Optimization of transesterification of Jatropha curcas Oil to Biodiesel using Response Surface Methodology and its Adulteration with Kerosene

Transcription

1 J. Mater. Envron. Sc. 4 () (013) Goyal et al. Optmzaton of transesterfcaton of Jatropha curcas Ol to Bodesel usng Response Surface Methodology and ts Adulteraton wth Kerosene P. Goyal 1, M.P. Sharma, S. Jan 3* 1 Bofuel Research Laboratory, Alternate Hydro Energy Centre, Indan Insttute of Technology Roorkee, Uttarakhand , Inda, Emal: prernagoyal8@gmal.com Bofuel Research Laboratory, Alternate Hydro Energy Centre, Indan Insttute of Technology Roorkee, Uttarakhand , Inda, Emal: mahendrapal.sharma@gmal.com 3 NUS Envronmental Research Insttute, Natonal Unversty of Sngapore, Sngapore Receved Oct 01, Revsed 7 Oct 01, Accepted 7 Oct 01 *Correspondng Author: - Emal: arthjan001@gmal.com, Tel: , Abstract Bodesel has emerged as an alternatve to non-renewable desel fuel. Growng demand for bodesel has gven brth to mal practces such as adulteraton whch degrade ts qualty. The present study has dealt wth the optmzaton of bodesel producton from Jatropha curcas ol usng a fve-level-four-factoral central composte desgn based on response surface methodology n 54 expermental runs. A bodesel yeld of 98.3% s obtaned wth methanol/ol rato (11:1 w/w) usng NaOH as catalyst (1% w/w) n 110 mn at 55 o C. ANOVA results revealed that catalyst concentraton, reacton tme and methanol/ol molar rato had a sgnfcant effect on JCB yeld. A model equaton for predctng the yeld of bodesel s formulated whch can be successfully adopted n ol ndustry to maxmze the yeld of methyl esters. The propertes of the bodesel, thus, produced conform to the ASTM and BIS specfcatons, makng t an deal alternatve fuel for desel engnes. The adulteraton of bodesel wth kerosene has been studed usng vscosty and densty measurement. A number of calbraton curves and correlatons are developed whch can be used to fnd out the extent of adulteraton n bodesel and determne ts qualty. Keywords: Adulteraton, Bodesel, Kerosene, Optmzaton, Response surface methodology, Transesterfcaton. 1. Introducton The world s presently confronted wth energy crss due to fossl fuel depleton and envronmental degradaton [1]. Ths has lead to the search for alternatve energy sources such as boethanol and bodesel. Growng demand of bodesel has gven rse to mal practce of adulteraton. Adulteraton s the addton of unwanted chemcals n bodesel to lower ts qualty and cost. The crtera for the addton of adulterants are that these should be mscble wth and cheaper than bodesel. The common adulterants used are kerosene and raw vegetable ols whch can have a negatve mpact on the engne performance wth regard to fuel consumpton, power output and engne lfe. It s a malcous practce and should be serously checked n order to meet the qualty standards of bodesel. Wth growng human populaton, more land s needed to produce food for human consumpton, whch poses a potental challenge to bodesel producton. Jatropha curcas ol (JCO) s a plant based feedstock that s unsutable for human consumpton and could be the best feedstock for bodesel producton []. For the converson of hgh free fatty acd (FFA) JCO to bodesel, two step acd-base catalyzed transesterfcaton method s used [3]. It conssts of acd catalyzed pretreatment step to reduce the FFA to less than 1% usng H SO 4 as acd catalyst and transesterfcaton of pretreated ol to bodesel usng alkal catalyst. Ths process nvolves many parameters that effect the reacton and optmzng so many reacton factors requre large number of experments, whch s laborous, tme consumng, and economcally non-vable. Response surface methodology (RSM) s a useful statstcal technque for the optmzaton of complex processes, as t reduces the number of experments requred to acheve ample data for a statstcally pertnent result [4]. So, optmzaton of bodesel producton usng RSM s of mmense nterest for meetng the growng fuel 77

