History of Soybean Crushing: Soy Oil and Soybean Meal - Part 6
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A Chapter from the Unpublished Manuscript, History of Soybeans and Soyfoods: 1100 B.C. to the 1980s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
Soybean Oil and Meal During World War II and the 1940s . World War II dramatically changed the position of the US in the world oils and fats economy, causing a major expansion of production and leading to a shift from importer to exporter. During the 1940s, as the war crippled Germany, Japan, and (to a lesser extent) Manchuria, the US emerged as the world's leader in soybean crushing and soybean oil and meal production (Fig. ??.?).
The outbreak of war between the US and Japan in December 1941 cut off US imports of oils and oilseeds (especially coconut and palm oils) from the Pacific islands (especially the Philippines) and other parts of East Asia controlled by the Japanese. These represented approximately two-thirds of the US prewar oil imports. Wars inevitably bring about a large increase in the need for edible oils and fats. Offsetting the curtailment of imports plus quadrupling exports to help supply the needs of US allies in Europe (especially the UK and the USSR) in addition to meeting increased basic demands required a dramatic increase in domestically produced oils. Most of this increase came from the soybean, since lard, cottonseed oil, and linseed oil were in short supply and thus high in price (Deasy 1941). Urgent US government programs were initiated to encourage farmers to grow more soybeans (Burtis 1950). In 1942 the USDA published a 4-page flyer entitled "Soybean Oil and the War: Grow More Soybeans for Victory." stating that
Uncle Sam needs soybean oil to win the war. We must replace a billion pounds (433,600 tonnes) of fats and oils cut off by the war in the Far East. Then, too, our Allies have asked us to send them a billion pounds or more of fats and oils this year . . . Most of this oil will go into food . . . Uncle Sam needs 9 million acres of soybeans for oil this year . . . Remember--when you grow more soybeans, you are helping America to destroy the enemies of freedom.
The response was amazing. Soy oil production increased by 62% the next year (1943) and passed second-place linseed oil. Then in 1944 it passed first-place cottonseed oil to become America's leading vegetable oil, a position it has maintained to this day (Fig. ??.?). During this period it also quickly replaced most of the coconut oil so widely used in margarines and cooking oils. The War Production Board restricted soy oil almost entirely to food uses. The most dramatic increase was in the use as shortening (Fig. ??.?). Large amounts of butter were also replaced by margarine, which included up to 30% soy oil.
Despite soy oil's expanding market, its flavor was singled out in 1945 as the "number one problem of the soybean industry," a problem that would limit future soybean production. Because of this flavor problem, it had wholesaled for 10-14% less than cottonseed oil during the period of 1935-1949 (Goldberg 1952). To deal with this crucial problem, in 1946 Edward Dies, Chairman of the Board of the National Processors Association, in conjunction with the Soybean Research Council, organized the first of what was to become annual Conferences on Flavor Stability in Soybean Oil. The historic conference, held 22 April 1946 at the Bismark hotel in Chicago, was attended by 28 top representatives of industrial, government, and academic research organizations. Research projects were initiated which yielded breakthrough discoveries during the next decade.
One major reason for calling the 1946 conference was to disseminate to the US soy oil industry a most unusual discovery, which had all the features of a cloak and dagger story. At the close of World War II, Warren H. Goss, a chemical engineer at the USDA Northern Regional Research Center (NRRC) at Peoria, Illinois, was assigned to follow General Patton's advancing tanks through Germany and to investigate the crippled German oilseed industry. He kept hearing about a secret process used to prevent "reversion" (called Umschlag in German) and the development of off flavors in the soy oil. In Hamburg he finally learned the keys to preventing soy oil reversion, a complex series of steps which he reported on in 1946. The basics were: (1) Use sound soybeans. (2) Use solvent extraction. Expellers, then widely used in the US, were said to burn the oil in such a way as to render subsequent treatments ineffective in preventing reversion. (3) Remove the lecithin by washing the oil twice with hot water, then removing the sludge by centrifugation. And (4) add a very small amount of citric acid to the oil during deodorization to inactivate any lecithin that remains in the oil. Goss (1946a, 1946b) stressed that nearly all German oil technologists considered lecithin to be the main cause of oil reversion. Linolenic acid, methyl- n -nonyl ketone, and traces of soap were also considered minor causes of quality deterioration. These steps were soon tested at the NRRC and found to work! But a great deal of research was done and some new breakthroughs were made before it was understood why they worked.
