform, and toluene). However, the size, geometry, and polarizability of the solvents also played a role in the solvent strength. To investigate the solvent strength of binary systems, one weak solvent was combined with one strong solvent and the selectivities were determined using the fatty acid methyl esters. Acetonitrile combined with equal amounts of dichloromethane or tetrahydrofuran led to equivalent solvent strengths. In contrast, dichloromethane-methanol mobile

DK4478_book.fm Page 50 Thursday, September 14, 2006 1:35 PM

50

Advances in Chromatography, Volume 45

TABLE 1.13 Methylene Selectivity (αCH2) and Eluotropic Solvent Strength (ε°) for Mobile Phases on Porous Graphitic Carbon Mobile Phase

αCH2

ε°

Methanol Acetonitrile Acetone 1-Propanol Ethyl acetate Heptane Dichloromethane Tetrahydrofuran Chloroform Toluene

1.70 1.66 1.48 1.45 1.38 1.29 1.26 1.23 1.20 1.15

0.00 0.01 0.06 0.07 0.08 0.11 0.13 0.13 0.15 0.17

Source: Adapted from Gaudin, K., Chaminade, P., and Baillet, A., Eluotropic strength in non-aqueous liquid chromatography with porous graphitic carbon, J. Chromatogr. A, 973, 61, 2002.

phases were stronger than the corresponding tetrahydrofuran-methanol phases. Moreover, toluene-methanol and chloroform-methanol were stronger than tolueneacetonitrile and chloroform-acetonitrile mobile phases. It was determined that the use of binary solvents can greatly affect the eluotropic solvent strength of the mobile phase and, thus, the retention and selectivity on the PGC stationary phase. The effect of water on retention behavior was examined by Guillarme et al. [71]. The authors observed linear van’t Hoff plots for small organic molecules with both methanolwater and acetonitrile-water mobile phases on a Hypercarb PGC column. These results demonstrated that both ΔH and ΔS were independent of temperature over the range from 20° to 180°C for all mobile phases, including pure water. The separation of a variety of solutes can help to elucidate the retention mechanism of PGC. Koivisto and Stefansson [103] examined the retention of sulfonated and acetylated disaccharides on Hypercarb PGC at two extreme pH values and temperatures from 30° to 80°C. The retention order for the disaccharides was reversed at the two pH levels. At low pH (2.2), the strongest interactions arose from the sulfate ion such that retention increased logarithmically with the number of sulfate groups. At high pH (12), hydrophobic interactions were dominant, as in conventional reversed-phase separations, and retention decreased with the number of sulfate groups. These results suggested a possible change in retention mechanism at different pH values. To examine this in more detail, linear van’t Hoff plots were obtained at both pH values and were used to evaluate the changes in molar enthalpy and entropy. At low pH (2.2), retention increased with temperature and both ΔH and ΔS values were positive. The unfavorable changes in enthalpy and favorable changes

DK4478_book.fm Page 51 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

51

in entropy characterized the retention of all disaccharides. At high pH (12), retention decreased with temperature and both ΔH and ΔS values were negative for all disaccharides. There was no evidence of enthalpy-entropy compensation within this system at low or high pH values. Another oligosaccharide that has been studied extensively as both a solute and a mobile phase additive is cyclodextrin. Kwaterczak and Bielejewska [104] examined native and permethylated derivatives of α-, β-, and γ-cyclodextrins as solutes on Hypercarb PGC. For the native cyclodextrins, retention increased logarithmically with the number of glucose units (6, 7, and 8 for α-, β-, and γ-cyclodextrin, respectively). The glucose selectivity was constant at 2.37 for 15% ethanol-water mobile phase. Values for ΔH and ΔS became progressively more negative with an increasing number of glucose units, and the adsorption of native cyclodextrins was determined to be enthalpy dominated. The ratio of entropic and enthalpic forces (TΔS/ΔH) for the PGC phase was compared to that for a representative octadecylsilica phase. The values for PGC were typically smaller than those for octadecylsilica (e.g., 0.82 and 0.94, respectively, for α-cyclodextrin), indicating that the unfavorable entropy term plays a lesser role in the adsorption mechanism for PGC. For the permethylated cyclodextrins, retention decreased with the number of glucose units. While the ΔH values were negative, the ΔS values were positive, and the adsorption of permethylated cyclodextrins was determined to be entropy dominated. As in the previous study, there was no evidence of enthalpy-entropy compensation for either native or permethylated cyclodextrins on the PGC phase. The addition of cyclodextrin to the mobile phase can modify retention based on the solute functional groups, size, and strength of the complex that can be formed. Clarot et al. [105] used this approach to evaluate the retention of four terpene derivatives on PGC. When the β-cyclodextrin concentration was zero, linear van’t Hoff plots were obtained over the temperature range from 25° to 45°C. Both ΔH and ΔS values were negative, and the retention process was enthalpy dominated. Plots of ΔH vs. ΔS were linear, which indicated that the terpenes exhibited enthalpyentropy compensation on PGC with a compensation temperature of 39.5°C. When β-cyclodextrin was added to the mobile phase, the retention factors for all terpenes decreased, as would be expected for complexation. The equilibrium formation constants were determined as a function of temperature, and from these data linear van’t Hoff plots were obtained. Negative values for ΔG confirmed that stable complexes were formed; however, positive values for ΔH indicated that this process was energetically unfavorable. Values for TΔS were larger than those for ΔH, which indicated that complexation was an entropy-dominated process. Plots of ΔH vs. ΔS were linear, with a compensation temperature of 19.1°C. Because the compensation temperatures were similar for the mobile phases with and without β-cyclodextrin, the authors suggested that complexation and transfer from the mobile phase to the PGC stationary phase had a comparable influence on retention. They also noted that this behavior was not observed for octadecylsilica stationary phases, where complexation played a much greater role. Other additives can be introduced into the mobile phase to influence retention. Andre and Guillaume [106] examined the effect of sucrose (saccharose) concentration in a 70% methanol-water mobile phase for both PGC and octadecylsilica phases.

DK4478_book.fm Page 52 Thursday, September 14, 2006 1:35 PM

52

Advances in Chromatography, Volume 45

Langmuir and bi-Langmuir models were used to evaluate the data under nonlinear conditions. The Langmuir model was appropriate under all conditions for the octadecylsilica phase, which suggested that the addition of sucrose did not alter the interaction of the phenolic solutes with the stationary phase. More complex behavior was observed for the PGC phase, where the bi-Langmuir model was necessary. The apparent retention factor decreased with increasing concentration of sucrose up to a critical value of 0.03 M, where it then increased. For low concentrations of sucrose, retention was dominated by the contributions from phenol adsorption on the PGC surface. Hence, the changes in molar enthalpy and entropy became progressively less negative with increasing sucrose concentration due to competition for the adsorption sites. For higher concentrations of sucrose, retention was dominated by the association of phenolic solutes with sucrose adsorbed on the PGC surface. Under these conditions, the retention factor increased with increasing sucrose concentration and the changes in molar enthalpy and entropy became more negative. Enthalpyentropy compensation was observed within both regimes, with a compensation temperature of 357 K for low sucrose concentrations and 340 K for higher sucrose concentrations. The authors noted that this difference, which was statistically significant, confirmed the different retention mechanism in the two regimes. Porous graphitic carbon is a very useful stationary phase for reversed-phase separations because of the distinct differences in selectivity from octadecylsilica and PS-DVB. However, the general retention characteristics of PGC for nonpolar, polar, and ionic solutes should be examined in much more depth and detail through thermodynamic and kinetic investigations. 3. Other Organic Adsorbents Other types of organic phases can be used as adsorbents in reversed-phase chromatography. Chen et al. [83] examined the changes in molar enthalpy and entropy for substituted benzenes on phenylsilica and pyrenylsilica phases, as summarized in Table 1.14. These values provide insight into the enthalpic and entropic contributions for individual functional groups, which can be compared with the values for octadecylsilica in Table 1.8. For solutes with a more negative change in molar enthalpy than benzene, their functional groups have a more energetically favorable interaction with the stationary phase or, equivalently, a less favorable interaction with the mobile phase. For the phenylsilica phase, the most negative values were observed for amide, bromo, iodo, nitroso, and hydroxyl functional groups. For the pyrenylsilica phase, the most negative values were observed for iodo, fluoro, nitrile, and phenylketone groups. Conversely, for solutes with a less negative change in molar enthalpy than benzene, their functional groups have a more energetically favorable interaction with the mobile phase or, equivalently, a less favorable interaction with the stationary phase. For the phenylsilica phase, the least negative values were observed for methylenehydroxyl, nitro, chloro, and methylketone functional groups. For the pyrenylsilica phase, the least negative values were observed for methylenehydroxyl, amide, and methylketone groups. Similar insight can be gained from the entropy term, provided that the accessible phase ratio remains constant for these solutes. For solutes with a more negative change in this term than benzene, their functional groups have

DK4478_book.fm Page 53 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

53

TABLE 1.14 Changes in Molar Enthalpy (ΔH) and Molar Entropy (ΔS/R-ln β) for Substituted Benzenes on Phenyl and Pyrenyl Phasesa Phenylsilica

Pyrenylsilica

Solute

ΔH (kJ/mol)

ΔS/R – ln β

ΔH (kJ/mol)

ΔS/R – ln β

Benzene Benzyl alcohol Phenol Benzamide Benzonitrile Nitrobenzene Nitrosobenzene Fluorobenzene Chlorobenzene Bromobenzene Iodobenzene Acetophenone Benzophenone

–8.734 –7.510 –9.185 –9.423 –9.163 –8.609 –9.249 –9.074 –8.681 –9.250 –9.154 –8.329 –8.978

–4.356 –4.724 –5.119 –5.576 –4.724 –4.410 –4.609 –4.510 –4.173 –4.314 –4.167 –4.418 –4.202

–9.176 –7.344 –8.932 –7.832 –9.276 –8.947 –8.853 –9.248 –9.049 –9.141 –9.542 –8.397 –9.290

–4.268 –4.201 –4.843 –4.726 –4.420 –4.069 –4.032 –4.343 –3.969 –3.878 –3.829 –4.043 –3.778

a

Experimental conditions: 80% acetonitrile-water mobile phase.

Source: Adapted from Chen, Z., Nakayama, T., Nakagama, T., Uchiyama, K., and Hobo, T., Thermodynamic approaches to intermolecular interaction and retention behavior in liquid chromatography, J. Liq. Chromatogr. Relat. Technol., 26, 2809, 2003.

a less favorable entropic contribution to retention. For the phenylsilica phase, the most negative values were observed for amide, hydroxyl, methylenehydroxyl, nitrile, and nitro functional groups. For the pyrenylsilica phase, the most negative values were observed for hydroxyl, amide, nitrile, and fluoro groups. Conversely, solutes with a less negative change in this term than benzene will have a more favorable entropic contribution. For both the phenylsilica and pyrenylsilica phases, the least negative values were observed for chloro, bromo, iodo, and phenylketone functional groups. By comparison of Tables 1.8 and 1.14, it is apparent that these aromatic phases provide greater overall retention and different functional group selectivity than octadecylsilica phases. 4. Silica Silica is by far the most widely used adsorbent and solid support for liquid chromatography. Silica is superior to other supports in terms of efficiency, rigidity, and general performance. The weakly acidic silanol groups (Si–OH) and weakly basic siloxane groups (Si–O–Si) are primarily responsible for the adsorption mechanism. There have been extensive reviews concerning the physical and chemical nature of the silica surface [67,68,107,108].

DK4478_book.fm Page 54 Thursday, September 14, 2006 1:35 PM

54

Advances in Chromatography, Volume 45

TABLE 1.15 Free Energy (ΔG kcal/mol) of Adsorption for Individual Functional Groups in Aliphatic (R) and Aromatic (Ar) Molecules on Silica Group X–CH3 X–CH2–Y X–C=C–Y X–F X–Cl X–Br X–I X–SH X–S–S–Y X–S–Y X–O–Y X2–N–Y X–(C=O)H X–NO2 X–CN X–(C=O)O–Y X–(C=O)–Y X–OH X–NH2 X–(S=O)–Y X–(C=O)OH X–(C=O)NH2

X = Ar Y = Ar

X=R Y = Ar

X=R Y=R

0.11 0.07 0.25 –0.15 –0.20 –0.17 –0.15 0.67

— 0.01 (0.25) — — — — — 0.94 1.29 1.83 2.52 — — — 3.45 4.69 — — 4.2 — —

0.07 –0.05 (0.25) 1.54 1.74 1.94 1.94 1.70 1.90 2.94 3.61 ~5.8 4.97 5.71 5.27 5.27 5.27 5.60 8.00 7.2 7.6 9.6

0.48 0.87 3.48 2.77 3.33 4.18 4.56 4.20 5.10 6.1 6.6

Source: Adapted from Snyder, L.R., Principles of Adsorption Chromatography, Marcel Dekker, New York, 1968.

The vast majority of separations using silica as an adsorbent are performed in the normal-phase mode. Snyder [109–112] has performed detailed and extensive studies of normal-phase adsorption for solutes with widely varying structure. The retention of these solutes was used to calculate the group adsorption energy, which is the change in free energy for a specific functional group minus that for an equivalent volume of alkane. Selected values of the group interaction energy are summarized in Table 1.15. The adsorption energy for methyl and methylene groups is nearly zero, whereas that for aromatic and higher bond orders is slightly greater. In general, the adsorption energy for specific functional groups is greater in aliphatic solutes than in aromatic solutes. This trend arises because the electron density of the functional group remains localized in aliphatic molecules, but is dispersed throughout the aromatic ring, thereby decreasing its availability for interaction with the silica surface. Accordingly, aliphatic halogen compounds have very high adsorption energy relative to their aromatic counterparts. This results from the high

DK4478_book.fm Page 55 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

55

polarizability and dipole moment of the halogens, leading to greater interactions from dipole induction and orientation [1]. Stronger interactions arise for functional groups that undergo proton donor-acceptor or electron donor-acceptor interactions. For acidic groups such as thiol, hydroxyl, and carboxyl, the adsorption energy arises primarily from interaction with siloxane groups. For basic groups such as carbonyl, nitrile, and amine, the adsorption energy arises primarily from interactions with silanol groups. If the functional groups of the solute act independently then, to a first approximation, the group interaction energies are additive. Hence, the overall interaction energy can be estimated by summation of the individual values. However, if the functional groups of the solute do not act independently, then secondary structural effects can arise [109]. For example, if the groups are in close proximity, they can both adsorb to the silica surface at the same time and can mutually enhance adsorption. Alternatively, if the groups interact with one another, they can diminish interaction with the surface. In addition, bulky groups can cause steric hindrance of neighboring groups, which can reduce their interaction. Other secondary effects can arise from intramolecular electronic induction, especially in aromatic solutes. The change in local electron density may increase or decrease adsorption, depending upon the nature of the functional group. The effect of mobile phase composition has also been examined by Snyder [109,112] for normal-phase adsorption on silica. The eluotropic solvent strength (ε°) was calculated for representative solvents relative to pentane, as summarized in Table 1.16. In contrast to the behavior on PGC (Table 1.13), the solvent strengths on silica increased with polarity. The weakest solvents were those of lowest polarity (pentane, benzene, carbon tetrachloride), whereas the strongest solvents were those of greatest polarity (acetone, dioxane, acetonitrile). To investigate the solvent strength of binary systems, one weak solvent was combined with one strong solvent. Mathematical

TABLE 1.16 Eluotropic Solvent Strength (ε°) for Mobile Phases on Silica Mobile Phase

ε°

Pentane Carbon tetrachloride Benzene Chloroform Methylene chloride Ethyl ether Ethyl acetate Acetone Dioxane Acetonitrile

0.00 0.11 0.25 0.26 0.32 0.38 0.38 0.47 0.49 0.50

Source: Adapted from Snyder, L.R., Linear elution adsorption chromatography. XIII. A narrow pore silica, J. Chromatogr., 25, 274, 1966.

DK4478_book.fm Page 56 Thursday, September 14, 2006 1:35 PM

56

Advances in Chromatography, Volume 45

TABLE 1.17 Changes in Molar Enthalpy (ΔH) and Entropy (ΔS) for Amino Acid Adsorption on Silica and Octadecylsilicaa Silica

Octadecylsilica

Solute

ΔH (kJ/mol)

ΔS (J/mol K)

ΔH (kJ/mol)

ΔS (J/mol K)

Glycine Alanine Valine Leucine Phenylalanine Glutamic acid Glutamine Tryptophan

–2.22 –1.39 –3.12 –2.49 –6.03 –1.59 –2.15 –2.19

–11.5 –8.6 –11.7 –9.4 –21.3 –15.5 –10.9 –12.3

–8.66 –8.95 –10.65 –14.11 –19.68 –11.32 –10.22 –31.36

–34.6 –35.1 –35.6 –39.1 –48.0 –49.8 –38.9 –79.7

a

Experimental conditions: water mobile phase.

Source: Adapted from Basiuk, V.A. and Gromovoy, T.Y., Comparative study of amino acid adsorption on bare and octadecyl silica from water using high-performance liquid chromatography, Colloid Surface A, 118, 127, 1996.

relationships were then developed to accurately predict the eluotropic strength of the solvent mixture [109]. More recently, separations in the reversed-phase mode have been investigated on silica as well. Basiuk and Gromovoy [113,114] compared the adsorption behavior of amino acids on silica and octadecylsilica with a water mobile phase. The amino acids had an unfavorable thermodynamic transition from water to silica (ΔG > 0), but a favorable transition to octadecylsilica (ΔG < 0). Van’t Hoff plots were obtained over the temperature range from 20° to 60°C and selected thermodynamic values are summarized in Table 1.17. The changes in molar enthalpy and entropy were negative for both silica and octadecylsilica, but significantly more negative for octadecylsilica. The unfavorable entropic contribution (T ΔS) was greater than the favorable enthalpic contribution on silica, indicating an entropy-dominated retention process under all conditions. In contrast, the enthalpic contribution was greater for some but not all amino acids on octadecylsilica. Enthalpy-entropy compensation was observed on both silica and octadecylsilica for all amino acids except glutamic acid and aspartic acid, probably owing to their second carboxyl group. Bidlingmeyer and Henderson [115] separated some polar lipophilic amines on silica using 70% methanol-water with 6 mM potassium phosphate (pH 7) as the mobile phase. The amines were separated with high efficiency and excellent peak symmetry in the reversed-phase mode. However, the van’t Hoff plots were observed to be nonlinear over the temperature range from 25° to 80°C. The authors concluded that this effect is most likely due to a mixed retention mechanism from both adsorption and ion exchange (electrostatic forces). The ion exchange mechanism will be discussed in more detail in Section IV.C.

DK4478_book.fm Page 57 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

57

Snyder [116] performed kinetic studies of the adsorption mechanism on silica in the normal-phase mode by means of the plate height method. For neutral solutes, the plate height remained constant with the change in solute retention volume. This suggested that the adsorption/desorption rate was relatively fast and did not contribute significantly to the plate height. The rate-limiting step involved slow diffusion in adsorbent pores or within the aggregate particle. For basic solutes, however, interaction with the acidic silanol sites caused severe broadening and asymmetry of the peak and decreased the column efficiency. Hence, many kinetic investigations have been strongly driven by the interest in eliminating undesirable peak tailing. Ohkuma and Hara [117] analyzed the experimental zone profiles by assuming that the kinetic origin of the tailing arose from slow adsorption/desorption processes. They separated the chromatographic peak into a Gaussian portion and a “tail” portion. They found the area of the “tail” portion decreased and its relative length with reference to the elution time increased with increasing flow rate. These observations suggested that the tail portion of the peak is strongly influenced by the kinetic effect of slow processes. Fornstedt et al. [118,119] investigated peak tailing phenomena and the influence of mass transfer kinetics in both linear and nonlinear chromatography. In linear chromatography, the kinetic origin satisfactorily explains the peak tailing [118]. The two-site model, which assumes two different types of adsorption sites having different equilibrium isotherms and different rates of mass transfer kinetics, has been utilized. The most pronounced tailing existed when the slow type of adsorption site had a smaller contribution to the retention than the fast type and the rate constant for the slow sites was 20 to 2000 times smaller than that of the fast sites. In nonlinear chromatography, both thermodynamic and kinetic origins contribute to tailing [119]. In addition to the heterogeneous mass transfer kinetics discussed above, the heterogeneous thermodynamics with overloading of the nonlinear isotherm may affect the tailing as well, especially when the kinetics of the slow type of site is relatively fast (e.g., larger than 100 min–1). Under these conditions, heterogeneous thermodynamics become the dominating contribution to tailing. In summary, the kinetics for adsorption/desorption of neutral solutes on silica is fast, however, the kinetics for desorption of basic solutes is much slower due to contributions from heterogeneous kinetics and heterogeneous thermodynamics.

C. ION EXCHANGE Ion exchange chromatography is a well-established technique for the separation of ionic species, including inorganic anions/cations and low molar mass acids/bases such as carboxylic acids, phosphoric acids, sulfuric acids, and amines. There have been many studies concerned with improvement of the separation process by innovation in the stationary phase and changes in mobile phase composition, pH, and ionic strength. Yet there have been relatively few thermodynamic or kinetic studies of ion exchange because of the small and relatively unpredictable dependence of selectivity on temperature. When ion exchange is the only mechanism involved, the separation may be either an endothermic or exothermic process. The process depends on the charge, size, and type of the ion. For example, the change in molar enthalpy was negative

DK4478_book.fm Page 58 Thursday, September 14, 2006 1:35 PM

58

Advances in Chromatography, Volume 45

for singly charged ions (iodide, thiocyanate, perchlorate) that were strongly retained on an anion exchange resin, but positive for multiply charged ions (sulfate, oxalate, phosphate) [120]. The process also depends on the type and concentration of the mobile phase, as there is competition with the solutes for the ion exchange sites. For example, the change in molar enthalpy was negative for many divalent metal ions on a cation exchange resin using an aqueous mobile phase containing 0.12 M perchloric acid, but positive using a mobile phase containing 3.0 mM phenylenediamine [121]. In general, the retention process occurs with little evolution or consumption of heat, typically less than 8 kJ/mol [122]. Temperature has not gained much attention in ion exchange chromatography, however, it can be a great tool for adjusting retention and selectivity in chelated ion exchange chromatography. Thus, it is interesting to examine the thermodynamic contributions to retention in chelated ion exchange chromatography. When silica bonded with glutamic acid was used as the stationary phase, both ion exchange and chelation played important roles in the separation of alkali, alkaline earth, and transition metal cations [123]. At lower ionic strength and pH 2.7, the ion exchange mechanism dominated and the retention factor decreased with increasing temperature. Under these conditions, the sorption was exothermic, with changes in molar enthalpy ranging from –3.97 to –9.29 kJ/mol. At higher ionic strength and pH 4.7, where the ion exchange mechanism was suppressed, the retention factor increased with increasing temperature. This was indicative of the large entropy contribution due to chelation, where several solvent molecules were released for each metal cation that was bound. Under these conditions, the sorption was endothermic with changes in molar enthalpy ranging from 10.1 to 30.6 kJ/mol. Other examples showed a similar chelating effect [124,125], which indicated temperature is an important parameter in chelated ion exchange chromatography. The retention models of ion chromatography with secondary equilibria were reviewed by Janos [126], with emphasis on acid-base and complexation equilibria. Detailed aspects of the retention mechanism and accurate determination of equilibrium constants are discussed. Regarding the kinetics for ion exchange chromatography alone, the sorption/ desorption process is usually fast and diffusion controlled [122]. Generally speaking, intraparticle diffusion is the greatest contribution to broadening when there is large particle size, low diffusivity in the particle, and low concentration of ion exchange sites. On the contrary, film diffusion controls the broadening with small particle size and high concentration of sites. When chelation occurs, however, these general considerations do not apply and kinetics are much slower, owing to the desorption process [124]. Recently, the application of ion exchange chromatography in the field of recovery, concentration, and purification of proteins has become more prevalent. The ability to characterize the thermodynamic and kinetic parameters involved in the ion exchange interaction of proteins is essential in scaling up for preparative separations. Bovine serum albumin (BSA) has been successfully utilized as a model protein in ion exchange chromatography for thermodynamic and kinetic investigations. The sorption of proteins is a relatively complicated process. It can be exothermic or endothermic, depending on the protein, the surface of the stationary phase, and the mobile phase composition. An exothermic process indicates the dominating force

DK4478_book.fm Page 59 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

59

to be an attractive interaction between the surface and the adsorbing proteins, whereas an endothermic process indicates favorable entropy-dominated interactions. An exothermic process is generally expected in ion exchange chromatography when there is electrostatic interaction between the oppositely charged surface and the protein. However, some noteworthy exceptions have been reported. Bowen and Hughes [127] investigated BSA sorption on two polymeric anion exchange stationary phases with calorimetric measurements. An endothermic heat of sorption was observed, which indicated that the driving force was probably the large change in entropy due to solvent (water) release. Similarly, Thrash et al. [128] obtained negative van’t Hoff plots over the temperature range of 25° to 40°C that were dominated by large entropy contributions. They used the preferential interaction model to estimate water release from the contact surface of the BSA protein and the anionic polyethyleneimine resin. At 25°C, the water release alone was a sufficiently large entropic contribution to overcome the unfavorable change in enthalpy. However, at a higher temperature of 37°C, other effects such as structural rearrangement must also contribute to the net change in molar entropy. The same speculation about structural rearrangement was suggested by Raje and Pinto [129], who reported an endothermic interaction between BSA and anionic polyethyleneimine resin with primary, secondary, and tertiary amine ligands. Repulsive electrostatic interactions between the sorbed molecules might be another important contribution to the endothermic process. Thrash and Pinto [130] compared the enthalpy contribution from repulsive interactions and the entropy contribution from water release, concluding that the former is more important than any other contributions. Thus, the cations in the mobile phase had a significant influence on the heat of sorption by shielding repulsive interactions between negatively charged protein molecules. In summary, the unusual endothermic interactions can be attributed to repulsive interactions, protein reconfiguration, or surface dehydration, depending upon the experimental conditions. The kinetics of protein sorption has gained attention recently because it has been realized that the poor separation efficiency of large biomolecules is due to their slow mass transfer kinetics. Weaver and Carta [131] investigated different types of resins for their equilibrium and kinetic properties using lysozyme as a model protein in batch experiments. These resins included a macroporous PS-DVB polymer (POROS 50 HS; Perceptive Biosystems, Inc., Boston, MA) and a composite material prepared by filling the pores of high-porosity polystyrene-coated silica particles with a polyacrylamide-based hydrogel (BioSepra S–HyperD M; Pall Corporation, East Hills, NY). While both sorbents could tolerate high linear velocity without compression or large pressure drop, they differed in their sorption capacity and kinetics. The gelcomposite sorbent had a 60% higher capacity than the macroporous sorbent because of their different means to reach loading capacity. The gel-composite sorbent was dependent upon the volume of hydrogel within the pores, while the macroporous sorbent was dependent upon the surface area. In their kinetic behavior, external film resistance played an important role at low protein concentrations for both sorbents. However, at higher concentrations, different models were utilized to describe the different sorbents. The pore diffusion model described the kinetics for the macroporous sorbent, with the concentration gradient in the pore fluid as the driving force and a pore diffusion coefficient of 1.1 × 10–7 cm2/s. The homogeneous diffusion

DK4478_book.fm Page 60 Thursday, September 14, 2006 1:35 PM

60

Advances in Chromatography, Volume 45

model described the kinetics for the gel-composite media, with the concentration gradient within the gel medium as the driving force and an apparent diffusion coefficient of 7.5 × 10–9 cm2/s. The authors concluded that the gel-composite sorbent was superior to the macroporous sorbent because of its higher dynamic capacity. Fernandez and Carta [132] and Fernandez et al. [133] also did similar research with BSA protein on the gel-composite HyperD resins using the batch and shallow bed techniques. In both cases, intraparticle mass transfer dominated at higher protein concentrations, but film resistance dominated at lower protein concentrations. The intraparticle mass transfer was independent of flow rate and independent of hydrodynamic conditions outside of the particles, which is consistent with a diffusionlimited transport mechanism. Heeter and Liapis [134] studied the sorption of BSA on a porous strong anion exchange resin using the frontal analysis method. Among the models examined, the dynamic nonlinear adsorption model for purely diffusive particles with a monodisperse porous structure was found to best describe the system behavior. The intraparticle diffusion coefficient was estimated to be 4.11 × 10–7 to 2.00 × 10–7 cm2/s for BSA concentrations of 1 × 10–5 to 5 × 10–5 g/cm3. The data were also evaluated to determine the contribution from intraparticle convection, which was found to be negligible. Conder and Hayek [135] came to similar conclusions about the dominant effect of intraparticle diffusion kinetics for BSA on polyethyleneimine, a weak anion exchange resin. Finally, Miyabe and Guiochon [136] provided more detail on the mass transfer kinetics for BSA on anion exchange resins. The lumped kinetic model showed a linear increase in the intraparticle mass transfer rate coefficient with BSA concentration. Further analysis showed that this dependence should be attributed to the concentration dependence of the surface diffusivity (Ds) and not the pore diffusivity (Dp). In summary, ion exchange chromatography of small ions is usually associated with small changes in molar enthalpy and entropy. The kinetic processes are usually relatively fast. However, when combined with chelating interactions, the kinetic rate decreases owing to slow desorption. Protein separations by ion exchange chromatography can be endothermic or exothermic processes. The kinetics are mainly limited by intraparticle diffusion, specifically surface diffusion, especially at high protein concentrations.

D. SIZE EXCLUSION Size exclusion chromatography, also known as gel filtration or gel permeation chromatography in aqueous or organic solvents, respectively, is utilized for the separation of industrial polymers, biopolymers, and other macromolecules as well as for the determination of molar mass distributions. Separation is achieved by the differential inclusion or exclusion of solutes as they pass through a column of porous particles. Solutes that are smaller than the pore diameter have a higher probability of entering the particles and, therefore, have a longer retention time than larger solutes that cannot enter the particles. From a thermodynamic perspective, size exclusion chromatography is distinguished from all other separation mechanisms. The interactions between the solute and the stationary phase surface are intentionally minimized or eliminated, such

DK4478_book.fm Page 61 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

61

that the changes in molar enthalpy are negligible. However, when solute molecules enter the pores, a number of solvent molecules are displaced which is accompanied by a change in the entropy of mixing. In addition, the mobility of the solute molecule may become more limited inside the pores, thereby decreasing the configurational entropy. Thus, size exclusion chromatography is thermodynamically associated with changes in entropy alone [137]. These entropy changes result from the concentration gradients as well as the flow and diffusion processes within the column, which lead to partial or full exclusion of macromolecules from the pores or other restricted regions. With a few noteworthy exceptions, the bulk of the theoretical and experimental evaluations of these phenomena were completed more than 30 years ago [138–142]. More recently, interest has focused on a technique known as liquid chromatography at the critical condition (LCCC). In this technique, separations are performed at the interface between the entropy-dominated size exclusion mechanism and the enthalpy-dominated adsorption mechanism. The entropic and enthalpic contributions are exactly balanced for a selected group of macromolecules (e.g., linear polystyrenes), such that enthalpy-entropy compensation is achieved. All molecules of the selected group are eluted at the void volume of the column, irrespective of their molar mass. Molecules that have slight differences in structure from the selected group (e.g., cyclic polystyrenes, PS-DVB copolymers, etc.) will have a different balance of entropic and enthalpic contributions and, hence, will be eluted before or after the void volume. The thermodynamic theoretical models of LCCC have been elucidated and reviewed by Gorbunov and Skvorcov [143]. The critical conditions can be tuned by adjusting the stationary phase composition, the mobile phase composition, or temperature, as summarized in an excellent review by Macko and Hunkeler [144]. Furthermore, at least one report demonstrates that pressure can also affect the critical conditions [145]. Among the papers searching for critical conditions, very few carried out systematic thermodynamic studies. Trathnigg et al. [146] performed a systematic study of polyoxyethylenes, including polyethylene glycols, fatty alcohol ethoxylates, and fatty acid polyglycol esters. For the oxyethylene unit in polyethylene glycol homopolymers, the critical conditions were realized on an octadecylsilica stationary phase with mobile phases of 85.8% methanol-water or 89.6% acetone-water. For both mobile phase systems, the van’t Hoff plots were linear over the temperature range from 15° to 35°C. The resulting changes in molar entropy and enthalpy were very small in the methanol-water system, but nearly tenfold larger in the acetone-water system. In the diblock and triblock polymers with hydrophobic alkoxy or ester groups, no compensation points were found. Philipsen et al. [147] performed thermodynamic studies on low molar mass polystyrenes and polyesters on an octadecylsilica stationary phase with tetrahydrofuran-water mobile phases. Linear van’t Hoff plots were obtained in almost all cases. The determined changes in molar entropy and enthalpy were negative for both polymers and became less negative with increasing concentration of tetrahydrofuran. They attributed the increase in enthalpy to the decreased affinity between the polymer and the stationary phase and the increase in entropy to the decreased number of polymer sites that interact simultaneously with the sorbent. Enthalpy-entropy compensation was identified for polystyrene oligomers in mobile phases containing different concentrations

DK4478_book.fm Page 62 Thursday, September 14, 2006 1:35 PM

62

Advances in Chromatography, Volume 45

of tetrahydrofuran. This indicated the same retention mechanism regardless of mobile phase composition. However, no enthalpy-entropy compensation temperature was identified for polystyrenes or polyesters at any mobile phase composition. Consequently, retention was not exactly independent of molar mass under the selected critical conditions. Kinetic studies of size exclusion chromatography have been relatively limited because the experimental zone profiles contain significant contributions from sample polydispersity and extracolumn dispersion. In general, sample polydispersity tends to dominate the observed broadening, making it difficult to extract the kinetic contributions. Investigations have shown that axial dispersion from multiple paths and slow mass transfer of the polymer within the stationary phase are the most important contributions to kinetic broadening [148]. Leypoldt et al. [149] noted that contributions from slow mass transfer are dependent on flow rate, whereas contributions from multiple paths and sample polydispersity are not. By judicious separation of these contributions, they demonstrated that restricted diffusion within the stationary phase depended upon the ratio of the radii of the solute molecule and the pore. In summary, size exclusion chromatography is traditionally treated as an entropydominated process and, thus, the solutes are separated based on their molar mass and size. Recently, the combination of entropy-dominated size exclusion with enthalpy-dominated adsorption has been used to facilitate group-specific separations that are independent of their mass and size. The kinetics aspects of size exclusion separations are often small compared to the effects of sample polydispersity.

E. CHIRAL Enantiomers are molecules that are mirror images of each other and nonsuperimposable. Because they have the same physical and chemical properties, enantiomers are very difficult to separate. As a consequence, chiral separations represent one of the most complicated classes of separation techniques. Chiral stationary phases (CSPs) may utilize multiple separation mechanisms, such as a combination of adsorption with ion exchange or complexation with size exclusion. These CSPs have been categorized according to the following types: brush (e.g., Pirkle-type), inclusion (e.g., cyclodextrin), polysaccharide (e.g., cellulose), and affinity (e.g., protein). In this section, the thermodynamic and kinetic properties of these different types of CSPs will be reviewed. 1. Brush CSPs Brush CSPs distinguish themselves from other CSPs by having a small chiral molecule covalently bonded to the solid phase, usually silica. The organic groups of the chiral molecule remain directed away from the silica surface, appearing as bristles of a brush. Among the brush CSPs, the most widely known and commercially successful are those developed by Pirkle and coworkers. Most of the Pirkle-type CSPs emphasize electron donor-acceptor interactions of the nonbonding (n) and aromatic (π) types. Chiral separations are achieved by establishing three points of interaction between an enantiomer and the CSP, where at least one of these inter-

DK4478_book.fm Page 63 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

63

actions is stereospecific [150]. Consequently, the other enantiomer is only able to sustain two points of interaction with the CSP and is less retained. Most separations with Pirkle-type CSPs are performed in the normal-phase mode. Although there have been few systematic thermodynamic studies of Pirkle-type CSPs, temperature has been successfully utilized as an optimization tool. Mazzo et al. [151] optimized the chiral separation of an antihypertensive drug, MK-286, with an N-(3,5-dinitrobenzoyl)phenylglycine CSP. The retention factors and enantioselectivity increased with decreasing temperature, however, the broadening (plate height) also increased. Because of these concomitant changes, the resolution increased from 0.75 at room temperature to 1.25 at 0°C. Although no thermodynamic parameters were reported, it was probably an enthalpy-controlled process that facilitated separation at a lower temperature. Pescher et al. [152] optimized the separation of tertiary phosphine oxides with N-(3,5-dinitrobenzoyl) derivatives of phenylglycine, serine, and alanine as CSPs. They observed a similar effect of temperature, where the enantioselectivity with the phenylglycine CSP increased from 1.52 to 2.33 with a decrease in temperature from 50° to –15°C. A nonlinear dependence of enantioselectivity on temperature was observed by Weaner and Hoerr [153]. They investigated a series of fatty acid esters and amide epoxides on an N-(3,5-dinitrobenzoyl)phenylglycine CSP. While the enantioselectivity of the ester decreased rapidly above 0°C, those of the amide reached a maximum at 10°C and remained constant up to 40°C. This behavior was attributed to the weak interaction forces between the ester and CSP and the relatively strong forces between the amide and CSP. Pirkle also observed some unusual temperature-dependent behavior [154]. In the separation of spirolactam enantiomers on an N-(3,5-dinitrobenzoyl)phenylglycine CSP, a nonlinear van’t Hoff plot was obtained. The curvature of these plots was dependent on the concentration of 2-propanol in the mobile phase, which indicated a temperaturedependent interaction of 2-propanol with the stationary phase or with the solute. However, the natural logarithm of enantioselectivity versus the inverse temperature gave a linear relationship, which indicated that this behavior affected both enantiomers equally and had little influence on the chiral recognition mechanism. Another example with the separation of N-(3,5-dinitrobenzoyl)-1-phenylethylamine enantiomers showed an initial decrease in enantioselectivity with a reduction in temperature, followed by an inversion of elution order and a subsequent increase in enantioselectivity [155]. The observed temperature effects showed a dependence on the type and concentration of organic modifier in the mobile phase. Pirkle and Welch [156] systematically investigated the role of solvation on the separation of N-(2-naphthyl)alanine derivatives on an N-(3,5-dinitrobenzoyl)leucine CSP. In general, both retention and enantioselectivity decreased with increasing mobile phase polarity in both normal- and reversed-phase modes. The van’t Hoff plots, obtained over a temperature range of 0° to 90°C, were used to evaluate the thermodynamic changes in molar enthalpy and entropy as a function of mobile phase composition. Under all conditions, the changes in molar enthalpy and entropy were negative, as were the differential changes (ΔΔH and ΔΔS) for each pair of enantiomers. The enthalpy and entropy changed in a similar manner with the polarity of the mobile phase, which indicated that the highly exothermic sorption was accompanied by a correspondingly large loss in entropy. The authors noted that this

DK4478_book.fm Page 64 Thursday, September 14, 2006 1:35 PM

64

Advances in Chromatography, Volume 45

behavior is analogous to enthalpy-entropy compensation, as discussed previously. The steric effects of the organic modifier were further investigated by Blackwell et al. [157]. Their detailed study used linear, branched, and cyclic alcohols, ranging in size from ethanol to octanol, in the normal-phase mode with amide derivatives of 1-phenylethylamine as model solutes. As shown in Table 1.18, the enantioselectivity varied slightly (1.15 to 1.35) as a function of the modifier structure for an N-(3,5dinitrobenzoyl)phenylglycine CSP. The enantioselectivity was higher and varied more (1.53 to 2.44) with modifier structure for an N-(3,5-dinitrobenzoyl)-4-amino1,2,3,4-tetrahydrophenanthrene CSP (Whelk-O; Regis Technologies, Inc., Morton Grove, IL). For both CSPs, the lowest selectivities were observed for 1-hexanol and cyclohexanol. The highest selectivities were observed for bulky modifiers, such as 2-propanol, 2-butanol, and 2-methyl-2-propanol. The highest selectivities were the result of a dramatic increase in retention of the second enantiomer. Based on detailed thermodynamic measurements, the authors suggested that the bulky modifiers cannot penetrate into the tight complex formed between the second enantiomer and the CSP. They also suggested that chiral separations giving large enantioselectivity (α > 2), such as those with the Whelk-O CSP, were not consistent with the classic three-point interaction model of Dalgliesh [150]. These separations exhibited a wide range of selectivities, were not well described by linear solvation energy relationships, and did not exhibit consistent enthalpy-entropy compensation behavior with varying modifier structure. Thermodynamic studies can also be used to examine the nature of the surface for covalently bonded CSPs. Pirkle and Readnour [158] designed a chromatographic system utilizing π-basic CSPs with a series of π-acidic solutes, bis-(2,4-dinitrophenyl)-α,ω-diaminoalkanes (bis-DNPs). The number of methylene groups in the bisDNPs was varied in order to influence their ability to interact with the CSP. The enthalpy of sorption, determined from van’t Hoff plots, was expected to be maximized (most negative) when the length of the bis-DNPs matched the distance between strands of the stationary phase. By using this approach, the authors found the maximum enthalpy when the number of methylene groups was equal to five. For CSPs having different surface coverage (0.36 to 1.36 μmol/m2), the maximum enthalpy occurred with the same number of methylene groups, but with different maximum enthalpy values. These interesting results suggested that the strands were not randomly distributed on the silica surface, but clustered with similar distributions of interstrand distances. The surface coverage thus affected the size of the cluster rather than the average distance between clusters. 2. Inclusion CSPs The most common examples of inclusion CSPs are the native and derivatized cyclodextrins. Cyclodextrins are cyclic oligosaccharides containing six or more Dglucose units connected through α-1,4-glycosidic bonds. The structure of β-cyclodextrin contains seven glucose units with 35 chiral centers, while α- and γ-cyclodextrins contain six and eight glucose units, respectively. In aqueous media, the hydroxyl groups are oriented toward the exterior surface, which makes the interior

–0.343 –0.445 –0.605 –0.486 –0.617 –0.433 –0.509 –0.391 –0.195 –0.195 –0.105 –0.505

ΔΔH (kcal/mol) –0.70 –0.97 –1.41 –1.11 –1.48 –0.96 –1.16 –0.86 –0.31 –0.32 –0.06 –1.37

ΔΔS (cal/mol K) 488 459 430 437 416 454 438 456 627 617 1814 369

Tiso (K) 2.01 2.06 2.44 1.92 2.28 1.89 2.25 1.78 1.71 1.67 1.53 1.60

α –0.891 –0.908 –1.242 –0.742 –1.087 –0.670 –1.027 –0.649 –0.604 –0.452 –0.265 –0.589

ΔΔH (kcal/mol) –1.56 –1.56 –2.32 –1.16 –1.96 –0.95 –1.77 –1.00 –0.93 –0.47 –0.02 –1.00

ΔΔS (cal/mol K)

573 581 534 638 555 706 578 650 651 957 13980 587

Tiso (K)

N-(3,5-dinitrobenzoyl)-4-amino-1,2,3,4-tetrahydrophenanthrene

Source: Adapted from Blackwell, J.A., Waltermire, R.E., and Stringham, R.W., The role of modifiers in high selectivity chiral separations on Pirkle-type columns, Enantiomer, 6, 353, 2001.

Experimental conditions: 20% alcohol–heptane mobile phase, 30°C.

1.25 1.29 1.35 1.27 1.32 1.27 1.30 1.24 1.18 1.18 1.15 1.16

Ethanol 1-Propanol 2-Propanol 1-Butanol 2-Butanol 2-Methyl-1-propanol 2-Methyl-2-propanol 1-Pentanol 3-Methyl-1-butanol 1-Hexanol Cyclohexanol 1-Octanol

a

α

Modifier

N-(3,5-dinitrobenzoyl)phenylglycine

TABLE 1.18 Enantioselectivity (α), Differential Changes in Molar Enthalpy (ΔΔH) and Entropy (ΔΔS), and Isoenantioselective Temperature (Tiso) for 1-Phenylethylacetamide Enantiomers Separated on Pirkle-Type CSPs with Various Alcohol Modifiersa

DK4478_book.fm Page 65 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography 65

DK4478_book.fm Page 66 Thursday, September 14, 2006 1:35 PM

66

Advances in Chromatography, Volume 45

of the cavity relatively nonpolar and hydrophobic. Consequently, solutes may enter the cavity and be retained by dispersion and hydrogen bonding interactions. Cyclodextrins were first immobilized to solid supports in 1978 [159]. Since then, extensive work has been done to improve the stability and versatility of cyclodextrin silica CSPs, which made possible their commercialization in the 1980s. However, despite the large number of publications, only limited information is available on the thermodynamic characterization of these CSPs, and even less information is available on the kinetic characterization. Both thermodynamic and kinetic information are important for elucidation of the retention mechanism. The earliest studies of temperature effects were focused on optimization of the separation process [160,161]. In these studies, lower temperature was shown to increase retention, due to stronger interactions between the solutes and the CSP, and also to increase enantioselectivity. A more detailed study of temperature was carried out by Cabrera and Lubda [162]. They separated two chiral pharmaceuticals, oxazepam and mephobarbital (Prominal), on native β-cyclodextrin silica operated in the reversed-phase mode. A decrease in temperature caused an increase in retention factor for both solutes, however, it caused different effects on their enantioselectivity. The enantioselectivity was increased by reducing the temperature for oxazepam, but by increasing the temperature for mephobarbital. The van’t Hoff plots were linear for both solutes over the temperature range of 5° to 40°C. From the thermodynamic data, the authors concluded that the differences in enantioselectivity were related to the enthalpy-controlled separation of oxazepam enantiomers (ΔΔH = –332 cal/mol, ΔΔS = 55 cal/mol) and the entropy-controlled separation of mephobarbital enantiomers (ΔΔH = 374 cal/mol, ΔΔS = 419 cal/mol). Morin et al. [163] also studied the effect of temperature on β-cyclodextrin silica with a series of six imidazole derivatives. Over the temperature range of 20° to 55°C, the van’t Hoff plots for all solutes were linear at pH 7.0 and 7.5, but curved at pH 6.5, 8.0, and 8.5. The curved plots showed a minimum in the logarithm of the retention factor between 35° and 40°C. The changes in molar enthalpy and entropy were determined from the van’t Hoff plots. An enthalpyentropy compensation analysis revealed that the retention mechanism was the same for all solutes at pH 7.0 and 7.5, and also the same for all solutes at pH 6.5, 8.0, and 8.5. From thermogravimetric analysis and differential scanning calorimetry, the authors speculated that a phase transition occurred at 43°C between the ordered and disordered state of the cyclodextrin cavity at pH 6.5, 8.0, and 8.5, but not at pH 7.0 and 7.5. In the disordered state, a gain in freedom of the hydroxyl groups on each edge of the cyclodextrin cavity was obtained with the minimum extent of hydrogen bond formation between them. In contrast, in the ordered state, a loss in freedom of the hydroxyl groups was obtained with the maximum extent of hydrogen bond formation. As a result, the hydrophobic character, surface tension, and structure of the cyclodextrin cavity were all affected, thereby altering the retention mechanism. Peyrin et al. [164] investigated the inclusion phenomena between β-cyclodextrin silica CSP and a series of dansyl amino acids with the salting-out agent sucrose. When the solute transfers from the mobile to stationary phase, sucrose is displaced from the cyclodextrin cavity. With a simple physicochemical model, the number of displaced sucrose molecules can be determined and the relative extent of solute inclusion can thus be calculated. The authors concluded that the L-enantiomers of

DK4478_book.fm Page 67 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

67

the dansyl amino acids displaced a constant number of 1.9 sucrose molecules regardless of the R group. In contrast, the D-enantiomers displaced a larger number of sucrose molecules that decreased from 2.7 to 2.2 as the accessible surface area of the R group increased. Hence, the chiral recognition and enantioselectivity decreased with increasing size of the R group. The theoretical value of the accessible surface area that corresponded to no chiral resolution was 191 Å2. It is noteworthy that the equivalent length of the R group (7.8 Å) is equal to the internal diameter of the native β-cyclodextrin cavity. The authors concluded that the inclusion process was largely dependent on steric hindrance. Temperature was also varied from 20° to 45°C to observe the thermodynamic effects on enantioselectivity. The negative ΔΔH and ΔΔS values were determined from the linear van’t Hoff plots, which indicated an enthalpy-controlled separation for all dansyl amino acids. The ΔΔH and ΔΔS values became less negative with increasing accessible surface area of the R group and more negative with increasing sucrose concentration. A similar approach was utilized by Guillaume et al. [165] using water as a surface tension modifier in the mobile phase. In this case, water was displaced when the dansyl amino acids were transferred to the β-cyclodextrin cavity. Similar results were concluded on the relative degrees of inclusion of the D- and L-enantiomers and the importance of steric effects in the chiral recognition mechanism. The effect of pressure on the separation process with a β-cyclodextrin silica CSP was investigated by Ringo and Evans [166]. With the positional isomers of nitrophenol as model solutes, a shift in retention with modest pressure change (300 bar) was observed and correlated to the change in partial molar volume for the complexation process. The same method was applied to chiral separations with barbiturates, β-blockers, and anticoagulants as model solutes [167–169]. Most chiral compounds showed a dependence of the retention factor, enantioselectivity, and separation efficiency on pressure. The retention factor dependence reflected the statistical nonzero changes of molar volume for the complexation processes. As illustrated in Table 1.19, the retention factors exhibited an increase or no change with pressure, and the change in molar volume was negative or negligible in the reversed-phase mode. However, in the polar-organic mode, the retention factors showed a decrease with pressure, and the change in molar volume was positive or negligible. The change in molar volume ranged from –12 to 17 cm3/mol. The enantioselectivity dependence reflected the unequal changes in molar volume between the enantiomers. Mephobarbital, metoprolol, and warfarin demonstrated a dependence of selectivity on pressure, with both positive and negative dependence observed. These pressure experiments provided insight into the differential volume of enantiomeric complexes formed with β-cyclodextrin. Recently, Wang and Ching [170] and Yu and Ching [171] investigated both the thermodynamic and kinetic behavior of derivatized β-cyclodextrin silica CSPs. The equilibrium and kinetic constants for the chiral separation were obtained by moment analysis on the basis of the equilibrium-dispersive and transport-dispersive (solidfilm linear driving force) models. The overall mass transfer coefficients, axial dispersion coefficients, and equilibrium constants were used to simulate the enantiomeric band profiles. Excellent correlation between simulated and experimental results confirmed the accuracy of the method as well as the results. The overall mass

DK4478_book.fm Page 68 Thursday, September 14, 2006 1:35 PM

68

Advances in Chromatography, Volume 45

TABLE 1.19 Effect of Pressure on Retention Factor (k) and Change in Molar Volume (ΔV) on a βCyclodextrin Silica Phase Operated in the Reversed-Phase Modea Δk/k (%) R-

Solute Hexobarbital Mephobarbital Ibuprofen Chlorthalidone Benzoin

12 11 2 1 –2

± ± ± ± ±

ΔV (cm3/mol)

S3 2 5 18 11

11 10 2 2 –2

± ± ± ± ±

2 3 4 18 8

R-

S-

–12 –11 –3 –2 0

–12 –10 –4 –3 +1

a

Experimental conditions: methanol-water–acetic acid– triethylamine mobile phase, 295 K. Source: Adapted from Ringo, M.C. and Evans, C.E., Pressure-induced changes in chiral separations in liquid chromatography, Anal. Chem., 69, 4964, 1997.

transfer coefficients were determined to be 58 min–1 and 73 min–1 for S- and Rfluoxetine; respectively, and 669 min–1, 846 min–1, and 106 min–1 for the coeluted SRS- and SSR-nadolols, RRS-nadolol, and RSR-nadolol, respectively. These kinetic data indicate that mass transfer processes in the derivatized β-cyclodextrin silica CSPs are relatively rapid. 3. Polysaccharide CSPs Polysaccharides such as amylose and cellulose are naturally occurring, optically active polymers. Amylose is composed of glucose units connected with α-1,4glycosidic bonds and adopts a helical structure in its native form. Cellulose is based on β-1,4-glycosidic bonds and adopts a more linear structure. Native polysaccharides have long been known to exhibit chiral selectivity [172], however, their chiral resolving abilities are not sufficient for practical usage. Poor resolution and broad peaks are obtained due to nonstereoselective binding and slow mass transfer processes, especially diffusion through the polymer network. In comparison, derivatized polysaccharide CSPs have much better chromatographic and enantiomeric properties, which has led to their successful commercialization. The chiral recognition mechanism for polysaccharide CSPs is generally thought to consist of cohesive attraction followed by steric fit of the chiral solute into the chiral surface. The structural differences between amylose and cellulose are responsible for the difference in their retention mechanisms. As a result, the thermodynamics and kinetics associated with the separation process are different as well. The temperature dependence of chiral discrimination was investigated by Smith et al. [173]. Two analogues of cromakalim, a potassium channel activator, were

DK4478_book.fm Page 69 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

69

separated on cellulose derivatized with tris(3,5-dimethylphenylcarbamate) (Chiralcel-OD; Daicel Chemical Industries, Ltd., Tokyo, Japan) over the temperature range from 0° to 42°C. Although the two analogs differ only by replacement of a benzoyl group with an n-pentanoyl group, they showed different temperature dependence. The benzoyl enantiomers exhibited an increasing trend in resolution with increasing temperature, whereas the n-pentanoyl enantiomers had a decreasing trend. The van’t Hoff plots were linear for all enantiomers, from which the changes in molar enthalpy and entropy were determined. A large difference was observed for ΔΔH (1.929 and 4.265 kJ/mol for the benzoyl and n-pentanoyl enantiomers, respectively), but relatively small difference for ΔΔS (10.3 and 13.6 J/mol K, respectively). The isoenantioselective temperature, which is analogous to the enthalpy-entropy compensation temperature, also differed greatly (–86°C and 41°C, respectively). Hence, the benzoyl enantiomers exhibited an entropy-dominated separation, whereas the n-pentanoyl enantiomers had an enthalpy-dominated separation. The authors concluded that increased temperature should be beneficial for any chiral separation that predominantly involves π-type electron donor-acceptor interactions and, conversely, that decreased temperature should be beneficial for separations that predominantly involve hydrogen bonding interactions. This general conclusion was confirmed by Kuesters and Spoendlin [174] for the separation of rolipram and a series of structurally related compounds on the same Chiralcel-OD CSP over the temperature range from 10° to 60°C. These aromatic compounds showed an unusual entropydominated separation with ΔΔH of 1.62 kJ/mol and ΔΔS of 5.72 J/mol K for rolipram, similar to the entropy-dominated separation of the benzoyl analog discussed above. More enthalpy-dominated separations were observed by Mesplet et al. [175] for some phosphoramidate derivatives of anti-human immunodeficiency virus (HIV) nucleosides. A detailed mechanistic investigation was undertaken by O’Brien et al. [176] with a diol intermediate for a leukotriene D4 antagonist separated on cellulose derivatized with tris(4-methylbenzoate) (Chiralcel-OJ; Daicel Chemical Industries, Ltd., Tokyo, Japan). A nonlinear van’t Hoff plot was observed over the temperature range of 5° to 50°C for both retention factor and selectivity, as shown in Figure 1.4. At low temperature (region II), the enantioselectivity was entropy controlled with positive values for ΔΔH and ΔΔS, but at high temperature (region I), it was enthalpy-controlled with negative values for ΔΔH and ΔΔS. This temperature dependence was very unusual. In general, the entropy contribution (T ΔΔS) should tend to increase with increasing temperature and overcome the enthalpy contribution (ΔΔH), thereby changing the separation mechanism from enthalpy-dominated to entropy-dominated with increasing temperature. The unusual temperature dependence observed by O’Brien et al. [176] suggested a phase transition in the CSP. The transition observed in the van’t Hoff plots occurred at a temperature of approximately 18°C, where a conformational change in the stationary phase was confirmed by infrared spectroscopy and differential scanning calorimetry. The thermodynamic behavior was also examined for different mobile phase compositions in the normal-phase mode. With an increase in concentration of the organic modifier, 2-propanol, an increase in ΔΔH was observed in both regions I and II, accompanied by an increase in ΔΔS. These two effects canceled each other by means of enthalpy-entropy compensation and

DK4478_book.fm Page 70 Thursday, September 14, 2006 1:35 PM

70

Advances in Chromatography, Volume 45

10

Retention factor

R enantiomer Region II R enantiomer Region I

S enantiomer Region II S enantiomer Region I

1 0.0030

0.0031

0.0032

0.0033

0.0034

0.0035

0.0036

−1)

1/temperature (K (a) 1.8

Region II

Selectivity

1.7

Region I 1.6

1.5 0.0030

0.0031

0.0032 0.0033 0.0034 1/temperature (K−1)

0.0035

0.0036

(b)

FIGURE 1.4 Van’t Hoff plots of (a) retention factor (k) and (b) enantioselectivity (α) for the diol enantiomers on the Chiralcel-OJ column (25 cm × 4.6 mm inner diameter). Mobile phase: 40% 2-propanol-hexane. (Adapted from O’Brien, T., Crocker, L., Thompson, R., Thompson, K., Toma, P.H., Conlon, D.A., Feibush, B., Moeder, C., Bicker, G., and Grinberg, N., Mechanistic aspects of chiral discrimination on modified cellulose, Anal. Chem., 69, 1999, 1997.)

DK4478_book.fm Page 71 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

71

resulted in a negligible change in molar free energy. This explains why an increase in the concentration of 2-propanol had virtually no effect on enantioselectivity. The authors suggested that the loss of interaction energy for the more retained R-enantiomer relative to the S-enantiomer with increasing 2-propanol concentration was balanced by a relative increase in the space available for the R-enantiomer when it entered the stationary phase. This was indicative of a swelling of the cellulose CSP with increasing 2-propanol concentration. The kinetic behavior was also evaluated by means of the plate height method. For the more retained R-enantiomer, the reduced plate height was large at temperatures up to 10°C, but showed a sharp decrease above 15°C. However, for the S-enantiomer, the reduced plate height showed a gradual decrease over the entire temperature range of 5° to 50°C. These data suggest that slow mass transfer processes occurred for the R-enantiomer at lower temperatures. This behavior was attributed to the relatively slow inclusion of the R-enantiomer at low temperature, where the cellulose chains are rigid. In closely related studies, Wang et al. [177–179] reported some unusual temperature effects with polysaccharide CSPs. They separated the enantiomers of dihydropyrimidinone (DHP) acid and its methyl ester on both amylose and cellulose derivatized with tris(3,5-dimethylphenylcarbamate) (Chiralpak-AD and ChiralcelOD; Daicel Chemical Industries, Ltd., Tokyo, Japan). A nonlinear van’t Hoff plot was observed for R- and S-DHP acid on Chiralpak-AD, but a linear relationship was observed on Chiralcel-OD. Moreover, the van’t Hoff plots obtained with Chiralpak-AD upon cyclic heating and cooling between 5° and 50°C were not superimposable, as shown in Figure 1.5. This suggested an irreversible, thermally induced conformational change of Chiralpak-AD at approximately 30°C. Solid-state NMR was utilized to verify and to characterize this conformation change. Booth and Wainer [180] investigated the separation of mexiletine and a series of structurally related solutes on Chiralpak-AD CSP. Although a nonlinear van’t Hoff plot was observed for one solute (which was subsequently eliminated), linear van’t Hoff plots were observed for all others over the temperature range from 0° to 30°C. The change in molar enthalpy was, thus, constant and negative for these solutes. An enthalpyentropy compensation plot of ln k vs. –ΔH was then constructed. The data for these solutes fell into two distinct groups, with linearity observed within each group. This linearity was indicative of an enthalpy-entropy compensation phenomenon within that group of solutes. The two groups had different compensation temperatures and, hence, different retention mechanisms. It was interesting to note that one group contained the solutes with hydroxyl substituents, while the other group contained all other solutes. The frontal analysis method has been utilized by Rearden et al. [181] and SeidelMorgenstern and Guiochon [182] to elucidate thermodynamics and mass transfer kinetics. The sorption isotherms of Troeger’s base were determined on a microcrystalline cellulose triacetate CSP over a temperature range of 30° to 60°C. Whereas the S-enantiomer exhibited a Langmuir isotherm, the R-enantiomer could best be described by a quadratic equation for an S-shaped isotherm. The sorption enthalpies of both enantiomers increased with increasing temperature. The mass transfer kinetics for the S-enantiomer were studied by two different models, the equilibriumdispersive and transport-dispersive (solid-film linear driving force) models. The two

DK4478_book.fm Page 72 Thursday, September 14, 2006 1:35 PM

72

Advances in Chromatography, Volume 45

Retention factor

100

10

1 0.0030 0.0031 0.0032 0.0033 0.0034 0.0035 0.0036 0.0037 1/temperature (K−1) (a)

Selectivity

10

1 0.0030 0.0031 0.0032 0.0033 0.0034 0.0035 0.0036 0.0037 1/temperature (K−1) (b)

FIGURE 1.5 Van’t Hoff plots of (a) retention factor (k) and (b) enantioselectivity (α) for DHP acid on the Chiralcel-AD column (25 cm × 4.6 mm inner diameter). Mobile phase: 15% ethanol-hexane with 0.1% trifluoroacetic acid. Methods: (a) R-(–)-DHP acid, heating (), cooling (); S-(+)-DHP acid, heating (), cooling (). (b) heating (), cooling (). (Adapted from Wang, F., O’Brien, T., Dowling, T., Bicker, G., and Wyvratt, J., Unusual effect of column temperature on chromatographic enantioseparation of dihydropyrimidinone acid and methyl ester on amylose chiral stationary phase, J. Chromatogr. A, 958, 69, 2002.)

DK4478_book.fm Page 73 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

73

approaches gave almost identical results. The mass transfer rate constant was determined to be around 7 to 11 min–1 for concentration steps from 0 to 3.2 g/L of Troeger’s base. 4. Affinity CSPs Proteins are naturally occurring, optically active polymers composed of L-amino acids. As a result, immobilized proteins on solid supports are capable of discriminating enantiomers. Protein-based CSPs are of special interest because of their unique enantioselective properties and wide range of applications. Chromatographic methods are distinguished on the basis of whether the proteins are soluble or immobilized, where the latter is known as affinity chromatography. The most commonly used proteins are as follows: (1) albumins, such as human serum albumin (HSA) and BSA; (2) glycoproteins, such as α1-acid glycoprotein, ovomucoid, and avidin; (3) enzymes, such as cellobiohydrolase I, lysozyme, and pepsin; and (4) other miscellaneous proteins, such as β-lactoglobulin. Although protein CSPs are widely used, the elucidation of their retention mechanism is relatively difficult due to their complex structures. These complex structures are the result of different amino acid sequences (primary structure) as well as different disulfide bridges, hydrogen bonding, and other types of intramolecular bonding (secondary and tertiary structure). Because of their structure, protein CSPs usually rely on different numbers or types of binding sites for discrimination of enantiomers. To retain the protein structure and avoid denaturation, most separations are performed in the reversed-phase mode. Both elution and frontal analysis methods have been used to characterize the binding constants, as well as thermodynamic and kinetic parameters. Loun and Hage [183,184] have systematically investigated the chiral separation mechanism of a HSA CSP by using warfarin as a model solute. From frontal analysis, it was found that both enantiomers have a single type of binding site in the same area of HSA. Temperature-dependent measurements of the equilibrium constant, obtained over the temperature range from 4° to 45°C, enabled the calculation of specific thermodynamic parameters. Although the changes in molar free energy were similar for R- and S-warfarin (–7.5 and –7.7 kcal/mol, respectively, at 37°C), the contributions from molar enthalpy and entropy were different. The R-enantiomer had a smaller contribution than the S-enantiomer from the enthalpy change (–3.5 and –5.6 kcal/mol, respectively), but a larger contribution from the entropy change (13 and 7 cal/mol K, respectively). The larger entropy change for R-warfarin indicated that more solvent molecules were released during the binding, while the larger enthalpy change for S-warfarin indicated stronger interaction during the binding. This is consistent with a model in which R-warfarin interacts with amino acid residues deep within the binding pocket of HSA mainly by hydrophobic interactions, whereas S-warfarin interacts with residues at or near the outer rim of the binding site mainly by polar interactions. The effect of the mobile phase was investigated by increasing the concentration of the polar organic modifier 2-propanol in aqueous phosphate buffer. Concomitantly, the retention factor for R-warfarin decreased, but that for S-warfarin increased. This result provided further support for the authors’

DK4478_book.fm Page 74 Thursday, September 14, 2006 1:35 PM

74

Advances in Chromatography, Volume 45

suggestion that R-warfarin interacted by nonpolar interaction, while S-warfarin interacted by polar interaction. From elution analysis, the plate height model was utilized to study the band broadening processes and to measure sorption/desorption kinetics. With an increase in temperature, both the sorption and desorption rate constants increased while the equilibrium constant decreased. The faster kinetics resulted in a decrease in plate height and an increase in column efficiency. The measurement of mass transfer kinetics at different temperatures allowed the calculation of activation energy, activation enthalpy, and activation entropy. Although the activation energy was similar for R- and S-warfarin (10.8 and 11.5 kcal/mol, respectively, at 37°C), the contributions from activation enthalpy and entropy were different. The R-enantiomer had a larger activation enthalpy (10.5 and 1.9 kcal/mol), but a smaller activation entropy (–0.9 and 31 cal/mol K) than the S-enantiomer. The large positive activation enthalpy for R-warfarin indicated more bonds breaking than forming during the complex-formation process, whereas the more negative activation entropy for S-warfarin indicated a higher degree of order. These results also corresponded well with the retention model presented previously. In this way, thermodynamic and kinetic information help to elucidate specific retention and separation mechanisms. Similar experiments were performed for another chiral solute, tryptophan, by Yang and Hage [185–187]. Both D- and L-tryptophan had a single type of binding site, but unlike the warfarin enantiomers, they were in different areas of HSA. The strength of L-tryptophan binding was enthalpy dominated, while Dtryptophan was entropy dominated. Consequently, L-tryptophan showed greater changes when varying the temperature, pH, ionic strength, and modifier concentration of the mobile phase. These changes were predominantly due to changes in the equilibrium constant, but also to changes in the moles of binding sites (phase ratio). From kinetic studies, L-tryptophan showed much greater broadening and, hence, slower kinetics than D-tryptophan. The primary contribution was the much slower desorption rate of L-tryptophan from the HSA phase. Fornstedt et al. [188–190] and Goetmar et al. [191] systematically investigated another type of protein CSP, immobilized cellobiohydrolase I (CBH I), with some β-andrenergic blocking agents as chiral solutes. The sorption isotherms for both Rand S-propranolol were fit well to the bi-Langmuir model. One of the contributions was the same for both enantiomers, whereas the other was different. The common contribution was considered to be the nonchiral (type I) interaction, while the other was the chirally selective (type II) interaction. It was suspected that the chirally selective interaction involved the formation of ion pairs on this CSP. Thus the saturation capacity of the type II sites was lower because of the limited number of ionic sites. As a result, the enantioselectivity decreased with an increase in solute concentration and eventually vanished when the nonchiral interactions became predominant. Also, due to the ionic nature of the chirally selective sites, retention behavior was pH dependent. At pH 4.7, the retention times of both enantiomers decreased with increasing temperature from 5° to 55°C. However, at pH 5.5, the retention time of the more retained S-propranolol increased, while that of R-propranolol decreased with increasing temperature. Thus, at this pH, an unusual increase in chiral selectivity was observed with an increase in temperature. Detailed thermodynamic parameters were determined at this pH to try to explain this phenomenon.

