8,505 1,670 51MB
Pages 900 Page size 432 x 648 pts Year 2009
Fats and Oils Handbook
Michael Bockisch Hamburg, Germany
Champaign, Illinois
This book is dedicated to my wife Gudrun to whom, in the course of doing this translation, revision, and update, I had to break my promise never to write a book again, and also to my son Benjamin and my daughter Valerie. AOCS Mission Statement To be a forum for the exchange of ideas, information, and experience among those with a professional interest in the science and technology of fats, oils, and related substances in ways that promote personal excellence and provide high standards of quality. AOCS Books and Special Publications Committee E. Perkins, chairperson, University of Illinois, Urbana, Illinois J. Endres, Fort Wayne, Indiana N.A.M. Eskin, University of Manitoba, Winnipeg, Manitoba T. Foglia, USDA-ERRC, Wyndmoor, Pennsylvania L. Johnson, Iowa State University, Ames, Iowa Howard R. Knapp, University of Iowa, Iowa City, Iowa J. Lynn, Edgewater, New Jersey M. Mathias, USDA-CSREES, Washington, D.C. M. Mossoba, Food and Drug Administration, Washington, D.C. G. Nelson, Western Regional Research Center, San Francisco, California F. Orthoefer, Monsanto Co., St. Louis, Missouri M. Pulliam, C&T Quincy Foods, Quincy, Illinois J. Rattray, University of Guelph, Guelph, Ontario A. Sinclair, Royal Melbourne Institute of Technology, Melbourne, Australia G. Szajer, Akzo Chemicals, Dobbs Ferry, New York B. Szuhaj, Central Soya Co., Inc., Fort Wayne, Indiana L. Witting, State College, Pennsylvania Copyright 0 1998 by AOCS Press. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means without written permission of the publisher. The paper used in this book is acid-free and falls within the guidelines established to ensure permanence and durability.
Library of Congress Cataloging-in-PublicationData Bockisch, Michael. [Nahrungsfette und ole. English] p. cm. Updated and revised translation of the original German work, Nahrungsfette und ole. Includes bibliographical references and index. ISBN 0-935315-82-9 (alk paper) 1. Oils and fats, Edible-Handbooks, manuals, etc. I. Title, TP670.B5713 1998 6 6 5 4 ~ 12 Printed in the United States of America with vegetable oil-based inks. 543 06 05 04
98- 11974 CIP
Preface Oils and fats have been constituents of human nutrition from ancient times. First, they contain the highest level of energy of all components of food; second, they supply essential elements for the body. However, the fundamental reason for their early and varied use was certainly the fact that they contribute to the development of flavor, making dishes tasty and giving them a good, smooth mouth-feel. In the course of his life, a human being living in the industrial nations consumes approximately three tons of fats and oiIs; about half of them are so-called invisible fats, hidden in other food, e.g., sausage and cheese. In the developing countries, the portion of fat in food intake lies far below the amount recommended by the World Health Organization (WHO). This is not due to limited world resources, but rather to problems of local purchasing power as well as logistics and distribution. Cultivation could easily rise considerably faster than the world population. Fats are an important factor in the economy because of their status as basic constituents of nutrition and the large amounts consumed. This becomes apparent in their ranking second in worldwide traded items. For some countries or regions, they represent an indispensable part of the gross national product and a source of foreign currency. Because of the development of new varieties of oil seeds, the areas under cultivation have been extended in the past decades from the earth’s sun belt to regions of temperate climate, which now deliver a substantial part of that crop. This has led to a shift in economic interests. ‘some of these factors will be discussed in this book to illustrate the large correlations that exist. This book will acknowledge the importance of fats and oils and give a survey of today’s state of the art in technology. Even when considering a topic that is relatively limited in scope, it is impossible for a single person to obtain more than a general view of the field. It follows that to give an adequate description, it is vital to have recourse to the knowledge base established over the years by the publications of many scientists and practitioners. I would like to express my thanks to all those colleagues whose findings and experiences formed a basis for my studies. Technology is not an end in itself. It is justified when it makes it possible to improve the foodstuff offered by nature in whatever way, whether in amount, cost, quality or other criteria that make possible its use or adaptation to our way of life today. To pursue food technology without knowing the “raw material” would mean working in a vacuum and performing “l’arrpour I’m.” In this book, great attention was thus paid to describing the sources of the oils and fats, and also the fats and oils themselves in such a way that the technological steps be well-founded and the purpose clear. Since the industrial revolution, there has been a great boom in the industry of edible fats and oils. The processes that remain in use today are founded on basic findings from approximately 100 years ago. Much has been improved since that time, and many facts already known are exploited today only on the basis of improved technology. Technology will continue to develop in parallel with man’s changing attitude to it. In the broadest sense, progress in biological science and biotechnology contributes to this. Because biotechnological processes are reputed to be more natural V
vi
Preface
than chemical ones, there is an attempt to use enzymic reactions for fat technology. Much more far-reaching effects can be achieved by cultivation or the application of gene technology. Plants containing fats and oils in a desired composition, structure and quality render superfluous certain steps of treatment or modifications. To date, agriculture has not yet fully realized the potential offered by the cultivation of tailormade plants for certain purposes, as opposed to mass cultivation. In addition to these aspects, the development of machines and equipment leading to a more responsible way of dealing with raw materials and the environment continues. New processes of refining are confined mainly to physical modes of operation and protection of energy and water resources, in keeping with the spirit of the period and the desire to keep costs low. The manufacturers of such equipment are constantly engaged in new and further developments. Here, I would like to thank the companies that are mentioned subsequent to the bibliography for their support in providing pictures and information. This book will survey the raw materials predominantly employed and the spectrum of processes used today. Man’s ability to absorb information visually, i.e., via pictures, is many times greater than through the other senses or through transposition from language. A diagram or figure conveys more than a thousand words. To impart information quickly and efficiently, a focus of importance in this book was the explanation of technological steps in the form of graphs, a form of presentation that offers the reader a quick orientation and conveys a general view. Sufficient detail is offered to highlight the critical points without obscuring the presentation of essential information. In that sense, this book can be considered a sort of picture-bookhopefully, in a good sense. Michael Bockisch Vienna, I993
Preface to the English Edition AOCS Press has decided to publish this (originally German) book in English, updated and revised, and I am very grateful for the opportunity to expand its distribution and readership. The book was written primarily for Europe, and especially for Germany; some parts of Chapter 1, in particular, focus on the home situation. These parts have been changed where possible to give a broader picture. However, some figures remain in their original form, describing the situation in Germany. This was the case whenever they were too specific to be changed or when they illustrated certain facts that may well be used as an example for other regions in the world or as representative for the European Union. It is an honor to be able to reach a much wider readership. I translated the book to the best of my ability; however, without the help of my son, Benjamin, who had to do a lot of proofreading, of Ralf Tonn, who assisted me with the translation of Chapter 1, and especially of Iain Gow, Greg Knoll and Michael Gude, who did their very best as co-readers to improve my style, I (or you as the readers) would have been worse off. Both my readers and I owe them our thanks. In the interim between the original German issue and this one, almost half a decade has passed, and some of the trends that could be seen on the horizon have intensified. The skepticism towards any form of technology has increased in some countries, especially Scandinavia and the German-speaking countries, coming very close to hostility at least in some parts of the population. This deepens the gap between the wealthy countries that can afford to reject useful technology and those parts of the world in which technology is urgently needed to feed the increasing and often poor population. It seems that part of today’s fat technology will disappear in Europe or that chemical processes will be replaced by physical ones regarded as more environmentally friendly or by enzymic ones regarded as biological and thus, natural. On the other hand, there is total rejection in some quarters of new technologies such as biotechnology whether the concern is enzymes and their methods of production or genetically modified organisms (GMO). The contradiction between the existence of less technology and the development of new plants that may save some processing steps will be difficult to resolve. The next two to three years will determine where these new developments will be accepted and where they will not. A start was made in 1996 with the introduction of GMO soybeans, with other modified oilseeds following mainly in 1998. Lastly, I hope my readers will follow the advice of the great German poet Wolfgang von Goethe, who said: “Also, we should not deny that we are willing to forgive one or the other typing error in a book because we feel flattered by the fact that we detected it.”
Michael Bockisch Hamburg, 1997
vii
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v
Chapter 1
The Importance of Fats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. A History of the Production of Oil and Fat 1.2. Fat in Food and Food’s Raw Materials 1.3. The Economic Importance of Oils and Fats 1.4 Fat in Nutrition 1.5. Fats and Oils in Legislation 1.6. Fats as Technical Raw Materials 1.7. Fats and Oils as a Source of Energy 1.8. New Sources of Raw Materials 1.9. Substitutes for Fat 1.10. References
1
Chapter 2
Composition, Structure, Physical Data, and Chemical Reactions of Fats and Oils and Their Associates . . . . . . . . . . . . . . . .53 2.1. Components of Fats and Oils 2.2. The Structure of Triglycerides 2.3. Physical Characteristics 2.4. Chemical Reactions 2.5. Lipids 2.6. References
Chapter 3
Animal Fats and Oils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Milk Fats 3.2. Rendering Fats 3.3. Marine Oils 3.4. References
121
Chapter 4
Vegetable Fats and Oils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. OiVFat-Containing Plants 4.2. Pulp Oils 4.3. Seed Oils 4.4. Nonedible Oils and Fats 4.5. Other Oil Sources 4.6. References
174
Chapter 5
Production of Vegetable Oils and Fats 5.1. Pulp Oils 5.2. Seed Oils and Fats
......................
345
Chapter 6
Modification of Fats and Oils . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1, Application and Combination of Modification Processes 6.2. Fractionation 6.3. Winterization 6.4. Interesterification 6.5. Hardening 6.6. References
446
ix
X
Contents
Chapter 7
Refining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1, Economic Importance of Refining 7.2. Neutralization 7.3. Bleaching 7.4. Deodorization 7.5. Physical Refining 7.6. Energy Consumption and Investment 7.7. Importance of Refining for Removal of Environmental Contaminants 7.8. References
Chapter 8
Fat as, or in, Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. Butter 8.2. Margarine 8.3. Edible Fats 8.4. Salad and Frying Oils 8.5. Mayonnaise 8.6, Vegetable Creams, Cream Substitutes 8.7. Peanut Butter 8.8. Margarine and Oils with Medium-Chain Triglycerides 8.9. Monoglycerides and Diglycerides 8.10. References
719
Chapter 9
Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1, Acid Number 9.2. Saponification Number 9.3. Iodine Value 9.4. Peroxide Value 9.5. Unsaponifiable Matter 9.6. Water Content 9.7. Phosphorus Content 9.8. Color 9.9. Hexane in Extraction Meal 9.10. Fibers in Extraction Meal 9.1 1. Protein in Extraction Meals 9.12. Ash Content 9.13. Solid Fat Content 9.14. Dilatation 9.15. Lipids Analysis
803
Chapter 10
Conversion Tables and Abbreviations. . . . . . . . . . . . . . . . . . . . . . .
809
Chapter 11
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
813
Chapter 12
Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
815
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
819
.613
Chapter 1
The Importance of Fats The importance of fats for humans, animals and plants lies in their high content of energy, which permits the greatest possible storage of energy in the smallest possible amount of food substance. In addition, fats allow humans and animals to consume fat-soluble vitamins and provide them with essential fatty acids, that is, those indispensable fatty acids that their bodies are unable to synthesize themselves. Fats are omnipresent in nature, although in the most diverse quantities. In the human body, they play a decisive role as well, beginning with the nutrition of the infant with breast milk. During the first 5 d, breast milk contains an average of 29.5% fat; from d 6 through 10, the amount is 35.2% and later 45.4% (Macy 1949). In the course of life, a human living in the industrial world satisfies an average of >40% of energy demand with fat. Metabolized in the human body, fats yield 38 kj/g of energy (9 kcal/g). In this exothermic reaction, -2000 mL of oxygen per gram of fat is consumed and -1400 mL of carbon dioxide is produced (Peters and van Slyke 1946). In addition to -63 ton of water, 0.5 ton of alcohol, 8 ton of carbohydrates and 2 ton of proteins, humans consume -3 ton of fat during their lives. The efficiency of fat as foodstuff is very high, because the fat contained in food is almost completely reabsorbed by the body; in the feces (in the course of one’s life -5 ton, plus 30 ton of urine) only 3.3% of lipids can be found (Pimparkar 1961). Thus, fats play an indispensable part in nutrition as supplier of energy, source of compounds that the body cannot synthesize by itself, and carrier of vital substances. Fats cannot be replaced by other substances. Apart from this physiological aspect, they are excellent carriers of flavors, and dishes prepared with fats are much tastier than others. Fats also provide a smooth, creamy consistency to many dishes, which translates into a good mouth-feel. This explains in part why the consumption of fat is still very high today, even though the segment of the population performing hard labor has diminished greatly compared with the past, rendering a very high supply of calories no longer necessary (see also Chapter 1.4). The improvement in flavor, in particular, is certainly the reason why fats and oils have been appreciated for a long time. However, only since the beginning of the present century has it been possible in the industrial nations to provide the population with sufficient quantities of fat at reasonable prices (see Chapter 1.3). Because of this increasing importance of fats and oils, governments have intervened to a great extent in their production and distribution in the last 100 years (see Chapter 1.5). European food legislation, in particular, has often been marked by protectionist objectives. The importance of fats and oils to the global economy (Chapter 1.3) becomes clear when considering the amount of oilseed and fruit grown worldwide. In 1995, -60 million ton of palm fruit and -1 1 million ton of olives as well as >200 million ton of oilseeds were harvested. From these amounts, >90 million ton of oils and fats were derived. Many countries are trying to enlarge their shares in the interna-
2
Fats and Oils Handbook
tional market, a strategy that is usually to the disadvmtage of others and leads to defensive measures. National interests play a part here. For example, 10 years ago, the European Union began to promote the cultivation of sunflowers and rape in the area of the Community. A simultaneous attempt to stabilize the Community’s budget deficit by introducing a tax on fat caused the U S . to fear for its soy exports, resulting in a threat of trade obstructions aimed at the European automobile industry. The confrontation was averted in 1987, but it will reappear again and again, unless the issue of a tax on fat is buried for good. In addition to the importance of oils and fats for human nutrition, there is a substantial market for technical fats. The importance of these oils and fats will increase considerably in the future because they represent a vast potential of naturally regenerating raw materials in which the chemical and pharmaceutical industries have a special interest. A short survey of these technical fats is given in Chapter 1.6. The importance of oils and fats for human nutrition, the animal feed produced from the processing of most oil plants and the economic importance of oils and fats, i.e., the fact that many millions of people worldwide make a living by the production and processing of oils and fats, all combine to give special importance to technology. This may even be enhanced if oil-bearing crops could be offered to the chemical industry as a source of regenerating raw materials. Only -1% of -300,000 existing species of plants has thus far been examined for their qualifications as useful plants. Only 300 of these, i.e., 0.1%, are being used agriculturally today. About 7% of these, -20, are oil plants. A considerable potential, which may be suitable for the recovery of oil, thus remains. This is especially true when plants with a special fatty acid pattern are desired. In addition, there are new methods of plant breeding that also may open up new prospects.
1.1 A History of the Production of Oils and Fats From ancient times, man unconsciously consumed fat in his food via plants, fish, and meat. However, the use of oils and fats required some simple techniques. For example, only when the ability to make fm was discovered was it possible to melt the fat of hunted animals and store the fat. Storage also required the ability to produce vessels made from clay or another materials. Mutton tallow and lard, and later butter, cream, and fish oils were known in prehistoric times (Hanssen and Wendt 1965). Vegetable fats from olives and sesame seed, and possibly flax were also known. Until the previous century, the utilization of fats and oils as food went hand in hand with their use as fuel, predominantly for purposes of illumination. Even today, the name “lampante” for certain qualities of olive oil refers to this. As a base for ointments and cosmetics, they are still in use today, just as in the earliest periods. It is known from pictures that food factories were in existence in early days. Wooley (1929) presented a picture of an Egyptian dairy-farm that illustrated, among other things, the process of churning. Erman and Ranke (1923) showed the sequence of operations of a large-scale Egyptian bakery of Ramses 111 in Thebes
Importance of Fats
3
around 1200 B.C. in which a dough resembling a Chelsea bun is being baked in oil. People in the Mediterranean region and in Asia used oils long before those in Central Europe. With olives in the Mediterranean area and sesame in the area of the Euphrates and Tigris, oil plants that do not grow in the temperate latitudes were readily available. The Bible mentions oil in many passages. Moses, for example, required oil as a donation for the Tabernacle’s lamps; cake and pancakes were baked with unleavened oil (2 Moses 29). It was customary to anoint with oil, and Jacob anointed a stone this way to sanctify it. Passages in Luke indicate that oil was a valuable commodity. For instance, there is a description of how a person owed 100 barrels of oil to somebody else (Luke 16: 5.6). The importance of the olive tree to the people in the Mediterranean region also became visible by its being consecrated to the goddess of wisdom, Athene, in ancient Athens; it later became the symbol of peace and promise. Thus, after the Deluge, the dove came to Noah with an olive leaf in its beak (1 Moses 8:1l), a symbol of the world’s survival. Plinius described how to extract olive oil by pressing ripe fruit in a squeezing vise, a procedure customary in his time. Butter (“dense, solid milk foam”) was also mentioned, but only as a replacement for olive oil in times of need, or for baking. Another familiar practice of fat technology was the remelting of lard for cleaning purposes. Roman fat technology spread throughout the Mediterranean region. Excavations in Tunis show its propagation in Northern Africa; in Pompeii and Herculaneum, entire processing facilities consisting of oil mills, oil presses, oil shops, and oil depots were uncovered (UNION 1959). Because the fatty acids corroded copper, oil was transported in vessels made of lead, but also in tanker vehicles. These were carts carrying animal skins enclosed in iron hoops in which the oil was kept. Poppy seed has also been discovered in Swiss lacustrine dwellings dating from the 25th century B.C. It is probable that the inhabitants already knew poppy oil and that people north of the Alps became familiar with oil mainly through the occupation by Roman troops. According to tradition, oil was also extracted from beechnuts that were mashed, then wrapped in cloths and pressed between plates of metal or stone. Every large farm extracted its own oil. Later, agriculture made further progress, and rape and linseed were added as oil fruit. In the 16th century, the profession of the oil miller, who processed the farmers’ seeds, evolved. The oil was extracted by grinding, bruising, or pressing. Later, people shifted to squeezing vises. Windpower was used as propulsion in windmills. In the course of further development, hydraulic presses were built, and from the middle of the 19th century on, it was possible to extract oils and fats from the seeds by means of solvents. When the seafaring European nations conquered the world, the resources were expanded and unknown types of oil fruit with much higher contents of oil and fat than those previously known were brought to Europe. Until the cultivation of soybeans was widely expanded, these oilseeds yielded the substantial part of vegetable fats and oils consumed in Europe. The explosive increase in population in the industrialized countries of the world during the industrial revolution, combined
4
Fats and Oils Handbook
with urbanization, led to a new situation. The population that gathered in the cities had to be fed. This required an entirely new system of food distribution, which adapted itself to the change from domestic supply in small units (farms, villages or small towns) to industrial production. New requirements for food arose, especially concerning its preservation. New products such as margarine (see Chapter 8) and novel techniques such as hardening (see Chapter 6.4) made important contributions to mastering these challenges. Although the main concern in the late 19th and early 20th centuries was the satisfaction of basic needs, in today’s industrial society, there is no longer a problem of quantity. After the decline in food production caused by the two World Wars and the Great Depression, during which the question of “mere nourishing” became important, the essential aspect of the 1960s and 1970s was that of enjoyment. The primary function of food was no longer the supply of calories, but the experience of taste. Accordingly, the focus shifted from the production of quantities, which were in fact available, to quality. In particular, the emergence of the trend toward health consciousness has stimulated the demand for quality. Here the fat industry delivered exceptional contributions. The connection between cardiovascular disease and nutrition was detected in the early days. This knowledge was used to develop special products that provide preventive measures (e.g., becel; see Chapter 1.4). The excessive fat consumption of most of the world’s population and the consequences of overweight have led to the development of reduced-fat and very-low-fat variants of the most diverse types of food. For margarines, for example, the law had to be changed to allow this. Starting with margarines, food groups developed that made possible low-calorie nutrition, or, as in the case of becel (a trademark of Unilever concern), furnish variants of diverse foods free of or low in cholesterol and high in polyunsaturated fatty acids. In recent years, a new trend could be observed developing in parts of the population. Sensitized by a growing awareness for environmental issues, “naturalness” of products has been given high priority. Although the demands resulting from this are partly exaggerated and no longer have a clear foundation, it will lead to changes in technology in some fields, at least in the wealthier countries. In poorer countries, which strive to produce food for mere nourishment and survival, there is little patience for these trends. The priority here is to feed the population. Although the strong increase in population is no doubt leading to problems that cannot be overestimated, the production of fats has always increased at a rate higher than the population growth (see Chapter 1.3).
1.2 Fat in Food and Food Ingredients Fats and oils, as such, are used for the production of food; in addition, unprocessed food or ingredients for food production contain fats and oil, sometimes in substantial quantities. These include fruits and vegetables but also meat and fish (Table 1.1). For oil and fat technology, these figures are important because they help identify the raw materials for oil and fat production. In processed food, oils and fats are
lmportance of Fats
5
TABLE 1.1 Approximate Fat Content of Unprocessed Food FruiVNuts
(O/O)
Bananas Oranges Pears Sw. chestnut Avocado Coconuts' Almonds' Hazelnuts' Walnuts' Pecan nuts*
0.2 0.2 0.4 2.0 17 34 54 62 64 71
VegetabledCereals
Fish
(%)
Meat
(%)
Potatoes Tomatoes Broccoli Cabbage Corn Barley Mushroomst Soybeans
Haddock Cod Mackerel Caviar Herring
0.1 0.3 11.5 15 19 26
Chic ken breast Roast chicken Boiling fowl
1 6 20
0.1 0.2 0.3 0.8 1 .o 1.5 3 20.9
Miscellaneous
Eggs Milk
Eel Shrimp Lobster Mussels
1.5 2 1.3
11.5 3.7
Rabbit
8
Fillet of beef Haunch of beef Beef tenderloin
4.5 7 10
Fillet of pork Pork chop Bacon (lean) (fat)
4.5 7.5 65 88
'kernels. +dried.
also found in diverse quantities. The variety of foods listed below (Table 1.2) ranges from those which are produced by a single treatment step (e.g., by grinding) to those that require many steps of processing because they are composed of many kinds of raw material (Table 1.2). The figures cited can vary considerably and thus serve only as a basis for the order of magnitude of the fat content of the food.
1.3 The Economic lmportance of Oils and Fats Nourishment is an indispensable and essential need of the human race. It is the task of agriculture and the food industry to satisfy this need. Considering the current develop ment of the world's population, a huge demand is arising. This demand has to be satisfied under the best possible conditions. There remains a shortage of supply in many parts of the world. Considering the fat-providing potential of agriculture and industry, this demand should be able to be satisfied easily. In past years, the production of fat has grown at a faster pace than the population (Fig. 1.1). A shortage of supply is thus rather a problem of distribution and purchasing power than one of shortage in the sense of lacking potential. In the coming years, the production of fat will also grow more rapidly than the world population. Mielke (1985) estimated a 1Cfold increase in oil production from the year 1958 to the year 2000. During this period, the amount of soybean oil will increase sevenfold and palm oil production will increase 2.5-fold. For palm oil, this prediction is already far outdated. In the past 10 years, the volume of oilseeds stored as surplus amounted to an average of 14% of production (Mielke 1990).However, as can be seen from Figure 1.2, there is a strong correlation between gross national product, reflecting the standard of living, and fat consumption. Chma, India, Pakistan and Bangladesh currently represent almost one half of the world's pop ulation. Demand for fat by these four countries will increase not only as their popula-
Fats and Oils Handbook
6
TABLE 1.2 Aproximate Fat Content of Processed Fooda Sausages
(YO)
Corned beef Wieners Bavarian veal sausages Cervelat sausage German Fleischwurst Mortadella Calf liver pate Salami Smoked s. spread
6 20 22 35 30 33 35 38 37
Baked products Rolls White bread Crisp bread Zwieback Rich tea biscuit
0.5 1.2 1.5 4 11
Spaghetti . _
1.2
sweets Sugar Caramel Cocoa (slightly defatted) Milk chocolate
Milk products
0 10 25
Yogurt, skim Condensed milk Cream (whippable)
0.1 4-8 27-33
33
Quark Fresh cheese Camembert cheese Edam cheese Parmesan cheese Emmenthal cheese Roquefort cheese Processed cheese
5-1 2 5-1 2 23 24 26 30 32 2 3-2 8
Cereal products Oat flakes Wheat bran Wheat flour Rye flour Corn flour Corn semolina Corn flakes Pop corn
6.5 10
1-2 1 -2 2.6 1.1 0.4 5
Snacks Potato chips Roasted peanuts
(W
(Oh)
40 48
Miscellaneous ~~~~
~
Mustard Salad mayonnaise Mayonnaise Margarine Butter Cod-liver oil
'?his analysis is from European f d ; all values are examples only and may differ from country to country according to local habits, local taste and local legislation.
World
fat production [MMT] 120 100
population [billion] 12
'
8060 -
40 -. World population
__
20 01936
194S
1956
1966
197s
1986
~
1 ' 1996
Fig. 1.1. Growth of world population and oilhat production.
2
0
6 52 82.5 80.5 82.5 99.8
7
Importance of Fats _.
Germany
us.
\
Gross National Product [lo00US$ per capita] Fig. 1.2. Average spending power and fat consumption in different countries (adapted from Leonhard 1995).
tions grow, but it can be predicted that a tremendous demand for fats and oils will be triggered as their gross national product improves. Statistics concerning the world production of oils and fats have existed only since 1942. Estimations at that time, as well as those concerning the years before, were coordinated by the International Agrarian Institute in Rome (today FAO; Schiittauf 1942). Since then, there has been a shift from animal to vegetable fats (Fig. 1.3). As for all food, consumer prices for oils and fats have dropped signfi-
Marine oils Lard, Tallow Butter Pulp oils Seed oils
Fig. 1.3. Proportion of different oil sources on the world fat production.
8
Fats and Oils Handbook
TABLE 1.3 Oil Seed and Oil Fruit Production in the World by Region (O/O of total) Region North America South America China India, Pakistan, Bangladesh USSR (successor states) Europe (without USSR) Africa Malaysia, Philippines, Indonesia All others
1935
1950
1960
1970
1980
1990
1995
17 7 15 11 9 21 7 6 7
24 8 13 9 8 18 8 7 5
24 9 9 9 9 19 9 6 6
27
28 11 7 8 1 15 6 9 6
26 12 9 6 1 0 9 15 5 13 5
24 13 12 1 4 14 4 14 4
10
7 8 1
2 17 7 5 7
cantly in relation to income. Rendered animal fats, and especially butter, were goods in short supply and thus very expensive. For example, in Germany in 1857, -15% of one’s daily food expense was needed to buy only 40 g of butter. In 1800, the entire expenditure for food amounted to -70% of a family’s entire income, in 1900 to -50% and in 1974 to -28%. Today in Western Europe, this fraction is clearly ~ 2 5 %(Gander 1984). The importance of oils and fats has increased during the past 100 years. Since the time before World War I, worldwide production has more than quintupled. The center of oil fruit cultivation was once situated in tropic regions, whereas today it is in temperate latitudes (Table 1.3). To counteract this trend, tropical countries, in particular, Malaysia, Indonesia and the Philippines strive hard to regain their former rank on the scale of producers. On the one hand, it allows them to feed their own growing population; on the other hand, it creates export possibilities and explains why the production of palm oil in these countries has been extended steadily and forcefully for about 15 years. The higher oil yields of cloned palms contributed to this in part. In Malaysia, production rose from 4 million ton in 1985, to -7 million ton in 1990 and to -10 million ton in 1995 (see Chapter 4.2.1). In that country alone, almost two million people make a living from palm oil, and in 1986, this branch of the industry yielded a revenue of almost $1.4 billion (US.). At the same time, these efforts are a prime example of the importance of oils and fats to certain regions of the world. Often, one region’s efforts to raise production are accompanied by countermeasures from another region. In the case of Malaysia, the countermeasure was an attempt to discredit tropic oils. The increase in the production of palm oil was regarded by the U.S. as a danger to its soybean oil market. Campaigns were started with the objective of pushing the tropic oils out of the country as completely as possible. In addition, new directives concerning labeling were passed, forcing the producers to declare the oils used. Public health policy was given as a reason; later, this proved to be true to some extent but that explanation was viewed by Malaysia as simply a pretext. Malaysia believed that concern about losses suffered in the production of soy seed
lmportance of Fats
9
was the actual reason (Anonymous 1987). After fractionation, palm olein is a good raw material (see also Chapter 1.4), and for many products and branches of industry, stearin is also used as a consistent fat. It is expected that the U.S. import of tropic fats will level off at -750,000 ton per year. Of the laurics (coconut and palm kernel oil), 70% go to oleochemical processing. The distribution of the individual types of oil fruit has changed considerably (Fig. 1.4) as shown by the comparison of their distribution in 1935, 1960, and 1995. A projection to the year 2000 (data from Mielke 1994) is also given. Figures 1.5-1.7 show the production of oils and fats during the last 80 years. The figures refer to the amounts of visible fats produced. A distinction must be made between butter, lard, and olive oil, which represent end products suitable for consumption (Fig. l S ) , vegetable oils and fats (Fig. 1.6), which have to be processed, and oils utilized mainly in industry (Fig. 1.7). To simplify the presentation, tallow, and fish oil, which are generally processed further, are listed together with butter, lard, and olive oil which are consumed as such. Figure 1.5 shows that the production of whale oil has virtually disappeared, whereas the production of fish oil is -10 times higher now. This can be attributed
(C = Corn oil; P = Palm kernel oil; S = Sesame seed oil)
Fig. 1.4. Proportion of different vegetable oil sources on the world fat
10
Fats and Oils Handbook
Fig. 1.5. World production of butter, lard, beef tallow, fish oil, and whale oil. mainly to various new catching methods. However, a stagnation of these figures has to be anticipated, also caused by international campaigns toward sustainable fishing. The high increase in tallow production mirrors the rise in beef production. The growth rates, however, are far lower than those of vegetable fats and oils; tallow is only a by-product. The rapid rise in the production of soybeans, which has increased 50-fold, is clearly visible. The growth rate of sunflower seed is even higher. Compared with these amounts, the production of oil-bearing plants utilized mainly in industry is modest; however, the fact that a considerable amount of the edible oils is used for industrial purposes must be noted (see Chapter 1.6). By looking at the production of fat and its products, the political development in the world or in particular countries can easily be inferred. Fat production in Germany (see Chapter 1.3.1), for example, was at its lowest level during both World Wars, but also during the great depression in the 1920s and 1930s. For a good supply at low prices and, related to that, an adequate profit for the farmer, the yield per hectare is essential because in the industrial nations only a high yield per hectare promises sufficient returns. However, during times of extensive shutdowns of areas (e.g., the US.),there are considerations if models exist that are based on less intensive tillage. The yield per hectare thus keeps rising, but at a far slower pace (Fig. 1.8). It is the yields per hectare that indicate the difference between developing and industrialized countries. Cultivation in Third World countries often takes place on very small farms in poor soil with an insufficient supply of water; for particular regions and their state of affairs, this clearly makes sense. Production in the industrialized countries can vary greatly as well (Fig. 1.9). Consider this comparison. By feeding cows the yield of 1 hectare (2000-3000 kg on average), only 2W300 kg of milk fat can be produced. Of course, one must
11
Importance of Fats
Soybean oil Palm oil
Rapeseed oil Sunflower seed oil
Cottonseed oil Peanut oil Coconut oil
Sesame seed oil Olive oil
Fig. 1.6. World production of important vegetable oils.
admit that, for milk, the production of fat is not the main objective. However, byproducts are also gained from most sorts of oil fruit. Not all influences on production can be controlled even when the means are available. The source of raw materials might be exhausted without the possibility of exercising any influence on it, or the opposite may be the case. For example, the production of fish oil in Peru (one of the largest producers, and for its size, by far the largest producer) decreased from 3 11,OOO ton year in 1976 to virtually zero in 1984. Excessive fishing was not the cause, as was speculated initially, but rather a change in the direction of the Humboldt stream, which shifted its flow 350 km to
Fats and Oils Handbook
12
1.5
E r 1.0 C
5 0
ze
Linseed oil Castor bean oil
10.5
P
Tung oil Sperm oil
0 1910
1930
1960
1970
19SO
Fig. 1.7. World production of main vegetable oils for industrial usage.
the west of the Peruvian coast. The fish followed the flow of the stream. The Peruvian boats, equipped for inshore fishing only, were not able to follow the stream, and the entire industry broke down completely. Chile profited from this: its catches in 1989 amounted to -6 million ton and -260,000 ton of fish oil were derived from it.
Fig. 1.8. Production yields of some selected vegetable oils.
Importance of Fats
13
Fig. 1.9. Production yield of some oilseeds in different countries.
The opposite occurred in the insular republics of South East Asia. There, devastating typhoons had gravely affected the coconut crop over many years. Without any identifiable reason, the typhoons changed direction, and the coconut industry began prospering again. World affairs interfere considerably with the production of oil but in different ways than the natural influences listed above. Their influence on oil prices is especially evident (Figs. 1.10 and 1.11). Wars, but also energy crises are mirrored here, although the supply of edible oil has hardly any relation to the production of mineral oil. Depending on the type, the price of vegetable oils is higher or lower than that of soybean oil, whereas the market value of animal fats lies, without exception, below that of soybean oil. However, it must be considered that fish oil, with very few exceptions, is used primarily in hardened form, so that in comparison with tallow and lard, the costs for hardening have to be taken into account. Tallow is usually fractionated before use. Because soybean oil has developed into the predominant oil, the prices of the other oils are to a certain extent dependent on that of soybean oil (Figs. 1.12 and 1.13). These figures show prices leveling off in years when edible oils strongly forced their way into the market and the difference in price for soybean oil, which has been relatively constant (except for some peaks). They also give an idea of the relative value of the various oils and fats. Considering the large volumes of oil produced, the high growth rates and the competitive situation, it is no surprise that oil prices have declined relatively. Depicted in U.S. $ against inflation in the U.S., it becomes obvious that oil prices
Fats and Oils Handbook
14
L I S M 1970 1975 1980 IS85 I990 1-
I I
1065 1970 1975 1980 1985 1990 1096
1400
1200
lo00
800
600400 200 Palm oil
0 1 1966
1970
1976
l&
1986
l&
l&
Fig. 1.10. World market prices of important vegetable oils.
in January 1990, in spite of the increase of the absolute amount in constant money, are only -36% of those in January 1965. During 1987, they dropped temporarily to only 25% of 1965 levels (Fig. 1.14). Taking into account the decline of the U.S. $ in proportion to the currencies of some European countries, the oil has become even cheaper. However, this is compensated partially by lower inflation rates (Figs. 1.15 and 1.24). Because of the historical development of the production of oil fruits in regions completely different from those in which they are consumed, the market has always been international. Oilseeds and oil are usually traded in U.S. $. Because
15
Importance of Fats
1200
Soybean oil
400
1985
1400
1970 1975 lSe0 1985 1990 1995
1985 1970 1975 1980 1985
lee0 1995
World market pnce [US. WMT]
1200 lo00
800
Ittoned oil
800 400 200
Soybean oil 0 '
1
19S5
I
1970
1975
IS60
1985
ID00
1996
Fig. 1.lo. Continued.
the European countries are not self-sufficient, they have to import oilseeds. For the oil milling industry outside of the U.S., which has to commercialize oil as well as meal, the deviation of the exchange rate of the U.S. $ to the respective currency (Fig. 1.15) is added to the deviation of the price for the raw material. This can exert quite a considerable influence on the market position because products made from vegetable oils and fats always compete with indigenous products (e.g., margarine with butter). In addition, extraction meals have to compete with indigenous fodder. The influence of the deviation of the U.S. $ exchange rates against the only stable European currency (German Mark) is made clear. One realizes that the deviation for the European
Fats and Oils Handbook
16
Edible Tallow
1965
1970 1975 1980 1985 1990 1995
1W5
1970 1975 1980 1985 1 W 1995
World market price [US. WMTJ I2Oo
I
1000
800
600
400
200 -.
0 ’
I
igss
1970
197s
is80
198s
isso
,
ims
Fig. 1.1 1. World market price of some animal fats (soybean oil as comparison).
market can thereby become even larger. In times of a powerful dollar, considerable price discrepancies arise that are difficult to pass on to the customer. Prices on the world market do not always depend on supply and demand alone, but also on expectations for the volume of the next harvest. Most deals are closed on futures long before the harvest amount is known. Besides the area that is going to be cultivated in the respective year, information about the expected weather is relevant. However, not even a major regional weather-related catastrophe would be capable of influencing prices substantially. The U.S. exhibits the greatest deviations in the area used for cultivation and thus the strongest influence
Importance of Fats
440
17
World market pnce dfference to soybcen 011 [US$/MT]
300
zoo
loo 0 -100
-100
300 400 1066
1970
1976
1980
1986
1990
1996
World market pnce dfference to soybean oil [US. $IMTl
Peanut oil
400
300
'
1066
1970
1976
1980
1966
1990
1996
Fig. 1.12. World market price difference of some vegetable oils and fats to the price of soybean oil.
on the amount of crop to be expected. According to the previous year's supply and achieved prices, areas of smaller or larger size will be shut down. Because soybean oil is by far the most-produced oil, its quantity influences the prices of the other oils and fats as well. Consequently, the number of acres cultivated in the U.S. in the respective year plays a very important role in the equation. During times of high supply, prices usually drop at harvest time and rise when the provisions run short. During this process, the quality usually declines as well; therefore Brazilian and Argentinean producers generally wait until the supply of the crop in the Northern hemisphere is exhausted, or the quality declines, before they commercial-
18
Fats and Oils Handbook
Fig. 1.13. World market price difference of some animal fats to the price of soybean oil. ize their crop. This is the only way to obtain good prices for relatively smaller producers with higher transportation costs. As a result of intensified cultivation in the Southern hemisphere, fresh soybeans are available twice a year. Apart from production and availability, demand also plays a decisive role in price. When comparing demand with availability, we see that the quantities pro-
Fig. 1.14. Price of soybean oil in current and constant money (US $).
Importance of Fats
19
Fig. 1.15. Price of soybean oil in U.S. $ and in a stable European currency (DM).
duced were almost without exception higher than the demand (Fig. 1.16). This leads to an accumulation, which depresses the prices. Malaysia alone is said to have had a stock of palm oil of -1 million ton in 1988-1989. The econoinic situation in the large countries that are not self-sufficient and are subject to monetary problems is closely linked to demand, and thus to a high
Fig. 1.16. Demand and supply balance for vegetable oils and fats (after Batterby).
Fats and Oils Handbook
20
degree, decisive for price development. On the one hand, there is the Soviet Union (respectively, the successor states), where the import of meals as fodder plays an important role. On the other hand, there is India, which represents a huge market with 9300 million people. In 1986-1987, -2 million ton of oils and fats were imported. The imports in the following years are estimated to be -0.5 million ton, but the most recent estimates suggest even lower amounts; however, the demand that cannot be met that way amounts to 1 million ton per year. In contrast to this, it is expected that Pakistan will increase its imports to almost 1 million ton. In view of these markets, the surplus production is within the range of the deviations of the quantities imported by these countries. The demands are thus not regulated by the actual demands, but by the availability of funds. When we relate the prices for oilseeds on the world market to those for oil, we have to consider that they are determined by the oil content as well as by the usability of the meal as fodder. We see that an intricate web of influences exists that determines the prices. This has direct influence on the supply, more so in some parts of the industrialized countries in which the price for seeds fluctuates around the limit of profitability despite subsidies. Without subsidies, production at today’s prices would not be possible. For the oil mills, the production of soy seed, for example, means that they have the following net profits (Prices Chicago, May 1990): Cost (US. $) lo00 kg 180-200 kg 820-800 kg
soybeans oil extraction meal
Cost of beans Gross profit from oil and meal*
Revenue (U.S. $)
322.12 193.68-2 15.20 134.55-13 1.27 322.12 328.23-336.47
*without any running and processing cost, without losses and depending on the oil content.
This margin (however simplistically calculated) is extremely small for a capital-intensive industry such as oil milling, so that small changes in the meal’s marketability or in the price would render the entire enterprise uneconomical. The economic importance of a raw material, however, reveals itself not only in the supply, but also in the demand. Because oils are traded internationally, their prices are not dependent on the demand in a single country. In addition, there are subsidies for agriculture in many regions. This has repeatedly given rise to international irritations when, for example, the U.S. reproached the European Community for violating the GATT-treaty by means of subsidies and by partially barring the market for agricultural imports (see Chapter 1.3.2). From Brussels’ point of view, these subsidies are necessary to assure some degree of self-supply for certain agricultural goods of the European Community.
lmportance of Fats
21
Subsidies can have other purposes, for example, as an incentive for structural reorganization. For instance, the Community supported the shift from the cultivation of rape with a high content of erucic acid (HEAR) to that with a low one (LEAR). It was intended to arouse the farmers’ interest in rape, which grows well in temperate climates, and to promote its cultivation instead of root crops and cereals, which are produced in excess. In 1988, the subsidy was 5.90 German marks (-3.30 U.S. $) per ton of rape (LEAR). The prices the producer can gain for his products on the market, however, are dependent on the demand in an individual country. In this respect, the price structure of butter and margarine in the individual countries is of interest. In some countries, for a large part of the population, margarine is a substitute for butter. The sales thus depend to a high degree on the relation between the prices for butter and margarine (Fig. 1.17). In Germany, for example, a portion of the consumers buy margarine whenever the price differential to butter increases beyond a certain value. This fact, which applies to other countries as well, was used by the EC-commission to diminish the “butter-mountain” by bringing butter onto the market at a reduced price. In the period from 1982 to 1985, heavily subsidized butter in the EC amounted to >2.2 million ton. To subsidize this quantity, which replaced 80% of the fresh butter, 1600 ECU per ton had to be paid (>3200 German marks or 1500 U.S. $ per ton at the exchange rates of 1986; Friedeberg 1986). In other countries, the image of margarine is completely different, and butter plays only a minor role. Interestingly, this also applies to the Netherlands and Denmark (buttedmargarine 1:7 and 1:3, respectively), countries that are commonly known as “milk countries.” 10
a
8
g
6
Butter
Vegetable
I
2 k
p.. margarine
Butter
4
.-8
Average margarine price
ti
2
0
,
i
I910
4
1930
I
,
1950
1970
1990
Fig. 1.17. Price differential of margarine and butter in Germany.
Fats and Oils Handbook
22
1.3.1 The Economic Importance of Oils and Fats as Well as of Fat Products in Europe and Germany
As stated in the preface, Germany is seen as a particular example of a western European country. In the past 50 years, nutrition habits in industrialized countries have changed drastically (Fig. 1.18), with Germany as an example. (Remark: all data in figures concerning Germany end with the year 1990 because data after the reunification would not be comparable.) In spite of diminishing hard physical labor, caloric intake has grown, and despite the findings and recommendations of nutritionists, the percentage of fat in the diet has increased. Since 1850, the consumption of fat in industrialized countries has risen constantly. It is assumed that the German population satisfies -40% of its energy demand with fats and oils. Currently, the amounts consumed consist of approximately equal shares of butter, margarine, and edible oil, as well as tallow and lard (Fig. 1.19). The proportion of margarine has dropped since the 1950s in favor of edible oil and fat. The market for emulsion fats altogether is currently dropping at a rate of -3% per year. The shift in proportions is disadvantageous for the visible fats, whose proportion or intake can be controlled consciously, and favors the invisible fats (proportion -1: I). Currently, the annual per capita consumption amounts to -30 kg of visible fats (Fig. 1.20). The history of fat policy in Germany (which is representative of other European countries) is depicted comprehensively by Schtittauf and Pischel(1978), whose work is referred to in part in the following paragraphs. Their overview reflects the political turmoil in Europe during the last 80 years. 200
0
I
6
eh)r
1940
I
w
* &--" #*-
1980
. .us*m --m -
1960
1970
1980
1990
Fig. 1.18. Per capita consumption of different food in Germany
Importance of Fats
23
Fig. 1.19. Proportion of some fats and oils on total fat intake.
Today, Germany’s self-sufficiency concerning vegetable oils and fats is quite low. About 90% of the raw materials are imported. In the early 1900s, Germany’s self-sufficiency for fats and oils was -50%; between 1945 and the present time, the overall self-sufficiency is -40%. Despite these shortcomings, there have always been attempts to impose special taxes on the import of oils and fats (fat-tax). Actually, this tax has never been directed against the fats themselves, but against fat products, especially margarine.
Fig. 1.20. Per capita consumption of visible oils and fats in Germany.
24
Fats and Oils Handbook
Toward the end of the last century, with the introduction of margarine, the first legal restraints were introduced; initially, these amounted only to sale restraints (e.g., no butter and margarine in the same room). However, a duty was imposed on the import of margarine or equivalent oil mixtures according to Bismarck’s policy of protective duties. Because this duty did not apply to individual oils but only to oil compounds or to finished products such as margarine, the first margarine plants were built directly along the border in The Netherlands, which were most progressive, so that transportation costs for the duty-free raw materials were as low as possible. During World War I, a considerable shortage of fat occurred. Naturally, imported fats and oils became scarce first, so that fats for margarine were not available in sufficient quantity. From 1917 on, only one third of the required volume of fat could be put on the market. In the respective figures, this becomes apparent through the lows in margarine production, and better yet, in the contrary development of the fraction of margarine and butter at that time. During the time of inflation, caused by the putative upswing during the postwar period, the number of margarine plants in Germany rose to several hundred (today the number is -10). On May 1, 1933, a fat-tax, combined with a fixing of quotas, came into effect in the Third Reich. The production of margarine was frozen at 60% of the 4thquarter output of 1932; moreover, a tax of 0.50 Reichsmark per kilogram of margarine, edible oil, hardened vegetable oil and whale oil was inflicted. An “Agency of the Reich for milk products, oils, and fats” was established. Soon, only a standardized margarine, packed in a simple, brown, unattractive wrapper was permitted to be marketed. As a consequence, 115 out of 148 margarine plants were forced to close. It was not until 1949 that branded margarine products were again permitted to be sold. However, at about the same time, the oil milling industry was financially hard hit because it was forced to sell to the agriculture sector meal at 50% of the world market prices. A fat-tax was again considered in 1950. It was prevented by protests from labor unions and social associations. Instead of the fat-tax, vegetable oils and fats were subsidized through the end of the Korean War. In later years, the fat-tax was considered several times, mainly during the times when there was an excess of butter, the so-called butter mountain, in the European Community. The fat-tax was consistently prevented. The primary reasons for the failure of the fat-tax included strong protests from consumer groups who feared higher prices. Another reason for the failure of the fat-tax was intervention by the U.S. (different reason than before), who did not want to see its export of soy products diminish because of the enormous rise in U.S. soybean production. At that time, the American soy farmers were feeling the effects of poor prices for their crop. Finally, this would have been the first tax that the EC could have inflicted, and increased independently from the individual governments. The EC certainly needed the revenues because subsidies of other agricultural products amounted to several hundred per cent (i.e., in 1977,220% for butter and 107% for olive oil).
Importance of Fats
25
The prices for fats and oils in Germany (as in other European countries with hard currency) in terms of constant money have developed differently from those in the U.S. Two opposing influences are operating here, i.e., the substantially lower inflation rate in Germany compared with that of the U.S. (over a period of 30 years from 1965 to 1995, only -56% of the U.S. rate) and the depreciation of the U S . $ with respect to the German mark (from 1965 to 1990, -45%). The effect is clearly visible in Figure 1.21. Presenting the price for soy on the world market in U.S. S as well as in German marks, according to the respective exchange rate and with an adjustment for inflation, one can see that soybean oil in Germany in terms of constant money costs only -30% of the amount in 1965 (U.S.: 36%; Fig, 1.21). This is reflected in the consumer prices for oils and products consisting mainly of oils and illustrates some of the problems of the fats and oil industry as a whole. After the reduction of the butter-mountain, which had temporarily reached >1.4 million ton, by means of regulative measures by the EC (quotas for milk), a certain pressure to intervene for regulating purposes has subsided. The next decisive step will be the integration of the different forms of agricultural producers in the new and old federal states of Germany into an all-German system. In the new states, a potential for the production of oilseeds is building up; in relation to the population, it is larger than that in pre-unification Germany. The European Union will be faced with the same problem when considering the membership of Eastern European countries such as Poland. Figures 1.22 and 1.23 reflect the situation in the German oil milling industry, which crushes about two thirds of the oil consumed. The development of the amounts of oil produced per variety reflects the European Union’s move towards rape and sunflower, and the crushings demonstrate the dominance of these two
Fig. 1.21. Price of soybean oil in current U.S. $, and current and constant DM.
Fats and Oils Handbook
26 I.o
I
0.8-
Er 0.6 -.-
0
0
3
Rapeseed oil
//I Palm kernel oil
Soybean oil
Oa4
T Peanut oil
\
I+
coconut oil
0.2
Sunflower seed oil
0 I910
1925
1940
1956
1970
1985
Fig. 1.22. Oil produced in German oil mills.
oilseeds over soybeans. Although more southern countries such as France are crushing relatively more domestically grown sunflower seed, this picture can be regarded as typical for Western Europe. In spite of the difficulties with eamings and profits and the strong competition from abroad, the quantities processed increased until 1980 and have since remained
r-
6
5
€ 4
f . P i
e
seeds (total)
Meals (total)
Soybeans Rapeseed Soy meal
2
a
Oil (total)
I
0 1910
1925
1940
1956
1970
1985
Fig. 1.23. Seed processed in German oil mills.
Importance of Fats
27
relatively stable. This is partly a result of the trend to move the mills closer to the large sea harbors, such as Rotterdam in The Netherlands. The quantity of extracted oil has increased because the oil content of rapeseed is higher than that of soybeans. In the processing of soft seeds (rape/sunflower), the Central European mills will thus have greater chances of competition than in crushing soybeans. 1.3.2 Oil Politics in the European Community
At the time when the guidelines of the EC-agricultural policy, which also encompasses the production of oil, were laid down, the degree of self-supply for vegetable oils (except olive oil, which is regulated separately) was ~ 1 0 % and that of vegetable proteins 4%. In the course of the past years, this figure has risen to >50% for vegetable oils (Fig. 1.24). Certain types (rapeseed, sunflower seed) are even exported (Friedeberg 1989). The production is subsidized and protected by duties; in 1962, the EC committed itself in the Dillon-round of the GATT-treaty to not raise the duties on oilseeds, oils, and meals. The subsidies rose from -0.1 billion ECU in 1977 to -3 billion ECU per year in 1988 and have remained on this level. Subsidies were granted without limits on volumes and represented the difference between a representative price on the world market and a desired price (both fixed by the EC-commission for one year). The subsidies are paid via the oil mills. As mentioned above, there are no limits on the acreage and the quantity produced and thus no limits on the total amount of subsidies for an individual producer or, equally, the EC. By changing the targeted price, an incentive to produce oil fruit was
Fig. 1.24. Production of soybeans, rapeseed, and sunflower seed in the European Union.
Fats and Oils Handbook
28
TABLE 1.4 Fatty Acid Composition of Fat in Human Adipose Tissuea and Difference in the Composition of Serum Lipid Extract of Vegetarians and Nonvegetariansb Fatty acid in fat of adipose tissue
(%)
Palmitic Palmitoleic Stearic Oleic Linoleic
25 7 6 45 8
-
All others
9
Vegetarians
Nonvegetarians
Fatty acid (YO) in serum lipids
1982
1986
1982
1986
Palmitic Palmitoleic Stearic Oleic Linoleic Linolenic Arachidonic All others
20.1 3 .O 6.3 19.2 36.5 1.5 5.9 7.5
20.7 3.4 6.6 19.1 34.5 1.1 5.6 9.0
21.1 3.4 6.4 22.3 30.9 1.1 6.3 8.5
22.5 3.7 6.3 22.4 28.4 0.9 5.8 10.0
aSource: Ceigy. bSource: Melchert (1988).
created that led to the explosion of quantities as illustrated above. Friedeberg assumed two motives by the EC-commission for this policy. On the one hand, there was the uneasiness in being dependent on others (low degree of self-supply), supported by a very brief embargo on soybeans by the U.S. in 1973; on the other hand, there was the attempt to reduce the weight of the subsidies on grains, which oppressed the Community’s budget. By means of this policy, the support for the production of oilseeds became the third highest item in the EC’s agricultural expenses. The mawtude of the subsidies unduly burdens both the budget and the relationships to nations TABLE 1.5 Cholesterol Content of Food Food Vegetable oildfats Fish oils Lard Milk Milk powder Butter Milk fat Pork (lean, U.S., UK) Pork (+ fat, U.S., UK) Beef (lean, U.S., UK) Beef (+ fat, US., UK) Calf‘s liver Finfish, low fat Finfish, high fat Shellfish, crustaceans E!% Egg yo1k
Cholesterol (ppm) 59C-670 >650-710 >650-710 720 4900 470-570 590-790 860-1 200 41 0 16,000-1 7,500
Reference %her 1987 Tucker 1993 Seher 1987 Taufel 1993 Taufel 1993 Taufel 1993 Taufel 1993 Larnbert 1993 Larnbert 1993 Lambert 1993 Lambert 1993 Seher 1987 Childs 1993 Childs 1993 Childs 1993 Taufel 1993 Stadelman 1993
lmportance of Fats
29
with large agricultural exports. In 1988, the U.S. issued a fonnal complaint for the first time concerning the violation of GAIT. As a mechanism of stabilization, the commission proposed a fat-tax, which created quite a stir internationally and provoked the U.S. to threaten counteraction. After great protests by most trading partners as well as by many organizations within the EC, the plans were suspended for the time being. Subsequently, the EC searched for alternative solutions and found a system of “stabilizers,” the elucidation of which would be beyond the scope of this book. It is doubtful that this system will lead to the desired outcome. It is also uncertain how these regulations can be made compatible with Article 110 of the EC-treaty, which states that the aim of the EC’s trade policy is to contribute to the general well being through the following: a harmonic development of world commerce, the progressive lifting of barriers in international trade and the reduction of customs barriers. It is certain that the problems cannot easily be solved, but do seriously threaten the budget. In particular, it is difficiult to make compatible the aims of the EC’s agricultural policy according to Article 39; these include a sufficient income for the farmer, stable (internal) markets, reasonable consumer prices, increased productivity, and a secure supply.
