Fluid Phase Equilibria 157 Ž1999.

29–40

Molecular thermodynamics of salt effect in vapor–liquid equilibrium

of ethanol–water systems
Ren Yi Sun
Department of Chemistry, Henan UniÕersity, Kaifeng, Henan, 475001, China
Received 1 October 1997; accepted 4 December 1998

Abstract

Scaled particle theory was used to derive a general expression for the salt effect parameter, K, of isobaric
vapor–liquid equilibrium for ethanol–water-1-1 type electrolytic systems, which appears in the Furter equation.
This expression was essentially a sum of two terms: 1, the hard sphere interaction term calculated by
Masterton–Lee’s equation, 2, the soft sphere interaction term calculated by Y. Hu’s molecular thermodynamical
model, in which the diameters of nacked ions were replaced by that of solvated ions, the solvation coefficients
Ži.e., in the radio of the latter to the former. were taken to be adjustable parameters, their magnitude implies the
ionic solvation rules. A correlation equation for the local dielectrical constant around central ions with liquid
concentration was obtained by mapping out experimental points. The calculated salt effect parameters of 9
ethanol–water-1–1 type electrolytic systems were in good agreement with the literature values within the wide
range of liquid concentration. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: Vapor–liquid equilibrium; Salt effect parameter; Scaled particle theory

1. Introduction

The original equation for salt effect in vapor–liquid equilibrium, proposed by Furter w7x, predicts
the effect of a salt dissolved in a mixed solvent on vapor–liquid equilibrium when the composition of
the solvent is held constant, that is
s
log Ž a 12ra 12 . s Kx e Ž1.
s
where a 12 and a 12 are relative volatility of nonelectrolyte Ž e.g., ethanol. to water without and with
salt present, respectively, in this paper, x i is mole fraction of a component i in liquid phase calculated
on a salt-free basis, i.e., x i s n irŽ n1 q n 2 ., n is the amount of substance, the subscript i may be one
of 1, 2, 3, 4 and e, which represents ethanol, water, univalent cation, univalent anion and 1–1 type

0378-3812r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 3 8 1 2 Ž 9 9 . 0 0 0 0 2 - 3
30 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

electrolytic molecule, respectively, so x 1 q x 2 s 1 and x 3 s x 4 s x e . According to the theoretical

background from which the equation was derived w11x, K, the salt effect parameter, remains constant
when the relative proportions of the two volatile components forming the mixed solvent are held
constant.
The Furter equation is different from the Setschenow equation w14x in the form. However, the
essence of salt effect predicted by two equations is same, that is the effect of salt to the activity
coefficients of the components in liquid phase. In the applicable range, the Furter equation is valid at
any solvents proportion, while the Setschenow equation can be applied only at very low nonelec-
trolytic concentration. Thus, in the calculation of the Setschenow salting coefficient, k s , the
self-interaction of nonelectrolytic molecules can be ignored. In the calculation of the Furter salt effect
parameter, K, however, various interactions related to nonelectrolytic molecules in the solution must
be taken into account. The theoretical calculation of k s has been reported many times Ž e.g., Refs.
w1,13,14,16x., however, the theoretical calculation of K has not appeared in the literatures up to now.
In this work a theoretical calculation method of the Furter salt effect parameter at any appointed
solvents proportion was proposed based on the scaled particle theory. Some significant information
about the microstructures of electrolytic solution were discovered from the calculation of the Furter
salt effect parameter.

2. Expression for the salt effect parameter

In practice, Eq. Ž1. is found to be valid only at low salt concentrations, i.e., as x e ™ 0. To obtain an
expression for K, it is convenient to differentiate Eq. Ž 1. with respect to x e at constant x 1. When x 1
is constant, the proportion of solvents is unchanged, it follows that
s
K s Ž E log a 12rE x e . x 1 , x e ™ 0 Ž2.
According to thermodynamics, the following equation can be deduced
s
ln a 12 q ln Ž w 1srw 2s . s Ž mr1 y m r2 . rRT Ž3.
where w s is the fugacity coefficient of the vapor phase in the system containing salt, mr is the
residual chemical potential. Eq. Ž 3. is inserted in Eq. Ž 2. , for the values of w s are almost unchanged
and may be regarded as a constant in the isobaric vapor–liquid equilibrium system when x 1 is fixed
and x e is changing, thus, the result is

