Fuel 107 (2013) 4751

Contents lists available at SciVerse ScienceDirect

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

Vapor pressure measurements of ethanolisooctane and

1-butanolisooctane systems using a new ebulliometer
Rama Oktavian, Vika Amidelsi, Agung Rasmito, Gede Wibawa
Department of Chemical Engineering, Faculty of Industrial Technology, Sepuluh Nopember Institute of Technology (ITS), Campus ITS Sukolilo, Surabaya 60111, Indonesia

h i g h l i g h t s

" A new simple ebulliometer has been developed in this work.

" The ebulliometer is reliable to measure vapor pressure of alcoholsisooctane mixture.
" The experimental data were well correlated with Wilson, NRTL, and UNIQUAC models.

a r t i c l e i n f o a b s t r a c t

Article history: A new, simpler ebulliometer has been developed to accurately measure the vapor pressure of ethanol
Received 8 November 2011 isooctane and 1-butanolisooctane systems at various mixture compositions. The reliability of the exper-
Received in revised form 31 January 2013 imental apparatus was validated by comparing the experimental data with published data for pure iso-
Accepted 2 February 2013
octane and ethanol and ethanolisooctane system. Our data were in good agreement with the published
Available online 22 February 2013
data, with a deviation of less than 1.9%. The vapor pressure of an ethanolisooctane mixture was initially
elevated by the addition of ethanol until the ethanol concentration reached 0.3 mass fractions, then
Keywords:
dropped at concentrations above 0.7 mass fractions. For a 1-butanolisooctane system, the vapor pres-
Ebulliometer
Ethanol
sure decreased with increasing 1-butanol concentrations. The experimental data were correlated using
Isooctane the Wilson, Non-Random Two-Liquid (NRTL), and Universal Quasi-Chemical (UNIQUAC) activity coef-
Vapor pressure cient models with an average absolute deviation in the vapor pressures of 3.5%, 3.3%, and 6.1%,
respectively.
2013 Elsevier Ltd. All rights reserved.

1. Introduction physical properties of a gasoline mixture because it refers to the

Fossil fuels, which remain the primary energy resources around
the world, cause signicant increases in the atmospheric CO2 con-
centration and have an effect on global warming [1,2]. Fossil fuels
combustion contributes 73% of CO2 production [3]. Thus, nding a
solution to greenhouse gas emissions is a major research focus for
solutions to the global warming problem.
Recently, the addition of oxygenates, such as simple alcohols,
that are primarily produced from renewable resources has received
tremendous attention because these compounds have an ability to
reduce pollution and enhance the octane number of gasoline. Nich-
olas Otto [4] rst used ethanol in internal combustion engines. Eth-
anolgasoline mixtures have been successfully used in various
machines and vehicles that typically run on gasoline [5]. However,
the mixture can have a higher vapor pressure than the base gaso-
line which raises concern regarding the evaporative emission from
the fuel mixture. Vapor pressure is one of the most important
Corresponding author. Tel.: +62 31 5946240; fax: +62 31 5999282.
E-mail address: gwibawa@chem-eng.its.ac.id (G. Wibawa).
volatility of the gasoline and can affect the evaporative emission
of the fuel mixture. Therefore, accurate experimental data are
important in designing fuel mixtures with specic vapor pressures.
The vapor pressures of gasoline mixtures with four alcohols at
37.8 C (100 F) of various compositions were measured by Pum-
phrey et al. [6]. Sun et al. [7] measured the bubble-point vapor
pressures for mixtures of an endothermic hydrocarbon fuel with
ethanol. The experimental results indicated that the addition of
ethanol had a critical effect on the vapor pressure of the fuel. An
ethanol-fuel blend with a low ethanol composition can have an in-
creased volatility.
The vapor pressures for both pure substances and mixtures are
traditionally measured using an ebulliometer, which was rst pro-
posed by Cotrell [8] to determine the boiling points of pure sub-
stances. Rogalski and Malanowski [9] developed a new
ebulliometer in 1980 using Cotrells principle for determining va-
porliquid equilibrium data. However, in these ebulliometers anal-
yses of both vapor and liquid-phase compositions are necessary
due to a difference between the liquid-phase composition at
equilibrium and the feed composition. Therefore, Li et al. [10]