1.4 Fat in Nutrition As previously mentioned, fat serves mankind as an energy supply, a reserve of energy, makes possible the intake of vital fat-soluble substances and supplies the body with essential fatty acids. The fat content of the human body is 16%in the embryo, with an adult body consisting of approximately the same percentage; deviations range from 8 to 50% (FriisHansen 1965). Fat is stored mainly subcutaneously and in the muscular tissue, as well as in deposits surrounding the inner organs such as the heart, kidneys, and intestines. In addition to its function as a quickly activated energy source,the subcutaneously deposited fat also serves as an insulating layer against hypothenria; the fat tissue surrounding the inner organs serves as a protective pad against physical injuries. The body can synthesize fat in part from carbohydrates, but to a large extent it is conveyed with food. The amount of fat in the diet cannot be precisely defined because of the influence of general living conditions (Gottenbos 1985 and 1988). Certain fat components, the essential fatty acids, are vital and must be supplied from outside. They are essential components of the cell membrane structure. Their metabolism is well known (Table 1.6; see Numa 1984, for example) leading to precursors of so-called eicosanoids that influence the behavior of the cell and are important for activities such as proper cholesterol transport. Fat conveyed with food passes through the stomach, is emulsified in the intestine by gall bladder secretions and is then hydrolized by lipases (pancrease), which are the enzymes of the intestine and the pancreas. The lipases present in the stom-
Fats and Oils Handbook
30
TABLE 1.6 Metabolism of Essential Fatty Acids Effect
Enzyme working
a-Linolenic acid ALA (18:3w3)
Linoleic acid (18:2w6) A6-Desaturase
.1
-2 H
Elongase
1
+2c
-1
-2 H
I
+2c
I Docosaheptaenoic acid DPA (22:5@6)
1 Docosa heptaenoic acid DPA (22:5w3j
Adrenic acid ADA (22:4w6) A4-Desaturase
I Eicosapentaenoic acid EPA (20:5w3)
Arachidonic acid AA (20:4w6) Elongase
1 Eicosatetraenoic acid (20:4w3)
Dihomo-y-linolenic acid DGLA (20:3w6) As-Desaturase
1 Stearidonic acid (18:403)
y-Linolenic acid GLA (18:3w6)
-2 H
1 Docosahexaenoic acid DHA (22:6w3j
ach separate fat that is hulled in protein from its protein hull. Hydrolysis is continued in the duodenum to yield ~ 1 0 % of triglycerides and diglycerides as well as 40-50% of monoglycerides, 40-50% of free fatty acids and glycerol. In the first 100 cm of the small intestine, the oily solution of triglycerides and phospholipids (chylomikrons) with a droplet diameter of 0.5 pm exists alongside the microchylons (0.05 pm), which consist of mono- and diglycerides and salts of gall acids. These are further broken down into the micellar fraction consisting of monoglycerides, fatty acids and gall acids. The particle size has then reached -0.005 pm, which is sufficiently small to pass through the intestinal wall. Passage is possible for particles smaller than 0.01 pm (Ludwig 1968). The fat is reconstituted after its components have passed through the intestinal wall (Langdon and Phillips 1961). Short-chain fatty acids can pass through the intestinal wall more easily; however, this is not of importance for healthy people (cf. also Chapter 8.8). The fat enters the body via the lymphatic system, and any unneeded surplus is stored in fat deposits; the remainder is conveyed to the liver metabolism. Fats with melting points 4 0 ° C are virtually completely digestible. For additional information on the metabolism see, for example, Welch (1993). For the nutrition physiology of fats, unsaturated fatty acids are especially important (Hunter 1989). They consist of three families and are characterized by the position of the first double bond of the fatty acid chain, counted from the methyl group. An n-x fatty acid has its first double bond between the xth and x + 1 C-atom of the chain counted from the end. The next double bond is usually situated three C-atoms further along the chain. Representatives of the three main groups are as follows:
Importance of Fats
n-3 linolenic acid n-6 linoleic acid n-9 oleic acid, erucic acid
31
(also 0-3) (also 0-6) (also 0-9)
The hydrolysis of fats is performed by the same enzyme irrespective of the fatty acids. Among unsaturated fatty acids, the enzyme has its highest activity for n-3 fatty acids. In the reesterification after passing through the intestinal wall, the composition of the triglycerides in relation to fat taken in with the food is changed because fats stored in different regions of the body exhibit typical fatty acids patterns. Their composition is relatively constant but can be changed by very unbalanced nutrition or high doses of fat. Human adipose tissue is composed essentially of only five fatty acids (Table 1.4). A considerable portion of linoleic acid must be accumulated in the body because linoleic acid, as an essential fatty acid, cannot be synthesized. The difference becomes visible when comparing the relative fatty acid composition of fat in vegetarians with that in nonvegetarians (Table 1.4). Mammals have the ability to convert saturated fatty acids, but only into those that are monounsaturated, with the location of the double bond only at C-9 (Thiele 1982). This chain can be prolonged toward the carboxyl-end, but not toward the methyl-end. Thus, the synthesis of linoleic acid is not possible in animal organisms (cf Chapter 2.1). In addition to linoleic acid, arachidonic acid (formerly called vitamin F; AaesJorgensen 1961) is also regarded as an essential fatty acid. For about 60 years, it has been known that these two fatty acids are vital (Burr and Burr 1929); for example, they comprise the initial stages of prostaglandins (Bergstrom and Samuelson 1965). Prostaglandins were discovered in sperm by von Euler in 1934, but they are present throughout the body. They are the building blocks of hormones and possess their own physiologic activity as well (hypotensive activity, stimulation of the sleek muscles, regulation of the release of fatty acid from fats). Moreover, essential fatty acids are necessary for growth, contribute substantially to the building of cell walls, and form a structurally essential component of phospholipids. They occur mainly in the brain and nerves and participate in many metabolic processes including those of mitochondria. If the supply of essential fatty acids is insufficient, other nonessential ones are built into cell walls, leading to disorders. Among others, Hansen et al. (1958), Thomasson (1953), Holman (1961), Holman et al. (1964), Aaes-Jorgensen (1966) and Vles and Gottenbos (1989) reported such deficiency symptoms. A deficit of essential fatty acids can result, for example, in reduced growth, lowered prostagladin synthesis and skin damage. The symp toms disappear or come to a halt when n-6 essential fatty acids are supplied. The amount of essential fatty acids (e.g., linoleic or arachidonic acid) that should be present in the diet were previously stated to be at least 2% of the entire calorie supply (Holman 1961), which corresponds to an intake of -2.4 g of linoleic acid4187 kJ (lo00 kcal). The recommendations were developed further, and in more recent recommendations by the FAO/WHO (1977), 3% (corresponding to 3.6 g/4187 kJ or loo0 kcal) is stated as the desirable quantity. This percentage should be increased to
32
Fats and Oils Handbook
during pregnancy and to 5 7 % in the lactation period. According to Adam et a!. (1958), infants should receive twice that amount. Wolfram (1987) surveyed the
metabolic effects of a diet rich in linoleic acid. Another important positive quality of essential fatty acids that is the subject of a growing number of studies is their lipid-lowering quality as well as their ability to exercise a favorable influence on an excessive cholesterol level in the blood. This has a special significance for health issues because both an increased lipid level and an increased cholesterol level are considered to be risk factors for heart attacks. Cardiovascular diseases are the number one causes of death in industrial nations throughout the world. The lowering of lipid levels as a precautionary measure has been known for more than 30 years (Ahrens 1957 and 1959, Groen et al. 1952, Kinsella et al. 1952); although it is undisputed among experts, it is repeatedly attacked by lobbies. Mertz (1983) conducted a survey of the current state of knowledge. These findings on the connection between cholesterol levels and cardiovascular diseases were supported by the awarding of the Nobel Prize for medicine in 1985 to Brown and Goldstein, who were pioneers in the studies on this subject. Arteriosclerosis, caused by cholesterol esters and elevated cholesterol level, has clearly been identified as one of the key risk factors for heart disease (LRCCPT 1984, Schlierf 1986). Today, a distinction is made between LDL (low-density lipoprotein), commonly labeled “bad cholesterol,” and HDL (high-density lipoprotein), which is considered “good cholesterol.” HDL is responsible for transporting surplus cholesterol from the body to the liver. Thus it represents a means of transportation. A high level of LDL is directly related to heart disorders. A survey of the influence of nutrition was given by Schettler (1984). The lowering of the cholesterol level is due to the direct supply of linoleic acid and to its relative quantity in the fat. Stamler (1966) showed that when raising the intake of polyunsaturated fatty acids from 9 to 15%, the amount of cholesterol in the blood serum decreases by only 1.2 mg/mL. When the supply of saturated fatty acids is reduced from 16 to 9%, which indirectly results in a relative increase in polyunsaturated fatty acids, the amount of cholesterol in the blood serum drops by 1.9 mg/mL. The fundamental factor is not the amount per se of polyunsaturated fatty acids, but the proportion between them and the saturated fatty acids. This proportion is also called the P/S ratio (P = polyunsaturated fatty acids, S = saturated fatty acids). According to Mertz (1983), the P/S factor in the food of the German population was 0.39. This is far below the desirable ratio of 1:1 (P/S = 1). As a consequence (e.g., for diet margarine), the legislature will probably abandon the exclusive specification of the content of linoleic acid in the declaration of diet products and require in the future that the portion of saturated fatty acids or the P/S quotient be stated. It is doubtful that reaching the positive effects of a higher P/S ratio as stated above through increasing the supply of linoleic acid to the diet will be achieved, especially considering the rivalry between margarine and butter. This is mainly a problem of agricultural politics (cf. Chapter 1.3) and not one of health politics. The fact that it is not the‘ origin of the fat (vegetable or animal), but rather the P/Sratio that is crucial
lmportance of Fats
33
is often neglected. The P/S ratio of butter is 0.05; that of vegetable coconut fat only 0.02. However, in spite of its vegetable origin, the latter is not considered suitable as an exclusive fat for healthy nutrition. The quantities consumed are low because it is not a basic food but one found only in specialty products. Sunflower oil, on the contrary, is especially suitable (P/S = 5.82;Wirths 1981). Its P/S ratio is so high that the sunflower oil can compensate for higher saturated fatty acid intake in the normal fat supply. More recent analyses led to the recommendation to adjust the proportions of polyunsaturated fatty acids, monounsaturated fatty acids and saturated fatty acids (PUFAIMUFAISAFA) to 1:l:l (MI3 1984). These findings do not call into question prevailing nutrition recommendations regarding linoleic acid as an essential fatty acid. Findings in this area have not changed. However, a stronger positive (lowering) influence of the monounsaturated acids (i.e., predominantly oleic acid) on the cholesterol level than estimated has been found. Oils rich in oleic acid (e.g., rapeseed oil, olive oil) can thus be suitable for a cholesterol-reducing diet as well as those with a high PUFA content (e.g., sunflower oil; see, for example, Laasko et al. 1989). However, because the invisible fats are rich in saturated fatty acids, diet margarines should have a content of saturated fatty acids ~ 2 0 % in order to reach the targeted proportion. Altogether, science today is able to correctly predict the average change of the cholesterol level in the blood when nutrition is changed. Findings to date led the U.S. Federal Health Agency (NIH 1984) to recommend that Americans should reduce their present intake of 40% of calories from fat in their diet to 30%. Moreover, they should limit the intake of saturated fatty acids to ~ 1 0 % of the calorie intake and raise that of polyunsaturated fatty acids to 10%of the calorie intake (but not more). The daily intake of cholesterol should be limited to a maximum of 25&300 mg. The European Consensus-Conference (1986) adopted the American values in their recommendations and added maximum values for the therapy of high-risk persons. (Schwandt 1987). The European Arteriosclerosis Society required further measures (Assmann and Schettler 1987). When examining the cholesterol level, it is essential to take into consideration the intake of the cholesterol itself. An important source is animal products, especially animal fats, which typically contain substantial amounts of cholesterol (Table 1.5). According to a general agreement derived from British legislation, substances with 4 0 ppm are regarded as cholesterol free. Vegetable oils not only have the advantage of reducing cholesterol levels, but they also do not add to the intake of cholesterol (Seher 1987). Thus, with an elevated cholesterol level, nutrition with the correct fat must be balanced with a suitable nutrition plan. Proposals were made to shift the P/S factor in the direction of one, but only sick and high-risk people need to take special care. At present, there are efforts to remove cholesterol from butter to avoid at least one of its detriments (extraction with supercritical COz, Kankare and Antila 1989). The findings in the cholesterol debate were surveyed by Goldberg and Schonfeld (1958), Grundy (1986), and McGandy and Hegsted (1973, among others.
34
Fats and Oils Handbook
Knowledge about the disadvantages of animal fats has continually led to attempts to reduce their presence by changes in animal feed (Leaf and Weber 1988). This applies particularly to the feeding of milk cows. It was possible, for example, to raise the content of n-3 fatty acids in milk fat to 6% by mixing fish oil (Menhaden oil) into the fodder (Hagemester 1989). The passage of the n-3 fatty acids from the fodder to the milk fat was between 35 and 40%. In this special instance, such milk fat is preferred to conventional milk fat from the point of view of nutrition physiology. However, it is uncertain whether it will provide a reasonable alternative considering the conversion factor mentioned above. For about 20 years, nutritionists have concentrated on n-3 fatty acids (a-3 FA), because Bang and Dyersberg (1975) observed that the Greenland Eskimos, in spite of their extensive consumption of fat, suffered less from heart disease than the Danes. Their biochemical values were very close to those of a Japanese population of fishermen, who also subsisted mainly on marine animals (Yamori et al. 1985). The Eskimo diet (marine oils) is rich in n-3 fatty acids. In marine oils, n-3 fatty acids are almost exclusively long-chain acids. It is estimated that the required daily quantity of these fatty acids is 0.2% (Benadt 1988). The daily intake with Western food is -1 g of n-3 fatty acids. Investigations concerning the advantages and possible disadvantages of these acids have not yielded any conclusive results to date and are still in progress. However, studies in Europe indicate (Kromhout et al. 1985, study in Zutphen, The Netherlands) that there is a connection. There appears to be definitive evidence that the daily consumption of n-3 fatty acids must exceed an average of 2 g to achieve an effect (Driss and Darcet 1988); this is equivalent to daily consumption of the relatively high amount of 200 g of fish. Findings are not sufficiently advanced to recommend the intake in concentrated form, e.g., in capsules, or to incorporate such fatty acids into fat products. Moreover, the quantity that would then have to be used to ensure the effect is rather high, although their efficiency is 20 times higher than that of linoleic or linolenic acid (Singer). It is also questionable whether a continuous nourishment with foods rich in n-3 fatty acids could be harmful. A survey of the state of the discussion about fatty acids from fish oils was given by Harris (1989). The concern that polyunsaturated fatty acids may be susceptible to oxidative damage and might develop into carcinogenic substances within the body is repeatedly expressed by scientific outsiders, but is unfounded. Dormandy (1983) determined that in-vivo oxidation of fatty acids does not take place. Within the body, as in the intact seed, they are protected by the body’s antioxidants. The consumption of fats oxidized (peroxides) by inappropriate handling is also not dangerous. Fats containing harmful concentrations of oxidized fatty acids are not edible as a result of their very bad taste. The fat supply itself constitutes another problem. Fat consumption in the industrial nations today satisfies between 35 and 45% of our energy demand, and in the developing countries between 10 and 20%. The worldwide consumption of fat during the mid-1970s was 12.5 kg per capita per year. In Germany, for exam-
lrnportance of Fats
35
ple, more than three times that amount of fat was eaten. The level of fat consumption parallels the rise in standard of living. Thus, the consumption of animal and vegetable fats and oils in the U.S. more than doubled between 1950 and 1985 (Hammond 1988). In contrast with Germany, however, there was a strong shift to vegetable oil consumption. This is disquieting in two respects. As a rule, the intensity of physical labor decreases as the standard of living rises. Recommendations for the daily intake of calories are as follows: 2000kcal 2300 kcal 3200 kcal
(8370k.l) (9620 k.l) (13390 k.l)
for light work, for medium work and for hard work.
This would mean that even a constant consumption of fat would be too high, because surplus energy is conveyed to the body. On the other hand, fat consumption is rising because fat makes food tastier. It is recommended that calorie demands met by fats in the diet range between 25% and a maximum of 35%. Adding to this negative trend is the fact that much of the fat in foods is in the form of hidden fats. Hidden fats are fats contained in other foods and consumed unconsciously. This applies especially to meat, sausage, and cheese (c$ Table 1.2). Studies in the Federal Republic of Germany in 1982 showed that >50% of the fats consumed was in the form of hidden fats. In addition to the fact that the population is not aware of which foods contain what quantities of fat, it becomes increasingly more apparent that the predominant part of hidden fats consists of animal fats with a high proportion of saturated fatty acids. In this manner, the P/S quotient is lowered (which is negative), and no essential fatty acids are consumed. When the calorie supply is reduced, which usually occurs by lowering the amount of visible fats, the negative tendency is further intensified. Attempts to exchange the saturated fats in foods for highly unsaturated ones are constantly thwarted by protectionist legislation (c$ Chapter 1.5). Greater freedom remains unattainable as a result of the harmonization of the respective laws and regulations within the European Union. Besides the fat-associated substances, minor components are important as well. Some of these have negative effects but are found only in the grain; in that case, they cannot be used as feed for all animals. Experiments have shown that gossypol from cottonseed causes pathologic changes in the testicles of mammals that can lead to sterility (Berardi and Goldblatt 1980, Xue 1980). For this reason, it was tested in China as a contraceptive for males. Its effect is attributed to the generation of oxygen radicals. Refined cottonseed oil does not contain gossypol, thus the attention is directed exclusively toward cotton grain. Fat-soluble vitamins have considerable positive influences. Vitamins A and the family of carotenes possessing vitamin A activity, as well as vitamins D and E, are fat-soluble and water-insoluble substances. Consequently, these substances occur together only with fat, i.e., they can be conveyed to the body only via fat-containing
36
Fats and Oils Handbook
food (Sebrell and Harris 1954). Vitamin A is necessary for regular growth, normal eyesight and procreative capacity. Moreover, it plays an important role in the stability of the cell membranes. Vitamin D ensures the correct calcium level in serum and is necessary for the normal growth of bones. Vitamin E protects vital substances such as unsaturated fatty acids, vitamins A and D, as well as thiol groups in enzymes from oxidation (c$ Chapter 2.2). In contrast to the essential fatty acids, vitamins occur in animal and vegetable fats, although vitamin E is present in larger quantities only in vegetable fats. Thus, with a normal food supply, deficiency is not an issue. Vitamin E represents one of the essential radical scavengers in lipid membranes (Pryor 1976). It was applied in clinical trials to combat illnesses caused by oxidation processes. These trials were rather successful (Bieri et al. 1983). A protective function towards DNA was observed (Beckmann et al. 1982), as was greater endurance in test animals (Davies et al. 1982). Other experiments suggest that vitamin E has anticarcinogenic effects as well (Wang 1982). p-Carotene also has antioxidant effects. It intercepts singlet oxygen, which has a strong mutagenic effect as a result of its high reactivity (Foote 1988, Krinsky and Deneke 1982). Experiments have also shown a protective function against the development of cancer (Mathews-Roth 1982, Rettura et al. 1983). A survey of the multiple fields of applications was made by Ames (1983). It is assumed that fatsoluble vitamins are helpful against oxidized metabolites of cholesterol that were observed to contribute to the development of heart disorders (Yagi et al. 1981). Apart from the anticarcinogenic effect of some of the minor components in fats, studies reporting a direct correlation between the fat intake and the frequency of breast and colon cancer continue to appear (Doll and Pet0 1981, Fink and Kritchevsky 1981, Kinlen 1983, NRC 1982). To date, it has not been possible to establish a direct connection. However, there seem to be more indications that the frequency of cancer generally rises with caloric intake. Only in this connection could fat be a role-playing factor.
1.5 Fats and Oils in Legislation The legislation in effect for this branch of industry applies to the products, thereby directly affecting the consumer and producer. The legislation also applies to the production process. 1.5.1 Product-Related Legislation
Many foods are narrowly defined by laws and regulations concerning their composition, mode of production and qualities. Not everything can (or should), however, be regulated by laws. In addition to legal directions, certain modes of behavior (principles of the responsible producer, good manufacturing practice) have emerged and various codes have been formulated, e.g., the Codex Alimentarius of the FAO, and the Leitsatze des Deutschen Lebensmittelbuches (Guiding Principles of the German Food Book) is an example for Europe. These guiding principles are
Importance of Fats
37
not legally binding but are consulted to define honest trading practice and are the foundations of legal decisions in case of controversy. In many cases, they fill the gaps where neither laws nor special regulations exist or give specifications which exceed legal limitation or description. In principle, there can be two motivations for product-related legislation in the domain of food, i.e., the protection of the consumer and citizen and aspects of economic policy. The protection of the consumer can include matters of health, but also protection from fraud. Aspects of economic policy might include emergency situations (war/postwar), partial attempts at selfsufficiency, or the preservation of an agriculture that is no longer competitive under the conditions of the world market. Concerning health risks, the protection of the consumer is always foremost and falls under the duties of the state within the obligation of public care. Frequently, however, the extent and the character of the measures to be taken are under discussion. Fats and oils, as such, fall under the common rules on maximum values of environmental pollutants. In contrast, there are very detailed rules for the products described in Chapter 8 (butter, margarine and mayonnaise). Legislative interventions motivated by health policy in the field of oils and fats have rarely occurred. An example is the regulation regarding the maximum content of erucic acid; however, this has been rendered superfluousby the cultivation of new types. For ingredients, processing aids, and additional substances, there are two basic approaches. One is that all substances that are not deemed harmful can be admitted, and consequently, harmful substances are prohibited or restricted in their quantity. The other is that everythrng that is not explicitly allowed is prohibited. The EC follows the second principle in many fields (regulation about the admission of food additives). This is a policy motivated by a desire for control rather than one of protection from danger, a fact that becomes immediately obvious when questioning why an ingredient is prohibited in one food but allowed in another. With few exceptions, one and the same substance cannot be at once both harmful and harmless. Trade policy is aimed primarily at the protection of agriculture. Due to their structure, the European States are not able to offer all agrarian products at world market prices. Therefore a certain protection is advisable to maintain at least partial independence.It must be considered, however, that excessive protection also prevents the seizing of opportunities. Legislation usually intervenes in times of emergencies. This happens primarily by means of regulations or guiding principles. Even more so than laws, guiding principles are a mirror of their times. During World War 11, they gave reliable information about the supply situation in Germany because regulations were in each case adapted to it. Thus, according to the German guiding principles for mayonnaise of 1941, salad mayonnaise, for example, had to contain only 20% oil instead of 50%, and milk and fish protein were allowed as substitutes for egg yolk. 1.5.2 Production-Related Legislation
In the production of oils and fats by means of mechanical or solvent extraction, in their processing or refining, and in the making of products containing fat, the producer
Fats and Oils Handbook
38
is subject to many general legal regulations concerning emissions or sewage, for example, but also to several highly specialized injunctions. These directives can differ widely depending on the location of the business, and even within one country from community to community. Dealing with these regulations in a detailed way and on a universally valid basis is possible but would exceed the scope of this book.
1.6 Fats as Industrial Raw Materials A relatively large portion of edible fats and oils is utilized for industrial purposes
(not nutrition). Often batches that do not comply with the strict demands for foodgrade raw materials are used for this purpose. Worldwide, the production of oleochemicals is -9-10 million ton (Seidel 1983), and a wide range of products is produced (Fig. 1.25). For some oils and fats, the portion not used for nutrition is considerable (Table 1.7). In 1981, after a continuous rise since the first energy crisis in 1973, the price of the raw material ethylene had climbed beyond that of soybean, coconut and palm oils, and tallow (Fig. 1.26). At that time, large chemical companies tried to secure their position by acquisitions that would provide a position in this raw material market. In the meantime, ethylene prices have fallen again. This interest was also motivated by a desire to remain in the market and exert influence on the types of new crops cultivated exclusively for the food industry. Erucic acid is relatively easily modified chemically. With the transition to rape OillFat
Olymml Resins EmulabIan C e i l U I o ~produdion o i for n ~ ~POlyOlS mineral oil produdion CosmsUcs Glymml Eaten Detergents
Auxiliary ~
Fatty h i n o s
Fatty Alcohols
Fatty Amldes
Adddiws for coal flotaUon Anti mnorives Sufladsntr Emubifiers
Esters with polyethybnglycd Fabric deanen Non-ionogenic emuI8hien
Detergents Emutsifien Fire extinguishen
Animal feed soaps Chemicsb fcfthe mineral oil and rubber industry
Fig. 1.25. Simplified flowchart of oleochemicals production.
Importance of Fats
39
TABLE 1.7
Uses of Vegetable Oil for Nonfood Purposesa Oil type and nonfood usage (YO) 0.25
Soybean oil Palm oil Palm kernel oil Rapeseed oil
10 10 40
Coconut oil Castor bean oil Linseed oil Tung oil
55 100 100 100
Type of usage (% of total nonfood usage) 36
Fatty acids Animal feed Soap Other
29 15 13
Paints Lubricants Polymers
3 2 2
aSource: Pryde and Rotfus (1 989).
species with low erucic acid content, the chemical industry was robbed of an important raw material. From then on, the prices for a low tonnage of rape with a high portion of erucic acid (HEAR) were above those of LEAR oil. Today crambe oil, which also possesses 55-60% of erucic acid, is considered to be a replacement for HEAR. Mustard seed oil can also be used. Taken as a whole, the production of raw materials in fat chemistry has developed from synthetic to natural raw materials, and this trend is continuing (Table 1.8). 70
-
.
0
Fig. 1.26. Price relation between soybean oil and ethylene.
Fats and Oils Handbook
40
TABLE 1.8 Oleo Chemicals and Their Raw Material Sourcesa Fatty acids
Amount (1 OOO ton) YO Natural YOSynthetic
Fatty alcohols
Glycerol
Fatty arnines
U.S.
Europe
U.S.
Europe
U.S.
Europe
U.S.
Europe
lo00
650
160
195
350
210
110
64
98 2
99 1
57 43
72 28
16 84
37 63
85 15
100 0
dSource:Seidel (1 983).
1.7 Fats and Oils as a Source of Energy In the course of history, it has been demonstrated repeatedly that fats are suitable as a source of energy. Rudolf Diesel had already determined that his engines could run on edible oil. As early as 1900, a Diesel engine powered by peanut oil was shown at the world’s fair in Paris (Nitske and Wilson 1965). However, the true potential of renewable raw materials may lie not in the combustion of the oils known today but in the utilization of new species. When we compare the qualities of vegetable oils with Diesel fuel, we notice that the caloric value is -10% below that of Diesel oil (Table 1.9). Among renewable raw materials, vegetable oils exhibit the most favorable relation between energy yield and energy investment (Table 1.10).Because vegetable oils have always been used in part as a source of energy, the idea is not new. However, the demands on the oils for use in modem engines have changed compared with those for illumination purposes. A study by Apfelbeck (1988) shows which fuel parameters must be met for today’s vehicles (Table 1.11). When comparing the prices for Diesel and soybean oil, we see that for the time being, there is no point in using edible oils and fats for combustion purposes. TABLE 1.9 Comparison of Diesel Oil, Rapeseed Oil, Sunflower Oil and Their Methyl Esters Sunflower
Caloric value (MJ/kg) Density (dcrn) Viscosity (cP,20°C) Cloud point (“C) Flash Point (“C) Ash (Yo) Sulfur (YO) Sulfur (rnol%) Reference
Diesel oil
Crude oil
Methyl ester
4246 0.835 3.9 -0.6 50-77 0.01 -0.27 -0.1 4 Shell
39.28 0.925 34.7 -6.6 215.5 0.04 0.1 2
40.16 0.880 4.22 0-1 183
-
-
0.01
-
Quick 1989
Rapeseed Crude oil
36.7-37.7 0.91-0.92 68-97.7
-
Methyl ester
37.02-37.20 0.86-0.90 6-9
-
31 7-324 1 1 1-1 75 12% and ca.2h)
1
M'C, aeparaton. 2nd shge of Concentn(ion >>> Sweet crcam butter d u c t i o n >>> (sa8.1.5.5)
Skim milk
euitrr
usually rtomd d.ep frozen I
Holding for cwnpbh, making
Butter oil I
I
(99.5%)
I
Heating
to pumpability
to W C , pMe h u t exchanger; removal of water Water >>>
1 C
1
Cooiina
I
I
Butter oil
tO 99.9%)
Fig. 3.5. Flow chart of butter oil production from butter.
rendered fat produced per year is enormous as a result of the immense number of animals killed (Table 3.5). In the table, the difference between the number of slaughtered animals and quantity of animal stock is a consequence of the time required for the animals to reach the desired weight.
Animal Fats and Oils
129
Fig. 3.6. Plants for the production of butter oil from butter and from cream (redrawn courtesy of Tetra Lava1 Food AB, Lund). The determination of the amount of fat is not simple because a significant part is not processed separately but eaten with the meat without the consumer’s awareness (so-called hidden fat). Another substantial part is consumed in sausages, corned beef and other meat products. These hidden fats account for considerably more than half of the fats obtained from animal sources. The fat content of animals for slaughter has substantially decreased during the past 40 years. In the 1950s (in Europe) emphasis was placed on calorie intake, i.e., fat meat, whereas now the consumers ask for lean meat. In the 1950s in Germany, for example, the price for fat pork belly was 90% of that for lean pork chops; in the mid 1970s, it was 55%; and today it is >> f FNB
~8-
Water < 6.1%
a
I
Burning material or brlngingbeck to p*ntotion aa hrtl(ber
Crude palm oil
Fig. 4.30. Flow chart of palm oil production.
I
Palm oil
IP r o P o [%I~ ~
Fatty add
MyMtic Palrnitii Palmitolek Stearic Oleic Linokic Linobnic Arachic
C 14:O C 16:O C 16:l C 18:O C 18:l C 18:2 C 18:3 C2O:O
< 0.4 2.0 47.0 0.6 3.5 6.0 36.0 -44.0 6.5 12.0 c 0.5 c 1.0
0.5 41.0
-
-
Proportion [%]
Fig. 4.31. Fact file of palm oil (fatty acid composition). 4.2.2 Olive Oil
4.2.2.1 Botany and History of Olives. Olive trees occur mainly around the Mediterranean where they have been established for centuries. Most likely the olive originated in Asia Minor. In the Bible alone, olives are mentioned more than 200
Vegetable Fats and Oils
21 3
Fig. 4.32. World market price development of palm oil.
times. For example, a dove brought Noah a branch of an olive tree as a divine sign that the flood was over. Most likely olives found their way from Asia Minor to Europe around 1700 B.c., approximately the time when Athens was founded. The Greek legislation of Solon, as recorded by Plutarch, fined anybody who cut more than two olive trees per year and per plantation; they recognized the importance of the trees. Around the year 700 B.C., olives came to today’s Italy, and by 200 B.C. they had spread over the whole of the Roman Empire. From Spain, at that time a Roman province, the oil reached as far as the British Islands in the second century. After that, it was almost completely consumed by the citizens of Rome and free trade became forbidden. TABLE 4.20 Fact File of Palm Oil German: Palmol
French: huile de palme
Relative density Refractive index Saponification value Iodine value Unsaponifiable matter
(at 50°C; ref. water 20°C) (nD40) (mg KOH/g oil) (Wijs method) (g/Kg oil)
Melting point Solids content at
0.891-0.899 1.449-1.455 190-209 50-55
Decanters (conunuowly)
Polishing
Separators
I
Avocado oil
Fig. 4.42.
Effluent
F l o w chart of avocado oil production.
TABLE 4.25 Fact File of Avocado O i I German: Avocado61
French: huile d'avocat
Relative density Refractive index Saponification value Iodine value Unsaponifiable matter
(at 25°C; ref. water 20°C) (nD40) (mg KOH/g oil) (Wijs method) (fig oil)
Melting point:
-C '
World market price per MT (mid-1995)
Spanish: aceite de aguacate
Solidification point
0.9104921 1.466-1.468 185-1 97 70-95 4 6 7-9°C
37OC-3900 U.S.$
The following section deals with those seeds that are used for oil extraction worldwide or at least in substantial parts of the world. In Chapter 4.4, three wellknown oils are briefly covered, even if they are not suitable for human consumption. For the benefit of completeness, a few oils are described that are of special interest although they occupy only a small market segment, e.g., walnut oil or grape seed oil.
Vegetable Fats and Oils
223
TABLE 4.26 Fat Content of NonoiI Seeds
Seed
Fat (Yo)
Apple Apricot Lemon Coniferae Tomato
20 40 50-54 25-30 25
4.3.1 Soybeans
4.3.1.1 Botany and History of Soybeans. Soybeans (Glycine soju) stem from China, where they were mentioned in the history of the emperor Chennung (2800 B.C.). Today, they are still mainly cultivated there, as well as in the U.S. and Brazil. Recent theories assume that the true origin of the bean is Australia, from where it has conquered the Pacific region (Broue et al. 1977 and 1978). Soybeans have been native for a long time in Japan where primarily their protein content (for tofu and similar food) plays an important role. The U.S., today's prominent producer, started cultivation at the beginning of the 20th century (Probst and Judd 1973). Soybeans came to Europe with missionaries (1740) who brought them to the Paris botanical garden. In 1872, the soy plant was exhibited as a plant with potential at the Vienna world's fair. It came to Central Europe (Germany) in 1872, brought home from the German-French war by an officer named Werhahn, who cultivated it on his farm at Meissen, a city close to Saxony's capital Dresden, mainly known for its china. Soybeans can stand moderate climate but are sensitive toward cold nights. The plants can be grown up to -3000 m (10,000 ft) above sea level. Optimal living conditions are days 12-16 h long and temperatures of 3&32"C (Salunkhe 1986). The higher the temperature, the more the proteidfat ratio is shifted towards fat (Hymowitz and Newall 1981). Soy plants can grow to 30-180 cm in height. Their fruit are hulls 2-10 cm in length and 2 4 cm in width that hold 1 4 beans (Fig. 4.43). These beans are oval to round, yellow to violet; they rustle in their hull when they are ripe. The weight of 1000 seeds is between 50 and 350 g. Besides the oil content of 17-25 % (usually 20% or a little below), the high protein content of 3 0 4 0 % is of special importance. Soybeans are one of the most prominent protein sources in the world.
4.3.7.2 Economic Importance of Soybeans. Within 50 years, soybeans have developed into the most important oilseed grown and comprise approximately one third of all oilseeds cultivated. More than 50% of the beans come from the U.S. (in the fall) (Table 4.27). In the 1970s, cultivation in South America started and now accounts for -25% of the world harvest (in the spring). Due to that southern hemisphere crop, fresh beans are now available twice a year. Maximum yields in 1994 have been reported by Italy (3606 kgha), Ethiopia (3090 kg/ha), Egypt (2883 kg/ha), and the U.S. (2815 kg/ha); Ethiopia and
Fats and Oils Handbook
224
Fig. 4.43. Soybeans. Zimbabwe, which ranked numbers one and two in 1990 (4356 and 3453 kgha, respectively), were down to 3090 and 2500 kg/ha, respectively. Western Europe (European Union) had an average of -2300 kgha. The lowest yields of those countries monitored by the F A 0 were reported for Kazakhstan (229 kgha), Ethiopia (309 kgha), and Malaysia (333 kgha) (see Table 4.28). TABLE 4.27 Soybean Production (-85% Used for Crushing; Crushing Oil Equivalent -1 8%) Soybeans(MMT)
1935
1950
1970
1980
1990
1993
1994 .-
1995
Total world
12.3 1.2 10.0
46.5 17.3 27.3 7.7 15.1 30.8 8.0 11.5 8.3 0.08 0.3 1.5 0.001 0.001 0.03
93.8 54.8 7.6 13.4 3.6 0.04
107.8 52.3 11.5 19.9 10.7 2.2
114.9 50.9 15.3 22.6 11.0 1.4
136.7 69.6 16.3 24.9 11.3 1.0
129.9 58.6 13.5 25.6 12.0 1.0
us.
China PR Brazil Argentina EU (10)
-
-
1960
-
-
TABLE 4.28 Production Yield of Soybeans Soybeans(kg/ha)
1935
1950
1960
1970
1980
1990
1993
1994
1995
Total world
1090 1160 1160
1180 1400 1670 930
1080 1600 780 1140 980
1330 1800 840 1250 1030
1480
2030
1900
1703 1989 1099 1578 2014 1891 2697 1690
1840 2180 1302 2094 1601 2359 3134 2585
1857 2194 1237 2123 2159 2456 3155 1577
2207 2781 1735 2164 2038 2610 3350 1805
2022 2347 1662 2196 2044 2578 3614 1704
us.
China PR Brazil Argentina France Italy Spain
-
800
-
-
-
-
-
-
-
Vegetable Fats and Oils
225
The increase in acreage in the US.,the dominant producer, occurred rapidly. It developed from cultivation in Iowa, Illinois, Indiana, and Idaho in 1940, with Missouri, Minnesota and Arkansas added in 1950, and Mississippi, Kansas, and North Carolina in 1960. Increases in yield were paralleled by increases in acreage. In 1927, an acre yielded -12.2 bushels (-700 m a ) , whereas it doubled by 1960 and increased by another 60% to date. As a result of its dominant position, soybean oil plays a leading role in pricing (Figs. 1.10 and 1.13). The share of the value of the oil as a percentage of the price of whole beans dropped from -55% to -30% between 1950 and late 1980 (Erickson 1990). The market is now driven more by demand for meal than by the demand for oil. 4.3.7.3 Composition and Properties of Soybeans and Their Oil. Soybeans have a thin hull that makes up -7% of the seed weight. Depending on the genotype, the protein or oil content can be above or below the average. Nondehulled beans have an oil content of -20% (Fig. 4.44). Soybean oil is a good all-round oil. Its iodine value lies between 120 and 143, its solidification point below -8°C. It is rich in unsaturated fatty acids, with a relatively high amount of linoleic acid and, compared with other seed oils, a very high content of linolenic acid (Fig. 4.45). In addition to the main fatty acids, the oil contains up to 0.5% myristic acid and up to 3.5% saturated fatty acids with more than 20 C-atoms. In its triglyceride composition, the high amount of triglycerides with six and more double bonds appears remarkable (Fig. 4.46). Extraction meal from soybeans has a protein content of 40-50%. After heat treatment (toasting; see Chapter 5.4), the meal is suitable for animal feeding. Digestibility is then >80% because antitrypsine-factors are destroyed. Toasting is regarded optimal if the urease activity is ~ 0 . mg 5 nitrogen/(g min) and the protein has a solubility of 2040% in water (Lennerts 1984). The meal should not be stored longer than 3 mo (meal properties are listed in Figure 4.47).
-
Fig. 4.44. Composition of soybeans.
226
Fats and Oils Handbook
Fig. 4.45. Main fatty acid composition of soybean oil.
Because soybeans are an important protein source, they play a paramount role in food technology, not only in fat technology. Overviews include Smith (1972), Circle (1972), and more recently, Erickson (1995). 4.3.7.4 Soybean Harvest, Storage, and Oil Extraction. Soybeans are harvested with harvest-threshers when the plant is dried and the leaves have already fallen
Fig. 4.46. Triglyceride composition of soybean oil (after Johanssonand Bergenstahl 1995).
227
Vegetable Fats and Oils
Fig. 4.47. Composition of expeller cake and extraction meal of soybeans.
(Table 4.29). After harvesting, the seed should be stored with a moisture of 95% at t = 40°C), there is a danger that the nuts or the copra could acquire mold infections (Salunkhe and Desai 1986). The growth rate of the molds depends on the moisture content. Child (1974) describes the dependency as follows:
H20 c 7%
green molds
8%c H20 c 2%
brown and yellow molds
H2O > 12%
black molds
(Aspergillus niger)
H2O >> 12%
white molds
(Rhizopus species)
(Penicillium glaucum, Aspergillus glaucus)
-
(various Aspergilli)
Graalmann (1990) reports that in 1987/88, 40% of the Philippine harvest had to be rejected because the aflatoxin B1-content was above the German limit. Following that, there was considerable investment in dryers to avoid such disasters in the future. The method of harvesting does not greatly influence the commercial benefit that can be drawn from the fruit. Nambiar (1983) describes a novel process that is said to have a 10 times higher yield. In this process, the nuts are carefully opened
265
Vegetable Fats and Oils
aCJuIlhg (5.2.1.1)
siu dudion (5.2.1.3)
I.
blla
PHCiar
Mnd pnuino (5 2 2)
PlUpmukq (6 2.2)
li
r
Cnrdo coconut dl
Fig. 4.90. Flow chart of coconut oil production. and the milk is collected and separated into oil and protein. Figure 4.90 shows a flow diagram of coconut processing. 4.6.3.5 Fact File of Coconut Oil. The fact file gives the fatty acid composition, the physical data as well as price development (Table 4.47, Figs. 4.91 and 4.92). Further literature is given at the end of this chapter.
Fats and Oils Handbook
266
TABLE 4.47 Fact File of Coconut Oil ~~~
~~
German: Kokosfett
French: huile de coco
Spanish: aceite de coco
Relative density Refractive index Saponificationvalue Iodine value Reichert value Polenske value Unsaponifiable matter
(at 2ooC; ref. water 20°C)
0.908-0.921 1.448-1.450 248-265 6-1 1 6-8.5 13-1 8 4 5
(nD40) (mg KOH/goil) Wijs method)
(fig oil)
Melting point:
20-28oC
Solidification point
Solids content at
(VF)
0132
10150
20168
25/77
30186
35/95
(YO)
>95
81
64
5
0
0
World market price
Price index (1995 average compared to average) 10 years ago 112% 167% 20 years ago 30 years ago 2 1 0%
1a 2 3 o c
(U.S. $/MTJ
min
0
max
1963-1995
144
501
1433
1963-1 969 1970-1979 1980-1989 1990-1995
144 188 220 275
293 532 598 498
41 1 1224 1433 755
-
propom [%I
Fatty pcid
-
0.4 0.8 5.0 10.0 4.5 8.0 4 . 0 51.0 18.0 -21.0 7.5 10.0 2.0 4.0 6.0 10.0 1.0 2.5 0.6
Cspronc C.~tylic cspric LSIJ~C Myrhtk Palmltic Stbark
C8:O C 8:O c lo:o C 120 C 14:O C 18:O C l8:O obic c 18:l Un&k C 1 8 2 C18:3 C24:l
-
-
-
propom WI
Fig. 4.91, Fact file of coconut oil (fatty acid composition).
4.3.7 Sesame Oil 4.3.7.7 Botany a n d History of Sesame. It has been proven that sesame (Sesamum indicum; Sesamum orientale) was already used by the Persians around 2100 B.C.;the assumption is that it had been in use since about 4000 B.C.Sesame was of immense importance. King Sargon 11, for example, introduced price control and paid out wages partly in sesame seed (Burkhill 1953). The plant was well known to all cultures in Asia and was later exported to other continents. The well-
Vegetable Fats and Oils
267
Fig. 4.92. World market price development of coconut oil. known saying “Open sesame” is derived from the plant and was an invocation to make the pod burst and release the seed. At all times, the oil has also been used as a remedy, especially for intestinal problems. Cosmetic use is also reported. The Greek goddess Hera rubbed herself from head to foot with sesame oil before she seduced the god Zeus (Homer, Iliad). A report from the times before our modem oil production in oil mills was given by Napoleon (war in Egypt, beginning of the 19th century). At that time, the seeds were soaked in water, then roasted for some hours, crushed between milling stones and subsequently extracted by means of lever presses. Today sesame is cultivated mainly in India and China, but also in Sudan and Mexico. The plant requires a temperature of 20°C in the beginning of its development, some humidity during growing, and dry weather during ripening. It can be sowed to altitudes of 1200 m (4000 ft) if a frost-free period of at least 5 mo is guaranteed. Sesame flowers after 45 d, and its oil forms between d 12 and 24 of maturation (Fig. 4.93). Normally, a sesame plant reaches aheight of 6CL120 cm and bears its seeds in capsules 2.5-8.0 cm length with a diameter between 0.5 and 2.0 cm. The thousandseed weight is 4-8 g of white to brownish seeds that are 4 mm long, 2 mm wide and 1 mm thick. 4.3.7.2 Economic lmportance of Sesame. In contrast to the other oilseeds, the main cultivation areas lie outside Europe and the Americas, where the oil is rarely found (Table 4.48). Sesame seed belongs to the high-fat seeds with more than 50% of its dry matter being oil. The yields are low because it is grown mainly in nonindustrialized regions or countries. On small farms, a yield of 300 kgha is normal; the highest yields reported
Fats and Oils Handbook
2 68
Fig. 4.93. Sesame. are 10 times that. The highest yields recorded in 1994 were from Ethiopia (3846 kgha), Israel (1515 kgha) and Lebanon (1455 kgha); the minimum yields were from Sudan, Guinea and Chad (197, 200 and 243 kgha, respectively; Table 4.49). The plant is regarded as risky for farmers because exceptionally low yields can be caused by too much rain; there can also be too much variation in ripeness times TABLE 4.48 Sesame Seed Production (-65% Used for Crushing; Crushing Oil Equivalent -43%) Sesarneseed(MMT)
1935
1950
1960
1970
1980
1990
1993
1994
1995
Total world India China PR Mexico Sudan
1.60 0.40 0.85 0.02 0.03
1.78 0.40 0.83 0.07 0.13
1.44 0.37 0.29 0.12 0.18
1.87 0.44 0.36 0.25 0.20
2.60 0.49 0.45 0.16 0.25
2.35 0.55 0.42 0.07 0.07
2.32 0.57 0.56 0.04 0.18
2.66 0.84 0.54 0.05 0.17
2.76 0.93 0.54 0.05 0.20
TABLE 4.49 Production Yield of Sesame Seed Sesameseed (kglha)
1935
1950
1960
1970
1980
1989
1993
1994
1995
Total world India , China PR Mexico Sudan
360 240 590 420 320
360 200 560 490 670
290 150
330 180 410 830 400
303 185 484 586 286
339 250 583 592 138
342 260 704 930 142
373 340 913 980 126
351 343 900 962 125
-
640 430
Vegetable Fats and Oils
269
Fig. 4.94. Composition of sesame seed.
of the individual capsules or easy bursting of the capsules during harvest. The oil is very expensive so that it is rarely traded worldwide and is used mainly in the countries of production. 4.3.7.3 Composition and Properties of Sesame Seed and Its Oil. Figure 4.94 shows the composition of sesame seed. The oil consists of >75% unsaturated fatty acids, with almost equal portions of oleic and linoleic acid (Fig. 4.95). The solidification point is -3 to -6"C, and the iodine value ranges from 104 to 120. Sesame oil may contain 0.5% palmitoleic acid and up to 0.1% (in special cases 0.9%) of
Fig. 4.95. Main fatty acid composition of sesame seed oil.
Fats and Oils Handbook
2 70
myristic acid. As can be predicted from the fatty acid composition, more than 60% of the triglycerides of sesame oil contain four or more double bonds (Fig. 4.96). The nontriglyceride components of the seed serve as fodder. They contain 35% protein. The plant itself is of no value and is plowed under, at least in largescale farming. Apart from the ash content, sesame meal is very similar to soybean meal (Fig. 4.97). The amount of meal fed to milk cows should be restricted to 1.2 kg/d to prevent a decrease in milk quality due to the carry over of fatty acids. Because the price is high, livestock feeding is rarely done. However, it is used to feed deer because it helps in the development of beautiful antlers. For human nutrition, the meal is of no use because the bitter hull cannot be satisfactorily separated. 4.3.7.4 Harvest and Storage of Sesame Seed. Harvest takes place approximately 100-110 d after sowing (Weiss 1983) with new cultivations after 70-75 d (Mantilla 1977).When harvested, the upper capsules are not yet ripe, but ripen postharvest during drying (Table 4.50). After drying, the seed is threshed. It can also be harvested by means of a combine harvester. Great care has to be taken then not to squeeze the seed. Figure 4.98 shows a processing flow chart.
Sesame oil TriglY-m type POP POS
Po0 SO0 PLP PLS
OOO PLO SOL SLO POL
OOL OLO SLL PLL LOL POLL OLL UL Proportion [% m/n
Fig. 4.96. Triglyceride composition of sesame oil (after Ouedraogo and Bezand 19811.
Vegetable Fats and Oils
Fig. 4.97. Composition of expeller cake and extraction meal of sesame seed.
Fig. 4.98. Flow chart of sesame seed oil production.
271
Fats and Oils Handbook
2 72
TABLE 4.50 Seeding and Harvesting Periods for Sesame Seed Jan
Feb
Mar
Apr
May
June July
s
India China Mexico Sudan
S S
H H
s
s
H
H H S
Aug
Sept
H
H H H
Oct Nov Dec
H H H H
H H H H
H
H H
TABLE 4.51 Fact File of Sesame Oil ~~
~~
~
German: Sesamol
French: huile de sesame
Relative density Refractive index Saponification value Iodine value Unsaponifiable matter
(at 20°C; ref. water 200C) (nD40) (mg KOH/g oil) W i j s method) (fig oil)
Melting point:
-"C
Solids content at
("(3°F)
0132
514 1
20168
(O/O)
75% linoleic acid gives safflower oil a great qualitative importance. Quantitatively, however, it is small and not only because of the high price. Depending on the planting density, safflower grows very branched or unbranched. It prefers temperatures of 24-32°C; -70 days after sowing, it begins to flourish. Seed formation begins 30 d later, and in another 20 d, the plant is ripe for harvesting. The roots of the safflower reach 2-3 m deep; its height varies between 50 and 200 cm, depending on variety and climatic conditions (Fig. 4.1 13). Safflower seeds are 6 1 0 mm in length and resemble those of the sunflower, but with a thicker hull. The yield is 1000-2500 seeds per plant. Earlier varieties had a hull content of some 70%; in recent varieties, this has been reduced to 45%. Because of the enormous proportion of hulls and the fact that the meal has no value at all, interest in safflower will remain limited. However, as a source of high-linoleic acid oil (highPUFA oil), it will be the plant of choice unless new breeds of other plants come up with similar contents. 4.3.10.2 Economic Importance of Safflower. The importance of safflower is in its linoleic acid content, which is required for products with high-PUFA claims. The amount produced is very limited (Table 4.58). The highest yield in 1994 was reported for Argentina (2667 kgha), the lowest for Pakistan (500 kgha). 4.3.7 0.3 Composition and Properties of Safflower Seed and Its Oil. Almost 50% of safflower seed is hull (Fig. 4.114). Where the oil is concerned, one could almost speak about single-acid triglycerides because linoleic acid makes up almost 80% of the fatty acids (Fig. 4.115). In addition to the main fatty acids, traces of linolenic acid (in special varieties up to 3%) and up to 1.5% myristic acid can be found. Because of the high proportion of linoleic acid, almost 90% of all triglycerides contain four double bonds or more, and more than half of it is trilinoleate (Fig. 4.1 16).
Fats and Oils Handbook
284
TABLE 4.58 Safflower Seed Production and Production Yield (-100% Used for Crushing; Crushing Oil Equivalent -1 7%) ~~
Production (MMT)
Production yield (kgha)
Safflower seed
1970
1980
1990
1995
1970
1980
1990
1995
Total world Mexico
0.6 0.25
1.0 0.58
0.9 0.16 0.49 0.17 0.01
0.9 0.11 0.46 0.19 0.01
680
693 1102 381 1267
720 900 549 1786 -708
808 1161 606 1984 700
India
us.
EU (10)
0.13
0.19
0.19
0.14 0.01
-
-
-
Fig. 4.115. Main fatty acid composition of safflower seed oil.
Vegetable Fats and Oils
285
Fig. 4.1 16. Triglyceride composition of safflower seed oil (after Jurriens 1968).
4.3.10.4 Harvest and Storage of Safflower Seed and Extraction of the Oil. The plant is harvested with combines -35-40 days after flowering when it is not yet totally dried (the outer leaves of the corolla are then brown). The moisture content of the seed must not exceed 8% (small storage facilities) and would be better at -5% (essential for silo storage); seeds with higher moisture must be dried. Usually the seed is not stored, but harvest dates are coordinated with the oil mills. The seed is partially dehulled; the hulls are used in the cellulose industry or for the production of insulating material. Its expeller cake and extraction meal are relatively rich in protein (Fig. 4.117). Because of the high portion of hulls, safflower seed contains only 15% oil. Figure 4.1 18 shows the production diagram of safflower seed oil extraction. 4.3.10.5 Fact File of Safflower Seed Oil. The fact file gives the fatty acid composition, the physical data as well as price development (Table 4.59, Fig. 4.119). Further literature is given at the end of this chapter. 6050
40-
30-
a m0 Wmtm
Rotah
Fat
Fiber
Kfm exbrct
Aeh
Fig. 4.1 17. Composition of expeller cake and extraction meal of safflower seed.
Fats and Oils Handbook
2 86
I
Urblctbn (5 2 3)
,,
I
B
t
I
I
Extmdbn mal
C r u d e ~ ~ o i l
Fig. 4.1 18. Flow chart of safflower seed oil production. TABLE 4.59 Fact File of Safflower Oil German: Distelol, SafloroI
French: huile de carthame
Spanish: aceite de cartamo
Relative density Refractive index Saponification value Iodine value UnsaDonifiable matter
(at 20°C; ref. water 20°C)
0.922-0.927 1.467-1.470 186-1 98 130-1 50 45
Melting point
-"C
Solids content at
("VF)
i%)
(mg KOHIg oil) (Wijs method) (dkg oil) Solidification point
-1 3 to -20°C
0132
10150
20168
30186
35195
0
0
0
0
0
World market price (US. $IMT)
1991
1992
1993
1994
-880
-1280
-1580
-1260
Vegetable Fats and Oils
287
Safflower seed oil FanY add
Pmpom1 ~x1 2.0 -10.0 4 0.5 1.0 .10.0 7.0 -42.0 64.0 -81.0 < 1.0 0.5 < 0.5
P w m n PA1
Fig. 4.1 19. Fact file of safflower seed oil (fatty acid composition).
