Response surface optimization and characteristics of rambutan (Nephelium lappaceum l.) kernel fat by hexane extraction

LWT - Food Science and Technology 44 (2011) 1946e1951<br /> <br /> Contents lists available at ScienceDirect<br /> <br /> LWT - Food Science and Technology<br /> journal homepage: www.elsevier.com/locate/lwt<br /> <br /> Response surface optimization and characteristics of rambutan (Nephelium<br /> lappaceum L.) kernel fat by hexane extraction<br /> Wanrada Sirisompong, Wannee Jirapakkul, Utai Klinkesorn*<br /> Department of Food Science and Technology, Faculty of Agro-Industry, Kasetsart University, 50 Ngam Wong Wan Road, Chatuchak, Bangkok 10900, Thailand<br /> <br /> a r t i c l e i n f o<br /> <br /> a b s t r a c t<br /> <br /> Article history:<br /> Received 12 October 2010<br /> Received in revised form<br /> 7 April 2011<br /> Accepted 26 April 2011<br /> <br /> Response surface methodology (RSM) was used to study the effect of moisture content (1.59e18.41 g/<br /> 100 g), extraction time (2.3e10.7 h) and particle size (0.09e2.11 mm) on the fat yield from rambutan<br /> kernels using hexane extraction. The physical and chemical characteristics of rambutan fat were also<br /> determined. The optimum conditions obtained from response surface analysis was 4.99 g/100 g moisture, 1.05 mm particle size and 9.2 h extraction time. Under these optimum conditions, the maximum fat<br /> yield was 37.35 g/100 g. The extracted fat was a white solid at room temperature. The physical and<br /> chemical characteristics of the extracted fat compared well with those of conventional fats. The high<br /> level of arachidic acid (w 34.3 g/100 g fat) and low iodine value in rambutan kernel fat permits the use of<br /> the fat, especially where oxidation may be a concern, without its being subjected to hydrogenation.<br /> Ó 2011 Elsevier Ltd. All rights reserved.<br /> <br /> Keywords:<br /> Rambutan kernel<br /> Fat extraction<br /> Response surface methodology<br /> Fat characteristics<br /> <br /> 1. Introduction<br /> Rambutan (Nephelium lappaceum Linn.) is a seasonal fruit native<br /> to west Malaysia and Sumatra. It is cultivated widely in Southeast<br /> Asian countries and Thailand has become the leading producer<br /> (0.6e0.7 million tons a year) and exporter (w 10e15 million dollars<br /> US) of the fruit (Salakpetch, 2000). This fruit is generally consumed<br /> fresh, although in the main producing countries like Malaysia and<br /> Thailand it is industrially processed to obtain juice, jams, jellies and<br /> marmalades (Morton, 1987). In addition, rambutan fruits are also<br /> processed as rambutan stuffed with a chunk of pineapple and<br /> canned in syrup. The rambutan fruits are deseeded during processing and these seeds (w 4e9 g/100 g) are a waste by-product<br /> of the canning industry (Tindall, 1994). Some studies have reported that rambutan seed possesses a relatively high amount of fat<br /> with values between 14 g/100 g and 41 g/100 g (Augustin & Chua,<br /> 1988; Kalayasiri, Jeyashoke, & Krisnangkura, 1996; Morton, 1987;<br /> Solís-Fuentes, Camey-Ortíz, Hernández-Medel, Pérez-Mendoza, &<br /> Durán-de-Bazúa, 2010; Winayanuwattikun et al., 2008), and the<br /> increasing demand for oils and fats, whether for human<br /> consumption or for industrial purposes, necessitates the search for<br /> new sources of novel oils and fats. Therefore, the extracted fat from<br /> rambutan seed not only could be used for manufacturing candles,<br /> soaps, and fuels, it also has a potential to be a source of natural<br /> edible fat with possible industrial use.<br /> * Corresponding author. Tel.: þ662 562 5031; fax: þ662 562 5021.<br /> E-mail address: utai.k@ku.ac.th (U. Klinkesorn).<br /> 0023-6438/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.