Although pioneering work in improving the flavor and stability of soy oil was done by researchers from private industry (major soy oil processors and refiners), universities, and the federal government, some of the best known and most important work was done at the USDA NRRC under the leadership of Drs. Cowan and Dutton. In 1947 they developed objective methods to assess oil flavor and color, combining taste panel tests with statistical analysis. Armed with these new tools, between 1947 and 1951, they showed that minute traces of metals (as little as 0.3 parts per million of iron or 0.01 ppm of copper) could catalyze oxidation in the soy oil and ruin its flavor. By 1952 iron and brass valves and other oil-contact parts had been removed from refineries and replaced by stainless steel ones. Also in 1947 the NRRC found the answer to the puzzle explaining why the complex German process largely eliminated off-flavor development. The key to the process was the addition of the citric acid which, the NRRC showed, inadvertently served to scavenge, complex, and inactivate traces of iron and copper that would otherwise have caused oxidation and off flavors. Thus the new focus was not on complete removal of lecithin but on complete removal of traces of prooxidant metals. Recognition that oxidation caused the off flavors and that citric acid inactivated the effect in prooxidant metals was slow in being accepted because often less than 0.03 ppm of copper or iron were involved. Before long, however, a variety of metal deactivators were being added to all soy oil products. Also in the late 1940s the NRRC showed that oxidation and hence off-flavor development could be reduced by blanketing soy oil with an inert gas (such as nitrogen) at all critical, high-temperature steps, including final packaging. Special care was taken to prevent oxidation during deodorization, since the oil is heated to as much as 260*C (500*F) and held there for 15-40 minutes. It was shown that there are at least two types of off flavors developed in soy oil: those caused by oxidative rancidity occurring at high levels of oxidation and those caused by "flavor reversion" at very low levels of oxidation. In flavor reversion, beany and grassy flavors develop at the early stages, then fishy and painty flavors at more advanced stages. However, since none of these flavors exist in crude soy oil, the concept of "reversion" is inaccurate and a misnomer, which nevertheless persists (Dimler 1968; Dutton 1981).
Despite the various new insights and the effectiveness of the several treatments described above in improving soy oil flavor, it was still not known by the early 1950s what mechanism was actually responsible for the off flavors. As early as 1936 Durkee had speculated that linolenic acid (a tri-unsaturated fatty acid that derives its name from "linseed" oil) in the soybean was the cause of the problem and that selective hydrogenation would be the solution. The linolenic question was debated for many years but in 1951, in a now-classic experiment, Herbert H. Dutton and co-workers at the NRRC proved convincingly that linolenic acid was the primary precursor of off flavor development and instability in soy oil. The high degree of unsaturation leads to oxidation, causing the off flavors and odors to develop. Today, the three major problems inherent in soy oil are said to be its high level of linolenic acid (7-8%, which can be reduced to below 3% by hydrogenation) and phosphatides (3%, which are removed by degumming), and the fact that the hydrogenated oil is a beta crystal former (why is this a problem??). In the early 1950s work began in the soy oil industry on three fronts to get rid of the newly identified culprit, linolenic acid: reacting, extracting, or breeding it out. Reacting it out was found to be the most practical short-range approach. After 1947 the researchers found catalysts that would react selectively during hydrogenation with the linolenic acid but not with the desired essential polyunsaturated fatty acid, linoleic. Thus the oil was deodorized with citric acid, mildly hydrogenated to reduce the linolenic content from an average 7 or 8% down to 3%, packed in a brown glass bottle to protect it from rancidifying fluorescent light and metals, and sealed blanketed with inert gas. Later copper catalysts were found that would hydrogenate linolenic acid some 15-20 times more rapidly than linoleic, thus making it possible to produce "copper hydrogenated soy oil" with essentially zero linolenic acid content; this was an important step in improving soy oil flavor (Dutton 1976). As of 1980 the flavor problems once associated with soy oil have been largely solved, except when the oil is used at high temperatures, as in frying. Attempts to breed out linolenic acid have been partly successful and levels were reduced by 50% between 1940 and 1980 (Erickson et al. 1980). Although as little as 4.3% linolenic acid is found in some soybean strains, further reductions from the present 7-8% average are expected to proceed slowly for a number of complicating genetic reasons.