2 requrements. Lterature has revealed that only few papers are avalable on the use of Response Surface Methodology (RSM) for the optmzaton of process varables to maxmze the bodesel yeld. Boonmee et al.[5] have studed the effect of three process varables vz. methanol/ol molar rato, catalyst concentraton and reacton tme on the methyl esters yeld of JCO. Central composte desgn (CCD) of 0 experments was employed and 99.87% bodesel yeld was acheved. Smlarly, Twar et al.[6] have optmzed methanol quantty and reacton tme usng RSM based on central composte rotatable desgn (CCRD) of 1 experments and obtaned bodesel yeld of 99%. In vew of the above, t can be seen that, no work s reported on the optmzaton of bodesel producton from JCO usng four process varables. The present paper, therefore, reports the results of the optmzaton of four process varables vz. catalyst (NaOH) concentraton (0-% w/w), reacton temperature (35 o -55 o C), reacton tme ( mn) and methanol/ol rato (w/w) (6:1 1:1) for the maxmzaton of Jatropha curcas bodesel (JCB) yeld. CCD of 54 experments based on RSM wth the help of Desgn Expert software has been used. A model to predct the response (JCB yeld (%)) has been formulated and valdated by Analyss of Varance (ANOVA). The model can be employed n the ol ndustry to maxmze the yeld of methyl esters. Further, the adulteraton of bodesel wth kerosene has been studed and correlatons have been developed to check the adulteraton of unknown samples of bodesel.. Materals and Methods JCO was procured from Jatropha Vkas Sansthaan, New Delh. All chemcals such as H SO 4, KOH, methanol, ethanol, Na SO 4 and NaOH were of Analytcal Reagent grade and 99% pure. NaOH n pellet form was used as a base catalyst. Kerosene was purchased from local market. The fuel propertes of JCO and kerosene determned by the methods descrbed by Jan and Sharma [3] are reported n Table 1. Table 1: Fuel propertes of JCO and Kerosene S.no Propertes JCO Kerosene 1 Densty (g/c.c at 15 o C) Vscosty (cst, at 40 o C) Flash pont ( o C) FFA contents (%) Gross calorfc value (MJ/kg) Acd- Base catalyzed transesterfcaton of hgh FFA JCO Raw JCO was fltered to remove all the nsoluble mpurtes followed by heatng at 100 o C for 10 mn to remove all the mosture. JCO had hgh FFA (14.60%), whch s far above the 1% lmt sutable for base catalyzed transesterfcaton reacton. FFAs were, therefore, frst converted to esters n a pretreatment process for the producton of JCB [3]. The reacton was carred out at a temperature of 50 o C for 15 mn. usng conc. H SO 4 (1.5 % v/v) as acd catalyst wth methanol/ol rato of 6.5:1. The hgh FFAs were reduced to 0.34% and the resultng reacton mxture was subjected to base catalyzed transesterfcaton process for bodesel producton usng NaOH as base catalyst [3]. The methyl ester layer was separated, washed wth water, heated to remove mosture and dred over anhydrous Na SO 4. The transesterfcaton of JCO has been optmzed usng RSM for the maxmzaton of JCB yeld whch was calculated usng the followng equaton (1): Yeld of JCB (%) (1) Total weght of methyl esters Total weght of ol n the sample *100% (1). Analyss of JCB The JCB was prepared n the laboratory under the operatng condtons optmzed by RSM and analyzed for fatty acd composton usng Gas chromatograph (Model-Netal). The process s already reported by Jan and Sharma [3]..3 Physo-chemcal propertes of JCB The physcal and chemcal propertes of JCB produced under optmum condtons were determned by usng standard methods [7]. 78

