. ..

Problems in
. .

Applied Thermodynamics
C. Bodsworth
Professor and Head of
Department of Metallurgy
designate, Brunel College

and

A. S. Appleton .
Head of Department of
Metallurgy designate, South
Australia Institute of Technology,
Adelaide

111111111 IIII IIIIII

NBL2366
Ill llUI 1111111111 llll llll

~LONGMANS
:)
I
LONGMANS, GREEN AND
48 Grosvenor Street, London W.1
CO LTD
Preface
Associated companies, branches and representatives
throughout the world
The student reading thermodynamics may choose from amongst many
© C. Bodsworth and A. S. Appleton 1965 textbooks presenting detailed expositions of the derivation and inter-
First Published 1965
relation of thermodynamic quantities. However, it is often found that
students have difficulty in applying the basic concepts of thermo-
Printed in Great Britain by dynamics to practical problems. Frequently too, the more advanced
Spottiswoode, Ballantyne and Co Ltd worker in research is unaware that thermodynamic techniques are
London and Colchester relevant to his particular problem and often can be applied using only
the available data' without recourse to experimentation. These de-
ficiencies of training can be avoided or corrected if the formal study of
thermodynamic theory is supplemented by a detailed examination of
problems covering a wide range of applications. If these problems
require the evaluation of quantities or conditions of practical applica-
,2 4 7 .lt,JI 1
-c~ tion, the student will obtain a deeper appreciation of the value of
thermodynamics as a useful discipline and will be more ready to strive

zt,1/_ (1
towards mastery of more complex relations. This book is intended to
assist and stimulate students and research workers in these ways.
Although many of the examples have been drawn from metallurgical
,ewt applications, the principles demonstrated apply to a wide range of situa-
\ 91t:; tions and any student or research worker concerned with a field amenable
to the application of chemical thermodynamics should find the methods
C•f of use to him.
The basic thermodynamic relations are presented in chapter 1. It is
assumed that the reader is, or has been, familiar with most of these,
either from lecture course notes or from one of the available textbooks.
They are presented only in sufficient detail to facilitate cross-reference
with the worked examples. The treatments of dilute solutions, in terms
of alternative standard states, and of dilute multicomponent solutions
are presented in more detail, as these concepts are covered adequately in
very few textbooks. Each of the six chapters following contain worked
examples which illustrate the application of a particular thermodynamic
concept or group of concepts, and are designed as a progressive course
for students taking either University or Technical College courses in
thermodynamics. Most of the examples are numerical and their evalua-
tion demonstrates the application of thermodynamics to a practical
situation or the derivation ef additional thermodynamic quantities either
from experimental results or from standard tabulated data. Generally,
the examples in each chapter are v of increasing c~plexity, and the
 •
1
I I t' ~i i ' . i I i ; I, ' ,I ! ~ 1:

arrangement of chapters follows the sequence in which thermodynamic

, ,·: ; .: , . , , ;:I.,
~u ,IJ'REFACE

concepts arc most commonly introduced in a textbook or lecture course. for some of the exercises are not so frequently encountered by many
Un worked exercises are given at the end of each chapter and the answers, science and engineering students. Treatments of these (adequate for tho
together where necessary with guidance in their solution, are given in an the purposes of this book) are given in an appendix.
appendix. The uncertaintyoj' thermochemical data is often large and precise
The worked examples begin with a statement of the problem, but in calculations are not warranted. Consequently the use of a good ten-inch
only a very few cases are the data required for obtaining the answer given slide rule is entirely satisfactory (and indeed recommended) for calcula-
in this statement. This. practice, also used in the unworked exercises, tion of most of the solutions to the worked examples and the exercises in
differs from that usually followed in examinations where, from necessity, this book. Where calculations require a higher precision, the necessity of
all the information required is given on the question paper. However, it using five-figure Jog tables has been noted. In general, the answers to tho
is more realistic of the situation which is encountered in the practical worked examples and to the exercises have been rounded off to the
application of thermodynamics for, in the latter, it is necessary first to nearest 100 cal for quantities up to IO' cal, and to 1 kcal for larger
decide on the method of solution and then to determine the availability quantities. In all cases, this practice is more than justified by the un-
of the data required for this solution. Finally, having obtained these certainties in the available data.
data, the numerical calculation can be performed. This style of presenta- The authors are indebted to their colleagues for comments on, and
tion of exercises is designed to develop a more analytical approach to the criticism of, the contents and presentation of the manuscript.
application of thermodynamics than that required when the necessary
data are presented in the statement of the problem.
The application of the Gibbs-Duhem relation to data for components
of a binary solution is widely practised, and various ways in wMlh such
application can be made are demonstrated in chapter 5. In recent years,
a number of procedures have been outlined for extension of this applica-
•
tion to ternary systems, but the detailed derivation and method of
application of these procedures are available only in the original papers.
In view of the wide potential application of these treatments, three of
· these procedures are outlined in chapter 8 and used to evaluate the
activity of a component in a ternary solution. From these examples, the
research worker should be able to decide on the most useful procedure
for application to a particular set of data or, alternatively, the most
suitable ternary compositions to examine experimentally for ease of
treatment of the data.
A number of compilations of standard thermodynamic quantities are
available, and some of these are referred t~iJ]. the text. However, all the
data used in the worked examples and r~q-wred for the unworked
exercises are tabulated in an appendix, most of these data being taken
from Metallurgical Thermochemistry by Kubaschewski and Evans.
Acknowledgement is made to these writers and their publisher, Pergamon
Press, for permission to reproduce the data.
For the greater part of the text a knowledge of mathematics to an
understanding of elementary differential and integral calculus is required.
A few of the mathematical procedures used in the examples and necessary

.) .
~
'
 ''
Contents
, I.:•,

Preface V

Notation xi

Chapter 1 Basic Thermodynamic Relations

Including Enthalpy, Heat Capacity, Entropy, Free Energy,
Activity, Gibbs-Duhem relations; applications to concentrated
and dilute solutions, alternativ« standdrdstates, multicomponent
solutions ' '' · · · .. · ·

Chapter 2 Enthalpy and Heat Capacity . 25

Worked examples of the application of Hess's Law, heats of
reaction at elevated temperatures, combustion problems
I
Chapter 3 Entropy and Free Energy :' 48
Calculation of entropies and free energies of formation and of
reaction at ambient and elevated temperatures 1 · ,

Chapter 4 The Equllibrlum Co~tant 61

Determination of equilibrium constants, standard free energies
of reaction and activities from experimental data. Estimation of
the equilibrium conditions for various types of systems !;
. I ' . , ~ Ii
I:
Chapter 5 Solutions
Calculation of the various integral and partial molar quantities
for ideal, regular and general solutions. Applications of the
Gibbs-Duhem relations
93
I;
! ,

Chapter 6 Dilute Solutions and Alternative Standard States 120

Determination of the free energy of dilution, conversion of
activities from one standard state to another, determination of
interaction coefficients and of equilibrium conditions for a
reaction when components are referred to various standard stales

Chapter 7 Evaluation of Experimental Data 134

Derivation of hasic relations and illustrations of their application
to calculate thermodynamic properties from (a) the reversible
e.m.]. of a galvanic cell, and (b) Iiquidus and solidus data from
constitutional diagrams

Chapter 8 The Ternary Gibbs-Ouhem E.quatlon IS8

Tireapplication of the Gibbs-Duhem relation to three component
solutions by (a) and (h) graphical methods and (c) an analytical
method

--·- .•. ----·-·-

x CONTENTS

Appeod,ix l Answen to Exerdla 180

Appeodlx 2 C-ODSUlats ud Ulefal ec.nnlon Facton 200 Notation
Appendix 3 Taba,-tllcl Thermochemical Data 201
The notation selected for the various quantities are those in most common use
Appeadlx 4 Atomic Welpts of. Stable E'Jements 206 and are largely based upon the recommendations of the International Union
of Pure and Applied Physics (1955) and the International Union of Pure and
Appeodlx 5 Mathematical Opendom 207 Applied Chemistry (1955). Some changes have been made, however, to avoid
confusion where a symbol has been commonly used to represent more than
one quantity.

Symbol , Meaning Common Units

A Helmholtz function cal mole-1
a Activity
C Molar heat capacity cal deg-1 moic-1
E Eloctromotlva. rcirce volt
F Faraday's constant cal volt-1 g-equiv-1
(coulomb g-equlv+')
I Fugacity attn
f Activity coefficient (Henrian)
(~ Interaction coefficient; the effect of ele-
ment X on the Heorian activity coeffi-
cient of element Y
G Gibbs free energy cal mole-•
H Enthalpy or Heat content cal mole-•
K Equilibrium constant
In Natural logarithm; log.
log Common logarithm; log10
Lr., Latent heat of fusion, transformation cal moJe-1
M Molecular weight gram
N Mole fraction
n Number of moles
P Pressure atm
p Partial pressure atm
Q Heat absorbed cal
R Gas consta. it cal deg-I
S Entropy cal deg-1 mole-•
T Temperature degree Kelvin
f./ lntamal Enero cal molo·•
V Volume crn3 or litre
W Work done by a system cal
z Electrochernlcal Valency
Xl Standard thermodynamic property of a
component A
24 Partial molar thermodynamic property of
a component A

x:
u
xil NOTATION CHAPTER 1
Symbol Meaning Common Units
XM Integral molar thermodynamic property of

X '1
mixing
Partial molar thermodynamic property of
mixing (relative partial molar thermo-
Basic Thermodynamic Relations
dynamic property) of a component A
X! Excess partial molar thermodynamic pro-
perty of a component A
y Activity coefficient (Raoultian) The internal energy ( U) of a system is defined through the First Law of'
yO Activity coefficient (Raoultian) at infinite
dilution Thermodynamics:
U-Q-W 1.1
t"~ Interaction parameter c3lnfx
i cl Ny
in terms of Q the heat transferred into the system and Wthe work done
v . 8logfx by the system. For an infinitesimally small change:
ex 1 nteracnon parameter owt % y
µ. Chemical potential cal mole"! dU = SQ-SW, 1.2
(g) Gas ~I •
(I) Liquid :1 • or, when only mechanical work is done against an external pressure P:
(s) Solid dU = SQ-PdV 1.3
where ( +) d Vis the volume increase of the system.
The enthalpy (H) of a system is defined as:
H = U+PV. 1.4
For an infinitesimally small change:
:\
·( dH = dU+PdV+ VdP 1.5
;; which, on addition of 1.3, gives:
:i
.,
·'
dH= SQ 1.6
··i'
,. at constant pressure. That is, the enthalpy change in any reaction at
i
; constant pressure equals the heat of reaction. Thus for tho reaction:
I
mA+nB = pC+qD
I! the enthalpy change is:
,l
JH = JQP = ~ H(Products)-I: H(Reactants) 1.7
= pHc+qHo-mHA -nHB
where He, H0, etc., are the enthalpies per mole of components C, D, etc.
Conventionally, the enthalpies of pure elements are zero at 298°K and
the heats of reactions in which compounds are formed from their com-
pone-it elements at 298°K are the Standard Heats of Formation of the
compounds. Values of these are listed, for example, by K. K. Kelley [l]
and Kubaschewski and Evans [2].
 l BASIC THERMODYNAMIC RELATIONS BASIC THERMODYNAMIC RELATIONS 3

Heats of reaction have additive properties. Thus, if two successive for a system with a generalized equation of state U =- f(P, T, V). For an
reactions A and B, with heats of reaction (or enthalpy changes) JHA and ideal gas, defined by the equation of state PV"" RT, it is observed
LI Hs, are equivalent to an overall reaction C, then the beat of reaction is experimentally that (o(./foV)r = 0 and substituting this in 1.13 yields:
given by: CP-C., = R. 1.14
AHc - AHA +AHB. 1.8
The variation of CP with temperature is adequately expressed for most
This relation is called Hess's law. substances by an empirical relation of the form:
The heat capacity ( C) of a system is defined as:
C, = a+bT+cT" 1.15
c- SQ 1.9 where a, b and c are constants for a particular substance.
dT The change in the enthalpy of a system due to a change in its tempera-
where iQ is the quantity of heat absorbed when a system increases in ture is given by integration of 1.11:
temperature by an infinitesimally small amount dT. Hence, from 1.3, the To

heat capacity at constant volume is given by: Hcr,-r,> = J CPdT 1.16

T,

c, -(~~t 1.10 which, on substitution of 1.15 giv--:

r,
and at constant pressure, from 1.6, by: Hcr,....• r,> = J (a+bT+cT")dT. 1.1'!

c,-(:;t 1.11
T,

The heat of a reaction (.d H) is dependent on temperature, the variation

being given by the change in heat capacity in the reaction. Thus it follows
_ (au)
ar
+P(av)
arp
. p
1.12 from 1.11 that:

( o.dH) = ..::lCp, 1.18

Subtracting 1.10 from 1.12 and substituting identical differentials" yields: er P

(the Kirchhoff' equation) where

;' I
A c, = L Cp (Products)- L CP (Reactants). 1.19
r
',~
I:
I ,

Substitution of: In indefinite integral form:

(X = }· (~;) P = temperature coefficient of volume expansion

LIH == J ..::lCpdT. 1.20
This equation is solved by substitution of the combined equations 1.15
and: and 1.19. The heat of reaction at temperature Tis:

f3 = -t·(~~t - coefficient of isothermal compressibility

'1Hr = J ACpdT
yields:
= 5rt;;~"""
L (a+bT+cT")-L (a+ bT+cT")]dr
ROM:Wl~
(X; VT '"" __
b P_4.J \T1
-"" __b-RJ_
c,-c"- -/1- = "1H0 +~ ap- L aiJ T +-'-""'

