- Email: [email protected]
Pressurized fluid extraction of pistachio oil using a modified supercritical fluid extractor and factorial design for optimization
ARTICLE IN PRESS
LWT 41 (2008) 1472–1477 www.elsevier.com/locate/lwt
Pressurized fluid extraction of pistachio oil using a modified supercritical fluid extractor and factorial design for optimization Ali Sheibani, Hassan S. Ghaziaskar Department of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Islamic Republic of Iran Received 20 February 2007; received in revised form 31 August 2007; accepted 6 September 2007
Abstract A pressurized fluid extraction (PFE) method for the extraction of pistachio oil was developed mainly as an analytical tool to determine oil content and/or its quality. The supercritical fluid extractor was modified to be able to pump liquid solvent and CO2 into the extraction vessel alternatively. The extraction yield was found independent of the pressure in the range 10–150 bar tested. The addition of crushed glass increased the extraction yield by more than 15 g/100 g, while the extraction reproducibility expressed by percentage RSD was improved from 4 to 1. Furthermore, the use of crushed glass reduced the solvent consumption from 35 to 20 mL. The effective variables of temperature (40–80 1C), solvent volume (5–25 mL), and crushed glass percentage (30–60 g/100 g) were optimized by a factorial design method. The model allows the prediction of the extraction yield at different conditions. The PFE yield (i.e. 52.6 g/100 g) and fatty acid composition of pistachio oil were found similar to Soxhlet extraction and their variations were within the experimental uncertainty verified by statistical t-test analysis. Two different solvents of n-hexane and ethanol were used for PFE of pistachio oil. The extraction yield was about one-third (i.e. 18 g/100 g) when ethanol was used as solvent. r 2007 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved. Keywords: Pressurized fluid extraction; Accelerated solvent extraction; Pistachio oil; Supercritical fluid extractor; Factorial design optimization; Soxhlet extraction
1. Introduction Pistachio oil is a highly unsaturated emollient providing superior moisturization and a high level of nourishingessential fatty acids in a dry non-greasy, rapidly absorbed, and very low odor form (www.desertwhale.com). Pistachio nut contains more than 50 g/100 g oil from which 51–54 g/ 100 g is oleic acid and 31–35 g/100 g is linoleic acid (Agar, Kafkas, Kaska, & Sheibani, 1997; Yildiz, Gurcan, & Ozdemir, 1998). Iran is one of the largest producers of pistachio. While a part of the harvested pistachios have no or only low marketable appearance and quality, they still have nutrition value. The oil of this kind of pistachio nuts has a good market and can be extracted and sold to increase the added value of pistachio cultivation.
Corresponding author. Tel.: +98 311 3913260; fax: +98 311 3912350.
E-mail address: [email protected] (H.S. Ghaziaskar).
Supercritical carbon dioxide (scCO2) has been previously used for extraction of oil from different nuts, such as almond (Marrone, Poletto, Reverchon, & Stasis, 1998), canola (Rezaei & Temelli, 2001), soybean (Friedrich & List, 1982), and peanut (Goodrum & Kilgo, 1987). scCO2 has also been used for the extraction of oil from hazelnut and pistachio at high pressures of 300–450 bar (Ozkal, Salgin, & Yener, 2005; Palazoglu & Balaban, 1998). In the case of pistachio oil extraction, even addition of 5–10 g/ 100 g of modifiers such as ethanol did not lead to lower operating pressure in the pressure range 207–345 bar and temperature range 50–70 1C. Maximum recovery of pistachio oil at 345 bar, 60 1C, and 10 g of ethanol/100 g sample was 66.14 g/100 g of its total oil content (Palazoglu & Balaban, 1998). Pressurized fluid extraction (PFE, Dionex trade name of accelerated solvent extraction (ASE)) is an efficient, rapid, selective, and reliable extraction method (Bjorklund, Bowadt, Nilsson, & Mathiasson, 1999). PFE technology and its applications have been developed and used by
0023-6438/$34.00 r 2007 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.lwt.2007.09.002
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
various researchers (Bautz, Polzer, & Stieglitz, 1998; Breithaupt, 2004; Draisci et al., 1998; Ezzell, 2003; Giergielewicz-Mozajska, Dabrowski, & Namienik, 2001; Hawthorne, Grabanski, Martin, & Miller, 2000; Kot-Wasik & Wasik, 2005; Lundstedt, Bavel, Haglund, Tysklind, & Oberg, 2000; Pecorelli et al., 2003; Richter & Covino, 2000). It has also been applied for the quantitative extraction of different samples of environmental organic compounds from soils (Richter & Covino, 2000; Saito et al., 2003) and lipids, and also for the analysis of food and biological samples (Schafer, 1998). The possibility of changing several extraction variables such as temperature, pressure, and volume of solvent is a promising characteristic of newly available automated PFE instruments (Li, Landriault, Fingas, & Llompart, 2003). Using a modified supercritical fluid extraction (SFE) apparatus, we have developed a new PFE method mainly as an analytical tool to determine oil content or as a sample preparation method for the oil quality measurements with full capability of automation when the workload is high (Wilkes et al., 2000). Special emphasis is placed on obtaining a rapid, selective, efficient, and reliable extraction process. A partial factorial design was used to design the experiment and a mathematical model was constructed to relate the extraction yield to the variables such as temperature, solvent volume (i.e. proportional to the dynamic extraction time), and the percentage of added crushed glass to the samples. The validity of the model was evaluated by analysis of variance (ANOVA) (Herrero, Martin-Alvarez, Senorans, Cifuentes, & Ibanez, 2005; Palazoglu & Balaban, 1998). The work is in progress in our research laboratory for using this technique for the large-scale extraction as a multipurpose extraction method. 2. Materials and methods 2.1. Materials Pistachio (Pistacia Vera L.) was obtained from Rafsanjan area of Iran. Ethanol, n-hexane, methanol, potassium hydroxide, hydrochloric acid, and 14 g boron trifluoride in 100 g methanol were supplied by Merck Co. 2.2. Apparatus The main components of the SFE apparatus had been described in detail elsewhere (Ghaziaskar, Eskandari, & Daneshfar, 2003). For this work, the apparatus was modified using a switching valve at the entrance of the pump to enable pumping of the liquid solvent and CO2 alternatively into the extraction vessel. A gas chromatograph (Chrompack CP 9001, The Netherlands) equipped with a capillary column (Sill 88, length i.d. ¼ 50 m 0.25 mm, a film thickness of 0.2 mm) and a flame ionization detector was used for the determination of fatty acid composition of extracted oil.