1 E Ž mr1 y mr2 . D Hv1 y D Hv2 ET

Ks
2.3 RT E xe T , x 1 , x e™0
q
2.3 RT 2 ž /
E xe x 1 , x e™0
Ž4.

where D Hv is the partial molar vaporation heat of components in the liquid phase. In the practice, the
last term on the right side of Eq. Ž4. may be regarded as zero for the isobaric vapor–liquid
equilibrium of ethanol–water–salt system because the difference of D Hv1 and D Hv2 is very small.
According to the scaled particle theory of Pierotti w20,21x, mri Ž i s 1,2. can be written as

mri s Gis q Gih Ž i s 1,2 . Ž5.

 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40 31

where Gih is the term of the hard sphere interaction and Gis is that of the soft sphere interaction. The
former corresponds to the process ŽI. , in which the molecules of the system expand into the hard
spheres such as actual molecules in diameter from geometrical points, while the latter corresponds to
the process Ž II., in which the hard spheres are charged with various potential energies and become the
soft spheres so that the actual solution can be imitated. Inserting Eq. Ž 5. into Eq. Ž 4. , the last term on
the right side of Eq. Ž4. is omitted for the isobaric vapor–liquid equilibrium of ethanol–water–salt
system, so that following results are obtained
K s K a q Kb Ž6.
1 E G 1h y G 2h
ž /
Kas Ž7.
2.3 RT E xe
T , x 1 , x e™0

1 E G 1s y G 2s
ž /
Kb s Ž8.
2.3 RT E xe
T , x 1 , x e™0

where K a and K b represent the contributions of the hard sphere interaction and the soft sphere
interaction to K, respectively. The problem now becomes one of deriving general expression for K a
and K b in terms of parameters characteristic to ethanol, water and the ions of salt.
Expression for K a : The hard sphere interaction term, Gih is the increment of the residual chemical
potential in the process ŽI. . In order to calculate Gih for the system of ethanol 1 –water2 –univalent
cation 3 –univalent anion 4 the Masterton–Lee equation was adopted
Gih 3t 2 si t 1 si 3t 2 si
s yln Ž 1 y t 3 . q 1q q Ž i s 1,2 . Ž9.
RT Ž1 y t 3 . t2 2 Ž1 y t 3 .
p 4
tn s
6
Ý r j sj n Ž 10.
js1

where s is the hard sphere diameter, and r is the number density of a solution species, which can be
represented as
r j s Lx jrV Ž j s 1,2,3,4 . Ž 11.
where L is Avogadro’s number, V is the volume of the solution just containing 1 mole of solvents
Žethanolq water. given by
4
Vs Ý x jVj Ž 12.
js1

where Vj is the partial molar volume of species j in the solution. According to the Gibbs–Duhem
equation, upon differentiating Eq. Ž12. with respect to x e at constant temperature and fixed x 1,
considering x 3 s x 4 s x e and x 1 q x 2 s 1, it is found that
EV
ž /
E xe T , x 1 , x e™0
sf Ž 13.
32 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

where f is the partial molar volume of 1–1 type electrolyte at infinite dilution. Utilizing Eqs. Ž 11.
and Ž13., the following results can be obtained
Er j Lx j f
ž /
E xe T , x 1 , x e™0
sy
Vm2
Ž j s 1,2 . Ž 14.

Er j L
ž /
E xe T , x 1 , x e™0
s
Vm
Ž j s 3,4 . Ž 15.

where Vm is the molar volume of the salt-free solution with concentration x 1. Upon differentiating
Eq. Ž10. with respect to x e at constant T and x 1, and utilizing Eqs. Ž14. and Ž15., it follows that
Etn
ž /
E xe T , x 1 , x e™0
s yn A n Ž 16.

pL x 1 s 1n q x 2 s 2n
yn s P Ž 17.
6 Vm
s 3n q s4n f
An s y Ž 18.
x 1 s 1n q x 2 s 2 n
Vm
and it is easily seen that
limtn s yn Ž 19.
x e™0

