Hindawi Publishing CorporationThe Scientific World JournalVolume 2013 Article ID 268385 7 pageshttpdxdoiorg1011552013268385

Research ArticleDynamic Modeling of Reversible Methanolysis ofJatropha curcas Oil to Biodiesel

Azhari M Syam12 Hamidah A Hamid1 Robiah Yunus1 and Umer Rashid1

1 Institute of Advanced Technology Universiti Putra Malaysia (UPM) 43400 Serdang Selangor Malaysia2 Department of Chemical Engineering Malikussaleh University Lhokseumawe 24351 Indonesia

Correspondence should be addressed to Robiah Yunus robiahengupmedumy

Received 4 September 2013 Accepted 16 October 2013

Academic Editors L Li and Z-W Liu

Copyright copy 2013 Azhari M Syam et alThis is an open access article distributed under theCreative CommonsAttribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Many kinetics studies on methanolysis assumed the reactions to be irreversible The aim of the present work was to study thedynamic modeling of reversible methanolysis of Jatropha curcas oil (JCO) to biodiesel The experimental data were collected underthe optimal reaction conditions molar ratio of methanol to JCO at 6 1 reaction temperature of 60∘C 60min of reaction time and1ww of catalyst concentrationThe dynamic modeling involved the derivation of differential equations for rates of three stepwisereactions The simulation study was then performed on the resulting equations using MATLAB The newly developed reversiblemodels were fitted with various rate constants and compared with the experimental data for fitting purposes In addition analysisof variance was done statistically to evaluate the adequacy and quality of model parameters The kinetics study revealed that thereverse reactions were significantly slower than forward reactions The activation energies ranged from 65 to 444 KJ molminus1

1 Introduction

Methanolysis is the most common and cheapest biodieselproduction route In general this process may be carried outsimultaneously by employing either homogeneous catalystssuch as alkalis [1ndash3] acids [4 5] and enzymes [6 7] or het-erogeneous catalysts [8ndash10] Borges and Dıaz [11] have alsoperformed the reaction using heterogeneous catalyst Thehomogeneous alkali catalysts often suffer problems such asdifficulty in removing the catalyst after the reaction processutilization of large amount of water for purification andemulsification as well [12]

Theoretically the methanolysis process takes place underthree consecutive reactions as shown in Figure 1 Firstlytriglyceridemolecule reactswithmethanol to producemethylester and intermediate molecule of diglyceride Then diglyc-eride reacts again with methanol and produces intermediatemolecule of monoglyceride and methyl ester Finally mono-glyceride reacts with methanol to produce the final productglycerol and methyl ester Thus based on the stoichiometricequation the overall reaction requires one mole of triglyc-eride molecule to react with three moles of methanol to pro-duce one mole of glycerol and three moles of methyl esters

Cheng et al [13] reported that the reactions were mainlyendothermic in nature

Themethanolysis reaction takes place very fast in partic-ular when the excess methanol is used to avoid the backwardreaction If the backward reaction happens the formationof product will take a longer time to complete because thereaction will never reach an equilibrium condition For thesereasons stepwise reactions cannot go on under a normal reac-tion mechanism Sometimes the formation of intermediateproducts can be detected by the analytical analysis such asgas chromatographyHowever inmost cases the intermediatecompound cannot be identified at all because of the nature ofsimultaneous reactions Therefore most of the kinetics stud-ies on methanolysis reaction assumed that the stepwise reac-tions proceed in an irreversible mode [14ndash17] thus requiringvery simple modeling approach Similar studies on modelingbiodiesel have been done also by other researchers [18ndash20]those of which focused particularly on the kinetics of homo-geneous alkali catalyzed methanolysis in batch reactor Thusthe existing kinetics models are only applicable under certainprescribed conditions that support irreversible assumptionHowever with the advent of technology in numerical compu-tation visualization and programming this assumption is no

