Fuel 108 (2013) 269–276

Contents lists available at SciVerse ScienceDirect

Fuel

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

Liquid–liquid and vapor–liquid equilibrium data for biodiesel reaction–separationsystems

Diogo I. Segalen da Silva a, Marcos R. Mafra a, Fabiano Rosa da Silva b, Papa M. Ndiaye a, Luiz P. Ramos b,Lucio Cardozo Filho c, Marcos L. Corazza a,⇑a Department of Chemical Engineering, Federal University of Paraná (UFPR), CEP 81531-990, Curitiba, PR, Brazilb Department of Chemistry, Federal University of Paraná (UFPR), CEP 81531-990, Curitiba, PR, Brazilc Department of Chemical Engineering, Maringá State University (UEM), CEP 87020-900, Maringá, PR, Brazil

h i g h l i g h t s

" We measured phase equilibrium data for the system involving biodiesel, methanol, ethanol, glycerol and soybean oil." VLE and LLE data were measured for binary and ternary systems." The boiling point temperatures were obtained using Othmer-type ebulliometer.

a r t i c l e i n f o

Article history:Received 2 April 2012Received in revised form 18 February 2013Accepted 25 February 2013Available online 13 March 2013

Keywords:BiodieselSoybean oilVapor–liquid dataLiquid–liquid dataMethanol

0016-2361/$ - see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.fuel.2013.02.059

⇑ Corresponding author. Tel.: +55 41 3361 3587.E-mail address: corazza@ufpr.br (M.L. Corazza).

a b s t r a c t

This work reports experimental vapor–liquid and liquid–liquid equilibrium data for binary and ternarysystems comprised of various mixtures of biodiesel (from soybean oil), methanol, ethanol, glycerol andsoybean oil. The binodal curves for biodiesel + methanol + glycerol, biodiesel + ethanol + glycerol, biodie-sel + methanol + soybean oil and biodiesel + ethanol + soybean oil systems were obtained at two differenttemperatures by titration. An Othmer-type ebulliometer was used for vapor–liquid equilibrium measure-ments at pressures ranging from 14.0 kPa to 92.0 kPa. Binodals curves (LLE) indicated that the tempera-ture range used in this study has no effect on the immiscibility region for thebiodiesel + methanol + soybean oil system. By contrast, for the biodiesel + ethanol + soybean oil system,the immiscibility region was larger at the lower temperature. For the vapor–liquid equilibrium data,the results showed that the addition of glycerol to the biodiesel + ethanol binary system does not signif-icantly affect the boiling temperature of the system while with the addition of methanol to the biodie-sel + soybean oil system the boiling temperatures of the ternary mixture are reduced. These results canbe used to enhance the reaction conversion and the purification processes associated with biodieselproduction.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Due to environmental problems and economic issues related tothe use of fossil fuels, intensive research has been carried out in thepast decade with the main objective of developing renewable andeconomically sustainable alternative energy sources. Derivatives ofvegetable oils and animal fats, known as biodiesel, obtainedthrough chemical transformation processes constitute an interest-ing alternative to fossil fuels, in so far as their use contributes toreducing emissions of the main gases related to global warmingand the dependence of certain countries on diesel [1,2].

ll rights reserved.

Biodiesel can be produced through the transesterification ofvegetable oils and animal fats or the esterification of free fattyacids by means of acid, alkaline or enzymatic catalysis, in homoge-neous or heterogeneous media [3–5]. An analysis of the BrazilianProgram for Production and Use of Biodiesel (PNPB) shows thatsoybean oil is the main raw material used for this purpose,accounting for 77.90% of the national production, compared with3.50% for cottonseed oil, 16.15% for beef tallow and only 2.45% ofother fatty materials [6].

Due to factors related to process simplification, biodiesel iscommonly produced by homogenous catalysis using methanoland NaOH as catalyst. Purification is carried out in a sequence ofunit operations which include removal of the free glycerol, excessalcohol and residual catalyst. The purification process includes sev-eral techniques, for instance, washing with distilled water and acid,

270 D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276

washing through a solid adsorbent, extraction with organic solventlike ether or hexane, and water and/or physical separation usingmembranes [7]. When ethanol is used instead of methanol, purifi-cation constitutes an important step in obtaining a product withmarket specifications [8,9]. Regardless of the process employed,knowledge of the phase diagram for the multicomponent systemis essential to designing an adequate, suitable and reliable processfor biodiesel production. Thus, obtaining the phase diagrams of thereaction medium represents a fundamental step for the optimiza-tion of the purification process. Liquid–liquid and vapor–liquidequilibrium (LLE and VLE, respectively) involving alkyl esters,methanol, ethanol and glycerol systems have been published inthe literature in recent years [10–12]. However, pure alkyl esterswere used in most of these studies and they did not take into ac-count the fact that biodiesel, in general, is a mixture of 5–10 ormore alkyl-esters. Only few studies available in the literature arerelated to the use of complete biodiesel samples (such as those ob-tained from soybean oil by transesterification) [13,14].

