3.1 Quality properties
Effects of different seed variety, extraction procedure and interaction of both factors on the physicochemical properties, oxidative stability indices and fatty acid profile of castor oils were obtained from general factorial analysis and the results (p-values) are presented in Table 1. Gibsoni castor variety yielded higher oil content on average between 40.12 and 53.51% than that of Carmenicita (38.18–44.97%) as shown in Table 2. These ranges were within those previously reported for castor seeds (34.6–56.6%) of varying genotypes 26). Solvent extraction yielded higher amount of oil when compared to screw press and traditional methods, which agrees with earlier observation of Panhwar et al. 27). The non-polar nature of the organic solvent used (n-hexane) contributed to its improved oil recovery capacity. However, a recent study of Díaz-Suárez et al. 28) indicated that more oil as high as 64% could be obtained from castor seed using optimized aqueous enzymatic extraction. The color index of the oils varied significantly between different varieties and extraction methods. Castor oil is expected to be colorless or pale yellow. Method of extraction had a considerable influence on the color. Oils from solvent extraction appeared more intense especially in Carmenicita variety. The capacity of organic solvent to dissolve various minor components such as carotenoids, gossypol, tocopherols, chlorophylls, and other pigments; thereby adding more hue to the oils, may account for higher color intensity in solvent extracted samples. Somewhat high temperature treatment during traditional procedure may have caused a significant thermal degradation of pigments 29) hence their low color index values. Specific gravity is one of many temperature-dependent physicochemical properties of oils that predict its secondary applicability. There was no significant variation among the cultivar, but traditionally extracted oils seemed denser than others. However, the values (0.96– 0.97) were in consonance with the current ASTM’s prescription for high quality castor oil 26). In the case of moisture content, both variety and extraction method were significant, but not their interaction. Despite, similar pre-extraction conditioning of the seeds, Carmenicita oil retained more moisture than that of Gibsoni. The rudimentary nature of the traditional method of processing could possibly increase the likelihood of the oil retaining more moisture, than semi-automated screw pressing and solvent extraction procedures. However, the range is still within allowable limit for castor oils according to ASTM recommendation (0.001–2.5%). The downside of high moisture-containing oil is it vulnerability to hydrolytic rancidity that stems from increased acidity 30) and eventual loss of quality.
Table 1
ANOVA p-values indicating the factors and interaction term (V: castor’s variety; E: extraction methods) effects on the quality parameters, oxidative stability indices and fatty acid profile of castor oils.
Parameters | Variety (V) | Extraction (E) | Interaction Term (V + E) |
Physicochemical properties | | | |
Yield (%) | 0.00 | 0.00 | 0.01 |
Color index | 0.00 | 0.00 | 0.07 |
Specific gravity | 0.23 | 0.10 | 0.66 |
Moisture (%) | 0.00 | 0.00 | 0.98 |
Oxidative stability indices | | | |
Peroxide value (meqO2/kg) | 0.00 | 0.00 | 0.00 |
Free fatty acid (% oleic) | 0.00 | 0.00 | 0.00 |
Fatty acid profile (%) | | | |
C14:0 | 0.28 | 0.83 | 0.67 |
C16:0 | 0.00 | 0.02 | 0.21 |
C18:0 | 0.00 | 0.00 | 0.02 |
C18:0® | 0.00 | 0.05 | 0.01 |
C18:1 | 0.95 | 0.00 | 0.00 |
C18:1® | 0.18 | 0.00 | 0.84 |
C18:2 | 0.00 | 0.00 | 0.00 |
C18:3 | 0.00 | 0.00 | 0.24 |
C20:0 | 0.00 | 0.02 | 0.00 |
C20:1 | 0.00 | 0.00 | 0.00 |
MUFA | 0.33 | 0.00 | 0.70 |
PUFA | 0.00 | 0.00 | 0.00 |
SFA | 0.06 | 0.00 | 0.02 |
MUFA/PUFA | 0.00 | 0.00 | 0.01 |
Quality parameters | | | |
Saponification value (mgKOH/g) | 0.34 | 0.11 | 0.31 |
Iodine value (gI2/100g) | 0.16 | 0.00 | 0.79 |
p-value < 0.05: significant; p-value < 0.01: very significant; p-value > 0.05: not significant |