<br /> doi:10.1016/j.lwt.2011.04.011<br /> <br /> The main process for the separation and recovery of oils and fats<br /> from seeds is solvent extraction. The fat yield from the seeds<br /> depends on the nature of the solvent, the extraction temperature<br /> and time, seed particle size and pretreatment conditions (Becker,<br /> 1978). Recently, a little information has become available on<br /> rambutan fat extraction (Azam, Waris, & Nahar, 2005; SolísFuentes, Camey-Ortíz, Hernández-Medel Mdel, Pérez-Mendoza, &<br /> Durán-de-Bazúa, 2010). However, no work has been reported on<br /> the effect of moisture content, extraction time and particle size on<br /> the extraction efficiency of fat from rambutan seed kernels. In order<br /> to study the effects of independent variables on fat extraction,<br /> central composite design (CCD) with response surface methodology (RSM) is often used (Mani, Jaya, & Vadivambal, 2007; Shao,<br /> Sun, & Ying, 2008; Tan, Jinap, Edikusnadi, & Hamid, 2008; Wei,<br /> Liao, Zhang, Liu, & Jiang, 2009). RSM has increasingly been used<br /> for optimizing purposes due to its efficiency and lower data<br /> requirements. This experimental methodology consists of a group<br /> of mathematical and statistical procedures that can be used to<br /> study the relationships between one or more responses and<br /> a number of independent variables. It also defines the effect of the<br /> independent variables, alone or in combination, in the process.<br /> Beside analysis the effects of independent variables, RSM creates<br /> a mathematical model that accurately describes the overall process<br /> (Hu, 1999; Montgomery; 2001). The main objective of the present<br /> study was to determine and explain the effects of moisture content,<br /> extraction time and particle size on the yield of rambutan kernel fat<br /> using hexane extraction. A model equation that would predict<br /> and determine the optimum conditions for total fat yield was<br /> <br /> W. Sirisompong et al. / LWT - Food Science and Technology 44 (2011) 1946e1951<br /> <br /> developed, and as well, the physical and chemical characteristics of<br /> the extracted fat were also determined.<br /> 2. Materials and methods<br /> 2.1. Materials<br /> Rambutan seeds (N. lappaceum L.) were obtained from Universal<br /> Food Public Company Limited (Nakornpathom, Thailand). The<br /> seeds were washed and the kernels were removed manually from<br /> the seeds. The kernels were then ground and dried using a tray<br /> dryer (Kan Seng Lee Machinery (1960) Ltd. Part, Thailand) at 55  C<br /> for 5 h (w 5 g/100 g moisture). The prepared kernels were stored<br /> at À5  C until used for experiments. The selected extraction solvent<br /> was hexane which was purchased from Mallinckrodt Baker, Inc.<br /> (Phillipsburg, NJ). All other chemicals were reagent grade or higher.<br /> 2.2. Sample preparation and extraction process<br /> Based on preliminary experimental results, the chosen independent variables were moisture content, extraction time and<br /> particle size. The moisture content of ground kernels was adjusted<br /> by slowly adding sprayed water while the ground kernels were<br /> continuously mixed in an Imarflex mixer (IF 309, Thailand). These<br /> samples were left sealed overnight at 4  C to reach equilibrium<br /> (Sandoval & Barreiro, 2007). The moisture content of the rambutan<br /> kernel samples was determined in triplicate according to the air<br /> oven method AOAC 931.04 (AOAC, 2000). The ground particles were<br /> separately by size using an AS 200 Control sieve shaker (Retsch,<br /> Germany) and different particle sizes ranging from 0.09 to 2.11 mm<br /> were obtained. The particles sizes based on sieve numbers used for<br /> analysis were: 10 (2.0e2.36 mm), 12 (1.7e2.0 mm), 18<br /> (1.0e1.8 mm), 35 (0.5e0.6 mm) and 170 (0.09e1.1 mm). A lab scale<br /> Soxhlet apparatus was used to extract fat from the rambutan<br /> kernels. About 20 g of ground and dried kernels was used for each<br /> combination of process parameters. The amount of solvent used for<br /> fat extraction was 150 ml for each sample. The extraction temperature was approximately 65  C. The extracted fat was expressed as<br /> a percentage, which is defined as the weight of the extracted fat<br /> over the dry weight of the sample.<br /> 2.3. Characterization of rambutan kernel fat<br /> 2.3.1. Color, refractive index and crystal polymorphism<br /> The color of the extracted fat was measured using a Miniscan XE<br /> (Hunter Association Laboratory Inc., USA) according to the modified<br /> method of Cheikh-Rouhou et al. (2007). The refractive index of<br /> kernel fat was determined using an instrument from Atago Co. Ltd.