4.3.11 Cocoa Butter
4.3.1 7.1 Botany and History of Cocoa. The cocoa tree (Theobroma cacao) grows mainly in the tropical forests of Africa. The plant stems from Central America and was exported to Africa, which is now responsible for 75% of the world production. In the plantations, the tree is kept at a maximum height of 4 m by clipping. It delivers 20-50 pods that contain 35- 50 seeds each. The weight of a single dried seed is -1 g. 4.3.7 1.2 Economic importance of Cocoa. Cocoa butter, as the fat is called, becomes available during further processing of cocoa. Although a by-product only, it is very valuable. As early as 1809/10, there have been international congresses in Paris that dealt with the protection of cocoa products against imitation. In the year 1970, the world production of cocoa butter amounted to only 130,000 MT, a tiny amount compared with other oils and fats (Table 4.60). However, the cocoa product industry is only a small branch of the food industry, and cocoa butter can relatively easily be substituted by cheaper fats (see Chapter 6.2.3.5). Apart from its role in sweets, cocoa butter is also a source material in the pharmaceutical industry. The naine cocoa butter cannot be used for fat that is produced by the extraction of whole nondehulled cocoa beans; the name is protected for fat that is won by press extraction of cocoa powder. After the common frst pressing, yielding -75% of the fat of cocoa, a hard, partly deoiled mass remains that still contains 22-24% fat. Further vigorous pressing leads to cocoa powders with not >lo% fat remaining. TABLE 4.60 Cocoa Bean Production Cocoa beans (MMT)
1935
1950
1960
1970
1980
1990
1993
1994
Total world Ivory Coast Brazil Ghana Nigeria Ecuador Malaysia
0.72
0.77
1.20 0.09 0.16 0.44 0.20 0.04
1.35 0.15 0.18 0.39 0.22 0.07
1.70 0.43 0.33 0.27 0.17 0.08 0.04
2.47 0.75 0.40 0.30 0.16 0.40 0.26
2.50 0.80 0.34 0.24 0.14 0.08 0.23
2.56 0.81 0.34 0.27 0.14 0.08 0.23
-
0.12 0.28 0.09' 0.02
-
0.14 0.27 0.10 0.02
-
-
-
288
Fats and Oils Handbook
TABLE 4.61 Cocoa Bean Production Yield Cocoa beans (kgha)
1960
1970
1980
1990
1993
1994
Total world Ivory Coast Brazil Ghana Nigeria Ecuador Malaysia
126 105 150 450 215 45
150 240 200 380 215 90
358 529 689 223 241 300 801
464 714 571 333 229 290 969
454 61 8 464 240 337 251 624
439 505 475 264 338 255 62 7
-
-
Highest 1990 yields have been reported from Haiti (2273 kgha), Indonesia (1625 kgiha) and the Solomon Islands (1389 kgha); 4 y later, Haiti was down to 1111 kgha and Indonesia was down to 1167 kg/ha. The record yield of 1994 was that of Sierra Leone (1667 kgha). The lowest yield of all countries that reported to the FA0 came from Liberia (31 kgha), followed by the Central African Republic with only 45 kgha (Table 4.61). 4.3.1 1.3 Composition and Properties of Cocoa Seed and Its Oil. After fermentation, 32-47 kg of fermented dried crude cocoa beans are left from the initial lo00 kg of fresh fruit. The water content of the fresh beans (30-35%) increases during fermentation to -60% and reaches 5 7 % after drying. Then the beans are roasted. Fermented, roasted cocoa kernels reach a fat content of -60% (Fig. 4.120). Apart from the main fatty acids shown in Figure 4.121, up to 1% arachic acid, 0.2% myristic acid and 0.3% palmitoleic acid can be found. The iodine value lies between 34 and 40.The point of gravity of cocoa butter triglycerides is with three configurations that have in common a molecule of an unsaturated fatty acid (oleic acid in nine out of ten cases) in the 2-position. The 1- and 3-positions are occupied by palmitic and stearic acid, single or in combination (Fig. 4.122). After fermentation, the hulls are separated from the kernels by sucking them off. They contain 2 4 % fat that is extracted with hexane for nonedible purposes. Besides technical cocoa fat, the hulls also contain 0.8-1.2% theobromine, a compound needed in the pharmaceutical industry. To obtain this, the extraction meal of cocoa fat extraction is heated to 60°C and again extracted, this time with a mixture of dichloroethane and ethanol. The residue is cleared of extraction solvents and used in animal feeding. The fat extracted from the hulls has an iodine value of 35-65 and a melting point of 28-35°C. It contains >20% unsaponifiable matter (for a comparison: cocoa butter 0.348%). 4.3.11.4 Harvest and Storage, Fermentation of Cocoa and Cocoa Butter Production. After harvest, the beans are left in the pulp of the pod, fermented and
roasted. Only fruit that is ripe enough contains enough sugar for fermentation. The process is the only really successful way to remove the pulp that sticks firmly to the
Vegetable Fats and Oils
289
Fig. 4.120. Cocoa fruit and composition of cocoa beans (photo: courtesy of Karlshamns, Karlshamn.
seeds. Beyond that, flavors or flavor precursors build in the beans that further develop into flavors during drying and roasting. Fermentation takes 4-8 d, with intermixing required after 24-48 h. Unripe or broken beans are sorted out and the rest is dried in the air or in dryers. Cocoa butter production is schematically shown in Figure 4.123.
Fig. 4.1 21. Main fatty acid composition of cocoa butter.
Fats and Oils Handbook
290
MOP. MOS
Fig. 4.1 22. Tryglyceride composition of cocoa butter.
I
c
P
LFalmkn7 Y
4
Fig. 4.123. Flow chart of cocoa butter production.
291
Vegetable Fats and Oils
TABLE 4.62 Fact File of Cocoa Butter German: Kakaobutter
French: beurre de cacao
Spanish: manteca de cacao
Relative density Refractive index Saponification value Iodine value Unsaponifiable matter
(at 15°C; ref. water 20°C) (nD40) (mg KOH/g oil) Wijs method) ( g k g oil)
0.945-0.976 1.456-1.458 19C-200 34-40 2 4
Melting point
32-36OC
Solids content at
("0°F) (Oh)
Solidification point 10/50
20/68
30/86
35/95
82
77
55
3
World market price per MT (mid-1995)
Cocoabutter
21-27OC
-3800 U.S. $
I Proportion [%I
Fatty acid
-
Linoleic Linolenic Arachic
22.6 30.4 0.1 0.5 30.2 36.0 29.2 36.4 C 18:2 1.3 4.0 < 0.5 C 18:3 < 1.2 C 20:O
-
I Proportion [%I I Fig. 4.1 24. Fact file of cocoa butter (fatty acid composition),
4.3.1 1.5 Fact File of Cocoa Butter. The fact file gives the fatty acid composition, the physical data as well as the world market price in 1990 (Table 4.62, Fig. 4.124). Further literature on cocoa and cocoa butter is given at the end of this chapter. 4.3.12 Corn Oil
Corn (Zea mays) is cultivated mainly in the temperate latitudes. It is rarely used for corn oil production. Oil is obtained from the germ, which is recovered during starch production (Fig. 4.125). Corn has very high yields (e.g., 8697 kgha in 1994 for the U.S.) that can be brought to records of 18,678 kg/ha (United Arab Emirates) by irrigation in warm climate. The germ from the corn flour industry contains -20% oil, that from the starch industry -50%. Because corn oil is a tiny by-product only, the worldwide tonnage is very limited (Table 4.63). Oil quality to a large extent depends on the pretreatment and the separation of the germ. The total amount of corn grown worldwide is -450 MMT.
Fats and Oils Handbook
2 92
Fig. 4.1 25. Corn. TABLE 4.63 Corn Oil Production ~~~~~
~
Corn oil (MMT)
1935
1950
1960
1970
1980
1990
1993
Total world
0.10
0.12
0.19
0.33
0.27
0.30
0.33
From Figure 4.126, it can be seen that corn has a total oil content of -6%, all of which is concentrated in its germ. Besides the main fatty acids (Fig. 4.127), it has up to 0.3%each of capric, myristic and palmitoleic acid as well as 0.4%each of arachic, arachidonic, behenic and lignoceric acid. Oils of the northern hemisphere contain 10-13% more linoleic acid than those from the southern. The iodine value is 100-135, and the solidification point lies below -10°C. Distribution of the fatty acids on the different triglycerides of corn oil shows that
Fig. 4.1 26. Composition of corn (maize).
Vegetable Fats and Oils
293
Fig. 4.127. Main fatty composition corn oil.
more than 65% of the triglycerides contain four or more double bonds; more than 40% even contain five or more (Fig. 4.128). Corn oil is regarded as a very stable oil. Depending on the source of the germ, the protein content of the cake can vary between 12 and 21%. Its digestibility is very good. In the market, mainly meal
Fig. 4.128. Triglyceride composition of corn (Strecker eta/. 1990).
Fats a n d Oils Handbook
294
Fig. 4.129. Composition of expeller cake and extraction meal of corn germ.
from solvent extraction can be found; however, it does not play a significant role (Fig. 4.129). The diagram shown in Figure 4.130 also shows the preceding process of starch manufacture. The fact file reports the fatty acid composition, physical data and price development (Fig. 4.13 1, Table 4.64). The low values for lauric and myristic acid given in the Codex Alimenturius can cause problems because unadulterated corn oils have been found that contained up to 0.3% of these acids. Portions that lie above these values suggest adulteration. More information on corn oil can be found in the literature given at the end of this chapter. 4.3.13 Olive Kernel Oil
Olive kernel oil is extracted from the kernels of olives (see Chapter 4.2.2),which contain -12%. Its composition is similar to the composition of olive (pulp) oil (Fig. 4.132). It is usually produced from the olive press cake. If pure olive kernel oil is wanted, the kernels have to be separated from the rest of the cake (rarely done). 4.3.14 Babassu Oil
4.3.14.1 Botany and History of the Babassu Palm. The babassu palm (Orbignya oleifera) grows mainly in some Brazilian provinces. Its lifetime is -200 y, and it begins to carry fruit after 8-10 y. The palm grows -20 m high and has leaves 6 m in length that are pinnated and strongly erect. Because they stand upwards at an angle of -25" to the stem, the plant can be recognized from a great distance. The fruit bunches are up to 100 cm in length and hold 200-600 fruit. They produce 1-6 bunches per year (Fig. 4.133). The fruits ripen between July and November, fall to the ground and are collected. The main harvest period is September-October. The 8-15 cm long and
Vegetable Fats and Oils
Corn
maize)
295
oii contents
5052’C, pH 3.5-4.0, -8 h
Water >>>
Oil contenl-48%
to water content < 50% In p m r
Watering
to water content < 3%, rnuttipb-hlbe dryer
EWadlon(523) I
I
I
CbrlRaUon
Evaporation of (5 2 3 3)
SObnt
!.
I CNde Maize (Corn) Oil
Fig. 4.130. Flow chart of corn oil production.
Corn oil (maize oil)
I Proportion [%I
Fatty acid
C 16 1
Palrnltoleic
Arachic
I
C 20 0
< 05
< 20 < 10 < 05 < 05
Fig. 4.1 31. Fact file of corn (fatty acid composition).
1-
Fats and Oils Handbook
296
TABLE 4.64 Fact File of Corn Oil (Maize Oil) _ _ _ _ _ _ ~ ~
~
German: M a i d Maiskeimol
French: huile de mais
Relative density Refractive index Saponificationvalue Iodine value Unsaponifiable matter
(at 20°C; ref. water 2oDc) (nDm) (mg KOHlg oil) (Wijs method) Cgncg oil)
Cloud point
--1 OOC
Solids content at
(VF)
0132
10150
(Yo)
>> Dehuihd rim
-.
Riw'bran (hulls and gem. oil content 12-18%)
85-1WC, 3 h, insctlvaticn ofenzymes (preferably al the mill) Pelbting
Cooking
Steam
Improve prccintiin during extraction
+
Flaker rolls
(Immersionextmdion)
t
c
I
Extraction solvent fE) >>*
I1
1 -
'I
/I t
I
I
I
Extraction (5.2.3) I
Residue (R)
Mismlla
Evaporation of
Evaporation of
I--
I
Extractibn meal
Crude rice bran oil
Fig. 4.152. Flow chart of rice bran oil production.
TABLE 4.72 Fact File of Rice Bran Oil ~
~
~
Spanish: aceite de arroz
German: Reisol
French: huile de riz
Relative density Refractive index Saponification value Iodine value Unsaponifiable matter
(at 25OC; ref. water 25°C)
0.91 6-0.921 1.465-1.470 181-1 94 98-1 10 3-5
(flD40)
(mg KOH/g oil) (Wijs method) (dkg oil)
Melting point
-5 to -1 0°C
Solidification point
Solids content at
("(2°F)
0132
10150
20168
30186
35/95
(Oh)
20% wt/wt of the fruit bunches entering the mill has to be disposed off as empty fruit bunches. The residues that were formerly burned had found use as a fertilizer. However, the organic part was lost. Now the empty bunches are brought back to the soil. The same is done with the digested sludge from effluent treatment.
350
Fats and Oils Handbook
5.1.2 Olive Oil Production
After harvesting, olives are exposed to enzymic processes that call for immediate processing to avoid fat splitting. Therefore, the oil is usually processed in small plants close to the farms. Olive oil is contained in the pulp as well as in the kernels. Its botany, economic importance and the harvesting methods used as well as its composition are described in Chapter 4.2.2 (pulp oil) and Chapter 4.3.13 (olive kernel oil). Figure 5.4 shows the improvement in the olive oil extraction process (Fig. 5.5) over time. In the old days, the fruit was crushed in the mills (1) and placed into bags (2) or frames (3). Compared with the time when the must was extracted with lever presses (4), today’s hydraulic presses (5) have drastically improved the process. After a time-consuming settling process, the oil was skimmed ( 6 ) and then placed in jars (7) or tanks (8). The oil was separated from the must water by sedimentation and decantation. Another advance in efficiency was made when centrifugal separators (12) were first introduced. Modem plants use automatic preparation of the fruit (13) and must separation via three-phase decanters (14) that separate oil, water and the accompanying solids. Because local processing is still used today, the process is on a much smaller scale than the extraction of oilseeds. There are two
Fig. 5.4. Olive oil production in the course of time (courtesy of Westfalia Separator AC, Oelde).
Extraction of Vegetable Oils
351
Olives Washing, removing foreign material (5.2.1) Crushing (5.1.2.2) (5.1.2.3)
1 Must separation
Expelling, centrifugation, percolation (5.1.2.4)
I Oil separation
Centrifugation (5.1.2.4.1)
I'
Olive oil Fig. 5.5. Processing flow chart of olive oil production. principal methods of oil production, i.e., pressing and separation via centrifuges. Depending on the separation method chosen, the fruit must be pretreated. The choice of method itself depends on whether the fruit were picked from the tree, picked up or mechanically harvested. If they are picked naturally, the proportion of leaves, stocks and other foreign matter is much less than with the two other methods. 5.1.2.7 Washing and Removal of Foreign Material. Washing and removal of foreign material are unnecessary processing steps if the oil is press extracted because the presence of leaves or twigs in the pressing process does not negatively influence oil quality and does not harm the press. If centrifuges are used it is extremely important to carry out these steps because foreign material (except the leaves) can easily damage the centrifuges. Leaves are sucked off by automatic machines. The olives are washed with circulating water.
5.7.2.2 Olive Crushing. The objective of this processing step is to break the cell walls in order to release the oil and to form larger droplets, which can more easily be separated. Metal mills require less energy for the drive than stone mills. Their disadvantage lies in trace metals which may be transferred to the oil, reducing its oxidative stability. Oil produced using metal crushers leaves the disintegration section at an -10°C higher temperature. A taste difference between the two oils is noticeable. 5.1.2.2.7 Stone mills. Stone mills consist of a round granite base block -2 m in diameter that carries a metal vessel consisting of a truncated cone that tapers off to the lower end where it is fixed to the granite (Fig. 5.6). In the upper part, two or three granite milling stones run in an upright position at -15 min-1. The milling
352
Fats and Oils Handbook
Fig. 5.6. Stone mill to crush olives.
stones have a diameter of -120 cm and a thickness of -40 cm. These milling stones mill down the olives, produce the pulp and break the stones. The advantage of granite milling stones is that almost no metal traces can reach the oil, thus prolonging its keepability. In addition, the pulp is not heated, the mechanical strain is low and there is no emulsion formation. The process is discontinuous. 5.1.2.2.2 Metal crushers. The principle of metal mills is that the olives are thrown against a metal wall where they burst. Usually, they are centrifugally flung away from a rotating disk and crushed at the wall of a conical vessel in which the disk is rotating. From there, they slide into the lower part and are conveyed off. This process can be carried out continuously. Therefore, metallic mills are usually used in oil mills that use continuous separation to extract the oil. 5.1.2.2.3 Hammer crushers. Continuous plants also use hammer crushers in which the fruit are crushed and beaten at the same time. Then 30% warm water is added in a mixer before the pulp is fed with an eccentric screw pump to the decanter for must production. Oils produced with a metallic crusher tend to have a more bitter taste (Martinez Moreno et al. 1957). 5.7.2.3 Mixing. After crushing, the paste has to be mixed to allow the fine oil droplets to coalesce into larger ones. This is done by stirrers that are equipped with blades or that are spiral shaped. The effect is shown in Table 5.1. Olive extraction
Extraction of
Vegetable Oils
353
TABLE 5.1 Olive Oil Droplet Size Distribution (%) Depending on Processing Step9 After Crushing Mixing
4 5
15-30
30-45
45-75
75-1 00
>150
6
49 18
21 18
14 18
4 19
6 25
2
aSource: di Ciovacchino (1989).
yield is proportional to mixing time and mixing temperature. Giovacchino (1996) reported a 10% increase in yield by increasing the temperature from -18 to 32°C and a 7% increase by doubling the mixing time from 30 to 60 min. After first mixing, the pulp can be further disintegrated by using machines that resemble meat mincers; they consist of a rotating knife behind a perforated plate.
5.7.2.4 Oil Extraction. In former times, oil extraction was done exclusively with presses. Today, this process is still very common (-50%). To improve the oil separation, processing aids that attack the cell walls and ease the oil flow may be added (Montedoro and Petruccioli 1974). These processing aids are used only when difficult raw materials have to be processed. The headings of the following sections always relate to the frrst (and essential) step of the total extraction process. A second step with centrifugation always follows to separate oil and must (Fig. 5.9). 5.1.2.4.1 Oil extraction with presses. Pressing is done discontinuously in a frame press, which holds a stack of alternately arranged mats and metal disks fmed in their position by a central spine. The olive pulp is smeared on the filter cloth. Such presses hold up to 500 kg of pulp per cycle. One pressing cycle takes up to 2 h; the pressure applied to the olive pulp is -15-20 MN/m2 (150-200 kg/cm2) in single pressing; for double pressing, it is 10-15 MN/m2 (100-150 kg/cm2) in the first step and up to 45 MN/m* (450 kg/cm2) in the second step. The vertical spine is hollow and perforated, thus allowing the must to drain off. A hydraulic pump drives a piston that puts pressure on the pulp by pressing the stack of frames toward the upper part of the press. If double pressing is applied, the pulp has to pass to subsequent presses; the first press applies only about half of the pressure of the second one. Recently, semiautomatic devices have been put in place that reduce labor cost for charging and discharging frame filter presses. 5.1.2.4.2 Oil extraction with centrifuges. Oil extraction can also be done in one step using decanters. This process was made possible by the development of centrifuges that allow for a ratio G of centrifugal acceleration to acceleration by gravity (see Chapter 7.2.6.1, Equations [7.6a] and [7.6b]).
354
Fats and Oils Handbook
Fig. 5.7. Three-phase decanter for olive oil recovery (Source: Westfalia Separator AG, Oelde).
Such machines allow direct oil extraction without prepressing. They manage to separate the solid from the liquid phase in one step and at the same time split the liquid phase into oil and water (Fig. 5.7). The second step of oil extraction, irrespective of prepressing or preseparation with centrifuges, is the clarification of the oil with a separator. In this process, solids are also separated. However, because these are particles that cause only turbidity, they are few in number and no decanter is necessary (Figs. 5.7 and 5.8). If processing is done via centrifuges and if two decanters and two separators that jointly feed one separator are used to polish the oil, capacities of 150 ton of olives per 24 h and per line can be achieved.
Extraction of Vegetable Oils
355
Fig. 5.8. Operational principles and constructional features of separators used in olive oil extraction (courtesy of Westfalia Separator AC, Oelde).
356
Fats and Oils Handbook
TABLE 5.2 Technical Data of Separators for Palm Oil and Olive Oil Clarificationa Polishing centrifuges Maximum throughput Bowl content Solids content (bowl) Bowl speed, max, Energy requ. drive Length Width Height Weight
6,600 7 3 7500 5.5 1265 860 1400 91 5
Separation centrifuges
10,000
12,000
25,000
50,000
150,000
12 6 6500 11 1400 860 1700 1210
12.5 6.5 6500 11 1650 980 1700 1000
26.5 16 5000 25 1500 1300 2200 2375
46 25 4500 45 1950 1400 2100 31 25
125 57 3300 70 2685 2400 2250 6500
Types SA ...006 and SA ...076.Source: Westfalia Separator AC, Oelde.
If centrifugation is used for the separation of the solid phase from the liquid phase, prior dilution becomes necessary. The amount of water added depends on the individual characteristics of the plant but usually lies between 75 and 125% wt/wt of the pulp. To preserve the oil quality, the water added should be tempered (2&25"C). Technical data of centrifuges are given in Table 5.2. Figure 5.10 shows a processing line for continuous olive oil production. This line is in contrast to the more traditional and artisanal approach to olive extraction. The olives (A) are delivered to a receiving hopper and conveyed (2) to the leaf blower (3). There, leaves and twigs are removed while the olives are further cleaned from sand and earth in a downstream washer (4).The clean fruit enter a hammer mill (6) where the entire fruit is ground. The olive paste from (6) is fed to a mixer (7) where the paste is heated to 35°C. Cellular tissues are broken up and the flow behavior of the paste is improved. The pulp is then pumped into a threephase decanter (12) by means of an eccentric screw pump (8). Hot water (up to 30%) is added to ease separation into oil (D), vegetation water (E) and solids (F). The two liquid phases pass through a vibrating screen (11) where light suspended particles are removed. The water (E) is deoiled in a separator (14). Residual water and solids are removed from the oil (D) also using a separator (13), and the clean oil is stored (G). Separated vegetation water may be reused or diluted after having been cleaned. 5 . I .2.4.3 Percolation extraction. The principle of percolation extraction is based on the different adhesion of the systems oiYmetal and watedmetal to the metal. A perforated steel plate is submerged into the must. Because of the different surface tensions, its holes are filled with oil, whereas water is rejected. If a large number of plates are immersed in and drawn off the must, they become coated with oil. Thus the oil is separated from the vegetation water. Such machines carry the name Sinoles system; its principle has been known since 191 1 (Martinez Moreno et al. 1957). Machines that can process up to 350 kg
Extraction of Vegetable Oils
357
pulp per hour have 5120 sheets with a total surface area of -6000 m2. They rotate 7-8 timedmin; they are immersed into the must and the oil is allowed to drip off. The advantage of this process is that it can be automated. However, the amount of unrecovered oil is high. Therefore, this process is most often combined with pressing or centrifugation (Petruccioli 1965; Fig. 5.11). The lower labor rates of this process are offset by the high investment cost so that it is not frequently used.
Fats and Oils Handbook
358
-
Fig. 5.10. Olive oil recovery with decanters and separators (courtesy of Westfalia Separator AG, Oelde).
Olives
Washing and removal of leaves and twigs
usually manyfold (5.1.2.3)
Sheets with 6,000 m' surface area, rotational speed 7-8 rnin.1, 350 kg of pulp I
Must (water content 55-65%)
I
c
c
ccc Water
Centrifugation I
Must
I
Oil Vegetation water
Vegetation water
-+
~
~
-
+
>>> vegetation water
Olive oil
Fig. 5.11. Extraction of olive oil by means of percolation and centrifuges.
Extraction of Vegetable Oils
359
5.7.2.5 Processing the Residues 5.1.2.5.7 Husk (pomace) processing. Smaller oil mills ship the pomace to larger plants for extraction of the residual oil. The husk, which holds up to 55% water, is first dried. After drying in horizontal rotary dryers, the husk consists of the following: water, 6-8%; oil, 5-9%; kernels (seeds), 40-50%; skins, 9-10%; and crude fiber, 20-30%. The dried husk is solvent-extracted to recover the residual oil. The extraction meal can be burned for energy generation or (without kernels) used as an additive to fodder. It may also serve as fertilizer (Fig. 5.12). In many mills, the kernels are separated before oil extraction. The oil content of the husk thereby increases to -18%. If the extraction meal is burned, solvent recovery may be less carefully conducted than when it is used as fodder. 5.1.2.5.2 Vegetation water. In the Mediterranean region, almost 10 MMT of olives is harvested every year. During processing, -4 MMT of vegetation water emerges, creating a major effluent problem. Bernardini (1985) gave details on the composition of the waste water as follows: water, -85% and organic components, -13%, of which oil comprises -0.15% and inorganic components -2%.
5.7.2.5.3Comparison of consumption. Table 5.3 clearly reflects that double pressing allows for the highest oil yield in combination with the lowest amount of
"'"I""
H20 C 15%
Hi0
Anirnalfeed
15%
Revolving furnace
t t t
Residue>>
Fats and Oils Handbook
360
TABLE 5.3 Yields in Olive Oil Production Using Different Techniquesa Waste water
Pulp
Oil yield Water content Oil content Pressing (D = 40 cm) Double pressing Centrifugation PercoIat ion Percolationkentrifugation
(YO)
(YO)
(%)
85-91 89-94 80-88 35-70 81-90
22-28 20-25 40-55 55-65 38-52
5-7 4-5 3.5-5.5 7-1 3 2.5-5
Amount Oil content (%) (Uton olives) 40-60 40-60 100-150 2-6 100-150
0.2-0.6 0.2-0.6 0.41.0 0 . 4 1 .o 0.3-0.8
aSource: di Ciovacchino (1989).
effluent. However, this cannot be the only parameter influencing the decision concerning which process is used. In addition, energy consumption and labor as well as investment cost have to be considered. Total consumption is given in Table 5.4.
5.2 Oilseed Extraction Unlike raw materials that deliver pulp oils, oilseeds can easily be stored over long periods and transported without major risks. Therefore, they are usually extracted close to the location where the crude oil is further processed. In principle, oilseed extraction consists of four basic steps:
1. Seed cleaning and pretreatment 2. Oil extraction (expelling or solvent extraction) 3. Working up the miscella or expelled oil 4. Working up the extraction meal Figure 5.13 shows these steps in more detail in a flow chart of seed oil extraction principles. The machines and plants shown and described in the following serve only as examples. Others exist from different manufacturers. Technical data given for some of the machines are meant to provide an overview rather than present the TABLE 5.4 Energy Consumption in Olive Oil Production Process step Washing and cleaning Stone milling Metal milling Beating Press extraction Centrifugation
(kWhlton) 3-5 10-20 12-1 4 15-1 8 12-20 60-80
Extraction of Vegetable Oils
I
361
CNda on
w-
-
+
+
Fig. 5.13. Flow chart of seed oil extraction.
latest technological developments. For the most current information, the suppliers should be contacted. Besides the technological aspects, running an oil mill consists mainly of the task of managing the mass flow. On its way through the mill, the volume stream is constantly changing (Fig. 5.14). Large oil mills have to manage mass flows of several million metric tons, i.e., up to 10,OOO MT/d. 5.2. I Seed Pretreatment
All oilseeds have to be prepared for extraction. The individual steps required depend on the kind of seed and on the technology chosen. The steps are schematically given in Figure 5.15. Special treatment for single seeds is given below.
Fats and Oiis Handbook
362
Fig. 5.14. Volume flow through an oil mill (after Weber 1987)
Figure 5.16 shows the common pretreatment steps. The dried seed (A) is conveyed with a screw conveyor, which feeds an iron separator ( 2 ) . Impurities are pneumatically removed and discharged via (C). The seed is then dehulled (4), if necessary, and further conveyed (D). In (6), it is reduced and flaked. From the flaker, it is conveyed to the extraction plant (B).
Seeds from Storage Metal separation (5.2.1.1.1) Sieving, air separation (5.2.1.1.2)
a Primary Breaking
Secondary Breaking
Cutting copra. breaking Seeds (5.2.1.3.1)
Corrugating (5.2.13.2)
I Conditioning
Heating (5.2.1.4)
I Flaking
Flaking (5.2.1.3.3)
+
1 structuring
t
Prepared Seed
I
ExtNSiOn (5.2.1.5) ALCON-Process
Oil Extradion
Fig. 5.1 5. Preparation of oilseeds for extraction.
Extraction of Vegetable Oils
3 63
Fig. 5.1 6. Seed pretreatment.
5.2.7.I Seed Cleaning. Oilseeds bear foreign material introduced during harvesting, transportation or storage. These materials may be seeds of weeds, leaves, twigs, and other plant material. Foreign seeds have no influence on the oil quality, whereas leaves and twigs disturb the subsequent processing. Impurities may also consist of sand, earth, stones, and metal parts from harvesting machines or transports. These damage the plant and therefore must be removed, even if they are inert toward the oil. According to U.S. standards (Official Grades of the U.S. Department of Agriculture), soybeans may contain 1% of foreign material to meet grade 1,2, and 3% to meet grades 2 and 3, respectively, and 5% of impurities to be graded 4. Figure 5.17 reflects the situation in a German oil mill during one year. The figure shows foreign seeds and impurities found in shipments of rapeseed and sunflower seed. The figures fluctuate according to the season, the quality of the seeds purchased and the region of origin. They cannot be seen as representative, but exemplary. Figure 5.18 offers a schematic of the cleaning steps through which the seeds have to pass. 5.2.7.1.1 lron separators. Iron separators separate magnetic metal particles from the seeds to protect the machines employed. A tramp iron separator consists of a drum equipped with an electromagnet. The material is fed onto a vibrating chute that is driven by an electric motor and an eccentric shaft. The drum, which consists of nonmag-
Fats and Oils Handbook
3 64
Average amount of foreign material [%]
4 n Sunflower seed
* Railway, Ship Q
Truck
Rapeseed *Railway,
Ship
+Truck
60.
v
50 40
30
20 10 0
netic hard steel, is fed constantly and driven by a geared motor. Inside the drum, a stationary electromagnet induces a magnetic field. This magnetic field captures magnetic metal particles that are separated from the main stream (Fig. 5.19). Iron separators have a capacity of up to 20 tonh (Table 5.5). The machines described in the table have a rotary speed of 1400revolutions/min and the drum diameter is always 300 111111. 5.2.1.1.2 Separation of nonmetallic impurities. Principally, there are two ways to separate nonmetallic impurities. The first is sieving, with the prerequisites of an even-sized seed and a clear size difference compared with the foreign material. The
3 65
Extraction of Vegetable Oils
seeds from storage, or transport
Sbve, collr80
(5.2.1.1.2.1)
(5.2.1 .I.2.1) 3
t
--t
4-
Iron separation
rnagnetkally (5.2.1.1.1)
Blowing or suction (5.2.1 .I .2.2)
Fig. 5.18. Removal of foreign Clean Seed
3
Dehulling or Bnaking
material from oilseeds.
non
Fig. 5.19. Iron separator (courtesy of Krupp MaschinentechnikCmbH, Hamburg).
Fats and Oils Handbook
366
TABLE 5.5 Technical Data
of Tramp Iron Separator+ Capacity (m3/h)
Length Width Width of vibrating chute Length of vibrating chute Magnetic drum length Diameter Drive power Speed Weight
(mm) (mm) (mm) (mm) (rnm) (rnm) (kw) (rnin-1) (kg)
3-5
4-7
6-10
a12
1995 825 300 350 350 300 0.55 1500 400
1995 1025 500 550 550 300 0.55 1500 450
1995 1170 650 700 700 300 0.75 1500 500
1995 1325 800
850 850 300 0.75 1500 550
10-15 2100 1525 1000 1050 1050 300 1.10 1500 650
-
Type MT, courtesy of Krupp MaschinentechnikCmbH.
second is wind sifting (air separation), which requires that the seed and the impurities to be removed have a different specific density. Sieving is done mainly with the use of vibrating screens that are perforated with holes larger than the seed being processed. Thus leaves, twigs and other larger impurities are held back and can be removed. Thereafter, the material passes through a sieve with holes smaller than the seed, which separates sand and other fine foreign material. If trough conveyors (see Chapter 4.1.6) with a sieve-like perforated trough are used, these impurities can automatically be removed during internal transport. Air separation consists of a large fan that blows or sucks off lighter foreign material. Parameters to be considered are the density of the sifting gas (in this case, air) and the material to be slfted. Furthermore, the resistance coefficient of the particles, their resistance area, the speed of the gas and the material and the particle diameter are of importance. Usually, the impurities to be removed are so different from the seeds in terms of density and flow resistance that their separation does not create any difficulties. Figure 5.17 also shows some graphs that indicate what material is sucked off as a function of the suction power. It can be seen that the suction power has to be well controlled; otherwise, too much of the seed is sucked off along with the impurities. To retain as much of the seed as possible, wind screening is therefore often done in two steps. In the first, the impurities are completely removed with a relatively high amount of seed carried along. In the second, the seed is recovered with as low an amount of impurities as possible. 5.2.1.2 Dehulling (Decorticating). To extract the oil of nuts such as palm nuts, coconuts and babassu nuts, the nuts have to be dehulled, i.e., they have to be cracked. For babassu and African palm nuts, this is still done by hand to a great extent. Most oilseeds that could be extracted after size reduction are also dehulled. Extraction without dehulling would be possible; however, the hulls usually do not contain fat and therefore reduce the capacity of the plant. Also, they may contain components that would have to be removed from the oil afterwards or reduce its
Extraction of Vegetable Oils
367
TABLE 5.6 Average Hull Contents of Oilseeds Oilseed
(%)
Soybeans Cottonseed Sunflower seed Peanuts Rapeseed Sesame seed Copra Palm nuts Safflower seed Babassu nuts
7 31 30 47 15 8 1
55 48 68
quality. The economic soundness of dehulling depends on the share of the hull in the total seed weight (Table 5.6). Soybeans, for example, have a very thin hull that does not greatly hamper extraction. Therefore, dehulling used to be uncommon. Today, however, soy beans are very often dehulled because this step is essential if high protein (HP) meal is required. Rape is not dehulled at all, but there are some findings (Schneider 1979, Schneider and Rass 1997, Signoret 1988) that may make it more common in the future. Cottonseed is always delinted and dehulled; the same holds true, naturally, for peanuts. Sunflower seed is also usually dehulled. There are different methods for dehulling, including hammer mills for nuts, rollers, disk attrition mills and many others. In disk attrition mills, the seed is fed via a funnel. It falls onto two vertical corrugated disks. Depending on the seed, the corrugation may be fine or coarse. Disks that are toothed or equipped with usually rectangular and blunt-edged bars also exist (see Fig. 5.20) for purposes such as
Fig. 5.20. Dehulling machine (courtesy of SKET, Magdeburg).
Fats and Oils Handbook
368
peanut dehulling. Depending on the construction principles, both disks rotate in a parallel direction (with a differential speed of --1:1.25; -lo00 seedmin) or one of the disks is static. The opening between the disks can be adjusted. The seed is fed into the center of the disks. From there, the particles are centrifuged to the outside from where they drop and are collected. Table 5.7 gives an indication of sizes and consumption. Rollers follow a different principle; however, they may be similarly constructed. One possibility is a pair of horizontal corrugated rolls that rotate with similar differential speed as described above for the disks. A second possibility is a roll that rotates inside a cylinder; a third is a pair of rolls that is cavitated instead of toothed. The capacities depend on the machine size but also differ from seed to seed. Another technique is dehulling by pneumatic impact. There the seeds are blown against a wall and crack.
5.2. I .2. I Hull separation. Hulls can be removed by screening, air separation and electroseparation. Screening is done via vibrating sieves. Air separation follows the same principle as described for the removal of impurities. Electroseparation is an interesting principle applied, for example, to sunflower hulls. The hullkernel mixture is fed with a vibration feeder to a roll from which it passes close to a corona electrode, which is a fine piece of electrical wire. Because of the different influence of the electrical field on hulls and kernels, the former are more deflected than the latter, thus falling into different boxes. Electroseparators (Fig. 5.21) with a capacity of 2 M T h have a connected load of -3 kW.Their outer dimensions are 2300 x 1750 x 1200 mm.
5.2.7.2.2Dehulling of the individual seeds and nuts. The dehulling process must be adapted to the individual seed or nut. It may differ considerably and may also influence the use of the meal. 5.2.1.2.3 Dehulling of soybeans. Soybeans have to be dehulled if high protein (HP) meal is required. Dehulling performed before extraction is called “head end.” TABLE 5.7 Technical Data of Dehulling Machinesa Peanuts output
MTh
2.1
Length Width Height Disk diameter Drive Weight
(mm) (mm) (mm)
1725 830 930 560 4 730
T y p e AS; SKET.
(mmj
(kWh)
(kgj
2.5 1900 1010 1140 710 7.5 1170
Cottonseed
2.9
900 11
1.7 1725 830 930 5 60 7.5 730
2.3 1900 1010 1140 710 11 650
3.3
900 19.5
Extraction of Vegetable Oils
3 69
Fig. 5.21.
Drawing of electric separation and electric separator (photo: courtesy of SKET, Magdeburg).
If the hulls are separated from the extraction meal, it is called “tail end.” For headend dehulling, the beans are broken and the hull, which is not fixed very tightly to the kernel, is loosened. The hulls are then separated by wind sifting. Beans that are not dehulled are separated from the main stream of broken beans by sieving. Residual hulls are removed from the broken beans in a second round. The hulls are then once more air separated to recover parts of beans that had been blown off with the hulls. The hulls are burned or blended with the meal to standardize it. Good head-end dehulling yields a hull content of dehulled seed el%(see Fig. 5.22). 5.2.1.2.4 Delinting and decorticating of cottonseed (Fig. 5.23). Cottonseed is a very special case because it has to be delinted before dehulling. Lint is made up of the residual woolly fibers that stick to the seed. To remove them, the seed passes rotating disks equipped with sawteeth. The lint is then brushed or blown off from the teeth of these disks. The lint (9697%) is removed in this process; the remain-
Fats and Oils Handbook
3 70 Nondehulledwhole beans
r+l
Dehulling
and Breaking
I
>>> Hulls (with m a l l amount of beans)
I ,
I Sieving
nondehulled
Hulls
broken beans (and hulls)
3
Fuel Additive to meal
Dehulled, broken beans
Fig. 5.22. Dehulling of soybeans.
der has the important function of holding together the hulls that would otherwise easily break into tiny particles during dehulling. Air classification is thus facilitated. The residual 3 4 % of lint is then removed together with the hulls. The separated lint passes a cleaning step to remove filth and bran; then it is pressed into bales. Chemical delinting is an alternative. Diluted acid is sprayed on the cottonseed, which is then dried. The acid is thereby concentrated and loosens the lint from the hull. The disadvantage of this process is that further use of the lint is limited. White seed
I
Black seed a d ISconcentrated and separates linters from seed >>> Linters
3
waste. FUOI
F
ccc Linters
Seed
I
seeds cutto half
pneumatlcslty, to &15% hull content -~
+
>>> Hulls ~
=
Filling nutonal for piastIcd
3
PlOdudlOn of xylore. furfurel, tannin
Delinted dehulled seed
Fig. 5.23. Delinting and dehulling of cottonseed.
Extraction of Vegetable Oils
371
In addition to the lint, dust and any other fine foreign material are removed before dehulling. The capacity and the efficiency of dehulling machines strongly depend on the proportion of lint. If it exceeds the commonly achieved 3 4 % and rises to -lo%, the capacity of dehulling is halved. In the early 1980s, the maximum daily output was -75 MT. Since then, new machines that allow the dehulling of 250 MT of delinted seed daily have been used (Lester 1987).
5.2.7.2.5 Dehulling of sunflower seed. Sunflower seed should be dehulled because the hulls make up 30% of the seed weight. Dehulling is also important because the seed holds a lot of waxes that would be transferred to the oil during extraction. The wax content of oil from nondehulled seed is approximately five times higher than from dehulled seed. However, a small proportion of hulls (60% of the soybeans and >50% of the cottonseed were extracted after extrusion. Extrusion applied after flaking has the advantage of a much higher bulk density. Flaked seed weighs around 550 g/L, whereas flaked extruded seed reaches 850g/L, i.e., -50% more. This significantly increases the plant capacity. Furthermore, extruded seed is mechanically more stable, allows for better solvent percolation and eases desolventizing by using even less energy. Combined with extrusion, a heat treatment can be conducted to inactivate lipases. Applying extrusion technology to soybeans, cottonseed and rice bran has been described by many authors, including Williams and Baer (1965), Bredeson (1983), Farnsworth et al. (1986) and Kemper (1989). Lusas and Wathuis (1988) gave an overview of the state of the art. The authors estimate the savings than can be achieved with this technology to be around $1/MT of seed. Because the seed cells are cracked under the mechanical stress, the extent to which flaking the seed improves solvent extraction has been debated. On the one hand, there are indications supporting this view (Schepinka 1934). On the other hand, the sheer stress applied on the individual cell cannot be that high because the cells are only one fifth the platelet size. Experience has shown, however, that seeds that have passed a corrugated roll have a higher content of oil that can be “washed out.” This free oil obviously plays a role, and there have been many attempts to explain the extraction process. 5.2.3.3.8 Heat treatment. Heat treatment influences the seed components. The water content is slightly reduced and the proteins are denatured. Thus the oil is pre-
Fats and Oils Handbook
406
sent in the cell in a sort of emulsion and can coalesce to form larger droplets. Increased temperature has an identical effect on oil finely dispersed in the cell. It can more easily coalesce as a result of the lower viscosity and eventual cracking of the cells. These larger droplets can be extracted much more easily. 5.2.3.4 Extraction Plants 5.2.3.4.7 Batch extractors. For a long time, batch extractors were the only equipment used for solvent oil extraction. Because of their widespread use and the many different Manufacturers, a great variety of batch extractors exist. Their design principle, however, is always the same. Batch extractors (Fig. 5.52) have a seed inlet (a) and an opening (b) for discharge. The extraction solvent is pumped in via (c) and the miscella sucked off via (d). To evaporate the solvent from the extracted seed, live steam can be blown in (e). The resulting vapors leave the vessel via (f); ventilation is possible (g). As a result of their processing principle, batch extractors are immersion extractors and therefore are equipped with a stirrer. The extraction solvent is always freshly added to the seed, stays in the vessel for the extraction period and is then pumped off. The ratio between solvent and extract (oil) becomes greater with every stage and thus more uneconomical. The meal is desolventized in the same extraction vessel. Batch extraction is no longer applied in high capacity plants; it continues to have some importance in special applications such as the extraction of special pharmaceutical oils or spent bleaching earth (see Chapter 7.3.6). However, viable commercial
B Drive
Fig. 5.52.
Batch extractor (after DGF 1975).
Extraction of Vegetable Oils
407
application would be only for processes with low throughput, i.e., those run only from time to time or in which the application of continuous processes has no payback. Because batch extractors do not allow running the process countercurrently or continuously, they have almost disappeared for bulk oil production.
5.2.3.4.2 Group extractors. Some disadvantages of batch extractors can be overcome if several batch extractors are combined into a group extractor. A semicontinuous process results that also allows for countercurrent extraction. The single batches that together make up the group extractor resemble the batch extractor from Figure 5.52. These single extractors (A-E; Fig. 5.53) are grouped in such a way that the extraction solvent can be pumped from A to B, from B to C and so forth. In addition, miscella can be pumped from the last extractor in the row E into the first one, A. Stage I is the stage that contains fresh nonextracted seed. Stage I1 contains extracted seed, whereas stage 111 contains the furthest extracted seed. The extraction solvent (miscella) is pumped in a countercurrent direction from I11 to I. During this process, A is refilled with fresh seed and extracted seed is discharged from B. Thereafter, in cycle 2, the fresh seed is in A, the most extracted in D; B is refilled and C is the discharge stage. The cycles then continue in the manner shown in Figure 5.53. 5.2.3.4.3 Continuous percolation extractors. As in group extractors, extraction is also done countercurrently in the various continuous extractors. With variations in the details, all of these extractors follow the same principles. After World War 11, the first basket-type extractor was constructed by Bollmann (Fig. 5.54). In this device, the extraction material is fed from the top to the baskets of the extractor. These baskets circulate clockwise. The extracted seed is discharged by turning the basket upside down after almost one cycle. The extracted meal is conveyed to desolventizing. The extraction solvent is pumped onto the fresh seed at two thirds of the height of the left vertical chain of baskets. It percolates through all baskets. This ensures that the furthest extracted seed is exposed to the freshest solvent. Having percolated the vertical left row of baskets, the miscella is collected in a bottom tub and pumped to the top of the right part. There it is sprayed onto the fresh seed. After percolating all right side baskets, it is collected in a bottom tub from which it is pumped to the oil recovery plant. The upper left third of the plant allows the miscella to drain off as much as possible. Another variation is the carousel extractor, manufactured, for example, by Krupp and Blaw Knox. This type of extractor (Fig. 5.55) consists of a sequence of cells that are arranged in the form of a carousel. The cell wheel is housed in a gas-tight cylindrical shell that is horizontally divided by a slit bottom. The cell wheel rotates above this bottom. The extraction material is continuously flowing into (A) from a feeding chute. It is conveyed clockwise over the stationary bottom to the discharge point (C), which is a sectoral opening in the bottom. From there, the extracted meal falls into the discharge chute from which it is continuously removed.
408
Fats and Oils Handbook
Fig. 5.53. Extraction sequence with a group extractor. The miscella collection chambers are formed by the weir-divided subcompartment just below the slitted bottom. They divide the different miscella concentrations from each other. The slitted bottom is formed from profile bars (Fig. 5.56) with wedge-shaped cross sections. As a result of the material gliding over it, the bottom is self-cleaning.The rotational speed of the extractor wheel is adjusted to the needs of the individual extraction material and may vary between 45 min and 30 h. The flow of the solvent miscella begins at the point at which fresh solvent flows onto the seed (B), i.e., close to the extracted material discharge. The solvent
Extraction of Vegetable Oils
409
Extraction
Fresh extraction solvent
Flnal mircelk
,
Fig. 5.54. Bollrnann basket type extractor (redrawn after DGF 1975).
becomes a low-concentration miscella after having passed the first cell of almost completely extracted seed; it is collected in the chamber below the cell. This also ensures that almost all miscella still contained in the extracted seed is washed out. This first miscella is then pumped through all other cells in a counterclockwise manner. The rich miscella is then pumped off (D). Before the extraction meal is discharged, it passes a draining area where as much of the bound solvent as possible is drained off. The lower part of Figure 5.55 shows the concentration of oil in the seed as well as in the miscella while passing the extractor. The darker the shade of gray in the seed, the higher the oil concentration; the lighter the shade of gray in the miscella, the richer it is. Figure 5.56 shows some details of carousel extractors. The equipment may be up to 15 m in diameter, allowing for a capacity of as much as 4000 MT soybeadd. The working volume may reach 380 m3 (-280 ton soybeans). The higher capacity extractors have two cell wheels mounted above each other. In this case, partly extracted seed is discharged from (C) of the upper wheel to (A) of the lower one in order to continue with a second extraction cycle. Table 5.22 gives the technical data of carousel extractors. If the cell wheel of a carousel extractor is uncoiled, in a sense, and brought into a straight line, belt, sliding bed or sliding cell extractors result. Figure 5.57 shows an example of a belt extractor, a deSmet type. This extractor is a horizontal unit equipped with a moving belt that transports the extraction goods countercur-
41 0
Fats and Oils Handbook
Fig. 5.55. Carousel extractor; sectional operating diagram of one-stage extractor and material flow (courtesy of Krupp Maschinentechnik GmbH, Hamburg).
rently to the incoming solvent. There are no walls in the belt to separate chambers from each other. The manufacturer claims that this increases seedsolvent contact, produces fewer fines and makes filtration of the miscella unnecessary. Rakes scrape the top of the bed after each miscella spray. As in the carousel extractor, the miscella from each area of the belt is collected and sprayed again onto less extracted seed. If a chain of connected bottomless frames moves parallel to a belt such as the one described above, a Lurgi system frame belt extractor results (Fig. 5 . 5 8 ; for technical data, see Table 5.23). The framebelt combination moves counterclockwise. The frames separate the seed batches from each other while they are trans-
Extraction of Vegetable Oils
A
= selfcleaning slitted bottom
= model photo = Shipment of 4,000MTtd extractor C D-G = stages of construction = adjustment of slitted bottom D E = collection chamber for different miscellas F = cell wheel with straight separation walls G = extractor finished and erected H = medium size outdoor installation
6
Fig. 5.56. Carousel extractor; details.
41 1
Fats and Oils Handbook
41 2
TABLE 5.22 Technical Data of Carousel Extractor+ Capacity (MTld with 0.8% residual oil and 12% seed moisture) 420
630
900
1250
2000
4000
Diameter (mm) 4500 5300 6000 Consumptiodton of seed Energy (kWh) 2.5 2.2 2.0 Steam (kg) 10.5 10.5 10.5 Including heating of flakes by 10 K in the extractor
7000
8000
9000
11,000
15,000
1.9 10.5
1.7 10.5
1.6 10.5
1.4 10.5
1.3 10.5
200
300
aCourtesy of Krupp Maschinentechnik CmbH.
Fig. 5.57. Continuous De Srnet belt extractor (courtesy of de Srnet S.A., Edegem).
Extraction of Vegetable Oils Chaln drlve
Slew, belt drum
Pumpa
Upper
rleve belt
Lower sievebelt
Frame cells
Dralnagm chambers
41 3 Drainage chambers
Feeding hoppor
Extraction meal discharge
Spraying heads
4'4"' 4'" 4'&$3'
$'"
+&on
Fig. 5.58. Frame belt extractor (courtesy of Lurgi GmbH, Frankfurt). TABLE 5.23 Technical Data of Solvent Extractorsa
Capacity (MT soybeandd) Length Width Height Weight Energy consumption (max) Steam consumption
100 (mm) (mm) (mm) (k@ (kWh) (kg/h)
Capacity (MT soybeandd) Length Width Height Weight Energy consumption (max) Steam consumption
12,600 1240 3525 30,000 8 25 100
(mm) (mm) (mm)
(k@ (kWh) (kg/h)
12,600 1400 3360 28,000 8 25
Frame belt extractor 500 1000
19,900 1830 4450 40,000 17 125
21,100 2775 4430 85,000 22 250
Sliding cell extractor 500 lo00
20,000 2000 4300 62,000 18 125
21,200 3000 4300 70,000 28
250
2000-2400 24,600 3275 5100 125,000 40 500600 2000-2400 26,000 3800 4320 120,000 60 500600
Energy consumption for filling, main drive and discharging; steam consumption without warming of fresh hexane. aCourtesy of Lurgi CmbH, Frankfurt.
41 4
Fats and Oils Handbook
ported on the belt. At the end of the belt, the frames are canied a little further horizontally. Because the supporting bottom of the belt is then missing, the seed falls out of the frame. The lower belt begins further left, so that the falling seed is collected in turned frames for a second extraction. At the right end of the belt, the extracted seed then falls out of the frame into the discharge hopper. Such equipment is no longer built by Lurgi, but still exists in many plants. These machines have been replaced by sliding cell extractors. By replacing the parallel moving belt with a static sieve plate, one obtains a sliding cell extractor. The extraction goods are conveyed by means of sliding cells that are open at their upper and lower sides, i.e., actual sliding frames. This equipment is mechanically much simpler and cheaper and has therefore replaced the frame belt extractors. Only cottonseed may cause some problems in the sliding cell reactor because the slits may become obstructed by cottonseed lint. The latter two types of extractors are usually constructed with nine extraction stages. They are able to handle up to 2400 MT of seedd. The extraction principle resembles that of the extractors described previously. Flakes are fed to a hopper (Fig. 5.59) and countercurrently extracted with the freshest solvent meeting the most extracted flakes; the miscella is pumped counterclockwise onto the seed. The C.M.B. basket extractor follows the same pattern; however, the flakes are conveyed in individual baskets. The baskets are fixed to a chain that passes the extractor in four levels. The miscella then subsequently percolates each of these layers, thus in effect following the countercurrent principle. Put into the context of the complete process, Figure 5.60 shows a plant designed by Lurgi. The seed (1) is fed into a sliding cell extractor (A) and the miscella leaves it at ( 2 ) . The extraction solvent is removed via the columns (B), (C) and (D). The crude oil is degummed in (E) and (F) if desired or necessary. The extraction meal is further processed in (H) and (J) to be marketable. The extraction solvent is condensed (K, L and M), and the vapors are also freed from solvent. The crude oil is pumped to storage (3). Such a plant with a capacity of 2400 MT/d requires a building -40 m in length, 25 m in width and 18 m high. The energy consumption is given in Table 5.24.
5.2.3.4.4 Continuous immersion extractors. The prototype of all extraction plants is the Hildebrandt extractor (Fig. 5.61). Its simplified construction principle is a Ushaped tube through which the extraction goods pass from one end to the other. On their way, the extraction goods pass through the solvent that is fed at the other arm of the tube. Following the principle of communicating tubes, the level of solvent is equally high in both arms of the tube. Therefore, the extraction goods are fully covered with solvent. The Hildebrandt extractor follows the immersion process. The solvent feed is located at a slightly higher position than the solvent outlet, thus guaranteeing that the miscella flows off without pumping. On its way out of the extractor, the miscella passes through a filter. The extraction goods are conveyed through the extractor by means of screws. The residence time in the apparatus is between 1 and 1.5 h. The oil concentration in the miscella never exceeds 13%.
Extraction of Vegetable Oils
41 5
Fig. 5.59a. Sliding cell extractor type Lurgi (courtesy of Lurgi GmbH, Frankfurt). Other extractors are V-shaped (olier extractor). The flaked seed sediments slowly on one side of the V, thus slowly passing through the solvent. The residence time is increased by deflecting sheets. From the bottom part, the flakes are continuously conveyed into a basket elevator by means of a screw. This elevator conveys the extracted flakes to the outlet. The solvent, as always, flows countercurrently. In the U.S., the Anderson-type extractor was frequently used. It resembles the V-
41 6
Fats and Oils Handbook
Fig. 5.59b. Sliding cell extractor type Lurgi (above) and 4000 MT/d extraction plant (courtesy of Lurgi GrnbH, Frankfurt).
shaped olier extractor. However, the extraction goods are fed into a vertical cylinder and removed from the bottom by means of a Redler conveyor. The solvent is fed close to the narrowest point and removed shortly before the place where fresh flakes are fed. A more recent construction employing this principle is the Bernardini Ih4MEX extractor (Fig. 5.62). It consists of a vertical cylindrical vessel with its largest diameter at the top where its walls are parallel. Close to the bottom, they are conical shaped. The extraction goods are fed from the top with a screw
Extraction of Vegetable Oils
41 7
Fig. 5.60. Extraction plant, schematic overview (courtesy of Lurgi GmbH, Frankfurt). feeder, thus ensuring even distribution over the solvent surface. From there, the goods sink through the solvent while being permanently agitated. After extraction, they are conveyed by means of a screw to a bucket conveyor, which discharges the extracted flakes at (B). The miscella flows off at (D) after having been fed at (C). Immersion extractors are suitable mainly for fibrous material, vegetable and animal materials. 5.2.3.4.5 Combined extraction plants. The advantages of percolation and immersion can be combined by sequential use of two plants. Such plants are offered by the C.M.B. company under the name Percolimm. A percolation extractor Simplex 40 (Fig. 5.63) is used in sequence with an immersion extractor IMMEX (Fig. 6.62). The flaked seed is percolation extracted, leaves the extractor with a residual oil content of 5% and is fed into the immersion extractor and then TABLE 5.24 Consumptions for Oil Extraction per Ton of Seeda Plant capacity (MT/d) Consumption/MT of seed
300
600
1500
240-260 260-280 250-270 (k@ Steam (kWh) 18 12 9 Electrical energy 18 18 . 18 Cooling water (m3, 7°C) (k@ 1.3 1.2 1.1 Solvent Data equivalent to plant in Figure 5.61; inclusive drying, cooling and degumrning. Tourtesy of Lurgi CmbH, Frankfurt.
2400 230-250 8 18 1.o
Fats and
41 8
Oils Handbook
Extraction
$4
E3a
Fig. 5.61. Extractor, Hildebrandt type.
desolventized. The combination of the processes avoids the disadvantageous low oil content of ~ 1 3 % in the immersion process, because the immersion miscella is used as a solvent in the percolation extractor. Seeds require 30 min to reduce their oil content to 6 5 % by percolation extraction. After that, an additional 150 min is required to reduce it further to between 0.3 and 0.5% by the immersion process. Because of the different residence times in both extractors, the combined process requires careful planning. 5.2.4 Processing the Miscella, the Meal and the Vapors
The miscella has to be carefully desolventized to produce an edible grade oil and for reuse of the solvent. Independent of the fact that traces of extraction solvent that might remain in the oil would be completely removed during refining, the oil has to leave the oil mill with as little residual hexane as possible. In addition to the solvent from the miscella, the solvent from the vapors that are collected from all over the plant also has to be removed. These vapors are condensed and consist mainly of vapors formed when the extraction solvent evaporates at 50-60°C, together with water from the seed. The meal also has to be freed from solvent not only to fulfill the legal requirements for animal feed but also to prevent explosion risks during storage, bagging and transport. Basically, solvent may escape in five different ways as follows: with the crude oil, with the extraction meal, with the air that leaves the plant, with the effluent (condensed water) and through leakage. The last-mentioned can be prevented by good design and careful maintenance. Losses from cases 1 through 4 can be minimized by
Vegetable Oils
41 9
Fig. 5.62. Immersion extractor (type IMMEX, courtesy of Bernardini 1985).