DK4478_book.fm Page 75 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

75

The changes in molar enthalpy and entropy on type I sites were calculated as –1.1 kcal/mol and 0.1 cal/mol K, respectively. On type II sites, R-propranolol had negative changes in molar enthalpy and entropy at –1.9 kcal/mol and –2.6 cal/mol K, respectively, whereas S-propranolol had positive changes at 1.6 kcal/mol and 11.6 cal/mol K, respectively. The authors attributed the large positive entropy change to a considerable decrease in the organization of water molecules around the chiral (type II) sites in the protein CSP. The positive enthalpy change suggested that this disorganization was associated with weakening of the hydrogen bonding of water molecules. The endothermic interaction for the S-enantiomer and exothermic interaction for the R-enantiomer accounted for the unusual temperature-dependent behavior. Peak tailing was observed at both low and high concentrations of solutes. At low concentrations, the type II sites were within the linear range of the isotherm. The tailing observed was a result of slow heterogeneous mass transfer kinetics. At high concentrations, the type II sites were saturated and no longer within the linear range. As a result, the more pronounced tailing observed was a combination of heterogeneous thermodynamics and heterogeneous kinetics. The transport-dispersive model using the solid-film linear driving force model with homogeneous kinetics failed to account for the pronounced peak tailing, while the same model with heterogeneous kinetics accounted well. For example, a sorption rate constant of 10,000 min–1 and desorption rate constant of 18/min–1 were determined for best fit to the experimental zone profile produced with 100 μl of 0.36 mM propranolol. The rate constants were found to be concentration dependent. With higher concentrations of 1.02 and 1.55 mM propranolol, the sorption rate constant was decreased to 90 min–1 and 60 min–1, respectively, when the desorption rate constant was held constant at 18 min–1. Some other interesting thermodynamic studies have been performed on other types of protein CSPs. Gilpin et al. [192] separated tryptophan enantiomers on a BSA CSP at varying temperatures and pH values. For D-tryptophan, the van’t Hoff plot was reasonably linear over the temperature range from 0.5° to 45°C. However, for L-tryptophan, a nonlinear plot was observed with a maximum value at approximately 20° to 24°C. The authors noted that the initial increase in the retention factor of L-tryptophan with temperature is thermodynamically inconsistent with a simple retention mechanism, where binding occurs at a fixed number of specific sites with a constant, negative change in molar enthalpy. They speculated that the observed nonlinearity may be due to an increase in the number of binding sites due to gross changes in protein surface orientation or an increase in the binding energy due to induced conformational changes of the protein. These changes are apparently not consequential for D-tryptophan, which does not have a specific binding site with BSA. Amyloglucosidase was immobilized onto the silica support by Strandberg et al. [193]. Both retention and enantioselectivity increased with column temperature. The ionic strength of the mobile phase and the concentration of the organic modifier 2-propanol both affected the chiral selectivity. The pH of the mobile phase affected the charges of the CSP and solute, thereby influencing their interactions. A pH above the isoelectric point of the protein CSP introduced more electrostatic interaction between the positively charged solute and the negatively charged protein. The thermodynamic results indicated that the interaction could be enthalpy or entropy dominated, depending on the mobile phase composition. An inversion of elution order

DK4478_book.fm Page 76 Thursday, September 14, 2006 1:35 PM

76

Advances in Chromatography, Volume 45

was observed for enantiomers of mosapride on an α1-acid glycoprotein CSP with a change in temperature and pH [194]. A linear van’t Hoff plot was obtained at a lower pH (4.2), but a nonlinear plot was obtained at a higher pH (7.4). Thermodynamic results showed the separation was enthalpy dominated at a lower pH, but entropy dominated at a higher pH. A vancomycin-based CSP was used to separate a series of dansyl amino acids in the temperature range of 0° to 28°C [195]. Linear van’t Hoff plots were obtained for all enantiomers. However, their thermodynamic parameters were distinguished in two groups, where the first group had a more negative ΔH with smaller molecular size (e.g., dansyl valine and leucine) and the second group had a much less negative ΔH with a larger molecular size (e.g., dansyl tryptophan). Enthalpy-entropy compensation was observed for the chiral compounds within the same group, indicating a similar retention mechanism. The size of the solute affected its retention and chiral recognition by affecting the hydrophobic interaction and steric hindrance with the CSP. Berthod et al. [196] observed linear van’t Hoff plots on four macrocyclic glycopeptide CSPs with 71 chiral compounds, which indicated that no conformational transitions were observed for these CSPs. Most chiral separations were enthalpy controlled, with a few exceptions, depending on the mobile phase composition and the solute structure as well. Enthalpy-entropy compensation was observed in most cases. In summary, protein CSPs have a much more complex structure than those CSPs containing only small chiral molecules. Temperature and pH are the most important parameters that control retention and separation behavior in this type of CSP. Broader peaks are usually observed due to the slow mass transfer kinetics. More investigations need to be performed to increase mass transfer kinetics and enable greater column efficiency for protein CSPs.

V. CONCLUSION Thermodynamic studies provide a detailed description of the equilibrium (or steady-state) aspects of the retention mechanism, including energetic, enthalpic, entropic, and volumetric contributions. Kinetic studies provide a complementary description of the nonequilibrium (or nonsteady-state) aspects, including the same energetic, enthalpic, entropic, and volumetric contributions. With alternative methods of data analysis, kinetic studies can also provide an account of the rank and magnitude of mass transfer processes. This detailed knowledge is necessary to transform routine applications of liquid chromatography from trial-and-error efforts to rational and efficient optimizations. Moreover, this knowledge is helpful to guide the development of new stationary phases with improved thermodynamic and kinetic performance. Among the common retention mechanisms in liquid chromatography, the most systematic and thorough thermodynamic and kinetic studies have been performed for partition using alkylsilica stationary phases operated in the reversed-phase mode. These studies have provided a cohesive picture of the retention event that, while not fully complete, includes the effects of stationary phase composition, mobile phase composition, solute structure, as well as temperature and pressure. Other mechanisms, such as adsorption, ion exchange, and size exclusion, are reasonably well

DK4478_book.fm Page 77 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

77

understood from a thermodynamic perspective, but not sufficiently well characterized from a kinetic perspective. The most complicated separations are those that combine multiple mechanisms, such as chiral separations. For these complex separations, the small number of studies that have been performed with unrelated stationary phases and solutes provide, at best, an unclear and incomplete description. Much more systematic work needs to be done to elucidate their thermodynamic and kinetic behavior more fully. In addition, much work remains to be done in the nonlinear regime of the isotherms so that all sites on these stationary phases can be adequately characterized.

REFERENCES 1. McGuffin, V.L., Theory of chromatography, in Chromatography, 6th ed., Fundamentals and Applications of Chromatography and Related Differential Migration Methods — Part A: Fundamentals and Techniques, Heftmann, E., ed., Elsevier, Amsterdam, 2004, pp. 1–93. 2. Giddings, J.C., Dynamics of Chromatography, Marcel Dekker, New York, 1965. 3. Atkins, P.W., Physical Chemistry, 6th ed., W.H. Freeman, New York, 1997. 4. Leffler, J. and Grunwald, E., Rates and Equilibria of Organic Reactions, Wiley, New York, 1963. 5. Krug, R.R., Hunter, W.G., and Grieger, R.A., Enthalpy-entropy compensation. 1. Some fundamental statistical problems associated with the analysis of van’t Hoff and Arrhenius data, J. Phys. Chem., 80, 2335, 1976. 6. Krug, R.R., Hunter, W.G., and Grieger, R.A., Enthalpy-entropy compensation. 2. Separation of the chemical from the statistical effect, J. Phys. Chem., 80, 2341, 1976. 7. Ranatunga, R., Vitha, M.F., and Carr, P.W., Mechanistic implications of the equality of compensation temperatures in chromatography, J. Chromatogr. A, 946, 47, 2002. 8. McGuffin, V.L. and Chen, S.H., Molar enthalpy and molar volume of methylene and benzene homologues in reversed-phase liquid chromatography, J. Chromatogr. A, 762, 35, 1998. 9. Miyabe, K. and Guiochon, G., Measurement of the parameters of the mass transfer kinetics in high performance liquid chromatography, J. Separat. Sci., 26, 155, 2003. 10. Bernasconi, C.F., Relaxation Kinetics, Academic Press, New York, 1976. 11. Harris, J.M. and Marshall, D.B., Direct measurement of sorption/desorption kinetics in reversed-phase chromatographic systems, J. Microcolumn Sep., 9, 185, 1997. 12. Dawson, A., Gormally, J., Wyn-Jones, E., and Holzwarth, J., Comparison of results obtained in kinetic studies of the interaction between cationic dyes and sodium carboxymethylcellulose using the Joule-heating temperature-jump and laser temperature-jump methods, J. Chem. Soc. Chem. Commun., 8, 386, 1981. 13. Waite, S.W., Marshall, D.B., and Harris, J.M., Temperature-jump investigation of sorption/desorption kinetics at reversed-phase chromatographic silica/solution interfaces, Anal. Chem., 66, 2052, 1994. 14. Ellison, E.H., Waite, S.W., Marshall, D.B., and Harris, J.M., Joule-discharge heating studies of pore connectivity in silica gel: Influence of pore diameter, chemical modification, and particle size, Anal. Chem., 65, 3622, 1993. 15. Waite, S.W., Harris, J.M., Ellison, E.H., and Marshall, D.B., Temperature-jump relaxation kinetics at liquid/solid interfaces: Fluorescence thermometry of porous silica heated by a Joule discharge, Anal. Chem., 63, 2365, 1991.

DK4478_book.fm Page 78 Thursday, September 14, 2006 1:35 PM

78

Advances in Chromatography, Volume 45 16. Caldin, E.F. and Field, J.P., Kinetics of complex formation between zinc mesotetraphenylporphyrin and some nitrogen bases in aprotic solvents, studied by an improved microwave temperature-jump method, J. Chem. Soc. Faraday Trans. I, 78, 1923, 1982. 17. Waite, S.W., Holzwarth, J.F., and Harris, J.M., Laser temperature jump relaxation measurements of adsorption/desorption kinetics at liquid/solid interfaces, Anal. Chem., 67, 1390, 1995. 18. Marshall, D.B., Burns, J.W., and Connolly, D.E., Direct measurement of liquid chromatographic sorption-desorption kinetics and the kinetic contribution to band broadening, J. Chromatogr., 360, 13, 1986. 19. Shield, S.R. and Harris, J.M., Triplet-state photoexcitation dipole-jump relaxation method to observe adsorption/desorption kinetics at a reversed-phase silica/solution interface, Anal. Chem., 74, 2248, 2002. 20. Gowanlock, D., Bailey, R., and Cantwell, F.F., Intra-particle sorption rate and liquid chromatographic bandbroadening in porous polymer packings. I. Methodology and validation of the model, J. Chromatogr. A, 726, 1, 1996. 21. Li, J., Litwinson, L.M., and Cantwell, F.F., Intra-particle sorption rate and liquid chromatographic bandbroadening in porous polymer packings. II. Slow sorption rate on a microparticle packing, J. Chromatogr. A, 726, 25, 1996. 22. Li, J. and Cantwell, F.F., Intra-particle sorption rate and liquid chromatographic bandbroadening in porous polymer packings. III. Diffusion in the polymer matrix as the cause of slow sorption, J. Chromatogr. A, 726, 37, 1996. 23. Bujalski, R. and Cantwell, F.F., Continuous desorption rate measurement from a shallow-bed of poly(styrene-divinylbenzene) particles with correction for experimental artifacts, J. Chromatogr. A, 1048, 173, 2004. 24. Howerton, S.B. and McGuffin, V.L., Thermodynamic and kinetic characterization of polycyclic aromatic hydrocarbons in reversed-phase liquid chromatography, Anal. Chem., 75, 3539, 2003. 25. Thomas, H.C., Heterogeneous ion exchange in a flowing system, J. Am. Chem. Soc., 66, 1664, 1944. 26. Wade, J.L., Bergold, A.F., and Carr, P.W., Theoretical description of nonlinear chromatography, with applications to physicochemical measurements in affinity chromatography and implications for preparative-scale separations, Anal. Chem., 59, 1286, 1987. 27. Golshan-Shirazi, S. and Guiochon, G., Comparison of the various kinetic models of non-linear chromatography, J. Chromatogr., 603, 1, 1992. 28. Guiochon, G., Preparative liquid chromatography, J. Chromatogr. A, 965, 129, 2002. 29. Tijssen, R., Schoenmakers, P.J., Bohmer, M.R., Koopal, L.K., and Billiet, H.A.H., Lattice models for the description of partitioning/adsorption and retention in reversedphase liquid chromatography, including surface and shape effects, J. Chromatogr., 656, 135, 1993. 30. Horvath, C., Melander, W., and Molnar, I., Solvophobic interactions in liquid chromatography with nonpolar stationary phases, J. Chromatogr., 125, 129, 1976. 31. Horvath, C. and Melander, W., Liquid chromatography with hydrocarbonaceous bonded phases; theory and practice of reversed phase chromatography, J. Chromatogr. Sci., 15, 393, 1977. 32. Martire, D.E. and Boehm, R.E., Unified theory of retention and selectivity in liquid chromatography. 2. Reversed-phase liquid chromatography with chemically bonded phases, J. Phys. Chem., 87, 1045, 1983.

DK4478_book.fm Page 79 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

79

33. Dill, K.A., The mechanism of solute retention in reversed-phase liquid chromatography, J. Phys. Chem., 91, 1980, 1987. 34. Tan, L.C. and Carr, P.W., Revisionist look at solvophobic driving forces in reversedphase liquid chromatography II. Partitioning vs. adsorption mechanism in monomeric alkyl bonded phase supports, J. Chromatogr. A, 775, 1, 1997. 35. Park, J.H., Lee, Y.K., Weon, Y.C., Tan, L.C., Li, J., Li, L., Evans, J.F., and Carr, P.W., Revisionist look at solvophobic driving forces in reversed-phase liquid chromatography. IV. Partitioning vs. adsorption mechanism on various types of polymeric bonded phases, J. Chromatogr. A, 767, 1, 1997. 36. Horvath, C. and Melander, W., Reversed-phase chromatography and the hydrophobic effect, Am. Lab., 10, 17, 1978. 37. Carr, P.W., Li, J., Dallas, A.J., Eikens, D.I., and Tan, L.C., Revisionist look at solvophobic driving forces in reversed-phase liquid chromatography, J. Chromatogr. A, 656, 113, 1993. 38. Ranatunga, R.P.J. and Carr, P.W., A study of the enthalpy and entropy contributions of the stationary phase in reversed-phase liquid chromatography, Anal. Chem., 72, 5679, 2000. 39. Krstulovic, A.M., Colin, H., Tchapla, A., and Guiochon, G., Effects of the bonded alkyl chain length on methylene selectivity in reversed-phase liquid chromatography, Chromatographia, 17, 228, 1983. 40. Tchapla, A., Colin, H., and Guiochon, G., Linearity of homologous series retention plots in reversed-phase liquid chromatography, Anal. Chem., 56, 621, 1984. 41. Tchapla, A., Heron, S., Colin, H., and Guiochon, G., Role of temperature in the behavior of homologous series in reversed-phase liquid chromatography, Anal. Chem., 60, 1443, 1988. 42. Jinno, K. and Ozaki, N., Enthalpy-entropy compensation of octylsilica stationary phase in reversed-phase HPLC, J. Liq. Chromatogr., 7, 877, 1984. 43. Miyabe, K. and Guiochon, G., Extrathermodynamic study of surface diffusion in reversed-phase liquid chromatography with silica gels bonded with alkyl ligands of different chain lengths, J. Phys. Chem. B, 109, 12038, 2005. 44. Sentell, K.B. and Dorsey, J.G., Retention mechanisms in reversed-phase chromatography: Stationary-phase bonding density and partitioning, Anal. Chem., 61, 930, 1989. 45. Sentell, K.B. and Dorsey, J.G., Retention mechanisms in reversed-phase chromatography: Stationary phase bonding density and solute selectivity, J. Chromatogr., 461, 193, 1989. 46. Sander, L.C. and Wise, S.A., Synthesis and characterization of polymeric C18 stationary phases for liquid chromatography, Anal. Chem., 56, 504, 1984. 47. Cole, L.A. and Dorsey, J.G., Temperature dependence of retention in reversed-phase liquid chromatography. 1. Stationary-phase considerations, Anal. Chem., 64, 1317, 1992. 48. Gill, S.J. and Wadso, I., An equation of state describing hydrophobic interactions, Proc. Natl. Acad. Sci. USA, 73, 2955, 1976. 49. Rimmer, C.A., Sander, L.C., Wise, S.A., and Dorsey, J.G., Synthesis and characterization of C13 to C18 stationary phases by monomeric solution polymerized and surface polymerized approaches, J. Chromatogr. A, 1007, 11, 2003. 50. Jinno, K., Nagoshi, T., Tanaka, N., Okamoto, M., Fetzer, J.C., and Biggs, W.R., Effect of column temperature on the retention of pero-pyrene-type polycyclic aromatic hydrocarbons on various chemically bonded stationary phases in reversed-phase liquid chromatography, J. Chromatogr., 436, 1, 1988.

DK4478_book.fm Page 80 Thursday, September 14, 2006 1:35 PM

80

Advances in Chromatography, Volume 45 51. Wheeler, J.F., Beck, T.L., Klatte, S.J., Cole, L.A., and Dorsey, J.G., Phase transitions of reversed-phase stationary phases: Cause and effects in the mechanism of retention, J. Chromatogr. A, 656, 317, 1993. 52. Morel, D. and Serpinet, J., Gas chromatographic evidence for phase transitions in very compact octadecyl bonded silicas, J. Chromatogr., 200, 95, 1980. 53. Morel, D. and Serpinet, J., Influence of the liquid chromatographic mobile phase on the phase transitions of alkyl-bonded silicas studies by gas chromatography, J. Chromatogr., 214, 202, 1981. 54. Claudy, P., Letoffe, J.M., Morel, D., and Serpinet, J., Long-chain alkyl grafts and mixed alkyl-alkane layers at the surface of macroporous silicas: Their gas chromatographic properties below and above the phase transition, J. Chromatogr., 329, 331, 1985. 55. Van Miltenburg, J.C. and Hammers, W.E., Transitions in chemically bonded organic phases on silica: Specific heat measurements by adiabatic calorimetry, J. Chromatogr., 268, 147, 1983. 56. Sander, L.C., Callis, J.B., and Field, L.R., Fourier-transform infrared spectrometric determination of alkyl chain conformation on chemically bonded reversed-phase liquid chromatography packings, Anal. Chem., 55, 1068, 1983. 57. McGuffin, V.L. and Lee, C., Thermodynamics and kinetics of solute transfer in reversed-phase liquid chromatography, J. Chromatogr. A, 987, 3, 2003. 58. Howerton, S.B. and McGuffin, V.L., Thermodynamic and kinetic characterization of polycyclic aromatic hydrocarbons in reversed-phase liquid chromatography, Anal. Chem., 75, 3539, 2003. 59. Vervoort, R.J.M., Ruyter, E., Debets, A.J.J., Claessens, H.A., Cramers, C.A., and de Jong, G.J., Characterization of reversed-phase stationary phases for the liquid chromatographic analysis of basic pharmaceuticals by thermodynamic data, J. Chromatogr. A, 964, 67, 2002. 60. Neue, U.D., Alden, B.A., and Walter, T.H., Universal procedure for the assessment of the reproducibility and the classification of silica-based reversed-phase packings. II. Classification of reversed-phase packings, J. Chromatogr. A, 849, 101, 1999. 61. Neue, U.D., Van Tran, K., Iraneta, P.C., and Alden, B.A., Characterization of HPLC packings, J. Separat. Sci., 26, 174, 2003. 62. Vailaya, A. and Horvath, C., Exothermodynamic relationships in liquid chromatography, J. Phys. Chem., 102, 701, 1998. 63. Layne, J., Characterization and comparison of the chromatographic performance of conventional, polar-embedded, and polar-endcapped reversed-phase liquid chromatography stationary phases, J. Chromatogr. A, 957, 149, 2002. 64. Liu, Y., Grinberg, N., Thompson, K.C., Wenslow, R.M., Neue, U.D., Morrison, D., Walter, T.H., O’Gara, J.E., and Wyndham, K.D., Evaluation of a C18 hybrid stationary phase using high-temperature chromatography, Anal. Chim. Acta, 554, 144, 2005. 65. Sander, L.C. and Wise, S.A., Influence of substrate parameters on column selectivity with alkyl bonded-phase sorbents, J. Chromatogr., 316, 163, 1984. 66. Nahum, A. and Horvath, C., Surface silanols in silica-bonded hydrocarbonaceous stationary phases 1. Dual retention mechanism in reversed-phase chromatography, J. Chromatogr. A, 203, 53, 1981. 67. Nawrocki, J., Silica surface controversies, strong adsorption sites, their blockage and removal. Part I, Chromatographia, 31, 177, 1991. 68. Nawrocki, J., Silica surface controversies, strong adsorption sites, their blockage and removal. Part II, Chromatographia, 31, 193, 1991.

DK4478_book.fm Page 81 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

81

69. Li, J. and Carr, P.W., Effect of temperature on the thermodynamic properties, kinetic performance, and stability of polybutadiene-coated zirconia, Anal. Chem., 69, 837, 1997. 70. Li, J. and Carr, P.W., Retention characteristics of polybutadiene-coated zirconia and comparison to conventional bonded phases, Anal. Chem., 68, 2857, 1996. 71. Guillarme, D., Heinisch, S., and Rocca, J.L., Effect of temperature in reversed phase liquid chromatography, J. Chromatogr. A, 1052, 39, 2004. 72. Coym, J.W. and Dorsey, J.G., Reversed-phase retention thermodynamics of purewater mobile phases at ambient and elevated temperature, J. Chromatogr. A, 1035, 23, 2004. 73. Grushka, E., Colin, H., and Guiochon, G., Retention behaviour of alkylbenzenes as a function of temperature and mobile phase composition in reversed-phase chromatography, J. Chromatogr., 248, 325, 1982. 74. Cole, L.A., Dorsey, J.G., and Dill, K.A., Temperature dependence of retention in reversed-phase liquid chromatography. 2. Mobile-phase considerations, Anal. Chem., 64, 1324, 1992. 75. Sentell, K.B., Ryan, N.I., and Henderson, A.N., Temperature and solvation effects on homologous series selectivity in reversed phase liquid chromatography, Anal. Chim. Acta, 307, 203, 1995. 76. Sander, L.C. and Field, L.R., Effect of eluent composition on thermodynamic properties in high-performance liquid chromatography, Anal. Chem., 52, 2009, 1980. 77. Barman, B.N. and Martire, D.E., Factors influencing retention and resolution of substituted alkylbenzenes in reversed-phase liquid chromatography, Chromatographia, 34, 347, 1992. 78. Wu, N., Yehl, P.M., Gauthier, D., and Dovletoglou, A., Retention and thermodynamic studies of piperazine diasteromers in reversed-phase liquid chromatography, Chromatographia, 59, 189, 2004. 79. Melander, W., Campbell, D.E., and Horvath, C., Enthalpy-entropy compensation in reversed-phase chromatography. J. Chromatogr., 158, 215, 1978. 80. Knox, J.H. and Vasvari, G., Performance of packings in high-speed liquid chromatography. III. Chemically bonded pellicular materials, J. Chromatogr., 83, 181, 1973. 81. Miyabe, K., Sotoura, S., and Guiochon, G., Retention and mass transfer characteristics in reversed-phase liquid chromatography using a tetrahydrofuran-water solution as the mobile phase, J. Chromatogr. A, 919, 231, 2001. 82. Zhao, J. and Carr, P.W., A comparative study of the chromatographic selectivity of polystyrene-coated zirconia and related reversed-phase materials, Anal. Chem., 72, 302, 2000. 83. Chen, Z., Nakayama, T., Nakagama, T., Uchiyama, K., and Hobo, T., Thermodynamic approaches to intermolecular interaction and retention behavior in liquid chromatography, J. Liq. Chromatogr. Relat. Technol., 26, 2809, 2003. 84. Carr, P.W., Tan, L.C., and Park, J.H., Revisionist look at solvophobic driving forces in reversed-phase liquid chromatography. III. Comparison of the behavior of nonpolar and polar solutes, J. Chromatogr., 724, 1, 1996. 85. Alvarez-Zepeda, A., Barman, B.N., and Martire, D.E., Thermodynamic study of the marked differences between acetonitrile/water and methanol/water mobile-phase systems in reversed-phase liquid chromatography, Anal. Chem., 64, 1978, 1992. 86. Opperhuizen, A., Sinnige, T.L., Van Der Steen, J.M.D., and Hutzinger, O., Differences between retentions of various classes of aromatic hydrocarbons in reversed-phase high-performance liquid chromatography: Implications of using retention data for characterizing hydrophobicity, J. Chromatogr., 388, 51, 1987.

DK4478_book.fm Page 82 Thursday, September 14, 2006 1:35 PM

82

Advances in Chromatography, Volume 45

87. Howerton, S.B. and McGuffin, V.L., Thermodynamics and kinetics of solute transfer in reversed-phase liquid chromatography: Effect of annelation in polycyclic aromatic hydrocarbons, J. Chromatogr. A, 1030, 3, 2004. 88. McGuffin, V.L., Howerton, S.B., and Li, X., Thermodynamic and kinetic characterization of nitrogen-containing polycyclic aromatic hydrocarbons in reversed-phase liquid chromatography, J. Chromatogr. A, 1073, 63, 2005. 89. Miyabe, K. and Guiochon, G., A kinetic study of mass transfer in reversed-phase liquid chromatography on a C-18 silica gel, Anal. Chem., 72, 5162, 2000. 90. Coutinho, F.M.B., Neves, M.A.F.S., and Dias, M.L., Porous crosslinked spherical resins for diversified applications: Packing materials for size exclusion chromatography, Macromol. Symp., 189, 27, 2002. 91. Tanaka, N. and Araki, M., Polymer-based packing materials for reversed-phase liquid chromatography, Adv. Chromatogr., 30, 81, 1989. 92. Andersen, T., Nguyen, Q.T., Trones, R., and Greibrokk, T., Studies on the long-term stability of particulate and monolithic poly(styrene-divinylbenzene) capillaries in reversed-phase liquid chromatography, Analyst, 129, 191, 2004. 93. Oberacher, H., Premstaller, A., and Huber, C.G., Characterization of some physical and chromatographic properties of monolithic poly(styrene-co-divinylbenzene) columns, J. Chromatogr. A, 1030, 201, 2004. 94. McNeff, C., Zigan, L., Johnson, K., Carr, P.W., Wang, A., and Weber-Main, A.M., Analytical advances of highly stable stationary phases for reversed-phase LC, LCGC North Am., 18, 514, 2000. 95. Yarita, T., Nakajima, R., and Shibukawa, M., Superheated water chromatography of phenols using poly(styrene-divinylbenzene) packings as a stationary phase, Anal. Sci., 19, 269, 2003. 96. Ells, B., Wang, Y., and Cantwell, F.F., Influence of solvent uptake and swelling by poly(styrene-divinylbenzene) column packings on sample sorption rate and band broadening in reversed-phase liquid chromatography, J. Chromatogr. A, 835, 3, 1999. 97. Knox, J.H., Kaur, B., and Millward, G.R., Structure and performance of porous graphitic carbon in liquid chromatography, J. Chromatogr., 352, 3, 1986. 98. Knox, J.H. and Ross, P., Carbon-based packing materials for liquid chromatography: Structure, performance, and retention mechanism, Adv. Chromatogr., 37, 121, 1997. 99. Hanai, T., Separation of polar compounds using carbon columns, J. Chromatogr. A, 989, 183, 2003. 100. Zhang, Y. and McGuffin, V.L., Thermodynamic and kinetic characterization of porous graphitic carbon in reversed-phase liquid chromatography, J. Chromatogr. A, submitted for publication. 101. Viron, C., Andre, P., Dreux, M., and Lafosse, M., Evaluation of porous graphitic carbon as stationary phase of the analysis of fatty acid methyl esters by liquid chromatography, Chromatographia, 49, 137, 1999. 102. Gaudin, K., Chaminade, P., and Baillet, A., Eluotropic strength in non-aqueous liquid chromatography with porous graphitic carbon, J. Chromatogr. A, 973, 61, 2002. 103. Koivisto, P. and Stefansson, M., Retention studies of sulphated glycosaminoglycan disaccharides on porous graphitic carbon capillary columns, Chromatographia, 57, 37, 2003. 104. Kwaterczak, A. and Bielejewska, A., Comparison of retention of native cyclodextrins and its permethylated derivatives on porous graphitic carbon and silica C18 stationary phases, Anal. Chim. Acta, 537, 41, 2005.

DK4478_book.fm Page 83 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

83

105. Clarot, I., Cledat, D., Guillaume, Y.C., and Cardot, P.J.P., Chromatographic study of terpene-β-cyclodextrin complexes on porous graphitic carbon stationary phase, Chromatographia, 54, 447, 2001. 106. Andre, C. and Guillaume, Y.C., Saccharose effects on surface association of phenol derivatives with porous graphitic carbon, J. Chromatogr. A, 1029, 21, 2004. 107. Nawrocki, J., The silanol group and its role in liquid chromatography, J. Chromatogr. A, 779, 29, 1997. 108. Stella, C., Rudaz, S., Veuthey, J.L., and Tchapla, A., Silica and other materials as supports in liquid chromatography: Chromatographic tests and their importance for evaluating these supports. Part I, Chromatographia, 53(suppl.), S113, 2001. 109. Snyder, L.R., Principles of Adsorption Chromatography, Marcel Dekker, New York, 1968. 110. Snyder, L.R., Linear elution adsorption chromatography. V. Silica as adsorbent. Adsorbent standardization, J. Chromatogr., 11, 195, 1963. 111. Snyder, L.R., Adsorption from solution. IV. Aromatic hydrocarbons on silica, J. Phys. Chem., 67, 2622, 1963. 112. Snyder, L.R., Linear elution adsorption chromatography. XIII. A narrow pore silica, J. Chromatogr., 25, 274, 1966. 113. Basiuk, V.A. and Gromovoy, T.Y., Comparative study of amino acid adsorption on bare and octadecyl silica from water using high-performance liquid chromatography, Colloid Surface A, 118, 127, 1996. 114. Basiuk, V.A. and Gromovoy, T.Y., Estimation of thermodynamic parameters for amino acid adsorption on silica from water using high-performance liquid-chromatographic technique, Pol. J. Chem., 70, 476, 1996. 115. Bidlingmeyer, B.A. and Henderson, J., Investigation of retention on bare silica using reversed-phase mobile phases at elevated temperatures, J. Chromatogr. A, 1060, 187, 2004. 116. Snyder, L.R., An experimental study of column efficiency in liquid-solid adsorption chromatography HETP values as a function of separation conditions, Anal. Chem., 39, 698, 1967. 117. Ohkuma, T. and Hara, S., Tail-producing slow adsorption-desorption process in liquid-solid chromatography, J. Chromatogr., 400, 47, 1987. 118. Fornstedt, T., Zhong, G., and Guiochon, G., Peak tailing and mass transfer kinetics in linear chromatography, J. Chromatogr. A, 741, 1, 1996. 119. Fornstedt, T., Zhong, G., and Guiochon, G., Peak tailing and slow mass transfer kinetics in nonlinear chromatography, J. Chromatogr. A, 742, 55, 1996. 120. Hatsis, P. and Lucy, C.A., Effect of temperature on retention and selectivity in ion chromatography of anions, J. Chromatogr. A, 920, 3, 2001. 121. Fortier, N.E. and Fritz, J.S., The effect of temperature on single-column ion-chromatography of metal ions, Talanta, 34, 415, 1978. 122. Helfferich, F., Ion Exchange, McGraw-Hill, New York, 1962. 123. Elefterov, A.I., Kolpachnikova, M.G., Nesterenko, P.N., and Shpigun, O.A., Ionexchange properties of glutamic acid-bonded silica, J. Chromatogr. A, 769, 179, 1997. 124. Kolpachnikova, M.G., Penner, N.A., and Nesterenko, P.N., Effect of temperature on retention of alkali and alkaline-earth metal ions on some aminocarboxylic acid functionalised silica based ion exchangers, J. Chromatogr. A, 826, 15, 1998. 125. Bashir, W., Tyrrell, E., Feeney, O., and Paull, B., Retention of alkali, alkaline earth and transition metals on an itaconic acid cation-exchange column: Eluent pH, ionic strength and temperature effects upon selectivity, J. Chromatogr. A, 964, 113, 2002.

DK4478_book.fm Page 84 Thursday, September 14, 2006 1:35 PM

84

Advances in Chromatography, Volume 45

126. Janos, P., Retention models in ion chromatography: The role of side equilibria in ionexchange chromatography of inorganic cations and anions, J. Chromatogr. A, 789, 3, 1997. 127. Bowen, W.R. and Hughes, D.T., Ion exchange of proteins: a microcalorimetric study of the adsorption of bovine serum albumin on anion-exchange materials, J. Colloid Interface Sci., 158, 395, 1993. 128. Thrash, M.E., Jr., Phillips, J.M., and Pinto, N.G., An analysis of the interactions of BSA with an anion-exchange surface under linear and non-linear conditions, Adsorption, 10, 299, 2005. 129. Raje, P. and Pinto, N.G., Importance of heat of adsorption in modeling protein equilibria for overloaded chromatography, J. Chromatogr. A, 796, 141, 1998. 130. Thrash, M.E., Jr. and Pinto, N.G., Characterization of enthalpic events in overloaded ion-exchange chromatography, J. Chromatogr. A, 944, 61, 2002. 131. Weaver, L.E., Jr. and Carta, G., Protein adsorption on cation exchangers: Comparison of macroporous and gel-composite media, Biotechnol. Prog., 12, 342, 1996. 132. Fernandez, M.A. and Carta, G., Characterization of protein adsorption by composite silica-polyacrylamide gel anion exchangers I. Equilibrium and mass transfer in agitated contactors, J. Chromatogr. A, 746, 169, 1996. 133. Fernandez, M.A., Laughinghouse, W.S., and Carta, G., Characterization of protein adsorption by composite silica-polyacrylamide gel anion exchangers. II. Mass transfer in packed columns and predictability of breakthrough behavior, J. Chromatogr. A, 746, 185, 1996. 134. Heeter, G.A. and Liapis, A.I., Model discrimination and estimation of the intraparticle mass transfer parameters for the adsorption of bovine serum albumin onto porous adsorbent particles by the use of experimental frontal analysis data, J. Chromatogr. A, 776, 3, 1997. 135. Conder, J.R., and Hayek, B.O., Adsorption and desorption kinetics of bovine serum albumin in ion exchange and hydrophobic interaction chromatography on silica matrices, Biochem. Eng. J., 6, 225, 2000. 136. Miyabe, K. and Guiochon, G., Kinetic study of the concentration dependence of the mass transfer rate coefficient in anion-exchange chromatography of bovine serum albumin, Biotechnol. Prog., 15, 740, 1999. 137. Yau, W.W., Kirkland, J.J., and Bly, D.D., Modern Size-Exclusion Liquid Chromatography, Wiley, New York, 1979. 138. Casassa, E.F., Equilibrium distribution of flexible polymer chains between a macroscopic solution phase and small voids, J. Polym. Sci. B, 5, 773, 1967. 139. Casassa, E.F. and Tagami, Y., Equilibrium theory for exclusion chromatography of branched and linear polymer chains, Macromolecules, 2, 14, 1969. 140. Casassa, E.F., Theoretical models for peak migration in gel permeation chromatography, J. Phys. Chem., 75, 3929, 1971. 141. Brooks, D.E., Hanes, C.A., Hritcu, D., Steels, B.M., and Mueller, W., Size exclusion chromatography does not require pores, Proc. Natl. Acad. Sci. USA, 97, 7064, 2000. 142. Netopilik, M., Relation between the kinetic and equilibrium quantities in size-exclusion chromatography, J. Chromatogr. A, 1038, 67, 2004. 143. Gorbunov, A.A. and Skvorcov, A.M., Statistical properties of confined macromolecules, Adv. Colloid Interface Sci., 62, 31, 1995. 144. Macko, T. and Hunkeler, D., Liquid chromatography under critical and limiting conditions: A survey of experimental systems for synthetic polymers, Adv. Polym. Sci., 163, 61, 2003.

DK4478_book.fm Page 85 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

85

145. Nefedov, P.P. and Zhmakina, T.P., Effect of temperature, solvent composition, and pressure on adsorption and chromatography of polystyrenes on macroporous glasses, Vysokomol. Soed. A, 23, 276, 1981. 146. Trathnigg, B., Rappel, C., Fraydl, S., and Gorbunov, A., Liquid chromatography of polyoxyethylenes under critical conditions: A thermodynamic study, J. Chromatogr. A, 1085, 253, 2005. 147. Philipsen, H.J.A., Claessens, H.A., Lind, H., Klumperman, B., and German, A.L., Study on the retention behaviour of low-molar-mass polystyrenes and polyesters in reversed-phase liquid chromatography by evaluation of thermodynamic parameters, J. Chromatogr. A, 790, 101, 1997. 148. Dawkins, J.V., High performance gel permeation chromatography of polymers, Pure Appl. Chem., 54, 281, 1982. 149. Leypoldt, J.K., Frigon, R.P., and Lee, W., Chromatogram broadening of proteins and dextrans in size exclusion chromatography, J. Appl. Polym. Sci., 29, 3533, 1984. 150. Dalgliesh, C.E., Optical resolution of aromatic amino acids on paper chromatograms, J. Chem. Soc., 137, 3940, 1952. 151. Mazzo, D.J., Lindemann, C.J., and Brenner, G.S., Subambient temperature highperformance liquid chromatographic determination of the enantiomers of ((6,7dichloro-2,3-dihydro-2-methyl-1-oxo-2-phenyl-1H-inden-5-yl)oxy)acetic acid, Anal. Chem., 58, 636, 1986. 152. Pescher, P., Caude, M., Rosset, R., and Tambute, A., Enantiomeric separation of tertiary phosphine oxides on Pirkle’s chiral stationary phase: Mobile phase and temperature optimization, J. Chromatogr., 371, 159, 1986. 153. Weaner, L.E. and Hoerr, D.C., Separation of fatty acid ester and amide enantiomers by high-performance liquid chromatography on chiral stationary phases, J. Chromatogr., 437, 109, 1988. 154. Pirkle, W.H., Unusual effect of temperature on the retention of enantiomers on a chiral column, J. Chromatogr., 558, 1, 1991. 155. Pirkle, W.H. and Murray, P.G., An instance of temperature-dependent elution order of enantiomers from a chiral brush-type HPLC column, J. High Resolut. Chromatogr., 16, 285, 1993. 156. Pirkle, W.H. and Welch, C.J., An investigation into the role of solvation in a well characterized chiral recognition system, J. Liq. Chromatogr., 14, 2027, 1991. 157. Blackwell, J.A., Waltermire, R.E., and Stringham, R.W., The role of modifiers in high selectivity chiral separations on Pirkle-type columns, Enantiomer, 6, 353, 2001. 158. Pirkle, W.H. and Readnour, R.S., Chromatographic approach to the measurement of the interstrand distance for some chiral bonded phases, Anal. Chem., 63, 16, 1991. 159. Harada, A., Furue, M., and Nozakura, S., Optical resolution of mandelic acid derivatives by column chromatography on crosslinked cyclodextrin gels, J. Polym. Sci. Polym. Chem., 16, 189, 1978. 160. Hinze, W.L., Riehl, T.E., Armstrong, D.W., DeMond, W., Alak, A., and Ward, T., Liquid chromatographic separation of enantiomers using a chiral β-cyclodextrinbonded stationary phase and conventional aqueous-organic mobile phases, Anal. Chem., 57, 237, 1985. 161. Aboul-Enein, H.Y., Islam, M.R., and Bakr, S.A., Direct HPLC resolution of racemic nomifensine hydrogen maleate using a chiral beta-cyclodextrin-bonded stationary phase, J. Liq. Chromatogr., 11, 1485, 1988. 162. Cabrera, K. and Lubda, D., Influence of temperature on chiral high-performance liquid chromatographic separations of oxazepam and prominal on chemically bonded β-cyclodextrin as stationary phase, J. Chromatogr. A, 666, 433, 1994.

DK4478_book.fm Page 86 Thursday, September 14, 2006 1:35 PM

86

Advances in Chromatography, Volume 45

163. Morin, N., Guillaume, Y.C., Peyrin, E., and Rouland, J.C., Retention mechanism study of imidazole derivatives on a β-cyclodextrin-bonded stationary phase: Thermal analysis contributions, Anal. Chem., 70, 2819, 1998. 164. Peyrin, E., Guillaume, Y.C., and Ravel, A., Degree of solute inclusion in native β-cyclodextrin: Chromatographic approach, Anal. Chem., 72, 1263, 2000. 165. Guillaume, Y.C., Robert, J.F., and Guinchard, C., Role of the water as a surface tension modifier for the retention of a series of dansylaminoacids on a β-CD stationary phase, Talanta, 55, 263, 2001. 166. Ringo, M.C. and Evans, C.E., Pressure-dependent retention and selectivity in reversed-phase liquid chromatographic separations using β-cyclodextrin stationary phases, Anal. Chem., 69, 643, 1997. 167. Ringo, M.C. and Evans, C.E., Role of modest pressures in chirally selective complexation interactions, J. Phys. Chem. B, 101, 5525, 1997. 168. Ringo, M.C. and Evans, C.E., Pressure-induced changes in chiral separations in liquid chromatography, Anal. Chem., 69, 4964, 1997. 169. Ringo, M.C. and Evans, C.E., Effect of mobile-phase composition on pressureinduced shifts in solute retention for LC separations using β-cyclodextrin stationary phases, J. Microcol. Sep., 10, 647, 1998. 170. Wang, X. and Ching, C.B., Kinetic and equilibrium study of the separation of three chiral center drug, nadolol, by HPLC on a novel perphenyl carbamoylated β-cyclodextrin bonded chiral stationary phase, Sep. Sci. Technol., 37, 2567, 2002. 171. Yu, H.W. and Ching, C.B., Kinetic and equilibrium study of the enantioseparation of fluoxetine on a new β-cyclodextrin column by high performance liquid chromatography, Chromatographia, 54, 697, 2001. 172. Kotake, M., Sakan, T., Nakamura, N., and Senoh, S., Resolution into optical isomers of some amino acids by paper chromatography, J. Am. Chem. Soc., 73, 2973, 1951. 173. Smith, R.J., Taylor, D.R., and Wilkins, S.M., Temperature dependence of chiral discrimination in supercritical fluid chromatography and high-performance liquid chromatography, J. Chromatogr. A, 697, 591, 1995. 174. Kuesters, E. and Spoendlin, C., Influence of temperature on the enantioseparation of rolipram and structurally related racemates on Chiralcel-OD, J. Chromatogr. A, 737, 333, 1996. 175. Mesplet, N., Saito, Y., Morin, P., and Agrofoglio, L.A., Liquid chromatographic separation of phosphoramidate diastereomers on a polysaccharide-type chiral stationary phase, J. Chromatogr. A, 983, 115, 2003. 176. O’Brien, T., Crocker, L., Thompson, R., Thompson, K., Toma, P.H., Conlon, D.A., Feibush, B., Moeder, C., Bicker, G., and Grinberg, N., Mechanistic aspects of chiral discrimination on modified cellulose, Anal. Chem., 69, 1999, 1997. 177. Wang, F., O’Brien, T., Dowling, T., Bicker, G., and Wyvratt, J., Unusual effect of column temperature on chromatographic enantioseparation of dihydropyrimidinone acid and methyl ester on amylose chiral stationary phase, J. Chromatogr. A, 958, 69, 2002. 178. Wang, F., Dowling, T., Ellison, D., and Wyvratt, J., Comparison study of Chiralpak AD-H with AD columns in chromatographic enantioseparation of dihydropyrimidinone acid and its methyl ester, J. Chromatogr. A, 1034, 117, 2004. 179. Wang, F., Wenslow, R.M., Jr., Dowling, T.M., Mueller, K.T., Santos, I., and Wyvratt, J.M., Characterization of a thermally induced irreversible conformational transition of amylose tris(3,5-dimethylphenylcarbamate) chiral stationary phase in enantioseparation of dihydropyrimidinone acid by quasi-equilibrated liquid chromatography and solid-state NMR, Anal. Chem., 75, 5877, 2003.

DK4478_book.fm Page 87 Thursday, September 14, 2006 1:35 PM

The Thermodynamic and Kinetic Basis of Liquid Chromatography

87

180. Booth, T.D. and Wainer, I.W., Mechanistic investigation into the enantioselective separation of mexiletine-related compounds, chromatographed on an amylose tris(3,5-dimethylphenylcarbamate) chiral stationary phase, J. Chromatogr. A, 741, 205, 1996. 181. Rearden, P., Sajonz, P., and Guiochon, G., Detailed study of the mass transfer kinetics of Troeger’s base on cellulose triacetate, J. Chromatogr. A, 813, 1, 1998. 182. Seidel-Morgenstern, A. and Guiochon, G., Modeling of the competitive isotherms and the chromatographic separation of two enantiomers, Chem. Eng. Sci., 48, 2787, 1993. 183. Loun, B. and Hage, D.S., Chiral separation mechanisms in protein-based HPLC columns. 1. Thermodynamic studies of (R)- and (S)-warfarin binding to immobilized human serum albumin, Anal. Chem., 66, 3814, 1994. 184. Loun, B. and Hage, D.S., Chiral separation mechanisms in protein-based HPLC columns. 2. Kinetic studies of (R)- and (S)-warfarin binding to immobilized human serum albumin, Anal. Chem., 68, 1218, 1996. 185. Yang, J. and Hage, D.S., Characterization of the binding and chiral separation of Dand L-tryptophan on a high-performance immobilized human serum albumin column, J. Chromatogr. A, 645, 241, 1993. 186. Yang, J. and Hage, D.S., Role of binding capacity versus binding strength in the separation of chiral compounds on protein-based high-performance liquid chromatography columns: Interactions of D- and L-tryptophan with human serum albumin, J. Chromatogr. A, 725, 273, 1996. 187. Yang, J. and Hage, D.S., Effect of mobile phase composition on the binding kinetics of chiral solutes on a protein-based high-performance liquid chromatography column: Interactions of D- and L-tryptophan with immobilized human serum albumin, J. Chromatogr. A, 766, 15, 1997. 188. Fornstedt, T., Zhong, G., Bensetiti, Z., and Guiochon, G., Experimental and theoretical study of the adsorption behavior and mass transfer kinetics of propranolol enantiomers on cellulase protein as the selector, Anal. Chem., 68, 2370, 1996. 189. Fornstedt, T., Sajonz, P., and Guiochon, G., Thermodynamic study of an unusual chiral separation: Propranolol enantiomers on an immobilized cellulase, J. Am. Chem. Soc., 119, 1254, 1997. 190. Fornstedt, T., Goetmar, G., Andersson, M., and Guiochon, G., Dependence on the mobile-phase pH of the adsorption behavior of propranolol enantiomers on a cellulase protein used as the chiral selector, J. Am. Chem. Soc., 121, 1164, 1999. 191. Goetmar, G., Fornstedt, T., and Guiochon, G., Retention mechanism of β-blockers on an immobilized cellulase: Relative importance of the hydrophobic and ionic contributions to their enantioselective and nonselective interactions, Anal. Chem., 72, 3908, 2000. 192. Gilpin, R.K., Ehtesham, S.E., and Gregory, R.B., Liquid chromatographic studies of the effect of temperature on the chiral recognition of tryptophan by silica-immobilized bovine albumin, Anal. Chem., 63, 2825, 1991. 193. Strandberg, A., Nystrom, A., Behr, S., and Karlsson, A., Use of immobilized amyloglucosidase as chiral selector in chromatography: Control of enantioselective retention and resolution in liquid chromatography, Chromatographia, 50, 215, 1999. 194. Karlsson, A., Skoog, A., and Ohlen, K., Effect of temperature on the reversal in the retention order of the enantiomers of mosapride on Chiral-AGP, J. Biochem. Biophys. Meth., 54, 347, 2002.

DK4478_book.fm Page 88 Thursday, September 14, 2006 1:35 PM

88

Advances in Chromatography, Volume 45

195. Slama, I., Jourdan, E., Villet, A., Grosset, C., Ravel, A., and Peyrin, E., Temperature and solute molecular size effects on the retention and enantioselectivity of a series of D,L dansyl amino acids on a vancomycin-based chiral stationary phase, Chromatographia, 58, 399, 2003. 196. Berthod, A., He, B.L., and Beesley, T.E., Temperature and enantioseparation by macrocyclic glycopeptide chiral stationary phases, J. Chromatogr. A, 1060, 205, 2004.

DK4478_book.fm Page 89 Thursday, September 14, 2006 1:35 PM

2

Applications of Ionic Liquids in Extraction, Chromatography, and Electrophoresis Colin F. Poole

CONTENTS I. Introduction..................................................................................................89 II. Extraction Solvents......................................................................................93 A. Extraction of Organic Compounds .....................................................94 B. Liquid-Phase Microextraction...........................................................100 C. Supported Liquid Membranes...........................................................101 D. Extraction of Inorganic Compounds.................................................101 E. Supercritical Fluid Extraction from Ionic Liquids ...........................103 III. Liquid Chromatography.............................................................................104 IV. Electrophoresis...........................................................................................106 A. Capillary Electrophoresis ..................................................................107 V. Gas Chromatography .................................................................................109 A. Retention Mechanisms ......................................................................109 B. Chromatographic Selectivity.............................................................111 C. Chemical Reactivity Problems..........................................................117 D. Open-Tubular Columns .....................................................................117 VI. Conclusion .................................................................................................119 References..............................................................................................................120

I. INTRODUCTION Under the umbrella of green chemistry, ionic liquids have emerged as potentially useful solvents for minimizing solvent waste from chemical processes [1,2]. For these applications, their favorable properties include low or negligible vapor pressure, an ability to dissolve a wide range of inorganic and organic compounds, high thermal stability, large electrochemical windows, and low flammability. Their negligible vapor pressure minimizes environmental contamination through evaporation,

89

DK4478_book.fm Page 90 Thursday, September 14, 2006 1:35 PM

90

Advances in Chromatography, Volume 45

a major concern for conventional organic solvents, allows their use in vacuum systems without loss, and facilitates product recovery by distillation or sublimation. In addition, ionic liquids are increasingly being studied for reasons of specific advantages in rate, specificity, and yield, as much as a simple replacement for volatile organic solvents [3,4]. These aspects of ionic liquids are reviewed elsewhere [1–3,5–7]. The use of ionic liquids as solvents in analytical chemistry has not received the same attention as has their use in organic chemistry [8–11]. The purpose of this review is to provide supporting material to define a role for ionic liquids in extraction and separations in addition to describing their applications in these fields. Ionic liquids are low-melting salts forming liquids composed entirely of ions. They are expected to have properties that differ from molecular liquids because of the presence of Coulombic forces, as well as properties in common with molecular liquids, because many of the ions contain nonionic structures and functional groups that interact with each other through conventional intermolecular interactions. The majority of useful ionic liquids are organic salts containing organic cations with organic or inorganic anions. Inorganic ionic liquids tend to have relatively high melting points, are usually chemically reactive, and are poor solvents for organic compounds, rendering them less useful as general purpose solvents for extraction and separations. The term ionic liquid is a suitable keyword for searching the recent literature. Its general adoption as a keyword though is quite recent, and for comprehensive coverage, particularly of the older literature, fused salts, molten salts, and liquid organic salts should also be used. In the past, the term ionic liquid was used quite broadly to describe studies in the liquid state for any thermally stable melt. For technical and laboratory applications, the emphasis has changed to salts with low melting points producing solvents that can be used over a wide temperature range. Today, more than 200 organic salts with melting points below room temperature are known (room temperature ionic liquids) [8] and these salts account for most of the current interest in ionic liquids as general purpose solvents for extraction and separations. Some typical ions used to prepare room temperature ionic liquids are shown in Table 2.1 and Figure 2.1. Methods to predict melting points and physical properties (e.g., density, viscosity, conductivity, thermal stability, etc.) of ionic liquids are poorly developed [3,12–15]. Given the enormous number of possible cation and anion combinations representing different ionic liquids, this limits approaches for the identification of further useful ionic liquids with tailored properties for specific applications to trial and error processes. Qualitatively it seems that low symmetry, weak hydrogenbonding interactions, and effective charge dispersion for either or both ions are properties likely to yield organic salts that are liquid at room temperature or have low melting points [3,16]. These factors reduce the lattice energy of the crystalline form of the salt and hence lower the melting point. Physical properties for some representative ionic liquids are summarized in Table 2.2. The largest number of known room temperature ionic liquids are from 1,3dialkylimidazolium salts containing tetrafluoroborate, hexafluorophosphate, bis(trifluoromethylsulfonyl)amide, tricyanamide and perfluoroalkanesulfonate anions [8]. The selection of the anion seems to have a strong influence on the viscosity of the

DK4478_book.fm Page 91 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

91

TABLE 2.1 Typical Cations and Anions Used in the Synthesis of Low Melting Point Ionic Liquids Cations

Anions

Alkylammonium Tetraalkylammonium Tetraalkylphosphonium 1,3-Dialkylimidazolium 1,2,3-Trialkylimidazolium 1-(Alkoxyalkyl)-3-alkylimidazolium 1-(Hydroxylalkyl)-3-alkylimidazolium N-Alkylimidazolium N-Alkylisoquinolinium N-Alkylpyridinium 1-Alkylpiperidinium 2,3-Dialkylindolinium 1-Alkyl-4-fluoroalkyl-1,2,4-triazolium Bis(N,N-dialkyl)dimethylguanidinium

Bis(trifluoromethylsulfonyl)amide Hexafluorophosphate Tetrafluoroborate Alkylsulfate Perfluoroalkylsulfonate Alkylcarboxylate Perfluoroalkylcarboxylate Dicyanamide Nitrate Dialkylphosphate Thiocyanate Diethyleneglycolmonomethylethersulfate

H3 C + N R

R N

+

+

+ N

N

R (a)

N

(b)

R (c)

R

(d)

FIGURE 2.1 Representative structures of organic cations used to prepare low melting point ionic liquids. (a) = 1,1′-alkylpyrolinium; (b) = 1-alkyl-3-methylimidazolium; C = N-alkylpyridinium; and D = N-alkylisoquinolinium.

ionic liquid [3,17,18]. Low viscosity is associated with small anions that have a diffuse negative charge and do not take part in hydrogen bonding. Low is a relative description, however, and typical ionic liquids have viscosities greater than 20 cP at room temperature. These viscosities are larger than desirable for many applications in extraction and separation, and this remains a significant factor restricting wider use. The density of 1,3-dialkylimidazolium ionic liquids is typically greater than water and decreases almost linearly with increasing alkyl chain length [1,16]. These density differences favor separation of immiscible ionic liquids from water by settling. A number of 1,3-dialkylimidazolium salts with weakly basic anions show exceptional thermal stability, allowing their use in an inert atmosphere at temperatures above 250°C, and in a few cases above 350°C [3,8,18–20]. Even if thermal stability determined by rising temperature thermal gravimetric analysis — the most common method — tends to overestimate thermal stability for practical purposes compared with equilibrium isothermal weight loss measurements, the lack of significant vapor pressure over wide temperature ranges remains one of the

DK4478_book.fm Page 92 Thursday, September 14, 2006 1:35 PM

92

Advances in Chromatography, Volume 45

TABLE 2.2 Physical Properties of Some Room Temperature Ionic Liquidsa

Ionic liquid Di-n-propylammonium thiocyanate Ethylammonium nitrate Tetra-n-butylammonium 2-hydroxy-4morpholinopropanesulsonate Tetra-n-hexylammonium bis(trifluoromethylsulfonyl)amide N-hexyl-N,N,N-triethylammonium triethyl-n-hexylboride 1-Butyl-3-methylimidazolium heptafluorobutyrate 1-Butyl-3-methylimidazolium hexafluorophosphate 1-Butyl-3-methylimidazolium tetrafluoroborate 1-Butyl-3-methylimidazolium trifluoroacetate 1-Butyl-3-methylimidazolium trifluoromethanesulfonate 1,3-Diethylimidazolium bis(trifluoromethylsulfonyl)amide 1,3-Diethylimidazolium trifluoromethanesulfonate 1-Ethyl-3,5-dimethylimidazolium bis(trifluoromethylsulfonyl)amide 1-Ethyl-3,5-dimethylimidazolium trifluoromethanesulfonate 1-Ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide 1-Ethyl-3-methylimidazolium dicyanamide 1-Ethyl-3-methylimidazolium heptafluorobutyrate 1-Ethyl-3-methylimidazolium tetrafluoroborate 1-Ethyl-3-methylimidazolium trifluoroacetate 1-Ethyl-3-methylimidazolium trifluoromethanesulfonate N-butyl-N-methylpyrrolidinium dicyanamide 2-Methyl-1-propylpyrrolinium bis(trifluoromethylsulfonyl)amide a

Melting Point (°C)

Density (g/ml)

Viscosity (cP)

Temperature Limit (°C)

5.5 12.5

0.964 1.122

85.9 32.1 4699

–6.8

1.11

435

0.847

258

1.333

182

10

1.373

450

349

–81

1.208

219

403

–78

130 120 170

130

1.209

73 (20°C)

16

1.290

90 (20°C)

14

1.452

35 (20°C)

23

1.330

53 (20°C)

–39

1.470

37

6

1.334

51 (20°C)

1.520 (22°C)

28

400

1.06

21 (20°C)

275

–3 –21

1.450 (22°C)

105 (20°C)

1.248 (20°C)

66.5 (20°C)

–14

1.285

35 (20°C)

–9

1.390

45 (20°C)

–55

0.930

50 (20°C)

1.460 (20°C)

57

5.8

19

Temperature is 25°C unless indicated otherwise.

400

DK4478_book.fm Page 93 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

93

outstanding properties of the 1,3-dialkylimidazolium salts. Flashpoints for 1,3dialkylimidazolium salts are generally at least 100°C higher than for conventional organic solvents [21].

II. EXTRACTION SOLVENTS Ionic liquids can simultaneously dissolve organic and inorganic substances and form immiscible solvent systems with molecular liquids having a wide range of polarities. Most ionic liquids are fully or partially miscible with organic solvents of intermediate polarity (e.g., methanol, acetonitrile, tetrahydrofuran, dichloromethane, acetone, etc.). Immiscible solvent systems are most commonly formed with organic solvents of low polarity (e.g., hexane, toluene) or water. A table of solvent miscibility with water and common organic solvents for ionic liquids is given in Poole [8]. These results are based mainly on simple physical observations of mixing. There is less information available for mutual solubility of solvents in biphasic systems [22–24]. For example, the solubility of water in 1-butyl-3-methylimidazolium hexafluorophosphate corresponds to 2.3% (w/w), which on a mole fraction scale is about 0.25 (or 1.1 M) [22,25]. The solubility of 1-butyl-3-methylimidazolium hexafluorophosphate in water is about 2% (w/w). For 1-octyl-3-methylimidazolium tetrafluoroborate, the solubility of water in the ionic liquid is about 10.8% (w/w) and the ionic liquid in water is about 1.8% (w/w). In addition, it is well known that many ionic liquids strongly absorb water from the atmosphere in various amounts (e.g., 0.2 to 2.0 M) [26,27]. The absorption of water is strongly influenced by the potential of either ion to form hydrogen bonds. Thus, it is quite likely that under conditions for liquid-liquid extraction, ionic liquids will absorb appreciable amounts of water when water is a countersolvent, and probably contain appreciable amounts of water when handled under typical laboratory conditions. This poses a problem for liquid-liquid extraction, namely, the possibility of carryover reducing the selectivity of extraction from an aqueous solution and variable selectivity when the countersolvent is organic. In the latter case, results might be expected to depend on the exposure history of the ionic liquid and variations in sample hydration. Gutowski et al. [28] demonstrated that mixtures of water and miscible ionic liquids can be induced to form aqueous biphasic systems by addition of a water-structuring salt (potassium phosphate), producing a mechanism to expand the number of available liquid-liquid systems for study and creating alternative extraction media to polymerbased aqueous biphasic systems. Phase diagrams for binary and ternary solvent systems in which one component is an ionic liquid are known for only a few systems: 1-alkyl-3-methylimidazolium hexafluorophosphate-butan-1-ol, where alkyl = butyl, pentyl, hexyl, heptyl, and octyl [29]; 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide with propan-1-ol, butan-1-ol, and pentan-1-ol [30]; 1-butyl-3-methylimidazolium hexafluorophosphate-water-ethanol [31,32]; 1-butyl-3-methylimidazolium hexafluorophosphate-water with acetonitrile, methanol, and propan-1-ol [25,33]; 1-octyl3-methylimidazolium chloride-benzene with heptane, decane, and hexadecane [34]; and 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide and 1-butyl3-methylimidazolium bis(trifluoromethylsulfonyl)amide with water, acetone, and

DK4478_book.fm Page 94 Thursday, September 14, 2006 1:35 PM

94

Advances in Chromatography, Volume 45

propan-2-ol [35]. In most cases the interest has been to record the physicochemical details of the systems and to make comparisons with molecular liquids. In this regard, the results are not remarkable, except perhaps for the ternary 1-butyl-3methylimidazolium hexafluorophosphate-water-ethanol system, which forms a single phase region for mixtures containing from 0.5 to 0.9 mole fraction ethanol, even though the ionic liquid is only partially miscible with either water or ethanol [31,32]. Berthod and Carda-Broch [33] took advantage of the biphasic ternary system 1butyl-3-methylimidazolium-water-acetonitrile (2:1:2) to overcome the problem of the high viscosity of the ionic liquid to facilitate separation by countercurrent chromatography. The biphasic ternary system afforded a good density difference between the two liquid phases and viscosities low enough to allow the mobile phase to be pumped through the apparatus. The phase diagram for ternary mixtures of 1octyl-3-methylimidazolium chloride and benzene with heptane, dodecane and hexadecane was used to assess the possibility of using the ionic liquid to selectively extract benzene from hydrocarbon fuels [34]. The authors speculate that the ionic liquid is suitable for removing small amounts of benzene from n-alkanes in the preparation of high-purity n-alkanes. The ionic liquids 1-ethyl-3-methylimidazolium tetrafluoroborate and 1-butyl-3-methylimidazolium tetrafluoroborate and hexafluorophosphate were shown to be effective for the selective removal of sulfurcontaining aromatic compounds from hydrocarbon fuels [36]. Sulfur compounds with a C5 aromatic ring (e.g., thiophene) were more favorably absorbed from hydrocarbon fuels by the ionic liquids, while sulfur-containing nonaromatic compounds and aromatic hydrocarbons and n-alkanes were only poorly absorbed. 1Octyl-3-methylimidazolium hexafluorophosphate was demonstrated to be useful for the extraction of n-butanol from fermentation broth (production of biofuels) with recovery of the alcohol from the ionic liquid by pervaporation [37]. These latter two studies are mainly of industrial interest. In the sections that follow, emphasis will be on analytical laboratory-scale applications.

