J. of Supercritical Fluids 61 (2012) 2 8

Contents lists available at SciVerse ScienceDirect

The Journal of Supercritical Fluids

j ourna l ho me page: www.elsev ier .com

The ph antemper no

Abdualn r MFaculty of Tech

a r t i c l

Article history:Received 18 MReceived in reAccepted 3 Oc

Keywords:VapourliquidSupercriticalSubcriticalEthanolysis of

iglycsion and psure aol mi

deteas 2

nd reulatio

overae begase (asing

2011 Elsevier B.V. All rights reserved.

1. Introduction

Biodiese(alcoholysisedible oils wmainly prothe currenttransportatfossil sourcthe supply products inthis limitatsources, suc

Presentlcan be foucommonly production exists, the biodiesel chas been dfor biodiesevide sufcibiodiesel pr

CorresponE-mail add

for its environmental advantages. Since process of transester-ication has several technological variants, like: homogeneous

0896-8446/$ doi:10.1016/j.l is currently produced by transestericationmethanolysis or ethanolysis) of various types ofith methanol or ethanol. However, since methanol is

duced from fossil sources (crude oil and natural gas) biodiesel production leads to only partly renewableion fuel production [1]. Furthermore, the depletinges have raised concerns related to uncertainty ofcoupled with rising costs of different petrochemicalcluding methanol. A possible solution to overcomeion is to use other alcohols produced from renewableh as ethanol.y, ethanol can be easily produced from biomass thatnd in abundance, via fermentation process, and it isknown as bioethanol. Furthermore, since large scaleof bioethanol as a substitute fuel for gasoline alreadysupply of bioethanol for the industrial production ofould be easily achieved. So far, very limited effortedicated to the research on the utilization of ethanoll production. This effort is needed in order to pro-

ent body of technical data needed by the commercialoducers, in order to promote application of bioethanol

ding author. Tel.: +381 113303707; fax: +381 113370473.ress: sglisic@tmf.bg.ac.rs (S.B. Glisic).

catalytic synthesis, heterogeneous catalytic synthesis and highpressure non-catalytic synthesis, in case of fatty acid ethyl esters(FAEE) these variants have to be investigated and optimized. Theresearch dedicated to biodiesel production from ethanol via non-catalytic transesterication under supercritical conditions is stillvery limited and much more information is required for pro-cess optimization which could result in maximum yields of FAEEbiodiesel [2].

The supercritical alcoholysis (SCA) has been identied as apotentially promising technique for future large-scale biofuel pro-duction, especially for biodiesel production from used vegetableoil and animal fat. Compared to conventional biofuel productiontechnology, the SCA has a number of advantages that includefast reaction rate, high yield of fatty acid ethyl esters (FAEE),and ease of continuous operation and elimination of catalysts(especially homogeneous ones which reduce the overall processefciency). Supercritical method of FAEE synthesis could alsoprovide improved phase solubility, decreased mass-transfer lim-itations, simpler separation and purication process steps, andcould also provide additional robustness regarding processingof various types of vegetable oils, even waste vegetable oils.The severe operating environment, i.e. the high temperature andpressure and the requirements related to the materials of con-struction and associated costs, are the main concerns for its widerapplication [3].

see front matter 2011 Elsevier B.V. All rights reserved.supu.2011.10.002ase equilibrium of triglycerides and ethature: The inuence on kinetics of etha

aser Muftah Almagrbi, Sandra B. Glisic , Aleksandanology and Metallurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia

e i n f o

ay 2011vised form 3 October 2011tober 2011

liquid equilibria (VLLE)

triglycerides

a b s t r a c t

Biodiesel synthesis by ethanolysis of trature can lead to high levels of converor vapourliquidliquid equilibrium aorder to determine the range of presequilibrium of the triolein and ethanmonitored in a view cell in order to195 C (the corresponding pressure wASPENplus and UniSim software awell with the experimental data. Simrium on the reaction mechanism andsince the two liquid phases exist at threactants. In case of single reaction phat high conversion levels due to incre/ locate /supf lu

ol at high pressure andlysis

. Orlovic

erides conducted under high pressure and at elevated temper-nd biodiesel yield over short reaction time. The vapourliquidhase distribution of ethanol and triolein were investigated innd temperature required for high oil conversion. The phasexture (1:40 and 1:26 molar ratio of triolein to ethanol) wasrmine the phase transition during heating of mixture up to5 and 26 bar). The experimental data were simulated usingsults obtained using RK-Aspen EOS were found to correlaten results show the important inuence of the phase equilib-ll kinetics under subcritical conditions (T < 270 C at 200 bar)inning of reaction, thereby limiting the contact between theT > 270 C at 200 bar) the initially high reaction rate is limited

extent of reversible reaction.