3 .4 Adulteraton of optmzed JCB wth Kerosene Dfferent blends were prepared by mxng JCB wth kerosene n varyng proportons and denoted by KBx,.e., kerosene bodesel blend and x: % of bodesel n blend..5 Expermental desgn of Transesterfcaton based on RSM A fve-level, four-factoral CCD was appled and the total number of experments were 54 ( k +k+6+4); where k s the number of ndependent varables [8]. Twenty four experments were augmented wth two replcatons at axal and factoral ponts and sx replcas at the centre pont to evaluate the error for carryng out the optmzaton studes to maxmze the JCB yeld. Catalyst (NaOH) concentraton (A) (% w/w), reacton temperature (B) ( o C), reacton tme (C) (mn) and methanol/ol molar rato (D) (w/w) were the ndependent varables selected to optmze the yeld of JCB. The coded and uncoded levels of the ndependent varables are gven n Table. Table : Independent varables and levels used for CCD n transesterfcaton process S.no Varables Symbols Levels Catalyst concentraton A (%w/w) Temperature ( o C) B Tme (mn) C Methanol-to-ol rato (w/w) D Statstcal Analyss The Desgn Expert software was used for the regresson and graphcal analyss of the data. The maxmum values of JCB yeld were taken as the response of the desgn experment. The expermental data obtaned by the above procedure was analyzed by the response surface regresson [9] usng the followng second-order polynomal equaton (): k y x x x x () o 1 k 1 where; y s the response (JCB yeld(%)), x and x j are the uncoded ndependent varables, and j are the lnear and quadratc coeffcents respectvely, β o s the regresson co-effcent, k s the number of factors studed and optmzed n the experment. Statstcal analyss of the model equaton and evaluaton of the analyss of varance (ANOVA) was carred out. Confrmatory experments were also performed to valdate the equaton. k j k j j j 3. Results and Dscusson 54 experments were performed to get the expermental values of JCB yeld. Expermental and predcted values for JCB yeld responses at the desgn ponts are gven n Table 3. Table 3: CCD arrangement and responses for JCB yeld Parameter 1 Parameter Parameter Parameter 4 Expermental Predcted Run A: Catalyst B: C: Tme D:Methanol/ol JCB yeld JCB yeld

4 The summary of ANOVA s provded n Table 4. The assocated Probablty (P) value for the model s lower than , mplyng the sgnfcance of the model. The value of regresson coeffcent R for the model s 0.945, ndcatng the good ftness of the model. Hgh values of predcted R (0.890) and adjusted coeffcent of determnaton (R Adj:0.93) and low value of coeffcent of varaton (C.V) (.09%), are an ndcaton of precson of ftted model [10]. P-values < 0.05 ndcate that the model terms are sgnfcant [11]. In ths case, A (catalyst concentraton), C (tme), D (rato of methanol/ol), nteracton effect of AB (catalyst concentraton wth reacton temperature), AC (catalyst concentraton wth reacton tme), BD (reacton temperature wth methanol/ol rato), CD (tme wth methanol/ol rato), B (quadratc effect of temperature), C (quadratc effect of tme) have sgnfcant effect on the JCB yeld. The regresson equaton (3) for the determnaton of predcted values of output parameter (.e. JCB yeld) s gven as follows: JCB (%) A B C D AB AC 0.58AD 0.003BC 0.5BD 0.04CD - 1.5A 0.075B C 0.003D 80 (3)

5 The graph between the predcted and actual JCB yeld (%) gven n Fgure 1 shows that the predcted values are qute close to the expermental values, thereby, valdatng the relablty of the model developed for establshng a correlaton between the process varables and the JCB yeld. Source Table 4: ANOVA for response surface quadratc model Sum of Degree of Mean F-value Squares freedom (df) Square P-value Prob>F Model < A-Catalyst concentraton < B-Temperature C-Tme < D- Methanol/ol molar rato AB < AC AD BC BD < CD < A B < C D Resdual C.V:.09% R model : 0.945, R Adj.: 0.93 Predcted R model: Fgure 1: Predcted versus actual JCB yeld (%) values 3.1 Effect of process parameters on JCB yeld (%) Fgure shows the effect of catalyst concentraton, reacton temperature, reacton tme and methanol/ol rato on JCB yeld. It can be seen from the fgure that the JCB yeld decreases sgnfcantly wth ncrease n catalyst concentraton. Ths may be due to the fact that addton of excessve catalyst causes more trglycerde to react wth the alkal catalyst leadng to the formaton of soap, whch decreases the bodesel yeld [1]. JCB yeld decreases wth ncrease n temperature tll the mddle pont s reached, after that t ncreases. Ths may be due to the fact that vscosty of ols decreases at hgh temperature resultng n an ncreased reacton rate and shortened reacton tme, thereby, ncreasng the bodesel yeld [1]. Reacton tme has a postve effect on the JCB yeld, as t ncreases wth ncrease n tme [13]. Yeld of JCB s found to ncrease wth the ncrease n methanol/ol rato; snce the transesterfcaton reacton s reversble n nature, so excess alcohol s added to ensure the total converson of trglycerdes [1]. Thus, the yeld of bodesel ncreases wth ncrease n temperature, tme and methanol quantty and decreases wth ncrease n catalyst concentraton. 81

