History of Soy Oil Hydrogenation and of Research on the Safety of Hydrogenated Vegetable Oils - Page 1
A Special Report on The History of Soy Oil, Soybean Meal, & Modern Soy Protein Products
A Chapter from the Unpublished Manuscript, History of Soybeans and Soyfoods: 1100 B.C. to the 1980s
by William Shurtleff and Akiko AoyagiCopyright 2007 Soyinfo Center, Lafayette, California
It has often been noted that hydrogenation is the single most important technical advance leading to the increased consumption of soy oil. The earliest food uses of soy oil were primarily in hydrogenated products (shortening and margarine), rather than as cooking or salad oils. Moreover hydrogenation is of immense importance in all modern soy oil and fat processing technology. And it made possible the development of the first two plant products (shortening and margarine) that replaced their animal counterparts (lard and butter), as the latter became increasingly expensive and relatively scarce.
By the late 1970s roughly 60% of all edible oils and fats in the US were partially hydrogenated (Dutton, in Emken and Dutton 1979). And an estimated 75% of the soy oil used in the US was hydrogenated to make shortening and margarine, as well as large amounts of lightly hydrogenated soy cooking and salad oils (Kromer 1976). Moreover, some soy oil has long been hydrogenated to make nonfood products, such as soaps.
Fats can be modified in at least three basic ways: by hydrogenation, interesterification, and fractionation. Of these, hydrogenation is by far the most flexible, versatile, and widely used. Since the other two processes are often used in conjunction with hydrogenation, we will also discuss them in this context.
WHAT IS HYDROGENATION?
The general term "hydrogenation" refers to the "reaction of hydrogen with an organic compound." The process used to modify oils is technically known as "catalytic hydrogenation," since it takes place in the present of a catalyst; however in common parlance the process is usually referred to as simply "hydrogenation." In catalytic hydrogenation, hydrogen gas is agitated in hot liquid oil, usually under pressure, in the presence of a catalyst, typically powdered nickel. Hydrogen atoms are added to some of the double bonds of the unsaturated fatty acids in the oil. The more unsaturated fatty acids (such as the triunsaturated fatty acid or triene linolenic) are generally hydrogenated first; each double bond can take up two hydrogen atoms. Linolenic (18:3) that has one double bond hydrogenated becomes linoleic (18:2). A saturated fatty acid (such as palmitic 16:0 or stearic 18:0) has all the hydrogen it can hold.
Hydrogenation can serve either or both of two important functions. First, it can be used to improve the flavor stability and keeping qualities of an oil, especially by reducing or removing the content of highly reactive (triunsaturated) linolenic acid, thus preventing much of the oxidative rancidity and off-flavor development that might otherwise occur, especially after the oil is used for frying. An unhydrogenated oil turns rancid by picking up oxygen at sites of unsaturation; hydrogenation blocks this by adding hydrogen at these points. In the early days hydrogenation was likewise widely used to rid off-grade or rancid oils of much of their bad odor and flavor (Weber and Alsberg 1934). Second, hydrogenation can change the physical character of an oil by converting it from a liquid into a semisolid, plastic fat, closely resembling butter or lard in texture, and suitable for use in making margarine or butter. Hydrogenation also generally lightens the color of an oil. There are many different degrees of hydrogenation, depending largely on the time, temperature, and pressure used for hydrogenation. Virtually all of the soy oil used for the dual purposes of cooking or salad oils is lightly hydrogenated to improve its flavor stability without changing its consistency. The soy oil used in shortenings and margarines is hydrogenated more completely, to a lower iodine value (see Chap. 47), transforming some or all of it from a solid into a semisolid. Although shortening and margarine appear to be solids, they generally contain less than 20% solids, and in some cases as little as 5% solids is sufficient to allow them to retain their form at 22-27°C (70-80°F). Soy oil is never hydrogenated complete, only partially. Complete hydrogenation would lead to a product containing only saturated fats (palmitic acid 16:0 and stearic acid 18:0), having an iodine value near zero, a solid waxy consistency something like candle wax or hard beef tallow, and a melting point above 60°C, so that it would not melt readily upon being consumed; body temperature is 37°C (98.6°F), but fats with a melting point below 50°C are readily digested.