Fig. 5.63. Basket extractor (Type C.M.B. Simplex-40; courtesy of C.M.B. Pomezia).
420
Fats and Oils Handbook
good manufacturing practice; this is of paramount importance because hexane losses are not only environmentally unfriendly but also costly and constitute a high explosion risk. Figure 5.64 gives an overview on the total process of solvent recovery. 5.2.4.1 Processing the Miscella. Special care has to be taken in working up the miscella because the oil has to leave the oil mill at edible grade quality and also to avoid any explosion risk (Fig. 5.65).
5.2.4.I. I Miscella filtration, Filtration of the miscella is necessary to separate small seed particles and fines. The fines content is significantly higher if the miscella originates from direct solvent extraction (no prepressing). The extraction principle also plays a role. Miscella from the immersion process contains more fibrous material than miscella from percolation plants. Additionally, the type of seed has to be considered. Cottonseed miscella, for example, is much richer in fibers than sunflower seed or soybean miscella. If a considerable amount of fiber has to be removed, a preclarification step with centrifuges may be useful. If filtration is not conducted carefully, fibers occupy the heating surfaces of the evaporation equipment. They act as insulation, reducing the heat transfer and thus the plant’s efficiency. Bernardini (1985) gave the filtration rate depending on the filtration time and the kind of seed (Fig. 5.66). Filtration can be carried out with the usual Niagara filters, which are also used for solvent fractionation. C.M.B. offers a special filter, shown in Figure 5.67. In this filter, the miscella is fed into (F) and passes through the filter plates (C). The filtered miscella leaves at the outlet (H). If the filter cake is thick enough to be discharged, solvent is sprayed (E) on rotating plates that are driven by the motor (I). The filter cake is washed off and collected at the bottom from where it is removed by the conveying screw (D) to leave the filter via (G). 5.2.4.I .2 Oil recovery from the miscella. The percolation miscella has an oil content of 20-30%, thus a solvent content of 70-80%; immersion miscella contain -87-93% solvent. It can easily be seen that considerable amounts of solvent have to be distilled off. With increasing oil content, the vapor pressure of the system increases and thus the boiling point of the miscella rises (see Chapter 5.2.3.2.2). This requires increasing amounts of energy. Solvent removal should be done at low temperature to protect the oil and save energy. Oil quality also suffers if contact time with the heat exchangers is too long, especially for oils of high lecithin content. Therefore a compromise has to be found between low contact time and low temperature. Principally, the solvent is removed in a three-stage operation. In the first step, the miscella passes a pre-evaporator for quick removal of the main part of the hexane. Afterwards, the residual hexane is removed in two steps under always sharper conditions. Figure 5.68 schematically shows a plant for solvent recovery from oil. After passing the filter (l), the oil is preheated (2). There are different designs for such equipment. One possibility is that the oil streams along hot surfaces formed, for
Extraction of Vegetable Oils 42 1
Fats and Oils Handbook
422
0
>>>vawn>>>
200 hPa. usualiy cp. 9%
Evaporation I
Drying
Crude oil
>>>vapors>>>
ahvays d 1 0 ' C
0
100 hPa
,
0
>>>vapors >>>
(Residual hexane < 50pprn; in good plontn < 10ppm)
0 sea figure 5.64
Fig. 5.65. Solvent recovery from miscella.
example, by bundles of steam-heated tubes. Then the oil is fed from the top to the main evaporatdr. There, almost 80% of the solvent is removed at -300 hPa vacuum and at -95°C. To save energy, these evaporators are currently heated by the hot vapors coming from the meal desolventizer. The second step is performed under similar conditions; there, evaporation reduces the solvent to 3 4 % of its initial amount. In a third step, the oil with a residual amount of solvent has to pass through a stripping column (4) with a lower vacuum (usually 200 hPa is suffi-
Filtration capacity [l/m2filtration area]
700_ I 600 Miscella -+
500
Soybean-
* Sunflower-
400
Rapeseed-
300
+ -
.. Peanut+Sesame-
\
200 100 1
2
3
4 . 5
6
7
0
Filtration duration [h]
Fig. 5.66. Filtration rate depending on filtration time and the type of seed (after Bernardini 1985).
Extraction of Vegetable Oils
423
s
Fig. 5.67. Filter for miscella filtration (courtesy of C.M.B., Pomezia). cient). Live steam injection supports evaporation. In a subsequent step, the almost solvent-free oil is dried (5) and the last parts of solvent are removed together with the water. The vapors from the plant-almost exclusively solvent and water-are condensed (6) with cold water or air.
Fig. 5.68. Miscella distillation and condensation plant (courtesy of Lurgi GmbH, Frankfurt).
424
Fats and Oils Handbook
Part of the processing is a steam distillation in which water and solvent are separated. The solvent is reused. Residual solvent is completely removed from the water in a postevaporation step. The vapors from that part of the plant are collected and cleaned (see Chapter 5.2.4.3). Bernardini (1985)calculated that -25% of the energy for solvent evaporation is required to heat the miscella to its boiling point; the rest is for evaporation. To process 1 MT of a 30°C miscella containing 30% oil (solvent boiling point 68"C), -80,000 kcal is necessary. 5.2.4.2 Solvent Recovery from the Meal. The meal coming from the extractor contains 25-35% solveat; it is, in a sense, impregnated with it. The meal must be carefully processed to remove the solvent as completely as technically possible. This desolventization is necessary to fulfill legal demands for animal fodder, to reduce pollution of the environment and to avoid any explosion risk (Fig. 5.69). Residual hexane may cause a problem with some seeds, such as rape. The solvent cannot be steamed out of the cells sufficiently because it is bound to remaining lipids (Schneider 1991). The nature of the solvent removal process is already predetermined in the pretreatment steps before oil extraction. For rapeseed, residual lipid (and thus solvent) is retained in the hull cells that are not broken up in the flaking rolls or in the screw press. Therefore, good seed pretreatment is not only a prerequisite for high oil yield but also vitally necessary for an efficient processing of the meal. In batch plants, this desolventizing step is conducted in the same batch vessels in which the extraction was done. In continuous extraction, special desolventizing plants are required. This has to be done with great care and under protective conditions because the meal can easily be damaged, thus reducing its quality. For some
I
Mal
(R..idulhau*
(n#y~rtmg.mdwm=u
-
b. < 3Qoppm)
Fig. 5.69. Recovery of solvent from extraction meal.
Extraction of Vegetable Oils
42 5
seeds such as soybeans, solvent removal is combined with a heat treatment step called toasting; it destroys the anti-trypsin factors with the aim of improving the digestibility of the meal. Older plants consist of several horizontal stacked cylinders (Fig. 5.70). The meal is fed (A) and screw-conveyed to the right relative to the left. Within the individual cylinders, the meal is conveyed by a system of mechanical paddles; at the end of each cylinder, it falls to the next lower one, thus zigzagging through the apparatus. The cylinders are equipped with steam jackets to heat the meal to -100°C and to evaporate the solvent. In the last cylinder, live steam may be injected to expel the last traces of residual solvent. The desolventized meal leaves the plant at (B); the vapors are condensed with water in (C). More modem equipment consists of a vertical cylinder holding a series of stacked trays. The meal is fed with a screw conveyor and heated on the double bottom of the trays over which it is transported by means of a sweeper. Once the residence time on the individual bottom is completed, a hole in the tray bottom opens and the meal is swept to the lower tray. As in the tube desolventizer described above, steam can be injected into the lowest tray. Figure 5.7 1 shows a desolventizer/toaster/dryer/cooler(for technical data, see Table 5.25). The apparatus consists of a heated, cylindrical housing containing a num-
UrtnCtlDnmufhed
A
I
-I
B orolm(iaddaut*(
Fig. 5.70. Drawing of a horizontal desolventizer (redrawn after Bernardini 1985).
42 6
fats and Oils Handbook
Fig. 5.71. Desolventizer-toaster-dryer-cooler; sectional operation diagram, perforated bottom for desolventizing stage (lower left) and outdoor installation (upper left, courtesy of Krupp Maschinentechnik CmbH, Hamburg).
ber of heated decks that are hermetically separated from each other. Agitator arms move the meal onto the decks and sweep it to the next deck. During the passage of the meal through the cylinder, countercurrent steam is blown in. Toward the end of the process, the meal undergoes hydrothermal treatment that destroys indigestible or harmful substances, thus improving digestibility. The lower part of the apparatus houses the drying and cooling section. This section is equipped with air baffle plates.
Extraction of Vegetable Oils
42 7
TABLE 5.25 Size of Desolvent izers Capacity (MTld)
Diameter (mm)
50 100 2 00 300
1400 1600 1900 2200 2500
5 00
Trays number
Height (mrn)
6 9 9
500 500
9 12
600 600
500
Hot and cold air is sucked in and pressed through the meal bed, removing moisture and heat and some of the meal bed as well. Toasting is necessary for soybean meal (to destroy antitrypsin factors). Before the introduction of 00-rape, the heat treatment during meal desolventizing also reduced the amount of glucosinolates. If combined equipment such as that described above is not applied, the meal should be dried after desolventizing. If it is toasted, it must be dried.
5.2.4.3 Vapor Treatment. Vapors originate from all steps of oil extraction and further processing, but mainly from desolventization. These vapors are collected from all over the plant and condensed. The resulting mixture is separated. The solvent (upper layer) is decanted from the water layer and pumped to storage. Traces of solvent that remain in the water are distilled off by direct steam heating of the water to 95°C. The resulting vapors are fed back to the condenser (Fig. 5.72). The air from the plant also has to be cleaned (Fig. 5.73). It is precooled and washed in an absorption column with pharmaceutical mineral white oil. This oil consists of aliphatic hydrocarbons; it absorbs the hexane, thereby removing it from Vapors
(loaded with extraction solvent)
from all over the plant
ccc Water P
Water
(with traws of extradion solvent)
Separation
settling, decanting
I
+
>>> Extraction solvent 3
Re-use
Waste' water (with traces of extradion solvent)
,
~
t
I apom
Effluent
Fig. 5.72. Solvent recovery
Fats and Oils Handbook
42 8
Waste air (with extraction sokent vapor) from all over the plant I
'I I
I I
I/
*>> mineral white oil (and extraction solvent)
I
Extraction sotvent vapors
Eftluent (free of extraction solvent)
Condensation
/ ~e-uw t Extraction solvent
Fig. 5.73. Waste air cleaning.
the vented air. The mineral oil is then heated in a stripping column where it releases the absorbed hexane. The hexane vapors are condensed and the absorptive is cooled and reused. An alternative to this process is the cleaning of exhaust air by open or catalytic combustion. However, the process is unreliable and expensive. Assuming that the hexane in the exhaust air (40°C) of a 480 todd extraction plant (2.12 kg/m3 air) was burned, a total of 550 ton hexane would be lost yearly, assuming 40m3h of exhaust air (Weber 1972). Based on the same data, Weber calculated the consumption levels required to recover hexane from the air as follows: 100 kg steam, 2 m3 cooling water and 2 kWh electrical energy. 5.2.5 Treatment of Crude Oils
Crude oils usually have to be refined. This can be done in the oil mill in an attached refinery or in refineries belonging to the plants in which the oil is processed further. Degumming, however, which is actually a refining step, is done mainly in the oil mills at present and therefore described in this chapter. One of the reasons for performing the degumming step at the oil mill is that the gums can be sprayed on the meal if the lecithin is not recovered. In refineries, however, the gums are a waste material. Degumming is essential for almost all uses of soybean oil.
5.2.5.7 Degurnrning. Phospholipids, proteins and carbohydrates, vegetable gums and colloidal components negatively influence the keepability of an oil. They are undesirable materials in refining because they increase the neutral oil loss and also hamper other operations. For example, they obstruct crystallization during the fractionation process and clog the pores of hardening catalysts. In salad oils, they would collect at the bottom of the bottle and make the oil appear spoiled.
Extraction of Vegetable Oils
429
For these reasons, oils that contain a certain amount of these substances have to be degummed. Degumming means the removal of the entire group of the abovementioned components irrespective of whether they really are gums or not. 5.2.5.2 The Theory of Degumming. Two kinds of phospholipids exist, those that are hydratable and those that cannot be hydrated. The relative rate of hydration was given by Seghers (1990; P represents phosphatidyl): P-choline (PC), 100; P-inositol (PI), 44; P-ethanolamine (PE), 16; and P-acid (PA), 8.5. Their calcium salts have an even lower hydration rate of 1. Hydratable phospholipids can be removed easily by the addition of water. The process can be conducted rapidly at elevated temperatures or slowly at low temperatures. The temperature, however, has to stay below the temperature at which the phospholipid hydrates begin to form liquid crystals (usually -40°C). Taking up water, phospholipids lose their lipophilic character, become lipophobic and precipitate from the oil. Nonhydratable phospholipids have to be converted to hydratable ones, usually by acidulation followed by neutralization. Traditionally, this conversion was done with acids that are sufficiently strong to hydrate phospholipids without hydrolyzing the triglycerides. Many different acids have been proposed, ranging from hydrochloric and citric acid (Hvolby 1971, Paulitz 1983), nitric, sulfuric and phosphoric acid (Paul 1968) and sulfurous acid (Merat 1955) to oxalic acid (Ohlson and Svensson 1976), which had been proposed for easier effluent treatment. A variety of other acids have also been tested. At present, citric and phosphoric acid are used almost exclusively. Even acid-treated PA and PE still have low hydration rates that may be increased by the addition of alkaline sodium solution. Acids are also often used to increase efficiency (better hydration); in this case, it is again citric and phosphoric acid that are recommended. If sulfuric acid is used instead, the oil can be used for non-food purposes without further refining. Some oils can also be heat-degummed. They are heated to 240-280°C and the gums precipitate. This is called “breaking” the oil. More recently enzymes, in this case phospholipase-A,, have been used to convert nonhydratable phosphatides. The enzyme attacks the 2-ester bond and splits off the relevant fatty acid, thereby easing hydration. The enzyme is sensitive toward heat and can therefore be destroyed by heating the oil. The reaction is best carried out at 6 0 ° C which is the temperature with the highest enzyme activity.
-
5.2.5.3 The Degumming Process. Until the 1920s, soybeans were extracted with a mixture of benzene and alcohol to produce lecithin. Bollmann then showed the way to produce lecithin from extracted soybean oil. Soybean oil contains 1-3% phosphatides. After being hydrated with water, they become lipophobic and precipitate from the oil. The precipitate can easily be separated from the oil by centrifugation because of its higher specific weight. In the past, degumming was a batch process, whereas today it is almost completely done continuously in the big oil mills. The lecithin sludge (also called break material) that is separated via centrifugation weighs about twice as much as the lecithin it contains. Pardun (1979) analyzed a water
Fats and Oils Handbook
430 Water degummlng WDG
T2%?1°'
(for saybans onb, w 5.2.1.3 3)
Crud? oil
Safinco degummlng Van de Moortde D.
Awn soybean oil
Dynamic m i n g -45 mh
3 0 6 0 min
2-3mln
Mino
Water degummed oil (P< 2ooppm)
Degummed ALCON oil
(P< bpm)
UF degummed oil (P < 5 r n )
Fig. 5.74. Simplified flow charts of different common degumming processes (1 = or water degumrned oil; p. = patent and/or process). content of 35-50% and found 15% soybean oil and soy fatty acids and 35-50% crude phosphatides. To recover lecithin from the sludge, it must be dried quickly to prevent further hydrolysis. Usually thin-film evaporation is chosen because a relatively moderate temperature of 80-95°C is sufficient at 70-400 hPa to evaporate the water within 1-2 min. Thus, the lecithin is protected against quality deterioration. If batch drying is applied, a temperature of 60-80°C is necessary at 25-75 hPa for 3-4 h. 5.2.5.4 Silicate Degumming. A new way to remove phosphatides and metals from oils and fats is the application of silicates (see, for example, Welsh and Bogdanor 1986). They reported several trials that yielded good results. The silicate is mixed with the oil at 70-100°C; the mixture is agitated for -30 min and then filtered. A single use of silicate replaces bleaching earth, filter aid and citric acid and is reported to be more efficient than bleaching earth at a lower rate of consumption. Table 5.26 shows a comparison based on the trials of the manufacturer.
5.2.6 New Processes There is ongoing effort to find new ways to extract oils and fats. The primary goal in this search is to eliminate the solvents currently in use. Solvents are not present in the refined oil, not even in minute traces. An increasing number of consumer groups, however, oppose the fact that they are applied at all. One way to avoid the
Extraction of Vegetable Oils Superdegumming Unilever p.
43 1 TOPd
umming
Van de%xtele p.
Crude oil '
h 70.C
CN& oil '
Crude oil ' Heating
TI). GS3, as the first stearin to crystallize, is then separated exclusively at TII. Then, at TIII(T, > TI,,), an additional second stearin, which is GS2U, can be obtained (Fig. 6.7), leaving behind GSU2 and GU3 in the second olein. If different methods of modification are combined, better results can often be achieved. Random or directed interesterification of palm oil before the fractionation step leads to new triglyceride classes so that subsequent fractionation yields
=
Modification of Fats and Oils
453
Palmoil I
Fractionation
GS,, I GSzU
GSU2, I GU3
Fatty acids: S = saturated, U = unsaturated
Fig. 6.6. Distribution of glyceride types in single-stage fractionation of palm oil.
totally different proportions of the fractions (Table 6.1). The very different relative amounts of GS, and GSzU after interesterification compared with native palm oil lead to very different stearins in two-stage fractionation, which allows the separation of these two stearins from each other. Similar results are possible when hardened fats are used. 6.2.1 The Theory of Fractionation
The above introduction shows that even carefully performed multistage fractionation does not allow sharp separation of fat fractions from each other because interactions in the oil/fat mixtures are extremely complex and effects during crystallization become superimposed. 6.2. I. 7 Phase Diagrams for the Solid and Liquid State. For the most part, the influences can be shown only by using model substances, in this case, model mixtures. To study interactions that triglycerides exert on each other, a hypothetical binary mixture of triglycerides that form a eutectic mixture is used. This model system clearly shows the properties and theoretical processes of fractionation. If a binary system of these substances (triglycerides) A and B is cooled down, mixed crystals MC appear (Fig. 6.8). The upper line representing the saturated solution of A in B, and vice versa, is supplemented by a second lower line. These two enclose the area of supersaturated solutions of A in B and respectively, B in A. Because they are unstable, they separate immediately into mixed crystals and saturated solution. The eutectic point separates two other areas from each other, namely, the mixtures of mixed crystals with Palmoil I
GSiU I
I 0u3 osu2,
Fatty acids: S = saturated, U = urwturated
Fig. 6.7. Distribution of glyceride types in double-stage fractionation of palm oil.
454
Fats and Oils Handbook
Fig. 6.8. Phase diagram of a binary mixture.
either the solid phase A or B. The lines between all adjacent areas represent the equilibrium between these areas, i.e., the adjacent phases they represent and the mixtures of these phases. Apart from three definite points t, (melting point of component A), t, (melting point of component B) and tE (eutectic point), the liquid mixtures have a different composition than the mixed crystals that are formed at the same temperature. Depending on the temperature t, the following crystal mixes are obtained: crystals A and mixed crystals MC (te > t >tx, M < E) crystals B and mixed crystals MC (t, > t > tx, M < E) mixed crystals MC ( t = rx, M = E) crystals A and mixed crystals E ( t < tx, M > E) crystals B and mixed crystals E ( t < rx, M > E) In many cases, the components A and B are not able to form a continuous range of mixed crystals. In such cases, a miscibility gap exists in which no mixed crystals of the eutectic type are formed in M, but a mixture of crystals of composition I and I1 (Fig. 6.9). In practice, fats more or less follow the above patterns with some deviation from the theoretical ideal. Nontriglyceride additions that cause an impurity of the system have great influence on the crystallization behavior. They diminish the speed of crystallization and influence the formation of crystal nuclei as well as crystal growth. Mixtures of two such substances A and B cannot be separated into the pure components by single-stage (cooling) crystallization. A sequence or cascade of
Modification of Fats and Oils
Fig. 6.9.
455
Phase diagram of a binary mixture with miscibility gap.
cooling (crystallization) and heating (melting) steps is necessary. One hundred percent purity can be achieved only by an infinite number of such steps. The principle of such a separation is represented in Figure 6.10, which illustrates a hypothetical two-component diagram. Following the diagram, the first step Temperature T (“K) Tx
cx
T2
r
53
100
Fig. 6.10.
Phase diagram of two triglycerides with eutectic point (after Bailey 1950).
Fats and Oils Handbook
456
is the crystallization of stearin S , crystals in the olein 0, liquid at the temperature T, from an initial melt of A and B (composition C,). The crystals S1 still contain some portion of A (as represented in the diagram). If stearin S , with the composition C2 is melted and brought to temperature T2, a second olein O2 and stearin S 2 are obtained. Point S , is closer to B, which implies that the purity of S2 is higher than that of S , . If continued via T2 with stearin S2, composition C3, and so forth to Tx,the pure component B can be obtained in an infinite number of steps. Figure 6.11 shows two binary systems that are formed from mixtures of two single-acid triglycerides. It can be seen that a eutectic point exists for the system tristearate/tripalmitate (-----), whereas the system tripalmitate/trilaureate (-) possesses a miscibility gap. If palm oil and palm kernel oil are mixed, instead of model pure substances, a system of natural oils is created that has a minimum of solids at room temperature for a palm oil concentration of 60%. At a palm oil concentration of 20%, this system has its lowest melting point (Fig. 6.12).
6.2.7.1 The Fractionation Tree. To separate mixtures of different substances, the scheme of the fractionation tree can be applied theoretically (Fig. 6.13). The process works by the repeated fractionation of the oleins (0)and stearins ( S ) of the parent generation of fractions. The neighboring stearins and oleins are combined (OS, SO) and also fractionated again. This process is repeated. For the separation into x fractions x - 1 stages are needed and i=x-1
C i fractionations have to be canied out
i=l
Temperature ["C]
i
75
. . . . . . . . . .
. . . . . . . . .
- - A = Stearate
B = Palmitate -A
= Palmitate
B = Laureate
0/100 20180 40160 60140 80120 Ratio in mtxture (NB)
10010
Fig. 6.1 1. Binary systems of single acid triglycerides.
Modification of Fats and Oils
Solids content [%I
457
Melting point ["C] 36
60
Solids at
4s-
-33
10'C v
30
-30
-
-27
-A-
.
15 -
....*-.._. * - ......... ....
A
20'C
30'C
3S'C ..@..
..-*-.-
This scheme enables good separation and yields high-purity end products. However, following this process is far too costly for the fractionation of edible fats. The aim here is not to produce pure fractions but to obtain end products with a specified functionality. Therefore, the method of limited separation is used as shown in Figure 6.14. This figure also shows a theoretical model because fractionaFat to be fractionated
I
S = Stemfin
0 = Obin
*?, S2
X
01
SY
03
X
#! y! F! $! f2y so + 0s
sso + S(SoC0s)
S7 I
O
S
'
qsoos)+ 00s
010
I
Fig. 6.13. Fractionation scheme for fractionation with high selectivity.
Fats and Oils Handbook
458
Fat to be fractionated
I*
11
SI
-,,
S = Stearin 0 = Olein
01
I Fig. 6.14.
Fractionation scheme for fats and oils.
tion is conducted in this way only in case relatively pure specialties have to be produced. The olein 0, of a previous stage is always combined with the Olein Ox+lthat originates from fractionating S,. This mix is again fractionated. The temperatures of fractionations 1 and 2, of 3 and 4 and so forth are equal. Products A, B, C, . . . are obtained and, at the end, a product that comprises the combined olein of the last and next to last fractionation stage. Fractionation is normally conducted in single or double stage only. This is reflected in Figures 6.13 and 6.14 by the limiting line x------x. For typical standard applications, only stearin I and olein I are separated. Then, in a second stage, without recombination, one of these fractions (rarely both) is fractionated again (Fig. 6.15). These principles do not change if dry or Lanza fractionation from the melt or wet fractionation from a solution is performed.
6.2.7.3 influencing Crystallization. Fractionation from a melt requires the following steps: complete melting, cooling for nucleus formation (Chapter 6.2.1.2.l) , transition into the area of crystal growth (Chapter 6.2.1.2.2)and separation of olein and stearin (Chapter 6.2.2). There are different ways of influencing mass crystallization (Fig. 6.16). The crystallization method can be chosen (either from the melt or from the solution), and the individual stages of the crystallization can be influ-
459
Modification of Fats and Oils
-
Fat to be fractionated
,
tn
Ftadiation S T I II
,
Frndknation O b y l 111
Obin O l S m
Stearin Si-Sn
,
t,
Olein OI-Om
Seatin S A i
Fig. 6.15. Scheme of fats fractionation.
enced. The separation operation itself follows the usual processing and influences the end products by the amount of olein that remains in the stearin. To judge the separation sharpness, the separation curve can be determined. This separation curve is a graph that shows the olein yield actually achieved vs. the olein yield that is theoretically possible. The theoretical amount of olein in such graphs is usually represented by a characteristic parameter of the oil, e.g., the iodine value. If the Liquid singb phase system Solution
I
Starting material
M ~ M
I
supersaturated or s u p e m b d system
Unstable (crystellking)
system
Stable (qstalliied) system
First transition prcdud
second transition produd
End prcduds
Separation
~ i6.1~ 6. .M~~~ of oils and fats (after Sattler 1977).
460
Fats and Oils Handbook
iodine value is plotted against the olein yield, the amount of olein obtained (represented by the iodine value) can be compared with the theoretical value. The graph for palm oil starts with an iodine value IV = 53, which is an average value for palm oil. The unprocessed oil is represented by a 100% olein yield. Increasing iodine values in the olein fraction are caused by separation of an increasing amount of saturated or partly saturated triglycerides and thus stearins so that the obtainable olein yield decreases. Curve (A) represents the theoretical olein fraction. which can be calculated from the iodine values obtained. Area (B) represents the yield that can be obtained by single-stage dry fractionation (6.2.2.1). The areas for Lanza (6.2.2.2) and wet fractionation (6.2.2.3) lie between A and B (Fig. 6.17). The fractionation success can be determined by calculating the ratio of the olein yield obtained and the theoretical maximal yield. This is the so-called separation factor. The most important conditions for optimal crystallization are: Stable crystal formation that can be influenced, Crystal growth that can be influenced, Heat transfer that occurs as quickly as possible. Fulfilling these requirements leads to crystals that combine filterability with minimum olein entrapment. 6.2.7.4 Crystallization. Crystallization is a process that is considerably more complex than melting in terms of the possibilities of influencing it and its course. Crystallization begins when a melt or a solution reaches an unstable condition. This unstable condition can be reached by supercooling or by partial removal of the solvent. The driving force for crystallization is that the system is beyond its equilibrium, and there is a strong irresistible drive to regain the equilibrium, which
Fig. 6.1 7. Separation curve for palm oil fractionation (after Stover et al. 1983).
46 1
Modification of Fats and Oils
is the nature of all processes. Thus, components crystallize as long as the equilibrium is out of balance. Crystal growth (see Chapter 6.2.1.4.2) begins with particles and crystallization nuclei; these may consist of the crystallization substance itself or of foreign material, When crystallization nuclei are seeded into the solution, crystallization takes place more quickly. If the melt is subcooled, i.e., if it is far outside its equilibrium, crystallization can occur like an explosion. Tammann (1932) was the first to research the relationship between crystal formation and crystal growth (see Chapter 6.2.1.4.1 and Fig. 6.18). It is important to steer nucleation in such a way that not too many nuclei are built because this would lead to too small crystals. If the temperature is regulated in an optimal fashion, crystal nuclei of high-melting triglycerides of the fat that build the seeds for crystal growth of the lower triglycerides are formed. To activate this process, moderate cooling rates are necessary and the mixture must be well stirred. Because the heat of crystallization is released, the solution must be sufficiently agitated to avoid local overheating and subsequent melting of the crystals. Optimal agitation is also important to ensure effective heat transfer on the cooled surfaces of the crystallization vessel. Mixing also ensures that, given the relatively high viscosity of the melt, enough material that can crystallize is transferred to crystals or crystal seeds. In the following, the basics of crystallization are summarized briefly. Detailed information on crystallization can be found in Tamman (1903), Matz (1969), McCabe and Stevens (195 1) and in Ullmann’s Encyclopedia of Technical Chemistry as well as in Perry and Chilton (1973) and Lysjanski et al. (1983).
6.2.7.4.7Nucleation. The formation of crystal nuclei in supersaturated solutions has been investigated mainly by Tammann (1903), who described it for different
’
- Viscosity
- Nucleation - - Crystallization
T-150
T- 100
T-50
T
Temperature [K]
Fig. 6.18. Dependency of nucleation rate, crystallization rate and viscosity on the temperature (after Organikum 1974).
Fats and Oils Handbook
462
I
Nucki fom\rtlon
Fig. 6.19.
I
Means of nucleation.
substances. Crystal nuclei can be formed just by themselves (primary nucleus formation) or by external influences (secondary nucleus formation, see Fig. 6.19). There is a maximum nucleus formation rate that is specific for every compound. This maximum is dependent on temperature (Fig. 6.20a). Following Tammann’s rule, the maximum nucleus formation lies -100 K and the maximal crystallization speed 20-50 K below the melting temperature. Similar curves can be obtained for the rate of nucleus formation of triglycerides, e.g., for trilaurin. The two curves for the p- and P’-modification (Fig. 6.20b) were experimentally determined. That for the a-form was obtained theoretically. As could be expected, the sequence in which crystals from the melt are built moves from the thermodynamically most unstable to the most stable form (Fig. 6.21a). Becker and Doring (1935) found an exponential dependency on the degree of supersaturation of the solution. Excessive rates of nucleus formation can thus be avoided by keeping supercooling and, consequently, supersaturation low. Ravich et al. (1946) showed nucleus formation in its relative course (Fig. 6.21b), because it could not be measured absolutely. Nucleus formation is superimposed by the transition of crystals from a more unstable to a more stable modification. The p+P’ transformation rate RT is represented by the dotted line. Crystal nuclei are intermediate structures on their way from an amorphous (unorganized) state to a crystallized (organized) one (Fischer et al. 1948, Frenkel 1947); the entropy of the system decreases. Crystal growth, however, is not a step toward a more organized state; it is the addition of molecules to an already wellorganized crystal surface (Ubbelohde 1937). Nuclei are spontaneously formed and can also disappear spontaneously. For a nucleus to grow to the size of a crystal, it must reach the size of a “critical nucleus.” Its size is determined by the temperature and the deviation of the state of the meltholution state from equilibrium. If a nucleus does not reach that size, it disappears again. Critical nuclei have a size between 0.1 and 1.0 pm (Mullin 1972). Bailey estimated that critical nuclei of triglycerides consist of as few as four to eight molecules. Van Putte and Bakker (1987) studied the nucleus formation of v-crystals of a palm oil fraction that consisted of only fully saturated triglycerides (8.8% of the initial oil). Figure 6.21 shows that a strong temperature dependency exists that is expressed not only in the induction period but also in the formation time of the
463
Modification of Fats and Oils
Number of nuclei / cm3
A
350 300
250
.................................................
200
..................................................
150
..................................................
100
.....................................
Glycerol
Piperine
50 0 -80
0
-40
Temperature
60
40
[“C]
Nucleation [nuclei/time]
B
.....................
* D’-Crystals
.....................
.....
.....................
.....
13 -Crystals
* a -Crystals +.RT = D --> 8’ RT = Rate of transformation
40
50
30
20
10
Temperature [“C]
Nucleation rate of glycerol and piperine depending on the temperature (after Tammann 1903) and nucleation rate of trilaurin (after Ravich eta/. 1946).
Fig. 6.20.
nuclei. The authors also proposed an equation that expresses the dependency on the degree of supersaturation defined as
a=- ‘ A
-S‘
L6.11
CS
where 6 is the degree of supersaturation, C, is the actual concentration, and Cs is the saturation concentration.
fats and Oils Handbook
464
The primary rate of nucleus formation Fp is a function of the degree of supersaturation as follows:
or
where k and A are individual constants. Starting from the nuclei, the crystals grow very rapidly until they have reached a size ten to a hundred times that of their nucleus (Matz 1970). In larger crystals. A
10” Nuclei /
I
d
Nuclei I s. IT? ,-I
og
1108 o7 -1 6 -1 d‘ 1102 Id
-I
1 30 60 90 Time [min]
B
120
,
10.2
0.4
0.6
Supersaturation
Crystal growth bm/h]
1000
I
I
I.. 8-Crystals ....................
-
0
0 20
30 40 Temperature rC]
0
30 Supersaturation
Fig. 6.21. Nucleation rate of fully saturated triglycerides of palm oil depending on temperature and supersaturation (above) and crystal growth of fully saturated triglycerides of palm oil depending on temperature and supersaturation (after van Putte and Bakker 1987)
Modification of Fats and Oils
465
slower mass transfer and restricted dissipation of heat of crystallization hinder rapid growth. This situation can be improved by agitation. Additionally, agitation decreases the initial time between the point at which supersaturation is reached and the initiation of crystal nucleus formation (Fig. 6.22). If the melt is inoculated with crystal nuclei, it must be noted that these can p r e mote crystal growth only if they consist of triglycerides that are also present in the oil to be crystallized. Additionally, they must have the same crystal modifcation that can crystallize at the temperature chosen. Reinders et al. (1932) found that cocoa butter does not crystallize if it is inoculated with tristearin crystals. Tristearin does not occw in cocoa butter. On the contrary, cocoa butter crystallizes spontaneously if oleostearin or palm kernel oil crystals, also present naturally in cocoa butter, are used. The explanation for p-crystals not being formed if inoculation is done with acrystals lies in their different crystal structure (see Chapter 2.3.1). A triclinic crystal cannot grow on the surface of a hexagonal crystal. To initiate growth, the same crystal modifications have to come together. As evidence, Nicolet (1920) showed that tristearin below 56°C crystallizes only in the a-form (m.p. 55"C), even if p-crystals (m.p. 72.5"C)are used for inoculation. In addition to the appropriate choice of inoculation crystals, the complete melting of crystal nuclei, the so-called resistant nuclei, has to be ensured by heating the system above its melting point and holding for an adequate time. Otherwise, these resistant nuclei would become effective during cooling and would prevent controlled crystallization of the melt from a virgin state.
6.2.7.4.2 Crystal growth. Crystal growth depends on many factors such as the temperature of the meltholution, relative to the AT of the solidification point of a Rate of crystal growth bm/min] 2.5 ..........................................
'Asymptote for R
= 00
2-
1 ' : System
/ Water at 27-28°C
CuS04 x 5 HO ,
Crystal growth rate without stirring
............................................
0
1
2
3
4
Stirrer speed R
Fig. 6.22. Dependency of crystal growth on the relative speed between crystal and solution (after McCabe and Stevens 1951).
Fats and Oils Handbook
466
melt and the actual temperature, the degree of supersaturation, the particle size of the crystals and the relative speed between meltholution and the crystal. Usually, one goal of a technical crystallization is to combine a low rate of nucleus formation with a high rate of crystal growth. McCabe and Stevens (195 1) showed the dependency on the relative speed between crystal and solution (Fig. 6.22). This relative speed, U, is of importance because mass transfer to the crystal surface has to be ensured. In addition, the heat of crystallization leads to local warming, which prevents further crystallization and has to be dissipated. The crystal growth curve runs asymp totically, with ro, rl and /3 the constants for each individual system. From the curve, it can be seen that the effect of stirring is limited (in this case to U = 2). This is less important because excessive agitation during crystallization shears the crystals. The fractions of these crystals have the same effect as that of nuclei, i.e., the rate of nuclei formation is unintentionally increased (secondary nucleus formation). Temperature has a much greater influence on crystal growth than supersaturation. Doubling supersaturation doubles the rate of crystal growth; halving temperature increases it a 100-fold (Fig. 6.22). Figure 6.23 shows the increase of viscosity and solids content of palm oil over the time of crystallization. The temperature of the cooling water is given as well as the stirrer speed. 6.2.2 Fractionation Techniques
At present, the following three main processing techniques are applied for fractionation: dry fractionation (without processing aids), Lanza fractionation (with wet100
1
100
Palm oil
.E 2o
8
55
0 0
2
4
8 8 1 Time [hours]
0
1
2
Fig. 6.23. Crystallization conditions of palm oil dependent on cooling and stirrer speed (courtesy of De Smet, Edegem).
Modification of Fats and Oils
467
ting agent), and wet fractionation (with solvents). Each process can be conducted in a very simple or in a highly sophisticated manner, thereby heavily influencing the yield, i.e., the sharpness of separation. Dry fractionation has the advantage that no processing aids are useqthat later have to be removed from the product. This is also environmentally more-friendly because there are no effluents or vapors. For the consumer, at least in countries with high awareness of such issues, there is a psychological advantage that the product has had no contact with processing aids (usually chemicals). Lanza and wet fractionation have the advantage of improved olein yields and superior separation sharpness. The price of this advantage is higher energy consumption (four times higher in the case of wet fractionation). Fractionation, as the total process is incorrectly referred to, can be split into the following two main steps: crystallization and the fractionation itself, i.e., separation of the olein from the stearin. All fractionation techniques consist of these two steps, allowing the combination of different methods of crystallization with different methods of separation. In the following, the three methods of fractionation are described followed by discussions of crystallization and separation of fractions. 6.2.2.1 Dry Fractionation 6.2.2.7.1 The principle of dry fractionation. Dry fractionation is based on the principles described in Chapter 6.2.1. Separation becomes possible because of the different melting points of the components forming the oil/fat. Olein and stearin are exclusively separated by filtration. 6.2.2.1.2 The process. The principle of dry fractionation is very simple. The oil/fat to be fractionated is heated above its melting point. Then it is cooled to the separation (fractionation)temperature and the fractions are separated from each other. The easiest way to do this is to heat the oil, cool it down in the same vessel and decant it. This method is applied, for example, to satisfy local demand in the poorer regions of Africa. There palm oil is heated to 75-90°C and allowed to cool down at ambient temperature to -32°C. Cooling takes a long time because ambient temperatures are between 20 and 30°C and AT is therefore low. Crystallized stearins sediment to the bottom of the vessel and the oil is decanted. Applying this method, the olein yield is -60%. This separation is not very sharp but satisfies local needs. Often the stearin from this operation is processed further by melting it at 80°C and, within 4-6 h, cooling it down in vessels with cooling jackets initially to 35°C and later to 22°C. The second olein, which is obtained in this manner, is pourable to 15°C and is used for cooking. The basic principle of fractionation as described above is also applied in all industrial processes. Only the technological expenditure is higher to achieve better separation via optimized temperature control and filtration techniques. Here also, simple and more sophisticated processes compete, depending mainly on the further use of the fractions (Fig. 6.24).
Fats and Oils Handbook
468
Fat
I
-
t > Fpb + 1OK (Fpb melting point of highest meking triglycsrldes)
Heating up
I I
I I
Cooling
stinlng slowly ( I W m i n ) during nuckation
Resting
at br(crystal-growth); cooling continued to conduct heat of crystallization
L
Filtration
+
-+
t
J
+
+ 4
vacuum-belt or dNm filter, p < p-30 hPa (p= ambknt pressure)
! c
L
c
1
Nz, 510min
1
1 t
t
F
R
cl -+
-+
F = Fittrate R = Solid residue b r = Fradination temperaturn
darin
0 !in Fig. 6.24.
Flow chart of dry fractionation.
In simplest form, crystallization is conducted in a vessel that is equipped with a slow stirrer and heating/cooling coils or jackets. The fat is heated to above its melting point so that it is completely melted and then cooled down. Temperature and cooling rate depend on the end products desired. The crystal mass suspended in the oil is separated into an olein and a stearin fraction. Separation has to be carried out quickly to avoid partial remelting of the crystals. Large plants are designed in such a way that crystal nuclei are formed in a precooling step in a large vessel that feeds several small crystallization vessels in which the crystals are allowed to grow. Thus, one achieves higher efficiency by separating the sensitive step of nucleus formation from the time-consuming step of crystal growth. 6.2.2.7.3Coolingprocesses. Commercial plants are run according to different modified processes that are usually connected with the name of the supplier of the equipment. The processes differ in the way the cooling step is carried out. Cooling influences the kind of crystallization, i.e., whether more or less olein is trapped in the stearin crystals, and also filterability. This in turn influences the fractionation yield,
Modification of Fats and Oils
469
meaning the amount of olein that could be separated in relation to the total. Deffense (1987), for example, managed to obtain 50% more olein in butter fat fractionation by slow cooling than with fast cooling (olein yield: fast cooling 50%, medium 65%, slow 75%). These figures may differ from oil to oil and may even be reversed. The processes recommended by C.M.B. and de Smet follow the principle of fast cooling. The oil is precooled for sufficient crystal nuclei formation. Then it is run via an intense cooler to quickly decrease the temperature. The surface of this cooler is large compared with its volume. This allows for quick cooling even if the temperature difference between the oil and the cooling medium is low. Cooling in the C.M.B. process is performed in three stages (see Chapter 6.2.2.5). Tirtiaux has patented a slow-cooling process usually carried out in vessels with slow stimng. These vessels are equipped with heatingkooling jackets and additional coils. The fat is first melted and then cooled following a special cooling pattern according to a cooling curve that has been elaborated for the individual fractionation operation. The amount of heat removed is varied according to the actual temperature. Heat of crystallization can thus be taken into account and quickly removed. Figure 6.25 shows cooling curves for different oils and fats. All curves show the expected shape, i.e., a temperature increase in the area of crystallization, caused by the heat of crystallization. It also becomes clear that a connection exists between the temperature at which this effect is observed and the duration of cooling.
6.2.2.1.4 Plants for dry fractionation. As noted above, Tirtiaux plants follow the route of slow cooling. The first cooling step with the formation of crystal nuclei takes place in special buffer tanks. From there, the oil is pumped into crystallizers Temperature [“C] Samples -A
= Soybean oil hardened I.V. Q6
. . . B = Fish oil hardened I.V. 103 - - C = Butter fat solidific. p. 28°C
50 40
30
20
- D = Palm oil I.V.
= 53
10
E = Edible tallow solidlfic. p. 46%
0 Fig. 6.25. Cooling curves for fractionation using the Tirtiaux process (after Tirtiaux).
470
Fats and Oils Handbook
in which the cooling is done under automatic control. A schematic of that process is shown in Figure 6.26. The crystallizers are vessels of up to 40 MT, equipped with cooling jackets and cooiing tube coils. Usually the crystallizers feed a vacuum belt filter (Florentine type, see Fig. 6.36). Of course, any other filter type can be used.
Fig. 6.26. Semicontinuous fractionation plant (above) and crystallizers (below; Kellens 1993).
Modification of Fats and Oils
471
The filters allow a throughput of up to 260 MT/d, depending on the type of stearin. To obtain such capacities, each filter has to be fed by several crystallizers. To equip a fractionation plant for 200 MT of palm oil, for example, six to eight crystallizers have to feed two filters. Unlike the process recommended by Tirtiaux, the de Smet process uses rapid cooling. A typical plant consists of four crystallizers feeding one filter (see Chapter 6.2.2.2.3). The oil is pumped from a buffer vessel (A) via a cooler (B) that is cooled by cold water into one of the four crystallizers (C). Once a crystallizer is filled, the cooling program is started. During the whole cooling period, the content of the vessels is kept permanently in motion by means of stirrer blades. The normal cooling time is -4 h, and it is claimed that the process produces medium- sized crystals that can be filtered off easily. The crystallizers feed a continuous rotating vacuum filter (E) where the olein is filtered off through a nylon cloth. Tanks are for hot (H), cold (W) and refrigerated (R) water. The olein is collected in the olein tank (F) and the stearin is melted in tank G (Fig. 6.26). To fractionate 1 MT of product, an indication of energy and water consumption is given by the following figures: Steam (6 bar) Electricity Cooling water (closed circuit, 32°C) Cooling water (used up)
70 kg 30 kwh 20 m3 1 m3
There is practically no product loss. For double-stage fractionation, these figures would be -20% higher. 6.2.2.2. Lanza Fractionation
6.2.2.2.1 The principle o f Lanza fractionation. The so-called "Lanza"-fractionation is based on a patent that was granted to Fratelli Lanza in 1905. It differs from the dry fractionation process by the way in which the olein is separated from the stearin. Separation is not carried out by means of a purely mechanical treatment such as pressing or sucking through a filter; rather, the crystal surface is wetted with a detergent solution or a wetting agent. This solution may contain salts as an additive to prevent emulsion formation. The wetting agent wets the surface of the fat crystals that have been precipitated from the melt. The wetted crystals become hydrophilic and sediment to the aqueous phase. The oil has thus, so to speak, been washed free of fat. Big oil droplets that are totally free of crystals are formed; these droplets coalesce to form even bigger droplets, and, at the end, they form an almost continuous oil phase (Fig. 6.27). The two phases (olein and aqueous phase with the stearin) can be separated via centrifuges as a result of the difference in specific density. Only in the 1970s has the process been picked up and revitalized by Alfa Lava1 who market it as the Lipofrac process. An example of a suitable wetting agent is sodium lauryl sulphate.
472
Fats and Oils Handbook
C: Formation of oil droplets without fat crystals
D: Feeding to separator
Fig. 6.27. Photomicrographs of crystal wetting in Lanza fractionation (courtesy of Tetra Lava1 AB, Tumba). 6.2.2.2.2 The process. The wetting agent is added in different portions. It is essential to stir vigorously during the addition of the first portion of sodium lauryl sulphate. Thus crystal agglomerates are crushed and the trapped oil is set free. Additionally, vigorous stirring is the only way to ensure complete wetting of the crystals. After the addition of the second portion, slow stirring is continued for a time. The oil droplets then coalesce, easing separation by centrifugation. The light (olein) phase is washed to remove all residues of detergent. If a multistage fractionation is conducted, it is not necessary to wash out the detergent completely because more is added again in the next stage. The heavy (crystavwater) phase is heated above the melting temperature of the stearin. The two resulting phases, namely, liquefied stearin and aqueous detergent solution are separated again by centrifugation. With the Lanza process, much smaller crystals can easily be separated from the olein than in the dry fractionation process. In general, this allows for much shorter crystallization times. The Lanza process can be applied to crude as well as
473
Modification of Fats and Oils Fat
t > Fpb+lOK (Fp= mMng point of h i g k t m M r g triglywride) to b, stinlng ( 1 W m l n )
to br (br fndknation-temprature) b = twd (b- tempraturn of detergent solution)
to crush agglomerates and to set fme trapped olein b = tsuoll
R
(twtemperature of Wetpent solution)
F
>>> Effluent
Stearin
Fig. 6.28.
Olein Processingflow chart of the Lama fractionation process.
refined oils and fats. Melting of the fat and crystallization follow the processing routes described before; only the separation step is different. (Fig. 6.28). The separation of the olein and the melted stearin from the detergent solution follows the principle of centrifugation described in Chapter 7. 6.2.2.2.3 Plants for Lanza fractionation. Several configurations of plants exist that all follow the same principle (Fig. 6.29). The oil is crystallized in crystallizers and is conveyed into a premix tank (A). There the wetting agent (sodium lauryl sulphate) is added, which may also contain an electrolyte (magnesium sulphate or sodium sulphate) to facilitate the agglomeration of the oil droplets and prevent emulsion formation. Initial mixing is done vigorously with a knife stirrer to destroy crystal agglomerates and ensure good wetting. Then, in a second tank (B) the mixture is gently stirred with paddle mixers while the oil droplets coalesce. In these two tanks, the lauryl sulphate solution, which is prepared in (C) and stored in (D), is added in portions. The crystal suspension and detergent solution must have the
Fats and Oils Handbook
474
Fig. 6.29.
Lanza fractionation plant (courtesy of Tetra Laval, Tumba).
same temperature to guarantee good results. After -2 min holding time, the mixture is separated via a centrifuge (E). The separated olein is carefully washed with water, and the water is then separated in a washing centrifuge. The steariddetergents solution mixture is melted (F) in a plate heat exchanger, stored ( G ) and also separated by centrifugation (H). The lauryl sulfate solution that has been separated is cooled countercurrently (6) and pumped into storage (4)to be reused. The melted stearin is also carefully washed and pumped into a storage tank. Table 6.2 lists consumption of energy and processing aids as well as the oil losses for a plant for the fractionation of 100 MT/d palm oil or lard. The consumption of electrical energy for palm oil fractionation is higher than for lard fractionation because lard can be cooled with water to 30-35°C; for the cooling of palm oil to 20°C however, electrical energy is required to dissipate the heat.
475
Modification o f Fats and Oils
TABLE 6.2 Energy Consumption for Fractionation Processes Consumption per ton for Lanza fractionation
Palm oil
Lard
Steam (4 bar)
150 30 0.2 0.24.6 0.2-0.7
150 18 0.2 0.2-0.6 0.2-0.7
Dry fractionation
Solvent fractionation
Electrical energy Cooling water Na-laurylsulfate Electrolyte, MgSO,,
(kg) (kWh) (m3)
Na2SO4
(kg) (kg)
Consumption for double-stage fractionation of palm oil Steam (6 bar) Electrical energy Water Solvent Yield: Olein Stearin I Stearin II
(kg)
(Oh)
85 36 1.2 60
(Oh)
40
18
40
12
(kWh) (m3)
(kg)
400 35 1
a 70
6.2.2.3 Wet Fractionation 6.2.2.3.7 The principle of wet fractionation. Wet fractionation is not based on the principle of different melting points but on different solubility of the oiYfat fractions in a solvent at a certain temperature. Although the melting points for POP (37"C), PPO (40°C) and PPP (65.772) differ by only 3 and -28"C, respectively, there is a greater difference in solubility, namely, POPPPOPPP is -1:3.5:0.002. Solvent fractionation leads to much higher separation sharpness, because crystallization can be influenced not only by changing the temperature but also by varying the amount of solvent. Beyond that, if there is olein entrained in the crystals, it is less pure olein because the entrained olein is diluted with solvent. The same holds true for the olein that wets the crystals. Additionally, the crystals may be washed olein free with the solvent if high-purity stearin is needed. Many specialty fats can be produced in the desired purity only by wet fractionation. It is the preferred method, for example, for the fractionation of triglycerides composed of two long-chain and one medium-chain fatty acids. The crystallized and the separated amount is dependent, of course, on the fractionation temperature and time (see Fig. 6.30). 6.2.2.3.2 The wet fractionation process. The main processing step in wet fractionation is the cooling step because this determines the separation and, with that, the final properties of the fractions. Like all other fractionation process, wet fractionation can also be conducted as a double-stage process (Fig. 6.31). The process for separation of solvent and oil is same as described in Chapter 5.2.3 for solvent extraction.
Fats and Oils Handbook
476
Removed solids
I
14
[%I
1
8-
JI
6-
"
1:
-8°C
t
-11°C
I
4-
- 6,5"C
20 0
2
4
6
8 10 12 14 16 18 20 22 24
Time [h] Fig. 6.30. Cooling curve for wet fractionation of palm oil (after C.M.B.). 6.2.2.3.3 Plants for wet fractionation. In principle, a wet fractionation plant consists of a mixing vessel (A) in which fat is dissolved in the solvent that is fed from a buffer. From there, the solution is pumped into a crystallizer (Ba or Bb) where it is cooled to the fractionation temperature. The cooled walls of the crystallizer are permanently scraped to ensure good heat transfer. The crystals precipitated from the solvent are separated by means of a hermetically closed filter (C) (otherwise the solvent would evaporate). Therefore, the whole plant must be constructed to be explosion proof. The stearin is separated, melted and the liquefied stearin is desolventized. The miscella (olein + solvent), intermediately stored in (D), is pumped to the second crystallization stage (E), where it is further cooled and then separated 0again. Stearin II is processed in the same way as stearin I. The solvent is distilled off from the miscella, leaving the olein behind. The solvent is pumped back into the buffer for reuse. Table 6.3 shows a comparison between energy and processing aids consump tion for the double-stage dry and the wet fractionation processes, assuming a plant capacity of 200 MTId. .The different fractionation techniques are compared in Chapter 6.2.2.7. 6.2.2.4 Other Processes. Again and again, attempts have been made to improve the traditional fractionation techniques. Hahn (1978), for example, conducted trials to selectively crystallize stearin on a chilled surface. The melt was kept at melting temperature while the higher melting fractions were crystallized on the chilled surface and the lower melting fractions were enriched in the liquid. To remove the crystallized stearin, the chilled walls were scraped. The trials were conducted with palm oil. The temperature of the chilled surface had to be adjusted in a way that prevented any part of the desired olein to precipitate (-20°C c tF c +20"C). To
Modification of Fats and Oils
477
Fat
I
SObbnt wapmtion I
I Solvent evaporstion I
I
I"
oldin II
Stairin 11
3 Dittlllstion Sbadn I
Fig. 6.31. Double-stage wet fractionation flow chart (above) and plant (below; after Bernardini 1985, courtesy of C.M.B. Pomezia).
measure success, a separation factorf, was defined as the quotient of the areas under the melting curves of the fractionated fat f, and the native fat fo: f, = f J f o
[6.31
Stirring was necessary to compensate for the depletion of the higher melting portion of the melt in the vicinity of the chilled surface and to prevent olein from being trapped in
Fats and Oils Handbook
478
TABLE 6.3 Triglyceride Classes of Dry Fractionated Palm O i l with Rapid or Slow Cooling and Lanza Fractionated Palm Oila ~~~
~
~
~
~~
~
~
~
Olein Fractionation
~~~
~~~~~~~~~~~~
Stearin
Dry, rapid
Dry, slow
Lanza
Dry, rapid
Dry, slow
Lanza
58.4 23.3 60
57.2 22.2 68
57.8 21.7 83
40.7 45.9 40
28.6 50.7 32
28.0 55.3 17
0.7 47.9 44.1 7.3
0.4 40.1 44.7 6.8
0.3 50.2 43.2 6.3
13.8 46.2 34.6 5.4
21.9 47.0 27.1 4.0
42.0 41.4 14.2 2.4
~~~
Iodine value Slip melting point ("C) Yield (%) Triglyceride class (%) s3
u
s2 su2
u3
aSource:Deffense (1 985).
the crystals. It was remarkable that the yield for refined palm oil was higher than for crude palm oil. Until now, this process has not been applied on a large scale. Its advantage would be that it could be used continuously with little effort for automation.
