Selectively increasing of polyunsaturated (18:2) and monounsaturated (18:1) fatty acids in Jatropha curcas seed oil by crystallization using D-optimal design

Background This study was done to obtain concentrated polyunsaturated fatty acid (PUFA) linoleic acid (LA; 18:2) and monounsaturated fatty acid (MUFA) oleic acid (OA; 18:1) from Jatropha curcas seed oil by urea complexation. Urea complexation is a method used by researchers to separate fatty acids (FAs) based on their molecular structure. Effects the ratio of urea-to-FAs, crystallization temperature and crystallization time on the final products of urea complexation were examined. D-optimal Design was employed to study the significance of these factors and the optimum conditions for the technique were predicted and verified. Results Optimum conditions of the experiment to obtain maximum concentration of LA were predicted at urea-to-FAs ratio (w/w) of 5:1, crystallization temperature of −10°C and 24 h of crystallization time. Under these conditions, the final non-urea complex fraction (NUCF) was predicted to contain 92.81% of LA with the NUCF yield of 7.8%. The highest percentage of OA (56.01%) was observed for samples treated with 3:1 urea-to-FAs ratio (w/w) at 10°C for 16 h. The lowest percentage of LA (8.13%) was incorporated into urea complex fraction (UCF) with 1:1 urea-to-FAs ratio (w/w) at 10°C for 8 h. Conclusions The separation of PUFA (LA) and MUFA (OA) described here. Experimental variables should be carefully controlled in order to recover a maximum content of PUFA and MUFA of interest with reasonable yield% with a desirable purity of fatty acid of interest.


Background
Linoleic acid (LA) [also called cis,cis,-9,12-octadecadienoic acid] is an example of a polyunsaturated fatty acid (PUFA), due to the presence of two carbon double bonds. The high content of LA makes Jatropha curcas seed oil very important for industry use. LA can be used in protective coatings, plastics, surfactant, dispersants, biolubricant, and a variety of synthetic and in the preparations of other long chain compounds. The high content of LA in seed oil of J. curcas is very important to the production of oleo-chemicals [1]. Oleic acid (OA) [also called (9z)-octadec-9-enoic acid] is an example of a monounsaturated fatty acid (MUFA). A small amount of OA is used in the pharmaceutical industry, as an emulsifying agent in aerosol products [2].
There are several methods which can be used to obtain polyunsaturated fatty acids (PUFA) including freezing crystallization, urea complexation, molecular distillation, supercritical fluid extraction, silver ion complexation and lipase concentration [3] as well as high-performance liquid chromatography [4]. The most economic and most efficient technique to obtain LA in the form of fatty acids (FAs) is urea complex fractionation. This is a well-established technique for the elimination of saturated fatty acids (SFAs) and MUFA [5].
Urea complexation has the advantage that the complex crystals are extremely stable, and filtration is not carried out at low temperatures which is required for solvent crystallization of FAs. This method is preferred by many researchers because complexation depends upon configuration of the FAs moieties due to the presence of multiple double bonds, rather than pure physical properties such as melting point or solubility [5,6]. The SFAs and MUFA easily form complexes with urea and crystallize out at cooling during urea complex fraction (UCF). These complexes can subsequently be removed by filtration. The liquid or non-urea complex fraction (NUCF) is enriched with PUFA and the crystals formed or UCF consists of SFAs and MUFA.
In this study, urea complex fractionation of a mixture of FAs of Malaysian J. curcas seed oil was carried out to obtain concentrated PUFA. The effects of urea-to-FAs ratio, crystallization temperature and crystallization time to the yield% of NUCF (Y 1 ), yield% of UCF (Y 5 ), percentage MUFA (OA) (Y 3 and Y 7 ) and percentage PUFA (LA) (Y 4 and Y 5 ) in NUCF and UCF were systematically studied.

