Bleaching and Purifying Fats and Oils: Theory and Practice, Second Edition

Bleaching and Purifying Fats and Oils Theory and Practice Second Edition

Editor Gary R. List

Urbana, Illinois

AOCS Mission Statement To be a global forum to promote the exchange of ideas, information, and experience, to enhance personal excellence, and to provide high standards of quality among those with a professional interest in the science and technology of fats, oils, surfactants, and related materials. AOCS Books and Special Publications Committee M. Mossoba, Chairperson, U.S. Food and Drug Administration, College Park, Maryland R. Adlof, USDA, ARS, NCAUR-Retired, Peoria, Illinois M.L. Besemer, Besemer Consulting, Rancho Santa, Margarita, California P. Dutta, Swedish University of Agricultural Sciences, Uppsala, Sweden T. Foglia, ARS, USDA, ERRC, Wyndmoor, Pennsylvania V. Huang, Yuanpei University of Science and Technology, Taiwan L. Johnson, Iowa State University, Ames, Iowa H. Knapp, DBC Research Center, Billings, Montana D. Kodali, Global Agritech Inc., Minneapolis, Minnesota G.R. List, USDA, NCAUR-Retired, Consulting, Peoria, Illinois J.V. Makowski, Windsor Laboratories, Mechanicsburg, Pennsylvania T. McKeon, USDA, ARS, WRRC, Albany, California R. Moreau, USDA, ARS, ERRC, Wyndoor, Pennsylvania A. Sinclair, RMIT University, Melbourne, Victoria, Australia P. White, Iowa State University, Ames, Iowa R. Wilson, USDA, REE, ARS, NPS, CPPVS-Retired, Beltsville, Maryland Copyright © 2009 by the American Oil Chemists’ Society. 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-Publication Data Patterson, H. B. W. (Henry Basil Wilberforce) Bleaching and purifying fats and oils : theory and practice / H.B.W. Patterson. p. cm. Includes bibliographical references and index. ISBN 978-1-893997-91-2 (alk. paper) 1. Oils and fats--Purification. 2. Bleaching. I. Title. TP673.P38 2009 665’.3--dc22 2008055038 Printed in the United States of America 00 99 98 97 96 95 94 5432

Contents Preface to the Second Edition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vi Preface to the First Edition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .viii 1 Basic Components and Procedures H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 Adsorption H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3 Adsorbents Dennis Taylor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4 Bleaching of Important Fats and Oils H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 5 Bleachers H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 6 Filtration and Filters H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 7 Oil Recovery H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 8 Safety, Security, and the Prevention of Error H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 9 Important Tests Relating to Bleaching H.B.W. Patterson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 10 The Freundlich Isotherm in Studying Adsorption in Oil Processing Andy Proctor and J.F. Toro-Vazquez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 11 Enzymatic Degumming of Edible Oils and Fats David Cowan and Per Munk Nielsen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 v

vi

G.R. List

Preface to the Second Edition Since the publication of this book in 1992, the bleaching process has continued to attract the attention of researchers and the edible-oil industry, and a number of excellent reviews appeared in the literature. They include: Hoadson, 1996; Taylor, 2005; Proctor and Brooks, 2005; Zschau, 1998, 2000; DeGreyt and Kellens, 2000; Erickson, 1995; and Nueman and Dunford, 2004. Although Dr. Patterson briefly discusses silica refining, much progress has been made in this area, and is found in both the open and patent literature: Brooks et al., 1991, 1992; Toeneboehn, 1994; and Siew et al., 1994. In addition, silicate refining was described by Hernandez and Rathbone. Many of the references to trace-metal analysis are obsolete, and the reader is directed to more modern techniques such as flame-atomic absorption, graphite furnace atomic adsorption, and atomic emission spectrometry involving direct current plasma (DCP) and inductively coupled Plasma (ICP) (Dijkstra & Meert, 1982; Holm, 1967; List et al., 1971; Mertins et al., 1971; Otijko, 1976). Similarly, phosphorus in edible oils can be quickly and easily measured by a turbidometer introduced by Sinram, and forms one basis for an Official AOCS Method. Sleeter (1985) discusses the pros and cons of these methods. During the 1990s, a considerable amount of work was done using the Freundlich Equation to study the bleaching of edible oils, which was summarized by Proctor and Foro-Vasquez (1996, 2005) and Dijskstra (2002). The recovery of oil from spent bleaching clay has attracted the attention of researchers. Among the techniques reported are high-temperature water extraction (Penninger, 1979), high-temperature oxidative aqueous regeneration (Kalim & Joshi, (1985), and extraction with supercritical CO2 (King et al., 1992; Waldmann & Eggers). Over the past several decades, improved degumming methods were developed. They include enzymatic degumming (Anon., 1992), total degumming (Dijsktra & Van Opstal, 1989), and soft degumming (Gibon & Tirtiaux, 1997)—all with the objective of the removal of phosphatides to very low levels as a prerequisite for physical refining. Compared to water degumming (100–250 ppm of phosphorus), these processes will remove phosphorus to levels in the 5–10-ppm range. Another process developed in the past decade is that of using membrane filtrates (Iwama, 1989; Lin Lan & Koseoglu, 1996). Press Effect in Bleaching The continuous bleaching process in a plant or in a series of batch filtrations can result in a substantial cake buildup on the filter cloth. This phenomenon is termed “press effect,” and refers to additional bleaching that can take place in the press cake. In effect, the filter press cake acts as a fixed-bed adsorption column. When the clay has additional capacity (i.e., it is not in total equilibrium with the oil it comes in contact with), it can continue to remove impurities and color bodies from the oil. The press effect was studied in the laboratory (Henderson, 1993). Gary R. List Washington, Illinois

Bleaching and Purifying of Fats and Oils, Second Edition

vii

Preface to the First Edition The title of this book was chosen to emphasize that some major operations, traditionally named “bleaching,” are much wider in scope than the mere removal of color. The selective removal of unwanted nonfat minor components from the parent fat is a form of separation which may be understood to include, in some cases, their destruction or chemical modification. Sometimes the process step is directed principally at the removal of pigment, but removes other minor components as a bonus. In other cases, the removal of components—such as remaining traces of gums, soaps, poisons of hydrogenation catalysts, and prooxidant metals—is the prime consideration; hence, we speak of purifying. Damage to the parent fat has to be avoided, and this relates to its intended use. A procedure employed in preparation of a technical or nonedible product may be quite unacceptable for edible material. The different physical and chemical operations now in use for bleaching and purifying are described in detail from a practical standpoint. Adsorption is a major technique; the way it works has come to be better understood since the instruments and methods of analysis have grown more sensitive. The theory of their operation and the structure and manufacture of adsorbent earths and carbons are simply explained for the benefit of processors already working in the fats and oils industry and for those entering it. The special use of adsorbent silica is also described, as well as other techniques. The chemical nature of the unwanted components is given; this serves to distinguish them from the fats in which they are found. Against this background, the processing of some twenty of the most important fats or oils is considered individually. Separate chapters deal with bleachers and filters—not forgetting the vitally important filter membranes. Finally, oil recovery, safety matters, and the significance of commonly used tests are considered where these have particular relevance to the subject. As with two earlier books, the author is much indebted to the painstaking work of Mrs. Marjorie Honor, who prepared the text during the years of its composition. H.B.W. Patterson, D.Sc. 9 The Wiend Bebington Merseyside L63 7RG England

viii

G.R. List

Acknowledgments Acknowledgments and sincere thanks are hereby extended to the following persons and organizations who have helped me with advice on their own specialties, whether of process detail or plant design. In a number of cases this has included ready agreement to the use of illustrations from their publications. K. Abe (Mizusawa Industrial Chemicals Ltd., Tokyo) R.F. Ariaansz (Engelhard De Meern BV) A. Athanassiadis (Extraction De Smet, SA) EMI Corporation R.C. Hastert (Consultant) P. Haynes (Schenk Filters GmbH) Y. Hoffmann (AB Pellerin/Zenith) S.S. Koseoglu (Texas A [amp] M) R. Leysen (American Soybean Assn., Brussels) Marcel Dekker, Inc., New York R.B. Morton (Grace GmbH) N. Phillips, M.E. Davies, and P.R. Shanks (Laporte Absorbents) J.B. Rossell (Leatherhead Food Research Assn.) J.R. Santhiapillai (PORIM) W.A. Starkie (P and S Filtration Ltd.) A. Sen Gupta (Walter Rau AG) F. Veldkamp (L.F.C. Lochem BV) A.J. Weir (Sutcliffe Speakman Carbons Ltd.) J.L.B. Wesselink and F. de Vooys (Norit NV) D.C. Wolfe (United Catalysts, Inc.) F.V.R. Young (Consultant) W. Zschau (Süd-Chemie AG)

A Note on In-text Reference Citations in the 2009 Edition: Bibliographic reference citations have been changed from the numbered format to the new AOCS standard (Author Name, Date) system. The reference list can be found at the end of this volume. To facilitate use of this revised Edition, not that all references to other sections of this book are indicated as follows: The section name appears, along with the chapter number if the section is from another chapter. It is our hope that these changes will allow you to navigate through this updated edition easily.

Chapter 1

Basic Components and Procedures H.B.W. Patterson

The Nature of Fats and Oils In the Preface, bleaching and purifying are shown to be forms of separation of unwanted minor components from oils which, ultimately, can mean the destruction of some of them. One must recognize that the method chosen must avoid damaging the oils themselves and, if possible, their beneficial minor components. Constraints on procedure are therefore inevitable, and compromises may have to be made: these are closely related to the intended use of the final product—whether edible or technical. Finally, one must always keep in mind the relative costs of different methods, as compared to the technical gains in quality achieved (James, 1958). Some description of the chemical and physical characteristics of fats and oils and of their common minor components will make any future discussion of the methods of bleaching or purifying more readily understandable. Triglycerides The basic unit of a fat consists of a molecule of glycerol combined with three molecules of fatty acid. When all the fatty acid molecules are of the same kind, the result is described as a simple triglyceride; if more than one kind is present, it is a mixed triglyceride (Fig. 1.1). When the melting point of a triglyceride is below ambient temperature, it is commonly called an oil, and if above, a fat. The same material may be referred to as one or the other depending on the zone in which it is being handled. Various classes of fatty acids exist, and have a marked influence on the triglyceride in which they occur. In a mixed triglyceride, the sequence in which different fatty acids are distributed over positions 1, 2, and 3 has an influence on its character, especially its melting behavior (Sonntag, 1979; Taylor, 1973). Positions 1 and 3 are slightly more

=

O H2C—OC—R1 =

O HC—OC—R2 =

O H2C—OC—R3 Fig. 1.1. Structure of triglycerides. 1

2

H.B.W. Patterson

exposed to chemical attack than position 2 (Drozdowski, 1977; Kaimal & Lakshminarayana, 1979; Paulose et al., 1978; Sebedio et al., 1981). These matters are not important in regard to bleaching. The size of the triglyceride molecule is 1.5–2.0 nm (15–20 Å, 0.0015–0.002 µmL), so it can pass readily into meso- and macropores (see Use of Carbon) of an adsorbent, but not into micropores (under 2 nm in width). When fats are partially hydrolyzed, mono- and diglycerides result. These are important for industrial use as emulsifiers, but normally are present in crude oils in very small amounts. Just as differing fatty acids markedly affect the character of an individual triglyceride in which they occur, so the presence of differing triglycerides affects the character of a natural oil or fat, which often contains several varieties.

Fatty Acids When a hydrocarbon chain is oxidized, terminally, so as to contain a carboxylic group, CH3-CH2—CH2CH2COOH the product is described as a fatty acid because many varieties of such acids occur in fats. The three simplest acids, formic, acetic and proprionic, are not included since they do not show the immiscibility-with-water characteristic of higher members. Butyric acid (CH3-CH2-CH2-COOH) is included only because it is found combined with butter. From the six-membered carbon chain (caproic acid), immiscibility with water grows as the chain lengthens. Because of the mechanism of their biochemical synthesis, the quantity of fatty acids containing an even number of carbon atoms in an unbranched chain vastly exceeds the total amount of other types (Taylor, 1973). However, because analytical techniques have improved since the 1950s, small amounts of many fatty acids with branched chains and chains containing an odd number of carbon atoms have come to light. For example, we may have a fatty acid with a total of 18 carbon atoms but containing a straight chain of 17 carbon atoms and an additional methyl group at one point or another along its length. The structural variants are positional isomers of one another. The 17-carbon atom unbranched-chain fatty acid, heptadecanoic (margaric) acid, occurs widely in animal fats, such as tallow, in very small amounts (around 1%), but is virtually absent from vegetable oils. Besides these, fatty acids were identified which contain epoxy, keto and hydroxy groups, and some which include in their chain three- (propenoid) and five- (furanoid) membered rings (Sonntag, 1979). These minor varieties usually account for less than 1% of the amount of fatty acid in common fats; occasionally a rare seed oil or fish-liver oil is found in which over one-half of the fatty acids are of exceptional types.

Saturated Fatty Acids When all carbons in the chain hold their full complement of hydrogen, the fatty acids are saturated. These are the most stable fatty acids; whether free or in combina-

Basic Components and Procedures

3

tion, their molecules pack together when solid more easily because of their regular contour. This causes a higher melting point, and as chain length increases, so does the melting point. The melting point of a fatty acid with an uneven number of carbon atoms in the chain is slightly lower than that of the even-numbered one immediately preceding it. The hydrophobic character increases with chain length. This means their sodium salts (soaps) become less soluble, but will form a more stable lather. Lauric (C12), palmitic (C16), and stearic (C18) acids are the most common saturated fatty acids (SFAs); chain lengths of up to 32 carbon atoms are found. SFAs are the most resistant to oxidation and other forms of chemical attack. They are not open to the same form of destabilization when heated with activated earth as may occur with fatty acids where considerable unsaturation is present in their composition.

Unsaturated Fatty Acids If a hydrogen atom is missing from each of two adjoining carbon atoms in the fatty acid chain, a double bond forms between them. Those fatty acids which contain only one such double bond are described as monounsaturated. Double bonds are potential points of oxidation and other forms of chemical attack; the vulnerability increases rapidly as the number of double bonds increases. Double bonds introduce an uneven feature into the chain. When the remaining hydrogen atoms on the two adjoining carbon atoms lie on the same side of the chain, the latter is seen as assuming an arc at that point, with the two hydrogen atoms lying toward the outside of the arc; this condition or isomer is known as a cis double bond. When the remaining two hydrogen atoms lie on opposite sides of the chain, a slight kink or dog-leg effect arises and is described as the trans isomer. These two forms exist because the rotational freedom of a single bond is lost, and the double bond now introduces a restriction or spatial rigidity. Unsaturation decreases hydrophobic character in comparison with a SFA, but as with the latter, an increase in chain length increases the hydrophobic effect. The trans isomeric form of fatty acid has the higher melting point, seemingly because the carbon chains pack together more easily to form a stable structure in the solid state. Cis isomers are associated with softness or liquidity, and this form is markedly dominant in the natural unsaturated fats and oils. Cis and trans forms are geometric isomers. If the double bond is located at different positions in the chain, these are positional isomers, and they also may have distinctly different physical characteristics, as Fig. 1.2 makes clear. Oleic acid is the most common unsaturated fatty acid found in plant and animal fats. Since it is not especially vulnerable to oxidation, oils in which it is the predominant unsaturated fatty acid, such as olive oil and groundnut oil, have good flavor stability. Palmitoleic acid (a 16-member carbon chain with a double bond at the C-9 position) also occurs widely.

4

H.B.W. Patterson

m.p. 16°C Oleic acid (cis-9 octadecenoic acid) Positional isomers Petroselenic acid (cis-6 octadecenoic acid) m.p. 30°C

Geometric isomers

Geometric isomers

m.p. 42°C Elaidic acid (trans-9 octadecenoic acid) Positional isomers Petroselaidic acid (trans-6 octadecenoic acid) m.p. 51.9°C

Fig. 1.2. Effects of geometric and positional isomerism (m.p.: melting point).

Polyunsaturated Fatty Acids By convention, if a fatty acid contains two or more carbon-carbon double bonds within the hydrocarbon chain, it is classed as polyunsaturated. Almost all polyunsaturated fatty acids (PUFAs) contain a methylene-interrupted configuration (-CH=CH-CH2-CH=CH-). As stated above, the double bond is vulnerable to attack and hence the focus of chemical reactivity. As the number of double bonds increases, the reactivity increases rapidly. The marine oils (where 20- and 22-carbon atom chains are substantial parts of the fats) have fatty acids with up to five and six methylene-interrupted double bonds. Figure 1.3 shows the well-known arrangements of double bonds in which unsaturation occurs. The possibility of an interaction between PUFAs and activated earths makes them of interest when bleaching or purifying is to be performed. Obviously, one must avoid process conditions which encourage harmful changes in the fat, as is discussed in later sections. Where a -CH2- or methylene group immediately adjoins a double carbon-carbon bond in the chain, it is known as an D-methylene group, and shows distinctly enhanced activity, which becomes even more marked if a second carbon-carbon double bond lies on its other side (see Fig. 1.3, 1,4 unsaturation). Oxidation at these sites leads to aldehydes, ketones, and free fatty acids, which are the sources of off-flavors and rancidity in edible fats. Rapid oxidation of very unsaturated oils can also yield epoxy-, oxy- and peroxy-groupings, and hence the hard, water-resistant skins derived from the so-called drying oils for paints and varnishes. Very often the natural long-chain PUFAs show a repetition of the skipped (-CH2-CH=CH-) group; this is consistent with the biochemical mechanism of their formation. This nonconjugated 1,4 unsaturation grouping may be transformed to the conjugated 1,3 unsaturation grouping of alternate double and single bonds (-CH=CH-CH=CH-) in certain reaction sequences. This conjugated arrangement is a most reactive system, and must be present, for example, for copper to act as a catalyst in the hydrogenation of oils. In spite of its instability, various examples of conjugated unsaturation occur in seed oils. Most of these are of little importance. The best known is 9,11,13 octadecenoic acid, or eleostearic acid, an isomer of the well-known natural linolenic acid (cis,cis,cis9,12,15 octadecenoic acid). Eleostearic acid is an example of 1,3,5 type of unsaturation because of the double-bond sequence. Tung-oil fatty acid is predominantly a cis,trans,trans form known as D-eleostearic acid, which melts at 49°C. This form easily converts to the trans,trans,trans or E-form, melting

Basic Components and Procedures

5

– –

H H –CH2–CH2–C=C–CH2–CH2–

cis

–

H trans

–

–CH2–CH2–C=C–CH2–CH2– H –CH2–CH=CH–CH=CH–CH2–

conjugated (1,3 unsaturation)

–CH2–CH=CH–CH=CH–CH=CH–CH2–

conjugated (1,3,5 unsaturation)

–CH2–CH=CH–CH2–CH=CH–CH2–

nonconjugated (1,4 unsaturation or skipped)

–CH3–(CH2–CH=CH)n...COO...

very common in nature; cis predominates

–CH2–CH=C=CH–CH2–

unknown fats

Fig. 1.3. Carbon-carbon double-bond arrangements.

at 71°C (Sonntag, 1979). In view of what was said earlier regarding the packing together of fatty acid chains, noteworthy is that here again the all-trans isomer has the highest melting point. The most important role of polyunsaturated oils is to act as building blocks from which the human metabolism is able to synthesize more complex molecules such as prostaglandins. These compounds have a wide variety of functions of the highest importance relating to blood pressure, male fertility, uterine contraction, nerve fibers, and so forth. Their usefulness is still being explored (Duffy, 1984; Frankel, 1984; Gunstone, 1984). Particular PUFAs—which the human metabolism needs for these purposes, but which it cannot synthesize itself—are classed as essential fatty acids (EFAs) and must be provided in the diet. The most common and important EFA is linoleic acid (cis9,cis12,octadecadienoic acid), CH3-(CH2)4 CH=CH-CH2-CH=CH-(CH2)7-COOH (11-14).

Pigments Already well-known is that moisture, suspended inorganic dirt, gums, and waxes can markedly affect the appearance of an oil, making it dull and altering its perceived shade or tint. When such unwanted components are removed by settling, degumming, alkali neutralization and washing, what is then accepted as the natural characteristic color of the oil becomes visible. This color may be derived from several pigments present in very different concentrations, and minor ones may yet remain to be precisely identified. As the chemical natures of these pigments differ, one type of processing is not equally effective in removing them; consequently the shade and intensity of color vary throughout refining. If the process step reduces or removes one color in a blend more than another, this, of course, creates the illusion that a

6

H.B.W. Patterson

fresh color is being generated in the product. Also important to add is that some color may also be derived from the breakdown of the principal pigments or their reaction with nonfatty components also present. This is often the cause of the brown discoloration and lack of clarity. Very often this defect stubbornly resists complete removal. It usually results from the mishandling and the neglect of seed or crude oil. To take precautions in harvesting and in expelling/extracting oil to minimize this degradation is prudent and economical. These processes are reviewed in some detail elsewhere (Patterson, 1989). Parcels of inferior seed are best kept separate to avoid jeopardizing the quality of one or more shipments. The important types of pigments present in edible oils are described in the following pages; some other minor components whose prior removal makes bleaching easier are also briefly described. For example, if the bleaching effect of heat is being used to destroy a particular pigment, one must first remove other materials which darken when the oil is heated; equally, if materials, such as gums, compete with pigments for a place on the surface of activated clays or carbon, they, too, are best first removed to enhance the efficiency of adsorptive decolorization (Andersen, 1962; Norris, 1982). Radiation absorption from one region in the spectrum gives rise to a complementary color associated with another region. In this connection, to consider this is convenient: Far ultraviolet = 220–320 nm. Near ultraviolet Blue = 320–500 nm. Blue green Remaining visible region = 500–720 nm.

Chlorophylls (Davies et al., 1989; Erickson, 1989; Rodd, 1959) A very important group of natural pigments arises from this basic porphin skeleton.

Four pyrrole nuclei (A, B, C, D) are linked by four methine groups (=CH-), together forming a 16-membered central ring. When the E-pyrrole positions are occupied by arrangements of methyl and ethyl groups, we have various porphyrin structures which show characteristic absorption bands in the visible region of the spectrum. Porphin rings form complex metallic derivatives with metals (e.g., Mg, Fe, Cu, Zn, etc.). The complexes containing magnesium (-phyllins) or ferrous iron

Basic Components and Procedures

7

(-haems) retain the metal less tenaciously. Chlorophylls contain a partially reduced porphin skeleton (the E positions of ring D) with additional groups as shown (Fig. 1.4). Chlorophylls constitute the green pigments of plants. The higher plants and green algae contain forms a (bluish green) and b (yellowish green) in proportions of about 3:1. Marine algae contain forms a and c. Parcels of Antarctic fish oil are found in which the marked green color shows resistance to refining and hydrogenation (Zschau, 1990). Red algae contain mainly a and d forms. These different chlorophylls have their own distinctive, but related, absorption spectra (Pfannkoch & Gill, 1990). The pure chlorophylls take the form of waxy blue-black crystals; in organic solvents, a red fluorescence is observed (Erickson, 1989). A loss of magnesium from chlorophylls a or b during extraction/processing operations generates pheophytins a or b, respectively. These are important olivegreen pigments; the magnesium is replaced by two hydrogen atoms. The loss of the carbomethoxy group experienced at high- process temperatures as well as the loss of the magnesium atom generates pyropheophytins a and b; the carbomethoxy group is replaced by a hydrogen atom. In contrast to the brilliant clear-green or yellowishgreen color of chlorophylls, the pheophytins present a duller brownish green. Chlorophyll acts as a photosensitizer for the production of singlet oxygen, and hence, possibly the initial oxidation of oils (Gunstone, 1984; Taufel et al., 1959; Usuki et al., 1984). After traces of chlorophyll (e.g., 4 ppm) are destroyed, free radical types of oxidation take over. Chlorophylls are reasonably unaffected by alkali refining and not especially thermolabile in deodorization. They are most diminished by acid-activated clay adsorption.

Carotenoids The carotenoids are easily the main source of yellow/red color in plant and animal fats, with the coloration of the latter being much affected by diet, hence, varying with season and location. Over 70 varieties of carotenoids are recognized. As a class, they are built up from isoprene units, and contain both cyclic and acyclic forma-

Fig. 1.4. Chlorophylls.

8

H.B.W. Patterson

tions; they also include derivatives of the polyisoprenic units thus obtained. The isomeric hydrocarbons—lycopene D-, E-, and J-carotene—are well-known, especially E-carotene. When the elements of water are added to each half of a E-carotene molecule that was divided at its central double bond, two molecules of vitamin A are produced (Fig. 1.5). An adult needs about 1 mg of vitamin A per day. Other noteworthy carotenoids are the ketonic or hydroxylic derivatives, the xanthophylls, such as lutein (Fig. 1.5). Absorption occurs in the near-ultraviolet and the bluegreen portions of the spectrum (420–475 nm), hence, the yellow/red appearance. The absorption is related to the extensive system of conjugated double bonds. Goodquality palm oils of different origins show peak absorptions in the region of 458 nm. Modest deviations from a precise absorption pattern are explicable by some variation in the carotenoids present in different species. Their outstandingly high-carotene contents are in the range of 500–2000 ppm. Where any particular palm oil has suffered damage, not only is the height of the peak (E 1%/1 cm 458 nm) diminished, but also a displacement toward 450 nm may be evident. Carotenoids, in general, are fat-soluble and water-insoluble, stable to alkali but unstable to heat, acids, and oxidation. Hydrogenation easily removes their color since the system of conjugated double bonds is attacked. Although in alkali neutralization some carotene may be occluded in the soapstock and removed, most of it concentrates preferentially in the oil layer. Unlike many other pigments, natural and synthetic, carotenes are singlet

Fig. 1.5. Carotenoids.

Basic Components and Procedures

9

oxygen quenchers, and therefore oppose the initiation of photooxidation (Carlsson et al., 1976; Gunstone, 1984; Patterson, 1989). In edible-oil processing, avoid oxidative bleaching since by the time substantial color is removed, the glycerides are well-advanced toward rancidity. Acid-activated clays and carbon readily adsorb carotenoids, but this is not true of synthetic silicas.

Flavines The basic flavinoid structure is shown in Fig. 1.6. When the 2,3 double bond is reduced, we have flavanones. Hydroxyl derivatives of the parent flavine are common in plants as yellow mordant pigments, and synthetic varieties are produced as dyes. Well-known flavine derivatives contain hydroxyl group(s) at the following position(s): 3; 5; 3,5,3°’,4°’; 3,5,7,3°’,4°’; and in the case of the ancient natural dye, luteolin, 5,7,3°’,4°’. Hydroxy flavines are frequently present in the plant as glycosides. Although many flavine colors possess absorption bands below 400 nm, some show distinct bands just above 400 nm, and this prompted the suggestion (O’Connor et al., 1949) that these may contribute, along with carotenoids, to the yellow part of the soybean pigment. Alkali refining does not affect the color, which is, however, readily removed by clay adsorption; hence, deodorizing brings no further change. The removal of flavinoid colors presents no problem in oil processing; interestingly, some flavines have a distinct antioxidant effect, evidently related to the phenolic-type hydroxyl groups present (Hudson & Lewis, 1983; Patterson, 1989). Anthocyanidins A great many of the familiar colors of flowers and berries come directly from the presence of anthocyanins. These result from a combination of anthocyanidins and a sugar. When the anthocyanin is hydrolyzed by an acid, the sugar and the salt of the anthocyanidin are produced. The basic structure of anthocyanidin is illustrated in Fig. 1.6. The overall similarity to the flavine structure is evident. The naturally occurring anthocyanins exhibit both phenolic and basic properties, forming salts with alkali or acid. Thus, a salt with an alkali may be blue (cornflower), or with an acid, red (rose, geranium), while the free anthocyanin is violet. Sugar units attach either

Fig. 1.6. Flavines and anthocyanidins.

10

H.B.W. Patterson

at the 3-position alone or at both the 3- and 5-position. Anthocyanins show marked absorptions in the 480–550-nm region; they may make a minor contribution to the color of an oil, such as soybean (O’Connor et al., 1949), which is not important. Like the flavines, they are not affected by alkali refining, but are readily adsorbed in clay bleaching.

Gossypols Gossypol is a phenolic substance present in the glands of traditional types of cottonseed; hence, the crude cottonseed oil may contain up to about 0.2% of it (Kenar, 2006). As might be expected from its structure, it has appreciable antioxidant properties.

Unfortunately, oxidation leads to a marked darkening of the oil in which it is present, and this is related, in part, to the formation of the pigment gossypurpurin. Pigs and poultry are badly affected by free gossypol remaining in the meal or cake after cottonseed oil is extracted; so this aspect has also come under strict control. In mild oxidative conditions, J-tocopherol present in small amounts in refined cottonseed oil may gradually be converted to the dark red chroman 5,6-quinone. As the oxidative induction period is exceeded, the chroman 5,6-quinone itself begins to oxidize, and the color lightens as it is destroyed. This effect, of course, is far too late to be of use to the manufacturer of an edible product who is committed to flavor stability as well as light color. Fortunately, gossypol and kindred substances are amenable to the alkali refining of the crude oil, and this is performed without allowing the temperature to rise too high. The much lighter color of the resultant semirefined or washed cottonseed oil may be further lightened by an activated clay treatment. A huge amount of research was done on these topics (Anon, 1978; Patterson, 1989), from which, obviously, the old adage “prevention is better than cure” applies. Hence, precautions commence at harvesting in avoiding conditions which allow seed to rot or overheat. Freshly expelled crude oil must not remain hot and in contact with air; this applies equally to fresh solvent-extracted oil. In each case, the risk is that color could be fixed by oxidation and, therefore, its subsequent removal by alkali refining, washing, and clay adsorption would be less effective (Hutchins, 1976; Patterson, 1989). Plant breeders have succeeded in producing cotton plants bearing glandless seeds which yield superior oil, meal, and lecithin; an improvement in their resistance to insect pests is being sought (Cherry, 1983; Patterson, 1989).

Basic Components and Procedures

11

Tocopherols and Chromans Tocopherols are best known as antioxidants widely distributed among plants, but much less common in animal tissue. Their basic structure is shown in Fig.1.7, in which one can see that they are derivatives of tocol, itself derived from chroman. D-Tocopherol is identified with vitamin E. The ranking of antioxidant potency is generally accepted as G > J > E > D, but to generalize on this matter is misleading since factors such as temperature and concentration can affect tocopherols’ behavior. The absolute and relative amounts of different tocopherols in different plant species vary enormously (Bieri, 1984). They are most effective when present at the same concentration levels found in nature. At higher concentrations, their efficiency drops steadily, and may actually reverse to become prooxidant. Lauric, palm, and olive oils contain less than 100 ppm of total tocopherols; sunflower, cottonseed, and groundnut, several hundred ppm; soybean and maize, around 1000 ppm and above; wheat-germ oil ca. 0.5% and rice-bran oil, between 2 and 4%, this being a major part of the total unsaponifiable matter present, which can itself range from 3 to 8%. The long side-chain induces such a high solubility in oil that, in spite of the phenolic hydroxyl group, tocopherols are not lost to more than a very small percentage in the soapstock and washes of alkali refining. Adsorption on activated clay is also low, but activated carbon at its usual level of usage up to 0.5% of carbon/oil may take out about one-half of the tocopherols present, and in the highly unlikely use of around 2% of carbon, the greater part of the tocopherol is likely to be adsorbed. Deodorization, especially at temperatures of 235°C and above and at higher vacuum, does strip out substantial amounts of tocopherol (Klagge & Sen Gupta, 1990) so that deodorizer condensates are processed to extract them. At the same time, the products in which the deodorized oil is incorporated may have tocopherol concentrate added to them to restore their oxidative and hence, flavor stability. Hydrogenation of oil causes little or no loss of tocopherols. At worst, normal refining leaves more than one-half of the original tocopherol content. Pure tocopherols are colorless or light yellow. Comparatively, mild oxidation can eventually break them down. In the case of J-tocopherol, as shown in Fig. 1.7, this produces the dark-red chroman 5,6-quinone, which resists bleaching stubbornly. This feature is true for many colors produced by oxidative degradation (Rich, 1964). As was previously mentioned, the avoidance of oxidative conditions where feasible, from harvesting to the end of processing, is a worthwhile precaution. Phosphatides (Gums) Complex esters which contain phosphorus, nitrogen bases, sugars, and long-chain fatty acids are classed as phospholipids. The phosphatides in oils are fatty acid esters of glycerol—which, at the same time, are also esters of phosphoric acid. The phosphoric acid may also be linked with a nitrogen base or a sugar, and a cation such as magnesium, calcium, or sodium.

12

H.B.W. Patterson

Fig. 1.7. The tocopherols.

Figure 1.8 shows a phosphatidic acid and the well-known ester, D-lecithin, in which the phosphatide group contains the base choline. If the phosphatide group is attached at the 2-glyceryl position, the compound is referred to as a E-lecithin. The D arrangement is more common. R1 and R2 are the long-chain hydrocarbon units of fatty acids. In natural phosphoglycerides, R2 is frequently unsaturated. The lysophosphatides are phosphatides in which one acyl group was removed by hydrolysis. Chemical hydrolysis is random, but enzymes are specific (e.g., phospholipase A removes the 2-acyl group). Their general behavior is comparable with phosphatides, notably regarding solubility in oil or water. Both D- and E-cephalin are other wellknown phosphoglycerides; they contain ethanolamine as the nitrogen base. The presence of phosphatides in many fats and oils is one of the most important reasons why the term “bleaching” is recognized as inadequate when describing the improvement in the quality of a crude oil which is being sought. Mere removal of pigment is not the only requirement; it may not even be the most important.

Basic Components and Procedures

13

Fig. 1.8. Typical phosphatides.

Phosphatides, as the principal constituent of gums in the crude oil, severely interfere with the efficiency of subsequent process steps if allowed to remain. Thus, in alkali neutralization, phosphatides’ presence causes increased amounts of neutral oil to be emulsified and hence, lost in the soapstock; diminishes the adsorptive action of clay and carbon by adhering to their surfaces; similarly poisons nickel catalyst; darkens the color of an oil if they become broken down by heat; and can lead to impaired flavor stability. The main responsibility for this last ill-effect, however, is linked with traces of prooxidant iron liable to persist in refined oil along with the gummy material rather than phosphatides themselves (Dijkstra & Van Opstal, 1989). A wide variation exists in the typical phosphatide content of different vegetable oils. For a long time, crude oils—such as olive, palm, palm kernel, babassu, and coconut—were known to contain a very small amount, perhaps a small fraction of 1%, of phosphatides; soybean, maize, sunflower, linseed, rapeseed, and cottonseed hold substantial amounts, best dealt with early in their processing; fats prepared from the specialized animal tissue containing large deposits of fat (tallow, lard, blubber) are low in phosphatides, but this is less certain since the whole carcass was processed. Fortunately, the composition and behavior of phosphatides have become much better understood since the 1950s; thus, effective and economical ways were discov-

14

H.B.W. Patterson

ered of getting rid of them before neutralization, bleaching, hydrogenation, or other oil-modification steps are attempted. Briefly, phosphatidic acid in its undissociated form is soluble in oils but not in water; if it can be caused to dissociate, it hydrates, flocculates to micelles or liquid crystals, and can be encouraged to pass to the aqueous phase by settling or centrifuging. Similarly, when the magnesium or calcium salts, which are insoluble in water, are treated with an acid or cation precipitant to remove the metal, they too will hydrate and can be separated (Zschau, 1990). This process we know as degumming. The water-binding capability increases by increasing the degree of dissociation (Braae et al., 1957; Hvolby, 1971). The relevant techniques are further described in the section on degumming (see section Degumming). Relatedly, more efficient and selective adsorbents of phosphatides are now attracting attention; these are the synthetic silicas. Preliminary aqueous degumming remains the obvious first step. This process has the further advantage of removing other unwanted minor components such as protein, sugars, and iron soap trapped in the micelles or liquid crystals (Segers, 1983, 1985).

Sterols Sterols are high-melting polycyclic alcohols of the general structure:

They are major components of the unsaponifiable portion of many vegetable and animal fats. They may occur in the free form, or the hydroxyl group may be esterified by a fatty acid such as linoleic acid, yielding a wax-like arrangement. They may also be present as glucosides. A very common sterol in animal fats is cholesterol: R = -CH(CH3) (CH2)3 CH(CH3)2. The sterols of vegetable oils are classed as phytosterols; one of the best known, E-sitosterol: R = -CH(CH3)CH2CH2CH(C2H5) CH(CH3)2, is prominent in the unsaponifiable portion of cottonseed oil and its isomer, J-sitosterol, in soybean oil (Sonntag, 1979). Some sterols occur in both animal and vegetable fats. In alkali refining, an appreciable loss of sterols is entrained in the soapstock. Deodorization at higher temperatures also removes substantial amounts of sterols, but the adsorption in clay during normal bleaching of oils is small. However, activated clay treatment may induce some chemical modification.

Basic Components and Procedures

15

Waxes Waxes are typically the high-melting fatty acid esters of long-chain fatty alcohols with a generally low solubility in oils. Other esters, such as certain sterols (as discussed in the previous section), where the alcohol is cyclic, conform broadly to wax characteristics. Several processing procedures to remove waxes are used, depending on how much wax is present and whether or not achieving a very low wax content in the final product is important (e.g., clear salad oil or solid margarine). The typical wax content of many vegetable oils may be only a few hundred ppm, rising in the case of a few oils to 2000 ppm and above. Much depends on whether the relatively high wax content of the seed shell was allowed to pass to the extracted oil. Sunflower seed is a good example since the dehulling of the seed prior to extraction is an obvious way of lowering the wax content of the final extracted crude oil. In practice, therefore, we have a situation where oils of peanut, rape, sesame, soy, palm, palm kernel, fish, tallow, and lard call for no special dewaxing step in processing. Sunflower and maize oils almost certainly will require dewaxing; crude rice-bran oil is notorious for a wax content of ca. 5% but sometimes exceeding 9%. The oil of the sperm whale contains about 75% of long-chain fatty alcohols, mainly C14–C20, saturated and unsaturated, combined with long-chain fatty acids. These waxes provide the basis of high-class durable lubricants (Patterson, 1983). For use as salad oils, oils need to remain clear for some minimal number of hours at a cold temperature, and this is generally achieved when the wax content is reduced to around 10 ppm. Chilling and filtering oils after bleaching or deodorizing them suffice for many seed oils which, in the crude state, have only a small wax content anyway. For those with a higher wax content (e.g., sunflower), a preliminary and substantial reduction of wax content is advantageous; this may be preceded by degumming, or degumming and dewaxing may be combined (see section Degumming). After neutralization, a final and more rigorous dewaxing may then follow bleaching or deodorizing, according to the intended end use (Fedeli, 1983; Forster & Harper, 1983; Haraldsson, 1983; Latondress, 1983; Pritchard, 1983; Sullivan, 1980). In the filtration steps at the reduced temperature of dewaxing, a siliceous precoat and body feed may often be used, but this is to hasten filtration and is not related to pigment removal. Trace Metals Many metals of the transition groups in the periodic table are active catalysts in promoting the breakdown of lipid hydroperoxy free radicals and, therefore, in accelerating the oxidative deterioration of fats and oils. Even in Roman times, holding olive oil in copper vessels was likely to shorten its useful life. Now that the whole topic of the stabilization of fats and oils is much better understood, naturally, particular attention is paid to traces of prooxidant metals as well as organic pigments and enzymes capable of hastening damage via oxidation or hydrolysis. Copper and iron attract the most attention because: (i) they occur widely as natural components at a very low level in various fats and oils and (ii) noncautious handling during

16

H.B.W. Patterson

harvesting/processing/shipment can raise their levels to a disastrous extent. While metals such as cobalt and manganese are sufficiently active to be chosen for use as prooxidants in the paint industry, their occurrence is too low and infrequent to be a significant hazard (Waters, 1971). Thomas (1982) makes a similar point in a comprehensive survey of other undesirable minor components to be removed from edible fats and oils. Obviously, the removal of trace metals to an acceptable level must depend on steps taken before deodorization. Table 1.1 illustrates this. Bearing in mind the current permissible upper limits (in Germany) for contents of these metals in fats and vegetable oils (As, 0.1; Pb, 0.25; Cd, 0.05; Hg, 0.05, ppm), evidently, further reduction is no longer required. Where oils were hydrogenated, one must monitor the final nickel content. The prooxidant effects of various metals (Cu, Mn, Fe, Cr, Ni, V, Zn, and Al) were compared quantitatively (Sonntag, 1979). The effectiveness of any prooxidant is limited by the type of fat against which it is acting and the conditions of its encounter (such as temperature, acidity or alkalinity, moisture, light, and catalyst concentration). As is being discovered, some organic catalysts can even exert a pro- or antioxidant effect according to conditions, especially concentration. The more highly unsaturated oils are potentially the most at risk. Next, the commercial issue is how much and what kind of harm is done. This could relate to flavor instability, worsening in color, or failure to bleach adequately when the oils are processed. Fortunately, a better understanding of the effect of trace metals has brought about improvements in the handling and, consequently, the quality of crude oils since the early 1970s. Whereas, in 1970, crude palm oil at a maximal content of 10 ppm of iron and 0.2 ppm of copper was commonly offered, by 1983, good crude palm oil at a maximum of 3 ppm of Fe and 0.02 ppm of Cu was available; some parcels at 0.1 ppm of Cu were still to be found. Keeping Cu below 0.05 ppm is desirable because even at levels as low as 0.02 ppm, one can detect some prooxidant influence. A deodorized palm oil is safest below 0.2 ppm of Fe (Patterson, 1989). List and Erickson (1980) reviewed the effects of iron and copper in crude and fully refined soybean oil. Normal beans yield oil up to 3 ppm of Fe and 0.05 ppm of Cu. For safety, the deodorized oil should not exceed 0.1 ppm of Fe and 0.02 ppm of Cu. Iron, even as low as 0.3 ppm, has a distinctly bad influence on flavor TABLE 1.1. Influence of Alkali Refining, Washing, and Clay Bleaching on the Content of Hazardous Metals (ppm) in Soybean Oil Refining step

As

Pb

Cd

Hg

Crude oil

0.02

0.06

0.005

<0.01

After neutralization (4 N NaOH)/washing

0.02

0.04

0.003

<0.01

After treatment, 1% active clay

0.01

0.04

0.005

<0.01

After deodorization

<0.01

0.04

0.004

<0.01

Basic Components and Procedures

17

stability. For animal fats like lard and tallow, to seek tighter limits than envisaged by the Codex Alimentarius is desirable and not difficult: Fe (1.0 ppm) (maximum) and Cu (0.05 ppm) (maximum) are preferable (Patterson, 1989). Generally, animal fats are deficient in antioxidants. When fish oils are selectively hydrogenated so that the remaining double bonds in the fatty acid chains are no longer so close to one another to provide an appreciable proportion of methylene groups of reinforced activity (Fig. 1.3), then the safe upper limits would be 0.12 ppm of Fe, 0.05 ppm of Cu, and 0.2 ppm of Ni. Good filtration and post-hydrogenation bleaching should easily achieve these, and in practice substantially smaller figures are likely to result. A good general rule is that filtered, bleached oils offered for deodorization should contain no more than 100 ppm of clay since part of the iron content of the latter is possibly accessible for prooxidant activity. Oils containing some PUFAs, such as soybean oil, are vulnerable. Again, with good filtration discipline, no difficulty arises in containing this threat (Patterson, 1983). Although several alloys are notably resistant to chemical attack, trace metals are likely to be present in a form in which, fortunately, they are vulnerable to attack by a variety of acids, including phosphoric, citric, ascorbic, tartaric, and ethylenediaminetetraacetic acid (EDTA); lecithin also acts in this way, which, considering its constitution, is not surprising. However, since phosphatides can cause some darkening in color during deodorization, one should make any lecithin addition after deodorization. Captured by these means, the trace metals are said to have been sequestered, and the complex thus formed is more effectively adsorbed by activated clay. Citric acid is a favorite sequestration agent (Law Kia Sang, 1984). For example, an addition of citric acid at 0.05% w/w on oil—added as a 50% solution in water and dispersed in the oil just before the addition of activated clay—will markedly enhance the sequestering action of the clay; even one-fifth of this dose of citric acid may be quite effective (Patterson, 1983). Enzymes known as lipoxygenases can catalyze direct oxidation of facts by molecular oxygen; only polyunsaturated fatty acids are attacked. Certain lipoxygenases attack particular parts of the carbon chain (regiospecificty) (Gunstone, 1984; Galliard, 1983; Grosch, 1972). When the protein of an enzyme is denatured by heating, it can release any metal which it may contain. If this is a transition group metal, such as iron or copper, a strongly catalyzed oxidative reaction, different in character from that promoted by the original enzyme, is then likely to result. Such an effect is termed pseudoenzymic oxidation. This possibility has to be borne in mind when heat inactivation of enzymes in foods is planned (Eriksson et al., 1971).

Miscellaneous Minor Components Besides the long-recognized minor components dealt with the preceding sections, a few others are described briefly. The precautions against the excessive formation of some are noted, or the point in the process sequence at which some are virtually eliminated is recognized.

18

H.B.W. Patterson

Dimers This term describes the situation where portions of triglycerides or fatty acids link together to form a larger molecule, and so may be regarded as an early step in a kind of polymerization. This effect may arise in various process steps, some mild, some severe. Obviously, in a reaction such as interesterification, where fatty acid groups detach themselves momentarily from their parent triglyceride, not only may detectable amounts of mono- and diglycerides remain behind as well as methyl or ethyl esters from the alcoholic sodium catalyst, but the remnants of the parent triglyceride molecules may continue to link together to a small extent until the catalyst is quenched. Again, where deodorization temperatures rise to 260°C and beyond, fragmenting and dimerizing of the triglyceride occurs (Eder, 1982; Rossell et al., 1981; Willems & Padley, 1985). Dimers are regarded more as potentially useless rather than toxic, and for edible purposes, conditions of processing are chosen so as to avoid or severely minimize their formation. Thus, Willems and Padley (1985) give 1% dimers as the sensible maximum acceptable for a good quality refined palm oil. Soybean oil, which because of its greater un-saturation is more at risk, is discussed in some detail by Rossell et al. (1981), who point out that an increase in viscosity is one of the most direct indications that polymerization of an oil has arisen from a temperature overshoot in processing. Cyclization of polyunsaturated fatty acids in the early stages of hydrogenation, if the temperature is not temporarily controlled below 150°C, is another risk. Here again, this is prevented from happening (Patterson, 1983) rather than attempting post-hardening removal of the polymers. It may be significant that one of the symptoms that this fault is occurring is an initial rise in refractive index before the usual drop when hydrogenation takes over. When oils are polymerized, an increase in refractive index is common. On the other hand, for technical use polymerization, especially of polyunsaturated oils, is sought by raising the temperature to 300°C under reduced pressure and in a current of steam to exclude air. The product is more viscous and paler. It has been used as a component in wool combing oil or as a soft oil in soapmaking. As would be expected, prior to this severe heat treatment the oil, usually a marine oil, is freed from organic components which would char and discolor at 300°C by an activated earth bleach; to be more secure the oil should be alkali neutralized and washed before being bleached (Patterson, 1989). The characterization and estimation of monomers, dimmers, trimers, and oligomers occurring in a variety of oils subjected to thermal and oxidative attack have been reviewed and also investigated by Christopoulou and Perkins (1989). Oils mentioned in this report include soybean and partially hydrogenated soybean oil, corn, olive, butter, and tall oils, as well as synthetic model lipids for comparison. The original paper should be consulted.

Basic Components and Procedures

19

Oxidized Fats and Oxidized Fatty Acids As oxidation of a fat progresses, a variety of oxidized compounds is produced, including epoxy-, hydroperoxy-, and peroxyglycerides. If the triglyceride begins to break up, then free oxidized fatty acids can also be present. Since oxidized fats act as stabilizers of water in oil emulsions, their presence in oil at 0.5% w/w and above blurs the separation of soapstock and increases refining loss. They are the equal of sulfur as poisons for nickel catalyst. Their presence indicates that minor components such as pigments may also have been oxidized and that a bleach-resistant product may have resulted. The oxidized fatty groups, although in some instances tasteless themselves, are the route by which potent off-flavors and rancidity result (Patterson, 1983; Hamilton, 1983). Oils containing polyunsaturated fatty acids are most at risk, although animal fats, being poor in natural antioxidants, are also vulnerable for that reason. It has been estimated that 2% of the triglycerides in crude soy oil may have suffered some amount of oxidation (Stage, 1981; Sen, 1974). Peroxide and anisidine values as indicators of primary and secondary degrees of oxidation do not tell the whole history of the oxidation which a fat may have suffered, but interpreted along with empirical experience of that species of oil they are most useful. Also, the question is: Is the oxidation spread evenly over the majority of the triglyceride molecules present or is it highly concentrated in minor polyunsaturated components? Thus, a peroxide value (PV) of 2 would suggest that approximately 0.1% of triglyceride molecules present have suffered a minimal degree of oxidation, but this is only in the worst case where the oxygen is shared out to the maximal extent (Patterson, 1989). Fortunately, alkali refining removes considerable oxidized fat, but clay bleaching is more powerful. For hydroperoxide removal, we depend on clay bleaching and deodorization (or steam stripping) (Thomas, 1981). Worth pointing out is that the breakdown of a hydroperoxide commonly results, at first, in a hydroxyl group, and activated clay bleaching of the latter leads to an increase of conjugated unsaturation as illustrated:

20

H.B.W. Patterson

Thus, the 13-hydroxy 9,11 iso-linoleic group gives the 9,11,13 isolinolenic group. The change from conjugated diene to conjugated triene is accompanied by a fall in the absorption at 230 nm and a rise at 270 nm (Mitchell & Kraybill, 1942; Ney, 1964; O’Connor et al., 1949; Pardun et al., 1968; Patterson, 1989; Rossell, 1983; Van den Bosch, 1973). Such a triene is then a candidate for further oxidation when given the chance and for the eventual production of off-flavors. This emphasizes the importance of keeping the interval between bleaching and deodorization as short as possible with a minimum of contact between warm oil and air (Patterson, 1989; Thomas, 1981). The clay bleaching presumably eliminated prooxidant catalysts and the oil was to be stored in the dark—all, of course, in its favor.

Pesticides, Nitrosamines, Extraction Solvents, Organo-Sulfur Compounds, and Polyaromatic Hydrocarbons Thomas (1982) recounts in some detail how organic pesticides persist during degumming, neutralization and bleaching, but are finally eliminated by prolonged deodorization around 190°C or very quickly at 270°C. Nitrosamines and extraction solvents such as hexane are also evidently removed by deodorization to safe limits. Certain organic sulfur compounds which have an appreciable poisoning effect on nickel catalyst are most effectively diminished in refining by the final step of deodorization. Thus, a fall in the sulfur content in rapeseed oil from 20 to 30 ppm to under 5 ppm is quoted, at which level it neither causes trouble during hydrogenation nor odors during heating. Polyaromatic hydrocarbons are discussed, but are dealt with more fully in the section on the use of activated carbon (see section Use of Carbon). Thomas stresses the need for caution in deciding what use, if any, may be made of deodorization condensate, since this contains all the unwanted classes of material removed from the parent oil. Settling The first portion of this chapter briefly described important differences which distinguish one triglyceride from another, as well as the characteristics which distinguish the various minor nonfatty components both from fats and one another. Most convenient is to deal with the several steps taken to separate these unwanted components more or less in the same order in which they take place industrially. As already stated in the Introduction, certain process steps which have long been acceptable in producing technical products are not used when operating for edible purposes, usually because of some damaging effect on the edible quality of the product. When crude oil is held in bulk storage, some separation of the different triglycerides may be possible, depending on their individual melting points. Whether or not this occurs, the deposition of nonfatty components— such as dirt, moisture, gums, and waxes—is a regular feature. The refiner has to decide whether he can profit by the separation or if he should take action to circumvent it. When even the best-quality crude whale oil was held undisturbed in store tanks (each tank held several thousand tons) for two or more years at ambient temperatures, evidently, the

Basic Components and Procedures

21

lower layer became richer in rather higher-melting material with foots concentrated on the floor. Subsequently, three days to a week of gradual heating rendered the bulk stock easily pumpable for subsequent refining and hydrogenation, the foots remaining. Crude palm oil held in bulk, even in the tropics, shows the same effect. If profitable, when emptying the tank, one can divert the lighter and heavier layers to different processing. Untreated soy oil readily deposits considerable amounts of phosphatides, sunflower oil deposits phosphatides and waxes, and rice-bran oil readily deposits the higher-melting portion of its rich wax content, even in the tropics. If a decision is made that a period of settling will encourage the separation of unwanted material from the crude oil, one may expect a period of 2–3 weeks to achieve most of the possible improvements. The supernatant bulk oil is drawn off, leaving the foots relatively undisturbed. Next, the foots are removed and processed as a batch, possibly by simple warming, to separate some useful oil, or they are stored separately until the amount accumulated justifies processing. Crude expeller or extracted soy oils contain nontriglycerides, such as gums, fines, moisture, seed fragments, meal, and proteinaceous material. One can remove these potential foots by passing the oil over a vibrating wire screen containing 0.01inch openings (Erickson & List, 1985). Cleaned, crude soy oil is still capable of depositing gums, especially if wet; hence, a deliberate first degumming is normal at the extraction plant so that subsequent storage and handling are made less troublesome and stability is more assured. A phosphatide content of around 0.3% is achieved. Of course, a limit exists as to the amount of foots that a tank can accommodate. Pumping relatively clean, dry crude oil on top of foots may disturb them, and in the foots, fat-splitting and oxidizing agents are likely found. Oil-storage tank design, operating, and cleaning are described in detail (Patterson, 1989). Dewaxing proceeds more effectively if gums are largely removed beforehand (see section Waxes), as in the case of sunflower oil; some wax may be lost with degumming, but what remains will separate more easily. After full refining, the oil is given a final winterizing and filtering if necessary. With rice-bran oil, final winterizing is a common practice. Refiners prefer to avoid storing crude cottonseed oil for several weeks because the oxidation of the gossypol fixes an inferior color. Promptly processed from crude oil, a semirefined or so-called washed cotton oil, adequately dried, is safer to store or transport (Patterson, 1989). Even moderately hydrogenated oils may commence to deposit higher melting fractions if stored with just enough applied heat to avoid complete solidification. When the preferable judgment is to keep the contents of a tank uniform before it is pumped to the next process stage or shipped elsewhere, the simplest method is to stir the contents by a side-mounted impeller or to pump the contents from the bottom and return them, always avoiding the entrainment of air (Leong & Berger, 1982; Patterson, 1989).

Degumming From the explanations stated in the sections on Phosphatides and Settling, clearly, good technical/commercial reasons exist for reducing the phospholipid content,

22

H.B.W. Patterson

early in the processing, in those crude oils where it is relatively high. A summary of the advantages of doing so follows: 1. The deposition of sludges, especially when the crude oil is wetted, is a nuisance in storage and transport. 2. In alkali refining, emulsification of neutral oil in soapstock and, hence, oil loss are significantly encouraged by phosphatides. 3. Phospholipids clog the surface of adsorbent clay or carbon and, therefore, compete with pigments and any other unwanted minor components which are desirable to remove. 4. Phospholipids obstruct and poison the surface of a nickel catalyst irreversibly. 5. At the temperatures of deodorization and, more surely, at those of physical refining, phospholipid breakdown leads to the darkening of color and impaired flavor stability. 6. Hydrated gums carry with them during their removal a proportion of unwanted minor components like trace metals. 7. Subsequent dewaxing, if needed, is all the easier in the absence of gums. 8. Degumming of soybean oil provides the principal supply of lecithin, for which an established market exists. The advent of physical refining has made the task of preliminary degumming all the more important. The matter is no longer seen as simply improving the removal of unwanted material from crude oils to produce a first-class edible product, but now the method chosen must keep environmental problems to a minimum. If energy and chemicals are saved at the same time, so much the better. As early as 1957, Braae (1957) reported on the advantage of treating certain vegetable oils with 0.05–0.2% w/w of strong phosphoric acid at 70–90°C for one minute with vigorous agitation, as a first step to precipitating phosphatides and getting rid of calcium and magnesium. Without such preliminary treatment, soaps of the latter metals persisted beyond neutralization and gave positive soap tests during the stages of washing. Crude oils of very low phosphatidic content, such as palm and lauric oils, when neutralized were much easier to wash, while rapeseed and linseed oils were notably more difficult. At lower temperatures and with less effective agitation, the acid-treatment step needed to be prolonged to as much as 10 minutes. Also, using soft water in both washes and treatment reagents was important. These recommendations are readily understandable from what was found in the last 30 years. Hvolby (1971) clearly states that the Mg/Ca phosphatidates and the undissociated phosphatidic acid are nonhydratable, whereas when broken down to dissociated phosphatidic acid, hydration proceeds in proportion to the extent of dissociation. Intensity and duration of agitation are crucial factors.

Basic Components and Procedures

23

Sen Gupta (1974) drew attention to the negative influence of any oxidized phospholipids present in crude soybean oil since these hydrate more slowly, yet he recognized (Sen Gupta, 1986) that if sufficient, easily hydratable phosphatides are present (or added), the subsequent formation of microemulsified gums drags some of the trapped oxidized phospholipids down into the aqueous layer. Even so, the presence of oxidized phospholipids in badly handled seed is a negative factor. Phospholipids in the seed are oxidized three times as quickly as the triglycerides during processing. Ong (1980, 1981) emphasizes the absolute necessity of obtaining a good-quality degummed palm oil for physical refining, and proposes this very secure standard: the phosphorous content of the crude degummed oil should not exceed 20 ppm, and the phosphorous content of the pretreated oil immediately prior to physical refining should not exceed 5 ppm. The reduction to these levels brings very low levels of trace metals such as copper and iron. Young (1989) reviews in detail the most recent quality requirements for oils being processed for ultimate physical refining. In the classical water-degumming process, one agitates about 2% of water into the oil at 70°C, avoiding air entrainment, for up to an hour before centrifuging out the lecithin sludge. One may use a shorter mixing time and a somewhat higher temperature in a continuous process (Brekke, 1980). Between 0.3% and 0.8% of phospholipids remain, with up to 3 ppm of iron. Several different techniques were used, and are still being explored to improve this standard. Practical experience in Germany showed Kock (1981a,b) and Penk (1981a,b, 1985) that the nonhydratable phosphatidic content of extracted soybean oil arose from the action of the lipoxygenases and phospholipases. If these were inactivated by steam heat and the dried bean flakes then hexane-extracted, the resulting crude oil could easily be water-degummed and then, with a light bleach of 0.5–1% of activated clay, brought to a standard suitable for successful physical refining. This process is known as the ALCON process and applies to soybean oil. Segers and van de Sande (1989) comment that whereas with average-quality soybeans the very satisfactory phosphorous content of 5–15 mg/kg after water degumming is obtained, from poor seed the phosphorous content may rise to 30–40 mg/kg. The so-called wet-acid degumming process mixes 0.07–0.15% of phosphoric acid (ca. 85% of H3PO4) with the oil at 60–80°C, and, after a 30-minute delay for the breakdown of nonhydratable phosphatides, water is added to complete hydration before centrifuging out the gums. As is now known, some phosphatides [phosphatidylcholine (PC) and phosphatidylinositol (PI)] hydrate much more rapidly than others [phosphatidic acid (PA) and phosphatidylethanolamine (PE)]. These more-easily hydrated phosphatides are capable, when hydrated, of encapsulating up to 80% of their own weight of the less-easily hydrated ones and entraining them at the time of flocculation and separation. This means that when the PC predominates at over 50% of phosphatides present, the removal of all types becomes nearly complete, but when only about 35% are of the PC class, only partial removal of PA

24

H.B.W. Patterson

plus PE types occurs, and other minor components are usually encapsulated with them. This accounts for the varied performance of acid degumming according to the makeup of gums present. Acid, usually phosphoric or citric, accelerates the breakdown and the hydration of the Mg and Ca salts, but the time allowed is likely to be insufficient to bring about the complete removal of items such as PA and PE. For oils where the phosphatidic content is low, a so-called dry degumming was used, and this is one of the simplest options. Basically, acid is added to the oil, and this serves not only to disrupt the nonhydratable phospholipids, but also to attack and sequester the trace metals. This means that the activated clay added next is adequate to adsorb phospholipids and metals. However, to obtain the best results in differing cases, some variations in procedure are preferred. Since the lauric oils contain, at most, only 0.07% of phosphatides, the added phosphoric acid assists the clay in achieving the easy set task of adsorption without difficulty. But if the common 85%-strength acid is used, this can also promote the splitting of some triglycerides and a reaction with the oil, known as phosphorylation. This, in turn, sometimes can lead to a darkening of the oil when it is being deodorized. This hazard is avoided by using instead a 0.05% w/w dose of citric acid as a 50% solution in water. This, although usually rather more expensive than phosphoric acid, works very satisfactorily (Young, 1989). If phosphoric acid is used, several researchers (Dijkstra & Van Opstal, 1989; Mag, 1980; Nilsson-Johansson et al., 1988) emphasize that the intense dispersion of the acid dose and the very brief contact time with the oil give optimal effect. Fortunately, the clay also plays a useful role in adsorbing the products of lipid oxidation. For palm oil and medium- or poor-grade tallow, a so-called dry process has been popular for about 30 years. In this process, a dose of 0.1% w/w of phosphoric acid (85% of H2PO4) is allowed to mix some minutes to disrupt nonhydratable phosphatides; a mixture of phosphoric and citric acids was also employed. Next, up to 1.5% w/w of activated clay is added, depending on the quality of the oil; a contact time of 20–30 minutes at 90–100°C is normal. The practice of then adding calcium carbonate to neutralize any remaining phosphoric acid before filtering was strongly queried (Hebendanz & Zschau, 1991), since if any calcium phosphatides form, these are nonhydratable and, therefore, the removal of phosphorus is then likely to be less satisfactory. When strong phosphoric acid (85%) was used, iron removal may have been impaired (Zschau, 1982). The persistence of phosphorous and iron compounds leads to color reversion at deodorization. The use of citric acid or dilute (10%) phosphoric acid circumvents these hazards (Hebendanz & Zschau, 1991; Zschau, 1982). In the case of dilute phosphoric acid, only sufficient acid is added so as to allow any surplus to be completely adsorbed by the following earth addition. The superdegumming process was designed in 1977, and accomplishes the following: 1. Breaks up nonhydratable phosphatides with the aid of acid at ca. 70°C and vigorous agitation.

Basic Components and Procedures

25

2. Gives hydratable material adequate time to gel. 3. Ensures, if necessary, that enough hydratable phosphatide is present (or added) to encapsulate any material particularly slow to hydrate, along with minor components such as trace metals. 4. Cools and adds water, keeping the temperature below 40°C, thus allowing hydrated phosphatides to form their typical lamellar liquid crystals. A similar process—in which a flocculating agent is added rather than additional hydratable phosphatide—is known as the special-degumming process (Bolicer et al., 1984). After about one hour, the liquid crystals have grown in size, and after judicious warming of the resultant mix, they are separated by centrifugation. Segers (1982, 1983, 1985; Segers & van de Sande, 1989) emphasizes the economic as well as the technological aspects of the process. Plant, reagent, and service requirements are moderate; a very good standard of degummed oil is achieved, not only with soybean oil but also with sunflower, rapeseed, groundnut, linseed, and grapeseed oils; the application of the method leads to physical refining and a minimum of environmental disturbance. Because of the cooling in the later stages, some waxes also are removed. A further development in the technology of degumming is presented as the total-degumming process (Dijkstra & Van Opstal, 1989). It operates on water-degummed oils, but may also be applied to undegummed oils. The true villain of the piece with regard to flavor reversion is identified as iron; hence, obviously, precautions against increasing the iron content during handling and processing are wellworthwhile. Simultaneous removal of copper by this same process is virtually certain. Degumming may be performed with citric acid or phosphoric acid of between 20 and 60% H3PO4, but no stronger, so as to keep well clear of the phosphorylation of the components of the oil (e.g., mono- and diglycerides). Very intense mixing is used for up to 2 minutes at 70–110°C; then the acid-treated emulsion is partially neutralized to pH 6 with dilute caustic soda so that no soap appears. Gums now need to be separated by the action of two successive centrifugations, after which the oil may be washed, dried, and bleached (ca. 0.5% of activated clay) before being physically refined, or it may be alkali-refined, with the prospect of a substantially improved refining factor because of the absence of gums. Bleaching and deodorizing would then follow in the conventional manner. Soapstock from alkali refining is found to be split much more easily. A key feature of total degumming is the removal of any undissociated phosphatidic and lysophosphatidic acids by forming their sodium salts, which then hydrate and will pass out with the aqueous layer. Totally degummed sunflower oil is easier to filter when it is winterized; rapeseed oil evidently loses some of its nickel-catalyst poisons when it is totally degummed.

26

H.B.W. Patterson

Purifying by Reverse Osmosis Although the employment of reverse osmosis by the method of ultrafiltration was directed in the first place at the removal of phospholipids from crude-oil miscella, it also removes other unwanted materials to a very useful extent; further, when alkali is added to the miscella, free fatty acid is converted to soap which, along with some pigments, is removed at the same time. Some basic explanation of the technology involved and the terminology presently in use is helpful at this point. Also note that clearly this technology is not yet widely accepted in the fats-and-oils-processing industry because it takes time to adopt effective and durable membranes as well as cleaning procedures which eliminate the necessity to replace the membranes too frequently (Koseoglu & Engelgou, 1990; Koseoglu et al., 1987, 1989). These problems are being confronted by much research and development in several countries since attractive gains are to be made in energy savings, greater environmental capacity, and hence, the cost advantages which these imply. Everyone is familiar with the idea of separating particles from a liquid or gas by the filtration of mixtures through a layer of porous material in the form of fine fibers, sintered metal, or glass particles. The extension of filtration to the separation of finer and finer particles, passing through the range of colloids to the separation of molecules, has met certain difficulties: (i) providing a material from which to make a filter with reasonably uniform pores becomes increasingly difficult as the required pore diameter decreases; (ii) the energy needed to drive the fluid through the filter mounts rapidly as the pore diameter is made smaller; and (iii) the material of the filter needs to be both compatible chemically with fluids being handled and sufficiently strong in the overall construction to ensure an economical working life. Fortunately, in the 1980s, new organic polymers and sometimes colloidal inorganic materials have made it possible to obtain extremely fine layers of the desired porosity supported on stronger, more porous materials. Thus, colloids and macromolecules can be separated from a continuous solvent fluid and its micromolecular solutes. Important to note is that the term “microfiltration” is taken to apply to the retention of particles in the size range 10–0.02 microns (i.e., 10,000 nm–20 nm), such as bacteria and colloidal silica, whereas “ultrafiltration” (UF) refers to the nominal pore sizes of 20 nm down to 1 nm (Meares, 1986). UF is therefore concerned with the separation of macromolecular solutes from solvent and micromolecular solutes. In describing UF, commonly given is the molecular weight of the smallest solute molecules retained by the membrane, that is, the so-called molecular-weight-cutoff (MWCO) value. Sen Gupta (1986) quotes the sizes of the mixed micelles of phospholipids present in the miscella of extracted crude soybean oil as ranging from 200–18 nm, corresponding with a molecular weight of around 500,000, whereas a triglyceride molecule approximates in size to 1.5 nm (15 Å), corresponding with a molecular weight of 850. Materials which succeed in passing through a membrane are named the permeate, and those held back the retentate.

Basic Components and Procedures

27

Bearing in mind the definitions just given, obviously, a fine or microfilter will not achieve UF and conversely, a UF membrane is quite unsuitable for fine or microfiltration. Evidently, therefore, before applying UF to a crude-oil miscella, an efficient conventional filtration must first be applied to remove suspended solids. In other words, “filtration” is the separation of a suspended solid phase from a liquid phase, and if the sieve or filter is very fine, we may correctly use the description “microfiltration”; some small entrainment of undissolved solid particles is always a possibility. A filter cannot separate dissolved material from solution, but a semipermeable membrane does separate one or more solutes from solution, and this is called “ultrafiltration” (e.g., the separation of albumin from whey). To understand how UF operates in purifying crude-oil miscella, we need to follow the description given by Sen Gupta (1986). As is well-known, in the case of a molecule such as sodium dodecyl sulfate, the sulfate part of the molecule is hydrophilic, and the dodecyl hydrocarbon chains are hydrophobic and lipophilic. In a phosphatide such as PC (Fig. 1.8), the choline phosphate group is the hydrophilic group, and the diglyceride group is hydrophobic and lipophilic. If the hydrophilic part of a molecule is represented by a small circle and the hydrophobic part as Vshaped, then Fig. 1.9 illustrates how such molecules position themselves on the surface of water, or if immersed by turbulence, how a tiny cluster forms with the hydrophilic ends facing outward and the hydrophobic parts hidden inside. This formation is described as a micelle, and the molecules as surface-active. If the continuous phase is a hydrocarbon (or other nonpolar fluid), then, of course, the hydrophobic part faces outward, and the situation is described as a reversed or inverted micelle. Next, a development called a swollen micelle is illustrated. Here, the micelle accepts insertions of foreign material insoluble in the continuous phase. This form of solubilization is a familiar phenomenon: oil or grease globules held dispersed in aqueous soap solution, sugar dispersed by phospholipid in hexane, or indeed our own blood are but three examples. Lastly, what is described as a microemulsion may form. In this, a tiny complete globule of oil may be held in water or a tiny drop of water in oil. Examples of such situations and how one type may be converted to the other are found in margarine manufacture. Sen Gupta asserts: 1. All phospholipids behave as ordinary mixed micelles, whether they are free, salts, or oxidized. 2. Metal derivatives and sugars—free or combined—are almost completely enclosed in these micelles. 3. Only a portion of particular pigments and other specific components of crude oil are taken into the micelle structures. These mixed micelles are the ones that were mentioned earlier as ranging in size from 18 nm to 200 nm. The separation of these micelles from the miscella of crude oil in hexane can

28

H.B.W. Patterson

Fig. 1.9. Micelle formations.

now be described. When a solution is divided from the same solvent by a semipermeable membrane, solvent molecules pass through the membrane into the solution, and some solute molecules may pass in the opposite direction. The pressure gradient from solvent to solution is the osmotic pressure, and it depends on the difference in the concentration of solute on either side of the membrane and the temperature at the time the movement is taking place. As the concentration of solute on each side of the membrane tends toward equality, so pressure gradient and flow rate (flux) tend toward zero. In the case of crude-oil miscella, hexane would be the solvent, glycerides and free fatty acids the permeable solutes, and phospholipid micelles and their encapsulated components the nonpermeable solutes (retentate). When a pressure is applied to the miscella greater than the natural osmotic pressure, solvent

Basic Components and Procedures

29

hexane, glycerides, and free fatty acids will then flow in the opposite direction, but the micelles are too large to penetrate the membrane, so their concentration on their side of the membrane increases. Figure 1.10 illustrates the situation as a crude-oil miscella flows down an inner tube separated from hexane solvent in a surrounding annular space. The miscella is separated from the solvent by a semipermeable membrane which is held firmly in position by a rigid porous support. Hexane, glycerides, free fatty acids, and certain components, such as tocopherols, pass into the annular space, leaving an increasing concentration of phospholipids and their occluded, unwanted components in the inner tube. Sufficient turbulence must be maintained in the miscella of the inner tube to prevent the micelles from clogging the surface of the semipermeable membrane. Sen Gupta then gives the analysis of soy oils (permeates) from which the hexane was removed by the usual distillation, and compares this with an analysis of the same oils before UF (Table 1.2). Particularly noteworthy is the great drop in metal content and the very small remaining content of green pigments. Hopefully, in developing UF to the full-plant scale, more than one UF stage could be employed, and some degree of recycling could be introduced at each stage so as to obtain the optimal technical/economic effect (Sen Gupta, 1986). This is shown diagrammatically in Fig. 1.11, and is sometimes called the feed and bleed method.

Fig. 1.10. Ultrafiltration or reverse osmosis.

30

H.B.W. Patterson

TABLE 1.2. Analysis of Soy Oils before and after Ultrafiltration (UF) of Miscellae Before UF

After UF

Phospholipids

2.0–3.0%

<0.03%

Glycolipids

0.15–0.3%

<50 ppm

Free sugar

0.10–0.15%

<1 ppm

Amino acids

ca. 0.005%

not measurable

Lipoproteins

ca. 0.1%

not measurable

Potassium

300–500 ppm

<5.0 ppm

Calcium

100–150 ppm

<5.0 ppm

Magnesium

80–120 ppm

<0.1 ppm

Iron

<0.1 ppm

<0.1 ppm

Copper

<0.1 ppm

<0.1 ppm

Carotenoids

ca. 25 ppm

<20.0 ppm

Green pigments

ca. 2 ppm

<0.5 ppm

FFA

0.5–1.6%

0.5–1.6%

Tocopherols

ca. 1250 ppm

ca. 1150 ppm

Fig. 1.11. Possible UF layout (Sen Gupta, 1986).

