A Specialist Periodical Report

Chemical Thermodynamics
Volume 2

A Review of the Recent Literature

Senior Reporter
M. L. McGlashan, Deparfmenf of Chemistry, Universify College
London

Reporters
mP
C m Hicks, Universify of Exefer
C. M. Knobler, Universify of California, Los Angeles, U.S.A.
T. M. Letcher, Universify of Witwafersrand, Johannesburg,
South Africa
1. A. McLure, University of Shefield
K. N. Marsh; Universify of New England, New South Wales, Ausfralia
G. M. Schneider, Ruhr-Universitaf, Bochum-Querenberg, Germany
R
m L. Scott, Universify of California, Los Angeles, U.S.A.
F. L. Swinton, New Universify of Ulsfer, Coleraine, Norfhern Ireland
A. G. Williamson, Universify of Canferbury, Chrisfchurch,
New Zealand
C. L. Young, Universify of Melbourne, Victoria, Australia

The Chemical Society

Burlington House, London W I V OBN
 ISBN: 085186
ISSN :0305-9731

Library of Congress Catalog Card No. 7243456

Copyright @ 1978
The Chemical Society
AN Rights Reserved
No part of this book may be reproduced or transmitted
in any form or by any means - graphic, electronic,
including photocopying, recording, taping or
information storage and retrieval systems - without
written permission from The Chemical Society

Organic formulae composed by Wright's Symbolset method

PRINTED W GREAT BRITAIN BY JOHN WRIGHT A N D SONS LTD., AT THE STONEBRIDGE PRESS, BRISTOL BS4 SNU
Fore word

The first volume of the Specialist Periodical Reports on Chemical Thermo-

dynamics dealt entirely with the range of measurements needed for the compila-
tion of tables of standard thermodynamic functions of pure substances. All
chapters after the first were written by members of the National Physical Labora-
tory, which was making a tremendous contribution not only to additions to
those tables and to refinement of values already in them but also to the develop-
ment of experimental methods for their measurement. Tragically that is no longer
so; the United Kingdom is no longer making any concerted effort in that field,
an effort that can come indeed only from a great national metrological laboratory.
Whereas in that field it is (or was) possible to foresee the time when tables of
standard thermodynamic functions would be complete, and be kept up to date
as new substances were prepared, in the field of thermodynamic properties of
mixtures it is scarcely possible even to contemplate our ever having a reasonably
complete set of properties even for mixtures of only two substances.
Both kinds of quantity are needed if the yield of a chemical reaction is to be
calculated under any particular conditons. Standard thermodynamic functions
lead to standard equilibrium constants. Except at infinite dilution for gases or
solutions, a condition of little importance in chemical industry, the properties of
real mixtures are needed to calculate the yield from the standard equilibrium
constant. No Haber-process chemist would keep his job for long who calculated
the yield at 50 or 100 MPa from the standard equilibrium constant. Nor do we
have even for mixtures of nitrogen + +
hydrogen ammonia measured values
of the properties needed.
Faced with this problem the scientist must look to improved theories from
which the properties of mixtures can be calculated from the properties of the
pure substances, and to the careful choice of mixtures and of experimental
methods by which to test those theories. The present volume is entirely devoted
to the thermodynamic properties of fluid mixtures of non-electrolytes. (Elec-
trolyte mixtures are amply covered elsewhere.)
A computerized bibliography of measurements of thermodynamic properties
of binary mixtures is included. I believe that such a bibliography is much more
useful than any attempt at compilation of smoothed and selected ‘critical’
tables. We hope that the bibliography will be kept up to date and reissued from
time to time, probably as computer output. In the meantime any of the group
who compiled it would be glad to receive notice of omissions or of new entries.
September 1977 M. L. McG.
Contents

Chapter 1 The Measurement of Thermodynamic Excess Functions

of Binary Liquid Mixtures 1
By K. N. Marsh
1 Introduction 1
2 Excess Gibbs Function - Data Reduction 3

3 Excess Gibbs Function - Experimental Methods 9

Dynamic Methods 9
Static Methods 12
Gas Chromatographic Methods 23
Isopiest ic Methods 25
Light Scattering Methods 25
Measurement of the Melting Temperature of Mixtures 26

4 Excess Enthalpies 28
Adiabatic Calorimeters 28
Isothermal Calorimeters 30
Flow Calorimeters 35
Test Systems 38

5 Excess Volumes 39

Chapter 2 Activity Coefficients at Infinite Dilution from Gas-

Liquid Chromatograp hy 46
By T. M. Lefcher

1 Introduction 46
2 Theory 47
The Unrefined Theory 41
Refinements to the Theory 48
The True Retention Time 50

3 Apparatus and Experimental Procedure 52

General 52
Columns 54
vi Contents
Temperature Control 55
Pressure Measurement and the Flow System 55
The Measurement of Retention Time 56
Flowmeters 56
Sample Injection, Detection, and Size 58
The Dead-space Volume 58
Adsorption Effects 59
Ancillary Data Requirements 60
4 Experimental Results 60
General 60
n-Alkane Mixtures and Related Systems 60
Systems containing Aromatic Compounds 62
Other Systems 62
5 Theoretical Treatment of Activity Coefficients 66
General 66
The Hildebrand-Scatchard Solubility Parameter Theory 67
Segment Theories 67
The Perturbation Treatment 68
Other Theories that have been applied to yr3 Results 69
6 Mixed Second Virial Coefficients 69
7 Conclusion 70

Chapter 3 Experimental Methods for Studying Phase Behaviour

of Mixtures at High Temperatures and Pressures 71
By C.L. Young
1 Introduction 71
2 Critical Properties 72
Critical Temperatures of Pure Substances 72
Critical Pressures of Pure Substances 73
Critical Volumes of Pure Components 75
Critical Properties of Mixtures 76
Factors affecting the Accuracy of the Measurement 82
Mercury Vapour 82
Effect of Dissolved Air 82
Decomposition of the Sample 83

3 High-pressure Vapour-Liquid Equilibrium 83

Flow Methods 83
Static Methods 90
Non-optical Methods 91
Optical Methods 98
4 Concluding Remarks 104
Contents vii
Chapter 4 High-pressure Phase Diagrams and Critical Properties
of Fluid Mixtures 105
By G.M. Schneider
1 Introduction 105

2 Two-phase Equilibria in Fluid Mixtures 107

A. Gas-Liquid Equilibria 107
Gas-Liquid Critical Curves of Binary Systems 107
Interference of Gas-Liquid and Liquid-Liquid
Equilibria 108
Critical Azeotropy 109
Interference of Solid Phases 111
B. Liquid-Liquid Equilibria 112
Results for Binary Systems 112
Thermodynamic Description 117
Effect of Solid Phases 119
Effect of a Third Component 121
C. Gas-Gas Equilibria 124

3 Results for Selected Systems 129

A. CO, + Hydrocarbon Systems 129
B. Water + Hydrocarbon Systems 134
C. CH4 + Hydrocarbon Systems 137
D. Inert Gas Mixtures 141

4 Conclusions 142

Chapter 5 Mixtures containing a Fluorocarbon 147

By F. 1. Swinton

1 Introduction 147

2 Mixtures containing a Non-aromatic Fluorocarbon 147

A. Liquid Mixtures 147
B. Gaseous Mixtures 150
C. Recent Fluid Mixture Theory 153

3 Mixtures containing an Aromatic Fluorocarbon 154

A. Liquid Mixtures 154
B. Solid Mixtures 161
C. Gaseous Mixtures 165
D. Miscellaneous Non-thermodynamic Measurements 167

4 Gas-Liquid Solubility Measurements 169

viii Contents
Chapter 6 Specific Interactions in Non-electrolyte Mixtures 174
By A. G. Williamson

Chapter 7 Volumetric Properties of Gaseous Mixtures 199

By C. M. Knobler

1 Introduction 199
2 The Virial Equation of State 199
Temperature Dependence and Magnitude of Virial
Coefficients 201
Temperature Dependence of 8 202
Thermodynamic Properties of Gas Mixtures 202
3 Experimental Methods for Determining Virial Coefficients 204
Gas Mixing Methods 205
Solubility in Compressed Gases 206
The Chromatographic Method 209
Calorimetric Measurements 210
4 Calculation of Interaction Virial Coefficients 21 1
Combination Rules 213
Corresponding States Relations 216
The Principle of Congruence 218
Non-central Potential Functions 219
Relation between Gaseous and Liquid Mixtures 220
5 Vapour-phase Complexes 221
6 Third Virial Coefficients of Mixtures 224
7 Excess Functions for Moderately Dense Gas Mixtures 226
Behaviour of the Excess Functions 228
8 Bibliography of Measurements of Interaction Virial
Coefficients 23 1

Chapter 8 Critical Exponents for Binary Fluid Mixtures 238

By R. L. Scott

1 Introduction 238
2 Critical Phenomena in Binary Mixtures 239
3 Non-analyticity. The Critical Exponents 243
Contents ix
4 Inequalities, Homogeneous Functions, and Scaling 246

5 Some General Experimental Problems 249

The Precision of Limiting Quantities 249
Gravitational Effects 250
Symmetry 252
Impurities 256

6 Experimental Methods and Results 258

Coexistence Curves 258
Other Thermodynamic Properties 263
Light Scattering 265

7 Some Tentative Conclusions 270

Chapter 9 A Bibliography of Thermodynamic Quantities for 275

Binary Fluid Mixtures
Edited by C.P. Hicks

Author Index 539

The Measurement of Thermodynamic Excess
Functions of Binary Liquid Mixtures
BY K. N. MARSH

1 Introduction
This chapter deals with experimental methods for determining the thermo-
dynamic excess functions of binary liquid mixtures of non-electrolytes. Most
of it is concerned with techniques suitable for measurements in the temperature
range 250 to 400 K and the pressure range 0 to 100 kPa. Techniques suitable
for lower temperatures will be briefly reviewed. Techniques for measuring the
molar excess Gibbs function GZ, the molar excess enthalpy ? and
I & the molar
excess volume V g will be discussed. The molar excess entropy Sg can only be
determined indirectly from either measurements of GZ and HE, at a specific
temperature I : = (HE, - G g ) / T ] , or from the temperature dependence of
S
Gg [SZ = - (3Gg/8T‘)J,l The molar excess functions have been defined by
McGlashan in the Introduction to Volume 1 of this series.l
Experimental excess functions of liquid mixtures are useful in that they provide
data to test theories of liquid mixtures and provide a guide for the formulation
of new theories. The data are also useful in the chemical and petroleum
industries. This chapter does not contain a summary of experimental data
since Chapter 9 of this volume consists of a bibliography of excess function and
related measurements on binary mixtures of non-electrolytes.
To date there has been no comprehensive review written on experimental
techniques. Williamson’s survey in 1967 is now out of date as a number of
new techniques have since been published. Moreover, his survey did not
include a detailed analysis of the accuracy of the various techniques.
The excess Gibbs function, unlike the excess enthalpy and excess volumes,
cannot be measured directly. The methods from which GE can be derived
include: the simultaneous measurement of the total vapour pressure p, mole
fraction y in the vapour phase and mole fraction x in the liquid phase at a fixed
temperature, or any two of these three, in either a recirculating still or a static
vapour-pressure apparatus; gas-liquid chromatography; isopiestic measure-
ments ; Iight-scattering measurements ; and freezing-temperature determinations
combined with excess enthalpies measured over a temperature range. In 1967
HAla et aLa gave a detailed review of techniques suitable for vapour-liquid
M. L. McGlashan, ‘Chemical Thermodynamics’ (Specialist Periodical Reports), The Chemical
Society, London, 1973, Vol. 1, Chap. 1.
* A. G. Williamson, ‘An Introduction to Non-electrolyte Solutions’, Oliver and Boyd, London,
1967.
a H. Hhla, J. Pick, V. Fried, and 0. V i l h , ‘Vapour-Liquid Equilibrium’, Pergamon, Oxford,
1967, 2nd English edn.
1
2 Chemical Thermodynamics
equilibrium measurement but their review was confined mainly to recirculating
stills. In recent years the dynamic method has been superseded by the static
method for measuring isothermal vapour-liquid equilibria data. The dynamic
method is still commonly used to obtain isobaric data. This review will include
only a limited discussion on recirculating stills and the reader is referred to
ref. 3 for further information on this method. Recent significant developments
in techniques for the determination of GZ include two stepwise dilution static
vapour-pressure systems,*#ti a compression apparatus for the measurement of
the dew pressure of a vapour mixture of known composition,s a precision freezing-
temperature apparatus,' and the extension of gas-liquid chromatography to
finite concentrations.*
Methods for measuring the excess enthalpy HE were last reviewed by
McGlashan in 1961.9 Since that review the most widely used calorimeters have
been modifications of that described by Larkin and McGlashan.lO Recently a
number of stepwise dilution calorimeters, based on the original design of Savini
et al.," have been de~cribed.l~-'~With this technique only two runs are required
to cover the whole composition range, as each run can give values of HE, for
more than half the volume fraction range. These calorimeters have a higher
precision than previous calorimeters and they have the added advantage that a
large number of results can be obtained quickIy. Flow calorimetric techniques
have been developed in recent years to give a precision equal to that of the
Larkin and McGlashan type ca10rimeter.l~ The main disadvantage of a majority
of flow calorimeters is that large quantities of pure liquids are required, some
requiring up to 300cm3 of each component for a few measurements. The
method has however served to confirm results obtained by other mefhods.l5-l7
Until recently, excess volumes VE have been determined by precision density
measurements on mixtures of known composition. If one or both components
are volatile, then the method can give values of VE which are quite wrong,
Great care must be taken both in the measurements and in applying the
corrections due to vaporization during any transfer process. All serious measure-
ments of VE should be made from direct measurements but this is difficult at
low temperatures. One method is to use a dilatometer in which known amounts
of each component are initially separated, usually by mercury, and then mixed.
The volume change is observed in a fine precision-bore capillary. A recent
R. E. Gibbs, and H. C. Van Ness, Ind. and Eng. Chem. (Fundamentals), 1972, 11, 410.
R. P. Tomlins, Royal Aust. Chem. Inst. 5th National Convention, Papers of Physical Chemistry
Division; R. P. Tomlins and K. N. Marsh, J. Chem. Thermodynamics, 1976, 8, 1185.
(I E. R. Brewster and M. L. McGlashan, J.C.S. Faruday I, 1973, 69, 2046.
R. H. Stokes, J. Solution Chem., 1974, 3, 673.
J. R. Conder and J. H. Purnell, Trans. Faraday SOC.,1969, 65, 824, 839.
M. L. McGlashan, in 'Experimental Thermochemistry', Vol. 11, Ch. 15, Heats of Mixing, ed.
H. A. Skinner, Interscience, London, 1967.
lo J. A. Larkin and M. L. McGlashan, J . Chem. SOC.,1961, 3425.
C. G. Savini, D. R. Winterhalter, L. H. Kovach, and H. C. Van Ness, J. Chem. and Eng. Data,
1966, 11, 40.
l2 D. R. Winterhalter and H. C. Van Ness, J. Chem. and Eng. Dara, 1966, 11, 189.
l3 R. H. Stokes, K. N. Marsh, and R. P. Tomlins, J . Chem. Thermodynamics, 1969, I, 211.
l4 B. blurakami and G. C. Benson, J. Chenr. Thermodynamics, 1969, 1 , 559.
l6 M. L. McGlashan and H. F. Stoeckli, J . Chem. Thermodynamics, 1969, 1 , 589.
l8 J. M. Sturtevant and P. A. Lyons, J. Chem. Thermodynamics, 1969, 1 , 201.
l' B. S. Harsted and E. S. Thornsen, J. Chem. Thermodynamics, 1974,6, 549.
Thermodynamic Excess Functions of Binary Liquid Mixtures 3
development, based on the method used by Geffcken et aZ.,18 for aqueous
solutions, is the stepwise dilution d i l a t ~ m e t e r . ~ With
~ - ~ ~this dilatometer one
component is added stepwise to the other component and the volume change is
measured directly. Two runs are required to cover the composition range.
Benson and co-workers have confirmed the results obtained by dilution dilato-
metry on cyclohexane + benzene with a magnetic float methode2 and with a
mechanical oscillator d e n ~ i m e t e r . ~ ~

2 Excess Gibbs Function-Data Reduction

The molar excess Gibbs function Gg(T, pe, x ) of a liquid mixture of A B +
containing mole fraction x of B in the liquid phase at temperature T and pressure
p* is defined by

where pz and p3 are the excess chemical potentials of component A and B

respectively. Provided p* is less than atmospheric pressure, where the gas-phase
behaviour can be represented by pV, = RT(1 +
Bp), the excess chemical
potentials are given with amply sufficient accuracy by the relations:

where y is the mole fraction of B in the vapour phase, p z and pfi; are the vapour
pressures of pure A and of pure B, p is the total pressure, Vi* and Vk* are the
molar volumes of pure liquid A and of pure liquid ByBAAand BBBare the second
+
virial coefficientsof A and of B, 8~ = BAB - (BAA BBB)/2 where BABis the
second virial coefficient for AB interactions, and V i and Vb are the partial
molar volumes of A and of B at mole fraction x of B. The last term of equations
(2) and ( 3 ) is generally less than 0.1 Jmol-l if the excess volume is less than
1.0 cm3mol-1 and ( p e - p ) is less than 100 kPa.
Measurements on second virial coefficients of pure substances and mixtures
made prior to 1969 have been compiled by Dymond and Smith 24 and by Mason
and S p ~ r l i n g . ~ ~
Cox and Lawrenson26 have listed more recent measurements on pure gases
in the first volume of this series and Knobler discusses gas mixtures in Chapter 7
l* W. Geffcken, A. Kruis, and L. Solana, 2. Phys. Chem., 1937, 35B, 317.
lo H. D. Pflug and G. C. Benson, Canad. J. Chem., 1968,46, 287.
2o R. H. Stokes, B. J. Levien, and K. N. Marsh, J. Chem. Thermodynamics, 1970, 2, 43.
21 R. Chareyron and P. Clechet, Bull. SOC.chim. France, 1971, 2853.
22 I. A. Weeks and G. C. Benson, J. Chem. Thermodynamics, 1973,5, 107.
23 0. Kiyohara and G. C. Benson, Canad. J. Chem., 1973,51,2489.
24 J. H. Dymond and E. B. Smith, ‘The Virial Coefficients of Gases. A Critical Compilation’,
Clarendon, Oxford, 1969.
2s E. A. Mason and T. H. Spurling, ‘The Virial Equation of State’, Pergamon, Oxford, 1969.
aa J. D. Cox and I. J. Lawrenson, ‘Chemical Thermodynamics’ (Specialist Periodical Reports),
Chemical Society, London, 1973, Vol. I, Chap. 5.
4 Chemical Thermodynamics
in this volume. Second virial coefficients for both pure components and mixtures
can also be estimated from various empirical c ~ r r e l a t i o n s . ~ ” - ~ ~
Until recently, the most common method for obtaining P m consisted of the
independent measurement of the three quantities p, x , and y by means of a
recirculating still. The method has the major advantage that @ can be
calculated directly from equations (l), (2), and (3) by straightforward arithmetic.
The two major problems in using recirculating stills are the difficulty in
establishing a steady state which differs insignificantly from true equilibrium
and the difficulty in analysing the liquid and condensed vapour samples.
A second advantage of measuring all three quantities is that one can test for
thermodynamic consistency because the three variables are not independent
but are related by the Gibbs-Duhem relationship, which at constants p and T
reduces to
(1 - x)dpz(T, p*, X) + xdpE(T, pg,X) = 0. (4)
This equation is rarely used for testing the consistency of p, x , and y data,
the claim being made that it is difficult to obtain the gradients with sufficient
accuracy except where there are a considerable number of points spread over the
entire composition range. The most common method for testing consistency is
to use an integrated form of the Gibbs-Duhem equation:

The equal-area consistency test, proposed independently by Coulson and

Herington 32 and Redlich and K i ~ t e r , involves
~~ plotting ln((1 - y)xpg/
[y(l - x)pX)>against x and if the areas above and below the x axis are equal,
then the data are considered to be consistent. Other equations, based on
equations (3), (4), and ( 5 ) but including a term for non-ideality of the gas phase,
have been However, a more detailed examination of equation (6)
shows that it is a consistency test only in the most restricted sense. The only
data needed to construct the plot are x , y, px, and pz. Except for the pure-
component values, the pressure measurements are not involved. The only test
27 R. C. Reid and T. K. Sherwood, ‘The Properties of Gases and Liquids’, McGraw-Hill,
New York, 1966, p. 132.
E. M. Danzler, C. M. Knobler, and M. L. Windsor, J. Chem. Phys., 1968,72, 676.
28 J. A. Huff and T. M. Reed,J. Chem. and Eng. Data, 1963,3, 306.
30 A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Trans. Faraday SOC.,1966,62,2341.
s1 G. A. Bottomley and T. H. Spurling, Austral. J. Chem., 1967, 20, 1789.
32 E. A. Coulson and E. F. G. Herington, Trans. Faraday Soc., 1948,44,629.
33 0.Redlich and A. T. Kister, Ind. and Eng. Chem., 1948,40,341.
a4 J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, Butterworths, London, 1969, 2nd edn.,
p. 128.
Thermodynamic Excess Functions of Binary Liquid Mixtirres 5
made by the area method is to determine whether the vapour and liquid
compositions are self-consistent and if the ratio pz/pg is appropriate for a given
set of x , y data. In other words, no matter how badly the pressure is measured,
as long as x , y, and the pure component vapour pressures are measured correctly,
the equal-area test would show the data to be consistent, even though the
individual values of pX and pB may be completely wrong. In fact any plot
which is based on the Gibbs-Duhem equation in the form
p m P*,x ) - P W , P*Y x ) = d@l(T,P*,x))/dx, (7)
will give rise to a ratio in which the total pressure will cancel. This statement is
only correct when the total pressure p and y are measured. If the partial pressures
were measured individually by some method which does not involve a knowledge
of the total pressure, then the equal-area test and other related tests would be a
valid test of consistency.
Recently Heringtons5 has extended the area test in a different form by con-
sidering the integrals

To analyse for consistency over parts of the composition range he considers

+
three properties of the integrals, (a) IA IB = constant; (b) the curves of IA
+
and I B against x are mirror images about (IA I B ) / ~and ; (c) the properties
(a) and (b) apply to any restricted range of composition.
The question arises as to the best method for reducing the simultaneous
measurements of p, x , and y to obtain GE. The advantage of simple arithmetic
in using p, x , and y directly is now minimal with the advent of high-speed
computers. There are three alternative methods available, each making use of
only two of the variables and the Gibbs-Duhem relation. With a perfect set of
data one should be able to calculate the same chemical potentials using all the
variables or any two of the variables. Alternatively one uses one pair of the data
to predict the third variable, which is then compared with the experimental value.
This method of data reduction and testing has been considered in detail by Van
Ness et aLS6
The excess Gibbs function GE and the total pressurep can be calculated from
( x , y ) data by assuming GE to have some functional form, for example:

whence:

m
pE/RT = ( 1 - x ) ~ [A,(1
Z - 2x)'-'(l
am0
- 2x(i+ l))]. ( 1 2)
3b E. F. G. Herington, J. Appl. Chem., 1968,18,285.
H. C.Van Ness, S. M. Byer, and R. E. Gibbs, Amer. Inst. Chem. Engineers J., 1973,19, 238.
6 Chemical Thermodynamics
For the special case of m = 1
P ~ R = +
T x ~ ( A ~ , ( 3- 4 4 1 , (13)
&/RT = (1 - x)"& + A,(1 - 443. (14)
From equations (2), (3), (13), and (14) and assuming ideal-gas behaviour, it
follows that
(P2 - P m T = 1n"l -Y)XPS/(Y(l - xlpX3I
= AO(2x - 1) - Al(1 - 6x + 62)). (15)
If y , x,p:, and p s are known, the coefficients A , and A , can be determined by
applying a least-squares analysis to equation (15). Once the coefficients are
determined, the total pressure p can be calculated from equations (2) and (13)
or (3) and (14). This approach can be readily extended to the general case as
m
(p2 - pB)/RT = x[[Aj(l - 2 ~ ) ~ - ~ { 2 i x-( lx) - (1 - 2 ~ ) ~ ) ] . (16)
i=O

The gas-imperfection terms can be easily incorporated as a rough estimate of

the total pressure would be sufficient. If pressure measurements have been made,
then consistency requires that the differences between the measured and calculated
pressures scatter randomly about zero. Where there is a distinct bias the data
can be judged to be inconsistent.
There are two distinct ways of reducing (p, y ) data. The first involves assuming
a specific form for GE(x), for example, equation (10). The resulting equations
(11) and (12) can be combined with equations (2) and (3) and the 2N equations
can be solved for the ( N + M ) unknowns by some systematic improvement of
trial values for the N values of x and M values of A . The usual method
minimizes the pressure residuals. The second method assumes no explicit form
for GE(x). Brewster and McGlashan6 make a least-squares fit to the pressure
in the form:
- pid = AjXid(l - + i ) (1 - 2X9+2 (17)
5=0

wherepid and xid are defined by the relations:

The derivative dp/dy is obtained by differentiation of equation (17) and the mole
fraction x is obtained by substituting equations (2) and (3) into (4)to give the
relation :
(1 - Y1-l - p-'dP/dy - PAdP/dY
x= - (2Y2aAB/RT)dP/dY - 4ysABP/RT
9 (20)
(1 - Y1-l -k Y-' - (PA - PBkbfdy
+ (2(1 - 2y)aAB/RT)dp/dy - 4aABP/RT
where /3c = (BCC- V p ) / R T . If experimental values of x have been deter-
mined they can be compared with the calculated values to determine consistency.
 Thermodynamic Excess Functions of Binary Liquid Mixtures 7
In the analysis of ( p , x) measurements, equations (1 1) and (12) can be combined
with equations (2) and (3) and the 2M equations can again be solved for the
( N + M ) unknowns by trial for the N values of y and A4 values of A.37
Alternatively equation (20) can be rearranged to:

which can be integrated by a marching procedure starting at an appropriate

limit. Usually p is expressed as a function of x over the whole or part of the
composition range by a series expansion. Some care must be exercised in selecting
the appropriate starting point for the integration. This method of analysing ( p , x)
data has been discussed in detail by Van N ~ s s . ~As * well as testing for consistency
of ( p , x,y ) data, these alternative methods can be used to calculate GE when only
(x,y ) , ( p , x), or ( p , y ) data are available.
For most binary mixtures, the excess Gibbs function can be presented within
experimental accuracy by a simple function of composition. Equation (10) is
the most commonly used expression and is known variously as the Redlich-
K i ~ t e r S, ~~ ~a t c h a r d and
, ~ ~ Guggenheim 41 equation. For binary systems where
the excess Gibbs function is approximately symmetrical, a good fit can usually
be obtained with two to four terms. The use of more than three terms in equation
(10) can lead to unreliable limiting values of the excess chemical potentials
because the successive terms have alternating symmetry, increasing in oscillation
the higher the term. Myers and Scott 42 have discussed this feature of the
equation and conclude that it is an unsatisfactory expression for highly skewed
data. They suggest the modification:

+
where k is a skewing factor, ranging from - 1 to 1. This Reporter has found
that for unsymmetrical systems such as benzene + octamethylcyclotetrasiloxane
the same standard deviation was obtained with 3 coefficients and a skewing
factor as was found with 5 or 6 coefficients using the unmodified equation (10).
Another equation in common use is the regular-solution equation of Scatchard
and Hildebrand,43
G 3 R T = A J L ( M 1 - $1, (23)
where V,(x) is the molar volume of the mixture at volume fraction 4 of B. If
V g is assumed to be zero, differentiation gives the chemical potentials
pf/RT = AVF/{(l - x ) V ~ / ( X V+~1l2,
) (24)
+
pE/RT = A V F / [ x V y / ( ( l - x ) V ~ } 112. (25)
37 J. A. Barker, Austral. J. Chem., 1953, 6, 207.
38 H. C. Van Ness, Amer. Inst. Chem. Engineers J., 1970, 16, 18.
3B 0. Redlich and A. T. Kister, Ind. and Eng. Chem., 1948, 40, 341.
G . Scatchard, Chem. Rev., 1949, 44, 7.
41 E. A. Guggenheim, Trans. Faraday SOC.,1937, 33, 151.
42 D. M. Myers and R. L. Scott, Ind. and Eng. Chem., 1963, 55,43.
43 (a) G . Scatchard, Chem. Rev., 1931, 8, 321; (6) J. H. Hildebrand and S. E. Wood, J. Chem.
Phys., 1933,1, 817.
8 Chemical Thermodynamics
A more general form of equations (24) and (25) has been proposed by Van
Laar 44
/J,~IRT= a/{l +
~ ( -l x ) / ~ x } ~ , (26)
pE/RT = b/[l + (bxla(1 - x ) } ] ~ , (27)
GE,/RT = abx(1 - x)/{a(l - x) + bx}. (28)
This particular two-parameter equation has been very successful in representing
highly unsymmetrical G g data.
Another equation often used by chemical engineers is the Wilson equation:46
GEIRT = - C[(1 - x)ln(l - Bx} - xln{l - A(l - x)}]. (29)
With C = 1, this two-parameter equation has been found to give a much better
fit to unsymmetrical GZ data than equation (10) with 4 or 5 parameters. Values
of C greater than unity are required to fit data which show liquid-liquid phase
separation.
Another equation frequently used is the Renon equation.46 This three-
parameter equation is also very successful for systems where GZ is highly
asymmetrical. Other equations due to Wohl and to Margules have been discussed
in detail by Hala et aL3 It should be noted that all the equations for GE are to
some degree arbitrary. For mixtures which show unusual composition
dependence, either the numerical integration method for ( p , x ) data or the
differentiation method for ( p , y ) data is to be preferred since it is difficult to
predict in advance if the deviations observed between the calculated and the
experimental values are due to errors in the data or are due to an inadequate
function for GE,. To determine the accuracy of GZ data it is necessary to consider
the magnitude of the various terms in equations (2) and (3). For the case where
pX = 100 kPa, p z = 60 kPa, p - p z = p - pg = 20 kPa, B u = B B = ~
-900cm3mol-l, 8~ = 1 0 0 ~ m ~ m o l - ~Vy, = Vy = 100cm3mot-l, and
T = 300 K, the second term contributes about 20 J mol-1 to Gg(0.5). The
third term contributes about 4 J mol-l, while the last term contributes generally
less than 0.1 Jmol-l. To obtain Gz(0.5) to an accuracy of 1 Jmol-l for the
above conditions, it is necessary to know the second virial coefficients for the
pure components and the mixture to within 20cm3mol-l. For non-polar +
polar gas mixtures and polar +
polar gas mixtures the usual methods for
calculating are far from reliable. For example, if the calculated values of
8- for acetone + nitromethane at 300 K listed by Reid and Sherwood 27 were
used in equations (2) and (3) to calculate Gg(0.5) an error of approximately
30 J mol-1 would be introduced. In circumstances where the second virial
coefficients are not known and the prediction methods are unreliable, it is
advisable that the coefficients for the pure components and the mixture be
measured. It should be sufficient for B ~ to B be determined from one measure-
ment at y = 0.5.
The accuracy of GZ(x) is, however, primarily determined by the accuracy of
p , x, and y as manifest in the first term of equations (2) and (3). Because of the
wide range of methods used to obtain vapour-liquid equilibria data and because
** J. J. van Laar, 2.Phys. Chem., 1910, 72, 723.
45 G. M. Wilson, J. Amer. Chem. Soc., 1964; 86, 127.
H. Renon and J. M. Prausnitz, Amer. Inst. Chem. Engineers J., 1968, 14, 135.
Thermodynamic Excess Functions of Binary Liquid Mixtures 9
errors differ for each individual point, it is impossible to be specific about errors.
As a generalization, the error in Gg(x) due to the first term in the equation is
SGS(x)/RT = Sp/p + + Sy/v + (1 - {Spz + ST(dp%dT)}/pz
~ X / X X)

+ x(sp$ + 6T(dp;5/dT)]fp& (30)

The temperature coefficient of GS(x) has been neglected in this error analysis.
The major effect of an error in T is related to the difference in temperature ST
between when p is measured and when p i or p;C;is measured. All of this error
has been included in the pX and pj!j term in equation (30).
This equation needs to be modified when only (p, x), (p, y ) , or (x, y ) is used
to derive GZ or when one component is involatile. In the former case, one of
the error terms drops out but a further error will be introduced because an
equation in some form, or an integration, or a differentiation, is necessary to
reduce the data. When component B is involatile, pf is calculated directly from
equation (2) using only ( p y x )data since the vapour phase consists only of the
volatile component; pg and hence Gg can only be calculated by making use of
the Gibbs-Duhem equation.

3 Excess Gibbs Function - Experimental Methods

Dynamic Methods.-In a normal recirculating equilibrium still, the compositions
of the vapour phase and the liquid phase are determined at a known temperature
and pressure. It is the determination of the vapour-phase composition which
presents difficulties, particularly for mixtures of organic compounds, because
usually neither of the components has any special characteristics which enable
an accurate chemical analysis. For such cases it is necessary to compare some
physical property of the unknown mixture, normally density or refractive index,
with that property of a mixture of known composition. Usually an accurate
measurement on the density or refractive index on a mixture requires volumes
varying from 1 to 5 cms. It is the problem of obtaining a sufficiently large sample
of the condensed vapour, along with the difficulty of establishing a steady state
which differs insignificantly from true equilibrium, which has led to a large
number of designs of recirculating stills. All recirculating stills are based on the
common principle shown in Figure 1.
The mixture in A is boiled and the vapours pass through B. After complete
condensation, the condensate is collected in C and returns, in a controlled
manner, to A for reboiling. After starting the still, and when C just fills with
condensate, the contents of C will be richer in the volatile components than the
vapour phase over the boiling mixture. On further operation the contents of C
returns to A, which becomes richer in the more volatile component, while C is
depleted. After continued recirculation a steady state is attained when the
compositions of both phases remain invariant with time. At this stage the
compositions of the liquids in A and C are determined. It can be shown that at
this steady state the composition of the liquid in the receiving flask must be
identical with the composition of the vapour phase, provided there is no entrain-
ment of droplets of boiling liquid. It is possible, however, to reach this steady
state when the vapour leaving the boiling flask is not in continuous thermo-
dynamic equilibrium with the boiling liquid. This occurs when there is super-
heating of the liquid and partial condensation of the vapour. It is these two
10 Chemical Thermodynamics
features of equilibrium stills which are difficult to check. A further problem with
some designs is the time taken to reach a steady state. In most cases a minimum
of 2 to 3 h is required but in some cases the time taken can exceed 40 h.
At normal pressures, when both components have roughly similar volatilities,
the problem of superheating and partial condensation can be overcome by
suitable design. However, when one component is much more volatile than the

Figure 1 Principle of recirculating still. A, boiler; B, condensor; C , receiver

other, or when measuring at low pressures, these problems increase and hence
the method is limited in its application. Techniques have been described which
are suitable under these extreme condition^.^^-^^ Hala et aL3 have made a
comprehensive review of the various recirculating st ills described in the literature
before 1965. Their review includes a detailed analysis of the advantages and
disadvantages of certain selected stills. One recirculating still which has been
found to behave satisfactorily over a wide temperature and pressure range is
due to Browns0 and is shown schematically in Figure 2. The liquid mixture is
boiled in flask C by an electric heater. To promote even boiling, finely powdered
Pyrex glass is sintered on to the walls of the boiler. The intimately mixed liquid
and vapour are forced up the taper and through a Cottrell pump D. The vapour
and liquid, after passing over a thermometer, are separated at R, the vapour
condensate being collected in trap F and liquid in G. The pressure is controlled
from above the condenser H. This still incorporates a number of advantageous
features. The magnetically controlled valves allow the traps to be isolated while
the liquid and vapour streams are still recirculating, while the traps are so
designed that contamination of the sample is avoided when liquid drains down
from the walls after circulation is stopped. Errors in the composition deter-
mination from evaporation losses can be reduced by cooling the traps before
opening to the atmosphere. The cooler I cools the part of the liquid stream that
mixes with the condensed vapour eliminating the possibility of flash boiling which
can occur when the more volatile condensate mixes with the hot returning liquid
phase. Because of this last feature, this instrument gives reliable results for
systems with a high volatility ratio. Some recent precision measurements of
47 K. Dvorhk and T.Boublik, CON.Czech. Chenz. Comm.,1963, 28, 1249.
F. E. Williams, Ind. and Eng. Chem., 1947, 39, 779.
E. S. Perry and R . E . Fuguitt, Ind. and Eng. Chem., 1947,39, 782.
I. Brown, Austral. J. of Sci. Research, 1952, SA, 530.
 Thermodynamic Excess Functions of Binary Liquid Mixtures 11
GEmade with recirculating stills have used modificationsof Brown’s apparatus.51’52
Benson and Boulik 5 3 54 ~ have recently described a modification of Dvorak
and Boublik’s still.*’ It incorporates a stirrer in both the liquid and condensate
receivers. Stirring reduced the possibility of density gradients within the receivers

Elevo tion
a t ‘A’

Figure 2 Recirculating still. C, boiling flask; D, Cottrell pump; R, thermometer well

and droplet separator; H, I, condensers; F, G, receivers with electromagnetic valves
(Reproduced by permission from Austral. J. Sci. Research, 1952, 5A, 530)

and hence reduces the time taken to reach a steady state. A similar design has
been described by Maripuri and R a t ~ l i f f . ~ ~
Since recirculating stiIIs are unsatisfactory for measuring vapour-liquid
equilibrium in systems of limited miscibility, various flow methods have been
developed. In the flow method a steady stream of boiling liquid or preheated
vapour of constant and known composition is fed into an equilibrium cell.
The liquid and the vapour phase are then separated and analysed. A good example
of this type of instrument has been described by Vilim et a1.,56and is shown in
61 T. J. V. Findlay and R. S. Kenyon, Austral. J. Chem., 1969,22,865.
52 S. N. Bhattacharyya and A. Mukherjee, J. Phys. Chem., 1968,72, 56,
53 T. Boublik and G. C. Benson, Canad. J. Chem., 1969,47,539.
54 J. Polak, S. Murakami, V. T. Lam, and G . C. Benson, J. Chem. and Eng. Data, 1970,15,323.
56 V. 0.Maripuri and G. A. Ratcliff, J. Chem. and Eng. Data, 1972, 17,366.
66 0.Vilim, E. Hhla, J. Pick, and V. Fried, Chem. Zisfy, 1953, 47, 1663,
12 Chemical Thermodynamics
Figure 3. The starting mixture is introduced into the supply vessel Z and the
float mechanism K maintains a constant level of liquid in the vaporizing tube V.
The liquid is preheated by heater Ht, boiled by heater HI, and the mixture of
liquid and vapour spurts on to the thermometer well T in the equilibrium

\
M

Figure 3 Flow still. V, vaporizing tube; H1, H2, heaters; K , float; T , thermometer
well; R, droplet separator; Z, supply vessel; Jk, liquid sampler; Jp, vapour sampler
(Reproduced by permission from Chem. Listy, 1953,47, 1663)

chamber. The liquid and vapour are separated and, after total condensation,
collected in the multi-receivers Jp and Jk. Equilibrium is established in 10 to
15 min and the liquid is at the boiling point for a few min only. This apparatus
would require modifications for isothermal measurements.
In principle it is necessary to measure only p and x or p and y with the
equiIibrium still method. However, because of the practical difficuIties in
checking the correct operation of a still, it is advisable to measure p, x, and y
as the results can then be checked by applying one of the various tests previously
described,
Static Methods.-When one of the components is involatile, the vapour phase
consists entirely of the volatile component; hence it is unnecessary to analyse
this phase. Under these conditions recirculating stills are of little use. The most
satisfactory method for these systems is the static method. In a static vapour-
 Thermodynamic Excess Functions of Binary Liquid Mixtures 13
pressure measurement, a mixture of known composition is contained in a vessel
attached to a manometer, both being immersed in a thermostat. The manometer
may be a nulling type, a differential type, or an absolute type. If an external
manometer only is used then it is necessary for the manometer temperature to
be above the thermostat temperature so as to prevent condensation.
The major difficulty with the static method is the need to remove completely
all traces of air from the system and the liquids. Until recently, volatile liquids
were degassed by multiple distillations in vacuum between two containers, the
condensate container being cooled in liquid nitrogen or a dry ice + acetone
slurry. The air is usually removed by pumping the system after all the volatile
liquid has condensed into the cold container. Eight or more distillations are
usually considered necessary to achieve sufficient degassing. Recently more
rapid and efficient degassing techniques have been investigated. Following the
suggestion by Barker 37 that the vapour-phase composition of a mixture of two
volatile liquids can be calculated from a knowledge of T, p , and x over the
composition range, the static method has also been used extensively for measuring
the vapour pressures above mixtures of volatile liquids. The method has the
disadvantage that it cannot be used when one of the components decomposes.
One of the early precision static systems described by Allen, Everett, and
Penney6’ is suitable for measurements at room temperature and above. The
unknown vapour pressure is balanced against a known pressure of air using a
metal-bellows nulling manometer. The null position is determined optically.
The method suffers in accuracy as the zero of the nulling device varies with
temperature and the metal parts of the bellows cannot be adequately degassed
prior to the introduction of the sample. Baxendale, Eniistun, and Stern58have
described a similar apparatus, the bellows being replaced by a mercury ‘cut-off’
nulling manometer. In the apparatus described by McGlashan and Williarn~on,~~
shown in Figure 4, the ‘cut-off‘manometer M, in the thermostat is not used as a
null device. The pressure required to transfer the measurement from M3 to the
main manometer M, is adjusted to within 6 kPa of the pressure above the
mixture. The technique used by McGlashan and Williamson will be described
in detail.
A known mass of degassed involatile material is contained in cell C of known
volume. This container has a break seal S3which separates the involatile material
from the manometer. The volume between the mercury meniscus in the ‘cut-off’
manometer and the break seal S, is determined as a function of the distance
between the meniscus and the pointer P, by compressing nitrogen in the space
and using the known diameter of the precision-bore tubing of the manometer.
An ampoule containing a known mass of degassed volatile material is contained
in D. By suitable manipulation of the vacuum taps and withdrawing the mercury
below the ‘cut-off’, this sample is distilled into the cell C . As the volatile material
distils through a grease tap TI3, it is necessary to keep the vapour pressure below
0.2 kPa to prevent losses by adsorption on the grease. The vapour pressure
above the mixture is then determined at the required temperatures, by making
67 P. W. Allen, D. H. Everett, and M. F. Penney, Proc. Roy. SOC.,1952, A212, 149.
68 J. H. Baxendale, B. V. Eniistiin, and J. Stem, Phil. Trans., 1951, A243, 169. .
5s M. L. McGlashan and A. G. Williamson, Trans. Faraday SOC.,1961, 57, 588.
14 Chemical Thermodynamics
measurements at known temperatures close to the required temperatures and
interpolating. The condition of equilibrium is established by the absence of
hysteresis in the measurements obtained after increasing and after decreasing the

i
T3‘

T -
9 (?b

MI M2 I 1 thermostat tank

I10

II IF

Figure 4 McGlashan and Williamson’s static apparatus

(Reproduced by permission from Trans. Faraday Soc., 1961, 57,588)

temperature. It was found that unless the surface of the liquid is continuously
disturbed by a stirrer, equilibrium can take up to 40 h. With efficient stirring,
equilibrium is normally achieved within 1 h of a major temperature change.
At the completion of the experiment a second sample of volatile material is
distilled into the cell and the measurements repeated. Usually the vapour
pressures of from four to eight mixtures are determined and a final check on the
composition may be made by distilling out the final mixture and determining its
mass. With this apparatus the pressure could be determined to rt 3 Pa and the
Thermodynamic Excess Functions of Binary Liquid Mixtures 15
temperature reproduced to k 0.003 K. G6mez-Ibhfiez and Shieh Go describe a
similar apparatus except that five ampoules could be loaded simultaneously.
The time spent in degassing the vacuum system was considerably reduced.
As with the McGlashan and Williamson apparatus, the volatile liquid had to be
distilled past a greased tap.

T5

CJ

Figure 5 Marsh’s static uapour-pressure apparatus for two volatile components.

A, vapour pressure cell; By mercury cut-ofl manometer; C , mercury reservoir;
T4, Teflon tap in thermostat to control mercury height in manometer; T1, T2, Teflon
taps; D, ampoule loaders
(Reproduced by permission from Trans. Faraday SOC.,1968, 64, 883)

Measuring the pressure primarily on an external manometer increases the

complexity of the apparatus and also increases the number of operations and
measurements required for a single pressure determination.
In this laboratory an apparatus containing an internal manometer has been
used 61 (Figure 5 ) . The two components are introduced to the vapour-pressure
cell A from weighed ampoules D via a single ‘cut-off’ manometer B contained
within the thermostat, the maximum measurable pressure on this manometer
being about 80 kPa. Adsorption during the transfer process was reduced by
using high-vacuum Teflon taps at T1 and T2. Such an arrangement leads to a
tall thermostat. The whole of the vacuum apparatus was arranged on a frame
with a winch so that it could be removed completely from the thermostat during
the transfer steps.

60 J. D. G6niez-Ibiiiiez and J. J. C. Shieh, J. Phys. Chem., 1965, 69, 1660.

K. N. Marsh, Trans. Faraday SOC.,1968,64,883.
16 Chemical Thermodynamics

Figure 6 Gibbs and Van Ness’s static vapour-pressure apparatus with injection of liquids
[Reproduced by permission from Ind. and Eng. Chem. (Fundamentals), 1972, 11,4101

Gaw and Swinton eliminated the external manometer by using a differential

technique. The difference between the vapour pressure of the mixture and the
vapour pressure of one of the pure components was measured directly.
Adsorption problems were reduced by using mercury ‘cut-off’ devices in place
of taps. With the differential method the error due to temperature differences
among a set of supposedly isothermal measurements is much reduced. Scatchard,

‘1
F

Figure 7 Degassing apparatus. A, reflux chamber; By cold finger; C , liquid storage

bulb; D, vacuum stopcock; E, Teflon needle valve; F, port to piston-injector
[Reproduced by permission from Znd. and Eng. Chem. (Fundamentals), 1972, 11, 4101
6% W. J. Gaw and F. L. Swinton, Trans. Faraday Soc., 1968,64, 637.
Thermodynamic Excess Functions of Binary Liquid Mixtures 17
Wilson, and Satkiewicz63 have described an apparatus using a bellows null
device but, like Allen et al.,57 they experienced difficulty with a shifting zero.
The zero of the null device has to be determined before each experiment. Their
equilibrium cell was contained within a large air thermostat capable of control
to rt 0.001 K over short periods. Other static vapour pressure apparatuses have
recently been d e s ~ r i b e d . ~ Singh
~ - ~ ~ and Benson 67 measured vapour pressures
to 323 K using a quartz spiral pressure gauge coupled directly to a vapour-
pressure cell. To overcome the condensation problem, the gauge and the
connecting tubes were controlled at 333 K. Smith and Robinson 68 connected
the cell to an absolute-pressure transducer immersed within the thermostat.
In the latter apparatus the liquids could be degassed in situ.
Gibbs and Van Ness4 prepared solutions by metering known volumes of
degassed liquids from accurate piston injectors. With such a system, measure-
ments on a complete system at one temperature can be made in one day. As it
is likely that modifications of their technique will eventually replace, to a large
extent, present methods of making ( p , x ) measurements with static systems, their
methods will be described in detail.
A schematic diagram of the apparatus is shown in Figure 6. Purified liquids
are degassed, then transferred to the two injector units where they are stored
under a positive pressure. One component is metered into the vapour-pressure
cell, and after equilibration, the vapour pressure of the pure component is
recorded. A small amount of the second component is then injected into the
cell and, after equilibration, the vapour pressure of the mixture is recorded.
The process is repeated until over half the composition range is covered, A
second run is made with the components added in the reverse order.
The degassing unit (Figure 7) is based on the design of Bell et aL70 The sample
to be degassed is added to bulb C . With tap E closed, A is evacuated through D.
+
The cold finger B is filled with an ice water mixture. With tap E opened,
heat is gently applied to C until the liquid refluxes. Heating is then removed,
E closed, and A evacuated. This sequence is repeated a number of times and is
then followed by a vacuum sublimation using either liquid nitrogen or dry ice
as the coolant in B. During the sublimation the system is pumped continuously.
The rate of sublimation is controlled through needle valve E. With E closed,
the thawed liquid is withdrawn through F into the piston injectors.
Figure 8 gives details of the piston injectors and vapour pressure cell. Only
one of the injectors is shown. The piston injector is a commercial 100cm3
capacity liquid metering pump capable of operating under vacuum conditions.
The displacement volumes can be read to 0.01 cm3. The vapour-pressure cell
consists of a 100 cm3 glass vessel connected to a metal lid E by an O-ring seal.
The lid supports a magnetic stirrer, a set of baffles, and an evacuation port H.
m G . Scatchard, G . M. Wilson, and F. G . Satkiewicz, J. Amer. Chem. Soc., 1964, 86, 125.
64 R. R. Davison, W. H. Smith, and K. W. Chun, Amer. Inst. Chem. Engineers J., 1967, 13,
590.
65 S. G. d’Avila and R. S. F. Silva, J . Chem. and Eng. Data, 1970, 15, 421.
66 G . R. Bruce and G. N. Malcolm, J. Chem. Thermodynamics, 1969,1, 183.
67 J. Singh and G. C. Benson, Canad. J. Chem., 1968,46, 1249.
68 V. C. Smith and R. L. Robinson, J . Chem. and Eng. Data, 1970, 15, 391.
68 H. Wolff and H. E. Hoppel, Ber. BunsengesellschaftPhys. Chem., 1968,12, 710.
T. N. Bell, E. L. Cussler, K. R. Harris, C. N. Pepela, and P. J. Dunlop, J. Phys. Chem.,
1968,72,4693.
18 Chemical Thermodynamics
The two small needle valves I, integral with the lid, are used to control the flow of
liquid from the pressurized injectors into the celI. The pressure is measured with
a quartz spiral gauge thermostatted at 322 K. The connecting line J is heated
to a temperature above the bath temperature to prevent condensation. With
A

K
a

"1 I-

T
O

Figure 8 Piston injector and vapour-pressure cell. A, line to degassing vessel; 3, piston-
injector body; C, packing nut; D, lead screw; E, cell cover; F, glass cup; G , Teflon-
coated magnet; H, port to vacuum system; I, needle valves; J, heated line; K, line to
reference vacuum; L, Texas Instruments pressure gauge; M, piston
[Reproduced by permission from Ind. and Eng. Chem. (Fundamentals), 1972, 11,4101
such an arrangement measurements cannot be made above 322 K. This restriction
has been overcome by interposing a differential pressure-null indicator between
the cell and the pressure gauge and using dry nitrogen on the pressure measuring
side to achieve a null. The null device and the line leading to the vapour-pressure
cell are controlled at a temperature just above the bath temperature. When the
liquids are withdrawn into the piston injectors, the piston M is advanced to put
the liquid under a pressure of above 500 kPa, determined by an accurate break-
point torque wrench. Unless the liquid is pressurized and the material flow
through the inlet I due to the vacuum in the cell is less than the flow rate of the
injector, it is possible for a vapour bubble to form in the injector.
Initially 25 cms of one component is injected into the cell and the vapour
pressure measured. A further 25 cm3 is added and the pressure redetermined.
Thermodynamic Excess Functions of Binary Liquid Mixtures 19
Disagreement between the two measurements usually indicates insufficient
degassing. A run consists of injecting a known volume of the second component
into the cell. The mixture is stirred until no pressure change is observed and the
stirrer is then stopped and a further reading is taken after an additional 10 min.
Further additions are made until 50cm3 of the second component is added.

Figure 9 Tomlin’s static vapour-pressure apparatus with injection of liquids. A, vapour-

pressure cell; B, capacitance null manometer; C, motorized valve; D, variable leak;
E, F, metal bellows valve; G , Teflon valves; H, degassing apparatus; M, nitrogen pressure
regulator; L, nitrogen ballast tank (35 1); V, to high vacuum line
(Reproduced by permission from Royal Australian Institute of Chemistry, 5th National
Convention, Papers of Physical Chemistry Division)
To calculate the masses of the liquids injected it is necessary to know the
densities of the liquids at the temperature and pressure in the injectors. The
injectors are not thermostatted but set above the bath in an air-conditioned
room. Because the injectors themselves have a large heat capacity, rapid
fluctuations in the room temperature will not have a major effect. Long-term
drifts in the temperature would be more serious. Compression of the liquids
to 500 kPa will cause an increase in density of about 0.05 per cent. Van Ness
points out that there will be a negligible error in the mole fraction calculated
from atmospheric data provided all volume measurements on both liquids are
made at the same applied pressure and the liquids have similar isothermal
compressibilities. When the isothermal compressibilities of the two components
differ appreciably a more detailed calculation is necessary to determine the mole
fraction.
An apparatus based on similar principles has been constructed in the author’s
laboratory.6 It is shown schematically in Figure 9. Only one of the piston
injectors is shown. The major differences from the Gibbs and Van Ness apparatus
are as follows. (a) The piston injector, capacity 20 cm3 graduated to 0.001 cms,
20 Chemical Thermodynamics
is immersed within the thermostat. It is driven by a motor through a clutch which
is designed to slip when the pressure on the liquid is slightly above 100 kPa.
(b) A capacitance nulling manometer B, immersed within the thermostat,
separates the vapour-pressure cell from the mercury manometer. The variation
in the capacitance which occurs when the sensing head is deflected is used to
operate a servo system comprising capacitance manometer, phase-advance
amplifier, pressure controller, and motorized valve C which controls the pressure
in the backing line. The capacitance manometer thus acts as a null device,
maintaining a pressure differential across the diaphragm of less than 0.7Pa.
Dry nitrogen, stored in a 35 dm3 container, is maintained at approximately
3 kPa above the backing line pressure by a vacuum regulator. Nitrogen is
continuously removed from the backing line through the needle valve D at a flow
rate adjusted to give optimum control. (c) The micro valve J consists of an
orifice of diameter 0.63 mm into which is inserted an 0.56 mm diameter wire.
This small orifice reduces the liquid flow sufficiently so that a pressure of only
100 kPa is required to prevent the formation of a vapour space within the
injectors. The apparatus is designed to be used over an extended temperature
range (up to 423 K) and the accuracy of the measured quantities at 303 K are:
pressure, 3 Pa; temperature (IPTS-68), 0.002 K ; volume, 0.002 cm3. Typically
a run covering half the composition scale involves the addition of 10cms of
each component.
In all the static methods so far described the total pressure and the total
amount of substance of each component in the cell are measured. To calculate
the liquid composition it is necessary to know the volume of the vapour space,
which should preferably be small. The method of transferring known masses of
material to the vapour-pressure cell involves extreme care in degassing the
samples and the vacuum line. Both McGlashan and Williamson5@and the
Reporter 61 have described in detail methods of preparing degassed samples and
the precautions necessary to obtain quantitative transfer. An alternative method,
described by both Harris and Dunlop71 and Hermsen and Prausnit~,'~ is to
sample the liquid phase and determine the composition by either refractive index
or density measurements. With this technique the mixture may be degassed
in situ. With the sampling technique, it is necessary to determine the property
for a synthetic mixture. Further, if one of the components is volatile, care must be
taken during the transfer process to prevent composition changes resulting
from evaporation. The technique of transferring known masses of degassed
liquids into the vapour-pressure cell is the most accurate method of establishing
the liquid composition, provided the volume available to the vapour is not
excessive. The technique of metering known volumes of liquids, though not as
accurate, is more convenient. It is important that the composition be known
accurately, as the Barker or integration methods invariably treat the mole
fraction of the liquid phase as the independent variable.
Gilmour et al.73have described an apparatus for measuring the vapour pressure
of mixtures at low temperatures. They used to 'click' gauge as the null device
'* K. R. Harris and P. J. Dunlop, J. Chem. Thermodynamics, 1970,2, 805.
R. W. Hermsen and J. M. Prausnitz, Chem. Eng. Sci., 1963, 18, 485.
73 J. B. Gilmour, J. 0. Zwicker, J. Katz, and R. L. Scott, J. Phys. Chem., 1967, 71, 3259.
Thermodynamic Excess Functions of Binary Liquid Mixtures 21
and the vapour-pressure cell was isolated from the gas burette by freezing mercury
into the capillary delivery line. Staveley and his colleagues 74* 75 have described
apparatus suitable for measuring the vapour pressure of mixtures of rare gases.
Measurements have been made up to 900 kPa. At such high pressures it is

Press u re+ 4Vacu u m

I
n

-G

Figure 10 Simon and Knobler’s static system for low temperatures. A, jilling tube;
B, diferential pressure transducer; C , coil; D, thermostatted bulb; E, valve; F, gas
burette system; G, mercury manometer
(Reproduced by permission from J. Chem. Thermodynamics, 1971, 3, 657)

necessary to include the third virial coefficient in the equation for calculating
the chemical potential.
Simon and Knobler 76 used the apparatus shown in Figure 10 to measure the
vapour pressures of mixtures of CH, +
CF, at 98 K. The cell, volume 1 cm3,
is connected to one side of a differential pressure transducer B by the capillary
tube A. The other side of the transducer is connected to a large insulated ballast
flask D. The pressure in the cell is determined by varying the pressure in the
flask D until a null is indicated on the pressure transducer. The pressure was
determined to k 3 Pa. The mixtures are prepared over mercury in the thermo-
statted calibrated gas burette F. A known volume of the gas mixture is condensed
into the cell through valve E. The mixture is stirred by activating a small magnet
in the cell by passing a current through the coil C.
When only one of the components is volatile most of the static methods
previously described can be used. An alternative is to use a McBain balance in
a closed system. The involatile component is placed on the balance in a thermo-
statted evacuated chamber and the mass of the sample is determined by the
74 V. Mathot, L. A. K. Staveley, J. A. Young, and N. G. Parsonage, Trans. Faraday SOC.,
1956,52, 1488.
7L R. H. Davies, A. G. Duncan, G. Saville, and L. A. K. Staveley, Trans. Faraday SOC.,1967,
63, 855.
76 M. Simon and C. M. Knobler, J. Chem. Thermodynamics, 1971, 3, 657.
22 Chemical Thermodynamics
extension of the quartz spiral. A small amount of vapour of the volatile compo-
nent is admitted and the pressure and the mass of volatile component in the
liquid mixture on the McBain balance is determined after the system has reached
equilibrium. Further additions of the solute are made to give the ( p , x )
isotherm. 77
An alternative method for obtaining CE when both components are volatile
is to determine the vapour pressure and vapour phase composition of the mixture.
Christian, Neparko, and Affsprung '* measured the vapour-phase composition
above a liquid mixture by a gas-density balance. With this technique it is not
necessary to remove any sample for analysis. Dixon and McGlashan79 have
proposed a method for determining GE for a mixture of liquids of comparable
volatility in which no precise measurement of the composition is necessary. The
method involves measuring the bubble pressure and the dew pressure of a
mixture of given, but only roughly known, composition. The dew and bubble
pressures can be readily determined by slowly decreasing the volume of an
initial gas mixture and observing the discontinuities in the pressure-volume
curve. An advantage is that the mixture second virial coefficient B- can be
determined from the ( p , V ) results obtained during the compression of the gas
phase. In the calculation of CE both x and y are unknown, but they are related
since the composition of the liquid phase at the bubble point has the same
composition of the gas phase at the dew point. The method, although appearing
attractive, is difficult because of problems in measuring the true bubble pressure.
The presence of even very small amounts of non-condensable gases in the initial
gas mixture can lead to quite spurious bubble pressures. Another alternative
is to measure the dew pressure of a gas mixture. Brewster and McGlashan6
have used this method for benzene + cyclohexane at 313.15 K.
They developed this method because, when measuring ( p , x ) data for tetra-
methylmethane + tetramethylstannane in an apparatus similar to that described
by the Reporter,61 they observed condensation of liquid on to the mercury
surface even when the mercury surface of the manometer was considerably
higher than the liquid-vapour interface in the equilibrium cell. This is a metastable
state which is probably brought about by slight temperature differences during
the equilibration period. They found that the return to equilibrium occurred
through extremely slow oscillations, which they ascribed to the cooling effects
caused by the liquid on the mercury surface as it distils back into the cell, coupled
with the slow rate of diffusion in the vapour phase. The composition of the
vapour phase must adjust itself to uniformity and be in equilibrium with the
liquid phase and this is a diffusion-controlled process. For the above system
they found the pressure still hunting in regular cycles after several days and the
effect could not be ascribed to variations in the thermostat temperature. Because
of this effect, they have questioned the accuracy of static ( p , x ) measurements
on mixtures of substances of comparable volatility.
When measuring the dew pressure it is necessary to make gaseous mixtures of
accurately known composition. Known masses of thoroughly degassed liquids
77 (a) D. E. Matire, R. L. Pecsok, and J. H. Purnell, Trans. Furaday SOC.,1965, 61, 2496;
(b) A. J. Ashworth and D. H. Everett, Trans. Faraday Suc., 1960,56, 1609.
S. D. Christian, E. Neparko, and H. E. Affsprung, J . Phys. Chem., 1960,64,442.
7s D. T. Dixon and M. L. McGlashan, Nature, 1965,206, 710.
Thermodynamic Excess Functions of Binary Liquid Mixtures 23
prepared by the method described by Marsh 61 are distilled into a 5 dm3 mixing
vessel thermostatted at 420 K. Mixing is achieved by forced convention using
an electrical heater inserted into a well in the bottom of the mixing vessel. A
baffle, consisting of a vertical sleeve placed over the well and sealed to the top
of the vessel with holes near the top, produces efficient mixing. Samples are
thoroughly mixed after 1 h. The uncertainty in the mole fraction of the vapour
mixture is about 7 x
A small sample of the gas mixture is trapped over mercury in a closed limb
of a U-manometer completely immersed in a 1.2 m high thermostat. The
manometer, made from 20 mm precision-bore tubing, consists of a short closed
limb with a reference mark and a tall limb open to the vacuum line. Dew
pressures are measured by compressing the gas mixture with mercury forced
from a reservoir at the base of the U-manometer. For small slow compressions
at pressures near the dew pressure, equilibration takes about 15 min. The dew
pressure is calculated graphically from a plot of the height difference in the two
mercury columns against the difference between the mercury height and the
reference mark in the closed tube. The overall uncertainty in the dew pressure
is about 7 Pa. No differences within this uncertainty were observed when
measurements were made by expansion from the two-phase region. For
benzene + cyclohexane at 313.15 K, GE calculated by the two methods previously
described agreed to within 0.2 Jmol-l. Brewster and McGlashan have re-
analysed the vapour-pressure measurements of Scatchard, Wood, and Mochel 8o
using modern second virial coefficients. Values of GE were calculated using the
( p , x , y ) method and the two ( p , y ) methods. The five sets of calculations fitted
the equation.
GE,l,/(Jmol-l) = x(1 - x)(1188.94 + 66.79(1 - 2x) + 44.78(1 - 2 ~ ) (31) ~ ) ~
with a root mean square deviation of 0.3 J mol-l.
The system benzene + cyclohexane is recommended for checking the correct
operation of vapour-liquid equilibrium systems.
The gas compression method has been used at low temperatures by Staveley
and colleagues 75 and by Levebvre and Guisset.81
Gas Chromatographic Methods.-In the total-pressure method, when only one
component is volatile, the accuracy of G 2 varies with mole fraction; the error
being greatest at low mole fractions of the volatile component. This region,
moreover, is often of most interest. The accuracy of measurement by the gas
chromatographic method is relatively independent of mole fraction from the
infinitely dilute region to relatively high mole fractions. The method is in fact
more precise at infinite dilution as the assumptions inherent in the calculations
are more likely to be valid in this region.82-86 Only a brief outline of the method
will be given here as the subject is extensively reviewed in Chapter 2.
G. Scatchard, S. E. Wood, and J. M. Mochel, J. Phys. Chem., 1939,43, 119.
C. Lefebvre and J. Guisset, Mol. Phys., 1961, 4, 199.
8a A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Proc. Roy. SOC.,1966, A295, 271.
C. L. Young, Chromatog. Rev., 1968,10, 129.
84 J. R. Conder, Adv. Analyt. Chem. Insfrumen., ‘Progress in Gas Chromatography’ (ed. J. H.
Purnell), 1968, 6, 209.
J. R. Conder and J . H . Purnell, Trans. Faraday SOC.,1969, 65, 824, 839.
8a R. Kobayashi, P. S. Chappelear, and H. A. Deans, Ind. and Eng. Chem., 1967, 10,63.
2
24 Chemical Thermodynamics
When a volatile solute is injected on to an involatile solvent spread along a
column it is necessary to pass a volume of inert gas V, (the retention volume)
along the column to move this zone of solute from the inlet to the outlet of the
column. The retention volume includes a volume V, due to that part of the
column occupied by the gas phase. The net retention volume VN, defined by
v, = VR - v,,
will vary with the concentration of solute injected, the pressure of the inert gas
carrier, and the pressure drop across the column.
+
Everett and S t ~ d d a r t , ~by' taking account of solute carrier gas interactions,
derived an equation relating the retention volume to the excess chemical potential
of the volatile component at infinite dilution (pf)" :
InVN = ln(nBRT/p3 - O m / R T - ( B a - VT)pYRT
-k 2po{(Pi/p0)~- 1){2BAC - - vka)/[3RT{(pk/pd2- 111- (33)
Here n~ is the amount of substance of stationary liquid, pX is the saturated
vapour pressure of the solute at temperature T, BACis the mixture virial coefficient
for solute + carrier gas interaction, BCc is the virial coefficient of the carrier
gas, Vy is the partial molar volume of the solute at infinite dilution in the
solvent, Vy is the molar volume of pure liquid A, and pi and po are the column
inlet and outlet pressures. The chemical potential at infinite dilution can be
calculated by measuring the retention volume of an infinitely small sample for
various inlet and outlet pressures and extrapolation to zero pressure drop across
the column. Everett and Stoddart proposed using equation (33) to determine
the mixture second virial coefficients. The precision in BACfrom this method is
nearly equivalent to the best static methods. The assumptions required to
derive the above equation have been examined by a number of 86, 89

Everett 88 considered the effect of the variation of the partition coefficient

(the ratio of solute concentration in the liquid phase and the solute concentration
in the gas phase) with the pressure drop across the column. Cruickshank,
Gainey, and YoungEvhave calcuIated the effects of carrier-gas solubility in the
stationary liquid. Conder and Purnell have extended gas chromatography
theory to measurements of finite concentrations. Their measurements on
n-hexane in squalane and n-heptane in dinonylphthalate, both at 303 K, agree
with static measurements in the volatile solute mole fraction range of 0.0 to
0.7 and 0.0 to 0.5 respectively. They conclude that the chemical potential can be
measured with an accuracy of approximately 25 J mo1-1 over the accessible
concentration range.
Apparatus for obtaining thermodynamic data from gas chromatography has
been described by a number of authors.82-86 Conder and Purnella6 described
a normal pressure apparatus suitable for working at finite concentrations.
The apparatus could be operated in three modes, viz.: (1) frontal analysis by a
characteristic point, (2) frontal analysis, and (3) elution on a plateau. The first
method is the most rapid; a complete isotherm can be measured and calculated
D. H. Everett and C. T. H. Stoddart, Trans. Faraday Soc., 1961, 57, 746.
88 D. H. Everett, Trans. Faraday Soc., 1965, 61, 1637.
88 A. J. B. Cruickshank, B. W. Gainey, and C. L. Young, Trans. Faraday Soc., 1968, 64,337.
 Thermodynamic Excess Fimctions of Binury Liquid Mixtures 25
in one day. The last two methods are more time consuming but give better
precision. The elution on a plateau technique requires no calibration of the
detector response. It is not possible to make measurements at solute
concentrations approaching unity as the retention volume tends to infinity.
The highest concentration that can be studied depends on the highest attainable
relative pressure p d p : of solute. This ratio depends on the temperature of the
saturator and the pressure drop across the column.
When using gas chromatography at normal pressures it is necessary to estimate
BAGvalues to obtain accurate thermodynamic data. Cruickshank and colleagues 46
have described a medium-pressure apparatus for obtaining BACas well as chemical
potentials at intinite dilution.
Isopiestic Method.-The isopiestic method can be used to determine GE,
provided only one of the components is volatile. The unknown solution and
reference solution containing the same volatile component but different involatile
components are placed in dishes on a metal block in an airtight enclosure. The
volatile component distils between the two solutions until their compositions
are such that they have the same chemical potential of volatile component.
After equilibration the two solutions are analysed. As the isopiestic method is a
comparative method it is necessary to know the variation of vapour pressure
with composition for the reference solution. The technique has the advantage
that no pressure measurements are required. The method has been frequently
used in the study of electrolytes and involatile non-electrolytes in aqueous
solutions. Corneliussen et aLoohave used the method to study polymer solutions.
An apparatus suitable for measurements with organic mixtures has been described
by Harris and D ~ n l o p . ~The
l method is suitable only when there is no possibility
of distillation between the involatile component in the reference solution and the
involatile component in the unknown solution.
Light-scattering Methods.-Thermodynamic properties of a binary liquid
mixture can be calculated from Rayleigh light scattering. The technique is
suitable when the refractive indices of the components differ by more than 0.1
and the excess Gibbs function is positive. Coumou and Mackor O2 have derived
a relation between GE and that part of the isotropic density scattering due to
concentration fluctuations which occur in a binary mixture. A number of
assumptions are required in deriving the relation but the method appears to
give GE accurate to about 40 Jmol-l. Apparatus suitable for Rayleigh light-
scattering measurements have been O4 Because of the approximations

required and the relatively low precision compared with static measurements,
this method has a limited application.
A more accurate light-scattering method involves analysing the ratios of the
intensities of the three closely spaced peaks (the Brillouin spectra) of the
Rayleigh scattering. A theoretical analysis by Fishman and Mountain O5

Q0 R. Corneliussen, S. A. Rice, and H. Yamahawa, J. Chem. Phys., 1963, 38, 1768.

K. R. Harris and P. J. Dunlop, J. Phys. Chem., 1967, 71, 483.
D. J. Coumou and E. L. Mackor, Trans. Faraday SOC.,1964, 60, 1726.
R. L. Schmidt and H. L. Clever, J. Phys. Chem., 1968,72, 1529.
O4 R. S. Myers and H. L. Clever, J . Chem. Thermodynamics, 1970, 2, 53.
O6 L. Fishman and R. D. Mountain, J. Phys. Chem., 1970,74,2178.
26 Chemical Thermodynamics
indicates that it is necessary to determine the intensity ratios at several scattering
angles in order to calculate the molar excess Gibbs function. The method has
been recently described by Miller and Leeg6 who studied acetone + carbon
disulphide and benzene + methanol at 298.15 K. They obtained agreement
with vapour-pressure measurements to within approximately 15 J mol-l. The
theory involves several approximations and refractive index measurements are
required over the whole concentration range.
The major application of this technique appears to be in determining the
excess Gibbs function for mixtures of non-volatile liquids.
Measurement of the Melting Temperature of Mixtures.-Melting-temperature
measurements have frequently been used to obtain excess Gibbs functions in
aqueous solutions. Few precise measurements have been made with mixtures of
organic liquids. However, in certain favourable systems the measurement of
the melting temperature of mixtures is far superior to vapour-pressure measure-
ments for obtaining excess Gibbs functions in the dilute region. For example,
cyclohexane, because of its low enthalpy of melting and hence large cryoscopic
constant, is a particularly suitable solvent. In cyclohexane, a solute forming an
ideal solution will have a melting temperature of 2 K below that of the pure
solvent at a mole fraction of solute of 0.01. This lowering can easily be measured
to 0.002 K giving an accuracy of 0.1 %. To obtain the same accuracy from
vapour-pressure measurements at say 298.15 K would require the partial pressure
of cyclohexane to be measured to 0.15 Pa or the partial pressure of the solute
to be measured to within of the vapour pressure of the pure solute. Such
precision is not obtainable and furthermore the measurements would require
corrections for the non-ideality of the gas phase. The virial coefficients of the
pure components and mixtures are in general not known to the accuracy required.
A disadvantage of the method is the possibility of the solid phase being a solid
solution.
A variety of techniques for measuring the melting temperature of mixtures
- l ~ ~some of these techniques one cannot be certain
have been d e ~ c r i b e d . ~ ~With
that the recorded temperature is the true equilibrium temperature unless a large
amount of the solid phase is present. It is then necessary to analyse the liquid
phase immediately after taking a temperature reading.loO
Stokes has used an isothermal dilution calorimeter l3 (to be described in detail
later) as an adiabatic calorimeter to follow the equilibrium melting curve of a
mixture until the last trace of solid disappear^.^ At this point the composition
of the mixture is precisely known from data on its preparation. A brief summary
of the method follows; pure solvent is thoroughly degassed, in situ, in the
calorimeter and the vapour space is removed by the addition of a known volume
of mercury. The calorimeter is placed in a thermostat controlled at about
0.5 K below the melting temperature of the solvent. It is stirred continuously.
After supercooling the solvent freezes as shown by the reversal of the time-
O6 G. A. Miller and C. S. Lee, J. Phys. Chem., 1973, 77, 2441.
97 R. J. M u m and F. Kohler, Monursh., 1960,91, 381.
E. Lieberrnann and F. Kohler, Monutsh., 1968,9!2, 2514.
O9 N. E. Vanderborgh, N. R. Armstrong, and W. D. Spall, J. Phys. Chem., 1970,74, 1734.

l o o R. H. Stokes, J. Phys. Chern., 1966, 70, 1199.

Thermodynamic Excess Functions of Binary Liquid Mixtures 27
temperature recording determined by a thermistor in the calorimeter. The
temperature rises to the equilibrium melting temperature and the thermostat is
adjusted to within 0.01 K of that temperature. Under these conditions the
energy loss from the calorimeter is about 1 Jh-l. A steady current of about
4 mA is then passed through a 100 Q heater in the calorimeter and the time-
temperature plot is recorded. On complete melting there is a rapid linear increase
in temperature. The melting temperature can readily be determined to within
5 x K from the intersection of the two linear portions of the curve. The
melting temperature of the solvent must be determined immediately prior to the
addition of the solute and air must be rigorously excluded from the system after
the degassing procedure because the solution of gases can have a marked effect
on the melting temperature. In the case of cyclohexane, the melting temperature
of an air-saturated sample is 0.17 K below that of a completely degassed sample.
A partially degassed sample open to the atmosphere would be useless for precise
measurements as the true melting temperature of the solvent will vary during an
experiment as the gases dissolve in the solvent.
To determine the melting temperature of a mixture, the calorimeter is heated
to approximately 0.05 K above the value obtained for the solvent and degassed
solute is injected into the calorimeter. The bath temperature is again lowered
and the melting curve is determined as before. The temperature is then restored
to 0.05 K above the melting temperature of the solvent and a further injection
made. The calorimeter is restored to this temperature to ensure that the mixture
which has been withdrawn along the connecting line between the burette and the
calorimeter due to thermal contraction is restored to the calorimeter. Further,
it is necessary for the connecting tube to be of sufficientvolume that the mixture
is never drawn back into the burette bulb at the lowest temperature reached.
Measurements on cyclohexane + ethanol were made in the range x(Et0H) =
0.015 to 0.08. To make measurements at lower mole fractions the burette is
disconnected, cleaned, and filled with mercury and the contents of the calori-
meter are transferred to the burette in such a manner that they never come in
contact with air.
A second dilution is then made by refilling the calorimeter with degassed
solvent and, after determining its melting temperature, the diluted mixture is
injected. For the above system this second run gave measurements in the range
x(Et0H) = 0.0015 to 0.021. A third-stage dilution gave melting temperatures
in the range x(Et0H) = 0.00035 to 0.005. At the lowest concentration the
depression in the melting temperature was 0.0811K.
The activity coefficient of the solvent, defined by 1nfA = pBRT at the melting
temperature of the solvent is given by

where AgHZ is the enthalpy of melting of pure A, Tf*is the melting temperature
of the pure solvent, Tf is the melting temperature of the mixture, ALC$,a is the
28 Chemical Thermodynamics
difference in heat capacity of the solid and liquid solvent, and HX(T, p, I ) -
H*(T, p, x) is the partial molar excess enthalpy of the solvent.
Stokes100calculated the activity coefficient of the solute, Inf& by defining a
mole-fract ion osmotic coefficient
g = - [In(fA(Tf*P,x)(1 - x)>]/x, (35)
and integrating the Gibbs-Duhem equation in the form

-ln(fB(Tf*, PIx)) = 1 -g -k gx + Jo'{(l - g)/x>dx. (36)

This activity coefficient is so defined that it becomes unity at infinite dilution of
the solute in the solvent, in contrast to the one commonly used for liquid mixtures,
which becomes unity for the pure liquid solute. The pure liquid-solute-based
activity coefficient can be calculated by combining the melting data with vapour-
liquid equilibria data at the melting temperature of the solvent. When vapour-
liquid equilibria data are known only at higher temperatures, it is necessary to
know the molar excess enthalpies of the mixture over the temperature range.
The enthalpy of melting can be readily determined by non-calorimetric
methods.lol One method involves the determination of the cryoscopic constant
of the solvent from the temperature depression observed by the addition of a
non-associating solute. The second method, using the Clapeyron equation,
involves measuring the volume change on melting for the solvent and the pressure
dependence of the melting temperature of the solvent.lol

4 Excess Enthalpies
In principle the molar excess enthalpy Hg may be derived from the temperature
dependence of the excess Gibbs free energy.
~E<TY
P*Y 4 = cE,(T,Pe,x) - w 2 3 T , Pe,x)lW,. (37)
In practice this is not a satisfactory method, as measurements of Gg rarely
extend over a temperature range of more than 15 per cent of the temperature.
Williamson has discussed the case where GE is determined at two temperatures
30 K apart in the region of 300 K. If GZ is assumed independent of temperature
then an error of i- 6 in GE will lead to an error of i- 156 in H g . To obtain
precise values it is necessary to measure HE, directly using either an adiabatic
or isothermal calorimeter. Until recently, most measurements have been made
with adiabatic calorimeters.
Adiabatic Calorimeters.-In an adiabatic experiment, the two liquids are mixed
in a vessel which is thermally isolated from its surroundings. If H E is positive
(endothermic) then there will be a lowering of the temperature. In practice,
electrical energy is usually supplied to the calorimeter partially to nullify the
temperature drop. If HEis negative (exothermic) then the temperature of the
calorimeter rises on mixing. A second experiment is necessary to determine the
amount of energy required to produce the same temperature rise. Alternatively
two identical calorimeters can be used.lo2 A known amount of electrical energy
R. H. Stokes and R. P. Tombs, J. Chem. Thermodynamics, 1974, 6, 379.
lo* D. A. Armitage and K. W. Morcom, Trans. Faraday Soc., 1969, 65, 688.
Thermodynamic Excess Functions of Binary Liquid Mixtures 29
is added to the second calorimeter in such a way that the temperature difference
between the two calorimeters is minimized.
The two main requirements of mixing calorimeters are: (a) the absence of a
vapour space and (6) some means to allow for the volume change which occurs
on mixing. McGlashan has reviewed the published mixing calorimeters prior

side view

Figure 11 Larkin and McGlashan's mixing vessel. A, B, two compartments containing

liquids; C,capillary containing mercury; D, expansion bulb; E, F, ground-glass joint;
H, heater; T,thermistors
(Reproduced from J. Chem. SOC.,1961, 3425)

to 1961. He concluded that most were deficient in terms of the above require-
ments. All the calorimeters reviewed were of the adiabatic type. One which has
been widely used and meets the above requirements is the calorimeter described
by Larkin and McGlashan.lO It consists of a glass vessel with two compartments
A and B in the upper half as shown in Figure 11. A bulb D connected to a
capillary C is attached to the vessel at F. The vessel is filled completely with
mercury and immersed in a bowl of mercury. Mercury is displaced from the
upper compartments by introducing weighted quantities of the liquids through
F by means of a hypodermic syringe with a bent needle. The capillary C is
half filled with mercury and attached to the vessel at F. The loaded vessel is
then placed in a vacuum enclosure in a thermostat. After temperature
equilibration, the liquids are mixed, in the absence of a vapour space, by turning
the vessel back and forth through 180" for a known number of times. The
direction of rotation is such that the liquid never comes in contact with the
greased ground joints E, F. The temperature change is measured by the four
thermistors T distributed over the surface of the vessel. In an exothermic
experiment the area Al under a temperature-time curve is determined. When
the calorimeter again reaches the thermostat temperature an amount of electrical
energy w is added and the calorimeter inverted for the same number of times as
30 Chemical Thermodynamics
before. The area A%from the temperature-time curve is determined. Finally,
the vessel is inverted in an identical manner to the previous two experiments to
give an area As. This represents the energy dissipated by stirring.
The molar excess enthalpy is related to the above quantities by

where n A and ~ Z are

B the amounts of substance of each component. This expression
is valid provided the pressure and the heat-leak path from the calorimeter to
the thermostat remain constant during the three stages of the experiment. The
first criterion is satisfied by means of the capillary and expansion bulb. The
volume change that occurs on mixing is taken up by the movement in the capillary.
If the volume of mixing is small compared with the volume of the expansion
bulb then the pressure change will be small. The second effect is difficult to
check and in many cases it can be the major source of error. For endothermic
systems, where electrical energy is added simultaneously with the mixing,
equations relating the excess enthalpy to the areas can be derived. The precision
of this apparatus appears to be about 0.7 J mol-1 or 0.5 per cent of the maximum
value of H E . The method is particularly advantageous when only small quantities
of purified liquids are available as measurements can be made with sample sizes
of the order of 1 cms. Most recent measurements of H E by the adiabatic method
have used slightly modified versions of the above c a l o ~ - i m e t e r . ~ ~ ~ - ~ ~ ~
Isothermal Calorimeters.-Malcolm and Rowlinson lo', lo* described an iso-
thermal apparatus based on the Bunsen calorimeter. The mixing vessel is
surrounded completely by a vessel containing a liquid, part of which is frozen
to give a mantle around the inner walls. Any energy released or absorbed during
the mixing process causes melting or freezing of the mantle and the volume
change that accompanies the process is measured by the movement of the liquid
or mercury in a capillary tube. Bunsen calorimeters can be used to measure
energy changes for slow processes. A major disadvantage is that measurements
can only be made at the melting temperature of the calorimeter fluid. Recently
Davies and Pritchard1O0 have discussed in detail the stability and accuracy of
Bunsen-ty pe calorimeters.
An alternative is the isothermal dilution calorimeter. For an endothermic
system, one component is slowly injected into the second component with the
simultaneous addition of electrical energy sufficient to maintain the calorimeter
isothermal. The addition is discontinued at any desired composition and the
excess enthalpy determined from the initial amount of substance of component
B in the vessel, the amount of substance of component A injected, and the
electrical energy added to maintain isothermal conditions. The apparatus is
normally designed so that the entire composition range can be covered in two
experimental runs. For exothermic systems it is possible to add electrical energy
Io3 L. A. Beath and A. G. Williamson, J. Chem. Thermodynamics, 1969,1, 51.
lo' P. J. Howell and D. Stubley, J. Chem. SOC.(A), 1969, 2489.
lo5 K. W. Morcom and D. N. Travers, Trans. Faraday SOC.,1965, 61, 230.
Io6 G . Lewis and A. F. Johnson, J. Chem. SOC.(A), 1969, 1816.
lo' G. N. Malcolm and J. S. Rowlinson, Trans. Faraday SOC.,1957,53, 921.
lo8 R. W. Kershaw and G. N. Malcolm, Trans. Faraday SOC.,1968,64, 323.
V. V. Davies and H. 0. Pritchard, J. Chem. Thermodynamics, 1972,4,9.
 Thermodynamic Excess Functions of Binary Liquid Mixtures 31
continuously to raise the temperature of the calorimeter a few tenths of a kelvin
above the thermostat temperature. On injection of liquid, the addition of
electrical energy is discontinued in such a way as to keep the calorimeter iso-
thermal. Operating the calorimeter in this mode requires a blank run to
determine the energy required to raise the temperature of the injected liquid to
that of the calorimeter. Alternatively, a thermoelectric cooling device can be
inserted directly into the calorimeter or connected to the calorimeter by a defined
heat-leak path. Energy is removed from the calorimeter at a constant rate and
the calorimeter is maintained isothermal, at the thermostat temperature, by the
continuous addition of electrical energy.
An apparatus suitable for endothermic systems was first described by Van Ness
and co-workers 110 in 1961. Several other calorimeters based on that original
design have been reported in the l i t e r a t ~ r e . l l -ll1-ll3
~ ~ ~ An apparatus lS which
has operated successfully in this laboratory is shown in Figure 12. The mixing
vessel A, of known volume, contains a heater B, thermistor D, and stirrers E, F.
The lower part of the vessel is filled with a known volume of mercury from the
pipette N. Degassed liquid is introduced via the inlet tube H from a syringe,
the air being displaced through the propeller shaft K. The drive chamber is also
filled with liquid and attached to the shaft. The vessel can be inspected to
ensure that no bubbles remain. The calorimeter is equilibrated, with the stirrer
operating, in a thermostat controlling to k 1 x K. A mechanically driven
piston burette 0, filled with the degassed second component, is connected to the
calorimeter by the coupling Q. The two components in the stainless-steel delivery
tube are separated by a small thread of mercury. As this amount of mercury is
usually not known, the effective zero reading of the burette has to be determined
by extrapolation. Injections of about 0.1 cm3 are made and the burette reading
against cumulative energy input is extrapolated to zero energy input. Temperature
swings during the addition process are never greater than 0.001 K for excess
enthalpies of about 200 J mol-l. Such variations produce a negligible error
in HE.
Two corrections to the measured energy are required. The power dissipated
in the vessel by the continuously running stirrer is of the order of 100 pW. This
causes the steady temperature of the vessel to be higher than that of the thermostat
by an amount AT z 0.01 K. The calorimeter therefore shows a small blank
correction which is evaluated from the results of a run where the same liquid is
used in both the burette and the mixing vessel. At first sight it would appear
that the energy input required to maintain constant temperature in a blank
run would be
h, = VCAT, (39)
where Y is the volume of liquid injected and C is the heat capacity per unit
volume. However, because injection occurs through one of the heat-leakage
paths, the injection has the secondary effect of carrying back into the calorimeter
R. V. Mrazek and H. C. Van Ness, Amer. Inst. Chem. Engineers J., 1961,7, 190.
ll1 R. S. Ramalho and M. Ruel, Canad. J. Chem. Eng., 1968, 46, 456.
lla R. Tanaka, S. Murakami, and R. Fujishiro, Bull. Chem. SOC.Japan, 1972,45, 2107.
113 M. B. Ewing, K. N. Marsh, R. H. Stokes, and C. W. Tuxford, J. Chem. Thermodynamics,
1970, 2, 751.
32 Chemical Thermodynamics

Figure 12 Isothermal displacement calorimeter. A, mixing vessel, capacity 38 cm3,

silvered except for viewing-strip, heat capacity (empty): about 20 J K-l; B, heater
(resistance about 1 0 0 R), nichrome wire 0.008 cm diameter on 2 cm x 2 cm flat mica
former; C, heater lead seals (tungsten-Pyrex); D, thermistor well: contains one drop of
silicone pump oil to improve thermal contact, thin-walled at bottom ; E, 2-blade propeller,
12 mm diameter; F, 4-blade propeller, 15 mm diameter; G, propeller shaft tube, stain-
less steel, 1.25 mm external diameter, 0.79 mm internal diameter, soldered into glass
capillary tube: propeller shaft (0.69 mm diameter) has sufficient clearance to permit
bubbles to be ejected up shaft tube during filling; H , stainless-steel inlet tube, 0.66 mm
internal diameter, cemented into glass capillary tube, inlet jet delivers upwards just
below lower propeller ; I , stainless-steel outlet tube, 1 mm diameter, with polythene
tube to H g pipette; J , vacuum jacket assembled by B5O cone and socket: silvered except
for viewing-strip, which is covered during runs; K , thermistor leads sealed through
B10 socket to permit changing thermistor: B10 cone on jacket is directly above thermistor
well; L, heater leads sealed through B14 socket; M, drive chamber; this contains a
brass rotor with minimum clearance all round: inset in the rotor is a magnet which is
driven by an external rotating magnet: the chamber is locked on the propeller shafr by
a conical Teflon gland: a flat on the propeller shaft fits into a matching socket in the
rotor shaft: the top rotor bearing is on a screw which is removed during filling; N,
mercury pipette, about 19 cm3 between fiducial marks; 0, piston-burette; P, 3-way tap,
Teflon-glass, with stainless-steel tubes attached; Q, inlet coupling: double Teflon
conical glands, brass body; R, soldered seals : glass platinized and copper-plated before
soldering; S , high vacuum Teflon tap; T , tap on mercurypipette
(Reproduced by permission from J. Chem. '17termodynamics,1969, 1, 211)
Thermodynamic Excess Functions of Binary Liquid Mixtures 33
some energy which otherwise would have been lost. An experiment is necessary
to establish the true blank correction. For the calorimeter shown, the blank
correction was found to be

The uncertainty quoted in the factor 0.7 would not contribute more than
0.07 J mol-1 to the excess enthalpy.
The second correction is required when the components differ appreciably in
density. When a less dense liquid is injected, the power dissipated by the constant-
speed stirrer falls progressively during the run. Since the temperature at the end
of an increment is always restored to its starting value, some of the electrical
energy must be used to compensate the reduced stirrer dissipation. Accordingly
at intervals during the run the injection is stopped and the slow cooling to the
new steady state is followed for about 1 h. The next injection then occurs with
a new lower baseline. From the temperature-time trace, average cooling rates
during each injection ri are obtained. Using the known heat capacity of the
calorimeter and contents, Zi, and the duration of each injection ti, the ‘cooling’
correction can be calculated. For the n’th injection the correction h, is

The largest correction found was for the injection of benzene into carbon
tetrachloride with a stirring rate of 270 min-l, where the cumulative correction
amounted to 0.27 J mo1-1 for x = 0.5. Even when the liquids have similar
densities the equilibrium temperature of the calorimeter alters during a run
because of the variation of the heat-leak path as the volume of mercury decreases.
A number of design improvements to the calorimeter shown in Figure 12 have
been made.113 The magnetically driven calorimeter stirrer has been replaced by
a directly driven stirrer. The stirrer shaft passes through a union A [Figure
13(a)] sealed by a conical gland G of Teflon impregnated with molybdenum
disulphide. The glass capillary inlet tubes have been replaced by stainless-steel
capillary tubes and instead of using mercury to separate the two liquids, the
inlet tube is branched at E as shown in Figure 13(b). One arm (P)which passes
through the surrounded vessel is closed by a small screw ball-valve X [Figure
13(c)]. A similar ball valve is connected to the coupling on the stirrer shaft.
After the calorimeter is completely filled with one component, the mercury
pipette tap is closed. The inlet tube R is then flushed with the second component,
this liquid escaping through valve X. The only part of the delivery tube which
now contains the calorimeter liquid is F, whose volume is 0.0045 cm3. Hence
the zero correction to the burette reading is fixed and no extrapolation is required.
Another modification includes replacing the outer glass vacuum jacket with a
metal jacket. Evacuation is not necessary if the jacket is spaced about 2 cm from
the walls of the mixing vessel since a time constant of about 1 h can be achieved
by filling the jacket with argon at atmospheric pressure.
One advantage of this apparatus is that it is not necessary to weigh any of the
liquids. The volume of the component in the calorimeter is defined from the
34 Chemical Thermodynamics

Yc 9
Q d GG
P PG

I
P
I1

Figure 13 Modification to isothermal displacement calorimeter.

(a) Stirrer shaft union: A, body of union stirrer shaft tube and shaft: bore diameters
1.5 mm and 0.9 mm; G, glands of Teflon impregnated with molybdenum disulphide;
C, screw caps which compress glands; Q, outlet tube, 0.90 mm 0.d. stainless-steel tube
terminating above thermostat lid in micro-valve V [part (b)].
(b) The calorimeter vessel. The sloping strokes indicate that sections of stainless-
steel tubing have been omitted from the drawing for compactness. X, V, micro ball-
valves [detail in part (c)];s, stirrer shaft, 0.9 mm diameter passing through stainless-
steel tube 1.5 mm inside diameter which is attached by brass cap B to calorimeter
body; F , inlet tube, forking at E into arms P and R, all of stainless-steel tube 0.68 mm
o.d , 0.4 mm i.d.; H , mercury outlet tube, 0.9 mm 0.d. stainless-steel tube; D , brass
bushes 3.1 mm diameter soldered to tubes. These pass through O-ring seals in the
vacuum jacket. Capacity of vessel about 38 cm3. The thermistor and heater have been
omitted from the Figure for clarity.
(c) Detail of micro ball-valve. L, 1 mm diameter steel ball seating in hole M ; K ,
groove in body [valve X,part (b)]to permit tubes R and P to be inserted together
through vacuum jacket seal. Body diameter of valves about 3.1 mm; hole N to fit
stainless-steel titbe
(Reproduced by permission from J. Chem. Thermodynamics, 1970,2, 751)
 Thermodynamic Excess Functions of Binary Liquid Mixtures 35
known volume of the vessel and the known amount of mercury added from the
pipette. Due allowances must be made for the effects of expansion of the glass
and mercury at temperatures different from the calibration temperature. A
further advantage is that the partial molar excess enthalpy can be estimated
directly provided the injections are such that they produce mole fraction changes
of the order of 0.05. The partial molar excess enthalpy of the injected component
is then closely approximated by AH/AnB, where A H is the enthalpy change
caused by the addition of an amount of substance AnB of component B. The
value obtained applies to the composition of the solution half-way through the
injection step. By using small increments in the initial stages of the experiment
the theoretically important limiting partial molar excess enthalpy of the injected
component can be determined with a precision of better than 1%.
An apparatus which is more suitable for measuring exothermic enthalpies
has been described by Van Ness (Figure 14). A thermoelectric cooling module
Z is used to remove energy at a constant rate. The cold junction of the module
is connected to a coin-silver bar Y and rod U which sits in a copper cup V
which is filled with mercury. The mercury acts as a heat-transfer agent and a
bearing for the stirrer, which consists of two coin-silver paddles. By this method
energy is removed uniformly from the calorimeter and no cold spots are produced.
The first correction discussed above may be eliminated by running the cooling
device at a power sufficient to negate the power input from the stirrer. The
apparatus consists of a vessel attached to a fixed plug. The plug contains the
injection tube M, stirrer K, temperature sensing device N, and the cooling
module. The process of lowering the vessel from the plug draws the second
component from a reservoir through a small mercury cup P into the mixing
chamber. The volume of liquid added is determined by adding a known amount
of mercury sufficient to refill the reservoir to a reference mark. Murakami and
Benson l4 have constructed a similar apparatus except that the cooling module is
located in a second plug at the base of the vessel. Tanaka et aL112describe an
apparatus which uses the same injection technique as the Van Ness type
calorimeter with a vessel of fixed dimensions in which mercury is displaced as
in the Stokes and Marsh type calorimeter.
A number of other isothermal dilution calorimeters which contain a vapour
space ll4* 116 have been described.
Dilution calorimeters normally require about 50 cm3 of each component for
the two runs necessary to cover the composition range, and H E is usually
measured at 20 to 30 compositions. This type of calorimeter appears to have a
higher precision than other calorimeters; HE can be measured to about
0.2 Jmol-l or 0.1 to 0.2 per cent of the maximum value of HE,whichever is the
greater in magnitude.
Flow Calorimeters.-In a flow calorimeter the two liquids are injected into a
mixing chamber at a steady and known rate. For an endothermic system the
power P in the heater is adjusted so that the temperature of the mixed liquid is
114 J. J. Christensen, J. W. Gardner, D. J. Eatough, R. M. Izatt, P. J. Watts, and R. M. Hart,
Rev. Sci. Instr., 1973, 44, 481.
llfi E. Kauer, D. Kirchner, D. Haupt, and H. J. Bittrich, Z . phys. Chem. (Leiptig), 1972, 250,
153.
36 Chemical Thermodynamics
restored to that of the unmixed liquid. The molar excess enthalpy and the mole
fraction x of component B are given by
HE = PKfi + fi), x = Mfl + fa), (42)
where fi and fs are the molar flow rates of the two components. Few measure-
ments have been made with flow calorimeters. The most accurate results are
those by McGlashan and Stoeckli.ls Their apparatus is shown in Figure 15.
The mixing chamber, constructed of Teflon, has a volume of 2.5 cms. The liquid
flowrates used are about 5 mg s-l giving a residence time in the mixing chamber

,/'

lk "

(a)
-2.5 cm

Figure 14 Isothermal displacement calorimeter with cooling module. A, stainless-steel

support tube; B, vent tube; C,current andpotential leads for heater; D , connector for
feed tube; E , Teflon plug; F, vent plug; G, heateriwire supports; H, bafles; I , Teflon
support; J , heater wires; K , stirrer magnet; L, stirrerpaddle; M , feed tube; N , thermistor;
P, Teflon feed cup; Q, water inlet tube; R, copper heat sink; S , 150 cms precision-bore
Dewar flask; T, O-rings; U , coin-silver cooling rod; V, copper cup; W, coin-silver
support rods; X , copper heat shield; Y,coin-silver bar; Z, cooling module
(Reproduced by permission from J. Chem. and Eng. Data, 1966, 11, 189)
 Thermodynamic Excess Functiorzs of Binary Liquid Mixtures 37
of about 250 s. They conclude that by suitably arranging the inlets, mechanical
stirring was not necessary. The major problem with flow calorimeters is in
producing a constant, but readily variable, flow rate. Using a vapour-pressure-
driven flow system, McGlashan and Stoeckli obtained a reproducibility of
0.25% in the flow rates. They concluded that H E could be measured to about
1%. However, 200 cm3 of each component were required adequately to cover the

0 1 2 Clll

3' b
Figure 15 Flow calorimeter. A, inlet ports (only twu of four shown); C , mixing chamber;
D, outlet port; E, heater plug; H, heater
(Reproduced by permission from J. Chem. Thermodynamics, 1969, 1, 589)

composition range. The calorimeter used by Sturtevant and Lyons required

considerably smaller samples, but the reproducibility was of the order of 2%.
Wormald ll8 has recently constructed a flow calorimeter which appears to give
a reproducibility of better than 0.2%. The commercial LKB isothermal flow
calorimeter which was initially designed for aqueous systems has been modified
for non-aqueous systems by various workers. Gustin and Renon 11' used 10 cm3
precision pumps which could be adjusted for flow rates from 0.01 to 0.4 cm3min-l
with an accuracy of 0.2%. They found it necessary to degass the solvent to
prevent bubble growth in the mixing cell. For the test system benzene + carbon
tetrachloride their results appear accurate to about 1 J mol-l.
Harsted and Thomsen l7 determined a calibration constant for their LKB
calorimeter by using various results determined by Marsh and co-workers. They
estimated that their modified flow calorimeter was accurate to about 2 Jmol-l.
Goodwin and Newsham118 have described a flow calorimeter for use with
alcohol + water systems.
11( K. Elliot and C. J. Wormald, J. Chem. Thermodynamics, 1976, 8, 881.
11' J. L. Gustin and H. Renon, J. Chem. and Eng. Data, 1973,18, 164.
S . R. Goodwin and D. M. T. Newsham, J. Chem. Thermodynamics, 1971,3, 325.
38 Chemical Thermodynamics
Picker et al.llDhave described an interesting calorimeter. By reducing the time
taken to reach a steady state in a flow calorimeter to less than 1 min and by
continuously varying the flow rates of the two liquids they managed to obtain
the complete excess enthalpy curve in less than 1 h. Becker et al.lZ0have described
a continuous method which requires two runs to cover the composition range.
One component is added at a constant rate to a container of fixed volume initially
containing the second component. The mole fraction of the mixture gradually
changes in a known way depending on the flow rate and the volume. Their
calorimeter was designed to have a very rapid response time so that the mole
fraction-excess enthalpy curve could be recorded in a continuous manner.
Comparisons with the test systems benzene +
carbon tetrachloride and
+
1,4-dioxan carbon tetrachloride indicate that the calorimeter is capable of a
precision of slightly better than 1 per cent. The major advantage of flow
calorimeters is that they are more readily adaptable for measurements over a
wide temperature and pressure range.
Test Systems.-Cyclohexane +
n-hexane at 298.15 K has recently been
recommended by the I.U.P.A.C. Commission on Thermodynamics and Thermo-
chemistry.121 Results for this system obtained in five different laboratories with
batch, dilution, and flow calorimeters show no systematic d i s c r e p a n c i e ~ . ~122
~-l~~
To ensure that the purity of the samples was not a source of error, all measure-
ments were made with Phillips Research Grade material. The recommendation
is that H Z should be calculated from the equation
HElJmol-l = x(1 - x){866.10 - 249.4(1 - 2x)
+
97.0(1 - 2 . ~ ) ~- 31.8(1 - 2 ~ ) ~ } , (43)
where x is the mole fraction of cyclohexane. This system is suitable for testing
calorimeters to be used for measuring positive excess enthalpies provided the
density difference between the components is not large. For large density
differences the efficiency of stirring is difficult to check. An alternative is
benzene + carbon tetrachloride at 298.15 K. Recent measurements l4 139

appear to have established H g for this system to within rt 0.5 J mol-l. However,
Watts et a1.l2l have suggested that a reaction between carbon tetrachloride and
mercury could lead to discrepancies.
There has been no systematic study on an exothermic system suitable for
testing calorimeters. The results of Murakami and Benson l4 and Tanaka
+
et aL112 for benzene dichloromethane [Hg(max) = - 84 J mol-1 at 298.15 K]
agree within 0.5 per cent. For 1,4-dioxan + dichloromethane [Hg(max) =
- 1100 J mol-1 at 303.15 K] the results of Murakami and Benson differ from
those of Winterhalter and Van Ness by up to 1.6 per cent. Measurements on
this system have been made by Becker et a1.120 at 298.15 K only. These two
systems should be investigated further as to their suitability for testing
calorimeters used for measuring exothermic systems. on behalf of the
P. Picker, J. Jolicourr, and J. E. Desnoyers, J. Chem. Thermodynamics, 1969, 1, 469.
lZo F. Becker, M. Kiefer, and H. Koukol, 2. phys. Chem. (Frankfurt), 1972, 80,29.
lal Bulletin of Thermodynamics and Thermochemistry, 1970, 13, 507.
laa H. Watts, E. C. W. Clarke, and D. N. Glew, Canad. J. Chem., 1968,46, 815.
las H. Watts, personal communication.
Thermodynamic Excess Functions of Binary Liquid Mixtures 39
Standards Committee of the Calorimetry Conference, has suggested tetra-
hydrofuran + water as a suitable standard.