Results and discussion
Non-urea complex fraction (NUCF) The original fatty acids (FAs) mixture was composed of 13.19% palmitic (16:0), 6.37% stearic (18:0), 43.33% oleic (18:1) and 36.71% linoleic (18:2) acid. Average molecular weight of the FAs was 203.36 as obtained from saponification test of the original oil. The results compared well with those of [7]. The PUFA (LA) concentrate was prepared by urea complex fractionation following the technique of [8], using the FAs that was previously obtained. The purpose of this procedure was to obtain a PUFA concentrate enriched in LA and simultaneously, maintain the highest yield% of LA. The crystallization process with urea preferentially selects SFAs and MUFA, and the tendency of FAs to combine with urea decreases with increasing chain lengths [9].
In this study, variations of factors that affect the urea complex fractionation such as the ratio of urea-to-FAs (w/w), crystallization temperature (°C) and crystallization time (h) were examined to obtain optimum conditions using the response surface method D-optimal design. Table 1 shows data obtained from the experiment on FAs composition in the NUCF of all the samples. Results showed that the percentage of LA has increased from 36.71% to as much as 92.81% while SFAs (0.33%) has been reduced considerably compared to the initial FAs mixture. In samples with high ratio of urea-to-FAs, the elimination of SFAs was near completion in NUCF. However, total removal of oleic, palmitic and stearic acids by urea complexation may be impossible because some of the SFAs do not form complexes with urea during crystallization [10]. These results demonstrate that oleic, palmitic and stearic acids have more tendencies to form urea adducts than LA.
The LA% derived from the NUCF phase was relatively high, and some even greater than 90% under certain experimental conditions (Table 1). This showed that the experimental conditions were suitable for the preparation of high purity LA. However, it is difficult to completely remove all the SFAs and MUFA to obtain 100% purity of PUFA in the concentrate. [11] reported that complete removal of SFAs and MUFA by urea complexation may be impossible since some of the SFAs do not bind with urea during crystallization.
The quadratic regression coefficient obtained by employing a least squares method to predict quadratic polynomial models for the yield% of NUCF (Y 1 ), percentage of SFAs (palmitic and stearic acids) (Y 2 ), percentage of MUFA (OA) (Y 3 ) and percentage of PUFA (LA) (Y 4 ) in NUCF are given in Tables 2, 3, 4 and 5 respectively.
Examination of these coefficients with a T-test shows that the percentage yield of NUCF (Y 1 ), percentage of MUFA (OA) (Y 3 ), percentage of PUFA (LA) (Y 4 ), the linear term of urea-to-FAs ratio (X 1 ) and quadratic term of urea-to-FAs were highly significant (p < 0.01), while the percentage of SFAs (Y 2 ), the linear term was significant at p < 0.05. Lastly, linear term of crystallization time (X 3 ) for the percentage of PUFA (LA) (Y 4 ) and percentage of MUFA (LA) (Y 3 ) in the concentrate were significant at p < 0.05.
The results suggest that the linear effect of urea-to-FAs ratio and crystallization time are the primary determining factors for FAs separation by urea complexation. [10] concluded that these two variables significantly influenced the results of their urea complexation study. Crystallization time was found to be the insignificant factor (P > 0.05). This finding is in agreement with the results reported by other researchers [5,6,10].
Notes: ** P < 0.05; *** P < 0.01. T: F test value. See Table 1 for a description of the abbreviations.  Figure 4). The relationships between the parameters and FAs percentages were linear or almost linear. High concentration of PUFA (LA) could be obtained by using high ratio of urea-to-FAs at low temperatures. However, this could also reduce the yield% of liquid NUCF in the final product as more LA would be lost into urea adducts. Experimental variables should therefore be carefully controlled in order to recover a maximum content of PUFA (LA) of interest with reasonable yield% [5].
Straight-chained molecules such as SFAs readily formed stable adduct with urea. SFAs formed complexes more readily than MUFA. MUFA formed more readily inclusion compounds than PUFA (LA). Similar complexation tendency patterns were also obtained by [12]. The addition of more urea could reduce the SFAs percentage in NUCF to a minimum level; it however results in indiscriminate FAs complexation and thus reducing the amount of MUFA (OA) and PUFA (LA). A lower urea-to-FAs ratio prevented indiscriminate FAs complexation. Lower crystallization temperature can facilitate formation of more stable urea adducts, that would reduce SFAs in NUCF. Longer periods of crystallization time would allow the crystals to further stabilize. However the parameters must be set at a level to achieve an acceptable yield% of product with high purity. Higher purity of PUFA (LA) will always give lower yield of NUCF.
Optimum conditions using D-optimal design to obtain maximum concentration of PUFA (LA) and minimum concentration of both SFAs (palmitic and stearic acids) and MUFA (OA) were predicted at a urea-to-FAs ratio (w/w) of 5:1, crystallization temperature of −10°C and 24 h of crystallization time. The final NUCF was predicted to contain 0.33% of SFAs (palmitic and stearic acids), 5.73% of MUFA (OA) and 92.81% of PUFA (LA) with the NUCF yield of 7.8%. The observed value was reasonably close to the predicted value as shown in Figures 5, 6, 7 and 8. Urea complex fraction (UCF) The crystallization process with urea complex fraction (UCF) selects SFAs (palmitic and stearic acids) and MUFA (OA), simultaneously maintain the highest yield% of SFAs and MUFA (OA). The new tendency of FAs to combine with urea decreases with increasing UFAs [9].    The quadratic regression coefficient obtained by employing a least squares method to predict quadratic polynomial models for the yield% of solid UCF (Y 5 ), percentage SFAs (palmitic and stearic acids) (Y 6 ), percentage MUFA (OA) (Y 7 ) and percentage PUFA (LA) (Y 8 ) are given in Tables 8, 9, 10 and 11 respectively.
Linear term of urea-to-FAs ratio was highly significant (p < 0.01) for the yield% of UCF (Y 5 ) and percentage SFAs (Y 6 ), while the linear term of urea-to-FAs was significant (P < 0.05) for the percentage PUFA (LA) (Y 8 ). The interaction between urea-to-FAs ratio and crystallization temperature were significant (p < 0.05) for the percentage PUFA (LA) (Y 8 ) and the percentage SFAs (Y 6 ). Quadratic term of urea-to-FAs ratio was       also significant (p < 0.05) for the percentage MUFA (OA) (Y 7 ). The coefficients of the independent variables (urea-to-FAs ratio; X 1 , crystallization temperature; X 2 and crystallization time; X 3 ) determined for the quadratic polynomial models are lists in Tables 8, 9, 10 and 11 respectively. Table 8 lists the yield% of solid UCF (Y 5 ), Table 9 lists the percentage SFAs (palmitic and stearic acids) (Y 6 ), Table 10 lists the percentage MUFA (OA) (Y 7 ) and Table 11 lists the percentage PUFA (LA) (Y 8 ).
The fitted models for ANOVA are summarized in Table 12. Examinations of the two models with an F-test Notes: ** P < 0.05; *** P < 0.01. T: F test value. See Table 7 for a description of the abbreviations. Notes: ** P < 0.05; *** P < 0.01. T: F test value. See Table 7 for a description of the abbreviations. Notes: ** P < 0.05; *** P < 0.01. T: F test value. See Table 7 for a description of the abbreviations. Notes: ** P < 0.05; *** P < 0.01. T: F test value. See Table 7 for a description of the abbreviations.
Figures 9, 10, 11 and 12 also represent the Design-Expert plots for all the responses. In the solid UCF, performing the technique using low amount of urea without cooling would give the desired high percentage of SFAs and MUFA (OA) as shown in Figures 10 and 11, respectively. PUFA (LA%) (Figure 12) was lower under these conditions. The observed value was reasonably close to the predicted value as shown in Figures 13, 14, 15 and 16.

