Synthesis and characterization of calcium methoxide as heterogeneous catalyst for trimethylolpropane esters conversion reaction

Page 1

Sf

Ha

b

c

a

ARRAA

KLHCT

1

btvo[cw

Mf

0d

Applied Catalysis A: General 425– 426 (2012) 184– 190

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General

jo u r n al hom epage: www.elsev ier .com/ locate /apcata

ynthesis and characterization of calcium methoxide as heterogeneous catalystor trimethylolpropane esters conversion reaction

assan Masooda, Robiah Yunusa,b,∗, Thomas S.Y. Choonga, Umer Rashidb, Yun H. Taufiq Yapc

Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, MalaysiaInstitute of Advanced Technology, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, MalaysiaDepartment of Chemistry, Faculty of Science, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

r t i c l e i n f o

rticle history:eceived 25 January 2012eceived in revised form 8 March 2012ccepted 12 March 2012vailable online 21 March 2012

eywords:ubricanteterogeneous catalystalcium methoxide

a b s t r a c t

Trimethylolpropane (TMP) esters are potential biodegradable basestock for biolubricant. In order to attainenvironmental benignity, attention has been focused on utilizing heterogeneous catalysts for productionof TMP esters. Alkaline homogeneous catalysts tend to react with free fatty acids to produce unwantedsoap, thus reducing the overall product yield. This study had focused on the synthesis of calcium methox-ide and investigating its potential as heterogeneous catalyst for the transesterification of TMP and palm oilmethyl esters (POME) to TMP esters. The performance of synthesized calcium methoxide as a catalyst wasexamined by characterizing it through some instrumental techniques. X-ray diffraction (XRD) showedcalcium methoxide has been successfully synthesized. Scanning electron microscopy (SEM) displayedthermally resistant surface structure with good porosity; BET showed high surface area; particle size

ransesterification analysis evidenced reasonable size of catalyst particles; and thermogravimetry (TGA) revealed good ther-mal stability of synthesized calcium methoxide. Moreover, the catalyst was found to possess mesoporoussurface by pore size analysis through Barrett–Joyner–Halenda (BJH) method. The results of transesteri-fication reaction indicated satisfactory catalytic activity of synthesized calcium methoxide and the TMPtriesters yield obtained was 80.35% after 2 h, 87.48% after 4 h, 91.30% after 6 h and 92.38% after 8 h reaction

H3

sters

time.

. Introduction

The concern and awareness of the harmful impact of mineralased lubricants on the environment have pushed the researchowards the production of environmental friendly lubricants fromegetable oils. Due to some performance limitations, vegetableils are not considered suitable to be used directly as lubricants1]; nonetheless, the limitations can be minimized by means ofhemical modification of vegetable oils through transesterificationith polyhydric alcohols [2]. In similar fashion, researchers have

CH2 – OH

CH3CH2 – C -CH2OH + 3 RCOO C

CH2 – OH

TMP Methyl E

∗ Corresponding author at: Institute of Advanced Technology, Universiti Putraalaysia, 43400 UPM Serdang, Selangor, Malaysia. Tel.: +60 3 89467531;

ax: +60 3 89467006.E-mail address: [email protected] (R. Yunus).

926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.oi:10.1016/j.apcata.2012.03.019

© 2012 Elsevier B.V. All rights reserved.

performed transesterification of vegetable oil methyl esters withtrimethylolpropane (TMP) to generate esters of TMP, which aregenerally utilized as base oil for various types of lubricating oils[3–9].

The transesterification reaction involves three consecutive reac-tions in the presence of a catalyst. Monoesters (ME) and diesters(DE) of TMP are formed as intermediate products towards the com-pletion of the reaction producing TMP triesters. The overall reactionstoichiometry requires one mole of TMP and three moles of methylesters. The overall reaction scheme is shown by Eq. (1).