0016-2361/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2013.02.005
48 R. Oktavian et al. / Fuel 107 (2013) 4751
developed an inclined ebulliometer that no longer required liquid
and vapor-phase composition analysis to measure the bubble point
for a chloroformethanolbenzene system. Li et al. [10] also noted
that they adopted the angle of inclination of 30 for their work to
ensure the accuracy of the equilibrium temperature measurement.
In this work, a new, simpler ebulliometer is proposed based on
the work of Li et al. [10] by removing the inclination and enlarging
the volume of the cell. We can ensure that the equilibrium and the
feed compositions do not change signicantly during the measure-
ments due to the small proportion of vapor formed during equilib-
rium by enlarging the volume of the cell. Therefore, the present
apparatus can be used to accurately measure the vapor pressure
of a gasoline mixture. We used isooctane in this work to represent
gasoline because it is the main component in gasoline.
2. Experimental
2.1. Materials
All materials used in this work (isooctane, ethanol, and 1-buta-
nol) were supplied by Merck, Darmstadt, Germany. These materi-
als were used without further purication. Table 1 lists the
properties of all materials used.
2.2. Experimental apparatus
A new, simpler ebulliometer based on the principle of the quasi-
static method is shown in Fig. 1b. The ebulliometer has a 5.5-cm
cell diameter, 14-cm height, and a 332.5-cm3 volume and is con-
nected with a vacuum pump. The ebulliometer cell was turned
on via moderate heating, which was controlled using a tempera-
ture controller to maintain equilibrium at the desired temperature.
The quasi-static method is based on the overall concentration
instead of the equilibrium composition of the liquid phase as de-
scribed by Li et al. [10] and shown in the inclined region in
Fig. 1a. The enlargement of the cell volume as shown in Fig. 1b al-
lows for an increase in the amount of substance being tested so
that the composition of the uid hardly changes during the vapor
pressure measurement. This can be calculated using the simple
mass balance in Eq. (1) as follows:
zi F xi L yi V 1 from Hull et al. [14].
In which, zi is the mole fraction of component i in the initial
mixture injected into the ebulliometer, F the total moles of mixture
injected into the ebulliometer, xi the mole fraction of component i
in the liquid phase at equilibrium, L the number of moles in the li-
quid phase at equilibrium, yi the mole fraction of component i in
the vapor phase at equilibrium, V the number of moles in the vapor
phase at equilibrium.
We assume, with the same value of V, the same mole fraction
for component i in the vapor phase (yi), and the same feed compo-
sition (zi), that a smaller ebulliometer means that the amount of
mixture injected into the ebulliometer (F) must be smaller, and sig-
nicant change can occur in the mole fraction of component i in li-
quid phase at equilibrium (xi). In our work, we developed a new,
Table 1
Properties of materials used in this work.
Property 1-butanol CH3(CH2)3OH
Molecular weight (g/mol) 74.12
Purity >99.5%
Density (d 20 C/20 C) 0.8090.812
Boiling range (C) 116119
simpler ebulliometer by removing the inclined portion from the
previous ebulliometer. In addition, a temperature detector was at-
tached at the liquidvapor interface to ensure the accuracy of equi-
librium temperature measurements.
The mixtures with varying isooctane and ethanol, isooctane and
1-butanol compositions were prepared by directly weighing the
constituent component on an Ohaus balance with a precision
of 0.0001 g. The temperature was measured using an RTD Pt
100 sensor connected to a Shimaden SD 15 temperature display
with a measurement accuracy of 0.1 K. The vapor pressures at
various temperatures and compositions were measured using a
mercury manometer with an accuracy of 0.1 mm Hg.
2.3. Experimental procedure
A sample containing approximately 220 cm3 of a liquid mixture
of known composition was charged in the ebulliometer cell. First,
the system was created under vacuum conditions to remove air
and impurities from the equilibrium cell. Then, the equilibrium cell
was heated to reach the desired equilibrium temperature. The tem-
perature was controlled by a PID controller (Shimaden SR 64). After
the ebulliometer reached a constant pressure at a desired equilib-
rium temperature, the pressure was recorded as the vapor pressure
at a desired temperature. The experimental procedure was re-
peated at different temperatures and compositions.
3. Results and discussion
3.1. Reliability test of the apparatus
The proposed ebulliometer was tested for the reliability by
comparing the experimental data for the vapor pressures of both
pure liquids and mixtures as measured by this apparatus with
the published data for pure ethanol, isooctane, and an ethanoliso-
octane mixture. The Antoine equation [11] for pure isooctane and
the Wagner equation [12] for pure ethanol were used to compare
the experimental data. The Antoine and Wagner constants were
obtained from Poling et al. [13]. For the ethanolisooctane mixture,
the experimental data were compared with the experimental data
A comparison of the experimental data obtained in this work
with the published data is presented in Tables 2 and 3 and Figs. 2
and 3 along with their deviations. The results indicate that the data
were in good agreement with the published data, with a deviation
of less than 1.9% calculated as an average absolute deviation (AAD)
in the vapor pressures as follows:
n  
1X 
Plit  Pexp  100%
AAD 2
n i1  Pexp 
i
in which Pexp is the vapor pressure obtained in the experiment
while P lit is the vapor pressure calculated using both the Antoine
[11] and Wagner [12] equations and n is the number of data.
Isooctane C8H18 Ethanol C2H5OH
114.23 46.07
>99.5% 99.9%
0.6910.696 0.7900.793
98100 78
 R. Oktavian et al. / Fuel 107 (2013) 4751 49

