A Balanites aegyptiaca sample was purchased from an Omdurman market (Souk Omdurman), Sudan, on January 2015. The epicarps of the Balanites aegyptiaca fruit were removed manually and the remaining edible flesh (mesocarp) plus the endocarps were soaked in tap water overnight. After complete removal of the edible flesh the endocarps were dried in sunlight for two days. Finally, the endocarps were removed manually using a hammer and the kernels were stored at 10 ºC for the next steps.

Moisture, ash, protein, and lipids were determined following the standard methods of the Association of Official Analytical Chemists (AOAC) [1990]. In each case triplicate analyses were done and the mean and the standard deviation were calculated (mean ± SD). Total carbohydrates were determined by difference and the standard deviation was calculated using propagation of error.

Acid, peroxide, and saponification values were determined following the standard methods of the AOCS Official Methods [AOCS, 2011]. In each case triplicate analyses were carried out and the mean and the standard deviation were calculated (mean ± SD).

The values reported in Tables 1 to 3 are the means ± the standard deviations (SD) of three replicates (two replicates for viscosity and refractive index). The statistical analysis was done with SPSS 20.0. Analysis of variance (ANOVA) was used to evaluate the significance between the raw and thermally processed samples. The variant means were separated using the least significant difference method (LSD). The level of significance was set at p < 0.05.

The results of the proximate analysis of the kernel seeds of raw and thermally processed samples are given in Table 1. As can be seen from the table, the raw sample contained 3.72% moisture, 2.88% ash, 42.41% crude protein, 39.98% crude fat, and 10.99% total carbohydrates (obtained by difference). These results fall within the range of previously reported values of the chemical composition of Balanites aegyptiaca kernel seeds (Hussain, et al. 1949; Nour et al. 1985; Mohamed, et al. 2002; Chapagain et al. 2009; Manji et al. 2013). In conclusion, the findings of these studies have revealed considerable variations in the chemical composition based on various genotypes of Balanites aegyptiaca and the different origins of the analyzed samples. The lowest and the highest values reported for crude fat were 39–49.9%, crude protein 26.1–50.37%, ash 3.3–6.3%, and moisture 3.1–5.7%. Compared to a raw sample, thermally processed kernel seeds did not show significant (p > 0.05) changes in crude protein, crude fat, or total carbohydrates. On the other hand, the findings of the study have shown a significant difference (p < 0.05) in the ash content between raw and boiled samples (2.88% for raw sample and 2.32% for boiled one). The leaching of minerals by water during the boiling process was reported by some authors (Arinola and Adesina, 2014; Hefnawy, 2011) as the main cause of this decrease in ash content. Furthermore, the moisture content of the roasted kernels changed significantly (p < 0.05) in comparison with the raw ones (Table 1). This variation in moisture content could be attributed to the effect of roasting which facilitates oil extraction by decreasing its viscosity, releasing oil from intact cells and removing moisture (Alenyorege, 2015).

The physical properties of the extracted oils from raw, boiled and roasted kernels are displayed in Table 1. As can be seen from the table, the crude oil of Balanites aegyptiaca is a liquid which has a light yellow color, a specific gravity of 0.9132, a viscosity of 63.4 cpoise, and a refractive index of 1.468. The values of the specific gravity, the color, and the refractive index of the oil are in good agreement with the reported values in the literature [Hussain, et al. 1949; Nour et al.1985; Manji et al. 2013; Chapagain et al., 2009]. However, only Chapagain et al. [2009] have reported the value of the viscosity and it was observably different from the present study; 49 cp aganist 63.5 cp repectively. Additionaly, the results demonstrated that both roasting and boiling did not have considerable effects (p > 0.05) on viscosities, refractive indices, and specific gravities of the extracted oils when compared to raw one. On the other hand, boiling has intensified the color (color changes from light yellow to deep yellow) as shown in Table 1. This could probably be due to the increase in α-carotene content as a result of thermal processing (boiling) which facilitates the distribution or extraction of α-carotene from kernels to oils. This was clearly observed from the color of the defatted cakes of the raw (yellowish) and the boiled (off white) samples. Previous study by Hussain et al. (1949) has shown that Balanites aegyptiaca oil contains α-carotene.

In the case of the roasted sample, the color of the oil has changed from light yellow to brownish yellow. Moreover, the color of the defatted cake was also changed to brown. The formation of brown color during roasting was noticed by some researchers (Akinoso, et al. 2011; Anjum, et al. 2006) and was explained by the formation of browning substance which results from Maillard-type non-enzymatic reactions, caramelization, and phospholipid degradation.

The chemical properties of the oils of Balanites aegyptiaca are presented in Table 1. As can be seen from the table, with the exception of peroxide value, the chemical properties of the oil extracted from the raw kernels compare well with the previous studies (Hussain, et al. 1949; Nour et al. 1985; Manji et al. 2013). Nour et al. (1985) have analyzed the oils of two different samples of Balanites aegyptiaca seeds originating from Sudan. They have found that the saponification values were 190.05 and 176.97, peroxide values 3.7 and 2.7, and acid values 0.40 and 0.50, respectively. Similarly, Manji et al. (2013) reported that the saponification and the peroxide values for Balanites aegyptiaca oil were 168.6 and 6.0, respectively.