5 Excess Volumes
The majority of excess volumes for liquid mixtures reported in the literature
have been calculated from the measurement of the density at known composition.
For most mixtures VE is usually no greater than 0.3 per cent of the total volume
of the mixture. To achieve a reasonable accuracy in YE, density measurements

cm

Figure 16 Po well and Swinton’s dilatometer

(Reproduced by permission from Trans. Faraday Suc., 1966, 62, 1090)

need to be made to an accuracy approaching one part in lo5. To achieve this

accuracy on mixtures containing a volatile component is very difficult as the
assumptions made when applying corrections for evaporation during the
preparation and transfer of the mixture are often inadequate. In a number of
cases, excess volumes calculated from densities have differed considerably when
compared with direct measurements. An extreme example is the liquid + solid
system benzene + diphenyl. Duff and using the density method,
calculated an excess volume at x = 0.5 of - 0.61 cm3mol-l at 303 K. Kortum
, ~ ~ ~a similar technique, obtained -0.12 cm3mol-1 at 308 K.
and S ~ h r i e b e rusing
Direct measurements by Powell and Swinton 126 gave - 0.22 cm3mol-1 at 303 K
and -0.29 cm3mol-1 at 323 K for x = 0.5. The determination of excess
volumes from density measurements should be discouraged.
A more satisfactory method is to measure VE directly in a dilatometer. In fact
the direct measurement of VE is the simplest and best method for obtaining the
density of mixtures containing volatile materials. The requirements for measuring
excess volumes directly are similar to those for measuring excess enthalpies,
except that the presence of a vapour space even as small as or cm3
may lead to considerable error. A simple apparatus used by Swinton and
colleagues is shown in Figure 16. The vessel is first filled with mercury while
124 G. M. Duff and D. H. Everett, Trans. Faraday Soc., 1956, 52, 753.
lZ5 G. Kortum and H. Schreiber, Z . Naturforsch., 1965, 20A, 1030.
126 R. J. Powell and F. L. Swinton, J . Chem. and Eng. Data, 1968, 13, 260.
127 W. A. Duncan, J. P. Sheridan and F. L. Swinton, Trans. Faraday Soc., 1966, 62, 1090.
40 Chemical Thermodynamics
attached to a vacuum system. Known masses, usually between 0.2 to 0.8 g, of
each liquid are injected into the limbs of the dilatometer. An appropriate
precision-bore capillary is fitted to the ground-glass joint, mating in such a way
that no air pockets are trapped within the dilatometer. After thermal
equilibration the mercury height is adjusted and its height relative to a reference
is measured with a cathetometer. The contents of the dilatometer are mixed by
rocking sideways and, after thermal equilibration, the mercury height is recorded.
To confirm that mixing is complete, the mixing is repeated until constancy in the
mercury height is obtained. The internal projection on the ground-glass joint to
the base of the U-tube is to ensure that the organic liquids do not enter the
capillary during the mixing operation. With a demountable capillary the
diameter of the capillary may be readily varied according to the expected size
of the volume change. A correction is necessary because the change in height
results in a change in pressure on the liquids. This correction will be discussed
later. With this type of apparatus the excess volume can be measured to 0.5 per
cent or 0.001 ~ m ~ m o 1 - l An
. apparatus similar to the calorimeter shown in
Figure 11 has also been used for excess volume measurements.*os The heater
and thermistor are removed and the capillary and expansion bulb are replaced
with a precision-bore capillary. Other modifications have been described.l2*,12@
Stookey et ~ 1have . discussed
~ ~ ~ in detail the potential sources of errors in the
‘single shot’ dilatometer.
Dilution dilatometers for measuring excess volumes offer the advantage that
the composition range can be covered in two runs and partial molar excess
volumes are readily obtained. Van der Waals and Desmyter 131 described an
apparatus which requires the liquids to be distilled under vacuum before sealing
by glass blowing. The operating procedure is complicated. Gomez-Ibaiiez and
Wang132used a modified version of that dilatometer but it suffered from the
same disadvantage. The most conveniently operated version of this type is
described by Williamson and Their apparatus is shown in Figure 17.
Known masses of two liquids are injected into the vessels A and B and are
confined over mercury, the two vessels being separated by a thread of mercury
in S. The mercury levels are adjusted to near R, and Rz in the precision-bore
capillary. The amounts of the two liquids are determined by weighing from a
hypodermic syringe. Mercury is withdrawn into the bulb Y resulting in the
transfer of liquid from A to B. The composition is established by the mercury
height in C and the known volume of the bulb X. Mercury is forced from B
to A when the vacuum is released at G . The liquids are mixed in vessel B and
readings of the mercury levels in C and D give the volume change on mixing.
The process is then repeated. A second run is made with the liquids interchanged.
One source of uncertainty arises from the difficulty in accurately calibrating the
volume of the connecting tube S. An error in the calibration leads to a
systematic error and shows in the lack of agreement in the overlapping results
las I. Brown and F. Smith, Austral. J. Chern., 1962, 15, 1.
1 3 R. K. Nigam and P. P. Singh, Trans. Faraduy SOC.,1969, 65, 950.
D. J. Stookey, H. M. Sallak, and B. D. Smith, J . Chem. Thermodynamics, 1973, 5, 741.
131 A. Desmyter and J. H. Van der Waals, Rec. Trav. chim., 1958,77, 53.
13% J. D. G6mez-Ibaiiez and T. C. Wang, J . Phys. Chem., 1966,70, 391.
n3 L. A. Beath, S. P. O’Neill, and A. G. Williamson, J . Chem. Thermodynamics, 1969, 1, 293.
Thermodynamic Excess Functions of Binary Liquid Mixtures 41
from the two runs. If the volume change is large then there is an uncertainty in
calculating the volume of liquid added from A. An alternative is to mix the
components completely in B prior to forcing mercury back to A. For this
situation the composition calculation is not as straightforward. Again com-
pression effects due to changes in the mercury heights need to be considered.

Figure 17 Beath, O’Neill, and Williamson’s dilution dilatometer. A, 30 cm3 bulb;

B, 50 em3 mixing bulb; C, D, calibrated capillary tubes; R, reference marks; S,fine-
bore connecting capillary; T1,T2, ground-glassjoints; X, Y , calibrated bulbs
(Reproduced by permission from J. Chem. Thermodynamics, 1969, 1, 293)

Mug and Bensonls have described an accurate dilatometer based on the

design of Geffcken, Kruis, and So1ana.l8 The volume of added component is
determined from the mass of mercury displaced in a piston-type arrangement.
Because the liquids pass through a greased tap its use is limited. Chareyron and
Clechet lS4 have described a similar apparatus.
184 R. Chareyron and P. Clechet, Bull. Soc. chim. France, 1971,2853.
 42 Chemical Thermodynamics
A modified version of the Geffcken et al. apparatus has been used in the
Reporter’s laboratory.20 Details of the dilatometer are shown in Figure 18.
It consists of a burette B1,Bz (containing one component) joined to a bulb A
(containing the other component) by a fine stainless-steel capillary tubing which
is sealed into bulb A through the closing plunger P. On opening tap T, mercury
flows from the bulb A to the burette and consequently liquid in the burette is
forced into the bulb through the coupling E and the stainless-steel tubing. The
volume change is observed on the calibrated tubing C and the composition is
determined from the change in mercury height in the burette which is constructed
from precision-bore tubing. The tap T, is to allow mercury to be added to, or
removed from, the system when the volume change on mixing is too large to be
accommodated in the capillary. This model has the advantage that it is not
necessary to weigh the component into A,. The volume of liquid in Al is defined
by the mercury heights M land M3 provided the volume of mercury is adjusted
to F,, F3, and F4 before loading with the liquids.
Compressibility corrections due to a change in the pressure head on the
liquids as the mercury levels change must be considered in all direct measure-
ments of VE. The corrections will be discussed in terms of the apparatus shown
in Figure 18. In what follows the mercury levels in the bulb A and capillary C
are understood to be taken relative to the reference marks. At any stage in the
experiment when T1 is closed, the pressure head of mercury acting on the system
is proportional to {h(M3) - h(M,)). The heights h(M3) and h(M,) are the
positions of the mercury surface in the capillary and bulb respectively. Now
h(Mz) is determined, to within & 0.005 cm solely by the level h(Ml) of the mercury
in the burette. The relation can be established by a blank run. Denoting the
levels before a given step of the process by primed symbols, the pressure change
for the step is proportional to [{h(MJ - h ( M 2 ) ) - (h(Mi) - h ( M i ) } ] . Before
the run, a measured air pressure is applied to the capillary C and the shift in the
mercury meniscus is measured. Denoting the meniscus shift per unit applied
pressure by 6, the volume change due to mixing is

where pm is the density of mercury, g is the local acceleration of free faII, and a
is the cross-sectional area of the capillary C. The factor 6 can be compared with
that calculated from the compressibility of the liquids and the mercury, and the
Young’s modulus and dimensions of the glass parts. The constancy of b for
various applied pressures is a useful indication of the absence of air bubbles in
the system. The apparatus is capable of a precision of -t 0.001 cm3mol-1 in the
excess volume. To achieve this precision it is necessary to control the bath
temperature to better than about 4 x K over the whole period of the run.
Normally the regulation is better than 5 x K. The compressibility
correction can be eliminated by adjusting the pressure on the capillary so that
the pressure on the liquid in the dilatometer remains constant. A suitable
device has been described by Pflug and B e n ~ o n . ~ ~
Thermodynamic Excess Functions of Binary Liquid Mixtures 43

El Il

Figure 18 The dilution dilatometer. A, glass mixing bulb, total capacity approximately
41.5 cm3; Al, initial liquid volume; A,, thick glass head with ground top and opening
to fit plunger P ; B,, B,, burette, consisting of 11 cm of 5 mm calibrated Veridiaprecision-
bore glass tubing (Bl), 24 cm of 10 mrn calibrated precision-bore glass tubing (B,),
and a small crossover section (VC). The total volume of the burette is approximately
22.5 cm3; F,, F,, F3, F,, F6,fiducial marks; MI, M,, M3, variable mercury meniscus
heights; T,, T,, 2 mm bore spring-loaded interkey taps operated by remote control
levers; S1,0.183 cm 0.d. stainless-steel tubing sealed to B, with Araldite; S z , 0.069 cm
0.d. stainless-steel tubing passing through the brass screw of E , and silver soldered to
the closing plunger P;X , iron-in-glass magnetic stirrer; El, coupling joining S1 and S2.
Double Teflon conical glands, stainless-steel body, brass nuts; E,, brass holder and
screw, with neoprene washer for sealing plunger P through A,; D, 2 mm capillary;
R, reservoir to accommodate mercury added to or subtracted from the system during
a run; C , 1 mm Veridia calibrated precision bore capillary; P,plunger having pressure
plate at top to seal bulb A. The stem of P is a close fit in the glass capillary tube of
A, so that there is no dead volume
(Reproduced by permission from J. Chem. Thermodynamics, 1970,2, 43)
44 Chemical Thermodynamics
Martin and Murray 136 have modified the above dilatometer by placing a tap
below the U-bend connecting the burette to the mixing bulb. This addition
makes it easier to fill the dilatometer with mercury under vacuum and also
facilitates the cleaning and calibration process.
Another modification by Bottomley and shown in Figure 19,
eliminates all grease taps. On tilting the diiatometer mercury flows from the tip

Figure 19 Bottomley and Scott’s tilting dilution dilatometer

(Reproduced by permission from J. Chem. Thermodynamics, 1974, 6, 973)
at A displacing an appropriate amount of the component in the burette into the
bulb through the upper capillary at B. The apparatus is sealed with two 1.25 mm
bore Teflon needle valves. Provided the temperature control is adequate, this
apparatus is the most precise of any dilatometer so far described, the precision
being about O.OOO15 cm3mol-l.
The direct measurement of excess volumes at low temperatures imposes severe
experimental problems. The usual method has been to determine the molar
volume of the pure components and the mixtures directly. The expansion method
generally used involves many potential sources of systematic error. The major
source of error is in the accurate determination of the volumes of the cell and the
expansion bulbs. Experimental techniques have been discussed in detail by
Liu and Miller.137
135 M. C. Martin and R. S . Murray, J. Chem. Thermodynamics, 1972, 4, 723.
136 G . A. Bottomley and R. L. Scott, J. Chem. Thermodynamics, 1974, 6, 973.
Y. P. Liu and R. C . Miller, J. Chem. Thermodynamics, 1972,4, 85.
Thermodynamic Excess Functions of Binary Liquid Mixtures 45
Powell and Swinton12s have suggested benzene + cyclohexane at 298.15 K
[ VE(x = 0.5) = 0.65 cm3mol-'1 as suitable for checking dilatometers. Reliable
measurements have differed by as much as 0.01 cms mol-l for x = 0.5. However,
recent measurements lso$138 with both the single shot dilatometers and dilution
2os

dilatometers indicate that VE(x = 0.5) for this system at 298.15 K is close to
(0.650 f 0.001) cm3moI-l.
In an attempt to resolve these discrepancies Benson has used two further
methods involving the direct determination of the density of mixtures of
benzene + cyclohexane. One method uses a magnetic float densimeter22
capable of a precision of 20 x in the density which corresponds to a
precision of about 0.002 cm3mol-1 in the excess volume. The results agreed
within experimental error with those of Stokes et uL20 The second method,
using an Anton Paar mechanical oscillator d e n ~ i m e t e ris
, ~capable
~ of about the
same precision as the first method and the results agreed to within 0.002 cm3mol-l.
13* M. L. Martin, personal communication; see also ref. 135.
2
Activity Coefficients at Infinite Dilution from Gas-
Liquid Chromatography
BY T. M. LETCHER

1 Introduction
Gas chromatography is primarily an analytical separation technique. However,
since the basic process is an equilibration of a solute between two immiscible
phases, the chromatographic technique may be used to measure such physical
properties as activity coefficients, second virial coefficients of gas mixtures,
partition coefficients, adsorption and partition isotherms, and complex formation
constants. Other properties which can be measured with less accuracy, from
secondary measurements or from temperature variation studies, include surface
areas, heats of adsorption, and excess enthalpies and excess entropies of solution.
A number of reviews and discussions on these measurements have appeared in
the 1iterature.l-ll The present work is restricted to a review of activity-coefficient
measurements.
The activity coefficient is the most important and fundamental property in
the thermodynamic study of liquid mixtures. It is a measure of the deviation of
the behaviour of a component in a mixture from ideality and it has been
interpreted by various theories of liquid mixtures. Gas-liquid elution
chromatography offers a rapid method of determining this property at infinite
dilution. Conder and Purnell have developed a method of determining activity
coefficients at finite concentrations I2-l5 and this has recently been used by other
workers.1s-18 To do this, the elution technique must be supplemented by
1 J. R. Conder, ‘Progress in Gas Chromatography’, in ‘‘Advances in Analytical Chemistry and
Instrumentation”, Vol. 6, ed. J. H. Purnell, Interscience, New York, 1968.
J. H. Purnell, Endeaoour, 1964, 23, 142.
D. E. Martire and L. Z . Pollara, Adu. Chromatog., 1965, 1, 335.
R. Kobayashi, P. S. Chappelear, and H. A. Deans, Ind. and Eng. Chem., 1967, 59, 63.
J. C. Giddings and K. L. Mallik, Ind. and Eng. Chem., 1967, 59, 19.
C. L. Young, Chromatog. Rev., 1968, 10, 129.
H. W. Habgood, ‘The Solid-gas Interface’, ed. E. A. Flood, Vol. 2, Marcel Dekker, New York,
1967.
H. Brusset, D. Depeyre, and M. Fromant, Chromatographia, 1972, 5, 576.
* (a) M. A. Kahn, Lab. Practice, 1961, 10, 547 and 709; (b) M. A. Kahn, Lab. Practice, 1962,
11, 120 and 195.
lo S. Kenworthy, J. Miller, and D. E. Martire, J. Chem. Educ., 1963, 40, 541.
l1 0. WiEarovi, J. Novik, and J. JBnak, J . Cfzromatog., 1972, 65, 241.
la J. R. Conder and J. H. Purnell, Trans. Foraday Soc., 1968,64, 1505.
J. R. Conder and J. H. Purnell, Trans. Faraday SOC.,1968, 64, 3100.
l4 J. R. Conder and J. H. Purnell, Trans. Faraday Soc., 1969, 65, 824.
l6 J. R. Conder and J. H. Purnell, Trans. Faraduy SOC.,1969, 65, 839.
P. A. Sewell and R. Stock, J. Chromatog., 1970, 50, 10.
C. J. Chen and J. F. Parcher, Analyt. Chem., 1971, 43, 1738.
la C. L. Hussey and J. F. Parcher, Analyt. Chem., 1973,45, 926.
46
Activity Coefficientsat Infinite Dilution from Gas-Liquid Chromatography 47
controlling the solute concentration in the carrier gas. In this type of
chromatography, the recorder no longer records a peak but an integral form of
the peak. This report will be concerned mainly with activity coefficients at
infinite dilution obtained by the elution method.
The nature of gas-liquid chromatography unfortunately limits the choice of
liquid mixture. The solute must be rather volatile if retention times are to be
reasonable and the solvent or stationary phase must be a liquid at the temperature
of the experiment with a sufficiently low vapour pressure so as not to ‘bleed’
off the column during the course of an experiment. The latter problem has to
a certain extent been overcome by the use of pre-saturators or by weighing the
column before and after an experiment, but a significant loss in accuracy is
inevitable.
Further limitations are that the choice of carrier gas is restricted to those
which are insoluble or nearly insoluble in the stationary phase and the range of
mixtures to those in which adsorption effects of any kind are negligible. For
systems which fall within these limitations accurate activity coefficients at infinite
dilution can be obtained, which would be very difficult to determine in other
ways. Furthermore, if the retention volumes are measured as a function of
pressure, the mixed second virial coefficients, BI2,of carrier gas (referred to in
this work as component 2) and solute (component 1) can be obtained simul-
taneously with the activity coefficient, yr3 (where 3 refers to the solvent).
Activity coefficients at infinite dilution are important and useful properties to
chemical engineers, solution chemists, and theoreticians. To the chemical
engineer, they are of interest in the design of plants that involve liquid-vapour
equilibrium. To the solution chemist they are important in understanding the
mixing process. Perhaps their most interesting and successful application has
been in the testing of various solution theories. This is possibly due to the large
number of systems that have been analysed (because the results can be obtained
so rapidly using g.1.c.) and to the infinitely dilute condition which in many cases
presents the theorist with simpler equations and fewer complications.

2 Theory
The Unrefined Theory.-The idea of gas-liquid chromatography goes back to
1941 when Martin and Synge, Nobel prize laureates for 1952, mentioned in a
paper on general chromatographylO that it should be possible to use a gas as a
mobile phase and a liquid as a stationary phase. They related the equilibrium
partition coefficient K to retardation properties by use of a plate theory. Their
general equation, when considered in relation to gas-liquid chromatography
(for zero pressure-difference across column), relates the retention volume of the
solute, YE,to the gas hold-up volume yC and the solvent volume V, according to
vH. = v, + KV,. (1)
Their idea was not taken up and it was left to Martin, this time in conjunction
with James,20to describe a separation using g.1.c. (They separated some volatile
fatty acids by partitioning between nitrogen and a mixed liquid phase of silicone
19 A. J. P. Martin and R. L. M. Synge, Biochem. J., 1941, 35, 1358.
ao A. T. James and A. J. P. Martin, Biochem. J., 1952, 50, 679.
48 Chemical Thermodynamics
oil and stearic acid.) More importantly, they presented the first theory specifically
applicable to g.1.c. by taking into account the compressibility of the mobile
phase. This involved applying a correction factor to the gas volumes of equation
(1). In terms of Everett‘s notation 21 this corrxtion term J : can be generalized
as
J? = n(lpi/pcdrn- 1I/m[Cpi~/o)~- 11, (2)
where pi and po refer to inlet and outlet pressures respectively.
In 1956 Martin22 again hinted at future developments in this field and
speculated that g.1.c. might be useful in studying the solution thermodynamics
of gas and liquid phases. In the same year Porter ef related the net reten-
tion volume VN to the activity coefficient of the sofute at infinite dilution, y&
according to
VN = nsRT/y?3”,p?, (3)
where n3 is the amount of liquid solvent on the column and p: the vapour pressure
of the solute at the temperature T of the experiment. The retention volume VN
is determined from the column outlet flow-rate Uo by

where tR and fG are the retention times for the solute and an unretained gas, and
VD is the dead-space volume or gas hold-up volume at mean column pressure
p&. This theory assumes that (a) idealized chromatographic conditions exist,
(b) the carrier gas and solute vapour behave as ideal gases, (c) adsorption effects
of any kind are absent, and (d) the carrier gas is insoluble in the solvent. The
measurements based on the above theory give activity coefficients for n-alkane
systems which are within a few per cent of the results obtained from the more
refined theory discussed below.
Refinements to the Theory.-The next major step in the evolution of deter-
minations of accurate activity coefficients came in 1961 when Everett and
Stoddard24 took into account the solute vapour and solute +
carrier gas
imperfections. An important outcome of this work was the possibility of
obtaining the mixed virial coefficient B12. Destya5 applied these ideas to the
determination of B12values and used an extrapolation procedure based on the
equation :
InVN = InVg + /3p& (5)
where VJ is the extrapolated retention volume at zero mean column pressure:
In V k = In(n3RT/y:8p,*) - (Bll - V r ) p f / R T , (6)
where V y is the molar volume of the solute and Bll the second virial coefficient
of pure solute. 18 is given by
B = (2Bl2 - VY)/RT, (7)
21 D. H. Everett, Trans. Furaduy Sac., 1965, 61, 1637.
2a A. J. P. Martin, Analyst, 1956, 81, 52.
2% P. E. Porter, C. H. Deal, and F. H. Stross, J . Amer. Chem. Sac., 1956,78, 2999.
t4 D. H. Everett and C. T. H. Stoddart, Trans. Faraduy SOC.,1961, 57,746.
D. H. Desty, A. Goldup, G. R. Luckhurst, and W. T. Swanton, ‘Gas Chromatography
1962’, Butterworths, London, 1962.
Activity Coefficientsat Infinite Dilution from Gas-Liquid Chromatography 49
where VF is the partial molar volume of solute at infinite dilution in the
stationary phase. Both of the above treatments assume that the partition function
does not change significantly along the column but remains constant at the mean
column-pressure value. Everett 21 attempted to avoid this assumption and
developed a detailed theory of the pressure-drop effect which led to a different
extrapolation procedure. The Bristol group z6-28 reformulated the differential
equation describing the local elution rate in the column, and suggested a third
extrapolation procedure :
1nVN = In V & + j?p&. (8)
This was tested using a numerical integration procedure and shown to be
superior to the previous extrapolation 25 Moreover, this theory

takes into account small imperfections in the carrier gas and is thus suitable for
carrier gases such as hydrogen, helium, nitrogen, oxygen, and argon.
For carrier gases which are appreciably nonideal they proposed :
In V h = In(VN(1 + bp&)/(l + bpo)) = In V & + Pp&, (9)
where b = B2,/RTand Bz2is the second virial coefficient of the carrier gas.
A further refinement by the Bristol groupz0 involved the solubility of the
carrier gas in the stationary liquid. They showed, neglecting second-order effects,
that the retention volume (for a pressure drop across the column of less than
200 kPa) is related to pressure, po, according to

where

and X is defined by the expansion of x2 (mole fraction of carrier gas in the solvent)
as a series in the local carrier-gas pressure:
x2 = xp, -k +p; 4- ..., (12)
where q5 is the coefficient of the second-order pressure term. Equation (11)
includes the two effects resulting from the solubility of the carrier gas, namely,
the increase in the amount of substance of stationary liquid and the change in
the activity coefficient due to the change in the nature of stationary liquid. The
importance of equations (10) and (11) is that it shows that because the quantity
(a In y?.Jax2) is virtually inaccessible, the property B12 cannot be obtained
unambiguously from g.1.c. measurements, with a solvent in which the carrier
gas is appreciably soluble. The significance of the carrier-gas solubility and
equation (1 1 ) in particular has been discussed in terms of experimental results,
by the Bristol group 30 and by Pecsok and W i n d s ~ r . ~ ~
2Ds

C. L. Young, Ph.D. Thesis, Bristol, 1967.

27 A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Proc. Roy. Suc. ( A ) , 1966,295,259.
28 M. L. Windsor and C. L. Young, J . Chromatog., 1967, 27, 355.
2n A. J. B. Cruickshank, B. W. Gainey, C. P. Hicks, T. M. Letcher, R. W. Moody, and C. L.
Young, Trans. Faraday SOC.,1969,65,1014.
D. H. Everett, B. W. Gainey, and C. L. Young, Trans. Faraday SOC.,1968, 64,2667.
s1 R. L. Pecsok and M. L. Windsor, Analyf. Chem., 1968,40, 1238.
50 Chemical Thermodynamics
The True Retention Time.-Existing theories of g.1.c. predict a unique retention
time. Experimentally, however, a peak spread is observed, so it is necessary to
speculate where on this peak the true retention time may be found. Figure 1
illustrates some of the peak properties that have been used to define this ‘true’
retention time. The peak initial time t~ and peak final time tF are determined
from the intersection of the base-line time with the tangents to the leading edge

Figure 1 Some peak properties that have been used to define the retention time: ( t l )
peak-initial time; ( t F ) peak-final riffle; (tT) peak-tangent t h e ; ( t M ) peak-maximum
time; ( t I E ) mean o f ( f r )and (tF)

and trailing edge respectively. The peak tangent time tT, on the other hand, is
determined from the point of intersection of the tangents to the leading edge and
trailing edge. The peak half-area time is defined as the net time that divides the
area under the peak into two equal parts. The peak maximum time tM is obtained
from the time of peak maximum, and the peak average retention time t l F is
obtained from the mean of t~ and fF.
Because of the lack of knowledge concerning the detailed processes taking
place in a g.1.c. column, the present-day theories cannot hope to unravel the
problems associated with peak asymmetry. Possible causes that have been given
by various workers $9 33 including eddy diffusion in packed columns, non-
329

equilibria between phases, sampIe size, surface heterogeneity, and non-zero

response time of the detecting system.
A somewhat surprising answer to the question of true retention time was given
in some of the earlier attempts to derive thermodynamic properties from g.1.c.
m e a s u r e m e n t ~ . ~35~ ~The peak initial time t~ was used, the justification being
349

that this gave the best agreement with data from static measurements. Later it
was shown 36 that if sufficient care was taken to achieve effective infinite dilution
of the solute and if the measurements were done over a range of pressure and
analysed by a theory which culminates in equation (8), then the peak tangent
gave good agreement with static measurements. This seems far more reasonable
as it is somewhere near the top of the peak although for skew peaks this may
not be true!
** A. B. Littlewood, ‘Gas Chromatography’, Academic Press, New York, 1962.
a* C. P. Hicks, Ph.D. Thesis, Bristol, 1970.
84 A. J. B. Cruickshank, D. H. Everett, and M.T. Westaway, Trans. Faraday SOC.,1965,61,235.
a6 D. E. Martire and L. Z. Pollara, J. Chem. and Eng. Data, 1965, 10,40.
36 A. J. B. Cruickshank, M. L. Windsor, and C. L. Young,Proc. Roy. SOC.( A ) , 1966,295, 271.
 Activity Coeficients at Infinite Dilution from Gas-Liquid Chromatography 51
Many workers 37, 38 have discussed the ‘first time moment’ or ‘centre of gravity’
of a chromatographic peak undergoing elution. In the absence of longitudinal
or eddy diffusion this property has been shown to be equal to the ideal thermo-
dynamic retention time for zero-pressure-drop columns. More recently, theories
regarding the fist time moment have been extended by Hicks 33 and by Buffham 39
to include pressure-drop columns and the first time moment has been related
to thermodynamic properties. Buffham used the ‘mean residence time’ i which
is equivalent to the ‘first time moment’:

i= IOw
tci(t)dt/joaCi(t)dt, (1 3)
where cl(t) is the response of the solute molecules i at time t after injection. He
related f to thermodynamic properties and showed that it can be used to obtain
results similar to those of Cruickshank et aLYz7 Stalkup and K ~ b a y a s h i and
,~~
Koonce et uZ.,*~ without the restrictions previously considered necessary.42
It is interesting that the first time moment for systems which do not include
longitudinal or eddy diffusion has been considered to be independent of the
43-46
kinetics of equilibration between the
Hicks 33 has recently arrived at a solution to the g.1.c. situation for which the
diffusion of solute in the liquid phase is the rate-determining step for equilibration.
In this work Hicks shows that for this situation the peak-average retention volume
f i can
~ be used to obtain thermodynamic properties. To support this, he showed
that the flow-rate dependence of retention volume, as observed in benzene +
glycerolYz9disappears, if peak-average retention volumes are used instead of
peak-tangent retention volumes. Experimentally tIF can be obtained directly
from the peak tangents without elaborate equipment or with the problems
associated with the ‘first time 47 The flow independence of 6~does
seem to point to tIF being an excellent estimate of the ideal retention time
although much more work is required on systems with asymmetric peaks to show
its range of validity. The first moment is possibly theoretically better because it
does not require assumptions about the nature of the kinetic processes governing
the equilibration between phases, but it is experimentally very inconvenient.
The problem of locating the true retention volume is, however, usually
important only for solutes which have short residence times and have very
asymmetric peaks. It is only then that the peak average retention times differ
significantly from the peak-maximum, peak-tangent, peak-half-area, or mean-
residence retention time.
Surface Adsorption Effects.-Surface adsorption is perhaps the most important
limitation of the g.1.c. method for determining activity coefficients. Martin 48
a7 D. A. McQuarrie, J. Chem. Phys., 1963, 38, 437.
a8 M. Kocirik, J. Chromatog., 1967, 30, 459.
*@ B. A. Buffham, Proc. Roy. SOC.(A), 1973,333,89.
40 F. Stalkup and R. Kobayashi, Amer. Inst. Chem. Engineers J., 1963, 9, 121.
41 K. T. Koonce, H. A. Deans, and R. Kobayashi, Amer. Inst. Chem. Engineers J., 1965,11,259.
J. A. Parcher. J. Chem. Educ., 1972, 49,472.
E. Kucera, J. Chromatog., 1965, 19, 237.
44 0. Grubner, Ado. Chromatog., 1968,6, 173.
45 0. Grubner, M. Ralek, and E. Kucera, Coll. Czech. Chem. Comm., 1966,31,2629.
46 S. Wicar, J. Novak, and N. Ruseva-Rakshieva, Analyt. Chem., 1971, 43, 1945.
47 P. R. Ronig and J. E. Funk, J. Chromatog. Sci., 1971, 9, 215.
48 R. L. Martin, Analyt. Chem., 1961,33,347.
52 Chemical Thermodynamics
suggested in 1961 that this adsorption at the gas-liquid interface can be related
to the retention volume and proposed :

where ka is the adsorption coefficient and As the area of the liquid surface.
Unfortunately there is no way of separately determining these two terms by
chromatographic experiments alone and an extrapolation procedure must be
used to obtain V,.*@

3 Apparatus and Experimental Procedure

General.-If the activity coefficients are not required to any great accuracy
( f 5 per cent uncertainty) then equation (3) will suffice, provided that adsorption

-
t
I 1 - 1
G

Figure 2 Atmospheric pressure g.1.c. apparatus. (A) Carrier gas cylinder; (B) pressure
regulator; (C) injection block; (D) column; (E) manometer; (F) katharometer detector;
(G) soap-filmflow meter. The dotted lines refer to the thermostat

effects are insignificant. For such measurements a simple gas chromatograph,

as illustrated in the flow diagram of Figure 2, is suitable. The type of detection
is not important so long as sample detection of 0.5 pmol is possible. The
advantage of katharometer detection is that the flow meter can be placed down-
stream of the column.
For more accurate determinations of activity coefficients, it is necessary to
take into account carrier gas and solute imperfections. Assuming no adsorption
of the solute on the solvent or solid support and assuming the carrier gas is not
appreciably soluble in the solvent, equation (9) is suitable:
In V& = In V$ + /3p&. (9)
The uncertainty in the activity coefficients determined in this way has been
estimated to be less than 0.4 per cent. Application of equation (9) usually
Activity Coefifcients at Infinite Dilution from Gas-Liquid Chromatography 53

2.80

2.78
i
2.76

2.74

2.72

1 I I L , I

0 0.2 0.4 0.6 0.8 1.0 1.2

Po J: / MPa
Figure 3 A typical plot of In V k versus p J i from medium-high-pressure g.l.c. This
example is for hexane + n-octadecane at 308.15 K

------- 9 -I

I
C
I
I
I
I
I
I
I
I
1,-

--
I
I
54 Chemical Thermodynamics
implies the determination of the retention volume as a function of pressure
(p&), from which the mixed second virial coefficient B12 and the activity
coefficient yY3 can be obtained. Figure 3 shows a typical plot of In V& versus
p& In the rare event of an accurate B12value being known, equation (9) need
be applied to only one experimentally determined value of retention volume VL.
Medium high-pressure g.1.c. does require more sophisticated apparatus as is
shown in the flow diagram of Figure 4. It must include a high-pressure injector,
pressure-control values, special metal-to-glass seals, and a flow meter capable
of operating at pressures of 1200 kPa. These will be discussed later in this
section.
Columns.-The accurate determination of the amount of solvent on the solid
support in the column is of paramount importance, even if moderate accuracy
is required. Experimental methods have been adequately described by P ~ r n e l I . ~ ~
Whatever variation of the technique is used, it should be possible to determine
the quantity of the stationary liquid phase (1 to 5 g) on the column to within
f 1 mg, thus reducing the error in this quantity to an insignificant level.
Solvent loadings (mass solute/mass solid support) of between 10 and 30 per
cent are normal in this work. Below this range, solute adsorption on the exposed
solid support becomes possible, whereas above this range, the column material
can become sticky and unmanageable and the retention times inconveniently
large. Much of course depends on the type of solid support and the solute-
solvent system under investigation. Celite 545 (BSS 80/100 mesh) has become
the commonly used solid support, although many others are available50 and
have been 51s52

In this work, copper columns of about 4 mm bore are convenient and most
commonly used. Copper tubing has the advantage that it can be bent into
suitable coils and is inert to most solutes. Capillary columns suffer from the
disadvantage that the amount of stationary liquid phase is difficult to determine
accurately. For solutes that react with copper, e.g. amines, glass columns may
be used. They are normally used in the form of U-tubes because of the difficulty
in filling them in coiled form.
Solvent loss by evaporation can very easily lead to incorrect results and must
be continually checked. The most convenient way to do this is to determine
periodically the activity coefficient of a particular solute chosen for its con-
venience. For some solvents it is necessary to use pre-columns or pre-saturators
to saturate the carrier gas with solvent upstream of the column. This has been
done by many workers 52-55 and is essential if the range of systems to be studied
by the g.1.c. method is to be extended. The loss of solvent has been determined
by the direct weighing of the column 56 but this has obvious drawbacks.
4eJ. H. Purnell, ‘Gas Chromatography’, John Wiley, New York, 1962.
Lo D. M. Ottenstein, ‘Advances in Chromatography’, Vol. 3, eds. J. C. Giddings and R. A.
Keller, Edward Arnold, London, 1966.
61 Y. B. Tewari, D. E. Martire, and J. P. Sheridan, J. Phys. Chem., 1970, 74,2345.
b2 R. K. Kuchhal and K. L. Mallik, J . Chem. and Eng. Data, 1972,17, 49.
6s B. L. Karger and A. Hartkopf, Analyt. Chem., 1968, 40, 215.
64 R. E. Pecsar and J. J. Martin, Analyt. Chem., 1966,38, 1661.
66 D. L. Shaffer and T. E. Daubert, Analyt. Chem., 1969, 43, 1585.
IeR. C. Pemberton and J. A. Larkin, paper 5/24 presented to the 3rd International Conference

on Chemical Thermodynamics, Baden, 3-7 September 1973.

 Activity Coefficients at Infinite Dilution from Gas-Liquid Chromatography 55
For limited accuracy, it is not necessary to purify the commercially available
stationary liquid phases (usually quoted as being better than 97 per cent pure)
but for more accurate measurements it may be necessary to zone refine before
use. Zone refining has become common practice in many laboratories and
efficient zone refiners can be made even with limited workshop facilities.57
As a test against possible column-loading errors it is advisable to measure the
activity coefficient on two different columns made up from different batches of
solvent-coated solid support. If the agreement is not satisfactory a third column
must be used.
Non-uniformity of the column packing does affect the retention volume
because of the pressure variation along the column. Measurements must there-
fore be made on both forward and reversed columns and the average taken,
if this effect is to be eliminated. If the difference between the activity coefficients
from the two sets of measurements is large the column must be repacked.
Temperature Control.-For moderate accuracy, the temperature of the column
should be controlled to within 0.2 K if the solute vapour pressure is known to
be better than 1 per cent. A good air thermostat is capable of this control.
It is, however, much easier to control the temperature of the water bath, and a
simple on-off relay in conjunction with a toluene regulator and backing heater
is capable of controlling the temperature to kO.01 K5* This is the order of
temperature control required for the accurate determination of activity coefficients.
Because it is necessary to have the column completelythermostatted and because
the injector cannot usually be immersed in a water bath, it is convenient to have
the bath mounted in a frame on pulleys so that it can be moved vertically. In
this way the column can be replaced without affecting other parts of the flow
line and, furthermore, the column can be taken out of the bath at the end of a
run, thus reducing unnecessary loss of solvent.
Pressure Measurement and the Flow System.-For moderately precise g.1.c.-
determinations of activity coefficients [as obtained from equation (3)], the
outlet pressure is usualIy considered as being atmospheric pressure and can be
measured on a normal Fortin barometer. The inlet pressure is usually deter-
mined by difference using a manometer placed in the carrier-gas flow line. An
excellent review by Ambrose 59 gives details of such pressure measurements.
For more accurate work, involving the determination of retention volume as
a function of pressure [equation (9)], inlet pressures ranging from 200 kPa to
over 1000 kPa must be measured. Such pressures have usually been measured
using a good Bourdon-tube dial-reading gauge.36 Unfortunately such gauges
can only be read to + 2 kPa thus introducing errors of up to 0.3 per cent in the
worst case.3@The outlet pressure is usually determined from the difference in
pressure across the column. This difference is usually less than 100 kPa and can
be measured on a conventional thick glass-walled manometer.
It is important to have no measurable fluctuations in inlet or outlet pressures
or in the carrier-gas flow rate. For atmospheric pressure g.1.c. a two-stage
I7 E. L. Anderson, Chem. and Ind., 1966, 1615.
6a W. J. Gaw and R. L. Scott, J. Chem. Thermodynamics, 1971,3, 335.
60 D. Ambrose, ‘Chemical Thermodynamics’, ed. M. L. McGlashan, Specialist Periodical
Reports, The Chemical Society, London, 1973, Vol. 1.
3
56 Chemical Thermodynamics
regulator attached to the carrier-gas cylinder is usually sufficient, but if not,
a more sensitive control valve such as the Negretti and Zambra or Ermeto
precision regulating valve should be used.
For medium-high-pressure g.1.c. the problem is more difficult as both the
inlet and outlet pressures must be controlled. The above types of regulating
valves have proved satisfactory when used in conjunction with a fine-control

Figure 5 The glass-to-metal coupling. (A) Metal tube; (B) rubber O-rings; (C) split
ring; (D) sauring bolts; (E) glass tube

needle valve (thermostatted in the water bath) placed at the column exit. The
order of pressure fluctuations with such valves is k 50 Pa which is well within
the limit of accuracy of the pressure measurements required.
The various parts of the apparatus can conveniently be connected using
copper tubing and demountable pipe couplings. The glass units should be
demountable and a simple O-ring seal, illustrated in Figure 5 and described by
Cruickshank and has proved to be satisfactory.
At the higher pressures, leaks pose a serious problem. The system should be
regularly checked for leaks, and the positioning of on-off valves in the line
should be arranged so that parts of the apparatus can be isolated and tested.
The Measurement of Retention Time.-The accurate measurement of retention
time TR [equation (4)] should not be determined from ruler measurements on the
chromatogram but should be obtained from good stop-watch measurements.
Paper stretching and mains-frequency fluctuations (which alter the rate of drive
of synchronous motors) are two important factors that are responsible for the
inaccuracies of the ruler method.
Flowmeters.-The measurement of gas-flow rate is usually carried out with a
soap-bubble meter although other methods involving rotameters and capillary
manometers are available. The soap-bubble meter 6o is an inexpensive, easy to
operate, direct method which has good accuracy.
The flowmeter may be placed upstream or downstream of the column depending
on the type of detection used. If thermal conductivity is used, then the flow-
IJO M. D. LaHue, R. J. Teck, G. Dombi, and H. D. Axelrod, J. Chem. Educ., 1973,50, 867.
Activity Coefficientsat Infinite Dilution from Gas-Liquid Chromatography 57
meter may be placed downstream of the detector, but if flame-ionization detection
is used, the flowmeter must be placed upstream of the column. For atmospheric-
pressure g.1.c. measurements, using thermal-conductivity detection, the soap-
bubble flowmeter need only consist of a graduated burette and a deformable
rubber reservoir. For medium-high-pressureg.1.c. this reservoir must be replaced,

Figure 6 The soap-film flow meter. (A) and (D) Copper tubing; (B) and (C) calibration
marks; (E) magnetic displacer; (F) magnet; (G) soap soZution

and a convenient alternative is the magnetic displacer 34 which is illustrated in

Figure 6.
Soap-bubble flowmeters placed upstream of the column have the disadvantage
that the soap bubbles must be removed from the flow line after travelling through
the burette and the gas must be dried before entering the column. This is best
done using two canisters, one empty (to break the bubbles) and the other filled
with a drying agent (silica gel). Freezing techniques are not always suitable and
can cause pressure fluctuations.
If the inlet flow rate Ui is measured, then the outlet flow rate Uo can be
calculated from an approximate equation of state of carrier gas, knowing the
temperature T’ of the flowmeter and the temperature T of the column from
uo = UiPiW + PoB22(T)I/Po[RT/3- PlB22(T?I (1 5 )
58 Chemical Thermodynamics
For accurate Aow-rate measurements, temperature T’ should be known to
within rf: 0.5 K. It is therefore advisable to place the flowmeter in an air thermostat
and to ensure that the carrier gas is at this temperature before entering the flow-
meter by using a long conditioning coil.
The range of flow rates usually encountered in this application of g.1.c. is
between 0.5 and 5 cm3s-l. Even with a good stop watch capable of reading to
k 0.02 s, errors of 0.2 per cent are possible at the faster flow rates. Improved

n C D

0
U E
0 F
Figure 7 Medium-high-pressure g.1.c. injector. (A) High-pressure gland; (B) displacer
piston; (C) to vaciiuin p i m p ; (D) to carrier gas supply; (E) to sample reservoir; (F) to
column; ( G ) spindle attached to ball calve
(Reproduced by permission from Chromatog. Rev., 1968, 10, 129)

precision can be obtained by eliminating the human element, using photocell

relays in conjunction with a counting device or electronic timer.61
Sample Injection, Detection, and Size.-For medium-high-pressure g.l.c., the
conventional microsyringe must be replaced by special gas-sampling valves or
‘locks’, because septum caps can withstand only relatively small pressure
differences. Much has been written about sample injection at high
pressure^.^^^ 369 63 Perhaps the most satisfactory method yet described is
39t 62p

that of Cruickshank et aZ.34which is shown in Figure 7. The storage cylinder

for the solute + carrier gas is separated from the carrier-gas stream by a
pressure-tight ball valve which can be momentarily opened to allow the solute
into the stream. This method has the advantage that the quantity of gas sample
can be adjusted very simply although this quantity may not be accurately known.
Sample sizes of under 1 pmol are possible for solutes of reasonably high vapour
pressure (over 5 kPa). This is sufficient for the infinitely dilute conditions that
must be satisfied if accurate and meaningful results are to be obtained.6p36
The pressure effect and the small sample size do limit the type of detector
used for medium-high-pressure g.1.c. The commonly used type is the flame-
ionization detector which at high pressures must be linked with a stream splitter.
The Dead-space Volume.-This is the volume of gas that occupies the flow line
from the point of injection to the point of detection. It is determined, at mean
column pressure pJf from
VD = JiuotG, (1 6)
R. W. Moody, Ph.D. Thesis. Bristol, 1973.
A. I. M. Keulemans, ‘Gas Chromatography’, Reinhold, New York, 1959.
6s R. P. W. Scott, ‘Gas Chromatography 1958’, ed. D. H. Desty, Butterworths, London, 1958.
 Activity Coeficients at Infinite Dilution from Gas-Liquid Chromatography 59
where ta is the retention time for an ‘unretained’ solute which may be hydrogen,
nitrogen, or methane depending on the liquid solvent. This property VD is
required in equation (4) to determine the net retention volume VN. It should be
independent of flow rate but in medium-high-pressure g.1.c. a small but
significant dependence is often observed. This is because there are several
components contributing to the dead-space volume, apart from the actual
column. These include the volume between the injector and column, the

A B C D E
r T t
I 1 I I I
Pressure I pi I I P, I pA I
I I I I I
I I I I I
I I I 1 I
Volume I Y1 I 1, I Y3 I V, I
I I I I I
I I I 1 1
I I I I I
Figure 8 Elements of the dead volume. (A) Injector; (B) and ( C ) refer to column ends;
(D) column outlet valve; (E) defector

volume y3 between the column and the outlet valve, and the volume V, between
the outlet valve and the detector. These components are illustrated in Figure 8.
Adjusting these volumes to the pressure poJI the dead-space volume may be
expressed as
VD = (Pi/PoJi)K + & (po/PoJ:)& + (pA/PoJ$& (1 7)
where P A is the pressure of the atmosphere. Thus to eliminate the pressure or
flow-rate dependence of the dead-space volume, the volumes q,V,, and V, must
be reduced to a minimum.
Adsorption Effects.-Although many workers have investigated the solute-
solvent adsorption effect in g . l . ~ . , ~ ~few ~ ~have* - ~taken
~ it into account in
determining activity coefficients. Cruickshank et aZ.29have argued that if there
is significant adsorption of the solute at the solvent-gas interface, then the zero-
pressure retention volume V g of equation (10) is a function of the mass ratio of
solid support to solvent. Extrapolation to zero mass-ratio should then give V k ,
corresponding to the bulk phase solvent-gas partition coefficient of the solute.
It is therefore necessary to measure retention volumes over a range of column
loadings. This has been done for benzene + glycerol using various carrier
gases and two different batches of celite s o l i d - s ~ p p o r t . ~ ~
Unfortunately many of the published activity coefficients from g.1.c. include
measurements done on polar solutes and solvents using only one column. As a
result, it is impossible to know whether adsorption is occurring and these results
must be treated with caution. The results obtained from heavily loaded columns
R. L. Martin, Analyt. Chem., 1963, 35, 116.
66 D. E. Martire, R. L. Pecsok, and J. H. Purnell, Nature, 1964, 203, 1279.
66 R. L. Pecsok, A . de Yllana, and A. Abdul-Karim, Analyt. Chem., 1964, 36, 452.
6’ R. L. Pecsok and B. H. Gump, J. Phys. Chem., 1967,71,2202.
60 Chemical Thermodynamics
(30 per cent solvent/solid support ratio) are however more reliable than the
results obtained from the lightly loaded columns (5 to 10 per cent) because the
bulk effect is greater in the former case and the surface effect is thereby relatively
diminished.
It is therefore very important, in g.1.c. work of this nature, to determine
retention volumes on a number of columns of different solvent/solid support
ratio. This will either point to the non-existence of surface effects or the results
can be used to determine a retention volume at zero mean column pressure,
V‘k, and hence an activity coefficient, which is independent of surface effects.
Ancillary Data Requirements.-Accurately known vapour pressures are required
in this work. For many compounds these are known and good compilations are
a ~ a i I a b l e . ~If~ -reliable
~~ data are not available, very pure samples must be
obtained in order to measure the vapour pressures. A review of the various
methods available has recently been published.59 Accurate measurements are
required because of the sensitivity of the value determined for an activity
coefficient to the value assumed for the vapour pressure. An error of 30 Pa in
a vapour pressure of 30 kPa will result in an error of 0.1 per cent in activity
coefficient.
Solute densities and the second virial coefficients of carrier gases and solutes
are also required. These properties need not be known to the same order of
accuracy as the vapour pressures. Good compilations are available for both
densities 68 and second virial coefficient^.'^ If they are not available, approxi-
mations can be made by a comparison with similar compounds.

4 Experimental Results
General.-Activity coefficients at infinite diIution of many systems have been
measured by g.1.c. and compared with static measurement^.^, 24# 3 s ~72-76 The
agreement has been excellent for those systems where no surface adsorption has
been present. Activity coefficients obtained from g.1.c. must therefore be held
suspect unless care has been taken to show the absence of surface effects.
The results prior to 1967 have been well discussed in reviews by Youngs and
Kobayashi et aL4 Many of these results were obtained from ill-defined sub-
stances such as silicone oils and apiezon greases and will not be discussed here.
n- Alkane Mixtures and Related Systems.-The activity coefficients for these
systems measured by g.1.c. have been the most useful and successful in testing
68 Selected Values of Properties of Chemical Compounds. Manufacturing Chemists Association
Research Project, College Station, Texas, 1974.
6s J. Timmermans, ‘Physico-chemical Constants’, Vols. 1 and 2, Elsevier, Amsterdam, 1950
and 1965.
70 R. R. Dreisbach, ‘Vapour Pressure-Temperature Data for Organic Compounds’, The Dow
Chemical Company, Midland, Michigan, 2nd ed., 1946.
71 J. H. Dymond and E. B. Smith, ‘The Virial Coefficients of Gases’, Clarendon Press, Oxford,
1969.
72 A. J. Ashworth and D. H. Everett, Trans. Furuduy SOC.,1960, 56, 1609.
73 G. F. Freegard and R. Stock, Nature, 1961, 192, 257.
i4 R. L. Pecsok and B. H. Gump, paper presented at: Div. Anal. Chem., 152nd ACS Meeting,
New York, Sept. 12-15, 1966.
75 R. K. Clark and H. H. Schmidt, J . Phys. Chem., 1965, 69, 3682.
i6 H. H. Schmidt, R. K. Clark, and C. F. Gay, Separation Sci., 1967, 2, 65.
Activity Coeficients at InJiPlite Dilution from Gas-Liquid Chromatography 61
solution theories. The repetitive nature of the carbon chains makes these
systems ideal in this respect. Lattice theories in particular have proved very
successful. Alkane systems are fortunately convenient to study because the
properties such as molar volumes, vapour pressures, and virial coefficients have
been well d o ~ u m e n t e d . ~77 ~ - ~ ~ ~
The pioneering work for these systems was done in 1958 by Kwantes and
Rijnders 78 who studied normal and branched chain alkanes in n-octane, n-decane,
n-hexadecane, n-tetracosane, and n-pentatriacontane. The results are consistent
with static mea~urements.~~, Measurements have also been reported by
LittlewoodY8lMartire and P ~ l l a r a and , ~ ~by Pearse and Thornburn,S2but these
do not agree well with static measurements.
The most reliable and comprehensive g.1.c. activity coefficient measurements
for n-alkane systems have been done by the Bristol group 30, 83--85 using
28s 369

medium-high-pressure g.1.c. and taking all carrier-gas and solute imperfections

into account. They have looked at the C4 to C8 n-alkane solutes in to C32
n-alkane solvents. Generally, the results indicate that the smaller the disparity
in carbon number between solute and solvent, the closer is the activity coefficient
to unity. The measured activity coefficients range from 0.930 for heptane +
hexadecane at 303 K to 0.695 for heptane + dotriacontane at 348 K. Activity
coefficients for many alk-l-ene + alkane systems have also been measured by
this group.86
Tewari, Martire, and Sheridan 61 have also measured the activity coefficients
of n-alkane and n-alk-1-ene solutes in long-chain n-hydrocarbon solvents (in
this case, CZ4,C3,,, and C3& Their results substantiate the general trends and
results obtained by the Bristol group.
Letcher and Marsicano 87 have measured the activity coefficients of other
unsaturated C , and c 6 straight-chain hydrocarbon solutes in n-octadecane,
n-octadec-l-ene, n-hexadecane, and n-hexadec-l-ene solvents. The activity
coefficients for these systems do not form simple trends because the various
types of double bonds (terminal, internal, and conjugated) inff uence inter-
molecular interactions in different ways. Definite trends can, however, be seen
in the interactional contribution Tx which is discussed in Section 5.
Branched-chain alkane systems are more difficult to fit into a lattice-theory
picture. NevertheIess, activity coefficients for such systems have been
83

77 M. L. McGlashan and D. J. B. Potter, Proc. Roy. SOC.(A), 1962, 267, 478.

'* A. Kwantes and G. W. A. Rijnders, 'Gas Chromatography 1958', Buttenvorths, London,
1958.
7p J. N. Bronsted and J. Koefoed, Kgl. Danske Videnskab. Selskab, Mat-Phys. Medd., 1946,
17, 22.
8o J. H. van der Waals and J. J. Hermans, Rec. Trav. chim., 1950, 69, 971.
A. B. Littlewood, Analyt. Chem., 1964, 36, 1441.
82 E. C. Pease and S. Thorburn, J. Chromatog., 1967, 30, 344.
83 A. J. B. Cruickshank, B. W, Gainey, and C. L. Young, Trans. Faraday SOC.,1968, 64,
337.
84 C. L. Young, Trans. Faraday SOC.,1968, 64, 1537.
C. P. Hicks and C. L. Young, Trans. Faraday Soc., 1968, 64, 2675.
86 A. J. B. Cruickshank, B. W. Gainey, C. P. Hicks, T. M. Letcher, and C. L. Young, Trans.
Faraday Soc., 1969, 65, 2356.
87 T. M. Letcher and F. Marsicano, J. Chem. Thermodynamics, 1974, 6, 501.
62 Chemical Thermodynamics
Tewari, Sheridan, and Martire88 have extended their n-alkane work to
halogeno-alkane solute + n-alkane solvent. The activity coefficients of these
systems increase with increasing number of halogen atoms. For example, the
activity coefficients at infinite dilution for solutes 1-chlorobutane, 1,2-dichloro-
propane, and 1,2,3-trichloropropane in solvent n-tetracosane at 349 K are 0.875,
1.114, and 1.571 respectively.
In most of the saturated and unsaturated normal hydrocarbon systems studied,
the effect of temperature on the activity coefficients is indistinguishable from the
experimental error for differences in temperature of as much as 20 K.
Systems containing Aromatic Compounds.-The earliest reliable results are again
by Kwantes and Rijnders 78 who determined the activity coefficients of a number
of hydrocarbon solutes in solvent 1,2,4-trichIorobenzene.
+
Reliable activity coefficients have been determined for benzene polyphenyl
systems by Clark and Schmidt 7s and by Schmidt, Clark, and Gray.76 No surface-
adsorption effect was noticed for these systems as the activity coefficients were
independent of the solvent/solid support ratio. It is interesting to note that much
of the work was done with the solvents in their supercooled states. The activity
coefficients at 363 K range from 1.105 in the solvent biphenyl to 0.888 in
m-quinquephenyl. The accuracy of their measurements, however, is limited by
the accuracy of the orthodox analytical g.1.c. apparatus.
The Bristol group have studied the activity coefficients of benzenee9 and
fluorobenzenes in various n-alkane long-chain hydrocarbons with the aim of
testing various solution theories. The activity coefficient of benzene at 323 K
varies from 0.993 in hexadecane to 0.644 in dotriacontane. The trend in the
fluorobenzene systems is for the activity coefficient to increase with increasing
fluorocarbon substitution. For example, in the case of n-octadecane systems at
323 K, the activity coefficient for fluorobenzene is 1.130 while for hexafluoro-
benzene it is 1.945. In these cases, the temperature variation of the activity
coefficients was sufficiently great to warrant the determination of the infinitely
dilute excess enthalpies.
Langer and Purnells1*92have measured the activity coefficients for a large
number of substituted benzenes in benzyldiphenyl, 7,8-benzoquinoline, phenan-
threne, and di-n-propyl tetrachlorophthalate. They also determined the
excess enthalpies from the temperature variation of the activity coefficients.
Other Systems.-Much of the earlier work on activity coefficients from g.1.c.
was done using squalane or dinonylphthalate solvents. Unfortunately, such
systems cannot be investigated by current solution theories. Many of these
earlier measurements were done specifically to test the method against static
results.32.36, 7 % 8 2
Sewell and Stock la have recently measured the activity coefficients for various
chlorinated hydrocarbons in dinonylphthalate and squalane and have critically
assessed earlier work done on these systems.
88 Y . B. Tewari, J. P. Sheridan, and D. E. Martire, J. Phys. Chem., 1970,74, 3263.
a@ B. W. Gainey and C. L. Young, Trans. Faraday SOC.,1968,64, 349.
A. J. B. Cruickshank, B. W. Gainey, and C. L. Young, ‘Gas Chromatography 1968’, ed.
C. L. A. Hatbourn, Institute of Petroleum, London, 1968.
S. H. Langer and J. H. Purnell, J. Phys. Chem., 1963, 67, 263.
S. H. Langer and H. Purnell, J. Phys. Chem., 1966,70, 904.
Activity Coeficients at Infinite Dilution from Gas-Liquid Chromatography 63
Recently Guermouche and Vergnaud D4 measured the activity coefficients
D39

of normal alkanes in squalane at various temperatures in order to determine the

excess thermodynamic properties of mixing. Unfortunately the results, obtained
from conventional atmospheric g.1.c. bi E 180 kPa), do not always agree with
the static and medium-pressure g.1.c. results quoted by Cruickshank et aZ.36
Harris and P r a ~ s n i t z using
, ~ ~ a conventional analytical gas-liquid chromato-
graph, have determined the activity coefficients of 22 polar solutes (including
ketones, acetates, and nitriles) in solvents squalane, squalene, octadecane, and
benzyldiphenyl. The results were used to obtain an approximate scale of relative
Lewis acidities similar to that of Brown.D6 The experimental accuracy is very
difficult to assess because surface effects were not determined. A comparison
with the literature values for three systems D7 indicates reliability within a few
34s

per cent.
Cadogan et aLD8have measured the activity coefficients of C3 to c 6 alcohols
in squalane at different solvent loadings. Their paper is a valuable contribution
to this field because the authors have consciously taken into account the surface
effect by plotting V& against l/V, and extrapolating to infinite liquid phase.
The basic ideas used in this paper were developed in the preceding paper by
Conder et aLBg
Activity coefficients of 1 1 hydrocarbon solutes (including n-alkanes, benzene,
cyclohexane, and cyclohexene) in three heterocyclic polar solvents - furfuryl
alcohol, furfuryl amine, and tetrahydro furfuryl alcohol were determined at
300 K by Kuchhal and Mallik.62The measurements were done using an analytical
gas chromatograph and no attempt was made to look at possible surface effects.
The column packing ratio, however, was high (28 per cent), thus reducing the
possible significance of a surface adsorption effect. Agreement with previous
work was reasonably good.loO The activity coefficients (some having values of
over 20) showed a decrease as the polarity of the solvent decreased.
Pecsok et aLB6and Martire Io1 have specifically investigated solute adsorption
at the gas-liquid interface in g.I.c., by varying the solvent loading. Pecsok et al.
have measured activity coefficients of 22 hydrocarbon and oxy-compounds in
the solvent p,P-thio-dipropionitrile while Martire has measured the activity
coefficients of alcohol solutes in glycerol.
Following on from this work, the Bristol group 29 have made careful measure-
ments on benzene + glycerol-a system which was known to involve surface
adsorption. The method used, which involves determining the extrapolated
retention volume V$ as a function of column loading, should serve as an example
to workers in this field. The activity coefficient at infinite dilution at 323 K was
found to be 2.084 f 0.005.
One of the main needs of chemical engineers today is for accurate parameters
(especially for binary systems) to feed into predicted equations. In evaluating
OS M.-H. Guermouche and J.-M. Vergnaud, J . Chromatog., 1973, 81, 19.
a4 M.-H. Guermouche and J.-M. Vergnaud, J. Chromatog., 1971, 58, 169.
96 H. G. Harris and J. M. Prausnitz, J. Chromatog. Sci.,1969, 7, 685.

96 1. Brown, I. L. Chapman, and G. J. Nicholson, Australian J. Chem., 1968, 21, 1125.

87 S. Evered and F. H. Pollard, J. Chrornutog., 1960, 4, 451.

D. F. Cadogan, J. R. Conder, D. C. Locke, and J. H. Purnell, J . Phys. Chern., 1969,73,708.

J. R. Conder, D. C. Locke, and J. H. Purnell, J . Phys. Chem., 1969, 73, 700.
looC. H. Deal and E. L.Derr, Ind. and Eng. Chern. (Process Res. and Development), 1964,3, 394.
lol D. E. Martire, Analyt. Chem., 1964, 38, 244.
64 Chemical Thermodynamics
these parameters there is an increasing trend to reduce to a minimum the amount
of experimental work and to increase the amount of mathematical work.lo2
Finite-concentration activity coefficients of gas or vapour-liquid systems in
particular are very important in the chemical industry and various theoretical
approaches have been used to determine these quantities from infinite-dilution
activity coefficients determined by g.1.c. Recently Bogeatzes and Tassios lo3have
measured the activity coefficient, at infinite dilution, for sulphur dioxide in
solvents NN-dimethylaniline and NN-dimethylacetamide, in order to test a one-
parameter equation for the determination of finite-concentration activity
coefficients. This work is also interesting from the experimental point of view,
as a technique is described for the determination of activity coefficients of
‘permanent’ gases.
Similar work was done by Hussy and Parcher l8with solutes propanol, acetone,
and hexane in solvents di-n-octylketone, di-n-octylether, and n-heptadecane.
Pecsar and Martin 54 have measured the activity coefficients of homologous
series of alkanes, chloromethanes, alkyl formates, aldehydes, amines, and
alcohols in the solvents H water, 2-pentanone, and 2,3,4-trimethylpentane from
293 to 313 K. The solvents used were rather volatile and pre-saturators were
essential. The work is not of high accuracy: the quoted average error was + 5
per cent but it is probably higher.6 The range of systems measured, however,
makes the work interesting. The results were adequately correlated by relation-
ships proposed by Pierotti et aL1O4~ lo5who have measured activity coefficients of
similar series.
Pecsar and Martin 53 were the first to measure activity coefficients using water
as a solvent. Their work was followed by that of Shaffer and Daubert55 who
measured the activity coefficients of volatile oxygenated organic solutes at
infinite dilution. Their results do not always agree with other measurements.
For example, recently Pemberton and Larkin 56 have measured the activity
coefficients of alcohol solutes in water at 298.15 K, using medium-high-pressure
g.1.c. over the range 270 to 670 kPa. Their results differ from those of Shaffer 55
and Martin54but agree, within a few per cent, with the static measurements of
Hansen,lo6 and the transpiration methods of Burnett lo’ and Butler.lo8 The
apparatus used by Pemberton and Larkin was a fully automated version of the
Bristol group’s a p p a r a t ~ s .Loss
~ ~ of solvent, however, was taken into account
by presaturating the nitrogen carrier gas and weighing the columns before and
after use.
The renewed interest in liquid crystals l o s o led Martire and his group to
determine the thermodynamic properties of a number of hydrocarbon solutes
loa A. K. Hilmi, S. R. M. Ellis, and P. E. Barker, Brit. Chem. Eng., 1970, 15, 1453.
loS A. S. Bogeatzes and D. P. Tassios, Ind. and Eng. Chem. (Process Res. and Development),
1973, 12, 274.
x04 G. J. Pierotti, C. H. Deal, E. L. Derr, and P. E. Porter, J. Amer. Chem. SOC.,1956,78,2989.
lo6 G. J. Pierotti, C. H. Deal, and E. L. Derr, Ind. and Eng. Chem., 1959, 51, 95.
lo6 R. S. Hansen and F. A. Miller, J. Phys. Chem., 1954, 58, 193.
lo’ M. G. Burnett, Analyt. Chem., 1963, 35, 1567.
lo* J. A. V. Butler, Trans. Faraday SOC.,1937, 33, 229.
lo* G. W. Gray, ‘Molecular Structure and the Properties of Liquid Crystals’, Academic Press,
New York, 1962.
110 G. H. Brown and W. G. Shaw, Chem. Rev., 1957,57, 1049,
Activity Coeficients at Infinite Dilution from Gas-Liquid Chromatography 65
in cholesteryl myristate.lll Retention volumes were measured over a range of
temperatures in the smectic, cholesteric, and isotropic-liquid phase regions of
the solvent. From the temperature variation of the activity coefficients other
thermodynamic properties were determined and the results were discussed in
terms of the molecular structure of the liquid-crystal phases and the solute-
solvent interactions. A detailed account of the degree of order in the nematic
mesophase has recently been given by Chow and Martire.l12
Hydrogen bonding is another interest of Martire’s group.l13 Using a con-
ventional analytical g.1.c. apparatus, they measured the activity coefficients of
eight low molecular-weight alcohols in solvents n-heptadecane, di-n-octyl ether,
di-n-octyl ketone at various temperatures. The non-polar solvent was chosen
as a ‘reference’ solvent having the same molecular polarizability, size, and shape
as the electron-donor solvents. The calculated equilibrium constants, standard
enthalpies, and standard entropies of hydrogen-bond formation were within
the ranges normally found for O H - 0 bonds.
Cadogan and Purnell 114 have also investigated hydrogen bonding using g.1.c.
and have measured formation constants and associated thermodynamic quantities
of the C, to C, alcohol solutes in didecyl sebacate. They took the adsorption
effect into account by ‘diluting’ the didecyl sebacate with squalane.
Much of the work on hydrogen bonding discussed above l l 3114 ~ was initiated

