J. Chem.

Thermodynamics 37 (2005) 297–303

www.elsevier.com/locate/jct

(Liquid + liquid) equilibria of (water + butyric acid +

isoamyl alcohol) ternary system
Mehmet Bilgin a, S _
ß . Ismail Kırbasßlar a,*
, Önder Özcan b, Umur Dramur a

a
Engineering Faculty, Chemical Engineering Department, Istanbul University, Avcilar-Campus, 34320 Istanbul, Turkey
b
Mustafa Nevzat Ilac San. AS., 34530 Yenibosna, Istanbul, Turkey

Received 20 July 2004; received in revised form 3 September 2004; accepted 30 September 2004
Available online 13 November 2004

Abstract

(Liquid + liquid) equilibrium (LLE) data for the ternary system (water + butyric acid + isoamyl alcohol) have been determined
experimentally at T = (298.15, 308.15 and 318.15) K. Complete phase diagrams were obtained by determining solubility and the tie-
line data. Tie-line compositions were correlated by Othmer–Tobias method. The UNIFAC method was used to predict the phase
equilibrium in the system using the interaction parameters determined from experimental data between groups CH3, CH2, CH,
COOH, OH and H2O. It is found that UNIFAC group interaction parameters used for LLE could not provide a good prediction.
Distribution coefficients and separation factors were evaluated for the immiscibility region.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: Butyric acid; Isoamyl alcohol; Liquid + liquid equilibria; UNIFAC

1. Introduction Solvent extraction techniques employed in down-

Butyric acid and some of its esters are used in the
food and perfumes industries because of theirs aromatic
properties. Butyric acid also used in the pharmaceutical
industry. Many processes have already been proposed
for butyrate production by fermentation. Nevertheless,
fermentation processes have not yet been used commer-
cially because of low product mass concentration in the
fermentation mash (20 to 30 g Æ l
1) and also because of
the acetate being produced simultaneously with buty-
rate. Presently, butyric acid is mostly produced as
synthetic from petrochemicals. However, the use of bu-
tyric acid or its esters as additives, in particular in the
food or cosmetic industries, makes its origin important
since consumers have preferences for natural products.
* ble thermodynamic models, which allow the calcula-
Corresponding author. Tel.: +90 2124737070; fax: +90 212 5911
997.
E-mail address: krbaslar@istanbul.edu.tr (S _
ß . Ismail Kırbasßlar).
stream separations or integrated with fermentation of
organic acids are of economic importance in the
chemical industry. (Liquid + liquid) equilibrium data
of the related systems are needed for the design of
an efficient and productive extraction system and
many studies have been carried out to improve such
techniques [1–6].
A detailed evaluation of solvents for the extraction of
butyric acid is not available in the literature. Neverthe-
less, Zigová et al. [7–9] carried out screening of 14 sol-
vents for the extraction of butyric acid and a tertiary
amine in different diluents, C8 to C18 alcohols, dibutyl
ether, toluene, and vegetable oils, has been tested.
The real behaviour of fluid mixtures can be calculated
with the help of activity coefficients. The correct descrip-
tion of the dependence on temperature, pressure and
composition in multicomponent systems requires relia-
tion of these properties from available experimental
data [10]. The UNIFAC method (UNIFAC, Universal

0021-9614/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jct.2004.09.022
298 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303