During the 1940s a sudden and extremely important change took place in the basic methods for removing oil from soybeans. Although continuous screw presses paced the expansion of soybean crushing during World War II, solvent extraction systems took the lead after that time, passing screw presses during 1949 in terms of total amount of soybeans crushed. By the 1950s the use of hydraulic pressing had decreased to less than 3% of the total soybean crush.
Before examining why solvent extraction suddenly became popular in the US during the 1940s, let us take a brief look at the characteristics and history of the three basic methods of oil removal used in the US (and worldwide) during that decade:
Hydraulic presses. The hydraulic press was invented in 1796 by Joseph Bramah, an English engineer. It quickly revolutionized the Western oil milling industry, replacing cider-type presses powered by draft animals. Starting in the late 1800s it was widely used by the Manchurian soybean industry, replacing hand-turned screw presses and wedge presses. The hydraulic press was the first press used to crush soybeans in the US (in 1911 in Seattle). Prior to that time it had been widely used in the US to crush cottonseed and flaxseed, but the operators of these mills soon discovered that it could be easily adapted to soybeans. For hydraulic pressing, the beans were ground or flaked, steamed and tempered in a cooker, then formed into a cake in press cloths and pressed at 500-4,000 psi. From about 1911 until the early 1930s hydraulic presses were fairly widely adapted to soy oil processing, although their percentage of total processing capacity during this period was probably less than 25%. Burtis (1950) stated that they never constituted more than a small fraction of total processing capacity. During World War II, when a serious shortage of processing capacity occurred in the northern and central states, soybeans from these states were transported to cottonseed areas of the south and the soy oil was extracted there by hydraulic presses as an emergency measure. By the late 1940s, hydraulic presses were almost obsolete in the US soybean industry, largely because of the high labor costs involved in loading and unloading the presses, the limited capacity of the batch process, and the low extraction efficiency.
Screw presses and expellers. Screw presses were developed in the US in the early 1900s to meet the need for continuous process oil extraction equipment. The first and still the most popular model was the Anderson expeller. It was developed in 19?? (the term "expeller" was coined by the V.D. Anderson Co. in Cleveland, Ohio to describe the screw presses they made), then followed by the French screw press (made by the French Oil Mill Machinery Co., in Piqua, Ohio). Anderson expellers were first used to crush soybeans experimentally in 1909, then commercially in 1915 in North Carolina. Expellers and screw presses dominated the US soybean crushing industry after this time, being used by most of the early crushers including Chicago Heights (1917-18), A.E. Staley (1922-) and Funk Brothers (1924-). From 1915-1945 an estimated 60-80% of the soybean crush was done with screw presses and expellers. In a screw press, crushed?? and presteamed (tempered) soybeans are fed into the machine so that a continuously revolving worm shaft (somewhat like that of a kitchen meat grinder) forces them through a strong metal barrel or cage, whose walls are perforated with many thin slots, each about 0.01 inch (0.2-0.3 mm) wide. Under pressure of 5-10 tons or more per inch and temperatures for friction and pressure reaching 150*C (302*F), the oil is pressed out and drains through the narrow slots. A choke nut at the discharge end of the press allows the operator to control the pressure and temperature. The higher the pressure, the higher the temperature and the more oil expressed. At the discharge end, the cake emerges in thin sheets, which are broken up by a revolving cake breaker.
Screw presses had numerous advantages over hydraulic presses: they processed continuously, were much less labor intensive, eliminated the need for press cloths, required less heating before pressing, had a larger capacity, and extracted more oil. Disadvantages: they required more mechanical power input and more skilled manpower. Compared with solvent extractors, they darkened the oil somewhat and extracted only 77% of the oil versus 95% for solvent. Even by the 1980s a few of the older cottonseed mills and some of the companies producing soy oil for the natural and health food trades still used screw presses for crushing soybeans.
Solvent extractors. The early history of solvent extraction and its widespread use in soy oil processing, first in England, then in Germany, is described under "Europe" above. Early use of solvent extraction and of alcohol solvents in Manchuria and Japan is also described earlier. Solvent extraction was much slower to catch on in the US than in Europe for two reasons: First, prior to the late 1930s, there was generally a lack of a sufficiently large quantity of soybeans in any given location to make the process cost effective. Europe, which imported large quantities of soybeans, could locate solvent plants in major port cities. But in the US, where soybeans were widely grown, small, decentralized mills typically processed the beans grown in their immediate vicinity. And second, several disastrous explosions in the early 1930s (as at the Glidden plant in 1936) caused by the ignition of hexane vapors, dampened interest in the solvent process. Several attempts, starting in 1923 (as described earlier), to use solvent extraction in the US proved unsuccessful. Henry Ford built small solvent extraction systems starting in 1932, and hoped that many farmers would install them in their barns to create a farm-based industry processing their soybeans. Work with this concept of "community solvent plants" was carried on by Iowa State College (Arnold 1941).