In this context, the main objective of this work was to enhancethe experimental databank by providing information on the VLEand LLE related to binary and ternary systems that could be in-volved in biodiesel production and purification processes. Thesesystems include biodiesel + ethanol, biodiesel + methanol, biodie-sel + ethanol + glycerol, biodiesel + methanol + glycerol, biodie-sel + methanol + soybean oil and biodiesel + ethanol + soybean oil.

2. Materials and methods

2.1. Chemicals

Methanol (0.998 mass fraction purity) and ethanol (0.998 massfraction purity) were supplied by Synth (São Paulo/SP/Brazil), andglycerol (P0.995 mass fraction) by Sigma–Aldrich (São Paulo, SP,Brazil). Soybean oil (Liza, Cargill Agrícola S/A, Mairinque, SP) waspurchased in a local store. All chemicals were used without furthertreatment.

The biodiesel (fatty acid methyl esters-FAMEs) was obtainedusing soybean oil by alkali-catalyzed transesterification with anoil to methanol molar ratio of (1:12), 0.5% of NaOH (mass basis)for the formation of alkoxide and 500 rpm at 65 �C. The reactiontime was 1 h. The methanol was removed by evaporation and theester phase was washed twice with hot water at 50 �C and thendried. The esters were then purified using a solid adsorbent(Perlimax�) with agitation for 30 min at 65 �C. The material wasthen filtered to remove solids and the biodiesel (esters) was driedusing anhydrous sodium sulfate.

After purification, the biodiesel by gas chromatography (GC)using the ASTM D 15342 standard method. The GC analysisshowed an ester content of 96.6 wt% while contaminants were par-tially characterized as free glycerol (0.153 wt%), triacylglycerols

Table 1Characterization analysis of biodiesel synthesized in this work.

Analysis Value Unit Method

Ester content 96.60 wt% ABNT NBR 1Acidity index 0.45 mg KOH g�1 ABNT NBR 1Glycerol total 0.153 wt% ABNT NBR 1Free glycerol 0.001 wt% ABNT NBR 1Monoacylglycerol 0.402 wt% ABNT NBR 1Diacylglycerol 0.308 wt% ABNT NBR 1Triacylglycerol 0.018 wt% ABNT NBR 1Methanol 0.50 wt% ABNT NBR 1Specific mass at 20 �C 0.877 kg m�3 ABNT NBR 7Water content 1198 mg kg�1 ASTM D 630Soap content 66.30 Sodium oleate % AOCS OfficiaMolecular weight 292.2 g gmol�1 National Bio

(0.018 wt%), monoacylglycerols (0.402 wt%), methanol(0.500 wt%) and water (0.0012 wt%). After characterization, themethyl esters were transferred to an amber flask, flushed withN2 and stored in the dark for no longer than 60 days until use inthe phase equilibrium experiments. Table 1 presents the resultsfrom the characterization of biodiesel produced in this work.

The soybean oil used for biodiesel production was analyzed byAOCS Method Ce 1e-91 and the fatty acid composition was linoleic53 wt%, oleic 23 wt%, palmitic 11 wt%, stearic 4 wt%, linolenic8 wt% and other minor fatty acids 1 wt% [15,16].

2.2. LLE apparatus and procedures – binodal curves

Binodal curves were determined considering the cloud pointobtained using the titration method under isothermal conditions.The experiments were performed in a liquid–liquid equilibriumjacketed-cell with temperature controlled by an ultrathermostaticbath. For the temperature monitoring inside the cell, a thermocou-ple with an uncertainty of around 0.5 K was used. The solution in-side the cell was kept continuously mixed using a magnetic stirrerto allow complete homogenization of the system. Fig. 1 shows theexperimental scheme used in this work. The binodal curves wereobtained from the mass fraction (w) and presented in triangulardiagrams.