1. Condenser tube joint

2. Vacuum jact
3. Inner casing
4. Thermometer well 1. Ebulliometer cell
5. Feed inlet 2. Thermocouple
6. Stirrer 3. Condenser
7. Boiler 4. Vacuum unit
8. Condenser 5. Mercury manometer
9. Vacuum pump 6. Heating system controller
10. Mercury manometer 7. Magnetic stirrer
(a) (b)
Fig. 1. Comparison of the ebulliometer apparatus: (a) Ebulliometer developed by Li et al. [10]; (b) new, simpler ebulliometer developed in this work.

Table 2
Pexp Isooctane
Comparison of experimentally measured isooctane vapor pressures (Pexp) with the 33
literature data (Plit). Pexp Ethanol
30 PLit
T (K) Plit (kPa) Pexp (kPa) AAD (%)
300.35 7.31 7.33 0.34 27
302.75 8.17 8.13 0.51
305.95 9.46 9.47 0.08 24
P (kPa)

308.25 10.48 10.80 2.94

310.15 11.39 11.73 2.89 21
313.15 12.97 13.07 0.76
316.25 14.77 14.80 0.18 18
318.95 16.51 16.40 0.68
Overall AAD 1.05 15

12

9
Table 3
6
Comparison of experimentally measured ethanol vapor pressure (Pexp) with the
300 303 306 309 312 315 318 321 324 327
literature data (Plit).
T (K)
T (K) Plit (kPa) Pexp (kPa) AAD (%)
301.35 8.86 8.93 0.86 Fig. 2. Vapor pressure data for isooctane and ethanol.
302.55 9.50 9.60 1.09
308.45 13.24 13.60 2.62
314.35 18.19 18.80 3.23
concentrations above 0.7 mass fractions. For the 1-butanolisooc-
318.95 23.07 23.46 1.66
321.15 25.78 26.26 1.86 tane, the vapor pressure of the mixture decreased with increasing
323.65 29.17 29.73 1.89 1-butanol concentrations.
325.85 32.46 33.06 1.83
Overall AAD 1.88 3.3. Correlation with the activity coefcient models

The experimental data were correlated using the Wilson [15],

3.2. Vapor pressure measurement
The vapor pressures at various temperatures and compositions
for the ethanolisooctane and the 1-butanolisooctane systems
obtained in this work are presented in Tables 4 and 5 and Fig. 4
in which w represents the mass fraction of the ethanol and 1-buta-
nol in the mixtures. For ethanolisooctane mixture, the vapor pres-
sure was initially elevated by the addition of ethanol until the
ethanol concentration reached 0.3 mass fractions. Then, the vapor
pressure stabilized as the ethanol concentration increased from 0.3
to 0.6 mass fractions. Finally, the vapor pressure dropped at
Non-Random Two-Liquid (NRTL) [16], and Universal Quasi-Chem-
ical (UNIQUAC) [17] activity coefcient models. The best t corre-
lation of the experimental data for ethanolisooctane and 1-
butanolisooctane mixtures was presented by Wilson activity
coefcient model as shown in Fig. 4. Comparisons of the experi-
mental data with the three activity coefcient models are pre-
sented in Fig. 5, which depicts one of the mass fractions
observed in our work with the parameters obtained from the cor-
relation of all experimental data in this work.
The average absolute deviations between the experimental
vapor pressures and the calculated vapor pressures with the
50 R. Oktavian et al. / Fuel 107 (2013) 4751

Table 5
20 Bubble point vapor pressure data for 1-butanol (1)isooctane (2) mixture.

