Fluid Phase Equilibria, 63 (1991) 141-150 141

Elsevier Science Publishers B.V., Amsterdam

MEASUREMENT OF LIMITING ACTIVITY COEFFICIENTS

FOR AQUEOUS SYSTEMS BY DIFFERENTIAL EBULLIOMETRY

DIANE L. BERGMANN
Department of Chemical Engineering, University of Illinois, Urbana, IL 61801 (U.S.A.)

CHARLES A. ECKERT *
School of Chemical Engineering, Georgia Tech., Atlanta, tiA 30332-0100 (U.S.A.)
(Received July 2. 1990; accepted in final form September 22, 1990)

ABSTRACT

Bergmann, D.L. and Eckert, CA., 1991. Measurement of limiting activity coefficients for
aqueous systems by differential ebulliometry. Fluid Phase Equilibria, 63: 141-150.

New data have been determined by differential ebulliometry for limiting activity coeffi-
cients of binary systems of water. Results are reported with acetone, N, N-dimethylfor-
mamide, ethyl acetate, isopropanol, methanol, nitromethane and tetrahydrofuran over a
modest range of temperatures. The results are used to demonstrate the limitations of several
current excess Gibbs energy expressions in representing both vapour-liquid equilibrium and
liquid-liquid equilibrium data from limiting activity coefficient measurements.

INTRODUCTION

Activity coefficients at infinite dilution (y”) provide a way to study the

forces between unlike molecules directly, thus giving insight into the reasons
behind solution behavior. These ym values can also be used to predict
vapor-liquid equilibrium (VLE) from any two-parameter excess Gibbs
energy expression. It has been shown by various authors (Lobien and
Prausnitz, 1982; Zou and Prausnitz, 1987; Lafyatis et al., 1989) that the
current excess Gibbs energy expressions, however, are inadequate in portray-
ing VLE and liquid-liquid equilibrium (LLE) simultaneously. More infor-
mation on such nonideal systems as water systems will help in the search for
more accurate expressions.
Although water is perhaps the most important industrial solvent, it is
complex and exhibits great nonideality with organics. In this study the

* Author for correspondence.

037%3812/91/$03.50 0 1991 ~ Elsevier Science Publishers B.V.

142

infinite dilution activity coefficients for aqueous systems of acetone, N,N-

dimethylformamide, ethyl acetate, isopropanol, methanol, nitromethane and
tetrahydrofuran were measured at various temperatures. The systems were
chosen because they are commonly used solvents and their volatilities
relative to water are small enough to be measured accurately with this
technique. The yoo data were used to predict VLE and LLE for some of
these systems.
The activity coefficients at infinite dilution for the systems in this study
were measured using the technique first proposed by Gautreaux and Coates
(1955). A differential apparatus was used to measure the change in boiling
point of a solvent with the addition of a small amount of solute at constant
pressure (U/ax)?. From this measurement ym was calculated using the
expression:

The use of this equation, including the determination of the fugacity

coefficients, saturation pressures, and liquid volumes, is discussed by Trampe
and Eckert (1990a).

EXPERIMENTAL

The activity coefficients at infinite dilution were measured using a modifi-

cation of the recent ebulliometer design by Scott (1986) which uses magnetic
stirring. The experimental apparatus and procedure are described in detail
by Trampe and Eckert (1990a). Working with water at dilute concentrations
presented some experimental challenges. Because the molecular weight of
water is small compared with the other chemicals studied, a solution of
water and the solvent had to be injected in order to obtain dilute concentra-
tions. The mole fraction regions studied were generally below 0.003, since at
large concentrations curvature was observed in the plot of temperature
change versus solute composition, exacerbating the uncertainties in
(ar/axg. B e f ore each run the test flask was either baked overnight or was
heated, and then flushed with dry nitrogen for approximately ten minutes.
More details are presented by Bergmann (1990).

MATERIALS

Solvents used in this experiment should have a purity of > 99.9%. All the
solvents except nitromethane were used as purchased. Nitromethane was
 143

distilled under a vacuum of 200 mm Hg with the middle 50% collected. All
solvents are reported to contain less than 0.05% water and were stored over
3A molecular sieves. Refractive index values are those at 20°C measured
with an Abbe model refractometer with stated accuracy of f O.OOOln,.