The binodal or equilibrium curves for the ternary systems, atconstant temperature, were determined by the titration of thetwo-component mixture (known concentrations) with a thirdcomponent until clouding of the solution. This is also known asthe turbidity point technique. Initially, known quantities of twomiscible components were weighed on an analytical balance witha precision of 0.0001 g (Marte, model AM220) and added directlyinside the equilibrium cell following the procedure used by Silvaet al. [17], Stragevitch and D’Ávila [18] and Ardila et al. [19]. Thetwo-component mixture is kept in a closed jacketed bottle to pre-vent evaporation. The device is connected to a water circulatingbath to allow rigorous temperature control during the experi-ments. Titration is then started by adding the third componentthrough an orifice at the top of the equilibrium cell and the solu-tion is continuously mixed by a magnetic stirrer until the visualappearance of a turbid mixture. The added volume is then recordedand the resulting mixture composition can be computed to obtaina point on the binodal curve. This procedure is repeated by chang-ing the solution composition until the obtainment of a completebinodal curve.

2.3. VLE apparatus and procedures

The vapor–liquid equilibrium data were measured using anOthmer-type ebulliometer as shown in Fig. 2, and presented inthe previous work [20]. The proposed all-glass ebulliometer is a

5342/EN 141034448/ASTM D 664/EN 141045344/ASTM D 6584/EN 141055341/ASTM D 6584/EN 14105, EN 141065342, ABNT NBR 15344/ASTM D 6584/EN 141055342, ABNT NBR 15344/ASTM D 6584/EN 141055342, ABNT NBR 15344/ASTM D 6584/EN 141055343/EN 14110148, ABNT NBR 14065/ASTM D 1298, ASTM D 4052/EN ISO 3675, EN ISO 121854/EN ISO 12937l Method Cc 17–95diesel Board

Fig. 1. Experimental scheme for liquid–liquid equilibrium measurements used inthis work. (a and b) are the sampling point, (c and d) are the inlet and outlet point ofultrathermostatic bath.

Fig. 2. Schematic diagram of experimental apparatus for vapor–liquid equilibriummeasurements used in this work.

Table 2Binodal data measured for the biodiesel(1) + ethanol(2) + glycerol(3) system (massfraction).

w1 w2 w3

T = 298 K0.2302 0.5516 0.21820.0893 0.5712 0.33940.4460 0.4367 0.11730.2950 0.5225 0.18240.1529 0.5777 0.26950.5789 0.3500 0.07110.0484 0.5459 0.40570.1244 0.5815 0.29410.3501 0.4942 0.15570.1112 0.5802 0.30850.0266 0.4989 0.47450.6260 0.3164 0.0576

T = 323 K0.2515 0.5054 0.24310.0942 0.5349 0.37090.4725 0.3988 0.12880.3155 0.4783 0.20610.1868 0.5298 0.28350.0556 0.5134 0.43090.6008 0.3214 0.07780.1417 0.5371 0.32120.3865 0.4431 0.17030.0319 0.4670 0.50110.6519 0.2873 0.0608

0.0

0.2

0.4

0.6

0.8

1.0

Ethanol0.0

0.2

0.4

0.6

0.8

1.0

Biodiesel0.0 0.2 0.4 0.6 0.8 1.0

Glycerol

Fig. 3. Binodal curve for the biodiesel(1) + ethanol(2) + glycerol(3) system at 298 K(j) and 323 K ( ) from this work and literature at 298 K (h) [19].

D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276 271

modification of the Othmer-type ebulliometer in which only thevapor phase is circulated. It is basically composed of an equilib-rium cell, a condenser, two sample ports (for the liquid and the va-por phase) and some auxiliary components such as a vacuumpump, heat resistor, magnetic stirrer and temperature and pres-sure sensors. The pressure is controlled using a needle valve andwas kept constant through a vacuum line. The pressure was mea-sured with a calibrated manometer (Greisinger, model GDH 12vacuum-meter), with a precision of 0.10 kPa and reading error of5% at the pressure value. A liquid solution with the desired compo-sition was introduced into the equilibrium cell, previously pre-pared gravimetrically by weighing the required amounts of thepure liquids and stirred before loading the cell with the mixture.The total volume occupied by the liquid solution in the equilibriumcell was approximately 140 mL. After loading the cell, the vacuumwas applied gradually up to the desired total pressure and thenslow heating was applied which was increased and adjusted toproduce the required boiling under vigorous mixing by a magneticstirrer. The boiling mixture was maintained until a drop count of50 drops per minute was achieved. When the equilibrium statewas achieved, which was characterized by a constant (equilibrium)

temperature and pressure and the uniformity of the drop rate for atleast 5 min, the temperature was measured using a calibratedPT100 temperature sensor. Subsequently, a new fixed pressure va-lue was set and kept constant and the ebullition temperature wasobtained as an equilibrium data point. To change the mixture com-position, different amounts of solvent (water, ethanol or ethyl stea-rate and palmitate) were introduced into the ebulliometer and theprocedure was repeated. For binary mixtures, at the end of eachprocedure, the liquid phase was sampled and its compositionwas analyzed using refractive index measurements taken on anAbbe-type refractometer, with a composition uncertainty of0.001. The adequacy of the ebulliometer used was demonstratedby measurements of vapor–liquid equilibrium data for the binarysystem water + glycerol as presented by Coelho et al. [20].