2 The Scientific World Journal

Consecutive reactions of methanolysis

Overall reaction of methanolysis

H2C-OH

CH-OCOR2

H2C-OCOR3

CH3OH

H2C-OH

CH-OH

H2C-OCOR3

R2-OCOCH3

Diglyceride Methanol Monoglyceride Methyl ester

H2C-OH

CH-OH

H2C-OCOR3

CH3OH

H2C-OH

CH-OH

H2C-OH

R3-OCOCH3

Monoglyceride Methanol Glycerol Methyl ester

H2C-OCOR1

CH-OCOR2

H2C-OCOR3

CH3OH

H2C-OH

CH-OCOR2

H2C-OCOR3

R1-OCOCH3

Triglyceride Methanol Diglyceride Methyl ester

H2C-OCOR1

CH-OCOR2

H2C-OCOR3

3CH3OH

H2C-OH

CH-OH

H2C-OH

R2-OCOCH3

R1-OCOCH3

R3-OCOCH3

Triglyceride Methanol Glycerol Methyl esters

OHminus

k2f

k5r

k3f

k6r

k1f

k4r

+ +

+ +

+ +

+ +

Figure 1 Mechanisms of consecutive and overall reaction ofmethanolysis

longer necessary Using MATLAB the complicated kineticsequations for the reversible reactions can be solved usingmultiple approaches and can reach a solution faster Usingthese models both forward and backward reaction ratesan Arrhenius activation energy can be determined at everyreaction temperature

This paper focuses on modeling of methanolysis processusing Jatropha curcas oil (JCO) as feedstockMost researchershad developed some models on biodiesel production as wellas biodiesel applications however literatures on modeling ofbiodiesel synthesis using JCO are still limited Another studyrelated tomodeling of biodiesel production process in partic-ular about the separation method of by-product (glycerol) toobtain a standardized biodiesel quality has been developed[21] In terms of reactor configuration the system has beenmodeled and simulated by Cheng et al [13] who used amembrane reactor and prereactor Their findings reportedthat the prereactor may be needed as the initial stage forcarrying out a substantial part of the methanolysis process

Thus the subsequent membrane reactor was controlled at aparticular methanol to oil molar ratio and catalyst concentra-tion under certain temperature condition Another proposedmodel was the work done by Brasio et al [22] for biodieselproduction using mechanical stirring reactor In view of theeconomic perspective of biodiesel production Hassounehet al [23] carried out the sensitivity study on the relationshipsof biodiesel with feedstock price through nonparametric andparametric modeling Modeling on algae-based feedstockand biocatalyst was conducted as well such as the studies onmass balance of nutrient model [24] and dynamic modelingof immobilized lipase catalyzed biodiesel production [25] Inrelation to biodiesel characteristics Gopinath et al [26] haveestablished the modeling of biodiesel properties with respectto its iodine value and saponification value

To date the kinetics study on methanolysis to producebiodiesel has been focused on irreversible reactions [16 17]However no systematic study on reversible methanolysis iscurrently available in the literature This drawback leads todifficulties in the explanation of the intermediate productsformed from the methanolysis reaction To overcome thisissue the aim of this work was to look into the mathemat-ical equations regarding the reversible kinetics in order toproduce a valid reversible model of methanolysis reactionThe modeling of the reversible reaction of methanolysis wasperformed using simulation software MATLAB and aimedto evaluate its kinetics model A statistical analysis of exper-imental data was also conducted to determine the accu-racy of the developed models as compared to that of theexperimental data An optimization routine was then used toadjust the parameters to obtain a good fit to the experimentalmeasurements The common approach is to minimize a sumof square errorsThemultiple range test included in the statis-tical programwas used to prove the existence of homogenousgroups within each of the parameters

2 Experimental

21 Methanolysis The samples were analyzed to determinethe conversion of triglyceride into intermediate and productcompounds using gas chromatography (GC) based on amethod described in our earlier study [27] Approximately003 plusmn 0005mL of sample was taken in a 20mL sample vialand diluted with 10mL of ethyl acetate and with 05mL ofBSTFAThe vial was then heated in a water bath for 15min at40∘C The sample was allowed to cool at room temperatureand it was then injected into the GC system The capillarycolumn (SGEHT5) 12m times 053mm and 015 120583m IDwas usedin the GC systemwith hydrogen as carrier gas at a flow rate of267mL minminus1 and a split ratio of 1 1 The oven temperaturewas set initially at 80∘C and was maintained for 3min thenwas increased at 6∘C minminus1 up to 340∘C and finally main-tained at 340∘C for another 6min The injector and detectortemperatures were set at 300∘C and 360∘C respectively Thefatty acid composition of JCO is shown in Table 1

22 Modeling To describe the model of biodiesel productionusingmethanolysis process themathematical equations were

The Scientific World Journal 3

Table 1 Fatty acid composition of JCO

Fatty acids Formula C D (wt)Myristic C14H28O2 14 0 01Palmitic C16H32O2 16 0 139Palmitoleic C16H30O2 16 1 08Margaric C17H34O2 17 0 01Stearic C18H36O2 18 0 77Oleic C18H34O2 18 1 480Linoleic C18H32O2 18 2 287Linolenic C18H30O2 18 3 01Arachidic C20H40O2 20 0 03Gadoleic C20H38O2 20 1 01Behenic C22H44O2 22 0 02C carbon number Y double bonds number

derived in order to produce a valid reversible model ofkinetics reaction Derivedmechanisms of JCOwere based onmethanolysis as shown in the following