The first successful solvent operation in the US was started in 1934 by the Archer Daniels Midland Co. (ADM) in Chicago. The William O. Goodrich Co., acquired by ADM in 1928, had been doing experiments with solvent extraction since 1926. Like other US soybean crushers, ADM drew heavily on German experience and technology, sending engineers to Germany to study the mills there (see Chapter 40). ADM purchased a German Hildebrandt continuous immersion extractor which could process 91 tonnes (3,000 bushels) of soybeans a day. Other early companies to use solvent extraction were The Glidden Company (1934; see chapter 68.2) and Central Soya, whose grand opening was in 1937 (see Chapter 40).
During the 1930s, most US solvent plants used benzene (then spelled benzine and also called benzol) or a special high-test gasoline as a solvent (Burlison 1936). In 1935, however, Sorenson and Beal were granted a patent for the use of hexane solvent (US patent 2,024,398) which, by 1941 served to establish hexane as the most widely used US solvent. A purified petroleum hydrocarbon fraction (C6H14), hexane is a colorless liquid with a pleasant odor, which was typically used at 58*C (136*F), boils at 63-70*C (146-158*F), has a relative density or specific gravity of 0.685, and is highly volatile and flammable, yet low in cost (Goss 1941). Hexane is toxic to humans. It is mildly depressant to the central nervous system and may irritate the lungs if inhaled. Because of its flammability and toxicity, its use requires extreme care. Two other solvents that attracted attention during the late 1940s were naptha and trichloroethylene (TCE). TCE, it will be recalled, had been shown to cause livestock deaths in Europe as early as 1916. In 1923 Piper and Morse had published prominent warnings against its use. But these problems had been largely forgotten by the 1940s when renewed interest in TCE developed; because it was nonflammable and nonexplosive, the cost of building oil mills that used it was very economical, although the solvent itself was considerably more expensive then hexane. By 1944 TCE was being used commercially in the US (Duncan 1948), and by 1952 TCE-extracted soybean meal had reached an estimated 2% of all US soybean meal. However at about this time its use was shown to cause aplastic anemia in cattle, and between 1955 and 1960 its use was discontinued (Altschul 1958). Thereafter hexane came to be virtually the sole solvent used in the US because of its low cost, ease of recovery, high oil yield, and selectivity for vegetable oils.
Another solvent that attracted considerable interest during the 1940s, especially on a noncommercial basis among US researchers, was ethanol (ethyl alcohol). As noted earlier research on its use had started in Manchuria in 1927 and it was being used there commercially by 1937. In 1948 Beckel, Belter, and Smith of the USDA Northern Regional Research Center reported that ethanol-extracted meal and oil both had better flavor and lighter (better) color than their hexane extracted equivalents, and that food products using flour ground from the meal (breads, meringues, whips, and candies) were superior in flavor, color, and nutritional value. Unlike hexane, ethanol was a natural, nontoxic food (the same found in beer or wine) and a renewable resource derivable from farm crops by fermentation. In cooperation with an industrial firm, the NRRC set up a pilot plant to further test the alcohol extraction process. The firm used the process commercially to make Gelsoy, a refined soy protein gelling substance?? Other advantages of ethanol were found to be the ease of by-product recovery (the oil and ethanol separate by just allowing them to cool) and the fact that ethanol can be removed from soybean meal at a lower temperature than hexane. Disadvantages were that soy oil is not very soluble in ethanol except at high temperatures, the high latent heat of evaporation of ethanol and the desired processing at 20 psig requires additional fuel for heating, the high cost of the solvent (33% more expensive than hexane in 1980), the necessity of drying the flakes to 3% moisture in the recovery process, the difficulty of maintaining high solvent purity, the high initial cost of equipment or changeover, and the stringent regulations established by the Internal Revenue Service on use of alcohol. However after 1970 there was renewed interest in using ethanol primarily for its ability to make superior quality meals and better tasting soy flours, concentrates, and isolates (Eldridge et al. 1971; Honig et al. 1976), but also because it may eventually become less expensive than petroleum-derived hexane. Trade sources say that by the 1970s A.E. Staley Mfg. Co. was using an ethanol-hexane solvent in a patented process.
