Optimization of a two-step process for biodiesel production from Jatropha curcas crude oil

Transcription

1 *Corresponding author: Optimization of a two-step process for biodiesel production from Jatropha curcas crude oil... Abderrahim Bouaid, Noureddin El Boulifi, Mercedes Martinez and Jose Aracil * Chemical Engineering Department, Faculty of Chemistry, University of Complutense, Madrid, Spain... Abstract In the present work, the process of synthesis of methyl esters from Jatropha curcas crude oil as an alternative vegetable oil, using a two-step catalyzed process is shown. In the first step, sulfuric acid was used as a catalyst for the esterification reaction of free fatty acid (FFA) and methanol in order to reduce the FFA content to 0.2%. In the second step, the product from the first step was further reacted with methanol using potassium metoxide as a catalyst. The two-step processes have been developed and optimized by application of the factorial design and response surface methodology. The optimum conditions for biodiesel production were obtained when using methanol to FFA contents of Jatropha crude oil molar ratio (MR) of 20:1, 5 wt% of sulfuric acid, at 608C with a reaction time of 60 min in the first step, followed by using an MR of methanol to product from the first step of 6:1, 0.95 wt% of KOCH 3,at458C with a reaction time of 60 min in the second step. The percentage of methyl ester in the obtained product was more than 98%. The model has been found to describe the experimental range studied adequately and allows us to scale-up the process. In addition, the fuel properties of the produced biodiesel were in the acceptable ranges according to EN14214 European biodiesel standards. Keywords: alternative fuels; FAME; Jatropha curcas; optimization; response surface methodology (RSM) Received 8 November 2011; accepted 5 December INTRODUCTION Diesel fuel plays an important role in the industrial economy of any country. These fuels run a major part of the transport sector and their demand is increasing steadily, requiring an alternative fuel which is technically feasible, economically competitive, environmentally acceptable and readily available [1]. Biodiesel, which is synthesized by transesterification of oils and fats from plant and animal sources, is a realistic alternative to diesel fuel because it provides a fuel from renewable resources and has lower emissions than petroleum diesel. Several studies have reported that the use of biodiesel has shown to be effective in reducing most regulated exhaust emissions, such as particulate matter (PM), unburned hydrocarbons (HC) and carbon monoxide (CO) [2, 3] as can be seen in Figure 1. The transesterification process combines oil with an alcohol; the alcohol employed in the transesterification is generally methanol. So the most common form of biodiesel is made with methanol and vegetable oils [4]. 1.1 Jatropha curcas as potential feedstock Currently, the most common feedstock for biodiesel production is edible oils such as soybean, rapeseed, canola, sunflower, palm, coconut and also corn oil. However, this practice has raised objections from various organizations, claiming that biodiesel is competing for resources with the food industry. In many countries, such as India or China, edible oils are not in surplus supply and therefore it is impossible to use them for biodiesel production as they are needed more for food supply [5, 6]. Among various oil-bearing seeds, J. curcas has been found to be more suitable for biodiesel production, as it has been developed scientifically to give better yield and productivity [7]. This non-edible oil is explored as a source for biodiesel production without compromising the food industry [8]. In addition, the oil percentage and the yield per hectare are important parameters in selecting the potential renewable source of fuel. Non-edible oils are not suitable for human consumption because of the presence of some toxic components in the oils. Therefore, Jatropha oil is considered a non-edible oil due International Journal of Low-Carbon Technologies 2012, 7, # The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com doi: /ijlct/ctr047 Advance Access Publication 8 February

