J Solution Chem (2012) 41:1282–1289

DOI 10.1007/s10953-012-9873-9

Estimation of the Activity Coefficients in Liquid Mixtures

by Using Volumetric Data

M.S. Páez · A.A. Portacio · L.C. Sánchez

Received: 29 June 2011 / Accepted: 20 October 2011 / Published online: 9 August 2012
© Springer Science+Business Media, LLC 2012

Abstract We present an equation that relates the partial molar volume of binary liquid
mixtures with the natural logarithms of the activity coefficients of solute and solvent. This
equation, in combination with one of the activity coefficient models such as those of Mar-
gules, Wilson, Van Laar or NRTL, can be used to estimate the activity coefficients of binary
liquid mixtures, knowing only the densities of the mixtures over the full range of concentra-
tion. In addition, we show a comparison of the estimated activities and activity coefficients
at infinite dilution with experimental values for aqueous solutions of 1,2-butanediol, 2,3-
butanediol, 1,3-butanediol, and 1,4-butanediol at 298.15 K. This method for the estimation
of activity coefficients can be applied to aqueous binary mixtures, because the equation pre-
sented is deduced from physicochemical principles.

Keywords Activity coefficients · Molar volume · Butanediols · Binary mixtures · Density

1 Introduction

Knowledge of activity coefficients is important for the following reasons. (i) They can be
used to correlate and predict vapor–liquid equilibrium (VLE) which is usually done using
activity coefficients for the liquid phase [1]. (ii) Infinite dilution activity coefficients charac-
terize the behavior of the solute molecule when it is completely surrounded by the solvent. It
is extremely valuable for both theoretical and practical purposes. For theoreticians, limiting
activity coefficients provide incisive information regarding solute–solvent interactions in the
absence of solute–solute interactions, which consequently are a measurement of the devia-
tions from ideal solution behavior. From a practical point of view, infinite dilution activity
coefficient data allow for the prediction of retention and selectivity for chromatographic

M.S. Páez ()

Departamento de Química, Universidad de Córdoba, Montería, Colombia
e-mail: [email protected]

A.A. Portacio · L.C. Sánchez

Departamento de Física, Universidad de Córdoba, Montería, Colombia
J Solution Chem (2012) 41:1282–1289 1283