Functional-Group Activity Coefficient) was developed TABLE 1

by Fredenslund et al. [11]. It is one of the best methods Densities (q) and refractive indexes (nD) at T = 293.15 K and boiling
temperatures at p = 101.325 kPa of the pure components [12]
in estimating activity coefficient to date. The UNIFAC
method for estimation of activity coefficients works on Compound q/kg Æ m
3 nD Tb/K
the concept that a liquid mixture may be considered a Expt. Lit. Expt. Lit. Expt. Lit.
solution of structural units from which the molecules Water 999.9 998.23 1.3324 1.3330 373.2 373.25
are formed rather than a solution of the molecules them- Butyric acid 957.6 957.70 1.3982 1.3980 436.67 436.68
selves. This procedure is known as group contribution Isoamyl alcohol 808.4 808.5 1.4075 1.4070 404.55 404.15
method. It has the advantage of being able to form a
very large number of molecules from a relatively small
system in order to observe the evaluation of the binodal
set of structural units. The structural units in the calcu-
curves and tie-lines. At each temperature, individual sol-
lation method are called subgroups. The mole fractions,
ubility curves were determined by the cloud-point
X Ei and X R
i of liquid + liquid equilibrium phases can be
method in the magnetically stirred equilibrium cell
calculated using the following equation:
[13]. The schema of the cell is given in figure 1. The
cEi X Ei ¼ cR R
i Xi ; ð1Þ end point was determined by observing the transition
where; E: extract (solvent) phase; R: raffinate (aqueous) from a homogenous to a heterogeneous mixture. The
phase; c: activity coefficient of the i component; i: com- mutual solubilities of the (water + isoamyl alcohol) sys-
ponent number. tem were determined by using cloud-point method, too.
In this study, the interaction parameters between A weighted amount (10 g) of one component was placed
CH3, CH2, CH, COOH, OH and H2O groups were used in the cell; than the other component added until a per-
to estimate the activity coefficients by the UNIFAC manent heterogeneity had been observed. Tie-line data
method. were obtained by preparing ternary mixtures (water +
This study is a part of a research program on the butyric acid + isoamyl alcohol) of known overall com-
recovery of butyric acid from dilute aqueous solutions positions lying within the two-phase region, in a flask.
using organic solvents. In this paper, we report the After being stirred vigorously, the mixture was allowed
LLE results for the ternary system (water + butyric to reach equilibrium in a shaker (4 h) at a constant tem-
acid + isoamyl alcohol) for which no such data were perature. Then the contents were immediately allowed
available. to enter the equilibrium cell equipped with an isothermal
jacket; after the complete separation of the phases, a
suitable amount (0.1 g) of each layer was removed for
detailed analysis. The temperature was measured out-
2. Experimental
side the cell by using an electronically controlled ther-
mostat with an accuracy of 0.1 K. All mixtures were
2.1. Materials
prepared by weighing with a Mettler scale accurate to
±10
4 g. The solvent was added by an automated micro-
Butyric acid and isoamyl alcohol were purchased
buret with an accuracy of 5 Æ 10
3 cm3.
from Merck and and were of 0.99 and 0.98 mass fraction
purity, respectively. Butyric acid and isoamyl alcohol
were used without further purification. GC analysis
did not detect any appreciable peaks of impurities.
Deionised water was further distilled before use. Refrac-
tive indexes were measured with Abbé-Hilger refractom-
eter; its stated accuracy is ±5 Æ 10
4. Densities were
measured with Anton Paar densimeter (Model 4500) in
±10
4 precision. Boiling point measurements were ob-
tained by using a Fischer boiling point apparatus. The
estimated uncertainties in the density and boiling point
measurements were ±10
4 g Æ cm
3 and 0.1 K, respec-
tively. The measured physical properties are listed in
table 1, along with literature values [12].

3. Procedure

Three different temperatures (298.15, 308.15 and

318.15) K were selected to study the ternary equilibrium FIGURE 1. (Liquid + liquid) equilibrium cell.
 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303 299