Basic Components and Procedures

31

Several interesting and potentially rewarding practical possibilities were brought to light and are described. 1. Seed oils such as rape, cotton, and sunflower can be fully degummed by UF, and made ready for physical refining, although pigments in rapeseed and cottonseed oils need to be reduced further by clay bleaching before the deodorization stage. The refiner must decide whether the phosphorous content calls for any activated clay or silica treatment before physical refining commences. Ong’s suggestion (1980, 1981) is that, for good flavor and color stability, the limit of 5 ppm of phosphorus should not be exceeded for oil immediately prior to physical refining. 2. The established possibility exists of the fractionation of micelle mixtures either by selective formation or by the destruction of particular kinds of micelles. By adding alcohols to the miscella and thereby increasing the polarity of the solvent, this kind of selection can be steered in the desired direction. 3. Free fatty acids, as already mentioned, escape encapsulation in the phospholipid micelles and pass through the membrane, but if organic or inorganic bases are added to the miscella, the soaps then form a link with existing micelles to produce new ones of considerably enhanced powers of occlusion. Some natural polymers found in oils, such as shea-nut fat and rice-bran oil, behave in a very similar way with soaps. In all such cases, UF achieves an even higher degree of purification of the oil than if no base addition had been made. 4. Certain examples of the improved effect of UF brought about by the addition of alkali to the crude-oil miscella are striking. Table 1.3 shows the effect of a caustic-potash addition to crude rapeseed-oil miscella. A marked reduction in metals as well as free fatty acids occurs. The oil from the caustic-potash-treated miscella was given a 1.5% Tonsil ACCFF clay bleach and deodorized. The flavor of the final oil remained satisfactory for over 3 months, and its frying performance was exceptionally good. Table 1.4 illustrates a similar improvement for cottonseed oil. Table 1.5 shows how the persistent green color of grapeseed oil, which causes severe problems in conventional refining, is substantially diminished when aqueous ammonia is added to it prior to UF. Notably, the ammonia soaps left in the retentate are easily decomposed by heat, and the NH3 obtained can be recycled for repeated use. Sen Gupta also lists some of the semipermeable membranes commercially available in countries such as the United States, France, Germany, Japan, the Netherlands, and the United Kingdom; he mentions the membranes used in trials recorded in Tables 1-C, 1-D, and 1-E. He adds that many more are constantly being developed in this very active field. Koseoglu et al. (Koseoglu, 1987; Koseoglu & Engelgou, 1990; Koseoglu et al., 1989, 1990) reported the results of employing a variety of membranes in edible-oil processing in the United States, where potential

32

H.B.W. Patterson

TABLE 1.3. Analysis of Rapeseed Oil before and after UF of Crude- Oil Miscella, with and without the Prior Addition of Caustic Potash P (ppm) Crude oil

FFA (%)

294 1.3

K (ppm)

Fe (ppm)

Cu (ppm)

S (ppm)

Lovibond 2”

39

3.2

0.3

19

8.2R + 80Y + 5.1B

UF without KOH addition

7 1.3

2

0.13

0.04

9

6.0R + 70Y + 1.2B

UF with KOH addition

3 0.03

0.7

0.01

0.01

4

4.2R + 50Y + 0B

TABLE 1.4 Analysis of Cottonseed Oil before and after UF of Crude-Oil Miscella, with and without the Prior Addition of Caustic Potash P (ppm)

K (ppm)

FFA (%)

Gossypol (%)

Lovibond 1”

E1% 1 cm 232/268 nm

Crude oil

630

210

6.8

0.79

70Y + 20R + 0.8Ba

24.8/7.3

UF without KOH addition

4

1.5

6.5

0.40

15.0/5.0 30Y + 6Ra 70Y + 60R + 1B

UF with KOH addition

2

0.8

0.2

<0.01

20Y + 4R

2.6/0.8

a

This set of readings taken in 1/8” cell.

TABLE 1.5. Analysis of Grapeseed Oil before and after UF of Crude-Oil Miscella, with and without the Prior Addition of Aqueous Ammonia Fe (ppm)

P (ppm)

Crude oil

4.0

FFA (%)

Green pigments (ppm) 57.6

21.7

65

UF without NH4OH addition

3.6

47.6

0.3

5

UF with 30% addition

0.5

16.8

0.4

5

cost-savings are very great. The interesting suggestion is made that a feed and bleed UF system (Fig. 1.11) could be combined with a distillation unit so that the oil-rich permeate is fed to the still for the removal of hexane while the solvent-rich retentate is fed back to the extraction unit along with hexane recovered from the still. Using this method, a 43% decrease in the amount of miscella needing to be distilled is indicated in trials. This is the most appropriate point at which to describe briefly two other technologies using a membrane as a filter, although their applications in industrial chemistry are more remote at present than in medicine. We may have a mixture of gases, or liquids, or even more than one kind of small molecule as a solute in some solution. A membrane or some other separating medium may then be chosen in which one molecule (gas, liquid, or solute) will dissolve and diffuse while another will not.

Basic Components and Procedures

33

A separation then becomes possible based on solution and diffusion, and not on the penetration of very fine passages in a rigid, solid-filter medium. The membrane is, therefore, to be regarded as an isotropic solvent for one or other of the components of the mixture or solution; it is not porous in the proper sense and it usually is an organic polymer. Concentration of the solute builds up from the feed side of the membrane according to partition coefficient, concentration gradient, and so forth. Lastly, the migrating solute escapes from the downstream side of the membrane into a solvent, or it is lost by evaporation (Meares, 1986). These membranes—which are able to separate molecules primarily because of the chemical nature and not solely on account of their differences in size—recall the biochemical operations associated with the transport of substances in living tissue. The migration of solute molecules across a thin layer of liquid which separates two otherwise miscible solvents is also possible. The intermediate layer or membrane exercises the selection process, and is immiscible with both feed and receiving phases on either side. Again, the solution–diffusion mechanism governs movement, and partition coefficient and concentration gradient play a part. The exact mechanism of transport—especially where the viscosity of the liquid membrane in one example may be markedly different from that in another example—is still the subject of research and speculation (Del Cerro & Boey, 1988). Two arrangements are possible in the application of liquid membranes. In one, the liquid membrane is held within a solid microporous inert support by capillary and surface forces, and is described as supported liquid membrane (SLM). The supports are similar to those used in conventional membrane processes; they may be hollow fibers or spirally-wound types. As much as 1000 m2/m3 of mass-transfer surface can be realized by these means. The other arrangement is the emulsion liquid membrane (ELM). In this case, the solution is emulsified in the immiscible membrane liquid by a high-speed stirrer; the membrane liquid is external to the solution droplets and, if helpful, it may contain some surfactant to facilitate emulsification. The emulsion is then dispersed in the continuous phase either by flowing it through a column contactor or by using a low-shear mechanical mixer. Figure 1.12 illustrates the situation at this point. Large interface areas for mass transfer up to 3000 m2/m3 are possible. When the globules of emulsion separate and coalesce, they are collected, and the emulsion is broken to free the internal-phase droplets. Extraction may be arranged to proceed in either direction across the membrane. In the current early stage of this technology, various operational procedures remain to be optimized (Del Cerro & Boey, 1988). Biochemical methods, such as the use of enzymes, were employed empirically for centuries, and are now being actively explored for the replacement of older methods of chemical processing. This development arose from the closer observation of biological methods. Supposedly feasible, in the same way, membrane technology now widely found in nature may also come to play an important part in fats-andoils processing as well as in other industries. However, to predict how rapidly this development will occur is difficult.

34

H.B.W. Patterson

Fig. 1.12. Emulsion liquid membrane.

Neutralization and Washing Neutralization, whether by alkali or steam stripping, batch or continuous, is part of the total refining process, as is deodorization, and is not the direct concern of this text. Truly, however, in the course of removing free fatty acid as soapstock and then washing the oil, some pigments, gums, and other components are carried away, either in solution, adsorbed on soap particles, or occluded like oil droplets in the aqueous soap emulsion. Gossypol is the outstanding example of the removal of pigment by alkali washing (Cavanagh, 1951), the cottonseed-oil soapstock being black; similarly, obviously, from the strong color of palm-oil soapstock, carotene is being removed; the repeated washing of neutralized fish oils shows a progressive lightening of the oil, and the washes themselves become paler in turn. After draining away soapstock, commonly, a very dilute caustic-soda solution is more efficient in washing than water, which one can use as a second wash. Purification of the fish oil may now be so adequate that after drying it to ca. 0.05% of moisture, it may be sent forward for hydrogenation without the usual activated clay bleaching. In any case, this treatment will follow hydrogenation before deodorization. Again, if the fish oil has experienced a light phosphoric-acid treatment (ca. 0.07% w/w) prior to neutralization, its chances of a satisfactory transit from neutralization/washing/drying directly to hydrogenation are almost certainly improved. Provided quality standards are maintained in the final product, it becomes a question of balancing a possible moderate increase in nickel usage and hydrogenation time against savings in clay, oil loss in clay, bleaching, and filtration time. Precisely, this same form of reasoning can be employed at later or downstream stages of processing. Specifically, the processor should consider whether it is better to destroy

Basic Components and Procedures

35

primary oxidation products (such as peroxides) and adsorb or destroy secondary oxidation products (such as aldehydes and ketones) with activated clay at the bleaching stage rather than rely on hotter and longer deodorization. The overall effect is the deciding factor. Seemingly, the processor must put his money where the customer’s mouth is. Typical alkali neutralization of animal fats (Patterson, 1975, 1976) and palm oil (Swoboda, 1985; Willems & Padley, 1985) was already published in some detail. Thermolabile pigments such as carotene are very substantially diminished in distillative neutralization, the decolorization taking place quite rapidly above 220°C. The fate of other minor components during different methods of neutralization is mentioned earlier in this chapter as they are considered individually (see sections on Pigments – Settling). Worth noting is that today’s much more efficient methods of degumming, especially of soybean oil, precede neutralization and improve its efficiency. In the 1960s, batch alkali neutralization of partly degummed oil was followed, after the removal of soapstock, by boiling 20–30 minutes with a solution of sodium carbonate or sodium carbonate plus sodium silicate to hydrolyze and flocculate nearly all the remaining phosphatides, which would then be drained off with the further loss of some neutral oil. Where oilseeds were damaged by oxidation and hydrolysis during prolonged unfavorable storage, Andersen (1962) notes that the interaction between protein and carbohydrate as well as the production of oxy-fatty acids result in a darker oil. Also, the higher temperature experienced in pressing leads to the oxidation of the oil and the extraction of some color from seeds and coats. This becomes more obvious on second pressings. The type of solvent used in extraction can also influence oil color. In general, the type of discoloration arising from damage of this sort is substantially removed in alkali neutralization and washing so the task of clay bleaching is made easier.

Clay Adsorption The selective removal of pigments from oils and fats by adsorption on clay or carbon, specially chosen and activated for specific effects, is the source of the description— “bleaching”. As explained in the Introduction section, for decades the technology was understood to cover many more effects than mere decoloration. Thus, in the 1980s the use of specially prepared silica for the purification of fats and oils by direct addition or by the treatment of miscella gained prominence. This section and those immediately following are intended to enumerate the different facets of adsorption, which are then examined in more depth in later chapters (Baldwin, 1949; Rich, 1964, 1967, 1970). Effective adsorption requires a large surface; for practical reasons, the high specific surface (m2/g) of a very porous solid is used. The channels by which molecules reach this surface must be negotiable by the molecules concerned. The nature of the surface must allow acceptably firm bonds, chemical or physical, between it and the adsorbate. The situation is not precisely that of an actively supported nickel catalyst, where hydrogen and unsaturated oil or other molecules all reach the active surface,

36

H.B.W. Patterson

and then, after reacting, escape back into the bulk oil to be replaced at once by other unsaturated molecules. True, of course, is that in the extremity of a long and narrow (20–40 Å width) pore, movement will be sluggish, and the concentration of completely hydrogenated molecules will approach a maximum (Patterson, 1983). In the case of bleaching or purification by adsorption, pigments or other components are selectively retained on the pore surface, and the triglyceride escapes. Gradually, the concentration of pigment on the available surface of the adsorbent and the concentration remaining in the oil come into balance, so further exchange is negligible. The best temperature for the oil/clay contact must be chosen, as must the duration of the contact; an excess of either will encourage undesirable side effects. A helpful action is to remove from the oil any material like gum or soap which will compete for room on the clay surface. This leaves the surface much freer to work on remaining traces of gum or soaps as well as on an adsorbing pigment. We must recognize chemical effects which activated clay may promote, such as the destruction of peroxides and hydroperoxides with a consequent change in the system of unsaturated bonds in fatty acid chains (see section Oxidized Fats and Oxidized Fatty Acids). Although air may destroy some pigment, it may fix or produce others as well as encourage flavor instability; it is therefore excluded by vacuum or inert gas, particularly where edible oil is concerned. A small amount of water helps the effect of adsorption. This amount may be as little as 0.1% w/w on oil, sometimes rather higher (Anon., 2007); the oil must not be so wet as to hinder the dispersal of the clay. Rich (1954) reports that 1% w/w of H2O added with active clay enhances the bleaching effect on tallow, but with natural earth a slight negative effect is likely. The interesting point is made by Sen Gupta (1988) that anhydrous soaps form spherical inverted micelles (see section Purifying by Reverse Osmosis) in oil with their polar group turned inside, and are, therefore, less easily adsorbed on the surface of bleaching clay. On the other hand, wet soaps with a water content of 0.2–0.4% in oil form laminar micelles with polar groups turned outward and are much more easily adsorbed. Further, these laminar micelles are frequently attached to other undesirable polar components, and, therefore, the latter are also removed along with the soap. Remember that normal clay contains around 10% w/w of moisture in temperate conditions, but this may rise to around 20% in the humid conditions of a monsoon. A small addition of a complexing or sequestering agent such as a citric-acid solution made to the oil before clay treatment not only splits remaining traces of soap, but also enhances the clay’s adsorption of trace metals which, if allowed to remain, would promote oxidative flavor instability. At least in the case of soybean oil, but presumably for many others, bleaching that reaches 0.5 PV or just below does not indicate the certain attainment of a low value for secondary oxidation (Ariaansz et al., 1989). The latter is commonly and empirically equated to anisidine value (AnV), while the so-called total oxidation (Totox) is shown as 2 PV + AnV (List et al., 1974). To reach the desired low value of AnV, a heavier dose of clay or a more active clay must be used.

Basic Components and Procedures

37

Having noted the importance of the degree of oxidation in relationship to the flavor stability of a particular species of oil, note also that flavor depends on what is being oxidized as well as how much oxidation has taken place. Some potent offflavors exist, and are said to have a low-threshold value; that is, they are detectable by the average person at a very small concentration (e.g., 10–9). On the other hand, a great many compounds are detectable only at a vastly greater concentration (highthreshold value). Still, most useful over a period of time is to relate empirically the results of chemical-oxidative tests with organoleptic findings for the same species of oil, and thus be able to offer a forecast of shelf life.

Use of Carbon The Egyptians used charred wood as early as 1500 B.C. for medicinal purposes and as a purifier in other connections. Similar early use was made of it in India. Charred defatted bone char was used for decolorizing sugar liquors around 1820 (Andersen, 1962). In 1855, Poll suggested its use for bleaching oils, and in 1899, Bornemann suggested charcoal for the same purpose (Feron, 1969). Early this century, R.V. Ostreyko (1902; Andersen, 1962) invented the activation of carbon. He showed firstly that a highly adsorbent product could be obtained by roasting charcoal or other carbonaceous raw material in steam, carbon dioxide, or other gas which made oxygen available, and secondly, that conditioning the raw material with reagents such as zinc chloride, alkali carbonates or others, prior to carbonization, also led to greatly enhanced activity. Regarded popularly as “physical” and “chemical” activation, respectively, these two processes provided the basis of the industry which developed rapidly at the time of World War I. Coconut shell or hard-fruit kernel were found to yield a charcoal very suitable for military respirators or other gas-adsorption duties. By the 1930s, other activated carbon was developed for the recovery of volatile organic solvents. Activated carbons now have a wide range of uses as purifiers, decolorizers, and catalysts. Employed as the minor portion of an earth/carbon mixture, they not only help to remove pigments from fats and oils, but also remove other unwanted nonfatty minor components. By the 1980s, the usage of activated carbon (powdered and granulated) in the purification of liquid-phase materials, including oils, had risen to 220,000 tons/year worldwide. This amounted to about 80% of the total. An additional 60,000 tons/year of granular carbon is used in gaseous-phase processing. The world’s largest usages are in Japan, the United States, and Europe. Depending on the starting material and the method of manufacture, a variety of types exists (Andersen, 1962; Anon., no date; Anon., no date; Bansal et al., 1988). Commonly, the internal surface range is 500–1500 m2/g, but much higher specific surfaces were achieved. Pore widths vary from under 2 nm to over 50 nm (see Chapter 3). Therefore, a very large adsorptive capacity can exist for components, such as pigment, which are desired to be removed from an oil. The following terminology was published by IUPAC in 1971 (Bansal et al., 1988):

38

H.B.W. Patterson

t


Macropores: widths exceeding 50 nm (0.05 µmL m or 500 Å)

t


Mesopores: widths 2–50 nm

t


Micropores: widths not above 2 nm

Worth noting is that the combined surface of mesopores, macropores, and the superficial or external particle surface is estimated to amount usually to no more than 10–200 m2/g (Bansal et al., 1988). This emphasizes the huge area within the micropores. Nevertheless, in the adsorption of larger molecules, such as those of pigments, the mesopores make an important contribution. If active carbon is being compared with active earth, one must state which type of each is being considered and for what task in what conditions. In general, what can happen is that in the case of a deeply colored oil (i.e., a high-pigment concentration), a well-activated carbon takes up more pigment, weight for weight, than an activated bleaching earth, but that at a low-pigment concentration the proportion of pigment adsorbed to weight of carbon used falls steeply, so as to become less than the corresponding figure for a good activated earth. Carbon is more expensive than activated earth; it also retains at least its own weight of fats, whereas a neutral earth may retain only 30% of its own weight and an active one 70% (Patterson, 1976). Further, a grade of carbon which acts efficiently as a decolorant on its own may be slow to filter. Some manufacturers produce certain of the activated carbons as coarser particles to overcome this difficulty. Scarcely surprising is that where carbon is considered advantageous, many refiners employ it blended with a larger proportion of the less costly activated earth. Such mixtures may be more efficient than carbon alone. A simple test program can be arranged in the laboratory where a set of earth/carbon mixtures is prepared in which one or the other predominates, and the decolorizing effect of a number of feasible dose levels is observed on the oil in question and in the conditions close to that of full-scale use. Figure 1.13 shows a typical result for such a program. The lowest residual color for any given dose level is easily seen. Possibly, a slightly higher final color is acceptable. In this case a mixture richer in earth can be chosen, and an economy in several directions realized. In practice, a 10–12% of a carbon blend would probably be chosen instead of the 40% of carbon which lies at the minimum in Fig. 1.13. Ready-made mixtures of this type recommended for use with particular fats and oils are commercially available (Anon., 2007) (see Chapter 3). Suppliers of activated carbon will probably be willing to perform tests for their customers. Where traces of mineral-oil contamination would give rise to the appearance of bloom on vegetable oil and fat products, a carbon treatment was found to be most effective. Other applications where carbon has yielded superior results are the removal of color from coconut and Palm Kernel oils, high-class tallows, and lard (Norris, 1964). It is an efficient adsorber of soap traces, and is free from the criticism that was at one time leveled against some bleaching earths—imparting a musty flavor and odor to oils treated by them. Also of practical importance is to recognize that activated carbon very readily adsorbs mucilage, gums, and so forth (Andersen,

Basic Components and Procedures

39

Fig. 1.13. Decolorization of oil by bleaching earth and carbon mixtures.

1962). Therefore, in any application, if one is relying on carbon to exert a final and vigorous removal of color or odor, the oil should already have been degummed. A particularly useful feature of superior grades of activated carbon (Chapter 3) is their capacity to adsorb polycyclic aromatic hydrocarbons (PAHs) from oil; activated earths are not similarly effective. Noteworthy is that while steam deodorization is capable of stripping out a substantial proportion of the so-called lighter PAHs (i.e., those whose molecules hold three or four rings), the heavy PAHs of five, six, and seven rings in the molecule persist throughout deodorization (see Fig. 1.14). It is, therefore, necessary to remove these heavy PAHs, some of which are known to be active carcinogens, by adsorbing them on activated carbon of established efficacy (e.g., Norit SA4) as part of the bleaching procedure, normally prior to deodorization. The occurrence of PAHs in foodstuffs, raw material for foodstuffs, and crude and refined edible oils has been the subject of wide investigations since the early 1960s (Biernoth, 1968; Biernoth & Rost, 1967, 1968; Grimmer & Hildebrandt,

40

H.B.W. Patterson

Fig. 1.14. Light- and heavy-polyaromatic hydrocarbons.

1967; Wendt, 1981). As far as edible fats and oils—vegetable and fish—are concerned, the field was reviewed in detail (Sagredos et al., 1988; Thomas, 1982) up to 1988. Evidently, in crude coconut oil extracted from smoke-dried copra as commonly obtained in the Philippines, the largest contamination with PAH occurs. Table 1.6 (Sagredos et al., 1988) shows the results to be expected from treating a typical Filipino coconut oil either by steam stripping at 200°C or by a combination of bleaching with activated

41

Basic Components and Procedures

TABLE 1.6. Removal of PAH from Crude Coconut Oil by Distillative Deacidification Compared with Active-Carbon Treatment Followed by Conventional Deodorization

PAH

Crude oil

PAH (ppb) after 200°C distillative deacidification

Anthracene

321

40

After 0.4% of active carbon plus active earth

Deodorization at 180°C

<0.1

<0.1

Phenanthrene

654

48

<0.1

<0.1

Fluoranthene

334

137

6.9

4.3

Pyrene

348

156

4.8

2.5

1,2-Benzanthracene

68

43

0.9

1.1

Chrysene

119

68

1..6

1.5

Total light PAH

1844

492

14.2

9.4

Benz(a)pyrene

33.1

32.5

0.1

0.1

Benz(e)pyrene

24.7

22.2

0.1

0.1

Perylene

4.7

4.4

<0.1

<0.1

Anthanthrene

3.5

0.3

<0.1

<0.1

1,12-Benzperylene

19.6

19.4

<0.1

<0.1

1,2,5,6-Dibenzanthracene

3.4

0.1

<0.1

<0.1

Coronene

7.7

0.1

<0.1

<0.1

Total heavy PAH

96.7

79.0

0.2

0.2

Combined total PAH

1940.7

571

14.4

9.6

carbon (0.4%) and activated earth followed by normal deodorization at 180°C. Table 1.7 gives the different levels of light and heavy PAHs found in various crude oils (119). For the more lightly contaminated crude oils (e.g., extracted from smokedried sunflower seed), a dose of 0.25% of an approved grade of activated carbon along with the use of activated earth is adequate (Wendt, 1981) when followed by deodorization at 180°C. Modern methods of drying, which avoid direct contact of the oil-bearing raw material with flue gases, lead to a greatly decreased contamination by PAHs (see also Chapter 4). In general, evidently, oils and fats ready for consumption are relatively lightly contaminated with PAHs, and the active-carbon treatment and deodorization are well able to keep them within the limits of a total PAH of 25 ppb and a heavy PAH of 5 ppb noted by Wendt (1981) (Sagredos et al., 1988). The situation regarding PAH in other foodstuffs (e.g., smoked meat, smoked fish, cereals, etc.) was reported in detail (Larsson, 1982, 1986). For example, Germany limits the overall amount of benz(a)pyrene (BaP) in smoked-meat products to 1 µmLg /k (ppb). Complete PAH analysis is used relatively (Sagredos et al., 1988), and for this

42

H.B.W. Patterson

TABLE 1.7. Comparative Levels of Contamination of Crude Oils with PAH PAH in µmLg/kg (ppb) Light Range

Heavy

Crude oil

Origin

Fish

South America

11–2383

486

Median

1–149

Range

29

Median

Fish

North America

3–23

13

1–6

3

Fish

Japan

17–1935

584

2–900

49

Fish

Scandinavia

38–109

85

58

6

Coconut

Philippines

234–4563

2264

15–130

79

Soybean

United States/Argentina

34–141

80

2–6

4

Sunflower

Eastern Bloc

8–118

39

1–6

3

Rapeseed

Europe

3–276

38

1–20

6

Palm kernel

Indonesia/Ivory Coast

6–81

28

0–10

4

Palm

Indonesia/Ivory Coast

2–42

15

0.7–10

3.5

Cotton

United States

40–59

50

2–6

4

reason, the total of all below BaP is classed as light PAH. Furthermore, in crude oils the proportion of total heavy PAH to BaP can vary from three to six times, thereby giving rise to parcels of appreciably different results. Accepting this limitation, the BaP value can be employed for the preliminary classification of crude fats and oils parcels. For refined oils, this proportionality is no longer valid since the various steps in refining have different effects on the individual PAHs (Sagredos et al., 1988). Although the usual doses of activated earth or carbon used in bleaching edible oils do not cause a significant loss of tocopherols, if carbon is substantially increased toward 2% w/w, much tocopherol may be adsorbed.

Use of Silica The point is made repeatedly in this chapter that the removal from the crude oil of material which would compete with pigment for space on the surface of bleaching clay or carbon is most beneficial and often more important than the removal of pigment itself. With the correct form of amorphous silica (Bogdanor, 1989; Bogdanor et al., 1989; Bogdanor & Toeneboehn, 1989; Morton & Griselli, 1990; Welsh et al., 1989), an agent was developed during the 1980s which is eminently effective in adsorbing soaps, phospholipids, and iron. The preliminary use of such silica clears the way for a greatly enhanced performance by activated bleaching clays, which then not only take up any remaining traces of these unwanted impurities, but also decompose peroxides and adsorb products of secondary oxidation, thus leading to higher Rancimat-stability test results. This silica, unlike the activated clay, is not an effective adsorbent of chlorophyll

Basic Components and Procedures

43

and its decomposition products. The simple addition of special silica just minutes prior to clay, then finally filtering both together, is referred to as dual addition. The continuous addition of silica to the continuous oil stream, followed by the passage of the silica-bearing oil first through a vacuum bleacher then through a filter already precoated with activated clay, gives the clay a still better opportunity to exert its maximal effect. This sequence is called modified bleaching. Finally, it is possible to omit the water-wash centrifuge in continuous refining, and transfer its task of final soap removal to an increased level of silica treatment. This is described as modified caustic refining. The economic choice depends on local circumstances (i.e., the shortage of water or a particular need to minimize liquid effluent). We shall return several times in this chapter and others to consider not simply the immediate direct effect of one adsorbent or another, but how to enhance their effects by using different sequences and modes of their employment, as well as combinations of the intrinsic advantages of the various adsorbents themselves. The above description of the different modes of silica usage provides an illustration of this. Economies in plant and labor utilization are available, as well as savings in material. Nickel catalyst which has lost a great deal of its hydrogenation activity still has a useful capacity for adsorbing soap, and this was exploited in the past by adding some old catalyst to an alkali-refined oil before adding new catalyst. This same approach was used in dealing with an oil which still contains sulfur (e.g., 20–30 ppm of S) after alkali refining. In this case, the nickel is actively sequestering sulfur compounds. Likely, both techniques have now become obsolete, but they demonstrate that the merits of a sequence of adsorptive effects have long been appreciated (Norris, 1979).

Chlorophyll Adsorption Not only does chlorophyll give rise to the need to use heavier doses of activated clay so as to remove it, but recently it was recognized as capable of promoting photooxidation and ultimately lowering flavor stability (Gunstone, 1984; Usuki et al., 1984). An agent which works particularly well in adsorbing chlorophyll and its common breakdown products (see section Chlorophylls) is, therefore, of special interest. Such an agent was developed and is reported (Davies et al., 1989) when acting on a superdegummed rapeseed oil of 20 ppm of phosphorous content to eliminate a concentration of 8000 ppb of chlorophyll. This is achieved by a dose of 1.25% w/w on oil. As might be expected, since phospholipids compete with chlorophyll for space on the surface of the adsorbent when substantial amounts of these are present (e.g., 250 ppm of phosphorus), they can easily double the necessary dose of the agent. To avoid this, a preliminary use of conventional synthetic silica was found to be helpful, but an even more valuable effect is secured by the preliminary use of a phospholipid-specific adsorbent from the same manufacturer. Other manufacturers have also developed chlorophyll adsorbents (Anon., 2007; Mag, 1989; Mizukalife; Taylor & Ungermann, 1984, 1987 ) (also see section Commercial Bleaching Clays in Chapter 3).

44

H.B.W. Patterson

Batch Bleaching This is the oldest and possibly still the most common procedure, especially in smaller plants, for removing pigment and other unwanted minor components from oils destined for edible or technical use. Traditional plants are well-known; some useful reminders covering design are given in Chapter 5. Regarding operating conditions, note the following: 1. While stirring under vacuum (ca. 50 mm Hg), the oil is dried to ca. 0.2% w/w of H2O at a temperature of 80°C. Some moisture hastens the bleaching process. At the same time, if the bleached oil is to be passed forward to a dead-end hydrogenation autoclave (Patterson, 1989 ), a final moisture content of not more than 0.05% of H2O is advantageous since this forestalls excessive water vapor arising early-on during hydrogenation in the gas space. The vacuum pipe becomes much cooler to the touch once most of the moisture has evaporated. 2. Clay normally is sucked into the bleacher by vacuum. Close the vacuum valve on the crown of the bleacher at this time to prevent clay from entering the vacuum pipe and eventually restricting it. The delivery pipe should dip some inches below oil level so as to trap incoming clay in the oil. If necessary, clay addition can be maintained by briefly closing the addition valve and restoring vacuum. The ideal is to hold the clay dose in a small vessel under inert gas at a reasonably high level with respect to the bleacher and with a relatively short, straight passage between the two. If the clay dose is being slurried in oil and then transferred, this may increase the chance of air contact unless precautions are taken. 3. If, for any reason, a second dose of the same or (usually) a second adsorbent is being added, an interval of 10 minutes allows the first to exert most of its effect, and this benefits the second. 4. The temperature is raised to 90–105°C while continuing agitation under vacuum. In particular cases, especially for technical products, a higher temperature may be specified. 5. A total clay/oil contact time of 20 minutes prior to filtration should be adequate; experience may show that one-half of this is feasible. 6. A filtration temperature of 90°C maximum is desirable, especially for polyunsaturated oils. Again, a closed system to point of delivery of the filtered oil is a worthwhile precaution to minimize oxidation. Vacuum may be broken by nitrogen, and the stirrer operated for a few seconds during filtration to minimize the amount of adsorbent settling onto the heating coils. 7. As was recognized for decades in many countries, in the short period that oil spends passing through the chamber of a filter press, a very useful degree of further adsorption of pigment (bleaching) and impurities (purifying) takes place. This is called the “press bleach” effect. One can calculate, for example, that by

Basic Components and Procedures

45

the time the first parcel of oil holding occurs,1% of clay is nearing the end of its filtration, the weight of earth held in a chamber represents a concentration of 30–40% on the oil held there, and as parcel follows parcel, this reaches 70– 80%. The contact time may also be around 5 minutes. Not surprisingly, such massive concentrations of clay can exert a considerable effect in some minutes. Taking a sample of slurry from the bleacher, filtering it through a Buchner funnel, and then comparing the color and clarity of the filtrate with what is issuing from the filter are most instructive, especially if this is done when a substantial cake has accumulated in the filter (Rich, 1964). 8. To exploit the several advantages arising from the press-bleach effect, many refiners reduced the amount of clay used in the second batch to filter on the same press, and progressively further reduced it on the third and subsequent batches until the press was filled. Depending always on experience, the clay dose on the second or third batches could easily be one-half of the initial standard dose. On any one batch of oil, of course, the final or cumulative color of the complete batch counts and not the color of the initial and final runnings. This procedure economizes in clay, oil lost in clay compared with what might have been the case, time and labor, and increases plant utilization. Many useful comments on bleaching practice were published (Andersen, 1962; Baldwin, 1949; King & Wharton, 1949; Norris, 1979; Rich, 1964, 1967, 1970; Singleton, 1956; Stout et al., 1949).

Multistage Procedures Two different but closely related considerations arise here. First, as appreciated for many years from theory, the most efficient color adsorption for a particular dose of adsorbent is achieved if the dose is applied in several successive portions with, ideally, a separation stage between each treatment. This is because adsorption continues until an equilibrium is reached between the concentration of pigment on the surface of the adsorbent and that remaining in the oil. Exposing the somewhat reduced concentration of pigment after the first treatment to a fresh adsorbent surface achieves yet another equilibrium, and so on as each portion is used. The expectation, then, is that the required final color can be reached with a reduced overall use of adsorbent, or that a lower color may be reached with the original total amount—all other factors remaining the same. When such procedures were explored in the past, the degree of oxidation arising from the necessary manipulation between stages very often gave rise either to increased color, fixation of existing color, and/or diminished flavor stability. The further point is that to follow such a procedure, substantial extra time and plant capacity become necessary. To have any hope of success, air contact must be excluded, and a more convenient handling routine must be found (Andersen, 1962; Baldwin, 1949; King & Wharton, 1949; Singleton, 1956). The second consideration is that a dose of cheaper adsorbent may remove a substantial amount of the more easily adsorbed components (soap, gums) and some

46

H.B.W. Patterson

color from the oil, and thus clear the way for a reduced dose of more expensive activated adsorbent to exert a greater effect. Thus, in the past, some refiners first introduced into the oil a cheap neutral clay and followed this, after a few minutes of contact elapsed, with some activated clay. For example, instead of a single treatment with 1% w/w of activated clay, a first step might be to add 0.4% of cheaper neutral clay; then 10 minutes later, add 0.6% of activated clay for the final bleaching while maintaining vacuum. Precise amounts need to be decided by experience. The advantages to be sought include a reduction in the cost of clay, a reduction of oil retained in clay, and an improvement in final color, stability, and final free fatty acid. Select the neutral clay so as not to introduce higher amounts of prooxidant metals. If a policy of decreasing clay dosage (as successive batches of oil commence to fill a press) is found effective, then possibly leaving the neutral clay dose unaltered and gradually reducing the expensive activated clay dose may be possible. This old technique of a two-stage treatment has gained a fresh relevance with the production of synthetic silica specially tailored to adsorb gums and soap (see section Use if Silica). The use of such material creates the usual advantages regarding the subsequent use of pigment adsorbers like activated clays. Further still, adsorbents particularly effective in taking up chlorophyll are now being made available (see section Chlorophyll Adsorption); these also give a rapid and improved performance when a phospholipid-specific adsorbent is first added to the oil. Because of these developments, refiners now have more readily within their grasp the means of offering to the final deodorization stage an oil more highly purified than ever before. Action in the bleaching field has become more lively.

Continuous, Countercurrent, and Fixed-bed Bleaching Methods Purification by adsorption is naturally complementary to the earlier steps of settling, degumming, and neutralizing. Like these, it is a stage through which the crude oil needs to pass in nearly all applications. The usual advantages of continuous processing have long been recognized: less bulky working units, more costly materials of construction (e.g., stainless steel) immediately feasible, less labor and space required, and tighter, consistent, and quickly responsive process control. Afterwards, the oils treated may pass individually to various modification processes such as hydrogenation, interesterification and fractionation, where it may be most convenient to handle batches, depending on the level of demand and the number of different product varieties. The continuous bleachers generally operate as a series of steps organized so that the semicontinuous processing procedure is able to accept feedstock and deliver product without interruption because of surge capacity. As always, the economic question is: When would additional effort and expense in the neutralization stage result in savings in bleaching or vice versa? No technological reason exists as to why, in continuous operation, advantage should not be taken of the press-bleach effect or two-stage adsorbent addition. One can program the automatic addition of clays and other purifiers to follow the situation at the filtration end just as the filters themselves are rotated in use. Plant manufacturers pay

Basic Components and Procedures

47

attention to the benefit of being able to change from one feedstock to another with a minimal cross-flow of oil and loss of time. The countercurrent procedure has the obvious advantage—partly exhausted adsorbent which can no longer take up pigment or other contaminants from partly treated oil is still able to adsorb substantial amounts from an unbleached oil; this is true of carbon as well as clay. This has, in fact, been put into practice, paying close attention to the importance of excluding air contact throughout (Baldwin, 1949; King & Wharton, 1949; Singleton, 1956; Singleton & McMichael, 1955). One can pass the neutralized, washed, and dried oil through filter cake before being dosed with fresh adsorbent, passed to the bleaching vessel, and then filtered on a second press. In due course, this second press, when filled, takes over the work of the first press, and a third empty press is then used for the last filtration. Rotation of these duties allows ample intervals for press cleaning. Although a substantial reduction in clay usage of 20–30% appears possible and security against oxidation is good, such systems have not gained the acceptance in the industry which they appear to deserve. Perhaps by increasing the costs of adsorbents, oil and solid-effluent disposal will force a revision of this attitude. Lastly, we come to the proposition of fixed-bed bleaching, which was considered and investigated intermittently over several years. As oil flowed through a massive concentration of adsorbent in the press chamber, or its equivalent in other filters, the leading face of adsorbent was the first to become saturated with adsorbate. Next followed an ever-moving zone of partial saturation, and at the downstream face emerged oil with nil or a minimal concentration of adsorbate. This situation persisted until the downstream zone began to lose its effect; the so-called spot concentration of adsorbate in the filtrate then rose. One may tolerate this situation for a limited period because a good quantity of filtered oil containing little or no adsorbate was already collected, and a specified limit of color, phospholipid, soap, or metal is set on the composite or overall composition of the whole filtrate. An acceptable filtration time is achieved by exposing many faces to the oil flow in the numerous chambers of the press. The proposition of using a battery of filter pads is not acceptable. However, an interesting alternative was advanced (Bogdanor, 1989). In modified bleaching, alkali-neutralized washed oil is led from the last centrifuge to a vessel where one or another form of amorphous silica is added continuously at a preset level to the oil flow. The flow is then passed through a vacuum bleacher during which time soap and phospholipid are adsorbed. Immediately following the bleacher is a filter precoated with activated clay. The clay, not hindered by the presence of gums and soap, rapidly exerts a powerful effect in adsorbing a pigment such as chlorophyll. For an oil such as soybean oil, a dosage of 0.12% w/w of silica on oil is suggested, and a precoat of 2.25 lbs of clay per square foot or the equivalent of 0.25% on oil filtered by the time the press is changed. As the total of solids used is reduced below what is needed if clay alone is used, the press remains in use considerably longer, the oil lost in press cake is smaller, and the disposal of press cake is easier. A final composite chlorophyll-A content of 50 ppb is easily achieved.

48

H.B.W. Patterson

This procedure combines convenience and the exclusion of air. The description of some bleachers is given in Chapter 5.

Heat Bleaching Caution is necessary both in the use of the term “heat bleaching” and the technique itself. In the precise first sense, only the removal of pigment by heat is intended, and not the influence of temperature on some other process such as hydrogenation or adsorption, which also happens to remove color. Secondly, experience shows that while sufficient heat can remove some color, it can promote the decomposition of other components present, such as gums, causing a darkening of the oil, or it can promote an interaction between components, which stabilizes color and subsequently makes it more difficult to remove. Lastly, and of particular importance to the processing of edible products, if oxidation is possible, heat will accelerate it; hence, flavor stability is damaged. In practice, the bleaching effect of heat is most notable where carotene and related carotenoids are present, as in the case of palm oil (see section Palm Oil in Chapter 4). Thus, in 20 minutes at 180°C, one-half of the carotene has disappeared, at 200°C about two-thirds, and at 240°C, only about 2% remains. The question was raised (Loncin, 1975) as to whether the thermal bleaching and deodorization of palm oil at temperatures which reach 270°C for a few minutes, or 220–240°C for somewhat longer, give rise to PAHs from the carotenoid pigments present. After a full investigation (Biernoth & Rost, 1968), this was shown to be quite a negligible effect, and is no more noticeable than with soy, corn, and groundnut oils (Biernoth & Rost, 1968; Rost, 1976). Traditionally, several procedures have been available for bleaching palm oil for edible use and others for technical use (see section Palm Oil in Chapter 4). The batch plant was steadily replaced by continuous units. For edible use, alkali neutralization and washing were followed by a batch bleach with activated clay up to 100°C, but up to 135°C is now envisaged (Willems & Padley, 1985). Suitably pretreated oil is physically refined at 260–270°C for less than 30 minutes (Willems & Padley, 1985). If the oil is going forward for technical purposes, the destruction of carotenoids with the catalytic assistance of acid-activated clay at about 150°C would be acceptable, but not for edible use. One procedure for edible use involved two batch-bleaching steps with an alkali neutralization preceding each, or at least the second one, after which the oil was deodorized. This double bleach is costly, and better procedures replaced it by physical refining. Chlorophylls are not nearly as thermolabile as the carotenoids. If, during processing, chlorophyll A loses its magnesium, a shift from intense green to the yellowgreen of a pheophytin occurs; heat causes the loss of a carbomethoxy group, yielding the distinctly yellow-green of a phyropheophytin (see section Chlorophylls). This change seems to underlie the loose suggestion of chlorophyll-containing oils losing color on deodorization. If an oil contains both carotenoids and chlorophylls, the temperature of deodorization destroys much of the red-yellow component, thus allowing the yellow-green of surviving chlorophylls to become more visible. Taylor and Ungermann (1987) noted that with both palm and soybean oils, adsorptive

Basic Components and Procedures

49

bleaching to lower the red colors encourages the heat-bleaching effect during deodorization to produce lower red colors in its turn; this effect is distinctly more noticeable in the case of soybean oil. At the same time, a deodorization heat bleach which achieves a low red color could mislead if oxidized components remain in the decolorized oil to undermine future flavor stability. This emphasizes the important function of adsorptive bleaching in the first place. The quality of the final oil, including the color, is strongly influenced by the original crude oil.

Air Bleaching Obviously, air bleaching cannot be applied in the production of edible fats since the fat as well as the pigment would be oxidized. In soapmaking an old procedure for decolorizing palm oil or tallow consisted of raising the temperature of the crude oil held in a vat or tank to about 120°C and then pumping air through a sparger mounted in the oil. The sparger consisted of a central feed pipe with several side arms upon the ends of which were located porous ceramic heads. A stream of fine air bubbles and then passes through the oil at all levels. The bleached palm oil was saponified shortly afterward. The resulting soap is a dull cream in color and has a characteristic odor—both features may be unimportant if masked by later additions (Andersen, 1962; Woollatt, 1985). Bleaching Effect of Light Intense light leads to the fading of many natural colors in manufactured products. It provides the energy for the catalytic initiation of photooxidation, which in turn yields products promoting the autooxidation of fats to establish itself; rancidity then results (Baldwin, 1949; Patterson, 1989). Processors and packers, therefore, have an interest in shielding their edible products from light, especially from ultraviolet radiation. Detergents, on the other hand, since the 1940s frequently include optical brighteners in their formulation. These have the ability to absorb ultraviolet light and then emit the energy so gained as visible radiation. This is optical brightening— not bleaching (Woollatt, 1985). Substituted ditriazinyl amino stilbene disulfonates have been popular as optical brighteners. In the illustrated example, R1 may equal R3, and R2 may equal R4. The R compounds can be aniline, diethanolamine, or a variety of other groups. Andersen (1962) describes how fats (such as palm oil, butter, and others) which contain some carotenoid pigments are bleached white by strong sunlight, and rapidly develop a rancid flavor. The bleaching and rancidity proceed inward from a solid surface by autooxidation. Beeswax and castor oil were bleached by exposing thin layers to sunlight.

50

H.B.W. Patterson

Steam Bleaching The expression “steam bleaching” may be met. It appears to be the procedure rather than a specific effect of the steam used. A good-quality palm oil is alkali-neutralized, washed, and dried. To a batch of the oil, a dose of 2.5% w/w of activated clay is added under vacuum, and during the next hour the temperature is raised from 90°C to 155°C with gentle agitation. During the next 45 minutes, open steam is drawn steadily through the oil held at about one-third of atmosphere pressure. After cooling, the oil is filtered. Physical refining would appear likely to replace such a procedure. SAFE Bleaching SAFE stands for sulfuric acid–Fuller’s earth. This old and relatively cheap treatment of darker or more highly colored fats and oils destined for technical outlets appears to rely on the sulfuric acid breaking up gummy and albuminous matter and then enhancing the action of the neutral clay in adsorbing pigment and other nonfatty components. The strength and amount of acid depend on the quality of the fat being treated and the extent to which the color must be improved. One procedure is to add up to 3% w/w of strong acid slowly to the oil with gentle stirring at ca. 40°C (Davidsohn et al., 1953); then after 30 minutes, add some 3–5% w/w of neutral clay. Next, the temperature is gradually raised to 100°C. Small additions of water at this point may improve the color as judged by filtered samples. The neutral clay soaks up the acid prior to filtration. Dark liver oils and poorer grades of tallow were processed in this way. Palm oil was bleached with activated clay or neutral clay plus more dilute acid (Andersen, 1962; Woollatt, 1985); in this case, the addition of small amounts of water later in the bleaching may progressively improve the color. Now that activated clays specially manufactured to improve lower-grade fats are available commercially, apparently, a worthwhile action would be to assess their cost-effectiveness and convenience before deciding to use the SAFE option. Chemical Bleaching In the great majority of chemical-bleaching processes, oxidation destroys the pigment. These processes are, therefore, not suitable for edible products since flavor stability would be damaged. For technical products such as soap, over many years, a wide variety of chemical procedures was employed on the fat before saponification and several on the soap itself. With the widespread use of soapless detergents since the 1950s, the chemical bleaching of soap-making fats declined in importance. What follows is a summary only, and for the working details, consult the references quoted. Hydrogen peroxide was used alone or in conjunction with other oxidizing agents such as sodium hypochlorite (Andersen, 1962; Davidsohn et al., 1953; Woollatt, 1985). Sodium perpyrophosphate was also mentioned as an example of the use of peroxy compounds. Benzoyl peroxide is said to be particularly effective

Basic Components and Procedures

51

with green-colored oils. Sodium or potassium chromate in conjunction with sulfuric or hydrochloric acid often was used to remove color, especially carotenoids. Na2Cr2O7 + 4 H2SO4 = Na2SO4 + Cr2(SO4)3 + 4 H2O + 30 Na2Cr2O7 + 8 HCl = 2 NaCl + Cr2Cl6 + 4 H2O + 30 also 2 HCl + O = H2O + Cl2 Potassium permanganate was used in the same way. Possibly the most popular chemical bleach has been the use of sodium chlorite, sodium hypochlorite, chlorine, or sodium chlorate (Andersen, 1962; Davidsohn et al., 1953; Norris, 1982; Woollatt, 1985), operating essentially through the action of chlorine dioxide. Thus, 5 NaClO2 + 2 H2SO4 = 4 ClO2 + 2 Na2SO4 + NaCl + 2 H2O (acid activation) 2 NaClO2 + Cl2 = ClO2 + 2 NaCl (chlorine activation) Bleaching by a reducing action was less popular, no doubt because of the risk of it being reversible. Although sodium bisulfite in acid solution (liberating SO2) was used directly on fats, the use of sodium hydrosulfite (blankite) added during soap-making is better known (Andersen, 1962; Davidsohn et al., 1953; Woollatt, 1985).

Hydrobleaching Hydrogenation is known to destroy or lighten the color of fats and oils. The unsaturated carotenoids are attacked by hydrogen, leading, in the case of palm oil, to a spectacular change from orange-red to pale yellow or off-white, depending on how far the hydrogenation is pursued. Significantly, the palm oil of the most intense red color responds best and produces the lightest colored fat. Evidently, this occurs because the carotene content is least damaged by oxidation. Adsorption of pigment by the siliceous support of the nickel catalyst and its destruction by heat no doubt assist the overall decolorization. In some oils the removal of orange-red pigment allows chlorophylls, if present, to become more easily visible, and suggests the oil has become greener. However, chlorophyll A (green) is changed to chlorophyll B (yellow-green) during hydrogenation (Norris, 1982), and further changes to chlorophyll-related compounds may ensue (see section Chlorophylls). A high degree of chlorophyll removal is best reached by specialized acid-activated earth or other adsorbents. If securing the greatest bleaching effect with a minimal alteration of the fat’s texture is important, then a light dose of around 0.05% of fresh active nickel/oil combined with an operating temperature of less than 150°C can be tried. The drop in iodine value can be kept extremely small in the case of a dead-end hydrogenation system by raising the hydrogen pressure in the headspace to 3–5 atm, closing the hydrogen valve, and then

52

H.B.W. Patterson

operating the stirrer until the pressure drops to about one atmosphere (gauge). One can repeat this operation easily; it is also useful in removing some off-flavors such as the rubbery flavor in some coconut oil, probably derived from smoke-dried copra. No change in melting point need arise. Such a procedure is known as flash, touch, or ultralight hydrogenation (Patterson, 1983). If a small drop in iodine value and some slight change in texture are acceptable, a dose of 0.5% of partly exhausted catalyst can substitute for the fresh catalyst mentioned above, and a higher temperature of 180–200°C can be used. Although other catalysts, notably copper, have the ability to remove color—and off-flavors—selectively, the problem remains of ridding the product of trace metals, particularly copper (Andersen, 1962; Patterson, 1983). In general, the disappearance of color during hydrogenation is valued as a bonus effect, and is not widely used specifically for bleaching.

Solvent Bleaching As would be expected, many organic solvents have a selective action not only toward the different triglycerides of a fat, but also toward its various minor components. Propane is a case in point (James, 1958). At lower temperatures, it is entirely miscible with fats, but near its critical temperature of 97°C, it is much less so; the saturated triglycerides are then the more soluble (so-called raffinate), and the unsaturated triglycerides, free fatty acids, and pigments are less soluble (so-called extract). Since propane at atmospheric pressure boils at -44°C, the propane extraction when performed on a commercial tallow (Solexol process) needs to be conducted under a pressure of several hundred pounds. The extract is then moderately richer in more unsaturated triglycerides but markedly richer in free fatty acid and pigment; the raffinate is richer in saturated fat and much poorer in free fatty acid and pigment. The volatile solvent is continuously recycled from both extract and raffinate fractions. The partition of components heavily depends on the polarity of the solvent; thus, the behavior of furfural is contrary to propane. The continuous purification of crude oil by the selective action of a volatile solvent and its subsequent recovery by distillation lost ground against other procedures due to cost, but are described here to illustrate a valid principle (Andersen, 1962; Feuge & Janssen, 1951; Moore, 1950; Norris, 1964, 1979; Patterson, 1983). The fractionation of triglycerides in refined oils with the aid of solvents is, however, commercially important (Thomns, 1985). As distinct from the action of solvent alone, the action of adsorbents is much-improved when they are made to operate on oil dissolved in solvent (Feuge & Janssen, 1951). The procedure lends itself to countercurrent flow of oil in hexane over packed-beds of adsorbent. One can recover continuously the solvent from the decolorized miscella. One can use a second solvent such as acetone to regenerate the adsorbent for many reuses. One can also recover the second solvent after each use. Again, the question is: Can that which is technically attractive prove to be commercially feasible in any given set of circumstances?

Chapter 2

Adsorption H.B.W. Patterson

Physical Adsorption and Chemisorption Adsorption is a phenomenon wherein the local concentration of a substance at the surface of a solid or liquid becomes greater than the concentration throughout the bulk. We thus have the well-known phenomenon of gas molecules concentrating on charcoal or pigments passing from solution in oil to deposit on clay. By contrast, absorption relates to the uniform penetration and dispersal of one item into another. For example, hydrogen rapidly penetrates and dissolves in palladium, light of a particular wavelength is taken up by a layer of liquid, and gases are selectively removed from an air stream by droplets of a scrubbing liquid. Thus, solute molecules which reduce surface tension will concentrate at an interface between a solid and a solution, and tend to be adsorbed on the solid. If, as happens in some cases, the solute increases surface tension, the concentration at the interface is diminished; this is known as negative adsorption. Water has a high surface tension; most solutes reduce this, so they are easily taken up by an adsorbent. Alcohol has a substantially lower surface tension than water, so most solutes are less able to lower it appreciably. Hence, they are not as readily adsorbed from an alcoholic solution. This thermodynamic aspect of adsorption is treated in physical chemistry textbooks under the heading “Gibbs adsorption equation.” When it comes to a question of the adsorption of gases on a solid, two types of forces are operating. Weak forces attracting the gas molecules to the solid surface are seen to be of the same kind as those holding the molecules together when the gas is in the liquid state. Notably, heats of adsorption and heats of condensation are broadly similar and relatively low. These weak forces are referred to as van der Waals forces, and promote a physical adsorption. They are mostly evident at low and moderately low temperatures. Not surprisingly, gases which are most easily adsorbed are also most easily liquefied. In other words, the volumes of gases adsorbed by a set weight of charcoal in moderate conditions such as 15°C and 1 atm are in the same order as their boiling points, with much more chlorine, hydrogen sulfide, or ammonia being adsorbed than carbon monoxide or oxygen. This weak physical adsorption is reversible at the same temperature simply by lowering the pressure. A second type of adsorption depends on the forces of chemical attraction between the surface and the surrounding gases or solute molecules in a surrounding liquid. This is called chemisorption. It is less common than physical adsorption; when it occurs, the heat of adsorption may be ten or more times greater. It is most evident at moderate temperatures, and to reverse it, a considerable rise in temperature is needed. If a mixture of hydrogen sulfide and oxygen passes over charcoal at room temperature, the hydrogen sulfide is adsorbed much more strongly. This is because 53

54

H.B.W. Patterson

the sulfur end of the H2S molecule is more negative than the hydrogen. That is, the molecule is polar, whereas the oxygen molecule is symmetrical and nonpolar. If, however, the temperature is markedly lowered, van der Waals forces come into play, and even nonpolar molecules are adsorbed. Oxygen is strongly adsorbed at 196°C. Furthermore, likely, molecules of a gas such as hydrogen may dissociate to atoms when adsorbed on the surface of metals such as nickel and copper. As is well-known, this condition facilitates their attack on the carbon-carbon double bonds in a surrounding liquid such as an unsaturated oil. As early as 1908, two mechanisms of adsorption operated together, one being electrochemical (chemisorption) and the other physical adsorption (Glasstone & Lewis, 1963). Seemingly, once chemisorption created a unimolecular layer on the surface available for that effect, the weaker van der Waals forces may make some further addition, depending in part on the concentration of pigment remaining in the solution. Various investigators approached the bleaching effect of activated clays as being partly an ion-exchange mechanism (Anon., Süd-Chemie). In the bleaching and purifying of fats and oils with adsorbents such as clays, carbon and special silica, van der Waals and chemical forces can play a part. This depends on the adsorbent, the nature of the minor component intended to be removed (the adsorbate), and the conditions of their contact. This brief exposition of the nature of adsorption should make clear that it is a complex phenomenon.

Adsorption Efficiency and Variation Chemical valence-type attraction falls off rapidly with distance, so it is likely to be responsible for only one layer of adsorbed molecules, whereas the van der Waals forces, though themselves weak, are considered capable of being exerted from the first layer to attract a further layer and so on, to several layers, especially if the temperature is fairly low and pressure rather high. In 1916, Langmuir advanced a mathematical relation which related the extent of the adsorption of a gas to the pressure for any constant temperature (the Langmuir isotherm). Assuming that the layer of adsorbed molecules does not exceed one molecule in thickness, an equilibrium is reached for any particular pressure between molecules evaporating from the surface (related to how much is already covered) and the molecules condensing onto it (related to the unoccupied space left available). At low pressures (lots of unoccupied space), the amount adsorbed per unit mass of adsorbent closely follows the pressure exerted. At the other extreme, as the surface becomes almost completely covered by a unimolecular layer, the amount adsorbed per unit mass approaches a constant limiting value at any temperature as pressure increases. At low temperatures (well below 0°C), this limit is approached gradually; at normal and higher temperatures (above 0°C), this limit is approached quite quickly. This unimolecular layer is a necessary condition for the fulfillment of the Langmuir isotherm. Many instances were found where it is followed closely; where deviations exist, one reason could be the intrusion of the weak van der Waals forces.

Adsorption

55

In the coverage of the surface which is intermediate between the two extremes just described, an empirical mathematical relationship provides a useful description for practical purposes of the relationship between the extent of adsorption per unit mass of adsorbent and the pressure of surrounding gas molecules or the concentration of solute molecules in a solvent. This is known as the Freundlich isotherm because, although not its inventor, Erwin Freundlich used it frequently to interpret his findings on the adsorption of solutes from solutions. This expression also applies to adsorption from colloidal dispersion. Although different textbooks vary slightly in their terminology, a popular expression of the Freundlich isotherm is

where x m c K

= = = =

amount of substance adsorbed (adsorbate) amount of adsorbent concentration of substance remaining unabsorbed a constant relating to the general capacity of that adsorbent for that adsorbate n = a constant relating to the manner in which the efficiency of adsorption changes as it progresses from a higher to lower color. Some authors write the value as 1/n, but regardless of the notation, the numerical value of the index ranges, in effect, between approximately 0.1 and 1.0.

It follows that log x/m = log K + n log C The implications of this relationship for practical purposes were clearly illustrated (Baldwin, 1949; Hassler & Hagberg, 1939) as shown in Fig. 2.1. Plotted on a log-log scale, the isotherms (within the limits for the test sequences) come close to forming a straight line, as was expected. By convention, the slope of any line is n and the intercept is K. The concentration of pigment is measured in Lovibond-color units. The value of the equation is independent of the particular color measurement chosen, as long as the results are additive. The following implications may be drawn: 1. Where K is larger, less of that adsorbent is needed to achieve the lower color in question. This means the adsorbent tolerates a larger amount of pigment being taken up in lowering the final color (residual unadsorbed pigment). Thus, for a final color of 1.0 in Fig. 2.1, the carbon–earth mixture is rather more efficient than the Fuller’s earth, but both are more efficient than carbon. 2. The index n relates to the rate of change in adsorption efficiency as bleaching progresses. Carbon tolerates a high load of pigment if the final residual color need not be low, but its efficiency drops rapidly as a low final color is sought. This

56

H.B.W. Patterson

Fig. 2.1. Bleaching of refined cottonseed oil (25Y-5,5R ).

suggests that a modest amount of carbon in a mixture with earth will effectively lower the color as bleaching commences, after which the earth is even better placed to continue the process and reach the required lower level. This means K for the mixture is raised because the whole isotherm is lifted by the early removal of color by the carbon component. 3. Obviously, from the previous statements that in selecting an adsorbent one must recognize over what range of removal of color—or other component—it is intended to operate. Equally, if two adsorbents are being compared, the comparison should be over the same range. Low values for n indicate that the slope of the isotherm is low and the performance does not change quickly across the range. Worth emphasizing is that in reaching these results, no gross competition from soap, gums, and so forth is being experienced and that comparisons are made at the same operating temperature, moisture content, absence of air, and so forth. Also, K

Adsorption

57

and n values hold good for a particular adsorbate–adsorbent (color–clay) system, and will show differences for other tasks and at other temperatures. If we look at the results obtained at extreme conditions where there is a departure from the theoretical Freundlich isotherm, the reasons for the divergence are not difficult to identify. For example, when the concentration of clay is already high and the color has become light, the amount of bleaching achieved by the use of an even greater clay dose falls rapidly. This means that the low concentration of pigment remaining in solution is able to be in equilibrium with a low concentration of pigment on the large dose of that particular clay. At the opposite extreme of the high loading of pigment onto the surface of a small clay dose, this loading could be rather higher if other minor components were not interfering by competing with the pigment for adsorption. As the dose is increased, this interference is overcome, and the color lightens as the dose increases in accordance with the Freundlich isotherm. Finally, the extreme of a very large clay dosage just described is reached. Evidence shows, at least in several instances, that adsorption from solution leads to the formation of a single layer of solute molecules on the surface of the solid adsorbent; this is closely analogous to the chemisorption of a gas as was mentioned earlier. We, therefore, should expect that an equation similar to the Langmuir isotherm should describe the effect of concentration on the extent of adsorption from solution. By replacing gas pressure by the concentration of solute in solution, we obtain the equivalent of the Langmuir isotherm (Achife & Ibemesi, 1989; Glasstone & Lewis, 1963). where C x m a b

= = = = =

concentration of substance remaining unabsorbed amount of substance adsorbed amount of adsorbent a constant giving a measure of the surface area of the adsorbent a constant giving a measure of the intensity of adsorption

Taking various predetermined amounts of rubber-seed oil and melon-seed oil in many bleaching tests and working with Fuller’s earth, activated carbon and a mixture (1:1) of these at 30, 55, and 80°C, Achife and Ibemesi (1989) found good agreement with Freundlich and Langmuir isotherms as were previously expressed. Thus, the adsorption of the pigments proceeded by the formation of monolayers on the surface of the adsorbent. The following conclusions were drawn from the experiments: 1. The specific adsorption (x/m) increased with oil concentration (i.e., higher oil to adsorbent ratios). This was taken to signify an increase in collision frequency between molecules of pigment and adsorbent, but this effect reached a maximum and then declined, possibly due to oil molecules competing with pigment for adsorption. 2. For both oils and all three adsorbents, a clear increase in specific adsorption was noted with an increase in temperature within the moderate temperature range

58

H.B.W. Patterson

investigated. This was attributed to the activation of more adsorption sites arising from increased porosity and total pore volume at the higher temperatures. This improvement depended mostly on the structure of the adsorbent, since it was clearly more pronounced with one adsorbent (Fuller’s earth) than another (activated carbon). The effect was also more obvious with the rubber-seed oil than with the melon-seed oil. 3. The value of 1/a decreased with a rise in temperature, again suggesting an increase in active adsorptive surface area as temperature increased. The adsorption of pure E-carotene from acetone solution by a variety of adsorbents was chosen as the means of studying adsorption phenomena because bleaching of palm oil has become a very large-scale operation in recent years (Khoo et al., 1979). Silica and kaolin were found to be of comparable activity, Fuller’s earth about ten times more active, and the activated bleaching clay Tonsil much more active than Fuller’s earth. With Tonsil and Fuller’s earth, the orange red of carotenoids with adsorption peaks at 478, 453, and 430 nm gave way to peaks at 360 and 330 nm, while the supernatant solution became green. This indicated more than physical adsorption had occurred. With silica, the peak at 453 nm decreased, but without any increase at 330 and 360 nm. This suggests only physical adsorption had occurred. Samples of the four adsorbents were each heated in air for 4 hours at 150, 400, and 600°C. Taking the activity of an unheated sample as one, the Tonsil showed activities of 1:3:9:5, the Fuller’s earth 1:1:7:7, the kaolin 1:1:2:2, and the silica 1:1:1:0.5. Since these adsorbents were calcined at ca. 500°C in the course of manufacture, not surprisingly, heating above that temperature could cause structural damage. The improvement at 400°C was seen as being principally due to driving off excess water and burning off organic matter to expose active sites. It apparently did not increase surface area. This interpretation raises the question of how excessive the excess moisture was in the untreated adsorbents, as a small amount is generally considered beneficial (Brimberg, 1981, 1982). The fact that heat pretreatment did not improve the silica at any stage was seen as an indication that its adsorptive effect was solely physical and did not involve chemisorption. Further experiments suggested that after an initial, very rapid physical adsorption of E-carotene from acetone solution onto the surface of Tonsil, a steady adsorption of more E-carotene proceeded as a first-order chemisorption reaction, and this increased in rate when performed at higher temperatures. Finally, as demonstrated, E-carotene can be destroyed by the presence of ferric ions, even when oxygen is rigorously excluded. Both Tonsil and Fuller’s earth gave positive results when tested on this oxygen-free basis.

Conditions Affecting Adsorption As already shown, the progress of adsorption depends, in the first place, on the characters of the adsorbent and the adsorbate. With some combinations, the interaction may be nil, and for others the result is highly favorable. Next, comes the important

Adsorption

59

question as to the effect of conditions in which the contact is made. At once it becomes evident that compromises have to be made between securing a favorable effect in one or more respects, such as removal of color, soap, gums and so forth, and doing damage to the oil itself or some minor component. Cost-effectiveness is another aspect which must always be kept in mind along with technical efficiency. The requirements can vary considerably in going from one oil to another (e.g., palm to soy oil, tallow to olive oil). What is best is to consider the general effect of each important factor in turn, and note where each interacts with another, favorably or otherwise.

Temperature Since several factors are at work in the phenomenon of adsorption, and more still if removal of pigment is the particular operation underway, not surprisingly, a balance must be struck between conflicting influences. We need to distinguish between the effect of a rise in temperature on the adsorption process itself and the possible destruction, formation or stabilization of color, which may be from the independent effect of heat itself (see sections Chlorophylls, Carotenoids, and Heat Bleaching in Chapter 1). For simple physical adsorption, van der Waals forces at a readily accessible surface with temperatures lower than 100°C probably favor the equilibrium condition in which a larger amount of pigment—or other adsorbate—is retained. As the temperature rises above 100°C, van der Waals forces are increasingly disrupted, and this may be only slightly compensated for by the viscosity of the oil falling and pores expanding so as to create a better access to the surface. As for chemisorption, an increase in temperature above 100°C helps to activate sites where it occurs. Where the chemical destruction of pigment is catalyzed by the adsorbent, as with acid-activated clays, this becomes an important factor as the temperature rises above 120°C. Other effects become noticeable above 95°C, especially with acid-activated clays: splitting of soap if present and some triglycerides to form free fatty acid is detectable. When the temperature is raised toward 160°C, the structural alteration of polyunsaturated fatty acids is likely (e.g., migration of double bonds). Finally, it is pertinent to recall (see section Miscellaneous Minor Components in Chapter 1) that when a good-quality crude whale oil was to be extensively polymerized for use as a component of wool-combing oil or a soft oil in soap manufacture, it would first be given a 2% activated-clay bleach and then raised to 300°C while a current of steam passed through the oil under reduced pressure, thus excluding air. In these circumstances, color improved and viscosity increased during polymerization. One can achieve the same effect with fish oil, but in this case a preliminary alkali neutralization, washing, and clay bleaching are needed if the color is to lighten rather than darken at 300°C. This clearly indicates the necessity of removing mucilage before the high-temperature treatment; other findings agree (Zschau, 1990). A very detailed investigation (Stout et al., 1949) of atmospheric bleaching based on neutralized cottonseed oil and neutralized soybean oil was reported in 1949. Those natural or so-called neutral bleaching clays examined achieved their maximal color removal at 118–132°C. The various commercial activated clays performed best in the

60

H.B.W. Patterson

narrow band of 100–106°C. The use of temperatures above 106°C in air served only to increase the well-known catalytic prooxidant action of the acid-activated clays; the net effect of this oxidation was an increasingly poorer color because the point had been reached where formation and stabilization of new color exceeded bleaching of the original color. Some adsorbents containing a particularly high proportion of silica needed to be heated to ca. 200°C to achieve their maximal effect of color removal, evidently because they exerted no catalytic prooxidant influence. This high temperature obviously stands in the way of their use for edible-oil air bleaching because of inevitable oxidation and damage to the fat itself. Certainly, for edible-oil bleaching, using adsorbents’ vacuum is the normal practice, and 95–105°C is also normal for optimal results. Another study (Rich, 1967) confirms this, and yet a further study (Rich, 1964) clearly shows that if color removal is the only criterion and the pigment present oxidizes readily without significant catalytic help, a neutral clay may prove moderately more effective than an activated one, provided new color is not formed or color-fixed. This would apply to an inferior tallow containing some carotenoid-type pigment but not to the removal of chlorophylls, which is more dependent on true adsorption.

Duration The discussion of what constitutes an adequate contact time between oil and adsorbent presupposes sufficient agitation. In the case of the conventional gate stirrer of batch bleachers, an rpm of ca. 40 suffices, or its equivalent in terms of sparger steam. Excessive agitation wastes power and may increase the incorporation of air. More than one-half of the color is likely to be adsorbed in the first 5 minutes, and after 15 minutes further improvement is slight. When we are primarily concerned with the adsorption of other minor components such as gums, soaps and so forth, the same holds true. An increase in temperature quickens initial adsorption; prolongation of contact in the bleacher beyond 30 minutes favors color reversion. As pointed out in the discussion of batch bleaching (see section Batch Bleaching in Chapter 1), contact is made with a massive concentration of clay in oil lasting for up to a few minutes in passing through the clay layers of a filter unit. Where oil is bleached as a solvent miscella, a rapid fall in color occurs during the first few minutes. For more information, refer to individual studies (Brimberg, 1981, 1982; Khoo et al., 1979; Patterson, 1976; Rich, 1964, 1967, 1970; Richardson, 1978; Singleton, 1956; Stout et al., 1949; Taylor & Ungermann, 1987). In several of these, one must note whether the author is talking about atmospheric or vacuum bleaching. Conditions of contact, time, and temperature for special adsorbent silicas are broadly the same as for clays. Dosage As explained (see section Adsorption Efficiency and Variation), the efficiency of a particular bleaching clay is related to the severity of the task given to it. A clay which readily takes up pigment when plenty is available may have to be used in very large doses if a very light final color is demanded. This means the relative efficiency of two

Adsorption

61

clays may even be reversed as a lower color is achieved. When it comes to the removal or destruction of components other than pigments, similar wide differences can exist; for example, a silica which is highly efficient in removing gums and soap can be useless for the removal of chlorophyll. An adsorbent should be chosen for factory use as a result of a practical test in the same circumstances and to the standard of its full-scale use. What may also happen is that having discovered the dose required to cope with a parcel of oil that is more difficult than usual to bleach, this becomes the routine factory practice for all oil of that species. Over months, this is very wasteful of clay, oil lost in clay, and plant and labor utilization. This neglect is the very opposite of the practice of making good use of the press-bleaching effect, in which dosage is decreased as the filter acquires an increasingly thick coating of adsorbent. The performance of three clays from the same manufacturer when bleaching soybean and canola oils is shown in Fig. 2.2 (Taylor & Ungermann, 1987). Comparable isotherms exist for the removal of red color from the same two oils. A more subtle point emerges when we consider the clay dosage in relation to future oxidative stability, not only for the bleached oil but also for the same oil after deodorization. Fig. 2.3 (Ariaansz et al., 1989) shows in the cases of both activated clay Gr 105 and the more highly activated Gr 160 that when bleaching the soybean oil of an original peroxide value of 6.9, the peroxide value had dropped well below 1.0

Fig. 2.2. Bleached-oil chlorophyll-adsorption isotherms in soybean and medium-chlorophyll canola.

62

H.B.W. Patterson

Fig. 2.3. Bleached-oil peroxide and anisidine values versus clay dosage.

by the time the anisidine value was in decline. This leads to the reasonable deduction that the catalytic decomposition of the peroxides is preferable to the adsorption of aldehydes at the active sites of the clay. Of particular interest is the observation that when the fall in color (Lovibond red or chlorophyll) is plotted against dosages of the same clays, clearly, the adsorption of pigments is preferred over the adsorption of aldehydes, a large part of which takes place after most of the pigments were removed. This leads to the important conclusion that at a dosage level where colors are removed to acceptable levels, the oil can still have a poor oxidative stability (i.e., high in aldehydes). The study was continued to the point where the total oxidation (Totox) value (2 peroxide value + anisidine value) of the bleached oil was compared with the clay dosage (upper Fig. 2.4) (Ariaansz et al., 1989); then the Totox value of the bleached–deodorized oil was compared with the clay dosage (lower Fig. 2.4) (Ariaansz et al., 1989). Plotting bleached–deodorized oil Totox against bleached-oil Totox gave a clear linear relationship (Taylor & Ungermann, 1987)—when the bleached Totox was 2.0, the deodorized was 1.0, and when the bleached Totox reached 12.0, the deodorized Totox became 6.0. If the policy is to produce deodorized oils of no more than Totox 3.0, the previously explained type of control suggests a way to achieve it. Less expected

Adsorption

63

Fig. 2.4. Bleached and bleached–deodorized Totox versus clay dosage.

was the result (Taylor & Ungermann, 1987) that if clay dosages were increased well above the level needed to obtain satisfactory color and oxidation values, the oxidative stability began to fall. These excessive doses of activated clay begin to diminish the tocopherol content of the oil to a substantial extent and, therefore, lower the natural resistance to oxidation of the oil. In the investigation (Ariaansz et al., 1989) of several crude soybean-oil feedstocks ranging from 2.0 to 10.0 in peroxide value, of which all had relatively low anisidine value, an increase of activated clay (Gr 160) from 0.2 to 2.0% of dosage was needed to ensure that the Totox of the final deodorized oil did not exceed 3.0. Although various soy-oil feedstocks were used in the studies just described, reasonably, the findings must apply in principle to many oils. Worth repeating is that the steps taken in the early stages of treatment to remove gums, soap, and so forth greatly assist the action of the adsorbent to be used later. Secondly, although oxidative stability (chemical test) is not the same as flavor stability (organoleptic assessment), if the former is satisfactory, this goes a long way toward ensuring the latter.

64

H.B.W. Patterson

Atmospheric and Nonatmospheric Bleaching From what was already written at the end of the discussion of temperature, in relation to particular features of bleaching in air (see section Temperature), it will be evident that this application is limited to certain nonedible products. The details are found in several references (Bogdanor & Toeneboehn, 1989; King & Wharton, 1949; Norris, 1982; Rich, 1967; Stout et al., 1949; Taylor & Ungermann, 1987; Zschau, 1990). The hazards to quality, arising from ample contact with air, from the extraction of the seed to the deodorization of the oil were reviewed in detail (Patterson, 1989). Some points merit emphasis: 1. By the end of the bleaching cycle, oils will have lost some of their natural antioxidant, and are then most vulnerable. 2. Activated earths are prooxidant catalysts as well as adsorbers of oxidized fat. 3. Air drawn into the oil along with clay and contact between hot filtering oil and air are sources of damage. 4. The polyunsaturated oils are most at risk to oxidation. The reaction between them and dissolved oxygen is extremely rapid (see section Oxidized Fats and Oxidized Fatty Acids in Chapter 1). 5. Reaction rates double for approximately each 10°C rise in temperature, so oil and dissolved oxygen will react about 40 times more quickly at 80°C than at 20°C. 6. Avoid entirely open filtration of edible oil after bleaching, and if oil is not blanketed by nitrogen during storage, this should be kept as short as possible, under eight hours. Unsaturated oils are obviously most at risk. Good filtration discipline will keep the clay content of filtered oil comfortably below 100 ppm.