systems and provide essential data in the design of separation equipment, while in environ-
mental sciences these infinite dilution activity coefficients allow the study of the fate and
distribution of organic chemicals in the environment [2]. (iii) The activity coefficient γ ∞ at
infinite dilution of a solute in an ionic liquid is important and can generally be used in the
preliminary selection of solvents for extraction or extractive distillation. Values of γ ∞ are
also useful for the reliable design of thermal separation processes when the last traces of im-
purities need to be removed [3]. (iv) Among the thermodynamic characteristics of solutions
that need be known in the development of processes for substance separation and impurity
extraction are the limiting activity coefficients of solutes. Values of these coefficients for
various substances in the same solvent furnish information about the solvent’s selectivity
toward this series of substances and enables evaluation of the extraction properties of the
solvent [4]. On the other hand, chemical activity has been used to understand, predict and
describe physico-chemical phenomena such as osmosis, mineral stability and cholesterol
biosynthesis [5]. Also, they represent the transition from pure thermodynamics to the the-
ory of intermolecular forces, and hence are needed in understanding the behavior of real
fluids [6]. Finally, values of activity coefficients are related to parameters of widely used
correlation equations for the description and calculation of vapor–liquid equilibria in binary
and multicomponent systems, which makes it possible to calculate and predict vapor–liquid
equilibrium at any concentration of the components, quantify the deviations from ideality
of a component in solution, and thus their practical and theoretical importance in many dif-
ferent areas is evident [4]. Several experimental methods based on properties of mixtures
are used to determine activity coefficients from vapor pressures, electromotive force (emf),
solubility, freezing point depression, surface properties, gas stripping, indirect headspace
chromatography and thermal measurements. Few theoretical methods have been proposed
and their applications are very limited in the case of aqueous solutions of nonelectrolytes
and cannot be extended to concentrated solutions [6–10].
There are several techniques for the experimental determination of infinite dilution ac-
tivity coefficients, such as the chromatographic method (gas–liquid chromatography (GLC),
headspace GLC, liquid–liquid chromatography (LLC)), differential ebulliometry method,
dew point method and the gas stripping method [1]. In many cases experimental determi-
nations are not easy and different empirical or semi-empirical predictions and correlative
models have been used to predict the relationship between activity coefficients and mole
fraction and, in particular, activity coefficients at infinite dilution. Many equations with two,
three or more parameters have been proposed such as those of Margules, Wilson, van Laar,
NRTL, and group contribution methods such as UNIFAC. Most of the methods used to
correlate thermodynamic properties are based on empirical adjustments and semi-empirical
methods that require many adjustable parameters. Even though in most cases the parameters
have been formally assigned a physical meaning, and are related to properties of the pure
compounds, the results show that these constants should be considered as empirical param-
eters [10]. For nonaqueous polar mixtures many models give satisfactory results, however
this is not the case for strongly nonideal mixtures such as aqueous solutions where the
large nonidealities are responsible for quite large variations in the infinite dilution activity
coefficients. We are concerned with the development of a method to determine activity coef-
ficients based on experimental thermodynamic properties of mixtures, which can be applied
to aqueous solutions. An examination of the available literature shows that equations used
to determine activity coefficients do not employ solution molar volumes and partial molar
volumes of the components as a function of concentration, even though Ben Naim’s work
suggests that this is possible by using volumetric data of good quality [11, 12].
1284 J Solution Chem (2012) 41:1282–1289

2 Experimental Section

Activity coefficients were obtained for butanediols in water at 298.15 K using the algorithm
to be described in the next section and volumetric data [13–17] from previous works. The
water activities were measured by using an AquaLab water activity meter (Model 3TE,
Decagon Devices, USA). This instrument applies the chilled mirror method to determine
the dewpoint temperature of air equilibrated with the sample.

3 Results and Discussion

The main approach assumed in this work is based on the relations between the derivative
of the chemical potential μi of the component i with respect to two different concentration
scales, molarity ci , and mole fraction xi , at constant temperature and pressure. The relation-
ship can be expressed by the following equations:
     
∂μ1 ∂μ1 ∂c1
= (1)
∂x1 T ,P ∂c1 T ,P ∂x1 T ,P
     
∂μ2 ∂μ2 ∂c2
= (2)
∂x2 T ,P ∂c2 T ,P ∂x2 T ,P
The partial derivative (∂ci /∂xi )T ,P is evaluated from the relation between the molar concen-
tration c (mol·dm−3 ), total volume V (cm3 ), and the molar volume Vm (cm3 ·mol−1 ), which
is given by:
ni xi
ci = = (3)
V Vm
From these expressions can be obtained, for a mixture, the partial derivative of the mo-
lar concentration of component i (i = 1, 2), ci , with respect to its molar fraction, xi , thus
obtaining:
   
∂ci 1 xi ∂Vm V̄j
= − = (4)
∂xi T ,P Vm Vm2 ∂xi T ,P Vm2

where V̄j is partial molar volume of component j, i = j , which is related to the solution
molar volume Vm by Eq. 5:
 
∂Vm
V̄j = Vm − xi (5)
∂xi T ,P
By inserting Eq. 4 into Eqs. 1 and 2 we have:
   
∂μ1 ∂μ1 V̄2
= (6)
∂x1 T ,P ∂c1 T ,P Vm2
and
   
∂μ2 ∂μ2 V̄1
= 2
(7)
∂x2 T ,P ∂c2 T ,P Vm

When the chemical potentials are written in terms of molar concentration, both Eqs. 8 and 9
satisfy the Gibbs–Duhem equation:
J Solution Chem (2012) 41:1282–1289 1285
   