Table 3Binodal data measured for the system biodiesel(1) + methanol(2) + glycerol(3) (massfraction).

w1 w2 w3

T = 298 K0.0291 0.4821 0.48880.0908 0.7978 0.11140.0214 0.1223 0.85640.0443 0.7446 0.21110.0086 0.2095 0.78190.0298 0.6826 0.28770.0086 0.2986 0.69270.0212 0.5955 0.38330.2940 0.6734 0.03260.0069 0.3178 0.67530.2355 0.7119 0.05260.3799 0.6005 0.01950.5793 0.4031 0.0176

T = 323 K0.1587 0.6956 0.14580.5916 0.3697 0.03870.4895 0.4558 0.05470.3285 0.5886 0.08290.7075 0.2680 0.02450.3774 0.5529 0.06980.0130 0.1036 0.88340.0926 0.7134 0.19400.0174 0.2292 0.75340.0042 0.3126 0.68330.0328 0.6687 0.29850.0085 0.5939 0.39760.0065 0.4264 0.5671

0.0

0.2

0.4

0.6

0.8

1.0

Methanol0.0

0.2

0.4

0.6

0.8

1.0

Biodiesel0.0 0.2 0.4 0.6 0.8 1.0

Glycerol

Fig. 4. Binodal curve for the biodiesel(1) + methanol(2) + glycerol(3) system at298 K (j) and 323 K ( ).

Table 4Binodal data measured for the biodiesel(1) + methanol(2) + soybean oil(3) system(mass fraction).

w1 w2 w3

T = 298 K0.4773 0.4784 0.04430.1052 0.8857 0.00910.1996 0.7824 0.01800.2906 0.6828 0.02670.3888 0.5844 0.02680.5646 0.3748 0.06060.6321 0.2809 0.08700.6475 0.1771 0.17550.6691 0.1974 0.13350.6600 0.2364 0.10360.6175 0.1552 0.22730.6093 0.1246 0.26610.6036 0.1201 0.27630.5219 0.0974 0.38070.4003 0.0659 0.53380.0310 0.0328 0.9362

T = 323 K0.5137 0.2435 0.24280.2332 0.7389 0.02800.4901 0.2037 0.30630.1835 0.1119 0.70460.4466 0.1654 0.38790.4351 0.4474 0.11750.3800 0.5510 0.06900.1239 0.0925 0.78360.4880 0.3540 0.15800.3612 0.5754 0.06340.2673 0.1206 0.61210.3373 0.1504 0.51230.5137 0.2435 0.24280.2332 0.7389 0.02800.4901 0.2037 0.3063

Table 5Binodal data measured for the biodiesel(1) + ethanol(2) + soybean oil(3) system(mass fraction).

w1 w2 w3

T = 298 K0.1620 0.6802 0.15770.2389 0.4093 0.35180.2073 0.3188 0.47400.1580 0.2177 0.62430.0875 0.1564 0.75610.0319 0.1378 0.83030.2004 0.5646 0.23500.1393 0.7482 0.11260.0816 0.8629 0.05550.0225 0.9424 0.0351

T = 323 K0.0807 0.6451 0.27410.0296 0.8418 0.12870.0536 0.7658 0.18050.0918 0.5689 0.33940.0924 0.4610 0.44670.0758 0.3432 0.58100.0615 0.2940 0.64450.0505 0.2811 0.66840.0260 0.2626 0.7114

272 D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276

3. Results

3.1. Binodal curves

Table 2 presents the measured experimental saturation point(binodal curve) for the biodiesel + ethanol + glycerol system andFig. 3 shows the ternary diagram at two temperatures (298 K and323 K) for this system.

In order to validate the solubility results (binodals) and checkthe reproducibility of the experimental apparatus and procedureused, the ternary system containing biodiesel + ethanol + glycerolat T = 298.15 K were compared to results from the literature [19],

as presented in Fig. 3. Besides, the phase transition – cloud pointdata, for this system at temperature of 323 K are also presentedin Fig. 3. As can be seen from this figure, the methodology andexperimental apparatus employed to determine the compositionsof equilibrium phases seem to be adequate since results obtainedin this work are in good agreement with those reported by Ardilaet al. [19].