The following are generally considered disadvantages or at least tradeoffs resulting from hydrogenation. It reduces the oil's content of essential linoleic acid, produces a variety of fatty acid isomers (which are not widely found in nature and about whose safety there is still some question), lowers the P/S (polyunsaturated-to-saturated-fatty acid) ratio and thus reduces the degree of unsaturation of the oil (but does not generally increase the proportion of saturated fatty acids in the oil), leads to the production of some undesirable flavors and odors (called "hardening flavors") and may leave traces of catalyst in the hydrogenated oil (legislation sets upper limits on these). The "hardening flavors," caused by minute amounts of aldehydes (especially 6- trans nonenal) formed by oxidation, are largely removed during subsequent deodorization. Hydrogenation also largely destroys the characteristic aroma of an oil, leaving it bland and almost tasteless; this may be either an advantage or a drawback, depending on the oil.
The actual process of hydrogenation, even in today's modern continuous-process vegetable oil refineries, is still usually a batch process and usually the only one. Typically, a mixture of refined oil and finely powdered nickel catalyst (comprising 0.05-0.1% of the weight of the oil) is pumped into a cylindrical pressure reactor of 5-20 tons capacity??. It is heated by heating coils to 120-188°C (248-370°F) at 1-6 atmospheres pressure. Hydrogen is pumped into the bottom of the reactor and dispersed by a stirrer, continuously, as bubbles into the oil. The internal conditions in the reactor may be either "selective" as is typically the case in making margarine (higher temperature and catalyst concentration, lower pressure and agitation rate, and sometimes use of a special nickel catalyst) or "nonselective" as is typically the case in making shortening (the opposite conditions and use of a nickel catalyst). For the same drop in iodine value selective hydrogenation gives a product which contains less linolenic acid (which is not wanted) and is harder, but which has a higher content of trans fatty acids (to be discussed below). After hydrogenation is completed to the desired degree, the oil is filtered to remove the catalyst (which may be reused) then pumped to a storage tank; it may later be blended with other harder or softer fats or oils to make margarine or shortening.
About Cis and Trans Fatty Acid Isomers . Isomers are two or more substances that are composed of the same elements united in the same proportions but different in molecular structure. The two important types of isomers in fatty acids are geometric isomers and positional isomers (Meyer et al. 1974).
Geometric Isomers. Unsaturated fatty acids can exist in either the cis or trans forms, depending on the configuration of the hydrogen atoms attached to carbon atoms joined by double bonds. If the hydrogen atoms are on the same side of the carbon chain, the arrangement is called cis , after the Latin prefix meaning "on this side." If the hydrogen atoms are on opposite sides of the carbon chain, the arrangement is called trans (meaning "on the other side"), as shown in the following simplified schematic diagrams.
H H H H H H H
... -C- C=C -C ... ... -C- C=C -C
H H H H H
Both trans and cis configurations can be denoted in several ways, such as " trans fatty acids," " trans acids," or " trans isomers."
With very few exceptions, the fatty acids in natural food oils and fats are in the cis configuration, not the trans . The main exception is vaccenic acid, the positional isomer of oleic acid, which occurs in small amounts in the fats of some ruminant animals. Most trans isomers are produced when oils and fats of any type are hydrogenated; portions of the fatty acid chain at a double band may be twisted through 180 degrees, thus placing a hydrogen atom on the opposite side of the chain.
Oleic acid (the common cis form) and elaidic acid (the generally man-made trans form) are geometrical isomers. When two double bonds are present in a fatty acid molecule, four isomers are possible: cis-cis , cis-trans , trans-cis , and trans-trans .
