The effects of cosolvents on homogeneously and heterogeneously base-catalyzed methanolysis of sunflower oil

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

The effects of cosolvents on homogeneously and heterogeneously base-catalyzedmethanolysis of sunflower oil

Zoran B. Todorovic, Olivera S. Stamenkovic, Ivica S. Stamenkovic, Jelena M. Avramovic, Ana V. Velickovic,Ivana B. Bankovic-Ilic, Vlada B. Veljkovic ⇑Faculty of Technology, University of Niš, Bulevar Oslobodjenja 124, 16000 Leskovac, Serbia

h i g h l i g h t s

" KOH and CaO catalyze the sunflower oil methanolysis in the presence of cosolvents." Tetrahydrofuran favors the triacylglycerol mass transfer by reducing drop size." The CaO-catalyzed methanolysis is not speeded up with the use of cosolvents.

a r t i c l e i n f o

Article history:Received 21 August 2012Received in revised form 6 November 2012Accepted 12 November 2012Available online 26 November 2012

Keywords:Calcium oxideCosolventDrop sizeMethanolysisPotassium hydroxide

a b s t r a c t

Homogeneously and heterogeneously base-catalyzed sunflower oil methanolysis was investigated with-out and with the presence of the cosolvents. In the former case, KOH and tetrahydrofuran (THF) wereused as a catalyst and a cosolvent, respectively, and the reaction was performed at the reaction temper-ature of 10 �C and the methanol-to-oil molar ratio of 6:1. In the latter case, CaO and various organic sol-vents such as THF, n-hexane, dioxane, diethyl ether, triethanolamine, ethyl acetate and methyl ethylketone were employed as a solid catalyst and a cosolvent, respectively, and the reaction was carriedout at the reaction temperature of 60 �C and the methanol-to-oil molar ratio of 6:1. The rate of KOH-catalyzed sunflower oil methanolysis increased with increasing the THF concentration up to 50% of theoil mass, which was due to the self-enhancement of the interfacial area as the result of decreasing the meandrop size. No effect of THF present at the concentration of 20% on the rate of CaO-catalyzed methanolysiswas observed, but at higher THF concentrations, the reaction was delayed and the final fatty acid methylesters (FAME) yield was decreased. Of all tested cosolvents, only n-hexane and THF slightly improvedthe methanolysis reaction in its initial period, triethanolamine and ethyl acetate had no effect, whilediethyl ether, dioxane and methyl ethyl ketone negatively influenced the reaction rate and the FAME yield.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The existing commercial biodiesel production processes mainlyinvolve the homogeneously base-catalyzed methanolysis of edibleoils. This reaction system is a two phase one, where mass transferlimitations due to immiscibility of the reactants determines theoverall reaction rate at the beginning of the reaction. This problemcan be overcome by using vigorous mechanical agitation, enhancedreaction temperature or cosolvents. The breakage of alcoholic reac-tant into a fine drop emulsion in the vegetable oil by intensive mix-ing increases the interfacial area and the triacylglycerol masstransfer rate [1]. The enhanced reaction temperature increasesboth miscibility of the reactants [2] and the reaction rate constants[3]. An organic solvent added as a cosolvent reduces immiscibility

of the reactants and produces a homogeneous solution, increasingthe transesterification reaction rate [4]. However, it should not beforgotten that the cosolvent addition will incur additional costsand will require an additional step for its removal from the finalreaction mixture. Guan et al. [5] warned that ‘‘excessive additionof cosolvent into the reaction system could reduce the transesteri-fication rate and increase the operating cost’’.

Since the work of Boocock et al. [4], many studies have beenperformed on the use of various cosolvents in transesterificationreactions. Most of the previous works have been related to homo-geneously catalyzed methanolysis [4–19]. Tetrahydrofuran (THF)has been mainly used as cosolvent [4–7,9–11,14,16,17], probablybecause of its low price, non-reactivity and easy separation fromthe reaction mixture after completion of the reaction, since meth-anol and THF have close boiling temperatures (65 �C and 67 �C,respectively). Some other solvents have been also investigated ascosolvents in transesterification reactions. Boocock et al. [7]

0016-2361/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.fuel.2012.11.049

⇑ Corresponding author. Tel.: +381 16 247 203; fax: +381 16 242 859.E-mail address: [email protected] (V.B. Veljkovic).

Fuel 107 (2013) 493–502

Contents lists available at SciVerse ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel


recommended the use of THF as the best cosolvent. According toGuan et al. [12], dimethyl ether as an environmentally friendly sol-vent could replace THF as a cosolvent for the transesterificationprocess. The same research team [5] used several cosolvents (THF,dimethyl ether, diethyl ether and methyl tertiary butyl ether) andfound that dimethyl ether was the best one. Park et al. [18] havesuggested the use of methyl esters as a cosolvent. Delgado [9] usedmethyl tertiary butyl ether and THF in the transesterification andinteresterification of triacylglycerols with a mixture of methanoland methyl acetate, to enhance the reaction rate and to obtainmethyl esters and triacetin. Encinar et al. [10] tested various cosol-vents: diethyl ether, diisopropyl ether, dibutyl ether, methyl ter-tiary butyl ether, THF and acetone, among which methyl tertiarybutyl ether and diethyl ether showing the best results. Pena et al.[19] studied the effect of three alkaline catalysts (CH3ONa, NaOH,and KOH) and hexane as a cosolvent on the transesterification ofcastor oil. Hancsók et al. [14] employed THF and dioxine as cosol-vents in a combined acid/base-catalyzed transesterification forthe production of fatty acid methyl esters (FAME) from used fryingoil having a high free fatty acids amount. Hernando et al. [15] stud-ied the transesterification of rapeseed and soybean oils combiningthe use of tert-butyl methyl ether and microwave irradiation to re-duce the mass transfer limitation in the first reaction stage.However, no study on the use of any of above-mentioned cosol-vents in transesterification reactions has included the drop sizemeasurement to visualize their effect on the overall reaction rate.