• Some of the . fundammtal relati~

appendix S. '
of partial differentiationare outlined in
· . i. + (L Cp- L ciJ 1""'0 .
11+1

l
V
4 BASIC THERMODYNAMIC RELATIONS BASIC THERMODYNAMIC RELATIONS S
The value of the integration constant AH0 is obtained if AH is known at must be incorporated at the transformation temperature. Combination
one . temperature (e.g. 298°K). Alternatively, equation 1.18 can be of J .27 and J .26 applied to the components of a reaction yield a relation
ell pressed as a definite integral: analogous to J .22 for the difference in the entropy of the reaction at
T, temperatures T1 and T2:
LJHr,-AHr, = J A~PdT 1.21 T, T,
T,

bp - L bR)
AS -AS = f Lf <?eiTi-+To)
T .
dT+
- T,
.S+f L1Cp(Tr-+To)
T .
dT.
.
J 28
.
= (L ap-L aR)(T2-T1)+( L 2 2 T, T,
-- 2-·-- (T2-T1) T, T,
The Helmholtz free energy (A) of a system is defined by the relation:
+(L c:~ ~ CR )c:rri>_T}->+I)).
A= U-TS 1.29
If any of the components undergo a phase change- (at temperature T,
which, when differentiated completely and added to 1.24, yields:
with a latent heat L, where T1 < T1 < T2) then 1.21 becomes:
n ~ 1' dA = -SdT-PdV. 1.30
,1 HT, - ,1 HT, = f L1 c,
Ti
(T,-+T,). dT ±Li+ f Ll
T,
~p (T,-+T,).
\
d1'. 1 .22 _Thus isothermally, dA = - Pd V, or the reversible change 'in the
Helmholtz free energy equals the isothermal work content of the system
The entropy (S) of a system is defined through, application of the (maximum work function).
Second Law ofTherf1odynamics, such that the change in entropy when The Gibbs free energy ( G) of a system is defined by the relation:
the system undergoes a reversible process is given by:
G=H-~ 1~
dS = oi 1.23 = U+PV-TS . 1.32

where SQ is the reversible heat transferred at the temperature T. Com- . Complete differentiation of 1.32 and addition of 1.24 yields:
bination with J .3 yields: dG = VdP-SdT. 1.33
dU = TdS-PdV 1.24 The isobaric variation of the Gibbs free energy with temperature
for a system undergoing a reversible process where only mechanical follows, for:

( aG) -S.
work is done. =
The entropy change when a system changes its temperature is given by ar p
integration of J .23 and substitution of 1.9. Thus, at constant volume: Substitution of 1.31 and rearranging differentials leads to the Gibbs--
To

S0(T1-+To) - f
T,
;.dT, 1.25
Helmholtz relation:

1.34
and at constant pressure:
To

Sp (T,-+To)= f i.
T,
dT. 1.26
or, in indefinite integral form:

i= -f ;.dT. 1.35
When a phase transformation (latent heat Li at T;) occurs in the system
This is solved on substitution of the indefinite integral form of 1.17 and
in the temperature range considered, the entropy change:
integration of the resultant relation:
LIS, -
Li
ft 1.27
G = -T fi 2[ Ho+ J (a+bT+cT").dT ldT 1.36
2
6 BASIC THERMODYNAMIC RELATIONS PARTIAL MOLAR QUANTITIES 7
to give: and the free energy per mole of the system (the integral molar free
G = f(T)+Ho+IT energy) is:
where H0 and I are integration constants determined from a knowledge 1.42
either of G at two temperatures, or of G and Hat one temperature. i
I where N1 is the mole fraction 'of component i.
Alternatively the variation of G with temperature is obtained by
substitution of 1.17 and 1.26 in the defining equation 1.31. Thus, in I' Subtraction of the complete differential of 1.41 from 1.40, expressed
as a partial differential equation, yields the Gibbs-Duhem equation:
indefinite form: l
'{ ~ n;dC; = ~ N;dCi = 0. 1.43
G ••• f C,dT-T.J 7-.dT. 1.37
Partial (and integral) molar quantities arc related by the equations
Again, integration yields two constants which are fixed by a knowledge previously derived for a system of fixed composition. For example:
of G at two temperatures, or of G and Hor Sat one temperature. 1.44
It may readily be shown that when a reaction or process takes place in and:
any system, the change in free energy of the system is given by:
1.38
ata (Ci)
T p =
B1
-T2•
1.45

where the inequality sign applies when the system reacts and tends For two and three component systems the values of the partial molar
spontaneously towards equilibrium, and the equality sign applies when quantities at a particular composition are readily found if values of the
the system reaches the equilibrium state. Thus in any real process, the
free energy of a system decreases, reaching a minimum value when
equilibrium is attained.

Partial molar quaatlties"

The total free energy of a multicomponent system (containing n 1,., "B, nc, ••
etc., moles of components A, B, C, etc., respectively) is given by:
G' -f(P, T,n>.,nB,nc, ... ). 1.39 f
e
When a small quantity 8n1 of component i is added to the system at fT

-tL
i'a
constant pressure and temperature, the total free energy of the system
changes by an amount 8G'. The ratio of these quantities (8G' and 8ni),
when both tend to zero, is defined as the partial molar free energy of
component i. That is:
A
i 11
8

(8G'
- ) (aG'
- )
Na-
lim - ••• G1 1.40 Fig. la. Graphical determination of partial molar quantities from
a,,,...+-0 Sn1 P,T,1t1 an1 P,r •••• - • integral molar quantity in a binary system.
Hence, for a system at constant temperature and pressure, the total free
energy of the system is given by: integral molar quantity are known over an adequate composition range
which includes that particular composition. Thus for a binary system,
1.41 complete differentiation of J .42 and subtraction of the Gibbs-Duhem
relation 1.43 yields:
• This section is given in terms of the Gibbe free energy (G), but it applies equally
to any of the state variables of the system.
 THE EQUILIBRIUM CONSTANT 9
II BASIC THERMODYNAMIC RELATIONS

Multiplication throughout by: For a ternary system (represented by an equilateral right triangular
prism) the intercept of the tangent plane to the integral molar quantity
-NA = --(as
dNB
NA
dNA
dNA = -dNB)
.
r surface with the axis N; = 1 gives the value of the patrial molar quantity
I for the component i at the contiguous composition (Fig. lb).
and elimination of CA by addition of 1.42 yields: !'·
In the study of multicomponent heterogeneous systems, the partial
dG molar free energy is of special significance. The necessary condition for
CB= G+(l-NB) dN~. 1.46
chemical equilibrium in such systems is best expressed in terms of an
independently defined quantity, the Gibbs chemical potential. If, for a
system containing an arbitrary number of moles, oU' is the change in
the internal energy of the whole system when on; moles of component i
are added, then:

Jim
&r,-o Sn;
(SU') = (~U'
on;
) = µ.; 1.48

where µ.1 is the Gibbs chemical potential of component i.

It may readily be shown that chemical equilibrium exists between
phases of a multicomponent system when equality of the Gibbs chemical
potential for any particular component exists between the phases. Thus

T
for a system of components I, 2, 3, etc., consisting of phases A, B, C, etc.,
at equilibrium:
I µ.f etc. }
I
I
( Tangent
µ.(' = µ.~ =
µ.f = µ. ~ = µ.<j etc.
1.49
I pion•
I xc Further, expression of the Gibbs chemical potential is possible in terms
I
of any of the extensive state variables. That is:
- I

cl
Xg I

NA I I
I
I
1
µ.; ( ou')
= an; s', V',n, =
(oH'
an:
) S',P,n, =
(oA')
On; T, v· ••• ,
I

= (~C!:.)
on; T,P,,.,
which by definition = C;. 1.50

~ Thus the Gibbs chemical potential is equivalent to the partial molar free
Fig. I b. Graphical determination of partial molar quantities from integral molar energy of the component, and equality of the partial molar free energy
quantity in a ternary system, of a component in the various phases of a heterogeneous system is a
Similarly: necessary and sufficient condition for these phases to be in chemical
1.47 equilibrium.

Thus, as shown in Fig. Ia, the partial molar quantity for a component i, The equilibrium constant
at a particular composition, is given by the intercept of the tangent to
From 1.33, for a system doing mechanical work only:
the curve for the integral molar quantity at that composition with the
axis N; = 1. · ' dG = VdP-SdT

\"
 ~ SOLUTIONS 11
IO BASlC THERMODYNAMIC RELATIONS
~ where KP is the equilibrium constant for the reaction and LlG0 is tho
or, isothermally for l mole of an ideal gas (defined by PV =RT): J standard free-energy change for the reaction (i.e. the free-energy change
dG - Vd.P
dP
- RTP = RTdlnP. 1.51 .I
:1
when all c .nponents are in their standard states).
The variation with temperature of the equilibrium constant for a
reaction is given by substitution of 1.58 in the Gibbs-Helmholtz
equation 1.34:
A quantity, the !ugacity (f) of a real, non-ideal gas is defined such
d In k ) · · Ll Ho
that JV= RT and hence:
1.52
1i .
( dT P P = RT2 (the van't Hoff equation) 1.59
dG"" RTdln/ '
where L1H0 is the standard enthalpy change for the reaction (i.e. the
lf-PasP-O)

'
enthalpy change when all components are in their standard states).
for a real gas. Solution of 1.59 is obtained by substitution of 1.20 and integration as an
Integration of 1.52 at constant temperature and pressure between the ~
0 ~p indefinite or, when possible, as a definite integral. As an indefinite
state of interest and a standard state designated by the superscript
1 integral:
yields:
I 1.53 lnKP = R AH 0+AalnT+AbT+LlcT" )
1 ( I+T 1.60
G-G0 = RTin I°. ~
liJ: where /JH0 and I are integration constants which are found either by
The ratio, /JI? for a substance i, is called the acthity (a) of the sub- II
insertion of known values of KP at two temperatures, or of KP and AH
\'
stance. Thus: 1! at one temperature, in similar manner to the solution of the free energy-
1.54
temperature relation 1.36.
where the partial molar free energy (Ou is the free energy of the substance Ii Over a small temperature interval, it is often permissible to assume
i in the state of interest, and the standard molar free energy o<I is the free it
~ that JH0 is independent of temperature. In this approximate form, l.59
energy of i in a standard state, usually (although not always) chosen as can be written as:
the pure component.
The applicability of the activity concept is demonstrated by considera-
tion of the generalized reaction:
! j
or:
1.61

bX+cY +dz+ ... ~ mP+nQ+lR+ •..

where b, c, d, m, n, etc., are the number of moles of components X, Y, Z,
P, Q, etc. The free-energy change for this reaction is given generally by:
Alternatively:
LlG = (mGp+nGQ+lCR+ ... )-(bCx+cCv+dCz+ ... ) 1.55
an» = 4·575 T1 T2(logK1 -logK:z) 1.62
or, more specifically, when all components are in their standard states: I T1-T2 ..
LlG0.,;, (mG~+n~+ldlt + ••. )-(b~+c~+dG~+ ... ). 1.56 Solutions
·1
Subtracting 1.56 from 1.55 and substituting 1.54 for each of the com- The difference between the free energy of a component i in a solution
.! and in the rho sen standard state is given by 1.54:
ponents gives, at equilibrium (when LlG = 0):

.d 00 = - RTln a'j: · Oo · 01it • •• 1.57 I .

J, C
U-x.Qy.Qz
d- .v .: For convenience, the difference on the left-hand side of this equation is
- -RTlnK ,, called the partial molar free energy of mixing or the relatin partial molar
12 BASIC THERMODYNAMIC RELATIONS SOLUTIONS 13

free energy of the component i, er

(or LlGJ. Partial tnolar entropies and and hence:
enthalpies of mixing are similarly defined, these being related through all
1.69
the expressions previously derived. For example:
er = R;M :-TS;M 1.63 Thus, the heat of mixing of an ideal solution is zero. Also, as from 1.33: )
( oG)
ap T = V,
and:

-
sr I 1.64 then:
T2.

Integral molar quantities of mixing are related to the partial molar ( acr)
s»
= vr
T
quantities of mixing by expressions of tile type 1.42. Thus the integral
molar entropy of mixing for a binary solution AB is given by:
a
= a"f,(RTlnNJ
SM= NASJ:!+NBS;t
for an ideal solution. But as the composition is independent of pressure,
= (NASA+NBSB)-(NAS1+N8Sg) 1.65 then:
from the definition of the partial molar entropy of mixing. The first term a .
in parentheses is the, integral molar entropy of the solution (S) and the oP (RTln N;) = O
second is the entropy of the unmixed pure components A and B. ,,
The study of solutions is facilitated by the use of the concept of the or: ·,i
ideal solution, originally defined by Raoult in terms of the equaiity of the 1.70
partial pressure of a component with its mole fraction-vapour pressure
product. Thus, for a component i in an ideal solution: and:

Pi= p~N;. v,., = LN;flt = o, 1.71

Replacing the partial pressure by fugacities (which removes the limitation that is, the molar volume of mixing of an ideal solution is zero.
of idea! gas behaviour of the vapours): From 1.68 and 1.31 applied to partial molar quantities of mixing:

f;
]? = N i = a; bdfi ..
y e rution. 1.66
s.M = _ er
I T
(Raoult's law)
for an ideal solution. Substituting 1.67 for er:
From this definition of an ideal solution, further special qualities are
derivable. Substituting 1.66 in 1.54: ·
S;M = -RlnN; 1.72

er= e;-~ = RTlnN; J.67

and the ideal entropy of mixing is given by:

for an ideal solution. Substituting this in 1.64: SM= L N;S;M

a n.M = - R L N; In N;. 1.73

arCR In N;) = --12 . The concept of the ideal solution is of limited usefulness as a means
of expressing the approximate behaviour of real solutions. However, the
As o/aT(RlnN;) = 0, t~ the composition of the ideal solution is
exact behaviour of real solutions may be expressed in terms of their
independent of tempefat~-;tll~
_...,,_ -;1,,...i:-..~ deviations from ideal behaviour. This expression may be made in several
/,iMI . .I\.,;~ 1.68 ways.
--- ·11 i
. .,,.
~-- "' r _;
,
• ! - -~-

. : ·' . ,
' ·•• J • - ;/.
It:,
. .. I
14 BASIC THERMODYNAMIC RELATIONS SOLUTIONS . 1!!