by the work of Littlewood and Wi1lm0tt.l~~ They made a detailed study of the
retention of organic solutes in l-dodecanol, lauronitrile, and their mixtures in
squalene. In some cases, equilibrium constants for hydrogen bonding were
calculated. Vernon 11*has recently used g.1.c. to investigate the hydrogen bonds
between ketone solvents and chloroparaffin, alcohol, and amine solutes.
Activity-coefficient determination by g.1.c. in tertiary systems is closely linked
to the study of donor-acceptor complexes; amongst other similarities, two non-
volatile stationary liquids are used. The main difficulty in both cases is
preferential adsorption of the solute on one of the stationary liquids at the gas-
liquid interface. The g.1.c. method of studying complexes through retention
volumes (or activity coefficients) was first mentioned by Chalkley 117 and later
extended by Purnell.lle Because of the dramatic effect on the retention volumes
of unsaturated hydrocarbons of adding silver nitrate to the stationary liquid
solvent, much of the earlier work was done on unsaturated hydrocarbon/Ag+
c o m p l e ~ e s . ~Other
~ ~ - ~systems
~ ~ also investigated in the last few years are alkyl
D. E. Martire, P. A. Blasco, P. F. Carone, L. C. Chow, and H. Vicini, J. Phys. Chem., 1968,
72, 3489.
L. C. Chow and D. E. Martire, Mol. Cryst. Liquid Cryst., 1971, 14, 293.
llS D.E. Martire and P. Riedl, J. Phys. Chem., 1968, 72, 3478.
11‘ D.F. Cadogan and J. H. Purnell, J. Phys. Chem., 1969, 73, 3849.
116 A. B. Littlewood and F. W. Willmott, Analyt. Chem., 1966, 38, 1031.
116 F. Vernon, J. Chromatog., 1971, 63, 249.
117 B. W.Bradford, D. Harvey, and D. E. Chalkley, J . Inst. Petrol., 1955, 41, 80.
118 J. H.Purnell, ‘Gas Chromatography 1966‘, ed. A. B. Littlewood, Institute of Petroleum,
London, 1967.
119 M.A. Muhs and T. F. Weiss, J. Amer. Chem. Soc., 1962,84,4697.
R . J. Cvetanovic, F. J. Duncan, W. E. Falconer, and R. S. Irwin, J. Amer. Chem. Soc.,
1965, 87, 1827.
121 E. Gil-Av and J. Herling, J. Phys. Chem., 1962, 66, 1208.
lZ2 A. N. Genkin and B. 1. Boguslavskaya, Neftekhimiya, 1965, 5, 897.
laa H. Schnecko, Analyt. Chem., 1968,40, 1391.
66 Chemrcaf Thermodynamics
benzenes with 2,4,Wrinitrofluorenone (TNF),l2*,125 aromatic hydrocarbons and
heterocycles with TNF,126substituted anilines with TNF,12’ aromatic compounds
with dinonyl tetrach1orophthaIate,l28olefins with 1,3,5-trinitroben~ene,l~~ metal
chlorocomplexes in fused ~ a l t ~ aromatic , ~ ~ ~heterocyclic
- ~ ~ ~ compounds with
dibutyltetra~hlorophthalate,~~~~ 13* halogenoalkanes with di-n-octyl ether and
di-n-octyl t h i ~ e t h e r ,aromatic
~~~ compounds and tetra-n-b~tylpyromellitate.~~~
Other systems investigated before 1967 have been discussed in a review by
Purnell .13?
Because liquid-liquid chromatography can be considered as an extension of
gas-liquid chromatography, it warrants mention in this context. The two
techniques differ only in the density of the respective eluants and in the degree
of solute-mobile phase interact ions. Unfortunately only activit y-coefficient
ratios can be determined directly, i.e. the ratio of activity coefficient of the solute
at infinite dilution in the mobile phase to the activity coefficient of the solute in
the stationary phase. Therefore, to obtain an activity coefficient at infinite
dilution one of the coefficients must be determined by another method. In many
cases, g.1.c. could supply this, thus making the two techniques complementary.
The 1.l.c. method is rapid and easy to operate and could make valuable con-
tributions to this field. Locke 138-141 has pioneered this work which he reviewed
in 1969.139

5 Theoretical Treatment of Activity Coefficients

General.-The activity coefficients at infinite dilution have been analysed in
many different ways, in attempts to understand the interactions of solute with
solvent. The earliest efforts based on empirical relations for homologous series
of solutes interpreted the results in terms of group interactions, structural effects,
polarity, and electron-donor and -acceptor capacities.2*lo4* lo5- 143 These ideas
142s

have served as foundations for the more sophisticated theoretical interpretations.

The fundamental approximation used in most of the theoretical treatments
considers the activity coefficient to be separable into two parts:
In y = In y,(configuration) + In yl(interaction). (1 8)
lZ4 C. Eon and B. L. Karger, J. Chromatog. Sci., 1972, 10, 140.
la6 D. L. Meen, F. Morris, and J. H. Purnell, J. Chromatog. Sci., 1971, 9, 281.
las A. R. Cooper, C. W. P. Crowne, and P. G. Farrell, Trans. Faraday SOC.,1967, 63,447.
lZ7 A. R. Cooper, C. W. P. Crowne, and P. G. Farrell, J. Chromatog., 1967, 29, 1.
l D 8 A. R. Cooper, C. W. P. Crowne, and P. G. Farrell, Trans. Faraday Soc., 1966, 62, 2725.
R. J. Cvetanovic, F. J. Duncan, and W. E. Falconer, Canad. J. Chem., 1964, 42, 2410.
I3O R. S. Juvet, V. R. Shaw, and M. A. Khan, J. Amer. Chem. SOC.,1969,91, 3788.
C. Pommier, C. Eon, H. Fould, and G. Guiochon, Bull. SOC.chim. France, 1969, 1401.
132 C. Pommier, C. Eon, H. Fould, and G. Guiochon, Compt. rend., 1969, ( C ) , 268, 1553.
133 C. Eon, C. Pommier, and G. Guiochon, Chromatographia, 1971, 4, 235.
136 C. Eon, C. Pommier, and G. Guiochon, Chromatographia, 1971, 4, 241.
136 J. P. Sheridan, D. E. Martire, and Y. B. Tewari, J. Amer. Chem. SOC.,1972, 94, 3294.
J. P. Sheridan, D. E. Martire, and Y. B. Tewari, J. Amer. Chem. SOC.,1972,94, 3298.
la7 J. H. Purnell, Ann. Reu. Phys. Chem., 1967, 18, 81.
138 D. C. Locke, J. Gas Chromatog., 1967, 5, 202.
13s D. C. Locke, ‘Advances in Chromatography’, Vol. 8, ed. J. C. Giddings and R. A. Keller,
Dekker, New York, 1969.
140 D. C. Locke and D. E. Martire, Analyt. Chem., 1967, 39,921.
161 D. C. Locke, J. Chromatog., 1968, 35, 24.
14a I. Brown, J. Chromatog., 1963, 10,284.
143 F. T. Eggertsen and H.S . Knight, Analyt. Chern., 1958,30, 15.
Activity Coefficients at InJinite Dilution from Gas-Liquid Chromatography 67
The configurational and interactional contributions can be considered independent
so long as the interactions are small. The configurational contribution, due to
the mixing of long-chain and short-chain molecules, can be related to simple
lattice properties.6 A simplified version of the theory, originally given
independently by Flory 144 and H ~ g g i n s , gives
l~~
Inyl(cOnfig.) = In[(l - +r)/xJ + (1 - 1/r)$s, (19)
where
4 3 = rx3/(x1 rx3)* (20)
The symbol 43 refers to the volume fraction of the long-chain molecules and r
to the ratio of sites occupied by the long- and short-chain molecules. For the
infinitely dilute condition that applies in g.l.c., equation (1 9) becomes
In y(config.) = In(l/r) + (1 - l/r).(21)
The ratio Y is often taken as the ratio of molar volumes of the long- and short-
chain compounds. For the infinitely dilute condition the interactional con-
tribution is given by 148
In y,(interaction) = x, (22)
where x is the interaction parameter. It is in the interpretation of x that the
various solution theories differ.
The Hildebrand-Scatchard Solubility Parameter The~ry.l~~-Accordingto this
theory the interaction parameter is given by
x = V f ( 6 , - 6J2/RT (23)
where VTj is the molar volume of component 1 and 6, and 6 , are the solubility
parameters which can be calculated from
6 = [(AH‘ - RT)/VT]t, (24)
where AHVis the enthalpy of vaporization of the particular component. This
theory has been used by many workers to discuss activity coefficients from g.1.c.
measurement^,^^^ 83 but it has not been very successful.
759

Weimer 148 has extended this theory, following the work done by van Arkel 149
and Blanks,15*by including the effects due to polar molecules. This theory has
been tested by Harris and Prausnitze5 for polar solutes in solvents octadecane
and benzyldiphenyl and shown to be an improvement on the original theory.
Segment Theories.-The original theory was presented by Tompa 151 and was
later developed by McGlashan et Various forms of it have been very
144 P. J. Flory, J. Chem. Phys., 1942, 10, 51.
145 M. L. Huggins, Ann. New York Acad. Sci., 1942, 1, 43.
1 4 ~E. A. Guggenheim, ‘Mixtures’, Oxford University Press, Oxford, 1952.
14’ J. H. Hildebrand and R. L. Scott, ‘The Solubility of Non-electrolytes’, Reinhold, New York,
1950.
R. F. Weimer and J. M. Prausnitz, Hydrocurbon Process, 1965, 44, 237.
l4@ A. E. van Arkel, Trans. Faraday SOC.,1946,42B, 81.
l50 R. F. Blanks and J. M. Prausnitz, Ind. and Eng. Chem. (Fundamentals), 1964, 3, 1.
161 H. Tompa, Trans. Faraduy SOC.,1949, 45, 101.
lS8 M. L. McGlashan, K. W. Morcom, and A. G. Williamson, Trans. Faraduy Soc., 1961, 57,
601.
68 Chemical Thermodynamics
successful in treating the interaction parameter x, obtained from experimental
activity coefficients at infinite dilution, and the calculated configurational
contribution obtained from equation (21). Applying this theory to mixtures
involving only two types of segments or contact points (A and B) and to data
obtained at infinite dilution, the interactional contribution is given by

where 8, and 8, are the A-type segment fractions of molecules of types 1 and 3,
and W is the interchange energy of the two types of segments defined in terms
of hypothetical molecules containing only one kind of segment. This inter-
change energy should be constant at a given temperature for all mixtures
containing only the two types of segments.
The Bristol group83-85 and Tewari et have been most successful in
applying this theory to n-alkane mixtures. Young26 has extended this theory
to include three types of segments. In this case the interactional parameter x
becomes

where ai, pi, and 6, are the fractions of segments of type A, B, or C of a molecule
i (i = 1 or 3), and W a , WBC,and WAC are the interchange energies of A and B,
B and C, and A and C respectively, and n, is the number of segments on the
smaller molecule. This theory has been applied by the Bristol group and
by Tewari et dS1 to n-alkene + n-alkane and to benzene + n-alkane systems.
Letcher and MarsicanoS7 have recently extended this to include n-alkene +
n-alkene systems. Tewari et ~ 1have . also
~ ~applied it to halogenoalkane solutes
in alkane solvents. This theory becomes unwieldy when more than three segments
are used.
The Perturbation Treatment.-Based on a statistical solution theory of Longuet-
H i g g i n ~ , l154
~ ~a *first-order 'perturbation approach' to the free energy of solution
has been developed by Luckhurst and Martire155and by Tewari et aLS1 This
theory appears useful for interpreting the interaction parameters within a series
of structurally related solute molecules in the same solvent. It is thus ideally
suited for testing data obtained from g.1.c. experiments. The theory involves a
comparison of two solutes - a reference one and a 'perturbed one' in a common
solvent. The interaction parameter x is shown to be linearly related to T:/(Vf)*
where T,C and V;F are the solute critical temperature and the solute molar volume
at temperature 0.6Tf respecively. The term T:/( V f ) &is a measure of the depth of
the interaction potential well.156
Plots of x against Tp/(Vy)* for solute families of n-alkanes, monosubstituted
alkanes, disubstituted alkanes, alk-1-enes, and monosubstituted benzenes in
solvents n-CsoHe2,and n-C3aH,4 gave straight lines which in these
cases gave an average linear correlation coefficient of 0.984.
lS3 H. C. Longuet-Higgins, Proc. Roy Soc., 1951 (A), 205, 247.
lS4 H. C. Longuet-Higgins, Discuss. Faraday Soc., 1953, 15, 73.
ls6 G. R. Luckhurst and D. E. Martire, Trans. Faraday Soc., 1969, 65, 1248.
IL6 A. Kreglewski. J. Phys. Chem., 1967,71,2860; 1968,72, 1897.
Activity Coeficients at Infinite Dilution from Gas-Liquid Chromatography 69
Other Theories that have been applied to y;f3 Results.-The principle of congruence
of Br~rnstedand Koefoed79 was first applied to g.1.c. activity coefficients by
Kwantes and R i j n d e r ~ . They
~ ~ showed fairly good agreement between the
theoretical and experimental results but their experimental work included none
of the refinements used in later work. Young,sa using more accurately deter-
mined activity coefficients, showed that the correlation of experimental with
theoretical results was poor when compared to other theoretical treatments. The
more sophisticated treatment of congruence by Bellemans and Maf,lS7however,
shows better correlation.86
The cell theory of Prigogine168has been applied by Clark and Schmidt76
and by Hicks and Youngs5 to benzene + polyphenyl systems and n-alkane
mixtures respectively. The Flory, Orwoll, and Vrij theory159 has also been
applied by Hicks and Young 85 to n-alkane mixtures.
Funke et aZ.lsOhas used a factor analysis technique to investigate activity
coefficients from g.1.c. In this method, the property in question (activity
coefficients) is resolved into a linear sum of the minimum number of factors
needed to reproduce the experimental results to the required accuracy. Using
these factors, it is possible to predict activity coefficients for systems not yet
measured. Funke et al. have also suggested that it should be possible to interpret
the elements of the matrix into meaningful parameters but this has not yet been
achieved.
Letcher and Marsicano 87 have applied factor analysis to unsaturated hydro-
carbon systems and factorized both the logloy13and Tx results. As expected
the fit for Tx was better than for log,,yl,.
Apart from gaining an insight into intermolecular forces, theoretical studies
have another important and practical use and that is predicting activity
coefficients. Especially in the chemical engineering and possibly analytical
fields such predictions are most useful. The accuracy of prediction varies from
one class of compounds to another. For example, in the case of n-alkane
mixtures by use of the segment theory combined with the configurational contri-
bution theory, the prediction is better than 1 per cent. On the other hand, it is
impossible to predict activity coefficients for systems involving surface adsorption.

6 Mixed Second Virial Coefficients

Little mention has been made of the mixed second virial coefficient B,, which is
often determined simultaneously with the activity coefficient^.^' Although the
medium-high-pressureg.1.c. technique used for determining B12is very restrictive
in regard to the systems that can be studied, the method is important as it is
independent of Bll or Bz2. Apart from the Bristol Dantzler et al.*62
and Vigdergauz and Semkin 163 have used the method for BI2determinations.
lS7 A. Bellemans and P. Mat, Mol. Phys., 1963, 6, 637.
lS8 I. Prigogine, ‘The Molecular Theory of Solutions’, North Holland, Amsterdam, 1957.
16@ P. J. Flory, R. A. Orwoll, and A. Vrij, J . Amer. Chem. SOC..1964, 86, 3515.
loo P. T. Funke, E. R. Malinowski, D. E. Martire, and L. Z . Pollara, Sepuration Sci., 1966,
1, 661.
lor A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Trans. Faraday SOC.,1966,62,2341.
163 E. M. Dantzler, C. M. Knobler, and M. L. Windsor, J . Chromatog., 1968, 32, 433.
los M. Vigdergauz and V. Semkin, J. Chromatog., 1971,58, 95.
70 Chemical Thermodynamics
7 Conclusion
It is only within the past 10 years that the main limitations and problems
associated with this technique have been appreciated or solved. Already new
applications and uses of the values of yy3 determined by the g.1.c. method are
being found,lo3*11* and as more reliable measurements are produced, these
applications will increase.
In conclusion, the Reporter suggests to those measuring activity coefficients
by g.l.c., that the limitations imposed on the nature of the mixtures by the
method be carefully examined, and that special attention be paid to accurate
column loadings and vapour-pressure measurements. In particular, surface
adsorption of the solute at the liquid-gas interface should be investigated by
measuring retention volumes over a range of column loadings. In any event a
number of columns should be made independently and used in measuring activity
coefficients because incorrect column loadings can very easily remain undetected.
The vapour pressure of the pure solute should be accurately known or should
be measured if there is doubt.
Provided that these points are observed, the g.1.c. method for determining
yy3 is capable of 0.2 per cent precisionle4 and the overall accuracy, when com-
pared with static measurement results, is better than 0.5 per cent. The results
can therefore be used with confidence when testing solution theories. This rapid
and easy-to-operate method could have important bearings on theoretical
studies of mixtures, and as more and more data are produced, a rapid growth in
our knowledge of intermolecular forces can be expected.
I wish to acknowledge the helpful discussions I have had personally with Dr J. D.
Bradley and Dr F. Marsicano, and in correspondence with Mr A. J. B.
Cruickshank and Dr C. P. Hicks.
le4 0.WiEarovB, J. Novik, and J. JanBk,J. Chromatog., 1970,51, 3.
3
Experimental Methods for Studying Phase Behaviour
of Mixtures at High Temperatures and Pressures
BY C. L. YOUNG

1 Introduction
The study of critical points and vapour-liquid equilibrium at high pressures and
temperatures has attracted interest since before the turn of the century. The
first studies, beginning with the classic work of Andrews, and greatly influenced
by van der Waals,I were mainly of an exploratory nature. The second phase
was the study of systems of practical importance, such as hydrocarbon and
cryogenic mixtures, and was almost exclusively undertaken by chemical engineers.
Major studies were undertaken by Sage and co-workers,2 Kay and co-wo~kers,~
and Hiza and co-worker~.~
The third phase of development corresponds to a renaissance in the theoretical
understanding of the nature of critical phenomena. On the one hand, the
realization that experimental critical indices are incompatible with those predicted
by simple theories of the van der Waals type has led to a number of very careful
studies of the behaviour of substances in the neighbourhood of a critical point.
On the other hand, new theories of liquid mixtures such as the generalized van
der Waals theories have given impetus to the study of systems suitable for
examining the accuracy and usefulness of these Also, recently,
increased attention has been paid to phase equilibrium of mixtures for which the
critical line between the two pure component critical points is not continuous.8
Such studies involve pressures considerably higher than the critical pressure of
either of the pure components.
In an article of this size it is not possible to discuss in detail all the experi-
mental techniques which have been used for the study of phase behaviour.
At least a hundred different types of equipment have been described in the
literature. Here we shall be primarily concerned with the experimental techniques
used for studying critical properties and vapour-liquid equilibrium at pressures
above 1 MPa.
J. D. Van der Waals, ‘Die Kontinuitat des Gasformigen und flussigen Zustandes’, Barth,
Leipzig, 1900.
a R. H. Olds, H. H. Reamer, B. H. Sage, and W. N. Lacey, Ind. andEng. Chem., 1949,41,474.
S . C . Pak and W. B. Kay, Ind. and Eng. Chem. (Fundamentals), 1972,11,255.
A. G. Duncan and M. J. Hiza, Adu. Cryogenic Engineering, 1970, 15, 42.
E. J. Partington, R. S. Rowlinson, and J. F. Weston, Trans. Faraday Soc., 1960, 56, 479.
M. T. Ratzsch, Z . phys. Chem. (Leipzig), 1970,243,212.
’ R. L. Scott, Ber. BunsengesellschaftPhys. Chem., 1972,76,296.
E. U. Frank, in ‘Physical Chemistry, an Advanced Treatise’, ed. W. Jost, Academic Press,
New York, 1971, vol. 1, p. 367.
I1
72 Chemical Thermodynamics
2 Critical Properties
It is convenient to consider the methods used to measure the critical properties
of pure substances before those methods used for mixtures. We follow this
procedure firstly because the critical properties of pure substances are needed in
any theoretical discussion of the results for mixtures. Secondly, many of the
techniques used for pure substances can, with minor modification, be used for
mixtures.
The critical point of a pure substance can be defined in terms of two different
types of observation, (a) the shape of the p , V isotherms or (b) the visual
observation of the appearance or disappearance of the phase boundary. The
method based on the first definition depends on establishing the highest tempera-
ture for which the isotherm satisfies the condition @p/a V ) , = 0 and determining
that point on the isotherm for which this condition is valid. At the critical
point
p > 0, ( a p p v - ) , = 0, (aZp/av2), = 0.
This method, originally used by Andrewsg in his famous experiments on
carbon dioxide in 1872, has been used recently,l* but very precise and elaborate
apparatus is needed to obtain accurate results. Inspection of pressure-density
graphs shows that it is very difficult to determine exactly which of the isotherms
corresponds to the critical isotherm even if very precise p , V,T measurements are
available. Hence critical data are usually obtained from such studies as a spin-
off, the major aim being the determination of the p,V,T behaviour of the
substances.
The experimental difficulty of determining critical behaviour from the shape
of the isotherm has led to a preference for determining the critical temperature
and pressure visually and the critical density by extrapolating liquid-vapour
equilibrium measurements to the critical temperature.
Critical Temperature of Pure Substances.-The determination of the critical
temperature of a substance is often a relatively simple experiment provided an
accuracy of 0.1 K is sufficient. The most commonly used method is the
observation of the appearance or disappearance of the meniscus when a sealed
tube containing some of the substance is slowly cooled or heated. The tube
must be filled so that the overall density of contents of the tube is approximately
equal to the critical density.
When a sealed tube containing a liquid and its vapour is slowly heated
uniformly, one of three things may happen. If the overall density is less than
the critical, the meniscus will fall and eventually all the liquid will evaporate.
If the overall density is greater than the critical, the meniscus will rise and
eventually all the vapour will condense. If the overall density is equal to the
critical the meniscus will rise or fall slowly until it is near the centre of the tube,
where it will become flat, fainter, and will eventually vanish at the critical
temperature. For measurements of moderate precision (0.1 K) there is no great
need for the overall density to be exactly the critical density. Owing to the
large compressibility of both liquid and gaseous phases near the critical point,
O T.Andrews, Phil. Trans., 1869,159, 1876.
lo D. R. Douslin, R. H. Harrison, and R. T.Moore, J. Phys. Chem., 1967,71, 3477.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 73
large density gradients develop and the meniscus will move up or down and
vanish at a height at which the local density is equal to the critical density. For
more precise measurements of the critical temperature the density must be
within 1 per cent of the critical density, the tube contents must be well stirred,
and the experimental tube must be short (less than 2cm). The final heating
(or cooling) through the critical temperature must be carried out very slowly.
Usually the reappearance of the meniscus on slowly cooling the system is some-
what ‘sharper’ l1 than the disappearance of the meniscus on heating.
The major requirement for measuring critical temperature by this technique
is a suitable furnace (or cryostat) in which (a) the experimental tube can be
heated or cooled without the introduction of significant temperature gradients,
and (b) accidental breakage of the tube does not constitute a major safety hazard.
Basically, there are three different types of furnace or cryostat which have been
used, vapour baths, liquid baths, and solid block-type furnaces.
Simple water or oil baths may be used between 280 and 480 K, provided that
the bath is adequately stirred. At higher temperatures the operation can become
hazardous if an experimental tube breaks unless suitable precautions are taken.
For measurements below room temperature, some form of Dewar vessel is
needed. Suitable coolant liquids are the non-flammable freons or hydrocarbons.
Liquid hydrocarbons have often been used as coolants in critical temperature
study, but their flammable nature is a major drawback.
Vapour baths may be used at temperatures up to about 600 K. They consist
of a boiler flask connected to a vapour condenser in which the sample tube is
suspended.12 The temperature is controlled by regulating the pressure at which
the vapour condenses. If a pure liquid is used in the boiler this type of equipment
gives good temperature control but is fairly complex to operate and can be
hazardous if a sample tube breaks.
The safest and probably the most satisfactory equipment for measurements
above 350K is the solid block furnace. A modified version of the aluminium
block furnace described by Powell et aZ.ls is illustrated in Figure 1. The furnace
consists of an aluminium cylinder 300 mm long and 160 mm in diameter. The
sample tubes can be inserted into a vertical hole drilled down the axis of the
aluminium cylinder and are viewed through a small horizontal hole. The tubes
are illuminated from behind by a small focused light. The aluminium block is
insulated by a layer of vermiculite (about 100 mm thick) and the whole enclosed
in a large asbestos cylinder. The original version was heated by externally
wound Nichrome wire. It has been found, however, that electric elements cast
in the aluminium block are more convenient, particularly when the furnace is to
be operated above 600 K for long periods.
Critical Pressure of Pure Substances.-The critical pressure may be determined
in several ways. As already indicated, the critical pressure may be determined
directly from p, Y isotherms lo but this method is not particularly accurate
because of the difficulty in establishing the critical isotherm.

l1 A. P. Kudchadker, G. H. Alani, and B. J. Zwolinski, Chem. Rev., 1968, 67, 659.

m W. B. Kay and G. M. Rambosek, Ind. and Eng. Chem., 1953,45,221.
la R. J. Powell, F. L. Swinton, and C.L. Young, J. Chem. Thermodynamics, 1970,2, 105.
74 Chemical Thermodynamics
Kudchadker et aZ.ll have pointed out that experimental methods which enable
visual observation of the critical temperature, the determination of the critical
density by the law of rectilinear diameters (see below), and subsequently the
direct determination of the corresponding critical pressure, should be preferred
over other methods.

Figure 1 Aluminium block furnace currently used b y the Reporter

The most commonly used techniques involve measurement of vapour pressure

up to the critical temperature. The sample is confined over mercury in a tube
sealed at the upper end. The mercury can be pressurized using a dead-weight
pressure tester and the pressure on the sample varied at will. The critical pressure
can either be obtained by measuring the vapour pressure as a function of
temperature up to a kelvin or so below the critical point and extrapolating to the
critical temperature, or by measurement of the pressure at which the meniscus
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 75
disappears at the critical temperature. In Ambrose and co-workers’ method,14
only the upper part of the experimental tube is maintained at an elevated
temperature. The mercury and some of the sample are kept at room temperature
and therefore it is not necessary to make any corrections for the partial vapour
pressure of mercury. This technique is suitable for pure substances but not for
mixtures, as the temperature gradients in the experimental tube would cause
concentration gradients to be set up in mixtures. The apparatus used by Kay3
and Young l5for mixtures is suitable for pure components and is discussed under
the critical properties of mixtures. However, the technique involves having hot
mercury in contact with the experimental sample and it is necessary to make
corrections for the partial vapour pressure of mercury.
Critical Volume of Pure Components.-The critical volume is the most difficult
of the three critical constants to measure and consequently critical volumes are
rarely measured to an accuracy of better than 0.5 per cent. The best method is
to measure orthobaric liquid and gas densities up to within a kelvin or so of the
critical temperature. The critical density is obtained using the law of rectilinear
diameters, i.e. by extrapolating the mean of these densities to the critical tempera-
ture. This technique was originally proposed by Cailletet and Mathias16 and
has been investigated carefully by Schneider and co-workers l7 and discussed in
detail by RowIinson.l*
Several techniques have been used to determine the orthobaric densities of the
liquid and gas, The simplest method is that proposed by van Eck lDand recently
used by the Reporter.15 A series of heavy-walled tubes are partially filled with
varying but known amounts of the substance under investigation. Each is heated
slowly in turn until the meniscus disappears out of either the top or bottom of
the tube, i.e. the vapour phase completely condenses or the liquid gas completely
evaporates. The temperature at which the tube becomes completely filled with
vapour or liquid is noted. The volume of each tube is then found by weighing
it in water and then cutting the tube open, emptying it and weighing the pieces
in water. This method enables the densities of liquid and vapour to be calculated
over a series of temperatures. This method is simple but time-consuming, and
requires considerable experimental care.
A superior method, in terms of speed, is the paired-tube method used by
Ambrose and c o - w ~ r k e r s . The
~ ~ method is suitable only for pure substances
and cannot be used for mixtures. The method involves sealing varying amounts
of the substance, as a liquid, in a series of thick-walled tubes, so that between
15 and 65 per cent of the internal volume of the tubes is occupied by liquid. The
volumes of the tubes are calculated from a knowledge of the bore length and the
dimensions of the sealed end. The fraction of this volume which is occupied by
liquid is measured at room temperature and from about 120 K below the critical
temperature to between 10 and 20 K below the critical temperature.
I4 D. Ambrose and R. Townsend, J . Chem. SOC.,1963, 3614.
Ib C. L. Young, J. Chem. Thermodynamics, 1972, 4, 65.
Is L. Caillet and E. Mathais, Compt. rend., 1886, 102, 1202.
l7 M. A. Weinberger and W. G. Schneider, Canad. J. Chem., 1952, 30, 422.
lS J. S. Rowlinson, ‘Liquid and Liquid Mixtures’, 2nd edn., Butterworths, London, 1969,
pp. 90-94.
l9 H. van Eck, Z . Physik., 1939, 40,13.
76 Chemical Thermodynamics
The theory of the method is as follows. Consider the effect of temperature
upon the proportion of the tube occupied by liquid. Let F be the fraction
occupied by liquid, V the internal volume of the tube, m the mass of substance
present, pv and p1 are the densities of vapour and liquid respectively, and a the
cubical coefficient of expansion of glass, and let the subscripts 0 and Trefer to
room temperature and temperature T respectively.
The volume of liquid = VFo (at room temperature)
= VFT(l + aT) (at temperature T),
m = m o p 1 0 + (1 - F0)pVol
= VfFTPlT + (1 - FT)PVTI[l + 4.
If pvo can be neglected

A plot of Fo against FT is a straight line for each temperature and if the density
of the liquid at room temperature is known, hTand plT can be calculated from
the gradient and the intercept at FT = 0.
A third technique, which is even quicker but requires slightly more sophisti-
cated equipment, has been used by several ~ o r k e r s . ~ ~Kayzf
- ~ * measured the
vapour and liquid volumes of a fixed mass of sample at the saturated vapour
pressure as a function of temperature. The critical volume was obtained by
extrapolation of these volumes to the critical point. This method is not very
accurate and also suffers from the disadvantage that the sample is in contact with
mercury.
Critical Properties of Mixtures.-The critical point of a mixture is not governed
by the same criticality conditions as a pure substance but by the conditions

In general, except for azeotropic mixtures,

A simple pressure-temperature projection of the pressure-temperature-

composition diagram for a mixture is given in Figure 2. It is necessary to define
the terms bubble point, dew point, maxcondentherm, and maxcondenbar. A
bubble point is a state of liquid mixture at which, if the pressure is decreased
slightly, a second phase, a vapour, appears. Similarly, a dew point is a state of a
vapour at which, if the pressure is increased slightly, a liquid phase appears.
The dew point locus and the bubble point locus are continuous curves meeting
at the critical point. The maxcondentherm and maxcondenbar are the maximum
temperature and maximum pressure respectively on the bubble point-dew point
2o D. Cook, Trans. Faraday SOC.,1953, 49. 716.
81 W. B. Kay and F. M. Warzel. Amer. Inst. Chem. Engineers J., 1958, 4, 296.
a2 G. Kaminishi and T. Toriumi, Rev. Phys. Chern. Japan, 1968, 38, 79.
23 W. B. Kay, Ind. and Eng. Chem., 1938,30,459.
24 H.Lu, D. M. Newitt, and M. Ruhemann, Proc. Roy. Soc., 1941, A , 178, 506.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 77

P PA'

r
Figure 2 Bubble-point dew-point locus of a mixture of constant composition. AA' and
BB' are paths along which retrograde condensation and retrograde evaporation take
place respectively; C , critical point ; D, maxcondentherm; and E, maxcondenbar

rI I1
temperature
readout

I
U
u cold

n
junction

4
surge tank
furnace

x
[FI
transformer
exwrimental
7
vent
' t u b e d to Hg
manometer

cornpress or block
Figure 3 Block diagram of Pak and Kay's apparatus for determining the critical
temperatures and critical pressures of mixtures
78 Chemical Thermodynamics
locus. The maxcondenbar, maxcondentherm, and the critical point of a mixture
are always different points, except when the mixture is an azeotrope. Com-
pression of the mixture at temperatures between those at C and D will cause
two phases to appear on crossing the dew curves; the volume of the denser phase
increases and then decreases and finally disappears on crossing the dew curve
again. This phenomenon is called retrograde evaporation. Similarly on heating
the mixture at a pressure between C and D two phases appear on crossing the
bubble curve and the volume of the lighter phase increases and then decreases
and finally disappears on crossing the bubble curve.
Figure 2 is the simplest pressure-temperature projection. More complicated
projections have been discussed elsewhere,** 25 and are reviewed in Chapter 4.
Critical temperatures of mixtures are usually determined in the same experi-
ment as critical pressures. However, some workers 26-32 have used the
5p 6p

sealed-tube technique discussed previously. The method is completely analogous

to the method used for pure substances. Critical temperatures and critical
pressures can be determined in the same experiment if an open-ended tube
method is used.
An apparatus suitable for measuring the criticaI temperatures and pressures
of mixtures and dew point-bubble point loci between 400 and 720 K has been
described by Pak and Kay and by Young.l**33 Figure 3 is a schematic diagram
of Kay’s apparatus. The sample of the mixture was confined over mercury in
the sealed end of a thick-walled tube. The tube was connected to a source of
high-pressure nitrogen gas, the pressure of which was measured on a high-
precision Bourdon gauge. The sample tube was heated in an electric furnace as
shown in Figure 4. This consisted of a copper bar about 40 mm in diameter and
300 mm long with a 12 mm hole drilled through the axis of the cylinder to take
the sample tube. One slit was cut right through the bar parallel to its axis and a
slot was cut at right angles to the slit. A beam of light could be passed through
the slit and sample and the sample could be observed at right angles in the slot.
Using this technique the intense light scattering, characteristic of the critical
state, could easily be detected. The copper block itself was enclosed in a furnace.
The temperature of the furnace could be controlled automatically or manually.
The complete furnace couId be raised or lowered over the experimental
tube. The temperature of the sample was measured using an iron-to-constantan
thermocouple located in a small well in the top of the experimental tube.
The sample could be agitated by a small stirrer which consisted of a magnetic
stainless-steel rod, of slightly smaller diameter than the bore of the glass capillary
tube. A small length of drill rod was spot welded on to the steel rod. The
stainless-steel rod was located in the tube just above the compressor block,
with the drill rod extending into the hydrocarbon sample at the top of the tube.
25 W. B. Kay, Accounts Chem. Res., 1968,1, 344.
86 M. T. Ratzsch and G. Strauch, Z . phys. Chem. (Leipzig), 1972,249,243.
27 C . L. Young, J.C.S. Faraday ZZ, 1972, 68, 452 and 580.
28 C. P. Hicks and C. L. Young, J. Chem. Thermodynamics, 1971,3, 899.
aD C. L. Young, Austral. J . Chem., 1972, 25, 1625.
8o C. P. Hicks and C. L. Young, Trans. Faraday Soc., 1971, 67, 1598 and 1605.
81 I. W. Jones and J. S. Rowlinson, Trans. Faraday Soc., 1963, 59, 1702.
8a D. E. L. Dyke, J. S. Rowlinson, and R. Thacker, Trans. Faraday SOC.,1959, 55,903.
83 C. P. Hicks and C. L. Young, J. C. S. Farady I, 1976, 122.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 79
The stirrer was operated by two solenoid coils energized alternatively by a
manually operated reversing switch.
About 0.15 cm3 of the (air-saturated) mixture under study was transferred to
the bottom of the inverted experimental tube through a very fine capillary

controlling
thermocouple

-
insulat i,ng
ceramic pluc

thermocouple.

viewing slit

experimental
tube-
i
copper black-

insulatincl
ceramic plug

Figure 4 Diagram of the critical point furnace used by Pak and Kay

funnel freshly made by heating and drawing an 8 mm glass tube. A small amount
of mercury was introduced into the tube and trapped air removed from between
the sample and mercury with a very fine capillary tube. The contents of the tube
were frozen in liquid nitrogen. The stirrer was inserted into position and the
tube evacuated, the sample and mercury being still frozen. The rest of the
sample tube was filled with mercury and the experimental tube assembled in
the compressor block. The furnace was preheated to near the critical tempera-
ture. The furnace was lowered on to the tube and the pressure on the sample
was varied so that the meniscus remained near the middle of the section occupied
by the sample.
80 Chemical Thermodynamics
The apparatus can be also used to estalilish the dew point-bubble point locus.
The great advantage of Kay's apparatus over most types of apparatus for
determining critical properties is that it is relatively simple to operate. However,
it does have the disadvantage that the samples are not degassed. This could
lead to errors of 0.01 to 0.06 MPa in the critical pressure and up to 1 K in the
critical temperature.

-
variac
thermocouple heating
element
7

digital
voltmeter
i

screw
' press
A experimental

determining the

Young33 has designed a slightly different sample tube and filling procedure
which enables the sample to be degassed thoroughly. However, the technique
has the disadvantage that the composition of the mixture is not known and has
to be estimated from critical temperature-composition data obtained in a
separate experiment. The apparatus is similar to that described by Kay, the
major differences being that the pressure is measured using a dead-weight tester
and a simpler experimental tube is used. The dead-weight tester enables the
equipment to be operated without having any high-pressure gas present (except,
of course, that under investigation) and is consequently less hazardous. A
diagrammatic representation of the apparatus is given in Figure 5.
There are three versions of the experimental tube.l4*33s s4 The first, which has
been used only for pure substances l4 consists of a well annealed thick-walled
J tube as illustrated in Figure 6. The bottom portion of the J tube is made of
8 mm internal diameter pyrex tubing with walls about 2.5 mm thick. The
remainder of the tube is made of 2.5 mm internal diameter pyrex tubing with
walls about 2.5 mm thick. To fill this tube about 0.5 to 0.8 cm3 of sample was
introduced into the tip of the J-tube. The sample was frozen in liquid nitrogen,
3* C. L. Young, unpublished work.
Studying Phase Behaviour of Mixtiires at High Temperatures and Pressures 81
evacuated, and then slowly thawed. The process was repeated until no bubbles
appeared in the liquid on thawing. Some of the sample was lost by this process.
If the total sample volume was more than 0.3 cm3 some sample was pumped off.
The sample was then frozen in liquid nitrogen and mercury added to the J-tube
in vacuo so that mercury filled the tube except for about 10 cm3. Air was then
admitted to the vacuum line with the J-tube held in an upright position so that

threaded
retaining
collar-

split -
collar

dive of rectangular cross

section forming part of glass- to-metal-to-

glass cone

thick-walled

Figure 6 Details of experimenfal J-tubes used by the Reporter

the degassed liquid was confined over the mercury. The J-tube was then attached
to the hydraulic fluid line.
The second and later version of the J-tube, which was used for mixtures was
similar except that a demountable glass-to-metal-to-glass seal was incorporated
in the sealed limb of J - t ~ b e The
. ~ ~glass-to-metal-to-glassseal was removed from
the sealed capillary and the appropriate mixture placed in the sealed tip. The
demountable coupling was reassembled and the degassing procedure was then
identical to that used in the first version.
The third version used slightly thicker capillary tubing for the sealed limb and
the bottom portion of the J-tube was replaced by stainless-steel tubes, designed
so that a larger change in the level of the mercury in the glass-capillary tube
caused a negligible change in the level of the mercury in the stainless-steel tube.
This enabled the difference in the height of the mercury level in the two limbs
of the J-tube to be measured to an accuracy of & 2 mm. This difference was
needed to calculate the hydrostatic pressure between the dead-weight tester
datum level and the position at which the critical phenomenon was observed.
The third version J-tube was operated up to 15 MPa.
82 Chemical Thermodynamics
Because the law of rectilinear diameters does not hold for mixtures, the
critical volumes of mixtures are very difficult to determine accurately. However,
if the maxcondentherm and maxcondenbar are fairly similar (not more than
1 K different in their temperatures) use of the law appears to lead to reasonable
values of the critical volume. The law cannot be recast by substituting pressure
in place of temperature as suggested by Kaminiski and Toriumi.12 The only
reliable method of obtaining the critical volume of a mixture is to extrapolate
the bubble-point volume and dew-point volume to the critical point.
Factors Affecting the Accuracy of the Measurement.-Mercury Vapour. Often
mercury is used as a pressure transferring medium in apparatus used for deter-
mining critical properties. If a sample is confined over mercury at temperatures
for which the vapour pressure is appreciable the measured pressure may be
considerably different from the pressure which would be measured in the absence
of mercury. Relatively little work has been undertaken on the effect of mercury
having a catalytic effect on slow reactions which take place in the sample.
Secondly, the partial pressure of mercury may differ considerably from the
vapour pressure of mercury at the same temperature. The usual procedure for
‘correcting’ for the mercury-vapour effect is to subtract the vapour pressure of
mercury from the experimental measured pressure. Jepson and Rowlinson 37 359

have questioned this procedure and Kay and co-workers 3s 38 have made measure-
369

ments with the sample confined over mercury and then over gallium. The vapour
pressure of gallium at the critical temperatures of hydrocarbons is negligible.
Kay and Pak3 suggested that for hydrocarbon mixtures the partial pressure of
mercury could be calculated from the equation*
log,,(p~/lbfi r 2 ) = 5.93075 - 3036.05(K/T),
where pm is the partial pressure of mercury. The partial pressures calculated
from this equation are about 85 to 90 per cent of the vapour pressure of mercury
at the same temperature. Kay and Pak3 pointed out that the mercury-vapour
effect would be less if the sample was not thoroughly stirred. The effect of
mercury on the critical temperature is small by comparison, amounting to
between 0.1 and 0.6 K.
Efect of Dissolved Air. The presence of dissolved air in a sample can give rise
to two effects (a) a reaction between the sample and the dissolved air, (b) a
‘multicomponent mixture’ effect. Kay and co-workers 38 have shown that, for
31

substances for which it is expected that the second effect is the most important,
usually the critical pressure is most affected and often the difference between
degassed and air-saturated samples is 0.02 to 0.03 MPa. The critical temperatures
of air-saturated substances were found to differ by up to 0.9 K from those of
degassed substances. However, many substances and mixtures react with
dissolved air and in such cases often the observed critical properties vary with
time.
86 W. B. Jepson and J. S. Rowlinson, J . Chem. Phys., 1955,23, 1599.
36 H. S. Rosenberg and W. B. Kay, J. Phys. Chem., 1974,78, 186.
37 W. B. Jepson, M. J. Richardson, and J. S. Rowlinson, Trans. Faraday SOC.,1957, 53, 1586.
38 W. B. Kay and D. W. Hissong, Proc. Amer. Petr. Inst., Refining Div., 1967, 41, 653.
W. B. Kay, personal communication.
* There is a typographical error in ref. 3, equation 1, which should read as above.a8
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 83
Decomposition of the Sample. Many substances decompose near their critical
points and a large number of mixtures which do not react at room temperature,
react at the critical point. It is obviously important that decomposition or
reaction of the components in a mixture is kept to a minimum and the extent
of reaction is known. A simple check is to measure the critical properties of
both of the pure components and one mixture as a function of the time the
sample has been kept near the critical point.40

3 High-pressure Vapour-Liquid Equilibrium

There are two principal methods of obtaining high-pressure vapour-liquid
equilibrium data, the ‘flow’ method and ‘static’ method. These methods are
extensions of the principles used at, or near, atmospheric pressure. However,
the details and materials of construction of the apparatus are often very different.
We will not be concerned with the details of standard high-pressure equipment
as reviews 42 and books 43n 44 on this subject are available.
41p

Flow Methods.-There are two types of ‘flow’ measurements. In the first, the
‘single-pass’method, a sample of pure gaseous component at a specified pressure
is passed through a liquid phase at a specified temperature. Some of the gaseous
component dissolves in the liquid phase and after a time, which in well designed
equipment is less than 15 min, the effluent gas and the liquid phase compositions
are essentially in equilibrium. After equilibrium is reached samples of the liquid
and gas phase are withdrawn and analysed. Usually the pressure is kept constant
during sampling. The major disadvantages of this technique are the difficulty
in ensuring that equilibrium is essentially reached and the fact that the liquid-
component sample is constantly being removed.
The other type of flow system is the vapour recirculation method. In this
method a two-phase mixture is placed in a vessel at a specific temperature.
Vapour is then withdrawn from the cell and recirculated through the liquid.
This contact between vapour and liquid provides a fairly rapid approach to
equilibrium. Again after equilibrium is reached samples of gas and liquid are
withdrawn and analysed. Some provision is usually made to keep the pressure
constant during sampling.
In practice the single-pass flow system is slightly easier to operate than the
recirculating vapour system but suffers from three major disadvantages. Firstly,
it is difficult to establish if equilibrium has been established in one pass. Secondly,
the technique can use considerable quantities of the gaseous component. While
this is not important for the common gases, the method cannot be used for less
common gases like xenon. Thirdly, the systems it is possible to study by this
method are restricted to those in which the partial pressure of the liquid is low
(usually below 0.01 MPa). Consequently the method is not suitable for deter-
minations near the critical point of the mixture. The recirculating method
40 C. P. Hicks and C. L. Young, Chern. Rev., 1975, 75, 119.
41 D. M. Newitt, ‘The Design of High Pressure Plant and the Properties of Fluids at High
Pressure’, Part 1, Oxford University Press, 1940, p. 1.
48 ‘High Pressure Physics and Chemistry’, ed. R. S. Bradley, Academic Press, London, 1963,
vols. 1 and 2.
43 P. W. Bridgeman, ‘Collected Experimental Papers’, Harvard University Press, Cambridge,
Mass., U.S.A., 1964, Vols. 1-VII.
44 S. D. Hamann, ‘Physico-chemicalEffects of Pressure’, Butterworths, London, 1957.
84 Chemical Thermodynamics
does not suffer from these restrictions, although it is not suitable very near the
critical point of the mixture. Some workers have used a single-pass method at
temperatures and pressures when the partial pressure of the liquid is low, and a

thermocouple sheath
-4

gas
diffuser-

sampling L e n t e r ing stream

Figure 7 High-pressure low-temperature uapour-liquid equilibrium cell used by Hiza

and co- workers

recirculating technique at high temperatures. The same equilibrium cell is often

used in both single-pass and recirculation methods.
Hiza, Herring, and co-workers 45-52 have discussed several single-pass and
recirculation methods for studying vapour-liquid and vapour-solid equilibrium
I6 M.J. Hi=, C. K. Heck, and A. J. Kidney, Ado. Cryogenic Engineering, 1968, 13, 343.
16 M. J. Hiza and R. N. Herring, Adv. cryogenic Engineering, 1963, 8, 158.
M.J. Hiza and A. J. Kidney, Adv. Cryogenic Engineering, 1966,11, 338.
48 R. N. Herring and P. L. Barrick, Ado. Cryogenic Engineering, 1965,10, 151.
4s C. K. Heck and P. L. Barrick, Ado. Cryogenic Engineering, 1967, 12, 714.
C. K. Heck and P. L. Bamck, Adv. Cryogenic Engineering, 1966, 11, 349.
61 M. J. Him and A. G. Duncan, Ado. Cryogenic Engineering, 1969,14, 30.
M. J. Hiza and R. N. Hemng, Adv. Cryogenic Engineering, 1964,10, 182.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 85
for a range of systems at cryogenic temperatures. A simplified diagram of their
vapour-liquid equilibrium cell is given in Figure 7.
Initially the less volatile component is condensed into the cell, which is then
cooled to the operating temperature. For single-pass operation pure volatile
component is then passed into the cell at some predetermined pressure and the
equilibrium gas phase leaving the cell passes through a pressure-reducing valve
and is continuously analysed using a flame-ionization detector. Liquid-phase

and vacuum pump

Figure 8 Block diagram offlow system used by Hiza and co-workers for the recirculating
vapour technique of studying vapour-liquid equilibrium at high pressures

sample can be withdrawn, at constant pressure, from the bottom of the cell.
The cell is fitted with a gas diffuser and liquid entrainment separator. The gas
diffuser at the bottom of the inside of the cell consisted of lightly packed copper
shaving. This diffuser created a large number of small gaseous bubbles, enhancing
mixing and hastening mass transfer between the two phases. The liquid-
entrainment separator was a cylindrical capsule filled with tightly packed copper
shavings and effectively removed any droplets of liquid mechanically carried
along with the gas. The cell was incorporated into an accurately controlled
cryostat, the details of which are discussed by Herring and B a r r i ~ k . ~ ~
In the recirculated vapour method these workers incorporated the cell into a
high-pressure flow system as illustrated in Figure 8. The gaseous phase was
pumped from the cell through the vapour sampling loop (or the vapour by-pass
loop) through a ballast volume and back into the cell. The gas-phase composition
was determined by gas chromatographic analysis. The provision of a vapour-
sampling loop by-pass made it possible to sample the vapour phase without
disturbing the: recirculating flow. Liquid samples were removed from the bottom
of the cell through a small capillary tube and analysed using gas chromatography.
Care was taken to ensure that the liquid sample obtained was a representative
86 Chemical Thermodynamics
sample from the cell by purging the capillary tube and chromatographic sampling
valve several times. The large ballast volume was incorporated in the system SO
that the pressure in the system would remain nearly constant even when liquid
samples were removed for analysis. The apparatus could be operated up to
pressures of 15 MPa.
A multipurpose cell which can be used in a recirculating or one-shot apparatus
for either solid-vapour or liquid-vapour equilibrium has been described by
Duncan and H i ~ a . ~
Kobayashi and co-workers 5 3 ~64 have designed a single-pass flow system for
the determination of the dew-point loci of methane + heavier hydrocarbons at
low temperatures and elevated pressures. The basic concept of the apparatus
is similar to that of Hiza and Herring 46 described above. The effluent gas in this
case was analysed using a thermal-conductivity cell, a flame-ionization detector
being unsuitable since it responds to the methane and the heavier hydrocarbons.
However, in order to calibrate the thermal-conductivity detector a specially
constructed high-precision micropump was used to introduce a suitable con-
centration of the heavier hydrocarbon into the methane gas stream. Dilute
mixtures as low as mole fraction of the heavier component in methane
could be accurately detected by this method. Kobayashi and co-workers 64-58
have described several modifications to their original apparatus including a
conversion to enable recirculation of the vapour phase.
Tully, De Vaney, and Rhodes59 have developed a windowed cell made of a
beryllium copper alloy for studying cryogenic systems. The basic concept is
again similar to that used in the flow-version of Hiza’s 46 apparatus. Liquid and
vapour samples were analysed using a high-precision thermal-conductivity
analyser. The method of obtaining samples for analysis is, however, somewhat
different from that used by Hiza and Herring.46 The system is allowed to
equilibrate with the recirculating pump operating. At the end of the equilibration
time the recirculating pump is isolated from the rest of the flow line and its
contents analysed. The liquid samples are then withdrawn through a fine
hypodermic tube at the same time as a small amount of gas is allowed into the
cell so that the cell pressure remains constant.
Kirk and Ziegler 6oe have developed a single-pass apparatus similar in
concept to that used by Dodge and Dunbar 62 in the original high-pressure flow
vapour-liquid apparatus. The cell consisted of a double-compartment vessel as
illustrated in the simplified diagram in Figure 9. Gas was passed in through the
base of the cell through a long spiral path in the lower compartment. The gas
then passed into the upper compartment which was partially filled with copper
m R. J. J. Chen, W. E. A. Ruska, P. S. Chappelear, and R. Kobayashi, Adv. Cryogenic Engfneer-
ing, 1973, 18, 202.
rid A. R. Price and R. Kobayashi, J. Chem. and Eng. Data, 1959,4,40.
bs I. Wichterle and R. Kobayashi, J. Chem. and Eng. Data, 1972, 17, 4.
l 6 D. G. Elliot, R. J. J. Chen, P. S. Chappelear, and R. Kobayashi, J. Chem. and Eng. Data,
1974, 19, 71.
b7 R. J. J. Chen, P. S. Chappelear, and R. Kobayashi, J. Chem. and Eng. Data, 1974, 19, 53.
sa L. C . Kahre, J. Chem. and Eng. Data, 1974, 19,67.
69 P. C. Tully, W. E. De Vaney, and H. L. Rhodes, Adv. Cryogenic Engineering, 1971, 16, 88.
6o 3. S. Kirk and W. T. Ziegler, Adu. Cryogenic Engineering, 1965, 10, 160.
81 J. C. Mullins and W. T. Ziegler, Adu. Cryogenic Engineering, 1965, 10, 171.
64 B. F. Dodge and A. K. Dunbar,J. Am. Chem. Soc., 1927,49,591.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 87
chips and glass fibre. The liquid was sampled through a capillary tube reaching
into the lower compartment and the effluent was sampled to determine the
equilibrium gas composition. The apparatus could be operated up to 12 MPa
at cryogenic temperatures.

,j
vapour outlet

vapour
inlet

liquid sampling
-capillary

insert with threadec'

-outside surface

Figure 9 High-pressure vapour-liquid equilibrium cell used by Kirk and Ziegler

Akers and Eubanks63 described a very simple cell for use in a recirculating
vapour apparatus. The cell consisted of a cylinder of type 303 stainless steel
50 mm outside diameter, 28 mm inside diameter, and 300 mm long. The ends
were stopped by two stainless-steel plugs. The seals between the plug and the
cell were made with annealed copper gaskets. A liquid sample outlet and an
inlet for recycled vapour entered through the bottom plug and an outlet for
vapour entered through the top plug. A differential pressure gauge was placed
across the cell to establish the liquid level in the cell.
(3 W. W. Akers and L. S . Eubanks, Adv. Cryogenic Engineering, 1960,3,275.
4
88 Chemical Thermodynamics
McKinley, Brewer, and Wangs4 have also used a single-pass flow system to
study solid-vapour equilibrium of the oxygen +
hydrogen system. However,
they used a modified technique (which could equally well be used in studies of
vapour-liquid equilibrium). A mixture of oxygen +
hydrogen (containing an
excessive amount of oxygen) was passed through the equilibrium cell where the
liquid analysis cell
pure component vapour analysis cell,
cylinder
gauge.
v
n .

.,
n

pure component
cylinder

L-- - - - -- ---;
thermostat

Figure 10 Schematic representation of the apparatus used by Akers, Burns and Fairchild

excess oxygen was frozen out. The effluent mixture was analysed for its oxygen
content. The solid phase was not analysed. Smith, Sonntag, and Van Wylen66
used a variation of this method to investigate the carbon dioxide + nitrogen
system. Equilibrium was attained by two alternate methods. Initially a carbon
dioxide + nitrogen mixture in which the carbon dioxide concentration was
greater than that expected at equilibrium was passed through the equilibrium
cell; during this process the excess carbon dioxide was condensed and trapped
in the equilibrium cell. A portion of the effluent gas was analysed using an i.r.
gas analyser. The system eventually was plugged by the carbon dioxide being
deposited. Pure nitrogen gas was then passed through the system at the specified
equilibrium pressure. In this case equilibrium was attained by picking up rather
than depositing carbon dioxide.
Akers, Burns, and Fairchild66 have described an apparatus which is a flow
system but differs from a normal recirculating flow apparatus in that the gaseous
and liquid phases are decompressed before recirculation. A simplified flow
diagram of the apparatus is given in Figure 10. The apparatus consists essentially
of two pure-component source cylinders and a compressor which can be connected
either to a storage cylinder or directly to the experimental cell. The experimental
cell was made from 2.5 cm stainless-steel pipe and had a volume of about 25 cm3.
64 C. McKinley, J. Brewer, and E. S. J. Wang, Ado. Cryogenic Engineering, 1962, 7 , 99.
O5 G . E. Smith, R. E. Sonntag, and G. J. Van Wylen, Adv. Cryogenic Engineering, 1964,9, 197.
66 W. W. Akers, J. F. Bums, and W. R. Fairchild, Ind. and Eng. Chem., 1954, 46, 2531.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 89
The inlet tube to the cell was passed through the cell wall midway between the
ends of the cell. The tip of the inlet tube turned downwards towards the bottom
of the cell. A Bourdon pressure gauge was attached to the top of the cell

liquid sampling
tube \ iF:ygr
vapour
inlet

L
Figure 11 Simplified diagram of the cell used by Michels and co-workers

together with a vapour outlet. A liquid outlet tube was attached to the bottom
of the cell. The cell was pressurized to the desired pressure at a predetermined
temperature. The compositions of the liquid and gaseous phases were estimated
by measuring the thermal conductivity of the samples which were continually
bled off from the equilibrium cell. After the thermal conductivity of the samples
had been measured (at atmospheric pressure) they were recompressed and
passed through the cell again.
Michels and co-workers 67-60 have described a flow apparatus which is suitable
for operation at pressures up to 80 MPa. The flow apparatuses described above
6'A. Michels, G. F. Skelton, and E. Dumoulin, Physica, 1950, 16, 831.
68A. Michels, E. Dumoulin, and J. J. Th. van Dijk, Physica, 1959, 25, 840.
a A. Michels, E. Dumoulin, and J. J. Th. van Dijk, Physica, 1961, 27, 886.
90 Chemical Thermodynamics
are not normally capable of being used at such a high pressure. A simplified
diagram of the cell is illustrated in Figure 1 1 . The pure gaseous component
enters the cell at the bottom and passes through the liquid. On entering the
liquid the gas stream is divided up into fine streams of bubbles by the grooves on
the circumference of the conical plug. The gas and liquid are then equilibrated
in a porous glass fibre filter. Fine throttling valves enable samples of gas and
liquid phase to be withdrawn from the cell continuously. The apparatus was
operated between 240 and 390 K.
Krichevskii and co-workers 70-74 have described several flow systems, A high-
pressure single-pass flow system which can be operated at pressures up to 300 MPa
has been described by Ipat’ev and c o - w ~ r k e r s . ~ ~
Flow methods have been extensively used in the determination of the solubility
of liquids (and solids) in compressed gases 76-79 and the solubility of compressed
gases in liquids.80-g0 As pointed out previously, one of the limitations of the
single-pass flow method is that the liquid component must not be too volatile
(i.e. too soluble) in the gaseous component. Precise results can, however, be
obtained for mixtures in which the liquid component is only slightly soluble in
the gaseous phase by passing a considerable quantity of gas through the apparatus.
Static methods are not particularly suitable for studying the slight solubility of
liquids in compressed gases. The two methods are complementary; which is the
more suitable depends on the system and the temperature and pressure range to
be studied.
Static Methods.-A wide range of equipment has been described for studying
vapour-liquid equilibrium at high pressures using static methods. A widely
used form of the static method consists of placing a two-phase mixture in a
container at a controlled temperature and agitating the mixture in some way,
such as rocking the container. After equilibrium is reached, portions of the
vapour and liquid are withdrawn and analysed. Often, but not always, no
provision is made for keeping the pressure constant while samples are withdrawn.
A drop in pressure during sampling disturbs the equilibrium and can lead to
70 I. R. Krichevskii and N. E. Kazanova, Z h u r . 3 ~ Khim.,
. 1939, 13, 1690.
71 I. R. Krichevskii and R. S. Kal’varskaya, Zhur. jiz. Khim., 1940, 14, 748.
78 I. R. Krichevskii and M. V. Koroleva, 2 h u r . j ~ Khim.,
. 1941, 15, 327.
7s I. R. Krichevskii and D. Y. Gamburg, Zhur. fiz. Khim., 1943, 17, 215.
74 I. R. Krichevskii, N. M. Zhavoronkov, and D. C. Tsiklis, Zhur.fiz. Khim., 1937, 8, 317.
7.5 V. V. Ipat’ev, V. P. Teodorovich, A. P. Brestkin, and V. S. Artemovich, Zhur. fit. Khim.,
1948, 22, 834.
76 E. Bartlett, J. Amer. Chem. SOC.,1927, 49, 65.
77 H. Braune and F. Strassman, Z . phys. Chem. (Leipzig), 1929, 143, 225.
78 H. Saddington and N. Krase, J. Amer. Chem. SOC.,1934, 56, 353.
70 B. Miller and B. F. Dodge, Ind. and Eng. Chem., 1940,32, 434.
Schen-Wu-Wan and B. F. Dodge, Ind. and Eng. Chem., 1940,32,95.
C . J. Sterner, Adu. Cryogenic Engineering, 1961, 6, 467.
88 F. P. Stein, C. J. Sterner, and J. M. Geist, Chem. Eng. Progr., 1962, 58, 70.
83 F. P. Stein, L. C. Claitor, and J. M. Geist, Adu. Cryogenic Engineering, 1962, 7 , 106.
W. B. Streett, Cryogenics, 1965, 5, 27.
W. B. Streett and C. H. Jones, Adu. Cryogenic Engineering, 1966, 11, 356.
W. B. Streett and C. H. Jones, J. Chem. Phys., 1965, 42, 3989.
87 W. B. Streett, Chem. Eng. Progr., Symp., 1967, 63, no. 81, 37.
W. B. Streett, Cryogenics, 1968, 8, 88.
W. B. Streett, Trans. Faraday SOC.,1969, 65, 696.
90 W. B. Streett, R. E. Sonntag, and G. J. van Wylen, J. Chem. Phys., 1964,40, 1390.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 91
considerable error. The other important type of static equipment, the bubble-
point dew-point apparatus, has already been described under critical properties
of mixtures. In that section, only apparatuses used for studies in which the
primary aim was to study the critical properties were discussed. Several more
complicated types of apparatus which can be used to determine critical properties
are described in this section.
Static phase-equilibrium cells which operate at pressures over 100 MPa have
been developed to study the phenomenon usually referred to as gas-gas
immiscibility.* Several of these are discussed in this section. A substantial
proportion of the work on gas-gas immiscibility has been carried out in Russia,
and the Russian work has been reviewed in detail in several books and reviews
by T s i k l i ~ . ~ ~ - l ~ ~
Firstly, in this section, we consider those static methods in which the gas and
liquid phases are analysed and there is no optical system whereby the contents
of the cell can be observed. Secondly, we consider those static methods which
have been used to study the one phase-two phase boundary using optical
techniques. Usually in such methods, the appearance of the second phase is
determined by direct visual observation of the resulting turbidity or meniscus.
The optical method usually provides fewer experimental data than the analytical
methods because the concentrations of co-existing phases are not usually
determined. However, there are some workers who have sought to combine the
advantages of both the analytical and optical methods by designing an optical
cell with provision for analysis of both phases.
Non-optical Methods. Rigas, Mason, and Thodos lol have described a variable-
volume static cell for studying vapour-liquid equilibrium. The basic features
of the cell are illustrated in Figure 12. The equilibrium cell consists of the main
body, a head, and a cap. The cap fits over the head and threads on to the main
body. The seal between head and main body is made by a Teflon ring. The
main body consists of a stirring chamber, an equilibrium chamber, and a mercury
chamber. The stirring chamber consists of a hollow cylinder along the axis of
the cell which is isolated from the mercury chamber but provided with radial
openings into the equilibrium chamber. The equilibrium and mercury chambers
are separated by an annular movable piston extending from the outside of the
hollow cylinder of the stirring chamber to the inside diameter of the main body.
G. M. Schneider, ‘Experimental Thermodynamics’, Vol. 11, Chapter 16, part 2, Butterworths,
London, 1975.
8a G. M. Schneider, Adv. Chem. Phys., 1970,17, 1.
O3 G. M. Schneider, Ber. Bunsengesellschaft.Phys. Chem., 1966, 70, 497.
94 G. M. Schneider, Ber. Brmsengesellschaft.Phys. Chem., 1972, 76, 325.
G. M. Schneider, this volume, Chapter 4.
G. M. Schneider, in ‘Topics in Current Chemistry’, Vol. 13, 1970, 559.
O7 D. S. Tsiklis and L. A. Rott, Russ. Chem. Rev., 1967, 36, 351.

D. S. Tsiklis, ‘Handbook of Techniques in High-pressure Research and Engineering’ (trans.

A. Peabody), ed. A. Bobrowsky, Plenum, New York, 1968.
D. S. Tsiklis, ‘Phase Separations in Gas Mixtures’, Khimiya, Moscow, 1969.
loo D. S. Tsiklis, ‘Phasentrennung in Gasgemischen’, VEB Deutscher Verlag fur Grundstoff-
industrie, Liepzig, 1972.
101 T. J. Rigas, D. F. Mason, and G. Thodos, Ind. and Eng. Chern., 1958,50, 1297.

* Schneider 91--O6 has discussed, both in this volume and elsewhere, the variety of phase
behaviour which can exist at high pressures.
92 Chemical Thermodynamics
The seals on this piston were made of O-rings and a Teflon guide attached to the
top of the piston permitted it to move smoothly inside the main body. The
stirrer mechanism consists of four two-bladed fans along the stirrer stem. A
small magnet is fitted into the lower end of the stirrer stem and the stirrer is

Figure 12 Static cell used by Rigas, Mason, and Thodos

rotated by magnetically coupling it to an externally located magnetic drive.

Four valves are incorporated into the equilibrium cell for withdrawal of vapour
and liquid samples.
In order to sample the vapour and liquid phases after equilibrium has been
set up the following procedure is adopted. The space between the two vapour
valves and the space between the two liquid valves are evacuated and then
sample is allowed to move into the spaces between the two valves by momentarily
opening the liquid and vapour valves which are nearest the equilibrium chamber.
The volume of the liquid and vapour samples were 0.2 and 0.75 cms respectively.
Although the pressure decreases slightly during withdrawal the authors claimed
 Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 93
that the trapped vapour and liquid are representative of the original conditions,
as sample volumes are small in comparison to the volume of the equilibrium
chamber and the time of withdrawal is small.
Tsiklis and co-workers loehave described a cell which is suitable for operation
up to 10 MPa. It consists of a cylindrical vessel closed with a packingless
spherical cap which is fitted with two membrane valves, one for removal of the
gas sample and the other connected to a membrane null-pressure instrument.

to vapour sampling valve

to eccentric drive +

Figure 13 Rocking autoclave cell designed by Bol’shakov and Linshits

The cap is also fitted with a valve for filling the cell. The contents of the cell
can be stirred by a stirrer whose shaft is moved up and down by a solenoid
placed over a high-pressure tube projecting out of the cap of the cell.
A simple apparatus has been used by Bol’shakov and Linshits.los The apparatus
can be operated at high pressures and is fairly typical of the rocking autoclave
system. The cell is illustrated in Figure 13 and consists of a steel vessel closed
by a steel head fitted with a valve for removal of liquid phase. A high-pressure
cross is fitted to the middle of the cell. A valve for removal of gas-phase sample
and valves to pressure gauges are fitted directly into the cross. The cell is sub-
merged in a liquid thermostat so that its axis is horizontal. The cell is pivoted at
the middle and can be rocked by an acentric drive at one end. Stirring is effected
by gravity on a steel ball in the cell.
Lindroos and Dodgelo4have designed a system which can be operated up to
nearly 400 MPa and over the temperature range 300 to 385 K. The essential
details of the equilibrium cell are shown in Figure 14. The system these workers
investigated was ammonia + nitrogen, which shows the barotropic effect at
about 200 MPa and 370 K. Initially ammonia was introduced into the cell as a
liquid at its vapour pressure. Nitrogen was then compressed into the cell and
lo2 D. S. Tsiklis, E. V. Mushkina, and L. I. Shenderei, Inzh. Fiz. Zhrir., 1958, 1, 3.
103 P. E. Bol’shakov and L. P. Linshits, Trudy. Gas. Inst. ,420. Prom., Gaskhimizdat, 1954,3, 18.
lo’ A. E. Lindroos and B. F. Dodge, Chem. Eng. Progr. Symp., 1953, 48, no. 3, 10.
94 Chemical Thermodynamics
the pressure adjusted to a specified value. After allowing time for the contents
of the cell to reach the temperature of the oil bath in which the cell was totally
immersed, a magnetic stirrer within the cell was operated. After 3 h of inter-
mittent agitation, the pressure within the equilibrium cell was measured by a
manganin resistance gauge and samples of each phase were taken and analysed.

1
solenoid coils

stirrer

sampling

liquid smiling port

Figure 14 Equilibrium cell suitable for pressures up lo 400 MPa used by Lindroos and
Dodge

The pressure after sampling was measured to establish the magnitude of the
pressure drop during sampling. The average pressure drop was about 0.6 per
cent of the initial pressure and the authors argued that such a pressure change
would have a negligible effect on the composition except near the critical region.
Krichevski and Tsiklis lo5 have described a static apparatus for investigating
gas-gas equilibria up to 2000 MPa. The basic design of the apparatus is
illustrated in Figure 15. It consists of a high-pressure cylinder containing a
I. R. Krichevskii and D. S. Tsiklis, Zhur. 32.Khim., 1943,17, 115.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 95
piezometer in which the mixture of known composition is compressed over
mercury. The top of the piezometer is connected to a sampling valve at the top
of the high-pressure cylinder. The bottom of the cylinder is connected to a
manganin pressure gauge and pressure intensifier. The piezometer is filled with

to pressure- intensifier
Figure 15 Static apparatus for investigating gas-gas equilibria used by Krishevskii and
Tsiklis

the gas mixture at a temperature and pressure where only one phase is possible.
The pressure-transmitting liquid is then forced into the cylinder by hydraulic
oil acting on the low-pressure side of the pressure intensifier. This causes mercury
to enter the piezometer and compress the gas. The system is allowed to
equilibrate at a predetermined pressure and temperature in the two-phase region.
Samples are then carefully withdrawn through the valve at the top of the cylinder,
the pressure being maintained constant by pumping mercury into the piezometer.
96 Chemical Thermodynamics
Firstly, the upper phase and then the lower phase will be forced out through the
valve and subsequent analysis enables the composition of co-existing phases to
be determined. One of the disadvantages of this apparatus is that it depends on
a homogeneous phase separating into two phases and there is no provision for

high-
vesse

high - pressure
cross

to thermostatted
manganin manometer

to valve and pressure

intensifier
Figure 16 Static apparatus for investigating gas-gas equilibria used by Tsiklis and
Madennikova

stirring. Tsiklis lo6has more recently designed a piezometer which incorporates

a stirrer and solenoid. Electric leads to supply the soienoid are passed through
the head of the high-pressure cylinder. The incorporation of a stirrer allows for
subsequent experiments at lower pressures than the first experiment.
Neither of these apparatuses can be used at temperatures much higher than
460K because of safety risks involved with using kerosene as a pressure-
lo* D. s. Tsiklis, Dokiady Akad. Nauk S.S.S.R., 1952,86, 993.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 97
transmitting fluid and because of the solubility of mercury in the compressed
gases. More recently, Tsiklis and Maslennikova lo7have described an apparatus
which does not involve mercury in contact with the experimental fluid and can
operate up to 770 K at pressures of several hundred MPa. The apparatus, which
is illustrated in Figure 16, consists of a high-pressure vessel fitted with an electro-
magnetic stirrer. This vessel is fitted to a high-pressure cross which has connections
to a manganin pressure gauge, sampling valve, and pressure intensifier. An
additional sampling valve is connected into the top of the vessel. The manganin
pressure gauge is insulated from the mixture whose pressure is being measured
by a bellows filled with dry kerosene. The components of the mixture are
measured into the high-pressure vessel at room temperature, the apparatus is
heated to a predetermined temperature, and the pressure is adjusted to near the
required pressure. After equilibrium is established the pressure is accurately
measured and samples are removed for analysis. When the cell contents are
equilibrating the pressure intensifier is isolated by means of a valve. During
sampling this valve is opened and one component of the mixture, such as water,
is added to the system to compensate for the sample being removed. There is
no pressure drop during sampling.
Todheide and Franck lo8have described a cell which is suitable for investigating
vapour-liquid equilibrium at high temperatures and pressures. The equilibrium
vessel consists of a thick-walled stainless-steel cylinder sealed at each end with
two Bridgman pistons. The cell is fitted with three capillary tubes which enter
the cell through the top piston, each of which terminates at a different level in
the equilibrium vessel; one terminates near the top, one near the bottom, and the
other near the middle of the cell. The contents of the cell can be introduced
through any of the capillary tubes. After the mixture has been allowed to
equilibrate, samples are withdrawn at different heights of the vessel through the
three capillary tubes and analysed. The pressure generated in the vessel is
controlled by the initial filling procedure, subsequent heating (needed to bring
the system to the experimental temperature), and the amount withdrawn for
sampling. The mixture in the cell can be agitated by a magnetically driven stirrer.
Sage, Lacey, and co-workers 1oB-112 have designed an apparatus for studying
the p , V,T relations of liquids and gases. The apparatus is attractive in that it
can be used to study one- or two-phase properties of either pure substances or
mixtures. These workers have published more p,V,T data on mixtures than
any other group.
In essence, the apparatus consists of a U-tube, closed at each end, which is
partly filled with a constant amount of mercury. The sample is confined in one
arm while air under pressure may be admitted to the other in order to change
the volume of the sample. The position of the mercury in the air-pressurized
arm can be measured accurately, and this enables the volume of the sample to
be estimated within about 0.05 per cent. The U-tube consists of two steel cylinders
lo7 D.S. Tsiklis and V. Y . Maslennikova, Doklady Akad. Nauk S.S.S.R., 1964, 157, 427.
K. Todheide and E. U. Franck, Z . phys. Chem. (Frankfurt), 1963,37, 387.
109 B. H. Sage and W. N. Lacey, Trans. Amer. Inst. Mining Met. Engineers, Petroleum Div.,
1940,136, 136.
H. C. Wiese, H. H. Reamer, and B. H. Sage, J. Chem. and Eng. Data, 1970, 15, 7 5 .
111 H.H.Reamer and B. H. Sage, J. Chem. and Eng. Data, 1964,9, 24.
112 H. H. Reamer, B. H. Sage, and W. N. Lacey, J. Chem. and Eng. Data, 1960, 5, 44.
98 Chemical Thermodynamics
of accurately known internal dimensions and a connecting tube. The sample
can be stirred by a stirrer which is driven from outside the cell by a magnetic
agitator. Pressure is measured using a pressure balance which was specially
designed for the purpose.
The apparatus has two disadvantages when used to measure vapour-liquid
equilibrium. Firstly, the sample is in contact with mercury at high temperatures.
Secondly, the co-existing vapour and liquid phases are not analysed. Therefore,
in order to establish the dew and bubble points, which were not the only aim of
the experiment, it was necessary to establish discontinuities in plots of volume
against pressure at constant temperature. This can, in some circumstances,
particularly near the critical point, lead to significant uncertainty in the dew-
and bubble-point pressures. Despite these disadvantages there are many systems
for which this group's data are the best currently available.
Several other non-optical methods have been described in the l i t e r a t ~ r e . ~ ~
OpticaZ Methods. Optical methods may conveniently be considered in three
categories. (a) Methods suitable up to 10 to 20 MPa in which the coexisting
liquid and vapour phases at predetermined pressures and temperatures are
analysed. Such methods combine the advantages of the optical and analytical
methods. (b) Methods suitable up to 10 to 20 MPa in which the composition of
the mixture studied is controlled by the filling of the cell at the outset of the
experiment. (c) Methods suitable for high-pressure studies above 10 MPa in
which the composition of the mixture studied is controlled by the filIing of the
cell at the outset of the experiment. Such methods are suitable for studying
either or both liquid-liquid or so-called gas-gas immiscibility.
Thodos and co-workers 1139 114 have used a visual observation cell to investigate
the system ethane +n-heptane. The high-pressure cell consisted of a stainless-
steel cylinder with an outside diameter of 15 cm and an internal space of 2.5 cm
diameter by 30 cm long. The contents of the cell could be viewed through a full-
length glass window. The cell was mounted on a pivoted support and was rocked
continuously with a motor-driven connecting rod. A representative sample of
the cell contents could be trapped in a volume of 2 cm3 between two valves in the
equilibrium cell. The experimental procedure involved degassing the mixture
in an auxiliary cell. The degassed mixture was then forced by mercury displace-
ment into the observation cell which was maintained at a specified temperature.
After charging, mercury was introduced into the cell until both a liquid and a
gas phase existed in the cell. The pressure was then increased by small increments
and after the cell had been rocked to establish equilibrium, volumetric data were
taken for the total mixture and liquid phase at each pressure. The increase in
pressure was continued until either complete liquefaction or complete vaporization
could be observed. (The state of the mixture at the highest pressure depends on
the temperature. The difference between the maxcondentherm and the critical
temperature can be quite large in ethane +
n-heptane.) The composition of the
mixture was determined by taking a sample from the cell at a pressure and
temperature above the critical point. The sample was analysed using a mass
spectrometer.
lla 0. Ekiner and G . Thodos, J . Chem. and Eng. Data, 1966,11, 154 and 457.
11' 0.Ekiner and G . Thodos, Cunad. J. Chem. Eng., 1966,43,205.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 99
McCurdy and Katz1I5 have described a simple visual cell for measuring
vapour-liquid equilibrium. The equilibrium cell had a volume of 92 cm3, was
fitted with double glass windows and could be rocked to agitate the fluid phases
and thus bring them into equilibrium. The sample was confined over mercury,
To sample the gas phase, mercury was driven into the cell at the bottom at the
same rate as gas sample was removed from the top. The liquid phase was
sampled by completely removing the vapour sample and then the liquid sample
by injection of mercury into the bottom of the cell. The gas and liquid samples
were analysed by chemical gas absorption methods.
Besserer and Robinson 116 have described an apparatus which can be operated
up to 8 MPa for measuring refractive indices and compositions of coexisting
phases in vapour-liquid equilibrium. The composition of each phase was deter-
mined by gas chromatographic analysis. The refractive index was determined
accurately and was used to calculate the density of each phase using the Lorentz-
Lorenz molar refractivity relation. The cell consists of three sections, a central
section and two cylinder-piston end sections. The central section is fitted with
one pyrex glass window which is positioned at 30" to a highly polished steel
mirror. A laser beam can be passed through the window, through the cell
contents, and reflected back along the same path by the mirror. An auto-
collimating telescope and precision rotary table enabled the refractive index of
the contents to be measured accurately in less than 1 min. The two cylinder-
piston assemblies enabled the liquid or gas phase to be placed in the central
section, It was also possible to move the pistons simultaneously up or down
maintaining a constant volume between them. Mixing and equilibrium were
attained by moving the pistons simultaneously. On the upstroke, an internal
check valve in the central section closed and the liquid and gas phases were
forced through an external circulation line and sprayed into the upper cylinder.
On the downstroke, a valve in the external circulation line closed and the
heterogeneous mixture passed through the check valve. Equilibrium was usually
attained in ten cycles. The pressure in the system was measured with a strain
gauge pressure transducer. The liquid and vapour were sampled by a micro-
metering valve. The average sample size was mol and the pressure drop
associated with taking a sample was always less than 0.01 MPa. Robinson,
Hughes, and Sandercock 117 have described a similar apparatus without facilities
for measuring the refractive index.
Roof and co-workers 11*-ll9 have designed a pressure vessel which has internal
windows. Basically the vessel consists of thick-walled glass tube of about 1 cm
internal diameter contained in a metal cell. The sample is confined in an annular
space between the outside of the glass tube and the metal cell. The contents of
the cell are viewed by a periscope which can be moved up and down inside the
glass tube. The design is attractive in that it enables visual observation at
pressures up to 50 to 60 MPa without the need for carefully ground optical

116 J. L. McCurdy and D. L. Katz, Ind. and Eng. Chem., 1944,36,674.

ll8 G . J. Besserer and D. B. Robinson, Canad. J. Chem. Eng., 1971, 49, 651.
117 D.B. Robinson, R. E. Hughes, and J. A. W. Sandercock, Canad.J. Chem. Eng., 1964,42,143
F. W. Wells and J. G. Roof, Rev. Sci. Instr., 1955, 26, 403.
J. G. Roof and J. D. Baron, J. Chem. and Eng. Data, 1967, 12, 293.
100 Chemical Thermodynamics
surfaces. Kurata and Kohn,120 Bloomer and Parent,121 and Lu, Newitt, and
Ruhemann24have also described optical apparatus which can be used to study
one-phase two-phase boundaries including the critical point for mixtures.
Most optical cells for studying phase behaviour at pressures in excess of
10 MPa do not use glass for the optical portion. While glass capillary tubes can
to pressure measuring
and generating apparatus

Figure 17 Optical cell of Schneider

be used at far higher pressures the behaviour of glass above 10 to 20 MPa is
unpredictable.122
Several high-pressure optical cells have been described in the literature. Most
of these use synthetic sapphire windows. Schneider123 has described a very
simple high-pressure cell which can be used to study liquid-liquid immiscibility
up to 5 0 0 MPa over a temperature range of 300 to 460 K. A simplified diagram
of the cell is given in Figure 17. The contents of the cell were viewed through
sapphire windows. A moving Teflon piston enables the contents of the cell to
be pressurized without the mixture under investigation coming into contact with
the pressurizing fluid.
Erlich and Kurpen 124 have described a simple high-pressure cell which was
used to study phase equilibria of polymer-solvent systems at high pressures.
These workers used large windows (2.5 cm diameter made from Pyrex, feurex,
12* F. Kurata and J. P. Kohn, Petroleum Processing, 1956, 57.
lZ1 0. T. Bloomer and 3. D. Parent, Chem. Eng. Progr. Symp., 1953, 49, no. 6, 11.
lZ2 Ref. 41, p. 92.
lZ3 G. Schneider, Z . phys. Chem. (Frankfurt), 1963, 37, 333.
P. Ehrlich and J. J. Kurpen, J. Polymer Sci., A , General Papers, 1963, 1, 3217.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 101
or plate glass which were sealed by two sets of Neoprene O-rings). The cell
operated up to 60 MPa and apparently did not involve the use of optically flat
windows.
Oeder and Schneider 125 have described a slightly more complicated optical
cell for operation up to 400 MPa. Although these workers used the cell at low

Figure 18 Optical cell of Buback and Franck

temperatures the basic design is suitable for operations at higher temperatures.

The cell was constructed from copper-beryllium and was fitted with two sapphire
windows. The fluid mixture in the cell could be agitated by a small magnetic
stirrer driven from outside the cell by a rotating magnet. The pressure was
transmitted to the mixture by a flexible bellows. The displacement of the bellows
could be detected and measured by a displacement transducer. This enables the
cell to be used for pressure-volume temperature measurements.
A slightly more complicated apparatus has been described by Buback and
Franck126 which only involves one sapphire window (Figure 18). The inner
125 D. Oeder and G. Schneider, Eer. Eunsengedschaft. Phys. Chem., 1969, 73, 229.
n6 M. Buback and E. U. Franck, Ber. Bunsengesellschaft. Phys. Chem., 1972, 76. 350.
102 Chemical Thermodynamics
surface of the sapphire window has a cavity in it and a gold-platinum bellows is
attached to the sapphire by a compression seal. The cell was originally used for
determining the vapour pressure of ammonium halides but can, with minor
modification, be used for pressure-volume temperature measurements and for
phase-behaviour studies on mixtures.

high-pressure cylinder
agnetic ring stirrer

and pressure measuring and

generating equipment.

Figure 19 Optical cell of Alwani and Schneider suitable for studies up to 400 MPa

A more complicated optical cell for studying phase behaviour which can
operate up to 400 MPa at 670 K has been developed by Alwani and f3~hneider.l~'
The cell, which is illustrated in Figure 19, is constructed from Nimonic 90 steel.
The phase transitions in the cell are observed through a single window of synthetic
sapphire. The pressure is transmitted to the fluid mixture under test by a piston
fitted in the base of a thick high-pressure tube. The moving seal between the
piston and the cylinder is made by neoprene or W o n O-rings. The displacement
of the piston could be determined by a displacement transducer enabling the cell
to be used for p,V,T studies. Stirring of the fluid mixture was achieved by a
magnetic ring in the cell which could be moved backwards and forwards by
making the inner face of the window assembly first a magnetic north pole and
then a south pole and repeating this cycle. This was achieved by passing a direct
current of slowly alternating polarity through a large solenoid attached to the
window plug. The cell was mounted in a large aluminum block fitted with
electrical heating. The temperature of the cell was measured directly with a
steel-sheathed thermocouple in the cell. The cell could be filled through a
capillary tube.
In order to study gas-gas immiscibility in helium + argon, De Swaan Arons
and Diepen12* used a glass cell which could be inserted into a high-pressure
autoclave which was fitted with windows. The cell is illustrated in Figure 20.
1%' Z . Alwani and G. M. Schneider, Ber. Bunsengesellschaft. Phys. Chem., 1969,73, 294.
J. De Swaan Arons and G. A. M. Diepen, J. Chem. Phys., 1966,44,2322.
Studying Phase Behaviour of Mixtures at High Temperatures and Pressures 103
Accurately known amounts of helium and argon were transferred from calibrated
glass bulbs by a mercury-displacement technique. The contents of the cell were
separated from the atmosphere by mercury in the U-portion of the cell. The
cell was installed in the high-pressure autoclave and the lower part immersed in
mercury. The top of the cell was positioned so that the compressed mixture

Cooling coil

Soft iron
I/
r”
Thermometer shaft

Water

Security window~-B Measuring vessel

Mercury

Figure 20 Optical cell of De Swaan Arons and Diepen

(Reproduced by permission from ‘The Physics and Chemistry of High Press1xes’, Socie:ty
of Chemical Industry, London, 1963, p. 11)

could be observed through the cell window. The mixture could be stirred by a
small permanent magnet which could be pushed up and down by means of a
magnetic solenoid outside the cell and a magnetic coupling. This cell has the
disadvantage that the experimental mixture is in contact with mercury. While
this may not be a grave disadvantage at low temperatures it is undesirable at
temperatures above 400 K.
Khodeeva 129 has developed a simple technique which can be used to determine
if a system exhibits gas-gas equilibria of the first kind (i.e. the critical locus of
the mixture starting at the critical point of the less volatile component
100 S. M.Khodeeva, Russ. J. Phys. Chem., 1966,40, 1061.
104 Chemical Thermodynamics
immediately moves to higher pressures and temperatures as the concentration
of the other component is increased). The less volatile component is introduced
into a glass tube such that the overall density of liquid and vapour phases is the
critical density. The critical temperature is measured and then a small quantity
of the second component is added. The critical temperature is again measured
and if it is greater than that of the pure component gas-gas equilibrium of the
first kind must be present.

4 Concluding Remarks
There will almost certainly be a significant increase in research into the phase
behaviour of fluid mixtures over the next decade. There are three areas in which
important advances may be expected: (a) The determination of accurate vapour-
liquid equilibrium data on simple systems up to the critical point undertaken
specifically for comparison with the generalized van der Waals model, (b) the
investigation of vapour-liquid equilibrium for systems which do not have a
continuous critical pressure-critical temperature projection such as Scott’s
type I11 and type IIIm projections, and (c) the investigation of gas-gas
immiscibility in more detail particularly for systems which are suitable for
comparison with the generalized van der Waals theory.
Some of the apparatus already described in the literature is capable of good
precision but often the systems studied are not suitable for comparison with
theory (i.e. they involve complex substances or gases for which quantum effects
are important) or the measurements have been carried out over a very restricted
range of temperature.
The helpful comments of Ian McKinnon were appreciated.
4
High-pressure Phase Diagrams and Critical
Properties of Fluid Mixtures
BY G. M. SCHNEIDER

1 Introduction
Whenever one looks into modern textbooks of thermodynamics or physical
chemistry in order to get information on the phase behaviour of fluid mixtures
one might think that most of the inspiring work of van der Waals, Kohnstamm,
Roozeboom, Timmermans, Kuenen, and the other pioneers in this field had been
completely forgotten and that practically nothing had happened since that time.
Only the most simple types of gas-liquid and liquid-liquid equilibria are dis-
cussed ;very rarely are the so-called gas-gas equilibria that were discovered some
thirty years ago additionally treated.
By the rapid development of experimental high-pressure techniques, however,
much progress has been made during recent decades in the investigation of
fluid systems at high pressures and of their high-pressure phase diagrams. In
Section 2 of this Report the results on binary and ternary mixtures will be briefly
reviewed for gas-liquid, liquid-liquid, and gas-gas phase equilibria at pressures
essentially above 30 MPa. It is a widespread misconception that these three
types of two-phase equilibrium in fluid mixtures are distinguishable in all cases.
Systematic investigations, however, in recent years have shown that the limits
between these three types are not well defined and that continuous transitions
occur. In Section 3 recent results on some selected systems will be discussed with
respect to this question and the continuity between gas-liquid, liquid-liquid,
and gas-gas equilibria will be deduced from the experimental data.
In Section 4 it will be shown that this concept of a continuity between all
kinds of phase equilibria in fluid mixtures will aid in understanding the great
variety of phase relationships and in giving a certain order to many different
types of phase behaviour and critical phenomena in fluid systems. Here practical
applications will be equally briefly reviewed. The presentation will be based on
some earlier essays and review papers of the author 1--5 to which the reader is
referred for examples, references, and details. A review of these problems has
also been given in a book by Rowlinson.6 Related problems have been treated

G. M. Schneider, Ber. Bunsengesellschaft.Phys. Chem., 1966, 70, 497.

G. M. Schneider, Ado. Chem. Phys., 1970, 17, 1.
G. M. Schneider, Fortschr. Chem. Forsch., 1970, 13, 559.
G. M. Schneider, Ber. Bunsengesellschaft.Phys. Chem., 1972,76, 325.
G. M. Schneider, in ‘Water, A Comprehensive Treatise’, Vol. 11, ed. F. Franks, Plenum Press,
New York, 1973, chap. 6.
* J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, 2nd edn., Butterworth, London, 1969.
105
106 Chemical Thermodynamics
in books and review articles by several a u t h o r ~ . ~ -For
l ~ the phase-theoretical
and thermodynamic foundations see references 15-21 and 22-25 respectively.
In Sections 2 to 4 critical phenomena will be of primary importance since they
make possible a systematic discussion of all types of phase behaviour in fluid
mixtures and of the relationships between them. The definition of a critical
point for a mixture is essentially the same as that for a pure component: at a
critical point all intensive properties of two phases in equilibrium become
identical. Whereas pure substances are characterized by a critical point for the
equilibrium gas-liquid, binary systems exhibit a critical line in the three-
dimensional p-T-x space (where x denotes mole fraction), and systems with It
components an (n - 1)-dimensional critical surface in the (n + 1)-dimensional
p-T-xl-x,-. ..-xnhl space for all kinds of fluid-fluid equilibria.
In all Figures in this Report the p(x) isotherms, the T(x) isobars, and the
p ( T ) isopleths [= p(T) cuts for x = const.] are represented by solid lines, all
kinds of critical curves by thick solid lines, thep(T) phase diagrams of the pure
components by dashed lines, and the three-phase curves (e.g. LLG) by -*-.-.-.
Two-phase regions are indicated by hatching.
The following symbols are used: T = temperature; p = (total) pressure;
x = mole fraction (of component I1 if not indicated); c = concentration;
M = mol dm-3; L = liquid phase, G = gaseous phase, S = solid phase;
CPI, CP I1 = critical point of the pure component I or 11; CP = critical
point of a mixture; UCST (LCST) = upper (lower) critical solution temperature;
A, B, C, D, E, K = critical end point; Q1, Qa = quadruple point.
D. S. Ciklis, ‘Phasentrennung in Gasgemischen’, VEB Deutscher Verlag fur Grundstoff-
industrie, Leipzig, 1972.
8 J. M. Prausnitz, ‘Molecular Thermodynamics of Fluid-phase Equilibria’, Prentice-Hall,
Englewood Cliffs, New Jersey, U.S.A., 1969.
I. R. Krichevskii, ‘Phase Equilibria in Solutions at High Pressures’, Goskhimizdat, Moscow,
1946 and 1952.
D. S. Tsiklis and L. A. Rott, Russ. Chem. Rev., 1967, 36, 351.
l1 W. B. Kay, Accounts Chem. Res., 1968,1,344; S . C. Pak and W.B. Kay, Znd. and Eng. Chem.
(Fundamentals), 1972, 11, 255.
la E. U. Franck, in ‘Physical Chemistry. An Advanced Treatise’, Vol. 1, ed. H. Eyring, D.
Henderson, and W. Jost, Academic Press, New York, 1971.
l3 R. L. Scott, Ber. Bunsengesellschaft. Phys. Chem., 1972, 76, 296.
l4 W. B. Streett, Asfrophys. J., 1973, 186, 1107.
lJ J. P. Kuenen, ‘Die Theorie der Verdampfung und Verfliissigung von Gemischen’, Barth,
Leipzig, 1902.
16 ‘Die heterogenen Gleichgewichte’, ed. H. W. B. Roozeboom; Systeme aus zwei Komponenten,
Vieweg, Braunschweig, Part 1, 1904. Part 2, 1918.
l7 J. Timmermans, ‘Les solutions concentrks’, Masson, Paris, 1936.
l8 J. E. Ricci, ‘The Phase Rule and Heterogeneous Equilibrium’, Van Nostrand, Toronto, 1951.
J. Zernike, ‘Chemical Phase Theory’, Kluwer, Deventer, 1955.
2o A. Findlay, ‘The Phase Rule’, 9th edn., revised by A. N. Campbell and N. 0. Smith, Dover
Publications, New York, 1951; A. Findlay, ‘Die Phasenregel und ihre Anwendungen’, Verlag
Chemie, Weinheim, 1958.
21 R. Vogel, ‘Die heterogenen Gleichgewichte’, Akademische Verlagsgesellschaft,Leipzig, 1959.
2o J. D. Van der Waals and Ph. Kohnstamm, ‘Lehrbuch der Thermodynamik‘, Vol. 1, Barth,
Leipzig, 1912.
2s I. Prigogine and R. Defay, ‘Chemical Thermodynamics’, translated and revised by D. H.
Everett, Longman Green and Co., London, 1954. I. Prigogine and R. Defay, ‘Chemische
Thermodynamik’, VEB Deutscher Verlag fur Grundstoffindustrie, Leipzig, 1962.
R. Haase, ‘Thermodynamik der Mischphasen’, Verlag Chemie, Weinheim, 1956.
A. V. Storonkin, ‘Thermodynamics of Heterogeneous Systems’, University Press, Leningrad,
1967.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 107
2 Two-phase Equilibria in Fluid Mixtures
A. Gas-Liquid Equilibria.-Because of their great industrial and scientific
interest, gas-liquid equilibria have been extensively studied as a function of
pressure up to the critical region, and much work has been done for the related
phenomena of the solubility of compressed gases in liquids and of liquid or
crystalline substances in compressed l l s 16#18-21
6s 8s

t
P

7- T-
Figure 1 Gas-liquid equilibria in binary systems (schematic representation; for symbols
see Section 1). a, Three-dimensional representation in the p-T-x space; b, p ( T )
projections of the phase diagrams

Since the critical pressures of pure components are rarely higher than 10 MPa
the pressure range necessary for these investigations is rather limited; thus these
phenomena are rather outside the scope of this Report. Representing the low-
pressure limit they are, however, important for the understanding of all types of
fluid-fluid phase equilibria at high pressures. For the thermodynamic and
theoretical background see references 8 and 28.
GasLiquid Critical Curves of Binary Systems. In Figure l a the pressure-
temperature-composition surface is represented schematically for the gas-
liquid equilibria of a binary system in a simple case. The dashed lines are the
vapour pressure curves of the pure components; they end at the critical points
CP I and CP I1 of the pure components I and 11. Some pressure-composition
cuts for constant temperature are given. The critical point of the binary system
is situated at the extreme value of eachg(x) isotherm or (not shown in Figure la)
at the extreme value of each T(x) isobar. The line that connects the critical
points of all binary mixtures is the critical curve; in a pressure-temperature
projection the critical curve is the envelope of all p ( T ) curves for constant
composition (so-called isopleths). At temperatures and pressures beyond the
critical curve the constituents are miscible in all proportions.
The phase equilibria and the critical phenomena are most easily understood
with the aid of the pressure-temperature projections of the critical lines. The
most important types are schematically represented in Figure lb. The critical
108 Chemical Thermodynamics
curve is not interruped and may run through a pressure maximum (curve l),
through a temperature minimum (curves 4 and 5 ) or a temperature maximum
(curve 61, or may run monotonically between the critical points CP I and CP I1
of the pure components I and I1 respectively (curves 2 and 3). The fact that the
critical p ( T ) curve represents nearly a straight line for type 3 indicates that such
systems are not far from ideal. Temperature maxima or minima on critical

a b

--3 1 1
- XI!
I1

Figure 2 Critical phenomena in binary systems where gas-liquid and liquid-liquid

equilibria interfere (schematic representation; for symbols see Section 1). a, b, and c,
p ( T ) projections of the phase diagram; d, p(x) isotherm for T = const. = TI of binary
systems corresponding to type 2b or 2c

p ( T ) curves are often but not necessarily combined with azeotropy. Some
examples for the types 1 to 4 in Figure l b are given in Figures 22 to 24 and 31 ;
for further examples the reader is referred to the 3sss G Y * , 11,1S-Z1

Interference of Gas-Liquid and Liquid-Liquid Eguilibria. In Figure 2 schematic

p ( T ) projections are represented for phase equilibria in binary systems where
gas-liquid and liquid-liquid equilibria interfere. Such types of equilibria will
be of great importance to the discussion in Sections 3 and 4.
For the equilibria in Figure 2a the gas-liquid critical curve (LG) is not
interrupted and runs through the usual pressure maximum. At much lower
temperatures separation into two liquid phases (LL) is also found, the upper
critical solution temperature (UCST) rising steeply with increasing pressure.
No superposition of the critical phenomena for gas-liquid and liquid-liquid
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 109
equilibria occurs in systems of this type which have been found, e.g. for COz +
27

For systems belonging to type 2b the critical curve starts at the critical point
CPII of component 11, runs through a pressure maximum and ends at a so-
called critical end point B on the three-phase line LLG where two liquid phases
and one gaseous phase coexist. The branch of the critical curve starting from B
first corresponds to LCST's for liquid-liquid equilibria and then merges con-
tinuously into the gas-liquid critical curve; the critical end point B can even be
situated at higher temperatures than the critical temperature of pure component
I.28 Many examples are known of this important type, e.g. ethane + propan01,~S
ethane + 2,6,10,15,19,23-hexamethyltetracosane,30 CH, + he~ane,3~-~3pro-
pane +polyi~obutene,~~ and it has been extensively discussed by several authors.
Type 2c resembles type 2b but the branch of the critical curve starting from
CPII does not end at the critical end point B but runs through an additional
pressure minimum. This types was found by Kuenen for ethane +
methan01.I~
Being a transition to other types of phase behaviour it will be of great importance
in the discussion of Section 3 where other examples will also be given (e.g.
CHI + methyl~yclopentane).~~~~~ For all systems belonging to types 2b and 2c
the branch of the critical curve starting from CP I ends at the critical end point
C on the three-phase line LLG.
In Figure 2d a schematic p ( x ) isotherm at a temperature TI between CP I and
C is represented for type 2b and 2c systems. The p ( x ) isotherms exhibits two
critical points CP' and CP"; here CP" can be situated at fairly high pressures
(see Section 3C).
Critical Azeotropy. Systems with critical azeotropy form azeotropes up to the
critical region. The type shown in Figure 3a corresponds to a positive azeotrope
and has been found e.g. for binary mixtures of H,O (component 11) with
ethan01,3~1 -propan01,~~ or acetone 37 (component I).? Several p ( x ) isotherms are
schematically represented for this type in Figure 3b, the cusp-like isotherm for
T = const. = T3being especially interesting since here the limits of material
and mechanical stability coincide.s
26 G. Schneider, Ber. Bunsengesellschaft. Phys. Chem., 1966, 70, 10.
G. Schneider, Z . Alwani, W. Heim, E. Horvath, and E. U. Franck, Chem.-Ing.-Tech., 1967,
39, 649.
J. M. Prausnitz and P. L. Chueh, 'Computer Calculations for High-pressure Vapor-Liquid
Equilibria', Prentice-Hall, Englewood Cliffs, New Jersey, U.S.A., 1968.
28 J. P. Kuenen, Phil. Mag., 1903,6,637; Y . H. Ma and J. P. Kohn J. Chem. and Eng. Data, 1964,
9, 3.
ao J. S. Rowlinson and P. J. Freeman, Pure Appl. Chem., 1961,2,329; E. Horvath, University of
Karlsruhe, 1965, Diplom. thesis.
31 A. J. Davenport and J. S. Rowlinson, Trans. Faraday SOC.,1963, 59, 78.
32 G. Schneider, 2.phys. Chem. (Frankfkrt), 1965, 46, 375.
33 D. Oeder and G. M. Schneider, Ber. Bunsengesellschaft. Phys. Chem., 1969,73,229; 1970,74,

580.
P. J. Freeman and J. S. RowIinson, Polymer, 1960, 1, 20.
T. Kleinert, Angew. Chern., 1933, 46, 18; D. F. Othmer and S. L. Levy, Chem. Eng. Progr.
Symp., 1953, 49, 64; H. Otsuki and F. C. Williamson, ibid., p. 55.
36 F. Barr-David and B. F. Dodge, J. Chem. and Eng. Data 1959, 4, 107; J. P. Kuenen and
W. G. Robson, Phil. Mag., 1902, 3, 149, 622.
37 J. Grisu~oldand S. Y. Wong, Chem. Eng. Progr., Symp., 1952, 48, 18.

+
t For ethanol f H,O and acetone HzO, however, the azeotrope is limited at a minimum
temperature of 32 and 85 "Crespectively.
110 Chemical Thermodynamics
Type 6 in Figure l b corresponds to critical negative azeotropy and exhibits a
maximum temperature on the critical line (e.g. HC1 + methoxymethanes8). In
Figures 3c to 3h p ( T ) projections and p ( x ) isotherms are given for binary systems
showing positive heterogeneous azeotropes up to critical region. Here the three-
phase line LLG is situated at higher pressures than the vapour pressure curve

I. c PI1
b I

-1 - XI1

Figure 3 Critical phenomena in binary systems with positive homogeneous (a and b)

and heterogeneous (c to e and f to h) azeotropes (schematic representation; see text)

of the more volatile component I, and it arises from C being displaced to lower
or higher temperatures with respect to C P I leading respectively to type 3c to
3e (e.g. hexane +
H,O sB) or type 3f to 3h (e.g. ethoxyethane + H,O 40). For a
detailed discussion and further examples see Rowlinson,G Zernike,l9 and others.
38 J. P . Kuenen, 2.phys. Chem. (Leipzig), 1901, 37, 485.
9s F. E. C. Scheffer, Koninkl. Ned. Akad. Wetenschap., Proc., Ser. B, 1913-14, 16, 404.
40 F. E. C. Scheffer, Koninkl. Ned. Akad. Wetenschap., Proc., Ser. B, 1912-13,15, 380.
High-pressure Phase Diagrams and Critical Properties of F h i d Mixtures II1
In Figures 3c and 3f the branch of the critical curve starting from CPII (not
shown in the Figure) may exhibit very different shapes, e.g. with a temperature
minimum corresponding to a gas-gas equilibrium of the second type (see
Sections 2C and 3). Thus the types 3c to 3e and 3f to 3h are of considerable

a b

1 ! :I

Q1 Q2 4 1
F'igure 4 Interference of crystallization and gadiquid critical phenomena in binary
systems (schematic representation; for symbols see Section 1 ; Q1 = LGSISII;
+
Qz = LLGSr). a, Type found for H20 NaCI; by type found for H 2 0 + SO,;
c, crystallization behaviour in systems of type 2b such as proposed in the literature 6

interest within the scope of this Report since they represent the low-pressure
limit of most gas-gas systems (see Section 2C).
Interference of Solid Phases. In Figure 4 the p ( T ) projections are given for some
important cases where the gas-liquid critical line and the crystallization diagram
interfere in the p-T-x space.
For type 4a the three-phase line LGSn runs through a pressure maximum
between the triple point of the pure component I1 and the quadrupole point Q
where a liquid and a gaseous phase coexist with the pure solid components I
and 11. This curve represents the 'vapour pressure curve' above a solution
saturated with solid component 11; it can be determined quite easily by experiment
in a closed autoclave filled with component I and an excess of solid component
I1 and has been found for aqueous solutions of NaCl, boric acid, borates, borax,
etcP1
E. U. Franck, Angew. Chem., 1961,73, 309.
112 Chemical Thermodynamics
For type 4b the pressure maximum of the LGSII curve is so pronounced that
the critical curve LG is cut twice at the critical end points C and E; between C
and E pure solid component I1 is in equilibrium with a homogeneous, highly
compressed, fluid phase. This type has been detected in the H 2 0 + SiO, system,
the critical end point E being situated at 1080 “C,970 MPa, and 74 mass per
cent Si02.441 45

Type 4c corresponds to type 2b mentioned above, the critical end point B,

however, being situated below the crystallization surface in the p T - x space.
The critical curve already ends at the critical end point D on the three-phase line
LLSI (or LLSII), and the critical end point B can be reached only in (strongly)
undercooled mixtures. Type 2c systems have been shown to exhibit a quite
+
similar phase behaviour e.g. CH4 heptane 52 (see Figure 31); here, of course,
no critical end point B exists in the undercooled region. Type 4c systems will
be important in the discussion of the limits of gas-gas equilibria in Section 2C.
B. Liquid-Liquid Equilibria.-For the pressure dependence of liquid-liquid
equilibria, only the classical work of Kohnstamm, Timmermans, and their
co-workers, and of a few other authors existed up to 1963; with only a few
exceptions their measurements were limited to the pressure range below
20 MPa.17- 47 New measurements of the effects of high pressures on liquid-
46s

liquid immiscibility were started in 1963; pressures up to 700 MPa were used
and some hitherto unknown types of pressure dependence of immiscibility
phenomena were observed.f-St * 6t

Results on Binary Systems. In Figure 5 all types that are presently known for
binary systems are schematically represented ; examples exist for all types
except those marked with a question-mark. On the left-hand side the differ-
ent types of temperature-composition diagrams for constant pressure are
represented, with upper critical solution temperatures (UCST; Figure 5a),
lower critical solution temperatures (LCST; Figure 5b), closed loops (Figure
5c), and a hyperbolic-like type (Figure 5d). The first horizontal row (Figures 5i
to 51) shows that with increasing pressure UCST’s will either decline (Figure 5i)
or rise (Figure 5j) or run through a temperature minimum (Figure 51). The
second horizontal row (Figures 5m to 5p) shows that with increasing pressure
LCST’s may either rise (Figure 5m) or decline (Figure 5n) or run through a
temperature maximum (Figure 5p). The pressure dependence of closed loops
(Figures 5q to 5t) is schematically shown in the third row: closed loops may
either shrink with increasing pressures and disappear completely at a definite
pressure in the three-dimensional temperature-pressure-composition space
(Figure 5q) or appear only at higher pressures such as shown in Figure 5r (so-
called high-pressure immiscibility) or resemble hyperboloids (Figure 3 ) .
Immiscibility surfaces of a saddle-like type such as those shown in the fourth
row in Figures 5u and 5v have also been found.
4a E. U. Franck, Ber. Bunsengesellschaft. Phys. Chem., 1969,73,135.
43 M. Sanchez and H. Lentz, High Temperature-High Pressure, 1974, 5, 689.
44 G. C. Kennedy, Econ. Geol., 1950, 45, 629.
45 G. C. Kennedy, G. J. Wasserburg, H. C. Heard, and R. L. Newton, Publ. No. 150, Institute
of Geophysics, U.C.L.A., Cal., U.S.A., 1960.
46 J. Timmermans, J . Chim. Phys., 1923, 20, 491.
’’ E. Biichner, ‘Systerne mit zwei fliissigen Phasen’, in ref. 16 (Part 2).
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 113

' p =const I k
p -->
1
=rl

THP

d
THP

-X
Figure 5 Pressure efects on immiscibilityphenomena in liquid binary systems (schematic
representation;see text)

a b

t / O C

1I
10

- 65
8

-75 6

4
- 85

2
-95
I -a6 oc

Figure 6 Pressure dependence of liquid-liquid equilibria in the system C,H, + CHF,.'*

a, T(x) isobars; b, p(x) isotherms
114 Chemical Thermodynamics

c .

0 2.L,6 -TMP 200 400

+ HO
,
__t p / MPa
Figure 7 Pressure dependence of upper (UCST) and lower (LCST) critical solution
tetnperatures?S ' a, Pressure dependence of upper solution temperatures for
38

x = const. z x&, at 0.1 MPa (K = appearance of a solid phase); b, pressure

dependence of lower solution temperatures for x = const. z xkCSTat 0.1 MPa
(DMP = dimethyipyridine, TMP = trimethylpyridine)
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 115
In Figures 5e to 5h pressure-composition diagrams for constant temperature
that belong to the types of pressure dependence shown in Figures 5i to 5v are
represented schematically. Their shapes correspond remarkably well to the
temperatur-omposition diagrams for constant pressure given in Figures 5a
to 5d.
Examples for all types represented in Figure 5 have been summarized else-
~ h e r e ; l - only
~ a few examples will be given here. In Figure 6 some recent

p/MPa
Figure 8 Pressure influence on the solution temperatures of closed solubility loops for
x = const. zi xbsr at 0.1 MPa 1p (HP = hypercriticaz solution point)
39

+
results of Peter 48 are shown for the CzH6 CHF3 system that corresponds to
the type 5j: in Figure 6a isobaric T(x) cuts exhibiting upper critical solution
temperatures, and in Figure 6b isothermal p ( x ) cuts with lower critical solution
+
pressures, are represented. The systems CF4 CHF3 and CF4 + CzH, have
been shown to belong to the same type.48 In Figure 7 some results are given for
the pressure dependence of UCST’s (Figure 7a) and LCST’S (Figure 7b).l In
Figure 8 some examples for type 5q are plotted: with increasing pressure the
loops become smaller and finally disappear at a so-called hypercritical solution
point (HP) in the p-T-x space where dTC/dp = fa;at higher pressures the
constituents are miscible in all proportions.
In Figure 9 the pressure dependence of the immiscibility phenomena in
methylpyridine + water systems 1-5s 49 is shown; they belong to the types of

+
Figure 5r and 5t. 2-Methylpyridine heavy water exhibits a closed loop at
normal pressure that disappears with increasing pressure at approximately
** K.Peter, Ph.D. thesis, University of Bochum, Germany, 1974; K. H. Peter, R. Paas, and
G. M. Schneider, J . Chem. Thermodynamics, 1976, 8, 731.
G. M. Schneider, 2. phys. Chem. (Frankfurt), 1963, 39, 187.
116 Chemical Thermodynamics

0 300 600 0 3 00 600

.
0
0
c

I I
0 300 600

>~- plMPa

Figure 9 Influence of high pressure on liquid-liquid immiscibility in binary mixtures of

rnethylpyridines with H 2 0 and D20: T(p) curves for constant mass fraction ( w = mass
fraction of H 2 0 or D20; see text) 49

4 P
Figure 10 Thermodynamic conditions for the pressure dependence of liquid-liquid
equilibria (see text) 1s 2*
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 111
20 MPa (Figure 9b). With further increasing pressures, however, a second
immiscibility surface is reached at approximately 200 MPa that is completely
separated from the first ; the phrase ‘high-pressure immiscibility’ has been
proposed for phase separation effects of this type. For binary mixtures of 2-, 3-,
and 4-methylpyridine with normal water (Figures 9a, c, e) and 4-methylpyridine
with heavy water (Figure 9f) the immiscibility surface at low pressures is
+
apparently displaced to negative pressures. For 3-methylpyridine heavy water
(Figure 9d) the mutual miscibility is lowest and the two immiscibility surfaces are
no longer separated but show a tube-like shape in thep-T-x space.
Thermodynamic Description. As shown by several authors 1s 6 s *s 23s 24s 138 6o
259

some knowledge of the thermodynamic mixing and excess functions can be

deduced from the demixing behaviour. According to a classical description the
following relations hold for critical solution points :

UCST: r ~
(gm)c =
2Hg
< O, ) ~
LCST: (%)c = (z)c > O,

and for the pressure dependence of critical solutions temperatures Tc (UCST’s

and LCST’s)

Here Vm, Hm, and Sm denote the molar volume, enthalpy, and entropy of the
binary mixture, V g and H g the molar excess volume and enthalpy of the binary
mixture, and TC the critical solution temperature; the index c means ‘at the
critical solution point’.
The relations (la) and (2a) hold only with the additional (very constricting)
condition that the isothermal H g ( x ) curves for a given pressure p are free of
inflection points, and for relation (3a) a simple Porter-type equation such as
Zg = A(T,p) (1 - x ) x is assumed for HE and V g .
In Figure 10 the predictions are summarized for all types of pressure dependence
of UCST’s, LCST’s, and closed loops in Figures 5i to 5t (except 5k, 50, 5s);
similar predictions for the types 5k, 50, 5s, 5u, and 5v are straightforward.
The thermodynamic behaviour of type 5t systems is especially interesting.
With increasing temperature at constant pressure a2Hg/ax2has to change sign
from plus to minus (or with the simplifying conditions mentioned, H E from
6O G. Rehage, 2. Naturforsch., 1955, 10a, 300.
118 Chemical Thermodynamics
minus to plus) and with increasing pressure at constant temperature a2Vg/ax2
has to change sign from plus to minus (or V z from minus to plus). These
findings have already been confirmed by experiment, e.g. for the system
3-methylpyridine + H20.61

43c

Y
"'

i 360

320
0 200 400
->pl MPa
Figure 11 T C ( p )curves for systems of type 5t (see also Figure 9) calculated from
equation (4).' IT%csT(O.~MPa) = 403 K; T k ~ ( 0 . 1MPa) = 322 K; u o l / ( 2 ~ ~ =
,)
20 K2 MPa-I; I, aO2= 0; 11, 11', aO2/(2RE1)= -0.05 K2 MPa-2; 111, III',
(2RE1) = -0.07 K2 MPa-2; IV, IV', ~02/(2RE1)= -0.06089 K2MPa-2; the data
correspond co the system butoxyethanol +
H20]

Equations for the excess functions that are consistent with the thermodynamic
relations (1) to (3a) and the types of immiscibility behaviour in Figure 5 have
been discussed in detail e1sewhere.l As an example TC(p)curves are plotted in
Figure 1 1 that have been calculated for the different types of closed-loop systems
in Figure 5t from
h ( ( 1 - 4 l x I = A(T,P)
1 - 2x RT

= 2 +

where x is the mole fraction of component I1 at the boundary of the miscibility

gap at temperature T and pressure P ; T%csT(O.~ MPa) and TEcsT(O.l MPa) are
the upper and lower critical solution temperatures at 0.1 MPa, and El, a,,, and
61 P. Engels, Ph.D. thesis, University of Karlsruhe, Germany, 1970; P. Engels and G . hi.
Schneider, Ber. Bunsengesellschaft. Phys. Chem., 1970,74, 93 1.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 1 19
aO2are constants. Equation (4) corresponds to a Porter-type equation for the
molar excess Gibbs energy G g of the general form
GE = A ( T , p ) ( l - X)X
= (aoo + aloT + a20T2+ aoG + aozp2)(1- x ) x .
For an extension to unsymmetric miscibility gaps and other types of equations
see reference 1.
a
T

4
L+L L+S,,
rtr . .-._.
.- .-.-. -. ._,
1, SPS,
I I1
-+In

I n 1 n
-In -XU
Figure 12 Interference of solid-liquid and liquid-liquid equilibria in binary systems
(schematic representation: see text: Ttr = triple point temperature LLS ;T, = quadruple:
point temperature LLSS)

The classical description is based on the assumption that Gm is an analytic

function of the mole fraction x and the temperature T for a fixed pressure p at
and near the critical point of the mixture. In spite of rather good qualitative
agreement with experimental results there is more and more evidence that the
classical theory seems to be quantitatively inadequate at critical solution points,
analogous with the failure of the classical theory at the critical point of a single
component 6 s l3

More experimental and theoretical work, which should include the new types
of phase behaviour recently found at high pressures,? seems to be necessary in
order to attain further progress on these problems.
Efect of Solid Phases. The crystallization behaviour of mixtures under high
pressure may be very complicated, e.g. owing to the formation of solid com-
pounds or mixed crystals. Here only the solidification of demixing binary
f For example, the exhibition of very small and infinitely small closed loops (near and at the
points HP in Figure 8; see also Figures 5q, 5r, 5t, 9, and lo), the transition type IV-IV' in Figure
11, the extreme values for Tc in TQ) curves (see Figures 51, 5p, 3, 7b, 9d, 10, 11) erc.
5
120 Chemical Thermodynamics
solutions where the formation of mixed crystals may be more or less neglected
will be discussed for some important and interesting cases.
The phase diagram for the normal case is well known and is represented
schematically in Figure 12a. The triple-point temperature Tb, has been determined

1 0
x (cyctohexane)<--

Figure 13 Crystallization in binary systems with liquid-Ziquid immiscibility underpressure;

T ( x ) isobars for cyclohexane +
methanol ( x = mole fraction) 62

as a function of pressure for several systems up to several hundred MPa.1-3p se 59

In some cases the liquid-liquid immiscibility range disappears below the

crystallization surface in the p T - x space, e.g. at the critical end point K in
Figure 7a; in Figure 13 some isobaric T(x) sections are represented for the
system cyclohexane + where the region of limited miscibility
disappears below the crystallization surface at approximately ,230 MPa and
92 0C.62
In Figures 12b to 12d a crystallization behaviour is shown that has been
found for aqueous solutions of some organic compounds (= component II),
6a K. Roth, G. Schneider, and E. U. Franck, Ber. Bunsengesellschaft.Phys. Chem., 1966,70, 5 .
 High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 121
e.g. acetonitrile, aniline, 1,2-dimethylben~ene.~~
Here the triple-point temperature
Gr decreases with increasing pressure whereas the eutectic temperature TE
increases with increasing pressure. At a pressure pq the temperatures & and TE
coincide at a quadruple-point temperature T , where two liquid phases, solid
component I, and solid component 11, coexist [for H20(I) + acetonitrile(I1)
p q = 124 MPa, t, = -24.3 OCS3].

- 60

Y
I
-120 CHF3(l)+CF4(2)

l
I
%
< 100 100 MPa
-70 3
0
LL
7-130.

110.
- 80

120.

Figure 14 Liquid-liquid equilibria in the ternary system C H F , +

CF, + C2H6:
solution temperature T as a function of x3 for xJx2 = const. z (xp/xi)bb at 0.1 and
100 MPa 48* 64

E'ect of a Third Component. Liquid-liquid equilibria in ternary systems have

been investigated most frequently at normal or low pressure, but for high pressures
experimental data are very scarce.
As an example Figure 14 shows the phase behaviour of ternary mixtures of
CF4 + CHF, + C2H6 according to measurements of Peter4* and P a a ~ .All ~~
three binaries exhibit miscibility gaps with UCST's. The ternary system belongs
to the so-called col-subcol type (see Francis 66). In Figure 14 the upper solution
temperature T is plotted as a function of x3 for small values of xs at 0.1 and
100 MPa and for (x,/x,)~,, = const. z (Xg/Xg)bin. All solution temperatures
in Figure 14 decrease by addition of small amounts of a third component, even
in the CHF, + CF4 system that shows the best mutual miscibility; this behaviour
~ ~ CHF, + CF4
can be qualitatively understood from a theory of P r i g ~ g i n e . ~For
with CaH6 added dT/dp is as small as 3 x K MPa-l (Figure 14a) and for
the two other systems is approximately 2 x 10-1 KMPa-1 (Figures 14b and
14~).~~
6a G. Schneider, 2.phys. Chem. (Frankfurt), 1964,41, 327.
R. Pass, K. H. Peters, and G. M. Schneider, J. Chem. Thermodynamics, 1976, 8, 741.
66 A. W. Francis, 'Liquid-Liquid Equilibria', Interscience, New York, 1963.
122 Chemical Thermodynamics
For aqueous solutions of organic compounds the influence of added salts on
the mutual miscibility has also been studied under pressure in some cases.
Figure 15 shows results obtained by Engels 56 for the 3-methylpyridine H 2 0
1p +
system with addition of potassium iodide and sodium sulphate, the mass ratio of
water to 3-methylpyridine being kept constant at a value of 70/30. The high-
pressure immiscibility of the salt-free system (see Figure 9c) is displaced to

0 100 200

>- p/MPa

Figure 15 Salt and pressure efect on liquid-liquid immiscibility in the system 3-methyl-
pyridine +
H,O [w(3-MP)/w(water) = 30/70;M = mol dm+; w = ~ussfruction]1, 5

higher pressures by the addition of a salting-in salt, e.g. sodium iodide, or

displaced to lower pressures by a salting-out salt, e.g. sodium sulphate. At a
sulphate concentration of 0.007 M the hypothetical immiscibility surface in the
negative pressure region has already reached positive pressures. With further
increase of salt content there is a continuous transition to the tube-like type of
the 3-methylpyridine + heavy water system; at a sulphate concentration of
0.032 M approximately the same solution temperatures are obtained as a
function of pressure as for the binary system 3-methylpyridine + heavy water
(see Figure 9d).
A very similar diagram is obtained by starting from the 3-methylpyridine +
HaO binary and replacing stepwise more and more H 2 0 by D 2 0 (Figure 16).49
Thus Figures 15 and 16 demonstrate the high degree of similarity between all
types of phase behaviour found for methylpyridine + water systems (see Figure 9).
Experiments under high pressure have shown that there are evident analogies
between the influence of pressure and of the amount of a third component added
66 P. Engels, diplomarbeit, University of Karlsruhe, Germany, 1966.
 High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 123
I

4 p/MPa
Figure 16 Pressure dependence of the liquid-liquid phase separation in the ternary
+ +
system 3-methylpyridine H 2 0 D204g for lo2w(3-MP) = 30 (curves I to V ) or
lo2w(3-MP) = 27.5 (curve VI): curves I , r = 0 (pure D20), curves 11, r = 1 ; curues
111, r = 3 ; curue IV, r = 4, curue V , r = 5 ; curve V I , r = co (pure H20) [where
r = w(H20)/w(D20);w massfraction]

(e.g. a dissolved salt) on the immiscibility phenomena in binary liquid systems,

An example is given in Figure 17 where upper and lower solution temperatures
are plotted for the system 1-propanol + water + potassium chloride according
to measurements by RUSSO,~' the mass ratio of water to l-propanol being kept

1 1 1
m(KCl)/~lH201=0.12E
p =0.1MPa
I
80-i-7?7 80
%
1
I % s.E,
y 40-1
: I I
2 liquid phases

9.3)"
9' 0
\
'.r
Lo- 2 liquid phases

Nk
fl
I 0 ,
xLL3'
I
1 liquid phase
l l 1
f
0
xdk>
I
1 liquid phase
I

67 G . M. Schneider and C. Russo, Ber. Bunsengesellschaft. Phys. Chem., 1966, 70, 1008.
124 Chemical Thermodynamics
constant at a value of 1.5. Figure 17a shows the solution temperature as a
function of the molality of potassium chloride at normal pressure, and Figure 17b
as a function of pressure for a constant salt content of 12.5 g potassium chloride
per 100 g water. There is a striking analogy between both curves, increasing
pressure showing the same effect as decreasing salt content. The phase behaviour
of aqueous solutions of non-electrolytes with added salts under high pressures,

a b

I
P

LLG 1-

. -
Figure 18 Gas-gas equilibria in binary systems (schematic representation; see text)

however, is not always as simple as shown in Figures 15 and 17; on the contrary
it can be extremely c o m p l i ~ a t e d . ~ ~
The rate of liquid-liquid phase separation at pressures up to 150 MPa had
been determined in pyridine +H,O + KCl mixtures by Jost by a temperature
jump method originally designed for the investigation of relaxation kinetics
under high pressure. Full turbidity was reached in milliseconds indicating that
diffusion-controlled processes are important in the formation of the two liquid
phases and that inhibition times greater than 1 ms are absent; the method can
also be applied to investigations at normal pressure.69
C. Gas-Gas Equilibria.-The term ‘gas-gas equilibria’ or ‘immiscibility of
gases’ has been attributed to some phase-separation effects in fluid mixtures at
high pressures and temperatures, especially above the critical temperature of the
less volatile s~bstance.l-~*
I n Figure 18a the different types of gas-gas equilibria are schematically given
for binary mixtures. The critical curve is interrupted and consists of two
branches. The branch starting from the critical point of the more volatile
component I ends at a critical end point C on the three-phase line LLG where
two liquid phases and one gaseous phase are in equilibrium, whereas the branch
beginning at the critical point of the less volatile component I1 either immediately
68 J. P. Novak and G. M. Schneider, Ber. Bunsengesellschafr. Phys. Chem., 1968, 72, 791;
Y . Hirose, P. Engels, and G. M. Schneider, Chem.-Ing.-Tech., 1972, 44, 857.
59 A. Jost and G. M. Schneider, J . Phys. Chem., 1975,79,858; U. Limbach, A. Jost, and G. M.
Schneider, J. Phys. Chern., 1976, 80, 1952.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 125
tends to higher temperatures and pressures (curve 1 ; so-called ‘gas-gas equilibria
of the first type’) or goes through a temperature minimum first and then runs
steeply to increasing temperatures and pressures (curve 2 ; so-called ‘gas-gas
equilibria of the second type’). A schematic three-dimensional representation of

0 20 40 60 80 100
4102r(Xe)

Figure 19 Gas-gas equilibria of the first type: p ( x ) isotherms of the system helium +
xenon according to de Swam Arms and Diepen 66 (1 atm x 0.1013 MPa)

these types of phase behaviour is given in Figure 18b. On the right-hand side of
the critical curve in figure 18a or of the two-phase surfaces in the p-T-x space in
Figure 18b the components I and I1 are completely miscible in all proportions.
These separation effects were predicted by van der Waals60 from the fold
theory and were discussed in detail by Kamerlingh Onnes and Keesom61 in
1907. Type 2 was found for the first time in 1940 by Krichevskii and co-workers 62
for nitrogen +
ammonia and in the meantime by different authors for about
60 J. D. Van der Waals, Zittingsuersl. Koninkl. Akad. Wetenschap. Proc. Ser. B, 1894, 133 (see
also J. De Boer, Physica, 1974, 73, 1).
H. Kamerlingh Onnes and W. H. Keesom, Comm. Phys. Lab. Univ. Leiden, Suppl. No. 15,
1907; Koninkl. Ned. Akad. Wetenschap. Proc. Ser. B, 1907, 9, 786; 1907, 10, 231.
I. R. Krichevskii, Acta Phys. Chim. U.R.S.S., 1940, 12, 480; 1. R. Krichevskii and P. E.
Bol‘shakov, Z h u r . 8 ~Khim.,
. 1941,15, 184; I. R. Krichevskii and D. S. Tsiklis, ibid., p. 1059;
1943, 17, 126; D. $. Tsiklis, Doklady. Akad. Nauk S.S.S.R., 1952, 86, 993.
126 Chemical Thermodynamics
fifty other systems; an especially remarkable example is Ne +
Kr investigated
by Trappeniers and Schouten 63 for which some characteristic p(x) isotherms are
represented in Figure 20. Type 1 was reported for the first time in 1952 by

-105.90 "C

a
d
2
1

I 160

12c

80

40
149.98

Figure 20 Gas-gas equilibria of the second type: p(x) isotherms of the system neon +
krypton according to Trappeniers and Schouten 63

Tsiklis for the system helium +

ammonia64and since that time by different
authors in about twenty other helium binaries and additionally in the system
Ar + H20;65 one of the most interesting systems of this type is helium + xenon
studied by de Swaan Arons and Diepen 66 for which some p(x) isotherms are
given in Figure 19. Detailed discussion, a list of the investigated systems, and
63 N. J. Trappeniers and J. A. Schouten, Physica, 1974, 73, 527, 546.
64 D. S . Tsiklis, Doklady Akad. Nauk S.S.S.R., 1952, 86, 1 1 59.
H. Lentz and E. U. Franck, Ber. Bunsengesellschaft Phys. Chem., 1969, 73, 28.
66 J. De Swaan Arons and G. A. M. Diepen, J. Chem. Phys., 1966,44,2322.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 127
references can be f~und.l-~s lo Sometimes barotropism is found in such
5-79

systems and even ternary systems have been ~ t u d i e d . ~

In Figure 21 some especially interesting examples are given. The dashed lines
represent the vapour pressure curves of pure ammonia and pure water respectively.
Whereas NH, + H , 0 s 7 and H 2 0 + NaCls8 exhibit non-interrupted critical
curves, He + NH364 and Ar + H 2 065 show gas-gas equilibria of the first type,
and all other systems in Figure 21 show gas-gas equilibria of the second type.

-tl"C
Figure 21 Critical ~ ( 7 ' curves
) of binary H20 3- and NH3 + systems (for discussion
and references see t e x t ) 3n 4 * '

The high-pressure limits of gas-gas equilibria have been recently discussed by

Streett.68 In principle there is no limit of the gas-gas critical curve to higher
temperatures and pressures.? Since dT/dp happens to be larger for a three-
phase line fluid-fluid-solid than for the gas-gas critical curve the most probable
behaviour, however, will be that the region of gas-gas equilibria disappears below
the crystallization surface at high temperature and pressure and that the gas-gas
critical curve ends at a critical end point on the three-phase line GGS (preferably
GGSII) which is more or less an extension of the three-phase line LLS.$ An

13' 9.
D. S. TsiMis, L. R. Linshitz, and N. Gorjunova, Zhur. fiz. Khim., 1965, 39, 2978.
S. Sourirajan and G. C. Kennedy, Nuclear Explosions - Peaceful Applications', UC 35,
TID-4500, Report UCRL-6175, 1965.
6 s W. B. Streett, Canad. J. Chem. Eng., 1974, 52, 92.

t The hypothesis of Tammann that the critical curve of all gas-gas systems will pass through
a temperature maximum and a pressure maximum successively and will finally end on a three-
phase line LLG (so-called Tammann's egg) could not be confirmed by experiment. Up to now
no temperature maximum and even less any pressure maximum has been found on any gas-gas
critical curve.
3 This crystallization behaviour resembles the solidification phenomena in the cyclohexane +
methanol system (see Figure 13) or in the methane 3- methylcyclopentane system (see Figure 31
and type 3 in Figure 34d; for the gas-gas systems the critical curve, however, exhibits a positive
slope before ending at D).
128 Chemical Thermodynamics
example is the system helium + hydrogen which is of considerable relevance to
planetary physics.14 In contradistinction to this behaviour, gas-gas equilibria,
however, may also persist to very high temperatures at very high pressures,
perhaps even to the limits at which atomic and molecular structure begin to
break down (see Streett se)

100 200 300 490

vapor pressure curve of CO, :-t/OC

+
Figure 22 Critical p ( T ) curves for binary C02 alkane mixtures (Go= COz +
2,6,10,15,19,23-hexarnefhy/tetracosane,other curues C , = CO, + n-C,H,,+,) 2~ 81

Theoretical approaches to predict or explain gas-gas equilibria have been

reviewed elsewhere;2p lo for more recent efforts in this field see references 70
‘9

to 75. The lattice-model calculations of Bartis and Hall ’* do not depend on the
classical assumption of G, being analytic (see the thermodynamic discussion in
Section 2B). For thermodynamic conditions see references 2, 6, 7, 10, 13, 22,
and 99.
The terminology ‘gas-gas equilibria’ has been lengthily discussed in the
literature; now it seems to be more or less established. It has to be kept in mind,
however, that the coexisting phases in these kinds of phase equilibria are not
gases in the common sense but highly compressed fluids of generally liquid-like
densities.a*
70 M. Rigby, B. J. Alder, A. M. Sapse, and C. E. Hecht, J. Chem. Phys., 1970,52, 3665.
71 M. L. McGlashan, K. Stead, and C. Wan, ‘Proceedings of the 3rd International Conference
on Chemical Thermodynamics’, September, 1973, Baden, Austria, Section 4/11b.
72 N. J. Trappeniers, J. A. Schouten, and C. A. Ten Seldam, Chem. Phys. Letters, 1970, 5, 541.
79 J. A. Schouten, C. A. Ten Seldam, and N. J. Trappeniers, Physicu, 1974, 73, 556.
74 J. T. Bartis and C. K. Hall, Physicu, 1974, 78, 1.
75 S. Peter and H. Wenzel, Ber. Bunsengesselschuft Phys. Chem., 1972, 76, 331.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 129
3 Results on Selected Systems
In this section the relationships between the different types of phase behaviour
in fluid mixtures which were described in Section 2 will be discussed and it will
be shown with the help of data for selected binary mixtures such as families of
hydrocarbons with CO, (Section 3A), with H 2 0 (Section 3B), with CH4 (Section
3C), and for some inert-gas systems that there is a logical continuity between

n---
too
I TI i

1
U
1 phase
n_

+-
T

f 5c
LL D

->r/"C
Figure 23 p ( T ) projections of the phase diagrams of binaryfluid CO, 3- alkane mixtures
(solid line = critical curve of binary system; dashed line = vapour pressure curve of
pure component; C8 = octane, C13 = tridecane, CI6 = hexadecane, C3, = 2,6,10,-
15,19,23-hexamethyltetracosane) 39 8149

phase diagrams with gas-gas type critical curves and those with the more familiar
liquid-liquid and gas-liquid critical curves.
A. CO, + Hydrocarbon Systems.-The p ( T ) projections of the critical curves of
some CO, binaries with alkanes of very different chain length (between 1 and
30 carbon atoms) are given in Figures 22 and 23 according to data taken from
the literature. The curves exhibit a whole pattern of different types of critical
behaviour in binary systems and cover many of the types given schematically in
Section 2 such as type 1 in Figure l b [e.g. methane + C02,76CO, + butane,77
76 G. Donnelly and D . L. Katz, Ind. and Eng. Chem., 1954, 46, 511.
77 H. Poettmann and D.L. Katz, Ind. and Eng. Chem., 1945,37, 847.
130 Chemical Thermodynamics
U b C
r
Z 8 8 8
i - 7
r,
- 6 6
.\'- \ L 6
i ti
1 G+L '1.
T 4
4 . G+L '\ 4
I
i i
2 2 \ 2 i
0 0I * /
. 0.2 0 . 4 0.6 0 7 1 1 0 - ~ 1 01- ~0 - ~ 1 0 - ~ 1o0 - ~
w --+p (decane)/g ~ r n - ~ --+ P (decane)/g cm'3
x (decane)
Figure 24 Pressure-composition diagrams for the system C02 decane at 37.8 "C +
according to Reamer and Sage78 (x = mole fraction, p = mass concentration).82
a, Pressure as a function of mole fraction x ; b, pressure as a function of mass concen-
tration p ; c, pressure as a function of mass concentration p (logarithmic scale)

120
0
P
\

i '0°
80

60

LO

20

a
0
-
30 LO
lo2 w K ~ o l
60

Figure 25 Pressure-composition isotherms for the system C02 + 2,6,10,15,19,23-

80 100

hexamethyltetracosane!' (w = mass fraction; the low-pressure part of the isotherms

is omitted)
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 131
CO, +~ e n t a n e ,CO,
~~ + 27 CO, + decane 78 (see also Figure 24),

+
C 0 2 undecane 26s type 2 in Figure l b (e.g. CO, + propane type 4 in
+
Figure l b (e.g. CO, ethane *O), the type of Figure 2a (e.g. COB octane,26s27+
CO, +undecane 27) and the type of Figure 2c (e.g. CO, + hexadecane,2*127
268

CO, +2,6,10,15,19,23-hexamethyltetracosane81), The p ( T ) curves of CO, +

hexadecane and CO, + 2,6,10,15,19,23-hexamethyltetracosane run through
temperature minima (see Figure 23) and thus resemble the critical behaviour
of gas-gas systems of the second type (see Figures 18, 20, and 21).
Experimental p(x) isotherms for CO, + decane and CO, + 2,6,10,15,19,23-
hexamethyltetracosane are represented in Figures 24 and 25. For the CO, +
decane system 789 the 40 "C isotherm is plotted as a function of the mole
fraction x of decane (Figure 24a), of the mass concentration p (Figure 24b) and of
log p (Figure 24c). Figures 24a to 24c demonstrate that the mutual miscibility
of CO, and decane increases rapidly with increasing pressure, and becomes
unlimited at pressures above the critical pressure (here approximately 8 MPa).
This fact is important for the use of CO, as a mobile phase in supercritical fluid
chromatography (SFC)82 (see Section 4); one has, however, to keep in mind
that the critical concentration of component I1 and thus the solubility of
component I1 (here decane) in component I (here CO,) at subcritical pressures
can be extremely small for systems of this type. The&) isotherms of the system
CO, + 2,6,10,15,23,29-hexamethyltetracosanein Figure 25 show an interesting
retrograde solubility behaviour. At 53 "C the mutual miscibility first increases
and then decreases again with increasing pressure; for temperatures above 55 "C
even two completely separated miscibility gaps exist, a first at low and a second
at high pressures.81 An analogous phase behaviour has been found for COz +
hexadecane 27 (see Figures 22 and 23).
261

+
The phase behaviour of systems such as CO, hexadecane or CO, +
2,6,10,15,19,23-hexamethyltetracosanecan be easily deduced from a super-
position of liquid-liquid and gas-liquid equilibria. In Figures 26a and 26c
schematic representations of the two-phase surfaces in the three-dimensional
pressure-temperatureomposition space, and in Figures 26b and 26d the
corresponding p ( T ) projections of the phase diagrams, are given. The type on
+
the left-hand side (see Figures 26a and 26b) corresponds to COz octane. Here
the gas-liquid critical curve runs through the usual pressure maximum. At
temperatures far below the critical temperature of pure CO, separation into two
liquid phases take place, The branch LL of the critical curve corresponds to
upper critical solution temperatures (UCST) as a function of pressure which
increase slightly with increasing pressure. For mixtures of COP with smaller
alkanes the branch LL is situated below the crystallization surface.
The more the mutual miscibility of the two components decreases the more the
branch LL of the critical curve is displaced to higher temperatures. It can finally
penetrate the ranges of temperatures and pressures for the gas-liquid critical
phenomena and may pass continuously into the gas-liquid critical curve LG
H. H. Reamer and B. H. Sage, J. Chem. and Eng. Data, 1963,8, 508; 1965,10,49.
70 H. H. Reamer, B. H. Sage, and W. N. Lacey, Ind. and Eng. Chem., 1951,43,2515.
J. P. Kuenen, 2.phys. Chem. (Leipzig), 1897,24, 667.
81 K. G. Liphard and G . M. Schneider, J. Chem. Thermodynamics, 1975, 7 , 805.
82 D. Bartmann and G. M. Schneider, J. Chromatog., 1973, 83, 135.
132 Chemical Thermodynamics
whereas the branch of the critical curve starting from the critical point of the
pure CO, ends at a critical end point C on the three-phase line LLG (Figures 26c
and 26d). CO, + hexadecane and CO, + 2,6,10,15,19,23-hexamethy1tetracosane
belong to this type. [For the calculation of critical curves from equations of
state see e.g. refs. 28 and 86.1

P
t I I ,

I
b
P

1 J
+ T - + J
Figure 26 Schematic representation of the phase behaviour of binary CO, + alkane
systems in the three-dimensional p-T-x space (x = mole fraction; HC = hydrocarbon).
a, b, Type found for C 0 2 +
octane; C , d, type found for CO, +
hexadecane and for
CO, +
2,6,10,15,19,23-hexamethyltetracosane

For still lower mutual miscibility of the components critical p ( T ) curves

without any pressure maximum or minimum may be obtained, as indicated in
Figure 26d by a dotted line. Such a phase behaviour is attributed to gas-gas
equilibria (GG) (see Figure 18) and may be found in mixtures of CO, with
alkanes having more than 30 carbon atoms; it has already been found for the
CO, + water system 83 (see Figure 21).
Figure 27 summarizes the most important types of phase behaviour in CO,
binaries (see ref. 82). In Figures 27a to 27d the critical p ( T ) curves are represented
schematically; characteristic p ( x ) sections for constant temperatures (T17 Td
slightly above the critical temperature of pure CO, are given in Figures 27e to
27k. Type 1 in Figure 27a is characteristic for binary CO, systems with non-
polar components of low and medium molecular weight and with some polar
low-molecular-weight substances showing weak association in the liquid state.
This type has been found in binary mixtures of CO, with alkanes (see above),
cyciohexane, methylbenzene, some unsaturated straight-chain hydrocarbons,
methoxymethane, methanol, acetic acid, acetaldehyde, acetonitrile, propan-1 -01,
a3 K. Todheide and E. U. Franck, 2.phys. Chem. (Frankfurt), 1963, 37, 387.
 High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 133
pentan-1-01, methyl chloride, and c h l ~ r o b e n z e n e85. ~ ~Additionally,
~ at lower
temperatures, regions of liquid-liquid immiscibility can exist (e.g. for CO,
mixtures with normal alkanes from octane to undecane). Type 2 in Figures 27b
and 27f corresponding to the type of figure 2b has been found for mixtures of

1 phase

I Type 2

1, cons t Tl =consf

2 phases 2ph-es

'L
.-.-.-.

11
T1 = const 12=const. 1, r c o n s t . 12=const.
Figure 27 Phase behaviour of binary CO, systems (schematic representation; see text;
x = mole fraction). a to d, p(T)projections of thephase diagrams; e to k,p ( x ) isotherms

CO, with nitrobenzene, halogenated nitrobenzenes, o-nitrophenol, thymol, and

urethane; for mixtures of aromatic amines, a more complicated phase behaviour
has been found because of the appearance of solid phases and the formation of
solid For type 3 in Figure 27c and for types 4 to 6 in Figure 27d,
the right-hand branch of the critical locus shows different shapes according to
+
the decreasing mutual miscibility, e.g. for C02 hexadecane 26* 27 (type 3),
CO, + 2,6,10,15,19,23-hexamethyltetracosane (type 4), CO, + water 83 (type
84 'Gmelins Handbuch der anorganischen Chemie', Vol. 14, Part C1, 8th ed., Verlag Chemie,
Weinheim, 1970.
s6 R. A. Snedecker, Ph.D. thesis, University of Princeton, 1956.
134 Chemical Thermodynamics
5), and He +
C 0 2 6 4(type 6). It has already been shown in Section 2C that
types 5 and 6 are related to the so-called gas-gas equilibria of the first type
(type 6) and of the second type (type 5). Figures 21 to 27 demonstrate that a
whole pattern of rather different phase diagrams has been found in binary
mixtures of CO, with organic compounds (even for rather similar compounds

Figure 28 CriticaZp(T) curves ofbinary hydrocarbon

-
+
t / O c

water systems (1 = benzene +

H20;89 2 = benzene +
D20;88 3 = methylbenzene +
H,0;87 4 = 1,Zdimethyl-
benzene + H20;875 = 1,3,5-trimethylbenzene +
H20;876 = cyclohexane + H20;88
7 = ethane +
H20;91 8 = butane +
H20;919 = naphthalene + H,O (LL);87
10 = biphenyl + H,O (LL);" 11 = fiuorobenzene + H,O)

such as the series of alkanes) and that there is evidence for continuous transitions
between all these types. The solubility of organic compounds in CO, will
certainly be of increasing importance in the near future, e.g. for extraction
processes, chromatography, etc. (see Section 4).
B. Hydrocarbon + Water Systems.-In Figure 28 the critical phase behaviour
of binary aqueous solutions of several hydrocarbons and fluorobenzene is shown
according to data taken from the literature. The dashed curve is the vapour
pressure curve of pure water and the solid lines are the branches of the critical
curves of the binary systems that start from the critical point of pure water
CP(H20). Except for naphthalene + H20 and biphenyl + H 2 0 the critical
86 D. Zudkevitch and J. Joffe, Amer. Znst. Chem. Engineers J., 1970, 16, 112; J. Joffe, G. M.
Schroeder, and D. Zudkevitch, ibid., p. 496; J. Joffe and D. Zudkevitch, Chem. Eng. Progr.,
Symp., 1967, 63, 43.
81 2. Alwani and G. M. Schneider, Ber. Bunsengesellschaft Phys. Chem., 1969, 73, 294,
88 K. Brollos, K. Peter, and G. M. Schneider, Ber. Bunsengesellschaft Phys. Chem., 1970, 74,
682.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 135
curves have qualitatively the same shape, with a temperature minimum and a
steep ascent to higher temperatures and pressures according to a gas-gas
equilibrium of the second type (see Figure 18). In Figure 29 some p ( x ) isotherms
are represented for the system cyclohexane + water 88 which are also characteristic

200

160
a
a
5Q

1 120

80

40

A , , , , , 1

20

Figure 29 p(x) isotherms for the system cyclohexane

40
-lo2+ 60 80
w (CgHi2)
H 2 088 (for t < 350 "C the
branches at low pressures are omitted; the thick solid line is the p(x) projection of the
critical curve; see analogous curves in Figures 19 and 20)

for the other systems. The curves for naphthalene +

H208' and biphenyl +-
H 2 0 8 *in Figure 28, however, have been shown definitely to belong to liquid-
liquid equilibria; they end on the three-phase line LLG but fit remarkably well
to the other curves at higher pressures.
In Figure 30 the phase behaviour of the systems in Figure 28 is deduced from
a superposition of liquid-liquid and gas-liquid Figures 30a and
30c correspond to a binary system with a gas-liquid critical curve LG showing a
temperature minimum and with a phase separation into two liquid phases at
lower temperatures, the UCST rising with increasing pressure. This type has
been found for the naphthalene +
H 2 0 s 7 and biphenyl H2OS8 systems. +
They are the first hydrocarbon +
H,O systems where complete miscibility in
all proportions has been found in ranges of temperature and pressure which
136 Chemical Thermodynamics
belong definitely to the liquid state; however, the gas-liquid critical curve LC
with a temperature minimum shown in Figures 30a and 30c has not yet been
determined experimentally.
The lower the mutual miscibility the more the critical curve LL in Figures 30a
and 3Oc is displaced to higher temperatures until finally the type of Figures 30b

a b

C
P

4 r
f
P

Figure 30 p-T-x surfaces and p ( T ) projections of phase diagrams for binary mixtures of
H,O with hydrocarbons (HC) (schematic representation; see text). a and c, Type found
for naphthalene +
H,O, biphenyl +
H,O; b and d, type found for benzene + H 2 0
and aqueous solutions of methylsubstituted benzenes; e, no aqueous hydrocarbon system
known; f, type found for cyclohexane +
HzO, butane H,O +
and 30d is observed, which corresponds qualitatively to the phase behaviour of
aqueous solutions of benzene,89alkyl-substituted benzene^,^' and fluorobenzene
(see Figure 28). For benzene + +
H20, fluorobenzene HzO, methylbenzene +
+
H20, lY2-dimethylbenzene HzO, and 1,3,5-trimethylbenzene + H,O the
origin of the steeply ascending branches from liquid-liquid equilibria seems to be
evident; the curves for naphthalene + +
H,O and biphenyl H,O that definitely
belong to liquid-liquid equilibria fit remarkably well to these curves. In Figures
30e and 30f is shown how the critical curves of systems such as cyclohexane +
H20,88 ethane + H2OYg1heptane +
H20,88 etc. that show typical gas-gas
equilibria of the second type, can be explained by the same hypothesis.
89 Z . Alwani and G. M. Schneider, Ber. Bunsengesellschaft Phys. Chem., 1967, 71, 633.
R. Jockers, Ph.D. Thesis, University of Bochum, Germany, 1976.
A. Danneil, K. Todheide, and E. U. Franck, Chem.-lng.-Tech., I967, 39, 816; K. Todheide,
Ber. Bunsengesellschaft Phys. Chem., 1966, 70, 1022.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 131
The qualitatively similar shape of all curves above 100 MPa demonstrates
that there are continuous transitions between all types of critical curves as shown
in Figure 28; additionally, a marked extractive effect of HzO on different hydro-
carbons is demonstrated.6 [For other examples see refs. 42 and 43.1
i5 0

a
Q
r
--.
9

1 100
1 phase

50

c
-150 CHL -100 -50 0 t50 +I00
__j tl"C
+
Figure 31 Phase behaviour of binary CH4 hydrocarbon systems 2, 329 33 (see text;
for CH, + propane 92 and CH4 i- rnethylcyclopentane the critical p ( T ) curves are
plotted; for all other systems p ( T) curves for 102w(CH4)= const. are given)

C. CH4 + Hydrocarbon Systems.-Liquid hydrocarbons of approximately

the same size are in general completely miscible in all proportions. For mixtures
of hydrocarbons differing considerably in size, however, immiscibility phenomena
have been found with the exhibition of LCST's, e.g. for methane with c 6 to C8
hydrocarbons, for ethane with C30to c36 hydrocarbons, and for hydrocarbon
solvents such as heptane, cyclohexane, or benzene with polymers (e.g. poly-
isobutene).2*
The decreasing mutual miscibility with increasing difference of chain length of
the constituents creates a series of very different types of critical behaviour,
e.g. for the methane +
hydrocarbon systems shown in Figure 31. The curves
plotted for CH, + propane 92 and CH, +
methylcyclopentane 33 are the critical
curves of these systems. The other curves in Figure 3 1 are p ( T ) cuts for constant
mass fraction of methane; they can be supposed to be also typical for the shape
of the p ( T ) projection of the critical curve of these s y ~ t e m s33. ~ ~ ~
Whereas CH4 + propaneg2exhibits a continuous critical curve equivalent to
type 1 in Figure lb, the phase behaviour of the system CH, +
hexane 31 corre-
sponds to the type in Figure 2b showing liquid-liquid immiscibility with the
92 W.W.Akers, J. F. Burns, and W.R. Fairchild, Ind. and Eng. Chem., 1954, 46, 2531.
138 Chemical Thermodynamics
exhibition of LCST’s. The critical behaviour of the system C H I +
methyl-
cyclopentane33corresponds to the type of Figure 2c. A temperature minimum
on the critical curve seems to be possible for this system, but a solid phase appears
at the critical end point D (84 MPa, - 143 “ C ) before this hypothetical tempera-
ture minimum is reached; this minimum was unattainable even in undercooled

I
25
- Hexene ( 4 8 . 9 )
A , CHL+l
0, CHL+l-Heptene (16.5)

20

a
0

-.
5
CL

t l5

10

CH, +
- tJ°C
Figure 32 p ( T ) curaes for x = const. in the binary systems CH, +
1-hexene and
l-heptene= (see text; compositions given as 102w(alkene); LS = liquid-
solid)

mixtures. Some T(x) isobars of this system are plotted in Figure 33 demon-
strating the existence of liquid-liquid immiscibility with the exhibition of LCST’s
and UCST’s; here the LCST’s are situated at higher temperatures than the
+
UCST’s. The system CHI methylcy~lohexane,~~ CH, +
isoprene,33 CH4 +
l-heptene33 (see also Figure 32), and CH, + heptane32s33have been found to
belong to a similar type; however, a solid phase appears before the pressure
minimum is reached in the CH, +heptane and the CH, +
2-methylhexane 33 329

systems. The system CHI + methylbenzene does not exhibit any extreme value
of pressure on the critical p ( T ) curve within the range of the experiments. The
critical behaviour of CHI + l-hexene33t93(see Figures 31 and 32) will be
discussed later.
 High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 139
Figure 34 presents a schematic compilation of thep(T) diagrams for the binary
CH4 systems given in Figure 31. Type 3 in Figure 34d corresponds to the
CH, + methylcyclopentane system already discussed. For a binary system with
even lower mutual solubility, the critical curve will be displaced to higher tempera-
tures and it may be possible that the critical curve no longer disappears below

10 w ( methyl cyclopent an el
Figure 33 T(x) isobars in the system CH, + methylcyclopentane 33 (thick soIid line =
T(x)projection of the critical curve)

the crystallization surface, but may be supposed to run through a temperature

minimum showing successively different types of gas-gas equilibria [Type 2,
1’ (e.g. He + CH3,94and 1 in Figure 34d]. With increasing mutual miscibility,
however, the critical curve of type 3 in Figure 34d will be displaced to lower
pressures and can cut the three-phase line LLG twice, e.g. at A and B in Figure
34c. The branch starting from A corresponds to UCST’s, whereas that starting
from B first corresponds to LCST’s for a liquid-liquid equilibrium (LL) and
merges then into the critical curve for the equilibrium gas-liquid (LG). Examples
+
for this type are CH, + l-hexene, CH, 2,3-dimethyl-l-butene, CH4 + 2-methyl-
+
l-pentene, and CH4 3,3-dimethyl~entane.~~p 95 If the branch of the critical
curve starting from A in Figure 34c lies at very low temperatures or below the
gs A. J. Davenport, J. S. Rowlinson, and G. Saville, Trans. Faraday SOC.,1966, 62, 322.
B4 W.B. Streett, A. L.Erickson, and J. L. E. Hill, Phys. Earth Planet Interiors, 1972,6, 69.
B6 R. L. Scott and P. H. van Konynenburg, Discuss. Faraday SOC.,1970, 49, 87.
140 Chemical Thermodynamics
crystallization surface, type 34b is found, e.g. CH, +
h e ~ a n e , ~ CH +
l - ~4 ~ 2,2,4-
trirnethyl~entane.~l-~~ For a further increase in mutual miscibility, no regions of
limited miscibility can any longer exist in the liquid phase and a phase diagram
+
of the simple type of Figure 34a is found, e.g. CHI propane.g2

Figure 34 p ( T ) projections of the phase diagrams of binary CH, systems (schematic

representation; see text; for symbols see Section 1 ; Q1 = SISIILG; Qa = SIILLG).
a, CH, + propane, CH, + pentane; b, CH, + hexane, CH, +
2,2,4-trimethyl-
pentane; c, CH, + 1-hexene; d, CH, + methylcyclopentane, CH4 + 1-heptene
(type 3 ) ; CH,+ +
methylbenzene (type 2'); CH4 NH, (type 2); He +CH4 (type 1');
no CH, system is known for type 1

It is evident that there are continuous transitions between all types of phase
behaviour shown in Figures 34a to 34d; the most important types (type of Figure
34c, type 3 in Figure 34d) are represented three-dimensionally in Figures 35a
and 35b. This continuity is equally demonstrated by the fact that many different
types of critical curves can be obtained from model calculations by varying
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 141
a b

- b T
Figure 35 Three-dimensional representation of the p-T-x surfaces or binary CH, +
hydrocarbon systems.2 a, Type found for CH, + l-hexene (corresponding to the type
- T'

represented in Figure 34c) ;b, type found for CH, +

methylcyclopentane (corresponding
to type 3 in Figure 34d)

parameters. Such calculations have been made by several authors, e.g. by Scott
and van Konynenburg with a modified van der Waals m0de1.l~~ 95

D. Inert Gas Mixtures.-Gas-gas separation effects have also been observed

in binary mixtures of inert gases.
In Figure 36 parts of the critical curves of some binary He systems and of the
vapour-pressure curves of the pure less volatile components are plotted making
use of a compilation by Streett.14~gsWhereas He Na,07 He +Ar,ge and +
-

860
0[

401
Het N2 He+ Ar

He t CHr, He+ CO;

-
He+ Ne

-
-+-4
2 8 32 36 40 44 120 124 150 192'290 294 298 302 306 310 3 4
TlK
Figure 36 Critical p(T) curves of some helium binaries according to a compilation of
Streett B6 (for references see text)

86 W. B. Streett, Trans. Faraday SOC.,1969, 65, 696; W. B. Streett and J. L. E. Hill, ibid., 1971,
67, 622.
B7 W. B. Streett, Chem. Eng. Progr., Synp., 1967, 63,no. 81, 37.
142 Chemical Thermodynamics
He +H2l4 exhibit gas-gas equilibria of the second type and He Ne 98 shows+
a certain tendency for such a phase behaviour, gas-gas equilibrium of the first
type has been found in the He + Xe system by de Swaan Arons and Diepen66
(see also Figure 19) and in the He + C02 system by T s i k l i ~ . The~ ~ system
He + CH, 94 belongs to an intermediate type. It has been shown by Trappeniers
and Schouten 63 that the critical behaviour of Ne +
Kr (see Figure 20) resembles
that of He + Ar. Other helium systems have been lo,96 It is ' 9

evident that there are continuous transitions between all types of critical curves
such as those represented in Figure 36.
The existence of gas-gas equilibria in mixtures of nonpolar substances and
even in rare gas mixtures is a very remarkable effect theoretically; some
theoretical treatments of gas-gas equilibria in inert gas systems have recently
been given. Interesting consequences for planetary atmospheres have been
extensively discussed by Streett.141 loo [For thermodynamic conditions see
969

ref. 101.1

4 Conclusions
In Section 2 the high-pressure phase diagrams and the critical phenomena were
briefly discussed for gas-liquid, liquid-liquid, and gas-gas equilibria, and the
discussion in Section 3 gave evidence that the hypothesis of continuous transitions
between all types of two-phase equilibrium in fluid mixtures is useful in under-
standing the great variety of phase relationships and in giving a certain order to
many different types of critical behaviour and phase diagrams in fluid mixtures.
Whereas gas-gas equilibria had been a curiosity of phase theory as lately as
10 years ago they have now proved to be as important as the classical types of
gas-liquid and liquid-liquid equilibria. They are not at all restricted to some
special cases but represent the normal type of two-phase equilibrium in systems
of components that differ considerably in size, shape, volatility, and polarity,
and consequently show a low mutual solubility even up to rather high tempera-
tures. Thus, fluid systems where the phase-separation effects have to be
attributed to the solubility of gas in a liquid lo2 (or of a liquid in a gas) at normal
conditions of temperature and pressure will frequently exhibit gas-gas critical
phenomena at higher temperatures; some examples for binary mixtures of He,
Nt, CHI, COz, etc. with organic liquids and liquid water are given in Sections 2
and 3.1-
O8 C. K. Heck and P. L. Barrick, Adu. Cryog. Eng., 1967,12,714.
D. S. Tsiklis, Doklady Akad. Nauk S.S.S.R., 1952, 86, 1159.
l o o W.B. Streett, H. J. Ringermacher, and G. Veronis, Icarus, 1971,14,319.
l o l R. L. Scott, Paper presented at the 3rd International Conference on Chemical Thermo-
dynamics, September, 1974, Baden, Austria, Section 4/1la.
loa R. Battino and H. L. Clever, Chem. Rev., 1966,66, 395.
lo3 A. S. Teja and J. S. Rowlinson, Chem. Eng. Sci., 1973, 28, 529.

f Characteristic isotherms are shown e.g. in Figures 19 (for T < T ~ I )24, , and 25 where the
right-hand branch of each isotherm corresponds to the pressure dependence of the solubility of
a gas (here He or CO, respectively) in a liquid (here Xe, decane, or 2,6,10,15,19,23-hexamethyl-
tetracosane respectively), and the left-hand branch to the solubility of the liquid in the coexisting
gaseous phase. The isotherms are of special interest for gas-solubility effects under very high
pressures where decreasing gas solubility with increasing pressure has frequently been found,los
such as in Figures 19 and 25.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 143
The rapid development of the experimental techniques 104-106 makes possible
the investigation of phase equilibria in fluid systems at high pressures, not only
at normal temperatures but over a large temperature range and in certain cases
under strong corrosion. Investigations of condensed gas mixtures at low tempera-
tures and eventually high pressures (see Sections 3C and 3D) will be interesting
from a theoretical point of view, since mixtures of small and highly symmetrical
molecules can be studied that are suitable for a theoretical treatment; measure-
ments at low temperatures may also have a certain practical and technical interest
for the petroleum and gas industries, for gas solubilities, cryogenic applications,
separation problems, space research, etc. Additionally phase equilibria,
miscibility reactions, solubilities, and even critical phenomena will be increasingly
studied in aqueous systems, molten salts and minerals, fluid metals, and other
substances at high temperatures and high pressures and it seems to be quite
probable that these phenomena will be rather similar to those discussed in
Sections 2 and 3 in view of the fact that all substances are in quasi-corresponding
states under critical conditions ; an example is the striking resemblance between
the critical phenomena in inert gas mixtures (see Figure 36) and in aqueous
mixtures of hydrocarbons (see Figure 28) within a comparable pressure range but
over as large a temperature range as more than 600 K.
All these measurements may have some practical or technological relevance to
mineralogy, geochemistry, and earth sciences (especially for problems connected
with hydrothermal syntheses and with the formation and migration of oil and
natural gases), to separation processes, reactor research, high-pressure chemistry,
desalination of salt water,lo7and other problems.
Solutions of high-boiling compounds (preferably organic substances with
low volatility and/or low thermal stability) in gases having critical temperatures
between 0 and 200 “C (e.g. C02, C2Hs,C3H8,NH3, etc.) will in particular be of
rapidly increasing importance, since such supercritical phases are promising
solvents for industrial extraction processes particularly in the pharmaceutical
and food industries; this is clearly demonstrated by an increasing number of
papers and patents in this field.10s-114
An interesting special application is supercritical fluid chromatography
(SFC). In this relatively new technique of partition chromatography, com-
pressed fluids at temperatures in the critical range are used as mobile phases
particularly for the separation of substances with too low a migration rate or
too low a thermal stability to allow separation by ordinary gas chromatography
lo4 D. S. Tsiklis, ‘Handbook of Techniques in High-pressure Research and Engineering’,
Plenum Press, New York, 1968.
lo6 G. M. Schneider, in ‘ExperimentalThermodynamics’, Vol. 2, ed. 13. Vodar and B. LeNeindre,
Buttemorth, London, 1975.
lo6 C. L. Young, see Chapter 3, p. 71.
lo7 Texaco Inc., U.S.P. 3 318 805/1967.
108 E. P. King, U.S.P. 1 166 160/1939.
loBT. P. Zhuze, Petroleum, 1960,23,298; T. P. Zhuze, G. N. Juskevic, and I. E. Gekker, Muslob.
Zhir. Prom., 1958, 24, 6.
T. P. Zhuze, Russ. P. 113 325/1958.
D. L. Katz and T. H. Whalley, U.S.P. 2 391 5761/945.
Studiengesellschaft Kohle mbH, Miilheim/Ruhr, Germany; Ned.€’. 6 404 127/1974; U.S.P.
880 475; G.P. 1 493 190.
113 0. Vitzthum and P. Hubert, G.P. 2 127 618/1971.
114 S. Peter, G. Brunner, and R. Riha, Dechema-Monogruphien, Vol. 73, Verlag Chemie, Wein-
heim, Germany, 1974, Chem.-Ing.-Tech., 1974,46, 623.
144 Chemical Thermodynamics
(GC).116-11sThe retention time tr of a substance under test is related to its
capacity ratio k and its partition coefficient K by

where

7-

Figure 37 Supercritical fluid chromatography (SFC) with COa as a mobile phase:

capacity ratio k as a function of density p and pressure p (see text; C;,H,, = decane,
G6H, = hexadecane, GoH12= eicosane, (&,H,,a = 2,6,10,15,19,23-hexarnethyl-
tetracosane) 11*

Here cStat and P O b are the concentrations of the substance under test in the
stationary and mobile phase respectively, Vsht and Fob are the volumes of the
stationary and mobile phase in the column, and t ois the retention time of an inert
substance. From measurements of tr and t o the capacity ratio k, and from this
the partition coefficient K , can be calculated. In Figure 37 experimentally
determined log k values are plotted against the density p and the pressure p of
supercritical carbon dioxide at 40 "C as a mobile phase for SFC experiments with
alkanes having between 10 and 30 carbon atoms; a chemically bonded stationary
phase (Carbowax 400 on Porasil C) was used and further details are given in the
caption of Figure 37.ll7, 11* The curves show that with increasing density
the retention time tr and the capacity ratio k decrease and consequently the
M. N. Myers and J. C. Giddings, Progr. Separ. Purif., 1970, 3, 133.
116 T. H. Gouw and R. E. Jentoft, J . Chromatog., 1972,68,303; Adv. Chromatog., 1975,13, 1.
11' D. Bartmann and G. M. Schneider, J. Chromatog., 1973, 83, 135.
11* U. Van Wasen and G. M. Schneider, Chromatographia, 1975,8,274; U. Van Wasen, Diplom.
thesis, University of Bochum, 1974.
High-pressure Phase Diagrams and Critical Properties of Fluid Mixtures 145
concentration of the substance under test in the mobile phase increases by factors
of ten (see equation 5). Thus, Figure 37 demonstrates particularly well the
increasing solvent properties of supercritical CO, and other fluid phases as a
function of increasing density and pressure.
The studies on phase relationships reviewed here must be completed in the
future by measurements of the thermodynamic and physicochemical properties
of fluid mixtures under high pressures. Thus, a large field for further investi-
gations and applications remains.
Note added in proof. Some selected references taken from the literature since
September 1974 are given where numerous additional citations can be found.
The most recent literature is regularly reported in the ‘Bibliography of High
Pressure Research’;llSfor safety problems see reference 120. Some related books
have appeared.lz1*lZ2 Considerable progress has been achieved in the theory of
critical p h e n ~ m e n a . l ~Review
~ - ~ ~ ~papers have been published on the pressure,
volume, and temperature properties of fluids (including critical properties of
fluid mixtures) (477 references), on high-pressure liquid-gas equilibria 129 (190
references), on critical phenomena in aqueous electrolyte solutions 130 (1 53
references) and on the high-pressure thermodynamics of fluid mixtures 131
(1 14 references). Phase equilibrium studies have been made on several selected
systern~;~ polymer
~ ~ - ~solutions
~~ have also been discussed. 13’ The calculation
1369

of phase diagrams has been treated by several a ~ t h o r s . ~ ~ Practical

* - ~ * ~ applications
llS L. Merrill, ‘Bibliography on High Pressure Research’, Brigham Young University, Provo,
Utah, Volume 10, 1977.
lao ‘High Pressure Safety Code’, High Pressure Technology Association, ed. B. G. Cox and G.
Saville, Miher & Sons Ltd., Brentford, 1975.
121 ‘Chemical Thermodynamics-HI’, IUPAC, ed. G. H. Findenegg. (Plenary lectures presented
at the 3rd International Conference on Chemical Thermodynamics, Baden, Austria,
September 1973.) Butterworths, London, 1974.
122 H. E. Stanley, ‘Introduction to Phase Transitions and Critical Phenomena’, Clarendon Press,
Oxford, 1971. H. E. Stanley (ed.), ‘Cooperative Phenomena near Phase Transitions. A
Bibliography with Selected Readings’, Massachusetts Institute of Technology, 1973.
123 C. Domb and M. S. Green, ‘Phase Transitions and Critical Phenomena’, Vols. 1 to 5,
Academic Press, London, New York, San Francisco, 1972 (start of series).
la4 R. B. Griffiths and J. C. Wheeler, Phys. Rev., 1970, A2, 1047.
la5 R. B. Griffiths and B. Widom, Phys. Rev., 1973, AS, 2173.
la6 R. B. Griffiths, J. Chem. Phys., 1974, 60, 195.
12’ M. 1. Guerrero and J. S . Rowlinson, J.C.S. Faraday Transactions, ZI, 1976,72, 1970.
la* D. R. Douslin, in ‘Thermochemistry and Thermodynamics’, MTP International Review of
Science, Physical Chemistry, Series Two, ed H. A. Skinner, Volume 10, Chapter 6, Butter-
worths, London and Boston, 1975.
la9 C. P. Hicks and C. L. Young, Chem. Rev., 1975,75, 119.
130 A. L. Horvath, Chem.-Ing.-Tech., 1975, 47, 859; 1976,48, 144.
lal G. M. Schneider, Pure Appl. Chem., 1976, 47, 277.
ls2 Z. Alwani and G. M. Schneider, Ber. Bunsengesellschaft Phys. Chem., 1976, 80, 1310.
133 E. U. Franck, H. Lentz, and H. Welsch, 2.phys. Chem. (Frankfurt), 1974, 93, 95.
lS4 R. Jockers, Ph.D. thesis, University of Bochum, Germany, 1976.
136 S. G. D’Avila, B. K. Kaul, and J. M. Prausnitz, J. Chem. andEng. Data, 1976,21,488.
136 R. Koningsveld, Brit. Polymer J., 1975, 7, 435; Makromol. Chem., 1973, 8, 197.
13’ G. Blaum and B. A. Wolf, Makromol. Chem., 1976, 177, 579; B. A. Wolf and G. Blaum,
Makromol. Chem., 1976,177, 1073.
G. J. F. Breedveld and J. M. Prausnitz, Amer. Znst. Chem. Engineers J., 1973, 19, 783.
la9 H. Hederer, S. Peter, and H. Wenzel, Chem. Eng. Sci., 1976, 11, 183.
140 J. Mollerup, Adv. Cryog. Eng., 1975, 20, 172.
141 L. J. Christiansen and A. Fredenslund, Amer. Inst. Chem. Engineers J., 1975, 21, 49.
Ira C. H. Twu, K. E. Gubbins, and C. G. Gray, J. Chem. Phys., 1976, 64, 5186.
143 U. Deiters and G. M. Schneider, Ber. Bunsengesellschaft Phys. Chem., 1976, SO, 1316.
 146 Chemical Thermodynamics
have been reviewed 14p concerning fluid e x t r a ~ t i o n146
, ~supercritical
~~~ fluid chroma-
tography (SFC) 147-149 and high-pressure liquid chromatography (HPLC).lS0
For an extensive compilation of experimental techniques see reference 151.
14* V. Pilz, Verfahrenstech., 1975, 9, 280.
145 R. Riha, Ph.D. thesis, University of Erlangen, Germany, 1976.
laa S. Peter, G. Brunner, and R. Riha, Chem.-Zng.-Tech., 1977, 49, 59.
147 R. E. Jentoft and T. H. GOUW, Adu. Chromatog., 1975, 13, 1.
14a E. Stahl and W. Schilz, Z. analyt. Chem., 1976, 280, 99; Chem.-Zng.-Tech., 1976, 48, 773.
14n E. Klesper and W. Hartmann, J. Polymer Sci., Part B, Polymer Letters, 1977, 15, 9.