4. Analysis di, for water (i = 1) and butyric acid (i = 2) and separa-

The liquid samples were analysed by a Gas
Chromatograph (Hewlett–Packard GC, Model 6890
Series), equipped with thermal conductivity detector
(TCD) for the quantitative determination of water, bu-
tyric acid and isoamyl alcohol. A 15 m long HP-Plot Q
column (0.320 lm i.d., 0.2 lm film thickness) was used
with a temperature-programmed analysis. Column tem-
perature, (343.15 to 493.15) K at 20 K Æ min
1, at
493.15 K (3 min) injection mode, split ratio 100:1;
detector, TCD; injector and detector temperature,
523.15 K; carrier gas, nitrogen 1 cm3 Æ min
1; injected
volume of 0.3 microliter of liquid sample.
5. Results and discussion
The experimental tie-line data of (water + butyric
acid + isoamyl alcohol) ternaries at T = (298.15, 308.15
and 318.15) K, were given in table 2. It was found that
isoamyl alcohol was very little soluble in water but mis-
cible with butyric acid. The experimental and predicted
tie-lines and binodal curves for the system at each tem-
perature were plotted in figures 2–4.
To show the selectivity and extraction strength of the
solvent to extract butyric acid, distribution coefficients,
tion factors, s, were calculated as follows:
d i ¼ wi3 =wi1 ; ð2Þ
distribution coefficient of butyric acid
s¼ ¼
distribution coefficient of water
ðw23 =w21 Þ=ðw13 =w11 Þ; ð3Þ
where w23 and w21 are the butyric acid concentrations in
solvent-rich and water rich phases, respectively, and w13
and w11 are the water concentrations in solvent-rich and
water rich phases, respectively.
The distribution coefficients and separation factors
for each temperature are given in table 3. The effective-
ness of extraction of butyric acid by isoamyl alcohol is
given by its separation factor, which is an indication
of the ability of isoamyl alcohol to separate butyric acid
from water. This quantity is found to be greater than 1
(separation factors varying between 38 and 159) for the
system reported here, which means that extraction of
butyric acid by isoamyl alcohol is possible. The separa-
tion factor is not constant over the whole two-phase re-
gion. The extracting power of the solvent at each
temperature, plots of d2 vs. w21 and s vs. w21 are shown
in figures 5 and 6, respectively.

TABLE 2
Experimental tie-line data of (water + butyric acid + isoamyl alcohol) ternary system
Water-rich phase (mass fraction) Solvent-rich phase (mass fraction)
Water Butyric acid Isoamyl alcohol Water Butyric acid Isoamyl alcohol
T = 298.15 K
0.9677 0.0082 0.0241 0.0852 0.1148 0.8000
0.9563 0.0195 0.0242 0.0849 0.2380 0.6771
0.9470 0.0286 0.0244 0.0938 0.3250 0.5812
0.9379 0.0377 0.0244 0.1100 0.4105 0.4795
0.9277 0.0477 0.0246 0.1253 0.4549 0.4198
0.9238 0.0515 0.0247 0.1420 0.5115 0.3465
0.9197 0.0554 0.0249 0.1601 0.5547 0.2852
T = 308.15 K
0.9716 0.0093 0.0191 0.0966 0.1124 0.7910
0.9613 0.0196 0.0191 0.1055 0.2185 0.6760
0.9469 0.0338 0.0193 0.1227 0.3108 0.5665
0.9420 0.0386 0.0194 0.1269 0.3851 0.4880
0.9334 0.0472 0.0194 0.1483 0.4332 0.4185
0.9298 0.0506 0.0196 0.1611 0.4984 0.3405
0.9252 0.0551 0.0197 0.1809 0.5351 0.2840
T = 318.15 K
0.9715 0.0093 0.0192 0.1135 0.1164 0.7701
0.9612 0.0195 0.0193 0.1250 0.2200 0.6550
0.9497 0.0310 0.0193 0.1359 0.3002 0.5639
0.9420 0.0386 0.0194 0.1559 0.3800 0.4641
0.9324 0.0481 0.0195 0.1693 0.4304 0.4003
0.9262 0.0543 0.0195 0.1946 0.4803 0.3251
0.9237 0.0567 0.0196 0.2236 0.5154 0.2610
300 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303

Butyric Acid

1.0
0

0
5

0.7
0.2

5
298.15 K

0
0.5

0.5
0
5
0.7

0.2
5
0
1.0

0
Isoamyl
Water Alcohol
0.00 0.25 0.50 0.75 1.00

– experimental
FIGURE 2. Ternary diagram for experimental LLE of {water(1) + butyric acid (2) + isoamyl alcohol (3)} at T = 298.15 K; –
solubility curve; –n– experimental tie-line data; –e– calculated (UNIFAC) tie-line data.

Butyric Acid
1.0
0

0
5

0.7
0.2

308.15 K
0
0.5

0.5
0
5
0.7

0.2
5
0
1.0

Water Isoamyl
0.00 0.25 0.50 0.75 1.00 Alcohol

– experimental
FIGURE 3. Ternary diagram for experimental LLE of {water(1) + butyric acid (2) + isoamyl alcohol (3)} at T = 308.15 K; –
solubility curve; –n– experimental tie-line data; –e– calculated (UNIFAC) tie-line data.