Positional Isomers. Two fatty acid molecules are positional isomers if the location of the double bond in each differs. Hydrogenation often causes double bonds to migrate to new positions along the carbon chain, where there were formerly no double bonds in the hydrogenated oil. When the double bond in oleic acid moves from the normal ninth position to the less common 11th position, oleic acid is transformed into vaccenic acid, a new positional isomer.
All positional isomers are also geometric, i.e. trans . [Only trans? Positional isomers also have trans or cis configurations? descriptions??] Hydrogenation typically changes the geometrical and/or positional configuration of 10-50% of the molecules in an oil and generally creates 6-12 new trans isomers that did not exist in the natural, unprocessed oil. Selective hydrogenation further increases the quantity of trans isomers.
The position of the double bond affects the melting point of a fatty acid to a limited extent, but the geometric configuration of the double bond affects it dramatically. For example, the most common trans isomer, trans -oleic acid or elaidic acid, has a melting point of 43.5°C, as compared to 13.5°C for its geometric counterpart cis -oleic acid. Thus oleic acid is a liquid at temperatures considerably below room temperature, while elaidic acid remains a solid even at temperatures above room temperature--as befits a margarine. A second difference is that trans forms are generally less reactive and thus more resistance to oxidation than cis forms. Third, the human body tends to metabolize trans isomers almost as if they were saturated fats (Meyer 1974 disagrees??), even though they are generally monounsaturated or polyunsaturated fatty acids, and may be listed as such in the product composition. Fourth, the cis isomer is physically more flexible than the trans (Irene disagrees??). It is folded back on itself in a sort of zigzag shape, whereas the trans isomer is more stretched out. The greater rigidity of the trans isomer is thought by some to affect its metabolism in cell membranes, as discussed later.
The amount of isomerization that occurs during hydrogenation can be regulated by adjusting the type of catalyst used, the degree of hydrogenation, and the conditions of time, temperature, and pressure. However it is not possible to eliminate trans isomers during hydrogenation. In some hydrogenated products trans acids are considered a plus factor, in others a drawback. For example, margarine and shortening makers have long known that increasing the proportion of trans isomers in their product gives the desired higher melting point with a lesser degree of hydrogenation. It is interesting to note that ordinary margarine contains only a trace of more saturated fats than the oil from which it is made. It is harder largely because of the higher level of trans -unsaturated fatty acids. Adding fairly large amounts of trans simplifies formulation problems by giving a quality product that melts at about mouth temperature, has a good mouthfeel, and spreads well. However, there is considerable controversy as to whether all this trans gives the most healthful product, as discussed below. On the other hand, makers of lightly hydrogenated soy cooking and salad oils try to minimize trans formation because of the higher crystallization losses they cause during winterization.
A very precise terminology has been developed for use in discussing trans and cis fatty acid isomers. Some of the more important saturated fatty acids and the cis and trans forms of unsaturated fatty acids are shown in Figure ??. Basic information on fatty acids is given in Chapter 47.