∂μ1 ∂μ2
c1 + c2 =0 (8)
∂c2 T ,P ∂c2 T ,P
   
∂μ1 ∂μ2
x1 + x2 =0 (9)
∂x2 T ,P ∂x2 T ,P
Moreover, according to Guerasimov et al. [18], the variation of chemical potential of one
mole of component i in going from a standard or reference state, denoted by index ∗ , to a
different state without index, does not depend on the choices of the concentration unit of the
component in solution or the corresponding activities. Thus:
 x  c
∗ ai ai
μi − μi = RT ln ∗x = RT ln ∗c (10)
ai ai
where μ∗i is the chemical potential of the infinitely dilute solution, aix and aic denote the
activities in the “rational” (mole fraction) and molar concentration systems, respectively. If
we choose the reference state so that ai∗x = xi∗ and ai∗c = ci∗ for both components taken as
the solute concentration at opposite ends of the composition scale (infinitely dilute solu-
tion), then, in the concentration limit of both components and different scales, the following
conditions must be met:
 ∗
a
lim γi = γi∗∞ = lim i∗ = 1 (11)
xi →0 xi →0 x
i
Equation 11 indicates that so far we are using an asymmetric normalization criterion.
To relate the standard symmetric normalization of activity coefficients with the standard
asymmetric ones, according to Prausnitz et al. [10]:
γ2
γ2∗∞ =  (12)
γ2
where γ2∗∞ is the activity coefficient at infinite dilution in the asymmetric reference sys-
tem, γ2 is the activity coefficient in the asymmetric reference system, and γ2 is the activity
coefficient in the symmetric system. From Eq. 10 the following expression can be obtained:
  ∗ 
γix ci xi
= (13)
γic xi ci∗
From to the definition of molar concentration is possible to obtain for the reference state
of a binary solution, with respect to component 1 and 2 at opposite composition ends, the
following expressions:
x1∗ M2
= ◦ (14)
c1∗ d2
x2∗ M1
= ◦ (15)
c2∗ d1
where di◦ and Mi represent the density and molecular weight, respectively, of the pure com-
ponents.
From Eqs. 13–15 the relations between activity coefficients γi on the two concentration
scales are given by:
Vm x
γ1c = γ (16)
V2◦ 1
Vm
γ2c = ◦ γ2x (17)
V1
Equations equivalent to Eqs. 16 and 17 are found in Prausnitz et al. [10–12, 18].
1286 J Solution Chem (2012) 41:1282–1289

Table 1 Activity coefficients at infinite dilution for this work at 298.15 K

System 12BD–water 13BD–water 14BD–water 23BD–water

Margules A 0.89212 0.92626 0.90548 0.92518

B −0.78867 −1.05636 −0.72181 −1.21627
σ 0.0012 0.0011 0.0018 0.0009
γ1∗∞ 1.1090 0.8780 1.2016 0.7474
γ2∗∞ 1.6451 1.4890 1.7239 1.3731

Van Laar A 0.398044 0.336616 0.426430 0.301012

B 0.597611 0.560792 0.623964 0.529756
σ 0.0011 0.0009 0.0016 0.0010
γ1∗∞ 1.4889 1.4002 1.5318 1.3512
γ2∗∞ 1.8178 1.7521 1.8663 1.6985

Wilson A 1.066259 1.194153 1.027703 1.266642

B 0.508404 0.465452 0.514551 0.447364
σ 0.0011 0.0009 0.0015 0.0010
γ1∗∞ 1.5333 1.4292 1.5811 1.3720
γ2∗∞ 1.8404 1.7693 1.8903 1.7121