lS0 L. R. Snyder and J. J. Kirkland, ‘Introduction to Modern Liquid Chromatography’, Wiley,

New York, London, Sydney, Toronto 1974.
151 B. Le Neindre and B. Vodar, ed., ‘Experimental Thermodynamics. Vol. 11. Experimental
Thermodynamics of Non-reacting Fluids’, Butterworths, London, 1975.
5
Mixtures Containing a Fluorocarbon
BY F. L. SWINTON

1 Introduction
The thermodynamic properties of mixtures containing an aliphatic or an alicyclic
fluorocarbon were reviewed by Scott in an article published in 1958 but actually
written some two years previously. No comparable review has appeared since,
although several recently published textbooks 2-4 and general review articles 5-7
on mixtures of non-electrolytes have sections devoted to fluorocarbon mixtures.
More recently Fenby * and Swinton have reviewed the properties of mixtures
containing an aromatic fluorocarbon as one component. The present article
attempts to review the existing literature on all classes of fluorocarbon-containing
mixtures and is based on a comprehensive literature survey extending from the
beginning of 1970 to late 1974 although many of the significant papers published
prior to 1970 are mentioned and thus appear as references.
The properties of fluid mixtures containing an aromatic fluorocarbon differ
completely in many respects from those of mixtures containing an aliphatic or
alicyclic fluorocarbon. It is therefore convenient to treat these two major classes
of mixture separately. The desire to interpret these extreme differences in thermo-
dynamic behaviour, at a molecular level, is one of the main reasons for the
continuing interest of many physical scientists in such systems.

2 Mixtures Containing a Non-aromatic Fluorocarbon

A. Liquid Mixtures.-Aliphatic and alicyclic perfluorocarbons first became
available in commercial quantities just over three decades ago. The impetus
towards their large-scale production came from the war-time Manhattan Project
when it was realized that the chemical stability and inertness of these materials,
combined with their unexpectedly low solubility in most other common organic
solvents, might prove useful in uranium isotopic separation processes based on
R. L. Scott, J. Phys. Chem., 1958, 62, 136.
J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, 2nd edn., Butterworths, London, 1969.
J. H. Hildebrand and R. L. Scott, ‘Regular Solutions’, Prentice Hall, New Jersey, U.S.A.,
1962.
J. H. Hildebrand, J. M. Prausnitz, and R. L. Scott, ‘Regular and Related Solutions’, Van
Nostrand Reinhold, New York, 1970.
R. L. Scott and D. V. Fenby, Ann. Rev. Phys. Chem., 1969’20, 111.
R. H. Stokes and K. N. Marsh, Ann. Rev. Phys. Chem., 1972,23, 65.
H. V. Kehiaian, in ‘Thennochemistry and Thermodynamics’, ed. H. A. Skinner (M.T.P.
International Review of Science), Physical Chemistry Series 1 , Butterworths, London, 1972,
Vol. 10.
D. V. Fenby, Rev. Pure Appl. Chem., 1972, 22, 55.
F. L. Swinton, in ‘Molecular Complexes’, ed. R. Foster, Paul Elek, London, 1974, Vol. 2.
147
148 Chemical Thermodynamics
UFB. In complete contradiction to the predictions of the existing liquid mixture
theory it was found that these fluorocarbons were only partially miscible with
hydrocarbons of comparable molar mass. The large volume of experimental
work that resulted from attempts to interpret this unexpected behaviour of
mixtures consisting of relatively simple non-polar molecules is covered compre-
hensively in Scott’s 1958 review.l
Binary mixtures of non-aromatic fluorocarbons with hydrocarbons are
characterized by large positive values of the major thermodynamic excess
functions @, the excess Gibbs function, HE, the excess enthalpy, SE,the excess
entropy, and VE, the excess volume. In many cases these large positive deviations
from ideality result in the mixture forming two liquid phases at temperatures
below Tgppr, an upper critical solution temperature. Experimental values of
the excess functions and of Tgpperfor a representative sample of such binary
mixtures are given in Table 1.
Table 1 Thermodynamicproperties of equimolar binary liquid mixtures containing
a non-aromatic fluorocarbon
-T - -GE- HE TSE VE Gpper
System K kJmol-l kJmol-I kJmol-l ~rn~rno1-l K Refs.
CF4 + CH4 110 0.36 (1.20) (0.84) 0.88 94.5 10, 11,
12
+ n-C,H,, 308
n-C6Fl, 1.329 2.157 0.828 5.38 296 13, 14,
15
+~C-GHi, 338
c-GF~ 1.35 2.07 0.72 - 316 16
c-CFS. GFI1 338 1.39 2.80 1.41 7.5 319 16
+ c-CH,*CgjH,,
The values of H Eand of TSEfor CF, + CH4shown in Table 1 were estimated
by combining the measurements of @ made by Thorp and ScottlO over the
temperature range 105.5 to 110.5 K with the recent value of this function
measured at 98 K by Simon and Knobler I2 and are thus subject to considerable
uncertainty. It is to be hoped that H E for this important binary system will be
measured by a direct calorimetric method in the near future.
The 1971 paper of Simon and Knobler described the only direct measurement
of an excess function for a mixture of this class to be made since 1970. The
exhaustive general reviews of Stokes and Marsh and of Kehiaian contain no
mention of such measurements. The virtual cessation of experimental work in
this area is due in part to the industry of Scott and his co-workers during the
fifties and sixties and in part to the fact that most binary mixtures of this type
exhibit great similarity in their thermodynamic behaviour. Equimolar mixtures
of alicyclic or aliphatic fluorocarbons with branched or unbranched aliphatic
hydrocarbons, with alicyclic or with aromatic hydrocarbons, or with other simple
lo N. Thorp and R. L. Scott, J. Phys. Chem., 1956, 60,670.
l1 1. M. Croll and R. L. Scott, J. Phys. Chem., 1958,62, 954.
l2 M. Simon and C. M. Knobler, J. Chem. Thermodynamics, 1971, 3, 657.
R. D. Dunlap, R. G. Bedford, J. C. Woodbrey, and S. D. Furrow, J. Amer. Chem. SOC.,1959,
81, 2927.
l4 A. G. Williamson and R. L. Scott, J. Phys. Chem., 1961,65,275.
l6 R . G. Bedford and R. D. Dunlap, J. Amer. Chem. SOC.,1958,80, 282.
D. E. L. Dyke, J. S. Rowlinson, and R. Thacker, Trans. Faruday Soc., 1959,55, 903.
Mixtures Containing a Fluorocarbon 149
non-polar molecules such as CC14 or Si(CH,), are found to possess an excess
Gibbs function of 1.0 to 1.5 kJ mol-1 and an excess enthalpy approximately
twice as large. All such mixtures exhibit a positive excess volume representing a
volume expansion of some 2 to 10 per cent and will separate into two partially
miscible liquid phases at temperatures below an upper critical solution tempera-
ture unless prevented from doing so by the formation of a solid phase. The size
of the excess functions and hence the tendency for phase separation are reduced
when a fully fluorinated aliphatic or alicyclic fluorocarbon is mixed with a
partially fluorinated hydrocarbon. A system composed of two perfluorocarbons
forms an almost ideal mixture. Mixtures of partially chlorinated fluorocarbons
with either fluorocarbons or hydrocarbons are also found to exhibit a decreased
tendency for phase separation. For example, no upper critical solution tempera-
ture is found for either CF4 + CFsCl or CH, + CF3C1.l7 Mixtures of fluoro-
chlorocarbons of low molar mass are often used commercially as refrigerants.
Vapour-liquid equilibrium data have been published recently for the systems
CHF3 + CF3Cl,18~1eCH2FCl + C2F4C12,20and CH2F2 + C02.21 The first
two systems were found to form minimum boiling azeotropes.
More recently Young 22 has measured the upper critical solution temperatures
of mixtures of n-C7Fl, with a series of n-alkanes (n-C, to n-CIS) and has shown
that, although solubility-parameter theory is unable to predict the experimental
critical temperatures with any accuracy, a modification of this theory due to
Reed 24 is more successful. Reed's improved theory takes into account the
23s

effect of the size and ionization potential differences between the two component
molecules and is thus similar in many respects to the theory of Hudson and
McCoubrey 2s discussed in the next Section.
The realization that there is little likelihood of further studies of these systems
resulting in the discovery of any new phenomena has removed much of the
incentive for additional experimental work even though experimental precision
and accuracy have improved considerably over the past decade. Modern high-
precision measurement of the excess functions of a few selected systems would
indeed be profitable, for such data are now required to an ever increasing extent
to test newly developed statistical theories of fluid mixtures.
It is now established both theoretically and experimentally that many thermo-
dynamic variables assume a simple power-law behaviour at or near critical
points in both pure and mixed fluids. The actual functional dependence of one
variable on another can be characterized by the so-called critical indices: a, fl,
etc. The critical index /3, for example, defines both the shape of the gas-liquid
coexistence curve for a pure fluid and the liquid-liquid coexistence curve of a
binary mixture in the vicinity of either an upper or a lower critical solution
temperature. The correspondence between critical phenomena in one-, two-,
l7 I. M. Croll and R. L. Scott, J. Phys. Chem., 1964,68,2853.
la J. V. Sinka, E. Rosenthod, and R. P. Dixon, J. Chem. and Eng. Data, 1970, 15, 73.
l9 F. P. Stein and P. C. Proust, J. Chem. and Eng. Data, 1971,16,389.
2o J. V. Sinka, J. Chem. and Eng. Data, 1970, 15, 71.
21 R. A. Adams and F. P. Stein, J. Chem. and Eng. Data, 1971,16, 146.
28 C. L. Young, Trans. Faraday SOC.,1969, 65,2639.
2a T. M. Reed, J. Phys. Chem., 1955,59,424.
24 T. M. Reed, J. Phys. Chem., 1959,63, 1798.
26 G. H. Hudson and J. C. McCoubrey, Trans. Faraday Soc., 1960,56, 761.
150 Chemical Thermodynamics
and three-component systems has been discussed recently by Scott.26 Blagoi
et aLZ7and Fannin and Knobler 28 have studied the effects of gravity-induced
density gradients on the shape of the liquid-liquid coexistence curve of
CF4 + CH4 mixtures. They have shown that these gradients greatly affect the
derived values of the critical exponent p . Gaw and Scott29 have studied
the behaviour of certain thermodynamic functions such as (a2V/3x2)T,,,,in the
neighbourhood of an upper critical solution temperature by measuring V E for
the binary system: n-C,F14 + n-C6H14as a function of mole fraction and at
temperatures to within K of the critical temperature.
B. Gaseous Mixtures.-The thermodynamic properties of a system in the
dilute gaseous state are a potentially fruitful source of information concerning
the intermolecular pair potential, u(r). Equation (1) is an exact relation for
spherically symmetric molecules between u(r) and By the second virial coefficient.

However, because of the relative insensitivity of B(T) to u(r) it is usually impossible

to arrive at a unique u(r) on the basis of second virial coefficients alone and other
data such as gas viscosities or collision cross-sections determined from molecular
beam experiments must be utilized to define u(r) unequivocally.
To date there have been only three experimental determinations in which B(T)
(and in some cases C(T) and D(T), the third and fourth virial coefficients) has
been measured for binary mixtures containing an aliphatic fluorocarbon where
the measurements are of sufficient precision and extend over a large enough
temperature range for meaningful conclusions to be made concerning the inter-
molecular potentials. These three sets of measurements are those of Douslin
+
et aL30 on CF4 CH4 from 273 to 623 K, those of Sigmund et aLsl on
CF4 + SF6 from 272 to 423 K, and those of Dantzler-Siebert and Knobler 32
on mixtures of the type n-CzF22+2+ n-CuH2v+a with x and y in the range 1 to 6.
The measurements of Douslin et al. are particularly valuable because the virial
coefficients of the two pure substances Bll and B22 and the interaction virial
coefficient B12were measured over such a range of temperature that it was possibIe
to make a direct determination of the Boyle-point parameters TB and
VB = (Td-B/dT)T=TB.BI1,BZ2,and BIZwere found to follow the same theorem
of corresponding states when T B and V B were used as the reduction parameters,
whereas this was not so when the more usual critical parameters, T C and V c
were employed. The experimental values of TB and V B are given in Table 2
together with the values calculated by assuming the Lorentz-Berthelot
combining rules :
TF2 = (TF1T&)'", (2)

2fi R. L. Scott, Ber. Bitnsengesellschaft Phys. Chem., 1972, 76, 296.

?' Y . P. Blagoi, V. I. Sokhan, and L. A. Pavlichenko, Zhur. eksp. teor. Fiz., 1970, 11, 291.
18 A. A. Fannin and C. M. Knobler, Chem. Phys. Letters, 1974, 25, 92.
29 W. J. Gaw and R. L. Scott, J . Chem. Thermodynamics, 1971,3, 335.
30 D. R. Douslin, R. H. Harrison, and R. T. Moore, J. Phys. Chem., 1967, 71, 3477.
31 P. M. Sigmund, 1. H. Silberberg, and J. J. McKetta, J. Chem. and Eng. Data, 1972, 17, 168.
32 E. M. Dantzler-Siebert and C. M. Knobler, J. Phys. Chem., 1971, 75, 3863.
Mixtures Containing a Fluorocarbon 151
Equations (2) and (3) are directly equivalent to the Lorentz-Berthelot rules in
their more familiar form using E and u, the well depth and collision diameter
associated with the intermolecular potential u(r):
-512 = (-511-522)1/2, (4)
u12 = (011 + U22)P- (5)
Also shown in Table 2 is the value of Tp2 estimated by assuming the Hudson and
McCoubrey combining rule:25

where 11,I2are the ionization potentials of the pure components.

Table 2 Boyle parameters for CF, + CH,

Interaction T~IK YB/cm3mo1-1
CH, + CH, experimental 509.3 54.34
CF, + CF, experimental 518.14 104.13
CF, + CH, experimental 467.0 78.75
CF, + CHI equations (2) and (3) 513.7 76.56
CF, + CH, equation (6) 496.0 -

This semi-empirical rule tries to account for the influence of differences in

size and in ionization potential on the intermolecular potential of the unlike
interactions although, as Douslin et al. point out, CF, and CH, possess similar
ionization potentials and so this is not a significant factor for this particular
fluorocarbon +
hydrocarbon pair. The use of equation (6) nevertheless leads
to a considerable improvement between the theoretical and experimental values
of TFz but there is still an unaccounted weakness in the unlike interactions. The
reality of this weakness is further emphasized by the fact that the experimental
value of VF2 is 2% larger than the value predicted by the Lorentz rule.
Deviations of real systems from the predictions of the Lorentz-Berthelot
relation can be accounted for empirically by incorporating the adjustable
parameters $, and x in equations (4) and (5) respectively. The values of these
parameters can be calculated explicitly for the present system: i$ = 0.9090 and
x = 1,0095. This is evidence of the most direct kind of the essential weakness of
the unlike interactions in such fluorocarbon + hydrocarbon systems.
In complete contrast the measurements by Sigmund e f aL31 of the second and
third virial coefficients of CF, +
SF6 showed that the Lorentz-Berthelot rules
predict the experimentally determined interaction parameters for the unlike
interactions within experimental error. Mixtures of spherically symmetric
fluorocarbons thus closely resemble similar hydrocarbon mixtures in this respect.
Lange and Stein33 reported measurements of the second and third virial
+
coefficients for CF3H CF, mixtures. A distinct weakness in the unlike inter-
actions was noted although no detailed calculations were made.
Knobler and his co-workers have measured the interaction virial coefficient,
B12, for mixtures of n-alkanes 34 and for n-alkane +
perfluoro-n-alkane
88 H. B. Lange and F. P. Stein, J. Chem. and Eng. Data, 1970, 15, 56.
84 E. M. Dantzler, C. M. Knobler, and M. L. Windsor, J. Phys. Chem., 1968,72,676.
6
152 Chemical Thermodynamics
mixtures35 by direct measurement of the excess virial coefficient 8 where
d = B12- (BI1 + B2,)/2. d was obtained by measuring the pressure change
that occurred when the two gases were mixed at constant temperature and
I n order to extract explicit values of B12from measurements of this
type reliable values of the virial coefficients of the pure components are required.
Knobler et al. noted that the virial coefficients of both the pure fluorocarbons
and pure hydrocarbons were well represented either by the empirical equation
of McGlashan and basically a three-parameter theorem of corre-
sponding states, or by using a three-parameter Kihara spherical-core potential,
where, in addition to E and u, the hard-core radius a is introduced as the third
adjustable parameter. The Kihara analysis was marginally more successful and
so this was used in the subsequent treatment of the interaction virial coefficients.
The experimental Blis for mixtures of n-alkanes and for mixtures of
perfluoro-n-alkanes were found to be predicted almost within experimental
error when the Lorentz-Berthelot combining rules were used to estimate the
potential parameters for the unlike interactions. This treatment was found to
fail completely for the fluorocarbon +
hydrocarbon mixtures. A value of
6 = 0.90 2 0.02 was required to ensure agreement between theory and experi-
ment, this 10% weakness in the energy of the unlike interactions agreeing well
+
with the analysis of Douslin et al. for CF4 CH,. These values of [ were found
to be in remarkably good agreement with values obtained by Scott by analysis
of the thermodynamic properties of the same mixtures in the liquid state using
solubility-parameter theory.
Use of the Hudson and McCoubrey combining rule together with the Kihara
potential yields an explicit expression for 4:

where pll = ull - 2all, etc. Dantzler-Siebert and Knobler found that the use
of equation (7) gave values of [ that were quite close to the experimental values
in some cases. Large uncertainties were introduced however, owing to lack of
reliable values for the ionization potentials of many of the perfluoro-n-alkanes.
In contrast to this relative success of the Hudson and McCoubrey combining
rule, Hicks and Young38found that the use of equation (7) with V c , the critical
volume, substituted for p3, completely failed to predict 6 for mixtures of per-
Auorocyclohexane + cycloalkanes. Hicks and Young derived values of f for
these systems from measurements of the gas-liquid critical temperatures of the
mixtures using the analysis developed by R ~ w l i n s o n*O. ~ Mixtures
~~ of this type,
being composed of pseudo-spherically symmetric molecules, should be more
amenable to theoretical treatment than mixtures of long-chain species. Although
the ionization potential for cyclo-C,F12 is not known, which prevented the theory
being tested rigorously, because the critical volumes of the various components
were quite comparable, equation (7) is unlikely to yield values of 6 lower than
0.98 compared with the experimental values of around 0.88.
36 E. M. Dantzler and C. M. Knobler, J. Phys. Chem., 1969, 73, 1335.
C. M. Knobler, Rev. Sci. Znstr., 1967, 38, 184.
w M. L. McGlashan and D. J. B. Potter, Proc. Roy. SOC.,1962,267, A, 478.
C. P. Hicks and C. L. Young, Trans. Furaduy SOC.,1971,67, 1605.
3s E. J. Partington, J. S. Rowlinson, and J. F. Weston, Trans. Furuday SOC.,1960, 56, 479.
40 I. W. Jones and J. S. Rowlinson, Trans. Furuduy SOC.,1963,59, 1702.
Mixtures Containing a Fluorocarbon 153
Additional evidence confirming the essential weakness of the fluorocarbon +
hydrocarbon interactions in these systems is apparent in the analysis of the
measurements by Battino and co-workers 41s 42 of the binary gas diffusion
coefficient, D12,of CH, and of CF, in a range of aliphatic and alicyclic hydro-
carbons. Application of Chapman-Enskog theory together with the assumption
of a Lennard-Jones 6-12 potential and the use of the Hudson and McCoubrey
combining rules gave theoretical values of DI2 for the CH, + hydrocarbon
systems that were within experimental error in most cases. In contrast the
experimental Dlz's for all the CF, +
hydrocarbon mixtures were consistently
10% higher than the theoretical values, once again indicating 4 < 1.0.
+
C. Recent Fluid Mixture Theory.-The system CF4 CH, is at present unique
among binary mixtures because (i) it is composed of simple non-polar
spherically symmetric molecules, (ii) the system exhibits considerable deviations
from ideality, and (iii) 8 and x are known explicitly and with high accuracy.
P and V E have been measured with high precision and H E (and hence TSE)
can be estimated with somewhat lower precision. This binary system has there-
fore been used to assess the merits of various rival statistical theories of fluid
mixtures. Some of these theoretical predictions are shown in Table 3. It is seen

Table 3 The system CF, +

CHI, comparison of theory with experiment for
equimolar mixtures at 110 K. = 0.9090, x = 1.0095
G ~ / Jmol-1 P/cm3 mol-1 H"/J mol-1 Refs.
Experiment 360 0.88 (+1200) 10, 11, 12
Random mixture 1260 13.9 - 43
Average potential 870 7.1 - 43
LRS, 1 fluid 280 0.90 +-
480 44
VdW, 1 fluid 171 1.11 45
Guggenheim, 1 fluid 298 1.22 46
Flory 376 0.39 +485 47

that the outmoded and unrealistic random-mixing approximation and average

potential model due originally to Prigogine 43 and his co-workers grossly over-
estimate both GE and VE. This is due primarily to the fact that these early
theories failed to account correctly for the effect of size differences between the
components on the thermodynamic properties of the mixture. There is a
significant size difference in the present case, UCF,/UCH, = 1.251.
The next three theories, listed in Table 3, all make use of the so-called van der
Waals approximation to relate the interaction parameters for the mixture to the
parameters associated with the individual like and unlike interactions. They
differ mainly in their choice of equation of state. The theory of Leland,
Rowlinson, and Sather 44 used the experimental equation of state of methane
whereas the two other theories use explicit analytic equations of state, those of
van der Wads and Guggenheim 46 respectively. The LRS theory appears to
469

E. Wilhelm and R. Battino, J. Chem. and Eng. Data, 1972, 17, 187.
E. Wilhelm,'R. Battino, and R. L. Carpenter, J. Chem. and Eng. Data, 1974, 19, 245.
43 I. Priogine, Molecular Theory of Solutions', North-Holland, Amsterdam, 1957.
T. W. Leland, J. S. Rowlinson, and G. A. Sather, Trans. Faraday Soc., 1968, 64, 1447.
4s M. L. McGlashan, Trans. Faraday Soc., 1970, 66, 18.
E. A. Guggenheh, Mol. Phys., 1965,9, 119.
154 Chemical Thermodynamics
be marginally superior but grossly underestimates H E . When f is adjusted to
give the experimental value of @ ( 5 = 0.887), H E is increased to + 600 J mol-1
but VE is increased from 0.90 to 1.22 cm3mol-l, thus worsening the agreement
with experiment. The use of the two-fluid versions of these theories leads to a
4g
slight but not really significant
The Flory theory 47 bears certain similarities to the van der Waals and Guggen-
heim theories in that it makes use of an explicit equation of state. It can be seen
from the results given in Table 3 that, although this theory is quite successful
in predicting GE,the theoretical values of VE and H E are too low by a factor of
more than two.
Although certain of the above-mentioned theories are moderately successful
in representing the experimental data of CF, +
C H I and other fluorocarbon +
hydrocarbon mixtures, experimental values of 5 and x are required. At present
there is no satisfactory method of obtaining these parameters a priori. Scott,
in his 1958 review, considered the various possible factors that could lead to a
weakening of the unlike interactions in such mixtures. He concluded that the
three most significant were the presence of non-central forces, differences in
ionization potential, and differences in size of the two component molecules.
The use of the Kihara potential together with the Hudson and McCoubrey rule
takes account of all these effects and thus the undoubted success of the Knobler
treatment is not surprising. Criticisms could be levelled at his use of a
spherically symmetric potential for substances such as n-hexane but the use of
a more realistic potential such as the Kihara line-core potential is hardly justified
until reliable experimental values for the ionization potentials of the fluoro-
carbons become available.
Scott’s statement made 16 years ago that ‘a large amount of information has
been obtained but we are yet far from a satisfactory explanation for all of it’,
is still pertinent. Our understanding of these systems has increased considerably
over the intervening years but a wholly convincing quantitative theory is still
lacking.

3 Mixtures Containing an Aromatic Fluorocarbon

A. Liquid Mixtures.-Aromatic fluorocarbons first became available com-
mercially just over 10 years ago and, in a paper published in 1960, Patrick and
Prosser noted that hexafluorobenzene formed an equimolar molecular com-
pound with benzene that melted some 18 K higher than the melting temperatures
of the two pure components. In the same year Brooke et aLb1observed that a
similar solid compound was formed between hexafluorobenzene and aniline.
The formation of solid compounds of this type implied the existence of unusually
strong fluorocarbon-hydrocarbon interactions in these mixtures, in complete
contrast to similar mixtures containing an aliphatic or an alicyclic fluorocarbon.
The large volume of research effort that has been undertaken in this area since
47 A. Abe and P. J. Flory, J. Amer. Chem. SOC.,1966, 88, 2887.
48 T. W. Leland, J. S . Rowlinson, G. A. Sather, and I. D. Watson, Truns. Furaduy SOC.,1969,
65, 2034.
ID K. N. Marsh, M. L. McGlashan, and C. Wan, Trans. Faraday SOC.,1970, 66,2453.
C. R. Patrick and G. S. Prosser, Nature, 1960, 187, 1021.
G. M. Brooke, J. Burdon, M. Stacey, and J. C. Tatlow, J. Chem. SOC.,1960, 1768.
 Mixtures Containing a Fluorocarbon 155
1960 has been stimulated primarily by a desire to understand the reasons for this
difference in behaviour between these two classes of fluorocarbon-containing
mixture.
The first systematic studies of fluid mixtures containing an aromatic fluoro-
carbon were those undertaken by Scott at U.C.L.A.62p6sand by Swinton at
S t r a t h ~ l y d e . ~Scott
~ - ~ ~and his co-workers measured HE for binary mixtures of
partially fluorinated aromatics with the general formula C,H,F,-, +
c6H,F6-, (0 < rn, n f 6, It # m) whereas the Strathclyde group studied
mixtures of CGF6with methylated aromatic and alicyclic hydrocarbons and
measured GE,HE, VE, and CF. A representative selection of experimental data
is given in Table 4.
A comparison of the data in Table 1 with those in Table 4 shows that the
values of the three principal excess functions GE, HE, and VE for aromatic
fluorocarbon + aromatic hydrocarbon mixtures are everywhere much less
positive than those for mixtures containing a non-aromatic fluorocarbon. The
experimental excess functions for aromatic fluorocarbons + alicyclic hydro-
carbon are seen to occupy an intermediate position. Another notable trend that
is obvious from the data in Table 4 is that there is a negative contribution to the
excess functions associated with an increase in the degree of substitution of the
hydrocarbon. This trend applies to mixtures of aromatic fluorocarbons with both
aromatic and alicyclic hydrocarbons.
The negative values of GE and HE and the positive values of CF found for
many of the aromatic + aromatic mixtures can only be interpreted in terms of
an enhanced fluorocarbon-hydrocarbon interaction. It was logical at the time
to suggest that charge-transfer forces were responsible and the presence of
z-n donor-acceptor complexes was postulated with the aromatic fluorocarbon
acting as the electron acceptor.60s52 Added weight was given to these ideas when
linear relations were noted between the values of the excess functions for a
particular series of mixtures and the ionization potentials of the aromatic hydro-
c a r b o n ~ .There
~ ~ was no evidence of charge-transfer absorption bands in the
U.V. however and no evidence of complex formation was forthcoming from
either proton or fluorine magnetic resonance experiments.
As more experimental evidence accumulated the need to invoke the existence
of charge-transfer forces receded. Powell and Swinton 6o compared the excess
volumes of C6F6 + methylated benzenes with the values in the series C6F6 +
methylated cyclohexanes. It was noted that, in both series, VE decreased with
increase in methyl substitution to virtually the same extent. A similar trend was
noted for the excess enthalpies.62 The polarizabilities of both aromatic and
b2 D. V. Fenby and R. L. Scott, J. Phys. Chem., 1967,71, 4103.
D. V. Fenby, I. A. McLure, and R. L. Scott, J. Phys. Chem., 1966,70, 602.
64 W. A. Duncan and F. L. Swinton, Trans. Faraday SOC.,1966, 62, 1082.
55 W. A. Duncan, J. P. Sheridan, and F. L. Swinton, Trans. Faraday SOC.,1966,62, 1090.
66 W. A. Duncan and F. L. Swinton, J. Phys. Chem., 1966,70,2417.
W. J. Gaw and F. L. Swinton, Nature, 1966, 212, 283.
68 W. J. Gaw and F. L. Swinton, Trans. Faraday SOC.,1968, 64, 637.
W. J. Gaw and F. L. Swinton, Trans. Faruday SOC.,1968, 64, 2023.
80 R. J. Powell and F. L. Swinton, J. Chem. Thermodynamics, 1970, 2, 87.
R. J. Powell, F. L. Swinton, and C. L. Young, J. Chem. Thermodynamics, 1970,2, 105.
A. Andrews, K. W. Morcom, W. A. Duncan, F. L. Swinton, and J. M. Pollock, J. Chem.
Thermodynamics, 1970, 2, 95.
Table 4 Thermodynamic properties of equimoiar mixtures containing an aromatic fluorocarbon
GE/J mol-1 HE/J mol-l CF/J mol-l K-l P / c m S mol-1 Refs.
-43.8 -433 1.8 0.801 55, 59, 61
- 184.2 -1146 16.7 0.414 55, 59, 61
- 394.5 - 1660 12.3 0.086 55, 59, 61
- - 2320 12.3 - 0.357 59, 61
- 79 0.6 0.86 52, 63
- 253 - 52
776.3 1517 - 2.7 2.567 55, 58, 61
- 1275 -2.1 2.214 60, 61
1165 -2.1 2.004 60, 61
- 970 - 2.7 1.635 60, 61
453 650 1.2 1.370 64, 65, 66
- 42 -0.3 0.811 95 9
22
k
Mixtures Containing a Fluorocarbon 157
alicyclic hydrocarbons increase by approximately the same amount as methyl
groups are added, so a reasonable explanation for the experimentally observed
trends in H E and VE is that the enhanced unlike interactions are due solely to
an increased contribution from dispersion forces. The strength of such forces is
directly proportional to the product of the polarizabilities of the interacting
molecules. Although this theory would account for the relatively small differences
between the various mixtures in a particular series, it cannot explain the gross
differences in the excess functions found experimentally between the aromatic
fluorocarbon +
aromatic hydrocarbon, aromatic fluorocarbon + alicyclic
hydrocarbon, and alicyclic fluorocarbon +
alicyclic hydrocarbon series.
Both Fenby and Scott 62 and Gaw and Swinton 59 attempted to rationalize
58s

the existing experimental data by assuming that the experimental excess function
XE (for an aromatic fluorocarbon +
aromatic hydrocarbon mixture) was
composed of two parts: XE = X s +
XE. XB represented the contribution due
to ‘physical’ or dispersion forces and X s the contribution resulting from
‘chemical’ or complexing interactions. XB was approximated by the value of
the excess function for the equivalent aromatic fluorocarbon +
alicyclic hydro-
carbon in which, it was postulated, no specific complexing forces would exist.
This is, of course, a rather dubious assumption as the non-specific interactions
are unlikely to be identical in the two binary mixtures owing to the considerable
differences in shape and polarizibility of the two hydrocarbons, so that the
numerical values of derived thermodynamic functions must be subject to con-
siderable uncertainty.
Values of GE as functions of temperature and composition can be analysed
to give the equilibrium constant and enthalpy of formation of the assumed 1-1
complexing reaction. This procedure was first used most successfully by
Andersen et a1.87 in 1962 when applied to mixtures of l-hydro-n-perfluoro-
heptane with acetone. An intercomponent H-bond is formed in this system and,
by combining experimental HE’s with equilibrium constants derived from n.m.r.
measurements, it was found that the derived values of H g were in close agreement
with the experimental enthalpy of the reference system n-C,Fle +
acetone in
which only physical interactions were assumed to be present.
Since 1970 Stubley and his co-worker at Lancaster 63s 68v 69 have measured
+
H E and VE for many systems of the type C8FSH alkyl benzene, and Fenby
and his group at Otago 70-72 have studied mixtures of the types C&X C6H6 +
and C6F5X + cyc~o-C,H,, where the substituent X is C1,Br, I, or CN. Stubley
found that the trend in the excess functions with the degree of substitution
of the hydrocarbon for binary mixtures containing pentafluorobenzene closely
B. J. Skillerne de Bristowe and D. Stubley J. Chem. Thermodynamics, 1973, 5, 865.
O4 B. I. Mattingley, Y.P. Handa, and D. V. Fenby, J. Chem. Thermodynamics, 1975, 7 , 169.
66 D. A. Armitage and K. W. Morcom, Trans. Faraday Soc., 1969, 65, 688.
M A. Chand and D. V. Fenby, J. Chem. Thermodynamics, 1975, 7 , 403.
67 D. L. Andersen, B. A. Smith, D. B. Myers, S. K. Alley, A. G. Williamson, and R. L. Scott,
J. Phys. Chem., 1962, 66, 621.
68 B. J. Skillerne de Bristowe and D. Stubley, J. Chem. Thermodynamics, 1973, 5, 121.
69 P. J. Howell, B. J. Skillerne de Bristowe and D. Stubley, J. Chem. Thermodynamics, 1972, 4,
225.
70 D. V. Fenby and S. Ruenkrairergsa, J. Chem. Thermodynamics, 1973, 5, 227.
71 S. Ruenkrairergsa, D. V. Fenby, and D. E. Jones, J. Chem. Thermodynamics, 1973, 5, 347.
C. Y. Leong, D. E. Jones, and D. V. Fenby, J. Chem. Thermodynamics, 1974,6,609.
158 Chemical Thermodynamics
paralleled the trends found previously for systems containing hexafluorobenzene.
H E and VE decreased in a regular manner as the extent of methyl substitution of
the aromatic or alicyclic hydrocarbon was increased. If, however, a series of
mixtures containing mono-substituted alkyl benzenes as one component was
considered, HE and VE were found to increase regularly as the length of the
alkyl chain was increased. This latter trend was due to the hydrocarbon changing
its character from aromatic to aliphatic with a resultant positive contribution to
both HE and VE.
Skillerne de Bristowe and Stubley 73 have analysed their results in terms of the
Barker 74 lattice model for mixtures with specific interactions. Mixtures of C6F6
and CBFsHwith monosubstituted aromatic hydrocarbons were treated success-
fully with constant values of the selected interaction parameters. The theory
failed to predict the thermodynamic functions for systems containing an aromatic
hydrocarbon with more than one methyl group and thus has little predictive
power.
Becker has applied his semi-empirical theory of associated mixtures 76 to 7s5

CsF6 + C6H6and has shown that this theory will reproduce the double azeo-
tropic behaviour found experimentally in this
Fenby et al.70-72followed the previously discussed practice of Andersen et aZ.67
and estimated the ‘chemical’ contributions to HE and VE of C6FBX+ CsH6
mixtures by subtracting the experimental values of the excess functions for
+
C6F6X cyclo-C,H,, mixtures. These ‘chemical’, or complexing HE and V z
contributions are given in Table 5.

Table 5 HZ and Vg for equimolar mixtures of the type C6FSX + C6H6

X H F c1 Br I CN
Hg/kJ mol-l -1.46 -2.07 -1.88 -1.90 -2.18 -
V&cm3mol-l -1.23 -1.62 -1.37 -1.33 -1.41 -1.49

The values of HE and VE given in Table 5 are listed, from left to right, in order
of the expected electron-withdrawing ability of the substituent group X. Although
Hg and YE seem to follow similar trends along the table, there is no systematic
increase in either function. The CN group, in particular, is known to be strongly
electron-withdrawing but YE for this system is not markedly more negative than
the values for the other mixtures. Fenby concluded72that ‘Charge transfer is
insignificant and the thermodynamic properties of aromatic fluorocarbon
systems can be understood in terms of shape and polarizibility differences’.
The presence of permanent dipole-induced dipolar forces in five of the six
systems listed above is another complicating factor that must be considered.
One of the most interesting findings of the earlier work in this area was the
observation of Gaw and Swinton 69 of multiple azeotropy in C6F6 + C6H6
67s

mixtures. The isothermal static vapour-pressure measurements of Gaw and

Swinton were performed at several temperatures over the range 303.15 K to
79 B. J. Skillerne de Bristowe and D. Stubley, J. Chem. Thermodynamics, 1974, 6, 581.
’‘ J. A. Barker, J. Chem. Phys., 1951, 19, 1430.
75 F. Becker, M. Kiefer, and P. Rhensius, 2. Naturfursch., 1972, 27a, 1611.
m F. Becker, M. Kiefer, H. Koukol, and H. D. Schgfer, 2. Nuturforsch., 1973, 28a, 772.
77 F. Becker, personal communication.
 Mixtures Containing a Fluorocarbon 159
343.15 K and the presence of both a maximum-boiling and a minimum-boiling
azeotrope was confirmed and was predicted to persist up to around 400K.
Recently two other research groups have examined this mixture using equilibrium
stills. Kogan and Morachevskii78 studied the system at pressures of 760, 500,
and 300 Torr, whereas the measurements of Chinikamala, Houth, and Taylor 7B
were restricted to 760 and 500 T0rr.P Both sets of measurements are in reason-
able agreement with one another and with the results of Gaw and Swinton.
(The azeotropic temperatures reported by Gaw and Swinton are misquoted in
the paper by Chinikamala et al.)
As far as this reviewer is aware, no other example of multiple azeotropy in a
binary mixture has been reported to date, although the phenomenon was first
predicted theoretically by Duhems0 at the beginning of the present century.
Van Konynenburgsl has studied the great variety of phase behaviour resulting
when the equation of state of van der Wads is applied to a binary mixture and
noted that, for certain restricted ranges of energy and size differences between
the two components, the existence of multiple azeotropy is correctly predicted.
Chinikamala et aL70 also studied isobaric vapour-liquid equilibria in binary
mixtures of C6F6 with toluene, methylcyclohexane, and n-hexane. The first
system mentioned was also studied earlier by Gaw and Swinton6Qand the new
measurements are found to be in substantial agreement but of rather lower
precision. Both methylcyclohexane and n-hexane were observed to form single
minimum-boiling azeotropes with CsFs, a consequence of the considerable
positive deviations from ideality resulting from weakened unlike interactions.
Wade and Taylor 82 have examined vapour-liquid equilibria in the ternary
+ +
system C6F6 CsHs c-CH,. C8Hll at a constant pressure of one atmosphere.
The system contains two minimum-boiling binary azeotropes and one maximum-
boiling binary azeotrope. The triangular temperature-composition diagram
contains a ridge structure extending between two ‘peaks’ representing pure
c-CH8.C6Hll (T = 373.95 K) and the maximum-boiling azeotrope in CsF6 +
C6H6(T = 253.45 K). The ridge between these two peaks contains a ‘bealach’
at XC,F, = 0.641, XC,H, = 0.291, T = 253.40 K. This constant-boiling mixture
is an example of a saddlepoint azeotrope and is neither the lowest-nor the
highest - boiling invariant point in the ternary system.
Beaumont and Davis 84 examined the U.V. spectra of mixtures of CsF6with
839

a variety of electron donors for evidence of charge-transfer absorption band^.^^-^^

No bands were observed for mixtures containing z-donors but moderately strong
n-donors such as NNN’N’-tetramethyl-p-phenylenediamineand several sub-
stituted anilines were found to exhibit quite intense U.V. absorption when mixed
I. V. Kogan and A. G . Morachevskii, Zhur. priklad. Khim., 1972, 45, 1885.
79 A. Chinikamala, G. N. Houth, and Z. L. Taylor, jun., J. G e m . and Eng. Data, 1973,18, 322.
P. Duhem, 2.phys. Chem., 1901,36, 605.
P. H. Van Konyenburg, Ph.D. Thesis, U.C.L.A., 1968.
J. C. Wade and Z. L. Taylor, J. Chem. and Eng. Data, 1973,18,424.
89 T. G. Beaumont and K. M. C. Davis, J. Chem. SOC.(B), 1967, 1131.
T. G. Beaumont and K. M. C. Davis, Nature, 1968, 218, 865.
H. A. Benesi and J. H. Hildebrand, J. Amer. Chem. SOC.,1949, 72, 2703.
86 G. Briegleb, ‘Elektronen-Donator-Acceptor-Komplexe’, Springer-Verlag, Berlin, 1961.
87 R. Foster, ‘Organic Charge Transfer Complexes’, Academic Press, London and New York,
1969.
t Tom = (101 325/760) Pa.
160 Chemical Thermodynamics
with C6F6 in solution in n-CsH1,. Measurements made over a range of tempera-
ture enabled both equilibrium constants and enthalpies of formation for the
complexing reaction to be determined. In later work with Morcomaa they
examined mixtures of C,F6 with both NN-dimethylaniline and NN-dimethyl-
toluidine. A combination of spectroscopic and thermodynamic measurements
was made and the excess enthalpy due to complex formation, El& was
found by using the system C6F6 + isopropylcyclohexane as the ‘reference
system’ within which only non-specific interactions were assumed to be
+
present. The value, for equimolar mixtures of C6F6 NN-dimethylaniline,
HE = -3040 Jmol-l, is in good agreement with the value -2816 J mol-1
derived entirely from the spectroscopic measurements. The spectroscopic and
thermodynamic estimates of HE are in even closer agreement for the other
system, CsF6 + NN-dimethyltoluidine. This surprising correspondence between
spectroscopic measurements made in dilute n-hexane solution and thermo-
dynamic measurements made in the neat binary mixtures must mean that the
complex is not solvated to any significant extent even at the low concentrations
used in the spectroscopic measurements. Recently Hall, Morcom, and
Brindley 89 have completed a similar study on C,F&N +
NN-dimethylaniline
using C6F5CN + isopropylcyclohexane as the reference system. HE and CF
were measured over a temperature range above the upper critical solution
temperature of 300 K while spectroscopic measurements enabled an independent
estimate to be made of HB. Once again almost exact agreement was found
between the thermodynamic and spectroscopic values.
Armitage and Morcom65 have measured H E for C6F6 + pyridine,
+ 2,6-dimethylpyridine, and +
2,4,6-trimethylpyridine and, more recently,
Meyer et aLBOmade similar measurements on C6F6 2-methylpyridine, +
+ 4-methylpyridine, and +
3,5-dimethylpyridine. H E was found to decrease
with the extent of methyl substitution in a manner reminiscent of the mixtures of
CsF6with methylated aromatic hydrocarbons discussed earlier. HE for a
particular mixture containing a methyl pyridine was always found to be less
negative than H E for the corresponding mixture containing a methylated benzene.
There is no spectroscopic evidence of charge-transfer forces in these systems
although the excess heat capacities are large and positive for all of the systems
containing a methyl-substituted pyridine, a fact indicative of strong unlike
interact ions.
Armitage and Morcom 65 also examined mixtures of C6F, with triethylamine,
a typical n-donor and measured HE and CF. Fenby et al. have recently quoted
values of GE and V E for this same s y ~ t e m .Experimental
~ ~ ~ ~ ~ values of the
principal excess functions are shown in Table 4. The equimolar values of GE,
HE,and VE are all positive and approximately half the magnitude of the excess
functions for C6Fs + cyclo-C,H,, mixtures. Cg is small and positive whereas
CF for CsF,+ CyClO-C,Hlz is small but of opposite sign. Handa and Fenby,*
suggest that the presence of a weak complex would account for this difference
88 D. A. Armitage, T. G. Beaumont, K. M. C . Davis, D. J. Hall, and K. W. Morcom, Trans.
Faraday SOC.,1971,67, 2548.
D. J. Hall, K. W.Morcom, and J. M. T. Brindley, J . Chem. Thermodynamics, 1974, 6, 1133.
so R. Meyer, M. Meyer, and J. Metzger, Thermochemica Acia, 1974,9, 323.
Mixtures Containing a Fluorocarbon 161
but the fact that triethylamine possesses a permanent dipole moment must be
considered as a possible alternative.
Chand and Fenby66 also measured VE for mixtures of two other n-donors,
diethylether and dimethylsulphoxide with CsFe, and Leong, Fenby, and Jones 91
studied V E for mixtures of acetone with both C,F, and C,F5H and noted that
although values for the equimolar mixtures were both large and positive the
value for C,F, +
acetone was three times larger than the corresponding value
+
for C6F5H acetone. This trend suggests the possibility of hydrogen bonding
in the latter mixture. Martin and Murrayg2 have published VE values for
mixtures of C6F6with four ethers : diethyl, di-isopropyl, di-n-propyl, and di-
n-butylether. The experimental VE’s are seen to become more positive as the
size of the hydrocarbon substituent increases, a not unexpected result.
Andrews and Morcom have continued their interesting series of studies on
binary mixtures containing cyclic ethers s3, 94 by measuring HE, C?, and VE for
C6FB+ 1,4-di0xan.~~ Their experimental values are given in Table 4. The inter-
molecular interactions in this system are particularly subtle because Ct was
found to depend strongly on temperature, and at all temperatures exhibited a
complex and irregular composition dependence. Andrews and Morcom have
compared the HE%for the four binary mixtures CBF, +
pyridine, C6F6 +
l,ddioxan, CgHs + +
1,4-dioxan, and CsFs CsH6. When H E is plotted as a
function of the mole fraction of ‘donor’ (the second-named component in each
case), H E was found to be strongly skewed in the same sense, i.e. towards mixtures
rich with respect to the ‘donor’ species.
The excess volumes of mixtures of CsF6 with n-donors : diethylether, diethyl-
sulphide, diethylamine, di-n-propylether, and di-n-propylsulphide have been
studied at several temperatures by James and M c L ~ r e .They ~ ~ also measured
VE for C6F6 +
n-pentane, +
n-hexane, and +
n-heptane and used these data in
the analysis of their results to make an estimate of the extent of donor-acceptor
+
complexing in the C6F6 ether mixtures.
B. Solid Mixtures.-This review is primarily concerned with the properties of
fluid mixtures containing a fluorocarbon but this section on the thermodynamic
properties of solid mixtures containing an aromatic fluorocarbon is included for
completeness. One of the most obvious features of many of these mixtures is the
formation of a solid congruently melting compound often with a melting tempera-
ture higher than the melting temperatures of both pure components. Although it
is convenient to describe the stoichiometry of these compounds as ‘1-l’, ‘2-l’,
etc. and this practice will be continued in the present article, it is well known that
the phase diagrams of many of these mixtures contain large solid-solution
regions where the ‘compound’ has a variable composition. The existence of
solid solutions was deduced in the early work in this area by the absence of
eutectic arrest points at compositions near those corresponding to integral mole

C. Y. Leong, D. V. Fenby, and D. E. Jones, J. Chem. Thermodynamics, 1974,6,611.

8a M. L. Martin and R. S. Murray, J. Chem. Thermodynamics, 1972, 4, 723.
s3 A. W. Andrews and K. W. Morcom, J. Chem. Thermodynamics, 1971,3, 513.
D4 A. W. Andrews and K. W. Morcom, J. Chem. Thermodynamics, 1971,3, 519.
96 A. W. Andrews, D. Hall, and K. W. Morcom, J. Chern. Thermodynamics, 1971,3, 527.
J. James, Thesis, University of Sheffield, 1974.
162 Chemical Thermodynamics
ratios of the two components. Recently Armitage 97 has established the limits of
the solid-solution region for the mixture C6F6 + NN-dimethylaniline in a most
elegant way by measuring the enthalpy of solution of the solid compound in
each of the two pure substances using a slightly modified Larkin and McGlashan
cal~rimeter.~~
Patrick and Prosser determined an approximate phase diagram for
+
C6F6 C6H6 and observed the 1-1 congruently melting compound melting
some 18 K higher than the melting temperature of either C6F6 or CsH6. Duncan
and Swinton 64 published a more precise phase diagram for this system at a later
date and, more recently, Goates, Ott, and ReederD9have studied this system
yet again. Their experiment results are in almost exact agreement with those of
Duncan and Swinton, who also established the presence of 1-1 congruently
melting compounds for the systems C6F6 + CH3 C3H5, + 1,4-(CH3),- C6Hp,
and + 1,3,5-(CH3)3-C6H3and showed that, in complete contrast C6F6 +
n-CBH14 formed a simple eutectic. The phase diagram of C G F3-~ cyclo-c6H,,
was found to be intermediate in character in that several weak incongruently
melting compounds were formed. Ott, Goates, and Reeder loo also studied
mixtures of C6F6 with cyclohexene, 1,3-~yclohexadiene, and with 1,5-cyclo-
octadiene. They found that the hydrocarbon with but one double bond formed
no compounds whereas the two other hydrocarbons each containing two double
bonds, conjugated and unconjugated respectively, formed incongruent 1-1
compounds.
McLaughlin and Messer Io1 noted the formation of 1-1 congruently melting
compounds in the mixtures C6F6 + naphthalene, perfluoronaphthalene + C6H6,
and perfluorobiphenyl + C6H6. In contrast C6F6 + biphenyl was found to
form an incongruent 1-1 compound, this being the only example so far observed
of a mixture of CsF6 with an aromatic hydrocarbon that fails to form a
congruently melting compound.
Phase diagrams for mixtures of C,F,H with hydrocarbons were published by
Duncan and Swinton 56 and by Kelly and Swinton.lo2 Incongruently melting
-
compounds of composition C6F5H C6H6 and 3C6F5H.2C6H6 were observed
for CBFSH+ CsH6 mixtures and the phase diagrams of Kelly and Swinton
showed the formation of 1-1 congruent compounds in the systems C,F,H +
O-(CHJ,C6H*, + m-(CH&C,H4, +
1,3,5-(CH&C6H3, and (CH&C6H. +
Incongruently melting compounds were noted for C6FsH + p-(CH3)&&4 and
for C6FsH + 1,3,5-(CH3)SC6H3,the former 1-1, the latter 3-2. Rather
surprisingly Kelly and Swinton found that C6FSH + C,(CH,), formed a simple
eutectic system with no evidence whatever of compounds.
The investigation of the phase diagrams of mixtures of C6F6 with a variety
oi electron donors has been undertaken recently by Goates, Ott, and their
collaborator^.^^^ loop lo4 Their results are summarized in Table 6.
lo39

D7 D. A. Armitage, unpublished data.

J. A. Larkin and M. L. McGlashan, J . Chem. SOC.,1961, 3425.
'* J. R. Goates, J. B. Ott, and J. Reeder, J. Chem. Thermodynamics, 1973, 5, 135.
l o o J. B. Ott, J. R. Goates, and J. Reeder, J . Chem. Thermodynamics, 1974, 6, 281.
lol E. McLaughlin and C. E. Messer, J. Chem. SOC. (A), 1966, 1106.
lo8T. Kelly and F. L. Swinton, J. Chem. Thermodynamics, 1974, 6, 435.
lo8 J. R. Goates, J. B. Ott, J. Reeder, and R. B. Shirts, J. Chem. Thermodynamics, 1974, 6, 489.
lo' J. B. Ott, J. R. Goates, J. Reeder, and R. B. Shirts, J. Chem. Soc., Faraday Trans. I, 1974,
70, 1325.
Mixtures Containing a Fluorocarbon 163
Table 6 Compound formation in mixtures of C6F6 and C6F&N with electron
donors
Mixture Compound T/K Refs.
C6F6 + C6H6 congruent
1-1 297.21 54,99
C6F, + dimethylsulphoxide incongruent
1-1 - 99
C6F6 + furan incongruent
1-1 - 99
c6F, + thiophen congruent
1-1 261.09 99
C6F, + tetrahydrofuran incongruent
1-1 - 103
C6F, + N-methylmorpholine congruent
1-1 216.18 104
GF, + N-methylpiperidine incongruent
1-1 - 104
C6F, + NN’-dimethylpiperazine congruent
1-1 264.84 104
C6F6+ NN-dimethylaniline congruent
1-1 314.99 106
C6F6+ NN-dimethyl-p-toludine congruent
1-1 315.6 106
C6F,CN + NN-dimethylaniline congruent
1-1 321.6 106
C,F,CN 4- c6H6 congruent
1-1 284.25 105
and 5-3 congruent 286.00 105

Goates, Ott, et al. have rationalized their extensive data by suggesting that a
cyclic structure for the second component is an important factor leading to
compound formation. Carbon disulphide appears to be the only non-cyclic
substance so far studied lo*that forms a compound with C6Fs. The presence of
multiple interaction sites on the ‘donor’ molecule is also seen to be important.
The stability of the compounds is enhanced as the number of active groups:
C=C double bonds, O-atoms, or N-atoms, is increased. The suggestion that the
existence of these solid compounds is due to donor-acceptor complexing can be
discounted when it is noted that triethylamine, NN-dimethylacetamide, tetra-
hydrofuran, and lY2-dimethoxyethanedo not form compounds with C6F8 but
are as efficient electron donors as many of the cyclic molecules listed in Table 6.
Hall and Morcomlo5 have recently published the phase diagram for
C,F5CN +
C,H6. This system is unique in that two congruently melting
compounds are observed: C6F5CN*C6H6melting at 284.25 K and 5CBF&N*-
3c6H6 melting at 286.00K. In contrast, the phase diagram for C,F,CN +
NN-dimethylaniline exhibits a single 1-1 compound similar to mixtures of this
amine with C6F6.106
These interesting intermolecular compounds have received very little attention
from X-ray crystallographers. Boyens and Herbstein,lo7in a preliminary study
made in 1965, established that the compounds were arranged in a regular
alternating layer structure with the fluorocarbon and hydrocarbon molecules
arranged in a face-to-face configuration, a structure reminiscent of systems
stabilized by x-x charge-transfer interactions. Since then Dahl has completed
detailed crystallographic studies of C6F6 + 1,3,5-(CH3)3C6H3~08and of
C6F6 + C6(CH3)6.10B1 He found evidence of considerable molecular motion
in the crystal lattices at temperatures well below the crystal melting temperatures.
lo5 D. J. Hall and K. W. Morcom, J. Chem. Thermodynamics, 1975, 7 , 405.
lo8 D. A. Armitage, J. M. T. Brindley, D. J. Hall, and K. W. Morcom, J. Chem. Thermodynamics,
1975,7, 97.
lo7 J. C. A. Boeyens and F. H. Herbstein, J. Phys. Chem., 1965, 69, 2153.
lo8 T. Dahl, Acta Chem. Scad., 1971,2!5, 1031.
lop T . Dahl, Acta Chem. Scand., 1972, 26, 1569.
110 T. Dahl, Acta Chem. Scand., 1973, 27, 995.
164 Chemical Thermodynamics
This thermally induced motion gave rise to considerable diffuse X-ray scattering
and thus reduced the precision of the deduced crystal structures. Dahl observed
that both compounds appeared to undergo a solid-solid transition at a well-
defined transition temperature below which the diffuse scattering was much
reduced.
Dahl studied the 1-1 C,F6*1,3,5-(cH,)3C6H, compound at 233 K, well below
the transition temperature and observed an interplane separation of 356 pm,
a value close to the sum of the van der Waals radii. The hexamethylbenzene
compound was found to form two completely different crystal structures above
and below its transition temperature of 273 K. This compound in the high-
temperature form is trigonal in structure with an interplane separation of 356 pm.
Dahl observed that this phase showed evidence of some disorder in the regular
alternate stacking sequence. Below 273 K the compound appeared to be
completely ordered, triclinic in crystal habit, and to possess a plane-to-plane
separation of only 343 pm.
The existence of orderdisorder transitions in compounds of this type has been
confirmed by Brennan, Brown, and Swinton who studied CGF6*C6Hs using a
differential scanning calorimeter. This mixture was found to undergo a solid-
solid phase transition at 249.2 K with a transition entropy of 2.65 J mol-1 K-l
and an associated volume increase of 2.1%. The enthalpy change associated
+
with the process C&&(s) C,F,(s) = CgHg'C6F6(S) Was found to be
(1.0 k 0.3) kJ mol-l. This small positive value further emphasizes that the
stability of these compounds is primarily due, not to the presence of strong
attractive intermolecular forces, but rather to entropic factors. It would also
be of considerable interest to measure the volume change associated with the
above solid-state reaction.
The face-to-face configuration adopted by the molecules in these 1-1 com-
pounds is unlike the molecular arrangement in many solid aromatic hydrocarbons
and in solid C6F6 in which the plate-like molecules are found to be aligned in a
mutually perpendicular manner. Andersen 112 measured T,, the spin lattice
relaxation time for solid benzene as a function of temperature, and observed that
Tl exhibited a pronounced minimum at 160 to 180 K. This was interpreted as
the onset of rotation about the six-fold axis. Similar measurements by Boden
et aL1l3on C6Fg have shown that TI exhibits two minima. Boden et al. have
successfully accounted for the presence of these two rotational processes by
noting that, whereas the C6H6unit cell contains four molecules in indistinguish-
able locations, the C6Fs unit cell contains six molecules, four equivalently
situated and the remaining two located differently at centres of
Boden has recently carried out a comparable study of the CsF6*C& com-
pound. lH, 2H, and 19Fspin-relaxation measurements at temperatures below
250 K are consistent with (a) the C6F6 molecules occupying two inequivalent
sites with relative populations of 2/1 as in pure C6F6, (b) the C&& molecules
occupying a single site with C6F6 molecules as nearest neighbours, and (c) the
J. S. Brennan, N. M. D. Brown, and F. L. Swinton, J.C.S. Furaduy Trans. I, 1974,70, 1965.
lla J. E. Anderson, J. Chem. Phys., 1965, 43, 3575.
N. Boden, P. P. Davis, C. H. Stam, and G. A. Wesselink, Mol. Phys., 1973, 25, 87.
11* N. Boden, P. P. Davis, C. H. Stam, and G. A. Wesselink, Mul. Phys., 1973, 25, 81.
ll6. N. Boden, unpublished measurements.
 Mixtures Containing a Fluorocarbon 165
molecular reorientation processes and kinetics being similar to those encountered
in pure C6F6and pure C6H6. Boden concludes: 'the stability of the solid com-
pound originates solely in the favourable packing of the molecules and is in no
way related to charge-transfer interactions'.
C. Gaseous Mixtures.-Powell, Swinton, and Young measured the gas-liquid
critical temperatures of mixtures of C,F, with methylated benzenes and with
methylated cyclohexanes and extracted values of the interaction parameter 5
using the Rowlinson theory 39s 40 mentioned earlier in Section 2B. Hicks and
Young 38 carried out similar measurements on C6F6,CsHs, and cyclo-C6F1, with
various cycloalkanes. A representative sample of 5' values is given in Table 7.

Table 7 values derived from gas-liquid critical temperatures

System f Refs.
C,F6 + cyclohexane 0.949 61
C6F6 + methylcyclohexane 0.968 61
C6F6 + 1-trans4dimethylcyclohexane 0.971 61
GF, + 1-trans-3,5-trimethylcyclohexane 0.98 1 61
GF6 + benzene 0.962 61
C6F6 + methylbenzene 0.975 61
C6F, + 1,4-dimethylbenzene 0.981 61
C6F6 + 1,3,5-trimethylbenzene 0.997 61
C6F6 + cyclopentane 0.947 38
GF6 + cycloheptane 0.952 38
c-GF1, + cyclopentane 0.896 38
c-GF,, + cyclohexane 0.898 38
c-C,F,, + cycloheptane 0.880 38
c6& + cyclopentane 0.995 38
C6H6 + cyclohexane 0.987 38
C6H6 + cycloheptane 0.991 38

The increase in f with extent of methyl substitution for both the aromatic and
alicyclic series is clearly seen confirming the pattern observed in the excess
functions at lower temperatures.
The interactions in the cyclohexane systems are everywhere weaker than the
interactions in the comparable methylated benzene system. The results of Hicks
and Young on cycloalkane mixtures show a comparable inter-system trend.
The ,f values lie in the order C,H, +
cycloalkane > c,F, + cycloalkane >
cyclo-C,F,, + cycloalkane. The Hudson and McCoubrey combining rule was
found to predict values of 5 in good agreement with experiment for the first
series but failed completely for the latter two.
Mousa et aZ.lls and Hicks and Young 117 studied the gas-liquid critical tempera-
tures of both C,F6 and C,H, with a range of n-alkanes; ( was found to be less
than unity for both series of mixtures and tended to decrease as the size difference
between the two components increased. This trend was especially marked for the
CgHg + n-alkane series and was correctly predicted by the Hudson and
McCoubrey combining rule.
ll6 A. el H. N. Mousa, W. B. Kay, and A. Kreglewski, J. Chem. Thermodynamics, 1972,4, 301.
11' C . P. Hicks and C . L. Young, J. Chem. Thermodynamics, 1971,3, 899.
166 Chemical Thermodynamics
Dantzler and Knobler 118 and Powell and Swinton ll9 have measured B12, the
interaction virial coefficient for C6Fs + CsH6 and for C6Fe cyclo-C6H12. +
Dantzler and Knobler measured BIZindirectly by measuring 8,the excess virial
coefficient, whereas Powell and Swinton measured the virial coefficients of a
mixture and of the two pure substances directly using a compression method.
They observed that, when TCand V c are used as the reduction parameters for the
pure components and the Lorentz-Berthelot rules used to estimate the reduction
parameters for the unlike interactions, the like and unlike interactions in both
binary mixtures follow different theorems of corresponding states. The most
striking observation is the fact that the intermolecular forces in pure C6F6 are of
a completely different character to those in the pure hydrocarbons.
This unusual behaviour was commented on by Douslin et a2.lz0who measured
B for C6F6 over a wide range of temperature and showed that neither a two-
parameter n-m potential nor a three-parameter Kihara potential could represent
the experimental virial coefficients. They observed that the three-parameter
Stockmayer potential was able to represent the experimental data within
experimental error. This potential assumes that the molecules possess a
permanent dipole moment and so its use is in no way justified for C6F6. Agnew
and Swinton 121 have attempted to fit the experimental virial coefficients using the
generalized angle-dependent potential of Pople 122 (of which the Stockmayer
potential is a special case). This potential includes terms involving the quadrupole
moment, departures from spherical symmetry, and anisotropy of the molecular
polarizability. C6F6 possesses a sizeable quadrupole moment and, although no
experimental value is yet available, reliable theoretical estimates have been made
by Gierke et aZ.lZ3 Although this potential was found to represent the experi-
mental data for C,H6 within experimental error, no comparable agreement could
be obtained for C6Fs and the superiority of the Stockmayer potential with a
fictitious permanent dipole was confirmed.
The fact that the pure components do not obey the same theorem of corre-
sponding states complicates the analysis of the mixed virial coefficients for both
C6F6 +CsHg and C6F6 +
cyclo-C,H1z. Dantzler and Knobler assumed the
existence of a 1-1 complex in the former gaseous mixture and estimated B,",,
the non-specific interaction virial coefficient, by the use of the previously-
mentioned McGlashan and Potter equation. This equation was found to predict
the interaction virial coefficient for C,F6 + cyclo-C6H12 with reasonable
accuracy, although this is a little surprising in view of the known weakness of
the unlike interactions in this mixture. The experimental interaction virial
coefficient was separated into specific and non-specific parts: Blz = &{ + By,,
and K,, the equilibrium constant for complex formation, estimated from the
well-known expression B z = -(RTln K,,)/2. K , was calculated at several
temperatures and AHc, the enthalpy of formation of the assumed complex
estimated: h H c = - 23.4 kJ mol-l. This unexpectedly large value is in good
118 E. M. Dantzler and C. M. Knobler, J. Phys. Chem., 1969,73, 1602.
ll9 R. J. Powell and F. L. Swinton, unpublished results.
120 D. R. Douslin, R. H. Harrison, and R. T. Moore, J. Chem. Thermodynamics, 1969, 1, 305.
121 F. J. Agnew and F. L. Swinton, unpublished calculations.
laa J. A. Pople, Proc. Roy. SOC.,1954, 221, A , 508.
lZ3 T. D. Gierke, H. L. Tigelaar, and W. H. Flygar, J. Amer. Chem. SOC.,1972,94, 330.
 Mixtures Containing a Fluorocarbon 167
agreement with the figure of -21 kJmol-l found by Baur et aZ.12* from an
analysis of measurements of the relative permittivity and refractive index of
this system in the liquid state. These figures can be contrasted with the value of
- 9.7 kJ mol-1 obtained by Gaw and Swinton 59 from their liquid-state Gibbs
free energy measurements.
D. Miscellaneous Non-thermodynamic Measurements.-Baur et aZ.124~ 125

measured Pm,the molar polarization, and R& the molar refraction at infinite
+
wavelength, for mixtures of C6F6 C6H6, 1,4-(CH3)2C6Ha,and 1,3,5-(CHJ3-
C6H3. In all three cases Pm was found to be a linear function of mole fraction
whereas considerable negative deviations from linearity were found for plots
of Rg against mole fraction. A similar pattern was observed for C6F6 +
cyclo-C6Hl, mixtures but the negative deviations were much reduced. These
experimental facts were interpreted by assuming the presence of associated
complexes resulting from strong specific interactions; the negative deviations in
Rg were attributed to a negative excess electronic polarizability. A strict linear
relation of Pm with mole fraction would have implied a complex with zero
permanent dipole moment but Baur et al. showed that their results were
consistent with the complex possessing a small dipole moment of up to 0.3 x
10-30 C m owing to scatter of the experimental results. In principle it is possible
to calculate the equilibrium constant and the dipole moment from such measure-
ments by use of the formalism developed by Baur et al. but, in practice, the
experimental results were not of sufficient precision in this case. The results were
incompatible with the complex possessing a large degree of charge-transfer
character.
In the late sixties Le Fkvre et aZ.126-128
carried out various dielectric relaxation
+
measurements on several mixtures including C6F6 C6(CH& They noted that
when either component was dissolved separately in CC14 or CyClO-C6H,, there
was no change in the dielectric loss but mixtures showed an appreciable increase
in loss. This increase was interpreted as being evidence of a complex with a
non-zero dipole moment, the complex being stabilized by electrostatic inter-
actions. In a related study 129 on C6F5H,it was noted that the permanent dipole
moment, the molar Kerr constant, and the dielectric relaxation time were all
increased when C6H6 rather than cyclo-C6H1, was used as diluent and this
was thought to be due to complex formation. The difference between the
experimental value of the molar Kerr constant of a solute in a complexing solvent
and the value for the same solute in a non-complexing inert solvent or in the
vapour state can be used to give information concerning the stereochemistry of
the assumed complexed species. It was considered probable that, in C6FsH +
C6H6 mixtures, the two mokules were arranged in a face-to-face, staggered
configuration with a maximum interplane angle of 35".
lZ4 M. E. Baur, C. M. Knobler, D. A. Horsma, and P. Perez, J. Phys. Chem., 1970, 74, 4595.
lZ6 M. E. Baur, D. A. Horsma, C. M. Knobler, and P. Perez, J. Phys. Chem., 1969, 73, 641.
126 K. E. Calderbank, R. J. W. Le FBvre, R. K. Pierens, and D. V. Radford, Austral. J. Chem.,
1968,21, 275.
127 R. J. W. Le Fbvre, D. V. Radford, G. L. D. Ritchie, and P. J. Stiles, Chem. Comm., 1967,
1221.
R. J. W. Le FBvre, D. V. Radford, and P. J. Stiles, J. Chem. Soc. (B), 1968, 1297.
lZB R. J. W. Le FBvre, D. V. Radford, G. L. D. Ritchie, and P. J. Stiles, J. Chem. SOC.(B), 1968,
148.
168 Chemical Thermodynamics
Watkins and Brey 130 measured l9F spin-lattice relaxation times in liquid
mixtures of C6Fg with cyc~o-C6H12, CsHe, 1,3,5-(CH3),C,H3, and with dimethyl-
formamide. They claimed that their experimental results showed evidence of
complexing for the two last-named mixtures but the evidence was not over-
whelming.
The suggestion of Le Fkvre et al. and of Baur et ul.12*that aromatic fluoro-
carbon + aromatic hydrocarbon mixtures form complexes in the liquid state
with small, but measurable, permanent dipole moments is at variance with the
recently published measurements of Dasgupta and Srnyth.l3l These authors
measured the dielectric constant and dielectric loss in the microwave region for
pure CeFs, pure C6H6, and their mixture. They observed that, although the
dielectric constant of C6F, is 10% lower than that of C6Hg,the dielectric loss is
appreciably greater, and the loss for an equimolar mixture is greater still. When
the loss is plotted against the logarithm of the frequency, curves with a typical
Debye shape result and it is indeed possible to analyse such data to give small
apparent dipole moments and relaxation times of around 1 ps. Dasgupta and
Smyth consider that this absorption is just the beginning of a wide and intense
absorption band reaching well into the far infrared. This band has been observed
for many other pure and mixed non-polar liquids132r133and is attributed to
collisional polarization. The present results can be explained by assuming that
the C-F bond dipoles interact with the benzene x-electron cloud thus giving
rise to an enhanced collisional absorption at higher frequencies without the
formation of a stable complex.
Barrett, Gill, and Steele 134 have made careful measurements of the intensity
and band shape of the infrared absorption band assigned to the a, C-F vibrational
mode of CgFgin cyc~o-C,H,, and in Cs&. The band in CsH6 moves to lower
wavelengths and broadens considerably. Their analysis of the band shape shows
that the molecular motion of C,F, in cyclo-C,H,, approximates to that of a
free rotor whereas there is restricted rotation in solution in C,H,. Although a
number of problems of interpretation still remain, the results are consistent
with the intermolecular interaction being short lived, stochastic, and due not
only to simple polar effects, and with the resulting forces being directed
perpendicular to the ring plane. There is a direct relation between atomic
polarizability and infrared band intensity and Barrett et al. note that their results
are inconsistent with the previously discussed finding of Baur et al. that the
atomic polarizability was a non-linear function of mole fraction for this binary
mixture.
Information concerning orientational correlations between anisotropic
molecules in the liquid state can be obtained from measurements of the Cotton-
Mouton effect and of depolarized Rayleigh light scattering. Le FBvre, Murthy,
and Stiles 135 studied C,F, +
C&g using the former technique and showed that
130 C. L. Watkins and W. S. Brey, J. Phys. Chem., 1970,74, 235.
S. Dasgupta and C. P. Smyth, J . Chem. Phys., 1974, 60, 1746.
13* M. Davies, G. W. F. Pardoe, J. E. Chamberlain, and H. A. Gebbie, Trans. Faraday Soc.,
1968, 64, 847.
133 A. M. North and T. G. Parker, Trans. Faraday SOC.,1971,67,2234.
13* R. M. Barrett, E. B. Gill, and D. Steele, J.C.S. Faraday Truns. 11, 1975,71, 532.
lS6 R. J. W. Le Fkvre, D. S. N. Murthy, and P. J. Stiles, Austral. J. Chem., 1969, 22, 1421.
Mixtures Containing a Fluorocarbon 169
their measurements were consistent with the mixture having pronounced
tendencies towards parallel alignment of neighbouring molecules. Exactly the
same conclusions were reached by Brown, Maguire, and Swinton136 who
examined the same system using depolarized Rayleigh scattering. They noted
that the experimental optical anisotropies for the bulk mixtures decrease with
increase in temperature whereas the opposite was true for the two pure compo-
nents. This behaviour indicates the presence of predominantly perpendicular
orientations in the pure liquids and face-to-face parallel arrangements in the
equimolar liquid mixture. An identical situation is observed in the solid sub-
stances as confirmed by X-ray crystallography. Brown and Swinton 13’ showed
recently on the basis of a crude calculation that these structures are predicted
when quadrupole-quadrupole interactions are taken into account. Although
quadrupole-quadrupole forces are likely to contribute little to the total inter-
molecular potential, it is probable that they play a major r61e in determining the
actual mutual orientations within an assembly of anisotropic molecules in both
the solid and liquid states. It is now generally accepted that electrostatic inter-
actions of this type, together with repulsive forces due to electron overlap,
largely determine the ground-state properties of many classical donor-acceptor
complexes, and also systems undergoing hydrogen-bonding i n t e r a c t i o n ~ . ~ ~ * - ~ ~ ~

4 Gas-Liquid Solubility Measurements

The dissolution of a gas in a liquid is basically a two-stage process. First a
cavity must be formed within the solvent of sufficient size to accommodate the
solute molecule, and this step is followed by the insertion of the solute within
the solvent cavity with a resultant change in the energy of the system which
depends primarily on the magnitude of the solvent-solute interactions. This
second stage of the process is now reasonably well understood in terms of the
statistical mechanical theories of fluid mixtures discussed briefly in previous
sections, although accurate predictions of thermodynamic excess functions are
only possible in a limited number of cases when auxiliary measurements enable
explicit values of 5 and x to be estimated. It is, as yet, not possible to predict
gas solubilities with comparable confidence.
Hildebrand and his co-workers have produced accurate gas solubilities for a
great variety of systems over a lengthy period, and many of their measurements
are tabulated in the comprehensive review of Battino and C 1 e ~ e r . lThey
~ ~ have,
however, persisted in the use of solubility parameter theory to correlate and
interpret their measurementsSs4 and in much of this work systems containing
fluorocarbons either as the gaseous solute or as the liquid solvent appear to
behave anomalously when compared with most other mixtures composed of
non-polar molecules.
When a comparison is made between fluorocarbons and hydrocarbons
containing equal numbers of carbon atoms, it is found that, under the same
198 N.M. D. Brown, J. F. Maguire, and F. L. Swinton, unpublished data.
197 N. M. D. Brown and F. L. Swinton, J. Chem. SOC.,Chem. Comm., 1974, 770.
138 M. J. S. Dewar and C. C. Thompson, Tetrahedron, 1966, Supp. 7, 97.
139 L. A. Carreira and W. B. Person, J. Amer. Chem. SOC.,1972, 94, 1485.
140 M. W. Hanna and J. L. Lippert, ‘Molecular Complexes’, Paul Elek, London, 1973, Vol. I, p. 1.
R. Battino and H. L. Clever, Chem. Rev., 1966, 66, 395.
170 Chemical Thermodynamics
conditions of temperature and pressure, fluorocarbon gases are less soluble than
the equivalent hydrocarbon when dissolved in the same non-polar non-
fluorinated solvent. The scanty experimental data seem to suggest that the
opposite is true when a fluorinated solvent is used. The other important experi-
mental generalization is that a fluorocarbon solvent will dissolve a much larger
amount of a particular gas solute than will the equivalent hydrocarbon. These
facts are illustrated by the representative selection of gas solubilities given in
Table 8.

Table 8 Selected gas solubility data 4~ 141-144 (the values given are 104x2where
31

x2 is the mole fraction of dissolved gas at p = 101.325 kPa and

T = 298.15 K)
Solvent
C-GH,, (C4F,)3N cs2 cat3 c6Ftl
32.7 68.83 12.72 20.9 38.42
10.34 148.1 2.06 5.75 45.61
- 227.4 107.9 148 -
24.51 474.8 11.1 156
- 52.0 - 8.15 24.18
15.2 61.05 4.87 8.8 24.0
- 42.36 2.22 4.42 17.95

Field et ~ 1 . have l ~ ~recently measured the solubility of six gases, including

CH4 and CF4, in methyl cyclohexane and in toluene. The solubility of CF4 was
found to be only one-third that of CHI under the same conditions. In contrast
P o ~ e 1 1 , ~working
4~ in Hildebrand's laboratory, has measured the solubility of
16 gases in CS, and in (C4F9),N. This latter compound closely resembles a
fluorocarbon in many of its physical properties and all the gases with the single
exception of hydrogen were found to have a higher solubility in it compared with
CS,. Powell noted that when the entropy of solution is extracted from the
experimental solubilities and is plotted against -Rlnx,, all the data for CSp
are well represented by a single straight line. When the data for the fluorinated
solvent are similarly treated it was found that the data for each class of gas
studied (noble gas, hydrocarbon, and fluorocarbon) followed different straight-
line relations.
The high solubility of many gases, including oxygen, in fluorocarbons has
led to these compounds being used as blood substitutes in cardiopulmonary
machines 1459146 and as potential carriers in liquid b~eathin8.l~'~
148 Tham et a1.149

have recently measured the solubility of 02,N,, and CO, in the three com-
mercially available fluoro-ethers: perfluoro-l,4-di-isopropoxybutane,perfluoro-
l-isopropoxyhexane, and perfluorobutyl-perfluorotetrahydrofuran and noted
142 R. J. Powell, J. Chem. and Eng. Dafa, 1972, 17, 302.
u3 E. Wilhelm and R. Battino, J. Chem. Thermodynamics, 1971, 3, 761.
lo4 L. R. Field, E. Wilhelm, and R. Battino, J. Chem. Thermodynamics, 1974, 6,237.
S. Howlett, D. Dundas, D. C. Sabistan, and M. C. Durham, Arch. Surg., 1965,91, 643.
146 H. A. Sloviter, H. Yamada, and S. Ogoshi, Fed. Proc., 1970, 29, 1755.
14' J. W. Sargent and R. J. Seffl,Fed. Proc., 1970,29, 1699.
148 J. H. Modell, E. J. Newby, and B. D. Ruiz,Fed. Proc., 1970, 29, 1731.
M. K. Tham, R. D. Walker, and J. H. Modell, J. Chem. andEng. Data, 1973, 18, 385.
Mixtures Containing a Fluorocarbon 171
that the solubility of O2 in these compounds is several times greater than the
solubility in many common organic solvents. The medical usage of fluorocarbons
is, of course, further encouraged by their chemical and physiological inertness.
Battino and his co-workers 150 have utilized the theory developed by Pierotti 161
to try to rationalize their experimental solubilities. Equation (8) is the basic
equation of this theory where xz represents the solubility of a solute dissolved
in a solvent of molar volume Vy at the standard pressurep*.