There were a number of reasons that solvent extraction (particularly hexane) systems replaced mechanical pressing systems for soy oil removal during the 1940s. Many of these reasons could be foreseen by the 1930s and 1940s (Stewart et al. 1932; Horvath 1927 and 1938a; Goldberg 1952). First, solvent extraction gave higher yields of oil. While mechanical pressing systems recovered only 70-80% of the oil present in the soybean (leaving 4-6% oil in the presscake), solvent systems were able to recover about 95% of the oil (leaving only 0.5-1% in the meal). This resulted in a gain of 1.5 pounds of oil from every 60-pounds bushel of soybeans or 2.5 pounds from every 100 pounds of soybeans. It also resulted in fewer pounds of meal from each bushel, but since oil was worth 4-5 pounds as much as meal on a weight basis, there was a clear net gain. Second, solvent systems had a larger processing capacity. Third, solvent systems, run at relatively high capacities, produced oil and meal at lower cost, largely because of lower labor and overhead costs. Fourth, solvent systems used less energy. And fifth, they gave oil that many considered to be superior in quality since it had superior bleaching qualities, lower refining losses, reduced susceptibility to rancidity, and better retention of fat-soluble vitamins. Problems with solvent extraction were the huge capital investment required to build a plant, the danger (with hexane) of explosions, the need for a large and steady supply of soybeans, the loss of 1-2 gallons of increasingly expensive solvent for each ton of soybeans processed, and the fact that small amounts of the heavy fractions of the solvent (especially in the early days) might remain in the meal and oil. On this last point Burlison wrote in 1936: "Objection has been made that some of the solvent remains in the meal. While efficiently operated modern plants can remove the last trace of solvent, mills in the US have not yet pushed the process to this stage of perfection."
The solvent process involves preparation of the beans and crushing to flakes, extraction of oil from the flakes with solvent, and reclamation of the solvent from the oil and meal. The two basic types of systems are percolation extractors (including rotary, stationary basket, chain and basket, perforated belt type, etc.) and immersion extractors. By the 1970s percolation extractors, in which a current of solvent percolated down through the flakes in a porous container, had largely replaced immersion extractors, in which the flakes are immersed in a bath of solvent.
Solvent extraction systems imported from Germany were the only ones used in the US until 1936. The Hansa-Muehle or Bollmann "paternoster" basket type was the most popular, followed by the Hildebrandt "U-type." In 1936 Henry Ford developed the first commercial American-made solvent system (see Chapter 46), but it was not used by large processors. In 1937 large vertical gravity extraction column extractors were introduced by Allis-Chalmers Manufacturing Co. and the V.D. Anderson Co. Also during the 1940s modifications of the Hansa-Muehle or Bollmann extractor were built in the US by the French Oil Mill Machinery Co. and the Blaw-Knox Co., which developed the Rotocel, an improved rotary-basket percolation type, now owned by Dravo. Crown Iron Works also had a system by 1948, and De Smet of Belgium entered the field later. All solvent extractors made after 1944 were continuous. Since solvent extractors first passed screw presses in 1950 to become America's most widely used method of removing soy oil, solvent extraction was still considered a "new process" during the 1950s, and hexane was still considered a relatively new solvent. In 1951 the solvent extraction systems processing the largest percentages of all US soybeans were French (26.5%), Blaw-Knox (11.0%), Anderson (5.3%), and Allis-Chalmers (5.3%) (Goldberg 1952). By 1957 more than 94% of the US soybean crush was solvent extracted, and after 1972 more than 95% was (Fig. ??.?). One additional advantage of solvent extraction, first recognized during the late 1950s and 1960s with the development of soy protein concentrates and isolates, was the mild heating involved, which caused relatively little protein denaturation.