Acetone; Mallinckrodt Nanograde

N, N-Dimethylformamide; Aldrich HPLC grade, > 99.9%
Ethyl Acetate; Mallinckrodt Nanograde n, = 1.3723
Isopropanol; Mallinckrodt Nanograde, n n = 1.3776
Methanol; Mallinckrodt Nanograde
Nitromethane; Aldrich HPLC grade, 96%, distilled at 200 mm Hg with
middle 50% collected
Tetrahydrofuran; Anhydrous Gold Label, 99.9%
Water; Mallinckrodt ChromAR HPLC or Aldrich HPLC grade

RESULTS AND DISCUSSION

The activity coefficients at infinite dilution measured for the seven binary
systems over temperature ranges varying from 35 o C to the solvent normal
boiling temperature are listed in Table 1. The errors in the data were
estimated by duplicate runs when possible or by maximum and minimum
values in (i3T/ax)p due to fluctuations in the measured temperature. The
data were compared to literature values where available as shown in Figs. 1
and 2. Slocum and Dodge (1964) measured the pressure difference between
a pure solvent and dilute mixture held in a constant temperature bath. The
mole fraction range for the water dilute range was between 0.01 and 0.05
and showed curvature. They made no vapor phase correction and used
laboratory distilled water in their experiments. Tochigi et al. (1977) used
differential ebulliometry and also did not make vapor phase corrections.
The magnitude of ym reflects the behavior of the solution. There are large
values of ym for the systems ethyl acetate-water and nitromethane-water,
which are partially immiscible in the temperature regions studied. In con-
trast, the ym values are much closer to ideal for the methanol-water system.
The ycc data for the N, N-dimethylformamide-water system show deviations
going from negative to positive as the temperature is raised. This behavior
suggests solvation at lower temperatures, with water association becoming
dominant as the temperature is raised.
A plot of In yM vs. l/T for the systems where water is a solvent is shown
in Fig. 3. Although these systems are not regular, the plots appear linear
within experimental error. One could use the slope of these plots to yield the
partial molar excess enthalpy at infinite dilution, h?. However, the temper-
ature regions studied here are small and it is preferable to measure this
144

TABLE 1
Experimental activity coefficients at infinite dilution

Solvent Solute Temp. “C ym

Acetone Water 34.7 6.02kO.15
44.9 5.68 f 0.1
55.3 5.3O-tO.02

N, N-Dimethylformamide Water 80.5 0.96kO.12

89.8 0.99 + 0.05
94.5 1.02kO.03
109.8 1.33kO.07
124.6 1.22kO.15

Ethyl-Acetate Water 36.3 12.2 t-o.1

45.4 10.8 20.9
55.3 10.1 -to.1
65.2 9.7 +0.2
76.3 8.98 t 0.05

Isopropanol Water 45.2 3.5 kO.4

55.0 3.39 + 0.07
64.8 3.0 +0.16

Methanol Water 34.6 1.6 +0.2

44.7 1.8 +0.2
54.7 1.7 kO.25
63.9 1.76 + 0.05

Nitromethane Water 41.1 13.3 kO.2

49.3 11.8 +0.5
59.9 11.8 +0.8
70.2 8.6 kO.4
80.5 8.7 kO.8

Tetrahydrofuran Water 35.2 11.1 kO.6

44.5 10.4 ,O.l
54.9 9.8 +0.3
64.9 9.4 +0.1

Water N, IV-Dimethylformamide 50.0 0.89 f 0.4

59.9 1.35 + 0.26
69.9 2.67 + 0.15

Water Isopropanol 44.7 8.8 -to.5

54.9 9.6 +0.5
64.4 9.5 kO.4
76.0 11.0 +0.8
84.6 11.6 kO.5
Water Methanol 44.7 1.46kO.16
55.3 1.49 f 0.05
64.5 1.59 + 0.05
75.1 1.52kO.12
84.6 1.78kO.14
 145

3.0

2.6 -

2.2 -

1.8 -

6+
c 1.4 -

1.0 -
THIS WORK, WATER DILUTE

: $&$?R~~s~~~~?k!j%TER DILUTE
0.6 - ;
SLOCuM ANo ““““:&&dOL DILUTE
(BEST FIT sTRAK;HT LINES)
0.2 I
0.0026 0.0029 0.f 33
K-’ l/T
Fig. 1. Comparison of experimental and literature ym data for the water-isopropanol system.

quantity directly. Trampe and Eckert (1990b) have used the technique of
asymmetric isothermal flow calorimetry for this purpose.
The yoo values measured for the ethyl acetate, isopropanol, methanol and
nitromethane systems were used to find the parameters of various two-
parameter excess Gibbs energy expressions ( gE(x)). Since yM was measured
for only the water-dilute end of the ethyl acetate and nitromethane systems,
a Van Laar and Margules fit of dilute VLE data was used to obtain yoo for
the other end respectively. Also ym for water in ethyl acetate at 50 o C was
interpolated from a linear fit of the ym data. The parameters are given in