A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8 3

In this study, the phase equilibrium of ethanol and triglycerideswas determined experimentally using the mixture of ethanol andvegetable oil (1:40 and 1:26 molar ratio of triolein to ethanol)at 195 C (the corresponding pressures were 25 and 26 bar). Theexperimentplus and Umum proctemperaturreaction kinusing recen

2. Experim

2.1. Materi

AnhydroMerck and by a local phexane, all The GC stastearic, oleiolein, diole(Saint Louis

2.2. Compo

For the each samplchromatogrFID detecto(5 m 0.53in n-hexane1 l. The inwas heatedmally at 501 min isoth4 C/min, 3340 C withdetector temow of 2.2 c

The ethyquantitativof FAEE andprepared anglycerides p

2.3. Appara

The obseformed usinG) [4]. The ature of 20shown in Fioil and ethawas graduathen kept aied tempeby CCD camthe top (vapfor composesters durining system.correspondfrom the cocompositioslight decre

F

ringe dev

samnst Hthe s

brinsampplin

ted fre, pessenge y, theng pawn. Therim

a we

mula

sim Plusilibr

ts, foic faccid ible ot is thglyculation (triolein), as well as glycerol mono- and di-oleatetially converted trioleins. RK-Aspen thermodynamic modelsed as it was previously determined to be the best one

similar system containing triolein and methanol at nearl conditions [4,5]. The rigorous ash option for VLL equi-was also used in the simulation. The ConstantinouGanicontribution method was used for estimation of thermody-properties [6,7]. The method was applied to the followingal and thermodynamic properties of pure compounds: thel boiling point, the normal melting point, the critical pres-he critical temperature, the critical volume, the standardpy of vaporization at 298 K, the standard Gibbs energy,ndard enthalpy of formation at 298 K, the acentric factoruid molar volume at 298 K. The results are presented in.al data were correlated using RK Aspen EOS (ASPENniSim software packages) in order to identify opti-

ess conditions over broader range of pressure ande. The inuence of phase equilibrium on ethanolysisetics at high pressure and temperature was analysedtly published data.

ental

als

us ethanol (A.R. grade, >99.5%) was purchased fromfood grade, rened edible sunower oil was suppliedroducer (Servo Mihalj, Zrenjanin). Chloroform and n-of GC grade, were obtained from Fluka (Dublin, Ireland).ndards for FAEE containing ethyl esters of palmitic,c, linolenic and linoleic acids (20% of each ester), tri-in and monoolein were purchased from SigmaAldrich, USA).

sition of reaction mixture

purpose of quantitative analysis, the composition ofe of the reaction mixture was estimated by GC. Gasaphic instrument (Varian 3400) was equipped with ar, on-column injector and fused silica capillary column

mm lm thickness 0.5 m). The samples were dissolved and chloroform and injected in column in amount of

jector temperature was set at 330 C, while the column according to a temperature program: 5 min isother-C, then heating from 50 C to 110 C with 50 C/min,ermally at 110 C, heating from 110 C to 170 C with

min isothermally at 170 C, heating from 170 C to 20 C/min and isothermally at 340 C for 15 min. Theperature was 345 C. The carrier gas was nitrogen withm3/min.l palmitate was used as an internal standard for the

e identication. By application of the standard mixture the standard for glycerides, the calibration curves wered used for the quantication of the FAEE and variousresent in the analysed samples.

