History of Palm Oil
From: The Cambridge World History of Food 2 volume boxed set
Edited by Kenneth F. Kiple
Bowling Green State University, Ohio
Kriemhild Coneè Ornelas
Published October 2000
The oil palm (Elaeis guineensis) is a native of West Africa. It flourishes in the humid tropics in groves of varying density, mainly in the coastal belt between 10 degrees north latitude and 10 degrees south latitude. It is also found up to 20 degrees south latitude in Central and East Africa and Madagascar in isolated localities with a suitable rainfall. It grows on relatively open ground and, therefore, originally spread along the banks of rivers and later on land cleared by humans for long-fallow cultivation (Hartley 1988: 5—7).
The palm fruit develops in dense bunches weighing 10 kilograms (kg) or more and containing more than a thousand individual fruits similar in size to a small plum. Palm oil is obtained from the flesh of the fruit and probably formed part of the food supply of the indigenous populations long before recorded history. It may also have been traded overland, since archaeological evidence indicates that palm oil was most likely available in ancient Egypt. The excavation of an early tomb at Abydos, dated to 3000 B.C., yielded “a mass of several Kilograms still in the shape of the vessel which contained it” (Friedel 1897).
A sample of the tomb material was submitted to careful chemical analysis and found to consist mainly of palmitic acid, glycerol in the combined and free state, and a mixture of azelaic and pimelic acids. The latter compounds are normal oxidation products of fatty acids, and the analyst concluded that the original material was probably palm oil, partly hydrolyzed and oxidized during its long storage. In view of the rather large quantity found, the oil was probably intended for dietary purposes rather than as an unguent.
A few written records of the local food use of a palm oil (presumably from Elaeis guineensis) are available in accounts of European travelers to West Africa from the middle of the fifteenth century. Red palm oil later became an important item in the provisioning trade supplying the caravans and ships of the Atlantic slave trade, and it still remains a popular foodstuff among people of African descent in the Bahia region of Brazil (Northrup 1978: 178—86; Hartley 1988: 1—3; R. Lago, personal communication, 1993).
The British Industrial Revolution created a demand for palm oil for candle making and as a lubricant for machinery. In the early nineteenth century, West African farmers began to supply a modest export trade, as well as producing palm oil for their own food needs. After 1900, European-run plantations were established in Central Africa and Southeast Asia, and the world trade in palm oil continued to grow slowly, reaching a level of 250,000 tonnes (metric tons) per annum by 1930 (Empire Marketing Board 1932: 117—23; Hartley 1988: 8—23; Lynn 1989: 227—31).
Meanwhile, the invention of the hydrogenation process for oils and fats in 1902 created the possibility of Western employment of palm products as, for example, in the making of margarine. Yet hydrogenation was more useful for liquid oils like groundnut, palm kernel, and coconut oils than for palm oil. After World War II, further improvements in palm oil refining technology and transport methods made it possible to use largely unhydrogenated palm oil in Western food products (Lim 1967: 130—2; Martin 1988: 45—8).
A rapid expansion of the palm oil export trade followed, accompanied by a marked growth in the plantation sector of production. Between 1962 and 1982, world exports of palm oil rose from about 500,000 to 2,400,000 million tonnes per annum, and Malaysia emerged as the world’s largest producer, accounting for 56 percent of world production and 85 percent of world exports of palm oil in 1982. Expanded production in Malaysia was achieved mainly by the privately owned estate sector, which increased its oil palm holdings more than tenfold in the 1960s and 1970s; and by the Federal Land Development Authority (FELDA), whose large-scale schemes organized oil production along plantation lines, although ownership was vested in the workforce of “smallholders” (Khera 1976: 183—5; Moll 1987: 140—62).
By 1990, world production had reached nearly 11,000,000 tonnes per annum, with a worldwide trade of 8,500,000 tonnes (Mielke 1991: 110). Although red palm oil is still used in soups and baked dishes in West Africa, elsewhere in the world, palm oil is eaten mainly in a highly refined form. Its food uses vary from the vanaspati and ghee of India to the margarine, cooking oils, and biscuits of Europe and the United States.
The Oil Palm: Wild and Planted
West Africa is the classic region of smallholder production, both of food and export crops. The oil palm, which has been both, flourishes in natural association with yam and cassava cultivation throughout the wetter parts of the region. In eastern Nigeria, which C. W. S. Hartley (1988: 7, 16) called “the greatest grove area of Africa,” densities of 200 palms per hectare (ha) were common in the late 1940s, and densities of more than 300 palms per hectare were not unknown.
These palms were typically self-seeded and tended (to varying degrees) by local farmers. Farther west, in the kingdom of Dahomey and in settlements established by the Krobo people near Accra, some deliberate plantings may have been made as the palm oil export trade developed from the 1830s (Manning 1982; Wilson 1991). However, as J. Reid (1986: 211) has noted, the word plantation was often used by contemporary European observers to mean a food farm on which oil palms happened to be growing. Moreover, although in Dahomey descriptions exist of seedlings being transplanted from the bush onto areas cleared for farming by slaves, this does not mean that the practice was universal. In any event, palm oil exports from Dahomey were much smaller than from the Niger Delta, where oil palms were planted deliberately in swampy regions outside their natural habitat, but where the bulk of production was carried out using natural groves. In the 1840s, Dahomey and the Niger Delta exported approximately 1,000 and 13,000 tonnes per annum respectively; by the 1880s these totals had risen to 5,000 and 20,000 (Manning 1982: 352; Reid 1986: 158; Martin 1988: 28—9; Lynn 1989: 241).
Beginning in the late nineteenth century, a number of experimental oil palm plantations were created by Europeans in West and west-central Africa. One of the earliest was founded in Gabon in 1870 by Roman Catholic missionaries (E. J. Mulder, personal communication, 1968). But like many of the other nineteenth-century plantations in West Africa, these ventures were unsuccessful. By comparison with African smallholders, European planters were highly specialized and vulnerable to the marked trading fluctuations of the late nineteenth century. Many also lacked capital but committed themselves to long-term investments of a type that African farmers sensibly rejected.
In the case of palm oil, money was spent paying laborers to produce, plant, and tend seedlings, often on marginal land, in regions where natural groves already contained more palms than local farmers could spare the time to harvest (Hopkins 1973: 212—14; Martin 1988: 46).