The activity coefficient for a component of any solution is defined by: on whether the solution in question shows positive (Yi> 1) or negative
(y; < l) deviations from ideal behaviour. These excess free-energy
a1 1.74 quantities are therefore very useful for describing the deviations of real
Y1= -·
Ni solutions from ideal behaviour.'
For an ideal solution a1 = N1 and 'Yi= 1, so the magnitude of the activity Many real solutions deviate from ideal behaviour in their enthalpy of
coefficient is a measure of the deviation of a non-ideal solution from mixing values (these being non-zero) but behave ideally with regard to
ideality. Solutions having values of y less than or greater than unity their entropy 'of mixing values. On this basis Hildebrand [3] defined a
deviate from ideal behaviour in a negative or a positive manner regular solution by the expressions:
respectively. HM= :E. NiHIM 'i' 0 l
It is sometimes more convenient to express the non-ideal behaviour
of a solution in terms-of one or other of the molar thermodynamic
and: ' I 1.80
properties of a component or of the whole solution. For this purpose,
s« = :E' NiSiM = stfca.1 = -R L NilnNi. J
partial and integral excess quantities are defined as the difference Substituting 1.54 and 1.81 in the definitional relation 1.31 applied to
between the value of the quantity in the real solution and the value the integral quantities o~;mixing yields, for a regular solution:
quantity would have in a hypothetical ideal solution of the same compo- HM.= GM+TSM
sition and under the same conditions of temperature and pressure. For
example, the excess partial molar free energy is defined as: : = RT'£, N11na1-RT1; N1lnNi

Cf (or Cf') = Ci -C{, 1.75

· = RTI, Ni In 'Yi 1.81
where y1 is the activity coefficient of component i. It follows, of course,
where C; is the partial molar free energy of the component i in a hypo- that for a regular solution:
thetical ideal solution. Thus, Cf= 0 when the component i is in an ideal
solution. Similarly, the excess integral molar entropy is defined as BiM ., RTlny1 1.82
1.76 which, by comparison with 1.79 yields:

where Sis the integral molar entropy of the solution under consideration
and S1 is the entropy of the hypothetical ideal solution of the same and:
B1M = Cf} 1.83
HM=(;£
composition. Thus SE= 0 for an ideal solution.
Excess partial and integral quantities are related through the thermo- for a regular solution. However, from 1.33 it follows that, generally:
dynamic relations previously derived, for example:
Cf= Bt-TSf, 1.77 ( oC.f)
oT P
= -S{
and (the Gibbs-Helmholtz. relation):
and so, for a regular solution, for which Sf, (from its definition and
o ((;£)
oT r
HE
= -72 ·
1.78 1.81) is zero, then:

Substituting 1.54 and 1.67 in 1.75 yields: (°;;jp = 0.

Cf= (c;4+RT1nau-(dl+RTlnN0 Hence it follows from 1.79 that for a regular solution the quantity
Rnn y1 is independent of temperature.
= RTlny1 1.79
The use of a functjon cz, defined by:
from which it is apparent that the sign of the excess partial molar (and lny,
hence excess integral) free energy will be positive or negative depending cz1 =.(1-NJ2 1.84

\~:
20 BASIC THERMODYNAMIC RELATIONS
ALTERNATIVE STANDARD STATES 21
solute as standard state) is a linear function of the weight-per cent
concentration of tho solute. For such a solute, the weight-per cent The extension of this relation to finite solute concentrations follows from
concentration can be used directly in thermodynamic calculations the non-dependence on the choice of standard state of the partial molar
without the necessity of converting analysed compositions into atom free energy of the solute, as described in the treatment of the infinitely
fractions (as is necessary when the standard state is defined in terms of dilute atom fraction standard state.
the infinitely dilute atom fraction). In dilute solution, the ratio Na/wt %B can be expressed as a real
The dilute weight-per cent standard state is defined such that: number for any given system, using the relation:
wt%B
. - 1 as wt,,0/"01- 0
01
wt %o1 MB
NB= wt%B + (100-wt%B)
or: a1 - ,vt%i(wtY.1__.0).
I Ms MA
If the activity equals the concentration up to I weight-per cent of com-
ponent i, then a;= I at wt %i-= 1 and this I weight-per cent solution is where MA and M8 are the atomic weights respectively of the solvent A
then the standard state. · and the solute B. For small concentrations of B, the first term in the
Deviations from this equality of activity and concentration are denominator is small compared to the second, and this relation may be
simplified to:
measured in terms of a Henry's law activity coefficient'{]"), defined by:
I Na~ wt%B
o .MA
i (;=-=~- 1.92 100 • M B '
I Wt%i Therefore:
As with the infinitely dilute atom fraction standard state, the free- ~= MA_. 1.9S
energy change accompanying the transfer of standard state from the wt%B 100.MB
pure solute to the dilute, weight-per cent solution is given by a relation Substituting 1.94 and 1.95 in 1.93 yields the free-energy change
identical to J .89. That is: accompanying the change in standard state:

L1G~ = RTin[0B(puro)] · , 1.93 B<Pure>- Bcwt x dilute>

aB(diluu) N11-comtant ';·

Proceeding in analogous manner to that used in the derivation of the 1.96

free-energy change to the infinitely dilute, atom fraction standard state, or:
at any chosen composition the activity of the solute is given by J .92, r..o r..o O MA
relative to the dilute, weight-per cent standard state, and by 1.74, relative V(diluto) = V(pure)+RTlny "JOO.MB 1.97
to the pure substance standard state. Hence the ratio of the activity Hence, by comparison with 1.93:
coefficients on the two scales is:
OB (pure) Y~ · MA •
YB (aBINB)pure _
[ l ---
aa (dil~te)
==
100. Ms
1.98
/; = (aBfwt %B)dilutj comtant composltlon
The term yg in 1.90, 1.97 and 1.98 is usually a function of temperature
and, at infinite dilution where fB = 1, YB= yt
the activity coefficient of and must be represented as such if free-energy changes are required over
the solute at infinite dilution relative to the pure substance standard a range of temperatures. However, frequently the value of~ is known
state. Therefore:
for only one temperature, and the assumption of regular solution
!:~_(/'.:'!.el]
[ Os (dilute) constant compoaltlon
:a [~. ':B ]
wt %B
•
COl\llalllcompoeltlon
1.94
behaviour must be made. At high dilution this is acceptable. Following
such an assumption, the partial molar heat of mixing of the solute at
infinite dilution [llBMJO can be regarded as a constant over a restricted
3
zz BASIC THERMODYNAMIC RELATIONS . MPLTI~OMPONENT SOLUTIONS 23
temperature range (see page 15). It follows from 1.82 that RTiny~ is a
constant, or In y~ is inversely proportional to the temperature. The
cientf* is found to be a logarithmic function of the concentration of Y,
but is independent of the concentration of X. Thus, this concentration-
standard free-energy change in 1.96 is then expressible as a function of dependence is expressed by:
temperature by:
1.102
1.99

where [ll~]0 = Rnny~.
If the pure solute normally exists in a different physical state to that
which it adopts in the solution at the same temperature and pressure, a
further term must be added to 1.90 or 1.97 to account for this change in
physical state, For example, if the pure solute normally exists as a solid
at a given temperature and pressure, but the solution of interest is a
liquid, the free-energy change accompanying the fusion process for this
solute must be added to the free-energy change for the standard state
transfer. As the free-energy of fusion (Ge) is ~ro at the true melting
point, therefore:
s, = Lr I.JOO
Tr
and:
Gr= Lr-TSr
where Sr and Lr, the entropy and enthalpy of fusion at the melting point
Tc, are assumed to be independent of temperature.
Multicomponent solutions
The thermodynamic behaviour of a particular solute in a solution is
usually affected by the presence of other solutes, even at low concentra-
tions. Thus the activity coefficient, fx, of a solute Xis both a function
of the concentration ofX and of the concentrations of the other solutes.
In a solution containing two solutes X and Y this interdependence can
be expressed as:
fx •• f~xf~ I.JOI
or:
logfx - logf~+logfJ
where fx is the activity coefficient of X at a given concentration in the
ternary alloy, f~ is the activity coefficient ofX at the same concentration
of X in the absence of Y · and f'£ called an interaction coefficient, is a
measure of the effect of Y on the activity coefficient of X. In most of the
dilute solutions which have been studied to date the interaction coeffi-
1.103
where the interaction parameters e~ and €* have values calculated on
the basis of the activity coefficients being defined respectively in terms
of the weight-per cent, or atom fraction, dilute solution standard states.
For dilute solutions containing more than two solutes, it is generally
found that values of the experimentally determined interaction para-
meters e~ or E* are unaffected by the presence in the solution of other
(third, fourth, etc.) solutes. Where this condition applies, Wagner [4]
has derived the following relation for the activity coefficient of a com-
ponent X in a multicomponent solution containing solutes X, Y, Z, etc.,
in terms of the interaction coefficients:
where the derivatives arc taken for the limiting case of zero concentration
of all solutes. However, at low concentrations, the terms involving
second- or higher-order derivatives become very small and can be
neglected. Making th.is approximation and substituting 1.102 or 1.103
for the first derivatives, the relation simplifies either to:
,' X Y Z
ln,x = Nx,Ex+Ny.*-x+Nz.€x+ ... 1.104
in terms of the atom fraction standard. state scale, or:
logfx wt%X.t{+wt%Y.e*+wti'oZ,e~
=;
,' 1.105
in terms of the weight-per cent standard state scale.
In practical systems, 1.104 and 1.105 can often be applied to represent
the interaction effects between solutes at concentrations up to at least
one atom- or weight-per cent These equations have commonly been
used to compute the activities or solutes in molten steels. However, the
errors arising from neglect of the higher order derivatives become
increasingly significant as - the concentration of solutes is raised and,

\V
24 BASIC THERMODYNAMIC RELATIONS CHAPTER 2
though equations 1.102 or 1.103 may continue to be applicable, equations
1.104 and 1.105 are usually not valid for other than relatively dilute
solutions.
Wagner has shown further that, at infinite dilution, the interaction
parameters are related by:
Enthalpy and Heat Capacity
1.106
X My y
ey = Mx·ex 1.107

where Mx and My arf the atomic weights of the solutes X and Y, which
relations may be made to yield values of the interaction parameters when
the latter are unknown or not readily amenable to experimental
determination. . against an external pressure (and as energy in a form other than heat,
References
I. K. K. KELLEY.Con1ributions to the data on theoretical metallurgy, X.,
U.S. Bureau of Mines, Bulletin 476, 1949.
2. o. K UBASCHEWSKI .and E. LL. !!VANS, Metallurgical Thermochemistry,
Pergamon Press, 1958.
3. J. H. HILDEBRAND, J. Am. Chem. Soc., 51, 1929, p. 66.
4. c. WAGNER, Thermodynamics of Alloys, Addison Wesley, 1952.
The heat change accompanying a reaction is an indefinite quantity, since
part of the energy of the reaction may be consumed as work of expansion
e.g. electrical energy). However, as enthalpy is completely defined in
terms of state variables ( equation 1.4), the enthalpy change in a reaction,
which is equal to the heat change at constant pressure when only
mechanical work is done (equation 1.6), is dependent only on the initial
and final states and not on the reaction path. The additive properties of
enthalpy, originally formulated empirically as Hess's law, follow as a
consequence of enthalpy being a variable of state. Thus, enthalpies or
enthalpy changes may be added and subtracted parallel with the same
manipulations performed on their respective components or reactions.
The enthalpies of the elements in their normal states are conventionally
made equal to zero at 298°K and one atmosphere pressure. The heats of
the formation reactions for compounds under the same conditions are,
therefore, the standard enthalpies, or heats of formation of these com-
pounds. Extensive tables of these values are available.
When heats of reaction are required at other than the standard
temperature it is adequate for some purposes to assume that the enthalpy
terms involved are independent of temperature and to use either the
standard ~alues or a mean value in the temperature range of interest. If
greater accuracy is required, heats of reaction at temperatures other than
298°K are calculated using the Kirchhoff equation (1.18).
The following examples are selected to illustrate various ways in which
these principles may be applied. As calorimetric work is usually carried
out under constant pressure conditions (or pressure effects are either
ignored or allowed for in the handling of the results) the readily available
quantities are enthalpy and heat capacity at constant pressure. However,
the principles used can, in theory at least, be applied equally well to
calculations in terms of internal energy and heat capacity at constant
volume,

I.I\
 c:gl 28 ENTHALPY AND HEA_T CAPACITY
HEATS OF REACTION AT ELEVATED TEMPERATURES 29

·fO The reactions 2.3.1 and 2.3.2 are the solution reactions for HCl and where JH0 is an integration constant, found by substituting:
y ZnCJ2, i.e. the enthalpy values are the heats of solution in water. For the LlH~81f6'r Cr2 6; :~- -270,000 cal mole=",
/ reaction between zinc and hydrochloric acid gas; i.e. i.e.
2.3.4

/~'f:r ,
/ Zn<•)+2HC1<s)~ ZnC12(•>+H2(s>
e heat of reaction is found (as in example 2.1) to be:
LlH298 = (-99·5)-2x(-22·0) = -55·5 kcal.
l:J Ho = -270,000-(2~·53 x 298)...:. (l · l
= - 279,900. ,
X

IO
;982)-(3·74 X

298
105)

~-
Reaction 2.3.3 can then be obtained by subtracting twice reaction 2.3.1 Therefore:
'i .
from the sum of 2.3.2 and 2.3.4. Hess's law gives: JHi = -279,900+Z8·53T+l-10x I0-3T2+3·74x I05T-1 cal mole-1
LlHciJJ>= ~H()JJ>+JHciJ."J-2iJH<2J.I> which may be solved ior: any v~l\.ie of T. Thus, at T = l ,900°C (2, 173°K):
••• -14·9-55·5-2x(-17·5) .f . ' . .
.dHf.173 = - 279\:900+(28·53 x 2,i 73)+(1 · 10 x 10-~ x 2,1732)
= -35·4 kcal mole=! (of zinc)
+(3·74,x_!05)
which is the heat of reaction when 1 mole of zinc is dissolved in 2 mole '2,173
of the HCl solution.
= .:...212,500cal moJ<::-1•
Heats of reaction at elevated temperatures
The following examples illustrate the application, in various forms, of METHOD 2:· Equation i.J7may be solved as a definite integral between
the Kirchhoff equation. the limits T= 298 an.a 2,173°1<..-::
Thus:
2.4 Evaluate (a) the enthalpy of Cr203 at l,900°C, (b) the heat required
2,173
io raise the temperature of 1 mole of Cr203from 10°C to 1,900°C.
(a) The solution of equation 1.16 to obtain the enthalpy of Cr203 at an
iJHtm-LJH~98 = f
298
CP.dT
elevated temperature requires knowledge of the standard heat of forma- 2,173
tion and the heat capacity of Cr 203• These are: = J (28·53+2·20x I0-3T-3·74x to5T-2]dT
Ll H~98 = - 270,000 cal rnole=! 298