119889 (119862JTG)

119889119905= minus1198961119891

(119862JTG) (119862MeOH) + 1198964119903 (119862JDG) (119862JME) (1)

119889 (119862JDG)

119889119905= + 119896

1119891

(119862JTG) (119862MeOH) minus 1198964119903 (119862JDG) (119862JME)

minus 1198962119891

(119862JDG) (119862MeOH) + 1198965119903 (119862JMG) (119862JME)

(2)

119889 (119862JMG)

119889119905= + 119896

2119891

(119862JDG) (119862MeOH) minus 1198965119903 (119862JMG) (119862JME)

minus 1198963119891

(119862JMG) (119862MeOH) + 1198966119903 (119862GL) (119862JME)

(3)

119889 (119862MeOH)

119889119905= minus 119896

1119891

(119862JTG) (119862MeOH) + 1198964119903 (119862JDG) (119862JME)

minus 1198962119891(119862JDG) (119862MeOH) + 1198965119903 (119862JMG) (119862JME)

minus 1198963119891

(119862JMG) (119862MeOH) + 1198966119903 (119862GL) (119862JME)

(4)

119889 (119862JME)

119889119905= + 119896

1119891(119862JTG) (119862MeOH) minus 1198964119903 (119862JDG) (119862JME)

+ 1198962119891

(119862JDG) (119862MeOH) minus 1198965119903 (119862JMG) (119862JME)

+ 1198963119891

(119862JMG) (119862MeOH) minus 1198966119903 (119862GL) (119862JME)

(5)

119889 (119862GL)

119889119905= +1198963119891(119862JMG) (119862MeOH) minus 1198966119903 (119862GL) (119862JME) (6)

where 119862 is the concentration of components such as Jat-ropha curcas triglyceride (JTG) Jatropha curcas diglyceride(JDG) Jatropha curcasmonoglyceride (JMG) glycerol (GL)

Jatropha curcas methyl ester (JME) and methanol (MeOH)and 119896 is the effective rate constant of each stepwise reaction

The resulting kinetics equations in terms of weightfraction (1)ndash(6) can be solved by implementing a numericalsolution via MATLAB version 77 For this study the ODEsolver function ode45 was selected because it employednonstiff solutions of the Runge-Kutta method of orders 4and 5 The rate constants for the three stepwise reactionswere determined from themaxima points of the intermediateproducts and the concentrations of the reaction products atfinal equilibrium The reaction products concentrations thatfit the product distribution data of concentration versus timefor JTG JDG JMG and methyl ester were obtained from theexperiment

23 Analysis of Variance The application of an analysisof variance (ANOVA) estimates any statistically significantdifferences based on the confidence level of 95 (Vega-Galvez et al [29]) The accuracy of the models used tocorrelate the experimental data was evaluated by means ofstatistical tests such as sum of square error (SSE) root meansquare error (RMSE) and Chi-square (1205942) as shown in (7)The lowest values of SSE RMSE and 1205942 are selected asoptimization criteria in order to evaluate the accuracy of themodels

SSE = 1119873

119873

sum

119895=1

(119910119890119895minus 119910119898119895)2

RMSE = [

[

1

119873

119873

sum

119895=1

(119910119898119895minus 119910119890119895

)2

]

]

12

1205942

=

sum119873

119895=1

(119910119890119895minus 119910119898119895)2

119873 minus 119886

(7)

where119873 119910119890119895 119910119898119895 and 119886 denote number of data weight per-

centage of glycerides by experiment and weight percentageof glycerides by model and number of constants

3 Results and Discussion

The results of analysis by gas chromatography (GC) detectedpeaks of JTG JDG and JME as shown in Figure 2 Thesimulation curves showed that the formation of methyl esters(biodiesel) was very slow at the beginning and reached theequilibrium value after only 25min of reaction As the for-mation of methyl esters was accompanied by the breakdownof JTG to JDG there was a sharp drop in concentration ofJTG in the first 10min of reaction The formation of JMEand breakdown of JTG to JDG were observed immediatelyafter 2min which indicated that the progressed reaction wasvery fast Consequently the formation and conversion of JDGwere also very fast at the beginning thus reaching a plateauand slowing down as the reaction progresses with time