Thus, when in 1907 William Lever sought large-scale land concessions in the British West African colonies in order to produce palm oil for his Lancashire soap mills, the Colonial Office was reluctant to help him. In a region characterized by small, fragmented, and often communally owned farms, it was felt that Lever’s scheme would be hard to administer, politically risky, and commercially unsound. Lever was left to pursue his dreams in the Belgian Congo, where the existing levels of both trade and population were far lower and where the colonial administration welcomed European enterprise (Hancock 1942: 188—200; Wilson 1954: 159—67).
Following the Lever debate, the West African palm oil industry remained in the smallholders’ hands. Few other entrepreneurs came forward to press the case for plantations, although a number of state-run estates were established under French influence in the Ivory Coast after 1960. By 1981 these estates covered a total of 52,000 ha, with a further 33,000 ha planted with oil palms in the surrounding villages (Hartley 1988: 31).
Yet even this development was relatively modest in scale, as shown in unpublished data from Nigeria, West Africa’s largest producer of palm oil. The area of wild palm groves, only partly harvested, was estimated at 2,400,000 ha, whereas there were 72,000 ha of estate plantations and another 97,000 ha of smallholder plantations. Estate plantations, which require large consolidated areas, are still difficult to create in Nigeria because the oil palm—growing regions are densely populated and the complex traditional land holding system has been carefully preserved. Elsewhere in West Africa, population densities are lower, but the problems of obtaining labor to sustain plantation developments are correspondingly greater.
In the late nineteenth century, both the German colonizers of Kamerun and the Belgian rulers of the Congo were keenly interested in applying European farming and processing techniques to the palm oil industry (Laurent 1912—13; Rudin 1938; Jewsiewicki 1983). But German botanical and mechanical trials were cut short by World War I, following which the German territories in Africa were divided between the French and the British.
In the Congo, however, Lever’s initial land- and produce-buying concessions (granted in 1911) proved to be the foundation for a long process of experimentation, which eventually revolutionized the palm oil industry worldwide. New planting materials led to dramatic increases in yields, thus cutting the cost of production; and improved machinery led to high oil quality at a competitive price. Alongside developments in European and American food-processing techniques, the Congo innovations paved the way for the entry of palm oil into Western diets.
Lever was originally more interested in setting up mills than plantations in the Congo, but his initial investments brought heavy losses (Fieldhouse 1978: 507—9). The fruit supply from wild trees proved hard to control, both in the amount brought to the mill and in its quality upon arrival. Overripe or bruised palm fruit made for highly acidic, low-quality oil, whereas unripe bunches gave low yields. Yet Lever Brothers (and its successor Unilever after 1929) was unwilling to incur the heavy initial costs of planting trees unless planting materials were improved to reduce the running costs. The Germans in Kamerun had identified an exceptionally thin-shelled palm fruit with a high oil content as early as 1902 (Hartley 1988: 50). But their “lisombe” palm, later to become known as the Tenera type, was found only rarely in the wild and failed to breed true.
In a renewed drive to encourage European investment in their colony and, in particular, in oil palm plantations, the Belgians began in 1922 to investigate this German discovery. An experimental plantation of Tenera palms was created at the Yangambi research station in the Congo, and in the 1930s these palms were subjected to a three-year testing program by M. Beirnaert. Meanwhile, private Tenera plantings had been made by Unilever and its subsidiary, the United Africa Company, in the British Cameroons and in the Belgian Congo itself (Courade 1977; Fieldhouse 1978).
Tenera seed also found its way to Sumatra and Malaya in the 1920s, although there, as in Central Africa, it failed to breed true. Beirnaert’s painstaking experiments finally showed why: The Tenera palm was actually a hybrid of two other types, the thick-shelled Dura and shell-less Pisifera, and when self-pollinated would produce 50 percent Tenera, 25 percent Dura and 25 percent Pisifera (Beirnaert and Vanderweyen 1941).
Beirnaert’s discovery was published at the height of World War II, and it was not until after 1945 that it could be turned to practical use with the establishment of large-scale and long-term Tenera breeding programs. It is ironic that the Congo was not the state that gained the most. Its oil palm plantations did expand from 52,000 ha in 1938 to 93,000 in 1945 and 147,000 in 1958, with a further 98,000 ha under smallholder cultivation by the end of that period (Mulder, personal communication, 1968; Hartley 1988: 30). But political unrest following independence in 1960 led to stagnation and decline in the industry.
Unilever, however, the most important single investor in 1960 with 47,000 ha under oil palms, remained loyal to the newly independent country of Zaire through two decades of intermittent losses and political uncertainty. Thus, Unilever managers remained in place following nationalization in 1975, and the company was allowed to take back full control of the estates two years later (Fieldhouse 1978: 530—45). But at a national level, the research effort was decimated, and new planting was very limited after 1960, in marked contrast to developments at the same time in Southeast Asia.
The oil palm was first introduced to Southeast Asia in 1848, when four seedlings, originating from West Africa, were planted in the botanical gardens at Buitenzorg (now Bogor) in Java (Hartley 1988, 21). But this introduction did not lead to a plantation industry for some time, although offspring of the palms were used as ornamentals by tobacco planters.
In 1905 a Belgian agricultural engineer, Adrien Hallet, arrived in Sumatra and noticed that its palms grew more quickly and bore a richer fruit than counterparts in the Congo, where he had previously worked. Just as the oil palms in southeastern Nigeria bore a fruit with more oily pulp and a smaller kernel than their counterparts farther west, so did the Deli Dura palms, descended from the four Buitenzorg seedlings, hold a distinct advantage over the ordinary Duras of West and Central Africa (Leplae 1939: 25—7; Martin 1988: 47).
This superiority probably reflected the optimal soils, rainfall, and sunshine conditions of Southeast Asia, rather than any special genetic quirks of the Buitenzorg palms. However, the fact that all the Deli Duras were descended from so few parents meant that the early planters could expect fairly uniform results (Rosenquist 1986). This lowered the risks associated with plantation cultivation, an effect reinforced by the absence of the palm’s usual pests and diseases in its new geographic setting.
The relatively high yields and low risks from planting oil palms in Southeast Asia helped the industry to grow quickly, following the pioneering plantings of Hallet in Sumatra and of his friend Henri Fauconnier in Malaya in the 1910s. By 1919 more than 6,000 ha had been planted in Sumatra, rising to 32,000 in 1925, by which time 3,400 ha had come under cultivation in Malaya. Over the next five years, a further 17,000 ha were planted in Malaya, while the Sumatran area doubled.