CPeno, = 28·53+2·20x 10-3 T-3·74x 105 T-2 cal deg-1 mole=!
The calculation of the enthalpy or enthalpy increment at, or between
temperatures other than, the standard temperature of 298°K can be
carried out in two ways.
= [28·53T+l·l0x I0-3T2+3·74x 10sr-1]~73
The value of LlH~98 is given as -270,000 cal mole-1. Therefore:
.1Ht173 = -270,000+28·53(2,173-298)+ l·lOx 10-3(2,1732-2982)

METHOD 1: A completely general expression for the enthalpy as a func- +3·74x 105(2.:73-2!8)
tion of temperature is first obtained by solving 1.17 as an indefinite
integral and finding the value of the integration constant by substitution = -270,000+53,500+5,100-1,100
of the standard heat of formation and the temperature of 298°K. Thus: = - 270,000 + 57,500
iJHi = J CP.dT r: 21i500
(',
cal mole:".
= J [28·53+2·20x 10-3T-3·74x 105T-2]dT • While, in general, i:Cmperaturcsin this book are quoted to the nearest degree
Kelvin, it should be appreciated that, in practice, the accuracy of temperature
= JH0+28·53T+ 1-IOx 10-3T2+3·74x 1osr-1 cal mole"! measurement is often much less than this at high temperatures.
 l I l-"i . .,: > :' ~: : ,•; . ' : ;"' . 'Ii:- I
26 ENTHALPY AND HEAT CAPACITY
l ,,; . :.:.:. _. _'. - ·,.-,.·,,.. . ''J APPLICATIONS OF HESS'S LAW 27
I ,i, ..• : :

· · : ;··, -' · . 2.2 Calculate the standard heat of formation of solid Pb02from solid
Applications of Hess's law L.~,±.;_;;~=,;.:.~;w,,,;'._,~-·-··--- · - ---~=t-·"~'"'"'''
lead and oxygen gas at 1 atmosphere pressure, given the following data
2.1 Determine the heat of reaction of pure solid Na20 with HCI gas at valid at 298°K:
I atmosphere pressure to form NaCl and water at 25°C.
Reaction Heat of Reaction
The standard heats of formation in kcal mole-1 are:
Pb<sl + ;o2u> ~ PbO<sl -52·4 kcal 2.2.1
NaCl<•> -98·6±0·2 HClw - 22·0 ± O· l
3Pb0<•> + ,Oica> ~ Pb304<•> -18·4 kcal 2.2.2
H20Cl> -68·32±0·01 Pb304+0ica> ~ 3Pb02u> -22·7 kcal 2.2.3
The enthalpy quantities involved in the reaction are therefore:
' This is an example of a reaction which can be represented as being
Na20<•> + 2HCJw ~ 2NaC~,> + H20c1> made up from a series of reactions for which data are available. The
'(-100·7) 2 X (-22·0) 2 X (-98•6) (-68·32) overall reaction is:

The heat of the reaction is equal to the difference between the enthalpies Pb<•> + 02{c) ~ Pb02C•> 2.2.4
of the products and of the reactants (equation 1.7), i.e.: this being equivalent to the sum of equation 2.2.J and one-third the sum
'1H2g8 = (2 X ( 198·6)+(-68·32))-((-100·7)+2 X (-22·0)] of 2.2.2 and 2.2.3. Following Hess's law, the heats of reaction are also
additive, i.e.
= -120·82 kcal.
LlH/Ji.u, = AHc~.:1.1J+tLlHc~.1.n+tiJH8.1J>
The errors in the individual enthalpies are additive," giving an overall
error of: -52·4-6·1-7·6

1·2+(2x0·1)+(2x0·2)+0·01 = ±1-81 kcal. -66·1 kcal.

Hence:
Historically, heats of formation are always quoted as exothermic, or
..1 H'k ""' - 120· 82 ± 1-81 kcal. · . negative heat, quantities. Thus the heat of the reaction for the formation
Because ..1H0 is negative, the 120·82 (± 1·81) kcal are evolved when of lead dioxide from its components at 298°K is -66· l kcal, arid the
I mole of Na20 and 2 mole of HCl gas are reacted completely at 25°C standard heat of formation of solid Pb02 is 66· l kcal mole-I,
to form 2 mole of NaCl and I mole of water. Thus the heat of reaction
is equal to: 2.3 Calculate the heat of reaction when l mole of zinc is dissolved in a
l,
HCI (IOOH;:O)solution, given the following data:
- I 20·82 ( ± l ·81) kcal per mole of Na20 or of H20
or: HClw+lOOH20c0~HCl(IOOH20)c!): ..1H298=-·17·5kcal 2.3.1
-60·41 ( ± 0·91) kcal per mole of HCl gas or of NaCl. ZnCl2(•> + 200H20co ~ ZnCi2(200H20)0> ..1H298 = -14·9 kcal
If the heat of reaction per gram of one of the components is required, 2.3.2
the molar heat is divided by the molecular weight of that component. ZnCJ2<•> LJH~98 = -99·5 kcal mole-I

• This is one of the major disadvantages of the application of Hess's law. In this
HClw LJH~8 = -22·0 kcal mole-I
example the error is only l ·5 per cent of the heat of reaction, qut when the latter is
small the error in its determination by this means may be of comparable magnitude, The reaction in question is:
and very much larger than that obtained from a direct experimental determination
of the enthalpy change. Znc,>+2[HCJ(IOOH20)Jcn ~ ZnCl2.(200H20)cn+H2(.r> 2.3.3
 30 ENTHALPY AND HEAT CAPACITY HEATS OF REACTION AT ELEV A TED TEMPERA TURES 31

The two methods are, of course, equivalent and require the same
information for solution. The first method is preferable when enthalpy
values are required at more than one temperature.
(b) The calculation of the enthalpy increment when a substance is
heated from one temperature to another does not require knowledge of
tho standard heat of formation. The amount of heat required to raise the
temperature of 1 mole of Cr203 from 10°C to l,900°C is:
2,173
proceeds only until the latent heat evolved ~uses the temperature of tho
bath to rise to the true melting point, when solidification ceases. Thus,
the heat balance arises between the heat evolved during solidification and
the heat required to raise the temperature of the bath (now a mixture of
liquid and solid copper) to the true melting point; i.e. through the 5°C
of supercooling. This quantity of heat is given by:
l,H6
f 7·5 dT = 7·5 x (1,361-1,356)

L1H0 = L1H~.113-AH'i,g3 = f
283
CP.dT
plus:
1,3,1
= 37·5 cal mole=! of liquid copper

2,173 I l,3S6
= f [28·53+2·20x 1Q-3T-3·74x J05T-2]dT !·
l,3SI
f (5·41 + l ·5 >;, I0-3 T)_dT
.
.
.
= [5·41T+0·75 x l0-3 T2]1•1,3,t
3.56
283
= 37·2 cal moie-1 of solid copper.
= 28·53(2,173-283)+ HO x I0-3(2,1732-2832)
As for every moleof copper which solidifies, 3,100 cal of latent heat
+3·74x 105(2,:73-2 3) ! are evolved, this will raise the temperature of:

= 53,900+5,100-1,200
= 57,800 cal mole=".
In the previous calculation (a), method 2, the enthalpy increment
between the standard temperature 298°K and 2, 173°K was calculated
(i.e. as above) and added to the standard heat of formation (AH'k) to
give the enthalpy of Cr203 at 2,173°K. The difference between the two
enthalpy increments (i.e. 57,800-57,500 = 300 cal mole=") is, of course,
equal to the enthalpy increment between 283 and 298°K.
/. , to form ferrous oxide (FeO) at l,350°C.
.. [3,100-37·2+1] = 82·7
: 37·5
mole of copper backto the melting point. That is, 81·7 mole of liquid
copper and 1 mole of solidified copper can be heated from 1,078 to
l,083°C by the laterit heat evolved on the solidification of I mole of
copper. Therefore, the fraction of the bath which solidifies is 1/82·7 or
1 ·21 per cent.
2.6 Find the heat of-reaction when y-iron is oxidized by pure oxygen gas

;;
~A b';;J~'-~~f ~6);~~J/dp]er is supercooled to 5°C below its true melting The tabulated data for the heat of ·oxidation of iron apply at 298°K,
point. Nucleation of solid copper then takes place and solidificationproceed, but between this temperature and the required temperature iron under·
under adiabatic conditions. What percentage of the bath solidifies? goes a ferromagnetic to paramagnetic change at '760°C and a phase
The following data are required: transformation (from er: to y form) at 9l0°C. To calculate the high
temperature heat of oxidation of iron it is necessary, therefore, to use
The latent heat of fusion of copper (Lr) = 3,100 cal mole:" at l,083°C. relations of the form of equation 1 .22. Between each of the transforma-
CPcuro = 7·5 cal deg-1 mole-1 tion temperatures the enthalpy increments are calculated by essentially
the same procedure as that outlined in example 2.4a, method 1, i.e., the
cl'Cu(J) = 5·41 + l ·5 x 10-3 Teal deg=!
integration constant, A H0, in the general equation is evaluated from a
mole-1•
known value of A Hi, valid in the temperature range of interest. Hence
This type of problem is essentially a variation on the classical heat these calculations are shown only in outline and emphasis is given to the
balance problem of dropping a hot object into a bucket of water, etc. procedure for adding or subtracting the latent heats at the various
Because no heat is exchanged with the surroundings, solidification transformation points.
 'I
32 ENTHALPY AND HEAT CAPACITY
lI HEATS OF REACTION AT ELEVATED TEMPERATURES 33
The standard heat of formation ofFeO.is given as -63·2 kcal mole-1,
i.e. for the reaction: I the 'latent heat of transformation' at 760°C is given as:
II
' -------·---------------(·· I LIH~ = + 660 cal.

the heat of reaction is:

~c«> + to2(a)
---.---·-·
~ FeO<a> 2.6.1
:! The heat of reaction 2.6.1 at 760°C(I,033°K) is, (from 2.6.2):
LJH?,033 = -62,600 cal.
iJH'jgg = - 63,200 cal.
Applying Hess's law at 1,033°K to reactions 2.6.1 and 2.6.3 yields the
The Kirchhoff equation 1.20 is used to extend this value to higher
reaction: ---·-· -- ·------------,
temperatures. The relevant heat capacity relations are: 2.6.4
/ Feca:, paf,..mq.)
L-- .
+ !02(.a) ~ FeOc,>/
.. ....
CPF.<a:. .....,,.uc, = 4·18+5·92x l(r3Tca1deg-1 mole~t for which the heat of reaction is:
Cl'P.ou, = 11 ·66 +2·00 x 10-3 T-0·67 x 105 T-2 cal deg-t mole-1
LJH?,033 .,. LlHc~.6.1>-AHc~.6J>
CPo,~, = 7·16+ 1·00 x I0-3T-0·40 x 105 T--:2 cal deg=! rnole-:'. = - 62,600- 660
Thus, for reaction 2.6.1:
= - 63,260 cal.
LICP = CPFe()(o)-cpP.(a:,mao,euc)-!C;.o,w As Feea, para.ma,.) has a different heat capacity-temperature relation to
= 3·90-4·42 x 10-3 r:....0.47 x 10s r=. Feca:, rerro-maa.>• i.e.
Substituting this in 1.20: CPFc <a:. not>-1nq.) = 9·0 cal deg-1 mole-1,.
the heat capacity change in reaction 2.6.4 is different to that in 2.6.1. The
L1H¥ = J .:1Cp.dT value for 2.6.4, obtained from the previously quoted values of CPF.ooo and
= LIH0 + 3·90T-2·21 x 10-3 T2 + 0·47 x 10s T-1 Cpo.~, with the Cpy <a: . .,.....•••••., is:
0

LJCP = -0·92+1·50xI0-3T-0·47xlOST-2.
where iJH0, the integration constant, is evaluated by substituting
L1H~98 = -63,200 cal. Hence the enthalpy change for reaction 2.6.4 is:
Hence: .dHi = f LJCP.dT I
iJHi = -64,300+3·90T-2·21 x I0-3T2+0·47x I05T-1 cal. 2.6.2
= .JHo-0·92T +0·75 x 10-3 T2+0·47 x 10s r-i.
At i@~_G) e-Fe changes from the ferro-magnetic to the para-
Solving for .JH0, the integration constant, by substitution of
magnetic state (i.e. the Curie point). This change is accompanied by an
.J H? 033 = - 63,260 cal gives:
absorption of heat over a range of temperature from about 300° to
900°C, which is evidenced by an abnormally high heat capacity for iron .dHi = -63,160-0·92T+0'.75x I0-3T2..-0·47x 105T-1.
over this temperature range and is characteristic of what is known as a At 1,183°K, the temperature of the oc-+y transformation in iron, the
second-order transformation. However, when the enthalpy of iron is heat change for reaction 2.6.4 is:
required at a temperature above 900°C, as in the present case, the heat
change associated with the magnetic change can be represented by a .JH?.1s3 = -:-63,160 cal.
'latent heat of transformation' (4) which is absorbed at a constant The transformation from oc to y iron is a typical 'first-order trans-
temperature of 760°C (the temperature of maximum heat capacity). formation" with a true latent heat of transformation of + 220 cal moie-1,
For the reaction: which is absorbed at a constant temperature of 910°C (1,183°K); l.e.
2.6.3 ' -----·-· ·_.· .... --·1. 0 - -1
(!eca:. para-ma_~:) r-_Fecyl_J· .1H - +220 cal mole • 2.6.5
34 ENTHALPY AND HEAT CAPACITY HEATS OF REACTION AT ELEVATED TEMPERATURES 3!!