1473
An analytical balance (Sartorius, model LP-1200S, Germany) was used in gravimetric determination of the extraction yields on a moisture-free basis of the sample weight. 2.3. Sample preparation for PFE The ground pistachio kernels passed through a sieve with mesh size of 16 and dried at 70 1C to a constant weight. It was kept within a sealed bag in a refrigerator until it was used. The pistachio kernel samples were ground prior to PFE to facilitate analytes mass transfer during the extraction process. Since extraction kinetic is controlled by the kernel particle size, a sieving step is important to obtain reproducible extraction yield. 2.4. PFE procedure The sample (4 g) was loaded into the 10 mL stainless steel cell. Cotton wool was packed at the exit end of the cell to prevent transfer of solid samples to the tubing and clogging of the system. Two different solvents of n-hexane and ethanol were used to investigate the influence of solvent polarity on the yield of oil extraction from pistachio. The extraction program consisted of static extraction, followed by passing the solvent at a flow rate of 1 mL/min at different temperatures and pressures. In the analytical scale extractions, the ground pistachio kernels were dispersed by different percentage of crushed glass (Pyrex broken laboratory glassware were crushed and the part falling between the 40 and 60 mesh sieve was used). However, the sample may be ground with the pistachio hard shell for dispersing the matrix in the large-scale extractions and benefit the potential subsequent use of the residue as feed source. The solvent residue in the cell and tubing was removed with purging the PFE system with liquid CO2 at the end of each extraction. The extraction solvent was evaporated at a temperature of 50 1C. The extraction yield was calculated as the weight of extracted oil to the dry weight of the dried sample loaded into the extraction cell. In this work, the effect of different variables such as extraction temperature (40–80 1C), pressure (10–150 bar), solvent volume (10–25 mL), and the percentage of crushed glass (20–60 g/100 g) on the extraction yield were investigated. Preliminary experiments showed that increasing the pressure had only a little effect but the other variables had significant impacts on the extraction yields. Ethanol and n-hexane were used as solvent for PFE of pistachio oil but the extraction yield with n-hexane was found higher. Therefore, the extractions were performed with n-hexane at different combinations of temperatures, solvent volumes, and the percentage of crushed glass according to the factorial design procedure detailed below. Conventional Soxhlet extraction (SE) was used as a reference method (Ozkal et al., 2005).
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
1474
Fatty acid methyl esters (FAME) of the extracted oils were prepared and analyzed by GC-FID (Beattie, Stafford, & King, 1993). Identification of FAME was based on retention times of reference compounds. Fatty acid composition was expressed as g/100 g of total FAME. 2.5. Factorial design A three-variables partial factorial design (temperature, T; solvent volume, V; and the percent crushed glass mixed with ground pistachio nut, G) was used for the optimization of the extraction as shown in Table 2. From experiments that were carried out and listed in Table 1, a total of 20 experiments were used including the central point that was repeated five times. The three variables were tested at three levels: T ¼ 50, 60, and 80 1C; V ¼ 5, 10, and 15 min; and G ¼ 30, 50, and 60 g/100 g. The program was written using MATLAB (the Math Works Inc., Version 6.0).
Table 1 The PFE experimental yields using n-hexane as solvent at different temperatures, crushed glass, and solvent volume at pressure 10 bar, static extraction time 5 min, and flow rate 1 mL/min. Temperature (1C)
Crushed glass (g/100 g)
Solvent volume (mL)
Experimental yield (g/100 g)
40 50
50 30
10 5 10 15
(31.26) 29.26 34.47 35.00
60
5 10 15
34.60 45.83 47.87
10 10 5 10 10 10 10 10 10 15 20 25 10
(33.90) (48.10) 39.30 48.75 50.00 49.50 49.40 48.50 48.60 50.50 (49.00) (49.50) (49.40)
30
5 10 15
41.57 48.93 50.43
60
5 10 15
42.34 50.26 52.58
60
20 40 50
60 80
The bracketed data was obtained as preliminary data and not included in the final model.
3. Results and discussion 3.1. The effect of extraction temperature Increasing temperature of the extraction increased the extraction yield. An extraction temperature of 80 1C was good enough to give 100% oil recovery. The increase in yield with temperature is due to increase in solvating power at higher temperatures. Moreover, the viscosity of organic solvents decreases at higher temperatures and the solvent penetrates better into the matrix and hence an improvement in the extraction yield is observed (Dean & Saim, 1998). The extraction yield can also be increased by increasing the solvent throughput (at constant flow rate). This method was preferred instead of working at the high temperatures. 3.2. The effect of the static extraction time and the solvent volume A 5 min period of static extraction time improved the overall yield of the oil extraction. While at the higher static extraction time of 10 min no improvement on the extraction yield were observed. Therefore, the static time of 5 min was used for all the experiments. The extraction yield was highly increased when the extraction solvent was passed through the sample cell for a period of time. Since a constant flow rate of solvent is used, by increasing the extraction time the amount of solvent passing through the extractor is also increased. As given in Table 1, when the solvent volume is increased from 5 to 15 mL (T ¼ 60 1C and G ¼ 50 g/100 g), the extraction yield (g/100 g) is increased from 39.3 to 50.5. 3.3. The effect of addition of crushed glass to the sample The extraction yield as a function of crushed glass percentage mixed with the sample (g/100 g) was investigated. The method of packing of the sample in the cell and the contact surface between the sample and the solvent is very important (Dean & Saim, 1998). Therefore, the pistachio kernels were ground and well mixed with the crushed glass. Addition of the crushed glass increased the extraction yield by more than 15 g/100 g, and decreased the RSD of the extraction from 4% to 1%. Furthermore, it Table 2 The PFE yield (g/100 g) of pistachio oil using n-hexane and ethanol as solvent at pressure of 10 bar, temperature 50 1C, static time 5 min, crushed glass 30 g/100 g sample, and flow rate 1 mL/min at different solvent volumes Type of solvent
Ethanol n-Hexane
Solvent volume (mL) 5
10
15
9.6 29.3
12.0 34.5
15.0 35.0
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
considerably decreased the solvent consumption. The crushed glass mixed with the ground pistachio disperses the matrix and thus increases the contact surface of the sample and the solvent and it improves the homogeneity of the sample packed into the extraction solvent. 3.4. The effect of extraction solvent
1475
each coefficient, the t-value for the null hypothesis (H0) were calculated. If t-value for each coefficient was greater than the critical value, the variable had significant effect in the model (at confidence level of 95%), and was preserved. After performing t-test, the following equation was obtained: Y ¼ 112:2294 þ 3:2101T þ 2:6751V
Two different solvents were used for the extraction. At similar conditions, the yield of extraction by ethanol was found to be about 66 g/100 g lower than the yield of extraction with n-hexane. Triglycerides and n-hexane are fully miscible while ethanol and triglycerides are not and this leads to higher extraction yields for n-hexane in comparison with ethanol. The other advantage of n-hexane in comparison with ethanol is its lower boiling point that leads to easier evaporation of n-hexane at the end of each extraction. The separation of solvent residue, at the end of extraction could be performed by liquid CO2. Both nhexane and ethanol have a high solubility in liquid or supercritical CO2 while triglycerides have relatively limited solubility. Therefore, the solvent residue could also be extracted from the pistachio oil as reported by Eller, Taylor, and Curren (2004) for the extraction of n-hexane residue from soybean oil. Depending on the solvent used, the color of oil was different. With using ethanol as extraction solvent, the color of oil was green, that is probably due to the coextraction of chlorophyll, while the color of oil was yellow when n-hexane was used as solvent.