This rapid expansion came not only because of growing confidence in the oil palm but also because of the grave postwar problems of the rubber industry. The oil palm was seen as a useful means of diversification to avoid a dangerous dependence on rubber. The pace of new planting slowed during the worldwide slump of the 1930s, but by 1938 Malaya had nearly 30,000 and Sumatra more than 90,000 ha under cultivation (Lim 1967: Appendix 5.2; Creutzberg 1975: Table 11; Hartley 1988: 21).
Developments in Sumatra hung fire for some time after 1945, as shown in Table II.E.3.1. Meanwhile, developments in Malaysia were more rapid, especially after 1960, when the replanting of old rubber estates with oil palms was stimulated by FELDA’s smallholder schemes.
At the same time, the Malaysian government and the estate sector launched several systematic Tenera-breeding efforts, in which high-yielding parents were selected and through which increasingly productive planting materials were generated. The new trees not only yielded more fruit but also produced a type of fruit that was ideally suited to the new screw presses which, having been tried out in the 1950s in the Belgian Congo, became widely used in Malaysia from the mid-1960s. These innovative developments have been described as “one of the world’s outstanding agricultural achievements” (Anwar 1981). The land area involved is shown in Table II.E.3.2.
A distinct species of the oil palm, Elaeis oleifera (also known as Elaeis melanococca), is indigenous to Latin America. Since the late 1960s, plant breeders have begun to take an interest in this variety because its oil has a high iodine value and unsaturated fatty acid content, making it especially suitable for food use. However, the fruit is often small, with a thin oil-yielding mesocarp surrounding a large, thick-shelled kernel. Harvested bunches often contain a low proportion of fruit of quite variable quality. Hence, the plant has not been cultivated commercially, although it is frequently found in riverside or swampy areas, and the oil is used locally for cooking, soap boiling, and lamp fuel (Hartley 1988: 85—9, 681—3).
Elaeis guineensis seeds were introduced to Central America by the United Fruit Company, which brought seeds from Sierra Leone to Guatemala in 1920, and from Malaysia to Panama in 1926 and Honduras in 1927. Other introductions from Java and the Belgian Congo followed, but the first commercial planting of 250 ha took place only in Guatemala in 1940. The United Fruit Company’s main interest was, traditionally, the production of bananas for export, but large banana-producing areas were destroyed by Fusarium wilt, and in consequence, oil palm and other crops were being tested as replacements.
The oil palm, however, proved vulnerable to disease in its new setting, and difficulties were encountered in identifying suitable growing conditions. Nonetheless, a successful development was founded on the northern coastal plain of Honduras, and in addition to developing plantations on its own land, the United Fruit Company stimulated oil palm cultivation by neighboring smallholders, whose fruit could then be processed in the company mills. Seed was also supplied to other Latin American countries (Hartley 1988: 33—6; D. L. Richardson, personal communication, 1993).
The beginnings of commercial planting in Latin America are summarized in Table II.E.3.3. By 1992 the total area planted to Elaeis guineensis in Latin America had grown to 390,000 ha — still a small fraction of the area in Africa or Southeast Asia. The distribution of plantings by country and sector is shown in Table II.E.3.4.
Until the early years of the twentieth century, palm oil was processed only by traditional village methods, by which loose fruits were collected from the ground or a few bunches were cut from the tree. Beginning in the 1920s, however, the United Africa Company and British colonial officials in Nigeria started experimenting with steam cookers and hand presses designed to make production at the village level more efficient in terms of labor use and oil yield. Yet a lack of cash prevented most farmers from trying the new machinery, with the exception of a few lucky recipients of free samples or government subsidies in the 1940s (Martin 1988: 64—6, 127—9).
A separate process of trial and error led to the development of the sophisticated factories required to deal with the volume of fruit produced on modern plantations and to produce oil of the high and standardized quality that would appeal to Western food processors. Such factories handle almost all the palm fruit of Southeast Asia, whereas in West Africa and Latin America, processing is carried out by a wide variety of methods, yielding oil for local consumption and for industrial as well as edible uses in the West.
Whatever the scale and sophistication of the process, the following main steps are required:
1. Separation of individual fruits from the bunch.
2. Softening of the fruit flesh.
3. Pressing out of the oily liquid.
4. Purification of the oil.
Traditional Village Process
Whole, ripe, fresh fruit bunches (FFB) are cut from the palm. With young trees this can be done from ground level. With older trees in West Africa, harvesting is still often accomplished by a man climbing the tree, secured to it by a loop of rope or other locally available materials, such as rattan and raffia fiber (Vanderyst 1920). But on plantations, a curved knife attached to a bamboo is used. After cutting, most of the fruits are still firmly attached to the bunches, which are divided into a few sections, heaped together, moistened, and covered with leaves. Natural fermentation during two to three days loosens the fruits so that they can be picked off the bunch sections by hand.
Following this step, two major variants in the process are used to produce two oils with different characteristics — those of soft oil and those of hard oil. The regions producing each type have changed since Julius Lewkowitsch (1895: 429) identified Saltpond in present-day Ghana as the cheapest source of hard oil, and Drewin on the Ivory Coast as the best place to buy soft oil. But the basic methods have changed little since they were first described by colonial officials in the 1910s and 1920s (Laurent 1912—13; Gray 1922; Faulkner and Lewis 1923; Martin 1988: 32—3).
For soft oil production, the fruits are separated as soon as they are loose enough and boiled with water for 4 hours to soften the flesh, which is very fibrous. The cooked fruit is emptied into a large container, which may be a pit lined with clay, an iron drum, or a large wooden mortar. It is then reduced to a pulp with pestles or by treading it under foot. The resulting mash may be diluted with water, and the oil is skimmed off or squeezed out of the fibrous mash by hand. In some instances, a sieve made of palm fronds is used to retain the fibers. At this stage the liquid product, which contains oil, water, and fruit fibers, is often boiled up with additional water and skimmed again, although this step is omitted in some cases. Finally, the oil is again heated to boil out the residual water.
Lewkowitsch (1922: 546) also reported on the preparation of small quantities of oil for kitchen use directly from freshly picked fruit, by boiling the fruit and skimming the oil. Such oil had good keeping properties and often a free fatty acid content below 2 percent; but yield was very low, and not available for export.