6.2.2.5 Crystallization. Crystallizers usually consist of a batch vessel that is equipped with a stirrer and can be cooled. Because crystal growth heavily influences filterability, separation sharpness and yield, crystallizers have been subject to a constant drive for improvement. A complete overview is given by Mersmann (1982). The simple vessels described in Chapter 6.2.2.1.1 have evolved into devices that are equipped with specially shaped low-speed stirrers. The walls are polished and can be cooled to best dissipate the heat of crystallization. Considerable effort has been made to shorten the crystallization time, to develop the process from a batchwise mode into a continuous one as well as to save energy. The main effort was to maintain the temperature as closely as possible according to the desired cooling pattern. To solve this problem, it was proposed, for example, to cool via external coolers. This allows the use of simple vessels and keeps the temperature difference narrow. With such a vessel (e.g., the Tetra Lava1 system), the oil is cooled in two stages via 35°C to 25-27°C. If the Lipofrac process is used, some detergent solution is added after stage one. The mixture is then pumped into crystallizer I in which it is kept for -30 min. It is then pumped into crystallizer I1 while passing through a plate heat exchanger (PHE) in which the total mixture is cooled down by 3-5°C.After holding in crystallizer 11, it is pumped into crystallizer I11 passing again through a PHE. A plant with this configuration has a capacity of 300 MT/d. This can be increased by adding further coolers/crystallizers. Because of a higher ratio of cooling surface and volume, the use of PHE ensures more uniform cooling compared with cooling in batch vessels. Additionally, the heat of crystallization can be dissipated more quickly and the cooling rate can be adjusted individually to
Modification of Fats and Oils
479
the oil. Alfa Lava1 claims that this form of cooling combined with the Lipofrac process allows for yields as high as 85%. C.M.B recommends three-stage crystallizers. The oil is first cooled with water of 32"C, then again with water of 12°C and, in the last stage, with water of 6°C. The water used is circulated via a chilling unit. The crystallization time of an oil fed at -58°C is reported to be -120 min. The temperature decreases by -20°C. In the second stage, an additional 150 min is needed to cool the oil another 15°C. After another 150 min in stage three, the final temperature of -8°C is reached. In typical plants, every filter is generally fed by several crystallizers. 6.2.2.6 Separation of Olein and Stearin. The separation of olein and stearin is usually done by filtration. Separation sharpness, on the one hand, is influenced by good crystallization, i.e., no entrainment of olein in the stearin crystals and sufficient crystal size. On the other hand, it very much depends on the efficiency of the filtration step. However, one must bear in mind that filtration is able to remove only solids from liquids. Stearin that is dissolved in the olein cannot be separated. Van Putte and Bakker (1987) showed the dependency of dissolved stearin for the case of tristearin in palm oil. There is an almost logarithmic dependency on the temperature. For P-crystals, 0.1% tristearin is soluble at 33°C (fY at 20°C) and 100% at 62°C (56°C). The dissolved part cannot be separated by any means. Figure 6.32 demonstrates the influence of the polymorphic form of the crystal on the separation. Separation by means of filtration is important not only for fractionation but also for bleaching in the course of refining (see Chapter 7.4). In other applications, the aim is simply to separate a processing aid from liquid oil, whereas in fractionation, the goal is the separation of two aggregation states of the same class of substance. The filters are therefore of different construction depending on the need. In this book, therefore, different types of filters are described in the context of their main usage. In fractionation, these are mainly membrane filter presses, vacuum belt filters and rotary drum filters. The construction and working principle of these types are dealt with in the following. The crystayoil suspension can be separated by means of normal filter presses. The filter cloth is typically made from natural fabrics ( e g , cotton) or synthetics (e.g., nylon). The filtration capacity depends on the surface and properties of the cloth, on the filtration pressure and, of course, on the amount, the type and the size of the crystals. One possibile way to increase the yield involves blowing for 5-10 min with nitrogen to blow as much olein from the stearin as possible. Although in this case, the olein is blown through the filter cloth, most of the common plants are designed to suck it through. Two types of filters are used for this purpose, those that use a drum and those with a belt. In a certain sense, the belt is a squeezed, stretched drum, so that the two techniques do not differ in principle. Figure 6.33 shows some membrane filter tests conducted with palm oil.
6.2.2.6.1 Membrane filter presses. Membrane filter presses resemble ordinary filter presses in their construction (see Chapter 7.4, bleaching). However, they are
480
Fats and Oils Handbook
Fig. 6.32. Electron photomicrographs of the polymorphic forms of palm oil SSS crystals; A, p (non-stirred); B, p’ (non-stirred); C, p’ (agglomerated) (after van Putte and Bakker 1987).
also equipped with an elastic membrane ( l ) , usually made from rubber or polypropylene. The filter is filled in the common way (2) with the suspension of stearin crystals and olein. The crystals are filtered off by the cloth. The olein is collected in the drains (3) and leaves the filter via the pipe (4)that is formed by the row of holes in the individual filter plates that are tightly pressed together. When the filter cake has reached a certain thickness, the membrane is blown with compressed air (5). In this way, the membrane is pressed onto the filter cake, squeezing out the residual olein. An almost olein-free stearin can be obtained. Hard and equal-sized crystals are a prerequisite for this filtration method (Fig. 6.34). Because such filters are basically constructed by coupling a number of segments (i.e., a number of filter plates), their capacity depends mainly on the number of filter plates that form the filter. For the fractionation of palm oil, 0.1 MT per hour and per filter plate can be expected. A press with 50 plates thus roughly equals the capacity of a vacuum belt filter (see Chapter 6.2.2.6.2).It must be mentioned that membrane filter presses can be operated only discontinuously because the stearin has to be removed once the cake has reached a certain thickness. For this purpose, the press is opened and the cake is removed mechanically.
Modification of Fats and Oils
481
n
-2
0
5
10 15 Filtration time [min]
20
25
Fig. 6.33. Membrane press filtration tests with palm oil fractions (Kellens 1993).
Brought into the perspective of a plant, the crystal/oil suspension (A) is pumped through the membrane filter press (1). The olein is collected in an intermediate storage vessel (2) from which it is pumped if needed. After the filter cake is thick enough, compressed air (3) is allowed to enter the membrane section. After the olein is squeezed out, the air is released (4) and the stearin is collected ( 5 ) . There it is melted with the use of steam and pumped off (Fig. 6.35). 6.2.2.6.2 Vacuum belt filters. Vacuum belt filters are widely known and used in the industry. They can be adjusted easily to varying needs and can be maintained and cleaned without great effort. The main characteristic of this type of filter is a horizontal filter belt that is stretched between two driving drums. This belt, usually made of stainless steel (but types with rubber belts also exist), passes over a horizontal opening in a.tank that is kept under vacuum. There are also some new versions that use sliding cells instead of the belt. These cells are tight against each other when passing the horizontal section. In the area of the drum, the vacuum is broken so that the filter cake falls off or can be scraped off. Today’s belt filters are housed in a climatized chamber. The belt itself is also kept at fractionation temperature. One of the most common filters is Tirtiaux’s Florentine filter. The suspension (A) is fed by a pump to the feeder (l), which feeds it to the belt. The first section (I) is variable and ensures that no stearin sucked through the filter belt is entering the olein tank. The suspension passing through the filter in this area is collected in (2) and fed to the filter again, jointly with fresh suspension coming from (A). In the next section (11), the filtrate (F) is
Fats and Oils Handbook
Fig. 6.34. Membrane filter press (above) and plant with membrane filter presses (below); courtesy of Tirtiaux S.A., Fleurus.
sucked through the filter into (3). The cake is thick enough in this section to ensure good separation sharpness. Sections I and II can be adjusted easily to individual runs md individual raw materials. In the first part of section II, the suspension is filtered; in &e second part it is “dried,” i.e., the stearin filter cake is sucked dry from the rest of he olein. At the turning point of the belt on the right drum, a scraper (4)removes the
Modification of Fats and Oils
483
Fig. 6.35. Simplified process flow sheet of filtration with membrane filter presses (above, courtesy of Tirtiaux S.A., Fleurus) and high-pressure hydrofilter press (below; courtesy of Krupp Maschinentechnik CmbH, Hamburg).
stearin cake from the belt. The stearin falls into a container (5). The belt, returning to the left drum, is heated by infrared lamps and its holes are blown free of melted stearin. The olein from (3) is pumped off. The stearin is also pumped off after it has been melted in (8) by hot water or steam (Fig. 6.36). The throughput of such filters very much depends on the materials being filtered. Typical throughputs are 4-6 MT/h for palm oil, butter oil and hardened oils and 9-1 1 MTih for beef tallow.
484
Fats and Oils Handbook
Fig. 6.36. Vacuum belt filter Florentine type (right) and plant with vacuum belt filter (left); courtesy of Tirtiaux S.A., Fleurus.
6.2.2.6.3 Rotary drum filters. Vacuum drum filters (Fig. 6.37) consist of a drum that is divided into sections. Its outer wall is perforated and covered with a filter cloth. Drums exist with a diameter of up to 5 m; as an addition to the vacuum system, they may be equipped with outer rolls that squeeze out the filter cake. The cake itself is separated from the cloth by means of scrapers, rolls, chains or strings. The suspension is pumped into the filter trough (F). The filling level is controlled
Modification of Fats and Oils
485
Fig. 6.37. Vacuum drum filter (courtesy of Amafilter b.v., Alkmaar).
by an overflow or sensors. The trough is equipped with a swing stirrer (S), which rotates with a frequency of 20 timedmin to prevent the sedimentation of solids from the suspension. The vacuum filter drum (T) is immersed to the depth of 25-33% of its diameter (1) into the trough, which holds the stearin suspended in the olein. The drum rotates through the trough. The rotational speed RS is variable (0.1 c RS c 1.0 turndmin) and is adjusted to the needs of different products. An adjustable vacuum is effective on different segments of the drum as long as it dips into the suspension (1). The olein is then sucked to the inside and the stearin is collected on the cloth. To prevent the filter cake from coming off and to suck in the olein completely, the vacuum is maintained until the filter cloth is tangentially guided away from the drum (+2 + 3). In the upper part of the drum, it is possibile to wash the stearin by means of a spray head or to squeeze out residual olein by means of a pressure roll (not included in the drawing) and to compress the cake (in area 3). Thorough separation of the cake from the cloth (4) is a prerequisite for good filter performance (Fig. 6.38). The figure shows loosening the cake with the aid of strings that are stretched around the drum together with the cloth. Diverting the strings around the scraping roll, the filter cake bursts off and falls into the collecting vessel. The strings are
486
Fats and Oi/s Handbook
Fig. 6.38. Vacuum drum filter, principal drawing and cake removal (courtesy of Amafilter b.v., Alkmaar).
adjusted to the appropriate distance from each other, passing through a comb. This construction minimizes wear on the filter cloth and allows the use of fine cloths because the stress applied to the cloth is low. Rotary drum filters are available in different sizes, i.e., different filter areas. AMA offers filter areas of up to 80 m3. The filtration capacity for palm oil with
Modification of Fats and Oils
* ;
487
Palm oil fractions Viscosity
Olein yield
SFC cake
3
5
7 9 Sampling time [hours]
11
Fig. 6.39. Vacuum filtration tests with palm oil fractions (Kellens 1993).
20% stearin, for example, is 200-300 L/(m2 ' h). Figure 6.39 gives the result of some vacuum filtration tests carried out by de Smet. 6.2.2.6.4 Comparison of filtration techniques. The different separation techniques result in different costs for investment or for operating the plant and also offer different sharpness of separation. Figure 6.40 shows the separation factors for
Fig. 6.40.
Separation sharpness in of palm oil fractionation (after Stover e t a / . 1983).
488
Fats and Oils Handbook
a dry-fractionated refined palm oil that has been separated by different means. As a comparison, some points reflecting the separation factor of the wet and the Lanza process are given. The higher separation factor with membrane filter presses must be paid for with a 10-15% higher investment. Automation that would lead to lower running cost is much more difficult to achieve than with continuously running belt or drum filters. The different separation sharpness is caused by the different pressures that can be applied. Suction filtration is promoted by a vacuum of 200 Wa (maximum), whereas the pressure in membrane presses is much higher. Trials with Krupp presses allowed for pressures on the membrane of up to 8 bar without the crystals being pressed through the filter cloths (Stover et al. 1983). In summary, it can be said that the advantage of rotary drum filters is their low investment cost combined with sturdy construction and easy handling, including easy cake discharge. Membrane filter presses allow for a sharper separation, i.e., higher olein yields with better quality. They have lower energy consumption with easy filtration and drying, combined with high-speed filtration. The fatty matter is protected from contact with oxygen from air. 6.2.2.7 Comparison of the Fractionation Techniques. Kreulen (1976) and Deffense (1985) gave a comparison of palm oil fractions that were produced with different fractionation methods (Fig. 6.41). The figure shows that the end products differ considerably. The solids content of Lanza-fractionated palm oil is approximately twice that of the dry-fractionated stearins. However, these also differ from
Fig. 6.41. Solids content of palm stearin depending on the fractionation process applied (after Deffense 1982).
489
Modification of Fats and Oils
TABLE 6.4 Olein Entrainment in Stearin Cake of Different Fatty Matters (Standard 6-bar Membrane Filter)a Solid fats content, SFC ( O h ) Origin of fattv matter Palm oil IV 52
Palm oil in mid-fraction IV 49
Soybean oil Hydrogenated 1V 75 Milk fat Tallow Lard
Slurry
Cake
14 20 29 15 24 31 31 33 11 16 13 12
58 55 51 73 61 55 64 74 54 42 44 41
Olein entrainment (YO)
42 45 49 27 39 45 31 26 46 58 56 59
aSource: Kellens (19941, courtesy of de Smet, Edegem.
each other because the slowly cooled product contains only 50% more solids than the rapidly cooled one. Therefore, the fractionation technique chosen heavily influences the end products. Because one cannot usually switch between these processing varieties inside a given plant, the decision how to build the plant must always be based on the desired properties of the end products. The different fractionation techniques also deliver different triglyceride composition in the olein and stearin. Duns (1985, Table 6.4) compared products obtained from dry and Lanza fractionation of palm oil. The results reported in the following originate from trials that were conducted on the equipment of Tirtiaux (dry fractionation), Alfa Laval (Lanza fractionation, Lipofrac) and C.M.B. (Bernardini, wet fractionation). Table 6.5 shows the processing conditions that have led to the results represented in Figures 6.42 and 6.43. In the figure, 0 refers to crude olein and S to crude stearin. Wet fractionation has been conducted as a two-stage process (indexed I and 11). TABLE 6.5 Process Conditions for Palm Oil Fractionation Trialsa Wet Process conditions Starting temperature Cooled down to Cooled down to
("C) ("C) ("C)
aSources: Tirtiaux. Alfa Laval and Bernardini.
DW
Double
Triple
70 40 20
45 30-33 20+10
7+4+2
490
Fats and Oils Handbook
TABLE 6.6 Data of Palm Olein from Different Fractionation and of Tallow from Different Separation Technique9 Olein characteristics Yield Palm oil (fractionation techniques) Dry fractionation Fast cooling during crystallization Slow cooling during crystallization Lanza fractionation Wet fractionation"
Iodine value
Solidification point ("C)
58.0
24.0
57.8 57.6
22.2 24.2
(Old
Iodine value
Solidification point ("C)
77-82 43-48 67-72 37-42
50.0 55.5 50.0 55.1
39.0 20.5 39.5 22.2
(Old
6C-63 67-72 77-83 60-63 Yield
Tallow (dry fractionation separation techniques) Membrane filter press (olein D.P. 37-38°C) (olein D.P. 20-22°C) Rotary drum filter (olein D.P. 37-38OC) (olein D.P. 20-22°C) aSource: Deffense (1 985) and Kokken (1 990).
Product characteristics changes during fractionation:
Fig. 6.42. influence of different fractionation processes on the fatty acid composition of the fractions and on the olein properties.
Modification of Fats and Oils
49 1
Fig. 6.43. Influence of different fractionation processes on the triglyceride classes of the fractions.
6.2.3 The Properties of Fractionated Oils and Fats
As mentioned in the introduction to this chapter, fats and oils are being modified to obtain products that are not available in sufficient quantity or to replace products that have a very high price. This also holds true for fractionation. Particular efforts have been made to find a substitute for cocoa butter. Attempts have been made to improve the use of palm oil (solidification point 2 7 4 3 ° C ) at lower temperatures, i.e., to enable its use as salad oil. Today, more than 50% of the palm oil production is fractionated. Taking the absolute tonnage, cocoa butter replacers are a nonplayer in comparison; however, this part of the process has great economic importance due to the high price of the CBE. There is a trend to produce tailor-made fats from “common” raw materials, without using processing aids. There is also a tendency to open new applications for traditional raw materials such as butter. However, the main focus clearly lies on palm oil. Many new fractionation plants have been built in the countries of origin, reflecting the dramatic increase in palm oil production. Because one of the aims was to supply the local markets with the olein, the amount of palm stearin offered in the world market has increased significantly as a consequence. Often modification techniques are combined, because this is the only way to achieve’thedesired results. In the trials described in the following, S, and 0, stand for the stearin or olein that result from a fractionation at temperature x. Saturated fatty acids are always represented by the letter S , whereas U refers to unsaturated ones. The opportunities in fractionation are almost endless, taking into account different fractionation temperatures and single to multistage processes. The deliberately chosen examples in the following can give only a rough impression of the range of
492
Fats and Oils Handbook
Fig. 6.44.
Properties of tallow fractions depending on the fractionation temperature (after Deffense).
products that can be obtained. The different ways of influencing the end products are illustrated for different raw materials. Figure 6.44 offers a view of the wide possibilities of changing the properties of Argentinean beef tallow, depending on the fractionation temperature. As can be expected, the iodine value of the olein is inversely proportional to this temperature. Olein yield and melting points must change proportionally. Naturally, the trend for stearins is the reverse of that for the oleins. 6.2.3.1 Lard. As described in Chapter 6.4, lard is often interesterified to improve its crystallization behavior. Improving these properties also increases its yield in fractional crystallization. Table 6.7 and Figure 6.45 show an example (Deffense 1987) of a double-stage fractionation and the properties of the resulting end products. Of course, the interesterification step before fractionation does not lead to a change in iodine value, but to a significant change in the way fatty acids are distributed over the glycerol backbone. The main shift is a decrease of the types POS and SPO by -40% and an increase of SOO, SLS and OLA by -60%. To give a clearer picture of the effect of these modification techniques, only four global triglyceride types are shown in Figure 6.46, namely, S3, S,U, SU, and U3.A dramatic change in the proportion of these types can be seen, making clear the influence of these processes. If the olein 03*is further fractionated, a whole range of secondary products can be obtained. If the second fractionation temperature is kept below 22'C, fully saturated triglycerides disappear from the product almost completely and S3
Modification of Fats and Oils
493
TABLE 6.7 Data of Lard and Its Fractions from Double-Stage Fractionationa Melting point
Iodine value (IV) Lard, native Lard, interesterified Stearin S32
63.9 63.4 50.4 67.3 55.9 58.7 59.6 69.2 70.6 73.2
Olein 03, 032-s22 032-s1
8
032-s14 O32-%2 O32-Ol8 032-01 4
Yield
("C)
(%F)
37.7 34.7 47.3
99.9 94.5 117.1 85.1 101.7 98.1 93.9 72.0 64.4 59.2
29.5 38.7 36.7 34.4 22.2 18.0 15.1
(O/o)
(% of initial)
20.2 79.8 18.3 31.4 43.1 80.7 68.6 56.9
Code N
U 20.2 79.8 15.4 25.0 34.4 64.4 54.9 45.4
H
A B C D E
F C
aSource: Deffense.
increases heavily in the stearin (Fig. 6.47). The properties of the different fractions become even clearer when their solids content is plotted against the fractionation temperature (Fig. 6.48).
6.2.3.2 Beef Tallow. Apart from lard, beef tallow is the most fractionated animal fat. The differences in the physical properties of the fractions are much greater than in lard. This becomes apparent in Figure 6.49, which shows the proportion of some fractions depending on the fractionation temperature. If a three-stage separation is conducted according to the fractionation tree principle, Argentinean beef tallow Lard
(m.p. = 38'~; I.V. = 63.9) Interesterlficstion
I.V. = Iodine value m.p. = Melting point i.a. = of initial amount
Intelustrrifkd lard (m.p. = 35'C; I.V. = 63.9) Fractionation32'C
(m.p. = 48%; I.V.
= 50.4)
(m.p. = W C i I.V. = 67.3)
I
Fradbwtlon 18%
I 31% (25% La.) Stearin @zSi:
-
(m.p. 36%;
I.V. = 58.7)
I
i
69% (55% La.) Olein 031-0011 (m.p. = 18'C; I.V. = 70.5)
Fig. 6.45. lnteresterification and double-stage fractionation of lard (after Ricci Rossi and Deffense 1984).
Fats and Oils Handbook
494
Fig. 6.46.
Proportion of triglyceride classes of lard (trials from Table 6.7).
can be separated into four products of considerably different properties (Fig. 6.50). Fractions obtained at different temperatures are shown in Figure 6.51, reflected by their fatty acid composition. Table 6.8 gives the fractionation results for doublestage fractionation of tallow at a constant first temperature and various second temperatures. The solids content of part of these fractions is given in Figures 6.52,
Fig. 6.47. Proportion of.triglyceride classes of lard and its double stage fractions (trials from Table 6.7).
495
Modification of Fats and Oils
Solids content
Lard
[%I
vv
Fraction. = interest.
-I
+ A = 032 = S32
-H
+ B = O32-S22
*D
= O32-S14
+E *G
= 032-022 = 032-014
5
10 15 20 25 30 35 40 45 50 Temperature
[“C]
Fig. 6.48. Fatty acid composition of lard and its double stage fractions (trials from Table 6.7). 6.53, and 6.54. With the use of this technique, products can be obtained that contain very different portions of solids at equal temperature. 00-fractions can be used as liquid shortenings, 0s and SO fractions as shortenings. The SS part is usually used up in the soap industry. Its melting temperature lies significantly above body temperature making it unsuitable for human consumption.
6.2.3.3 Butter Fat. The only fractionation process suitable for butter fat fractionation, provided that the butter flavors have to be retained, is dry fractionation. Edible tallow
PerCBntaws based on M P I amount
224% StaannV
Fig. 6.49. Multistage fractionation of edible tallow (after Gander 1969)
Fats and Oils Handbook
496
-
EdiMe tallow (AfgenUnian) (m.p. = 45.1%; I.v. = 51.2)
68%Obin 0 1 6 (m.p. = 38.2'C; I.V. = 49.0)
(m.p. = 41.9.C:
34% SteatInSw (m.p. 50.8.C I.V. = X.0)
I.V. = 42.2)
63% (42% 1,s.) Obln C k d h (m.p. = 22.3.C: I.V. = 54.2)
Fig. 6.50.
I.V. lOdhmVdue m.p. =Melting point 1.a. = of I n W amount
(m.p. = u.9.C;
I.V.
= 42.0)
59% (20%I.&) Stearin Swslr (m.p. = 0.2'c;
I.V. = 30.6)
Three-stage fractionation of edible tallow (after Deffense 1984).
Often, this is conducted as a double-stage process (Fig. 6.55 and Table 6.9). Unlike most vegetable fats, the properties of butter fat can vary greatly, depending on the season as a result of a different fatty acid composition (see Fig. 3.3). The stearin fraction can be one third higher in winter, with a predictable effect on the olein. Deffense (1987) described the olein yield as a function of the fractionation temperature and the cooling method. Figures 6.56 and 6.57 show the proportions of important fatty acids and the solids content for different fractionation temperatures. The cooling rate (At/t) also influences the olein yield and thereby the composition. The above examples reflect processing conditions using a medium cooling rate. With the application of rapid cooling techniques, the olein yield decreases by -10%; with slow cooling, it increases by the same amount.
Fig. 6.51,
Fatty acid composition of different fractions of Argentinean tallow.
497
Modification of Fats and Oils
TABLE 6.8
Characteristics of Tallow and Its Fractions from Double-Stage Fractionationa Melting point
Yield
Iodine value, IV
("C)
("F)
Tallow
45.0
45.1
113.2
Olein 03,
49.0
36.2 40.8
31.6 24.6 20.0 16.7 12.5 9.8 34.4 41.4 46.0 49.3 53.5 56.2
(% of initial)
Code
-
-
n
66.0 47.9 37.4 30.4 25.3 19.0 14.8 52.1 62.6 69.6 74.7 80.9 85.2
-
a Aa Ab Ac Ad Ae Af Ag Ah Ai Aj Ak Al
036-s22
42.2
41.9
036-s24
41.7 41.3 39.6 38.3 55.2 54.2 53.8 52.4 51.8 51.4
42.7 43.4 44.2 44.7 10.5 T 22.3 24.5 26.6 29.0 30.4
97.2 105.4 107.4 108.9 110.1 111.6 112.5 50.9 T 72.1 76.1 79.9 84.2 86.7
36.1 30.6 30.6 29.4 27.1 42 .O 42.0 40.9 40.0
50.8 53.2 53.3 54.0 54.7 44.9 45.4
123.4 127.8 127.9 129.2 130.5 112.8 113.7
34.0 58.8 57.1 47.6 32.6 41.2 42.9
20.0 19.4 16.2 11.1 14.0 14.6
Hb Hc Hd He Hf
46.4 47.5
115.5 117.5
52.4 67.4
17.8 22.9
Hg Hh
O 3 6-s2 0
036-s26 036-s28 O 3 6-s2 9 O36-O20 O36-O22 O36-O24 O 3 6 -O 2 6 O 3 6 -O 2 8 O36-O29
Stearin s 3 6 s36-s45 s36-s46 s36-s47 s36-s48 s36445 s36446 s36-047 s36-048
-
-
H Ha
aSource: Deffense.
The stearin fraction of butter oil can be used in bakery products such as puff pastry and croissants; its plasticity is high. The olein fractions can be used in cake and (legislation permitting) cheese making. Use as the oil portion in melanges is also possible as well as applications in cream fillings for cakes and sweets.
6.2.3.4 Palm Oil. The fractionation of palm oil has increased in line with its greater availability in recent years (see Chapter 4.2.1). Worldwide, -5 MMT are fractionated. Palm oil offers fractionation opportunity because it contains approximately equal proportions of low- and of high-melting triglycerides. It can be separated into an olein, which can serve as a salad oil, and a stearin. In addition to this separation, cocoa butter-like fractions can be produced. Depending on the fractionation technique, different oleins and stearins are obtained. Some typical end products are listed and discussed in the following (Fig. 6.58). Interesterification before fractionation yields an even larger number of different products because of the change in triglyceride composition during interesterification (Table 6.10).
Fats and Oils Handbook
498
Edible tallow
Solids content [%]
100
Fraction 'A- Ha
= s36-s45
*H
= s36
+N
= Native
'A
=Os
80
.
..
A..
.*H*
~
-.:'.
.. Ha
:
,
.
. ..
.
.
.
."Ag = 036-020
5
10
15
20
25
30
35
40
Temperature ["C] Fig. 6.52. Solids content of tallow olein 0 3 6 and its double-stage fractions (trials from Table 6.8).
In many applications, slightly hardened palm oil is preferred because hydrogenation is much cheaper than fractionation. However, hydrogenated products have some practical drawbacks. Combining hardening and fractionation, the advantages of both processes can be merged, and interesterification can also be
Solids content [%] I
5
10
15
20
25
30
35
40
Temperature ["C] Fig. 6.53.
Solids content of tallow and its double-stage fractions (trials from Table 6.8).
Modification of Fats and Oils
499
-Solids content
[%I
100
2nd stearins
60 -
40 -
20 -
0
I 15 5
20
10
25
30
35
40
Temperature ["C]
Fig. 6.54. Solids content of tallow stearin s36 and its double-stage fractions (trials from Table 6.8). Butter oil
(m.p. = 32'C;
I.V. = 37.8)
I.V. = Iodine value m.p.= Mfflng point i.a. =of infflal amount
FraCtiOnstion 2O'C
40% Stearin S20
60% Obin 0x0 (m.p. = 1vci 1.V. = 42.5)
(m.p. = 40'C; I.V. = 31.3)
Fractlonatlon 1O'C
39Oh (23% i a ) ' Stearin 02oS10
61% (37% i.8.j Olein Ozc-010 (m.p. = 1o'C; I.V. = 47.8)
(m.p. = 24'C; 1.V. = 37.5)
Fig. 6.55.
Double-stage fractionation of butter oil (after Deffense 1987).
TABLE 6.9 Characteristics of Butter Oil and Its Fractions from Double-Stage Fractionationa Iodine value (IV)
Melting point ("C)
(OF)
Yield (%)
(% of initial)
-
N
23 77
A B
Butter oil
37.8
32
90
-
Olein 02,
42.5 37.5 47.8 31.3
18
64 75
60 39 61 40
020-s20 020-s20
Stearin S,, aSource: Deffense (1987).
24 10
50
40
104
-
Code
C H
Fats and Oils Handbook
500
Proportion [%]
Fig. 6.56. Table 6.9).
Fatty acids content of butter oil and its double-stage fractions (trials from
conducted. The stearin Sz0 of interesterified palm oil is 225% of its proportion in native palm oil. Melting point and cloud point of the olein are decreased, whereas the iodine value increases (Fig. 6.59 and Table 6.1 1). Figure 6.60 demonstrates the influence of fractionation on the melting point. This is caused by the change in the proportion of the triglyceride classes. In the olein fraction, the S3 triglycerides disappear for the most part (Fig. 6.61). The super olein has by far the highest proportion of triglycerides with two or three Butter oil
-N =
Native
- - H = S20
Solids content
ou
[%I
7
50 -
‘\
H
30 20 ,
,
100
5
\\
‘. 10
15
20
25
30
35
40
Temperature [“C]
Fig. 6.57.
Solids content butter oil and its double-stage fractions (trials from Table 6.9).
Modification of Fats and Oils
Palm OH (nrtive) (m.p. = 3O.S’C I.V. = 55.7) Fndknrtkn 3O’C
501 I.V. = lodim V d U m.p.= W n g point c.p. 8 ckud point 1.r. =dfn#*lunamt
20% Stearin Sm
801(Oblnom
(m.p. = 51.5’C; I.V. = 38)
(C.P. E 5’c;, 1.v.= 80.5)
irc
I
~ndbnrtfon I
-
wdll -17 (rn.p. = 2 T C ; I.V. 53)
30% (24% I.. )
Fig. 6.58.
(md&‘)
-
-17
(ap. 5’C: I.V. e3)
Double-stagefractionation of palm oil (after Haraldsson 1987).
unsaturated fatty acids. The solids content of the fractions listed in Table 6.11 is shown in Figure 6.62. For the mid-fraction of palm oil, PORIM demands an iodine value of 32-55, a melting point of 2340°C and triglyceride compositions as follows: C5d(C48+ C54) > 4 and C5*< 43. Naturally, these figure are valid only far Malaysian products. To serve as a starting material for cocoa butter replacers (see Chapter 6.2.3.5), these mid-fractions also require further modification. The fractionation trials shown previously start with equal fractions from stage I and reflect the influence of stage II. Table 6.12 shows the influence of different stage I temperatures with constant \ stage I1 separation temperature.
6.2.3.5 Cocoa Butter Replacers (Cocoa Butter Equivalents). The paramount property of cocoa butter is its melting behavior, which results from its triglyceride composition (see Fig. 4.122). CBE (cocoa butter equivalent) or CBS (cocoa butter substitute) offer choices in substituting for cocoa butter. A cocoa butter equivalent is defined as a fat that can be used in any proportion to replace cocoa butter in any recipe. Its physical and chemical properties are identical to those of cocoa butter. A cocoa butter replacer is a fat that must be used to replace at least 75% of cocoa butter in “chocolate” recipes. Its chemical and physical properties differ from those of TABLE 6.1 0 Palm Oil Triglyceride Classes Before and After Directed lnteresterification Proportion (mol YO)
s3
s2
u
su2 u3
Native
lnteresterified
7 49
32 13
38
31
6
24
Fats and Oils Handbook
5 02
Palm oil (native) (m.p. = 30.5'C; I.V. = 55.7)
lnteresterified palm oil (m.p. = 52°C; I.V. = 55.7)
I.V. = Iodine value m.p.= Melting point c.p. = Cloud point i.a. = of initial amount
Fractionation 20°C I
45% Stearin szo (m.p. = 58'C; I.V. = 34.4)
55% oiein 020 (c.P. = 2.7'C; I.V. = 76) I Fractionation0'C 73% (40% iaj Olein O2o-08 (C.P. = 4 ' C ; I.V. = 81.7)
27% (15% i.a.j Stearin Ozo-Sa (m.p. = 35'C; I.V. = 62.3)
Fig. 6.59.
Fractionation of interesterified palm oil (after Haraldsson 1987).
cocoa butter. Loders Croklaan company have listed the following requirements for CBE: a melting range equivalent to that of cocoa butter a fatty acid and triglyceride composition close to that of cocoa butter compatibility with cocoa butter processing of chocolate products is identical to that for cocoa butter-based products a polymorphic behavior that does not hinder tempering. It should therefore crystallize in the same polymorphic form as cocoa butter, i.e., in the P-modification the appearance and bloom-free shelf-life of chocolate products containing CBE should at least be identical to products based on cocoa butter alone adequate color and good flavor stability TABLE 6.1 1 Characteristics of Palm Oil and Its Fractions from Double-Stage Fractionationa Iodine value, IV
Melting point ("C)
Palm oil Olein O,, 020-S17 mid fraction O,,-S,, mid olein
55.7 60.5 53 63
30.5 5T 27 3T
Stearin S,,
38
51.5
aSource: Haraidsson (1 9871, T = thaw point.
Yield (%)
( % of initial)
Code
-
-
25 25
I\: A
37
80 31 69
125
20
-
(OF)
87 41T 81
B C H
Modification of Fats arid Oils
Solids content
503
[%I
80 60 40
-
20 -
,
0 0
10
20
30 40 50 Ternperatu re ["C]
I
60
70
Fig. 6.60. Solids content of the triglyceride classes of palm oil depending on the temperature (after Klein 1979).
To make a fat suitable as a starting material for CBE or CBS production, the fatty acid distribution on the glycerol is as important as the fatty acid composition itself. Because fractionation is a physical process that does not allow change in the triglyceride structure, only those fats that contain such structured triglycerides native in the starting material are suitable as raw materials for CBE production. These are primarily shea fat, Illi@ butter (borne0 tallow) and palm oil. All of them contain a relatively high portion of symmetrical SUS-type triglycerides. For very high quality CBE, wet fractionation is the method of choice because the products exhibit no tailing in melting
: r-----I Proportion [%]
:~u 50
20
10 0
Palm oil Fractions:
c Super olein A
Olein 0 2 0
N = Native
0 H = Stearin S20
s 3
Triglyceride class
Fig. 6.61.
Triglyceride classes of palm oil and its fractions (trials from Table 6.1 1).
Fats and Oils Handbook
504
Palm oil
Solids content
[%I
100
Fraction
-C = Super olein
80
- - A = Olein 020
- N = Native
60
- - B = Mid fraction - H = Stearin S20
40
20 0
5 10 15 20 25 30 35 40 45 50 55 Temperature
["C]
Fig. 6.62. Solids content of palm oil and its double-stage fractions (trials from Table 6.1 1).
TABLE 6.12 Characteristics of Palm Oil and Its Fractions from Double Fractionation
Fraction Palm oil s35 s33 s3 1 s29 s23
sl 9 035-s18 033-s18 031-s18
029-s18 035-s14 033-s18 031-s18 029-slS
O23 0 19
aSource: Deffense (1985).
Iodine value
53 59-6 1 59-6 1 59-6 1 59-61 60-62 60-62 60-62 60-62 58 59
Melting Point
Yield
("C)
('/OF)
(%)
(%of initial)
55 54 53 52 49 48 44 43 42 41 36 33 30 27
131 129 127 126 120 118 111 109 108 106 97 91 86 81
14 17 20 22 33 36 32 32 32 32 59 59 59 59
-
67 64
-
-
27.5 26.5 25.5 25 51 49 47 46
Modification of Fats and Oils
Liquid / solid content
505
[%I
100 80 60
Cocoa butter 40 20 0 0
5
10
15
20
25
30
35
Temperature [“C]
Fig. 6.63. Solids and liquid content of cocoa butter at different temperature (after Steiner 19551.
behavior. Today, improved processing may also makes such products available from dry fractionation (Willner et al. 1989). From the compositions shown in Table 6.13, it becomes clear that palm oil is suitable for CBE production because it has a relatively large content of SUS-type triglycerides. Its mid-fraction is used. This fraction contains mainly triglycerides with palmitic acid as the saturated component. From the table, it can also be seen that shea fat is much more suitable because its SUS triglyceride composition is much closer to that of cocoa butter. The search for fats that could be used in CBE production is commercially very interesting and has resulted in many patents. Table 6.14 lists some of these patents to make clear that a wide variety of raw materials have been studied and found useful for CBE production. The cooling curves obtained from dry fractionated palm oil mid-fractions and cocoa butter are quite similar (Fig. 6.64). If the solids curves of such fractions are compared with those of wet fractionated cocoa butter equivalents, it becomes clear that dry fractionation is scarcely able to achieve adequate products. Wet fractionation is superior. At present, the main solvents used in wet fractionation are hexane, acetone and 2-nitro-propane (Walkden 1988). The best separation sharpness can be achieved with acetone. The most unsaturated triglycerides are separated as olein at 0°C. The stearin is dissolved in acetone, recrystallized between 18 and 20°C or washed with acetone until the desired iodine value has been reached. Thus three fractions can be obtained, -60% of an oleic acid-rich fraction (IV 62-64), slightly more than 10% of a stearin (IV 12-14) and a mid-fraction. This mid-fraction (IV 32-36) contains >70% of symmetrical SUS triglycerides of which 60% is POP. A typical CBE does not exist; the composition varies depending on the desired use and the raw material chosen. Two examples are given in Table 6.13.
Fats and Oils Handbook
506
TABLE 6.1 3 Triglyceride Composition and Distribution of Cocoa Butter, Shea Butter, Palm Oil, Illipe Butter and Some Cocoa Butter Equivalents (CBE) Cocoa butter
Shea butter
Palm oil
lllip6 butter
CBE I
CBE II
32 28 36 2
39 22 35 2
Proportion (YO)
Fatty acid
C16:O palmitic C18:O stearic C18:l oleic C18:2 linoleic C2O:O arachic
25 37 34 3 1
7 39 50 4
2 81 >> Misceila >>> R=Soiiiresldue F = Filtrate
I
I
R
R
6.67 10-6 L2 3-1 and < 10 mg/Ldry matter as the minimal requirements for conducting a smooth reaction and avoiding corrosion. Energy consumption for electrolytic hydrogen is immense: 2400 Ah are needed at 0°C to produce 1 Nm3 of hydrogen; at 20°C only 2180 Ah/Nm3 are required. Plants with medium efficiency or high-pressure plants with high efficiency therefore consume, respectively, for the production of 1 Nm3 hydrogen (water saturated):
2.60 V x 2180 Ahlm = 5.68 kWh/Nm3 1.85 V x 2180 Ah/m = 4.04 kWh/Nm3 They produce roughly 0.2 Nm3 hydrogen per kilowatt-hour. Regarding cost, it must be considered that electrolytic hydrogen is very pure and requires almost no postpurification. In addition, electrolytic oxygen can be sold. The cost of electrical energy is also important. Taking Europe as an example. such costs in Norway and Sweden, for example, are 50% of the cost in Germany; in the Netherlands it is 115% of the German cost (Haraldson 1985). In other parts of the world, the costs may be only a fraction of that. 6.5.4.3 Steam Reforming Process (Hydrocarbon Refining). Hydrocarbons can be reformed with water to produce carbon dioxide and hydrogen. Although the reaction has been known since the turn of the century, sufficient progress was made to use it efficiently for industrial purposes only in the 1950s. The reaction has to be accelerated by a catalyst and is described by Equation [6.19], which represents the sum of the two successive equations.
CnHZn+,+ 2n H 2 0 + n CO,
+ (3n + 1) H,
[6.19]
It consists of two steps and includes large-scale purification. These steps are as follows: CnH2n+2+ 12 H,O.
+
n CO + (2n + 1) H2
[6.20]
nCO+nH20
+
nC02+nH2
[6.21]
In the first step, the alkane, usually natural gas, reacts with excess water at high temperature (850°C). The endothermic reaction is catalyzed by nickel. The second (exothermic) step is conducted initially at 450°C and then at 200°C; water gas is converted into the end products hydrogen and carbon dioxide (Fig. 6.114). In total, the process is much more complicated because purification has to be included. The most important catalyst poison in hardening is sulfur. Almost all hydrocarbons are suitable for reforming. They occur in mineral oil and natural gas and are always accompanied by sulfur and sulfur compounds. These have to be
Fats and Oils Handbook
570
Hydrocarbons e.g. natural gar Hydrogen >>>
3 w c , CahlystcaMO
+ 15 kcPVMol I
W ' C , FdCr catalyrt, cooling
ROfOfming
I
L
2 W C , Culzn catalyst, cooling (counter c u with~madion water)
Shining
1
I
WC, Ethanol amine solution
250'C, Nbtalyst (Hydrogenation of COand COz-rests)
I
Hydrogen
Fig. 6.1 14. Hydrogen production via steam hydrocarbon reforming.
removed before the reaction because they not only poison the catalyst but also hamper later hydrogen purification. Because hydrogen sulfide can be much more easily removed than organic sulfur compounds, the hydrocarbon used for refining is hydre genated at -350°C over catalysts (e.g., cobalt-molybdenum). The resulting hydrogen sulfide can be separated by absorption on zinc oxide or washed out. Only then does the reforming process start. Crude hydrogen from the refining process must then be purified further. It is washed with a solution of ethanol amine to remove Cq.The CO and
571
Modification of Fats and Oils
TABLE 6.32 Composition of Crude Hydrogen from the Iron Carbon and from the Hydrocarbon Reforming Process Iron carbon process Hydrogen Carbon dioxide Carbon monoxide Oxygen Hydrogen sulfide Nitrogen
98 -99% 0.5 -1.0% 0.2 -0.4% 0.0 -0.1% 0.05-0.2% 0.25-1 .O%
Hydrocarbon reforming process Hydrogen Carbon dioxide Carbon monoxide Oxygen Methane Nitrogen
99.968% 0.001 Yo 0.001%
0.005%
w0
0.01 0.007%
remaining traces of CO, are hydrogenated to methane. C02 is evaporated from the ethanol amine solution and reused. If this process is conducted carefully, hydrogen purities of almost 99.97% can be reached (Table 6.32). Other methods allow for even higher quality. Molecular sieves partly combined with active carbon treatment yield 99.99% purity. Using palladium silver membranes, which allow only the small hydrogen molecules to penetrate, makes possible purification to 99.9999%. However, such qualities do not add any advantage to the hydrogenation process, its yield or the quality of the end products (Charlesworth and Schmidt 1965, Hack and Hall 1965, Serfass and Silman 1965). The Proximo1 process (Lurgi Apparatebau) uses methanol as a buffer substance for hydrogen production. In the Proximol process, methanol is formed from water and carbon monoxide. This can be split and converted under copper catalysis at moderate temperatures (compared with steam reforming) in one step (Fig. 6.116). 6.5.4.4 Other Sources of Hydrogen. As mentioned earlier, hydrogen can also be purchased; if the producer is nearby, it can be delivered under pressure and liquefied via pipelines. Storage under pressure is at -200 bar in steel cylinders that are usually bundled. Delivery is done mainly via trailers that are left on the premises of the hydrogenation plant. Hydrogen is liquefied at a temperature of about -253°C. Hydrogen technology will quickly develop further when pressure increases to develop cars whose motors bum hydrogen. Should this succeed, technologies for hydrogen production and purchasing possibilities would further improve. 6.5.5 Influence of Processing Conditions
Hydrogenation of oils and fats is influenced mainly by six parameters. The task is to find the optimal combination for each end product and to adequately control the
572
Fats and Oils Handbook
Fig. 6.1 15. H2 production plant for the methanol cracking process (750 Nm3/h), principle and photo. Source: J.P. Daurn, 1993.
process. These six parameters are time, temperature, hydrogen pressure (and thus solubility), mass transfer, catalyst (kind, state and concentration) and substrate (i.e., oil and fat). They all flow into a seventh parameter, namely, processing, the sum of all possible influences. Time as a process parameter can easily be understood because any reaction needs a certain reaction time. However, its main influence is the different products that can be obtained by incomplete (i.e., time-dependent) hardening. Figures 6.117 and 6.1 18 show the dependence of the characteristics of end products (solids content and fatty acid composition) on the reaction time. It becomes evident from these figures that, as expected, the solid content rises with time. C,,-fatty acid composition is characterized by a decrease in linolenic acid first, followed by a quick increase in oleic acid parallel to a decrease in linoleic acid. Only after 40 min does the fraction of oleic acid, which is always fed from saturation of linoleic acid, decrease in favor of stearic acid. The course of the reaction becomes clearer in Figures 6.119 and 6.120. The influence of all other parameters on the reaction and the end products is effective over the entire reaction time (Table 6.33). These influences are described in the following sections. Influences of temperature, hydrogen pressure, amount of catalyst and concentration of catalyst poison (sulfur) are shown for the example of canola rapeseed oil hardening. The trials
Modification of Fats and Oils
5i3
Fig. 6.1 16. H, production plant for the steam reforming process (300 Nm3/h), principle and photo. Source: J.P. Daum, 1993.
have been conducted with an identical oil sample under laboratory conditions to keep the selected conditions as equal as possible (Table 6.34). 6.5.5.1 Temperature. Hydrogenation is an exothermic process with a little more than 30 kcal being set free per mole of double bond. This is equivalent to a temperature increase of 1.6-1.7 K per lowering of each IV unit. The heat of reaction on the one hand makes heating unnecessary after the activation energy has been introduced. On the other hand it makes it necessary to cool quickly and in a well-controlled manner so that the reaction does not “run away” and follows the temperature pattern foreseen. Like any other chemical reaction, hydrogenation is temperature dependent and the reaction rate increases in direct proportion to temperature. The temperature dependence of the reaction rate is true not only for the reaction itself but also for all side reactions, including such effects as an increase in reaction rate so that a considerable amount of by-products are formed, or a temperature higher than that at which they start running. Besides this direct influence on the reaction, an increase in temperature manifests itself in higher solubility of hydrogen (see Chapter 6.5.5.2) and a decrease in viscosity, which improves mass transfer (see Chapter 6.5.5.3).
5 74
Fats and Oils Handbook
Fig. 6.117. Solids content of hydrogenated soybean oil dependent on the time (after Unichema International, Emrnerich). Below an excerpt of a table is reproduced in which Patterson (1994; courtesy of AOCS) showed the temperature influence on hydrogenation. 100-1 10°C Partial hydrogenation of vegetable oil to reduce the majority of linoleic acid; minimum formation of trans-isomers. The first of two stages in a hydrogenation that seeks minimum solids 120°C content with flavor stability. (a) Not to be exceeded until a certain IV drop has been attained for oils 150°C containing substantial amounts of linolenic acid and even more unsaturated groups so as to avoid cyclization of hydrocarbon chain. (b) The popular level at which to conduct hardening of the all-hydrogenated vegetable shortening with prolonged melting range. (a) Above this temperature nickel carbonyl is completely unstable; 160°C hence, the poisoning effect of CO on Ni ceases. (b) Above this temperature, migration of double bonds and formation of trans-isomers are encouraged to reach their equilibrium level. The usual level for edible oil hydrogenation, which may follow a set 180°C amount at a lower temperature for reasons given above. If a relatively quick melting range is needed, this temperature should be used as much as possible after any other control requirements have been met. At this level, polyunsaturates diminish markedly. Should not be exceeded for edible product hardening. Above this, the 200°C risk of worsening color and increase in free fatty acids grows.
-
575
Modification of Fats and Oils
70
-{
g 60 5
EM, 8 P 2 30 LL
10
0 0
30 40 20 Resctlon tlme [mln]
10
60
Soybean oil; 3 bar hydrogen pressure; 180°C; 0.03% PRICAT 9910
Fig. 6.118. Fatty acid composition of hydrogenated soybean oil dependent on the time (after Unichema International, Emmerich). 210°C 240°C
Maximum for the most technical or nonedible hydrogenations; above this temperature, hardening rate may even be increasingly retarded. Acceptable maximum for hydrogenation of dimer and trimer fatty acids.
The various processing parameters mutually influence each other. Therefore the influence of temperature is always also dealt with in the following chapters. As a Solids content [%I
at
60 -
1ooc
60 -
15OC
40
20°C 25OC
30
30°C 20 3s0c 10
40°C
0 0
16
45 30 Hardeningtlme [mln]
60
Fig. 6.119. Solids content of hardened soybean oil depending on the hardening time.
Fats and Oils Handbook
576
Palm kernel oil Iodine value
Iodine value
160
120 PRICAT 9920,150-200"C, 3 bar hydr press.
80
40
0 0
20
40
60
80
100
120
140
160
Reaction time [min] Fig. 6.120. Decrease in iodine value during hardening, depending on the reaction time (after Unichema International).
model reaction, the hydrogenation of rapeseed oil at 140 and 200°C is given in Figures 6.121 and 6.122, respectively. The solid fat content of the end product from trial I (200°C) is lower at temperatures below 3 0 T , but higher at temperatures above 30°C as a result of the increased stearine formation at low hydrogenation temperature. Coenen (1976) gave the starting point for the temperature increase dependent on the iodine value (RID = refractive index drop): R*D(start for temperature increase)
= Oeoo2
'
(I'unhardened
triglyceride)*
[6.22]
6.5.5.2 Hydrogen Pressure and Solubility. The higher the solubility of hydrogen, the higher the probability that it is available at the point of reaction, i.e., the TABLE 6.33 Influence of Different Process Parameters on the Reaction and the Hardened Products Increase in Temperature Pressure Agitation Amount of catalyst
Reaction rate
Selectivity
lsomerization
++
t++
+++
+ ++
++ +
++ +
+t+
tt
+
+influence ++great influence +++very great influence.
Modification of fats and Oils
577
TABLE 6.34 Process Conditions of Trials to Illustrate Influences of Different Process Parameters on the Hydrogenation Reaction and on the Hardened Products Trial
Catalyst concentration ("C)
Temperature ("C)
0.1 0.2 0.1 0.1 0.1
200
Hydrogen pressure (bar1
-
I II 111 IV V
200 140 2 00 2 00
Catalyst: PRICAT 9906; Unichema International. Stirring: (R) = 750/min.
contact surface of catalyst and oil. Because hydrogen solubility is directly proportional to temperature, an increase in temperature must be positive in this sense. The solubility of hydrogen in oil can be described by the following equation (Anderson et al. 1974):
where S is the amount of dissolved hydrogen (NmVMT of fat), t is the temperature ("C), p is the pressure (bar; 1 c p < lo), afat= 40.06, bf,, = 0.334, soil = 47.04 and boil= 0.249.
Melting point ['C] 66
-
Reaction time R [min] Trans f.a. content ph]
*----* H
60 50 60 45 40 40
30 35
20 30
10 26
90
80 70 Hardening progress [I.V.]
90
80 70 Hardening progress [I.V.]
Rapeseed oil; 2 bar hydrogen pressure; 0.1% wlw PRICAT 9906
Fig. 6.121. Reaction time, melting point and trans content of hardened rapeseed oil dependent on the reaction temperature.
5 78
fats and Oils Handbook
Rapeseed oil; 2 bar hydrogen pressure; 0.1% whv PRICAT 9906
Fig. 6.122. Solids content of hardened rapeseed oil dependent on reaction temperature. The equation shows that the influence of pressure predominates over the above-mentioned influence of temperature. Pressures >10 bar, however, do not seem to make sense on the basis of current knowledge. Hydrogenation is usually carried out between 2 and 5 bar. The benefits of increased pressure were investigated by Mounts et al. (1978) and Koritala et al. (1980), who conducted the reaction at 210 bar in autoclaves resistant to pressure up to 350 bar. They showed that hydrogenation was twice as fast at 70 bar as at 35 bar; at 210 bar, it took only 10% of the time necessary at 35 bar. The extrapolated graph for hydrogen solubility in oil for three temperatures shows that doubling of the temperature results in 20% higher solubility, whereas doubling of the pressure accounts for 60% higher solubility (Fig. 6.123). The effect of the theoretical considerations from Figure 6.123 was investigated by Ray (1985). He worked out graphs that show the speed of reaction (IV unitdmin) of soybean hardening dependent on pressure and temperature. In contrast to theory, practice proved that the solubility of hydrogen is not the predominant factor for the reaction rate. At low pressure, increasing temperature by 50% results in a reaction rate that is 2.5 times higher. An increase in pressure does not remarkably accelerate the reaction at low temperature (T < 150°C); at 2OO'C, it is almost 10 times faster. Generally, there is a shortage of hydrogen compared with what would be required for a fast reaction. Assuming a vegetable oil with an iodine value of 135 (soybean oil, for instance), the concentration of double bonds is a little higher than 5000 mol/m3 of oil. Hardening of this oil at 3 bar hydrogen pressure at 180°C(see Fig. 6.1 18, for instance) allows for a hydrogen concentration only slightly less than 10 m0l/m3 of oil, which is only a small fraction of the double bond concentration..
5 79
Modification of Fats and Oils
-
-
CoUonreed 011
I.4-
Cotton seed oil
natlve
hardened
I-
0.12
1.2
I.o
0.10
9.7 bar 0.8-
1.0 bar
0.08
4.8 bar 0.6
0.4
I
loo
I
I60
l
l
’
loo
b.os I60
200
Fig. 6.123. Solubility of hydrogen in cottonseed oil (after Anderson 1974).
It should be noted that the higher availability of hydrogen caused by temperature or pressure increases is not desired in all cases. The selectivity S,,, for example, decreases with increasing hydrogen concentration in the reaction mixture because excess hydrogen is then always available to enable further reaction to stearins. Increasing the hydrogen pressure increases the melting point, the hydrogenation time and the cram content. The solids content is also higher at most temperatures (Figs. 6.124 and 6.125).
6.5.5.3 Mass Transfer. Hardening takes place in the three-phase system gas/liquidsolid. The gadliquid reaction partners hydrogen and oil have to be brought together on the solid catalyst surface for reaction. The reaction products have to be removed as quickly as possible to enable new reactions, and consumed hydrogen in the reaction has to be replaced. This is particularly important because the concentration of double bonds in the oil is always very much higher than the concentration of hydre gen (see Chapter 6.5.5.2).Mass transfer therefore plays an important role: 1. H, is brought into the oil. 2. H, is dissolved in the oil. 3. H, travels to the catalyst surface surrounded by an oil film. 4. Oil molecules and H2 travel into the catalyst’s pores toward the nickel surface. 5 . H, is absorbed onto the active portion of the Ni catalyst. 6. H, dissociates to hydrogen atoms capable of reaction. 7. Oil molecules are absorbed at their double bonds by the catalyst. 8. H, is added to the double bond of the oil molecule.
580
Fats and Oils Handbook
Hardening progress [I.V.]
Hardening progress [I.V.]
Rapeseed oil; 200°C reaction temperature; 0.1% whv PRiCAT 9908
Fig. 6.124. Reaction time, melting point and trans content of hardened rapeseed oil dependent on the hydrogen pressure.
9. Desorption of the (partially) hardened oil molecule. 10. Diffusion from the catalyst’s pores. 11. Return to the bulk oil through the oil film surrounding the catalyst particle. In the immediate vicinity of the catalyst, diffusion plays the most important role. Similar to the equation for hydrogen solubility, Anderson (1974) has developed a representation for this process as follows: N = A DH.c/L
[6.24]
where N is the rate of transported hydrogen, A is the surface of pores, L is the length of pores, DH is the diffusion coefficient for hydrogen in oil (cmVs) and c is the concentration of hydrogen in the oil (L H2/kg oil). The diffusion rate is only very weakly correlated to the iodine value; therefore it changes during reaction, but only slightly and without any practical consequence (Fig. 6.126). Contrary to mass transfer on a small scale by diffusion at or close to the catalyst surface, mass transfer on a large scale has to be supported by stirring and permanent circulation. In addition to increasing mass transfer, stirring also helps to keep the temperature in the reaction mixture as homogeneous as possible to avoid local overheating. It also helps thermoregulation by transfemng the heatingkooling effect quickly through the mixture. Usually stimng helps for a good intermixing. New processes also use pump circulation with permanent injection of the reaction partners into the reaction chamber (see Chapter 5.56). Design and arrangement of the stirrer are of great
Modification of Fats and Oils ""
581
I
10%
so
2O0C
3OoC
4OoC
90
ao
70
Hardening progress [I.V.]