0.0

0.2

0.4

0.6

0.8

1.0Biodiesel0.0

0.2

0.4

0.6

0.8

1.0

Methanol0.0 0.2 0.40 0.8 1.0

Soybean Oil0.6

Fig. 5. Binodal curve for the methanol(1) + biodiesel(2) + soybean oil(3) system at298 K (j) and 323 K (h).

0.0

0.2

0.4

0.6

0.8

1.0Biodiesel0.0

0.2

0.4

0.6

0.8

1.0

Ethanol0.0 0.2 0.4 0.6 0.8 1.0

Soybean Oil

Fig. 6. Binodal curve for the ethanol(1) + biodiesel(2) + soybean oil(3) system at298 K (j) and 323 K (h).

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0w1

300

320

340

360

380

400

420

T / K

Fig. 7. Comparison of boiling point temperature measured in this work withliterature data reported for the system ethanol(1) + biodiesel(2) at 14.0 kPa ( , thiswork; h, Guo et al. [14]) and 92 kPa (N, this work; 4, Guo et al. [14]).

Table 6Boiling point temperature for the ethanol(1) + biodiesel (2) system (mass fraction).

p (kPa) T (K) p (kPa) T (K)

w1 = 1.0000 w1 = 0.261613.8 308.20 13.9 310.5723.8 318.40 24.1 321.7733.8 325.50 34.0 329.0644.2 331.10 44.4 335.1654.2 335.60 54.5 340.2564.5 339.60 64.4 344.4574.7 342.90 74.5 348.2590.8 347.60 91.7 353.65

w1 = 0.8597 w1 = 0.106913.7 308.77 13.8 311.3724.3 319.77 23.9 322.8634.0 326.66 34.1 330.8644.3 332.26 44.2 336.9654.5 336.86 54.4 342.2564.2 340.55 64.5 346.7574.7 344.15 74.7 350.5591.2 349.05 91.4 355.94

w1 = 0.6638 w1 = 0.075913.7 309.07 13.8 315.0723.9 320.37 23.9 330.4634.3 327.86 34.3 339.7544.3 333.46 44.2 348.0554.6 338.26 54.5 355.9464.4 342.05 64.4 364.2474.6 345.65 74.6 370.5491.2 350.25 91.7 377.93

w1 = 0.4575 w1 = 0.034713.9 309.67 14.1 321.4724.0 320.67 24.0 338.7634.2 328.26 34.1 358.5444.3 333.96 44.1 373.9354.3 338.66 54.4 392.0264.4 342.75 64.2 408.0174.5 346.25 74.4 412.9191.2 351.35 91.5 425.50

D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276 273

Results for the biodiesel + methanol + glycerol system at 298 Kand 323 K are shown in Table 3 and Fig. 4. A comparison withFig. 3 shows that the liquid–liquid equilibrium region is smallerthan in the case of the methanol system. With regard to the biodie-sel purification process, this behavior explains to some extent whythe purification process is more complex when ethanol is used in-stead of methanol during the reaction.

Tables 4 and 5 present the experimental binodal curves for thebiodiesel + methanol + soybean oil and biodiesel + ethanol + soy-bean oil systems, respectively. Data were obtained at 298 K and323 K and the corresponding binodal curves are presented in Figs.5 and 6, respectively.

Analysis of Figs. 5 and 6 shows that the extension of theliquid–liquid immiscibility region is higher for the biodiesel +methanol + soybean oil system than for the biodiesel +methanol + soybean system at 298–323 K. For the latter system,

the binodal curve at 323 K is very short, suggesting that the sepa-ration of the three components is more effective at 298 K.

3.2. Vapor–liquid equilibrium

In order to check the reliability and reproducibility of the exper-imental scheme, preliminary experiments were carried out mea-suring the boiling temperatures of the ethanol + biodiesel binary

Table 7Boiling point temperature for the methanol(1) + biodiesel(2) system (mass fraction).