NRTL A 0.46016 0.44719 0.29166 0.01260

B 0.11352 0.09682 0.21755 0.29338
C 1.27834 1.22605 1.23699 1.23699
D 2.20024 2.45070 1.77307 1.77307
σ 0.0009 0.0010 0.0013 0.0010
γ1∗∞ 2.0173 1.9062 1.7831 1.3620
γ2∗∞ 2.0338 1.9827 1.9687 1.7037

According to this construction, the activity coefficients involved in Eqs. 16 and 17 are
expressed in the asymmetric system of reference. Finally, from the use of the definition of
chemical potential and Eqs. 16 and 17, we find exact expressions for the partial derivatives
(∂μi /∂xi )T ,P and (∂μi /∂ci )T ,P , given by:
     
∂μi 1 ∂ ln γix
= RT − (18)
∂xi T ,P xi ∂xj T ,P

and
       
∂μi RT ∂ ln(Vm /Vj◦ ) ∂xi ∂ ln γiz
= + RT − RT (19)
∂ci T ,P ci ∂ci T ,P ∂ci T ,P ∂xj T ,P

Equation 7 can be rewritten as:

 2
V̄1 ∂μ
∂c T ,P
Vm2 =  ∂μ 2 (20)
2
∂x2 T ,P

Finally, by using Eqs. 12 and 18–20, the following equation is obtained:

 ∂ ln(Vm /V1◦ )   ∂x2   ∂ ln γ2 
V̄1 ∂c T ,P
− V̄1 ∂c T ,P ∂x1 T ,P V̄1
Vm2 = 2
 1  ∂ ln γ2 
2
+  ∂ ln γ2  (21)
x2
− ∂x1 T ,P 1
x2
− ∂x1 T ,P c2
J Solution Chem (2012) 41:1282–1289 1287

Table 2 Activity coefficients at infinite dilution obtained from literature data [19]

System T /K γ2∗∞ Method

12BD–water 299.15 2.0 ± 0.2 Dewpoint

13BD–water 299.05 2.2 ± 0.2 Dewpoint
14BD–water 299.35 2.8 ± 0.2 Dewpoint
23BD–water 298.95 1.6 ± 0.2 Dewpoint

Fig. 1 Water activities, measured using an AquaLab water activity meter (Model 3TE, Decagon Devices,
USA) [20], are compared with those obtained in this work. The insets show data in the expanded diluted
region

Equation 21 is expressed in the symmetric system of reference. According to this expression,

the molar volume of the mixture can be expressed in terms of the partial molar volume of
component 1, and of certain partial derivatives of the molar volume and activity coefficient
γi in the “rational” system. This expression suggests that from density data of quality, it is
thermodynamically possible to evaluate the activity coefficients on the molar fraction scale.
Equation 21 represents a new approach for estimating activity coefficients in binary so-
lutions for the entire concentration range. It can be used in connection with appropriate
expressions to represent the activity coefficients, such as those of Margules, Van Laar, Wil-
son, and the NRTL equation. The key to its implementation lies in using the left side of
Eq. 21 as a convergence parameter.
1288 J Solution Chem (2012) 41:1282–1289

Table 3 Water activities of the investigated diol/water mixtures calculated in this work at 298.15 K

1,2-Butanediol 1,3-Butanediol 1,4-Butanediol 2,3-Butanediol

x(H2 O) a(H2 O) x(H2 O) a(H2 O) x(H2 O) a(H2 O) x(H2 O) a(H2 O)