(1)

CH 2OCOR

CH3CH2 – C – CH2OCOR + 3CH3OH

CH2OCOR

TMP tri esters Methan ol (1)

Catalyst holds a prominent role in the transesterification reac-tion of TMP and vegetable oil methyl esters, in order to obtain abetter yield of TMP triesters in less time. Reported catalysts for

the production of TMP esters include homogeneous bases such assodium methoxide [10], homogeneous acids such as sulfonic acid[11], enzymes such as lipases [4,5] and heterogeneous base cal-cium methoxide [6]. Alkaline homogeneous catalyst was highly

Page 2

is A: G

ceaaa[bedTtUepl

aowbpTmtpi

2

2

c2mBfpmT

2

1r5ptmf

TA

H. Masood et al. / Applied Catalys

orrosive in transesterification between TMP and palm oil methylsters (POME) under vacuum, and also tend to react with free fattycids to produce unwanted soap [10,12]. Homogeneous acid cat-lysts are difficult to recycle and operate at high temperatures,nd also give rise to serious environmental and corrosion problems13]. Lipases, used as enzymatic catalysts in the transesterificationetween rapeseed oil methyl esters and TMP, were reported toxperience problems with free fatty acids and methanol which hadecreased the operational ability of lipases in the reaction [5,14].he need of developing heterogeneous catalysts is gaining atten-ion due to non-toxic nature, ease of separation and reusability [15].nfortunately, heterogeneous calcium methoxide developed in anarlier study by direct reaction between calcium and methanolroved inadequate for the transesterification reaction due to pro-

onged reaction time [6].In previous literature, calcium methoxide has also been served

s a catalyst for transesterification reaction between methyl estersf olive oil and rapeseed oil with neopentyl glycol [6]. The catalystas also utilized for biodiesel production by methanolysis of soy-

ean oil [16,17] and tributyrine [18], and for synthesis of aliphaticolyesters by polymerization of �-caprolactone and l-lactide [19].he current study aims to report the detailed synthesis of calciumethoxide and its effectiveness as a heterogeneous catalyst in the

ransesterification of TMP with POME. The physical and chemicalroperties of synthesized calcium methoxide are also character-

zed.

. Experimental

.1. Chemicals

Precipitated calcium carbonate (analytical grade) was pur-hased from Systerm Chemicals Malaysia. Trimethylolpropane,-ethyl-2-hydroxymethyl-1,3-propanediol (purum grade) andethanol (GC grade) were obtained from Sigma Aldrich Sdn.

hd. (Malaysia). Palm oil methyl esters (POME) were obtainedrom Carotech Sdn. Bhd. Malaysia. The average fatty acid com-ositions of POME were determined according to the standardethods of oil and fats analysis [20]; the results are presented in

able 1.

.2. Synthesis of catalyst

An amount of 5 g calcium carbonate was calcined at 900 ◦C for.5 h under helium flow of 50 mL min−1. The calcined product waseacted with 100 mL methanol under reflux at 65 ◦C for 2 h with0 mL min−1 nitrogen flow. A continuous stirring of 700 rpm wasrovided to facilitate sufficient contact between the reagents. After

he reaction, the resultant precursor was filtered for the removal of

ost of the methanol and then dried in the vacuum oven at 105 ◦Cor 1 h.

able 1verage fatty acid compositions of palm oil-based methyl esters.

Fatty acid Composition (wt%)

Capric (C10:0) –Lauric (C12:0) 0.9Myristic (C14:0) 1.5Palmitic (C16:0) 41.5Stearic (C18:0) 2.7Oleic (C18:1) 40.6Linoleic (C18:2) 11.9Linolenic (C18:3) 0.3Others 0.2

eneral 425– 426 (2012) 184– 190 185

2.3. Characterization of catalyst

X-ray diffraction (XRD) was performed to examine the con-stituents of the synthesized catalyst on Shidmazu Model XRD-6000diffractometer having Cu tube and scintillation counter detector.The analysis was made over a 2� range from 5◦ to 40◦ along witha step size of 0.04◦ and scanning speed was set at 2 min−1. Mea-surement of IR spectrum was performed using attenuated totalreflection-Fourier transform-infrared (ATR-FTIR) on PerkinElmerSpectrum 100 FTIR spectrometer to know the surface groups exist-ing on catalyst surface. The catalyst sample in powder form wasanalyzed over the scanning range of 650–4000 cm−1 at a resolutionof 4 cm−1.

