CCHHEEMMIICCAALL EENNGGIINNEEEERRIINNGG TTRRAANNSSAACCTTIIOONNSS VOL. 45, 2015 A publication of The Italian Association of Chemical Engineering www.aidic.it/cet Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Sharifah Rafidah Wan Alwi, Jun Yow Yong, Xia Liu Copyright © 2015, AIDIC Servizi S.r.l., ISBN 978-88-95608-36-5; ISSN 2283-9216 DOI:10.3303/CET1545253 Please cite this article as: Yusuf N.R., Kamil R.N.M., Yusup S., 2015, Response surface methodology approach for optimisation of alkaline hydrolysis of jatropha curcas oil via microwave assisted, Chemical Engineering Transactions, 45, 1513-1518 DOI:10.3303/CET1545253 1513 Response Surface Methodology Approach for Optimisation of Alkaline Hydrolysis of Jatropha Curcas oil via Microwave Assisted Nur’aini R. Yusuf a , Ruzaimah N. M. Kamil* a , Suzana Yusup b a Fundamental and Applied Sciences Department Universiti of PETRONAS, Perak. b Biomass Processing Lab, Center of Biofuel & Biochemical Research, Green Technology Mission Oriented, University of Teknologi PETRONAS, Perak. ruzaimah_nikmkamil@petronas.com.my The production of fatty acid by alkaline hydrolysis method from Jatropha Curcas oil using microwave assisted irradation system has been investigated. This fatty acid is widely used as raw materials for biochemicals industry. The optimisation of alkaline hydrolysis variables such as temperature, catalyst concentration loading, reaction time and oil to solvent ratio was carried out using central composite design (CCD) in response to surface methodology (RSM) to achieve maximum free fatty acid yield (% FFA). Analysis of variance (ANOVA) analysis shows that the catalyst concentration are the most significant variance effecting the yield of % FFA, with R 2 = 0.9827 indicating good fitting of the suggested model. The highest fatty acids yield was 98 % achieved at the optimum condition of 95 °C, 1.75 M of ethanolic KOH concentration, 1:68 oil to solvent ratio and 15 min of reaction time in 3,000 W microwave power system. This study reports under optimised condition, fatty acids yield was significantly increased and this also showed the potential of Jatropha Curcas oil as feedstock for fatty acids production. 1. Introduction Microwaves are electromagnetic energy, which has frequency and wavelength in the ranges of 300 MHz to 300 GHz, and 0.001 m to 1 m (Costa et al., 2013). Lately microwave irradiation is becoming worldwide popular method of heating which replace the conventional method because it proved to be clean, convenient and most important it affords higher yields and used in shorter time (Palma et al., 2013).. Microwave radiation causes molecular friction of electric dipole, which results in heating. Lipids have low specific heat, which makes them susceptible to this radiation. Microwave preheating can be used to substitute conventional heating results in fast pre-treating with less energy consumption. This also generates permanent pores in the oil that makes higher yield. This method of heating has been extended to almost all area of chemistry with the exception of the carbohydrate i.e. fatty acids but it is testified by the limited number of applications (Chiavaro et al., 2010). Fatty acids (FA)s is an important components and precursors used in the preparation of a wide variety of products such as soaps, surfactants, lubricants, plasticizers, paints, coatings, pharmaceuticals, foods, agricultural, industrial and personal care products (Gunstone and Hamilton, 2001). Currently, these fatty acids have also become a low-cost and biodegradability raw material for polyol esters synthesis (Schuchardt et al., 1998). Conventional, fatty acids are obtained from the reaction of vegetable oils and/or animals fats with superheated steams (Patil et al., 1988). This reaction is carried out at 100 - 260 °C and 100 – 7,000 kPa using a 0.4-1.5 wt % water-to-oil ratio. Different variations of this technology have been used by industry (e.g. Twitchell process, the Colgate-Emery synthesis and the Eisenlohr process) (Twitchell, 1898). These processes are normally referred to as fat splitting. The main drawbacks of this process are the requirement of a distillation unit to further clean free fatty acids (FFA). Researchers have employed several methods to prepare FAs such