2 A. Bouaid et al. Experiments were carried out in a stirred batch reactor of 500 cm 3 volume. This reactor was provided with temperature and speed control, and immersed in a thermostatic bath capable of maintaining the reaction temperature to within +0.18C by means of an electrical device connected to a PID controller. Figure 1. Emissions of biodiesel with respect to diesel fuel. to the presence of these toxic phorbol esters [6, 9]. Jatropha curcas is a drought-resistant tree belongs to the Euphorbiaceae family, which is cultivated in Central and South America, South-east Asia, India, Africa and many other countries. It is easy to establish, and grows almost everywhere even on gravelly, sandy and saline soils. It produces seeds for 50 years with a high oil content of 37% or more. The oil from the seeds has valuable properties such as low acidity and good stability when compared with soybean oil, low viscosity when compared with castor oil and better cold properties when compared with palm oil. Besides, Jatropha oil has a higher cetane number compared with diesel, which makes it a good alternative fuel with no modifications required in the engine [7, 8]. However, most non-edible oils contain a high level of free fatty acids (FFAs), which is undesirable as it lowers the yield of biodiesel. This is because a high FFA (.1%w/w) will promote more soap formation and the separation of products will be difficult during alkali-catalyzed transesterification. The objective of the present work was to evaluate the different variables affecting the two-step catalyzed processes of J. curcas crude oil. The optimum value for the variables affecting the process will be determined by application of the factorial design and response surface methodology (RSM). The factorial design of experiments gives more information per experiment than unplanned approaches; it allows us to see interactions among experimental variables within the range studied, leading to better knowledge of the process and therefore reducing research time and costs [10] 2 MATERIALS AND METHODS Jatropha curcas oil was supplied by Agricultura E Pecuaria (Rio de Janeiro, Brazil). All chemicals used in the experiments such as methanol of 99.8% purity and sulfuric acid (99%) were of analytical reagent grade and were supplied by Panreac (Spain). The potassium methoxide (KOCH 3, 30%) was supplied by BASF, Spain, and used as a base catalyst for transesterification reactions. 2.1 Pre-treatment: first step in biodiesel production The acid value of crude Jatropha oil was mg KOH/g; FFA content of the oil was determined according to AOCS official method. The fatty acids composition and physicochemical properties of the oil are summarized in Table 1. Jatropha curcas oil contains 9.48% of FFA, which is far beyond the acceptable limit of a 1% FFA level. Thus, a pretreatment step to reduce the FFAs of feedstock is required for a better biodiesel yield. Therefore, FFAs were first converted to esters in a pretreatment process with methanol using H 2 SO 4 as an acid catalyst. In this step, the Jatropha oil was poured into the reactor and heated. The calculated amounts of sulfuric acid and methanol were then added to the oil. The mixture was heated at 608C, working pressure was fixed at atmospheric pressure and the impeller speed was 600 rpm. Different molar ratios (MRs) of methanol to FFA contents of raw oil were used to investigate their influence on the acid value of Jatropha oil. After 1 h of reaction, the mixture was allowed to settle for 30 min and the methanol water fraction at the top layer was removed. The acid value of the product separated at the bottom was determined. The product having an acid value of,0.3 mg KOH/g was used for the transesterification reaction. 2.2 Transesterification: second step in biodiesel production The transesterification reaction was carried out in the same reactor with methanol and using KOCH 3 as the base catalyst. Table 1. Characteristics of Jatropha oil used in this study and fatty acids composition. Characteristics Jatropha Oil Acid number (mg KOH/g) Iodine number (I 2 /100 g) 115 Peroxide number (meq Per/kg) 7.01 Viscosity (408C) (mm 2 /s) Water content (mg/kg) 700 Fatty acid compositions (%) Palmitic (C16:0) Palmitic (C16:1) 0.64 Stearic (C18:0) Oleic (C18:1) Linoleic (C18:2) Arachidic (C21:0) 0.10 Behenic (C22:0) 0.16 Other minor components Rest to International Journal of Low-Carbon Technologies 2012, 7,

3 Optimization of a two-step process for biodiesel production Experiments were performed according to the following procedure: the product from the first step was added to the reactor and fitted with a reflux condenser. When the set temperature was reached, the KOCH 3 catalyst diluted in methanol was introduced in the reactor. Samples were taken at regular intervals and analyzed by gas chromatography. The impeller speeds between 500 and 1200 rpm were tested; a stirring speed of 600 rpm was found to be appropriate to avoid external mass transfer limitation [4, 11]. During the experiments, the pressure and impeller speed were maintained constant. The total reaction time was 60 min and the products were allowed to settle for 120 min before removing the glycerol layer from the bottom in a separating funnel to get the ester layer on the top. 2.3 Analytical methods Reaction products in the first and second steps were monitored by capillary column gas chromatography, using a Hewlett-Packard 5890 series II equipped with a flame ionization detector (FID). The injection system was split splitless. The carrier gas was helium at a flow rate of 1 ml/min; analysis operating conditions have been described in detail in a previous work [4]. The internal standard technique has been used in order to quantify the amount of the chemical species. The fuel properties were analyzed according to the following procedures acid value, AV (AOCS Ca 5a-40), peroxide value, PV (AOCS Cd 8 53), iodine value, IV (AOCS Cd 1 25), moisture content by the Karl Ficher method and viscosity n (ISO 3104). The flash point was measured by PMA4, Protest Analyzer and the cetane number by IROX diesel instruments. The oxidation stability of methyl esters was analyzed according to the Rancimat method using Metrohm 743 Rancimat (Herisau, Switzerland). The cloud point (CP) and pour point (PP) of methyl esters were measured by an automatic analyzer: CP and PP measurements CPP 97 2, according to ASTM D97 and ASTM D2500 methods. 2.4 Experimental design and statistical analysis The two steps, reduction in FFAs of Jatropha oil using H 2 SO 4 as a catalyst and the synthesis of methyl esters by transesterification of Jatropha oil using KOCH 3 as a catalyst, were studied and optimized using the factorial design of experiments. The experimental design applied to the two steps of this study was a full two-level factorial design 2 2 (two factors each, at two levels). Application of this method requires the adequate selection of response, factors and levels Pretreatment process The response selected, Y, was the AV of the Jatropha oil. The selection of factors was made considering the chemistry of the system and the practical use of factorial design and to optimize the process from an economic point of view. The factors chosen were initial catalyst concentration, X C, and MR of methanol to FFA contents of Jatropha crude oil, X MR. Table 2. Factorial design matrix of the two variables in the pretreatment step and AV of the final product. Experiment Coded design levels Real values X RM X C RM C (%) : þ : þ1 20: þ1 þ1 28: : : : : AV (mg KOH/g) Selection of the levels was carried out on the basis of results obtained in a preliminary study, considering the experimental installation limits, and the working conditions limit for each chemical species. The amount of catalyst was progressively increased; AV was monitored versus time. The levels chosen were 1 and 5 wt% based on the oil weight. MR levels were selected according to reactant properties and on the basis of preliminary studies; so, the lower value was 20:1 and the higher was chosen as 28:1. Once these values were selected, the statistical analysis was performed. The experimental matrix for the factorial design is shown in Table 2. The first two columns of data give the +1 coded factor levels in the dimensionless co-ordinate and the next two give the factor levels on a natural scale. All the runs were performed at random. Four experiments were carried out at the centerpoint level, coded as 0, for experimental error estimation Transesterification process The experimental design applied to this step was a full twolevel factorial design 2 2 and amplified to RSM. The response selected, Y, was the yield of methyl ester. The factors chosen were reaction temperature, X T, and initial catalyst concentration, X C. An excess of methanol is necessary to drive the equilibrium toward methyl ester formation. In this sense, initial alcohol/oil molar ratio was fixed at 6:1, working pressure was fixed at atmospheric and the impeller speed was fixed at 600 rpm to avoid mass transfer limitations on the process [4, 11]. Temperature levels were selected according to reactant properties and on the basis of other studies [11, 12]; so, the lower value was set at 308C and the higher was chosen as 608C. The levels of catalyst concentration were chosen on the basis of preliminary experiments [13], the amount of catalyst was progressively increased and the ester yield was monitored versus time. The levels chosen were 0.8 and 1.2 wt%, referring to the whole mass reaction. The experimental matrix for the factorial design is shown in Table 3. The use of analysis and factorial design of experiments allowed us to express the AV of the pretreated International Journal of Low-Carbon Technologies 2012, 7,

4 A. Bouaid et al. Table 3. Full 2 2 central composite design and experimental results for the transesterification process. Experiment Jatropha oil and the yield of methyl ester as polynomial models. We can write the responses, AV, and methyl ester yield, as functions of the significant factors. 