In the hard-oil process, the fruit is allowed to ferment for 3 or more days longer than in the
soft-oil process, until the flesh is soft enough. It
is then pulped by treading underfoot in an old canoe or pounding in a mortar. Oil is allowed to drain out for up to 3 days, then water is added, and the mix is trodden again. Further oil rises to the surface and is skimmed. The oil is boiled up with water in another container and finished as described for soft oil.
These two processes differ in some important respects. The prolonged fermentation in the hard-oil process results in a much greater enzymic breakdown of the neutral fat and, therefore, in a much higher free fatty acid content. The yield obtained by this process is also much lower. However, it has a substantial advantage in that the labor and firewood requirements are also much lower. Table II.E.3.5 summarizes these differences.
The strong characteristic flavors developed during both of these processes, as well as the naturally strong red color of the oil, are appreciated by local cooks and visiting gourmets, but they present severe limitations in the export markets. The high free fatty acid content and solid consistency of hard oil limits its range of uses, making it well suited to soap boiling but not to food processing. The solid consistency of hard oil is not due directly to the free fatty acids formed during the fermentation step but rather to the diglycerides, the other fragments obtained when one fatty acid is split from the neutral triglyceride molecule. M. Loncin and B. Jacobsberg (1965) have demonstrated the formation of a eutectic between triglycerides and diglycerides, resulting in a minimum melting point and maximum softness at a diglyceride content of 7 percent.
Mechanization of the Small-Scale Process
With the rapid twentieth-century growth in West African exports came the introduction of simple machines to reduce labor requirements and increase oil yield from a given quantity of fruit. Early machines, before and after the 1914—18 war, as described by Hartley (1988: 694—703), included a cylinder fitted with manually operated beaters, which was fed with softened fruit and hot water. After “beating,” an oil-water mixture was run off through a sieve. Another system used a special cooker and a press as adjuncts to the soft-oil process.
The first device to become widely adopted, however, was a modified wine and cider press: the Duchscher press. This consisted of a cylindrical cage of wooden slats, held in place by iron hoops, and a ram on a screw thread. The screw thread was turned manually by means of long bars (in the manner of a ship’s capstan), forcing the ram onto the pulped fruit. The exuding liquid was collected in a trough surrounding the cage.
Similar presses, but using a perforated cylindrical metal cage, are still in use today, giving yields of 55 to 65 percent of the oil present. A recent analysis of the needs for mechanization in the village has concluded that this is still the most practical implement, because it can be made and maintained locally and is inexpensive by comparison with other presses (C. E. Williams, personal communication, 1981). However, farmers in Nigeria (which was once the world’s largest exporter of palm oil) have, since the 1950s, been reluctant to invest in this or other improvements because of the low producer prices offered by the state-controlled marketing boards. It is to be hoped that recent reforms of marketing structures in Nigeria and elsewhere in Africa will encourage renewed innovation at the village level (Martin 1988: 126—36).
The next development in pressing was the introduction in 1959 of the hand-operated hydraulic press by Stork of Amsterdam. This was capable of processing 600 to 1,000 pounds of fruit per hour and could recover 80 percent of the oil present. The hydraulic mechanism was later motorized.
A different approach to mechanization brought forth the Colin expeller (first patented in 1904), which in essence is similar to a domestic mincer. It consists of a perforated cylindrical cage, fitted with a spiral screw or “worm,” which is turned manually through a gear. Cooked fruit is fed to the worm through a hopper, and the pressure developed as the worm pushes the fruit forward forces oil out through the perforations. Spent fiber and kernels are discharged at the end of the cage. The machine has a capacity of 100 kg cooked fruit per hour, or 250 kg per hour if motorized. The Colin expeller became popular after 1930, mainly in Cameroon. Its limitations were a reduced efficiency with Dura fruit, which forms the bulk of the wild oil palm crop; rapid wear of the screw; and a relatively high cost. The principle of the expeller, however, has been further developed into the screw press found in all modern oil mills (Hartley 1988: 703—5).
The presses described here provided a relatively efficient process for the step of pressing out the oily liquid during oil production and led researchers to seek improvements in the other steps. Several innovations have resulted from a project begun by the Nigerian Institute for Oil Palm Research (NIFOR) during the 1950s in cooperation with the Food and Agriculture Organization of the United Nations (FAO) and the United Nations Development Program (UNDP).
The following unit operations and equipment are involved in palm oil production.
1. Fruit bunch cookers, which are wood-fired cylindrical tanks. They are loaded with cut-up fresh fruit bunches (FFB).
2. A bunch stripper operated by hand, which consists of a cylinder made up of slats and turns on a horizontal axis that tumbles the cooked bunch sections until the individual fruits separate from the bunch and fall between the slats.
3. A digester (to break up the softened fruit and release its oil from its cells), consisting of a horizontal cylinder in which beater arms rotate, driven by a small diesel motor.
4. A hydraulic press, which was introduced in 1959.
5. A clarification unit consisting of two linked tanks, whereby heating with water causes the oil layer in the first tank to overflow at the top into the second tank. There it is dried by the waste heat from the fire under the first tank.
Extraction efficiencies of 87 percent at a free fatty acid (FFA) level of below 4 percent are routinely attainable by this process. Between a quarter and a half ton of fresh fruit bunches per hour can be processed, depending on cooker capacity.
A number of variants of this process are in use:
1. Bunches are allowed to ferment so that only loose fruit is loaded into the cooker. This variant yields oil of higher FFA.
2. The hydraulic presses may be driven by a small diesel engine.
3. Clarification can take place in simple tanks with direct heating.
4. Cooking of bunches may be by steam, whereby whole bunches are loaded into a tank fitted with a perforated plate about 15 centimeters (cm) from the base. Water is boiled under the plate, and the steam penetrates through the bunches.
5. In Ghana, an interesting operating procedure has been developed, in which the mill owner provides mill facilities to the farmers, who are then responsible for the bunch stripping and cooking of
the fruit. Mill operatives carry out digesting and pressing procedures, after which the farmers take away the oil from their own fruit for clarification (G. Blaak, personal communication, 1989).
The advantages for the farmers are numerous: They need no capital investment in mill equipment; there are no arguments regarding purchase price of FFB; if farmers produce high-quality Tenera fruit, they retain the full benefit; and farmers pay only a processing charge. Their net profits are higher than those obtained selling FFB, even if they employ labor to carry out their share of the processing.