0.9954 0.9954 0.9958 0.9958 0.9979 0.9979 0.9977 0.9977

0.9943 0.9943 0.9946 0.9946 0.9957 0.9957 0.9972 0.9972
0.9921 0.9922 0.9927 0.9927 0.9945 0.9945 0.9949 0.9949
0.9916 0.9917 0.9920 0.9920 0.9920 0.9921 0.9909 0.9910
0.9908 0.9909 0.9911 0.9912 0.9907 0.9908 0.9894 0.9895
0.9895 0.9896 0.9896 0.9897 0.9887 0.9888 0.9888 0.9889
0.9871 0.9872 0.9869 0.9871 0.9870 0.9872 0.9874 0.9875
0.9828 0.9831 0.9829 0.9832 0.9849 0.9851 0.9864 0.9866
0.9737 0.9743 0.9740 0.9746 0.9828 0.9831 0.9851 0.9853
0.9698 0.9706 0.9697 0.9705 0.9809 0.9812 0.9784 0.9788
0.9591 0.9605 0.9580 0.9595 0.9720 0.9727 0.9727 0.9733
0.9546 0.9563 0.9527 0.9546 0.9635 0.9646 0.9652 0.9662
0.9438 0.9463 0.9438 0.9464 0.9494 0.9515 0.9446 0.9471
0.9337 0.9372 0.9267 0.9309 0.9224 0.9271 0.9438 0.9464
0.9249 0.9293 0.9153 0.9208 0.8931 0.9016 0.9370 0.9402
0.9143 0.9199 0.8653 0.8778 0.8918 0.9005 0.9288 0.9328
0.8653 0.8778 0.7847 0.8118 0.8653 0.8781 0.8621 0.8747
0.7847 0.8122 0.6981 0.7434 0.7847 0.8128 0.8204 0.8397
0.6981 0.7447 0.6085 0.6733 0.6981 0.7451 0.7821 0.8080
0.6085 0.6758 0.5196 0.6024 0.6085 0.6752 0.7583 0.7884
0.5196 0.6065 0.4349 0.5318 0.5196 0.6036 0.7304 0.7653
0.4349 0.5375 0.3572 0.4625 0.4349 0.5314 0.6973 0.7379
0.3572 0.4695 0.2885 0.3957 0.3572 0.4600 0.6574 0.7044
0.2885 0.4037 0.2297 0.3329 0.2885 0.3910 0.6087 0.6627
0.2297 0.3414 0.1805 0.2751 0.2297 0.3266 0.5476 0.6089
0.1805 0.2836 0.1405 0.2238 0.1805 0.2679 0.4692 0.5366
0.1405 0.2318 0.1084 0.1793 0.1405 0.2164 0.3648 0.4336
0.1084 0.1865 0.0756 0.1302 0.1084 0.1723 0.2195 0.2755
0.0756 0.1361 0.0756 0.1242 0.0999 0.1311

If in Eq. 21 the activity coefficients are represented by an expression of concentration

using one the selected models, then the optimum parameters are obtained by minimizing the
objective function OF defined as:
 2 2
OF = Vm,exp − Vm,calc
2
(22)

where Vm,exp is the molar volume obtained from experimental density data and Vm,calc is the
molar volume obtained from Eq. 21.
The activity coefficients at infinite dilution obtained in this work are shown in Table 1,
while those reported in literature [19] are shown in Table 2. Figure 1 shows the measured
water activities [20] compared with those obtained in this work. Additionally, these data are
listed in Table 3.
J Solution Chem (2012) 41:1282–1289 1289

4 Conclusions

The theoretical basis of a new method for the estimation of the activity coefficients of binary
liquid mixtures has been discussed. Good agreement is observed among the results estimated
by the new method and experimental results. From this analysis, it is concluded that this
method can be applied to a wide variety of aqueous mixtures.

Acknowledgements This work was supported by the Universidad de Córdoba, Montería, Colombia.

