Index | Search | Home | Table of Contents

Koziol, M.J. 1993. Quinoa: A potential new oil crop. p. 328-336. In: J. Janick and J.E. Simon (eds.), New crops. Wiley, New York.

Quinoa: A Potential New Oil Crop*

Michael J. Koziol

    1. Maize Oil Content and Yields
    2. Quinoa Oil Content and Yield
    3. Oil Composition
    1. Saponins
    2. Oil Press Cake as a Dietary Supplement
    3. Carbohydrate Cream Substitute
  5. Table 1
  6. Table 2
  7. Table 3
  8. Table 4
  9. Table 5
  10. Table 6
  11. Table 7

Quinoa (Chenopodium quinoa Willd.) is an Andean pseudocereal grain about 1.0 mm thick which ranges in diameter from 1.0 to 2.5 mm and in seed weight from 1.9 to 4.3 g/1000 seeds (Alvarez et al. 1990). From earlier investigations (White et al. 1955) to more recent reviews (Coulter and Lorenz 1990; Galwey et al. 1990), attention has focussed primarily on the content and quality of the protein in quinoa. This absorbing interest in quinoa's profile of essential amino acids has completely overshadowed another equally important nutritional characteristic of the grain when compared with cereals, namely a relatively high content of an oil which is rich in the essential fatty acids, linoleate and linolenate (Koziol 1990a). (As arachidonic acid can be synthesized from linoleic acid it will not here be considered as essential.)

The use of non-traditional oil crops, such as maize, as sources of edible vegetable oils depends partially on their oil content and composition but more importantly, on the commercialization of other major products derived from the grain; oil production is best viewed as a byproduct in such cases. Thus, as early as 1943, Jamieson stated: "Were it not for the fact that in the preparation of hominy, starch, glucose, and other corn products, the germ is almost completely separated from the rest of the kernel, corn oil would not have become an important commercial product." Production of maize oil increased proportionally with that of high-fructose corn syrup, and its consumption has been boosted by an increasing public awareness of the importance of polyunsaturated fats in the diet (Leibovitz and Ruckenstein 1983). Worldwide, the production of maize oil increased from 0.4 million t in 1965 to 1.0 million t in 1985, with a further increase to 1.5 million t predicted by 1995 (Gunstone 1989). The production of maize oil as a commercially viable byproduct serves as an appropriate model with which to compare the potential of quinoa as a new oil source, with special emphasis on Ecuadorian data.


Maize Oil Content and Yields

Maize cultivated in the United States contains 3 to 4% oil (Mounts and Anderson 1983). The germ, which accounts for 5 to 14% of the weight of a maize kernel (Lásztity 1984; Patterson 1989; Alexander 1989) contains about 85% of the oil (Mounts and Anderson 1983). In the starch industry, the maize germ as removed by the wet degermination (milling) process and presented for oil extraction contains about 50% oil, as opposed to the 10 to 24% oil available in germs removed by the dry process used in the production of hominy, grits, and corn flakes (Leibovitz and Ruckenstein 1983).

Breeding programs for maize with a higher oil content have successfully produced hybrids that have 6 to 8% oil with yields equivalent to those of other commercial cultivars (Weber 1983). Hybrids with higher oil contents tend to have lower yields in terms of tonnage per hectare; for example, within 22 cycles the oil content of 'Alexho Synthetic' was increased from 6.2 to 12.9%, but the yield of the maize was reduced from 8.50 to 5.69 t/ha (Alexander 1989). Given the higher oil content, oil yield per hectare actually increased from 0.53 to 0.73 t/ha. Increases in oil production, however, are accompanied by increases in germ sizes and by decreases in grain and endosperm weight and hence in starch production (Weber 1983; Alexander 1989). Despite the successes of the maize breeders, the only inducement for farmers to cultivate the higher oil content maize for milling up until the 1980s was special contracting (Weber 1983). More recently, the increased production of high-fructose sugars in the United States suggests that higher oil content maize might not be welcome by the wet millers, even though the value of maize oil is at least three times higher than that of the corn starch (Alexander 1989). Some of the dry milling processes cause excessive damage to the larger embryos of the higher oil content maize resulting not only in contamination of the endosperm by oil but also in the release of lipolytic enzymes from the embryo which can accelerate rancidity. The future of the higher oil containing hybrids of maize depends upon three factors: (1) a reassessment of the value of maize oil as a byproduct in the production of other maize products; (2) the costs of modifying existing processes and equipment to deal with the changes in the morphology of the kernels; and (3) the productivity of the new hybrids in terms of yield per hectare.