18 T (K) P (kPa) T (K) P (kPa) T (K) P (kPa) T (K) P (kPa)

w1 = 0.10 w1 = 0.20 w1 = 0.30 w1 = 0.40
16 302.75 7.73 300.95 7.33 306.55 8.80 305.75 8.53
306.15 9.20 307.05 9.87 308.15 10.40 308.15 9.60
14 308.15 10.67 308.15 11.33 308.95 10.27 309.05 9.65
309.45 10.80 309.35 11.47 316.45 14.40 313.05 11.73
316.65 14.53 316.05 15.20 320.05 16.27 316.75 14.00
12
319.35 15.60 322.35 19.33 321.65 17.47 322.35 17.47
P (kPa)

Ref [14] w1 = 0.50 w1 = 0.60 w1 = 0.70 w1 = 0.80

10 304.05 8.53 301.95 6.27 301.25 5.07 302.15 4.13
Exp Data at T = 308.15 K 315.05 14.00 307.05 8.13 308.15 7.60 308.15 6.27
8 322.75 19.47 308.15 8.93 308.85 7.47 310.95 6.80
325.45 21.33 308.75 8.80 311.15 8.13 315.25 8.40
6 Overall AAD = 1.23% 328.15 23.60 312.35 10.27 319.65 12.13 318.95 9.20
316.15 12.67 326.45 16.00 322.25 10.53
4 w1 = 0.90
305.85 3.20
308.15 4.53
2
308.35 4.00
320.95 7.47
0 0.2 0.4 0.6 0.8 1 323.45 8.13
326.65 9.20
ethanol mol fraction

Fig. 3. Vapor pressure data for the ethanolisooctane mixture.

21
Table 4
Bubble point vapor pressure data for ethanol (1)isooctane (2) mixture. 18
T (K) P (kPa) T (K) P (kPa) T (K) P (kPa) T (K) P (kPa)
w1 = 0.10 w1 = 0.20 w1 = 0.30 w1 = 0.40 15
305.65 16.13 302.95 15.47 306.75 19.20 302.15 14.13
308.15 18.80 307.05 19.20 308.15 20.27 302.95 14.53
12
P (kPa)

311.95 21.60 308.15 20.00 313.45 27.06 303.65 15.20

314.45 24.00 310.15 22.40 316.15 30.80 304.95 15.87
316.35 26.26 311.75 23.73 319.45 35.46 308.15 20.00 9
w1 = 0.50 w1 = 0.60 w1 = 0.70 w1 = 0.80
303.15 15.33 304.65 15.20 308.15 19.00 302.55 12.87
6
305.45 16.67 305.85 15.73 310.75 20.73 306.65 15.67
308.15 20.00 308.15 19.20 312.85 22.46 308.15 18.73 Pexp ethanol-isooctane at T = 308.15 K
311.65 22.53 310.95 20.93 313.55 23.40 310.85 19.67 3 Pexp 1-butanol-isooctane at T = 308.15 K
312.15 23.06 313.05 23.20 315.15 24.73 314.45 22.73 Wilson eq.
w1 = 0.90
0
306.95 15.87
0.0 0.2 0.4 0.6 0.8 1.0
308.15 16.67
309.25 17.33 w
310.35 17.73
313.15 20.40 Fig. 4. Vapor pressures of the ethanolisooctane and 1-butanolisooctane
mixtures.

interaction parameter pairs of the systems tted to the experimen- 24

tal data are presented in Table 6. The binary interaction parameters
22
for the mixtures studied in this work were obtained by minimizing
the average absolute deviation in the vapor pressures dened in 20 Pexp ethanol(1)-isooctane(2) at w1 = 0.5
Eq. (2) as the objective function. Pexp 1-butanol(1)-isooctane(2) at w1 = 0.7
18
The experimental data were well correlated with the Wilson, Wilson eq.
NRTL eq.
NRTL, and UNIQUAC activity coefcient models with average abso- 16
P (kPa)

UNIQUAC eq.
lute deviations in the vapor pressures of 3.5%, 3.3%, and 6.1%,
14
respectively.
12