References

1. Murotomi, A., Yokota, M., Miyano, Y., Sakane, T., Fukuchi, K.: Measurement and prediction of infinite
dilution activity coefficients for alkane in alkane within wide temperature. J. Therm. Anal. Calorim. 95,
393–396 (2009)
2. Estrada, E., Díaz, G.A., Delgado, E.J.: Predicting infinite dilution activity coefficients of organic com-
pounds in water by quantum-connectivity descriptors. J. Comput.-Aided Mol. Des. 20, 539–548 (2006)
3. Zhu, J., Yu, Y., Chen, J., Fei, W.: Measurement of activity coefficients at infinite dilution for hydrocar-
bons in imidazolium-based ionic liquids and QSPR model. Front. Chem. Eng. China 1, 190–194 (2007)
4. Maiorov, E.A., Dobryakov, Yu.G.: Limiting activity coefficients of hydrocarbons in phenol. Russ. J.
Appl. Chem. 80, 2040–2044 (2007)
5. Reichenberg, F., Smedes, F., Jönsson, J., Mayer, P.: Determining the chemical activity of hydrophobic
organic compounds in soil using polymer coated vials. Chem. Cent. J. 2, 1–10 (2008)
6. Gracia-Fadrique, J.: Activity coefficients at infinite dilution from surface tension data. Langmuir 18,
3604–3608 (2002)
7. Badarayani, R., Kumar, A.: A simple model for estimation of activity coefficients of salts in aqueous and
nonaqueous solutions and their mixtures up to high temperatures. Ind. Eng. Chem. Res. 40, 1996–2003
(2001)
8. Dobrjakov, Y.G., Balashova, I.M., Maurer, G.: Experimental results for the limiting activity coefficients
in some binary and ternary mixtures of organic components. J. Chem. Eng. Data 45, 185–193 (2000)
9. Alessi, P., Fermeglla, M., Kikic, L.: Significance of dilute regions. Fluid Phase Equilib. 70, 239–250
(1991)
10. Prausnitz, J.M., Lichtenthaler, R., Gomes, E.: Molecular Thermodynamics of Fluid-Phase Equilibria.
Prentice-Hall, Englewood Cliffs (1969), Chap. 6
11. Ben-Naim, A.: Water and Aqueous Solutions. Plenum Press, New York (1974)
12. Ben-Naim, A.: Molecular Theory of Solutions. Oxford University Press, Oxford (2006)
13. Romero, C., Páez, M.: Volumetric properties of aqueous binary mixtures of 1-butanol, butanediols, 1,2,4-
butanetriol and butanetetrol at 298.15 K. J. Solution Chem. 36, 237–245 (2007)
14. Checoni, R.F.: Experimental study of the excess molar volume of ternary mixtures containing {water
+ (1,2-propanediol, or 1,3-propanediol, or 1,2-butanediol, or 1,3-butanediol, or 1,4-butanediol, or 2,3-
butanediol) + electrolytes} at a temperature of 298.15 K and atmospheric pressure. J. Chem. Thermodyn.
42, 612–620 (2010)
15. Geyer, H., Ulbig, P., Gornert, M.: Measurement of densities and excess molar volumes for (1,2-
ethanediol, or 1, 2-propanediol, or 1,2-butanediol + water) at the temperatures (278.15, 288.15, 298.15,
308.15, and 318.15) K and for (2,3-butanediol + water) at the temperatures (308.15, 313.15, and
318.15) K. J. Chem. Thermodyn. 32, 1585–1596 (2000).
16. Piekarski, H., Pietrzak, A.: Molecular interactions of alkanediols in methanol and in water: density and
heat capacity measurements. J. Mol. Liq. 121, 46–52 (2005)
17. Hawrylak, B., Grade, K., Palepu, R.: Thermodynamic properties of binary mixtures of butanediols with
water. J. Solution Chem. 27, 17–31 (1998)
18. Guerasimov, Ya., Dreving, V., Eriomin, E., Kiseliov, A., Lebedev, V., Panchenkov, G., Shliguin, A.:
Course in Physical Chemistry. Mir, Moscow (1987)
19. Kojima, K., Zhang, S., Hiaki, T.: Measuring methods of infinite dilution activity coefficients and a
database for systems including water. Fluid Phase Equilib. 131, 145–179 (1997)
20. Marcolli, C.P., Peter, Th.: Water activity in polyol/water systems: new UNIFAC parameterization. At-
mos. Chem. Phys. 5, 1545–1555 (2005)