Figure 2 depicted the breakdown of triglyceride slowingdown sharply after 10min of reaction This phenomenon can

4 The Scientific World Journal

0

01

02

03

04

05

06

07

08

09

0 10 20 30 40 50 60

Mas

s fra

ctio

n of

com

poun

ds (w

w)

JDG expJME expJTG exp

Time (min)

Figure 2 Experimental results from methanolysis process

0 10 20 30 40 50 600

01

02

03

04

05

06

07

08

09

Time (min)

Mas

s fra

ctio

n of

com

poun

ds (w

w)

TGDGMGGL

METG expDG expME exp

Figure 3 Comparison between experimental and simulation ofmethanolysis process

be explained later based on the kinetics data The breakdownof JDG to JMG could not be detected thus only the modelingresult was presented in Figure 3 Likewise glycerol couldnot also be detected by GC After 25min the JTG peakdisappeared and the amount of JME stabilized immediatelyafter that

The experimental weight fractions of each reaction com-ponent obtained from the GC analysis were fitted with thekinetic models using the numerical method via Runge-Kuttafourth- and fifth-order method [30]The rate constants of theproposed kinetics models were determined by minimizationof errors based on the optimum criteria of statistical analysisand by comparing the component concentrations at maxi-mum and equilibrium The simulated concentration profile

Table 2 Rate constant 119896 for reversible reaction model at 60∘C

Rate constant minminus1 Values Yunus and Syam (2011) [28]1198961119891

79 times 10minus2

15 times 10minus1

1198962119891

15 times 10minus1

18 times 10minus1

1198963119891

17 times 10minus1 mdash

1198964119903

56 times 10minus4 mdash

1198965119903

75 times 10minus4 mdash

1198966119903

14 times 10minus3 mdash

curves for the kinetics models were then plotted againstthe experimental values as shown in Figure 3 The dottedpoints represented experimental data and the solid curvescorresponded to the simulated results Therefore the kineticmodels developed by considering the three reactions indi-vidually as indicated by (1)ndash(6) would be able to describethe progress of the reaction better than the overall reaction(1) After plotting all experimental data matched with thesimulation results This fact was supported by the statisticalanalysis of the reaction

The irreversible kinetics model published by Yunus andSyam [28] predicted that the 119896 values are significantly higherthan the 119896 values from the present study (Table 2) FromTable 2 it can also be seen that the rate constants (119896) alsoincreased as the reactions progressed through the three step-wise reactions The rate constants for the forward reactionswere considerably higher than those for the reverse reactionsThis phenomenon also showed that the rate of first stepwisereaction took place slower than that of the last two reactionsThis confirms that the reactionsmust form an initial complexof compounds before proceeding to form the tetrahedralintermediate with low activation energy consumption Atransition state was required by reactions to break the com-plex of compounds to be tetrahedral intermediate and thefinal product Furthermore the results also confirmed thatthe assumption of irreversible reactions for methanolysis ofJCO is satisfactory for all conditions

Figure 4 demonstrated the concentration profiles fordiglyceride (JDG) andmethyl ester (JME) calculated from thefirst- order model at 60∘C using the 119896 values given in Table 2which was then compared against the experimental values Itwas observed that the concentration profile correlated usingthe current model with MATLAB was fitted well with theexperimental data and with the correlation coefficient 1198772 at099 After optimum curve fitting the average values of rateconstants 119896

1119891and 119896

4119903from the current kinetics models were

79 times 10minus2 and 56 times 10minus4 with levels of uncertainty at 00005and 00006 respectivelyOn the contrary the concentrationprofile for ME was obtained using irreversible model [28]which did not correlate well with the experimental valuesConsequently the resultant correlation coefficient for theintegral model was only 097 significantly lower than thevalue obtained from the current kineticmodel (099)The rateconstant 119896

2119891value for the irreversible model was 18 times 10minus1

with a level of uncertainty of 001 as compared to 0150from the current model Because the earlier method assumedan irreversible reaction 119896

2119903is zero whereas 119896

5119903from the

The Scientific World Journal 5

0

002

004

006

008

01

012

014

016M

ass f

ract

ion

of co

mpo

unds

(ww

)

JDG expJDG model

0 10 20 30 40 50 60Time (min)

(a)

0010203040506070809

JME expJME model

Mas

s fra

ctio

n of

com

poun

ds (w

w)

0 10 20 30 40 50 60Time (min)

(b)