Quinoa Oil Content and Yield

On a fresh weight basis quinoa shows an oil content ranging from 1.8 to 9.5% (Table 1) with a calculated global mean of 5.8%, an oil content higher than that of normal maize. As with maize, the oil is concentrated in the germ which in quinoa represents 25 to 30% of the weight of the grain (Cardozo and Tapia 1979; Fuentes 1972). As the quinoa germ encircles the endosperm, we have found that it can easily be removed by a modified polishing procedure to give a fraction containing 19% oil.

Although special breeding programs were necessary to achieve an oil content of 6 to 8% in maize, several cultivars of quinoa already show oil contents in this range (Table 2). Unlike maize, in which an increase in oil content resulted in a decrease in starch content, increased oil content in the quinoa grain showed no significant correlation with total carbohydrate content (r = 0.348) and was negatively correlated with protein content (r = -0.910, P < 0.025) (Table 3).

Field trials were conducted with six Ecuadorian cultivars of quinoa with oil contents between 7.2 and 8.7% (fresh weight) sown at an altitude of 3,100 m and using experimental plots ranging in size from 0.2 to 0.5 ha (Burgasi et al. 1990). Converting the data from Table 4 to a dry matter basis gave no significant correlation (r = 0.648) between yield and oil content, unlike maize in which yield decreased with increasing oil content (Alexander 1989).

The potential yields of quinoa and maize oils were estimated (Table 5) for Ecuador by multiplying the ranges of average oil content by data for average crop yields as reported by the Ministerio de Agricultura y Ganadería (MAG 1985). The potential yield of quinoa oil given under the general case was calculated using the data of Nuñez and Morales (1980), who reported quinoa yields in Bolivia of 3,960 kg/ha without fertilization and of 5,420 kg/ha with fertilization. This latter value was used to calculate a maximum possible yield of quinoa. Nuñez and Morales (1980) extrapolated a yield per hectare on the basis of very small experimental plots (4 rows of 6 m spaced 0.4 m apart), and their experimental plots received eight treatments against mildew. Commercially, quinoa would likely receive one or two treatments against mildew (M. Alvarez pers. commun.). Thus, the 488 kg oil/ha reported in Table 5 is best cautiously interpreted as a maximum yield obtainable under exceedingly favorable conditions.

Conversely, the exceedingly low yield of quinoa grain of 449 kg/ha reported by MAG (1985) reflects small-scale traditional agricultural practices; hence the calculated oil yields for Ecuador based on that data from MAG (1985) are thus also low. Better estimates of quinoa yields in Ecuador are those given by Burgasi et al. (1990) and reported in Table 4, which would then correspond to potential oil yields of 102 to 306 kg/ha (Table 5). This oil yield for quinoa compares favorably with that estimated for maize by Pryde and Doty (1981) in the United States, namely 254 kg/ha assuming an average oil content of 4.8% for maize.

Oil Composition

The fatty acid composition of quinoa oil is similar to that of maize oil (Table 6). The high concentrations of linoleic and linolenic acids normally make such oils susceptible to oxidative rancidity but both oils have relatively high concentrations of natural antioxidants, namely tocopherol isomers. The mean concentration of alpha-tocopherol reported for three cultivars of quinoa on a dry weight basis was 52 ppm (De Bruin 1964), which corresponds to a concentration of 754 ppm in the oil. Further analyses have shown quinoa oil to contain 690 to 740 ppm alpha-tocopherol and 790 to 930 ppm gamma-tocopherol; upon refining these concentrations fall to 450 and 230 ppm, respectively (U. Bracco pers. commun.). In comparison, refined maize oil contains 251 ppm alpha-tocopherol and 558 ppm gamma-tocopherol (Souci et al. 1986). As optimum antioxidant activity of the alpha- and gamma-isomers of tocopherol has been reported at 100 to 200 ppm (Hudson and Ghavami 1984) quinoa oil would be expected to show a stability towards oxidative rancidity similar to that of maize oil.