C14:0: Myristic acid, C16:0: Palmitic acid, C18:0: Stearic acid, C18:0® : 9, 12-Dihydroxy stearic acid, C18:1: Oleic acid, C18:1®: Ricinoleic acid, C18:2: Linoleic acid, C18:3: Linolenic acid, C20:0: Arachidic acid, C20:1: Cis-11-Eicosenoic acid, SFA: Sum of all saturated fatty acids, MUFA: Sum of all monounsaturated fatty acids, PUFA: Sum of all polyunsaturated fatty acids |
Table 2
Physicochemical properties of castor oils of two varieties under three extraction procedures
Parameters | Gibsoni | Carmencita | |
Traditional | Screw | Solvent | Traditional | Screw | Solvent |
Yield (%) | 46.86 ± 1.18bA | 40.12 ± 1.21cA | 53.51 ± 1.82aA | 40.56 ± 1.66bB | 38.18 ± 1.07bB | 44.97 ± 1.20aB |
Color index | 2.04 ± 0.07cB | 2.82 ± 0.21bB | 3.95 ± 0.09aB | 3.19 ± 0.06cA | 3.59 ± 0.13bA | 4.96 ± 0.14aA |
Specific gravity | 0.97 ± 0.002aA | 0.96 ± 0.001aA | 0.96 ± 0.002aA | 0.97 ± 0.001aA | 0.96 ± 0.002aA | 0.96 ± 0.003aA |
Moisture content (%) | 0.68 ± 0.03aB | 0.50 ± 0.05bB | 0.37 ± 0.05cB | 0.92 ± 0.11aA | 0.72 ± 0.06bA | 0.60 ± 0.02cA |
a−c Mean values with different lowercase superscript letters in the same row indicate significant different in oils of different extraction methods (P < 0.05) |
A−B Mean values with different capital superscript letters in the same row indicate significant difference in oils of different varieties (P < 0.05) |
3.2 Oxidative and quality indices
Seed oils are ranked into different quality categories using some oxidative stability indices such as peroxide value (PeV) and free fatty acids (FFA) (Fig. 2). When these parameters are scaled against standard values, then the application of the oils can be predicted. Inedible oil as in the case of castor oil, have shown a remarkable promise as biofuel in petrochemical industry. Therefore, these quality parameters are of particular importance in determining both storage stability of the oils and their behavior under subsequent treatments. PeV of the oils differ significantly (P = 0.00) with respect to variety, extraction technique and their interaction (Fig. 2a). Carmenicita oils exhibited overall higher PeV than Gibsoni, thus indicating low accumulation of primary oxidative products (hydroperoxide) in Gibsoni oils and less vulnerability to rancidity. More importantly, in both varieties, traditionally processed oils had the lowest PeV (2.31–4.50 meqO2/kg), followed by solvent extracted (3.94–6.42 meqO2/kg), while screw pressed oils (5.98–7.21 meqO2/kg) were the least desirable in this regard. However, these ranges did not deviate from the usual reports in literature on castor oils 14, 31). One of the criticisms of PeV as an accurate predictor of oil stability is that the measurement is a non-selective analytical technique that quantifies all substances that can oxidize potassium iodide and take these oxidized products as peroxides 32). As imprecise as this might be, high PeV is never a desirable quality parameter for both edible and inedible oils. The free acidity of the oils varied with both factors (P = 0.00) and followed the same pattern as PeV (Fig. 2b). FFA of the oil reflects the intrinsic dissimilarities between the two seed varieties as Carmenicita oil showed higher FFA than Gibsoni. Screw press produced oils of more FFA (3.08–3.94%) content than those of solvent (2.18–3.46%) and traditional extraction (0.99–2.91%) methods. In addition to monounsaturated nature of castor oils, combined effects of high temperature treatment and decantation during traditional extraction method present a partial refining process that may have contributed to the low free acidity of the oils. In the same vein, several hours of refluxing during solvent extraction may have reduced FFA content in solvent extracted oils. A gradual reduction in FFA was recorded in kenaf seed oil during refining unit operation 29). Similarly, thermally pretreated castor seeds resulted into a substantial reduction in FFA of its oil 33). Free acidity mitigates against the catalytic effects of NaOH during biodiesel production. However, for a good biodiesel fuel to be produced from this high FFA oils, esterification must take place in the presence of acids 34). Unlike refined oils, storage stability of crude castor oil may be compromised due to it comparatively higher acidity.