The results also showed that the peroxide values of the oil from thermally processed seeds changed significantly (p < 0.05). Compared to the oil from the raw sample, the peroxide value of the oil from the roasted sample decreased drastically (p < 0.05); whereas a considerable increase (p < 0.05) in the peroxide value of the oil from the boiled sample was noticed. The drastic change in peroxide value in the boiled sample is probably due to the increase in oxidation of the oil caused by heating (boiling) which accelerates the oxidation processes as well as the hydrolysis in an aqueous medium which also increases the rate of rancidity. In contrast to boiling, the considerable decrease in peroxide value for the oil of the roasted seeds could be due to the formation of some materials which inhibit the oxidation of the oil. Roasting of seeds prior to oil extraction has been reported in some articles (Wijesundera, et al. 2008; Abou-Gharbia, et al. 1996) to enhance the oxidative stability of oils. In contrast, some authors have found that roasting has a negative effect on the oxidative stability of oils extracted from roasted seeds (Mariod, et al. 2012; Anjum, et al. 2006). Wijesundera et al. (2008) have reported that the roasting (at 165 ˚C for 5 min) of canola and mustard seeds increased the oxidative stability of the extracted oils and did not affect the content of tocopherols. The enhancement in oxidative stability of canola oil was attributed to the formation of 2,6-dimethoxy-4-vinyl phenol (DMVP) during roasting.

Table 2 displays the fatty acid composition of Balanites aegyptiaca oils. It is apparent that linoleic (average 31.92%), oleic (average 39.17%), stearic (average 13.56%), and palmitic (average 12.68%) acids are the major fatty acids of the oils which collectively represent about 97% of the total acids. Moreover, the unsaturated fatty acids constitute about 71% of all the fatty acids present in the oils. The monounsaturated fatty acids represent an average value of 32.52%, whereas the polyunsaturated fatty acids reach an average value of 39.43%. The higher content of the unsaturated fatty acids is favorable from a nutritionist’s point of view, although it is inferior with respect to oxidative stabilities of the oils because higher levels of unsaturation (level of polyunsaturated fatty acids specifically) is more susceptible to oxidation. Generally, the rate of oxidation was found to increase with the increase in the number of double bonds in the fatty acids (Kamal-Eldin, 2006; Savage, et al. 1999). Savage et al. (1999) have studied the oxidative stability of different samples of Walnut oil. The results revealed the presence of a correlation between the levels of C18:2 fatty acid in the oil and the reduction in the oxidative stabilities of the oil. Compared to the oil extracted from the raw sample, the fatty acid composition of the oil extracted from the processed samples remained roughly unchanged. The present results of the fatty acid composition of Balanites aegyptiaca oil are in good agreement with previous studies (Al Ashaal, et al. 2010; Chapagain, 2009; Hussain, et al. 1949; Mohamed, et al. 2002; Nour et al.1985) regarding the major fatty acids in the oil, however, except for Nour et al. (1985), significant variations in the percentages of these acids could be noticed. The noticeable variations in the percentages of fatty acids were clarified by Chapagain et al. (2009) who analyzed the fatty acid composition of six Balanites aegyptiaca genotypes. The percentage ranges of the four major fatty acids were found to be: palmitic acid 12.7–16.0%, stearic 10.2–12.1%, oleic acid 23.5–43.7%, and linoleic acid 31.50–51.6%. Moreover, the present study is the first study to report the presence of cis-vaccenic (0.80–0.88%), margaric (0.11–0.12%), and eicosenoic (gondoic) (0.1%) acids in Balanites aegyptiaca oil.

The tocopherol and tocotrienol (tocols) compositions of the oil of Balanites aegyptiaca are given in Table 3. It is obvious that α- and γ- tocopherols represent the major constituents of tocols although traces or very little quantities of δ-tocopherols were also detected. The total tocopherol contents of the oil extracted from the raw, boiled, and roasted samples were found to be 42.2 mg/100g, 39.85 mg/100g, and 40.05 mg/100g, respectively. Tocopherols were reported in the literature (Kamal-Eldin, 2006) as one of the essential components of vegetable oil which have activity against oxidation or the development of rancidity. Compared to the values of tocopherol contents in vegetable oil and industrial fats investigated by Schwartz et al. (2008), the tocopherol content of Balanites aegyptiaca oil is higher than coconut oil, refined olive oil, and extra virgin olive oil and is comparable to sesame oil. In addition, it is observably lower than sunflower oil, rapeseed refined and cold pressed oil, camelina oil, and wheat germ. Furthermore, the results have revealed the absence of tocotrienols as well as plastochromanol–8 (P8) although trace quantities of α- and γ-tocotrienols were detected in some samples (Table 3).

The oxidative stability of the oil is expressed as the induction period (IP) and was determined by Rancimat at 110 ˚C. The oxidative stability of the oil extracted from the raw kernels was 10.64 h. The highest stability was shown by the oil from the roasted sample (14.50 h) whereas the lowest stability was displayed by the oil from the boiled sample (7.14h). It has been reported in the literature that the oxidative stability of vegetable oil depends on its fatty acid composition and the antioxidants (mainly tocopherols) (Kamal-Eldin, 2006). However, in the present study, these variations were not seen because of the differences in fatty acid compositions or tocol contents as their values remain fairly unchanged between the oil from the raw and the processed samples. Possibly these differences could be due to the presence of other antioxidants or might be due to the effect of boiling which accelerates the oxidation of the oil. The peroxide values for the oil confirm this assumption.