Quinoa can be classified according to its saponin concentrations as either "sweet" (saponin free or having less than 0.11% saponins on a fresh weight basis) or "bitter" (containing more than 0.11% saponins) (Koziol 1990b). The saponins in quinoa are glycosidic triterpenoids (Burnouf-Radosevich et al. 1985; Mizui et al. 1988, 1990; Ma et al. 1989; Meyer et al. 1990; Ridout et al. 1991) and represent the major antinutritional factor found in the grain (Koziol 1992). Fortunately, most of these saponins are concentrated in the outer layers of the grain (perianth, pericarp, seed coat, and a cuticle-like layer) which facilitates their removal industrially by abrasive dehulling (Reichert et al. 1986) or traditionally by washing the grains with water.

The toxicity of saponins depends upon their type, method of absorption, and target organism (for a comprehensive review, see Price et al. 1987). Because of their differential toxicity to various organisms saponins have been investigated as potent natural insecticides which would have no adverse effects on higher animals and man (Basu and Rastogi 1967). Other interest in saponins is in their antibiotic, fungistatic, and pharmacological properties (Basu and Rastogi 1967; Agarwal and Rastogi 1974; Chandel and Rastogi 1980; Nonaka 1986). The pharmacological interest in saponins lies with their ability to induce changes in intestinal permeability (Gee et al. 1989; Johnson et al. 1986) which may aid the absorption of particular drugs (Basu and Rastogi 1967), and with their hypocholesterolemic effects (Oakenfull and Sidhu 1990). As the saponins in quinoa have been relatively little studied their potential commercial uses remain unknown.

Oil Press Cake as a Dietary Supplement

Both bitter and sweet quinoa are currently subjected to abrasive dehulling before export from Ecuador, resulting in one case in a material concentrated in saponins and ready for extraction and in the other a high fiber bran. A second, modified "polishing" gives the germ fraction which can be used for oil extraction. In preliminary trials, we found this germ fraction to contain 40% protein. Quinoa protein is of an exceptionally high quality. Koziol (1992) summarized the results of four different studies on the protein efficiency ratio (PER) of quinoa in feeding trials with rats, expressing the PERs as percentages of the casein control diets. Raw quinoa (both sweet and bitter) exhibited PER values from 44 to 93% and cooked quinoa PER values from 102 to 105%; in comparison, raw and cooked wheat exhibited PER values from 23 to 32% of the casein control. It is rare for a vegetable protein such as that from quinoa to approximate so closely the quality of casein.

Comparing the profile of the essential amino acids (in human nutrition) of quinoa with that of maize, rice, and wheat shows that quinoa protein is particularly rich in lysine and contains more histidine and methionine + cystine (Table 7). Of the non-essential amino acids quinoa contains more arginine and glycine but less glutamic acid and proline than the cereals. The protein in the quinoa oil press cake would be an important complementary protein for improving the nutritional quality of both human and animal foodstuffs.

Carbohydrate Cream Substitute

The endosperm remaining after degerming the quinoa grain contains a starch with rather unusual qualities. The majority of the starch grains are less than 3 µm in diameter (Wolf et al. 1950; Scarpati de Briceño and Briceño 1982; Atwell et al. 1983). Small granule starches generally exhibit gelatinization temperatures higher than those of large granule starches (Kulp 1973; Swinkels 1985) but quinoa starch initiates gelatinization at temperatures similar to those for the larger granule wheat and potato starches, i.e. 56° to 58°C (Scarpati de Briceño and Briceño 1982; Swinkels 1985). Although quinoa starch initiates gelatinization at a temperature similar to that of wheat starch, its pasting behavior is considerably different and at equal starch concentrations shows higher viscosities than wheat starch when measured with a Brabender amylograph (Atwell et al. 1983).