4. Conclusion 10

8
The new, simpler ebulliometer proposed in this work is a reli-
able tool to accurately measure the vapor pressure of both a pure 6
liquid and a mixture. This experimental apparatus was used to
4
measure the vapor pressure of an ethanolisooctane and a 1-buta- 300 303 306 309 312 315 318 321 324 327
nolisooctane mixture with deviations of less than 1.9%. The
T (K)
experimental data were well correlated with the Wilson, NRTL,
and UNIQUAC activity coefcient models with average absolute Fig. 5. Vapor pressure data for the ethanolisooctane and l-butanolisooctane
deviations in the vapor pressures of 3.5%, 3.3%, and 6.1%, mixtures: measured and correlated with three activity coefcient models.
 R. Oktavian et al. / Fuel 107 (2013) 4751 51

Table 6
Summary of models comparison and models parameter calculated for the studied mixtures.

Mixture Wilson NRTL UNIQUAC

a12 a21 a b12 b21 u12 u21
Ethanol(1)isooctane(2) 1630 231 0.29 8475 1231 364 1470
Overall AAD (%) 3.48 3.29 6.09
1-butanol(1)isooctane(2) 1952 163 0.18 8537 1314 214 662
Overall AAD (%) 4.19 4.74 4.82

respectively. For further research, the new ebulliometer developed
in our work will be employed to measure and generate vapor pres-
sures or vaporliquid equilibrium data for various mixtures.
Acknowledgments
This work was supported by the Directorate Research and Pub-
lic Services, Directorate General of Higher Education, Department
of National Education, in the Republic of Indonesia (Contract No:
10473/I2.7/PM/2009 April 1, 2009) and Institute of Research and
Public Services ITS.
References
[1] Yu J, Corripo AB, Harrison OP, Copeland RJ. Analysis of the sorbent energy
transfer system (SETS) for power generation and CO2 capture. Adv Environ
2003;7:33545.
[2] Demirbas F, Bosbas K, Balat M. Carbon dioxide emiision trends and
environmental problems in Turkey. Energy Explor Exploit 2004;22:33565.
[3] Wildenborg T, Lokhorst A. Introduction on CO2 Geological storage-
classication of storage options. Oil Gas Sci Technol Rev IFP 2005;44:93108.
[4] Rothman H, Greenshields R, Calle FR. The alcohol economy: fuel ethanol and
Brazilian experience. London: Francis Printer; 1983.
[5] Balat M. Current alternative engine fuels. Energy Sources 2005;27:56977.
[6] Pumphrey JA, Brand JI, Scheller WA. Vapour pressure measurements and
predictions for alcoholgasoline blends. Fuel 2000;79:140511.
[7] Sun H, Fang W, Guo Y, Lin R. Investigation of bubble-point vapor pressures for
mixtures of an endothermic hydrocarbon fuel with ethanol. Fuel
2005;84:82531.
[8] Cottrell FG. On the determination of boiling points of solutions. J Am Chem Soc
1919;41:7219.
[9] Rogalski M, Malanowski S. Ebulliometer modied for the accurate
determination of vapor liquid equilibrium. Fluid Phase Equilib 1980;5:97112.
[10] Li H, Han S, Teng Y. Bubble points measurement for system chloroform
ethanolbenzene by inclined ebulliometer. Fluid Phase Equilib
1995;113:18595.
[11] Antoine CCR. Tensions des vapeurs; nouvelle relation entre les tensions et les
tempratures. Comptes Rendus des Sances de lAcadmie des Sciences
1888;107:6814.778-80.836-7.
[12] Wagner W. New vapour pressure measurements for argon and nitrogen and a
new method for establishing rational vapour pressure equations. Cryogenics
1973;13:47082.
[13] Poling BE, Prausnitz JM, OConnell JP. The properties of gases and liquids. 5th
ed. Singapore: McGraw-Hill International Editions; 2001.
[14] Hull A, Golubkov I, Kronberg B, Van Stam J. Alternative fuel for conventional
spark ignition engines based on a standard gasoline, ethanol and other
oxygenates. Int J Engine Res 2005;7:20314.
[15] Wilson GM. Vaporliquid equilibrium. XI. A new expression for the excess free
energy of mixing. J Am Chem Soc 1964;86:12730.
[16] Renon H, Prausnitz JM. Local compositions in thermodynamic excess functions
for liquid mixtures. AIChE J 1968;14:13544.
[17] Abrams DS, Prausnitz JM. Statistical thermodynamic of liquid mixture: a new
expression for the excess gibbs energy of partly or completely miscible system.
AIChE J 1975;21:11628.