A. EXTRACTION

OF

ORGANIC COMPOUNDS

Partition coefficients for several organic compounds were reported for the biphasic systems containing the ionic liquids ethylammonium nitrate, n-propylammonium nitrate, or di-n-propylammonium thiocyanate and hexane, toluene, n-octanol, or dichloromethane (Table 2.3) [38–40]. Low polarity compounds are strongly distributed to the organic solvents, but those compounds capable of acting as either hydrogen-bond acids or bases are selectively extracted by the ionic liquids. Dichloromethane and n-octanol saturated with ethylammonium nitrate competes more effectively with the ionic liquid for polar solutes in general. Di-n-propylammonium thiocyanate extracts a larger amount of all solutes than ethylammonium and npropylammonium nitrates from hexane. Isolated compounds can be effectively recovered from the ionic liquids by back-extraction into an organic solvent after dilution with water or buffer. When gas chromatography is used for the determination step, extractive derivatization (e.g., alkylation, acylation, and silylation) is the preferred technique, avoiding the accumulation of ionic liquid on the column, resulting in poor detector baseline stability [39]. A number of applications were presented for

DK4478_book.fm Page 95 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

95

TABLE 2.3 Partition Coefficients for Organic Compounds in Biphasic Systems Containing Ethylammonium and n-Propylammonium Nitrate and Din-Propylammonium Thiocyanate Ionic Liquids and Organic Solventsa Ionic Liquidb EAN

Organic Solvent n-Hexane

Toluene

Octanolc

Dichloromethaned

PAN

Hexane

Toluenee

Solute Naphthalene 1-Nitronaphthalene 4-Ethylnitrobenzene N-Methyl-2-nitroaniline N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline N,N-Diethyl-4-nitroaniline 2-Naphthol Pyridine 1-Nitronaphthalene 4-Ethylnitrobenzene N-Methyl-2-nitroaniline N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline N,N-Diethyl-4-nitroaniline 2-Naphthol Pyridine 1-Nitronaphthalene 4-Ethylnitrobenzene N-Methyl-2-nitroaniline N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline N,N-Diethyl-4-nitroaniline 2-Naphthol Pyridine 4-Ethylnitrobenzene N-Methyl-2-nitroaniline N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline N,N-Diethyl-4-nitroaniline 2-Naphthol Naphthalene 1-Nitronaphthalene N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline 2-Naphthol Pyridine 1-Nitronaphthalene N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline 2-Naphthol Pyridine

Partition Coefficient 0.09 0.71 0.27 1.73 7.47 4.00 1.37 28.8 6.16 0.05 0.03 0.15 1.64 0.09 0.07 3.93 0.18 0.22 0.18 0.61 0.85 0.58 0.19 0.19 0.85 0.03 0.07 0.40 0.07 0.03 2.10 0.29 1.72 78.5 9.00 33.0 5.83 0.12 2.61 0.24 8.44 0.14 Continued.

DK4478_book.fm Page 96 Thursday, September 14, 2006 1:35 PM

96

Advances in Chromatography, Volume 45

TABLE 2.3 (Continued) Partition Coefficients for Organic Compounds in Biphasic Systems Containing Ethylammonium and n-Propylammonium Nitrate and Din-Propylammonium Thiocyanate Ionic Liquids and Organic Solventsa Ionic Liquidb

DPAT

Organic Solvent

Solute

Partition Coefficient

Dichloromethanef

Naphthalene 1-Nitronaphthalene N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline 2-Naphthol Pyridine Naphthalene 1-Nitronaphthalene N-Methyl-4-nitrolaniline N,N-Dimethyl-4-nitroaniline 2-Naphthol Pyridine

0.07 0.08 0.35 0.10 2.88 0.67 2.45 9.22 132 18.0 1.67 4.26

Hexane

a

Calculated with the concentration of the solute in the ionic liquid as the numerator. DPAT = di-n-propylammonium thiocyanate; EAN = ethylammonium nitrate; PAN = n-propylammonium nitrate. c Octanol/EAN = 100:11 (v/v). d Dichloromethane/EAN = 100:18 (v/v). e Toluene/PAN = 100:9 (v/v). f Dichloromethane/PAN = 100:29 (v/v). b

the isolation and determination of polar compounds in complex matrices, including shale oil, dust extracts, soap, tobacco condensate, and urine samples. The largest number of partition coefficients are available for the biphasic systems formed by 1-butyl-3-methylimidazolium hexafluorophosphate and water or heptane (Table 2.4) [1,9,25,41,42]. In addition, Liu et al. [42] list partition coefficients for 15 polycyclic aromatic hydrocarbons in the 1-octyl-3-methylimidazolium hexafluorophosphate-water biphasic system and Berthod and Carda-Broch [33] list partition coefficients for some simple aromatic compounds in the biphasic system formed by 1-butyl3-methylimidazolium hexafluorophosphate-acetonitrile-water. These partition coefficients were determined by classical shake flask or slow stir methods and by countercurrent chromatography. N-alkylisoquinolinium hexafluorophosphate ionic liquids (alkyl = octyl and tetradecyl) were shown to be more effective than 1-butyl-3methylimidazolium hexafluorophosphate at extracting simple aromatic compounds from water [43]. Water-containing 1-butyl-3-methylimidazolium hexafluorophosphate was shown to form 1-butyl-3-methylimidazolium fluoride on heating [44]. This reaction should be taken into account when considering methods to recover isolated compounds from this, and probably other, hexafluorophosphate-containing ionic liquids [44]. The partition coefficients in Table 2.4 are for the neutral form of ionizable compounds. Additional data for the distribution constants of partially and fully

DK4478_book.fm Page 97 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

97

TABLE 2.4 Partition Coefficients for Biphasic Systems Formed by 1-Butyl-3Methylimidazolium Hexafluorophosphate and Water or n-Heptane log P Counter Solvent Solute

Water

Acenaphthene N-Acetyl cysteine 2-Aminobenzamide 3-Aminophenol Aniline Anthracene Ascorbic acid Benz[a]anthracene Benzamide Benzene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene Benzoic acid 4-Bromoaniline 2-Bromobenzoic acid 1-Bromo-2-nitrobenzene 4-Bromophenol Caffeic acid Catechol 2-Chlorobenzoic acid 3-Chlorophenol 4-Chlorophenol Chrysene 2,3-Diaminotoluene Dibenzo[a,h]anthracene 2,3-Dihydroxybenzoic acid 2,5-Dihydroxybenzoic acid 3,4-Dihydroxybenzoic acid 3,5-Dinitrobenzoic acid 2,4-Dinitrophenol 2,6-Dinitrophenol 1,3-Diphenylenediamine Ferulic acid Fluoranthene Fluorene 2-Fluorophenol Glutathione 4-Hydroxybenzoic acid Indeno[1,2,3cd]pyrene

3.71 –0.155 0.851 0.613 1.568 4.15 –2.097 4.22 0.681 2.146 4.31 4.30 4.32 1.724 2.301 1.322 2.633 1.940 0.301 0.613 1.176 2.013 1.859 4.22 1.114 4.31 0.342 0.000 –0.244 1.580 1.431 1.481 0.792 1.146 4.20 3.82 1.398 –0.921 0.362 4.36

n-Heptane

Solute Descriptors E

S

1.604

1.05

1.556

1.130 0.955 2.290

2.079 –0.194

1.613 1.633 1.929 0.959 2.114

1.973 1.881

A

B

V

0

0.22

1.259

1.15 0.96 1.34

0.65 0.26 0

0.79 0.41 0.28

0.875 0.816 1.454

0.990 0.610

1.50 0.52

0.49 0

0.67 0.14

0.973 0.716

0.730 1.190

0.90 1.19

0.59 0.31

0.40 0.35

0.932 0.991

1.180 1.080

1.32 1.17

0 0.67

0.26 0.20

1.066 0.950

0.970

1.07

0.85

0.52

0.834

0.909 0.915 3.027

1.06 1.08 1.73

0.69 0.67 0

0.15 0.20 0.33

0.897 0.898 1.823

1.260 1.220

1.40 1.62

0.05 0.05

0.54 0.59

1.124 1.124

2.377 1.604 0.660

1.55 1.04 0.69

0 0 0.61

0.24 0.20 0.26

1.357 1.259 0.792

0.930

0.92

0.87

0.53

0.990

2.079

1.944

1.447

Continued.

DK4478_book.fm Page 98 Thursday, September 14, 2006 1:35 PM

98

Advances in Chromatography, Volume 45

TABLE 2.4 (Continued) Partition Coefficients for Biphasic Systems Formed by 1-Butyl-3Methylimidazolium Hexafluorophosphate and Water or n-Heptane log P Counter Solvent Solute Naphthalene 1-Naphthol 2-Naphthol 3-Nitrobenzoic acid 2-Nitrophenol 4-Nitrophenol Phenanthrene Phenol Phthalic acid Pyrene Pyrocatechol Resorcinol Theaflavine Toluene 4-Toluic acid 4-Toluidine

Water 3.34 1.756 1.989 1.342 2.041 1.538 4.06 1.204 1.053 0.041 4.21 0.845 –0.046 0.000 2.380 1.301 1.544

n-Heptane

2.708 0.653

1.799

Solute Descriptors E

S

A

B

V

1.340 1.520 1.520 0.996 1.015 1.070 2.055 0.805

0.92 1.05 1.08 1.41 1.05 1.72 1.29 0.89

0 0.61 0.61 0.70 0.05 0.82 0 0.60

0.20 0.31 0.40 0.44 0.37 0.26 0.26 0.30

1.085 1.144 1.144 1.106 0.949 0.949 1.454 0.775

2.808

1.71

0

0.28

1.585

0.980

1.00

1.10

0.58

0.834

0.601 0.730 0.923

0.52 0.90 0.95

0 0.60 0.23

0.14 0.38 0.52

0.857 1.073 0.957

2.362

–0.420 1.568 1.398

ionized compounds are available in Carda-Broch et al. [25] and in the biphasic system formed by 1-butyl-3-methylimidazolium hexafluorophosphate-acetonitrilewater in Berthod and Carda-Broch [33]. In general terms, the distribution constants for the neutral form of an ionizable compound are more favorable for extraction by the ionic liquid than the ionic form. Manipulation of the pH of the aqueous phase is an effective means of adjusting selectivity, in agreement with common practice for molecular liquids [25,41,45]. Subsets of the partition coefficient data in Table 2.4 were analyzed by CardaBroch et al. [25] and Abraham et al. [46] using the solvation parameter model. The model, in a form suitable for liquid-liquid extraction, is set out below [47–49]: log P = c + eE + sS + aA + bB + vV,

(2.1)

where log P is the partition coefficient for a varied group of neutral compounds between the ionic liquid and countersolvent. The capital letters (E, S, A, B, V) are solute descriptors, available for more than 4000 compounds, and are required to fit the experimental data (log P) to the model. The complementary properties indicated by the lowercase letters (e, s, a, b, v) are the system constants. These constants characterize the difference in the contribution of cavity formation and intermolecular interactions to solvation in the ionic liquid and countersolvent. Thus, e is an indication of how the difference in electron lone pair interactions between the solvent-

DK4478_book.fm Page 99 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

99

TABLE 2.5 System Constants for the Solvation Parameter Models for Liquid-Liquid Partitioning between 1-Alkyl-3-Methylimidazolium Hexafluorophosphate and Water or Heptane 1-Butyl-3-Methylimidazolium

1-Hexyl-3-Methylimidazolium

System Constant

Water [25]

Water [46]

Heptane [23]

Water [46]

e s a b v c

1.29 –0.73 –0.76 –2.39 0.64 1.79

0.45 0.23 –1.76 –1.83 2.15 –0.17

3.28 –0.75 2.77 2.46 –2.80 –0.10

0.05 0.40 –1.48 –2.11 2.30 –0.13

a

a Model for the full data indicated in the table: e = 0.93 (± 0.78); s = 1.00 (± 0.60); a = 2.24 (± 0.32); b = 0.004 (± 0.02); v = 0.48 (± 0.89); and c = 0.87 (± 0.78). Multiple correlation coefficient = 0.972, standard error = 0.25, Fischer statistic = 31, and number of solutes = 15.

saturated ionic liquid and ionic liquid-saturated solvent contributes to extraction; s indicates the contribution of the difference in dipole-type interactions; a (hydrogenbond basicity) and b (hydrogen-bond acidity) indicate contributions to the difference in hydrogen-bonding interactions; and v is the difference in cohesion between the two phases. The c term is an equation constant and is not assigned any chemical meaning. The system constants for the 1-butyl-3-methylimidazolium hexafluorophosphate-water, 1-hexyl-3-methylimidazolium hexafluorophosphate-water, and 1butyl-3-methylimidazolium hexafluorophosphate-heptane biphasic systems are summarized in Table 2.5 [25,46]. The agreement for the 1-butyl-3-methylimidazolium hexafluorophosphate-water system is poor. The system constants from Blanchard and Brennecke [25] are difficult to rationalize in chemical terms and are based on a small dataset (12 compounds) with high cross-correlation between the E and S solute descriptors (r = 0.85) and a relatively poor model fit (overall multiple correlation coefficient ρ = 0.954). A larger number of partition coefficients with known solute descriptors are now available, as indicated in Table 2.4, and the system constants for the 1-butyl-3-methylimidazolium hexafluorophosphate-water system can be recalculated, as indicated in Equation 2.2: log P = 1.19 (± 0.87) + 0.81 (± 0.51)E – 0.76 (± 0.43)S – 1.58 (± 0.35)A – 0.01 (± 0.05)B + 1.29 (± 1.24)V.

(2.2)

The multiple correlation coefficient for this model is 0.913, the standard error of the estimate is 0.53, the Fischer statistic is 27, and the number of solutes is 33. This model is not very good either chemically or statistically. Since we know that the solute descriptors are not in large error, the problem must be errors in the experimental log P values. In which case, the data in Table 2.4 should be used with caution. The model proposed by Abraham et al. [46] makes chemical sense, and is based on

DK4478_book.fm Page 100 Thursday, September 14, 2006 1:35 PM

100

Advances in Chromatography, Volume 45

a set of partition coefficients determined by a single group, but only published in the form of a figure, so they cannot be compared directly with the values in Table 2.4. Interpretation of Abraham et al.’s model indicates that 1-butyl-3-methylimidazolium hexafluorophosphate saturated with water is less cohesive and hydrogenbond acidic and basic than water, but slightly more dipolar/polarizable and has a greater affinity for electron lone pair interactions. Solute properties that favor extraction into the ionic liquid phase are increasing size and dipolarity/polarizability (V, E, and S). Solute properties that diminish extraction by the ionic liquid are the capacity of the solute for hydrogen-bonding interactions (A and B). The model for partitioning of neutral solutes in the 1-butyl-3-methylimidazoliumheptane system [25] exhibits the same problems as the ionic liquid-water system, and must be considered suspect. The indication that the heptane phase is more dipolar/polarizable than the ionic liquid is difficult to defend. The addition of further data from Table 2.4 to those used originally results in a more sensible chemical model, but with poor statistics. To obtain a model with the confidence needed for reliable interpretation, we must await the availability of higher quality partition coefficients.

B. LIQUID-PHASE MICROEXTRACTION Ionic liquids are suitable for use as suspended droplets for headspace and directimmersion extraction from aqueous samples [11,50–54] and as fiber-supported films for solid-phase microextraction [55]. The absence of vapor pressure, adequate viscosity, the ability to dissolve a wide range of compounds, and the availability of ionic liquids immiscible with water are favorable properties for these applications. Enrichment factors of about 5 to 175 for a wide range of compounds (e.g., benzene, toluene, ethylbenzene, and xylene (BTEX), polycyclic aromatic hydrocarbons, phthalate esters, phenols, amines, herbicides, organometallics) have been obtained using the ionic liquids 1-alkyl-3-methylimidazolium hexafluorophosphates (alkyl = butyl, hexyl, or octyl) and drop sizes of 3 to 5 μl. Extractions by liquid-phase microextraction usually occur under nonequilibrium conditions and enrichment factors depend in a complex way on the sampling time, partition coefficients, temperature, analyte diffusion coefficients, and the viscosity of both the ionic liquid and sample solution. The relatively high viscosity of ionic liquids allows larger drop sizes to be used compared with typical organic solvents. This is expected to increase the extraction yield, but in practice, this advantage is often offset to a large extent by slow mass transfer of analytes at the droplet interface. The studies described above are largely exploratory, but it is clear that the enrichment factors obtained are generally inadequate for trace analysis. Ionic liquids can be injected onto reverse-phase liquid chromatographic columns without problems. For gas chromatography, direct injection is not feasible because ionic liquids are nonvolatile. This difficulty can be circumvented by using the ionic liquid as a fiber-supported film coating [55]. Some ionic liquids are sufficiently stable to resist drainage from the fiber support during analyte recovery by thermal desorption. However, the thin film of ionic liquid limits sample detectability compared with conventional immobilized poly(siloxane)-coated fibers.

DK4478_book.fm Page 101 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

101

C. SUPPORTED LIQUID MEMBRANES Supported liquid membranes are prepared from a thin porous membrane with an ionic liquid immobilized inside the porous structure. Transport of material from the sample side of the membrane (donor) to the other side (acceptor) is mediated by the properties of the ionic liquid and the membrane. It is important that the ionic liquid has low solubility in the sample and acceptor phases to ensure stability and longterm performance. 1-Alkyl-3-methylimidazolium (alkyl = butyl, octyl, decyl) hexafluorophosphate and 1-butyl-3-methylimidazolium tetrafluoroborate have been used as the immobilized liquid in various polymeric membranes for the separation of amines and amino acid esters in their neutral and ionized form [56–60]. The transport of organic compounds through the membrane seems to involve two mechanisms [59,60]: an initial phase in which transport is largely regulated by the selectivity of the ionic liquid toward each solute, and a secondary phase mediated by the formation of water microenvironments. Since the solubility of water in the ionic liquids is not negligible (even though their mixtures form biphasic systems), over the course of time, water microenvironments are formed by diffusion of water into the ionic liquid. The mobility of these water microenvironments eventually dominates the transport mechanism, with a complete loss of selectivity. This is unfavorable for selective extraction, and if these preliminary studies are confirmed as a general transport mechanism, then ionic liquids will have a limited role in supported liquid membrane extraction. Miyako et al. [58] presented an interesting example of the use of an ionic liquid-supported membrane for the separation of the enantiomers of ibuprofen. These authors used an enzyme to selectively methylate (S)-ibuprofen in the presence of (R)-ibuprofen, the (S)-ibuprofen methyl ester was then selectively transported through an ionic liquid-supported membrane to the acceptor phase, where the methyl ester was hydrolyzed by a different enzyme to release the (S)-ibuprofen. Since (S)-ibuprofen is more hydrophilic than its methyl ester, it is not transported back through the supported ionic liquid membrane. This method has possibilities for the isolation of pure enantiomers on a preparative scale. Since ionic liquids are nonvolatile, they do not add any contamination to a gas stream. This quality gives ionic liquids an innate advantage over traditional solvents used for absorbing gases. A porous alumina membrane saturated with 1-butyl-3methylimidazolium bis(trifluoromethylsulfonyl)amide demonstrated a selective permeability for carbon dioxide over nitrogen of 127 [61]. It was speculated that with further technical development, ionic liquid-supported membranes would be competitive with amine scrubbers for the recovery of carbon dioxide from flue gas streams.

D. EXTRACTION

OF INORGANIC

COMPOUNDS

Common ionic liquids are noncoordinating solvents. In the absence of a complexing agent to facilitate the extraction of metal ions, ionic liquids do not exhibit useful extraction properties for metal ions from aqueous solution [62]. In this regard, they behave like conventional molecular solvents used for metal extractions. There is no evidence to date for facilitated extraction through ion exchange where metal ions are selectively attracted into the ionic liquid with displacement of the ionic liquid

DK4478_book.fm Page 102 Thursday, September 14, 2006 1:35 PM

102

Advances in Chromatography, Volume 45

cation to the aqueous phase. The purpose of the complexing agent is to dehydrate the metal ions and to offer a more hydrophobic environment to facilitate solvation of the metal ions by the ionic liquid phase. To ensure complete transfer of metal ions from the aqueous phase, the complexing agent should reside preferentially in the ionic liquid with only limited distribution to the aqueous phase. A number of complexing agents in the biphasic systems containing 1-butyl-3-methylimidazolium or 1-octyl-3-methylimidazolium hexafluorophosphate or bis(trifluoromethylsulfonyl)amide and water adjusted to a suitable pH have been shown to fulfill this condition. The applications reported include extraction of a variety of transition metals using 1-pyridylazo-2-naphthol (PAN) and 1-thiazolylazo-2-naphthol (TAN) [62]; extraction of Cd2+, Hg2+, Pb2+, Zn2+, Cu2+, and Ag+ by dithizone [63]; extraction of actinides by octyl(phenyl)-N,N-diisobutylcarbamoylmethyl phosphine oxide (CMPO) [64–66]; extraction of group I and II metals by crown ethers [67–70]; and extraction of group I and II metals by calix[4]arenes [71,72]. The extraction of metal ions by 1-pyridylazo-2-naphthol, 1-thiazolylazo-2-naphthol, and dithizone shows the expected pH dependence typical of extractions with conventional solvents [62,63]. The recovery of metal ions from dithizone complexes dissolved in ionic liquids by back extraction into an aqueous solution of suitable pH has been demonstrated. Ionic liquids containing octyl(phenyl)-N,N-diisobutylcarbamoylmethylphosphine oxide in the presence of tri-n-butyl phosphate show an unusually large extraction efficiency for lanthanides, uranyl complexes, and 137Cs and 90Sr compared with conventional water-immiscible organic solvents [64–66]. Octyl(phenyl)-N,N-diisobutylcarbamoylmethylphosphine oxide is a neutral complexing agent and requires an anion such as nitrate to transfer metal ions into low polarity organic solvents. A useful practical advantage of ionic liquids is that the high extraction efficiency extends to much lower concentrations of nitric acid compared with conventional solvents. Addition of tri-n-butyl phosphate to nitric acid solutions containing Sr2+ was shown to markedly increase the extraction of Sr2+ by dicyclohexano-18-crown-6 into 1alkyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amides [70]. This is accounted for by adduct formation between strontium, the crown ether, and tri-nbutyl phosphate, resulting in enhanced extractability by the ionic liquids. Increasing the alkyl chain length of the 1,3-dialkylimidazolium cation reduces this enhancement effect. The selectivity order for the extraction of group I and II metal ions into 1alkyl-3-methylimidazolium hexafluorophosphate by dicyclohexano-18-crown-6 was K+ > Rb+ > Cs+ > Na+ > Li+ [69]. This order reflects the relative complexing ability of the crown ether for alkali metal cations and is similar to results for conventional solvents. Solutions of calix[4]arene-bis(tert-octylbenzo-crown-6) in 1-alkyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amides provide efficient extraction of Cs+ ions from aqueous solution under conditions that gave negligible extraction with traditional organic solvents [71]. Pyridinocalix[4]arene in 1-alkyl-3-methylimidazolium hexafluorophosphate demonstrated enhanced extraction efficiency for Ag+ compared to chloroform and high selectivity over Cu2+, Zn2+, Co2+, and Ni2+[72]. Task-specific ionic liquids containing both ionic and complexing groups in a single structure have been developed for the extraction of metal ions [73,74]. Attaching a metal ion coordinating group directly to the imidazolium cation makes the complexing agent an integral part of the hydrophobic phase and greatly diminishes

DK4478_book.fm Page 103 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

103

the chance for loss to the aqueous phase. Ionic liquids containing imidazolium groups and thiol, thiourea, or urea substituents with hexafluorophosphate counterions were shown to be effective for the extraction of Hg2+ and Cd2+ ions from aqueous solution when used alone as the hydrophobic phase or as a mixture with a conventional ionic liquid. An ethylene-glycol functionalized bis-imidazolium ionic liquid with bis(trifluoromethylsulfonyl)amide counterions showed high selectivity for the extraction of Hg2+ in the presence of Cs+ [74]. Mercury could be recovered from the ionic liquid by changing the pH of the aqueous phase. Extending the ethylene glycol monomer chain length or introducing other chelating atoms (e.g., sulfur or nitrogen) might extend the range of task-specific ionic liquids for metal extractions.

E. SUPERCRITICAL FLUID EXTRACTION

FROM IONIC

LIQUIDS

While distillation remains a reasonable option for the recovery of volatile compounds from ionic liquids, supercritical fluid extraction is an attractive alternative for the recovery of thermally labile compounds and offers a suitable alternative to liquidliquid extraction [74,75]. There are several reports providing details of the solubility of gases and simple organic compounds as a function of temperature and pressure in 1-alkyl-3-methylimidazolium ionic liquids with hexafluorophosphate, tetrafluoroborate, nitrate, and ethylsulfate anions and for N-butylpyridinium tetrafluoroborate [75,77–79]. A phase diagram for the binary system of fluoroform and 1-ethyl-3methylimidazolium hexafluorophosphate is also available [80]. For product recovery from ionic liquids, most attention has focused on the use of supercritical fluid carbon dioxide [75–78]. The phase behavior of carbon dioxide with ionic liquids differs significantly from molecular liquids. While large amounts of carbon dioxide dissolve into the ionic liquids, this occurs without appreciable expansion of the liquid. It is assumed that this is due to the strong Coulombic forces between ions that resist changes to the ion spacing. Once the void pace within the ionic liquid is saturated, even under very high pressure, no more carbon dioxide can dissolve into the ionic liquid. On account of this unique behavior, ionic liquid–carbon dioxide systems remain two-phase, even at high pressures. Although large quantities of carbon dioxide dissolve in the ionic liquid-rich phase, reducing the ionic liquid viscosity, virtually no ionic liquid dissolves in the carbon dioxide-rich phase. This allows for product recovery by fluid expansion with minimal contamination by the ionic liquid. The lack of any appreciable solubility of ionic liquids in the carbon dioxide phase can be attributed to the extremely low vapor pressure of ionic liquids and the inability of carbon dioxide to adequately solvate ions in the gas phase. The nonextractability of ionic liquids by carbon dioxide was demonstrated in early studies of supercritical fluid chromatography with silica-supported ionic liquids as the stationary phase [81]. In a study of 20 various model compounds dissolved in 1-butyl-3-methylimidazolium hexafluorophosphate, it was shown that more than 95% recovery was possible for all compounds using supercritical fluid carbon dioxide [75]. As would be expected, compounds requiring the most carbon dioxide for extraction were those with the lowest volatility and strongest interactions with the ionic liquid. It should be noted, however, that small amounts of water in the ionic liquid dramatically reduce the solubility of carbon dioxide and can shift the pH toward the acid side, both factors

DK4478_book.fm Page 104 Thursday, September 14, 2006 1:35 PM

104

Advances in Chromatography, Volume 45

being potential problems for extracting real samples [77]. Also, it has been shown that open-tubular columns coated with 1-butyl-3-methylimidazolium hexafluorophosphate can be used to determine partition coefficients with supercritical fluid carbon dioxide over a wide range of temperature and density conditions [82].

III. LIQUID CHROMATOGRAPHY Ionic liquids are not widely used in liquid chromatography [38–40,83–85]. They are generally too viscose to use as mobile phases, typically one to three orders of magnitude higher than desirable. In addition, they are more expensive than traditional solvents and often exhibit poor transmittance of low-wavelength ultraviolet (UV) light (sometimes due to the presence of impurities). The viscosity of some ionic liquids can be reduced to manageable ranges by dilution with water or an organic solvent and by increasing the column temperature. Aqueous solutions of alkylammonium thiocyanates corrode stainless steel surfaces as well as demonstrating poor baseline stability [38]. The aqueous solutions of tetra-n-alkylammonium sulfonates are strongly basic and rapidly degrade silica-based column packings [84]. The aqueous and organic solutions of alkylammonium nitrate [38,39,83] and acetate [85] ionic liquids are generally compatible with silica-based column packings. Except for hydrogen-bond acids, most polar compounds are not eluted from typical reversephase columns by alkylammonium nitrate-water mobile phases. For hydrogen-bond acids, retention factors (log k) are usually a linear function of the volume fraction of ethylammonium or n-propylammonium nitrates for the composition range of 20 to 60% (v/v) ionic liquid [38]. Ethylammonium nitrate has a similar solvent strength to tetrahydrofuran for hydrogen-bond acids, while n-propylammonium nitrate is even stronger. A typical separation of a mixture of hydrogen-bond acids is shown in Figure 2.2. The large change in relative retention for 2,4-dinitrophenol in changing the mobile phase from methanol-water to n-propylammonium nitrate-water is particularly notable. The solvent strength of ethylammonium acetate is indicated as similar to methanol, with a wider range of potential applications when compared with the alkylammonium nitrate ionic liquids [85]. The alkylammonium nitrate and acetate ionic liquids can be used as strengthadjusting solvents with miscible organic solvents (e.g., methanol, ethanol, 2-propanol, acetonitrile, tetrahydrofuran, etc.) [83]. This facilitates elution of a wider range of polar and hydrogen-bond acid compounds from chemically bonded column packings. Except for 2-propanol, most solvent modifiers exhibit a linear change in the retention factor (log k) with the volume fraction of organic solvent for the composition range of 50% to 90% (v/v) organic solvent. The solvent strength of organic solvents in alkylammonium nitrate solutions is generally different from their values in water. Liquid-liquid chromatographic separations have been performed with alkylammonium nitrate ionic liquids coated on silica supports with mobile phases containing various amounts of chlorinated hydrocarbons or ethers in hexane [39,40]. The columns were prepared by the dynamic coating technique and were stable for many hours of operation using mobile phases saturated with the ionic liquid. One appli-

DK4478_book.fm Page 105 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

105

4 2 1 4 1 5

2 3

5

3 (a)

(b)

FIGURE 2.2 Separation of hydrogen-bond acids by reverse-phase chromatography on a 25 cm × 1 mm inside diameter column packed with Spherisorb ODS 1 (10 μm) (Waters Corp., Milford, MA) with a mobile phase flow rate of 50 μl/min. The mobile phase was (a) watermethanol (3:2) and (b) water-n-propylammonium nitrate (3:2). Identification: 1 = anthranilic acid; 2 = 2,4-dinitrophenol; 3 = 1,2-diamino-4-nitrobenzene; 4 = 4-nitroaniline; and 5 = 3nitroaniline. (From Shetty, P.H., Youngberg, P.J., Kersten, B.R., and Poole, C.F., J. Chromatogr., 411, 61, 1987. With permission.)

cation of these columns is the facile determination of liquid-liquid partition coefficients. Coated columns have also been used for the separation of structural isomers (Figure 2.3) [39]. Ionic liquids at concentrations of 2 to 60 mM have been used as mobile phase additives for the separation of amines on chemically bonded stationary phases [86–89]. In this instance, it is unclear whether the fact that the additives are ionic liquids is of importance. Weak bases and ammonium salts, in general, are commonly used as mobile phase additives for the same purpose and are just as effective. 1Alkyl-3-methylimidazolium ionic liquids are assumed to be adsorbed onto the stationary phase surface, suppressing the activity of free silanol groups. This results in improved peak shapes and decreased retention for amines and increased retention for acids. In addition, the repulsion between the imidazolium cation and ionized amines may play a role in minimizing silanol-group interactions. Ionic liquids are also useful deactivating agents for the separation of peptides by normal-phase thinlayer chromatography [90]. Liquid chromatographic techniques are suitable for detecting impurities in ionic liquids. Ion exchange chromatography with suppressed conductivity detection (ion chromatography) is suitable for detecting halide impurities at low parts per million concentrations in ionic liquids [91]. Reverse-phase liquid chromatography can be used to determine UV-absorbing impurities in 1,3-dialkylimidazolium-based ionic liquids [92,93]. Cation exchange chromatography provides higher selectivity for the separation of short-chain 1,3-dialkylimidazolium and N-alkylpyridinium cations

DK4478_book.fm Page 106 Thursday, September 14, 2006 1:35 PM

106

Advances in Chromatography, Volume 45

5 6 7

1

2

3 4

(a)

(b)

(c)

FIGURE 2.3 Separation of structural isomers by liquid-liquid chromatography using an ionic liquid as the stationary phase. The column was 20 cm × 1 mm packed with LiChrospher Si 100, 10 μm particle size (Merck, Darmstadt, Germany). The ionic liquid was coated on the silica support by passing a solution of 25% (v/v) n-propylammonium nitrate solution in acetonitrile through the column overnight (at least 50 column volumes) at a flow rate of 5 μl/min. The acetonitrile was then washed out of the column with methylene chloride (about 3 h at 50 μl/min) and the column equilibrated with the mobile phase for the separation. The volume of ionic liquid adsorbed by the support was typically 0.035 ml and the phase ratio about 0.2. The mobile phase was (a,c) hexane-methyl t-butyl ether (4:1) and (b) hexanemethylene chloride (4:1). The mobile phase was saturated with n-propylammonium nitrate before use. Identification: 1 = 2-nitrophenol; 2 = 4-nitrophenol; 3 = 1-naphthol; 4 = 2-naphthol; 5 = norethindrone acetate; 6 = norgestrel; and 7 = norethindrone.

than reverse-phase chromatography. Ion-pair chromatography is preferred for more hydrophobic cations [93]. Cation exchange sorbents are suitable for the isolation of ionic liquid cations by solid-phase extraction [94].

IV. ELECTROPHORESIS Electrophoresis in fused inorganic salts and their eutectic mixtures was developed in the 1960s for the separation of inorganic ions and isotopes [95–97]. Later, electrophoresis in ionic liquids was used to study fundamental aspects of ion solvation, transport, and complex formation in the absence of competition from water molecules. Electrophoresis in inorganic ionic liquids requires a special apparatus that can tolerate the high temperatures and corrosive nature of these solvents and their electrolysis products. Separations were generally performed in the flatbed mode using glass fiber or asbestos paper strips as well as powdered thin-layer supports. As a separation method, inorganic ionic liquids are virtually

DK4478_book.fm Page 107 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

107

unused today, while interest in organic ionic liquids, largely in capillary electrophoresis, is just beginning.

A. CAPILLARY ELECTROPHORESIS High viscosity, variable purity, and interference in absorbance detection at low wavelengths are the main problems inhibiting the wider use of organic ionic liquids in capillary electrophoresis at present. Ionic liquids have been used as background electrolytes in aqueous [98–100] and nonaqueous [101–104] capillary electrophoresis. The most common cosolvents are water, acetonitrile, and mixtures of acetonitrile and methanol. Ionic liquids are promising candidates for use as background electrolytes, where dilution with a cosolvent reduces their viscosity. The addition of an ionic liquid to organic solvents changes the electroosmotic flow of the system. The change is most dramatic for acetonitrile and less so for methanol. Changes in the cation have a greater effect on the electroosmotic flow than changes in the anion. In solvents of low polarity, a lack of free ions diminishes the effectiveness of ionic liquids as background electrolytes. 1,3-Dialkylimidazolium-based ionic liquids are adsorbed onto the capillary wall from an aqueous solution. In so doing, they reverse the direction of the electroosmotic flow, more or less independent of pH [98,105]. Ionic liquids can be covalently bonded to the capillary wall in a several-step process that achieves the same result as physical coating [98,106]. A column of this type with an electrolyte solution containing 7.5 mM lactic acid, 0.6 mM 18-crown-6, 12 mM α-cyclodextrin, and 100 mM 1-hexyl3-methylimidazolium hydroxide to adjust the electrolyte pH to 4 provided baseline separation of eight alkaline and alkaline earth metal ions plus nickel, lead, and ammonium in less than 14 min [106]. A number of different samples, including DNA fragments [107], proteins [105], metal ions [98,106], ionic dyes [102], anthraquinones [100], phenols [99,103], poly(phenols) [108,109], etc., were separated using 1,3-dialkylimidazolium-based ionic liquids as a major component of the electrolyte solution without significant operational difficulties. The separation of poly(phenols) is shown in Figure 2.4 [109]. A contributing factor to this separation is the association of the poly(phenols) with 1,3-dialkylimidazolium cations both at the capillary wall and in solution. The migration order of halophenols in an aqueous electrolyte solution indicates that the 1,3-dialkylimidazolium cations of the ionic liquid are likely associated with the phenols, since they migrate as if carrying a partial positive charge [99]. β-Cyclodextrin was used as an additive to achieve baseline separation of four anthraquinones in a plant extract with a background electrolyte containing 60 mM 1-butyl-3-methylimidazolium tetrafluoroborate in pH 10 buffer [100]. Model proteins were separated in buffers containing 90 mM 1-ethyl3-methylimidazolium tetrafluoroborate [107]. The ionic liquid was responsible for reversing the direction of the electroosmotic flow and repelling basic proteins from the silica surface, improving peak shapes. Ionic liquids were used as modifiers in micellar electrokinetic chromatography with polymeric surfactants for the separation of both chiral and achiral compounds [110]. The main effect of adding from 1 to 20 mM 1-alkyl-3-methylimidazolium tetrafluoroborate ionic liquids to the background electrolyte was adjustment of the

DK4478_book.fm Page 108 Thursday, September 14, 2006 1:35 PM

108

Advances in Chromatography, Volume 45

Absorbance (mAU)

∗

A

∗

B

C

10

21

31

42

FIGURE 2.4 Separation of poly(phenols) using (A) 1-ethyl-3-methylimidazolium tetrafluoroborate, (B) 1-butyl-3-methylimidazolium tetrafluoroborate, and (C) 1-ethyl-3-methylimidazolium hexafluorophosphate. The electrolyte solution contained 150 mM ionic liquid in water. The fused-silica capillary column was 50 cm × 50 μm inside diameter, operating voltage 16 kV, and absorbance detection at the anode end at 240 nm. (From Yanes, E.G., Gratz, S.R., Baldwin, M.J., Robinson, S.E., and Stalcup, A.M., Anal. Chem., 73, 3838, 2001. With permission.)

electroosmotic flow, with little evidence that the ionic liquids modified the selectivity of analyte-surfactant interactions. Capillary electrophoresis is a suitable technique for determining UV-absorbing ionic impurities in ionic liquids [111]. Citrate buffers provide improved separation and peak shapes for 1,3-dialkylimidazolium cations. A pH 4 citrate buffer of 200 mM concentration was indicated as optimum with respect to peak shapes and resolution of 1,3-dialkylimidazolium cations. High buffer concentrations likely diminish interactions with the silica surface, thus reducing tailing. For 1,3-dialkylimidazolium cations, ion mobility was shown to be a linear function of the mass of the cation, regardless of the type of alkyl or aryl substitution. Consequently cations with different structures, but the same mass, are not separated. Capillary electrophoresis was used to detect chloride ion impurities in 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide and bromide impurities in 1-butyl-3methylimidazolium hexafluorophosphate [112]. The background electrolyte consisted of 2-amino-2-methyl-1,3-propanediol and sodium dichromate of various concentrations, but with a constant 6:5 molar ratio in water. The halide ions are detected by indirect absorption at 254 nm with sodium dichromate as the indicator electrolyte. Using both electrokinetic injection and the water plug method for online concentration, quantification of chloride and bromide ions at the parts per million level was achieved.

DK4478_book.fm Page 109 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

109

V. GAS CHROMATOGRAPHY Early studies of ionic liquids as stationary phases in gas chromatography are reviewed in Poole et al. [113]. In addition, gas chromatography is widely used to determine the physicochemical properties of ionic liquids, reviewed in Poole [8] and Poole et al. [114]. These articles contain a large amount of data useful for the selection of ionic liquids for particular applications. In this section, studies that extend our understanding of the characteristic properties of ionic liquids as stationary phases for gas chromatography are highlighted. Low melting point, high thermal stability, the absence of significant vapor pressure, the capability of wetting inorganic oxide supports and glass surfaces, and the ability to dissolve a wide range of organic compounds are favorable properties of ionic liquids for use in gas chromatography. Some examples of ionic liquids with a useful temperature operating range are summarized in Table 2.6. In most cases, the lower operating temperature is indicated as the melting point or room temperature and the upper temperature limit by some variant of the baseline rise or temperature equilibrium method determined by gas chromatography [114]. Only a fraction of known ionic liquids have been used as stationary phases for gas chromatography and the data in Table 2.6 are no more than representative of a much larger pool of potential ionic liquids available for study. As well as favorable operating temperature ranges, most ionic liquids demonstrate acceptable kinetic properties for separations, as indicated by similar plate numbers to conventional nonionic stationary phases.

A. RETENTION MECHANISMS The dominant retention mechanism for most organic compounds on ionic liquids is gas-liquid partitioning [114–123]. Adsorption at the gas-liquid interface is important for compounds of low solubility in ionic liquids, such as hydrocarbons. In a few cases, for example, 1-methyl-3-ethylimidazolium chloride, ethylpyridinium bromide, and triethanolammonium thiocyanate interfacial adsorption is the dominant retention mechanism for most compounds [115,116]. When determining physicochemical properties of ionic liquids by gas chromatography, it is necessary to first establish the retention mechanism, otherwise the properties deduced from experimental gas-liquid partition coefficients (or retention factors, retention indexes, etc.) will result in incorrect and system-dependent property estimates. It is relatively straightforward, if somewhat time consuming, to determine gas-liquid partition coefficients from systems exhibiting a mixed retention mechanism [8,114,121,124]. For compounds retained by a mixed retention mechanism, the relative contribution of interfacial adsorption and gas-liquid partitioning is usually temperature dependent. At higher temperatures, gas-liquid partitioning is favored at the expense of interfacial adsorption, which diminishes and may fall to zero. An application that exploits the difference in retention mechanism for n-alkanes and n-alkylaromatic compounds on an ionic liquid stationary phase is the packed column separation of aromatic compounds in gasoline on tetra-n-ethylammonium 4-toluenesulfonate [125]. The primary requirement for this separation is the elution of benzene after dodecane. Benzene and the alkylbenzenes present in gasoline are retained predominantly by gas-liquid

DK4478_book.fm Page 110 Thursday, September 14, 2006 1:35 PM

110

Advances in Chromatography, Volume 45

TABLE 2.6 Temperature Operating Range for Some Representative Ionic Liquids Used in Gas Chromatography Temperature Operating Range (°C) Ionic Liquid Tetra-n-butylammonium Perfluorooctanesulfonate Tris(hydroxymethyl)methylamino-2-hydroxy-1propanesulfonate 4-Morpholinepropanesulfonate Octanesulfonate Perfluorobenzenesulfonate 2-[Bis(2-hydroxyethyl)amino]ethanesulfonate 4-Toluenesulfonate 3-(Cyclohexylamino)propanesulfonate Benzenesulfonate Tetrafluoroborate Trifluoromethanesulfonate Picrate Butanesulfonate Methanesulfonate Tris(hydroxymethyl)methylaminopropanesulfonate Tetra-n-butylphosphonium 4-Toluenesulfonate Chloride 1-Butyl-3-methylimidazolium Bis(trifluoromethylsulfonyl)amide Hexafluorophosphate Trifluoromethanesulfonate Chloride 1-Benzyl-3-methylimidazolium Trifluoromethanesulfonate 1-(4-Methoxyphenyl)-3-methylimidazolium Trifluoromethanesulfonate 4-Toluenesulfonate Tetra-n-pentylammonium Tri-n-propylammonium Tetra-n-ethylammonium Di-n-ethylammonium n-Ethylammonium Tri-n-ethylammonium Tri-n-butylammonium Trimethylammonium a

r.t. = salt is a liquid at room temperature.

Lower Limita

Upper Limit

Liquid Range (°C)

r.t. r.t.

220 180

> 200 > 160

r.t. r.t. 51 r.t. 55 r.t. 78 162 112 90 50 79 110

180 180 210 170 200 160 210 290 240 200 160 180 210

> 160 > 160 159 > 150 145 140 132 128 128 110 110 101 100

44 83

230 230

186 147

r.t. 10 16 r.t.

185 170 175 145

> 165 160 156 > 125

r.t.

220

> 200

r.t.

250

> 230

55 75 85 105 121 78 82 93

190 180 190 210 225 180 180 190

135 105 105 105 104 102 98 97

DK4478_book.fm Page 111 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

111

3

1

4

2

5 6

0

5

FIGURE 2.5 Separation of aromatic compounds in premium unleaded gasoline on a 3.5 m × 2 mm inside diameter column of 25% (w/w) tetra-n-ethylammonium 4-toluenesulfonate on Chromosorb P-AW (Celite Corp., Lompoc, CA) at 140°C with a nitrogen carrier gas flow rate of 15 ml/min. Identification: 1 = n-alkanes ≤ C-14; 2 = benzene; 3 = toluene; 4 = pxylene and ethylbenzene; 5 = o-xylene and n-propylbenzene; and 6 = n-butylbenzene.

partitioning on the ionic liquid, while the n-alkanes are retained largely by interfacial adsorption. The desired separation (Figure 2.5) is achieved through optimizing the separation conditions (column length, temperature, support type, phase loading, and temperature) to enhance the retention of aromatic compounds while minimizing the retention of n-alkanes.

B. CHROMATOGRAPHIC SELECTIVITY The solvation parameter model is probably the most useful tool to assign values for the contribution of individual intermolecular interactions and cavity formation to retention in gas chromatography [48,49]. Compared with liquid-liquid distribution processes, the cavity term, V in Equation (1), is replaced by the new solute descriptor, L in Equation (3), with all other solute descriptors retaining their original meanings and values. Transfer between condensed phases occurs with more or less complete cancellation of dispersion interactions, which is not appropriate for transfer from the gas phase. Since the gas phase is assumed to be ideal, dispersion interactions only occur when the solute is solvated in the stationary phase. This is accounted for by the L solute descriptor, the gas-liquid partition coefficient for the solute on hexadecane at 25°C. The complementary system constant, l, is then equivalent to the contribution of cohesion (cavity formation) and dispersion interactions in the stationary phase. In Equation (3), KL is the gas-liquid partition coefficient, but it can

DK4478_book.fm Page 112 Thursday, September 14, 2006 1:35 PM

112

Advances in Chromatography, Volume 45

be replaced by the retention factor, retention volume, or relative retention (but not the retention index) without affecting the system constant values. The system constants are truly properties of the solvation characteristics of the stationary phase and only the value and interpretation of the equation constant (c) will change: log KL = c + eE + sS + aA + bB + lL.

(2.3)

There are a several relatively large compilations of system constants for ionic liquids. Poole and Poole [126] calculated system constants for 38 ionic liquids (mainly tetra-n-alkylammonium and tetra-n-alkylphosphonium salts) at about 120°C. The dependent variable in this case is the gas-liquid partition coefficient fully corrected for interfacial adsorption. For the same conditions, Kollie and Poole [127] provided system constants for 10 tetra-n-alkylammonium alkanesulfonate and perfluoroalkanesulfonate ionic liquids. This database can be compared directly with a similar collection of system constants for 23 nonionic stationary phases commonly used for gas chromatography [128]. Anderson et al. [129,130] provide system constants for 21 ionic liquids (mainly 1,3-dialkylimidazolium and alkylammonium ionic liquids) at two or three temperatures (40°C, 70°C, and 100°C). The dependent variable in this case is the retention factor and the system constants are not corrected for contributions from interfacial adsorption. A comprehensive review of the above studies is given by Poole [8], together with additional system constants for five alkylammonium nitrate and thiocyanate room temperature ionic liquids at about 80°C, corrected for interfacial adsorption. These studies demonstrate the unique solvent behavior of ionic liquids and justify the current interest in their use in analytical chemistry. All ionic liquids have a significant capacity for dipole-type interactions and are strong hydrogen-bond bases, but usually not strong hydrogen-bond acids. The alkylammonium nitrate and thiocyanate salts [8], 1-ethyl-3-hydroxypyridinium bromide and 4-toluenesulfonate salts [127], tri-n-butylammonium sinapate and α-cyano-4hydroxycinnamate [129], and 1-butyl-3-methylimidazolium and 1-butyl-2,3-dimethylimidazolium bis(trifluoromethylsulfonyl)amide [129] are exceptions. Hydrogenbond acidity is not a common property of nonionic stationary phases, nor for tetran-alkylammonium salts with anions containing hydrogen-bond acid substituents, which are also hydrogen-bond bases (e.g., alcohol and phenol groups). In these cases, it seems that self-association is preferred to solute-solvent hydrogen-bonding interactions where the solvent acts as the hydrogen-bond acid [131]. The weak hydrogen-bond acidity of the ionic liquids is generally associated with cation NH groups (OH groups in the case of the 1-ethyl-3-hydroxypyridinium salts). The typical range of values for the system constants of the ionic liquids and nonionic solvents at 120°C is summarized in Table 2.7. There is a tight grouping of properties, with most ionic liquids having s constants between 1.5 and 2.1 and a constants between 3.0 and 4.0. The exceptions are the tetra-n-alkylammonium salts with pentacyanopropionide, picrate, perfluoroalkanesulfonate, and pentafluorobenzenesulfonate anions (low a and s system constants). They have properties similar to polar nonionic stationary phases. Tetra-n-alkylammonium and tetra-n-alkylphosphonium chloride, bromide, and nitrite salts have larger than average a system

DK4478_book.fm Page 113 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

113

TABLE 2.7 Typical Range for System Constants of Ionic and Nonionic Liquids at 120°C Range System Constant

Ionic Liquids

Nonionic Liquids

e (electron lone pair interactions)a s (dipole-type interactions) a (solvent hydrogen-bond basicity) b (solvent hydrogen-bond acidity) l (cohesion and dispersion interactions)

0.07–0.50 1.4–2.1 1.4–5.4 0 0.26–0.37b 0.44–0.55c

0–0.37 0–2.1 0–2.1 0 0.37–0.58

a b c

Fluorine-containing stationary phases have negative values. Value typical of ionic liquids with associated anions. Value typical of ionic liquids with nonassociated anions.

constants and average or above average s system constants. The hydrogen-bond basicity of ionic liquids is mainly a property of the anion and is influenced primarily by the size and charge localization on the anion. Ions that can delocalize a charge (e.g., picrate, pentacyanopropionide, perfluoroalkanesulfonate, etc.) are weaker hydrogen-bond bases and less dipolar than other ionic liquids. The halide anions have relatively small atomic radii and no mechanism for charge delocalization. They are the most basic of the ionic liquids. The above general trends are illustrated in Figures 2.6 and 2.7 for a series of tetra-n-butylammonium ionic liquids with different anions [125,132,133]. Selectivity judged from the point of view of the anion is affected mostly by dipole-type and solvent hydrogen-bond base interactions. The anions in Figure 2.6 are ranked in terms of increasing solvent hydrogen-bond basicity. Changing the anion type has only a small affect on the retention of solutes with a large dipole moment (e.g., 1nitropropane and pyridine), but a profound effect on the retention of solutes acting as hydrogen-bond acids (e.g., n-butanol). Figure 2.6 illustrates the separation of a varied group of solutes with different functional groups on three similar columns prepared with tetra-n-butylammonium trifluoromethanesulfonate, 4-toluenesulfonate, and methanesulfonate. The dipolar/polarizable solute (pyridine; none of the salts are hydrogen-bond acids) elutes at almost the same time on the three columns, whereas those solutes that act as hydrogen-bond acids (n-butanol and 2methyl-2-propanol) are shifted substantially in good qualitative agreement with the relative basicity of the anions. Varying the type of cation for a series of alkylammonium 4-toluenesulfonate ionic liquids has only a limited affect on selectivity and influences retention primarily by an increase in nonspecific interactions (dispersion and cohesion) (Figure 2.8) [133]. There is a general increase in retention with an increase in the molecular weight of the cation, but hardly any noticeable change in selectivity as cations with N-H groups are replaced by N-R groups.

DK4478_book.fm Page 114 Thursday, September 14, 2006 1:35 PM

114

Advances in Chromatography, Volume 45

PIC PCP FBuS FMS BFS MOPSO BZS SCN PTS SuL NAT MOPS CHES CAPS MES ETS HS BR NO2 Cl 1.5

A

2.5

B

C

3.5 4.5 Free energy of solution

D

5.5

FIGURE 2.6 Variation in the retention of (A) benzene, (B) pyridine, (C) 1-nitropropane, and (D) 1-butanol on tetra-n-butylammonium ionic liquids with different anions at 120°C. Retention of each compound is indicated as its partial molar Gibbs free energy of solution in kilocalories per mole. Identification of anions: PIC = picrate; PCP = pentacyanopropionide; FBuS = perfluorobutanesulfonate; FMS = trifluoromethanesulfonate; BFS = pentafluorobenzenesulfonate; MOPSO = 2-hydroxy-4-morpholinopropanesulfonate; BZS = benzenesulfonate; SCN = thiocyanate; PTS = 4-toluenesulfonate; SuL = sulfamate; NAT = nitrate; MOPS = 4-morpholinopropanesulfonate; CHES = 2-(cyclohexylamino)ethanesulfonate; CAPS = 3-(cyclohexylamino)-1-propanesulfonate; MES = methanesulfonate; ETS = ethanesulfonate; HS = hexanesulfonate; Br = bromide; NO2 = nitrite; and Cl = chloride.

The ionic liquids with nonassociating ions have surprisingly large l system constants, 0.44 to 0.55, compared with polar nonionic solvents and are unique among polar solvents in their ability to separate the members of a homologous series. For anions believed to be associated as hydrogen-bond complexes, the l system constants are significantly smaller, 0.26 to 0.37, and equivalent to the values observed for the most polar of the nonionic liquids (Table 2.7). This unexpected ability to separate nonpolar compounds while demonstrating strong polar interactions with polar solutes has been referred to as the “dual nature selectivity behavior” of ionic liquids [129,130,134]. There is a simple explanation, however, for this unusual behavior. Since ionic liquids are composed entirely of ions, the equilibrium distances between ion shells are controlled primarily by Coulombic forces. These distances are comparatively large when compared with molecular solvents and result in solvents of low cohesion. When either or both ions are capable of strong ion-ion intermolecular interactions, such as association by hydrogen bonding, there is an increase in the

DK4478_book.fm Page 115 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

115

1, 8 3,9 7 10

6 2

A

5 4

1 7, 8 3 B 5 9 10

6 4

2

7, 8 1

3 C 9 10

5

4 6 2

0

10

5

15

Min

FIGURE 2.7 Variation of retention for a varied group of compounds with different functional groups on three matched columns of (A) tetra-n-butylammonium trifluoromethanesulfonate, (B) 4-toluenesulfonate, and (C) methanesulfonate. Each column was 3.5 m × 2 mm inside diameter packed with 10% (w/w) of ionic liquid on Chromosorb W-AW (Supelco, Inc., Bellefonte, PA). Temperature 120°C and nitrogen flow rate 15 ml/min. The ionic liquids are arranged in order of increasing basicity of the anion determined by the solvation parameter model. Identification: 1 = benzene; 2 = n-butanol; 3 = n-pentanone; 4 = 1-nitropropane; 5 = pyridine; 6 = 2-methyl-2-pentanol; 7 = 1-iodobutane; 8 = 1-octyne; 9 = dioxane; and 10 = cis-hydrindane.

DK4478_book.fm Page 116 Thursday, September 14, 2006 1:35 PM

116

Advances in Chromatography, Volume 45

EA DEA DAA DPR TEA QEA DMB QPR QBA QPE 1.5

A

B

C

D

2.5 3.5 4.5 Free energy of solution

FIGURE 2.8 Variation in the retention of (A) benzene, (B) dioxane, (C) 1-nitropropane, and (D) 1-butanol on 4-toluenesulfonate ionic liquids with different cations at 120°C. Retention of each compound is indicated as its partial molar Gibbs free energy of solution in kilocalories per mole. Identification of cations: EA = ethylammonium; DEA = diethylammonium; DAA = diallylammonium; DPR = di-n-propylammonium; TEA = triethylammonium; QEA = tetraethylammonium; DMB = dimethyl-n-butylammonium; QPR = tetra-n-propylammonium; QBA = tetra-n-butylammonium; and QPE = tetra-n-pentylammonium.

cohesion of the ionic liquids resulting in retention properties for low-polarity solutes quite like those observed for polar nonionic solvents. For polar solutes, additional intermolecular interactions with individual ions are possible and result in increased retention characteristic of polar interactions with polar nonionic liquids. In addition, it is possible that dipole-type interactions are affected by the strong Coulombic fields that only exist in ionic liquids. For ions that can delocalize charge, Coulombic interactions are expected to be weaker. These ionic liquids are also expected to be more cohesive. For tetra-n-alkylammonium ionic liquids with a common anion, the l system constant increases approximately linearly with increasing size of the cation [116,126,127,133]. It is noteworthy that the l system constant for tetra-n-pentylammonium 4-toluenesulfonate is about the same as for a conventional poly(dimethylsiloxane) stationary phase. In initial experiments to extend the application range of ionic liquids to the separation of enantiomers, it was shown that 1-butyl-3-methylimidazolium chloride or hexafluorophosphate were useful solvents for dissolving permethylated and 2,6-dimethylated β-cyclodextrins [135]. However, the enantiomer recognition capabilities of the chiral selector were reduced by dissolution in the ionic liquid, probably through formation of inclusion complexes between the ionic liquid and cyclodextrins. Greater success was achieved with a chiral ionic liquid [N,Ndimethylephedrinium bis(trifluoromethylsulfonyl)amide], which has two chiral centers resident on the cation [136]. Enantiomer selectivity was observed by gas chromatography for chiral alcohols, sulfoxides, epoxides, and acetylated amines. After use for several weeks at temperatures greater than 140°C, the ionic liquid lost selectivity for some enantiomer separations, perhaps through racemization of one of the chiral centers.

DK4478_book.fm Page 117 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

117

C. CHEMICAL REACTIVITY PROBLEMS Since ionic liquids contain ions spanning a wide range of properties, there is a reasonable probability of chemical transformations occurring under the conditions required for gas chromatography, namely, high temperatures, significant residence times in the stationary phase, and high concentration ratios of ionic liquid to sample. Reactions that might occur include nucleophilic displacement, acid/base catalyzed transformations, and oxidation by easily reduced anions (e.g., nitrate). In early studies with tetra-n-alkylammonium tetrafluoroborate salts, it was noted that alcohols, phenols, and alkylamines were retained excessively and in many cases irreversibly [115,137]. Similar observations were made for imidazolium-based ionic liquids containing trifluoromethanesulfonate, hexafluorophosphate, and bis(trifluoromethylsulfonyl)amide anions [129,130,134]. Columns prepared with tetra-n-butylammonium methanesulfonate were rapidly degraded by injection of 1-bromoalkanes [138]. On columns prepared with alkylammonium thiocyanates, peaks were not always observed for pyridine and 1-iodobutane [116]. The recovery of injected mass for hexanol and benzaldehyde was shown to depend rather critically on the sample size and temperature for columns prepared from ethylammonium 4-toluenesulfonate [113]. For tetra-n-butylammonium chloride, picrate, methanesulfonate, trifluoromethanesulfonate, and 4-toluenesulfonate salts and the n-butylammonium, di-nbutylammonium, and tri-n-butylammonium 4-toluenesulfonate salts, the only significant chemical reactions observed were nucleophilic displacement of halogens from saturated carbon and the degradation of alkanethiols on some salts [139]. Strong nitrogen bases exhibited poor peak shapes on these ionic liquids. In general, with the exceptions noted above, the mass recovery of samples (e.g., hydrocarbons, halobenzenes, nitro compounds, weak nitrogen bases, ketones, aldehydes, ethers, esters, alcohols, phenols, carboxylic acids, etc.) injected onto ionic liquids is in agreement with expectations.

D. OPEN-TUBULAR COLUMNS In general, standard laboratory methods are suitable for the preparation of packed columns with ionic liquid stationary phases [114,124]. These columns usually show the expected efficiency, and because of the excellent deactivation properties of ionic liquids, normal peak shapes for even polar compounds. The most cohesive of ionic liquids, like ethylammonium nitrate, are an exception in that stable films are formed only when the concentration of the coating solution exceeds a minimum value required to achieve a complete film on the support surface [118]. Most measurements of the physicochemical properties of ionic liquids by gas chromatography use packed columns, and in general, the majority of reported applications have used packed columns also. This is probably due to the simplicity of column preparation and the robustness and durability of these columns. The preparation of open-tubular columns with ionic liquids is desirable because of the greater intrinsic efficiency that these columns can provide. In common with polar molecular solvents, ionic liquids do not in general form stable films on smooth glass surfaces. In most cases they lack sufficient viscosity to avoid film migration with a variation of temperature and in

DK4478_book.fm Page 118 Thursday, September 14, 2006 1:35 PM

118

Advances in Chromatography, Volume 45

some cases only poorly wet glass surfaces. So far, no routes have been found to immobilize or form cross-linked films with ionic liquids, mimicking the success of poly(siloxane) and poly(ethylene glycol) stationary phases that dominate the practice of modern gas chromatography today. Various researchers [129,130,134–136] have demonstrated that 1,3-dialkylimidazolium-based ionic liquids can be coated onto fused-silica capillary columns by the static coating method. These columns exhibited reasonable efficiency at temperatures of 100°C (approximately 1500 to 2500 plates/m), and in the case of 1-(4methoxyphenyl)-3-methylimidazolium trifluoromethanesulfonate, an upper operating temperature of about 250°C. This behavior is quite exceptional, however, and most columns statically coated with 1,3-dialkylimidazolium-based ionic liquids exhibit upper temperature limits in the range of 150°C to 200°C. It is unclear whether temperature limits expressed by the baseline rise method are the result of decomposition of the ionic liquid itself, slight volatilization of the ionic liquid, or leaking of the stationary phase into the detector due to the mobility of the ionic liquid film. None of the above reports contain an indication of column durability. In addition, the static coating method is only suitable for coating solid phases and liquid phases of high viscosity, and may not be applicable to all ionic liquids, or for coating long columns with ionic liquids of modest viscosity. To date, there is no indication that the static coating method can be used to prepare columns with a range of film thicknesses to facilitate studies of the effect of film thickness on efficiency, useful temperature operating range, and sample capacity. Anderson and Armstrong [130] note three deficiencies in ionic liquid-coated columns prepared as discussed above: the maximum operating temperature that can be employed, the poor efficiency and peak shape for some solutes (e.g., strong hydrogen-bond acids), and the tenacious retention of some analytes (e.g., acids and bases). They correctly point out that these are problems that require a solution before open-tubular columns coated with ionic liquids will find routine applications. A common method of increasing film stability on glass surfaces is to increase surface interactions with the liquid film by roughening the glass surface. Surface preparation methods used to prepare open-tubular columns with alkylammonium ionic liquids include whisker formation [140,141], sodium chloride deposition [134,142], barium carbonate deposition [137], and deposition of a low surface area silica support [137]. The performance of these column types as a function of temperature for the ionic liquid tetra-n-butylammonium tetrafluoroborate is summarized in Figure 2.9. The barium carbonate columns are the most efficient, but the least durable. The columns coated with a layer of low surface area silica are neither efficient nor durable, and are therefore of little practical utility. The efficiency of the whisker-walled columns varies with the amount of ionic liquid applied to the surface [141]. The optimum amount corresponds to the minimum concentration of coating solution required to form a complete and homogeneous film above the melting point of the salt. At other, both higher and lower concentrations, unstable films result. Consequently, retention cannot be varied by changing the film thickness. Sodium chloride deposition on either a smooth or roughened (whisker formation) glass surface provides columns of acceptable performance and durability. In addition, these columns are quite straightforward to prepare and might well be preferred for

DK4478_book.fm Page 119 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

A

Hmin (mm)

1.5

B

1

C D E

0.5

0 160

119

180 200 Temperature (°C)

220

FIGURE 2.9 Variation of the minimum plate height with temperature for open-tubular columns coated with tetra-n-butylammonium tetrafluoroborate. The columns were prepared using different surface roughening techniques: A = whisker surface with a layer of Chromosorb R6470-1; B = smooth surface with a layer of sodium chloride; C = whisker surface; D = whisker surface with a layer of sodium chloride; and E = smooth surface with a layer of barium carbonate.

use in routine applications. An example of a separation on a column of this type is shown in Figure 2.10. The viscosity of the tetra-n-butylammonium tetrafluoroborate ionic liquid over the temperature range shown in Figure 2.9 is about 2 to 8 cP. This is significantly lower than desirable for stable film formation. Changes in local film homogeneity accompanying a decrease in viscosity with increasing temperature are the probable cause of the trends indicated in Figure 2.9. At temperatures greater than 240°C, breakup of the film occurs with the formation of droplets and mechanical transport of the ionic liquid from the column. Ionic liquids of higher intrinsic viscosity than tetra-n-butylammonium tetrafluoroborate are expected to exhibit greater durability as well as higher practical operating temperatures.