p (kPa) T (K) p (kPa) T (K)

w1 = 1.0000 w1 = 0.178013.7 293.80 14.3 295.7824.0 304.80 23.9 306.1734.0 312.10 34.2 313.9744.3 317.80 44.3 319.7754.3 322.40 54.3 324.6664.5 326.40 64.5 328.8674.7 329.80 74.6 332.5691.1 334.80 91.8 337.66

w1 = 0.8696 w1 = 0.093913.8 293.88 13.8 295.1824.1 305.28 24.1 306.5734.1 312.87 34.2 314.2744.3 318.77 44.2 320.2754.4 323.46 54.3 325.2664.5 327.36 64.3 329.4674.6 330.76 74.5 333.2692.3 335.86 92.1 338.76

x1 = 0.7359 x1 = 0.040414.2 295.48 13.9 297.5823.9 305.77 24.0 310.1734.0 313.47 34.1 318.7744.3 319.47 44.3 325.8654.3 324.16 54.4 331.9664.4 328.26 64.4 337.4674.6 331.86 74.5 344.2591.9 337.16 92.4 352.14

w1 = 0.5486 w1 = 0.017514.0 294.98 15.9 304.9824.0 305.77 24.0 316.6734.2 313.57 34.0 329.8644.3 319.47 44.2 344.4554.3 324.36 54.3 355.9464.5 328.46 64.4 364.4474.6 331.66 74.5 374.9392.1 336.96 92.0 389.62

w1

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0300

320

340

360

380

400

420

440

T / K

Fig. 8. Boiling point temperatures for the ethanol(1) + biodiesel(2) system atdifferent pressures ( , 14.0 kPa; h, 24.0 kPa; , 34.2 kPa; }, 44.3 kPa; +, 54.5 kPa;j, 64.4 kPa; �, 74.6 kPa; N, 91.4 kPa).

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0w1

280

300

320

340

360

380

400

T / K

Fig. 9. Boiling point temperatures measured for the methanol(1) + biodiesel(2)system at different pressures ( , 14.3 kPa; h, 24.0 kPa; , 34.1 kPa; }, 44.3 kPa; +,54.5 kPa; j, 64.5 kPa; �, 74.6 kPa; N, 92.1 kPa).

Table 8Boiling point temperature for the ternary biodiesel(1) + ethanol(2) + glycerol(3)system.

System 1 System 2 System 3

p (kPa) T (K) p (kPa) T (K) p (kPa) T (K)

13.8 310.97 14.1 312.07 14.1 312.3724.0 322.76 24.4 323.16 24.1 322.9634.0 330.46 34.1 330.46 34.3 330.5644.5 335.96 44.1 336.16 44.2 336.2654.3 340.55 54.5 341.25 54.4 341.1564.5 344.75 64.4 345.25 64.5 345.3574.6 348.35 74.8 349.05 74.8 349.0590.9 353.45 90.6 353.95 90.4 353.85

The following compositions were used for the systems in the table:System 1: biodiesel (49.1 wt%) + ethanol (44.5 wt%) + glycerol (6.4 wt%).System 2: biodiesel (41.3 wt%) + ethanol (46.2 wt%) + glycerol (12.5 wt%).System 3: biodiesel (26.8 wt%) + ethanol (54.2 wt%) + glycerol (19.0 wt%).

274 D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276

systems at 14 kPa and 92 kPa and the results were compared tothose obtained from the literature as shown in Fig. 7.

The T–x diagram presented in Fig. 7 shows that, in general,slight discrepancies are observed between the data obtained here-in and those reported in the literature. These discrepancies may berelated to differences in the raw materials used to produce biodie-sel, since Guo et al. [14] used sunflower seed oil to carry out theirwork.

Tables 6 and 7 present the VLE data for the ethanol + biodieseland methanol + biodiesel systems, respectively, and Figs. 8 and 9show their respective T–x diagrams.

From these figures, one can observe that the pressure has a sig-nificant effect on the mixture boiling temperature. Also as ex-pected, lower pressures correspond to lower mixture boilingtemperatures. For the composition effect two trends can be high-lighted. For an ethanol or methanol mass fraction value higher than10%, the boiling temperature remains practically constant (at boil-ing temperature of pure alcohol). On the other hand, at low ethanolor methanol concentrations, small variations in the mass fractionlead to significant increases in the boiling temperature.

3.3. VLE data for ternary systems

The VLE data were obtained from homogeneous ternary sys-tems with fixed chemical compositions.

Results are shown in Tables 8 and 9 for the biodiesel + etha-nol + glycerol and biodiesel + ethanol + soybean oil systems,respectively. Figs. 10 and 11 present the corresponding pressure–temperature diagrams. These figures show that, within the rangeof pressures investigated in this work, the addition of glycerol(Fig. 10) or soybean oil (Fig. 11) has no significant effect on theboiling temperature of the system independently of the biodieselto ethanol molar ratio. As a result the boiling temperature of themixture is close to that of pure ethanol.

Table 9Boiling point temperature for the ternary biodiesel(1) + ethanol(2) + soybean oil(3)system.