The increasing demand for biodiesel requires improvement ofits manufacturing process, which means the use of large produc-tion capacities, new catalytic systems, non-edible vegetable oilsand simplified operations ensuring a high biodiesel yield and smal-ler amounts of wastes. For instance, the use of heterogeneous (so-lid) catalysts in biodiesel production bears a number of benefits,compared to that of homogeneous catalysts. The processes of sep-aration and purification of solid catalyst and products is greatlysimpler and cheaper than those related with the use of a homoge-neous catalyst. Also, the solid catalyst can be used repeatedly,reducing the operating costs. Despite the above benefits, the useof cosolvents in heterogeneously catalyzed methanolysis has beendescribed only in a few papers [20–23]. A disagreement betweendifferent researchers on the importance of cosolvents, such asTHF and n-hexane, for the performance of heterogeneoulsy cata-lyzed methanolysis reactions is present in the available literature.The positive action of THF was reported by Gryglewicz [20] andYang and Xie [23], while Ilgen et al. [21] and Kim et al. [22] ob-served an inhibitory effect of THF. Kim et al. [22] found that meth-anolysis reaction was improved in the presence of n-hexane, whileIlgen et al. [21] and Yang and Xie [23] reported a negative effect ofthis cosolvent. The observed disagreement on the effect of cosol-vents on heterogeneously catalyzed transesterifications should beinvestigated in order to be clarified.

The present paper deals with homogeneously and heteroge-neously base-catalyzed sunflower oil methanolysis with the pres-ence of some cosolvents at the 6:1 methanol-to-oil molar ratio. Inthe former case, KOH and THF were used as a catalyst and a cosol-vent, respectively. The reaction was conducted at 10 �C to lower thereaction rate and to facilitate the observation of mass transfer lim-itation in the beginning of the reaction. The goal was to show vari-ations in the Sauter mean drop size and the drop size distributionduring the reaction carried out in the presence of various concen-trations of THF (from 0% to 100% of the oil mass). In the latter case,CaO and various organic solvents such as THF, n-hexane, dioxane,diethyl ether, triethanolamine, ethyl acetate and methyl ethyl ke-tone were employed as a heterogeneous catalyst and a cosolvent,respectively. The initial amount of catalyst was 5% (based on theoil mass), as it was used in a previous study [24], and the amountof THF was from 0% to 100% of the oil mass. The reaction was con-

ducted at 60 �C. The main aim was to reflect the effect of THF andsome other cosolvents having different hydrophobicity and polar-ity on the CaO-catalyzed sunflower oil methanolysis.

2. Experimental

2.1. Materials

Commercial sunflower oil (Sunce, Serbia) was used. Methanol(99.8%) and KOH (85%) were purchased from Lachema (CzechRepublic) and Merck–Alkaloid (FYR of Macedonia), respectively.CaO (extra pure) was from Centrohem (Serbia). Immediately beforeuse CaO was activated by calcination at 550 �C for 2 h, as recom-mended elsewhere [24]. To avoid adsorption of water and carbondioxide from the air on catalytic active sites on the catalyst surface,the activated CaO was cooled and then stored in a well closed, glassbottles in a desiccator containing CaCl2 and KOH. Also, the follow-ing solvents were used: diethyl ether (Lachner), n-hexane (99%,Promochem, Germany), ethylacetate (99.5%, Merck), n-butylacetate(99%, Zorka–Šabac, Serbia), dioxane (p.a., Merck–Alkaloid), methylethyl ketone (HPLC grade 99.5%, JT Baker), propylene carbonate,triethanolamine (99%, Aldrich–chemic), n-hexane and 2-propanol(HPLC grade, JT Baker). The properties of the employed cosolventsand methanol, as well as minimum molar ratios required forachieving a complete dissolution of methanol and sunflower oil,are presented in Table 1. Standards of triolein, diolein, monooleinand a mixture of methyl esters of palmitic, stearic, oleic, linolenicand linoleic acids (20% of each ester) were purchased from SigmaCo.

2.2. Catalyst characterization

Morphological and textural properties of commercial (raw) CaOand calcined CaO were characterized by the nitrogen physisorptionmethod. The morphology of samples was examined using a JEOLJSM 530 scanning electron microscope. High-resolution adsorptionnitrogen adsorption–desorption isotherms were determined usinga Sorptomatic 1990 Thermo Finnigan at �196 �C. Typically, 4 g ofsamples was used in analysis after outgassed under at 110 �C dur-ing 18 h. Data space volume was determined using helium. Thetextural parameters were calculated according to common meth-ods using software package Advanced Data Processing Ver. 5.17.

Using the part of the adsorption isotherm which includes bestlinear fit for relative pressure up to 0.2, the specific surface areawas calculated using the Brunauer, Emmett and Teller method[32]. Total pore volume was calculated at maximal adsorptionpressure. Mesopore volume and pore size distribution were ob-tained according to the Barrett, Joyner and Halenda method [33]from desorption branch of isotherm, using standard isothermLecloux and Pirard [34]. The Dubinin–Radushkevich method wasused for the calculation of micropore volume [35].

