Fluid Phase Equilibria 259 (2007) 6670

Isobaric vaporliquid and vaporliquidliquid equilibrium data

for the waterethanolhexane system
V. Gomis , A. Font, R. Pedraza, M.D. Saquete
Departamento de Ingeniera Qumica, Universidad de Alicante, Ap. 99, E-03080 Alicante, Spain
Received 29 November 2006; received in revised form 2 April 2007; accepted 4 April 2007
Available online 19 April 2007

Abstract
Isobaric vaporliquid and vaporliquidliquid equilibria were measured for the waterethanolhexane system at normal atmospheric pressure.
The apparatus used for the determination of vaporliquidliquid equilibrium data was an all-glass dynamic recirculating still with an ultrasonic
homogenizer coupled to the boiling flask.
The experimental data demonstrated the existence of a ternary heterogeneous azeotrope at 329.2 K with a composition of 0.105, 0.236 and
0.658 mol fraction of water, ethanol and hexane, respectively. The experimental vaporliquid and vaporliquidliquid equilibria data were found to
be thermodynamically consistent. The correlation of these data was carried out using NRTL and UNIQUAC models and predicted with UNIFAC.
The equilibrium data of the system were compared with other hydrocarbon systems previously analyzed and commonly used as entrainers in
the dehydration of the ethanol by azeotropic distillation such as benzene and cyclohexane.
2007 Elsevier B.V. All rights reserved.
Keywords: Vaporliquid equilibria; Vaporliquidliquid equilibria; Ethanol; Hexane

1. Introduction
Heterogeneous azeotropic distillation is a widely used
technique to separate binary azeotropic mixtures into their components. For many years, the traditional process to obtain pure
ethanol from a waterethanol mixture has been heterogeneous
azeotropic distillation using benzene as entrainer [1]. In the last
few years, because of the carcinogenic effect of benzene, this
entrainer has been substituted by other compounds [2,3].
The absolute ethanol produced with this process can be used
as a carburant mixed with gasoline. Since hexane is a common compound found in gasoline and is suitable to use it as an
entrainer in the azeotropic distillation of ethanol, any remaining
amount of hexane in the anhydrous ethanol will not be a problem
for its subsequent use as fuel. Many studies have been carried
out in this field, either by experimentation or by simulation.
However, it is not possible to find a complete experimental
study in literature of the isobaric vaporliquidliquid equilibrium (VLLE) and vaporliquid equilibrium (VLE) equilibrium

Corresponding author. Tel.: +34 965 90 38 67; fax: +34 965 90 38 26.
E-mail address: vgomis@ua.es (V. Gomis).

0378-3812/$ see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.fluid.2007.04.011

of the waterethanolhexane system at atmospheric pressure.

Usually, the equilibrium data for this system, which are necessary for the design and simulation of the heterogeneous
azeotropic distillation process are substituted by predictions of
these data using activity coefficient models calculated using
group contribution models like UNIFAC or by the UNIQUAC or
NRTL model with parameters based on the correlation of binary
VLE and LLE data.
The objective of this research is to present VLE and VLLE
data for the waterethanolhexane system at a constant pressure
of 101.3 kPa, to study the accuracy of the activity coefficients
models for this system and to compare the results obtained with
those of other systems commonly used in azeotropic distillation.
2. Experimental
2.1. Chemicals
All chemicals were used as supplied after gas chromatography failed to show measurable amounts of organic impurities.
The nominal purities of ethanol and hexane obtained from Merck
were 99.8 and 99.5 mass%. For 2-propanol, which was used as
the internal standard for chromatographic analysis, the nomi-

V. Gomis et al. / Fluid Phase Equilibria 259 (2007) 6670

nal purity was 99.5 mass%. The Karl-Fischer titration method

was used to determine the water content in mass% for hexane,
ethanol and 2-propanol and was found to be 0.005, 0.06, and
0.10, respectively. Ultrapure water was obtained using a Milli
Q-Plus system.