Conclusion
Optimum conditions of the experiment to obtain maximum concentration of PUFA (LA), were predicted at urea-to-FAs ratio (w/w) of 5:1, crystallization temperature of −10°C and 24 h of crystallization time. The final NUCF At this condition was predicted to contain 92.81% of PUFA (LA) with a NUCF yield of 7.8%. The highest percentage MUFA (OA) (56.01%) was observed for sample treated with a urea-to-FAs ratio (w/w) of 3:1 at 10°C for 16 h. The lowest percentage PUFA (LA) (8.13%) was incorporated into the UCF with a urea-to-FAs ratio (w/w) of 1:1 at 10°C for 8 h. All of the above mentioned factors have to be controlled to yield a reasonable amount of product with a desirable purity of FAs.

Experimental and Methods
FAs were obtained by the hydrolysis of J. curcas seed oil, as carried out by [13]. The separation of PUFA from the hydrolyzed FAs of J. curcas seed oil was carried out using the technique describe by [5]. FAs of J. curcas seed oil (10 g) were mixed with urea in 95% aqueous ethanol and heated at 60°C with stirring until the mixture turned into a clear homogeneous solution. The ratio of urea-to-FAs (Table 13) was changed by using different amounts of urea (1-5 g Figure 10 Response surface (a) and contour plots (b) for the effect of the urea-to-FAs ratio (X 1 , w/w) and crystallization temperature (X 2 ,°C) on the SFAs% (Y 6 ) of UCF.
were separated from the liquid (NUCF) by fast filtration. The liquid (NUCF) was diluted with an equal volume of water and acidified to pH 2-3 with 6 N HCl; an equal volume of petroleum ether was subsequently added and the FAs were extracted. The non-aqueous phase (top phase), containing liberated FAs, was separated from the aqueous layer containing urea by filtration. The petroleum ether layer was washed with 5% NaCl solution to remove any remaining urea. The petroleum ether fraction was dried over anhydrous Na 2 SO 4 . The solvent was removed using a rotary evaporator with water bath temperature at 65°C and the FAs converted to FAME for gas chromatography (GC) analysis according to [6].       respectively. An initial screening step was carried out using the technique describe by [5] to select the major response factors and their values. The independent variables were X 1 , X 2 and X 3 representing the urea-to-FAs ratio (w/w), crystallization temperature (°C), and crystallization time (h). The settings for the independent variables were low and high values: urea-to-FAs ratio of 1 and 5; crystallization temperature of −10 and 10 and crystallization time of 8 and 24. Each variable was coded at three levels: -1, 0, and +1. A quadratic polynomial regression model was assumed for predicting individual Y variables. The model proposed for each response of Y was: Where B 0 ; Bi; Bii and Bij are constant, linear, square and interaction regression coefficient terms, respectively, and xi and xj are independent variables. The goodness of fit of the model was evaluated by the coefficient of determination R 2 and the analysis of variance (ANOVA).
Open access provides opportunities to our colleagues in other parts of the globe, by allowing anyone to view the content free of charge.