The basic strength of the catalyst was determined by the Ham-mett indicators method and expressed by an acidity function [36].The following Hammett indicators were used: phenolphthalein(H_ = 9.3), thymolphthalein (H_ = 10.0), thymol violet (H_ = 11.0),2,4-dinitroaniline (H_ = 15.0) and 4-nitroaniline (H_ = 18.4). Thecatalyst basic strengths were quoted as being stronger than theweakest indicator that exhibited a color change, but weaker thanthe strongest indicator that produces no color change.

2.3. Methanolysis

2.3.1. Homogeneously base-catalyzed methanolysisThe homogenously base-catalyzed methanolysis without and

with adding THF as a cosolvent was performed in a 1 L glass

494 Z.B. Todorovic et al. / Fuel 107 (2013) 493–502


reactor, equipped with a two-flat blades stirrer (diameter: 75 mm;placed at the one third of the height of the reaction mixture fromthe reactor bottom; agitation speed: 200 rpm), under the atmo-spheric pressure and at 10 �C. The reactor was immersed in athermostated chamber kept at 10 ± 0.1 �C. Sunflower oil (600 g)was poured into the reactor and thermostated for 30 min. In themeantime, the catalyst (6 g KOH) was dissolved in methanol(132 g; methanol-to-oil molar ratio: 6:1) in a glass beaker placedin the chamber at 10 �C. Then, a predetermined amount of a cosol-vent (0.1%, 1%, 5%, 10%, 20%, 30% and 50% of the oil mass) wasadded to methanol and the catalyst, while the glass beaker wasthoroughly shaken. Finally, the mixture of methanol, the cosolventand the catalyst was added to the sunflower oil in the reactor, andthe reaction was timed.

2.3.2. Heterogeneously base-catalyzed methanolysisThe apparatus for heterogeneously-catalyzed methanolysis

consisted of a 500 mL three-neck round-bottomed flask, placed ina glass chamber, and equipped with a magnetic stirrer and a con-denser. Water was circulated through the chamber from a thermo-stated bath (60 ± 0.1 �C). Sunflower oil (91.92 g) was added to thereactor. In the first series of experiments, the amount of CaO(4.60 g; 5% of the oil mass; the mass concentration of 3.93% basedon the total reaction mixture) was held constant at differentamounts of THF. In the second series of experiments, to avoid thedilution of the catalyst by adding different amounts of THF, theamount of CaO was varied to keep the CaO mass concentrationconstant at 3.93%. In the third series of experiments, where the ef-fect of different cosolvents (20% based on the mass oil) was stud-ied, the amount of CaO was 5.35 g, corresponding to the massconcentration of 3.93%. In all experiment, the catalyst was sus-pended into methanol (20.28 g; methanol-to-oil molar ratio: 6:1)in a glass beaker. The predetermined amount of a cosolvent wasadded to the mixture of methanol and the catalyst, while the glassbeaker was thoroughly shaked. The overall mixture was added tosunflower oil in the reaction flask, and the reaction began to betimed.

2.4. Analytical method

Samples (1 mL) were removed from the reaction mixture duringthe methanolysis reaction neutralized by adding a solution ofhydrochloric acid (1:4) to the catalyst, and centrifuged (5000g,5 min). The upper layer was withdrawn, dissolved in 2-propanol/n-hexane (5:4 v/v) in an appropriate ratio and filtered through a0.45 lm Millipore filter. The resulting filtrate was used for theHPLC analysis, which was described in details by Avramovic et al.[25].

2.5. Drop size measurements

The effect of the amount of THF on the Sauter mean drop sizeand drop size distribution during the KOH-catalyzed methanolysiswas investigated. The photos of the agitated reaction mixtureswere taken during the reaction. Not less than 200 drops were mea-sured for each operational condition. The photographic systemconsisted of a 500 W light source and an OLYMPUS SP-550UZ cam-era with ultra-zoom lens 28–504 mm (equiv) and Bright lens1:2.8–4.5 with Super Macro capability of 1 cm. The photos werethen processed using image software to determine drop size [1].The Sauter-mean drop diameter, d32, was calculated as follows:

d32 ¼P

nid3i

Pnid

2i

where di is the drop diameter and ni is the number of drops withdiameter di.

3. Results and discussion

3.1. KOH-catalyzed sunflower oil methanolysis with the presence ofTHF

Fig. 1 shows variations of the FAME yield during the meth-anolysis reaction catalyzed by KOH (1.0% of the oil mass) inthe presence of different amounts of THF (from 0% to 50% basedon the oil mass) at 10 �C and the methanol-to-oil molar ratio of6:1. The low reaction temperature was chosen to reduce the