The small-scale processes just described are suitable for the processing of FFB from wild oil palm groves or from smallholdings. The main objective is to produce red palm oil for traditional food use.
The processing of the large quantities of fruit produced by plantations or by large smallholder cooperatives, however, requires a progressively greater degree of mechanization and mechanical handling as the quantity increases. Furthermore, since oil produced on a large scale is usually intended for export or for local refinery processes, the ultimate objective is a neutral oil of bland flavor and nearly white color. To attain this quality, the processes (including the handling of FFB) are designed to minimize the development of free fatty acids and oil oxidation.
A simple factory process of intermediate scale, in which the material is still handled manually between processing stages, is the Pioneer mill, which was developed by the United Africa Company around 1939. It is designed to process about three-quarters of a ton of fruit per hour, which is the equivalent of about 1 ton of fruit bunches, following the removal of the fruit from the bunch stalks. The process consists of the following steps:
1. Autoclaving — 200 kg of fruit is loaded into a vertical batch autoclave mounted on a gantry and cooked under steam pressure of 20 pounds per square inch for 15 minutes.
2. After cooking, the fruit is discharged by gravity into a digester fitted with a stirrer, which breaks it up and releases the oil from the cells.
3. The resulting mash is treated in a basket centrifuge, operating at 1,200 revolutions per minute.
4. The oil flowing from the centrifuge passes through a screen to remove the fiber, and then to a settling tank.
5. The settling tank contains a layer of hot water, and the oil is pumped in below water level. The water is boiled for 15 minutes and then allowed to settle. The oil layer is decanted through a hot-water layer in a second settling tank.
6. The tank is heated to boiling point for 15 minutes and allowed to settle. The clean oil is put into drums.
7. The sludge from the two settling tanks is further treated by boiling and settling, and the residual oil is recovered.
An oil mill of essentially the same design, with a capacity of 2 to 3 tons of fruit per hour, was featured in the Wembley Exhibition of 1924 by Nigerian Products Ltd., Liverpool, and was apparently demonstrated in operation there (Elsdon 1926: 316—22). In 1950 there were 13 Pioneer mills in operation in Nigeria. The numbers increased to 65 in 1953 and more than 200 in 1962, producing about 25,000 tons. But, subsequently, their use has declined (Mulder, personal communication, 1968).
The Pioneer mill cannot meet the needs of well-established plantations generating large volumes of fruit. To keep costs down and output up, it is vital to have a fully mechanized power-operated mill. The development of such mills began in Kamerun and in the Congo before World War I. Mills using centrifuges for oil extraction were in operation in the Congo in 1916, in Sumatra in 1921, and in Malaya in 1925 (Hartley 1988: 703—5). Centrifuges were largely replaced by hydraulic presses in the 1930s, although they were still being operated at Batang Berjuntai, Malaysia, in 1982. Batch-fed hydraulic presses were, in turn, replaced by continuous screw presses, which saved labor and handled much larger volumes of fruit. At this final stage of innovation, the development of agricultural and processing technology went hand in hand. The screw press tended to mangle the fruit of the Dura palm, with its relatively thin layer of oil-bearing mesocarp, but proved ideally suited to the Tenera variety (Maycock, personal communication, 1991).
The principal steps involved in the production of palm oil today are the following:
1. Harvest at optimum ripeness.
2. Transport FFB to an oil mill with minimum bruising.
3. Transfer FFB to sterilizing cages.
4. Sterilize FFB by steam under pressure.
5. Transfer cooked FFB to a bunch stripper.
6. Transfer fruit to a digester.
7. Press in single-screw or twin-screw press.
8. The oily discharge from the press, containing water and fruit debris, is passed through screens and settling tanks.
9. The oil phase from the settling tanks is passed to a clarifying centrifuge. The sludge, or heavy phase, from the settling tanks is centrifuged and the recovered oil returned to the settling tanks.
African Food Uses
In West Africa, palm oil has a wide range of applications. It is employed in soups and sauces, for frying, and as an ingredient in doughs made from the various customary starch foods, such as cassava, rice, plantains, yams, or beans. It is also a condiment or flavoring for bland dishes such as fufu (cassava). A basic dish, “palm soup,” employs the whole fruit. The following dishes from Ghana are illustrative of the wide range of palm oil use (Wonkyi Appiah, personal communication, 1993).
In the case of palm soup, first wash and boil palm fruits. Next, pound the fruit and mix with water to a paste. Add meat or fish, vegetables, onions, spices, and salt. Boil for 25 minutes and simmer for a further 15. Serve the soup with cooked rice, yam, plantain, fufu, or kpokpoi. (The latter is a corn dough, steamed and cooked, with okra and palm oil stirred in.)
Palm oil is also used in baked dishes, and one popular dish has different names according to the local language. When Ofam in Twi, or Bodongi in Fanti, is prepared, ripe plantains are pounded and mixed with spices, some wheat or corn flour, beans, and perhaps eggs. Palm oil is stirred into the mixture and the whole is then baked in the oven. The dish is served with ground nuts. Apiti is a similar dish, baked in leaves, without beans or eggs.
The characteristic flavor of palm oil prepared by village methods is an important feature of these dishes. Indeed, it is one of their most “traditional” features. Several of the other key ingredients, such as salt, wheat, or (in popular eastern Nigerian dishes) stockfish, became widely available only in the nineteenth and early twentieth centuries, when they were imported from Europe in exchange for palm oil itself (Martin 1988: 28—9, 50).
Early Western Food Uses
The fully flavored red palm oil produced by West African village methods has not proved suitable for food use in the importing countries of the West, where the consumer requires a bland cooking fat, near white in color, or a margarine, similar in appearance to butter. Today’s plantation-produced palm oil can be bleached and neutralized to meet Western requirements, but in the nineteenth and early twentieth centuries, the high FFA content even of “soft”
West African palm oils made them too difficult and uneconomic to neutralize (Andersen 1954: 27). Even before loading aboard ship, they fell far short of the current quality standard of less than 5 percent FFA, 0.5 percent moisture, and 0.5 percent dirt; and a slow ocean voyage did little to improve matters, as the acidity tended to increase en route (Vanneck and Loncin 1951).