Inserting Eq. Ž 9. into Eq. Ž7. and utilizing Eqs. Ž16. and Ž19., we can obtain the results as follows

1° ~ Ž s12 y s 22 . 3 y1 3 y2 3 y2
2

Ka s
2.3 ¢ 1 y y3
A1 q Ž s 1 y s 2 .
1 y y3
qŽs 2
1 y s 2
2
. ž 1 y y3 / A2

y3 3 y1 3 y2 3 y2
2 ¶
A 3•
q
1 y y3
Žs 2
1 y s 2
2
. 1yy
3
q Ž s1 y s 2 .
1 y y3
qŽs 2
1 y s 2
2
. ž 1 y y3 / ß Ž 20.

where yn and A n are defined as Eqs. Ž17. and Ž18., respectively.

Expression for K b : In order to calculate the soft sphere interaction term Gis , Hu and Prausnitz w10x
proposed a simplified radial distribution function in which the first coordination circle around central
molecules was taken as in order, out of the circle as random mixing. Then it follows that
ri j )) 3 y si 3j
g i j Ž r . s H Ž r y ri j )) . q d Ž r y ri j ) . Ž 21.
3ri j ) 2

where H represents the Heaviside function, d represents the Dirac function, r is the radius about
central molecule i, si j is the collision diameter of i–j pair of molecules, ri j ) is the center to center
distance between the central molecule and the molecules in the first coordination shell, ri j )) is the
distance between the center of the central molecule and the out boundary of the first coordination
 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40 33

shell. Both of them are proportional to the collision diameter of i–j pair of molecules, i.e.,
ri j )) s K )) si j , ri j ) s K ) si j , the proportional coefficients K )) and K ) may be regarded as
universal constants under the condition of liquid density. Hu et al. w9x suggested that K ) s 1.150,
K )) s 1.575.
In the liquid mixture of ethanol–water-1–1 type electrolytic systems the interaction between
molecules of ethanol and water includes dipole–dipole and dipole-induced dipole in addition to the
Lennard–Jones energy. Thus the interaction energy of the molecules of ethanol and water can be
written as

si j 12 si j 6 a i j m2j q a j m2i 2 m2i m2j

´ Ž r . s 4´ i j
Ž1.
ij ž r / yž r / y
r6
y
3kT r6
Ž i s 1,2; j s 1,2 . Ž 22.

(
where ´ i j s ´ i ´ j , ´rk is the energy parameter of molecules, k is the Boltzmann constant, m and a
are the dipole moment and the polarizability of molecules, respectively. The first, second and third
terms on the right side of Eq. Ž22. represent the Lennard–Jones, dipole-induced dipole and
dipole–dipole interaction energy, respectively.
For the interaction of nonelectrolytic molecules and ions the following energy function was
adopted, in which charge-dipole and charge-induced dipole interaction energies are involved besides
the Lennard–Jones energy:

si j 12 si j 6 e 2 m 2i e 2a i
´ iŽ2.
j s 4´ i j žr/ žr/ y y
3kTr 4 D 2
y
2 r 4D2
Ž i s 1,2; j s 3,4 . Ž 23.

where e is the charge of electron, D is the local dielectrical constant around a central ion. The first,
second and third terms on the right side of Eq. Ž23. represent the Lennard–Jones, charge-dipole and
charge-induced dipole interaction energy, respectively.
In the process Ž II. , the increment of the Helmhotz energy of the system, D A, is approximately
equal to the increment of its internal energy, DU. It follows that

1 2 2 ` 2 4 `
2 2
D As LÝ Ý n iH r j ´ Ž r . g i j Ž r . 4p r d r q L Ý
Ž1.
ij Ý n iH r j ´ iŽ2.
j Ž r . g i j Ž r . 4p r d r Ž 24 .
2 is1 js1 0 is1 js3 0

We can obtain the following equation according to Eq. Ž 21.