Table 1The properties of the employed cosolvents and methanol.a

Cosolventb TEA THF ETAC HEX DEE DIOX MEK MET

Density (g/ml) 1.130 0.886 0.894 0.655 0.713 1.034 0.805 0.791Viscosity at 20 �C (mPa) 0.36 0.55 0.45 0.31 0.24 1.37 0.43 0.59Partition coefficients at 25 �C (logP) �2.3 0.46 0.73 4.00 0.89 �0.49 0.29 �0.76Dipole moment, D 3.48 1.63 1.78 0.08 1.15 0.45 2.78 1.70Dielectric Constant at 40 �C 5.14 7.58 6.02 2.02 4.33 2.25 18.5 32.7Minimum molar ratioof methanol/oil/cosolvent for obtaining a single phasec –d 6:1:1.80 6:1:2.49 6:1:3.53 6:1:2.11 6:1:1.71 6:1:2.16 –

a All properties are considered uncertain by the original Refs. [27–31].b TEA – triethanolamine, THF – tetrahydrofuran, ETAC – ethyl acetate, HEX - n-hexane, DEE – diethyl ether, DIOX – dioxane, MEK – methyl ethyl ketone and MET –

methanol.c Determined in the present work by a titration method at 60 �C. A cosolvent was added to the oil–methanol mixture, and its amount which changed the mixture from

turbid to transparent was recorded.d It was not possible to determine the point when the mixture changed from turbid to transparent.

Fig. 1. Variations of the FAME yield with the progress the sunflower oil methan-olysis reaction catalyzed by KOH at different THF concentrations: 0% – s, 1% – D, 5%–r, 10% – e, 15% – �, 20% – N, 30% – . and 50% – � (catalyst amount, based on theoil weight: 1%; the reaction temperature: 10 �C; and the methanol-to-oil molarratio: 6:1).

Z.B. Todorovic et al. / Fuel 107 (2013) 493–502 495


reaction rate making more observable the effect of the cosol-vent. The sigmoidal curves were observed when the methanol-yisis reaction was performed without and with the presence ofTHF at low concentrations. With no THF present, the methanol-ysis was initially slow because of mass transfer limitation, thenthe FAME yield increased, due to self-enhancement caused byincreased interfacial area [1] and finally, the reaction reachedthe equilibrium. At the THF concentrations of 1% and 5%, actu-ally the same sigmoid variations of the FAME yield with timewere observed. However, at the THF concentration of 10%, the

sigmoid shape was hardly observed, indicating no initial masstransfer limitations, which agreed with the observance ofBoocock et al. [4,7], Ataya et al. [6] and Kumar et al. [17].The FAME yield higher than 30% was observed in the first reac-tion minute. With further addition of THF to the reaction mix-ture, the initial reaction rate was accelerated, so that the FAMEyield of 50% was observed in the first 15 s at the THF concen-tration of 50%. The curves representing the variation of theFAME yield with time became exponential, and the equilibriumwas reached in less than 10 min. This observance confirmed theresults of Çaglar [26], which showed that 50% THF was neces-sary to make the reactants completely miscible at the reactiontemperature of 20 �C. According to Guan et al. [5], the oil con-version reached its maximum near the point of the minimummolar ratio of THF-to-oil which was required for the completedissolution of the reactants. When THF was added to the reac-tion system in excess, the methanolysis rate decreased becauseof diluting the reactants.

Although THF has been widely used cosolvent in transesterifica-tion reaction systems, only Guan et al. [12] studied the synthesis ofbiodiesel from sunflower oil in the presence of this cosolvent atroom temperature. Their results showed the same trend as thepresent ones. In their experiments, sunflower oil was almost com-pletely converted into biodiesel after 20 min reaction while only aconversion of 78% was reached in the absence of THF. All other re-ported studies on the use of THF as a cosolvent have been relatedto transesterification of some other oils, for instance canola oil [6],soybean oil [7], as well as mahua and jatropha oils [17].

The effect of THF on the methanolysis rate could be explainedby the self-enhancement caused by increasing the mass transferFig. 2. Variation of the Sauter-mean drop diameter (mm) during the methanolysis

reaction at performed at different THF concentrations (agitation speed: 130 rpm).

Fig. 3. Drop size distributions during methanolysis when 10% tetrahydrofuran was added after: (a) 0.5 min, (b) 1 min, (c) 2 min and (d) 4 min (temperature: 10 �C; methanol-to-oil molar ratio: 6:1; and agitation speed: 130 rpm).



rate and the change in polarity of the reaction mixture. These ef-fects were due to the effective agitation due to the smaller viscos-ity of the reaction mixture containing a cosolvent and betterglycerol separation from the FAME–oil phase to the alcoholicphase. It could be also assumed that the change in polarity of thereaction mixture affects the ionization of KOH and consequentlythe reaction rate [4].

The positive effect of the self-enhancement of the interfacialarea on the mass transfer rate in the presence of THF was expectedto be the same as that in its absence. This was documented by pho-

tos of the reaction mixtures taken during methanolysis reactionscarried out at different THF concentrations. The same conclusionswere withdrawn from the Sauter-mean drop diameter and thedrop size distribution curves.

The variations of the Sauter-mean drop diameter with the reac-tion time for the methanolysis performed at various THF concen-trations are shown in Fig. 2. The Sauter-mean drop diameterdecreased both with increasing the THF concentration and thereaction time. The reduction of methanol amount due to the meth-anolysis reaction and the stabilization of dispersed drops by the

Fig. 4. Drop size distributions at different THF concentrations after one minute methanolysis reaction (temperature: 10 �C; methanol-to-oil molar ratio: 6:1; and agitationspeed: 130 rpm).



emulsifiers formed (mono- and diglycerides and soaps) were thereasons for the decrease of the Sauter-mean drop diameter withthe reaction time. Also, the amount of methanol to be emulsifieddecreases due to methanol dissolution into THF, causing the reduc-tion of the Sauter-mean drop diameter.