þ 1:1235G þ 0:0009T V 0:0094T G þ 0:0141V G 0:019T 2 0:1225V 2 0:0052G 2 .
ð2Þ
As shown in Eq. (2), all variables proposed in Eq. (1) have significant effects in the model. The t-values which show the effect of the variables on the PFE yield are calculated for the solvent volume as (t) ¼ 5.22, for temperature as (t) ¼ 4.57, and for crushed glass g/100 g as (t) ¼ 2.27, at the 95% confidence level, and are regarded as important variables for optimization of the experiment. Nevertheless, the second-order interactions are less important. Response surfaces showing the effects of different variables on extraction yields are given in Figs. 1–3. ANOVA was used to evaluate the validity of the model. There was no lack of fit at the confidence level of 95% (calculated F lack of fit is less than the critical value, 3.39o4.95). The regression equation for the extraction yield calculated from the model (Ycal) versus the experimental extraction yield (Yexp) is Ycal=0.90+0.98Yexp and the correlation coefficient of the data R2=0.9879, which
3.5. The effect of pressure The PFE was performed at the pressure range of 10–150 bar, but a small change was observed on the extraction yield. Therefore, pressure of 10 bar was applied only to keep the solvent as liquid at the high temperatures and drive the solvent through the sample in order to improve the extraction yield. 3.6. Modeling A second-order polynomial equation is proposed for prediction of the PFE yields of pistachio oil by n-hexane as a function of independent variables as follows: Y ¼ a0 þ a 1 T þ a 2 V þ a 3 G þ a 4 T V þ a5 T G þ a6 V G þ a7 T 2 þ a8 V 2 þ a9 G 2 ,
ð1Þ
where a0 is the intercept; a1, a2,y, a9 are the coefficients of the polynomial; Y is the extraction yield (g/100 g); T the temperature; V the solvent volume; and G denotes the grams of crushed glass mixed with 100 g of the sample. The coefficients of these variables were obtained by multiple least-squares regression using the MATLAB program. For
Fig. 1. Response surface of the extraction yield (g/100 g) versus temperature and solvent volume using 50 g crushed glass/100 g sample.
ARTICLE IN PRESS 1476
A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477 Table 3 Comparison of PFE with SE and SFE methods for the extraction of pistachio oil Parameters
SEa
SFEb
PFE
Type of solvent Solvent/sample weight (g/g) Temperature (1C) Process time (h) Extraction yield (g/100 g)
n-Hexane 15 68 2.5 52.6
scCO2 30 60 2.5 34
n-Hexane 7.5 60 0.5 52.6
a
This work. Palazoglu and Balaban (1998) (at 345 bar and 60 1C).
b
Table 4 The main fatty acid composition of pistachio oil (g/100 g7S.D.a) extracted by PFE and SE methods
Fig. 2. Response surface of the extraction yield (g/100 g) versus temperature and crushed glass (g/100 g) with the solvent volume 15 mL.
Fatty acid
PFE
SE
C16:0 C18:0 C18:1 C18:2 C18:3
9.4670.27 1.3670.07 54.7371.21 29.0670.8 0.3770.01
9.1770.21 1.2370.06 55.1171.6 29.4570.88 0.3770.02
a Each value is an average and standard deviation of three determinations.
Table 3. The PFE yield is similar to SE, but takes only 20% of the process time and consumes only 50% of the solvent/ sample weight (g/g). The PFE yield is 18.6 g/100 g higher compared with SFE while the process time and the solvent/ sample weight (g/g) are only 20% and 25% of the ones for SFE, respectively (Palazoglu & Balaban, 1998). Moreover, matrix effects in PFE are much less prominent compared with SFE (Wilkes et al., 2000). As shown in Table 4, the oil main fatty acid composition expressed as g/100 g extracted by the PFE and SE methods are similar. The t-test did not show any significant difference between the data obtained from the methods. The calculated t-values for the data listed in Table 4 are smaller than the critical t-values at the 95% confidence level and four degrees of freedom. 4. Conclusion Fig. 3. Response surface of the extraction yield (g/100 g) versus solvent volume and crushed glass (g/100 g) at the temperature 60 1C.