-RTlnx, = G, + Gi + RTln(RT/V;p*). (8)

G, is the partial molar Gibbs function associated with cavity formation and Gi
is the partial molar Gibbs function for the solutesolvent interaction. Pierotti
used an expression for G, derived from scaled-particle theory assuming a hard-
sphere potential. G, is thus a function of plythe solvent number density, and of
the diameters of the solvent and of the cavity. Ui and hence Gi were estimated
from the expression :

where q2,cI2 are the Lennard-Jones parameters for a solvent-solute interaction

obtained by assuming the Lorentz-Berthelot relations.
Battino et al. have found that the Pierotti theory is quite successful in
predicting the experimental solubilities of a wide range of binary mixtures
including many of those listed in Table 8. The wide range of experimental
solubilities is correctly reproduced and this is particularly satisfactory bearing in
mind the somewhat unrealistic model employed. The fact that a different inter-
molecular potential for the solvent is used in the calculation of G, and of Gi is
unfortunate and the theoretical equations do not account for the fact that the
size (and shape) of the solvent cavity will be affected by the presence of the solute
molecule. Not surprisingly partial molar enthalpies and entropies of solution
are not reproduced with the same precision as is the overall solubility.
The analysis shows, however, that the large solubility of gases in the aliphatic
and alicyclic fluorocarbons results from the relative ease of cavity formation in
these liquids. G,, for example, for perfluoromethylcyclohexane is only half that
for methylcyclohexane for a given solute molecule. Hexafluorobenzene is found
to occupy a position midway between these fluorocarbons and hydrocarbons in
the ease of cavity formation. Surprisingly Battino et a1.l6O find that the theory
is better at predicting the solubility of fluorocarbons such as CF, or SF, in
hydrocarbons than in fluorocarbon liquids. They suggest that this may be due
to a failure of the geometric mean rule used explicitly in the derivation of equation
(9). This explanation appears most unlikely in the light of current understanding
of the intermolecular interactions in fluorocarbon-fluorocarbon mixtures.
Note added in proof. This short appendix extends the literature coverage of the
chapter by including papers published up to late 1976.
The plea made by the present reviewer in Section 2A concerning the direct
calorimetric determination of H E for CF4 and CH, mixtures has been answered
160 F. D. Evans and R. Battino, J. Chem. Thermodynamics, 1971, 3, 753.
lS1 A. Pierotti, J. Phys. Chem., 1963, 67, 1840.
172 Chemical Thermodynamics
by Lewis and S t a ~ e l e ywho
l ~ ~ have measured this important function at 98.1 K.
Their experimental value for the equimolar mixture, 537 J mol-l, is at variance
with the value (1200 J mol-l) listed in Tables 1 and 3 which was estimated from
existing measurements of GE. The GE and HE measurements are obviously not
thermodynamically consistent unless HE itself proves to have an unusually
strong temperature dependence. Lewis and Staveley comment on this possibility,
and obviously more measurements of both HE and GE are required before the
matter can be finally resolved. When the new experimental value of HE is
inserted into Table 3 it is seen that both the LRS 1-fluid theory and the Flory
theory are now in reasonable agreement with experiment, with the former theory
slightly superior especially with respect to VE. Recent perturbation theory is
now also in fair accord with experiment for this sytsem, for, using the same values
of the interaction parameters listed in Table 3, Grundke et ~ 1 . predict l ~ ~
CE = 318 J mol-l, H E = 581 J mol-l, and VE = 1.76 cm3 mol-l.
Schneider and his co-workers have studied the phase diagrams for the three
binary mixtures CHF, + C2&, CF4 + C2H6, and CF, + CHF3, together with
+
the related ternary, CHF3 + CF4 C2H6, in the temperature range 120 to
200 K and at pressures up to 170 MPa.154s155 All three binaries exhibit upper
critical solution temperatures that increase with increase of pressure, and the
ternary phase diagram is a most interesting one with certain regions where three
partially miscible liquid phases can coexist. Mollerup and F r e d e n ~ l u n d have
l~~
studied low-temperature vapour-liquid equilibria in the commercially-important
refrigerant system CCl,F2 + CCIF,.
Dickinson and McLure 15’ have measured the excess volumes of four binary
mixtures C-CGF~,+ C-C~HI~,4- 193,5-Me3C6H3, CeFc, and + + MeeSi404
(octamethylcyclotetrasiloxane)at 338.2 K. All the YE’s are large and positive
the equimolar value for the last-named mixture exceeding 7 cm3 mol-l. Interest-
+
ingly, the values for c-C6F12 C,F6 are within 10% of the values for the corre-
sponding hydrocarbon mixture. A comparison of the excess functions of certain
of these mixtures with current theory is made in a subsequent publication156
and modest agreement is obtained.
Ott, Goates, and Cardon 159 have studied the thermodynamics of the formation
of the solid compound C6F,*CsHsin a manner very similar to that of Brennan,
Brown,and Swinton.lll Their basic conclusions are identical for they find a small
positive enthalpy of formation and also note that the compound is non-
stoichiometric. They make no mention of the low-temperature phase transition
noted in the earlier paper. Ott ef al. also examinedp-MeC,H, +CCI, and found
in complete contrast, good evidence for a strong stable equimolar compound
with a relatively large and negative enthalpy of formation.

lS2 K. L. Lewis and L. A. K. Staveley, J . Chem. Thermodynamics, 1975,7,847.

lSs E. W. Grundke, D. Henderson, J. A. Barker, and P. J. Leonard, Mol. Phys., 1973, 25,
883.
lS4 K. H. Peter, R. Paas, and G. M. Schneider, J. Chem. Thermodynamics, 1976, 8, 731.
ls6 R. Paas, K. H. Peter, and G. M. Schneider, J. Chem. Thermodynamics, 1976,8, 741.
lb6 J. Mollerup and A. Fredenslund, J . Chem. and Eng. Data, 1976, 21, 299.
lS7E. Dickinson and I. A. McLure, J. Chem. Thermodynamics, 1975,7,725.
lS8 E. Dickinson, D. C. Hunt, and I. A. McLure, J. Chem. Thermodynamics, 1975,7,731.
lS9 J. B. Ott, J. R. Goates, and D. L. Cardon, J. Chem. Thermodynamics, 1976,8, 505.
Mixtures Containing a Fluorocarbon 173
Masood et aE.leohave studied liquid mixtures of C,F, + CsHBby examinillg
dielectric and ultrasonic absorption. Their measurements can be interpreted by
assuming the presence of a significantconcentration of oriented pairs of dissimilar
species although the pair-order is relatively short lived. The same group has also
measured 161 the excess isentropic compressibility and the excess thermal pressure
coefficient for the same system over a short temperature range. They show how
probable differences in the structure of the pure liquids and of the binary mixture
can yield excess functions in qualitative agreement with experiment.
By far the most important recent paper concerning the nature of the inter-
actions in mixtures of hexafluorobenzene with aromatic hydrocarbons is that of
Bauer et al.ls2 who have examined depolarized Rayleigh scattering from
C,FB + C6H, and + 1, 3,5-Me3C6Ha. Both total-intensity and line-width
measurements were made and no evidence of long-lived complexes was obtained.
The results could be rationalized by assuming a strong unlike intermolecular
interaction that manifested itself both in the static and in the dynamic correlation
functions and in the individual molecule reorientation times. The single particle
reorientation times of each anisotropic component in both binary mixtures were
found to be slowed considerably when compared with their values in neutral
solvents such as CCI,.
Wilhelm and Battino 163 have extended their measurements of binary gas
diffusion coefficients using the Stefan technique 48 to mixtures of SF, with
41y

four aliphatic hydrocarbons. Jaster and Kosky 164 have completed a series of
measurements of the solubility of SF6 in four commercially-availablefluorocarbon
mixtures and Pierotti166 has published a recent review on the application of
scaled-particle theory to the estimation of gas solubilities.
l60 A. K. M. Masood, A. M. North, R. A. Pethrick, M. Towland, and F. L. Swinton, Adu. Mol.
Relaxation Processes, 1976, 9, 153.
161 A. K. M. Masood, A. M. North, R. A. Pethrick, M. Towland, and F. L. Swinton, J. Chern.
Thermodynamics, 1977, 9, 133.
D. A. Bauer, J. I. Braurnan, and R. Pecora, J. Chem. Phys., 1975, 63, 53.
lBS E. Wilhelm and R. Battino, J. Chern. and Eng. Data, 1975, 20, 399.
164 H. Jaster and P. G. Kosky, J. Chem. and Eng. Data, 1976,21, 66.
166 R. A. Pierotti, Chem. Rev., 1976, 76, 717.
6
Specific Interactions in Non-electrolyte Mixtures
BY A. G. WILLIAMSON

From the beginning of the study of non-electrolyte mixtures there have been two
extremes of approach to the question of how the contribution of the inter-
molecular interactions to the mixture properties should be handled. These two
approaches, exemplified by the work of Van Laarl and Dolezalek,2 are often
called ‘physical’ and ‘chemical’ or ‘non-specific’ and ‘specific’ interaction models.
Real systems whose behaviour is readily accountable in terms of one or other of
these approaches are well known. For example, the behaviour of mixtures of
simple spherical molecules can be fairly well described by theories based on the
principle of corresponding states and various simple equations of ~ t a t e . ~The
-~
physical models are in many ways more satisfying than are the specific inter-
action models. The relations between the physical system and the model are
better defined and the treatment tends to concentrate attention on the relations
between the unlike molecule interactions ~3 and the like interactions Eii, Ejj. The
major success of such treatments has been in relating the excess thermodynamic
functions GE, HE, and VE to one another with but one adjustable parameter for
any given mixture. However, these treatments have not reached the stage where
the properties of one mixture may be deduced from a knowledge of the pro-
perties of some other mixture. Recent developments in the non-specific theories
are discussed fully elsewhere.lO The remainder of this essay will be concerned
chiefly with the specific interaction models and the attempts that have been
made to use such models in correlating and predicting the properties of mixtures.

As a starting point for this discussion it is convenient to establish simple

references of mixing behaviour. Many of the models which will be discussed
make the simplification of separating the excess Gibbs energy GE into two parts,
an ‘athermal’ contribution which is independent of the molecular interactions
but depends on the molecule geometries and an energetic part which can depend
J. J. van Laar, Z . phys. Clzem., 1910, 72, 723; 1911, 76, 67; 1913,83, 599.
a F. Dolezalek, 2. Phys. Chem., 1908, 64, 727.
T. W. Leland, J. S. Rowlinson, and G. A. Sather, Trans. Faraday SOC.,1968, 64, 1447.
T. W. Leland, J. S. Rowlinson, G. A. Sather, and I. D. Watson, Trans. Faraday Soc., 1969,
65, 2034.
I , M. L. McGlashan, Trans. Faraday SOC.,1970, 66, 18.

J. S. Rowlinson and I. D. Watson, Chem. Eng. Sci., 1969, 24, 1565.

I. D. Watson and J. S. Rowlinson, Chem. Eng. Sci., 1969, 24. 1575.
A. J. Gunning and J. S. Rowlinson, Chem. Eng. Sci., 1973, 28, 521.
D. K. Astin and I. D. Watson, Austral. J. Chem., 1973, 26, 2071.
lo I. R. McDonald, in ‘Statistical Mechanics’, ed. K. Singer (Specialist Periodical Reports),
The Chemical Society, London, 1973, Vol. 1, Chap. 3.
174
Specific Interactions in Non-electrolyte Mixtures 175
on both molecular geometries and molecular interactions. The athermal part is
usually assumed to be adequately described by Flory’s l1 approximation
GE(ath) = E:ixiRTIn t#i,
which in turn gives
SE(ath) = &xiR In +i,
where +i is the volume fraction of component i. The complete expression for
the excess Gibbs energy is then constructed by adding an enthalpy or energy
term, the form of which depends on the details of the model used. There are a
number of particularly simple forms for the energy term which recur in dis-
cussions of specific interactions. These relate to simple assumptions about the
nature of the interactions and are outlined below.
If one assumes that interactions between molecules are randomly distributed
throughout the mixture on a number basis then the change in enthalpy H E
(or energy*) on mixing is given by the expression
HE = Ax1x2(2w12 - W11 - WZZ), (3)
where Wij is an energy characteristic of a contact between molecules i and j and
A is a parameter, dependent on the liquid ‘geometry’, which describes the number
of ‘contacts’. A is assumed to be independent of composition.
If one assumes that the interactions are randomly,distributed on a surface-
area basis then the expression for H E becomes
HE= xlx2ala2(2wl2 - W11 - W22)Kxl.l + xz4, (4)
where alp2are the number of units of area and the Wij are energies of interaction
per unit area of ‘contact’.
Finally, if one assumes that the interactions are randomly distributed on a
volume basis the expression for HE becomes
HE= xlxa;v,v,(2wl2 - w11 - w22)/(x1K + x2v,), (5)
where and yZ are the molar volumes and the wij are the energies of interaction
per unit volume.
The formulae can be generalized to
HE = Axlx2V: Yt/(xl V r + x2V;), (6)
where n has the values for special cases given in Table 1. Figure 1 illustrates the
effect on equation (6) of the assumptions n = 0, 8, and 1 for Y,/Y, = 3. The
remaining feature which is common to many discussions of specific interactions
is the quasi-chemical approximation. In its most primitive form this is just
Dolezalek‘s theory in which all deviations from ideal-mixing behaviour, defined
by
pi = py RTlnxi, + (7)
l1 P. J. Flory, J . Chern. Phys., 1942, 10, 51.
* Whether these equations refer to enthalpy of mixing at constant pressure or energy of mixing
at constant volume depends on the assumptions one makes about the volumetric behaviour of
the system on mixing.
176 Chemical Thermodynamics
Table 1 Special forms of equation (6)
Common name of
n Ruation approximat ion Source
0 3 simple solution Guggenheim l2
3 4 surface interaction Langmuir l3
spherical molecules
1 5 regular solution Hildebrand l4

and implying
H E = 0,
are assumed to arise from the formation of involatile complexes which them-
selves mix ideally with the primary components. In this way Dolezatek explained
negative deviations from ideality (e.g. in mixtures of chloroform and acetone)
as the result of complex formation of the type:

A + D = AD,
with an equilibrium constant:
K, = XD~XAXD.

Positive deviations (e.g. in alcohol +

hydrocarbon mixtures) are assumed to be
caused by self-association of one or both components according to equations

L
a
W
\
s

*2
Figure 1 Comparison of composition dependence of H E for random interactions by
number, surface, and volume for spheres of relative volumes 3 : 1

E. A. Guggenheim, ‘Mixtures’, Oxford University Press, 1952, Chap. 4.

lS I. Langmuir, Colloid Symposium Monograph, Chemical Catalogue Co., New York, 1925,
Vol. 3, p. 48.
l4 J. H. Hildebrand and R. L. Scott, ‘Solubility of Non-electrolytes’, Reinhold, New York, 1950,
3rd edition.
Specific Interactions in Non-electrolyte Mixtures 177
such as
A + A = A%.
Although the ideal associated model fell into disrepute, largely because of
attempts to use it to account for all deviations, it clearly has considerable merit.
This is particularly so in cases, such as that of mixtures of chloroform + acetone,
where a very specific complexing interaction is indicated by ‘molecular’ properties
such as the n.m.r. spectrum, which has been used to establish16 an equilibrium
constant for the reaction
Me,CO + HCCl, = Me,C=O.-H-CCl,, (12)
of K , = 1.8 +_ 0.6. Following the pioneering work of Huggins, Pimentel, and
Shoolery the method has been used to determine both equilibrium constants
and enthalpies of formation for a number of ‘complexes’. Alley and Scott16
studied a number of mixtures of C,Fl5H with electron-donor solvents such as
ketones and amines. Watson and Williamson l7 examined the hydrogen bonding
for the case of formation of two complexes:
A+ D = AD, (1 3)
A + AD = AaD, (14)
in mixtures of C,F,,H + 1,4-dioxan. Barker, Watson, and Williamson l8
determined equilibrium constants and from their temperature coefficients, the
enthalpies of formation of complexes in a number of mixtures of di-n-alkyl
ethers + chloroform. They then used these data to estimate the enthalpies of
mixing of the ethers with chloroform and compared their results with independent
calorimetric measurements as shown in Table 2. A similar study has been

Table 2 Enthalpies of mixing of chloroform and ethers at 298.15 K

Ether Di-n-propyl Di-n-pentyl Di-isopropyl Di-tert-butyl
ether ether ether ether
1.2 1.02 2.41 1.33
23;/umol-1
HE(x 0.5)/J mol-l calc.
=
- 10.8
-1770
- 10.0
-1490
- 11.7
- 2650
-7.9
- 1360
calori- -1940 -1720 - 2960 - 1490
metric

carried out by Jolley, Hughes, and Watson19 in which they have, using only
n.m.r. measurements, reproduced the calorimetric enthalpies of mixing for a
number of mixtures of chloroform and thioethers within about 15 per cent.
It is interesting to note that in all cases the n.m.r. predictions are numerically low.
This almost certainly results from the fact that the n.m.r. ‘sees’ only the hydrogen-
bonding interaction whereas the enthalpy of mixing reflects the effects of other
interactions. This is perhaps made clear by the fact that the enthalpies of mixing
C.M.Huggins, G. C. Pimental, and J. N. Shoolery, J. Chem. Phys., 1955,23, 1244.
1s S. K. Alley and R. L. Scott, J. Phys. Chem., 1963, 67, 1182.
1’ I. D. Watson, R. J. Knight, I. R. McKinnon, and A. G. Williamson, Trans. Furaday Suc.,
1968,64, 1763.
1% J. R. Barker, I. D. Watson, and A. G . Williamson, Austral. J. Chem., 1971, 24, 2047.
la K. W. Jolley, L. M. Hughes, and I. D. Watson, Austral. J. Chem., 1974,27, 287.
I78 Chemical Thermodynamics
of ethers and carbon tetrachloride are negative and about 10 to 20 per cent of the
enthalpies of mixing of the corresponding chloroform + ether mixtures. 2o
More sophisticated treatments of complex formation take two further effects
into account. It is possible to formulate the expressions for the excess functions
500

400

30c
‘d

E
d

3!
\

wz 2oc

1oc

Figure 2 Enthalpies of mixing of acetone + chloroform at 298.15 K; - -, curve for

1840x,xz cal mol-I; -, curuefor system with two equilibria: Kl = 0.967, Kz = 1.117,
Ill = 2.42 cal m o F , and h2 = 3.29 cal mol-I; 0, experimental results

so as to allow for more than one complex. McGlashan and Rastogi 21 developed
such a treatment to account for the excess Gibbs energies and excess enthalpies
of mixing of chloroform +
1,4-dioxan by assuming two equilibria:
/CH,--CH, /CH,--CH2
1
\
Cl&’H + 0\ /
0 = C13CH-..0
\ /
0 (15)
CH,-CH, CH,-CH,
and
CH2-CH /CH2-CH,
/ \ 1
2C13CH 0 +
/ /
0 = C13CH*.+0
\ /
O...HCCI, (16)
CM,--CH, CH,--CH,
The existence of two complexes in this case is clearly implied by the chemical
structures and by the unsymmetrical behaviour of the excess functions. Kearns 22
L. A. Beath and A. G . Williamson, J. Chem. Thermodynamics, 1969, 1, 51.
81 M. L. McGlashan and R. P. Rastogi, Trans. Faraday Sac., 1958, 54, 496.
E. R. Kearns, J. Phys. Chem., 1961,65,314.
Specific Interactions in Non-electrolyte Mixtures 179
later used the same model to account for the unsymmetrical enthalpy of mixing
of chloroform and acetone shown in Figure 2. As was shown in Figure 1, the
simple models lead to the expectation that the maximum in the curve of enthalpy
of mixing us. mole fraction would be displaced towards the component of lower
molar volume. However, in the chloroform +
acetone mixtures the maximum
is displaced towards the component of higher molar volume. Although Kearns'
physical explanation of the cause of this (the formation of complexes of the type
c1 CI
I I
Me&= 0 ...H -C-Cl*** H-C-CI)
1 I
c1 c1
is open to doubt the mathematical structure used is certainly adequate to describe
the shapes of the excess functions.
The extreme forms of the simple association model envisage chain association
of the type
A f A = A2, (1 7)
As f A = At+l, (1 8)
with equilibrium constants K1,2,K2,3, .., . In the Mecke-Kempter model
Kl,f+l = K (all i 2 1). (1 9)
This model has been used recently by Renon and Prausnitz23to interpret the
properties of alcohol +
alkane mixtures.
A somewhat more productive extension of the association model has been
. ~ ~ for the enthalpies of mixing of C7F15H+
that used by Anderson et ~ 1to account
acetone shown in Figure 3. In this case it is abundantly clear that simple com-
plexing alone cannot account for the shape of the curve. However the assumption
of non-ideal association with a 1 : 1 complex being formed exothermically in a
mixture which also contains non-specific endothermic interactions (equation 5 )
provides an overall expression capable of yielding both a maximum and a
minimum. This separation of effects is partially confirmed by the estimates of
the hydrogen bonding (complex formation) by n.m.r. spectroscopy and the
estimates of the non-specific term obtained from enthalpies of mixing of the
+
partially miscible system C7F16 acetone. This approach was extended by
+
Watson et a1.l' to the system C7F15H 1,4-dioxan in which, following McGlashan
and RastogiY2ltwo complexes were assumed and, following Anderson et al.,24
a non-specific term was introduced. Estimates of the excess thermodynamic
functions GE and H E based on n.m.r. measurements for the specific contributions
and the upper consolute temperature of mixtures of C7F16 + 1,4-dioxan for the
non-specific term agreed with the directly measured values within a few per cent.
One of the most interesting applications of the non-ideal associated model is
in the work of McGlashan, Stubley, and Watts 25 and of Howell and Stubley 2*
H. Renon and J. M. Prausnitz, Chem. Eng. Sci., 1967, 22, 299.
a4 D. L. Anderson, R. A. Smith, D. €3. Myers, S. K. Alley, A. G. Williamson, and R. L. Scott,
3. Phys. Chem., 1962, 66, 621.
2b M. L. McGlashan, D. Stubley, and H. Watts, J. Chem. SOC. (A), 1969, 673.
a6 P. J. Howell and D. Stubley, 3. Chem. SOC.(A), 1969, 2489.
180 Chemical Thermodynamics
on mixtures of carbon tetrachloride with methyl-substituted benzenes. The
enthalpy of mixing curve of carbon tetrachloride +
toluene which, over a wide
temperature range has two maxima and a minimum, is explained in terms of an
exothermic contribution from a 1 : 1 complex and an endothermic contribution
from the simple mixture formed by the primary components and the complex.

1 I I I I I I I I I
0.5 1

x(OCMe2)

Figure 3 Enthalpies of mixing of C,F15H + OCMe,

Although both of these effects are symmetrical in mole fraction their com-
position dependences differ as described in equations (20) to (24) and illustrated
in Figure 4. The equilibrium constant for the simple mixture with complex
formation can be written

where w is a free energy of interchange defined in the usual way l2 for the lattice
model and K a is an equilibrium constant. The enthalpy of mixing is given by
H E = [(AH& - L 4 2 ) + ~ ( -l x)Lu, (21)
Specific Interactions in Non-electrolyte Mixtures 181

0.2

0.1

0
0 0.5 1
X
Figure 4 Physical and chemical interactions for simple system; numbers of complexes
2 per mole of mixture from Z/(l - 2)= Kx(1 - x)/(l + K ) A with A chosen to
normalize 2 at 0.25 for x = 0.5. - - K = 3; - K = 03 ; -, x(1 - x ) vs. x
-y -y

I I I 1 I I I I I I 1
0 0.5 1
x ( C6H,CH,)
Figure 5 Enthalgies of mixing of carbon tetrachloride +
toluene: 0,experimental
results; -, calculated from equilibrium constants and interchange energies given &y
McGlashan, Stubley, and Watts 25
182 Chemical Thermodynamics
where
u = w - Taw/aT,
and AHiB is the enthalpy of formation for the complex. When the two effects
are similar in magnitude and opposite in sign as would be expected for mixtures
of carbon tetrachloride with aromatics then it becomes possible for the enthalpy
of mixing to develop the shape shown in Figure 5. Furthermore, in a mixture of
this kind one would expect to find that the variation in KABwith temperature
would lead to a rapidly varying shape for the curve of H E . That this is indeed
so is shown in Figure 5.
Fenby and Scott 27 expanding on the earlier studies of Swinton 28 and of
Morcom29studied the enthalpies of mixing of 34 pairs of fluorinated benzenes.
They interpreted their results in terms of ‘physical’ (non-specific) and ‘chemical’
(specific) interactions which were assumed to obey regular-solution and chemical-
equilibrium equations respectively. In interpreting the behaviour of these
systems two kinds of ‘chemical’ interaction were invoked. One of these was
believed to be a charge-transfer30 interaction and one was believed to be a
specific interaction between matching hydrogen and fluorine atoms on adjacent
rings. This latter interaction was introduced to account for the fact that the
mixtures of pairs of compounds exhibiting a matching fluorine-hydrogen ‘lock
and key’ relationship were very much more exothermic than those of otherwise
similar systems. For example, mixtures of p-difluorobenzene with 1,2,4,5-
tetrafluorobenzene are formed more exothermically than are mixtures of
p-difluorobenzene with any other tetrafluorobenzene. Although the two
‘chemical’ effects have not been separated quantitatively the relation between
the specific (mole-fraction symmetrical) term and the non-specific (volume-
fraction symmetrical) term is dramatically illustrated in Figure 6 taken from
Fenby and Scott’s paper. In mixtures of C,H5F (Vm = 110 cm3mol-l) with
benzene (V, = 89 cm3mol-l) the fluorocarbon has the larger molar volume and
the positive contribution to HEhas its maximum displaced towards pure benzene.
In mixtures of C6H5F(Vm = 110 cm3mol-l) with CIHO*C8H6 (V, = 156 cm3mol-l)
the relation between the molar volumes is reversed, and because the chemical
and physical effects are similar in magnitude this results in a complete reversal
of the shape of the curve of HE us. x . Fenby 31 has recently reviewed the extensive
work carried out over the past few years on fluoroaromatic mixtures. The
original view that the chemical effect was a charge-transfer interaction is now in
considerable doubt. Recent measurements by Fenby and Ruenkrairergsa 32 of
the enthalpies of mixing of chloro-, iodo-, and bromo-pentafluorobenzene with
benzene are consistent with the previous measurements on fluorobenzenes and
support the view that the complexing is not of the chargetransfer type.

27 D. V. Fenby and R. L. Scott, J. Phys. Chem., 1967,71,4103.

28 W. A. Duncan and F. L. Swinton, Trans. Faraday SOC.,1966,62, 1082.
29 A. Andrews, K. W. Morcom, W. A. Duncan, F. L. Swinton, and J. M . Pollock, J. Chem.
Thermodynamics, 1970, 2, 95.
so R. S . Mulliken and W. B. Person, ‘Molecular Complexes’, Wiley-Interscience, New York,
1969.
31 D. V. Fenby, Rev. Pure and Appl. Chem., 1972,22, 55.
aa D. V. Fenby and S. Ruenkrairergsa, J . Chem. Thermodynamics, 1973,5,227.
Specific Interactions in Non-electrolyte Mixtures 183
The models discussed so far are ones in which each mixture is treated as a
special case and quantitative information is generally not transferable from one
mixture to another. Between the limiting cases of the very strong interactions
which can be approximated by chemical equilibria and the overall (non-specific)
physical interactions there lies a whole spectrum of interactions varying widely

0 0.5 1
x C&H)
Figure 6 Enthulpies of mixing of benzene and of n-butylbenzene with pentufluorobenzene:
A, benzene mixtures at 298.45 K; B, n-butylbenzene mixtures at 298.15 K

in both strength and specificity. To deal with this situation a number of models
have been evolved, all of which aim (1) to deal with the possibility of more than
one kind of specific interaction in a given mixture and (2) to enable the transfer
of interaction data from one system to another.
All of the models attempt to achieve these aims by envisaging the molecules as
being composed of strings of identifiable segments or groups between which
interactions occur to an extent determined by the relative areas of the groups in
given molecules which are exposed to contact with other molecules. In this
sense these models trace their origins back to Langmuir.lS The most compre-
hensive treatment of this kind is that developed by Barker 33 on the basis of earlier
ss (a)3. A. BaTkeT, J. Chem. Phys., 1952, 20, 1526; (b) J. A. Barker, 3. Chem. Phys., 1953, 21,
1391; (c) J. A. Barker and F. Smith, J. Chem. Phys., 1954, 22, 375.
7
184 Chemical Thermodynamics
work by T ~ m p aGuggenheim,l2S
,~~ 35 and others. Where possible we shall discuss
other treatments in terms of Barker's treatment.
Barker's model assumes (1) that all molecules can be represented as open-
chain structures composed of freely jointed identifiable segments, (2) that the
available configurations of the molecuIes in the liquid mixtures can be adequately
represented by arrangements of the segments on a fixed lattice, (3) that the only
significant interactions are those between nearest-neighbour segments of different
molecules, and (4) that the changes in thermodynamic properties on mixing are
adequately represented by changes in the configurational partition function.
Barker assumes that the distribution of contacts between segments is controlled
by a quasi-chemical relation which he writes in the form

where

and EiB is the 'interchange energy' la

EP = + c;").
E P- &(EP (25)
The quantities NiB are the numbers of contacts between segments of types i
and j on molecules A and B respectively. Using the approximation described by
Guggenheim,12 Barker's treatment leads to the final equations for the thermo-
dynamic properties.
p 3 R T = C,[Qf In(Xf/Xf'xA>] + r&Z - 1) In &(rBxs/ra) (26)
and
U" = 2C,Ci>j(XfX4 - ~ a X p ' X f ' ) q p H F+ 2CA2BCi>jXPXF&BH$B,
(27)
where
H i B = E;"fB- T(2EfiB/aT), (28)
Qf denotes the number of contact sites of type i on molecule A, and
ZjQf = ra(Z - 2) + 2, (29)
where 2 is the coordination number of the lattice.
The Xf and Xf' are obtained by solving the equations:

Although the lattice model of the liquid state, upon which equations (26) and
(27) are based, is open to criticism the treatment does describe a well-defined
procedure for dividing a molecule into segments, counting the segment inter-
actions of various kinds, and relating these to the thermodynamic properties of
the mixture. The lattice model has frequently been applied at any one of three
(a, H. Tampa, J. Chem. Phys., 1953, 21, 250; (b) H. Tompa, Trans. Faraday Soc., 1949, 45,
101.
(u) E.A. Guggenheim. Proc. Roy. SOC.( A ) , 1944, 183, 203; (6) E. A. Guggenheim, Proc.
Roy. SOL (Aj, 1944,183, 213.
Specific Interactions it? Non-electrolyte Mixtures 185
levels of approximation. These are (1) the quasi-chemical approximation as just
described, (2) the quasi-chemical approximation with temperature-independent
interaction energies i.e. H i B = E p , (3) the zeroth approximation, which assumes
that the segment contacts are randomly distributed despite their non-zero Eli.
This approximation corresponds to putting qij = 1 in equations (30) and (31).
The values of X so obtained are then used to calculate UE(zeroth) using equation
(27) with qeB = 1, H i B =I= 0, and pz(atherma1) using equation (26). The excess
Gibbs free energy is then obtained from
GE(zeroth) = &xApE(ath) + UE(zeroth).
One of the chief difficulties in using Barker’s approach is that it is possible to
invent unduly large numbers of segment types in an attempt to account for the
behaviour of a particular system. This has two consequences. Firstly it introduces
a great deal of coniputation and secondly it obscures the physical significance of
the various parameters. The ready availability of computers has reduced the
first of these problems to a triviality but the second can be overcome only by the
introduction of simplifying assumptions and by the exercise of a rigorous
economy in the choice of interaction parameters. One simplifying assumption
which is common to nearly all versions of Barker’s treatment, but to which
attention is rarely drawn, is that E p and H P are taken to be independent of
environment (i.e. independent of A and B). This assumption not only reduces
the number of energy parameters drastically but is also the key assumption in
the use of the group-contribution method for predicting data for one system
from measurements on others. The choice of the co-ordination number 2
appears to be non-critical. It has been shown in a number of studies12s36that
choices of 2 between 4 and 12 have little effect on the shapes of the curves of
GE and H E 11s. x for binary mixtures, and that the effects of changes in 2 for a
given system can be largely oKset by corresponding changes in the Eu. This
implies that the lattice aspect of the model is relatively unimportant and that the
value of 2 may be chosen for convenience. However, since the values of 2 and
Eg are closely linked it is important that the same value of 2 be maintained
throughout any given correlation scheme. In dealing with mixtures of organic
compounds the choice 2 = 4, though not reasonable on physical grounds, is
particularly convenient in relation to the structure and segmentation of the
molecules. The choice of Y is important in deciding the segmentation of the
molecules, and different choices of r lead to significant differences in the shapes
of the curves of excess functions for a given mixture. In terms of the physical
model it is reasonable to choose Y to be integral and to reflect the relative molar
volumes of the components. However, if equations (26) to (31) are not regarded
as describing a model to be adhered to rigidly but rather as the basis of a
correlation scheme, then the choice of Y becomes more flexible,37 and some
workers38 have even allowed the relative values of rA and YB to change with
temperature. The choice of the Qi by division of the molecules into segments is
M. L. McGlashan, K. W. Morcom, and A. G. Williamson, Trans. Faraday SOC.,1961, 57,
601.
G. R. Bruce and G. N. MalcoIm, J. Chem. Thermodynamics 1969,1, 183.
D . H. Everett and R. J. Mum, Trans. Faraday SOC.,1964, 60, 1951.
186 Chemical Thermodynamics
usually based on a blend of intuition and molecular geometry. This choice is,
like those of 2 and r, flexible and is itself dependent on the choices made for 2
and r. It is not always possible to be consistent in the choice of 2, r, and Qi.
For example, in dealing with mixtures of n-alkanes with di-n-alkyl ethers one
may choose say -CH3 and -C2H4- as i and j segments for the n-alkanes to
give consistency between r and the molar volumes of the alkanes. However, if
one chooses to regard di-n-propyl ether as i-j-k-j-i with r = 5 and
V, = 138 cm3mo1-1 then one creates an inconsistency with n-hexane for which
r = 4 and Vm = 132 cm3mol-l. A similar problem arises in many ether mixtures
and it is notable that Malcolm et ~ 1and. other~ ~authors *O found that they could
improve the fit of Barker’s equations to their results for polyether solutions by
arbitrarily aItering the value of Q for the ether oxygen.
When z, r, and the Qi have been chosen, by whatever means, one still has to
find the set of Ed and Hu which best fit the measured results. For mixtures
composed of simple molecules such as n-alkanes only it may be possible to
envisage but one ij interaction, and in such cases it is usually found that the
choices of r and Qi have more effect on the shape of GE or UE us. x than do
EG and f l u . When more complicated molecules are dealt with the number of
Eu (and Hu) parameters increases rapidly. For molecules containing a total of
only four unlike groups it is necessary to evaluate 6Eij and 6Hjj in order to fit
GE and UE. Choosing the ‘best’ set of parameters is complicated by the fact
that the relative numbers and magnitudes of the contributions to the energy of
any particular ij pair are dependent on all the Eu and Hg. Of particular interest
in this respect is the zeroth approximation in which the relative numbers of
interactions are independent of the Eu so that the contributions of the number
of interactions and their energies are separable. This in turn means that it is
possible to examine more or less separately the effects of the different ij inter-
actions on the shape of the HE or GE us. x curve. In the limiting case of the
zeroth approximation for mixtures with only one ij interaction the shape and
magnitude of the curve are independent. Whichever approximation is used the
idealized method of using Barker’s equations as a correlation-prediction scheme
would consist of a value of 2, a set of Eg, Hg, and a prescription for choosing Y
and the Qi for any particular molecule. Ideally the values of Eu, Hij would have
been established from measurements on a set of systems each of which involved
only one ij pair, or from measurements on a series of systems of increasing
complexity, each of which involved previously determined interactions and one
new pair. In either case an unequivocal set of parameters would be established
which correlates the properties of the reference systems and which may be used
for the prediction of properties of new mixtures involving the same groups.
The second of the methods described for establishing the parameters is the only
practically worthwhile one. Unfortunately it has the disadvantage that errors
in the determination of r, Qi, Eu, Hd will accumulate and, though this has little
effect on the correlation of data within the reference set it can cause serious
errors in the prediction of values for systems not included in the reference scheme.
sB G. N. Malcolm, C. E. Baird, G. R. Bruce, K. G. Cheyne, R. W. Kershaw, and M. C. Pratt,
J. Polymer Sci., 1969, 7, 1495.
40 J. A. Barker, I. Brown, and F. Smith, Discuss. Faruday SOC.,1953, 15, 188.
 Specific Interactions in Non-electrolyte Mixtures 187
In the past several attempts have been made to set up tables of interaction
parameters to be used for predictive purposes. Barker and co-workers applied
33g

+
the method to mixtures of alcohols benzene, alcohols + carbon tetrachloride,
and alcohols + chloroform. In their earlier work it was assumed that equations
(26) to (31) could be applied directly to CE and HE at constant pressure and the
equations were used with Hij = EG. Although the agreement between the theory
and experimental results was not outstanding it was clear that the method was
capable of dealing with very unsymmetrical GE and H Ecurves. Goates, Snow,
and James41 obtained a fairly good correlation of the enthalpies of mixing of
methanol + benzene, ethanol + benzene, and ethanol +
cyclohexane using
temperature-independent Eij and assuming only four significant interaction types
H-0, H-S, I-S, and 0-S where H = alcohol hydrogen, 0 = alcohol oxygen,
S = aromatic hydrogen, and I = aliphatic hydrogen. Goates, Snow, and Ott 42
obtained reasonable correlations of the enthalpies of mixing of five alcohols
with benzene, toluene, and ethyl benzene using five interaction energies. In their
correlation only one value of the O-H interaction was used but the values of
some of the other Eg’s were changed from system to system to improve the fit.
Dacre and Benson’s study 43 of mixtures of methanol, ethanol, n-propanol,
n-butanol, and n-octanol, with carbon tetrachloride represents the first attempt
to apply Barker’s equations systematically to the excess Gibbs energy CE and
the more appropriate 44 excess energy at constant volume UF rather than the
369

excess enthalpy at constant pressure. Though they used temperature-independent

Eu’s Dacre and Benson did carefully restrict the number of such parameters to
three. They found that the major interaction (O-H) could be kept constant but
that changes in the minor interactions were required to improve the fit. The
major defect of their correlation was the failure of the equations to reproduce
the fall in UF at high mole fractions of alcohol. Later experimental work on
ether + carbon tetrachloride 20, 45 suggests that the oxygen-chlorine interaction
should be negative. This and the inclusion of temperature-dependent Eij’s do
lead to rather better correlations 46 than those obtained by Dacre and Benson.
Ott, Goates, and Snow 47 constructed a table of interaction parameters
applicable to the enthalpies of mixing of a wide range of compounds including
alkanes, aromatics, and halogenocarbons. However, they did not attempt to
use these parameters to predict enthalpies of mixing of any systems not used in
constructing the correlation. Beath, Watson, and Williamson 48 in their study
of mixtures of di-n-alkyl ethers + carbon tetrachloride + chloroform used
temperature-dependent Eij’S and also compared the calculated energies of mixing
with experimental values of U$. As in the work of Dacre and B e n s ~ nsome- ,~~
what arbitrary assumptions were made. In order to generate a physically

41 J. R. Goates, R. L. Snow, and M. R. James, J. Phys. Cbem., 1961,65, 335.

J. R. Goates, R. L. Snow, and J. B. Ott, J. Phys. Chem., 1962, 66, 1301.
43 B. Dacre and G . C. Benson, Canad. J. Chem., 1963, 41,278.
44 R. L. Scott, J . Phys. Chem., 1960, 64, 1241.
45 D. F. Gray, 1. D. Watson, and A. G . Williamson, Austral. J. Chem., 1968, 21, 379.
46 D. F. Gray and A. G. Williamson, unpublished calculations.
*’ J. B. Ott, J. R. Goates, and R. L. Snow, J. Phys. Chem., 1963, 67, 515.
48 L. A. Beath, I. D. Watson, and A. G . Williamson, Proceedings of the 1st International Con-

ference on Calorimetry and Thermodynamics, P.W.N., Warsaw, 1971, p. 773.

188 Chemical Thermodynamics
reasonable model of the molecules it was necessary to envisage three ij inter-
actions in the carbon tetrachloride mixtures and three more in the chloroform
mixtures. The interaction energies shown in Table 4 reproduced GF and Ug for
the reference systems di-n-propyl ether +
carbon tetrachloride and di-n-propyl
ether + chloroform very well. However, the prediction of the thermodynamic
properties of mixtures of other ethers with chloroform and with carbon tetra-
chloride was disappointing. It appears that di-n-propyl ether is an unfortunate

Table 3 Interaction parameters for ether -k chiorocarbon mixtures: Z = 4, I =

alkyihydrogen, H = chloroform hydrogen, r(CC1,) = 5 , r(CHC1,) = 4,
r = n + 1 for ether with n carbon atoms
Interaction type Cl-I C1-0 0-1 CI-H H-I H-0
E/J mol-’ +lo0 -699 0 0 +415 -5417
H/J mol-l 112.5 -276 0 0 +418 -8533

choice of reference system. Similar difficulties were encountered by Kershaw and

Malcolm 49 in their studies of mixtures of polypropyleneglycol dimethyl ether
with carbon tetrachloride and with chloroform. Malcolm et and Bruce and
continuing the study of polyether +
chlorocarbon mixtures, found
that better agreement with experiment could be obtained if the number of
oxygen contacts was arbitrarily reduced. Pasco, Fenby, and Bissel 51 have
recently examined the thermodynamics of mixtures of di-n-hexyl ether with
carbon tetrachloride and with chloroform and have correlated their results with
those of Bruce and Malcolm.60 The general conclusions arising from the work
on ether + chlorocarbon mixtures are that Barker’s treatment is capable of
correlating thermodynamic properties of mixing of systems containing the same
functional groups in similar intramolecular environments, but that minor
changes in the molecular structure can cause significant changes in the values of
Eij. Recently Bissel and Williamson52have measured GE and H E for mixtures
of n-alkanes with carbon tetrachloride and with chloroform. These measure-
ments were part of a programme to establish a systematic set of interaction
parameters applicable to all mixtures of alkanes, ethers, and chlorocarbons.
Of particular interest were the questions whether or not the Eij’s are transferable
from one molecular class to another, and whether or not the quasi-chemical
approximation is significantly better than the In dealing with a wide
range of systems it is difficult to arrange to evaluate all the Eij’s separately and
it often becomes necessary to evaluate two or more Eij’s simultaneously. In the
Reporter’s laboratory we have examined several procedures for the simultaneous
choice of a best pair of Ejj’s. Of the optimization methods studied only Rosen-
brock’s 53 was moderately successful. However, the most satisfactory procedure
we have found is that illustrated in Figure 7. A computer program was written
to find pairs of EU,E k l for which the sum of squares of deviations, u2, between
4B R. W. Kershaw and G. M. Malcolm, Trans. Faraday SOC.,1968,64, 323.
50 G. R. Bruce and G . N. Malcolm, J . Chem. Thermodynamics, 1969, 1, 182.
61 N. F. Pasco, D. V. Fenby, and T. G. Bissel, J. Chem. Thermodynamics, 1974, 6, 1075.
52 (a) T. G. Bissel, G. Okafor, and A. G . Williamson, J . Chem. Thermodynamics, 1971, 3, 393;
( b ) T. G. Bissel and A. G. Williamson, J . Chem. Thermodynamics, 1975, 7 , 131.
s3 H. H. Rosenbrock, Computer J., 1960, 3, 175.
Specific Interactions in Non-electrolyte Mixtures 189
the computed and experimental excess functions is constant. These Eij, Ekl are
then plotted as a set of contours, one for each of a number of values of 0,and
from the contour map so produced it is possible to choose visually the best pair
Of Eij, Ekl.

-40
I I
-250 0 + 250
E (a-0) / J mol-’
Figure 7 Contours of constant deviation for carbon tetrachloride + di-n-propyl ether
with three energy parameters: E(1-0) = 10 J mol-1

Figure 7 illustrates the results of this procedure for the choice of E1-0 and
EC1-0 for di-n-propyl ether + carbon tetrachloride. This shows clearly the
existence of two sets of optimum values of opposite signs. The reason for this
becomes clear when one recognizes that there are similar numbers of contacts
of the two kinds and therefore inversion of the labels will have little effect on the
computed values of GE and H E . This situation is not uncommon and it is likely
that an optimization involving the simultaneous determination of more than two
Eij’s would have several saddles in the a2 us. Eij, ..., Ekl surface. Although the
particular choice of the optimum set of Eij’s does not matter for the system being
fitted it could be very important in the application of these parameters to other
systems in which the relative numbers of interactions of different kinds are very
different from those in the test system. In the case illustrated we have chosen to
assign the negative value to EO-CI. The set of Eu parameters was built u p
according to the pattern shown in Table 4 for Z = 4 and lattice units -CH,,
-CzH4-, -0-, -C1, and -CH. Interaction energies for both the quasi-
chemical approximation and the zeroth approximation are shown in Table 5 .
With these parameters it is possible to correlate the excess Gibbs energies and
the energies of mixing of the eight systems from which the parameters are derived
within better than 10 per cent. As has been noted in previous studies, there was
little to choose between the zeroth and the quasi-chemical approximations.
Attempts to predict GpE and UF for other systems met with mixed success. The
190 Chemical Thermodynamics
Table 4 Stepwise procedure for solving for the interchange free energies, Eu, and
the interchange energies, Hu
Key to the systems: (1) n-hexadecane + n-hexane
(2) chloroform + carbon tetrachloride
(3) n-hexane + carbon tetrachloride
(4) n-heptane + carbon tetrachloride
(5) n-hexane + chloroform
(6) n-heptane + chloroform
(7) di-n-propyl ether + carbon tetrachloride
(8) di-n-propyl ether + chloroform

Sysfern Contact types to be Contact types required from previous

Step niinrber solved steps
A 1 CH3-C2H4 -
2 CH-C1 -
B 3 and 4 CH3-C1 C2H4-Cl CH3- C2H4
C 5 and 6 CHS-CH C2H4-CH CH3-C2H4 CH3-Cl C2H4-Cl
CH-C1
D 7 0-c1 CH3-C2H4 CH3-C1 CH3-0 *
C2H4-C1 C2H,-0 *
E 8 0-CH CH3-C2H4 CH3-Cl CH3-CH
CH3-O * C2H4-Cl C2H4-CH
C2H4-O* 0-C1 CH-C1
* EG and Hg for these contacts were assumed to equal zero.

prediction for GF for n-octane +

carbon tetrachloride was fair (10 per cent) but
for t-butyl chloride and carbon tetrachloride with /3P’-dichlorodiethyl ether the
predicted GF’s were in error by 30 and 64 per cent respectively, confirming the
conclusion of earlier work that the molecules tend to acquire unique characters
with relatively small changes in structure. In terms of the original model the
assumption that EfjBis independent of A and B is, in the cases examined above,
highly suspect. Recko and Sadowska 5 4 have also examined the applicability of

Table 5 Optimum interchange free energies and interchange energies determined

by the first and zeroth approximations
Contact type First approximation Zeroth approxiniation
ij EuJJ mol-l HgJJ mol-1 EulJ mol-l HuJJ mol-l
CHS- C2H4 362.5 1010 339 883
CH-Cl 2412 245 5 4358 6282
CH,-Cl 244 444 22 1 378
C2H4-Cl - 4.0 200 14.9 242
CH3- CH 4072 777s 5076 7838
C2H4-CH 2033 2055 4539 7040
0- c1 - 454 - 1032 - 462.5 - 1081
CH3-O * 0 0 0 0
C2H4-0 * 0 0 0 0
0-CH - 3751 - 6887 - 10814 - 35263
* These parameters were assumed to equal zero.
O4 W. M. Recko and K. W. Sadowska, Bull. Acad. polon. Sci.,S b . Sci. chirn., 1972, 20, 73.
Specific Interactions in Non-electrolyte Mixtures 191
Barker’s equations to mixtures of carbon tetrachloride with n-hexane and
n-hexadecane but with temperature-independent Eij. Because of this and because
of a different prescription for the segmentation of the molecules their results
cannot be compared with those reported above.
De Bristowe and Stubley55used the lattice theory to interpret the properties
of mixtures of fluoroaromatics with hydrocarbons for which the models used by
Fenby and Scott were qualitatively correct but not amenable to quantitative
calculations. However, they were unable to achieve quantitative agreement
between the lattice theory and the experiment. The failure of the lattice theory
to account for differences between mixtures involving different isomers of the
same fluorinated benzenes serves to emphasize the failure of the assumption that
E t B is independent of the intramolecular environment.
Smith and Brown56have recently completed an extensive study of alcohol +
alkane mixtures in which Barker’s method is compared directly with ‘chemical’
association models. They conclude that the equations of Barker’s theory are
more satisfactory than those of Kretschmer and Wiebe 57 with concentration
equilibrium constants and those of Redlich and Kister 58 with mole fraction equili-
brium constants. However, even Barker’s equations are inadequate for predicting
accurate excess Gibbs energies and enthalpies. Sosnkowska-Kehiaian, Hryniewicz,
and Kehiaian K g accounted for their measurements of enthalpies of mixing HF of
alkanes and other hydrocarbons with n-alkyl ethers using a zeroth approximation
formula with one temperature-independent energy parameter.
In addition to correlations based directly on the lattice model there have been
a number of simplified schemes proposed, many of which can be thought of as
having their origins in lattice theory. Huggins 6o has recently published a number
of papers describing a treatment which has many features in common with the
lattice theory as derived by Barker. Rodger, Hsu, and Furter6l have used
Barker’s equations with temperature-independent energy parameters for
correlating liquid-vapour equilibria for a number of hydrocarbon + carbon
tetrachloride mixtures and for predicting isobaric liquid-vapour equilibria in
these systems. They also point out the possibility of using the equations simply
as empirical smoothing relations with the Qi determined by curve fitting rather
than by molecular geometry. However, they did not examine this possibility
having found that ‘the number of available parameters U(Eij) was more than
sufficient to meet the needs of their investigation’.
Recently the equations :
GE/RT = - Cjxi In(1 (33)
and
(34)
65 B. J. Skillerne de Bristowe and D. Stubley, J. Chem. Thermodynamics, 1974, 6, 581.
56 F. Smith and I. Brown, AustraZ. J. Chem., 1973, 26, 691, 705.
67 C. B. Kretschmer and R. Wiebe, J . Chem. Phys., 1954, 22, 1697.
58 0. Redlich and A. T. Kister, J. Chem. Phys., 1947, 15, 849.
69 K. Sosnkowska-Kehiaian, H. V. Kehiaian, and R. Hryniewicz, (a) Bull. Acad. polon. Sci.,
S&r.Sci. chim., 1969, 17, 185; (b)J. Chim.phys., 1971, 68, 922.
0Q M. L. Huggins, Polymer. J., 1973, 4, 511, and other papers referred to therein.
61 A. J. Rodger, C. C. Hsu. and W. F. Furter, Canad. J. Chem., 1969, 47,490.
192 Chemical Thermodynamics
which were first proposed by Wilson 62 have become popular for smoothing data
and for deriving parameters Ail, All from binary data which can be u s e d k the
prediction of properties of niulticomponent systems. Wilson’s equation for HE
can be derived from the lattice-theory equations (23) to (31) by a set of approxi-
mations which lead to intermediate relations which may themselves be useful
for smoothing or correlating data. Equation (30) may be written for two
components A and B with groups 1,2,3, etc.
x*(xt 4- 7 l z X e 4- xp + 712xF 4- ...) = Q ~ X A / ~ , (35)
xt(ql2xt 4- X t 4- 7)12xgf xr 4- ...) = Q ~ X A / ~ , (3 6 )
etc.
In the zeroth approximation (all q i ~= 1):

xf/x? = (Qfpxa)/(Qpxrd, (37)

and if this approximation is returned to equation (30) we obtain the set
(X f x ? c A x i q i j Q * ~ A ) / Q P ~ B= Qfpx~/2, (3 8)
from which for any particular pair Xf,XF we can obtain two equations
Q~QT~AxB/~(QAxA4- QBXB) = - (QAXA 4- QBXB)-lcACj+i’
Q$xA(~- vi.j)]XtXp, (39)
and
Q~@XAXB/~(QAXA -!- QBXB) = 11 - (QAXA f QBxB)-’EAzj+i’
@xA(1 - qj’i)]xfxp.
(40)
Similar sets of expressions can be obtained for all ij pairs. If it is assumed further
that all qi’j may be replaced by a generalized Ti, and all qj’i by a generalized qj,
then the excess enthalpy can be written
HE = XAXBCizj+ii[Bij/(l - AijXB) + Bu/(1 - Ajixa)], (41)
where
Bij = Q ~ Q ~ B H U T ~ (+
QA XA
QBXB), (42)
and
Ai.j = (1 - qi>h/(QaxA QBXB), (43)
Aji = (1 - .Ij)QA/(QAxA f ~BXB). (44)
Wilson’s equation corresponds to the special case of equation (41) with only
one ij pair. The extended form has apparently never been used. Wilson’s
equations are useful as smoothing equations and because the parameters derived
for binary mixtures can be applied directly to multicomponent mixtures. Their
chief defect is their inability to give CE large enough to lead to phase separation.
This defect is usually remedied when the equations are used near a liquid-liquid
region by including an arbitrary third parameter to give
G E / R T = - C[x,ln(l - A12xz) + x21n(l - AZ1x1)]. (45)
A good account of the uses of Wilson’s equations has been given by Prausnitz.6S
62 G . M. Wilson, J . Amer. Chem. SOC.,1964, 86, 127.
03 J. M. Prausnitz, ‘Molecular Thermodynamics of Fluid Phase Equilibria’, Prentice-Hall, New
Jersey, 1969, Chap. 6.
 Specific Interactions in Non-electrolyte Mixtures 193
The slightly more complex equations of Renon and Prausnitz 64 offer little
advantage over those of Wilson except near the two-liquid region. A much more
empirical approach to group-contribution correlations has been taken by
Pierrotti and others65 in a series of papers dating back to 1959. These corre-
lations, formulated for mixtures of homologues RIXl with R,X,, use equations
of the type:
RTln y; = A,, + B2nl/n2+ Cl/nl + D(nl - n2)2 + F2/n2, (46)
where nl, n2 are the numbers of carbon atoms in R,, R2; A12is a coefficient which
depends on x1 and x,; B, is a coefficient which depends only on x,; C, is a
coefficient which depends only on x,; D is a coefficient independent of x, and
x,; F2 is a coefficient which depends only on x,; 7; is the infinite-dilution activity
coefficient of component 1. Equation (46) can be simplified in a number of ways
for special combinations of homologues. For example, for mixtures of n-alkanes
with n-alkanes it reduces to
RTln yi = D(n, - n2),, (47)
which was originally proposed by Brsnsted and Koefoed.66
Perhaps the most clearly enunciated empirical group-contribution correlation
is that expounded by Wilson and Deal 67 and by Scheller.68 Both of these treat-
ments use a Flory-Huggins l1, l2 expression for the excess entropy to allow for
size differences and a group-interaction term which sums the contributions of
the various group interactions. Each term in this latter sum is dependent on a
specific group-interaction parameter and on the group concentration.
Models of this type lead to simple correlations of systems in which certain
sets of groups appear in different relative concentration but in what may be
regarded as generally similar environments. These models cope reasonably well
with the composition dependence of the properties of a given molecular pair
and with the properties of mixtures of different pairs of compounds containing
the same groups. Scheller was able to correlate the enthalpies of mixing of a
series of alcohol + water and alcohol + alkane mixtures in terms of the ratio
of CH, to OH groups.
Correlations of this kind which assert that properties of mixing depend only
on the ratio of numbers of groups are in effect restatements of Brsnsted and
Koefoed’s simple principle of congruence.66 This principle assumes that the
properties of a mixture of homologues are determined by the ‘index’ of the
mixture and that mixtures with the same index v (congruent mixtures) will have
the same properties. The index of a mixture is given by

13‘ H. Renon and J. M. Prausnitz, Amer. Inst. Chem. Engineers J., 1968, 14, 135.
a (a) G . J. Pierrotti, C. H. Deal, and E. L. Derr, Znd. and Eng. Chem., 1959, 51, 95; (6) 0.
Redlich, E. L. Derr, and G. J. Pierrotti, J. Amer. Chem. Suc., 1959, 81, 2283; (c) M. N.
Papadopoulos and E. L. Derr, J. Amer. Chem. Suc., 1959, 81, 2285; ( d ) C. H. Deal. E. L.
Den, and M. N. Papadopoulos, Znd. and Eng. Chem. (Fundamentals), 1962, 1, 17; (e) C . H.
Deal and E. L. Den, lnd. and Eng. Chem., 1968,60,28.
J . N . Brnnsted and J. Koefoed, Kgl. Danske Videnskab Selskab Mat.-Fys. Medd., 1946,22, 1.
G . M. Wilson and C. H. Deal, Znd. and Eng. Chem. (Fundamentals), 1962, 1, 20.
W. A. Scheller, Ind. and Eng. Chem. (Fundamentals), 1965, 4, 459.
194 Chemical Thermodynamics
where the 01 are the indices of the pure components. In Brarnsted and Koefoed’s
formulation for n-alkanes the index was taken as the number of carbon atoms,
The principle extends directly to homologous series of the type RX, where X is
some group or fixed set of groups.
r

I I I I I 1 I I I I I
0 0.5 1
(C16H34)
Figure 8 Enthalpies of mixing of n-CBH,, + n-C,,H,; A, 349.15 K; B, 358.15 K;
C, 374 K
In Brarnsted and Koefoed’s empirical formulation the alkane mixtures were
assumed to be regular with
GE = Axlxa, (49)
and this combined with the principle of congruence leads to
A = K(n, -n2)? (50)
The failure of this expression is illustrated by the changes in sign with composition
and temperature of the enthalpies of mixing of n-alkanes of widely different chain
lengths illustrated by Figure 8.69,70However, a specificationof the form of equation
to be used is not essential to the principle and various other equations have
been used. The principle itself can be justified in various ways71 and can be
arrived at as a consequence of the lattice model if one takes the zeroth approxi-
mation with number averages for both r and Qj in a mixture of RX,, and RXnZ.
Other empirical forms 73 have been used, and for volumes the expression
72p

V(n) = i-1Cai(n - 2)-f+a, (51)

leads to a very compact description of the volumes of mixing.
8p Th. Holleman, Ph.D. Thesis, Amsterdam, 1964.
J. A. Friend, J. A. Larkin, A. Maroudas, and M L. McGlashan, Nature, 1963, 198, 683.
71 H. C. Longuet-Higgins, Discuss. Furuduy Suc., 1953, 15, 207.
72 M. L. McGlashan, Mol. Phys., 1961,4, 87.
f3 J. Hijmans and Th.Holleman, Mol. Phys., 1961, 4, 91.
Specific Interactions in Non-electrolyte Mixtiires 195
An even more general approach to the principle is that devised by Hijmans 74
and illustrated in Figure 9. Measurements of H E or V E of two homologues of
widely separated indices are plotted us. index as shown. The property of any
mixtures of molecules of intermediate indices should then be obtainable as the
difference between the curve and the straight line joining the limiting indices for

the mixture. As shown in Figure 9 in which data for several mixtures are plotted,
the consistency of the principle between mixtures is excellent. The advantage of
the graphical procedure is that it can be used without any assumption about, or
knowledge of, the precise functional form of the HEor VE us. composition relation.
Even though the graphical correlation works well and is simple to use for
binary systems, attempts have continued to find suitable functional forms to
describe the properties of homologous mixtures in terms of composition and
chain length. McGlashan and co-workers 76 used the zeroth approximation
369

of the lattice theory 76 for molecules of different sizes to account for their measure-
ments of liquid vapour equilibria and enthalpies of mixing of n-alkanes. The
relations used were:

+ 41~2x,xzw(a1- az)2/(x141+ x,qz)W, (52)

74 J. Hijmans, Mol. Phys., 1958, 1, 307.
76 M. L. McGlashan and K. W. Morcom, Trans. Faraday SOC.,1961,57, 907.
E. A. Guggenheim, ‘Mixtures’, Oxford University Press, 1952, Chaps. 10, 11.
196 Chemical Thermodynamics
and
U E / R T = ~ l ~ , X l X 2 ~-( ~a2>2/(x141
1 + X292)kTY (53)
where 2q1, Zq2are the numbers of nearest-neighbour sites to molecules 1 and 2,
rl,r2 are the numbers of sites occupied by molecules 1 and 2, 2 is the co-
ordination number of the lattice, w is an interchange free energy and
u = w - Taw/aT. These were found to be better than the simple congruence
relations
GE = Ax,x2(nl - t ~ z ) ~ , (54)
and
HE = Bx,x2(n1- ngyy (55)
but are still not entirely satisfactory since they do not permit changes in the sign
of H E with composition.
Stoeckli, Fernandez-Garcia, and Boissonas 77 derived relations for the enthalpy
of mixing based on a van der Waals equation. Cruickshank et aZ.78have used
the lattice theory zeroth approximation with 2 = co and with different lattice
specifications leading to
1nyr = InC(1 - 55dx11 + (1 - G 4 2 l T P )
+ K(n1 + l"2 - n1>/(n1 + w 2 + UI2, (56)
and
Iny? = InK1 - 42)/x11 + (1 - r1421f-2)

+ K(n, + 2"2 - nl)/(nl + 2 b 2 + 2)12Y (57)

to correlate the infinite dilution activity coefficients of n-alkanes in n-octadecane,
determined from gas-liquid chromatography. Williamson 79 has proposed an
empirical modification of Hildebrand's regular-solution equation :14
UE = Wl55A~l - w, (58)
in which the solubility parameters of homologous series are expressed as functions
of chain length
UE = kxlxz(n1 + 2)(n2 +
2)[(n,/(n1 3- 2))4
- {n,/(na + 2)>fI2/(x1n1+ x2n2 + 2). (59)
Williamson and Scott have compared these various correlations and have
shown that all these treatments are a great improvement on the simple principle
of congruence. In particular they can predict both the drift in values of UF for
mixtures of various alkanes with a common second component and the sudden
changes in energies of mixing when both nA and n~ are changed but (na - n ~ )
is kept constant. However, they still do not predict the changes in sign of
H E us. x at higher temperatures for systems with very large chain-length
differences.

H. F. Stoeckli, J. G. Fernandez-Garcia, and Ch. G . Boissonas, Trans. Fcraday Soc., 1966,62,

3044.
A. J. B Cruickshank, B. W. Gainey, and C. L. Young, Trans. Faraday SOC.,1968, 64, 337.
7B A. G . Williamson, Ann. Rev. Phys. Chem., 1964, 15, 63.
A. G . Williamson. Trans. Furaday Soc., 1970,66, 335.
SpeciJic Interactions in Non-electrolyte Mixtures 197
Recently Looi, Mayhew, and Williamson *l have examined the extension of
the principle of congruence to ternary alkane mixtures. As shown in Figure 10,
the enthalpy of mixing of mixtures of alkanes of end-point indices 6 and 16 is
very nearly independent of whether or not the end-point index (n = 16) is that
of a pure substance (n-hexadecane) or of a mixture (n = 13 n = 19, +
x = 0.5).

c
120

d
0
E
7

n
Figure 10 Enthalpies of mixing of ternary mixtures of alkanes of indices 6 and 16;
a, experimental measurements for pure C,H14 with mixture of C13H28 C2aH40 +
(n = 16); - - -, Larkin et aE.;82- - -, McGlashan and M o r ~ o m-,
; ~ HE/J
~ mol-l =
~ ( -l ~)(440- 48.2(1 - 2 ~ ) )

A similar approach along even more general lines has been proposed and
examined by Woycicki who has taken the view that if the enthalpy of a system
depends only on the group composition then the enthalpy of mixing of the same
groups in the same proportions will be independent of their intramolecular
arrangements in the starting materials. Furthermore, if two sets of groups are
present in different ‘components’ in the same proportions then the enthalpy of
mixing of those components should be zero. A selection of Woycicki’s results
illustrating these points is shown in Table 6 for the formation of multicornponent
mixtures from equimolar binary mixtures. There is clearly some merit in group-
contribution methods. For simple molecules such as n-alkanes and other low-
polarity homologous series the assumption that the properties depend essentially
on the group composition seems to be a very good one. Indeed the quality of the
graphical 74 correlations and the success of the null experiments 84 suggest that
the group-contribution approach is better than the equations derived from any
C. K. Looi, C. J. Mayhew, and A. G . Williamson, J. Chem. Thermodynamics, 1974, 6, 1171.
82 J. A. Larkin, D. V. Fenby, T. S. Gilman, and R. L. Scott, J. Phys. Chem., 1966, 70, 1959.
s8 M. L. McGlashan and K. W. Morcom, Trans. Farudu-v SOC.,1961,57, 581.
s4 W. Woycicki, Proceedings of the 1 st International Conference on Calorimetry and Thermo-
dynamics, P.W.N., Warsaw, 1971, p. 797.
198 Chemical Thermodynamics
Table 6
Component 1 Component 2 “xl” HE/J mol-*
+
cyclohexane toluene +
methylcyclohexane benzene 0.51 1 1.2
methylcyclohexane + ethyl- ethylcyclohexane + toluene 0.497 0.2
benzene
+
cyclohexane 4-methyl- methylcyclohexane + pyridine 0.502 - 0.2
pyridine
+
cyclohexane ethylcyclo- benzene + ethylbenzene 0.500 599
hexane
methylcyclohexane toluene 0.50 52 1

specific model would indicate. For more strongly polar interactions such as
those involved in the ether and alcohol systems it appears that the group-
contribution approach has limitations resulting from the fact that the groups in
such molecules rapidly acquire individuality which depends on their intra-
molecular environment. However, it has not yet been shown conclusively that
the failures to account successfully for such systems result from failure of the
principle of group contributions or merely from the inadequacies of the equations
used to represent these group contributions.
 7
Volumetric Properties of Gaseous Mixtures
BY C. M. KNOBLER

1 Introduction
The inclusion in this volume of a chapter on the properties of gaseous mixtures
may appear inappropriate; there are, however, close connections between the
properties of liquid and gaseous mixtures. An obvious but nonetheless important
point of contact is the study of vapour-liquid equilibrium. The thermodynamic
analysis of vapour-pressure measurements requires a knowledge of the chemical
potentials of the components of the gas mixture in equilibrium with the liquid.
Even in experiments in which the properties of the vapour are not directly
involved, the volumetric properties of a vapour phase must often be known in
order to determine with precision the composition of the liquid mixture.
Studies of the equations of state of pure substances have played a significant
role in the development of our knowledge of intermolecular forces. Similar
measurements on gas mixtures can provide important information concerning
interactions between unlike molecules, the interactions that must be specified in
all theories of liquid mixtures.
Volume 1 contains a chapter by Cox and Lawrenson on the p , V, T behaviour
of single gases,l and the relatively recent monograph by Mason and Spurling2
provides a detailed discussion of the volumetric properties of pure gases and
mixtures. A thorough discussion of the thermodynamics of gases has been
given by Beattie and St~ckmayer.~ Some of the material presented here will be
found in these earlier reviews, to which the reader is referred for details.

2 The Virial Equation of State

The volumetric properties of a gas at low and moderate densities are best
described by the virial equation of state
pY,/RT = 1 + B(T)/Y, + C ( T ) / V L + D(T)/V% 3- ..., (1)
where p represents the pressure, Vm the molar volume, T the temperature, and R
the gas constant. The temperature-dependent coefficients B, C, D, ..., are known
respectively as the second, third, fourth, ...,virial coefficients. The series diverges
at densities approaching that of the liquid, and in practical applications equation
(1) is rarely extended beyond the term in V G ~In . the low-density limit (Vm -+ 03)
J. D. Cox and 1. J. Lawrenson, ‘Chemical Thermodynamics’, ed. M. L. McGlashan (Specialist
Periodical Reports), 1973, Vol. 1, p. 162.
E. A. Mason and T. H. Spurling, ‘The Virial Equation of State’, Pergamon, New York, 1969.
J. A. Beattie and W. H. Stockmayer, ‘A Treatise on Physical Chemistry’, eds. H. S. Taylor
and S. Glasstone, Van Nostrand, New York, 1951, 3rd edn., Vol. 11.
199
200 Chemical Thermodynamics
equation (1) clearly reduces to the perfect-gas equation. It can be shown
rigorously that the second virial coefficient term represents the contribution to
gas imperfection arising from the interaction between pairs of molecules. The
third virial coefficient term stems from the interactions of triplets and the
succeeding terms represent the interaction between groups of four, five, etc.,
molecules at a time.
When the interactions between molecules have no angular dependence, the
relation between the virial coefficients and the intermolecular potential u(ru) is
very simple in form

1d
B(T) = ( N A P Y ) [iexp[- u(r12)/k7l - 11 dr1 dr2
*

= - 2 z N .~Cm{exp[-
o u(r)/kT] - l}r2dr, (2)

x (exp[- u(r2,)/kT] - 1) dr, dr2dr,. (3)

For a gas mixture in which the composition is given by the mole fractions Xi,
the virial coefficients, B(x), C(x), are constructed from averages of all possible
interactions between groups of molecules

Three second virial coefficients are therefore needed to describe the binary
interactions in a two-component mixture: BI1 and BZ2,the virial coefficients of
the pure components, and B12,the interaction virial coefficient, which depends
on the intermolecular potential between molecules 1 and 2. The third virial
coefficient of the mixture is composed of four terms: Clll, C222,ClI2,and ClZ2.
It is convenient to define an excess second virial coefficient g, and two excess
third virial coefficients Sland S2
8 = Bl2 - *(&l + &,I,
91 = c11 2 - W C l l l + Cz22),

F 2 = c 1 2 2 - * G l l + 2C222).

With these definitions, equations (4) and ( 5 ) become for a binary mixture
B(x) = XlBll + x& + 2x1x28, (8)
C(X) = Xlclll + X2C222 + 3X;XFl + 3X&F2. (9)
Although the virial expansion in density, equation (l), is the equation of state
directly relatable to theory, an expansion in terms of pressure sometimes proves
useful
pv, = RT +
B’(T)p +
C’(T)p2 ....+ (10)
 Volumetric Properties of Gaseous Mixtures 201
The pressure and density virial coefficients are related by
B’ = B, (1W
C’ = (C - B2)/RT. ( m
Temperature Dependence and Magnitude of Virial Coefficients.-A useful way to
gain insight into the behaviour of virial coefficients is to evaluate B and C for
the square-well potential
u(r) = m, 0 < r < a,
u(r) = - E, a < r < Aa, (12)
u(r) = 0, r > Aa.
When this potential is substituted into equations (2) and (3) we obtain
- (R3 - 1)A],
B = bofl (13a)
C = bi[a, - a2A -I- asA2 - a4As], (13b)
where A = exp (&/kT)- 1, bo = (2/3)xN*a3 and the coefficients ai are functions
of A only. Often A is chosen to be 1.5. At low temperature A -+ 00 and both
B and C are large in magnitude and negative. At high temperature A -+ 0 and,

1 I i I

-1

-2

-3

-4

-5
0.5 1 2 5 10 20 50 100
T*
Figure 1 Reduced second and third virial Coefficients vs. reduced temperature for the
Lennard-Jones 6-1 2 potential

for this potential, the virial coefficients approach a constant positive value. As a
result of the change in the relative magnitudes of the alternating positive and
negative terms in equation (13b), C passes through a maximum.
From equations (13a) and (13b) we see that it is possible to define reduced
virial coefficients B* = B/b, and C* = C/&,which, for a given A, are functions
202 Chemical Thermodynamics
only of a reduced temperature T* = kT/&. This principle of corresponding
states for virial coefficients is of wide applicability and will be discussed in a
later section.
Figure 1 shows reduced virial coefficients calculated for a more realistic two-
parameter potential, the Lennard-Jones 6-12 potential. The only significant
difference between the Lennard-Jones and square-well second virial coefficients
occurs at high reduced temperatures, where experimentally B is observed to have
a maximum value. This maximum is related to the softness of the repulsive
portion of the potential; it therefore cannot be reproduced by the infinitely steep
square well. The third virial coefficients calculated from the square well are
qualitatively correct but show significant departures from the Lennard-Jones
values.
The reduced temperatures that correspond roughly to the melting, boiling,
and critical points are 0.52, 0.87, and 1.29, respectively. At these temperatures
the reduced second virial coefficients are - 8.1, - 3.2, and - 1.6. Typical values
of b, are similar to molar volumes of liquids, hence for simple substances the
second virial coefficient varies between - 1000 and - 200 cm3mol-1 from the
melting point to the critical point. The Boyle point, the temperature at which
B = 0, lies at T* z 3.4, or roughly four times the boiling point.
Temperature Dependence of &.-An approximate formula for B can be obtained
by substituting the square-well formula for B [equation (13a)I into equation (6)
?
!
f = bo(12) - *[b,(ll) + b,(22)] - (R3 - 1)
x CbO(12)AlZ - HbO(1 1 )A11 + b0(22)A22)1.
The exponentials in equation (14) can be expanded to give
8 = bo(12) - *[b,(ll) + b,(22)] - (R3 - 1)

- b,(22)&
+ 2b,( 12)& - b,( 1 1)&
2(kT)2

At high temperatures d approaches the value Ab = b,(12) - &[b,(ll) +

b,(22)], which, for a hard spherical core potential, is negative. At intermediate
temperatures 8 is roughly linear in 1/T. As we will see (Section 4), often to a
good approximation c12can be replaced by ( E ~ ~ E ~ ~In) * this
. case, for molecules
that do not differ much in size the temperature-dependent contribution to d is
positive. This is the usual case. Some typical values of &' for simple systems are
shown plotted against 1/T in Figure 2. A detailed discussion of the relative
magnitudes of B12,Bll, and BZ2has been given by Khoury and R o b i n ~ o n . ~
Thermodynamic Properties of Gas Mixtures.-The residual thermodynamic
properties of a gas, that is the differences between the measured thermodynamic
F. Khoury and D. B. Robinson, J. Chem. P h p , 1971,55, 2071.
 Volumetric Properties of Gaseous Mixtures 203
properties and the perfect gas values, can be expressed in terms of the equation of
~ t a t e .The
~ residual enthalpy is given by

H,R = fop[& - m&/aT>,ldP (1 6a)

= B’ + (C-B2)p2/2RT + ...
+ T((dB/dT)P + [dC/dT - 2B(dB/dT)]p2/2RT + ...}. (16b)

200

150

c
‘L.

E
m
fi 100
\

50

0
3.0 4.0 5.0 6.0

lo3 r-W
Figure 2 Typical values of &m3 mol-l vs. lo3 T-l/K-l. Data for CH, + C,Hlo
and CH, + C3H8from Dantzler et al. ;sl Ar + He from Brewer and Vaughn l1

As in the case of liquids, it is possible to define excess thermodynamic functions

for gas mixtures. The enthalpy of mixing of two gases at constant temperature
and pressure,
HE, = HE(x) - x,H2(1) - XzH,R(2), (1 7)
can be obtained in the form of a pressure series by using (16b) and the definitions
(4), (519 (8), and (9)
HZ = 2 ~ 1 ~ 2 (-Bdg/dT)p + + Td/3/dT)p2/2RT+ ...,
X I X ~ ( / ~ (18)
Thus, at low pressures the enthalpy of mixing is simply related to B and dB/dT.
The higher-order terms, however, are rather involved functions of the virial
coefficients of the pure components and the excess virial coefficients.
From the relation between H Z and the excess Gibbs free energy and excess
entropy, GE and Sg,
H E = GE T Sg,+ (20)
it is clear that Sg is given by the terms in (18) that are multiplied by Tand G g
is given by the remaining terms.

3 Experimental Methods for Determining Virial Coefficients

A wide variety of experimental methods have been used to determine the virial
coefficients of pure substances, and they are discussed in detail in previous
reviews.'* These techniques can also be applied to mixtures, and interaction
virial coefficients can be obtained from equations (4) and ( 5 ) if the virial
coefficients of the pure components are known. To determine a second virial
coefficient to 5 1 cms mol-1 from volumetric studies at room temperature
requires that measurements be made to a precision of a few parts in lo5. Under
ideal conditions, this standard of precision on B(x), BI1, and Bzz leads to an
uncertainty more than twice as large in Biz.
High precision is especially hard to obtain in measurements on vapours.
Pressure ranges in expansions are restricted by proximity to saturation conditions,
and high-accuracy measurements of low pressures are difficult. Adsorption also
complicates studies on vapours. For example, in measurements by the Burnett
method on pure Ar and Kr near saturation, Weir et aL5 determined the effect
of adsorption in their copper apparatus by performing experiments in vessels of
different surface-to-volume ratios. Corrections for adsorption first became
significant at temperatures at which the saturation pressure was just in excess of
atmospheric, and rose rapidly at lower temperatures. For Ar at its triple point,
where the saturation pressure is 70 kPa and B = - 280 cm3mol-l, the correction
due to adsorption was 8 cm3mol-I. Hall and Eubank recommend that Burnett
measurements be combined with isochoric data to avoid systematic errors due
to adsorption.
Measurements of adsorption isotherms of organic vapours on clean glass
surfaces have been performed by Bottomley and co-worker~.~. * They find that,
in general, monolayer coverage (based on liquid densities) is reached at approxi-
mately 75 per cent of the saturation pressure. For typical pV-apparatus they
calculate that adsorption can produce errors in B as large as 30 to 100 cm3mo1-1
when the adsorption occurs on a clean surface. Minute traces of grease
ti R. D. Weir, I. Wynn Jones, J. S. Rowlinson, and G. Saville, Trans. Furaduy Suc., 1967, 63,
1320.
K. R. Hall and P. T. Eubank, Physicu, 1972, 61, 346.
' G. A. Bottomley and C . G. Reeves, Trans. Furuduy SOC.,1957, 53, 1455.
G. A. Bottomley, I. H. Coopes, G . L. Nyberg, and T. H. Spurling, Austral. J. Chem., 1966,
18, 1105.
Volumetric Properties of Gaseous Mixtures 205
introduced into the apparatus by repeatedly passing mercury through a greased
stopcock raise the adsorption many fold.
Adsorption can be enhanced in vapour mixtures. Pavlychenko reports that
the adsorption of mixtures of acetone and chloroform on graphitized carbon
black is substantially greater than that of either of the pure components.
Absorption of organic vapours can also be significant in elastomers. O-rings
sample
inlet
manometert I

Figure 3 Schematic diagram of apparatus for the determination of d from the pressure
change on mixing gases at constant volume
(Reproduced by permission from Rev. Sci. Instr., 1967, 38, 184)

must therefore be excluded from those parts of the apparatus in which pressure
changes cannot be tolerated.
In apparatus in which mercury can flow rapidly in a dry atmosphere, another
complication can arise: static electricity can build up on glass surfaces until a
discharge occurs. (Such a phenomenon is frequently observed in glass mercury
diffusion pumps.) Vapours with low ionization energies can be made reactive
by the discharge and a significant contamination of the sample can occur. In
addition, substances thought to be inert are sometimes found to react with parts
of apparatus. A notorious example is CCI4, which apparently reacts slowly
with rner~ury.~
Gas Mixing Methods.-Since the properties of a gas mixture depend on the like
as well as the unlike interactions, no pV-technique has been devised by which
interaction virial coefficients can be determined directly. On the other hand, at
low pressures the change in pressure on mixing gases at constant volume and the
change in volume on mixing gases at constant pressure are directly proportional
to 8.
The pressure-change technique was utilized by Beenakker and co-workers to
study gas mixtures1° at 9 0 K ; a virtually identical apparatus was utilized by
Brewer and Vaughn l1 at temperatures from 148 to 323 K. An adaptation of
this apparatus more suited to vapour mixtures l2 is shown in Figure 3.
N. M. Pavlychenko, Russ. J. Phys. Chem., 1970, 44, 152.
lo C. M. Knobler, J. J. M. Beenakker, and H. F. P. Knaap, Physica, 1959, 25,909.
l1 J. Brewer and G. W. Vaughn, J. Chem. Phys., 1969,50,960.
la C. M. Knobler, Rev. Sci. Instr., 1967, 38, 184.
206 ChemicaZ Thermodynamics
Three 2dm3 glass bulbs are immersed in a large thermostatic bath; bulbs I1
and I11 have cold fingers sealed into them. The bulbs are interconnected by
metal lines and can be isolated from each other and the sample-filling system
by metal bellows valves A, A’, B, B’, C , D, and E. T is a sensitive differential
pressure transducer.
With valves A’, C , and D closed, bulbs I and I1 are filled through valve A
with one of the substances to be studied. The filling pressure is measured on an
external manometer. Valve A is closed and bulb I11 is now filled through valve
A’ with the other gas to exactly the same pressure, as can be observed on the
pressure transducer. Valves A’, B’, and E are now closed and valve C is opened,
connecting bulbs I and 11.
The two gases are mixed by adding liquid nitrogen to one of the cold fingers
for about 15 min and then closing valve C. After the system has returned to the
bath temperature, valves B’ and C are reopened and the pressure difference is
observed. The mixing process is repeated until a constant pressure difference is
obtained.
The relation of Ap the pressure change on mixing, p the filling pressure, and
T the temperature to the excess second virial coefficient is given by

from which d can be determined by successive approximations. The apparatus

is normally operated under conditions for which the second term in (21) is only
a few per cent of the first, and higher-order terms are negligible. Values of Ap
are typically 0.05 to 0.5 kPa at filling pressures below 100 kPa.
This technique offers several advantages over direct measurements of the
second virial coefficient. There is no need to determine volumes with high
accuracy, temperature stability is not critical because of the differential nature
of the pressure measurement - only uniformity of temperature is required - and
there are no stringent requirements on the measurement of the filling pressure.
Although there is only a small change in the total pressure when the gases are
mixed, errors caused by adsorption must still be considered because the partial
pressure of each component is halved. The apparatus can easily be constructed
with a large volume-to-surface ratio, however, and this minimizes adsorption
effects.
Although in favourable cases B can be determined to 1 to 2 cm3mol-l, the
uncertainty in B12depends on the precision with which the virial coefficients of
the pure components are known. On the other hand, for some applications, such
as the calculation of enthalpies of mixing or the correction of vapour-liquid
equilibria data, d is the quantity of interest. Pressure-change measurements are
not likely to be of utility for obtaining information about interaction third virial
coefficients because the contribution of the second virial coefficients to the
 Volumetric Properties of Gaseous Mixtures 207
p / R T term in equation (21) is usually of the same order of magnitude as that of
Fland S2.
Edwards and Roseveare13 originated the technique in which the volume
change on mixing gases at constant pressure is measured. An improved version
of their apparatus was developed by Gorski and Miller.14 The operation of the
apparatus is similar to that of the pressure-changedevice. Three bulbs are required;
one serves as a pressure reference and the others contain the gases to be mixed.
At the start of the experiment the pressures in the bulbs are equal. When the
gases are mixed there is a change in pressure which can be observed on a sensitive
pressure transducer. The mixing was accomplished in the Gorski-Miller
apparatus by means of a pump, but the simpler mixing method using a cold
finger could also be used.
The volume change is measured by adding or withdrawing mercury from a
burette that is part of the system until the transducer returns to the original null.
For bulbs of equal volume the relation between Q and AV,, the molar change,
is
Q = 2AJ& - [(&I - B2a)2/2 - Q(Bi1 +
&)
+ +
3(p1 S2)/4 - g2/2]p/RT ..., + (22)
which differs from (21) by the term - b2p/2RT because the mixing occurs at
constant pressure rather than at constant volume.
Recently Hall and Eubank l6$ l6 proposed that B12be determined in a Burnett
apparatus by mixing at constant volume two gases not necessarily initially at
the same pressure. The interaction virial coefficient is then determined from the
initial and final pressures, the compression factors of the pure substances, and
the temperatures. It is claimed that the virial coefficients of the pure substances
are not required; to specify the compression factors, however, much the same
information is needed. The difference between the Hall and Eubank method and
other gas-mixing techniques lies in the treatment of the results; for a simulated
experiment the proposed analysis leads to a smaller uncertainty in B12.
Solubility in Compressed Gases.-When an inert gas (1) is added to a system in
which a liquid or solid (2) is in equilibrium with its vapour, the vapour pressure
of the condensed substance changes. In other words, if the total pressure in the
system is p and the mole fraction of (2) in the vapour phase is y,, the partial
pressure p y 2 is not equal top:, the vapour pressure of the pure substance. If the
inert gas is not appreciably soluble in the solid or liquid, the change in vapour
pressure is attributable to two effects. (i) The chemical potential of the solid or
liquid is increased because of the rise in the applied pressure, the so-called
Poynting effect.17 (ii) The chemical potential of the vapour is changed because
of interactions between the vapour and the inert gas.
Thermodynamic analysis of these effects leads to the relation

la A. E. Edwards and W. E. Roseveare, J. Amer. Chem. SOC.,1942,64,2816.

l4 R. A. Gorski and J. G. Miller, J. Amer. Chem. Soc., 1953, 75, 550.
l5 K. R. Hall and P. T. Eubank, J. Chem. Phys., 1973, 59, 6691.
l6 K. R. Hall and P. T. Eubank, J. Chem. Phys., 1973, 59, 709.
J. S. Rowlinson and M. J. Richardson, Ado. Chern. Phys., 1959,2, 85.
208 Chemical Thermodynamics

When the condensed phase is assumed incompressible, the virial equation is

used to evaluate the integrands, and the integrations are carried out, a relation
involving a power series in the applied pressure can be obtained
RTln(PY2lP;) = V!XP - P t ) + B2zPZ
- 12Y;LBlZ + 3 1 - Y W 2 2 - Yf&lP + .-.- (24)
Under the conditionsp % p t , y 2 < y1 equation (24) reduces to
R T In (py2/p3 = [ V,* - 2BlZ - BJp .... + (25)
Hence, it is possible from measurements of y z as a function of p to obtain
values of Blz if the molar volume of the pure condensed phase and the equation
of state of the inert gas are known.
Equation (24) can be written in a variety of equivalent ways. One form
frequently used as a working equation is
RTWYzlPZ) = GYP - P t ) + B2ZPZ
- 2RT(Y142 - YzB22~/Y,(u)
+ RTlnrPKl(Y)/RTl. (26)
The molar volume of this mixture, Vm(y), is obtained by successive approxi-
mations. Higher-order terms can be analysed to obtain information about
C11z,19and equations can be modified to include the effect of solubility of the
inert gas in the condensed phase.
Gas solubility has sometimes been determined in static systems, with the
concentration of the solute in the gas phase measured spectrophotometrically.20$ 21
More commonly, a flow technique is employed. A schematic diagram of a flow
apparatus 22 for determining virial coefficients by the gas-solubility method is
shown in Figure 4. The inert gas flows under high pressure from a storage
cylinder A through a pressure regulator B. A Bourdon gauge C is used to read
the pressure. The gas is bubbled through a fritted disc D into the bottom of a
thermostatted pressure cell E containing the liquid phase. After leaving the cell
the gas mixture passes through a heated section of tubing F and is expanded
through a heated needle valve G into a series of three cold traps H. The section
of tubing and valve are heated to prevent condensation occurring before the
mixture reaches the cold traps. A wet-test meter I at the outlet is used to measure
the volume of inert gas that has flowed through the system during a measurement.
The quantity of liquid phase trapped is determined gravimetrically. Measure-
ments are performed at a variety of flow rates to ensure that equilibrium has been
reached.
R. C. Miller, A. J. Kidnay, and M. J. Hiza, J . Chem. Thermodynamics, 1972, 4, 807.
Is F. H. Kate, jun. and R. L. Robinson, jun., J. Chem. Thermodynamics, 1973, 5, 259.
2o G. C. Najour and A. D. King, jun., J . Chem. Phvs., 1966, 45, 1915.
21 A. D. King, jun., J. Chem. Phys., 1968, 49, 4083.
22 C. R. Coan and A. D. King, jun., J. Chromatog., 1969, 44,429.
Volumetric Properties of Gaseous Mixtures 209
In studies of the solubility of solids a closed-loop gas-recirculation system
has been used. The gas passes over solid deposited in trays and gas
chromatography is used to determine its composition,
The Chromatographic Method.-The chromatographic method for the deter-
mination of interaction virial coefficients is related to the gas-solubility

Figure 4 Schematic diagram of apparatus for determining the soIubiIity of a liquid in a

compressed gas [adapted from Coan and King (ref. 22)]

technique.* The thermodynamic description of the equilibrium distribution of

a solute between a stationary phase and the carrier gas in a chromatographic
column leads to a power series in the pressure in which the coefficients depend
on the virial coefficients describing the interaction between the solute and the
carrier gas. As a result, studies of chromatographic separations as a function of
column pressure can provide equation of state data. The most detailed investi-
gation of this technique has been by Cruickshank and c o - w o r k e r ~ . ~They
~-~~
express their results in an equation for the net retention volume VN,the volume
of gas that has flowed through the column between sample injection and the
emergence of a peak maximum
InVN = In V & + PpoJ4, + & ~ i ( J 3 ~ , (27)
where
P = w 1 2 - J,%(Ul/m (28)
f = (3c122 - 4Bl,B2,)/2(RT)2. (29)
The subscripts 1 and 2 refer respectively to the carrier gas and the solute.
Vg(1) is the partial molar volume of the solute in the stationary phase at infinite
dilution. The quantity J i is defined by
J f = 3[(pi/po)* - 1]/4[(pi/~o)~
- 11, (30)
23 A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Proc. Roy. Soc., 1966, A , 295,259.
24 A. J. B. Cruickshank, M. L. Windsor, and C. L. Young, Proc. Roy. Soc. 1966, A , 295, 271.
25 A. J. B. Cruickshank, B. W. Gainey, C. P. Hicks, T. M. Letcher, R. W. Moody, and C. L.

Young, Trans. Faraday Soc., 1969, 65, 1014.

* A fuller account of the chromatographic method with the emphasis on its uses to determine
activity coefficients is given in the following Chapter.
210 Chemical Thermodynamics
and arises from a correction for the pressure drop in the column from the inlet
pressure pi to the outlet pressure po. The zero-pressure value of equation (27),
In V k , depends on the molar volume V: and vapour pressure pT of the pure
solute, the infinite-dilution value/? of the activity coefficient of the solute in the
solvent, and the amount of substance nl of solvent in the column

An essential assumption in the derivation of equation (27) is that the carrier is

insoluble in the stationary phase.
The experiments utilize a chromatograph designed to operate at pressures up
to about 2500kPa. This requires that special provision be made for sample
injection and 26 The choice of stationary phase is dictated primarily

by the requirement for insolubility of the carrier gas, but it is also necessary that
the stationary phase be a good solvent for the solute in order that adsorption of
solute on the column packing be minimized.25
Retention volumes are determined at a variety of pressures and the term
B1, - Vg(1) is determined by least-squares analysis of the 1nVN us. p curve.
It is important in such an analysis to examine the effect of the p 2 term on the
error in the coefficient of the linear term.27 A check on the self-consistency of
the method is the requirement that the intercept In V g be independent of the
carrier gas. 28
Under optimal conditions values of B,, can be determined by the chromato-
graphic method with a precision of about k 2 cm3mol-l, but typical uncertainties
are about 10 to 20 cm3mol-l. Virial coefficients determined by chromatography
have been shown to agree within experimental error with values obtained from
gas-rni~ing,~~ gas-solubility,22and p , V, T-methods3* When the solubility of the
, ~ ~precision of the measure-
carrier gas in the stationary phase is a p p r e ~ i a b l ethe
ments decreases and the agreement with other results is poorer.
Calorimetric Measurements.-The relation between the equation of state and the
isothermal Joule-Thomson coefficient p = ( S H / ~ ~ can
P ) ~be obtained from
equation (16b). In the low pressure limit it is simply
p = B - T(dB/dT). (32)
An integration constant must be evaluated in order to calculate B directly from
measurements of p ; usually a functional form for B ( T ) or an intermolecular
potential function is specified.
A ffow calorimeter for measurement of the isothermal Joule-Thomson
coefficient of vapours has been described by Francis, McGlashan, and Wormald 32
and an apparatus said to be useful for studies on mixtures has been described by
, ~ ~very few measurements of p for mixtures have been
Dawe and S n ~ w d o n but
26 B. K. Kaul, A. P. Kudchadker, and D . Devaprabhakara, J.C.S. Faraday I , 1973, 69, 1821.
R. L. Scott and R. D. Dunlap, J. Phys. Chem., 1962, 66, 639.
** M. Vidergauz and V. Semkin, J. Chromafog., 1971, 58, 95.
29 E. M. Dantzler, C. M. Knobler, and M. L. Windsor, J . Chromatog., 1968, 32, 433.
30 D . H. Everett, B. W. Gainey, and C. L. Young, Trans. Furuday SOC.,1968, 64, 2667.
31 R. L. Pecsok and M. L. Windsor, Anulyt. Chenr., 1968,40, 1238.
s2 P. G. Francis, M. L. McGlashan, and C. 3. Wormald, J. Chewt. Thermodynamics, 1969, 1,441.
33 R. A. Dawe and P. N. Snowdon, J . Chem. Thermodynamics, 1974,6,65.
Volumetric Properties of Gaseous Mixtures 21 1
reported. Francis and Luckhurst 34 analysed the isenthalpic data of Sage,
Lacey, and co-workers on mixtures of methane with ethane,35n - b ~ t a n e and ,~~
propane,37 and of Charnley et aL3* on mixtures of nitrous oxide with nitrogen
and ethylene in terms of the empirical McGlashan-Potter equation.39 The
accuracy of the data is too low to provide an adequate test of the equation.
Isenthalpic Joule-Thomson measurements on N2 + CHI + C2H, mixtures
have been reported by Ahlert and WenzeL40 They compared their results with
the predictions of the virial equation of state with virial coefficients calculated
from the Lennard-Jones 6-12 potential.
A disadvantage of Joule-Thomson measurements is that they give information
on B(x) and not BI2. Williamson41 has suggested that measurements of the
enthalpy of mixing of vapours (Section 10) at low pressures are a distinctly
better source of data on interaction virial coefficients. It can be seen from
equation (18) that such studies give Q - Tda/dT directly. As in the case of
Joule-Thomson data, some functional form for d must be used to allow
integration of the data. To date no measurements on vapour mixtures have
been reported.

4 Calculation of Interaction Virial Coefficients

A number of empirical rules have been proposed for calculating an interaction
virial coefficient from the virial coefficients of pure components at the same
temperature. None is very satisfactory, but the best of the combinations appear
to be12
Bij = (BiiBjj)1’2, (33)
and
Cijk = (ciiicjjj Ck&)”’. (34)
More reliable procedures are based on the principle of corresponding states.
As noted in the earlier section, it is possible to define reduced virial coefficients
that are universal functions of the reduced temperature. Pitzer I 3 has shown that
this corresponding states principle is strictly valid for substances that interact
with a potential of the form
u(r) = ef(r/a), (3 5 )
where E and u are constants and f(r/u) is a universal function. The interactions
must be classical and the translational degrees of freedom must be separable
from the internal degrees of freedom.

34 P. G. Francis and G. R. Luckhurst, Trans. Faraday SOC.,1963, 59, 667.

35 R. A. Budenholzer, B. H. Sage, and W. N. Lacey, Ind. and Eng. Chem., 1939,31, 1288.
36 R. A. Budenholzer, B. H. Sage, and W. N. Lacey, Ind. and Eng. Chem., 1940,32, 384.
s7 R. A. Budenholzer, D. F. Botkin, B. H. Sage, and W. N. Lacey, Ind. and Eng. Chem., 1942,
34, 878.
3s A. Charnley, J. S. Rowlinson, J. R. Sutton, and J. R. Townley, Proc. Roy. SOC.,1955, A ,
230, 354.
sn M. L. McGlashan and D. J. B. Potter, Proc. Roy. Soc., 1962, A , 267, 478.
40 R. C. Ahlert and L. A. Wenzel, Amer. Inst. Chem. Engineers J., 1969, 15, 256.
41 A. G. Williamson, personal communication.
42 Ref. 2, p. 257.
43 K. S. Pitzer, J. Chem. Phys., 1939, 7 , 583.
212 Chemical Thermodynamics
The most severe of these restrictions is that on the form of the potential.
Only a few molecules might be expected to interact through centrally symmetric
forces - the rare gases and possibly small nearly spherical molecules such as
H,, N2, 02,and CH,. Numerous tests of the principle of corresponding states
have been performed for these substance^.^^ Among the most recent is the study
by Dymond45 who showed that second virial coefficients and dilute-gas
transport coefficients for the inert gases neon to xenon could be interpreted
satisfactorily on the basis of a universal reduced pair potential. Kestin and
co-workers 46-50 have made careful systematic studies of the viscosities of simple
gases. They find50*51 that the principle of corresponding states applies to the
second virial coefficients, the diffusion coefficients, and the viscosities of the rare
gases and their mixtures. (Helium deviates at low temperatures because of
quantum effects.) The transport properties of simple polyatomic molecules can
also be correlated by corresponding states, but the virial coefficients cannot.
Although Kestin’s analysis is carried out in terms of empirical universal reduced
functions, the 11-6-8 potential of Hanley and Klein 5 2

4 = E [ A ~ ( ( T /-
T )A
~~, ( o / ~ )-~ A,(o/r)*],

also provides a good representation of the data.50

The results just cited seem to demonstrate that the potentials of the rare
gases have the same forms, yet crossed molecular beam studies 53 show that the
potentials are not strictly conformal. To put these results into perspective it is
useful to make a distinction between ‘utility’ and ‘validity’ employed by Hanley
and Klein54in a discussion of the m-6-24 potential. They define utility as
‘the ability to correlate particular kinds of data’ and validity as ‘the ability . . .
to fit all possible kinds of data and, at the same time, to be compatible with
independent fundamental ideas on the nature of the forces between molecules . .’. .
The principle of corresponding states has very great utility even if its validity
may in some sense be challengeable.
For substances that follow simple corresponding states, the interaction virial
coefficients can be calculated at any temperature once the potential parameters
E~~ and o12have been specified. Obviously it would be advantageous to be able
to calculate the parameters for the unlike interaction from the potential parameters
for the pure components, although there is no fundamental reason why this
should be possible.
44 J. 0. Hirschfelder, C. F. Curtis, and R. B. Bird, ‘Molecular Theory o f Gases and Liquids’,
Wiley, New York, 1964.
45 J. H. Dynond, J. Chem. Phys., 1971,54, 3675.
J. Kestin, W. Wakeham, and K. Watanabe, J . Chem. Phys., 1970, 53, 3113.
47 J. Kestin, S. T. Ro, and W. A. Wakeham, J . Chent. Phys., 1972, 56, 4036.
J. Kestin, S. T. Ro, and W. A. Wakeham, J. Chem. Phys., 1972, 56, 4086,
49 J. Kestin, S. T. Ro, and W. A. Wakeham, J . Chem. Phys., 1972,56, 5837.
6o J. Kestin, S. T. Ro, and W. A. Wakeham, Physica, 1972, 58, 165.
61 J. Kestin and E. A . Mason, ‘Transport Phenomena - 1973’, American Institute o f Physics
conference Proceedings, ed. J. Kestin, New York, 1973, p. 137.
62 M. Klein and H. J. M. Hanley, J . Chem. Phys., 1970, 53, 4722.
6s J. M. Farrar. T. P. Schafer, and Y. T. Lee, ‘Transport Phenomena - 1973’, American Institute
of Physics Conference Proceedings, ed. J. Kestin, New York, 1973, p. 279.
H. J. M. Hanley and M. Klein, J . Phys. Chem., 1972, 76, 1743.
Volumetric Properties of Gaseous Mixtures 213
The London dispersion interaction between two spherical atoms can be written
as a series 55
4 = - c6/r6- C,/r-B - ClO/r-lo .... + (37)
Accurate values for the coefficient of the induced-dipole-induced-dipole term,
c a y have been obtained for a number of like and unlike interaction^.^^ Kramer
and Herschbach 57 used three empirical combination rules and approximate
London and Kirkwood-Miiller formulas to calculate values of C, for 153 unlike
interactions from the C6-coefficientsof the pure components. They find that a
combination rule suggested by Moelwyn-Hughes58

where the a’s are the polarizabilities of the pure substances, predicts the unlike
interaction essentially within experimental error.
When ( ~ l ~ / a=b )Cy/Ctb,
~ equation (38) reduces to the geometric mean
combination
cp = (cpc:”)’/2, (39)
and when a, = ab it becomes the harmonic mean

Neither of these simpler rules is as reliable as equation (38). Tang59has shown

that the geometric mean gives an upper limit to C2b.
Relatively little is known about the short-range repulsive forces between
molecules. Experimental evidence from molecular beam studies 5 5 is compatible
with a geometric mean relation for unlike repulsive interactions

Combination Rules.-The relation between simple empirical potential functions

and the more fundamental and more detailed descriptions of intermolecular
interactions is tenuous. Most simple potentials provide a convenient means by
which the gross details of the interactions can be specified. As a result, one
should not expect that the combination rules that provide the best representation
for the dispersion or repulsive interactions will prove the most satisfactory for
a given potential. A variety of empirical rules for combining pure-component
potential parameters has been proposed; the most widely used have been

Although equations (42) and (43) work reasonably well for simple molecules
not too different in size, deviations from them become evident in analyses of
ss E. A. Mason and L. Monchick, A h . Chem. Phys., 1967,12, 329.
66 A. Dalgarno, Adv. Chem. Phys., 1967, 12, 143.
H. L. Kramer and D. R. Herschbach, J. Chem. Phys., 1970, 53, 2792.
68 E. A. Moelwyn-Hughes, ‘Physical Chemistry’, Pergamon, New York, 1961, 2nd revised
edn., p. 339.
6s K. T. Tang, J. Chem. Phys., 1968,49,4727.
214 Chemical Thermodynamics
careful measurements made over substantial temperature ranges. The deviations
are usually attributed to the failure of the geometric mean rule and equation (42)
is, therefore, modified by the introduction of the parameter f : *

Less commonly, deviations from equation (43) are also considered, and
departures from the arithmetic mean are expressed by the parameter 7

012 = q(a11 + a22)/2. (45)

Values of 6 and 9 for rare gas mixtures are given in Tables 1 and 2, respectively.
They are derived from the work of Lin and Robinsonso who obtained potential
parameters for like and unlike rare-gas interactions by fitting second virial
coefficients with the Dymond-Alder potentiaLsl Quantum corrections were
included in their analysis. For helium the correction procedure is valid only above
40 K, hence they excluded data for He taken at lower temperatures. Different
parameters would be found if the analysis were based on another potential, but
the relative values would be little changed.s2
A study of the unlike interactions in rare-gas mixtures has also been performed
by Kestin and ~ o - w o r k e r s Their ~ ~ which is based on empirical corre-
. ~ ~ ~work,
sponding states functions, leads to significantly different values of 6 for many
of the mixtures. In particular, they find f ' s larger than unity for Ar + Kr,
Ar + Xe, and Kr +
Xe. The parameters they obtained for like interactions
have been chosen to optimize the fit to measured second virial coefficients,
diffusion constants, and viscosities, but most of the unlike parameters have been
based on diffusion data alone. This may be the reason for the disparity between
their results and those of Lin and Robinson.
The values of 4 shown in Table 1 for He Ar, He + Kr, He + Xe, Ne + Kr, +
and Ne + Xe systems are unusually small. Departures from the geometric
mean of a few per cent, such as those for Ar +
Kr and Ar Xe, are more +
typical. In almost all cases f is found to be smaller than unity.
Trends in the departures from the geometric mean are discernible; large
differences in o and/or E generally lead to small values of 4. Accordingly, several
formulae have been devised for the prediction of 4 from the properties of the
pure components

s=
5= (47)

5= 1 + (0.17/eP)(Ii - I# ln(li/Zj), (li > Ij).

6o H. M. Lin and R. L. Robinson jun., J . Chern. Phys., 1971, 54, 52.

J. H. Dymond and 3.J. Alder, J. Chenz. Phys., 1969, 51, 309.
6 2 R. L. Scott, 'Liquid State', Volume VIIIA of 'Physical Chemistry, An Advanced Treatise',

eds. H. Eyring, D. Henderson, and W. Jost, Academic Press, 1971, p. 32 shows E A ~ / E I ( ~is
essentially constant for 10 potentials.
* Another common representation of deviations from the geometric mean uses (1 - klz)
in the place of (.
Volumetric Properties of Gaseous Mixtures 215
Table 1 Comparison of experimental and calculated values of 4 for rare-gas
systems
System 5* Eq. (46) Eq. (47) Eq. (47) 1- Eq. (48)
He + Ne 0.930 (0.036) 0.996 0.884 0.886 0.962
+ Ar 0.788 (0.022) 0.93 1 0.915 0.664 0.776
+ Kr 0.677 (0.068) 0.895 0.890 0.791 0.689
+ Xe 0.61 1 (0.107) 0.824 0.848 0.648 0.577
Ne + Ar 0.860 (0.008) 0.960 0.950 0.968 0.872
+Kr 0.777 (0.007) 0.930 0.932 0.967 0.798
+ Xe 0.661 (0.009) 0.868 0.876 1.028 0.915
+
Ar Kr 0.986 (0.029) 0.995 0.997 1.000 0.973
+ Xe 0.961 (0.01 7) 0.969 0.978 0.991 0.915
Kr + Xe 1.000 (0.032) 0.988 0.991 0.990 0.967
* Numbers in parentheses are standard deviations obtained from least-squares analysis of
second virial coefficients.
t With all = ua2,x1 = xa.
Equation (46), usually attributed to Hudson and M ~ C o u b r e y ,was~ ~ also
proposed by Srivastava and Madan.64 It was obtained by equating the attractive
portion of the Lennard-Jones potential to the London formula for c6. The
quantities 4 are the ionization energies of the pure components. A similar
derivation based on the Kirkwood-Muller treatment of dispersion forces 65 leads
to equation (47), which requires knowledge of the diamagnetic susceptibilities
xi of the pure components. When xi = xj and (iii = q j , equation (47) reduces
to the harmonic mean for EU.
Although equations (46) and (47) predict a weakening of the geometric mean
(4 < l), neither of them is capable of reproducing the very small values of 8
found for unlike interactions involving He and Ne, Hiza and Duncan 66 obtained
equation (48) in a purely empirical procedure designed to correlate the data for
these substances.
The values of 8 calculated for the three rules are also shown in Table 1.
Generally speaking, all of the rules are an improvement over the simple geometric
mean, but no one rule is clearly superior. Of significance is the fact that even
for closely similar spherical molecules, c12 cannot be predicted to within 1 or
2 per cent, the level of accuracy needed for the calculation of the properties of
liquid mixtures.
Combination rules for (i other than the arithmetic mean have also been
proposed. If the geometric mean rules for C, and $12(rep.) are compared to the
corresponding terms in the Lennard-Jones 6-1 2 potential, geometric-mean
rules for both c12and ul2 are ~btained.~'Good and Hope 68 tested a variety of
combination rules for paired with both the arithmetic and geometric mean
rules for u12. They feel that the results prove that the geometric mean for u12is
superior to the arithmetic mean. This conclusion is open to question, however.
None of the 29 systems that they examined would be expected to conform to the
G. H. Hudson and J. C. McCoubrey, Trans. Faraday SOC., 1960,56, 761.
64 B. N. Srivastava and M. P. Madan, Proc. Phys. SOC., 1953, A , 66,278.
65 B. E. F. Fender and G . C. Halsey, jun., J. Chem. Phys., 1962,36, 1881.
66 M. J. Hiza and A. G . Duncan, Amer. Inst. Chem. Engineers J., 1970, 16, 733.
R. J. Good and C. J. Hope, J. Chem. Phys., 1970, 53, 540.
R. J. Good and C. J. Hope, J. Chem. Phys., 1971,55, 111.
8
216 Chemical Thermodynamics
Lennard-Jones 6-12 potential, which was used to obtain E~~ and ul2 from virial
coefficients. Further, few of the systems have been studied over a sufficiently
broad range of temperature to assure that the fitting procedure leads to un-
equivocal results.”~7 0 As shown in Table 2, the geometric mean gives poorer
results than the arithmetic mean for most of the u12’sof the rare-gas mixtures.