67

computer by using the Shimadzu CLASS-VP Chromatography

Data System. Separation of the components was obtained on
a 2 m 3 mm column packed with Porapack Q 80/100. The
oven temperature was 210 C, and the helium flow rate was
40 mL min1 .
Sampling was carried out using different methods depending
on the phase treated:

2.2. Apparatus and procedures

The equipment used for the determination of VLLE data was a
commercial unit (Labodest model 602) built in Germany by Fischer Labor und Verfahrenstechnik, that was modified by Gomis
et al. [4] by coupling an ultrasonic homogenizer to the boiling flask. The still allows good mixing of the vapor and liquid
phases and good separation of the phases once they reach equilibrium. The ultrasound system ensures a good dispersion of
partly miscible liquid phases, and consequently this equipment
works perfectly for the determination of VLLE data. For VLE
determinations, the same equipment without modifications was
used.
A Pt-100 sensor was employed to measure the equilibrium
temperatures. The probe was connected to a Cropico thermometer (model 3002) with an uncertainty of 0.006 C according to
certificate of calibration (scale ITS 90 [5]). The estimated accuracy of the measures was 0.05 C. A Fischer M101 phase
equilibrium control system was used to measure and control
the pressure and the heating power. The pressure in the still was
101.3 kPa, measured and controlled with an accuracy of 0.1 kPa.
To guarantee the correct operation of the equipment, the boiling point of pure water was measured and compared with the
accepted value.
Most of the analytical work was carried out by gas chromatography on a Shimadzu GC-14B coupled with a personal

(a) The sampling of the gaseous phase was carried out using a
UW Type, 6-port valve from Valco Instruments Co., which
injected the samples automatically into the GC.
(b) For the sampling of the liquid phase in the heterogeneous
region, a small amount of the liquid coming from the separation chamber of the instrument was diverted to a tube
using a solenoid valve. In this tube, the dispersed liquid
phases enter and separate into two layers at their bubble
point, since the tube is placed in a thermostatic bath at the
boiling point temperature of the mixture. A sample of each
layer was taken and placed in a vial with a small amount of
an internal standard.
(c) Liquid samples in the homogeneous region were withdrawn
with a syringe of the liquid coming from the separator chamber and put into a vial.
To obtain quantitative results in the analysis of the liquid
phases, an internal standard method was used. For this reason,
2-propanol (which is completely miscible with water, ethanol
and hexane) was also added to the sample vials. Furthermore, the
addition of the standard prevents phase-separation effects when
changing the temperature after the separation of the phases.
A further description of the sampling procedure is given in a
previous paper [4].

Table 1
Vaporliquidliquid equilibrium data (mole fraction) for the water (1)ethanol (2)hexane (3) ternary system at 101.3 kPa
Organic phase

BIN
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

Aqueous phase

Vapor phase

Tb (K)

x1

x2

x3

x1

x2

x3

y1

y2

y3

0.002
0.001
0.002
0.002
0.003
0.003
0.004
0.005
0.006
0.007
0.008
0.010
0.012
0.015
0.016
0.022
0.029
0.037
0.047
0.055
0.061

0.000
0.004
0.015
0.021
0.028
0.039
0.049
0.060
0.065
0.080
0.087
0.101
0.113
0.133
0.142
0.172
0.203
0.235
0.272
0.297
0.316

0.998
0.995
0.983
0.976
0.969
0.958
0.947
0.935
0.930
0.913
0.905
0.889
0.875
0.852
0.842
0.806
0.768
0.728
0.680
0.648
0.622

1.000
0.965
0.893
0.869
0.791
0.754
0.732
0.704
0.674
0.616
0.555
0.524
0.489
0.431
0.379
0.353
0.324
0.296
0.248
0.234
0.219