` 4 `
H0 r ´ j ij Ž r . g i j Ž r . 4p r 2 d r s p ž ri j )) 3 y si 3j / r j ´ i j Ž r ) . q H r j ´ i j Ž r . 4p r 2 d r Ž 25.
3 r i j ))

In Eq. Ž25. ´ i j Ž r . can be replaced by ´ iŽ1.

j r or ´ i j r . Inserting Eqs. 22 and 23 into Eq. 24 ,
Ž . Ž2. Ž . Ž . Ž . Ž .
utilizing Eq. Ž25., and according to the following thermodynamical relation

Gis 1 ED A
RT
s
LkT ž / E ni T ,V , nw i x
Ž i s 1,2 . Ž 26.
34 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

we can obtain

Gis p 4 2 rj 2
sy 5.140 Ý r i ´ i j si 3j q 2.017 Ý 3 ž m2i m2j q a j m2i q a i m2j /
RT kT js1 js1 s ij 3kT

4 rj e 2 ai m2i
q2.216 Ý
js3 D 2si j ž 2
q
3kT / Ž i s 1,2 . Ž 27.

In Eq. Ž27. the energy parameters of cations and anions are evaluated from the Mavroyannis–Stephen
theory w15x. It follows that
´ jrk s 2.28 = 10y8a j3r2 Z j ) 1r2sy6
j Ž j s 3,4 . Ž 28.
where Z j ) is the total number of electrons in the ion j. Inserting Eq. Ž27. into Eq. Ž8. and utilizing
Eqs. Ž14. and Ž15. , the result is

pL 2 2 xj 2 f
Kb s y
2.3kT ½ Ý Ý Žy1.
is1 js1
i
5.140 x j ´ i j si 3j q 2.017
s 3
ij
ž 3kT
m2i m 2j q a j m2i q a i m2j / Vm2

2 4 5.140 ´ i j si 3j 2.216 e 2 a i m2i

yÝ Ý
is1 js3
Ž y1.
i

Vm
q
Vm D 2si j ž 2
q
3kT /5 . Ž 29 .

The salt effect parameter, K, can be calculated by adding K a and K b . Thus it is necessary to
know: Ž1. f Ž the partial molar volume of 1–1 type electrolyte at infinite dilution., Vm Žthe molar
volume of the salt-free solution. and D Žthe local dielectrical constant around central ions. at
appointed liquid concentration x 1; Ž2. the energy parameters ´rk and the dipole moments m of
ethanol and water; Ž 3. the diameters s and the polarizabilities a of various molecules and ions.
According to thermodynamics, Vm should be calculated by using the partial molar volumes of
ethanol and water in the salt-free solution. It has been found practically, however, that the substitution
of the molar volumes of the pure solvents for their partial molar volumes in the calculation of K do
not cause its obvious deviation. Consequently, a linear equation was taken to calculate Vm approxi-
mately so as to avoid the complication in the calculation of K, which is given by
Vm s x 1Vm1 q x 2Vm 2 Ž 30.
where Vm1 and Vm2 represent the molar volumes of ethanol and water, respectively. In order to
calculate K, we need to know the values of f in ethanol–water mixture. Since there are short of
these data in the literatures, on the other hand, it has been found practically that the values of f do
not influence K b almost, and do K a weakly. All things considered, an interpolating correspondence
method, as a first approximation, was adopted in this paper so as to provide a general estimate of the
values of f in ethanol–water system. Millero w17x has used the correspondence method to estimate
f Žion. s Žthe partial molar volume of ions in a certain solvent. form f Žion. w Žthe partial molar volume
of ions in water.. This empirical method can be represented by
s w
f Ž ion . s af Ž ion . q b Ž 31.
 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40 35

Table 1
Molecular parameters and molar volume of solvents for salt effect parameters
Moleculesa VmŽml moly1 .
˚.
s ŽA a Ž10y24 ml. ´ r k ŽK. m rDebye
EtOH 4.38 5.130 412.97 1.68 60.98
H 2O 2.75 1.470 150.00 1.84 18.54
a
Taken from Ref. w23x.