1 Fig. 6.125. pressure.
Rapeseed oil; 200°C reaction temperature; 0.1% w/w PRICAT 9906
I
Solids content of hardened rapeseed oil dependent on the hydrogen
importance for an effective distribution of hydrogen and an even temperature in the reaction chamber. Care has to be taken to avoid damage to the catalyst by mechanical stress as a result of poor design of the stirrer or excessive stirrer speed, which can drastically lower the efficiency.
Fig. 6.126. Diffusion of hydrogen in cottonseed oil (after Anderson e t a / . 1974).
Fats and Oils Handbook
5a2
Higher temperatures as well as increased stirrer speed thus improve diffusion and speed up the reaction. Both effects add up but not in a linear fashion; increasing the stirrer speed at high temperatures accelerates the reaction more than at low temperatures. The same is true in reverse; at a high stirrer speed, a temperature increase is more effective than at a low stirrer speed. The choice of these two parameters depends on the end products wanted because those are influenced by both (Fig. 6.127 and 6.128). By keeping the catalyst concentration constant, the selectivity c& be increased by lowering the stirrer speed. The less hydrogen that is available at the point of reaction, the lower the formation of stearic acid. Increasing the amount of catalyst logically works the other way around. This influence is described in Chapter 6.5.4. 6.5.5.4 Kind, Condition, and Concentration of Catalyst. The kind of catalyst has already been discussed in Chapter 6.5.3. The previous figure showed that there is an optimum catalyst concentration that is influenced by the other reaction processing parameters (here stirrer speed); this is also true for selectivity. Similar considerations can be made for the reaction rate (Fig. 6.129). As could be expected, increased stirrer speed improved mass transfer with the effect of a higher reduction rate of IV (decrease of IV units per time). For a given catalyst concentration, however, an optimum that is dependent on stirrer speed can be observed.
Stirrer speed [min
-’ 10001 3
1950
1450
450
Decreas of I.V. units Der minute
160
180
200
220
240
Temperature [“C] Soybean oil; 3 bar hydrogen pressure; 0.08% Ni-catalyst
Fig. 6.127. The influence of stirrer speed and temperature on the hardening of soybean oil (after Allen 1982).
Modification of Fats and Oils
Stirrer speed [min
583
-' 10001 *
2000
1500
1000
500
0 0.01
0.05
0.13
0.09
Catalyst concentration [% Nil Soybean oil; 185°C; 2.4 bar hydrogen pressure Fig. 6.128. The influence of stirrer speed and catalyst concentration on the selectivity in hardening of soybean oil (after Allen 1982). Reaction rate [Decrease in I.V. units/min]
5 4-
321-
0.02
R = Stirrer speed (min - l )
0.04
0.06
0.08
0.10
0.12
0.14
Catalyst concentration [% Nil Soybean oil; 192°C; 3 bar hydrogen pressure
Fig. 6.129. Reaction rate of soybean oil hardening dependent on stirrer speed and catalyst concentration (after Allen 1982).
Fats and Oils Handbook
584
The influence of different catalyst concentrations with all other process parameters kept constant is shown in Figures 6.130 and 6.13 1 . As noted earlier (see Chapter 6.5.3), the structure of the catalyst is of great importance. Too narrow pores can be blocked by oil molecules and render adsorption and desorption more difficult. Too large pores reduce the total catalyst surface. The same can be said for the particle size of the catalyst as such; too large particles reduce the surface area, too small particles make filtration more difficult (Table 6.35). The structure of the catalyst also influences the trans fatty acid formation as shown for the reaction of a nontriglyceride model substance (Fig. 6.132). Apart from the physical properties that the catalyst carries by nature of its production, there is also an influence from the number of reuses. In addition to the cost-saving effect of multiple use, reuse of catalyst usually leads to more stable (uniform) quality of the end products. The catalyst condition is stabilized at a constant performance level after some reaction cycles, a level that can deviate substantially from the initial one. With a catalyst that has already been used two or three times before, products with constant quality are obtained, i.e., equal level of trans fatty acids, iodine value and solid fat content at different temperatures. In addition to the amount applied, the kind of catalyst plays an important role (as can be expected). Different catalysts for different applications are being produced, three of which are compared below in their influence on the reaction products (Table 6.36). All three are nickel catalysts and have been used under laboratory conditions to ensure constant processing conditions. Table 6.36 gives these conditions, and the results of the trials are illustrated in Figures 6.133 and 6.134.
I
Rapeseed oil; 200°C reaction temperature; 2 bar hydrogen pressure
Fig. 6.130. Reaction time, melting point, and trans content dependent on the catalyst concentration.
I
of hardened rapeseed is
583
Modification of Fats and Oils
Hardening progress [I.V.] Rapeseed oil; 200°C reaction temperature; 2 bar hydrogen pressure
Fig. 6.131. Solids content of hardened rapeseed o i l dependent on the catalyst concentration.
Catalyst XI (Pricat 9900) is a very active catalyst for universal use; catalyst XII (Pricat 9906) is highly selective and is applied when the amount of trienes and saturated triglycerides in the end product must be kept low; catalyst Xm is used when a steep dilatation curve is desired. More examples for the characteristics of products hardened with different catalysts are given in Chapter 6.5.7. At present, these catalysts have been replaced by new types exhibiting better activity and sensitivity, lower sensitivity toward catalyst poison and an improved filtration rate. In the table, results of trials with Pricat 9910 and Pricat 9920 are shown. The former is a multipurpose, sulfur-resistant, medium-pore catalyst with a high nickel surface area; the latter is a very selective, wide-pored catalyst, especially developed for vegetable oil hardening. TABLE 6.35 Influence of Structural Parameters of the Catalyst on the Hardening Processa ~~
Structural Parameter Total surface area Nickel surface area Pore structure Particle size Poisoning
Activity
Influence on Selectivity
** *
* **
* **
* **
*,**influence proven; (*) influence assumed aSource: Klauenberg( 1 986).
Consumption
(*I
** * *
Fats and Oils Handbook
586
Fig. 6.132. Amount of trans isomers in methyl oleate hydrogenation depending on the pore size of the catalyst (after Linsen 19711.
Finally, the effect of catalyst poisoning is shown in Figures 6.135 and 6.136. They represent the effect on soybean oil hardening with catalysts that were poisoned intentionally with 2 and 4 ppm sulfur, respectively.
6.5.5.5 The Substrate. The influence of by-products or impurities will not be discussed in this section but rather the influence of the oiYfat itself. The most prominent characteristic for an oiYfat to be hardened is its degree of unsaturation, i.e., the number of double bonds represented by its iodine value. The IV is a measure for the amount of hydrogen that is consumed during reaction (Fig. 6.137). Fats and oils are natural products that undergo certain fluctuations in their composition. These fluctuations are caused either by their genetic blueprint (different cultivation) or by environmental influences such as climate, weather or location TABLE 6.36 Process Conditions for Soybean Oil Hardening, Comparing Three Catalyst9
Temperature Hydrogen pressure Catalyst concentration
Pricat trial
9900 A
9906 B
9908 C
(“C) (bar)
180 3.5 0.09
140 3.0 0.09
180 2.0 0.45
(Old
aCourtesy of Unichema International, Emmerich.
587
Modification of Fats and Oils
Share C18:O [%I 80
Share C18:l [%]
I
I
I
120
80
100
25
0 60
Iodine Value
Fig. 6.133. Influence of catalyst and reaction conditions on the c18:o and acids (catalysts from Table 6.36).
c1&1fatty
of cultivation, and harvest and postharvest conditions. Therefore, to obtain optimal results, hardening conditions have to be adjusted to the raw materials. The composition of poultry fat is very much dependent on the feed; the more grains in the diet, the more unsaturation. Milk fat differs in composition when
50
-b
Soybean oil
\,.\18:2
120
100
80
60
Iodine Value
Fig. 6.134. Influence of catalyst and reaction conditions on the acids (catalysts from Table 6.36).
c1&2 and
c,&3 fatty
Fats and Oils Handbook
5aa
1
Rapeseed oil; 200°C react. temp.: 2 bar hydr. press.; 0.1% w/w PRICAT 9906
1
Fig. 6.135. Reaction time, melting point, and trans content of hardened rapeseed oil dependent on the sulfur content of the oil.
cows are grazed on pastures compared with winter feed inside. However, these two fats are rarely hardened. Fish oil is more highly unsaturated when produced from fish that are about to spawn or starting on their way to the spawning grounds. The IV value of fish oil increases with the coldness of the water in which the fish are caught.
1
Rapeseed oil, 200°C react. temp.; 2 bar hydr. press.; 0.1% wiw PRICAT 9906
1
Fig. 6.136. Solids content of hardened rapeseed oil dependent on the sulfur content of the oil.
Modification o f Fats and Oils
5 8'1
Soybean oil Sunflower seed oil Cottonseed oil Rapeseed oil Groundnut oil
Fig. 6.137. Melting point of hardened oils and hydrogen consumption (after Rudischer 1959).
Vegetable oils and fats also exhibit such influences. The linoleic acid content of sunflower oil depends on the difference between day and night temperatures in the area of cultivation; it can range from 30 to 70% (Morrison et ul. 1978). Similar differences are found for Argentinean (IV 103-105) and Nigerian (IV87-95) peanut oil. The many areas of origin, the many different kinds of stress during harvesting, crushing, storage and transport and the variety of end products desired lead to a variety of processing conditions adapted to the specific case; many of these are described in detail by Patterson (1994). 6.5.5.6 Influence on Isomerization. Isomerization reactions that occur during hardening (see Chapter 6.5.2.1) have been explicitly researched. Although all isomerization products from hardening of vegetable and mammal fats also occur naturally in fats and oils and are harmless according to today's knowledge, much attention and publicity have been paid to them in some parts of the world (see Chapter 6.5.7.1). It must be said that hardened fats, as such, cannot be equated with rruns fatty acids. The amount of TFA can be kept low by special hardening techniques. Generally the TFA content increases with increasing S, selectivity. Ray (1985) researched the influence of some processing parameters on trans fatty acid content on a laboratory scale (Table 6.37). 6.5.6 Hardening Techniques
Hardening is influenced mainly by the parameters described in Chapter 6.5.5. A great variety of end products can be obtained from the same starting material by varying the reaction parameters. Besides these variations, the quality of the oiVfat plays an important role.
Fats and Oils Handbook
590
TABLE 6.37 Influence of Process Conditions on the Trans Fatty Acids Content of Hardened Fats and Oils Trans content increasing with
Keeping constant ~~~
~
~
Lowering stirrer speed, increasing temperature Falling pressure, increasing catalyst concentration Fa1Iing pressure
Pressure Stirrer speed Temperature
6.5.6.1 Preparation of Fats and Oils (Purification). To ensure efficient hydrogenation and good quality fats, the oildfats to be hardened have to be prepared carefully. Usually they are prerefined (see Chapter 7, refining). Recently, however, processes have been developed for direct hardening, i.e., without prerefining. Different authors articulate different opinions on what the quality of a well-pretreated fat/oil for hardening (Table 6.38) must be. An excessively high water content contributes to deactivation of nickel and to fat splitting (hydrolysis). Too high FFA content favors the formation of nickel soaps and negatively influences filtration. Soap, as a surface active substance, covers the catalyst’s surface and disables it. All other impurities are usually removed during bleaching (see Chapter 7.4 bleaching). These include seed (chlorophyll), blood pigments from the crude oils and fats that can block the catalyst’s surface. The really important catalyst poisons, however, are nitrogen, phosphorous, sulfur and chlorine compounds. Table 6.39 gives typical ranges for such compounds in crude oils as well as in oils pretreated for hardening. Obviously, the complete removal of such catalyst poisons is not possible for cost reasons. However, they have to be brought to levels that represent the commercial compromise between cost for removal and benefit for hydrogenation. Different catalysts (catalyst metals), of course, differ in their sensitivity toward catalyst poisons. The most damaging of these poisons, sulfur, which is present in some form in all oils and fats, heavily deactivates almost all catalysts. A possible explanation for the negative influence of sulfur on different catalysts is given by Coenen (1976). Sulfur poisoning also indirectly shows the advantage of support catalysts over nonsupport catalysts. Assuming a surface of 0.15 m2/g of catalyst, 5 ppm of sulfur block about 13 m2 of catalyst surface area in the oil. This means that the equivalent of TABLE 6.38 M i n i m u m Quality Requirements for Oils/Fats to Be Hardened Free fatty acids Soap Water Sulfur Source
I
Soap decantation or centrifugation
Washing Soap free
1
+
>>>Aqueous soap solution >>w
+ I
I
' i Dlyw
Neutralized oil
2040 hPa
Soap splming
1:
Separation
1
-I
t
heavihl diluted Fatty acids
aqueous soiution of NaJSO,
Fig. 7.5. Processing flow chart of neutralization with alkali lye.
Oil Purification
61 9
TABLE 7.1 Factors Influencing the A m o u n t of Lye Needed for Neutralization Influencing factor Proportion of free fatty acids before neutralization Specific density of the oil Average molecular weight of the oil's fatty acids Lye strength
influence
+ + -
+ = proportional; - = inversely proportional
In this formula, an average molecular weight is used for practical purposes. Because of the different average molecular weights of different oils and fats and their different specific densities (see Chapter 2.3.2.3), which enter the formula via the percentage of FFA, different amounts of caustic soda are calculated for the neutralization of different fats (Table 7.2). Baumt is an old measure for density (0' BaumC is equivalent to the density of a 10% solution of sodium chloride and 60" B t corresponds to a density of 0.745; see also Fig. 7.6). It can be seen that the amounts of lye needed to neutralize the common vegetable oils are all of the same order of magnitude and that only palm fats (laurics) and rape oil with high erucic acid content (not common any more) differ substantially. Thus it becomes possible to cover the whole spectrum with two lump numbers. To ensure complete neutralization, an excess of lye has to be used. Table 7.3 shows the factors influencing this excess. Application of excess lye is limited because some effects are positive but can easily become negative. For example, intense contact of the lye with the oil is positive on the one hand to keep the excess of lye small; on the other hand, the danger of emulsion formation rises with increased stirring speed. Because emulsions are difficult to break and cause trouble TABLE 7.2 A m o u n t of Lye Necessary for Neutralization of Different O i l s a n d Fats
Oilfat Lard Tallow Palm oil Soybean oil Cottonseed oil Sunflower seed oil Peanut oil Rapeseed oil (LEAR) (HEAR) Coconut oil Palm kernel oil
Amount of lye needed per % of FFA to neutralize one MT of oil with caustic soda
Density at 15%C (g/cm3)
Average molecular weight
(kg NaOH)
(L 4n-NaOH)
(L NaOH, 14 Be)
0.920 0.945 0.935 0.928 0.925 0.923 0.91 8 0.915 0.913 0.923 0.930
2 73 2 70 2 70 280 2 75 280 2 78 2 85 308 215 225
1.35 1.40 1.39 1.33 1.35 1.32 1.32 1.28 1.19 1.72 1.65
8.4 8.8 8.7 8.3 8.4 8.2 8.3 8.0 7.4 10.7 10.3
12.3 12.7 12.6 12.1 12.3 12.0 12.0 11.6 10.8 15.6 15.0
Fats and Oils Handbook
620
Amount of lye needed for neutralization [%I
Lye strength [W
I
lo
30 27 24 21 I8
16 12
9 8 3
0
1 1
2
3
4
5
Oil’s FFA [%I
0 1
1
2
3
4
5
8
Lye normality
Fig. 7.6. Amount of caustic soda needed for neutralization (after Brimberg 1981).
in production, it is more efficient in the end to increase the amount of lye. Less excess lye is needed for unsaturated fatty acids because the soaps from unsaturated fatty acids are liquid and do not entrain lye (Rudischer 1959). After soap separation, the oil is washed with water until soap free and dried under vacuum. Neutralization is then followed by (interesterification, if needed) bleaching and deodorization. In the old days, it was also quite common to neutralize oil using unslaked lime instead of caustic soda. It is considered fraudulent to add unslaked lime to crude oils with an FFA above the specification in order to reduce the FFA content, thus bringing it back within the limits of the specification. In addition, this process harms the client who receives the oil because sludge sediments, which TABLE 7.3 Several Factors Influencing the Excess of Lye Needed in Alkali Neutralization ~
Factor Free fatty acids (FFA) after neutralization Neutralization temperature Share of short-chain FFA Share of unsaturated FA Proportion of gums Portion of FFA (i.e., FFA < 1%) Intensity of contact between oil and lye (distribution) Color (brightness as an indication for impurities) Lye strength
+ proportional; - inversely proportional.
Influence
Oil Purification
62 1
are difficult to remove, are deposited in oil storage tanks. If oils are difficult to refine with common neutralization techniques, a soda water glass boiling step can be included. The silicates that precipitate bind undesired particles. S o d a water glass boiling is conducted at temperatures of 1W105"C and takes -15-30 min. The approximate amounts of energy required for common neutralization processes per ton of oil are as follows: discontinuous process, 150 kg steam and 4 kwh electrical energy; centrifugal process, 85 kg steam and 13 kwh electrical energy.
7.2.2.2.7 Discontinuous neutralization. A neutralizer for discontinuous neutralization is a vertical cylindrical vessel of up to 75-ton content; it is conical in its lower part. It is equipped with an outer heating mantle and an inner heating coil to ensure proper heating. Also, there is a stirrer for good agitation and some shower heads to dose finely dispersed caustic soda solution onto the oil. In the lower part, there is a view glass. Figure 7.7 shows a neutralizer with heating mantle, stirrer, inlet for steam and outlet for the condensate. After being heated to -6o"C, the lye is finely sprayed onto the oil. It is then heated to 70430°C. Because of its higher specific weight, the caustic soda solution percolates through the oil, thereby neutralizing the FFA on its way. Stirring supports this process. The aqueous soap solution collects in the lower cone and is decanted. The completeness of soap separa-
Fats and Oils Handbook
622
tion is optically controlled using the view glass (manually) or by ultrasonic or conductivity measurements (automatically). The soap is collected and the oil is washed soap free. Then the oil is dried in the neutralization vessel. Discontinuous neutralization has advantages if small batches of different oils have to be neutralized or if the daily throughput does not exceed 10 MT. The investment for such a plant is low. On the other hand, it cannot be automated easily, thus incurring higher labor costs. This makes discontinuous neutralization of particular interest for countries with low wages even for higher throughput. In addition, such plants can be manufactured and maintained locally, which is very important in countries with high import duties or with no reserves of foreign currency. Compared with these advantages, in special situations, the high energy costs usually do not matter. Cycle times of a batch neutralizer are given in Figure 7.8.
7.2.2.2.2 Semicontinuous neutralization. If neutralization in batch vessels (discontinuous) is combined with soap separation via centrifuges (see Chapter 5.2.5.3), a semicontinuous process is born. If the centrifuge is sufficiently large, several neutralizers can feed one machine. Because no settling time is needed for the soap cycle, times can be reduced. To break the emulsion, an electrolyte must sometimes be added. The Zenith process is another semicontinuous process that is not common in Europe (Bergmann and Johnsson 1964, Cavanagh 1990, Hoffmann 1973 and 1974). In this process, the degummed oil is fed dropwise from the bottom to a vertical cylindrical vessel that is filled with weak lye. The oil percolates through the lye (from the bottom to the top) and is neutralized along its way. It then drains off on top. The entire Zenith process includes semicontinuous degumming and bleaching with the described neutralization step in between the two ( B r h 1976). 7.2.2.2.3 Continuous neutralization. Using centrifuges for separation, a fully continuous neutralization plant can be designed. These plants require a high Crude oil Batch filling
30 rnin
I
I
t
Cooking @
20min
I
10 min
60 rnin
30 rnin
Soap separation I
I
I Neutralized oil
10 rnin
225 rnin (e255 rnin)
@ for soda waterglau cooking
Fig. 7.8. Cycle times for batch neutralization
Oil Purification
62 3
investment but can be run at very low cost. Plants of this type were fxst erected in 1948 by Alfa Laval, although the idea as such was already quite old (Hefter 1906). The processing steps given in Figure 7.5 are not actually performed in the same vessel one step after the other as described in Chapter 7.2.2.2.1, but in specialized equipment that is subsequently passed through. A processing flow chart is given in Figure 7.9, w h l e Figure 7.10 shows continuous neutralization plants for both the short and the long mix process. The plant is fed from a storage tank.The oil passes through an oil meter and a plate heat exchanger where it is indirectly heated to the reaction temperature of 80-9O0C. From there, it is pumped into a mixer where it is dosed with an amount of caustic soda equivalent to the acid value of the oil. This dosing step is coupled with the oil meter to ensure the addition of the correct amount of lye. Then the oillsoap mixture is pumped into a reaction vessel and a second mixer from which it passes through a self-discharging separator. Less than 15 s after addition of the caustic soda the soap is already separated by centrifugation. The bowl rotates at 400-5000 rotations/min and the soap solution is continuously discharged. If needed, water can be added to the bowl to improve soap separation.
Oleometer
Metenng oil
1
optional
>>Aqueous soao solution
-9
=.
Soap splitting
1
1
7 c t
e i Wing
>>> Aaueous saao solution
Soap splitting
Vacuum dryer (-20 hPs), thin oil film
Neutralized oil
Fig. 7.9. Flow chart of continuous neutralization with separation centrifuges.
624
fats and Oils Handbook
Fig. 7.10. Continuous neutralization plants for the long-mix (above) and short-mix (below) processes (courtesy of Alfa Lava1 AB, Tumba).
Bad quality oils may require a second treatment (rerefining), following the same principle as the first process. The plate heat exchanger for the second stage is much smaller because the oil leaving the first separator is at a sufficiently high temperature so as to require only postheating. The oil is then heated to washing temperature in a plate heat exchanger and washed with hot water (-10% of the amount of oil). The washing water is also separated by a centrifuge. After separation of the washing water, the oil is dried. It is subsequently fed into a vacuum dryer where it flows in a thin film under reduced pressure (-25 hPa) over a series of cascades. The water easily evaporates from that thin film, and the oil is collected at the bottom of the vessel to be pumped into the neutral oil storage. The soaps are also collected and are worked up or sold as such.
Oil Purification
62 5
Such plants are designed for a capacity of up to 300 todd of crude oil. In modem installations, the plant can be cleaned in place (CIP), which allows for continuous operation for several months without external interference. In addition to the above-mentioned short mix process, which is applied mainly in Europe, there is a long mix process that is preferred in the U.S. This process differs in that the caustic soda is added to the oil at ambient temperature and the mixture is heated to 70-80°C after a long reaction time of 5-15 min. The separation temperatures in the long mix process are between 55 and 90°C (Hendrix 1990). Figure 7.11 shows a model of the neutralization plant with separation centrifuges. Haraldsson (1983) compared the results of refining soybean oil (0.4% FFA, 0.5% nonhydrated phosphatides) for both processes (Table 7.4). 7.2.2.4 The lntegration of Degurnrning. Before the continuous neutralization described above, a degumming step can be carried out (see also Chapter 7.1). To do so, the amount of acid required (typically 75% phosphoric acid) is added to the oil. The acid supports the hydration of phospholipids followed by their precipita-
Fig. 7.11. Model of a plant for continuous neutralization (courtesy of Westfalia Separator AG, Oelde).
Fats and Oils Handbook
626
TABLE 7.4 Results of Soybean O i l Refining Using the Short (s) and the Long (I) M i x Process, W i t h (+) or Without (-) Prior Phosphoric Acid TreatrnenWb ~~
~
Process
Mixing time (5)
Mixing temperature ("C)
Washing water (L)
P (pprn)
90 1 x10 72 -20 1 x10 4 S+ 5 90 1 x10 6 s+r 5 90 2x4 1 Soybean oil: 0.4% FFA, 0Soh nonhydrated phosphatides. Caustic soda 4n, 25% excess; soap analysis according to AOCS method.
S-
I-
5 240
Resulting in Ca (ppm)
Soap (ppm)
a4 2 6 >z Washlna water (very low in SOED~
_-+_
Oil, degummed. neutralized
Fig. 7.12. Processing flow chart of degumming, refining, dewaxing and rerefining vla centrifuges.
7.2.4 Other Processes
Many other processes have been tested but failed to succeed in the end. For example, Bengen and Schlenk (1940) patented a process using urea, which forms adducts with the fatty acids. Rigamonti and Riccio (1952 and 1953) developed the process further, but the results achieved were not good enough for upscaling. Bhattacharya et al. (1989) proposed a technique for enzymatic neutralization. They demonstrated in detail that the stoichiometric addition of glycerol yielded the best results. With the enzyme, the use of a temperature of 70°C and 10% enzyme addi-
Fats and Oils Handbook
62 8
FUtnUon
Refined rnimtk
F l b r eid (0.g. b*.cMng wfthr) ( F u m pmcedng ma 5.2.4.1)
Fig. 7.1 3. Processing flow chart of miscelI a refin ing .
tion were optimal. There is still a long way to go for such a process to be upscaled because usually oils are also hydrolyzed by the enzyme.
7.2.5 Soap Splitting Soap is usually split with sulfuric acid as follows:
2 R-COONa + H2S04+ 2 R-COOH + Na2S04
V.51
In conventional plants, soap splitting is done batchwise. The soap is then separated by decantation. The separation line is detected visually with a view glass installed in the conical lower part of the vessel. The refining fatty acids or soap stock fatty acids also contain the neutral oil that was entrained during soap formation as well as some impurities. The soap stock fatty acid has to be refined. The losses, which consist of entrained non-fatty acid substances, usually oils and fats, form the refining loss and are calculated in the refining factor. This factor determines the ratio (wt/wt) of the refining fatty acids compared with fatty acids present in the crude oils before neutralization. An ideal neutralization process, which removes only the fatty acids, has a factor of 1.0. A factor of 1.7, for example, means that components that were unintentionally removed from the oil together with the fatty acids account for 70% of the fatty acid weight in the nonneutralized oil. At present, continuous soap splitting is applied. In such plants, sulfuric acid is continuously dosed into the soap stream. The dependency on the pH-value follows the usual rules of neutralization with acids. That means that low amounts of acids or lye close to the equivalence point cause a large pH shift, whereas in the regions of high or low pH higher amounts of acid or lye are necessary. Figure 7.14 shows that the titration curves for soap splitting are very steep between pH 3 and pH 5. Small amounts of acid result in a large pH change. For good results, a pH of 3 is
Oil Purification
62 9
required in soap splitting. After the mixture of soap and sulfuric acid has passed through the reaction chamber, conductivity is measured to decide whether free acid is present. The dosage of sulfuric acid is controlled such that no free acid occurs. Currently, FFA can also be separated from the water via centrifugation mainly because materials have been developed that can stand the aggressive fatty acids and the sulfuric acid (pH 2 to 3) at temperatures of 80- 90°C. 7.2.6 The Principle of Centrifugal Separation
Separation via centrifuges occurs not only in refining but also in oil extraction (see Chapter 3). It enables high throughputs in continuous operation. Separators are also called centrifuges; if they also allow the separation of large amounts of solids, they are called decanters. Diipjohann and Hemfort (1975) have compiled the principles in an article. Brunner (1984) reported new developments. An excellent paper on centrifugation was published by Hemfort as a brochure of Westfalia Separator AG (1984). 7.2.6.7 The Theory of Centrifugal Separation. A separator consists of a set of disks that are stacked and adjusted at an angle cp toward the horizontal axis. The set of disks rotates in a bowl with an angular velocity o around the separator axis. The liquid fed to the separator either contains solids or it consists of two liquids that are insoluble in each other. The liquid is fed via (3) (Fig. 7.15). As a result of their higher density, the solids are driven toward the inner wall of the bowl shell where they collect in the sediment holding space (11). From there, the solids are ejected
'1
-
--- ---_
-,
7
............
11.9%
TFM
Fats and Oils Handbook
630
17
15
" I
14
'I
1
\
\
l6
l3
Fig. 7.1 5 . Drawing of a centrifuge (courtesy of Westfalia Separator AG, Oelde).
(12) via the sediment ejection ports (13), once the sliding piston (12) opens. The heavy liquid phase leaves the separator via (2), the lighter, via (4). They are separated by the centripetal pump (7). Water flushing is possible via (5). Other details include disks (2), rotometer (6) and annular piston (17). The liquid feed is evenly distributed over all separating spaces between the disks. It is thus split into many thin layers that pass through these spaces. To enable this separation, the disks have rising channels (Fig. 7.17), which are vertical passages through the whole set of disks, formed by holes in the disks that are positioned one above the other. In a separator with, for example, z disks, the quantity passing one separating space is Q/z (where Q is the hourly flow of feed stock; Fig. 7.16). The solid particles being introduced with the liquid at E are carried with the liquid phase at a separation velocity v, in the direction of the arrow. They are forced to the outside, in this case the upper disc U with a radial velocity vr, which is also the sedimentation velocity. The solid particles then have the vector velocity vv, resulting from the sum of v, and v,. They are considered to have been removed from the liquid once they have reached the upper conical surface of the individual separating space. They then slide down in a cohesive layer to the bowl's sludge space, because, having reached the wall, the force 1 becomes negligible so that they follow vr The liquid flows with a separation velocity vs in the direction of the arrow, exiting through the center of the bowl. To describe the separation mechanism in a centrifuge, many authors have developed equations (see, e.g., Cowan 1976 and Hemfort 1960 and 1984). Particles
63 1
Oil Purification
F,=rn.b
[7.6a]
F c = r n . w2 . r
[7.6b]
where rn is the mass of the particle, b is the acceleration (in the case of gravity, b = g = 9.8 m/s*), o is the angular velocity, and r is the radial distance (radius of the Disk set with rising channels near the periphery
Disk set with rising channels near the center
Disk set with central rising channels
Fig. 7.17. Position of rising channels in a set of separator disks (redrawn; courtesy of Westfalia Separator AC, Oelde)
632
Fats and Oils Handbook
centrifuge). For immiscible liquids or solid particles in a liquid, the driving force for separation is the difference of the forces applied to either the two liquids or the solid particles and the liquid. This force differs depending on the mass of the liquids and/or the solids. The mass m from Equation [7.6b] then becomes the mass difference.
F , = (mliquid - msolid) . 02 . r or F, = (Am) . 02 . r
P.71
Assuming spherical particles, their mass may be replaced by the product of volume times density and Am may become the difference in density Ap. The volume of a spherical particle is characterized by its diameter (D). Am=--- A P
6 nD3
~7.81
By substituting Equation [7.8] in [7.7],
According to Stoke’s law, exposed to the force Fs, a particle is sedimenting with the sedimentation speed V, (Equation L7.101). This force works opposite to F,:
If Fs and F, are equal, the following result is obtained: [7.11] The sedimentation speed thus becomes:
[7.11a] If the centrifugal acceleration is replaced by gravity, the speed of sedimentation without additional forces is obtained as:
v, = D2 .p2. g
[7.1Ib]
18.77
The sedimentation distance n of a particle (the product of the speed V, and residence time in the bowl) can be calculated when time is expressed as the quotient of the volume V of the liquid and its flow rate Q.
[7.12]
Oil Purification
633
If n is set to s/2, thus assuming that half of the particles with the diameter D are sedimenting during the residence time, Equation [7.12] becomes Q=- D
~ . A PV . 0 2 . r
9.v
[7.13]
S
If the volume V is substituted by a function of the radius and machine specific parameters are brought in, Equation [7.14] is obtained after allowing for some simplifications: [7.14] In this equation, the first part means Stoke's sedimentation speed, the second the equivalent clarification area of the bowl, where q is the dynamic viscosity, D is the particle diameter, g is the acceleration due to gravity, Ap is the density difference and z is the number of disks; see Figure 7.16 for x,rl and r2. Hemfort further developed Equation [7.14] by reintroducing the acceleration due to gravity and the clarification area A (effective rotor area):
[7.14a] Defining 6 = r . w2/g as the centrifugation or acceleration factor, and substituting the first part of the equation by VStokes, Equation [7.14a] becomes
Q = VStokes 6 ' A '
[7.14b]
where VStokesis Stoke's sedimentation speed, 4 is the centrifugation factor, A is the clarification area, and 6 . A is the equivalent clarification factor. The number of disks (z), the angular velocity 0,and the angle x as well as rl, r2 can be influenced by the construction of the centrifuge, whereas q, A, and Ap are dependent on the products to be separated. Within the limits set by the substances to be separated and the technical restrictions for construction, these parameters influence the throughput as shown in Table 7.5. Factors relating to the product may sometimes also be changed, thereby improving centrifugal separation. If, for example, solids have to be removed from oil, the density difference may be increased by a water treatment. If the solids are able to soak, the density difference is usually increased. The viscosity and density of the oil can be reduced by heating it up; because lowering the viscosity of the carrier liquid increases efficiency, this will help. Dilution with a solvent also helps, but is not done because the solvent has to be removed later. Increasing the particle diameter also has
Fats and Oils Handbook
634
TABLE 7.5 Influence of Different Parameters on the Throughput of Centrifuges ~~
~~
Dependency on influencing factor Influencing factor
Single
Square
Cubic
+
Particle diameter D Density difference Ap Viscosity q Angular velocity w Tilt angle of disk (tan cp) Number of disks z Effective radii r, and r,
+ t
+ + + ~~
~
+ proportional; - inversely proportional
positive effects. Addition of flocculates or agglomeration of the particles improves the performance. Adjustment of the pH helps with proteins to reach their isoelectric point. In centrifugation of products of Lanza fractionation (see Chapter 6.2.2.2),the addition of surface active detergents and electrolytes helps to reduce surface tension and to avoid electrical charging of fine floating substances. Both measures, initially done for other reasons, help to improve separation. Enlarging design factors can also improve centrifugal separation. Increasing the bowl speed would help. However, there is a limit dictated by the permissible stress. Hemforth (1960) mentioned a 00,2 limit 5 500 N/mm2. Only very few materials fulfill this criterion. The bowl diameter and with it (r13- r23) are determined by the required sediment-holding capacity of the bowl. The tilt angle of the disks, cp, is limited by the slope of the separated solids. It must be smaller than the slope angle. The number of disks in the stack is limited by the bowl as a housing. All factors together must fulfill the equilibrium or better the optimal compromise between the best technical solution and commercial viability. After having separated a solid phase, the remaining two liquids (in our case oil and water) may subsequently be separated. The position of the rising channels is determined by the density difference of the two liquids. The rising channels must always lie within the separating zone. The clarification area determines the separation sharpness. For the system watedoil, this means that in case A (rising channels near the periphery), the larger separation area is on the side of the lighter component, i.e., on the oil side. Consequently, almost complete water separation can be ,achieved. If the larger area is on the side of the heavier (denser) component, water (central rising channels), oil-free water can be obtained. The component with the larger separation area is always more cleanly separated. Equal separation areas yield products without preference for one of the components. The setting of the rising channels cannot be changed if separators are used for different separation work. The exit diameter of the heavier phase is adjusted by a regulating ring such that the separation line is in the area of the rising channels.
Oil Purification
635
7.2.6.2Separators. Separators are used in the fats and oils industry to achieve the following: (i) separate liquids, for example, soap from oil after neutralization, water from oil after melting emulsion fats or skimming of milk; (ii) to separate small amounts of solids, for example, polishing of palm and olive oil or polishing of fish oil. If the task is to separate very small amounts of solids (Solids = S < 0.1% wdwt), discontinuous small separators are the choice. The solids are then collected on the inner wall of the bowl after being separated from the liquid and after centrifugation. If the sediment holding space is filled, the separator must be opened and the solids removed. Compared with self-discharging centrifuges, the advantages include a simpler construction combined with lower investment costs as well as much lower complexity and handling. The disadvantages are that more handling is needed and the process is run discontinuously. If the amount of solids is higher (0.1%< S < 20%), separators with self-discharging bowls are used. If the solids holding space is filled, the desludging mechanism starts. A piston opens and the solids are removed through the solids’ ejection port by the centrifugal forces of the rotating disks. Opening of the piston is done while working at full rotation speed. For polishing of pulp oils (palm and olive oil), separators with liquid discharge are used. Residual water can easily be removed in this manner. Tables 7.6 through 7.8 give some indication of the technical data for centrifugal separators for different applications and with different capacities. For the extraction of oils and fats in the fat industry, separators that enable the separation of the solids as well as the separation of oil and water phase in one go are usually used. In this application, the solids content is much higher than for polishing (Fig. 7.18) and thus other types must be used (Fig. 7.19). Centrifuges that are usually used for the extraction of slaughter fats have dimensions and capacities shown in Table 7.6. In addition to oil extraction and polishing of oils, separators are also used for refining (see Chapter 7.2) and for degumming (see Chapter 7.1). Special separators that have a capacity of up to 25 ton/h and are suitable for CIP are built for these purposes. Small amounts of solids are collected in the sediment holding TABLE 7.6 Capacities of Edible Oil Refining Centrifuges in Different Processesa Capacity (ton/d) for sepatator type Process Water degurnrning Acid degurnrning Neutralization (FFAc3%) Washing Winterization Cold refining Miscella refining (50% oil) aSource: Westfalia Separator AC, Oelde
RTA40
RTA60
RTA 140
75 75 75 100 50 50 60
150 120 150 150 60 60 100
240 240 240 42 0 150 150 200
RSE 100 RSE2OO
300 300 300 300 150 150 200
600 600 600 600 300 300 300
RSE250 1000 800 1000 1000 420 420 360
Fats and Oils Handbook
63 6
TABLE 7.7 Technical Data of Refining-Separatorsa ~~
~~
Type
RSA40
- 10,000
Rated capacity Bowl content Solids space Bowl speed, max. Motor Length Height Weight
20 5 6540 15 1645 1730 1420
RSA60
RSE 100
RSE2OO
RSE250
-1 5,000 20 5 6500 15 1645 1730 1430
-25,000 60 15 4050 30
-35,000 60 15 4550 37
-,60,000
2050
2050
2110
2350 1920
2350 2280
2080 3200
60 15 4800 60
aSource: Westfalia Separator.
space and are periodically discharged. The aqueous soap solution and the oil are separated under pressure with the aid of regulating rings (Table 7.7). 7.2.6.3 Decanters. If the solids content is >20%, decanters are used instead of the separators described above. They allow the separation of suspensions with a solids content 260%.In principle, they are solids-oriented separators. They are used mainly to separate fish oil from stick water and solids (see Chapter 3) and slaughter fats from greaves and water (see Chapter 3). As indicated earlier in the case of separators, the feed of the decanter and the discharge of liquid and solid are continuous. In addition, it is possible to split a liquid into a lighter and a heavier phase (3-phase decanter). In principle, a decanter is a horizontally fixed centrifuge. The separated solids are transported by a screw to the outlet (Fig. 7.20). Clarification of the feed liquid is performed in a separation zone where the solids (dark shaded) are separated on the bowl wall after they have been centrifuged. The liquid (light shaded) leaves the decanter in the right part, and the solids are conveyed off by a scroll, which is rotating slightly faster or slower than the bowl shell. That part of the bowl in which liquid is no longer found is called the drying zone. The separation line can be adjusted by means of regulating rings, comparable to those in separators. The drying zone is also called the “beach TABLE 7.8 Technical Data of Decanters (Type CA)a Maximum throughput
(Uh)
Bowl diameter Bowl speed, max. Energy requ. drive Length Width Height Weight (without drive)
(mm) (min-1) (kWh) (mm) (mm) (mm) (kg)
Westfalia Separator AC.
10,000
25,000
65,000
90,000
220 4500 11 1925 1180 600 700
354 4000 22-30 2275 1520 895 1680
650 2950 45-1 10 3 700 2645 1400 7500
650 2950 75-1 10 4300 2645 1400 8500
Oil Purification
63 7
1 Feed 2 Discharge 3 Turbidity meter 4 Disks 5 Sediment holding space 6 Sediment discharge 7 Operating-water valve 8 Drain hole 9 Opening chamber 10 Closing chamber 11 Annular piston 12 Timing unit 13 Discharge pump
Fig. 7.18. Clarifier (courtesy of Westfalia Separator AG, Oelde).
zone,” and the separating zone the “pool zone.” Decanter types with flat and steep angles can be distinguished (Fig. 7.21). Flat cone decanters have a larger drying zone ensuring a higher drying rate. In contrast, the polishing rate is higher for steep cone decanters. The slope of the scroll is determined by the kind of solids to be removedthe smaller the slope, the less the danger of stirring up sedimented solids. The same holds true for too high a differential speed between the scroll and the bowl shell. This difference is usually on the order of 20-80 rotations/min, depending on the product (-30/min for the preclarification of slaughter fats). Clarification decanters that polish only liquids have a throughput of up to 90,000 L/h (Table 7.8).
638
Fats and Oils Handbook
1 Feed, product 2 Fine tuner 3 Vapor seal 4 Hydrohermetic feed 5 Special disk S
W
6 Hydraulic ejection mechanism 7 Noise reduction 8 Gabtight frame 9 Three phase AC motor 10 Discharge, heavy phase product 11 Bowl flush 12 Discharge, light phase product
Fig. 7.19. Refining separator type RSE (courtesy of Westfalia Separator AG, Oeide).
7.3 Bleaching Oils and fats are bleached to remove undesired colorants in part because these colorants would negatively influence the taste of the oil and in part because the color would disturb the consumer. Therefore, on the whole, these colorants limit use and marketability. In addition, some of the p d c l e s that are removed during bleaching promote deteriora-
Oil Purification
639
Fig. 7.20. Decanter (courtesy of Westfalia Separator AC, Oelde).
Fig. 7.21. Schematic drawing of separator and clarifier bowls and of flat and steep angle decanters (courtesy of Westfalia Separator AC, Oelde).
Fats and Oils Handbook
640
tion of the oil mainly via their prooxidative properties. During bleachmg, the oil is brought into contact with a surface-active substance that adsorbs the undesired particles. The adsorbent and the adsorbed particles are filtered off, and the oil leaves the plant with the desired color. With the naked eye, the success of bleaching can be checked only via the color of the oil. This color is often measured to evaluate the success of bleaching (for example, Lovibond color, see Chapter 9.8). For more detailed investigations, the extinction of light of certain wavelengths is used. Rudischer (1959) described the course of bleaching, dependent on different parameters (Fig. 7.22). 7.3.1 The Theory of Bleaching
The colored particles (or substances) that should be removed during bleaching are present in the oil either dissolved or in a colloidal form. In addition to being a physical process, i.e., adsorption, bleaching is also a chemical process by the interaction of the colorants and the chemically active centers of the bleaching earths. In both cases, the process is conducted on the surface of the bleaching agent. In no case is the bleaching used for edible oils and fats similar to the usual bleaching with chemicals (for instance hypochlorite). The colored particles are bound only adsorptively (chemisorption or physical adsorption) and are removed by filtration together with the bleaching agent. This mechanism is also supported by the fact that the colorants can easily be removed from the bleaching earth if they do not partly decompose (observed with chlorophyll and carotenes). Mathematically, the bleaching process follows the Freundlich adsorp tion isotherm, which is an extension of the Langmuir equation. These equations are
Bleaching
................
A: time [min]
.....
B: temperature PC] C: earth p! ]..................
0
40 20 0.4
80
40 0.8
120 60 1.2
160 80 1.6
200 100 2.0
Fig. 7.22. Bleaching effect of an oil dependent on different processing parameters (after Rudischer 1959).
Oil Purification
64 1
valid for constant temperature and describe the dependency of the adsorbed amount of a substance ( k ) on its residual amount (c) in the solvent (in this case in the oil).
-=+)
b
k
[7.15]
k0
In simplified form, k, = a .c,b
[7.15a]
or, log k, = log a + log b log cr where the indices 0 and 1 indicate initial and relative amount, respectively, k , is the ratio of the adsorbed components, c is the amount of nonadsorbed component, and b are system-specific constants. If Equation [7.15] is plotted logarithmically, a straight line with slope b is obtained. However, in reality, the isotherms of Freundlich (1930) and Langmuir describe only the beginning of the adsorption process satisfactorily. If an amount rn of adsorbent (bleaching agent) is included, Equation [7.16] results: k = rn or
,
a . cb
[7.16]
log k - log rn = log a + b . log c
[7.16a]
If these bleaching trials are conducted with different bleaching earths, a is a measure of the (relative) amount of bleaching earth that has to be used to achieve the same bleaching result. To keep the product of rn and a constant, at a = 0.25. for example, the amount of bleaching earth must be four times as-high as at a = 1.0. Different researchers have determined the system constants a and b for different bleaching agents and oils (see, for example, Bailey 1951). For all of these trials, a was between 0.1 and 7.6. The numbers found for b were between 0.33 and 4.0. The difference in these numbers is caused in part by different methods to determine the bleaching success. Brimberg (1981) described the bleaching process using a formula developed by Berg and Ohlsson (1982) on the basis of experiments of Alfa Laval. [7..17] c
log-=10gk+0.5
log t
[7.17]
CO
In this equation, c is the concentration of the pigments, i.e., the components to be removed at the time t (index 0 is the start); t is the time passed since the addition of the bleaching agent, and k is a system-specific constant. The equation has been tested for chlorophyll and carotene.
Fats and Oils Handbook
642
7.3.1.1 The influence of Time. Brimberg (1982) published trials on the kinetics of bleaching. In the following, the influence of different process parameters on extinction values of the bleached oil is shown. The trials were conducted with rapeseed oil. The bleaching temperature was 80°C, and bleaching was performed with 1% bleaching earth, Tonsil Standard FF (Fig. 7.23). Both curves follow the findings indicated in Equation [7.17]. It can be seen that the negative slopes of the curves become flatter after a very steep decrease of the chlorophyll or carotene content; after 30 min, no further lightening can be achieved. The graph can be divided into three areas with very different slopes. According to recent theories, in the first phase, the stability of the colloid is broken when bleaching earth is added and the substances stabilizing the colloid are adsorbed. In the second phase, the colloidal particles form aggregates, which are also adsorbed; such particles could be made visible under a microscope. 7.3.1.2 The influence of the Amount of Bleaching Earth. According to Equation E7.161, apart from the dependency on time, there must also be a dependency on the amount of bleaching earth (Fig. 7.24). It can be shown that increasing the amount of bleaching earth improves the bleaching result for all bleaching times shown. However, these results, provided that a sufficient amount of bleaching earth is present, can also be achieved by increasing bleaching time. Here the commercial equilibrium balance between lower cycle times and higher cost for bleaching earth must be evaluated. 2.4
1% Bleaching earth, TONSIL Standard
2.0
5 E w
1.6
1.2
0.8
0
I
2 3 Bleaching time
4
[a, min]
5
6
Fig. 7.23. Bleaching result depending on the time (after Brimberg 1982).
Oil Purification
643
14
at 480nm (chlorophyll:
t 480nm (carotene)
2.0
2.0
-
Bleaching time [min]
1.5
1.5
1.o
1.o
0.5
Bleaching time [min] 1
0
2
3
0
0.5
2
1
Bleaching earth [% wtlwtof the oil]
Fig. 7.24.
Bleaching result depending on the amount of bleaching earth (after Brimberg 1981, Tollenaar and Hockmann 1964).
7.3.1.3 The lnfhence of Temperature. Bleaching temperatures are usually between 90 and 110°C (Fig. 7.25). Temperatures higher than 150°C must be avoided because changes in the structure of the fatty acids,may occur; isomerization reactions also might start. This does not hold true for the heat bleaching of palm oil (see Chapter 7.3.7).
-
Lovibond Color
ao
a
60
6
B
=40
4 8
%
2
20
0
0 0
20
40
60 80 100 I20 Bleaching temperature pC]
140
160
Fig. 7.25. Bleaching result depending on the temperature (after Bernardini 1985).
Fats and Oils Handbook
644
7 . 3 . 7 . 4 The lnfluence of Humidity. The amount of water in the oils to be bleached must not be too high or the oil would be hydrolyzed, catalyzed by the bleaching earth. However, a certain amount of humidity is required to optimize the activity of the earth. Undried bleaching earth can contain up to 10% water. Figure 7.26 shows the dependency of the reaction constant k from Equation [7.17] on the water content based on data determined by Brimberg using rape oil. It becomes clear that the speed of reaction increases with increasing humidity. The water comes from the oil (-0.2%) and also from the earth.
7.3.2 The Bleaching Agent As noted above, bleaching with bleaching earth is not bleaching in the original sense of the word, that is, by destruction of the colored particles; rather, it is removal of the colorants by adsorption on an adsorbent with a large surface area such as bleaching earth. As bleaching agents, natural bleaching earths can be used such as diatomic earth or Fuller’s earth. In addition, activated or synthetic bleaching earths as well as active carbon may be used.
7.3.2.7 Natural Bleaching Earth. The best known natural bleaching earth is Fuller’s earth. The name comes from a former use in wool processing (scarring or fulling). Fuller’s earth is an aluminum hydrosilicate that can be used without activation. Its mineralogical properties were described in detail by Kerr (1932) and Nutting (1933). Diatomic earth is also suitable for bleaching. Nonactivated earths play only a minor role and are not described further.
Fig. 7.26. Velocity constant for rapeseed oil bleaching (after Brimberg 19811.
Oil Purification
645
7.3.2.2 Activated Bleaching Earth. Activated bleaching earths are of natural origin. They are different from the nonactivated in that they have enhanced properties caused by an acid treatment. Bleaching earth used in Europe is mainly montmorillonite, an aluminum hydrosilicate (Si02:A120,, -4: 1). The aluminum in this earth can be partly substituted by magnesium or by iron. Zschau (1985) and Mag (1990) described the structure of bleaching earths (Fig. 7.27).
OH OH
Water and exchangeable cations
4
1 0
H2O
1 ‘ 1 Fig. 7.27. Structure of Montmorillonite (redrawn; after Zschau 1985).
Fats and Oils Handbook
646
The individual layers are negatively charged because Si atoms are partly replaced by A1 or Fe atoms. The A1 atoms may then be partly replaced by Mg atoms. Interlayer cations create the electrical charge, making that mineral a natural ion exchanger. To be used as bleaching earth, the earth must be pretreated. First, the raw mineral is cleansed of impurities. Second, the surface of the bleaching earth is increased by an acid treatment to enable better properties for the bleaching process. For this activation, the earth is suspended in water and treated with a mineral acid. The acid attacks the octahedral layer of the montmorillonite, thereby increasing surface area. From in between the layers, cations are released and replaced by protons, thus creating acidic exchange sites. In addition, a part of the metal ions from the layer structure is dissolved and the surface is thus increased. The acid is removed from the solid suspension via filter presses and the filter cake is washed almost acid free with water. Then the filter cake (water content -40%) is dried to a humidity of 6-10% and then ground. In this process, water that is chemically bound has to be retained in the earth, whereas water that is adsorptively bound has to be removed. During the drying process, particles smaller than 1 pm agglomerate to larger particles (Fig. 7 . 2 8 ) . A typical bleaching earth (Tonsil, Siidchemie AG, Miinchen) has the following approximate composition: 70% as Montmorillonit
(Aluminum hydrosilite)
U Blending Standardizing
Clays ofdifferent origin
HCI or HaSG (-30%). ca. 105'C
Washing
to acid level < 0.1% (calculated as HCI) Water content -40%
Burning
350400'C in gas current to 5 8 . 5 % humidity
I Airclarification I Actiiated bleaching earth
Fig. 7.28. Flow chart of bleaching earth production and activation and 30,OOOX rnicrophotograph of bleaching earth before (above) and after (below) activation (photos: courtesy of Sudchernie AG, Munchen).
Oil Purification
647
SiO,, 2% MgO, 15% A1,0,, 1% CaO, 4% Fe,O, and -7% calcination loss. By activation, the surface area of the bleaching earth can be increased from 10-20 to 200-300 mYg. 7.3.2.3 Synthetic Bleaching Earth. Natural bleaching earths have a given structure that can be changed only within certain limits. For example, activation can increase their surface area but the pore size and pore size distribution remain unchanged. For the bleaching process, an optimum pore size deviating only within narrow limits would be very advantageous. However, this can be achieved only if bleaching earths are synthesized. In addition, recycling of naturally activated bleaching earth for repetitive use is a problem because this process is very costly and the active surface decreases considerably. Disposal of bleaching earth becomes increasingly expensive; therefore, interest is rising in synthetic bleaching earth, which can be recycled. Such bleaching earths have been developed but with no real breakthrough as yet because either they are not very convincing in their properties or their application is commercially nonviable. However, it is obvious that in the near future synthetic bleaching earths that combine good bleaching properties with acceptable cost will be available. The synthetic bleaching agents that have continuously been developed further include TriSyl (Grace Davison) and SorbSyl (Crosfields). These are amorphous silica adsorbents with an average particle size of -19 Fm and are claimed to achieve a bleaching effect equivalent to that of clay at a much lower dosage; phospholipids are much better removed with silica than with clay. 7.3.2.4 Active Carbon. Active carbon is very rarely used alone as an adsorbent. In oils and fats that are very difficult to bleach, active carbon is used in combination with bleaching earth (carbon:earth, 1:9). Using this combination of adsorbents, polycyclic aromatic hydrocarbons (PAH) can almost completely be removed from the oil even if present in higher quantities. Many qualities of active carbon are available commercially. If used for edible oils, it is very important to check these qualities thoroughly to avoid the adsorbent itself containing aromatic hydrocarbons.
-
7.3.2.5 Silicates. Some manufacturers have launched amorphous silicates for which good bleaching properties are claimed. Such silicates have an average particle size of -20 p m and 99% of their dry matter consists of S O 2 . If blended with bleaching earth, it is claimed that they reduce the amount of bleaching earth needed to -50% of current needs. The other 50% can be replaced by 15-20% silicates. Up to now, there is an insufficient number of large-scale trials to determine if silicates are an alternative. 7.3.2.6 Comparing the Bleaching Agents. Natural bleaching earth can adsorb up to
15% of its own weight in colorants (Siddiqui and Hasnuddin 1968). It binds up to 30% of its own weight in oil. Because of its increased surface area, activated bleaching earth adsorbs 70-100% of its own weight. Using active carbon it can be up to 170%.
648
Fats and Oils Handbook
As a result of the higher activity, activated bleaching earth can split soaps (ion exchanger effect), thus increasing the FFA. This is one of the reasons why the oil should be soapfree for bleaching. The other reason is that all bleaching agents adsorb soap, which can clog their surface and thus inactivate a considerable part of the earth. Above 150°C, there is a danger with all types of activated earth that isomerization of the fatty acids will occur (Patterson 1976). This temperature, however, is so far above the common processing temperatures that structural changes cannot occur provided processing is done properly. Table 7.9 gives properties of typical bleaching agents. 7.3.3 The Bleaching Process and Bleaching Plants
As for aLl processing steps in refining, discontinuous and continuous bleaching operations exist. There is a development toward continuous processing provided frequent changeovers can be avoided. Figure 7.29 shows the bleaching process schematically. The oil requires careful pretreatment (see Chapter 7.3.4) to achieve optimal results with a minimum of bleaching earth. It has to be taken into consideration that not only the earth itself is a cost factor, but also the loss of the oil that is bound to it. The amount of bleaching earths used is between 0.5 and 2.0% (wt/wt of the oil/fat) with a typical value of -1%. For the processing of oils of poor quality or high concentrations of environmental contaminants, active carbon can be added. The carbon then accounts for 10% of the total bleaching agent. If raw materials have been dried over an open fire (copra at times, see Chapter 4.3.6), the fat may contain fiveringed polycyclic components. These can be removed satisfactorily from the fat with the help of active carbon (0.4%; Biernoth and Rost 1968, see Chapter 7.5.7.). Bleaching earth flows easily; thus, it can be dosed under vacuum without difficulties. Plants also exist in which an ofileaching earth slurry is produced, which is then TABLE 7.9 Properties of Bleaching Agents, Typical Value9 Bleaching earth Property p H in aqueous suspension Bulk density (g/L) Surface area (rnVg) Particle size (pn) >80 (%) (Yo) 4C-80 20-40 (Yo) >20 (O/O) 50 (Yo) aSource: Patterson (1976).