<br /> Series No. 11506, Japan. The crystal polymorphism of the fat sample<br /> was analyzed using an X-ray diffractometer (JEOL JDX-3530, Japan)<br /> and the modified method of Reshma, Saritha, Balachandran, and<br /> Arumughan (2008).<br /> <br /> 1947<br /> <br /> G1530N, Agilent Technologies, Inc., USA). The gas chromatography<br /> apparatus was equipped with a Supelco SP-2560 capillary column<br /> 100 m  0.25 mm id with the film thickness of 0.2 mm (Supelco Inc.,<br /> USA) and FID detector, and operated in a split mode with a split<br /> ratio of 100:1. The injector and detector temperatures were 250  C.<br /> The column temperature was held at 140  C for 5 min, and then<br /> programmed to rise to 250  C at 3  C/min and then held for 17 min.<br /> The carrier gas used was helium set at a flow rate of 1.1 ml/min. The<br /> area percents were used to determine the relative amounts of each<br /> fatty acid. The response factors for each fatty acid according to<br /> AOAC 996.06 method were used to correct area percents. The fatty<br /> acid content was reported as grams per 100 g of fat.<br /> 2.5. Analysis of phytosterol content<br /> The phytosterol content in rambutan fat was determined<br /> according to the method of Schwartz, Ollilainen, Piironen, and<br /> Lampi (2008). Gas chromatography analysis was conducted using<br /> an HP 6890 (Hewlett Packard Co., USA) instrument equipped with<br /> an HP-5 capillary column 30 m  0.32 mm id with the film thickness of 0.25 mm and FID detector, and operated in a splitless mode.<br /> The injector and detector temperatures were 300  C. The column<br /> temperature was held at 245  C for 1 min, and then programmed to<br /> rise to 275  C at 3  C/min and then held for 20 min. The carrier gas<br /> used was helium set at a flow rate of 3 ml/min. The relative amount<br /> of each phytosterol was reported as milligrams per gram of fat.<br /> 2.6. Analysis of a-tocopherol content<br /> The a-tocopherol in rambutan fat was measured using an Agilent 1100 Series reverse-phase high performance liquid chromatography apparatus with a diode array detector (Agilent<br /> Technologies, Inc., USA). A Hypersil ODS column (250  4.0 mm,<br /> Alltech Associates, Inc., USA) was used as the stationary phase. The<br /> mobile phase with this column was 98 percents methanol and 2<br /> percents water and a flow rate of 1 ml/min was used. Column<br /> temperature was maintained at 25  C and the a-tocopherol was<br /> detected at a wavelength of 292 nm.<br /> 2.7. Thermal behavior of rambutan fat<br /> The thermal behavior experiment was conducted using a differential scanning calorimeter (Model DSC1, Mettler-Teledo International Inc., USA) and a method modified from Saloua, Eddine, and<br /> Hedi (2009) and Besbe, Blecker, Deroanne, Drira, and Attia (2004).<br /> Approximately 4e5 mg of the unrefined fat sample was placed in an<br /> aluminum sample pan and an empty aluminum pan was placed on<br /> the reference platform. A linear heating and cooling rate of 5  C/min<br /> over a temperature range of À50 to 90  C was used. The thermogram<br /> peak was used to provide an estimate of enthalpy (DH) and the<br /> thermogram peak points were used to determine the melting point.<br /> 2.8. Experimental design and statistical analysis<br /> <br /> 2.3.2. Iodine value, sponification number and unsponification<br /> matter<br /> The iodine value of fat samples was determined according to<br /> AOCS Cd 1d-92 (AOCS, 1997). The sponification number and<br /> unsponification matter of rambutan fat samples were measured<br /> according to AOCS Cd 3e25 and AOCS Ca 6a-40, respectively.