Common Name Systematic Name Symbolic Notation
Palmitic acid Hexadecanoic 16:0
Stearic acid Octadecanoic 18:0
Oleic acid cis-9-Octadecenoic 9c-18:1
Elaidic acid trans-9-Octadecenoic 9t-18:1
Vaccenic acid trans-11-Octadecenoic 11t-18:1
Linoleic acid cis-9,cis-12-Octadecadienoic 9c,12c-18:2
Trans diene trans-9,cis-12-Octadecadienoic?? 9t,12c-18:2
Conjugated diene cis-9,trans-11-Octadecadienoic?? 9c,11t-18:2
Linolenic acid cis-9,cis-12,cis-15-Octadecatrienoic 9c,12c,15c-18:3
First find oleic acid. Our terminology tells us that this is a monounsaturated fatty acid with 18 carbons in the basic chain. The single double bond is a cis double bond in the ninth position, i.e. nine carbon bonds from the terminal methyl (CH3) group on the carbon chain. Second, note that we have listed only two common trans forms; there are many other monounsaturated and diunsaturated trans forms, but two will suffice here to illustrate basic principles of terminology. Third, note that elaidic acid is the trans isomer of oleic acid; it is not a positional isomer, since in each case the double bond is in the ninth position. Vaccinic acid, however, is a positional trans isomer??, since the double bond is two positions away from the natural ninth position. Fourth, note what happens when linolenic acid is hydrogenated to linoleic, then linoleic to oleic. In each case, the double bond with the highest number (i.e. furthest from the methyl end of the carbon chain) is hydrogenated and thus eliminated as a double bond; how orderly! Fifth, the cis-trans configuration of diunsaturated fatty acids (dienes) is more reactive with oxygen than the usual cis-cis form, so that during hydrogenation one usually attempts to keep the former to a minimum. The conjugated diene, in which the two double bonds are separated by a single group of one carbon and two hydrogen atoms, with only one single bond (i.e. 9-11, or 10-12), is much more reactive toward oxygen than double bonds separated by the usual two single bonds (i.e. 9-12) (Bai?ley 1951, p. 9).
H H H H H H H H H H H H H
H H H H H
Nonconjugated fatty Conjugated fatty
acid chain acid chain
Although trans fatty acids are not generally found in unhydrogenated vegetable oils or in plants, and they are not produced or synthesized in the bodies of humans and most other mammals, they are found to exist naturally in the fat of various ruminants, such as cattle, sheep, and goats; here they are thought to be formed via biohydrogenation by microorganisms in the animal's rumen. The butter and fat of cattle typically contain 5-10% trans fatty acids, and these are consumed by humans who drink milk or eat butter, tallow, or beef. It is important to note, however, that the predominant trans fatty acid in ruminant fat (vaccenic acid; 11 t- 18:1) is not the same as the predominant trans fatty acid produced by hydrogenation of vegetable oil (elaidic acid; 9 t- 18:1). Also the trans fatty acids in ruminant fats and hydrogenated oils have different distributions of positional isomers. However, minute amounts of all the trans fatty acids found in hydrogenated oils are also found in ruminant fats, such as milk (Patton and Jensen 1975) and minute amounts of trans fatty acids have even been found in the leaves and seeds of several types of plants. The body metabolizes each type of trans acid differently. Strictly speaking, then, the trans fatty acids produced in hydrogenated oils are not "unnatural" in the sense that minute amounts are found here and there in nature, but they are "unnatural" in the sense that they are not found in the natural oils in which they are produced during processing.
The main source of trans fatty acids in most diets is hydrogenated vegetable oils; trans acids constitute roughly 30-35% of the fatty acids in stick margarine, 14% in soft-tub margarine, and 11% in soy cooking and salad oils (ref??). Note that in hydrogenating soy oil to make a salad or cooking oil, the iodine value is reduced from 130-110, polyunsaturates are reduced by about 30%, monounsaturates are roughly doubled, but saturated fatty acids are essentially unchanged. About 11% of the fatty acids are trans isomers and in 24% of the fatty acids (not shown), the positions of the double bonds migrate (ref??).
In summary, then, the typical Westerner consumes trans fatty acids from two sources: hydrogenated oils and ruminant animal fats. A 1974 study in Germany estimated that 35-45% of the trans intake originated in ruminant fats (Applewhite 1981). It was estimated in 1975 that the trans fatty acid content of the fat in the average American diet was about 8%; most of this was contributed by hydrogenated vegetable oil products (Kummerow 1975; FASEB 1976). Rizek et al. (1974) estimated that in the period from 1937 to 1972 per capita annual consumption of trans fatty acids increased by 81%, from 6.3-11.4 gm. During the same period per capita consumption of vegetable oils and fats increased by only 64% (from 36-59 gm).