Native
Activated
Active carbon
8 0.684.90 68
2.8-6.0 0.32-0.68 165-310
6.0-1 0.0 0.384.43 500-900
19 20 19 42
10-15 20-25 25-30 30-40 30-40 40-50
30-1 0
Oil Purification
649
Neutralized Oil
I
Heatiy Up
1 7MIO.C Vacuum (-XI hPa), humidm < 0 2%
w t b i of bleaching earth)
Plate-, cartndge-, dirk- filters
Polishing filter
(removal of finest bbaching earth pskkts)
Bleached oil
Disposal
(Filter oil)
Fig. 7.29. Processing flow chart of bleaching.
pumped into the bleaching vessel. For batchwise processing, this process is too complicated and not necessary. In continuous bleaching, however, it is necessary because only in this way can the exact dosing of bleaching earth be ensured. Oxygen from the air must be absent during bleaching or the oil may be oxidized. Activated bleaching earth can thus act as a catalyst; for that reason, bleaching is normally conducted under vacuum. Typical temperatures applied for bleaching are between 90 and 110°C. Temperatures exceeding 150°C must be avoided because they are sufficiently high to change the structure of the fatty acids by isomerization reactions. For heat bleaching of palm oil, this does not hold true. After bleaching, the oil is filtered to remove the bleaching earth (see Chapter 7.3.5.), which is then discarded after the adsorbed oil has been extracted. Sometimes this extraction is not worthwhile and the spent earth is disposed of (see Chapter 7.3.6)To remove the finest bleaching earth particles, the oil has to pass through polishing filters before being pumped to storage.
7.3.3.7Plants for Discontinuous Bleaching. Plants for discontinuous bleaching are the same as those for discontinuous neutralization. Most often, these processing steps are carried out one after the other in the same batch vessel after the neutral-
650
Fats and Oils Handbook
Fig. 7.30. Plant for discontinuous bleaching.
ized oil has been prepared for bleaching (see Chapter 7.3.4.). The vessels themselves are therefore often called neutralizershleachers. If the oil is additionally interesterified, this can also be done discontinuously in the same vessel after neutralization but before bleaching. Figure 7.30 shows a plant for discontinuous batchwise bleaching of oils and fats. The plant shown has different vessels for neutralization (1) and bleaching (3). The oil is pumped from the neutralizer into the bleacher after the soap has been separated and pumped into storage (2). After the oil is dried, the bleaching agent is sucked into the bleaching vessel (3) by means of a vacuum. After being stirred for some time, the mixture is pumped with a pump (4) into the filter presses ( 5 ) where the bleaching agent is removed. The oil is collected in the filter oil vessel and intermediately stored before deodorization. The sizes of bleachers are between 10 and 40 ton. As an example, Figure 7.31 shows the processing cycles for bleaching of an oil in a neutralizerhixer. 7.3.3.2 Plants for Semicontinuous Bleaching. As shown in Figure 7.31, the process consists of four major steps, namely, heating and drying, settling and stirring, cooling and filtration. If these four steps are carried out one after the other in four stacked trays housed in a vertical vessel, a semicontinuous process can be performed (Fig. 7.32). The neutralized oil is pumped into the upper tray of the bleacher and is steam-heated to bleaching temperature. The tray is evacuated during this operation. After the oil has been dried and heated up, it flows into the second tray where the bleaching earth/oil slurry is dosed. Here it is held for the reaction period
Oil Purification
65 1
Stirring, -40min”
dependkg on
Filter oil
(Z5@500 L/rr+h)
(nwimbd,bbschKl)
Fig. 7.31. Processing flow chart of batch bleaching.
while being stirred. After that, the oil falls into tray three where it is cooled to filtration temperature. The lowest tray is used as a buffer for filtration. Athanassiadis ( 1980) gave the average energy consumptions for bleaching including neutralization using 1% bleaching earth and hermetic filters (Table 7.10). If hermetic filters are steamed before discharge of the bleaching earth, the amount of steam has to be increased by 15 kg/ton of oil. The figures in the table do not include the part of the energy that is required to extract the adsorbed oil from the bleaching earth.
7.3.’3.3Plants for Continuous Bleaching. Continuous bleaching plants are commercially viable if long runs of one and the same oil must be bleached. Their work-
Fig. 7.32. Semicontinuous bleaching plant (courtesy of Lurgi GmbH, Frankfurt).
Fats and Oils Handbook
652
TABLE 7.1 0 Energy Consumption for Neutralization and Bleaching, Inclusive of Washing and Dryinga Heat recovery Energy consumption (kg steamlton of oil) for Heating of crude oil (20°C Hot water production Vacuum production Total
+ 110°C)
Without
With
88 30 50 168
38 18 35
91
aSource: Athanassiadis (1980).
ing principle is based on the fact that the oil to be bleached is blended with an oiybleaching earth slurry and passes through the reaction vessel under predetermined conditions. During this passage, the four steps of bleaching outlined above are carried out consecutively. Like all continuous processes, this process can also be easily automated and, as a matter of routine, can be run by one person. However, this is also possible with semicontinuous or discontinuous plants if they are well designed. Figure 7.33 shows the configuration of a continuous bleaching plant. The oil is pumped from a vacuum dryer (not shown) into the bleaching vessel while passing through a heat exchanger on its way. An aliquot of the oil is pumped into a vessel where bleaching earth is dosed from the bleaching earth storage. A vacuum is applied to degas the oiybleaching earth slurry, which is pumped into the bleaching vessel. The slurry falls on a rotating disk or passes through stirred compartments in order to be well agitated, thereby ensuring intense contact. The bleaching vessel is separated by a vertical wall to force a zig-zag flow, thus increasing the time required to pass through it. Afterwards the oil passes through the plate heat exchanger in a countercurrent manner to the incoming oil for heat transfer, in this case for cooling. It is pumped to a pair of filters that are operated alternately to avoid standstill during the time of earth discharge. Different types of filters (see Chapter 7.3.5) can be used. After passing through a polishing filter, the oil is stored (filter oil storage). The earth is collected and discharged or the adsorbed oil is extracted. 7.3.4 Pretreatment of the Oils to Be Bleached
To ensure an optimum effect of the bleaching earth, it is necessary to avoid reducing its activity as a result of impurities in the oil or inadequate processing conditions. This is especially important because it is the only way to minimize the consumption of bleaching earth. High amounts of bleaching earth-in addition to the cost of the earth itself-always cause high oil losses. This adsorbed oil suffers quality deterioration even if recycled via extraction; it is therefore often no longer considered as food grade. In addition, extracting the oil is very cost intensive. One of the factors that negatively influence the activity of the earth is the oil’s soap content. The acid groups of the activated montmorillonite are able to promote
653
Oil Purification
BLEACHER
Fig. 7.33. Continuous bleaching plant (left; redrawn; courtesy of Lurgi CmbH, Frankfurt) and continuous bleacher (redrawn; courtesy of De Smet S.A., Edegem).
soap splitting. On the one hand, FFA are formed, which adversely affect the value of the oil and reduce the activity of the earth by blocking the active centers of the earth. On the other hand, the acid groups are inactivated by a neutralization via ion exchange of H+vs. Na+. The soap content should therefore be as low as possible, meaning that the oil has to be washed carefully. Careful neutralization is indeed a prerequisite for successful bleaching. The water content of the oils also plays a role because fats can be hydrolyzed under bleaching conditions (90-1 10°C, bleaching earth catalysis). However, a certain humidity is necessary to increase the activity of the earth. It should not exceed 0.1%. Because bleaching earth is also a good oxidation catalyst (Zschau 1985), the working atmosphere should be oxygen free. It is therefore preferred to work under vacuum. This is the only way to ensure that conjugated dienes and polyenes are not formed. If present, they are formed by decomposition of hydroperoxides. Most bleaching plants are run under a vacuum of 30-40 hPa. 7.3.5 The Filtration of the Bleaching Agent
Filtration through a porous medium depends on the filtration area (F),the filtration time (0,the pressure difference (Ap), the viscosity of the filtrate (E), the thickness
654
Fats and Oils Handbook
(T) of the porous medium, and its permeability (P).Bringing all of these parameters together in one equation results in the following (Darcy’s Law):
[7.18]
The quotient of AV and At is the specific flow rate, which is volume per time and per area. Multiplied by the area F, this leads to the flux rate of the filter as follows: [7.19]
If the quotient of permeability and thickness of the filter medium PIT is replaced by the sum of the resistance of the apparatus and the cake RA + RK, Equation [7.20] results:
[7.20] Assuming a homogeneous distribution of the liquid, Equation [7.20] can further be developed by introducing it into Equation [7.21] V RK = r . m F
[7.21]
where r is the specific resistance of the filter cake and m is the proportion of solids per volume of filtered suspension. The equation is valid under the assumption that the filter cake is incompressible. Equation [7.20] becomes AV -F . At
-
Ap (RA+ r . mV/F) . &
[7.22]
From this equation, the parameters of filtration and their influence can be read. An overview of recent developments of filtration techniques was given by Esser et al. (1985) and Gosele (1987); a special review on filtration techniques in the bleaching process was given by DGF (1996). Figure 7.34 shows the formation of the cake with and without filter aid. The filter has to be precoated (“black run”); then particles smaller than the cloth’s mesh size are retained by the bridges of larger particles sticking together, holding back the smaller ones.
Oil Purification
655
Cake forrnation (precoating:
.Bleachingeiaith partidea 'I
Cake formation with filter aid
Fig. 7.34. Cake formation with
and without filter aid (after DCF 1996).
7.3.5.7 Plate Filter Presses. In many factories, bleaching earth is still filtered via plate filter presses. Plate filter presses consist of a horizontal stack of hollow plates equipped with filter cloths. The surface of the plates is porous so that the oil can be pressed into the inside of the plate and the filter cloth holds back the bleaching agent. With old plates, which are usually manufactured from cast iron, every single plate has a side exit that enables the oil to drain off. The oil leaves the plate via a tap and flows into an open drain that runs along the whole side of the press (Fig. 7.35). More recent plates are constructed with a central hole, which forms a pipe when many plates are tightly pressed together. The oil then drains off via that pipe without coming into contact with the oxygen of the air. After the filter plates are fully loaded with bleachtng earth so that it has to be removed, the plates from the stack are separated from each other. At one time, this was done manually; today, the plates are separated via a motor-driven chain. This chain saves the hard manual labor and ensures that the plates are an equal distance from each other. Underneath the plates is a large container into which the bleaching agent falls. This container can be opened on the lower side, and the bleaching agent can thus be discharged, falting from the plate press via the container into small carts, or it is removed by means of a screw. If the bleaching agent does not loosen from the plates, it is manually removed by means of a knife-like device. To ensure good functionality, it is essential to use a suitable filter cloth. This cloth must have the following properties: high mechanical resistance, suitable pore size, low tendency to becoming stuck and long standing time.
656
Fats and Oils Handbook
Fig. 7.35. Plate filter press (schematic)and traditional filtration plant with open outlet. Mechanical resistance is necessary because, on the one hand, the cloth has to withstand the forces imposed on it by filtration; these are not too high because the plate acts as a support. On the other hand, however, it has to withstand the mechanical drawing off of the earth and repetitive cleaning. If monofile filter cloths are used, it is easily possible to achieve mesh sizes as small as 30pm.Such cloths can withstand pressure up to 500 bar (Hermann 1988). Such high pressures, of course, are not necessary for the filtration of bleaching earth or hardening catalyst; however, they are a measure of the stability of the fabric. A suitable pore size reflects a compromise between high filtration capacity (large pore size) and short black run (small pore size). In the phase of filter cake formation via bridge building of the solid particles, many of these solid particles pass through the filter. This is the so-called black run because the bleaching earth is dark colored due to the adsorbed color particles. If active carbon is added, it is really black. Every filter must allow recirculation until the black run phase is completed and those parts that lie behind the filter cloth are washed clear again. The actual filtration is made possible by solid particle bridges that are formed in the course of the black run. As a rule of thumb it can be said that the pores of a filter cloth should be two to four times larger than the average size of the solid particles to be filtered off. If they are smaller than that, the pores easily become blocked and the filtration capacity is decreased. If pores are larger, the black run is prolonged and sudden increases in pressure will lead to a breakthrough of the filter cake and additional black run during filtration. Clogging of the filter cloths can be minimized if fibrous fabrics are not used; such materials have a tendency to trap solid particles, leading to stuck filters. Materials that are well dited for a filter cloth are linen as well as nylon and perlon,
657
Oil Purification
Filtration capacity [kg/m2.h]
1200
I
800-
1 \ -- -. \ 2.6 ba;-
600 -
- -\
I
400 -
5.2 bar
6.1 bar
which have mainly replaced linen today. Figure 7.36 shows the filtration capacity of an earth I which contains almost 50% solid particles with a size 4 y m and an earth I1 having only 10% solids of that size. In the beginning, earth I has a high filtration capacity that quickly decreases as the pores become stuck; at a cake thickness of 10 mm, no further satisfactory results can be obtained. Earth I1 can be filtered to a cake thickness almost twice that. High standing time has double importance. On the one hand, this means that the filter cloth can withstand many cycles of filtration and cleaning. On the other hand, the number of cycles before the cloth has to be removed for cleaning is high because the plant stands idle during changing of the cloths. Generally, working with plate presses has the disadvantage that it is very labor intensive and cannot be automated. However, these filters can be operated by less skilled personnel and have relatively low investment costs. 7.3.5.2 Continuous Filters. The filters that are called continuous are not actually continuous in reality because the bleaching agent has to be discharged when the filter is filled. This means that, in this case also, filtration has to be intermpted to discharge the solids. Compared with the filters described above, this needs only a little time to be completely automated. This represents not only great commercial progress but also a substantial improvement in the working conditions because working with the old filter presses normally took place under difficult circumstances, i.e., hard labor at high temperatures. As noted above, the continuous filters also need discharging time and, after restarting, there is a black running time. Figure 7.37 shows the cycle times of filters. To bridge the time span required for filtration, two filters can be fed by one bleacher. They then work alternately, which means that while one filter is emptied and restarted, filtration is done via the other one and vice versa.
Fats and Oils Handbook
658
Mixture of
Duration of prcceoa rhpa [min] Cartridget.
I
Fibrprassl
Dirk filter
Filter filling
5 1
Dhcha~ingcake
3 14-1s
Oil
~
12
6
2-3
36
27-33
TOM q ~ kna k excluding kitration mtf
Fig. 7.37. Cycle time of different filter types for bleaching earth'filtration.
7.3.5.2.7 Candle filters. Candle filters consist of a number of porous tubes that are closed on their lower end and hang vertically in a hermetically closed vessel. The upper end of all of these tubes is open and connected to a piping system, which allows the liquid inside the tubes to drain off, but is closed toward the vessel. As an example, the Fundabac filter (see Fig. 7.38 and Table 7.11) consists of a number of bags stiffened by multitube elements that are systematically build into a pressure vessel. A single bag support system consists of seven tubes of equal dimensions. Six of these tubes (C) are arranged concentrically around one central tube (B). The six tubes are provided with horizontal draining slots (D). The tube bundles are covered with a suitable highpressure weave @), which fits snugly around the curved surfaces during filtration. The filter cake is built up on this weave (G).During the back wash, cycle gas is blown through the central tube (Aa) into the outer tubes (H) and through their slots (Da). Thus the weave is lifted off (Ea) the tubes, and this movement causes the cake residue to be thrown off (Fa). The filter medium is of circular weave, i.e., the medium forms a seamless hose which withstands a back wash pressure of several bar and can thus be thoroughly cleaned. The yarn is either monofile and/or multifile and meets all process conditions. The bag is clamp-futed to the tube assembly at both ends. The filtrate exits via the central tube, which also serves as the feed line for the back wash medium. The filtrate flows vertically down the external surface of the concentric tubes (H) toward the central tube through which the filtrate exits upwards (A) into the register pipe system. As a filter cloth surrounding the tubes, Fundabac, for example, uses a wire of 100-pm mesh size for the filtration of crude oil and 2 0 0 - p mesh for the filtration of bleaching earth. Amafiiter uses a cotton cloth with filter channels of l-lOqLm; however, the smaller mesh sizes serve other purposes than the filtration of bleaching earth. For filtration, the vessel is filled with the oil/bleaching earth slurry and the filter is thus precoated. After a thin filter cake has formed on the cloth, filtration can be started. The oil passes through the cloth and enters the central tube via the porous tubes. It is col-
Oil Purification
659
Fig. 7.38. Candle filter Fundabac type (courtesy of DrM, Dr. Muller AG, Mannedorf).
lected in the central tube from where it flows through the register and drains off. Lf increasing back pressure indicates that the filter cake is thick enough to require discharge, feeding of the oil/bleaching earth slurry to the filtration chamber is stopped and the chamber is emptied. Back washing is usually done with pulsed blows of nitrogen. Pulsing significantly amplifies the washing effect on the hose weave. The pulsating
0
0 0
TABLE 7.1 1 Technical Data of Candle Filter& Filtration area
(m2)
12
15
23
23
31
Total height Volume Weight Diameter Number of candles Number of registers Candle length
(mm)
2641 2342 975 1200 36 6 1250
3041 2794 1070 1200 36 6 1650
3891 3755 1271 1200 36 6 2500
3068 4520 1717 1600 72 10 1250
3568 5324 1874 1600 72 10 1650
(L)
(k@ (mm)
36 6 (mm)
46 4318 7032 2210 1600 72 10 2500
48
63
95
3636 9305 3298 2200 148 14 1250
4036 10,759 3553 2200 148 14 1650
4886 13,841 4093 2200 148 14 2500
2 i:
T a k e thickness, 2&30 mm; throughput 40&500 Um2; filter type Fundabac.
Iu
9
TABLE 7.1 2
9 z
Technical Data of Plate Filter&
1: Vertical: maximum capacity (Uh)
2.5 Filter area Tank diameter Tank capacity Number of plates Length Width Height Weight (no drive) aSoybeanoil with 1% of bleaching earth. bNiagara type filter, courtesy of Amafilter b.v., Alkmaar.
Horizontal: maximum capacity (Uh)
5.0
7.5
12.5
2.5
7.5
15.0
25.0
20 1070 1.6 13 2 700 1900 3280 1400
30 1220 2.4 15 3000 2150 3540 2200
50 1500 4.3 17 2500 2600 3850 2650
10 914 1.1 15 3730 1650 2135 1300
30 1220 2.6 22 5590 2200 2650 21 50
60 1220 4.5 44 8890 2220 2650 2780
100 1530 8.0 49 10,010 2500 3050 4000
8 0 2.
Oil Purification
661
effect is brought about by a motor-driven rotating ball valve in the air feed line. The washing effect results from flutter of the hose caused by pulsating the gas stream. Instead of nitrogen, steam can also be used. Air is strictly forbidden because back washing with air could lead to immediate ignition of the hot oil that is finely dispersed over the large surface area of the bleaching agent. This kind of filtration can be repeated, fully automated, as often as desired and therefore constitutes in effect a continuous process. One of the advantages of that kind of filter and the vertical Niagara filter (see Chapter 7.3.5.2.2.)is that the candle filter has no movable or motor-driven parts. 7.3.5.2.2 Tank filters (leaf filters). Tank filters follow similar construction principles as those applied for candle filters. Equivalent to the so-called candles, filter leaves hang in the tank. The best known filter of that type is the Niagara filter of Amafilter (Fig. 7.39). This type of filter is operated in the same way as described for the candle filters. The difference is in the cake discharge, which is achieved by means of a pneumatically operated leaf vibrator. In the horizontal types, discharge is achieved via a manual cake door, which may also be operated by a hydraulic drive. The cover of the tank is closed by eyebolts and hand wheels and can be fully opened, ensuring easy access to its interior. Tank filters are delivered with filtration areas of up to 50 m2, Normal working pressure is 4.5 bar with a maximum of up to 10 bar. Under suitable processing conditions, the filters can be operated for up to 2.5 h without bleaching earth discharge. The separated cake contains up to 25% oil. If such filters are used for crude oil filtration, the solids content should not exceed 7% and the operation temperature should be 4 0 ° C . If nickel hardening catalysts are filtered, a cake can be formed if the oil contains 9 . 3 % solids (ten Hage 1986). Such filter cakes hold 40% oil. Applying winterization filtration, the standing time of the filter can be up to 8 h between two cake discharges. Niagara filters are also available as horizontal filters in which case the bleaching earth is mechanically separated from the filter leaves. As a result of the lighter construction made possible by the horizontal design, the filtration area can be up to 100 m* (Fig. 7.39). For discharge, the filter is opened at the side and the stack of leaves is driven out of the tank. Technical data of vertical and horizontal tank filters are summarized in Table 7.12 (p. 660). 7.3.5.2.3 Centrifugal discharge filter. Centrifugal discharge filters are also called disk-filters. The filter consists of a pressure tank with a central hollow shaft on which a series of disk-filter elements are arranged at specific intervals. The tank is usually pressure tight to 5 bar (special make to 30 bar) and can be used up to 110°C (Fig. 7.40). The filter stack, comprising the hollow shaft and the elements, is located so that it can rotate within the tank. The filter elements are fitted with woven wire, textile material, sintered metal or perforated plates, depending on the requirements. The drive of the filter stack is either mechanical, using hydraulic coupling and reduction gear box, or a steplessly variable hydrostatic drive.
662
Fats and Oils Handbook
Fig. 7.39. Vertical and horizontal tank filters with filter leaves (type Niagara; courtesy of Amafilter b.v., Alkrnaar).
As with candle filters, the filter chamber is filled with the oivbleaching earth slurry until the filter-disks are precoated. If the filter cake is thick enough to be discharged, the oiyearth slurry not yet filtered is pumped off the filter chamber. Then the whole stack is spun by means of a drive system. The cake is then thrown off by the centrifugal forces and crashes against the wall of the tank where it cracks and falls to its bottom. Emptying the filter takes. 1-2 min. Then, the filtration chamber is filled again and
Oil Purification
663
Fig. 7.40. Centrifugal discharge filters (disk filters); types Funda (upper right) and Schenk; A = hydrodrive or electromotor; 6 = shaft suspension; C = upper bearing and seal; D = filter plates; E = lower bearing and seal; F = filtrate nozzle; G = slurry discharge; (courtesy of Schenk GmbH, Schwabisch Cmund and Dr. Muller AG, MannedorD.
Fats and Oils Handbook
664
TABLE 7.1 3 Technical Data of Schenk Disk Filtersa Filter area
(-m*)
5
10
15
30
40
50
Total height with drive Vessel volume Weight with drive Drive Vessel diameter Disk diameter Number of disks Turbid volume
(mm) (L) (k!4
1735 350 1200 7.5 900 805 10 0.15
2100 700 1600 11
2510 1000 2000 15 900
2850 1800 4700 37 1350 1200 28 0.9
3230 2200 5050 37 1350 1200 37 1.2
3580 2900 5350 37 1350 1200 47 1.5
(kw) (mm) (mm) (m3)
900
805 21 0.3
805
32 0.45
aCourtesy of Schenk GmbH, Schwabisch Gmund.
filtration is continued. The thickness of the filter cake is at most 30 mm with a spacing between the disks of roughly 40 mm. This is equivalent to -20 kg of bleaching earth per square meter of filter area. Mesh size of the filter is 5CL130"m. This leads to a filtration capacity depending on the earth and thickness of the cake of 300-600 U(h . m2) of filtration area if run continuously. Approximately twice that flow rate is needed to ensure good precoating (duration -5 min). If individual batches of oil have to be filtered, the filtration capacity is much lower. The technical data of such filters are summarized in Table 7.13. 7.3.6 Recovery of Oil from Spent Bleaching Earth
Assuming that the average amount of bleaching earth applied is 1% (wt/wt of the oil) and that bleaching earth adsorbs approximately its own weight in oil, 1% of the bleached oil would be lost if not recovered. There are two principal ways to recover the oil, namely, to extract the earth or to drive out the oil from the earth with water (steam) and a surface-active agent. To ensure that the extracted oil is edible grade, extraction must be carried out immediately after bleaching. Otherwise, the oil quality worsens because of the large surface exposed of the adsorbed oil to the oxygen of the ambient air (Table 7.14). After recovery of the oil, the bleaching earth is mainly deposited (special waste). Following recent developments, it can also be used as a raw material for the cement and brick industry. These industries, however, are interested in the earth only if it still contains the oil that then serves as a fuel to heat up the ovens, Recently, it has also been found that if bleaching earth is composted, it serves well to loosen the compost. 7.3.6.1 Water Treatment. If the oil is driven out of the earth with hot water, oils of restricted quality are obtained because this process extracts not only the oil but also adsorbed polar substances such as oxidation products. The advantage of this processes is that, contrary to solvent extraction, explosive solvents are avoided. To achieve good results, a processing aid, usually a detergent, has to be added to the hot water. A second advantage of this process is that the solid earth that is obtained can easily be
Oil Purification
665
TABLE 7.14
Influence of Storage Time on Quality and Quantity of Extracted OilaTb Storage time of bleaching earth (d) Soybean oil extracted from bleaching earth with 30% oil content
0 1 2
Extraction yield (YO) 95 92 83
5
50
10
25
Comparison: fresh oil, neutralized, bleached
Color
FFA (%I
1 .O d 4.2 d 13 i 23 i 100 i
6.01 13.9
2.2 d
0.09
0.11 0.54 3.14
ad,dichromate;
i, iodine color; FFA, free fatty acids. bSource:Gander (1 969).
deposited, but this advantage has to be paid for by a high effluent load that requires purification. In this process, residual oil in the bleaching earth lies between 5 and 50% (wt/wt of the earth). 7.3.6.2 Extraction. Like oilseeds, bleaching earth can also be extracted with solvents. If bleaching is done in the oil mill, bleaching earth is often combined with the crude seed and extracted again. In stand-alone refineries, the earth is extracted as such. In principle, the process follows the same pattern as that for the extraction of seeds (see Chapter 5 ) . Because of the much smaller amount, discontinuous batch extractors are used almost exclusively. If solvent extraction is done in a refinery, usually the building for that process is detached and erected at some distance to the rest of the plant. Not only would damage be minimized if the plant were to explode, but also the complete refinery plant would otherwise have to be built explosion-proof and would have to be run under difficult circumstances. Another possibility is to extract the earth directly in the filter after the oil has been filtered off. To do so, the oil-free filter is filled with hexane, extracted, the miscella is pumped off and the filter is steam blown to evaporate the residual hexane. The extracted earth should be removed under inert gas atmosphere. 7.3.6.3 Other Processes. Weber (1980) reported a new process, offered under the name of Contiblex by Extechnic (now Krupp Maschinentechnik). This process is a combination of extraction and hot water treatment. The aim is to combine the advantages of both systems without their disadvantages. The process is run in a combined extraction separation column. The bleaching earth is suspended in a solvent (usually hexane) and fed to the column in a finely dispersed form. The column is half filled with water and covered with the same volume of hexane (Fig. 7.41).The suspension of finely dispersed bleaching earth is dosed under the hexane surface. The bleaching earth sediments and is extracted on its way downward. From the bottom, fresh hexane is fed and, at the top, the miscella is drained off. In
Fats and Oils Handbook
666
Oily bleaching earth
Fig. 7.41. Plant for bleaching earth recovery with the Contiblex process (courtesy of Ex Technik, now Krupp).
countercurrent extraction. During sedimentation of the extracted bleaching earth through the water layer, the solvent is dnven out of the earth. The sedimented layer, once it has reached a certain thickness, is impermeable to water and can be removed with a screw. The miscella is stripped, and the water vapor condensate formed is used to refill the water reservoir in the column because part of the water is carried out together with the bleaching earth. The separation of hexane and oil follows the principles described in Chapter 5. The above-described plant delivers a miscella with 1&15% oil and extracted bleaching earth with a residual oil content of 0.1% (optimum). The working temperature lies between 40 and 50°C. The manufacturers claim a consumption of 800 kg of steam, 30 m3 of cooling water and 30 k W h of electrical energy to process loo0 kg of bleaching earth. The advantage of that process is that good oil quality is combined with a bleaching earth that can easily be deposited without creating effluent problems. 7.3.7 Heat Bleaching
The deodorization step carried out at 250-260°C during physical refining of palm oil (see Chapter 7.5.3.) is called heat bleaching.
Oil Purification
667
7.3.7.7 Theory of Heat Bleaching. At high temperature, carotenes are thermally decomposed. According to Young (1978), temperatures of 260°C are required for this process. Carotenes are present in palm oil in substantial amounts (500-600 ppm) and are responsible for the orange red color of crude palm oil. 7.3.7.2 Pretreatment of Palm Oil for Heat Bleaching. Heat bleaching of palm oil requires some pretreatment. Although the amount of phosphatides in palm oil is very small, the oil should be degummed to ensure good refining results. Because of the low amount of gums, no filterable precipitate can be formed. Therefore, some bleaching earth is added to make the phosphatides filterable after an excess of phosphoric acid has been neutralized. As a second effect, the bleaching earth prebleaches the oil. At 1 1O-12O0C, part of the carotenoids is adsorbed. The process conditions of palm oil heat treatment are described in more detail in Chapter 7.5 because bleaching is carried out simultaneously with physical refining. The flow chart (Fig. 7.42) shows mainly the pretreatment steps. In some Asian countries, a combined process for palm oil bleaching is applied. In this process, a first bleaching is carried out at 150-160°C with 3-5% bleaching earth. Then the oil is deodorized batchwise at 190-200°C. As can be expected, the bleaching result depends on time and bleaching temperature. High temperature ensures good success even within 5 min (Fig. 7.43).
7.4 Deodorization Oils and fats contain undesired odor and flavor components that have to be removed. These are minor natural components or components that have formed
Preheating Drying, Degassing
150'C, Zmin, 2hPa
I Heat bleaching Deodorization
25C-260°C, 30-45 min, 3-8 hPa
Fats and Oils Handbook
668
100
Palm oil Bleaching temperature ["C]:
9)
5 g 0
20
0
0
10
20 30 40 Bleaching time [min]
50
60
Fig. 7.43. Bleaching result depending on time and temperature (after Loncin 1970).
during storage or transport. The main representatives of this group of components are the hydrocarbons, aldehydes, ketones, lactones, and FFA. Grosch (1987) gave an overview of the enzymic formation of such flavor and odor substances from lipids. Eichner (1986) reported on carbonyl components with low threshold values for odor or taste. The substances he researched were oxidation products of lipids. Dependent on the concentration these substances have an unpleasant repellent character. The number of substances that are responsible for offtaste is very large. Smouse and Chang (1967) and Chang er al. (1967) found the following volatile components in aged soybean oil. All of these components added to an off-taste: 22 acid components, 18 aldehydes, 8 ketones, 8 alcohols, 2 esters, 6 hydrocarbons, 3 lactones and 4 other components. It must be noted that hydrocarbons add to taste or smell only if they are unsaturated. Their proportion is very, very small. Marcelet (1936) identified 70 ppm in olive oil and 20 ppm in peanut oil of unsaturated hydrocarbons that were responsible for a repulsive taste. Aldehydes and ketones have a much larger influence than unsaturated hydrocarbons. The threshold values (i.e., the concentration at which 50% of the members of a test group recognize the substance) of these substances in oils is far below 1 ppm (Table 7.15). As Table 7.16 indicates, there is quite a difference between taste and smell and the matrix from which these components arise. The taste impression of fatty acids is rancid for C4-C6, above which it is dull and soapy. Aldehydes and ketones are formed mainly autocatalytically from hydroperoxides (see Chapter 2.4.3) that decompose to form them after a certain time. Table 7.17 shows some aldehydes and ketones with the off-taste they create. As a total, odoriferous components generally add up to not more than 200 ppm. This is very low compared
Oil Purification
669
TABLE 7.1 5 Threshold Values for Some Substances Causing Off-Taste in Oils and Fatsa Dissolved in Caprylic acid Capric acid Octanal Dodecanal Hexenal
Nonenal Nonadienal
Threshold value (DDrn)
Vegetable oil
(trans-Z-) (cis-2-) (trans-3-) (cis-34 (trans-6-) (cis-64 (cis-24 (cis-2-)
350 200 0.9 0.9 2.5 0.15 1.2 0.1 1 0.0003 0.002 0.01 8 0.002 0.65 0.1 9
Paraffinic oil
Water
2-Heptanone 2-Nonanone aSource: DGF (1981).
with the fatty acids that may be present on the order of some percentages. However, the effect on the quality is just the reverse. 7.4.1 The History of Deodorization
The first application of deodorization for edible fats and oils occurred in the middle of the 19th century by Cassgrand (1854) long after it had been known for the purification of organic compounds. In 1855, the first patent was granted to Bardies. Detailed overviews of the beginnings of deodorization are given by Markley (1961), Lude (1962), Swern (1964) and the DGF (1981). In the year 1893, the first German patent to be actually applied in practice was granted. It describes the deodorization of coconut oil. A product produced according to a similar process since 1921 remains on TABLE 7.1 6 Threshold Values for the Taste of Short-Chain Fatty Acid9 ~~
Coconut oil
Dairy cream Fatty acid (pprn) in Butyric Caproic Caprylic Capric Lauric Myristic
C40 C, C, C, C,, C14
Smell
Taste
Sweet cream butter
Smell
Taste
50 85 2 00 >400 >400 >400
60 105 120 90 130 >400
40 15 455 250 200 5000
35 25 >loo0 >loo0 >loo0 21 000
160 50 25 15 35 75
aSources: Crosch (1987), Halsbeck eta/. (1986), and Pfannhauser (1994).
670
Fats and Oils Handbook
TABLE 7.1 7 Components Causing Off-Taste in Fats and Oils ~
~
Component
Taste imDression
Reference
cis-3-Heptenal 2,4-Heptadienal 2,4-Decadienal Vinylamylketon trans-2-, cis-6-Nonadienol Diacetyl 2 -Penty lfuran cis-4-Heptenal trans-2-, cis-4-, cis-7-Decatrienal
Green taste of soybean oil Green taste of soybean oil Aged soybean oil Aged soybean oil Aged soybean oil Buttery Green beans Fishy Fishy
Hoffmann (1961) Hoffmann (1961) Hoffmann (1961) Hill and Harnrnond (1965) Hill and Harnmond (1965) Seals and Hamrnond 11 966) Chang eta/. (1 967) Meijboom and Stroink (1972) Badings (1973)
the market. Later, better materials such as special steels were found, the deodorization temperature was increased and better plants allowed the working pressure to be reduced, thus making the plants much more effective. 7.4.2 The Theory of Deodorization
The vapor pressure of a liquid is described by Clausius-Clapeyron's equation: [7.23] where p is the vapor pressure, R is the universal gas constant, AHvapis the molar heat of evaporation, and T is the temperature (K). If In P is plotted against 1/T, a straight line is obtained from which the boiling point of the components (at the boiling point: vapor pressure = ambient pressure) can be read. The vapor pressure of some relevant components is given in Chapter 2.3.2.4. Dalton's Law describes the vapor pressure of a mixture as the sum of the partial pressures of its single components: [7.24] The partial pressure of the single components can be calculated following Raoult's law: [7.25] where pci is the vapor pressure of the pure component i and x is the mole fraction of the component in the total system { 1, 2, 3, ... ,n}. Because the total vapor pressure of a system (ideal, nonaceotropic system) according to Equation [7.24] must always be greater than the partial pressure of each single component, the boiling point of every system must be lower than the boiling point of its highest boiling component. Despite that, of course, a higher vapor pressure caused by a higher
Oil Purification
671
temperature speeds up the distillation. From Equation [7.26], it can be concluded that the ratio of the vapor pressure of the single components of the system equals the mole fraction of these components in the system: nA
nB
-PA
[7.26]
P B
where n is the number of moles. From the above equations, one can calculate to what extent the vapor pressure of different fatty acids and minor components contributes to the vapor pressure of the total system (Figs. 7.44 and 7.45). It becomes clear that in an oil with FFA and some aldehydes and ketones, the FFA contribute predominantly to the vapor pressure of the system, even more so than do the aldehydes and ketones. Assuming 0.1% ( d m ) of FFA, equivalent to 0.3% (wdwt), the vapor pressure of the triglycerides accounts only for > vapors Cooling externally, usually plate heat exchanger
Oil / Fat
(deodorized)
S.C.D
Fat (molten,60-90'C)
Oil (-4o'c) Tray
I
+
t I
170-180'C, 2550 hPa Drying
220.240% 5-10 hPa
3
(max 270%)
4
Deodorizing
225240'C, 5-10 hPa (max 270'C)
I
Deodorizing
220.240'C, 5-10 hPa (max 270'C) optional (with 2. tray)
Cooling 5'6
80-110'c
Oil / Fat (deodorized)
Fig. 7.48. Processing steps in batch deodorization (B.D.) and in semicontinuous deodorization (S.C.D.). Deodorizers of the Lurgi design consist of a stack of six trays. In principle, semicontinuous deodorization is an automatically controlled batchwise process. If the retention time of tray 1 is elapsed, tray 1 will be drained off into tray 2 by the automatic opening of its bottom valve. A special interlocking system is provided to assure that tray 2 is completely empty before tray 1 is drained. The circulation of
Oil Purification
679
the oil in the individual trays is done by a special mammoth pump. This is to ensure high efficiency of the injected steam and good heat transfer. The oil is heated to -170°C in tray 1 after the oil has passed through a heat transfer tray. After heating up to 240°C (normally), the deodorization is started. Deodorization requires -2.5 cycle times. Consequently, at least 3 trays are needed for deodorization. In the last tray, the oil is cooled to 110-120°C. The trays themselves usually hold 2-7 ton. Figure 7.49 shows a semicontinuous deodorizer with six trays and the peripheral equipment for vacuum generation. Because the equipment design makes it inevitable that the residence times in all trays are equal (if volume is the same), the cycle time of one cycle must be as long as the residence time required for the longest process step for which only one tray is foreseen (cycle times in Fig. 7.50). All other residence times are equal or a multiple of that one. Part of the total time of all cycles is used up for draining the trays in the order tray 6 + tray 5 + ... tray 1. For each tray, this requires -6 min, i.e., one seventh of the total cycle time of - 4 0 4 4 min. Such plants require a square space of -25 m* for the deodorizer itself and auxiliary equipment. The deodorizer is 20 m high, the auxiliary equipment 10 m. A Lurgi plant is shown in Figure 7.51. Other manufactures offer different designs; however, the principle of a continuous process imitated by batch processes split into the processing steps remains untouched. In the de Smet deodorizer, the trays are not vertically stacked, but horizontally arranged in groups of six (Fig. 7.52) with one central cell and five to surround it symmetrically. All six compartments are equally equipped with sparge steam distributors at the bottom, steam jets and lifting pumps for good agitation. They also have the same capacity. The plant capacity is up to 200 t o d d or double
Fully refined oil Cooling
c3
Fig. 7.49. Semicontinuous deodorizer (redrawn; courtesy of Lurgi CmbH, Frankfurt).
Fats Oil andPurification Oils Handbook
680
673
pressure pD, which also takes into account the height of the oil layer (especially important for batch plants): Deodor. Deodor. Wing Heating
z:.
400.
E
L
5-10 hPa
3550_ hPa
[7.28]
c
em
where p D = p + p . g . h/2, h is the height of the oil layer, g is the acceleration due to gravity, p is the density of the oil, and p is the vapor pressure of the system. At 250°C the thickness of the oil layer adds -0.6 mm Hg (1.33 hPa) to the total pressure for each centimeter of oil height. According to Bailey (1941), an evaporation factor E can be introduced, which may also be called the efficiency of deodorization (it reflects the ratio between I ' 1 , actual and 0theoretical partial pressure); the equation ,is integrated after introducing 0 20 40 80 80 loo 120 140 180 180 200 220 some assumptions and Equation [7.29] is obtained: P m(DGF t n g 1988) time [mln] I
i
'
,
I
I
I
Fig. 7.50. Cycle times and temperature profile of a five-tray semicontinuous deodorizer.
[7.29] that if two units are stacked. Stacks of three are also available. Figure 7.53 shows a plant one x2 group vessels. Theend oil enters the plant via A after being wherewith x1 and are of thethree initial and the mole fractions, respectively, ofprethe heated the hot oil, which leaves the the plant (D). In the upper part ofat A, it is volatileby component i. The equation gives amount of steam necessary a given degassed and furtherx1heated up. Itthecollects theconcentration bottom and is heated to initial concentration to achieve desiredatend x2.further This equation is deodorization From timesatisfactorily to time, the bottom of A opens andthat the theoretical andtemperature. can describe reality only if valve an empirical factor oil is be drained off into one of the deodorizing chambers B. After is must determined experimentally is added. Szabo Sarkadi (1958)deodorization added this faccompleted, theischamber the oil flows into tanksystem) C, which is kept at the tor Ai (which related toopens massand transfer properties of the to the numerator same vacuum as B.This After the oil through the cooler D countercurrently to of Equation [7.29]. factor canpasses be found in standard tables. the feed oil, it is finally cooled in E. Vapors are condensed in F. of plant is suitable for small batches of different oils or blends, for 7.4.3This Thekind Process Conditions example, those attached to margarine halls. Table 7.19 gives the average consumpApart theand factauxiliary that certain processing conditions aredeodorization. necessary to conduct the tion offrom energy material for semicontinuous process, they also have a large influence on the quality of the products and the processing The influence of the different variables processing shownoil in As stated above,on good contact is between 7.4.4.3 costs. Continuous Deodorization. Table 7.18. Most important are, of course, the vapor pressures of the individual and steam is essential for effective deodorization. In this context, continuous processing offers more opportunities than noncontinuous working. At fxst, much shallower TABLEoil 7.1 8layers can be acliieved than in noncontinuous plants because a steady managed.influencing Thus better contact can be ensured between the oil and the flow can beParameters Processing Deodorization sparge steam. The steam may be injected by means of different systems (Fig. packed columns and falling films. In 7.53). The finest films can be achieved withInfluencing addition the thin filmPressure produced, the liquid that passes packed column Influencingtofactor: Temperature neededthrough aSteam consumption is forced by the packing material to continuously change its direction of flow. + Pressure + It is thus well agitated. of falling film vessels, turbulence is created, + which Deodorization time On the walls [-I [+I + droplets, thereby ensuring + good FiIIing height effects. A rotating [-I disk creates fine oil has similar Temperature Intermixing inI-] the three other cases is achieved by allowing the - oil to intermixing. (+ = directly proportional, - = indirectly proportional, = indirect flow around or through obstacles such as[Iholes of influence. sieve plates or over steps. It can
Oil Purification
Fig. 7.51. Model photo of a semicontinuous deodorization plant.
681
682
Fats and Oils Handbook
TopvlrwdthOCblb 01 tha r h o dwdorku
Fig. 7.52. Semicontinuous deodorizer with heat recovery by direct oil-oil heat exchange (redrawn; courtesy of De Srnet S.A., Edegern). also be achieved by the injection of steam. If the oil film is sufficiently thin, enough degassing takes place automatically soon after the oil is exposed to reduced pressure. Figure 7.54 shows some details of a horizontal continuous deodorizers. The deodorization step itself follows the same principles as semicontinuous deodorization. The difference is that the discrete processing steps in the latter are spread over the entire length of the continuous deodorizer without sharp boundaries. Plant designs exist with integrated heat transfer and with external heat exchangers. Figure 7.55 shows a Lurgi plant with integrated heat transfer. The neutralized bleached oil is warmed up (W) and dried (A) before it passes a second heat exchanger for preheating. Then it passes the bottom part of the deodorizer (H) to be heated up by the deodorized hot oil, which leaves the apparatus. Indirect highpressure steam serves to reach the processing temperature. The oil then passes the different deodorizer (D) stages, which each contain direct steam-operated oil circulators. It is then cooled in the heat exchange section and some citric acid may be added if required (detail omitted in the drawing). A fine filter (F) is passed through to polish the oil before it leaves the plant to be stored. A vacuum is generated in V.
Oil Purification
B Trays
633
+ Bubble heads
Sieve plate
4
Steam, Vapors
1 Oil, Fat
Packed bed
Falling film
Rotating disk
Fig. 7.53. Technical means to ensure contact between oil and steam (redrawn from Sjoberg 1987).
Figure 7.56 shows a photo of a model of a complete plant; Table 7.20 gives average consumption. The de Smet deodorizer shown in Figure 7.57 is of similar design. The oil is fed via (1) and then consequently passes ring-shaped trays. The trays can overflow TABLE 7.19 Consumptions for Semicontinuous Deodorizationa Consumption per MT of oil De Smet, heat recovery Lurgi Steam for deodorization for vacuum generation Heating energy (gas, oil, ...) Cooling water for product cooling for vacuum generation Electrical energy Steering air (7 bar)
(kg)
20-30
(kg)
70-1 00
(MI)
2 94 5-6 12-1 6 2
(m3) (m3)
(kWh) (Nm3)
1
At 25-50 hPa working pressure in stage 1, 4-7 hPa in stages 2-6; deodorization temperature 240-270'C. Tourtesy of Lurgi CmbH, Frankfurt and De Srnet, Edegern
Without
125 500 25 13
With
17 125 150 2.8 13
684
Fats and Oils Handbook
Fig. 7.54. Details of deodorizers (UL integrated heat exchanger; UR heating tray; LL view of a heat exchanger tray through the inspection glass during fabrication; LR heating chamber; courtesy of Krupp Maschinentechnik CmbH, Hamburg).
via (3) and can be emptied via (7). Steam is injected (8) and vapors are sucked off (6,9). Shell oil can drain off via valve (1 1). Most of the deodorizers in the world follow a vertical design. The continuous KirchfeldiTirtiaux deodorizer, however, consists of horizontal cylinders. The manufacturers claim some advantages, including the great oil surface that allows for a large contact area with the steam and also the shallow layer. Figure 7.58 shows one cylinder of such a plant and a sectional drawing. The cylinders have a diameter of 95 cm and a length of 4.5-12.5 m. Double-pass cylinders may also be ordered, which are horizontally separated into two trays internally. In the lower part of the cylinder, a bank of heating tubes compensates for the heat losses during evaporation. Heating steam is fed via C and the condensate leaves via B. Sparge steam, which enters via D, is finely distributed by means of a sieve plate. The sparging gas also constantly renews the top oil layer. The product is fed via (A) and drains off via (E). The vapors are collected until they are removed through the vapor drain pipe. Splash oil is reduced by the design of that pipe. The large space over the oil ensures a large oil surface exposed to the vacuum. The complete set of pipes
Oil Purification
685
Neutralized
v Stea
I 4 Shell oil
Fig. 7.55. Continuous deodorization plant (redrawn; courtesy of Lurgi CrnbH, Frankfurt).
can be drawn off the cylinder for repair and maintenance (see upper part of Fig. 7.58). Because one cylinder is insufficient, several can be arranged sequentially to form a deodorizer. Such plants, which differ significantly in design, can be set up for capacities between 100 and 600 MT/d; two to eight cylinders are required, depending on the capacity (Fig. 7.59). 7.4.4.4 Thin-film Deodorization. The first plant for large scale (250 MT/d) thinfilm seed oil deodorization was started in 1996. It had been known for some time that thin-film stripping theoretically had considerable advantages over tray stripping. Because the steam passes over a very thin oil film in a true countercurrent operation, the contact between steam and oil is optimized. The surface area of the oil can be further increased by a special design of the packed column. Other means to enhance mass transfer are eddy currents caused by wall heat transfer in the falling film design or centrifugal action if rotating disks are used. However, there are still some major drawbacks, one of which is the increased trans fatty acid ("FA) level (Stenberg 1996). A second is the removal of antioxidants and vitamins as a result of the very high effectiveness of the operation. A third is that the heat bleaching effect may be reduced
686
Fats and Oils Handbook
Fig. 7.56. Continuous deodorization plant (courtesy of Lurgi, Frankfurt) Oelde).
68 7
Oil Purification
TABLE 7.20 Consumptions for Continuous Deodorizationa Consumption per MT of oil
(kg) (kg)
Steam for deodorization (4 bar) for vacuum generation (1 1 bar) Heating energy (gas, oil, ...)* Cooling water for product cooling for vacuum generation
ikJ) (m3) (m3)
(kWh) (Nm3)
Electrical energy Steering air (7 bar)
15-20 50-70 105,000 2-3 10-12 3-5 3
Working pressure 4-7 hPa, deodorization at 240-270OC; *indicates heat recovery aCourtesy of Lurgi CmbH, Frankfurt.
---3
7
t .
11
9
Fig. 7.57. Continuous deodorizer (courtesy of de Smet).
688
Fats and Oils Handbook
Fig. 7.58. Horizontal deodorizer model photo with drawn cylinder, sectional drawing, principle design (redrawn; courtesy of Kirchfeld GmbH & Co KG, Dusseldorf and Tirtiaux s.a., Fleurus).
Oil Purification
689
Fig. 7.59. Horizontal deodorization plant type DEOTEST, model photo (courtesy of Kirchfeld CmbH & Co KG, DusseldorD.
690
Fats and Oils Handbook
because of lower temperatures. In the case of TFA increase, pilot plant data showed promising results. Hence, there is more to do to enhance the advantages of the process. 7.4.5 Auxiliary Equipment 7.4.5.7 Heating the Oil. Deodorization is conducted at relatively high temperatures. Therefore it is important to have equipment available that allows these temperatures to be reached reliably and quickly to keep cycle times short. Because the oil has to be cooled after deodorization, heat integration is essential to save energy and keep the cost low. 7.4.5. 7 . 1 Steam heating. The most common way of heating is by indirect steam. To heat deodorizers, which are run at -260"C, steam of 50-70 bar is necessary (264-285°C). The boilers normally used in factories are not able to supply such pressures. Therefore, high-pressure boilers that supply the pressure needed are required. The advantage of steam is its high heat transfer, caused by the large heat of e v a p ration that is set free again during condensation. The heat transfer is about four times that of thermal oils. The second advantage, which is important in the case of a leakage, lies in the fact that the product can come into contact only with water, not with a chemical. To avoid any unacceptable risk for the consumer, responsible food producers therefore have changed their plants to run with steam only. Saturated steam of -50 bar is required to heat the oil to 240-260°C. Usually there is a preheating step to 180-200°C using steam of 15-18 bar so that high-pressure steam boilers can be kept small. 7.4.5.7.2. Heating with thermal oil. Using thermal heating oil, the investment in the plant is considerably lower. Zehnder (1976) estimated the savings at around 50%. For this reason, thermal heating oil systems have been installed in many plants. Usually, the systems use eutectic mixtures of 27% biphenyl and 73% biphenyl oxide (e.g., Dowthenn). The liquid (vapor pressure at 30O0C, -2.5 bar) can be evaporated at considerably lower pressure than water. It is nonflammable at 115°C and is flammable from 138°C upward. In 1973, Japan reported illness and fatalities that were said to be caused by the consumption of rapeseed oil containing thermal heating oil (Brekke 1980). In addition, the thermal heating oil was said to be heavily polluted. As a consequence, such plants were forbidden in Japan. Imai (1974), however, assumed that thermal heating oil that contaminates the product through a leak is removed in the deodorization process together with the undesired components. This can be the case only if the leakage does not affect the processing conditions so that they are maintained long enough after the contamination to ensure complete purification of the oil. Generally, avoiding thermal heating oil is the only completely safe way to circumvent this risk. 7.5.4.7.3 Direct heating. Kuroda and Young (1989) showed that it is also possible to heat the oil directly. Up to now, this approach has always been excluded
Oil Purification
691
because it was assumed that local superheating could not be avoided. In Asia in 1989, at least ten deodorization plants with direct heating were in operation. These plants obviously produced oil that was in no way inferior to conventionally heated oil. A special construction of the burner avoids superheating. 7.4.5.2 Vacuum Generation. Jet ejectors produce a vacuum by passing a motive fluid successively through a motive nozzle and a diffuser (Fig. 7.60). The static
Fig. 7.60. Drawing of a steam jet vacuum ejector and plant with ejector vacuum system (courtesy of Korting Hannover AG).
692
Fats and Oils Handbook
pressure energy of the motive flow is converted into kinetic energy. This energy is transferred to the suction flow, thus sucking off gaseous particles. In the case of deodorizers, the motive flow usually consists of steam, i.e., the equipment used is a steam jet ejector. The main functional parts of the pump are the motive nozzle and the diffuser. Because of the design of the motive nozzle (2), i.e., decreasing throat diameter, the pressure decreases as the velocity of the fluid rises. Conversely, in the diffuser section (4,5), the flow is retarded and the pressure increases to reach discharge (atmospheric) pressure again. Pressure and speed are displayed in Figure 7.61. The lowest static pressure p s exists between motive nozzle and diffuser. The supersonic motive flow (steam) possesses high kinetic energy, which can be released to the suction flow by impulse transfer where the two flows mingle. Reaching the diffuser throat, the velocity is reduced to sonic. In its diverging section, it further decreases to reach discharge pressure pd at the end. The ratio of these two pressures p i p , is called the compression ratio of a jet ejector. Assuming the atmospheric pressure at sea level to be lo00 hPa and the working pressure at present to be between 2.5 and 4.0 hPa (sometimes lower), a compression ratio between 400 and 250 is required. This cannot be achieved by a single ejector and thus requires multistage equipment. E
I
P
Fig. 7.61. Velocities and pressures in a steam jet vacuum ejector (courtesy of Korting Hannover AC).
Oil Purification
693
7.4.5.3 Vapor Condensation. All vapors from the plant, whether steam, fatty acids or other lipids, have to be condensed. Usually condensation is done with injected water so that the condensate consists of 8690% water. To save water, the vapors can be precooled by passing through cooled surfaces. Figure 7.62 shows a multistage steam jet vacuum pump. The possibility of condensing the water vapors and the motive steam depends on the coolant (water in most cases) temperature. This means a dependency on ambient temperature and the influence of the seasons. Condensation efficiency may be improved in the following sequence of technical development: surface condensation; direct contact condensation, alkaline loop at normal cooling water temperature 20°C; direct contact condensation, alkaline loop at low cooling water temperature 4 4 ° C ; direct contact condensation, brine loop 20°C; ice condensation 2% product weight, this value is reached in the aqueous phase because the water content of margarine is 20% at maximum. In addition to these properties, salt levels ~ 0 . 2 % work as an antispattering agent during shallow frying. 8.2.5.6 Flavors. The flavor cocktails used in margarine making work as flavors and flavor enhancers; they also mask off-flavors. The flavors used depend on the geographical region, i.e., the local taste, and can range from bland to over-buttery to cheesy. If butter is normally slightly rancid in the respective country, this can also be reflected by the flavor notes used. Dosage also is done according to local preference.