VI. CONCLUSION Interest in ionic liquids has grown rapidly over the last few years and a striking number of applications have appeared in organic synthesis and electrochemistry. Their exploitation in separation science and extraction is not as well developed. There are a number of reasons for this, such as a lack of fundamental property information, the need to synthesize and purify salts for laboratory use, poor systems for aiding the correct choice of ionic liquids for a particular application, and the need for ionic liquids with a wider range of properties for some applications than those currently available. Time and curiosity will take care of many of these issues. Fundamental studies that can direct the identification of task-specific ionic liquids should be considered a high priority. In many of the applications discussed here, the ionic liquids available to the authors were pressed into service and were sometimes found wanting for predictable reasons. An individual ionic liquid is incapable of representing the diversity among all ionic liquids and future workers with a larger lexicon of ionic liquids at their disposal will likely convert passed failures into useful

DK4478_book.fm Page 120 Thursday, September 14, 2006 1:35 PM

120

Advances in Chromatography, Volume 45

0

16

32

48

Min

FIGURE 2.10 Separation of Aroclor 1254 (Spectrum Laboratories, Ft. Lauderdale, FL) on an 8 m × 0.125 mm inside diameter whisker-walled, open-tubular column coated with a layer of sodium chloride. The column was dynamically coated with a solution of 100 mg/ml tetran-butylammonium tetrafluoroborate in methylene chloride previously saturated with sodium chloride. The separation was carried out isothermally at 190°C with a hydrogen carrier gas flow rate of 1 ml/min.

applications. There is a lot to be done in this regard and analytical chemists, with their familiarity of measurement techniques and adaptability, are well positioned to play a key role in advancing the science of ionic liquids and finding practical uses for these novel solvents.

REFERENCES 1. Abraham, M.A., and Moens, L., Clean Solvents. Alternative Media for Chemical Reactions and Processing, American Chemical Society, Washington, DC, 2002. 2. Wasserscheid, P. and Welton, T.E., Ionic Liquids in Synthesis, VCH-Wiley, Weinheim, Germany, 2002. 3. Forsyth, S.A., Pringle, J.M., and MacFarlane, D.R., Aust. J. Chem., 57, 113, 2004.

DK4478_book.fm Page 121 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis 4. 5. 6. 7. 8. 9. 10. 11. 12. 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.

121

Earle, M.J., Katdare, S.P., and Sedden, K.R., Org. Lett. 6, 707, 2004. Wasserscheid, P., and Keim, W., Angew. Chem. Int. Ed., 39, 3772, 2000. Gordon, C.M., Appl. Catal. A Gen., 222, 101, 2001. Olivier-Bourbigou, H., and Magna, L., J. Mol. Catal. A, 182, 419, 2002. Poole, C.F., J. Chromatogr. A, 1037, 49, 2004. Berthod, A., and Carda-Broch, S., J. Liq. Chromatogr. Rel. Technol., 26, 1493, 2003. Stalcup, A.M., and Cabovska, B., J. Liq. Chromatogr. Rel. Technol., 27, 1443, 2004. Liu, J.-F., Jonsson, J.A., and Jiang, G.-B., Trends Anal. Chem., 24, 20, 2005. Turner, E.A., Pye, C.C., and Singer, R.D., J. Phys. Chem. A, 107, 2277, 2003. Zhao, H., Phys. Chem. Liq., 41, 545, 2003. Marsh, K.N., Boxall, J.A., and Lichtenthaler, R., Fluid Phase Equilib., 219, 93, 2004. Eike, D.M., Brenneeke, J.F., and Maginn, E.J., Green Chem. 5, 323, 2003. Huddlestone, J.C., Visser, A.E., Reichert, W.M., Willauer, H.D., Broker, G.A., and Rogers, R.D., Green Chem., 3, 156, 2001. Okoturo, O.O., and VanderNoot, T.J., J. Electroanal. Chem., 568, 167, 2004. MacFarlane, D.R., Forsyth, S.A., Golding, J., and Deacon, G.B., Green Chem., 4, 444, 2002. Baranyai, K.I., Deacon, G.B., MacFarlane, D.R., Pringle, J.M., and Scott, J.L., Aust. J. Chem., 57, 145, 2004. Kosmulski, M., Gustafsson, J., and Rosenholm, J.B., Thermochim. Acta, 412, 47, 2004. Fox, D.M., Awad, W.H., Gilman, J.W., Maupin, P.H., De Long, H.C., and Trulove, F.C., Green Chem., 5, 724, 2003. Anthony, J.L., Magnin, E.J., and Brennecke, J.F., J. Phys. Chem. B, 105, 10942, 2001. Blanchard, L.A., and Brennecke, J.F., Ind. Eng. Chem. Res., 40, 287, 2001. Wong, D.S.H., Chen, J.P., Chang, J.M., and Chou, C.H., Fluid Phase Equilib., 194, 1089, 2002. Carda-Broch, S., Berthod, A., and Armstrong, D.W., Anal. Bioanal. Chem., 375, 191, 2003. Cammarata, I., Kazarian, S.G., Salter, P.A., and Welton, T., Phys. Chem. Chem. Phys., 3, 5192, 2001. Tran, C.D., Lacerda, S.H.D., and Oliveira, D., Appl. Spectrosc., 57, 152, 2003. Gutowski, K.E., Broker, G.A., Willauer, H.D., Huddleston, J.G., Swatloski, R.P., Holbrey, J.D., and Rogers, R.D., J. Am. Chem. Soc., 125, 6632, 2003. Wu, C.-T., Marsh, K.N., Deev, A.V., and Boxall, J.R., J. Chem. Eng. Data, 48, 486, 2003. Heintz, A., Lehmann, J.R., and Wertz, C., J. Chem. Eng. Data, 48, 472, 2003. Swatloski, R.P., Visser, A.E., Reichert, W.M., Broker, G.A., Parma, L.M., Holbrey, J.D., and Rogers, R.D., J. Chem. Soc. Chem. Commun., 2070, 2001. Najdanovic-Visak, V., Esperanca, J.M.S.S., Rebelo, L.P.M., Nunes da Ponte, M., Guedes, H.J.R., Seddon, K.R., and Szydiowski, J., Phys. Chem. Chem. Phys., 4, 1701, 2002. Berthod, A., and Carda-Broch, S., Anal. Bioanal. Chem., 380, 168, 2004. Letcher, T.M., and Deenadayalu, N., J. Chem. Thermodyn., 35, 67, 2003. Doker, M., and Gmehling, J., Fluid Phase Equilib., 227, 255, 2005. Zhang, S., and Zhang, Z.C., Green Chem., 4, 376, 2002. Fadeev, A.G., and Meagher, M.M., J. Chem. Soc. Chem. Commun., 295, 2001. Shetty, P.H., Youngberg, P.J., Kersten, B.R., and Poole, C.F., J. Chromatogr., 411, 61, 1987. Shetty, P.H., Poole, S.K., and Poole, C.F., Anal. Chim. Acta, 236, 51, 1990.

DK4478_book.fm Page 122 Thursday, September 14, 2006 1:35 PM

122

Advances in Chromatography, Volume 45

40. Shetty, P.H., Liquid organic salts as stationary and mobile phases for modern liquid chromatography, PhD dissertation, Wayne State University, Detroit, Michigan, 1989. 41. Khachatryan, K.S., Smirnova, S.V., Torocheshnikova, I.I., Shvedene, N.V., Formanovsky, A.A., and Pletnev, I.V., Anal. Bioanal. Chem., 381, 464, 2005. 42. Liu, J.-F., Chi, Y.-G., Peng, J.-F., Jiang, G.-B., and Jonsson, J. A., J. Chem. Eng. Data, 49, 1422, 2004. 43. Visser, A.E., Holbrey, J.D., and Rogers, R.D., J. Chem. Soc. Chem. Commun., 2484, 2001. 44. Swatloski, R.P., Holbrey, J.D., and Rogers, R.D., Green Chem., 5, 361, 2003. 45. Visser, A.E., Swatloski, R.P., and Rogers, R.D., Green Chem., 2, 1, 2000. 46. Abraham, M.H., Zissimos, A.M., Huddleston, J.G., Willauer, H.D., Rogers, R.D., and Acree, W.E., Ind. Eng. Chem. Res., 42, 413, 2003. 47. Abraham, M.H., Chem. Soc. Rev., 73, 1993. 48. Poole, C.F., and Poole, S.K., J. Chromatogr. A, 965, 263, 2002. 49. Abraham, M.H., Ibrahim, A., and Zissimos, A.M., J. Chromatogr. A, 1037, 29, 2004. 50. Liu, J., Jiang, G.-B., Chi, Y.-G., Cai, Y.-Q., Zhou, Q.-X., and Hu, J.T., Anal. Chem., 75, 5870, 2003. 51. Liu, J.-F., Peng, J.-F., Chi, Y.-G., and Jiang, G.-B., Talanta, 65, 705, 2005. 52. Liu, J.-F., Chi, Y.-G., and Jiang, G.-B., J. Sep. Sci., 28, 87, 2005. 53. Peng, J.-F., Liu, J.-F., Jiang, G.-B., Tai, C., and Huang, M.-J., J. Chromatogr. A, 1072, 3, 2005. 54. Andre, M., Loidl, J., Laus, G., Schottenberger, H., Bentivoglio, G., Wurst, K., and Ongania, K.H., Anal. Chem., 77, 702, 2005. 55. Liu, J.-F., Jiang, G.-B., Liu, J.-M., Jonsson, J.A., and Wen, M.-J., J. Chromatogr. A, 1066, 27, 2005. 56. Branco, L.C., Crespo, J.G., and Afonso, C.A.M., Chem. Eur. J., 8, 3865, 2002. 57. Branco, L.C., Crespo, J.G., and Afonso, C.A.M., Angew. Chem. Int. Ed., 41, 2771, 2002. 58. Miyako, E., Maruyama, T., Kamiya, N., and Goto, M., J. Chem. Soc. Chem. Commun., 2926, 2003. 59. Fortunato, R., Afonso, C.A.M., Reis, M.A., and Crespo, J.G., J. Membr. Sci., 242, 197, 2004. 60. Fortunato, R., Gonzalez-Munoz, M.J., Kubasiewicz, M., Luque, S., Alvarez, J.R., Afonso, C.A.M., Coselhoso, I.M., and Grespo, J.G., J. Membr. Sci., 249, 153, 2005. 61. Baltus, R.E., Counce, R.M., Culbertson, B.M., Luo, H.M., DePaoil, D.W., Dai, S., and Duckworth, D.C., Sep. Sci. Technol., 40, 525, 2005. 62. Visser, A.E., Swatloski, R.P., Griffin, S.T., Hartman, D.H., and Rogers, R.D., Sep. Sci. Technol., 36, 785, 2001. 63. Wei, G.T., Yang, Z.S., and Chen, C.J., Anal. Chim. Acta, 488, 183, 2003. 64. Visser, A.E. and Rogers, R.D., J. Solid State Chem., 171, 109, 2003. 65. Visser, A.E., Jensen, M.P., Laszak, I., Nash, K.L., Choppin, G.R., and Rogers, R.D., Inorg. Chem., 42, 2197, 2003. 66. Nakashima, K., Kubota, F., Maruyama, T., and Goto, M., Anal. Sci., 19, 1097, 2003. 67. Dai, S., Ju, Y.H., and Barnes, C.E., J. Chem. Soc. Dalton Trans., 1201, 1999. 68. Visser, A.E., Swatloski, R.P., Reichert, W.M., Griffin, S.T., and Rogers, R.D., Ind. Eng. Chem. Res., 39, 3596, 2000. 69. Chun, S., Dzyuba, S.V., and Bartsch, R.A., Anal. Chem., 73, 3737, 2001. 70. Stepinski, D.C., Jensen, M.P., Dzielawa, J.A., and Dietz, M.L., Green Chem., 7, 151, 2005.

DK4478_book.fm Page 123 Thursday, September 14, 2006 1:35 PM

Ionic Liquids in Extraction, Chromatography, and Electrophoresis

123

71. Luo, H.M., Dai, S., Bonnesen, P.V., Buchanan, A.C., Holbrey, J.D., Bridges, N.J., and Rogers, R.D., Anal. Chem., 76, 3078, 2004. 72. Shimojo, K. and Goto, M., Anal. Chem., 76, 5039, 2004. 73. Visser, A.E., Swatloski, R.P., Reichert, W.M., Mayton, R., Sheff, S., Wierzbicki, A., Davis, J.H., and Rogers, R.D., Environ. Sci. Technol., 36, 2523, 2002. 74. Holbrey, J.D., Visser, A.E., Spear, S.K., Reichert, W.M., Swatioski, R.P., Broker, G.A., and Rogers, R.D., Green Chem., 5, 129, 2003. 75. Blanchard, L.A. and Brennecke, J.F., Ind. Eng. Chem. Res., 40, 287, 2001. 76. Dzyuba, S.V. and Bartsch, R.A., Angew. Chem. Int. Ed., 42, 148, 2003. 77. Blanchard, L.A., Gu, Z., and Brennecke, J.F., J. Phys. Chem. B, 105, 2437, 2001. 78. Anthony, J.L., Maginn, E.J., and Brennecke, J.F., J. Phys. Chem. B, 106, 7315, 2002. 79. Husson-Borg, P., Majer, V., and Gomes, M.F.C., J. Chem. Eng. Data, 48, 480, 2003. 80. Shariati, A. and Peters, C.J., J. Supercrit. Fluids, 25, 109, 2003. 81. Dean, T.A. and Poole, C.F., J. Chromatogr., 468, 127, 1989. 82. Planeta, J. and Roth, M., J. Phys. Chem. B, 108, 11244, 2004. 83. Poole, C.F., Kersten, B.R., Ho, S.S.J., Coddens, M.E., and Furton, K.G., J. Chromatogr., 352, 407, 1986. 84. Poole, S.K., Shetty, P.H., and Poole, C.F., Anal. Chim. Acta, 218, 241, 1989. 85. Waichigo, M.M., Riechel, T.L., and Danielson, N.D., Chromatographia 61, 17, 2005. 86. He, L., Zhang, W., Zhao, L., Liu, X., and Jiang, S., J. Chromatogr. A, 1007, 39, 2003. 87. Zhang, W.Z., He, L.I., Gu, Y.L., Liu, X., and Jiang, S.X., Anal. Letts., 36, 827, 2003. 88. Xiao, X.H., Zhao, L., Liu, X., and Jiang, S.X., Anal. Chim. Acta, 519, 207, 2004. 89. Kaliszan, R., Marszall, M.P., Markuszewski, M.J., Baczek, T., and Pernak, J., J. Chromatogr. A, 1030, 263, 2004. 90. Baczek, T., Marszall, M.P., Kaliszan, R., Walijewski, L., Makowiecka, W., Sparzak, B., Grzonka, Z., Wisniewska, K., and Juszczyk, P., Biomed. Chromatogr., 19, 1, 2005. 91. Villagran, C., Deetlefs, M., Pitner, W.R., and Hardacre, C., Anal. Chem., 76, 2118, 2004. 92. Stepnowski, P., Muller, A., Behrend, P., Ranke, J., Hoffmann, J., and Jastorff, B., J. Chromatogr. A, 993, 173, 2003. 93. Stepnowski, P. and Mrozik, W., J. Sep. Sci., 28, 149, 2005. 94. Stepnowski, P., Anal. Bioanal. Chem., 381, 189, 2005. 95. Alberti, G. and Allulli, S., Chromatogr. Rev., 10, 99, 1986. 96. Lederer, M., Chromatography for Inorganic Chemistry, Wiley, Chichester, U.K., 1994. 97. Poole, C.F., Electrophoresis in ionic liquids, in Encyclopedia of Analytical Science, 2nd ed., vol. 2, Worsfold, P., Townshend, A., and Poole, C.F., eds., Elsevier, London, 2004, p. 392. 98. Qin, W.D., Wei, H.P., and Li, S.F.Y., J. Chromatogr. A, 985, 447, 2003. 99. Cabovska, B., Kreishman, G.P., Wassell, D.F., and Stalcup, A.M., J. Chromatogr. A, 1007, 179, 2003. 100. Qi, S.D., Cui, S.Y., Chen, X.G., and Hu, Z., J. Chromatogr. A, 1059, 191, 2004. 101. Vaher, M., Koel, M., and Kaljurand, M., J. Chromatogr. A, 979, 27, 2002. 102. Vaher, M., Koel, M., and Kaljurand, M., Chromatographia, 53, S302, 2001. 103. Vaher, M., Koel, M., and Kaljurand, M., Electrophoresis 23, 426, 2002. 104. Vaher, M. and Koel, M., J. Chromatogr. A, 1068, 83, 2005. 105. Jiang, T.F., Gu, Y.L., Liang, B., Li, J.B., Shi, Y.P., Ou, Q.Y., Anal. Chim. Acta, 479, 249, 2003. 106. Qin, W.D. and Li, S.F.Y., J. Chromatogr. A, 1048, 253, 2004.

DK4478_book.fm Page 124 Thursday, September 14, 2006 1:35 PM

124

Advances in Chromatography, Volume 45

107. Qin, W. and Li, S.F.Y., Analyst, 128, 37, 2003. 108. Yanes, E.G., Gratz, S.R., Baldwin, M.J., Robinson, S.E., and Stalcup, A.M., Analyst, 125, 1919, 2000. 109. Yanes, E.G., Gratz, S.R., Baldwin, M.J., Robinson, S.E., and Stalcup, A.M., Anal. Chem., 73, 3838, 2001. 110. Mwongela, S.M., Numan, A., Gill, N.L., Agbaria, R.A., and Warner, I.M., Anal. Chem., 75, 6089, 2003. 111. Markuszewski, M.J., Stepnowski, P., and Marszall, M.P., Electrophoresis, 25, 3450, 2004. 112. Berthier, D., Varenne, A., Garell, P., Digne, M., Lienemann, C.P., Magna, L., and Olivier-Bourbigou, H., Analyst, 120, 1257, 2004. 113. Poole, C.F., Furton, K.G., and Kersten, B.R., J. Chromatogr. Sci., 24, 400, 1986. 114. Poole, C.F., Furton, K.G., Pomaville, R.M., Poole, S.K., and Kersten, B.R., Determination of physicochemical properties of liquid organic salts using gas chromatographic techniques, in Molten Salt Techniques, vol. 4, Gale, R.J. and Lovering, D.G., eds., Plenum Press, New York, 1990, p. 41. 115. Poole, C.F., Butler, H.T., Coddens, M.E., Dhanesar, S.C., and Pacholec, F., J. Chromatogr., 289, 299, 1984. 116. Coddens, M.E., Furton, K.G., and Poole, C.F., J. Chromatogr., 356, 59, 1986. 117. Furton, K.G. and Poole, C.F., Anal. Chem., 59, 1170, 1987. 118. Kersten, B.R. and Poole, C.F., J. Chromatogr., 399, 1, 1987. 119. Pomaville, R.M., Poole, S.K., Davis, L.J., and Poole, C.F., J. Chromatogr., 438, 1, 1988. 120. Pomaville, R.M. and Poole, C.F., Anal. Chem., 60, 1103, 1988. 121. Poole, S.K. and Poole, C.F., J. Chromatogr., 500, 329, 1990. 122. Poole, S.K., Kollie, T.O., and Poole, C.F., J. Chromatogr. A, 664, 229, 1994. 123. Furton, K.G. and Morales, R., Anal. Chim. Acta, 246, 171, 1991. 124. Poole, C.F., The Essence of Chromatography, Elsevier, Amsterdam, 2003. 125. Poole, S.K., Furton, K.G., and Poole, C.F., J. Chromatogr. Sci., 26, 67, 1988. 126. Poole, S.K. and Poole, C.F., Analyst, 120, 289, 1995. 127. Kollie, T.O. and Poole, C.F., Chromatographia 33, 551, 1992. 128. Poole, S.K. and Poole, C.F., J. Chromatogr. A, 697, 415, 1995. 129. Anderson, J.L., Ding, J., Welton, T., and Armstrong, D.W., J. Am. Chem. Soc., 124, 14247, 2002. 130. Anderson, J.L. and Armstrong, D.W., Anal. Chem., 75, 4851, 2003. 131. Martin, S.D., Poole, C.F., and Abraham, M.H., J. Chromatogr. A, 805, 217, 1998. 132. Furton, K.G., Poole, C.F., and Kersten, B.R., Anal. Chim. Acta, 192, 255, 1987. 133. Furton, K.G. and Poole, C.F., J. Chromatogr., 399, 47, 1987. 134. Armstrong, D.W., He, L., and Liu, Y.-S., Anal. Chem., 71, 3873, 1999. 135. Berthod, A., He, I., and Armstrong, D.W., Chromatographia, 53, 63, 2001. 136. Ding, J., Welton, T., and Armstrong, D.W., Anal. Chem., 76, 6819, 2004. 137. Dhanesar, S.C., Coddens, M.E., and Poole, C.F., J. Chromatogr., 349, 249, 1985. 138. Furton, K.G. and Poole, C.F., J. Chromatogr., 349, 235, 1985. 139. Poole, S.K. and Poole, C.F., J. Chromatogr., 435, 17, 1988. 140. Dhanesar, S.C. and Poole, C.F., Anal. Chem., 56, 2509, 1984. 141. Dhanesar, S.C., Coddens, M.E., and Poole, C.F., J. Chromatogr., 324, 415, 1985. 142. Dhanesar, S.C., Coddens, M.E., and Poole, C.F., J. Chromatogr. Sci., 23, 320, 1985.

DK4478_C003.fm Page 125 Friday, September 22, 2006 8:39 AM

3

Applications of Capillary Electrophoresis to Molecular Recognition and Analysis of In-Capillary Enzyme-Mediated Transformations Jose Zavaleta, Dinora B. Chinchilla, Abby Brown, Alejandra Ramirez, Violet Calderon, Taguhi Sogomonyan, Ruth Montes, and Frank A. Gomez

CONTENTS I. Introduction................................................................................................126 II. Results and Discussion..............................................................................129 A. Affinity Capillary Electrophoresis ....................................................129 B. Partial Filling Affinity Capillary Electrophoresis .............................130 C. Flow-Through Partial Filling Affinity Capillary Electrophoresis ..................................................................................137 D. Competitive Flow-Through Partial Filling Affinity Capillary Electrophoresis ..................................................................................137 E. Multiple-Injection Affinity Capillary Electrophoresis Techniques .........................................................................................142 1. Multiple-Step Ligand Injection Partial Filling Affinity Capillary Electrophoresis ............................................................142 2. Multiple-Injection Affinity Capillary Electrophoresis ................144 F. Enzyme-Mediated Microreactions ....................................................149 1. Single On-Column Enzyme-Catalyzed Microreactions .............149

125

DK4478_C003.fm Page 126 Friday, September 22, 2006 8:39 AM

126

Advances in Chromatography, Volume 45

2. Double On-Column Enzyme-Catalyzed Microreactions............152 3. On-Column Enzyme-Catalyzed Microreactions Using Indirect Detection........................................................................155 G. ACE Coupled to On-Column Microreactions ..................................159 1. On-Column Ligand Derivatization Affinity Capillary Electrophoresis ............................................................................159 2. On-Column Receptor Derivatization Affinity Capillary Electrophoresis (OCRDACE)......................................................163 III. Conclusion .................................................................................................167 Acknowledgments..................................................................................................167 References..............................................................................................................168

I. INTRODUCTION The first five years of the 21st century have demonstrated the synergism between multiplex synthesis, detection, and analysis. The reliance of the synthetic chemist on analytical techniques for the development of new compounds — just 10 years ago, frequently an arduous and time-consuming task — has been reduced to routine, allowing for high-throughput synthesis, screening, and analysis with similar precision and accuracy. Not limited to the analysis of large libraries of compounds, the biochemist too has profited from advancements in analytical instrumentation, as it is now common to study macromolecules by a variety of techniques, including mass spectrometry and X-ray diffraction. Of equal importance is the development of novel molecular biology techniques that has allowed for the discovery of a myriad of biological interactions and processes. All together, these advances have provided great insight into life processes and, in particular, cell division, cell death, and cell transformation. The past two decades have seen the development of highly robust, versatile, and accurate instruments capable of analyzing minute quantities of material under high-throughput conditions. One technique that has been used in research laboratories is capillary electrophoresis (CE). Primarily because of its small sample size requirement and high speed of analysis, CE has become a staple in the analysis of many types of compounds including proteins, DNA, cations and anions, and sugars [1–4]. CE separates molecules based on their mobility under the influence of an applied voltage. The value of the mobility is directly related to the charge of the species and is inversely related to its hydrodynamic drag. Since the inception of the first commercial CE instruments in the late 1980s, a number of versatile CE techniques have been developed to explore the physical biochemistry of biomolecules. Of these techniques, affinity capillary electrophoresis (ACE) and enzymemediated microanalysis (EMMA) have shown great promise in the analysis of biophysical parameters. Since the first papers in 1992 [5–9] documenting the use of ACE to measure affinity parameters between biological species, its use in probing a variety of receptor-ligand interactions has greatly expanded and includes, but is not limited to, protein-drug, protein-DNA, peptide-peptide, peptide-carbohydrate, carbohydrate-

DK4478_C003.fm Page 127 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

127

drug, and antibody-antigen [10–45]. For example, Kaddis et al. [10] used ACE to estimate binding constants for the substrate and activator of Rhodobacter sphaeroides adenosine 5′-diphosphate-glucose pyrophosphorylase. Li et al. [11] used capillary isoelectric focusing and ACE to determine binding constants between antibodies to the prion protein. Finally, Lewis et al. [12] described the screening of antimicrobial targets using ACE. Unlike more established analytical techniques (e.g.,, size exclusion, equilibrium dialysis, sedimentation, slab gel electrophoresis, and fluorescence quenching) used to estimate affinity parameters between receptors and ligands that frequently require the separation of free or complexed species in an equilibrium mixture, standard ACE uses changes in the migration time of a receptor (or ligand) induced upon complexation with a ligand (or receptor) relative to noninteracting standards in the electrophoresis buffer [4]. Subsequent Scatchard analysis yields a value for the binding constant. As long as an equilibrium is established between the ligand and receptor at the point of detection, a Kb can be estimated. Enzyme-mediated microanalysis has been used to monitor on-column reactions in particular enzyme-catalyzed microreactions [46–63]. For example, Zugel et al. [51] used an on-column technique to assay leucine aminopeptidase on a microchip. Jin et al. [52] used an on-column approach to monitor the oxidation of glucose by glucose oxidase in submicroliter samples. Finally, Kwak et al. [56] related electrophoresis conditions of in-capillary enzyme-catalyzed microreactions to product distribution profiles. In EMMA, differential electrophoretic mobility is utilized to merge zones of analyte and reagent(s) under the influence of an electric field. Upon electrophoresis, zones of sample overlap, yielding product which is then transported to the detector. Much of the current assay work on microfluidic formats, and especially those using multiplex strategies, use early EMMA methodologies as their basis. Affinity capillary electrophoresis and EMMA are complimentary to each other in sample analysis requiring only small amounts of material and fast turnaround times. Because of these similarities, it seems only natural that both techniques would have found wide use in a coupled format. Yet there have been only a few reports describing the use of ACE and EMMA-like methods employed in the same experiment [17,24,25]. Herein, we describe the use of CE for probing biological interactions and in examining enzyme reactions. Furthermore, direct coupling of ACE and EMMA techniques is developed, demonstrating the versatility of CE to effect reactions with subsequent binding assays. To demonstrate the efficacy of ACE we focus on our work examining the binding of the glycopeptide antibiotics vancomycin (Van), ristocetin (Rist), and teicoplanin (Teic) (Figure 3.1) to D-Ala-D-Ala terminus peptides and carbonic anhydrase B (CAB, E.C. 4.2.1.1) to arylsulfonamides. To demonstrate the wide potential of enzyme-mediated microreactions we detail our work involving single- and double-enzyme microreactors and specifically the enzymes glucose-6-phosphate dehydrogenase (G6PDH, EC 1.1.1.49), hexokinase (HK, EC 2.7.1.1), apyrase (APY, EC 3.6.1.5), and fructose-biphosphate aldolase (ALD, EC 4.1.2.13). Finally, we will expand on our efforts in coupling on-column ligand and receptor derivatization to ACE using the glycopeptide antibioticpeptide system.

DK4478_C003.fm Page 128 Friday, September 22, 2006 8:39 AM

128

Advances in Chromatography, Volume 45

(a)

Me

OH

O NH2 HO HO

O O

O HO

Me

O

Cl

O

Cl HO

H

O2C HO Me

O N

O

NHX O

O O

OH

CH3

Cl

O

O

Cl

O O H OH

O

Me

O

HO HO

HN

−

H N

H

HO HO

N

H N

O

H N

H

O

H HO2C

H

O

N

N H O

H

H

O

H HO

HO

H

O

N H

NHY H

O

OH O

O

OH

OH OH

HO

OH

X Teic A2-1 Teic A2-2 Teic A2-3 Teic A2-4 Teic A2-5

O

OH

HO

(b)

H

O

H2N

+

NH2

N

H

N H

H

Me H

N

H

H

O

OH

O

O

N

H

O −

H N

O

H N

Y

+

Teic Teic-acetylTeic-succinyl-

(Z)-4-decanoic acid 8-methylnonanoic acid n-decanoic acid 3-methyldecanoic acid 9-methyldecanoic acid

H2 C(O)CH3

C(O)CH2CH2CO2−

2 OR

(c) O HO H3C

O O + NH3 H

H N

O

O

O

H N

H

H O

N H H H3CO2C HO

H

H

OR2

H

N H

HO HO

OH

O

O

N

N

O

H

NHY H

O CH3 HO

R1 = R2 = sugars

Y Rist Rist-acetylRist-succinyl-

H2+ C(O)CH3 C(O)CH2CH2CO2–

FIGURE 3.1 Structures of (a) Van complexed with N-Ac-D-Ala-D-Ala (3), (b) Teic, and (c) Rist.

DK4478_C003.fm Page 129 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

129

II. RESULTS AND DISCUSSION A. AFFINITY CAPILLARY ELECTROPHORESIS At present, there are a number of analytical techniques in use to measure affinity parameters of biological interactions, including fluorescence quenching, radioimmunoassays, ammonium sulfate precipitation, and slab gel electrophoresis techniques. These techniques frequently require the separation and quantitation of free or complexed molecules in an equilibrium mixture. If the amount of bound and free ligand in solution can be distinguished, these techniques can provide reasonable estimates of binding constants (Kb) for the interaction in question. In more than a decade of work on ACE, we have examined several receptorligand combinations and specifically the glycopeptide antibiotics Van, Teic, and Rist and their binding to D-Ala-D-Ala terminus peptides and CAB and its binding to arylsulfonamides [4,14–29]. These systems are amenable to CE, given the lack of adsorption of receptor onto the capillary column, readily available materials, relative ease of synthesis of peptides, and binding constants in the range conducive to CE. In classical ACE, changes in the migration time of a receptor on complexation with a ligand present in the buffer can be correlated to Kb. Analysis of this change in migration time as a function of the concentration of the ligand yields a valve for Kb. Figure 3.2 demonstrates the principle of ACE using electrophoretic mobility (μ) as the basis for the analysis. It is common to derive binding constants in the range of 103–108/M–1. It is apparent from Figure 3.2 that changes in the charge of the receptor on complexation to the ligand induce a greater change in μ than the change in mass. For example, for a protein of mass (M) equal to 50 kDa and pI of 6 in buffer at physiological pH, the charge of the protein would be slightly negative and it would migrate through the capillary at a velocity less than electroosmotic flow. Upon complexation to a ligand of mass (m) equal to 0.5 kDa and a charge of –1, the change in charge of the receptor-ligand complex would be greater than the change in mass (M + m). In other words, the value for M + m approaches M and any change in μ of the receptor-ligand complex compared to the receptor alone would be caused by the change in charge. We recently used classical ACE to examine the binding of Rist to several peptides [4,14,17]. Rist is a glycopeptide antibiotic of the Van group that inhibits the growth Kb + Kd Receptor

Ligand

Complex

Charge

Z

z

Z+z

Mass

M

m

M+m

μ~ Z/M 2/3

μ~ z/m 2/3

μ~ (Z + z)/(M + m)2/3

Electrophoretic mobility

FIGURE 3.2 Schematic of a typical receptor-ligand interaction and its relationship to charge, mass, and electrophoretic mobility.

DK4478_C003.fm Page 130 Friday, September 22, 2006 8:39 AM

130

Advances in Chromatography, Volume 45

of gram-positive bacteria by hindering cell wall peptidoglycan biosynthesis and has been the standard route for treating bacterial infections [64–72]. These molecules are naturally produced by the fermentation of microorganisms and inhibit the growth of gram-positive bacteria by hindering the biosynthesis of cell-wall peptidoglycan. They bind to the D-Ala-D-Ala portion of peptidoglycan intermediates and inhibit the transglycosylation reaction needed for cross-linking of the cell wall, thereby resulting in bacteriostasis or bacterial cell death. Figure 3.3A shows a representative series of electropherograms of Rist in capillaries partially filled with increasing concentrations (0 to 300 μM) of ligand 1. Upon electrophoresis a dynamic equilibrium is achieved between the plug of Rist and ligand 1, resulting in a shift in migration time of the Rist-ligand 1 complex. The complexation between ligand 1 and Rist resulted in an increasing negative charge and the complex is detected later than the uncomplexed form. As can be seen in Figure 3.3, all of the electropherograms had the same elution pattern. Each concentration of ligand was run in four repetitions at least twice over two separate days. The protein peak shifts to greater migration times on increasing concentrations of adenosine triphosphate (ATP) in the buffer. Figure 3.3B is a Scatchard plot of the data. A Kb of 41,400 was obtained for the interaction of Rist and ligand 1 using the relative migration time ratio (RMTR) (Equation 3.1): RMTR = (tr – ts′)/(ts′ – ts),

(3.1)

where tr, ts, and ts′ are the measured migration times of Rist and two noninteracting standard peaks (Fmoc-Ala-OH [an impurity in the peptide synthesis] and horse heart myoglobin [HHM]), respectively. The RMTR is a dual-marker form of analysis that has been shown to give better results in obtaining Kb than other forms of analysis, especially in instances where electroosmotic flow is variable [16]. A Scatchard plot can be obtained using Equation 3.2. Here, ΔRMTRR,L is the magnitude of the ΔRMTRR,L/[L] = KbΔRMTRR,Lmax – KbΔRMTRR,L

(3.2)

change in RMTR as a function of the concentration of ligand 1. Equation 3.2 allows for the estimation of Kb on a relative time scale using two noninteracting standards and compensates for fluctuations in the capillary column induced by electrophoresis. The values obtained by ACE were in good agreement with previous binding constant studies using other analytical techniques.

B. PARTIAL FILLING AFFINITY CAPILLARY ELECTROPHORESIS Although ACE has been demonstrated as a versatile technique in estimating binding parameters, the great demand imposed by the chemical and pharmaceutical industry for robust analytical techniques utilizing ever decreasing amounts of material has warranted the development of other methods of analysis. Partial filling techniques in CE were originally developed in the mid-1990s specifically for the separation of enantiomers by cyclodextrins [43,44]. Unlike standard ACE, which generally requires the entire capillary be filled with ligand species, in partial filling affinity

DK4478_C003.fm Page 131 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

(1), uM

∗

300 200

∗

150 100 75 50

∗

20

∗

0.0

70 t (s)

Rist A 9780

4900

40

∗

55

14660

ΔRMTR/(1) (104 M−1)

Rist A MO HHM CAB

131

20 0.1

0.2

0.3 ΔRMTR

0.4

0.5

(10−1)

(b)

85

(a)

FIGURE 3.3 (a) A representative series of electropherograms of Rist in 192 mM glycine, 25 mM Tris buffer (pH 8.3) containing various concentrations of ligand 1 using the standard ACE technique. Mesityl oxide (MO) and CAB were used as internal standards. The total analysis time was 2.0 min at 25 kV (current 7.7 μA) using a 30.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. (b) Scatchard plot of the data for Rist A according to Equation 3.2.

capillary electrophoresis (PFACE), a much smaller zone of ligand is injected into the capillary column. Until our work in 1999, no other group had extended the use of partial filling to receptor-ligand interactions encompassing proteins and peptides [18]. In that year we developed PFACE to measure binding constants of ligands to receptors (Figure 3.4A). In this technique, the capillary column is first partially filled with a ligand and then a sample of receptor and noninteracting standard is introduced and electrophoresed. Analysis of the change in the migration time of the receptor, relative to the standard, as a function of the concentration of ligand, yields a value for Kb. Using this technique, we examined the binding of D-Ala-D-Ala terminus peptides to Van. In our studies using R-D-Ala-D-Ala peptides (R = N-succinyl-(2) and N-acetyl-(3)) we demonstrated that PFACE could be used to estimate binding

DK4478_C003.fm Page 132 Friday, September 22, 2006 8:39 AM

132

Advances in Chromatography, Volume 45

Detector A Sample plug

Ligand plug

50 μm

Inject

Detector B Ligand plug

Sample plug

50 μm

Inject Neutral marker

Protein peak

Ligand plug

Time (s) (a)

FIGURE 3.4 (A) Schematic of a PFACE experiment. (B) Schematic of a FTPFACE experiment. (C) Schematic of a CBFTPFACE experiment. The sample plug is enlarged to best pictorially represent the technique. Continued.

constants of peptides to Van. Figure 3.5 shows a representative series of electropherograms of Van in a capillary partially filled with ligand 2. Upon increasing the concentration of ligand 2 in the capillary column, a shift in the migration time of Van is observed. The Van-ligand 2 complex is more negative than Van, and upon binding, shifts to the right (longer migration time). The neutral marker, MO, is unaffected by the change in ligand concentration, and its migration time does not vary significantly during the course of the experiment, hence MO can be used as a marker in the analysis of Kb. As shown in Figure 3.5, the change in concentration of ligand 2 in the column is visualized as an increased height in the ligand plateaus. The boxlike structure of the ligand peak at all concentrations of ligand denotes both a uniform injection of ligand into the column and a stable concentration of peptide in the capillary column. For the sample plug to elute on top of the ligand boxes, care was taken to ensure

DK4478_C003.fm Page 133 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

133

Detector A Sample plug

Ligand plug

50 μm

Inject

Detector B Sample and ligand plugs

50 μm

Inject

Detector C Ligand plug

Ligand plug

Inject Protein Marker peak Marker

Time (s) (b)

FIGURE 3.4 Continued.

Sample plug

50 μm

DK4478_C003.fm Page 134 Friday, September 22, 2006 8:39 AM

134

Advances in Chromatography, Volume 45

Detector A Neutral ligand plug

Sample plug

Negative ligand plug

Inject

Detector B Negative ligand plug

Inject

Charged Neutral marker ligand CAB Neutral marker Negative ligand

Time (s) (c)

FIGURE 3.4 Continued.

Neutral ligand plug

Sample plug

DK4478_C003.fm Page 135 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

135

Van MO (2), μM 1150

825

600

425

300

175

100

50

0

60

150 t (s)

240

FIGURE 3.5 A representative set of electropherograms of Van in 192 mM glycine, 25 mM Tris buffer (pH 8.3) containing various concentrations of ligand 2 using the PFACE technique. The total analysis time in each experiment was 4.0 min at 25 kV (current 5.2 μA) using a 60.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. MO was used as an internal standard.

DK4478_C003.fm Page 136 Friday, September 22, 2006 8:39 AM

136

Advances in Chromatography, Volume 45

0 100000

−400 −600 −800

80000 60000 40000 20000

−1000 −1200 −12

CAB

Van ΔRMTR/(4) (M−1)

Δ 2M/(2) (M−1)

−200

−10

−8.0 −6.0 Δ 2M(10−2)

−4.0

0 0.15

−2.0

0.2

0.25 0.3 ΔRMTR

(a)

0.35

0.4

(b)

30.0 CAB ΔRMTR/(5) (104 M−1)

25.0 20.0 15.0 10.0 5.0 0 6.0

6.5 7.0 7.5 8.0 8.5 9.0 (1 − ΔRMTRR,L) / [1 + (L) (Kb−)] (10−2) (c)

FIGURE 3.6 (a) Scatchard plot of the data for Van according to Equation 3.4. (b) Scatchard plot of the data for carbonic anhydrase B according to Equation 3.2. (c) Scatchard plot of the data for CAB according to Equation 3.8.

that ligand was injected into the capillary for a long enough period of time, otherwise incomplete overlap will occur, making analysis of the interaction problematic. Further proof that binding is occurring is shown as an increase in the peak area of the Van peak on increasing the concentration of ligand 2 in the capillary column. Figure 3.6A is a Scatchard plot of Van using varying concentrations of ligand 2 in the running buffer using the mobility ratio, M, as the basis for the analysis [16]. In this form of analysis, a noninteracting standard was used in estimating the binding constant using Equations 3.3 and 3.4: M = (teo/(tR) + 1,

(3.3)

DK4478_C003.fm Page 137 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

ΔMR,L/[L] = KbΔMR,Lmax – KbΔMR,L.

137

(3.4)

Here teo and tR are the measured migration times of the reference peak MO and Van, respectively. A Scatchard plot can be obtained using Equation 3.4. ΔMR,L is the magnitude of the change in the migration ratio as a function of the concentration of ligand. Table 3.1 lists the values for Kb using PFACE. These values are in agreement with those obtained by standard ACE techniques and other types of assays.

C. FLOW-THROUGH PARTIAL FILLING AFFINITY CAPILLARY ELECTROPHORESIS In cases where it is not possible to use conventional PFACE techniques, that is, sample and ligand plugs do not elute simultaneously at the point of detection, other means of obtaining binding constants for the interaction must be used. We recently described the use of flow-through partial filling affinity capillary electrophoresis (FTPFACE) to estimate binding constants of ligands to receptors [19]. In this technique, a smaller plug of ligand than that found in standard PFACE was used. Here the sample containing receptor and standards flows completely through the ligand plug during electrophoresis. Figure 3.4B shows a schematic of a typical FTPFACE experiment. In this work we examined the binding of carbonic anhydrase B (CAB, EC 4.2.1.1.) to charged arylsulfonamides. CAB is a zinc protein of the lyase class that catalyzes the equilibration of carbon dioxide and carbonic acid. It is strongly inhibited by sulfonamide-containing molecules. In this experiment a plug of ligand 4 at increasing concentration was injected for 0.1 min into the capillary column at high pressure, followed by a plug of sample containing CAB and noninteracting standards and electrophoresis. Upon electrophoresis, the sample plug flows into the domain of the ligand plug which is migrating at a slower velocity through the capillary column. A dynamic equilibrium is quickly reached between CAB and ligand 4. Continued electrophoresis causes the sample plug to flow through the ligand plug, which is detected first. The ligand plug is detected second as a rectangular-shaped box. Figure 3.7 shows a representative series of electropherograms of CAB in capillaries partially filled with increasing concentrations of ligand 4. Complexation between CAB and ligand 4 results in an increasing negative charge on CAB, hence it migrates later than the uncomplexed form. Figure 3.6B is a Scatchard plot of the data for CAB obtained using Equation 3.1. Using Equation 3.2, a Kb of 0.48 × 106/M–1 was obtained for ligand 4, similar to that obtained using other assay techniques.

D. COMPETITIVE FLOW-THROUGH PARTIAL FILLING AFFINITY CAPILLARY ELECTROPHORESIS We extended the use of the FTPFACE technique by examining the binding of uncharged ligands to CAB [22]. A basic premise of ACE is that the migration time of a receptor will change upon binding to a ligand. This change in migration time can then be related to the noninteracting standards and a new μ. Changes in μ are then used to obtain Kb. Small, neutral ligands, when bound to a protein, do not

DK4478_C003.fm Page 138 Friday, September 22, 2006 8:39 AM

138

Advances in Chromatography, Volume 45

TABLE 3.1 Experimental Values of Binding Constants (Kb) (104 mol/L–1) of Ligands 1–3 and 6–15 to Glycopeptide Antibiotics Measured by the Different ACE Techniques Ligand

Antibiotic

Technique

Kb (104 mol/L–1)

1

Vancomycin

2

Vancomycin

3

Vancomycin

6 7

Vancomycin Vancomycin

8

Vancomycin

9

Vancomycin

10 11 12 13 14 15 1

Vancomycin Vancomycin Vancomycin Vancomycin Vancomycin Vancomycin Teicoplanin

2 3

Teicoplanin Teicoplanin

6

Teicoplanin

OCLDACE MIACE PFMIACE MIACE PFACE FTPFACE MSLIPFACE PFACE ACE FTPFACE MSLIPFACE MIACE PFMIACE MIACE OCLDACE MIACE MIACE PFMIACE MIACE MIACE MIACE PFMIACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE ACE MIACE MIACE ACE ACE MIACE PFMIACE ACE PFMIACE

4.16 [24] 2.2 [26] 1.95 [73] 0.98 [27] 1.34 [18] 0.91 [19] 0.99 [29] 0.39 [18] 0.5 [23] 0.36 [19] 0.5 [29] 0.37 [26] 0.72 [73] 4.2 [26] 17.5 [24] 1.53 [26] 4.0 [26] 17.1 [73] 10.2 [27] 2.59 [26] 2.43 [27] 1.91 [73] 1.49 [24] 1.51 [24] 3.1 [24] 2.26 [24] 3.43 [24] 3.67 [24] 2.18 [17] 22.0 [24] 14.4 [24] 23.7 [24] 3.4 [25] 4.4 [25] 2.1 [26] 2.5 [73] 26.0 [25] 7.8 [73]

DK4478_C003.fm Page 139 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

139

TABLE 3.1 (Continued) Experimental Values of Binding Constants (Kb) (104 mol/L–1) of Ligands 1–3 and 6–15 to Glycopeptide Antibiotics Measured by the Different ACE Techniques Ligand

Antibiotic

7

Teicoplanin

8

Teicoplanin

9

Teicoplanin

10 11 12 13 14 15 1

Teicoplanin Teicoplanin Teicoplanin Teicoplanin Teicoplanin Teicoplanin Ristocetin

2 3 6 7

Ristocetin Ristocetin Ristocetin Ristocetin

8

Ristocetin

9

Ristocetin

10 11 12 13 14 15

Ristocetin Ristocetin Ristocetin Ristocetin Ristocetin Ristocetin

Technique OCLDACE ACE MIACE MIACE PFMIACE MIACE PFMIACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE ACE MIACE ACE ACE OCLDACE ACE MIACE ITC ACE MIACE ITC ACE MIACE ITC OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE OCLDACE

Kb (104 mol/L–1) 18.5 [17] 42.0 [25] 2.1 [26] 15.5 [26] 6.24 [73] 21.5 [27] 12.2 [73] 1.79 [17] 16.5 [17] 7.58 [17] 4.42 [17] 13.7 [17] 3.50 [17] 0.82 [17] 4.14 [14] 1.17 [27] nd 1.52 [14] 3.34 [14] 5.24 [17] 0.91 [14] 1.71 [27] 2.93 1.64 [14] 2.55 [27] 4.69 2.53 [14] 1.3 [27] 4.0 2.19 [17] 2.96 [17] 2.40 [17] 4.64 [17] 5.18 [17] 4.72 [17]

Note: ITC, isothermal titration calorimetry; nd, not determined.

DK4478_C003.fm Page 140 Friday, September 22, 2006 8:39 AM

140

Advances in Chromatography, Volume 45

MO

HHM CAB

(4), μM 80 60 30 22 14 8.0 6.0 4.0 3.0 2.0 0.0

240

300 t(s)

360

FIGURE 3.7 A representative set of electropherograms of CAB in 192 mM glycine, 25 mM Tris buffer (pH 8.6) containing various concentrations of ligand 4 using the FTPFACE technique. The total analysis time in each experiment was 6.5 min at 28 kV (current 5.8 μA) using a 60.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. MO and HHM were used as internal standards.

appreciably change the migration time of the newly formed complex due to a insignificant change in mass between uncomplexed and complexed receptor, thereby prohibiting the use of standard ACE techniques. To remedy this problem we have developed a competitive assay using FTPFACE to examine the binding of neutral arylsulfonamides to CAB. Figure 3.4C shows a schematic of a competitive flowthrough partial filling affinity capillary electrophoresis (CFTPFACE) experiment. In this technique, the capillary is first partially filled with the negatively charged ligand 4, followed by injections of CAB and noninteracting standards and a separate plug of ligand 4. Upon application of a voltage, CAB and the standards flow into the plug of ligand 4, where a dynamic equilibrium is established. Ligand 5 then penetrates the CAB-ligand 4 domain and a new equilibrium is established between ligand 5 and CAB. Continued electrophoresis results in the complex and standards flowing through ligand 4, which is detected last. Figure 3.8 shows a representative series of electropherograms of CAB in capillaries partially filled with increasing concentrations (0 to 64 μM) of ligand 5 and

DK4478_C003.fm Page 141 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

141

MO HHM CAB

(5), μM 64

CAB MO HHM 5 4

48 32 24 20 16 8.0 6.0 4.0 2.0 1.5 0.75 0.0

300

360 t(s)

420

FIGURE 3.8 A representative set of electropherograms of CAB in 192 mM glycine, 25 mM Tris buffer (pH 8.3) containing various concentrations of ligand 5 using the CBFTPFACE technique. The total analysis time in each experiment was 8.5 min at 28 kV (current 5.8 μA) using a 60.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. MO and HHM were used as internal standards.

electrophoresed at a constant concentration of ligand 4 (20 μM). The height of the ligand plateaus increases, as seen in the series of electropherograms, due to the increase in concentration of ligand 5 partially filled in the capillary column. In this work we modified a Scatchard analysis used in previous ACE work [22]. Rao et al. [42] showed that a dissociation constant between a receptor and a charged ligand can be estimated using Equation 3.5: Rf /[L] = (1/Kd) – (Rf/Kd).

(3.5)

Here Rf is defined as the fraction of the total concentration of R ([R]T) present as R · L, where Rf = [R · L]/[R]T. In the work of Rao et al., [L] is the concentration of a charged ligand. They showed that a similar analysis could be used to estimate binding between a receptor and a neutral ligand (L0) in a competitive binding assay using Equations 3.6 and 3.7:

DK4478_C003.fm Page 142 Friday, September 22, 2006 8:39 AM

142

Advances in Chromatography, Volume 45

Rf = [R · L0]/[R]T,

(3.6)

Rf/[L0] = (1 Rf)/((Kb0–1)(1 + [L+][Kb±])).

(3.7)

Here Kb± and [L+] are the known values for the binding constant of a charged ligand and the concentration of charged ligand, respectively. Equation 3.7 was used to obtain binding constants of neutral ligands to a receptor. In the present work, we used Equation 3.4 to estimate the ΔRMTR and substituted RMTR for Rf in Equation 3.7 to estimate binding constants between neutral arylsulfonamides and CAB using Equation 3.8 [5]. We showed that by using Equation 3.8, ΔRMTRR,L/[L0] = (1 – ΔRMTRR,L)/((Kb0–1)(1 + [L–](Kb–)),

(3.8)

the value for Kb0 can be obtained for the interaction of charged ligands and a receptor. Equation 3.8 shows the general equation used for Scatchard analysis of this competitive binding system. Here Kb– and L– are the known values for the binding constant and concentration of the negatively charged ligand in the running buffer, respectively. In this work, Kb– and L– are 4.16 × 105 mol/L and 20 μmol/L, respectively. L0 is the concentration of the neutral ligand used in the experiment. Subsequent Scatchard analysis of the value of ΔRMTR measured by FTPFACE as a function of the concentration of L0 gives the value for the binding constant (Kb0) of the neutral ligand to CAB. Figure 3.6C is a Scatchard plot of the competitive binding assay of ligand 5 for CAB.

E. MULTIPLE-INJECTION AFFINITY CAPILLARY ELECTROPHORESIS TECHNIQUES 1. Multiple-Step Ligand Injection Partial Filling Affinity Capillary Electrophoresis The human genome project and advances in proteomics have resulted in the discovery of many molecular interactions. However, the small amount of material available sometimes precludes the use of standard analytical techniques to measure the extent of their interactions. Hence, versatile ACE techniques are needed to determine the strength of these interactions. To offset the use of large quantities of material, multiple-step ligand injection partial filling affinity capillary electrophoresis (MSLIPFACE) was developed to examine the binding between receptors and ligands [29]. Figure 3.9A is a schematic of the ACE technique. In this technique, a plug of Van and noninteracting standards is injected and electrophoresed in buffer containing a given concentration of peptide. The sequence is repeated at increasing concentrations of peptide until all concentrations of ligand are run. Analysis of the change in the RMTR provides Kb. Figure 3.9B shows a representative set of electropherograms of Van in increasing concentrations of ligand 3. Periodic injections of ligand 3 at higher concentrations results in the Van peak shifting to greater migration time for any concentration of ligand 3

DK4478_C003.fm Page 143 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

143

Detector

Ligand plug

Sample plug

Ligand plug

Sample plug

Ligand plug

Sample plug

50 μm

Inject

Time (s) (a) 1200 800

Van MO CAB CAA

300 50

100

400

600

150

0.0

350

675

1000

1325

1650

t(s) (b)

FIGURE 3.9 (a) Schematic of a multiple-step ligand injection ACE experiment. (b) A representative electropherogram of Van in 192 mM glycine, 25 mM Tris buffer (pH 8.3) containing various concentrations of ligand 3 using the multiple-step ligand injection ACE technique. The total analysis time in each experiment was 27 min at 24 kV (current 4.0 μA) using an 80.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. MO and CAB (containing CAA and CAB isozymes) were used as internal standards. The number above each set of sample peaks refers to the concentration of ligand 3 (in μM).

DK4478_C003.fm Page 144 Friday, September 22, 2006 8:39 AM

144

Advances in Chromatography, Volume 45

3.0 Van

ΔRMTR/(3) (103 M−1)

2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.1

0.2

0.3 ΔRMTR

0.4

0.5

0.6

FIGURE 3.10 Scatchard plot of the data for Van according to Equation 3.4.

in the electrophoresis buffer. Figure 3.10 is a Scatchard plot of the data for Van using Equation 3.4. We have also developed a competitive assay using the MSLIPFACE technique in examining the binding of uncharged ligands to CAB. 2. Multiple-Injection Affinity Capillary Electrophoresis Recently we expanded on the multiple-injection concept by developing multipleinjection affinity capillary electrophoresis (MIACE) to determine binding constants between receptors and ligands using Van and Teic [26]. In this technique, a plug of sample containing a noninteracting standard is injected first, followed by multiple plugs of sample containing the receptor and then a final injection of sample containing a second standard. Between each injection of sample is injected a small plug of buffer containing an increasing concentration of ligand to effect separation between the multiple injections of sample. The electrophoresis is then carried out in an increasing concentration of ligand in the running buffer. Continued electrophoresis results in a shift in the migration time of the receptor in the sample plugs upon binding to their respective ligand. Analysis of the change in the RMTR or μ of the resultant receptor-ligand complex relative to the noninteracting standards as a function of the concentration of ligand yields a value for Kb. In our initial work, we first examined the binding of ligand 1 to Van using the MIACE technique. In the MIACE technique, a plug of sample (0.5 psi at 3 sec) containing MO was injected into the capillary column, followed by five plugs (0.5 psi at 3 sec) of sample containing Van (Figure 3.11). Between each injection of Van was placed a small plug (0.5 psi at 18 sec) of ligand 1 to aid in the separation of all Van peaks. A final plug (0.5 psi at 3 sec) of the second noninteracting standard, 4-carboxybenzenesulfonamide (CBSA), was then injected and electrophoresed in a solution of ligand 1.

Marker

Ligand plug

Ligand plug

∗ Van plugs

Time (s)

Van 5

FIGURE 3.11 Schematic of a MIACE experiment.

Inject

Ligand plug

∗ ∗

∗∗ ∗

Van 4

Ligand plug Van 3

Ligand plug Van 2

Ligand plug Van 1

Ligand plug

Marker

Detector

50 μm

DK4478_C003.fm Page 145 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis 145

DK4478_C003.fm Page 146 Friday, September 22, 2006 8:39 AM

146

Advances in Chromatography, Volume 45

Upon electrophoresis, individual plugs of sample migrate through the capillary column to produce seven peaks (five for Van and two for the standards) at the point of detection. Figure 3.12 shows a representative series of electropherograms of Van in a capillary filled with increasing concentrations of ligand 1 at 200 nm. The peaks for Van are not baseline resolved, but can easily be differentiated from each other. As the concentration of ligand 1 was increased (0 to 300 μM) in the running buffer, the peaks for Van shifted to longer migration times, as the Van-ligand 1 complexes are more negative than Van alone. The inverted peaks to the right of the Van peaks are due to the dilution of ligand 1 in the running buffer upon complexation to Van. These negative peaks are commonly observed in ACE studies and are particularly pronounced when the ligand or receptor in the running buffer is chromophoric or when high concentrations of the ligand or receptor are used for the binding assay. Due to the higher mass of the newly formed complex upon increasing the concentration of ligand 1, the height of the peaks for Van increase in comparison to the marker MO. Analysis of the change in the RMTR or μ of the resultant complex yields a value for Kb. (1), μM Δ 300 O

Δ O

75

Δ

O

0

150

200

250 t(s)

300

350

FIGURE 3.12 A representative set of electropherograms of Van (darkened diamond, triangle, square, circle, and open square) in 192 mM glycine, 25 mM Tris buffer (pH 8.3) containing various concentrations of ligand 1 using the MIACE technique. The total analysis time in each experiment was 6.0 min at 25 kV (current 6.8 μA) using a 40.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. MO (open circle) and CBSA (open triangle) were used as internal standards.

DK4478_C003.fm Page 147 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

147

Figure 3.13 shows Scatchard plots of the data for the five Van peaks using Equation 3.2. As can be seen, all five Van peaks result in similar Scatchard plots. The average binding constant of ligand 1 to Van was determined to be 22.3 × 103/M–1. We then examined the binding of Van to other ligands. Using a similar injection sequence, we determined the binding affinities of ligands 3, 6, 7, 8, and 9 to Van. VAN 1

2.0 1.5 1.0 0.5 0.0

2.0 1.5 1.0 0.5 0.0

0.0

2.0

4.0 6.0 8.0 ΔRMTR (102)

0.0

10.0 12.0

VAN 3

2.0

4.0 6.0 8.0 ΔRMTR (102)

2.0 1.5 1.0

10.0 12.0

VAN 4

2.5 ΔRMTR/(1)/(10−2 M1)

2.5 ΔRMTR/(1)/(10−2 M1)

VAN 2

2.5 ΔRMTR/(1)/(10−2 M1)

ΔRMTR/(1)/(10−2 M1)

2.5

2.0 1.5 1.0 0.5

0.5

0.0

0.0 0.0

2.0

4.0 6.0 8.0 ΔRMTR (102)

10.0 12.0

2.0

4.0 6.0 8.0 ΔRMTR (102)

10.0 12.0

VAN 5

2.5 ΔRMTR/(1)/(10−2 M1)

0.0

2.0 1.5 1.0 0.5 0.0 0.0

2.0

4.0 6.0 8.0 ΔRMTR (102)

10.0 12.0

FIGURE 3.13 Scatchard plots of the data for Van and ligand 1 according to Equation 3.2.

DK4478_C003.fm Page 148 Friday, September 22, 2006 8:39 AM

148

Advances in Chromatography, Volume 45

Me H

O R

N

N H

O−

O

O Me

R 1 Fmoc-Gly2 N-succinyl3 N-acetyl6 Nα, Nε-diacetyl-Lys7 Fmoc-Ala8 Fmoc-Phe9 Fmoc-Val10 Fmoc-Gly-Ala11 Fmoc-Gly-Ala-Ala12 Fmoc-Gly-Gly-Ala-Ala13 Fmoc-Ala-Ala-Ala14 Fmoc-Ala-Ala15 Fmoc-Ala-Gly-Ala-AlaO H2 N

S

R

4 R = CH2NHC(O)(CH2)4CO2– 5 R = CO2–

SCHEME 3.1 Compounds used in these studies.

Table 3.1 summarizes the binding data obtained for Van and the ligands. The binding constants obtained using MIACE were comparable to those determined using standard ACE techniques. A similar study was conducted with Teic, the results of which are shown in Table 3.1. Although CE is a very powerful technique for separating charged species, compounds with the same charge and only minimal differences in molecular weight are not always able to be separated, thus resulting in a single peak or, at best, peak overlap. A modification in MIACE has just recently been developed in our laboratories [27]. In this technique, two peptides with similar charges and mass can be separated and their binding to glycopeptide antibiotics assessed simultaneously. Here separate plugs of sample containing noninteracting standards, ligand 1, buffer, and ligand 2 are injected into the capillary column and electrophoresed. Peptides migrate through the column at similar electrophoretic mobilities but remain as distinct zones due to the buffer plug between peptides. The electrophoresis is then carried out in an increasing concentration of antibiotic in the running buffer. Continued electrophoresis results in a shift in the migration time of the peptides upon binding to the antibiotic, and a Scatchard plot is derived using RMTR analysis. The advantages of MIACE are several: One, smaller quantities of ligand are needed to conduct the studies in comparison to other assay techniques. Two, the

DK4478_C003.fm Page 149 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

149

number of receptor injections can be increased and is dependent on the capillary length and applied voltage. Third, MIACE allows for the estimation of binding affinities between biological interactions on a timescale faster than that found for standard ACE. Finally, multiple binding constants can be obtained in a series of ACE experiments, shortening the amount of time required to conduct the assay.

F. ENZYME-MEDIATED MICROREACTIONS 1. Single On-Column Enzyme-Catalyzed Microreactions In our initial microreactor experiments we examined a single enzyme-catalyzed reaction (Figure 3.14A) [50]. In these experiments, separate plugs of substrate and enzyme at increasing concentrations were introduced into the capillary column and a voltage was applied. The model system used to demonstrate quantitation of reaction products was the conversion of nicotinamide adenine dinucleotide (NAD) to NADH by G6PDH in the oxidation of Glc-6-P to 6-phosphogluconate. G6PDH is critical in the pentose phosphate pathway and is obtained from yeasts, Escherichia coli, and various mammalian tissues. G6PDH was initially injected, followed by NAD, since G6PDH has a smaller electrophoretic mobility than NAD. Glc-6-p was present as part of the buffer. Upon electrophoresis, the zone of NAD migrates into the zone of G6PDH, forming NADH. Figure 3.15A shows the electropherograms obtained for oxidation of Glc-6-P to 6-phosphogluconate using G6PDH and NAD as cofactors. Both NAD and NADH are negatively charged species and are observed as two distinct peaks in the series of electropherograms. At pH 7.8 and 30 kV, the conversion of NAD to NADH by G6PDH is rapid and the progression of the reaction is readily observed in the series of electropherograms. Figure 3.16A shows the response of the electropherograms to changes in the concentration of G6PDH in the plug. Increasing consumption of NAD and generation of NADH correlated with increasing concentrations of G6PDH in the plug. Calibration curves for NAD and NADH were run prior to the on-column microreactions and were used to generate Figure 3.16A. As can be seen from Figure 3.16A, the total amount of NAD and NADH remains constant throughout the experiment and is observed as a horizontal line when plotted versus G6PDH concentration. This demonstrates the ease of quantitation of reaction products and validates the use of the microreactor technique in the analysis of microscale reactions. In a second type of experiment, we used the G6PDH system to probe the effect an increasing plug length (injection time) of enzyme has on the conversion of a substrate to product and the ease of quantitating the reaction. Figure 3.15B shows a representative series of electropherograms obtained for the oxidation of Glc-6-P to 6-phosphogluconate using G6PDH and NAD as cofactors in buffer. Figure 3.16B shows the response of the electropherograms to changes in the injection time of G6PDH in the plug and thus the contact time between enzyme and substrate. Increasing the injection time of G6PDH allows for a greater conversion of NAD to NADH in the microreactor. In addition, the injection time of the enzyme can be easily correlated to consumption of NAD and generation of NADH, as shown by the near horizontal line in Figure 3.16B.

DK4478_C003.fm Page 150 Friday, September 22, 2006 8:39 AM

150

Advances in Chromatography, Volume 45

Detector

Substrate plug

Enzyme plug

50 μm

Inject (a) Detector

Enzyme plug

Product plug

Substrate plug

50 μm

Inject Product Substrate

Time (s) (b) Detector

Enzyme plug

Enzyme plug

Substrate plug

50 μm

Inject (c) Detector

Prod. 2 plug

Prod. 1 plug

Substrate plug

Enzyme plug

Enzyme plug

50 μm

Inject Substrate Prod. 1 Prod. 2

Time (s) (d)

FIGURE 3.14 Schematics of two types of on-capillary enzyme-catalyzed microreactions. (a,b) Schematic of an in-capillary enzyme-catalyzed single microreaction. (c,d) Schematic of an in-capillary enzyme-catalyzed double microreaction.

DK4478_C003.fm Page 151 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

151

NADH (G6PDH), mg/mL NAD

NADH G6PDH Injection time, (s)

0.80

NAD

10.0

0.60

7.0

0.40

5.0 4.0

0.20 3.0 0.10 2.0 0.05

1.0

0.00

60

100 t (s) (a)

140

0.0

70

95 t (s)

120

(b)

FIGURE 3.15 A representative series of electropherograms showing the conversion of NAD to NADH on (a) increasing concentrations of G6PDH in the enzyme plug and (b) on increasing the injection time of G6PDH in the enzyme plug. Electrolyte: 30 mM Tris, 200 μM Glc-6P (pH 7.8); detection 260 nm; temperature 30 ± 0.1°C. Analysis time 8.0 min at 30 kV (current 39.8 μA) using a 47.0 cm, 50 μM inside diameter uncoated quartz capillary.