The Sauter-mean drop diameter does not always fully describethe emulsion. Furthermore, it is possible that emulsions with thesame Sauter-mean drop diameter have different drop size distribu-tions. Figs. 3 and 4 show variations of the drop size distributionduring the methanolysis with presence of the 10% THF and afterone minute reaction at different THF concentrations, respectively.The photographies of the reaction mixtures shown in Figs. 5 and6 illustrate the variation of drop size during the methanolysis withpresence of the 10% THF and after one minute methanolysis reac-tion at different THF concentrations, respectively. These figuresillustrate the positive effect of THF on the drop size distribution.During the methanolysis reaction at a constant THF concentration(10%), the drop size distribution curve became narrorer and itspeak shifted towards lower drop sizes (Fig. 3). As can be seen fromFig. 4, drop size distribution curves corresponding to the reactionmixtures after one minute reaction were bimodal for the reactionmixture without and with 1% THF, while at higher THF concentra-tions they were unimodal. At medium THF concentrations (up to10%), drop size distribution curves were asymmetric with a highpeak on the low drop size side and a short tail on the large dropsize side. Generally, the drop size distribution curves became nar-rower and shifted towards lower size with increasing the THF con-centration above 10%. At higher THF concentrations (20% and 30%),

the drop size distribution curves became more symmetric, and thevalue of the Sauter-mean drop size was close to the peak of thecurve.

3.2. Catalyst characterization

Scanning electron microphotographs of raw and calcined CaOare presented in Fig. 7. Thermal treatment did not inducesignificant change in morphology. The particles of both samplesdiffer in shape and size. The size of agglomerates decreased inthe calcined CaO.

Textural analysis of calcined CaO was performed using adsorp-tion–desorption isotherms of nitrogen at �196 �C. The softwareenabled analysis of textural parameters, and the obtained resultsare given in Table 2. Isotherm of raw CaO can be classified as TypeII according to IUPAC nomenclature [36]. This type of isotherm istypical for nonporous or macroporous materials. Low values of to-tal pore volume clearly demonstrated nonporous nature of rawCaO. Although the isotherm of CaO calcined at 550 �C is also typeII, the change of its shape is obvious almost in the whole relativepressure region (Fig. 8). Especially, the hysteresis loop in the regionof relative pressures higher than 0.7 indicates the altered texturalcharacteristics compared to raw CaO. Calcination caused an in-crease of the specific surface area, the total and mesopore volumeand even the micropore volume. Obviously, this can be associatedwith the evolution of gaseous products, expulsion of which duringthe calcination of CaO generating new pore system and moreporosity of the calcined sample.

Fig. 5. Photographies of the reaction mixtures during methanolysis when 10% tetrahydrofuran was added (temperature: 10 �C; methanol-to-oil molar ratio: 6:1; andagitation speed: 130 rpm).



The basic strength of the raw and calcined CaO is also tabulatedin Table 2. It was observed that basic strength increased after cal-cination of CaO. This might be attributed to the fact that the orderof basicity for the alkaline earth metal compounds is as follows[37]: oxide > hydroxide > carbonate, so the increase of basicstrength could be attributed to the formation of oxide form. Thisresult is similar to that reported by Kouzu et al. [38] and Mootabadiet al. [39]

3.3. CaO-catalyzed methanolysis of sunflower with the presence of acosolvent

Fig. 9 presents the change of the FAME yield during thesunflower oil methanolysis catalyzed by CaO at different

concentrations of THF used as a cosolvent. The initial catalyst load-ing of 5% (based on the oil mass) was selected from a previousstudy [24]. THF was added in the amount of 20%, 30%, 50% and100% (based on the oil mass) to the CaO suspension in methanol.The liquid part of all reaction systems were single-phase at 60 �C.The methanolysis reaction was almost completed in 3 h when noTHF had been added to the reaction mixture. No effect of THF pres-ent at the concentration of 20% was observed. However, at higherTHF concentrations, the reaction was delayed, the curves repre-senting the change of the FAME yield with time were shifted atlonger times, and the final FAME yield was decreased. For instance,when THF was added 100%, the FAME yield was only 11% after 4 h.Kim et al. [22] have also reported that the biodiesel yield in thesoybean oil methanolysis catalyzed by Na/NaOH/c-Al2O3

Fig. 6. Photographies of the reaction mixtures after 1 min methanolysis reaction at different THF concentrations (temperature: 10 �C; methanol-to-oil molar ratio: 6:1; andagitation speed: 130 rpm).



decreased with addition of THF but it increased in the presence ofn-hexane. Several researchers have, however, reported that THFsubstantially increased the methanolysis rate or the final conver-sion. The positive effect of THF (10%) on the rate of soybean oilmethanolysis catalyzed by CaO was observed by Gryglewicz [20].Yang and Xie [23] have published that THF increased the conver-sion of soybean oil in the methanolysis reaction catalyzed bySr(NO3)2/ZnO, while it was reduced by n-hexane and DMSO. Ilgenet al. [21] have reported that the conversion did not change in thecanola oil methanolysis catalyzed by Amberlyst-26 in the presenceof THF (10%) but decreased with the addition of n-hexane (10%).The positive effect of THF was explained as the result of mutualsolubility of vegetable oil, methanol and THF [20]. No explanationof the negative effect of THF or other cosolvent on the methanoly-sis rate and the final conversion degree has been found in the liter-ature. Ilgen et al. [21] speculated that the cosolvents did not

Fig. 7. SEM microphotographs of: (a) raw CaO and (b) calcined CaO.