tus and procedures

rvation of phase transition and equilibrium was per-g high pressure view chamber (Eurotechnica PD-E350maximum working pressure was 350 bar and temper-0 C. The schematic presentation of view chamber isg. 1. The view cell was led with mixture of vegetablenol up to 3/4 of cells volume (20 ml). The temperaturelly increased to 195 C (heating time about 10 min) andt isothermal conditions for 1 h. Upon reaching the spec-rature the phase distribution in the cell was recordedera. At the same time samples were withdrawn fromour phase) and bottom part (liquid phase) of the cell,ition analysis. The evaporation of ethanol and ethylg sampling was suppressed by an adequate condens-

The observed phase distribution was calculated usinging image. Ethanol content was determined by removalllected samples using Rotary Vacuum evaporator. Then of reaction mixture was obtained by GC analysis. Onlyase of pressure and temperature was detected inside

cell duaverag

Thesel (Erusing out bytional for samcalculaperatularge vno chaquentlsampliwithdrappliedthe exptal datphase.

2.4. Si

TheASPENfor equponenacentroleic avegetasince ierties, the simfor parwas ufor thecriticalibria group namic physicnormasure, tenthalthe staand liqTable 1ig. 1. Schematic diagram of the experimental apparatus.

sampling. The experiment was repeated twice withiation of less than 1.51%.e experiment was repeated in the high pressure ves-aage), applying previously described procedure [5],

taticanalytic method. The experiments were carriedging the system to equilibrium conditions. The addi-ling tube with varying length was added to the reactorg of the intermediate phase. The exact height wasrom the phase observation in view cell at desired tem-ressure and composition in the reactor. Due to thel volume (2 dm3) and slow sampling (5 min), there wasof the system pressure and temperature and, conse-

system phase equilibria remained undisturbed by therocedure. Once the samples of each phase have been

from the vessel, gravimetric and GC analysis weree reproducibility of the method was tested by repeatingent twice and the average deviations of the experimen-

re: 2.78% for the liquid phases and 0.09% for the vapour

tion of phase equilibrium

ulation of phase equilibrium was performed using v2006 and UniSim software. The procedure usedium simulation was performed by dening the com-llowed by calculation of the critical parameters andtor using the group contribution method [4,5]. Sinces one of the most abundant fatty acids in majority ofils (canola, rapeseed, soybean and sunower oil) ande component with well dened thermodynamic prop-erol tri-oleate was chosen as the key component in

4 A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8

Table 1The physical and thermodynamical parameters of pure components used in the simulation.

Parameter Triolein Diolein Monoolein Ethyl oleate Ethanol Glycerol

M g/mol 885.45 621.00 356.55 310.51 46.07 92.09TBVB Vl (298 K) TCPCVC ZC 4 Gf (298 K) + 05 Hf (298 K) + 05 Hv (298 K) + 04

3. Results

3.1. The phapressure and

The analmation of etemperaturful kinetic mof di- and together wreaction, mtion. In thismixture (1:itored in a during heatwere 25 aninvestigatedshown in Fminimize tbetween et25 bar was ing the condesired temstopped and

At the band oil) arecell was slivapour phatity. The inand triglycephase is thtem and emimages in Frich phase vegetable oslightly incAfter heatinand 26 bar)in reactionof mono- aintermediat100 C). At rich phase equilibriumthe top pha(L3) sample2 dm3) usintal error samas well. MeTable 2.

L1 2) wnol dl to ohasel estd m

steri phasor m), thety of5.90 mses i

ulatiobtaed ex2.39phas:1 mons the 4viatiositioanol

Aspdyn

on ase [4]-comgate.h the

ethainedhe pf elelein

fromK 827.40 765.03 674.82 m3/kmol 2.708 1.106 0.533m3/kmol 0.974 0.677 0.378K 977.88 920.20 835.06 bar 3.34 5.05 10.56 m3/kmol 3.250 2.252 1.254