where a and b are solvent-dependent constants given by plotting f Ž ion. s against f Žion. w. By
utilizing the results of Kawaizumi and Zana w12x, and Mukerjee w19x we obtained a s 0.96, b s y8.3
for the monovalence ions in ethanol. Assuming that Eq. Ž31. is still valid for ethanol–water mixture
with a fixed concentration x 1, it is evident that a and b should be two variations with x 1, and
moreover, a s 1, b s 0 when x 1 s 0 and a s 0.96, b s y8.3 when x 1 s 1. And further, as a first
approximation, assuming that both of a and b are the linear variations with x 1, i.e., a s 1 y 0.04 x 1,
b s y8.3 x 1. Then the partial molar volume of monovalence ions in ethanol–water solution with a
specified composition can be estimated by Eq. Ž 31. and the partial molar volume of 1–1 type
electrolytes can be obtained by using the additivity, viz.,
f s f 3 q f4 Ž 32.
Tables 1 and 2 show the values of Vm of the pure solvents Ž ethanol and water. , the values of
f Žion. w at 758C and the relevant particle’s parameters s , a , m and ´rk. The temperature of
ethanol–water–salt systems in isobaric vapor–liquid equilibrium at atmosphere waves at about 808C
except in the end of water-rich region. Therefore the values of Vm were calculated using the densities
of ethanol and water at 808C.
Considering the solvation of ions in the ethanol–water solution, the solvation coefficient h was
taken to represent the ratio of the solvated ion to the nacked ion in diameter and the following
assumptions were adopted.

Table 2
Ionic parameters and conventional partial volume in water for salt effect parameters
Ions a f c Žion. w
Žml moly1 .
˚.
s ŽA a Ž10y24 cm3 . Z) h Žh .
Lq
i 1.42 0.075 2 3.55 3.55 y1.8
q
Na 1.90 0.210 10 2.16 2.18 0.8
Kqi 2.66 0.870 18 1.41 1.39 10.3
NHq 4 2.96 b 1.709 10 1.32 19.5
Cly 3.62 3.020 18 1.27 1.12 17.4
Bry 3.96 4.170 36 1.10 1.07 25.1
Iy 4.32 6.280 54 1.00 1.02 37.2
a
Taken from Ref. w4x.
b
Taken from Ref. w6x.
c
Taken from Ref. w18x.
36 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

Ž1. Different species of ions in the solution have different solvation coefficients. Generally, the
solvation coefficient of the small ion is greater than the large ion.

Table 3
Calculated salt effect parameters of isobaric VLE for ethanol–water–1–1 type electrolytic systems at 100 kPa
x1 K exp K cal Ka Kb Relative deviation %
a
LiCl 0.192 2.66 2.71 5.24 y2.53 1.87
0.293 2.13 a 1.96 4.79 y2.83 y7.98
0.494 1.50 a 1.46 4.13 y2.67 y2.66
0.700 1.60 a 1.60 3.82 y2.22 0.00
LiBr 0.192 1.57 a 2.00 5.06 y3.06 27.39
0.494 1.43 a 1.40 4.06 y2.66 y2.09
NaCl 0.240 3.46 b 3.68 5.22 y1.54 6.36
0.248 3.62 b 3.46 5.19 y1.73 y4.42
NaBr 0.192 2.79 a 3.26 5.26 y2.00 16.85
0.248 2.87 b 2.91 5.03 y2.12 1.39
0.494 2.41a 2.27 4.22 y1.95 y5.81
0.597 2.40 a 2.30 4.04 y1.74 y4.17
0.700 2.36 a 2.51 3.97 y1.46 6.36
KCl 0.193 3.28 a 3.88 5.36 y1.48 18.29
0.487 2.77 a 2.82 4.35 y1.53 1.81
KBr 0.206 2.89 c 3.09 5.12 y2.03 6.92
0.311 3.02 c 2.98 4.78 y1.80 y1.32
KI 0.196 2.19 a 1.95 5.08 y3.13 y10.96
0.309 2.33 d 2.30 4.73 y2.43 y1.29
0.494 2.79 a 2.60 4.24 y1.64 y6.81
0.597 2.77 a 2.83 4.07 y1.24 2.17
0.700 3.03 a 3.26 4.09 y0.83 7.59
NH 4Cl 0.143 2.50 b 2.21 4.69 y2.48 y11.60
0.223 2.20 b 2.25 4.51 y2.26 2.27
0.313 2.19 b 2.23 4.27 y2.04 1.83
0.367 2.29 b 2.29 4.14 y1.85 0.00
0.451 2.35 b 2.24 3.96 y1.72 y4.68
0.527 2.23 b 2.16 3.83 y1.67 y3.14
NH 4 Br 0.154 1.44 d 1.30 4.45 y3.15 y9.72
0.206 1.61c 1.47 4.36 y2.89 y8.69
0.246 1.67 d 1.71 4.26 y2.55 2.40
0.278 1.67 d 1.75 4.22 y2.47 4.79
0.305 1.83 c 1.79 4.15 y2.36 y2.19
0.309 1.94 c 1.81 4.15 y2.34 y6.70
0.378 1.99 d 2.01 4.02 y2.01 1.00
0.520 2.45d 2.24 3.75 y1.51 y8.57
0.662 2.75d 2.75 3.70 y0.95 0.00
a
Taken from Ref. w22x.
b
Taken from Ref. w11x.
c
Taken from Ref. w2x.
d
Taken from Ref. w3x.
 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40 37