Throughout the nineteenth century, exported oil from West Africa was placed in wooden casks usually supplied from Europe in the “knocked down” state and put together before being filled. Sailing ships became much larger in size and were gradually displaced in the second half of the century by steam ships, which were able to call at a greater number of ports and make more regular voyages (Lynn 1989). This development probably improved the overall quality of oil arriving in Europe, but as the oil was still made on a small scale by different methods and carried in casks, there was plenty of variation.
This quality problem could have been resolved in the late 1920s, when bulk storage tanks were installed at some African loading ports, initially in Nigeria and the Belgian Congo. It was then possible for incoming oil to be washed and cleaned before bulking, with an improvement in quality (Iwuchukwu 1965; Mulder, personal communication, 1968). However, hardly any Nigerian palm oil was suitable for the European food industry until the 1940s.
When Sumatran and, later, Malayan plantations started to export oil in the 1920s, their fruit was harvested systematically from the beginning. It was transported with minimal bruising to the factories and processed in a standardized way. Bulk shipment was developed from the outset. The first shore tanks were installed at Belawan in North Sumatra in the 1920s, and oil from Malaya was taken there by steamer from 1931 onward. In 1932 the Malayan planters set up their own Palm Oil Bulking Company with an installation at Singapore (Shipment of palm oil in bulk 1931; United Plantations Berhad, unpublished documents; T. Fleming, personal communication, 1993).
It thus became possible to develop and maintain the quality standards that are now current worldwide. The planters aimed to produce oil of 3.5 percent FFA, which would then fall well within the limit of 5 percent FFA on arrival in Europe or America. Oil arriving at above 5 percent FFA was sold at a discount, depending on the excess acidity (Hartley 1988: 687).
European food manufacturers could now begin to introduce palm oil on a commercial basis, drawing on earlier experiments and fitting it into two long-standing patterns of fat use. In central and northern Europe, indeed in cool weather regions generally, the traditional fats are the products of the farm yard — butter, beef tallow, and lard. In southern Europe, with its dry hot climate, olive oil has been in general use for thousands of years. Thus, consumers have had available either a plastic product of solid appearance or a clear liquid oil, and the cooking and eating habits developed accordingly.
Respect for these traditions led to the invention in 1869 of margarine and its development as a replacement for butter, when the latter was in short supply. Margarine was originally made from beef fat, and the plastic nature of butter was attained by blending in a liquid fraction separated from beef fat by crystallization. Margarine proved so popular that European supplies of beef fat did not suffice. Imports from the New World were important in the nineteenth century, but various imported vegetable oils gradually took the place of beef fat margarine blends as refining techniques developed. The fact that even “soft” palm oil is a solid fat in temperate climates, with a consistency quite similar to butter, made it an obvious candidate for such experiments, and the first recorded trial took place in 1907 (Feron 1969).
The refining and bleaching process required to render suitable palm oil involved a great deal of research and empirical know-how. Illustrative is some unpublished correspondence (copies held by K. G. Berger) between Dr. Julius Lewkowitsch, a consultant chemist in oils and fats, and a Liverpool trading house, the African Association Limited. Dr. Lewkowitsch had invented a process for rendering palm oil into an edible product and had entered into an agreement, dated January 24, 1905, to share the costs of development of the process with the African Association.
Evidently the work proceeded rather slowly, because in September 1907, Lewkowitsch received a letter from the Vice-Chairman of the African Association, saying: “I have sent you under separate cover a sample of refined beef suet. . . . Would it be possible to have the samples of the palm oil products made up to appear like this sample? I am afraid I shall never satisfy my Co-Directors until I can show them a palm oil product they can eat.” A successful prototype was probably produced eventually, because in 1910 a small manufacturing concern, V. B. Company, was incorporated and the African Association was paying Lewkowitsch a regular salary as managing director from 1910 to 1912.
The first decade of the twentieth century also saw the introduction of hydrogenation of oils, a process by which liquid oils could be turned into plastic or hard fats to a controlled degree. As a result, vegetable oil—based “shortenings” were produced to replace lard and beef tallow as ingredients for cakes, pastries, and biscuits and as frying fats. Once adequately refined, palm oil was easily introduced in blend formula for these types of products and had the advantage of not requiring hydrogenation. By the mid-1930s, the relatively clean and less acidic plantation-produced palm oil of Malaya and Sumatra was finding a ready market in the United States, where it was used not only in fine toilet soaps but also in the making of compound lard. Over 50,000 tonnes per annum were used in the American compound lard industry between 1935 and 1939 (Lim 1967: 130—2).
Wartime interruptions of supplies from Asia during the 1940s forced American manufacturers to find substitutes for palm oil, and the market was slow to revive afterwards. However, in Britain, wartime shortages of butter encouraged the use of palm oil in both margarine and compound lard, and this market continued to grow in the 1950s (Lim 1967: 131). British manufacturers, through the home Ministry of Food and the West African Produce Control Board, were able to corner the market in British West African palm oil (Meredith 1986). The Produce Control Board and, from 1949, its successor, the Nigerian Oil Palm Produce Marketing Board, played an important role in bringing the quality of this oil up to the standards set in Southeast Asia.
A grading system was set up as follows:
Grade I under 9 percent FFA
Grade II 9 to 18 percent FFA
Grade III 18 to 27 percent FFA
Grade IV 27 to 36 percent FFA
Grade V over 36 percent FFA
Higher prices were paid for the better grades, and there was an immediate response from the village producers, which enabled a Special Grade palm oil to be specified in 1950 with maximum 4.5 percent FFA at time of purchase. A significant premium was paid for this oil, with the result that Special Grade oil, which was only 0.2 percent of production in 1950, jumped to over 50 percent by 1954. In 1955, the specification was tightened to 3.5 percent FFA, and by 1965 Iwuchukwu (1965) reported that more than 80 percent of material for export had reached this quality.
Market Developments Since 1950
The development (mainly since the 1950s) of convenience foods and of snack food manufacture on an industrial scale opened up additional uses for palm oil, because of its good resistance to oxidative deterioration and its better ability to withstand the high temperatures used in frying than most alternative oils (Kheiri 1987; Berger 1992). The market developed especially rapidly after 1970, as the trees planted during the 1960s in Malaysia matured and as the Malaysian government and estate sector began to promote their product more actively in the West. Asian markets had also become important by 1980, as Western processing techniques were applied to meet local needs.