0.107 0.200 0.2041.9782 1.7632 1.532

kJ/kmol 2.23E + 05 3.00E + 05 3.77EkJ/kmol 1.78E + 06 1.38E + 06 9.78E

kJ/kmol 1.07E + 05 6.83E

and discussion

se equilibria of ethanolvegetable oil system at high temperature

ysis of phase equilibrium and composition during for-thyl esters of fatty acids (FAEE) at high pressure ande is very important for describing an applicable and use-odel which might be used for reactor design. Formationmono-glycerides during ethanolysis of triglycerides,ith FAEE and glycerol, which are the nal products ofakes this system very complex and difcult for simula-

study, the phase equilibrium of the triolein and ethanol40 and 1:26 molar ratio of triolein to ethanol) was mon-view cell in order to determine the phase transitioning of mixture up to 195 C (corresponding pressuresd 26 bar). The phase transition during heating of the

reactive mixture is shown in Fig. 2. Each of the imagesig. 2 was taken in a separate experiment in order tohe disturbance of phase equilibrium due to reactionhanol and triglycerides. The time to reach 195 C andthus kept at minimum (around 20 min) thereby reduc-version to FAEE and intermediates. Upon reaching theperature and pressure in the view cell the heating was

the system was left for 510 min to stabilize.eginning of heating, the two liquid phases (ethanol

in equilibrium. At that moment the pressure in viewghtly above the vapour pressure of ethanol and thusse of pure ethanol was not present in a remarkable quan-itially present intermediate phase containing ethanolrides gradually disappears since the existence of thise consequence of addition of components to the sys-ulsion like behaviour of ethanol and triglycerides (rstig. 2a and b). Upon heating, the amount of ethanol(L1) increases while at the same time the volume ofilethanol phase (the second stable phase L2) onlyreases due to dissolution of ethanol in vegetable oil.g of up to 195 C (corresponding pressures were 25 bar, a small amount of FAEE was formed, around 3 mass%

mixture (Table 2). With relatively small quantities

TheTable of ethaethanodiate pof ethyerol antransetion ofratio. F195 Cquantitions (L3 phaused.

Simsition detect1.38%, esters the 26deviatiwhen age decompofor eth

Thethermotributimixturpseudoinvesti(Fig. 3)

Witcriticalin obtamine trange othe trioaturesnd di-glycerides, FAEE was mainly concentrated ine L3 phase (ethanolFAEE phase initially appearing at195 C and 25 or 26 bar, three liquid phases: ethanol(L1), TGethanol (L2) and ethanolFAEE (L3) were in

(LLL). Two samples were taken from the view cell,se (L1) and bottom phase (L2). The intermediate phase

was collected from the reactor system (Ernst Haageg sampling tube. In order to minimize the experimen-ples of L1 and L2 were taken from the 2 dm3 reactor

asured compositions of L1, L2 and L3 are shown in

procedure oequilibria slished as thof triolein a

The resuthe systemonly one suatures. Theand trioleinof vapour pthe liquid p617.00 351.44 561.000.489 0.063 0.0870.341 0.058 0.073

764.00 514.00 850.0012.80 61.37 75.001.060 0.168 0.2640.214 0.241 0.0871.0494 0.6436 0.5127

1.17E + 06 1.68E + 05 4.47E + 056.26E + 05 2.35E + 05 5.78E + 056.36E + 04 3.89E + 04 6.57E + 04

phase consisted of almost pure ethanol (>99 mass%,hile the bottom phase (L2) contained certain amountissolved in oil (2126 mass%), depending on the initialil molar ratio. The third liquid phase (L3, the interme-) consisted mainly of ethanol, in which some amounters (FAEE) were dissolved together with traces of glyc-onoglycerides. The latter were also generated by thecation reaction. As indicated in Table 2 the composi-es was almost independent of the ethanol to oil molarolar ratio of ethanol to oil of 40 (Fig. 2b, frame taken at

intermediate phase (L3) increases in volume but low FAEE is dissolved in this phase under specied condi-ass% of L3). The FAEE distribution between the L2 and

s different when 26:1 molar ratio of ethanol to oil was

on results for phase equilibrium and phase compo-ined using Aspen software correspond well to thoseperimentally (Table 2), with the standard deviation of% and 2.62% for ethanol rich phase (L1), ethanolethyle (L3), and oilethanol phase (L2) respectively, whenolar ratio of ethanol to oil was used. The standardwere 0.52%, 1.19% and 2.78% for L1, L2 and L3 phases0:1 molar ratio of ethanol to oil was used. The aver-n between experimental and calculated values of phasen was around 0.5% for the ethanol phase and 1.53.0%ethyl esters phase and oilethanol phase.en simulation, with the appropriate physical andamic parameters, correlates accurately the phase dis-