The BET surface area, total pore volume, pore diameter and poresize distribution were measured with BELsorp-mini II (BEL Japan,Inc.) automatic specific surface area/pore size distribution analyzer.A weighed sample of calcium methoxide was prepared for analysisby outgassing it at 150 ◦C for 6 h under vacuum through BELprep-vac II. The isotherms were generated by dosing nitrogen onto thecatalyst in a bath containing liquid nitrogen at 77 K. Equipment waspurged with helium gas during operation. The relative pressurerange for adsorption was set at 0.020–0.990 bar and for desorp-tion the range was set at 0.970–0.306 bar. BET specific area, totalpore volume and average pore diameter were evaluated from theBET plot, while the pore size distribution was evaluated by usingBarrett–Joyner–Halenda (BJH) plot. The particle size distributionwas measured using Malvern Mastersizer Hydro 2000MU laser par-ticle size analyzer. These analyses were useful in understanding thereactivity of catalyst in the reaction.

Scanning electron microscopy (SEM) was performed on Ther-moscientificInfravision Quasi-S (Scanning Electron Microscope,Model No.: S3400N, Hitachi Co., Japan) to observe the shape ofcatalyst particles. Energy-dispersive X-ray spectroscopy (EDX) wasconducted on ThermoscientificNoran System 7 with Nanotracedetector to know chemical composition of elements inside the cat-alyst. Finally, thermogravimetric and differential thermal analysis(TG–DTA) was carried out with Mettler Toledo TGA/SDTA851e inan air flow, with heating up to 800 ◦C at the rate of 10◦ min−1.

2.4. Transesterification reaction procedure

The transesterification reaction was carried out in a 500 mLflat bottom three-necked flask made of borosilicate glass (as thereactor) having a PTFE coated magnetic stirrer bar, a reflux con-denser, a mercury thermometer and a sampling port. A rotaryvane vacuum pump was coupled with the reflux condenser to pro-vide the necessary vacuum conditions, and the vacuum level wascontrolled by an air leakage valve located in between the refluxcondenser and the vacuum pump. An amount equal to 6.5 g TMPwas introduced into the reactor and was heated up to 100 ◦C. Thetemperature was maintained at 100 ◦C for 15 min with constantstirring at 1000 rpm under 20 mbar vacuum in order to removemoisture trapped by the hygroscopic TMP. At this point, 86.2 g ofmoisture free POME was added to the reactor and mixed with TMP.The molar ratio TMP:POME was 1:6 and the amount of TMP andPOME to be introduced in the reaction was calculated from theirrespective molecular weights accordingly. The purpose of usingexcess POME was to suppress the reverse reaction [21]. The reac-tion mixture was heated to 180 ◦C and 0.28 g of calcium methoxide(0.3%, w/w of reaction mixture) was added. Then, vacuum pressurewas changed to 50 mbar and temperature was maintained at 180 ◦C

for 8 h. Samples were collected through sampling port after every1 h and were analyzed by gas chromatography (GC) to know thecomposition of mono-, di-, and tri-esters of TMP in the reactionproduct.

Page 3

186 H. Masood et al. / Applied Catalysis A: General 425– 426 (2012) 184– 190

2

p[sNtaisaa6T3

3

3

abccravi[cacdob

C

Fig. 2. FTIR spectrum of calcium methoxide. (a) Peak assigned to CO stretching vibra-

part of surface is occupied by mesopores, having diameters withina range of 20–30 nm, which confirms mesoporous nature of thecatalyst. Mesoporous materials can serve as effective catalysts in

Fig. 1. XRD patterns of: (a) calcination product (b) reaction product.