Moisture Moisture in the bleaching operation may arise from three sources. First, it may arise from the moisture naturally present in the clay as it is used; secondly, from the oil as it is brought forward to the bleaching operation; and thirdly, it may be deliberately added in certain bleaching procedures of technical-grade fats (Brimberg, 1981; Zschau, 1990). Usually, activated clays as supplied in temperate climates are in equilibrium with their surroundings at around 10% w/w of free moisture. During storage in monsoon conditions, this may reach nearly 30%. Activated clays, natural clays, and carbon affected in this way are quite able to perform satisfactorily, although it may be prudent in calculating the dosage to make an allowance for the exceptionally high-moisture content of the adsorbent. Some manufacturers will state the optimal moisture range (e.g., 15–23% of H2O) for their products for given bleaching tasks (Taylor & Ungermann, 1987). A common practice has been to dry the oil to about 0.1% of H2O w/w before adding bleaching clay. If the oil is to go forward to a dead-end hydrogenation system, it is advantageous for it to enter the autoclave not above 0.05% of H2O. This low level

Adsorption

65

cuts down the volatiles in the headspace early in the hydrogenation. Since the clay addition to the oil will be on the order of 1% w/w, the moisture contributed by the clay is insignificant compared with what is appropriate for the oil. Oil which has just been alkali-neutralized, washed, and settled will hold ca. 0.6% of H2O. If, through operator error, adsorbent clay is added to oil whose moisture content is well above this level, the clay may form coarse aggregates. This necessitates prolonged drying before filtration is practical. One study (Baldwin, 1949) compared two procedures. In the first procedure, once the wash was separated from the oil, the clay was added immediately to the oil, and then drying and bleaching under vacuum proceeded simultaneously. In the second procedure, the oil was first dried under vacuum, then clay was added, and bleaching was completed under vacuum. No significant difference in color was found. Lastly, when low-grade fats are being bleached with a mixture of sulfuric acid and Fuller’s earth, some further small additions of water during the bleaching cycle (see section SAFE Bleaching in Chapter 1) may extend the bleaching action to a useful extent. Although clays may have been heated to 500°C in the course of manufacture, and therefore dehydrated, they will have had the opportunity to reestablish a moisture content before use. Indubitably, some water will be bound more firmly than the superficial moisture lost at 105°C, and this water is likely to play an important part in chemisorption. Water molecules are held between layers in the lamellar structure of bleaching clays (Rich 1964, 1967). If this water is removed by rigorous or sudden drying, a collapse of the layers is encouraged, and the bleaching efficiency is appreciably less because the amount of accessible active surface is less. A related point is that clay should be added to oil at a temperature below the boiling point of water at the pressure being employed; thus violent frothing can be avoided. The oil can then displace moisture from between the laminates in a steady fashion as the temperature is raised without the loss of bleaching capacity.

Acidity How acidity arises in bleaching clays, how it may be enhanced, and how this affects adsorption performance are questions intimately related to their structure and activation. The description of them is contained in Chapter 3. At this time, it is sufficient to say that natural bleaching clay contains a high proportion of complex aluminum silicate in which a number of other metals such as iron, magnesium, calcium, sodium, and potassium are present in varying degrees as impurities. Weathering produces acid-reacting salts such as: Fe2(SO4)3 and Al2(SO4)3. The leaching out of the iron or aluminum and the substitution of an H cation, in effect, produce a form of silicic acid, and this, in turn, is capable of donating a proton to the carbon-carbon double bonds present in pigments or other minor components. A carbonium (organic cation) forms. The result is an electrostatic attraction (chemisorption) which holds the adsorbate to the clay. Acidity is expressed as the titratable acid of a water extract of the clay or as a pH of a suspension of it in water. Notably, in some clays, a considerable buffering

66

H.B.W. Patterson

effect occurs on the pH from the salts present, so that the obvious inverse relationship between acidity and pH can be disturbed. Hence, two clays of the same acidity (mg KOH/g) may exhibit different pH and vice versa (Rich, 1967). Although the activation of clays is highly beneficial regarding increased bleaching performance, it carries with it certain penalties which must be considered. Such clay will also retain more oil after use. It will more readily split soap and triglyceride to form free fatty acid, it is a more efficient prooxidant catalyst, and it is more prone to encourage the isomerization of unsaturated oil. It is also more expensive. A bonus is its enhanced ability to trap trace metals (Baldwin, 1949; Patterson, 1976; Rich, 1964; Richardson, 1978; Taylor & Ungermann, 1987).

Sequestering To get rid of prooxidant trace metals in refining fats and oils is important; the most common are iron and copper. Iron at 0.1 ppm and copper at 0.01 ppm can lower the stability of a deodorized oil such as soy oil. Various acids readily combine with these trace metals forming chelate complexes, and this renders the metals inactive as catalysts. A citric-acid solution is most popular for this purpose, which is described as sequestering. Other acids which behave in a similar way are ascorbic (vitamin C), phosphoric, tartaric, and EDTA.

The complexes formed when trace metals are sequestered are very readily adsorbed by activated clays; hence, this provides a double security. A typical precaution is to add ca. 0.05% w/w of citric acid/oil as a 50% solution in water to an oil about to be bleached. Some sequesterants do no more than capture trace metals, but others, notably citric acid, will scavenge free oxygen, split soap, and convert nonhydratable phosphatides to a hydratable form.

Chemical Side Effects Already explained was that the acid condition of activated clays leads to their promoting oxidation, hence the need to bleach in vacuum. Furthermore (see section Oxidized Fats and Oxidized Fatty Acids in Chapter 1), activated clays promote the rapid change from hydroperoxy- to hydroxy- to conjugated unsaturation, and this was even adopted to compare the activity of one clay with another (Pardun et al.,

Adsorption

67

1968). Although more than 100 ppm of activated clay present in an oil as vulnerable as soybean oil after deodorization could exert a damaging prooxidant effect, good filtration discipline should easily keep the clay level well below this (see section Trace Metals in Chapter 1). Any isomeric changes which tend to occur in the absence of air at high deodorization temperatures (230–270°C) are promoted by the presence of activated clay. Neutral clay does not promote such isomerization. At normal bleaching temperatures (i.e., ca. 100°C), neither kind of clay nor carbon causes significant isomerization in the absence of air unless some fat was already oxidized.

Particle Size (Brimberg, 1981; Morgan et al., 1985; Patterson, 1976; Rich, 1970; Taylor & Jenkins, 1901) “Adsorption” is by definition a phenomenon related qualitatively to the kind of surface and quantitatively to the extent of that surface. In practice, other requirements are important and make themselves felt through costs. Textbooks give various illustrations of how a cube of an element such as carbon, nickel, or sulfur weighing about a gram and having a surface area of a few square centimeters, when divided into particles of a 5–10-nm width, then has a combined superficial area of around 100 m2/g or more. Particles of this extremely small size present severe filtration problems. The answer in the case of a nickel catalyst is to deposit these nickel crystallite particles throughout the pores of some inert support whose larger porous particles are ca. 10,000 nm and above in size. In the case of clays and carbon, the basic material is rendered extremely porous by the manufacturing process (Chapter 3), and then ground and sieved to a convenient range of particle sizes which no longer pose filtration hazards. Naturally, this process increases cost. In this way, commercial activated carbons of 1000 m2/g are produced. The atoms sitting on the surface and along the broken edges have unsatisfied valence bonds, and, therefore, these same atoms are most capable of exerting a chemisorptive effect. This is exploited for purposes of bleaching and other forms of purification. A high proportion of pores must be negotiable by adsorbate molecules. A further advantage is that filtration is acceptably rapid when a good proportion of medium-sized particles is present. The complete removal of small particles increases the filtration rate further, but also begins to reduce bleaching activity. This, in turn, increases the necessary dose; hence, a compromise gives the optimal result, which translates directly to the final cost. Having overcome the filtration problem by depending more on porosity than small particle size, a different problem is presented by the more porous solids retaining more oil after filtration was completed. Whereas a natural clay will retain about 30% of its own weight of oil, an activated clay will hold about 70% and activated carbon possibly 150%. Although a useful proportion of these retained oils can be recovered, the cost has to be recognized. Fortunately, for technical reasons explained earlier (see section Use of Carbon in Chapter 1), only a modest proportion of activated carbon is necessary in clay/carbon blends.

Chapter 3

Adsorbents Dennis Taylor

Introduction Mineral Types A vast amount of research on the nature and action of clays, carbons, and synthetic silicas was completed, especially in the 1980s, which gives us a much better understanding of them as adsorbents and catalysts. Information is given here for those who use them in the processing of fats and oils. For more detailed information on structure and uses, including those outside of the fats and oils field, consult the several references to excellent original papers. The natural, neutral, or nonactivated bleaching clays are derived from claymineral deposits, which are merely dried, milled, and sieved to obtain a desired range of particle sizes. Natural weathering over a very long period has rendered the original mineral partly porous, and has given it some power of adsorption for pigments, soap, and so forth. Such material came into early use to cleanse fat from sheepskins (i.e., fulling); therefore, in England it became known as Fuller’s earth. This class of material is particularly mild in action, not promoting chemical changes in triglycerides, showing little tendency to split soap, and generally responding only modestly when bleached with acid to make it a more active adsorbent. Throughout the world (with the exception of Antarctica), deposits of different forms of the clay mineral bentonite are found, usually near the surface and in seams up to a 15-feet thickness. The name “bentonite” is derived from the large deposits associated with Bentonite shale at Rock River, Montana. At least 85% of this material consists of forms of an aluminum silicate known as montmorillonite, which was identified as long ago as 1847 at Montmorillon, France. Basically, montmorillonite is characterized by its three-layer structure. It is composed of a sheet of silica (tetrahedral structure) lying above and below a central sheet of alumina (octahedral structure), hence the description of a 2:1 layered mineral. Electrochemical binding attracts the sheets to one another. Some early explanation of nomenclature is necessary as confusion can easily arise in the minds of nonmineralogists. Subdivisions of this material are of particular interest to organic chemists. First, a divalent magnesium ion (Mg+2) or ferrous ion (Fe+2) may substitute itself for a trivalent aluminum ion (Al+3) in the octahedral layer, and this gives rise to a weak disseminated electronegative charge throughout the layer. In montmorillonite in particular, about one magnesium ion appears for every five aluminum ions. Neutrality and overall stability are restored by positive mobile hydrated alkali cations such as Ca+2, Na+, or K+ inserting themselves between neighboring triple layers. Because the attraction between the electronegativelycharged triple layers and the interlayer positive cations is weak, the neighboring trilayer 69

70

D. Taylor

sheets are able to move slightly apart (expand), and the cations themselves are easily exchangeable by other available cations. This gives rise to the description of swelling bentonite, western (United States) bentonite, and so forth, the replacement cation commonly being sodium and the material having uses as a flocculating or clarification agent (Rich, 1964; Taylor & Ungermann, 1987; Taylor & Jenkins, ; Taylor et al., 1989 ). The activity of this form of bentonite changes little with acid leaching, and it is not the source of acid-activated bleaching earths. A much more important variant, calcium montmorillonite or non-swelling bentonite [southern (United States) bentonite, Texas bentonite, etc.], lends itself very readily to leaching with mineral acid (HCl or H2SO4); its power of adsorption and catalysis is thereby dramatically increased. An important detail to add is that the entire group of expandable 2:1-layered silicate minerals is classed as “smectite,” and this term is not confined to montmorillonite. Other well-known 2:1 layered clay minerals exist that are incapable of expansion, either because of an extremely high layer charge (mica, vermiculite) or no layer charge (talc, pyrophyllite). Acid-activated clays were first produced in Germany around 1905; then several other countries around the world followed—this extension is continuing today. A currently approximate world production of 700,000 tons per year [P. Crossley, Industrial Minerals, No. 408, 69 (2001)] must be seen against a world production of total bentonite-type minerals of 6 million tons per year. The acid-activated clays are used in cleaning mineral oils as well as vegetable and animal fats; they are also used as catalysts in promoting organic reactions such as polymerization and isomerization, and may take the form of fixed beds or powdered additions. Finally, they have a role as proton donors in developing color from microencapsulated colorless leuco bases in copy paper (Fahn & Fenderl, 1983; Taylor & Jenkins, 1901).

Acid Activation and Adsorption of Pigment As might be anticipated, individual deposits of nonswelling bentonites, although predominantly composed of calcium montmorillonite, show sufficient variation in character to make their acid treatment a matter of judgment by individual manufacturers (Morgan et al., 1985) around the world. For example, the surface area of one bentonite may increase easily from 100 m2/g to a peak of 540 m2/g, then fall back to around 300 m2/g if treatment continues. Other bentonites reach 280–320 m2/g, and then steadily fall. A swelling-type bentonite undergoing the same treatment only rises from about 30 m2/g to 140 m2/g and then slowly falls (Taylor & Jenkins), thus illustrating how unsuitable it is for activation. Briefly, then, how is acid activation performed and what are its results? The clay is separated from foreign material such as limestone at the mine site, and then transported to the activation plant, where it is crushed to convenient size and, when necessary, excess moisture is dried off. The granulated clay is then slurried with a predetermined proportion of water plus acid, usually HCl or H2SO4. The agitated slurry is then heated near its boiling point for a few hours to achieve the desired quality of product from that particular bentonite. Experience and sophisticated

Adsorbents

71

testing contribute to the effectiveness of the process regime. Next follows a washing of the clay held in a filter until the appropriate residual acid is reached, then follow drying, grinding, sizing, and packing. The proportion of acid used and the duration of the extraction are important factors in the success of the process. When an increased degree of extraction is necessary, this adds to the chemical and energy costs, increases cycle time, and reduces the yield of the finished product. Sizing to a particular distribution of particle sizes may also reduce yield and contribute to the finished cost. Nevertheless, the activated clays are so much more effective for many purposes than the neutral ones that many bleaching and purifying steps are rendered more economical, practical, and convenient by their use. The important physical and chemical changes brought about by activation are seen as taking place in a distinct sequence: 1. Aluminum, magnesium, and iron cations are leached by acid from the octahedral structure at the center of the triple layers. Gaps are created in the crystalline structure, confirmed by x-rays (Taylor & Jenkins, 1901), and a large increase occurs in internal surface area. 2. The liberated, more acidic cations (Al+3, Mg+2, and Fe+2) now replace the cations Ca+2, Na+, and K+. Now an increase occurs in the concentration of sites of strong surface acidity—not merely total surface acidity. Furthermore, an increase in activity relates particularly to the increase in the number of 50–200 Å pores, and not just pore volume of any sort (Swoboda, 1985; Taylor et al., 1989). As suggested, both of these features should be present simultaneously for high-quality performance. Lastly, from the very beginning of acid treatment, a pronounced falling-off of cation-exchange capacity occurs (Taylor & Jenkins). As mentioned earlier, the capacity for an easy exchange of interlayer mobile cations (Ca+2, Na+, and K+) is a feature of some clay minerals. Since acid leaching steadily removes these cations, naturally the number of sites open to such exchange falls. When pigment-bearing oil contacts activated clay, the cations at the strongly acidic sites are ready to donate a proton to the pigment molecules, which usually contain electrophilic bonds ready to accept it, and thus form a positively charged carbonium ion (i.e., organic cation). The pigment molecule is then held to the clay surface by electrostatic attraction (Fahn & Fenderl, 1983). Evidently, this whole interaction is only able to occur if the channels or pores leading to the active sites within the clay are easily able to allow the passage of pigment molecules. Although the interaction on the outer surface of the clay particles may be the same, its contribution to the whole is small. Also (Khoo et al., 1979), the chemisorption effect can lead to a reaction proceeding on the surface which radically alters the pigments: for example, the ferric ion (Fe+3) has the ability to convert orange-red carotenoid to a green color. As has been known for many years, acid activation steadily increases the surface area and the bleaching efficiency until a maximal value for both is reached, after which both decline. In many instances, the maxima are reached at virtually the same

72

D. Taylor

amount of acid treatment, but, in certain cases, activity (i.e., bleaching efficiency) continues to increase for an appreciable interval after the surface area has begun to fall. This simply means that the decreasing specific surface area is still able, for a period, to maintain an increasing intensity of activation. All this must enter into the calculations of the manufacturer. In connection with a specific surface area (m2/g), as already explained (see section Particle Size in Chapter 2), this increases as the particle size is made smaller; both filtration time and oil retention increase so that the manufacturer is obliged to strike a compromise for reasons of economy. This is further examined in Chapter 6. The manufacturer is faced with four main factors determining the success of the product. The quality of the raw clay as mined is the first and most obvious influence. Next comes the degree of acid activation and the recognition that one can pursue this beyond the optimum. Third, once the activated clay is washed free of surplus acid, it must be dried to the moisture level at which it operates best. Finally, the activated clay must not be ground so fine that it creates difficulty in filtration after use. Later in this chapter, several well-established commercial products are described. In view of what was learned concerning the structure of natural montmorillonite, which is preeminent in the manufacture of activated clays, several varieties of synthetic montmorillonite were produced for experimental comparison (Taylor et al., 1989). First, a pure montmorillonite was made in which all the metal atoms of the central octahedral layer were aluminum. Two varieties were produced: one with a low Si:Al ratio and the other one with a high Si:Al ratio. In addition, a so-called iron montmorillonite and a so-called nickel montmorillonite were prepared in which was a controlled substitution of Al by Fe or Ni, respectively. Several other minerals, natural and synthetic, were also included in the comparison program. Each selected material, natural or synthetic, was subjected to a preliminary or pilot acid treatment, with the duration of extraction varied on each until an optimum was found. This condition was then used, appropriate to each material, to prepare a larger amount so that bleaching tests on refined soy oil could be performed. The final report found that the activated natural-calcium montmorillonite had the best overall performance. However, the activated products of iron, low Si:Al, and high Si:Al montmorillonites were very close contenders in the removal of carotene, chlorophyll, trace metals, and phospholipids.

Commercial Bleaching Clays Most of the major manufacturers make products derived from natural (neutral) to heavily acid-activated montmorillonites; however, Oil-Dri uses a natural attapulgite/ montmorillonite blend to produce neutral and lightly acid-activated products. Particular adsorbents are offered which user experience shows are most likely to meet required oil specifications, not only regarding color, but also oxidative stability, trace metal content, and so forth. Ease of filtration and minimal oil retention are now universally recognized as unavoidable economic constraints. Manufacturers are very ready

73

Adsorbents

to discuss which of their products best suit their customers’ needs and, if necessary, to conduct tests. Some bleaching clays are produced whose principal application is outside the field of fats and oils. Significant changes have occurred in the ownership of many of the major bleachingclay manufacturers over the past decade. The Galleon brand is now manufactured and distributed by Ashapura Volclay Ltd. (Bhuj, India) through a joint venture between Ashapura, AMCOL, and Mizusawa. The Fulmont brand, for many years produced by Laporte Industries, Ltd., is (still) manufactured in Widnes, England, by Rockwood Additives, Ltd. (Princeton, NJ), and distributed by AMC (UK) Ltd. Rockwood is owned by KKR (the private equity firm). Engelhard Corporation, including its Filtrolbrand bleaching clays, was acquired by BASF (Florham Park, NJ) in June 2006; its bleaching clays are still manufactured in Jackson, MS.

AMC (UK) Ltd. AMC (UK) Ltd. (Wirral, Mereyside, UK) is the sole distributor for a wide range of high-quality “Fulmont” bleaching earths produced by Rockwood Additives, Ltd. A chart of Recommended Applications was taken directly from their Web site (www. amcuk.ltd.uk). According to the Web site, Fulmont bleaching earths possess moisture contents in the range of 8–12%, and densities in the range of 0.5–0.65 g/cc. More detailed information on typical properties of these products could not be obtained. Recommended Applications Bleaching demand Oil or fat type

Easy

Normal

Challenging

Difficult

Fulmont “AA”

Fulmont Premiere

Fulmont XMP-3

Fulmont XMP-4

Specialty oils, sunflower, safflower, oleins, cottonseed Soy, rape, palm, animal fats, coconut, hydrogenated fats, waxes, linseed, castor, palm stearine High-color oxidized crude oils (e.g., rape, palm, tallow, rice bran) Fulmont grade

Ashapura Volclay Ltd. Ashapura Volclay Ltd. (AVL) (Bhuj, India), an Indo–American–Japanese joint venture, was established to set up a state-of-the-art plant for manufacturing acidactivated bleaching earth in India. The joint venture is a technical collaboration with Mizusawa Industrial Chemicals Ltd., the largest producer of activated bleaching earth in Japan, and AMCOL International Corporation of USA, a worldwide pro-

74

D. Taylor

ducer of value-added bentonite products. Previously, Mizusawa’s GALLEON brand was manufactured only in Japan. The product is well-known internationally, and is widely used as a refining and bleaching agent for various vegetable oils, fatty acids, petroleum, lubricating oils, and waxes. The Recommended Applications (below) and the Typical Properties as reported in Table 3.1 were taken directly from the company Web site (www.avlgalleon.com). Recommended Applications Galleon V-2

High-quality special grade for difficult-to-bleach general fats and oils, in particular, castor oil, palm oil, soybean oil, rapeseed oil, linseed oil, Rice Bran oil, cottonseed oil, fish oil, and beef tallow.

Galleon V-2

Bleaching earth of the highest bleachability and filterability for difficultto-bleach oil fats and oils, such as castor oil, palm oil, linseed oil, rapeseed oil, fish oil, soybean oil, cottonseed oil, and olive oil. It exhibits high performance in both chemical and physical refining for palm oil.

TABLE 3.1 Ashapura Volclay Ltd./Bleaching-Clay Adsorbents Typical properties

V2

V2  Super

Moisture (%) (loss on drying at 110°C)

6.5

8.0

Fineness (%) (amount passed through standard sieve) 100 mesh

99.2

99.7

150 mesh

98.3

98.5

200 mesh

93.6

92.9

325 mesh

64.2

78.3

<5 µm (%)

15.8

10.9

Average particle size (µm)

22.2

29.3

Residual acidity (mg KOH/g) (according to extraction method)

1.1

1.1

pH (at 25°C) (5% suspension)

3.02

2.69

Bulk density (g/cc) (according to Iron Cylinder method)

0.55

0.53

Particle size (according to laser diffraction method)

2

Specific surface area (m /g) (according to BET method) Filtration rate (filtration period)

—

348

<20 min

<5 min

Adsorbents

75

BASF Catalysts LLC BASF Catalysts LLC (Florham Park, New Jersey, USA) (www.catalysts.basf.com) supplies bleaching clays or “adsorbents” of the montmorillonite type for applications to purify fats and oils for edible and inedible applications. These types of mineral adsorbents are also used for purification applications in petrochemical, paper, and other industries. The montmorillonite mineral-activation process includes the careful control of acid activation, washing, drying, and particle-size adjustment, resulting in products having high surface area, optimal porosity, bleaching activity, and filterability for most fats and oils bleaching applications. The BASF Catalysts LLC product line encompasses a range of particle sizes and surface-acidity levels to fit the adsorbent needs of the fats and oils industry. Table 3.2 provides basic data for several commonly used products. The following lists some of the products with recommendations for their use. Recommended Applications Grade F-1

General-purpose adsorbent for the decolorization of animal, vegetable, and mineral oils as well as waxes.

Grade F-2, Grade F-100

Neutral adsorbent for mild-decolorization applications, dimerization reactions, and other applications requiring slight surface acidity.

Grade F-20

Higher acidity adsorbent suitable for the decolorization of inedible fatty oils and fatty acids. Suitable as solid catalyst for acid-catalyzed reactions.

Grade F-105, Grade F-160

High-performance adsorbent for hard-to-bleach oil applications. Suitable for excellent chlorophyll and carotenoid removal in vegetable oils (soybean, canola, rapeseed, palm, and others).

Grade F-115FF, Grade F-105SF

High-performance adsorbent for hard-to-bleach and milder oil applications. Suitable for the decolorization of a variety of oils in applications requiring an intermediate level of filterability.

Nevergreen™

Specialty adsorbent suitable for maximal level of chlorophyll removal.

BleachAid™

Bifunctional specialty adsorbent having filter-aid and adsorbent properties. Suitable as a filter pre-coat to enhance packed-bed bleaching performance.

76

TABLE 3.2 BASF Catalysts LLC/Bleaching-Clay Adsorbents F-1

F-2, F-100

F-20

F-105, F-160

F-115 FF

F-105 SF

Nevergreen™

BleachAid™

% Free moisture (105°C)

15

14–15

16

15–16

16

16

15

16

Particle size % < 200 mesh

85

87–90

85

88–90

70

55

88

47

Acidity (mg KOH/g)

5

<1

20

4

4

4

4

4

pH

3

7

2

3

3

3

3

3

Apparent bulk density (g/cc)

0.6

0.8

0.5

0.7

0.6

0.6

0.7

0.5

Surface area (m2/g)

275

75

350

325

325

325

325

310

Filter rate*

15

10–20

15

12–14

32

48

11

90

* BASF Internal Test

D. Taylor

 Typical properties

Adsorbents

77

Oil-Dri Corporation of America Oil-Dri Corporation (Chicago, Illinois, USA) manufactures a full spectrum of bleaching sorbents developed for purifying oleochemical fluid products ranging from natural triglycerides to biodiesel esters. It supplies these products under the Pure-Flo®, Perform® (Table 3.3) and Select® labels (Tables 3.3–11). Additional information is available on their company Web site (www.oildri.com). Oil-Dri utilizes a unique, naturally-occurring blend of two adsorptive clay minerals, attapulgite and montmorillonite. As a result of innovative acid-modification technologies, key adsorptive properties are enhanced while maintaining the natural porosity that contributes to the excellent filtration characteristics of these products. Pure-Flo Product Line TABLE 3.3 Natural and Acid-Modified Bleaching Clay Specifically Developed to Manage Easy/ Normal-To-Bleach , Color-Sensitive Oils Product name

Suggested application

Comments

Pure-Flo B80

Palm, coconut, tallow, corn, olive, cottonseed, and other Oils

Natural adsorbent efficient at removing color bodies and metals. Choice for organic oils. Extremely fast filtering.

Pure-Flo Supreme B81

Palm, coconut, corn, olive, cottonseed, sunflower, safflower, and other oils

Efficient at bleaching low chlorophyll and color-sensitive oils, extremely fast filtering, and excellent at removing metals.

Pure-Flo Supreme Pro-Active

Soy, sunflower, corn, tallow, rapeseed, canola, cottonseed, safflower, and other oil

High-activity acid surface- modified adsorbent efficient at removing chlorophyll and color bodies. Extremely fast filtering.

Perform Product Line TABLE 3.4 Specifically Developed to Manage Tougher-to-Bleach, Chlorophyll-Sensitive Oils Product name

Suggested application

Comments

Perform 4000

Soy, rapeseed, canola, linseed, and other oils

High-activity acid-modified adsorbent, efficient at removing chlorophyll and color bodies; offers fast filtering characteristics and efficient metals adsorption.

Perform 5000

Soy, rapeseed, canola, linseed, and other oils

Higher activity for demanding oils; aggressively removes color bodies and chlorophyll while offering good metals adsorption and filtration characteristics.

Perform 6000

Soy, rapeseed, canola, linseed, and other oils

Most active for hard-to-bleach oils while maintaining good filtration characteristics.

78

D. Taylor

TABLE 3.5 Oil-Dri Corporation/Bleaching-Clay Adsorbents

Product

Mesh % -325

Acidity pH 5% slurry

% Free moisture

Densitypacked (lb/ ft3)

Densitypacked (kg/ m3)

Pure-Flo B80

85

<7.8

15.5

36–48

577– 769

Pure-Flo B81

77

4.2

15.5

36–48

577– 769

Pure-Flo Supreme Proactive

77

2.9

15.5

36–48

577– 769

Perform 4000

77

2.7

10.5

36–48

577– 769

Perform 5000

77

2.5

10.5

36–48

577– 769

Perform 6000

70

2.3

10.5

36–48

577– 769

Recommended Applications for Tonsil Bleaching Earths Supreme 110 FF Supreme 112 FF

Well-suited for the removal of polar compounds like chlorophyll, carotenoids, phospholipids, and peroxides, via chemisorption and acid catalysis. Especially recommended for hard-to-bleach oils and fats, or if a very low bleaching-earth consumption is desired. Supreme 112 FF has higher acidity value than 110 FF.

Optimum 210 FF, 214 FF, Standard 310 FF, 314 FF, 3141 FF

Possesses outstanding adsorptive capacity for polar compounds, like peroxides, chlorophyll, carotenoids, and phospholipids, via chemisorption and acid catalysis. Very suitable for refining vegetable and animal oils and fats. Slight differences in physicochemical properties between the various grades (see Table 3.F).

Standard 3191 FF

Mainly used for refining mineral oils and solvents, or for gentle refining and treatment of vegetable oils and hydrogenated fats.

Tonsil 416 FF

Manufactured by the acid activation of calcium bentonite; it has a neutral pH-value and outstanding adsorptive capacity for polar compounds like chlorophyll, carotenoids, phospholipids, peroxides. Very suitable for extremely mild refining of vegetable and animal fats and oils, and for finishing and/or reprocessing numerous types of mineral oils, paraffins, and waxes.

Terrana 510

Natural bleaching earth used in a wide scope of applications manufactured by the thermal activation of calcium bentonite. Mainly utilized for smooth refinery treatment of vegetable oils and hardened fats.

Sϋd-Chemie AG (Munich, Lenbachplatz 6, 80333 München, Germany) Important characteristics of various special grades of bleaching clays popular with edible-oil refiners are shown in Table 3.6. The suffix FF (fast filtration) indicates clays of particularly uniform grain-size distribution, ensuring the shortest possible filtra-

Adsorbents

79

tion times. Standard grades of the same activity are available (e.g., Tonsil Standard). Recommended Applications and data for Table 3.6 were taken from the company Web site (www.sud-chemie.com).

Powdered Activated Carbon In Chapter 1 a brief account was given of the earlier uses of charcoal as a purifying agent and how so-called physical and chemical procedures for its manufacture developed in the nineteenth century. A closer look at these procedures and their products will help us understand how powdered activated carbon (PAC) can be used to the greatest advantage in the fats and oils industry. As early as 1924, Mecklenberg (Mecklenberg, 1924 ) drew attention to the various properties which characterize activated carbon. Among the most important were the surface area, capillary volume per unit weight, cross-sectional area of various capillaries, particle size, chemical character of the surface, and the nature of the adsorbent itself. All activated carbons contain some micro-, meso-, and macropores (Chapter 1), but their proportions to one another vary substantially, depending on the kind of starting material and the production procedures. At one extreme are carbons prepared from coconut shell, which are likely to possess 95% of their available internal surface as micropores and are, therefore, ideally suited for the adsorption of small molecules (not pigments) and in circumstances where the concentration of contaminant is low. The ash content of such carbons is usually small. At the other extreme, wood- and peat-based carbons contain mainly meso/macropore structures. These are well-adapted to adsorb larger molecules (e.g., pigments); hence, such carbons are of particular interest for removing color bodies or other unwanted minor components in what we term “cleansing.” Carbons derived from a variety of coals show pore structures intermediate in type between the two mentioned. A striking feature in both wood and coal charcoals is the clear evidence of the influence of the pore structure of the original plant material. Even the characteristically higher proportion of small pores in carbons derived from anthracite may relate to different botanical structures present in contemporary wood and the plants which were the original precursors of the coal concerned. Naturally, processing methods are carefully chosen so as to conserve the highest proportion of advantageous pore structure, while taking into account the final yield of activated carbon from raw material and the economic duration of activation (Norit Co.; Speakman Carbons). The forms (granulated, powdered, extruded, and so forth) in which the products are presented to customers are described later; PAC for the treatment of edible fats and oils forms only a very minor part of the total market. Apart from the classic raw materials of wood, peat, coal and nut shell, a very wide range of carbonaceous raw materials was used in local circumstances in which they were inexpensive and the product was adequate for certain purposes. These less common raw materials range from waste tires to rice husks or coffee beans (Speakman Carbons). Yields can be as low as 10% of the weight of raw material; passage through a continuous retort or kiln can take up to almost half a day.

80

TABLE 3.6 Sϋd-Chemie AG/Bleaching-Clay Adsorbents Supreme 112 FF

Optimum grades*

Standard 314 FF

Standard 3191 FF

Apparent bulk density (g/L)

460

400

520–570

600

580

Free moisture (2 h/110°C) %

≈10

≈10

≈10

≈10

≈5

LOI (pre-dried, 2 h/1000°C) %

6.2

11.8

7.2–8.1

7.4

8.5

pH (10% suspension)

2.8

2–4

2.2–4.8

4.0–7.0

7–8.5

Acidity (mg KOH/g)

3.5

7–13

4.5–6.8

0.3

Chloride content (mg Cl/g)

0.5

0.1

0.5

0.5

Surface area (B.E.T.) (m /g)

285

180

165–220

171

230

Micropore volume

0– 80 nm (mL/g)

0.41

0.26–0.31

0.25

0.27

0–25 nm (mL/g

0.37

0.22–0.28

0.22

0.23

0–14 nm (mL/g)

2

<0.1

0.32

0.17–0.24

0.19

0.19

Particle- size

>150 µm %

6

5-6

5

1

distribution

>100 µm %

18

15–17

16

6

>63 µm %

28

28–30

30

20

>45 µm %

38

40–41

45

31

>25 µm %

58

59–62

62

54

* 210 FF, 214 FF, 310 FF, 3131 FF

36 65

D. Taylor

Supreme 110 FF

Typical properties Tonsil

TABLE 3.6 (CONT’D.) Sϋd-Chemie AG/Bleaching-Clay Adsorbents Standard 410 FF

Typical properties Tonsil

Standard 416 FF

Terrana 510

Apparent bulk density (g/L)

520

550

800

Free moisture (2 h/110°C) %

≈10

≈10

≈5

LOI (pre-dried, 2 h/1000°C) %

10–12

7–9

7–10

pH (10% suspension)

3.6

6.0

Acidity (mg KOH/g)

0.5

8.0 <0.3

0.4

0.6

<0.1

Surface area (B.E.T.) (m2/g)

220

250

60

0–80 nm (mL/g)

0.35

0.35

0.09

0–25 nm (mL/g

0.31

0.29

0.07

0–14 nm (mL/g)

0.27

0.27

0.06

Micropore volume

Particle-size

>150 µm %

6

9

distribution

>100 µm %

16

25

>63 µm %

30

39

>45 µm %

41

51

>25 µm %

60

70

Adsorbents

Chloride content (mg Cl/g)

81

82

D. Taylor

Activation Procedures Although macropores contribute only a small amount to the total internal surface, they are important in the sense that they lead to meso- and micropores, where the bulk of the adsorption occurs. As stated above, both raw materials used and the methods of manufacture determine the type and extent of the final porous surface. The wood, coal, or other carbonaceous raw material is heated according to a programmed temperature regime, the maximum not usually exceeding 800°C, in a stream of inert gas. A multiple hearth furnace, rotating kiln, or continuous vertical retort are types of equipment commonly used (Houghton & Wildman, 1971; Speakman Carbons). The first stage of carbonization drives off moisture up to 170°C. At slightly higher temperatures, carbon dioxide, carbon monoxide, and other volatile degradation products are evolved. During pyrolysis, tar, methanol, and other products appear around 270°C. The basic microstructure forms by 500°C, and by 600°C the softening process has given place to the production of a hardened, somewhat shrunken skeleton. The carbon content is now about 80%. Controling the rate of temperature rise is important— tarry matter can be readily volatilized without choking the pores of the char on the one hand, while not disrupting the porous mass on the other. In these circumstances, one can easily see why manufacturers must exercise a judicious choice of raw material and subsequent treatment, even in the early stages, to obtain a class of product best suited to their customers’ needs. This was emphasized by Bansal et al. (1988). In some cases the char from the first stage may be roasted for a further period at 1000°C in the absence of gas. At the end of carbonization or charring, much so-called disorganized carbon remains packed at random between molecular aromatic sheets. So-called physical activation is performed by the further heating of the char in the range of 800–1100°C in a flow of steam or carbon dioxide or in a mixture which includes air. The object of this activation is to burn off (oxidize) much of the disorganized carbon, and further develop pores by increasing their number and the diameter of a portion of them. The sheets of aromatic carbon are now exposed; active carbon atoms are located at their edges and at flaws in their structure. These in effect are regarded as having residual unsatisfied valencies. Such carbons may have an alkaline reaction. Alternatively, the activation of the char may be procured by what is termed “chemical activation.” The raw material is impregnated with the activating agent in solution, extruded, and pyrolyzed at 400–600°C in the absence of air. The cooled product is washed to release most of the agents for reuse. When dried and calcined (400–800°C), the agent promotes the dehydration of the char, leading to a very highly porous skeleton. In chemical activation, the carbonization and activation are regarded as a single uninterrupted progressive step. By the influence of the agent, the formation of tar, methanol, and so forth is inhibited; this improves the yield of activated carbon. Phosphoric acid is a common activating agent, but a long list of others is available (Bansal et al., 1988). Slight residual acidity is too tightly held to hydrolyze triglycerides. The dried product is ground to a useful size under 100 µ;

Adsorbents

83

the range 25–50 µ is popular. Thus, the particles have a large external surface, and diffusion distances for adsorbates become quite modest. They are, therefore, suitable for the cleansing of liquids, including oils. Surface areas of up to 4000 m2/g were claimed. Powdered, chemically activated carbon has a high adsorption capacity, and is popular for removing larger molecules, like pigments. An adsorption capacity of 0.6–0.8 cc/g is quite possible. Pore volumes’ range is 0.5–1.5 cc/g; for a true density of about 2.2 g/cc, apparent densities of 0.3–0.5 g/cc are found. Fig. 3.2 is an impression of the haphazard arrangement of fragmented sheets of carbon between whose slit-like crevices are found the micropores. At the same time, we must understand that the behavior of an activated carbon, especially one derived by chemical activation, is not completely defined by internal and external surface area and pore-size distribution, important though these may be. The precise steps taken by manufacturers to obtain particular activated carbons are commercial secrets. Since a considerable variety in the raw material chosen is available, the rate of heating, maximal temperature, duration of heating, and activating agents as well as further variation in secondary processes such as sizing, the complexity of the whole business lends itself to secrecy. The sequence in Fig. 3.3 gives only an overall impression of the wealth of options open to the manufacturer. While the specification of the product is vital in guiding the customer when making his or her selection, most important is that trials be performed under the conditions in which the selected carbon is to be employed so as to find the most cost-effective answer. A blend of activated earth and activated carbon (Fig. 3.3) is almost certain to be the answer. The ways in which carbonized char develops into a useful activated charcoal were the subject of experiments by international groups of researchers for decades (Bansal et al., 1988). Broadly, the activation following carbonization occurs in two stages. At first, what may be called disorganized carbon is burned off, clearing the mouths of pores; this does not exceed 10% of the total carbon present. As mentioned earlier, the subsequent development of the number and the width of pores is greatly influenced by the kind of gas flowing through the kiln—carbon dioxide, water vapor or air, together or singly being the popular choices. In experiments by Caron (Bansal et al., 1988), pure wood char was activated by steam at 950°C. As more and more carbon was burned off, the activation was monitored by taking samples at intervals and measuring the specific surface area as seen by nitrogen adsorption capacity BET (N2), the benzene index, the methylene blue index, and the molasses index (see Chapter 9). The BET (N2) index result comes close to measuring the effective surface area, including the great majority of micropores, while the other indices show what is happening when larger and larger molecules are involved. Two series of measurements were made. In the first, the test was based on the unit mass of carbon obtained at each stage, and in the second it was based on the unit mass of carbonized material at the beginning of activation. As the burn-off proceeded from 10 to 80%, the adsorptive capacity increased markedly in the first series for all tests, but whereas the BET (N2) result increased in a ratio of 1:2 and the benzene index by 1:2.5, the methylene blue index increased by 1:6, and the molasses index by 1:11. Clearly, as the molecular dimensions

84

D. Taylor

Fig. 3.2. Schematic representation of the microstructure of active carbons (Bansal et al., 1988).

concerned are N2 3–4 Å, benzene 5–6 Å, methylene blue 8–9 Å, and molasses 12–20 Å, the proportion of wider pores increased most. In the second series of tests based on what happens to a unit mass of original char over the same range of burn-off, the BET (N2) very rapidly gained a maximum around 12% of burn-off, and then from 35% of burn-off declined rapidly, reaching less than half its early value by the time 80% of burn-off was achieved. For benzene, the results were very similar, but for methylene blue and molasses a steady increase occurred up to 50–60% of burn-off, followed by an equally steady decline up to 80% of burn-off. The above series of results demonstrate that an initial increase exists in micropores which gives way to the formation of macropores, and then as the walls between cavities are themselves burned off, the adsorptive capacity of all types falls. Increased burn-off means lower yield, increased cycle time, and diminished utilization or productivity of the plant item. The producer must budget all this in comparison with possible improvement in quality. The parallel with the activation of bleaching earths by leaching with mineral acids is obvious. Tourkow et al. (1977) conducted a progressive study on the activation of two brown coals, carbonizing them at 900°C with three different gases—water vapor, carbon dioxide, and oxygen. As might be expected, the three experiments produced different distributions of porosity. At low burn-off, only micropores were produced, with oxygen giving the greatest volume. As burn-offs increased, differences

Adsorbents

85

Fig. 3.3. Manufacturing sequences in the production of activated carbons.

produced by the three gases became obvious. Water vapor progressively yielded all sizes of pores so that by 70% of burn-off a wide distribution of almost all pore sizes resulted, with a surface area of 920 m2/g and a total adsorption volume of 0.83 cc/g. Carbon dioxide predominantly produced micropores with a total-pore volume of only 0.49 cc/g, although the effective surface area was 900 m2/g. Oxygen showed a rapid micropore-type development of surface area and total-pore volume at the lower burn-offs (e.g., 25%), but by 70% it gave the lowest effective surface area (650 m2/g) and total adsorptive volume (0.45 m3/g). These poorer results are attributed to surface oxygen structures blocking the entrances of micropores and hindering further activation, as well as some burning of the outer parts.

86

D. Taylor

Forms of Activated Carbons (Bansal et al., 1988; Norit; Speakman Carbons) Powdered Activated Carbons Reference already was made to PACs in which the particle size was often reduced to around 25 µ. At this stage, the external surface makes some useful contribution to the adsorptive effect, diffusion distances within the particle are small, mass-transfer barriers are low, adsorptive action in use is rapid (5–30 min), and the filtration of such particles is not very difficult. They are most often used in combination with a greater proportion of activated earth as far as oil processing is concerned. Chemical activation of wood (sawdust) is a convenient method of manufacture. Granulated Carbons Granulated carbons are produced in the shape of small, irregular lumps which are ground and screened before or after activation. Employed in fixed beds, they are preferred for the adsorption of gases and vapors, the diffusion of these being faster than for liquids. Particles must not be small enough to allow entrainment in gas flow, and must be coarse enough in relation to the height of the adsorbent bed so as not to cause excessive pressure drop. Coal, petroleum-oil wastes, and rubber can serve as raw materials. Extruded Carbons Extruded carbons result from mixing powdered raw material with a binder such as pitch, extruding this as a fine cylinder and then carbonizing and activating it. Typical cylinder diameters are 1–4 mm, with lengths two to four times the diameter. This form also is used in fixed beds, so the resistance to the flow of the gas or liquor being purified is a vital consideration. In this application, periodic regeneration of activity is most worthy of consideration, either in situ or by return to the producer. Spherical Carbons Porous spheres of raw material are carbonized and activated to form spherical carbons. They are found to be very efficient in adsorbing gases such as sulfur dioxide and nitric oxide. Impregnated Carbons Several carbons exist in which the activated carbons were impregnated with finely divided metal or organic compounds. The need to counter the possible use of more sophisticated war gases accelerated research in this field during World War II, and now nuclear reactors are equipped with impregnated carbon for the adsorption of radioactive- iodine compounds. Of direct interest to the fat-hydrogenation industry is the purification of hydrogen generated from water and sodium amalgam, the latter being a by-product of the electrolysis of brine. Most contaminated mercury is thrown

Adsorbents

87

down by the act of compression; the hydrogen is then passed through a fixed-bed scrubber packed in succession with activated carbon/iodized carbon/activated carbon. This results in a mercury content of less than 10 µg/m3, which is much below the required safety limit (Patterson, 1983).

Commercial Powdered Activated Carbon Products The discussion which follows and the products mentioned refer to the treatment of edible oils, but in several cases they are also suitable for the processing of other liquors. For oils, the usual method is to unite the adsorptive action of the carbon with that of activated earth, usually by adding the PAC shortly after the earth during the same process step. Sometimes ready-made mixtures of earth and carbon which are the most cost-effective for their particular application are available. In general, storage and dosage systems allow for the separate handling of the carbon, a dust-free operation being a much valued facility. Small-scale trials are essential before deciding on the choice of adsorbents and their dose rates; test conditions must be close to those of full-scale use. Likely, suppliers will be willing to perform preliminary tests before advising their clients. In Chapter 4 suggestions are made as to what kind and dose of adsorbents should be used in treating about twenty of the more important fats and oils, and where appropriate, carbon is included. Powdered Activated Carbon in Processing Edible Oils Edible-oil purification utilizing activated carbon is typically done with PAC. Granular activated carbon is usually not applied because of operational constraints. The use of PAC for edible-oil refining is always in conjunction with bleaching earth. PAC dosing can occur at the same time as bleaching earth is added to the bleacher. For relative small-scale operations, ready-for-use bleaching-earth/activated-carbon mixtures are offered by established bleaching-earth manufacturers. Performance improvements may be obtained by dosing PAC after bleaching earth in a two-stage bleaching process whereby the majority of the carbon is added in the second stage. A two-stage bleaching process may be required when the spent bleaching earth and spent carbon have to follow separate disposal routes. t


The use of PAC for detoxification of crude fish oil is done without bleaching earth, usually following active silica pre-treatment. The purpose of the silica is for the selective removal of phospholipids from the crude oil as they are inhibitors for the activated carbon. The purpose of the carbon is for the selective removal of persistent organic pollutants (dioxins, polychlorinated biphenyls, etc.).

t


The operation is usually under vacuum to avoid the oxidation of the oil; nitrogen can also be used. In any case, the presence of oxygen during the adsorption process should be minimized.

88

D. Taylor

t


Contact time is 20–30 min for the decolorization/purification of vegetable oils and fish oils; for vegetable oils heavily contaminated with polyaromatic hydrocarbons (PAHs), a minimal contact time of 45 min is recommended.

t


Dosing rates of PAC: · For decolorization: 5–10% portion in a mixture with bleaching earth. Dosage for the adsorbent mixture is usually 0.5–1.5% (5–15 kg/ton). · For PAH removal: PAC dosing up to 0.5% (5 kg/ton) for coconut oils. Other PAH-contaminated oils (except olive pomace oil) usually require less to meet the regulated limits on PAH. The bleaching-earth dosage is at a minimum 1.5 times higher than of that of the PAC dosage. · For dioxin/PCB removal from fish oils: 0.1–0.3% (1–3 kg/ton)

t


Recommended temperatures for refining with activated carbon plus bleaching earth: from 80 to 100°C, based on the recommended bleaching-earth treatment for the corresponding type of oil. For crude fish-oil detoxification (without bleaching earth): 70–80°C.

t


Filtration of the adsorbent mixture is usually done with pressure-leaf filters, membrane presses, or pulse-tube filters in special cases where a minimum of filter aid is to be used.

Commercial Powdered Activated Carbon Companies AgriTecSorbents, LLC AgriTecSorbents , LLC (Stuttgart, Arkansas, USA) (www.agritecsorbents.com), which began commercial production in 2008, uses a patented process to convert rice hulls into rich-hull ash. The resulting rice-hull ash is rich in high-purity silicas and PACs, which are further processed into commercial products. Typical properties for this product are listed in Table 3.7. AgriTecSorbents’ Activated Carbon AgriCarb DC-600 is recommended for processing fats and oils; it is considered an ideal choice for the removal of chlorophyll from rapeseed oil (canola), as well as generalcolor removal. It also removes PAHs and color bodies from fish and vegetable oils.

Chemviron Carbon, Ltd. Chemviron Carbon, Ltd. (Ashton-in-Makerfield, Lancashire, UK) (www.chemvironcarbon.com) acquired 100-year-old Sutcliffe Speakman Ltd. (www.sutcliffespeakman. com) and its line of carbon products via a purchase by its parent company, Calgon Carbon Corporation, in March 2004. Typical properties of activated carbons offered by Chemviron for processing edible oils and fish oils are listed in Table 3.8.

89

Adsorbents

Recommended Applications for Chemviron Carbons 207AP PAH (powder) Pulsorb HF (powder)

Removal of polyaromatic hydrocarbons (PAHs) from edible oils and fish oils.

Cal / Cal 1 12x40

Removal of PAHs from edible oils and fish oils.

Cal-TR (granular)

Low-dust removal of PAHs from edible oils and fish oils.

TABLE 3.7 AgriTecSorbents, LLC/Powdered Activated Carbon General characteristics of AgriTecSorbents AgriCarb DC 600 Iodine number (mg/g)

625 min

Molasses-decolorizing efficiency (%)

95 min

Molasses-decolorizing index

20 min

Moisture, as packed (%)

8 max

pH, Water extract

7–9

Water solubles (%)

1 max

Apparent density (tamped) (kg/m3)

391

2

Surface area (m /g)

600

Sieve analysis: Thru 100 mesh (150 µm) %

95–100

Thru 200 mesh (75 µm) %

85–95

Thru 325 mesh (45 µm) %

80–90

TABLE 3.8 Chemviron Carbon/Activated Carbons 207AP PAH

Pulsorb HF

Cal/Cal 1 12x40

Cal-TR (granular)

Surface Area (m2/g)

>950

>1000

>1050

>1050

Moisture

<10%

<5%

<2%

<2%

Typical properties

pH

9–11

7–9

9–11

9–11

Ash

<15%

<14%

<10%

>5%

0.35– 0.45

0.35– 0.45

0.44– 0.55

0.44– 0.55

65–85%

60–80%

—

—

Mesh size >12 (1.70 mm)

—

—

5% max

5% max

Mesh size >40 (0.425 mm)

—

—

4% max

4% max

Apparent density (g/cc) Sieve analysis* (–325 mesh)

* U.S. Mesh (0.045 mm)

90

D. Taylor

Norit Nederland B.V. (Zenderen, The Netherlands) Norit Activated Carbon Two grades of PAC are recommended by Norit. The recommended grades are frequently necessary to control the presence of higher hydrocarbons and dioxins, thereby ensuring a food-safe operation according to the principles of HACCP guidelines. Typical properties are reported in Table 3.9. Norit SA 4 PAH, standard grade. t


Features: Proven performance on PAH, dioxin, and PCB adsorption. Specified values on PAH adsorption, good filtration properties, guaranteed purity to ensure a food-safe operation.

t


Norit SA 4 PAH-HF, high-filterability grade. Same features as for Norit SA 4 PAH but with a manipulated particle-size distribution to ensure excellent filtration properties.

TABLE 3.9 Norit B.V./Powdered Activated Carbon General characteristics of Norit SA 4 PAH-HF Light-PAH adsorption index

5.1

Heavy-PAH adsorption index

11.0

Total surface area (BET) (m2/g)

1150

Particle size D10 (µm)

7

Particle size D50 (µm)

34

Particle size D90 (µm)

100

Filtration time (min)

8 3

Apparent density (tamped) (kg/m )

530

Ash content (mass-%)

12

Chloride (acid extr.) (mass-%)

0.1

Moisture (as packed) (mass-%)

2

pH

Alkaline

Activated Earth/Carbon Mixtures Sϋd-Chemie AG(Munich, Lenbachplatz 6, 80333 Munchen, Germany) Two Tonsil bleaching earths are available already mixed with activated carbon and recommended for the decolorization of certain fats or oils which usually present some difficulty in bleaching (see Table 3.10). Süd-Chemie AG seems to be the only bleaching-clay manufacturer who currently offers bleaching-earth/activated-carbon mixtures.

91

Adsorbents

Recommended Applications for Tonsil Bleaching-Earth/Activated-Carbon Mixtures Tonsil 412 FF Tonsil 4121 FF

Manufactured by the acid activation of calcium bentonite and blended with activated carbon. These products possess an outstanding adsorptive capacity for the removal of polar compounds like chlorophyll, carotenoids, phospholipids, peroxides via chemisorptions and acid catalysis, as well as nonpolar components like polyaromatic hydrocarbons (PAHs). 412 FF contains 5% of carbon; 4121 FF contains 10% of carbon.

TABLE 3.10 Sϋd-Chemie AG Bleaching-Clay/Activated-Carbon Adsorbents Typical properties Tonsil Activated carbon, %

Standard 412 FF

Standard 4121 FF

5

10

Apparent bulk density (g/L)

520

508

Free moisture (2 h/110°C) %

≈10

≈10

10–12

14.7

pH (10% suspension)

Loss on ignition (predried, 2 h/1000°C) %

3.6

4.1

Acidity (mg KOH/g)

0.5

0.3

Chloride content (mg Cl/g)

0.4

0.4

Surface area (BET) (m2/g)

220

284

0–80 nm (mL/g)

0.35

0.32

0–25 nm (mL/g)

0.31

0.30

0–14 nm (mL/g)

0.27

0.26

Micropore volume

Amorphous Silica Hydrogel The most recent addition to the arsenal of sorbent materials used for purifying fats and oils is the amorphous silica hydrogels. Their use in this application was originally discussed in a mid-1960s paper by Bogdanor and Welsh (1967). These materials were subsequently patented (Parker & Welsh. U.S. Patent 4,734,226; Welsh & Parent. U.S. Patent 4,629,588) and commercialized. The basic processes for employing silica hydrogel in the adsorptive purification of fats and oils were detailed (Welsh & Bogdanor. Paper 90, 77th AOCS Annual Meeting, Honolulu, Hawaii, May 14–18,1986). Silica hydrogels are quite selective for the adsorption of phospholipids and soaps (including associated metal ions), but not for color bodies. As a consequence, amorphous silica hydrogels are used in conjunction with bleaching clays in the adsorptive purification of fats and oils. The presence of significant moisture (i.e., >30 wt%) keeps the structure open during the initial stages of contact, and aids filtration from the bleached oil. Silica hydrogels possessing average

92

D. Taylor

pore diameters greater than about 60 Å exhibit the highest capacity and selectivity for the adsorption of phospholipids and soaps (and associated metals ions) from glyceridic oils.

Preparation Typically, amorphous silica can be prepared by destabilizing an aqueous sodium silicate solution (“water glass”) by the addition of acids or certain inorganic salts to yield silica hydrogels. The product silica gels and/or precipitated silicas are derived from these hydrogels after appropriate washing and drying. Conditions employed during preparation are extremely important in determining final-product properties; generally, these are more important than the actual manufacturing process. Specifically, superior phospholipid adsorption is attained when the amorphous silica hydrogel is prepared in such a way that it possesses a high proportion of its total porosity (and associated surface area) concentrated in pores whose average diameter falls in the range of 60–350 Å. Practically speaking, such silicas will generally have surface areas in the range of 100–1200 m2/g and total-pore volumes in the range of 0.9–1.9 cc/g. A second important requirement of silica hydrogel used for adsorptive purification of fats and oils is that it should possess substantial quantities of water (typically >60 wt%) in its pores. The presence of the water, while it seems not to affect phospholipid adsorption itself, significantly improves the filterability of the silica from the oil, even after the use at oil temperatures that would cause the water content to be substantially lost during the treatment step.

Commercial Amorphous Silica-Hydrogel Products INEOS Silicas INEOS Silicas (Joliet, Illinois, USA) (www.ineossilicas.com), specializing in silica and alumina technology, produces liquid and solid silicates, gel and precipitated silicas, and a range of zeolites for detergent and other applications. According to the manufacturer, SORBSIL R® is carefully classified to reduce fine particles (<5 µm) which is said to reduce packing during filtration and greatly improve filtration rates. Typical properties are listed in Table 3.11. Supporting product literature points out that the best performance is obtained when the adsorption of target impurities is conducted at temperatures in the range of 70–95°C, and the moisture content of the oil is in the range of 0.1–0.2%. Recommended Applications for INEOS Amorphous Silica Hydrogel Product name

Suggested application

Comment

SORBSIL R ®

Amorphous silica hydrogel

Specially modified adsorbent with an affinity for soaps, metals, and phospholipids.

93

Adsorbents

TABLE 3.11 INEOS Silicas/Amorphous Silica Hydrogel Typical properties Total volatiles (1000°C, wt%) Mean particle diameter (Malvern, μm) Ave. Pore Diameter (Å) B.E.T. surface area (m2/g) Oil retention (blown, wt%)

SILSORB R ® 67 18 80–100 800 15–25

Chemical analysis (wt%)* SiO2 Al2O3, Fe2O3, MgO, CaO, K2O, Na2O

99.6 <0.1 (individually)

* Volatile free basis

W.R. Grace & Co. W.R. Grace & Co. (Columbia, MD, USA, 21044) (www.grace.com), the secondlongest continually operating chemical company in the United States (founded 1854), produces TriSyl® silicas at its Grace Davison division (Curtis Bay, Baltimore, MD). A variety of processes using the TriSyl silica family of adsorbents is available. Basically, TriSyl Silica Refining processes fall in one of two categories: bleaching processes which include Sequential Addition and Packed Bed and refining processes which include Modified Caustic Refining (MCR) and Modified Physical Refining (MPR). The company declined to provide information on typical properties of its TriSyl® products. Recommended Applications for Grace Davison Amorphous Silica Hydrogel For oils that have little or no chlorophyll, TriSyl silica can be used as a replacement for clay in both caustic and physical refining. Chlorophyll-containing oils can be processed by using TriSyl silica in combination with clay. All TriSyl silicas are acceptable for most bleaching applications, but higher capacity TriSyl 300 and 600 silicas are required for the modified refining processes.

Specialty Products That Act Like Amorphous Silica Hydrogel While the preceding amorphous silica hydrogels are basically high-purity synthetic silicas, other products were offered by bleaching-clay companies which perform the same task (selective adsorption of soaps, metals, and phospholipids). Although details of their manufacture are scant, they are probably derived from clays which were treated to enhance their selectivity for these particular contaminants. Suggested applications and typical properties for these products are given in Tables 3.10 through 3.12.

94

D. Taylor

BASF Catalysts LLC Recommended Application for BASF PreSorb™ Specialty Adsorbent PreSorb™

Specialty adsorbent designed for maximal soaps, metals, and phospholipids removal in the initial step of two-step bleaching processes. Suitable also as an adsorbent for biodiesel purification (Table 3.12).

TABLE 3.12 BASF PreSorb™ Typical properties

PreSorb™

% Free moisture (105°C)

14

Particle size % <200 mesh

87

Acidity (mg KOH/g)

21

pH

2

Apparent bulk density (g/cc)

0.6

Surface area (m2/g)

325

Filter rate*

15

* BASF Internal Test

Oil-Dri Corporation of America Select Product Line Magnesium aluminum-based silicates are specifically developed for vegetable-oil refining and biodiesel applications (Table 3.13). Product name

Suggested applications

Comments

Select 350

Use in place of traditional silica-gel adsorbents.

Specially modified adsorbent with an affinity for soaps, metals, and phospholipids.

Select 450

Use in place of traditional silica-gel adsorbents.

Specially modified adsorbent offering improved filtration characteristics.

TABLE 3.13 Oil-Dri Select Products: Typical Properties Product

Mesh % -325

Acidity pH 5% slurry

% Free Moisture

Densitypacked (lb/ft3)

Densitypacked (kg/m3)

Select 350

62

3.2

4.5

30–38

481–609

Select 450

63

3.2

15.5

39–53

625–849

95

Adsorbents

Sϋd-Chemie AG Recommended Applications for Süd-Chemie Tonsil Specialty Adsorbents Tonsil 419 FF

Specialty adsorbents particularly recommended for the refining of fats and oils and for the removal of phospholipids (gums), soaps, and trace metals .

Tonsil 4191FF

Possesses excellent bleaching effect for adsorption of chlorophyll A.

TABLE 3.14 Sϋd-Chemie Tonsil Specialty Adsorbents: Typical Properties Typical properties of Tonsil

Standard 419 FF

Standard 4191 FF

Apparent bulk density (g/L)

300

330

Free moisture (2 h/110°C) %

max 15

≈15

LOI (pre-dried, 2 h/1000°C) %

8.6

10.2

pH (10% suspension)

6–8

≈3

Acidity (mg KOH/g)

0.2

2

Chloride content (mg Cl/g)

0.2

0.1

Surface area (B.E.T.) (m2/g)

220

220

Micropore volume

Particle-size distribution

0–80 nm (mL/g)

0.49

0–25 nm (mL/g)

0.34

0–14 nm (mL/g)

0.23

>150 µm %

13

>100 µm %

24

>63 µm %

36

>45 µm %

48

>25 µm %

61

Chapter 4

Bleaching of Important Fats and Oils H.B.W. Patterson

General Principles Now well-understood is that bleaching may actually mean the adsorptive purification of an oil prior to an important process step such as hydrogenation, physical refining, or deodorization. As we know, these later steps then become more effective regarding both the quality of product and the economy in services, time, and reagents. Thus, the oxidative stability of a refined–bleached–deodorized soybean oil is substantially greater than the same oil merely refined and deodorized (Ariaansz et al., 1989). Equally, adsorption itself is more effective when an earlier, relatively simple cleaning of the oil is performed, such as described in previous sections (see sections Settling–Neutralization and Washing and Use of Silica in Chapter 1). In this chapter specific recommendations for the bleaching/purifying of about twenty fats and oils are set forth. Apart from giving the basic fatty acid constitution, reference is made to important features of the oil which may be present, such as particular vulnerability to oxidation, richness in natural antioxidants, or danger of so-called color reversion. From what was said in the first three chapters of this work, several general principles emerge which are set forth here before going on to make recommendations for particular oils. Each detailed recommendation which follows includes the reminder to consult this section. 1. Keep opportunities for oxidation, thermal damage, and hydrolysis to a minimum. 2. Choose the lowest practical process temperature. 3. Bleach under vacuum or an inert gas. 4. Thirty minutes at the bleaching temperature is enough; sometimes appreciably less is feasible. 5. Reduction of soap in the neutralized, washed oil helps the subsequent action of the adsorbent; a soap content less than 50 ppm is attainable with centrifugal refining. 6. Adding the adsorbent below 80°C and then raising the temperature of the oil with agitation under vacuum to the peak temperature is the preferred procedure. 7. Phosphoric and citric acids are the popular agents for scavenging trace metals and substantially increasing the oxidative stability of oils. Free phosphoric acid remaining in the oil at a level of over 5 ppm of phosphorus is likely to promote some increase in free fatty acid (Taylor & Ungermann, 1987). Free phosphoric 97

98

H.B.W. Patterson

acid must be removed before hydrogenation or deodorization. 8. For best results, especially with poor oils, one aims at zero peroxide value, under 50 ppb of chlorophyll or under 5 ppm of phosphorus rather than a Lovibond color alone. 9. Often most economical is to use the smaller dose which a more active clay makes possible. This reduces oil lost in clay and increases filter utilization, as well as brings about other less obvious indirect savings. 10. If a blend of bleached oils is to be made, the best idea is probably to bleach any difficult oil separately. Special circumstances, such as the oil’s characteristics (free fatty acid, color, etc.), exceeding the capacity of the processing unit may justify blending beforehand. 11. Do not forget to exploit the press-bleaching effect (see section Batch Bleaching in Chapter 1). 12. Since starting color, desired final color, temperature (to some extent), purity of oil, and the relevant activity of the many grades of clay considered are all variables, the recommendations must be considered as a trustworthy guide forming the basis of adjustment in the light of results.

Lard Selected tissues of clean healthy pigs are rendered by either dry or wet heat to produce lard; the latter is more popular and may be assisted by the maceration of the tissue (Anon., Am. Meat Digest; Norris, 1979). Various parts of the carcass are excluded by Codex Alimentarius 28-1981, but Codex 29-1981 allows a greater range in choice of healthy tissue for rendered pork fat. If any processed pig fat is incorporated in otherwise unrefined material, the label must state this. Both the diet of the animal and the location of the organ in the carcass from which the fat was rendered affect its characteristics. Fat from interior organs is firmer [lower iodine value (IV)] than subcutaneous fat (higher IV). U.S. lard tends to be softer (slip m.p. 33°C) than European lard (slip m.p. 35°C), possibly because less than a random proportion of trisaturated triglycerides exists in the former and more in the latter. Thus, a rise can occur in the slip point of the former and a fall in the latter upon interesterification. Although scores of very minor fatty acid components, including odd-numbered and branched-chain items, were identified in recent years, Table 4.1 shows a typical and simplified composition of a lard sample of 62 IV (Patterson, 1983, 1989). A more detailed composition is available in Supplement 1 to Codex Alimentarius Vol. XI., First Edition (1983 ). Table 4.2 summarizes the characteristics of raw edible-class lard which, at best, is fit for immediate consumption, or, if inferior, is capable of being refined for edible

99

Bleaching of Important Fats and Oils

TABLE 4.1 Typical Fatty acid Composition of Lard (Simplified) Acid type

%

C14:0

2

C14:1

0.5

C16:0

24

C16:1

4

C18:0

14

C18:1

43

C18:2

9

C18:3

1

C20:0

0.5

C20:1

1

C20:2

1

TABLE 4.2 Characteristics of Unprocessed Edible Grades of Lard (Patterson, 1975, 1976, 1983) Superior

Normal

Inferior

Free fatty acid (% maximum) (mol wt 282)

0.5

1.0

1.5

Unsaponifiable (%)

<1.0

<1.0

<1.0

Moisture plus impurity (%)

0.1

Color maximum (Y + R 1”)

10 + 1

12 + 1.2

15 + 1.5

—

63–71 usual

—

—

46–77 possible

—

Iodine value

Saponification value

1.0

190–203

Peroxide value (meq)

5 maximum

BoumL mer (acetone)

74 minimum

up to 10

up to 10

71

71

m.p. (°C)

31–37

Titer (°C)

32–43

100

H.B.W. Patterson

use. The maximal percentage of free fatty acid of 1.5% for inferior lard depends in practice on what refining effort is economical; in some circumstances, it might be taken as average. Phosphatides are expected to be less than 0.05%. Lard is an example of a fat in which distinguishing between virgin and processed material is important. The work of H.P. Kaufmann and others (1956 a,b; 1958) relates particularly to whether the fat may be regarded as bleached or not. Bleaching increases the E268 and the triene content (see section Oxidized Fats and Oxidized Fatty Acids in Chapter 1). Thus, if the values of E1% of 1 cm are found for 264 nm, 268 nm, and 272 nm, then one-half of E264 and E272 (i.e., the mean value) is subtracted from the E268 value, and the difference is multiplied by 100; the result is called the T value. For nonprocessed lard, T is less than 1.0, but following bleaching, it becomes substantially greater than 1.0. A correction may be made to allow for the fact that when E268 is higher, T tends also to be greater. A value Q = T/E268 is easily calculated; then for nonprocessed lard, Q lies between 3 and 4, but following bleaching, it lies between 7 and 10, and is sometimes higher. In practice, when T is >1 and Q is >6, the lard is judged to have been bleached. Desirable is to seek lower trace metal contents than allowed by the Codex; an Fe of a 1.0 ppm maximum and Cu of 0.05 ppm should be available. Animal fats are well-known to contain very little natural antioxidants, and lard stability is considerably improved by the addition of up to 250 ppm of tocopherol. For butylated hydroxyanisole (BHA) or butylated hydroxytoluene (BHT), the optimal effective dose is 200 ppm, and their effect is further improved by the addition of 100 ppm of citric acid if prooxidant trace metals are known to be present. Tertbutylhydroquinone (TBHQ) gives good results with lard; plastic packing material, if present, should be eliminated in the same way as for tallow (see section Beef Tallow, subsection Recommendations in Chapter 4).

Recommendations The neutralized–washed–dried lard will have free fatty acid n> 0.1% and soap n>0.05%. A maximal dose of 0.5% of mildly activated clay (Tonsil Standard FF, Fulmont 300C, Engelhard F4, or Galleon NS) or 0.25% of moderately activated earth (Tonsil L80FF, Fulmont AA, Engelhard F105, Galleon N, or NF2) is adequate. A contact time of 15 minutes at 95–100°C is sufficient. Only for exceptionally poor qualities should a heavier dose of very active clay (Tonsil Optimum FF, Fulmont Premiere, Engelhard F160, or Galleon V2) and a temperature up to 110°C be needed. A drop in color from 15 Y 1.5 R (5¼”) to 7 Y 0.4 R (5¼”) is typical. (See section General Principles in Chapter 4.)

Beef Tallow Fat derived from cattle is usually softer than that from sheep or goats, and, as with pigs, inner organs yield the firmest fat (lower IV) and subcutaneous tissue the softest. Codex Alimentarius (Codex Stan. 30-1981) for Premier Jus (oleo stock) limits the fat to be rendered at low heat from certain inner organs of healthy bovine animals collected at slaughtering time (Baldwin, 1985; Bates, 1968; Zucker, 1968). Climate

101

Bleaching of Important Fats and Oils

and diet have a marked effect on the fatty acid makeup of tallow. One may derive edible tallow or dripping, according to Codex Stan. 31-1981, from a considerably wider range of organs, and one may include sheep. As with lard, a huge array of minute proportions of fatty acid with branched or uneven-numbered carbon chains was identified. Table 4.3 shows a simplified range of typical fatty acid compositions for Premier Jus. TABLE 4.3 Typical Ranges of Fatty acid Composition of Premier Jus Acid type

%

Acid type

%

C <14

<2.5

C17:0

0.5–2.0

C14:0

1.4–6.3

C17:1

<1.0

C14:1

0.5–1.5

C18:0

6–40

C15:0

0.5–1.0

C18:1

26–50

0 iso

<1.5

C18:2

C16:0

20–37

C18:3

<2.5

0 iso

<0.5

C20:0

C16:1

0.7–8.8

C20:1

<0.5

C16:2

<1.0

C20:4

<0.5

0.5–5.0 <0.5

A detailed fatty acid composition covering Premier Jus and edible tallow is provided in Codex Alimentarius Vol. XI, Supplement 1 (1983). Table 4.4 indicates the standards to be followed for tallows which are for edible use. The better qualities within the superior grade are the only ones likely to be used in such a way without refining. Procedures for this refining were described in detail (Patterson, 1975, 1976). The processors must decide whether the 2% max. of free fatty acid for inferior tallow represents a rigid requirement or not, because they may be prepared to meet an increased refining cost to achieve specification. If a refined edible tallow is included in edible tallow, the Codex requires this to be stated on the label. Tallow hydrogenated to an iodine value of <3 provides a source of stearic acid for conversion to monoglyceride emulsifiers of edible grade. If these are distilled, an even lighter product is obtained. One is unwise to judge tallow simply on the basis of color because some parcels respond to bleaching much more than others. Furthermore, a parcel may be made to look better by the vendor simply by blending in a proportion of already-bleached tallow. In this case, the response to bleaching is virtually certain to be less than that of an untreated tallow of the same color. Tallow provides a good example of how some preliminary cleaning of feedstock makes subsequent steps—such as bleaching, hydrogenation, and so forth—so much more effective. If a genuine superior-grade tallow is subjected to 2% of activated-earth bleach prior to hydrogenation, the ultimate hardened fat and the monoglycerides are likely to have the required very light color. The case may well be, however, if a poorer and cheaper grade is chosen, neutralized,

102

H.B.W. Patterson

TABLE 4.4 Characteristics of Unprocessed Beef Tallow—Top Grade Superior,

Normal,

Inferior

1.0

1.5

2.0

Unsaponifiable (% maximum)

0.8

0.8

Moisture plus impurities (%)

0.1

Free fatty acid (% maximum) (mol wt 282)

Color maximum (Y + R 1”) Iodine value Saponification value Peroxide value (meq) m.p. (°C)

10 + 1

1.5 1.0

18 + 1.8

25 + 2.5

33–58 190–203 2–3 40–48

Titer (°C) a

Swift life (minimum hours to give a peroxide value of 10)

40–47 14

7

a

Swift life is the number of hours required to reach a stated peroxide value when a sample is aerated at ca. 150 cm3 of air/minute at 98° 0.5°C.

washed, and then bleached and hardened, the color of the hardened oil in this case will also prove satisfactory for the production of monoglycerides. In such a case, the following progression was quoted (Patterson, 1983, 1989): neutralized + washed, 5–18 R (5¼”); bleached, 2–4 R (5¼”); fully hardened and post-treated, 0.2 R (1”). Remnants of polyethylene packaging are a nuisance in some parcels of animal fats. If a sample kept for 5 hours at 60°C remains completely clear, negligible or no plastic is present. If cloudiness is believed to be due to the separation of some stearin, the sample must be dissolved in a solvent, filtered, and the infrared spectrophotometric absorption checked. Usually, after bleaching, fat is filtered at ca. 90°C; if, however, filtration temperature is lowered below 70°C, any polyethylene will agglomerate, and, hence, filtration reduces plastic content below 20 ppm.

Recommendations For a tallow which was neutralized and washed, the color should fall in the range of 160 Y 18 R (5¼”) to 50 Y 5.0 R (5¼”). For top grades, a choice is available between up to 1% of mildly activated clay (Fulmont 300C, Tonsil Standard FF, Engelhard F4, or Galleon NS) or about 0.3% of well-activated clay (Fulmont AA, Tonsil L80FF, Engelhard 105 or 105SF, Galleon N, or NF2). A contact time of 20 minutes at 95–100°C is normal. Beyond these choices, the limiting dose of clay should be 2%. Very active clays such as Fulmont XMP3, Tonsil Optimum FF, Engelhard F160, and Galleon V2 could be used. Sometimes 0.5–1% of activated carbon in addition to the clay may improve color and clarity. (See section General Principles in Chapter 4.)

Bleaching of Important Fats and Oils

103

Butterfat A huge range of fatty acids, saturated and unsaturated, some with branched chains and some with odd numbers of carbon atoms, was reported over the years (Sonntag, 1979). Usually, chain length falls between C4 and C22 as shown in Table 4.5. Monounsaturated disaturated (US2) is generally 33–39%, but as trisaturates (S3) fall from 40 to 27%, monosaturated diunsaturateds (SU2) increase correspondingly. The iodine value varies widely from 30 to 43, depending on breed, season, and diet. A typical phospholipid content is only 0.3% with sulfur around 8 ppm. Carotenoids are the principal pigment, and vary around 20 ppm, again depending on breed and season; a typical color would be 50 Y 5 R (5¼”). Neutralized, washed butterfat is easily brought to <0.1% of free fatty acid, 0.05% of soap, and 0.01% of phosphatide. TABLE 4.5 Typical Fatty Acid Composition of Butterfat (Simplified) Acid type

%

Acid type

%

C4:0

3

C16:0

29

C6:0

1

C16:1

4

C8:0

1

C18:0

11

C10:0

3

C18:1

25

C12:0

4

C18:2

2

C14:0

12

C14:1

2

>C18:0

2

C20 and C22 unsat.

1

Recommendations The usual choice between a generous dose of ca. 1% of a mildly activated clay (Engelhard F4, Tonsil Standard FF, Laporte 300C, or Galleon NS) or a much reduced dose of ca. 0.4% of well-activated clay (Engelhard 105 or 105SF, Tonsil L80FF, Fulmont AA, Galleon N, or NF2) exists. A contact time of 20–30 minutes is ample. For the most active clays, a temperature of no more than 100°C is advisable; for mild clays 110°C is permissible. A final color of 20 Y 2.0 R (5¼”) to 4.0 Y 0.8 R (5¼”) is to be expected. (Se section General Principles in Chapter 4.)