System 1 System 2 System 3

p (kPa) T (K) p (kPa) T (K) p (kPa) T (K)

14.0 309.47 13.9 309.47 14.1 309.6724.3 320.37 24.1 320.27 24.0 320.0734.1 327.56 34.4 327.66 34.2 327.5644.5 333.46 44.2 333.16 44.2 333.1654.4 338.16 54.3 337.96 54.4 337.9664.4 342.05 64.2 341.95 64.3 341.9574.7 345.75 74.4 345.45 74.8 345.6591.6 350.85 91.3 350.75 91.1 350.55

The following mass compositions were used for the systems in the table:System 1: biodiesel (50.1 wt%) + ethanol (44.4 wt%) + soybean oil (5.5 wt%).System 2: biodiesel (47.3 wt%) + ethanol (41.9 wt%) + soybean oil (10.8 wt%).System 3: biodiesel (41.6 wt%) + ethanol (36.9 wt%) + soybean oil (21.5 wt%).

305 310 315 320 325 330 335 340 345 350 355 360

T / K

10

20

30

40

50

60

70

80

90

100

p/ k

Pa

Fig. 10. Boiling point temperatures measured for the biodiesel(1) + etha-nol(2) + glycerol(3) system at different compositions (h, w1 = 0.491, w2 = 0.445; +,w1 = 0.413, w2 = 0.462; d, w1 = 0.268, w2 = 0.542).

305 310 315 320 325 330 335 340 345 350 355

T / K

10

20

30

40

50

60

70

80

90

100

p/ k

Pa

Fig. 11. Boiling point temperatures measured for the biodiesel(1) + etha-nol(2) + soybean oil(3) system at different compositions (h, w1 = 0.501,w2 = 0.444; +, w1 = 0.473, w2 = 0.419; d, w1 = 0.416, w2 = 0.369).

Table 10Boiling point temperature for the ternary biodiesel(1) + methanol(2) + soybean oil(3)system.

System 1 System 2 System 3 System 4

p (kPa) T (K) p (kPa) T (K) p (kPa) T (K) p (kPa) T (K)

14.1 313.37 13.9 295.18 – – – –24.0 342.75 23.8 306.27 24.1 311.27 – –34.0 364.14 34.1 314.07 34.0 320.27 – –44.2 383.73 44.3 319.97 44.3 327.86 45.2 319.97– – 54.4 325.06 54.3 334.16 54.5 324.66– – 64.3 329.36 64.4 339.95 64.4 328.86– – 74.7 333.06 74.5 346.85 74.7 332.56– – 91.2 338.16 91.6 353.25 91.4 337.66

The following mass compositions were used for the systems in the table:System 1: biodiesel (71.9 wt%) + methanol (9.3 wt%) + soybean oil (18.8 wt%).System 2: biodiesel (63.4 wt%) + methanol (20.0 wt%) + soybean oil (16.6 wt%).System 3: biodiesel (45.5 wt%) + methanol (9.2 wt%) + soybean oil (45.3 wt%).System 4: biodiesel (41.8 wt%) + methanol (16.6 wt%) + soybean oil (41.6 wt%).

280 290 300 310 320 330 340 350 360 370 380 390

T / K

10

20

30

40

50

60

70

80

90

100

p/ k

Pa

Fig. 12. Boiling point temperatures measured for the methanol(1) + biodie-sel(2) + soybean oil(3) system at different compositions (h, w1 = 0.093,w2 = 0.719; +, w1 = 0.200, w2 = 0.634; d, w1 = 0.092, w2 = 0.455; 4, w1 = 0.166,w2 = 0.418).

D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276 275

The results for the ternary biodiesel + methanol + soybean oilsystem are shown in Table 10 and Fig. 12. In contrast to the sys-tems with ethanol the boiling temperature is strongly influencedby the addition soybean oil, as can be observed in Fig. 11. Forexample, at a fixed pressure of 40 kPa, the boiling temperature var-ies from 320 K (at soybean oil mass fraction of 0.167) to 380 K (atsoybean oil mass fraction of 0.195).

4. Conclusion

Liquid–liquid and vapor–liquid equilibrium data for systems re-lated to alkyl esters (biodiesel) production and purification werepresented. The results show that, in both cases using ethanol ormethanol, the purification process is more effective at low temper-ature when two liquid phases are present. The immiscibility re-gions for systems based on methanol are larger than thoseobserved for systems based on ethanol. The vapor–liquid equilib-rium data show that the addition of soybean oil or glycerol tothe biodiesel–ethanol binary system has no effect on the mixtureboiling point temperature, in contrast to the biodiesel–methanolsystem where a significant variation in the boiling temperaturewas observed. This observation suggests that a high biodiesel qual-ity is easier to attain for the reaction using methanol.