indicates good performance of the model. The maximum extraction yield under different conditions was predicted using the program written by MATLAB. 3.7. Comparison of PFE with SE and SFE methods Methods of SE, SFE, and PFE have been compared in terms of solvent type, the solvent/sample weight (g/g), temperature, process time, and the extraction yield in
The oil from low market value pistachio nuts was extracted by PFE method. The overall yield was 52.6 g/ 100 g that is equal to 100% oil recovery with lower solvent consumption and the extraction time in comparison with the SE method. The fatty acid composition of pistachio oil extracted by PFE method was found similar to that of the oil extracted with SE method. The proposed model can predict the maximum extraction yield at different conditions. Mixing of crushed glass with the sample increased the extraction yield by more than 15 g/ 100 g, and decreased percentage RSD of the extraction from 4 to 1 and lowered the extraction solvent consumption from
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
about 35 to 20 mL. Extraction yield was about one-third when ethanol was used as the extraction solvent. Acknowledgments This work has been supported by the Research Council and the Center of Excellence in Green Chemistry Research of IUT. We are thankful to Dr. T. Khayamian for helping in factorial design and modeling. References Agar, I. T., Kafkas, S., Kaska, N., Sheibani, A. (1997). Lipid characteristics of Turkish and Iranian pistachio kernels. In Proceedings of the second international symposium on pistachios and almonds. Acta horticulture, Vol. 470, pp. 378–384. Bautz, H., Polzer, J., & Stieglitz, L. (1998). Comparison of pressurized liquid extraction with soxhle extraction for the analysis of polychlorinated dibenzo-p-dioxins and dibenzofurrans from fly ash and environmental matrices. Journal of Chromatography A, 815, 231–241. Beattie, S. E., Stafford, A. E., & King, A. D. (1993). Triacylglycerol profiles of Tilletia controversa and Tilletia tritici. Applied and Environmental Microbiology, 59, 1054–1057. Bjorklund, E., Bowadt, S., Nilsson, T., & Mathiasson, L. (1999). Pressurized fluid extraction of polychlorinated biphenyls in solid environmental samples. Journal of Chromatography A, 836, 285–293. Breithaupt, D. E. (2004). Simultaneous HPLC determination of carotenoids used as food coloring additives: Applicability of accelerated solvent extraction. Food Chemistry, 86, 449–456. Dean, J. R., & Saim, N. (1998). Modern alternatives to supercritical fluid extraction. In E. D. Ramsey (Ed.), Analytical supercritical fluid extraction techniques (pp. 392–417). London: Kluwer Academic Publishers. Draisci, R., Marchiafava, C., Ferretti, E., Palleschi, L., Catellani, G., & Anastasio, A. (1998). Evaluation of musk contamination of freshwater fish in Italy by accelerated solvent extraction and gas chromatography with mass spectrometric detection. Journal of Chromatography A, 814, 187–197. Eller, F. J., Taylor, S. L., & Curren, M. S. S. (2004). Use of liquid carbon dioxide to remove hexane from soybean oil. Journal of the American Oil Chemists’ Society, 81, 989–992. Ezzell, J. L. (2003). Accelerated solvent extraction of organometallic and inorganic compounds. Comprehensive Analytical Chemistry, 41, 343–352. Friedrich, J. P., & List, G. R. (1982). Characterization of soybean oil extracted by supercritical carbon dioxide and hexane. Journal of Agricultural and Food Chemistry, 30, 192–193. Ghaziaskar, H. S., Eskandari, H., & Daneshfar, A. (2003). Solubility of 2ethyl-1-hexanol, 2-ethyl hexanoic acid, and their mixtures in supercritical carbon dioxide. Journal of Chemical and Engineering Data, 48, 236–240. Giergielewicz-Mozajska, H., Dabrowski, L., & Namienik, J. (2001). Critical Reviews in Analytical Chemistry, 31, 149–165.
1477
Goodrum, J. W., & Kilgo, M. B. (1987). Peanut oil extraction with SCCO2: Solubility and kinetic functions. Journal of the American Society of Agricultural Engineering, 30, 1865–1868. Hawthorne, S. B., Grabanski, C. B., Martin, E., & Miller, D. J. (2000). Comparison of Soxhlet extraction, pressurized liquid extraction, supercritical fluid extraction and subcritical water extraction for environmental solids: Recovery, selectivity and effects on sample matrix. Journal of Chromatography A, 892, 421–433. Herrero, M., Martin-Alvarez, P. J., Senorans, F. J., Cifuentes, A., & Ibanez, E. (2005). Optimization of accelerated solvent extraction of antioxidant from Spirulina platensis microalga. Food Chemistry, 93, 417–423. Kot-Wasik, A., & Wasik, A. (2005). Determination of robenidine in animal feeds by liquid chromatography coupled with diode-array detection and mass spectrometry after accelerated solvent extraction. Analytica Chimica Acta, 543, 46–51. Li, K., Landriault, M., Fingas, M., & Llompart, M. (2003). Accelerated solvent extraction of environmental organic compounds in soils using a modified supercritical fluid extractor. Journal of Hazardous Material, 102, 93–104. Lundstedt, S., Bavel, B. V., Haglund, P., Tysklind, M., & Oberg, L. (2000). Pressurized liquid extraction of polycyclic aromatic hydrocarbons from contaminated soils. Journal of Chromatography A, 883, 151–162. Marrone, C., Poletto, M., Reverchon, E., & Stasis, A. (1998). Almond oil extraction by supercritical CO2: Experimental and modeling. Chemical Engineering Science, 53, 3711–3718. Ozkal, S. G., Salgin, U., & Yener, M. E. (2005). Supercritical carbon dioxide extraction of hazelnut oil. Journal of Food Engineering, 69, 217–223. Palazoglu, T. K., & Balaban, M. O. (1998). Supercritical CO2 extraction of lipids from roasted pistachio nuts. Journal of American Society of Agricultural Engineering, 41, 679–684. Pecorelli, I., Galarini, R., Bibi, R., Floridi, Al., Casciarri, E., & Floridi, A. (2003). Simultaneous determination of 13 quinolones from feeds using accelerated solvent extraction and liquid chromatography. Analytica Chimica Acta, 483, 81–89. Rezaei, K., & Temelli, F. (2001). On-line extraction-reaction of canola oil using immobilized lipase in supercritical CO2. Journal of Supercritical Fluids, 19, 263–274. Richter, B. E., & Covino, L. (2000). New environmental applications of accelerated solvent extraction. LC.GC, 18, 1068–1073. Saito, K., Takekuma, M., Ogawa, M., Kobayashi, S., Sugawara, Y., Ishizuka, M., et al. (2003). Extraction and clean up methods of dioxins in house dust from two cities in Japan using accelerated solvent extraction and a disposable multi-layer silica-gel cartridge. Chemosphere, 53, 137–142. Schafer, K. (1998). Accelerated solvent extraction of lipids for determining the fatty acid composition of biological material. Analytica Chimica Acta, 358, 69–77. Wilkes, J. G., Conte, E. D., Kim, Y., Holcomb, M., Sutherland, J. B., & Miller, D. W. (2000). Sample preparation for the analysis of flavors and off-flavors in foods. Journal of Chromatography A, 880, 3–33. www.DesertWhale.com. Yildiz, M., Gurcan, T. T., & Ozdemir, M. (1998). Oil composition of pistachio nuts (Pistacia vera L.) from Turkey. Journal of Lipid Science and Technology, 100, 84–86.