Coconut Oil The coconut palm is cultivated in coastal areas within 20 latitudinal degrees either side of the equator around the world. In particular, several countries have formed the Asian and Pacific Coconut Community (APCC), with headquarters in Jakarta; this provides a consultancy service which is supported by UNIDO. They have developed a use for every part of the plant (1980). The dried flesh, or meat, of the nut—“copra”—contains 64–70% of oil. The relatively simple fatty acid makeup of

104

H.B.W. Patterson

the oil, shown in Table 4.6, is dominated by short-chain fatty acids, with lauric acid being the most important. Because of the similarity of their chemical constitution and related botanical origin, palm kernel, babassu, tucum, murumuru, ouricuri, and cohune oils, along with coconut oil, are known as lauric oils. The high content of short-chain fatty acids gives the oils a much-valued rapid melting point. Coconut oil melts quickly at around 25°C, and the fully hardened oil at around 35°C (Patterson, 1983). Worth noting is that even a small contamination (1–2%) with nonlauric oil elevates the slip point of the contaminated oil when fully hardened by a few degrees. This can be useful to the user and a source of confusion between the manufacturer and the customer. The Codex Stan. 124-1981 applies to coconut oil; possible variations and current quality standards were described (Graalmann, 1989; Norris, 1979; Young, 1983). The small amounts of unsaturated fatty acids suggest a resistance to oxidation, but a wide variation in AOM stability tests on the refined oil of 30–250 hours was, in fact, reported (Norris, 1979); the addition of BHA and citric acid makes results ca. 350 hours possible. Phosphatides are normally absent, and only up to 4 ppm of sulfur is likely to be encountered. Precautions against phosphorylation (if degumming with phosphoric acid is employed) are mentioned in the discussion of degumming (see section Degumming in Chapter 1). If copra was contaminated by sulfurous compounds—presumably if dried by direct contact with hot flue gas—a most unpleasant flavor can occasionally develop within seconds of the freshly deodorized oil being exposed to sunlight and fresh air. This is prevented by a brief contact of the original oil with nickel catalyst and hydrogen (Patterson, 1983). TABLE 4.6 Typical Fatty Acid Composition of Coconut Oil Fatty acid

Type

%

Caproic

C6:0

trace

Caprylic

Fatty acid

Type

%

Palmitic

C16:0

9

C8:0

8

Stearic

C18:0

2.5

Capric

C10:0

6

Oleic

C18:1

7.0

Lauric

C12:0

47

Linoleic

C18:2

2.5

Myristic

C14:0

18

If the hydrolytic splitting of the lauric oils’ triglycerides occurs, this quickly leads to an objectionable soapy taste. Smoke drying of copra leads to the contamination of coconut oil with up to 3 mg/kg of polycyclic aromatic hydrocarbons (PAHs) (see section Use of Carbon in Chapter 1). While the lighter PAHs are removed by deodorization, 5-, 6- and 7-membered-ring compounds need to be adsorbed by a dose of 0.4% of well-activated carbon (see section Commercial Activated Carbon Products) added during the bleaching cycle. Fortunately, several producers of copra and oil appear to have avoided this risk (Sagredos et al., 1988). The danger of PAHs in foodstuffs needs to be considered against the danger of PAHs in petroleum-engine exhaust fumes. A

Bleaching of Important Fats and Oils

105

crude oil of Fe (2 ppm) and Cu (0.2 ppm) should fall to the maxima of Fe (0.1 ppm) and Cu (0.02 ppm) after refining.

Recommendations Commercial contracts for edible-grade processing may demand a color of 50 Y 9 R (5¼”) maximum (Pritchard, 1983 ). In practice, a very good crude oil may be as little as 1.5 Y 0.8 R (1”), while 7.5 Y 1.5 R (1”) is common. Industrial grades of crude coconut oil envisage free fatty acid of 4–10% and a color of (Y + 5 R) 9–30 (UNIDO, 1980). A neutral or very lightly activated earth such as Tonsil ACC FF, Engelhard F4, Galleon NS, or Fulmont 300C or 237 at a dose of 0.2–0.4% should be adequate for top-grade edible oil. For a more vigorous action, Tonsil Standard FF or L80 FF, Engelhard F105, Fulmont 700C or Galleon N, or NF2 at about the same dose rate should prove adequate to meet normal specifications. More highly activated earths are unlikely to be necessary. Whatever the dose rate of earth, 0.4% of a well-activated carbon such as Norit FND should be added 10 minutes after the earth, and the oil should then be maintained at 90–95°C with full agitation for another 45 minutes. This ensures satisfactory adsorption of the heavy PAHs. Carbon also contributes to the appearance of a light, bright oil, and might achieve the necessary light color on its own, but some precoating of the filter is advisable. Süd-Chemie (see section Activated Earth/Carbon Mixtures in Chapter 3) produces a range of earth/carbon mixtures (Tonsil 25C FF—70 CC FF), and one of these could be investigated as an alternative to the above mixtures. Fulmont and Engelhard earth/ carbon mixtures are also available. After bleaching, an exceptionally light-colored oil would read 4.5 Y 1.2 R (5¼”), while 7.0 Y 1.3 R (5¼”) would be more usual. If an advantage is taken of the pressbleaching effect to reduce earth doses after the first batch on the filter, the carbon dose should remain unaltered. If it has been found adequate to use a very small earth dose (e.g., 0.1%), leaving this unaltered in subsequent batches will act against the blinding of the filter; that is, it is acting as a form of “body feed.” (See section General Principles in Chapter 4.)

Cottonseed Oil Cotton is grown in warm, temperate, and tropical climates around the world primarily for its fiber, and yields—for every 100 pounds of fiber—180 pounds of seed. The seed contains some 16–25% of oil after the complete removal of the fiber (clean whole seed) and drying. In 1978, UNIDO published (Anon., 1978) a detailed review of different varieties of cotton plants as well as commercial and technical information on seed, oil, and cake. Codex Stan. 22-1981 is devoted to cottonseed oil, and the Federation of Oils, Seeds, and Fats Association [FOSFA (Rule 54)] and the National Cottonseed Products Association (NCPA, of the USA) are two of the most important organizations concerned with trade in cottonseed and oil (Patterson, 1989). More recently, the same field was reviewed by L.A. Jones of NCPA (1989).

106

H.B.W. Patterson

The oil belongs to the important oleic–linoleic family, as is obvious from the simplified fatty acid composition quoted in the Codex and shown here in Table 4.7. The paucity of more unsaturated fatty acids plus the presence of natural antioxidants, especially tocopherols, give cottonseed oil oxidative stability. Over 2% of the crude oil is likely to be made up of phosphatides (see section Phosphatides (Gums) in Chapter 1), gossypol (see section Gossypols in Chapter 1), sterols (see section Sterols in Chapter 1), and about 1000 ppm of various tocopherols (see section Tocopherols and Chromans in Chapter 1) besides carbohydrates and pigments. These result in a crude oil of a deep-brown color. By correct management of harvesting and storage of seed, a crude oil of 1 ± 0.5% of free fatty acid can be obtained; when seed is allowed to rot in damp storage, the end result may well be 8–11% of free fatty acid, which represents a very serious loss of at least double that percentage when the oil comes to be refined. Fortunately, conventional alkali neutralization and washing bring down the levels of phosphatides and gossypol to under 100 ppm each, leaving the way open for active clay bleaching to reduce this level even further (Norris, 1979). Physicalrefining methods have not yet been found which deal with the removal of gossypol so effectively, so chemical refining has retained its popularity (Young, 1989). Because the contents of tocopherols and the phenolic-type pigment, gossypol, are greater in the crude oil than in the refined oil, this was claimed as a reason for its greater oxidative TABLE 4.7 Fatty Acid Composition of Cottonseed Oil Fatty acid type

Typical (%)

Guideline range of variation (Codex Stan. 22) (%)

C <14

trace

<0.1

C14:0

1

0.4–2.0

C16:0

25

17–31

C16:1

0.5

0.5–2.0

C18:0

2.5

1.0–4.0

C18:1

18

13–44

C18:2

51.5

33–59

C18:3

0.5

0.1–2.1

C20:0

0.5

<0.7

C20:1

trace

<0.5

C22:0

trace

<0.5

C22:1

trace

<0.5

C24:0

trace

<0.5

Cyclopropenoid

0.5

Bleaching of Important Fats and Oils

107

stability. Also noted (Hudson & Ghavani, 1984) is that phosphatides act as synergists in crude cottonseed oil. Immediately, this must be said: since the removal of color from the crude oil is found to become more difficult when it is allowed to stand for a few weeks, customarily, at least a partial neutralization and washing are performed promptly; hence, both domestic and international trade are carried on with “washed,” that is, semirefined oil. For closely related reasons, if, after expelling, the crude oil is not cooled promptly, a poorer response to bleaching becomes likely. Traces of sulfur and metals are unlikely to be met. An unusual feature of cottonseed oil is the presence of cyclopropenoid fatty acids, even reaching the 1% level in exceptional cases. Sterulic acid (Fig. 4.1) is one of the best known of these, and serves to identify cottonseed oil in the Halphen test by giving a distinct red color when the oil is warmed gently with carbon disulfide and amyl alcohol. Heating or hydrogenation destroys the intensity of the color, and may well render the test useless. Fully refined cottonseed oil is used in margarine and cooking oil; after winterization (cold test 7°C), it is used as a salad oil, and when hydrogenated, as a shortening or vegetable ghee (vanaspati). Both free fatty acid and color are involved in defining the grades of oil for trading purposes so processors must consult the latest trading rules of the organization under which they propose to operate. For crude cottonseed oil, feasible standards would be:

Scheme 4.1. Sterulic acid.

3.25% of free fatty acid (FFA) maximum capable of yielding a semirefined (washed) oil of (a) 35 Y 7.6 R (5¼”) maximum with loss between 9 and 12%; or (b) 35 Y 12 R (5¼”) maximum with loss between 9 and 20%. The titles—prime crude oil, basis prime crude oil, off crude oil, reddish off crude oil, and low-grade crude oil—may be met. For semirefined oil, these standards could apply: Prime summer yellow-semirefined (PSY), FFA of 0.25% maximum, 7.6 R (5¼”) maximum, moisture and volatiles 0.1% maximum. Prime bleachable summer yellow (PBSY), FFA of 0.25% maximum, 2.5 R (5¼”) maximum. Other semirefined grades were described as choice summer yellow, prime winter yellow, good off summer yellow, summer yellow, off summer yellow, and reddish off summer yellow. Bleached (neutralized) oil may be classed as prime summer white and prime winter white (Anon., 1978). A general standard for neutralized washed cottonseed

108

H.B.W. Patterson

oil would be 35 Y 3 R (1”) maximum with free fatty acid of 0.2% maximum, but in European experience 5.5 R (5¼”) is quite normal. Alkali refining and subsequent clay bleaching eliminate aflatoxins if present.

Recommendations ( Before bleaching, the oil should be neutralized and washed, probably to well below 0.1% of free fatty acid and 0.05% of soap. Typically, such an oil would then show 115 Y 5.5 R (5¼”). If a moderately activated clay is used (Fulmont 300C, 237, or 700C; Tonsil ACC FF; Standard FF; Engelhard F4; or Galleon NS), a dose of upward from 1% would probably ensure a color of around 30 Y 3.0 R (5¼”), suitable for use in margarine blends. A contact time of 20 minutes at 90–95°C is normal. For hydrogenation, a somewhat higher color would be acceptable. To obtain lighter colors, more active clays were employed (Fulmont XX, Tonsil L80 FF, Engelhard F105 and F105 SF, and Galleon V2 and V2-Super). Obviously, if more than 2% of earth is used, lower filter utilization and greater oil-in-earth loss are incurred. (See section General Principles in Chapter 4.)

Grapeseed Oil After crushing out the juice from grapes, about one-quarter of their weight remains as pomace (Patterson, 1989). The same term is applied in processing olives and tomato seed. The grape pomace is pressed further, dried, and the seed separated and cleaned. The seed amounts to about 4% of the original weight of the grape; its oil content is usually 14–20%, but some black varieties have as little as 6%. Cold-pressing yields golden oil, hot-pressing dark green or brown. Cooking, flaking, and solvent extraction (Bernardini, 1973) yield a green oil difficult to bleach. Table 4.8 shows the fatty acid composition suggested by Codex 127-1981. The oil is of the oleic–linoleic family; although the Codex suggests an iodine value range of 130–138, a wider range of 124–143 seems possible with an unsaponifiable of less than 2%. To follow conventional phosphoric-acid degumming with alkali neutralization, washing, and drying is customary so that the phosphatide content is less than 0.1% (4 ppm of phosphorus) before bleaching. Because the oil has a slip melting point ca. 10°C, it has potential as a salad oil as well as for hydrogenation to medium-melting-point fats (Patterson, 1983) and as a component in soapmaking. Recommendations At one time, the bleaching of grapeseed oil after alkali refining depended on a heavy (1.5–2.0%) dose of an acid-activated clay which had received more than the average degree of activation. In addition, a highly activated carbon such as Norit SA4 at a dosage of 1% was added. Possibly an interval of up to 10 minutes between the two additions would allow the clay to adsorb some of the nonpigment bodies and allow the carbon to be most effective. A bleaching time of 30 minutes at 105°C is sufficient. A final color of 40 Y 4.0 R 1 B (5¼”) is feasible, but depends on the quality of the original oil. Omission of the carbon makes it necessary to increase the clay dose by

Bleaching of Important Fats and Oils

109

TABLE 4.8 Ranges of Fatty Acid Composition of Grapeseed Oil Fatty acid C12:0

Composition (%) <0.5

C14:0

<0.3

C16:0

5.5–11

C16:1

<1.2

C18:0

3.0–6.0

C18:1

12–28

C18:2

58–78

C18:3

<1.0

C20:0

<1.0

C22:0

<0.3

C24

<0.1

more than an additional 1%. In the 1980s, adsorbents appeared with a superior performance in removing chlorophyll and its related compounds. These do not depend on any help from carbon, although efforts made at an earlier stage to reduce the concentration of nonpigment impurities (soap, gum, etc.) are worthwhile. One can now achieve a target of 50 ppb of a chlorophyll maximum on the bleached oil starting at 500 ppb (soybean oil), but when prebleached chlorophylls are around 21,700 ppb (canola oil), the dose rises (Fig. 2.2) considerably to 1.8% (Taylor & Ungermann, 1987). Ultimately, it is a question of balancing the expense of a possible multistage treatment and special adsorbents against the premium for this or that specification, bearing in mind that chlorophyll above 50 ppb will initiate oxidation (Gunstone, 1984; Lee, 1987; Mag, 1989; Usuki et al., 1984), apart from the question as to whether a certain shade of green/yellow is acceptable. Adsorbents now commercially available which are designed to have a high capacity for the removal of chlorophyll compounds are Tonsil Optimum FF, Fulmont XMP2 and XMP3, Engelhard NevergreenTM and F160, and Mizukalife F-1 and DC. Depending up the initial chlorophyll content, a dosage of 0.5–2.0% should first be investigated as a means of meeting specifications. (See section General Principles in Chapter 4.)

Groundnut (Arachis, Peanut) Oil This well-known nut of South American origin is now cultivated in the warm temperate regions of Asia, Africa, and the United States. The kernel is rich in oil (ca. 50%), and is such a valuable source of protein that in many places it is consumed as

110

H.B.W. Patterson

such. The oil is one of the oleic–linoleic family. As with other oils and fats, material which develops in the cooler location tends to the higher iodine value and vice versa (Ory & Flick, 1989), the range for groundnut oil being 80–106. Two older fatty acid compositions, one for Nigerian and the other for Argentinian oil, and the present mandatory Codex 21-1981 (amended 1987) composition are shown in Table 4.9. Some points of special interest follow: 1. Around 8% of long-chain fatty acids is usual. 2. The particularly low content of linolenic acid confers oxidative and flavor stability. 3. Argentinian oil as shown is more unsaturated than Nigerian, yet has a slightly higher melting point because of its higher content of long-chain fatty acids. Groundnut oils melt around 2°C and have a jelly-like consistency even at 5°C. The unsaponifiables are low at 1% or less, sulfur is negligible, and tocopherols up to 500 ppm provide a useful antioxidant effect. Trading associations such as FOSFA and NCPA provide standards covering color and free fatty acid. To settle crude oil for about three weeks is highly beneficial (Patterson, 1989) so as to drop phosphatides to the foots and preferably leave under 0.2% in the bulk oil above. In this case, a

TABLE 4.9 Fatty Acid Compositions of Groundnut Oil Fatty acid (chain length)

Nigerian (%)

Argentinian (%)

C <14

<0.4

C14:0

<0.6

C16:0

10

11

6.0–16

C16:1

<1.0

C17:0

<0.1

C18:0

3.5

3

1.3–6.5

C18:1

59

39

36–72

C18:2

20

38

13–45

C18:3

0.5

0.5

<0.3

Arachidic C20:0 (Eicosanoic)

1.5

1.5

1.0–3.0

Gadoleic C20:1 (Eicosenoic)

1.5

1.5

0.5–2.1

Behenic C22:0 (Docosanoic)

2.5

3.5

1.0–5.0

1.5

2.0

0.5–3.0

Erucic C22:1 (Docosenoic) Lignoceric C24:0 (Tetracosanoic) a

Codexa

Mandatory 1987 (Mounts, 1987).

<0.3

Bleaching of Important Fats and Oils

111

refined oil of less than 0.01% of phosphatide is attainable. As explained previously (see section Oxidized Fats and Oxidized Fatty Acids in Chapter 1), bleaching may convert dienes to trienes with a fall in E232 and rise in E270. Subsequent deodorization above 180–220°C has the reverse effect. If data are available on peroxide value, anisidine value, and oxidized fatty acid content of the oil, a valuable indication is then given of its likely flavor stability and whether an inferior crude oil was processed so as to disguise its deficiencies. Thus, a vegetable oil with an E232 at 5 or more is regarded as poor, and an E270 above 5 suggests advanced oxidation. A large amount of groundnut oil is used as table oil in France, and there the government regulations require that the ratio E232/268 exceeds 2; similarly, for several other vegetable oils, it must exceed 1. Such data are also linked with color specifications to indicate whether heavy processing was used to come within specifications (Patterson, 1989). Although the carcinogenic aflatoxins develop on nuts which were allowed to become moldy, these drop to ca. 1000 ppb in crude oil, to less than 15 ppb on alkali refining, and to under 1 ppb on bleaching (Parker & Melnick, 1966).

Recommendations A wide range of colors is met in crude oil; 35 Y 4 R (1”) would be very high, whereas 30 Y 1.5 R (1”) to 15 Y 1.0 R (1”) or 70 Y 6 R (5¼”) would all be considered typical. The alkali-refined, washed oil prior to bleaching will contain less than 0.05% of soap—possibly less than 0.005%—and less than 0.01% of phosphatides. If soluble soap is a nuisance, an addition of 93 parts of MgSO4:7 H2O for every 100 parts of soap renders the latter insoluble and assists adsorbents. A contact time of 20 minutes with clay at 90–95°C is normal. A dose of 1% of medium-activated clay such as Galleon NS, Engelhard F4, Tonsil ACC FF or Standard FF, and Fulmont 300C should be adequate for normal oil, and a final color of 20 Y 2 R to 5 Y 0.5 R (5¼”) is attainable. (See section General Principles in Chapter 4.)

Illipe Oil, Borneo Tallow, and Other Vegetable Butters Illipe oil (mee oil), obtained from the seeds of the Indian plant Madhuca longifola, is similar to mowrah fat, obtained from another Indian plant, Madhuca latifolia. Both have been used for technical purposes and, when refined, also for edible use (Sonntag, 1979). Illipe and mowrah fats more closely resemble shea butter in their proportion of unsaturated fatty acids (oleic and linoleic) than cocoa butter or Borneo tallow. Borneo tallow (green butter) is obtained from the plant Shorea stenoptera, common in the East Indies and Malaya, where about one-hundred related members of the species grow. In some varieties, the size of the nut produced makes harvesting it unprofitable. The composition and the properties of the above, along with other vegetable butters, were discussed in some detail (Hilditch & Williams, 1964; Sonntag, 1979). Table 4.10 illustrates the fatty acid compositions for the samples of the iodine value shown. When the crude oil has deteriorated, a free fatty acid content in the range of 12–25% may easily be met. A liquid/liquid extraction was used to remove the bulk of free fatty acid and partial glycerides. The semirefined oil of ca. 1.0% of free fatty acid

112

H.B.W. Patterson

may then be blended with similar natural or artificial vegetable butters, depending on availability, and alkali refining is used to produce an oil of 0.1% of free fatty acid and preferably under 0.01% of soap prior to bleaching. For example, the original crude shea butter may have color values approaching 70 Y 2.6 R 1.0 B (1”) and a phosphorous level of ca. 15 ppm. Both solvent and alkali deacidification will reduce these substantially prior to bleaching, with the phosphorous content possibly falling below 4 ppm. TABLE 4.10 Composition of Selected Vegetable Butters (Sonntag, 1979) Cocoa butter

Borne tallow

Shea butter

Mowrah fat (illipe fat)

36.7

33.2

59.1

63.9

C16:0

24.4

18.0

5.7

23.7

C18:0

35.4

43.3

41.0

19.3

Analysis Iodine value Fatty acid, wt%

C20:0

1.1

Saturated

59.8

62.4

46.7

43.0

C18:1

38.1

37.4

49.0

43.3

C18:2

2.1

0.2

4.3

13.7

40.2

37.6

53.3

57.0

Unsaturated

Recommendations A moderately acid-activated clay with a good capacity for removal of soap and residual gummy material also will give a satisfactory color. A contact time of 20 minutes at 90°C with moderate agitation is adequate. A dose of 1% (Fulmont 300C, Tonsil ACC FF, Engelhard F4, or Galleon NS) normally is adequate to bring the color below 15 Y 1.5 R (1”). For a cocoa-butter equivalent just prior to deodorization, a color of 6 Y 1 R (1”) is satisfactory. If the filtered oil clarity/color is disappointing, as an alternative to employing a larger clay dose, say 1.5%, the option of employing two successive clay additions, first 0.4% then 0.6% after an interval of 10 minutes, may succeed. (See section General Principles in Chapter 4.)

Linseed Oil (Flax) One variety of flax gives a high yield in seed and another in fiber. Dual-purpose varieties are said to give inferior seed and inferior fiber. The oil content of the seed is around 40%; the dominant fatty acid is linolenic acid, which may reach 62%. The plant thrives in warm temperate or subtropical zones. A range in IV of 130–205 exists, and this is heavily influenced by variety and climate. Table 4.11 shows a typical fatty

113

Bleaching of Important Fats and Oils

acid composition. With high unsaturation, the oil is used as a drying oil in the paint industry and in the production of linoleum and printing inks. Hydrogenated oil has been used for edible purposes and soapmaking (Patterson, 1983). Since gums and waxes are both present, settling crude oil for up to three weeks drops some 1% of the crude oil into the foots. Traditionally, the decanted oil’s temperature is then raised to 110°C, when as much as another 0.5% drops out. A free fatty acid of 2% is normal for separated crude oil, with a maximum of 1.5% of unsaponifiables. The degumming, neutralization, and washing procedures are very similar to those used with soybean oil. Chilling may succeed in depositing an additional small amount of wax. A problem with hardened linseed oil is that the hydrogenation of linolenic acid followed by oxidative breakdown is likely to produce small amounts of trans 6-nonenal (Keppler et al., 1965, 1967), the so-called linolenic hardening flavor (LHF). This nauseous sweet flavor is detectable at 0.0003 ppm (Patterson, 1983). The use of the hydrogenated oil for edible purposes or soapmaking is of decreasing importance, and even as a drying oil, linseed now suffers competition from the petrochemical industry.

Recommendations As this oil is very prone to oxidize, obviously an advantage is to degum, neutralize, and wash it as promptly as possible after milling/extraction. Alkali refining of a good crude is capable of producing an oil of only 0.1% of phosphatides; poorer crude may increase this substantially and thereby increase the burden on bleaching. A typical crude oil is 40 Y 4.0 R (1”), and an acceptable maximum is 70 Y 6 R (1”). Manufacturers clearly indicate which of their activated clays have the highest capacity for improving difficult-to-bleach oils, and linseed generally falls into this class. Naturally, the least doses are required when using the most active clays. If, however, the initial quality of the crude oil is good and the product specification is not the most demanding with respect to color, a fairly heavy dose of moderately activated earth may be adequate. Particularly active clays which are likely to be effective at 1.0–1.5% of dosage at 95°C and 20–30 minutes contact time include Engelhard NevergreenTM or F160, Galleon V2 Super, Fulmont XMP2, and Tonsil Supreme. Other active clays produced by these manufacturers which are likely to perform well TABLE 4.11 Fatty Acid Composition of Linseed Oil Fatty acid

a

%

Fatty acid

C16:0

6

C18:0

4

C16:1

trace

C18:1

22

C20:0

trace

C18:2

16

C20:1

trace

C18:3

52a

Variation in linolenic acid (%), 35–62.

%

114

H.B.W. Patterson

(although at a higher dose level) on good oil include Engelhard F105 or F105 SF, Galleon N or NF2, Fulmont AA or Premiere, and Tonsil L80 FF or Optimum FF. The color to be expected on the bleached oil is lighter than 30 Y 3 R (5¼”); for some applications, a 20 Y 1.5 R (1”) maximum is allowable. The hydrogenation of the refined, bleached oil further lightens the color; to hydrogenate it with a minimal delay after bleaching is advantageous. After the oil receives a post-hardening light bleach, a color as light as 1 Y 0.2 R (1”) can result. (See section General Principles in Chapter 4.)

Corn Oil (Maize) Over 80% of the oil in the corn grain is concentrated in the germ (embryo); the germ itself accounts for only around 5% of the whole grain. The corn germ, when manually separated from the grain, contains some 33% of oil. When the germ is separated by the dry-milling process, however, enough endosperm is entrained with the germ to lower the overall oil content to 20–25%. Alternatively, one may soak the grain in water, and the germ is separated by the so-called wet milling. This latter procedure also extracts some soluble material such as sugar, starch, and protein from the germ at the same time so that when it is separated and dried again to about 3% of moisture, the oil content stands close to 50%. In this situation, the most economical action is to solvent-extract dry-milled grain, whereas wet-milled grain may be fullexpelled (89–94% of oil recovery), or it may be pre-expelled and then the cake is solvent-extracted (combined oil yield 97–99%) (Strecker et al., 1989). The wet-milled material produces the better-quality oil. Corn oil is another member of the oleic–linoleic family, as Table 4.12 shows. Not only is corn oil particularly stable because of its very low linolenic-acid content; it also contains up to 1200 ppm of tocopherols and some esters of ferulic acid (Fig. 4.2), whose phenolic structure confers antioxidant properties. In addition, Strecker et al. (1989) and Mounts (1987) pointed out that, in corn-oil triglycerides, 98% of the sheltered 2-positions are occupied by unsaturated fatty acids (i.e., a nonrandom distribution), and that the remaining fatty acids, saturated and unsaturated, are left to occupy the more exposed 1- and 3-positions. Therefore, not surprisingly, after interesterification, when fatty acid distribution is randomized, the oxidative stability of the interesterified oil is only about one-quarter that of the natural corn oil (Lau et al., 1982). Corn oil melts at 11°C, so it is excellent for salad oils. But to ensure clarity, the small amount of wax present (ca. 0.05%) and a small amount of associated disaturated triglyceride must be removed by a winterizing step sometime during refining. A wide range of 103–133 is in iodine value, and unsaponifiables are usually less than 2.0%, although the Codex permits a maximum of 2.8%. Phosphatides range from 1 to 3%, and are removed by degumming methods employing phosphoric or citric acid which resemble those used with soybean oil. Free fatty acid between 1.5 and 4.0% is common in crude oil, but much higher figures are treated. Degumming, alkali neutralization, and washing have long been the normal methods of removing free fatty acid, phospholipids, soap, and trace metals to a level

115

Bleaching of Important Fats and Oils

Fig. 4.2. Ferulic acid.

which made the subsequent task of activated-clay bleaching relatively simple before deodorization. Forster and Harper (1983), however, reported that, provided the wetdegumming process is used, even a high (8%) free fatty acid dry-milled crude corn oil can be made fit for physical refining by active earth bleaching. Dewaxing may be included, and the same technique can be applied successfully to sunflower oil [see also (Young, 1989)]. Table 4.13 shows typical data at various process stages. Leibovitz and Ruckenstein (1983) developed their own special wet degumming (SWD) procedure, followed by a 2% activated clay bleach prior to physical refining. This SWD avoided the earlier need to give the oil a further alkali refining after the first steam stripping of the free fatty acid. They chose to place the winterizing step immediately after bleaching; therefore, working with a dry degummed oil made this step easier and Strecker et al. (1989) do likewise. Of the eight examples quoted by Leibovitz and Ruckenstein (1983), only two are shown in Table 4.14 to illustrate the extent to which a flexible procedure can still produce very useful oil from inferior feedstock. Yields were fully competitive or better than what might have been obtained from conventional alkali refining. Superdegumming (see section Degumming in Chapter 1) may also be applied to corn oil (Segers, 1982, 1983, 1985; Segers & van de Sande 1989). TABLE 4.12 Fatty Acid Compositions of Maize (Corn) Oil Fatty acid

Typical%

C12:0

<0.3

C14:0

<0.3

C16:0

10.0

C16:1

0.5

<0.5

C18:0

3.0

0.5–4

C18:1

33.0

24–42

C18:2

52.0

32–62

C18:3

0.5

9–14

<2.0 b

C20:0

trace

C20:1 trace

C24:0 Mandatory 1987 (Mounts, !987).

<1.0 <0.5

b

C22:0 a

Codex range (%)a

<0.5 <0.5

b

Trace = 0.05%.

116

H.B.W. Patterson

TABLE 4.13 Physical Refining of Maize Oil (Forster & Harper, 1983) % Free fatty acid

% Phosphorus

Color Y/R

5.24

0.033

40/4.4 (1”)

5.30

0.0012

36/4.2 (5¼”)

0.0011

35/4.0 (5¼”)

Feedstock Degummed Bleached

0.005

Dewaxed Steam-refined

0.03

34/2.5 (5¼”)

Recommendations As with other oils, starting from a good-quality crude already affords a particular advantage when seeking the highest specification for the finished product. The principal pigments in corn oil are xanthophyll and carotene; little or no chlorophyll is present. The nonhydratable proportion of the phosphatides is also small. Crude oil, especially from dry milling, may reach 70 Y 7.0 R (1”). A dose of 0.75 to 2.0% of well-activated clay such as Tonsil Optimum FF, Engelhard F105, Galleon N or NF2, and Fulmont AA or Premiere at a temperature of 105°C for at least 20 minutes should produce a color under 35 Y 3.5 R (5¼”), and probably much lighter from a good wet-milled oil. One must note that alkali neutralization followed by earth bleaching, or earth bleaching performed before physical refining, was demonstrated by various investigators (Strecker et al., 1989) to destroy completely any aflatoxin derived from moldy grain. (See section General Principles in Chapter 4.)

TABLE 4.14 Processing Maize (Corn) Oil Dried milled oil (solvent extraction)

Wet milled oil (prepressingsolvent extraction)

Crude oil FFA%

8.4

2.1

P (ppm)

391

285

Y 70 R 7.4

Y 40 R 5

2.9

2.85

0.07

0.03

Lovibond (1”) After SWD and bleaching (2%) P (ppm) After physical refining FFA% PV (meq/kg) Lovibond

0

0

Y 20 R 2 (1°)

Y 15 R 1 (5¼°)

Bleaching of Important Fats and Oils

117

Olive Oil The olive tree flourishes in warm temperate climates free from prolonged frosts, and is predominantly associated with the Mediterranean. It has been successfully introduced into California and Argentina. Over the centuries many varieties were developed. Ripe fruit contains 10–60% of oil (wet basis) or 35–70% (dry basis). Harvesting of fruit is confined to about four months, and processing (removal of light trash by air, then washing to remove dirt) must commence within three days to avoid fermentation, rising free fatty acid, and so forth. The olives are crushed in a plant which contributes nothing to metallic contamination, and then mixed to an emulsion-type paste, at which point oil may separate from the paste. Such oil is of the highest quality. The paste is then pressed to release more oil. If substantial amounts of oil have not already separated at the mixing stage, probably, two successive pressings at 60 kg/cm2 and 425 kg/cm2 will be used. The addition of inert agents such as talc to the paste enhances the efficiency of the separation of the oil from the solid; enzymes also achieve the same effect. As an alternative to pressing, the addition of hot water to the paste followed by the centrifugal separation of the oil is becoming popular because it improves quality and yield (Mendoza et al., 1989). Press residue or pomace may be pressed yet a third time, and finally, with an 8–10% oil content, it is solvent-extracted (olive-residue oil). Even this oil, when refined, may be incorporated into edible grades. But when carbon disulfide was used as a solvent, this gave rise to the terms “sulfur olive oil” and “olive-oil foots,” and such oil went to technical outlets. The phospholipid content of olive oil is only about 100 ppm. Fideli (1983) mentions three different methods of refining according to the quality of the crude oil: (i) conventional centrifugal refining; (ii) neutralization and separation of the soapstock by the selective action of two immiscible solvents; and (iii) degumming, bleaching, and steam stripping most of the free fatty acid, leaving the remainder to be removed by conventional alkali refining, ending in deodorization. Winterization of the oil has been practiced since Roman times. The Codex Stan. 33-1981 makes it clear that only oil resulting from the first separation from the olive is to be classed as virgin oil, and this must not have been processed in any way other than to filter it to ensure clarity. Specifications for refined olive oil and refined olive-residue oil follow, a cardinal point in each case being that refining should be “by means of methods which do not lead to alterations in the initial glyceride structure” The fatty acid structure put forward is as shown in Table 4.15. The additional proviso in Supplement 1 (1983) is that the amount of saturated fatty acids (effectively, palmitic and stearic) at position 2 of the triglycerides shall not exceed 1.5% for virgin oil, 1.8% for refined oils, and 2.2% for refined residue oils, of the total fatty acid of the oils concerned. As would be expected, the oil has outstanding oxidative and flavor stability. For refined oils, the iodine values range from 75 to 94 since, as with other oils, climate and species have a marked influence: oil from cooler regions tends to have a higher iodine value. Refined oil should have no more than 1.5% of unsaponifiables, except that refined olive-residue oil is allowed 2.5%

118

H.B.W. Patterson

TABLE 4.15 Fatty Acid Composition of Olive Oil—Codex Stan. 33-1981 Fatty acid

%

Lauric

C12:0

not detectable

Myristic

C14:0

0–0.05

Palmitic

C16:0

7.5–20.0

Palmitoleic

C16:1

0.3–3.5

Stearic

C18:0

0.5–3.5

Oleic

C18:1

56.0–83.0

Linoleic

C18:2

3.5–20.0

Linolenic

C18:3

0.0–1.5

Arachidic

C20:0

minute only

Gadoleic

C20:1

minute only

Behenic

C22:0

minute only

Erucic

C22:1

not detectable

Lignoceric

C24:0

minute only

maximum. Whereas for virgin oil, maxima for E232/270 = 3.50/0.25 are defined; for refined oil, the maximum E270 becomes 1.10; and for refined olive-residue oil, the maxima E232/270 = 6.00/2.00. Blends have intermediate maxima. With the exception of virgin oil, tocopherol may be added to other grades so as to restore a total tocopherol content of 200 ppm. Numerous other standards are given in the Codex. Most olive oil is used for cooking and table oil, and a small proportion for toilet soap. Very little is hydrogenated since in most circumstances no economic incentive is indicated, but when this is done to medium slip points (Patterson, 1983), the results are typical of the oleic–linoleic class of oils.

Recommendations To improve clarity and to lighten the color of darker crude oils and, therefore, provide a component which may be more easily blended, various bleaching clays are available. Tonsil 13 and Tonsil 13C (includes 5% of carbon) and Engelhard F4 are offered for use with olive-oil. Fulmont grades from Premiere to XMP2 are suitable, and the Galleon grades such as V2, DC, and the adsorbent Mizukalife F-1 also should be effective. The size of the clay dose of 0.5–2.0% depends on starting and finishing color. A contact time of 20 minutes at ca. 90°C is normal. The E232/270 maxima mentioned above must be noted (see also section Oxidized Fats and Oxidized Fatty Acids in Chapter 1 and section Groundnut (Arachs, Peanut Oil). (See section General Principles in Chapter 4.)

Bleaching of Important Fats and Oils

119

Palm Oil The present annual world production of palm oil is ca. 9 million tons. Two striking features of the history of this oil are: (i) the phenomenal growth in production in Malaysia since the early 1970s, where considerably more than one-half of the world’s production now takes place and (ii) through the logical development of what we know as physical refining applied to this oil, the same technique was steadily applied to many other oils which previously were refined by classical alkali (i.e., chemical) processes. The oil palm Elaeis guineensis was introduced from West Africa to East Asia in 1848 for ornamental reasons. The original mother oil-palm tree still grows in the Bogor Botanic Gardens, some miles inland from Jakarta. Nigeria and Indonesia have become net importers of palm oil. Only a tiny fraction of world production now takes place in Zaire where, early in the twentieth century, W.H. Lever (later Lord Leverhulme) rented territory from Belgium for one centime per hectare to develop the extensive plantation cultivation of the oil palm (Huileries Congo Beige). Observing the decline in demand for natural rubber and in revenue from tin mining, the Malaysian government resolved to back oil-palm planting. Research initiatives by the Unilever company greatly enhanced the fertilization of female palms. Of even greater importance is that Unilever Research succeeded in cloning a high-yielding variety. By early in the next century, a worldwide annual production of palm oil of over 21 million tons is forecast (Jones, no date; Law & Thiagarajan, 1989; Patterson, 1989). Apart from Elaeis guineensis, some hundreds of species of palm tree exist. The oil palm prefers a humid atmosphere within 10 degrees of the equator. Palm oil makes up ca. 56% of the pulp of the fruit, and in the kernel is some 47% of a very different type of oil, much more akin to coconut and babassu oils (see section Palm Kernel Oil in Chapter 4). Palmitic and oleic acids are the dominant fatty acids of palm oil. Their relative proportion to one another influences the character of the oil derived. Thus, West African palm oil typically melts at about 35°C, whereas Malaysian palm oil usually melts at 35–38°C. Table 4.16 shows typical fatty acid compositions of palm oils from different regions along with the Codex Stan. 125-1981 advisory standard. The sequence in which the fatty acids are grouped in the various triglycerides and what proportions of these triglycerides are present have been the subjects of years of investigation. The following is evident: 6–9% of trisaturates constitutes less than a random distribution; oleodi-palmitin and dioleopalmitin make up about one-half of the triglycerides; and linolenic acid is virtually absent. The composition of palm oils of different national origins (Rossell et al., 1985) and the nutritional value (Clegg, 1978) were reviewed in detail. An oil of such composition has high inherent oxidative and flavor stability if properly handled. However, a powerful hydrolytic enzyme is present on the exterior of the palm fruits, and if this is not deactivated promptly after harvesting, any oil escaping from damaged fruits or during processing splits rapidly to free fatty acid. Thus, prior to the 1970s, crude oil approaching 50% of free fatty acid was classed as hard and fit only for soapmaking and other technical outlets; oils around 12% of free fatty acid were known as soft, and could certainly be refined for

120

H.B.W. Patterson

TABLE 4.16 Fatty Acid Composition of Palm Oils Fatty acid

Nigeria (%)

C12:0

Zaire (%)

Sumatra (%)

Codex Stan. 125 (%)

trace

trace

<1.2

C14:0

1

1

1

0.5–5.9

C16:0

40

44

46

32–59

trace

trace

<0.6

C16:1

0.5

C18:0

5

5

4

1.5–8.0

C18:1

40.5

40

39

27–52.0

C18:2

12

10

10

5.0–14

C18:3

0.5

C20:0

0

<1.5 0.5

trace

<1.0

edible use. An intermediate class of oil around 35% of free fatty acid was classed as semisoft. Standards have improved immensely since those times. As has been recognized for several years, if freshly expelled crude palm oil is promptly vacuum-dried and then cooled from ca. 110 to 55°C, hydrolysis and oxidation are much less. Crude oil can be produced at less than 2.5% of free fatty acid and shipped at 0.15% of H2O, at which moisture level neither hydrolysis nor oxidation is particularly favored (15,214); nitrogen blanketing of ships’ storage tanks during a voyage retards oxidation. The addition of tertiary butylhydroquinone (TBHQ) (110 ppm) and citric acid (365 ppm) combined is also very helpful (Anon., 1978; Patterson, 1989; Sherwin, 1976), not only in assuring a better color in the final refined oil, but also in reducing the amount and, therefore, the cost of the bleaching clay needed, because the amount of oxidized crude oil present was made substantially less. Furthermore, Coppen (1983) noted that the early addition of TBHQ to crude oil results in a superior quality of refined and deodorized oil, even though the TBHQ may be removed by the final deodorizing. Along with the improvement in the free fatty acid content of crude oil on offer, now accepted is that other characteristics should have desirable values as low as peroxide value of 4 meq/kg, anisidine value of 4, Fe of 3 ppm, and Cu of 0.1 ppm (Jacobsberg & Jacmain, 1973; Johansson, 1975; Landon, 1975). In fact, an upper limit of Cu of 0.05 ppm is desirable since as little as 0.02 ppm is capable of encouraging oxidative attack. A high red color [e.g., 26 Y 26 R (1”)] indicates the carotene present—which may be as high as 0.2% (2000 ppm) of the crude oil—is largely undamaged and, therefore, readily amenable to heat-bleaching (i.e., thermal destruction). Figure 4.1 (Macellan, 1983) shows clearly how residual carotene drops more and more rapidly as the treatment temperatures rise above 220°C (see also section Heat Bleaching in Chapter 1). Hiscocks & Raymond (1964) and Macellan (1983) have remarked that the oxidation of carotene proceeding in the presence of some free fatty acid may cause the resultant brown color which is so difficult to remove. The up-to 0.1% of tocopherols

Bleaching of Important Fats and Oils

121

and tocotrienols present in palm oil provides potent scavengers of oxygen-free radicals (Law & Thiagarajan, 1989), but a high-carotene content with its prooxidant behavior at that level may overcome their influence until after neutralization, bleaching, and deodorization are performed. By then, the carotene level has fallen greatly and, consequently, the oxidative stability rises (Meara & Weir, 1976). Johansson (1975) emphasizes that light-colored crude palm oil suggests some carotene was oxidized, and therefore, preferably, one can assess the concentration of undamaged carotene present by spectrophotometric analysis. Palm oils of different origins and the use of different solvents in measuring the peak absorptions give somewhat different characteristic values. Normally, a peak close to 450 nm using n-hexane as solvent exists, and multiplying by an empirical factor of 427 was used to express the concentration of carotene in ppm (Patterson, 1983) (see also section Adsorption Efficiency and Variation in Chapter 1). At one time, some users were accustomed to check peak E458 of crude Nigerian palm oil in chloroform as a solvent, and they considered the readings of 2.5 to 3.0 to be quite satisfactory, but below 1.5 indicated a poor quality. At the same period, crude Malay sian palm oil showing readings of 1.5–1.8 performed well on refining. More recently the deterioration of the bleach ability index (DOBI) found favor (Law & Thiagarajan, 1989). In this test the crude palm oil is dissolved in isooctane or n-hexane, and the adsorbance at 269 µm and 446 nm is checked. DOBI = Adsorbance at 446 nm DOBI = Adsorbance at 269 nm The result is expressed to two decimal places. An arbitrary scale based on experience suggests: DOBI = 3.25 and above—very satisfactory = ca. 2.75—average quality = below 2.0—very poor oil Phosphatides in crude palm oil are usually less than 0.1%, and sulfur is negligible. The typical iodine value is 54 with a range of 52–58; unsaponifiables are normally under 0.8%. As indicated above, in the 1980s the condition of crude oils on arrival at the end of a lengthy sea voyage greatly improved; a peroxide value of 4 and an anisidine value of 4 are both feasible (Gapor & Ong, 1982; Johansson, 1975; Willems & Padley, 1985). When the alkali refining of crude palm oil was the universal method, numerous procedures were in use by different refiners, and these were influenced by the condition of the crude oil (Swoboda, 1985). The aim was to obtain a soapstock which separated readily from the oil when using the minimal excess of alkali so as not to lose neutral oil by saponification and emulsification. In batch procedures, initially deaerating the oil, avoiding undue local excess of added caustic-soda solution by efficient dispersal, maintaining a temperature of ca. 90°C, and, in some procedures, adding brine secured efficient neutralization. Since phosphatides were so low,

122

H.B.W. Patterson

Fig. 4.1. Thermal destruction of E-carotene (Macellan, 1983 )

degumming was not essential (Willems & Padley, 1985), but was sometimes used. Phosphoric acid or citric acid was an effective degumming agent, and subsequently, in centrifugal refining, the same option of using a small addition of phosphoric acid (0.05%) prior to neutralization served the same purpose, making more certain that the nonhydratable phosphatides were disrupted, hydrated, and then expelled along with the soapstock. After drying, what remained was to remove/disrupt pigment, trace metals and any remaining sulfur, phosphorus, or oxidized compounds on an efficient adsorbent. After deodorization, preferably at a temperature a little above 220°C, oils of reliable flavor stability resulted. For the bleaching step, the following is appropriate.

Recommendations The earth dosage at one time might have exceeded 2%, but with the earths available in the 1990s, less than that would be adequate for many purposes, and 1–1.5% would suffice for the removal of both pigment and impurities. Similarly, although bleaching temperatures up to 140°C were envisaged as acceptable on occasion, even for products with an edible end use, a safer action is to keep within an upper limit of 110°C when highly active clay is also present (Zschau, 1997, 1990). This minimizes

Bleaching of Important Fats and Oils

123

the probability of physical as well as chemical changes arising in the triglycerides. Palm oil is, of course, less inclined to these than more unsaturated oils. A contact time of 20–30 minutes is adequate. Following chemical refining, Fulmont 237 was suggested, but now Fulmont AA or Premiere would probably be chosen, especially for poorer quality crude oils. The physical refining certainly requires Fulmont AA or Premiere following the addition of phosphoric or citric acid in the presence of small amounts of water (Young, 1989). Similarly, Tonsil Optimum FF and Supreme FF are the most powerful Süd-Chemie adsorbents for palm-oil treatment; less demanding tasks would be met by the range Tonsil ACC FF/Standard FF/L80FF. Engelhard suggests F105 for palm-oil bleaching, but the more powerful F160 would be the choice for dealing with poorer qualities. Galleon grade V2 is put forward for palm-oil bleaching, but grade V2-Super is specifically mentioned if physical refining is to be used (see section Degumming in Chapter Basic Components and Procedures). Where physical refining failed to secure adequate flavor stability, refiners would adopt a routine of following the high-temperature steam deacidification with a light chemical refining and final conventional (180°C) deodorization. This is expensive and, particularly as the quality of crude palm oil has improved, the need for it has diminished. For many purposes a fully refined oil color of 20 Y 2.0 R (5¼”) to 28 Y 2.8 R (5¼”) is adequate (Pritchard, 1983). From a first-class crude-oil degumming, chemical refining concluding with deodorization (220–240°C), or physical refining (270°C) yields a color of 10 Y 1.0 R (5¼”). If the expense is justified, a combination of both refining techniques can yield an even lighter color (Patterson, 1989). As mentioned in the discussion of heat bleaching (see section Heat Bleaching in Chapter Basic Components and Procedures), the formation of PAH in the thermal bleaching/physical refining at 270°C for a few minutes or at 220–240°C for two hours has been very thoroughly checked (Biernoth & Rost, 1968; Rost, 1976) and was shown to be negligible; it is clearly no more than with other oils such as soy, corn, and peanut. Much palm oil is fractionated (Thiagarajan, 1989). The olein is easier to bleach; about 1.5% of moderately activated earth is enough. Many of the impurities carry over to the stearin fraction, so as much as 2% of highly activated earth may be needed to bleach it; a contact time of 30 minutes at 100°C followed by deodorization at 260–270°C applies in each case. (See section General Principles in Chapter 4.)

Palm Kernel Oil The kernel of the oil-palm fruit usually contains 46–48% oil (dry basis), but as high as 53% has been met. As palm-oil production rises, so does that of Palm Kernel oil, which amounts to about 12% of the former (Graalmann, 1989). Table 4.17 shows how the fatty acid composition of Palm Kernel oil compares with that of other lauric oils. Although Palm Kernel oil’s content of C6, C8, and C10 saturated fatty acids is somewhat less than that of coconut oil, it contains more oleic acid. This means it shows a greater change when hydrogenated (Patterson, 1983).

124

H.B.W. Patterson

Typically, coconut oil melts at 25°C and Palm Kernel oil at 26°C, but different parcels show overlapping slip points. As these and other lauric fats melt sharply, leaving no fatty sensation or cling on the palate, they have been popular for a long time in the formulation of ice cream, margarine, hard butters, biscuit dough, and filling cream. When saponified, they produce a very soluble and quick-lathering soap, so they are useful components in toilet-soap production. The valued sharp-melting characteristic can be accentuated by: hydrogenating the Palm Kernel oil itself, fractionating then using the stearin fraction, and hydrogenating the olein (Patterson, 1983; Shukla, 1989; Young, 1983). As might be expected, because of their low degree of unsaturation, the lauric oils and their hydrogenated products have good oxidative stability, but on hydrolysis give a soapy off-flavor. For commercial practice, maxima peroxide value of 7 and anisidine value of 3 were proposed (Pritchard, 1983). Crude oil should not exceed Fe of 2 ppm and Cu of 0.2 ppm; these values should fall to Fe of 0.1 ppm and Cu of 0.02 ppm or even lower after refining. Phosphatides are very low and can be depressed still further by quite a mild phosphoric- or citric-acid treatment which, of course, also helps get rid of prooxidant metals prior to either alkali or physical refining; a dose of 0.05% w/w of citric acid added as a 50% aqueous solution at 70°C and vigorously dispersed is most popular (Young, 1989). The sulfur content is negligible. A maximal color for filtered crude oil of 20 Y 1.5 R (1”) is undemanding. In practice, poor crude oil may show 35 Y 3.7 R (5¼”), 10 Y 1.0 R (1”), while good crude may be no more than 22 Y 1.3 R (5¼“), 4 Y 0.4 R (1”). This depends mainly on the way the kernels were pressed/extracted, shipped, and stored. More and more are being processed in the country of origin, especially Malaysia. Obviously, the chance exists of carotene penetrating the thinner shells of certain kernels or by the breakage of kernels so that a normal content (Pritchard, 1983) of 4.4 ppm is exceeded. More troublesome is the reversion of color which may sometimes appear after deodorization. If this is simply due to prooxidant metals, these are effectively removed by more thorough citric-acid sequestering and active earth bleaching before the deodorization, as mentioned previously. As noted, reversion is more common when the oil was stored for about half a year, sometimes above 40°C, or the oil was extracted from very old kernels. This may well call for a heavier bleaching of the oil. Whereas a good crude may have a peroxide value of 2, a badly oxidized oil may reach a peroxide value of 25. A final deodorized oil color of 10 Y 1.0 R (5¼”) maximum may be acceptable for many purposes; on the other hand, less than 4 of Y 0.4 R (5¼”) is quite feasible from good feedstock.

Recommendations A contact time of 20 minutes at 90°C is normal. With good-quality feedstock, a dose of up to 0.4% of Galleon NS, Fulmont 237, Tonsil Standard FF, or Engelhard F4 is appropriate. With either poorer feedstock or a lighter final color specification, a slightly heavier dose of a more active earth such as Galleon NF2, Fulmont 700, Tonsil

TABLE 4.17 Fatty Acid Composition of Palm Kernel and Other Laurie Oils Babassua

Fatty acid Caproic (%)

C6:0

Coconutb

Cohune

Murumura

Palm kernelc

Tucum

<1.2

0.3

0.1

<0.5

0.2

Caprylic (%)

C8:0

2.6–7.3

3.4–15

8.7

1.3

2.4–6.2

2.9

Capric (%)

C10:0

1.2–7.6

3.2–15

7.2

1.5

2.6–7.0

2.3

Undecanoic (%)

C11:0

Lauric (%)

C12:0

40–55

41–56

47.3

46.2

41–55

51.8

Myristic (%)

C14:0

11–27

13–23

18.2

32.4

14–20

22.0

5.2–11

4.2–12

7.7

5.6

6.5–11

6.8

1.3–3.5

2.3

C16:0 C16:1

Stearic (%)

C18:0

0.1 1.8–7.4

1.0–4.7

3.2

2.2

Oleic (%)

C18:1

9.0–20

3.4–12

8.3

6.9

10–23

9.3

Linoleic (%)

C18:2

1.4–6.6

0.9–3.7

1.0

1.5

0.7–5.4

2.4

Arachidic (%)

C20:0

Iodine value

10–18

6–11

8–14

8–13

13–23

10–14

Saponifiable value

245–256

248–265

250–260

237–247

230–254

240–250

0.2

Bleaching of Important Fats and Oils

Palmitic (%) Palmitoleic (%)

0.1

a

Codex Stan. 128-1981.

b

Codex Stan. 124-1981.

c

Codex Stan. 126-1981.

Remainder (Thomas & Paulika, 1975). 125

126

H.B.W. Patterson

L80FF, or Engelhard F105 may be necessary. For oil that is to be hydrogenated, a slightly higher color is acceptable. (See section General Principles in Chapter 4.)

Rapeseed Oil (Colza) For 4000 years, oils from plants of the Brassica family were highly appreciated, more particularly in Asia, for their spicy flavor. This arises from sulfur compounds including glucosinolates and about fourteen others (Thomas, 1982). Traditional varieties of the plant are variously known as turnip rape, colza ravison, sarson, and torea. Summer and winter crops are known, and flourish in cool temperate climates. The oil content of the undried seed is ca. 40%. The fatty acid composition of traditional rapeseed oil includes an unusually high (ca. 50%) proportion of the long-chain erucic acid (C22:1), giving the oil valued lubricant properties but causing it to fail cold tests if it is intended to be used as a table oil. Since the early 1970s, when Canadian scientists recognized a rapeseed cultivar of particularly low erucic- acid content (Bronowski specimen), a new variety of rapeseed oil has evolved in stages. Today, a much-reduced erucic-acid content (LEAR) of under 3% readily allows the oil to be used for the production of table oil, even without being winterized, but the old variety (HEAR) still has a certain market for industrial use. The low erucic-acid content also avoids the suspicion that this fatty acid encourages heart disease. Furthermore, the natural glucosinolate content has been vastly reduced, falling below 3 mg of glucosinolate/g of dry meal (Ackman, 1983; Anon., 1981) and then to below 2 mg/g or 18 µmLmol of glucosinolate expressed as 3-buterylisothiocyanate (Thomas, 1982). Typically, the new oil contains only 3–5 ppm of sulfur after water degumming and 2 ppm after alkali refining (Shukla, 1989). The traditional rapeseed oil contained ca. 60 ppm of sulfur, which frequently remained at ca. 10 ppm after alkali refining and active clay bleaching. Today, the new rapeseed meal after the extraction of the oil no longer poses the dietary hazard to poultry, pigs, and young ruminants that it once did. Sulfur is no longer present in the oil at a level which seriously poisons nickel catalyst and hinders hydrogenation (Patterson, 1983). Thus, we have the so-called double-zero rapeseed oil now known as canola. This offers the Canadian farmer a second crop to wheat and by 1990–91, the enormously increased production in France, Germany, the United Kingdom, and Denmark had also been completely converted to the “00” type. The preceding paragraph is a highly condensed general account of a complex and very successful development program already reviewed in more detail (Patterson, 1989); particular facets are considered in other references (Carr, 1989; Forster & Harper, 1983; Mag, 1983; Ohlson, 1983, 1989; Patterson, 1983; Pritchard, 1983; Taylor & Ungermann, 1987; Thomas, 1982; Young, 1989; Zschau, 1990). The fatty acid compositions of different brassica oils are shown in Table 4.18. In Table 4.19 (Taylor & Jenkins, 1901), some of the typical trace constituents of soybean, canola, and palm oils are compared. The maximum of 0.1% of gums quoted for canola evidently applies to a crude oil on which the producers have already

127

Bleaching of Important Fats and Oils

TABLE 4.18 Fatty Acid Composition of Brassica Oils Typical traditional rapeseed oil (Patterson, 1989) (%)

Low erucic-acid rapeseed oil Codex Stan. 123-1981a (%)

Mustardseed oil Codex Stan. 34-1981 (Suppl.) (%)

C <14

—

<0.1

<0.5

Myristic

C14:0

0.5

<0.2

<1.0

Palmitic

C16:0

3.5

2.5–6.0

0.5–4.5

Palmitoleic

C16:1

—

<0.6

<0.5

Stearic

C18:0

1.0

0.9–2.1

0.5–2.0

Oleic

C18:1

13.0

50–66

8.0–23

Linoleic

C18:2

14.0

11–23

10–24

Linolenic

C18:3

9.0

5–13

6.0–18

Arachidic

C20:0

1.0

0.1–1.2

<1.5

Eicosenoic

C20:1

7.5

0.1–4.3

5.0–13

Eicosadienoic

C20:2

1.0

—

<1.0

Behenic

C22:0

0.5

<0.5

0.2–2.5

Erucic

C22:1

47.5

<5.0b

22–50

Docosadienoic

C22:2

1.0

—

<1.0

Lignoceric

C24:0

0.5

<0.2

<0.5

Tetracosenoic

C24:1

—

—

0.5–2.5

a

As made mandatory 1987 (Mounts, 1987). bLess than 2% desirable but not mandatory.

performed appreciable degumming [see also (Young, 1989)]. Ackman (1983) quotes an D-tocopherol content of 190 ppm, and this boosts the antioxidant components. Because the proportion of C18 chain-length fatty acids is high (over 90%), this gives rise to crystallization difficulties if shortening or margarine fats are formulated with a very high proportion of hydrogenated and soft canola oil. This is also true of other oils where the great majority of fatty acids are of the same chain length, sunflower and soybean oils being well-known examples. E-phase crystals (25–100 microns) develop within a few weeks rather than the desired E’ (beta prime ) crystals (less than one micron), and this causes sandiness or grittiness in the texture. The greater the degree of hydrogenation, the more noticeable is this defect. Happily, the inclusion in the formulation of 15%, or in some cases less, of some other oil rich in fatty acids of a different chain length (e.g., cottonseed or palm oil) overcomes the difficulty. Swedish plant breeders have even succeeded in growing a variety with 10% of palmitic acid already in the oil composition (Anon., 1981; Carr, 1989; Mag, 1983; Patterson, 1983; Teasdale, 1981; Thomas, 1982). A comparison of typical approximate

128

H.B.W. Patterson

degrees of unsaturation in canola-oil fatty acids and those of some other important vegetable oils is instructive (Table 4.20). Although the 10% of linolenic acid in normal canola oil may contribute to oxidative instability as in other oils, its location in the 2-position of the triglycerides protects it to a degree (Carr, 1989). A variety of canola oil with linolenic acid reduced to less than 2% was produced by plant breeders. TABLE 4.19 Typical Trace Constituents in Some Crude Vegetable Oils (Taylor & Jenkins, 1901) Type of vegetable oil Component

Soybean

Fatty acids (free) (%) Phosphatides (gums) (%)

Canola

Palm

0.3–0.5

0.3–0.5

2–5

1–3

0.05–0.1

0.05–0.1

Triterpene alcohols/sterols (%)

0.4–0.7

0.4–0.6

0.1–0.2

Tocopherols (natural antioxidants) (%)

0.06–0.2

0.06–0.1

0.06–0.1

25–40

25–70

500–2000

1–2

20–30

nil

Carotenoids (color pigments) (ppm) Chlorophyll/phenophytin (color pigments) (ppm)

TABLE 4.20 Typical Approximate Fatty acid Compositions

Saturated (%)

Canola

Olive

Soybean

Palm

6

16

15

55

64

71

22

34

Diunsaturated (%)

20

12

54

10

Triunsaturated (%)

10

1

9

0

Monounsaturated (%)

Canola oil has an iodine value range of 110–126 as against 94–120 for HEAR. Codex Stan. 123-81 for LEAR gives a maximum peroxide value of 10, but a very good commercial oil may be as low as a peroxide value of 1.0 and an anisidine value of 0.8. Again, the Codex Stan. 123-81 gives a maximum Fe of 1.5 ppm and Cu of 0.1 ppm, but substantially less may be demanded. In any case, degumming and heavy activated clay bleaching employed to remove chlorophyll effectively control trace metals at the same time. Although processors of crude canola have had the option of various combinations (Carr, 1989) of degumming, alkali refining, heavy bleaching, deodorization, acid/water degumming, heavy bleaching, and physical refining, much emphasis has come to be placed on the removal of chlorophyll derivatives to a maximum of 50

Bleaching of Important Fats and Oils

129

ppb, not merely to ensure a good color but also to minimize the catalytic initiation of photochemical oxidation leading to flavor instability. These aspects were mentioned in previous sections of this publication (see sections Use of Silica, Heat Bleaching, and Dosage in Chapter 2). Seemingly, established is that super degumming (see section Degumming in Chapter 1), followed by active clay bleaching, makes physical refining reliable. The preliminary effort to adsorb remaining phosphatides, soap, and trace metals on adsorbent silica reduces the level of activated clay needed (see section Use of Silica in Chapter 1). The application of this technique—to canola oil in particular, although it also applies to soybean oil—was described in some detail (Toeneboehn & Welsh, 1990; Welsh et al., 1989). Working on a plant scale, the optimal technical efficiency is claimed by mixing some clay and silica with the oil flowing to a bleaching vessel, and then passing the slurry through a filter precoated with clay. An increasingly deep layer develops in the filter, thus achieving a maximum of press bleaching. This arrangement can be combined with conventional or physical refining.

Recommendations A poor-quality rapeseed oil may show a color of 40 Y 5 R (1”). Conventional alkali refining and activated clay bleaching drop the color to a 50 Y 5.0 R (5¼”) maximum. A further decrease after deodorization gives a final color between 20 Y 2.5 R (5¼”) and 10 Y 1.0 R (5¼”). A bleaching temperature of 105°C is popular with a contact time of 30 minutes. A trial will show if a prolongation of the contact achieves any lasting benefit. A dose of 2% of activated clay may be necessary, depending on the color of original oil, the effect of preliminary treatment, and the ability to achieve 50 ppb of chlorophylloids. A color of ca. 2.0 R (5¼”) prior to deodorization is feasible (see Fig. 2.2). The earths specially recommended by their manufacturers to achieve low chlorophylloids and carotenoids are Engelhard F160 and Filtrol NevergreenTM, Tonsil Optimum FF, Fulmont XMP2, and Mizusawa’s Mizukalife F-1, alone or jointly with Galleon DC, depending on the chlorophylloids needing to be removed. (See section General Principles in Chapter 4.)

Rice Bran Oil Rice of various kinds is one of the oldest and largest crops; it is found in the tropics and warm temperate zones. Some explanation of the terms used in connection with the crop and its processing will enhance understanding. Threshing produces the rice grains, “rough rice” or “paddy” A sequence of milling operations detaches and separates first the hull or husk (ca. 20%) and then the inner brown skin, or bran and the germ (ca. 10%), leaving ca. 70% for the production of the familiar polished white rice (head rice), plus any broken rice grains, “second head,” “screenings,” “brewer’s rice,” and any polishings. If only the hull is removed (hulled), the result is brown rice or cargo rice. Simple mills often remove the hulls and bran and return them to the grower. In the large mills, the bran is produced as a separate item, and this is the source of Rice Bran oil. According to the type of rice and degree of milling, the bran contains 9–23% of oil, the remainder being protein, fiber, and inorganic material.

130

H.B.W. Patterson

In Asia especially, the whole rice grain is often steeped in water, steamed, and then dried at a lower temperature. This is parboiled rice. It is exported as such and used locally in conventional milling. This has been the technique for preserving the quality of the rice, but the parboiled rice bran is the major by-product of the parboiledrice milling industry, and such bran is much more stable. Obviously, where bran is separated from the rice paddy without parboiling, the bran itself must be stabilized by steaming (Anon., 1985, 1987). If this is not done, the lipolytic enzymes present in the bran cause an extremely rapid increase of free fatty acid within hours, possibly 5% per day, in the warm and humid ambient conditions. Therefore, the best practice is to steam the freshly separated bran for a few minutes at 90–95°C and then dry it to ca. 4% of moisture and cool. The parallel with the prompt sterilization of oil-palm fruit bunches before the expulsion of the oil is evident. In spite of an annual production of well over 400 million tons of rice (paddy) per year with a potential production of 7 million tons of Rice Bran oil, the estimate is that only under 90,000 tons per year are produced for edible use. In contrast, in Japan, where the infrastructure in the rice fields exists to connect the facilities for sterilization of the bran, about two-thirds of the oil produced is available for edible use at home or export. Without such facilities, the oil, if recovered at all, is almost certain to be fit only for technical outlets as soap (Koenig, 1978). Mostly crude oil under 15% of free fatty acid is processed for edible use, with a possible upper limit of 20% of free fatty acid. Generally, high free fatty acid oil is darker in color, and processing has to take this into account (Bhattacharyya et al., 1983) lest it becomes unbleachable or fixed. Now, apparently, the most cost-effective way of extracting oil from bran is to pelletize it and solvent-extract the oil (Abhay Sah et al., 1983; Koenig, 1978). Usually, the wax content of rice bran oil approaches 9%, although extremes of ca. 20% were found. About one-half of this wax melts at 72–84°C, and provides a useful by-product (Harris, 1971; Lasztity, 1971). Up to 4% of phosphatides may be present (Abhay Sah et al., 1983). Both tocopherols (see Tocopherols and Chromans in Chapter 1) at ca. 0.47% and ferulic-acid esters (see Scheme 4.2) are present, and these, combined with the fact that less than 1% of linolenic acid is present in the triglycerides (see Table 4.21), contribute to considerable oxidative stability, especially when prooxidant trace metals and pigments were reduced to a very low level. A little carotene may contribute to the dark color of crude oil, and as much as 20 ppm of chlorophyll was found. Squalene content is between 0.2 and 0.5% (Harris, 1971; Reddi et al., 1948). The iodine value ranges from 92 to 115; there is no Codex Standard. The considerable potential of rice bran oil for use as a salad oil—when a final winterization has removed any remaining traces of wax and also has taken out the very small proportion of higher-melting triglyceride—can be gauged from the comparison between it and the fatty acid composition of several other well-known members of the oleic–linoleic family in Table 4.21. The winterization would be done just prior to deodorization, and the separated solid would be used in margarine. The most modern process for a high recovery of the oil, the X-M process, entails

TABLE 4.21 Fatty Acid Compositions of Some Common Soft Oils Compared with Rice Bran Oil

(in Percentage)

Fatty acid

Rice bran oil

Sunflowerseed oil

Corn oil

Sesame oil

Cottonseed oil

Groundnut oil

Soybean oil

Myristic

(14:0)

0.4–1.0

(1)

0–1.7

(0)

0

(0)

0

(0)

0.5–1.5

(1)

0–1

(0)

0–1

(0)

Palmitic

(16:0)

12–18

15)

8–12

(8)

3–6

(4)

7–9

(8)

20–23

(21)

6–9

(7)

7–11

(8)

Stearic

(18:0)

1–3

(2)

2–5

(4)

1–3

(3)

4–5

(4)

1–3

(2)

3–6

(4)

2–6

Olei

(18:1)

40–50

(45)

14–49

(46)

14–43

(34)

Linoleic

(18:2)

29–42

(35)

34–62

(42)

44–75

(57)

Linolenic

(18:3)

0.5–1.0

(1)

0

(0)

0

(0)

0

0

(0)

0

(0)

0–4

(1)

0–1

Arachidonic (20:0)

37–49

(45)

23–35

(29)

35–47

(41)

42–54

(45)

(0)

0–11

(3)

0

(1)

0.2–1.5

(1)

2–4

(4)

53–71

(62)

15–33

(28)

13–27

(23)

43–56

(54)

(0)

2–10

(5)

(3)

0–2

(0)

Bleaching of Important Fats and Oils

Range (typical analysis)

131

132

H.B.W. Patterson

mixing the whole rice (paddy) with a minimum of hexane and then separating the loosened bran, still in hexane, from the rice (Fedeli, 1983). Presumably, the lipase has already been deactivated. At 60–70°C, wax is easily soluble in the triglyceride/ hexane miscella, but at 20–25°C, it is practically insoluble. This fact permits one to reduce the proportion of waxes in oil (Harris, 1971). When desolventized, the rice grain is sold as such. The bran is next separated from the miscella; then the hexane is finally distilled. Universally agreed is that any refining of the rice bran oil must be preceded by dewaxing and then degumming. In the author’s experience, freshly extracted batches of oil left to stand in tanks at even ambient tropical temperatures deposit around one-half of their wax content. Sensibly, therefore, one should decant the oil and, if helpful, filter out any entrained solid. Subsequent deliberate cooling would cause further deposition of wax. Degumming follows, and is improved by the use of citric or phosphoric acid to disrupt nonhydratable phosphatides (see section Degumming in Chapter 1), after which hydrated gums can be separated centrifugally in the same way as with other oils. The efficiency of the subsequent refining process is related to the success of the preliminary purifying of the crude oil. Conventional batch alkali refining has frequently shown extremely high refining factors. Centrifugal refining has brought improvement. In cases in which the free fatty acid is high (ca. 15%), a steam distillation followed by a mild alkali refining was reported to give an economically low refining loss (Yokochi, 1979). As is evident from the fatty acid composition of Rice Bran oil (Table 4.21), if gums and trace metals were effectively eliminated by an activated clay bleaching, physical refining is practical (Sullivan, 1976). The separation and purification of Rice Bran waxes were reviewed in detail (Harris, 1971; Lasztity, 1971).

Recommendations In spite of the oil being dewaxed to a great extent, degummed and alkali refined, quite large doses of 3% and even 5% of activated bleaching clays were mentioned (Anon., 1987), especially in earlier reports. However, those amounts are not specifically associated with carotenoids; chlorophyll as high as 20 ppm seems possible (Rao & Murti, 1961; Reddi et al., 1948). A filtered crude oil of 35 Y 5.3 R (1”), which gave an alkali neutralization of 70 Y 4.4 R (1”) at best, was described (Reddi et al., 1948). The particularly active clays of the 1990s at a 3% dose rate and a contact time of 20–30 minutes at 100–110°C should ensure an adequate effect. Fully refined Rice Bran oil, commercially available in Japan, showed a color range of 30 Y 3.0 R (5¼”) to 20 Y 1.7 R (5¼”). Activated clays giving good results with Rice Bran oil are Galleon V2 and DC, Tonsil Supreme FF, Engelhard F160, and Fulmont XMP2. (See section General Principles in Chapter 4.)

Safflower Oil (Cartamo, Kusum) The cultivation of this plant, established for many centuries in Asia and Europe, has spread to the Americas and Australia. Originally, greater importance was attached

133

Bleaching of Important Fats and Oils

to the florets as a source of yellow/orange dye, with the seeds left to ripen on the plant, each of which carried up to 200 seeds on a thistle-like head. Seeds of the latest variety contain 48% of oil. Spines make manual harvesting difficult, but it can be done mechanically; care has to be taken to choose times when the natural habit of shattering (dehiscence) is at its lowest. Spineless varieties were bred. The oil is excellent for edible purposes, being the richest in linoleic acid, but since the content of natural antioxidants (tocopherols) is low at ca. 400 ppm, its oxidative stability is poor and needs to be reinforced with synthetic antioxidants (Patterson, 1989; Sonntag, 1979). This led to the successful breeding of a new variety with greatly increased oleic-acid content (Table 4.22), which begins to resemble olive oil in fatty acid composition (Smith, 1985; Sonntag, 1979). As expected, the tendency to produce oxidative polymers in cooking operations is avoided in the new variety. The Codex Stan. 27-1981 Suppl. 1 (1983) recognizes the new variety. For classic safflower oil, the Codex suggests an iodine value range of 135–150, an unsaponifiable of 1.5% maximum, a peroxide value of 10 maximum, and Fe of 1.5 ppm and Cu of 0.1 ppm maximum. Phosphatides, sulfur compounds, and trace metals have offered no particular problems in the refining of this oil. Prior to bleaching, the neutralized oil derives its golden-yellow color mainly from a E-carotene content of ca. 12.6 mg/L. Every incentive is to use this oil as a table oil, but it would hydrogenate readily to either a margarine or shortening fat if required (Patterson, 1983). In such a case, since the C18 carbon atom chains predominate to over 90%, to include ca. 10% of some other oil rich in fatty acids of a different chain length in the formulation is important (see also section Rapeseed Oil (Colza)). This prevents the development of grittiness in the product within very few weeks. Safflower oil has TABLE 4.22 Fatty Acid Composition of Safflower Oils High linoleica (%) Shorter than

C14

<0.1

C14:0

<1.0

C16:0

2–10

C16:1

<0.5

C18:0

1–10

4–8

C18:1

7–42

74–79

C18:2

55–81

11–19

C18:3

<1.0

C20:0

<0.5

C20:1

<0.5

C22:0

<0.5

a

Codex Stan. 27-1981 Suppl. 1 (1983).

b

High oleicb (%)

Experimental high-oleic variety (Sonntag, 1979).

4–8

134

H.B.W. Patterson

also been employed since ancient times in the manufacture of enamels and glazes (Patterson, 1989).

Recommendations After alkali refining, a choice is made for bleaching between the use of about 0.25% of a well-activated clay such as Tonsil L80 FF, Engelhard F105, Galleon N or NF2 or Fulmont AA, or a rather large dose of Tonsil ACC FF, Engelhard F4, Galleon NS, or Fulmont 300C. Much depends on how light a color is desired. A contact time of 20–30 minutes at 95–100°C is sufficient. A final color of 30 Y 3.0 R (5¼”) should be obtainable if required. Unlikely will the oil from smoke-dried safflower now be met, but if a precaution to remove heavy PAH (see section Use of Carbon in Chapter Basic Components and Procedures) needs to be taken, include 0.25% of well-activated carbon (e.g., Norit SA4) with the clay and prolong the contact time to 45 minutes. (See section General Principles in Chapter 4.)