Table 2Selected textural properties and basic strength of raw and calcined CaO.a

Sample SBET (m2/g) V0.999(cm3/g) Vmes (cm3/g) Dmax (nm) Dmed (nm) Vmic (cm3/g) Basic strength (H_)

Raw CaO 5.9 0.023 0.016 3.6 19.9 0.002 9.3 < H_ < 10Calcined CaO 13.7 0.138 0.077 3.5 15.2 0.004 15 < H_ < 18.4

a SBET – specific surface area, V0.999 – total pore volume at maximal adsorption pressure, Vmes – mesopore volume, Dmax – the most abundant pore diameter, Dmed – medianpore diameter, Vmic –micropore volume.

Fig. 8. N2 adsorption/desorption isotherms of raw and calcined CaO.

Fig. 9. The variations of FAME yield with the progress of sunflower oil methanolysiscatalyzed by CaO at different concentrations of THF: 0% – �, 20% – s, 30% –D, 50% –h and 100% – e (oil: 91.92 g; catalyst loading: 5% of the oil weight; the reactiontemperature: 60 �C and the methanol-to-oil molar ratio: 6:1).

Fig. 10. The variations of FAME yield with the progress of sunflower oil methan-olysis catalyzed by CaO at different concentrations of THF: 0% – �, 20% – s, 30% –D,50% – h and 100% – e (oil: 91.92 g; catalyst loading: 5% of the total volume of thereaction mixture; the reaction temperature: 60 �C and the methanol-to-oil molarratio: 6:1).



effectively homogenize the reactants and did not speed up the rateof formation of mono- and diacylglycerols.

The observed negative effect of the cosolvent on the methanol-ysis reaction catalyzed by CaO might be attributed to the dilutionof both reactants and the catalyst with the addition of THF. The cat-alyst concentration was held constant at 5% (based on the totalmass of the reaction mixture) to avoid its dilution. However, thesame effect of THF was observed in these experiments as can beseen in Fig. 10, where the variation of FAME yield with time waspresented. Thus, it could be concluded that an inhibition of the cat-alyst at higher THF concentrations was responsible for its negativeeffect on the rate of methanolysis catalyzed by CaO and the finalFAME yield. As can be seen in Fig. 7, the initial methanolysis rateat THF concentrations of 20% and 30% was greater than that forthe reaction performed in the absence of the cosolvent. In the laterperiod of the reaction, the rate of methanolysis was practically thesame as that in the absence of THF at the concentration of 20% andthe same equilibrium FAME yields were achieved. At the THF con-centration of 30%, however, the FAME yield was significantly lowerthan that achieved in the reaction without THF, indicating aninhibitory effect of the cosolvent. At THF concentrations of 50%and 100%, the inhibitory effect was even enhanced.

In order to check how the other cosolvents, employed in theprevious studies of the effect of cosolvents on heterogeneously cat-alyzed methanolysis of vegetable oils, impact on the CaO-catalyzedmethanolysis, this reaction was performed in the presence of sev-eral organic solvents: THF, diethylether, dioxane, methyl ethyl ke-tone, triethanolamine, n-hexane and ethyl acetate. The amount ofcosolvents was 20% of the mass oil, and the catalyst amount was5.35 g (corresponding to the mass concentration of 3.93%). The li-quid part of the reaction systems in the presence of THF, dioxane,methyl ethyl ketone and diethylether were single-phase, while thereaction systems with n-hexane and ethyl acetate were two-phase.The changes of the FAME yields during the methanolysis in thepresence of different cosolvents are presented in Fig. 11. The FAMEyields achieved in the presence of different cosolvents in 4 h wereas follows: THF – 98.1%, diethyl ether – 88.3%, dioxane – 94.8%,methyl ethyl ketone – 14.7%, triethanolamine – 97.9%, n-hexane– 98.8% and ethyl acetate – 97.1%; the FAME yield in the reactionwithout a cosolvent was 98.0%. Of all tested cosolvents, onlyn-hexane and THF slightly improved the methanolysis reaction inits initial period. Triethanolamine and ethyl acetate had no effect

on the methanolysis reaction, compared to the reaction withouta cosolvent. Diethyl ether, dioxane and methyl ethyl ketone nega-tively influenced both the reaction rate and the FAME yield. TheFAME yields with methyl ethyl ketone, dioxane and diethyletherachieved in 1 h were only 0.4%, 5.8% and 9.4%, respectively, whichwere much lower than the FAME yield (18.8%) obtained for thesame reaction time without a cosolvent. It is unclear why thesecosolvents, which are very good for homogenous methanolysis,have an inhibitory effect on heterogeneously-catalyzed methanol-ysis. A possibility could be blocking the active sites on the catalystsurface by a cosolvent, which will prevent the formation of calciummethoxide that catalyzes the methanolysis reaction. However, thefact that polar THF and nonpolar n-hexane behave similarly makesunlikely that the blockage of active sites really occurs.