.5. Analysis of product

The collected samples were analyzed by gas chromatogra-hy (GC) through a method described in an earlier research22]. Approximately 1 �L of sample was taken in a 2 mL auto-ampler vial and diluted with 1 mL ethyl acetate and 0.5 mL,O-bis (trimethylsilyl) trifluoroacetamide (BSTFA). The vial was

hen heated in a water bath for 10 min at 40 ◦C. The sample wasllowed to cool down at room temperature and it was then injectednto the GC system. The capillary column (SGE HT5) was used in GCystem with hydrogen as carrier gas at a flow rate of 26.7 mL min−1

nd a split ratio of 1:1. The oven temperature was set initiallyt 80 ◦C and was maintained for 3 min, then was increased at◦C min−1 up to 340 ◦C and was maintained for another 6 min.he injector and the detector temperatures were set at 300 ◦C and60 ◦C respectively.

. Results and discussion

.1. Characterizations of the synthesized catalyst

The XRD patterns of the substrates obtained after calcinationnd synthesis reaction are demonstrated in Fig. 1. The obviousroad peaks at 2� values of 32◦ and 37◦ in Fig. 1(a) are attributed toalcium oxide (JCPDS 37-1497) which was obtained after the cal-ination of calcium carbonate. After reacting with methanol undereflux, this calcium oxide was transformed into calcium methoxides evident from the characteristic peak of calcium methoxide at 2�alue of 11◦ in Fig. 1(b). Similar results were reported in earlier stud-es confirming the synthesized catalyst to be calcium methoxide23]. There was no indication regarding the presence of unreactedalcium oxide or calcium carbonate in the final product. However,n insignificant peak at 2� value of 18◦ hints the existence of cal-ium hydroxide (JCPDS 1-84-1264) which has possibly appearedue to the interaction of catalyst with water formed as a by-product

f the synthesis reaction. The reaction which has feasibly occurredetween calcium oxide and methanol is represented in Eq. (2).

aO + 2CH3OH65 ◦C−→Ca(OCH3)2 + H2O (2)

tion of methanol; (b) Peak resulting from CH bending; (c) Peaks derived from CH3

stretching vibrations; (d) Peak assigned to OH stretching vibration of primaryalcohol.

Fig. 2 shows the FTIR spectrum of synthesized catalyst atroom temperature. A distinct peak observed at around 1070 cm−1

was assigned to C O stretching vibration of primary alcohol,while another peak at around 3650 cm−1 was attributed to OHstretching vibration of primary alcohol [16]. Other peaks indicat-ing the important features were 2800–3000 cm−1 derived from CH3stretching vibrations and 1465 cm−1 concerning C H (alkane)bending [17]. Earlier researches suggest the peak appeared at3650 cm−1 due to adsorption deriving from hydration of the sur-face [24]; it was peculiar to calcium methoxide. This peak indicatesthe existence of OH groups isolated on calcium cation [25]. Sincewater was produced as a by-product in the catalyst synthesis reac-tion, these isolated OH groups might have produced from waterfacilitated by the strong basic property of calcium methoxide.

The reactivity of a catalyst is directly related with the surfacearea of the catalyst, therefore BET surface area is considered to bean important characterization of solid catalyst. The analyzed resultsfor BET surface area, total pore volume and average pore diameterof synthesized catalyst were found to be 38.6 m2 g−1, 0.24 cm3 g−1

and 25.7 nm respectively. Due to high BET surface area, the catalystappears favorable for use in the liquid phase reactions, and canprovide an enough reaction area in stirred reactor [23]. The poresize distribution of synthesized catalyst in Fig. 3 reveals that a large

Fig. 3. Pore size distribution of calcium methoxide.

Page 4

H. Masood et al. / Applied Catalysis A: General 425– 426 (2012) 184– 190 187

ibutio

td[

fiptFttr

t

Fig. 4. Particle size distr

ransesterification reactions for adsorbing large organic moleculesue to uniform pore structure and extensively high surface area26].

Particle size distribution can markedly affect the settling andltering characteristics of solid catalyst in a stirred slurry reactor;articles within a size range of 5–200 �m are generally favorable inhis regard [27]. The particle size distribution of catalyst is shown inig. 4, indicating that a large amount of catalyst particles are withinhe size range of 2–100 �m. These particles can be separated fromhe reaction products through filtration or centrifugation, after the

eaction.