LWT 41 (2008) 1472–1477 www.elsevier.com/locate/lwt
Pressurized fluid extraction of pistachio oil using a modified supercritical fluid extractor and factorial design for optimization Ali Sheibani, Hassan S. Ghaziaskar Department of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Islamic Republic of Iran Received 20 February 2007; received in revised form 31 August 2007; accepted 6 September 2007
Abstract A pressurized fluid extraction (PFE) method for the extraction of pistachio oil was developed mainly as an analytical tool to determine oil content and/or its quality. The supercritical fluid extractor was modified to be able to pump liquid solvent and CO2 into the extraction vessel alternatively. The extraction yield was found independent of the pressure in the range 10–150 bar tested. The addition of crushed glass increased the extraction yield by more than 15 g/100 g, while the extraction reproducibility expressed by percentage RSD was improved from 4 to 1. Furthermore, the use of crushed glass reduced the solvent consumption from 35 to 20 mL. The effective variables of temperature (40–80 1C), solvent volume (5–25 mL), and crushed glass percentage (30–60 g/100 g) were optimized by a factorial design method. The model allows the prediction of the extraction yield at different conditions. The PFE yield (i.e. 52.6 g/100 g) and fatty acid composition of pistachio oil were found similar to Soxhlet extraction and their variations were within the experimental uncertainty verified by statistical t-test analysis. Two different solvents of n-hexane and ethanol were used for PFE of pistachio oil. The extraction yield was about one-third (i.e. 18 g/100 g) when ethanol was used as solvent. r 2007 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved. Keywords: Pressurized fluid extraction; Accelerated solvent extraction; Pistachio oil; Supercritical fluid extractor; Factorial design optimization; Soxhlet extraction
1. Introduction Pistachio oil is a highly unsaturated emollient providing superior moisturization and a high level of nourishingessential fatty acids in a dry non-greasy, rapidly absorbed, and very low odor form (www.desertwhale.com). Pistachio nut contains more than 50 g/100 g oil from which 51–54 g/ 100 g is oleic acid and 31–35 g/100 g is linoleic acid (Agar, Kafkas, Kaska, & Sheibani, 1997; Yildiz, Gurcan, & Ozdemir, 1998). Iran is one of the largest producers of pistachio. While a part of the harvested pistachios have no or only low marketable appearance and quality, they still have nutrition value. The oil of this kind of pistachio nuts has a good market and can be extracted and sold to increase the added value of pistachio cultivation.
Corresponding author. Tel.: +98 311 3913260; fax: +98 311 3912350.
E-mail address: [email protected] (H.S. Ghaziaskar).
Supercritical carbon dioxide (scCO2) has been previously used for extraction of oil from different nuts, such as almond (Marrone, Poletto, Reverchon, & Stasis, 1998), canola (Rezaei & Temelli, 2001), soybean (Friedrich & List, 1982), and peanut (Goodrum & Kilgo, 1987). scCO2 has also been used for the extraction of oil from hazelnut and pistachio at high pressures of 300–450 bar (Ozkal, Salgin, & Yener, 2005; Palazoglu & Balaban, 1998). In the case of pistachio oil extraction, even addition of 5–10 g/ 100 g of modifiers such as ethanol did not lead to lower operating pressure in the pressure range 207–345 bar and temperature range 50–70 1C. Maximum recovery of pistachio oil at 345 bar, 60 1C, and 10 g of ethanol/100 g sample was 66.14 g/100 g of its total oil content (Palazoglu & Balaban, 1998). Pressurized fluid extraction (PFE, Dionex trade name of accelerated solvent extraction (ASE)) is an efficient, rapid, selective, and reliable extraction method (Bjorklund, Bowadt, Nilsson, & Mathiasson, 1999). PFE technology and its applications have been developed and used by
0023-6438/$34.00 r 2007 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.lwt.2007.09.002
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
various researchers (Bautz, Polzer, & Stieglitz, 1998; Breithaupt, 2004; Draisci et al., 1998; Ezzell, 2003; Giergielewicz-Mozajska, Dabrowski, & Namienik, 2001; Hawthorne, Grabanski, Martin, & Miller, 2000; Kot-Wasik & Wasik, 2005; Lundstedt, Bavel, Haglund, Tysklind, & Oberg, 2000; Pecorelli et al., 2003; Richter & Covino, 2000). It has also been applied for the quantitative extraction of different samples of environmental organic compounds from soils (Richter & Covino, 2000; Saito et al., 2003) and lipids, and also for the analysis of food and biological samples (Schafer, 1998). The possibility of changing several extraction variables such as temperature, pressure, and volume of solvent is a promising characteristic of newly available automated PFE instruments (Li, Landriault, Fingas, & Llompart, 2003). Using a modified supercritical fluid extraction (SFE) apparatus, we have developed a new PFE method mainly as an analytical tool to determine oil content or as a sample preparation method for the oil quality measurements with full capability of automation when the workload is high (Wilkes et al., 2000). Special emphasis is placed on obtaining a rapid, selective, efficient, and reliable extraction process. A partial factorial design was used to design the experiment and a mathematical model was constructed to relate the extraction yield to the variables such as temperature, solvent volume (i.e. proportional to the dynamic extraction time), and the percentage of added crushed glass to the samples. The validity of the model was evaluated by analysis of variance (ANOVA) (Herrero, Martin-Alvarez, Senorans, Cifuentes, & Ibanez, 2005; Palazoglu & Balaban, 1998). The work is in progress in our research laboratory for using this technique for the large-scale extraction as a multipurpose extraction method. 2. Materials and methods 2.1. Materials Pistachio (Pistacia Vera L.) was obtained from Rafsanjan area of Iran. Ethanol, n-hexane, methanol, potassium hydroxide, hydrochloric acid, and 14 g boron trifluoride in 100 g methanol were supplied by Merck Co. 2.2. Apparatus The main components of the SFE apparatus had been described in detail elsewhere (Ghaziaskar, Eskandari, & Daneshfar, 2003). For this work, the apparatus was modified using a switching valve at the entrance of the pump to enable pumping of the liquid solvent and CO2 alternatively into the extraction vessel. A gas chromatograph (Chrompack CP 9001, The Netherlands) equipped with a capillary column (Sill 88, length i.d. ¼ 50 m 0.25 mm, a film thickness of 0.2 mm) and a flame ionization detector was used for the determination of fatty acid composition of extracted oil.