Trying to explain the effects of the cosolvents used on the rateof sunflower oil methanolysis catalyzed by CaO, the FAME yieldachieved in 1 h was correlated with hydrophobicity (logP, whereP is the partition coefficient, its values for the cosolvents usedbeing given in Table 1) in Fig. 12. The FAME yield generally in-creased with increasing logP; the only exception is the FAME yieldachieved with methyl ethyl ketone. The FAME yields in the pres-ence of the cosolvents having different functional groups and sim-ilar logP-values were different, which implied that specificfunctional groups might have an effect on the methanolysisreaction. At the first look, it is difficult to understand how twoco-solvents, such as n-hexane and THF, with very different polarityand hydrophobicity, give rise to similar FAME yields. The sameeffects of THF, a polar solvent, and n-hexane, a nonpolar solvent,might be due to the nature of the reactants, methanol and sun-flower oil, which are polar and nonpolar, respectively. It might beexpected that the polar solvent favors the miscibility of the polarreactant, while the nonpolar solvent helps the dissolution of thenonpolar reactant. Also, n-hexane with a much lower viscositythan sunflower oil dramatically reduces the viscosity of the reac-tion mixture, thereby decreasing the mass transfer resistance andincreasing both the reaction rate and the FAME yield in the initialperiod of the reaction.

4. Conclusion

The KOH- and CaO-catalyzed methanolysis of sunflower oil wasinvestigated without and with the presence of different cosolvents,

Fig. 11. The variations of the FAME yields with the progress of sunflower oilsmethanolysis catalyzed by CaO in the presence of different cosolvents (20% basedon the oil weight): no cosolvent – �; methyl ethyl ketone (dielectric constant, 18.4)– s; THF –D; ethyl acetate (6.0) –r; triethanolamine (5.14) – h; diethyl ether (4.3)– e; dioxane (2.25) – ⁄; and n-hexane (1.9) – + (reaction conditions: oil, 91.92 g;catalyst, 5%; 60 �C; and methanol-to-oil molar ratio, 1:6).

Fig. 12. Correlation of FAME yield achieved in 1 h with logP of cosolvents(methanol – MET, tetrahydrofuran – THF, n-hexane – HEX, dioxane – DIOX, diethylether – DEE, triethanolamine – TEA, ethyl acetate – ETAC, and methyl ethyl ketone –MEK).



trying to reflect the effect of cosolvent presence on the final FAMEyield and the reaction rate. The study confirmed the positive effectof THF on the KOH-catalyzed sunflower oil methanolysis at its con-centrations up to 50%, which was attributed not only to the bettermiscibility of the reactants but also to the better mass transfer asthe result of the self-enhancement of the interfacial area. Themethanolysis reaction in the presence of THF was very fast, com-pared to the reaction in its absence, it being completed in a coupleof minutes at the THF concentration of 50%. However, only a slightpositive effect of THF and n-hexane, present in the reaction systemcontaining CaO as a solid catalyst at the concentration of 20% wasobserved. Furthermore, triethanolamine and ethyl acetate had noeffect on the methanolysis reaction, and diethyl ether, dioxaneand methyl ethyl ketone negatively affected the FAME yield andthe reaction rate. Thus, the use of cosolvents to speed up the meth-anolysis is of no practical importance if CaO is used as a solid cat-alyst. However, further studies are needed to clarify the sameimpact of THF and n-hexane having different polarity on theCaO-catalyzed methanolysis of sunflower oil.

Acknowledgment

This work has been funded by the Ministry of Education andScience of the Republic of Serbia (Project III 45001).

References

[1] Stamenkovic OS, Lazic ML, Todorovic ZB, Veljkovic VB, Skala DU. The effect ofagitation intensity on alkali-catalyzed methanolysis of sunflower oil.Bioresource Technol 2007;98:2688–99.

[2] Vicente G, Martinez M, Aracil J, Esteban A. Kinetics of sunflower oilmethanolysis. Ind Eng Chem Res 2005;44:5447–54.

[3] Stamenkovic OS, Todorovic ZB, Lazic ML, Veljkovic VB, Skala DU. Kinetics ofsunflower oil methanolysis at low temperatures. Bioresource Technol2008;99:1131–40.

[4] Boocock DGB, Konar SK, Mao V, Sidi H. Fast one-phase oil-rich processes for thepreparation of vegetable oil methyl esters. Biomass Bioenergy 1996;11:43–50.

[5] Guan G, Sakurai N, Kusakabe K. Synthesis of biodiesel from sunflower oil atroom temperature in the presence of various cosolvents. Chem Eng J2009;46:302–6.

[6] Ataya F, Dube MA, Ternan M. Single-phase and two-phase base-catalyzedtransesterification of canola oil to fatty acid methyl esters at ambientconditions. Ind Eng Chem Res 2006;45:5411–7.

[7] Boocock DGB, Konar SK, Mao V, Lee C, Buligan S. Fast formation of high puritymethyl esters from vegetable oils. J Am Oil Chem Soc 1998;75:1167–72.

[8] Casas A, Fernández CM, Ramos MJ, Pérez Á, Rodríguez JF. Optimization of thereaction parameters for fast pseudo single-phase transesterification ofsunflower oil. Fuel 2010;89:650–8.

[9] Delgado J. Procedimiento para producir combustibles biodiesel conpropiedades mejoradas a baja temperature. SP patent 2201894; 2002.

[10] Encinar JM, González JF, Pardal A, Martínez G. Transesterification of rapeseedoil with methanol in the presence of various co-solvents. In: ProceedingsVenice 2010, third international symposium on energy from biomass andwaste, Venice, Italy, 8–11 November 2010. Italy: CISA, Environmental SanitaryEngineering Centre; 2010.

[11] Guan G, Kusakabe K. Synthesis of biodiesel fuel using an electrolysis method.Chem Eng J 2009;153:159–63.