The shape and topology of catalyst particles was observedhrough SEM, shown in Fig. 5. The surface appears to be formed

Fig. 5. SEM images of surface of calcium methoxide.

n of calcium methoxide.

out of the clusters of thin plates which were similar in shape tothose reported in an earlier research [23]. A large number of poresare visible on the surface. The shape of the particles gives no indi-cation of the chemical composition for constituents of catalyst, soEDX analysis was performed and the spectrum can be seen in Fig. 6.The mass% of calcium, oxygen and carbon on a particular area atthe surface of synthesized catalyst were determined to be 41.38%,52.79% and 5.82% respectively. The atomic ratio between calciumand oxygen (Ca:O) was 2.83, while the atomic ratio between cal-cium and carbon (Ca:C) was 3.04. Peaks at 2.1 keV and 2.9 keVrepresent the sputter coating of gold/palladium alloy, and the val-ues were excluded from the final results. The actual formula forcalcium methoxide is Ca(OCH3)2. The surface of catalyst has strongbasic character and therefore holds tendency to absorb moisture.The peaks attributed to OH on XRD and FTIR results explain thepresence of more oxygen atoms on surface. Furthermore, carbondioxide adsorbed within the pores on the basic catalyst surface hasinfluence on both Ca:O and Ca:C. Hydrogen was not quantified byEDX because of small orbital diameter (atomic number = 1), whichreduces catch probability for spectroscopy to almost zero.

In thermogravimetric analysis (TGA), the weight of synthesizedcatalyst was measured as the function of temperature while it wassubjected to a controlled heating program. The amount of weightloss provided a quantitative indication about the composition ofsample. In addition, material behavior was also investigated interms of thermal degradation and reaction with oxygen. Temper-ature of thermal and oxidative degradation was also measured.Differential thermal analysis (DTA) curve showed the physical andchemical transitions of the sample by measuring exothermic andendothermic effects.

Fig. 7 corresponds to the thermal decomposition of synthe-sized calcium methoxide by heating under air flow. The TGA curveremained almost constant from 25 ◦C to 396 ◦C, with the slightweight loss due to loss of moisture or volatile matters. After that,there was a sudden weight loss of around 4%, possibly due to achemical reaction, and an exothermic DTA peak at 450 ◦C furtherconfirms it. A small DTA peak at 405 ◦C can be attributed to theloss of water and removal of some adsorbed carbon dioxide. Thereaction at 450 ◦C might resulted in the decomposition of calciummethoxide with oxygen to calcium carbonate (Eq. (3)), and minorweight loss was observed due to very little difference in molecu-lar weights between calcium methoxide (102 g/mol) and calciumcarbonate (100 g/mol) [16,17].

Ca(OCH3)2 + 3O2 → CaCO3 + CO2 + H2O (3)

After 600 ◦C, the TGA curve had evidenced thermal decomposi-

tion of calcium carbonate with the formation of gaseous products.At 760 ◦C, the DTA peak showed endothermic transformation ofcalcium carbonate to stable calcium oxide, and 40% weight lossmatches with the stoichiometric values of the calcination reaction

Page 5

188 H. Masood et al. / Applied Catalysis A: General 425– 426 (2012) 184– 190

Fig. 6. EDX spectrum on surfac

Fig. 7. TG–DTA thermogram of calcium methoxide.

Fig. 8. Formation of M

e of calcium methoxide.

for calcium carbonate. The analysis suggested that the synthesizedcalcium methoxide was stable below 400 ◦C.

3.2. Reaction results

To study the effectiveness of calcium methoxide as a catalyst forproduction of trimethylolpropane triesters, the transesterificationreaction was conducted between TMP and POME by the methoddescribed earlier. The conditions selected for the reaction were180 ◦C temperature, 50 mbar vacuum, molar ratio TMP:POME 1:6,0.3% calcium methoxide by weight of reaction mixture and 8 h reac-tion time. The transesterification reaction involved the cleavage of

an ester group RCOO , from palm-based methyl esters, by an alco-hol group of TMP to produce new palm-based esters. Since there arethree alcohol groups in TMP, the process yielded intermediate for-mation of monoesters (ME), diesters (DE) and triesters (TE) of TMP.

E, DE and TE.