3.3 Fatty acid distribution
The fatty acids profile (% weight of methyl esters) was characterized in the oil samples and comparisons were made between different varieties and extraction techniques. As shown in Table 3, ten fatty acids were identified in total, 5 of them are saturated (SFA), 3 monounsaturated (MUFA) out of which ricinoleic acid represents over 90%, and 2 polyunsaturated (PUFA) fatty acids. Generally, fatty acid distribution of the oils from the two varieties appears similar with occasional variations based on their descriptive statistics. In addition to ricinoleic acid (85.93–89.19%), other major fatty acid acids that characterized castor seeds oils are oleic acid (4.73–5.84%), stearic acid (1.41–2.50%), linoleic acid (1.08–3.41%), and palmitic acid (0.60–1.29%). These distributions are similar to those reported in literature 17, 35− 36) with slight variations that could be attributed to geographical origin differences. For instance, Salimon et al. 35) reported 84.2, 90.2 and 94.0% ricinoleic acid contents for castor oils of Malaysia, Brazil, and India origins, respectively. From ANOVA p-values (Table 1), both factors had no significant effect on myristic acid (C14:0) contents of the oil, whereas other fatty acids varied with respect to one or both factors. They were significant for palmitic acid (C16:0) and stearic acid (C18:0) – the two major saturated fatty acids of the oil in addition to a minor C18:0®. Oils of Gibsoni had higher C16:0 than those of Carmenicita, while the opposite was true for C18:0. Screw pressed, and solvent extracted oils were slightly more saturated than those of traditional methods. Comparatively, traditional extraction involved higher temperature treatment which may have influenced the oil’s saturation. Machmudah et al. 37), reported a slight decrease in saturated fatty acids of rosehip oils especially C18:0; with an increase in extraction temperature. The contents of the two major monounsaturated fatty acids of the oil; ricinoleic (C18:1®) and oleic (C18:1) acids did not vary significantly with variety (P > 0.05). However, method of extraction influenced these fatty acids, with traditionally extracted oils showing higher contents of C18:1® and screw pressed oils slightly richer in C18:1. However, the amount of these acids in both castor seed varieties (C18:1®:85.93–89.19; C18:1: 4.73–5.84) fall within the range reported from other studies 38 − 39). Studies have attributed break thermal efficiency and low fuel consumption of diesel engine containing up to 15% castor oil substitution, to its high ricinoleic content 40), thus making these oils suitable for biofuel production. Generally, legumes’ oils are known for their polyunsaturated fatty acids (PUFA) and vulnerability to oxidation. However, the monounsaturated nature of castor oil makes it less susceptible to oxidation while providing hydroxyl functional group that allows for easy modification 41). C18:2 is the major polyunsaturated fatty acid characterized and its variation with respect to variety and extraction techniques were significant (P < 0.05). Carmenicita had higher value of the acid especially in solvent extracted oils. The disparities between the samples with respect to calculated MUFA (total monounsaturated fatty acids), PUFA (total polyunsaturated fatty acids), MUFA/PUFA ratio, and SFA (total saturated fatty acids), were due to the factors considered.