Table 2 Comparison of experimental and calculated vaIues of 9 for rare-gas

systems
System 9* 7 calc. t
He + Ne 1 .OOO (0.004) 1 .ooo
+ Ar 1.054 (0.003) 0.992
+ Kr 1.031 (0.010) 0.987
+ Xe 1.106 (0.029) 0.978
Ne + Ar 1.023 (0.006) 0.995
+ Kr 1.024 (0,006) 0.992
+ Xe 1.023 (0.009) 0.983
Ar + Kr 0.999 (0.017) 0.999
+ Xe 0.988 (0.011) 0.996
Kr + Xe 0.997 (0.018) 0.998
* Numbers in parentheses are standard deviations obtained from least-squares analysis of
second virial coefficients.
t Based on 0 1 9 = ( U ~ ~ U ~ J ” ~ .
Geometric-mean combining rules for both E~~ and u12have also been proposed
by Calvin and Reed 71 who use a Mie (n,6) potential and treat the repulsive
index n as an additional parameter. For the unlike interaction the geometric
mean is also used for n. They have also suggested similar rules for the three
parameters in the e x p 6 p~tential.’~ Kong 73 has proposed combination rules
involving the effective number of outer-shell electrons.
Corresponding States Relations.-Given values of E , a, 5, and 7,one can calculate
the second virial coefficientsof simple substances and their mixtures by application
of the principle of corresponding states. Values of reduced virial coefficients
have been tabulated 44 as a function of the reduced temperature for a number of
potentials of the form of equation (35). Alternatively, virial coefficients can be
calculated without recourse to specific potential functions by the use of empirical
relations devised to fit the properties of simple fluids. Critical constants are
generally used as reduction parameters. An example is the equation of McGlashan
and Potter (38)
B / V c = 0.430 - 0.886(TC/T)- O.694(Tc/n2. (49)
Combination rules for Y c and T Cequivalent to equations (44) and (45) are

*O H. M. Lin and R. L. Robinson jun., J . Chem. Phys., 1970,52, 3727.

’* H . J. M. Hanley and M. Klein, US. Nat. Bur. Stand. Tech. Note, no. 360; M . Klein and
H. J. M. Hanley, Trans. Faraday SOC.,1968,64,2927.
D. W. Calvin and T. M. Reed, J. Chern. Phys., 1971, 54, 3733.
’I2 D. W. Calvin and T. M. Reed,J. Chem. Phys., 1972, 56, 6262.
7s C. L. Kong, J. Chem. Phys., 1973,59, 1953.
 Volumetric Properties of Gaseous Mixtures 21 7
Other common reduction parameters are the Boyle temperature TB, and the
Boyle volume
VB = TB(dB/dT),. (52)
Values of B for simple substances can be calculated from corresponding
states with accuracy approaching that of a very careful experiment. This is
not the case for larger and more complex molecules. Deviations from the simple
principle of corresponding states are significant for pseudospherical molecules
such as CF, and C(CH& small dipolar and quadrupolar molecules such as
CH&l and COz, and non-spherical molecules such as C3H8 and C,H,. The
failure of the corresponding states principle for these substances is attributable
to the need for a potential with a form more complex than equation ( 3 5 ) to
describe the molecular interactions. The forces between large spherical molecules
are non-central; the interactions are not simply proportional to the distance
between molecular centres. Interactions between polar molecules are angle-
dependent, as are the interactions between non-polar non-spherical molecules.
If the departure from equation (35) is not too large, it can be treated as a
perturbation. The effect of non-central interactions on the second virial
coefficient is second order and it is very difficult to distinguish between any of
the non-central interactions on the basis of the behaviour of B.74 In essence,
it is possible to fit the properties of many pure substances to equations based on
simple corresponding states and additional terms whose magnitude is pro-
portional to a perturbation parameter that is a measure of the deviation from
central forces. One such perturbation parameter is Pitzer’s acentric factor
~ ; 76 another
~ ~ iss Rowlinson’s In a homologous series the number of carbon
atoms in the chain constitutes yet another measure of the 77

It can be shown that there is a simple relation between the different f a c t o r ~77. ~ ~ ~
A principle of corresponding states based on three parameters can be
enunciated for substances to which the perturbation approach applies : all
substances that have the same value of the perturbation factor have the same
reduced equations of state. For example, the reduced second virial coefficient
can be written in terms of the acentric factor 74s ’*
BpC/RTC= (0.1445 + 0 . 0 7 3 ~ - ) (0.330 - 0.46w)(TC/T)
- (0.1385 + 0.5)(TC/T)2- (0.0121 + 0.0970)(T~/T)~
- 0.0073~(T~/T)~. (53)
If w = 0, equation (53) reduces to an equation applicable to simple substances
such as Ar.
McGlashan and Potter 38 could correlate the second virial coefficients of the
n-alkanes by adding a term to equation (49)
B/Vc = 0.430 - 0.886(TC/T)- O.694(Tc/T)’ - 0.0373(n - l)(TC/T)4.6.(54)
74 K. S . Pitzer and L. Brewer (revision of Lewis and Randall), ‘Thermodynamics’, McGraw-Hill,
New 1-ork, 1961, p. 605.
76 K. S. Pitzer, D. Z. Lippman, R. F. Curl, jun., C. M. Huggins, and D. E. Petersen, J . Amer.
Chem. Soc., 1955,77, 3433.
78 J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, Butterworths, London, 1969, 2nd edn.,
p. 270.
77 E. A. Guggenheim and C. J. Wormald, J. Chem. Phys., 1965’42, 3775.
K. S. Pitzer and R. F. Curl, jun., J. Amer. Chem. SOC.,1957, 79, 2369.
218 Chemical Thermodynamics
Here n is the number of carbons in the chain. If (n - 1) is treated simply as a
perturbation parameter, the equation is applicable to molecules other than the
alkanes.77* 8 o79t

To apply three-parameter corresponding states to mixtures, it is necessary to

have a combination rule for the third parameter. A linear combination of the
third parameters for the pure components is usually assumed, but there is some
evidence that the geometric mean is more satisfactory.81 Examples of the
application of a three-parameter principle of corresponding states to interaction
virial coefficientsare the work of Dantzler et al. who used the Potter-McGlashan
equation to correlate B,,’s for hydrocarbon,81 f l u ~ r o c a r b o n ,argon
~~ hydro- +
+
carbon,2g and hexafluorobenzene benzene 82 mixtures. Rowlinson’s 6
parameter was employed by King and co-workers in their analyses of the
solubility in gases of CC14,83~ h e n a n t h r e n eand
, ~ ~ anthracene.86
An extension of corresponding states to highly polar substancess6 has been
based on the acentric factor correlation. A fourth parameter related to the
dipole moment must be introduced. A similar four-parameter description of
the virial coefficients of polar fluids and their mixtures has been developed by
Halm and Stiel 87 and for mixtures of quadrupolar fluids by Ramaiah and Stie1.88
The Principle of Congruence.-In 1946, Brarnsted and Koefoed 8 9 demonstrated
that there is a simple empirical relation between the properties of the normal
alkanes and their mixtures. The relation, termed the principle of congruence,
is formulated in terms of Fi, the average chain length of the mixture, which is a
function of the mole fractions and the number ni of carbon atoms of each of the
components

To a very good approximation the properties of a mixture are the same as the
properties of the pure alkane of chain length f i .
Barker and Lintonso were the first to apply the principle of congruence to
virial coefficients. They treated data on three hydrocarbon mixtures and a
number of binary mixtures of alkanes with hydrogen. (The chain length n = 0
was assigned to H2.) Within experimental error the virial coefficients of the
mixtures were equal to the virial coefficient of the hydrocarbon with chain
length E. Newer measurements made it possible to test the principle of con-
gruence more thoroughly for mixtures of n-alkanes 81 and also for n-perfluoro-
alkanes.79 The principle holds reasonably well for the hydrocarbons and some-
what more poorly for the perfluorocarbons. Stryjek 91 applied the principle of
79 E. M. Dantzler and C . M. Knobler, J. Phys. Chem., 1969, 73, 1335.
M. V. Pandya and A. G . Williamson, Austral. J . Chem., 1971, 24,465.
E. M. Dantzler, C. M. Knobler, and M. L. Windsor, J . Phys. Chem., 1968, 72, 676.
82 E. M. Dantzler and C . M. Knobler, J. Phys. Chem., 1969, 73, 1602.
83 S. K. Gupta and A. D. King, jun., Canad. J. Chem., 1972,50, 660.
84 H. Bradley jun. and A. D. King, jun., J. Chem. Phys., 1970,52, 2851.
85 G . C. Najour and A. D. King, jun., J. Chem. Phys., 1970,52, 5206.
J. P. O’Connell and J. M. Prausnitz, Znd. and Eng. Chem. (Process Design), 1967, 6, 245.
87 R. L. Halm and L. I. Stiel, Amer. Inst. Chem. Engineers J., 1971, 17, 259.
8s V. Ramaiah and L. Stiel, Ind. and Eng. Chern. (Process Design), 1973, 12, 305.
89 J. N. Bransted and J. K. Koefoed, Kgl. danske Vidensk. Sclsk. (Mat. Fys. Ski-.), 1946, 22, 1.
Ha J. A. Barker and M. Linton, J. Chem. Phys., 1963, 38, 1953.
g1 R. Stryjek, Bull. Acad. Polon. sci., ski-. Sci. chint., 1967, 15, 355.
Volumetric Properties of Gaseous Mixtures 219
congruence to alkyl formates with fair success, and Pandya and Williamsonso
have tested it for n-alkanes, n-perfluoroalkanes, and n-alk-1-enes. All of these
recent studies show that the principle of congruence, while useful, is not as
reliable a predictive method for virial coefficients as three-parameter corre-
sponding states.
Non-central Potential Functions.-Non-central interactions can also be handled
in terms of specific multi-parameter potential functions such as the Stockmayer
and Kihara potentials.lS 44 Although there is an extensive literature on the
2l

subject of such potentials, it is only rarely that virial coefficients are sufficiently
precise or extensive to warrant their use. For this reason we will deal only
briefly with the subject.
When multi-parameter potentials are applied to mixtures, the combination
rules for parameters other than E and cr often follow from the model underlying
the potential. For example, the Kihara spherical-core potential is based on a
model in which each molecule has a molecular core of characteristic radius aii.
+
The combination rules aij = (aii ajj)/2 follows directly from the hard-core
a s ~ u m p t i o nD2. ~ ~ ~
Prausnitz and MyersD3have used the Kihara potential for the calculation of
for mixtures of simple molecules, and Blanks and PrausnitzD4employed
the Stockmayer potential in studies of mixtures involving polar molecules. A
variety of specific non-central interactions were included in the potentials that
Lichtenthaler and Schafer D 5 employed in their analysis of their measurements
of virial coefficients of Ar with polar and non-polar molecules.
More recently, Khoury and Robinson D6 calculated virial coefficients for
ethane + hydrogen sulphide mixtures from an effective Lennard-Jones 6-1 2
potential that includes a contribution from the dipole-induced-dipole interaction.
The parameters were evaluated by fitting their measurements on C,H, and H,S
to a 6-12 potential and to a Stockmayer potential respectively. Bradley and
King 84 have used several potential functions to analyze interactions of phen-
anthrene with a number of small polyatomic molecules.
The many models used for the calculation of virial coefficients of globular
molecules have been reviewed by Johnson and S p ~ r l i n g .They~ ~ calculate B1,’s
forCH, + CF4,CH4 + SF,,CH4 +
C(CH,),,CH, +
Si(CH,),,andCF, + SF,
from the Mamann-Lambert m0de1.~~~In this model all atom-atom inter-
actions are assumed to be of the Lennard-Jones type. The interaction energy is
calculated by summing all pairwise atom interactions and then averaging equally
over all relative orientations of the molecules. A similar approach for mixtures
of globular molecules loois based on interactions between spherical shells. I n
effect, the outer atoms of a molecule are assumed to be uniformly distributed
over the surface of a sphere.
92 E. M. Dantzler-Siebert and C. M. Knobler, J. Phys. Chem., 1971, 75, 3863.
93 J. M. Prausnitz and A. L. Myers, Amer. Inst. Chem. Engineers J., 1963, 9, 5.
g4 R. F. Blanks and J. M. Prausnitz, Amer. Inst. Chem. Engineers J., 1962, 8, 86.
g5 R. N. Lichtenthaler and K. Schafer, Ber. Bunsengesellschaft Phys. Chem., 1969, 73, 42.

96 F. Khoury and D. B. Robinson, J. Chem. Phys., 1971, 55, 834.

97 C. H. J. Johnson and T. H. Spurling, Austral. J. Chem., 1971,24,2449.
98 S . D.Namann and J. A. Lambert, Austral. J. Chem., 1954, 7 , 1.

g9 J. A. Lambert, Austral. J. Chem., 1959, 12, 109.

l o o R. R. Dorsch and A. G . DeRocco, J. Chem. Phys., 1971,54,2727.

220 Chemical Thermodynamics
Relation Between Gaseous and Liquid Mixtures.-As we have indicated, second
virial coefficients can be calculated by corresponding-states methods with
sufficient accuracy to be useful in predicting the properties of pure gaseous
substances and their mixtures. This is of considerable practical importance for

0.15

0.10 0 2
u=,

0.05

Figure 5 Departures from the geometric mean determined from vnpour-phase measure-
ments,tV, vs. values from liquid-phase measurements, &
1. CO2 C4Hlo 2. CO2 C3H6 3. CO2 C3H8 4. CF4 C2H6
+ +
++
5. C2F6 4- &Ha 6. C4F10 C4Hlo 7. C5F12 C5H12 8. C ~ F S C3H8
9. GF14 +
GH14 10. H,S +C3H6 11. HZS c3H8 12. CO2 C2He +
13. C02 + C2H4 14. H,S +C&I2 15. CF, + CH, 16. Ar Kr +
17. Ar + CO 18. CH, + CO 19. Ar +
CH, 20. Ar N2 +
workers interested in the properties of liquid mixtures because equation-of-state
data are needed for the analysis of many thermodynamic measurements. A more
significant relation between liquid mixtures and the equation of states of gases is
the use of virial coefficients to obtain values of f .
All theoretical calculations of the properties of binary liquid mixtures depend
very sensitively on (2q2 - cll - E ~ ~ )When . information concerning the unlike
interaction is lacking, it has been customary to assume that <
= 1, but even
small departures from the geometric mean can have radical effects on the
calculated propert ies. For example,lo1in solubil it y-parameter theory the activity
coefficient at infinite dilutionf,”, is given by
RTInfF = V T [ ( 6 , - 82)2 + 2(1 - 5)8,6,]. (56)
lol J. H. Hildebrand, J. M. Prausnitz, and R. L. Scott, ‘Regular and Related Solutions’, Van
Nostrand Reinhold, New York, 1970, p. 99.
 Volumetric Properties of Gaseous Mixtures 221
In a typical mixture the solubility parameters of the components, 6, and a,,
might be 7.0 and 8.5 ca11’2cm-3/2, respectively, and VT, the molar volume of
component 1, 100cm3mol-l. For this mixture at 300K, f? = 1.46 when
e = 1. A 1 per cent reduction in f changes fr
to 1.78, and for = 0.95,
f? = 3.96.
It is clear that no significant comparison between theory and experiment can
be made unless 6 is known. That values of 6 determined from gas-phase measure-
ments are relevant to the properties of dense fluid mixtures is demonstrated in
Figure 5 ; shown plotted are the values of 6 determined from virial data on
mixtures vs. &values that give the best fit to the properties of liquid mixtures
of the same substances. The gas-phase values are probably no more reliable than
0.01 to 0.02. In fitting data on liquid mixtures, variations in f of f0.002 may
be significant, but the (-values derived depend on the theory chosen to represent
the data. As a result, the uncertainties in liquid and gaseous values of f are
probably comparable.
Given the precision of the data, the correlation between the low- and high-
density values of f is quite good. For some mixtures, however, the liquid-phase
values are markedly smaller than those obtained from the gas. These are
typically systems such as H2S + C2H4for which a specific acid-base interaction
in the liquid phase might be expected. Generally speaking, however, gas-phase
determinations of 4‘ provide valuable information for the prediction and
correlation of the properties of liquid mixtures,

5 Vapour-phase Complexes
An early approach to the problem of gas imperfection was to treat deviations
from ideality as the result of chemical association. Consider a gaseous mixture
of substances A and B that can react
A + B = AB.
If nA and ng are the formal amounts of substance of each component and nAB
is the amount of substance of complex AB, at equilibrium the application of the
perfect gas law gives
PV/RT = n~ -I-n~ - n o = ( n ~ + - ~ A E / ( ~4-A nB)]
= (nA -k nB)[1 - xAxBR%(nA -k nB)/Vl, (57)
where we have introduced K , = P A B f P A P R , the equilibrium constant written in
terms of partial pressures. Comparison of equation (57) with equations (1) and
(8) shows that if complex formation is the only source of nonideality, B12 is
simply related to K , :*
Bl2 = - RTKJ2. (58)
For a binary mixture of associating imperfect gases there will be ‘physical’ as
well as ‘chemical’ contributions to B12. If the concentration of complex is small,
B12 = B,,(phys.) + B12(chem.). (59)
* For dimerization of a pure substance equation (58) becomes B1, = -RTK,. Self-association
in the gas phase has been studied for a number of systems, e.g. R. L. Scott, Mol. Phys., 1966,
11,399; J. 0. Lambert, G. A. H. Roberts, J. S. Rowlinson, and V. J. Wilkinson, Proc. Roy. Soc.,
1949, A, 196, 113.
222 Chemical Thermodynamics
B,,(phys.) is the interaction virial coefficient arising from non-chemical inter-
actions between the components and B,,(chem.) is given by equation (58). Since
B,,(chem.) is negative, evidence for association would be an unusually large
negative BIZor a negative Q.
If BIZfor an associating mixture is measured and B,,(phys.) is calculated by a
corresponding-states procedure, B,,(chem.) can be obtained by difference. The
enthalpy change in the association reaction can also be derived from studies of
the temperature dependence of B,,(chem.).
Among the first studies of vapour phase association by equation of state
measurements was the work of Alexander and Lambertlo2 who studied the
dimerization of pure acetaldehyde. They calculated B,,(phys.) from the
Berthelot equation. Hydrogen bonding in mixtures of chloroform with carboxylic
acids, esters, and diethylamine was investigated by Lambert et a1.,lo3and Lambert
also studied the chloroform +diethyl ether system.lo4 Complex formation
between acetonitrile and acetaldehyde was reported by Prausnitz and Carter.lo6
In their study of association in ammonia +
acetylene mixtures, Cheh,
O’Connell, and Prausnitz lo6 calculated the physical contribution to B12 from
potentials that included hard-core as well as multipole interactions. The
existence of vapour-phase complexes of ethylene with ammonia and methanol
and of methanol with pentane has been inferred from virial coefficient data.lo7
King and co-workers have obtained association constants for COz with
naphthalene,20 ethanol,log and diethyl ether,l1° and for H 2 0 with
C 0 2 and N20.111
A specific interaction in the vapour phase between hexafluorobenzene and
benzene was reported by Dantzler and Knobler.82 Recently, Belousova and
Verkhova 112 pointed out that interaction virial coefficients for benzene +
n-hexane have an unusual temperature dependence and are inconsistent with
B12’sfor benzene with other alkanes. They suggest that this abnormal behaviour
is the result of association. Vapour-phase amine complexes with methanol have
been studied by i.r. spectroscopy and p, V , T-methods,ll3~11* Equilibrium
constants were calculated from the volumetric data under the assumption that
the vapours are ideal.
We have derived equation (59) by ‘handwaving’; it is useful to attempt a more
fundamental derivation.l15 For a potential that has no angular dependence, the
separation of the interaction between two molecules into a physical portion and a
lo* E. A. Alexander and J. D. Lambert, Trans. Faruduy SOC.,1941, 37, 421.
Io3 J, D. Lambert, J. S. Clarke, J. F. Duke, C. L. Hicks, S. D. Lawrence, D. M. Morris, and
M. G. T. Shone, Proc. Roy. SOC.,1959, A , 249,414.
lo* J. D. Lambert, Discuss. Furuday Soc., 1953, 15, 226.
lo6 J. M. Prausnitz and W. B. Carter, Arner. Inst. Chem. Engineers J., 1960, 6, 61 1.
lo6 H.Y. Cheh, J. P. O’Connell, and J. M. Prausnitz, Cunud. J. Chem., 1966, 44, 429.
lo’ M. Ratzsch and H. Freydank, J. Chem. Thermodynamics, 1971, 3, 861.
lo* B. Hemmaplardh and A. D. King, jun., J. Phys. Chem., 1972,76,2170.
lo9 S. K. Gupta, R. D. Leslie, and A. D. King, jun., J. Phys. Chem., 1973, 77, 2011.
R. Massoudi and A. D. King, jun., J. Phys. Chem., 1973, 77, 2016.
ll1C. R. Coan and A. D. King, jun., J . Amer. Chem. Suc., 1971, 93, 1857.
112 Z. S. Belousova and V. f.Verkhova, Russ. J. Phys. Chem., 1973, 47, 236.
113 M. Fild, M. F. Swiniarski, and R. R. Holmes, Inorg. Chem., 1970, 9,839.
114 D.J. Millen and G. W. Mines, Trans. Faraday Soc., 1974, 70, 693.
115 L. S. Moore and J. P, O’Connell, J . Phys. Chem., 1972,76, 890.
Volumetric Properties of Gaseous Mixtures 223
chemical portion is rather arbitrary. If the potential is assumed to be a super-
position of a short-range chemical part and a weaker, longer-range physical part,
a distance could be chosen as the limit of the chemical interaction, but even then
the separation into two contributions is not clear cut. Consider the potentials
illustrated in Figure 6. The one associated with complex formation has a well

Figure 6 Schematic potential-energy functions representing chemical (upper) and

physical (lower) interactions

depth u1 and contributes negligibly to the interaction beyond a distance rl. The
physical potential falls off more slowly and has a depth u2. To calculate the
second virial coefficient, we substitute the combined potentials into equation (2)

B = - 2xN.( /o"{exp[ - (u, + u,)/kT] - l}r2 dr

+ Ir:[exp(- u2/kT) - l]r2dr)

= - 27cN~(Iof1(exp[- (ul + u2)/kT] - 1 - exp( - u2/kT)+ l ) r 2dr

+ Jom[exp(- u,/kT) - l]r2dr)

= B(phys.) - 2 n N ~Jorl[exp(- u,/kT) - l]exp( - u,/kT)r2dr. (60)

The second term in equation (60) is the part of the virial coefficient that would
be termed B(chem.). In general, the separation of the chemical and physical
portions is not complete; the physical interaction weights the chemical term
224 Chemical Thermodynamics
[exp(- uJkT) - 11. As a further consequence of the factor exp(- u,/kT) the
temperature derivative of B(chem.) does not relate simply to the enthalpy of
association.
The exponential in equation (60) containing u2 can be expanded

B(chem.) = - 2 x N ~. ["[exp( - u,/kT) - 1111 - (u,/kT) + ...I r2 dr, (61)

o

so that for the conditions u1 B u2 and uJkT < 1, B(chem.) depends only on
the chemical interaction. This is the case of a strong association for which the
simple derivation of equation (59) should be valid. Thus, when there is strong
association, interaction virial coefficients can be used to determine equilibrium
constants and enthalpies of reaction. As the complex becomes weaker, however,
the analyses become increasingly suspect. It can be said that when - B I 2 is
unusually large and has a marked temperature dependence, chemical association
may be the cause. The delineation of the specifics of the interaction can be
obtained only from a detailed analysis based on equation (60), a procedure that
has never been performed.
Rowlinson 116 makes the point that when the chemical interaction is limited to
a narrow range of relative orientations between the molecules, the assumption
that the chemical and physical forces are separable is likely to be valid. Such
highly directional interactions are usually of high energy and therefore correspond
in the case of central forces to the condition u1 %- Ua.

6 Third Virial Coefficients of Mixtures

Reliable methods for estimating and correlating third virial coefficients of
mixtures have not yet been developed. Even for pure substances, the extent of
our knowledge concerning C is very limited. In part this ignorance is due to a
lack of reliable data. No techniques for the direct determination of C have been
devised, and third virial coefficients can be extracted from p , V, T-data only if
they are extensive and highly precise. The temperature dependence of the third
virial coefficient has been determined for relatively few pure substances; inter-
action virial coefficients are known for only a handful of mixtures.
Theoretical treatments of the third virial coefficient must deal with an effect
not encountered in calculations of B. The interaction between three molecules
is not strictly additive, that is, it is not the sum of the pair interactions. In recent
years the effect of non-additive contributions to the attractive forces has received
considerable attention 11' and it is clear that for simple molecules such as Ar the
non-additive portion of C can be as large as 10 to 15 per cent of the total. The
contribution of any non-additive repulsive force has not yet been ascertained.
Crude quantum mechanical estimates 118 indicate that it is of opposite sign to the
non-additive dispersion interaction and similar in magnitude. Its effect on the
third virial coefficient is presumably small, however, because there is good
11* J. S. Rowlinson, Trans. Faraday SOC.,1949, 45, 974.
11' For example see C. T. Chen and R. D. Present, J . Chem. Phys., 1972, 57, 757; C. H. J,
Johnson and T. H. Spurling, ibid., 1972, 57, 1804; H. J. M. Hanley and R. D. McCarty.
ibid., 1972, 51, 3023.
A. E. Sherwood, A. G. DeRocco, and E. A. Mason, J. Chem. Phys., 1966,44,2984.
 Volumetric Properties of Gaseous Mixtures 225
agreement between experimentally determined C’s and those calculated assuming
non-additive contributions only to the dispersion forces.117
The exact calculation of interaction third virial coefficients is very complicated.
It was not until 1968 that Stogryn 119 obtained C’s for mixtures of molecules that
interact with a Lennard-Jones potential. In later work, he extended the
calculations to include multipolar interactions 120 and also treated non-additivity
in the multipolar terms.121
A semi-empirical method for calculating the additive portion of the interaction
third virial coefficient from a knowledge of the reduced third virial coefficient
of a pure substance was proposed by Rowlinson, Sumner, and Sutton.122Their
equation can be written in the form

where Cfyd(Tf2)is the reduced third virial coefficient of a pure substance at the
reduced temperature Tf2 = k T / ( ~ & ) l ’ ~ .
A similar approximation, due to Orentlicher and P r a ~ s n i t zuses
,~~~

in place of CTtdd. Each of the reduced virial coefficients in equation (63) is to be

calculated from a specific potential, and the unlike parameters are to be obtained
from the appropriate combination rules. Stogryn 119 compared approximations
(62) and (63) to his exact Lennard-Jones calculations. He found that neither is
reliable near the maximum in C unless the potential parameters of all the inter-
acting molecules are nearly the same. At higher reduced temperatures both
approximations are satisfactory, and equation (62) gives somewhat better results
than that of Orentlicher and Prausnitz.
Chueh and Prausnitz 124 analysed third virial coefficients for pure substances
and obtained a generalized three-parameter corresponding-states equation for C.
Interaction virial coefficients can be calculated from their correlation by either of
the methods already described. When Pope et aZ.125attempted to analyse their
low-temperature measurements of the third virial coefficients of Ar, CHI, and
C2Hs, they found that even when non-additivity was taken into account they
were unable to fit the data with a specific three-parameter potential (Kihara
spherical core) for which Sherwood and Prausnitz 126 had obtained third virial
coefficients. To correlate their data, Pope et al. derived a corresponding-states
relation in terms of the acentric factor and based on an empirical equation for C
originally formulated for CHp.lZ7Their treatment is applicable only for values
of o between -0.002 and 0.10. It is claimed that this equation is superior to
that of Chueh and Prausnitz primarily because it has been obtained by fitting
data over a larger range of reduced temperature.
ll9 D. E. Stogryn, J . Chem. Phys., 1968, 48,4474.
lao D. E. Stogryn, J. Chem. Phys., 1969, 50, 4967.
lal D. E. Stogryn, J. Chem. Phys., 1970, 52, 3671.
12z J. S. Rowlinson, F. H. Sumner, and J. R. Sutton, Trans. Furaday Sac., 1954,§0, 1.
las M. Orentlicher and J. M. Prausnitz, Canad. J. Chem., 1967, 45, 373.
lz4 P. L. Chueh and J. M. Prausnitz, Amer. Inst. Chem. Engineers J., 1967, 13, 896.
la6 G. A. Pope, P. S. Chappelear, and R. Kobayashi, J. Chem. Phys., 1973, 59,423.
lZ8 A. E. Sherwood and J. M . Prausnitz, J. Chem. Phys., 1964,41, 429.
lZ7 A. J. Vennix, T. W. Lelai 1, and R. Kobayashi, Amer. Inst. Chem. Engineers J., 1969,15,926.
226 Chemical Thermodynamics
7 Excess Functions for Moderately Dense Gas Mixtures
Although the virial equation of state can be applied to moderately dense gases,
the difficulty in determining third and fourth interaction virial coefficients makes
its application to mixtures impractical. It has become increasingly common in
dealing with moderately dense gas mixtures to use the experimental approaches
and theories usually associated with liquid mixtures. For example, one can
describe the volumetric properties of gas mixtures by the excess volume rather
than in terms of virial coefficients. The focus is on excess functions rather than
on the properties of the mixture.
Methods for Measuring V g and H&-Beenakker and co-workers were among
the first to use this approach. The Edwards and Roseveare technique13 was
adapted to high pressure by Zandbergen and Beenakker,12*who measured the
volume change required to maintain constant pressure when two gases were
mixed. Analysis of the experiment gives
V g = An(xlVf + x2Vz)/(nl + n2 - An),
where n, and n2 are the amounts of substance of each gas that is mixed, and An
is the amount of substance of gas that must be withdrawn or added after mixing
to keep the pressure constant. The equations of state of the pure components
are used to calculate Vr and Vg at the experimental conditions and n, and n2
are calculated from the calibrated volumes in the apparatus. At low pressures
the V g is a strictly parabolic function of the mole fraction [equation (S)], hence
measurements need be made at only one composition. Since this symmetry is
not inherent in at higher pressures, it is necessary to make measurements at
a number of compositions. This is accomplished by using vessels of different
sizes.
Low-pressure enthalpies of mixing calculated from B12data have been reported
for hydrocarbon and fluorocarbon 7B mixtures, and for mixtures of benzene
with hexane, heptane, and 0 ~ t a n e . l Direct
~~ measurements of enthalpies of
mixing of moderately dense gases have been reported by a number of
groups.130-'35
A typical flow calorimeter 134 designed for studies of HE is shown schematically
in Figure 7. The gases to be mixed are supplied from high-pressure cylinders and
pass through flow controllers, A, A'. Heat exchange coils, B, B', serve to bring
the gases to the temperature of the thermostatic bath or cryostat C that defines
the temperature of the experiment. The gases mix in the calorimeter D. Electrical
power supplied to heater E is used to return the gas mixture to the bath tempera-
ture. This isothermal condition is determined by the response of a thermopile F
that measures the temperature difference between the inlet and outlet gas streams.
12* P. Zandbergen and J. J. M.Beenakker, Physica, 1967, 33, 343.
lZ9 V. M. Prokhorov, Z. S. Belousova, and V. P. Verkhova, Rum. J. Phys. Chem., 1972,46, 1083.
130 B. van Eijnsbergen and J. J. M. Beenakker, Physica, 1968, 39, 499.
131 A. V. Hejmadi, D. L. Katz, and 3. E. Powers, J. Chem. Thermodynamics, 1971, 3, 483.
132 M. Knoester, K. W. Taconis, and J. J. M. Beenakker, Physica, 1967, 33, 389.
lSs J. I. Lee and A. E. Mather, J. Chem. Thermodynamics, 1970, 2, 881.
134 C. J. Wormald and K. L. Lewis, Proceedings of the 3rd International Conference on Chemical
Thermodynamics, Baden, 1973, Vol. 11, p. 212.
135 J. C. W. Bosschieter, Doctoraal Scriptie, University of Leiden, 1971.
 Volumetric Properties of Gaseous Mixtures 227

3
Figure 7 Schematic diagram of enthalpy of mixing apparatus
(Adapted from Wormald and Lewis, ref. 134)

The pressure is held constant by the controller G and is measured by the Bourdon
gauge, H. Flow rates are determined at atmospheric pressure by a gasometer I
and the mixture is analysed by a gas chromatograph J. The calorimeter must be
carefully designed to ensure complete mixing without producing a significant
pressure drop. Trial measurements in which the same substance ir used in both

I I I

10 v/ \\

0 0.5 1
X
Figure 8 Volume change on mixing Vg/cm3 mol-l vs. x(H& for the system Ar + H,
at 171 K. The upper curve is at p = 8.5 MPa, the middle at p = 5.0 MPa, and the
lower at 0.5 MPa [data froin Zandbergen and Beenakker (ref. 128)]
228 Chemical Thermodynamics
gas streams must be performed at a variety of conditions to check for systematic
errors.
Behaviour of the Excess Functions.-Some characteristic VE data are shown in
Figure 8. At low pressures both VE and H E / p are consistent with equation-of-

I I

0 5 19
p/MPa
Figure 9 Vg/cm3 mol-1 vs. p/MPa for the system Ar + H, at 171 K (adapted from
ref. 128)

state determinations of d and 8 - Tda/dT. As the pressure is increased VE

becomes pressure dependent and H E is no longer linear in p . The mole-fraction
dependence of the excess functions also departs from a parabolic dependence as
p is increased. The pressure dependence of the excess functions can be more
readily seen when the excess functions at constant composition are plotted
against p, Figures 9 and 10.
The shapes of the HE us. p curves in Figure 10 have been explained in
qualitative terms by Knoester and B e e n a k k e ~ . At ~ K the HB + H2and
~ ~ 231
CH, + Ha interactions can be assumed to be small compared to the CHI + CHI
interaction. As a result the major contribution to H E is the change in H,R* the
residual enthalpy of CHI when, on mixing, its partial pressure goes from p to
WH4lP
H%,P, T ) x W * (XP,T ) - H W P , TI>. (65)
M.Knoester and J. J. M. Beenakker, Physica, 1967, 33,410.
Volumetric Properties of’Gaseous Mixtures 229
The molar residual enthalpy of CH4 as a function of the pressure at 231 K is
shown in Figure 11. Also included in the plot is a line giving HF* at 0 . 5 6 ~ .
From equation (65), H g at this temperature should be roughly proportional to
the vertical distance between the two lines. We see that H i increases with

I I
I

0 5 10
p/MPa
Figure 10 (Hg/’)J mol-l MPa-l vs. p/MPa for the system CH, +
H, at 231 K
[data from van Eijnsbergen and Beenakker (ref. 130)]. The dashed line has been
calculated from equation (65)

pressure and then goes through a maximum because of the flattening of the HR*
curve at high pressures. The dotted line in Figure 10 was calculated by the
approximation given in equation (65).
The applicability of corresponding-states theories of liquid mixtures has been
tested for all the dense gas systems for which the excess functions have been
directly determined.133,135-139 The results are inconclusive. For many of the
mixtures studied there is a large difference in the critical temperatures between
the components. In such cases, as shown above, the excess functions at moderate
pressures depend primarily on the properties of the component of higher critical
temperature - they are insensitive to the details of the mixed interactions. At
higher pressures, where the test of theory is more significant, the calculated
properties depend markedly on the choice of reference Even simple
lS7 B. van Eijnsbergen and J. J. M. Beenakker, Physica, 1968, 39, 519.
lS8 B. Malesinska and J. Stecki, Physica, 1973, 66, 380.
B. Malesinska and J. Stecki, Proceedings of the 3rd International Conference on ChemicaI
Thermodynamics, Baden, 1973, Vol. 11, p. 146.
230 Chemical Thermodynamics
substances do not adhere to the principle of corresponding states over broad
ranges of the reduced temperature.
In general the two-fluid average potential model 1409 141 gives better results
than the one-fluid model. The three-fluid model counts all binary inter actions
and is, therefore, properly used at low densities. The van der Waals one-fluid

-1

*U

-2 '

0 5 10
p /MP8
Figure 11 Reduced molar residual enthalpy of CH, vs. p/MPa at 231 K. The upper
curve is H E * at 0.569 [datafrom van Eijnsbergen and Beenakker (ref. 130)]
model142 is generally in better agreement with experiment than the average
potential models, particularly for molecules that differ substantially in size;
however, its superiority is not as marked as in the case of liquid mixtures.
Measurements at higher pressures and for a wider variety of systems must be
made before more definitive tests of theories can be made.
A summary of H E and VE measurements is given in Table 3.
Table 3 Summary of direct measurements of excess functions of moderately
dense gases. Numbers indicate references in body of text
HE VE
Ar + H, 132 128
+ He 130
+ N2 132 128
+ CH4 130
H2 + N2 132 128
+ CH, 130
+ co, 131, 133
He + N, 135
+ CH, 130
N2 + CH4 130, 134
140 I. Prigogine, 'Molecular Theory of Solutions', North-Holland, Amsterdam, 1957.
lal R. L. Scott, J . Chem. Phys., 1956, 25, 193.
142 T. W. Leland, J, S. Rowlinson, and G. A. Sather, Trans. Furuday Soc., 1968, 64, 1447.
Volumetric Properties of Gaseous Mixtures 23 1
8 Bibliography of Measurements of Interaction Virial Coefficients
A literature survey covering measurements of interaction virial coefficients up
to 1967 is included in Mason and Spurling’s monograph., Listed in Table 4 are

Table 4 Literature survey

System Refs. System Refs.
Ar + H2 7 + C8H18 47
+ He 3,42,* + CsH,, (2,2,4- 47
7 trimethylpentane)
+Kr 8 F6S + CClF, 40
+ N2
+
7 + CF, 14
Ne 7 H, + He 7
+ CC14 19, 47 + N2 7
+ CF4 14 + Ne 7
+ CHCl3 47 + CCl, 19
+ CH,Br 33 + C2H60(ethanol) 20
+ CH3Cl 33 + C6H6(benzene) 18, 10,
+ CH,N02 47 17
+ CH4 8,33 + C8H18 17
+ CH40 (methanol) 22 + GH1, (2,2,4- 17
+ CO, 4, 33 trimethylpentane)
+ c2H4 5 H2O + N2O 9
+ C2HbNO2 47 + co, 9
+ C2H6 33 + GH8 9
+ C2H60(ethanol) 20, 47 H2S + C2H4 29
+ C3H8 33 He + Kr 25
+ C;H80 (propan- 47 + N2 32, 7,
2-01) 21
+ C4H80(methyl- 47 + Ne 7
ethyl ketone) + Xe 26
+ C4H80, (dioxan) 47 + CC14 19
+ C4H10 33 + C6H6(benzene) 10,17
+ C4H100 (butan- 35 Hg + CH,O (methanol) 46
1-01) + C3H60(acetone) 46
+ C4HlOO (ethyl 35 Kr + Ne 37
ether) + CH4 8
+ C5H5N(pyridine) 47 N, + CCl, 19,47
+ C5H12 15, 35, + CHCI, 47
47 + CH3NO2 47
+ C5H12 (2-methyl- I5 + CH40 (methanol) 22
butane) + C2H5N02 47
+ c&&(benzene) 18, 10, + C,H,O (ethanol) 47
17, 47 + C4H80(methyl- 47
+ C6H12 (cycle- 47 ethylket one)
hexane) + C4H802(dioxan) 47
+ C6H1202 (butyl- 47 + C4H10 48,23
acetate) + C4H100 (ethyl 35
C6H14 15,47 ether)
+ C6H14(2,2-di- 15 + C4HlOO (butan- 35
methylbu t ane) 1-01)
+ C6H142-methyl- 15 + C5H5N(pyridine) 47
pentane) + C5H12 35, 48,
+ C7H8(toluene) 47 23, 47
C7H16 47 + C5H12(2-methyl- 48
+ C8H8(styrene) 47 butane)
232 Chemical Thermodynamics
Table 4 (cont.)
System Refs. System Refs.
+ c6F6 48 + GH8 I6
+ C6H6(benzene) 18, 12, + C4H10 16
47, 10, + C4HlOO (butan- 35
17, 48 1-01)
+ GHl, (Cycle- 48, 47 + C4H100 (ethyl 35
hexane) ether)
+ GH1203 (butyl 47 + GH12 16, 41,
acetate) 35
+ C6H14 48, 23, +C5H12 (2-methyl- 41
47 butane)
+ C,H8 (toluene) 47 +C6H6(benzene) 10, I7
+ GH14 (hept- 48 + C6H14 41,16
l-ene) +CeH14 (2-methyl- 41
+ GH16 48, 23, pentane)
47 +C6H14(2,2-di- 41
+ GH8 (styrene) 47 methylbutane)
f C8H18 48,47 +CloH8 (naph- 38
+ GHl8 (2,2,4- 47 thalene)
trimethylpentane) +C34H10 (anthra- 39
f C9H20 47 cene)
NH3 + GH4 44 + C34Hio (phen- 6
+
N 2 0 CH40(methanol) 22 anthrene)
+ G H 6 0 (ethanol) 20 CH,O(methanol) + C 0 2 22
+ GOH,(naph- 30 +
C2H4 44,22
thalene) + 22
+ G4H10 @hen- 6 + C5H,2 44
+
ant hrene)
0, C6H6(benzene) I7
+ CeH, 31
(benzene)
+
CC14 CH4 19 CO + GHe (benzene) 17
+ co, 47, I9 + C8H18 17
+ GH4
+
19 +
C 0 2 &H5N02 47
G H , (benzene) 45 + C2He0(ethanol) 47,20
CF4 + CHF3 32* + C,H,O(acetone) 47
+ CH4 14 + GH80 47
+ GF6 12 (propan-2-01)
+ C2H6 14 + C4H80(methyl- 47
+ GF8 12 ethyl ketone)
+ C*ho I2 + C4H,02(dioxan) 47
+ C4H10 14 + C4H10O(butan- 35
+ C5FU I2 1-01)
+ C,H,, (2,2-di- 14 + C4Hlo0 (ethyl 35
methy 1butane) ether)
+(314
+ Ck3H14
12 + C5H,N (pyridine) 47
14 + C5H12 47,35
CHCI, + CO, 47 + C6H, (benzene) I1,27,
+ C2H60(ethanol) 34 47
+ C6H14 34 f C6Hi.a (CyClO- 27,47
CH,Cl, + C,H,Cl 43 hexane)
CH,Br + C2H5Cl 43 + C6Hl2O2(butyl 47
CHBNOz + C02 47 acetate)
CH4 + CH40 (methanol) 22 + C6H14 27,47
f C2H4 36* + C,H,(toluene) 47
+ C2H6 24,* 16 + GHl6 47
+ C2H60(ethanol) 20 + C8H8 (styrene) 47
Volumetric Properties of Gaseous Mixtures 233
Table 4 (cont.)
System Refs. System Refs.
+ C8H18 47 CjH,O(acetone) + c6H6 31
+ 5H18 (2,2,4- 47 (benzene)
trimethylpentane) GH8 + C4H10 16
+ CloH8 (naphtha- 35 + C6H12 16
lene) + C6H14 16
+ C14H10 39 c4F10 + cH4 14
(anthracene) + GHl3 14
+ C&O (phen- 6 + C4H10 14
anthrene) + C5F12 12
c2F6 + CH, 14 + GFl4 12
+ GH6 14 + GH14 14
+ GF8 12 C4H10 + C5Hl'd 16
GH4 + GH60(ethanol) 20 + GH14 16
+ C6H6(benzene) 10 C4HlOO (ethyl ether) 31
+ GOHS(naph- 38 +
c&& (benzene)
thalene) + C6H14 44
+ C14Hlo(phen- 6 CSFl, + CH4 14
anthrene) + GH12 14
C2H5Cl + GH,Cl 43 C5Hl, + GH,F(fluoro- I
+ C4H,CI (l-chloro- 43 benzene)
2,Z-dimethyl- + GH,(benzene) 50
butane) + GHl4 16
GH6 + C2H60(ethanol) 20 c6F14 + cH4 14
+ GH* 16 + C6H14 14
+ C*HlO 16 C6H6 (benzene) f c6F6 (hexa- 13
+ C4Hlo0(butan- 35 fluoro-
1-01) benzene)
+ C4H1,0 (ethyl 35 f C&&,Br 28
ether) (bromo-
+ CSH12 35, 41, benzene)
16 +
C&f,Cl 28
+ C6H14 16 (chloro-
+ c
1 &8 (naph- 30 benzene)
t halene) +
C6HsF 52
+ C14H10 (anthra- 39 (fluoro-
cene) benzene)
C,H,O(ethanol) -I- CBH, 31 + GHl2 13
(benzene) (cyclo-
C6H14 34 hexane)
GF8 + GH8 I6 f C6H14 49, 2
+ C8H18 51

References for Table 4

1 Z. S. Belousova, T. D. Sulimova, and V. M. Prokhorov, Russ. J. Phys. Chem., 1973,47,235.
2 Z . S. Belousova and V. P. Verkhova, Russ. J. Phys. Chem., 1973, 47, 236.
3 A. L. Blancett, K. R. Hall, and F. B. Canfield, Physica, 1970,47, 75.
4 T. K. Bose and R. H. Cole, J. Chem. Phys., 1970,52, 140.
5 T. K. Bose and R. H. Cole, J. Chem. Phys., 1971,54,3829.
6 H. Bradley, jun. and A. D. King, jun., J. Chem. Phys., 1970, 52, 2851.
7 J. Brewer and G. W. Vaughn, J. Chem. Phys., 1969,50,2960.
8 M. A. Byrne, M. R. Jones, and L. A. K. Staveley, Trans. Faraday SOC., 1968,64, 1747.
9 C. R. Coan and A. D. King,jun., J. Amer. Chem. SOC.,1971,93, 1857.
I0 C. R. Coan and A. D. King, jun., J. Chromatog., 1969,44,429.
234 Chemical Thermodynamics
11 A. J. B. Cruickshank, B. W. Gainey, C. P. Hicks, T. M. Letcher, R. W. Moody, and C. L.
Young, Trans. Faraday Soc., 1969,65, 1014.
12 E. M. Dantzler and C. M. Knobler, J . Phys. Cheiir., 1969, 73, 1335.
13 E. M. Dantzler and C. M. Knobler, J. Phys. Chem., 1969, 73, 1602.
14 E. M. Dantzler-Siebert and C. M. Knobler, J. Phys. Chem., 1971, 75, 3863.
15 E. M. Dantzler, C. M.Knobler, and M. L. Windsor, J. Chromatog., 1968, 32, 433.
16 E. M. Dantzler, C. M. Knobler, and M. L. Windsor, J. Phys. Chem., 1968, 72, 676.
17 D. H. Everett, B. W. Gainey, and C. L. Young, Trans. Faraday Soc., 1968, 64, 2667.
18 B. W. Gainey and C. L. Young, Trans. Faraday Soc., 1968, 64, 349.
19 S. K. Gupta and A. D. King, jun., Canad. J. Chem., 1972, 50, 660.
20 S. K. Gupta, R. D. Leslie, and A. D. King, jun., J. Phys. Chem., 1973, 77, 2011.
2Z K. R. Hall and F. B. Canfield, Physica, 1970,47, 219.
22 B. Hemmaplardh and A. D. King, jun., J. Phys. Chem., 1972,76,2170.
23 C. P. Hicks and C. L. Young, Trans. Faraday SOC.,1968, 64, 2676.
24 A. E. Hoover, I. Nagata, T. W. Leland, jun., and R. Kobayashi, J. Chem. Phys., 1968, 48,
2633.
25 F. H. Kate, jun. and R. L. Robinson, jun., J. Chem. Thermodynamics, 1973, 5, 259.
26 F. H. Kate, jun. and R. L. Robinson, jun., J. Chem. Thermodynamics, 1973, 5, 273.
27 B. K. Kaul, A. P. Kudchadker, and D. Devaprabhakara, J.C.S. Faraday I, 1973, 69, 1821.
28 M. P. Khosla, B. S. Mahl, S. L. Chopra, and P. P. Singh, Indian J. Chem., 1972, 10, 1098.
29 F. Khoury and D. B. Robinson, J. Chem. Phys., 1971,55, 834.
30 A. D. King, jun., J. Chem. Phys., 1968,49, 4083.
31 D. H. Knoebel and W. C. Edmister, J. Chem. and Eng. Data, 1968,13, 312.
32 H. B. Lange, jun. and F. P. Stein, J. Chem. and Eng. Data, 1970, 15, 56.
33 R. N. Lichtenthaler and K. Schiifer, Ber. BunserigesellschaftPhys. Chem., 1969, 73, 42.
34 N. P. Markuzin and V. N. Bardin, Vestnik Leningrad. Univ. (Fiz.Khim.), 1973, 77.
35 R. Massoudi and A. D. King, jun., J. Phys. Cheni., 1973, 77, 2016.
36 H. G . McMath, jun. and W. C. Edmister, Amer. Inst. Chem. Engineers J., 1969, 15, 370.
37 R. C. Miller, A. J. Kidnay, and M. J. Hiza, J. Chem. Thermodynamics, 1972, 4, 807.
38 G. C. Najour and A. D. King, jun., J. Chem. Phys., 1966,45, 1915.
39 G. C. Najour and A. D. King, jun., J. Chem. Phys., 1970, 52, 5206.
40 R. D. Nelson and R. H. Cole, J. Chem. Phys., 1971, 54,4033.
41 R. L. Pecsok and M. L. Windsor, Analyt. Chem., 1968,40, 1238.
42 J. A. Provine and F. B. Canfield, Physica, 1971, 52, 79.
43 M. RBtzsch, Z . phys. Chem. (Leipzig), 1968,238,321.
44 M. Ratzsch and H. Freydank, J . Chem. Thermodynamics, 1971, 3, 861.
45 M. RBtzsch and E. Rasenberger, Z. Chem., 1968, 8, 156.
46 H. S. Rosenberg and W. B. Kay, J. Phys. Chem., 1974, 78, 186.
47 M. Vidergauz and V. Semkin, J. Chromatog., 1971, 58,95.
48 C. L. Young, Trans. Faraday Soc., 1968, 64, 1537.
49 Sh. D. Zaalishvili, Z. S. Belousova, and V. P. Verkhova, Russ. J. Phys. Chem., 1971,45, 149.
50 Sh. D. Zaalishvili, Z. S. Belousova, and V. P. Verkhova, Russ. J. Phys. Chem., 1971, 45, 894.
51 Sh. D. Zaalishvili, 2.S. Belousova, and V. P. Verkhova, Russ. J. Phys. Chem., 1971,45,902.
52 Sh. D. Zaalishvili, Z. S. Belousova, and V. P. Verkhova, Russ. J. Phys. Chem., 1972,46,291.

systems that have been studied since then. The listing is reasonably complete
up to 1973, and the early 1974 literature has been covered in part. No references
to these or other sources not generally available have been included.
The order of listing is similar to that used by R 0 ~ 1 i n s o n . l ~
It~is subject to
these conventions. (1) Inorganic components precede organic. (2) Carbon is
placed first in all organic substances and is followed immediately by hydrogen
(if present). (3) All substances with one carbon atom, etc. precede those with
two. (4) References to interaction third viriaf coefficients are indicated with an
asterisk.
I wish to thank Professor R. L. Scott for his many suggestions and comments
and Dr N. Pasco for his critical reading of the manuscript.
J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, Butterworths, London, 1969, 2nd edn.
p. 367.
Volumetric Properties of Gaseous Mixtures 235
Note added in proof. Table 5 is a summary of interaction virial coefficient measure-
ments reported since the manuscript was first submitted. A review of the
chromatographic method has appeared 144 and a compilation of B12’sfor hydro-

Table 5 Literature Survey

System Rejs. System Refs.
Ar F6s 2,17 + C6H11 (truns-hexa- 13
+
Si(CH3), 2 1,4-diene)
+
CHF3 9 + CBHll (trans-hexa- 13
+
CHSF 9 1,3-diene)
+
CH, 2,ll + C6H12 (trans- 13
+
co,* 7 hex-3-ene)
+ C2H6 11,18 + GH,, (trans-hex- 13
+ GH, 11, I8 2-ene)
+ C4H10 18 + CeH12 (hex-l-ene) 13
+ C5H12 (tetra- 1,2 + C6H14 13
methylmethane) + ClOHl4 10
+ CQH20(2,294- 10 + ClOH22 10
trimethylhexane) + C12H26 10
+ ClOH22 10 Si(CH3)4 + CH, 2
+C10H14 (t-butyl- I0 + C5H12 (tetra- 2
benzene) met hylmethane)
+ C12H26 10 CH, + C5H12(tetra- 192
F6S f Si(CH,), 2 methylmethane)
+Kr I7 + C9H2o (2,2,4-tri- 10
+CH4 2 methy 1hexane)
+C5H12 (tetra- 2 + GOHI4 (t-butyl- 10
methylmethane) benzene)
He +CO, 12,20, + ClOH22 10
21 + C12H26 10
+ C5H12 1 c02 f C3H8 8
+
N2 N 2 0 22 CClF,+ CC12F2 8
+ CH40 (methanol) I6 + CClSF 8
+C2H,0 (ethanol) 16 + CHCIF, 8
+C,H,O (propan- 16 + CHF3 8
1-01) CC12F2+ CC1,F 8
+ C4H100 16 + CHClF, 8
(butan- 1-01) + CHF3 8
+ C,H, (cis-penta- 13 CC1,F + CHCIF, 8
1,3-diene) + CHF3 8
+ C5H1, (trans- 13 CHClF2 + CHFS 8
pent-2-ene) C,H,F (fluorobenzene) 6
+ C5Hl2 13 + C,H8 (toluene)
+ C5H12(tetra- 1 + C7H16 5
met hylmet hane) + C8H18 19
+ C6H10 (cis-hexa- 13 C,H,. (benzene)
4
1,3,5-triene)
.
+
CsHla (cyclohexane)
+ CsHll (truns,truns- 13 +
C,H8 (toluene) 3
hexa-2,4-diene) C7H7F(o-fluorotoluene) 14
+ C6H11 (truns-hexa- 13 +
C,H, (toluene)
1,5-diene) C7H7F (p-fluorotoluene) 15
+
C,H, (toluene)

144 R. J. Laub and R. L. Pecsok, J. Chromatog., 1974,98, 511.

236 Chemical Thermodynamics
I G. L. Baughman, S. P. Westhoff, S. Dinar, D. D. Duston, and A. J. Kidnay, J. Chem.
Thermodynamics, 1975,7, 875.
2 5. Bellm, W. Reineke, K. Schiifer, and B. Schramm, Ber. Eunsengesellschaft Phys. Chem.,
1974, 78, 282.
3 Z. S . Belousova and T. D. Sulimova, Russ. J. Phys. Chern., 1974,48, 254.
4 Z. S. Belousova and T. D. Sulimova, Russ. J. Phys. Chem., 1976,50,292.
5 Z . S. Belousova and T. D. Sulimova, Russ. J. Phys. Chem., 1976,50, 585.
6 Z. S. Belousova and T. D. Sulimova, 2hur.fiz. Khim., 1976,50,2122.
7 K. Bier, W. Schoenmann, and R. Waibel, Chem.-Ing.-Tech., 1975, 47, 349.
8 J. Bougard and R. Jadot, J. Chim. phys., 1976, 73,415.
9 T. G. Copeland and R. H. Cole, J. Chem. Phys., 1976, 64, 1747.
10 S. G. D’Avila, B. K.Kaul and J. M.Prausnitz, J. Chem. and Eng. Data, 1976,21, 488.
11 R. Hahn, K. Schgfer and B. Schramm, Ber. Bunsengesellschaft Phys. Chem., 1974,78,287.
12 L. R. Linshits, I. B. Rodkina, and D. S. Tsiklis, Russ. J. Phys. Chem., 1975,49, 1258.
13 T. M. Letcher and F. Marsicano, J. Chem. Thermodynamics, 1974,6,501.
14 L. V. Mozhginskaya and L. E. Kolysko, Russ. J. Phys. Chem., 1975,49, 860.
15 L. V. Mozhginskaya and L. E. Kolysko, Russ. J. Phys. Chem., 1975,49,982.
16 P. Neogi and A. P. Kudchadker, J.C.S. Faraday I, 1977,73, 385.
17 J. Santafe, 5. S. Urieta and C. G. Losa, Chem. Phys., 1976, 18, 341.
18 K. Schiifer, B. Schramm, and J. S. Urieta Navarro, Z. phys. Chem. (Frankfurt), 1974,93, 203.
19 T. D. Sulimova and Z. S. Belousova, Russ. J. Phys. Chem., 1974, 48, 620.
20 D. S. Tsiklis, L. R. Linshits, and I. V. Rodkina, Russ. J . Phys. Chem., 1974,48,906.
21 D. S. Tsiklis, L. R. Linshits, and I. V. Rodkina, Russ. J. Phys. Chem., 1974, 48, 908.
21 N. P. Yakimenko, G. M. Glukh, M. B. Iometov, L. I. Polishchukova, V. E. Dobrovol’skaya,
and A. S. Fik, Zhur. fiz. Khim., 1976, 50, 1909.

carbon mixtures 145 has become available. Experimental studies of interaction

third virial coefficients remain scarce. Caligaris and Henderson 148 have
calculated third virial coefficients for Ar +
Kr and Kr + Xe and included the
effects of three-body interactions. The literature on enthalpies of mixing of gases
has grown significantly. Wormald and co-workers have recently reported H E
measurements on: CH, + Ar, + M,, and + 14*-, N2 + Ar1*7*149
N21479 and
+ Hz147. Other investigators have studied N2 + 0 2 1 5 0 and N, + CO,l5l.
New combination rules have been proposed by Kong 153 that are derived

by requiring that the attractive portion of the potential be constructed from the
geometric mean of the attractive terms for the pure components while the unlike
repulsive interaction is that suggested by Smith :15*
$12 (rep., t) = (#I1 (rep., 2 4 + $A2 (rep., 2 h ) } / 2 ,
where r = rl +
r2 and id& (rep-)/dr]r-zrl = (d~z~(rep.)/dr},,a,,.
The state of knowledge of unlike interactions for the rare-gas pairs has been
reviewed by Bobetic and Barker 155 who calculated the vibrational levels of
heteronuclear van der Waals dimers. Their choices for ‘best’ potentials are those
116 A. Pompe and T. H. Spurling, ‘Virial Coefficients for Mixtures of Gaseous Hydrocarbons’,
C.S.I.R.O. Austral. Div. Appl. Organic Tech., Paper No. 3, 1976.
146 R. E. Caligaris and D. Henderson, MoZ. Phys., 1975,30, 1853.
14’ C. J. Worrnald, K. L. Lewis, and S. Mosedale, J. Chem. Thermodynamics, 1977,9, 27.
148 K. L . Lewis, S. E. Mosedale, and C. J. Wormald, J. Chem. Thermodynamics, 1977,9, 121.
14* K. L. Lewis, S. E. Mosedale, and C. J. Wormald, J. Chem. Thermodynamics, 1977, 9,221.
lS0 L. B. Ba, V. P. S. Nain, S. C. Kaliagune, and R. S. Ramalho, J. Chem. and Eng. Data, 1977,
22, 171.
161 V. V. Altunin, V. F. Bondarenko, and D. 0. Kuznetsov, Teplofiz. Vys. Temp., 1973, 11, 533;
1974,12, 513.
161 C. L. Kong, J. Chem. Phys., 1973,59, 2464.
15a C. L. Kong and M. P. Chakrabarty, J. Phys. Chem., 1973, 77, 2668.

lS4 F. T. Smith, Phys. Rev. A, 1972, 5, 1708.

16( M. V. Bobetic and J. A. Barker, J. Chem. Phys., 1976,64,2367.
Volumetric Properties of Gaseous Mixtures 237
determined from a variety of properties such as molecular-beam scattering cross-
sections, second virial coefficients,and transport coefficients. Bobetic and Barker
+
note that the potential parameters for Ar Kr, Ar + Xe, and Kr + Xe deter-
mined by Lin and Robinson from virial coefficients alone60 agree well with the
‘best’ The agreement is much poorer for other rare-gas pairs. Chen
et aZ.lS7found that no single reduced potential could fit the unlike interactions
involving He. Their estimates of the well depths for He + Ar, + Kr, and
+ X are virtually identical and are consistently shallower than the Lin and
Robinson values.
Well depths for Ne + Ar, + Kr, and +
Xe are also found to be closely
similar,168but in this case deeper than those obtained from virial coefficients
alone. On the other hand, the He + Ar scattering cross sections of Smith et
aZ.,lSs when combined with virial coefficients and transport properties, lead to a
well depth 30 per cent deeper than that obtained by Chen et aZ.lS7 The larger
clZis in excellent agreement with the Lin and Robinson estimate. More recently,
Derderian and Steele I6O have found that their measurements of the adsorption
of He on Xe are consistent with a value of el, in good agreement with that derived
by Lin and Robinson. Clearly even for optimal systems the determination of
unlike interactions remains difficult.
lS6 J. K. Lee, D. Henderson, and J. A. Barker, Mol. Phys., 1975, 29, 429.
lS7 C. H. Chen, P. E. Siska, and Y. T. Lee, J . Chem. Phys., 1973,59, 601.
lSa C. Y . Ng, Y . T. Lee, and J. A. Barker, J. Chem. Phys., 1974,61, 1996.
lS9 K. M. Smith, A. M. Rullis, G . Scoles, R. A. Aziz, and G . Duquette, J. Chem. Phys., 1975,
63, 2250.
160 E. J. Derderian and W. A. Steele, J. Chem. Phys., 1977, 66, 2831.
8
Critical Exponents for Binary Fluid Mixtures
BY R. L. SCOTT

1 Introduction
It has been obvious for a long time that the traditional ‘mean-field’ treatments
of gas-liquid equilibria (e.g. van der Waals-like theories *) and of magnetism
(e.g. the Weiss theory2) fail to describe adequately the details of thermodynamic
behaviour in the critical region. It is now well known that the difficulty lies in
the assumption that the appropriate thermodynamic function (e.g. the molar
Helmholtz free energy A , for the one-component fluid) is analytic, i.e. can be
expanded in a Taylor series around the critical point (e.g. for the fluid, in powers
of the difference in temperature T - TC and the difference in molar volume
Vm - VL). Although the arguments were originally developed for particular
models the same predictions arise quite generally s-6 and invariably disagree
with the experimental observations.
In particular the coexistence curve (temperature T us. density p) is not
parabolic as predicted but more nearly cubic, the critical isotherm (pressure
p us. p ) is of distinctly higher order than cubic, and the molar heat capacity at
constant volume C v , m diverges at the critical point instead of remaining finite.
The modern treatment of critical points summarizes these and other properties
by assigning them ‘critical exponents’ (a,p, y, 8, etc.) which describe how they
converge to zero or diverge to infinity as one approaches the critical point.
Formally we may define the exponent f governing the limiting behaviour of a
property Y as it changes with, say, T - TC,along a specified path, by the equation:

If modern phenomenological ‘theories’ are valid, the exponents associated with

different properties are interrelated by a series of equations (the ‘scaling laws’)
which will be detailed in Section 4.

l J. D. van der Waals, ‘On the Continuity of the Gaseous and Liquid States’, Ph.D. Thesis,
Leiden, 1873; see also R. L. Scott, in ‘Physical Chemistry’, ed. by H. Eyring, D. Henderson,
and W. Jost, Academic Press, New York and London, 1971, vol. 8A, ch. 1, pp. 66-80.
P. Weiss, J. phys. (France), 1907,6,667; see also C. Kittel, ‘Introduction to Solid State Physics’,
3rd edn., Wiley, New York, 1966, ch. 15.
J. D. van der Waals, Z. phys. Chem., 1894, 13,657.
J. J. van Laar, Proc., k. ned. Akad. Werenschap., 1912, 14, 1091.
L. D. Landau, Physik. 2. Sowjetunion, 1937, 11, 26, 545; see also L. D. Landau and E. D.
Lifshitz, ‘Statistical Physics’, 2nd edn., Pergamon Press, Oxford, 1969, pp. 263-270.
R. H. Fowler and E. A. Guggenheim, ‘Statistical Thermodynamics’, Cambridge University
Press, London, 1939, pp. 314-318.
238
 Critical Exponents for Binary Fluid Mixtures 239
Less widely appreciated is the fact that substantially the same departures from
the behaviour predicted by mean-field theories occur in fluid mixtures in the
region of the critical solution point. The same failures of analytic theories are
found and similar critical exponents are deduced from experiment. It is the
purpose of this chapter to review the thermodynamic principles underlying
critical phenomena, to emphasize the analogies with pure-fluid gas-liquid
equilibrium, to outline some of the difficulties encountered when one tries to
determine critical exponents precisely, and to summarize the presently available
experimental results from binary fluid mixtures. (For a review of the related
situation in magnetism and in pure fluids, see Stanley and Levelt Sengers.8)
Certain questions will arise:
1. What are the exact values of the critical exponents for binary mixtures?
2. Are they the same for all mixtures (the principle of ‘universality’)?
3. Are they the same as the best values of the corresponding exponents for a
one-component fluid?
4. Do the exponents satisfy exactly the relations deduced from the ‘scaling’
hypothesis ?
5. Are the exponents the same as those deduced theoretically for the Ising
model in three dimensions?
6. Does the lack of complete symmetry in the binary mixture lead to a failure
of the ‘law of the rectilinear diameter’ in the critical region?
These and other questions will be considered in the pages that follow. How-
ever, the present state of our understanding of fluids and fluid mixtures-from
the standpoint of both theory and experiment-is such that no completely
definite answer can be given to any of them.

2 Critical Phenomena in Binary Mixtures

Critical phenomena in binary fluid mixtures are substantially analogous to those
in one-component fluids when proper allowance is made for the increase in the
number of independent thermodynamic variables by one. Thus any extensive
property of a one-component system (e.g. the energy U ) can be expressed as a
function of three variables, for example the entropy S, the volume V, and the
amount of substance n; but the binary mixture requires four (e.g. S, V, n,, and
Its). For most purposes the size of the system is irrelevant and one reduces the
number of variables by one by using molar quantities Um,Sm,Vm; for the binary
mixture three independent variables (e.g. Sm,Vm, and a composition variable
such as the mole fraction x of the second component) then suffice.
All the thermodynamic properties then derive from the refations embodied in
the differential equation for dUm :
dUm = TdSm - PdVm - d dx, (2)
where T is the thermodynamic temperature, p the pressure, and d = p1 - p2,
the difference of the chemical potentials of the two components.
’I H. E, Stanley, ‘Introduction to Phase Transitions and Critical Phenomena’, Oxford University
Press, 1971.
J. M. H. Levelt Sengers, Physica, 1974, 73, 73. This article includes an excellent history of the
experimental failure of the ‘classical’ theory, shown as early as 1900 by Verschaffelt.
240 Chemical Thermodynamics
In this molar formulation, all the variables are intensive but they are all
generalized ‘densities’ in the sensea that, like the density itself, the values differ
in two coexisting phases. Unlike these density variables, a potential or ‘field’
variable (e.g. T,p, A) must have the same value in every phase at equilibrium.
Equation (2) can be transformed into the generalized Gibbs-Duhem equation
in which the independent variables are fields:
SmdT - Vmdp + (1 - x)dp, + xdp2 = 0, (3)
or, selecting pl as the (dependent) variable of interest:
4%= - SmdT + Vmdp + x d A . (4)
For other thermodynamic quantities the appropriate variables are a mixture
of ‘densities’ and ‘fields’. Thus the molar enthalpy H m = Urn + pV, =
Hm(Sm,p, x), and the molar Helmholtz free energy A m = Um - TS, =
Am(T,Vm, x), while the molar Gibbs free energy G , = Hm - TSm = Gm(T,p, x).
For binary mixtures G m = (1 - x ) p l +
xp2 is especially important and its
general variation can be written
dG, = - SmdT + Ymdp - d dx. (5)
Quite generally any critical point may be defined as a point where the differences
in properties between two coexisting phases all vanish. In a binary mixture this
occurs lo whenever the second and third derivatives (at least) of the molar Gibbs
free energy Gm with respect to the mole fraction x are simultaneously zero:t
( a * ~ ~ p x 2=) -(adlax),,.
~,~ = 0, (64
(a3Gm/aX3)T,, = - (a2d/ax2)T,, = 0. (6b)
[Actually, if one abandons the notion that analytic equations can be extended
into ‘metastable’ one-phase regions, any point in a one-phase region which
satisfies equation (6a) is a critical point and equation (6b), while also satisfied
at such a point, is superfluous.]
Equations (6a, b) for the binary mixture are analogues of the equations that
define the gas-liquid critical point in a pure fluid:
(a2Am/aV& = -(ap/a&)T = 0, (7a)
(a3Am/aV3T = -(a2p/wg)T = 0. (7b)
An important difference between equations (6a, b) and (7a, b) is the additional
variable needed for a binary system. While equations (7a, b) define a single
critical point in the two-dimensional T, Vm space (or T,p space or T, p space) of
a one-component system, equations (6a, b) define a critical line in the three-
dimensional space (e.g. T, p , x or T,p, d) appropriate for a binary mixture. Thus
R. B. Griffiths and J. C. Wheeler, Phys. Rev., 1970, A2, 1047.
10 J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, Buttenvorths, London, 1969, 2nd edn.,
ch. 5.
t A perhaps more familiar definition of a critical point in binary mixtures involves the chemicaI
potentials: (&/ax)T = 0 and (a2pz/i3x2)T = 0, or the analogous pair of equations for
component 1. Applickions of the Gibbs-Duhhm equation (3) shows these to be exactly equivalent
to equations (6a, b).
Critical Exponents for Binary Fluid Mixtures 241
there is a critical solution temperature TCfor each fixed pressure, and a critical
solution pressure p c for each fixed temperature.
Figure 1 shows a schematic diagram l1 of the coexistence surface for a binary
mixture in T , p , x space. Any constant-pressure section of this diagram yields a

Figure 1 Schematic three-dimensional T,p , x diagram showing the coexistence surface

(shaded), constant pressure T, x sections and one constant temperature p , x section.
The critical line is shown dashed (- - - -)
(Reproduced by permission from Ber. Bunsengesellschaft Fhys. Chem., 1972, 76, 296)

T, x curve entirely analogous to the T, Vm (or T, p) curve for a pure fluid. A

constant-temperature cross-section yields a similar p , x coexistence curve. One
can take these constant-pressure or -temperature sections including two-phase
regions without complication because these variables are ‘fields’, variables that
have the same values in the conjugate phases on opposite sides of the coexistence
surface. In contrast a section at constant x would have no simple interpretation
because the curve representing the intersection of the constant-x plane with the
coexistence surface represents points on tie-lines that do not lie in the plane and
that terminate at a second phase with a different value of x.
The alternative T , p , d diagram (Figure 2), emphasized by Griffiths and
Wheeler,O poses no such problems for all co-ordinates are ‘fields’. There are no
tie-lines for the two-phase region has shrunk to a single surface, since A (like T
and p ) is the same for both conjugate phases. The coexistence surface is bounded
by the critical line. (The analogous one-component diagram - the two-
dimensional T,p diagram - is more familiar. The coexistence line is the ‘vapour-

R. L. Scott, Ber. Bunsegesellsch?ift Phys. Chem., 1972, 76, 296.

242 Chemical Thermodynamics

Figure 2 Schematic three-dimensional T, p , A diagram corresponding roughly to the

system shown in Figitre 1. The two-phase region has shrunk to a single coexistence
surface (shaded). Constant temperature or pressure sections yield single coexistence
lines (as worild a constant A section - not shown). The critical line which bounds the
coexistence surface is shown dashed (- - - -)

pressure curve’, which ends at (is bounded by) the critical point.) A constant-
pressure section of the T , p , d diagram yields a T, A coexistence line ending in a
critical point T C Ac.
,
We see that the formalism appropriate to a one-component system with its
two variabIes is equivalent to that for a two-component system with three
variables if to the original two variables one adds a ‘field’ that is held constant.
More generally the analogous functions for one-, two-, and c-component fluids

Table 1 Corresponding thermodynamic functions

‘Density’
uariables One component
2 UdSrn, ’V
 Critical Exponents for Binary Fluid Mixtures 243
are those with the same number of independent variables that are densities; the
number of ‘field’ variables is irrelevant because any ‘field’ can be held constant
with the same value in coexisting phases.
Table 1 lists some of the corresponding functions for one- and two-component
systems. The correspondences are not unique, for as one increases the number of
variables there are alternative choices, some of which are shown in Table 1.
Here Cm is the molar heat capacity T(aSm/aT) with the designated quantities
held constant, K is the compressibility -(l/Ym)(aYm/8p), a is the expansivity
(l/Vm)(aYm/aT), and Q is -(ax@).
Extensions to fluids of three or more components are straightforward but more
complicated and cumbersome. Details may be found in recent

3 Non-analyticity. The Critical Exponents

If Gm(T,p , x), the appropriate thermodynamic function for a binary mixture, is
analytic, the deductions about the behaviour of the various thermodynamic
properties are entirely analogous to those for the one-component fluid. The
same conclusions arise from any general Taylor series expansion in which all
the coefficients are non-zero except those two required to define the critical point
[equations (6a, b)]. In particular, the coexistence curve (T us. x at constant p )
should be parabolic, the critical isotherm (p us. x at constant T and p ) should
be cubic, and the molar heat capacity Cg,z,m should be everywhere finite.
However, just as is the case for many other kinds of critical phenomena
(e.g. the one-component fluid, magnetism, order-disorder transitions in solids,
etc.) such predictions do not agree either with the results of careful experimental
measurements or with simple theoretical models that can be treated nearly
exactly. The coexistence curve is more nearly cubic than parabolic, the critical
isotherm is of distinctly higher order than cubic, and the heat capacity Cp,z,m
diverges at the critical point.
1. At constant pressure, the difference in the mole fractions (x” - x‘) of the
conjugate phases varies as (TC - T)@as the critical solution temperature is
approached.
( p constant): X“ - x‘ = B[(TC- T)/Tc]@. (8)
[It is customary to cast equations like the above in ‘reduced form’; hence the
variable is not (T - Tc),but rather (T - TC)/To,which we shall later symbolize
by t.] The exponent /3 is approximately 1/3 and almost certainly lies between
0.31 and 0.36, well below the ‘classical’ (i.e. analytic) 1/2. The other property
of the coexistence curve, the midpoint of the two-phase region or ‘diameter’
(x’ + x“)/2 describes a very nearly straight line, but recent theoretical studies l2* l3
suggest a small curvature close to TCas described by the equation:
(p constant): (x’ + x”)2 = xc + C [ ( P - T)/TCl1”,(9)
where a is the exponent for the heat capacity C9rz,m,which (see below) is
believed to be small but positive, between 0.00 and 0.15. If so, there is a weak
divergence in the derivative of the diameter at the critical point.
la B. Widom and J. S . Rowlinson, J. Chern. Phys., 1970, 52, 1670.
13 N. D. Mermin and J. J. Rehr, Phys. Rev. Letters, 1971,26, 1155; Phys. Rev., 1971, A 4,2408.
244 Chemical Thermodynamics
2. Along a line of constant temperature T = TCand p = p c (i.e. the critical
isotherm-isobar), A - A c (and so also pl - p: or pa - pi taken separately)
approaches the critical composition as ( x - xc)&
(T z= T C , p = p"): Id - Acl = l@,l - p3 - @,z - /A$) I = D ( x - PI8,
(10)
where 6 lies between 4 and 5, much higher than the classical 3. An alternative
way of stating this result is that the quantity Q T , P = - (a~/ad)~,,,a binary
mixture analogue of the isothermal compressibility KT of a one-component fluid,
diverges as I A - dc 1" where E = (6 - I)/& i.e. 0.75 to 0.80 rather than the
'classical' 2/3. One expects that along a p a t h t T = TC, A = AD, K ~ will~ , ~
diverge as I p - p c 1" and along a path p = p c, d = dc,Cp,A,m and a , , ~will
diverge as I T - TC1" and that all these E'S will be nearly, if not exactly, the
same :
(T = TC,p = p"): (ah0 Cp,A,m, KT A, C X ~ , Adiverge
Qp,p ) as I A - A" I-?',
(T = TC,A = A"): (also Cp,A,m,
KT,A a p , ~diverge
) as I p - p c I",
( p = p c , A = Ac): Cp,A,m, a 9 , A (also K T , A , Q T , v ) diverge as I T - PI4.
(1 1)
3. In the one-phase region along a line of constant pressure p = p c and
constant mole fraction x = xc, (a2Gm/aX2)F,p = -(ad/a~)~,, = l/QT,@, which
must be zero at the critical point, is proportional to I T - T C I Y + (i.e. Q T , p
diverges along this path as I T - TCI++) where y+ lies between 1.2 and 1.4,
higher than the 'classical' value of 1. Similar behaviour will be observed along
other paths where one 'density' is held constant; thus:$
( p = pc, X = X c ) : Cp,A,m, K T , A , Q T , 9 , a u , diverge
,~ as I T - TCI-Y',
(T = TC,x = x"): C p , A , m , K T , A , Q T , v , a p ,diverge
~ as I p - p c I-Y',
(T = TC,Vm = V L ) : C D . ~ , KF>A,
m , QTIV, a p ,diverge ~ as I A - Ac I-)"+, etc.
(12)
A corollary of the proportionality of (a2Gm/ax2)T,vto (T - Tc)Y+ is the
conclusion that (a2Vm/ax2),, and (a2S,/ax~)T,,approach the critical point as
(T - TC)Y+-l;i.e. since y+ > 1, they must go to

l4 J. S. Rowlinson, in 'Critical Phenomena', ed. M. S. Green and J. V. Sengers, U.S. Natl. Bur.
Standards Misc. Publ., No. 273, 1965, p. 9.
R. D. Dunlap and S. D. Furrow, J. Phys. Chem., 1966,70, 1331.
If) D. B. Myers, R. A. Smith, J. Katz, and R. L. Scott, J . Phys. Chem., 1966,70, 3341.
l7 W. J. Gaw and R. L. Scott, J. Chem. Thermodynamics, 1971, 3, 335.
t The 'natural' thing to do is to consider how Q T , varies
~ along a path T = T C , p= pc, and
how KT,A varies along a path T = TC,A = A c ; however, the same property may be considered
along different paths and the divergences may be different. Griffiths and Wheeler * have stressed
the differences between a path that is asymptotically parallel to the coexistence curve (e.g.
constant mole fraction 9)and one that is not (e.g. constant pressure or constant temperature).
A third kind of path is one 'asymptotically parallel to the critical surface' (the critical line in the
binary mixture); the critical isotherm at a maximum or minimum temperature on a critical line
is an example.
Note that these are all 'unnatural' paths (in the sense of the previous footnote) for the
quantities in question. The actual differentiation is not along a direction corresponding to the
path.
Critical Exponents for Binary Fluid Mixtures 245
Along the coexistence curve ‘below’ the critical temperature7 relations similar
to equations (12) apply. QTy9and its counterparts should diverge as I TC - T 19-
where y- is expected to be approximately, if not exactly, the same as y+.
4. The molar heat capacity Cp,z,m,evaluated in the one-phase region along
the ‘natural’ path of constant pressure p = p c and constant mole fraction xc is
expected to diverge weakly as 1 T - TC.1’“ Along the same path or along other
paths in which one ‘density’ and one ‘field’ are kept constant at their critical
values, the weak divergences of and should be governed by similar
exponents :.I:
( p = pc, x = Y): Cp,z,m,K~,,, ao,,diverge as I T - TCI++,
(T = To, x = x“): Cp,z,m, K T , ~ a,,,
, diverge as I p - p c Id+,
etc. (13)
Along the coexistence curve, ‘below’ the critical solution temperature and at
constant pressure p = p c the molar heat capacity CDtztm (and the analogous
quantities) should diverge as I Tc - TI-”.- while in the two-phase region with
the overall mole fraction constant at xC it should diverge as I Tc - TI-”;. As
in the case of y+ and y-, a+,a:, and a; should be about the same.
The difficulty in the determination of a, to which we shall return later, is that
it is so near to zero that it is not easy to distinguish between a weak divergence
with 0 < a < 0.1 or a logarithmic singularity (a = 0) as the system crosses
from the two-phase to the one-phase region. What certainly is not observed as
the system crosses from the two-phase to the one-phase region is the simple
finite jump predicted by the analytic treatment. However, the plot of C p , z , m
vs. T [or the corresponding quantities in the alternate forms of equation (13)]
is far from symmetrical around T C ;the divergences (with exponents a+ above
and %- below Tc) are superimposed upon a finite discontinuity.
5. Properties which are derivatives in which two ‘densities’ are held constant,
: , do not diverge at all, but come to a finite maximum at a
e.g. CV,z,m, ~ g , ~etc.,
sharp cusp. Thus
(V, = VL,x = P,T > TC): Cv,,,m
(Vm = v;, x = xc, T < TC): cv,+.m

where C&,m is the finite maximum at the critical point. It is the change in sign
of the exponent which keeps Cv,lF,m finite;§ only its derivative (aCv,z,m/aT),,,
diverges.
We see that, in general, we are interested in several types of derivatives, which
can be classified according to the number of ‘densities‘ held constant. Thus, for
t. For simplicity of discussion we assume the more common situation in which the two-phase
region lies below the critical solution temperature T!. In the binary fluid mixture this will not
always be the case; there are numerous systems with lower critical solution temperatures.
$ The exponents a+,cry, and 01; must not be confused with the coefficient of thermal expansion
a p 2.
The change in sign from - ar to a arises because of the change from Cp,z,m to Cv m. The
denominator (1 - a)in the exponent is an example of ‘renormalization’; it arises becausd C V , ~ , ~
is plotted as a function of T along a path of constant volume. Along the ‘unnatural’ path of
constant pressure, C V , ~would
, ~ have an unrenormalized exponent +a.
246 Chemical Thermodynamics
binary fluid mixtures, those involving no ‘densities’, i.e. C,,alm, KT,A, QT,p,
a P , ~correspond
, to CP,M,K ~ and
, a p for the one-component pure fluid, while
those for which only one ‘density’ is held constant, i.e. C,,%,m, K T , ~ ,( L ~ ,etc.,
~,

correspond to Cp,m,K ~ etc.,

, for the pure fluid.
Figure 3 summarizes on a T, x diagram the various quantities and the various
paths, and their associated exponents.

I x-xc

Figure 3 Diagram illustrating the various directions of approach to the critical point and
the exponents appropriate to the carious quantities. Negative exponents (e.g. - a) indicate
the quantity becomes infinite at the critical point; positive exponents (e.g. p) indicate
the quantity vanishes at the critical point. The coexistence curve and the diameter
correspond to 16 = 0.35; 1 - a = 0.90; note that the curvature of the diameter is
barely noticeable. ([ is the long-range correlation length; it and its exponent v are
discussed in Section 6 )

Exceptions to these general rules about behaviour in the critical region occur
at special points on a binary mixture critical lS examples are extrema in
pressure or temperature, points where azeotrope lines join critical lines, or double
points (intersections of two critical lines).lB Moreover, in the special higher-
order critical points (‘tricritical points’) found in systems with a greater number
of variables (three- and four-component fluid mixtures; SHe + *He), different
exponents may be found.20*

4 Inequalities, Homogeneous Functions, and Scaling

With the failure of any analytic theory of the critical point, no unequivocal
conclusions can be drawn about the precise values of the exponents a, p, y, 8
for fluids and fluid mixtures, since no statistical-mechanical theory of fluids is

l8 W.F. Saam, Phys. Rev., 1970, A4, 1461.

G . W. Mulholland and 3. J. Rehr, J. Chem. Phys., 1974, 60, 1297.
eo R. B. Griffiths, Phys. Rea. Letters, 1970,24, 715.
21 R. B. Griffiths, J. Chem. Phys., 1974, 60, 195.
Critical Exponents for Binary Fluid Mixtures 247
reliable enough to treat the critical region accurately. The only plausible physical
model for which reasonably accurate exponents are known is the highly over-
simplified three-dimensional Ising lattice.
However, some inequalities can be deduced from the general requirements of
thermodynamic ~ t a b i l i t y , ~of~which
- ~ ~ the most important are
a, + 213 + y- z 2,
a; + /3(S + 1) z 2,
y- - p(6 - 1) z 0,
a, + y-(S + l)/(S - 1) 2 2.
Table 2 lists the various exponents, with their classical values, the values
calculated from the theory of the three-dimensional Ising and the
presently accepted ‘best’ values for one-component fluids.*,26

Table 2 Critical exponents a

‘Classical’ Ising model
analytic (3 dimensions) Fluids

k
Y+
1 /2
3
1
0.312 & 0.004
5.1
1.2500 0.0002z
I- 0.1
0.350
4.4
1.22
&
&
&
0.015
0.4
0.05
Y- 1 1.27 rt 0.04 1.2 & 0.2
a+ O (finite jump) 0.125 L- 0.002 0.05 & 0.05
0 (finite jump) 0.12 & 0.06 ?
a;- 0 (finite jump) 0.12 L- 0.06 0.08 0.08
8 0
The exponent 9 applies to the limiting curvature of the vapour-pressure line (a2p/aT2)o,
which should diverge as (T, - T)-e.Analogous quantities in the binary fluid are (azSm/aTax),
and (a Vm/apax)T,,, about which no experimental information is available. The scaling hypothes;:
leads to the conclusion that B = a.
The symmetry of the Ising model requires that a; = a;.

No one has yet succeeded in proving that these inequalities must in fact be
equalities, or that y+ = y-, or even that the limiting exponent y for
CP,Armcc I T - TCIY must be the same as that for KT,A M: I p - p c IY. How-
ever, all these equalities are features of the ‘classical’analytic treatment, and most,
if not all, appear to be satisfied within the small uncertainties of the exponents
calculated for the three-dimensional Ising model.
If one assumes that the appropriate thermodynamic function [e.g. G,(T, p , x )
for the binary mixture], or more accurately the singular part thereof, is a homo-
geneous function of the appropriate variables (e.g. T - TC or p - p c and
x - xc), one can show 27 that inequalities (15) to (18) become equalities and that

aa G . S. Rushbrooke, J. Chem. Phys., 1963,39, 842; 1965,43, 3439.

28 M. E. Fisher, J. Math. Phys., 1964, 5, 944.
24 R. B. Griffiths,J. Chem. Phys., 1965, 43, 1958; Phys. Rev. Letters, 1965, 14, 623.
26 J. Stephenson, in ‘Physical Chemistry’, ed. by H. Eyring, D. Henderson, and W. Jost, Academic
Press, New York and London, 1971, Vol. 8B, ch. 10, p. 742.
28 J. S. Rowlinson, Ber. Bunsengesellschaft Phys. Chem., 1972, 76,281.
B. Widom, J. Chem. Phys., 1965, 43, 3898.
9
248 Chemical Thermodynamics
y+ and y- are the same, as are all the a’s. In other words:
a+ = = %- = 2 - cscs f 11, (19)
yf = y- = p(s - l), (20)
and two independent exponents fi and 6 suffice.
A phenomenological argument from a scaled mode1,28~29 in which the range
of statistical correlations greatly exceeds the range of real intermolecular inter-
actions, leads to identical conclusions. This permits one to write a ‘scaled
equation of state,3o that can be tested by taking experimental results over the
entire critical region 32 rather than along a few special paths (critical isotherm,
31p

critical mole fraction, coexistence curve, etc.).

In the application to binary mixtures, one way of writing this ‘scaled equation
of state’ is
( p = p”): A(T,x) - d(T,XC) = (x - P ) l x - xCIW7(y),
(21)
where y = (T - TC)/Ix- xcl1/P and the function h(y) must satisfy a series of
condition^.^^ The variable y runs from a constant negative value on the coexistence
curve through zero on the critical isotherm to infinity at x = x,.
Schofield 3p has proposed a parametric transcription of the scaled equation
338

of state which, in the same binary mixture notation as equation (21), may be
written as
A(T,x) - A(T,xc) = d V ( 1 - O), (22)
(T - Tc) = r(l - b202), (23)
( x - XC) = r h ( 9 ) . (24)
Here the parameter Y is a measure of the ‘distance’ from the critical point; the
analytic equations fail as r goes to zero. The parameter 8, which defines the
‘direction’ from the critical point, is - 1 on the left-hand branch of the coexistence
+ +
curve and 1 on the right-hand branch, - l / b and l / b on the left-hand and
right-hand branches of the critical isotherm, and zero when x = xc. The
constants a and b may vary from system to system but must satisfy the conditions
a > 0, b > 1. The simplest acceptable form for the function m(8) is g 8 where
g is a positive constant; moreover, Schofield et suggest that b2 be set equal
to (6 - 3)/[(6 +1)(1 - 2/3)].
Essential to equations (21) and (22) is the assumption that d(T,x ) - d(T, xc)
is exactly antisymmetric around x = xc, a feature of uncertain validity. (We
shall examine the problem of symmetry in a subsequent section.) More general
expressions (‘extended scaling’) have been examined by Green, Cooper, and
Levelt S e n g e r ~ . ~ ~
28 L. P. Kadanoff, W. Gotze, D. Hamblen, R. Hecht, E. A. S. Lewis, V. V. Palciauskas, M. Rayl,
J. Swift, D. Aspnes, and J. Kane, Rev. Mod. Phys., 1967, 39, 395.
ZD L. P. Kadanoff, Physics, 1966, 2,263.
R. B. Griffiths, Phys. Rev., 1967,158, 178.
31 M. Vicentini-Missoni, J. M. H. Levelt Sengers, and M. S. Green, J. Res. Nut. Bur. Stand.,
Sect. A , 1969, 13, 563.
3p M. Vicentini-Missoni, R. I. Joseph, M. S. Green, and J. M. € Levelt
I. Sengers, Phys. Rev.,
1970, B l , 2312.
33 P. Schofield, Phys. Rev. Letters, 1969, 22, 606.
34 P. Schofield, J. D. Litster, and J. T. Ho, Phys. Reu. Letters, 1969, 23, 1098.
35 M. S. Green, M. J. Cooper, and J. M. H. Levelt Sengers, Phys. Rev. Letters, 1971, 26, 492.
 Critical Exponents for Binary Fluid Mixtures 249
Equations such as (21) and (22) regularize much of the experimental information
on critical phenomena, but they are nonetheless empirical. More recently,
Wilson36 has cast the Kadanoff scaling model in the form of the differential
equations of the renormalization group, and has shown that the form (symmetry)
of the Hamiltonian for the Ising ferromagnet leads to exponents in excellent
agreement with those derived by numerical methods. Impressive as this result is,
it is still based upon a scaling hypothesis; a rigorous statistical mechanical
derivation has yet to be found.

5 Some General Experimental Problems

Superficially one might expect that the determination of critical exponents for
binary mixtures should entail only the selection of an appropriate property, its
measurement as a function of temperature (or composition) with sufficient care
and precision, and a log-log plot of the results according to equation (7)- aided
perhaps by a least-squares analysis. Unfortunately, the situation in reality is
much more complicated. Several problems must be recognized and resolved
before one can reliably infer exponents from experimental results in the critical
solution region.
The Precision of Limiting Quantities.-As defined by equation (l), a particular
critical exponent 4 is not the average value of the slope when In I Y - Y c I is
plotted us. In I T - Tc I but the limiting value as T approaches Tc. Moreover,
since Y - Y c is not analytic, the problem cannot be dissected (as, for example,
the evaluation of virial coefficients can be) by evaluating successive coefficients
in a power series as intercept, limiting slope, limiting curvature, etc. There is no
completely convincing evidence - either from theory or experiment - as to the
range of T - TC[or rather the reduced quantity t = (T - TC)/TC)] over which
the exponent 4 can be expected to be sensibly constant.
Needless to say, as one approaches the limit, T - TCbecomes smaller and,
for the exponents ,8 and 6 at least, so does Y - Yc. Assuming that the precisions
of T and Y themselves are approximately constant, the uncertainties in
In I T - TCI and In I Y - Y c ( are growing as one gets closer to the desired
limit. Moreover, as we shall see, the closer one gets to the critical point, the more
serious the perturbations due to gravity become.
Statements that an exponent f has been determined with a precision (standard
deviation) of, for example, (T = 0.002 are suspect. Sometimes all this means is
that the experimenter has plotted all the measured points, weighted them equally,
and used a linear least-squares program to evaluate the best value of the slope
and its standard deviation; the 4 reported is then not a limiting value, but an
average value, and the (T may be nearly meaningless. The actual uncertainty
may be far larger, as more precise measurements and more careful analysis
may show.7
36 K. G. Wilson, Phys. Rev., 1971, B9, 3174, 3184.
The problem of weighting is non-trivial, especially so when, for as apparently simple an
exercise as fitting y(x) to a straight line, there are experimental regions where the major source
of error is in x and other regions where the major source of error is in y. The careful work of
Levelt Sengers and co-workers s2 illustrates the attention which needs to be paid to this
subject.
250 Chemical Thermodynamics
Today's experimentalists are more sophisticated and error analyses are now
both more detailed and more circumspect. Fewer claims of high precision are
being made; estimates of exponents and their uncertainties are surely now more
realistic, but the added caution leads to the superficial misimpression that the
experimentalists have made no progress.
Gravitational Effects.-In one-component fluids, the existence of sizeable
density gradients in equilibrium systems near the critical point are well known
and their effect on the determination of critical exponents is r e c ~ g n i z e d . ~ ' - ~ ~
An analogous situation involves composition gradients near the critical solution
point in binary mixtures whenever the densities of the two pure liquids are
appreciably different.
Fannin and Knobler 40 have recently presented an analysis of gravity-induced
concentration gradients and their influence upon the experimental determination
of p .
The mole fraction gradient for a binary mixture may be written

Here 11 is the height in the cell, g is the acceleration of gravity, and

where MI and M2, and V, and V, are the molar masses and the partial molar
volumes of the two components. When the volume change on mixing is small,
the partial molar volumes can be replaced by the molar volumes of the pure
components Mtlpl and M 2 / p 2 ,yielding an expression proportional to p2 - pl:

Fannin and Knobler 40 computed mole fraction profiles for a (classical) simple-
mixture model using a value of 2 appropriate for the rather extreme example
of CH4 + CF4 at T C where the density ratio p2/p, = 4.2 (Figure 4). We see
that the composition may vary appreciably in a cell whose height is no greater
than 1 cm, and that only at one particular level in the cell is the system at its
critical p0int.t Critical behaviour at x = xc can be observed somewhere in the
cell even though the average mole fraction 3 (i.e. the prepared composition) may
be sensibly different from xc.
Thus, at the critical temperature phase separation will be observed at the
bottom of the cell for a mixture for which 3 < xc, while at the top of the cell
phase separation will be observed when f > xc. If these observations are
recorded on a T, x diagram, they will produce not only a spurious flat top to the
coexistence curve, but also changes in the shape of other regions.
Using the scaled equation of state (21), Fannin and Knobler constructed an
'observed' coexistence curve for CH4 + CF4 in a 2 cm cell, assuming a 'true'
37 H. W. Habgood and W. G. Schneider, Canad. J. Chem., 1954,32,98.
38 C . Edwards, J. A. Lipa, and M. J. Buckingham, Phys. Rev. Letters, 1968, 20, 496.
3s P. C. Hohenberg and M. Barmatz, Phys. Rev., 1972, A6, 289.
'O A. A. Fannin, jun. and C. M. Knobler, Chem. Phys. Letters, 1974, 25, 92.
t Obviously only here is x = XC, T = TC ,p = pc. If there is a volume change on mixing
(dT/dp)c is non-zero along the critical line (typical values may be as great as 0.5 K MPa-l);
thus, although at equilibrium the entire cell is at uniform temperature, the pressure varies with
height and only at one level is T = Tr.
Critical Exponents for Binary Fluid Mixtures 251
TC = 94.300KandP = 0.350. Log-log plots of Ax = X” - x’vs. (T - TC)/TC
cease to be linear, and the experimenter would be tempted to adjust Ta in order
to make the curves more nearly linear. If this were done (Figure 5), unweighted

E
SO

-2

-4

-6
0.49 0.50 0.51
X

Figure 4 Mole fraction proj2e as a function of height h for a simple mixture at its critical
solution point (excess Gibbs free energy G g =2RTCx(l - x), xc = 1/2) for Z z
33 g mol-l, g = 9.8 m s - ~(normal gravity) [based upon data from Fannin and Kmbler
(ref. w1
least-square fits for TC = 94.301, 94.303, and 94.305 K would yield /3 = 0.321,
0.357, and 0.382, respectively. They conclude that experimental points obtained
close to Ta should be excluded from analyses which attempt to extract /3 from
coexistence-curve measurements.
Similar problems will arise in the analysis of other properties. For example,
the ‘observed’ heat capacity CVtm will not display an actual divergence. Only
one layer in the cell is at the true critical point and the total enthalpy H i s averaged
over all heights and will not show singular behaviour. The phenomenon is
entirely analogous to the gravitational perturbation of the measurement of Cv
for a pure
+
The composition gradients in CH4 CF, are of comparable magnitude to
the density gradients in Xe,31so we see that critical phenomena in binary mixtures
can be quite as sensitive to gravity effects as are gas-liquid critical phenomena
in pure fluids.t
t The detailed conclusions drawn above apply strictly only in the equilibrium limit. In fact,
as has recently been shown (S.C. Greer, T. E. Block, and C. M. Knobler, Phys. Rev. Letters, 1975,
34,250), strong concentration gradients develop rapidly at the top and bottom of the system, but
the gradient in the middle layers remains substantially unchanged (e.g. effectively zero) for very
long times. This feature makes analysis of, and correction for, gravity effects even more difficult.
252 Chemical Thermodynamics
In general, the magnitude of the gravitational perturbation will scale as
HgZC/RTCwhere H is the total height of the cell and Z c is the value of 2 at the
critical point. In an ultracentrifuge, where g can exceed lo5 normal gravity,
concentration gradients can be very large and very striking effects can be observed
in the critical r e g i ~ n . 42~ l ~

1 1

10-4 t 10-3 10-2

Figure 5 Log-log plots of x" - x' vs. t from a simulated coexistence curve. T c adjusted
from 'true' T C = 94.300 K to 94.301 K (circles and dashed line); 94.303 K (triangles
and full line); and 94.305 K (crosses and dot-dashed line) [based upon data from Fannin
and KnobZer (ref. 4011

Symmetry.-The Ising model, like the phenomenon of ferromagnetism which it

was designed to simulate, has as an essential feature an exact symmetry. All the
thermodynamic functions are symmetric or antisymmetric with respect to an
axis of zero magnetic field. Transcription of this model to that of a one-component
fluid leads to the highly artificial 'lattice-gas' in which the critical density pc is
exactly half the density po of the close-packed lattice.
A real fluid, of course, has no such natural and exact symmetry. Consequently,
in analysing real data to extract critical exponents, it is essential to select
variables in such a way that at least an approximate symmetry appears over as
wide a range as possible; only then can the asymptotic behaviour be examined.
It is well known that the coexistence curve is much more nearly symmetric in a
T, p representation than in a T, Vm representation; only the former yields, at
least approximately, the rectilinear diameter.
Levelt Sengers 43 has examined the antisymmetry character of argon isotherms
with respect to the critical isochore ( p = pc or Vm = V&)and has found that
J. H. Hildebrand, B. J. Alder, J. W. Beams, and H. M. Dixon, Y. Phys. Chew., 1954,58, 577.
E. Dickinson, C. M. Knobler, V. N. Schumaker, and R. L. Scott, Phys. Reu. Letters, 1975,34,
180; T. E. Block, E. Dickinson, C. M. Knobler, V. N. Schumaker, and R. L. Scott, J. Chem.
Phys., 1977,66, 3786.
43 J. M. H. Levelt Sengers, Ind. and Eng. Chem. (Fundamentals), 1970, 3 , 470.
Criticul Exponents for Binary Fluid Mixtures 253
the p, p representation is strikingly superior to either the p, p or p, Vm represen-
tations; only when I p - pc I is plotted against I p - pc I do the two branches of
the isotherm lie on approximately the same curve over an appreciable range of
densities. At most only one unique set of variables will display the appropriate
symmetry, so it must be chosen carefully. The structure of thermodynamics
alone is no guide, although special models (e.g. the van der Waals equation,
the lattice gas> are suggestive.

701 8 m r

60

50

40
9
30

20

10

0
0 01 0.2 03 0.L 0.5 0 02 0.L 0.6 0.8 1.0
X d
Figure 6 Test of symmetry variables for the system CH, + CF, at 98 K [based upon
data from Simon, Fannin, and Knobler (ref. 46)]

The same situation applies to the binary mixture. From the Ising model (or
from a van der Waals mixture with the volume parameter b the same for both
components), one concludes that the appropriate variables are A and x. But
these models have a special symmetry (V, = V,) such that XC = 1/2 and there is
no distinction between the mole fraction x and the volume fraction +:t

No actual binary mixtures have such exact symmetry (except for mixtures of
dl-enantiomers where no liquid-liquid phase separation has yet been found).
It is well known that coexistence curves for binary mixtures are frequently
more nearly symmetric in a T,9 representation than in a T, x repre~entation.~~~
46

Simon, Fannin, and Knobler 46 found, in analysing their vapour-pressure

+
measurements on the system CH4 CF,, that plots of d us. 4 [where their
definition of C$ takes the molar volumes in equation (28) as those of the pure

44 J. H. Hildebrand and R. L. Scott, 'Regular Solutions', Prentice Hall, Englewood Cliffs, New
Jersey, 1962, ch. 10.
46 J. B. Gilmour, J. 0. Zwicker, J. Katz, and R. L. Scott, J. Phys. Chem., 1967, 71, 3259.
46 M. Simon, A. A. Fannin, jun., and C. M. Knobler, Ber. BrinsengeseIIschaftPhys. Chem., 1972,
76,321.
-f The very definition of the volume fraction is ambiguous in a more general case. Are the
volumes which appear in equation (28) the partial molar volumes Vl(x) and Vz(x)or are they the
pure-fluid molar volumes Vl(0) and Y,(l)? Unless the volume change on mixing is zero, these
are not the same.
254 Chemical Thermodynamics
liquids] were much more nearly antisymmetric than plots of d us. x (Figure 6).
They suggest in the scaled equation of state (21) that the variable x - xc should
be everywhere replaced by 4 - p. Indeed it is plausible to regard d, 4symmetry
as the two-component analogue of the approximate p, p symmetry observed in
one-component fluids; just as p is proportional to n / V (where n is the amount of
substance), so also is 4 proportional to n,/V (only approximately so if there are
volume changes on mixing).
Similar conclusions as to the superiority of volume fractions over mole fractions
have been drawn by Zollweg47from a measurement of the coexistence curve of
n-C,H1, +n-CGFI1.
The choice of an inappropriate set of variables makes a substantial perturbation
in the extraction of critical exponents from experimental data. We illustrate this
by assuming that the expected asymptotic symmetry (or antisymmetry) will be
exact when a particular composition variable y [not the variable y of equation
(21)] is used and examine the complications introduced by the choice of a
different composition variable z, such thatt

Along the critical isotherm we assume that in the d , y representation, the

curves are exactly antisymmetric.

For the d,z representation, the 'observed' critical exponent 6* will be given by
the expression

where g = (f - 1)/2 is a parameter whose usefulness wilI appear below. We

see that for (f - 1 ) positive, one will deduce values of &* larger than the 'true'
6 from slopes of curves in the region y > yc and smaller values when y < yc
If data from both branches of the isotherm are available, the mismatch seen
in Figure 6 will be apparent.
The analysis of the coexistence curve is more complex. We assume 'exact'
equations for y' and y"
+ Ctl-" - Bts),
y' = + ( A
Y" = *(A + Ctl" + Bts),

where t = (TC - T)/Tc. Thus y" - y' will equal Btg and the diameter
(y' + y")/2 = +(A + Ctl-) includes the predicted 1 - a singularity. For the
t, z representation, one deduces from the coexistence curve 'observed' exponents

47 J. A. Zollweg, personal communication.

t Were z the mole fraction x, and y the volume fraction 4, f would be V,/V2 (again ignoring
volume changes on mixing).
Critical Exponents for Binary Fluid Mixtures 255
,8* and (1 - a)*
aln(z" - z')
a In t
- 2(1 -a)[l + g(A + Ctl-u)]gCt'-u
B+ [l + g(A + Ct1-)]2 - (gBtS)S
B+

+ (12/3g4B4t4s
+ gA)4
+ O(t28+1-u), (33)

(1 - a)* =

- 2/3[1 + g(A + Ct*d)l(gBt~)2]}

{[l + g ( A + Ct1-u)]2- (gBtB)2}

It is apparent from equation (33), especially when it is cast in the form of terms
of increasing (non-integral) powers of t, that the 'observed' p* should always
be slightly greater than the 'true' /3 whenever g is non-zero, i.e. whenever any but
the exactly appropriate symmetry variable for composition has been chosen.
However, the effect is usually small and, in any event, as t approaches zero, fl*
must approach fl.
Quite the opposite situation applies when we examine the 'observed' (1 - a)*
in an attempt to test the validity of the law of the rectilinear diameter.t According
to equation (34), unless g is exactly zero, the limiting value of (1 - a)* as t
approaches zero is not (1 - a), but rather 2p7 since 2/? c (1 - a). (This
spurious 2p exponent arises quite generally whenever any variable without the
correct symmetry properties is chosen; it does not depend on the particular
transformation o f equation (29). The effect appears to have been pointed out
explicitly first by Buckingham48for the transformation from p, p or T, p to
T, V' co-ordinates for the lat tice-gas coexistence curve.)
As we shall see, this feature effectively defeats the test of the rectilinear
diameter;$ we obtain virtually no information about a. The magnitude of the
M. J. Buckingham, in 'Phase Transitions and Critical Phenomena', ed. C. Domb and M. S.
Green, Academic Press, London and New York, 1972, vol. 2, p. 18.
t It should be emphasized that (1 - a) and (1 - a)* are not the exponents associated with the
+
diameter (x' x")/2 itself, but with its deviation from verticality, i.e. from XC. Thus an exponent
significantly different from unity does not always produce a diameter which obviously 'looks'
strongly curved.
3 This difficulty is non-existent in the 'classical' analytic formulations. Since @ = 1/2 and
01 = 0, 2 p = 1 - a = 1 and the distinction between the two vanishes.