In a related study, we applied on-column microreaction techniques to correlate electrophoresis conditions to product distribution profiles [56]. Using the G6PDH system, we showed that a mathematical relationship could be obtained to relate voltage (V), enzyme concentration (E), and mixing time of the reaction (M) to product ratios. In these experiments, G6PDH and NAD were introduced sequentially into the capillary column by pressure injection. Upon application of a voltage (1.0 kV) (we define this initial voltage as the contact voltage), mixing of enzyme and cofactor occurs and a reaction takes place. After 2.0 min (we refer to this as the mixing time at the respective contact voltage), the voltage is increased to 30 kV and the sample is electrophoresed until completion. In this type of enzyme microreactor, the amount of product formed is related to the applied voltage. Figure 3.17 shows a representative series of electropherograms obtained for the conversion of NAD to NADH. Upon completion of the series of contact voltages, higher concentrations of enzyme were used and the experiments run as before. Product distribution profiles were obtained at two other mixing times and were used to generate Equation 3.4, correlating V, E, and M to N (Equation 3.9). Table 3.2 lists the values of N obtained experimentally (Nexp) and theoretically (Ntheor) using Equation 3.9:

DK4478_C003.fm Page 152 Friday, September 22, 2006 8:39 AM

152

Advances in Chromatography, Volume 45

Component amount, (fmol)

600 500 400 NAD

300

NADH NAD + NADH

200 100 0.0 0.0

1.6 3.2 4.8 6.4 (G6PDH), (10−1 mg/mL)

8.0

(a)

Component amount, (fmol)

600 500 400 NAD

300

NADH 200

NAD + NADH

100 0.0 0.0

2.0 4.0 6.0 8.0 G6PDH, injection time (s) (b)

10

FIGURE 3.16 Response of the electropherograms to (a) changes in the concentration of G6PDH in the enzyme plug and (b) changes in the injection time of G6PDH in the enzyme plug.

N = 0.37952 – 0.016095V + 0.00016466V2 + 2.9450E – 3.2775E2 – 0.031144M + 0.10626M2.

(3.9)

As can be seen, Equation 3.9 provides an accurate means of calculating product ratios obtained from in-capillary enzyme-catalyzed microreactions using CE. 2. Double On-Column Enzyme-Catalyzed Microreactions In this series of experiments, we examined a double enzyme-catalyzed microreaction when multiple plugs of substrate and enzymes were introduced into the capillary

DK4478_C003.fm Page 153 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

153

NADH V (kV) 1.0 NAD

2.5

5.0 7.5

10

15 20

25

30

70

125 Time (s)

180

FIGURE 3.17 A representative series of electropherograms showing the conversion of NAD to NADH by G6PDH on increasing voltage. Electrolyte 30 mM Tris, 200 Glc-6-P; pH 7.8; detection 260 nm; temperature 30 ± 0.1°C. Analysis time 4.0 min at 30 kV (current 42 μA) using a 37.0 cm, 50 μM inside diameter uncoated quartz capillary.

(Figure 3.14B) [59]. We examined the conversion of ATP to adenosine diphosphate (ADP) and adenosine monophosphate (AMP) by the enzymes HK and APY, respectively. HK is a relatively nonspecific enzyme contained in all cells that catalyzes the phosphorylation of hexoses such as D-glucose, D-mannose, and D-fructose. APY is a widely distributed enzyme in plant and animal tissues that catalyzes the bivalent metal ion-dependent hydrolysis of ATP and ADP with sequential release of inorganic phosphate. Before conducting the on-column microreactions we first needed to determine the order in which the species were to be injected onto the column. We found that ATP had the smallest electrophoretic mobility, followed by HK and finally APY. Plugs of ATP and HK in buffer at pH 7.1 were introduced sequentially into the capillary followed by an increasing concentration of APY. Mixing of the two sample zones of HK and ATP is achieved by electrophoretic mixing, forming ADP. Continued electrophoresis not only causes overlap of ADP and APY to form AMP, but

DK4478_C003.fm Page 154 Friday, September 22, 2006 8:39 AM

154

Advances in Chromatography, Volume 45

TABLE 3.2 Comparison of the Values of N as Obtained by Experiment and by Equation 3.9 Contact Voltage (kV)

Mixing Time (min)

[E] (mg/ml)

Nexp

Ntheor

Error (%)

10.0 2.5 7.5 30.0 25.0 15.0 15.0 30.0 20.0 2.5

1.5 0.60 1.2 0.30 1.00 0.80 0.30 0.5 0.20 0.6

0.10 0.20 0.15 0.40 0.14 0.15 0.35 0.25 0.35 0.10

0.682 0.825 0.772 0.703 0.492 0.567 0.812 0.564 0.757 0.602

0.686 0.818 0.752 0.699 0.503 0.586 0.805 0.587 0.751 0.622

0.56 0.92 2.67 0.66 2.28 3.30 0.99 3.99 0.81 3.22

also overlap of ATP and APY, forming both ADP and AMP. Figure 3.18A shows the electropherograms obtained for the conversion of ATP to ADP to AMP using HK and APY. The phosphates are negatively charged species and are observed in the electropherograms as three separate peaks. At 260 nm, glucose, Glc-6-P, and inorganic phosphate are weakly absorbent and are not observed in the series of electropherograms. At the highest concentrations of APY, near-total conversion of ATP to ADP and AMP is observed. Because the quantities of APY and HK used in the experiments are catalytic, peaks associated with them are not observed in the electropherograms. Figure 3.19A shows the response of the electropherograms to changes in the concentration of APY in the plug. Calibration curves of ATP, ADP, and AMP were run prior to the on-column microreactions and were used for quantitating reaction products. Increasing consumption of ATP and generation of ADP and AMP correlated with increasing concentration of APY in the plug. Throughout the duration of the experiment, the total amount of ATP, ADP, and AMP is nearly constant and is observed as a horizontal line when plotted vs. APY concentration. In a similar type of experiment, a plug of ATP at a constant concentration and a plug of HK at an increasing concentration were introduced sequentially into the capillary followed by a plug of APY. Figure 3.18B shows the electropherograms obtained for the conversion of ATP to ADP to AMP using HK and APY. Figure 3.19B shows the response of the electropherograms to changes in the concentration of HK in the plug. Increasing consumption of ATP and generation of ADP and AMP correlated with increasing concentrations of HK in the plug. Throughout the experiment, the total amount of ATP, ADP, and AMP is constant and is observed as a horizontal line when plotted versus HK concentration. At the concentrations used in this experiment, APY converts ADP to AMP at rates greater than either HK or APY can convert ATP to ADP.

DK4478_C003.fm Page 155 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

155

ATP

ATP (APY), mg/mL

(HK) mg/mL

0.00

0.00

AMP ADP

80

120

0.15

160

0.10

AMP ADP

0.31

0.30

0.46

0.50

0.62

0.70

80

120

t (s)

t (s)

(a)

(b)

160

FIGURE 3.18 A representative series of electropherograms showing the conversion of ATP to ADP and AMP on (a) increasing concentration of APY at a constant concentration of HK and (b) increasing concentration of HK at a constant concentration of APY. Electrolyte 100 mM phosphate, 200 mM D-glucose, 10 mM magnesium acetate; pH 7.1; detection 260 nm; temperature 38 ± 0.1°C. Analysis time 4.0 min at 15 kV (current 205 μA) using a 27.0 cm, 50 μM inside diameter uncoated quartz capillary.

3. On-Column Enzyme-Catalyzed Microreactions Using Indirect Detection In our final series of investigations, we examined the quantitation of reaction products using indirect detection by means of two experiments [60]. In the first experiment, we examined the enzyme-catalyzed reaction when an increasing concentration of enzyme in the plug was introduced into the region occupied by a plug of substrate. We used as our model system the ALD-catalyzed conversion of fructose 1,6-diphosphate (FBP) to 3-phosphoglyceraldehyde (GAP) and dihydroxyacetonephosphate (DHAP). This reaction is an aldol cleavage and is characterized by the formation of a ketimine Schiff base intermediate with the substrate DHAP. FBP is present in all animal and plant tissue and in most microorganisms.

DK4478_C003.fm Page 156 Friday, September 22, 2006 8:39 AM

156

Advances in Chromatography, Volume 45

Component amount, (pmol)

3.0 2.5 ATP ADP AMP Sum

2.0 1.5 1.0 0.5 0.0 0.0

1.55

3.10

(APY), (10

−1

4.65

6.20

mg/mL)

(a)

Component amount, (pmol)

2.5 2.0 1.5 ATP ADP AMP Sum

1.0 0.5 0.0 0.0

1.4

2.8

4.2

5.6

7.0

(HK), (10−1 mg/mL) (b)

FIGURE 3.19 Response of the electropherograms to changes in (A) the concentration of APY at a constant concentration of HK in the enzyme plug and (B) the concentration of HK at a constant concentration of APY in the enzyme plug.

In these experiments, neither substrate nor product is ultraviolet (UV) light active, thereby requiring the use of indirect detection techniques. Cetyl trimethyl ammonium bromide (CTAB) was used as the electroosmotic flow modifier at a concentration of 0.1 mM. The use of CTAB in indirect detection is well documented and has been used successfully to analyze many inorganic and organic ions. Sorbic acid was chosen as the electrolyte because it has a high molar absorptivity coefficient, is singly charged, and because its mobility matches the migrating solute zones of the substrate and products. Plugs of FBP and ALD in reaction buffer were introduced sequentially into the capillary based on their different mobilities in an electrical field. Plugs of sample

DK4478_C003.fm Page 157 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

157

Detector

Tris plug

Substrate plug

50 μm

Tris plug

Enzyme plug

Inject (a)

Detector

Enzyme plug

Prod.2 plug

Prod.1 plug

Substrate plug

Tris plug

50 μm

Inject

Prod. 2 Prod. 1

Substrate Time (s)

(b)

FIGURE 3.20 Schematic of an in-capillary enzyme-catalyzed plug-plug microreaction (a) before reaction and (b) after reaction.

containing Tris buffer at pH 7.6 were injected before ALD and after FBP to facilitate the reaction (Figure 3.20). The reaction does not occur if the Tris plugs are not injected into the capillary. Figure 3.21A shows a representative series of electropherograms obtained for the conversion of FBP to GAP and DHAP using ALD. Increasing consumption of FBP and generation of GAP and DHAP correlated with increasing concentrations of ALD in the plug. FBPase is used in catalytic amounts and is not observed in the electropherograms. At pH 7.6 and 12 kV, the conversion of FBP to GAP and DHAP is fast and is readily observed in the electropherograms using indirect detection. A small peak is also observed at increasing concentrations of ALD. We believe this is from the buffer salts that were used to stabilize ALD as purchased from the manufacturer. Figure 3.22A shows the response of the electropherograms to changes in the concentration of ALD in the plug. In theory, both GAP and DHAP should form in equal quantities. Unfortunately, the amount of GAP formed is less than the amount of DHAP. We believe GAP is observed in smaller quantities than DHAP because of its instability in aqueous solution and at pH 7.6. At low concentrations of ALD,

DK4478_C003.fm Page 158 Friday, September 22, 2006 8:39 AM

158

Advances in Chromatography, Volume 45

(ALD), mg/mL ∗ GAP

2.0

∗

∗

GAP DHAP

FBP

FBP DHAP

∗

1.3 ∗

1.0

∗

∗

∗

0.5

1.0

∗

0.0

0.0

260

315 t (s) (a)

370

5.0

3.0

∗

∗

∗

ALD, Injection Time, (s) 7.0

255

305 t (s)

355

(b)

FIGURE 3.21 A representative series of electropherograms showing the conversion of fructose-1,6-bisphosphate to dihydroxyacetone phosphate and glyceraldehyde-3-phosphate on (a) increasing the concentration of aldolase and (b) increasing the injection time of aldolase. Electrolyte 30 mM sorbic acid, 0.1 mM CTAB (pH 9.5); detection 254 nm; temperature 25 ± 0.1°C. Analysis time 8.0 min at 12 kV (current 13.3 μA) using a 57.0 cm, 50 μM inside diameter uncoated quartz capillary. The asterisk (*) indicates the position of unidentified charged species from the enzyme plug carried through the capillary by electroosmotic flow.

the amount of DHAP formed does not equate to the amount of FBP consumed. However, at high concentrations of enzyme, the amounts are quite similar, thereby validating the use of indirect detection to quantitate in-capillary microreactions. In the second type of experiment we examined the enzyme-catalyzed reaction when a plug of substrate is introduced into the region occupied by an increasing plug length (injection time) of enzyme in the plug at a constant concentration. Figure 3.21B shows a representative series of electropherograms obtained for the conversion of FBP to GAP and DHAP using ALD. Figure 3.23B shows the response of the electropherograms to changes in the injection time of ALD in the plug and thus to

DK4478_C003.fm Page 159 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

159

Component amount, (pmol)

14.0 12.0 10.0

FBP DHAP GAP Sum

8.0 6.0 4.0 2.0 0.0 0.0

0.5 1.0 1.5 (ALD), (mg/mL)

2.0

(a)

Component amount, (pmol)

10.0 8.0

FBP DHAP GAP Sum

6.0 4.0 2.0 0.0 0.0

1.0

2.0 3.0 4.0 5.0 6.0 ALD, injection time (s)

7.0

(b)

FIGURE 3.22 Response of the electropherograms to changes in (a) the concentration of ALD and (b) the injection time of the ALD plug.

the contact time between enzyme and substrate. Increasing consumption of FBP and generation of GAP and DHAP correlated with increasing injection time of the enzyme in the capillary at constant voltage. Similar to the previous experiment, the amount of GAP formed is less than the amount of DHAP formed. Also, the amount of DHAP formed was found to be greater than the amount of FBP consumed.

G. ACE COUPLED

TO

ON-COLUMN MICROREACTIONS

1. On-Column Ligand Derivatization Affinity Capillary Electrophoresis The past few decades have seen an almost exponential growth in the development of new drugs and drug targets because of new synthetic and analytical approaches

DK4478_C003.fm Page 160 Friday, September 22, 2006 8:39 AM

160

Advances in Chromatography, Volume 45

Detector

Van plug

Buffer plug

Fmoc plug

50 μm

Sample plug

Inject (a) Detector

Sample and van plugs

50 μm

Inject New species Markers Reaction sideproducts

Van

Time (s)

(b)

FIGURE 3.23 Schematic of an on-column ligand derivatization PFACE experiment.

to rational drug design. This increased output of potential drugs compared to traditional techniques of drug design has made expeditious and facile analysis of new drugs a must in any new analytical technique. ACE has been used to examine the binding affinities of ligands to receptors resulting from a high-throughput screening and combinatorial approach to rational drug design [5,16,34,39]. However, these approaches examined ligands to receptors that were synthesized off-column and prior to analysis by CE. We recently documented the coupling of on-column synthetic techniques and ACE in the analysis of binding parameters. It is our hope that this method will allow for high-throughput synthesis and analysis in the development of CE for the analysis of receptor-ligand interactions. In this work we examined the on-column ligand derivatization of multiple Fmoc-derivatized peptides and their respective analysis by PFACE [24]. Here a plug of sample containing Ala-D-Ala-D-Ala, D-Ala-D-AlaD-Ala-D-Ala, and Gly-Ala-Ala-D-Ala-D-Ala, NAD, and CBSA was injected into the capillary followed by separate plugs of Fmoc-Gly-NHS ester and buffer (Figure 3.23). The capillary was subsequently partially filled with increasing concentrations

DK4478_C003.fm Page 161 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

161

of Van and electrophoresed. The overlap of the first two zones of material allow for the formation of three different Fmoc-derivatized species, Fmoc-Gly-D-Ala-D-AlaD-Ala (ligand 10), Fmoc-Gly-D-Ala-D-Ala-D-Ala-D-Ala (ligand 11), and FmocGly-Gly-Ala-Ala-D-Ala-D-Ala (ligand 12), each eluting at distinct migration times. Figure 3.24 shows a representative series of electropherograms of ligands 10 to 12 in buffer containing increasing concentrations of Van. All three Fmoc species elute at earlier migration times than in the absence of Van in the running buffer. At the highest concentration of Van, a 40 sec shift in the ligands is observed between the uncomplexed and complexed forms. In these experiments, approximately 16.8 pmol of peptide and 74 pmol of Van are used in the binding assay. The height of the Van B 10 11 12 NAD CBSA

C

A

Van, (μM) 40

B C A

32

B C A B

24 C

A B

18 C

A

12

B A

C

6.0

B A 0.00

C

134

169

204 t (s)

239

274

FIGURE 3.24 A representative series of electropherograms of Fmoc-Gly-D-Ala-D-Ala-DAla, ligand 10, Fmoc-Gly-D-Ala-D-Ala-D-Ala-D-Ala, ligand 11, and Fmoc-Gly-Gly-AlaAla-D-Ala-D-Ala, ligand 12, in 20 mM phosphate buffer (pH 7.5) at 205 nm containing various concentrations of Van using the on-column synthesis PFACE technique. The total analysis time in each experiment was 5.0 min at 24 kV (current 35.4 μA) using a 40.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary. NAD and CBSA were used as internal standards. A–C are explained in the text.

DK4478_C003.fm Page 162 Friday, September 22, 2006 8:39 AM

162

Advances in Chromatography, Volume 45

4.5

3.6

2.7

1.8

2

RMTR/(10, 11, 12) (104 M−1)

10 11 12

0.9 0.0

3.0

6.0 9.0 RMTR (10−1)

12

15

2

FIGURE 3.25 Scatchard plot of the data for ligands 2 through 4 according to Equation 3.2.

plateaus in Figure 3.25 increase due to the increased concentration of Van partially filled in the capillary column. Peaks A to C are identified as (A) unreacted Fmoc-Gly-NHS ester, (B) D-AlaD-Ala-D-Ala, D-Ala-D-Ala-D-Ala-D-Ala, and Gly-Ala-Ala-D-Ala-D-Ala, and (C) Fmoc-Gly-acid. These species do not interfere in the PFACE binding studies. Based on the peak heights of A, B, and C, we estimate the yield of ligand 5 to be approximately 15%. Figure 3.25 is a Scatchard plot of the data for ligands 10 to 12 using RMTR analysis. Overall, eight different D-Ala-D-Ala terminus peptides were examined, the results of which are summarized in Table 3.1. To our knowledge, binding constants for ligands 7 and 11 through 15 have not yet been reported. To assess the correctness of the on-column ligand derivatization PFACE technique, we conducted a separate ACE experiment between ligand 13 and Van. Ligand 13 was initially synthesized off-column and used for the ACE studies. A sample of ligand 13, NAD, and CBSA were injected onto the column and electrophoresed to obtain the RMTR for ligand 13. Increasing concentrations (0 to 80 μM) of Van were subsequently injected onto the column, thereby inducing changes in the migration time of ligand 13. Measurement of the RMTR due to complexation of ligand 13 and Van resulted in a binding constant of 17.7 × 103/M–1. This value is slightly smaller than that obtained using the on-column ligand synthesis PFACE technique. A factor that may have contributed to the observed differences in binding constants between the on-column PFACE and standard ACE techniques is the small variation in Fmoc species formed between runs of the former technique. If too great a deviation in the amount of peptide formed on-column occurs, such variations may influence the ratio of bound to unbound peptide and hence the migration time of the peptide. Still, we believe that since multiple electrophoresis runs were used for the Scatchard analysis, such deviations caused by variable synthetic yields are minimized over the totality of the analysis.

DK4478_C003.fm Page 163 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

163

2. On-Column Receptor Derivatization Affinity Capillary Electrophoresis (OCRDACE) Binding constants between Teic and Rist and their derivatives to D-Ala-D-Ala terminus peptides were determined by on-column receptor derivatization coupled to PFACE or ACE (Figure 3.26) [15]. In this technique, derivatives of the glycopeptides Teic and Rist are first synthesized on-column before analysis by ACE or PFACE. After the column has been partially filled with increasing concentrations of D-AlaD-Ala terminus peptides, a plug of buffer followed by two separate plugs of reagents are injected. The order of the reagent plugs containing the antibiotic and two noninteracting standards and the anhydride varies with the charge of the glycopeptide. Upon electrophoresis, the antibiotic reacts with the anhydride, yielding a derivative of Teic or Rist. Continued electrophoresis results in the overlap of the derivatized antibiotic and the plug of D-Ala-D-Ala peptide. Analysis of the change in RMTR of the new glycopeptide relative to the noninteracting standards as a function of the concentration of the D-Ala-D-Ala ligand yields a value for the binding constant. Detector

Running buffer

Anhydride plug

Teic plug

Buffer plug

D-Ala-D-Ala plug

Inject

Detector

Running buffer

Teic and D-Ala-DAla

Inject Teic A2-2 Teic A2-X Teic-acetyl A2-2 Teic-acetyl A2-X Teic-succinyl A2-2 Teic-succinyl A2-X MO NAD

Time (s)

FIGURE 3.26 Schematic of on-column receptor synthesis coupled with PFACE experiment (top) before reaction and (bottom) after reaction.

DK4478_C003.fm Page 164 Friday, September 22, 2006 8:39 AM

164

Advances in Chromatography, Volume 45

In this series of experiments we examined the binding interaction between Teic and its derivatives to ligand 3. In these studies, PFACE was used to estimate binding constants. In PFACE, the capillary is partially filled with peptide. As long as a dynamic equilibrium is established between the ligand and receptor prior to the point of detection, a binding constant can be estimated. In the present studies, a plug containing increasing concentrations of ligand 3 was initially injected to partially fill the capillary. A buffer plug was then vacuum injected into the capillary to separate the reagents from the ligand plug. The third injection contained Teic and MO and NAD as standards. The final injection contained a mixture of succinic and acetic anhydride. Upon electrophoresis, the anhydrides and Teic plugs overlap and react, forming new Teic derivatives. The buffer plug serves as a barrier between Teic and the D-Ala-D-Ala terminus peptides, so mixing does not occur prior to ACE analysis. The Teic derivatives then migrate into the zone of ligand 3 and a dynamic equilibrium is achieved as electrophoresis continues. Figure 3.27 shows a representative series of electropherograms of Teic and its derivatives. The plug of ligand 3 is observed as a box in the electropherogram that increases in height with increasing concentrations of ligand 1 (0 to 500 μM) in the plug. The glycopeptide, its derivatives, and the D-Ala-D-Ala terminus peptides used in this study are all negatively charged at pH 6.9 and thus elute after the neutral marker MO. The difference in charge neutralized upon acetylation and succinylation dictates the difference in mobility between Teic and its acetylated and succinylated derivatives. The pKa of the N-terminus amine of Teic is 7.1. This equates to an approximate charge difference between Teic and its acetylated and succinylated forms of 0.4 and 1.4, respectively. Upon addition of increasing concentrations of ligand 3 in the running buffer, the migration times of Teic and its derivatives shift to greater migration times. The complexation between ligand 3 and the Teic derivatives resulted in an increasing negative charge on the compounds and the complexes are detected later than the uncomplexed form. At the point of saturation, the Teic peaks no longer shift to the right, despite increasing concentrations of ligand 2 in the running buffer. Further proof that derivatization of Teic occurs is that Teic and Teic-acetyl-A22 elute at the same place, relative to the internal standards, as do Teic-acetyl-A2-2 and Teic-succinyl-A2-2, respectively. These results also demonstrate that derivatization of Teic has a greater effect on changing the charge than the mass. Teic-A2X exists in much smaller concentrations than the major form of Teic and upon derivatization yields much smaller peaks for its acetylated and succinylated forms. Figure 3.28 is a Scatchard plot of the data for Teic and its derivatives. In this form of analysis, Kb is estimated using the RMTR form of analysis (Equation 1) [16]. Table 3.3 details the binding of Teic and its derivatives to ligands 3, 6, and 1. Using similar experimental protocols, we also examined the on-column acetylation and succinylation of Rist. Unlike Van and Teic, Rist possesses an esterified C-terminus, thereby making it approximately neutral at the pH of the study (6.9). Upon derivatization, the Rist derivatives become negatively charged and elute at greater migration times than Rist (data not shown). Table 3.4 summarizes the binding data for Rist and its derivatives to ligands 3 and 6.

DK4478_C003.fm Page 165 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

165

(3), μM

Teic A2-2 Teic A2-X Teic-acetyl A2-2 Teic-acetyl A2-X Teic-succinyl A2-2 Teic-succinyl A2-X MO NAD

500

400

300

200

100

75.0

50.0

25.0

0.00

100

150

200

250

300

t (s)

FIGURE 3.27 A representative series of electropherograms of Teic and its derivatives in 20 mM phosphate buffer (pH 6.9) containing various concentrations of ligand 3 using on-column receptor synthesis coupled to PFACE. The total analysis time in each experiment was 5.0 min at 20 kV using a 46.5 cm (inlet to detector), 50 μm inside diameter open uncoated quartz capillary.

DK4478_C003.fm Page 166 Friday, September 22, 2006 8:39 AM

166

Advances in Chromatography, Volume 45

6.0 Teic A2-2 Teic A2-X Teic-acetyl A2-2 Teic-acetyl A2-X Teic-succinyl A2-2 Teic-succinyl A2-X

Δ 2RMTR/(3) (104 M−1)

5.0 4.0 3.0 2.0 1.0 0.0 1.0

2.0

3.0 4.0 5.0 6.0 Δ 2RMTR (10−1)

7.0

8.0

FIGURE 3.28 Scatchard plot of Teic and its derivatives with ligand 1.

TABLE 3.3 Experimental Values of Binding Constants (Kb) (104/M–1) of Teic-A2-2 and A2-X with Ligands 1, 3, and 6 Measured by the On-Column Receptor Derivatization PFACE Technique Kb (104/M) (Correlation Coefficient) Antibiotic Teic Teic-A2-X Teic-acetyl-A2-2 Teic-acetyl-A2-X Teic-succinyl-A2-2 Teic-succinyl-A2-X

Ligand 1

Ligand 3

Ligand 6

140 (0.94) 140 (0.99) 16 (0.98) 7.6 (0.99) 5.0 (0.96) nd

37 20 13 6.2 6.2 3.4

65 130 7.9 8.7 12 11

(0.94) (0.98) (0.98) (0.99) (0.99) (0.99)

(0.95) (0.98) (0.89) (0.97) (0.99) (0.93)

Note: nd, not determined.

For the on-column ligand and receptor derivatization ACE technique to be accepted and generally used, several criteria must first be satisfied. One, the reaction must be kinetically favorable, since the length of time both the substrate and derivatization reagent are in contact may be only a few seconds. Two, the electrophoretic mobilities of the ligand and receptor must be different in order for the zones to overlap upon electrophoresis. Three, knowledge about the electrophoretic mobilities of the receptor, noninteracting standards, and ligand is necessary in order to determine if overlap is sufficient at the point of detection. Four, a sufficient amount of ligand must be injected onto the capillary to ensure a dynamic equilibrium is

DK4478_C003.fm Page 167 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

167

TABLE 3.4 Experimental Values of Binding Constants (Kb) (104/M–1) of Rist and Its Derivatives with Ligands 3 and 6 Measured by the On-Column Receptor Derivatization PFACE Technique Kb (104/M–1) (Correlation Coefficient) Antibiotic

Ligand 3

Ligand 6

Rist Rist-acetyl Rist-succinyl

4.8 (0.99) 1.7 (0.99) 0.9 (0.99)

10 (0.96) 3.1 (0.97) 1.8 (0.98)

achieved between it and the receptor. Five, receptors, noninteracting markers, and ligands must not absorb onto the walls of the capillary.

III. CONCLUSION The application of CE in the analysis of molecular interactions and enzyme-catalyzed microreactions has been documented. It is apparent from the techniques derived herein that ACE is both a versatile and viable approach to examine bimolecular noncovalent interactions and is complementary to other traditional assay techniques. In some cases, ACE is a far superior approach for analyzing small molecules. For example, only small quantities of receptor and ligand are required. Purification of the sample is not required as long as CE can distinguish the impurities from the analyte of interest. Radiolabeling of the molecules is not necessary. Automated CE instrumentation is widely available. Data are reproducible and expeditiously obtained. Finally, a wide range of molecular interactions can be characterized in free solution. The modifications in ACE and enzyme-mediated transformations described herein utilize much smaller quantities of sample than other assay techniques and expedite the analysis of the respective interaction. It is easy to see that these techniques have great potential in the pharmaceutical industry and in industries where high-throughput analysis of drugs and drug targets is critical. Future work is focused on expanding the range of receptor-ligand interactions studied, in miniaturizing the techniques described herein in a microfluidic format, and in further developing more novel CE techniques.

ACKNOWLEDGMENTS The authors gratefully acknowledge financial support for this research from the National Science Foundation (grant nos. CHE-0136724, CHE-0515363, DMR0351848, and MCB-0448676), and the National Institutes of Health (grant nos. 1R15 AI055515-01 and 1R15AI65468-01).

DK4478_C003.fm Page 168 Friday, September 22, 2006 8:39 AM

168

Advances in Chromatography, Volume 45

REFERENCES 1. Clohs, L. and McErlane, K.M., Development of a capillary electrophoresis assay for the determination of carvedilol enantiomers in serum using cyclodextrins, J. Pharm. Biomed. Anal., 24, 545, 2001. 2. Guzman, N.A., Immunoaffinity capillary electrophoresis applications of clinical and pharmaceutical relevance, Anal. Bioanal. Chem., 378, 37, 2004. 3. Simal-Gándara, J., The place of capillary electrochromatography among separation techniques — a review, Crit. Rev. Anal. Chem., 34, 85, 2004. 4. Azad, M., Silverio, C., Zhang, Y., Villareal, V., and Gomez, F.A., On-column synthesis coupled to affinity capillary electrophoresis for the determination of binding constants of peptides to glycopeptide antibiotics, J. Chromatogr. A, 1027, 193, 2004. 5. Chu, Y.-H., Avila, L.Z., Biebuyck, H.A., and Whitesides, G.M., Using affinity capillary electrophoresis to identify the peptide in a peptide library that binds most tightly to vancomycin, J. Org. Chem., 58, 648, 1992. 6. Chu, Y.-H. and Whitesides, G.M., Affinity capillary electrophoresis can simultaneously measure binding constants of multiple peptides to vancomycin, J. Org. Chem., 57, 3524, 1992. 7. Baba, Y., Tsuhako, M., Sawa, T., Akashi, M., and Yashima, E., Anal. Chem., 64,1920, 1992. 8. Chu, Y.-H., Avila, L.Z., Biebuyck, H.A., and Whitesides, G.M., Use of affinity capillary electrophoresis to measure binding constants of ligands to proteins, J. Med. Chem., 35, 2915, 1992. 9. Heegaard, N.H.H. and Robey, F.A., Use of capillary zone electrophoresis to evaluate the binding of anionic carbohydrates to synthetic peptides derived from human serum amyloid P component, Anal. Chem., 64, 2479, 1992. 10. Kaddis, J., Zurita, C., Moran, J., Borra, M., Polder, N., Meyer, C.R., and Gomez, F.A., Estimation of binding constants for the substrate and activator of Rhodobacter sphaeroides ADP-glucose pyrophosphorylase using affinity capillary electrophoresis, Anal. Biochem., 327, 252, 2004. 11. Li, G., Zhou, X., Wang, Y., El-Shafey, A., Chiu, N.H.L., and Krull, I.S., Capillary isoelectric focusing and affinity capillary electrophoresis approaches for the determination of binding constants for antibodies to the prion protein, J. Chromatogr. A, 1053, 253, 2004. 12. Lewis, L.M., Engle, L.J., Pierceall, W.E., Hughes, D.E., and Shaw, K.J., Affinity capillary electrophoresis for the screening of novel antimicrobial targets, J. Biomol. Screen, 9, 303, 2004. 13. Handwerger, S., Pucci, M.J., Volk, K.J., Liu, J., and Lee, M.S., Vancomycin-resistant Leuconostoc mesenteroides and Lactobacillus casei synthesize cytoplasmic peptidoglycan precursors that terminate in lactate, J. Bacteriol., 176, 260, 1994. 14. Azad, M., Hernandez, L., Plazas, A., Rudolph, M., and Gomez, F.A., Determination of binding constants between the antibiotic ristocetin A and D-Ala-D-Ala terminus peptides by affinity capillary electrophoresis, Chromatographia, 57, 339, 2003. 15. Silverio, C.F., Azad, M., and Gomez, F.A., On-column derivatization and analysis of the antibiotics teicoplanin and ristocetin coupled to affinity capillary electrophoresis, Electrophoresis, 24, 808, 2003. 16. Mito, E., Zhang, Y., Esquivel, S., and Gomez, F.A., Estimation of receptor-ligand interactions by the use of a two-marker system in affinity capillary electrophoresis, Anal. Biochem., 280, 209, 2000.

DK4478_C003.fm Page 169 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

169

17. Azad, M., Brown, A., Silva, I., and Gomez, F.A., Estimation of binding constants between ristocetin and teicoplanin to peptides using on-column ligand derivatization coupled to affinity capillary electrophoresis, Anal. Bioanal. Chem., 379, 149, 2004. 18. Heintz, J., Hernandez, M., and Gomez, F.A., Use of a partial-filling technique in affinity capillary electrophoresis for determining binding constants of ligands to receptors, J. Chromatogr. A, 840, 261, 1999. 19. Mito, E. and Gomez, F.A., Flow-through partial-filling affinity capillary electrophoresis can estimate binding constants of ligands to receptor, Chromatographia, 50, 689, 1999. 20. Brown, A., Silva, I., Chinchilla, D., Hernandez, L., and Gomez, F.A., Partial-filling techniques for affinity capillary electrophoresis to probe receptor-ligand interactions, LCGC Eur., 1, 2004. 21. Villareal, V., Kaddis, J., Azad, M., Zurita, C., Silva, I., Hernandez, L., Rudolph, M., Moran, J., and Gomez, F.A., Partial-filling affinity capillary electrophoresis, Anal. Bioanal. Chem., 376, 822, 2003. 22. Kaddis, J., Mito, E., Heintz, J., Plazas, A., and Gomez, F.A., Flow-through partialfilling affinity capillary electrophoresis can estimate binding constants of neutral ligands to receptors via a competitive assay technique, Electrophoresis, 24, 1105, 2003. 23. Gomez, F.A., Mirkovich, J.N., Dominguez, V.M., Liu, K.W., and Macias, D.M., Multiple-plug binding assays using affinity capillary electrophoresis, J. Chromatogr. A, 727, 291, 1996. 24. Zhang, Y., Kodama, C., Zurita, C., and Gomez, F.A., On-column ligand synthesis coupled to partial-filling affinity capillary electrophoresis to estimate binding constants of ligands to a receptor, J. Chromatogr. A, 928, 233, 2001. 25. Silverio, C.S., Plazas, A., Moran, J., and Gomez, F.A., Determination of binding constants between teicoplanin and D-Ala-D-Ala terminus peptides by affinity capillary electrophoresis, J. Liq. Chromatogr. Rel. Technol., 25, 1677, 2002. 26. Chinchilla, D., Zavaleta, J., and Martinez, K., Multiple-injection affinity capillary electrophoresis to estimate binding constants of receptors to ligands, Anal. Bioanal. Chem., 2005. 27. Zavaleta, J., Chinchilla, D., Martinez, K., and Gomez, F.A., Multiple-injection affinity capillary electrophoresis to examine binding constants between glycopeptide antibiotics and peptides, J. Chromatogr. A, 1105, 59, 2006. 28. Gomez, F.A., Avila, L.Z., Chu, Y.-H., and Whitesides, G.M., Determination of binding constants of ligands to proteins by affinity capillary electrophoresis: compensation for electroosmotic flow, Anal. Chem., 66, 1785, 1994. 29. Zhang, Y. and Gomez, F.A., Multiple-step ligand injection affinity capillary electrophoresis for determining binding constants of ligands to receptors, J. Chromatogr. A, 897, 339, 2000. 30. Erim, F.B. and Kraak, J.C., Vacancy affinity capillary electrophoresis to study competitive protein-drug binding, J. Chromatogr. B Biomed. Sci. Appl., 710, 205, 1998. 31. Gao, J., Mammen, M., and Whitesides, G.M., Evaluating electrostatic contributions to binding with the use of protein charge ladders, Science, 272, 535, 1996. 32. Neubert, R.H.H. and Ruttinger, H.H., Affinity Capillary Electrophoresis in Pharmaceutics and Biopharmaceutics, Marcel Dekker, New York, 2003. 33. El-Shafey, A., Zhong, H., Jones, G., and Krull, I.S., Application of affinity capillary electrophoresis for the determination of binding constants and thermodynamic constants of enediynes with human serum albumin and histone, Electrophoresis, 23, 945, 2002.

DK4478_C003.fm Page 170 Friday, September 22, 2006 8:39 AM

170

Advances in Chromatography, Volume 45

34. Chu, Y.-H., Dunayevskiy, Y.M., Kirby, D.P., Vouros, P., and Karger, B.L., Affinity capillary electrophoresis mass spectrometry for screening combinatorial libraries, J. Am. Chem. Soc., 118, 7827, 1996. 35. Varenne, A., Gareil, P., Colliec-Jouault, S., and Daniel, R., Capillary electrophoresis determination of the binding affinity of bioactive sulfated polysaccharides to proteins: study of the binding properties of fucoidan to antithrombin, Anal. Biochem., 315, 152, 2003. 36. Qian, X.-H. and Tomer, K.B., Affinity capillary electrophoresis investigation of an epitope on human immunodeficiency virus recognized by a monoclonal antibody, Electrophoresis, 19, 415, 1998. 37. Kiessig, S., Bang, H., and Thunecke, F., Interaction of cyclophilin and cyclosporins monitored by affinity capillary electrophoresis, J. Chromatogr. A, 853, 469, 1999. 38. Dunayevskiy, Y.M., Lyubarskaya, Y.V., Chu, Y.-H., Vouros, P., and Karger, B.L., Simultaneous measurement of nineteen binding constants of peptides to vancomycin using affinity capillary electrophoresis-mass spectrometry, J. Med. Chem., 41, 1201, 1998. 39. VanderNoot, V.A., Hileman, R.E., Dordick, J.S., and Linhardt, R.J., Affinity capillary electrophoresis employing immobilized glycosaminoglycan to resolve heparinbinding peptides, Electrophoresis, 19, 437, 1998. 40. Colton, J.J., Carbeck, J.D., Rao, J., and Whitesides, G.M., Affinity capillary electrophoresis: a physical-organic tool for studying interactions in biomolecular recognition, Electrophoresis, 19, 367, 1998. 41. Busch, M.H.A., Carels, L.B., Boelens, H.F.M., Kraak, J.C., and Poppe, H., Comparison of five methods for the study of drug-protein binding in affinity capillary electrophoresis, J. Chromatogr. A, 777, 311, 1997. 42. Rao, J., Colton, I.J., and Whitesides, G.M., Using capillary electrophoresis to study the electrostatic interactions involved in the association of D-Ala-D-Ala with vancomycin, J. Am. Chem. Soc., 119, 9336, 1997. 43. Amini, A. and Westerlund, D., Evaluation of association constants between drug enantiomers and human a1-acid glycoprotein by applying a partial-filling technique in affinity capillary electrophoresis, Anal. Chem., 70, 1425, 1998. 44. Tanaka, Y. and Terabe, S., Separation of the enantiomers of basic drugs by affinity capillary electrophoresis using a partial filling technique and α1-acid glycoprotein as chiral selector, Chromatographia, 44, 119, 1997. 45. Harmon, B.J., Patterson, D.H., and Regnier, F.E., Mathematical treatment of electrophoretically mediated microanalysis, Anal. Chem., 65, 2655, 1993. 46. Patterson, D.H., Harmon, B.J., and Regnier, F.E., Electrophoretically mediated microanalysis of calcium, J. Chromatogr., 662, 389, 1994. 47. Avila, L.Z. and Whitesides, G.M., Catalytic activity of native enzymes during capillary electrophoresis: an enzymatic microreactor, J. Org. Chem., 58, 5508, 1993. 48. Chang, H.-T. and Yeung, E.S., Variability of intracellular lactate dehydrogenase isoenzymes in single human erythrocytes, Anal. Chem., 65, 2947, 1993. 49. Xue, Q. and Yeung, E.S., Variability of intracellular lactate dehydrogenase isoenzymes in single human erythrocytes, Anal. Chem., 66, 1175, 1994. 50. Zhao, D.S. and Gomez, F.A., Enzyme-catalyzed microreactions using capillary electrophoresis: a quantitative study, Chromatographia, 44, 514, 1997. 51. Zugel, S.A., Burke, B.J., Regnier, F.E., and Lytle, F.E., Electrophoretically mediated microanalysis of leucine aminopeptidase using two-photon excited fluorescence detection on a microchip, Anal. Chem., 72, 5731, 2000.

DK4478_C003.fm Page 171 Friday, September 22, 2006 8:39 AM

Applications of Capillary Electrophoresis

171

52. Jin, Z., Chen, R., and Colon, L.A., Determination of glucose in submicroliter samples by CE-LIF using precolumn or on-column enzymatic reactions, Anal. Chem., 69, 1326, 1997. 53. Cole, L.J. and Kennedy, R.T., Selective preconcentration for capillary zone electrophoresis using protein G immunoaffinity capillary chromatography, Electrophoresis, 16, 549, 1995. 54. Patterson, D.H., Harmon, B.J., and Regnier, F.E., Dynamic modeling of electrophoretically mediated microanalysis, J. Chromatogr. A, 732, 119, 1996. 55. Harmon, B.J., Leesong, I., and Regnier, F.E., Moving boundary electrophoretically mediated microanalysis, J. Chromatogr. A, 726, 193, 1996. 56. Kwak, E.-S., Esquivel, S., and Gomez, F.A., Optimization of capillary electrophoresis conditions for in-capillary enzyme-catalyzed microreactions, Anal. Chim. Acta, 397, 183, 1999. 57. Lillard, S.J. and Yeung, E.S., Capillary electrophoresis for the analysis of single cells: laser-induced fluorescence detection, in The Handbook of Capillary Electrophoresis, 2nd ed., Landers, J., ed., CRC Press, Boca Raton, FL, 1996, p. 523. 58. Xue, Q. and Yeung, E.S., Differences in the chemical reactivity of individual molecules of an enzyme, Nature, 373, 681, 1995. 59. Zhao, D. S. and Gomez, F.A., Double enzyme-catalyzed microreactors using capillary electrophoresis, Electrophoresis, 19, 420, 1998. 60. Zhang, Y., El-Maghrabi, R., and Gomez, F.A., Use of capillary electrophoresis and indirect detection to quantitate in-capillary enzyme-catalyzed microreactions, Analyst, 125, 685, 2000. 61. Swanek, F.D., Ferris, S.S., and Ewing, A.G., Capillary electrophoresis for the analysis of single cells: electrochemical, mass spectrometric, and radiochemical detection, in The Handbook of Capillary Electrophoresis, 2nd ed., Landers, J., ed., CRC Press, Boca Raton, FL, 1996, p. 495. 62. Mechref, Y. and Rassi, Z.E., Capillary enzymophoresis of nucleic acid fragments using coupled capillary electrophoresis and capillary enzyme microreactors having surface-immobilized RNA-modifying enzymes, Electrophoresis, 16, 2164, 1995. 63. Emmer, A. and Roeraade, J., Capillary electrophoresis combined with an on-line micro post-column enzyme assay, J. Chromatogr., 662, 375, 1994. 64. Williams, D.H., Davies, N.L., Zerella, R., and Bardsley, B., Noncovalent interactions: defining cooperativity. Ligand binding aided by reduced dynamic behavior of receptors. Binding of bacterial cell wall analogues to ristocetin A, J. Am. Chem. Soc., 126, 2042, 2004. 65. Heck, A.J.R., Bonnick, P.J., Breukink, E., Morris, D., and Wills, M., Modification and inhibition of vancomycin group antibiotics by formaldehyde and acetaldehyde, Chem. Eur., 7, 910, 2001. 66. Chiosis, G. and Boneca, I.G., Selective cleavage of D-Ala-D-Lac by small molecules: re-sensitizing resistant bacteria to vancomycin, Science, 293, 1484, 2001. 67. Kerns, R., Dong, S.D., Fukuzawa, S., Carbeck, J., Kohler, J., Silver, L., and Kahne, D., The role of hydrophobic substituents in the biological activity of glycopeptide antibiotics, J. Am. Chem. Soc., 122, 12608, 2000. 68. Gasper, M.P., Berthod, A., Nair, U.B., and Armstrong, D.W., Comparison and modeling study of vancomycin, ristocetin A, and teicoplanin for CE enantioseparations, Anal. Chem., 68, 2501, 1996. 69. Pearson, A.J. and Heo, J.-N., Approaches to the fully functionalized DEF ring system of ristocetin A via highly selective ruthenium-promoted SNAr reaction, Org. Lett., 19, 2987, 2000.

DK4478_C003.fm Page 172 Friday, September 22, 2006 8:39 AM

172

Advances in Chromatography, Volume 45

70. Beauregard, D.A., Maguire, A.J., Williams, D.H., and Reynolds, P.E., Semiquantitation of cooperativity in binding of vancomycin-group antibiotics to vancomycinsusceptible and resistant organisms, Antimicrob. Agents Chemother., 41, 2418, 1997. 71. Shiozawa, H., Chia, B.C.S., Davies, N.L., Zerella, R., and Williams, D.H., Cooperative binding interactions of glycopeptide antibiotics, J. Am. Chem. Soc., 124, 3914, 2002. 72. Lee, J.-G., Sagui, C., and Roland, C., First principles investigation of vancomycin and teicoplanin binding to bacterial cell wall termini, J. Am. Chem. Soc., 126, 8384, 2004. 73. Chinchilla, D., Zavaleta, J., Ramirez, A., and Gomez, F.A., unpublished results.

DK4478_book.fm Page 173 Thursday, September 14, 2006 1:35 PM

4

Aptamers as Molecular Recognition Elements in Chromatographic Separations Daniela Hutanu and Vincent T. Remcho

CONTENTS I. Introduction................................................................................................173 II. Aptamers as High-Performance Liquid Chromatography Sorbents......... 175 III. Aptamers as Micro Liquid Chromatography Sorbents............................. 182 IV. Aptamers as Capillary Electrochromatography Sorbents .........................185 V. Conclusion ................................................................................................. 193 References..............................................................................................................194

I. INTRODUCTION In the last few years, aptamers have become a practical alternative to antibodies as affinity-related separation and purification stationary phases. Aptamers are ribonucleic acid (RNA) or single-stranded deoxyribonucleic acid (DNA) molecules synthesized in vitro and especially selected for a desired compound based on their recognition of a specific epitope on that molecule. The affinity binding to a certain target is a combination of hydrophobic and steric interactions, hydrogen bonding, and electrostatic forces [1], which can also be exploited for partitioning chromatography of nontarget compounds. Aptamers are selected from a large oligonucleotide pool via a succession of elution chromatography steps and amplified to yield the desired quantity. The selection process is called systematic evolution of ligands by exponential enrichment (SELEX), which was first used for aptamer selection in 1990 [2,3]. Aptamers can be selected for a variety of targets, and several have found use in chromatographic applications (Table 4.1). In 2003 Bowser and Mendonsa [4] pioneered the use of capillary electrophoresis (CE) in the aptamer selection process, managing to reduce the selection time from the few weeks required in the SELEX process to a few days. In CE-SELEX, the

173

DK4478_book.fm Page 174 Thursday, September 14, 2006 1:35 PM

174

Advances in Chromatography, Volume 45

TABLE 4.1 Targets for Aptamers Used as Stationary Phases in Chromatography Aptamer Type

SELEX Target

Reference

Single-stranded DNA Single-stranded DNA Single-stranded DNA

Human L-selectin D-arginine-vasopressin D-adenosine ATP L-tyrosinamide L-arginine L-tyrosine Flavin mononucleotide Flavin adenine dinucleotide Thrombin

23 25 26

Single-stranded DNA RNA RNA RNA Single-stranded DNA

27 28 31 33 35

aptamer-target complexes are separated based on the shift in their migration time compared to the unbound sequences. CE-SELEX has recently become a potent selection method for aptamers specific for human IgE [5], HIV-1 reverse transcriptase [6], and neuropeptide Y [7]. Drabovich et al. [8] used another approach — equilibrium capillary electrophoresis of equilibrium mixtures (ECEEM) or affinity capillary electrophoresis (ACE) — for selecting aptamers with predefined equilibrium dissociation constants (Kd) of aptamer-target complexes. In ECEEM, the target concentration in the run buffer is equal to the equilibrium concentration and the aptamer is more or less bound to the target, or in free form, depending on the Kd. The Kd value determines the migration time of the aptamer-target complexes. An aptamer for MutS protein was selected via this method [8]. Prior to aptamers, the affinity separation field involved the use of antibodies and molecular imprinted polymers (MIPs). These two approaches, despite notable successes, exhibit certain drawbacks. The large molecular size of the antibody leads to small column capacities due to reduced surface loading. The antibody production involves in vitro experiments on live organisms, thus requiring nontoxic and highly immunogenic targets. The antibody molecule is easily denaturated and degraded, making tailoring of the elution conditions and structural modifications very difficult. Imprinted polymers lack recognition of all compound classes, exhibit a relatively modest selectivity and efficiency, and are typified by low sample loading capacity. On the other hand, aptamers are custom synthesized in a only a few weeks by automated processes, have a relatively small size that allows for greater surface coverage, and they can be easily modified to facilitate attachment to surfaces, labeling, linking to other reagents or changes in conformation for selectivity alteration. Aptamers are stable molecules, compatible with a wide range of solvents and separation conditions. The molecular recognition abilities of aptamers are comparable to those of antibodies and MIPs. Aptamers have been exploited in various analytical applications

DK4478_book.fm Page 175 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

175

and their use has been reviewed [9,10]. Recently (as early as 1999) DNA and RNA aptamers have been used effectively as stationary phases in liquid chromatography [11–18] and capillary electrochromatography [13,19–22] for the purification or separation of compounds as large as proteins to compounds as small as amino acids (Table 4.2). The separation mechanism was based both on target epitope recognition by the aptamer [11,13–18] and on the simple partitioning of nontarget molecules on the stable aptamer stationary phase [19–22].

II. APTAMERS AS HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY SORBENTS Romig et al. [18] were the first to assess the performance of an aptamer ligand as a high-performance liquid chromatography (HPLC) affinity stationary phase component. They immobilized a 5′-biotinylated DNA aptamer (Table 4.2) on a streptavidin-linked resin and subsequently used the aptamer sorbent for packing a 0.5 cm inside diameter × 5 cm chromatographic column. The 36-base aptamer was initially selected for binding the lectin domain of human L-selectin [23], a cell adhesion protein that mediates the attachment of lymphocytes to lymph node high endothelial venules [24]. The Kd for this particular protein-aptamer complex was found to be 2 × 10–9 M [23]. The authors demonstrated the successful application of the aptameric sorbent in the affinity purification of Lselectin receptor globulin (LS-Rg), a recombinant fusion protein that contains the extracellular domain of L-selectin. The fusion protein was produced by selected LSRg-expressing Chinese hamster ovary (CHO) cells and presented to the HPLC column in the clear supernatant obtained by centrifugation of the growth medium. In a single step, the achieved purification was 63%, with 83% recovery. Figure 4.1 presents the chromatographic results of LS-Rg purification from an unfractionated CHO cell medium. Brumbt et al. [11], Michaud et al. [15,16], and Ravelet et al. [17] were the first to apply aptamer sorbents to chiral separations of target and nontarget molecules by HPLC. The streptavidin-biotin interaction was employed to immobilize biotinylated aptamers on POROS streptavidin chromatographic media. The DNA and RNA aptamer chiral stationary phases (CSPs) successfully discriminated between enantiomers of vasopressin [15], adenosine and tyrosinamide [16], arginine [11], tyrosine and tyrosine-related [17] compounds (Table 4.3). Several chromatographic properties acquired for the HPLC aptamer chiral sorbents are also included in Table 4.3. In 2003 Michaud et al. [15] exploited the stereoselective affinity of a 55-base DNA aptamer (Table 4.2) for D-arginine-vasopressin to create a CSP capable of separating the vasopressin enantiomers. Williams et al. [25] showed that the 55-base oligonucleotide does not exhibit affinity for the L-dipeptide, but interacts with the D-dipeptide with an association constant close to 1.1/μM–1. The aptamer CSP was capable of binding D-vasopressin while, as expected, the L-enantiomer eluted in the void volume under all experimental conditions (Figure 4.2). The effect of pH, the ionic strength of the mobile phase, and the column temperature on the separation ability of the column was explored. The retention factor for the D-enantiomer

Note: HPLC, high-performance liquid chromatography; OTCEC, open tubular capillary electrochromatography.

′5-GGGGTTGGGGTGTGGGGTTGGGG-3′

5′-GGTTGGTTTGGTTGG-3′ ′5-GGTTGGTGTGGTTGG-3′

5′-GGTTTTGGTTTTGGTTTTGG-3′

Separated Compounds

L-selectin receptor globulin Vasopressin enantiomers Adenosine enantiomers Tyrosinamide enantiomers Arginine enantiomers Tyrosine enantiomers Enantiomers of tyrosine-related compounds Micro liquid Adenosine and adenosine-related chromatography compounds Micro liquid Flavin mononucleotide and thiourea, chromatography flavin mononucleotide and anthracene OT-CEC D-trp and D-tyr D-trp and L-trp OT-CEC Alanyl dipeptides OT-CEC Naphthalene and benzo[a]pyrene Naphthalene and benzo(ghi)-perylene OT-CEC Trp-Arg and Arg-Trp OT-CEC Alanyl dipeptides Homodipeptides OT-CEC Bovine β-lactoglobulin variants A and B OT-CEC Trp-Arg and Arg-Trp OT-CEC Alanyl dipeptides, homodipeptides

HPLC HPLC HPLC HPLC HPLC HPLC

Separation Method

21 19 22

19 22

22 20

20

13

14

18 15 16 16 11 17

Reference

176

5′-GGCGUGUAGGAUAUCGUGUUCGAGAAGGACACGCC-3′

5′-GTGCTTGGGGGAGTATTGCGGAGGAAAGCGGCCCTGCTGAAG-3′

5′-GCGGTAACCAGTACAAGGTGCTAAACGTAATGGCGC-3′ 5′-TCACGTGCATGATAGACGGCGAAGCCGTCGAGTTGCTGTGTGCCGATGCACGTGA-3′ 5′-ATTATACCTGGGGGAGTATTGCGGAGGAAGGTATAAT-3′ 5′-AATTCGCTAGCTGGAGCTTGGATTGATGTGGTGTGTGAGTGCGGTGCCC-3′ 5′-GACGAGAAGGAGCGCUGGUUCUACUAGCAGGUAGGUCACUCGUC-3′ 5′-GGGCAGUCAACUCGUAAGAUGGCCUUACAGCGGUCAAUACGGGGGUCAUCAGAUAGGGAGGCC-3′

Immobilized Aptamer

TABLE 4.2 Aptamers Employed as Chromatographic Stationary Phases, the Separation Method Used, and the Target Analytes

DK4478_book.fm Page 176 Thursday, September 14, 2006 1:35 PM

Advances in Chromatography, Volume 45

DK4478_book.fm Page 177 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

177

Absorbance (280 nm)

2.0

EDTA elution

1.0

0.0 0

20

40

60

80 100 Time (min)

120

140

160

FIGURE 4.1 HPLC chromatogram of LS-Rg purification on the L-selectin-specific aptameric sorbent; 694 mg total protein from unfractionated CHO cell medium was loaded on the column. LS-Rg was eluted with a linear ethylenediaminetetraacetic acid (EDTA) gradient (0 to 50 mM EDTA in 10 min). (From Romig, T.S., Bell, C., and Drolet, D.W., Aptamer affinity chromatography: combinatorial chemistry applied to protein purification, J. Chromatogr. B, 731, 275, 1999. With permission.)

decreased with the potassium chloride (KCl) concentration in the mobile phase, but the pH and temperature variation did not lead to a significant variation in kd. One year later, Michaud et al. [16] expanded the application of aptamer CSPs to HPLC separations of small molecule enantiomers. The immobilized DNA aptamers were specific for D-adenosine [26] and L-tyrosinamide [27] (see Table 4.2 for their sequences). Both stationary phases exhibited affinity for the particular enantiomer they were selected for by the SELEX process (Figures 4.3 and 4.4). The sorbents were evaluated in different temperature conditions, as it was previously determined [15] that temperature plays an important role in the chromatographic resolution of the racemic mixture. Table 4.3 summarizes the reported chromatographic parameters for the two aptamer CSPs. By decreasing the column temperature, the resolution was enhanced at the expense of the total analysis time. Brumbt et al. [11] attached an RNA aptamer to a chromatographic support and employed it in target-specific HPLC separations of arginine enantiomers. A 44-mer anti-L-arginine D-RNA aptamer [28] (Table 4.2) was initially used as the affinity component on the stationary phase. The CSP proved successful in separating Larginine from D-arginine (Figure 4.5), but its target retention decreased in 8 days by 65% due to enzymatic cleavage by RNase present in the environment and reagents. The authors decided to examine the practical application and stability of CSPs of the mirror image of the initial aptamer, given that enzymes generally show affinity for only D-enantiomers. The arginine enantiomers eluted in reverse order on the L-aptamer chiral sorbent (Figure 4.6). The L-RNA showed affinity for the mirror image of the initial target due to chiral inversion. As predicted by the authors, the new aptameric CSP proved

DK4478_book.fm Page 178 Thursday, September 14, 2006 1:35 PM

178

Advances in Chromatography, Volume 45

TABLE 4.3 Target Analytes and HPLC Chromatographic Parameters for the Chiral Aptamer Stationary Phases Used in Separations

Separated Enantiomers Vasopressin Adenosine

Tyrosinamide

Arginine

Tyrosine α-methyl-tyrosine DOPA Tryptophan 1-methyltryptophan 5-hydroxyltryptophan N-acetyl-tryptophan 3-benzothienylalanine 2-quinolyl-alanine 2-naphtyl-alanine 1-naphtyl-alanine

Kd/chr Hours, α Chromatographic Reduced Temperature EnantioRetention Rs Plate (°C) selectivity Factor Resolution Height Reference — 12 16 20 24 28 32 20 23 26 29 32 4 8 11 14 17 10 10 10 10 10

— 3.72 3.62 3.57 3.44 3.27 3.18 78 50 31 19 11 — — — — — 28.06 21.91 7.20 4.09 6.98

— 7.66 5.60 3.91 2.46 1.65 1.10 2.59 1.72 1.14 0.51 0.28 2.20 1.73 1.43 0.99 0.83 0.57 0.59 0.83 1.42 1.50

— 1.88 1.41 1.26 1.23 1.03 0.94 3.13 1.36 1.14 0.82 — 1.70 1.30 1.26 1.18 1.02 — — — — —

35–40 — — — — — — — — — — — 35–60 — — — — — — — — —

15 16

10

3.62

1.09

—

—

17

10 10

9.48 1.73

0.49 2.13

— —

— —

17 17

10 10 10

1.88 3.06 1.32

1.43 5.29 3.03

— — —

— — —

17 17 17

16

11

17 17 17 17 17

DK4478_book.fm Page 179 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

179

L

FIGURE 4.2 HPLC separation of vasopressin enantiomers on a D-arginine-vasopressin-specific aptameric stationary phase. Conditions: 2.1 × 30 mm column, 20°C; mobile phase: 100 mM KCl, 3 mM MgCl2, pH 7.0; flow rate: 150 μl/min; injection: 100 nl at a concentration of 0.9 mM; detection at 195 nm. (From Michaud, M., Jourdan, E., Villet, A., Ravel, A., Grosset, C., and Peyrin, E., A DNA aptamer as a new target-specific chiral selector for HPLC, J. Am. Chem. Soc., 125, 8672, 2003. With permission.)

D

0

5 Time (min)

10

L D

28°C

0.003

FIGURE 4.3 HPLC separation of adenosine enantiomers using a D-adenosine-specific DNA aptamer as CSP. Conditions: 370 × 0.76 mm inside diameter column; mobile phase: 20 mM phosphate buffer, 25 mM KCl, 1.5 mM MgCl2, pH 6.0; injection: 70 pmol D-,Ladenosine in 100 nl; flow rate: 50 μl/min; detection at 260 nm. (From Michaud, M., Jourdan, E., Ravelet, C., Villet, A., Ravel, A., Grosset, C., and Peyrin, E., Immobilized DNA aptamers as target-specific chiral stationary phases for resolution of nucleoside and amino acid derivative enantiomers, Anal. Chem., 76, 1015, 2004. With permission.)

AU 0.000

L

0.003

16°C

AU

D

0.000 0

20 Time (min)

40

DK4478_book.fm Page 180 Thursday, September 14, 2006 1:35 PM

180

Advances in Chromatography, Volume 45

D 0.008

29°C L

AU 0.000 D 0.008

20°C

AU L 0.000 0

20

40

Time (min)

FIGURE 4.4 HPLC separation of tyrosinamide enantiomers using an L-tyrosinamide-specific DNA aptamer as CSP. Conditions: 250 × 0.76 mm inside diameter column; mobile phase: 20 mM phosphate buffer, 25 mM KCl, 1.5 mM MgCl2, pH 6.0; injection: 70 pmol D-,Ladenosine in 100 nl; flow rate: 20 μl/min; detection at 224 nm. (From Michaud, M., Jourdan, E., Ravelet, C., Villet, A., Ravel, A., Grosset, C., and Peyrin, E., Immobilized DNA aptamers as target-specific chiral stationary phases for resolution of nucleoside and amino acid derivative enantiomers, Anal. Chem., 76, 1015, 2004. With permission.)

D

L D

0

5

L

10

15

0

5

10

15

Time (min)

Time (min)

FIGURE 4.5 HPLC separation of arginine enantiomers on a D-RNA CSP. Conditions: 10 ng D-,L-arginine injected; 370 × 0.76 mm inside diameter column; mobile phase: 25 mM phosphate buffer, 25 mM NaCl, 5 mM MgCl2, pH 7.3; column temperature: 4°C; 100 nl injection volume; 50 μl/min flow rate; detection at 208 nm. (From Brumbt, A., Ravelet, C., Grosset, C., Ravel, A., Villet, A., and Peyrin, E., Chiral stationary phase based on a biostable L-RNA aptamer, Anal. Chem., 77, 1993, 2005. With permission.)

FIGURE 4.6 HPLC separation of arginine enantiomers using an L-RNA CSP. Conditions: 50 ng D-,L-arginine injected; 370 × 0.76 mm inside diameter column; mobile phase: 25 mM phosphate buffer, 25 mM NaCl, 5 mM MgCl2, pH 7.3; column temperature: 4°C; 100 nl injection volume; 50 μl/min flow rate; detection at 208 nm. (From Brumbt, A., Ravelet, C., Grosset, C., Ravel, A., Villet, A., and Peyrin, E., Chiral stationary phase based on a biostable L-RNA aptamer, Anal. Chem., 77, 1993, 2005. With permission.)

DK4478_book.fm Page 181 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

181

D+L

0

30

15 Time (min) (a) L D

0

15 Time (min)

30

(b)

FIGURE 4.7 HPLC chromatograms of a racemic mixture of tyrosine injected onto (a) nonmodified streptavidin and (b) tyrosine-specific L-RNA aptamer columns. Conditions: 350 × 0.76 mm inside diameter column; mobile phase: 8 mM Tris-HCl buffer, 25 mM NaCl, 5 mM MgCl2, pH 7.4; column temperature: 10°C; injection: 100 nl at 0.50 mM concentration; 15 μl/min flow rate; detection at 220 nm. (From Ravelet, C., Boulkedid, R., Ravel, A., Grosset, C., Villet, A., Fize, J., and Peyrin, E., A L-RNA aptamer chiral stationary phase for the resolution of target and related compounds, J. Chromatogr. A, 1076, 62, 2005. With permission.)

stable in time after 30 days of use and about 1600 volumes of buffer passed through the column. Ravelet et al. [17] further employed the mirror-image strategy [29,30] in the development of an RNA CSP for chiral separation of tyrosine and some related compounds. A 63-base RNA aptamer (Table 4.2) specific for L-tyrosine [31] was used as the chiral selector on the HPLC stationary phase. The same streptavidinbiotin bridge was used for immobilization of the oligonucleotide on a streptavidin media slurry. The L-RNA stationary phase discriminated the tyrosine enantiomers, showing affinity for the D-enantiomer (Figure 4.7). Furthermore, the authors demonstrated the ability of the L-RNA CSP to resolve racemates of several compounds structurally related to tyrosine. Figure 4.8 presents representative HPLC chromatograms of tyrosine-related compounds on the L-RNA stationary phase. By examining the collected chromatographic data (shown partially in Table 4.3), the authors concluded that the anti-D-tyrosine CSP can be exploited

DK4478_book.fm Page 182 Thursday, September 14, 2006 1:35 PM

182

Advances in Chromatography, Volume 45

L

L D

0

15

30 Time (min)

45

60

0

15

D

30 Time (min)

(a)

45

60

(b)

L

L D D

0

15 Time (min) (c)

30

0

15

30 Time (min)

45

(d)

FIGURE 4.8 HPLC separation of (a) tryptophan, (b) 2-quinolyl-alanine, (c) N-acetyl-tryptophan, and (d) 1-methyl-tryptophan using a tyrosine-specific L-RNA aptamer CSP. Conditions: 350 × 0.76 mm inside diameter column; mobile phase: 8 mM Tris-HCl buffer, 25 mM NaCl, 5 mM MgCl2, pH 7.4; column temperature: 10°C; injection: 100 nl at 0.50 mM concentration; 15 μl/min flow rate; detection at 220 nm. (From Ravelet, C., Boulkedid, R., Ravel, A., Grosset, C., Villet, A., Fize, J., and Peyrin, E., A L-RNA aptamer chiral stationary phase for the resolution of target and related compounds, J. Chromatogr. A, 1076, 62, 2005. With permission.)

for chiral separation of amino acid-related compounds with a large aromatic side chain in the β position of the asymmetric carbon atom.