Table 3
Fatty acid profile of castor oils from two varieties under three extraction procedures
Fatty acids (%) | Gibsoni | Carmencita |
Traditional | Screw | Solvent | Traditional | Screw | Solvent |
C14:0 | 0.10 ± 0.03aA | 0.12 ± 0.05aA | 0.09 ± 0.02aA | 0.09 ± 0.02aA | 0.08 ± 0.02aA | 0.09 ± 0.04aA |
C16:0 | 0.94 ± 0.10bA | 1.24 ± 0.16aA | 1.29 ± 0.12aA | 0.60 ± 0.06bB | 0.81 ± 0.05aB | 0.72 ± 0.11aB |
C18:0 | 1.41 ± 0.13cB | 1.87 ± 0.20bB | 2.50 ± 0.17aA | 2.06 ± 0.12bA | 2.31 ± 0.14abA | 2.39 ± 0.10aA |
C18:0® | 0.99 ± 0.05aA | 1.03 ± 0.03aA | 0.75 ± 0.10bA | 0.55 ± 0.06aB | 0.59 ± 0.12aB | 0.61 ± 0.10aB |
C18:1 | 4.94 ± 0.09bA | 5.55 ± 0.08aA | 5.39 ± 0.21aA | 5.33 ± 0.26bA | 5.84 ± 0.10aA | 4.73 ± 0.20cA |
C18:1® | 88.93 ± 0.30aA | 85.93 ± 0.06cA | 86.53 ± 0.43bA | 89.19 ± 0.20aA | 86.30 ± 0.65bA | 87.23 ± 1.38bA |
C18:2 | 1.08 ± 0.03bB | 2.12 ± 0.03aB | 2.16 ± 0.11aB | 1.21 ± 0.09cA | 2.89 ± 0.16bA | 3.41 ± 0.12aA |
C18:3 | 0.66 ± 0.05bA | 0.91 ± 0.03aA | 0.95 ± 0.02aA | 0.53 ± 0.04cB | 0.79 ± 0.03bB | 0.89 ± 0.05aA |
C20:0 | 0.25 ± 0.03cB | 0.67 ± 0.04aA | 0.39 ± 0.05bB | 0.68 ± 0.07bA | 0.47 ± 0.04cB | 0.83 ± 0.08aA |
C20:1 | 0.42 ± 0.09aA | 0.18 ± 0.05bA | 0.07 ± 0.01cA | 0.04 ± 0.01bB | 0.15 ± 0.06aA | 0.03 ± 0.01bB |
MUFA | 94.29 ± 0.44aA | 91.66 ± 0.16bA | 91.99 ± 0.38bA | 94.56 ± 0.22aA | 92.29 ± 0.50bA | 92.00 ± 1.31bA |
PUFA | 1.74 ± 0.05bA | 3.02 ± 0.06aB | 3.11 ± 0.13aB | 1.73 ± 0.13cA | 3.68 ± 0.17bA | 4.30 ± 0.14aA |
SFA | 3.69 ± 0.24bA | 4.92 ± 0.36aB | 5.02 ± 0.14aA | 3.97 ± 0.24aA | 4.26 ± 0.26aA | 4.64 ± 0.23aB |
MUFA/PUFA | 54.22 ± 1.57aA | 30.32 ± 0.56bA | 29.61 ± 1.23bA | 54.64 ± 4.04aA | 25.14 ± 1.23bB | 21.40 ± 0.49cB |
C14:0: Myristic acid, C16:0: Palmitic acid, C18:0: Stearic acid, C18:0® :9,12-Dihydroxy stearic acid, C18:1: Oleic acid, C18:1®: Ricinoleic acid, C18:2: Linoleic acid, C18:3: Linolenic acid, C20:0: Arachidic acid, C20:1: Cis-11-Eicosenoic acid; SFA: Sum of all saturated fatty acids, MUFA: Sum of all monounsaturated fatty acids, PUFA: Sum of all polyunsaturated fatty acids |
a−c Mean values with different lowercase superscript letters in the same row indicate significant different in oils of different extraction methods (P < 0.05) |
A−B Mean values with different capital superscript letters in the same row indicate significant difference in oils of different varieties (P < 0.05) |
3.4 Saponification (SaV) and iodine values (IoV)
In contrast to other quality parameters, saponification values (SaV) of the oils were not affected by varietal difference and methods of extraction. However, the ranges recorded in both varieties (Gibsoni: 175–185 mgKOH/g; Carmenicita: 177–184 mgKOH/g) (Fig. 3a) regardless of the extraction techniques, agree with ASTM standard limits (175–187 mgKOH/g) and other authors 14, 27). Díaz-Suárez et al. 28) observed higher SaV (194.12–194.34 mgKOH/g) for castor oils extracted using solvent and aqueous enzymatic extractions while Arawande & Akinnusotu 42) reported lower values (120–124 mgKOH/g) from castor oils of different geographical locations. However, an extremely high SaV value for non-edible oil like castor oil makes it less useful as biodiesel due to the competing effects of soap formation during neutralization process 43). Therefore, since both varieties produced oils of acceptable SaV according to ASTM