Acknowledgments

The authors thank CNPq, Fundação Araucária, FINEP (Grant01.07.0488.00 – 3433/06), PRH24/ANP and MEC/Reuni (BrazilianAgencies) for the financial support and scholarships.

References

[1] Demirbas A. Progress and recent trends in biofuels. Prog Energy Combust2007;33:1–18.

[2] Stephenson AL, Dennis JS, Scott SA. Improving the sustainability of theproduction of biodiesel from oilseed rape in the UK. Proc Saf Environ Prot2008;86:427–40.

276 D.I. Segalen da Silva et al. / Fuel 108 (2013) 269–276

[3] Peterson CL, Hustrulid T. Carbon cycle for rapeseed oil biodiesel fuels. BiomassBioenergy 1998;14:91–101.

[4] Zhang Y, Dubé MA, McLean DD, Kates M. Biodiesel production from wastecooking oil: 1. Process design and technological assessment. Bioresour Technol2003;89:1–16.

[5] Cordeiro CS, Silva FR, Wypych F, Ramos LP. Heterogeneous catalysts forbiodiesel production. Quim Nova 2011;34:477–86.

[6] ANP – Agência Nacional do Petróleo, Gás Natural e Biocombustíveis (BrazilianAgency). Disponível em <http://www.anp.gov.br/biocombustiveis/biodiesel.asp> [In Portuguese].

[7] Atadashi IM, Aroua MK, Aziz ARA, Sulaiman NMN. Refining technologies for thepurification of crude biodiesel. Appl Energy 2011;88:4239–51.

[8] Meher LC, Sagar DV, Naik SN. Technical aspects of biodiesel production bytransesterification – a review. Renew Sustain Energy 2006;10:248–68.

[9] Marchetti JM, Miguel VU, Errazu AF. Possible methods for biodiesel production.Renew Sustain Energy Rev 2007;11:1300–11.

[10] Oliveira MB, Teles ARR, Queimada AJ, Coutinho JAP. Phase equilibria of glycerolcontaining systems and their description with the Cubic-Plus-Association(CPA) Equation of State. Fluid Phase Equilib 2009;280:22–9.

[11] Csernica SN, Hsu JT. Reverse-phase ternary phase diagram, tie lines, and plaitpoint for commercial biodiesel–glycerol–methanol. Ind Eng Chem Res2011;50:1012–6.

[12] França BB, Pinto FM, Pessoa FLP, Uller AMC. Liquid–liquid equilibria for castoroil biodiesel + glycerol + alcohol. J Chem Eng Data 2009;54:2359–64.

[13] Kuramochi H, Maeda K, Kato S, Osako M, Nakamura K, Sakai S. Application ofUNIFAC models for prediction of vapor–liquid and liquid–liquid equilibria

relevant to separation and purification processes of crude biodiesel fuel. Fuel2009;88:1472–7.

[14] Guo Y, Zhong J, Xing Y, Li D, Lin R. Volatility of blended fuel of biodiesel andethanol. Energy Fuels 2007;21:1188–92.

[15] Pinto LF, Silva DIS, Silva FR, Ramos LP, Ndiaye PM, Corazza ML. Phaseequilibrium data and thermodynamic modeling of the system(CO2 + biodiesel + methanol) at high pressures. J Chem Thermodyn2012;44:57–65.

[16] Araújo OAS, Silva FR, Ramos LP, Lenzi MK, Ndiaye PM, Corazza ML. Phasebehavior measurements for the system (carbon dioxide + biodiesel + ethanol)at high pressures. J Chem Thermodyn 2012;47:412–9.

[17] Silva LHM, Coimbra JSR, Meirelles AJA. Equilibrium phase behavior ofpoly(ethylene glycol) + potassium phosphate + water two phase systems atvarious pH and temperatures. J Chem Eng Data 1997;42:398–401.

[18] Stragevitch L, D’Ávila SG. Application of a generalized maximum likelihoodmethod in the reduction of multicomponent liquid–liquid equilibrium data.Braz J Chem Eng 1997;14:41–52.

[19] Ardila YC, Pinto GMF, Machado AB, Maciel MRW. Experimental determinationof binodal curves and study of the temperature in systems involved in theproduction of biodiesel with ethanol. J Chem Eng Data 2010;55:4592–6.

[20] Coelho R, Santos PG, Mafra MR, Cardozo-Filho L, Corazza ML. (Vapor + liquid)equilibrium for the binary systems water + glycerol and ethanol + glycerol,ethyl stearate, and ethyl palmitate at low pressures. J Chem Thermodyn2011;43:1870–6.