<br /> 2.4. Analysis of fatty acid composition<br /> The fatty acids were hydrolyzed from fat and then methylated to<br /> fatty acid methyl esters (FAMEs). The fatty acid composition of fat<br /> was investigated using gas chromatography (Model 6890N<br /> <br /> The effect of three independent variables: moisture content (X1;<br /> 1.59e18.41 g/100 g), extraction time (X2; 2.3e10.7 h) and particle<br /> size (X3; 0.09e2.11 mm) on the response variable (Y, percent fat<br /> yield) was evaluated using a three factor central composite design<br /> (CCD). The five coded levels of moisture content, extraction time<br /> and particle size were incorporated into the design and were<br /> analyzed in 20 combinations (Table 1). The central point of the<br /> design was repeated six times to calculate the reproducibility of<br /> the method (Montgomery; 2001). For each combination of the<br /> independent variables in the experimental design, the dependent<br /> parameter percent for extracted fat was determined. The effect of<br /> <br /> 1948<br /> <br /> W. Sirisompong et al. / LWT - Food Science and Technology 44 (2011) 1946e1951<br /> <br /> Table 1<br /> Uncoded and coded levels of independent variables in the experimental design and<br /> percent extracted fat.<br /> Treatment Moisture<br /> runs<br /> content<br /> (X1, g/100 g)<br /> <br /> Extraction time<br /> (X2, hours)<br /> <br /> Particle size<br /> (X3, mm)<br /> <br /> Extracted<br /> fat (Y, g/100 g)a<br /> <br /> Uncoded Coded Uncoded Coded Uncoded Coded<br /> 1<br /> 2<br /> 3<br /> 4<br /> 5<br /> 6<br /> 7<br /> 8<br /> 9<br /> 10<br /> 11<br /> 12<br /> 13<br /> 14<br /> 15<br /> 16<br /> 17<br /> 18<br /> 19<br /> 20<br /> <br /> 10<br /> 5<br /> 10<br /> 10<br /> 15<br /> 10<br /> 10<br /> 5<br /> 10<br /> 15<br /> 10<br /> 5<br /> 18.41<br /> 10<br /> 15<br /> 10<br /> 1.59<br /> 10<br /> 5<br /> 15<br /> <br /> 0<br /> 10.705<br /> À1<br /> 4<br /> 0<br /> 6.5<br /> 0<br /> 6.5<br /> 1<br /> 9<br /> 0<br /> 6.5<br /> 0<br /> 6.5<br /> À1<br /> 9<br /> 0<br /> 6.5<br /> 1<br /> 4<br /> 0<br /> 6.5<br /> À1<br /> 9<br /> 1.682 6.5<br /> 0<br /> 2.295<br /> 1<br /> 9<br /> 0<br /> 6.5<br /> À1.682 6.5<br /> 0<br /> 6.5<br /> À1<br /> 4<br /> 1<br /> 4<br /> <br /> 1.682<br /> À1<br /> 0<br /> 0<br /> 1<br /> 0<br /> 0<br /> 1<br /> 0<br /> À1<br /> 0<br /> 1<br /> 0<br /> À1.682<br /> 1<br /> 0<br /> 0<br /> 0<br /> À1<br /> À1<br /> <br /> 1.1<br /> 0.5<br /> 1.1<br /> 1.1<br /> 1.7<br /> 2.109<br /> 0.091<br /> 1.7<br /> 1.1<br /> 0.5<br /> 1.1<br /> 0.5<br /> 1.1<br /> 1.1<br /> 0.5<br /> 1.1<br /> 1.1<br /> 1.1<br /> 1.7<br /> 1.7<br /> <br /> 0<br /> À1<br /> 0<br /> 0<br /> 1<br /> 1.682<br /> À1.682<br /> 1<br /> 0<br /> À1<br /> 0<br /> À1<br /> 0<br /> 0<br /> À1<br /> 0<br /> 0<br /> 0<br /> 1<br /> 1<br /> <br /> 35.53<br /> 35.74<br /> 34.89<br /> 33.57<br /> 23.13<br /> 23.58<br /> 34.52<br /> 35.40<br /> 33.63<br /> 29.83<br /> 34.75<br /> 35.39<br /> 20.63<br /> 28.17<br /> 32.46<br /> 33.88<br /> 35.34<br /> 34.48<br /> 30.07<br /> 18.86<br /> <br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> <br /> 0.47a<br /> 0.38a<br /> 0.40a<br /> 0.24ab<br /> 1.48d<br /> 0.63d<br /> 0.55ab<br /> 0.84a<br /> 0.66ab<br /> 0.73c<br /> 0.09a<br /> 0.62a<br /> 1.11e<br /> 1.49c<br /> 1.34b<br /> 0.68ab<br /> 1.10a<br /> 1.09ab<br /> 0.61c<br /> 1.63e<br /> <br /> a<br /> Means Æ standard deviation having the same letters are not significantly<br /> different at the 5% level.<br /> <br /> these independent variables X1, X2 and X3 on the response Y was<br /> investigated using the second-order polynomial regression equation with stepwise elimination. This equation, derived using RSM<br /> for the evaluation of the response variables, is as follows:<br /> <br /> The estimated regression coefficient of the response surface<br /> models for the extracted rambutan kernel fat along with the corresponding coefficient of determination values (R2) and lack of fit<br /> test are shown in Table 2. The linear coefficient indicates that the fat<br /> yield is positively correlated with all independent variables. On the<br /> other hand, the quadratic coefficient of all independent variables<br /> shows a negative correlation with extracted fat yield. The R2 and<br /> adjusted R2 were 0.974 and 0.967, respectively. These results imply<br /> that the response surface model can explain more than 96 percents<br /> of the variation in the response variables studied. Therefore, the R2<br /> values of the response models