limit, they could be better suited in biofuel production. Another important quality parameter that qualifies a non-edible oil for this purpose, is iodine value (IoV) (Fig. 3b). The value denotes the degree of unsaturation of the oils, measured as the grams of iodine (I2) require to completely saturate 100 g of oil 44). Only the extraction method had a significant (P < 0.05) influence on this parameter. While solvent extraction produced oils of slightly higher iodine value (83–87 gI2/100g) than others, the values are still within ASTM (82–88 g I2/100g) and EN 14214 (120–140 g I2/100g) tolerant limits 27, 45). The relatively low IoV of castor oils could be attributed to the monounsaturated nature of its fatty acid (mainly ricinoleic acid) and minimum number of fatty acids with conjugated double bonds.
3.5 PCA evaluation of Castor Oil chemical data
A PCA model was built on the combined physicochemical data of the oils (18 samples x 22 variables). The score and loading plots representing the projection of observations and variables based on the first two principal components are presented in Fig. 4. The model had in total seven latent variables (LVs), and the first 2 explained 65% variance with 0.96 and 0.67 as coefficient of determination (R2) and cross-validation (Q2), respectively. PCA is particularly appropriate here, because it’s an unsupervised method of extracting realistic relationship between extraction conditions and chemical data of oils 46). From the score plot (Fig. 4a), there were clearly recognizable patterns corresponding to oils of different varieties (upper and lower parts of the ellipse) produced using different extraction methods. On the upper and lower left side of the quadrant (positive and negative PC 2), oils from traditional extraction method clustered together and the weighted variables predominantly influencing the separation were MUFA/PUFA and C20:1 for Gibsoni variety, and MUFA, C18:1®, specific gravity, moisture content, saponification, and iodine values for Carmenicita variety. If properly harnessed and optimized, traditional extraction methods could maximize production of higher ricinoleic-containing castor seed oils. On the right side of the ellipse, there are two distinct clusters (upper and lower quadrants) occupied separately by oils from the two castor varieties. Interestingly, these clusters are mixture of screw pressed and solvent extracted oils, which indicates similarities in some of their quality and chemical properties. For Gibsoni, both screw press and solvent extraction produced castor oils that are best defined by their high contents of C18:1, C18:3, C18:0®, C16:0, C14:0, SFA and yield, while Carmenicita oils from both extraction techniques are best distinguished by their peroxide values, free fatty acids, and total polyunsaturated fatty acids (PUFA). These two oxidative stability indices (PeV and FFA) that were significantly higher in screw press and solvent extracted Carmenicita oils, in addition with polyunsaturated fatty acids suggest high susceptibility of the oils to oxidative degradation. An earlier study observed a more rapid oxidative degradation of n-hexane extracted tree oil than other non-polar solvent 47). Similarly, high PUFA, PeV and FFA were found to have negative contributions to oxidative stability of vegetable oils 46).