Applying Hess's law to equations 2.6.4 and 2.6.5 at 1,183°K, for the calculated. The heat in these gases ii partly transferred-to the-furnace and
reaction: ,,-------- partly carried off in. the exhaust. Calculation of the latter quantity will
cfo 118'?>"(.''< 2.6.6 also give, by difference; the former - the heat supplied to the furnace.
. . \ -~~~~~,Ol(a) ~ FeO~~j The required data are:
the heat of reaction 1s: · ,·
For CO: Standard heat of formation = - 26,400 cal mole+!
LJH?,113 -= .dH<~.6.,()-LJH&.6.5) Heat capacity CP = 6·79+0·98 x I0-3T-O·ll x 105T-2 cal
deg-1 mole-1
= -63,160-220
For CO2: Standard heat of formation = -94,050 cal mole-1
= - 63,380 cal./
Heat capacity CP= l0·55+2·16x I0-3T-2·04x 105T-2 cal
The heat of capacity of y-iron is given by: deg-1 mole=!
cPF.(y) = 1·84+4·66 x 10-3 Teal deg-1 mole=! For02: Heat capacity Cp=7·16+1·00xI0-3T-0·40xl05T-2 cal
dcg-1 mole=!
and the heat capacity change for reaction 2.6.6 is therefore:
For N2: Heat capacity CP = 6·66+ 1·02x I0-3Tcal deg-1 mole=!
LlCP = 6·24-3·16x 10-3T-0·47x 105T-2•
Hence: For the reaction:
2.7.J
and, since J H?,183 = - 63,380 cal, the heat of reaction at 298°K is obtained by application of Hess's law to
therefore: the formation reactions for CO2 and CO.
Thus: ·
L1Hi = -68,580+6·24T- l·58 x I0-3 T2+0·47 x 10s r-»
L1H~8 = ,;[2 x (-94,050)-2 x (-26,400)-0]
which is valid from 1,183°K up to 1,651°K, the melting point of FeO. = ·-135,300 cal.
At the temperature of l,350°C (1,623°K):
Now, following 1.20:
Ll H?,623 = - 62,600 cal.
Thus, when y-iron is oxidized to form ferrous oxide at l ,350°C, 62,600 cal
'
I
L1H0 = f LlCP'dT
of heat are evolved per mole of iron oxidized. and for the above reaction:
.JC,= 2<:;'PCO.<a>:._2Cpeo(a)-Cp()e(a)
2.7 Blast-furnace gas containing 20 per cent CO, 10 per cent CO2 and = 0·36+1·36x 10-3T-3·46x IO'T-2.
70 per cent N2 by volume is burned with the correct amount of air in a Therefore:
•,
r'
furnace. If the gases enter the furnace at 600°C and leave at l,OS0°C,
calculate the maximum flame temperature and the heat supplied to the L1Hi = LlH0+,;0·36T+0·6~x 10'"'31'2+3·46x 105T-1
furnace in B.Th. U. per cubicfoot of exhaust gases. where LlH0, the integration constant ( = -136,600), is found by substi-
This is a very common type of problem of the beat-balance type. The tuting LlH'18 = -135,300 cal.
heat evolved in a reaction is used to heat up the products of the reaction Hence:
and the surroundings, the problem being to calculate one or other of the L1Hi = .•. 136,600,+0·36T+0'.68 x'I0-3 T2+3·46 x 105 T-1.
heat terms or increases in temperature. In this particular problem, .the
gas of known analysis burns to give up a calculable amount of heat which Solving this for a temperature of 600°C (873°K), the temperature of the
is used to heat the product (and any diluting) gases. Assuming adiabatic gas entering the furnace; the beat liberated by reaction).7.1 is:
conditions, the maximum temperature of these gases in the flame can be LIH~13 = -135,400 cal.
36 ENTHALPY AND HEAT CAPACITY HEATS OF REACTION AT ELEVATED TEMPERATURES 37

This heat of reaction heats the products, together with the diluting Thus, the maximum flame temperature is l,654°C.
nitrogen and carbon dioxide in the incoming gas, to the maximum flame The heat supplied to the furnace is the difference between the heat
temperature. Note that this heat will be a positive quantity with respect evolved in the reaction {135,400cal at 873°K) and the beat carried away
to the exhaust gas. in the exhaust gases. The latter is the quantity of beat required to raiso
In reaction 2.7.J, 2 mole of CO combine with I mole of oxygen from the exhaust gases to the exit temperature of 1,0S0°C {1,323°K) from the
the air to form 2 mole of CO2• Therefore, the exhaust gas will consist of: inlet temperature of 600°C. Thus, the heat carried away is:
2 mole of CO2 from reaction 2.7.1 } 1,323
, 3 mole CO2
I mole of CO2 in the blast furnace gas ;
f
873
L Cp.dT
. exh&IUI
= [104·91T+8·85x J0-3T2+6·12x 105T-1 J!;23
7 mole of N2 in the blast furnace gas } = 47,100+8,740-240
11 mole N2
4 mole of N2 in the air · = 55,600 cal,

and these are heated to the maximum flame temperature (TM) by the and the heat supplied to the furnace is:
heat supplied from the combustion of the 2 mole'of CO at 873°K. 135,400- 55,600 ••• 79,800 cal.
r«
Thus, available heat = 135,400 cal= J [3CPco. +II CPN,]dT. This is the heat corresponding to 3 mole of CO2 and 11 mole of N2
.( = 14 mole of exhaust gas). Assuming these gases behave ideally, the
j V3
volume of this gas at N.T.P. is 14 x 22·4 litre. At 1,050°C (1,323°K) and
For 3 mole of CO2: Cp = 31·65+6·48 x 10-3 T-6·12 x I05T-2
cal deg=! · a constant pressure of 1 atmosphere, the gas expands to a volume of:

For 11 mole ofN2: CP = 73·26+ 11·22x I0-3Tcal deg-1• 1,323 .

I 4 x 22· 4 x ----· = l 5 2 l litre
273 '
Hence:
[104·91T+8·85 x 10-.3 T2+6·12 x 105 T-1(" sn = 135,400 = 53·7 cu ft of exhaust gas.
.
and: Therefore, the heat supplied to the furnace per cu ft of exhaust gas is:
104·9ITM+8·85x I0-3~+6·12x 105T~i = 234,350
which can be solved for TM· However, to simplify the solution it is 79,800 = 1,485 cal '
I

possible to neglect the third term on the left-hand side of the equation. 53·7
Because TM must be greater than 1,320°K (the exit temperature of the =-= 5·89 B.Th.U.
gases), then:
6· 12 x 105 T;;.,1 < 500 cal, or < 0·25 % of the available heat. 2.8 Pure chromic oxide (Cr203) reacts with the stoichiometric amount of
pure aluminium, both initially at 25°C, to produce plumina and pure liquid
Thus: chromium. If the maximum temperature attained in· the crucible is 1,900°C,
8·85 x 10-3 Tk+ 104·91TM-234,350 ~ 0,
calculate the heal lost to the surroundings per kilogram of aluminium.
a quadratic in TM• which is solved by the formula method:"
This problem can be approached from two different viewpoints.
-104·91 + y[(l04•91)2+(4 X 8·85 X 234·35)] Possibly the simpler method is to assume that the reaction takes place at
TM ~ . 2 X 8·85 X 10-3 . -
25°C, part of the heat generated being used to heat the products of
139"0) reaction to l,900°C and the remainder being lost to the surroundings .
. = (--104~~~ x i03 = l,927~K or l,654°C. The reaction at room temperature is:

• Sec appendix 5. 2.8.1

4
 :t.
3S ENTHALPY AND HEAT CAPACITY
HEATS OF REACTION AT ELEV A TED TEMPERA TURES 39
The standard heats of formation of Cr203 and Al203 are - 270 and
-400 kcal mole=! respectively. Hence the heat of reaction at 298°K is: This heat will correspond to 2 mole of aluminium reacted. Thus, the heat
loss per kilogram of aluminium is:
.dH~8 = -400,000-(-270,000) \: '
32,500 . 1,000 = OOO . I
= - 130,000 cal (heat evolved). . , i ' X 26·97 603' ca
This heat is used partly to heat 2 mole of chromium and 1 mole of AI203 where 26·97 is the atomic weight of aluminium .
. to l,900°C (2, 173°K). In practice, of course, the overall process would not take place so
Heat required for Al203. discretely as this.' For once the reaction is initiated, the temperature of
2,173 the reacting mass would rise and at any time during the reaction, un-
,:jH<°298-2,173"K) "" J
298
CP.u.o.. dT consumed reactants and just-generated products would be at some
temperature less than :J,900°C. However, as enthalpy is' a.variable of
and: state of any thermodynamic system, changes in enthalpy as the system
CPAi.o.c.> = 27·38 +3·08 x 10-3 T-8·20 x 105 T-2 cal deg-1 moler". passes from one state to another will be independent of tho way in which
Therefore: this happens. Hence in this instance, the difference between the heat
generated by the reaction and the sensible heat requirements. of tho
.dH&,;a ..• 2,173"X) - 51,350+7,140-2,375 products, i.e. the heat evolved, will be a constant, independent of just
,.. 56,115 cal mole-1• how or at what temperature the reaction takes place. To illustrate this,
Heat required for chromium. consider the 'opposite extreme' from the process stipulated for tho
Chromium melts at l,850°C (2,123°K) and the heat required to raise calculation just made. Let heat be supplied to the reactants to raise their
the two mole of chromium from 298°K to 2,173°K will include the latent temperature to J ,900°C and let the reaction take place at that temperature
heat of fusion. Noting that: to give alumina and liquid chromium as the final product. Then, if tho
initial heat supplied is subtracted from the heat of reaction to give tho
CPcrc.> = 5·84+2·36 x 10-3T-0·88 x 105 T-2 cal deg-1 mole+!
net heat evolved, the latter should be equal to the value just calculated.
CPc,o, = 9·40 cal deg-1 mole-1 Consider reaction 2.8.J at I,900°C.
Lr= 4,600 cal mole=! at 2,123°K, The heat required to raise the temperature of 2 mole of aluminium to
1,900°c is obtaned from: I
then, following 1.22:
CPAJ<,> = 4·94+2·96 10-3 T cal deg-1 mole-1
1
2,123 x

.dH&,;a.:.2.m"K> = [11·68+4·72x 10-3T-1-76x 105T-2]dT CPAJO> ;,,.• 7·00

(
.
.
cal. deg-1 mo!e-1
2,173 L, at 659°C ~ 2·5 kcal mole"!
+ (2 .X 4,600) + J 18·80.dT and is calculated as sh~wn previously to obtain:
2,123
Ll H&s-+2,t7J"K) = 30,940 cal. •
= 21,300+ 10,430-508+9,200+940
The heat required to raise the temperature of I mole of chromium
= 41,360 cal.
oxide to J,900°C has been calculated in example 2.4a, method 2 (p. 29),
The heat lost to the surroundings (for each unit of reaction) is the as:
difference between the heat of reaction and the sensible heat of the .dH(°z<Js-+2,11J'K> = 57,500 cal.
·. ·,'.
products. Thus heat loss Thus the total heat required to raise the temperature of the reactants
= 130,000-56,100-41,400 to 1 900°C is:
,
•'. ;!

-= 32,500~. ! H
}/i ·. 0,. == 30,940 + 57;500
· = 88,440 cal.
-40 ENTHALPY AND HEAT CAPACITY THE USE OF THE CLAPEYRON EQUATIONS ,1

Because of the occurrence of phase transformationsin the temperature calculate the pressure variation of an equilibrium transformation
range 298--+2, 173°K (i.e. aluminium melts at 932°K'. chromium melts at temperature or, alternatively, the temperature variation with pressure
2, 123°K) calculation of the variation with temperature of the heat of may be used to deduce the heat of reaction. The following two examples
reaction 2.8.1 necessitates the use of 1.22. Following the procedure illustrate typical applications of these equations. Further applications to
elaborated in example 2.6: transformations in condensed systems are involved in the derivation of
932
thermodynamic data from constitutional diagrams considered in
chapter 7. ·
J Htm -LlH~ = f [2CPcr<>> + CPAJ,o.c,>-2CPAJ(I>- CP(;r,O,<>>]dT
298 . .
2.9 At I atmosphere pressure, sodium melts at 97·8°C with a latent heat
2,123
of fusion of 630 cal mole-" and an increase in specific volume of 0·0279
-(2 x 2,500) + f l2CPctu,+ cPAJ.o.CJ> - 2cPAJro - cPo.o,<o,JdT cc g-1• Calculate the melting point of sodium at 10 atmosphere pressure.
9~2
2,173 The effect of pressure on a phase change in a one-component system is
+ (2 x 4,600) + J
2,123
[2CPo<l> + CPA,ltO,(Jl - 2CPAJO) - CPO.O.(I)] dT.
given by the Clapeyron equation. One derivation of the relation (applied
to the fusion reaction considered in the problem) is as follows.
At equilibrium:
Using the CP values given previously and:
G(I) = G(I)
CPer.o,<,> = 28·53+2·20 x I0-3 T-3·74 x 105 r-» cal deg-I mole-I, where G is the Gibbs free energy of the phase. Hence, if the pressure and
the heat of reaction at 2, 173°K is calculated as: temperature of the system are changed infinitesimally, such that the
system stays at equilibrium, then:
LlHt173 = -121,000 cal. dGcl) = dGco
The difference between this and the sensible heat requirements of the or, from 1.33:
reactants will be the heat lost to the surroundings, i.e,
i Hence:
heat loss= 121,000-88,440 ·I dT _ Vin- v<•> _ L1 V
= 32,560 cal per unit of reaction. dP - Scl)-Sc,> - L1S
Thus the heat loss per kilogram of aluminium where L1 V and L1 Sare the volume and entropy 4ifferences, respectively,
between the two phases at equilibrium (in this case liquid and solid).
3\08·x However, as at equilibrium, .1S = L1H/T, from 1.27, then:
= 32,560
dT TL1 V
= 603,000 cal 2.9.l
dP = AH
' !
which is equal to the value calculated previously, when the reaction was where tiH is the enthalpy change at the equilibrium transformation
assumed to take place at 25°C. temperature T (in this case, the heat of fusion (Lr) at the I atmosphere
melting point, 97·8°C or 371 °K).
The use of the Clapeyron equations Particular care should be taken in selecting the units of the quantities
The preceding examples have involved the Kirchhoff equation, giving the substituted in equation 2.9.J. In the present problem, for consistency,
temperature dependence of the heat of a reaction at' constant pressure. the !atent heat of fusion must be expressed in cal g-1• As the atomic
However, if pressure is introduced as a second independent variable, in weight of sodium is 23, then:
addition to temperature, useful relations may be deduced expressing the
interdependence of these three quantities. In particular, the Clapeyron, LlH = -630
23
= 27·4 cal g-1•
or the approximate Clausius-Clapeyron, equation may be used to
42 ENTHALPY AND HEAT CAPACITY