are sufficiently high, to indicate that<br /> the response surface model is workable and can be used for estimation of the mean response and the subsequent optimization<br /> stages. The lack of fit, which measures the fitness of the model, was<br /> found to be non significant (p > 0.05), indicating that the number of<br /> experiments were sufficient for determining the effect of independent variables on percentage fat yield (Montgomery; 2001). As<br /> shown in Table 2, the extracted fat yield was significantly (p < 0.05)<br /> influenced by the main effects of moisture content, extraction time<br /> and particle size and the quadratic effects of all independent variables, as well as the interaction effect between moisture content<br /> and particle size and extraction time and particle size. The<br /> following response surface model (Eq. (2)) was fit to the three<br /> independent variables (X1, X2 and X3) of the response variables (Y):<br /> 2<br /> 2<br /> Y ¼ 24:97 þ 1:39X1 þ 1:31X2 þ 5:56X3 À 0:08X1 À 0:12X2<br /> 2<br /> À4:76X3 À 0:53X1 X3 þ 0:77X2 X3<br /> <br /> (2)<br /> <br /> 3.3. Effect of independent variables on percent extracted fat<br /> <br /> 2<br /> 2<br /> 2<br /> Y ¼ b0 þ b1 X1 þ b2 X2 þ b3 X3 þ b11 X1 þ b22 X2 þ b33 X3<br /> <br /> þ b12 X1 X2 þ b13 X1 X3 þ b23 X2 X3<br /> <br /> 3.2. Response surface analysis<br /> <br /> (1)<br /> <br /> where Y is the response variable (fat yield, g/100 g); bo, b1, b2, b3, b11,<br /> b22.. are regression coefficients and X1, X2 and X3 are uncoded<br /> values for moisture content, extraction time and particle size,<br /> respectively. An analysis of variance (ANOVA) was performed to<br /> determine the lack of fit and the effect of linear, quadratic and<br /> interaction terms on fat extraction. The analysis of data and the<br /> optimizing process were generated using Minitab statistical software v.15 (Minitab Inc., USA).<br /> 3. Results and discussion<br /> 3.1. Percent extracted fat<br /> Fat from rambutan kernels (N. lappaceum L.) was extracted with<br /> hexane using a Soxhelt apparatus. The effect of different levels of<br /> the independent variables namely moisture content, extraction<br /> time and particle size on percent fat yield was determined and the<br /> results are summarized in Table 1. The yields of extracted fat ranged<br /> from 18.86 to 35.74 g/100 g with a mean value of 31.19 g/100 g.<br /> Augustin and Chua (1988) analyzed the seed composition of<br /> rambutan grown in Malaysia and reported that the seed contained<br /> 37.1e38.9 g/100 g crude fat (petroleum ether extracted), which is<br /> higher than the value (33.4 g/100 g) reported by other researchers<br /> in Mexico (Solís-Fuentes et al., 2010). In Thailand, it has been shown<br /> that rambutan seed contains a wide range of fat, between 14 and<br /> 41 g/100 g (Kalayasiri et al., 1996; Winayanuwattikun et al., 2008).<br /> Those differences in the fat content of the seed may be attributed to<br /> the variability of the studied cultivars, a diversity in the maturity of<br /> the seeds used and the agricultural conditions in the area cultivated<br /> (Augustin & Chua, 1988; Cheikh-Rouhou et al., 2007).<br /> <br /> The response surface plot of fat yield using hexane for various<br /> combinations of moisture content, extraction time and particle size<br /> is shown in Fig. 1. The percent fat yield was higher at lower moisture content and vice versa (Fig. 1a). Similarly, the fat yield was<br /> higher for smaller particle size and vice versa (Fig. 1b). The fat yield<br /> decreased when the moisture content and particle size increased.<br /> Higher moisture content samples have more resistance to penetration by the solvent into the samples, which resulted in low fat<br /> yield. As the particle size decreased, the surface area increased and<br /> this enhanced fat extraction, resulting in a higher fat yield (Shahidi<br /> & Wanasundara, 2002). The fat yield increased in particles up to<br /> 1.1 mm and then gradually decreased as particle size increased. As<br /> shown in Table 2, the fat yield was significantly (p < 0.05) influenced by the interaction effect of moisture content and particle<br /> Table 2<br /> Regression coefficients and p-value of the response surface models and statistical<br /> analysis.