The amount of oil recovered from a given weight of soybeans has gradually increased because of both improved extraction technology and new soybean varieties bred to contain more oil. In 1924, 100 pounds of soybeans yielded 12.3 pounds of crude soy oil, which increased to 14.2 pounds in 1930, 15.5 pounds in 1938, 16.3 pounds in 1948, and 17.8 to 18.5 pounds during the 1980s. This represents an increase of 48.8% during this 57-year period (Fig. ??.?). In 1981 100 pounds of soybeans yielded typically 72.2 pounds of soybean meal (44% protein), 18.3 pounds of crude soy oil, 7.0 pounds of hulls, and 2.5 pounds of shrink or manufacturing loss (DWB??).
Modern Soybean Crushing and Oil Refining Processes . Having discussed the shift from mechanical pressing to solvent extraction and the advances made in soy oil processing, let us take a more detailed look at the steps involved in the solvent extraction process and the subsequent degumming and refining processes, now used by most of the industry. This will help us to understand a number of key developments from 1950 on.
1. Preparation of Soybeans for Extraction. Soybeans are cleaned and dried in a drier to reduce their moisture to about 10%. Using a series of corrugated steel rolls each bean is cracked into 2-8 pieces, keeping the pieces as large as possible. The hulls, representing about 6-1/2% of the total weight of the beans, are removed by aspiration (usually suction); they are then called "soybean mill run." They are usually toasted, ground, and used separately in animal feeds, or they may be added back into the defatted flakes later. The cracked cotyledons are now conditioned/tempered with mild dry heat at about 75*C (167*F) for 15-20 minutes, to make them more plastic, less brittle. Then they are passed through flaking rolls to form very thin (0.2-0.3 mm-thick) flakes, each about the size of the lengthwise cross section of a peanut. During flaking, most of the soybean cells are disrupted, liberating most of the oil and facilitating its extraction with solvent.
2. Extraction of Oil from Flakes with Hexane Solvent. The flakes are run into a percolation type extraction, typically made up of a series of cells or baskets that rotate about a vertical axis and have perforated, openable bottoms. (A typical large extractor can handle 2,000-4,500 tonnes of dry beans per 24 hours.) A current of hexane solvent is percolated down through the flakes and soy oil dissolves in it to form a miscella (mixture of oil and solvent), which is drained off then used to wash the flakes three more times. The final miscella contains 25-30% oil. The flakes, given a final wash with raw solvent, have had about 95% of their oil extracted.
3. Reclamation of Solvent from Oil and Flakes. The solvent is stripped from the crude soy oil at about 135*C (275*F) under a strong vacuum, then the solvent vapor is condensed and recycled for further extraction use. The crude oil, which has a dark color and oily mouthfeel, is now ready for degumming and refining.
For typical feed use, the solvent-laden flakes are passed down through a desolventizer-toaster (DT), which takes out the hexane and "toasts" the flakes with moist heat. In the DT, a tower of steam-jacketed compartments, the hexane is stripped off by heating the flakes, first with live steam in the top compartment, then with dry heat at up to 110*C thereafter. The heat also inactivates antinutritional factors (primarily urease and lipoxygenase) in the flakes. After being dried and cooled, the flakes can be mixed with the soybean hulls (seed coats, removed during cracking) then, using a hammermill, ground to a meal, which contains 44% protein. Adding back the hulls lowers the crusher's costs per unit weight of meal produced. If the hulls are not added back, the meal will contain 49% protein, and sell for a slightly higher price. The meal is then sold to livestock feed compounders; the 49% protein meal is used mostly in broiler feeds.
If the solvent-laden flakes are to be converted to foods, they are desolventized and toasted slightly differently, depending on the nitrogen solubility index (NSI) required, as explained in Chapter 29.
The crude soy oil is now ready to be degummed and refined. Each step in these processes is designed to remove undesirable components from the oil, leaving a finished product containing more than 99% pure triglycerides.
4. Oil Degumming and Lecithin Separation. Most soybean crushers degum their crude soy oil before it is shipped to a refinery, unless the refinery produces commercial lecithin. Shipping is usually done using bulk rail cars in the US or tankers overseas. Degumming removes phosphatides (including lecithin) and mucilaginous gums, which would form a heavy sludge during storage and shipment of the crude oil; in a refined oil they would inhibit hydrogenation and cause smoking during frying. Moreover, lecithin is now a valuable by-product. In water degumming about 1-3% water is added to the crude oil at a temperature of 60-70*C. The phosphatides form an oil-insoluble sludge, which is continually separated from the lighter oil by centrifugal force. The phosphatides in the sludge are refined to make lecithin, as discussed in the next chapter. There are also other methods of degumming. A small proportion of the degummed oil is used directly as drying oils in paints, varnishes, and printing inks. But most is sent to oil refiners; major refiners include Unilever, Procter & Gamble, Hunt-Wesson Foods, and Kraftko's Humko subsidiary. These food firms refine the "crude degummed" oil then convert it into a variety of brand-name soy oil products ready for sale to consumers.