III. APTAMERS AS MICRO LIQUID CHROMATOGRAPHY SORBENTS Deng et al. [14] immobilized a biotinylated DNA aptamer (Table 4.2) isolated for adenosine and ATP [26] on streptavidin-POROS media and streptavidin-porous glass beads. The particles were then packed in fused silica capillary columns and further used for frontal analysis experiments and weak affinity [32] liquid chromatography separations of adenosine and adenosine-related compounds. The dissociation constants of the aptamer complex with cyclic AMP (cAMP), AMP, ATP, ADP, and adenosine (138 ± 18, 58 ± 2, 38 ± 2, 28 ± 6, and 3 ± 1 μM, respectively, on glass, similar on POROS beads except for a higher Kd for adenosine) were determined by frontal analysis. The difference in Kd values allows the selective retention and separation of these compounds on the aptamer sorbents (Figure 4.9). After evaluating the effects of mobile phase composition change on sorbent resolution and selectivity, the authors concluded that a phosphate buffer with a lower pH, with a low concentration of sodium chloride (NaCl) and fair Mg2+ is the most

DK4478_book.fm Page 183 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

183

cAMP

0.018

A260 nm

0.016 ATP ADP

0.014

AMP

Adenosine

0.012 AA

0.010 0.008 150

300

450 600 Time (s)

750

900

(a) 0.06 0.05

GTP, GDP, GMP, AA cAMP

A260 nm

NAD 0.04 0.03

AMP ATP ADP

Adenosine

0.02 0.01 300

600

900

1200

1500

Time (s) (b)

FIGURE 4.9 Liquid chromatography separation of adenosine-related compounds on adenosine/ATP-specific aptamer sorbents. Conditions: mobile phase 20 mM KH2PO4, 20 mM NaCl, 10 mM MgCl2, pH 6.6. (a) 20 cm long, 150 μm inside diameter capillary column packed with POROS streptavidin aptamer; flow rate: 0.072 cm/sec, injection volume: 80 nl. (b) 7 cm long, 50 μm inside diameter capillary column packed with controlled pore glass (CPG) streptavidin-aptamer porous glass beads; flow rate: 0.050 cm/sec, injection volume: 10 nl. (From Deng, Q., German, I., Buchanan, D., and Kennedy, R.T., Retention and separation of adenosine and analogues by affinity chromatography with an aptamer stationary phase, Anal. Chem., 73, 5415, 2001. With permission.)

efficient mobile phase for separation due to alterations of ion exchange and Coulombic interactions between analyte and aptamer, and also modifications of the aptamer secondary structure. Clark and Remcho [13] developed an aptameric stationary phase by binding a 35-base RNA aptamer specific for flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD) [33] on the inner surface of fused silica capillaries. The threestep aptamer immobilization method modified from Maskos and Southern [34] consisted of derivatization of the silanol groups with (glycidoxypropyl)trimethoxysilane (GOPS), followed by reaction with 1,1′-carbonyldiimidazole (CDI), and finally by replacement of the imidazole group with the amine-modified aptamer (Figure 4.10).

DK4478_book.fm Page 184 Thursday, September 14, 2006 1:35 PM

184

Advances in Chromatography, Volume 45

OCH3 Si OH

+

H3CO Si

O

O

Xylene, Hunig’s

OCH3

OCH3 Si O

base, 80°C

Si

O

O

OCH3

GOPS HClaq (pH 3.5), 80°C

OH Si O

Si

O O

O

CH3CN N

OH

OH

N

OH

O N

N

N

N

80°C

+

Si O

Si OH

OH O OH

CDI

+ O O P

NH2

Aptamer −

O

pH = 8.0 to dryness under Ar O

O Si O

Si

O

O OH

N

O P

Aptamer −

O

FIGURE 4.10 RNA aptamer derivatization of the inner surface of a fused silica capillary. (From Clark, S.L. and Remcho, V.T., Electrochromatographic retention studies on a flavinbinding RNA aptamer sorbent, Anal. Chem., 75, 5692, 2003. With permission.)

The chromatographic properties of the aptameric sorbent were evaluated in liquid chromatography separations of a mixture of FMN and FAD, and also FMN with anthracene or thiourea, small molecules lacking the flavin moiety involved in the affinity recognition. Although the new stationary phase was unable to resolve the two biological cofactors, it discriminated among FMN and the unrelated molecule, anthracene (Figure 4.11). Figure 4.12 shows that the RNA sorbent retains FMN more strongly than thiourea in a high organic content mobile phase, but exhibits more affinity for thiourea under aqueous conditions, therefore being selective, but not specific for FMN. Chung et al. [12] attached an RNA aptamer selected by SELEX for hepatitis C virus (HCV) RNA replicase to PS-co-PEG (CutiCore) beads via a photocleavable linker (Figure 4.13). The aptamer beads were packed in a silicon-on-glass microfluidic channel and the sorbent was evaluated for purification of HCV RNA replicase from a complex protein mixture. The proteins, previously labeled with FITC were injected in the channel by pressure-driven flow, then cleaved by exposure to 360 nm ultraviolet (UV) light, eluted with buffer, and fluorescently detected at 500 nm. The fluorescence signal intensity for HCV RNA replicase was significant compared with

DK4478_book.fm Page 185 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

185

15 1 10 5

Absorbance (mAU)

0 0

1

2

3

−5 (a) 15 1

10 5

2

0 −5

0

1

2

3

Time (minutes) (b)

FIGURE 4.11 Liquid chromatography chromatograms of (a) anthracene (1) only and (b) a mixture of anthracene (1) and FMN (2) on an RNA aptamer-derivatized capillary. Conditions: 50 mbar constant head pressure at the inlet, mobile phase: 50 mM Tris-HCl, pH 7.6. (From Clark, S.L. and Remcho, V.T., J. Sep. Sci., 26, 1451, 2003. With permission.)

the other labeled proteins, demonstrating the specificity of the aptamer sorbent for its SELEX target (Figure 4.14).

IV. APTAMERS AS CAPILLARY ELECTROCHROMATOGRAPHY SORBENTS McGown and her coworkers [19–22] were the first to report the separation of nontarget molecules on an aptamer sorbent. They immobilized aptamers on silica surfaces and applied the aptameric sorbents in capillary electrochromatography (CEC) separations of compounds not related to the SELEX target. Kotia et al. [20] used two aptamers (Table 4.2) as stationary phases for CEC separations of amino acids (D-trp and D-tyr), enantiomers (D-trp and L-trp), and polycyclic aromatic hydrocarbons (naphthalene and benzo[a]pyrene or benzo(ghi)perylene). One aptamer was initially selected for thrombin [35] and the second was a related single-stranded DNA molecule with 20 bases, both forming G-quartets in the presence of K+ ions [36]. The authors demonstrated that the presence of Gquartets is essential for polynuclear aromatic hydrocarbon (PAH) separation.

DK4478_book.fm Page 186 Thursday, September 14, 2006 1:35 PM

186

Advances in Chromatography, Volume 45

3

−2

1

0

1

2

3

4

5

(a) 2

15 1

10

Absorbance (mAU)

5 0 −5

1

0

2

3

4

5

(b) 10

1, 2

5 0 −5

0

1

2

3

4

5

4

5

(c) 2

10

1

5 0 −5

0

1

2 3 Time (minutes) (d)

FIGURE 4.12 Liquid chromatography chromatograms of FMN (1) and thiourea (2) on an RNA aptamer stationary phase. (a) FMN in 50 mM Tris-HCl at pH 7.6; (b) FMN and thiourea in 50 mM Tris-HCl at pH 7.6; (c) FMN and thiourea in 50%v acetonitrile mobile phase; (d) FMN and thiourea in 80%v acetonitrile mobile phase. (From Clark, S.L. and Remcho, V.T., J. Sep. Sci., 26, 1451, 2003. With permission.) O N H

O MeO

NO2 H N

O

O

N H

O

O N H

O 2

O

N O

Reactive site

FIGURE 4.13 Structure of the photocleavable site on the bead, linked to a long spacer with an active ester group for aptamer coupling. (From Chung, W.-J., Kim, M.-S., Cho, S., Park, S.-S., Kim, J.-H., Kim, Y.-K., Kim, B.-G., and Lee, Y.-S., Electrophoresis, 26, 694, 2005. With permission.)

DK4478_book.fm Page 187 Thursday, September 14, 2006 1:35 PM

800 (e)

600 400 (b) (a)

200

(d) (c)

0 0

50

187

Peak area (arb. unit)

Fluorescence intensity (arb. unit)

Aptamers as Molecular Recognition Elements

6000 4000 2000 0

100 150 200 250

(a)

(b)

(c)

(d)

(e)

Time (s)

FIGURE 4.14 Microaffinity liquid chromatography chromatograms on RNA aptamer microbeads: (a) streptavidin-FITC; (b) ovalbumin-FITC; (c) human serum albumin (HSA)-FITC; (d) HCV RNA replicase-FITC; (e) mixture containing all the proteins above at a concentration of 68 nM, except for ovalbumin-FITC (380 nM). Conditions: flow injection (5 μl/h) at 4°C for 30 min; mobile phase: 25 mM Tris-HCl, 10 mM MgCl2, pH 8.0. (From Chung, W.-J., Kim, M.-S., Cho, S., Park, S.-S., Kim, J.-H., Kim, Y.-K., Kim, B.-G., and Lee, Y.-S., Electrophoresis, 26, 694, 2005. With permission.)

The aptamers were attached to the inner walls of fused silica capillaries using three different linkers [37–39], shown in Figure 4.15. Figure 4.16 shows a separation of D-trp and D-tyr achieved on the sorbent with the 20-mer aptamer immobilized using the adapted linker from Guo et al. [37]. The compounds were baseline resolved, but the retention times changed over time due to stationary phase degradation in aqueous conditions. O S

CH3 Si

O

Si

HN

S

C

N H

N H

O

C

N H

P

O

aptamer

−

O

O CH3 CH3 O Si

O

Si O

O

O

H N

O

O

O

HO

P

O

aptamer

−

O

CH3 CH3 O

O Si

O

Si

NH

C

O NH

C

CH2

S

(CH2)6

O

aptamer

O CH3

FIGURE 4.15 Structures of linkers used to attach DNA aptamers to fused silica capillaries inner walls. (From Kotia, R.B., Li, L., and McGown, L.B., Separation of nontarget compounds by DNA aptamers, Anal. Chem., 72, 827, 2000. With permission.)

DK4478_book.fm Page 188 Thursday, September 14, 2006 1:35 PM

188

Advances in Chromatography, Volume 45

0.020 0.015 0.010 0.005 0.000 10

15

20

25

30

35

(a) 0.05 0.04 0.03 0.02 0.01 Absorbance

0.00 10

15

20

25

30

35

40

45

30

35

40

45

(b) 0.05 0.04 0.03 0.02 0.01 0.00 10

15

20

25 (c)

0.025 0.020 0.015 0.010 0.005 0.000 −0.005 10

12

14

16

18

20

(d)

FIGURE 4.16 (a–c) CEC separation of amino acids using a capillary coated with the 20mer aptamer: (a) 2 mM D-trp and 1.5 mM D-tyr; (b) 4 mM D-trp; (c) 3 mM D-tyr. (d) CEC of the same mixture as in (a) on an unmodified capillary. Conditions: capillaries 375 μm outside diameter × 75 μm inside diameter, 47 cm total length, 40 cm to the detection window; detection at 280 nm; mobile phase: 50 mM Tris amine buffer, 5 mM KCl, pH 7.2; 15 kV, positive polarity, 5 sec pressure injection, temperature 16°C. (From Kotia, R.B., Li, L., and McGown, L.B., Separation of nontarget compounds by DNA aptamers, Anal. Chem., 72, 827, 2000. With permission.)

DK4478_book.fm Page 189 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

189

0.02

0.01

0.00 0.02

Absorbance

0.01

0.00 0.03 0.02 0.01 0.00 0.02

0.01

0.00 0

1

2

3

4

5

Migration time (min)

FIGURE 4.17 CEC separation of enantiomers using a capillary coated with the 20-mer aptamer. From top to bottom: 0.5 mM D-trp and 0.5 mM L-trp; repeated run of the same mixture; 0.5 mM D-trp and 0.1 mM L-trp; 0.1 mM D-trp and 0.5 mM L-trp. Conditions: capillaries 375 μm outside diameter × 75 μm inside diameter, 47 cm total length, 40 cm to the detection window; detection at 280 nm; mobile phase: 50 mM Tris amine buffer, 5 mM KCl, pH 10; 15 kV, forward polarity, 2 sec pressure injection, temperature 16°C. (From Kotia, R.B., Li, L., and McGown, L.B., Separation of nontarget compounds by DNA aptamers, Anal. Chem., 72, 827, 2000. With permission.)

A CEC separation of tryptophan enantiomers (Figure 4.17) was accomplished on the 20-base DNA sorbent made by attaching the aptamer via the linker in Potyrailo et al. [39]. Fluorescence batch experiments showed that the aptamer interacts more strongly with L-trp and D-tyr than with D-trp, thus explaining the elution order. It was also demonstrated that the potential has an important effect on the interaction since pressure-driven separation was not achieved. The effect of mobile phase organic content and capillary inner diameter was investigated for the 15-base aptamer stationary phase in the separation of PAHs. A baseline separation of naphthalene and benzo(ghi)-perylene was obtained when the mobile phase contained 30% in methanol. The decrease in the inner diameter

DK4478_book.fm Page 190 Thursday, September 14, 2006 1:35 PM

190

Advances in Chromatography, Volume 45

of the capillary improved the resolution of PAH separation, but decreased the analysis sensitivity. Rehder and McGown [21] took advantage of the stability of the G-quartet conformation and the ease of immobilization on silica material to create an aptamer stationary phase (see Table 4.2 for the sequence) capable of separating bovine βlactoglobulin variants A and B (LgA and LgB) by CEC. These whey proteins encountered in bovine milk differ by only 2 of their 162 amino acid residues. 5′thiol-modified oligonucleotide molecules were covalently attached to the inner wall of fused silica capillaries via sulfosuccinimidyl 4-(N-maleimidomethyl) cyclohexane-1-carboxylate using a method published in Phillips and Chmielinska [40]. Figures 4.18 and 4.19 show the CEC separations of the two lactoglobulin variants achieved on both the aptameric stationary phase and the stationary phase created by 0.001 0.0005 0 −0.0005 −0.001 (a)

Absorbance (AU)

0.0013 0.0008 0.0003 −0.0002 −0.0007 (b) 0.0013 0.0008 0.0003 −0.0002 −0.0007

0

5 15 10 Retention time (min)

20

(c)

FIGURE 4.18 CEC separations of LgA and LgB mixtures on a DNA aptameric stationary phase: (a) 50 μM LgA and 50 μM LgB, (b) 75 μM LgA and 25 μM LgB, and (c) 25 μM LgA and 75 μM LgB. Conditions: mobile phase 25 mM Tris, 1 mM KCl, pH 7.3; 75 μm inside diameter capillary, 47 cm long; 15 kV; 5 sec low-pressure sample injection; temperature 25°C. (From Rehder, M.A. and McGown, L.B., Open tubular capillary electrochromatography of bovine beta-lactoglobulin variants A and B using an aptamer stationary phase, Electrophoresis, 22, 3759, 2001. With permission.)

DK4478_book.fm Page 191 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

191

0.013 0.0125 0.012 0.0115

Absorbance (AU)

(a) 0.0122 0.0117 0.0112 0.0107 (b) 0.0117 0.0112 0.0107 0.0102 0

5

20 10 15 Retention time (min)

25

30

(c)

FIGURE 4.19 CEC electropherograms of LgA and LgB mixtures on a capillary coated with linker only. Concentrations of LgA and LgB and other conditions as in Figure 4.18. (From Rehder, M.A. and McGown, L.B., Open tubular capillary electrochromatography of bovine beta-lactoglobulin variants A and B using an aptamer stationary phase, Electrophoresis, 22, 3759, 2001. With permission.)

binding only the linker molecule to the silica surface. However, the separation times and the peak shapes are improved on the aptameric sorbent due to less interaction of the protein with the aptamer than with the organic linker, leading to a less denaturing separation. Charles and McGown [19] further demonstrated the ability of the G-quartet conformation of the 15-base thrombin aptamer [35] and the 23-mer aptamer sequence in Rehder and McGown [21] to separate Trp-Arg and Arg-Trp, two isomeric dipeptides. The examination of the nonspecific interaction of oligonucleotides with peptides is important in elucidating the mechanism of protein binding by Gquartet-forming molecules. The attachment of 5′-thiol-modified oligonucleotides to silica capillary inner walls was based on the same immobilization method from Phillips and Chmielinska [40] used by Rehder and McGown. Of the two dipeptides, Trp-Arg eluted first in all CEC experiments. This may be due to better access of the guanidinium group of arginine to the binding site of the aptamer, since it has been previously shown that this group interacts strongly with certain guanine-rich DNA molecules, destabilizing their conformation [41].

DK4478_book.fm Page 192 Thursday, September 14, 2006 1:35 PM

192

Advances in Chromatography, Volume 45

Absorbance (AU)

0.059 0.049 0.039

40°C 25°C

0.029 0.019 0.009 −0.001 (a)

Absorbance (AU)

0.0014

0.0009

40°C 25°C

0.0004 −0.0001

0

2

4 6 Retention time (min)

8

10

(b)

FIGURE 4.20 CEC separation of a Trp-Arg/Arg-Trp mixture on a 15-base aptamer sorbent. (a) Conditions: 0.5 mM each peptide; mobile phase: 25 mM Tris, 2 mM KCl, pH 7.2, 75 μm inside diameter capillary, temperature 25°C and 40°C, 10 kV separation voltage; (b) same conditions as (a) except for a 25 μm inside diameter capillary. (From Charles, J.A.M. and McGown, L.B., Separation of Trp-Arg and Arg-Trp using G-quartet-forming DNA oligonucleotides in open tubular capillary electrochromatography, Electrophoresis, 23, 1599, 2002. With permission.)

Figure 4.20 shows the CEC separations of the peptides on the 15-base aptamer sorbent in the best determined conditions of temperature and mobile phase composition. This stationary phase proved more efficient in resolving the Trp-Arg and ArgTrp elution peaks than the 23-mer aptamer sorbent. The authors hypothesize that partial destabilization of the G-quartet is responsible for this behavior, since the thrombin aptamer reached the melting point of 41°C in aqueous solution in the same conditions, while the second aptamer had a Tm greater than 70°C. Vo and McGown [22] also investigated stationary phases containing the thrombin aptamer that forms a two-plane quartet [35], the 23-base aptamer used in Rehder and McGown [21], with a four-plane quartet, and a different sequence named “G8T” resembling the thrombin aptamer, having the –G– in position 8 replaced by a –T– base. They performed OT-CEC separations of 14 homodipeptides and 10 alanyl dipeptides on aptameric sorbents prepared by the previously mentioned immobilization method [40] and compared the separation profile with the CEC electrochromatograms on a bare silica capillary and a scrambled oligonucleotide sorbent that lacks G-quartets. An example of CEC separations of 10 alanyl dipeptides on the thrombin aptamer stationary phases compared with the separation on a bare fused silica capillary and scrambled aptamer sorbent is depicted in Figure 4.21.

DK4478_book.fm Page 193 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

193

AH

0.002

AV 0.001

2-Plane EOF = N/A

AM

AA

AR

AP

AQ AW

AS

AE

0

Absorbance (AU)

10

15

20

0.0049

30

AS + AW

0.0039

AH

AM

35

Bare EOF = 7.6 min

AP + AV

0.0029

AQ AE

0.0019

AR

0.0009 −0.0001

25

6

AA 7

8

9

10

11

0.005

12

13

14

Scrambled EOF = N/A

0.004 0.003 0.002 0.001 −0.001

5

10

15 20 Time (min)

25

30

FIGURE 4.21 Open tubular capillary electrochromatography (OTCEC) separation of 10 alanyl dipeptides on a two-plane aptamer sorbent and control capillaries. Conditions: 75 μm inside diameter capillaries; mobile phase: 25 mM Tris-HCl, pH 5.0, 2 mM KCl; 10 kV separation voltage; temperature 25°C. (Reprinted partially from Vo, T.U. and McGown, L.B., Selectivity of quadruplex DNA stationary phases toward amino acids in homodipeptides and alanyl dipeptides, Electrophoresis, 25, 1230, 2004. With permission.)

The authors demonstrated that by varying the mobile phase composition, the pH, and the temperature, the aptameric sorbents were able to resolve certain peptides otherwise unresolved on bare silica columns by capillary zone electrophoresis, proving that the separations on the aptameric sorbents consist of both a chromatographic and an electrophoretic component. Differences between electrochromatograms on the G-quartet stationary phases and the scrambled nucleotide sorbent demonstrate that the presence of a G-quartet is essential in separating specific peptides.

V. CONCLUSION Aptamers have proven to be powerful affinity reagents in chromatographic separations and stand as viable alternatives to other affinity techniques, including antibodyantigen interactions and MIPs, that have limitations of their own. Moreover, aptamer

DK4478_book.fm Page 194 Thursday, September 14, 2006 1:35 PM

194

Advances in Chromatography, Volume 45

stationary phases show extended applicability, since they are also successful in separating compounds unrelated to the SELEX target compound. While much remains to be done to move aptamers to a position of easy adoption for chromatographic separations, their promise is such that a continued presence in the literature of separations is assured.

REFERENCES 1. Hermann, T. and Patel, D.J., Adaptive recognition by nucleic acid aptamers, Science, 287, 820, 2000. 2. Ellington, A.D. and Szostak, J.W., In vitro selection of RNA molecules that bind specific ligands, Nature, 346, 818, 1990. 3. Tuerk, C. and Gold, L., Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase, Science, 249, 505, 1990. 4. Bowser, M.T. and Mendonsa, S.D., Screening combinatorial libraries for functional RNA and DNA ligands of target molecules using electrophoretic selection CE-SELEX, Patent #WO 2003102212 (Dec. 11, 2003); US 2004/0018530 A1 (Jan. 29, 2004). 5. Mendonsa, S.D. and Bowser, M.T., In vitro selection of high-affinity DNA ligands for human IgE using capillary electrophoresis, Anal. Chem., 76, 5387, 2004. 6. Mosing, R.K., Mendonsa, S.D., and Bowser, M.T., Capillary electrophoresis-SELEX selection of aptamers with affinity for HIV-1 reverse transcriptase, Anal. Chem., 77, 6107, 2005. 7. Mendonsa, S.D. and Bowser, M.T., In vitro selection of aptamers with affinity for neuropeptide Y using capillary electrophoresis, J. Am. Chem. Soc., 127, 9382, 2005. 8. Drabovich, A., Berezovski, M., and Krylov, S.N., Selection of smart aptamers by equilibrium capillary electrophoresis of equilibrium mixtures (ECEEM), J. Am. Chem. Soc., 127, 11224, 2005. 9. Clark, S.L. and Remcho, V.T., Aptamers as analytical reagents, Electrophoresis, 23, 1335, 2002. 10. Tombelli, S., Minunni, M., and Mascini, M., Analytical applications of aptamers, Biosens. Bioelectron., 20, 2424, 2005. 11. Brumbt, A., Ravelet, C., Grosset, C., Ravel, A., Villet, A., and Peyrin, E., Chiral stationary phase based on a biostable L-RNA aptamer, Anal. Chem., 77, 1993, 2005. 12. Chung, W.-J., Kim, M.-S., Cho, S., Park, S.-S., Kim, J.-H., Kim, Y.-K., Kim, B.-G., and Lee, Y.-S., Microaffinity purification of proteins based on photolytic elution: toward an efficient microbead affinity chromatography on a chip, Electrophoresis, 26, 694, 2005. 13. Clark, S.L. and Remcho, V.T., Open tubular liquid chromatographic separations using an aptamer stationary phase, J. Sep. Sci., 26, 1451, 2003. 14. Deng, Q., German, I., Buchanan, D., and Kennedy, R.T., Retention and separation of adenosine and analogues by affinity chromatography with an aptamer stationary phase, Anal. Chem., 73, 5415, 2001. 15. Michaud, M., Jourdan, E., Villet, A., Ravel, A., Grosset, C., and Peyrin, E., A DNA aptamer as a new target-specific chiral selector for HPLC, J. Am. Chem. Soc., 125, 8672, 2003. 16. Michaud, M., Jourdan, E., Ravelet, C., Villet, A., Ravel, A., Grosset, C., and Peyrin, E., Immobilized DNA aptamers as target-specific chiral stationary phases for resolution of nucleoside and amino acid derivative enantiomers, Anal. Chem., 76, 1015, 2004.

DK4478_book.fm Page 195 Thursday, September 14, 2006 1:35 PM

Aptamers as Molecular Recognition Elements

195

17. Ravelet, C., Boulkedid, R., Ravel, A., Grosset, C., Villet, A., Fize, J., and Peyrin, E., A L-RNA aptamer chiral stationary phase for the resolution of target and related compounds, J. Chromatogr. A, 1076, 62, 2005. 18. Romig, T.S., Bell, C., and Drolet, D.W., Aptamer affinity chromatography: combinatorial chemistry applied to protein purification, J. Chromatogr. B, 731, 275, 1999. 19. Charles, J.A.M. and McGown, L.B., Separation of Trp-Arg and Arg-Trp using Gquartet-forming DNA oligonucleotides in open tubular capillary electrochromatography, Electrophoresis, 23, 1599, 2002. 20. Kotia, R.B., Li, L., and McGown, L.B., Separation of nontarget compounds by DNA aptamers, Anal. Chem., 72, 827, 2000. 21. Rehder, M.A. and McGown, L.B., Open tubular capillary electrochromatography of bovine beta-lactoglobulin variants A and B using an aptamer stationary phase, Electrophoresis, 22, 3759, 2001. 22. Vo, T.U. and McGown, L.B., Selectivity of quadruplex DNA stationary phases toward amino acids in homodipeptides and alanyl dipeptides, Electrophoresis, 25, 1230, 2004. 23. Hicke, B.J., Watson, S.R., Koenig, A., Lynott, C.K., Bargatze, R.F., Chang, Y.-F., Ringquist, S., Moon-McDermott, L., Jennings, S., Fitzwater, T., Han, H.-L., Varki, N., Albinana, I., Willis, M.C., Varki, A., and Parma, D., DNA aptamers block Lselectin function in vivo: inhibition of human lymphocyte trafficking in SCID mice, J. Clin. Invest., 98, 2688, 1996. 24. Kansas, G.S., Selectins and their ligands: current concepts and controversies, Blood, 88, 3259, 1996. 25. Williams, K.P., Liu, X.H., Schumacher, T.N., Lin, H.Y., Ausiello, D.A., Kim, P.S., and Bartel, D.P., Bioactive and nuclease-resistant L-DNA ligand of vasopressin, Proc. Natl. Acad. Sci. U.S.A., 94, 11285, 1997. 26. Huizenga, D.E. and Szostak, J.W., A DNA aptamer that binds adenosine and ATP, Biochemistry, 34, 656, 1995. 27. Vianini, E., Palumbo, M., and Gatto, B., In vitro selection of DNA aptamers that bind L-tyrosinamide, Biorg. Med. Chem., 9, 2543, 2001. 28. Famulok, M., Molecular recognition of amino acids by RNA-aptamers: an L-citrulline binding RNA motif and its evolution into an L-arginine binder, J. Am. Chem. Soc., 116, 1698, 1994. 29. Klussmann, S., Nolte, A., Bald, R., Erdmann, V.A., and Furste, J.P., Mirror-image RNA that binds D-adenosine, Nat. Biotechnol., 14, 1112, 1996. 30. Schumacher, T.N., Mayr, L.M., Minor, D.L., Jr., Milhollen, M.A., Burgess, M.W., and Kim, P.S., Identification of D-peptide ligands through mirror-image phage display, Science, 271, 1854, 1996. 31. Mannironi, C., Scerch, C., Fruscoloni, P., and Tocchini-Valentini, G., Molecular recognition of amino acids by RNA aptamers: the evolution into an L-tyrosine binder of a dopamine-binding RNA motif, RNA, 6, 520, 2000. 32. Leickt, L., Bergstrom, M., Zopf, D., and Ohlson, S., Bioaffinity chromatography in the 10 mM range of Kd, Anal. Biochem., 253, 135, 1997. 33. Burgstaller, P. and Famulok, M., Isolation of RNA aptamers for biological cofactors in in vitro selection, Angew. Chem. Int. Ed. Engl., 33, 1084, 1994. 34. Maskos, U. and Southern, E.M., Oligonucleotide hybridizations on glass supports: a novel linker for oligonucleotide synthesis and hybridization properties of oligonucleotides synthesised in situ, Nucleic Acids Res., 20, 1679, 1992. 35. Bock, L.C., Griffin, L.C., Latham, J.A., Vermaas, E.H., and Toole, J.J., Selection of single-stranded DNA molecules that bind and inhibit human thrombin, Nature (Lond.), 355, 564, 1992.

DK4478_book.fm Page 196 Thursday, September 14, 2006 1:35 PM

196

Advances in Chromatography, Volume 45

36. Macaya, R.F., Schultze, P., Smith, F.W., Roe, J.A., and Feigon, J., Thrombin-binding DNA aptamer forms a unimolecular quadruplex structure in solution, Proc. Natl. Acad. Sci. U.S.A., 90, 3745, 1993. 37. Guo, Z., Guilfoyle, R.A., Thiel, A.J., Wang, R., and Smith, L.M., Direct fluorescence analysis of genetic polymorphisms by hybridization with oligonucleotide arrays on glass supports, Nucleic Acids Res., 22, 5456, 1994. 38. O’Donnell, M.J., Tang, K., Koster, H., Smith, C.L., and Canton, C.R., High-density, covalent attachment of DNA to silicon wafers for analysis by MALDI-TOF mass spectrometry, Anal. Chem., 69, 2438, 1997. 39. Potyrailo, R.A., Conrad, R.C., Ellington, A.D., and Hieftje, G.M., Adapting selected nucleic acid ligands (aptamers) to biosensors, Anal. Chem., 70, 3419, 1998. 40. Phillips, T.M. and Chmielinska, J.J., Immunoaffinity capillary electrophoretic analysis of cyclosporin in tears, Biomed. Chromatogr., 8, 242, 1994. 41. Harada, K. and Frankel, D., Identification of two novel arginine binding DNAs, EMBO J., 14, 5798, 1995. 42. Clark, S.L. and Remcho, V.T., Electrochromatographic retention studies on a flavinbinding RNA aptamer sorbent, Anal. Chem., 75, 5692, 2003.

DK4478_book.fm Page 197 Thursday, September 14, 2006 1:35 PM

5

Advances in HighThroughput HighPerformance Liquid Chromatography and “Purification-Friendly” Combinatorial Library Design Strategies Sabine Schefzick

CONTENTS I. II. III. IV. V.

Introduction................................................................................................197 New High-Throughput Purification Strategies .......................................... 198 How to Increase Throughput.....................................................................199 Which HPLC Methods Are Used in High-Throughput Mode? ...............200 Recent Advances in High-Throughput HPLC...........................................201 A. UV-HPLC ..........................................................................................202 B. MS-HPLC..........................................................................................203 VI. Combinatorial Library Design with Purification Considerations .............205 A. Introduction to QSAR and QSRR ....................................................205 B. QSRR Approach for Combinatorial Library Design........................ 208 VII. Conclusion .................................................................................................210 References..............................................................................................................210

I. INTRODUCTION Pharmaceutical companies are under pressure to rapidly identify leads for new targets and to focus on better quality leads with a decreased likelihood for attrition. To meet these expectations, most large pharmaceutical companies apply combinatorial and parallel synthesis to rapidly identify, optimize, and develop new chemical entities. The key advantage of combinatorial chemistry is the ability to rapidly synthesize a 197

DK4478_book.fm Page 198 Thursday, September 14, 2006 1:35 PM

198

Advances in Chromatography, Volume 45

large collection of compounds. The advancements of combinatorial chemistry produced follow-up analysis methods, such as purification, to increase throughput to meet the demands of combinatorial libraries. In addition, the demand to identify better quality leads highlights the importance of purification. The identification of better leads depends on the requirement of generating reproducible screening data, which are dependent on the absolute purity of the combinatorial library compounds. Improving compound purity is an effective way to decrease the chance of obtaining false-positive and negative data. Impurities severely complicate the interpretation of structure-activity relationships. Therefore, high-throughput purification has become an important technology for combinatorial and medical chemistry in an effort to decrease ambiguity in the screening results, leading to high quality leads. Traditional purification techniques (e.g., highperformance liquid chromatography [HPLC]) are well understood and commonly applied by chemists in “open-access laboratories.” However, many of these techniques are not directly applicable to the purification of large numbers of compounds, where parallel processes are necessary to increase productivity. Nevertheless, because of its simplicity and the capability of separating a wide range of diverse compounds, reverse-phase HPLC has become a popular method for preparative separation. High-throughput purification methods depend on several variables. The ability to perform multiple parallel processes and the ease of automation are principal factors when considering different purification methods. Finally, the cost and the difficulty of solvent removal are deciding factors in selecting a high-throughput purification method. High-throughput reverse-phase HPLC has become a good universal purification method for combinatorial libraries because it is possible to perform multiple parallel processes in an automated fashion with relatively low costs. However, compounds are commonly lost because of insufficient resolution during the separation process. In general, the recovery rate can range from 90 to 60% of the samples. Better recovery rates are achieved when the chromatographic conditions are known prior to the purification step. The chromatographic conditions are identified by either analytical scale HPLC analysis or “universal” gradient conditions with peak detection devices for the particular sample collection. In either case, the method development involves significant effort to ensure the optimal conditions for each compound. An enhancement of these methods would be to estimate the chromatographic parameter, such as retention time, computationally and to identify which of the available methods is the best option for all compounds in the library. Moreover, these predictions could be performed on virtual compounds, thereby facilitating the plating of compounds for purification and analysis. In this review, I discuss combinatorial library design strategies to facilitate a more efficient purification process and elaborate on recent advances in highthroughput HPLC.

II. NEW HIGH-THROUGHPUT PURIFICATION STRATEGIES High-throughput purification methods can be divided into two different subcategories: 1) Analytical HPLC methods are applied to quantify or identify a compound.

DK4478_book.fm Page 199 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

199

These analyses are usually performed on a small scale compared to preparative HPLC. 2) Preparative HPLC’s goal is the isolation and purification of a specific compound within the sample. In contrast to analytical HPLC, the sample will be collected in a fraction detector. In the following paragraphs I will focus on preparative HPLC unless otherwise mentioned.

III. HOW TO INCREASE THROUGHPUT The development of combinatorial chemistry has challenged HPLC methods to increase throughput to meet growing demands. Many conventional techniques are not directly applicable to the purification of large numbers of compounds where parallel processes are necessary to increase productivity. High-performance liquid chromatography time efficiency can be improved by increasing the amount of automation during each step of the purification. The use of autosamplers and fraction collectors in conjunction with computer and controller systems that deal with the distribution of tasks and data collection has increased the speed of HPLC. Different software packages with graphical interfaces are used to support the chemist in setting up the purification run. Since computers are available at a reasonable cost, this automation is common in most analytical laboratories. It is therefore important to look at the separation step itself to increase throughput. One possible solution is to decrease the separation time by retaining the same resolution [1]. Equation 5.1 illustrates that the capacity factor (k′), that is, the resolution of the separation, is directly proportional to the gradient time (tg) and the flow rate (F), but indirectly proportional to the void volume (V0), the gradient range ΔΦ expressed as a fraction, and S, a constant. Therefore, the only option to decrease tg while keeping the capacity factor constant is to increase the flow rate and decrease the void volume. However, since both of the parameters increase the backpressure of the column, the separation time can only be decreased by using shorter columns (decrease in gradient time by increasing flow rate) with a lower void volume, which is achieved by decreasing the particle size of the stationary phase. k′ = tg ⋅

F 1 ⋅ Vo ΔΦ ⋅ S

(5.1)

To scale up a reaction from the analytical to the preparative level, one needs columns with a larger diameter. The mobile phase linear velocity (u) is proportional to the flow rate, but indirectly proportional to the column diameter (d) and the column porosity (ε0) (Equation 5.2). To keep the retention time and thus the velocity constant during scale-up, the flow rate must be adjusted according to the column diameter (Equation 5.3). As observed in Equation 5.4, the retention time depends on the length of the column, the flow rate, and the diameter of the column. To increase the throughput of the purification, the column length can be reduced for faster separation. Moreover, the flow rate can be increased when utilizing a shorter column, as the backpressure of the column decreases with shorter column length. The drawback is mass overload of the column. The shorter the column and the faster the flow rate,

DK4478_book.fm Page 200 Thursday, September 14, 2006 1:35 PM

200

Advances in Chromatography, Volume 45

the lower the sample load a column can accept before overloading happens. It is usual in preparative HPLC separation to except light overloading signs, however, overloading increases peak tailing and will eventually diminish the selectivity of the separation. u=

Fprep Fanal

rt prep rtanal

=

L prep Lanal

4F πε0 d 2

⎛ d prep ⎞ =⎜ ⎟ ⎝ d anal ⎠

(5.2)

2

(5.3)

2

⎛ d prep ⎞ ε prep F ∗ ∗ anal ∗ ⎜ Fprep ⎝ d anal ⎟⎠ ε anal

(5.4)

Greater throughput can also be achieved by using multiple purification systems in a parallel fashion. Parallelization can be accomplished by using either multiple complete HPLC systems that are controlled by one computer or by multiplexing several building blocks of system components, for example, the pump, autosampler, and detection system, and performing the separation on different HPLC columns. The later option has the advantages that the system is more compact and lower in cost. However, when only one pump is used in a parallel setting, it is nearly impossible to obtain the same flow rate for all HPLC columns, since the backpressure varies from column to column. One can therefore utilize a flow control unit, which produces a steady stream of eluent to a multichannel injection system. A flow control unit [2] ensures constant flow by metering the pressure decrease in small capillaries. After the separation, the stream is passed through detection cells of a multiplexed photodiode array (PDA) detector or a multiplex mass spectrometer [3].

IV. WHICH HPLC METHODS ARE USED IN HIGH-THROUGHPUT MODE? High-throughput purification methods depend on several variables. Obviously, sample throughput has become a high priority criterion. The ability to perform multiple parallel processes and the ability to automate the purification process are main factors in considering purification methods. Finally, cost and the possibility of solvent removal are deciding factors in selecting a high-throughput purification method [4]. The chromatographic processes can be defined as separation techniques involving mass transfer between stationary and mobile phases. HPLC utilizes a liquid mobile phase to separate the components of a mixture. These components (or analytes) are first dissolved in a solvent and then forced to flow through a chromatographic column under high pressure. In the column, the mixture is resolved into its components. The amount of resolution is important and depends upon the extent of interaction between the solute components and the stationary phase. The stationary

DK4478_book.fm Page 201 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

201

phase is defined as the immobile packing material in the column. The interaction of the solute with the mobile and stationary phases can be manipulated through different choices of both solvents and stationary phases. In reverse-phase chromatography, the mechanism is hydrophobic interaction between the alkyl and aryl groups on the analyte with the bonded phase. As a result, HPLC acquires a high degree of versatility not found in other chromatographic systems, and it has the ability to easily separate a wide variety of chemical mixtures. The chemical and physical properties of the stationary phase are one critical factor in the success of the separation. The most frequently used stationary phase in preparative high-throughput HPLC is columns packed with spherical C18. The silica packing is covalently bond to hydrophobic octadecylsilyl (C18) groups. Particle size is an important parameter for analytical and preparative HPLC. Generally, smaller particle sizes allow greater efficiency and permit the use of shorter columns to increase separation speed. In preparative chromatography, the particle size is important because the column is often used in an overloaded state. In general the particle size can range from 5 to 10 μm [5]. Since the pressure decrease is inversely proportional to the particle diameter squared, larger particles produce smaller pressure decreases, allowing higher flow rates, which in turn increases the throughput of preparative columns. However, larger particles negatively influence the purity and recovery rate of the separation. The amount of sample per injection depends on the column dimensions. Typical column dimensions seen in preparative HPLC are diameters ranging from 4 to 50 mm with a length of 5 to 25 cm. Nonlinear effects are observed when the column is overloaded, which causes the resolution to decrease. However, a study of the percentage of recovery on overloaded columns has show a nearly 100% recovery rate if the separation resolution is greater than one [6]. The stationary phase is one way to change the separation condition. Another critical factor is the mobile phase. The mobile phase can influence the efficiency, recovery rate, resolution, and purification cost. In general, the same mobile phase as was used for analytical HPLC can be linearly transferred to preparative HPLC conditions. The best mobile phase exhibits a low viscosity and a low boiling point. The mobile phase must be inert and has to be a good solvent for small molecules. For the sake of easy handling and storage, the mobile phase should be nontoxic. The most frequently used solvent mixture is a water/acetonitrile or water/methanol mobile phase. Up to 0.1% of TFA is used as a modifier for the mobile phase. However, it is known that trifluoroacetic acid (TFA) increases the ambiguity of the screening assay, is expensive, and will affect the baseline at low ultraviolet (UV) wavelengths. Moreover, it suppresses ion formation in electrospray mass spectroscopy (MS) and must be stored carefully. Therefore, other additives such as formic acid, acetic acid, and propanol are used instead.

V. RECENT ADVANCES IN HIGH-THROUGHPUT HPLC Preparative high-throughput HPLC has the goal of purifying the submitted samples to obtain the wanted compound with the highest purity and yield possible in the fastest time possible. To purify the samples, reverse-phase HPLC is coupled with a

DK4478_book.fm Page 202 Thursday, September 14, 2006 1:35 PM

202

Advances in Chromatography, Volume 45

fraction collection. The fraction collection can be triggered in multiple ways. One can imagine a manual mode, where the technician has to press a button (software or hardware) to start and stop collection of the fraction. Another mode is a timebased collection mode, where fractions are collected based on a fixed time interval. Equation 5.4 demonstrates that it is possible to calculate the retention time of the analyte under preparative conditions based on the retention time recorded in an analytical setting. However, such scale-up calculations are never completely accurate. The best method for fraction collection is based on the signal of a detector. UV-triggered or peak-based fraction collection are activated if the signal for the UV detector reaches a set threshold. UV detection fails if the samples are insoluble in water. Generally such compounds are eluted in dimethyl sulfoxide (DMSO), which can elute in the void volume and will pass uncollected. Moreover, UV-triggered fraction collection will fail if the sample’s chromophore is too weak to detect. UVtriggered detection is a commonly used technique where every fraction is collected. On the other hand, the mass-based fraction collection approach enables compoundspecific fraction collection. Here, fractions are only collected if the specified mass is present. MS-triggered fraction collection can fail if the sample does not ionize to a degree that exceeds the minimum threshold for MS detection. In the following sections, specific UV-HPLC and MS-HPLC systems will be reviewed.

A. UV-HPLC The first high-throughput reverse-phase HPLC was introduced by Weller et al. [7] at Bristol-Meyers Squibb (New York, NY). It is a UV threshold-triggered fraction collection system that is operated as an open access system. This rapid preparative and analytical system is capable of purifying up to 200 samples a day at weights of up to 200 mg. Similar systems, but without open access, have also been reported from Abbott Laboratories (Abbott Park, IL) [8]. They use systems that are connected to a UV or electron light-scattering detector (ELSD)-triggered fraction collector and another connected to a MS-triggered fraction collector [9]. MS-HPLC is also commercially known as the Agilent 1100 series [10] LC/MSD (Agilent Technologies, Palo Alto, CA), which includes a PDA detector and mass spectrometer. Parke-Davis (Morris Plains, NJ) researchers reported on a UV-triggered HPLC system for the purification of combinatorial libraries that can be operated in the reverse-phase or normal-phase HPLC mode [11]. A multichannel UV-HPLC system called Parallex that can purify up to four samples (up to 200 mg/sample) in parallel per cycle is available from Biotage (Uppsala, Sweden) [12,13]. This system consists of four independent HPLC columns and four independent fraction collectors, each one holding up to eight deep-well plates for fraction collection. The system allows for fraction collection simultaneously triggered by two different wavelengths. Edwards and Hunter [14] reported that the Parallex HPLC system is able to purify at least 200 samples generated via solid-phase synthesis in 10 hours. All of the above mentioned detection systems are based on a predefined UV threshold that triggers the fraction collector to start collecting once the signal intensity exceeds the threshold. This method generates an intense analog signal that cannot be used to distinguish product from impurities. Therefore, all substances that exceed

DK4478_book.fm Page 203 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

203

the UV threshold are collected and a follow-up analysis step, most likely flow injection MS, is necessary to determine the location of the product. In contrast, MStriggered fraction collection is specific to the product’s mass:charge ratio, which allows collection of the desired product fraction.

B. MS-HPLC The first MS-HPLC in a parallel operation was introduced in 1998 when a fast gradient separation was coupled with MS-triggered detection. The first method [15], called PrepLCMS, incorporates liquid chromatography and intelligent fraction collection using electrospray ionization mass spectroscopy (ESI-MS). The power of this system lies in its ability to distinguish between wanted product and unwanted impurities. The fraction collection is initiated on a real-time threshold reconstructed ion current signal being observed for a particular m/z input value. This system is capable of analyzing 100 compounds per day and is available for chemists in an “open-access”-like environment. In a further modification, the same group improved the throughput by creating a parallel analytical and preparative MS-HPLC workstation [16]. Rather than sequentially analyzing samples, Zeng and Kassel [16] introduced parallel HPLC by using one solvent pump and a splitter tee that transfers solvent to two HPLC columns for either analytical analyses or purification. After separation, the effluent is sprayed into a modified dual-atmospheric pressure ionization (API) ion source interface of a quadrupole mass spectrometer. Kiplinger et al. [17] reported a workstation that is capable of separating mixtures in the 10 to 20 mg range. The fraction collection is triggered either by UV or by MS. A novel multiplexed four-way electrospray interface for MS enabled liquid chromatography/MS to dramatically increase sample throughput in a truly parallel approach [3]. The new MUX technology (multiplexed interfaces) allows four HPLC columns to be multiplexed in parallel with one orthogonal electrospray time-of-flight (TOF) detector (Figure 5.1). The MUX ion source is comprised of separate electrospray needles around a rotating disc, allowing independent sampling from each individual sprayer by MS. A rotor within the ion source allows each electrospray channel to be monitored exclusively, eliminating cross talk between sample inlets (see Figure 5.1). In practice, each electrospray is sampled sequentially in very quick succession with a total duty cycle of less than 1 sec. Molecular weight-triggered collection from each of the four HPLCs is directed to each of four fraction collectors. Currently this technology can be expanded to a total of eight HPLC systems [18]. Utilizing a conventional approach for the analysis of a 96-well microtiter plate (MTP) would take approximately 4.5 h using a liquid chromatography gradient technique. The new MUX technology can decrease the analysis time by a factor of eight, to just over half an hour [19]. Eckers et al. [18] described the use of two separate electrosprays for introducing sample and calibrant for accurate mass measurement. The experiments, carried out using a MUX-MS, illustrate that the mass can be measured accurately within a few millidaltons of the theoretical value for both protonated and deprotonated molecules and fragment ions. Moreover, during their work, the researchers did not observe any interchannel cross talk. Another workstation used three switching valves and four

DK4478_book.fm Page 204 Thursday, September 14, 2006 1:35 PM

204

Advances in Chromatography, Volume 45

1

Ion block Sampling rotor Hexapole ion bridge

2

1

Electro??

OA-TOF Mass spec. Sampling rotor

2 Desolvation gas

Ions

3

Sampling cone

Liquid flow

Sampling cone

3 Gas 4

4

FIGURE 5.1 Schema from the MUX technology. (Courtesy of Waters Corporation, MUX technology, www.waters.co.jp/product/ms/lcms/mux_sys.html, 2005.)

HPLC columns to stagger injections onto the columns, allowing the mass spectrometer to continuously analyze the chromatographic window of interest [21]. Using this approach, the optimized run time is greater than the sum of the widths of the desired peaks. Lui et al. [22] used a multichannel device with an array of 96 electrospray tips for MS analysis with the ability to analyze 720 samples per hour. MUX technology utilizing a fast scanning quadrupole mass spectrometer was reported in 2002 [23]. To maximize sample recovery, the MUX purification application was modified so as to allow the user to establish the optimum fraction collector valve-switching delay time for each channel. This delay time was determined by measuring the reduction in postfraction collector UV signal at various valveswitching delays. Applying the optimal valve-triggering delay time, the application was capable of purifying sample sizes from 1 to 10 mg. The scientists achieved recovery values greater than 80% and 90% purity of materials after optimizing the fraction collector’s valve-switching timing. However, it has been shown that doubling the number of purification channels does not double the throughput in the purification step because of other factors such as sample tracking, data processing and archiving, postpurification quality control, etc. Realistically one can increase the throughput by approximately 30% [24]. Recently Isbell et al. [25] outlined key improvements in automation, solvent handling, and sample handling to sustain a mean throughput with a parallel, fourchannel HPLC/MUX/MS purification system of 528 samples/day over a multimonth time period. Similar date handling and automation issues were described by Koppitz et al. [26]. They state that the developed workstation can be supervised by one person to process up to 200 compounds on a 150 mg scale. In summary, the MUX technology has been applied successfully to highthroughput analysis of combinatorial libraries [19–21,23–25] and bioanalytical samples [22]. Currently molecular weight-triggered fraction collection HPLC systems are commercially available from Agilent (Palo Alto, CA) [27], Waters/Micromass (Milford, MA) [20], and Thermo Electron Corp. (Waltham, MA) [28].

DK4478_book.fm Page 205 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

205

High-throughput HPLC methods have made dramatic steps toward increasing throughput while delivering pure samples with sufficient yield. However, prepurification analysis steps and chromatographic method development steps are a bottleneck for preparative methods. If the number of samples is high, the purification method cannot be optimized for each compound and one needs to rely on generic chromatographic methods. However, one generic method might not be sufficient for a 96-well MTP and will result in loss of product. To counteract the loss of sample, Schefzick et al. [29] proposed that the purification conditions be considered during the combinatorial library design process.

VI. COMBINATORIAL LIBRARY DESIGN WITH PURIFICATION CONSIDERATIONS For library purification, HPLC in particular has received the most attention because of its adaptability to automation and widespread experience with this instrumentation for separation of a variety of structural types. If HPLC is applied as a high-throughput purification method, no prior knowledge of the chromatographic properties of the sample exists. The chromatograph conditions are either identified by analytical-scale HPLC analysis or by identifying “universal” gradient conditions with peak detection devices for the particular sample collection. This involves significant effort in method development to ensure the optimal conditions for each compound. Since extensive method development defies the purpose of high-throughput analysis, a limited number of generic methods are available, and the “best” method is often selected based on the analyst’s knowledge and experience. A reasonable change in the work flow would be to estimate which of the available methods would give the best result for each compound in the library. Moreover, this prediction could be performed on virtual compounds, thereby guiding the plating of compounds for purification and analysis. In a recent study, researchers applied quantitative structure activity relationship (QSAR) methods to predict the chromatographic properties (e.g., retention time) prior to the purification and analysis step [29]. Moreover, the quantitative structure retention relationship (QSRR) models can be used to design “analysis-friendly” library purification plates, where compounds are arranged based on their predicted separation condition. Furthermore, enumerated products not amenable to the separation methods can be identified during the library design phase.

A. INTRODUCTION

TO

QSAR

AND

QSRR

The quantitative structure property relationship (QSPR) (or QSAR) approach was developed by Hansch et al. [30] in 1962 and has been successfully applied to correlate chemical structure to a wide range of physiochemical, biological, toxicological, ecotoxicological, and technological properties as an alternative to more expensive prediction methods. QSAR represents an attempt to correlate structural or property descriptors of compounds with activities. These physicochemical descriptors, which include parameters to account for hydrophobicity, topology,

DK4478_book.fm Page 206 Thursday, September 14, 2006 1:35 PM

206

Advances in Chromatography, Volume 45

14000

Number of publications

12000 10000 8000 6000 4000 2000 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 Publication year

FIGURE 5.2 Illustration of the rapid growth of publication in the area of QSAR.

electronic properties, and steric effects, are determined empirically or more commonly by computational methods. Activities used in QSAR include chemical and physicochemical measurements, biological assays, and absorption, distribution, metabolization, and excretion/toxicity (ADME/Tox) data. QSAR is currently applied in many disciplines, many of which pertain to drug design and environmental risk assessment. The underlying concept is that a close relationship between bulk properties of compounds and the molecular structure of those compounds exists. This idea allows one to provide a clear connection between the macroscopic and the microscopic properties of matter. Therefore, it is an attempt to identify these assumed relationships between molecular structure and physical chemistry properties and then to quantify them. Once a satisfactory correlation between structure and property is found, any number of compounds, including those not yet synthesized, can be readily screened on the computer in order to select chemical structures with the desired properties. It is then possible to select the most promising compounds to synthesize and test in the laboratory. Accordingly, the QSAR approach conserves resources and accelerates the process of development of new molecules for use as drugs, materials, or additives or for any other purpose. The recent exponential growth in the number of papers dealing with QSAR studies clearly demonstrates the rapid progress in this area (Figure 5.2). Several approaches have been published to predict the solute retention behavior on a selected column. Every effort is based on the linear solvation energy relationship (LSER) [31–33], also known as the solvatochromic equation, which relates reactivity parameters with solvent-solute interactions based on physicochemical properties. log k ′ = c + rR2 + s π 2H + a

∑α

H 2

+b

∑β

2

+ vVx

(5.5)

DK4478_book.fm Page 207 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

207

Equation 5.5 defines the relationship between the capacity factor k′ and solute descriptors, where R2 is excess molar refraction, π2H is the solute polarizability/dipolarity, Σα2H and Σβ2 are the solute hydrogen-bond acidity and basicity, and Vx is the solute volume. The constants c, r, s, a, b, and v are specific for the system condition employed. This approach describes the contributions of individual intermolecular interactions that are responsible for the partition behavior of neutral molecules in octanol-water or reverse-phase separation systems. The solute properties are empirical descriptors that are only available for about 4000 compounds. Hence, this approach is limited to several thousand compounds and therefore not feasible for our study. Kaliszan [34–39], Kaliszan et al. [40,41], and Baczek and Kaliszan [42–44] demonstrated the prediction of solute retention under a given set of linear, reversephase gradient HPLC conditions by generating a QSRR model. Specifically, they have shown that the retention time of a solute under gradient reverse-phase HPLC can be determined from the following equation: tR = k1 + k2μ + k3δmin + k4Awas,

(5.6)

where μ is the total dipole moment, δmin is the electron excess charge of the most negatively charged atom, and Awas is the water-accessible molecular surface area of the solute. The constants k1, k2, k3, and k4 are related to specific stationary phase and mobile phase gradients employed in the separation. Another simple quantitative structure retention time model correlates the retention behavior with the logarithm of the n-octanol/water partition coefficient log P, which can be calculated with a variety of computer programs, retention parameter = k1 + k2log P.

(5.7)

Baczek and Kaliszan [44] presented a study that compared the latter two approaches for predicting gradient retention. In this study, Equation 5.6 and 5.7 revealed statistically significant QSRR models. However, they also showed that the predictive power of these models is limited. Thus, they conclude, “a suitable translation which would reveal the properties encoded into the structure in a reliable manner is still lacking.” A more recent approach is to classify the compounds and build a decision tree to predict the retention time under isocratic conditions [45]. Here the response (retention factor log kw) is used to build a decision tree based on 266 molecular descriptors (0D, 1D, and 2D). Despite the fact that a statistically relevant model with good predictive power is achieved, the author feels that a “more diverse set of substances with a more diverse retention time” might be needed to predict the chromatographic behavior. Another QSRR classification model was derived by Luan et al. [46] using semiempirical, two-dimensional and three-dimensional descriptors. Before generating the support vector machine (SVM) models, a heuristic method in CODESSA was applied to identify the five most descriptive variables out of 445 calculated descriptors. The overall mean square error in prediction for the SVM was 0.844. Recently Baczek et al. [47] compared clogP-based QSRR models with descriptor-based models and hydrophobic-subtraction models [48]. The descriptors for

DK4478_book.fm Page 208 Thursday, September 14, 2006 1:35 PM

208

Advances in Chromatography, Volume 45

hydrophobic-subtraction models are column-specific parameters derived by subtracting the hydrophobic contribution from the LSER equation (Equation 5.5). All three models were compared using principal components for nine reverse-phase HPLC columns and provide similar classification criteria. The authors concluded that all models resulted in similar classification of HPLC columns; however, they stated that QSRR-based approaches are simpler and require less labor. There are several commercially available software programs that can optimize the HPLC separation conditions or predict HPLC retention times [49–55]. However, none of these programs was developed for rapid analysis of combinatorial libraries. Therefore, the predicted retention times are typically longer than 5 min. Schefzick et al. [29] employed a QSRR approach to generate statistical models that are used to predict a set of reverse-phase gradient HPLC conditions best suited for the characterization of combinatorial compounds. A database containing all 75 models is used to determine the predicted retention time of every compound (using the solute’s QSRR descriptors) in the library under each of the chromatographic conditions. The calculated retentions of the solute under the analytical chromatographic conditions are assigned to one of three bins: tR < 1.5, not retained; 1.5 < tR < 4.5, moderately retained; tR > 4.5, highly retained. A chromatographic method that yields moderate retention for a solute receives a score of one for that solute. Otherwise, the chromatographic method receives a score of zero for that solute. The chromatographic method with the highest score for all solutes in the virtual library is selected as the recommended analytical method for the library. A similar approach is applied to recommend a preparative chromatographic method for library purification. In this case, analytical HPLC columns with the same stationary phase and lengths as the preparative HPLC columns are used to obtain the retention time information. The retention time of the solute under preparative HPLC conditions is proportional to the retention time of the solute under analytical HPLC conditions and the flow rates of the mobile phase, and indirectly proportional to the square of the column diameters (Equation 5.4).

B. QSRR APPROACH

FOR

COMBINATORIAL LIBRARY DESIGN

The QSRR approach was applied to generate statistical models to predict a set of reverse-phase gradient HPLC conditions best suited for the characterization of combinatorial compounds. A database containing all 75 models is used to determine the predicted retention time of every compound (using the solute’s QSRR descriptors) in the library under each of the chromatographic conditions. The retention time was measured for 62 compounds under five different gradient conditions on 15 different columns. Therefore, it was necessary to generate 75 QSAR models that are able to predict the retention time of new samples under these 75 conditions. To identify the best descriptor set possible, nearly 2500 descriptors were calculated for the dataset of 62 samples. After the removal of zero variance and correlated variables, several genetic algorithms where utilized to identify the most descriptive properties. Many of the identified descriptors represent molecular parameters that are known to influence the separation in reverse-phase HPLC (see Table

DK4478_book.fm Page 209 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

209

TABLE 5.1 List of the 20 Most Frequently Selected Descriptors by Genetic Algorithm Descriptor #amine CLogP nNHRPh vsa_other Atype_N_68 DCASA C-027 Donors DASA

QplogKp FASA+ PDsol (mcg/ml) Group_count_for_chain_c=n nCOOHPh H8m HATS3u estate_sCH3 Group_count_for_Phenyl CLogP_error SlogP_VSA9

Description Number of nonconjugated amines (QikProp) Biobyte’s log P (Sybyl6.9) Number of secondary amines (aromatic) (Dragon) Approximation to the sum of VDW surface areas of atoms typed as “other” (MOE) AlogP N in:Al3N (Cerius2) Absolute value of the difference between CASA+ (positive-charge weighted surface area, ASA+ times max{qi > 0 }) and CASA (MOE) Response to R–CH–X, Ghose-Crippen atom centered fragment (Dragon) HIVol Donor (Sybyl6.9) Absolute value of the difference between ASA+ (water accessible surface area of all atoms with positive partial charge [strictly greater than zero]) and ASA– (MOE) Predicted skin permeability (QikProp) Fractional ASA+ calculated as ASA+/ASA (MOE) Aqueous solubility (webPK) Group_count_for_chain_c=n (TSAR) Number of carboxylic acids (aromatic) (Dragon) H autocorrelation of lag 8/weighted by atomic masses GETAWAY (dragon) Leverage-weighted autocorrelation of lag3/unweighted GETAWAY (Dragon) Estate for CH3 group (Sybyl6.9) Group_count_for_Phenyl (TSAR) Biobyte’s log P (Sybyl6.9) Sum of vi such that Li > 0.40 (MOE)

5.1). For example, it is well known that the ionization state (neutral or charged) of a compound will affect the retention behavior of the solute. Moreover, under the HPLC conditions used in this study, amines will be charged, whereas the carboxylic group will be neutral. Therefore, it makes sense that the number of amines (#amine) and aromatic amines (nNHRPh) impact the prediction behavior more than the number of carboxylic acids (nCOOHPh). Likewise, it is not surprising that ClogP was selected as a significant variable, since the partitioning of a compound between liquid aqueous and organic phases is related to the solute’s partition equilibrium between the mobile and stationary phase. In particular, the nitrogen in aliphatic substructures is of importance, as the AlogP of nitrogen (:A3N) is ranked as fourth most important descriptor identified by the genetic algorithms [56] (Table 5.1). In addition, we observed a relevant contribution of different water accessible surface descriptors (DCASA, DASA, FASA+), which is in agreement with the results of Baczek and Kaliszan [44], who found that a water accessible molecular surface area descriptor, calculated in Hyper-Cube, also showed a significant contribution in their QSRR studies.

DK4478_book.fm Page 210 Thursday, September 14, 2006 1:35 PM

210

Advances in Chromatography, Volume 45

After identifying a suitable set of variables, stepwise multiple linear regression was chosen to generate QSRR equations for all HPLC conditions tested. Stepwise multiple linear regression produces a multiterm linear equation; however, not all independent variables are used. Step-by-step variables are added to the equation and a new regression is performed. If the new variable contributes significantly to the regression equation, the variable is retained, otherwise the variable is excluded, thus preventing overfitting. The model with the best R2 (best predictably) found for each dataset was used to identify the best predictive QSRR model for a specific HPLC condition. Overall it was possible to generate 71 of 75 statistically significant and predictive QSRR models, with an average r2 = 0.76 and q2 = 0.62. The average predictability of all 75 QSRR models is quite good, with R2 = 0.86. The average number of descriptors for all models was 7.4. Currently, scientists utilize these models in designing combinatorial libraries as well as selecting analytical methods for purification and characterization of singleton compounds in an open access analytical lab, where expert analysis may not be readily available.

VII. CONCLUSION Parallel HPLC approaches in analytical and preparative high-throughput HPLC are promising approaches to increase throughput at desired levels of purity and recovery rates. This chapter reviewed the recent advances in UV-triggered and MS-triggered HPLC. Moreover, it highlighted the importance of analytical considerations during the library design process. Unfortunately, attrition during the purification process is very hard to control. A recent approach to predict the retention time of a given sample for several chromatographic conditions could be a beginning. Based on the predicted retention time, the methods appropriate for the sample is chosen, followed by the next appropriate one, and so on. The compounds are arranged on the microtiter plates according to the chromatographic method to increase throughput.

REFERENCES 1. Wehr, T., Fast LC for high-throughput LC-MS, LCGC N. Am., 20(1): 40, 2002. 2. Müeller-Kuhrt, L., God, R., Gumm, H., and Binkele, J., Device and method for the parallel separation of substances by liquid chromatography, U.S. patent no. 6,911,151 B1, 2000. 3. de Biasi, V., Haskins, N., Organ, A., Bateman, R., Giles, K., and Jarvis, S., High throughput liquid chromatography/mass spectrometric analyses using a novel multiplexed electrospray interface, Rapid Commun. Mass Spectrom., 13, 1165, 1999. 4. Zhao, J., Zhang, L., and Yan, B., Strategies and methods for purifying organic compounds and combinatorial libraries, in Analysis and Purification Methods in Combinatorial Chemistry, Yan, B., ed., Wiley, New York, 2004, p. 253. 5. Majors, R.E., The role of the column in preparative HPLC, LCGC Eur., 17, 512, 2004. 6. Zhu, J. and Coscolluella, C., Chromatographic assay of pharmaceutical compounds under column overloading, J. Chromatogr. B Biomed. Sci. Appl., 741, 55, 2000.