well as a component distribution for such complex. RK-Aspen EOS and data for triolein as vegetable oilponent were used to obtain the Pxy diagram of the

d system for a temperature range from 100 to 400 C

aim to understand the kinetics of subcritical and super-nolysis and in order to be able to explain the difference

yields at these conditions, it is also necessary to deter-hase equilibria of the oil and ethanol over the broadervated pressure and temperature. The phase equilibria ofand ethanol mixture at 200 bar and for different temper-

200 to 400 C were calculated using the ash separation

f ASPENplus and UniSim software. The ash type VLLystem was analysed with RK-Aspen EOS already estab-e most appropriate thermodynamic model in the caselcoholysis [4].lts obtained by this simulation show that below 270 C,

consists of the two phases in equilibrium (Fig. 4) andpercritical or dense phase is present at higher temper-

similar phase equilibrium was observed for methanol as shown in Fig. 4 [4]. Ethanol is the main constituenthase (more than 99.9 mol%) and content of ethanol inhase, i.e. in triolein, increases with temperature. The

A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8 5

Fig. 2. The phase transition during the hating of mixture of vegetable oil and ethanol up to 195 C and 26 bar with ethanol to oil molar ratio of 1:26 (a) and 1:40 (b).

Table 2The chemical composition of the existing phases at 195 C.

Ethanol to oil molar ratio Phases Mass% of phase The mass% of component in the phase

Glycerol EtOH FAEE MG DG TG

Experimental data1:2625 bar

Liquid 1 51.49 0.02 99.48 0.50 0 0 0Liquid 2 36.50 0.20 20.51 4.37 1.67 0.60 72.66Liquid 3 12.01 0.64 83.02 11.36 5.04 0.45 0

1:4026 bar

Liquid 1 32.62 0.03 99.37 0.60 0 0 0Liquid 2 38.54 0.19 23.17 1.80 0.41 1.17 73.27Liquid 3 28.84 0.76 89.51 5.90 3.82 0.01 0

Simulation data1:2625 bar

Liquid 1 50.79 0.09 99.81 0.14 0 0 0Liquid 2 37.48 0.13 20.38 3.36 1.04 1.49 73.60Liquid 3 11.73 1.88 75.89 14.39 7.74 0.09 0.01

1:4026 bar

Liquid 1 32.39 0.03 99.91 0.05 0 0 0Liquid 2 37.89 0.05 22.21 1.60 0.50 1.10 74.54Liquid 3 29.72 0.83 88.92 6.62 3.59 0.04 0

6 A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8

Fig. 3. The Pxy diagram for triolein and ethanol in temperature interval100400 C.

Fig. 4. VLE at supercritical p

similar datamethanol tosure (Fig. 4phase) andcase of mepressure.

The distrature increa200 bar is pmight be exsince ethan

Fig. 5

Fi

thatangen [4percnol aroxime (vaexpectwill chpositioand sumethaby appmixturdifferent temperatures and content of ethanol in vapour, liquid andhase at 200 bar (molar ratio of methanol and triolein of 42:1).

was obtained for methanol and triolein system (40:1 triolein molar ratio) at the same temperature and pres-). However, the content of ethanol in oil phase (liquid

oil in ethanol (vapour phase) are both lower than inthanoltriolein system at the same temperature and

ibution of vapour and liquid phases during the temper-se from 200 to 400 C under the constant pressure ofresented in Fig. 5. This is an idealized situation whichpected if triolein and ethanol do not react. However,olysis will occur under investigated conditions one can

. The phase distribution in mass% at different temperatures.

and triolein

3.2. Ethano

Ethanol 20 bar lowelar structurgroup, therphysical prinvestigateunder supereactivity oethanol in to obtain thtime for thcase of metobtained atFig. 3] whilepressure ortemperaturfor methanyield of estthe yield ofThe processFAEE or conysis of vegeapplicationof ester proof vegetabldeterminedof approxim(40:1), as ca