Sesame Oil (Gingili, Sim-sim, Til) The sesame plant is cultivated in many warm countries, and exists in several varieties, growing one to seven feet in height. The dehiscent pods contain the seeds whose oil content averages 50%. So-called white, red, brown, and black types of seed occur— the kernel of the first is the favorite as a confectionery item, while that of the red holds more oil. Sesame oil, another obvious member of the oleic–linoleic family (Table 4.23), has outstanding oxidative and flavor stability which has widened its use to pharmacy and cosmetics (Sonntag, 1979). The C18 atom chains account for around 90% of the fatty acids. The oil’s very low content of linolenic acid and the substantial presence of tocopherols (which may reach 500 ppm) together do not account for the oil’s stability. Lengthy research shows sesamol (Fig. 4.2) to be a potent antioxidant released on the breakdown of sesamolin present in the oil at ca. 0.6%. Sesamin is also present (Budowski, 1964), and it enhances the effect of pyrethrum in insecticides. Further derivatives (sesamol dimer and sesamol-dimer quinone) were later identified as contributing to the overall antioxidant potency (Kikugawa et al., 1983). Sesangolin was isolated (Fig. 4.2). Alkali neutralization, washing, and the deodorization steps diminish the amount of released sesamol present, but bleaching earth and the action of dilute acids release more sesamol from sesamolin. Therefore, after bleaching, some rise in antioxidant activity is noted. Hydrogenation causes some breakdown of sesamolin and sesamin. Sesamolin or free sesamol gives a striking cherry-red color with strong hydrochloric acid and furfurol. This is the basis of the long-known Villavechia test or its modified form, and the Baudoin test. In some countries the inclusion of around 5% of sesame oil in the formulation of margarine or vegetable ghee (vanaspati) is mandatory to enable its detection if it is used as an adulterant of butter or ghee. Also, the seame oil contributes additional polyunsaturates to the formulation as well.

Bleaching of Important Fats and Oils

135

Fig. 4.2. Structural formulas of sesamolin, sesamin, sesmol, sesamol dimer, sesamol-dimer quinone, and sesangolin.

Codex Stan. 26-1981 Suppl. 1 (1983) advises a fatty acid composition as shown in Table 4.23, an iodine value range of 104120, unsaponifiables of a 2.0% maximum, and for virgin oil, a maximal acid value of 4. The phosphatide content is less than 0.1%, and the sulfur content is negligible. Such an oil is easily alkali-refined by the same method as groundnut oil. For many years this oil, remarkably free from catalyst poisons, was the choice for making comparisons of the hydrogenating activity of different nickel catalysts. More demanding tests were eventually found desirable. Now,

136

H.B.W. Patterson

more satisfactory is to retain a standard unreduced nickel base and then reduce a portion for testing against the unknown catalyst, using as a medium of comparison some oil with which the hardener is involved on a day-to-day basis. In this way, test conditions approach much more closely the process regime (Patterson, 1983). Sesame oil has been used as an adulterant of cottonseed or groundnut oil and vice versa, according to profitability.

Recommendations A dose of 0.5% of moderately activated clay (sometimes less) during a contact time of 20 minutes at 90–95°C should be adequate. Such clays are Fulmont 300C, Tonsil ACC FF, Engelhard F4, and Galleon NS. A normal crude oil is not expected to exceed 30 Y 2.5 R (1”), while a crude showing 35 Y 4.3 R (5¼”) will probably drop to 35 Y 1.5 R (5¼”) after neutralization and washing alone. The subsequent bleach as described above should then show further lightening to the range of 3 Y 0.6 R (5¼”) to 1Y 0.2 R (5¼”). Some refiners (Goebel, 1976) have added a little activated carbon along with the clay. (See section General Principles in Chapter 4.)

Soybean Oil Well-known is that soybeans are the world’s largest oilseed crop, and over 15 million metric tons (MMTs) of oil are produced each year (palm, 8.4; rape, 7.5; sunflower, 7; MMTs, respectively) (Erickson & Wiedermann, 1989). Originally a native of China, its cultivation is now widespread—the principal producers are the United States, Brazil, China, and Argentina. Whereas the oil content of the bean is only some 20% (dry basis), the protein available is 40%, and the demand for the latter sustains the market. Soy protein is both valuable human food and animal feed; note that about seven pounds of vegetable protein in livestock feed is needed to produce one pound of animal protein for human consumption. The crop has the advantage TABLE 4.23 Fatty Acid Composition of Sesame Oil Fatty acid C <14

Range (%) <0.1

C14:0

<0.5

C16:0

7.0–12

C16:1

<0.5

C18:0

3.5–6.0

C18:1

35–50

C18:2

35–50

C18:3

<1.0

C20:0

<1.0

C20:1

<0.5

C22:0

<0.5

Bleaching of Important Fats and Oils

137

that the beans handle and store well in the correct conditions. The fatty acid composition is not particularly subject to environmental factors, although many varieties of plants were tested and are in use on a gigantic scale (Hutchins, 1976). The Codex Stan. 20-1981 Suppl. 1 (1983), mandatory from 1987 (Mounts, 1987), gives the composition shown in Table 4.24. In an effort to increase the oxidative and flavor stability of the oil itself or its lightly hydrogenated products, varieties were bred with linolenic fatty acid contents below 4%. This does not yet appear to have secured the desired improvement (Erickson & Wiedermann, 1989). Comments made at the same time stated that linolenic acid as an Z-3 polyunsaturated fatty acid may be important in human metabolism not earlier appreciated (Erickson, 1983). The Codex quotes an iodine value range of 120–143; in fact, 133 is very typical. In practice, the unsaponifiable is under 1%, although it quotes 1.5% maximum. The oil contains 0.15–0.21% of tocopherols. While these are useful antioxidants, it appears (Brekke, 1980) that an increase in red color of mishandled soybean oil may be due in part to the formation of chroman 5,6 quinone (Fig. 1.7) by the oxidation of tocopherol. Brown tints arising from the degradation of protein and carbohydrate are notoriously difficult to eliminate. Commercial crude oil at 0.3–0.7% free fatty acid meets the standard required by the National Soybean Processors Association, U.S.A. (NSPA) (see Table 4.25). Phosphatides in crude oil vary in the range of 1.5–3%, but such oil is customarily supplied after being partly degummed at the mill to ca. 0.3% of phosphatides (ca. 110 ppm of phosphorus); the NSPA accepts up to 200 ppm of phosphorus. This reduction in gummy matter at the start makes subsequent handling, storage, transport, and processing simpler;

TABLE 4.24 Fatty Acid Composition of Soybean Oil Fatty acid C <14

a

Range (%)a <0.1

C14:0

<0.5

C16:0

7.0–14

C16:1

<0.5

C18:0

3.0–5.5

C18:1

18–26

C18:2

50–57

C18:3

5.5–10

C20:0

<0.6

C20:1

<0.5

C22:0

<0.5

C24:0

<0.5

Mandatory 1987 (Mounts, 1987).

138

H.B.W. Patterson

it also saves expense. The degumming of soybean oil alone yields enough lecithin to meet market demand. If, for example, 0.01% of citric acid/oil is used to treat the nonhydratable phosphatides and all this passes to the 1% of lecithin present in the oil, this would be equivalent to 1% of citric acid on that same lecithin when it was separated for use. For many purposes, this concentration is likely to be quite acceptable. Standards of crude oil, methods of processing, and the quality of the finished products have all advanced considerably in the last 30 years. Progress still continues, although further improvement on already high standards naturally becomes more difficult and possibly more expensive. A choice of three approaches to processing still exists. The decision rests on the quality of crude oil, type and standard of products to be produced, and, as always, the cost involved. Now well-understood is that whatever refining route is adopted, that to ensure lasting stability of the deodorized oil, a strict control of phosphatides, prooxidant trace metals, degree of oxidation, and the presence of chlorophyll compounds must be exercised. Degumming by various means was already described (see section Degumming in Chapter 1). The earliest alkali refining dealt with oil from which a substantial proportion of the phosphatides was already removed by settling and water degumming so that only some 0.3–0.5% remained. The neutralization of the free fatty acids with caustic soda was and is intended to remove nearly all of these remaining phosphatides along with the soapstock formed. Those which escape, ca. 0.015% of gum or 5 ppm of phosphorus, are to be trapped finally by adsorbents at the bleaching step. Centrifugal continuous refining enhances performance, and a small dose of citric or phosphoric acid is added to the oil to ensure that all the phosphatides become hydratable (see section Settling in Chapter 1). This also assists, to some extent, with the removal of trace metals. The hydrated gums may then be expelled on their own by the first centrifuge (ideally self-cleaning) prior to neutralization, or they may be left to be expelled along with the soapstock. Washing and drying follow; then come bleaching and deodorizing. In place of centrifugal refining, the Zenith semicontinuous degumming and neutralization plant offers an alternative. But here again phosphatides are acidified, hydrated, and expelled before finely divided oil is caused to flow up through a column of dilute caustic soda to neutralize the free fatty acids with a minimum of neutral-oil loss (Cavanagh, 1989; Erickson & Wiedermann, 1989; Hoffman, 1973; Young, 1978, 1981). The duplication of semicontinuous components makes a continuous operation possible. The plant is made of stainless steel; automatic and manual operations are possible. Lastly, we come to the possibility of physically refining soybean oil. Up to the early 1980s, centrifugal refining had come to enjoy a dominant position as the process for obtaining a first-class product from a range of crude-oil qualities. It still does this (Erickson & Wiedermann, 1989). From then on, a search was conducted by many workers to establish what criteria needed to be satisfied to operate a physical-refining facility. Apart from the increasing thermal efficiency of physical-refining plants operating on other oils, notably palm and lauric oils, it became more and more desirable

TABLE 4.25 Soybean-Oil Quality Parameters Water degummed for: NSPA water degummeda

Chemical refiningb

Physical refiningd

Free fatty acid (%)

0.75

0.75

0.75

Volatiles (%)

0.20

Insolubles (%)

0.10 c

e

Maxima

Physical refiningf

Neutralized bleached oilh

Neutralized bleached deodorized oilh

0.75

<0.15

<0.05

8-hour minimum

A O M to 35 meq/kg 200

50 Y 5 R

50 Y 5 R

<4

<4

2.0

1.5

1.5

<4

<2

3.0

2.0

2.0

nil

nil

E232

2.5

2.5

1.5

3.1i

E268

0.4

0.3

0.2

2.1i

Iron ppm

2.0

2.0

1.0

Copper ppm

0.1

200

Anisidine value Peroxide value meq/kg Unsaponifiables %

Flash point Soap

1.5

<0.1

<0.1

<0.03

<0.03

nil

nil

Bleaching of Important Fats and Oils

150

50 Y 5 R

Phosphorus ppm Lov. (1”)

20

g

121°C minimum

139

100 ppm Phosphorus = 0.01% Phosphorus = 0.3% of phosphatides (approx.) a (Forster & Harper, 1983; Norris, 1979). b Quality Assurance Data for crude, water-degummed soybean oil for caustic-soda refining (Young, 1981). c This semidegummed crude is to have more gum removed along with or just prior to the removal of soapstock. Bleaching then follows. d (Forster & Harper, 1983). e Further acid/earth treatment to attain 10 ppm of a phosphorus maximum prior to deodorizing. f Ong, J.T.L.; Fette Seif. Ans. 1980, 82, 169 and (Young, 1989). g Further acid/earth treatment to attain 5 ppm of a phosphorus maximum prior to deodorizing. h (Young, 1981). i Pardun, II., Analyse der Nahrungs fette. P. Parey (1976) p. 223. Also see Young (1989).

140

H.B.W. Patterson

from an environmental point of view to avoid the effluent problems associated with soapstock treatment. The most readily comprehensible way of summarizing this research is to set out in tabular form a selection of currently recommended quality parameters (Table 4.25). A good-quality crude soybean oil after preliminary degumming is expected to show no more than 50 Y 5 R (1”). After alkali neutralization, washing, and activeearth bleaching, a color around 28 Y 2.6 R (5¼”) is feasible, and 3.5 R is a normal maximum. When fully refined, a color of 20 Y 2 R (5¼”) maximum is expected (Norris, 1979); in the best circumstances (80,257,264), a color less than 10 Y 1.0 R (5¼”) is obtained. Also required is a cold test at 0°C of 5.5 hours (60,272). In the 1980s, a more rigorous removal of chlorophyll was sought. Crude normal soybean oil contains between 500 ppb and 3000 ppb of chlorophyll and its derivatives (Young, 1989). Their removal is not simply on account of color, but because chlorophyll encourages photooxidation. When soybean oil is bleached to 4.0 R (5¼”), normally it will show only 0.5 R (5¼”) after deodorization (Hastert, 1988). This drop in red color will be less for oil from damaged beans and very slight for oils such as cottonseed and corn. For oil showing 0.5 R (5¼”), a green-color component is detectable only at ca. 100 ppb of chlorophyll and above (Hastert, 1988). Nevertheless, many processors seek a 50 ppb maximum, and some even lower, so as to escape its prooxidant effect. An instructive commentary on some commercial quality standards in 1989 was given by Leysen (1989).

Recommendations A contact time of 20–30 minutes at 90–105°C is usual. Several earths specially designed to remove chlorophyll while maintaining normal efficiency in bleaching other pigments are available. Those are Engelhard NevergreenTM and F160, Tonsil Optimum FF, Fulmont XMP2 and Galleon DC, possibly in conjunction with Mizukalife F-1. A dose rate of 0.5–1.5% of activated adsorbent should be adequate when working with normal oil. (See section General Principles in Chapter 4.)

Sunflower Oil (Tournesol, Girasol) Originally a native of Central America, the sunflower was brought to Europe in 1569 for ornamental purposes. Its cultivation has spread to warm climates all around the world; the Ukraine produces about one-quarter of the world’s crop. The production of sunflower oil at over 7 million tons per year is the fourth-largest edible-oil production in the world, following soybean, palm, and rapeseed oils. Plant species vary from 2.5 m to 6 m in height; the largest flower head is 35 cm in diameter. All parts of the plant can be used, since after the removal of the seeds the flower heads can be fed to cattle, and the other parts of the plant ploughed back, with or without burning, as a source of potash. Breeding has raised the oil content of the seeds to 50% and even higher (Sullivan, 1980); the hull accounts for 25% of the seed, and contains only some 2% of oil, hence, the oil content of the kernel is around 66%. Since some 21% of protein is also present in the kernel, it becomes the source of a high-protein meal after the

Bleaching of Important Fats and Oils

141

extraction of the oil (Gohl, 1975; Veldstra & Klere, 1989). The popular course is to dehull the seed and then prepress/solvent extract the kernels. The hull contains about 83% of all the wax present in the complete seed, and the seed coat 17%, leaving only traces of wax in the kernel (Sullivan, 1980). The advantages of dehulling are that the crude oil then contains very much less wax, the meal is protein-rich, and plant utilization is higher. Hulls can be used as a fuel, or a proportion can be mixed into animal feed. Some mills rely solely on pressing or extraction. The wax content of crude sunflower oil varies considerably with species, origin, and method of milling as reported in the literature over the years. With 100% extraction, the wax content of crude oil could reach 1.25% (Campbell, 1983), the wax content of seed is usually less than 1% (268), and crude sunflower oil contains waxes from 50 ppm to above 1500 ppm, but oil from dehulled seed always contains less than 300 ppm (Denise, 1987). Sullivan (1980) comments regarding wax measurement, “The analytical methods are involved and time-consuming.” The phosphatide content is generally low at around 0.2% of lecithin (ca. 66 ppm of phosphorus) in first-class oil, but sometimes may reach 1.0%. Now generally recognized is that in the case of sunflower oil, besides the advantage in improving soapstock separation or safeguarding color and stability in physical refining, the removal of gums markedly assists the separation of waxes. This is because the presence of phosphatides inhibits crystallization and the separation of wax. When little phosphatide is present, waxes are sufficiently hydrophilic to concentrate when crystallized at an oil–water interface. Noteworthy is that when crude sunflower oil is settled in a storage tank with as little as 0.16% of moisture present (Pritchard, 1983), substantial hydration of gums occurs; as these settle into the foots, they bring down considerable amounts of wax. The supernatant oil may then be pumped to processing. One option is to then dose the continuous oil flow in the cold with dilute citric acid and a dilute solution of a wetting agent such as sodium dodecyl sulfate which, within a short interval (residence time), will bring about further separation of gums and crystallized wax. These are removed by centrifugal washings. A filtration step may follow centrifuging to get rid of more wax. Refining follows. To reduce the phosphorous content below 5 ppm prior to the deodorization step of physical refining is advisable (Veldstra & Klere, 1989). (Note also footnote g, Table 4.25.) Sunflower oil may be refined chemically or physically and, if intended for use as a salad oil, winterized to pass the cold test. Different options are described in the literature (Campbell, 1983; Forster & Harper, 1983; Pritchard, 1983; Sullivan, 1980; Veldstra & Klere, 1989). If foots rich in wax and gums, as mentioned previously, are themselves mixed with cold dilute citric acid and a wetting agent and allowed to settle, the lower aqueous layer may be withdrawn along with the gums and waxes; then the supernatant oil layer is returned to crude oil. This procedure holds back a heavy proportion of wax and gum in the foots from the main process stream. Table 4.26 shows the Codex Stan. 23-1981, Suppl. 1, Vol. XI (1983) amended 1987 (Mounts, 1987) ranges of fatty acid composition of sunflower oil. As with other oils, examples can be found where these limits were exceeded, especially in

142

H.B.W. Patterson

experimental production. The oil is much valued for its high linoleic-acid content, which can range from 7 to 80% at extremes and conversely, an oleic-acid content of up to 88% (Knowles, 1968; Robbelen, 1991). As expected, the warmer climates produce higher oleic-acid and lower linoleic-acid contents, and those oils are, therefore, particularly well-suited for salad oils. The Codex advises an iodine value range of 110–143; a composition for an iodine value of 132 is shown as typical in Table 4.26. The exceptionally low linolenic-acid content makes for oxidative stability, and furthermore, when oxidative degradation does occur, the immediate off-flavors do not have nearly as low a threshold value as those arising from oils of higher linolenicacid content. Groundnut oil shares this advantage. Many parcels of crude oil have a peroxide value under 10 (as proposed by the Codex) and an anisidine value under 5; an anisidine value above 12 indicates a very poor oil. As with groundnut oil, various authorities set limits for E232 and E268 and their proportion to one another, but these are empirical in the sense that they relate to previous experience of flavor instability of the species of oil in question (sunflower, groundnut, rapeseed, etc.). An E232 less than 5 suggests a good, little-damaged oil. The Codex gives 1.5% of unsaponifiable as a maximum; usually this is less than 0.8%. Like the Codex, the AFOA (Campbell, 1983) put a maximum of 2% on free fatty acid in crude oil; in practice 1.0–1.5% is common (Patterson, 1989); tocopherols (mostly D form, therefore with vitamin-E TABLE 4.26 Fatty Acid Composition of Sunflower Oil Fatty acid

Possible rangea (%)

C <14

a

Typical composition at iodine value of 132 (%) <0.4

C14:0

<0.5

C16:0

3.0–10

6.5

C16:1

<1.0

0.5

C18:0

1.0–10

4.5

C18:1

14–35

23.0

C18:2

55–75

63.5

C18:3

<0.3

<0.3

C20:0

<1.5

0.5

C20:1

<0.5

0.5

C22:0

<1.0

0.5

C22:1

<0.5

C24:0

<0.5

C24:1

<0.5

Mandatory 1987 (Mounts, 1987).

Bleaching of Important Fats and Oils

143

activity) normally amount to 700 ppm. Sulfur compounds are negligible. If PAHs from smoke-dried seed are suspected, the procedure is to add 0.25% of carbon/oil of a well-activated carbon (see section Commercial Activated Carbon Products in Chapter 3) to the bleaching clay and prolong the contact time to 45 minutes. Sullivan (1980) remarks that the preparation of sunflower oil for physical refining follows much the same pattern as for soybean oil; the crude oil should be water-degummed to 50 ppm of phosphorus, and this in turn should be reduced to 15 ppm of phosphorus (preferably less) by a 1% dose of well-activated clay bleach, and any iron present should be reduced to no more than 0.1 ppm at the same time. As with several other oils in the oleic–linoleic family, the C18 fatty acids dominate the composition to the extent of 90%. To avoid graininess developing in the margarine or shortenings which rely heavily on hydrogenated sunflower oil, about 10% of some other oil that is not so predominantly of the C18 fatty acid composition should be introduced into the formulation (see section Rapeseed Oil (Colza) in Chapter 4). Lastly, when considering the question of bleaching as a means of removing color, note that after refining a sunflower oil—including the step of bleaching to adsorb unwanted impurities such as traces of phosphatides and prooxidant metals— important in some markets is to add some permitted golden-yellow color to meet the customer’s perception of how a sunflower salad oil should appear. This, of course, is not universally necessary. A top-quality crude sunflower oil is likely to have a paleyellow color of 45 Y 5 R (5¼”) or 22 Y 1.4 R (1”); a poor-quality crude oil may be amber-colored and show 50 Y 4 R (1”). Neutralization and bleaching usually improve the color to 20 Y 2 R (5¼”); AFOA and NCPA call for a 2.5 R (5¼”) maximum (Pritchard, 1983). The final product as a salad oil may easily show no more than 6 Y 0.6 R (5¼”). Obviously, for margarine making and hydrogenation, the required color need not be nearly so pale.

Recommendations Not pigment, phosphatides, nor chlorophyll present difficulties at the bleaching step. A dose of 0.5% of the more active earths such as Tonsil Standard FF, Tonsil L80 FF, Engelhard F105 or F105 SF, Fulmont AA, and Galleon NF2 is indicated. Less active earths such as Galleon NS, Tonsil ACC FF, Fulmont 300C or Engelhard F4 might require a dose of up to 1.5%. As usual, the quality of the crude oil is an important factor. A bleaching temperature of 95–100°C is adequate. If 0.25% of activated carbon is added along with the earth to remove PAHs, the contact time should be extended from 30 to 45 minutes. (See section General Principles in Chapter 4.)

Marine Oils For decades, whale oil dominated the marine-oil market, looking to technical outlets, while residues from the processing of seasonal surpluses of fish were used as fertilizers. The arrival of industrial hydrogenation (Anon., 1914; Margarit, 1913; Normann, 1902, 1903) opened the way for whale oil to be converted to a substitute for tallow in the manufacture of soaps and glycerine and for conversion to cooking fat and margarine.

144

H.B.W. Patterson

The rapid decline of the whale population by the 1960s accelerated the production of fish oil to take the place of whale oil for edible purposes. Also important is the fact that certain amino acids present in fish protein can supplement vegetable proteins when fed to animals. In fact, fish protein can be prepared for direct human consumption. This is scarcely surprising since only about one-third of the fish caught in the world each year is converted to oil and meal (Patterson, 1989). World consumption of fish oil is around 1.6 million tons per year, with the United Kingdom taking over 200,000 tons, followed closely by Germany, Japan, and the Netherlands (Opstvedt et al., 1989). Sperm whale oil had long been recognized as being in a separate class since three-fourths of it consists of long-chain fatty alcohols (mainly C14 to C20, saturated and unsaturated) esterified by long-chain fatty acids. These components are, of course, waxes, and provide the basis of high-class lubricants—hydrogenation again being involved. As methods of structural analysis became quicker, easier and more accurate, the differences in the fatty acid composition of oils from different species of fish and indeed from different organs of the same fish became very evident and coincided with their much-increased use. As with other living things, diet, environment, and degree of maturity markedly affect this composition. Table 4.27 shows the approximate proportions of saturates, monoenes, and polyunsaturates in the most common fish oils and in whale and seal oils. This is of considerable interest to the oil hardener since the greater the required fall in iodine value, the greater will be the hydrogen required. Also the more subtle point is that if few highly unsaturated fatty acids are present, as in capelin oil, careful selective hydrogenation of these may quickly reduce them to a level where the fat now has oxidative and flavor stability but a texture still soft enough to enable it to form tub margarines, spreadable when taken from the domestic refrigerator (Patterson, 1983, 1989). The typical fatty acid makeup of capelin oil is shown in Table 4.28, and may be compared with examples of the opposite kind in Table 4.29. Menhaden oil and its hydrogenation were described at length (Bimbo, 1987). Where whales were hunted for human consumption of their flesh—as in Japanese whaling—the oil obtained from their blubber was of the highest quality since the animal would be flensed very shortly after being killed. If, on the other hand, the inflated carcass is left to float for some days before being hauled on to a factory ship for extraction of the oil, putrefaction has already set in and typically the sulfur content of the oil may be around 30 ppm (as compared to 18 ppm in the first example). Again, better oil is produced when the blubber is cooked separately from bone and tissue. The moral for fish oils is the same. The chances of producing a good oil are enhanced if the fish is processed within a week of being caught. At a late stage in the processing, the oil is centrifuged from the press liquor, and the watery portion known as the stickwater is concentrated by evaporation. If the concentrated stickwater is itself centrifuged to obtain a little oil, the latter will be much higher in sulfur content than the main bulk of the oil separated earlier, so it should not be mixed back. The European Economic Community recommended in July 1979 that as a health precaution the proportion of erucic acid in the fatty component of diet be

145

Bleaching of Important Fats and Oils

TABLE 4.27 Fatty Acids of Marine-Oil Groups Iodine value

Fatty acid type

(%)

ca. 100

saturated

19

monoene (very high)

76

polyunsaturated (very low)

5

Group A Capelin

Herring Cod (liver) ca. 130

saturated (moderate)

Halibut (liver)

ca. 160

monoene (high)

under 30 ca. 50

Haddock (liver)

ca. 120

polyunsaturated

15–30

Whale

ca. 120

ca. 35

Seal, etc. Group B Anchovy

ca. 200

Pilchard

ca. 200

saturated

Sardine

ca. 170

monoene (low)

Menhaden

ca. 165

polyunsaturated (moderately high)

under 30 35–45

TABLE 4.28 Principal Fatty Acid Composition of Capelin Oil Fatty acid type

%

Fatty acid type

%

C14:0

8

C20:0

2

C16:0

11

C20:1

23

C16:1

10

C22:0

1

C22:1

24

C18:0

2.5

C18:1

13.5

C18:2

0.7

C20 and C22

1.7

polyunsaturated

1.7

total

97.4

kept within 5% by weight. Strictly interpreted erucic acid is cis-13 docosenoic acid: C22:1n-9. Some countries and manufacturers may prefer to apply the restriction to all C22:1 isomers. This means that hardened fish oil rich in C22:1 may have to be blended with other oils such as the hardened oils derived from the type of fish oil shown in Table 4.29 where the proportion of C22:1 is very low. The Player and Wood test (Player & Wood, 1980; Rossell, 1991) is used to identify

146

H.B.W. Patterson

TABLE 4.29 Principal Fatty Acid Compositions of Anchovy and Pilchard Oils Fatty acid type

Anchovy oil (Peruvian) (%)

Pilchard oil (South African) (%)

C14:0

7.5

7.3

C16:0

17.5

16

C16:1

9

9

C16:2

1

2

C16:3

2

2

C16:4

2

3

C18:0

4

4

C18:1

12

10

C18:2

2

2

C18:3

1

1

C18:4

3

—.

C20:1

2

1

C20:2

1

—.

C20:3, 4, 5

19

21

C21:5

—.

1

C22:1

1

2

11

12

C22:3, 4, 5, 6 C24:1

0.5

0.7

totals

95.5

94.0

erucic acid specifically; the early nutritional research was described by FAO (1980). More recently, much attention was given to the importance in human metabolism of cis 5,8,11,14,17 eicosapentaenoic acid (EPA); C20:5n-3 and cis 4,7,10,13,16,19 docosahexaenoic acid (DHA); C22:6n-3. These are also referred to as Z-3 fatty acids: that is the group in which the first double bond is located three carbon atoms from the terminal methyl group itself counted as number one. These occur in fish oils. Vegetable oils such as soybean and canola contain useful amounts of another Z-3 fatty acid, D-linolenic acid, cis G9,12,15 octadecatrienoic acid: C18:3n-3. This fatty acid is capable of being metabolized to EPA. Latta (1990) has contributed a very useful review of the discussion concerning the need to supplement diet with EPA and DHA if ample D-linolenic acid is already being consumed. Crude fish oil must be well-refined before it can be used in an unhydrogenated form. This was described along with the limitations of such use (Opstvedt et al., 1989). In the process of separating oil from the fish body tissue and considering

Bleaching of Important Fats and Oils

147

that oils are prone to oxidize, normally copious amounts of mucilage pass into the soapstock and wash during refining with caustic soda. In the classic batch refining process, the soapstock may be dark brown; subsequent washes grow progressively lighter. This alone indicates how the quality of the oil is being improved. Continuous centrifugal refining has long been popular for fish oils, and in this case a small (up to 0.1%) amount of phosphoric acid intensively mixed into the oil flow prior to neutralization leads to better separation of gums which may be present, and probably helps in the removal of traces of Fe and Cu, as well as subsequently producing a paler, clearer oil. The first centrifuge should be of the self-cleaning type; if it is of the solid-bowl type, employing bowl flush water, it will have to be cleaned at intervals. Two further centrifugal washing steps follow; the first of which may employ a weak, caustic-soda solution. Very full operating instructions are given by plant suppliers and the International Association of Fish Meal Manufacturers. When the soapstock and combined washes come to be split by mineral acid, tarry deposits form so that the splitting plant’s acid-oil-holding vats need to be cleaned in rotation after about two weeks’ use. A test sample of acid oil poured into petroleum ether separates its oxidized fatty acids, and provides an indication of quality to those processing it further as an oleochemical. Crude fish-oil free-fatty acid content usually lies between 2 and 5%; unsaponifiable is normally less than 2%, but in the case of South African pilchard oil, the unsaponifiable may rise from the usual 2% to around 6% when the oil content of the fish is at its lowest just after spawning. Presumably, a similar related effect is in other fishes. The phosphorous content is unlikely to exceed 100 ppm (0.3% of phosphatide), and the sulfur content is 18 ppm for properly handled oil. Although a peroxide value up to 20 meq/kg and an anisidine value as high as 60 may be met in some fish oils, for herring oil a peroxide value of 6 maximum, anisidine value of 12 maximum, E232 of 12 maximum, and E269 of 4 maximum are quoted as satisfactory (Patterson, 1983). Again, although Fe up to 7 ppm and Cu up to 0.3 ppm are possible, an Fe 1.5 ppm and Cu 0.2 ppm in anchovy and sardine oils (Patterson, 1989) would be rather high, yet could easily be reduced in processing. No Codex Standard exists for fish oils. Refining with caustic soda followed by activated clay bleaching is the obvious pretreatment prior to hydrogenation since this method effectively gets rid of the unwanted minor components of the oil, including catalyst poisons, without placing too great a burden on the bleaching step. Physical refining was avoided, and this trend is likely to continue. Nevertheless, variations of the popular pretreatments were explored by some processors, at least for a period, primarily on the basis that costs could be reduced. Partly exhausted nickel catalyst was used as a sulfur adsorbent at the prehydrogenation step to make subsequent hardening easier. Some suppliers of fish oil (e.g., Peruvian anchovy) provide the service to the export customers of semirefining crude oil to ca. 0.3% of free fatty acid before shipment. Alkali refining may, of course, be completed on arrival, or the semirefined oil merely dried before hydrogenation. Alternatively, the dried, semirefined oil may simply be passed through a filter press partly filled with residue from previous bleaching batches, or it may itself be given a moderate (ca. 0.5%) adsorptive bleach before hydrogenation. The ultimate appears

148

H.B.W. Patterson

to have been proposed by those who added bleaching clay to crude whale oil along with nickel catalyst in hardening, recognizing that the soapstock produced in the posthardening refining is hardened-oil soapstock. Cutting down on pretreatment pushes the burden on to post-treatment. If any oil-modification process such as hydrogenation, interesterification, or fractionation is to give its best, this is achieved by offering a well-cleaned oil to the modification step. The choice becomes either taking the initiative early in the process or being forced to do so later to keep up with a competitor. All fish oils contain a proportion of polyunsaturated fatty acids, and in several oils it is quite high. The iodine value of anchovy oil, for example, sometimes slightly exceeds 200. Opportunities for contact with air during storage and processing are best kept to a minimum. During neutralization and washing, especially in centrifuges, whether hermetically sealed or not, the risk of oxidation is not as great as at other stages. The oil is next dried to ca. 0.1% of moisture under vacuum, and usually passed to bleaching with little delay and minimal contact with air. At this stage, the oil has lost most of its phosphatides and some of its prooxidant trace metals; little free dissolved oxygen remains, the peroxide value is beginning to fall, and the anisidine value is beginning to rise; E232 and E268 probably show only slight changes. Handling in stainless-steel vessels, holding under vacuum (free from under-surface air leaks), and blanketing with nitrogen all diminish risks to quality; oil cooled below 50°C reacts more slowly than oil held at 80–90°C. These precautions need to be seen in relation to the fact that following process steps themselves improves flavor stability. As is generally agreed, when bleaching is complete, oil is most vulnerable to oxidative attack. Then, during, and after filtration, the exclusion of air is most worthwhile (Patterson, 1989). When hardened and deodorized, the color of fish oils normally falls below 20 Y 2 R (5¼”).

Recommendations) A contact time of 20 minutes at 90°C is usually adequate. Difficult cases will normally be met by increasing the clay dose from the normal 0.5–1% to 2%, according to results. Activated earths recommended by suppliers are Tonsil ACC FF up to Tonsil Supreme FF, Galleon N and NF-2 up to V2-Super, Engelhard F105 SF and NevergreenTM, and Fulmont 700C up to Premiere. (See section General Principles in Chapter 4.)

Hydrogenated Oils Alkali refining of hydrogenated oils followed by a light bleach prior to deodorization is not necessary if trace metals can be adsorbed on a light dose of activated clay and the free fatty acid is removed in deodorization. Either way, the aim is to achieve a nickel content to <0.2 ppm, iron to <0.12 ppm, and copper to <0.05 ppm. The extent to which color removal/destruction is required varies widely. A crude hardened lauric oil may show only 2 Y 0.1 R (5¼”) and a crude hardened palm oil 28 Y 2.8 R (5¼”).

Bleaching of Important Fats and Oils

149

The ability of even mildly activated clays to adsorb metals is markedly enhanced by the introduction of 0.05% w/w of citric acid/oil as a 50% solution.

Recommendations A contact time of 15 minutes at 90°C is sufficient. In easy cases of color improvement and the removal of metals, a dose of 0.5% of moderately activated clay may achieve all that is necessary before deodorization. In cases in which a better removal of color is sought, a dose of 0.5–1.0% of highly activated clay is normal. The final color sought may be between 2 Y 0.1 R (5¼”) and 25 Y 2.5 R (5¼”). Experience shows how much the heat-bleaching effect of deodorization can be taken into account. Certain activated clays were put forward as being particularly effective adsorbers of trace metals. Tonsil phosphorus is offered for the removal of Fe, Cu, Ni, and As, Engelhard F160 achieves low levels of metal, Laporte mentions Premiere and Fulmont Super A as suited for metal removal, and Mizusawa offers Galleon DC activated earth for the treatment of hardened oils and adds that the SiO2/MgO agent Mizukalife F-1 adsorbs trace metals as well as chlorophyll. (See section General Principles in Chapter 4.)

Interesterified Oils Prior to interesterification, oils should have been degummed (if necessary), neutralized, washed, and dried. Some may have been bleached and hydrogenated as well. Many such fats and oils are quite light in color. After interesterification, the oil is washed to a low level of soap—50 ppm. Recommendations A contact time of 20 minutes at 95–105°C is quite adequate. The commencing and final required color cannot be defined here, but in general, the final color is such that the product can easily be incorporated in a margarine formulation. These tasks are accomplished by up to 0.5% of Fulmont 300C, Tonsil 13, Galleon NS, and Engelhard F4. When a greater bleaching effect is necessary, this is achieved by 1% of Fulmont R169 or 700C, Tonsil ACC FF or L80 FF, Galleon NF 2, and Engelhard F 105. (See section General Principles in Chapter 4.)

Castor Oil Castor oil appears to have been in use in Egypt as a lamp oil 6000 years ago. Varieties of the plant, annual and perennial, from one to twelve meters high, occur in warm and temperate climates around the world. Although many varieties were bred with the goal of making harvesting easier and more regular, a significant annual production of oil is derived from wild trees. Three seeds, about twice the size of a coffee bean, are held in what resembles a rather small fruit of the chestnut tree. Inside the hard hull of the seed is the castor bean. If the hull of the seed is damaged, lipolysis of the oil content begins. World production of castor seed varies around one million tons per year. The oil content of different varieties of seed ranges from 40 to 60%. Castor oil

150

H.B.W. Patterson

is remarkable for containing more hydroxy fatty acid than any other vegetable oil. This is ricinoleic fatty acid, cis G9,12 hydroxyoctadecenoic acid. CH 3 (CH2)5CHOH-CH2CH=CH(CH2)7COOH. Through the years, ricinoleic-acid contents of the oil from 80 to 95% were reported, no doubt depending on variety, climate and so forth, but 89% is common. Table 4.30 shows the typical range of composition of castor oil; no Codex Standard exists, but national pharmacopoeias define the quality for pharmaceutical use, and a FOSFA contract may be used to regulate commercial transactions. Oil for pharmaceutical use is of the highest quality, and is taken from the first pressing. Some deacidification via caustic-soda refining may be required. As this use accounts for only about 1% of the market, its production would not sustain the industry. It may be given a clarification with a dose of activated clay or carbon before being filtered; correctly harvested and processed, beans commonly produce some almost colorless oil (Janson, 1974; Sonntag, 1979). The bulk of production is classed as technical or commercial castor oil, and is divided into three qualities. The first is supposed to consist solely of pressed oil, whereas castor “seconds” and “thirds” may contain some oil which was solvent-extracted from press cake, which typically contains about 12% of oil. Janson (1974) describes pressing, solvent extraction, and refining of castor oil at length. Pressing is kept at a steady 60°C because at even a few degrees higher the color suffers. The crude oil should be dried as it leaves the presses or refined immediately because moist oil increases in free fatty acid and deepens in color within 24 hours. Castor oil is denser (0.945–0.965, 25°C/25°C) and more viscous (9.5–11.0 poise @ 20°C) than nearly all vegetable oils. It maintains its viscosity over a wide temperature range from cold to hot; the pour point is 16°C. On combustion it leaves very little ash; therefore, it is a valuable lubricant. The iodine value ranges from 81 to 91, and first quality oil should not exceed 2% of free fatty acid (oleic) nor 2.2 Y 0.3 TABLE 4.30 Fatty Acid Composition of Castor Oil (Janson, 1974) Fatty acid type

%

C16:0

0.9–1.2

C18:0

0.7–1.2

C18:1

3.2–3.3

C18:2

3.4–3.7

C18:3

0.2

C18:1 (Δ9, 12-OH)

89–89.4 (ricinoleic)

C18:0 (9,10 di-OH)

1.3–1.4 (dihydroxy stearic)

Bleaching of Important Fats and Oils

151

R (1”). Hydroxyl values of 155–169 were found, but usually the range is 162–165. The accepted maximum for unsaponifiable is generally 1%, but U.S. No. 1 grade requires a 0.5% maximum. When some deacidification is needed, it is done on the batch procedure at 90–100°C to discourage emulsion formation. Caustic-soda solution is sprayed onto the hot oil with gentle stirring. After settling and the withdrawal of soapstock, hot water, and brine, if necessary, are sprayed onto the oil, preferably without stirring. The soap-free oil is then dried and bleached. Substantial amounts of castor oil are hydrogenated, usually to below 3 iodine value at 110°C and 5 atm pressure, the aim being to avoid a drop in hydroxyl value above two units and to obtain a slip point of 86°C (Patterson, 1983). Castor oil of up to 1% of free fatty acid hydrogenates comparatively easily. The castor bean contains a toxic protein ricin and an allergen which arises in the dust of castor meal. The oil after pressing/extraction is free from these, but their presence has heavily influenced the use of the meal as a fertilizer. Janson (1974) describes how these considerations must affect the siting and staffing of castor-bean processing plants.

Recommendations As mentioned previously, castor firsts are accepted up to 2.2 Y 0.3 R (1”]); Janson describes how “cold” (60°C) -pressed crude oil may show 5 Y 1.5 R, whereas screwpressed oil using temperatures around 120°C shows 50 Y 2.8 R from the same feedstock. A dose of 0.5–1.0% of Fulmont XMP3 is considered suitable by Laporte, depending on the color required. Corresponding highly active clays are Tonsil Supreme, Engelhard F160, and Galleon V2 or V2-Super. In all cases a temperature of 100°C should be maintained so the oil filters hot. Also desirable is to organize the program so that the filter operates in virtually continuous use until filled or cleaned, since if allowed to stand for 12 to 24 hours and cool, changes would take place in the press cake comparable to a kind of resinification. Pharmaceutical-grade castor oil is given a final polishing filtration (Janson, 1974). (See section General Principles in Chapter 4.)

Chapter 5

Bleachers H.B.W. Patterson

Batch Bleachers For batch processing, the basic design of a neutralizer/bleacher was altered little over the last 30 years (Young, 1986). Capacities of 5–60 tons were used. Figure 5.1 indicates the important features. The cylindrical (straight wall) dimension of the vessel is close to 1.5 times the diameter. This usually means that, after the removal of the soapstock, only the very top stirrer arm (33 feet) may be exposed. Stirrer arms 44 feet and 66 feet are in the plane of the paper and 33 feet and 55 feet perpendicular to it, so that the flat strips connecting the tips from upper to lower arms are helical in shape. Both flat arms and helical strips are so inclined that the upper stirrer stirs down and the lower stirrer stirs up. The internal angle of the cone is about 80 degrees. If a coil surface of 3 m2 per ton of oil is provided, a steam pressure of 12 atm will raise the oil temperature about 2°C per minute. Two-speed stirring is provided, top speed being double lower speed; 40/20 rpm would suit a 25-ton oil charge, rather slower for larger vessels and faster for smaller sizes. Baffles may be provided to prevent vortexing if necessary. Bleaching clay is delivered, as indicated, below the oil surface at ca. 80°C. Not shown in the figure are a safety valve, a bursting disc, a sight glass and a light glass, vacuum connecting and pivoted manhole lid, all of which are located on the crown. The vessel and vacuum system must be able to maintain a vacuum of 50 mm of Hg absolute. No leaks of air must occur below the oil surface. If the vessel is called upon to process half-size charges at times, an upper and lower coil system are provided, the space between the systems corresponding with the space between the upper and lower stirrers. Whether the heat-up time taken from 80°C to the recommended bleaching temperature is counted as part of the earth/oil contact time is at the discretion of the operator in the light of results. Obviously, filtration time and the press-bleach effect count toward the final result. In a truly continuous system, this type of consideration is less complicated. (See sections Batch Bleaching–Continuous, Countercurrent and Fixed-Bed Bleaching Methods in Chapter 1 and section General Principles in Chapter 4) Semicontinuous and Continuous Bleachers Several advantages (see section Continuous, Countercurrent and Fixed-Bed Bleaching Methods in Chapter 1) of continuity apply to bleaching as well as to many other process steps in the industry. King and Wharton (1949; Norris, 1982) proposed a system in 1947 in which the oil clay slurry was first sprayed into a vessel under reduced pressure at 55°C to flash off free moisture and dissolved air, and then transferred via a heat 153

154

H.B.W. Patterson

Fig. 5.1. Batch neutralizer/bleacher (not drawn to scale ).

exchanger to the lower half of the same vessel at 110°C. Agitation in each section is provided by a bleed of stripping steam. A bank of filters follows which allows for continuous filtration. EMI (Norris, 1982) employs a system in which the slurry is deaerated as it is sprayed into the head of a tall vessel under vacuum; thereafter, the slurry flows down through two heated compartments provided with paddle agitators to ensure freedom from forward mixing and to provide adequate residence time at bleaching temperature. Lastly, the slurry is pumped through a cooler to a filter bank. The Pellerin-Zenith system of a semicontinuous bleacher (Young, 1978) is shown in Fig. 5.2. The simple basic principle depends on a well-engineered, automatic, and

Bleachers

Fig. 5.2. Pellerin-Zenith semicontinuous bleacher.

155

156

H.B.W. Patterson

variable timing system, which has gained it considerable popularity. Since the bleachedoil slurry is drawn continuously from a final buffer section, the semicontinuous or stepwise procedure provides, in effect, a continuous service. Neutralized dried oil is passed via a heat exchanger to the spray shown at the top of the small vessel, which contains a heater and is under permanent vacuum. Free moisture is flashed off, and the oil pours into the top compartment of the bleacher. Here it is dosed with a 15% of citric-acid solution in water, or in the case of physical refining, possibly with phosphoric acid. The heat provided in this compartment increases the temperature about 20°C to the required bleaching temperature. At a preset time interval, the top compartment automatically discharges to compartment 2, where the bleaching clay is added automatically, either all at once or in two or three lots which may include a carbon dose if needed. Vacuum is maintained the whole time. Compartment 1 fills again as soon as compartment 2 is empty. Clay and carbon are held in nitrogen-blanketed bins. After the same time interval, compartment 2 discharges to compartment 3, where the bleaching is completed and the slurry finally dropped to compartment 4, still under vacuum. In normal running, oil slurry is pumped continuously to a bank of filters (line 1); manual provision is available for the recirculation of filtered oil until it becomes clear, and filter aid can be added as indicated if a clean filter is being precoated (line 4); oil which is not being filtered for some reason is recirculated (line 2); and to avoid compartment 4 emptying too far (FLC4R control), unfiltered oil is caused to recirculate if necessary (line 3). A common size bleacher would hold 1100–1500 kg of slurry per compartment, and the residence time per compartment could be 10 minutes (or as required). Hence, a throughput of 9 tons/hour can be accomplished. The float levels in compartments 1 and 4 control the starting and stopping situations; manual intervention is always possible. The outlet pump is protected by a sieve at the suction side, and delivers through a throttle followed by a sight glass. Compartments drain to empty, and a gap of a compartment full of slurry can be fed into the timing sequence to allow a clean change from one feedstock to another within minutes. Alfa Laval provides a bleacher operating in the same manner. The De Smet fully continuous bleacher (Fig. 5.3) also has proved popular. Here a flow of neutralized dried oil is dosed continuously with bleaching clay prior to entering the upper compartment of the bleacher (a), the whole being held under reduced pressure. Both closed and open steam are provided so that not only is the remaining free moisture flashed off, but also a very intense mixing occurs before the slurry overflows down a central duct into the lower compartment (b), impacting a baffle plate as it enters. Sparge steam is provided here also, causing the slurry to flow up and down over a series of vertical baffles. This creates the necessary residence time, and prevents forward mixing. A pump finally directs the emerging flow of slurry to a bank of enclosed filters which provide full continuity of operation. Oil or slurry is capable of being recycled (c) to the bleacher should this be required for a time. Finally, the clear oil is pumped via a polishing filter to a deodorizer. The intense mixing created by the sparge steam is claimed to maximize the bleaching effect of the clay and hence makes some worthwhile economy in the latter possible.

Bleachers

157

Several other plant suppliers offer bleaching plants. It is felt that the five mentioned here give an adequate illustration of the practical possibilities open to the refiner.

Fig. 5.3. De Smet continuous bleacher.

Chapter 6

Filtration and Filters H.B.W. Patterson

Factors in Filtration The removal of the bleaching clay from the oil is an extension of the bleaching step itself, since not only are these steps joined physically, but the filtration, if poorly performed, undoes some of the benefits of the bleaching. The membranes chosen, the filtration equipment selected, the overall speed of the operation, and the labor needed are all noteworthy economic factors. Filtration and the provision of the many different devices for performing it constitute both a separate study and an industry. In this chapter attention is drawn only to those practical considerations which make for success and the explanations of the reasons for their importance in the field of fats and oils technology. The Oil The more easily one layer of oil moves over another, the quicker it flows. Resistance to flow or viscosity is measured in centipoises (cP)1, and for most oils it falls considerably as the temperature rises. Therefore, at a filtration temperature of 90°C for nearly all of the oils considered here, the viscosity lies between 7 cP and 11 cP, which are the extremes. At 40°C, however, the viscosity increases threefold. Viscosity also increases as carbon chain length and degree of saturation increase. If, however, impurities such as soap and gums persist in the oil, these can clog the pores of the filtration membrane in addition to increasing the bulk of the clay particles upon and in which they are adsorbed. Today, degumming, neutralization, and washing are well capable of offering to the adsorbent an oil of less than 50 ppm of soap and 4 ppm of phosphorus (as gum) content, so that on their account no real difficulty need arise. In soapmaking and other processes where the crude oil itself may be bleached and heavier doses of clay are necessary, the situation is more difficult; the loading of the filter can be several times heavier per ton of oil treated, and it therefore fills more rapidly. 1

1 centipoise (cP) = 10–3 Newton seconds/m2

The Adsorbent In Chapter 3, numerous adsorbents are already described in some detail. Although clay and carbon particles are of an irregular shape, they can be classified on a practical basis by sieving and sedimentation tests. A typical particle-size distribution for a bleaching clay could be taken as:

159

160

H.B.W. Patterson

Size (microns)

Percentage (by weight)

above 80

10

40–80

25

20–40

30

under 20

35

As we shall see, when a suspension of such particles passes a textile woven from wool, cotton or man-made fiber, particles of all sizes become trapped in the net of yarns and even between the filaments making up the yarn. When the suspension meets an appropriately sized woven metal gauge, particles quickly form bridges across the pores in the weave, which are now of a standard size. If the pores are between two and four times the median size of the particles, most likely, in a very short time, from two to four particles of sufficient size will have collided in the pore and obstructed, but not necessarily blocked, the flow of oil. On the thin layer of cake which then forms, the majority of the filtration (i.e., the separation of clay from oil) finally takes place. The unaided eye can distinguish individual particles down to 40 µmL, but sufficiently 10 µmL particles can create a distinct lack of clarity in a fluid. If the median size of the particles is in the 20–40 µmL range, obviously, an 80 µmL pore will be most useful. If pore size is smaller than this, the rate of filtration will be lower unless the number of holes per square meter (open area) is increased; if pore size is larger, the time of the blackrun—until sufficient bridges form—will be longer. Furthermore, in the latter case, the chances of the extended bridge collapsing under pressure fluctuations will be greater, and this, at least temporarily, brings back some degree of blackrun. Therefore, a fair degree of tolerance exists on either side of 80 µmL for pore size, depending on the range of particle size to be filtered. The removal of fines up to 5 µmL improves the rate of filtration, but decreases bleaching activity, so more clay is needed. If the larger particles, over 100 µmL, are removed, activity is restored, and the improved filtration rate remains. This sort of screening must add to the cost of the final clay, but it is part of the technique used by clay manufacturers to market a more competitive product. This means that depth filtration provided by a bed of sand 0.5 m deep is now provided by a layer of activated clay only 1 cm deep. Although activated carbons and synthetic silicas may have smaller particle size, their filtration benefits when performed in the presence of a clay cake.

The Membrane Paper, textiles—natural or synthetic—and wire gauze or so-called metal cloth are the commonly used filter fabrics in the fats-and-oils industry. Certain requirements common to all membranes are mentioned, in turn, before going on to look more closely at each kind of membrane and how it, in particular, meets these requirements.

Filtration and Filters

161

The stability of the material of construction of the membrane under the conditions of use is an obvious necessity. This implies the recognition of any limitations regarding working temperature and the presence of water, acid, or alkali. Support for the membrane covers the question of the maximal pressure expected in operation. Next come the mechanical strength of the membrane’s construction and how well it resists distortion and wear in repeated use. Implicitly, in the design of the filter unit, mechanical support of the membrane is quite feasible without seriously detracting from its efficiency. The size and shape of the pore openings and how they are presented to the oncoming flow of slurry influence the ability of the membrane to encourage a rapid buildup of cake on its surface before allowing too great a penetration of fines and, therefore, early blinding or rapid accumulation of particles between filaments during successive filtrations. For example, acceptable filter rates may certainly be obtained from wire gauzes of 60 µmL to 130 µmL mesh aperture sizes, but the latter may be fouled more easily by the penetration of the smaller particles into the weave; the pore aperture being larger, the particle bridges which form are more vulnerable to pressure fluctuations which could cause them to collapse. A smooth surface allows easy detachment of the cake at cleaning time; it also allows cake to slide down more easily if the membrane is operating in a vertical position. In the 1970s, experience pointed to a wire-gauze membrane of an 80 µmL aperture being the best all-round solution to the filtration of popular bleaching clays if 350 kg of oil/m2/hour filtration rate was to be sustained for a working life of 3000 tons of oil/m2 over two years at least; much longer is possible, and flow rate is affected by many factors (see section Filter Units). The weave pattern, which confers rigidity and stability as well as porosity, is very important as later is noted in the discussion of textiles and wire gauzes (see sections Textiles–Fabric Finishing). These figures are compared with the results achieved on plate-and-frame presses with textile membranes of 200 kg of oil/m2/hour where 40–100 tons of oil/m2 are expected to pass the cloth before replacement is needed. In the 1990s, a modified weave of metal gauze gave an enhanced filtration rate since, although the individual apertures were reduced to 55 µmL, more existed per square meter; stability and strength also increased, and the continued selection of correct stainless steel overcame the corrosion problem. Brewing, pharmaceutical, and similar trades as well as fats-and-oils refining all benefited by these improvements. The smoother surface of the 55-µmL weave made the expulsion of the filter cake easier and more effective. At the same time, man-made-fiber textile overlays were developed to make the cleaning of plate-and-frame presses quicker and easier. Sharp focal points of pressure between membrane and support should be avoided because they are likely to encourage fatigue and rupture. The cost of installation and renewal, as always, is most important. Expensive membranes with a long life associated with big savings in labor may easily increase profits. In the case of plate-and-frame filter presses and candle-type polishing filters which depend on cloth membranes, their use continues

162

H.B.W. Patterson

until a cloth or candle is so blinded by repeated use that, independent of the cake deposited on them, they begin to offer an unacceptably high resistance to oil flow. They are then discarded; laundering cotton cloths is unsatisfactory. In the case of the metal-gauze membrane, held firmly in position, gentle countercurrent steaming, dilute caustic-soda washing, and a final hot-water washing, at intervals as appear necessary, not only remove particles which have lodged within the gauze, but also the films of oxidized and polymerized oil on the wire. The need to carry out this cleaning is judged by the resistance to flow (back pressure) developing earlier in a filtration cycle than experience shows to be normal.

The Procedure By commencing the filtration in a steady, gentle fashion and then continuing it within recognized pressure and flow-rate limits, the filter plant reaches its maximal utilization. If an unnecessarily high pressure is employed, particularly at the beginning of filtration, it can lead to the layer of cake forming at that time being compacted. This depends partly on the character of the clay in use, but the general effect would be to blind the filter medium earlier, to attain the limiting pressure more quickly, and therefore, lower the overall throughput or utilization by creating an unduly short filtration cycle. Usually, a 15-minute recirculation of the initial filtrate to obtain clarity is adequate; one-half of this time may suffice. During this period, although the flow rate may be quite brisk (little resistance is present), the pressure required will be below the average for the whole cycle. The thin layer of cake (starting with a clean filter) which forms may be between 0.25 and 3 mm thick, depending on the circumstances (Patterson, 1983). The flow rate may now be maintained or steadily increased to normal for the cloth or gauze being used. Flow rates (kg/m2/hour) for different membranes are quoted in the discussion of filter units (see section Filter Units). The pressure will increase steadily; if it is made to rise abruptly, clarity may be lost for a time (i.e., blackrun returns). Where batches of oil are being filtered in succession on the same filter, starting the second and later batches is easier and quicker because of the clay coating already established on the membrane. The total cycle time includes an allowance for unproductive time or downtime, most of which is devoted to cleaning. When a continuous service is required, the obvious answer is to have two filters—the filtration area of each being easily large enough to contain the cake resulting from the output of the bleacher while the other is being cleaned. Cleaning is likely to include: a period for the recovery of oil from the cake in situ, a discharge of the deoiled cake, and, in the case of a gauze membrane, a final brief backflush of hot water. For older-style batch operations with cloth-clad plate and frame presses, the calculation of the filtration capacity to cope with a succession of batches, possibly from a number of bleachers working synchronously, rests on the same basic approach. Since the tolerable loading of cake per square meter of filter is already well-known from experience with various membranes and since the throughput of the bleaching plant is central to factory planning, the calculation of the filtration area needed is not a sophisticated problem. One highly experienced manufacturer of filters (Veldkamp,

Filtration and Filters

163

1991) gives two readily understood criteria: required flow through filter (m3/hour) = m2 required tolerable filtration rate (m3/m2/hour) Well-known is that flow rate is influenced by the type of solid in suspension, liquid viscosity (temperature), pressure allowable, and the nature of the filter membrane. A large volume of case history is available to help here, and, if this is insufficient, laboratory and pilot-plant tests can be done. Another aspect is that even if the filter area is large enough to cope with the required flow, the filter itself may become full too quickly. Therefore, one must take into account the desired cycle length. If the cake volume associated with solid loading and the permitted cake thickness are known, a determination can be made by using this equation to calculate: Cake volume per cycle (m3) = m2 required cake thickness (m) The filtration time on one filter must allow enough time for the other to be cleaned. Whichever of these two calculations gives the larger answer, that area must be chosen. What ought to never happen, but sometimes has, is that the flow of filtered oil is impeded as it moves to the store tank or next stage by the layout of the line concerned (e.g., repeated sharp bends, brief constrictions, etc.). This, of course, contributes artificially to the failure to achieve the maximal throughput of which the filter is capable (see section Filter Units). Before making a final decision, one should consider the contingency of future changes in the type of feedstock. Also, in a large operation where several classes of oil (low-iodine value hardened lauric oils, soft marine oil, soybean oil, etc.) are all to be handled during the week, one must consider the segregation of plant items to handle specific classes; it can be essential. Some less-difficult segregation may be achieved by always scheduling the production of an item not in heavy demand to be carried through on freshly cleaned filters and so forth and then following it with a compatible oil. The ultimate precaution is to arrange production, mechanically or electronically, so that one isolated system cannot be made to function in conjunction with another without the help of engineering modification. This can begin with the couplings to offload tank cars.

Filter Aids The best types of bleaching clays operating on an adequately pretreated fat or oil ought not to require any filter aids. These have a cost: they hold back some oil; they increase the bulk of the press cake, which entails more frequent cleaning; and they increase the solid effluent. The reason for their use in edible-oil filtration is that not all bleaching clays are carefully graded to a range of particle sizes which achieves both

164

H.B.W. Patterson

high adsorption of unwanted material from the oil and also maintains a sufficiently porous structure in the cake, therefore permitting acceptably rapid filtration during a convenient working cycle. Some earths have a high proportion of fines, which may well help in rapid adsorption, but probably prolong the lack of clarity in the filtrate (blackrun) and then encourage early blinding. Similarly, an activated carbon which contains more than 40% of particles (w/w) less than 10 µmL in size raises problems of clarification if used on its own. A nickel catalyst, generated by the progressive thermal decomposition of nickel-formate crystals in an oil medium, contains a high proportion of nickel particles less than 1 µmL; it depends completely on the help of a support/filter aid in some form during use. Filter aids are finely divided and graded granular solids or, in some instances, thin fibrous solids. They must be chemically inert and insoluble in the fluid in which they are used.

Diatomaceous Earths Diatomaceous earth (diatomaceous silica, diatomite, kieselguhr, infusorial earth) is the most widely used filter aid. It is a nonmetallic mineral, and represents the skeletal remains of tiny aquatic organisms deposited in many parts of the world millions of years ago. Their outstanding characteristic is that their skeletons were based on silica and not carbon. Thousands of varieties existed, and, naturally, these deposits of amorphous silica may include minor amounts of crystalline silica (quartz), oxides of Ca, Mg, Fe and Al, and even some organic matter, depending on their location. After being mined, the deposits are processed by combinations of drying, milling, calcining, size classifying and so forth, as required by their end use. Their use extends far beyond fats-and-oils processing. One of the principal aims of the processing is to obtain some degree of standardization of particle-size distribution within the different grades which best suits their application. Table 6.1 shows this distribution for three of the several grades produced. Note how, in grade C, the proportion of medium (10–50 µmL) particles has greatly increased. Perlites Perlites are small pebbles of natural glass which contain a small amount of occluded water, and are found in volcanic deposits. When smashed and softened by heat, the

TABLE 6.1 Particle-Size Distribution of Diatomaceous Earths (Percentage Weight Basis) Grade

0–5 µmL

5–10 µmL

10–20 µmL

20–30 µmL

30–50 µmL

50–100 µmL

>100 µmL

A

23

47

25

3

2

0

0

B

5

14

43

15

10

7

6

C

4

9

35

19

13

13

7

165

Filtration and Filters

water vaporizes, leaving behind enlarged hollow spheres. These are again smashed, and give a filter aid of very irregularly shaped particles whose bulk density ranges 0.05–0.10 g/mL (3–6 lb/ft3). This bulk density is between one-tenth and one-fifth of many bleaching clays. The particle-size distribution of two different grades of perlite is shown in Table 6.2. As with other filter aids, be prudent and make a full-scale trial in working conditions to assess the full value—economic and technical—of any perlite grade chosen.

Fiber Cellulose fibers as a filter aid do not figure prominently in oil filtration except either as a paper overlay on a stronger support in polishing filters or as an overlay upon textiles in the main filtration which protects the cloth and makes cleaning easier. Man-made fibers have also entered this field successfully. An advantage of paper is that when it is discarded from the filter along with cake, it presents no operational difficulty. Because it tolerates acid and alkali conditions, cellulose fiber is used in the beverage-and-food industry as a precoat on a coarse support. It, in turn, may have a diatomite or perlite layer laid upon it to enhance its action in quickly giving a clear filtrate. Method of Use The two established methods of use of a filter aid are described as precoat and bodyfeed. In the precoat method, the fixed membrane is coated with a very thin layer, perhaps only 0.25–3.0 mm thick, of the precoat material by circulating a rather rich suspension of the material normally made up in the same fluid as is about to be filtered through the membrane. The deposit bridges the apertures in the fixed membrane with an assortment of particles of the correct size and shape to trap fines and still allow further porous cake to form easily. This operation lasts 5–15 minutes, and is conducted easily at only a moderate filtration pressure because the membrane or septum is clean and also because one must not compact the precoat layer being laid down. When the filtrate appears clear, the interstices in the thin cake should be adequate both in size and shape to retain the proportion of fines in the clay about to be filtered. The filtration proper may now commence, and the pressure gradually is increased to maintain the flow rate. To precoat the membrane by first circulating a moderate amount of the slurry to be filtered is feasible and indeed normal practice. For example, in a case in which the filter aid chosen was inappropriate, a polishing TABLE 6.2 Particle-Size Distribution of Expanded Perlites (Percentage Weight Basis) Grade

0–5 µmL

5–10 µmL

10–20 µmL

20–30 µmL

30–50 µmL

50–100 µmL

100–150 µmL

>150 µmL

A

7

9

29

12

17

12

13

1

B

3

4

8

6

19

34

14

12

166

H.B.W. Patterson

filter designed to remove 150-ppm particles down to a maximum of 10 µmL in size showed, after being coated with only 0.25 mm of so-called filter aid, a rapid rise in back pressure and a fall in flow rate during polishing from 8 to 2.5 tons/hour. Subsequently, in this case, merely circulating the first filtrate through the polisher for 5–10 minutes gave the desired result, both trapping fines and allowing polishing to continue at a normal rate and pressure. When the bodyfeed method is used, a steady addition of filter aid is mixed into the slurry as it passes to the filter. This solid, because of its high volume-to-weight ratio (i.e., low bulk density), maintains the ever-increasing layer of cake in a porous condition. Sometimes only a precoat is needed, sometimes only bodyfeed, and sometimes both. They do not necessarily have to be of the same material. Generally, the specification for bodyfeed is less demanding, and a considerably cheaper grade may be adequate. Definitely remember that if bodyfeed is used, not only must its direct cost and the cost of the oil it helps to adsorb be acceptable, but its bulk is helping to fill the filter. Hence, although the filtration rate is increased, the filter becomes full sooner. The optimum is when the longest time between successive cleanings is matched with the fastest flow rate. For example, if a good filtration rate results from the addition of bodyfeed at only 10% by weight of the solids already present in the slurry, this is satisfactory. The percentage may increase to 50%; as the added bodyfeed nears equality with the solids already present, the choice of this type of filter aid becomes increasingly questionable. One study (Purchas, 1967) showed how, when 1% of its own weight of bodyfeed was added to a slurry, an overall filtration rate of 20 gal/ft2/hour resulted. At a 2% addition, the overall rate rose to 30 gal/ft2/hour; at 3% a maximum of 37 gal/ft2/ hour was reached. Thereafter, the filter filled more rapidly, the overall rate dropped quickly, and a further addition of filter aid became counterproductive because of its own bulk. One way to assess a filtration problem is to filter about a 5-mm depth of a selected filter aid onto a Buchner funnel, and then filter some slurry. After gently draining the cake by suction, examine it to see if the filtered solids have formed a distinct upper layer or if they have penetrated the full depth. In the first case, probably a suitable cloth or gauze could be found which, after the usual limited recirculation of the first filtrate, would itself retain the slurry solids. What remains, then, is to discover how long the filter cake retained a useful porosity and whether some cheap bodyfeed would prolong this economically. If the slurry solids penetrated the depth of the precoat, this suggests that a bodyfeed will be needed and perhaps other grades of bodyfeed should be tested in the same way.

Filter Membranes Although already stressed (see section Factors and Filtration) was that it is the cake itself which quickly becomes the filter medium during filtration, the membrane plays a very important part in bringing this about. The durability of the membrane

Filtration and Filters

167

and its ability to give a repeat performance over many filtrations confer technical and economic success in most instances; only in the case of paper is the membrane discarded after use. Naturally, the nature of the filtration task varies considerably from brewing to oil refining to chemicals’ manufacture, and this gives scope for variety in both membranes and the filter machines which use them. One industry learns from another. In the discussion of filtration (see section Factors and Filtration), the basic conditions which affect the choice of filter were already listed. At this point, paper, cloth, and gauze are discussed in more detail, and the cases in which they are most useful in processing oil are indicated. More exotic materials—such as sintered powdered stainless steel, porous ceramics, interlocking fine metal discs, fiber-covered cartridges, candle filters, and so forth—belong to the field of polishing filters. For every edible oil, the golden rule applies—keep it clean, dry, dark, as cool as feasible, and out of contact with air.

Paper Paper sheets have the property of retaining very fine particles (10 µmL or smaller when required) while allowing high rates of flow of liquor. Consequently, they are used in many industries. Traditionally, the paper fibers are derived from wood or cotton, and their effect depends on the extent to which they were deliberately broken during manufacture. Asbestos and kieselguhr are incorporated for particular applications. In the fats-and-oils industry, paper overlays were placed over textile filter cloths in the main filtration to assist in the rapid buildup of what might reasonably be termed the “precoat layer” of the first bleaching clay to be retained; they protect the cloth and, since they are discarded along with the cake, the cleaning operation is much easier. One may use finely woven man-made fibers to serve the same purpose. Later in the process sequence, paper membranes have a use in removing fines so that less than 100 ppm pass on to deodorization, saponification, or other stages. One may also use them as the last precaution before loading oil for delivery to customers. Obviously, the paper membrane is supported upon cloth or gauze while cake accumulates upon it. Oil has little or no effect on its bursting strength, but when the paper is wetted with water, as when washing or steaming in situ, this strength collapses. This, in fact, is an advantage because, on the final opening and cleaning of the filter, no resistance is offered. Presumably no difficulty arises in the disposal of the deoiled filter cake because of its presence. The paper overlays used in the main filtration are of a fairly cheap, coarse grade without special additions made to the pulp from which it is manufactured. The paper may weigh 50–100 g/m2 (ca. 80 g/m2 is common); crimping increases its bulk. The dry bursting strength’s range is 90–175 kiloPascals (kPa) [A Pascal is a unit of pressure equal to a force of one Newton acting over one square meter. The Newton (N) = kg m s–2]. The wet bursting strength is about one-half of the dry bursting strength. Among industrial filter papers, the retentivity of this grade is low-to-average. Some manufacturers express the porosity of their papers as the milliliters of water passing through one square meter in one minute under a head of 1 cm. Thus, a paper of 75

168

H.B.W. Patterson

g/m2 might have a porosity of 230–340 mL/m2/minute, and a heavier paper of 95 g/m2 has the lower porosity of 140–230 mL/m2/minute. A paper’s permeability or air resistance is defined as the pressure differential (in kPa) across the paper which sets up an air-flow speed of 10 m/minute. Thus, the expression representing this permeability is Gp10. A light overlay or facing paper of 56 g/m2 may show a G10 of 0.45, whereas a heavier paper of 90 g/m2 requires a Gp10 of 1.12 to achieve the same result. Such papers are expected to show no more than a faint acidity. When it comes to the task of polishing, the emphasis shifts to a good retention of fine particles. Thus, a heavy, strong paper may be supported on a coarse, metal screen. On the other hand, when a tightly woven textile is used at the polishing stage in a second filtration, this textile is sometimes overlaid by some processors with a cheap paper as a measure of protection. In a polishing task, the buildup of cake is very gradual; therefore, a dozen or more batches of oil may be filtered before it is deemed advantageous to renew the overlay. In cases in which the paper itself is the main polishing agent and is expected to retain particles in the 10 µmL range because they can create a haze if present in sufficient concentration, then a paper of about a 220-g/m2 weight, with a dry bursting strength of 390 kPa, a wet bursting strength of 160 kPa, and an air resistance of 5.0 kPa is likely to be used. In those industrial processes in which the filtered cake is to be given some treatment after discharge or where the presence of paper would be a nuisance, a finely woven synthetic textile is used instead.

Textiles The usefulness of a textile as a filter membrane depends on: (i) the physical and chemical properties of the basic material; (ii) how the filaments are made up into yarn (when this is done); and (iii) how yarn is woven into sheets. Some degree of compromise between conflicting requirements is necessary. A need exists to retain fine particles, yet avoid early blinding; a high flow rate of oil per unit area is welcome, and the detachment of the final cake from the cloth should be easy. The performance in use decides final acceptability. Some of the properties of popular textiles are set out in Table 6.3 (P and S Filtration Ltd., United Kingdom). Cotton Cotton once claimed a dominant position in oil filtration, but this was eroded first by synthetic fibers and later by sophisticated stainless-steel gauze. It has good strength and resistance to wear, and its ability to retain fines is also good because of its hairy filaments. Its working life at 90–100°C amounts to several months. It withstands cold, diluted acids but not hot, weak acids or cold, strong acids. Alkali causes some swelling but not a disruption. Polyamide (Nylon) Polyamide (nylon) is very popular, having good resistance to abrasion, high tensile strength when dry or wet and good elasticity, and it provides a smooth surface for cake

169

Filtration and Filters

TABLE 6.3 Fiber Properties Cotton

Wool

Polyamide (nylon 66)

Polyester (terylene)

Polyolefin (propylene)

Teflon

Specific gravity

1.5

1.31

1.17

1.38

0.92

2.1

Melting point (°C)

—-

—

255

255

165

350

Acid

no

yes

no

yes

yes

yes

Alkali

yes

no

yes

no

yes

yes

Organic solvents

yes

yes

yes

yes

no

yes

Resistance to:

Dry heat (°C)

100

110

110

150

100

260

Wet heat (°C)

100

100

110

100

100

260

Flexing wear

good

very good

excellent

excellent

good

fair

discharge. In time, oxidizing agents and mineral acids can degrade it, but it resists alkalis and organic acids. As a continuous working temperature, 110°C is suggested, but occasional exposure to 140°C is acceptable.

Polyester (Terylene and Many Other Trade Names) Polyester makes up strong, abrasion- and flex-resistant cloth which, during finishing, may be shrunk so as to give a tight fabric with excellent retention of solids. It tolerates oxidizing agents and most acids, but is degraded by hot, strong alkalis (10% of caustic soda at 100°C); it has fair resistance to weak alkalis (3% of sodium carbonate at 70°C). Since alkali attacks oil before terylene, the situation in oil filtration is easy. Yarn Production The synthetic fibers nylon and terylene are extruded as a smooth continuous monofilament of solid cross section. Natural fibers such as cotton and wool have fibers of limited length (staple) and irregular surface contours which contain voids. The rough surface helps to trap fine particles. Not being completely solid, they are more compressible. Monofilaments can be and are woven directly into cloth. Obviously, the diameter of the filament and the pattern of the weave have a big influence on the filtration characteristics of the cloth. In general, such a cloth would be best used in filtering coarse or fibrous solids. It allows a good cake discharge and may blind easily, but is easily cleaned.

170

H.B.W. Patterson

Multifilament yarns are produced by twisting together two or more filaments, producing a yarn of great tensile strength. The degree of twisting, which may be from 2 to 15 complete turns in as little as 2.5 cm, affects the retentivity. Thus, as the degree of twist increases, the ability to trap fines decreases. Lower-twist yarns trap fines better to the extent that fine particles become caught between the filaments in the yarn as well as in the spaces between the strands of yarn in the weave. In the textile trade, the expression “bunding” describes a situation in which the individual strands of yarn have become so swollen by the solids trapped among their filaments that the interyarn spaces are almost closed. Of course, filtration in any weave can be greatly impeded when the open spaces of the weave are almost blocked by particles. Filter-cloth users speak of all cloths which have lost their porosity as bunded, but the distinction just explained is worth noting if an attempt to improve performance is being considered, either by cleaning the existing cloth or replacing it with one of different yarn or weave construction. What was just described does not prevent the multifilament yarns from being woven to give a cloth of smooth surface, thus making cake discharge and cleaning comparatively easy.