1473
An analytical balance (Sartorius, model LP-1200S, Germany) was used in gravimetric determination of the extraction yields on a moisture-free basis of the sample weight. 2.3. Sample preparation for PFE The ground pistachio kernels passed through a sieve with mesh size of 16 and dried at 70 1C to a constant weight. It was kept within a sealed bag in a refrigerator until it was used. The pistachio kernel samples were ground prior to PFE to facilitate analytes mass transfer during the extraction process. Since extraction kinetic is controlled by the kernel particle size, a sieving step is important to obtain reproducible extraction yield. 2.4. PFE procedure The sample (4 g) was loaded into the 10 mL stainless steel cell. Cotton wool was packed at the exit end of the cell to prevent transfer of solid samples to the tubing and clogging of the system. Two different solvents of n-hexane and ethanol were used to investigate the influence of solvent polarity on the yield of oil extraction from pistachio. The extraction program consisted of static extraction, followed by passing the solvent at a flow rate of 1 mL/min at different temperatures and pressures. In the analytical scale extractions, the ground pistachio kernels were dispersed by different percentage of crushed glass (Pyrex broken laboratory glassware were crushed and the part falling between the 40 and 60 mesh sieve was used). However, the sample may be ground with the pistachio hard shell for dispersing the matrix in the large-scale extractions and benefit the potential subsequent use of the residue as feed source. The solvent residue in the cell and tubing was removed with purging the PFE system with liquid CO2 at the end of each extraction. The extraction solvent was evaporated at a temperature of 50 1C. The extraction yield was calculated as the weight of extracted oil to the dry weight of the dried sample loaded into the extraction cell. In this work, the effect of different variables such as extraction temperature (40–80 1C), pressure (10–150 bar), solvent volume (10–25 mL), and the percentage of crushed glass (20–60 g/100 g) on the extraction yield were investigated. Preliminary experiments showed that increasing the pressure had only a little effect but the other variables had significant impacts on the extraction yields. Ethanol and n-hexane were used as solvent for PFE of pistachio oil but the extraction yield with n-hexane was found higher. Therefore, the extractions were performed with n-hexane at different combinations of temperatures, solvent volumes, and the percentage of crushed glass according to the factorial design procedure detailed below. Conventional Soxhlet extraction (SE) was used as a reference method (Ozkal et al., 2005).
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
1474
Fatty acid methyl esters (FAME) of the extracted oils were prepared and analyzed by GC-FID (Beattie, Stafford, & King, 1993). Identification of FAME was based on retention times of reference compounds. Fatty acid composition was expressed as g/100 g of total FAME. 2.5. Factorial design A three-variables partial factorial design (temperature, T; solvent volume, V; and the percent crushed glass mixed with ground pistachio nut, G) was used for the optimization of the extraction as shown in Table 2. From experiments that were carried out and listed in Table 1, a total of 20 experiments were used including the central point that was repeated five times. The three variables were tested at three levels: T ¼ 50, 60, and 80 1C; V ¼ 5, 10, and 15 min; and G ¼ 30, 50, and 60 g/100 g. The program was written using MATLAB (the Math Works Inc., Version 6.0).
Table 1 The PFE experimental yields using n-hexane as solvent at different temperatures, crushed glass, and solvent volume at pressure 10 bar, static extraction time 5 min, and flow rate 1 mL/min. Temperature (1C)
Crushed glass (g/100 g)
Solvent volume (mL)
Experimental yield (g/100 g)
40 50
50 30
10 5 10 15
(31.26) 29.26 34.47 35.00
60
5 10 15
34.60 45.83 47.87
10 10 5 10 10 10 10 10 10 15 20 25 10
(33.90) (48.10) 39.30 48.75 50.00 49.50 49.40 48.50 48.60 50.50 (49.00) (49.50) (49.40)
30
5 10 15
41.57 48.93 50.43
60
5 10 15
42.34 50.26 52.58
60
20 40 50
60 80
The bracketed data was obtained as preliminary data and not included in the final model.
3. Results and discussion 3.1. The effect of extraction temperature Increasing temperature of the extraction increased the extraction yield. An extraction temperature of 80 1C was good enough to give 100% oil recovery. The increase in yield with temperature is due to increase in solvating power at higher temperatures. Moreover, the viscosity of organic solvents decreases at higher temperatures and the solvent penetrates better into the matrix and hence an improvement in the extraction yield is observed (Dean & Saim, 1998). The extraction yield can also be increased by increasing the solvent throughput (at constant flow rate). This method was preferred instead of working at the high temperatures. 3.2. The effect of the static extraction time and the solvent volume A 5 min period of static extraction time improved the overall yield of the oil extraction. While at the higher static extraction time of 10 min no improvement on the extraction yield were observed. Therefore, the static time of 5 min was used for all the experiments. The extraction yield was highly increased when the extraction solvent was passed through the sample cell for a period of time. Since a constant flow rate of solvent is used, by increasing the extraction time the amount of solvent passing through the extractor is also increased. As given in Table 1, when the solvent volume is increased from 5 to 15 mL (T ¼ 60 1C and G ¼ 50 g/100 g), the extraction yield (g/100 g) is increased from 39.3 to 50.5. 3.3. The effect of addition of crushed glass to the sample The extraction yield as a function of crushed glass percentage mixed with the sample (g/100 g) was investigated. The method of packing of the sample in the cell and the contact surface between the sample and the solvent is very important (Dean & Saim, 1998). Therefore, the pistachio kernels were ground and well mixed with the crushed glass. Addition of the crushed glass increased the extraction yield by more than 15 g/100 g, and decreased the RSD of the extraction from 4% to 1%. Furthermore, it Table 2 The PFE yield (g/100 g) of pistachio oil using n-hexane and ethanol as solvent at pressure of 10 bar, temperature 50 1C, static time 5 min, crushed glass 30 g/100 g sample, and flow rate 1 mL/min at different solvent volumes Type of solvent
Ethanol n-Hexane
Solvent volume (mL) 5
10
15
9.6 29.3
12.0 34.5
15.0 35.0
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
considerably decreased the solvent consumption. The crushed glass mixed with the ground pistachio disperses the matrix and thus increases the contact surface of the sample and the solvent and it improves the homogeneity of the sample packed into the extraction solvent. 3.4. The effect of extraction solvent
1475
each coefficient, the t-value for the null hypothesis (H0) were calculated. If t-value for each coefficient was greater than the critical value, the variable had significant effect in the model (at confidence level of 95%), and was preserved. After performing t-test, the following equation was obtained: Y ¼ 112:2294 þ 3:2101T þ 2:6751V
Two different solvents were used for the extraction. At similar conditions, the yield of extraction by ethanol was found to be about 66 g/100 g lower than the yield of extraction with n-hexane. Triglycerides and n-hexane are fully miscible while ethanol and triglycerides are not and this leads to higher extraction yields for n-hexane in comparison with ethanol. The other advantage of n-hexane in comparison with ethanol is its lower boiling point that leads to easier evaporation of n-hexane at the end of each extraction. The separation of solvent residue, at the end of extraction could be performed by liquid CO2. Both nhexane and ethanol have a high solubility in liquid or supercritical CO2 while triglycerides have relatively limited solubility. Therefore, the solvent residue could also be extracted from the pistachio oil as reported by Eller, Taylor, and Curren (2004) for the extraction of n-hexane residue from soybean oil. Depending on the solvent used, the color of oil was different. With using ethanol as extraction solvent, the color of oil was green, that is probably due to the coextraction of chlorophyll, while the color of oil was yellow when n-hexane was used as solvent.