0.000
0.035
0.106
0.131
0.208
0.244
0.266
0.293
0.322
0.376
0.433
0.461
0.492
0.541
0.575
0.593
0.613
0.624
0.620
0.609
0.604

0.000
0.000
0.001
0.001
0.001
0.002
0.002
0.003
0.004
0.008
0.012
0.014
0.019
0.029
0.046
0.053
0.063
0.079
0.132
0.157
0.178

0.214
0.159
0.139
0.137
0.128
0.124
0.123
0.122
0.120
0.117
0.114
0.112
0.110
0.107
0.104
0.103
0.101
0.100
0.096
0.096
0.095

0.000
0.076
0.150
0.162
0.185
0.195
0.197
0.202
0.207
0.214
0.221
0.225
0.229
0.235
0.239
0.242
0.246
0.248
0.253
0.254
0.256

0.786
0.765
0.711
0.702
0.688
0.681
0.680
0.676
0.673
0.670
0.665
0.663
0.661
0.658
0.657
0.656
0.653
0.653
0.652
0.650
0.649

334.44
332.57
330.80
330.54
330.05
329.92
329.82
329.77
329.71
329.56
329.51
329.48
329.45
329.42
329.26
329.28
329.48
329.50
329.53
329.55
329.56

68

V. Gomis et al. / Fluid Phase Equilibria 259 (2007) 6670

Depending on the composition of the samples, detection was

carried out by different techniques: in the organic phase, water
was analyzed by the Karl-Fischer titration method (Metrohm
737 KF Coulometer) and the ethanol and hexane by gas chromatography using a TCD (thermal conductivity detector); in the
aqueous phase, water and ethanol have been analyzed using TCD
and hexane with FID (flame ionization detector). The temperature of the detector was 220 C, and the current for the TCD was
100 mA.
The accuracy of the mole fraction measurements was estimated at 0.002 for all the compounds except for the water in
the liquid organic phase and hexane in the aqueous phase, where
the accuracy was approximately 0.005.
3. Results
The experimental VLLE data of the ternary system studied
are summarized in Table 1. VLE data for the homogeneous
region are shown in Table 2. In these tables, the bubble point
(Tb , K) and the composition (mole fraction) of the liquid phases
(xi ) and the vapor phase (yi ) are presented.
The experimental ternary VLLE and VLE data were tested
by the point-to-point L/W Wisniak [6] consistency test and were
found to be thermodynamically consistent. All the values of L/W
are between 0.97 and 1.00. Vapor pressures were calculated with
the Antoine equation, whose parameters Ai , Bi , and Ci for water,
ethanol and hexane were taken from literature [7] and are given
in Table 3. This test does not reveal any substantial inconsistency
in the data.
Fig. 1 represents the VLE data of liquid mixtures in the homogeneous region of the multicomponent system that has been
studied. Several liquid mixtures of the homogeneous region in
Fig. 1 have a vapor phase, that once condensed, would belong to
the heterogeneous region of the diagram representing the multicomponent system. Samples of these vapor phases were directly
injected into the chromatograph to avoid the problem of having
to sample from a mixture of two liquid phases.

Table 2
Vaporliquid equilibrium data (mole fraction) for the water (1)ethanol
(2)hexane (3) ternary system at 101.3 kPa
Liquid phase

Vapor phase

Tb (K)

x1

x2

x3

y1

y2

y3

0.317
0.139
0.234
0.135
0.408
0.253
0.138
0.390
0.246
0.263
0.460
0.428
0.344
0.156
0.154
0.123
0.263
0.267
0.291
0.035
0.030
0.029
0.035
0.073
0.272
0.103
0.190
0.019
0.020
0.016
0.013
0.013
0.013
0.014
0.012
0.009
0.013
0.013
0.018
0.094
0.110
0.094
0.072
0.055
0.054
0.049
0.048
0.145
0.122