Substituting in 2.9.1: and:

dT 371 x 0·0279
dP = 27·4
This is one form of theCla~sh:is':-Cla~yron' equation.
== 0·378 deg cc ca1-1. For this problem: .' '· . . .. . '
The units deg cc cal-1, deduced from consideration of the substituted . · .. '., · •. '5·30 . : .'
quantities, must now be converted into more conventional and con- + 2,4QOT-2- 2· 303 r-i, deg='. • ,
venient ones, for example deg atm-1• To do this, it is necessary to
multiply by a quantity in units of ca,lcc-1 atm-1 (i.e. deg cc car-' x cal cc-1 Hence:
.'f','

atm-1 = deg atm:"). . '1Hf = 4·575T2.(2,400T-2-2·3T-1)

As I cal= 41·293 cc atrn, therefore:
the units of which are cal mole-1, as R = 1 ·987 cal deg-1 mole-1 or RT2
dT 1 is in units of cal deg moler". Thus at 77°C (350°K):
dP = 0'378 x 41·293
AH= 7,300 cal mole-1•
- 0·00915 deg atm-1
This value for the latent heat of vaporization of carbon tetrachloride
which, if assumed to be independent of pressure, yields an increase of
at the boiling point compares well with the value of 7,150 cal mole-1
0·09 degrees in the melting point of sodium when the pressure is increased
quoted in the literature.
to 10 atmospheres. This value is comparable with the limits of accuracy
of the determination of the melting point and is fairly typical of the small
effect of pressure on the melting points of pure metals. • The units of d ::.! u:e deg-1, for it sho.uld~ recalled that logP is dimensionless,

2.10 The vapour pressure of liquid carbon tetrachloride is given as a or alternatively that:
function of temperature by: d logP 1 dP pressure
-<iT- = 2·303P.dT ""pressure x temperature
Iogp ••• -2,400T-1-5·30logT+23·60
where p is in mm Hg. /.-----
1 temperature
Calculate the latent heat of vaporization of CC14 at its boiling point, ·''
Note also that Pis given 'in 'mm Hg. However, the units are immaterial when (logP) ia
11°c. differentiated for substitution in the Clausius-Clapeyron equation. For example:
The Clapeyron equation was derived in the previous example for any logP = -2,400T-1-5·3logT+23·6
phase change in a single component system. For the particular case of the or:
vaporization of a liquid, thevolume of the latter is small relative to that P =- ant[log(-2,400T-1-:5·3logT+23-6JmmHg.
I
of the vapour and the volume change on vaporization will approximate Converting to atmosphc1a:
I.
closely to the' volume of the vapour. If it is further assumed that this }antilog I - 2;400T-1-5·3 log T + 23·6]
vapour behaves as an ideal gas (defined by PV= RT) then, from p =,·------------
.,\' 7(1J
equation 2'.9.1: or:
logP i - 2,4007'-I- 5-3,log T + 23 ·6- log 7(1J
and: I•

or: ,·
dlogP
d7'- ·r ·\'·2'•""
I'' .

1' ~
·'."""Tf:. - 2·303
·, ' .
~·3
'·
r-Id' cg-
~- '
l I fi
as be ore.

---
dlnP
dT
AH
RT2
That is, conversion of thl (logP) expression to any system of units changes only the
temperature independent~t~rm.wh.icl\ ~i5ap~fl on differentiation.

·.·.,
 44 ENTHALPY AND HEAT CAPACITY.
Exercises
2.A Twenty gram of ice at a temperature of -2°C is dropped into a
calorimeter containing 100 gram of water at 25°C The water equivalent
of the calorimeter is 18 gram and the final te~pefature of the bath,
corrected for heat losses, is 9·7°C. Find the specific ~cat of ice.
2.B Steam at I00°C is blown into .a swimming bath of dimensions
80 n >< 30 ft x 5 ft to raise the temperature of the water from l 5°C to
20"C. Find the weight of steam required. (Latent ~eat of steam - 960
B.Th.U. Jb-1, I B.Tll.U. = heat required to raise 1 lb of water l°F,
! cu ft of water weighs 62·5 lb.) ' ~ii
2.c Assuming no heat is lost to the surroundings, calculate the amount
of heat required to just melt 1,000 gram of lead initially at l 5°C.
2.o Fifty gram of steam at l00°C are passed into an adiabatic calori-
meter containing 200 lram of ice and 100 gram of water at 0°C. Find the
final temperature of the water. (Use data given in 2.B.)
2.r: A bath of pure molten aluminium is cooled towards its melting
point of 659°C at a constant rate of 5·0°C min-1• Assuming that the rate
of heat loss from the bath remains constant and that equilibrium is
maintained throughout the solidification process, calculate the time for
isothermal solidification of the bath and the cooling. rate immediately
after solidification.
/
2.P A skater weighing 120 lb glides in a straight line on one skate,
decelerating uniformly to rest in 20 yd from a speed of 10 m.p.h. on ice
at - 5°C. If the width of her skate is 5 mm, what is the maximum depth
of groove in the lee that she could leave behind her? The specific heat and
density of ice are 0·5 cal deg-1 g:-1 and 0·917 g cc-1 respectively.
2.C An aircraft is fitted with an auxiliary' ramjet engine of 2,000 cm2
cross-section which is ignited at an air speed of 400 m.p.h. atan altitude
of 6,500 ft. At this altitude the air density is f,.000g,.m-3 and .its pressure
0·8~ atm and temperature 0°C. Purified kerosine fuel is burnt in the
ramjet, being supplied at 3· 11 .Iitre sec-1 which isjust sufficient to
combine completely with the oxygen of the air.The kerosine is of analysis
86·4 per cent C and 13 ·6 per cent H2 by weight, of density O· 79 g cc- 1,
and has a heat of combustion of l 1·14 kcal g-1• ffthe weight of the
aircraft is 2,000 kg and 5 per cent of the theoretical th;ust is available to
·~ ' .
.i
'
EXERCISES ~
drive it forward, calculate the instantaneous acceleration of the aircraft
in level flight when the ramjet is fired.: It should be assumed that the
exhaust gases behave ideally and that the error arising from the use of
CP rather than C., data is negligible.
2.H The mean value of the coefficient of cubical expansion (a:) of liquid
mercury in the temperature range 0-100°C is 18·1 x lQ-5 deg-1 and the
Coefficient of isothermal Compressibility (/3) is 45 X 10-6 atnr-'. If the
density of mercury is 13·35 g cc=", calculate the difference in its molar
heat capacities at constant pressure and constant volume.
2.1 The latent heat of vaporization of zinc is 27·3 kcal mole-! at the
boiling point, 907°C. Find the vapour pressure over pure zinc; at 850°C.
2.1 Calculate the approximate pressure required to distil mercury at
I00°C if the latent heat of vaporization is 14· 14 kcal mole-! at the boiling
point 357°C.
2.K A drop calorimeter employs boiling liquid nitrogen as the calori-
metric fluid, values of the integrated heat capacity of a metal sample
dropped into the calorimeter from various temperatures being calculated
from the extra volume of nitrogen gas boiled off. If this boil-off is to be
metered by a simple U-tube type of ma no metric flow-meter, calculate the
approximate correction (per mm Hg) which must be applied to the
integrated heat capacity calculation to take into accountthe necessary
pressure drop across this flowmeter. The latent heAt of vaporization of
nitrogen is I· 33 kcal mole-1 at the boiling point of 77·6°K (760 mm
pressure) and the specific volume difference between liquid and gaseous
nitrogen is 6·39 Iitre mole-1 at that temperature.
2.L Given the following data, calculate the heat of reaction when solid
magnesium is dissolved in a solution consisting of MCI (100 H20)
M~ >+2HClc,> = MgCl2e,>+H2e,>; AH= -109·4 kcal
1
HCic,>+ l00H20 = HCl(IOOH20); AH= -17·4 kcal
MgCl2<s>+200H20 = MgC!i(200H20); AH= -35·5 kcal.
2.M Calculate the heat of formation of Fe203 at 298°K, given that the
heat of reaction of Fe203 with carbon monoxide to form solid iron and
carbon dioxide is -6·65 kcal mole-1 of Fe203 •
 .CS ENTHALPY AND HEAT CAPACITY
EXEkCISES 47
'2.N Given the following data, calculate the standard heat of formation \, ' .
2.T The most commonly used Thei mit or exothermic material for
of solid aluminium carbide (Al4C3), and the limits of error inthis value.
fusion welding is afinely divided mixture of aluminium and iron oxide
Heat of formation of: COea> = -26·40 ± 0·03 kcal molc-1 (Fe304) in the ratio i :3·2 respectively which, after initiation, reacts to
CO2,> ,.;; -94·05 ± 0·01 kcal moJe-1 give 66,300 B.Th.U. fpr every lb.ofaluminium used. In preparing a melt
or pouring on to a buited steel joint, it is found that a 60-lb charge of the
Al203:a> = -400·0± 1·5 kcal mole-". mixture in the above ·'proportions would reach a maximum temperature
The heat of reaction between I mole of Al4C3 and 9 mole of CO2 to of 2,600°C in the reaetion crucible. A.s this temperature is too high for
produce alumina and carbon monoxide at 25°C is -223·7 ± 11 ·7 kcal. pouring, pieces of s_teel scrap are mixed with the charge at room tern-
perature (25°C), befo1et,he i~1tia'tionof the reaction, to give a maximum
2.o Calculate the heat of formation of ferrous oxide (FeO) from pure temperature of 2,400°,C.Assuming that heat losses from the charge and
iron and oxygen gas atI ,390°C, making use of the calculation given in the heat capacity of the.cruoible are independent of temperature in the
example 2.6. temperature range 2.ooo,,'..~.t).()0°C, .calculate the approximate weight of
steel scrap to be. added to th~;;charg~:,The heat capacity and heats of
2.P Calculate the heat of formation of calcium oxide (CaO) at 1,000°c transforrnation.of the .steel may be tak6n as equal to those for pure iron.
from the appropriate enthalpy and heat capacity data of appendix 3.

2.Q If a mass of ferrous oxide is heated to I ,000°C and then exposed to

a gas mixture of carbon monoxide and hydrogen in suitable proportions,
the oxide can be reduced to the metal without any further heat supply
(i.e. the process is autogenous or thermally self-supporting). Find the
, optimum ratio PcolPH, (where Pco +PH,= I) for the inlet gas in order to
\ achieve this condition, assuming a heat loss from the charge to the
surroundings of 5,000 cal per lb of oxide reduced.

2.R Regenerated producer gas of composition 5 per cent CO2, 25 per

cent CO, 2 per cent H20, 14 per cent H2, and 54 per cent N2 by volume
is fed into a furnace at 700°C. It is desired to heat up the furnace as
rapidly as possible, but the maximum flame temperature must not. .. ~i .
exceed I ,600"C or the furnace refractories would soften. What is the
minimum volume of air to be fed into the furnace with each 1,000 cu ft
of producer gas to achieve this? (For the purposes of this calculation,
the heat capacities of the gases should be assumed to be linear functions
r-
of temperature: i.e. the terms in 2 should be ignored.)

2.s During the grinding of steels, sparks are generated as a result of ...
particles of molten iron, ejected at high speed, combining with the t· '!,:.

oxygen of the air to form molten FeO. Assuming that the grinding action
raises the temperature of the iron particles to l,600°C, and using the
relations obtained in example 2.6, what is the maximum temperature
attainable in the spark i~ air at 25°C7 What assumptions are made
concerning the data used in this calculation 7

i!
 -,--
i
CHAPTER 3
i
ENTROPY AND FREE ENERCY .49
/
I .dG0 can be obtained from knowledge of AH0 and As0• Expressing

Entropy and Free Energy I JH0 for the reaction in similar form to that in problem 2.1:
I AH0 = (.dH~0+3AH,&.)-(A~H,+AHft.o)

I ""(-26·40)+(0)-(-17·89)-(-57·80)
= +49·29 kcal mole-I or 49,290 cal mole-I.
Calculations of the entropies or free energies of substances or reactions !· Similarly:
from heat capacity data follow similar lines to the calculations of As0 = (S~o+3S~.)-(SgH,
+s~.o)
enthalpy detailed in the last chapter. For example, equations 1.26 for
Entropy and 1.37 for Gibbs Free Energy correspond roughly to equation = (47·3) +(3 X 31 ·2)-(44•5)-(45• 1)
I .20 used previously. The solution of these equations may be obtained
••• 51 · 3 cal deg-1 mole-I.
in a manner similar to the two methods used proViously to solve the Hence:
enthalpy equations, either treating the integrals as indefinite and solving
.dG~a ::a +49,290-51·3T
for the integration constants, or as definite integrals between temperature
limits. As before, both methods require knowledge of the appropriate +49,290-15,280
values at one of the limits. :
+34,000 cal moJe-1•
As the third law of-thermodynamics proposes that the entropy of all
substances in complete internal equilibrium is zero at 0°K, the' standard' The value of the standard free-energy change for the reaction at298°K
entropies of elements at 298°K are non-zero. In calculations involving is positive. As any' going' process is always accompanied by a decrease in
standard entropies of reactions, these non-zero values for the elements free energy (equation 1.38), this value suggests that methane will not be
must be substituted together with those for compounds. oxidized by water vapour at 298°K, but rather that carbon monoxide
,'. . :
and hydrogen will react to give methane and watervapour,i.e. the reaction
3.1 Determine the standard free-energy change for the oxidation of as written will go from right-to-left.* However, it should be noted that as
methane by water. vapour to form carbon monoxide at 25°~. both JH0 and .ti.s0 for the reaction are positive at 298°K then, if these
• I•
.
•
: ~ ' .~ .
••
·,-
w
. values do not change much with temperature, A G0 ~ = AH0-T.dS0) will
This is an illustration of the extension of' the principle outlined in
become negative at sufficiently high temperatures/For example, on the
example 2.1 (Hess's law) to the calculation of the free-energy change. basis of JH0 and .ds0 being temperature independent, AG0 becomes
The reaction may be written as: · ' zero at 961 °K, and at higher temperatures the reaction will proceed
CH4<s)+H20w ~ COw+3H2~; spontaneously from left to right. In fact the values of both .dH0 and
LJS0 are temperature dependent, and the accurate evaluation of the
for which the following data: are required:' ·t minimum temperature of this reaction involves t}1e determination of
AHg9s. {S~9s i LJG0 as a function of temperature. Such determinations are involved in
kcal mole=! caldeg " mole=! some of the following examples, and evaluations of minimum reaction
. )44·5 temperatures are illustrated in chapter 4•
CH4<B> -17·~9
CO(g> -26·40 ·147'.3 • It will be recalled that .JGO is the free-energy change when all the components are
H21.sJ 0 :31·2 in their standard states (equation 1.56). Thus, the sign of .JG0 is not' an infallible
H20ee> -57·80 '45·1 indication of the direction of a reaction in practice, for one or more of the components