Ž2. Ionic solvation is involved in the soft-sphere interaction. Therefore the diameter of ions in the
calculation of K b should be replaced by the diameter of the solvated ions, i.e., the collision diameter
of molecules and ions in Eq. Ž29. should be calculated by

1
si j s Ž s i q hj s j . Ž i s 1,2; j s 3,4 . . Ž 33.
2

In water, the molecules of water around a central ion will certainly make uniform arrangement
because there is a very strong electrostatic field near the central ion so that it faces the dipole of water
to be orientated along with itself and a dielectrical saturation state is formed. Thus the local
dielectrical constant D around central ions is greatly reduced to 2–6 w5x. In the solution of ethanol and
water, the values of D should be a function of liquid concentration x 1. We found the values of D
wave at about 3 in most regions of liquid concentration. However, the values of K are very sensitive
to the chosen values of D. Therefore the solvation coefficient, h , was taken to be an adjustable
parameter. Thus, the values of D at various experimental points can be obtained from the initial
values of h by the experimental values of K. Then, the initial values of h were adjusted so that the
values of D of all the experimental points form the most smooth possible curve and the equation of
the curve may be represented by a quadratic function. The result is

D s 2.554 q 2.530 x 1 y 2.195 x 12 Ž 34.

Thus, the values of D and x 1 were correlated by Eq. Ž34..
According to the above analysis, the calculated salt effect parameters of 9 ethanol–water-1–1 type
salt systems in the isobaric vapor–liquid equilibrium by means of Eqs. Ž 6. , Ž 20. , Ž 29. – Ž 32. and Ž 34.
are shown in Table 3. It is seen that the calculated salt effect parameters, K, are in good agreement
with the experiments.

3. Discussion

In Table 2 the values of h were used in this work, and the values of Žh . are the hydration
coefficients, which were obtained by means of calculation with the ionic primary hydration numbers
based on the assumption that the hydration ion is a solid sphere w8x. The comparison of h and Žh .
represents the values of h very close to the values of Žh . except the obvious deviation of Cly, The
solvation rule as above mentioned in the assumption 1 is roughly reflected from the values of h
considering the difference of the hydration of ions in water and the solvation of ions in ethanol–water
solution.
Eq. Ž34. indicates that the values of D increases along with x 1 in the end of the water-rich region
of ethanol–water solution. The result is shown in Fig. 1, in which the points represent the values of D
calculated by the experimental values of K, and all of the points form the most smooth possible curve
from adjusting the values of h. There is a very strong electrostatic field around a central ion in the
solution. It will force the molecules of water in the vicinity of the central ion to make directional
arrangement so that the local dielectrical constant D decreases greatly. However, the molecules of
38 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

Fig. 1. The local dielectrical constant of ethanol–water solution.

ethanol are very strong polar molecules, and it is very difficult for ethanol molecules to be driven out
from the vicinity of a central ion. The directional arrangement of water molecules will certainly be
disturbed by ethanol molecules. Thus, the dielectrical saturation will be weakened so that the values
of D gradually increase with the increase of ethanol concentration in the end of the water-rich region
of ethanol–water solution.
In all cases, K a is positive. Physically, this can be interpreted to mean, in terms of Eq. Ž 7. , that the
chemical potential of ethanol in the process ŽI. increases faster than water as the concentration of salt
increases. The K b values are negative, indicating that the first term on the right of Eq. Ž8. is
out-weighed by the last term. Physically, it means that the chemical potential of water in the process
ŽII. decreases more slowly than ethanol as the concentration of salt increases.