0.679
0.844
0.756
0.831
0.579
0.724
0.803
0.594
0.693
0.670
0.528
0.557
0.609
0.729
0.559
0.624
0.718
0.713
0.708
0.964
0.936
0.880
0.848
0.676
0.628
0.574
0.653
0.697
0.799
0.828
0.665
0.592
0.524
0.456
0.371
0.309
0.247
0.750
0.815
0.658
0.678
0.635
0.700
0.783
0.774
0.676
0.734
0.666
0.727

0.004
0.016
0.010
0.034
0.014
0.023
0.059
0.016
0.061
0.066
0.013
0.015
0.046
0.115
0.287
0.253
0.019
0.020
0.001
0.001
0.033
0.090
0.117
0.251
0.100
0.323
0.157
0.284
0.181
0.156
0.323
0.395
0.463
0.530
0.616
0.682
0.740
0.237
0.167
0.248
0.212
0.272
0.228
0.162
0.172
0.275
0.218
0.189
0.151

0.215
0.105
0.159
0.089
0.209
0.135
0.078
0.189
0.102
0.100
0.216
0.175
0.110
0.077
0.103
0.098
0.080
0.154
0.220
0.036
0.025
0.020
0.025
0.054
0.102
0.079
0.090
0.017
0.012
0.012
0.012
0.013
0.014
0.015
0.015
0.017
0.017
0.008
0.012
0.062
0.070
0.068
0.051
0.037
0.037
0.039
0.033
0.080
0.068

0.683
0.669
0.636
0.563
0.529
0.485
0.438
0.494
0.340
0.297
0.504
0.435
0.282
0.313
0.253
0.257
0.279
0.536
0.752
0.957
0.686
0.484
0.437
0.313
0.250
0.277
0.270
0.355
0.403
0.408
0.351
0.338
0.333
0.327
0.324
0.319
0.313
0.375
0.419
0.301
0.300
0.298
0.324
0.377
0.371
0.332
0.351
0.289
0.316

0.103
0.227
0.205
0.347
0.261
0.380
0.484
0.317
0.558
0.604
0.280
0.389
0.608
0.610
0.644
0.645
0.641
0.310
0.029
0.008
0.289
0.495
0.539
0.634
0.648
0.644
0.640
0.628
0.586
0.580
0.637
0.649
0.653
0.657
0.661
0.665
0.670
0.617
0.570
0.637
0.631
0.633
0.625
0.586
0.593
0.629
0.616
0.632
0.616

349.66
346.18
346.66
342.64
345.91
341.71
338.04
344.36
335.28
332.55
345.88
342.20
332.71
332.66
329.51
329.71
330.34
343.86
351.10
351.20
329.73
329.73
329.73
329.71
330.23
330.25
330.63
332.36
334.33
334.64
332.08
331.72
331.50
331.39
331.29
331.25
331.26
333.16
334.59
331.09
331.21
330.75
331.84
333.91
333.95
331.79
331.90
332.03
332.22

Table 3
Antoine equation parametersa of the pure substances

Fig. 1. VLE (mol%) diagram for the water (1)ethanol (2)hexane (3) ternary
system at 101.3 kPa: () liquid phase and (+) vapor phase.

Compound

Temperature
range ( C)

Water
Ethanol
Hexane

8.07131
8.11220
6.91058

1730.630
1592.864
1189.64

233.426
226.184
226.280

+1/+100
+20/+93
30/+170

Antoine equation: log(P) = A B/[t + C], with: P/mmHg and t/ C.