6 Fgure : Effect of catalyst concentraton, temperature, tme and methanol/ol rato on JCB yeld (%) 3. Optmzaton of response parameters Optmzaton of ndvdual responses was performed to acheve the desred maxmzaton of JCB yeld based on equaton 3. Desgn Expert software was used to optmze the response and the value of the response (JCB yeld) was set at maxmum. The optmal value of nput process parameters s gven n Table 5. An experment was carred out at the optmal parametrc settngs for JCB yeld to obtan the targeted value of response parameter. Table 5: Optmzed nput process parameters and optmum value of JCB yeld Response Optmum value of process parameters Predcted value Expermental value A B C D JCB yeld (%) Analyss of JCB The Fatty acd (FA) composton of JCB prepared usng the above optmum parameters gven n Table 5 and determned by Gas Chromatography (GC) s gven n Table 6 whch shows that JCB manly contaned Olec and Lnolec acd. The FA composton s n good agreement wth the composton reported by Jan and Sharma [7]. The physo-chemcal propertes of JCB are reported n Table 7. The propertes of JCB were found to be n good accord wth ASTM and BIS specfcatons and wth the work of Rashd et al. [14]. Table 6: Fatty acd composton of JCB S.no Fatty acd Formula % Composton 1 Palmtc acd C 16 H 3 O 16. CH 3 (CH ) 14 COOH Stearc acd C 18 H 36 O 8. CH 3 (CH ) 16 COOH 3 Olec acd C 18 H 34 O 38.4 CH 3 (CH ) 7 -CH=CH-(CH ) 7 COOH 4 Lnolec acd C 18 H 3 O 36.8 CH 3 (CH ) 4 CH=CH-CH -CH=CH- (CH ) 7 COOH 5 Lnolenc acd C 18 H 30 O CH 3 (CH ) 4 CH=CH-CH -CH=CH-CH - CH=CH-(CH ) 4 COOH 0.4 8

7 Table 7: Physo-chemcal propertes of JCB S.no Property (unt) ASTM D6751 IS JCB ASTM D6751 lmts IS lmts 1 Vscosty (cst; 40 C) ASTM D445 IS Densty (g/c.c at 15 C) ASTM D405 IS Flash pont ( o C) ASTM D93 IS Mn Ester content (%) - EN Mn Adulteraton of JCB wth Kerosene The JCB prepared under optmum condtons was adulterated wth kerosene n dfferent proportons and change n the propertes of JCB was observed wth respect to kerosene. The varous propertes that were analyzed are gven as follows: Vscosty of Blends: Fgure 3 shows the plot of vscostes of bodesel blends. The plot can be used as a calbraton curve to fnd out the extent of adulteraton of kerosene n bodesel on the bass of vscosty. The fgure ndcates that the vscosty of bodesel blends decreases from KB 100 to KB 10 due to lower vscosty of kerosene than bodesel. Based on the curve shown n Fgure 3, a correlaton (Equaton (4)) s developed to measure the vscosty of any adulterated sample of bodesel blend. The correlaton can be used to keep a check on the mal practce of adulteraton by fndng out the extent of blendng of bodesel wth kerosene. The value of the regresson coeffcent (R ) s qute hgh for equaton 4, thus, ndcatng the valdty of the correlaton developed. v b ; R (4) where; v Vscosty of blends n cst Fgure 3: Vscosty of bodesel blends wth kerosene 3.4. Densty of Blends Fgure 4 shows the plot of densty of bodesel blends that can be used as calbraton curve to know the densty of unknown adulterated bodesel sample. The densty of bodesel blends also decreases from KB 100 to KB 10. Ths s agan due to lower densty of kerosene than bodesel. Based on the curve shown n Fgure 4, a correlaton (equaton 5) s developed to measure the densty of any unknown adulterated sample of bodesel blend. The value of the regresson coeffcent s qute hgh for correlaton, thus, ndcatng the valdty of the correlaton developed b e ; R (5) where; ρ - densty of blends n g/c.c The present study has provded the optmum range of parameters to maxmze the JCB yeld to a hgh value of 98.3%. Moreover, the low reacton temperature condton and less reacton tme, that has been optmzed usng RSM, s a new work n the feld of transesterfcaton of JCO to produce bodesel. The esterfed ol wth reduced FFA of < 1% was subjected to base-catalyzed transesterfcaton to produce bodesel. A bodesel yeld of 98.3% has been acheved wth methanol/ol molar rato of 11:1 (w/w) usng NaOH as catalyst (1% w/w) n 110 mn at 55 o C. Further, the adulteraton of bodesel wth kerosene has been studed and correlatons have been developed to check the adulteraton of unknown samples of bodesel. 83