4. List of symbols

D local dielectrical constant

e charge of electron
g radial distribution function
Gh term of the hard sphere interaction in Eq. Ž5.
Gs term of the soft sphere interaction in Eq. Ž5.
H Heaviside function
D Hv partial molar vaporation heat
K salt effect parameter
k Boltzmann constant
ks salting coefficient
L Avogadro constant
R Gas constant
r intermolecular distance
T temperature
 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40 39

x mole fraction of a component in liquid phase, calculated on a salt-free basis

Z) total number of electrons
a polarizability of molecules
a 12 relative volatility of component 1 to component 2
d Dirac function
´ energy parameter
h solvation coefficient used in this work
Žh . hydration coefficient taken from literatures
m dipole moment
mr residual chemical potential
r number density
s diameter of molecules
si j collision diameter of i–j pair of molecules
w fugacity coefficient
f partial molar volume of electrolytes at infinite dilution
Superscripts
s with salt present
Subscripts
1 ethanol
2 water
3 univalent cation
4 univalent anion
e 1–1 type electrolytic molecule
a contribution of the hard sphere interaction
b contribution of the soft sphere interaction

References

w1x J.O’M. Bockris, J. Bowler-Reed, J.A. Kitchener, Trans. Faraday Soc. 47 Ž1951. 184–192.
w2x J.A. Burns, W.F. Furter, Adv. Chem. Ser. 155 Ž1976. 99.
w3x J.A. Burns, W.F. Furter, Adv. Chem. Ser. 177 Ž1979. 11.
w4x B.E. Conway, Electrochemical Data, American Elsevier Publishing, New York, NY, 1952.
w5x B.E. Conway, J.E. Desnoyers, A.C. Smith, Phil. Trans. Roy. London A 256 Ž1964. 389–437.
w6x R.C. Evans, An Introduction to Crystal Chemistry, 2nd edn., Cambridge Univ. Press, Cambridge, 1964.
w7x W.F. Furter, PhD Thesis, University of Toronto, Toronto, Ontario, 1958.
w8x Z.Q. Hung, An Introduction to Theory of Electrolyte Solution, 2nd edn., Science Press, Peijing, 1983, p. 68.
w9x ¨
Y. Hu, D. Ludecke, J. Prausnitz, Fluid Phase Equilibria 17 Ž1984. 217–241.
w10x Y. Hu, J.M. Prausnitz, J. Chem. Ind. Eng. China 38 Ž1987. 22–37.
w11x D. Jaques, W.F. Furter, Ind. Eng. Chem. Fund. 13 Ž1974. 238–241.
w12x F. Kawaizumi, R. Zana, J. Phys. Chem. 78 Ž1974. 1099–1105.
w13x W.L. Masterton, T.P. Lee, R.L. Boyington, J. Phys. Chem. 73 Ž1969. 2761–2763.
w14x W.L. Masterton, T.P. Lee, J. Phys. Chem. 74 Ž1970. 1776–1782.
w15x C. Mavroyannis, M.J. Stephen, J. Mol. Phys. 5 Ž1962. 629–638.
w16x W.F. McDevit, F.A. Long, J. Am. Chem. Soc. 74 Ž1952. 1773–1777.
w17x E.J. Millero, J. Phys. Chem. 73 Ž1969. 2417–2420.
40 R.Y. Sun r Fluid Phase Equilibria 157 (1999) 29–40

w18x F.J. Millero, Chem. Rev. 71 Ž1976. 147.

w19x F. Mukerjee, J. Phys. Chem. 65 Ž1961. 740–744.
w20x R.A. Pierotti, J. Phys. Chem. 67 Ž1963. 1840–1845.
w21x R.A. Pierotti, J. Phys. Chem. 69 Ž1965. 281–288.
w22x R.Y. Sun, J. Chem. Eng. China 47 Ž4. Ž1996. 401.
w23x Y.N. Xu, Y. Hu, G.J. Liu, J. Chem. Ind. Eng. China 38 Ž2. Ž1987. 137.