as enzymatic hydrolysis using lipases from Aspergillus niger, Rhizopus javanicus and Penicillium solitum (Carvalho et al., 2009), C.rugosa (Serri et al., 2008) and mailto:ruzaimah_nikmkamil@petronas.com.my 1514 subcritical water (Pinto and Lantas, 2006). Nowadays, researchers used potassium hydroxide catalyzed hydrolysis of esters sometimes known as saponification because of its relationship with soap making. There are two big advantages of doing this. The reactions are one-way rather than reversible, and the products are easier to separate (Salimon et al., 2011). On laboratory scale, typically potassium or sodium hydroxide in ethanol, refluxing for 1 h, and the FAs recovered after acidification of the reaction mixture. This is a sufficiently mild procedure that most FAs, including polyunsaturates, epoxides, and cyclopropenes, are unaltered. The earliest known study involving microwave is pre-treatment of C18 fatty acids esterification (Melo-Junior et al., 2009). The results showed conversion up to 60 % in 60 min of reaction under relatively mild conditions and non-catalytic system. It has been proposed that in case of low power of high-frequency electromagnetic field the non-thermal activation of enzyme may be observed. The application of microwave in alkaline hydrolysis has so far rarely found. Therefore, the aim of this study was to investigate the efficiency of pre-treatment alkaline hydrolysis under microwave irradiation. The Response surface methodology (RSM) was applied to design the optimum operating conditions for the fatty acids production. 2. Methodology 2.1 Materials and chemicals Jatropha curcas oil was purchased from Bionas Sdn Bhd, Malaysia. Potassium hydroxide pellets (purity of 99 %), isopropanol, phenolphthalein, sodium hydroxide and n-hexane were obtained from Merck, Darmstadt, Germany. Ethanol (purity of 96 %), anhydrous magnesium sulphate, Hydrochloric acid (HCl), Methanol dry (MeOH dry) and Toluene were purchased from Fisher Chemicals (UK). The Jatropha oil and all chemicals are of analytical grade and were used directly without any purification. Microwave treatments were carried out using a domestic microwave synthesis reactor (Fabricated by Dixson Company) with variable power levels of 3,000 W. 2.2 Preparation of fatty acids by microwave system The fatty acid process was carried out as described by Salimon et al. (2011). Briefly, the batch mixture containing Jatropha curcas seed oil and ethanolic KOH was mixed in a 500 mL temperature-controlled reactor. The reactor vessel was placed in a microwave synthesis reactor. The contents in the reaction flask was stirred using mechanical stirrer. In addition, due to high temperature, a reflux system was established for condensing sample into reactor. After the reaction was completed, the sample was cooled and 200 mL water was added to the mixture to enhance separation of fatty acids. A 100 mL of hexane was added to extract the unsaponifiables. The aqueous alcohol phase, containing the soaps, was acidified to pH 1 with HCl 6N, and the FFA was recovered by extraction with hexane. The extract was washed with distilled water until neutral. The FFA-containing upper layer was dried with anhydrous magnesium sulphate, and solvent was evaporated in a vacuum rotary evaporator at 50 °C. 2.2 Determination of the free fatty acids (FFA) Samples were taken from the reaction mixture according to experimental design and was used for analysis of free fatty acid concentration (FFA %). The FFA% from alkaline hydrolysis of Jatropha curcas seed oil was determined according to acid value test as reference method (AOCS, 1999). The FFA % was calculated by using Eq(1). % FFA = MW x N x (A-B) / W (2) where N is normality of titrant, MW is molecular weight of KOH, W is the weight of sample, A is volume of titrant used for sample, B is volume of titrant used for blank. 