Recently, the Nutrasweet Company exploited the properties of quinoa starch and filed a European patent for making a carbohydrate cream substitute from it (Singer et al. 1990). Although the procedure for obtaining the starch followed the method described by Atwell et al. (1983) for whole quinoa grains, there is no obvious reason why degermed quinoa could not be used instead.


Quinoa offers an oil rich in polyunsaturated fatty acids, a protein whose quality approaches that of casein and a starch that can be converted into a cream/fat substitute, all of which are easily marketable as products or as natural additives that should appeal to today's health conscious consumer. The saponins removed from bitter quinoa may find niches in pharmaceutical preparations or in programs of integrated pest management.

Despite being a promising "rediscovered" crop with important nutritional characteristics, the current industrial use of quinoa is limited by small-scale production which serves to keep prices for the grain too high to be commercially competitive with wheat, rice, and barley, especially in the Ecuadorian market. Fomenting further interest, research and the development of improved methods of commercial cultivation both locally and worldwide (see National Research Council 1989; Wahli 1990) will help ensure that quinoa regains the prominence it once enjoyed under the Incas.

In Ecuador, quinoa is already adapted for cultivation at altitudes from 2,300 to 3,500 m, too high for maize yields to be commercially viable (upper limits for maize for subsistence, not commercial, cultivation being 2,800 to 3,000 m; yields of wheat and barley also decline notably above 3,000 m). Quinoa should therefore be viewed as a versatile cash crop which would extend the range of commercially arable hecterage in Ecuador.


*Gratefully acknowledged are the efforts of U. Bracco and his research group at the Nestlé Research Centre, Vers-chez-les-Blanc, Switzerland, for performing the fatty acid and tocopherol analyses on samples of quinoa oil.
Table 1. Moisture and oil content of quinoa grains.

Moisture (%) Fat (%)
No. determinations Mean (Range) No. determinations Mean (Range) Reference
3 10.2 (9.8-10.5) 3 6.2 (5.5-6.7) De Bruin (1964)
58 12.7 (6.8-20.7) 60 5.0 (1.8-9.3) Cardozo and Tapia (1979)
58 12.9 (5.4-20.7) 54 4.6 (1.8-8.2) Romero (1981)
127 9.6 (6.2-14.1) 92 7.2 (4.3-9.5) Koziol (1990a)

Table 2. High oil content cultivars of quinoa (data from Alvarez et al. 1990 for quinoa grown at Cumbayá, Ecuador).

Source Accession Moisture (%) Oil (%)
Cambridgez Chilena-B 8.9 6.7
Chilena-T 9.6 6.8
No. 63 8.7 7.7
No. 63-1 6.8 6.9
INIAPy Ecu Sep 17-0271 11.0 6.9
V-8 9.4 7.5
V-10 8.7 7.9
V-11 9.6 8.0
San Juan 0036 7.1 7.8
Latinrecox Potoroc 7.2 7.2
011Pn 10.1 7.6
011Pr 9.7 7.7
012 7.7 7.8
012Pn 8.1 7.8
012Pr 10.1 8.7
013 9.4 7.5
013Pn 9.7 8.0
013Te 9.5 8.5
zSeed supplied by N.W. Galwey, Department of Genetics, University of Cambridge, England.
ySeed supplied by the Ecuadorian Instituto Nacional de Investigaciones Agropecuarias.
xEcotypes isolated by the Department of Agronomy, Latinreco, S.A., with the exception of 'Porotoc'.

Table 3. Composition of six genotypes of quinoa on a dry matter basis (data from Alvarez et al. 1990 for quinoa grown at Cumbayá, Ecuador).