From the defining relation 1.31:

or the reaction may not be in its gtandard stat~. In an a/1.ga.g rtaction such as that
presently considered, any deductions from calculated values of .JGO must necessarily
A c;o = AH0-T.ti.s0 involve the proviso that the gases are present in the system in their standard states
i.e, at one atmosphere pressure. '

·'
 !10 ENTROPY AN:P FREE ENERGY
ENTROPY AND FREE ENERGY 51
(3.2) The freezing point for pure copper is generally accepted as l,083°C, (J]) Using the data given in the previous problem, determine the Gibbs
\huf small droplets of liquid copper have been supercooled to 847°C before free-energy change for the reaction C1JoJ = Cvc,> at the temperature of
spontaneous solidification occurs. Calculate the entropy change for the spontaneous solidification, 847°C.
isothermal solidification of copper at the latter temperature.
This prob: ... m is most readily solved by application of the indefinite
At the true melting point {l,083°C) the entropy difference between the integral form of the Gibbs-Helmholtz relation, equation 1.35, applied
solid and liquid forms of copper is equal to the entropy of fusion (Sc). to the reaction in question. That is:
That is, for the reaction:

ClJm ~ ~•>
·

3.2.1
~_q-~·=
T ,
-J tJHO.dT
T2
or:
Lls8:....> - -Srat 1,356°K.
The latent heat of fusion of copper is 3,100 cal mote-1, and therefore,
i 0
LIGr;=_-T J [.dH~
T -+T, 1 J ACpdT
2 2 · J dT.
J.
from 1.27:
Substituting the .d CAI-elationfor reaction 3.2.1:
:~ .' .
< -:ff-+·;j,"°1 2 J (-2·09+1·5x
f'.[Aifo
LIS? 356 = - (3,lOO) •• -2·28 cal deg-1 mole-1
· 1,356 . ,1
LJGr
o
= -T 10- 3 T)dT ] dT
The heat capacities of solid and liquid copper are: ... '· fft·'·<:: . ·-..... ·,. '
J
CPeuo> = 7·50 cal deg-1 moie-1 = -T
Ji.[.dH 2·09 ·
t_--;;;-/::--f"+0·75xI0-3 dT

CPeueo> = 5·41 + l ·50 x I0-3 T cal degr! mole-1• = .4H0,+\i-09ffr{;r"""'.0'75 x 10~3 T2 + fT 3.3.J
Hence: where LI HO and I are i~tegratfo; cJnstants, the former arising from the
ACP<>.2.1> = -2·09+1·50 x 10-3 Teal deg-1 mole=! integration of the Kirchhoff equation and the latter from the integration
of the expression for LIH0.Theseare determined, as detailed in chapter 1,
and the change in the entropy of the reaction between l,083°C (1,356°K) from a knowledge either of-Lld0 at two tempcra/ures or ofAG0 and
and 847°C (l , 120°K) is given, following equation 1.26, by: AH0 at one temperature. The latter information s presently available
for reaction 3.2.1,. for at the equilibrium melting point (I ,083°C or
t,120
A O A
.uS1.120-.uS1.3S6
O
= J L!Cp d
"i": T
1,356°K):
.dH0 = -Lr= -3,100 cal
1,356
and: L!Go = 0 (from 1.38).
= [-2·09lnT+l·5x I0-3T)1'120
1,356
Hence:
= -2·09ln
1 120
1·5 X 10-3(1,120-1,356) .dH?,356 = -3,100 = .dH0+ J (-2·09+1·50x I0-3T)dT
1:356+

= +0·046 cal deg-1 mole='.

= AH0-2'.09 x 1,356+0·75 x IQ-3(1,356)2

Therefore: and: iJHo = -1,650.

L!S?,120,.. -2·28+0·046 Substituting this value in 3.3.1:

- -2·23 cal deg-1 mole='. ao; == -l,650+2·09TlnT-0·75x I0-3T2+1Tcal mole-'.

~(
 !12 ENTROPY AND FREE ENERGY
· ENTROPY AND FREE ENERGY !3
But .JG?,3v. = 0. Therefore, substituting T = 1,356°K and equating to
zero yields: Substituting this in equation 1.26, the entropy change over the tempera-
I= -12·85. ture interval 298 to 1,033°K is given by:
Hence:

/ ..~_Gi = -1,650+2·09Tln T-0·75 x I0-3T2-12·85Tcal mole=!

.:.---,
LJ
A o =
Aso1,oH-""Sm J
1,033
Acp
T.dT
298
and the required free-energy change at 847°C(I,120°K) is: 1,033

.JG?.120 = -540cal mole-1• 1

J
298
(3·90T-1-4·42x I0-3-0·47x 10sr-3JdT

This calculation of the free-energy difference between two coexisting = 4·85-3·25-0·25.

phases of a pure component from data for the latent heat of the trans- Therefore:
formation is used again in chapter 7, where it is applied to the derivation AS?_o33+16·95= 1·35
of thermodynamic data from constitutional phase diagrams (cf. or:
examples 7.5 and 7.7). · · ~
A s?.033 = -15·60 cal deg-1•
Alternatively; .JH?,rzo-x. can be determined in the manner described in
At 1,033°K the ferro- to para-magnetic change can be regarded as
example 2.4 and .JG? 120"K is then obtained by, inserting this value and
being accompanied by a latent heat of transformation of+ 660cal mole-1
that for .JS? 120'K, det~rmined in the pr~viousexample, in equation 1.31.
(cf. example 2.6, p. 32). Thus, for the reaction:
The procedure is demonstrated in the f91lo\\'.ing.e,,::ample.
Fe<«, rerro-m-,net1c> ~ Fe<«, P&ta-maanetlc> 3.4.2
3.4) Evaluate the entropy and Gibbsfree-energ/cha~ge at I ,350°C for at 1,033°K the entropy change is:
-, __The formation of ferrous oxide from pure iron and_.oxjgengas at I atmo-
sphere pressure (cf. example 2.6). · · 660
.dso - .., 0·64 cal deg-1 moie-1•
1,033
The reaction at 298°K is:
Applying Hess's law to the entropy changes for reactions 3.4.J and 3.4.2,
for the reaction:
Fe(cx)+-rOiw ~ FeO<•>· 3.4.I I
The standard entropies of the components are: Fee«. para-maaneuc> + toiw· ~ Fed<•> 3.4.J
the entropy change at 1,033°K is:
S~, cal deg-I mole-I
6·49 AS?,on - Js&.,.n-AS<~-'.2>
14·05 = -15·60-0·64
. 49·02 ••• -16·24 cal deg-1• • •
. . ~..
Hence, at 298°K, the entropy change for reaction 3.4.J is: Reaction 3.4.3 applies up to 910°C (1,183°K), at which temperature
:,".
the reaction:
ASO
298 =- SOFe0(1) ,-. soFe(«) . ".""."I
'.1sO ..
Oo(a)
LJ Fee«. para.masnettc> ~ Fe<y> 3.4.4
= -16·95 cal deg-1• occurs, with a latent heat of transformation of + 220 cal mole-1, or an
entropy of transformation of 220/1,183-0·19 cal deg-1 mo!e-1• The
At temperatures up to 760°C (1,033°K), the Curie temperature for entropy change· for reaction 3.4.J at this temperature is obtained by
iron, the value of LlCP for reaction 3.4.1, given in example 2.6, is: substituting JC, for this reaction, from example 2.6:
..::1cp = 3·90-4·42 x 10.:.3r:...o.4i x 1osr-2. ,
JC - -0·92+1•50x I0-3T-0·47x 105T-2
5
~" F.NTROPY A'.'.D FREE ENERGY
EN rROPY .AND FREE ENERGY !SS
11110equation 1.26 and determining the entropy increment in the tern-
The Gibbs free-energy change for the formation of ferrous oxide at
!>crature range 1,033 to 1,183°K, i.e.:
l,350°C is calculated from the value of the entropy of formation and the
LlS?,i83-.:::lS?.o33 = -0·92ln !:~!! + 1-5 x I0-3(1,183-1,033).
entha!py of formation at the same temperature. That is:

LJG'?,623 = LJH?,623-l,623LIS?,62l·
+0·24x 1osx(c1.1~3)2-(1,0133)i). Substituting the value of LJH0 (from example 2.6) and the value of LIS0
just calculated:
· fherefore:
LlS?,1s3+ 16·24 = 0·10 Ll<;?,623 = -62,600-(1,623 x -15·76)
LlS?,1sJ= -16·14caldeg-1• - 37,000 cal mole-1 of FeO.
Applying Hess's law to reactions 3.4J and 3.4.4 at 1,183°K, for the Alternatively, the Gibbs free-energy change can be evaluated by the
reaction: use of the Gibbs-Helmholtz relation 1.34- as in example 3.3. That is, the
3.4.5 variation of enthalpy with temperature, calculated in example 2.6, may
be substituted in the integral form 1.35,. the resulting integration con-
the entropy change is: stants being found by substitution of the known enthalpy and free-energy
L!S?,1u = LlS(~.,.1)-Js&.M) data at one temperature in.each temperature range. The solution by this
method is illustrated again in .the next example.
' ,t ' '
-16·14-0·19
-16·33 cal deg-1• 3.5 At 882°C, titanium transformsfrom.the low-temperature hexagonal,
close-packed (a)form to.a body-centred cubic (j3) structure. Calculate the
For this reaction (from example 2.6): hypothetical melting point of the. «form .of titanium.
LICP = 6·24-3·16 x 10-3 T-0·47 x 105 r-2 Thi·s type of calculatiot can be made for any substance which has more.
than one allotropic forrrrand for which adequate data exist. The accuracy
which applies up to l,378°C(l,651 °K), the melting point of FeO. Hence
of the answer depends very largely on the magnitude of the temperature
the entropy change at 1,623°K, the temperature for which the answer is
extrapolation which is made in the calculation (i.e. the/difference between
required, is:
the temperatures of theiallotropic transformation and the hypothetical
LI O 1,623 melting point). ; · ,. .
O
LI s,,623 = s,,1u+6·24ln 1183-3·1 6 x 10- 3 (I,623-1,183) The principle used is s}¢ple:' Because the free energy is a function only
' of the state of a system, (ts variation inany process is independent of the
· ( 1 I ) path of the process, Thus the free-energy change accompanying the
+0·24 x IO' x (I,623)2- (1,183)2
change from a to liquid titanium can. be written as: ,
-16·33+0·57 JG?~•>-~ JG?~P>~·L!G?ti-o
= -15·76 cal deg-1• which is equal to zero at the hypot'he:ical equilibrium melting point of
Alternatively, the integrated heat capacity equations could be solved . a titanium. The problemis to find the temperature at which: : ·
as indefmite integrals, the integration constants being found by substitu- LIG'?lP> = -:.1~-I) '= LIG'?,-P>· . 3.5.J
tion· of the known values of entropy at known temperatures, as (for
enthalpy) in example 2.6. The two procedures, as emphasized previously,
by expressing both LJ G0 .~erms as functions of temperature.
The reaction: ·
are equivalent and yield the same answer. 3.5.2
 ~6 F.NTROPY AND FREE ENERGY
ENTROPY AND FREE ENERGY st
occurs normally at 882°C (1,155°K) with a latent heat of
'
transformation and:
of 830 cal mole:". The heat capacities.ofthetwo allotropes arc:
-·· ~
C;(Tl. cz> = 5·28 + 2·4 x 10-3 T cal deg-1 mole-!
LI G~ = - T f [LI; o + ;2 I J
1 ·09 dT dT

Cp(n.i, = 6:91 cal deg=! mole+'. = .1H0-1·09TlnT+IT.