2.3 Design of experiments The free fatty acids yield of alkaline hydrolysis was taken as response and influenced by various potential variables. The Design Expert software version 8 (INC,USA) was used to generate experiment value from central composite rotatable design (CCRD) procedure and analysed by Response Surface Methodology (RSM) and also for determined the effect of independent variables on response and optimise the alkaline hydrolysis. A 3 4 full factorial experiment design, 24 experiments with 6 replication at design centre were carried out randomly to estimate the error (total of 30 runs) was employed to generate the experiment data for RSM models. The variables include temperature (A), catalyst concentration (B), time (C) and oil to solvent molar ratio (D). The coded values were listed in Table 1. 1515 Table 1: Experimental range and level coded of independent variables for the central composite design Variables Units Type Range and level (coded) -2 -1 0 + 1 +2 Temperature Deg Celcius ( o C) Factor A 50 65 80 95 110 Catalyst concentration Molarity (M) Factor B 1.5 1.75 2.00 2.25 2.50 Time Minute (min) Factor C 7.5 15.00 22.50 30.00 37.50 Oil to Solvent ratio Factor D 58 68 78.00 88 98.00 FFA Percentage (%) Response 67.1 98 3. Result & Discussion 3.1 Optimisation of alkaline hydrolysed for increased fatty acids yield. The optimal condition of the alkaline hydrolysis was chosen based on reaction of triglyceride with ethanol and KOH to produce glycerol and fatty acids. The initial results of optimal conditions are as follow: oil/ ethanol ratio, 1:68; KOH concentration, 1.75 M; time, 15 min; temperature 90 °C which is summarized in Table 2, which also shows experiment design and results of CCD for alkaline hydrolysis. The overall second-order as shown in Eq(2) describes the relationship between the variables and the free fatty acids yield in terms of coded values; where Y is the predicted response, A-D are actual variables of predictor; temperature (A), catalyst concentration (B), time (C) and oil to solvent molar ratio (D). The positive sign in front of terms reveals synergistic effect while negative sign reveals the antagonistic effect that would indicate the influence of independents variables on the process (Saidi et al., 2013). 3.2 Analysis of Regression model and Statistical Analysis of variance (ANOVA) for fatty acid yield was used in order to check the adequacy of the empirical model and the results were summarized in Table 2. To measure how well the suggested model fit the experiment data, the parameter such as R 2 , p-value, F-value, and lack of fit were used (Ceylan, 2008).The actual values are obtained from response data of experimental runs while the predicted values were generated by using the approximation functions based on the quadratic model of CCD (Mazaheri, 2001) and tabulated in Table 3. Table 2: Analysis of variance (ANOVA) for response surface quadratic model Source of variance Sum of squares Degree of freedom (df) Mean Square F-Value P-Value Prob>F Model 3,352.67 14 239.48 60.96 < 0.0001 A-Temperature 5.56 1 5.56 1.41 0.2527 B-Catalyst Concentration 172.00 1 172.00 43.78 < 0.0001 C- Time 111.16 1 111.16 28.29 < 0.0001 D- Oil to alcohol ratio 39.91 1 39.91 10.16 0.0061 AB 89.54 1 89.54 22.79 0.0002 AC 0.66 1 0.66 0.17 0.6877 AD 437.33 1 437.33 111.32 < 0.0001 BC 21.51 1 21.51 5.47 0.0335 BD 120.18 1 120.18 30.59 < 0.0001 CD 2.00 1 2.00 0.51 0.4870 A 2 192.84 1 192.84 49.09 < 0.0001 B 2 987.94 1 987.94 251.47 < 0.0001 C 2 706.88 1 706.88 179.93 < 0.0001 D 2 1,292.31 1 1292.31 328.95 < 0.0001 Residual 58.93 15 3.93 Lack of fit 58.67 10 5.87 112.83 < 0.0001 Y = +67.50 – 0.48A -2.68B – 2.15C -1.29 D + 2.37AB +0.20AC - 5.23AD + 1.16 BC +2.74BC + 0.35 CD +2.65A 2 + 6.00B 2 +5.08C 2 +6.86D 2 (2) 1516 As according to Table 2, the p-value of the model is less than 0.0001 demonstrating high significance in predicting the response value and the suitability of the deduced model. The significance of all coefficients was established by P-values also shown in same table. Catalyst loading was the most important operating variable based on higher F-value compared to other factorial (Koochecki et al, 2009). Another critical determination are the corresponding coefficients of determination R 2 = 0.9827, adj-R 2 = 0.9666. They were consider relatively high, which indicate good agreement between experimental and predicted data as reported by Little and Hill (1978). Table 3: Experimental design and results of CCD for alkaline hydrolysis Run order Temperature (°C); A Catalyst concentration (M); B Time (min); C Oil to solvent ratio; D Free Fatty Acids ( % FFA) 1 80.00 2.0 22.50 78 67.6 2 80.00 2.0 22.50 