þ 1:1235G þ 0:0009T V 0:0094T G þ 0:0141V G 0:019T 2 0:1225V 2 0:0052G 2 .
ð2Þ
As shown in Eq. (2), all variables proposed in Eq. (1) have significant effects in the model. The t-values which show the effect of the variables on the PFE yield are calculated for the solvent volume as (t) ¼ 5.22, for temperature as (t) ¼ 4.57, and for crushed glass g/100 g as (t) ¼ 2.27, at the 95% confidence level, and are regarded as important variables for optimization of the experiment. Nevertheless, the second-order interactions are less important. Response surfaces showing the effects of different variables on extraction yields are given in Figs. 1–3. ANOVA was used to evaluate the validity of the model. There was no lack of fit at the confidence level of 95% (calculated F lack of fit is less than the critical value, 3.39o4.95). The regression equation for the extraction yield calculated from the model (Ycal) versus the experimental extraction yield (Yexp) is Ycal=0.90+0.98Yexp and the correlation coefficient of the data R2=0.9879, which
3.5. The effect of pressure The PFE was performed at the pressure range of 10–150 bar, but a small change was observed on the extraction yield. Therefore, pressure of 10 bar was applied only to keep the solvent as liquid at the high temperatures and drive the solvent through the sample in order to improve the extraction yield. 3.6. Modeling A second-order polynomial equation is proposed for prediction of the PFE yields of pistachio oil by n-hexane as a function of independent variables as follows: Y ¼ a0 þ a 1 T þ a 2 V þ a 3 G þ a 4 T V þ a5 T G þ a6 V G þ a7 T 2 þ a8 V 2 þ a9 G 2 ,
ð1Þ
where a0 is the intercept; a1, a2,y, a9 are the coefficients of the polynomial; Y is the extraction yield (g/100 g); T the temperature; V the solvent volume; and G denotes the grams of crushed glass mixed with 100 g of the sample. The coefficients of these variables were obtained by multiple least-squares regression using the MATLAB program. For
Fig. 1. Response surface of the extraction yield (g/100 g) versus temperature and solvent volume using 50 g crushed glass/100 g sample.
ARTICLE IN PRESS 1476
A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477 Table 3 Comparison of PFE with SE and SFE methods for the extraction of pistachio oil Parameters
SEa
SFEb
PFE
Type of solvent Solvent/sample weight (g/g) Temperature (1C) Process time (h) Extraction yield (g/100 g)
n-Hexane 15 68 2.5 52.6
scCO2 30 60 2.5 34
n-Hexane 7.5 60 0.5 52.6
a
This work. Palazoglu and Balaban (1998) (at 345 bar and 60 1C).
b
Table 4 The main fatty acid composition of pistachio oil (g/100 g7S.D.a) extracted by PFE and SE methods
Fig. 2. Response surface of the extraction yield (g/100 g) versus temperature and crushed glass (g/100 g) with the solvent volume 15 mL.
Fatty acid
PFE
SE
C16:0 C18:0 C18:1 C18:2 C18:3
9.4670.27 1.3670.07 54.7371.21 29.0670.8 0.3770.01
9.1770.21 1.2370.06 55.1171.6 29.4570.88 0.3770.02
a Each value is an average and standard deviation of three determinations.
Table 3. The PFE yield is similar to SE, but takes only 20% of the process time and consumes only 50% of the solvent/ sample weight (g/g). The PFE yield is 18.6 g/100 g higher compared with SFE while the process time and the solvent/ sample weight (g/g) are only 20% and 25% of the ones for SFE, respectively (Palazoglu & Balaban, 1998). Moreover, matrix effects in PFE are much less prominent compared with SFE (Wilkes et al., 2000). As shown in Table 4, the oil main fatty acid composition expressed as g/100 g extracted by the PFE and SE methods are similar. The t-test did not show any significant difference between the data obtained from the methods. The calculated t-values for the data listed in Table 4 are smaller than the critical t-values at the 95% confidence level and four degrees of freedom. 4. Conclusion Fig. 3. Response surface of the extraction yield (g/100 g) versus solvent volume and crushed glass (g/100 g) at the temperature 60 1C.