Page 6

H. Masood et al. / Applied Catalysis A: G

Fig. 9. Composition of ME, DE and TE obtained from transesterification betweenT

MrrpgtF

catalyst, while Oı− had extracted Hı+ and Caı+ adsorbed O CH3from alcohol. These two neighboring adsorbed species had then

Fc

MP and POME catalyzed with calcium methoxide.

ethanol was produced as a by-product and it was continuouslyemoved to ensure the completion of process aiding by forwardeaction. After eliminating the unreacted methyl esters, the finalroduct consisted mainly of TE, minor quantity of DE, and negli-ible amount of ME. These three consecutive–competitive steps of

ransesterification reaction between TMP and POME are shown inig. 8.

ig. 10. Possible mechanism for transesterification of POME with TMP catalyzed by calatalyst; (c) reaction between adsorbed neighboring species.

eneral 425– 426 (2012) 184– 190 189

Fig. 9 shows the composition of ME, DE and TE when calciummethoxide synthesized in this study was used in the transesterifi-cation reaction between TMP and POME. Thirty minutes after thestart of the reaction, the product contained 3.73% ME, 61.86% DEand 34.40% TE. Thereafter, the composition of TE started to increasewhile composition of DE and ME kept on decreasing. The compo-sition of TE reached 80.35% after 2 h and the increase in yield of TEwas less prominent afterwards. After 8 h reaction time, the productcontained 92.38% TE and the reaction was still in progress due to thepresence of unreacted ME i.e., 0.40%. In comparison with the earlierstudies, the reaction had generated a better yield of TE in less time[4–6]. However, the yield was lower as compared to Yunus et al.[10] who managed to get 98% triesters in 1 h by having better masstransfer in the reaction provided by homogeneous sodium methox-ide catalyst. The only setback was the formation of unwanted soapwith the use of homogenous catalyst.

The proposed mechanism of the catalysis process is shown inFig. 10. The mechanism is based on an earlier study where calciumethoxide showed a similar behavior in another transesterificationreaction as calcium methoxide did for the current reaction [27]. Itwas assumed that the reaction took place on the surface of catalystwith Ca and O being the two active catalytic sites participating in thereaction. Ester and alcohol were adsorbed on these two neighboringfree catalytic sites as indicated by step (a) and step (b) in Fig. 10.The adsorbed ester had formed an intermediate on the surface of

reacted with each other, in step (c), which led to the formationof ME of TMP and a molecule of methanol. The ME followed the

cium methoxide: (a) adsorption of ester on catalyst; (b) adsorption of alcohol on

Page 7

1 is A: G

smt

4

bCaXfhsfgBtme5c

A

ngo

R

[

[

[

[

[

[[

[[[

[

[

[

[

[[25] T. Iizuka, H. Hattori, Y. Ohno, J. Sohma, K. Tanabe, J. Catal. 22 (1971)

130–139.

90 H. Masood et al. / Applied Catalys

ame process on the surface of catalyst and reacted with anotherolecule of POME to form DE, which was further converted into

riester of TMP in a similar fashion.

. Conclusion

Developing an environmental benign process for production ofiolubricant using a solid base catalyst has become a global need.alcium methoxide was synthesized to serve as heterogeneous cat-lyst for production of TMP esters as biolubricant basestock. TheRD, FTIR and EDX results showed that the catalyst was success-

ully synthesized with sufficient purity; however, the surface hadydroxide functional group indicating the influence of water. Theynthesized catalyst was also found to acquire mesoporous sur-ace, high surface area, moderate particle size distribution, andood thermal stability as heterogeneous catalyst through SEM,ET, particle size analysis and TG–DTA characterizations. Finally,he experimental results showed good catalytic ability of calcium

ethoxide for the transesterification of TMP and POME to TMP tri-sters, with the experiment conducted under 180 ◦C temperature,0 mbar vacuum, molar ratio of TMP:POME at 1:6 and 0.3% (w/w)atalyst respective to reaction mixture.

cknowledgments

The authors are grateful to the Ministry of Science and Tech-ology, Malaysia, for the financial support under the Techno fundrant and Science Fund (vot: 5450511). The valuable contributionf Prof. Dr. Masato Kouzu for this work is also acknowledged.

eferences

[1] B. Wilson, Ind. Lubr. Tribol. 50 (1998) 6–15.[2] B.K. Sharma, A. Adhvaryu, Z. Liuand, S.Z. Erhan, J. Am. Oil Chem. Soc. 83 (2006)

129–136.