3 RESULTS AND DISCUSSION 3.1 Acid-catalyzed esterification ( pre-treatment step) Linear stage The experimental design applied in the first step was a 2 2 factorial design, to which four central points were added, to evaluate the experimental error. The results obtained are shown in Table 2. A statistical analysis was performed on these experimental values, and the main and interaction effects for the two variables were calculated. The test for statistical significance is shown in Table 4. Concentration catalyst and molar ratio effects, and concentration of catalyst molar ratio interaction were fitted by multiple regression analysis to a linear model. The response function for the significant main effects and interactions is: Y AV ¼ 0:322 þ 0:057X MR 0:037X C þ 0:028X MRC r ¼ 1 Coded design levels Real values Y JOME (%) X T X C T (8C) C (%) þ þ þ1 þ a a a a From the statistical analysis, it can be concluded that within the experimental range, the molar ratio of methanol to FFA contents of Jatropha crude oil (X MR ) is a significant factor in the range studied (20:1 28:1) affecting the process of AV reduction in Jatropha oil. It has a positive influence on the response, at higher methanol to oil molar ratio (28:1), the AV in the final product was increased, and this could be due to the fact that higher amount of methanol in the system could dilute the system, resulting in a reduction in H 2 SO 4 efficiency. The effect of MR is greater than that of the catalyst concentration. The initial catalyst concentration (H 2 SO 4 acid) influence is ð1þ Table 4. Statistical analysis for pretreatment process. Y(AV (mg KOH/g)) Main effects and interactions: I MR ¼ 0.114, I C ¼ , I MRC ¼ Significance test (confidence level: 95%) Mean responses Y ¼ 0.32 Standard deviation S ¼ 0.02, t ¼ Confidence interval: Significant variables: MR(þ), C(2), MR-C(þ) Response equation Y ¼ 0:322 þ 0:057X MR 0:037X C þ 0:028X MRC r ¼ 1 Table 5. Statistical analysis for the transesterification process. ME yield (%wt) Lineal model: Main effects and interactions I T ¼ 2.08, I C ¼ 1.88, I TC ¼ Significance test (confidence level: 95%) Mean responses Y ¼ Standard deviation S ¼ 0.15, t ¼ Confidence interval: Significant variables: T(þ), C(þ), TC(2) Significance of curvature C ¼ Y2Y C ¼ 0.55 Confidence curvature interval: Significance: Si Response equation Y ¼ 98:34 þ 1:04X T þ 0:94X C 0:62X TC r ¼ 1 Quadratic model Main effects and interactions I T ¼ 0.86, I C ¼ 1.02, I TC ¼ 21.48, I T 2 ¼ 20.28, I C 2 ¼ Response equation Y JOME ¼ 98:97 þ 0:43X T þ 0:51X C 0:74X TC 0:14X 2 T 0:15X2 C r ¼ 1 statistically significant in the range studied (1 5%). This effect has a negative influence on the process. Interaction of the main significant effects molar ratio and catalyst concentration (X MR-C ) is significant and positively affects the esterification process of Jatropha oil. The minimum acid value is achieved working at the minimum level of molar ratio (20:1) and at the maximum level of catalyst concentration (5%). 3.2 Alkaline-catalyzed transesterification (second step) Linear stage The results obtained are shown in Table 3. A statistical analysis was performed on these experimental values, and then the statistically significant and interaction effects for two variables were calculated. The test for statistical significance is shown in Table International Journal of Low-Carbon Technologies 2012, 7,

5 Optimization of a two-step process for biodiesel production Temperature (X T ), concentration catalyst (X C ) effects and their interactions (X TC ) were fitted by multiple regression analysis to a linear model. The response function for the main significant effects and interactions can be expressed as: Y JOME ¼ 98:34 þ 1:04X T þ 0:94X C 0:62X TC r ¼ 1 ð2þ As observed in the statistical analysis, the concentrations of the catalyst and temperature are significant factors. The statistical analysis of experimental results also indicates that there is a significant curvature effect for the Jatropha oil methyl ester (JOME) process. It was therefore necessary to consider a different design, which allows us to fit the data to a second-order model Non-linear stage According to the central composite design methodology, a second-order model is required for JOME synthesis; the experiments have been amplified using an RSM. Four additional runs, called star points and coded +a, were added to the 2 2 factorial plus centre-points to form a central composite design, where a, the distance from the origin to the star point, is given by a ¼ 2 n/4, in the design, a ¼ The full central composite design, adapted from Box and Wilson [10], includes factorial points, centre points and star points, and is shown in Table 3. The corresponding model is the complete quadratic surface between the response and the factors, as shown by Equation (3): Y ¼ a 0 þ X2 k¼1 a k X k þ X2 k¼1 a kk X 2 k þ X2 k=j a kj X k X j where a 0 is intercept, a k first-order model coefficient, a kk quadratic coefficient for the ith variable, a kj interaction coefficients for the interaction of variables k and j. The influence of parameters on the quadratic model is shown in Table 5. The coefficients of Equation (3) were determined by multiple regression analysis. This analysis includes all the independent variables and their interactions, regardless of their significance levels. The best-fitting response surfaces can be expressed by the following statistical model: Y JOME ¼ 98:97 þ 0:43X T þ 0:51X C 0:74X TC 0:14X 2 T r ¼ 0:97 0:15X 