<br /> Regression Term<br /> <br /> Extracted Fat (Y, g/100 g)<br /> Coefficient<br /> <br /> Intercept (b0)<br /> Moisture (b1)<br /> Time (b2)<br /> Particle size (b3)<br /> Moisture2 (b11)<br /> Time2 (b22)<br /> Particle size2 (b33)<br /> Moisture/Particle size (b13)<br /> Time/Particle size (b23)<br /> R2<br /> R2 (adj)<br /> Regression<br /> Lack-of-fit<br /> a<br /> <br /> Probability (p-value)a<br /> <br /> 24.9732<br /> 1.3885<br /> 1.3091<br /> 5.5557<br /> À0.0836<br /> À0.1157<br /> À4.7579<br /> À0.5284<br /> 0.7722<br /> 0.9743<br /> 0.9673<br /> e<br /> e<br /> <br /> 0.000<br /> 0.000<br /> 0.004<br /> 0.004<br /> 0.000<br /> 0.000<br /> 0.000<br /> 0.000<br /> 0.000<br /> e<br /> e<br /> 0.000<br /> 0.122<br /> <br /> The p-value more than 0.05 is not significantly different at the 5% level.<br /> <br /> W. Sirisompong et al. / LWT - Food Science and Technology 44 (2011) 1946e1951<br /> <br /> 1949<br /> <br /> content and with medium particle size but with a long extraction<br /> time (Fig. 1). To determine the exact optimum points for all the<br /> independent variables necessary to achieve the optimized condition, a numerical optimization was utilized. The results showed<br /> that the extraction using 1.05 mm particles at 4.99 g/100 g moisture<br /> for 9.2 h of extraction time provided the maximal fat yield.<br /> Under these optimum conditions, the predicted fat yield was<br /> 37.35 g/100 g. The adequacy of the response surface equations<br /> was demonstrated by a comparison between the experimental<br /> values and predicted values based on a response regression<br /> (Mirhosseinia, Tan, Hamid, & Yusof, 2008). Under the recommended optimum conditions, the experimental value for fat yield<br /> was 37.20 Æ 0.69 g/100 g which was not significantly different<br /> (p > 0.05) from the predicted value (37.35 g/100 g). Good agreement must exist between the values calculated using the model<br /> equations and the experimental values at the points of interest, and<br /> no significant (p > 0.05) difference was reported between the<br /> actual and the predicted values (Predicted value ¼ 1.0016 Experimental value). The high correlation coefficients (z0.978) also<br /> confirmed that a close agreement between experimental data and<br /> predicted values calculated using the models had been obtained.<br /> The closer the experimental and predicted results, the better they<br /> explain the adequacy of the regression equation (Rossa, de Sá,<br /> Burin, & Bordignon-Luiz, 2011).<br /> 3.5. Physical and chemical characteristics of rambutan kernel fat<br /> 3.5.1. Color<br /> The rambutan kernel fat is consistently a white solid at room<br /> temperature (25 Æ 2  C). The L*, a* and b* values of solid fat were<br /> 86.87, À2.06 and 3.55, respectively (Table 3). When the solid fat was<br /> heated (w 60  C), the melted fat had a golden yellow color with L*,<br /> a* and b* values of 66.34, À2.31 and 6.62, respectively. The L*, a*<br /> and b* values of other vegetable oils, such as palm, soybean,<br /> sunflower, olive, and corn range from 63.4 to 69.5, 3.8 to 4.4 and 9.2<br /> to 10.4, respectively (Hsu & Yu, 2002). This shows that the melted<br /> fat has b* values lower than those of other vegetable oils. It may<br /> suggest the presence of less yellow pigments, e.g. carotenoids, in<br /> rambutan kernel fat (Cheikh-Rouhou et al., 2007).<br /> Fig. 1. Response surface plots showing the interaction effects of extraction variables on<br /> the extracted fat yield; a ¼ extraction time-moisture effect, b ¼ particle size-moisture<br /> effect, and c ¼ particle size-extraction time effect.