The reliability of experimentally measured tie-line tion (4). Othmer–Tobias plots were constructed of the
data can be ascertained by applying the Othmer–Tobias ln ((1
w33)/w33) vs. ln ((1
w11)/w11). The parameters
correlation [14] at each temperature as shown in equa- of this correlation were given in table 4 and the correla-
 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303 301

Butyric Acid

1.0
0

0
5

0.7
0.2

5
318.15 K

0
0.5

0.5
0
5
0.7

0.2
5
0
1.0

0
Water Isoamyl
0.00 0.25 0.50 0.75 1.00 Alcohol


FIGURE 4. Ternary diagram for experimental LLE of {water (1) + butyric acid (2) + isoamyl alcohol (3)} at T = 318.15 K; – – experimental
solubility curve; –n– experimental tie-line data; –e– calculated (UNIFAC)

TABLE 3
Distribution coefficients (di) of (water and butyric acid) and separation 19
factors (s)
17
d1 d2 s
T = 298.15 K 15
0.088 14.00 159
0.089 12.21 137 13
d2

0.099 11.36 115

0.117 10.89 93 11
0.135 9.54 71
0.154 9.93 65 9
0.174 10.01 58
7
T = 308.15 K
0.099 12.03 121
5
0.110 11.16 102
0.130 9.20 71 0.00 0.02 0.04 0.06
0.135 9.98 74 w21
0.159 9.18 58
FIGURE 5. Distribution coefficient d2 of butyric acid as a function
0.173 9.85 57
of the mass fraction w21 of butyric acid in aqueous phase. T =
0.196 9.70 50
(–e, 298.15; –h–, 308.15; –n–, 318.15) K.
T = 318.15 K
0.117 12.55 107 lnðð1
w11 Þ=w11 Þ ¼ a þ b  lnðð1
w33 Þ=w33 Þ; ð4Þ
0.130 11.28 87
0.143 9.67 68 where; w11: weight fraction of water in the water-rich
0.165 9.84 59 phase; w33: weight fraction of isoamyl alcohol in the sol-
0.182 8.95 49
vent-rich phase; a and b are constants of the equation
0.210 8.85 42
0.242 9.09 38 (4).

5.1. Prediction of equilibria by the UNIFAC method

tion was shown in figure 7 at each temperature. The cor-
relation factor (R2) being approximately unity indicates The equilibrium data of the ternary mixture were
the degree of consistency of the related data. predicted by UNIFAC method [11] using the interac-
302 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303

TABLE 5
150 UNIFAC group parameters for prediction tie-lines data [15]
m n amn anm
130 CH, CH2, CH3 COOH 663.5 315.3
COOH OH
151 199
110 OH CH, CH2, CH3 156.4 986.5
s

H2O CH, CH2, CH3 300 1318

90 H2O COOH
14.09
66.17
H2O OH
229.1 353.5
70

50
TABLE 6
30 UNIFAC Rk and Qk parameters [15]
0.00 0.01 0.02 0.03 0.04 0.05 0.06 Group Main group no. Rk Qk
w21 CH3 1 0.9011 0.848
CH2 1 0.6744 0.54
FIGURE 6. Separation factor, s, as a function of the mass fraction of COOH 20 1.3013 1.224
butyric acid in the aqueous phase, w21. T = (–e–, 298.15; –h–, 308.15; OH 5 1 1.2
–n–, 318.15) K. H2O 7 0.92 1.4

tion parameters between CH3, CH2, CH, COOH, OH cause the experimental and predicted data are practi-
and H2O obtained by Magnussen et al. [15]. The values cally independent of temperature, the discrepancy
of the UNIFAC parameters for LLE prediction are between these values was observed not to change with
summarised in tables 5 and 6. As shown in figures 2– temperature.
4, LLE relations predicted by the UNIFAC method The root mean square deviations (rmsd) are
cannot adequately fit the experimental LLE data. Be- calculated from the difference between the experimental
data and the predictions of the UNIFAC model at each
temperature according to the following formula:
TABLE 4 8 " #, )1=2
Constants of Othmer–Tobias equation for the (water + butyric <X X X
2
acid + isoamyl alcohol) ternary system (R2: regression coefficient) rmsd ¼ ðX i;exp
X i;calcd Þ 6n ;
: k i j
T/K a b R2
298.15 0.4321
2.7649 0.9833 ð5Þ
308.15 0.4593
2.8546 0.9653
318.15 0.4779
2.8876 0.9702 where; j is water or butyric acid, i is the solvent or
aqueous phase, and k = 1, 2, 3,. . .,n (tie-lines). The