DK4478_book.fm Page 211 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

211

7. Weller, H.N., Young, M.G., Michalczyk, S.J., Reitnauer, G.H., Cooley, R.S., Rahn, P.C., Loyd, D.J., Fiore, D., and Fischman, S.J., High throughput analysis and purification in support of automated parallel synthesis, Mol. Divers., 3, 61, 1997. 8. Routburg, M., Swenson, R., Schmitt, B., Washington, A., Mueller, S., Hochlowski, J., Maslana, G., Minin, B., Matuszak, K., Searle, P., and Pan, J., Implementation of an automated purification/verification system, in International Symposium on Laboratory Automation and Robotics, Zymark Corporation, Hopkington, MA, 1996. 9. Hochlowski, J., High-throughput purification: triage and optimization, in Analysis and Purification Methods in Combinatorial Chemistry, Yan, B., ed., Wiley, New York, 2004, p. 281. 10. Agilent Technologies, Agilent 1100 series, http://www.chem.agilent.com/Scripts/ PDS.asp?lPage=89, 2005. 11. Kibbey, C.E., An automated system for the purification of combinatorial libraries by preparative LC/MS, Lab. Robot. Automat., 9, 309, 1997. 12. Coffey, P., Ramieri, J., Garr, C., and Schultz, L., An open, multivendor specification for high-throughput organic chemistry (HTOC): a flexible production process for pure, characterized, quantified organic compounds, Am. Lab., 31, 57, 1999. 13. Schultz, L., Garr, C.D., Cameron, L.M., and Bukowski, J., High throughput purification of combinatorial libraries, Bioorg. Med. Chem. Lett., 8, 2409, 1998. 14. Edwards, C. and Hunter, D.J., High-throughput purification of combinatorial arrays, J. Comb. Chem., 5, 61, 2003. 15. Zeng, L., Burton, L., Yung, K., Shushan, B., and Kassel, D.B., Automated analytical/preparative high-performance liquid chromatography-mass spectrometry system for the rapid characterization and purification of compound libraries, J. Chromatogr. A, 794, 3, 1998. 16. Zeng, L. and Kassel, D.B., Developments of a fully automated parallel HPLC/mass spectrometry system for the analytical characterization and preparative purification of combinatorial libraries, Anal. Chem., 70, 4380, 1998. 17. Kiplinger, J.P., Cole, R.O., Robinson, S., Roskamp, E.J., Ware, R.S., O’Connell, H.J., Brailsford, A., and Batt, J., Structure-controlled automated purification of parallel synthesis products in drug discovery, Rapid Commun. Mass Spectrom., 12, 658, 1998. 18. Eckers, C., Wolff, J.-C., Haskins, N.J., Sage, A.B., Giles, K., and Bateman, R., Accurate mass liquid chromatography/mass spectrometry on orthogonal acceleration time-of-flight mass analyzers using switching between separate sample and reference sprays. 1. Proof of concept, Anal. Chem., 72, 3683, 2000. 19. Spickermann, J., Between automatized synthesis and high-throughput screening, LaborPraxis, 24, 45, 2000. 20. Waters Corporation, MUX technology, www.waters.co.jp/product/ms/lcms/mux_ sys.html, 2005. 21. Van Pelt, C.K., Corso, T.N., Schultz, G.A., Lowes, S., and Henion, J., A four-column parallel chromatography system for isocratic or gradient LC/MS analyses, Anal. Chem., 73, 582, 2001. 22. Liu, H., Felten, C., Xue, Q., Zhang, B., Jedrzejewski, P., Karger, B.L., and Foret, F., Development of multichannel devices with an array of electrospray tips for highthroughput mass spectrometry, Anal. Chem., 72, 3303, 2000. 23. Xu, R., Wang, T., Isbell, J., Cai, Z., Sykes, C., Brailsford, A., and Kassel, D.B., Highthroughput mass-directed parallel purification incorporating a multiplexed single quadrupole mass spectrometer, Anal. Chem., 74, 3055, 2002.

DK4478_book.fm Page 212 Thursday, September 14, 2006 1:35 PM

212

Advances in Chromatography, Volume 45

24. Isbell, J.J., Xu, R., Cai, Z., and Kassel, D.B., Realities of high-throughput liquid chromatography/mass spectrometry purification of large combinatorial libraries: a report on overall sample throughput using parallel purification, J. Comb. Chem., 4, 600, 2002. 25. Isbell, J.J., Zhou, Y., Guintu, C., Rynd, M., Jiang, S., Petrov, D., Micklash, K., Mainquist, J., Ek, J., Chang, J., Weselak, M., Backes, B.J., Brailsford, A., and Shave, D., Purifying the masses: integrating prepurification quality control, high-throughput LC/MS purification, and compound plating to feed high-throughput screening, J. Comb. Chem., 7, 210, 2005. 26. Koppitz, M., Brailsford, A., and Wenz, M., Maximizing automation in LC/MS highthroughput analysis and purification, J. Comb. Chem., 7, 714, 2005. 27. Agilent Technologies, 1200 Series LC/MSD, http://www.chem.agilent.com, 2005. 28. Thermo Electron Corporation, http://www.thermo.com/com/cda/category/category_ lp/1,2152,64,00.html, 2005. 29. Schefzick, S., Kibbey, C., and Bradley, M.P., Prediction of HPLC conditions using QSPR techniques: an effective tool to improve combinatorial library design, J. Comb. Chem., 6, 916, 2004. 30. Hansch, C., Maloney, P.P., Fujita, T., and Muir, R.M., Correlation of biological activity of phenoxyacetic acids with Hammett substituent constants and partition coefficients, Nature (Lond.), 194, 178, 1962. 31. Abraham, M.H., Rosés, M., Poole, C.F., and Poole, S.K., Hydrogen bonding. 42. Characterization of reversed-phase high performance liquid chromatographic C18 stationary phases, J. Phys. Org. Chem., 10, 358, 1997. 32. Platts, J.A., Abraham, M.H., Butina, D., and Hersey, A., Estimation of molecular linear free energy relationship descriptors by a group contribution approach, J. Chem. Inform. Comp. Sci., 39, 835, 1999. 33. Platts, J.A., Abraham, M.H., Butina, D., and Hersey, A., Estimation of molecular linear free energy relationship descriptors by a group contribution approach. 2. Prediction of partition coefficients, J. Chem. Inform. Comp. Sci., 40, 71, 2000. 34. Kaliszan, R., Quantitative Structure-Chromatographic Retention Relationships, John Wiley & Sons, New York, 1987. 35. Kaliszan, R., Computer aids to chemistry, in Trends in Analytical Chemistry, Vernin, G. and Chanon, M., eds., Chichester, Ellis Horwood, 1986. 36. Kaliszan, R., Multivariate chemometrics in QSAR (quantitative structure-activity relationships): a dialogue, in Chemometrics and Intelligent Laboratory Systems, Mager, P.P., ed., Wiley, New York, 1988. 37. Kaliszan, R., Quantitative structure-retention relationships applied to reversed-phase high-performance liquid chromatography, J. Chromatogr. A, 656, 417, 1993. 38. Kaliszan, R., Retention data from affinity high-performance liquid chromatography in view of chemometrics, J. Chromatogr. B Biomed. Sci. Appl., 715, 229, 1998. 39. Kaliszan, R., Chromatography and capillary electrophoresis in modelling the basic processes of drug action, Trends Anal. Chem., 18, 400, 1999. 40. Kaliszan, R., Haber, P., Baczek, T., Siluk, D., and Valko, K., Lipophilicity and pKa estimates from gradient high-performance liquid chromatography, J. Chromatogr. A, 965, 117, 2002. 41. Kaliszan, R., van Straten, M.A., Markuszewski, M., Cramers, C.A., and Claessens, H.A., Molecular mechanism of retention in reversed-phase high-performance liquid chromatography and classification of modern stationary phases by using quantitative structure-retention relationships, J. Chromatogr. A, 855, 455, 1999.

DK4478_book.fm Page 213 Thursday, September 14, 2006 1:35 PM

High-Throughput HPLC and Combinatorial Library Design Strategies

213

42. Baczek, T. and Kaliszan, R., Quantitative structure/retention relationships in affinity chromatography, J. Biochem. Biophys. Meth., 49, 83, 2001. 43. Baczek, T. and Kaliszan, R., Combination of linear solvent strength model and quantitative structure-retention relationships as a comprehensive procedure of approximate prediction of retention in gradient liquid chromatography, J. Chromatogr. A, 962, 41, 2002. 44. Baczek, T. and Kaliszan, R., Predictive approaches to gradient retention based on analyte structural descriptors from calculation chemistry, J. Chromatogr. A, 987, 29, 2003. 45. Put, R., Perrin, C., Questier, F., Coomans, D., Massart, D.L., Vander Heyden, Y., Classification and regression tree analysis for molecular descriptor selection and retention prediction in chromatographic quantitative structure-retention relationship studies, J. Chromatogr. A, 988, 261, 2003. 46. Luan, F., Xue, C., Zhang, R., Zhao, C., Liu, M., Hu, Z., and Fan, B., Prediction of retention time of a variety of volatile organic compounds based on the heuristic method and support vector machine, Anal. Chim. Acta, 537, 101, 2005. 47. Baczek, T., Kaliszan, R., Novotna, K., and Jandera, P., Comparative characteristics of HPLC columns based on quantitative structure-retention relationships (QSRR) and hydrophobic-subtraction model, J. Chromatogr. A, 1075, 109, 2005. 48. Snyder, L.R., Dolan, J.W., and Carr, P.W., The hydrophobic-subtraction model of reversed-phase column selectivity, J. Chromatogr. A, 1060, 77, 2004. 49. Galushko, S.V., Software for method development in HPLC, GIT Spez. Chromatogr., 2, 88, 1996. 50. Snyder, L.R. and Quarry, M.A., Computer simulation in HPLC method development: reducing the error of predicted retention times, J. Liq. Chromatogr., 10, 1789, 1987. 51. Dolan, J.W., Snyder, L.R., Djordjevic, N.M., Hill, D.W., and Waeghe, T.J., Reversedphase liquid chromatographic separation of complex samples by optimizing temperature and gradient time: I. Peak capacity limitations, J. Chromatogr. A, 857, 1, 1999. 52. Dolan, J.W., Snyder, L.R., Djordjevic, N.M., Hill, D.W., and Waeghe, T.J., Reversedphase liquid chromatographic separation of complex samples by optimizing temperature and gradient time: II. Two-run assay procedures, J. Chromatogr. A, 857, 21, 1999. 53. Dolan, J.W., Snyder, L.R., Wolcott, R.G., Haber, P., Baczek, T., Kaliszan, R., and Sander, L.C., Reversed-phase liquid chromatographic separation of complex samples by optimizing temperature and gradient time: III. Improving the accuracy of computer simulation, J. Chromatogr. A, 857, 41, 1999. 54. Williams, A., Kolovanov, E., and Hofmann, H., Chromatographic data system with integrated molecular structure management, GIT Labor Fachzeit., 44, 154, 2000. 55. Delaurent, C., Brenier-Maurel, C., Galushko, S.V., Tanchuk, V., Shishkina, I., et al. Chromatography and chemometry: a software for computer assisted method in high performance liquid chromatography, Spectra Anal., 31, 34, 2002. 56. Ghose, A.K., Viswanadhan, V.N., and Wendoloski, J.J., Prediction of hydrophobic (lipophilic) properties of small organic molecules using fragmental methods: an analysis of ALOGP and CLOGP methods, J. Phys. Chem. A, 102, 3762, 1998.

DK4478_book.fm Page 214 Thursday, September 14, 2006 1:35 PM

DK4478_book.fm Page 215 Thursday, September 14, 2006 1:35 PM

6

Biological, Clinical, and Forensic Analysis Using Comprehensive Two-Dimensional Gas Chromatography Nicholas H. Snow

CONTENTS I. Introduction................................................................................................216 II. History........................................................................................................216 A. Overview of Multidimensional Separations .....................................217 B. Classical Heart-Cutting in GC ..........................................................217 C. Development of Comprehensive GCxGC.........................................219 D. Overview of Complex Separations in Clinical, Biological, and Forensic Analysis ....................................................................... 219 III. GCxGC Fundamentals and Instrumentation .............................................220 A. Modulation Techniques .....................................................................220 B. Injection and Sampling .....................................................................221 C. Column Selection ..............................................................................224 D. Detectors............................................................................................225 E. Data Handling ...................................................................................227 1. Peak Integration and Identification ............................................. 229 2. Target Compound Analysis .........................................................230 3. Fingerprinting and Group Analysis ............................................231 IV. Additional Applications .............................................................................231 A. Biological ..........................................................................................231 B. Clinical and Drug Testing .................................................................232 C. Forensic .............................................................................................236 D. Pharmaceutical ..................................................................................236 E. Natural Products................................................................................237 V. Conclusion ................................................................................................. 238 Acknowledgments..................................................................................................238 References..............................................................................................................238 215

DK4478_book.fm Page 216 Thursday, September 14, 2006 1:35 PM

216

Advances in Chromatography, Volume 45

I. INTRODUCTION This chapter presents current research and discusses the potential for comprehensive two-dimensional gas chromatography (GCxGC) to have an impact on the analysis of samples with biological origins. GCxGC has seen a tremendous upswing in interest over the past 5 years and developments are happening quickly. While most GCxGC applications have been for petroleum and related mixtures, interest in other areas is advancing as the technique becomes more developed. Early challenges, such as reliable sample modulation between the two columns and handling of the large amounts of generated data, have generally been overcome, but work remains before GCxGC will become a mainstream technique. The fundamental ideas and instrumentation behind GCxGC and recent applications are presented through the lens of the needs of clinical, biological, pharmaceutical, and forensic analysts, with an eye toward exploring GCxGC as an effective tool in those fields.

II. HISTORY As the demand for high-resolution separation techniques in the analysis of complex mixtures has increased, so has interest in multidimensional separations. In traditional column chromatography, separations occur as the analytes are transported by a mobile phase through a column that contains a stationary phase and are detected as they elute from the column. Minute differences in intermolecular interactions between the different analytes and the stationary phase material drive the separation process. In gas, liquid, and supercritical fluid chromatography, there are myriad column configurations, dimensions, and chemistries available for solving complex separation problems. Multidimensional chromatography usually involves the serial combination of two or more separation chemistries in an effort to generate additional peak capacity or selectivity. Comprehensive two-dimensional gas chromatography has received increasing attention over the past decade, with more than 250 references resulting from a search on the keyword “comprehensive two-dimensional gas chromatography” in SciFinder Scholar [1]. As commercial instrumentation has become more readily available, the publication numbers have increased dramatically from fewer than 10 per year in the late 1990s to 73 in 2004, and more than 50 in 2005, for a total of 276 publications. Finally, nearly all of the references cited in this article and indeed the bulk of references are from journals dedicated to separation science. GCxGC has reached the status where it is of great interest to chromatographers, but it has not yet reached the general use that would see articles published in journals of other fields. This article begins with a brief overview of multidimensional chromatography in general and of the development and fundamental theory of GCxGC. Instrumentation for GCxGC is described with a focus on the differences between GCxGC and traditional single-dimensional gas chromatography (GC). Finally, the clinical, pharmaceutical, biological, and forensic applications of GCxGC are reviewed through the lens of a hypothesis that GC is underutilized in these research areas and that GCxGC provides excellent potential for a wide range of separation problems in these fields.

DK4478_book.fm Page 217 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

A. OVERVIEW

OF

217

MULTIDIMENSIONAL SEPARATIONS

The family of multidimensional separation techniques is quite large, including nearly all chromatographic and extraction methods. All of them were developed according to the basic goal of separation science: to achieve optimum resolution of two or more sample components. In his introduction to the recent text Multidimensional Chromatography [2], Bartle uses the master resolution equation, shown below, to state that dramatic improvements in separation performance, or resolution, must be obtained by increasing selectivity: ⎛ N ⎞ ⎛ α − 1⎞ ⎛ k ⎞ 2 Rs = ⎜ ⎟⎜ ⎟⎜ ⎟ ⎝ 4 ⎠ ⎝ a ⎠ ⎝ 1 + k2 ⎠

(6.1)

Increases in theoretical plates (N) may be obtained by optimizing instrumental and column parameters, such as column dimensions, sampling techniques and detector settings, but the changes are incremental in most cases. Increasing resolution through retention factors (k) is limited by k/1 + k approaching 1 as k becomes large. This leaves selectivity (α) as the remaining variable from which the dramatic resolution gains needed for increasingly complex samples may be obtained. Most work in multidimensional chromatography, whether through serially coupled columns, column switching, selective detection, or enhanced sample preparation, has focused on increasing selectivity. Most of the work that is traditionally termed multidimensional chromatography involves the serial coupling of columns with differing chemistries, whether on- or off-line, to produce increased selectivity. The second critical variable, which is especially evident in GC, is the peak capacity,

n=

N ⎛ t2 ⎞ ln +1, 4 R ⎜⎝ t1 ⎟⎠

(6.2)

which relates the number of peaks that may be obtained in a given one-dimensional separation space to the resolution, theoretical plates, and a retention time window between t1 and t2. It follows that increases in N and even very large retention time windows (long analysis times) will only incrementally improve peak capacity in one-dimensional separations. The major advantage in using a two-dimensional separation space is that the peak capacities for each dimension, which are always greater than one, are multiplied to gain the peak capacity of the coupled system, generating dramatic increases in separating power.

B. CLASSICAL HEART-CUTTING

IN

GC

Prior to the development of GCxGC, and still in use today, heart-cutting techniques were commonly employed for the generation of two-dimensional gas chromatographic separations. Heart-cutting involves the switching of a selected portion of

DK4478_book.fm Page 218 Thursday, September 14, 2006 1:35 PM

218

Advances in Chromatography, Volume 45

the effluent from a traditional column using a six-port valve, onto a second column of differing stationary phase chemistry. Before and after this heart-cutting period, the traditional column is directly connected to a detector. This allows a portion or portions of a complex chromatogram to be further evaluated on the second column. However, heart-cutting is not considered a comprehensive two-dimensional separation in that only a few small portions of the first dimension chromatogram are sampled into the second column. Several runs are therefore required to generate a full, comprehensive two-dimensional chromatogram. An example heart-cut GC separation is shown in Figure 6.1. Heil et al. [3] separated chiral urinary acids (pyroglutamic acid isomers) using an achiral first column, followed by a chiral second dimension. Note the complexity of the first-dimension separation and the simplicity of the second dimension. This nicely illustrates the most common goal, achieving greater selectivity and resolution, of using the second-dimension separation.

A

Backflush Cut 0

15

30

45

60

Time (min)

188% B

(D) PGA

84

4400 73:21

(L)

4500 75:01

4600 76:41

4700 78:21

4800 80:01

4900 81:41

s min

FIGURE 6.1 Example of classical heart-cut GC. The small section of the first-dimension (achiral) chromatogram is injected into the (chiral) second-dimension column. Only one second-dimension chromatogram is obtained. (Reprinted with permission from Heil et al., J. Chromatogr. B., 714, 119, 1998. Copyright 1998 Elsevier Science.)

DK4478_book.fm Page 219 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

C. DEVELOPMENT

OF

219

COMPREHENSIVE GCXGC

The first reference shown from a SciFinder Scholar search of “comprehensive twodimensional gas chromatography” is a 1991 article “Comprehensive two-dimensional gas chromatography using an on-column thermal modulator interface,” published in the Journal of Chromatographic Science [4]. This initial article provides many of the fundamental ideas present in today’s application of the technique, including the use of thermal modulation and the combination of numerous short second-dimension chromatograms into a high peak capacity retention plane. For nearly its first decade, development in GCxGC was relatively slow, with only a few researchers working on the technique. Over the next several years, application areas were explored, including the petroleum-related samples that dominate the field today and the beginnings of clinical-related analysis. In 1993 Venkatramani and Phillips [5] reported the separation of 6000 peaks from kerosene, and in 1994 Liu et al. [6] showed the analysis of pesticides from human serum and saw 15 pesticides in less than 4 minutes. Through the second half of the 1990s, developments continued at the pace of a few papers per year, dealing with additional applications, mostly in petroleum analysis and in instrumental improvements: more reliable modulation, column selection, initial attempts at quantitative analysis, and software development; detection remained primarily with fast flame ionization detection (FID) [7–14]. The development of GCxGC is somewhat unusual in the present research environment. Although a potentially revolutionary development, it remained in relative obscurity for nearly a decade, with only a few publications per year, mostly produced by closely collaborating research groups. The untimely death of its inventor, John Phillips, in 1999, coincides roughly with the beginning of increased attention to comprehensive GCxGC, although these likely have little to do with one another. More likely, the vast improvements in computing technology and solid-state electronics within gas chromatographs were the primary movers behind the surge in GCxGC. As the instrumentation has become reliable and mainstream, attention to the technique has increased.

D. OVERVIEW OF COMPLEX SEPARATIONS IN CLINICAL, BIOLOGICAL, AND FORENSIC ANALYSIS Two-dimensional GC has been mostly applied in the petroleum and petrochemical industries and has been somewhat ignored in most others. This is not surprising, as petroleum and related samples provide some of the most complex separation problems and thus especially difficult qualitative and quantitative analysis. High selectivity is needed to produce the needed resolution, as the samples often include large numbers of closely related structural isomers. In other industries, it is perhaps believed that the additional separation power in GCxGC is not necessary. In their detailed review, Dalluge et al. [15] point out that this assumption is false: GCxGC will find applications in many industries. In clinical, biological, forensic, and pharmaceutical analysis, there are numerous examples of complex single-dimension separations that could benefit from the availability of the second dimension.

DK4478_book.fm Page 220 Thursday, September 14, 2006 1:35 PM

220

Advances in Chromatography, Volume 45

I

D

M

1st

2nd

FIGURE 6.2 Schematic of GCxGC instrumentation showing the inlet, first-dimension column, modulator, second-dimension column, and detector. (Reprinted with permission from Dalluge et al., J. Chromatogr. A, 1000, 69, 2003. Copyright 2003, Elsevier Science.)

III. GCxGC FUNDAMENTALS AND INSTRUMENTATION One important advantage in GCxGC is that most GCxGC instruments are built by modifying traditional single-dimension instruments. A simple schematic of a GCxGC system is shown in Figure 6.2. The instrument is modified by adding an independently heated secondary oven as a fourth heated zone, usually within or attached to the outside of the main oven. This additional oven may be much smaller than the main oven, as the second-dimension column is usually much shorter and has a smaller inside diameter than the first-dimension column. The two columns are typically joined using a simple butt connector. The modulating device, described by most authors as the “heart” of GCxGC, is placed at the beginning of the seconddimension column. It serves to trap and focus analytes as they elute from the firstdimension column and inject them as a narrow band onto the second-dimension column at regular intervals, typically of a few seconds. A traditional detector, such as FID, is coupled to the end of the second column. Recently rapid mass spectrometers, such as time-of-flight (TOF), have been used in a variety of applications, providing yet another dimension of selectivity.

A. MODULATION TECHNIQUES The heart of a GCxGC system is the modulation device that provides the interface between the first- and second-dimension columns. The modulator must perform several tasks involved in transferring the first-column effluent into the second column. The modulator acts as a combined fraction collector and injector. First, the effluent from the first-dimension column must be divided into a series of fractions, which are continuously injected as the narrowest possible bands into the second column. The modulation timing must be coordinated with the temperature conditions of both columns to avoid wraparound: elution of compounds from the seconddimension column after the next modulation cycle has begun. These peaks will show up in the “next” chromatogram and possibly interfere with it. Several modulation schemes have been described: valve modulation, thermal modulation, and cryogenic modulation using liquid carbon dioxide or liquid nitrogen. In today’s commercially available systems, cryogenic modulation has emerged as the technique of choice.

DK4478_book.fm Page 221 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

221

Valve modulation was developed by Bruckner et al. [16] and Seeley et al. [17], and involves the placement of a simple multiport valve between the first- and seconddimension columns. Small fractions of the first-dimension effluent are transferred to the second column at regular intervals using a splitting process. This system has been used in extensive chemometric studies of GCxGC retention behavior, but was not generally applicable to quantitative analysis. One advantage is that additional cryogens and the associated handling systems are not needed. More recently, Bueno and Seeley [18] developed a simple flow-switching device for mechanically performing modulation that allows the use of two second-dimension columns. Thermal modulation was used in much of the early fundamental work of Phillips and Beens [19]. Originally this consisted of a two-stage, metal, painted or coppercoiled capillary that was heated by a constant-voltage power supply [20–22]. This was replaced by a moving device containing a separate heating element that came to be known as the “sweeper” [23,24]. The sweeper operates by moving a small heated element rapidly over the modulator capillary at regular intervals. Heating provided by this element desorbs analytes rapidly into the second-dimension column. Recently, attention and instrument development has focused on cryogenic modulation, using either liquid nitrogen or liquid carbon dioxide. As with other cryogenic applications in GC, cryogenic modulation involves cooling the end of the firstdimension column or the beginning of the second-dimension column with the cryogen to collect and refocus the analyte within the effluent, and then heating the same column portion to desorb the trapped analytes into the second-dimension column. Timing of the second-dimension chromatograms is achieved by controlling the cooling and heating times. The main advantage of focusing and modulation using liquid nitrogen or carbon dioxide is that even small molecules, such as gaseous hydrocarbons, can be easily modulated. A schematic diagram of a modern, typical two-stage modulator is shown in Figure 6.3. In this example, the two columns are connected using a press-fit connector. The jets are aligned with the head of the second-dimension column. First, the right-hand jet is activated to trap analytes at the head of the second column. To make the injection into the second column, the right-hand jet is turned off; the column heats either by simple heating to the oven temperature or by a stream of heated gas. Simultaneously the left jet is activated to prevent analytes from continuing to elute during the injection process. Finally, the right jet is reactivated and the left jet turned off to begin another collection cycle. Detailed reviews and discussions of cryogenic modulation are provided in articles by Dalluge et al. [15], Pursch et al. [25], Harynuk and Gorecki [26], and Cavagnino et al. [27].

B. INJECTION

AND

SAMPLING

Injection and sampling are the aspects of GCxGC that are most similar to classical single-dimension GC, as most GCxGC instruments are based on their classical counterparts. The first-dimension column is nearly always an open tubular column of traditional dimensions. Online sampling and sample preparation techniques can add further selectivity beyond the gains already realized by the two-dimensional column configuration. Most sampling and injection techniques that are used with

DK4478_book.fm Page 222 Thursday, September 14, 2006 1:35 PM

222

Advances in Chromatography, Volume 45

A

CO2 liquid or cold N2, gas

Valves Jets Press-fit connector

1st dim. column

2nd dim. column Cooling B1

Cooling

Injection in 2nd column B2

Cooling B3 1st dim. column

2nd dim. column

FIGURE 6.3 Schematic of two-jet cryogenic modulator. (A) Modulator design. (B1) Focusing the first-dimension eluent using the right-hand jet. (B2) Rapidly desorbing the focused band into the second column. (B3) Restarting the cycle by starting the right-hand jet. (Reprinted with permission from Dalluge et al., J. Chromatogr. A, 1000, 69, 2003. Copyright 2003, Elsevier Science.)

GC have been attempted with GCxGC; choices have been application dependent and most of the GCxGC literature therefore focuses on applications, columns, and detectors rather than inlets and sampling. A summary of inlets and sampling techniques that are commonly used with GC and that can be applied to GCxGC is presented in Table 6.1, with general references to assist users in better understanding these techniques. Examples of their use with GCxGC are included throughout this chapter. The addition of a selective injection or extraction process prior to GCxGC analysis adds yet another dimension to the selectivity of GCxGC-based analytical methods. For example, Cavagnino et al. [27] used a novel large-volume injection technique, performed using a customized glass sleeve within a classical splitless inlet, to determine polycyclic aromatic hydrocarbons (PAHs) in complex mixtures such as composite diesel fuel. They detected parts-per-million-level PAHs in both neat composite diesel fuel and in fuel that was diluted 2500:1. This demonstrates a powerful combination for trace analysis that could also be very useful in examining natural products and in the analysis of drugs and pharmaceuticals in biological samples. Using a programmed temperature vaporization (PTV) inlet with GCxGC time-of-flight mass spectrometry (TOFMS), Focant et al. [28] performed

DK4478_book.fm Page 223 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

223

TABLE 6.1 Summary of Inlets and Common Injection and Sampling Techniques with Key Technique Split injection

Splitless injection

On-column injection

PTV injection

Large volume injection Headspace extraction SPME and sorptive microextraction Pressurized liquid extraction and SFE

Membrane-based extraction Liquid-liquid extraction Single-drop microextraction

Description Liquid sample injection with splitting to reduce sample size. Common uses in group separations, fingerprinting, and neat mixtures. Can be optimized for very rapid injection at the cost of smaller sample volume. Uses same instrumentation as split, with split purge vent closed. Nearly entire injected sample reaches the column. Most common method for trace quantitative analysis. Liquid sample injected directly into the capillary column. Entire sample (and interferences) reaches the column. Generally performed cool. Most inert injection technique. Cool injection into inlet liner followed by rapid temperature increase to desorb sample into column. Usually applied to large volume injections. Uses PTV inlet for rapid injection of samples up to 100 μl or more. Sampling of vapors above a solid or liquid sample. Use of sorbent material such as PDMS to extract analytes from liquid or headspace. Use of high-pressure heated liquids (e.g., methanol/water mixtures) and gases (e.g., carbon dioxide) for static and continuous extraction. Commonly used in analysis of soils, foods, and other very complex sample matrices. Use of semipermeable membrane to transfer analytes between liquid or vapor matrices based on size or functionality. Classical extraction using immiscible liquid phases followed by liquid injection such as split or splitless. New liquid-liquid extraction technique using a single drop suspended from a syringe needle followed by split or splitless injection.

References 71

72

73

74

75 76 77, 78 79, 80

81 82 83

qualitative analysis of 58 halogenated compounds, with baseline resolution of 57 of the compounds. Headspace solid phase microextraction (SPME), in combination with chiral and achiral GCxGC, was used by Williams et al. [29] to analyze the volatile components from strawberries to differentiate between berries from different growing regions and to examine the effects of storage. They noted that SPME extraction and optimization procedures remain essentially the same when coupled with GCxGC as with traditional GC. It is interesting to note, however, that they used the less selective polydimethylsiloxane (PDMS) fiber coating, while their literature search seemed to recommend the more selective polyacrylate (PA) coating. This presents an interesting and unexplored possible benefit of GCxGC: the possibility of revisiting sampling and sample preparation to make it less selective. This thinking could greatly simplify the often difficult and time-consuming sample preparation involved with complex

DK4478_book.fm Page 224 Thursday, September 14, 2006 1:35 PM

224

Advances in Chromatography, Volume 45

biological and natural samples. In a similar vein, Ong et al. [30] demonstrated pressurized liquid extraction (PLE) with GCxGC for the analysis of PAHs in soil. They also showed that method development on the extraction side was no different in GCxGC than in traditional GC. Their method compared favorably to a traditional gas chromatography/mass spectrometry (GC/MS) method for quantitative analysis, but showed the superior qualitative analysis for complex mixtures exhibited in GCxGC. Additional SPME-GCxGC applications are provided by Shao et al. [31], who used SPME with on-fiber derivatization to analyze trans-resveratrol in wine, and by Zini et al. [32], who examined eucalyptus clones and leaf volatiles.

C. COLUMN SELECTION With the addition of the second-dimension column, column selection in GCxGC becomes somewhat more complex than in traditional GC. In order for a separation technique to be truly two-dimensional, the two columns must be orthogonal: they must exhibit significantly different intermolecular interactions that drive analyte separation. In the early work on GCxGC, which primarily involved analysis of hydrocarbons and related compounds from petroleum and related samples, column selection was straightforward: the first-dimension column was a nonpolar column typically used in hydrocarbon separations and the second-dimension column was typically a moderately polar column that would exhibit selectivity for various functional groups or aromatic properties. With GCxGC beginning to be employed in far broader applications, column selection becomes an important, but mostly unexplored aspect of method development. In the process of developing a GCxGC column test mixture along the same lines as the classical Grob test mixtures for traditional GC [33], Dimandja et al. [34] performed a comprehensive survey of the column choices used in GCxGC work dating to 2003. They found that about 90% of applications at that time used a nonpolar column in the first dimension, with about 70% of those using a semipolar second-dimension column and most of the remainder using a polar second-dimension column. This is not surprising given that the majority of applications involve petroleum and related analyses. The test mixture that they developed includes several homologous series, providing a wide range of functionalities that cover a large portion of the two-dimensional separation space. Figure 6.4 shows an apex plot, which simplifies the retention plane by showing just a data point at the apex of each peak instead of the entire peak, of an analysis of the test mix, with the analytes sorted by compound class. It is interesting to see that even with the wide range of compounds chosen, the separation space remains relatively unfilled; most of the peaks are closely grouped. Recently Ryan et al. [35] addressed the issue of orthogonality in GCxGC with an eye toward maximizing use of the available separation space in two dimensions after many authors observed that much of the available separation space is unused. They showed that altering the polarity of the first-dimension column affects elution temperature from the first column, and thus the first-dimension retention time and retention on the second column. They use these ideas to accurately predict twodimensional separations, but still most of their experimental and predicted separa-

DK4478_book.fm Page 225 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

225

4

3.5

3

2.5 ??

??

Alkanes Alkenes Alkynes Aldehydes ?? Ketones Alcohols FAMEs Diethylethers Carboxylic acids ?? Alkylbenzenes Cycloalkanes Naphthalenes Xylenes Grob mix compounds

2 ??

1.5 0

200

400

600

800 1000 1200 1400 1600 1800 2000

FIGURE 6.4 Apex plot of GCxGC test mixture shows regular patterns for the elution of the homologous series and the classical Grob test mix components. (Reprinted with permission from Dimandja et al., J. Chromatogr. A, 1019, 261, 2003. Copyright 2003, Elsevier Science.)

tions do not make use of the full separation space. Fully using the separation space will continue to be a challenge in GCxGC. Thus, column selection in GCxGC continues to be based on traditional GC ideas: the first-dimension column is usually chosen to match previous traditional analyses of the analyte mixture; the seconddimension column is usually more polar, or is selected to match the particular chemistry of the analytes or a portion of them. With the differences in stationary phase polarity, orthogonality is usually assumed, but is rarely optimized. In an interesting study of orthogonality, Adahchour et al. [36] reversed the traditional column combination. They used more polar columns in the first dimension and less polar columns in the second. They found that in a number of cases, including diesel oil, olive oil, and vanilla extract, improved peak shapes and sensitivity were obtained, especially for a number of the usual target analytes from these samples.

D. DETECTORS Most of the common detectors, including flame ionization, electron capture, and mass spectrometers, have been used with GCxGC. The rapid second-dimension chromatogram necessitates that the detector have extremely fast response and equilibration times and the usual small inside diameter of the second-dimension capillary column requires an extremely small detector cell volume. Consequently, most GCxGC work has been performed using flame ionization detectors, the most common detector used for organic compound analysis in GC, which has nearly instantaneous response, and with most data systems is capable of data acquisition rates of 200 Hz. MS provides an additional high degree of selectivity, however, traditional quadrupole mass spectrometers are criticized for scanning too slowly.

DK4478_book.fm Page 226 Thursday, September 14, 2006 1:35 PM

226

Advances in Chromatography, Volume 45

With rapid scanning, TOFMS has emerged as the detection method of choice for GCxGC, although it represents considerable additional expense. In this section, unique detector applications with quadrupole mass spectrometers and other selective detectors are described. FID and TOFMS applications are described elsewhere throughout this chapter. Several authors have coupled GCxGC with quadrupole MS, which is much less expensive than TOFMS and more convenient because of the much smaller data handling and analysis requirements. Song et al. [37] stated that full data analysis of a single analysis of cigarette smoke required 7 h of data processing time. They compared GCxGC-quadrupole MS with GCxGC-TOFMS for the semiquantitative analysis of 77 underivatized drugs as standards in methanol solution. By using a reduced scan range (42 to 235 μm) that scanned at 6.97 Hz and a truncated spectral search library for the quadrupole mass spectrometer, they obtained acceptable results for most of the drugs they studied. They compared full-scan spectra obtained with quadrupole MS to TOFMS and found that spectral skewing effects, possible with rapid eluting peaks in quadrupole MS, were minor. Figure 6.5 shows the comparison of GCxGC-quadrupole MS data for chlorpheniramine with GCxGC-TOFMS. The TOF results showed no spectral skewing in TOFMS, but the quadrupole MS data were sufficient for library searching and for semiquantitative screening. (II) GC × GC-TOFMS

(I) GC × GC-qMS B

4 × 105

C

A

0

E

Ion abundance

Ion abundance

2 × 105

F

D 0.0

31.56

31.60

31.64

1680

1678

1676

It (S)

It (min)

R

R

Ion abundance × 104

4

0

B

500

203

C 203

20

58 58 203 58

40

0 240 40

0 240 40

1100 58 Ion abundance

A

240

D

1100 58 203

0

40 140 240

E

1100 58

203

203

0

40 140 240

F

0

40 140 240

m/z

FIGURE 6.5 Comparison of mass spectra from GCxGC-quadrupole MS and GCxGCTOFMS. Plots (A) through (C) are the spectra obtained from the front, middle, and back of the GCxGC-quadrupole MS peak. Plots (D) through (F) are the spectra obtained from the front, middle, and back of the GCxGC-TOFMS peak. Note the artifactual differences in the quadrupole MS plots and the close similarity between the TOFMS plots. (Reprinted with permission from Song et al., J. Chromatogr. A, 1058, 223, 2004. Copyright 2004, Elsevier Science.)

DK4478_book.fm Page 227 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

227

Adahchour et al. [38] coupled GCxGC to a rapid-scanning quadrupole mass spectrometer for the analysis of polychlorinated biphenyls (PCBs) and food allergens, with a focus on understanding the effects of scan rate on spectral quality and quantitation. Their instrument had a scan speed of 10,000 da/sec and they found that they could monitor a 200 da mass range at a frequency of 33 Hz with no loss of spectral quality. At larger scan ranges, spectral quality decreased. A contour plot and the mass spectra they obtained for two of the compounds are shown in Figure 6.6, along with the library search results. The contour plot shows the multidimensional equivalent of a traditional total ion chromatogram, with the experimental mass spectra for two compounds shown on the left and the library search results on the right. The spectra obtained with GCxGC-quadrupole MS compared favorably with library search results. Element-specific detection in GCxGC has been demonstrated by van Stee et al. [39] for petrochemical analysis. While not specifically biological, clinical, or forensic analysis, the use of metal-containing compounds is increasing in clinical and pharmaceutical venues. GCxGC with element-specific detection may also find applications in forensic analysis of petroleum and related samples. Using GCxGC with atomic emission detection (AED), they identified several nitrogen- and sulfurcontaining compounds from complex petroleum samples. In a related study, Blomberg et al. [40] used sulfur chemiluminescence detection (SCD) with GCxGC to study the distribution of sulfur-containing compounds in petrochemical samples. They provide a detailed analysis of the effects of detector cell dimensions vs. the speed of the detector electronics in determining the data acquisition rate, critical for GCxGC. They show the identification of a number of naphthobenzothiophenes, dibenzothiophenes, benzothiophenes, thiophenes, sulfides, and thiols and they demonstrate how these show the grouping behaviors typically seen in GCxGC. A representative GCxGC-SCD analysis from this work is shown in Figure 6.7. Two-dimensional gas chromatography can be effectively used with detectors other than the typical flame ionization and TOF mass spectrometer types, which are obviously applied because of their rapid acquisition rates. For the other detectors, which typically respond more slowly, the detector is either physically modified or the conditions are carefully optimized. In any application of a new detector with GCxGC, additional detector optimization experiments are necessary in order to avoid excessive peak broadening and the resulting overlap and lower response. Analysis of biological, clinical, and forensic samples will typically require mass spectrometers or other selective detectors.

E. DATA HANDLING The continuous nature of GCxGC generates extremely high data throughput, with flame ionization and mass spectrometer detectors typically operated at data collection rates 10 or more times greater than in traditional GC. Two- and three-dimensional data presentation has provided challenges to the researchers and designers that develop the instrumentation, and there is much literature describing approaches to data processing. Two-dimensional chromatographic data processing presents challenges in peak integration for quantitative analysis, target and group compound

DK4478_book.fm Page 228 Thursday, September 14, 2006 1:35 PM

228

Advances in Chromatography, Volume 45

Second dimension retention time (s)

4.0

2.0

14

3.5 1

3.0

1.5

10 12

2.5

17

16

1.0 0.5

7 15

2.0

11

16.0

13

16.5

25

10.0

15.0

22.5

2 7.5

15.5

23 24 19 17 18 22 20 16 21

6

0.5

15.0

9 3 4 8 5

1.5 1.0

0.0

12.5

17.5

20.0

First-dimension retention time (min) %

%

135

100.0

107

75.0

107

75.0

150

91

50.0

135

100.0

150

91

50.0

123 55 53 65

25.0

79 77

106 95

59 72 83

0.0

75.0

50.0

25.0

206 149 191 145 163 173 188 205 221 235

121

98 100.0

125.0

150.0

175.0

200.0

55

79 77 83

0.0

225.0

105

65

50.0

75.0

149

95 100.0

125.0

(a)

163

150.0

175.0

200.0

225.0

%

131 147

131 147

100.0

188

188

91

91

117

117

57

75.0

O

57

204

204 105

106

50.0

50.0

65 25.0

191 206

(b)

% 100.0

75.0

O

123

55

0.0 50.0

77 129 145 79 107 161 134 71 89 ?? 156 171 184 206 215 83 94 ?? 75.0

100.0

125.0

150.0

(c)

175.0

200.0

65 77

25.0

89 237

225.0

107

129

145 134

161

0.0 50.0

75.0

100.0

125.0

150.0

175.0

200.0

225.0

(d)

FIGURE 6.6 GCxGC-quadrupole MS of a food allergen standard. The contour plot is shown on top with additional detail showing a perfume sample. Mass spectra (plots a and c) with library search results (plots b and d) for compounds 16 and 17 are shown below. (Reprinted with permission from Adahchour et al., J. Chromatogr. A, 1067, 245, 2005. Copyright 2005, Elsevier Science.)

DK4478_book.fm Page 229 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

229

Second dimension relative retention times (seconds)

7.50

6.25 Naphthobenzothiophenes 5.00 2–(3-thienyl) ethanol internal standard

Dibenzothiophenes

3.75

Benzothiophenes

2.50

1.25

0.00 0

25

50 75 First dimension retention time (minutes)

100

125

FIGURE 6.7 GCxGC-SCD of a Liverpool Bay crude oil sample shows several analyte groups. (Reprinted with permission from Blomberg et al., J. Chromatogr. A, 1050, 77, 2004. Copyright 2004, Elsevier Science.)

identification for qualitative analysis, and in the handling of very large volumes of mass spectrometric data in GCxGC-TOFMS. The commercially available instruments today include dedicated customized data systems that perform these tasks, along with additional instrument control functions both for traditional GC and the additional functions needed for GCxGC. Designing these complex data systems for the regulated environment found in pharmaceutical and forensic laboratories continues to be a challenge for data system designers. 1. Peak Integration and Identification The most common approach to peak integration involves using traditional integration techniques to generate peak areas in the second-dimension peaks and then summing the peak areas for each peak. Dalluge et al. [41] provide an excellent description of this process applied to the analysis of pesticides in food extracts. Although a food analysis example, this is of interest in clinical and other biological analyses, as food also provides a very complex matrix from which to extract trace analytes. Pesticides are also among the most difficult analytes to extract, as they are highly reactive with extraction materials such as glassware and the gas chromatographic inlet and column itself. The data visualization process is shown in Figure 6.8. First, the effluent at the first-dimension column outlet, which if detected would show broad, extremely overlapped peaks, is subjected to the modulation process to generate the family of related peaks shown in the “Raw 2D chromatogram.” Knowing the modulation timing, these raw data can be transformed or sliced

DK4478_book.fm Page 230 Thursday, September 14, 2006 1:35 PM

230

Advances in Chromatography, Volume 45

1D-chromatogram (at first column outlet) 1

Contour plot 3

2 3

n

sio

1 1st d imen

sion

1. Modulation

2

en

m

di

d

2n

3. Visualization

2. Transformation n Second-dimension chromatograms stacked side by side

sio

Raw 2D chromatogram (at second column outlet)

1st

dim

ensi

m

en

on

d 2n

di

FIGURE 6.8 Diagram of the data visualization process shows each of the steps in data processing. (Reprinted with permission from Dalluge et al., J. Chromatogr. A, 965, 207, 2002. Copyright 2002, Elsevier Science.)

into the separate second-dimension chromatograms, which are then visualized side by side. Finally, the final data are transformed and visualized as a three-dimensional contour plot. This is the method of choice for visualizing most GCxGC data today. Users new to GCxGC will quickly encounter one of its challenges: finding the useful data within the very large amount of generated data. While the major vendors of GCxGC instrumentation have developed sophisticated peak recognition and identification algorithms, many chromatograms still require significant manual data analysis. One challenge is illustrated in Figure 6.8: how are peaks summed from the stacked second-dimension data so that the entire volume of a peak is expressed and how is the true peak maximum located to define the retention time in both dimensions? In a complex analysis, there will easily be hundreds or thousands of peaks requiring identification and integration. 2. Target Compound Analysis Many applications in clinical and pharmaceutical analysis, such as doping control, quality assurance, and residual solvent analysis, involve the determination of a target compound whose one-dimensional retention times and mass spectra are well known. Since the two-dimensional separation begins with a one-dimensional analysis, the same peak identification techniques used in one-dimensional GC and GC/MS can be applied to GCxGC. Target compound analysis usually also requires precise and accurate quantitation, which has been a challenge in the development of GCxGC due to difficulties in the summation of peak areas in two-dimensional data. In addition to traditional integration, techniques based on principal component analysis have been developed to assist in peak deconvolution and quantitation [42,43]. A comparison of traditional integration with principal component analysis for quanti-

DK4478_book.fm Page 231 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

231

tation is presented by van Mispelaar et al. [44]. They analyzed six target essential oils from 32 different GCxGC chromatograms of perfume samples and found that traditional integration provided better accuracy and precision, while the principal component analysis was much easier to implement, faster and easier to automate. When MS is used for detection, isotope dilution is often used, especially in combination with selected ion monitoring, to enhance quantitative analysis of known target compounds. Ryan et al. [45] recently reported the first use of isotope dilution with GCxGC-TOFMS for quantitative analysis, with methoxypyrazines from wine as the analytes. With sample preparation by headspace SPME, they reported detection limits in the low nanogram per liter range. As there continue to be relatively few publications involving target compound analysis and purely quantitative applications of GCxGC, this remains a fruitful area for research, especially as so much of clinical, forensic, and biological analysis involves target analysis. 3. Fingerprinting and Group Analysis In petroleum and environmental applications, GCxGC has attracted special interest due to its fingerprinting and group analysis capabilities. Very often the need is more to identify whole classes of compounds that may be present in a sample (e.g., hydrocarbons, alcohols, thiols, etc.) than to identify individual components. As seen in Figure 6.4, each homologous series of compounds present appears in a predicable and reproducible pattern on the retention plane, making elucidation of the presence of the group members straightforward. In fingerprinting applications, reproducible retention patterns in two-dimensions produce many more peaks than in a single-dimension separation. See the discussion of oil spill fingerprinting in Section IV.C below.

IV. ADDITIONAL APPLICATIONS Comprehensive GCxGC is beginning to see wider application in research beyond petroleum and petrochemical analysis, mostly in the past 2 to 3 years. Additional applications in biological, clinical and drug testing, forensic, pharmaceutical, and natural products (including food) analyses are described in this section. The status of GCxGC for these analyses is indicated by most of the references being in specialized chromatography journals. GCxGC is still primarily practiced by chromatography specialists; it has not made the mainstream yet, as GC and highperformance liquid chromatography (HPLC) have been for decades.

A. BIOLOGICAL Biological analytes and sample matrices present some of the most complex separation and analytical problems. Often there are numerous analytes to be identified and quantified at once, following separation from highly complex sample matrices such as urine, blood, or tissues. In traditional GC, mass selective detection is often required to assist in identifying compounds of interest from very complex chromatograms. GCxGC should lend itself well to these analyses.

DK4478_book.fm Page 232 Thursday, September 14, 2006 1:35 PM

232

Advances in Chromatography, Volume 45

An excellent comparison of traditional GC/MS and GCxGC for the analysis of sterols from dried feces and green lip mussel tissue is provided by Truong et al. [46]. Determination of fecal contamination in water systems is one of the most important biological and environmental analyses. Using a 30 m × 0.25 mm × 0.25 μm 5% phenyl PDMS column in the first dimension and a 2 m × 0.1 mm × 0.1 μm 50% phenyl PDMS column in the second dimension, they obtained nine detectable sterols, compared to five with a traditional GC method. The key result in this study was the increased resolution of the sterols of interest from the numerous matrix components. In a plant biology example, Hope et al. [47] used GCxGC to study the small molecules, including amino and organic acid derivatives, for analysis of metabolites in rye grass. They examined 28 amino acids and 18 organic acids using both standards and extracts from lawn grass samples by GCxGC-TOFMS. They also used a combination of a nonpolar column in the first dimension with a more polar seconddimension column. Figure 6.9 shows total ion chromatogram contour plots for two rye grass samples, indicating six metabolites and numerous matrix components. It is obvious that compounds c through f would not be separated in a single dimension separation. It is clear, and stated by the authors, that GCxGC-TOFMS shows promise for metabolite analysis, but it is also stated by the authors that limitations in quantitation and sample visualization must be overcome. These authors have also extensively studied the use of multivariate statistics for peak deconvolution and the separation of overlapping spectra in GCxGC-TOFMS, applying the results to metabolite studies [48].

B. CLINICAL

AND

DRUG TESTING

Two-dimensional gas chromatography is promising for many applications in clinical analysis and forensic toxicology. These include drug testing and doping control in both human and animal testing situations. As in other applications, addition of the second-dimension separation allows for far more fingerprinting and analysis possibilities without spectroscopic detection than traditional single-dimension GC. GCxGC also shows promise for routine screening applications that are currently performed, perhaps unnecessarily, by GC/MS. The promise of GCxGC in drug analysis from urine and blood has been demonstrated by Kueh et al. [49], who first reported GCxGC in doping control in 2003, and by Song et al. [50], who reported drug screening and confirmation by GCxGCTOFMS in 2004. The Kueh et al. work is especially interesting, as the authors propose the use of GCxGC-FID for the separation and screening of 27 drugs from urine. This is potentially simple compared to other methods for routine screening. They used urine samples obtained from a greyhound (orally administered 10 mg of prolintane hydrochloride) and spiked blank horse urine. Most of these drugs were amino compounds, so they were derivatized to form the acylates. As in most other applications, a nonpolar first-dimension column was followed by a moderately polar second-dimension column. Chromatograms demonstrating the separation of a drug standard mixture three ways are shown in Figure 6.10. First is a traditional singledimension separation and second is the one-dimensional representation of the GCxGC plot. Note that these are very similar. The final plot is the contour plot,

DK4478_book.fm Page 233 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

233

Column 2 time (seconds)

1.4 1.2 1.0 a 0.8 b

0.6 0.4 0.2 12

16

20 24 28 32 Column 1 time (minutes)

36

(a)

Column 2 time (seconds)

1.4 1.2 1.0 c

0.8 0.6

d e

0.4 0.2 12

f 16

28 32 20 24 Column 1 time (minutes)

36

(b)

FIGURE 6.9 Total ion chromatogram contour plots of two rye grass samples. The lettered peaks are individual characteristic organic acids. Note the alignment of peaks c through f, showing that traditional GC would not have been adequate for this separation. (Reprinted with permission from Hope et al., Talanta, 65, 380, 2005. Copyright 2005, Elsevier Science.)

showing both dimensions. If the contour plot were viewed “on edge” the middle chromatogram would be obtained. It is clearly seen that a large number of the compounds would not be separated in a single-dimension separation. The authors further discuss a targeted approach to data analysis that assists in validating the second-dimension retention times, allowing for orthogonal peak identification. This has great potential benefits for screening and may reduce false-positive results in FID-based methods. Until recently, the need for mass spectral data in forensic toxicology presented a challenge to the use of GCxGC for drug analysis. GCxGC-TOFMS provides a ready solution to this problem. The Song et al. paper [50] describes the first use of GCxGC-TOFMS in drug screening and confirmation for 78 different drugs of abuse that did not require derivatization prior to gas chromatographic injection. The drugs were spiked in four groupings into drug-free blood and extracted by standard

DK4478_book.fm Page 234 Thursday, September 14, 2006 1:35 PM

234

Advances in Chromatography, Volume 45

Response (pA)

30 25 20

2 A 1 12

18 24 16 23 7 4 8 6 17 21 5 22 25

14 13

20

B

9

15 10

10

20

30 Time (min)

40

50

60

Response (pA)

(a) 140 120 100 80 60 40 20

2 4 20

1 12

3, 11, 19

6

22

13

10, 15

16

14 21

23 18 17 7 24 8

5

25 26 10

20

30 Time (min)

40

9

50

60

D2 retention time (s)

(b) 2.8

12

20

2

11

2.2

19

1.7

21 13

1.1

3

0.6 0.0

24

14

4 5.0

10.5

16.0

21.5

22

16

23

25 7 8 26 9 10, 15

6 17 18

5

27.0 32.5 38.0 43.5 DI retention time (min)

49.0

54.5

60.0

(c)

FIGURE 6.10 GCxGC analysis of drug standard. (A) Chromatogram from a single-dimension GC. Note that several peaks overlap. (B) Pulsed GCxGC analysis shows the one-dimensional chromatogram obtained by reconstructing the second-dimension chromatograms. (C) Twodimensional contour plot shows complete separation of the drugs. (Reprinted with permission from Kueh et al., J. Chromatogr. A, 1000, 109, 2003.)

procedures. As in most other applications, the first-dimension column was nonpolar (5% phenyl PDMS) and the second dimension was moderately polar (50% phenyl PDMS). These authors present a thorough discussion of the optimization process, including sample preparation, column selection, modulation temperature and rate, data acquisition rate, and mass spectral conditions. Figure 6.11 shows the GCxGC analysis of a real case sample showing the one-dimensional separation GCxGCTOFMS contour plot from total ion chromatogram data and a two-dimensional extracted ion chromatogram with masses characteristic of each analyte shown. Note the disappearance of the large matrix peaks between plots B and C.

DK4478_book.fm Page 235 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

235

9

400

600

800

1000 1200 Time (s)

1400

1600

1800

(a) 4.0 3.0 3 2.0

4

9

5

1

250

500

750

1000 (b)

1250

1500

1750

4.0 3

2D

retention time (s)

7 8

2

1.0 0.0

6

4

9

3.0

7 8

5 6

2.0 1.0

2

1

0.0 250

500

750 1000 1250 1D retention time (s)

1500

1750

(c)

FIGURE 6.11 GCxGC-TOFMS of a drug case sample. (a) One-dimensional total ion chromatogram. (b) Two-dimensional contour plot total ion chromatogram. (c) Two-dimensional contour plot extracted ion chromatogram. Peak identification (extracted ion): 1) methamphetamine (58), 2) ketamine (180), 3) doxylamine (58, 167), 4) metropolol (72, 107), 5) promethazine (72,180), 6) diazepam (256), 7) nordiazepam (242), 8) temazepam (271), 9) chloresterol. (Reprinted with permission from Song et al., Forensic Sci. Int., 143, 87, 2004.)

DK4478_book.fm Page 236 Thursday, September 14, 2006 1:35 PM

236

Advances in Chromatography, Volume 45

These are currently the only two reports on doping analysis using GCxGC. Clearly GCxGC has the potential to be highly useful in separating drugs of interest from the complex matrices of urine and blood. Both papers report quantitative enhancement due to the rapid second-dimension separation and both show separations of compounds that did not separate in the single-dimension analysis. Applying GCxGC to standard drug testing methods appears straightforward now that effective software for quantitation and mass spectral analysis exists. The methods for using quadrupole MS reported in Section III.D of this chapter also show promise for drug analysis applications.

C. FORENSIC Two-dimensional gas chromatography has shown much promise in several aspects of forensic chemical analysis. Given its beginnings in petroleum analysis, natural extensions have included forensic identification of oil spills and fingerprinting of arson accelerants. These provide some of the best and most thorough work in GCxGC to date for fingerprinting applications. Addition of the second-dimension separation allows much more detailed fingerprinting of these samples than is possible with traditional GC. Forensic toxicology applications are discussed above. Two-dimensional gas chromatography is a natural technique for the fingerprinting of complex samples and early on in its development, GCxGC was applied to the forensic analysis and identification of the source of samples from oil spills. GCxGC was first applied to specific problems in petrochemical fingerprinting by Blomberg et al. [9] in 1997. They showed several examples of the characterization of hydrocarbon process streams and light oils and demonstrated the powerful potential for fingerprinting of complex samples offered by GCxGC. The forensic use of GCxGC for fingerprinting and identifying the source of oil spills from collected samples has been performed primarily in the laboratory of Frysinger and Gaines at the U.S. Coast Guard Academy (New London, CT). They used GCxGC to examine several samples obtained from oil spill cases at the U.S. Coast Guard Marine Safety Laboratory, along with samples from suspected sources [51]. Using group separations of alkanes, cycloalkanes, alkylbenzenes, alkylnaphthalenes, and PAHs as fingerprints, the GCxGC results compared favorably with their previous GC- and GC/MS-based methods. They have continued to develop GCxGC for forensic analysis of oil spills, including more specific analyses of compound classes found within them, such as benzene, toluene, and xylene (BTEX) and petroleum biomarkers, and they were the first to apply GCxGC with MS for the analysis of petroleum [52–54]. They have also used GCxGC to characterize and identify arson accelerants and weathered accelerants using analysis of fire debris pyrolyzates and study of the chemical changes that occur during weathering of the samples [55].

D. PHARMACEUTICAL Two-dimensional gas chromatography has seen little direct introduction and use in pharmaceutical applications. As with GC, which many agree is underutilized in pharmaceutical analysis, GCxGC will likely be used for the analysis of organic

DK4478_book.fm Page 237 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

237

volatile impurities (OVIs) in raw materials and finished products as required by international regulations [56]. As with environmental analysis of OVIs, this is one of the more complex target analysis applications, as there are more than 80 regulated compounds and the samples may include the complex matrix interferences found in tablets, capsules, ointments, creams, gels, and devices. GCxGC has the potential to significantly reduce analysis time and improve resolution for OVIs, which often requires single-dimension separations of 30 to 60 min, with incomplete resolution of all possible components. These methods typically require proof of method selectivity by separating the large numbers (80 or more) of OVIs that can possibly be present, while any actual sample may only contain a few (3 to 6) of them. A few unique applications of OVI analysis by GCxGC have been demonstrated, including the use of headspace SPME-GCxGC for OVI emissions from wounded plants and a dual second-dimension column GCxGC separation of more than 100 OVIs [57,58].

E. NATURAL PRODUCTS Natural products and related analyses are highly complex and GCxGC lends itself well to these problems. Most of the work has centered on the identification and fingerprinting of volatile oils and other compounds from a variety of samples, including herbs, medicines, and foods. The addition of substances, such as pesticides, to food is also of great interest and has been explored by GCxGC. Outside of petroleum and related products, perhaps the biggest interest in GCxGC has been in the analysis of volatile oils from foods, flavors, spices, and other natural products. Much of the work on essential oil analysis has been performed by Shellie and Marriott [59], including a thorough review in 2003 in which they describe both conventional and enantioselective essential oils analysis. Some of the applications include essential oils from Pelargonium gravoleans using GCxGCquadrupole MS [60], volatile oils in Chinese herbs such as ginseng by GCxGC with headspace SPME, and Chinese medicines by GCxGC-TOFMS [61–63]. Wild coriander essential oils, especially as odorants, were identified by Eyers et al. [64] using GCxGC-TOFMS and compared favorable with GC olfactometry. They were able to identify the most abundant odorants and discovered several previously unreported compounds in Coriandrum sativum and Eryngium foetidum. Essential oils from tobacco were identified and quantified by Zhu et al. [65] using GCxGC-TOFMS, again with many more components identified than in traditional GC/MS analysis. Shellie et al. [66] separated and identified the eight representative components of sandalwood oil using GCxGC-TOFMS. In each case, GCxGC provides far more resolving power than traditional GC or GC/MS. All of these authors reported identification of previously unknown components of these various natural samples. In food analysis, GCxGC development has focused on the complex problems of flavors, fatty acid methyl esters, and additives, which have all proven difficult in many cases with traditional GC and GC/MS. In the analysis of pesticides from foods, Dalluge et al. [41] used GCxGC-TOFMS to identify 58 pesticides from lettuce extracts using full-scan MS. They noted that this was not possible with a singledimension system. Additional work on adulterants has involved the analysis of allergens in fragrances using GCxGC-TOFMS by Shellie et al. [67]. They examined

DK4478_book.fm Page 238 Thursday, September 14, 2006 1:35 PM

238

Advances in Chromatography, Volume 45

the 24 components listed by the European Commission’s Scientific Committee on Cosmetics and Non-Food Products, which is an especially difficult analysis due to the complexity of cosmetic samples. Coffee bean volatiles were examined by Ryan et al. [68]. They used GCxGCTOFMS in comparison with a previous report on GCxGC-FID and demonstrated the superior qualitative and quantitative analysis capabilities of TOFMS. Fatty acid methyl ester analysis is of major interest in the food industry, as there are stringent health and labeling requirements for various types of fatty acids in food products. Hyoetylaeinen et al. [69] used GCxGC to analyze the dietary fatty acids in milk. They evaluated several column combinations and demonstrated the superior separating power of GCxGC with acceptable quantitation. This is one of the few applications that employed GCxGC-FID. Mondello et al. [70] used GCxGC to explore the analysis of fatty acid methyl esters, which are nearly always very complex samples. They reported on column optimization and fingerprinting that is not available in traditional GC.

V. CONCLUSION Comprehensive two-dimensional gas chromatography is an up-and-coming analytical technique. With straightforward instruments and commercial software now available, applications are beginning to be seen outside of the traditional chromatography literature. As a relatively new method, there are several advantages, especially since the required modifications to GC can be made to a traditional instrument. Although most of the application development has been in the petroleum and related industries, there is great potential for GCxGC in other areas where GC has traditionally not been the technique of choice, including clinical and pharmaceutical analysis. GCxGC also shows great promise in biological and food analysis, especially with complex mixtures of volatile compounds such as flavors, essential oils, and fatty acids. GCxGC shows its greatest power when combined with TOFMS, as the separations are complex; the MS is needed for rapid peak identification. It is clear that any laboratory that performs complex mixture analysis or expects to do so can benefit from GCxGC and should examine its possibilities.

ACKNOWLEDGMENTS The author gratefully acknowledges the support of LECO Incorporated for providing instrumentation to begin his GCxGC research efforts.

REFERENCES 1. SciFinder Scholar 2004 CAS, American Chemical Society, 2004, accessed August 2005. 2. Bartle, K., Introduction, in Mondello, L., Lewis, A.C., and Bartle, K.D., eds., Multidimensional Chromatography, John Wiley & Sons, Chichester, 2002, p. 3.

DK4478_book.fm Page 239 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

239

3. Heil, M., Podebrad, F., Beck, T., Mosandl, A., Sewell, A., and Bohles, H., Enantioselective multidimensional gas chromatography-mass spectrometry in the analysis of urinary organic acids, J. Chromatogr. B Biomed. Sci. Appl., 714, 119, 1998. 4. Liu, Z. and Phillips, J., Comprehensive two-dimensional gas chromatography using an on-column thermal modulator interface, J. Chromatogr. Sci., 29, 227, 1991. 5. Venkatramani, C. and Phillips, J., Comprehensive two-dimensional gas chromatography applied to the analysis of complex mixtures, J. Microcol. Sep., 5, 511, 1993. 6. Liu, Z., Sirimanne, S.R., Patterson, D., Needham, L., and Phillips, J., Comprehensive two-dimensional gas chromatography for the fast separation and determination of pesticides extracted from human serum, Anal. Chem., 66, 3086, 1994. 7. Venkatramani, C., Xu, J., and Phillips, J., Separation orthogonality in temperatureprogrammed comprehensive two-dimensional gas chromatography, Anal. Chem., 68, 1486, 1996. 8. Grainger, J., Green, V., Liu, Z., Barr, J., McClure, C., Patterson, D., Jr., Holland, J., and Gardner, B., Organohalogen Compounds, 27, 354, 1996. 9. Blomberg, J., Schoenmakers, P., Beens, J., and Tijssen, R., Comprehensive twodimensional gas chromatography (GC×GC) and its applicability to the characterization of complex (petrochemical) mixtures, J. High Resolut. Chromatogr., 20, 539, 1997. 10. Beens, J., Boelens, H., Tijssen, R., and Blomberg, J., Quantitative aspects of comprehensive two-dimensional gas chromatography (GC×GC), J. High Resolut. Chromatogr., 21, 47, 1998. 11. Beens, J., Tijssen, R., and Blomberg, J., Comprehensive two-dimensional gas chromatography (GC×GC) as a diagnostic tool, J. High Resolut. Chromatogr., 21, 63, 1998. 12. Phillips, J. and Xu, J., Environmental applications of comprehensive two-dimensional gas chromatography, Organohalogen Compounds, 31, 199, 1997. 13. Gaines, R., Ledford, E., and Stuart, J., Analysis of water samples for trace levels of oxygenate and aromatic compounds using headspace solid-phase microextraction and comprehensive two-dimensional gas chromatography, J. Microcol. Sep., 10, 597, 1998. 14. Kinghorn, R. and Marriott, P., Comprehensive two-dimensional gas chromatography using a modulating cryogenic trap, J. High Resolut. Chromatogr., 21, 620, 1998. 15. Dalluge, J., Beens, J., and Brinkman, U.A.T., Comprehensive two-dimensional gas chromatography: a powerful and versatile analytical tool, J. Chromatogr. A, 1000, 69, 2003. 16. Bruckner, C., Prazen, B., and Synovec, R., Comprehensive two-dimensional highspeed gas chromatography with chemometric analysis, Anal. Chem., 70, 2796, 1998. 17. Seeley, J., Kramp, F., and Hicks, C., Comprehensive two-dimensional gas chromatography via differential flow modulation, Anal. Chem., 72, 4346, 2000. 18. Bueno, P. and Seeley, J., Flow-switching device for comprehensive two-dimensional gas chromatography, J. Chromatogr. A, 1027, 3, 2004. 19. Phillips, J. and Beens, J., Comprehensive two-dimensional gas chromatography: a hyphenated method with strong coupling between the two dimensions, J. Chromatogr. A, 856, 331, 1999. 20. Venkatramani, C. and Phillips, J., Comprehensive two-dimensional gas chromatography applied to the analysis of complex mixtures, J. Microcol. Sep., 5, 511, 1993. 21. Liu, Z. and Phillips, J., Sensitivity and detection limit enhancement of gas chromatographic detection by thermal modulation, J. Microcol. Sep., 6, 229, 1994.