Figure 4 Plot of experimental data and model for (a) JDG and (b) JME compounds

Table 3 Values of statistical parameters for JDG and JME

Parameters JDG JMESSE 43 times 10

minus6

47 times 10minus4

RMSE 21 times 10minus3

22 times 10minus2

1205942

53 times 10minus6

52 times 10minus5

currentmodel is 75times 10minus4This indicated that the assumptionof an irreversible reaction in the integral model is reasonablebecause the value of 119896

5119903was close to zero

The average optimum values of 1198963119891

and 1198966119903were found to

be 17 times 10minus1 and 14 times 10minus3 with the levels of uncertainty at00003 and 0004 respectively This confirmed the accu-racy of the current kinetics model and also the rate constantsdetermined from this work However Table 2 showed theevidence of a reversible reaction as indicated by the value of1198966119903from the currentmodelThe value of 119896

6119903was slightly larger

than that of 1198965119903This confirmed the significance of the reverse

reaction The reverse reaction also led to an accumulationof JDG and a small presence of JMG at the end of thereaction This problem had been successfully addressed bythe current model which enabled the predication of the JMGconcentration profile for the entire 1 h reaction period

The statistical experimental design has been proposedto allow for adequate variation of operation variables andprocess responses allowing for more precise estimation ofmodel parameters and identification of experimental effectsIn this context the use of differential methods in order toavoid the numerical integration of balance equations shall bejustified Most kinetics studies did not take into account thevariance of the measured values and in this matter statisticaltests cannot be applied properly to evaluate the modeladequacy and the quality of model parameters Additionalevidence supported that the validity of the current kineticsmodels was the statistical evaluation of the models based

minus4

minus35

minus3

minus25

minus2

minus15

minus1

minus05

0295 3 305 31

JTG forward JDG forwardJMG forwardJTG reverse

JDG reverse JMG reverse

logk

1T times 103 (K)minus1

Figure 5 Arrhenius activation energy at various temperatures andrate constants

on curve fitting as depicted in Table 3 The model exhibitedgood statistical correlation with a sum of squares error (SSE)of 43 times 10minus6 RMSE of 21 times 10minus3 and Chi-square (1205942) of53 times 10minus6 compared to the experimental JDG values Thecorrelation coefficient 1198772 was relatively higher at 098 Theresults indicated that the mean values of statistical analysesapplied to the kinetic models for the reversible reactionsexpressed a good fit with values close to zero for all statisticaltests as shown in Table 3 In general the statistical analysisindicated that the current model improved the estimation ofthe concentration profiles of components for the methanoly-sis process

Figure 5 showed the plot of Arrhenius activation energyunder several reaction rate constants and reaction temper-atures The values of activation energy for the reversible

6 The Scientific World Journal

Table 4 Activation energies for reversible reaction of methanolysis

Scenario of reaction Rate constants Activation energies (KJmolminus1) 1199032

JTG + MeOH 997888rarr JDG + JME 1198961119891

65 091MeOH + JTG JME + JDG 119896

4119903

395 092JDG + MeOH 997888rarr JMG + JME 119896

2119891

61 092MeOH + JDG JME + JMG 119896

5119903

24 097JMG +MeOH 997888rarr GL + JME 119896

3119891

24 092MeOH + JMG JME + GL 119896

6119903

444 097

reaction of methanolysis were found to be within the rangeof 65ndash444 KJmolminus1 as presented in Table 4 The activationenergies of forward reactions are generally lower than thoseof the reverse reactions thus the forward reactions are morefavored Based on the underlying principle of activationenergy the mechanism of reversible reaction can be under-stood in greater detail For the forward reaction betweenJCO andmethanol the reaction should acquire the activationenergy level before the reactants can form the activatedcomplex

Similarly for the reverse reaction glycerol and methylester must also acquire the activation energies level prior toforming the activated complex These trends showed that theforward reactions required less energy to form the complexas compared to the reverse reactions which needmore energyto form the activated complex As a result the rate of forwardreactions proceeds faster than the rate of reverse reactions dueto the lower level of activation energy

4 Conclusion

In this study the dynamic modeling of reversible methanol-ysis reaction for JCO using MATLAB was successfully per-formed The kinetics models incorporated reversible reac-tions which enable better prediction of reaction rate param-eters The results indicated that the homogenous alkaline-catalyzedmethanolysis reaction took place relatively fastTheforward reactions occurred at a much faster rate than thereverse reactions The rate of breakdown of JTG into JDGwas slower than the breakdown of JDG into JMG and finallythe formation of methyl ester These were evidenced fromthe values of the rate constants for the individual stepwisereactions The activation energies for the reverse reactionswere generally higher than those for the forward reactionswhich indicated that the reverse reactions were more difficultto take place than the forward reaction In addition the ratesof reverse reaction were also significantly slower than thoseof the forward reactionThe statistical analysis demonstratedthat the models correlated well with the experimental dataThe values of statistical parameters were very close to zerowhich demonstrated that the kinetics models developedusing MATLAB and its corresponding parameters werereliable and accurate