Spun-Staple Yarn Natural fibers such as cotton and wool have a very limited length (the staple) and so must be twisted together (spun) to provide a continuous strand. Synthetic filaments may also be cut to staple lengths and then spun to continuous strands. In each case, the spun staples are a class of yarn. As far as synthetics are concerned, these spun staples can themselves be twisted together before going on to be woven into a fabric, and in this way the necessary bulk for the cloth is ensured. The expressions “cotton spun” and “linen spun” refer to tightly twisted fibers spun into yarn, whereas “woolen spun” indicates less twisting so that considerable filtration goes on between the single filaments of a single yarn as well as between the interyarn spaces of the weave. Retentivity and flow rate with spun-staple weaves can be good; because their surface is coarser than that of mono- and multifilament cloth, cake discharge is not as easy. As an indication of the behavior of different fibers in working conditions, the following broad assessment in Table 6.4 was made by a cloth manufacturer (Anon., no date).

TABLE 6.4 Filter-Cloth Loading of Trapped Solids (Percentage by Weight) Type of yarn

Washing advisable

Complete blinding

Monofilament

5

10

Multifilament

10

20

Spun staple

20

40

Filtration and Filters

171

Weaving Only a brief description of basic types of weave of interest to filter cloth users is given here. Fortunately, the terminology is the same in weaving textiles and the increasingly popular wire gauzes (metal cloths). Warp As the fabric moves steadily from the loom, those strands (yarns, filaments, and wires)—which run in the same direction along the length of the sheet—form the warp. Spacings of the warp strands are according to the spacings of notches cut in the reed at the head of the loom. Individual strands or groups of strands can be lifted momentarily above their neighbors as required. Weft As the warp strands are lifted, a shuttle rapidly drags a strand across the fabric, passing under and over the warp strands or groups of these in sequence. This is the weft. Both warp and weft strands can be tightly or loosely spaced; they may be of different diameters and even of different materials. Plain Weave Plain Dutch or Hollander weave [Leinen bindung (Ger.)] is the simplest of all weaves—the warp and weft strands pass over one another from top to bottom and side to side, one up, one down. The even, widely separated warp and weft in Fig. 6.1 give a “straight through,” virtually square, opening. Such an arrangement would be

Fig. 6.1. Plain weave.

172

H.B.W. Patterson

mechanically weak and, with openings sufficiently large, of poor retentivity; hence, it would have a high flow rate and a low tendency to blind. Obviously, if the warp and weft are more tightly spaced, retentivity will improve, porosity (flow rate) will decrease, and bunding will tend to increase. In the case of wire weaves, if the warp and weft are each slightly bent at the set intervals to help them pass over each other, this is known as double crimping. Twill Weave. [Köper bindung (Ger.)] As is shown in Fig. 6.2, weft is made to pass over and under pairs of warp, but a progressive shift occurs diagonally regarding which warp the weft passes over. The openings are of the straight-through kind, and a gain is made in mechanical strength. When not single, but groups of strands—warp and weft—are made to behave in this way, a braided or basket weave results. Twill-woven textile fabrics have medium retentivity and resistance to blinding; they resist abrasion very well and show high flow rates. This good balance in characteristics accounts for their wide popularity in textile-fiber cloths. Satin Weave [Funfschaftköper bindung (Ger.)] In Fig. 6.3, warp and weft are each shown as passing over four strands, under one, then back to the surface again. The number passed over could be three or five if desired; the name in German or French indicates how many. In textiles, this weave has low retentivity, high resistance to blinding, and, as might be expected because of its smooth surface, an excellent cake release.

Fig. 6.2. Twill weave.

Filtration and Filters

173

Fig. 6.3. Satin weave.

Needle Loom or Batt-on-Base Fabric This is a nonwoven fabric which has some use in the chemical industry, but was not found to give any advantage in filtering oils.

Wire Gauzes (Metal Cloths) Plain Dutch Weave [Tressengewebe (Ger.)] is very widely adopted for wire cloth or gauze. In some constructions, the warp wires may have twice the diameter of the weft, and they are spaced at intervals as are the warp strands in Fig. 6.1; however, the weft wires are placed next to one another so that the open-square apertures, instead of being square, become roughly triangular and oblique to the surface of the membrane. Since the wire filaments rigidly maintain their position relative to one another throughout the use of the membrane, their diameters, spacing, and manner of weaving are all vital to fixing the size, shape, and inclination of the opening across which the bridges of clay particles form very early in filtration. The diameter and spacing of the wires govern the strength of the gauze, and the free-space per-unit area affects the porosity or flow rate. The surface of the gauze must be smooth enough so as not to impede the removal of the cake, and the metal chosen should resist corrosion, which may arise when acid-activated clay was used, and the cake comes to be water-washed and/ or steamed. For this purpose, the A1S1 316 SS alloy with a high nickel content has proved very effective (Young, 1983). Some explanation of the terms used in describing a wire gauze will aid in understanding the way in which it is constructed. A mesh is the distance from the center of

174

H.B.W. Patterson

one filament to the center of its neighboring filament running in the same direction. That is, a mesh is the distance from warp to warp running along the length of the fabric and from weft to weft across its width. If these distances in each direction are the same, the weave is said to be square mesh, and when longer in one direction, it is called rectangular mesh. A mesh size is given as the number of such distances which are found in a lineal inch of 25.4 mm in each direction. Other trades have employed a different “inch,” such as the silk weavers’ inch of 27.7 mm. By convention, the warp number comes first, then the weft. Immediately following, come the respective diameters (mm) in the same order. The workhorse of metal-gauze filtration membranes for many years—and still widely used—is a Plain Dutch or Hollander weave: 24 × 110; 0.314/0.254. The retentivity of such a gauze was classed as approximately 140 µmL, depending somewhat on wire diameters and the manufacturer, being the dimension of the largest particle capable of penetrating the open channel between warp and weft on its own. Further development went on in the 1970s to improve retentivity to 60–80 µmL with an equal or even higher porosity. The brewing industry was especially interested, and the gauze manufacturers Gebr. Kufferath of Düren played an important part. An armored Hollander weave resulted in which the retentivity was reduced to 80 µmL. This is a Panzertressenge-webe (PZ80), and a typical weave would be 132 × 36; 0.193/0.396. This weave then gained considerable popularity in many industries, including the fats-and-oils industry. Important to note is that this weave is exceptional because: the weft is much thicker than the warp, and such a strong gauze is produced by packing these strong weft wires as closely as possible (i.e., 36 in a 25.4-mm length with the distance between centers 0.71 mm). Unusually, now the thinner warp wires must bend abruptly as they pass over and under alternate wefts. When examining a sample under magnification, the viewer must be prepared for this; otherwise, it may be mistaken for a plain Dutch weave (which it closely resembles) with the basic pattern simply rotated through 90 degrees when viewed from above. Further advances have been made since the 1970s with a view to improving retentivity, flow rate, strength, and so forth. Not surprisingly, in view of earlier remarks concerning the benefit of twill-pattern weaving in textiles, a twill pattern in wire gauze is now being used with great success. This is the Köper Panzer-tressengewebe 55µmL (KPZ55), (Twill-armored weave 55 µmL). On the basis of the nomenclature described above, a typical weave would be 170 × 46; 0.15 × 0.30. Here again the weft is the thicker wire, and the thinner warp bends around it. Figure 6.4 shows this and how a twill pattern results. Figure 6.5 shows the three gauze weaves just described. Not only does the KPZ55 have a higher retentivity because of its smaller apertures, but also because the number of these was increased to provide a larger total open space per square meter, the flow rate (porosity) per square meter increases considerably. Thus, for a given task, a filter of smaller area becomes feasible. Advances in the comparatively minute detail of the filter weave have heightened the success of the automatic horizontal-plate self-cleaning filter, and huge examples of up to 200 m2 are now available (Schenk Filterbau Gmbh, Germany/Maryland). The gauzes are

Filtration and Filters

175

Fig. 6.4. KPZ55.

sufficiently strong and rigid to maintain a consistent retentivity from one year to the next. The recommendation is that the mesh be well-supported and constructed as part of a multiple unit so that whenever the leaf assembly is lifted out of the filter, the risk of distortion/damage is minimal. Although warp and weft numbers are widely quoted on the same basis as used here, wire diameters may be quoted from different national gauge systems. Only

176

H.B.W. Patterson

Fig. 6.5. Important weave types.

millimeters were used here for the sake of simplicity. Sometimes, when the mesh number is multiplied by the warp or weft diameter, the result exceeds 25.4 m. This is not necessarily an error in specification, but means simply that all wires no longer lie strictly in one plane but that alternate wires are forced to lie above and below the mean position because of the close packing. A common practice is to improve the smoothness of the surface of a wire gauze by carefully controlled rolling between steel rollers. Obviously, the distortion of the weave must be avoided. The process is known as calendering. The gauze rests on coarser screens whose thicker wires were crimped to pass easily over one another. At the crimped points, the wire forms knuckles, and these must be flattened to a degree by calendering so that they do not form focal points of metal fatigue or rupture by sharp pressure on the gauze. Some textile cloths are rolled in this way after being woven.

Fabric Finishing After the most suitable material, size of filament, and pattern of weave are selected, the resulting fabric is often given a final conditioning to prepare it for use. Thus, synthetic fabrics are likely to be heat-set by increasing their temperatures for a period to make sure of their dimensional stability up to the recommended maximal operating temperature. While this is taking place, the shrinking of the warp and the weft may be freely allowed. This procedure can be used to produce very dense fabrics when required. Similarly, cotton-filter fabrics are always shrunk by being dipped one or more times in boiling water (shrunk, double shrunk). Even after this, allow some further shrinkage, during use, of about 3% in width and 5% in length. Sometimes this is applied to synthetic fabrics. Blends of separate natural and synthetic fibers, such as cotton and nylon, can be used for some filter cloths used in oil filtration. In plate-and-frame or recessed plate presses, the edge of the filter cloth is most subject to mechanical abrasion. Protection is obtained by the impregnation of the edges of the prepared cloth piece with neoprene, and this also reduces the seepage of the oil through the fabric at this point. Manufacturers of textiles may offer further conditioning of the cloth to enhance its resistance to blinding, ease cake discharge, and even improve washability. The increased cost must be evaluated against the increased performance.

Filtration and Filters

177

Filter-Cloth Selection The grading of bleaching clays to a well-defined particle-size distribution and the fact that densities and viscosities of the oils handled at the usual filtration temperature of 90–100°C nearly all lie tolerably close to one another suggest that the definition of a membrane which will obtain an efficient separation of solid from liquid can be made easily and with precision. Accepted is that various requirements conflict with one another, so the result has to be a compromise which ensures high quality of filtrate and economic costs in time, material, labor, and energy. Just as experience has shown what qualities of paper are best suited for different filtration tasks (see section Filter Membranes), similarly, indicating the qualities of cloth most successful in oil filtration is possible. Wire gauze was already considered (see section Wire Gauzes (Metal Cloths)). The weight of the cloth is an important item in its specification, and relates to the type of yarn and weave: 1 oz/sq yard = 33.90 g/ m2; 1 g/m2 = 0.025 oz/sq yard. The porosity (permeability) of the cloth is a second important item, and also is affected by the type of yarn and weave. Permeability may be expressed as the cubic meters of air flowing through one square meter per minute under a pressure drop of 25 mm WG (water gauge). On this basis, different weaves would fall into three groups:

Close weave Medium weave Open weave

under 3 m3/m2/minute (Separation of fine particles). = 3–18 m3/m2/minute (The usual grade employed on filter presses). = 18–305 m3/m2/minute (Separation of coarse particles, especially on rotary drum filters).

Some manufacturers continue to quote permeability measured with a pressure drop of 12.7 mm of WG. As is evident, for oil filtration, natural and synthetic cloths come in the groups of close or medium porosity. For decades, cotton cloth has been used for oil filtration. A typical weight would be 680 g/m2 (20 oz/sq yard), and a double-shrunk twill weave has been normal. A paper overlay is often used since it helps delay blinding and eases cleaning. Where a considerable proportion of fines (5 µmL) is present, as in catalyst filtration, a double thickness of cloth improves retentivity, but it is still essential to lay down a thin layer of solid on the membrane by recirculation at a low pressure. Once a clear filtrate is established, it will not contain more than 30 ppm of fines; this is well below the concentration needed to promote the oxidation of unsaturated edible oils. Terylene staple in twill weave at 540 g/m2 and with a permeability of 6–9 m3/ m2/minute at 25 mm of WG works well as a filter and is often used with an overlay. When an overlay is not used, a heavier twill-weave terylene cloth of 680–720 g/m2 and a permeability of only 3 m3/m2/minute would be chosen. As an alternative to a

178

H.B.W. Patterson

paper overlay, a light, very closely woven terylene multifilament plain-weave cloth of 240 g/m2 and low permeability (3 m3/m2/minute at 25 mm of WG) may be used. This can make cleaning even quicker than when using paper, and possibly, after some time in use, it can be reversed so as to obtain a cleaning action which prolongs its useful life. With such a procedure, the filtrate of the first batch of oil after reversal contains some fines flushed from the interyarn spaces of the overlay. This first filtrate, therefore, needs to be refiltered, after which normal routine is resumed. The specifications of weight and weave for nylon are very similar to those for terylene. Synthetic cloths have a much longer life than natural fiber, and their huge production has improved their competitive costs.

Filter Units A wide variety of items is described in this chapter which relate to the choice of a filter unit. Before going on to consider a limited number of units, some further general requirements must be mentioned. Filtration Rates Versus Cake Loading As already suggested (see section The Membrane), a typical flow rate for a cloth-clad filter press is 200 kg of oil/m2/hour, whereas for gauze 350 kg/m2/hour would be quite feasible. However, some estimates for the gauze would place a rate for a caustic refined oil as 250–300 kg/m2/hour and for an acid-degummed/bleached oil due for physical refining, as 200–250 kg/m2/hour. Such estimates envisage a clay dosage of 0.5–3.0% of clay/oil, up to a loading of 10–12 kg of original clay/m2. If for any reason, such as coping with a quick succession of batch filtrations, a high specific flow rate is sought (i.e., almost double that when serving a steady continuous flow), then accepting a final loading of only 4–5 kg of original clay/m2 would be more appropriate. Again it appears that where heavy doses of 3–6% clay are used, as in bleaching soapmaking oil, it may indeed be possible to reach 18–20 kg of original clay/ m2 because of the rapid buildup of cake and continuing porosity. But in bleaching edible oils where the dose is only 0.5–3.0% of clay and the buildup is much slower, a loading of 12–14 kg of original clay/m2 is a safer planning figure. All that was just said refers to the carefully-sized graded clays described in Chapter 3. When it comes to clays manufactured in other areas of the world, the sizing may not be so advantageous. In this case it may well be that a loading of only 8 kg of original clay/m2 for a flow rate of ca. 200 kg/m2/hour must be accepted. An alternative approach would be to add to such a clay some 10–20% of its own weight of filter aid (see section Filter Aids). Obviously, in planning filter requirements, the quality of oil, quality of clay, and size of clay dosage must be taken into account. As mentioned previously (see section The Procedure), in addition to flow rate and cake loading, the dead time in a filter’s cycle must be recognized and an appropriate arrangement made. For example, in instances in which the clay dosage is low, possibly 20 hours would be needed to fill the filter and 2 hours to clean it at the national standard rates. In a very different situation, the filter might fill in 4 hours and still require 2 hours for

Filtration and Filters

179

cleaning. In the first case, an agitated insulated buffer tank could hold the oncoming slurry for 2 hours while the filter was being cleaned. A filter whose capacity allowed for the modest increase in the performance of 22 hours of slurry output being filtered in 20 hours might be chosen. In the second example, the obvious choice is to have two similar filters. Dead time is taken to include recovery in situ of oil from the cake in the full filter, discharge of the cake, and preparation of the filter for its next use. In the case of a rotary-drum filter, this would entail precoating; in the case of a gauze filter, this might entail a brief countercurrent steaming. Remember that all press cakes remain undisturbed during filtration because of the danger of blackrun. At least with filter presses, in which every chamber outlet was provided with a little inspection window in the flexible tube which connects it to the common exit manifold, a sudden lack of clarity can be seen and the offending chamber shut off. Most automatic self-cleaning filters do not offer that facility. Cakes which are held horizontally are, however, more secure. If lack of clarity is detected, the first option is to return to recirculating until the crack in cake on the gauze is made up, and clarity is restored. The smoother a vertical membrane is made to assist discharge, the greater is the risk of cake sliding down and leading to an irregular cake thickness. This, in turn, may adversely affect the efficiency of any in situ washing which may be used as a routine. A common precaution against such troubles is to position a guard-cum-polishing filter immediately downstream of the main filter. This must not impede the flow. In untroubled circumstances, its cleaning may be necessary only after relatively long intervals of use, and is simple. Double cloths on a press reduce risk since damage to the upper cloth may be noted during cleaning. Planned maintenance (replacing the cloths at a fixed interval which is shorter than their average life) increases cloth consumption and reduces the risk of failure. For automatic solid discharge filters, troubles were experienced in the early 1970s with certain main discharge valves, due to particles lodging in the valve seating and later giving rise to leakage. Understandably, the modern design of the slide valve has overcome this, and, in many cases, butterfly valves have also proved successful.

Filter Presses For very many years, filter presses have been used around the world in one or other of three designs: (i) the plate-and-frame press; (ii) the recessed-plate press, in which extra thickness around the periphery of the plates does away with the need for an intervening frame; and (iii) the washing press, in which an inlet above every alternate plate enables the wash to enter behind the cloths of two adjoining chambers, pass right through the cake in each, then through the opposite cloth and out. This latter design is, of course, favored when the cake is the valued commodity. A cake chamber is considered full when seven-eighths of the space is occupied, the remaining eighth being the narrow distance between facing cakes. Manual operation is heavy, dirty work when the press is made from cast iron or some form of steel, but is less arduous when it is made from aluminum, wood, rubber, or other nonmetallic material. The procedure has now been greatly eased by the introduction of electrically-

180

H.B.W. Patterson

driven mechanical aids which open and close the press and move the plates in either direction, causing cake discharge under the direction of the operator, whose safety is well-guarded. Automation was also developed to start and stop filtering, blowing, washing, and drying. Table 6.5 is a summary of the principal improvements in filterpress design apart from the material of construction. Considerable variety exists in the detail of all three types. In conclusion, synthetic cloths and automation are the principal items sustaining the trade in filter presses.

Self-Cleaning Enclosed Horizontal-Leaf Filter Units A filter outstanding in its saving of labor, affording obvious cake stability, excluding air contact for edible oils, resistant to the penetration of fines (high retentivity), having acceptable ease of cleaning, available with steam/hot-water jacket, with a long life and modest demands on floor space is a filter which requires attention and evaluation, not only in the field of the filtration of edible and technical oils, but also much more widely. This, in fact, has been taking place progressively since the 1970s. Just as in the case of filter presses, different types of such a filter have evolved to meet the needs of a wide range of industrial filtration steps. The description which follows is naturally biased toward the needs of the fats-and-oils industry, particularly where edible products are concerned. Filters of this type are available from several companies; the one described here is made by Schenk (no date), and is designed to discharge the cake as a solid, possibly somewhat pasty in texture. The disposal of the cake in this form is almost certainly easier, but another model will eject it as a slurry. Figure 6.6 shows the important structural details of the filter with solid discharge; the operational sequence follows. Oil/clay slurry is pumped into the filter at the bottom (Patterson, 1989) and top (Sonntag, 1979) while a vent is open to allow air (or inert gas) to escape. Pumping TABLE 6.5 Filter-Press Improvements Manual press For

Against

Wide application Cheap

Semiautomatic improvement (About double cost)

Strong No rest volume filtration (blow clear) Floor space per m2 filter surface acceptable Individual chamber closure

Unattractive work Laborintensive Appreciable dead time

Cloth damage

Reduced cloth damage

Much easier work Laborhalved Reduced dead time

Filtration and Filters

181

Fig. 6.6. Schenk self-cleaning enclosed horizontal-leaf filter (solid discharge type). (1) Upper inlet (shared with lower inlet); (2) Shaft compression device; (3) Inlet distribution and spacing ring; (4) Supported filter elements; (5) Supported filter elements; (6) Filtrate discharge shaft; (7) Sight glass; (8) Cake discharge plough; (9) Cake discharge; (10) Seal flush; (11) Hydraulic drive; (12) Central entry area for whole-shaft assembly; (13) Optional spray assembly (included as may suit the employment intended); (14) Four scavenge plates; (15) Lower inlet; (16) Shaft seal; (17) Scavenge filtrate outlet; (18) Main filtrate outlet; (19) Hydraulic motor.

182

H.B.W. Patterson

slurry to the top simultaneously in this way offsets the tendency for solids to settle and preferentially load the lower plates more heavily. As the last air escapes the main filtrate (Davies et al., 1989) and the scavenge filtrate (Andersen, 1962) valves open, the vent closes. The filtrate, now commencing to flow through the gauzes [4, 5, etc.] down the hollow central shaft [6], is directed back to the feed vessel and, although the flow is ca. 700 kg/m2/hour, the pressure inside the filter is only about 2.5 atm, dropping some 0.5–0.8 atm in passing the filter gauze. Possibly, 26 gauzes [4, 5] and four rest-volume (RV) or scavenge gauzes are involved (FAO, 1980). Main gauzes are 30 mm or more apart. Filters of 5 m2 to 200 m2 are made. As soon as the filtrate becomes clear (after about 10 minutes including filling), it is directed to the filtered oil tank. The recirculation line and the exit from the scavenge plates are then closed. The scavenge plates, like the main plates, have by now become coated with a thin layer of clay. In some industries, a precoat may be applied in this way. Filtration continues for 30 minutes–10 hours, according to the concentration of clay in the oil, while the pressure starting at about 1.5 atm rises slowly to about 5 atm until 80%, at most, of the depth between plates is filled. The general rule is to have at least a 10-mm depth of cake at the end so that it can easily be flung off, and always to leave at least a 10-mm free space above the cake so that it will not be obstructed when leaving. This spacing of the plates can be more or less according to the user’s requirements. Typically, the flow rate will have fallen from ca. 450 kg/m2/ hour at the outset to ca. 200 kg/m2/hour at the end, so averaging 350 kg/m2/hour for the usual dosages of 0.5–3.0% of clay used in edible-oil bleaching. In this situation, a final cake loading of 12–14 kg of original earth/m2 is typical, as mentioned previously. The feed pump is stopped, and the feed valves are closed. If required, the line to the filter may be blown clear with nitrogen or drained, if a different oil is to follow. Nitrogen pressure is applied at the top of the filter, causing the RV to continue to filter. This pressure is not allowed to exceed the maximum reached at the end of filtration since that might lead to an earlier penetration of the exposed upper cakes, thus prolonging RV filtration. In a few minutes, when one-half of the plates are exposed and bubbles begin to appear in the filtered oil, the exit of the four scavenge plates is opened and the exit of the main plates are closed. Filtration now continues on the scavenge plates until, in a few more minutes, the lowest one is almost exposed. Then, automatically, the heel of the RV is drawn off via the lowest point by a special small pump provided for the purpose and rapidly pumped to a spreader plate above the top main filter plate, which becomes flooded. In other makes of filter, the main filter pump may be used for this purpose. From there the slurry flows down over one leaf after another. The exit of the main filter shaft opens at this point so that some of the RV is filtered on the main plates as it descends, and any cracks which may have formed there can be sealed. Although experience shows that the four scavenge plates are, in fact, adequate to take all the rest volume, this maneuver of making use of the main plates again speeds the operation to the maximum. The pumping rate of this transfer

Filtration and Filters

183

should be above the average main pumping rate so the cakes do not dry and crack while it is taking place. The loading in these circumstances probably reached only ca. 7 kg of original clay/m2 on the scavenge plates as against ca. 13.5 kg of original clay/ m2 on the main plates above. If a heavy dose of clay has been used (e.g., above 3%, as in bleaching technical oils), shortly before the main filtration is due to end, some clear filtrate can be diverted back to the feed vessel, thus diluting the concentration of clay in it. This diluted slurry is, then, what forms the RV. Final draining of the filter is by a brief blowing with nitrogen with main and scavenge valves open. This step may take the form of merely allowing the internal nitrogen pressure to dissipate via these filter lines (0.5 minutes) or continuing to blow up to ca. 5 minutes if there is any advantage in so doing. Since the loosely held oil is now displaced, a number of options are open to the refiner. The oily cake may be expelled by first opening the discharge valve (Duffy, 1984) and then rotating the stack of filter plates at 240–300 rpm (Burr & Burr, 1929, 1930). With gauze being flat and not dished and probably being of the KPZ55 smooth surface twill weave, everything is in favor of the cake leaving the plate within 5 minutes. The Schenk filter has a low flat dished base (Fig. 6.6) swept by plough blades (Gunstone, 1984) as the shaft rotates; further, the KLK model is slightly conical in shape (divergent toward the base), so there is increasing space between the edge of the plates and the wall of the filter as the cake slides down the wall to the exit. The cake may be removed for solvent extraction of fat elsewhere or within the same workplace, in which case safety regulations governing the use of solvent (probably hexane) will apply. This releases the filter after a brief cleaning procedure to filter more slurry. The extraction may be done by pumping a hexane wash through the filter. Instead of hexane, hot water may be used and the same water recirculated a number of times from the base of the wash-separating tank. If a hexane wash has been used, the cake must be freed from hexane by blowing it with nitrogen, or if hot water has been used, the cake is dried by a current of steam to a moisture content of ca. 25% of H2O, after which it is expelled as described above. When gas, steam, or air has been used under pressure, a period of 30 seconds of decompression must always be allowed before opening the filter. More details concerning the use of hot water to extract filter cakes follow in Chapter 7. If a batch procedure is in use and the same kind of oil is to follow because the filter has not yet been filled with cake to capacity, there is no need to go through the RV filtration step; the feed continues if available. If it is delayed, the filtrate should be temporarily recirculated to the feed vessel until the next batch is ready. This avoids disturbing the cake, and prevents the clay in the RV from settling onto the floor of the filter. Maintaining the gauze in a clean and, therefore, efficient condition avoids much harassment and loss of time. The measures to achieve this relate to the conditions of use. Blinding depends on weave, grading of clay, particle sizes, and the sequence of pressures used. A real but longer-term hazard is the fouling of the mesh with gummy deposits. Longer term still is the corrosion and pitting of the gauze by the choice of an inadequate

184

H.B.W. Patterson

steel; the latter question also relates to whether the manufacturer of the activated clay has chosen the proper amount and type of mineral acid for its activation. The chloride ion bleached from clay particles lodged around the periphery of the gauze can be a source of corrosion since steaming is probably featured in the complete cycle. After what has been said, what practical steps are open to users of gauze filters to obtain the best results over time? The following cautions and suggestions may be taken to apply to gauze pressure leaves in general, and should be checked with their suppliers. If hot water has been used to extract part of the oil in the cake, a period of forward steaming follows; this partially dries the cake. A saturated steam flow of 50 kg/m2/hour at a maximum of 3.5 atm (ca. 140°C) lasting 10 minutes will probably be adequate; 20 minutes is the maximum. A maximum of 140°C safeguards valve membranes; prolonged steaming bakes cake onto the gauze. The cake is then flung off the gauze plates as described previously; some detachment of adhering particles may now begin if this is found worthwhile. It may be sufficient as a matter of routine between each cycle to pass steam via a reducing valve into the filtered oil line, up the hollow central shaft, under the filter leaves, and out of the atmospheric vent at the top of the filter. A condensate drain valve open at the base of the filter allows water to escape; the leaves are rotated. The reduced pressure of the steam as it enters the filter is no more than 0.7 atm, and falls rapidly inside the filter to 0.2–0.5 atm. Within just a few minutes (less than 10), the steam is closed, and decompression is allowed its 30 seconds. The cake-discharge valve is opened, and out falls any clay expelled from the gauze. As a regular routine, this may suffice to keep the filter at its top performance for weeks. Such experience gained with a plain Dutch-weave gauze of 130 µmL retentivity may well be much improved with KPZ55, so that countercurrent steaming is much less frequent than once per cycle. Other refiners may opt to wait until the operator observes an unusually rapid rise in pressure (resistance to flow) during filtration. This may come a week or a month after rigorous cleaning, depending on the oils processed and the clays used. In this event a dilute (0.2 N) hot (95°C) caustic soda is circulated forward for 10–30 minutes. Immediately after, the filter is drained and then filled with hot water, and this circulated forward for 10 minutes; it drains from there. Next, fresh hot water is pumped backward through the filter to the drain with leaves rotating. This wash is not recirculated since it could return detached particles from one side of the gauze to the other. Leaves should never be rotated when the filter is full of water, unless this is explicitly authorized by the manufacturer, because they could be damaged by the stresses then created. After the reverse wash has ended, the leaves are rotated for about a minute to eject loosely held water and displaced clay particles. Finally, a steam or air blow is used to dry the interior. Perhaps very unsaturated oils such as pilchard or linseed have deposited a particularly persistent skin on the gauze, and in this case the strength of the caustic soda wash should be increased to 0.5 N. Important is that facilities for in situ cleaning be included at the time the filter is installed. The central hollow shaft must be exactly vertical and the plates horizontal. If a slurry discharge type of filter is to be used, the construction and operating details will be provided by the manufacturer where these differ from those just described.

Filtration and Filters

185

Self-Cleaning Enclosed Vertical-Leaf Filter Units A variety of filters of this type exists; they are very often fitted with wire gauzes. Rectangular vertical leaves of the same size may be positioned in succession along the length of a horizontal shell, or rectangular leaves of decreasing width may be positioned across a vertical cylindrical shell, the widest being across a diameter. Usually, a filtered oil manifold connects the interiors of all leaves (Veldkamp, 1991), but in one model each leaf has its own exit so clarity may be observed (Avon Ind. Filter). In vertical-leaf filters, cake discharge may be by vibration, scraping, or sluicing. In another model, the filtration manifold is central in the circular leaves (vertical), so that when discharge becomes due the leaves are slowly rotated while scraper blades detach the cake, which then falls to a sump below provided with an Archimedean screw discharge (Veldkamp, 1991). The point was made that the more the surface of a gauze is made smooth to ease discharge, the greater is the risk of the cake sliding down that surface when it is vertical. This means the very successful KPZ55 appears an unlikely choice for vertical leaves. There is also the difficulty of using vertical filter leaves to deal with RV filtration. If one kind of oil continues to be filtered, it is simple to expel the unfiltered RV to the feed vessel or small drop tank to be processed later (Veldkamp, 1991); otherwise, an attempt is made by siting two wedge wire scavenge tubes on either side of the main manifold in the floor of the shell (Anon., Ind. filter). Yet another attempt to overcome the difficulty of RV filtration appears in the form of a tilting filter. Here circular leaves in a vertical cylindrical shell are horizontal while filtering, then when the discharge of the cake is finally to be done, the whole unit is rotated through 90 degrees and becomes horizontal, so the leaves become vertical and the cake can be discharged as usual by vibration, scraping, or sluicing (Veldkamp, 1991). A filter size of 30 m2 appears to be the largest of this type.

Filter Economics The costs of installation vary from one situation to another; therefore, only general guidance is offered here. Automatic filters can provide a high degree of automation and central control. Plan work to achieve a minimum of dead time, while avoiding a restriction on the output of other units. Filter oils which are compatible in a sequence which ensures the number of occasions when a less-than-full filter has to be cleaned is at a minimum consistent with quality. In a refinery, handling six or more species of oil (e.g., soy, palm, lard) in a week, zoning, and segregation become inevitable. In making cost comparisons, take the following into account: 1. Is a new or existing building to be used? 2. Will long continuous runs on some qualities be likely? 3. Is labor at moderate cost readily available or not? 4. Large floor spaces filled with manual presses were replaced by a few centrifugal

186

H.B.W. Patterson

Self-Cleaning Enclosed Horizontal-Leaf Filters Positive

Negative

Versatile Very labor saving Very good working conditions Minimal floor space needed

Needs more headroom than press

Higher flow rate/m2 than press

Costs more per m2 than press

RV filtration proven

Press blowing short and simple

Cleaning/discharge time is low (others longer or more difficult)

Not fully continuous Has several moving parts Requires higher-quality maintenance

No air/oil contact Gauze lasts longer than cloth

Cloth cheaper than gauze

Horizontal cake is secure

filters requiring much-less labor and supervision. Determine equivalent filtration areas (i.e., 30 m2 in filter A may equal 50 m2 in filter B). 5. Specific filtration rates (including cleaning/discharge and so forth) may be 170 kg of oil/m2/hour for A and 100 kg of oil/m2/hour for B. Hence, overall rate for A is 5.1 tons/hour and for B 5.0 tons/hour. 6. Filter A may cost more to buy and install than B, but building costs with A could be significantly less. Although wire gauze is more expensive than cloth, the cloth may have to be renewed for every 40–100 tons of oil/m2 filtered, whereas a square meter of gauze may easily filter 3000 tons of oil over several years of use before renewal (perhaps 3–5 years). Filter aids may not be needed or may be equal in each case. Labor costs for A may be far less than for B and, thus, completely outweigh the slight chemical costs for occasional cleaning of A. Other very different filtration systems may have their cost heavily increased if they need to be enclosed to prevent ready access of air to the warm oil, or if the relatively large filtration areas for oil/clay slurries (40 m2 and larger) are not easy to accommodate in the floor area.

Polishing (Patterson, 1973) Normal unaided vision detects particles down to 40 µmL in size as individual specks. However, particles as small as 5–10 µmL, if present in sufficient concentration, cause dullness or a haze in what otherwise should be a brilliantly clear liquid. Although we are now concerned with fines derived from bleaching clay or carbon, remember that droplets or crystals of water, stearin, gum, or wax are much more common causes of

187

Filtration and Filters

lack of clarity. To remove these, the oil is filtered at temperatures around 5°C, 15°C, or whatever is appropriate to the specification. While oil at 95°C has as a viscosity around 10 cP, by 25°C this has risen to 75 cP; hence, the pressure differential across the membrane would increase. However, polishing to remove inorganic specks is usually performed around 90°C, and this takes away the nuclei on which crystals of gums and so forth would have begun to grow. This merely delays crystallization, so to pass the cold test for use as table oils, several oils have to be filtered at a low temperature. Some risk exists that small traces of bleaching clay may catalyze the oxidation of polyunsaturated fatty acids in oils such as soybean. Bleaching clays may contain about 4% of iron but probably combined in a form much less active than that derived from corroded equipment. As a precaution, the recommendation is that clay traces be kept below 100 ppm. Fortunately, the normal, properly conducted main filtration of the bleached oil yields no more than 25–30 ppm of clay fines. Polishing filters may be graded to remove particles of 10 µmL, or even less, but such a filter is facing a very different particle-size distribution than that confronting a main filter. Table 6.6 illustrates the difference. Since the weight of clay and not the number of particles is the factor (in the first place) which decides the size of the catalytic effect, we are on relatively safe ground when a normal filtration yields only some 25–30 ppm. Even if the entire 87.52% of 2–5 µmL particles were to pass the polishing filter, they account for less than half, by weight, of the fines present.

Polishing Filters Essential characteristics of a polishing filter are its abilities to retain fine particles and to permit an acceptable oil-flow rate per square meter. Weaves with many rough and hairy filaments are advantageous, so paper and wool have a long history of use with edible oils; finely-woven synthetic textiles are now used for the same purpose. Extremely fine-wire weaves, sintered powdered metal, and ceramics were used for polishing filtrations, but not with edible oils as far as is known. One can make a distinction between those membranes which polish by retaining on their surface all but the very fine particles (below the micron-size grading of the filter) and those which depend on the fines being trapped in the labyrinth of channels within the membrane. Both types blind eventually; this may take several days or even weeks because the loading is small. Some, such as the synthetic textiles, may be repeatedly cleaned and returned to use; others, such as paper, will be discarded and replaced. Paper is used as an overlay on a coarse textile in a small press or as a heavy paper overlay on a TABLE 6.6 Particle-Size Distribution by Numbers (After Filtering, Before Polishing). Micron size (µmL) % (in numbers)

2–5

5–10

10–25

25–50

above 50

87.52

11.11

0.82

0.22

0.09

188

H.B.W. Patterson

horizontal metal-gauze filter. With edible oils, polishing paper grades with 10 µmL or 5 µmL of retentivities are usual, but in other fields such as pharmacy, even finer retentivities are in use. Granted that the membrane on offer retains particles down to the specified size and permits acceptable flow rates, the user must choose according to cost, life, and ease of manipulation. Continuous self-cleaning filters are available which consist of a stack of small circular rings narrowly separated from one another by spaces. These may not profess to retain less than 40 µmL of particles (the width of the spacer) and, therefore, they are not strictly polishing filters. The spaces between the discs retain the solid while liquid flows through. The rotation of the stack of discs against a stationary set of fine blades scrapes out the collected dirt, and allows it to fall into a small sump. However, if these limitations in retentivity are acceptable, such filters provide a cheap, simple, and compact method of controlling the appearance of oils and were so used. A widely used polishing filter, the cartridge filter, contains several rather small cylinders a few inches long, through whose porous walls the oil flows. The cylinders are grouped in layers, one layer above another, with manifolds providing a common inlet and outlet for the whole unit. The porous walls may be made from wool, cellulose fibers, or two types of fiber wound on the same former. Some membranes may be cleaned and reused. A choice of lower retention limit is available—between 75 and 5 µmL. These filters are not expensive, are not bulky, and are easily serviced. The so-called candle filter is very similar in operation to the cartridge filter, but relies probably on only one porous cylinder that is a few feet long. This will effectively remove particles down to 10 µmL and even lower at very acceptable flow rates. Although a common practice is to polish oils on their way to deodorization, some manufacturers include a further polishing step immediately before the tanker, as the product is being loaded out to customers.

Chapter 7

Oil Recovery H.B.W. Patterson

The Changing Situation At the end of oil filtration, some oil is retained in the filter cake and some more loosely held in the membrane and associated pipework. Recovering this oil is valuable— leaving behind virtually all of the adsorbed impurities and using as much minimal expenditure and filter downtime as may be feasible. Another consideration has gained increasing importance since the 1970s: the greater cost of disposing of unwelcome liquid and solid effluent. Bleaching clays in technically advanced countries have improved in quality, and new kinds of filters have come into more widespread use; fats-and-oils processing units have concentrated more intensively within their regions; and several oil-recovery methods were known and used in different places over many years. While much of what follows is well-known to experienced operators, a general review in some detail is given in the belief that it may lead to the checking of earlier assumptions or to the prompting of the exploration of possible improvements. Because of differing circumstances in one factory or another, the most economical solution will not be the same in every case. Filter Cake Before describing the different methods of oil recovery and some of the important factors affecting them, most useful is to consider the makeup of the filter cake itself. A first step in any recovery process is to expel the loose oil from the filter and to some extent from the cake by blowing, preferably with an inert gas, or if the oil is not particularly unsaturated, with a limited amount of air. Observation will show for how long appreciable streams of oil are being expelled, but 10 minutes is probably the maximum and a shorter time may be quite adequate. Unsaturated oils such as linseed and marine oils are very vulnerable to oxidation, and therefore, in such cases, air blowing is best avoided as it is almost certain to prevent the recovered oil from reaching generally accepted standards for processed edible oil. For technical outlets, the standard would not be as high. Safety is a further consideration. In old plants with filter exits open to the air, lachrymatory fumes from oxidizing marine-oil cake give the first warning of excessive air blowing; in any plant a scorched cake clearly shows air blowing was heavily excessive. For nitrogen blowing, a fixed maximal volume may be designated as approximately 10–15 m3, the reasoning being that more would not be cost-effective. Excessive blowing also creates cracks in a cake, and encourages irregular extraction if this is done in situ. At this point, the blown cake (supposedly mostly resulting from the use of activated clay) is likely to contain about 70 parts oil 189

190

H.B.W. Patterson

per 100 parts “dry” original clay (i.e., approaching 41% since other organic material is present). Two points arise here. First, dry original clay as purchased is likely to contain in a temperate climate 4–8% of free H2O and more in a humid tropical climate. This level of moisture is estimated by drying the clay at 105°C. If such clay were then calcined at 850–900°C, the loss of firmly held moisture would then amount to a further 8–10%. When reporting results on a dry original clay basis, use the first result obtained at 105°C. Some manufacturers state how moist their activated clay should be to achieve the best results, and for certain oils, some specify adding a small amount of water at the time of bleaching. Secondly, as stated earlier, activated carbon, activated clay, and natural nonactivated clay will retain by adsorption about 150%, 70%, and 30%, respectively, of their own dry weight as oil. Besides pigments and oil, a whole range of other compounds was adsorbed, depending on what was present in the oil. Broadly speaking, the adsorbate on the cake is likely to be comprised of 85% fat (hexane-soluble), and the balance is oxyacids, polymers, soap, and other insoluble organic substances. If unextracted, cake is left exposed to the air, oxidation will commence within a day, especially if the oil is unsaturated, and it is thinly spread over a large solid surface as when less than 30% is in the cake. For appreciably heavier oil contents, the onset of oxidation is delayed, but eventually takes over. Warmth and air movement hasten the reaction. Cake stored under nitrogen or in an inert organic solvent remains unaffected for months. The effects on recovered oil of storing the cake in air are: (i) recovered oil yield becomes less; (ii) color of oil is worse; (iii) free fatty acid (FFA), oxidized FFA, saponification value, and E232 and E268 increase; and (iv) iodine value falls.

Oil Recovery by Solvent The attractions of this method of oil recovery were recognized for decades, and units such as the Kelly and Sweetland filter presses have catered to it (Andersen, 1962). The enclosed horizontal-leaf filters now provide the same faculty. The advantages are a high yield of extracted oil (over 90%) with a corresponding very low residual fat content of about 3% in the remaining cake. Other methods employ hot water and/or steam, and might show a (petroleum ether-extracted) fat residue on the dried extracted cake of approximately 18%. If the other organic matter (soap, gum, pigment, and so forth) also amounts to 18%, then this corresponds with 18 parts of oil per 64 parts of original clay or 28 parts of oil per 100 parts of original clay (i.e., a drop from 70 to 28 or a recovery of 60%). With a cake predominantly from activated clay, a final result of 20 parts of oil per 100 parts of original clay is the best that can be expected. Also important is the fact that the popular solvent hexane is particularly selective in removing triglycerides and fatty acids but not pigments, oxidized fats and so forth, so that the quality of this recovered oil easily matches that of the filtered oil itself. The disadvantage is the cost. Although hexane is the cheapest solvent, it brings with it those precautions which must be observed in all plants when handling a flammable solvent. The description

Oil Recovery

191

of the use of hexane as presenting a tremendously hazardous situation (Watson & Meierhoefer, 1976) must be read alongside the fact that it has found use in some European countries where plant discipline and safety standards are high. Were this otherwise, another method would be preferable. Apart from hexane, trichloroethylene can be used, and it has the advantage that it is nonflammable. Its disadvantages are that it is narcotic, toxic, and because of its much lower selectivity toward triglycerides, darker, poorer-quality recovered oils are obtained. Petroleum ether (b.p. 65–75°C) gives a high yield, and is very elective but flammable; the 75–90°C b.p. grade was used to reduce explosion hazard and solvent loss. Other solvents for fat include carbon tetrachloride, methylene chloride, and ethylene chloride. The lower-boiling aliphatic hydrocarbons give the best-quality recovered oil; chlorinated hydrocarbons give high yields but the darkest oil. Three possible procedures present themselves. The cake may be extracted with solvent in the filter and then the residue desolventized by steaming. This prolongs downtime and, therefore, involves additional filter capacity for the plant. A possible variant would be to complete the solvent extraction in the filter and then, if the filter is of the solid-discharge type, to discharge the cake for desolventizing in a separate unit nearby. The oily cake may be discharged directly to an extractor, and the filter quickly allowed back on duty. However, when the extractor has finished its work, a filter must be provided to separate the miscella from the residue. A variant of this procedure is to drop the oily cake into a container, and move this to a solvent plant. The container can be purged free from air if desirable. Finally, the oily cake, suitably protected, may be sent to another company in the area which is in business for this kind of task. The first option will be of greatest interest if a new plant is being built. Protecting an existing area from explosion or toxic hazard can be expensive even if enclosed filters are already in use. If a solvent extraction unit is to be built to serve several neighboring factories, an assured minimal supply of oily cake is essential to make the project viable. A figure for cake corresponding to six tons of original clay per 24 hours was mentioned (Taylor & Ungermann, 1987). This estimate depends on, among other things, the replacement cost of the oil, the transportation cost to and from, and the cost of residue disposal in one form or another, assuming a choice is available. Whereas fresh oily cake can be extracted in a few minutes, cakes which were stored for a long period may require an hour. Especially with unsaturated oils, if some contact with air was experienced, the color of the oil recovered will suffer. As bleaching clays become more efficient, smaller doses are used. This situation means it becomes more difficult to provide enough cake to justify building a communal solvent plant. Table 7.1 illustrates the possibility of using reduced doses of more costly earth (Taylor & Ungermann, 1987). Presuming it is more active, a direct cost-savings may be possible. In any case, no cost of spent-earth disposal is included, so a reduction of this is another savings. The third option of using another company must be examined clearly if one exists. Transportation costs and how promptly the extraction is performed are important.

192

H.B.W. Patterson

TABLE 7.1. Approximate Cost (U.S.$) per Short Ton of Oil Bleacheda Percent of bleaching earth required Delivered cost per ton (U.S.$) of bleaching earth

0.25

0.50

1.00

1.50

2.00

400

1.375

2.750

5.500

8.250

11.000

500

1.625

3.250

6.500

9.750

13.000

600

1.875

3.750

7.500

11.250

15.000

700

2.125

4.250

8.500

12.750

17.000

800

2.375

4.750

9.500

14.250

19.000

900

2.625

5.250

10.500

15.750

21.000

1000

2.875

5.750

11.500

17.250

23.000

a

Assumptions: (i) The value of refined and bleached oil is $500 (U.S.) per short ton. (ii) Percentage of oil retention is 30% by weight of spent earth. (iii) Does not include cost of spent-earth disposal.

The recovered oil quality can be good if the work is handled promptly. Because the yield is high and the oil quality is very good, solvent recovery of the higher-priced oils may easily pay the direct operating costs when performed on a reasonable scale. This question arises: Can the capital charges for the installation also be met? Penalties now attached to discarding a high-fat residue support the solvent-plant case or the use of another company.

Oil Recovery by Hot Water in situ One can apply in situ oil recovery by hot water to the conventional cloth-clad filter press or to automatic filters using metal gauzes. Important is that the cake should be stable, and therefore, a variation in thickness is minimal. Horizontal metal gauzes have an advantage here over vertical ones (Veldkamp, 1991). If a filter press is in use, fill it to 85–95% of its capacity; 87.5% is ideal. Metal gauze filters also require a minimal specified free space above the cake. At the end of filtration, the filter is lightly blown to displace loosely held oil; nitrogen is preferable for this purpose when the oil is very unsaturated or scheduled for edible use. Excessive blowing causes cracks in the cake and inefficient extraction. Hot water at 95°C is then pumped briskly through the cake. A pressure that is a little above that of the final filtration pressure is suitable; a high flow rate reduces the influence of variation in cake thickness on the degree of extraction. The weight of hot water passing through the cake will vary between 5 and 20 times the weight of the cake, but this water will be the same water circulated several times; normally, a ratio of earth/water of 1/10 is enough. This is judged by observation and experience. For example, a flow rate of 0.5–1.0 tons/m 2/hour is equivalent to 166–333 kg/m2 in 20 minutes, and this passes through approximately 14 kg of cake on a square meter.

Oil Recovery

193

The wash water and the oil it carries with it flow to a nearby tank (preferably insulated) where the oil separates on the top. This tank is kept warm by a closed steam coil. Hot water may be drawn from the base of this tank and recirculated. A superior arrangement is to allow the water from the base of the tank to level up into the top of an adjacent tank, where its temperature is fully restored to 95°C as it continues to circulate. The duration of the extraction depends on filter, clay, oil, and packing. Activated clays give up their oil much less readily than natural ones; regard 40 minutes as a maximum—shorter times are likely. A repeated sampling of the filter effluent shows that in a situation where 20 minutes achieves all the extraction that is going to occur, over 80% of the recovered oil is obtained within the first 5 minutes, and extraction then commences to fall quickly. Further obvious is that much of the better-colored oil comes out in this first flow, and as washing continues, it darkens appreciably. This affords the possibility of diverting the better oil to its own tank so that one may use it ultimately for an edible or superior technical product. The same stock of water may be used repeatedly until it is too dirty; this economy minimizes liquid-waste disposal. When extraction has finished, the cake is dried by blowing a current of air through it for up to 30 minutes. In the case of marine or linseed oil, this time may have to be shortened because the oxidation of the unsaturated oil can cause overheating. Steam sometimes is featured in the drying routine. One can achieve a final content of 25% of H2O. Finally, the filter is emptied and made ready for reuse. In connection with this process the following points are worth noting: 1. If an opportunity arises to inspect the extracted cake—as in the opening of a filter press—well-extracted areas appear light brown or slightly green, depending on the kind of oil, whereas poorly extracted areas appear dark brown. 2. Acidity extracted from activated clays may shorten the lives of cotton cloths a little. One may counteract this by making the wash water slightly alkaline with sodium carbonate. If the water is made more strongly alkaline, although the oil yield increases, the quality is worse. Terylene resists both acid and mechanical wear better than cotton. 3. On resuming filtration after a hot-water extraction of a filter press, the first ton (approximately) of bleached oil may be discolored by impurities leached from the damp cloths, and this should be recycled to the bleacher until the color is satisfactory. This recycling is in any case a normal feature of filtration in order to lay down an adequate precoat layer of clay. 4. Leaving too much loose oil in the filter and pipework may give an exaggerated impression of the effectiveness of the hot-water wash.

194

H.B.W. Patterson

Oil Recovery by Hot Water Advantages

Disadvantages

Equipment and operating costs (including labor) are low.

Filter downtime increased (as for all in-situ recovery).

Yield of 60–70%.

Slight increase in cloth usage.

Satisfactory quality of early fraction.

Yield (65%) less than solvent extraction (95%). Solvent-oil quality is better.

Minimal contact time of earth/oil before extraction commences.

Oil left in clay increases the disposal problem.

The economics of this recovery depend on the value allotted to the recovered oil. As a generalization, the first 85% of an extraction would at least have the value of the parent crude oil, and probably nearer to the bleached oil, since processing lowered its FFA, degree of oxidation, gum content, and color. The balance of 15%, if kept separate, will best be passed to a nonedible outlet. As it is first obtained, the 85% yield is warm and moist. Both conditions favor lipolysis and oxidation; therefore, the sooner it is passed to its intended destination, the better. Although it is well-known that a variety of organic solvents extracts slightly different percentages of total oil from a filter cake, as far as the hot-water extraction is concerned, no selective action on one class of triglycerides as against another is apparent; the oil extracted is substantially the same, regarding fatty acid makeup, as the oil remaining in the cake. Some further comments on the testing of extracted cake are included in Chapter 9.

Oil Recovery by Separate Aqueous Solution This is probably the oldest method of oil recovery, predating pumping hot water through the filter press where the cake was reasonably well held as a kind of sandwich in the press chamber. With a vertical leaf (gauze) pressure filter, the opportunity exists for the wetted cake to slide from the gauze surface. Many old variants of the freshly discharged press cake are processed with an alkaline solution or wetting agent and even employing an autoclave (Andersen, 1962). A simple basic method is to transfer the fresh oily cake to roughly four times its own weight of hot water (90°C) in a conical-bottomed vessel fitted with a steam jacket and paddle agitator. The mixture is stirred steadily to keep the cake in suspension for at least 6 hours, and the temperature is kept just above 90°C. A sample in a beaker toward the end of this period will begin to show the separation of an oily layer and a deposit of clay slurry below. A satisfactory extraction is indicated by seeing no more than a few dark specks of oily cake in the lower slurry layer. Hot water is sprayed onto the oil, and stirring is stopped. Settling is allowed for 12 hours; the temperature is kept above 90°C. One may improve extraction and separation if enough sodium carbonate or caustic soda is added during the stirring to keep the pH just above 7. The oil is finally decanted

Oil Recovery

195

and cleaned according to requirements. The yield of oil is around 70%, but the color is not as good as obtained by extracting the cake in situ. The stirring must not be so rapid as to create an emulsion. Process times and amount of alkali added are adjusted according to experience. The disadvantages of this method are the disposal of the bulky slurry of partly deoiled clay and the inferior color of the recovered oil. On the other hand, filter downtime is not affected, and equipment and operation costs are moderate.

Chapter 8

Safety, Security, and the Prevention of Error H.B.W. Patterson

Bleaching-Plant Safeguards In an oil-bleaching plant, as in any other, in assessing risk, the general principle applies of asking: “What is the chance the fault will occur?” and “How serious will the consequence be if it does?” Probably neither question can be quantified mathematically, but the product of the two answers gives a worthwhile guidance as to how far we must go in preventing the fault. Primarily, the errors particularly associated with a bleaching plant are considered here. Two other oil-processing textbooks (Patterson, 1983, 1989) included safety chapters dealing with a wider range of operations in oil processing, yet the risks of spillage, contamination, fire, and explosion can reach out to threaten operations in the bleaching zone in particular circumstances. Safety begins with the considered attempt to “design out of trouble.” If the design is such that the user must seek help from a mechanic to modify the equipment before the risk situation can be introduced, then the design can be regarded as safe so far as that risk is concerned. Fire is the number one risk as shown by experience with bleaching plants. If the discharged oily cake or oily cloth is allowed to remain some hours, especially in a draft, the odds are very much in favor of it beginning to smolder and then burn, more especially if the oil is a highly unsaturated marine oil. Conveyor belts may be out of easy view for much of their travel. The failure to clean them adequately of the filter cake sets the stage for a plant fire during the 48 hours the plant may be closed for a weekend or holiday. Even when soft marine-oil filter cake was subjected to in-situ hot-water oil recovery and then some blowing to dry it, a lachrymatory fume is detectable after some hours’ exposure to air. Dumps of spent clay, if they are tolerated, are best located where any fire cannot spread from them to the refinery building, and easy access for firefighters must be possible. A similar risk exists if oil leaks onto insulating material which is porous by nature, and thus, the oil is drawn along capillary fibers as in a wick. Noted is that cellular nonporous, noncombustible material, such as glass foam, reduces such risk very substantially. A great deal of difference exists between exposing a portion of oils of different degrees of unsaturation in dishes to warm air and exposing the same oils spread thinly on a porous material. In the latter case, as the iodine value increases from 100 to 200, a corresponding increased readiness to smolder and then burn occurs. Such conditions correspond more closely to those in a plant. Porous insulation treated with a proprietary sealant or boric acid offers a much-reduced risk. The filtered-oil store in a bleaching plant, in many cases, serves a hardening unit close by; arrangements must be made to prevent hydrogen under pressure from 197

198

H.B.W. Patterson

leaking back along feed lines or being drawn back by a preferential draft (chimney effect) from a leak in the hardening unit. The same condition holds good for any neighboring unit employing flammable solvent; then the risk is more insidious since the vapors will be heavier than air and very much less inclined to disperse than hydrogen. Press cake must be thoroughly cleaned from plate-and-frame presses so that when they are put back into use, hot oil (90°C at 4–6 atm) does not spurt out from an imperfect closure. As in other parts of the refinery, where corrosive fluids (caustic alkali or mineral acid) are handled, the fabrication of the pipework and the protection of flanges must guard against sudden leaks under pressure. Glass content gauges fitted to tanks should be guarded. Before personnel enter a bleacher or tank even briefly, the temperature must be brought below 40°C; oil and chemical feed lines and steam and water connections must be blanked rather than relying on closed valves. In addition, removing the fuses from circuits supplying power to agitators is safer. In instances in which a large bleaching or other plant is handling, in the same week, a number of incompatible oils on isolated sequences of production units, the prudent action is to arrange that an oil cannot cross from one sequence to another without some slight engineering modification having been made. Lastly, well-known is that human errors are most common at or immediately after a change of shift. A system can therefore be established in which supervision changes an hour after or before the change of operators. This means that operators, on taking over a process, are under the supervision of someone who was directing the work for several hours already or, alternatively, of someone who has had ample opportunity to learn the current situation before exercising control himself.

Chapter 9

Important Tests Relating to Bleaching H.B.W. Patterson

Purpose and Validity of Tests Some tests relate to the intrinsic nature of the adsorbent and others to the extent to which it exerts its influence on other materials. For example, particle-size distribution and degree of acidity indicate important features of the adsorbent itself; how effectively it removes carotene and chlorophyll from an oil shows how useful it can be. The latter tests involve a convention defining the conditions in which the pigment removal is to be performed. The test conditions must be the same as those for the full-scale process, or at least capable of being related to them to a reliable degree. Finally, in comparing one adsorbent with another, both types of test may be involved, depending on the basis of comparison. Again, as was already emphasized (see Fig. 2.1), if we are comparing the ability of two clays to bleach an oil, we must decide the range of color removal about which we are particularly interested. Different starting and finishing points can affect, even reverse, the result (Andersen, 1962; Taylor & Ungermann, 1984). Again, the comparison must be made on the basis of the relative weights required of the adsorbents to obtain the same given effect, and not the relative effects of the same weights of adsorbents. The same is true in comparing the activities of a catalyst and, no doubt, in many other fields of comparative testing. The underlying reason for this is that as a test progresses from its starting point, conditions (concentration of pigment or other adsorbate or reactant) are steadily changing, so we make the required overall change the target. Also true is that an adsorbent or catalyst is itself affected by progressive interaction with the oil (adsorption, hydrogenation, interesterification), and therefore, its performance is judged by the minimal dose needed to achieve the target. This is the appropriate point at which to remind readers that a particular species of oil may vary from parcel to parcel. If an unknown clay is being evaluated, it must be done in comparison with the performance of a standard adsorbent, both acting on exactly the same oil as test medium—again, very similar to evaluating an unknown nickel catalyst for hydrogenation purposes. In this instance, usually, three tests are done spanning the range from just below to just above the anticipated clay dosage required. A simple plot of dosage (x-axis) versus oil-quality parameter [Lovibond R.Y., chlorophyll (ppb), phosphorus, etc.] (y-axis) for the unknown and the standard enables a reliable comparison to be made (Taylor & Ungermann, 1987). This chapter is written not to give all the details of the analytical methods, but to give an insight into the usefulness of several tests and possibly their limitations. Often the case is that individual manufacturers have established at least some of their own test methods, and are quite ready to explain these to prospective clients. Many test methods are laid down by IUPAC or various national standards authorities. If 199

200

H.B.W. Patterson

some specific recommended or authorized test was used, this should always be made clear in reporting the result; if some other method was used, this itself should be described. This simple precaution is the means of avoiding much confusion.

Evaluation and Comparison of Bleaching Clays and Other Adsorbents The crucial importance of adopting a valid basis of comparison between adsorbents and applying test methods which allow results to be related to full-scale work has been recognized for decades, and Andersen (1962) is one of those who considered this in detail. After widespread consultation, the American Oil Chemists’ Society (AOCS) obtained a consensus between the manufacturers and the users of adsorbents (clays, carbon, and synthetic silica) that enabled it to produce an agreed-upon official method in 1991 which embodies what was found to be good practice (Cc8f-91). This method is published along with a list of AOCS evaluation methods which relate to the various quality parameters by which the effectiveness of the adsorbents are most commonly judged (e.g., oil color, chlorophyll content, degree of oxidation, and so forth). Although national standards may differ in evaluating these individual quality parameters, the method of making the comparison between adsorbents seems wellfitted to have universal application. The basic procedure is as follows. Procedure Apparatus The apparatus is a glass reactor (250–1000 mL) equipped with a variable-speed stirrer, a thermometer and automatic temperature control (±2°C preferred), a vacuum facility which can maintain a reduced pressure between 20 mm and 100 mm of Hg absolute, a nitrogen (or other inert gas) supply, and a jacketed porcelain or stainlesssteel Buchner funnel capable of filtering the test batch in one filling and fitted with paper to retain particles down to 2.5–5.0 microns. Test Conditions The test vessel is not more than half-full; the agitated oil is dried/deaerated at 60 ± 5°C for 15 minutes (unless adsorbent suppliers indicate otherwise) under vacuum; vacuum is broken with nitrogen; and the set weight of adsorbent (treatment material) is added. With vacuum and agitation restored, the temperature is quickly (under 5 minutes) increased to the predetermined level (usually 95–120°C), and maintained there for 30 minutes. Next, the test vessel and contents are cooled rapidly to 70°C, and the vacuum is broken with nitrogen. The oil is filtered through the warm Buchner funnel, and the sample filtrate is sealed under nitrogen until required. For those who wish, a shorter contact time may be used; bleaching in air is also possible. Evaluation Procedure Before commencing the program, an estimate is made of the amount of oil later required to do all the particular quality tests needed (color, peroxide value, E232/268,

Important Tests Relating to Bleaching

201

etc.). The recommendation is that three times this amount should be bleached. As mentioned previously (see section Purpose and Validity of Tests), it is advisable to do three bleachings in a test—one below, one at and one above the anticipated optimal dose—and thus obtain a contour or isotherm. At least one test should be done in duplicate to establish reproducibility. The system described by Taylor and Ungermann (1987) is comprised of two vessels side by side so that all tests can easily be run in duplicate if desired, or one vessel may contain one level of dose and the other vessel a different one. Possibly, the level of dose is not the process parameter to be investigated, but some other obvious one, such as contact temperature or duration. The basic procedure is very similar. The initial values of the quality parameters are obtained for the unbleached oil. The bleaching tests are then done on the preselected conditions, a contour being obtained for each condition. Each bleached–oil sample is then analyzed on each of the quality parameters (color, peroxide value, etc.). The initial values and the postbleached values for each treatment (process parameter) sample are next graphed against the treatments concerned. The resulting contours then indicate which treatment is best or which should be changed to seek further technical or economic improvement or both. The AOCS Uniform Methods Committee end their study by emphasizing that regardless of the laboratory results obtained, the final decision rests on plant experience derived from the test procedure. Also mentioned (Taylor & Ungermann, 1987) was that where stability is an issue, the necessity will probably be to follow bleaching with deodorization and then perform an AOM or Rancimat test. As distinct from discovering how best to employ an unknown adsorbent or how best to treat a parcel of oil of uncertain quality with existing adsorbents, a continuing need exists to check successive production batches or deliveries of adsorbent of ostensibly the same activity. In this fairly simple case, one contour derived from three dose levels can be compared with the current standard. This quickly shows what dose level is required to achieve the same effect as, for example, a 1% dose of the standard. If less, the activity is higher, and vice versa.

Important Quality Characteristics Very complete accounts are given in many national and international standards and textbooks of the methods used to determine the important quality characteristics of fats and oils. Here, appropriately, an enumeration is made of only those associated with bleaching evaluation and the character of adsorbents. Color To distinguish between clarity and color is necessary (Patterson, 1973, 1976). More particularly in the case of a crude oil, moisture and dirt need to be removed before measuring color. Since heating may induce color changes, moisture is removed by drying the oil over anhydrous sodium sulfate and then filtering it through a paper retaining particles down to 2.5–5.0 microns; Whatman No. 15 or its equivalent is regarded as suitable. Solutions of potassium dichromate or iodine in potassium

202

H.B.W. Patterson

iodide solution (Cocks & van Rede, 1966; Mehlenbacher, 1960) found wide favor in Europe, especially for works’ practice as visual standards of comparison, and had the advantage of being expressed as a single number. The present most popular system of color grading for a great many fats and oils depends on matching clean-oil samples held in 1” or 5¼” cells (0.5”, and 2” are also sometimes quoted) with a combination of red (R) and yellow (Y) tinted-glass slides held in a color comparator (Lovibond Tintometer) and all viewed simultaneously. For some users, a provision is made for including blue (B) tinted slides if this improves the match; (Ca 13b-45) Corr 77 and BS 684; Sec. 1.14 (1987) refer. Numerous precautions were proposed from time to time with the intention of maintaining accuracy and agreement between different users (Cocks & van Rede, 1966; Hamilton & Rossell, 1986; Macmillan, 1949; Sonntag 1982). Slides of a particular tint are additive (2 × 5 Y = 10 Y); colors at or below 3 R are more easily differentiated. Staring too long into the tintometer (much above 10 seconds) fatigues the vision, and if a match cannot be obtained, the test should be repeated after a few minutes. Obtaining a match when the ratio Y/R = 10/1 is recommended is feasible; obtaining the findings of more than one reader helps in some way to reduce individual error. Substantial amounts of green or brown pigments in the oil make readings less reliable. The description of early attempts by the U.S. Department of Agriculture (USDA) to find a satisfactory system makes worthwhile reading (Macmillan, 1949; O’Connor et al., 1949). Naturally, attempts would be made to avoid a dependence on human judgment, and this has led to the development of the Automatic Lovibond Tintometer AF960, in which three beams of colored light are passed through the oil and inspected by a photoelectric cell. One is a reference beam to compensate for variation in light intensity, the presence of dirt and so forth, while the other two evaluate red and yellow. Facilities are provided for reading levels of carotene and chlorophylls a and b (see also Pfannkoch & Gill, 1990). Hamilton and Rossell (1986) point out that in those instruments in which the reference beam operates at a wavelength which can be affected by blue or green colors in the oil, any over- or undercompensation for blue or green in the reference beam could affect the accuracy of the red/yellow readings. They add that automatic colorimeters have proved useful for factory use on partially or fully processed oils where colors are paler and seldom possess green or blue tinges, but for crude oils having a greater variety of hues, they may be less suitable. The Color in Oils Committee of AOCS (Berner, 1991) noted that an established trend for the Lovibond Automatic Colorimeter AF960 was to give higher readings than the manual method, more especially with darker oil, and that this should be rectified before use in AOCS approved methodology. The McCloskey Scientific automatic colorimeter also has facilities for assessing carotene and chlorophylls a and b (Sleeter, 1985). For fats having markedly light or dark green, yellow and dark-red tones, another system which depends on matching with colored discs (there being 26 subdivisions) has gained some popularity among those handling such fats. This is the FAC system (AOCS Cc 13a-43) described by Sonntag (1982), who also gives details of the Gardner scale used for certain technical products (AOCS Td la-64).

Important Tests Relating to Bleaching

203

The spectrophotometric system is another attempt to avoid dependence on the human eye. For cottonseed, peanut and soybean oils, absorbances are measured at 460, 550, 620, and 670 nm at 25–30°C, and the values are converted by a simple formula to an index, AOCS Cc-50 (Corr. 77). This system appears capable of being extended to other vegetable oils (Sonntag, 1982). Extinction Values When light of a certain wavelength passes through a medium, its intensity when entering (Io) and leaving (I) can be monitored by a photoelectric cell. Absorbance (optical density, extinction), E, is expressed as log (Io/I). Within limits, the absorbance in passing through a solution varies both with the molecular concentration of the solute and the path length through the solution, assuming no chemical reaction occurs between the solute and the solvent. Truly, different solvents can give rise to somewhat different absorbances with the same solute. Since we often may not know the molecular weight of the solute, as a matter of convenience, a 1% w/w solution is used, and the path length is made of that of a 1-cm cell. Hence, the expression, “E-one, one,” or as we commonly see it in relation to some particular wavelength (e.g., 262 nm, 268 nm, 450 nm), E262 and so forth. As the wavelength changes, so does the absorbance. Peak values are characteristic of a particular substance; these may weaken or drift to another value if an oil is damaged by oxidation. Examples of the use made of extinction values are provided in this volume (see section Oxidized Fats and Oxidized Fatty Acids in Chapter 1, and sections Groundnut (Arachis, Peanut) Oil and Palm Oil in Chapter 4), and the effect of processing is also discussed by O’Connor and others (1949). [AOCS Cd7-58, Sec. G (1981) and BS. 684 Sec. 1.15 (1978) refer.] Peroxide Value (PV) Fat hydroperoxides and peroxides are flavorless, but later degenerate to other compounds which are the most potent causes of rancidity—the fatty aldehydes and ketones; many of these have a much lower flavor threshold value than fatty acids. This initial stage of oxidation is called primary oxidation. The PV is most often expressed as milliequivalents of oxygen, present in such a form in one kilogram of fat that they are able to liberate iodine from potassium iodine solution in the set conditions of the test. The iodine is then titrated. Numerically, the value is a kind of chemical snapshot of how much of the fat is in this early state of this primary oxidation; other parts of the fat may already have reached a further stage. Some workers expressed the value in millimoles of oxygen (Lea, Wheeler); the result, therefore, was halved numerically. The test is empirical, but with experience it can be related to the future flavor stability of the particular species of oil on which it was performed (Allen & Hamilton, 1983; Cocks & van Rede, 1966; Patterson, 1989; Sonntag, 1979). Possible sources of error were reviewed (Allen & Hamilton, 1983; Gray, 1978), and include the adsorption of iodine at double bonds in fatty acids, the liberation of iodine by oxygen in the solution being titrated, the weighing of samples, a variation in solvent, time, and temperature of reaction and in the behavior of the peroxides themselves.

204

H.B.W. Patterson

Several modifications of the method were explored, and a comparison of these was given by Barthel and Grosch (1974). Usually found is that a freshly deodorized oil has a PV of nil and certainly less than 1.0; before an off-flavor becomes obvious, the PV may reach 10; by this time oxidation beyond peroxides (secondary oxidation) may well exist. Anisidine Value (AnV) [306 B.S. 684. 2.24 (1984)] This is also a chemical snapshot employed to trace the later oxidative degeneration of fat (secondary oxidation), and depends on the condensation of anisidine (p-methoxy aniline, CH3O-C6H4-HN2) with aldehydes present. High temperatures and high acidity are avoided. The condensed products show strong absorbance in the 350-nm band, the exact pattern depending on the variety of aldehydes present (Allen & Hamilton, 1983; Gray, 1978; Patterson, 1989). As accepted since Holm and Ekbom-Olsson (1972) introduced this test, interpretation is empirical, and numerical values must be related to established experience with the flavor stability of the same class of oil. Totox Value (Allen & Hamilton, 1983; Krishnamurthy, 1982; List et al., 1974) The combination of primary and secondary oxidation was once called the oxidation value (OV), and gave a useful, if not complete, picture of the present state and past history of an oil in oxidation terms. Since a peroxide is a potential source of two aldehydes, the convention OV = 2 PV + AnV was adopted. Unfortunately, this became known as the Totox value, but since neither PV nor AnV represents all the possible oxidative damage at their respective stages, this term is misleading. A Totox under 10 suggests a good crude oil, but a very unsaturated crude oil may have a Totox well above this, as in the case of a fish oil, yet after neutralization, bleaching, hydrogenation and deodorization, it may be satisfactory. The Totox is often met with in the case of crude palm oils (See section Dosage in Chapter 2). Finished-Oil Stability As is pointed out (Taylor & Ungermann, 1987), in the evaluation of a bleaching clay’s performance, as well as PV and AnV tests on the bleached oil, a fat-stability test may be an issue; this is usually done on the bleached and deodorized oil. The fat stability determined by the Swift or Active Oxygen Method (AOM) measures the time taken in defined conditions by a fat sample to reach a predetermined PV—or rancid state [AOCS Cd 12-57 (1981) and B.S. 684.2.25 (1976)]. Several authors caution that this is an empirical test, and is best used by groups who already have established relationships between AOM results and oil quality. Sleeter (1985) also describes some of the limitations of the test. The automated Swift Test is described and discussed by Rossell (1983) along with precautions to ensure reproducibility. It is marketed as the Rancimat apparatus, and has come into wide use (McGinley, 1991).

Important Tests Relating to Bleaching

205

Phosphorus The first step consists of destroying the organic matter of the sample by ashing it and then dissolving the ash in hydrochloric acid; after filtration (if necessary), the solution is neutralized, and the ammonium molybdate solution is added to form a blue solution of ammonium phosphomolybdate (Cocks & van Rede, 1966) (AOCS Ca 12-55, Reapp. 73). The phosphorus is estimated by reading the absorption of the solution with a spectrophotometer. Estimation was done visually by comparison with standards (Cocks & van Rede, 1966). This routine was automated, and an instrument was produced capable of testing 4320 samples in a day (Sleeter, 1985). Trace Metals Numerous methods exist for the determination of trace metals including iron and copper, which have special relevance as prooxidant catalysts. Dry ashing of the fat is tedious and risks loss; wet ashing (sulfuric and perchloric acids) requires careful and skilled supervision to avoid explosion. The transfer of trace metals to an aqueous medium with the aid of acids and special solvents, although difficult, is preferred by Cocks and van Rede (1966) for routine work. The solution of the metals is finally used for colorimetric analysis; BS 684. 2.17 (1988) applies to iron, and BS 684.2.16 (1988) to copper. AOCS Ca 18-79 applies to iron. Sonntag (1982) and Sleeter (1985) describe briefly (with extensive references) the many developments in the field of instrumental analysis for trace metals, some of which would be expensive for use in routine work.

Fat Content of Filter Cake Several good reasons are noted for checking the amount of fat remaining in a filter cake, presuming it has at least been blown with an inert gas or air to expel loosely held oil in it and its surroundings. If no oil recovery is being attempted, we learn (approximately) what we are throwing away and, possibly, what we may be charged for so doing. If recovery is being performed, we are able to monitor with acceptable accuracy how well the recovery process is working as against what it could do. To calculate the results to a common basis and to state the method of analysis of the cake are essential actions. If a filter press is being sampled as it is being opened and cleaned, a strip of cake from the bottom to the top of each fifth plate approaches a representative sample. An equivalent procedure must be arranged for filters which discharge the cake in bulk without any opportunity for sampling individual plates. If nearly all the fat was extracted by organic solvent, the latter will presumably have been expelled to a tiny proportion on safety grounds. If hot-water extraction was used (steam, etc.), the sample must be dried to constant weight at 105–110°C and reported on this basis. If the dried cake is then extracted with a solvent such as petroleum ether, 40:60°C b.p. fraction (or hexane), this solvent is quite selective in removing triglyceride oil. A simple calculation then shows what oil is being lost when the hot water-extracted press cake is jettisoned. Worth noting is that several

206

H.B.W. Patterson

organic solvents such as ether, acetone, chloroform, and methylene chloride are less selective than petroleum ether or hexane, and will extract some part of the remaining organic matter in a press cake (e.g., phospholipid, oxidized fat, pigment, soaps). A most searching solvent coming close to giving the total fatty matter in the cake is a 3:1 mixture of benzene/ethyl alcohol, but regular use of benzene as a laboratory reagent is now discouraged because of its toxicity. The total organic matter can best be determined by ashing a sample of the cake and knowing what amount of ash is given by the dry original clay. Soap will contribute to the ash content of a cake, but considering the very low level of soap (0.005% and less) remaining in oil nowadays and that this contributes only about one-tenth of its own weight to the ash in any case, correcting for this does not seem significant among other sampling/analytical experimental errors. While this searching analysis may be undertaken as the basis of estimating a typical total fatty-matter balance for a process, it will be sufficient to monitor a recovery process by extracting triglyceride. Obviously, if the performance of a solvent plant is being monitored, the same solvent should be used in the laboratory test as in the plant.