78 67.7 3 80.00 2.0 7.50 78 93.3 4 50.00 2.0 22.50 78 78.0 5 80.00 2.5 22.50 78 84.8 6 95.00 2.25 15.00 68 91.7 7 95.00 1.75 15.00 68 98 8 80.00 2.00 22.50 78 67.1 9 65.00 1.75 15.00 88 95.5 10 95.00 2.25 30.00 88 83.9 11 65.00 2.25 15.00 88 91.0 12 80.00 2.00 22.50 98 91.7 13 95.00 1.75 30.00 88 73.8 14 80.00 2.00 22.50 78 67.5 15 65.00 2.25 30.00 88 91.5 16 95.00 2.25 30.00 68 91.9 17 80.00 2.00 22.50 78 67.7 18 65.00 1.75 30.00 68 88.9 19 80.00 2.00 22.50 78 67.6 20 80.00 2.00 22.50 58 97.1 21 65.00 1.75 30.00 88 93.8 22 95.00 1.75 30.00 68 95.4 23 110.00 2.00 22.50 78 77.1 24 80.00 1.50 22.50 78 97.1 25 65.00 2.25 30.00 68 74 26 65.00 1.75 15.00 68 97.1 27 65.00 2.25 15.00 68 80 28 95.00 1.75 15.00 88 84.3 29 80.00 2.00 37.50 78 81.2 30 95.00 2.25 15.00 88 83.1 R 2 = 0.9827 R 2 (Pred) = 0.9008 R 2 (adj) = 0.9666 3.3 Interaction between process variables ANOVA results in table 2 shows interaction between variables and have significant effect on the yield of FFA. This were determined by P-Value < 0.0001 which are interaction between temperature and oil to solvent ratio and interaction between catalyst concentration and oil to solvent ration give most significant effect to the FFA yield. This also can be seen at Figure 1(a), the parallel lines connecting each other in interaction plot shows significant interaction. Therefore, instead of study single variable (as in conventional) the interactions will be investigated which is significance and important for a comprehensive optimisation study. 1517 Design-Expert® Software Factor Coding: Actual FFA CI Bands X1 = A: Temperature X2 = B: Catalyst concentration Actual Factors C: Time = 22.50 D: Oil to solvent ratio = 78.00 B- 1.75 B+ 2.25 B: Catalyst concentration 65.00 71.00 77.00 83.00 89.00 95.00 A: Temperature F F A 65 70 75 80 85 Interaction Design-Expert® Software Factor Coding: Actual FFA 98 67.1 X1 = A: Temperature X2 = D: Oil to solvent ratio Actual Factors B: Catalyst concentration = 2.00 C: Time = 22.50 68.00 73.00 78.00 83.00 88.00 65.00 71.00 77.00 83.00 89.00 95.00 65 70 75 80 85 F F A A: Temperature D: Oil to solvent ratio Design-Expert® Software Factor Coding: Actual FFA 98 67.1 X1 = B: Catalyst concentration X2 = D: Oil to solvent ratio Actual Factors A: Temperature = 80.00 C: Time = 22.50 68.00 73.00 78.00 83.00 88.00 1.75 1.85 1.95 2.05 2.15 2.25 65 70 75 80 85 90 F F A B: Catalyst concentration D: Oil to solvent ratio Figure 1(b) shows the effect of different reaction temperature and oil to solvent ratio on the FFA yield in three dimensional surface response. From the figure, at any point quantity of molar ratio from 68 to 88, the yield of FFA increase proportionally with reaction temperature but the yield augmented when the molar ratio was reduced from 88 to 68 at any constant temperature. The observed phenomena shows increasing reaction temperature enhance the reaction rate therefore enhance the FFA yield (Goyal et al., 2012). The effect of catalyst concentration and temperature shows in Figure 1(c). At low temperature (65 °C) and low catalyst concentration (1.75 M), shows the high yield but at higher temperature the yield become almost the same. Reaction temperature plays a crucial role in determination of reaction rate (Saka and Isayama, 2009). Figure 1(d) represents the effect of oil to solvent ratio to catalyst concentration. Within the studied of molar ratio, increment in the ratio slightly enhance the yield but at some point the yield decrease when molar ratio increase. This figure also shows influence of catalyst concentration on FFA yield. It can be seen the yield increase at low amount of catalyst. Figure 1: Interaction between (a) Interaction plot (b) Oil to solvent ratio and temperature (c) Catalyst concentration and temperature (d) Oil to solvent ratio and catalyst concentration Conclusion Microwave assisted for fatty acids production was carried out to determine the effect of various parameters on FFA yield. In order to optimise the yield, numerical feature of Design Expert software was applied to find the optimum combination condition which result in maximum yield. The highest FFA achieved (98 %) under optimum conditions of 95 °C, 1.75 M of ethanolic KOH concentration, 1:68 oil to solvent ratio and 15 min of reaction time. The