V. Gomis et al. / Fluid Phase Equilibria 259 (2007) 6670

Fig. 2. VLLE (mol%) diagram for the water (1)ethanol (2)hexane (3) ternary
system at 101.3 kPa: () liquid phase and (+) vapor phase.
Table 4
Ternary azeotrope of the water (1)ethanol (2)hexane (3) system at 101.3 kPa
obtained by different authors
x1

x2

x3

Tb (K)

0.144
0.210
0.105

0.215
0.168
0.236

0.641
0.622
0.658

329.19
329.15
329.21

Kudryavtseva
Fritzweiler
Experimental

Fig. 2 represents some of the VLLE data and includes the nonisothermal binodal curve and the vapor line. This figure shows
the presence of a ternary azeotrope in the heterogeneous region.
The points of the vapor line corresponding to the liquidliquid
equilibrium lines 113 are under the liquidliquid tie line. However, the points corresponding to lines 1420 are above the tie
line. Therefore, there is a point between lines 13 and 14 where
the composition of the vapor coincides with that of a liquid heterogeneous mixture. The composition of the ternary azeotrope
determined by numerical interpolation is x1 = 0.105, x2 = 0.236,
and x3 = 0.658 mole fraction of water (1), ethanol (2), and hexane
(3), respectively, and the temperature is 329.21 K. The composition of the two liquid phases in equilibrium is x1 = 0.401,
x2 = 0.561, and x3 = 0.038 in the aqueous phase, and x1 = 0.016,
x2 = 0.138, and x3 = 0.847 in the organic-rich phase. It is worth
noting that establishing a composition of the ternary azeotrope
by numerical interpolation is difficult. For this reason, the accuracy of these compositions is less than that of the experimental
data. The experimental data obtained previously by Kudryavtseva et al. in 1973 [8] and Fritzweiler and Dietrich in 1933 [9]
(compiled by Gmehling et al. [10]) that can be seen in Table 4
differs significantly from the azeotrope composition obtained in
this work, especially that obtained by Fritzweiler. However, the
bubble temperature is the same.

69

was done. The models used were NRTL and UNIQUAC and
the correlated data were the bibliographic VLE data of the
ethanolhexane [11] and waterethanol [11] systems and the
homogeneous and heterogeneous VLE data of this work. All
the correlations and calculations have been made using the
process simulator CHEMCAD V [12]. However, the parameters obtained do not properly predict the heterogeneous region.
Consequently, an LLE data correlation has been done to try to
improve the calculations of that region. The parameters obtained
and the mean deviations between experimental and calculated
compositions are presented in Table 5.
To properly simulate an azeotropic distillation column to
dehydrate ethanol with hexane, it is important to know the size
and shape of the heterogeneous region of the system and the
slope of the liquidliquid tie lines. These characteristics have a
great influence on the compositions obtained of the two liquids
produced in the decanter after the condensation of the vapor that
leaves the top of the distillation column.
For this reason the non-isothermal binodal curve and some
tie lines calculated with UNIFAC LLE and with the parameters
obtained with NRTL and UNIQUAC are compared in Fig. 3.
The size of the heterogeneous region calculated with the different models is larger than the experimental one, especially in
the right-hand region of the diagram. The liquidliquid region
calculated with the models even includes points that are homogeneous. Furthermore, the tie lines of this region, which are those
with the greatest concentrations of ethanol, are those necessary
for the calculations of the decanter and therefore very important
to the simulation of the column.
In conclusion, in spite of the fact that the parameters obtained
appear to predict small deviations from the experimental data,
these small differences have a great influence on the simulation results. For this reason, an improvement in the calculation
method of the equilibrium compositions is necessary.
On the other hand, in order to compare the equilibrium experimental results obtained with those of other hydrocarbon systems

4. Discussion and conclusion

In order to use the experimental data to do simulations of
an ethanol dehydration process, a correlation of the VLE data

Fig. 3. Comparison of the non-isothermal binodal curve and some tie lines
calculated with UNIFAC LLE and with the parameters obtained with NRTL and
UNIQUAC.