The obsmethanolysof solubilitysive energyand vegetabg. 6. The temperature dependence of density at 200 bar.

formation of different compounds during ethanolysis the mass ratio between phases, as well as their com-]. The calculated values of densities for vapour, liquidritical phase are presented in Fig. 6. The densities ofnd oil mixture, for both liquid and vapour phase, areately 1020% higher than densities of ethanol and oil

pour and liquid phase), for 40:1 molar ratio of alcohol.

lysis of triglycerides

has a slightly higher critical temperature but almostr critical pressure than methanol. Since the molecu-es of ethanol and methanol differ only by one methyle is no substantial difference between the chemical andoperties of these alcohols. Lower alcohols have beend as the source of alcohol for production of biodieselrcritical conditions [811] and it was found that thef alcohol decreases when ethanol is used. By usingthe reaction of vegetable oil alcoholysis, it is possiblee FAEE yield higher than 90 mass%, but the necessary

is level of conversion is three times longer than in thehanol. It is obvious that much lower reaction rates are

subcritical conditions of alcohol [10, Figs. 2 and 11, the reaction rate could be gradually increased as either

temperature rises. Moreover the increase of reaction

e and pressure, especially to and above critical valuesol or ethanol, results in a favourable inuence on theer [10,11]. The effect of various process parameters on

FAEE was studied by several researchers [1,2,1014]. parameters studied, are shown in Table 3. The yield ofversion of vegetable oil during methanolysis or ethanol-table oil depends also on the type of vegetable oil. The

of saturated vegetable oils resulted in the higher yieldduct. The optimum process conditions for ethanolysise oils producing FAEE, with 9095 mass% yields, were

as: temperature in the vicinity of 300 C, reaction timeately 30 min and high molar ratio of ethanol to oiln be seen in Table 3.erved differences in reaction rates of triglyceridesis and ethanolysis could be explained by the values

parameters, dened as the square root of the cohe- density. Solubility parameters for ethanol, methanolle oils are available in standard handbooks [16]. Since

A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8 7

Table 3Review of process parameters used for ethanolysis of vegetable oil under supercritical conditions and production of FAEE.

References Type of oil Optimum reaction conditions Yield (mass%)

Pressure (bar) Temperature (C) Time (min) EtOH:oil molar ratio

Warabi et al., 2004 [1] Rapeseed 165a 300 45 42 99.0Madras et al., 2004 [12] Sunower 200 350 40 40 99.0Varma and Madras, 2005 [13] Linseed 200 300 100 40 99.0

Castor 200 300 100 40 96.0Vieitez et al., 2008 [14] Soybean 200 350 42 40 77.5Demirbas, 2008 [10] Cottonseed 200 250 8 41 85.0Demirbas, 2009 [11] Linseed 200 250 8 41 88.0Gui et al., 2009 [2] Palm 295a 349 30 33 79.2Valle et al., 2010 [15] Fodder radish 319 22 39 97.5

a Calculated by the vapour pressure.

the solubility parameter for methanol (29.7 MPa1/2) is higher thanfor ethanol (26.2 MPa1/2) and more similar to solubility parame-ter of oil (37.4 MPa1/2), the mutual solubility of oil in methanoland consequently conversions in methanol, are higher comparedto those obtained in ethanol. Contact of reactants for the systemalcoholtriglyceride certainly plays a crucial role in increasing reac-tion rate and it can be more quantitatively described by phaseequilibrium at investigated pressures and temperatures.

Ethanolysis reaction proceeds in three consecutive steps andeach of thereaction is mare convertthe intermesmall amoumonoglycerstage, the rglycerol. Ba[4] it can binuence dconversion rium or coep,T conditiophase concthe phases reaction). Stal data shothe rst or kinetic analconducted peratures o(200400 C

rst order kinetics in a batch reactor and in Fig. 7b the same isshown for irreversible second order kinetics. In the initial phaseof reaction at subcritical conditions of ethanol (200 C at 200 bar),the mass transfer of reacting species plays a signicant role in lim-itation of the overall reaction rate so the expected second orderreaction kinetics turns into rst order kinetics (rst order kineticsat 200 C in Fig. 7a better ts the experimental data than secondorder kinetics shown in Fig. 7b). This fact is conrmed by the molarratio of oil to ethanol in oil rich phase (L2 in which the reaction

lacen dad by

phainethow

reacperimThe smentond MEtOe (aboden ph

rsxperseqution.