8.2.5.7 Preservatives. Preservatives are rarely used in 80% fat margarines in countries with moderate climate, households equipped with refrigerators and welldeveloped logistic chains. In reduced-fat margarines, they are not necessary for production, but are needed to protect the product during open shelf life. Detailed information on preservatives is given by Luck (1984). 8.2.5.7.1 Benzoic acid. Benzoic acid works as a preservative in the undissociated form only. A minimum acidity is necessary therefore to block dissociation and to dissolve the acid from its salts when benzoates are used. The dissociation constant is 6.46 x 10-5 (Bosund 1960 and 1962), which means that the pH that is reached applying lactic acid is already sufficient. However, the pH of common margarine is only a little lower than that needed to ensure the efficacy of benzoic acid. Above that, the distribution coefficient between the fat and water phase is very unfavorable. The working mechanism of benzoic acid is based on the inhibition of enzymes belonging to the acetic acid cycle. It also inhibits oxidative phosphorylation as well as the citric acid cycle and has a negative influence on cell walls. In butter, benzoic acid (cO.OOl%, Rudischer 1959) can be found naturally, stemming from the fodder. This amount, however, is not sufficient for preservation. Benzoic acid is not allowed in all countries. When it is used, the amount is usually 10.1%.
8.2.5.7.2Sorbic acid. Sorbic acid usually is allowed up to 0.12%, but the amount used is normally lower. Sorbic acid is -50% more effective than benzoic acid and is used mainly in reduced fat margarines. The distribution coefficient between the oil and water phase is favorable. In the human body, it decomposes into water and carbon dioxide.
Fat as F o o d
747
The working mechanism of sorbic acid is based on the inhibition of enzymes of the carbohydrate cycles and the citric acid cycles (Azukas 1962, Rehm and Wallhofer 1964 and 1967, York and Vaughn 1964). Sorbic acid forms covalent bonds with SH-groups of the enzymes (Marsoadiprawito and Whitaker 1963), thus inactivating them. In addition, it has negative influence on the cell walls (Eklund 1980 and 1985). Furthermore, undissociated acid infiltrates the cell, and 40% of the enzymes are undissociated at a pH of 3.15. Because the dissociation constant is 1.73 x 10-5, sorbic acid also operates in ranges of low acidity. Detailed information on its mechanism to preserve margarine is given by Becker and Roeder (1957). 8.2.5.8 Thickening Agents, Stabilizers. In half-fat margarines or margarines with an even lower amount of fat, the emulsion on its own is no longer able to stabilize the product. Therefore, thickening agents or stabilizers, (for example, gelatin), are used to stabilize the water phase. A good overview of possible thickening agents is given by Trudso (1988). An overview of thickeners or stabilizers that are allowed for half-fat margarines in some countries is given in Chapter 8.2.10.
8.2.5.9 Colorants. In Europe, margarines usually are not colored with artificial colors. Often, however, P-carotene is added; it dissolves in the oil to give a reddish color. In many countries, Bixin and Annatto are allowed in margarines. The addition of p-carotene is also allowed for butter and is used to color winter butter. If unbleached palm oil is used in the blends, a natural yellow-reddish color can be obtained by the carotenes without later addition. 8.2.5.70 Vitamins. Vitamins are added to almost all household margarines. In some countries, the addition at the level that is usually present in butter is mandatory. Usually, fat-soluble vitamins A, D and E are used. In geographical regions with deficiencies in water-soluble vitamins, vitamination with vitamins B and C is also common.
8.2.5.77 Water. Water usually stems from wells on the margarine factory sites or is taken from the municipal net. Water quality must be monitored constantly. 8.2.5.12 Air, Nitrogen. Some margarines and fats are whipped with nitrogen or air to soften the products (so-called soft margarines). In most countries, the gas used for whipping has to be declared on the ingredients list.
8.2.5.13Starch. Until some years ago, in some European countries, starch had to be added to margarine to enable adulteration of butter with margarine to be detected. This legal requirement stems from the last century, because current modem analytical techniques allow easy detection of adulteration without the addition of marker substances such as starch or sesamol. Starch has no function at all in margarine. On the contrary, it negatively influences its spattering behavior.
Fats and Oils Handbook
748
8.2.6 Physical Properties
Margarine is plastic, which means that its internal repulsive forces are not sufficient to bring it back to its original shape when it is deformed under outer pressure. Kroll (1978) gives models for explaining the viscosity of plastic fats compared with a Newton liquid (Fig. 8.18). Because the inner structure of margarine can be softened under mechanical influence (e.g., kneading), the structural bonds are then destroyed so that only the molecular bonds between the crystals stay in place. Hardness and flow limit of margarine are closely connected. Hardness is expressed by the following equation:
where w is the weight of the cone of a penetrometer including the falling stick suspension, p is the depth of penetration (l/lO mm), and n is a constant (1.6 for butter, margarine and bakery fats.) The factor k depends on the angle of the cone. This dependency can be derived with a very high correlation coefficient in exponential regression from the given values. If it is double logarithmically plotted against k, a straight line results.
k = u . ab'
(U
= 6.2 . 106 and b' = -1.906)
[8.21
or
k = d a b (U = 6.2 . lo6 and b = 1.906)
[8.2a]
10 8 6 4
2
0 0
1
2
3
4
5
6
7
8
91011121314
Shear stress [N/m2] Fig. 8.18. Viscosity of a Newton fluid and two plastic fats (after Kroll 1978).
Fat as Food
749
The complete formula to calculate firmness then reads as follows:
where a = 6.2 . 106, b = 1.906 and n = 1.6 for butter, margarine and bakery fats. Given a penetrometer, only one variable @) is then left in the equation, i.e., the depth of penetration itself; all other variables can be reduced to a constant that is specific for this penetrometer. C-values are measured because of their importance for the functionality of some plastic fats (Table 8.13). To reach the above requirements, i.e., specific C-values that ensure the desired functionality of the margarine, special processing conditions have to be maintained that allow for controlled crystallization. If margarine is warmed up too much after p r e duction so that the fats crystals are partly or completely melted, uncontrolled crystallization can occur during cooling. The C-value of such margarine can deviate enormously from the one that has been manufactured by the producer under these special processing conditions. Another reason for totally different C-values after recrystallization is that the amount of mixed crystals can change, and pure triglycerides may crystallize. Usually this leads to a deterioration in melting behavior. The crystal modification can also change. During production, the aim is to obtain p' crystals (-1 pm long and 0.25 pm thick). At a certain temperature, there is a transition to p crystals, which have a much higher melting point (see Chapter 2.3.1) and are larger (20-30 pm). The coarse crystals give a sandy impression on the tongue. To retard p-crystal formation, crystal inhibitors can be used (Krog 1977); an example is sorbitan tristearate, which is legal in Canada. However, careful blend composition, processing and good logistics make their use unnecessary. The physical properties of margarine are also determined by the amount of solid triglycerides that margarine has at a certain temperature. These values can deviate from the typical values that are shown in Figure 8.19. Heart health margarines with a guaranteed high amount of polyunsaturated fatty acids (e.g., in Europe, Becel) are much softer that other margarines due to their fatty acid composition. Hardness of margarine is determined by many factors. If it crystallizes in the pack, i.e., after production, this results in a much harder product TABLE 8.13 Importance of C-Values for the Plasticity of Fat Emulsions C-value (dcm2) 1500
Properties Not spreadable because too soft ; flows away Very soft, just spreadable Very spreadable, good plasticity; also spreadable at refrigerator temperatures Hard, but still spreadable Very hard, barely spreadable; not spreadable at refrigerator temperatures Too hard, no longer spreadable
Fats and Oils Handbook
750
Solids content
80
[%I
1
0
5
10
15
20
25
Temperature
30
35
40
45
[“C]
Fig. 8.19. Solids content of typical margarines.
because the margarine can no longer be overworked mechanically. Compositions that are slowly crystallizing or those that are not sufficiently cooled have a special tendency toward postcrystallization. This is shown in Figure 8.20 for a fat that is completely crystallized in the scraped surface heat exchanger and a fat that has a high rate of postcrystallization. The share of fat (m.p. 58°C) in the fat/oil mixture Stress [
/kPa]
(a) fully crystallized in cooler 25
(b) with high postcrystallizationi
--f
5
01 0
I
I
I
1
2
I
,
I
3
4
5
5
6
Strain [ 10. ?In(ho/h)]
Fig. 8.20. Stress-strain curve of a fully crystallized fat and a fat with high postcrystallization (courtesy of Unilever, see also Heertje e t a / . 1988).
Fat as Food
751
of the above trial was 14%. When leaving the processing unit, it was 100% crystallized in case (a) and only 40% in case (b), which means that 60% of the total fat blend postcrystallized. This leads to an inhomogeneous structure with long (p-) crystals. In such a structure, fewer bonds can break under stress than in the case of the fully crystallized blends. Of course the fat composition itself also has a large influence on hardness. The lower the oil content, the harder the product. As noted before, primary bonds between the crystals stay intact even during the mechanical overworking; secondary bonds are broken up. Strong primary bonds therefore contribute greatly to hardness. Figure 8.21 shows typical C-values for German margarines of the composition types shown in Figure 8.16. The higher the shear stress during overworking, the more the continuous structure turns into a grainy structure. This has been proven by Heertje et al. (1988), who recorded scanning electron microscope pictures of the crystal structure. Air or gas, whipped in intentionally or unintentionally, also influences hardness. Proportional to the amount of gas, the number of crystals per volume decreases; margarine becomes softer. By comparison, for butter, churning leads to an air content of -5% (Mulder and Walstra 1974); soft margarines contain 10-20% air or nitrogen. 8.2.7 Margarine Production
Production of margarine consists of some principal steps starting from refined oils and fats. These principal steps are shown in Figure 8.22. Apart from the fat blend, the properties of the finished product are really influenced only by steps C and D.
Fig. 8.21. C-values of typical German margarines (after Schleenvoigt 1989).
Fats and Oils Handbook
752
OivFet blend
I
A B
Mixing
Ingrsdii I Emulsifying
I C @
Cooling
stomge
PortayataHkstknand hardening @ StapaCandDinmany
Margarine
different conRguntknr
Fig. 8.22. Principal steps of margarine production.
The composition of the ingredients has only a minor influence. Posthardening after packaging cannot be influenced and is based on the properties that are inherent from steps C and D.
8.2.7.7 Ingredient Preparation. There are two extreme variations in the way that ingredients are prepared and added. On the one hand, it is possible to prepare single aqueous or oily solutions of single ingredients and then blend them with the fat blend, the water and the milk by means of a multihead proportioning pump (proportioning process); the mixture is then emulsified by static mixers built into the tubes. On the other hand, the complete margarine composition including fat and water phase and all ingredients can be mixed in one batch from which the emulsion is then directly processed through a scraped surface heat exchanger (premix process). In between those two extremes, all intermediates of ingredient and emulsion preparation are possible, and there are examples in the industry for all of these intermediate steps. Using the premix process, the water content of the emulsion is checked only once per batch (continuously good agitation assumed). If proportioning pumps are used, the water content has to be checked at intervals on a statistical basis over the period of production. The aqueous phase should be pasteurized. This can be done with the whole emulsion as well as with single solutions. Soured milk as such cannot be pasteurized because the protein would coagulate and become sandy. This would be possible only after the addition of protection colloid. However, pasteurization is not necessary if the milk has been soured under hygienic conditions and the source sweet milk has been pasteurized. This means that the soured milk has the same bacteriological status as a postpasteurized soured milk except for the souring cultures. These stay alive, but they are no longer active in margarine because the temperature is too low and the size of the water droplets does not allow for it.
Fat as Food
753
8.2.7.1.7Milk souring. As a source for the production of soured milk, pasteurized fresh milk or milk that is reconstituted from milk powder and water is used. The milk is pasteurized and soured with cultures such as Streptococcus lactis or S. cremoris to an acidity of about 40"SH. Souring can be carried out with parts of a mother culture that is inoculated or by using deep-frozen or freeze-dried cultures purchased from a third party. Cultures are reproduced according to the scheme that is shown for butter (see Figure 8.3). Souring itself also follows this pattern. Ripening of the milk takes about 12 h. The soured milk is stored in hygienic, cooled batches and transported via a piping system. Cleaning of this system requires special care. When proportioning pumps are used, milk is dosed via a dosing cylinder. In addition to bacteriological souring, there is also the possibility of chemical souring. For that purpose, citric acid or lactic acid is added to the sweet milk. The influence on the milk protein is similar to bacteriological souring; however, none of the flavors that develop in bacteriologically soured milk will be present. If milk, buttermilk or whey powder is used, it is dissolved in water and stored in a batch. For cleaning, there are the same requirements as for milk. 8.2.7.7.2 Water-soluble ingredients. The water-soluble ingredients are prepared singly or as a mixture. Usually, they are rinsed into the storage batch by circulation over a Venturi-tube or collected from mother solutions. Such solutions can also be purchased as such (e.g., brine). 8.2.7.7.3 Oil-soluble ingredients. Oil-soluble ingredients are dissolved either in a single oil or in the relevant margarine fat blend. In both cases, the amount of fat from the prepared ingredients has to be taken into account when calculating the total composition. If one-oil margarines (e.g., sunflower margarine) are produced, the oil for dissolving the ingredients must be the same oil as for the margarine, unless legislation allows for a certain amount of foreign triglycerides. To avoid crystallization of the emulsifier (monoglycerides and monodiglycerides), the ready-prepared ingredients mixture has to be stored warm (T > 55°C). Depending on the ingredients, it may also be necessary to agitate to avoid sedimentation. 8.2.7.7.4 Oil/fat blend. There are two principal possibilities for preparation of the fat blend. One is to blend fully refined oils and fats (see Chapter 7.4); the other is to compose the blend from crude oils and to refine the complete blend. Of course, it is also possible to combine both alternatives, which means blending fully refined oils and fats with fully refined fat partial blends. The assembly of the oil in a batch is done gravimetrically over a composition weigh scale or volumetrically over a rotary piston meter or a mass flow meter. With today's methods of electronic process control, it is possible to compensate for the volume differences of oils and fats of different temperature so that flow meters can provide the same accuracy as weigh scales if the amounts are not too small.
Fats and Oils Handbook
754
TABLE 8.14 Hydrophilic Lipophilic Balance (HLB) Necessary to Emulsify Certain Oils and Fat9 Water emulsified with
Oil in water
Water in oil
6
3 5 3 4
Cocoa butter Corn oil Cottonseed oil Palm oil Rapeseed oil Soybean oil Beef tallow
a 6
7 7 6 6
4
3 3
aSource: Griffin (1979).
Bacteria cannot grow in the oiYfat composition. It is therefore not necessary to pasteurize the oil blend. Pipes should be flushed regularly with hot oil to remove the fat sediments on their walls. In fact, it is important to exclude water from these vessels and pipes to maintain quality. 8.2.7.2 €mu/si&ing. To emulsify the oil and water phases, they have to be mixed care fully. Depending on the process chosen, this is done at least partly in the premix batch or later in the proportioning pump. Usually additional static mixers built into the pipes complete the task. Emulslfying is easily accomplished because all margarine compositions contain sufficient emulsifiers that support it. At first, an emulsion of the type oil in water is built; this is converted by phase inversion into an emulsion water in oil. Table 8.14 shows the HLB-values that are necessary for different oils and emulsion types. Lynch and Griffin (1974, Table 8.15) compared the emulsifying properties of different apparatus that can be used. Apart from static mixers that promote emulsification in the pipes, only scraped surface heat exchangers (SSHE) are presently well suited for margarine making. They alone guarantee sufficient heat transfer to ensure cooling down of the emulsion and removal of the heat of crystallization. TABLE 8.1 5 Properties and Performance of Emulsification Devicesa ~
Device Anchor stirrer Paddle stirrer Scraped surface heat exchanger Disk mill Homogenizer Colloid mill
~~
Agitation speed
Carried in mechanical energy
Suitable for . . . viscosities
Heat transfer
low medium medium
low low to medium low to medium
high medium to high high
medium poor very good
high low high
low low low
low to medium low to high low to medium
medium poor
aSource:Lynch and Griffin (1974).
poor
Fat as food
755
Dosing the ingredients before emulsion preparation can be done via low- or high-pressure pumps. If low-pressure pumps are used, the SSHE/crystallizer combination itself has to be fed via a high-pressure pump. 8.2.7.3 Cooling (Crystallization) and Working of the Emulsion. To achieve the desired consistency, the margarine fadoil blend must of course be properly composed to achieve these properties. Within the given limitations of the specific fat blend, the consistency of the margarine can be heavily influenced by processing. In principle, processing is a sequence of cooling steps that start crystallization at different temperature levels, i.e., holding zones that allow for further crystallization without cooling and application of mechanical stress to break up secondary bonds to the degree desired. Heat has to be deducted as well to cool down the emulsion and to remove heat of crystallization. Fats and oils crystallize relatively slowly, some very slowly. If slowly crystallizing fats such as palm oil at 60°C are set in a bath of 0°C and are cooled down, it takes up to 2 h until they are fully crystallized. For quickly crystallizing fats, for example, palm kernel oil, the same process takes only -10 min (see Fig. 8.23). If the cooled mixture is agitated (compare lines A and B in Fig. 8.23), crystallization occurs much more quickly, in this case -10 times more quickly. Except for palm oil compositions, a fat blend that is continuously scraped off (i.e., agitated) in SSHE and crystallizers takes -7 min to fully crystallize given cooling medium temperatures below -10°C and an inlet temperature of -40°C. In the following, two different processes for margarine making that are still applied today are described. The chum-drum process that was the standard method
25
Agitated
,B-
20 -
, f C
A
--
L _ _ _ _ _ - _ - - - - - - - - - - -
Hardened fish oil/ soybean oil (50:50)
I I I
15 -
Palm kernel oillsoybean oil (40:60)
Fats and Oils Handbook
756
TABLE 8.1 6 Weaknesses and Strengths of Scraped Surface Heat Exchangers Compared with Cooling Drum9 Type of product
SSHE
Drum
Normal household margarines Soft household margarines Half-fat margarinedlow-fatspreads Bakery and cream margarines Puff pastry margarines White fatdshortenings Water in oil emulsions (high water content)
+++ +++ +++ +++ ++ +++ +++
+++ + + +++ +++ + +
+++ very good; ++ medium; + poor. aSource: Crindsted.
until the 1950s and a process that was introduced then and is used almost exclusively today, namely, the process via scraped surface heat exchangers. Today, the cooling part of the churn-drum process, i.e., cooling over a drum and working in a complector, is used only for special margarines. The scraped surface heat exchanger is much more universal. It also has its weaknesses for specialized products (Table 8.16); however, these do not play a large role.
8.2.7.3.7 Churn-drum process (discontinuous). Up to -1955, margarine was emulsified in churns and subsequently cooled down in cooling drums. The socalled chum-drum process was applied (Fig. 8.24). Chums are cylindrical vessels with a double housing for cooling. They are equipped with a stirrer or beater. These beaters consist of flat, perforated shafts that are rotating against each other, driven by two motors. The rotational speed lies between 30 and 60 rotationdmin. The vessels, which can hold up to 4OOO L, are used only to a maximum of 70%of their volume. The emulsion is cooled (also by addition of ice water). Then it is brought onto the cooling drum in a layer 0 . 1 4 2 5 mm thick. The cooling drum is cooled from the inside to -12 to -18°C with liquid ammonia or brine. It rotates horizontally and the cooled margarine is scraped off after 300" rotation. Margarine is then stored for further crystallization in open vessels and is subsequently overworked mechanically to improve its structure (breaking of secondary bonds). Later, it is packed. Using the so-called drumcomplector process (complector is a brand name of the Gerstenberg company) (Fig. 8.25), the line starts with the churns (A) and (B) used alternately to make the process semicontinuous. The cooling drum (H) is fed with churned product via the pipe @) and scraped off as flakes with a knife (Qafter having passed through the drum. Via a buffer (K), these extremely thin margarine flakes are conveyed into the convector (L-T) after being allowed to crystallize for some minutes in a buffer (K). The complector itself consists of two screws, (L) and (M),which convey the margarine into a kneader 0 where it is overworked. The apparatus is kept
Fat as Food
757
Fig. 8.24. Churn for margarine production (after Stuyvenberg 1969).
under light vacuum. Complectors have a throughput of up to 4.5 ton/h. Today, only special (mainly bakery) margarines are produced via cooling drums. 8.2.7.3.2 Scraped surface heat exchangers process (continuous). In 1938, margarine production was revolutionized by an invention of the Votator company. It is a process that for the frst time allowed continuous processing of emulsions in a scraped surface heat exchanger. Single-ingredient solutions and the fat blend are combined via a multiproportioning pump, emulsified via static mixers and pumped into the scraped surface heat exchanger where the combination is cooled, crystallized and overworked. In the case of the premix process, a simple pump to convey the emulsion is needed. These SSHE (A-units) in combination with crystallizers (C-units) are well known under different brand names (Votator, Kombinator, Perfector, Unitator). The sequence of coolers and crystallizers in the plant depends on the oil and fat composition and the product properties desired. Figure 8.26 shows an arbitrary example of such a combination, suitable for producing a household margarine, SSHE consist of a tube that is cooled down from the outside to -25°C (liquid ammonia, frigen, brine). This tube has a maximum length of 3 m (10 ft.), and its
Fats and Oils Haihdbook i
758
Fig. 8.25. Margarine line for the drum complectwprocess.
inner diameter is up to 250 mm (10 in.), allowing for good heat transfer. In this tube is a shaft, leaving a space of 7-12 mm (0.3-0.5 in.) between the shaft and the tube's inner wall. The shaft rotates at high speed (up to 800 rotations/min). The margarine emulsion is pumped through this annular space. It is cooled on the inner surface of the tube and solidifies (Fig. 8.27). Emulsion tp
tp = Product temperature -4O'C
Scraped cooler
tp
- 28.c
Scraped cooler
tp
-1O'C, 3OC-600 min-i
- W C , 300-600 min-i
- 18'C
Crystallizer
-
B Unit
Margarine, packed
100-200 min-i
for slowly crystslliiing blends
Fig. 8.26. Margarine production with scraped surface heat exchangers SSHE (number and sequence of the tubes are an example only).
Fat as F o o d
759
Fig. 8.27. Cross section of a scraped surface heat exchanger and a crystallizer. The number of A- and C-units and their sequence depend on the margarine composition and the plant’s throughput. Because the size of the cooling tubes is limited by physical constraints (handling and modular composition of the plant), the cooling area needed has to be achieved by a series of units. Well controlled cooling creates many crystal nuclei of the p’ type, and crystal growth is slow. To ensure overworking of the solidifying emulsion and prevent the tube from being blocked by solidified product, the shaft that rotates in the tube has two to four rows of knives. These knives are flexible so that, driven by the centrifugal force, they touch the inside of the tube when rotating and scrape off the cooled solid margarine emulsion. The cooling tubes of the plants are also called A-units. To enable crystallization of the emulsion, A-units are used in combination with so-called C-units (crystallizers). Coolers and crystallizers are used in different configurations that depend on the fat blend and the product properties intended. Applying blend-specific cooling temperatures on the one hand ensures that the heat of crystallization is removed; on the other, controlled crystallization is guarateed. Wild crystallization must in any case be avoided because it can lead to uncontrolled changes in the product properties. A crystallizer is also a tube, containing an inner rotating shaft. However, in comparison with the diameter of the shafts in A-units and also to the diameter of the tube, its diameter is very small. That means that the opening between shaft and tube is very large (100-200 mm; 4-8 in.). Three rows of pins that are regularly distributed on the tube (stator) jut out from the tube wall interior. The shaft (rotor)
760
fats and Oils Handbook
also carries two rows of pins. These rotate with the shaft through the gaps left by the pins fixed to the tube. The sheer stress from overworking the product ensures the homogeneity of the emulsion and its plasticity. Crystallizersare not cooled (Fig. 8.28). Because of its low volume, the residence time in an A-unit is only 5-10 s. Because a crystallization time of -7 min is needed, the plant has to be supplemented by parts with higher residence times. To pass through a C-unit takes the emulsion -2-3 min. To ensure sufficient crystallization time for slow crystallizing blends, holding tubes (B-units) can be added at the end of the SSWcrystallizer combinations; their length and volumes are adjusted to the blends that are to be processed. An impression of a total plant (Kombinator, trademark of SchrMer company) with scraped surface heat exchangers and crystallizers is given in Figure 8.29. The emulsion is pumped into the plant at -40°C. There is a rule of thumb that 2.0 ton/h of normal household margarines can be produced per square meter (per 10 ft2) of cooling surface. Thus, the number of scraped surface heat exchangers and crystallizers depends not only on the composition and the properties of the product but also on the output of the plant. In addition to the heat of crystallization of the fat (-160 Idkg), the specific heat of the oil, the fat, the water and the ingredients has to be removed. In addition, the problem arises of how to dissipate the heat of friction and the mechanical heat equivalent of the rotational work. The emulsion leaves the plant at 10-20°C. From these figures, configurations for plants from two well-known European manufacturers can be derived (Table 8.17). The volume of a processing unit with a throughput of 3.5 ton/h is -360 L, accounting for a residence time in the unit of -400 s. The internal pressure rises to 35 bar. The most recent Kombinators, for example, allow throughputs of more than 10 ton/h and pressures of up to 120 bar. These plants are also available for sterile processing, allowing sterilization with steam of 140°C. For some special margarines, usually for bakery margarines with high amounts of palm oil, B-units that allow additional residence time are added, thus
Fig. 8.28. Crystallizer with shaft drawn (courtesy of Schrijder, Lubeck).
761
Fat as Food
Fig. 8.29. Schematic view of margarine processing unit (courtesy of Schroder, Lubeck).
ensuring a more complete crystallization. Such tubes can have a length of several meters. Apart from the process described above, it also possible to divert part of the emulsion stream just before it enters the B-unit. This part is then fed into the uncooled emulsion before its entry into the first A-unit (recirculation process). TABLE 8.1 7 Configuration of Processing Lines Consisting of SSHE and Crystallizers (Line Layout from Two Major European Manufacturers)
Number of A-units Number of C-units Cooling area (m*) per A-unit total A-units per t o n h installed per ton/h needed Installed electrical capacity (kWh) A-units C-units Total Size of total unit (mm) length (incl. drive) width (incl. NH,-unit) height
3.6 t o n h
7.2 t o n h
10.5 t o n h
2-3 2
4-5 2
6 2
0.80-1.05 2.1 -2.4 0.55-0.66 0.5
0.80-1.05 4.0-4.2 0.55-0.66 0.5
0.80-1.05 4.8-6.3 0.55-0.66 0.5
50-52 20-22 70-74
95-1 04 20-22 115-126
125-1 50 20-22 145-1 52
3300-3800 1450-2 100 22 OO-3000
3300-3 800 145c-2 100 2900-4700
3300-3800 1450-2 100 35004000
762
Fats and Oils Handbook
When fed into the warm emulsion, it is melted down with some crystal nuclei remaining that inoculate the emulsion to achieve better crystallization. However, this process is applied only rarely today because it has few advantages. In the event of a stoppage in the packaging line with no further product stream, the scraped surface heat exchangers would be immediately blocked by rock-hard solidified emulsion. In this case, it would be necessary to warm it up in order to melt down the solid blocking product. This means that the unit has to be emptied of ammonia, filled with hot water, the water then removed and the unit refilled with liquid amme nia. To avoid this time-consuming procedure, the throughput of the SSHE is drastically reduced in case of a packaging machine breakdown. Via a valve placed between the last SSHE or crystallizer and the packaging machine, the reduced amount of margarine leaves the machine. It is called rework and is either collected in a vessel (if a premix system is applied, this may be the premix vessel), remelted and later dosed again into the stream of warm emulsion or it is continuously reprocessed. If the rework is remelted and dosed into the system just before the entry into the first SSHE (thus circulating), the process is called closed rework. Figure 8.30 shows a Kombinator plant, i.e., a plant consisting of A-units and C-units. Usually the processing plant is set up according to the capacity of the packaging line(s). The largest packaging machines that are common today are six-lane machines with 360 tubs per minute. The equivalent processing units therefore have a throughput of more than 10 ton/h. 8.2.7.3.3 The influence of temperature, residence time and rotational speed. Three parameters, namely, temperature, residence time and rotational speed influence the product properties. In the cooling units (A-units), temperature and residence time are coupled to each other, because higher temperature and longer residence time have the same effect as lower temperature and shorter residence time. The circumstances in the C-units (crystallizers) are somewhat different because they are not cooled; this reduces to two the degrees of freedom that can be influenced. Because the temperature difference between the SSHE-surface and the emulsion is limited as a result of the limited heat transfer that can be achieved, sufficient cooling area is essential. As noted before, a rule of thumb states that 0.5 m* of cooling area is needed to produce 1.O ton/h of household margarine. To ensure good heat transfer, it helps that the scraping knives in the A-units are metal; this improves heat transfer between the knives and the surface of the cooling tube and also between the knives and the emulsion passing through the Aunit. The disadvantage of metal knives is that abrasion is higher than with plastic. The number of scrapings A that happen per minute in the cooler is determined as follows: A=K.R
~3.41
where K is the number of rows of knives of the rotating shaft and R is the rotational speed (min-'). Again, as a rule of thumb, A should be -1OOO. Because A is a
Fat as Food
763
Fig. 8.30. Photo of Schroder Kombinator (Trademark of Schroder Company, Lubeck).
measure for the shear stress that the cooled emulsion is exposed to, it very much influences the properties of the final product. 8.2.7.4 The Whole Process in an Overview. Figure 8.31 shows the complete margarine manufacturing process as it would be carried out with a five-tube combination of A-units and C-units that are arbitrarily composed.
8.2.8 Packing Today, margarine is immediately packed after it has left the processing unit. The packaging material’s task is to give optimal protection to the product during transport, shelf life and use. For household margarine, an appealing design would improve sales promotion. In addition to the mechanical protection that the product needs as it leaves the processing unit in a soft state, protection from light and oxygen is necessary. Beyond that, the packaging should be as water tight as possible to prevent margarine from drying out. Surface water evaporates from the product
Fats and Oils Handbook
764
I,
w
c
Ic.dho.rarm PastwrLing
+
knit
I
-
+
Cryrtrflizing
Gunk
Fig. 8.31. Flow chart of margarine production.
Fat as Food
765
15pm PETG Copolyester lOpm ABS = Acrylbutadienstyrene 250pm HIPS = High Impact Polystyrene
Fig. 8.32. Possible substitutes for PVC in margarine packing.
1Ovm PETG
15pm
until the head space between product and packaging (lid) is saturated. If the pack is tight, drying out is low. If not, water vapor can escape the pack so that it evaporates more and more from the product surface. A dark yellow surface, the characteristic color of the dried out fat phase, remains. Kroll (1978) demonstrated the dependency of drying out on the duration of storage. Polyvinylchloride (PVC), which was used mainly until the late 1980s, has very low water and oxygen permeability (see Chapter 8.4). In addition, it can easily be imprinted and is shape-retaining. Because of its bad environmental image, however, it has been replaced more and more by polypropylene, polyethylene and polyester as well as laminates. The properties of these substitutes (same material strength assumed-) are much worse. A five-layer laminate that was promoted as a substitute for PVC by different companies in the late 1980s (but was not successful) is shown in Figure 8.32. Table 8.18 demonstrates that some materials &e superior to PVC in single properties, but that PVC is paramount in the combination of very low water and oxygen permeability. The search for an equal substitute will have to continue. Preformed tubs or containers that are molded from a foil (form-fill-seal process) can be filled with margarine. The table also shows the forming temperatures. Furthermore, it is common to wrap margarine. The necessary consistency TABLE 8.1 8 Permeability of Different Materials Used for Packaging Permeability for Material strength 275 pm
Working temperature
PVC (polyvinylchloride) Material X Material strength 100 pm Hard PVC HIPS (high-intensitypolystyrole) ABS (acrylbutadienstyrene) LDPE (low-density polypropylene) HDPE (high-density polypropylene) PP (polypropylene) PA (polyacryl)
Oxygen (cmVmz) 32 94
Water vapor Ig/(d pack)] 0.083 0.12
("C)
krnV(n-12bar)]
[g4d crn2)l
70 70-90 95 80 90-1 10 120-1 40 120
30 1200 1000 1600 800 600 9
3 20 20 1 0.4 0.5 13
Fats and Oils Handbook
766
Marg/nnne
Coil of wrapper
(hard)
2Y Filling
c
%-
c
t
Folding dow
Wrapped margarine
a
to casepadtor
Fig. 8.33. Flow chart of margarine wrapping.
depends on the way of packaging. Tubs can be filled semiliquid. The same holds for wrappers if the wrapping foil is folded in the shape of a rectangular or square tub before filling.
8.2.8.7 Wrapper Machines. Parchment remains the wrapper material in some countries, but now paper and/or plastic-coated aluminum are mainly used; plastic laminates are also used. In the past, wrapped margarine was produced exclusively by pressing a hard block of margarine from the packaging machine and wrapping the wrapper around it. The margarine itself then is the forming element (Fig. 8.33). In later machines (Fig. 8.35), the foil enters the machine (A) and is pressed by a piston through a connecting link, thus forming a hollow squared block that is open to the upper side. Over the dosing heads (C), margarine (or shortenings or butter) is filled while the pack is lifted toward those dosing heads. The pack is closed by folding in (D), and the fold stamped flat (E). The product is expelled via F. Different folds are possible (Fig. 8.34). The characteristics of a wrapper machine are given in Table 8.19. If margarine is wrapped, the packaging machine is fed directly by the processing unit. For butter, this is usually collected in a trough from which it is conveyed by a screw into the machine. Such packaging machines exist up to 2 kg per pack; twin machines have been built up to 2 . 250 g. The above-described machine is suitable for all products that can be filled in semiliquid form because the material must be distributed evenly in the preformed pack. 8.2.8.2 Tub-Filling Machines. In Europe, shortenings are usually sold in wrappers, whereas the majority of household margarine is produced in preformed tubs. Principally there are two different types of tub-filling machines, stroked machines
767
Fat as Food
Fig. 8.34. Bottom and side folds in margarine wrapping (courtesy of Benhil, Dusseldorf)
and continuous machines. With stroked machines, the tubs stop under the dosing head for the filling. All other operations (Fig. 8.37) are then also carried out with each stroke. With continuous machines, filling takes place by means of a dosing head that swings with the tub while this is conveyed on the belt. The same is true for all other stations of the machine. Two types of stroked machines exist, rotary fillers and lane fillers. Rotary fillers have all stations mounted on or above a rotating round plate that carries the tub from station to station. In lane fillers, all stations are placed on a straight line one behind the other. The latter has the advantage that the capacity can theoretically be extended to the infinite, whereas rotary fillers consume much less space and are much easier to handle. Because of the high labor cost, there is a strong move toward lane machines in industrialized countries. TABLE 8.19 Technical Data of a Wrapper Packaging Machine (Type Multipack 8362, Trademark of Benhil Company) output (250-g packs) Pack format, minimal maximal Width of pack material roll Machine size (L x W x D) Weight Energy demand Compressed air (8 bar) Main drive Vacuum pump Transport drive
(packs/m in) (ton/h)
12C-240 of 6CL250 g 1.8-3.6 60 x 60 x 21 100 x 75 x 30 1 10-242 (double for twin-machines) 3000 x 1800 x 1950 4600 (net) 10 4.5 0.37 0.25
768
Fats and Oils Handbook
Fig. 8.35. Wrapping machine (Type Multipack 8362, trademark of Benhil Company Dusseldorf; courtesy of Benhil).
The most important part (for the product) is the dosing system. On the one hand, dosing is responsible for accuracy in product weight (underfilling must be avoided, overfilling causes economic damage); on the other hand, the dosing head is the site at which the pasteurized products leaves the hermetic piping system and comes in contact with the environment. Great care has to be taken in the construction of the dosing unit to ensure that back-growing infection is avoided and that the unit is easily cleanable, preferably by cleaning in place (CIP). Filling takes place in the following steps (Fig. 8.36): 1. Starting position after filling of a tub 2. The valve rod closes the outlet, the margarine flow pushes the piston upwards 3. The required quantity of margarine is in the piston chamber 4. The valve rod closes the upper valve seat, the outlet is opened 5 . The margarine flow pushes the piston downwards, the margarine is pressed out of the piston chamber and through the outlet into the tub
Fat as Food
IProduct infeed 3 Piston rod 5 Cutting device 7 Bypass valve
769
2 Valve rod 4 Outlet nozzle 8 Cleaning connection 8 Cleaning pipe
Fig. 8.36. Principal positions of margarine dosing heads (courtesy of Harnba, Neunkirchen)
6 . Position after filling of the tub 7. Margarine filler in cleaning position, outlet nozzles and cutting device are replaced by a cleaning pipe A complete tub filler is shown in Figure 8.38 with the technical data given in Table 8.20. Lids (A) and tubs (B) are conveyed into the machine. If the machine is fed by only one lane of tubs and lids, this one lane has to be split into as many stacks of tubsflids as the machine has lanes. The tubs are denested from the feeding
Fats and Oils Handbook
770 Tub from stack
I Soparati:
I
single tub
lnwrting into cell pkte
tauoff
I
C
Q
I
checking whaher p m n t
heated w i n
I Sealing lid
1
t
Cover lids from st&
=$
to caw packer
Fig. 8.37. Flow chart of tub margarine packing.
stack. The dosing heads are fed with product via a compensator (C). A probe (D) tests whether a tub is inserted and, if so, activates the filler. If required, a lid foil is laid on (E) and pressed on (F), or the tub may be sealed instead. Lids fed via (A) are separated into the respective number of stacks. The tubs are closed with the lids (G) and the finished packs expelled (H). Each dosing head can be adjusted individually via a motor. Thus they can be steered via a check weigher that is integrated into the machine, allowing the tubs of every lane to be filled with the statistically exact weight. For sensitive margarines with high water content, these machines are also built hermetically closed to allow running under sterile con'ditions. Tubs fed are sterilized with hydrogen peroxide, and the clean chamber housing the dosing heads is flushed with sterile air. All new machine types are equipped for CIP. In tropical countries, ambient temperatures may be so high that margarine melts. In these cases, it may be packed in tins. In countries with such low spendable income that only a day's need can be purchased (day-laborer), it is also sold in plastic sachets of 50 g, for instance. 8.2.8.3 Packing of Margarine for Artisanal and Industrial Use. Margarine for industrial or semi-industrial purposes (e.g., feeding centers or bakeries) is also filled into boxes with inserted plastic bags. Semiliquid margarines for feeding centers are also delivered in buckets; puff pastry margarine is produced in sheets that can easily be used.
Fat as
food
771
Fig. 8.38. Four-lane tub filling machine (Hamba Type B K 8004 M, courtesy of Hamba, Neunkirchen).
Machines for packing bakery margarines are therefore constructed differently. On the one hand, their consistency is much firmer; on the other hand, they are packed in bars or sheets of 2-5 kg each. Usually, slowly crystallizing blends require a B-unit (holding tube) between the cooler and the packaging machine, thus allowing for further crystallization time. Figure 8.39 shows a machine to pack 5-kg blocks. In the background, the long Bunit can be seen. The wrapper foil is hung up on a roll (A) and is wrapped around the margarine block that is extruded from the B-unit (rectangular profile) and is then cut off. The machine output is automatically adjusted to the throughput of the processing and the B-unit. The weight accuracy of the wrapped margarine blocks is f 0.1%. The length of such a machine is 1.8 m (without B-unit), compensating cylinder and dosing unit. In total, the length of the filling unit is 4.5 m. The dosing
Fats and Oils Handbook
772
TABLE 8.20 Technical Data of a Tub Packaging Machine (Type Bk 8004 M, Trademark of Hamba Company) output
(500-g packs) Machine length Including tub feeder Energy demand Compressed air (8bar) Main drive Vacuum pump Dosing adjustment Steering
(packs/min) (to&)
25C-5OOg 7.5
(mm) (mm)
4350
(Nm3h)
20
(kW (kW (kW) (kW)
5500
4.0 1.85 0.13
2 .o
heads can be mounted and adjusted respectively for blocks of 1-25 kg with a maximum speed of 25/min. Energy consumption of a machine that weighs 2.3 ton is 5.8 kW electrical energy and 8 bar compressed air for the pneumatic steering. The throughput depends on the weight of the individual packs and is 5.5 tonh (12 ton/h) with 2.5-kg (20-kg) blocks; 2.5 tonh (4.5 ton/h) can be achieved when packing 1.0-kg (2.5-kg) sheets of margarine.
Fig. 8.39. Packaging machine for bakery margarine bars (Type BKS of Bock & Sohn -Company, Norderstedt; courtesy of Bock).
Fat as Food
773
For the last-mentioned, the dosing head is mounted vertically to the machine (Fig. 8.40). Paper is fed from roll A, and the wrapped block leaves the machine to the front of the picture. This machine is 3.3 m in length, 2.4 m in width and weighs 1.6 ton. 8.2.9 Special Margarines
Margarines for special purposes differ in their properties of use, characterized by their C-value and their solids content (Fig. 8.41).
8.2.9.1 Cream Margarine. Usually, cream margarines have low melting points (30-34°C). They are used for cake fillings and decoration and must be whippable, i.e., hold the air whipped in, which is achieved mainly by the enormous number of fine crystals @'-type). In spite of their low melting point, they must have a high fat content. Despite that, a quick melting in the mouth (steep dilatation curve) is required, combined with a cooling effect. Coconut oil is ideal for such products. The good creaminess of such margarines is based on quick crystallization. Therefore, interesterified lard is also a good starting material (see also Chapter 6.4.5). The product must be easily stinable in bakeries at an ambient temperature of -22-26°C. The composition of the fat blend can be as shown in Table 8.21. These compositions guarantee a soft consistency with and without coconut oil. Coconut oil has a very fluctuating price and is usually very expensive. Therefore,
Fig. 8.40. Packaging machine for bakery margarine bars (Type BPM 200 of Bock & Sohn Company, Norderstedt; courtesy of Bock).
Fats and Oils Handbook
774
Solids content [%]
50 40
30 20 10
0 0
-
C Value l
400
'
"
!
'
"
I
800
"
'
,
'
1,200
'
'
1,600
2,000
Fig. 8.41. C-value and solids content of margarines.
it is used mainly for premium-brand cream margarines. In cases in which a firmer consistency is desired, the proportion of hardened oil 41 has to be increased at the expense of oil or hardened oil 34. 8.2.9.2 Bakery Margarine. Bakery margarines have a higher melting point (35-38°C) than those products produced for direct consumption. They do not have to melt in the mouth, but are designed to separate the crumbs as long as possible by breaking the continuity of the protein starch structure. This characteristic and their function as nuclei for boiling (steam formation that puffs the baked goods) ensure tender cake. The composition of bakery margarines resembles that of fm cream margarines. 8.2.9.3 Puff Pastry Margarine. Puff pastry margarine has to meet high demands on its tenacity. In puff pastry, it has to ensure that the many layers of the laminate stay separated. In the lamination process (whether hand or machine made), the thin layer of margarine must not break, but has to adapt to the laminate in smooth, very TABLE 8.21 Typical Formulations for Cream Margarines A(%)
Coconut oil Vegetable oil hardened to 34'C hardened to 41°C Vegetable oil
15-25 40-50 10-15 15-25
B(%) 70-80 2C-30
Fat as F o o d
775
thin layers, i.e., it has to have high plasticity. Its C-value is very high (up to 2600), and its melting point of 40-44"C is considerably higher than that of bakery margarine. Yet even today, puff pastry margarine is sometimes produced on cooling drums (see also Table 8.14). The plasticity of margarines is inversely related to the amount of product that postcrystallizes after having left the processing unit and B-unit. Madsen (1981) compares three puff pastry margarines produced under different processing conditions (Table 8.22). The processing units are Perfectors (trademark of Gerstenberg Company) that consist of A and C-units. All margarines of this series of trials contained 16% water, 0.1% of a distilled monoglyceride (Dimodan PM, trademark of Grindstedt Company) and 0.1% soy lecithin as well as the fat blend. The trials were run on a processing unit of experimental size. Although the results cannot in any case be transferred directly to production-sized units, clear trends can be detected. Only the product run via the cooling drum (trial 5 ) and the one run with low throughput, i.e., high residence and also crystallization time (4), had sufficient plasticity and were not greasy. Apart from the formulations based on tallow (that had P-crystals, as would lard), all blends crystallized in P'-modification. The postcrystallization behavior of the different margarines was also very different. Only cooling with a drum allows the emulsion to be supplied to the cooling surface in such a thin film that heat is immediately carried off. By comparison, the thickness of that film in a scraped surface heat exchanger (A-unit) is 7-12 mm, which is the size of the gap between the wall of the cooling tube and the rotating shaft. Here also the product is overworked during cooling; while on the drum-cooling, it is stress free and kneading is done later. This ensures a more plastic end-product. This is why producing puff pastry margarines on SSHE requires special know-how and many companies remain with the cooling-drum process. Before distribution, puff pastry margarine has to ripen for -4 d in an environment of -15°C. It is also possible to produce puff pastry with shortenings. In that case, however, there must be some water present in the shortening (Hoffmann 1989) to enhance the "puff' in order to obtain a good product. TABLE 8.22 Processing Conditions for Puff Pastry Production Trialsa ~
Sequence of units Throughput Residence time Exit temperature with product base Tallow Palm oil Soybean oil aSource: Madsen (1 981 1.
(kgh) (5)
Trial 1
Trial 2
Trial 3
Trial 4
Trial 5
A-C-A 50 90
A-A-C
A-A
50 40
A-A 20 1 00
Drum
50 90
9 10 12
13 12 18
7 8 10
1
("C) 10
8 10
0 0 0
Fats and Oils Handbook
776
8.2.10 Half-Fat Margarine
Most margarine laws allow for half-fat margarines with fat contents prescribed -40%. The demand for these products, which were first introduced in 1964, has increased considerably over the past 30 years, because they offer a real opportunity for reduced calorie intake while maintaining good taste. In 1981, in the whole of Europe, only 28 such products were marketed (Madsen 1984), whereas eight years later in the United Kingdom alone, there were more than 200. Table 8.23 gives some details concerning the amount of half-fat margarine, half-fat butter (see also Chapter 8.1.7) and melanges (see Chapter 8.2.12) sold in individual countries in 1988 and 1995. In such reduced-fat products, the water phase has to be stabilized with thickeners, because the emulsion and the c,rystal network alone are not able to guarantee temperature stability and good shelf life properties. In addition, dry matter is increased by the addition of milk proteins as milk powder. Half-fat margarines normally contain preservatives, because the water droplet distribution is much coarser than with normal margarine, and therefore they are much more sensitive toward microorganisms that find ideal conditions in which to grow. The production of sterile half-fat margarines is not a problem, however, because machinery for aseptic filling exists. The only drawback is some expenditure for the sterilization of the packaging material. The crucial point is the period of open shelf life, i.e., the period of use at home. Even if produced and delivered in a sterile condition, a nonpreserved low-fat margarine is exposed during use to ambient conditions that can cause quick spoilage. In such cases, the supplier is made responsible for the damage although the contamination with microorganisms occurred in the home by contact, for example, TABLE 8.23 Half-Fat Margarine, Half-Fat Butter, and Melanges Production in Selected European Countriesa ~
Half-fat margarines
Half-fat butter
~~
Melanges
Production (1 000 MT)
1988
1995
1988
1995
1988
1995
Belgium Denmark Germany Finland France
15.0 13.7 22.5 4.9 26.5 49.0 4.3 34.5 0.5 2.5 28.6 2.4
15.1 2.6 90.0 1.5 23.5 69.0 6.6 40.3 1.2 20.0
5.5 10.0 2.5 4.0 2 .o -
1.7 0 11.0 5.0 0.5 1.4 1.7 0 0 -
7.8 -
0.1 24.2 5 .O 11.5 6.1 45.3 13.9 0.3 0 0.4 37.7 3.0
Great Britain Ireland Netherlands Portugal Spain Sweden Switzerland
aSource: 1988 data after Madsen (1 990).
-
-
28.5 4.5
t
777
Fat as Food
TABLE 8.24 Typical Formulations for Half-Fat Margarines
c (Yo)
B (%)
A (%) _______
~
Water phase (water to 100) Whey powder Butter milk powder Na-Caseinate Gelatine Sorbate Salt Oil phase Emulsifier Monoglycerides Monodigl ycerides Fab‘oil blend Vitamins, carotenes, flavors
1 .o-1.5
2.3 1.4 1.3
7.0 1.5-2.5 0.1
0.1
0.2
pm.
p.m.
39.3-39.8
0.2 39.1
0.2 39.3
p.m.
pm.
p m
p.m.
0.2-0.5
aSources:A, B from Madsen (1 990) and C from Nichols (1 989).
with air, dust, “knives with marmalade” and so on. In that sense, open shelf life represents a factor of uncertainty that can be avoided only by preservation. Typical formulations for half-fat margarines are given in Table 8.24. Legislation in this field has been very illogical. Stabilizers allowed in one country were disallowed in a neighboring country and vice versa. Table 8.25 shows the situation in that field for the year 1989. Meanwhile, at least for the European Community, this seems to have TABLE 8.25 Stabilizers Allowed for Half-Fat Margarine in 1989 in Some Selected European Countries (with Some Limitations)
Monoglycerides Esters of Monoglycerides Acetic acidLactic acidCitric acidTartaric acidPropylene glycol Sorbitol G a r gum Locust bean gum Carragheenans Alginates Pectins Gelatin Xanthan gum Methylcellulose (MC) Carboxymethylcellulose (CMC)
D
DK
NL
S
CH
F
X
X
X
X
X
X
X X X X
X X X X X X X
X X X
X X X X
X X
X
X
X X X X X
X X
X
X X X X X X X X X X X X X
X
X X X X X X
X
X
X
X
Fats and Oils Handbook
778
been overcome. Also, protein concentrates can be used as stabilizers (e.g., Hawley 1977 and Wallgren and Nilsson 1979). The protein content of the aqueous phase can then increase to 25%. 8.2.11 Other Emulsion Fats. With liberalized legislation in many countries, all
fat levels are legal in emulsion fats; however, the name margarine can be used only for some distinct fat levels, i.e., usually 80% fat for the full-fat version and 75, 50 and 25% of it. Six to seven years ago, the limit for stable emulsions was -20%, whereas virtually fat-free products can be produced today; there is still room for taste improvement, however, with such products. The technical requirements for producing reduced-fat spreads are given by Darrington (1988), for example. The formulations are based on those for half-fat margarine. In addition to these products, melanges and others, i.e., blends of butterfat and vegetable fats have entered the market after the recent lifting of strict legal limitations. Usually they are produced from butter with the addition of vegetable oil (Madsen 1990). Because they do not offer real advantages over the pure traditional products (butter and margarine), their success in the market in most countries (except Sweden, for example) is very limited. 8.2.12 Energy and Auxiliary Material Consumption. Energy consumption for margarine differs from type to type; 80% fat margarine accounts for approximately the following energy consumption:
Water Steam, low pressure Electrical energy for liquefying ammonia
0.75495 150-200 . 50-80 15-20
(mVton) (kg/ton) (kWton) (kWton)
For the production of shortenings, similar figures can be expected.
8.3 White Fats, Shortenings White fats and shortenings are used in large amounts as ingredients in the food industry; both are heat transfer media for shallow frying in households and feeding centers. If used for baking and in non-Enghsh speaking countries, they are called shortening. That name has usually been used for lard products that were used in baking and that made the structure of the baked good “shorter” by hulling the m b with a fat film, preventing the gluten from sticking together. Such shortenings have a water content of 0.2%maximum without any water being added. They are composed of animal or vegetable fat blends that may be partially hardened, depending on the declaration. 8.3.1 White Fats in Wrappers or Tubs
Shortenings are suspensions of fat crystals (usually of the stable p-modification) in oil or semiliquid fats. Solid fat is only -1% soluble in semiliquid fat or oil, which
Fat as Food
779
means in fact that they can only exist as a suspension and do not dissolve. Depending on the structure, one speaks of pumpable or plastic shortenings (for the production process, see Fig. 8.42). 8.3.1.1 Pumpable (Liquid) Shortenings. Liquid shortenings are produced by cooling melted fat and destroying crystals that are formed by sheering or beating. The desired crystal size is 5-10 pm. Liquid shortenings contain 5-30% solid triglycerides with high melting points. The liquid state makes it more applicable in industrial baking or the food industry because it is pumpable and much more easily dosed than solid fat. If such shortenings are blended with 3-5% monodiglycerides (see Chapter 8.9), many small crystals are built as a result of their crystallization behavior and melting point. These work as crystal seeds, promoting crystallization. Added to bakery fats, the baking-promoting properties of monodiglycerides are then already incorporated into the shortening. Such shortenings should be liquid above 15°C and keep their properties of use up to 35°C. Pumpable shortenings are also used in feeding centers and the food industry for large-scale shallow and deep frying. They are packed in units ranging from buckets up to whole tank cars.
8.7.3.2Plastic Shortenings. Plastic shortenings are produced according to the margarine making process but without emulsification with a water phase. In the plant, therefore, 25% more heat of crystallization has to be deducted. This is partly compensated for by the fact that water with its high specific heat need not be cooled down. Crystallization is more difficult in shortenings because no water droplets are present on their interface to the fat to promote crystallization. If a shortening is whipped with nitrogen (preferred to air, because oxygen free), its use is improved becuse it is softer and much easier to portion. In addition, its color, which may appear gray in most solidified fats that also contain oil, Fat blend
I I
I
Melting
I anahour to margarine (see Flgm8.24)
plastic: wrappers, tubs, boxes semiliquid: buckets, containen
Fat (Shortening)
Fig. 8.42. Flow chart of shortening production,
Fats and Oils Handbook
780
whitens. There are different opinions concerning the point at which to add nitrogen in the plant. Working models for any injection position between the pump feeding the processing unit and the packing machine can be found. Plastic shortenings can also be blended with monodiglycerides, with a common amount of 510%. 8.3.1.3 Fats in Other Forms. In addition to the above described forms, fats are also delivered in flakes, as a granulate and as fat powder (see Chapter 8.3.3). It is then essential that they be stored below their melting point so as not to bake together. 8.3.1.4 Fat in the Form of Slabs. Today, fat in slabs is almost exclusively coconut based. The liquefied fat is cooled down in a SSHE or other agitated vessel to a point at which -5% is crystallized. It is then still pumpable though already highly viscous. It is poured into forms resembling those of the chocolate industry. These forms are further cooled in a cooling tunnel. Then the fat plates are beaten out of the molds and packed. The forms can be made from metal or plastic material. Metal has the advantage of easier heat transfer; the plastic molds are lighter, less noisy and can be cleaned more easily (for the production process, see Fig. 8.43). Such fats are used to produce confectionq covertures. They are also excellent for deep frying, with the disadvantage of possibly heavy foaming. 8.3.2 Fat Specialties
These fats serve special purposes. Their properties have to be adjusted exactly to the technologically founded specifications of the industrial clients. Such fats include, for example, cocoa butter substitutes (see Chapter 6.2.3.5), coatings for ice cream or fats for coffee whiteners (see Chapter 8.6.2). If these fats are not used
molten until 5% a n crystallized Filling into forms I
I
i cooling in cooling tunnel I
Pouring machine -15 to -20% -20 min
Wrapper or sachet
(Coconut-) Fat in slabs
Fig. 8.43. Flow chart of fat production in slabs.
Fat as Food
781
TABLE 8.26 Council Regulation 2991/94 of the European Community Concerning Fats Composed of Plant and/or Animal Origin
Fat group
Product categories Additional description of the category with an indication of the fat content by weight
Definition
Sales description
C. Fats composed of plant and/or animal products Products in the form of a solid malleable emulsion principally of the water-inoil type, derived from solid and/or liquid vegetable and/or animal fats suitable for human consumption, with a milk-fat content between 10 and 80% of the fat content
1. Blend
The product obtained from a mixture of vegetable and/or animal fats with a fat content of not