<br /> <br /> size. From the results (Fig. 1b) it is apparent that with a smaller<br /> particle size (less than 1.1 mm) the fat yield slightly increased with<br /> an increase in moisture content up to w 10 g/100 g and the fat yield<br /> decreased when the moisture content was further increased. On<br /> the other hand, for the bigger particle sizes (more than 1.1 mm) the<br /> fat yield slightly decreased with increasing moisture content and<br /> when the moisture content was increased from 5 to 15 g/100 g<br /> a marked decrease in fat yield was observed. In Fig. 1c, the fat yield<br /> was increase with an increase in extraction time and did not<br /> increase after 6.5 h for small particle size (w 0.5 mm). At intermediated particle size (w 1.1 mm), the fat yield increased as the<br /> extraction time was increased and reached a maximum at 9 h. Any<br /> further increase in extraction time did not increase the fat yield.<br /> 3.4. Optimization and validation of regression model<br /> Response optimizations were performed to measure the<br /> optimum levels of independent variables required to achieve the<br /> desired fat yield. For hexane extraction, the process providing<br /> the maximal fat yield would involve samples with low moisture<br /> <br /> 3.5.2. Crystal polymorphism<br /> In regard to polymorphic forms of rambutan kernel fat as<br /> determined by XRD, it can be seen that the extracted fat contains<br /> <br /> Table 3<br /> Physical and chemical characteristics of rambutan kernel fat.<br /> Properties<br /> Color (CIE Lab)<br /> Solid fat<br /> Melted fat<br /> Polymorphic forms (g/100 g)<br /> Beta (b)<br /> Beta-prime (b0 )<br /> Refractive index<br /> Iodine value (g/100 g oil)<br /> Saponification value (mg KOH/g oil)<br /> Unsaponifiable matter (g/100 g)<br /> Phytosterol (mg/g)<br /> Stigmasterol<br /> b-Sitosterol<br /> Campesterol<br /> a-Tocopherol (mg/100 g)<br /> a<br /> b<br /> <br /> Valuea<br /> L* ¼ 86.87 Æ 0.97;<br /> a* ¼ À2.06 Æ 0.48; b* ¼ 3.55 Æ 0.35<br /> L* ¼ 66.34 Æ 0.18;<br /> a* ¼ À2.31 Æ 0.46; b* ¼ 6.62 Æ 0.56<br /> 84.7 Æ 1.2<br /> 15.3 Æ 1.2<br /> 1.469 Æ 0.001<br /> 41.6 Æ 1.2<br /> 166 Æ 3<br /> 0.19 Æ 0.04<br /> 0.32 Æ 0.03<br /> 0.61 Æ 0.06<br /> NDb<br /> 0.103 Æ 0.001<br /> <br /> Mean Æ standard deviation of duplicate analysis.<br /> ND ¼ Not detected.<br /> <br /> 1950<br /> <br /> W. Sirisompong et al. / LWT - Food Science and Technology 44 (2011) 1946e1951<br /> <br /> mixtures of b (84.7 g/100 g) and b0 (15.3 g/100 g) polymorphic<br /> forms (Table 3). The structure, composition and polymorphic forms<br /> of fatty crystals are most important criteria for determining the<br /> functional properties of fat (Reddy & Jeyarani, 2001). From the<br /> results, it appears that rambutan kernel fat tends to crystallize in<br /> b form which is also seen in cocoa butter. For cocoa butter,<br /> a representative confectionery fat, the most functional polymorph<br /> (form V) is a b type, since this form results in optimal density,<br /> melting behavior, and surface appearance (Sato, 1999). Therefore,<br /> rambutan kernel fat has the potential to be used for confectionery<br /> products.<br /> 3.5.3. Refractive index, iodine value and saponification number<br /> The refractive index of rambutan kernel fat is 1.469 Æ 0.001<br /> (Table 3). This value is consistent with the values of other<br /> vegetable oils (Padley, Gunstone, & Harwood, 1986) and the<br /> results of Solís-Fuentes et al. (2010), who reported that rambutan<br /> fat had a refractive index of 1.468. The iodine value of the fat,<br /> 41.6 Æ 1.2 g per 100 g fat (Table 3), places it in the non-drying<br /> group of oils which also included palm oil, palm kernel oil and<br /> coconut oil (Tan & Che Man, 2000). This value is similar to the<br /> previous studies (Azam et al., 2005; Solís-Fuentes et al., 2010),<br /> which reported that rambutan fat had an iodine value of about<br /> 44e47 g per 100 g fat. The saponification number, which is an<br /> indication of the average molecular weight of the fat, was<br /> 166 Æ 3 mg KOH per gram fat for rambutan kernels (Table 3). The<br /> low saponification value suggests that rambutan fat contains<br /> a long chain fatty acid and a relatively high average molecular<br /> weight (Onyieke & Acheru, 2002).