to r exp, form steps is reversible reaction. Initial phase of the overallainly resulting from the rst step in which triglycerides

ed to diglycerides and fatty acid ethyl esters (FAEE). Indiate phase diglycerides, unconverted triglycerides andnt of monoglycerides are being converted to mainlyides, small amount of glycerol and FAEE. In the nalemaining monoglycerides are converted to FAEE andsed on the analogy with methanolysis of triglyceridese concluded that pressure and temperature levels willifferent phase distribution scenarios with increasingduring reaction. Existence of vapourliquid equilib-xistence of one vapour and two liquid phases at certainns will inuence the reaction rate through local singleentrations, mass transfer of reactive species betweenand reaction equilibrium (the magnitude of reversibleimplied kinetic analysis of the published experimen-wn in Table 3 [1,2,1014] was performed, assuming

second order reaction kinetics for overall reaction. Theysis for the rst and second order irreversible reaction,in a batch reactor, is shown in Fig. 7 for several tem-f interest for high pressure non-catalytic ethanolysis). In Fig. 7a the kinetic data are tested for irreversible

takes ppositiorequireinto L2order k(data sfor thethe exin L2. experifor secratiomixturrium mreactiopseudowith ethe conof reac

DuekineticsystemFig. 7. Analysis of the experimental data [1,13] using the pseudo rst orde), which has a value of 1:2 at 200 C (liquid phase com-ta is shown in Fig. 4). This excess of ethanol is less than

the reaction stoichiometry and mass transfer of ethanolse limits the overall reaction resulting in apparent rstics. At 250 C the oil to ethanol ratio in L2 phase is 1:16n in Fig. 4) which provides sufcient excess of ethanoltion. Consequently the rst order kinetics can describeental data due to more than vefold excess of ethanolecond order kinetics are also in good agreement withal data at 250 C since the mole balance equation usedorder kinetics contains local L2 ethanol to oil molarH. At conditions of supercritical ethanoltriglyceridesove 270 C at 200 bar as predicted by the phase equilib-l calculations, Figs. 4 and 5), corresponding to the singlease and the absence of mass transfer limitations, thet order reaction kinetic model is in better agreementimental data than second order kinetic model. This isence of high ethanol to oil molar ratio at the beginning

elatively high ethanol to oil ratio used in the analysederiments and increasing concentration of FAEE in theation of a single uid phase could be expected at higherr (a) and second order (b) kinetic models.

8 A.M. Almagrbi et al. / J. of Supercritical Fluids 61 (2012) 2 8

triglycerides conversions [4]. Moreover, the increasing amount ofmonoglycerides, diglycerides and glycerol at higher conversionstrigger the reversible reaction and the assumed model (irreversiblerst or second order) can not t well with the experimental data(data at 300 C for 100 min in Fig. 7). This trend occurs after 70% oftotal triglycerides conversion and corresponds to the appearance ofthe initial amount of glycerol in the system (Fig. 7a and b, and [4]).Similar deviations could have been observed at other temperaturesif the higher conversion levels had been obtained in the analysedexperiments (200 C and 250 C). This kinetic analysis shows that itis not possible to describe such complex system by simple kineticmodels and that the phase distribution and the phase compositionplay an important role in the overall reaction kinetics.

4. Conclusion

The phase equilibria and kinetics of triglycerides ethanolysis atelevated pressure and temperature were analysed in this study.Phase equilibrium data were obtained using a high pressure viewcell followed by GC analysis of the collected samples. The sim-ulation of phase distribution using RK-Aspen equation of state,correlated well with the experimental data with high precision incalculation of a phase and component distribution for the inves-tigated complex system. The results also indicate that the phaseequilibrium of the system plays an important role in the reactionmechanism and kinetics of the non-catalytic synthesis of biodieselunder conditions of high pressure and temperature. Simulation ofphase distribution results conrms that the rate of triglyceridestransesterication, or overall reaction rate, depends strongly onthe oil to ethanol ratio in the oil rich phase (corresponding to thetwo phase system below 270 C at 200 bar). The reversible reactiontakes place if one phase system exists at the end of reaction, therebydecreasing

Acknowled

The naof the Repu

acknowledged, as well as the support of CHEN Department at TexasA&M University at Qatar in use of Aspen software. The authorswould like to express their gratitude to Prof. D. Skala for his supportin the experimental work on this study.