[

[

eneral 425– 426 (2012) 184– 190

[3] R. Yunus, A. Fakhrul-Razi, T.L. Ooi, S.E. Iyuke, A. Idris, J. Oil Palm Res. 15 (2003)42–49.

[4] Y.Y. Linko, T. Tervakangas, M. Lämsä, P. Linko, Biotechnol. Tech. 11 (1997)889–892.

[5] E. Uosukainen, Y.Y. Linko, M. Lämsä, T. Tervakangas, P. Linko, J. Am. Oil Chem.Soc. 75 (1998) 1557–1563.

[6] S. Gryglewicz, W. Piechocki, G. Gryglewicz, Bioresour. Technol. 87 (2003)35–39.

[7] G. Hillion, D. Proriol, OCL: Ol. Corps Gras Li. 10 (2003) 370–372.[8] S.Z. Sulaiman, A.L. Chuah, A. Fakhru’l-Razi, J. Appl. Sci. 7 (2007) 2002–

2005.[9] R.N.M. Kamil, S. Yusup, U. Rashid, Fuel 90 (2011) 2343–2345.10] R. Yunus, A. Fakhrul-Razi, T.L. Ooi, S.E. Iyuk, A. Idris, J. Oil Palm Res. 15 (2003)

35–41.11] V. Eychenne, Z. Mouloungui, A. Gaset, J. Am. Oil Chem. Soc. 75 (1998)

293–299.12] M. Di Serio, R. Tesser, M. Dimiccoli, F. Cammarota, M. Nastasi, E. Santacesaria,

J. Mol. Catal. A: Chem. 239 (2005) 111–115.13] E. Lotero, Y. Liu, D.E. Lopez, A. Suwannakaran, D.A. Bruce, J.G. Goodwin, Ind.

Eng. Chem. Res. 44 (2005) 5353–5363.14] M.H. Zong, Z.Q. Duan, W.Y. Lou, T.J. Smith, H. Wu, Green Chem. 9 (2007)

434–437.15] J.H. Clark, Acc. Chem. Res. 35 (2002) 791–797.16] M. Kouzu, J. Hidaka, K. Wakabayashi, M. Tsunomori, Appl. Catal. A: Gen. 390

(2010) 11–18.17] X. Liu, X. Piao, Y. Wang, S. Zhu, H. He, Fuel 87 (2008) 1076–1082.18] I.N. Martyanov, A. Sayari, Appl. Catal. A: Gen. 339 (2008) 45–52.19] Z. Zhong, M.J.K. Ankoné, P.J. Dijkstra, C. Birg, M. Westerhausen, J. Feijen, Polym.

Bull. 46 (2001) 51–57.20] W.L. Siew, T.S. Tang, Y.A. Tan, PORIM Test Method, PORIM, Bangi, 1995,

p. 181.21] R. Yunus, A. Fakhru’l-Razi, T.L. Ooi, D.R.A. Biak, S.E. Iyuke, J. Am. Oil Chem. Soc.

81 (2004) 497–503.22] R. Yunus, O.T. Lye, A. Fakhru’l-Razi, S. Basri, J. Am. Oil Chem. Soc. 79 (2002)

1075–1080.23] M. Kouzu, T. Kasuno, M. Tajika, S. Yamanaka, J. Hidaka, Appl. Catal. A 334 (2008)

357–365.24] H.D. Lutz, H. Moller, M. Schmidt, J. Mol. Struct. 328 (1994) 121–132.

26] P. Shah, A.V. Ramaswamy, K. Lazar, V. Ramaswamy, Appl. Catal. A: Gen. 273(2004) 239–248.

27] X. Liu, X. Piao, Y. Wang, S. Zhu, Energy Fuels 22 (2008) 1313–1317.