<br /> 3.5.4. Unsaponification matter<br /> The estimated unsaponification matter in rambutan kernel fat<br /> found to be 0.19 Æ 0.04 g/100 g, as compared to 0.2e0.6 g/100 g in<br /> cocoa butter (Padley et al., 1986). In the present study, the<br /> phytosterol along with a-tocopherol content were also determined<br /> (Table 3). The results show that the rambutan kernel fat had<br /> 0.61 Æ 0.06 and 0.32 Æ 0.03 mg per gram fat of b-sitosterol and<br /> stigmasterol, respectively. The fat from rambutan kernel has lower<br /> level of total phytosterol than do commercial oils and fats (Padley<br /> et al., 1986). The extracted fat also contained 0.103 mg per 100 g<br /> fat of a-tocopherol. The low content of a-tocopherol in rambutan<br /> kernel fat was similar to that of cocoa butter, cod liver oil and beef<br /> fat (Swern, 1964). These results suggest that rambutan kernel fat<br /> may not be a good source of the natural antioxidants, phytosterol<br /> and a-tocopherol.<br /> <br /> Table 4<br /> Fatty acid composition of rambutan kernel fat.<br /> Fatty acid composition<br /> <br /> Content (g/100 g)a<br /> <br /> Saturated fatty acid<br /> Myristic acid (C14:0)<br /> Palmitic acid (C16:0)<br /> Stearic acid (C18:0)<br /> Arachidic acid (C20:0)<br /> Heneicosanoic acid (C21:0)<br /> Behenic acid (C22:0)<br /> Tricosanoic acid (C23:0)<br /> Lignoceric acid (C24:0)<br /> Monounsaturated fatty acid<br /> Palmitoleic (C16:1u7)<br /> Trans-9-Elaidic acid (C18:1u9t)<br /> Cis-9-Oleic acid (C18:1u9c)<br /> Erucic acid (C22:1u9)<br /> Polyunsaturated fatty acid<br /> Cis-9,12-Linoleic acid (C18:2u6)<br /> a-Linolenic acid (C18:3u3)<br /> Cis-11,14-Eicosadienoic acid (C20:2)<br /> <br /> 49.57<br /> 0.02<br /> 4.69<br /> 7.03<br /> 34.32<br /> 0.05<br /> 3.10<br /> 0.03<br /> 0.33<br /> 37.97<br /> 0.49<br /> 0.03<br /> 36.79<br /> 0.66<br /> 7.89<br /> 1.37<br /> 6.48<br /> 0.04<br /> <br /> a<br /> <br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> Æ<br /> <br /> 0.14<br /> 0.00<br /> 0.15<br /> 0.08<br /> 0.01<br /> 0.00<br /> 0.04<br /> 0.01<br /> 0.06<br /> 0.22<br /> 0.04<br /> 0.00<br /> 0.16<br /> 0.03<br /> 0.01<br /> 0.02<br /> 0.03<br /> 0.00<br /> <br /> Mean Æ standard deviation of duplicate analysis.<br /> <br /> 3.7. Thermal properties of rambutan kernel fat<br /> The thermal behavior, melting and crystallization of rambutan<br /> kernel fat is shown in Fig. 2a. These results show three well-defined<br /> peaks with shoulders that correspond to groups of triglycerides<br /> with high, middle, and low temperatures of melting and crystallization. These results illustrate the complex nature of triglycerides<br /> in fat samples (Tan & Che Man, 2000). For the melting behavior, the<br /> first peak had the lowest melting temperatures, with the temperature peak at 4.2  C. The results suggest that this region includes<br /> a group of triglycerides with a greater abundance of unsaturated<br /> fatty acids. The second and third peaks represent intermediate and<br /> high melting temperatures, with a peak temperature at 37.4 and<br /> 49.8  C, respectively. These regions might be represented by groups<br /> <br /> 3.6. Fatty acid composition<br /> The fatty acid composition of rambutan kernel fat is described in<br /> Table 4. The total saturated and unsaturated fatty acids found in<br /> rambutan fat were 49.6 and 45.9 g/100 g, respectively. The most<br /> abundant fatty acids in rambutan seed fat was arachidic acid<br /> (C20:0) for the saturated fatty acids (34.3 g/100 g), and oleic acid<br /> (C18:1) was the main unsaturated fatty acid (36.8 g/100 g). These<br /> two fatty acids comprised more than 70 g/100 g of the total fatty<br /> acids, which is in agreement with previously published data<br /> (Augustin & Chua, 1988; Azam et al., 2005; Solís-Fuentes et al.,<br /> 2010). Likewise, measurable amounts of palmitic (C16:0), stearic<br /> (C18:0) and behenic (C22:0), for saturated fatty acids as well as<br /> palmitoleic (C16:1), linoleic (C18:2n-6) and linolenic (C18:3n-3) for<br /> unsatutated fatty acids were detected (Table 4). The high level of<br /> arachidic acid is the main reason for the low iodine value in<br /> rambutan seed fat (Table 3) would allow the fat to be used without<br /> being subjected to hydrogenation, especially where autoxidation<br /> may be a concern.<br /> <br /> Fig. 2. Melting and crystallization curves of rambutan kernel fat (a) and commercial<br /> cocoa butter (b).<br /> <br />