2 C The statistical model was obtained from coded levels. Equation (4) can be represented as a dimensional surfaces plot (Figure 2), revealing the predicted yields for JOME within the investigated range of temperature and initial catalyst concentration Analysis of factors affecting the transesterification process From the statistical analysis, it can be concluded that within the experimental range, initial catalyst concentration is a ð3þ ð4þ Figure 2. Response surface plot of JOME yield as a function of temperature and catalyst concentration. significant factor affecting the process of JOME production. The effect of catalyst concentration has a positive influence on the response. The temperature influence is statistically significant in the range studied (30 608C). This effect has a positive influence on the process. Interaction of the main significant effects temperature and catalyst concentration (T C) is significant and negatively affects the transesterification process of JOME production, possibly due to the formation of emulsions and byproducts, such as soaps Analysis of response: ester yield The ester yield generally increases with increasing catalyst concentration and temperature, but it progressively decrease at high levels of both reaction temperature and catalyst concentration. This finding may be explained by the formation of byproducts, possibly due to saponification processes, side reactions which are favored at high catalyst concentrations and temperatures. This side reaction produces potassium soaps and thus decreases the ester yield. The FFAs neutralization could not be substantial since the AV for the pretreated Jatropha oil was only 0.2 mg KOH/g. Consequently, triglyceride saponification must be the only possible side reaction. This is due to the presence of the metoxide group that originated soaps by triglyceride saponification. Owing to their polarity, the soaps dissolved into the glycerol phase during the separation stage after the reaction. In addition, the dissolved soaps increased the solubility of methyl ester in the glycerol phase, and this involved an additional loss of methyl ester yield. The surface and the contour plot of JOME yield versus temperature and catalyst concentration obtained when individual International Journal of Low-Carbon Technologies 2012, 7,

6 A. Bouaid et al. molar ratio. According to these conditions, maximum conversion rates.98% for JOME could be obtained. Figure 3. Residual plot of JOME yield for the second-order model. Table 6. Quality control of JOME compared with EN Properties JOME EU Standard, EN Density (kg/m 3 )at158c Viscosity (mm 2 /s) at 408C 4.70 Max Flash point (8C) 165 Min. 120 Acid value (mg KOH/g) 0.24 Max Iodine value (mg I 2 /g) 113 Max. 120 Cetane number 56.8 Min. 47 Water content 200 Max. 500 mg/kg Ester contents (wt%).98.0 Min. 96.5% (m/m) Monoglyceride content (wt%) 0.40 Max. 0.80% (m/m) Diglyceride content (wt%) 0.15 Max. 0.20% (m/m) Triglyceride content (wt%) 0.00 Max. 0.20% (m/m) Free glycerol (wt%) Max. 0.02% (m/m) Oxidative stability (h) 2.83 Min. 6 h Cloud point (8C) 3.00 a Pour point (8C) 1.00 a Cloud filter plugging point (CFPP) Depending on the country a Not specified. EN uses time- and location-dependent values for the CFPP instead. experimental data are plotted is shown in Figure 2. The comparison among these plots shows that the maximum ester yield is achieved at the medium level for both the operation temperature and catalyst concentration. Figure 3 presents a plot of the residual distribution, defined as the difference between calculated and observed values, over the observed values for the response studied JOME yield. The quality of the fit is good because the residual distribution does not follow any trend with respect to the predicted variables. All the residuals are smaller than 1.2%, which indicates that the model adequately represent the methyl esters yield over the experimental range studied. Lower temperature and insufficient amount of catalyst resulted in incomplete conversion of triglycerides into esters. Higher temperature would lead to methanol losses, causing catalyst concentrations larger than 1.2 wt%, which is not recommended because undesirable soap formation may occur, leading to product loss and purification problems. However, from an economic point of view, the best conditions for the JOME process are a catalyst concentration of 0.95% and an operation temperature of 458C working with 6:1 methanol/oil Quality control of JOME Some of the important quality parameters of biodiesel, viscosity, AV, ester contents, CP, PP and oxidative stability for the optimum reaction conditions, are shown in Table 6. The measured values were in agreement with European Union Standard EN The kinematics viscosity of JOME was 4.7 mm 2 /s at 408C and is within the range specified in EN The acid value was 0.24 mg KOH/g, well within the maximum 0.5 mg KOH/g set in EN The flash point of JOME was 1658C