9+
256 Chemical Thermodynamics
Table 3 ‘Observed’ exponents from the coexistence curve
a=O.O5, p=O.35, A=0.8, C=O.2, B = 2
g = -0.20 g = -0.01 g = -0.001
t p* (1 - a)* (1 - a)* (1 - a>*
10-6 0.3500 0.702 0.734 0.853
10-5 0.3501 0.703 0.756 0.884
10-4 0.3503 0.705 0.783 0.908
10-3 0.3514 0.710 0.8 17 0.925
10-3 0.3575 0.722 0.853 0.935
10-1 0.3933 0.763 0 884 0.94 1

problem is illustrated in Table 3 which shows values of the ‘observed’ exponents

/?* and (1 - a)* that would be derived from the slopes of the log-log plots at
various values of t . For these calculations [equations (33) and (34)] the ‘true’
a and p are assumed to be 0.05 and 0.35,values are chosen for A , C, and B that
would not be unreasonable for a hydrocarbon +
fluorocarbon mixture, and
several values of g are examined.
We see that for g = -0.20, which might represent the change from volume
fractions to mole fractions in a hydrocarbon +
fluorocarbon mixture, p* is
only slightly greater than the true except for very large values of t
or T C - T > 3 K for T C = 300 K), and that over the same range of temperature,?
(1 - a)* does not deviate much from 28. Of course the poor choice of a
composition variable would be signalled by a mismatch of the A , z plots or by
the very curved diameter. For a much better choice, g = -0.01, the ‘observed’
p*’s (not shown) are still closer to p, but (1 - a)* is still far from the ‘true’
1 - a = 0.95. Even for g = -0.001, a closer fit to the true symmetry co-
ordinate than there would be any expectation of achieving,$ the observed
(1 - a)*’s still cover quite a range. Moreover, the ‘worst’ values are obtained as
t approaches zero, and the ‘best’ values at large t where there is no reason to
expect that the actual experimental data (in any set of co-ordinates) fit the simple
asymptotic forms. As suggested above, this kind of analysis is unlikely to yield
meaningful evidence about the rectilinear diameter. If, as has been suggested,lg
the term Crl-” in equation (32) is masked by a higher term Dt, the prospect of
finding meaningful experimental evidence for a singularity in the ‘true’ diameter
is even less promising.
Impurities.-lf a fluid mixture has been prepared from two components that are
not entirely pure, what is thought to be a binary mixture will in fact consist of
three or more components. If so, the experimental measurements will normally
be made under conditions such that the mole fractions xi of the impurities are
held constant or varied in a linear manner; since this hidden constraint is a
49 B. Widom, quoted in ref. 8.
t There is no reason to expect asymptotic behaviour for t > so the numbers reported for
t = 10-l are not very meaningful. At the other end of the scale t = means a temperature
less than 1 mK below T“, close to the limit of most experimental precision.
$ In principle one could vary f and calculate a series of log-log plots from the experimental
data, finding the one that maximizes (1 -a)*. However, there is no guarantee that the trans-
formation to the correct symmetry variable is of the form of equation (29); indeed it is not clear
how to handle volume changes o n mixing.
Critical Exponents for Binary Fluid Mixtures 257
‘density’, not a ‘field’, the way in which different properties approach the critical
point is perturbed. The general theory of these perturbations60-62leads to the
phenomenon of ‘exponent renormalization’.
In what is thought to be a pure fluid, a measurement of Cv,mas a function of
Tmay, because of an impurity, turn out to be a measurement of CFrZtm for a
binary mixture and the expected divergence reduces to a finite cusp. In a
purported binary mixture an impurity changes CBIZIm to CP,x,,xr,rn
and again
a divergence is replaced by a finite cusp [i.e. equations like (14) instead of
equations (191.
No matter how slight the impurity, it must ultimately influence the limiting
behaviour at the critical point. However, a very small amount of impurity (or
an impurity like an isotopic variant with very similar properties) might not
produce a measurable difference in exponent except very close indeed to the
critical point, perhaps only at a value of (T - TC)well beyond our present
experimental ability to achieve. Only within a relatively narrow region around
the critical point could one observe the fully renormalized exponents. In terms
of the simple exponents, these renormalized exponents (distinguished by a
subscript x) have been shown6*to be
ax+ = -a+/(l - a+),

The magnitude of the effect of an impurity depends upon its amount and its
nature. Fisher and Scesney63 have made detailed calculations for a soluble
model (a mobile-electron ferromagnet), utilizing a dilution parameter X which
is approximately I T,C - TCl/T,C where TC is the critical temperature of the
system with the impurity while T,C is that of the ‘pure’ system. They find that
appreciable values of X are required in order to observe a renormalization effect.
For example, they find as a useful approximation for the ‘observed’ /3* deter-
mined? by a log-log fit of experimental data in the range of t from to lo-$:

Thus full renormalization would require an impurity that shifted the critical
temperature by approximately a factor of two.
The evidence to support renormalization comes not from studies of systems
with small amounts of impurities, but from measurements of ternary systems
with large amounts of the third component. Thus Bak, Goldburg, and Pusey 65 649

so J. W. Essam and H. Garelick, Proc. Phys. Suc., 1967, 92, 136.

51 B. J. Lipa and M. J. Buckingham, Physics Letters, 1968, 26A, 643.
62 M. E. Fisher, Phys. Reu., 1968, 176,257.
63 M. E. Fisher and P. E. Scesney, Phys. Reu., 1970, A2, 825.
64 C. S. Bak, W. I. Goldburg, and P. N. Pusey, Phys. Rev. Letters, 1970,25, 1420.
66 W. I. Goldburg and P. N. Pusey, J. Phys. (France), 1972, 33, C1-105.

t ”e /3* of equation (36), like the B* of equation (33), shows the effect of an unanticipated
complication upon the ‘observed’ value. The two effects are different and might both need to be
allowed for in any actual analysis of an experimental measurement.
258 Chemical Thermodynamics
have found (in light-scattering studies on the system bromobenzene water + +
acetone), an enhanced y+ = 1.50, while Zollwegs6 has deduced from the
T , x 2 , x Qcoexistence surface for ethanol + water +
chloroform, an enhanced
/3 = 0.38.
In carefully designed experiments on ‘binary’ systems of reasonably pure
components, it does not seem likely that the presence of small amounts of
impurities (particularly those which are hard to detect or remove) can appreciably
perturb the determination of critical exponents.

6 Experimental Methods and Resdts

In this section we try to summarize what is known experimentally about the critical
exponents in binary mixtures, together with some comments on the experimental
techniques that have been used or that could be used. In the tables, estimates of
range or uncertainty ( + ) are the opinions of the experimenter or of those who
analyse the data, not of this reviewer. Sometimes these estimates are based upon
a least-squares fit; sometimes they are little more than intuition. In any case,
because of one or more of the complications outlined in the previous section, the
‘true’ value of the exponent may lie well outside the error limits reported.
Coexistence Curves.-The most numerous and most extensive measurements on
binary mixtures in the critical region are, of course, those of T,x phase diagrams.
Early studies not designed to yield critical exponents are rarely useful because
they include few if any precise measurements close to the critical point
( t G 10-3).
Experimentally the easiest procedure is to prepare samples of binary mixtures
of known composition, warm each until it is in the homogeneous one-phase
region and then cool it slowly until separation into two phases occurs. Fortu-
nately under most of the usual experimental conditions supercooling is slight
in liquid mixtures. (Substantial supercooling has been observed in C7Hl4+ C,FL4
under conditions designed to test current theories of nucleationP7,68) Thus, if
one can be confident of the uniformity of the temperature throughout the system
and if proper allowance is made for the concentration gradients due to gravity
discussed in the previous section, one can safely record the T , x observationt
as a point on the equilibrium coexistence curve.
This method introduces some difficulties into the analysis: (1) one does not
have experimental points x’ and x” at the same temperature for both branches
of the coexistence curve; and (2) one has to infer from the set of experimental
points the ‘best’ co-ordinates of the critical point (TC, xc). The second difficulty
is probably inherent in any method, but the &st difficulty could be avoided by
sampling isothermal conjugate phases and analysing them for x’ and X I by any
appropriate method (e.g. density or refractive index). In fact this has rarely
been done. Thompson and Rice and Wims, McIntyre, and Hynne 6o measured
56 J. A. Zollweg, J. Chem. Phys., 1971, 55, 1430.
57 B. E. Sundquist and R. A. Oriani, J. Chem. Phys., 1962,36,2604.
68 R. B. Heady and J. W. Cahn, J. Chem. Phys., 1973,58, 896.
6D D. R. Thompson and 0. K. Rice, J. Arner. Chem. Soc., 1964,86, 3547.
A. M. Wims, D. McIntyre, and F. Hynne, J. Chem. Phys., 1969,50,616.
t The gravity correction arises because the mole fraction x at the level in the vessel where the
initial phase separation occurs is not exactly equal to X , the overall composition of the system.
 Critical Exponents for Binary Fluid Mixtures 259
densities of coexisting phases by float techniques;61theirs are probably the most
accurate determinations to date of binary mixture coexistence curves.61 A
magnetic float device capable of even higher precision has recently been
described.62
In analysing the results, a common practice is to construct a Cox-Herington
diagram 63 in which7 (T - TC)ll3is plotted against x, and is adjusted until the
data fall on two reasonably good straight lines intersecting at (T - TC)lis= 0
and x = xc. With TCand xc established, one can construct a kind of tentative
diameter f ( T ) , and then plot I x - 2 I us. I TC - 7' I on a log-log scale to deter-
mine a better value of /lthan the initially assumed 1/3. In the final stages, xc
and TCmay be adjusted further to improve the fit, although this procedure may
mask other problems as Fannin and Knobler *O have stressed (see Figure 5).
Figure 7 illustrates the treatment of the data for 3-methylpentane nitro- +
ethane by Wims et CZZ.,~O showing A, the coexistence curve; B, a Cox-Herington
plot, and C, a log I p" - $ 1 us. log I t l plot. (Densities p rather than mole
fractions x are shown, because there were the quantities directly measured by
a magnetic-float technique.)
Table 4 presents a selection of determinations, mainly recent, of the exponent
/Ifrom coexistence curves; also shown is the coefficient B in equation (8). For a
variety of binary mixtures, including non-polar substances, water, and metals,
the /3's lie between 0.315 and 0.38, with a strong preference for 0.33 to 0.35;
indeed Stein and Allen conclude that the nine systems which they analysed are
all consistent with /3 = 0.342 f: 0.016. One is tempted to conclude that the
constant is universal for fluids and fluid mixtures; if so, why is it not the same as
that for the three-dimensional king model, 0.31 or at most 0.32? Levelt Sengers
regards the evidence from pure fluids that fl is larger (about 0.35) as compelling;
the measurements on binary mixtures seem somewhat less conclusive.
The coefficient By while not required to be the same for all mixtures (unless
one assumes a conformality of intermolecular interaction energies, which produces
a principle of corresponding states), seems to be roughly the same for most
mixtures. $
Up to now, much less effort has focussed on the question of the rectilinear
diameter in mixtures than in gas-liquid one-component However,
Stein and Allen 61 have examined this as part of their analysis of coexistence data
on nine systems and have been unable to draw meaningful conclusions about the
possible singularity. For example, they find that analysis of the coexistence curve
results of Wims, McIntyre, and Hynne 6o on 3-methylpentane + nitroethane
yields (1 - a)* = 0.62 f 0.21 if volume fractions are used, 0.67 f 0.20 if

61 A. Stein and G . F. Allen, J. Phys. Chem. Ref. Data, 1973, 2,443.

62 S. C. Greer, M. R. Moldover, and R. Hocken, Rev. Sci. Instr., 1974,45, 1462.
a3 J. D. Cox and E. F. G. Herington, Trans. Faraday SOC.,1956, 52, 926.
43' J. M. H. Levelt Sengers, J. Straub, and M. Vicentini-Missoni, J. Chem. Phys., 1971,54, 5034.
-
The original observation of Cox and Herington was that if (T Tc)1/3was plotted against
ln[x/(l - x ) ] a very wide range of experimental points could be represented by two straight lines
intersecting at the critical point. For x - XC small, ln(x/[l - +
x ) ] reduces to ln[fi/(l - XC)]
( x - *)/[fi(l - - x",'.
xC)l*+* OCX
$ Some of the variation in B is coupled to the variation in 6. Assuming a fit at t =
a change of 0.01 in the exponent fi produces a 10 per cent change in the coefficient B.
9-
Table 4 The exponent /3 for binaryfluid mixtures
Range of
System t = I T - TC]/To 18 B Reference
CC14 + C7F14 6 x to 5 x 0.3 Zimm 60, 6s
2 x 10-7 to 2 x 10-6 0.333 k 0.018 0.31 & 0.05 Thompson and Rice 6 s s 61
CH4 + CF, 1 x 10-4 to 2 x 10-2 0.35 Fannin, Knobler, and Simon 6o
n-C,H,, +
n-C,F,, 2 x 10-3 to 3 x 10-2 0.34 2.0 Bedford and Dunlap O7
4 x 10-4 to 3 x 10-3 0.36 1.7 Gaw and Scott l7
3 x 10-6 to 3 x 10-4 0.38 1.4 Judd and Scott 68
3 x 10-6to 3 x 10-2 0.35 1.9 above, combined
Ne + Xe 2 x 10-6 t o 4 x 10-8 0.35 i: 0.02 Deerenberg, Schouten, and
Trappeniers
2 x to 8 x 0.332 f 0.007 1.81 f 0.05 Wims, McIntyre, and Hynne *I
3 x 10-6 to 7 x 10-3 0.342 If: 0.004 2.02 f 0.03 Irani and Rice 61* 70
5 x 10-5 to 1 x 10-2 0.34 Atack and Rice 6os 71
1 x 10-6 to 1 x 10-2 0.324 i-0.009 1.78 5 0.01 Kao and Chu 61s 72
7 x 10-6 t o 4 x 10-4 0.34 k 0.04 2.0 f 0.7 Loven and Rice * l s 73
2 x 10-6 to 4 x 10-3 0.33 f 0.03 3.4 k 0.3 Kohler and Rice 61* 74 (7
3 x 10-6 to 1 x 10-2 0.34 t 0.03 2.7 f 0.2 copp 81s 76 x3
1 x 10-3to3 x 10-2 0.346 0.02 0.49 f 0.04 Rudd and Widom 61s 76 2
1 x to 8 x 0.336 & 0.03 0.70 f 0.10 Woerman and Sarholz 61*77
1 x 10-5 to 1 x 10-2 0.36 f 0.02 *
1.9 0.1 Govindarajan, Subramanyam, and
Gopal 78
5
2
1 x 10-6t04 x 1 0 - 3 0.38 & 0.02 1.7 f 0.1 Viswanathan, Govindarajan, and 2
+ CH,OH
Gopal 79 2
CS2
CS2 + CH,N02
1 x to8 x
1 x 10-ato 2 x 10-1
0.35 i- 0.03
0.315 k 0.004
2.1
1.67 k 0.01
Viswanathan, Gambhir, and Gopal 8o
Gopal, Ramachandra, and
9
Chandra Sekhar 5
F'
CSa+ (CH3CO)aO 1 x to 1 x 0.34 1.7 Lele, Subramanyam, and Gopal 9
rr.

Na + NH, 1 x 10-4 to 1 x 10-2 0.34 0.02 Chieux and Sienko $.

Ga + H g 1 x 10-4 to 1 x 10-2 0.34~ Schurmann and Parks 13' .4

a Fitted using density p as composition variable. Fitted using volume fraction 4 as composition variable. A t large values of t (> 10-3, the
coexistence curve is classical ( @ x OS), which suggests that here the range of intermolecular interaction exceeds the range of correlations. % a
(L)Coexistence curve shows a lower critical solution temperature.
3
66 B. H. Zimm, J . Phys. Coll. Chem., 1950, 54, 1306. z:
l6A. A. Fannin, jun., C. M. Knobler, and M. Simon, personal communication, quoted in ref. 11.

67 R. G. Bedford and R. D. Dunlap. J. Amer. Chem. Soc., 1958, 80, 282.

3
N. F. Judd and R. L. Scott, to be published, quoted in ref. 11. 9
69 A. Deerenberg, 5. A. Schouten, and N. J. Trappeniers, Chem. Phys. Letters, 1974, 28, 316.
7 0 N. F. Irani and 0. K. Rice, Trans. Faraday Soc., 1967, 63, 2158.
s
4
D. Atack and 0. K. Rice, Discuss. Faraday Soc., 1953, 15, 210.
W. P. Kao and B. Chu, J. Chem. Phys., 1965,50, 3986.
7s A. W. Loven and 0. K. Rice, Trans. Faraday SOC.,1963, 59, 2723.
3c
l a F. Kohler and 0.K. Rice, J. Chem. Phys., 1957, 26, 1614.
76 J. L. Copp, Trans. Faraday SOC., 1955, 51, 1056.
's:
D. F. P. Rudd and B. Widom, J. Chem. Phys., 1960,33, 1816. g
77 D. Woerman and W. Sarholz, Ber. Bunsengesellschaft Phys. Chem., 1965, 69, 319. F
78 K. Govindavajan, S. V. Subramanyam, and E. S. R. Gopal, J . Chem. Phys., 1972,56,4235.
2
79 A. Viswanathan, K. Govindavajan, and E. S. R. Gopal, Ind. J. Pure and Appl. Phys., 1973, 11, 157.

A. Viswanathan, R. D. Gambhir, and E. S. R. Gopal, J. Chem. Phys., 1970, 53, 4405; see also ref. 91.
61 E. S. R. Gopal, R. Ramachandra, and P. Chandra Sekhar, Pramiina, 1973,1,260; see also E. S. R. Gopal, R. Ramachandra,
P. Chandra Sekhar, K. Govandarajan, and S. V. Subramanyam, Phys. Rev. Letters, 1974, 32, 284.
M. V. Lele, S. V. Subramanyam, and E. S . R. Gopal, Chem. Phys. Letters, 1971, 11, 542.
P. Chieux and M. J. Sienko, J. Chem. Pfiys., 1970, 53, 566.
H. K. Schurmann and R. D. Parks, Phys. Rev. Letters, 1971,26, 367.
262 Chemical Thermodynamics

0
'

*'

298

i
I I I
-

-
q
'c, 0.10 -

4'
/
&*
,.p -
-

-
d
P
/*
I I

p / g cm-3
0.20 1 I
1
C
m
0.10
Ecn 0.95
-
i
\
/3=0.34
Q
- 0.02
:I
Q

0.01 I I

-10-5 1 o-4 10-3

c
I

Figure 7 Experimental determination of 16 for 3-methyipentane -k nitroethane.

A, Coexistence curve. B, Cox-Herington plot. C, log I p" - p' I - log I t I plot
[based upon data from Wims, Mclntyre, and Hynne (ref. 60)]
Critical Exponents for Binary Fluid Mixtures 263
mole fractions are used, and 0.85 k 0.06 if the directly measured densities are
used. (The tempting conclusion that density is the best of the three variables is
suspect because the conversion to mole fraction and volume fraction was made
using the incorrect assumption that the volume change on mixing is zero.)
Other Thermodynamic Properties.-Determination of the other critical exponents
from thermodynamic measurements is less common and usually less precise.
The exponent 6 is obtainable from the slope of the critical isotherm, but
unlike the p , p plot for a one-component fluid the d, x (or A , #) diagram is not
constructed from directly measured quantities; pl and p2 (and hence d) must
be deduced from vapour-pressure meas~rements.~~ The few values reported so
far (Table 5 ) lie between 4 and 5 , and are consistent with either the Ising value
(5.1 f 0.1) or the best pure-fluid value (4.4 2 0.4).

Table 5 The exponent 6 for binary fluid mixtures

System 6 Reference
CC14 C7F14 + z4 Zimm 65, 85
C2H6 CF4 + 4 to 5 Croll and Scott 86
CH, + CF* 4.6 Simon, Fannin, and
Knobler 46

The exponent y+ is deduced from the quantity ( a d / a ~ )evaluated

~, at x = xc
for a series of temperature^;^^ again d must be extracted from measurements
of vapour pressure. Alternatively measurements of V g or H g can be 68

a log-log plot of (a2Vm/ax2)T,p,r-zc or (a2Hm/a~2)T,S,2--zc against t yields a

slope (y+ - 1). The majority of y+ values are obtained from light scattering
measurements, as we shall see in the next subsection; the few thermodynamic
studies are summarized in Table 6 . No experimental information about y- for
binary mixtures is yet available.
Table 6 The exponent y+ for binary fluid mixtures
System Measurement Range oft Y+ Reference
CH, + CF4 vapour pressure 1 x to 4 x 1.4 Simon, Fannin, and
1x to 2 x 10-1 1.3 Knobler48
n-GHI4 + volumechange 5 x lo6 t o 4 x 1.3 GawandScottl’
n-C6F14 4 x to7 x 1.37 JuddandScott68

The exponent a is the most elusive of all. In principle it can be determined

[see equation (12)] by measurements at constant x = xC of the enthalpy H or
the volume V as a function of temperature at constant pressure (yielding
CB,z,m or a,,J or the volume as a function of pressure at constant temperature
(yielding K ~ , ~ )While
. volumetric measurements leading to ag,=may ultimately
yield the best values of a, early experiments 88 (before the concept of critical
*#’

exponents was fully developed) yielded little useful information, and most
86 P. Heller, Reports Progr. Phys., 1967, 92, 730.
I. M. Croll and R. L. Scott, J. Phys. Chern., 1964, 68, 3853; see comment by R. L. Scott in
‘Critical Phenomena’, ed. M. S. Green and J. V. Sengers, U.S. Nat. Bur. Stand. Misc. Publ.
273, 1965, p. 25.
87 G. Jura, D. Fraga, G. Maki, and J. H. Hildebrand, Proc. Nut. Acad. Sci. U.S.A., 1953,39, 19.
J. S. Rowlinson, ‘Liquids and Liquid Mixtures’, Buttenvorths, London, 1959, 1st edn., p. 170.
 N
Q\
P

Table 7 The exponent a for binarypuid mixtures

System Measurement Range of I t I 01 Reference
+
CHaOH c-GHIS Clb0,rn 3 x 104 to 5 x 10-3 0.0 d < 0.2
a! Anisimov, Voronel, and Ovodova 8D
CHF3 + CF4 Cn,e,m 1 x 10-3 to 1 x 10-1 a+ = 0.03 Cope, Reamer, and Pings
Czwrm 2 x 104 to 1 x 10-1 a!; = 0.06
CS2 + CH,OH clhar,m 3 x 10-4 to 1 x lo-* a! = 0 or 0.1 Gambhir, Viswanathan, and Gopal Dl
a!,, 2 x 10-4 to 2 x 10" 01 =0 Lele, Subramanyam, and Gopal 92
CH30H + n-C,H,
0
G,
0,m 8 x lod to a!+ = 0 or 0.13 Viswanathan, Govindarajan, and Gopal 79
7 x 10" to 5 x lo-* a!- = 0 or 0.14
3He + 4He CV,B,rn a 1 x to 1 x a!* = 0.09 0.02 Brown and Meyer D3
Brown and Meyer conclude that in the binary fluid (near its gas-liquid critical point) Cv appears to diverge like CV for pure We. They attribute
this to the small slope (aT/ad)of the critical line and conclude that truly asymptotic behaviour will not be observed unhl a much smaller value of I is
reached.

M. A. Anisimov, A. V. Voronel, and T. M. Ovodova, Soviet Phys. J.E.T.P., 1972,34, 583.

** A. F. G. Cope, H. H. Reamer, and C. J. Pings, Eer. Eunsengesellschaft Phys. Chem., 1972, 76, 318.
R. D. Gambhir, B. Viswanathan, and E. S. R. Gopal, Indian J . Pure Appl. Phys., 1973, 9, 787.
M. V. Lele, S. V. Subramanyam, and E. S. R. Gopal, Indian J. Pure Appl. Phys., 1973,11,228.
s3 G. R. Brown and H. Meyer, Phys. Rev., 1972, A6, 1578.
Critical Exponents for Bitiary Fluid Mixtures 265
subsequent workers have concentrated on the heat capacity C’l)rz,m.The
difficultiesin the precise determination of a are several:
(1) Before one makes a log-log plot, one must subtract from Cv,I,mor aP,,
a non-singular ‘background’, a contribution that would be present even in the
absence of special critical phenomena; it is not easy to estimate this term.
(2) Unless the measurements are carried out in a horizontal cell of negligible
height, gravity corrections can be serious.
(3) Measurements need to be carried out over several decades in t ; otherwise
it is not possible 48 to distinguish between (a) a weak divergence (0 < a < 0.1 5),
24s

(b) a logarithmic singularity? (a! = 0), or even (c) a sharp but finite cusp
[equation (13)].
Table 7 reports some recent attempts to evaluate a for binary mixtures.
It is clear from the reports of these experimenters, many of whom tested for
both positive a (log-log plots) and zero a! (semi-log plots), that they could not
distinguish between the two; evidently all that one can say is that 0 d a < 0.1.
Finally, there have been at least three attempts to fit the properties of binary
mixtures over a range of T, p , and x by a scaled equation of state. These assume
scaling, and usually fit 18 to a coexistence curve and choose a second exponent
by a least-squares or visual fit. Table 8 summarizes these. Of these, the most

Table 8 Critical exponents from a scaled equation of state

System Measurements 18 6 a Y Reference
CH, + CF, vapour pressure 0.35 4.6 (0.04) (1.26) Simon, Fannin, and
Knobler 42
n-C,H,, + volume change 0.35 4.8 (-0.03) (1.33) Judd and Scott 68
n-CP14
3He + ,He vapour-liquid 0.361 4 27 (0 098) (1.18) Leung and
equil. Griffiths O4

extensive results and the most careful analysis are those by h u n g and Griffiths 9 4
on the mixture 3He + ,He; the fact that this system, in which quantum corrections
must be substantial, yields exponents almost indistinguishable from those reported
for most other systems, is certainly strong evidence for a principle of universality.
Light Scattering.-It is well known that the scattering of light by matter arises
because of inhomogeneities in density on the scale of the wavelength of the
incident radiation. In the neighbourhood of the critical point, large density
fluctuations (concentration fluctuations in a binary mixture) produce large
local fluctuations in the refractive index and in the electron density, causing
visible light to be scattered strongly, a phenomenon called ‘critical opalescence’.
Quantitative measurements of scattering from fluids and fluid mixtures, made
increasingly precise by the use of lasers, yield much detailed information
84 S. S. Leung and R. B. Griffiths, Phys. Rev., 1973, AS, 2670.
t A logarithmic singularity, i.e. one that satisfies the equation Cq,F,m= CF ;z,m - E In I t 1,
where C$,x,m is the ‘background’ term, can yield spurious positive values of a because
a ln(Cv,z,m - C;,,,,)/a In I t I yields an observed a* = l/ln I t I. Thus at t = a* = 0.144;
at t = a* = 0.086; and at t = lo-’, a* = 0.062. Only the extreme limit at t + O does
a* approach the ‘true’ exponent zero for a logarithmic singularity. Only if the observed slope
a* is essentially constant over several decades can the logarithmic singularity be excluded.
266 Chemical Thermodynamics
concerning both static and dynamic properties. Details of the theory and the
experiments can be found elsewhere;g5,g6 a brief discussion will suffice here.
The differential scattering cross-section, or the Rayleigh ratio 9, is defined
as the flux of light scattered into a unit solid angle divided by the incident intensity
and by the scattering volume, and may be written as
w = I(R, e, $)Rv~v, (37)
where I(R, 8,+) is the average scattered intensity measured at a point with
spherical polar co-ordinates R, 8, and (5, I. is the incident intensity, and V is
the scattering volume.
The effect of long-range fluctuations upon light scattering in the critical
region was first accounted for by Ornstein and Zernikeg7who assumed that the
long-range tail of the total correlation functiont ldr) decays as a function of
intermolecular separation as
=
h(r) exp(- r/&/r, (38)
where E is the long-range correlation length. With this assumption, Ornstein
and Zernike derived an expression for 9 for a one-component fluid :
9(K) = (kt/l 6 x 2 )sinz+p2(a&/ap)~kTKT/(1 + f2K2), (39)
where k , = 2n/X,, A, is the incident wavelength in vacuum, is the angle between
the polarization direction of the incident plane polarized light and the direction
of observation, p is the fluid density, E is the dielectric constant, k is the Boltzmann
constant, is the isothermal compressibility, and K is the magnitude of the
momentum transfer vector equal to (27;11/h0)(2sin(8/2)), where n is the index of
refraction of the fluid and 6 is the scattering angle. Given fixed angle and wave-
length, the only quantities varying markedly as one approaches the critical
point are K~ and (.
Equation (39) is exact in the limit K -+ 0 where the factor (1 + t 2 K 2 )becomes
unity, a result originally due to Einstein,gebut it is now believed to be somewhat
incorrect when f 2 K 2 is substantial. FishersQ has proposed a modification of
equation (38), such that
h(r) z exp(- (/r);rl+v, (40)
where 7 is another critical exponent believed to be small and positive. This
leads to a modification of equation (39):
+ (2K2)1-"2. (41)
9(K) = (k$/16x2)sin2~p2(~&/~p)*kTKT/(1
Moreover the long-range correlation length is expected to vary with
t = (T - TC)/TC
with an exponent v such that
E = f*t-v+, (T > TC, p = p"); (4W
6 = So( - r p - (along the coexistence curve). (42b)
*I B. Chu, Ber. Bunsengesellschaft Phys. Chem.. 1972, 76, 202.
B. Chu, 'Laser Light Scattering', Academic Press, New York and London, 1974.
97 L. S. Ornstein and F. Zernike, Proc. k . ned. Akad. Wetenschap., 1914,17,793; 1915,18, 1520;
1916, 19. 1312, 1321.
8E A. Einstein, Ann. Physik. 1910, 33. 1275.
M. E. Fisher, J . Math. Physics, 1964, 5, 944.
i-The total correlation function h(r) = g ( r ) - 1 where g(r) is the pair distribution function or
radial distribution function. For a detailed discussion of these functions, see ref. 10.
Critical Exponepits for Binary Fluid Mixtures 267
The scaling hypothesis 28 leads to the predictions that
27p

(43)
and
(44)
or, rewriting equation (44)to yield an expression for q :
r] = 2 - (y/v) = (5 - S)/(S + 1). (45)
The procedure for extracting the exponents y and v from experimental light-
scattering data (and information on TCand pc from other measurements) would
be straightforward? were y = 0 and the Ornstein-Zernike equation (39) valid.
One would then plot the reciprocal intensity I-l as a straight-line function of
K 2 for a series of temperatures. The intercept at K 2 = 0 is proportional to
1 / ~ so
~ , a log-log plot of the intercept us. t yields y. Similarly the slopes are
proportional to [ 2 / ~ T so
, a log-log plot of C2 us. t yields v. Actually the experi-
mental results so plotted do yield very good straight lines in many cases, so much
so that the exponent r] has proved to be very elusive. Only in a few cases have
estimates of 7 been feasible.
The binary mixture situation is more complex. The differential scattering
cross-section resulting from concentration fluctuations alone may be written as
W&) .kTQT,,/(1
= (G/l 6x2)sin24x2((a@x)$. + f2K2)1-q”2. (46)
In addition however, there is a background due to density fluctuations which
must be subtracted from the total differential cross-section.lOO A further
complication, that of depolarization by double scattering near the critical point,
has been pointed out recently by Oxtoby and Gelbart;lol its effect on the analysis
of results from binary mixtures has yet to be assessed.
Figure 8 illustrates a reciprocal intensity plot for the system n-decane +
P,P’-dichloroethyl ether (Kao and Chu106). Note that the slopes are approxi-
mately the same for all temperatures; the slopes are proportional to [2/QT,3,i.e.
to t2v’-‘“+ which is the same as t2q”if the scaling hypothesis is correct.
Table 9 summarizes the results of some recent light-scattering measurements
on binary mixtures. We see a gratifying uniformity about the values of y and v.
Indeed Chu has concludeds5 that if the exponents y and v are the same for all
fluid systems (one-, two-, and three-components), y = 1.23 A 0.02 $ and
Y = 0.63 rt 0.02. Assuming scaling (equation (45)), this yields q = 0.05 L- 0.06,
hardly a useful result in itself.
loo B. Chu. J. Chern. Phys., 1964, 41,226.
lo1 D. W. Oxtoby and W. M. Gelbart, J. Chern. Phys., 1974,60,3359.
t This comment ignores all the experimental difficulties in obtaining accurate data to use in
the analysis.
8 This value is in good agreement with the Ising model prediction, but not with the few thermo-
dynamic measurements on binary mixtures from which decidedly higher values of y were
extracted (Table 6). Since these thermodynamic measurements, although of different properties
of different systems, were all made in one laboratory, this anomalous result has been referred to
as ‘the UCLA effect’. While my colleague Professor Knobler and I are not aware of any major
sources of error in our measurements, we regard the weight of evidence from other sources as
indicating fairly overwhelmingly that the value of y must be appreciably lower.
Table 9 Critical exponents from light scattering of binary mixtures
System Range o f t Y+ V ?1 Reference
CCl4 + C7Fl4 6 x 10-4 to 5 x z 1.2 Zimni 67
3 x 10-4to7 x 10-3 1.15 f 0.02 0.54 _t 0.02 Chu, Thiel, Tscharnuter, and
Fenby lo2
0.10 f 0.01 a Brady, McIntyre, Myers, and
Wims lo3
1.24 & 0.04 0.62 f 0.03 Chu, Schoenes, and Kao Io3
3 x 10-6 to 2 x 10-3 1.25 ? ? Kao and Chu
4 x 10-6to 1 x 10-9 1.oo x 0.44 0.055 k 0.037 Jaeckel and Majer lo6
5 x 1O6to6 x 1.14 rf: ? 0.2 It ? Kao and Chu lo'
7 x 10-6 to 1 x 10-8 1.22 7z 0.40 0.07 k 0.05 Jaeckel and Majer IO8
GHSOH + H2O 1.32 k 0.03 0.6 f 0.1 Pusey and GoIdburg lo8
y- = 1.36, 1.28
7 x 10" to 1 x 10-2 1.22 f 0.01 0.63 5 0.01 0.08 f 0.05 Calmettes, Lagues, and Laj loo
1.13 to 1.18 0.57 t o 0.67 Jaeckel and Majer 110
0.2 f 0.1 McIntyre and Sengers ll1
2 x to 8 x 1.23 f 0.04 0.616 5 0.013 0.006 f 0.040 Keyes, Chang, Sengers, and
Alley 118
polystyrene + c-qH12 1 x 10-1 to 1 x 10-2 1.26 f: 0.08 0.62 f. ? Kuwahara, Fenby, Tamsky,
and Chu 113
0 From small-angle X-ray scattering; all other measurements are for visible light. From their light-scattering measurements at different mole fractions,
Jaeckel and Majer deduce 6 = 4.7 rt 1.2.
lo2 B. Chu, D. Thiel, W. Tscharnuter, and D. V. Fenby, J. Phys. (France), 1972, 33, Cl-111. cl
loS G.W.Brady, D. McIntyre, M. E. Myers, and A. M. Wims, J. Chern. Phys., 1966,44,2197.
B. Chu, F. J. Schoenes, and W.P. Kao, J. Amer. G e m . Soc., 1968,90,3042.
irs
l o S W. P. Kao and B. Chu, J. Chem. Phys., 1969,50, 3986. r;'
lo@ K. Jaeckel and U. Majer, Ber. Bunsengesellschaft Phys. Chem., 1972,76, 232. %
!
B. Chu and W. P. Kao, J. Chem. Phys., 1965,42,2608. 2
lo* P. N. Pusey and W. I. Goldburg, Phys. Rev., 1971, A3, 766.
log P. Calmettes, I. Lagues, and C. Laj, Phys. Rev. Letters, 1972, 28, 478; J. Phys. (France), 1972,33,C1-121.
K.Jaeckel and U. Majer, Ber. Bunsengesellschaft Phys. Chem., 1973,77, 364. $
111 D. McIntyre and J. V. Sengers, in 'Physics of Simple Liquids', ed. H. N. V. Temperley, J. S. Rowlinson, and
2
G. S. Rushbrooke, North-Holland, Amsterdam, 1968, pp. 473-474.
118 P. H.Keyes, R. F. Chang, J. V. Sengers, and C. 0. Alley, Univ. Maryland Dept. of Physics and Astronomy, Technical

Report 73-018, 1972.

-s
n'
b
llS N. Kuwahara, D. V. Fenby, M. Tamsky, and B. Chu, J. Chem. Phys., 1971,55, 1140.
Critical Exponents for Binary Fluid Mixtures 269
Additional information is yielded by studies on three-component systems,
some of which have been mentioned in the earlier discussion on i m p i i r i t i e ~ . ~ ~ - ~ ~
Some show substantial renormalization; others do not.
Lin, Thiel, and Chu 11* have made a careful light-scattering study of the three-
component system ethanol + water + chloroform (the system for which Zollweg 65

1 I 1 I

0 20 40 60
106(2/A l2 sin2(8/2)/nrn2
Figure 8 Typical light-scattering plot of reciprocal relative intensity I;' vs. (2/h)2sin2(8/2).
Measurements on the system n-decane + p, p'-dichloroethylether containing mole
fraction 0,3952 of n-decane corrected for density fluctuations and for attenuation.
(The diferent symboIs indicate different wavelengths and different cells.) The intercept
of each line is proportional to Q F : ~while
, the slope is proportional to f 2 / Q T , ,
(Reproduced by permission from J. Chem. Phys., 1969, 50, 3986)

measured the coexistence curve precisely) and extracted exponents yx = 1.50 f

0.03,v, = 0.765 f 0.015, and = 0.08 f 0.02. The exponent 7 does not undergo
renormalization, but yx and v, appear to be fully renormalized. Indeed to reduce
yx = 1.50 k 0.03 to y = 1.23 f 0.02 requires [equation (35)] a = 0.19 0.02,
somewhat larger than the Ising model prediction of about 0,125.

11* F. L. Lin, D. Thiel, and B. Chu, Phys. Letters, 1974, A47,479.

270 Chemical Thermodynamics
7 Some Tentative Conclusions
We return now to the questions posed at the end of the Introduction. Given the
present experimental situation, and the uncertainties recognized (or unrecognized),
no unequivocal conclusions should be drawn. What follows are the tentative
personal opinions of one reviewer.
1. Within experimental error the critical exponents /!I, y , and u appear to be
the same for all binary mixtures. The same statement (within wider limits) can
be made about the exponents 6, 01, and 7, but the evidence is more fragmentary.
When allowance is made for the analogies between one, two, and three compon-
ents, the exponents appear to be the same for all fluids.?
2. Within the appreciable experimental error, the scaling relations [equations
(19), (20), (43), and (44)] appear to be satisfied.
3. Tt is hard to reconcile the exponent /3 = 0.35 -t- 0.02 for fluids and fluid
mixtures with the three-dimensional Ising model prediction of 0.312 f 0.004.
It is barely possible that systematic and unrecognized errors have produced
experimental values of 18 which are consistently slightly too high; it is also
possible, although unlikely, that the limiting value as one approaches still
closer to TCwill prove to be somewhat lower than the average value extracted
from measurements in the range lo-' -=
t < What is presently lacking is
any convincing reason why the fluid exponents should differ from the Ising model
ones.
4. The theoretical argument for a PUsingularity in the diameter of a not-
exactly-symmetric coexistence curve seems convincing. It seems unlikely that
experimental measurements (especially on liquid mixtures) will either confirm
or disprove this prediction in the foreseeable future.
To these personal opinions, one other should be added. Only very carefully
designed experiments or experiments on entirely new kinds of systems are likely
to add much to our present experimental understanding of the critical region.
It does not seem profitable to continue to pursue experiments on more and more
systems simply to make tables like Table 4 longer. Coexistence curves are, of
course, needed for other purposes, but only an extraordinarily well-designed,
well-executed, and well-analysed experiment will improve upon the estimate that
/3 = 0.35 k 0.02. More information is needed about the exponents 8, 01, and
7, but again only very careful experiments are likely to yield this.

I wish to thank my colleague Professor C. M. Knobler for many helpful discussions

and suggestions. The research on critical phenomena at UCLA, some of which
is reported in this review, is supported by the National Science Foundation.

Note added in proof. The typescript for this chapter was submitted to the senior
Reporter in July 1974. The subsequent three years have seen considerable further
progress, both experimental and theoretical. No attempt has been made to bring
up to date all the tables of experimental determinations of the various exponents.
Three developments however deserve comment : (i) the unexpected convergence

7 The theory behind the scaling hypothesis suggests that this should be true for all situations
in which the range of correlations greatly exceeds the range of intermolecular energies. For all
the fluids studied the energies are of course shore range.
Critical Exponents for Binary Fluid Mixtures 27 1
of theoretical and experimental values of the exponent p, ( i i ) further attempts to
measure the exponent a including reports of the (1 -a) divergence of the co-
existence curve diameter, and (iii) the recent theoretical and experimental interest
in tricritical points where ‘classical’ exponents are found.
(i). The exponent /3 for the coexistence curve. It now appears that neither the
experimental measurements on binary systems nor the theoretical evaluations
of the limiting behaviour of the Ising model yielded as precise values of the
exponents as had been believed.
Two new calculations have been based upon renormalization group theory.
Kadanoff et aL115 using variational approximations report /3 = 0.3367 and
y = 1.214, while Baker et aZ.116-11s obtain from the Pade summation of series
derived from the Callen-Symanzik equation /3 = 0.320 -t 0.015, y = 1.241
It 0.002. Both new values of are substantially higher than the previously
accepted one (Table 1); evidently there is still considerable uncertainty about the
exact figure.
Critical exponents are properly defined only in the asymptotic limit T -+ T,;
thus, at non-zero values of t, there can be higher terms in a series which yield
(e.g. from In Y against In t plots) an average value of an exponent for a particular
range of t which differs from the limiting value. Such ‘range effects’ have been
looked for but until very recently not found. Two new experimental studies of
fluids suggest that they are real.
Hocken and Moldover 117 carried out precise optical measurements of the
equations of state of Xe, SFs, and C 0 2very near their critical points (- 1.5 x
-c t < 5 x Extensive data analyses yield values of best values of /3 lying
between 0.321 and 0.329. The fit for /3 = 0.34 is markedly poorer.
Greer 11* has made careful new measurements of the coexistence curve of
isobutyric acid + water and has reanalysed earlier measurements on the system
CS, + CH3N02.813 119 She concludes that the volume fraction 4 is the preferred
composition variable, that the limiting values of p are appreciably lower
(/3 = 0.324 -t 0.018 for the first system and /3 = 0.316 2 0.008 for the second)
than previously believed, and that the behaviour further from Tc is consistent
with ‘extended scaling’.120
Further work, both experimental and theoretical, is still needed, but it is much
easier now to believe that the experimental exponents for fluids and fluid mixtures
are identical with those for the three-dimensional king model than it was when
Section 7 with its ‘tentative conclusions’ was written.
(ii). The exponent a. The exponent a for the weak divergence of the heat
capacity Cp,T,l,l,the thermal expansion and the diameter of the coexistence
( Y ~ , ~ ,

curve continues to be elusive. Most experiments still fail to distinguish between

a logarithmic singularity (01 = 0) and that with a small positive exponent
115 L. P. Kadanoff, A. Houghton, and M. C. Yalabik, J. Stat. Phys., 1976, 15, 263.
116 G. A. Barker, jun., B. G . Nickel, M. S. Green, and D. I. Meiron, Phys. Rev. Letters, 1975,36,
1351.
11’ R. Hocken and M. R. Moldover, Phys. Rev. Letters, 1976, 37, 29.
118 S. C. Greer, Phys. Reo., 1976, A14, 1770.
ll@ E. S. R. Gopal, P. Chandra Sekhar, G . Ananthakrishna, R. Ramachandra, and S. V.
Subramanyam, Proc. Roy. SOC.,1976, A350, 91.
l 3 O F.J. Wegner, Phys. Reu., 1972, R5, 4529.
272 Chemical Thermodynamics
(a z 0.1). Two systems which have been studied carefully are nitroethane + 3-
methylpentane l2* and isobutyric acid + water 121-ln3, mainly by volumetric
methods. Although an anomaly is clearly seen, an adequate analysis must include
an adequate treatment of the normal background. The divergence of the thermal
expansion is best fitted by an equation for the volume of the system at its critical
composition
V(xc) = A +
C t l - * + Dr. (47)
C is normally much smaller than D,which compounds the difficulty. Only the
measurements of Morrison and Knobler 123 are extensive enough (1 x <
t < 1.6 x to give a clear preference to the positive a over the logarithmic
divergence; they find that the ‘best fit’ value of 01 is consistently between 0.08 and
0.14.
Two recent papers 124, 125 report observations of a divergence of the diameter
of the coexistence curve of the pure fluid SF6, although no analysis is given.
Gopal et ll9 have reported similar divergences in the systems CS2
0 1 . ~ ~ 9 CH3N02 +
+
and cyclohexane acetic anhydride. They fit their data over a range 3 x <
t < 10-1not to an equation like equation (47) with x’ + X” substituting for V
[e.g. equation (9) with a Dt term added], but rather to
x’ + X” = A c C P a e x p (-G* t”) + Dt. (48)
The values of (1 - a) deduced range from 0.87 to 0.92, corresponding to a
from 0.08 to 0.13, and they see no difference between mole fraction and volume
fraction although the magnitude of g = (K - V,)/2% in equations (31-34) is
about 0.07. The conclusion (pp. 255,256 and Table 3) that the ‘213 effect’ swamps
any chance of observing the (1 - a) divergence unless exactly the right composi-
tion variable is chosen (a conclusion which is strengthened by including a
substantial term Dt for the ‘rectilinear’ contribution to the diameter) evidently
does not apply here. The mathematical reason for this is that when the data are
fitted to equation (48) very large values of C ( I 574 and 616, respectively, according
to our definition?) are deduced. The term CP-“dominates the form of the
diameter in the range < t c but the ‘attenuation factor’ exp (- G* t h )
causes it to be nearly negligible outside the critical region. Without this factor,
the substantially rectilinear diameter observed outside the critical region requires
C to be small (as was assumed for the calculation reported in TabIe 3). No
theoretical explanation for the attenuation factor has been offered and the
magnitude of C is surprising. In the absence of a detailed error analysis of these
measurements, careful studies on other systems, and a better understanding of the
special form of equation (48), the experimental evidence for the (1 - a) diver-
gence of the diameter must be regarded as tentative.
I2l S. C. Greer and R. Hocken, J . Cliein. Phys., 1975, 63, 5067.
lI2 H. Klein and D. Woermann, J . Chem. Phys., 1975, 62, 2913; Ber. Burisengesellschaft Phys.
Chem., 1975, 79, 1180.
G. Morrison and C. M. Knobler, J . Chem. Phys., 1976, 65, 5507.
J. Weiner, K. H. Langley, and N. C. Ford, j m . , Phys. Rev. Letters, 1974, 32, 879.
z“s D. Y. Ivanov, L. A. Makarevich, and 0. N. Sokalova, J.E.T.P. Letrers, 1974,20, 121.
lx Ph. Kohnstamm, ‘Handbuch der Physik’, Springer Verlag, Berlin, 1926, Vol. 10, Chap. 4,
Section 45.
t Note that in this review t = (Tc - T ) / T , whereas in reference 119 r = T, - T.
 Critical Exponents for Binary Fluid Mixtures 273
(iii). Tricriticalpoints. Fifty years ago Kohnstamm 126 suggested the possibility
of a critical point ‘of second order’ in a ternary fluid mixture, a point at which
three coexisting phases become identical in contrast to an ordinary ‘first-order’
critical point at which two coexisting phases become identical. Experimental
evidence for the existence of such points in ternary and quaternary mixtures was
found much later by research groups in the Soviet Union, in particular by
Efremova and her colleagues, starting 12’ in 1963. An analogous phenomenon
in symmetric systems was predicted, from a phenomenological theory, by
Landau 128 in 1937 ; the corresponding phase transition observed experimentally
(as the junction of three critical lines) in mixtures of 3He and 4He and in certain
solids (metarnagnets) was called a ‘tricritical point’ by Griffiths.12DSuch tricritical
points have a special symmetry which is absent from fluid mixtures except for a
+
few unusual cases (3He 4He or mixtures of optical enantiomers 130), but the
more general case of a ‘second-order critical point’ is now generally regarded as
an ‘unsymmetrical tricritical point’.
Unlike a symmetrical tricritical point, the unsymmetrical variety is not the
junction of three critical lines, but the limit of a three-phase line between two
critical end points. At the tricritical point itself, the three-phase line has shrunk
to a point on a single smooth critical line; moreover, because of the lack of
symmetry in the phase diagram, the tricritical point lies at a composition outside
any of the nearby three-phase triangles of which it is the limit.
Griffiths 131 has given a phenomenological (Landau) treatment of tricritical
points which expresses the free energy as a sixth-order polynomial in an order
parameter (which is some suitable linear combination of the physical ‘densities’,
e.g. the mole fractions). The scaling properties of the singular part of the poly-
nomial lead to four numbers = 5/6, $2 = 4/6 = 2/3, $ J = ~ 3/6 = 1/2,
#4 = 2/6 = 1/3, in terms of which various critical exponents are expressed.
Because this is an analytic (mean field) formulation, these exponents are ‘classical’,
but it is believed 132p 133 that for experimental tricritical points in three dimensions
they should be. (‘Nonclassical’ logarithmic factors may exist, but these do not
alter the exponents.)
This prediction has now been subjected to several experimental tests. Lang
and Widom 134 have studied at saturation pressure the three-phase region of the
+
quaternary system benzene ethanol + water + ammonium sulphate. At
each temperature the three-phase region consists of a stack of triangles lying
between two critical end points, and the locus of the compositions of the three
phases is defined by a single line in the isothermal composition tetrahedron. In
the phenomenological theory there are three characteristic dimensions f l , f 2 ,
lZ7 I. R. Krichevskii, G. D. Efreniova, R. 0. Pryanikova, and A. V. Serebryakova, Zhur. fiz.
Khim., 1963, 37, 1924 (Russ. J . Phys. Chem., 37, 1046). For more nearly complete biblio-
graphies of the Russian literature see ref. 134.
lZ8 L. D. Landau, Physik. Z . Sowjetunion, 1937, 11,26. See also L. D. Landau and E. M. Lifshitz,
‘Statistical Physics’, 2nd edn., Pergamon, London, 1965, p. 193.
lZ9 R. B. Griffiths, Phys. Rev. Letters, 1970, 24, 715.
130 R. L. Scott, J . Chem. Soc. Faraday ZI, 1977, 73, 356.
131 R. B. Griffiths, J. Chem. Phys., 1974, 60, 195.
132 E. K. Riedel and F. 5. Wegner, Phys. Reu. Letters, 1972, 29, 349; F. J. Wegner and E. K.
Riedel, Phys. Rev., 1973, B7, 248.
133 B. Widom, in ‘Fundamental Problems in Statistical Mechanics HI’, Proceedings of the 1974
Wageningen Summer School, ed. E. G. D. Cohen, North Holland, Amsterdam, 1975.
134 J. C. Lang, jun., and B. Widom, Physica, 1975, 81A, 190.
274 Chemical Thermodynamics
and t3of the coexistence region which shrink as the tricritical temperature is
approached with a proportionality to powers of (& - T) which are
(1 - = 1/2, (1 - + 2 ) / ~ 4 = 1, and (1 - 43)/44= 3/2, respectively. Lang
and Widom found the exponents to be 0.42 f 0.02,0.98 f 0.03, and 1.51 f 0.03.
Creek et al.136have studied the three-phase regions of the ternary hydrocarbon
systems methane + n-pentane + 2,3-dimethylbutane and methane + 2,2-di-
methylbutane + 2,3-dimethylbutane. These are ‘quasi-binary’ systems 136 in
the sense that the relative proportions of the two higher hydrocarbons are very
nearly the same in all three phases. To this approximation the relative volume
fraction 2 of the two higher hydrocarbons is a ‘field’ variable playing the same
role as a (physically impossible) continuous variation of the solvent-solute
molecular interaction in a binary system. The phenomenological theory predicts
that the difference AT between the temperatures of the upper and lower critical
end points (or the corresponding pressure difference Ap) should be proportional
to (Zt - Z)&”k The observed exponents for the first system were 1.51 f 0.02
(for AT) and 1.42 k 0.03 (for Ap) in good agreement with the ‘classical’
+31+4 = 312.
It should be noted that the combination of these two sets of experiments
suffices to determine all four of the numbers $2, c # ~and
, 44. While more
experiments are obviously desirable, there seems little reason to doubt that the
exponents governing the change of appropriate properties as the tricritical point
is approached are indeed the classical ones deduced from the phenomenological
theory.
lz5 J. L. Creek, C. M. Knobler, and R. L. Scott, J . Chenr. Phys., 1977, 67,366.
130 R. L. Scott and P. H. van Konynenburg, Discuss. Foraday SOC.,1970,49, 87.
9
A Bibliography of Thermodynamic Quantities for
Binary Fluid Mixtures
EDITED BY C. P. HICKS*

This bibliography brings together contributions by different workers each

responsible for a different property as follows:
Excess Gibbs energy K. N. Marsh
Excess enthalpy A. G . Williamson
Excess volume I. A. McLure
High-pressure C . P. Hicks* and C . L. Young
phase equilibrium and
critical properties
* Present address; National Physical Laboratory, England.

1 Introduction
Mixtures are entered in the main body of the bibliography in the alphabetical
order of the first component according to the Hill system as employed by
Chemical Abstracts. An index is provided so that all the binary mixtures involving
a particular substance can be located. The literature has been covered to the
end of 1974, with some 1975 publications also included.
The Hill empirical formula is obtained for a particular compound by placing
the elements in their alphabetical order unless C i s present, in which case C is
placed first, H second, and the remaining elements in alphabetica1 order. The
system is extended to mixtures and lists of substances by placing the substances
in the alphabetical order of their Hill formulae. Substances with the same Hill
formula are ordered alphabetically by their name.
For GE,HE, and V E the approximate temperature (or pressure) range of
measurement is given. In many cases the specified property is not actually given
in the quoted paper, but can be derived from the results given there. Isobaric
studies of vapour-liquid equilibrium at low pressure are not quoted as a source
of GEunless the temperature change across the composition range is small.
The high-pressure phase-equilibrium references contain a wide variety of
studies of fluid-phase equilibria. References have only been included if they
give some values above 1.5 MPa. The actual variables studied are noted and
short explanatory comments on the nature of the system or results are occa-
sionally given.
2 Bibliography
275
276 Chemical Thermodynamics
m-
I
n
a > X
=I c Y
zr x
-
0 U
Y u
0 c
t
o
a
E z -.
a 0
Y
N z
> c
0
a 0
*
z 0
Y E
L
c.
4
a x n
X
YI
a
Y
z
YL
t
0
a
f
W
U
-
Y
n
II
X
0
z
x
0
4 Y 4
Bibliography 277
I- c c
a a yo
0 0 w
-
a
m
n
yo
z
a
I
--
c C
a a
L * n- t - -
ri,
w
n m - cam
-...
V
-
S
z c - f e
-B

- t
UNS-

e
c1

e
UN
W
w
z
-0
S N 0
-
rn
8 Y
W I
U
TO-u
I) N
* -
O L
U
H
c.
Y
a Y
a
3 0 3
N
f <
z I
N
n
w - u
a m o
w
-. m
w .
a m
0"-
W E
t-I-.-I
U I X C
Z W I -

-
Y
t
3
0

*m
cc
u a
L

W
W
a n -. an
IUO- m wc.
r-.ar D
.
>
n
S
I #-
t
W Z
-
0

a
W
a a
>
I
3
5 t
W
'i,
a
I-
8
J
N
N-.
-
- Y
O D U z
OD
7
*
I
V
I
u
00
0
N
I
L
n
I
.d
U
Y
a
U
 Chemical Thermodynamics
z
N z
W
U c
R 02 ( O X Y G E N )
280 Chemical Thermodynamics
L L
w W
z
u "
S
-
f
N Y
9Y
. -I
< U c
u
c c
n *
e
* 0 n
N
Q
-
--
E
x
N u e8 u
n
0
3
z
W x
Y 7
W -0
ma=
9 Y
--
(0
u -
n
PI
I -
E I s--
C J 00.
00 0 y1 W w
c x 0 Y
S I
U -0
0 . I-
. - # .-
h 0 a
m 0
I-
3
x
Y
c
w

* w
o
.
-
-
m
n
P
Y
Y O
Z
- zF*
x a m
-
0 -
u w w
el- a - 1
o w
a a u
w r
w -Y w Y Y
Q < U eu 3 A J
Y uu u Y
(L
r
e- *.
x
2
-
" W
. n
n
<
x
U

a
*f
U
-"u Ie -
b- n-
u
n
a n
-
X
a
U
- T X
aa
<*
WW
Y Iw I
.
a
*
c
-
7.1
0
n
N
w T xc-
w u -0.
z -3 c
a x
-
a a
-
w w w
0 >> >
W >
Y I
b u -t > >
UIW >
VI try s
w
I
T In
0
a
0
3
J
Y
U
+
z
1 YI
Y C L Y
7 a
3
E D
aa
ao
O d
3 r
.AU
.. W
z
Y
2

a
9,
I.-
-
3 d UY
L I-
0 (vz
-.
-
W
r
Y

Y I4
nr
L-.

I
m
T W
S z
W "
0
N
.c 4 Lb
Bibliography 28 I
U Y
n
0 a
0 U
N
0 n
c
>
N <
n z
z
Y
J
c
N
m
0
Y
U
0
.
b
!
l
W
U
S
a
w
z
4
-n
a
0
z
Y
i u
I
6
m
rl
0
c
z
n
a
<
I -I
Qu
o e
v u
Y
0
n
N
c
0
0
0
c N
Y
Y
I
Ly
u - c
- s -
O F Z
- w -
0
u u
s
u s
u
282 Chern ical Therrn ody narn ics
--
m
m c
-J
Y O
c
CI
00 an
t m
-
0
z
Y
J
*
t
Y
V
<
Y
.
L O N
c-8 E
a
axv N
V
Bibliography 283
U
N
N

D
z
a
-
c
u
-
a
u

0
m
o
m -

N
n
n
W
c
0
0
0

T
Y
I
U -
n
4
U

W
N
0 n
0
m
L
J
v-
n 0-u
rZCYlc
-- >
I
n z
Y
Y
u
t
m
I
W
z
r
r
t
a
Y
a
W 0
t -I
I
N I
U
I
N 0
-n D
N
F Y
-7
- 4 Y
9-
V S
6-
.tY)
u0
2 a N N
J 0
u z
eu -8 n
zz
I
N
U
I
N
U "
N
a m CI
284 Chemical Therrnodynamics
-
n
U
z
U a
z w
s-
-
A
o n
Y
0
E
a u
Y
W
c
I
V
r a
u
U
X
*
z
f 0
L
-
-
4
I
n
I
c
w
A
v
I
I
Bibliography 285
.-
t c
w-
9- I
Q
n
N
c
7
n
C I .
NW-
n n
N 1u
-8
0
N
z
-
&
a
0
n
a
Y
0 zc
I
(D 0.t
X Z
n nu
0 u u
u a U 0.00
Y
286 Chemical Thermodynamics
r
U
t
v
-
N u
N
0 N
c
n
a U
Y)
c
c N
n
*
UL
-4
.-
o
l-
-
I
P
a
Y
N
lL n
u
3 0.0
a
0 f
R
U U W W Y uluw Y Y W W W Y
I>>Us5 x x w >>>>>>
W
c
4
I-
Y
U
*
-4
-
I
c
Y
N
0
D
I
-s
v
4
Bibliography 287
U
R
>
n
I
C b b 6 - - .-
t- -I-
CNOW *
n n Y
.N .eu I
YIC U k r v1
..
U
-
W
X
Y
u
I-
W
I-0 0
n o
L f c
O X S
m u u
au u
-
<
Y c .
II
"
J
-1
288 Chem ical Therrnody namics
w w ww w
e I S > >
Bibliography 289
-
r-
s
N
c
a
m
>
a
-I
I
U
e
c
Y z
z w
W N
N r
z w
YI m
a 0-
0 a w
Q C Z
0 -yI
3 Z N
290 Chemical Thermodynamics
<
u
v c
U
2
0
U
Bibliography 29 1
0
m
0
a
0
N
-m
-
Q
CIC)---
-N
0
0.0 0 0
N
-a any. -Q m
c
N 0 -0.
c - c
0 -
u
m
0
cr DY
292 Chemical Thermodynamics
Bibliography 293
T
N Y)
llD I CI
0
N U 0
Iy
J L
*
N
Q
0
0
r
U
n
0
*
c
D
4.
a
4
.
IL
r
<
e
L
n
a
(u
>
.c
c
p:
z
a
*
0
z
r0:
u
a
0
I
3
Y
-
Y
Q
o
N
v
Y Y W Y WWLU
w
I I T 5 2- we
;r
a
D -
o w
z n Ly
2 - z
% a <
- 0 X
W A Y
T
0
-.
u
*
u
2
2-
S
F
5
PI
rvz
Y Y Y
1 7
U
t-
I
P O 9 c c
U
au u u
*
vaa Y .c aa.0 4.
Y
-.
J
U
U
N
\o
P
II C 8 H 18 ( Z ,2 40 T R I M E 1 H Y 1P E N T
ANE)
ML

HE
VE
VE
VF
WE
VC
VE

RE
VE

GE
He

He
VL

vc
VC
V t

GE
UP
ME
VE
VF
VE
Vf
Bib Iiography 297
nJ
Q
*?
u
r
- -0
n
W
L
N
a
-
u
N
a
c.
w
2
Y
0
In
5
w
r-
N
I
U
8-
n n
za
< z >
0
f
S
nI
L
0 Y
U N
€ a
U < - -
% u
f a 0 s
0 m
4
z
<
U
0
a
m
(b
u0 0-
0 0
-w
n
.
I Y
0 Yc
.
U <
S
a 0
E 0
In
0 a
0
1
-
U
c
A
<
-
Y
0
Y
a
c1 z
0 .c
a
N
Y
Y
I>
Y W >
Y u
6 B
PI w
I s
Y 5
U 0
W Id
W Y 5
Y
IL: I
0
a a
a
..
*
3
m
i

A -
-1
0
2
0

I
z
0
E.
*
i
e
L
--
ww
Z Z
0
-I
8
U
a -.
*
zU
O v
-
4-3 nw
c- L
m I0
.L
U 0-x
oa
u < ri Y Y a N t
-
Iy
u U n n 0
0 c 7UI
0 n a
0 0 Y Y F
-
ia
z 0
u -
Qz8
Y Y
Y
a.
U O
w OY 0
0 O X 0 0 0 0
-- -
r-c
-
n ry WYC N 0 9 d 00
rY I
--
cn
N N N N N U
S N N <
I I z S T x x s
0 0 C 0
c N N N N nu
r
0 8- u u u
v
U V V u u U u v
4
a€
QE
HI!
nE
VL
Vt
OE
Bibliography 299
W I I
Y I I
-n
az
A
Y4
w w w had w
o>a -c" 0
- - -
W
w - w E
z w L 4
I
c
UI
W
T
0 L
4
.
z
LYC
L-.
< S
A
0
z
- Y I
z
3 m
0 S
c
YI
a
c n
A
w
I " r,
L
Y Y
m
I z
N
au
c
.E I H U
n
a
0
z
u
300 Chemical Thermodynamics
N
f
X L
w e
D - Q
ru
N . .
.
(P (0
U
N .
I I 7-
n n w
0 .
00.
nC
O
-- --
n
D - e
* a
c
u u
0
U

m
A
. " V

n
)
a a
*
E u
Z?
. I
I Y W
a -
* - Dw
a
LII na
b
c).-Q
0.0c~0D.ooo.o 0
NNmNNNNNN CI

-
u
>W
W > S
Y I
Y -
I
w
I I
ry
0
a
0
3
A
n
a
F
L
9
n
0
a
0
2
x-
-=
uu
*
-.t
LnX
A
U
N
U
*
Bibliography 301
..
Y
U Y
0 0
ao a
ov I
s v
-1E N
U
-ea
CI w
n
J
V
S
U
302 Chemical Thermodynamics
m
-0 m
z m
u r n
N
F
c
x
9
n
0
Y
Bibliography 303
U
0
IA
u
uo
0 5
Ly
U
T
u W oN
u
4
>
0
Y
- xan YlyIYyl
>
YT s- >> w
a
Y
4
b
w
U
U
1
2
t U
I x
0
t
w 3
y
Y mu.
02
N N
0 0 0 0
U 'Q
S I -0 Q
T
I
mn m
0 U
V
u u
W
 w
0
P
V t

Ct
GF

rE
VE
VL
8 C L H 8 0 2 t t i ~ Y lA C E T A T C I HE
VE
b c 4 M l O O ( D I E T H Y 1 ETHER) CE
OE

HE
nt
HE

YE
ME
nE
HE
ve
vc
VE
VC

Vf

VL
Bibliography 3 05
E
W
I
U
W -J
m
0
w
U 0
U
c 4.
x z
w U
U
f CT
0 m
N
T
n
U
a
0
Y
rc
c
3
a-
N
Y
a=
w
I
L
306 Chemical Thermodynamics
r
0
a
I)
e
z
X
U
u
I
-w
-
0
0
r
U
0
Y
U
0
Y
z-x
Y u
YY >
u
Bibliography 3 07
#
.- Y I V Y I Y
CI
a
Y
I
L
Y c
m
I
n
-.
a-
c
-. 6 c
W
> E Y 3
X -c m
F c
-a
%I-
a-

Ly
L 0
w
x
z
Y

U
*
J
X
a
-
z
0
Y
0 Y
m.z 0
0- Q
a-
c.x
r
0
0- 9
I I
--lQ r
U- u
Q u
0
x u
-
Um
U u 4
CL
J
U
u
E
308 Chemical Thermodynamics
Y
Y
0
3
Y
c
n
P
N
n
. . IN
Y
Y w u w
x s I >
I-
- I u
J
Bibliography 309
-.
ur
z
*
T
c
YI
0
a
-
0
-.
c
Y
N
I
N
T
U
rd
310 Chern ical Thermodynamics
7
n a
.r c
-
I
n
W
If
L
W
&=
OD-
-14
L
W L
f -0
a bcr
L
0
w -0
0
n
1
3
J
2
U Y W Y
S T > I >
Y
-
a
U

*
rr
a
Bibliography 311
L
0 m
r)
..-
Y onura
s - u a 5
n
c
312 Chemical Thermodynamics
r-
c)
P
e
z
c
L
A
0
3
n
N
CT
I-
.-.
5
N
v
n
Y
W
YI
W
z
*
S
*
Y
0
IL
0
J

-
W
c
z z
U
0
M
I
c
Bibliography 313
Y I Y w
L W i
l
-
x
0
c

-a*
- 0
I 5
8-0
314 Chemical Thermodynamics
0
U
c
z
0
N
n
N

-
-
u
i
m
Q a
c
N
L
e
c-
I Z
n c N
w -
-4w
.-
0.3
U
x4
-
a-
1 U
01
rt0
4*

YlMt
ZZ8-
Y Y - -
a
u
-
X
us<
t u
U

I
I

3 - w
2-u
0
tN-
-0
TI --
N
UU
z
O N NW
0 0 0 - c c
Q Q
x x XaZT 5
Nc.
v v cL0.0 0
uuu u
Bibliogruphy 315
I
Iy
I
U
a
Y
yt
n
I
U
3
<
F
L
a
0
a
.L
Y
0.
-
x
z
0
a
z
7
0
(D
5
4
w
0
I
0
V N
U
Bibliography 317
>
0
Y
*
u
X
u
X u
X
U
I
w .
I
0
I 0
L
I
Y
I
-~
X X Y W X X
0 0 0
9 ID O r
I s L X
c1 c) 9u
u u uu
m u
w
c
00
Bibliography 319
Y)
>
a
L
N I
U
Y)
w
0 u
z
Y n
n
L
w-
r
0
I
3
(P
c.
L
YT
L
m
Y)
a 0
u
m L
E m
w m
0
X X
Q 0
L1
n
W
n
ry
v
T
-
B-
YI
2
a
m
Y
03
a a
W
S YI
a
Y -w-
U
-
rru
X N
*
Y
%
4
c
L
Y
n.
0
A
u
>
0
-I
*
r
c
z
w
Y
N
r
8
u
Q
rz
320 Chemical Thermodynamics
n
7
D
*
7
*
I
z
w
a
Y
3
- .
m u
P I
I D
u
U z
Y
Bibliography 32 1
N
c
0
F
N
c
0
(r
0
2
.
0
W
0
-8
a
U
w
0
*
U
W >
a
a *
N
W
w
c
f c
S
322 Chemical Thermodynamics
--
4 ...
7
0
0
+
m
I
f
4 *
ec
Y
Y
E
Y
u
2
Y
U
u
S
Bibliography 3 23
>
c
3
c
0 m
m
U CL
0
CIL 8- 2
a f
3 .t u >
n
z
4
0.
0 -. 0
*
6 0
c a
0 0
E
2
U
>
0
c
ol
0
*)
n
2
UI
0
L
z-
< a
t-
U W
L
Y
"
U
S
324 Chemical Thermodynamics
F
n
N
-
y.
rn
-
0
N
w-
r.
P
9- N
a
. 9
U
0
s-
n
a
3 I
I
-
I L
f
I
U
0
0
I

.-
N
L
n
2- ff
T
La 3
U
I
c
4
0
4 -
a
a
n
4
>
w 0
-.
0
Y
a
3 0
Y
L
4
n
an
z
w
7. A
0 t
Q
OL T
c
Y
v
U
x
N
U
Bibliography 325
a
*
0
F
Y
<
c
I
m
- -
L I
y
*I
c
a x x
u v a
L *
-.
L
W
w
Y
c.
9 -
W
W
a -w
z+o
Y
Y)w u WYI Iy UU YY)
x > u X I > +TI>
-
n
m
0
X
-1
I 0
LL 0
3 >
n
3 0
d 0.
0
2
w
r
W
A
t
-*
a-
W
Ye
Y
N on
O L" 0
0 9 z
E- ss
9 4
T
"
N
S
CI
O n
U
*+n
uv
n? c
326 Chemical Thermodynamics
z
m
Y
.
>
d

- m - m
x
U
s
U -
Y Y F
I N
-
a s
s-
E W Y I
L U .

--
cc
Ykl
w
W
Y V Y -
Y
I n
-
n
N.U N N
0 N
0
0 0 0
.DQ Q 6 a
x s I
X I
uu 3 u
Q - u
.
vu u
Bibliography 3 27
< u -
u v u
m m m
e?
Y
I
c
w 3
r
c
2
328 Chp m ical Thermodynamics
Bibliography 329
n
Y S
Y
V
z
=s
0
U
n
>
E
-J r n
W
u
I
I
u
Y
U
.
c
t
0
U
-.
L
Q
z
N N
I 4
r> t F
.Y
L n
4
L
X
OL
w
X wI
A
Y rr
z
N
Y t
YT
I
w
in
X
0
a
K -
w
z--
Y
a
U
u
I
330 Chemical Thermodynamics
Bibliography 331
332 Chemical Thermodynamics
n
w-
k
c
N
c
U
0
v-
-
S
I
Y
...
N
.L
E
3
N
I
<
>
Y
Y
P
m
Y
1
---
nn n
w
w
w
334 Chemical Thermodynamics
n
0
J
m
9 0
0 0
0
9- 0
c
S 5
ma Y
u
f S
U
ml
>
m
).
h
L a
L
N N
L 5
Y W
f z
T I-
u U
a
3 m
3
I I;
- .--.
Y
"
Y
- -
z
e
P
0

e
L
0
a
*
--.
I II
--
a
a a
3
3
Q a
0 0
-
I
D
3 3
0
-
X X
I- w.
Y
336 Chemical Thermodynamics
xI
.
L
-.
u
t I
U
C a
L Y b
C =I L
Y
F
V w
0
S
Bibliography 337
0
X
0
3
0
t
L
c.
z
Y
338 Chemical Thermodynamics
c
Y
n
c.
-
A
0
Y
W
z
0
I-
<
Y
Y
0
x
CI
U
e
Bibliography 339
Ln
N
19
N
n u
m
0
93
N
#
c
*
<
a
<
a
c
U
r
u n
"
I.
n
*L
0
r
-ll
Q
c
n
z
340 Chem ical Thermodynamics
c +
Y Y
Y

-
Y - -
I
a
0
- a d
- 0
U N
U U
Bibliography 341
”
m 0
-
c
0
0 n
c - o -a
L .no* w LI c.
-
U -coLL U
c. r
U
a e m
Lp
01 a L
I a
Y
a
0
0
4
c -
U
a
I
-
0.
L
Ly
x
V
m
IYIL
X
*
0
X
J I M
-
X
Y
c
u
ow
ar
-
v
I
0-
sz
-I
un
-9s

0
,“ -
D
f
-
a-
r-u
4

CI
u
Y 4 Y u w ..=BY

m
U
r)
3
U
I
N
U
342 Chemical Thermodynamics
a R.
3 3
x
N N
I N
I
U
n
4
c
U
1.
--
Iy N N X
YI
Lp 0
m a m U E
X
Q c.
0
c
0 -m X
f
ry
CI 0
OD
r L
W
U
L
a
r
4 Q
<
a
u
z -
-S 4
Y
Y)
I >
yl a
y1
u
T
N Z
Q Y
-I
0

u -
U
*u -.
m r n a0
W
.-
-
a z
a -n
s
4
a
.
PI -
I
I
- 0
-
PI
a zu
- 0I
Y w a
n
c- c. zo
UJA
-
a
a
L -.I
I
t .
U-
a
-
y1 0
m
0 z M
r
W
A W
U
"
-8
X ?-a
7
U f
.
I n W
0 .B)
W
z
Y
-
n
a
a

*
t
L
2
U
z
0
fc
Y

-
c
Y
w
T

z
w
U
a
0
*
a
w
c
a 0 *
I
e
3
v
N Y
-8 N
0
t a
I P
0
I
I 0
az
n n Q
V w
U I
fI n
u U
BibIiograp hy 343
.
1
3
N
f
.I
U.
c
0
n
PI
...."
X Y
<
z z*
3 3
U
z
n a
-1
w
Y
U
0
U
u
on
a m ..
-u-
4
a
a
W
s: J
I - t S
e a m
..c
9
I
0 03 U
U YO
.I
c
0
Y -0
9
Y Y
Y 1y wnl ty w w w m U Y
> > T I 7 > = x x z x
a -
-
bs
- Uu
I W
0
rc
0
-1
I Y
u
I
N
I
L
--
0 0
U
-uI
0-
0
U
Y
2- N
0
f
0 c
.
-
f0f
US I
U
U an- 0
U
-u UU
-44 4 4
3
V
N
I
N
U
344 Chemical Thermodynamics
Q (0.C D M
Y Y Y

P
N
Y

D O
NN
Y Y

P D
NN
- Y -
e
N
--
Y
z
W
Bibliography 345
t-
I .
TU
W
t

-
z
e
X
1
Y
0 X
k W
I
m 0,
u
4 0
Y I
Y m
Y
N
0 0
z
II
0

0
*
t
U
*
9
t
346 Chemical Thermodynamics
I - c n
J
c
u
0 a .>
- 0
n
E 0 *
z r t
O f .
r.
0
L
-!
P
F
v
Y
Q
9
x
N
U
4
w
P
4
w
P
oa
Bibliography 349
.-&
W W
W x
z U
n
A
a
Y
t
c
3
Y
S
. .
0
t.
Iy
N
V
350 Chemical Thermodynamics
*
0
c\
n
m
N
c
0 c
m
v
c
U
E
b- y
I -
-
Y
N
w-
*
x
U
Bibliogruphy 35 1
Q

-
E
Y
I
fu
c
I
N 8
I
Ir 0
nr
N n
Q
-
m m C
U U
0
r 0
v
a
CI
c
f I L
I I c.
t
Y I
X Y
I
N
I
L
-
N
n
I
m
I

-
m -w
.s-
0 - w
r z
Y y .
u u sa M
uo -I-
n
z. z
n s>
U Y
N=
-0l
CII
z
00r
-
*
Y
u
10u
0
352 Chemical Thermodynamics
c
n 0 0
0
c)
n
e
0
c-
*-
0
s-
c
o
0.
111
Y
0
W
I
* x
0 -
Y
z a =
-
N
YI F
U
C

W
I
F
w
W
m z
-I
t
+
3
m
1
7
1
U
c
0
F
"
c
-4 .
I) 4
u 0
.* 4
X
N
Y "
X
N
Bibliography 353
354 Chemical Thermodynamics
>Y
U c- 3.
nl
I
a
0
I
0
J
a
*
e
*
2
T - w
B -
-a T
9 r u
- I H
uz-7
u-ww
- 0 0 -
-t
F z
? c
Y T I
u m 9
( 0 v
Y
.a v
N
o
-Y
I
N
U
C211402 ( A C E T I C 4CIl))
8 HZO (UATER)

w
v,
v,
356 Chemical Thermodynamics
.
a
*
X
D
w r x s
H L Y Y Y
m x x z
Y U U
.
I
--
W
J
a
1
u
a
*
0
z
-
z
Y
w
/
4
X
U
I
T
Bibliography 357
0
ar
I
*
x
"
u
T
c
N
N
*
D
=
t
*
0
U
L
Y
0
L
w
m
z
I
t
c
n
z
n
-a
n
*
z
a
I
.
1
L
Y
z
U
u
m
L
"
8 #
Y V
.. m m
-
0 0
I-
E I.
-r
Y
<
z "
n 0
;c x
0
0 a a
a
a 0
-I
0
J
A x
U U
I;
r
s B a
c S
Y 0
ut Y
2- n
-
8
N NU
O Z
Z Y
JH
V T

u w
O Z
Z W
-.U
YT
0 -8 -t
f
J
S f z
0
U 0
U
u
* W 4
358 Chemical Therm ody namics
-
II
u
X
>
c
.
I
L
f
0
Y
E
Ir
0
z
m
*
a
Iy
L
Y
Q
0
r
U
n
A
Y
7
4
"
U
n
7
Y
w
h
I
0
V
8-
Bibliography 359
360 Chemical Thermodynamics
U
z
P
0
-- - -
an
I-*
-I
u
o
w
x

a
s-
Y
-
Y
1c
0
c-
U
c
U
Y
0
I
n
0
I
Bibliography 361
m
. .
.N
D
L
Y
I)
a
0 -

m
n
z
Y
S
0 -
-
-cU a
3 I
a E r
n N m U
'N0t N t
V L
U
-n a
YI 0
Y
0
Y 0
X
c-
> .1
u
I=x P
U
-
aa
t-
un .-
c-
Y
(Y
00 N
c
2
cea Q
I S
u
U
..
.u *
I
u
362 Chemical Thermodynamics
V
m
L
L1
Y
u
f
*
O-J
m -
.-P
-il
V
z
Y
Y
z
V
m
0
m
0
A
-
v
I
-I
U
m
E
9
U
Bibliography 363
n
nI
0
YI
0.
n
U

-
c
0
a 0
0rm
0
V
w w w w
X I I I
364 Chemical Thermodynamics
n
-n
.y
0
R
N U
(Y
0
0
c
n
)r
T
0
N
x
u
4
3
a
t-
Q
Q
a
z
-0
0
N
Y
m
I
I
0
(Y
I
z
~
c
c
z
(Y
a
- I -
0
a
z
N
U
Bibliography 365
n
c
0.
a-
-
>
I
3
Y
c
<
c
1)
- uv) u
n o 0
I
0
T
W W Y w
x x x 5
-
Y
z
Y

-
U
z
t
a
Y
I
3
-
4
0
t 0
L
Y
0
I
C
U
o@
366 Chemical Thermodynamics
n
- ..I
Q
e4
.
I
Y)
U
E
I
9
J
-. c
uu
-
m
z -a
- Y c
Y
0 w
M
3
0.
N
0
- 8 -
2:
-
am Y
a WJ
L 3
z
w 0
Y
z a
>
4 1
z
0
U
*<
3.I.
n r
n u
u.
<
N
a
z
b-
z
a
la
c -
*
=
t
m
S
y1 I
n
4 c-
t S
4
a
w
s
f
-cr 06 ac CYIUUOE
0
0
T
"
lw
Bibliography 367
0
0
a
N
HE
GL
tt
HE
YE ZOl.
GE AND

Pt
fit
nE
370 Chemical Thermodynamics
.
0
N
r
s
v
Y
n
*
Q0
-CI
-a0-
-n
a
I CWYC-N
n n m
0
I
Q
I

-
Y
L
X 0
c
T I
rv
a
z
U
b-
73
m
Y
0
L m
r o
-
I
O
-NI 4
U
d
n
0
Q
I
ht
0
Bibliography 371
-
yc
.f
0
c
nr
0
0
7
*
n
N
8-
N
Ln
II
n
f
*
A.
c
P
a
0
Y
L

a
m
rn
I

.
II
I
3
W Y
I
L z
n
a P
0
I 0
I
u
3
10
a
*
7
n o n6.0
cococ
NNNNN
372 Chemical Thermodynamics
l
c
u
-

-
- 2
r
N
Y
Y W U W Y Y
0
W
> a x > > w
c
z
Y
m
N
0
d
Y
n
U
Bibliography 373
0
0
UJ
-
-
u
a
0
- u
r
-
o a o
n a n
* L r
a - a
o n o
n x
m
-.a -
W W L U
w
W - t u -z 5
I r
z 0
4
a
0 l
(u
Q
.m
-
a
I). Y
t C
c
3 3
9
Y "
0 0
N
I T
CI
c -8
3
U
"
I
n
3 74 Chemical Thermodynamics
z
W
"
*n
7.
e
a t c
Y
UJ *
Y
n
a
U
a
-7
1u
( 1
m
7
U
L
w
z
*
Y .M
*
W
-x-
a
c
a
-
Y
7
Bibliography 375
Y Y Y
0 s -
A
-
I)
c
a
A
LL

ff
L
0 . n
L
a y.
a- I
I w
t
z
U
-
U
c-

d
-
0 Y
<
u
--
w
Z Q

a
0
0
*) 0
0
oai c
I-x I
. -x
UP! u
-8 -8
U
*
UU
vat
4 4
0
9
I
n
u
376 Chemical T h e mody na rn ics
--
c
ObC -
DP
c- e
L
n
s
c
7
Bibliography 377
0
n m
0
c eu
f
a
z
t
a
0

--
I
a
t
c
3
0
7
U
Y
w ln
0
-r L
6
x
"
-1
t
PT
f
Y e s - - * 0
Y I
T I " z_
Y
0
ry
v 7
T S
0
U d
U
0
378 Chemical Thermodynamics
e
y.
z
"
4.
W
0
I
7
Y
N
N
T
h
U
c
u
Bibliography 379
- --
0
R
a
C
9 0
m
N
9
0 n
..-
Y \ J
9 0 -c
N*

r *
z
Yz
a
Q
-.O
90-
L s-
Y
380 Chemical Thermodynamics
a
L;
N
Y IV
Y n
v
-
_I
0
I
N w
b z
;z
4 U
X
n Y
9 s
X
a T
T O N N - c
U r r - l 1 0
O ~ X P O C I X C

- Y
4
0
c

Y
N
r
S
O e O O N -ON n 0
V
U U U P I r u x U
a a -
-4P-Ic -=- ¶ . I 4
N
0 0 0
Q Q Q
I x x
ct n n
v " U
Bibliography 38 1
-
Y
Y
I

<
>
0
a
=
n-
m
0
n
Q I
0
m
10 a m
m
a
0 Y N
e
z
a z
z
4
n
<
L
w m Y Y IU w w
0 s >ff w 0 Y7
382 Chemical Thermodynamics
I
a
*
S
a
In
0 N
OI
U
-
2
n
w
W
YI
0
<
?
=
m
t
I-
<
T
-
c
I-
A
MI
UI
7
I W
t a
I-
U c n
z
W Y
N X T
7 Y 7
Iy I
Y
e
Y Y
.o
Q
0
r .-
T X E
Q u
0 c
u
"
w -B er
Bibliography 383
-0
N h
0
-
N
O
-- I
- 0 - -6
n r u cu
F - c
T I Y
J I I \ ul
v u 0
384 Chemical Thermodynamics
.
I
c
n
8-

r
Y
f
.
.r)c
0 -
S
Y
S
U f
x
I
*
A
z
(I
-
I:
m
-
S a
m
5
U
ol
z I I
m t
a Z n
c
a -
t
=
0
x
W
z
.0
c
-
x
a a
I- c
.-
I 4 *I
W
4
e
3

U
I-
P
U
r
0
..
0
0
z
4

U
I -
W
n
m a c.
-4
> _.
0 "
Y L
X 3
-
U
e
U
x

0
r
-I
UL
a f
t z
I
U
0 1 Y 1
a
0
L
N
e z
v
NNY
a
>
a-
-
J
f
0
U

m
c
I
.V
D
-
N
N
I
0

ooa-
OQ
CIQ
L f
O*
u u)
(YC

a
t
Y
O V u
Bibliography 385
/-
w
c
0
a
c
rui
T N
8 - 0
-Iyw
m
c
m
U
386 Chemical Thermodynamics
m
0
U
J
0
U
L
.* - Q
- r
- -
- =
M I
- 3
0 - f
O N U
-
W
c
4
c
Y
"
4
t
2
n
I=
U
0
v
-
N
0
L
c
Y
P
IF

-- c

w
u
S
n
u
Bibliography 387
-
n
U
, Y
m
Y,
LI
a
--
7
4u
I
V
*
Y
m L
U
o
-8
N
CI 7
0
*
D
P
8-
--I
cn
0 0
f. Y
N I
>N U (cz
6" D U
N .
-a
Y
Q
OT.
c
%- .
I
0
I
N
388 Chemical Thermodynamics
Do
0
u -
- r
n n -
Q c x
"
W
1
"
4
Y
13
0
0
-1
Y
.a
N
.-
N
T
v
c
.- U
;
r
Ic
5-
n
U
C 3 Y 9 Y ( 1*'l f T h Y L E 1H y L P '',I h ?I
390 Chernical Therrnodynamics
a
0 -
c .
u
U
Y
W
7
?.
Y
3
c
I
Q
0
-5
c
00
X
n
u
Bibliography 391
Y O
m 7:
>
Y -
X m
a x
4
#-- I8-Y
.-
0
u,
a *nJu
=
w Y
...
.
Ii
..
YI
<
F
P
-Ma
a
w
c-
U
3
v
0
z
N
-
am
392 Chemical Therrnody namics
J
c
Q
0
re
0
3
IL
A
n a
LL
W
Y a
1-
8-
II
-
-
w
00
Y
a-
.
0 Q
Y
N
I
IU
4
Bibiiograp hy 393
0
U 0 n
c
U
P
tc
r
Y
z
W
?
0-
a
w
-<
Y..
. . I
Y
O Y
u u u
Chemical Thermodynamics
. -- r
m
n
n
0 -
Y)
N
z
-
Y
L
LU X
4
as
3
a
3
2
I
O
Y
2
U
In
-c
0
0
Bibliography 395
N
c
n
0
.-.
w
4
*
Y
I.
0
9
"
t
2
0
I U
zL>:
4
Y Y
h)
c
I
0
U
m
X
J
U
Bibliography 397
x
w
8
'L
- 7 L
Y
-
N
f
u
0
.-
U
Q
I
"
id
VE
VL
HE
HE
VC
VC

Gt
QE
HE.

ME
ME
HE
VF
Vt
VL
Vl:
V€
HE

HE
YE
6L
Vt
nE
tt
VE
YE
G I
Gt

GE
HE
YE
Bibliography 399
n
n
m i n
;
c
.)
0
I f
y.
X
U
J
n N
> c
n I 0
v
.r) a
0.
c
- 0
-a uIr.*
. n
(Lt
z
-
u
;c

c
X
w
S
0
-
Y A
z
N u
>
n U
0
c
N
c
9
X I
0
U 9
0
* m
X
U
U
400 Chemical Thermodynamics
0
n
c
Y w
I a
0
0
f
.D
*3 a
7
I e r
I D
I -N
I L I
-
tp
4
L
-
N
0
.o
I
U
0
Bibliography 40 1
n n CI
n
0 n
Q
r
U
W
a
..
v)
u
n

z
U
a
t
I- m
n
m
U V
z
0 0 --
b
c n
>
1- <
E
n z
a c
0 L
-a
F A
-1 u
-07
LL
a
y1
b
J y
a
W
Y
.
L
4
0
z z
z
-
0 w
u
- C I I
Wr-
Y
5
U

z
-
a
-L
-L
T
4
L'
..
W W W J W
W
W YI
5 >
5 >'z
w
W
3
J
0
c
v
.-
r.-
N m
T I I
0
U -
" L
0
4
402 Chemical Thermodynamics
Bibliography 403
r
x r
f
0
U 0
U
a
U
*
L.
r.
F
.
0
x
0
a
e
t
--A
0 0 4
uu-
me=-
--
I
a

4
u
5
.- -
cc z
I 6 I
- I I - I
uu u ou u u
00 0 0 0 0 0
rrn n nn a
l-l- * ** F *
.I I I . I .
-
y1 Y) WYI
I > -53.
.
I
ru
c?

I
0
J
n
w - - 8 - 8 -
u u u u u
404 Chemical T h e mody nam ics
- -.
*
CL
0
I
m
N
a
w
c
YI
X
E
-
0
r nc
uu
N
I
u
m
0
>
Y
a
0
0
0
.
L
Y
L
z
Y Y w wm w
x > r >> >
D D D
-
B
u L

-
A
0
o x c
:
Y
f r
"
4
0
U
U
Y
a
- -
n a
- -
z
0
7
I
m
a
Y

-
b-
us
OD-
10
N
Q
T
r(E

VE

4E

nE
YE

YE
ME
ne
CL
VIE

ne
HE

P
8
406 Chemical Thermodynamics
u u a
a
m
c
a
Y
2
x
0
Y
a
U
Bibliography 407
U
t S
<
e u n
-
I L
.1
Y

W
m
D- -z
ov
r 6
U
I
w
"
L
u *)
c
H
2:
I
a
w
x
U
*
C
Y
m
3
n
-
G
0
.-.
u-
x
.
V

D
9-X
.--
-*
U
-t
0
en
D -
u
-%
m
n-l-
.-
x
2- = x Y >% Y x x
yi
I
V
w
a
I
7
Y
N
N
T
r
U
sc
408 Chemical Thermodynamics
A
Bibliography 409
L
4.
0 - 0 c
22
z o z
> N U
I-nz
a.--
a m-a
T
W
N
0
Q
I
Q
0
410 Chemical Thermodynamics
u
L
> 3
e- >
0
3
E 0
c
z
U
3 7
<
n -
-
7
r a
a -
a x
a u
T
C .
w u
n
W
. I
-
o
&
4
>
7
0 u
0
-
t-
W
-
0
W
4

a
A
2-
5(
Ly
X
0
u
2
8-
1
*
u
I
z u a
U >
-
L
W
x X
Y c
f T
W
A
t
0
Y
Y
Po
b
I
u
-5
Bibliography 41 1
n
0
n Ic
CI
9
I -
r
n *.
*I
UYI 0
u
X
U
m Q 6
-
T
X
n
8-
-
Y
n
J 2
3 0
I I;
c N
I;
z z
4 e
a Y
0 0
a a
p. n
-I -I
* *
T. T
L- c
.-
w LY
-S .r
A *
O N
-
rn
N
412 Chemical Thermodynamics
.y
N
U
0
0

-
* 1
n
a
c.
*
n
>

r-
N
U
*
c
c.
3
Y
0
S
c
0
e
C
m
0
a
a
Y
a
0
L
r
U
4
c
I
<
Y
a
4
2
Bibliography 413
-
z
I
Y
#n
c
J
*
a
3
-
m
t
L
P
c E
CI
b-L
0
2?-
c
r
0 0
<V
w I--
*3 --.
0 X
L i l Ir
a
L
F
z
4
Y

m
D O
mm
u-
-u
a
-
a
# W
Nr I-
N 0 1
urn
*3
0
Q nz v
-
T
U
U

I
m
.
V- 0
N
I
414 Chemical Thermodynamics
a
p.
0
.-
0. 0
t
Y
-
n
U lL
Y
 8 C6H7N ( A N I L I N E )
Ic C ~ H ~( C YZC L O H E X A N E )

C&HlOOi 10L)
416 Chemical Therm ody nam ics
9 .
--W
00
- 9
m-
n-
N
N
u
.
Y
w
u u - x
x .
I -
% 5 P
W Y W W YI Lu
>xITx I >
YI
L
U
e
a
I
f
8
7
v
0 u t
c CQ
x 0 I U
.
Iu
L N -0
S OV
v
c -? I
0
8-
Bibliography 417
-
T
0
L
c
I
Q
-.
f
n
z
W
L
U
N
z
yt
iTI
0
a
0
-
J
w
Y
-
n
Y
0
U
.c
418 Chemical Thermodynamics
U
N -
U-
a 0.
m
c1
c
0
e
N
n
>
I
a
. -
*
a
S
W
t
I
L
t
I
b
7 -
n
x
u
*
I u
X
N
I
U
- a
Y
u
n -
N
c.
t
U
3

-
c
a
I
t
0 W
I I
>>
Y PL >
Y
-
W
4
m
X - 5
Y
Bibliography 419
- z
yi
L
a
YI
z
W
-
ry
Y a
a 0 a
m,
U 4
E U
E
a
D .
I
m
0
-
n
U
L
4

n
t-

a
m
n

a
m
n
0 0
-
2
> N
a
0 0
P 0 0
(r
W
*
n
.-
m
0
8-
c
T 4
c-
n
4
r t
W n
W
x
U I
U
I
U r?r
c
Y
0.
w - 8 - I a .D
Y Y L w
P F
? d z -z 5 0
I .r
X
u c 0
0
u " -0
z u
Y)
-
I
0
a
c
a
z a a
a 0 0
W
2
-
z 3
c
/ n
0 *
<
c
n c-
LI
n 0
m
a b-
I Y u
r +
c
0
ff 0 a
U I
4 2.
U
a
0
0 n P
rc
4 w
W u.
w
a
a
3
4
W r x
2
,.. -
3 c c
a
0
Y Y
3
6 o[)
c
- r
n
3

v
n
3
Y I- c
-
Kl
0-
N
00
0
N
v -
02
0.
N
2-
W Y
% 1
pt
- W
I
c
I 1 8 1 - I -
L
4
a
a
w
-an
Z-I
U Z
ZI
Z C
U k
#-7
w-
JO
F l
LN
c -
Y N
z-
U
aN
+--
Y X
con
-U
z m
h
UYl
f
f
u
J
420 Chemical Thermodynamics
90
n 0
U c
F U
r-
U
9 N -
"
n
a
w
n
-
W
m
;c
0
A
n
0
D
I
U
0
2
-
-I
3
Q
YI
4 az
X U
u A
I
I
0 m
0 0
n
U
Y
Y w W U W YU 4L
I ma
T W
f Y
T U
D
w u W W S 11.
-
I
0 - -
-
w
2 u
k 0
N Y O
- u *
Y X I
m . n m
V U Y
.. . .
e n i t . , 166R; 16, 141.
422 Chemical Thermodynamics
2
t
2
N
d H E ( 2 9 8 JOI K, A Y A Y A , BULL, CHEM, SOC, JAPAN, 1961, 5 6 , 1271, 1278,
? HE (298 K ) I K a A M A Y A I BULL. C l t f l a S O C a J A P A N , l 9 6 l 1 341 1271, q278a
6
d
1
2
424 Chemical Thermodynamics
r - c O
X I S
Q O h
u u u
Bibliography 425
-c
0)
x
U
*
-
V
I
<
f
I
Y
z
U
e
c
m
.
-B e
n
w
I
0 m
r
t, W
X L
Y
0
c
a
I
a z
p1
t S
U
0 a
.I
<
z
- 8 m -
r r z - -
LI

-B
"
3 W
.
Y a
U
0. z
0 Y
t a
1.
1
I
0
*
z
z
0
#-
a
W
-3
ou
c-
.-m
0 .-
E
c
S
m
Y
426 Chemical Thermodynamics
- a -
* u
I u-
l - Y
T I
m m
u Y
C 5 H 1 0 0 2 (N-PROPVL ACETATE)
t H23 ( W h T E R )
P
N
00
C5U12
430 Chemical Thermodynamics
432 Chemical Thermodynamics
L
n - - a
Y YY Y Y W W Y
> >I s o x >>
-
n
t
I
Bibliography 433
W m
W
a? L
w
< V 0
z
W 'T
Q
(Y YI -.
c U
V -. L I
I
T
L *
0
z a
2-
<
a
-
Y c
*)c-
n
N.-
0
1 - w-
O N Y
I-*
L l0
r
CI
0
.N
D
Y
n
Y
Q
Y >
W
UI
z
*
0
0
7

rn
I-
YI
fr -
Y
z
X
Y Y
- r N
z
I
P
0
0 tx-
Q c
S WI-
I -.YO
C
U a O N
U Y -1s
t
m 4 c)S*
0
N
.-
I
m
u
P
w
P
Bibliography 435
m
- . .
W Y
-
Y V V
Y
c w -
s s s
Q
v
u ....
9
Y
9
V
436 Chemical Thermodynamics
c?
I I
U -a
a -u
*
S a
nu*
P
w
4
438 Chemical Thermodynamics
a
1
W N -
4 O D ~
N U
- -
e n
Q U
. -
0 . -I
-. I -z
Q
m
0
9-
W y1 YYI UI
w
t
T > >
u -
-.-
Y
0 4
a =
t Q
w v
cn c
“
440 Chemical Therm ody numics
U
9
0
n
Y W Y Y W W
X X I I S X
e n
m
Y
-
Y -
I
-
@- r-
Y d

I
Y
H
n
x
Y
r(leuomu nu
A
IL U
N
s
n
9 9
U U
Bibliography 44 1
-
I
X
U
c.
b
U
.
c
m
CI
0
LD
L
0
-4
a
0
Y YLUIw
I =LIZ
442 Chemical Thermodynamics
Y W I Y Y W
>>>x 'Lx
Bibliography 443
I I
L r r
-
Q
L
c a u
<
CIO W
c
m n
V >
-a -m m n n
I
I
S
L U
I *
I t
c a
-* sI -
- w <
YY"Y
<
*t=
-
>or
-a
.
3
x a-s
.>0
%3 1
x . .
m m
tmw
f l W
T w ww
5 2 0
W 0
yt w
x o w
W U 0 Y Y
Q IwI S
-
A

-
a
0

Y
L
U.
X
Y
0
S
m L
az <
X
U
-.
V
c.
L
0
H
z
N
0
-
m
U
t
Y m
X I
9
V u
Is
-.
U
m
X
a
V
P
P
P
Bibliography 445
-.e
cQ Y
.-m
0
N
Y
I
YI
*I
a
0
w
W 7
z a Y
YI a N
-1
H
7 -
*
z
U
3
a
-w w Y Y L
rym 7 YI z z
2 3
w o
UI
N
zc
a
*
-1
z
YLU
NN
N O 27 a
N c
Y tz
2 - w ww
W- 0 Y Y o-
0 m
a -
N 0
0 n
a - o w- v
b - 0
Q
O X z L f X -S
3-n 0 c
0 0
U 0
U Z
UUQ
2 u u
-m
Y t.8
k m 4 Nt.0
z?
0
L n
YI
L X
u
0 0
U
 P
P
o\

4 C6H7I ( A N I L I N E )
448 Chemical Thermodynamics
a
W
r
P
*
m
-.
0
1
Y u Y Y Y u Y Y YIV W Y Y W Y Y
I x I I S x L I S L L L X x s w w
v f (288 1
V E CSGS I
19 3
VL (296 I(
VE (538 1
VE (298 N
VP (SO1 c
19t
VE (291 c
19t
ve (298 c
19t
VE (296 c
VE (298 c
VO (2V8 1
CH I
V E (298 1
VS ( 3 0 8 1
vt (
V L (291,
V E (298
tva
s
191
v e t 278 1
19 ?
V f (291 1
(C 8
V E ( 298 U
v e (298 1
197
V t (SO1
VI
VE KLti

ve
VI Slt.
VL
VI
VI
VO
VI

V l
ve

P
P
\o
450 Chemical Thermodynamics
-
S
U
I
z r
I

0
"
x
. .
>
YI Y
S s
Y wI Y
S Y
f m V
L
>
GC
GE

GE

re
HE
HE
HE
HE
ME
HE

HE
HE

*I
Vt
YE

VE
YE
YE

VE
YE
VE
VE

ve
VE

P
m
c
P
ul
t
4
Bibliography 453
0 .
N N N N
N N N N N N
n W
UI n
U n VI
U n
U n n n
U n
J
- -
U
U J U
n n n n n n n n n n
--
n
.o
Q
0
O
n
r
-
Y
r
U "
W
T
Y Y wY)wYw.nwWwIy Y Y Y Y
T T X C 9 W T W X W W X O W U ( D 0
-
n
0
U
u
I
v
U
u
-
4
U
n
0

0
*
3
N

-
U
c
N

U
n
U
N

Y
L
w
3

Y
z
Y
3
N w f
Y z
UL z
Y A
0 c
A
z 2
a
- -
Y w
m
0 I-
n
0
z
-
n
0
N
z
Y
ID
*
-
Y Y
Y
0
N
0 Y
-u I u
2
z
m a Do
f
I c
S I T
Ic
U
U
c
ur u
- t
454 Chemical Thermodynamics
c - c - c
o w - m- c
C I N 00 N
c M N c
v
c -0
0 u - -n
*
N
N
N
0
no L
-n
m c r
a
0
r.D
-
c n
em
9 0
c
Q
0 c -
-
U
9- 0 -
n
<
-.
Y
IL
3
H
T
I
W
a
J
Q
9
Y
BibIiogra p hy 455
0
(Y L
I
I
u
) -
a
I:
m
*
m - i
U
n
I
c
e
-
--
n
*nc
456 Chemical Thermodynamics
0
-
o n
-
W 4 0
N.- c
P-
YI
7
c1
m
N - *
z
z
I -C
N
Y
cz
W c S
3 r-- 3
W
w
0
A -0
Q
-.
- .I-
0.- 0 0
-
a a
c
v)
e
Y
yt
mo
nre
N -T-
- c
w
0
3
W
- I
c
0
c
u
0
UI
Y
.G
U
0
n
a
Y
-
x
I
0
Y
co 0
c
x
u
(D
4
Bibtiograp hy 457
b
c - -a
- 0
I
c
z
w
n
3
P-
I
#-
r
Y
L.
a
-.
#-
a
N
N-
U z
Y
c
.
s
a
u
9
S
Q
U
 P
v1
00

VI
VE
VE

L(E
VE
YE

HE

Vt
GE

at
HE
nE
VC
VC
P
ul
W
460 Chemical Thermodynamics
G re

. c .
o
N
m
m - C
c C O
0 0
7 7
. c
0 - 0
< Q
z -
u
U c
N m
-.- e n
0
a#-rr-0
C Y Y Y I -
-...
I-
x
- -rl -N
rr-n-n
-
0.a n 00
0 N O 0
Iv*.)NN

Q
-8W
0
N
xx- x- .a"" Y Y
V Y Y
U U W U Y Y l U l W Y
w
> -s
W >>
W Y 5
W Y W W Y W W Y w
t - t > c w C . f U W S W T W 1 . x>
Bibliography 46 1
a
3 r
N
L Iy
- -
- a c - - 0
I . -

x
U
u
0
a
u r n < a 3
0 Y F a
a - a I
b W
.
)
C
-J
n
a
-
a n
a
-
w W S
(9
u
z
W z
W
n I
a
*
I
n az
s
Y
z
m
3
m
=
Y
3
0 - t
L u K
w a
t
u
0 W
N
Q
=
w
L
U
a
I3
w
c
-
0
S
x
0
a
m j.1
r
3
3
-
m
U
c 0
.
.I U
<
L
r n
m
a e
E
0
z
U
0
W
a
0
-1
I
V
4
a
t
I
W-
A
2
F Y
z
0 z
a
-..
4
1
Y a
0 K
Y
t
<
u
3
N
0
r
0
0 0
n n N
X
V U
W
462 Chemical Thermodynamics
U
N
Y
>
Y D
Y Z
Y W
W
-
Y
z
Y
I
f
U
Y
a
-z
3
N
N
X
w
U
P
V
Bibliography 463
-
D
a
e
m
.
f
n
UI
U
5
w
.-
c
0.