Composition of Hydrogenated Soy Oil Products, Lard, and Butter
Product Saturated Monosaturated Polyunsaturated
Fatty Acids Fatty Acids Fatty Acids P/S IV Source
(Percent) (Percent) (Percent)
Unrefined soy oil 15.5 25.0 60.5 (53.2% linoleic) 4.0?? FASEB 1976
Lightly Hydrogenated 14.3 44.6 (9.3% trans) 40.6 (38% linoleic) 2.3?? Carpenter et al.1976
Winterized soy oil 2.0% trans
Soy oil Margarine 16.0 39.0 (ca.22% trans) 21.0 1.3 78-90 Weihrauch 1977
(Stick type) & Carpenter 1973
Soy oil Margarine 13.0 36.2 (ca.14% trans) 27.0 2.1 92-130 Carpenter 1973
Soy oil shortening 25 47 (14% trans) 28 (5% trans) 70-81 Mattson 1975
Lard 44 46 10 (10% dienes) 63-69 Meyer et al. 1974
Butterfat 67 30 2.3 (2% dienes) 30-40 Meyer et al. 1974
Notes: Jones et al. (1965) found the trans content of hydrogenated winterized soy oil to be 15.1% and that soy oil shortening to be 36.8%, both much higher figures than those listed above. Perkins et al. (1977) found the 18:1 trans content of 12 commercial margarines to average 18% (range 6.9-30.8%).
A BRIEF HISTORY OF HYDROGENATION
The origins of the modern process of catalytic hydrogenation are found in the early 1800s. In the following sections we will introduce a number of references which do not mention soy oil, but which are important to understanding hydrogenation. Those which are not cited in our bibliography have the the symbol ^ following the author or date. Each of these is cited in references cited in our bibliography, and often noted at the end of the paragraph.
The earliest known use of the term "hydrogenate" was by Sir Humphrey Davy in 1809^??, in the sense of a substance combining with hydrogen, but not in the presence of a catalyst. The first scientific observation of a catalytic transformation (one that occurs because of the presence of a catalyst, which is not itself altered by the reaction) seems to be in 1811, when Kirchoff^?? showed that mineral acids in hot water solution change starch to dextrine and sugar without being themselves altered by the reaction. In 1817 Davy^?? observed the catalytic action of platinum in a gas. In 1845 Berzelius^?? in his masterful Treatise on Chemistry discussed these phenomena and first applied the term "catalytic" to them. He pointed out that this was an important yet largely unrecognized type of chemical reaction. The first recorded catalytic hydrogenation of an organic compound was by Debus in 1863^??; he produced methylamine by passing the vapor of hydrocyanic acid, mixed with hydrogen, over platinum black (Ellis 1930).
Hydrogenation's Early Years (1897-1939). But it was not until 1897 that the hydrogenation process began to be recognized as one of the major new chemical techniques. This was due to the brilliant research of two French food chemists, Paul Sabatier and his pupil, J.B. Senderens, who are usually credited as the founders of the modern hydrogenation process. In 1897^?? they discovered that the fluid consistency of vegetable oils is due to their having a lower hydrogen content than solid fats such as butter, tallow, and lard. Due to their work, nickel came to be the most widely used catalytic agent in all types of hydrogenation. Their investigations are recorded in a long series of articles, most of which were published between 1899 and 1902 in the Comptes Rendus of the Academie des Sciences and in the Bulletin de la Societe Chimique de France . In 1901 Sabatier was granted what was probably the first patent relating to the reduction of organic substances by hydrogen in the presence of a (nickel) catalyst (German patent y139,457^??; July).
Sabatier and Senderens did their hydrogenation work largely with vapors and substances that could be vaporized. The difficulty of volatizing liquid fatty acids and the practical impossibility of volatizing the oils themselves prevented them from converting liquid oils into plastic fats by the addition of hydrogen in the presence of a catalyst. Sabatier's main interest was to find a new way to harden liquid oils for use in soaps. In 1912 he received the Nobel Prize in chemistry for his work on catalysis. Sabatier summarized all this pioneering work in his book La Catalyse en Chimie Organique , which was translated into English by E.E. Reid in 1922 as Catalysis in Organic Chemistry . Other early pioneering work on fat hardening was done by Guido Goldschmidt of Vienna.