and the cetane number was 56.8; both fulfilled the requirement in EN Cold flow: JOME displayed a CP of 38C, a PP of 18C and a cold filter plugging point (CFPP) of 238C; these values are relatively high. However, JOME is suitable to be used as biodiesel in hot climate conditions, even in cold weather, the cold flow properties could be improved by many kind of treatments as described by Nestor et al. [14]. It may be noted that the CP is the parameter contained in the biodiesel standard ASTM D6751, while the European standard EN prescribes the CFPP. Oxidative stability of JOME was determined by the Rancimat method EN14214, and the average of two tests was 2.8 h. The biodiesel sample does not meet the oxidative stability requirements in the EN14214 standard. JOME sample showed poor oxidative stability behavior, possibly due to the higher degree of unsaturation (unsaturated compounds ¼ 66.79%) of JOME. However, the nature and physicochemical properties of the JOME composition, and the presence of mono-, diglycerides (intermediates in the transesterification reaction) and/or glycerol, may play a major role in oxidative stability and cold flow properties. According to the biodiesel standard EN 14214, the monoglycerides content should be lower than 0.8 wt%, with diglycerides and triglycerides contents each lower than 0.2 wt%. In addition, the ester content should be 96.5 wt%. For JOME, the contents of ester were more than 98% and individual glycerides (MG, DG and TG) were within the three specifications, which implies that the transesterification reaction was complete. 4 CONCLUSIONS In the present work, design of experiments has been applied to optimize the synthesis process of FAME from crude Jatropha oil via two steps. The optimum conditions for biodiesel production were obtained when using methanol to FFA contents of Jatropha crude oil molar ratio of 20:1, 5 wt% of sulfuric acid, at 608C with a reaction time of 60 min in the first step, followed by using the molar ratio of methanol to product from the first step of 6:1, 0.95 wt% of KOCH 3,at458C with a reaction time of 60 min in the second step. According to this study, the maximum yield of methyl ester.(98%) can be obtained. 336 International Journal of Low-Carbon Technologies 2012, 7,

7 Optimization of a two-step process for biodiesel production These models are useful to determine the optimum operating conditions for the industrial process using the minimal number of experiments with the consequent benefit from economical point of view. The FAME (biodiesel), produced from Jatropha oil through the two-step catalyzed process, can be used as a diesel fuel substitute since it conforms to European Biodiesel Standard EN These results make Jatropha oil a promising oil feedstock for cultivation in areas of (centralsouthern) Spain, and could offer the possibility of exploiting the Mediterranean marginal areas for energy purposes. ACKNOWLEDGEMENTS Financial support from the (CICYT) Spanish project CTQ is gratefully acknowledged. REFERENCES [1] Zhang Y, Dube MA, McLean DD, et al. Biodiesel production from waste cooking oil: economic assessment and sensitivity analysis. Bioresour Technol 2003;90: [2] Staat F, Vallet E. Vegetable oil methyl ester as a diesel substitute. Chem Ind 1994;7: [3] Knothe G. Analyzing biodiesel: standards and other methods (review). JAOCS 2006;83: [4] Vicente G, Coteron A, Martínez M, et al. Application of the factorial design of experiments and response surface methodology to optimize biodiesel production. Ind Crop Prod 1998;8: [5] Shuit SH, Lee KT, Kamaruddin AH, et al. Reactive extraction and in situ esterification of Jatropha curcas L. seeds for the production of biodiesel. Fuel 2010;89: [6] Shah S, Gupta M. Lipase catalyzed preparation of biodiesel from Jatropha oil in a solvent free system. Process Biochem 2007;42: [7] Jain S, Sharma MN. Prospects of biodiesel from Jatropha in India: a review. Renew Sust Energy Rev 2010;14: [8] Tapanes NCO, Gomes Aranda DA, de Mesquita Carneiro JW, et al. Transesterification of Jatropha curcas oil glycerides: theoretical and experimental studies of biodiesel reaction. Fuel 2008;87: [9] Leung DYC, Wu X, Leung MKH. A review on biodiesel production using catalyzed transesterification. Appl Energy 2010;87: [10] Box J, Wilson W. Central composites design. J R Stat Soc 1951;1:1 35. [11] Bouaid A, Martinez M, Aracil J. Pilot plant studies of biodiesel production using Brassica carinata as raw material. Catal Today 2005;106: [12] Vicente G, Martinez M, Aracil J. Optimization of Brassica carinata oil methanolysis for biodiesel production. JAOCS 2005;82: [13] Vicente G, Martínez M, Aracil J. Integrated biodiesel production: a comparison of different homogenous catalysts systems. Bioresour Technol 2004;92: [14] Nestor U, Soriano J, Migo VP, et al. Ozonized vegetable oil as pour point depressant for neat biodiesel. Fuel 2006;85: International Journal of Low-Carbon Technologies 2012, 7,