2.6

2.6 -

52.2

*+ 241 /

2.0 -
l TIIIS WORK
0 TCCHIGI ET AL., (1977)
- LINEAR BEST FIT

1.6
134
0.0026 0.0026 1 /T0.0030 K-’ om32 ‘A

Fig. 2. Comparison of experimental and literature ym data for water dilute in ethyl acetate.
 2.0

WATER AS SOLVENT
1.5
I --vnO

o ISOPROPANOL
q N.N-OIMETHYLFORM4MlOE
A Mm+ANOL
- LINEAR EST FIT

-1.0 I 1
0.0026 0.0028 p30 0.0032
l/T K-

Fig. 3. The natural log of ym vs. inverse temperature for several systems where water is the
solvent.

TABLE 2
Parameters for gE(x) expressions from ym data and from soluhility data

Parameters from ym
System ij Margules Van Laar NRTL UNIQUAC Wilson
A,, Al, A&, Au,, Ah,,
(cal mol - ’ ) (cal mol-t) (cal mol-‘)
a = 0.3
Isopropanol(l)/ 12 2.266 2.266 30.84 369.71 309.36
Water(2) 55 o C 21 1.189 1.189 1445.56 10.97 1199.34
Methanol(l)/ 12 0.378 0.378 1135.88 - 599.54
Water(2) 45 o C 21 0.569 0.569 - 631.17 964.85
Ethyl acetate(l)/ 12 4.202 4.202 732.46 840.98
Water(2) 50 o C 21 2.367 2.367 2177.53 153.83
Water(l)/50 oC 12 2.468 2.468 1391.60 317.26
Nitromethane(2) 21 3.063 3.063 858.42 658.68
Parameters from Solubility
System ij Margules Van Laar NRTL
Ai, Al, Aglj
(cal mol-‘)
a = 0.3
Ethyl acetate(l)/ 12 4.306 4.413 606.73
Water(Z) 50 o C 21 1.531 2.047 2409.48
Water(l)/50 oC 12 2.309 2.387 317.26
Nitromethane(2) 21 3.335 1.313 658.68
 147

OXCOME DATA
- WRCULES
+++++ “AN LAM
.____ NFgL

Fig. 4. VLE predicted from ym data for the water-isopropanol system at 55 ’ C. VLE data of
Tunik and Zharov (1980). The Van Laar and NRTL predictions overlap.

Table 2. As shown in Figs. 4-6, these gE(x) expressions do not give a good
representation of relatively nonideal VLE data, as can be seen in Fig. 4
where the Margules equation even predicts a two phase region in the
isopropanol-water system.
The immiscibility regions of the ethyl acetateewater and nitromethanee
water systems were investigated. The solubility at 50 o C was predicted using
gE(x) expressions with parameters obtained from the yoo data. These

Fig. 5. VLE predicted from ym data for the water-ethyl acetate system at 50 o C. VLE data of
Kharin et al. (1968).
148

240 ,

200 -

s?

8
t.lSO-

2
u-l 1201 CCOOOME OATA ,

-z “%iYS
E .____ NWL
- - IJNIOUAC
60-

Fig. 6. VLE predicted from ym data for the water-nitromethane system at 50°C. VLE data
of Schuberth (1964).

calculated solubilities are shown in Table 3. While all the gE(_x) expressions
predict immiscibility regions at this temperature, the compositions of the
two phases are predicted only qualitatively. The difference between predic-
ted and actual solubility is as large as 68% for the ethyl acetate dilute phase.

TABLE 3
Solubility determined from ym data, and ym values determined from solubility data at 50 o C

Solubility from ydo

System Measured Margules Van Laar NRTL UNIQUAC
Solubility
Ethyl acetate(l)/ X, = 0.815 a 0.901 0.819 0.859 0.868
Water(2) x; = 0.013 a 0.018 0.020 0.022 0.022

Water(l)/ xi = 0.143 b 0.120 0.124 0.151 0.143

Nitromethane(2) x; = 0.951 b 0.927 0.935 0.903 0.910
ym from Solubility
System Measured Margules Van Laar NRTL
Yrn
Ethyl acetate(l)/ vr = 66.8 74.2 82.5 86.8
Water(2) yz” = 10.7 4.6 7.8 8.7

Water(l)/ yp” = 11.8 10.07 10.88 12.15

Nitromethane(2) y? = 21.4 27.1 28.1 30.4
a Merriman (1913).
b Riddick et al. (1986).
 149