8 Fgure 4: Densty of bodesel blends wth kerosene Conclusons Optmzaton of transesterfcaton process of JCO was acheved by four-factoral CCD usng RSM n 54 expermental runs. A second-order model has been obtaned to predct the JCB yeld as a functon of process varables. A bodesel yeld of 98.3% has been acheved wth methanol/ol molar rato (11:1) usng NaOH as catalyst (1% w/w) n 110 mn tme at 55 o C temperature. The prepared JCB conformed to the ASTM and BIS specfcatons. On the bass of ANOVA; the catalyst concentraton, reacton tme and methanol/ol rato had a sgnfcant effect on JCB yeld. Based on the analyss of blends of bodesel wth kerosene by vscosty and densty measurement, curves and correlatons have been developed to check the extent of adulteraton of kerosene n unknown samples of bodesel. The study postulates that the regresson coeffcent (R ) of correlatons has a value above 0.93 ndcatng ther usefulness n curbng the mal practce of bodesel adulteraton. Ths study wll be helpful n characterzng the bodesel for ts adulteraton wth kerosene. Ths would lead to customer acceptance, standardzaton and qualty assurance of bodesel and ts blends n the market. The model for predctng the bodesel yeld can be successfully employed n the bodesel ndustry to maxmze the yeld of methyl esters. Acknowledgements The authors acknowledge the fnancal support from the Mnstry of Human Resource Development (MHRD), Government of Inda, n the form of a scholarshp to carry out ths work. References 1. Fazal, M.A., Haseeb, A.S.M.A., Masjuk, H.H. Renew. Sust. Energ. Rev. 15 (011) Syam, A.M., Yunus, R., Ghaz, T.I.M., Choong, T.S.Y. Ind. Crop. Prod. 37 (01) Jan, S., Sharma, M.P. Boresour. Technol. 101 (010) Jeong, G.T., Park, D.H. Appl. Bochem. Botechnol.156 (009) Boonmee, K., Chuntranuluck, S., Punsuvonand, V., Slayo, P. Kasetsart. J. (Nat. Sc.). 44 (010) Twar, A.K., Kumar, A., Raheman, H. Bomass. Boenerg. 31 (007) Jan, S., Sharma, M.P. Energ. Fuel. 5(3) (011) Yuan, X., Lu, J., Zeng, G., Sh, J., Tong, J., Huang, G. Renew. Energ. 33 (008) Ghadge, S.V., Raheman, H., Boresour. Technol. 97 (006) Rashd, U., Anwar, F., Ansar, T.M., Arf, M., Ahmad, M., J. Chem. Technol. Botechnol. 84 (009) Shandlya, P., Jan, P.K., Jan, N.K. Inter. J. Eng. Sc. Techn. 3 (011) Leung, D.Y.C., Wu, X., Leung, M.K.H. Appl. Energ. 87 (010) Freedman, B., Pryde, E.H., Mounts, T.L. J. Am. Ol Chem. Soc. 61 (1984) Rashd, U., Anwar, F., Jaml, A., Bhatt, N.H. Pakstan. J. Bot. 4(1) (010) (013); 84