During the late 1800s, as solid fats became increasingly scarce and expensive (for reasons described later), it became the dream of the oil chemist to find a solution to the problem of converting oleic acid (a liquid) into stearic acid (a solid), or olein into stearin, simply by the addition of hydrogen, so as to make valuable hard fats from relatively inexpensive raw materials (Ellis 1930). The first actual catalytic hydrogenation of liquid oil to make solid fats, based on the preceding work of Sabatier and Senderens, was accomplished in 1902 by Wilhelm Normann, a German research scientist living in England. On 14 August 1902 (ref??) he applied for a patent on the process in Germany under the name of Herforder Maschinenfett und Oelfabrik Leprince & Sieveke (German patent 141,029) and on 21 January 1903 he applied for a patent in England under his own name (British Patent 1,515). In 1904 Ipatieff^?? helped to introduce modern hydrogenation methods by demonstrating that reactions were possible with high-pressure hydrogenation which could not occur under the conditions formerly employed.
Normann's patent would eventually revolutionize the world of oils and fats and change the eating habits of the world. But before it could do so, considerable work had to be done to make it commercially applicable. In 1905 Joseph Crosfield and Sons, Ltd. of Warrington, one of the three largest soap makers in England, made an agreement with Normann enabling them to work on perfecting the patented process, with Normann's help. By 1906 the first small runs of hydrogenated oils for soaps were made using the Normann process. From 1909, when the process was recognized as having vast commercial possibilities, Crosfield's began to buy the patent rights, first for the UK and then for many other countries. They hoped to sell these at a profit when the time was ripe. In 1909 the Dutch oil processing firm of Anton Jurgens N.V. bought the patent rights from Crosfield's for Germany and (for edible use only) for Great Britain. Along with these they also took Normann, who was transferred to Jurgens' research station. In 1911 Jurgens' Oelwerke Germania at Emmerich, Germany, began to produce hardened fats from whale oil, probably for soap. This was the first viable hydrogenated product on the European market.
Meanwhile other scientists were at work on the fat-hardening problem and by 1912 several other processes had been developed and patented, as by Testrup (a Swede), Erdmann (an Englishman), and Wilbuschweitz (a Russian). Licenses to these competing patents were quickly purchased by competing firms such as Lever Brothers and others. A great and complex scramble for control took place between major European soap and margarine manufacturers seeking inexpensive substitutes for traditional hard fats. From 1912 on there were major legal battles and complex treaties and agreements, all designed to set up a master patent and monopoly to control the powerful new fat hardening process; all of these failed (Wilson 1954; Stuyvenberg 1969). In 1913 Lewkowitsch (ref??), a leading authority on oils and fats in Europe, was able to write: "The production of solid fats from liquid oils by the direct addition of hydrogen has grown enormously of late years." He also mentioned the activity of copper as a hydrogen catalyst.
The greatest early potential for the use of hydrogenation lay in the United States, where a vast production of cottonseed oil was looking for new uses. In 1909 Procter & Gamble in Cincinnati acquired the US rights to the Normann patent from Crosfield's and in 1911 they began marketing Crisco, the first hydrogenated shortening, which contained a large amount of cottonseed oil. In America, however, six other firms had been working since 1915 according to the patents of C.E. Kayser (1910^??) and Carleton Ellis (1912^??), and with a number of other processes, most of which were never published. After a long period of litigation, initiated by Procter & Gamble, for alleged infringement of patent rights, a US court decision held the 1915^?? Burchenal patent (US Patent 1,135,351), under whose broad claims P&G's shortening was then being made, to be invalid. This opened the way for a number of firms to begin manufacture of hydrogenated shortenings and, from 1915, margarines. Eventually most of these came to be made with soy oil (Bailey 1951).