Figure II.E.3.1 shows that the world production of palm oil, together with its share in world supplies of oils and fats, increased dramatically from 1970 onward. Yet by this time, in many parts of Africa the industry had declined (Table II.E.3.6). Nigeria, for example, had no exportable surplus of oil after 1970 and, in fact, became a net importer of palm olein in the 1980s. Exports from Zaire became very limited in the 1980s, and the Ivory Coast, with exports of 60,000 to 100,000 tonnes per annum, was left as the only significant African supplier. Meanwhile, as shown in Table II.E.3.6, new Asian producers were emerging, in particular Papua New Guinea and Thailand. By 1990, exportable surpluses of 10,000 to 30,000 tonnes per annum were also reaching the world market from Honduras and Costa Rica (Mielke 1991: 111).
As Malaysian production grew, both the planters and the government realized that it was vital to improve processing methods and to encourage the growth in demand for palm oil. The estate sector took the lead during the 1960s, developing higher grades of crude palm oil to suit European needs. Later, the government joined in the development of refineries and new products to suit Asian, as well as Western, tastes.
The old standard of 3.5 percent FFA on leaving the factory continued to apply to the bulk of crude Malaysian palm oil. However, in the last 30 years it has been recognized that the production of a stable refined palm oil of good color is also dependent on characteristics other than FFA content, as shown in Table II.E.3.7. In particular, the degrees of oxidation and contamination with catalytic traces of metals are important. But surveys indicate that the peroxide value (a measure of the state of oxidation) of standard palm oil arriving at Malaysian ports or refineries fell from 3.9 milligrams per kilogram to 2.3 milligrams per kilogram between 1974 and 1991 (Jacobsberg 1974; Wong 1981; V. K. Lal, personal communication, 1991).
Planters, both in the Belgian Congo (Zaire) and in Malaysia, also sought to develop a premium product with exceptionally low FFA, obtained through stricter harvesting routines and processing with minimum delay. “Special Prime Bleach” (SPB) grade was developed in the Belgian Congo (Loncin and Jacobsberg 1965), having a maximum of 2 percent FFA and reduced levels of iron and copper contamination, while in Malaysia two special grades became available, namely “SQ” and “Lotox.” The SQ and Lotox specifications include limits for oxidation characteristics as well as trace metals and in practice satisfy the most stringent requirements of the European market (Johansson and Pehlergard 1977).
A separate development, which also improved the quality of palm oil arriving in Europe, was the introduction in the 1960s of “parcel tankers.” These are specialized ships of up to 30,000 tons. The cargo space is subdivided into a number of separate tanks, generally with a capacity of between 100 and 1,000 tons. Tanks are fitted with separate pumps and pipelines so that different liquid cargoes can be carried without contamination. With the very large export trade from Southeast Asia, parcel tankers are capable of economically carrying palm oil and other oils of different grades to destinations all over the world. Appropriate shore installations have been developed since 1960 in the exporting ports of Southeast Asia and in most receiving ports in Europe, United States, and Japan, and are being developed in countries that have only recently become large importers. Like the development of bulk shipment from Malaya in the 1930s, this innovation was fostered by cooperation among the planters, who marketed their oil through a common Malayan Palm Oil Pool from 1953 to 1974 (Allott and Wong 1977).
By 1974 the volume of Malaysian oil exports and the diversity of markets had grown to the point at which a free marketing system was more appropriate. The range of palm oil products exported was also growing, following the application of the fractionation process first developed for beef tallow in the 1870s. This technique separated crude palm oil into olein and stearin. The olein remains a liquid oil in hot climates and, therefore, readily fitted into the large Indian demand for cooking oil.
From 1970 the Malaysian government encouraged and licensed private enterprises to set up refineries that could both fractionate palm oil and use it in a more traditional manner to produce shortenings for Asian markets. In India, for example, there is a large consumer demand for vanaspati, a shortening developed as a replacement for butterfat ghee. Similar shortening products are traditional in Pakistan, Egypt, and other Middle East countries. Often lacking their own refining facilities, such countries have tended to import palm oil products from Malaysia in fully processed, ready-to-eat form.
The Malaysian government encouraged this trend by offering tax concessions for refineries in their early years and by progressive remission of the export tax on crude palm oil, graded according to the degree of processing of the end product. This development received a mixed reaction in Western Europe, which had ample processing capacity and extensive technical know-how (Berger, personal observation). However, it proved successful in stimulating the growth of new Asian markets, as shown in Table II.E.3.8.
Although private enterprises had an excellent record in developing new processing techniques, they often felt hampered by their distance from major markets, which posed difficulties in designing and introducing new products. In 1979 the Malaysian government set up a specialized Palm Oil Research Institute (PORIM) as a statutory body, financed by a levy on palm oil production. A major part of its mission was the development of application technology for palm oil and the propagation of the information to end users anywhere in the world. By studying consumption patterns of oils and fats, the Institute’s staff was able to identify potential new markets and provide the technical input needed for their development (PORIM 1981: 1—5). Their work proved useful to producers worldwide, especially from the late 1980s on when a debate arose over palm oil’s nutritional value.
Nutritional Properties of Palm Oil
Briefly, the general nutritional functions of fat are to:
1. Provide energy efficiently.
2. Supply the essential linoleic and linolenic acids.
3. Carry the fat soluble vitamins A, D, and E.
4. Improve the palatability of foods.
The specific nutritional properties of palm oil may be considered in relation to its chemical composition. Typical values are given in Table II.E.3.9.
The unsaturated acids present are mainly oleic, with a useful level of linoleic and a small amount of linolenic acid. In consequence, palm oil has a high stability to oxidation. Palm oil is readily absorbed and shows a digestibility of 97 percent or greater, similar to that of other common edible oils.
As in other vegetable oils, the middle 2-position is mainly occupied by unsaturated fatty acids. This is different from animal fats, where the 2-position is usually occupied by a saturated fatty acid.
Unrefined, or “virgin,” palm oil is one of the richest natural sources of carotenoids. Regrettably, these are removed during the industrial refining process so that their nutritional benefits are lost, except to populations who traditionally use palm oil in the unrefined state.
The tocopherol content (see Table II.E.3.10) is one of the most interesting features in palm oil because it consists mainly of the tocotrienols, with an unsaturated side chain. These are not found in the other common vegetable oils.