indicates good performance of the model. The maximum extraction yield under different conditions was predicted using the program written by MATLAB. 3.7. Comparison of PFE with SE and SFE methods Methods of SE, SFE, and PFE have been compared in terms of solvent type, the solvent/sample weight (g/g), temperature, process time, and the extraction yield in
The oil from low market value pistachio nuts was extracted by PFE method. The overall yield was 52.6 g/ 100 g that is equal to 100% oil recovery with lower solvent consumption and the extraction time in comparison with the SE method. The fatty acid composition of pistachio oil extracted by PFE method was found similar to that of the oil extracted with SE method. The proposed model can predict the maximum extraction yield at different conditions. Mixing of crushed glass with the sample increased the extraction yield by more than 15 g/ 100 g, and decreased percentage RSD of the extraction from 4 to 1 and lowered the extraction solvent consumption from
ARTICLE IN PRESS A. Sheibani, H.S. Ghaziaskar / LWT - Food Science and Technology 41 (2008) 1472–1477
about 35 to 20 mL. Extraction yield was about one-third when ethanol was used as the extraction solvent. Acknowledgments This work has been supported by the Research Council and the Center of Excellence in Green Chemistry Research of IUT. We are thankful to Dr. T. Khayamian for helping in factorial design and modeling. References Agar, I. T., Kafkas, S., Kaska, N., Sheibani, A. (1997). Lipid characteristics of Turkish and Iranian pistachio kernels. In Proceedings of the second international symposium on pistachios and almonds. Acta horticulture, Vol. 470, pp. 378–384. Bautz, H., Polzer, J., & Stieglitz, L. (1998). Comparison of pressurized liquid extraction with soxhle extraction for the analysis of polychlorinated dibenzo-p-dioxins and dibenzofurrans from fly ash and environmental matrices. Journal of Chromatography A, 815, 231–241. Beattie, S. E., Stafford, A. E., & King, A. D. (1993). Triacylglycerol profiles of Tilletia controversa and Tilletia tritici. Applied and Environmental Microbiology, 59, 1054–1057. Bjorklund, E., Bowadt, S., Nilsson, T., & Mathiasson, L. (1999). Pressurized fluid extraction of polychlorinated biphenyls in solid environmental samples. Journal of Chromatography A, 836, 285–293. Breithaupt, D. E. (2004). Simultaneous HPLC determination of carotenoids used as food coloring additives: Applicability of accelerated solvent extraction. Food Chemistry, 86, 449–456. Dean, J. R., & Saim, N. (1998). Modern alternatives to supercritical fluid extraction. In E. D. Ramsey (Ed.), Analytical supercritical fluid extraction techniques (pp. 392–417). London: Kluwer Academic Publishers. Draisci, R., Marchiafava, C., Ferretti, E., Palleschi, L., Catellani, G., & Anastasio, A. (1998). Evaluation of musk contamination of freshwater fish in Italy by accelerated solvent extraction and gas chromatography with mass spectrometric detection. Journal of Chromatography A, 814, 187–197. Eller, F. J., Taylor, S. L., & Curren, M. S. S. (2004). Use of liquid carbon dioxide to remove hexane from soybean oil. Journal of the American Oil Chemists’ Society, 81, 989–992. Ezzell, J. L. (2003). Accelerated solvent extraction of organometallic and inorganic compounds. Comprehensive Analytical Chemistry, 41, 343–352. Friedrich, J. P., & List, G. R. (1982). Characterization of soybean oil extracted by supercritical carbon dioxide and hexane. Journal of Agricultural and Food Chemistry, 30, 192–193. Ghaziaskar, H. S., Eskandari, H., & Daneshfar, A. (2003). Solubility of 2ethyl-1-hexanol, 2-ethyl hexanoic acid, and their mixtures in supercritical carbon dioxide. Journal of Chemical and Engineering Data, 48, 236–240. Giergielewicz-Mozajska, H., Dabrowski, L., & Namienik, J. (2001). Critical Reviews in Analytical Chemistry, 31, 149–165.
1477
Goodrum, J. W., & Kilgo, M. B. (1987). Peanut oil extraction with SCCO2: Solubility and kinetic functions. Journal of the American Society of Agricultural Engineering, 30, 1865–1868. Hawthorne, S. B., Grabanski, C. B., Martin, E., & Miller, D. J. (2000). Comparison of Soxhlet extraction, pressurized liquid extraction, supercritical fluid extraction and subcritical water extraction for environmental solids: Recovery, selectivity and effects on sample matrix. Journal of Chromatography A, 892, 421–433. Herrero, M., Martin-Alvarez, P. J., Senorans, F. J., Cifuentes, A., & Ibanez, E. (2005). Optimization of accelerated solvent extraction of antioxidant from Spirulina platensis microalga. Food Chemistry, 93, 417–423. Kot-Wasik, A., & Wasik, A. (2005). Determination of robenidine in animal feeds by liquid chromatography coupled with diode-array detection and mass spectrometry after accelerated solvent extraction. Analytica Chimica Acta, 543, 46–51. Li, K., Landriault, M., Fingas, M., & Llompart, M. (2003). Accelerated solvent extraction of environmental organic compounds in soils using a modified supercritical fluid extractor. Journal of Hazardous Material, 102, 93–104. Lundstedt, S., Bavel, B. V., Haglund, P., Tysklind, M., & Oberg, L. (2000). Pressurized liquid extraction of polycyclic aromatic hydrocarbons from contaminated soils. Journal of Chromatography A, 883, 151–162. Marrone, C., Poletto, M., Reverchon, E., & Stasis, A. (1998). Almond oil extraction by supercritical CO2: Experimental and modeling. Chemical Engineering Science, 53, 3711–3718. Ozkal, S. G., Salgin, U., & Yener, M. E. (2005). Supercritical carbon dioxide extraction of hazelnut oil. Journal of Food Engineering, 69, 217–223. Palazoglu, T. K., & Balaban, M. O. (1998). Supercritical CO2 extraction of lipids from roasted pistachio nuts. Journal of American Society of Agricultural Engineering, 41, 679–684. Pecorelli, I., Galarini, R., Bibi, R., Floridi, Al., Casciarri, E., & Floridi, A. (2003). Simultaneous determination of 13 quinolones from feeds using accelerated solvent extraction and liquid chromatography. Analytica Chimica Acta, 483, 81–89. Rezaei, K., & Temelli, F. (2001). On-line extraction-reaction of canola oil using immobilized lipase in supercritical CO2. Journal of Supercritical Fluids, 19, 263–274. Richter, B. E., & Covino, L. (2000). New environmental applications of accelerated solvent extraction. LC.GC, 18, 1068–1073. Saito, K., Takekuma, M., Ogawa, M., Kobayashi, S., Sugawara, Y., Ishizuka, M., et al. (2003). Extraction and clean up methods of dioxins in house dust from two cities in Japan using accelerated solvent extraction and a disposable multi-layer silica-gel cartridge. Chemosphere, 53, 137–142. Schafer, K. (1998). Accelerated solvent extraction of lipids for determining the fatty acid composition of biological material. Analytica Chimica Acta, 358, 69–77. Wilkes, J. G., Conte, E. D., Kim, Y., Holcomb, M., Sutherland, J. B., & Miller, D. W. (2000). Sample preparation for the analysis of flavors and off-flavors in foods. Journal of Chromatography A, 880, 3–33. www.DesertWhale.com. Yildiz, M., Gurcan, T. T., & Ozdemir, M. (1998). Oil composition of pistachio nuts (Pistacia vera L.) from Turkey. Journal of Lipid Science and Technology, 100, 84–86.