model is validated by RSM with R 2 value = 0.9827. Acknowledgments The authors gratefully acknowledge Ministry of Higher Education Malaysia (MOHE) LRGS Grant 0153AB-I41 and University of PETRONAS STIRF Grant 0153AA-D02 for providing financial support to carry out this research. Authors also thankful to help and co-operation received from project supervisor and laboratory technologists of the Biomass Processing Laboratory, Universiti Teknologi PETRONAS. (c) F F A ( % ) (b) (a) (d) (c) F F A ( % ) F F A ( % ) F F A ( % ) 1518 References AOCS 1999. Acid Value. Official Methods of Analysis. Association of Official Analytical Chemist, Arlington, Cd 3d-63. Carvalho P.D.O., Campos P.R., Noffs M.D.A., Fregolente P.B., Fregolente L.V., 2009. Enzymatic hydrolysis of salmon oil by native lipases: optimisation of process parameters, J. Braz. Chem. Soc., 20, 117- 124. Ceylan H., Kubilay S., Aktas N., Sahiner N., 2008, An approach for prediction of optimum reaction conditions for laccase-catalyzed bio-transformation of 1-naphthol by response surface methodology (RSM), Bioresour Technol, 99, 2025-203. Chiavaro E., Rodriguez-Estrada M.T., Vittadini E., Pellegrini N., 2010, Microwave heating of different vegetable oils: Relation between chemical and thermal parameters, Lwt-Food Sci Technol, 43, 1104- 1112. Costa S.S., Arumugamb D., Gariepyb Y., Rochaa, S.C., Raghavanb, V., 2013. Spilanthol extraction using microwave: calibration curve for gas chromatography, Chemical Engineering Transactions, 32, 1783- 1788. Goyal P., Sharma M.P., Jain S., 2012, Optimisation of conversion of high free fatty acid Jatropha curcas oil to biodiesel using response surface methodology, ISRN Chemical Engineering, 2012, 1-8, DOI: org/10.5402/2012/327049. Gunstone F.D., Hamilton R.J., 2001, Oleochemical Manufacture and Applications, Sheffield Chemistry and Technology of Oils and Fats, Blackwell, CRC Press, Oxford, UK. King J., Holliday R., List G., 1999, Hydrolysis of Soybean Oil in a Subcritical Water Flow Reactor, Green Chem., 1, 261–264. Koocheki A., Taherian A.R., Razavi S.M., Bostan A., 2009, Response surface methodology for optimisation of extraction yield, viscosity, hue and emulsion stability of mucilage extracted from Lepidium perfoliatumseeds, Food Hydrocolloid, 23, 2369-2379. Little T.M., Hills F.J., 1978, Agricultural experimentation: Design and analysis, John Wiley & Sons, New York, USA. Mazaheri H., Lee K.T., Bhatia S., Mohamed A.R., 2010, Subcritical water liquefaction of oil palm fruit press fiber for the production of bio-oil: effect of catalysts, Bioresour. Technol., 101, 745-751. Melo- ni r . . u u r u .E., Fortuny M., Dariva C., Egues S., Santos A.F., Ramos A.L., 2009, Use of microwave irradiation in the noncatalytic esterification of C18 fatty acids, Energy Fuels, 23, 580- 585. Palma V., Ciambelli, P., Meloni E., 2013, Catalyst Load Optimisation for Microwave Susceptible Catalysed DPF, Chemical Engineering Transactions, 32, 799-804. Pinto J.S.S., Lanças F.M., 2006, Hydrolysis of corn oil using subcritical water, J. Braz. Chem. Soc., 17, 85- 89. Patil T.A., Butala D.N., Raghunathan T.S., Shankar H.S., 1988, Thermal Hydrolysis of Vegetable-Oils and Fats Reaction-Kinetics, Ind. Eng. Chem. Res., 27, 727–735. Schuchardt U., Sercheli R., Vargas R.M.J., 1998. Transesterification of vegetable oils: a review, J. Braz. Chem. Soc., 9, 199–210. Twitchell E., 1898, Twitchell Process of Decomposing Fats or Oils into Fatty Acids and Glycerin Thereof. U.S. Patent .601-603. Saka S., Isayama Y., 2009, A new process for catalyst-free production of biodiesel using supercritical methyl acetate, Fuel, 88, 1307-1313. Salimon J., Abdullah B.M., Salih N., 2011, Hydrolysis optimisation and characterization study of preparing fatty acids from Jatropha curcas seed oil, Chem. Cent. J., 5, 1-9. Saidi S., Belleville M.P., Deratani, A., Amar R.B., 2013. Optimisation of peptide production by enzymatic hydrolysis of tuna dark muscle by-product using commercial proteases, Afr. J. Biotechnol., 12, 1533- 1547. Serri N.A., Kamarudin A.H., Abdul Rahaman S.N., 2008, Preliminary studies for production of fatty acids from hydrolysis of cooking palm oil using C. rugosa lipase, J Phys Sci.,19, 79-88. http://dx.doi.org/10.5402/2012/327049 http://dx.doi.org/10.5402/2012/327049