It is not clear?? when and for what purpose soy oil was first hydrogenated. It is known that in 1912 in the US 717 tonnes of soy oil were used to make shortening and 322 tonnes were used to make margarine; it is not stated whether or not the oil was hydrogenated. It probably was not. Also in 1912 Ellis (US Patent 1,047,013) recommended the use of hydrogenated soy oil to make shortenings of varying composition, but we do not know if his recommendation was put into practice. It may, however, have been the first recommendation. In 1914 Thompson, an American, reported that soy oil was widely hydrogenated in Europe to make soap and candles; this was confirmed the same year by a report in Seifensieder Zeitung (ref?? Soapboiler's Newspaper), concerning hydrogenated soy oil in German soaps. Thus soy oil was definitely being hydrogenated commercially in Europe by 1914 and probably several years earlier--but not yet for food use?? In 1918 Fox^?? noted that semi-drying oils such as soy oil could be "hydrogenized" and converted to non-drying oils for use in lubricants. He prepared a lubricating oil from hydrogenated soy oil mixed with 14% mineral oil (Ellis 1919).
By 1914 the hydrogenation industry was booming. That year Thompson reported that the capacity of European hydrogenation plants was estimated at 250,000 tons, which, however, was two or three times as much as had ever been treated. Located chiefly in England, Norway, Germany, and France, these plants hardened linseed, whale, soy, and cottonseed oils for use mainly in soap and candles. The great increase in demand for margarine in Europe and shortening (lard compounds) in the US, both of which traditionally used hard fats, and new techniques for deodorizing vegetable oils, would soon lead hydrogenated vegetable oils to be widely used in these two food products.
Before the use of hydrogenation, the production of shortening and margarine had been entirely dependent on animal fats as a source of raw materials. Increased demand soon caused these to grow scarce and expensive. Thus hydrogenation liberated shortening and margarine from their dependence on animal fats and made it possible for cooks to have products resembling lard and butter made from vegetable oils. Nevertheless it was not until after 1920 that hydrogenated vegetable oils were widely used in margarine and shortening. During the 1930s the use of hydrogenation worldwide took a quantum leap forward, as production increased greatly. Eventually advances in hydrogenation technology made it possible to produce a wide variety of hydrogenated products bearing some characteristics that were superior to those of butter or lard: lower P/S ratios, easy spreadability, low calorie products, lower cost, less cholesterol, a range of melting temperatures, and the like.
Hydrogenation was used in two ways to make margarine and shortening. Either all of the oil could be hydrogenated to the desired firmness, or a portion of it could be hydrogenated to a very firm consistency, then mixed with unhydrogenated oil. The former process was preferred in the early 1900s since it was thought to improve the oil's flavor and stability (Ellis 1919), yet later the latter came to be preferred, as it still is, since it is less expensive and more flexible, however, the unhydrogenated oil retains some of its flavor and it does not keep as well.