-2.0
1

-2.5
0.8
ln((1-w11)/w11)

w23/(w23+w13)

-3.0 0.6

0.4
-3.5

0.2
-4.0
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 0
ln((1-w33 )/w33 ) 0 0.05 0.1 0.15
w21 /(w21 +w11 )
FIGURE 7. Othmer–Tobias plot of the (water + butyric acid + isoa-
myl alcohol) ternary system at T = (–e–, 298.15; –h–, 308.15; –n–, FIGURE 8. Selectivity diagram at investigated temperature values
318.15) K. (free-solvent basis) T = (e, 298.15; h, 308.15; n, 318.15) K.
 M. Bilgin et al. / J. Chem. Thermodynamics 37 (2005) 297–303 303

UNIFAC method was used to correlate the Acknowledgement

experimental data at T = (298.15, 308.15 and 318.15)
K with rmsd values of 0.3132, 0.2720 and 0.2861, This work was supported by The Research Fund of
respectively. Istanbul University. Project number: 83/15052003.

5.2. Selectivity References

[1] A.S. Kertes, C.J. King, Biotechnol. Bioeng. 28 (1986) 269–282.

Selectivity diagrams on a solvent-free basis are ob- [2] J.M. Wardell, C.J. King, J. Chem. Eng. Data 23 (1978) 144–148.
tained by plotting w23/(w23 + w13) vs. w21/(w21 + w11) [3] U. Dramur, B.J. Tatli, J. Chem. Eng. Data (1993) 23–25.
for each temperature in figure 8. The effect of a temper- [4] M.A. Fahim, S.A. Al-Muhtaseb, I.M. Al-Nashef, J. Chem. Eng.
ature change on the selectivity values was found to be Data (1997) 183–186.
[5] J.A. Briones, J.C. Mullins, M.C. Thies, Ind. Eng. Chem. Res. 33
insignificant.
(1994) 151–156.
[6] A. Arce, A. Blanco, P. Souza, I.J. Vidal, Chem. Eng. Data 40
(1995) 225–229.
[7] D. Vandák, J. Zigová, E. Šturdik, Š. Schlosser, Process Biochem.
6. Conclusion 32 (1997) 245–251.
[8] J. Zigová, E. Šturdik, J. Ind. Microbiol. Biotechnol. 24 (2000)
153–160.
It is concluded that isoamyl alcohol may serve as an [9] J. Zigová, D. Vandák, Š. Schlosser, E. Šturdik, Sep. Sci. Technol.
adequate solvent to extract butyric acid from its dilute 31 (1996) 2671–2684.
aqueous solutions although isoamyl alcohol shows [10] M.F. Hegazi, J. Chem. Tech. Biotechnol. 34A (1984) 251–256.
some solubility in water. It is shown that butyric acid [11] A. Fredenslund, R.L. Jones, J.M. Prausnitz, AIChE J. 21 (1975)
1086–1099.
is more soluble in solvent phase than that in water [12] D.R. Lide, Handbook of Chemistry and Physics, CRC Press Inc,
phase. It was found that the UNIFAC method good- Boca Raton FL, 1992, pp. 3–470.
ness was not satisfactory to the experimental data. Fi- [13] S. Cehreli, Braz. J. Chem. Eng. 19 (2002) 45–53.
nal conclusion is that the complete break of conjugate [14] D.F. Othmer, P.E. Tobias, Ind. Eng. Chem. 34 (1942) 690–692.
phases was never hindered by density, viscosity and [15] T. Magnussen, P. Rasmussen, A. Fredenslund, Ind. Eng. Chem.
Process. Des. Dev. 20 (1981) 331–339.
interfacial tension related phenomena during the set-
tling process. JCT 04-164