Genotype Oil (%) Protein (%) Fiber (%) Ash (%) Carbohydrate (%) Saponins (%)
Porotoc 7.8 19.0 3.3 2.6 67.1 0.2
V-8 8.3 18.1 3.1 3.1 66.4 1.0
012 8.5 19.0 4.2 3.4 64.7 0.2
V-10 8.7 17.6 3.7 3.6 65.7 0.7
013Te 9.4 16.7 3.5 3.1 67.1 0.2
013Pr 9.7 16.6 3.0 3.0 67.5 0.2

Table 4. Grain yield of quinoa (data from Burgasi et al. 1990; seed moisture and fat content as per Table 2).

Grain yield (kg/ha)z
Accession 1986 1987 1988 Mean
Porotoc 2200 --- --- 2200
V-8 1360 --- --- 1360
012 --- 4500 3063 3782
V-10 2200 --- --- 2200
013Te 2700 4500 --- 3600
012Pr 3000 4000 2956 3319
zYield data are from 0.2 to 0.5 ha experimental plots at an altitude of 3,100 m.

Table 5. Comparison of oil yields from quinoa and maize.

Oil yield (kg/ha)
Source Range of fat content (%)z General Ecuador
Maize 2-5y 254x 34-85w
Quinoa 2-9v 108-488u 9-40w
zAt normal seed moisture content.
yDuke and Atchley (1986).
xPryde and Doty (1981).
wMAG (1985).
vFrom Table 1.
uCalculated on the basis of yields reported in Nuñez and Morales (1980).
tCalculated on the basis of oil content from Table 2 and yield data from Table 4.

Table 6. Comparison of the compositions of quinoa and maize oils.

Variable Quinoa Maize
Fatty acids (as % of lipid fraction):
Myristic (C14:0) 0.2z --- 0.2x
Palmitic (C16:0) 9.9z 11y 11.2x
Palmitoleic (C16:1) 0.1z --- 0.1x
Stearic (C18:0) 0.8z 0.7y 2.1x
Oleic (C18:1) 24.5z 22y 29.8x
Linoleic (C18:2) 50.2z 56y 55.0x
Linolenic (C18:3) 5.4z 7y 0.9x
Arachidic (C20:0) 0.7z --- 0.4x
Specific gravity 0.891w 0.918-0.925v
Refractive index 1.464w 1.464-1.468u
Saponification value 190w 189-191v
Iodine value (Wijs) 129w 125-128v
Unsaponifiable matter (%) 5.2w 0.8-2.9u
Sterols (%) 1.51w 0.85-1.42t,s
Lecithins (%) 1.8w 1-3r
zU. Bracco (pers. commun.).
ySánchez Marroquín (1983).
xData compiled from Leibovitz and Ruckenstein (1983) and Weiss (1983).
wDe Bruin (1964).
vMounts and Anderson (1983).
uEckey (1954).
tSouci et al. (1986).
sWeber (1983).
rPatterson (1989).

Table 7. Comparison of amino acid profiles (Koziol, 1992).

Content (g amino acid/100 g protein)
Amino acid Quinoa Maize Rice Wheat
Essential (for humans):
Histidine 3.2 2.6 2.1 2.0
Isoleucine 4.4 4.0 4.1 4.2
Leucine 6.6 12.5 8.2 6.8
Lysine 6.1 2.9 3.8 2.6
Methionine + Cystine 4.8 4.0 3.6 3.7
Phenylalanine + Tyrosine 7.3 8.6 10.5 8.2
Threonine 3.8 3.8 3.8 2.8
Tryptophan 1.1 0.7 1.1 1.2
Valine 4.5 5.0 6.1 4.4
Alanine 4.5 7.3 6.0 3.6
Arginine 8.5 4.2 6.9 4.5
Aspartic acid 7.8 6.9 10.0 5.0
Glutamic acid 13.2 18.8 19.7 29.5
Glycine 6.1 4.0 4.7 4.0
Proline 3.3 9.1 4.9 10.2
Serine 4.1 5.1 6.3 4.8

Last update April 17, 1997 aw