DK4478_book.fm Page 240 Thursday, September 14, 2006 1:35 PM

240

Advances in Chromatography, Volume 45

22. de Geus, H.-J., de Boer, J., and Brinkman, U.A.T., Development of a thermal desorption modulator for gas chromatography, J. Chromatogr. A, 767, 137, 1997. 23. Phillips, J. and Ledford, E., Thermal modulation: a chemical instrumentation component of potential value in improving portability, Field Anal. Chem. Tech., 1, 23, 1996. 24. Phillips, J., Gaines, R., Blomberg, J., van der Wielen, F.W.M., Dimandja, J., Green, V., Granger, J., Patterson, D., Racovalis, L., deGeus, H.J., DeBoer, J., Haglund, P., Lipsky, J., Sinha, V., and Ledford, E., A robust thermal modulator for comprehensive two-dimensional gas chromatography, J. High Resolut. Chromatogr., 22, 3, 1999. 25. Pursch, M., Eckerle, P., Biel, J., Streck, R., Cortes, H., Sun, K., and Winniford, B., Comprehensive two-dimensional gas chromatography using liquid nitrogen modulation: set-up and applications, J. Chromatogr. A, 1019, 43, 2003. 26. Harynuk, J. and Gorecki, T., New liquid nitrogen cryogenic modulator for comprehensive two-dimensional gas chromatography, J. Chromatogr. A, 1019, 53, 2003. 27. Cavagnino, D., Magni, P., Zilioli, G., and Trestianu, S., Comprehensive twodimensional gas chromatography using large volume injection for the determination of polynuclear aromatic hydrocarbons in complex matrices, J. Chromatogr. A, 1019, 211, 2003. 28. Focant, J.-F., Sjodin, A., and Patterson, D., Qualitative evaluation of thermal desorption-programmable temperature vaporization-comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry for the analysis of selected halogenated contaminants, J. Chromatogr. A, 1019, 143, 2003. 29. Williams, A. Ryan, D., Olarte Guasca, A., Marriott, P., and Pang, E., Analysis of strawberry volatiles using comprehensive two-dimensional gas chromatography with headspace solid-phase microextraction, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 817, 97, 2005. 30. Ong, R., Lundstedt, S., Haglund, P., and Marriott, P., Pressurised liquid extractioncomprehensive two-dimensional gas chromatography for fast-screening of polycyclic aromatic hydrocarbons in soil, J. Chromatogr. A, 1019, 221, 2003. 31. Shao, Y., Marriott, P., and Hugel, H., Solid phase microextraction — on-fibre derivitisation with comprehensive two dimensional gas chromatography analysis of transresveratrol in wine, Chromatographia, 57(suppl.), S349, 2003. 32. Zini, C.A., De Assis, T.F., Ledford, E.B., Jr., Davina, C., Fachel, J., Christensen, E., and Pawliszyn, J., Correlations between pulp properties of eucalyptus clones and leaf volatiles using automated solid-phase microextraction, J. Agric. Food Chem., 51, 7848, 2003. 33. Grob, K., Jr. and Grob, K., Evaluation of capillary columns by separation number or plate number, J. Chromatogr., 207, 291, 1981. 34. Dimandja, J.-M., Clouden, G., Colon, I., Focant, J.-F., Cabey, W., and Parry, R., Standardized test mixture for the characterization of comprehensive two-dimensional gas chromatography columns: the Phillips mix, J. Chromatogr. A, 1019, 261, 2003. 35. Ryan, D., Morrison, P., and Marriott, P., Orthogonality considerations in comprehensive two-dimensional gas chromatography, J. Chromatogr. A, 1071, 47, 2005. 36. Adahchour, M., Beens, J., Vreuls, R.J.J., Batenburg, A.M., and Brinkman, U.A.T., Comprehensive two-dimensional gas chromatography of complex samples by using a “reversed-type” column combination: application to food analysis, J. Chromatogr. A, 1054, 47, 2004. 37. Song, S.M., Marriott, P., and Wynne, P., Comprehensive two-dimensional gas chromatography — quadrupole mass spectrometric analysis of drugs, J. Chromatogr. A, 1058, 223, 2004.

DK4478_book.fm Page 241 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

241

38. Adahchour, M., Brandt, M., Baier, H.-U., Vreuls, R.J.J., Batenburg, A.M., and Brinkman, U.A.T., Comprehensive two-dimensional gas chromatography coupled to a rapid-scanning quadrupole mass spectrometer: principles and applications, J. Chromatogr. A, 1067, 245, 2005. 39. van Stee, L.L.P., Beens, J., Vreuls, R.J.J., and Brinkman, U.A.T., Comprehensive two-dimensional gas chromatography with atomic emission detection and correlation with mass spectrometric detection: principles and application in petrochemical analysis, J. Chromatogr. A, 1019, 89, 2003. 40. Blomberg, J., Riemersma, T., van Zuijlen, M., and Chaabani, H., Comprehensive two-dimensional gas chromatography coupled with fast sulphur-chemiluminescence detection: implications of detector electronics, J. Chromatogr. A, 1050, 77, 2004. 41. Dalluge, J., van Rijn, M., Venís, J., Vreuls, R.J.J., and Brinkman, U.A.T., Comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometric detection applied to the determination of pesticides in food extracts, J. Chromatogr. A, 965, 207, 2002. 42. Fraga, C., Prazen, B., and Synovec, R., Comprehensive two-dimensional gas chromatography and chemometrics for the high-speed quantitative analysis of aromatic isomers in a jet fuel using the standard addition method and an objective retention time alignment algorithm, Anal. Chem., 72, 4154, 2000. 43. Fraga, C., Prazen, B., and Synovec, R., Enhancing the limit of detection for comprehensive two-dimensional gas chromatography (GC×GC) using bilinear chemometric analysis, J. High Resolut. Chromatogr., 23, 215, 2000. 44. van Mispelaar, V., Tas, A., Smilde, A., Schoenmakers, P., and van Asten, A., Quantitative analysis of target components by comprehensive two-dimensional gas chromatography, J. Chromatogr. A, 1019, 15, 2003. 45. Ryan, D., Watkins, P., Smith, J., Allen, M., and Marriott, P., Analysis of methoxypyrazines in wine using headspace solid phase microextraction with isotope dilution and comprehensive two-dimensional gas chromatography, J. Sep. Sci., 28, 1975, 2005. 46. Truong, T., Marriott, P., Porter, N., and Leeming, R., Application of comprehensive two-dimensional gas chromatography to the quantification of overlapping faecal sterols, J. Chromatogr. A, 1019, 197, 2003. 47. Hope, J., Prazen, B., Nilsson, E., Lidstrom, M., and Synovec, R., Comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry detection: analysis of amino acid and organic acid trimethylsilyl derivatives, with application to the analysis of metabolites in rye grass samples, Talanta, 65, 380, 2005. 48. Sinha, A., Hope, J., Prazen, B., Fraga, C., Nilsson, E., and Synovec, R., Multivariate selectivity as a metric for evaluating comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry subjected to chemometric peak deconvolution, J. Chromatogr. A, 1056, 145, 2004. 49. Kueh, A., Marriott, P., Wynne, P., and Vine, J., Application of comprehensive twodimensional gas chromatography to drugs analysis in doping control, J. Chromatogr. A, 1000, 109, 2003. 50. Song, S., Marriott, P., Kotsos, A., Drummer, O., and Wynne, P., Comprehensive twodimensional gas chromatography with time-of-flight mass spectrometry (GCxGCTOF-MS) for drug screening and confirmation, Forensic Sci. Int., 143, 87, 2004. 51. Gaines, R., Frysinger, G., Hendrick-Smith, M., and Stuart, J., Oil spill source identification by comprehensive two-dimensional gas chromatography, Environ. Sci. Technol., 33, 2106, 1999.

DK4478_book.fm Page 242 Thursday, September 14, 2006 1:35 PM

242

Advances in Chromatography, Volume 45

52. Frysinger, G., Gaines, R., and Ledford, E., Quantitative determination of BTEX and total aromatic compounds in gasoline by comprehensive two-dimensional gas chromatography (GC×GC), J. High. Resolut. Chromatogr., 22, 195, 1999. 53. Frysinger, G. and Gaines, R., Separation and identification of petroleum biomarkers by comprehensive two-dimensional gas chromatography, J. Sep. Sci., 24, 87, 2001. 54. Frysinger, G. and Gaines, R., Comprehensive two-dimensional gas chromatography with mass spectrometric detection (GC × GC/MS) applied to the analysis of petroleum, J. High Resolut. Chromatogr., 22, 251, 1999. 55. Frysinger, G. and Gaines, R., Forensic analysis of ignitable liquids in fire debris by comprehensive two-dimensional gas chromatography, J. Forensic Sci., 47, 471, 2002. 56. ICH Guidelines, International Conference on Harmonisation, Geneva, Switzerland. 57. Ranjini, P., Marriott, P., and Galbally, I., Headspace solid phase microextraction — comprehensive two-dimensional gas chromatography of wound induced plant volatile organic compound emissions, Analyst, 127, 1601, 2002. 58. Seeley, J., Kramp, F., Sharpe, K., and Seeley, S., Characterization of gaseous mixtures of organic compounds with dual-secondary column comprehensive two-dimensional gas chromatography (GCx2GC), J. Sep. Sci., 25, 53, 2002. 59. Shellie, R. and Marriott, P., Opportunities for ultra high resolution essential oils analysis using comprehensive two-dimensional gas chromatography. A review, Flavour Fragr. J., 18, 179, 2003. 60. Shellie, R. and Marriott, P., Comprehensive two-dimensional gas chromatographymass spectrometry analysis of Pelargonium graveolens essential oil using rapid scanning mass spectrometry, Analyst, 128, 879, 2003. 61. Di, X., Shellie, R., Marriott, P., and Huie, C., Application of headspace solid-phase microextraction (HS-SPME) and comprehensive two-dimensional gas chromatography (GC×GC) for the chemical profiling of volatile oils in complex herbal mixtures, J. Sep. Sci., 27, 451, 2004. 62. Wu, J., Lu, X., Tang, W., Kong, H., Zhou, S., and Xu, G., Application of comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry in the analysis of volatile oil of traditional Chinese medicines, J. Chromatogr. A, 1034, 199, 2004. 63. Shellie, R., Marriott, P., and Huie, C., Comprehensive two-dimensional gas chromatography (GCxGC) and GCxGC-quadrupole MS analysis of Asian and American ginseng, J. Sep. Sci., 26, 1185, 2003. 64. Eyers, G., Dufour, J.-P., Hallifax, G., Sotheeswaran, S., and Marriott, P., Identification of character-impact odorants in wild coriander leaves using gas chromatography olfactometry (GCO) and comprehensive two-dimensional gas chromatography time of flight mass spectrometry (GCxGC-TOFMS), J. Sep. Sci., 28, 1061, 2005. 65. Zhu, S., Lu, X., Xing, J., Zhang, S., Kong, H., Xu, G., and Wu, C., Comparison of comprehensive two-dimensional gas chromatography — time of flight mass spectrometry and gas chromatography-mass spectrometry for the analysis of tobacco essential oils, Anal. Chim. Acta, 545, 224, 2005. 66. Shellie, R., Marriott, P., and Morrison, P., Comprehensive two-dimensional gas chromatography with flame ionization and time-of-flight mass spectrometry detection: qualitative and quantitative analysis of West Australian sandalwood oil, J. Chromatogr. Sci., 42, 417, 2004. 67. Shellie, R., Marriott, P., and Chaintreau, A., Quantitation of suspected allergens in fragrances: evaluation of comprehensive two-dimensional gas chromatography for quality control, Flavour Fragr. J., 19, 91, 2004.

DK4478_book.fm Page 243 Thursday, September 14, 2006 1:35 PM

Comprehensive Two-Dimensional Gas Chromatography

243

68. Ryan, D., Shellie, R., Tranchida, P., Casilli, A., Mondello, L., and Marriott, P., Analysis of roasted coffee bean volatiles by using comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry, J. Chromatogr. A, 1054, 57, 2004. 69. Hyoetylaeinen, T., Kallio, M., Lehtonen, M., Lintonen, S., Peraejoki, P., Jussila, M., and Riekkola, M.-L., Comprehensive two-dimensional gas chromatography in the analysis of dietary fatty acids, J. Sep. Sci., 27, 459, 2004. 70. Mondello, L., Casilli, A., Tranchida, P., Dugo, P., and Dugo, G., Detailed analysis and group-type separation of natural fats and oils using comprehensive twodimensional gas chromatography, J. Chromatogr. A, 1019, 187, 2003. 71. Grob, K., Split and Splitless Injection in for Quantitative Gas Chromatography, Wiley-VCH, Weinheim, 2001, p. 149. 72. Grob, K., Split and Splitless Injection in for Quantitative Gas Chromatography, Wiley-VCH, Weinheim, 2001, p. 257. 73. Grob, K., On-Column Injection in Capillary Gas Chromatography: Basic Technique, Retention Gaps, Solvent Effects, 2nd ed., Wiley-VCH, Weinheim, 1991. 74. Janssen, H.-G., Sample Introduction Techniques for Capillary Gas Chromatography, Gerstel, Baltimore, 2001, http://www.gerstelus.com. 75. Snow, N., in Modern Practice of Gas Chromatography, 4th ed., Grob, R. and Barry, E., eds., John Wiley & Sons, New York, 2004, p. 461. 76. Kolb, B. and Ettre, L., Static Headspace Gas Chromatography, John Wiley & Sons, New York, 1997. 77. Pawlisyn, J., Solid Phase Microextraction: Theory and Practice, John Wiley & Sons, New York, 1999. 78. Vercauteren, J., Peres, C., Devos, C., Sandra, P., Vanhaecke, F., and Moens, L., Stir bar sorptive extraction for the determination of ppq-level traces of organotin compounds in environmental samples with thermal desorption-capillary gas chromatography — ICP mass spectrometry, Anal. Chem., 73, 1509, 2001. 79. Brachet, A., Rudaz, S., Mateus, L., Christen, P., and Veuthey, J., Optimisation of accelerated solvent extraction of cocaine and benzoylecgonine from coca leaves, J. Sep. Sci., 24, 865, 2001. 80. Imma, F. and Furlong, E., Accelerated solvent extraction followed by on-line solidphase extraction coupled to ion trap LC/MS/MS for analysis of benzalkonium chlorides in sediment samples, Anal. Chem., 74, 1275, 2002. 81. Jonsson, J. and Mathiasson, L., Membrane-based techniques for sample enrichment, J. Chromatogr. A, 902, 205, 2000. 82. Snow, N., Inlet systems for gas chromatography, in Modern Practice of Gas Chromatography, 4th ed., Grob, R. and Barry, E., eds., John Wiley & Sons, New York, 2004, p. 549. 83. Psillakis, E. and Kalogerakis, N., Developments in single-drop microextraction, Trends Anal. Chem., 21, 54, 2002.

DK4478_book.fm Page 244 Thursday, September 14, 2006 1:35 PM

DK4478_C007.fm Page 245 Friday, September 22, 2006 8:41 AM

7

Determination of Endocrine Disrupting Chemicals Found in Environmental Samples by Gas Chromatography/ Mass Spectrometry James D. Stuart

CONTENTS I. II. III. IV. V.

Introduction................................................................................................245 Pertinent GC/MS Research Papers............................................................247 Polychlorinated and Polybrominated Compounds ....................................259 Analysis of Natural Phytohormones by GC/MS.......................................263 Special Gas Chromatographic Methods: Online SPE Concentration, Solid Phase Microextraction, and Stir Bar Extraction..............................265 VI. Newer Developments in Chromatographic Methods................................268 References..............................................................................................................271

I. INTRODUCTION In this review chapter I have endeavored to include the more recent developments in the analyses of the various groups of organic compounds reported or alleged to be endocrine disrupting chemicals (EDCs), as they are found in environmental, primarily aquatic samples. This review chapter follows the same format and uses the same compound abbreviations (acronyms) as used in a widely referenced 2002 review paper on the same topic [1]. Aside from the various review articles published by various authors, the annual and biennial reviews in Analytical Chemistry by Susan D. Richardson and her coauthors provide excellent summaries of the newer and emerging issues that are important to contaminant and environmental research [2–4].

245

DK4478_C007.fm Page 246 Friday, September 22, 2006 8:41 AM

246

Advances in Chromatography, Volume 45

Pollution of the marine environment caused by the discharge of wastewater from sewage treatment plants has become an important international topic. Recent scientific reports have documented that a wide variety of both natural and man-made chemicals are being added to the environment by these allowed discharges. Certain of these chemicals or their by-products have been shown to have strong hormonal effects on the endocrine system of living organisms and that these chemicals are persistent and recalcitrant. Endocrine disrupting chemicals consist of a wide variety of different groups of chemicals that may alter, compete, or displace in living species certain of their important natural estrogens from their receptor sites, thereby changing the functions of those natural hormones [5]. EDCs have been defined by the European Commission as follows: “An endocrine disrupter is an exogenous substance or mixture that alters function(s) of the endocrine system and consequently causes adverse health effects in an intact organism, or its progeny, or (sub)populations” [6]. Studies have shown that EDCs can be absorbed from water and solid sediments into the marine life and then may be bioaccumulated by orders of magnitude up the food chain. Increased body loads of EDCs may cause birth defects, altered immune functions, sexual dysfunctions, and even cancers and possibly heart disease [5]. It was been demonstrated that in laboratory studies with the model fish species Japanese medaka (Oryzias latipes), either the natural estrogens 17β-estradiol or the synthetic estrogen 17α-ethynylestradiol can cause total sex reversal of male to female fish at levels of 5 to 10 μg/L, and can cause the secretion of the female-specific protein vitellogenin by male fish at a concentration of 1 μg/L [7]. The chromatographic analyses of various groups of chemicals reported or alleged to have endocrine disrupting activity are currently of international environmental concern. Modern chromatographic methods, especially when they are interfaced to a mass spectrometer for detection, allow for the simultaneous determination, often at nanogram per liter or picogram per gram levels, of various different groups of chemical compounds. It needs to be emphasized that proper sample collection, storage, and workup prior to the actual analysis step is most important. Many of the literature references have reported using gas chromatography (GC) with either a selective detector or with a mass spectrometric detector for the analyses of specific groups of EDCs. To improve the sensitivity and selectivity, especially for the semipolar and more polar compounds, such polar compounds were derivatized off-line. More recently, the combination of liquid chromatography (LC) and mass spectrometry (MS) interfaced by the newer electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI) methods has allowed for the determination of various of the more polar compounds without prior chemical group derivatization. Often minimal sample preparation, except for perhaps sample preconcentration, is required for LC/MS and liquid chromatography/tandem mass spectrometry (LC/MSMS). This review addresses only the new advances that have occurred when GC/MS is used to analyze EDCs in environmental samples. As is often the case, the actual method of analysis that a laboratory uses is dictated by various factors, including instrumentation availability, operator expertise, and requirements to follow mandated governmentally decreed methods.

DK4478_C007.fm Page 247 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

247

II. PERTINENT GC/MS RESEARCH PAPERS In a 2001 research article, which in the opinion of this reviewer should be consulted first, Espejo et al. [8] reported the determination of nineteen 4-alkylphenol EDCs by both GC/MS and LC/MS in Geneva, Switzerland’s municipal sewage wastewater. Valuable background information is given into the manufacture, degradation processes, and analyses of 4-alkyl polyethoxylates. The final degradation products of alkyl polyethoxylates are various alkylphenol isomers, such as octylphenols and nonylphenols (NPs), whose endocrine disrupting capabilities in various marine species are well documented. The authors point out that 4-n-nonylphenol (4-n-NP) can be used as a convenient surrogate standard, but that it should not be used as a representative compound, as its precursor has not been found in the technical industrial alkyl polyethoxylate products. Espejo et al., as others after them have done, report developing several analytical methods by both GC/MS and LC/MS to simultaneously determine both alkyl polyethoxylates as well as their different by-products, including the various alkylphenols. As Espejo et al. note, in contrast to most publications, since their wastewater samples formed very fine emulsions which tended to clog ordinary filtration media and solid phase extraction (SPE) cartridges, they chose to do multiple stages of liquidliquid extractions. Both their GC/MS and LC/MS were done by analyzing final extracts that were not further derivatized, which explains why their reported detection limits (s/n > 3) for the various alkylphenol isomers were in the range of 0.4 to 6 ng/L and their quantification limits (s/n > 10) were in the range of 2 to 22 ng/L. They reported that after the primary wastewater treatment step (crude solid filtration), the wastewater contained on average 2.5 μg/L (range 1.0 to 6.8 μg/L) of free 4-alkylphenols and up to 0.66 mg/L of those compounds were present as 4-alkyl polyethoxylates. In 2001, Bester et al. [9] used both GC/MS and LC/MS to analyze NPs, nonylphenol ethoxylates (NPEs), and various linear alkylbenzene sulfonates (LASs) in water and sediment samples from the German Bight of the North Sea. They noted that while NPEs are no longer used in detergents, they may still be used in metal and textile processing, in the paper industry, and in the production of paints and formulations of pesticides. As NPEs have been shown to be transformed to NPs in wastewater treatment processes, it is expected that a large amount of the different NP isomers may still be entering the environment. These authors used a simple solvent extraction of 1 L of n-pentane to 100 L of sea water or estuary water, followed by the addition of sodium sulfate to dehydrate the n-pentane layer and then volume reduction to 1 ml. This extract was further concentrated and fractionated using normal phase high-performance liquid chromatography (HPLC). Analysis of the various HPLC extracts was accomplished by GC/MS. For the combination of various NPs in a technical mixture of NP isomers, the limits of detection (LODs) were 0.01 ng/L with a recovery rate of 41%. Liquid chromatographic/atmospheric pressure chemical ionization tandem mass spectrometry (LC-APCI-MSMS) was employed for the analyses, and LODs of 10 ng/L for NPEs and 5 ng/L for the LASs were reported with precisions of 6% and 3% expressed as relative standard deviations (RSDs), respectively. In general the concentrations for NP were found to vary from 27 to 36 ng/L near the River Elbe estuary of Germany’s North Sea coast and 1 to 6 ng/L when sampled several hundred kilometers further into the central part of the North Sea. As

DK4478_C007.fm Page 248 Friday, September 22, 2006 8:41 AM

248

Advances in Chromatography, Volume 45

might be expected, higher concentrations of NPs (range 11 to 153 ng/g) were found in the inshore sediments collected from various estuaries and marinas in the area. Also in 2001, Kuch and Ballschmitter [10] reported analyzing surface and drinking waters for bisphenol A (BPA), 4-t-octylphenol (4-t-OP), the various isomers of NPs, estrone, 17α-estradiol, 17β-estradiol, and 17α-ethinylestradiol. Water samples of 1 to 5 L were extracted by SPE cartridge and the eluents blown down and derivatized with pentafluorobenzoyl chloride, which was reconstituted with hexane, further volume reduced, and analyzed by capillary column GC with a mass spectrometer using the negative chemical ionization mode (methane as the reagent gas) and the selected ion mode. All samples were also analyzed by GC-electron capture detection (ECD). LODs reported in spiked treated wastewaters were BPA, 20 pg/L; 4-t-OP, 50 pg/L; technical NPs and the various estrogens, in the range of 200 ng/L. BPA concentrations were found in all river waters in southern Germany, ranging from 16 to 500 ng/L, 4-n-NP was found ranging from 6 to 135 ng/L, and the various estrogens ranged from 200 pg/L to 5 ng/L. In 2002 Jeannot et al. [11] reported using both GC/MS, GC/MSMS, and LC/MS to detect a variety of alcohol polyethoxylates, NPEs, octylphenol ethoxylates, 4-nNP, 4-t-OP, BPA, and a variety of natural and synthetic estrogens in both surface and wastewater samples as well as in certain domestic sludge samples. Different extraction methods and types of SPE cartridges were used for sampling. Compound derivatization with N,O-bis(dimethylsilyl)trifluoroacetamide (BSTFA) followed by both split mode, large volume (40 μl) injection GC/MS and splitless (2 μl) GC/MSMS were used. The authors reported quantitation limits (s/n > 10) in the lower nanogram per liter range for most of their analytes in spiked surface and wastewater samples. The sludge samples were Soxtec extracted prior to SPE cleanup with C18 cartridges. BPA-d16 was used as the internal standard. Table 7.1 summarizes their reported limits of quantitation (LOQs) for select EDCs from different matrices and analysis methods.

TABLE 7.1 Comparison of Quantitation Limits for Select Analytes LOQ (s/n = 10)a

Compound

40 μl Volume Injection/GC/MS (FS) (in Spiked Lab Water)

2 μl Volume/GC/MSMS (FS) (in Spiked Lab Water)

LC-MS (SIM) (in Spiked Wastewater)

4-t-OP 4-n-OP BPA 17β-estradiol

5 ng/L 2 ng/L 0.5 ng/L 2 ng/L

1 ng/L 1 ng/L 0.5 ng/L 3 ng/L

100 ng/L 100 ng/L ND ND

FS, full-scan total ion; SIM, selected ion monitoring; ND, not determined. a

LOQs calculated from 10 injections of the lowest concentration of the standard solutions.

Source: Data from Jeannot, R., Sabik, H., Sauvard, E., Dagnac, T., and Dohrendorf, K., J. Chromatogr. A, 974, 143, 2002.

DK4478_C007.fm Page 249 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

249

TABLE 7.2 Summary of Validated Parameters Compound

Estrone

17β-estradiol

Estriol

Ethinylestradiol

Detection limit (s/n = 3) Quantification limit (= 10) Repeatabilitya (10 ng/L, CV) Reproducibilitya (10 ng/L, CV) Accuracyb (10 ng/L)

0.03 pg 0.04 ng/L 2.4% 2.6% –2.8%

0.02 pg 0.04 ng/L 0.7% 1.5% –1.3%

0.05 pg 0.08 ng/L 5.9% 5.9% +4.9%

0.24 pg 0.32 ng/L 5.6% 5.6% –4.7%

a b

Refers to Mouatassim-Souali et al. [12], describing statistical validation of the method. Defined as relative error.

In 2003 Mouatassim-Souali et al. [12] reported an analysis method to measure both natural estrogens and their conjugated forms (sulfated and glucuronide) in both wastewater, river, and surface water samples collected around Paris, France. First, they used an enzymatic hydrolysis step to convert the conjugated forms of any of the estrogens to their free form. After glass filtration, solid phase speedisk (Speedisk C18, 59 mm, 1.2 μm bed height; J.T. Baker, Phillipsburg, NJ) was used to extract the target analytes, estrone (E1), ethinyl estradiol (E2), estriol (E3), and the synthetic ethinyl estradiol (EE2). After elution from the disk and solvent reduction, pentafluoropropionic acid anhydride derivatization was done, followed by GC/MS using both the full-scan and selected monitoring mode. Table 7.2 shows the validated parameters that they reported. Their validated method was used for the quantitative determination of the four targeted estrogens and their conjugates in samples taken in January 2002 from both the influent and effluent of four wastewater treatment plants, two located upstream and two located downstream from the City of Paris. In all of their analyses of the wastewater samples, significant concentrations of estrogens at concentration levels ranging from 2 to 18 ng/l were observed. However, the authors reported that no estrogen conjugates were detected in these wastewater samples. In the river water samples, the concentrations ranged from 0.5 to 3 ng/L, with the highest concentrations found at sites near the wastewater treatment plants [12]. Also in 2003, Zafra et al. [13] reported developing a simultaneous determination for BPA as well for four of its chlorinated adducts in wastewater samples after doing liquid-liquid extraction, organic solvent concentration, and derivatization with BSTFA followed by GC/MS. The authors explain that since chlorination is the most widely used method of water disinfection, it makes sense that a method be developed to analyze for BPA and its chlorinated by-products. It was reported that their method provided the following detection limits: BPA, 0.3 ng/L; Cl-BPA, 0.6 ng/L; Cl2-BPA, 2.0 ng/L; Cl3-BPA, 4.5 ng/L; Cl4-BPA, 13.0 ng/L. Wastewater samples from urban Granada, Spain were found to be free of BPA and any of its chlorinated byproducts at the LODs reported. In 2004 Liu et al. [14] reported the simultaneous determination of certain alkylphenols and estrogens from river and ocean water using SPE and BSTFA + 1%

DK4478_C007.fm Page 250 Friday, September 22, 2006 8:41 AM

250

Advances in Chromatography, Volume 45 160 140

Recovery (%)

120 100 80 4-tert-octylphenol

60

4-nonylphenol Bisphenol A

40

Estrone Estradiol

20

Hydroxyestrone Ethynylestradiol

LB H

V+

as is

EN 8/ C1

ut e-

O

8 C1 ut eol Is

ol

Is

-1

SC -1 8 D SC -S i D PA -δ S

Sl at a

D

X St r

St ra ta

St ra ta

CN

0

SPE cartridges

FIGURE 7.1 The recovery of EDCs from different SPE cartridges using methanol as eluent. (Reprinted from Liu, R., Zhou, J.L., Wilding, A., J. Chromatogr. A, 1022, 179, 2004. With permission from Elsevier.)

trimethylchlorosilane (TMCS) derivatization followed by GC/MS. Their target compounds included 4-n-NP, 4-t-OP, BPA, estrone, 17β-estradiol, 16α-hydroxyestrone, and 17α-ethynylestradiol. Figure 7.1 shows data on the percent recovery of their target compounds when nine different SPE adsorbents were compared. Methanol was used as the elution solvent when spiking levels of 200 ng/L for BPA and 400 ng/L for all of their target compounds were used. Highest recoveries (80 to 120%) were obtained from the polymeric (Oasis HLB, Waters Corporation, Milford, MA) and C18 SPE phases when either ethyl acetate or methanol was the eluting solvent. They reported other analyte recovery studies that compared different SPE elution solvents, the effect of adding different concentrations of sodium chloride to the water wash solution, the effect of the water’s pH, and the effect of humic acid concentrations in the water wash. For their seven target analytes spiked in ultrapure laboratory water, they reported that LODs ranged from 0.3 to 5.3 ng/L and LOQs were between 1.0 to 17.4 ng/L. Also in 2004, in a follow-up paper, Liu et al. [15] reported using microwaveassisted extraction (MAE) followed by a small cartridge silica gel column, and then the same derivatization with BSTFA (1% TMCS) and GC/MS to analyze for the same seven target analytes in river sediments. The advantage of the MAE technique is that there is a considerable reduction in both extraction time and organic solvent volume required in comparison to either liquid-liquid or Soxhlet extractions. Their second paper reported careful optimization of the conditions of the MAE (110°C,

DK4478_C007.fm Page 251 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

251

15 min, 100% power at 600 W), the organic solvent (methanol) used in the MAE, how the silica columns were conditioned (percent water); the best eluting solvent proved to be a mixture of ethyl acetate-hexane (4:6 [v/v]). Figure 7.2 shows the excellent separation of their seven derivatized analytes, along with internal standards, BPA-d16 (peak 4) and 17β-estradiol-d2 (peak 6) in a series of selected ion monitoring (SIM) chromatograms. Panel (a) shows the separation of their nine analyte standards. Panel (b) is a SIM chromatogram of a sediment extract spiked with 20 ng/g of their target analytes. Panel (c) is of an extract of a wastewater sediment sample showing peaks for BPA (peak 3), esterone (peak 5), 17β-estradiol (peak 7), and 17α-ethynylestradiol (peak 8). Table 7.3 reports the LODs (s/n = 3) and LOQs (s/n=10) for their seven target analytes by their reported method. Further, Liu et al. [14,15] reported applying their optimized method for the determination of their targeted compounds in river water samples taken from East and West Sussex, U.K. The concentrations of the targeted compounds were found to be relatively low, often below their LODs. For the compounds that were detected, their levels were found to be greater in water sampled closer to the wastewater sewage outfall than from either upstream or further downstream. In 2003 Lerch and Zinn [16] published a paper that compared the derivatization of 21 natural and synthetic EDCs, including 17β-estradiol, 17α-ethinylestradiol, 17β-testosterone, and BPA, using different MS ionization modes, electron impact, or chemical ionization while employing different reagent gas mixtures. The work of others reported that LOQs for most of the derivatized estrogenic hormones ranged between 0.05 and 1 ng/L in surface water and between 0.1 and 1 ng/L in wastewater. In their research, Lerch and Zinn found that most of the natural estrogens and their derivatized products underwent a large amount of fragmentation in the electron impact process. Their approach was to use chemical ionization to reduce the fragmentation, thus enabling MSMS and multiple MS techniques (MSn) to obtain lower LODs. Their paper reported using the following eight different fluoride-containing derivatizing agents: pentafluorobenzyl bromide (PFBBr), 3,5-bis-(trifluoromethyl)benzyl bromide (BTMBBr), octafluorotoluene (OFT), pentafluoropyridine (PFPy), fluoro-phenylalanine (FPA), heptafluorobutyrylimidazole (HFBI), heptafluorobutyric acid (HFBA), and trifluoroacetic acid anhydride (TFAA). The general conclusion from the paper was that the TFAA derivatization reagent with positive ion electron impact (not negative electron impact) or positive chemical ionization and with methane as the chemical ionization gas gave the best compromise of highest characteristic TFAA adducts, good GC peak separation, and MS spectral quality data. In 2004 Hernando et al. [17] published an interesting comparison of the determination of the following five compounds: 4-tert-octylphenol, BPA, estrone (E1), 17β-estradiol (E2), and 17α-ethynylestradiol (EE2) in wastewater samples doing both compound derivatization versus nonderivatization, both followed by GC/MSMS. Both methods included solid phase extraction with Oasis HLB cartridges, which according to the authors, gave an enrichment factor of about 100fold for the wastewater samples. The derivatization reagent, BSTFA, was used as the silylation reagent because it had rapid and complete reactivity with the hydroxyl groups, higher volatility, lower interaction with the GC column’s stationary phase, and produced fewer trimethylsilyl derivatized molecular ions. They reported a ten-

DK4478_C007.fm Page 252 Friday, September 22, 2006 8:41 AM

Advances in Chromatography, Volume 45

Relative abundance

252 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

M/z 3 368, 369

2

1

M/z 207, 208

M/z 179, 292

M/z 6 287, 418

4

M/z 357, 358 M/z 342 257 218

10

11

12

13

15 14 Time (min)

16

7

M/z 285, 416 326 M/z 285 425 232

5

17

18

M/z 286 430 8 9

19

20

Relative abundance

(a) 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

3

4 2

1

6 7 5

8

10

11

12

13

14 15 Time (min)

16

17

18

19

9

20

Relative abundance

(b) 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

3

5

4

10

11

12

13

14 15 Time (min)

16

17

18

8

6 7

19

20

(c)

FIGURE 7.2 Selected ion monitoring chromatogram of target EDCs in (a) standard solution (1 ng injection), (b) sediment sample spiked with 20 ng/g dry mass of the target compounds, and (c) sediment sample. Peak numbers refer to (1) 4-tert-octylphenol, (2) 4-n-NP, (3) BPAd16, (4) BPA, (5) estrone, (6) 17β-estradiol-d2, (7) 17β-estradiol, (8) 17α-ethynylestradiol, and (9) 16α-hydroxyestrone. (Reprinted from Liu et al., J. Chromatogr. A, 1038, 19, 2004. With permission from Elsevier.)

DK4478_C007.fm Page 253 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

253

TABLE 7.3 LODs and LOQs for Seven Target Analytes Compound

LOQ (s/n = 10) (ng/g)a

LOD (s/n) (ng/g)a

4-t-OP 4-n-OP BPA Estrone 17β-estradiol 17α-ethynylestradiol 16α-hydroxyestrone

1.7 1.7 3.4 0.9 0.9 1.4 0.5

0.5 0.5 1.0 0.3 0.3 0.4 0.2

a

Based on nanograms per gram of dry sediment spiked.

fold increase in the signal:noise ratio when their derivatized analytes were analyzed by GC/MS-ion trap in the selected ion mode in comparison to when the spiked wastewater samples were not derivatized. But what is most interesting in this paper is that Hernando et al. reported similar LODs, repeatability, and linearity when the same five compounds were not derivatized, but the extracts were analyzed by GC/MSMS. The MSMS parameters were optimized individually for each compound by using spiked SPE extracts of wastewater samples at a concentration of 10 μg/L. The most intense fragment ion in the electron impact spectrum was used as the precursor ion and up to three fragment ions were monitored to allow for enhanced selectivity, especially in the complex wastewater matrix, which was important for identification of 4-tert-octylphenol. The authors noted that for the routine analysis of wastewater samples, their GC/MSMS method was able to reduce the sample preparation step by avoiding the derivatization and avoided the addition of internal standards that are required in the GC/MS method. Comparable LODs, inter- and intraday reproducibilities, and linearities were obtained for the five target analytes by both the derivatization GC/MS (SIM) and nonderivatization GC/MSMS methods, except for BPA, which had a LOD of 26.5 ng/L by the former and better than a factor of 10 lower, or 2.5 ng/L, by the latter method. When different influent wastewater samples were analyzed, 20% showed high levels of BPA (range 885 to 1105 ng/L), 4-tert-octylphenol (40 to 44 ng/L), estrone (60 to 87 ng/L), and 17βestradiol (49 to 93 ng/L) as measured by the GC/MSMS method. Wastewater samples collected at the effluent from the same treatment plants gave lower concentrations: BPA (13 to 19 ng/L) and 4-tert-octylphenol (17 to 19 ng/L). In 2003 Ding and Chiang [18] published an important paper that compared the effectiveness of various GC derivatizing reagents (primarily trimethylsilylating agents) on the following natural and synthetic estrogens: estrone (E1), 4-hydroxyestrone (4-OH-E1), 17β-estradiol (E2), estriol (E3), 2-hydroxy estradiol (2-OH-E2), 17α-ethynylestradiol (EE2), mestranol (MeEE2), testosterone (TS), and diethylstilbestrol (DES). The difference in the percent derivatized of the various derivatizing

DK4478_C007.fm Page 254 Friday, September 22, 2006 8:41 AM

254

Advances in Chromatography, Volume 45

agents on these nine target estrogens are summarized in Table 7.4. The authors postulated that the difference in the percent derivatized was due to the steric hindrance of multiple hydroxyl groups or the presence of ethynyl groups in the compound’s structure. It was noted that the addition of a “stimulator,” for example, a small percentage of TMCS (say 1% TMCS with BSTFA), was found to increase the derivatization yield, especially for compounds with multiple hydroxyl groups. It needs to be emphasized that the formation of a single derivative with a strong molecular ion for the compound’s adduct is important to give the maximum sensitivity and selectivity. Informative discussions of their MS peak assignments and the relative abundances of each compound’s mass spectral pattern of the TMCS derivatives were presented. The derivatizing agent BSTFA + 1% TMCS was found to provide the highest percentage yield (refer to Table 7.4) and was used by the authors for the rest of their study. When BSTFA + 1% TMCS was employed, seven of the nine targeted estrogens were reported to have 100% derivatization, with the exceptions being EE2 (6%) and MeEE2 (3%), presumably due to the hydroxyl group at position 17 being protected from derivatization by the steric hindrance of the ethynyl group. They reported LOQs (s/n > 10) that ranged from 5 to 10 ng/L for the nine estrogens from SPE extracts of 1000 ml of spiked laboratory water samples (Table 7.5). Sample enrichment was achieved using Oasis HLB SPE cartridges (3 ml, 60 mg). They reported a calibration curve ranging from 0.1 to 5.0 ng/μl with excellent linearity when chrystene-d12 was used as the internal standard. Recoveries of 87 to 102% with 1% to 6% RSDs at a spiking level of 100 ng/L for three replicate 1000 ml deionized and river water samples were reported. However, none of their targeted estrogen compounds were detected in their local river samples. Within a year of Ding and Chiang’s paper [18], Shareef et al. [19] reported that the derivatizing reagent BSTFA + 1% TMCS caused the conversion of 17αethynylestradiol (EE2), with a 42% yield, to its estrone (E1) derivative and N-methylN-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) caused an almost 100% conversion to its respective estrone (E1) derivatives. Hence, for at least the analysis of EE2 in environmental samples, these two derivatizing reagents need to be used with caution. In 2005 Lin et al. [20] published a paper on the determination of primarily acidic pharmaceutical compounds using SPE, with the widely used SPE phase (Oasis HLB [polystyrene-divinylbenzene-N-vinyl pyrrolidone terpolymer; Waters Corporation). This was followed by online derivatization with tetrabutylammonium hydrogen sulfate and large volume (10 μl) injections onto a DB-5MS capillary column (30 m × 0.25 mm, 0.25 μm film) with a 2 m deactivated retention gap. The gas chromatograph was a Varian 3400CX connected to a Saturn ion trap mass spectrometer (Varian Inc., Palo Alto, CA). The authors reported that the basic Oasis HLB SPE phase produced greater recoveries of five of the acidic target drugs in comparison with the use of either RP-18 (ENVI-18 SPE, Supelco) or PS-DVB (LiChrolut EN, Merck), two other widely used SPE phases. Mass spectral interpretations of the butylated molecular ions and their fragmentation ions were discussed. LOQs (s/n > 10) of the pharmaceutical drugs clofibric acid, ibuprofen, carbamazepine, naproxen, ketoprofen, and diclofenac ranged from 1.0 to 8.0 ng/L in 500 ml of tap water. Recovery

O-TMS (100%)

O-TMS (100%)

O-TMS (100%)

O-TMS (100%)

O-TMS (100%)

BSTFA + 1% TMCS

BSTFA + 5% TMCS

BSTFA + 10% TMCS

N-methyl-N(trimethylsilyl) trifluoroacetamide (MSTFA)

E1

BSTFA

Derivatizing Agents

Bis-O-TMS (100%)

Bis-O-TMS (100%)

Bis-O-TMS (100%)

Bis-O-TMS (100%)

Bis-O-TMS (100%)

4-OH-E1

2-OH-E2

Bis-O-TMS (100%) O-TMS (12%)

Bis-O-TMS (100%) O-TMS ( 3), and LOQ (s/n > 10) for their detected or investigated compounds by both GC/ECD and GC/MSMS. While in most cases the GC/ECD method gave higher slopes in the calibration plots, the authors noted that GC/MSMS provided for higher selectivity and avoidance of false positives, and that the MSMS configuration provided for better matrix interference discrimination and allowed for specialized library searching. The chemical most commonly detected in the women’s blood proved to be endosulfan-α, detected because of its known low biodegradability and its expected ability to be bioconcentrated. The high incidences of endosulfan-α detection are believed to be due to its incorporation into the food chain and not from direct exposure to the pesticide. The authors advise that future studies should be designed for prioritize those EDCs for which hormonal activity has been previously determined or when overexposure is suspected. In a 2004 review article, Rieck [29] elaborated on the development of a method for the extraction, cleanup, and analysis of polybrominated diphenyl ethers (PBDEs) in fish tissue, soil, and water samples. Just as polychlorinated biphenyls (PCBs) were identified because unidentified peaks were found in environmental extracts being analyzed for chlorinated insecticides, so too were PBDEs first detected as GC peaks eluting after the Arochlor 1260 peaks. However, GC with an atomic emission detector (AED) set to measure the bromine emission line, had indicated that the compounds contained bromine. Subsequently virtually all of approximately 70 fish tissue extracts examined were found to contain PBDEs. Rieck reported that PBDEs were found in microgram per gram to nanogram per gram quantities in many different fish tissues. Interesting results from the analysis of the blubber of two different types of orca (killer) whales indicated that a whale from a transient pod had total levels of PCBs

DK4478_C007.fm Page 263 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

263

TABLE 7.8 Retention Time Windows (RTWs) and Calibration Data (n = 8) of the GC/ECD Method Pesticide

RTW (min)

Linear Ranges (μg/L)

r2

RSD (%)

LOD (μg/L)a

LOQ (μg/L)b

LOQ (μg/L)b

Ethoprophos Dichloran ISc Chlorothalonil Parathion-m Vinclozolin Fenitrothion Malathion Captan Procymidone Dieldrin Buprofezin Pyrazophos

23.80–23.86 28.69–28.78 32.94–32.98 34.21–34.25 40.73–40.77 41.85–41.89 45.21–45.27 47.56–47.60 53.23–53.25 55.65–55.67 58.85–58.87 59.47–59.49 74.30–74.32

1–100/100–1000 0.5–100/100–1000 — 0.5–100/100–1000 0.5–100/100–1000 0.5–50/50–1000 1–100/100–1000 0.5–100/100–1000 1–100/100–1000 0.5–100/100–1000 0.5–100/100–1000 10–100/100–1000 10–100/100–1000

0.998/0.990 0.997/0.996 — 0.999/0.997 0.999/0.994 0.993/0.997 0.999/0.990 0.999/0.990 0.996/0.999 0.992/0.992 0.994/0.997 0.996/0.990 0.992/0.998

3.1/2.8 3.6/3.4 — 3.8/3.5 5.7/4.9 1.8/1.6 2.2/1.8 2.3/2.1 2.3/2.0 1.5/0.9 1.2/1.1 5.4/4.9 3.5/3.1

0.1 0.1 — 0.1 0.2 0.1 0.2 0.1 0.1 0.1 0.1 3.0 2.0

0.5 0.4 — 0.4 0.8 0.5 0.8 0.4 0.5 0.4 0.4 10.0 10.0

1.0 0.5 — 0.5 0.5 0.5 1.0 0.5 1.0 0.5 0.5 10.0 10.0

Note: Calibration obtained using relative heights to that of IS. a b c

Based on the values of the blank at the tR of the analytes. Based on the lowest concentration where the RSD is estimated to be less than 10%. Where IS, or the internal standard, was pentachloronitrobenzene.

Source: Reproduced with permission from Vidal et al., J. Chromatogr. A, 867, 235, 2000.

(1254 and 1260) of 900,000 μg/kg (wet weight basis) and total PBDEs of about 8500 μg/kg, while a whale from a resident population had total levels of PCBs (1254 and 1260) of 8200 μg/kg (wet weight basis) and total PBDEs of about 800 μg/kg. The order of differences in concentration in body burdens were explained by the different types of foods that these orca whales are known to consume. A similar sample preparation and analysis method was used for the analysis of activated and dewatered sludge samples taken from three water treatment plants along the same rivers from which certain of the fish and bottom samples were collected. High milligram per kilogram to microgram per kilogram levels of the various PBDE congeners were reported. Rieck concluded that the high concentrations of PBDEs being discharged by the water treatment plants may be a significant source of PBDEs assumed to be adsorbed onto the sediments, and that the PBDEs could ultimately undergo biological uptake into the food chain to the various fish species.

IV. ANALYSIS OF NATURAL PHYTOHORMONES BY GC/MS In 2003 Birkemeyer et al. [30] reported developing a multitargeted quantification of various phytohormones, some of which might be expected to be strong EDCs

11.14–11.20 12.74–12.78 13.12–13.19 13.96–14.15 15.43–15.47 15.53–15.59 16.41–16.49 16.76–16.83 18.86–18.92 18.94–19.00 20.85–20.89 21.03–21.09 28.07–28.17

Ethoprophos Dichloran IS Chlorothalonil Vinclozolin Parathion-m Fenitrothion Malathion Captan Procymidone Dieldrin Buprofezin Pyrazophos

158 206 265 266 212 263 260 173 114 283 277 175 221

(94) (176) (237) (231) (115) (136) (138) (99) (79) (255) (206) (193) (210)

Ion 10–400 (1–400) 20–400 (5–400) — 10–800 (0.6–400) 1–800 (0.5–200) 20–400 (10–400) 5–800 (5–800) 10–400 (1–200) 80–800 (40–800) 10–800 (0.5–200) 5–400 (0.5–400) 10–800 (0.5–200) 20-200 (2–800)

Linear Range (μg/L) 0.992 (0.996) 0.997 (0.997) — 0.997 (0.998) 0.998 (0.992) 0.984 (0.999) 0.991 (0.997) 0.996 (0.996) 0.981 (0.995) 0.998 (0.993) 0.996 (0.993) 0.997 (0.999) 0.995 (0.998)

r2 5.4 (16.2) 3.9 (7.3) — 4.5 (5.7) 0.8 (2.3) 3.2 (5.1) 2.7 (7.3) 5.6 (8.7) 11.2 (13.3) 2.2 (5.5) 4.3 (8.3) 5.7 (11.6) 12.3 (15.2)

RSD (%) 1.0 (0.3) 6.0 (1.0) — 3.0 (0.2) 0.3 (0.1) 1.5 (3.0) 1.0 (1.5) 2.0 (0.4) 25.0 (15.0) 1.0 (0.1) 1.0 (0.1) 1.0 (0.1) 5.0 (0.6)

LOD (μg/L–1)

4.0 (1.0) 20.0 (5.0) — 10.0 (0.6) 1.0 (0.5) 5.0 (10.0) 5.0 (5.0) 6.0 (1.0) 80.0 (40.0) 4.0 (0.5) 5.0 (0.5) 4.0 (0.5) 20.0 (2.0)

LOQ (μg/L–1)

Source: Reproduced with permission from Vidal et al., J. Chromatogr. A, 867, 235, 2000.

264

Note: Calibration data in GC/MS obtained using relative heights to that of IS, except for dichloran and pyrazophos, and using relative areas in GC/MSMS.

RTW (min)

Pesticide

TABLE 7.9 Retention Time Windows (RTWs) and Calibration Data (n = 8) of the GC/MS and GC/MSMS Methods

DK4478_C007.fm Page 264 Friday, September 22, 2006 8:41 AM

Advances in Chromatography, Volume 45

DK4478_C007.fm Page 265 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

265

to animals and humans. Many of the different commercially available derivatizing reagents were used according to their manufacturer’s recommendation. Their Table I (not reproduced in this review; see the original publication) summarizes the derivatizing conditions and molar response ratios to the reference standard 5αcholestane for the 13 known phytohormone standards selected to cover most phytohormone classes. All were derivatized with the 15 different derivatizing reagents. The authors noted that while no derivatization strategy allowed for analysis of all 13 of the phytohormone standards, the combination of silylation and methylation afforded by MTBSTFA provided for the highest sensitivity, was less prone to the formation of by-products, and gave good resistances to up to 2% water. LODs (s/n = 5) ranged from 0.01 to 1.0 pmol among the different compounds because there was so much variation in the chemical groups and structures of the various phytohormone standards. The authors used derivatization with MTBSTFA followed by GC/MS to successfully different phytohormones from tobacco roots and seedlings.

V. SPECIAL GAS CHROMATOGRAPHIC METHODS: ONLINE SPE CONCENTRATION, SOLID PHASE MICROEXTRACTION, AND STIR BAR EXTRACTION In 2003 Brossa et al. [31] reported finding several specific compounds, known to be EDCs, in irrigation ditch water samples using an automated, online, programmed temperature vaporizer (PTV) injection GC/MS method. In what was documented to be a fairly elaborate analytical setup, a series of three six-port valves and a supplementary HPLC were used to deliver 7.5 ml of various types of water samples to a short, 10 mm × 2 mm SPE column packed with styrene-divinylbenzene copolymer. The sorbed analytes were back-flushed using small volumes of ethyl acetate from a syringe pump onto a short Tenax column cleverly packed into the splitless PTV liner of the gas chromatograph’s injection port. The trapped analytes were then quickly thermally desorb separated and analyzed by the GC/MS. The authors reported that their new online method, in comparison to their previously reported conventional off-line SPE preconcentration GC/MS method, provided for greater recoveries but similar results. In general, LODs (s/n = 3) for about 10 of the most commonly used pesticides ranged from 1 to 30 ng/L and had RSDs between 1% and 8% for hexachlorobenzene (HCB), p,p′-DDT, and p,p′-DDE spiked at levels of 50 ng/L in laboratory water. In all water samples, including laboratory water (Milli-Q, Millipore, Billerica, MA), various phthalates (butylbenzyl phthalate [BBP], and di(2-ethyl hexyl) phthalate [DEHP]) and an adipate were detected. It was only in the irrigation water samples collected in the spring, and not in tap, sea, or waste water samples, that atrazine (at 0.16 μg/L) and p,p′-DDE (at 0.040 μg/L) were found. In 2004 Wang et al. [32] reported developing an online, in-tube solid phase microextraction (SPME)/GC instrument for the trace analyses of various groups of organic compounds from water samples. They used a six-port valve to siphon in the water sample by back-pressure from a miniature water circulating pump onto a

DK4478_C007.fm Page 266 Friday, September 22, 2006 8:41 AM

266

Advances in Chromatography, Volume 45

conventional capillary column, 5 m × 0.53 mm inside diameter, 1.2 μm film of polydimethylsiloxane (PDMS), which was their online, in-tube concentrator. After the organic compounds were sorbed to the PDMS film, the valve was switched and the compounds ballistically heated at a rate of 290°C/min up to 320°C with the desorbed compounds directed via a capillary transfer line onto an on-column injector of a gas chromatograph equipped with different detectors. Using a series of alkanes, n-C10 to n-C19, as test compounds, the authors reported up to a 30-fold concentration increase with the online tube configuration in comparison to normal fiber SPME desorption. Wang et al. reported that the GC baseline for the in-tube configuration was considerably smoother and that detection limits (s/n = 3) ranged from 10 to 70 ng/L for the n-C13 to n-C19 compounds. The method was reported to be linear for three orders of magnitude and precisions from 5.8 to 9.2% RSD were reported at the 500 ng/L level for the n-C13 to n-C19 compounds. In addition, Wang et. al. reported the following applications. An analysis of drinking water spiked with a mixture of polycyclic aromatic hydrocarbons (PAHs), when analyzed by their intube concentrator and GC/FID, gave detection limits of approximately 10 ng/L. A mixture of chlorinated pesticides, spiked into a drinking water sample, was reported to have detection limits of approximately 1 ng/L with a 5 min extraction of a 30 ml aqueous sample by GC/ECD. A mixture of phosphorus-containing pesticides was spiked into a drinking water sample and a GC-pulsed flame photometric detector (FPD) operated in the phosphorus mode produced an estimated detection limit of approximately 50 ng/L for most of the compounds, with an in-tube extraction time of only 5 min. Solid phase microextraction, which some have criticized as not being completely quantitative and certainly not allowing for 100% recovery from a liquid or solid matrix, has been shown to avoid much of the labor-intensive sample preparation. In 2002 Diaz et al. [33] made a compelling case by showing that results obtained by SPME-GC/MS compared well to similar results by the more conventional SPEGC/MS. They reported developing simultaneous analysis from a water matrix of various NPs, nonylphenol mono- and di-ethoxylates (NP1EO and NP2EO), and their known acidic metabolites, nonylphenoxy mono- and diethoxyacetic acid (NP1EC and NP2EC). They found that the addition of this derivatizing agent in excess to the aqueous solution gave better recoveries of their analytes than presaturating the SPME fiber with the derivatizing reagent and then introducing the modified SPME fiber into the headspace or the sample’s solution. In the former case, the chromatographic profiles gave only the desired derivatized peaks, whereas in the latter case, a combination of the desirable derivatized, but also the underivatized product peaks were noted. Of the various derivatizing agents evaluated, dimethyl sulfate in the presence of sodium hydroxide gave the highest conversion of the respective methyl esters. These authors reported comparing the recoveries of their five target analytes using seven different commercial SPME fibers. Different SPME variables (e.g., derivatizing/extraction temperatures and times, reagent concentrations, and matrix salt concentrations) were optimized. Tables 7.10 and 7.11 summarize their results. In 2002 Liu et al. [34] reported dipping an SPME fiber into a methanolic solution containing a micromolar mixture of various long chain, saturated and unsaturated fatty acids, and C10 and C18 acids to which an excess of different quaternary

DK4478_C007.fm Page 267 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

267

TABLE 7.10 Results Using the In-Sample Derivatization Headspace (HS)-SPME-GC/MS

Compound NP NP1EO NP1EC NP2EO NP2EC

Linear Range Studied (μg/L)

LODa (μg/L)

0.06–3.5 0.50–4.5 0.9–6.0 1.2–6.0 4.0–14

0.02 0.20 0.30 0.4 1.5

Precisionb (RSDs %, no. of trials) Run-to-Run 8 8 7 12 18

(5) (3) (5) (3) (5)

Day-to-Day 12 15 16 22 25

(10) (13) (7) (20) (19)

a

LOD defined as concentration that gives three times the standard deviation of the blank. b Determined by using spiked laboratory water at NP 0.08 μg/L, NP1EO and NP2EO 2.1 μg/L, and NP1EC and NP2EC 4.2 μg/L, n = 3 and n = 3 replicated over 3 days.

TABLE 7.11 Comparison of the Results of the HS-SPME-GC/MS and Conventional SPE-GC/MS Methods Using Actual River Water, Quantitated Using the Following as Internal Standards: n-NP, n-NP1EO, and n-Nonyloxybenzoic Acid HS-SPMEGC/MS

SPE-GC/MS

Significance Level

Compound

Mean (μg/L)

% RSD (n = 3)

Mean (μg/L)

% RSD (n = 3)

p-value; Student’s t-test

NP NP1EO NP1EC NP2EO NP2EC

1.8 0.7 2.5 1.0 11

13 17 11 19 25

1.6 0.8 3.1 0.8 9

6 5 16 15 7

0.065 0.153 0.072 0.141 0.141

ammonium hydroxide derivatizing agents were added. Phenyl trimethyl ammonium hydroxide (PTMAH) was used to cause the transesterification of the acid group, but should work as well for a HO- or a HN- group. After a suitable extraction time in the stirred methanolic solution, the SPME fiber was transferred into the heated gas chromatograph injection port to cause pyrolytic methylation (5 to 10 min) and thermal desorption of the derivatized products. Linearity of the C18:0 fatty acid was from 3.3 × 10–6 to 3.3 × 10–4 M, with a detection limit of 1 μM. Excellent separation of the derivatized peaks for C18:1, C18:2, and C18:3 fatty acids were obtained. The use of the PTMAH derivatization and online thermal demethylation of the fatty acids in olive oil was demonstrated.

DK4478_C007.fm Page 268 Friday, September 22, 2006 8:41 AM

268

Advances in Chromatography, Volume 45

One of the important new developments in GC and GC/MS analyses of EDCs has been the development and commercial availability of stir bar sorptive extraction (SBSE) and thermal desorption of targeted components pioneered by P. Sandra. SBSE avoids a considerable number of the tedious sample preparations steps and provides for quantitative thermal desorption of the analytes into the GC/MS instrumentation. In its simplest form, an aqueous sample (water, wastewater, even urine or blood samples; about 5 ml) is placed in a 20 ml sealed vial to which is added a stir bar coated with a thick layer (24 μl) of PDMS gum (Twister; Gerstel, Müllmeim a/d Ruhr, Germany). For analyte derivatization, the Twister stir bar can be easily preimmersed into an appropriate derivatizing reagent before being exposed to the sample. After a predetermined stirring rate and stirring time (e.g., 1000 rpm for 60 min), the stir bar is taken out of the vial with tweezers, or better yet robotically, gently rinsed or wiped to removed excess sample or reagent, and then placed into a special glass thermal desorption tube (TDS-2, thermal desorption unit; Gerstel) mounted on an appropriated GC/MS instrument. In a 2003 article, Tienpont et al. [35] reported the profiling of many different compounds, including barbiturates and certain pharmaceutical drugs in urine. In situ derivatizations with either acetic anhydride or ethyl chloroformate allowed for enhanced extraction efficiency to the PDMS from the aqueous phase and subsequent better thermal desorption and gas chromatographic and mass spectral responses. They reported that barbiturates could be quantified in the concentration range between 5 and 500 μg/L, with LODs of approximately 1 μg/L in the MS ion extraction mode and 10 ng/L in the selected ion monitoring mode. In 2004 Serôdio et al. [36] reported using SBSE followed by organic solvent desorption using acetonitrile, then solvent blow-down and reextraction and then large volume (20 μl) injection GC/MS to screen for more than 60 EDCs including those from the following classes: alkylphenols, biocides, herbicides, organochlorine and organophosphorus pesticides, PAHs, PCBs, and phthalates from water samples. All of the above conditions were optimized, and high recoveries and excellent linearities were obtained in the 25 to 400 ng/L range. The authors noted that their method provided a suitable protocol for screening for a large number (more than 60 different organic compounds of various chemical groups) in drinking water samples that would comply with the European Union directive on water quality. Also in 2004, Kawaguchi et al. [37] reported the trace analysis of various phenolic EDCs in fresh and wastewater treatment samples using the SBSE technique with in situ derivatization with acetic anhydride, thermal desorption, and GC/MS analyses. No sample preparation was done to the 10 ml water samples. Table 7.12 summarizes some of their results.

VI. NEWER DEVELOPMENTS IN CHROMATOGRAPHIC METHODS Comprehensive two-dimensional gas chromatography (GCxGC), especially when it is interfaced to a rapid scanning time-of-flight (TOF) mass spectrometer, offers the possibility of extremely high resolution and identification of closely eluting GC peaks

DK4478_C007.fm Page 269 Friday, September 22, 2006 8:41 AM

Endocrine Disrupting Chemicals

269

TABLE 7.12 Summary of Analysis Results for the Stir Bar Sorptive Extraction (SBSE) with In Situ Derivatization and Thermal Desorption (TD) then GC/MS

Compound

LOD (ng/L) (s/n > 3)

LOQ (ng/L) (s/n > 10)

Percent Recovery (Spiked at 100 ng/L)

Percent RSD of Recovery

4-t-butyl phenol (BP) 4-t-octylphenol (OP) 4-nonylphenol (NP) Pentachlorophenol (PCP) Bisphenol A (BPA)

2 0.5 5 2 2

10 2 20 10 10

102.2 93.9 113.0 107.8 103.0

7.2 6.1 5.9 6.0 5.3

in complex environmental samples. Employing the pioneering concepts of J.B. Phillips [38], a slower, longer, first-dimension column is attached by a multistage thermal modulator to a faster, shorter, second-dimension column. The thermal modulator chops effluent from the first column into a series of focused plugs, then transfers and thermally launches them into the second column. A typical GCxGC configuration uses a nonpolar stationary phase on the first column and a polar stationary phase on the second column. This configuration produces a two-dimensional retention time plane where the first-dimension separation is primarily by carbon number and the second-dimension separation is by molecular polarity. Compounds that coelute on the first column are often separated on the second column. Because the separation mechanisms of the two dimensions operate independently of each other, the total peak capacity in the two-dimensional plane is the product of the peak capacity of each single GC dimensional plane. This is often enough to separate components of interest from complex matrices found in environmental samples. GCxGC separations of a variety of chemical contaminants such as PCBs, PAHs, and a wide variety of petroleum types and biomarkers have been published by Gaines, Frysinger, and others [39,40]. Figure 7.6 provides a comparison of the higher peak capacity with significantly better resolution when the same mixture of PCBs and isomers are separated, the top panel being an overlay of two single-dimension GC/electron capture chromatograms of PCBs and PBDEs and the bottom panel being a single two-dimensional chromatogram of the same PCB and PBDE mixtures.

DK4478_C007.fm Page 270 Friday, September 22, 2006 8:41 AM

270

Advances in Chromatography, Volume 45

500 450 400

PCBs

350 300 250

PBDEs

200 150 100 50 0 1200

1400

1600

1800

2000 Time (s)

S1 PBDE Mix No Mod: 1

2200

2400

2600

2800

S1 Aroclor Mix No Mod: 1

Signals: S1 5.5 >

4.5

3.5

>

2.5