Q
c
9
N

?
- -
0
w-
"
'L .-
0 .
I v
N D I n m
n
-J n J 0.
n
*
r
X U
c. .-
lK
U
N
n
Y
*
7
I
a n
0 e
a- a
a P z
- Y) 0
I
H
I c
x
a
4 I W
w W . -I
0 ==&
U N I
z 0
Y
+
2. U
-
U
-
a
a
0
m
W
m
a
-
-I
I
L
UI
X
U
/ v
n w w u
-*ew
w
w
W Y W
u s s --
C fL
-
W
z
0
z
w
x
a
*
0
Y
..
a
U
U
w
e
Y 4
3
0 Y
m
I 0
a x
N
0
C a@
0
0
I
Q
U
464 Chemical Thermodynamics
c
n
0 0
X
I)
.
u
-8
YL
a
0
<
c
0
I
0
v
n
z - -.
Y
a
c
.r<
m a
0 -
cc)
o X
w
0
2
--
c n ww
w e
Y
u
w w w
=>> wu.mtwu
e)U)UCDx
Y
I
w
x
~ w u wY
I>)> >
Y
f
Y Y
wx U
xJx
wxw >
Y
- - a
a
W
t
t
U -
Y
-
Y
L
-
Y
c

0
a
U
z X
Y C
A
* - #
- +
U
Z L
Y I Z
3
h -< .
V Y
Bibliography 465
-
n
U
a C
n
U Y
0
L.
n
0
z
0
0
A
-
a
<
U
A
A
m
3
I
t
w o x z o x
-a
-
U ? O 0
z v
-
W * T
m c.3
-.a-
-
>a
a au T X W a
Y F> es-+ Y
c nv wU
wY3 c
c
L
n
00- v
-.-
X ' E O 0
a
um w
-a N
S
4 a74 I
c c c
S I I
- 9 9
U
u u
466 Chemical Thermodynamics
L
;c
c rr
Ic
a
m
a
I
n
h m
c
.t r
u
n
c
-0 c
m.- e
7
r
-0
r-n
a
* -
-n
-
a
U
-I
I
L
-0. *
0.- 0 0
r . r . 0 0
I- mrrn z I
vb-
S-
0 0 0 9 a
NNN Y W
a t
I
-3
U Y
.-.L
m z r
4
w
a*-a U
X
m..
n -
3
-w
r
* S
"U
n
U r *
V
r
n U
Z S
0 - W
A
U
on
o
A r
.-
n w
-a
*ti- 0
z
q w
v c
u -
u
.Z
AS
2-
311)
NI1
N
N
-m
0 - 0
z
ZIn
0
N
w
w Y Y YIUIUUUJ w w Y u w w m w w
5- I, > o x s > > L I I z f I > f
m a m
a W L n n
a I -
n J
F
J
0
I
<
X
m
X
0
J
U
F
U
Y
0
n,
w
S
d
U
4
Bibliography 467
a
t
I-
.-
m -
-
.
I
- >
- s x
a a
I z
- a
o w
z
4 - 2
I <
.
UI
I C T
n a
u
>
x
Y
I
0
A
"
r
u
Y
F
S
P
U-
4-08.. Y
HE
HE
HE
HE
HE
HE
HE
HE
He

Mt
nE
UE
ne
ne
HE
HE
WE
RE
HE
nr

P
QI
W
 p.

n
-I
L
0

-4
0
r
E
I*
z
m
Y

X I < < < < c z

nmm m rn m m m
-
N
0
-m
Q
o V
Q x

- I

N
*uI

U C
3
ua
S D
f
D

-
"
L

n
m
e-
r

rn
L
D

T
Q

I
c
-s
I
0
-
L
U
0
-
L

X
"
n
f

rn 0 W Q
z 0 u-4
a n 0 0

0
D
D
-4

OLP
 . -
CHI.< F R A N C ~ , 1968, i360.
HE (325 K); K, S O S N K O W S K 4 = U k r l I A I A h A 4 D H.V. KEHIAIAM, BULL, ACADI POLOH. S C l e r SLRe

P
4
U
 P
4
N
GE
HE

Mi?
*t

ME
VE
G t
IiE
Ht
*t
I E

YE
YE

(IE
ME
VL

VE
G I
GL
HE
HL
HE
Hk
VE
VE

HE
a!!
UE

HE
VE

YE
YE
1973 2; i68,
TC(X)I C.L: YOUUG, J . CkE'l,r
AUSTRAL. 1972. 2 5 , 1625.
HE ( 2 9 8 K I I 4. ~ O Y C I C K ~J,. c H E M .
T H E R M O D Y N A M I C S , 1971, 6 , l b l .
HE (2Vlr32S K > I G.W. LUNDBEREI J . CMEM. EHG. D A T A , 1964, 9 1 195,
OE ( 2 9 8 TO 328 K ) [ D.V.S. J A I R A N D O,P, Y A D A V , l N D l A N J . CHEM.. 1971, 9 . 342.
H E (298 I 0 8 I. V E B E L Y AND J . P I C K , C O L L . C Z E C H . C n E H . COHM,, 1969. 5 b , 1 t P Z .
Bibliography 473
>u
30
oa
-
N -
r-Q
a rn 0 -
a -.uZ& a -
n a
U
L
L
0
m
a
W
e
c
- 4 CI c-L.
Y*-CX Y
0 -JJ
m .-m da
-
u.
n4
u

0.x
N -
-a
w c
I >-
I e
I- W
--I
U
z
Y
474 Chemical Thermodynamics
c
Y .
.--
n
m-

m
.-
c -u
8
n
X
r)
Q .
- x
Nrr
rr
W
L
Ic
0
W
n
0
a
I
7
Y
a
N
S
N
"
e
e
 Vt
PE

Ge

HE
nL
YE

P6

nt
vt
GE
Gt
I c S O H 6 2 OQIJALANC) GE
6L
8 C L 4 S l ( S I L I C O V TETRACHLORI-
DE) Gf

HE
Vt
Gl
HE
JE

G I

HE
ME
476 Chem ical Thermodynamics
N
w
L - 0
- e m
< m -
R
a
N
I
>
4
a
r
W
9
W
L
0
-
0
0
7
>
- 0 0
OZI-
0
-2
.
-II
D
0
m
0

I
a
0
f

-
I
c
I
P U
-. e
t - E
If u - c
I- Z W W
 6
CdHl2OZ (OUTYL ACETATE)

C6HlZ02
8
&

8
6
c
6
C4H1202
6
C6H1202
6

C 6 H 1 2 0 2 (N-HEXANOIC ACID)
1120 ( W A T E R )
L 6 W l Z O Z (2-YETHYLPROPY L ACETATE)
~8W16 (OCTmlmLNC)
CbWl202
6

CbHlSCk
478 Chemical Therm ody na mics
*
W
a
.
*
n
.-
U
E
m
Y

I
a
u a
Y
a
.
C
-
.-
X f
0 c
P e L
a8
r w
n m
Y
.- -
w N
7 -
X
U
f
U I
Q
U
n
- -
a
c a
0 -
0 .y 0.
U F I- t m
I- n
c c e 0 z
0-.0
- f 0
A
m
0
-
a
a
c
U
2
A
m
3
4
X
PI
3
0
a
a
n
3
Y
-*
0
-
w
t
Y
3
-
-#
D-
0
9
I
c
U
u
6 C l Z I Z I N (PCRFLUOROTRIBUTYL-
AWINE)
b C12'126 ( N - D O D E C A N F )
L Zcr c L O O C T A N E )
c e n i i-
Bibliography 48 1
D!
Q 0
U
E
*
> a
0
W
a 0 u
L
u r - u
0
0
a.
a .-
(0

II
0
T
u
I
W
w
u
a
z r
"
U
I 0
Y
>
W
LL
a
5
z
4
3
r
0
c1 m
4
3 I-
<
3
U
L
wm YI-
>s > u
I 8
a - 2L
c w
o n
Y I -
= a
0 0
 Chemical Thermodynamics
0
z
Ic
z
0
w.
a
Y)
a
n
n
Y
U
0
a
0
f u%
v
w
U
I
J
u
z
w N
P-
D
I-
u
a
-a-
Y T
z n
I
-
YI
n
2
>
L
I-
-
E
W
e
#-
a
4
N
N-
U L
Y
m a
c
I
u
10
w
Bibliography 483
I
L
L. c
.-
r. m m
m
N m
U
I
t
Q
-
Y
W
0
N
Y Y
wwml w Y Y w YI \u w u lywu 111 w Y Y Y w In (Yw Y W wm U Ytu
>>> > x w z w r >c >>+ > r o w Q z x S E 11 3cs L >>
I 8
Q
U

-*
t
484 Chemical Thermodynamics
--
*
m a
L
w a a
S
4 m 0
a Q
vl 0
w- a
3-
m UI Y Y u
> > w w >
B
<
U
m
a
c
t
U
m
z 0
0
C
Y I:
Y 0
-1
c1
c v
f
Y
(I
S a
7
x D
*
Y
* Y
w
z1 u
Ic
en en N
Q
It II
p 0
S
U10
rtm ru c
U
::
*
C 6 H 1 4 (NmHEXANE)
6 Y204 ( D l N I T R O t E N T E T R O X I D C )
C 6 n i 4 (ZBMETHYLPENTA~E)
6 C7t14 ( T L T R A D E C A F L U O R O ~ E T H -
YLCYcLOnExANE)
& C t H 8 (TOLUENE)

P
00
m
3
Bibliography 487
+
a Q
3 c
0
P
Y
Y
<
2
a 0
N
&
Y
c
f
w
"
x
-
n
0
v3
> S
a
T W
x
U
w W Y w w
I W I T >
488 Chemical Thermodynamics
0
&
D N
0
n
c I
r
 V E ( 2 9 0 a ) ~A.N. CAMPBELL AND S a y , L A Y , CANAD, J m CHEr., 1973, 51, 4005.
TC, P C I J e Y e K U E h E l J r P M l L e M 4 G . 8 1 9 0 3 1 6 , 6371
(T,P)I PI K O H N S T A M Y A N D JIT I N M E R ? l A N S , V E R S L A G . AKAO, WETENSCHAP., 1914, 211 763
(CHe~.AB1.,1914,8,1016).

C C l O H 2 2 (N-DBCANE)
8 ClOH3OOJSIC (PECAMETPYLTET-
RAStLOXANO)
 P
\o
0

HI!
VL
VE
C T n l C L O (ECNZOYL CHLORIDE1
6 C 7 H l (TOLUENE)

nE
VC
vc
VE
VL
ne
492 Chemical Thermodynamics
9 -
c o
a r u
L I r
z
3
w
J
0
E
k
0
Y
a
E
Y
u
0
n
a
CI
z
0
-
m
I
P
\o
w
P
\o
P
 8 ClJHlOO (BLNZOPHENONE)

C7H80 (4mMiTHY LPHENOL)

6 C l l H l O (2rMCTHYLNAPHTHALENE)
C7*1L
ctn1r
C?H 14
CTHl b
CtH14

tfH 1 1
P
W
4
498 Chem ical Thermodynamics
Y
n
T a
U n
c
M *
az r
a
S
n
=
N N
Ca
-
z
c
c
a a
B-
-
L.
N S f
2 z X
n
U 0
U
n
c
2
W
c)
w
0
Y T y.
0 <
U u
0
D z
0
m
-1
Y
wn
W Y
c
WJY>
I
I-
w
L
n
J
>
w I
z I-
Q
a
n
,r
-
I
W
a
I-
< UB
c
0
0
U N
N-
0 u w
7 mL
<
r.
P
T
4)
u
-5
4
4

6
Y
L
6

L
L

L
2
II

&
L

L
VI
0
0
 H
il

Jm

AS
502 Chemical Thermodynamics
m
u
u u Y Y C Y w
Z > S W " s T
L
U
S
+
Y
w
L
-.
t
S a
e
Y 0
n
a
A
t
2
z
Y
a
-
8
N
.y
c
L
f
U
Bibliography 503
m -
nr
C O * -
xo lvxsc
l u I l U U
E O
-w
504 Chemical Thennody na mics
* . -
J; j
0 0 0
vlm m
** >
aa u
C E n
*-
u a
u
a
uu 4
.. ..
m n
z u
<
a m
z i
v)

.LY
.I
mcn

Y
-
n
.I
a n
z z
I .
v)cn
<
z* z
.I
an
zz
.
)
7
u <
4
I <
z<z U I< aa
z
a
a a a +
c 4 a aa a arc
t-8- arc ac
c a c
a
C C c cc t +I-
I IL
c 0
a
f Q
p.
2
4
>
D- I
e
3 &-
z-
w
m
-
a
sUI
(Uc
m c
.+ -3
Q
w
U Y
Ot
- a-
V S
U ET
oa
-
U-x N
-
N-. N
d
7. S
0
8-
0
U
W g
Bibliography 505
I
A
Y
u
w
0
n
c
L
I
a
- - t
o o m - N N Y U
r - c a
-
*
u
a
a
.
r
w
I
I
w
l
U
0
a

u
D
N
w
a
1 x 0 L
u u 4
L
a
.L
Iy
. - I X
u
a
* >> >>
.
4
G
a
c

u-s
n o
u-
a a
u u

<-
o n
uu
IXLL
a <
.
Y
ol
L
.. . -
LLIL
v)vJ
z z

Ye&
an-
z z
u u u 4.5
a a x a a
* ++ CD-
0
n
..
n o
I
=a
.
n n Il-r
m Y x n
- w
a~ a 2 a
D b X
f U N -
U - n
UL w Y Y Y I Y I -
I T. >> - 2 -
D I D @
-2 u < - .
W U * x u
L > u w >
au J X U
Y
;r
4
YI
V
n
U
X
nl
I
z
v
U
CI
S
Q
v
U
- N
Y
CIbHSbSN (TPTRABUIILSTANNA-
ME)
BibZiography 507
x
YL
U
I
0
0 a
0
z
yt
*
4
a n
L c
rn a
e
(L
a
u
W
a n
z
4
a
4 8-
U
W
a A
z e
a 3
MI
z w
4
0
0
f Y
n
-. z
4
-.
I.
U
3
U

-
t
I-
J
L
Y
a
I-
m
-8
U
2-
#
.
c
..
I-
u1
N
T
c 8.3
508 Chemical Thermodynamics
w Y I L Y
T I T
-I >
3
24 4.11
a
0
N
S
Q
0
 1974, 61 895.
V E ( 2 9 8 K ) 1 J,-D, OROLXER, D, B A L L E T t A N D A. VIALCARD, J. CHEM, THERMODYNAMICS,
1 9 7 4 , 61 8 9 s .
510 Chemical Thermodynamics
L
W
I
U
n
*
2
a
wl
w
W
u
z
m
z
3
m
E
L
Y
a
--
0
W
W
a
n -
o
-- - _
a x aa
w w Y L Y
-3 3 3
v
W) m
N NN
mm
a a a a
Lyw w w
zz 7 2
C C Ce
z
wzw z z
w w
"U u u uu vu
I
.
I
Bibliography 511
I I 0
5 n t
P
0 2
2.
n w
0 I
I
I
R
u a -
-- "
-
a a
Y
z
<
m
z
e
U

Y
z
c
u
-
Q Y
a L
t
" Y
Q
0
a
D
*.-
a.
o w *
hi- f
Y
- 4
N
I
.
)
v I
U
0
4 a m
-
0
X
u
512 Chernical Thermodynamics
n
--
s-r
N
. I
H
s -
0 -
*s-
0
<
Y Y c a
U
L
m
-
u
A
>
Y
-.. A
-
r w a
n a uz a
3
UJ
n
a
4
I
z
U
I
-
0
a
I
c
z
n
a
n
U
a
Bibliography 513
z
r, V
I
a
u
Y
U
w
T
c
.- .
CI
ww Y Y I . I
zz z z 22 il
el- +I-
W zz
WW
uu
z z
W y l
0 0 -.. -
Y W Y
c
- -
-
A
Y
z m
"
4
w
0 0 z
Y
-z *
I
514 Chemical Thermodynamics
lL
Y
Y
X
N
w
u-
a
3
f
N
w
0
n
=a
U
>
n
z
4
-
z
x
m
3
b-
n
a
z
c.
I a
nnu 3
Y
z
uuzo
a
a
uuu
I -
5a
w w a
11.

- f O
I .
.
.I
-
z
W
a
-
0
a
c
2
Y
z
N
.
I
CLH (HYDROGEN C H L O R I D E )
0 HZO (UATLR)

a KR (KRYPTOY)
XP ( X E N O N )

SLK (POTASSIUM CHLORIDE)

4 H20 (WATER)

CLNA
 V E ( 1 4 TO 2 0 K ) l V.N. G R f 0 0 R ' E V AND V.S. RUOLNKOn ZHUR. LKSP. TLOR. r I 2 . 1 1 9 6 1 , 4 0 1 ?$?

RtROX I Dh)
 I
N

c.
;t
.
(
p.0
a
N
1 ZD
t0

om
t
0.-

4- 40 -4 - x
-4 monrr r n 0 C rn
m
Y
It z- n o Rrr w
-m -0-r
v
- X
- 0
t
-
n A
-
Y

0 h
m -4
- U
N x
.o
m
-t-
0-
x 0.-
Y
- -'3. 0

W
*
-
GOT
c

.z
-. m
W -a
-
DI N h - l
I
O -X
rn
8Y
,
C N *>
D I I oz
c - 0
U B W N
W Z
0 0.

LIS
518 Chemical Thermodynamics
w
v
--
W n J t F
s>--
z
w
u
0
- 2
z-
-
a
c
U
--
O N
o
P x
a-
T
z
N
Bibliography 519
I420 ( W A T E R )
6 02SI (SILICON DIGxIDEI

4 031 ( S U L P H U R T R I O X I D E )
R OSu C f u N O S l E N T R I O X I D E )
8 XE (XENON)
Bibliography 521
0
I
'c.
c
Q
z
P -
O
I-- 4
N - 0
o u oz
-
Tern
-mm
=a-
**
V Y
t
v
a
-
a -
O I
U L I I
C f W
- 0
t u x
- L O
 VI
N
me L u t L ?Uft) N
02 ( O X Y G E N )

I 01s ( W L P l t u I DIOXIDE)

X I (NINON)

KR

oe
GL
VE
V€
NO

I t
Bibliography 523
N .
a
*
r
H N I
01E &
C I
- 0 -> 9
0
u
0 -r*
0
u
m
..
A
0
W
%
0
0
L
C
z
0
Q
*
x
-
T"0
Y
3N
0 0
R
c m
z
n Y
0
z
.*
5 24 Bibliography Index
3 Index

A L L W I X T U R E S I N V O L V I N O SUBSTANCES
O F H I L L MOLECULAR FORMULA HAY B E
FOUND BY R E F C R E I i C E TO T H € FIRST
COMPONENTS G I V E N If4 T H E SECOND COLUMN

AR SEE AR.

8CL3 SEE BCi3.

BF3 SEE BF3.

8216 SEE 62H6.

8203 SEE 8203.

BRn SEE ERU.

BRZ SEE BRZ.

BR3P SEE BR3P.

BR4S I SEE BR4SI.

BRCSN SEE BRISI, 8R4SN.

CBRF3 SEE CBRF3.

CBR4 SEE CIRC.

CCiF3 SEE CCLFJ.

CCLZFZ SEE FCCZFZ.

CCLZO SEE CCL2O.

CCL30 SEE CCL3D.

CCL4 SEE AR, BRZr B R C S I r B R C S H r CCLC.

CD3N02 SE€ C03N02.

C05h SEE CDSN.

C F4 SEE CF4.

CHBR3 SEE CCLF3. CHERJ.

CHCLF2 SEE AR, CBRFS, C C L Z F Z i CUCLFZ.

CHCL2F SEE CHCLZF.

CHCL3 SEE CCL4, CHCL3.

CHF3 SEE CCLF3. C F 4 , CHF3.

CHI3 SEE CH13.

CdZ SEE CH2.

C H ~ ~ R Z SEE CHZBRZ.

CHZCLF SEE cn~ci~.

C~ZCLZ SEE CCL4, CntrR3, CHCL3. CHZBRZ, CNZCLZ.

CH209N SEE CHZD3N.

CH2F2 SEE CHF3. CHZFZ.

CHZlZ SEE CHZBRL, CHZCLZ.

cwo S€€ two.

cnzoz SEE CHZOZ.

CH5BR SEE CH3tiR.

C ~ ~ C L SEE CNJBR, CIi3CL.

 Bibliography Index 525
CH3CL3S1 SE€ CH3CL3S.I.

CH3DO SEE CH3DU.

CH3DZN SEE CH3DZN.

CH3 I SEE CCl4, CHCL3e CHZCLZ. CH31.

CH3NO SEC CH3NO.

CH3N02 SEE CCL4, CHCL3, CH3NOZ.

cn4 SEE AR, CCLL, C F 4 o CHCLFZ, CH4.

CH4CLZSI SEE CR4CLZSI.

CHLN2O SEt CH4NZO.

CH40 SEE AR, C C L 4 r CHBRJ, CHCL.3, CHZCLZ, CH3C1, CH3N0, CH3N02, CWLv.CHIO*
CHSN SEE CCL4, CHSN.

. cc) SEt AR, cn4, C H ~ O , CO.

COZ S€E AR, CCL4, CHBR3, CHCLFZ, CHZFZ, CH3C1, CH4, CHCO, CO. COZ.

CSZ set C.CL4r CHCL3, CHZCLZ. CH40r C S Z .

CZBRZCLF3 SEE C ZB RZC L F 3 .

C2BRZF4 SEE CZHRLF4.

CLCLFS SEE CCLZF2, C H C L F ~ I CHZCLZ, CHZFZI CZCCFS.

CLCLZF4 SEE CiiZCLF, CZCLZF4.

CZCL3F3 SEE CZCL3f3.

CL'CL4 SEE CCL4. CS?. CZCL4.

CZCL6 SEE CZCL6.

CZD7N SEE C 2 D 714.

CZF3N SEE CLF3N.

CZF4 SEt CH40-

CZF6 SEE CHF5, CZF6.

C2HCLF4 SEt CZHCLF4.

C~r(C13 SEE CZHCL5.

C2HCLS SEE C Zrl C L5.

CZHD6N SEE C 2ll D 6 I I .

C2HF302 SEE CCL4r CZHF3OZ.

cznz SEE CCL4r CH40, COZ, CZYc!.

C2HZBRZ SEE CHCL3, C H 3 l r CZHZRRZ.

CZHZBR4 SEE CZHZBR4.

CZHZCLZ SEE C C L 4 r CZHZBRZI CZHZCLZ.

CZHZCL4 SEE CCL4r CZCL4, CZHZCL4.

CZHZFZ SEE CZHZFZ.

~21420 SEE CZHZU.

C Z ~ ~ C L SEE CZH3CL.

CZH3CLO SEE BCL3.

CZH3C13 SEE CCL4, CZH3CL3.

CZH3F30 SEE C ZH3 F30.

CZH3N SEE CCL4, CHCL3. CH3N02, CX40, C2HCL3, CZHSN.

CZH4 SEt CCL4, CHCL3. CH4, CH40. COZ, CZCL6. CZH2, CZH3C1, C2H6.

CZH4BRZ SEE CCL4, CHCL3, CSZ, CZH4BR2.

526 Bibliography Index
CZHCCLZ SEE CCL4, CtIZOZq CH40, C Z H 4 r CZHCBRZ, CZHZCLZ.

C2HCF2 SEE CCLZFZ.

C2HCO SEE CCLL, CHCL3, CH4, CHCO, C O 2 i C2H40.

C2HC02 SEE CCL4, ClICL3, CHZCLZ, CH40r C021 C Z H F 3 0 2 , CZH402.

CZHSER SEE CZH5ER s

C2HSCL SEE CZHSCL.

CZHSCLO SEE CZHSCLO.

CZHSDO SEE CZHSOO.

CZHSDZN SEE CZH5OZN.

C2H51 SEE C C L 4 r C Z H ~ E R I CZHSI.

CZHSN SEE CZHSN.

CZHSNO SEE C C L b i C 2 H G O Z i CZHSNO.

CZH5N02 SEE CCL4, CZH5NO2.

C2H5N04 SEE CZHSN04.

C2H6 SEE CH4, ChCO, CO, C 0 2 1 C2CL6. CZF6, CZH2, C2H4, CZH6.

CZH6C LGA SEE tZH5CLGA.

CZH6CLZSI SEE Ct13CL3SI. C2tiS:LZSI.

CZHLON SEt C Z H 4 D ti.

CZH6NZO SEE CCL4, CZHLNZU.

CZH6O SEE AR, CCLC, CHER3. C H C L 3 , C H 3 N 0 2 . C H 4 0 tH40r COZ. CSZ, C t M S F 3 0 .

ctnc~~
C Z H J ~ L C Z ~ ~ C, Z H ~ C L Z . C Z H C O . C Z H L O Z , C ~ I I S D R , C Z H ~ X ,c m ,

C2H60S SEE C C L 4 r CHBR3. C H C L 3 , CHZERZ, t W 2 C 1 2 1 CH40, CZH3Cl3, C2W3Hi L2W4021

C2tlbQ1 CZHOOS.
CZH60S1 SEt C2H6USI.

C2H602 SEE CH40, C 2 H 6 0 r CZH602.

C2H6S SEE CCL4.

CZH?N SEE CCL4, C2H7N.

CZH7NO SEE CZH7NO.

C2H704P SEE CHCL3.

C2H8NZ SEE CZHOOZI C2HBNZ.

C3F6O SEE COZ, C3F60.

C3F8 SEE CZHGr t3FB.

C3HFSO2 SEE C3HFS02.

C3H2Fb0 SEE C 3 H 2 F 60.

C3H3F3 SEE C2H4.

C3H3F302 SEE C3H3F302.

C3H3N SEE CZHbOSr C3H3N.

C3H3NS SEE CCL4, C3H3NS.

E3H4 SEE c3n4.

C3H4CLF30 SEE CHCL3.

C3HCCLC SEE CZH4.

C3HLF40 SEE C H C L 3 r CHZCLZ, C3HCFCO.

C3HCO SEE t3H40.

C3HC03 SEE CZH60S. C3H403.

C3HCOC SEE C3HCOC.

Bibliography Index 527
CJHSBR SEE

C3HSF30 SEE
C3H5N SEE

C3H6 SEE
C3H68RZ SEE

C3H6CL2 SEE

t3H6N6 SEt

c3n60 SEE

C3H602 SEE

C3H603 SEE
C3H7BR SEE

t3H7N SEE

c3n7~o SEE

C ~ ~ T X O Z SEE
C3H8 SEE

C3H80 SEE

t3H802 SEt

t3H803 SEt
C3HPCLSI SEE

C3HPGA SEE

C3H9N SEE

C4C L3F7 SEE

C4F8 SEE

C4FlO SEE

C4H3F303 SEE

t4H40 SEE

C4H404 SEE

C4H4S SEE

C4H5N SEE

C4tl6 SEE

C4H6CLF3J SEC

C4H602 SEE

C4H603 SEE

C4H6iJ4 sit
c 4 H 7r40 5EC

C4H8 StE

C4HSCLZ SFt

C 4 H 8 C L2 0 StE

C41180 SEC
528 Bibliography Index
C4H8OZ SEt C C L 4 , C H C L 3 , C H Z C L Z . C113E10~ C H 3 N O 2 , CHCO, C S 2 , C Z H C L 3 . C t H S N a
CZMLBR2, CLlISUe CZH51. CZH60, CZH60S, CZH602, c 3 H 6 0 2 0 C3H7NO'
C3HOO. C 3 H 8 0 3 , C 4 H S H . C 4 H 6 0 C 4 H 6 0 Z . C C H 8 0 . ~ 4 ~ 8 0 2 .

C4H802S SEt CCLC, CtIZCL2, CH40, CZH60r C3H80, C4H802, CCH8OZS.

CCH9ER StE CCL4r C2HJCLSr CbH8CL20, C4H96R-

CCH9CL SEE CCLC, CS2, CZH3CL3, C3H&CLZ, CCH98Re C4H9CL.

CCHQN SE€ C4H80, C4H802, CCHTN.

C4HPNO SEE CZHZCL4. C3H3N. CCH602, CCH9NO.

CbHlO SEE CHC, C H 4 0 , C H S N . COZ, C2HC. C Z H 4 0 2 . C Z H S D Z N , C Z H 6 . CZH7N. CSF8,

CJM6, C 3 H 6 0 . C J H T N O I C S H 8 , C 4 F 1 0 q C C H 8 , C b H l O .
C4HlOCLGA SEE C C H OCLGA.
~

C4HlONZO SEE CCHlUNZO.

CCH100 SEE A R I C B R 4 , C C L 4 , C H B R 3 . C H C L Z F , cHCt.3, CH13, CHZCLZ. C t l 3 1 0 CH4, C H 4 0 0

COZ. C S Z , C Z C L 6 . C Z H C L 3 , C Z H C L S r C Z H Z C L C , C Z H 3 C L 3 , C Z H 3 N v C Z H 4 C L Z i
C Z H 4 0 , C 2 H 5 1 , C Z H 6 , C t H 6 C L G A e C Z H 6 0 , C Z H ~ O S I C2H402, C 3 H 4 0 , C3H60r
C3H6OZe C 3 H 8 O v C3H802. C 3 H 8 0 3 , C 3 H 9 G A r C 4 H 6 e C 4 H 8 C L Z 0 , C 4 H 8 0 ~ C4N8020
CGH8OZS, CCHPBR, CCHlOCLGA, C Z H 1 0 0 .

C4HlOOZ SEE CHCLZF, CHCL3r C4H80, CCU802, CGH100Z~

CCH1003 SEE C3HlNO. C4H1003.

CCHlOS SEE C4HlOS.

CZHllBRSI SEE C4111BRSI.

C4HllN SEE CCL4, CHCO, C2H60, CZHlN, C3H7NOv C 3 H 8 0 0 CCU80, CCNlOO, t4HllN.

CCHllNOZ SEE C4HllN02.

C4HlZSl SEE CLHllBRSI, C4H12SI.

C4HlZSN SEE C4HlZSI I C4HlZSfI.

CCH13N3 SEE C4H13N3.

CSCLZF6 SEE C4C L 3 F 7 * C5C LZFC.

C5F10 SE t CSF10.

CSFl2 SEE C3F60, C 4 H l Z S I ~C 5 F 1 0 e C5FlZ.

C5H4F80 SEt C 5H4F80.

CSH402 SEE CCLC, C4H8r C5H402.

C5HSN SEt CCL4. EHBR3r CHCL3, CHZCLZI CH40, CZCL4, CZH402, C2H60, CSH60,
C5H602, C 3 H 8 0 , C 4 H l l O 2 1 C 4 H 1 0 0 , CSHSN.

CSHbN2 SEE CH3foO2, CZH60, C3HI:JO. CCH802.

CSH60 SEE CtlCL3r CSH60.

CSH7N SEE C4H50. CCH802, C5H7N.

CSH8 SEE CH4, CfH7NO. C4MPHO. C5H8.

CSH8CLF30 SEE CHCLJ.

CSH8CL4 SEE CZH4.

CSHB!, SCE cs2.

C5Ha02 SEE C3H60. C3Ha0, CSHROZ.

C5H804 S t t C 5H804.

C5H806 SEE C2rl402. C4H603.

CSH9N SEt C 5 H Y PI .
C>n9NO SEt CLi1402, CZtI60, CZH60S, CZH7:IO. C4H802, CSH9NO.

cbnlr) 3Fr C C L L , CfiL, C O 2 , CZCL4, CZH3N. CZH60. C3H7N0, C4F8r C 4 H 8 0 1 CCHPEIOr

C 4 I I l J O . C5H10.

CSHlOCLZ SEE c 1, L I 6 0 I t S H 1U C L 2 .
Bibliography Index 5 29
C5HlUO SEE CIICL3, CZHZCL4, C31160, C4HSN. C 4 H 8 0 0 C4H8OZ. C 4 H 9 N e CLHPNO, CSHTN,
CSHlU, CSH100.

CSH1002 SEE CHZOZ, CZH5N, C3H80. CSH100Z.

C S H l 1B R SEt C51111 BR.

C5H11CL SEE C4HB0, CSHllCL.

C S H ~ ~ N SEt C 4 H d 0 ~ C 4 H 8 0 2 1 CSH5’1, CSHlO, C5H100, C5HlIN.

C 5 H 1 1 NO SEE CH3:10, C Z H 6 0 r C3H60, CbH802, CSHllNOm

C5HlZ SEE A R , C C L 4 , CH4, C H 4 0 . C 0 2 s C S Z , C 2 H 3 C L 3 , CZHSN, C 2 H 6 , C 2 H 6 0 , C 3 H 3 H ,

C 3 H 6 C L 2 , C3H60, C 3 H 7 N O e C 3 H 8 , C 3 H 8 0 , C 4 H 8 r C4HPNOo C l H l O , CLH100,
C 4 H l l N a C 4 H l L S I r C4H12SNa C 5 F 1 2 , C 5 H l O i C S H I I B R , C5H12.

CSH12NZO SCt CSHlZN20.

CSHlZO SEE C C L 4 , CHCL3, C H 4 0 , C Z H I C L Z , C Z H 6 0 r C3H40, C3ll60r C3H80, Csll8OZe

C4H8CL2Oe C4H802, CbH802Sa C4H100, CSHSN, CSMlZO.

CSHlZOZ SEE CHCL3.

CSH1203 SEE CSHl203.

CSH13N SEE CSHlJN.

C6BRFS SEE C6ERFS.

C6CLF5 SEE C6CLFS.

C606 SEE C 4 H 6 r C6D6.

C601Z SEE C6D12.

C6F5 I SEE C6FSI.

C6F6 SEE CCL4, CZHIOS, C3HbOo C3H8, C4H802, C4H100, C4HlQ8, C 4 H l f N r CJWSN,
C5W0, C 5 H l L . C6F6.

C6FlZ SEE CSH10, C6F6, C6FlZ.

C6FlZO SEE C6F120.

C6F14 SEE C3F6Os C3HFS02, C3H8, CSF10, C S H l Z i C6F6, C6Fl20, C6Fl4.

t6HF5 SEE C3H601 C6F6, C6HF5.

C6HZF4 SEE C6F4, C6HF5, C6HZF4.

C6HJCLZNOZ SEE COZ.

C6H3CL3 SEE CZHGI C2H6r C6H3CL3m

C6H3F3 SEE C6F6, C6HFSa C6HZFb. C6H3F3.

C6HGRRCL SEt CZH4, C 2 H 6 r C6H4ERCL.

C6H48RZ SEE C 2 I l b i C6HkERCL, C6H4BR2.

CbH4CLI SEt CZH4, CZH6.

C6t14 C LNOZ SEE C02. CCH6, C6H4BRCL. C6H4ERZ.

CbH4CC2 StE Ci140, COL, CZH3CL3. C2H4, C2H6, C6H4BRCLe C6H48R21 C6H4CL2.

CbH4FZ SEE C6F6, C6IiFS. C6HZF4, C6H3F3, C6H4FZ.

C 6 ti 4 N Z 0 4 SEt CZH4.

CbH5BR SEt CHCL3, CZH60, C4H8CLZOI C4HXOZr CSH12, C6HSER.

C6HSCL SEE C C L 4 , CHCL3, CH4

E3H7N021 C3H80r
C6HSBR, C 6 H S C L .

CGHSCLO SEE CZHbO. C J H 6 0 0 C4H80, C4H802, CSHSN, CSH100. C6HSCLO.

C6H5F SEE CCL4, C4H802, C4HllP1, C6F6, C6HF5, C6HZF4r C6H3F3r C6HLF2, C6HSBRt
C b H 5 C L . C6HSF.

C6H5I 5EE C6H5URt C b l t 5 C L 1 C6115Fa C6HSI.

C6HSNO2 5Et CCI.4, COZ, C Z H S l ,

-
CSHl2 * C b I l 5 t l O Z
CLH60S, C3HPNr CLHdCLZO, C4H802, C4HllNi C5H5Ne
530 Bibliography Index
CbHSN03 SEE

C6H6 SEE

SEE

SEt

SEt

SEE

C6H8 SEE

C6H8NL SEE
C6HON20 SEE

C6H10 SLt

C6H100 SEE

C6ttl00Z SEE

CbHl004 SEE

C6H11)06 SEE

C6HllBR SEt

C 6 H 1 1C L SEt

C6H11N SEE

C6Hl ?NO SEE

CbHlI N 0 2 SEE

CbHl2 SEE

C6ttlZCLZ SEE

C6Hl2U SEt C C L C r C Z H 3 F 3 0 , C 2 H C a C Z H 6 0 , C 2 H 6 0 S t C3HCF40, C 3 H 6 0 , C 3 H 8 0 , C 4 H 1 0 0 ,

C S H 4 F 8 0 r C S H 1 0 , C S H l O O , C 5 H l 2 0 , C 6 H S B R a C b H S C L , C 6 H S C L O . C6tit00, f6Wl2,
C a H 1 20.

COHlZO2 SEE CCL4, CH40. CZH3N, CLHC, C4H1002. C6H6, COH60, C6H1202.

C6l11 L O 3 SFE CHCL3, CSZ. CGH100.

CGH13BK SEE C4HcCLLO.

C6H13CL 5EE ChY13CL.

CGH13U S t t C 4 4 ~ 0 ,C 4 H 8 0 2 , C S H S ' J , C S H l O O , CSH11N, C6HSCLI C6H6, C6H7Nq C 6 H l l C l q

C6H12. C I H l Z O , CbM13N.

C6H13NO S t t Cblil3NO.
Bibliography Index 53 1
C6H14 SEt

C6H140 SEE

C6H14O2 SEE

C6H1403 SEE
CbH1404 SEE

CBHl4S SEE

COHl5AS SEE

C6H15GA StE

C6Hl5N SEE

C6HlSN03 SEE

C6HlSSB SEE

C61116N2 SEE

C6Hl8N30P SEt CCL4, Ctl8R3, CHCL3, CHZBRZ, C H Z C L Z , CCHPCL, CSHllCL* CSHlZ, C~HSCLI
C6H6r C6H10, C6H12. C6H14, C b H l 8 N 3 0 P .

C6H18N4 SEE C6H18N4.

C6H180S I 2 SEE CZHQO, C Z H 6 0 S I n CSHlO, CSH12, C6F6, C 6 F l Q t CbH14, C61180Sl2.

CbH1803SI3 SEE C6H12.

C 6 H l 9 N S I2 SEE CCL4.

C7F5N SEt C6H6, C6H12, C7F5N.

C7FB SEE C6F6.

C7F14 SEE AR. CCI.4, C6F6, C6F12, C6HlZr C6H14r C7F14.

C7F16 SEt CCL4, C2H6, C3F60, C3H60, C3H8, C4H10, C5CL2F6, CSHlZr C6Hl4, C7f16.

C7HF15 SEE CCL4, C3H60, C4H802, C6H6.

C7HSHRO2 SFE C6H6.

c 7H5C LU SEt BCLJ, CCL4, CHCL3r CZHJN, C3H60, C4H802, C6H5CL, C6HSNO2, C6H6, C 6 H l 4 ,
C7HSCLO.

C 7H5C LO2 SEE C61i6.

ClHSF02 SEt C6H6.

C7H5101 SEE C6Ho.

C7H5N SEE C6HSCLr C6H6r C6H12. C7HSN.

C7HSN04 SEE C6H6.

C7H60 SEE CZHbOS, C4H3CL20.

C7H602 SEE CbHt.

C7H7CL SEE C 6 P t ~ t C 6 H l R I J 3 0 P I C7H7CL.

c 7 t17 Y 0 2 SEE Cli40.

C7ti6 SCE AR, C C L F J , CCLLO, C C L 4 , C H B K 3 r C H C L 3 r C H 2 C L Z v CH3NOZv CH4, C H 4 0 , C 0 2 ,

CSZ, C L H 3 C L 3 , CZH311, C Z H 4 , CZHOBRZ, C Z H 4 C L 2 , C 2 H 4 0 2 , CZHSNOZ, C 2 H 6 0 ,
C Z H 6 0 S , C 3 H 6 , C 3 t l o 0 , C3tI7NO. C 3 H 8 0 , C 3 H 9 N e C 4 H 6 . C 4 H 6 0 2 o C 4 H 8 , C 4 H 8 C L t O e
C 4 l l j U , C 4 t i S O Z t CCHPCL, C ~ H P N O I C 4 H 1 0 0 r C 4 H 1 1 N e CSHSqu, C S H 8 0 2 , CSHPN,
C S t i l t J , C S H 1 1 N a C5H11:10, C 5 H 1 2 , C 5 H l 2 0 , C 6 F 6 , C6HF5, C6H4CL2, C6HSBRo
C b t I 5 i L , C6tlSF. C 6 H 5 1 1 C6H5NO2, C6H6, C6H7N, C 6 H 1 0 0 , C b H l l C L , C 6 H 1 2 1
C D d 1 2 0 . t 6 H 1 3 N . C l H 1 4 , C 6 H 1 4 0 , C6H15N. C 6 H 1 8 N 3 O P t t b H 1 8 0 S I 2 , C~H~CLOI
C 7 t f S t l . C71i8.
532 Bibliography Index
C7IidO SEE

C7llOO2 SEE

C7HVfi SEE

C I H l CC L 4 SEE

c I t 1 1 LO SEE

c 7 tl 1201 StE

C11113N(I SEE

Cllllb SEt

c 7 t l t 40 SEt

C7H 1 4 0 2 SEt

C7H15tl SEE

c 7 11 1 5 ti :I SEE

C7i116 SEE

C 7 ti1 6 0 Ski

C7H16U3 SEi

CbF160 SEE

C8FlO SEE

CdHb SEE

Con60 SEE

C8M7N SEE

CLH8 SEE

C8H8O SEE

CbH(IO2 SEE

CLHYCL SEE

C8H10 SEC

SEL

S t t

SEt

Cdtil40L SEt

CtlH1404 SEt

CdHl5NO SEE

C6Ml6 SEE
Bibliography Index 533
C8H160 SEE C3H4O. CbH80, C5H100.

CaH1602 SEE CCL4, C6H6.

C9Hl603 SEE C4H100, CBH1402.

CdHl7CL SEE C6H14. C8H17CL.

C8H17NO SEE CCL4r C 7 H l 6 r C8H17NO.

CllH18 SE€ AR, CCL4, CH4, C 0 2 1 C S 2 i CZHF302, C2H402, C2H6, C2H60r C3H60, CSM602,
C 3 H 8 r C 3 H 8 O e C 4 H 6 , C 4 H 8 . C 4 H 8 C L 2 0 , C 4 H 8 0 , C L H 8 0 2 , C 4 H l O m C4HlOO.
C 4 H 1 0 0 2 , C 4 H l l N , CSCLZFbr C S H 4 0 2 r CSHSN, C S H 1 0 t C S H l 2 , C5H120, C6F6.
C6H6, C6H12, C6H14, C6H140, t 6 H 1 8 0 S I 2 , C7F16, C7H8, C7Hl4, C7H16,
CBFlbO, C8H8021 C8H10, C8H18.

CgHl8O SEE CCL4, CHCL3r CH40, C2H2i C2H4, CZH60, C3H80, C3H802, C4H100, C 6 F 6 1
C6M6, C6H12, C6H14, C6H140, C7H801 C7H16, C 8 H 1 4 0 2 , C8H180.

C8H1803 SEE ChH6. C6H140, C7H16. C8H180, C8H1803.

C8H1804 SEE C5HlZO. C8H1804.

C8H19N SEE C1140, C Z H 6 0 , C 3 H 8 0 , C 4 H 1 0 0 , C 6 H S C L e C 6 H S N 0 2 , C6H6r CIHllCl, C6HllNOZ,

C6HlZe C6H14, C7H8, C7H14r t 7 H 1 6 r C7H160.

CaHZOSI SEE C2H60, C 3 H 6 0 r C S H 1 0 , C6H6, C6H12. C7H14, CBH16.

CaHZUSN SEE CSHl2, C6Hl4, C8HZOSN.

CUH23NS SEE C8HZ3NS.

c 611it402 s I 3 SEE CZH60SIe CSH12, C6F14, C6H14, C6H180SI2, C7H16, C8H18, C8HZCOZSI3.

CRHL404SI4 SEE CCL4, C4H12S11 CSHlO, C5HlZv C6F6. C6F12, C6H6, C6H121 C6H14, C7hl4,
C8H16, C8H18, C8H2404SIb.

C9H7N SEE CCL4r CHBR3, CHCL3, CZH402, C ~ H S C L O IC6HSN03, C6H12, C7H80, C8H100,
C8H18. C9H7N.

C9H6 SEE C8Hh0, C9H8.

CYHlUOZ SEE CHCL3. CH3NOZ. C 3 H 6 0 , C4HBCL20, C 4 H 8 0 2 , CSHSNv C6H6. C6H7Ns C6H14,

C 7 H 8 r C7HdOa C7H16, C8H10, C8H18, CPHlOO2.

C9Hl2 SEE CCL4, CHCL3, CH3N02. CZH3CL3, CZH4BR2, EZMCCLZg CZMSNOZ, C 3 H 7 1 0 2 r

C 4 H 8 C L 2 0 , C 4 H P C L - C o F 6 p C 6 H F S e C 6 H 5 8 R e C 6 H S C L v C 6 H S C v C 6 H 5 1 r C6H6,
C 6 H 6 0 , C f j H I N , C 6 H 1 2 . C 6 H l S N p C 7 H 8 , C 7 H 8 0 , C7H16, C8H60. C 8 H l 0 1
CIMlIR, C 9 H S . C 9 H l Z .
c v H i 20 SEt C5H120, C7H16, CBH10.

CYHl202 SEE C 611 1 00.

C9H13N SEE C6F6.

CYH16CL 4 SEE C2H4.

CYH18 SEE C6F6, C6F12, C6H601 C7F5Ne C7F14.

C9H180 SEE C2H60S, C6H6, CtH12, C6Hl4.

C9HL0 SEE C C L 4 , C H 4 , C H S N , CZHb-, C Z H 4 0 2 , C Z H 6 0 , C 3 H 6 0 , C 4 H 1 0 0 2 , C S H ~ N I C 5 H 1 0 ,

C 5 H 1 2 , C 6 F 6 , C 6 H 6 , C 6 H 1 2 , C 6 H 1 4 , C 6 H l 4 0 , C 7 F 1 6 , C 7 H 1 6 r C 8 H 1 8 , C8tl180,
~ 8 ~ 1 8 0 3C ,~ H Z O .

CYH2OO SEE CZHZ, C3H801 C4H100.

C9H21N SCE CCHSCL, C6HStJ02. C6H6, C6H11CL, C6HllNO2, C6Hl21 C6H14r CIMUI. C 2 I l l C
~ 7 ~ 1 6 .

C1UH8 SEE CO2, CZH2, CZH4, C6H6, C6H12, ClOHB.

C10H80 SEt ClOH80.

ClUH9N SEE ClOH9N.

ClOH1004 SEE CH40, C4H802, C8H1404.

ClUHlZ SEE CSHlZr C6H6, C6H60, C6H7Nv C6H12, C7H16, ClOH72.

ClOHl4 SEE CCL4, CHCL3, CHCi C2H3CL3r CZHbBRZ, C Z H 4 t L Z v C ~ H B C L Z O I C ~ W P C L It 6 f L

C6HFSv C7H16, ClOHl4.
ClOHl4II2.

CCL4.
534 Bibliography Index
ClOH16 SEE CSZ, C6H6.

ClUH160 SEE Cl0Hl60.

ClUH18 SEE C C L L , C H C L 3 , CHL, C S 2 . C 3 H 6 0 , C 5 H 1 0 . C S H l 2 , C 6 H 6 , C6H60, C6M12,

CoH14. C7H8, C7H16, CBHlO, CBH18r CPH20, C10M18.

c 1on1 B O L SEE CCH802, C8H1404.

ClQH2O SEE ChHl2.

ClUH2OO SEE C2H4.

CIOH2002 SEE C2HO.

C l U t i L l CL SEt C6I114, C 1O H 2 1 C L.

ClOH22 SEE C C L 4 , CH2, CHC, CO, C O 2 r C 2 H 4 r C 2 H 4 0 2 , C 2 H 6 , C S H 6 0 , C 3 H 6 0 2 , CSH8,

C 5 M 8 0 1 CCHG, C 4 H U O 2 , C 4 H 9 C L , C 4 H l 0 , C 4 H I U O e C 4 H l Z S N , C S H s N a C S H l 2 r
C b f 6 , COH5CLI C 6 H S N 0 2 , C 6 H 6 , C 6 H 6 0 , C 6 H 1 2 , t 6 H 1 4 , C 6 H 1 4 0 3 r
C 6 H 1 8 0 S I 2 , C I H 8 0 , C 8 H B 0 2 , C 8 H 1 0 0 e CBHZOSNI C a H 2 4 0 2 S I 3 r C 9 H 1 0 0 2 ,
C10H72.

c iQ n 2 z 11 SEE C C L b r C H C L 3 , C H 4 0 , CZHZ, C Z H C , C Z H 6 0 , C 3 M 8 0 , C C H 1 O O r C 6 H 6 r C4H124

C 6 H 1 4 0 , C 7 H 1 6 , C S H 1 7 C L . C S H 1 8 0 , C ~ O H ~ ~ CCL1 0I H 2 2 0 .

ClOHLZ05 S E ‘i CCL2FZv CHCLF2, CHCLZF, CHCL3.

C10H3003514 SEt C5H12. CbHls;, C6H18US12, C7H16, CBH18r C8H2402SI3, ClOH221

ClOH3003SI4.

ClUH30OSSl5 SEt coni2, c8ni6.

CllHlO SEE CU2, C6H60, C 7 H U 0 , CBH100.

CllH1402 SEE C6H14, ~ 7 ~ 1 6 .

CllH16 SEE CbHF5.

C11HZ004 SEE C4HdO2, CSH1404.

CllHZ2 SEE C6H60, C7H80.

C 1 1H2ZO SEE C 3 M 6 0 r C4HSOr C5HlOO.

CllH24 SEE CCLC, C02. C2H4021 C3H80, C 4 H l O O r E 5 H S N r CSU12, C6Y6, C64111U.s

crorum
C12F27N SEE C2CL3F3. C2HF3021 C5Hl21 C 6 f 1 4 a C6Hl4.

ci2nio SEE C2HCr C6H6r ClZH10.

C12H100 SEE CCLF3r C12H100.

C 1 2 H 1 1 I4 SEE c02. C12HllN.

CldH12CLZ04 SEE C4H802r C8H1404.

C1 2 H 1 2 0 4 SEE CCHSOZ, CIHICOC.

C lZH1404 SEE C H ~ O , ~ 2 ~ 6 0C ,~ H ~ O Z c, 8 n i 4 0 4 .

C12H16 SEE clznl6.

ClZHlB SEE C2H4, CZHCCL2, C4HBCL20. C6F6. C12H18.

C12H22 SEE C6Hl2, ClZM22.

ClZH2204 SEE c4na02, C8H1404.

C12H24 SEE CtMB, C12W24.

C17H258R SEE C2H2BR4.

ClZH2SCL SEE C6H14r ClOH220, C12H2SCL.

CIZH20 C C L L , C U L , c z i i 4 , C L ~ I . O Z . C L H O . C Z H ~ O . ~ 3 ~ 6 0 2 c3nsn.
. ~4~802.
C 1 H ? O O , C ’ t Y l Z S Y r f 5 H > i J e C 5 H I 2 , C 6 F 6 r C6H6, C b H l L , C 6 H 1 4 , C 6 H 1 4 0 ,
c 7 n a o . ~ 7 ~ 1 c6 ., 1 1 6 , c 8 t 1 ~ l I s t ~ ,c l u 1 6 , ~ 1 0 ~ 1 3c ,i u t ~ z z , C 1 2 H l 6 , C l 2 H l 8 ,
Cl2H22. Cl2H24, ClLiiZb.

Stt CELL. CHCL3, CHLO. CSti17CL. ClJMLlCL, C12H25CL.

SEE C 1Zh2635.

SEE C641Jbr CL’ilSN.

Bibliography Index 535
C 12H 2 7 I4 SEt ChllSCL, CbfISN02, C6tI0, C6tillCL. C 6 H l l N 0 2 1 C6H12, C6H141 C7H8, C7H14,
C7Hlo.

c 12 fic 7 0 4 k' SEC CCL4, CIlCL3, C6H6, C6H14, C6H140, C7H16, C8H180. C9HZ0, ClOH22,
C12H26.

C 12 H 3 6 0 4 S I 5 3Et C5t11Z8 C6ti14, C6ll18OSIZ, C7iI16, C81118, C8H240ZSI3, ClOHZZ,

ClOH5003S14* ClLH3604SI5.

C13ti150 StE CCL4, CLH4a C7H80.

~ 1 3 ~ 1 2 SEE C2YF302r CbH6r ClZH100, C13H12.

C I 3 H 13N SEE Cbti 5 C LO.

C131i20 S t t CZ114CL2, C4H8ClZLl.

C13H24 SEE ~ 6 ~ 1 2 .

C 13H26 SEC ClL'HZ6.

C13H28 SEE CH4, COZ, C3H60, CZHl2. C6H61 CbHl2, C6Hl4, CPtIZOi C l 3 t I Z l .

C14H802 SEE C4H100e C14H802.

C14HIO SEE CZH4, C6H6, C14H10.

C14H1002 SEE C14H1002.

C14H12 SEE CZH4.

C14HlZO2 SEE C4H802. C6H6.

C14H14 SEE C6H6.

Cl4H1804 SEE C4H802, C8H1404.

C14H22 SEE CZH6CL2, C4H8CL2Om

C14H26 SEE C6H12.

Cl4H2604 SEE C4HLIO2, C8H1404.

C14HZ9CL SEE C6H14, C14H29CL.

C14H30 SEE CCLF3, CCL4, CZH60, CGHPBR, C6H6, C 6 H 1 4 . C6H18OSIZ~ c 7 H l 2 0 ,

c 8 ~ 2 4 0 2 S I f , ClOHZZ, C l O H 3 ~ 0 3 S I 4 . c17HZ6, C1ZH3601S15r C14lt30.

C14M300 SEE C8H17CLe ClOM21CL, C12HZSCL.

C14H4205SI6 SEE C6H18OS12, C8H2402S13, ClOH3003SX4.

C15HZ4 SEE CbF6.

C15dZ8 SEE C6Hl2.

c i 5n32 SEE C6H6.

C16H10 SEE C16HlU.

C16HZ204 SE€ CCLLF2, CHCLFZ, C4HdOtp CBMl404.

C16H3202 SEE C3ti8.

C16H33CL SEE C6H14, C16H33CL.

C16H34 SEE CCL4, CHCL3, CH4r C H ~ O I CO2r C2HZe C2H3Cl3* C2H4, C3M60, C3H80, CCltPBR,
C4HYCLv C4H1001 CGHlZSN, CSHllCL, CSHl2, C6M6, C6H12, CbH13CL, C6P14,
CbH140, C7H8. C7HlZ0, C7H14, C7H16, C8H802, C8H10, C8M16, CBltlTCL,
CYHldr C8H2dSN, C8H2404SI41 CPHlZ, CPHZO, ClOHlZ, C l O f l l C i ClOHl8,
ClOHZlCL. CtSHZ2, ClZHZSCLt C12H26, C14H29C1, C14H30, ClbH33CLI ClbNS4.

Cl6H340 SEE CZIIZ.

Cl6H36SN SEE C5H12, C6H14. C7H16. C8H18, CPH20, C10H22t ClZHZ6t C16M34r C16H36SN.

ClbH4806SI7 SEE C6H180S 12.

C17H36 SE€ C6H6, ClbH36SN.

ClLIH14 SEE C6H6.

C18H2604 SEE C4H802, C8H1404.

C18H30 SEE C2H4.

ClilH3402 SEE CSZ. C2H4. C3H60, C3H8, C4H100.

536 Bibliography Index
c18H34o4 SEE

C18ti360Z SEE

c18n37C~ SEE

ClllH38 SEE

ClaH38u1 SEE

C18H54U7SI9 SEt
Cl9Hl6 SEE

ClYH40 SEE

CZUHZh SEE
CZWH3002 SEE

C2UU42 SEE

tZUH4207 SEE

C2ZH46 SEt

C2 2 H 4 6O 7 SEE

C23H48 SEt

CL41148 SEE

CZ4d50 SEk
C L G l l s l :I SEE

C24HSZO4Sl SEE

CZuH4204 SEt
CLIHZO SEE

C27n54 SEt
c26Hs8 SCE

C30HSO hi€

c 3 ‘3 d 6 2 >EE CCLItr CtICL3. CHZCL2, CZH6, CSHlZ, C6H6, C6H12. C6H14, CtH16, C8Hl8,
C16d34.

C3Ld66 SEE CZH6, CSHl8, C9H2O.

~$4~6802 SEE C3Hd.

C36H74 SEE C2HCl CbH14, CIH16, C8H18, C9H20, C 1 3 M 2 8 , ClbH34.

c36~75N SEE C6H5CL. CbH6, CBH18.

C37U76 SEE CZM6.

cb2nl26 SEE C8H18, CPHZO. C13HZ8, C16H34.

CLF3 SEE CLF3.

CLH SEE CHCJ. CZHZ, CZHSCL, CZH6r CZH6Ot C316, C3H8, C4Hl0, C6HSCL, ClH.

C LK SEE CLK.

C LNO SEt CHCL3. C4Hd02, CSHSN.

C LNA SEE CCNA.

cLL SEE BR2.

CL3HSl SEE C6H6.

CLfDP SE € CL30P.

CL3P SEE C3ti9H, CL30P.

E14SI SEE BR4SNo CCL4. CHSCL’JSI, CH4CLZSl. C Z H 6 C L Z S I e C6H12, C6H14, CtHl6.

CBtIlO,

SEE CCL.6, CH3CL351r C6H12, CLCSI, CL4SN.

SEE CCL4. CLH8OZ. C6H4, C6Ml2, C681202a C7N16, CLISI, CLISN.

Bibliography Index 537
CUD SEE CUO.

I1 tI SEt DH.
DL SEt ARe C7H16. C8H18, Dtlr 02.

Di(l StE CIl31)0, CZH40, CZHSDO, CZH60r C b H 8 0 1 C4H802, C6H60, C6H7Ne C 6 H l S N , 020.

DL'OZ SEE 020.

I) 3 H SEE C3:1n, 03N.

FA SEE 8F5, C L F 3 r FH.

F i SEE F2.

F6S SkE CO2, C5;18, CSF12. FbS.

l6!d SEE C5F13.

6CU2 SEE GEUL.

II 1 StE HI.

'i'JU5 SEt llNU3.

tIilA0 SEE ti tl AU .
HL S tE AR, 82116, CCLGI CH4. C H 4 0 , COa COZ, C Z H 4 , C 2 H 4 0 , C Z H 6 , C Z H 6 0 , C 3 H 6 r
C J H 8 , C3HCIU. C 4 H l G t C 4 H 1 0 0 , C S H l Z , C 6 H 6 1 C 6 H 1 2 , C 6 H 1 3 N e C 6 H 1 4 . C 7 H 8 ,
C71114. C 7 H l 6 , C B t i l u , C B H l 8 , C l J t i Z Z , D H I D Z , F Z , H z .

HZu >Et AR, 6203, R R H , C H Z C L Z , C H Z O i CtI2021 C H 3 N 0 , C H 3 N O z r CH4, C H 4 0 , CHSN, C 0 2 1

C Z H Z , C 2 H 3 F 3 O o CZH3N, C Z H 4 , C Z H C O i C Z H 4 0 2 , CLHSNO, C Z H 6 , C Z H 6 0 , C2H6OS,
C 2 H 6 0 2 , C Z H ~ N I CZH7t40, C Z H 8 N Z o C 3 F 6 0 s C'JHZF60, C 3 H 3 N , C 3 H 3 N S a C 3 H 4 ,
C 3 H 4 F 4 0 r C 3 H 1 0 , C 3 H 4 0 3 , C 3 H 4 0 4 , C 3 H 5 F 3 0 r C3H6, C 3 H 6 0 , C S H 6 0 2 , C 3 H 7 N 0 ,
C 5 H 6 r C3H80, C 3 H 8 0 2 o C3H803, C5H9N, C4HC04, C4H6, C 4 H 6 0 3 , C4H604,
C C H 7 N O t C 4 H 8 . C4H80, C 4 H 8 0 2 , C 4 H 8 0 Z S t COH9N0, C 4 H l O e C 4 H l O O e C4H10021
C 4 H 1 0 0 3 , C S H l l N , CCH111402r C C H 1 3 N 3 . C S H 4 0 2 . CSHSN, C S H 8 0 4 , C S H 9 N 0 ,
C 5 H l 0 , r CSHlOO, CSHlOOZ, CSH11Ne C S H l l N O , C5H12. C S H l Z N Z O ~CSH120,
C S H l Z 0 3 , C5H13N, C6H6, C6H60, C6H602, C6HIN. C6H100, C6H12, C 6 H 1 2 0 2 ,
C6H13NI CbHlSNO, C6H14, C 6 H 1 4 0 2 , C 6 H 1 4 0 3 , C6H1404, C6H15N, C6H15N03,
C Q H 1 8 N S O P s C 6 H 1 8 N 4 , C7H8, C 7 H 8 0 r C 7 H 9 N , C 7 H 1 3 N 0 , C 7 H 1 4 , C 7 H l 4 0 2 ,
C7H15N0, C 7 H 1 6 r C8H10, C 8 H l l N , C8H15N0, C8H17NOv C 8 H l 8 , C 8 H 1 8 0 3 ,
C 8 H 1 8 0 4 , C B H 2 3 N 5 , C 9 H 7 N , C 9 H l Z e CPHZO, C 1 0 H 8 , C l O H 1 4 N 2 , C l O H Z Z , C 1 2 H 1 0 ,
C 1 2 H 2 6 . C 1 2 H Z 6 0 5 , C l 6 H 3 4 , C Z O H 4 2 r C Z O H L Z O 7 , C Z Z H 4 6 0 7 , CLH, C L K r ELHA,
C U O , DZO, F H I GEOZ. H I , H N 0 3 , H N A O i H Z , H2O.

HLU2 SEi HZO.

HLUCS SEt CH3NI)ZI CCH100.

rl 2 s SEE CH4, C t i 4 0 r COZI CZH6, C3H6, C3H8, CCHIO, CSHlZ, ClOHZZ, HZO, HZS.

'I 5 1.1 SEE AR, CH4, CHCNZO, C O Z i C Z H Z i C Z H 6 , C 3 H 6 N 6 , C3H8, C 4 H 1 0 , C 6 H 6 0 2 r C 6 H 8 N 2 ,

C6H1006, C6H12, C6H16NZ. C8H18, C l O H 8 r C 1 0 H 8 0 r ClOHPN, ClOH160, C l Z H l l N ,
C 1 3 A 1 2 , C 1 4 H d 0 2 0 C ~ C H ~ O CI l C H 1 0 0 2 1 C 1 9 H 1 6 , D 3 N v H2, H Z O , H3N.

SEE HZO.

SEt CZHBNZ.

SEE At?, C E R F 3 , C C L F 3 r C C L Z F Z , C C L 4 , CF4, C H C L F Z , C H F 3 r CH4, CH40, C O i COZ,

C Z O R L C L F 3 , C Z B R Z F C i C Z C L F 5 , C Z C L Z F 4 r C Z C L 3 F 3 , CZHZ, C Z H Z F Z , C 2 H 4 . C 2 H 6 ,
CL'HOL), C 3 H 8 , C 4 Y l O t C 5 t i l Z 1 C 6 H 6 , C 6 H 1 4 , F 6 S i H Z , H Z O , H3N, HE.

!I b SEt ARr CH40, C3H60, C3d8, C4H10, HZO.

I2 SE t CCL4, CO2, csz.

KR SEE AR, CH4, C3H6, CLH, ll3;*, H E , KR.

'Id03 SEE H LLI.

'I 3 SEE Cllb, NO.

'1 2 SEE AR, C C L Z F Z , C C L 4 , CF4, C H C L F Z i CH4, C H 4 0 , CO, C o z i C Z H 4 0 , CZHCOZI C 2 H 6 ,

C 2 H 6 0 , C 3 H 8 . C 3 H R O I C b H l O r C 4 H 1 0 0 , C S H 1 2 , C 6 H 6 , C 6 H 1 4 , C7H8, C 7 H l 4 r
C 7 H 1 6 , C 8 H 1 8 r C l O H Z Z t C l h H 3 4 , D Z r H Z r H Z O , H 3 N e H E , NZ.

NZO SEE CH4, CH4Oe C O Z r C Z H 4 r C 2 H 6 i C Z H 6 0 , C 6 H l Z t HZO, HE, NZI NLOs

NL03 SEE CHZCLZ. C 4 H 8 0 2 1 C7H8, NZ05.

538 Bibliography Index
MZOC SEE

MA SEE

14 E SEt

LIZ SEE

OLS SEE

02Sl SEE

03 s SEE
05W SEE

04US SEt
S SEE

12 SEE

XE SEk
 Author Index

Abdul-Karim, A., 59 Beaumont, T. G., 159, 160 Brown, N. M. D., 164, 169
Abe, A., 154 Becker, F., 38, 158 Bruce, G. R., 17,185,186,188
Adams, R. A., 149 Bedford, R. G.,148,261 Brunner, G.,143, 146
Affsprung, H. E.,22 Beenakker, J. J. M., 205, 226, Brusset, H.,46
Agnew, F. J., 166 228,229 Buback, M., 101
Ahlert R. C.,211 Bell, T..N., 17 Buckingham, M. J., 250, 255,
Akers,'W. W.,87, 88, 137 Bellemans, A., 69 257
Alani, G.H.,73 Bellm, J., 236 Budenholzer, R. A., 21 1
Alder, B. J., 128,214, 252 Belousova, Z. S., 222, 226, Buchner, E., 112
Alexander, E. A., 222 233,234,236 Buffham, B. A., 51
Allen, G.F., 259 Benesi, H.A., 159 Burdon, J., 145
Allen, P. W,, 13 Benson,G. C.,2,3,11,17,187 Burnett, M. G., 64
Alley, C.O., 268 Besserer, G. J., 99 Burns, J. F., 88, 137
Alley, S. K.,157, 177, 179 Bhattacharyya, S. N., 1 1 Butler, J. A. V., 64
Altunin, V. V., 236 Bier, K.,236 Byer, S. M., 5
Alwani, Z.,102, 109, 134, 136, Bird, R. B., 212 Byrne, M. A., 233
145 Bissel, T. G.,188
Ambrose, D., 55, 75 Bittrich, H.J., 35 Cadogan, D. F., 63,65
Ananthakrishna, G.,271 Blagoi, Y. P., 150 Cahn, J. W.,258
Anderson, D.L., 157, 179 Blancett, A. L., 233 Caillet, L., 75
Anderson, E. L., 55 Blanks, R. F., 67,219 Calderbank, K. E., 167
Anderson, J. E., 164 Blasco, P. A., 65 Caligaris, R. E., 236
Andrews, A. W.,155,161,182 Blaum, G., 145 Calmettes, P.,268
Andrews, T., 72 Block, T. E., 251, 252 Calvin, D.W.,216
Anisimov, M.A., 264 Bloomer, 0.T.,100 Canfield, F. B., 233,234
Armitage, D. A., 28, 157, 160, Bobetic, M.V., 236 Cardon, D. L., 172
162, 163 Boden, N.,164 Carone, P. F., 65
Armstrong, N. R., 26 Boeyens, J. C. A., 163 Carpenter, R. L.,153
Artemovich, V. S., 90 Bogeatzes, A. S., 64 Carreira, L. A., 169
Ashworth, A. J., 22, 60 Boguslavskaya, B. I., 65 Carter, W. B., 222
Aspnes, D., 248 Boissonas, Ch. G., 196 Chakrabarty, M. P., 236
Astin, D.K.,174 Bol'shakov, P. E.,93, 125 Chalkley, D.E., 65
Atack, D.,261 Bondarenko, V. F., 236 Chamberlain, J. E., 168
Axelrod, H. D., 56 Bose, T,K.,233 Chand, A., 157
Aziz, R. A., 237 Bosschieter, J. C. W., 226 Chandra Sekhar, P., 261, 271
Botkin, D. F.,211 Chang, R. F., 268
Ba, L. B., 236 Bottomley, G. A., 4,44, 204 Chapman, I. L., 63
Baird, C. E., 186 Boublik, T., 10, 11, Chappelear, P. S., 23, 46, 86,
Bak, C. S., 257 Bougard, J., 236 33C
4.22
Bardin, V. N.,234 Bradford, B. W.,65 Chareyron, R., 3,41
Barker, G. A., jun., 271 Bradley, H.,jun., 218,233 Charnley, A., 21 1
Barker, J. A., 7, 158, 172,177, Brady, G. W.,268 Cheh, H. Y.,222
183, 186,218,237 Brauman, J. I., 173 Chen, C.H.,237
Barker, P. E.,64 Braune, H.,90 Chen. C. J.. 46. 224
Barmatz, M., 250 Breedveld, G. J. F., 145 Chen; R. J.'J., 86
Baron, J. D.,99 Brennan, J. S., 164 Cheyne, K. G., 186
Barr-David, F., 109 Brestkin, A. P., 90 Chieux, P.,261
Barrett, R. M., 168 Brewer, J., 88,205, 233 Chinikamala, A., I59
Barrick, P. L., 84, 142 Brewer, L., 217 Chopra, S. L., 234
Bartis, J. T., 128 Brewster, E. R.,2 Chow, L. C.,65
Bartlett, E., 90 Brey, W.S., 168 Zhristensen, J. J., 35
Bartmann, D.,131, 144 Bridgeman, P. W.,83 hristian, S. D.,22
Battino, R., 142, 153, 169, Briegleb, G.,159 Zhristiansen, L,J., 145
170, 171, 173 Brindley, J. M.T., 160, 163 Zhu. B.. 261. 266. 267 268,
,'

Bauer, D.A., 173 Brollos, K.,134 269 . .

Baughman, G. L., 236 Brsnsted, J. N.,61, 193,218 Clhueh, P. L., 109,225
Baur, M. E., 167 Brooke, G. M.,154 Zhun, K. W.,17
Baxendale, J. H.,13 Brown, G. H.,64 Zikhs, D.S., 106
Beams, J. W., 252 Brown, G. R., 264 ?laitor. L. C.. 90
Beath, L. A., 30,40,178, 187 3rown, I., 10, 40, 63, 66, 186, "ark, R. K.,'60
Beattie, J. A., 199 191 Zlarke, E. C. W., 38
539
540 Author Index
Clarke, J. S., 222 Dombi, G., 56 Francis, A. W.,121
Clechet, P., 3, 41 Donnelly, G., 129 Francis, P. G., 210, 211
Clever, H. L., 25, 142, 169 Dorsch, R. R., 219 Frank, E. U., 71,97, 101, 106,
Coan, C. R., 208, 222,233 Douslin, D. R,,72, 145, 150, 109, 111, 112, 120, 126, 132,
Cole, R. H., 233, 234, 236 166 136, 145
Conder, J. R., 2, 23, 46, 63 Dreisbach, R. R., 60 Fredenslund, A., 145, 172
Cook, D., 76 Duff, G. M., 39 Freegard, G. F., 60
Cooper, A. R., 66 Duhem, P., 159 Freeman, P. I., 109
Cooper, M. J., 248 Duke, J. F., 222 Freydank, H., 222,234
Coopes, I. H., 204 Dumoulin, E., 89 Fried, V., 1, 11
Cope, A. F. G., 264 Dunbar, A. K., 86 Friend, J. A,, 194
Copeland, T. G., 236 Duncan, A. G., 21,71, 84,215 Fromant, M.,46
Copp, J. L., 261 Duncan, F. J., 65,66 Fuguitt, R. E., 10
Corneliussen, R., 25 Duncan, W. A., 39, 155, 182 Fujishiro, R., 31
Coulson, E. A., 4 Dundas, D., 170 Funk, J. E., 5 1
Coumou, D. J., 25 Dunlap, R. D., 148, 210, 244, Funke, P. T., 69
Cox, J. D., 3, 199, 259 26 1 Furrow, S. D., 148, 244
Creek, J. L., 274 Dunlop, P. J., 17, 20, 25 Furter, W. F., 191
Croll, I., M., 148, 149, 263 Duquette, G., 237
Crowne, C. W. P,, 66 Durham, M. C., 170 Gainey, B. W., 24, 49, 61, 62,
Cruickshank, A. J. B., 4, 23, Duston, D. D., 236 196, 209,210,234
24, 49, 50, 61, 62, 69, 196, Dvorak, K., 10 Gambhir, R. D., 261,264
209, 234 Dyke, D. E. L., 78, 148 Gamburg, D. Y., 90
Curl, R. F., jun., 217 Dymond, J. H., 3,60,212,214 Gardner, J. W., 35
Curtiss, C. F., 212 Garelick, H., 257
Cussler, E. L., 17 Eatough, D. J., 35 Gaw, W. J., 16, 55, 150, 155,
Cvetanovic, R. J., 65, 66 Edmister, W. C., 234 244
Edwards, A. E., 208
Dacre, B., 187 Edwards, C., 250
Dahl, T., 163 Efremova, G. D., 273
Dalgarno, A., 213 Eggertsen, F. T., 66
Danneil, A., 136 Ehrlich, P., 100
Dantzler, E. M., 4, 69, 150, Einstein, A., 266 Gelbart; W. M.,267
151, 152, 166. 210, 218. 219, Ekiner, O., 98 Genkin, A. N., 65
234 Elliot, D. G., 86 Gibbs, R. E., 2, 5
1Dasgupta, S., 168 Elliot, K., 37 Giddings, J. C., 46, 144
1Daubert, T. E., 54 Ellis. S. R. M., 64 Gierke, T. D., 166
IDavenport, A. J., 109, 139 Engels, P., 118, 122, 124 Gil-Av, E., 65
IDavies, M., 168 Eniistun, €3. V., 13 Gill, E. B., 168
IDavies, R. H., 21 Eon, C., 66 Gilman, T. S., 197
IDavies, V. V., 30 Erickson, A. L., 139 Gilmour, J. B., 20, 253
CYAviIa, S. G., 17, 145, 236 Essam, J. W., 257 Glew, D. N., 38
1Davis, K. M. C., 159, 160 Eubank, P. T., 204, 208 Glukh, G. M., 236
Ilavis, P. P., 164 Eubanks, L. S., 87 Goates, J. R., 162, 172, 187
IDavison, R. R., 17 Evans. F. D., 171 Gotze, W., 248
Ilawe, R. A.. 210 Evered, S., 63 Goldburg, W. I., 257, 268
1Deal, C. H., 48, 63, 64,193 Everett, D. H.,13, 22, 24, 39, Goldup, A., 48
IDeans, H. A., 23,46, 5 1 48, 49, 50, 60, 185, 210, 234 Gomez-Ibifiez, J. D., 15,40
1De Boer, J., I25 Ewing, M. B., 31 Good, R. J., 215
IDeerenberg, A., 261 Goodwin, S. R., 37
IDefay, R., 106 Fairchild, W. R., 88, 137 Gopal, E. S. R., 261, 264, 271
IDeiters, U., 145 Falconer, W. E., 65, 66 Gorjunova, N. P., 127
13epeyre. D., 46 Fannin, A. A,, jun., 150,250 Gorski, R. A., 208
1Derderian, E. J., 237 253. 261 GOUW,T. H., 144, 146
1De ROCCO, A. G . , 219, 224 Fartar, J. M.,212 Govindavajan, K., 261
13err, E. L., 63, 64,193 Farrell, P. G., 66 Gray, C. G., 145
1Desmyter, A,, 40 Fenby, D. V., 147, 155, 157, Gray, D. F., 187
Ilesnoyers, J. E., 38 161, 182,188,197,268 Gray, G. W., 64
IDestv. D. H,. 48 Fender, B. E. F., 215 Green, M. S., 145, 248, 271
De Swaan Arons, J., 102, 126 Fernandez-Garcia, J. G., 196 Greer, S . C., 251,259,271,272
De Vaney, W. E., 86 Field, L. R., 170 Griffiths, R. B., 145, 240, 246,
Devaprabhakara, D., 210,234 Fjk, A. S., 236 247, 248, 265, 273
Dewar, M. J. S., 169 Fild, M., 222 Griswold, J., 109
de Yllana, A., 59 Findlay, A., 106 Grubner, O., 51
Dlckinson, E., 172, 252 Findlay, T. J. V., 11 Grundke, E. W., 172
Dieoen. G. A. M.. 102. 126
I ,
Fisher, M. E., 247,257, 266 Cubbins, K. E., 145
Dincer,?S., 236 Fishman, L., 25 Guermouche. M.-H.. 63
Dixon, D. T., 22 Flory, P. J., 67, 69, 154, 175 Guerrero, M: I., 145-
Dixon, H. M., 252 Flygare, W. H., 166 Guggenheim, E. A., 7,67,153,
Dixon, R. P., 149 Ford, N. C., ‘un., 272 176, 184, 195,217,238
Dobrovol’skaya, V. E., 236 Foster, R., 119 Gujochon, G., 66
Dodge, B. F., 86,90,93, 109 Fould, H., 66 Guisset. J.. 23
Dolezalek. F.. 174 Fowler, R. H., 238 Gump, B . *H., 59, 60
Domb, C.; 145 Fraga, D., 263 Gunning, A. J., 174
Author Index 54 1
Gupta, S. K., 218, 222, 234 HIryniewicz, R., 191 Keyes, P. H., 268
Gustin, J. L., 37 HIsu, C. C., 191 Khan, M. A,, 66
H ubert, P., 143, Khodeeva, S. M., 103
Haase, R., 106 H udson, G. H., 149, 215 Khosla, M. P., 234
Habgood, H. W., 46,250 H uff, J. A., 4 Khoury, F., 202,219, 234
Hahn. R.. 236 H uggins, C. M., 177, 217 Kidnay, A. J., 84, 208, 234,
H uggins, M. L., 67, 191 236
H ughes, L. M., 177 Kiefer, M., 38, 158
H ughes, R. E., 99 King, A. D., jun., 208,218,222,
234 H unt, D. C., 172 233, 234
H ussey, C. L., 46 King, E. P., 143
H ynne, F., 258 Kirchner, D., 35
Kirk, B. S., 86
Iometov, M. B., 236 Kirkland, J. J., 146
Ipat’ev, V. V., 90 Kister, A. T., 4, 7, 191
216, Irani, N. F., 261 Kittel, C., 235
Irwin, R. S., 65 Kiyohara, O., 3
Hanna, M. W., 169 Ivanov, D. Y., 272 Klein, H., 272
Hansen, R. S., 64 Izatt, R. M., 35 Klein, M., 212, 216
Harris, H. G., 63 Kleinert, T., 109
Harris, K. R., 17, 20, 25 Jadot, R., 236 Klesper, E., 146
Harrison, R. H., 72, 150, 166 Jaeckel, K., 268 Knaap, H. F. P., 205
Harsted, B. S., 2 James, A. T., 47 Knight, H. S., 66
Hart, R. M., 35 James, J., 161 Knight, R. J., 177
Hartkopf, A., 54 James, M. R., 187 Knobler, C. M., 4, 21,69, 148,
Hartmann. W.. 146 Janak, J., 46, 70 150, 151, 152, 166, 167, 205,
H arvey, D:, 65. Jaster, H., 173 210, 218, 219,234, 250, 251,
H aupt, D., 35 Jentoft, R. E., 144, 146 252, 253, 261, 272, 274
H eady, R. B., 258 Jepson, W. B., 82 Knoebel, D. H., 234
H eard, H. C., 112 Jockers, R., 136, 145 Knoester, M., 226, 228
H echt, C. E., 128 Joffe, J., 134 Kobayashi, R., 23, 46, 51, 86,
H echt, R., 248 Johnson, A. F., 30 225, 234
Heck, C. K., 84, 142 Johnson, C. H. J., 219, 224 Kocirik, M., 51
H ederer, H., 145 Jolicourr, J., 38 Koefoed, J. K., 61, 193, 218
H eim, W., 109 Jolley, K. W., 177 Kogan, I. V., 159
H ejmadi, A. V., 226 Jones, C. H., 90 Kohler, F., 26, 261
H eller, P., 263 Jones, D. E., 157, 161 Kohn, J. P., 100, 109
H emmaplardh, B., 222, 234 Jones, I. W., 78, 152 Kohnstamm, Ph., 106,272
H enderson, D., 172,236,237 Jones, M. R., 233 Kolysko, L. E., 236
H erbstein, F. H., 163 Joseph, R. I., 248 Kong, C. L., 216,236
H erington, E. F. G., 4, 5, 259 Jost, A., 124 Koningsveld, R., 145
Herling, J., 65 Judd, N. F., 261 Koonce, K. T., 51
H ermans, J. J., 61 Jura, G., 263 Koroleva, M. V., 90
H ermsen, R. W., 20 Juskevic, G. N., 143 Kortum, G., 39
H erring, R. N., 84 Juvet, R. S., 66 Kosky, P. G., 173
H erschbach, D. R., 213 Koukol, H., 38, 158
H icks, C. L., 222 Kadanoff, L. P., 248, 271 Kovach, L. H., 2
H icks, C. P., 49,50,61,78,83, Kahn, M. A., 46 Kramer, H. L., 213
145, 152, 165,209,234 Kahre, L. C., 86 Krase,. N., 90
H ijmans, J., 194, 195 Kaliagune, S. C., 236 Kreglewski, A., 68, 165
H ildebrand. J. H.. 7. 67, 147, Kal’varskaya, R. S., 90 Kretschmer, C. B., 191
159, 176,220, 252,‘ 255,265 Kamerlingh Onnes, H., 125 Krichevskii, I. R., 90, 94, 106,
Hill, J. L. E., 139, 141 Kaminishi, G., 76 125, 273
Hilmi, A. K., 64 Kane, J., 248 Kruis, A., 3
Hjrose, Y.,124 Kao, W. P., 261, 268 Kucera, E., 5 1
Hmchfelder, J. O., 212 Karger, B. L., 54, 66 Kuchhal, R. K., 54
Hissong, D. W.,82 Kate, F. H., jun., 208, 234 Kudchadker, A. P., 73, 210,
Hiza, M. J., 71, 84, 208, 215, Katz, D. L., 99, 129, 143, 226 234, 236
234 Katz, J., 20, 244, 253 Kuenen, J. P., 106, 109, 110,
Ho, J. T., 248 Kauer, E., 35 131
Hocken, R., 259, 271, 27‘2 Kaul, B. K., 145,210,234,236 Kurata, F., 100
Hohenberg, P. C., 250 Kay, W. B., 71, 73, 76, 78, 82, Kurpen, J. J., 100
Holleman, Th., 194 106, 165, 234 Kuwahara, N., 268
Holmes, R. R., 222 Kazanova, N. E., 90 Kuznetsov, D. O., 236
Hoover. A. E.. 234 Kearns, E. R., 178 Kwantes, A., 61
Hope, C. J., 215 Keesom, W. H., 125
Hoppel, H. E., 17 Kehiaian, H. V., 147, 191 Lacey, W. N., 71,97, 131,211
Horsma, D. A., 167 Kelly, T., 162 Lagues, I., 268
Horvath, A. L., 145 Kennedy, G. C., 112, 127 La Hue, M. D., 56
Horvath, E., 109 Kenworthy, S., 46 Laj, C., 268
Houghton, A., 271 Kenyon, R. S., 11 Lam, V. T., 11
Houth, G. N., 159 Kershaw, R. W., 30, 186, 188 Lambert, J. A., 219
Howell, P. J., 30, 157, 179 Kestin, J., 212 Lambert, J. D., 221, 222
Howlett, S., 170 Keulemans, A. I. M., 58 Landau, L. D., 238,273
542 Author Index
Lang, J. C., jun., 273 McLure, 1. A., 155, 172 viountain, R. D., 25
Lange, H. B., jun., 151, 234 VcMath, H. G., jun., 234 viousa, A. el H. N., 165
Langer, S . H., 62 McQuarrie, D. A., 51 Mozhginskaya, L. V., 236
Langley, K. H., 272 Madan, M. P., 215 Mrazek, R. V., 31
Langmuir, I., 176 Maguire, J. F., 169 ufuhs, M. A,, 65
Larkin, J. A., 2, 54, 162, 194, Mahl, B. S., 234 Mukherjee, A., 11
197 Majer, U,, 268 Mulholland, G. W., 246
Laub, R. J., 235 Makarevich, L. A., 272 Mulliken, R. S., 182
Lawrence, S. D., 222 Maki, G., 263 Mullins, J. C., 86
Lawrenson, I. J., 3, 199 Malcolm, G. N., 17, 30, 185, Munn, R. J., 26, 185
Lee, C. S., 26 186, 188 Murakami, B., 2
Lee, J. I., 226 Malesinska, B., 229 Murakami, S., 11, 31
Lee, J. K., 237 Malinowski, E. R., 69 Murray, R. S., 44, 161
Lee, Y. T., 212, 237 Mallik, K. L., 46, 54 Murthy, D. S. N., 168
Lefebvre, C., 23 Maripuri, V. O., 11 Mushkina E. V., 93
Le Fevre, R. J. W., 167, 168 Markuzin, N. P., 234 Myers, A.’L., 219
Leland, T. W., 153, 154, 174, Maroudas, A., 194 Myers, D. B., 157, 179, 244
225, 230, 234 Marsh,K. N.,2,3, 15,31, 147, Myers, D. M., 7
Lele, M. V., 261, 264 154 Myers, M. E., 268
Lentz, H., 112, 126, 145 Marsicano, F., 61, 236 Myers, M. N., 144
Leonard, P. J., 172 Martin, A. J. P., 47, 48 Myers, R. S., 25
Leong, C. Y., 157, 161 Martin, J. J., 54
Leslie, R. D., 222, 234 Martin, M. L., 44, 45, 161 Nagata, I., 234
Letcher, T. M., 49, 61, 209, Martin, R. L., 51, 59 Nain, V. P. S., 236
234, 236 Martire, D. E.,46, 50, 54, 59, Najour, G. C., 208, 218, 234
Leung, S. S., 265 62, 63, 65, 66, 68, 69 Nelson, R. D., 234
Levelt Sengers, J. M. H., 239, Maslennikova, V. Y., 97 Neogi, P., 236
248, 252, 259 Mason, D. F., 91 Neparko, E., 22
Levien, B. J., 3 Mason, E. A., 3, 199,212,213, Newby, E. J., 170
Levy, S. L., 109 224 Newitt, D. M., 76, 83
Lewis, E. A. S., 248 Masood, A. K. M., 173 Newsham, D. M. T., 37
Lewis, G., 30 Massoudi, R., 222, 234 Newton, R. L., 112
Lewis, K. L., 172, 226, 236 Mat, P., 69 Ng, C. Y.,237
Lichtenthaler, R. N., 219, 234 Mathais, E., 75 Nicholson, G. J., 63
Liebermann, E., 26 Mather, A. E., 226 Nickel, B. G., 271
Lifshitz, E. D,, 238, 273 Mathot, V., 21 Nigam, R. K., 40
Limbach, U., 124 Matire, D. E., 22 North, A. M., 168, 173
Lin, F. L., 269 Mattingley, B. I., 157 Novak, J., 46, 51, 70, 124
Lin, H. M., 214, 216 Mayhew, C, J., 197 Nyberg, G. L., 204
Lindroos, A. E., 93 Meen, D. L., 66
Ljnshits, L. R., 93, 127, 236 Meiron, D. I., 271 O’Connell, J. P., 218, 222
Linton, M., 218 Mermin, N. D., 243 Oeder, D., 101, 109
Lipa, B. J., 257 Merrilt, L., 145 Ogoshi, S., 170
Lipa, J. A., 250 Messer, C. E., 162 Okafor, G., 188
Liphard, K. G., 131 Metzger, J., 160 Olds, R. H., 71
Lippert, J. L., 169 Meyer, H., 264 O’Neill, S. P., 40
Lippman, D. Z., 217 Meyer, M., 160 Orentlicher, M., 225
Litster, J. D., 248 Meyer, R., 160 Oriani, R. A., 258
Littlewood, A. B., 50, 61, 65 Michels, A., 89 Ornstein, L. S., 266
Liu, Y. P., 44 Millen, D. J., 222 Orwoll, R. A., 69
Locke, D. C., 63, 66 Miller, B., 90 Othmer, D. F., 109
Longuet-Higgins, H. C., 68, Miller, F. A., 64 Otsuki, H., 109
194 Miller, G. A., 26 Ott, J. B., 162, 172, 187
Looi, C. K., 197 Miller, J., 46 Ottenstein, D. M., 54
Losa, C. G., 236 Miller, J. G., 208 Ovodova, T. M., 264
Loven, A. W., 261 Miller, R. C., 44, 208, 234 Oxtoby, D. W., 267
Lu, H., 76 Mines, G. W., 222
Luckhurst, G. R., 48, 68, 21 1 Mochei, J. M.,23 Paas, R., 121, 172
Lyons, P. A., 2 Modell, J. H., 170 Pak, S. C., 71, 106
Moelwyn-Hughes, E. A,, 213 Palciauskas, V. V., 248
Ma,Y. H., 109 Moldover, M. I?., 259, 271 Pandya, M. V., 218
McCarty, R. D., 224 Mollerup, J., 145, 172 Papadopoulos, M. N., 193
McCoubrey, J. C., 149, 215 Monchick, L., 213 Parcher, J. F., 46, 51
McCurdy, J. L., 99 Moody, R. W., 49, 58, 209 Pardoe, G. W. F., 168
McDonald, I. R., 174 234 Parent, J. D., 100
McGlashan, M. L., 1, 2, 13, Moore, L. S., 222 Parker, T. G., 168
22, 61, 67, 128, 152, 153, Moore, R. T., 72, 150, 166 Parks, R. D., 261
154, 162, 174, 178, 179, 185, Morachevskii, A. G., 159 Parsonage, N. G., 21
194, 195, 197,210, 211 Morcom, K. W., 28, 30, 67 Partington, E. J., 71, 152
Mclntyre, D., 258, 268 155, 157, 160, 161, 163, 182 Pasco, N. F., 188
McKetta, J. J., 150 185, 195, 197 Patrick, C. R., 154
McKinley, C., 88 Morris, D. M., 222 Paviichenko, L. A., 150
McKinnon, 1. R.,177 Morris, F., 66 Pavlychenko, N. M., 205
Mackor, E.L., 25 Morrison, G., 272 Pease, E. C., 61
McLaughlin, E., 162 Meosdale, S. E., 236 Pecora, R,, 173
 Author lndex 543
Pecsar, R. E., 54 Rehr, J. J., 243, 246 Scheffer, F. E. C., 110
Pecsok, R. L., 22, 49, 59, 60, Reid, R. C., 4 Scheller, W. A., 193
210, 234, 235 Reineke, W., 236 Schen-Wu-Wan, 90
Pemberton, R. C., 54 Renon, H., 8, 37, 179, 193 Schilz, W., 146
Penney, M. F., 13 Rhensius, P., 158 Schmidt, H. H., 60
Pepela, C. N., 17 Rhodes, H. L., 86 Schmidt, R. L., 25
Perez, P., 167 Ricci, J. E., 106 Schnecko, H., 65
Perry, E. S., 10 Rice, 0.K., 258, 261 Schneider, G., 100, 101, 105,
Person, W. B., 169, 182 Rice, S. A., 25 109? 121
Peter, K. H., 115, 121, 134, 172 Richardson, M. J., 82, 208 Schneider, G. M., 91, 102,
Peter, S., 128, 143, 145, 146 Riedel, E. K., 273 115 118 120 123, 124, 131
Petersen, D. E., 217 Riedl, P., 65 134: 136: 143: 144, 145, 175
Pethrick, R. A., 173 Rigas, T. J., 91 Schneider, W. G., 75, 250
Pflug. H. D., 3 Rrgby, M.. 128 Schoenes, F. J., 268
Pick, J., 1, 11 Riha, R., 143, 146 Schoenmann, W., 236
Picker, P., 38 Rijnders, G. W. A., 61 Schofield, P., 248
Pierens, R. K., 167 Ringermacher, H. J., 142 Schouten, J. A., 126, 128, 261
Pierotti, G. J., 64,193 Ritchie, G. L. D., 167 Schramm, B., 236
Pierotti, R. A., 171, 173 Ro, S. T., 212 Schreiber, H., 39
Pilz, V., 146 Roberts, G. A. H., 221 Schroeder, G. M., 134
Pimentel, G. C., 177 Robinson, D. B., 99,202,219, Schumaker, V. N., 252
Pings, C. J., 264 234 Schurmann, H. K., 261
Pitzer, K. S., 21 1, 217 Robinson, R. L., jun., 17,208, Scoles. G.. 237
PolBk, J., 11 214, 216, 234 , 67,
Polishchukova, L. I., 236 Robson, W. G., 109 148,
Pollara, L. Z., 46, 50, 69 Rodger, A. J., 191 177,
Pollard, F. H., 63 Rodkina, I. V., 236 214,
Pollock, J. M., 155, 182 Ronig, P. R., 51 2%
Pommier, C., 66 Roof, J. G., 99 , 274
Pompe, A., 236 Rosenberg, H. S., 82, 234 Scott,’R. P. W.; 58
Pope, G. A., 225 Rosenbrock, H. H., 188 Seffl, R. J., 170
Pople, J. A., 166 Rosenthod, E., 149 Semkin, V., 69, 210, 234
Porter, P. E., 48, 64 Roseveare, W. E., 208 Sengers, J. V., 268
Potter, D. J. B., 61, 152, 211 Roth, K., 120 Serebryakova, A. V., 273
PowelI, R. J., 39, 73, 155, 166, Rott, L. A., 91, 106 Sewell, P. A., 46
I70 Rowlinson, J. S., 4, 30, 71, 75, Shaffer, D. L., 54
Powers, J. E., 226 78, 82, 105, 109, 139, 142, Shaw, V. R., 66
Pratt, M. C., 186 154,145, 147, 148, 152, 153, Shaw, W. G., 64
Prausnitz, J. M., 8, 20, 63, 67, 174, 204, 208, 211, 217, 221, Shenerei, L. I., 93
106, 109, 145, 147, 179. 192, 224, 225, 230, 234, 240,243, Sheridan, J. P., 39, 54, 62, 66,
193, 218,219,220, 222, 225, 244, 247, 263 155
236 Rudd, D. F. P., 261 Sherwood, A. E., 224,225
Present, R. D., 224 Ruel, M., 31 Sherwood, T. K., 4
Price, A. R., 86 Ruenkrairergsa, S., 157, 182 Shieh, J. J. C.. 15
Prigogine, I., 69, 106, 153, 230 Ruhemann, M., 76 Shirts, R. B., 162
Pritchard, H. O., 30 Ruiz, B. D., 170 Shonee, M. G. T., 222
Prokhorov, V. M., 226, 233 Rullis, A. M., 237 Shoolery, J. N., 177
Prosser. G. S., 154 Ruseva-Rakshieva, N., 51 Sienko, M. J., 261
Proust, P. C., 149 Rushbrooke, G. S., 247 Sigmund, P. M.,150
Provine, J. A., 234 Ruska, W. E. A., 86 Silberberg, I. H., 150
Pryanikova, R. O., 273 Russo, C., 123 Silva, R. S. F., 17
Purnell, J. H., 2,22, 23,46, 54, Simon,M., 21, 148, 253, 261
59. 62, 63, 65, 66 Saam, W. F., 246 Singh, J., 17
Pusey, P. N., 257, 268 Sabistan, D. C., 170 Singh, P. P., 40, 234
Sabramanyam, S. V., 264 Sinka, J. V., 149
Radford, D. V., 167 saddington, H., 90 Siska, P. E., 237
Ratzsch. M., 222, 234 gadowska, K. W., 190 Skelton, G. F., 89
Ralek, M., 5 1 Sage, B. H., 71, 97, 131, 211 Skillerne de Bristowe, B. J.,
Ramachandra, R., 261, 271 iallak, H. M., 40 157, 158, 191
Ramaiah, V., 2 18 Sanchez, M., 112 Sloviter, H. A., 170
Ramalho, R. S., 31, 236 iandercock, J. A. W., 99 Smith, B. A., 157
Rambosek, G. M., 73 iantafe, J., 236 Smith, B. D., 40
Rasenberger, E., 234 iapse, A. M., 128 Smith, E. B., 3, 60
Rastogi, R. P., 178 ;argent, J. W., 170 Smith, F., 40, 183, 186, 191
Ratcliff, G. A., 11 jarholz, W., 261 Smith, F. T., 236
Ratzsch, M. T:, 71, 78 iather, G. A., 153, 154, 174, Smith, G. E., 88
Rayl, M., 248 230 Smith, K. M., 237
Reamer, H. H., 71, 97, 131, iatkiewicz, F. G., 17 Smith, R. A., 179, 244
264 javille, G., 21, 139, 204 Smith, V. C., 17
Recko, W. M., 190 iavini, C. G., 2 Smith, W. H., 17
Redlich, O., 4, 7, 191, 193 scatchard, G., 7, 17, 23 Smyth, C. P., 168
Reed, T. M., 4, 149, 216 icesney, P. E., 257 Snedecker, R. A., 133
Reeder, J., 162 ichafer, H. I)..158 Snow, R. L., 187
Reeves, C. G.. 204 ichafer, K., 219, 234, 236 Snowdon, P. N., 210
Rehage, G., 117 ichafer, T. P., 212 Snyder, L. R., 146
544 Author Index
Sokalova, 0.N., 272 'hodos, G., 91, 98 Neeks, I. A., 3
Sokhan, V. I., 150 'hompson, C. C., 169 Kegner, F. J., 27 1, 213
Solana, L., 3 'hompson, D. R., 258 Neimer, R. F., 67
Sonntag, R. E., 88,90 'homsen, E. S., 2 Neinberger, M. A., 75
Sosnkowska-Kehiaian, K., 191 'horburn, S., 61 Neiner, J,, 272
Sourirajan, S., 127 'horp, N., 148 Weir, R. D., 204
Spall, W. D., 26 'igelaar, H. L., 166 Neiss, P., 238
Spurling, T. H., 3, 4, 199, 204 'immermans, J., 60, 106, 112 Neiss, T. F., 65
219, 224, 236 'odheide, K.. 97, 132, 136 Nells, F. W., 99
Srivastava, B. N., 215 'omlins, R. P., 2, 28 Nelsch, H., 145
Stacey, M., 154 'ompa, H., 67, 184 Nenzel, H., 128, 145
Stahl, E., 146 'oriumi, T., 76 Nenzel, L. A., 211
Stalkup, F., 51 'owland, M., 173 Nesselink, G. A., 164
Stam, C. H., 164 'ownley, J. R., 211 Westaway, M. T., 50
Stanley, H. E., 145, 239 'ownsend, R., 75 Nesthoff, S. P., 236
Staveley, L. A. K., 21, 172, 'rappeniers, N. J., 126, 128, Neston, J. F., 71, 152
233 26 1 Nhalley, T. H., 143
Stead, K., 128 'ravers, D. N., 30 Nheeler, J. C., 145, 240
Stecki, J., 229 'scharnuter, W., 268 Nlicar, S., 51
Steele, D., 168 'siklis, D. S., 90, 91, 93, 94, N'iEarova, O., 46, 70
Steele, W. A., 237 97, 106, 125, 126, 127, 142, Nichterle, I., 86
Stein, A,, 259 143, 236 Widom, B., 145,243,247,256,
Stein, F. P., 90, 149, 151, 234 *ully, P. C., 86 261, 273
Stephenson, J., 247 kxford, C. W., 31 Niebe. R.. 191
Stem, J., 13 :wu, C. H., 145 Niese,?H. *C.,97
Sterner, C. J., 90 Nilhelm, E., 153, 170, 173
Stiel, L. I., 218 Jrieta Navarro, J. S., 236 Nilkinson, V. J., 221
Stiles, P. J., 167, 168 Nilliams, F. E., 10
Stock, R., 46, 60 ran Arkel, A. E., 67 Williamson, A. G., 1, 13, 30,
Stockmayer, W. H., 199 ianderborgh, N. E., 26 40, 67, 148, 157, 177, 178,
Stoddart, C. T. H., 24, 48 ran der Waals, J. D., 71, 106, 179, 185, 187, 188, 196, 197,
Stoeckli, H. F., 2, 196 125, 238 211. 218-
- I ~

Stogryn, D. E., 225 ran der Waals, J. H.. 40, 61 Williamson, F. C., 109
Stokes, R. H., 2, 3, 26, 28, 31, 'an Dijk, J. J. Th., 89 Willmott, F. W., 65
147 ran Eck, H., 75 Wilson, G. M., 8, 17, 192, 193
Stookey, D. J., 40 ran Eijnsbergen, B., 226, 229 Wilson, K. G., 249
Storonkin, A. V., 106 ran Konynenburg, P. H., 139, Wims, A. M., 258, 268
Strassman, F., 90 159, 274 Windsot. M. L., 4, 23, 49, 50,
Straub, J., 259 ran Laar, J. J., 8, 174, 238 69, 151, 209, 210, 218, 234
Strauch, G., 78 ?an Ness, H. C., 2, 5, 7, 31 Winterhalter, D. R., 2
Streett, W. B., 90, 106, 127, ?an Wasen, U., 144 Woerman, D., 261
139, 141, 142 (an Wylen, G. J.. 88, 90 Wolf, B. A., 145
Stross, F. H., 48 v'aughn, G. W., 205, 233 Wolff, H., 17
Stryjek, R., 218 ?ennix, A. J., 225 Wong, S. Y., 109
Stubley, D., 30, 157, 158, 179, Jergnaud, J.-M.. 63 Wood, S . E., 7, 23
191 Jerkhova, V. P., 222, 226, Woodbrey, J. C., 148
Sturtevant, J. M., 2 233, 234 Wormald, C. J., 37, 210, 217,
Subramanyam, S. V., 261,271 iernon, F., 65 226, 236
Sulimova, T. D., 233, 236 ieronis, G., 142 Woycicki, W., 197
Sumner, F. H., 225 iicentini-Missoni, M., 248, Wynn Jones, I., 204
Sundquist, B. E., 258 259
Sutton, J. R., 211, 225 U'icfni, H., 65 Yakimenko, N. P., 236
Swanton, W. T., 48 U'igdergauz, M., 69, 210, 234 Yalabik, M. C.,271
Swift, J., 248 Vilim. 0.. I . 11 Yamada, H., 170
Swiniarski, M. F., 222 Viswanathan, B., 261, 264 Yamahawa, H., 25
Swinton, F. L., 16, 39, 73, 147, Vitzthum, O., 143 Young, C . L., 4, 23,24,46,49,
155, 162, 164, 166, 169, 173, Vogel, R., 106 50, 61, 62, 69, 73, 75, 78, 80,
182 Voronel, A. V., 264 83, 143, 145, 149, 152, 155,
Synge, R. L. M., 47 Vrij, A., 69 165, 196, 209, 210, 234
Young, J. A., 21
Taconis, K. W., 226 Wade, J. C., 159
Tamsky, M., 268 Waihel, R., 236 Zaalishvili, Sh. D., 234
Tanaka, R., 31 Wakeham, W. A., 212 Zandbergen, P., 226
Tang, K. T., 213 Walker, R. D., 170 Zernike, F., 266
Tassios, D. P., 64 Wang, E. S. J., 88 Zernike, J., 106
Tatlow, J. C., 154 Wang, T. C., 40 Zhavoronkov, N. M., 90
Taylor, 2. L., jun., 159 Warr, C., 128, 154 Zhuze, T. P., 143
Teck, R. J., 56 Warzel, F. M., 76 Ziegler, W. T., 86
Teja, A. S., 142 Wasserburg, G. J., 112 Zimm, B. H., 261
Ten Seldam, C. A., 128 Watanabe, K., 212 Zollweg, J. A., 254, 258
Teodorovich, V. P., 90 Watkins, C. L., 168 Zudkevitch, D., 134
Tewari, Y. B., 54, 62, 66 Watson, 1. D., 154, 174, 177 Zwicker, J. O., 20, 253
Thacker, R., 78, 148 187 Zwolinski, B. J., 73
Tham, M. K., 170 Watts, H., 38, 179
Thiel, D., 268, 269 Watts, P. J., 35