5. Alkali Refining. Also called "caustic refining," "alkali deacidification," or "neutralization and water washing," this continuous process removes undesirable free fatty acids, color bodies (which color the oil), prooxidant metals, and traces of phosphatides, carbohydrate, and protein. Unfortunately, it also removes 50-75% of the tocopherols (vitamin E), natural antioxidants found in soy oil that make it quite stable, retarding rancidification. The oil at 60-70*C (140-160*F) is washed with alkaline water, typically containing sodium hydroxide (caustic soda) and/or sodium carbonate. The alkali combines with and neutralizes the free fatty acids to form soapstock, which is removed by centrifugation. Accounting for about 6% of the volume of the original oil, the soapstock is typically acidulated then used in livestock feeds or to make soaps, fatty acids, or glycerine. A small amount of soy oil is physically refined or steam refined, as described later.
6. Bleaching. Bleaching oils is not a chemical process like bleaching cloth with Clorox. It is a physical process, resembling clarification. Typical "bleaching" agents are fuller's earth (hydrated aluminum silicate) or clays (composed mainly of bentonite or montmorillonite), which are usually "activated" by treatment with sulfuric or hydrochloric acid to increase their adsorptive capacity. The acid-activated earths or clays are added to the oil, usually under vacuum and 105-115*C (220-240*F), in a continuous process. Color bodies or pigments stick in a very thin layer to the surface of these bleaching earth particles, which are then centrifuged out of the oil. Bleaching also removes traces of soap and improves the oil's color, flavor, and oxidative stability.
7. Hydrogenation. Although not technically considered a part of the refining process, this step which involves blowing hydrogen gas into the hot oil in the presence of a nickel catalyst, reduces the oil's linolenic acid content, which greatly improves the flavor and stability. Since hydrogenation also imparts off flavors, it is usually followed by deodorization.
8. Deodorization. During deodorization the oil is heated to its highest temperatures found during the entire crushing and refining process, 220-260*C (392-500*F). Steam is bubbled and blown up through the hot oil under a high vacuum; the higher the temperature of the oil, the shorter the processing time. This strips off undesirable volatiles, including beany flavors and odors, remaining free fatty acids, peroxides, and most or all of any residual pesticides or herbicides. (Hunter reported in 1981 that crude soy oil has been found to contain 220 parts per billion (i.e. per 1,000 million) of Dieldrin, 40 ppb of Heptachlorepoxide, 19 ppb of DDT, 10 ppb of Aldrin, or 10 ppb of Heptachlor. These pesticides are translocated from the soil to the seeds). Deodorization also improves the oil's stability. The oil is then cooled and treated with citric acid, which acts as a metal scavenger (mainly for iron and copper) to prevent oxidative rancidity. Finally the oil is filtered using diatomaceous earth (daitomite). Typically an antioxidant such as BHA (butylated hydroxyanisole) or BHT (butylated hydroxytoluene) and sometimes a defoaming agent (methyl silicone) is added. The resulting bland product tastes quite similar to many other refined oils and can be used interchangeably with them.
Soy Oil Products. The various processes described above are used in various combinations to produce the basic popular soy oil products.
To make salad and cooking oils , refined and bleached soy oil is lightly hydrogenated to reduce its iodine value from 130 down to about 110-115. It is then winterized by chilling it for 3 days so that hard fats such as stearine that are formed during partial hydrogenation settle out and can be removed by filtration. Winterization prevents the oil from clouding in home refrigerators. The stearine is used later in making shortenings. After deodorization, the finished oil, known as "lightly hydrogenated winterized soy oil" (LHWSO) is used for salad and cooking oils. It has a polyunsaturated/saturated fatty acid ratio of about 2.3 and contains about 12% trans fatty acids. This is the most popular liquid oil in the US.
To make margarine or shortening , hydrogenated soy oil is more fully hydrogenated to lower the iodine value to 55-100, depending on the desired application. A blend of oils are then deodorized and made into margarine or shortening.