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgment

The authors acknowledge the Higher Education Ministry ofMalaysia for the financial support through the FRGS Scheme(via vot 5523422)

References

[1] A M Syam R Yunus T I M Ghazi and T C S Yaw ldquoMeth-anolysis of jatropha oil in the presence of potassium hydroxidecatalystrdquo Journal of Applied Sciences vol 9 no 17 pp 3161ndash31652009

[2] J M Dias M C M Alvim-Ferraz and M F Almeida ldquoCom-parison of the performance of different homogeneous alkalicatalysts during transesterification of waste and virgin oils andevaluation of biodiesel qualityrdquo Fuel vol 87 no 17-18 pp 3572ndash3578 2008

[3] AM Syam R Yunus T I M Ghazi and T S Y Choong ldquoSyn-thesis of Jatropha curcas oil-based biodiesel in a pulsed loopreactorrdquo Industrial Crops and Products vol 37 no 1 pp 514ndash519 2012

[4] E Lotero Y Liu D E Lopez K Suwannakarn D A Bruceand J G Goodwin Jr ldquoSynthesis of biodiesel via acid catalysisrdquoIndustrial and Engineering Chemistry Research vol 44 no 14pp 5353ndash5363 2005

[5] J MMarchetti V UMiguel and A F Errazu ldquoPossiblemetho-ds for biodiesel productionrdquo Renewable and Sustainable EnergyReviews vol 11 no 6 pp 1300ndash1311 2007

[6] N R Sonare and V K Rathod ldquoTransesterification of usedsunflower oil using immobilized enzymerdquo Journal of MolecularCatalysis B vol 66 no 1-2 pp 142ndash147 2010

[7] Y Liu H-L Xin and Y-J Yan ldquoPhysicochemical propertiesof stillingia oil feasibility for biodiesel production by enzymetransesterificationrdquo Industrial Crops and Products vol 30 no 3pp 431ndash436 2009

[8] O Ilgen ldquoDolomite as a heterogeneous catalyst for transesteri-fication of canola oilrdquo Fuel Processing Technology vol 92 no 3pp 452ndash455 2011

[9] E Li Z P Xu and V Rudolph ldquoMgCoAl-LDH derived hetero-geneous catalysts for the ethanol transesterification of canolaoil to biodieselrdquo Applied Catalysis B vol 88 no 1-2 pp 42ndash492009

[10] Y H Taufiq-Yap H V Lee R Yunus and J C Juan ldquoTranses-terification of non-edible Jatropha curcas oil to biodiesel usingbinary Ca-Mg mixed oxide catalyst effect of stoichiometriccompositionrdquo Chemical Engineering Journal vol 178 pp 342ndash347 2011

[11] M E Borges and L Dıaz ldquoRecent developments on hetero-geneous catalysts for biodiesel production by oil esterification

6 The Scientific World Journal

Table 4 Activation energies for reversible reaction of methanolysis

Scenario of reaction Rate constants Activation energies (KJmolminus1) 1199032

JTG + MeOH 997888rarr JDG + JME 1198961119891

65 091MeOH + JTG JME + JDG 119896

4119903

395 092JDG + MeOH 997888rarr JMG + JME 119896

2119891

61 092MeOH + JDG JME + JMG 119896

5119903

24 097JMG +MeOH 997888rarr GL + JME 119896

3119891

24 092MeOH + JMG JME + GL 119896

6119903

444 097

reaction of methanolysis were found to be within the rangeof 65ndash444 KJmolminus1 as presented in Table 4 The activationenergies of forward reactions are generally lower than thoseof the reverse reactions thus the forward reactions are morefavored Based on the underlying principle of activationenergy the mechanism of reversible reaction can be under-stood in greater detail For the forward reaction betweenJCO andmethanol the reaction should acquire the activationenergy level before the reactants can form the activatedcomplex

Similarly for the reverse reaction glycerol and methylester must also acquire the activation energies level prior toforming the activated complex These trends showed that theforward reactions required less energy to form the complexas compared to the reverse reactions which needmore energyto form the activated complex As a result the rate of forwardreactions proceeds faster than the rate of reverse reactions dueto the lower level of activation energy