Oil Bleachability In general terms, this might be considered simply a complementary test to the evaluation of bleaching clay described at the beginning of this section. In such a case, a sample of a parcel of oil could be treated with the clay currently used by the refiner, and then the test repeated to find how much more—or less—clay was needed to achieve the desired result. In the case of crude palm oil, both adsorption and heat bleaching were involved for a very long time. If the carotene present was damaged little by oxidation, a bleaching procedure sensitive to this fact will show a marked drop in color correlated with a low degree of oxidation and, therefore, an oil very desirable for edible purposes. The Bernam test (Bek-Nielsen, 1974; Krishnan, 1975) gives such indications. Those who required palm oil for technical outlets such as soapmaking were prepared to use more drastic bleaching methods and, hence, were content with a test less sensitive to degree of oxidation. Truly, heavily oxidized palm oil could present a more difficult bleaching task for the soapmaker. Fortunately, the long-standing difficulty in obtaining a widely acceptable routine for testing the bleachability of crude palm oil was overcome in the form of the SCOPA test, BS 684. 2.27 (1987) (Anon., no date; McGinley, 1991). This simulates the procedure in physical refining as degumming, adsorption, and heat bleaching are represented. Crude palm oil is agitated at 90°C under nitrogen for 10 minutes with 0.5% of a 20% solution of citric acid. Then a 2% dose of Tonsil Standard FF clay is added, and the temperature is raised to 105°C; the agitation continues for 15 minutes, and the oil is filtered. A sample of the oil is next raised to 260 ± 5°C (within 10 minutes) and held there for 20 minutes. The color of a cooled sample is read in 5¼” and 1” cells. The test is completed in 2.5 hours; satisfactory oil reads 3.5 R or less (5¼”). Several variations of this test are reported by McGinley (1991), who also mentions a so-called Direct Bleachability Test (DBT) in which a sample of crude oil is agitated at 150°C under carbon dioxide with 5% of Tonsil Standard FF (Hoffman et al., 1975).

Important Tests Relating to Bleaching

207

Acidity of Bleaching Clay This important characteristic of a clay, especially if activated, is expressed in either or both of two ways. In the first way, a stated weight (1 g or 100 g) is boiled with distilled water, filtered, and the required amount of alkali (mg KOH, mL 0.1N, NaOH, etc.) needed to neutralize the extract is determined. In the second way, a suspension of the clay in water is made (1, 2, 10% of clay w/w), and the pH is measured. Different suppliers have their own way of doing these tests. One must remember, as pointed out previously (see section Acidity in Chapter 2), that salts present in some clays exert a marked buffering effect on pH, so that the expected inverse relationship of high acidity showing low pH is distorted. Clays of the same titratable acidity (mg KOH/g) may have a different pH and vice versa. Particle-Size Distribution The effects of particle-size distribution are explained in the discussion of temperature (see section Particle Size in Chapter 2) and at the end of the discussion of acid activation and pigment adsorption (see section Acid Activation and Adsorption of Pigment in Chapter 3). At one time, particle-size distribution was given over a wide range as percentage by weight passing through apertures, meshes, and screens related to national standards, but also in some cases using sedimentation tests. This information could be related to size grade in microns, and could be obtained from the manufacturer. Easily seen by reference to the discussion of commercial bleaching clays (see section Commercial Bleaching Clays in Chapter 3) is that particle-size distribution is now often less lengthy and detailed, but is confined mainly to the important range of 20–80 microns. This simplification fits well—and may be influenced by—the standardization of aperture sizes of 55–80 microns described in detail in Chapter 6 and especially as applied to wire gauze filters (see section Wire Gauzes (Metal Cloths) in Chapter 6). Pore-Size Distribution For activated clays, carbons, and synthetic silicas, the internal-pore surface is much more important than that of the outer surface of all the particles concerned. Furthermore, as with a supported nickel catalyst, access to this huge internal surface via a labyrinth of connecting pores must be adequate in relation to the size of the molecules seeking to reach it. Many authors have made this clear in their own way. One of the most readily understandable expositions of molecular transport in pore channels is given by Coenen (1976). In 1910, Freundlich employed the concept that at constant temperature the concentration of a solute on the surface of an adsorbent depended on its concentration in solution; similarly with gases, the amount of gas adsorbed per unit of surface depends on the pressure above the surface. In 1916 Langmuir suggested that a unimolecular layer of gas was held on the surface of an adsorbing solid and that the equilibrium pressure of the gas at any moment varied with the proportion of the surface then covered until saturation was reached. Chapter 2 of this volume

208

H.B.W. Patterson

examines the bearing of these theories on our understanding of the removal from the oils of pigments and other minor components. Brunauer, Emmet, and Teller in 1938 developed Langmuir’s work to a concept including multilayer adsorption. When the adsorption isotherm (pressure of adsorbed gas at constant temperature compared with pressure at complete saturation = condensation) is interpreted by the BET equation, the volume of the first unimolecular layer is calculable and hence, knowing the molecular diameter, the area of the adsorbent surface can be derived. If nitrogen is the gas used in the experiment to measure the surface area of the adsorbent, the term “nitrogen isotherm” is used. The loose bonding apparently existing between the molecules of the first layer adsorbed and the molecules of the adsorbing surface is referred to as chemisorption. The forces retaining further layers are very much weaker and more in evidence at low temperatures and high pressures. They take their name, Van der Waals forces, from the scientist who proposed their existence (see also section Physical Adsorption and Chemisorption in Chapter 2). Further refinements lead to the calculation of pore-volume distribution versus pore diameter as described by Taylor, Jenkins, and Ungermann (1984). Using nitrogen adsorption and desorption cycles, they obtained the distribution of pore volumes in six different pore-diameter regions of 30–50 Å, 50–100 Å, 100–200 Å, 200–300 Å, 300–400 Å, and 400–600 Å.

Activated Carbon-Adsorption Tests Apart from the nitrogen adsorption/desorption procedures being applied to activated carbon, the specific adsorptions (w/w) of a range of materials of increasing molecular size are used for production control and research purposes to track the presence of pores of different sizes as these appear or disappear during the course of processing of the raw charcoal. The specifications given previously (see sections Commercial Activated Carbon Products–Activated Earth/Carbon Mixtures in Chapter 3) include iodine, phenol, methylene blue, and molasses indices; benzene is also used (see section Activation Procedures in Chapter 3). Adsorptive capacity is also measured by the specific capacity (w/w) to take up carbon tetrachloride (see section Activated Earth/ Carbon Mixtures in Chapter 3). The methodology of the respective tests is available from the manufacturers concerned. The discussion of the activation procedures (see section Activation Procedures in Chapter 3) makes clear how these tests can be used as an investigational tool.

Chapter 10

The Freundlich Isotherm in Studying Adsorption in Oil Processing Andy Proctor1 and J.F. Toro-Vazquez2 1

Department of Food Science, 2650 Young Ave., University of Arkansas, Fayetteville, AR 72704; 2Centro de Investigación y de Estudios de Posgrado de la Facultad de Ciencias Químicas, Universidad Autónoma de San Luis Potosí, Zona Universitaria, 78240 San Luis Potosí, México

Introduction The objectives of this review are: (i) to discuss the history of the use of the Freundlich isotherm in investigating the adsorption processing of vegetable oils and (ii) to evaluate its current value in adsorption studies. The Freundlich isotherm was originally developed to explain the adsorption of a single solute from a solution. However, a similar adsorption pattern was observed when studying a more complex system of adsorption of vegetable oil pigments onto bleaching clay during the commercial bleaching of vegetable oils. The Freundlich isotherm has been useful for decades in determining the commercial value of adsorbents providing a narrow experimental interval of adsorbate is used. More recent studies show that a complex series of interactions controls the adsorption process. While the isotherm summarizes these reactions, investigating them is vital to understanding the physicochemical factors involved during adsorption. In addition, statistical modeling and spectroscopy are also useful in understanding the vegetable oil bleaching/refining process as a multiple-component adsorption system. Adsorption involves the separation of a substance from a liquid, or gas phase, and its accumulation on the adsorbent surface. Adsorption bleaching of vegetable oils was originally designed to remove oil chlorophyll and carotenoid pigments. However, it is now used to adsorb a number of other oil components that adversely affect oil quality and stability, including free fatty acids, phospholipids, peroxides, and carbonyl compounds. Bleaching, the only adsorption operation used in vegetable oil refining, nevertheless, is an effective operation in improving oil quality. Researchers have used the Freundlich isotherm for more than 60 years to describe the adsorption of pigments onto adsorbents in vegetable oil bleaching. During this time, the understanding of the mechanisms and physical chemistry of the adsorptionbleaching process significantly increased. However, limitations still exist in explaining the adsorption of other oil components (e.g., fatty acids and phospholipids) from a relatively complex multi-component system, such as vegetable oils. The objectives of this chapter are to review the history of the use of the Freundlich isotherm in studying adsorption bleaching, and to evaluate the current value of the Freundlich isotherm in studying vegetable oil lipid adsorption.

209

210

A. Proctor and J.F. Toro-Vazquez

The Freundlich Isotherm Adsorption is a physical chemical process involving the mass transport of an adsorbate from the solution phase (i.e., bulk solution, film, and intraparticle transportation) to the interior surface of the porous adsorbent where the adsorption occurs. When the thermodynamic equilibrium of adsorbate concentration is established between the solution and the adsorbent, no further net adsorption occurs (Weber, 1985). This equilibrium is defined by the concentration of adsorbent and adsorbate in the system, and conditions of temperature, viscosity, and pH. Adsorption equilibrium is the most fundamental property of the adsorbate–adsorbent interaction. Therefore, the theoretical and empirical models that describe reversible adsorption were developed on the basis of a thermodynamic equilibrium. Several equilibrium models were developed to describe adsorption–isotherm relationships (Table 10.1). However, no single model was generally descriptive of the process. Thus, Langmuir and BET isotherm models often fail to describe experimental data in vegetable oil systems, despite their sound theoretical basis. Although the Freundlich equation is an empirical model, it is widely used to describe vegetable oil adsorption bleaching (Brown & Snyder, 1989; Mingyu & Proctor, 1993; Palaniappan & Proctor, 1990, 1991; Patterson, 1992). Freundlich (1924) described the adsorption of a single adsorbate in an aqueous solution as a reversible equilibrium, when the equilibrium was established in a few seconds or minutes at a fixed temperature. The Freundlich equation (Table 10.1) states that, at a constant temperature, the amount of adsorbate bound per unit weight of adsorbent, Qe (adsorption efficiency of the adsorbent) is a logarithmic function of the residual concentration in the fluid phase at equilibrium, Ce. At low-solute concentrations, the amount of adsorbate adsorbed increases greatly with an increase in solution concentration. However, at larger concentrations of adsorbate, the amount adsorbed approaches a constant value (Fig. 10.1). Thus, the gradient of the curve is greatest at TABLE 10.1 Common Isotherm Adsorption Models Based on Thermodynamic Equilibrium Adsorption isotherm

Equationa

Freundlich

Qe = kfCe(1/nf) also used as Qe = kfCen as used by oil processors x/m = KCn

a

Brunauer-Emmet-Teller (BET)

Qe = (BCeQ)/(Ce–Cs)[1 +(B – 1)(Ce/Cs)]

Langmuir

Qe = (QbCe)/(1 + bCe)

Linear

Qe = KpCe

Legend: Qe = x/m, the amount of substance adsorbed per mass of adsorbent at equilibrium; Ce = C, residual adsorbate concentration in solution at equilibrium; x, amount adsorbed per mass of adsorbent; m, amount of adsorbent; K = kf, intercept and n = 1/nf , slope of Freundlich isotherm when expressed as log(x/m) = log K + nlog(C); Q, adsorbate surface concentration when all available adsorption sites are occupied; b, adsorption equilibrium constant (ratio of adsorption and desorption rates); Cs, adsorbate concentration to saturate the solvent (solubility limit); B, energy adsorption constant; Kp , the partition coefficient of the adsorbate.

The Freundlich Isotherm in Studying Adsorption in Oil Processing

211

Fig. 10.1. Freundlich isotherm in its logarithmic (A) and linear (B) forms. The isotherm is shown using two different nomenclatures, Qe = kfCe(1/nf ) and x/m = KCn. The definitions of the terms are described in Table 10.1.

low-solute concentrations, and decreases with solute concentration. An important limitation of the Freundlich equation is that it does not describe a limit in adsorption capacity; theoretically, the amount adsorbed may be infinite as solute concentration increases (Suzuki, 1990) (Fig. 10.1). In addition, paradoxically, the Freundlich isotherm was successfully used to describe the adsorption bleaching of oils to bleaching clays, as the equation describes a reversible adsorption, whereas the adsorption bleaching itself is irreversible. One can justify the widespread use of the Freundlich equation in the oil-bleaching process and other industrial processes for three reasons: (i) for practical purposes, the equation is adequate to describe a nonlinear adsorption in a narrow range of adsorbate concentrations, (ii) the mathematical simplicity of the equation enables it to be used easily, and (iii) the Freundlich model describes the adsorption process on surface adsorp-

212

A. Proctor and J.F. Toro-Vazquez

tion sites which are energetically heterogeneous (Suzuki, 1990), a condition commonly found in adsorption systems. In contrast, the Langmuir and the BET models are based on the assumption that adsorption takes place on energetically uniform adsorption sites, which is seldom achieved in oil-bleaching systems. Although the Freundlich model (Table 10.1) is based on empirical concepts, the parameters of the equation, kf (or K) and 1/nf (or n), are relative indicators of adsorption capacity and energy of adsorption, respectively (Palaniappan & Proctor, 1991; Suzuki, 1990; Weber, 1985). The values kf (or K) and 1/nf (or n)—obtained from the fitting of the experimental data with the logarithmic form of the equation [i.e., Log(Qe) = Log(kf) + 1/nfLog(Ce) or Log(x/m) = Log(K) + nLog(C), Fig. 10.1]—depend on factors other than the concentration of the adsorbate, namely, temperature, nature of the adsorbent, adsorbate, the solvent characteristics, and the liquid-phase viscosity (Paulose et al., 1978; Frankel, 1984). The oil-processing industry uses the Freundlich equation with a different nomenclature (Table 10.1); the equivalence among the nomenclatures utilized is shown in Table 10.1 and Fig. 10.1.

Conventional Adsorption Bleaching Hassler and Hagberg (1939) pioneered the application of the Freundlich isotherm to describe adsorption bleaching. The work described the industrial adsorption of cottonseed oil “color bodies” by Fuller’s earth and active carbon using Lovibond units. The adsorption isotherm in a log/log plot showed a linear behavior allowing the comparison between the adsorbents. The authors indicated that the slope of the log/log isotherm (i.e., n or 1/nf ) was a useful index of the adsorption efficiency of adsorbents: that is, the steeper the slope, the greater the adsorption efficiency of the adsorbent (Hassler & Hagberg, 1939). The work was important in establishing the potential of the Freundlich isotherm as a means of evaluating adsorbents and distinguishing between them. However, the Freundlich-isotherm model was developed in a single-adsorbate system measuring the residual-adsorbate concentration, whereas the Hassler and Hagberg system (Hassler & Hagberg, 1939) was a multi-component lipid system which used subjective Lovibond-color measurements. Nevertheless, assuming that oil’s Lovibond color is proportional to pigment concentration, adsorption data resembling a Freundlich isotherm were reported (Hassler & Hagberg, 1939). This work indicated that since the isotherm represents the concentration equilibrium for the adsorbate, an adsorbent in contact with a lightly colored oil will have reserve adsorptive capacity for additional pigment from darker oils. Hence, a countercurrent procedure whereby fresh oil is first contacted with partly exhausted adsorbent which can no longer take up pigment (or other oil compounds) from partly treated oil is still capable of adsorbing a significant amount of pigments from an unbleached oil (Hassler & Hagberg, 1939). This was later used to explain the so-called “press bleaching effect,” a phenomenon observed during the continuous filtration of the oil–adsorbent mixture during bleaching, which results in a substantial cake building-up on the filter cake. This filter press cake acts as a fixed-bed adsorption column, and if the adsorbent has additional adsorption capacity, additional pigment binding by the adsorbent occurs during the filtration of the oil (Henderson, 1993; Patterson, 1992).

The Freundlich Isotherm in Studying Adsorption in Oil Processing

213

By using different bleaching earths to bleach soybean oil, Hinners et al. (1946) showed that the adsorptive capacity of the adsorbents can not be adequately evaluated by measuring Lovibond color. In most cases, the color of oils is the result of a mixture of different pigments, and the Lovibond color is unable to obtain a satisfactory oil-color match that directly represents the concentration of each particular pigment. Furthermore, the ratio of various pigments changes after bleaching as a function of adsorbent concentration and type of adsorbent. Thus, Lovibond measurements are insufficient to explain differences in pigment adsorptive capacity among adsorbents (Hinners et al., 1946). In general, the technique of color measurement used to evaluate bleaching efficiency must represent the relative proportion of each color to the overall color of the oil. Additionally, the color units must be proportional to pigment concentration (Hinners et al., 1946). In fact, the authors used spectrophotometry to determine chlorophyll to evaluate the adsorption capacity of different bleaching earths by means of the Freundlich isotherm (Hinners et al., 1946). The isotherm equations obtained for the different bleaching earths indicated that the slope, n, was the same among the earths, suggesting that the nature of the adsorption sites was probably similar (Fig. 10.2). The analysis of the equations and the respective log/log plots indicated that the constant K in the Freundlich equation was directly proportional to the bleaching capacity of the earth for chlorophyll (Fig. 10.2). In general, a large K value for an adsorbent is always desirable because it indicates the ability to bleach oils to a low residual pigment concentration. Thus, according to Hinners et al. (1946), K is the number of “color

Fig. 10.2. Chlorophyll-adsorption isotherms obtained by bleaching soybean oil (30 minutes at 110 oC in vacuo) with different commercial bleaching earths (A, B, and E) using adsorbent concentrations within 1.0 to 6.0%. The chlorophyll concentration was estimated in relative units (i.e., chlorophyll density) by using spectrophotometric data. The Freundlich equations that fit the experimental data are shown for each particular earth using the format x/m = KCn. From Hinners et al. (1946).

214

A. Proctor and J.F. Toro-Vazquez

units” that each gram of earth would hold if the oil was reduced to one “color unit.” A large n value is also beneficial because it indicates the ability to efficiently bind pigment from highly colored oils (e.g., n indicates the interval over which the adsorbent is most effective for color reduction). However, a high n value should not be at the expense of K. Within this framework, the Freundlich equation and its constants allowed the prediction of the earth doses and earth types necessary to produce an oil of a desired color (Henderson, 1993). Stout et al. (1949) used the log/log plot technique to survey a large number of clays and silicates to estimate their relative effectiveness to decolorize cottonseed and soy oils. King and Wharton (1949) noted that reactions other than adsorption may be occuring during industrial adsorption bleaching, which may affect pigment binding. Atmospheric adsorption bleaching produced Freundlich log/log plots with K and n lower than those obtained with the same adsorbent in vacuum bleaching. In atmospheric bleaching, oil oxidation catalyzed by the adsorbent was responsible for darkening the oil and affecting the Freundlich constants. As with previous studies (Hassler & Hagberg, 1939), subjective Lovibond values were used as units of pigment concentration, and the study could not distinguish between natural-oil pigments and pigments produced by oxidation during bleaching. Gutfinger and Letan (1979) used Freundlich log/log plots to measure phospholipid (PL) and pigment binding by adsorbent clays from phosphoric-acid-degummed soybean oil. PL concentration was measured at the level of µg/g of oil with a specific technique, but pigment was reported as Lovibond color. PL binding by Tonsil adsorbents and Fuller’s earth conformed to a Freundlich isotherm. However, for pigment binding as evaluated by Lovibond color, only the Tonsil adsorbent follows the Freundlich isotherm. This was explained as being due to an increase in oil color due to oxidation by Fuller’s earth as previously mentioned by King and Wharton (1949). Unfortunately, the data were fitted to the logarithmic form of the Freundlich equation with only three experimental data points. With such a small number of data points in a narrow experimental interval, a high correlation coefficient might be obtained by random chance alone. In this work, the Freundlich isotherm was used for its mathematical simplicity and the significance of its empirical constants; nevertheless, little attempt was made to understand why the Freundlich equation was effective in describing adsorption in a multi-component oil system, prone to oxidation and catalysis, which are very different conditions from those described in Freundlich’s original work. The significance of this work was to show the empirical value of the Freundlich isotherm to allow comparisons between bleaching clays, and other potential adsorbents, used for commercial oil bleaching. More recent studies used the Freundlich isotherms to evaluate: the bleaching of novel oils (Achife & Ibemisi, 1989; Bayrak, 2003), adsorbents (Huang et al., 2007; Rossi et al., 2003), and the effect of temperature on the isotherms to increase maximal adsorption (Achife & Ibemisi, 1989; Boki et al., 1992; Rossi et al., 2003).

Dilute-Miscella Processing Miscella vegetable oil refining has been successfully utilized in industrial operations for over 35 years. The main advantages of continuous miscella refining compared to

The Freundlich Isotherm in Studying Adsorption in Oil Processing

215

continuous conventional refining of vegetable oils are: the lower refining losses; the possibility of dewaxing, winterizing, and hydrogenation on the continuous operation right after miscella refining; and the lower degree of oil oxidation, color fixation, and flavor reversion in the refined oil because miscella refining is performed right after the extraction process (Cavanagh, 1990). Particularly, oil miscella (i.e., oil:organic solvent mixtures) bleaching is used to avoid the use of high temperature needed to reduce oil viscosity (Feuge & Janssen, 1951; Toro-Vasquez & Mendez-Montealvo, 1995), thus favoring film and intraparticle transportation of the adsorbate to the interior surface of porous adsorbent where the adsorption takes place. Thus, effective bleaching is possible at lower temperatures, avoiding oxidation problems experienced at high temperatures, such as the production of new pigments. Feuge and Janssen (1951) studied the reduction in Lovibond-red values in cottonseed oil in a hexane solution by adsorption with a commercial-grade activated clay. In general, a greater reduction in the color of refined oil was obtained by miscella–oil bleaching (i.e., 30% of oil–miscella concentration at 25°C) than with conventional adsorption (i.e., oil bleaching at 25 or 110°C), independent of the clay dosage utilized (0.5–8.0%). The authors also showed that the bleaching efficiency decreased linearly when the concentration of oil in the miscella increased from 10 to 100%. In this work, the Freundlich equation was used to compare the different adsorption procedures; thus, bleaching oil with activated clay at 25°C, with and without hexane, produced similar n values (5.10, 4.55, respectively), but processing in hexane solution furnished a greater K value (0.002, 0.0006, respectively). These results indicated that the presence of solvent did not significantly change the mechanism of adsorption (i.e., similar n value), but did significantly increase maximal pigment adsorption approximately 3.3 times (i.e., ratio between the respective K values). Under conventional bleaching conditions at 110°C, a larger K value was obtained (0.07), showing a greater decolorizing power in contrast to miscella bleaching; nevertheless, a significant lower n value was achieved (2.39), which indicated that miscella bleaching was more effective in removing the first portion of color than conventional bleaching. On the basis of these results, the authors proposed a countercurrent bleaching procedure by using 30% of oil miscellas (Feuge & Janssen, 1951). Proctor and Snyder (1987) used miscella refining to investigate the nature of the Freundlich isotherm, and how it could be used to better understand the theoretical aspects of pigment adsorption from vegetable oils. The introduction of polar species into the soy oil–hexane miscella (e.g., isopropanol) and the water deactivation of the silicic-acid adsorbent reduced the maximal amount of pigment binding. They proposed that lutein, the major soy-oil carotenoid, did not occupy all the available adsorption sites because the pigment was competing for adsorption sites with isopropanol or with oil components that were present in larger concentrations than lutein (i.e., s). This hypothesis was further tested in adsorption studies with lutein-rich–low-triglyceride mixtures in hexane prepared from crude soybean oil. In this case, the majority of the pigment was bound, producing a straight-line isotherm close to the vertical axis (Fig. 10.3). Pure lutein in hexane was totally bound by silicic acid; therefore, triglyceride and other substances in the miscella limited pigment binding. Competitive adsorption was

216

A. Proctor and J.F. Toro-Vazquez

proposed as an important factor in controlling the Freundlich isotherm in miscellas, and probably plays an important role in oil systems (Proctor & Snyder, 1987). The polarity of oil components is probably an important factor in this competition (Mingyu & Proctor, 1993). Earlier, Brown and Snyder (1985) showed that PL was adsorbed from a soy oil– hexane miscella according to a Freundlich isotherm. Because the Freundlich isotherm is a reversible process, the reversibility of PL was examined by incubating silicic acid with two soy-oil–hexane miscellas, the second miscella with a lower PL concentration than the first. However, the second incubation resulted in additional binding of PL, and not PL desorption as one would expect if the adsorption was irreversible. PL is a polar–ionic lipid molecule which could be an important competitor of soy-oil pigment for adsorption sites. “Irreversible” adsorption may be caused by the polar–ionic PL being more thermodynamically stable as a bound complex than in a nonpolar solution; the ionic portion of the molecule may be involved at the adsorption site (Adhikari et al., 1995). Hence, in addition to thermodynamics, the chemical equilibria of the molecular species involved may affect the nature of the isotherm in a soy oil–hexane system. In this system, maximal PL adsorption was enhanced by the addition of a polar solvent (i.e., isopropanol) to the soy oil–hexane miscella. Later research showed that the enhanced PL adsorption could be due to the disruption of PL miscellas by isopropanol-enabling individual PL molecules to adsorb more freely (Brown & Snyder, 1989). This is in contrast to the reduction in lutein binding from soybean oil by the same polar solvent. However, the incomplete adsorption of PL was associated with the competition of triglycerides for adsorption sites (Brown & Snyder, 1989), as was also proposed for soy-oil pigments (Boki et al., 1992). Although triglyceride is not as polar

Fig. 10.3. Adsorption of lutein from a miscella of soybean oil in hexane (miscella) and an extract of a lutein-rich–low-triglyceride mixture in hexane prepared from crude soybean oil (extract). From Proctor and Snyder (1987).

The Freundlich Isotherm in Studying Adsorption in Oil Processing

217

as PL, it is present at a much greater concentration, which, in turn, results in a higher thermodynamic drive for adsorption. Similar findings were reported using rice-hull ash in adsorbing soy-oil pigment (Brown & Snyder, 1989; Palaniappan & Proctor, 1990) and PL (Palaniappan & Proctor, 1992 ) from soy-oil–hexane miscellas. Chapman and Pfannkoch (1992) used single-component isotherms and the ideal dilute-solution theory (IDST) to study the co-adsorption of protoporphyrin IX dimethyl ester, fatty acids (i.e., stearic and linolenic acids), and triglycerides (i.e., tripalmitin and trilinolein) on acid-activated bleaching clay by using dichloromethane as the solvent. IDST predicts a two-solute adsorption from single-solute adsorption data, providing that the solutions are diluted (Radke & Prausnitz, 1972). The experimental relative affinities to bleaching clay from single-component isotherms were protoporphyrin >> fatty acids and triglycerides. The isotherms for binary mixtures of low protoporphyrin concentrations with higher concentrations of triglycerides or fatty acids were calculated from single-component isotherms’ data by using IDST. The resulting isotherm showed dramatically the ability of the lipids to suppress protoporphyrin adsorption. This investigation pointed out the importance of competitive adsorption in determining the performance of adsorbents in oil bleaching.

Concentrated-Miscella Processing Investigations based on the Freundlich isotherm equation were performed with dilute miscellas of vegetable oils (i.e., <40% of oil in solvent) and short contact times (i.e., <20 minutes) while assuming thermodynamic adsorption equilibrium. However, in miscella-oil refining, the functional-oil concentration utilized is within 40 to 80% (Cavanagh, 1990). At such miscella-oil concentrations, additional kinetic and diffusion factors (i.e., oil oxidation and viscosity) might limit the achievement of a real thermodynamic adsorption equilibrium (Toro-Vazquez & Rocha-Uribe, 1993; Toro-Vazquez & Mendez-Montealvo, 1995). Furthermore, oil miscellas are mixtures of compounds with a great diversity in physicochemical properties, chemical reactivity, molecular structure, and concentrations. Consequently, each organic component has a different degree of adsorption on the adsorbent. Thus, the adsorption process during miscella-oil bleaching/refining must be considered as an adsorption process of mixtures of organic compounds (i.e., a multi-component adsorption system). Consequently, the adsorption profile of a given oil component would depend on the level of adsorption of other components and their adsorption mechanism, both affected by the respective concentration of the different oil components in solution as well as on the adsorbent. Through infrared spectroscopy, Adhikari et al. (1994) showed that isopropanol, an oil solvent with potential use in the industry, reduced free oleic-acid adsorption (i.e., free fatty acids) on silicic acid. This effect was explained by a combination of competitive adsorption of isopropanol, as well as an interaction between isopropanol and oleic acid in solution. Toro-Vazquez and Rocha-Uribe (1993) demonstrated the complex nature of interactions among different oil components in determining the adsorption-isotherm behavior. These authors utilized a particular experiment design (i.e., split-plot design) to evaluate the effect

218

A. Proctor and J.F. Toro-Vazquez

of an oil–miscella concentration (60–-100% in hexane) and the presence of different levels of ethanol (0, 5, and 25%) on the vegetable-carbon adsorption at 50°C of peroxides, unsaturated carbonyls, free fatty acids, and carotenoids from sesame-oil miscellas. In general, adsorption was promoted as solvent concentration increased in the miscella (i.e., lower viscosity) and a pro-oxidant effect of ethanol was observed. Furthermore, Toro-Vazquez and Mendez-Montealvo (1995), using multiple regression analysis, studied the competitive adsorption on vegetable carbon among free fatty acids, carotenoids, and unsaturated carbonyls. This study utilized different sesame-oil batches at oil–miscella concentrations normally found in industrial miscella refining (40 and 60%). The behavior of the adsorption isotherms suggested that unsaturated carbonyls promoted free fatty acids’ adsorption, mainly in the pores that were readily accessible for fatty acids. However, in the presence of high concentrations of carbonyls, the less accessible adsorbent pores were occupied by carbonyls, decreasing fatty acids’ adsorption. Thus, highly oxidized oils would decrease the adsorption efficiency for fatty acids due to the competing effect of carbonyls (Toro-Vazquez & MendezMontealvo, 1995). In contrast, carotenoids’ adsorption was substantially reduced by free fatty acids, in agreement with the results for protoporphyrin IX and fatty acids’ competing adsorption obtained by Chapman and Pfannkoch (1992). The Freundlich isotherm failed to fit the experimental data. However, in the mathematic expression of the Freundlich isotherm, the effect of components’ interactions on adsorption efficiency is not considered. The interpretation of adsorption isotherms becomes complex when two or more adsorbable components can occupy the same adsorption sites [Fig. 10.4 (Suzuki, 1990; Toro-Vazquez & Mendez-Montealvo, 1995]. Equations that deal with the isotherm adsorption of a mixture of compounds were shown by Suzuki (1990). Some of these isotherms are mathematical extensions of the Freundlich and Langmuir equation with relatively large numbers of empirical parameters and a lack of thermodynamic background (Suzuki, 1990). To the authors’ knowledge, these equations were not used in oil-bleaching studies.

Summary The Freundlich isotherm was originally presented to explain the adsorption of a single solute from solution. However, a similar adsorption pattern was observed when studying the much more complex system of adsorption of vegetable oil pigments onto bleaching clay during the commercial bleaching of vegetable oils. The Freundlich isotherm was useful in comparing the commercial value of adsorbents if a narrow experimental interval of adsorbate was utilized. More recent studies showed that a complex series of interactions control the adsorption process, and while the isotherm summarizes this interaction, investigating these factors is vital to understand the physicochemical factors involved during adsorption. Statistical modeling and experiment design are proving effective in evaluating vegetable oil bleaching/refining as a multiple-component adsorption system. Furthermore, the use of physical–chemical techniques to elucidate the nature of the solute–adsorbent complex and the solute–solvent interaction process provided

The Freundlich Isotherm in Studying Adsorption in Oil Processing

219

Fig. 10.4. Behavior of the Freundlich isotherm using the dilution and batch method in a vegetable oil with a hypothetic mixture of two free fatty acids with concentrations C1 and C2. A, isotherm in the form of x/m = KC1/n; B, isotherm in the form of Log (x/m) = Log (K) + (1/n)Log(C). x/m is the concentration of free fatty acids adsorbed per mass of adsorbent, and C is the residual concentration of free fatty acids in the oil. Modified from Suzuki (1990).

the basis for understanding the interactions contributing to adsorption (Adhikari et al., 1994, 1995). These techniques support the development of new isotherm models applicable to oil-bleaching/refining. However, isotherm studies using the Freundlich model do have value in summarizing the main factors controlling the adsorption of vegetable oil in a conventional or miscella system.

Chapter 11

Enzymatic Degumming of Edible Oils and Fats David Cowan and Per Munk Nielsen Novozymes A/S, DK-2880 Bagsvaerd, Denmark

Introduction Vegetable oils need to to be refined, and this is partly to eliminate the phospholipids. These compounds will have a negative effect on the taste and shelf life of the oil. Numerous different refining/degumming processes were developed, and are classified as physical or chemical processes (Dijkstra, 1992). All methods aim to reduce the phospholipid content of the oil, measured as phosphorus, to below 10 ppm. One of the most recent methods is enzymatic degumming. The use of enzymes in the degumming process dates back to 1994 when they were used for the first time in large-scale production (Anonymous, 2001). The enzymatic method is based on the conversion of nonhydratable phospholipids to hydratable lyso-phospholipids, which are easy to eliminate with the water phase by centrifugation. The method is applied to a wide range of oils, and the number has grown since the method was first introduced. Phospholipids in Vegetable Oil Several different types of phospholipids are in vegetable oils (Fig. 11.1). Phospholipids are part of plant-cell walls including those containing the lipids. The amount of phospholipids that are extracted together with the oil depends on the extraction temperature. Both the oil-extraction rate and the content of phospholipids in the oil are increased by increasing the extraction temperature. A temperature increase from 55 to 57°C can result in a change of phospholipid content in soybean

Fig. 11.1. Structure of the most common phospholipids: R1, R2: Fatty acid residues, PA = phosphatidic acid, PI = phosphatidylinositol, PE = phosphatidylethanolamine, PC = phosphatidylcholine, and PS = phosphatidylserine. 221

222

D. Cowan and P.M. Nielsen

oil from 0.70 to 0.85% (Bockisch, 1998). Thus, processing systems which maximize oil yield also increase the phospholipid level in the resulting oil. The ability of these phospholipids to be hydrated governs the ease with which water can wash them out of the oil. A large difference is noted in the rate of hydration of the different phospholipids. The order of the hydration rate of the different phospholipids is PC > PI > PE > PA (phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine, phosphatidic acid, respectively). This means that the nonhydratable phospholipids (NHP) mainly consist of PA. Figures for the composition of NHP in different types of oil were reported (Buchold, 1995), and are shown in Fig. 11.2. PA is mainly produced in the oil seed by the conversion of the other types of phospholipids as a result of enzymatic activity in the seed, mainly by phospholipase type D, (which is described in Fig. 11.4). The amount of PA varies not only with the type of seed but also with the treatment of the seed. If poor storage conditions (e.g., high-moisture content and elevated temperature) have already allowed the phospholipase D to act, the adverse effect of the storage cannot be overcome by the inactivation of the enzyme before extraction (List & Mounts, 1993). The approximate content of phospholipids in different types of oils is shown in Table 11.1. Commonly, one can use a conversion factor of 28 to calculate the content of phospholipids (ppm) from ppm phosphorus.

Fig. 11.2. Composition of nonhydratable phospholipids (NHP) from different oils. PI = phosphatidylinositol, PC = phosphatidylcholine, PE = phosphatidyl-ethanolamine, and PA = phosphatidic acid. Data from Buchold, (1995).

223

Enzymatic Degumming of Edible Oils and Fats

TABLE 11.1 Content of Phospholipids in Different Oils Type of oil Soybean

Phospholipids in oil, %

Phosphorus in oil, ppm

1.8–3.2

640–1140

Rapeseed

0.7–1.7

250–600

Sunflower

0.5–1.0

170–360

Corn

1.0–2.0

360–720

Ricebran

0.8–1.1

300–400

Linseed

1.8

640

Figures are partly from Pardun (1989) and Nielsen (2003).

Phospholipids are molecules with an amphiphilic character. This means they have a high preference for distribution at the oil–water interface, and act as emulsifiers. Temperature has an influence on the structure and dispersibility of phospholipids in oil–water mixtures. At low temperature, the phospholipids are organized as laminar double layers, and are relatively easy to disperse in water, where they locate at the interface between water and oil. At an elevated temperature, the tendency to make globular structures develops. These structures are not readily dispersible in the water phase (Fig. 11.3) (T.H. Callisen, personal communication, 2003 ). The transition temperature is in the range of 45–55°C, depending on the fatty acid composition.

The Biochemistry of Phospholipases Different types of phospholipases are classified according to where the catalyzed hydrolysis reaction takes place on the phospholipid molecule (Fig. 11.4). Of these, only the two A-types were used by the industry on a regular basis. The reaction scheme for the conversion of the phospholipids to lyso-phospholipids by means of a phospholipase type A1 is illustrated in Fig. 11.5. More recently, a phospholipase C was developed and used on a limited basis by itself and in combination with phospholipase A1 (Dayton, 2008). In enzymatic degumming, phospholipase A1 converts phospholipids to the lyso form as is shown in Fig. 11.6 (a and b) (Clausen, 2001). When the enzymatic treatment

Fig. 11.3. Organization of phospholipids in the oil dependent on temperature. Enzymes for the degumming process

224

D. Cowan and P.M. Nielsen

Fig. 11.4. Types of phospholipases (A1, A2, B, C, and D) distinguished according to which bond they are hydrolyzing. R1 and R2 represent fatty acids, and X is hydrogen, choline, ethanolamine, serine, or inositol.

is started, PA plus a small amount of PE is found in the oil phase. The amount of PC and PI in the oil phase is often lower than the detection limit for high-performance liquid chromatography (HPLC) detection. PA from both the water phase and oil phase is gradually converted to lyso-phosphatidic acid (LPA) which is then found in the water phase. Lyso-phosphatidylcholine (LPC) and lyso-phosphatidylinositol (LPI) are accumulated in the water phase. Also, the PC and PE in the water phase are gradually disappearing and becoming converted to LPC and lyso-phosphatidylethanolamine (LPE), respectively. In vegetable oils, the packing of the phospholipids at the oil–water interface is very dense, but despite this the reaction does occur. To create such a system, the oil and the water phase must be emulsified efficiently. This requires high-shear mixing to distribute the low amount of water used evenly throughout the much larger mass of oil.

Comparison of Enzymatic and Other Degumming Methods In Fig. 11.7, chemical refining is compared to physical refining. Enzymatic degumming is a special case of physical refining. The main difference between the chemical

Fig. 11.5. The conversion of phospholipids to lyso-phospholipids ensures that all the phosphatides in the oil become hydratable.

Enzymatic Degumming of Edible Oils and Fats

225

Fig. 11.6. Distribution between hydratable (a) and nonhydratable (b) phospholipids/lyso-phospholipids in rapeseed oil and the sludge phase after reaction with phospholipase A1. The oil is water-degummed rapeseed oil hydrolyzed at 45°C. PA = phosphatidic acid, PI = hosphatidylinositol, PE =phosphatidylethanolamine, and PC = phosphatidylcholine. LPA = lyso-phosphatidic acid..

and physical refining is the way the free fatty acids (FFAs) are eliminated. In chemical refining, the FFAs are converted to soapstock by the addition of sodium hydroxide. At the same time, all the phospholipids are hydrated and are removed with the soapstock. The elimination of the soapstock requires several washing/centrifugation steps; this causes a significant oil loss. In physical refining, different methods are used to assure the conversion of NHP to hydratable phospholipids, which one can then eliminate normally by washing followed by centrifugation. This leaves the FFAs in the oil until the deodorizer. In the deodorizer, the FFAs are stripped off, together with sterols, tocopherols, etcetera.

226

D. Cowan and P.M. Nielsen

Crude oil

Alkali treatment

Water degumming

Enzymatic reaction

hydratable P-lipids

Acid degumming

3 times neutralization and washing/ centrifugation

non-hydratable P-lipids

Centrifugation

Centrifugation

Bleaching

Bleaching

Bleaching

Deodorization

Deodorization

Deodorization

FFA and P-lipids as soapstock

FFA

RBD Oil Fig. 11.7. Flow chart comparing chemical refining to physical refining, including enzymatic degumming. RBD, refined, bleached, deodorized; FFA, free fatty acid.

As a result, the physical-refining methods require more capacity on the deodorizer compared to chemical-refining methods, but oil loss is reduced. Gibon and Tirtiaux (2000) made a review of these different processes. The key advantage for enzymatic degumming is that the process converts all the gums to a hydratable form, and therefore virtually eliminates oil losses. No oil is entrapped in the gums; therefore, the physical yield of oil can be 1.0–1.5% higher, depending on the process and the oil to be degummed. This makes enzymatic degumming the process with the lowest yield loss of all the different degumming methods. All of the different types of degumming have to reach a low content of phosphorus, but only enzymatic degumming is able to do this, while at the same time reducing oil losses to a minimum. Normally, physical degumming, while being preferred because of the use of less aggressive chemicals, is not seen as being robust enough to tackle as wide a range of conditions as chemical refining. Enzymatic degumming combines this with the ability to cope with a very wide range of oils and oil qualities. A mass balance for the different processes was calculated (Cowan & Holm, 2007), and illustrates the anticipated loss in oil, starting from 1,000 kg of crude soybean oil. Enzymatic degumming of crude oil gives the highest output of oil (Table 11.2), and is suitable for both crude and already water-degummed oils.

Enzymatic Degumming of Edible Oils and Fats

227

TABLE 11.2 Overall Oil Output (Deodorized Oil) Obtained via Different Degumming Methods Degumming method

Kg of oil obtained

Oil loss %

Water degumming + chemical refining

942.7

5.73

Water degumming + semi physical refining

944.3

5.57

Water degumming + acid degumming

945.9

5.41

Water degumming + enzyme 948.2 degumming

5.18

Full enzyme degumming

4.79

952.1

As oil prices increase, so does the attractiveness of this form of degumming as is evident by the increasing take-up of this technology.

Process Details 1. The process equipment for enzymatic degumming must provide a system where the enzyme is able to work at optimal conditions. Enzymatic degumming is seen as a three-step operation in which it is necessary to: Create an emulsion with the phospholipids distributed at the water–oil interface, 2. Provide the conditions for the enzyme to work efficiently, and 3. Separate the phospholipids/gums from the oil. The process is shown in Fig. 11.8, starting with crude or water-degummed oil. The first part of the process is the citric-acid step, in which a small amount of citric acid (0.04–0.1%) is added to the oil as a concentrated solution (45–50%). The citric acid is distributed with a high–shear mixer, and allowed to react in a holding tank with a retention time of 10–30 minutes at a temperature of 70°C. The citric acid helps to chelate the metals in the oil, and opens the phospholipid micelles to allow enzyme action upon the phospholipids. Following the citric-acid treatment, the oil is cooled to 50–55°C to bring the temperature to an optimal level for the enzyme reaction. Sodium hydroxide (NaOH) is added to adjust the pH of the water phase. The optimal amount is 1.5 mol of NaOH for each mol of citric acid. Water (up to 1.5–2.5% of total water) is added together with the enzyme, and includes the water added as part of the citric acid, NaOH, and enzyme solutions. In the first-generation processes, an enzyme dosage of 30 ppm (30 g/1000 kg of oil) of Lecitase Ultra is used or the corresponding amounts of other phospholipases. A second high-shear mixer is employed to assure a complete distribution of the ingredients and the creation of a water-in-oil emulsion with the phospholipids distributed at the water–oil interphase.

228

D. Cowan and P.M. Nielsen Citric Acid

NaOH

50 % Citric Acid in water

4N NaOH

Enzyme Lecitase Ultra water

shear mixer

shear mixer

centrifuge

heating 70°C

cooling 60–68°C

cooling 70°C

Crude Oil or Water Degummed Oil

Separated gums

Degummed vegetable oil

Fig.11. 8. Continuous enzymatic-degumming process.

The use of a high-speed shear mixer to ensure that the enzyme is evenly distributed in very small water droplets throughout the oil is a key element in achieving good degumming. The conversion of the phospholipid to lyso-phospholipid takes place at the oil–water interface, and clearly, the more finely the droplets are divided, the greater the interface surface area will be, and the more efficient the degumming process. Besides the enzyme, some chemical additions are still required, and these are summarized in Table 11.3 for the degumming of 10 MT/hour of oil.

TABLE 11.3 Ingredients for Enzymatic Degumming of 10,000 kg of Oil per Hour Dosage

Kg per hour

Citric acid

0.065%

7.1 kg of citric acid monohydrate as a 45% solution

Sodium hydroxide (NaOH)

1.5 molar equivalents

2.03 kg of NaOH as a 4N solution

Water

2.5% of total water

226 kg beside what comes in with the other ingredients

Enzyme (Lecitase Ultra)

30 ppm

0.3 kg

Oil

10,000

Enzymatic Degumming of Edible Oils and Fats

229

The second phase of the process is to allow enough time for the enzyme to react. This is done in a continuous stirred-tank system (CSTR), where the typical tank volume is 4–6 times the hourly flow from the oil extraction to the production line. This provides a contact time of 4–6 hours, and is determined by a laboratory batchdegumming determination. A CSTR system is preferred for continuous production because this minimizes the tank volume required. The enzyme reaction was modeled as a first-order reaction, and by using normal design principles for reaction tanks, the size of the reaction vessels is calculated to ensure the correct contact time. Table 11.4, shows the volume increase in percentage for a system based on a single chamber tank or multiple chamber tanks. A CSTR setup with six compartments is the best configuration, giving a lowpercentage volume increase but not having an overcomplex tank design. Once the enzyme reaction is completed, the oil is heated to 70–80°C to inactivate the enzyme and to facilitate the aggregation of the gums and their removal by the centrifuge in the next step of the process. After the reaction, the phospholipids/gums are separated from the oil. As all the gums are now hydratable, they are eliminated with the water phase by the centrifugation. The viscosity of the obtained gum phase (consisting partly of lyso-phospholipids) is significantly lower than that containing nonhydrolyzed phospholipids. This eases the handling and pumping of the resulting gum phase in comparison to that obtained by water degumming By using such a system, a wide range of oil types is degummed with a very low yield loss, and the range of oils is extended to include oils not previously considered. Continuous centrifuges were developed for the separation of the oil from the gum/ water phase, and are capable of operating for long periods of time reliably although the manufacturers recommend preventive maintenance and occasional shut-downs. Following the centrifugal separation, the oil passes to bleaching and deodorization, where both silica and bleaching earth are used to remove any remaining phospholipid and adjust the color to the desired level. TABLE 11.4 Number of Tanks in CSTR System and the Volume Increase Required Number of tanks

Volume increase, %

1

208

2

69

3

41

4

29

5

22

6

18

230

D. Cowan and P.M. Nielsen

Testing Degumming Efficiency at the Laboratory Scale An evaluation of the usefulness of enzymatic degumming normally starts with a laboratory investigation of enzyme efficiency. In the laboratory, establishing a continuous process is difficult. However, Claussen (2001) described an assay system where the oil is recirculated through an in-line mixer to a one-liter reactor. One disadvantage of this system is the relatively intensive shear mixing applied as the mixer is running all the time during the reaction, which is not in accordance with a full-scale operation. Because of this, laboratory experiments are often carried out in a batch setup to simulate more precisely the situation in a plant. Nielsen and Husum (2003) demonstrated a laboratory system which was used in the development of a new phospholipase for enzymatic degumming. The principle is to treat the oil sample with a high-shear mixer (Ultra Turrax T25 mounted with dispersion tool S25N-18G from Janke & Kunkel, Germany) for a short period (90 seconds) after the addition of each of the ingredients. The emulsified sample is incubated with a propeller stirring device in a water bath at the reaction temperature between each episode of high-shear mixing and during the enzyme reaction. Another important aspect of degumming is the separation step. In a laboratory degumming using 2.5% total water and degumming rapeseed oil with 143 ppm of phosphorus, a test was made with different centrifuge times (0.5–20 minutes) (Nielsen & Husum 2003). The water content in the oil after centrifugation decreased from 0.21 to 0.10% when the centrifuge time was increased. The resulting phosphorous content was 12 and 2 ppm, respectively, after centrifugation at a short and a longer time. A linear model for the relationship between water and phosphorous content showed a highly significant correlation. This demonstrates the importance of the ratio of water content in the oil before and after centrifugation, and thus the efficiency of the centrifugation. Any laboratory model developed for testing degumming must be validated against plant experience to prevent the laboratory method from giving an overly favorable impression of the degumming efficiency. Two critical parameters are the centrifuge time and speed because laboratory centrifuges can exert a bigger centrifugal effect than that seen in continuous plant-scale centrifuges. If this is not allowed for then, the laboratory results will not be reflective of what is achieveable on the larger scale. Plant-Scale Enzymatic Degumming Published data on plant-scale operations are limited, but the available information confirms the laboratory findings that one can degum a wide range of oils at low yield losses (Cowan et al., 2005). Dayton (2005) described an enzymatic-degumming installation for soybean oil capable of running with either crude or water-degummed oil. Using Lecitase Ultra, with the use of meticulous mass-balance studies, the oil recovery was documented over an extended period. Table 11.5 summarizes the results that were obtained, and proves that yield loss was virtually eliminated by the enzymatic process.

231

Enzymatic Degumming of Edible Oils and Fats

TABLE 11.5 Results from the Plant-Scale Operation of Enzymatic Degumming (Dayton, 2005)  

Crude oil

Water-degummed oil

Caustic refining

Enzymatic refining

Caustic refining

Enzymatic refining

P level in oil

525 ppm

525 ppm

150 ppm

150 ppm

P level after centrifuge

2 ppm

2 ppm

2 ppm

2 ppm

Soapstock (%)

3.19

1.7

1.51

0.5

Refining loss (%)

3.08

1.57

1.42

0.45

Yield of oil (%)

96.6

97.8

98.3

99

 

Plant-scale degumming of rapeseed oil was carried out following laboratory determinations of the optimal processing conditions (Yang et al., 2006). Two different rapeseed-oil qualities with initial phosphorous levels of 212.4 and 120.5 ppm were used. The results of their optimization suggested an optimal pH of 4.9 and temperature of 48°C for degumming with an optimal dosage of 40 ppm of Lecitase Ultra. They also noted that higher dosages of enzyme reduced the time required to reach a particular phosphorous content in the oil. A plant-scale evaluation, using a 400 ton/ day plant, was established using two three-stage CSTR tanks as the enzyme-reaction vessels. The output phosphorous level was maintained at below 10 ppm as long as the operating pH was between 4.6 and 5.1. After bleaching, the residual phosphorous level was maintained at 3 ppm. In a similar setup, soybean oil was degummed by using the same pH, temperature, and enzyme dosage as used for the rapeseed trials (Yang et al., 2007). A reaction time of 6 hours in the six chambers of the CSTR was also used for this test. From a starting level of 121.5 ppm of phosphorus, they obtained less than 10 ppm at the exit of the reactor chain and 1–2 ppm following bleaching and deodorization. The resulting gums from the degumming process were more fluid than those from acid degumming, and the oil loss was significantly reduced. They concluded that the increased oil yield was due to the conversion of the gum phospholipid to not only lyso-phospholipid but also to glycerol-phospholipids which would be even less oil binding than the lyso form. The separation of the gums is carried out by continuous centrifugal separators, and Dayton et al. (2003) suggested that the addition of citric or other similar organic and mineral acids can improve the efficiency of this separation. The addition of 100–200 ppm of citric acid to the oil following the enzyme reaction was made to reduce the fouling of heat-exchange surfaces, the centrifuge plate stack and other associated equipment. The aim of this addition was to increase the time interval between stopping the process for the cleaning of surfaces fouled by the gums.

232

D. Cowan and P.M. Nielsen

Process Developments in Enzymatic Degumming The existing plant-scale operations described above are based on having relatively lowenzyme dosages and extending the reaction time for degumming. This is a reflection of the original Lurgi process, where enzyme recycling and extended reaction times were used to compensate for the high cost of the pancreatic phospholipase in use at that time (Cowan et al., 2006 ). One of the few disadvantages of this degumming method is that to ensure a complete enzyme reaction, a large reaction volume and contact time are required. From basic chemical-engineering principles, for a single tank operated as a CSTR, to ensure that all parts of the oil are converted, a volume increase of 208% is required compared to a batch-use operation. This translates for a 100-ton/hour line and a 6-hour contact time into a required tank volume of ≈1200 tons, which is not realistic. The normal answer would be to construct a CSTR in which case a tank of six compartments would only require a volume increase of 18%. Results presented in the literature demonstrate the differences found for different types and qualities of oils degummed in a six-stage CSTR system (Dahlke, 1998). The figures are from a 500-ton/day plant, and include water-degummed rapeseed oil, canola oil, and two different qualities of soybean oil (Fig. 11.9). Total mean holding time in the reactor system is 6 hours. Evidently, the full holding time of 6 hours is not required by all the oils if the target is to achieve 10 ppm of phosphorus after degumming. One of the soybean oils needed all 6 hours to be satisfactory, whereas the other soybean (already after the first stage) was below 10 ppm. This clearly reflects the difference between a good and a bad oil quality, which itself is an indication of how the original seeds or beans were stored and subsequently processed. Already the water-degummed crude oils indicated large differences in oil quality, as one of the soybean oils contained 65 ppm of phosphorus, and the other contained 180 ppm after water degumming. Recently, an increased enzyme dose was examined as an additional route for reducing treatment time and the associated size of reaction vessels. Degumming at the normal (30 ppm) and double dosing of Lecitase Ultra demonstrate that at the higher level, degumming is accomplished in a much shorter time (Fig. 11.10). One of the main benefits of enzymatic degumming is the increased oil yield, and in any short-time degumming process (STDG), one should not compromise this. Laboratory tests to measure both degumming efficiency and yield indicate that a successful STDG is achieved without a drop in oil yield. Reducing contact time to ≈90 minutes and still maintaining oil yields will be an important option for reducing the equipment footprint of enzymatic degumming. This can result in a required tank volume of approximately 25% of that needed for standard enzymatic degumming. Oil Recovery from Gums In caustic refining, water degumming is applied as a first-stage process to remove the hydratable phospholipid fraction, but the extracted gum normally contains equal amounts of entrained oil and gum as well as water. The quantity of oil lost is not the

Enzymatic Degumming of Edible Oils and Fats

233

Fig. 11.9. Reduction of phosphorus in a series of six CSTRs with different qualities of oil (Dahlke, 1998).

Fig. 11.10. Comparison of phosphorus removal and enzyme dose in the degumming of rapeseed oil.

234

D. Cowan and P.M. Nielsen

same as one can recover in full enzymatic degumming, but it is approximately 80% of this amount, as some oil is associated with the NHP left in the oil. When this gum is placed onto the protein meal, its value is considerably lower than refined oil which makes oil recovery of considerable economic interest. Enzymatic treatment of gums is a process by which one can recover oil from gums coming from water or acid degumming. In this process, gums from the water- or acid-degumming centrifuge are transferred to a treatment tank where the pH and temperature are adjusted to give optimal enzyme performance (pH 5.0–5.5 and 55°C). As in degumming, the phospholipid is hydrolyzed, becoming hydrophilic, and the gum can no longer hold the oil. A heating step is used to break the emulsion, and centrifugation separates the resulting oil. Oil yield as a function of incubation time and pH is shown in Fig. 11.11. Also, clearly, one must avoid overhydrolysis of the gums with extended contact time because it results in the formation of an overstable emulsion system from which the oil cannot be released. The progress of the gum hydrolysis was followed by using 31p-nuclear magnetic resonance (NMR) measurements, which can track the conversion of the individual phospholipids to their lyso form (Fig. 11.12). One can apply this process to any situation where a significant gum amount exists in the oil (e.g., soybean and rapeseed). The main advantage is that because the gum itself is processed, the treatment volumes are only 5–6% of the total flow through the unit, resulting in significant reductions in tank-size requirements. Where gums are

Fig. 11.11. Influence of pH and time on oil yield from soy gums (conditions: T = 55°C, 400 ppm of Lecitase Ultra on gum dry weight, internal data from Novozymes A/S).

Enzymatic Degumming of Edible Oils and Fats

235

Fig. 11.12. Conversion of a phospholipid to a lyso form in degumming (conditions: T = 55°C, 400 ppm of Lecitase Ultra on gum dry weight, internal data from Novozymes A/S).

extracted for the production of lecithin, the resulting gum from this process (deoiling) is a lyso-phospholipid, which has enhanced emulsification properties compared to the normal material.

Degumming of Exotic Oils Today, enzymatic degumming is established in many factories. The oils processed are soybean, rapeseed, sunflower, and Rice Bran oils. Beside these types of oils, corn and cottonseed oils were also successfully tested in the laboratory. The demand for vegetable oil is increasing, and the recent growth in Bio-Fuels sparked an interest in non-food oil crops, such as Jatropha and similar plants. Extraction trials with Jatropha oils were made by Rietzler and Brandt (2007), and they observed that to get economic yields, hot pressing was required which increased the phosphorous levels above that needed for biodiesel production (Fig. 11.13). One could use enzymatic degumming here, particularly the STDG process, to produce a well-degummed nonfood feedstock for biodiesel, and no doubt other similar oils will become available. Future Perspectives in Enzymatic Degumming For some time, work has been going on in the development of the enzymatic process.

236

D. Cowan and P.M. Nielsen

Fig. 11.13. Effect of pressing temperature on oil yield and phosphorous content.

The fact that the enzyme can work efficiently in a water content less than 0.5% suggests a variant of the degumming process shown in Fig. 11.8. The changes in the process are to lower the total water content to 0.5% and to substitute the centrifuge step with a silica adsorption. This saves water and reduces the oil loss. Also investment and energy costs are positively influenced by these changes. Successful trials with the new enzymatic degumming/adsorption process were made in a laboratory scale using water-degummed oil from soybean and rapeseed (Cowan et al., 2004). The results indicate that water-degummed oils are successfully processed to low phosphorous values by using a combination of enzyme treatment and silica absorption. Pilot-scale testing also showed that the filtration of the silica from the oil is achieved if the oil is first dried to remove entrained water. This adaptation of the process is particularly interesting where water is limited and in situations where the partially refined oil is brought to another location for final refining. Also a considerable development occurred in the range and types of phospholipases available for enzymatic degumming. The original Enzymax process used a pancreatic phospholipase which required enzyme recycling for good process economy, and in addition was not kosher- or halal-approved. Also, the required water addition was >5% to facilitate the removal of the gums and their partial recycling to the next batch of oil to reuse the enzyme. These limitations restricted the adoption of enzymatic degumming, and encouraged the development of more practical enzyme processes. The first microbial phospholipase was introduced in 2000, and a second-generation enzyme followed in 2002 which was more stable during storage and application than the other product. The characteristics of these enzymes are shown in Table 11.6. Recent studies with rice bran oil demonstrated that one can also use Lecitase Ultra successfully to degum this material. This provides a major benefit for the Rice

237

Enzymatic Degumming of Edible Oils and Fats

TABLE 11.6 Characterisitics of Different Phospholipases LecitaseTM 10L

Lecitase Novo

Lecitase Ultra

Source

Porcine pancreas

Microbial

Microbial

Specificity

A2

A1

A1

Molecular weight

12–14 kDa

≈28 kDa

≈35 kDa

2+

Ca -dependent

Yes

No

No

Td(DSC), °C

70–80

50

60

Oil-degumming Temp. optimum, °C pH optimum

65–70 5.5

45 4.5

55 4.5

Effective at 1% of total water

No

Yes

Yes

Kosher/Halal

No

Yes

Yes

Bran-oil-processing industry since this enzyme is more robust than the previous product (Lecitase Novo) used for this purpose. Other phospholipases are also under development to try and improve the existing products and their mode of action in degumming. The main focus is on the development of additional A1 and A2 phospholipases which one can use in the already existing enzymatic-degumming processes. At least two other products are commercially available at the current time. Lyso-phospholipases were also developed to further degrade lyso-phospholipids to glycero-phospholipids, which would be even more hydrophilic than the lyso form. These enzymes may play a role in oil recovery from gums since they would increase the separation of gum and oil by further reducing the ability of the gum to bind oil. In enzymatic degumming using the A1 or A2 phospholipase, a release also occurs of fatty acid coming from the hydrolysis of the phospholipid. Stochiometric calculations suggest that the release of 0.1% of FFAs should accompany a reduction in phosphorus of 100 ppm. On that basis, the degumming of crude oil with ≈750 ppm of phosphorus would result in an increased level of FFAs of 0.75%. This would need to be removed in the deodorizer, and would increase the load on this equipment. Laboratory- and plant-scale tests reveal that the amount of generated FFAs is lower than this because not all the phospholipid needs to be hydrolyzed to obtain the desired reduction in phosphorus. Table 11.7 shows the results from the degumming of canola oil. Under normal enzymatic-degumming conditions, the increase in FFA is limited, but attempts are being made to reduce this still further. Phospholipase C was suggested as a route to avoid FFA production because the mode of action of this enzyme (Fig. 11.4) would result in the release of the phosphate group and no generation of FFA.

238

D. Cowan and P.M. Nielsen

TABLE 11.7 FFA Production from the Degumming of Canola Oil Enzymatic degumming of canola oil Reduction in P (ppm)

175

Calculated increase in FFA

0.175%

Observed increase

0.08%

Unfortunately, the currently available form of phospholipase C is unable to degrade all the phospholipid types found in oil, and PA is not hydrolyzed. As this can be a significant part of the total phospholipid, total degumming with phospholipase C alone is not currently possible. One solution to this problem was proposed by Dayton (2008), who combined the A1 phospholipase with the C type to maximize the reduction in phosphorus and to minimize the generation of FFA. The three methods of degumming are compared in Table 11.8, and demonstrate how two phospholipases working in conjunction can provide even higher oil yields. Research is underway to develop phospholipases that have a broader action than the current phospholipase C, and these may offer even better degumming options. Another approach was to attempt enzymatic degumming by a completely different approach, using an acyl transferase. In theory, this enzyme would transfer one fatty acid from the phosphatide to an acyl acceptor, rendering the phospholipid water hydratable without the production of FFA. The acceptor molecule proposed by Hemmingsen et al. (2004) would be tocopherol or other sterols in the oil. Model studies using soybean oil—to which phosphatidylcholine was added—demonstrated that the phospholipid was removed, but no measurements of residual phosphorus were provided. In addition, the process requires sufficient sterol to be present, and results in the production of a sterol ester, which reduces the value of the tocopherol, which is normally recovered. However, the concept is interesting, and may well be developed further. TABLE 11.7. Degumming with Sodium Hydroxide and Phospholipasest

Caustic refining

PLA Enzymatic refining

PLA/PLC enzymatic refining

Starting Phos level in crude oil

500 ppm

500 ppm

500 ppm

Phos level after centrifuge

2 ppm

2 ppm

2 ppm

Centrifuge discharge (dry%)

3.19

1.13

0.62

Yield of oil (%)

96.5

97.4

98.3

Enzymatic Degumming of Edible Oils and Fats

% Contribution to Savings

Fig. 11.14. Environmental savings from enzymatic degumming compared to chemical refining.

Yield Loss

NaOH

Phosphoric Acid

Waste Treatment

Fig. 11.15. Contributing factors to reduce environmental impact. EDG, enzymatic degumming.

239

240

D. Cowan and P.M. Nielsen

Environmental Benefits of Enzymatic Degumming Compared to chemical refining with NaOH, enzymatic degumming gives higher yields due primarily to the avoidance of soapstock, and if practiced on crude oil, it eliminates losses in the lecithin fraction. As the enzyme process uses lower amounts of chemicals and runs with lower energy consumption, we can anticipate that it will have a reduced environmental impact. To look more closely at this, a Life Cycle Assessment of enzymatic and chemical (NaOH) refining was carried out. In this system a “cradle to grave” approach is used with all changes—from the growing of the soybeans to the final production of the oil being considered. All of the different inputs and outputs of the two production processes are considered in a side-by-side analysis to arrive at the possible savings that then the enzyme process can make (Nielsen, 2008). The environmental impact of the enzyme production is also included both in terms of the standard environmental indicators and also the direct consumption of agricultural raw materials. The end result of the analysis is shown in Fig. 11.14. In this analysis, two figures are calculated for each of the environmental indicators: One is the savings that are obtained by running an enzymatic compared to a chemical process, and the other is the used amount of these resources to produce the enzyme itself. In each case, the savings obtained are considerably higher than the amount of resources consumed (used) in producing and delivering the enzyme itself. The analysis also shows the areas of the degumming process which contribute most to the environmental savings. Figure 11.15 indicates where in the process the savings arise from and that yield increase plays the biggest part. Optimization of the yield increase can therefore be expected to give further reductions in environmental impact. Conclusions and Summary Enzymatic degumming has now become an accepted tool in the refining of vegetable oils, and is unique in that it is the only process where oil losses are minimized. Since the initial introduction of the technique, the process and the enzymes themselves were optimized, and the original extended incubation times and high enzyme costs are avoided. New enzyme concepts to extend the range of process options are being studied, and at least one of those was tried on an industrial scale. The future for this process—which reduces the environmental impact of oil refining—should be promising and should also be extended into the processing of oils for biodiesel as well as for normal food use.

References A Abhay, Sah; B.K.D. Agrawal; L.S. Shukla. J. Am. Oil Chem. Soc. 1983, 60, 465, 466. Achife, E.C.; J.A. Ibemesi. Ibid. 1989, 66, 247. Ackman, R.G. Chemical composition of rapeseed oil. High and Low Erucic Acid Rapeseed Oils: Production, Usage, Chemistry and Toxicological Evaluation; J.K.C. Kramer, F.D. Sauer, W.J. Pigden, Eds.; Academic Press: London, 1983. Allen, J.C.; R.J. Hamilton. Rancidity in Foods; Elsevier: London and New York, 1983; Chap: Measurement of Rancidity, J.B. Rossell. Andersen, A.J.C. Refining of Oils and Fats for Edible Purposes, 2nd ed.; Pergamon Press Ltd.: London, 1962; pp 107, 108, 112, 130. Anon. Eastman Chemical Products Inc. A Study of the Stabilization of Whole— Crude, Once Refined and Oxidized—Crude Palm Oil with TBHQ Antioxidant, Publ. No. ZG-215B (Feb. 1978). Anon. Industrial Filter and Pump Manufacturing Company, Cicero, Illinois, U.S.A. Anon. Palm Oil Bleachability Test, 3rd revision; (Leaflet) Seed Crushers and Oil Producers Association: London, 1986; also British Standard 1987, 684, 2, 27. Anon. Technical Handout, Aktiebolaget Pellerin/Zenith, Helsingbors, Sweden. Anon. Enzymatic degumming. Oils Fats Int. 1992, 6, 22–24. Anon. (a) Margarit, De Nordiske Fabriker, Oslo, 1913; (b) Arbeiten ausdem Pharmazeutischen Institut der Universität, Berlin, 1914. Anon. Industrial Processing of Cottonseed, compiled by Société pour l’étude et le developpement de l’industrie, de l’agriculture et du commerce (SEDIAC) Paris, for INTIP, UNIDO/IOD, 190/Rev. 1 (1978). Anon. Technical handout of Süd-Chemie AG: Munich, Highly Active Bleaching Earths “Tonsil,” p 18. Anon. Mizukalife F-1, Technical handout of Mizusawa Industrial Chemicals Ltd.: Tokyo. Anon. Galleon Earth, Technical handout Ibid. Anon. J. Am. Oil Chem. Soc. 1981, 58, 723A. Anon. Rice Bran: An Underutilized Raw Material; UNIDO publication, ID/320, sales no. E84, II.B.3, New York, 1985. Anon. The Rice Bran Oil Refining Technology, UNIDO ref. IO. 10. Vienna., 23 Nov. (1987), Compiled by U. Myint Pe. Anon. Trisyl® Silica for Edible Oil Refining http://www.gracedivision.com/products (1999). Anon. IUPAC. Manual of Symbols and Terminology for Physicochemical Quantities and Units, App II, Pt. I, 1971. Anon. Applications of Tonsil Bleaching Earths in the Food Industry http://sudchemie. com (2007). 241

242

G.R. List

Anon. Making Better Lard; American Meat Institute: Chicago. Anon. Technical Handouts of Stucliffe Speakman Carbons Ltd.: Leigh, U.K.: (a) Active Carbon Specifications; (b) Activated Carbon—Its Manufacture, Properties and Uses (Colin Moore); (c) Macroporosity in Activated Carbons: Its Origins and Functions (J. Wildman and F.J. Derbyshire); (d) Laboratory Test Methods for Active Carbons. Applewhite, T. (Ed.); Wiley Interscience: New York, 1985; Vol. 3, pp 273–309. Ariaansz, R.F.; C.B. Ungermann; D.R. Taylor. Bleaching of Fats and Oils: Effects on Oxidative Oil Properties in Fats for the Future II; Engelhard De Meern B.V.. 1989.

B Bailey’s Industrial Oil and Fat Products, 4th ed.; D. Swern, Ed.; John Wiley & Sons: New York, 1979; Vol. 2, pp 152, 746. Ibid. Vol. 1, pp 313, 315, 332, 361–362, 401, 506. Ibid. 1982; Vol. 2, p 254. Ibid. 3rd ed.; 1964; p 1018. Baldwin, A.R. J. Am. Oil Chem. Soc. 1949, 26, 610. Baldwin, A.R. (Ed.) Proceedings of the World Conference on Emerging Technologies in the Fats and Oils Industry, Cannes, Nov. 1985, AOCS. (a) Alsaker, J., New Rendering processes and their influence on tallow qualities; (b) Berge, J.L., Energy conservation in continuous rendering; (c) Swan, J.E., Low temperature rendering of animal materials and its application to reducing fat in edible meat. Bansal, R.C.; J.B. Donnet; F. Stoeckli. Active Carbon, Marcel Dekker Inc.: New York and Basle, 1988; pp 20, 23, Chap 3, Introduction. Barthel, G.; W. Grosch. J. Am. Oil Chem. Soc. 1974, 51, 540. Bates, R.W. Ibid. 1968, 45, 420A. Bek-Nielsen, B. Technical and Economic Aspects of the Oil Palm Fruit Processing Industry, United Nations Publication (UNIDO), ID/123, New York, 1974. Berger, K.G. Ibid. 1983, 60, 206. Bernardini, E. The New Oil and Fat Technology; Technologie s.r.l.: Rome, 1973. Berner, D.L. Color in Oils, AOCS 81st meeting, Chicago, May 1991. Bhattacharyya, D.K.; M.M. Chakrabaty; R.S. Vaidyanathan; A.C. Bhattacharyya. J. Am. Oil Chem. Soc. 1983, 60, 467. Bieri, J.G. Ibid. 1984, 61, 1917. Biernoth, G.; H.E. Rost. The occurrence of polycyclic aromatic hydrocarbons in coconut oil and their removal. Chem. Ind. 2002 (1967). Biernoth, G.; H.E. Rost. Arch. Hyg. Bakt 1968, 152, 238. Biernoth, G. Fette Seif. Ans. 1968, 70, 217. Bimbo, A.P. J. Am. Oil Chem. Soc. 1987, 64, 705. Bogdanor, J.M. Modified Bleaching; W.R. Grace & Co.: Baltimore, 1989.

Bleaching and Purifying of Fats and Oils, Second Edition

243

Bogdanor, J.M.; F. Griselli; F. Foschi. Studio sull utizzo di silice sintetica nel processo di raffinazione degli oli alimentari. Risultati e prospettive. La Rivista Italiana Delle Sostanze Grasse. Vol. LXVI, Jan. 1989. Bogdanor, J.M.; G.J. Toeneboehn. Silica Refining of Oils Containing Little or No Chlorophyll, technical presentation by W.R. Grace & Co.: Baltimore, MD, 1989. Boki, K.; M. Kubo; N. Kawasaki; H. Mori. Adsorption isotherms of pigments from alkali refined vegetable oils with clay minerals. J. Am Oil Chem. Soc. 1992, 69, 372–378. Boki, K.; H. Mori; N. Kawasaki. Bleaching rapeseed and soy bean oils with synthetic adsorbents and attapulgites. Ibid. 1994, 71, 595–601. Bolicer, Z.; Z. Leibovitz; C. Ruckstein. Fette Seif. Ans. 1984, 86, 537. Boukerroui, A.; M. Ouali. Edible oil bleaching with bentonite activated by microwave irradiation. Ann. Chim Sci. Mat. 2002, 27, 73–81. Braae, B.; V. Brimberg; M. Nyman. J. Am. Oil Chem. Soc. 1957, 34, 293. Brekke, O.L. Bleaching. Handbook of Soy Oil Processing and Utilization; Dr. Erickson et al., Eds.; ASA/AOCS: Champaign, IL, 1980; pp 72, 105–130. Brimberg, U.I. Fette Seif. Ans. 1981, 83, 184. Brimberg, U.I. J. Am. Oil Chem. Soc. 1982, 59, 74. Brooks, D.D.; S.A. Borphy; G.R. Goss. Oil Bleaching Method and Composition for Same. U.S. Patent 5,004,570 (1991). Brooks, D.D.; S.A. Borphy; G.R. Goss. Oil Bleaching Method and Composition for Same. U.S. Patent 5,151,211 (1992). Budowski, P. J. Am. Oil Chem. Soc. 1964, 41, 280. Burr, G.O.; M.M. Burr. J. Biol. Chem. 1929, 82, 345 and 1930, 86, 587.