References

[1] Y. Warabi, D. Kusdiana, S. Saka, Reactivity of triglycerides and fatty acidsof rapeseed oil in supercritical alcohols, Bioresource Technology 91 (2004)283287.

[2] M.M. Gui, K.T. Lee, S. Bhatia, Supercritical ethanol technology for the produc-tion of biodiesel: process optimization studies, J. Supercritical Fluids 49 (2009)286292.

[3] D. Wen, H. Jiang, K. Zhang, Supercritical uids technology for clean biofuelproduction, Progress in Natural Science 19 (2009) 273284.

[4] S.B. Glisic, D.U. Skala, Phase transition at subcritical and supercritical conditionsof triglycerides methanolysis, J. Supercritical Fluids 54 (2010) 7180.

[5] S. Glisic, O. Montoya, A. Orlovic, D. Skala, Vaporliquid equilibria oftriglyceridesmethanol mixtures and their inuence on the biodiesel synthe-sis under supercritical conditions of methanol, J. Serbian Chemical Society 72(2007) 1327.

[6] R. Gani, L. Constantinou, Molecular structure based estimation of properties forprocess design, Fluid Phase Equilibria 116 (1996) 7586.

[7] R. Gani, B. Nielsen, A. Fredenslund, A group contribution approach to computer-aided molecular design, AIChE J. 37 (1991) 13181332.

[8] A. Demirbas, Recent developments in biodiesel fuels, International J. GreenEnergy 4 (2007) 1526.

[9] D. Kusdiana, S. Saka, Effects of water on biodiesel fuel production bysupercritical methanol treatment, Bioresource Technology 91 (2004) 289295.

[10] A. Demirbas, Studies on cottonseed oil biodiesel prepared in non-catalytic SCFconditions, Bioresource Technology 99 (2008) 11251130.

[11] A. Demirbas, Production of biodiesel fuels from linseed oil using methanoland ethanol in non-catalytic SCF conditions, Biomass and Bioenergy 33 (2009)113118.

[12] G. Madras, C. Kolluru, R. Kumar, Synthesis of biodiesel in supercritical uids,Fuel 83 (2004) 20292033.

[13] M.N. Varma, G. Madras, Synthesis of biodiesel from castor oil and linseed oilin supercritical uids, Industrial & Engineering Chemistry Research 46 (2007)

.ieitezchmaturesalle, Aercritercrit.M. Bass, Flothe overall reaction rate.

gements

ncial support by the Ministry of Education and Scienceblic of Serbia through Grant No. 45019 is gratefully

16[14] I. V

I. Jamix

[15] P. VsupSup

[16] A.FPre, C. Da Silva, G.R. Borges, F.C. Corazza, J.V. Oliveira, M.A. Grompone,nian, production of soybean biodiesel in supercritical ethanolwater, Energy Fuels 22 (2008) 28052809.. Velez, P. Hegel, G. Mabe, E.A. Brignole, Biodiesel production using

ical alcohols with a non-edible vegetable oil in a batch reactor, J.ical Fluids 54 (2010) 6170.rton, Handbook of Solubility and Other Cohesion Parameters, CRCrida, 1985.

The phase equilibrium of triglycerides and ethanol at high pressure and temperature: The influence on kinetics of ethanolysis1 Introduction2 Experimental2.1 Materials2.2 Composition of reaction mixture2.3 Apparatus and procedures2.4 Simulation of phase equilibrium

3 Results and discussion3.1 The phase equilibria of ethanolvegetable oil system at high pressure and temperature3.2 Ethanolysis of triglycerides

4 ConclusionAcknowledgementsReferences