Two types of hydrogenation processes are widely used: selective and nonselective. The earliest studies of selective hydrogenation were done by Vavon (1911^??) in France and Paal (1912^??) in Germany. As early as 1917, Moore, Richard, and van Arsdel pointed out that hydrogenation under any conditions tends to be somewhat selective, preferentially hydrogenating the most unsaturated fatty acids, and that the final proportions of fatty acids depend on the pressure, temperature, amount of catalyst, and degree of agitation used. Theoretically, perfect selectivity in hydrogenation would convert, in orderly stepwise progression, all of the most unsaturated fatty acids in a fat to the next most unsaturated fatty acid. For example, all of the triunsaturated linolenic acid would be converted to the diunsaturated linoleic acid before it would hydrogenate any of the linoleic acid to monounsaturated oleic acid. Complete lack of selectivity would result in random hydrogenation of all double bonds. Selective hydrogenation is especially valuable for use on soy oil, since it permits the hydrogenation and thus elimination of a large proportion of the 5-11% linolenic acid contained in soy oil (undesirable because it causes flavor and oxidative insatability), without much reduction of the amount of essential polyunsaturated linoleic acid, and with little or no increase in saturated fatty acids (which must be removed by winterization or fractionation from cooking and salad oils). The most elaborate and painstaking research to date on selective hydrogenation was published by Richardson, Knuth, and Milligan in 1924 and 1925, together with a critical review of previous work. They noted problems with the formation of iso-oleic acid, now called trans oleic. Selective hydrogenation expanded in use during the 1940s and in 1949 Bailey gave an excellent description of the kinetics of the reaction, showing that the relative reaction rates under selective conditions for linolenic, linoleic, and oleic acid in soy oil were 100, 50, and 1 respectively. The nonselective rates were 12.5, 7.5, and 1 respectively. Thus under selective conditions, linolenic acid is hydrogenated twice as fast as linoleic and 100 times as fast as oleic. In 1954 Melnick and Deuel reported that most margarine was being made by selective hydrogenation and most shortening by nonselective. Selective conditions involved higher temperatures (160-205°C, vs. 121°C for nonselective), lower hydrogen pressures (6 vs. 50 psig), higher catalyst concentrations, special catalysts, and reduced agitation rates. Selective conditions were also known to produce about 50% more trans fatty acids. Margarine made by selective hydrogenation was reported to be more resistant to oxidative rancidity (because of its small proportion of linolenic acid and large proportion of trans fatty acids), to stand up well on the table at room temperature (because of the higher melting point of the trans acids), but to melt readily in the mouth.
During the 1960s there was renewed interest in the use of copper catalysts, which hydrogenate linolenic acid 15 (rather than two) times as fast as linoleic. In 1965 De Jonge^?? was granted the first patent for the use of copper catalysts with soy oil and that same year their selectivity was discovered and evaluated independently at the USDA Northern Regional Research Center. It was found that they could yield soy oil with only 20-25% as much linolenic acid as nickel catalysts in salad and cooking oils (Dutton 1982). Dutton noted that copper catalysts have been used commercially with soy oil in both the US and Europe, but use is not expanding; the higher concentrations (10 times as much as nickel) of this less active catalyst make it less economical. Moreover, some processors are reluctant to add copper as a catalyst when they have worked so hard to eliminate it from their processing systems (see Chapter 47). Dutton believes that copper will be widely used when continuous hydrogenation becomes popular.
A process that can also be used (in place of hydrogenation) to harden liquid oils or change the consistency of fats is interesterification, also called "molecular rearrangement." The process involves treating a fat with an excess of glycerol in the presence of a catalyst at relatively low temperature, causing the rearrangement or redistribution of the fatty acids on the glycerol portion of the molecule, and thus producing fats with different melting and crystallization characteristics than the parent. This rearrangement process does not change the degree of unsaturation of the fatty acids, nor convert cis into trans . Some of the earliest work on this process was done by Menschutkin in 1883^??, but the first patents for interesterification were granted to C. van Loon in 1920^?? and 1924^??. Prior to the late 1940s most of the interesterification was the undirected, random type, but thereafter it also came to be directed, and then began to receive increased attention (Bailey 1951). Interesterification came to be more widely used in Europe than in the US (where it was used to some extent in making shortenings); it is not permitted in France. In 1973 Unilever^?? filed a British patent application for a hardened, randomized margarine that reportedly contained only 3.2% trans fatty acids. By the mid-1970s European oil chemists considered large scale interesterification as "the greatest technological advance of recent years. It permits truly tailor-made fats to be produced, at the same time allowing almost limitless variation of the raw materials used" (Gander 1976). By 1976 some European margarines with zero trans fatty acid content and high polyunsaturate content were being produced by interesterifying a non-hydrogenated soy oil with about 20% of a soybean hardfat (fully hydrogenated), followed by the usual steps of emulsion preparation and solidification. At least one margarine was specifically advertised as being free of trans .