70

V. Gomis et al. / Fluid Phase Equilibria 259 (2007) 6670

Table 5
Parameters and mean deviations of the LLE correlation
i

Bij

Bji

Uij Ujj

Uji Uii

Water
Water
Ethanol

Ethanol
Hexane
Hexane

765.2826
2130.053
421.547

889.5103
1661.666
275.1408

0.3031
0.2
0.3827

821.2323
1057.531
199.0427

347.0423
1823.642
564.5045

Mean deviation
NRTL
UNIQUAC

D x11

D x21

D x12

D x22

0.003
0.007

0.012
0.013

0.010
0.010

0.010
0.012

Bij binary interaction parameters for the NRTL model (cal/mol). Uij Uii binary interaction parameters for UNIQUAC (cal/mol). D xij (i = component, j = phase)
mean deviations in mole fraction of water (1) and ethanol (2) in organic phase (1) and aqueous phase (2).

As can be seen in Fig. 4, the water content in the aqueous

phase in the system with benzene is greater than in those with
cyclohexane and hexane. For these two systems, it is necessary
to recirculate a greater amount of all the components to separate
the same amount of water with the consequent increase in the
production cost.
Acknowledgement
The authors thank the DGICYT of Spain for the financial
support of project PPQ2003-03385.
References

Fig. 4. Comparison of the non-isothermal binodal curve, the vapor line

and the tie line containing the azeotropic point for each one of the
waterethanolhydrocarbon systems.

previously analyzed and commonly used as entrainers, Fig. 4

has been drawn. This figure shows the non-isothermal binodal
curve, the vapor line and the tie line containing the azeotropic
point for each one of the waterethanolhydrocarbon ternary
systems, being the hydrocarbon hexane, benzene and cyclohexane. The waterethanolbenzene data have been obtained from
literature [13,14] and the experimental data for cyclohexane
were published in a previous paper [15].
From the point of view of the use of these systems in the
dehydration of ethanol using an azeotropic distillation column,
the efficiency of the process is improved when the amount of
water eliminated in the condenserdecanter of the column is
greater.
As the vapor leaving the top of the column has a composition
closed to the azeotropic one, the composition of the aqueous
phase of the tie line containing this point is related with the
amount of water that can be eliminated after condensation and
decantation.

[1] S. Young, J. Chem. Soc. 81 (1902) 707.

[2] S.A. Leeper, P.C. Wankat, Ind. Eng. Chem. Process Des. Dev. 21 (1982)
331334.
[3] I.V. Prikhodko, T.M. Letcher, T.W. de Loos, Ind. Eng. Chem. Res. 36 (1997)
43914396.
[4] V. Gomis, F. Ruiz, J.C. Asensi, Fluid Phase Equilib. 172 (2000) 245
259.
[5] B.W. Mangum, G.T. Furukawa, U.S. Department of Commerce, National
Institute of Standards and Technology, Springfield, 1990.
[6] J. Wisniak, Ind. Eng. Chem. Res. 32 (1993) 15311533.
[7] R.H. Perry, D.W. Green, Perrys Chemical Engineers Handbook, 7th ed.,
McGraw Hill, New York, 1997.
[8] L.S. Kudryavtseva, M.Yu. Toome, M.P. Susarev, Deposited publ. VINITI
DEP.NO. 6975-73, 1973.
[9] R. Fritzweiler, K.R. Dietrich, Beihefte Angewandte Chemie und Die
Chemische Fabrik, Heft 4, Berlin, 1933.
[10] J. Gmehling, J. Menken, J. Krafczyk, K. Fischer, Azeotropic Data, VCH,
Weinheim, 1994.
[11] J. Gmehling, U. Onken, Chemistry Data Series, DECHEMA, Dortmund,
1977.
[12] V. Chemcad, Process Flow Sheet Simulator, Chemstations Inc., Houston,
2002.
[13] W.S. Norman, Trans. Inst. Chem. Eng. 23 (1945) 6675.
[14] J. Barbaudy, Rec. Trav. Chim. Pays-Bas 45 (1926) 207213.
[15] V. Gomis, A. Font, R. Pedraza, M.D. Saquete, Fluid Phase Equilib. 235
(2005) 710.