Analytical work has shown that an average of 50 to 60 percent of the tocopherol content remains after refining, but the extent of removal depends on the refining conditions used. The tocopherols are important natural antioxidants, although their antioxidant activity is somewhat lower than the synthetic phenolic antioxidants permitted in foods. They are less volatile and, therefore, more persistent in high-temperature conditions, such as in deep-fat frying. The tocopherol content is a major factor in stabilizing palm oil against oxidation. The nutritional benefit of tocopherols in a number of disease conditions in which free radicals or oxidation are implicated has become a very active field of research, although to date little has been done on tocotrienols as such.
The three major component fatty acids of palm oil — palmitic, oleic, and linoleic acids — are also the most common fatty acids found in vegetable oils. Palm oil has been used as a traditional food in West Africa probably for thousands of years, which provides some evidence that it has good nutritional properties.
Research into coronary heart disease in relation to diet led to the general hypothesis of A. Keys, J. T. Anderson, and F. Grande (1957) that saturated fatty acids raised blood cholesterol levels, whereas linoleic acid reduced them. Subsequent refinements of the hypothesis (Hegstedt et al. 1965) indicated that saturated acids did not all have the same effect. In particular, D. M. Hegstedt and colleagues concluded that myristic acid was 3 to 4 times more effective than palmitic acid in raising blood cholesterol levels.
The early work leading to these hypotheses did not use palm oil in the experimental diets. However, between 1985 and 1987, concern was expressed in the media, principally in the United States, that the saturated fatty acid content of palm oil meant that it would raise blood cholesterol levels and was, therefore, an undesirable food component. This assertion was not based on any direct experimental evidence. Instead, a review of the few dietary experiments in which palm oil had been used (as a control, not as the subject of investigation) showed that, in general, a small reduction in blood cholesterol level was experienced (New findings on palm oil 1987).
Subsequently, a study in Nigeria, a principal traditional consumer of palm oil, was published (Kesteloot et al. 1989). Serum lipid levels were measured in 307 men and 235 women, whose ages were 15 to 64 (mean 38.8) for men and 15 to 44 (mean 31.4) for women. Mean values for total serum cholesterol were 156.3 and 170.9, respectively, and for HDL cholesterol 46.0 and 49.0. Subjects consumed their normal diet, with 84 percent of the fat intake from palm oil. These serum lipid levels compared very favorably with black and white populations in the United States, where total fat intake is much higher, and where palm oil comprises only 1 to 2 percent of total fat intake (Park and Yetley 1990).
A number of new dietary studies have addressed the nutritional properties of palm oil, particularly in its effect on blood lipids. T. K. W. Ng and colleagues (1991), for example, found that when palm olein formed 75 percent of fat intake in a normal Malaysian diet, total serum cholesterol was significantly reduced, by 9 percent from the level at entry, and that the reduction was almost entirely in the undesirable LDL cholesterol. The study was carried out on 20 men and 7 women of average age 24.
A. Marzuki and colleagues (1991) used 110 residential high school students of both sexes as subjects (ages 16 to 17). Although the normal menu was provided, palm olein was the sole cooking oil for 5 weeks. This was followed by a “washout” period of 6 weeks on regular cooking oil and a second experimental 5 weeks in which only soya bean oil was used. There was no difference in plasma total LDL or HDL cholesterol between the two trial periods.
K. Sundram and colleagues (1992) carried out a double blind crossover study on 38 men, in which 70 percent of the fat in a normal Dutch diet was replaced by palm oil. There was no effect on total serum cholesterol, but a significant increase of 11 percent in HDL 2 cholesterol, resulting in a beneficial decrease in the LDL/HDL 2+3 ratio. G. Hornstra and co-workers (1991) also measured plasma lipoprotein (a), which is strongly associated with an increased risk of ischemic cardiovascular disease. They found a highly significant 10 percent decrease in this component during the test diet period, and the decrease was greatest in subjects with an initial high level of lipoprotein (a), that is, those with an enhanced risk.
R. Wood and colleagues (1993) examined the effect of six different fats used as components of items of a normal American diet on 30 middle-aged men. When refined palm oil formed 60 percent of the dietary fat intake, there was no significant effect on total cholesterol, but HDL cholesterol was increased compared with the baseline diet.
Ng and colleagues (1992) studied 33 subjects consuming a Malaysian diet containing 34 percent of calories as fat. When palm olein was 23 percent of energy (that is, two-thirds of the fat intake), there was no significant difference in serum total — LDL or HDL cholesterol contents from the levels at entry. The use of olive oil in place of palm olein gave almost identical results, although the ratio of Thromboxane B2 to Prostacyclin PGFI alpha was significantly lower during the palm olein dietary period.
D. Heber and colleagues (1992) found no increase in the plasma total cholesterol level of 9 subjects, but a small rise in plasma HDL cholesterol when one-half of the dietary fat was palm oil. The diet contained 35 percent energy as fat.
A. S. Truswell and co-workers (1992) conducted 2 trials (21 and 30 subjects, respectively) in which one-half of the dietary fat was palm olein. He found that a 10 percent increase in HDL cholesterol accounted for the 3 percent rise in total cholesterol observed.
The conclusion is that palm oil, used as a dietary fat at a high level — 10 to 20 times that usual in a Western diet — does not raise serum total cholesterol. However, the level of serum HDL cholesterol (popularly described as the “good” cholesterol, because in this form cholesterol is catabolized and removed) was significantly increased in several of the recent studies.
Mention might also be made of two additional studies, by R. C. Cottrell (1991) and C. E. Elson (1992). These authors reviewed 139 and 180 scientific publications, respectively, and both concluded that palm oil was a nutritionally satisfactory component of the diet. Cottrell wrote that “the decision to use palm oil in food products should be based on a rational appraisal of its technical performance value rather than on a misconceived view of the health implications of its use” (Cottrell 1991: 989S—1009S).
Nonetheless, the view still persists in some circles that palm oil is an unhealthy tropical grease, and it is difficult for palm oil producers to counter this perception because the product had little or no public image among Western and Asian consumers before the start of the recent media debate. Processed until it has become an anonymous ingredient, and used in a wide variety of compound fats and such other items as biscuits, its original flavor and feel have been lost to most consumers. But now that the wider nutritional benefits of palm oil’s natural carotenoids are becoming more generally recognized, perhaps it is time to rediscover the fully flavored red oil and promote its use, not only in Africa and Latin America but also in Asia and the West.
K. G. Berger
S. M. Martin