Conversely, solubility data for the ethyl acetate and nitromethane systems
were also used to determine the parameters of the g”(x) expressions and
thus estimate the values of ym at 50 o C for these systems. The parameters
are given in Table 2 and the calculated yM values are listed in Table 3. All
the gE(x) expressions tested tend to underpredict the water-dilute end and
overpredict the water-concentrated end. The errors in the predictions are
much larger in the water-concentrated end. This may be due to errors in
estimating ym from dilute VLE data. The water-dilute region of the
nitromethane-water system seems to be predicted better by all the expres-
sions than the same region in the ethyl acetate-water system. The NRTL
equation seems to represent best the water-dilute region, the region mea-
sured in this study.

CONCLUSIONS

Activity coefficients at infinite dilution were measured for various aque-

ous-organic systems using the technique of differential ebulliometry. It was
shown that VLE can be predicted accurately from these values. The ym data,
however, demonstrate that the current gE(x) expressions are inadequate in
representing both VLE and LLE behavior. Potentially these new data can be
used in the development of future gE(x) models.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the financial support of the DuPont

Company and the very useful advice and assistance of Mr. Leon Scott.

LIST OF SYMBOLS

activity coefficient at infinite dilution

partial molar enthalpy at infinite dilution
fugacity coefficient of the vapor
fugacity coefficient at saturation pressure
total pressure
saturation pressure
gas constant
temperature
“, liquid molar volume
X, composition

Subscripts

1 solute
2 solvent
150

REFERENCES

Bergmann, D.L., 1990. Measurement of Activity Coefficients at Infinite Dilution for Aqueous
Solutions, M.S. Thesis, University of Illinois, Urbana, IL, Slpp.
Gautreaux, M.F. and Coates, J., 1955. Activity coefficients at infinite dilution. AIChE J., 1:
496.
Kharin, M.S.E., Perelygin, V.M. and Volkov A.G., 1968. Liquid-vapor phase equilibrium in
ethanol-ethyl acetate and water-ethyl acetate systems. Izv. Vyssh. Ucheb. Zaved., Pishch.
Tekhnol., 4: 136.
Lafyatis, D.S., Scott, L.S., Trampe, D.M. and Eckert, C.A., 1989. A test of the functional
dependence of gE(x) in liquid-liquid equilibria using limiting activity coefficients. Ind.
Eng. Chem. Res., 28: 585.
Lobien, G.M. and Prausnitz, J.M., 1982. Infinite dilution activity coefficients from differen-
tial ebulliometry. Ind. Eng. Chem. Fundam., 21: 109.
Merriman, R.W., 1913. The mutual solubilities of ethyl, acetate and water and the densities of
mixtures of ethyl-acetate and ethyl-alcohol. J. Chem. Sot., 103: 1774.
Riddick, J.A., Bunger, W.B. and Sakaro, T.K., 1986. Organic Solvents, Physical Properties
and Methods of Purification, 4th edn., Wiley, p. 1042.
Schuberth, H., 1964. Vorausberechmtng und Messung von Phasengleichgewichtsdaten am
System Nitromethan/Wasser sowie erglnzende Bemerkungen zur Verwendung des zwei-
seitigen Porterschen Ansatzes auf Systeme mit relativ geringer gegenseitiger Liislichkeit. Z.
Phys. Chem., 225: 305.
Scott, L.S., 1986. Determination of activity coefficients by accurate measurement of boiling
point diagram. Fluid Ph. Equil., 26: 149.
Slocum, E.W. and Dodge, B.F., 1964. Activity coefficients at infinite dilution: I-propanol-
water system. AIChE J., 10: 364.
Tochigi, K., Minami, S. and Kojima, K., 1977. Prediction of vapor-liquid equilibria with
chemical reaction by analytical solutions of groups. J. Chem. Eng. Jpn., 10: 349.
Trampe, D.M. and Eckert, C.A., 1990a. Limiting activity coefficients from an improved
differential boiling point technique. J. Chem. Eng. Data, 35: 156.
Trampe, D.M. and Eckert, C.A., 1990b. Calorimetric measurement of partial molar excess
enthalpy at infinite dilution. J. Chem. Eng. Data, submitted for publication.
Tunik, E.E. and Zharov, V.T., 1980. Phase and chemical equilibria in the formic acid-water-
isopropyl formate-isopropanol system. VINITI, 1242.
Zou. M. and Prausnitz, J.M. 1987. Vapor-liquid and liquid-liquid equilibria in binary
aqueous systems. J. Chem. Eng. Data, 32: 34.