4 Conclusion

In this study the dynamic modeling of reversible methanol-ysis reaction for JCO using MATLAB was successfully per-formed The kinetics models incorporated reversible reac-tions which enable better prediction of reaction rate param-eters The results indicated that the homogenous alkaline-catalyzedmethanolysis reaction took place relatively fastTheforward reactions occurred at a much faster rate than thereverse reactions The rate of breakdown of JTG into JDGwas slower than the breakdown of JDG into JMG and finallythe formation of methyl ester These were evidenced fromthe values of the rate constants for the individual stepwisereactions The activation energies for the reverse reactionswere generally higher than those for the forward reactionswhich indicated that the reverse reactions were more difficultto take place than the forward reaction In addition the ratesof reverse reaction were also significantly slower than thoseof the forward reactionThe statistical analysis demonstratedthat the models correlated well with the experimental dataThe values of statistical parameters were very close to zerowhich demonstrated that the kinetics models developedusing MATLAB and its corresponding parameters werereliable and accurate

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgment

The authors acknowledge the Higher Education Ministry ofMalaysia for the financial support through the FRGS Scheme(via vot 5523422)

References

[1] A M Syam R Yunus T I M Ghazi and T C S Yaw ldquoMeth-anolysis of jatropha oil in the presence of potassium hydroxidecatalystrdquo Journal of Applied Sciences vol 9 no 17 pp 3161ndash31652009

[2] J M Dias M C M Alvim-Ferraz and M F Almeida ldquoCom-parison of the performance of different homogeneous alkalicatalysts during transesterification of waste and virgin oils andevaluation of biodiesel qualityrdquo Fuel vol 87 no 17-18 pp 3572ndash3578 2008

[3] AM Syam R Yunus T I M Ghazi and T S Y Choong ldquoSyn-thesis of Jatropha curcas oil-based biodiesel in a pulsed loopreactorrdquo Industrial Crops and Products vol 37 no 1 pp 514ndash519 2012

[4] E Lotero Y Liu D E Lopez K Suwannakarn D A Bruceand J G Goodwin Jr ldquoSynthesis of biodiesel via acid catalysisrdquoIndustrial and Engineering Chemistry Research vol 44 no 14pp 5353ndash5363 2005

[5] J MMarchetti V UMiguel and A F Errazu ldquoPossiblemetho-ds for biodiesel productionrdquo Renewable and Sustainable EnergyReviews vol 11 no 6 pp 1300ndash1311 2007

[6] N R Sonare and V K Rathod ldquoTransesterification of usedsunflower oil using immobilized enzymerdquo Journal of MolecularCatalysis B vol 66 no 1-2 pp 142ndash147 2010

[7] Y Liu H-L Xin and Y-J Yan ldquoPhysicochemical propertiesof stillingia oil feasibility for biodiesel production by enzymetransesterificationrdquo Industrial Crops and Products vol 30 no 3pp 431ndash436 2009

[8] O Ilgen ldquoDolomite as a heterogeneous catalyst for transesteri-fication of canola oilrdquo Fuel Processing Technology vol 92 no 3pp 452ndash455 2011

[9] E Li Z P Xu and V Rudolph ldquoMgCoAl-LDH derived hetero-geneous catalysts for the ethanol transesterification of canolaoil to biodieselrdquo Applied Catalysis B vol 88 no 1-2 pp 42ndash492009

[10] Y H Taufiq-Yap H V Lee R Yunus and J C Juan ldquoTranses-terification of non-edible Jatropha curcas oil to biodiesel usingbinary Ca-Mg mixed oxide catalyst effect of stoichiometriccompositionrdquo Chemical Engineering Journal vol 178 pp 342ndash347 2011

[11] M E Borges and L Dıaz ldquoRecent developments on hetero-geneous catalysts for biodiesel production by oil esterification

The Scientific World Journal 7

and transesterification reactions a reviewrdquo Renewable andSustainable Energy Reviews vol 16 no 5 pp 2839ndash2849 2012

[12] Z Helwani M R Othman N Aziz J Kim and W J N Fer-nando ldquoSolid heterogeneous catalysts for transesterification oftriglycerides with methanol a reviewrdquo Applied Catalysis A vol363 no 1-2 pp 1ndash10 2009

[13] L-H Cheng S-Y Yen Z-S Chen and J Chen ldquoModeling andsimulation of biodiesel production using a membrane reactorintegrated with a prereactorrdquo Chemical Engineering Science vol69 no 1 pp 81ndash92 2012