C Campbell, E.T. J. Am. Oil Chem. Soc. 1983, 60, 387. Carlsson, D.J.; T. Suprunchuk; D.M. Wiles. Ibid. 1976, 53, 656. Carr, R.A. Rapeseed/canola. AOCS World Conference Edible Oils and Fats Processing— Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Champaign, IL. Cavanagh, G.C. Neutralization II: theory and practice of non-conventional caustic refining by miscella refining and by the Zenith Process. Ibid. Cavanagh, G.C. J. Am. Oil Chem. Soc. 1951, 28, 377. Chapman, D. The Structure of Lipids; Methuen & Co.: London, 1965. Chemistry of Carbon Compounds; E.H. Rodd, Ed.; Amsterdam, 1959; Vol. IVb. Cherry, J.P. J. Am. Oil Chem. Soc. 1983, 60, 360. Christidis, G.E.; P.W. Scott; A.C. Dunham. Acid activation and bleaching capacity of bentonites from the islands of Milos and Citios Aegean Sea, Greece. App Clay Sci. 1997, 12, 329–348.

244

G.R. List

Christopoulou, C.N.; E.G. Perkins. J. Am. Oil Chem. Soc. 1989, 66, 1338; Ibid. 1989, 66, 1344; Ibid. 1989, 66, 1353; and Ibid. 1989, 66, 1360. Clegg, A.J. Ibid. 1978, 50, 321. Cocks, L.B.; C. van Rede. Laboratory Handbook for Oils and Fats Analysts; Academic Press: New York, 1966. Coconut Processing Technology (7 parts); Asian and Pacific Coconut Community: Jakarta, Indonesia, UNIDO/IOD.377, Sept. (1980). Coenen, J.W.E. J. Am. Oil Chem. Soc. 1976, 53, 385. Coppen, P.P. The use of antioxidants. Rancidity in Foods; J.C. Allen, R.J. Hamilton, Eds.; Elsevier Applied Science Publishers: London, 1983.

D Davidsohn, J.; E.J. Better; A. Davidsohn. Soap Manufacture; Interscience: New York, 1953; Vol. 1, p 229. Davies, M.E.; P.R. Shanks; D.S. Anderson; R.S. Taylor. Investigations into the Measurement and Removal of Chlorophyll Derivatives from Edible Oils; Laporte Industries R and D: Widnes, U.K.; also in Edible Fats and Oils Processing: Basic Principles and Modern Practices; D.R. Erickson, Ed.; American Oil Chemists’ Society: Champaign, IL, 1989. DeGrey, T.W.; M. Kellens. Refining Practice in Edible Oil Processing; W. Hamm, R.J. Hamilton, Eds.; CRC Press: Sheffield, England, 2000; pp 79–128. Del Cerro, C.; D. Boey. Chem. Ind. 7 Nov. 1988, p 681. Denise, J. RFCG 1987, 34, 133, and Veldstra, J.; J. Klere. Sunflower seed oil. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Champaign, IL. Dijstra, A.J.; M. Van Opstal. The total degumming process. J. Am. Oil. Chem. Soc. 1989, 66, 1002–1009. Dijkstra, A.J.; M. Von Opstal. Process for Producing Degummed Vegetable Oils and Gums of High Phosphatidic Acid Content. U.S. Patent 4,698,185 (1987). Drozdowski, B. J. Am. Oil Chem. Soc. 1977, 54, 600. Duffy, M. Ibid. 1984, 61, 476.

E Eder, S.R. Fette Seif. Ans. 1982, 84, 186. Erickson, D.R. Bleaching/Adsorption Treatment in Practical Handbook of Soybean Processing and Utilization; AOCS Press/United Soybean Board: Champaign, IL and St. Louis, MO, 1995; pp 203–217. Erickson; D.R.; G. List. Bailey’s Industrial Oil and Fat Products, 5th ed.; T.H. Applewhite, Ed.; John Wiley & Sons: New York, 1985; Vol. 3, p 278, and List, G.R.; D.R. Erickson. Storage, stabilization and handling. Handbook of Soy Oil Processing and Utilization; American Soybean Assn. and American Oil Chemists’ Society: Champaign, IL, 1980; pp 271, 279–282.

Bleaching and Purifying of Fats and Oils, Second Edition

245

Erickson, D.R.; G.R. List. Storage, handling and stabilization of edible fats and oils. Bailey’s Industrial Oil and Fat Products; Erickson, D.R.; L.H. Wiedermann. Soybean oil modern processing and utilization. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989, AOCS Press, Champaign, IL. Erickson, D.R. J. Am. Oil Chem. Soc. 1983, 60, 35. Eriksson, C.E.; P.A. Olsson; S.G. Svensson. Ibid. 1971, 48, 442.

F Fahn, R.; K. Fenderl. Clay Minerals; 1983; Vol. 18, pp 447–458. FAO. Dietary Fats and Oils in Human Nutrition, FAO Food and Nutrition Series No. 20, FAO: Rome, 1980. Fedeli, E. J. Am. Oil Chem. Soc. 1983, 60, 404. Feron, R. Margarine; J.H. van Stuyvenberg, Ed.; Liverpool University Press: Liverpool, 1969; Chap 3. Feuge, R.O.; H.J. Janssen. J. Am. Oil Chem. Soc. 1951 28, 429. Forster, A.; A.J. Harper. Ibid. 1983, 60, 265. Frankel, E.N. Ibid. 1984, 61, 1908. Fuller, G.; G.O. Kohler; T.H. Applewhite. Ibid. 1966, 43, 477.

G Galliard, T. Rancidity in Foods; J.C. Allen, R.J. Hamilton, Eds.; Elsevier Applied Science Publishers: London, 1983. Gapor, A.; A.S.H. Ong. PORIM Bull. 1982, 5, 39. Gavin, A.M. J. Am. Oil Chem. Soc. 1981, 58, 176. Glasstone, S.; D. Lewis. Elements of Physical Chemistry; Macmillan: London, 1963; pp 558–567. Goebel, E.H. J. Am. Oil Chem. Soc. 1976, 53, 342. Gohl, B.O. Tropical Feeds; FAO Agricultural Studies No. 96, FAO: Rome, 1975. Graalmann, M. Laurics (coconut/palm kernel). AOCS World Conference on Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989. Gray, J.I. J. Am. Oil Chem. Soc. 1978, 55, 539. Grimmer, G.; A. Hildebrandt. Content of polycyclic hydrocarbons in crude vegetable oils. Chem. Ind. 2000 (1967). Grosch, W. Fette Seif. Ans. 1972, 74, 375. Gülsah Kirali, E.; O. Lacia. Statistical modeling of acid activation on cotton oil bleaching by Turkish bentonite. J. Food. Eng. 2006, 75, 137–141. Gunstone, F.D. J. Am. Oil Chem. Soc. 1984, 61, 441, 442.

246

G.R. List

H Hamilton, R.J. Rancidity in Foods; J.C. Allen, R.J. Hamilton, Eds.; Elsevier Applied Science Publishers: London, 1983; p 13. Hamilton, R.J.; J.B. Rossell. Analysis of Oils and Fats; Elsevier: London, 1986; Chap Classical analysis of oils and fats, J.B. Rossell. Haraldsson, G. J. Am. Oil Chem. Soc. 1983, 60, 251. Harris, R.V. Rice Bran Oil and Wax; UNIDO ID/WG. 89/8. Feb. 1971. Hasnuddin Siddiqui, M.K. Bleaching Earths; Pergamon Press: London, 1968. Hassler, J.W.; R.A. Hagberg. Oil and Soap 1939, 16, 188. Hastert, R.C. Effective Adsorptive Treatment of Edible Oils, AOCS Short Course, Phoenix, May 1988. Hebendanz, N.; W. Zschau. Impurities—How to Get Rid of Unwanted By-Products; World Conference on Oleochemicals, AOCS, Kuala Lumpur, Oct. 1990 and AOCS meeting, Chicago, May 1991. Henderson, J. H. A laboratory study of the press effect in adsorptive bleaching. J. Am. Oil Chem. Soc. 1993, 70, 831. Hernandez, E.; S.J. Rathbone. Refining of Glyceride Oils by Treatment with Silicate Solutions and Filtration. U.S. Patent 6,448,423 (2002) (Silicate Refining). Hilditch, T.P.; P.N. Williams. The Chemical Constitution of the Natural Fats, 4th ed.; John Wiley and Sons: New York, 1964. Hill, S.E. Effectiveness of Bleaching Earths on Expelled Oil: A Comparative Study; 81st AGM; AOCS Press: Champaign, IL, 1990. Hiscocks, E.S.; W.D. Raymond. J. Am. Oil Chem. Soc. 1964, 41, 224. Hodgson, A.S. Refining and bleaching. Bailey’s Industrial Oil and Fat Products, 5th ed.; Wiley Interscience: New York, 1996; pp 157–212. Hoffman, Y. J. Am. Oil Chem. Soc. 1973, 50, 260. Hoffman, G.; T. Nijzink; J.H. Recourt. Rev. Franc. Corps Gras. 1975, 22, 511. Holm, U.; K. Ekbom-Olsson. Anisidine Value and Oil Quality, Abstracts ISF Congress, Göteborg, Sweden, 1972. Houghton, F.R.; J. Wildman. Manufacture and Uses of Activated Carbon; Chemical and Process Engineering, May 1971. Huang, J.Y.; Lui yue Lui; X. Wang. Effect of Attapulgite pore size distribution on soy bean oil bleaching. J. Am. Oil Chem. Soc. 2007, 84, 687–692. Hudson, B.J.F.; M. Ghavani. Lebensm. Wiss. u. Technol. 1984, 17, 191. Hudson, B.J.F.; J.I. Lewis. Food Chem. 1983, 10, 47. Hutchins, R.P. J. Am. Oil Chem. Soc. 1976, 53, 279. Hvolby, A. Ibid. 1971, 48, 503.

Bleaching and Purifying of Fats and Oils, Second Edition

247

I Industrial Processing of Cottonseed, compiled by Société pour l’etude et le developpement de l’industrie, de l’agriculture et du commerce (SEDIAC) (Paris) for INTIP, UNIDO/IOD, 190/Rev. 1 (1978).

J Jacobsberg, B.; D. Jacmain. Oléagineux. 1973, 28, 25. James, E.M. J. Am. Oil Chem. Soc. 1958, 35, 76. Janson, H. Castor Oil Production and Processing; United Nations Publications (UNIDO), 1D/125: New York, 1974. Jebe, T.A.; M.G. Matlock; R.T. Sleeter. Collaborative study of the oil stability index analysis. J. Am. Oil Chem. Soc. 1993, 70, 1055–1061. Johansson, G. Chem. Ind. 1 Nov. (1975): 902. Jones, L.A. Understanding cottonseed oil. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989, AOCS Press: Champaign, IL. Jones, L. Clonal Oil Palm; Unilever Information: London.

K Kaimal, T.N.B.; G. Lakshminarayana. J. Am. Oil Chem. Soc. 1979, 56, 578. Kaufmann, H.P. Analyse der Fette und Fett Produckte; Springer, Berlin, 1958; p 763. Kaufmann, H.P.; J.G. Thieme; F. Volbert. Fette Seif. Ans. 1956, 58, 505, 995, 1056. Kaynak, G.; M. Ersoz; H. Kara. Investigation of the properties of oil at the bleaching unit of an oil refinery. J. Colloid and Interface Sci. 2004, 280, 131–138. Kenar, J.A. Reaction chemistry of gossypol and its derivatives. J. Am. Oil Chem. Soc. 2006, 83, 269–302. Keppler, J.G.; M.M. Horikx; P.W. Meijboom; W.H. Feenstra. Ibid. 1967, 44, 543. Keppler, J.G.; J.A. Schols; W.J. Feenstra; P.W. Meijboom. Ibid. 1965, 42, 246. Khoo, L.E.; F. Morsingh; K.Y. Liew. Ibid. 1979, 56, 672. Kikugawa, K.; M. Arai; T. Kurechi. Ibid. 1983, 60, 1528. King, J.W.; G.R. List; J.H. Johnson. Supercritical extraction of spent bleaching clays. J. Supercritical Fluids 1992, 5, 38–41. King, R.R.; F.W. Wharton. J. Am. Oil Chem. Soc. 1949, 26, 201, 389. Klagge, P.; A.K. Sen Gupta. Fat Sci. Technol. 1990, 92, 315. Knowles, P.F. J. Am. Oil Chem. Soc. 1968, 46, 130. Kock, M. Fette Seif. Ans. 1981, 83, 552. Kock, M. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn. Koenig, H.R. Small Capacity Rice Bran Oil Extraction Plants, Fifth International

248

G.R. List

Congress of Food Science and Technology, Kyoto, Sep (1978), UNIDO: Vienna. Koseoglu, S.S. Membrane Degumming, Refining and Bleaching of Crude Vegetable Oils, AOCS AGM: Urbana , IL, 1987. Koseoglu, S.S.; D.E. Engelgou. J. Am. Oil Chem. Soc. 1990, 67, 239. Koseoglu, S.; W.E. Fark; W. DeGreyt; M. Kellens. Degumming of Edible Oils by Ultrafiltration. U.S. Patent 6,797,172 (2004). Koseoglu, S.S.; J.T. Lawhon; E.W. Lusas. J. Am. Oil Chem. Soc. 1990, 67, 315. Koseoglu, S.S.; K.C. Rhee; E.W. Lusas. Membrane processing of crude vegetable oils: laboratory-scale membrane degumming, refining and bleaching. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Krishnamurthy, R.G. Cooking oils, salad oils and salad dressings. Bailey’s Industrial Oil and Fat Products, 4th ed.; D. Swern, Ed.; John Wiley & Sons: New York, 1982; Vol. 2, p 315. Krishnan, S. J. Am. Oil Chem. Soc. 1975, 52, 23. Kühnau, J. Developpments recents en chimie et biochimie des acides grasses essentiels. Margarine Today; E.J. Brill, Ed.; International Federation of Margarine Associations: Leiden, 1970.

L Landon, J.W. Chem. Ind. 1 Nov. (1975): 892. Larsson, B.K. Polycyclic Aromatic Hydrocarbons in Swedish Foods; Swedish University of Agricultural Sciences: Uppsala, 1986. Larsson, B.K. Z. Lebens. Unters. Forsch. 1982, 174, 101. Lasztity, R. Rice Bran Wax Recovery and Utilization; UNIDO ID/WG 89/10. March 1971. Latondress, E.G. J. Am. Oil Chem. Soc. 1983, 60, 257. Latta, S. INFORM 1990, 1, 239. Lau, F.Y.; E.G. Hammond; P.F. Ross. J. Am. Oil Chem. Soc. 1982, 59, 407. Law, K.S.; T. Thiagarajan. Palm oil—edible oil of tomorrow. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Law Kia Sang. Oléagineux 1984, 39, 89. Lee, E.C. J. Am. Oil Chem. Soc. 1987, 64, 635. Leibovitz, Z.; C. Ruckenstein. Ibid. 1983, 60, 395. Leong, W.L.; K.G. Berger. Storage, Handling and Transportation of Palm Oil Products; PORIM Tech. Series No. 7 (1982), revised Sept. 1986. Leysen, R. Edible soybean oil quality in Western Europe. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Lin, L.; S.S. Koseoglu. Membrane processing of fats and oils. Bailey’s Industrial Oil and

Bleaching and Purifying of Fats and Oils, Second Edition

249

Fat Products, 6th ed.; Fereidoon Shahidi, Ed.; Wiley Interscience: New York, 2005; Vol. 5, pp 433–458. Lin, L.; K.C. Rhee; S.S. Koseoglu. Bench scale membrane degumming of crude vegetable oil. J. Memr. Sci. 1997, 134, 101–108. List, G.R. Anisidine Test (1974). List, G.R.; C.D. Evans; W.F. Kwolsk; K. Warner; B.K. Boundy. J. Am. Oil Chem. Soc. 1974, 51, 17. Loncin, M. Ibid. 1975, 52, 144A.

M Macellan, M. Ibid. 1983, 60, 368, also in Loncin, N.; B. Jacobsberg; G. Evrard. Rev. Fr. Corps Gras 1971, 18, 69. Macmillan, D. J. Am. Oil Chem. Soc. 1949, 26, 615. Mag, T.K. Process of Converting Rendered Triglycerides Oil from Marine Sources into Bland, Stable Oil. U.S. Patent 7,149,491 (2007). Mag, T.K. U.S. Patent 4,240,972 (1980). Mag, T.K. Bleaching—theory and practice. Proceedings of World Conference on Edible Fats and Oils Processing; D.R. Erickson, Ed.; AOCS Press: Champaign, IL, 1990; pp 107–116. Mag, T.K. Bleaching—theory and practice. AOCS World Conference on Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989; AOCS Press: Urbana, IL. Mag, K.T. J. Am. Oil Chem. Soc. 1983, 60, 380. McGinley, L. Quality control for processing and processed fats. Analysis of Oilseeds, Fats and Fatty Foods; J.B. Rossell, J.L.R. Pritchard, Eds.; Elsevier: London and New York, 1991. Meara, M.L.; G.S.D. Weir. Rev. Ital. Sostanze Grasse 1976, 53, 178. Meares, P. The Chemical Engineer; Feb. 1986, pp 38–40. Mecklenberg, W. Angew. Chem. 1924, 37, 873, and Andersen, A.J.C. Refining of Oils and Fats for Edible Purposes, 2nd ed.; Pergamon Press Ltd.: London, 1962; p 108. Mehlenbacher, V.C. The Analysis of Fats and Oils; Garrard Press: Champaign, IL, 1960. Mendoza, J.A.; M. Gomez; A. Casado. Technological evolution of the different process for olive oil extraction. AOCS World Conference on Edible Oils and Fats Processing— Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Mitchell, J.H.; H.R. Kraybill. J. Am. Chem. Soc. 1942, 64, 988. Moore, E.B. J. Am. Oil Chem. Soc. 1950, 27, 75. Morgan, D.A.; D.B. Shaw; M.J. Sidebottom; T.C. Soon; R.S. Taylor. Ibid. 1985, 62, 292. Morton, R.B.; F. Griselli. Stability and Quality Aspects in Edible Oil Refining with Thiyl Silica, Technical Presentation by Grace GmbH, Worms, Germany, at AOCS

250

G.R. List

Annual Meeting, Baltimore, MD, 1990. Mounts, T.L. J. Am. Oil Chem. Soc. 1987, 64, 686.

N Nebergall, R.S.; D.R. Taylor; C.J. Kucharz. Process for Regenerating Spent Acid Activated Bentonite Clays and Smectite Catalysts. U.S. Patent 5,468,701 (1995). Neumann, T.E.; N.T. Dunford. Edible Oil. Bleaching in Nutritionally Enhanced Edible Oil and Oil Seed Processing; N.T. Dunford, H.B. Dunford, Eds.; AOCS Press: Champaign, IL, 2004; pp 148–160. Ney, K.H. Fette Seif. Ans. 1964, 66, 512. Nilsson-Johansson, L.; U.I. Brimberg; G. Haraldsson. Fat Sci. Technol. 1988, 90, 447. Normann, W. German Patent 139457 (1902). Normann, W. British Patent 1515 (1903). Norris, F.A. Bailey’s Industrial Oil and Fat Products, 4th edn.; D. Swern, Ed.; John Wiley & Sons: New York, 1982; Vol. 2, pp 295, 306–308. Norris, F.A. Refining and bleaching. Ibid. 3rd ed.; 1964; pp 719–792. Norris, F.A. Refining and bleaching. Ibid.; pp 292–314. Norris, F.A. Ibid. 1979; Vol. 2, pp 256, 261. Norris, F.A. NSPA quality standards for soybean oil and meal. Proceedings of Quality Control Seminar in Soybean Crushing Plants and Soybean Oil Processing Plants; American Soybean Assn., Brussels, 1979.

O O’Connor, R.T.; E.T. Field; M.E. Jefferson; F.G. Dollear. J. Am. Oil Chem. Soc. 1949, 26, 710. Ohlson, J.S.R. J. Am. Oil Chem. Soc. 1983, 60, 385. Omar, S.; B. Girgis; F. Taha. Carbonaceous materials from seed hulls for bleaching of vegetable oils. Food Res. Inst. 2003, 36, 11–17. Ong, J.T.L. Fette Seif. Ans. 1980, 82, 169. Ong, J.T.L. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn. Opstvedt, J.; N. Urdahl; J. Pettersen. Fish oils—an old fat source with new possibilities. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press, Urbana, IL. Ory, R.L.; G.F. Flick, Jr. Peanut oil—chemistry and properties. Ibid. Ostreyko, R.V. German Patent 136, 792 (1902).

P P and S Filtration Ltd., Haslingden, U.K. Pardun, H.; E. Kroll; O. Werber. Fette Seif. Ans. 1968, 70, 531 and II Ibid. 1968, 70,

Bleaching and Purifying of Fats and Oils, Second Edition

251

643. Parker, W.A.; D. Melnick. J. Am. Oil Chem. Soc. 1966, 43, 635. Patterson, H.B.W. Filtration and Separation 1973, 1, 68. Patterson, H.B.W. Handling and Storage of Oilseeds, Oils, Fats and Meal; Elsevier Science; Essex, 1989; pp 10–35, 76–77, 86, 99, 109, 194, 320, 349, Chap 2, Chap 7. Patterson, H.B.W. Hydrogenation of Fats and Oils; Elsevier Applied Science: London and New York, 1983; pp 48, 91, 117, 204. Patterson, H.B.W. Ibid. p 43, and List, G.R.; D.R. Erickson. Storage, Stabilization and Handling, Handbook of Soy Oil Processing and Utilization ; American Soybean Assn.; AOCS Press: Urbana, IL, 1980. Patterson, H.B.W. (a) Chem. Ind., 18 Sep 1976, p 771; (b) Fette Seif. Ans. 1975, 77, 330. Patterson, H.B.W. J. Am. Oil Chem. Soc. 1976, 53, 339. Paulitz, B.; J. Segers; A. Spits. Process Relating to Triglyceride Oils. U.S. Patent 4,584,141 (1986). Paulose, M.M.; K.D. Mukherjee; I. Richter. Chem. Phys. Lipids 1978, 21, 187. Penk, G. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn., Part 2. Penk, G. Fette Seif. Ans. 1985, 87, 499. Penk, G. Ibid. 1981, 83, 558. Pfannkoch, E.A.; P.J. Gill. Identification and Characterization of Chlorophyll Derivates in Edible Oils by Reversed Phase HPLC, Presentation by W.R. Grace & Co.: Baltimore, MD, 1990. Player, R.B.; R. Wood. J. Assoc. Public Analysts 1980, 18, 77. Pritchard, J.R. J. Am. Oil Chem. Soc.1983, 60, 322. Proctor, A.; D.D. Brooks. Adsorptive separation of oils. Bailey’s Industrial Oils and Fat Products, 6th rev. ed.; F. Shahidi, Ed.; John Wiley & Son: New York, 1979; Vol. 5 pp 267–284. Proctor, A.; J.K. Toro-Vasquez. The Freundlich isotherm in studying adsorption in oil processing. J. Am. Oil Chem. Soc. 1996, 73, 1627–1633. Purchas, D.B. Filter Aids, Chemical and Process Engineering, June 1967, p 95.

R Rade, D.; D. Strucell; Z. Mokrovcak. Effect of soybean pretreatment on the phospholipid content in crude and degummed oils. Fat Sci. Technol. 1995, 97, 501–507. Reddi, P.B.V.; K.S. Murti; R.O. Feuge. J. Am. Oil Chem. Soc. 1948, 25, 206. Rich, A.D. Ibid. 1964, 41, 315 ; Ibid. 1967, 44, 298A; Ibid. 1970, 47, 564A; Ibid. 1954, 31, 374. Richardson, L.L. Ibid. 1978, 55, 777. Ringers, H.J.; J. Segers. Degumming Process for Triglyceride Oils. U.S. Patent

252

G.R. List

4,049,686 (1977). Robbelen, G. Chem. Ind. 7 Oct 1991, p 713. Rossell, J.B. Vegetable Oils and Fats in Analysis of Oilseeds, Fats and Fatty Foods; J.B. Rossell, J.L.R. Pritchard, Eds.; Elsevier: London and New York, 1991. Rossell, J.B. Rancidity in Foods; J.C. Allen, R.J. Hamilton, Eds.; Elsevier Applied Science Publishers: London, 1983; pp 32, 41, and Hudson, B.J.F. Ibid., p 54. Rossell, J.B.; S.P. Kochhar; I.M. Jawad. Chemical Changes in Soy Oil During High Temperature Processing; Proceedings Second ASA Symposium, Antwerp, June 1981, American Soybean Assn. Rossell, J.B.; B. King; M.J. Downes. J. Am. Oil Chem. Soc. 1985, 62, 221. Rossi, M.; M. Gianazza; C. Alamprese; F. Stanga. The effect of bleaching and physical refining on color and minor components of palm oil. Ibid. 2001, 78, 1051–1055. Rossi, M.; M. Gianazza; A. Alamprese; F. Stanga. The role of bleaching clays and synthetic silica in palm oil physical refining. Food Chem. 2003, 82, 291–296. Rost, H.E. Chem. Ind. 1976, 17, 612 (July).

S Sagredos, A.N.; D. Sinha-Roy; A. Thomas. Fat. Sci. Tech. 1988, 90, 76. Sanchez-Martin, M.J.; M.S. Rodriguez-Cruz; M.S. Andrades; M. Sanchez-Camazano. Efficiency of different clay minerals modified with a cationic surfactant in the adsorption of pesticides: influence of clay type and pesticide hydrophobicity. App Clay Sci. 2006, 31, 216–228. Schenk Filterbau Gmbh, D7076 Waldstetten, Germany and 414 Ridge View Court, Arnold, MD., U.S.A. Sebedio, J.L.; M.F. Langman; C.A. Eaton; R.G. Ackman. J. Am. Oil Chem. Soc. 1981, 58, 41. Segers, J.C.; R.L.K.M. van de Sande. Physical refining, AOCS World Conference Edible Oils and Fats Processing--Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS: Urbana, IL. Segers, J.C. J. Am. Oil Chem. Soc. 1983, 60, 262. Segers, J.C. Fette Seif. Ans. 1985, 87, 541. Segers, J.C. Ibid. 1982, 84, 543. Segers, J. Oil Purification by Adding Hydratable Phosphatides. U.S. Patent 4,162,260. Sen Gupta, A.K. Fette Seif. Ans. 1986, 88, 79; Ibid. 1988, 90, 251; Ibid. 1974, 76, 440. Sherwin, E.R. J. Am. Oil Chem. Soc. 1976, 53, 430. Shukla, V.K.S. Confectionery fats. AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Siew, W.L.; Y.A. Tan; T.S. Tang. Silica refining of palm oil. J. Am. Oil Chem. Soc. 1994, 71, 1012–1016.

Bleaching and Purifying of Fats and Oils, Second Edition

253

Singleton, W.A.; C.E. McMichael. Ibid. 1955, 32, 1; Ibid. 1956, 33, 477. Sinram, R.D. Nephelometric determination of phosphorus in soybean and corn oil processing. Ibid. 1986, 63, 667–670. Sleeter, R.T. Instrumental analysis for quality control and quality assurances. Bailey’s Industrial Oil and Fat Products, 1985; Vol. 3, p 182. Smith, J.R. J. Am. Oil Chem. Soc. 1985, 62, 1286. Sonntag, N.O.V. Composition and characteristics of individual fats and oils. Bailey’s Industrial Oil and Fat Products, 4th ed.; D. Swern, Ed.; John Wiley & Sons: New York, 1979; Vol. 1, pp 34, 387, 453. Sonntag, N.O. Structure and composition of fats and oils. Ibid. pp 1–98. Sonntag, N.O.V. Analytical methods. Ibid. 1982; Vol. 2, p 407. Stage, H. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn. Stout, L.E.; D.F. Chamberlain; J.M. McKelvey. J. Am. Oil Chem. Soc. 1949, 26, 120. Strecker, L.R.; A. Maza; G. Wennie. Corn oil—composition, processing and utilization. AOCS World Conference on Edible Oils and Fats Processing—Basic Principles and Modern Practice, Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Sullivan, F.E. J. Am. Oil Chem. Soc. 1976, 53, 358; Ibid. 1980, 57, 845A. Swoboda, P.A.T. Ibid. 1985, 62, 287.

T Taufel, K.; U. Franzke; G. Heder. Fette Seif. Ans. 1959, 61, 1225. Taylor, R.J. Essential Fatty Acids—A Review; Unilever Information Division: London, 1968. Taylor, D.L. Bleaching. Bailey’s Industrial Oil and Fat Products, 6th ed.; Fereidoon Shahidi, Ed.; Wiley Interscience: New York, 2005; Vol. 5, pp 285–340. Taylor, D.L. Cooking oils, salad oils and salad dressings. Ibid. R.G. Krishnamurthy, Ed.; pp 315–342. Taylor, D.L. Structure and composition of fats and oils. Ibid. 4th ed.; Sonntag, Ed.; 1979; Vol. 1, pp 1–98. Taylor, D.L. Composition and characteristics of individual fats and oils. Ibid. D. Swern, Ed.; pp 289–478. Taylor , D.L. Refining and bleaching. Ibid. 1982; Vol. 2, pp 253–314. Taylor, R.J. The Chemistry of Glycerides; Unilever Educational Booklet, Advanced Series No. 4, Unilever Information Division: Lon Taylor, D.R.; D.B. Jenkins. Acid-Activated Clays, Trans. Soc. Mining Engineers AIME; Vol. 282, 1901/10. Taylor, D.R.; D.B. Jenkins; C.B. Ungermann. J. Am. Oil Chem. Soc. 1989, 66, 334. Taylor, D.R.; C.B. Ungermann. Effective Adsorptive Bleaching of Oil in Practical Short Course in Soybean Extraction and Oil Processing Manual; L.R. Watsins, E.W. Lucas, S.S. Koseoglu, Eds.; Texas A & M (1987), also from Engelhard De Meern B.V.: Netherlands.

254

G.R. List

Taylor, D.R.; C.B. Ungermann. Process for Making Acid Activated Bleaching Earth Using High Susceptibility Source Clay and Novel Bleaching Earth Product. U.S. Patent 5,008,226 (1991). Taylor, D.R.; C.B. Ungermann. Process for Making Acid Activated Bleaching Earth Using High Susceptibility Source Clay and Novel Bleaching Earth Product. U.S. Patent 5,008,226 (1991). Taylor, D.R.; C.B. Ungermann. A Chlorophyll-Selective Bleaching Clay: Contrast in Chlorophyll Removal from Soya, Sunflower and Canola Oil, Engelhard Corp Technical Handout; De Meern: Netherlands. Presented at AOCS Annual Meeting, Dallas, TX, 1984. Taylor, D.R.; C.B. Ungermann. Laboratory Bleaching System; Technical Handout, Engelhard Corp, Menlo Park, New Jersey, U.S.A. and De Meern, The Netherlands. Teasdale, B.F. Canada Packers, Private communication, 1981. Technical Handout, Süd-Chemie AG, Highly Active Bleaching Earths “Tonsil”, Munich. Technical Handout, Norit Co., Amersfoort (NL), Glasgow (UK) and Jacksonville (USA), Introduction to Activated Carbon. Technical Handout, De Smet USA Corporation: Atlanta, USA and n.v. Extraction De Smet, s.a., B. 2520 Edegem, Antwerp, Belgium. The Merck Index—An Encyclopedia of Chemicals, Drugs and Biologicals; 11th ed.; S. Budavari, Ed.; Merck and Co. Inc.: Rahway, 1989. The Rice Bran Oil Refining Technology; U Myint Pe. UNIDO/IO.IO: Vienna, 1987. Thiagarajan, T. Developments in the Malaysian palm oil refining industry. AOCS World Conference on Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Urbana, IL. Thomas, A.E.; F.R. Paulika. Solvent Fractionated Fats (Glidden-Durkee, S.C.M. International) Symposium on Recent Advances in Chemistry and Technology of Oils and Fats, Liverpool Polytechnic, 1975. Thomas, A. Fette Seif. Ans. 1982, 84, 133. Thomas, A.E. Bailey’s Industrial Oil and Fat Products, 5th edn.; T.H. Applewhite, Ed.; John Wiley & Sons: New York, 1985; Vol. 3, p 12. Thomas, A. J. Am. Oil Chem. Soc. 1982, 59, 1. Thomas, A. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn. Toeneboehn, G.J.; W.A. Welsh. Packed-Bed Bleaching of Canola and Rapeseed Oils, 81st AOCS meeting, Baltimore, 1990. Toeneboehn, G.; W. Cheek; W.A. Walsh; J.M. Boganor. MPR Process for Treating Glyceride Oils, Fatty Chemicals and Wax Esters. U.S. Patent 5,298,639 (1994). Toeneboehn, G.; W. Cheek; W.A. Walsh; J.M. Boganor. MPR Process for Treating Glyceride Oils, Fatty Chemicals and Wax Esters. U.S. Patent 5,298,639 (1994). Topallar, H. Bleaching kinetics of sunflower seed oil. J. Am. Oil Chem. Soc. 1998, 75, 531–533.

Bleaching and Purifying of Fats and Oils, Second Edition

255

Tourkow, K.; T. Siemieniewska; F. Czeckowski. Fuel 1977, 56, 121, and Bansal, R.C.; J.B. Donnet; F. Stoeckli. Active Carbon; Marcel Dekker Inc.: New York and Basle, 1988; p 19. Transfeld, P. Electro filtration cuts bleaching earth USF. Inform 1998, 9, 502–506.

U Usuki, R.; T. Suzuki; Y. Endo; T. Kaneda. J. Am. Oil Chem. Soc. 1984, 61, 785.

V Veldkamp, F.G. Selecting the Right Filter; Technical handout (1991), Lochem B.V. 7240 A.A. postbus 35, Holland. Veldstra, J.; J. Klere. Sunflower seed oil. AOCS World Conference Edible Oils and Fats Processing Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS Press: Urbana, IL.

W Waldmann, C.; R. Eggers. Deoiling contaminated bleaching clay by high-pressure extraction. J. Am. Oil Chem. Soc. 1991, 68, 922–930. Waters, W.A. Ibid. 1971, 48, 427. Watson, K.S.; C.H. Meierhoefer. Ibid. 1976, 53, 437. Welsh, W.A.; J.M. Bogdanor; G.J. Toeneboehn. Silica refining of oils and fats. Edible Fats and Oil Processing: Basic Principles and Modern Practices; D.R. Erickson, Ed.; American Oil Chemists’ Society: Champaign, IL, 1989. Wendt, H.H.R.H. Fette Seif. Ans. 1981, 83, 541. Wiedermann, L.H. Degumming, refining and bleaching soy bean oil. J. Am. Oil Chem. Soc. 1981, 58, 159–166. Willems, M.G.A.; F.B. Padley. Ibid. 1985, 62, 454. Woollatt, E. Lite Manufacture of Soaps, Other Detergents and Glycerine; Horwood: Chichester, 1985.

Y Yokochi, K. Rice Bran Processing for the Production of Rice Bran Oil and Rice Bran Protein Meal; UNIDO ID/WG/120/9: New York, 1979; pp 47–50. Young, F.V.K. Physical refining, AOCS World Conference Edible Oils and Fats Processing—Basic Principles and Modern Practice; Maastricht, Oct. 1989, AOCS: Urbana, IL. Young, F.V.K. The Refining and Hydrogenation of Fish Oil; Bulletin No. 17 (1986), Technical Handout from International Assn. of Fish Meal Manufacturers, Potters Bar: U.K., 1986. Young, F.V.K. J. Am. Oil Chem. Soc. 1983, 60, 374.

256

G.R. List

Young, F.V.K. Chem. Ind. 16 Sep 1978, p 692. Young, F.V.K. Caustic refining, trace impurities, refining methods and edible oil quality. Proceedings Second ASA Symposium; Antwerp, June 1981, American Soybean Assn.

Z Zschau, W. Bleaching: theory and practice. Proceedings World Conference on Oilseed and Edible Oil Processing; S.S. Koseoglu, K.C. Rhee, R.F. Wilson, Eds.; AOCS Press: Champaign, IL, 1998; Vol. I, pp 64–76. Zschau, W. Bleaching. In Introduction to Fats and Oils Technology, 2nd ed.; R.D. O’Brien, W. Farr, P.J. Wan, Eds.; AOCS Press: Champaign, IL, 2000; pp 158–178. Zschau, W. Fette Seif. Ans. 1985, 87, 506; Ibid. 1982, 84, 493. Zschau, W. INFORM 1990, 7, 638. Zschau, W. Versuche zur Reinigung und Bleichung von Palmöl als Verstufe zur physikalischen Raffination unter Verwendung von To Opt FF und Tonsil P., DGF– Tagung, Wurzburg, Sep 1977. Zucker, F.J. Fette Seif. Ans. 1968, 70, 529.

INDEX

Index Terms

Links

All page numbers followed by the letter “f” refer to figures, and all page numbers followed by the letter “t” refer to tables.

A acetone solution

52

58

206

activation procedures

70

82

104

141t

201

204

53

97

199

208

111

117

25

35

117

Active Oxygen Method (AOM) measures adsorption

209 aflatoxins AgriTecSorbents, LLC Alfa Laval

108 88 156

alkali refining chlorophylls

7

carotenoids

8

degumming

5

dimers

18

fats and oils corn oil

115

116

cottonseed oil

106

108

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

alkali refining (Cont.) grapeseed oil

133

groundnut oil

111

hydrogenated oils

149

linseed oil

113

marine oils

148

olive oil

117

palm oil

122

rapeseed oil

126

rice bran oil

132

safflower oil

134

sesame oil

134

136

soybean oil

138

140

vegetable butters

112

flavines

9

gossypols

10

heat bleaching

48

metals and

129

16f

moisture and

65

oxidized fatty groups

19

phosphatides

13

sterols

14

tocopherols

11

22

AMCOL International Corporation of USA AMC (UK) Ltd.

73 73

American Oil Chemists’ Society (ACOS)

200

201

202

This page has been reformatted by Knovel to provide easier navigation.

Index Terms amorphous silica anchovy oil

Links 42

47

91

164

148

anisidine value (AnV) clay adsorption

36

dosage and adsorption

62

fats and oils groundnut oil

111

marine oils

147

palm kernel oil

124

palm oil

120

rapeseed oil

129

sunflower oil

143

oxidized fatty groups

19

Antarctic fish oil

7

anthocyanidins

9

148

122

antioxidants atmospheric bleaching

64

fats and oils corn oil

114

cottonseed oil

106

groundnut oil

111

lard

100

rapeseed oil

127

safflower oil

133

sesame oil

134

soybean oil

137

135

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

antioxidants (Cont.) flavines

9

gossypols

10

oxidized fatty groups

19

tocopherols

11

ascorbic acid (vitamin C)

17

66

ash

79

91t

205

206

Ashapura Volclay, Ltd.

92t

151

59

64

153

185

192

73

Asian and Pacific Coconut Community (APCC) atmospheric bleaching

103 49 214

automatic filters

179

Automatic Lovibond Tintometer AF960

202

autooxidation

49

50

babassu oil

13

104

127t

BASF Catalysts, LLC

73

75

94

147

151

B

batch method basic procedures

44

duration and adsorption

60

enzymatic degumming

232

fats and oils

132

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

batch method (Cont.) filtration Freundlich isotherm

162

178

221f

heat bleaching

48

neutralization

34

Baudoin test

136

beef tallow

74

100

bentonite

69

74

benzene

83

206

208

biodiesel

77

94

235

239

blackrun

179

bleachability

121

bleachers

153

169

172

206

blinding coconut oil

105

filter aids

164

filter membranes

161

168

177 filters

183

187

13

145

bodyfeed method

105

165

borneo tallow

111

Brassica family

126

129t

Buchner funnel

45

166

bunding

170

172

bursting strength

167

168

2

18

124

136

blubber

butter

166

49

103

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

C cake countercurrent procedure

47

fats and oils

114

151

filter aids

165

166

filter presses

180

182

filtration rate

178

fixed-bed bleaching

185

48

formation of

160

Freundlich isotherm

212

monofilaments

169

oil recovery

189

paper overlays

167

polyamide

168

safety

197

satin weave

172

spun-staple yarn

170

tests

205

calcium

22

Calgon Carbon Corporation

88

canola oil

capelin oil

152

161

168

198

24

30t

65

61

126

127

232

237

240t

144

carbon activated, forms of

86

activation procedures

82

adsorbents

69

87

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

carbon (Cont.) adsorption

35

carotenoids

9

castor oil

54

67

151

countercurrent procedure

47

deodorization

41

fats and oils coconut oil

104

grapeseed oil

108

safflower oil

134

sesame oil

136

sunflower oil

143

filtration

159

Freundlich isotherm

212

manufacturing sequences

87f

microstructure of

86f

moisture

164

64

oil recovery

190

particle size

67

phosphatides

13

phospholipids

22

pigments

6

powdered

79

semicontinuous bleachers

156

tests

200

tocopherols

11

use of

37

carbon dioxide

105

82

83

207

208

83

84

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

carbon-earth mixtures. See earth-carbon mixtures carbon monoxide

53

82

carcinogens

39

111

carotene acid activation and

72

adsorption

58

carotenoids

8

fats and oils

116

121

124

202

206

129

132

124f

132 heat bleaching

48

hydrobleaching

51

neutralization and washing

35

tests

199

carotenoids characteristics of

7

concentrated-miscella processing

218

dilute-miscella processing

215

fats and oils

103

Freundlich isotherm

209

heat bleaching

48

hydrobleaching

51

temperature and adsorption

60

ultrafiltration

30t

castor oil catalyst poisons caustic potash

49

50

73

136

148

31

32t

74

150

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

caustic soda fats and oils

122

138

150

151

filter units

184

membranes

162

neutralization and washing

147

148

225

229

169

34

oil recovery

194

cellulose fibers

165

centrifugal filters

185

centrifugal refining degumming

25

221

230

236

marine oils

146

147

olive oil

117

palm oil

122

rice bran oil

132

soybean oil

138

sunflower oil

141

fats and oils

oil recovery

148

140

234

char

82

83

charcoal

37

53

79

82

86

36

66

83

208 chemical activation

37

chemical bleaching

50

chemical changes

14

69

123

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

chemical refining enzymatic, versus

224

242f

fats and oils

107

123

142

NaOH

239

chemisorption

53

58

59

71

208

Chemviron Carbon, Ltd. chimney effect

65

88 198

chlorophyll acid activation and

72

adsorption

43

amorphous silica hydrogel

93

characteristics of

6

fats and oils corn oil

116

grapeseed oil

109

rapeseed oil

129

rice bran oil

132

soybean oil

138

sunflower oil

143

fixed-bed bleaching

47

Freundlich isotherm

209

general principles

98

heat bleaching

48

hydrobleaching

51

140

213

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

chlorophyll (Cont.) multistage procedures

46

silica

43

temperature

60

tests chlorophyll-adsorption isotherms

61

199

202

63f

215f

citric acid clay adsorption

36

degumming

24

25

231

122

123

fats and oils coconut oil

104

corn oil

115

lard

100

palm kernel oil

124

palm oil

120

rice bran oil

132

soybean oil

138

sunflower oil

141

general principles semicontinuous bleachers

142

97 156

sequestering and adsorption

66

trace metals

17

clays activation

84

adsorbents

69

70

adsorption and

35

59

atmospheric bleaching

64

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

clays (Cont.) batch bleaching

44

45

153

carbons and

38

86

87

carotenoids

9

chemical side effects

66

chlorophylls

7

43

commercial

72

80t

countercurrent procedure

47

degumming

24

229

215

217

dilute-miscella processing dimers

90

93t

18

fats and oils beef tallow

101

butterfat

103

castor oil

151

coconut oil

105

corn oil

115

116

cottonseed oil

106

108

grapeseed oil

108

109

groundnut oil

111

hydrogenated oils

149

lard

100

linseed oil

114

marine oils

148

olive oil

118

palm kernel oil

124

palm oil

120

123

rapeseed oil

126

129

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

clays (Cont.) rice bran oil

132

safflower oil

134

sesame oil

135

soybean oil

140

sunflower oil

143

vegetable butters

112

133

136

filter aids

163

filter units

182

184

filtration

159

162

173

211

213

fixed-bed bleaching

47

Freundlich isotherm

209

general principles

98

hydrobleaching

51

membranes

161

167

metals and

16f

17

moisture and

64

multistage procedures

46

neutralization and washing

34

35

oil recovery

189

191

193

oxidized fats

19

particle size and adsorption

67

phosphatides

13

phospholipids

22

physical and chemical adsorption

53

54

57

pigments

6

71

72

polishing

187

polyunsaturated fatty acids

178

58

4

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

clays (Cont.) SAFE bleaching saturated fatty acids semicontinuous bleachers

50 3 156

silica and

42

sterols

14

tests

199

91

200

206

207

161

171

172t

177

79

82

86

cocoa butter

111

112

coconut oil

13

38

88

103

tocopherols

11

wire gauzes

173

cloth membranes coal

coconut shell

40

52

140

142

79

cohune oil

104

127t

cold press

108

238f

cold test

107

126

187 color adsorption

59

bleaching effect of light

49

carbon

38

39

8

9

carotenoids copper enzymatic degumming

62

52 229

fats and oils

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

color (Cont.) beef tallow

101

butterfat

103

castor oil

151

corn oil

117

118t

cottonseed oil

107

108

groundnut oil

111

hydrogenated oils

149

interesterified oils

149

linseed oil

114

marine oils

148

olive oil

118

palm kernel oil

124

126

palm oil

120

121

rapeseed oil

129

rice bran oil

130

133

safflower oil

134

143

vegetable butters

112

Freundlich isotherm

55

gossypols

10

multistage procedures

45

oil recovery

190

phosphatides

13

pigments tests trace metals

102

123

57

213

215

191

193

194

217

218

5 201 16

Color in Oils Committee of AOCS

202

competitive adsorption

215

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

concentrated-miscella processing

217

conjugated unsaturation constraints

contamination

4

19

66

10

16

87

41

146

204

38

86

104

117

232

235f

18

197 continuous bleaching methods

46

156

229

231

continuous stirred-tank system (CSTR) copper chemisorption

54

color

52

degumming

23

hydrogenated oils

149

polyunsaturated fatty acids

4

sequestering and adsorption

66

tests trace metals copra

25

205 15 104

105

11

13

15

48

114

235

173

183

bleaching of

58f

105

degumming

31

235

corn oil (maize)

corrosion cottonseed oil

dilute-miscella processing

215

Freundlich isotherm

212

214

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

cottonseed oil (Cont.) gossypols

10

neutralization and washing

34

phosphatides

13

sesame oil and

136

settling

21

sterols

14

temperaturee and adsorption

59

tests

203

tocopherols

11

ultrafiltration

32t

countercurrent bleaching procedure

47

215

52

64

D dead-end hydrogenation system degumming basic procedures enzymatic

21 221

fats and oils coconut oil

104

corn oil

115

grapeseed oil

108

interesterified oils

149

linseed oil

113

palm oil

122

123

rapeseed oil

126

127

rice bran oil

132

118t

129

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

degumming (Cont.) soybean oil

21

filtration

159

Freundlich isotherm

214

organic pesticides

20

phosphatides

14

pigments reverse osmosis

138

140

122

123

5 31

deodorization activated carbon and chlorophyll dimers enzymatic degumming

41 7 18 237

fats and oils coconut oil

104

groundnut oil

111

hydrogenated oils

149

marine oils

148

palm kernel oil

124

palm oil

121

rapeseed oil

129

sesame oil

134

flavines

9

heat bleaching

49

oxidized fatty groups

19

phospholipids

22

physical refining

20

225

polycyclic aromatic hydrocarbons

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

deodorization (Cont.) (PAHs)

39

sterols

14

tocopherols

11

trace metals

16

De Smet continuous bleacher dewaxing

17

156 15

21

115

132

180

215 diatomaceous earths

164

diatomaceous silica

164

dilute-miscella, processing

214

dimers Direct Bleachability Test (DBT)

18 206

disorganized carbon

82

83

disposal

47

48

167

191

193

197

64

148

dissolved oxygen Dutch-weave gauze

184

E earth-carbon mixtures carbon versus coconut oil

38 105

commercial products

87

decolorization of oil

39f

particle size EDTA (ethylenediaminetetraacetic acid)

55

90

67 17

66

This page has been reformatted by Knovel to provide easier navigation.

Index Terms emulsion liquid membrane (ELM) energetically heterogeneous absorption

Links 33

34f

212

Engelhard Corporation

73

environment

26

enzymatic degumming

221

Enzymax

236

erucic acid

126

European Economic Community

146

exotic oils

235

explosion hazards

191

extinction values

203

extruded carbons

86

239

242f

146

197

205

F fatty acids characteristics of dilute-miscella processing dimers enzymatic degumming

2 217 18 223

237

fats and oils beef tallow

101

brassica oils

129t

butterfat

103

castor oil

150

coconut oil

104

corn oil

114

cottonseed oil

106

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

fatty acids (Cont.) grapeseed oil

111t

groundnut oil

110

111

98

101t

lard laurie oils

127t

linseed oil

113

marine oils

144

olive oil

117

palm kernel oil

124

palm oil

119

rapeseed oil

126

rice bran oil versus others

133t

safflower oil

133

134

sesame oil

134

136

soybean oil

137

sunflower oil

142

vegetable butters

112

Freundlich isotherm

209

oil recovery

190

oxidation

19

solvent bleaching

52

sterols

14

tests triglycerides and

147

127

143

194

203 1

2

105

111

29

32

Federation of Oils, Seeds, and Fats Associations (FOSFA) feed and bleed method

150

This page has been reformatted by Knovel to provide easier navigation.

Index Terms feedstock

filter aids

Links 46

47

116

126

151

163

163

186

148

151

161

178

186

168

170

177

178

180

186

filter cake. See cake filter membranes. See membranes filter presses

179

filter units

178

filtration basics of

159

bleachers

154

chemical side effects

67

degumming and

25

fats and oils

102

moisture

65

powdered activated carbons

86

reverse osmosis

26

trace metals

17

waxes

15

filtration rate

74

Filtrolbrand

73

fines

fish-liver oil

141

88

2

fish oil bleaching of marine oils chlorophylls neutralization and washing

144

146

7 34

polycyclic aromatic hydrocarbons

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

fish oil (Cont.) (PAHs)

40

powdered activated carbon

87

temperature and adsorption

59

tests

204

trace metals

17

waxes

15

fixed beds

47

70

86

87

212 flammable solvents flavines

190

198

9

flavor stability air and

36

chemical bleaching

50

chlorophyll adsorption

43

clay adsorption

37

dilute-miscella processing

104

215

dosage and adsorption

63

earths

38

fats and oils groundnut oil

110

111

marine oils

144

148

olive oil

118

palm kernel oil

124

palm oil

120

rapeseed oil

129

sesame oil

134

soybean oil

137

122

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

flavor stability (Cont.) sunflower oil

143

gossypols

10

heat bleaching

48

hydrobleaching

52

oxidized fatty groups

19

phosphatides

13

phospholipids

22

polyunsaturated fatty acids tests

4 203

tocopherols

11

trace metals

16

unsaturated fatty acids

20

204

3

flax. See linseed oil (flax) foots

20

111

113

111

150

117

142 FOSFA (Federation of Oils, Seeds, and Fats Associations)

105

free fatty acids (FFAs) acidity and adsorption

66

concentrated-miscella processing

217

218

enzymatic degumming

225

237

fats and oils beef tallow

104t

canola oil

240t

coconut oil

105

corn oil

115

cottonseed oil

106

107

108

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

free fatty acids (FFAs) (Cont.) groundnut oil

111

hydrogenated oils

149

lard

100

marine oils

147

palm oil

120

121

rice bran oil

130

132

soybean oil

138

sunflower oil

143

vegetable butters

112

Freundlich isotherm

221f

general principles

97

neutralization and washing

34

oil recovery polyunsaturated fatty acids

190 4

reverse osmosis

26

temperature and adsorption

59

Freundlich isotherm

194

55

28

31

57

207

58

209

Fuller’s earth adsorbents

69

adsorption

55

57

212

214

Freundlich isotherm moisture and adsorption

65

SAFE bleaching

50

Fulmont brand

73

123

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

G Galleon brand

73

74

gossypols

10

21

34

106

31

32t

108

15

117

120

190

205

121

123

107 Grace Davison division

93

granulated carbons

86

grapeseed oil

25

groundnut (arachis, peanut) oil bleaching of

110

color

203

degumming

25

heat bleaching

48

oxidative stability

143

tocopherols

11

waxes

15

gums. See phosphatides

H harvesting

10 151

hazards

126

183

HEAR

126

128

48

107

heat bleaching

149 herring oil

147

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

hexane degumming

23

deodorization

20

dilute-miscella processing

215

216

fats and oils

121

132

filter cake

205

206

filter units

183

oil recovery by solvent

190

reverse osmosis

27

solvent bleaching

52

32

Hiscocks & Raymond

121

Hollander weave

171

174

horizontal-leaf filter units

180

190

hot flue gas

104

hydrobleaching

217

51

hydrogenation carotenoids dilute-miscella processing dimers

8 215 18

fats and oils beef tallow

101

castor oil

151

cottonseed oil

107

grapeseed oil

108

interesterified oils

149

linseed oil

113

114

marine oils

144

145

olive oil

118

108

148

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

hydrogenation (Cont.) palm kernel oil

124

126

rapeseed oil

126

127

safflower oil

134

sesame oil

135

sunflower oil

143

hydrobleaching

51

neutralization and washing

34

polyunsaturated fatty acids

4

tests tocopherols hydrolysis

199 11 12

97

120

124

114

148

149

234

I ideal dilute-solution theory (IDST)

217

illipe oil (mee oil)

111

impregnated carbons

86

INEOS Silicas

92

interesterification

98 199

International Association of Fish Meal Manufacturers

147

iodine value fats and oils butterfat

103

castor oil

151

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

iodine value (Cont.) corn oil

114

grapeseed oil

108

groundnut oil

110

marine oils

144

olive oil

118

palm oil

121

rice bran oil

132

safflower oil

134

sesame oil

136

soybean oil

137

sunflower oil

142

vegetable butters

112

filter cake hydrobleaching

148

190 52

safety

197

tests

201

203

208

iron acidity and adsorption

65

degumming

23

hydrogenated oils phosphatides polishing sequestering and adsorption

149 13 187 66

soybean oil

141t

tests

205

trace metals isopropanol

25

15 215

216

217

This page has been reformatted by Knovel to provide easier navigation.

Index Terms IUPAC

Links 37

199

J Janke & Kunkel

230

Jatropha

235

K kaolin

58

Kelly and Sweetland filter presses

190

King and Wharton

153

Köper Panzer-tressengewebe (KPZ55) wire gauze

174

177f

184

laboratory-scale tests

230

236

237

lachrymatory fumes

189

197

Langmuir isotherm

54

57

212

218

185

L

Laporte Industries, Ltd.

73

lard

13

207

210

15

17

38

98 lauric acid

3

104

127t

lauric oils

11

22

24

127t

149

LEAR

126

128

Lecitase Novo

237

124

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

Lecitase Ultra

227

230

231

232

22

138

234

16

49

53

203

5

14

133

142

213

215

236 lecithin

17 239

Life Cycle Assessment light linoleic acid

239

linolenic acid dilute-miscella processing

217

fats and oils corn oil

114

groundnut oil

110

linseed oil

113

palm oil

119

rapeseed oil

128

rice bran oil

130

sesame oil

134

soybean oil

137

sunflower oil

143

linseed oil (flax) bleaching of

113

degumming

22

filter units

184

oil recovery

189

phosphatides

13

lipolysis

150

liver oils

50

Lovibond color

55

25

193

194

212

This page has been reformatted by Knovel to provide easier navigation.

Index Terms lubricants lutein

Links 15

74

144

151

8

215

216

2

38

79

82

22

24

65

144

189

M macropores

84 magnesium

7

manganese

16

margarine fats and oils cottonseed oil

107

interesterified oils

150

marine oils

144

palm kernel oil

124

rapeseed oil

127

safflower oil

134

sesame oil

136

sunflower oil

143

108

reverse osmosis

27

waxes

15

marine oil

4

18

193

197

McCloskey Scientific automatic colorimeter

202

melon-seed oil

57

58

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

membranes filters

32

184

159

160

163

166

reverse osmosis

26

28

29

31

technology

33

2

38

79

82

162

187

192

23

24

filtration

menhaden oil mesopores metal gauzes

145

metals acid activation

72

acidity and adsorption

66

amorphous silica hydrogel

92

characteristics of

15

clay adsorption

36

degumming

22

227

fats and oils corn oil

115

cottonseed oil

107

hydrogenated oils

149

lard

100

marine oils

148

palm kernel oil

124

palm oil

122

rapeseed oil

129

safflower oil

134

soybean oil

138

general principles

97

hydrobleaching

52

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

metals (Cont.) impregnated carbons porphin rings

86 6

reverse osmosis

27

sequestering and adsorption

66

specialty products

93

tests microfiltration micropores

mineral acids

mineral oils mixed triglycerides

29

31

205 26

27

2

38

84

85

84

147

198

231

70

75

79

82

169

184

78

1

Mizusawa

73

Modified Caustic Refining (MCR)

93

Modified Physical Refining (MPR)

93

moisture activation procedures

82

adsorption

58

amorphous silica hydrogel

91

batch bleaching

44

bleachers

153

fats and oils

104t

oil recovery

190

pigments

64

120

141

151

5

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

moisture (Cont.) semicontinuous bleachers

156

settling

20

trace metals

16

molasses index

83

208

111

117

mold

21

molecular-weight-cutoff (MWCO) value montmorillonite mowrah fat

26 69

72

75

111

143

77

111

N National Cottonseed Products Association (NCPA)

105

National Soybean Processors Association, U.S.A (NSPA) needle loom (batt-on-base fabric)

138 173

neutralization basic procedures

34

continuous bleaching methods

46

fats and oils cottonseed oil

107

interesterified oils

149

linseed oil

113

marine oils

148

palm oil

121

sesame oil

136

108

122

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

neutralization (Cont.) soybean oil

138

sunflower oil

143

filtration organic pesticides

141t

159 20

nickel catalyst degumming

25

deodorization

20

fats and oils

104

filter aids

164

oxidized fatty groups

19

particle size and adsorption

67

phosphatides

13

silica

43

tests

199

126

136

148

22

nitrogen batch bleaching

44

edible oils

64

filter units

182

marine oils

148

oil recovery

190

powdered activated carbons tests nitrogen adsorption capacity nitrogen blanketing

192

87 200

208

83 120

nitrosamines

20

nonatmospheric bleaching

64

non-swelling bentonite

70

156

This page has been reformatted by Knovel to provide easier navigation.

Index Terms Norit Nederland B.V. nylon

Links 90 178

O Oil-Dri Corporation of America oil recovery

oil-water interface oleic acid

oleic-linoleic family

72

77

94

189

197

205

234

237

223

224

228

3

119

124

133

142

217

106

108

110

114

100

101

144

232

134 olive oil bleaching of

117

dimers

18

phosphatides

13

powdered activated carbons

88

tocopherols

11

trace metals

15

unsaturated fatty acids

3

optical brightening

49

organ meats

98

organo-sulfur compounds

20

oxidation atmospheric bleaching carotenoids chemical bleaching

64 8 50

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

oxidation (Cont.) chemical side effects chlorophylls clay adsorption

66 7 37

concentrated-miscella processing

217

dilute-miscella processing

215

dosage and adsorption fats and fatty acids

61

63

2

19

fats and oils corn oil

114

cottonseed oil

106

grapeseed oil

109

groundnut oil

110

linseed oil

113

marine oils

144

olive oil

118

palm kernel oil

124

palm oil

120

rapeseed oil

128

rice bran oil

130

safflower oil

133

sesame oil

134

soybean oil

137

sunflower oil

143

filter-cloth selection

179

Freundlich isotherm

214

general principles

97

gossypols

10

147

148

121

132

138

140

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

oxidation (Cont.) heat bleaching

48

multistage procedures

45

oil recovery

189

190

193

polishing

187

204

206

53

84

87

13

15

38

104

49

51

119

polyunsaturated fatty acids

4

saturated fatty acids

3

settling

21

temperature and adsorption

60

tests unsaturated fatty acids oxygen

203

194

3

P palm kernel oil

124 palm oil bleaching of

48

carotenoids

8

degumming

22

dimers

18

hydrogenated oils

24

149

neutralization and washing

34

phosphatides

13

physical refining

23

SAFE bleaching

50

settling

21

35

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

palm oil (Cont.) steam bleaching tests

50 204

tocopherols

11

trace metals

16

paper overlay

160

206

165

167

177

187 particle size acid activation

72

activated carbon

79

conditions affecting adsorption

67

expanded perlites

165

filter aids

163

filtration

159

natural clays

86

189t

69

paper overlays

167

polishing

186

tests

199

textiles

168

207

peanut oil. See groundnut (arachis peanut) oil Pellerin-Zenith system

154

Perform product line

79t

perlites

164

peroxide value (PV) dosage and adsorption

61

fats and oils beef tallow

104t

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

peroxide value (PV) (Cont.) groundnut oil

111

marine oils

147

palm kernel oil

124

palm oil

120

rapeseed oil

128

safflower oil

134

sunflower oil

143

general principles

98

oxidized fatty groups

19

tests

148

122

203

pesticides

20

petroleum

74

86

petroleum-engine exhaust fumes

105

petroleum ether

190

191

205

206

pharmaceutical use

134

150

152

188

7

48

49

pheophytins phosphatides acid activation

72

carbon

38

characteristics of

11

chlorophyll

43

degumming

21

duration and adsorption

60

39

221

fats and oils butterfat

103

coconut oil

104

corn oil

115

116

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

phosphatides (Cont.) cottonseed oil

106

grapeseed oil

108

groundnut oil

111

lard

100

linseed oil

113

marine oils

147

olive oil

117

palm kernel oil

124

palm oil

121

122

rapeseed oil

127

129

rice bran oil

130

132

safflower oil

134

sesame oil

136

soybean oil

138

sunflower oil

141

142

filtration

159

194

Freundlich isotherm

209

214

heat bleaching

107

148

143

48

pigments

5

polishing

186

6

reverse osmosis

26

27

sequestering and adsorption

66

settling

20

21

silica

61

91

specialty products

93

29

92

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

phosphoric acid activation procedures

82

degumming

22

23

25

fats and oils coconut oil

104

corn oil

115

grapeseed oil

108

marine oils

147

palm kernel oil

124

palm oil

122

rice bran oil

132

soybean oil

138

Freundlich isotherm

214

general principles

97

phosphatides

11

prior to neutralization

34

semicontinuous bleachers

123

156

sequestering and adsorption

66

trace metals

17

phosphorus biodiesel production

235

chlorophyll adsorption

43

enzymatic degumming

23

221

232

236

237

238

235f

fats and oils corn oil

118t

marine oils

147

palm oil

122

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

phosphorus (Cont.) soybean oil

138

sunflower oil

141

vegetable butters

112

filtration general principles plant-scale enzymatic degumming reverse osmosis tests

143

159 97 231 31 205

230

9

43

physical activation

37

82

physical adsorption

53

54

22

224

photooxidation

49

140

58

59

physical refining degumming fats and oils corn oil

118t

cottonseed oil

106

marine oils

148

palm oil

23

119

rapeseed oil

129

soybean oil

140

141t

sunflower oil

141

143

phospholipids

22

reverse osmosis

31

semicontinuous bleachers steam bleaching testing oil bleachability

123

156 50 206

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

pigments acid activation

70

activated carbon

83

adsorption

53

air bleaching

49

carbon

37

characteristics of

50

chemisorption

54

clay adsorption

36

dosage and adsorption

57

38

5

chemical bleaching

dilute-miscella processing

55

215

216

60

62

fats and oils butterfat

103

corn oil

116

cottonseed oil

106

107

palm oil

122

123

rice bran oil

130

132

soybean oil

140

sunflower oil

143

Freundlich isotherm

209

heat bleaching

48

multistage procedures

45

oil recovery

46

190

reverse osmosis

27

solvent bleaching

52

temperature and adsorption

59

tests

212

199

208

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

pilchard oil

147

184

Plain Dutch weave

171

173

plant-scale enzymatic degumming

230

237

plastic

100

102

plate and frame presses

161

179

198

Player and Wood test

147

p-nuclear magnetic resonance (NMR)

234

polishing filters

151

156

165

168

105

134

143

144

174

186 polyamide (nylon) filter membranes

168

polycyclic aromatic hydrocarbons (PAHs) activated carbon

39

characteristics of

20

fats and oils heat bleaching

104

88

48

polyester (terylene)

169

polyethylene packaging

102

polyunsaturated fatty acids (PUFAs) characteristics of dimers marine oils oxidized fatty groups polishing

4 18 148 19 187

temperature and adsorption

59

trace metals

17

polyunsaturated oils

44

64

pomace

88

108

This page has been reformatted by Knovel to provide easier navigation.

Index Terms powdered activated carbons

Links 79

83

86

87

precoat layer

165

167

182

193

preferential draft (chimney effect)

198

Premiere

123

Premier Jus (oleo stock)

100 45

61

98

105

153

212

64

66

124

130

132

138

143

148

press-bleaching effect

press cake. See cake prooxidant catalysts prooxidant metals fats and oils

multistage procedures

46

trace metals

13

prostaglandins Pure-Flo product line

15

16

4

9

19

203

204

42

201

204

231

5 77

R rancidity

Rancimat test

49

rapeseed oil (colza) bleaching of

126

chlorophyll adsorption

43

degumming

22

25

235

236

oil recovery from gums

232

234

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

rapeseed oil (colza) (Cont.) phosphatides phosphorus removal

13 235f

reverse osmosis

31

sulfur content

20

tests

230

ultrafiltration

32t

waxes

15

reverse osmosis

26

reversible adsorption

53

210

211

rice bran oil bleaching of

129

enzymatic degumming

235

phospholipids

225t

reverse osmosis

31

settling

21

waxes

15

rice hulls

79

ricin

236

88

217

151

Rockwood Additives, Ltd.

73

rubber

86

179

rubber-seed oil

57

58

182

183

185

186

183

189

197

RV filtration

S SAFE bleaching

50

safety

20

This page has been reformatted by Knovel to provide easier navigation.

Index Terms safflower oil salts

Links 73

133

9

65

sardine oils

148

satin weave

172

175f

2

52

180

183f

saturated fatty acids (SFAs)

207

118

124

180

185

186

23

26

27

29

31

36

Schenk self-cleaning enclosed horizontal-leaf filter seal oil

144

sedimentation tests

159

Select label

77

self-cleaning filters

179

semicontinuous bleachers

153

Sen Gupta

separate aqueous solution

207

194

Sequential Addition

93

sequestering agents

17

36

66

sesame oil

15

133t

134

218

5

20

65

113

138

141

31

111

112

6

15

124

148

107

127

134

143

settling

shea butter shipment of oils shortening short-time degumming process (STDG)

232

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

silica adsorption

54

clay adsorption

35

enzymatic degumming fish oil Freundlich isotherm multistage procedures rapeseed oil use of silicic acid simple triglycerides

229

58

60

61

216

217

236

87 214 46 129 42 65

215

1

smoked meats

41

smoke-dried copra

40

smoke-dried safflower

134

smoke-dried seed

143

52

soaps amorphous silica hydrogel

91

carbon

38

92

fats and oils butterfat

103

corn oil

115

groundnut oil

111

interesterified oils

149

palm kernel oil

124

rapeseed oil

129

rice bran oil

130

soybean oil

141t

vegetable butters

112

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

soaps (Cont.) filtration oil recovery

97 190

reverse osmosis

26

silica and

61

specialty products

93

temperature and adsorption

59

tests

159

31

206

soapstock batch bleachers

153

carotenoids

8

degumming

25

duration and adsorption

60

225

239

fats and oils castor oil

151

marine oils

147

palm oil

122

soybean oil

138

sunflower oil

141

neutralization and washing

34

phosphatides

13

sterols

14

148

140

sodium carbonate

193

194

sodium hydroxide (NaOH)

225

227

Solexol process

239

240t

52

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

solvent extraction fats and oils

108

114

117

130

141 oil recovery

191

194

solvents adsorption

55

bleaching

52

characteristics of

20

concentrated-miscella processing

217

dilute-miscella processing

215

216

fats and oils

112

121

neutralization and washing

35

oil recovery by

190

reverse osmosis

28

tests

205

SORBSIL R

92

203

soybean oil acid activation

72

anthocyanidins

10

bleaching of

137

chemical side effects

67

degumming

22

23

215

216

dilute-miscella processing dimers

18

dosage and adsorption

61

63

enzymatic degumming

221

230

235

238

fixed-bed bleaching

25

231

232

47

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

soybean oil (Cont.) Freundlich isotherm gums hazardous metals

213 236f 16f

heat bleaching

48

neutralization and washing

35

oil recovery from gums phosphatides phosphorus pigment polishing

214

49

234 13

225t

235f 9 187

polycyclic aromatic hydrocarbons (PAHs) quality parameters

42t 141t

reverse osmosis

26

sequestering and adsorption

66

settling

21

sterols

14

temperature and adsorption

59

tests

29

203

tocopherols

11

trace metals

16

ultrafiltration

30t

waxes

15

special-degumming process

25

specialty products, adsorbents

93

special wet degumming (SWD) procedure

115

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

spectrophotometry

203

205

spectroscopy

209

217

sperm whale

15

144

split soap

59

66

69

3

101

118

stearin

123

124

186

sterols

14

106

238

141

148

90

95

spun-staple yarn filter membranes

170

stainless steel membranes

161

STDG process

235

stearic acids

213

217

storage enzymatic degumming

222

fats and oils

106

124

20

21

subcutaneous fat

98

100

Süd-Chemie AG

78

80t

settling

sulfur fats and oils butterfat

103

coconut oil

104

cottonseed oil

107

groundnut oil

110

marine oils

146

palm kernel oil

124

palm oil

121

122

rapeseed oil

20

126

safflower oil

134

sesame oil

136

147

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

sulfur (Cont.) sunflower oil

143

particle size and adsorption

67

silica

43

sulfuric acid

65

205

sulfuric acid–Fuller’s earth See SAFE bleaching sulfur olive oil

117

sunflower oil (tournesol, girasol) bleaching of

141

degumming

25

phosphatides

13

235

polycyclic aromatic hydrocarbons (PAHs)

41

reverse osmosis

31

settling

21

tocopherols

11

waxes

15

sunlight

49

superdegumming process

24

supported liquid membrane (SLM)

33

Sutcliffe Speakman Ltd.

88

Suzuki swelling bentonite Swift Test synthetic silicas

116

129

9

14

69

200

207

218 69 204 159

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

T table oil

111

118

126

134

187 tallow air bleaching

49

clay adsorption

36

degumming

24

fatty acids

2

phosphatides

13

SAFE bleaching

50

solvent bleaching

52

temperature and adsorption

60

trace metals

17

use of carbon

38

waxes

15

tertiary butylhydroquinone (TBHQ)

100

120

terylene

177

178

193

tests activation procedures bleaching

83 199

carbon

38

87

clays

37

72

enzymatic degumming

230

filtration

163

perlites

165

textiles

160

166

165

167

168

172

This page has been reformatted by Knovel to provide easier navigation.

Index Terms thermal bleaching

Links 48

thermodynamic adsorption equilibrium Tirtiaux

210

216

217

226

tocopherols carbon

42

characteristics of

11

crude vegetable oils

130t

enzymatic degumming

238

fats and oils corn oil

114

cottonseed oil

106

groundnut oil

110

lard

100

olive oil

118

palm oil

121

rapeseed oil

127

rice bran oil

130

safflower oil

133

sesame oil

134

soybean oil

30t

sunflower oil

143

gossypols

10

oxidation

12f

reverse osmosis Tonsil

total oxidation (totox)

107

137

29 58

78

95

214

36

62

90

91

204

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

toxicity

151

191

206

triglycerides acidity and adsorption characteristics of clay adsorption dilute-miscella processing dimers

66 1 36 215 18

fats and oils

114

natural clays

69

oil recovery

190

oxidized fatty groups

19

settling

20

solvent bleaching

52

temperature and adsorption

59

tests TriSyl silicas tucum oil tung-oil fatty acid twill weave

216

118

119

191

194

123

206 93 104

127t

4 172

174

177

U ultrafiltration (UF) Ultra Turrax T25 ultraviolet radiation

26 230 49

UNIDO

103

Unilever Company

119

105

This page has been reformatted by Knovel to provide easier navigation.

Index Terms unsaturated fatty acids

Links 3

104

106

111

114 unsaturated oil acidity and adsorption

66

chemisorption

54

fats and oils

113

124

oil recovery

189

190

191

192

54

59

208

193 open filtration

64

safety

197

totox value

204

trace metals

16

U.S. Department of Agriculture (USDA)

202

V van der Waals forces

53

vegetable butters

111

vertical-leaf (gauze) pressure filter

194

Villavechia test

135

virgin oil

117

118

136

viscosity

59

151

159

217

229

210

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

W water adsorption

53

amorphous silica hydrogel

92

carotenoids clay adsorption

65

8 36

enzymatic degumming

222

230

filters

183

184

neutralization and washing

34

oil recovery

190

olive oil

117

oxidized fatty groups

236

192

205

19

polishing

186

safety

197

water degumming degumming

23

enzymatic degumming

226

227

232

fats and oils

126

138

143

oil recovery from gums

232

234

water-oil interface

236

227

waxes characteristics of fats and oils

15 113

114

141

144

pigments

5

polishing

186

settling

20

130

132

21

This page has been reformatted by Knovel to provide easier navigation.

Index Terms weathering weaves

Links 65

69

161

169

171

177

59

144

145

172

185

187 wet ashing

205

wet milling

114

whale oil

20 148

wheat-germ oil

11

winterization degumming dilute-miscella processing

25 215

fats and oils corn oil

114

cottonseed oil

107

olive oil

117

rapeseed oil

126

rice bran oil

21

sunflower oil

142

115

132

wire gauzes (metal cloths) filter membranes

weaving and wood

160

161

186

187

171

173

79

82

86

167

169

170

187

179 wood char wool W.R. Grace & Co.

83 159 95

This page has been reformatted by Knovel to provide easier navigation.

Index Terms

Links

X xanthophylls

8

116

159

168

169

138

156

157

Y yarns

177

Z Zenith process

This page has been reformatted by Knovel to provide easier navigation.