[12] Guan G, Kusakabe K, Sakurai N, Moriyama K. Rapid synthesis of biodiesel fuelsat room temperature in the presence of dimethyl ether. Chem Lett2007;36:1408–9.

[13] Guan G, Kusakabe K, Sakurai N, Moriyama K. Transesterification of vegetableoil to biodiesel fuel using acid catalysts in the presence of dimethyl ether. Fuel2009;88:81–6.

[14] Hancsók J, Kovács F, Krár M. Production of vegetable oil fatty acid methylesters from used frying oil by combined acidic/alkali transesterification. PetrolCoal 2004;46:36–44.

[15] Hernando J, Leton P, Matia MP, Novella JL, Alvarez-Builla J. Biodiesel and FAMEsynthesis assisted by microwaves: homogeneous batch and flow processes.Fuel 2007;86:1641–4.

[16] Karmee SK, Chadha A. Preparation of biodiesel from crude oil of Pongamiapinnata. Bioresource Technol 2005;96:1425–9.

[17] Kumar GR, Ravi R, Chadha A. Kinetic studies of base-catalyzedtransesterification reactions of non-edible oils to prepare biodiesel: theeffect of co-solvent and temperature. Energy Fuel 2011;25:2826–32.

[18] Park JY, Kim DK, Wang ZM, Lee JS. Fast biodiesel production with one-phasereaction. Appl Biochem Biotechnol 2009;154:246–52.

[19] Pena R, Romero R, Martinez SL, Ramos MJ, Martinez A, Natividad R.Transesterification of castor oil: effect of catalyst and co-solvent. Ind EngChem Res 2009;48:1186–9.

[20] Gryglewicz S. Rapeseed oil methyl esters preparation using heterogeneouscatalysts. Bioresource Technol 1999;70:249–53.

[21] Ilgen O, Akin AN, Boz N. Investigation of biodiesel production from canola oilusing Amberlyst-26 as a catalyst. Turk J Chem 2009;33:289–94.

[22] Kim H-J, Kang B-S, Kim M-J, Park YM, Kim D-K, Lee J-S, et al. Transesterificationof vegetable oil to biodiesel using heterogeneous base catalyst. Catal Today2004;93:315–20.

[23] Yang Z, Xie W. Soybean oil transesterification over zinc oxide modified withalkali earth metals. Fuel Process Technol 2007;88:631–8.

[24] Veljkovic VB, Stamenkovic OS, Todorovic ZB, Lazic ML, Skala DU. Kinetics ofsunflower oil methanolysis catalyzed by calcium oxide. Fuel 2009;88:1554–62.

[25] Avramovic JM, Stamenkovic OS, Todorovic ZB, Lazic ML, Veljkovic VB.Empirical modeling of ultrasound assisted base-catalyzed sunflower oilmethanolysis kinetics. Chem Ind Chem Eng Q 2012;18:115–27.

[26] Çaglar E. Biodiesel production using co-solvent. Book of Abstracts. In:European congress of chemical engineering (ECCE-6), Copenhagen, 16–20September, 2007.

[27] Sangster J. Octanol–water partition coefficients of simple organic compounds.J. Phys Chem Ref Data 1989;18(3):1121–77.

[28] 1,4-DIOXANE – Caledon Labs. <www.caledonlabs.com/.../msds/4300-1e.pdf>.[29] Verschueren K. Handbook of environmental data on organic chemicals. 3rd

ed. New York: Van Nostrand Reinhold; 1996. p. 1823–25.[30] Dean J. A handbook of organic chemistry. New York: McGraw-Hill; 1987.[31] Howard PH, Maylan WM. Hand book of physical properties of organic

chemicals. 1st ed. New York: CRC-Press; 1997.[32] Rouquerol F, Rouquerol J, Sing K. Adsorption by powders and porous

solids. London: Academic Press; 1999.[33] Barrett EP, Joyner LG, Halenda PP. The determination of pore volume and area

distribution in porous substances. I. Computations from nitrogen isotherms. JAm Chem Soc 1951;73:373–80.

[34] Lecloux A, Pirard JP. The importance of standard isotherms in the analysis ofadsorption isotherms for determining the porous texture of solids. J ColloidInterface Sci 1979;70:265–81.

[35] Gregg SH, Sing KS. Adsorption, surface area and porosity. New York: AcademicPress; 1967.

[36] Xie W, Li H. Alumina-supported potassium iodide as a heterogeneous catalystfor biodiesel production from soybean oil. J Mol Catal A: Chem 2006;255:1–9.

[37] Yoosuk B, Udomsap P, Puttasawat B. Hydration–dehydration technique forproperty and activity improvement of calcined natural dolomite inheterogeneous biodiesel production: structural transformation aspect. ApplCatal A: Gen 2011;395:87–94.

[38] Kouzu M, Kasuno T, Tajika M, Sugimoto Y, Yamanaka S, Hidaka J. Calcium oxideas a solid base catalyst for transesterification of soybean oil and its applicationto biodiesel production. Fuel 2008;87:2798–806.

[39] Mootabadi H, Salamatinia B, Bhatia S, Abdullah AZ. Ultrasonic-assistedbiodiesel production process from palm oil using alkaline earth metal oxidesas the heterogeneous catalysts. Fuel 2010;89:1818–25.


The effects of cosolvents on homogeneously and heterogeneously base-catalyzed methanolysis of sunflower oil

Documents