[14] D Darnoko andM Cheryan ldquoKinetics of palm oil transesterifi-cation in a batch reactorrdquo Journal of the American Oil ChemistsrsquoSociety vol 77 no 12 pp 1263ndash1267 2000

[15] G Vicente M Martınez and J Aracil ldquoKinetics of Brassicacarinata oil methanolysisrdquo Energy and Fuels vol 20 no 4 pp1722ndash1726 2006

[16] O S Stamenkovic Z B Todorovic M L Lazic V B Veljkovicand D U Skala ldquoKinetics of sunflower oil methanolysis at lowtemperaturesrdquo Bioresource Technology vol 99 pp 1131ndash11402008

[17] H J Berchmans K Morishita and T Takarada ldquoKinetic studyof methanolysis of Jatropha curcas-waste food oil mixturerdquoJournal of Chemical Engineering of Japan vol 43 no 8 pp 661ndash670 2010

[18] H Noureddini and D Zhu ldquoKinetics of transesterification ofsoybean oilrdquo Journal of the American Oil Chemistsrsquo Society vol74 no 11 pp 1457ndash1463 1997

[19] G Vicente M Martınez J Aracil and A Esteban ldquoKinetics ofsunflower oil methanolysisrdquo Industrial and Engineering Chem-istry Research vol 44 no 15 pp 5447ndash5454 2005

[20] H Chen and J Wang ldquoKinetics of KOH catalyzed trans-esterification of cottonseed oil for biodiesel productionrdquo Journalof Chemical Industry and Engineering vol 56 no 10 pp 1971ndash1974 2005

[21] A Abeynaike A J Sederman Y Khan M L Johns J F David-son and M R Mackley ldquoThe experimental measurement andmodeling of sedimentation and creaming for glycerolbiodieseldroplet dispersionsrdquo Chemical Engineering Science vol 79 pp125ndash137 2012

[22] A S R Brasio A Romanenko L O Santos and N C P Fer-nandes ldquoModeling the effect ofmixing in biodiesel productionrdquoBioresource Technology vol 102 no 11 pp 6508ndash6514 2011

[23] I Hassouneh T Serra B K Goodwin and J M Gil ldquoNon-parametric and parametricmodeling of biodiesel sunflower oiland crude oil price relationshipsrdquo Energy Economics vol 34 pp1507ndash1513 2012

[24] C Rosch J Skarka and N Wegerer ldquoMaterials flow modelingof nutrient recycling in biodiesel production from microalgaerdquoBioresource Technology vol 107 pp 191ndash199 2012

[25] S Al-Zuhair A Dowaidar and H Kamal ldquoDynamic modelingof biodiesel production from simulated waste cooking oil usingimmobilized lipaserdquo Biochemical Engineering Journal vol 44no 2-3 pp 256ndash262 2009

[26] A Gopinath S Puhan and G Nagarajan ldquoTheoretical model-ing of iodine value and saponification value of biodiesel fuelsfrom their fatty acid compositionrdquo Renewable Energy vol 34no 7 pp 1806ndash1811 2009

[27] R Yunus O T Lye A Fakhrursquol-Razi and S Basri ldquoA simplecapillary column GC method for analysis of palm oil-basedpolyol estersrdquo Journal of the American Oil Chemistsrsquo Society vol79 no 11 pp 1075ndash1080 2002

[28] R Yunus and A M Syam ldquoKinetics of the transesterification ofJatropha curcas triglyceride with an alcohol in the presence ofan alkaline catalystrdquo International Journal of Sustainable Energyvol 30 no 2 pp S175ndashS183 2011

[29] A Vega-Galvez E Notte-Cuello R Lemus-Mondaca L Zuraand M Miranda ldquoMathematical modelling of mass transferduring rehydration process of Aloe vera (Aloe barbadensisMiller)rdquo Food and Bioproducts Processing vol 87 no 4 pp 254ndash260 2009

[30] H Abd Hamid R Yunus and T S Y Choong ldquoUtilization ofMATLAB to simulate kinetics of transesterification of palm oil-based methyl esters with trimethylolpropane for biodegradablesynthetic lubricant synthesisrdquo Chemical Product and ProcessModeling vol 5 no 1 article 13 2010

Impact Factor 173028 Days Fast Track Peer ReviewAll Subject Areas of ScienceSubmit at httpwwwtswjcom

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawi Publishing Corporation httpwwwhindawicom Volume 2013

The Scientific World Journal