Therefore, for the reaction 3.5.2: Solving for the integration constants in a similar manner to before:

.dCP = 1·63-2·4x 10-3Tcal deg-t moJ~-1. .1//0 = 4,500-(1·09 x 1,933) = 2,400

Substituting this in 1.36, the free-energy change for $.5j is: I = -2,400+ 1·09 x 1,933 x 7.57,,.. 7·01
1,933 .
AGr- -rJ(A;•+M cHJ-2-4,do-,~dT]dT
1·.ol

J L +7
-T
rL1Ho
T2 t1·63
-1·2x
.] · ·· ;;
10-3 dT 1·2
1·0
- .dHo,-1 ·63Tln T+ 1·2 x '10-3 T2+1T l 0-8
0'6l
where LI HO and r are I integration constants which ..k,e obtained as r_...tur.
0·.ol ol interMCt~
follows. · 11so·K
As LJ H?.153 for reaction 3.5.2 is 830 cal mole-1 and
-- ...• _
AGa.--.p - ... -- -
therefore:
830 •.• LIH0+(1 ·63 x 1,155)-' l ·2 x (1,155)2 x 10-3
or: 1·2
L;JJ{0 = 548
and:
1·.ol I
1·
.dGJ = 548-1·63TlnT+l·2x I0-3T2+IT.
But .1 G0 = 0 at the equilibrium transformation temperature of l, l 55°K. 1·
Therefore: 1Kx103
I ••• 9·65 Fig. 3.Sa. Determination of hypothetical melting Point of ex-titanium.
and:
' Therefore:
. LlG'?cr-,8) = 548- I·63Tln T+ 1·2 x I0-3T2+9·65Tcal mole-1•
For the reaction: Llqe ..• n =- 2,400-1·09TlnT+7·01Tcalmole-1
or:
Ti<P> .= Tiro 3.5.3 JGg=.P> = -2,_400+ 1·09TlnT-7·01Tc:al ~oi~-1._
the latent heat of fusion is 4,500 cal mole-1 at l,660°C (1,933°K), and
Equation 3.5.1 is applicable at the hypothetical melting point of
the heat capacity of liquid titanium, Cp (Ti. 0 = 8·00 ~I deg-1 mole:".
«-titanium. The temperature at which the two free-energy values,
The change in the heat capacity for r;eacti_on3.5.3is:
Lle?«-P> and ..1Gg ..• P> are equal is best found graphically. Plots of the two
L1 C == 1 ·09 cal deg-1 inoJe-1 free-energy functions arc shown in Fig. 3.5a, from which the temperature
? .. '
 (' .·.',, ·'
EXERCISES·
58 ENTRo.-_y AND FREE ENERGY
,.'· cw . with the estimate: ma4l ~;': Ka:ufril~ri [l] of 1,730 ( ± 30)°K, f9r the
of intersection of the two curves is-rea<;i-,offas1,863°K or 1,590 (± 7)°C,
the·};y-poth-et/calmelting point of ex-titanium. However, although the hypothetical melting pQint, .based on free~~ri~rgycalculations using the
thermodynamic treatment is perfectly valid, the hypothetical melting Debye'theory of specifi~ beats~ · • · · ' · ' ·
• ~ . ' :·:· !
point may be grossly in error, for it is apparent from the graph that the
valves for JG?"' ...• ,8) for T > 1,400°K are very much in doubt. This error Exerc;i~
arises from the extrapolation of the C, data for the e-phase to higher 3.A Using the· appropriaie•data ;giveil in, appendix 3, calculate the
temperatures than those to which the empirical relation is applicable, enthalpy, entropy and free~energy i::\i'ariges at 1,000°K for ill~water-gas
and also probably from the assumed linear temperature dependence of reaction: · ,· · · ·
the CP relations for T~ and Tio>·
If the enthalpies and entropies of transformation and fusion are
3.s Assuming that vaiues of the standllfd entropies offonxuition of the
assumed to be independent of temperature (which corresponds to
a svurni ng constant and equal heat capacities for all three phases), the
components are not available, recalculate the free-energy change in the
strail<',iit, broken lines of Fig. 3.5a are obtained. The equations ofthese previous example, giveri that the free-energy change at S00°K is 15·2 kcal.
1in-:, arc obtained as follows:
3.c Derive relations for the variation of the free-energy change for the
L161,. ·fi> = L1Hi"' ...• fJ>-TAS&-.fJ> oxidation of aluminium with temperature applicable over the range
900-1,000°K which includes the melting point of the metal.
= Oat 1,155°K. the transformation temperature (T,).
Therefore, at any other temperature, assuming ACP = 0: 3.o Given that the free energy of formation of ammonia from nitrogen
and hydrogen is 8·35 and 13·95 kcal rri.ole-"-1 at 500 and 700°C, respee-
tively, calculate the standard heat of formation of ammonia at 298°K.

Tx 830 3.E If the latent heat of vaporization ofzinc is 27·3 kcal rnole=! at the
= 830---·-
l,155 boiling point 907°C, and using the data given in appendix 3, draw a
graph of the variation; of the entropy of'zinc between 25 and 920°C.
= 830-0·719T. • ,' I •

3.F On slow ~ool.ing,.: iron containing 0·01 .weight-per cent carbon

Ar-0 4,5()() transforms at aro\.md'9,08°Cfrom the high-temperature F.C.C. austenite
LJv(, ...• f!J = -4, 500 +1933 T phase to the B.C.C. ferrite phase. However, as is well known, rapid
' quenching of the austenite suppressbs.the transformation until at some
-4,500+2·33T. lower temperature th,e austenite transforms by a shear mechanism
T!1c intersection of these lines gives a value for the hypothetical to the phase know~ as
.martensite. The thermodynamics of this
mcl:mg po int of <X-titaniumof 1,750°K or l,477°C. This value is probably process are of obvi¢.us importance, one quantity which must be
mvrc accurate, in this instance, than that calculated using the CP data. considered being the tt.ee~energy·'changefor the suppressed austenite-«
·111i, <uuation is unusual, however, and arises from the large extra- ferrite transformatio·~~ Calculate this free-energy change as a function
\'•'':,ti,)!1 (if the latter. Generally, the assumption of non-temperature- of temperature below'.}neCurie temperature.
,!q,cn,',.·nt enthalpies and entropies of reaction leads to a less accurate '·
\·a'.,~e than that obtained by taking into account their variations with 3.G Pure solid manganese has four allotropes, designated «, {3, y and 8,
temperature, but in some circumstances, as in the present case, the stable in the tempera{tire ranges 'as shown i·, appendix 3, Heat Capacity
arpr(1\imaiion can be justified. It is interesting to note that the value of data. Tt has been fou~d ,tpa~ they-phase cannot be completely retained
1,750 'K obtained frorn the linear G/Trelations is in excellent agreement by quenching to rooni\~rnperature, but transforms, at least partially, to
' .. '. '· .. ·: d ·.,'. '· .. '., '.'·' . .
 ;· I

.,
'{
;:
·,r.,
 •;

ir:

CHAPTER 4·
60 ENTROPY AND FREE ENERGY . ij.
\ '

the low-temperature e-form. Calcu.).ate the theoretical transformation •..

temperature y'....i.cx, assuming that:
(a) the enthalpies and entropies oftransfomiation are independent of The Equilibri4111 Cohstallt
' ,. '.':·· ·,

temperature, ·'.,
(b) the CP data for the ex- and y-pbases can be extrapolated to higher
and lower temperatures respectively. (Use 5-figure logs.) ... :{Y ,\:. ;·\.·\"- , . . .
Possibly· the commonest 'application of;_ the principles of chemical
3.H The following data are smoothed values of the temperature varia- thermodynamics is, tct·be .foundin consideration of systems in which all
tion of the heat capacity for an ordered alloy of composition Cu3Au on the components or phas,es,ar~- in complete thermodynamic equilibrium.
heating from 100°C to above the critical disordering temperature, 390°C. T)1e equilibrium contposition,of.tiJ.e· phases of a heterogeneous system
Calculate the approximate 'entropy of ordering' of this alloy as given by under a given set bf prescribed 'conditions, or the.e'ffects of changes in
these data: the applied conditions on: the -equilibriurn. · compositions., is' readily
I 100 150 200 250 275 300 325 350 calculable from a knowledge of certain. simple theirliodyhamic .data of
T°C
-----·
I
I the system. When, after reaction between the cornponents'or phases of a
C 1 24·8 25·5 26·2 26·9 27·3 30·0 32-9 36·5 system, thermodynamic equilibrium is attained, the concentrations of
C::1 deg-1 rnole " \ the components or {he compositions of the phases then present are
375 395 425 450 500 550 related (irrespective ,of their initial concentrations) through the com-
T°C I 385
I ponent fugacities or activities at equilibrium. This relation is inherent in
c, : 50·4 62·0 33·0 30·2 29·2 29·9 30·7 the definition of the equilibrium constant for the reaction in terms of the
cal deg-1 mole-1 \ standard free-energy change, equation J;58, for the latter is ~ function
of temperature only 'and, being an intensive property, is independent of
3.1 On heating, the metal thallium undergoes a phase transformation
the amount of the reacting species present.
from a hexagonal to a body-centred cubic structure at 234°C. Calculate
The thermodynamic approach can only yield information relating to
the entropy of fusion of hexagonal thallium at its hypothetical melting
the equilibrium condition, giving no indication of the rate of approach
point, using the Cp data in appendix 3. (Use 5-figure logs.)
of the system towards equilibrium'. As.an illustration of this limitation,
consider that in air at atmospheric pressure all the common metals
References should completely oxidize at room temperature. Their commercial use
1. L. KAUFMAN, Acta Met., 7, No. 8, 1958, p. 575. in metallic form is possible only because the equilibrium condition is
approached at a very slow rate at this temperature .. ,
Consideration of heterogeneous systems· at equilibrium (i.e. systems
comprising more than one phase) involves an important principle of
therrnodynamics.. for the conditions for equilibrium will, in general,
apply to the ·whole,Lsystem,and not only to one phase thereof. For
example, if a liqui(or solid metal and its oxide are exposed to a gas
containing oxygen, then at equilibrium the partial pressure of oxygen in
equilibrium with the condensed phases must also equal that in equilibrium
with the other components of the gas phase (cf. examples 4.5, 4.11, 4.12).
As shown in earlier chapters, often a reaction can be regarded as
arising from a series. of more.
,
elementary reactions. Thus, following
' '

Hess's law, the stan~ard· free-energy changes for the more elementary
t ; '
' !~
 62 THE EQUILIBRIUM CONSTANT .. THE EQUILIBRIUM CONSTANT 63
reactions may be manipulated to yield the value Joi: the reaction of For reaction 4.1.J, equation 1.58 is expressed as:
interest. From the logarithmic relation 1.58 between the standard free-
r1/2 a';
energy change and the equilibrium constant, the value of the latter for L1 G0 = - RTin.J_Q~
such a reaction can be obtained by appropriate multiplication and acu,O
division of the equilibrium constants for the elementary reactions, -4·575Tlogp~~ 4.1.3
corresponding to the addition and subtraction of the standard free-energy
. changes for the latter (cf. example 4.11).
assuming that the Cu and Cu20 are in the pure state (i.e. a= 1), and the
oxygen behaves as an ideal gas (i.e.f0, = PoJ·
Although, to be thermodynamically precise, the equilibrium constant
In a vacuum of 10-~mm Hg, taking the partial pressure of oxygen in
must be expressed in terms of the fugacities or activities of the components
air as 0·2 atmosphere:
of the reaction at equilibrium, it can often be assumed that a gas behaves
ideally when its pressure is not greater than 1 atmosphere and the 1
temperature is well in excess of the boiling point of the gas. In such
Po. = 0·2 x 10-5 x
760
= 2·635 x 10-9 atmosphere.
circumstances, the expression for the equilibrium constant can be simpli-
(This, of course, ignores vacuum pump vapours, etc.) Therefore:
tied if the fugacity is replaced by the partial pressure of a gaseous com-
poricnt." In addition, it is sometimes justifiable to assume that little or LlG0 = -2·288T!og(2·635 x !0-9)
no reaction occurs between solid components in a system and, if these
= + 19·6T. 4.1.4
arc present initially as pure elements or compounds, their activities may
be taken as unity (i.e. the component is in its pure standard state). For The minimum temperature (Td) for dissociation of Cu20 is given by
example, in a gaseous oxidation reaction the metal and its oxide may the equality of relations 4.1.2 and 4.1.4. That is:
appear at unit activity in the expression for the equilibrium constant for 40,500+3·92TdlogTd-29·5Td = 19·6Td
the reaction. or:
These princip)es are demonstrated in the following examples. 3·92Tdlog Ts = 49· 1 Td -40,500

4. ! Determine thelowest temperature at which copper oxide (Cu20) can which is best solved graphically. Thus by plotting (49· lT-40,500) and
dissociate in a vacuum of I'O-' mm Hg. (3·92Tlog T) versus T, the minimum temperature of dissociation of Cu20
is obtained as the intersection of the line and the (very slight) curve at
There are several ways in which this example could be solved, all of T= 1,085°K, or 312( ± 3tC.
which arc essentially equivalent. The basic step in calculating the mini- This value can be checked by solving 4.1.3 for the value of Po, at this
mu rn temperature for a reaction to proceed is to determine the free-energy temperature. At 1,085°K, LlG0 (from 4.1.2) is +21,400 and hence, from
change for the reaction as a function of temperature. Under a particular 4.1.3, Po, is 2·38 x 10-9 atmosphere. This is somewhat smaller than the
set of conditiors, the minimum temperature is that a,t which the free- value calculated earlier and it follows that the equilibrium temperature
energy change is related to the value of the equilibrium constant for the is too high by about 2°C, which is, however, within the error limits of
reaction (under the same conditions) by equation 1.58:: the value determined graphically.
L!G0 = -RTlnKr .. ,. 4.2 In separate experiments, pure nitrogen at atmospheric pressure is
The dissociation reaction: passed over either mercury or a mercury-thallium amalgam (68·4 atom-per
Cu20<•> ~ 2Cu<•) + !Oi<s> 4.1.1 cent thallium). At temperatures less than l00°C the vapour pressure of
thallium is negligible, relative to that of mercury. The flow rate of the
is accompanied by a standard free-energy change:
nitrogen is controlled such that the gas is just saturated with mercury
L1G0 = 40,500+3·92TlogT-29·5T'cal. 4.1.2 vapour as it passes over the pure mercury or the amalgam. The vapour
• Indeed, sometimes there is no alternative to this replacement, for the fugacity of carried over in the gas stream is condensed in a cold trap and weighed. In
the gas may be unknown. · · · experiments with the samples at 26°C, the weights of mercury trappedfrom

t.l ti
64 THE EQUILIBRIUM CONSTANT
THE EQt)ILIBRIUM CONSTANT ss
the pure mercury and the amalgam were respectively 2·20 and0·75 mg per When passed over the pure mercury, this volume of nitrogen contains
hundred litres of nitrogen, the nitrogen volume being measured at 20°C and 2·20 mg mercury. As 200·~ g of mercury . (= gram atomic weight of Hg)
I atmosphere pressure. Assuming that equilibrium is continuously main- occupies 22·4 litre,yol~;~e at·N.T.P.,·th:eiefore: .
tained over the samples, calculate the partial pressure of mercury over the : ; .•· ., . •J < 200.
2·2 mgHg octhp1es·
. 6;x 22·'4·, ...
2·4 = 0·239 cc at N.T.P.
amalgam and the activity coefficient of mercury in the amalgam at 26°C.
This example is concerned with the carrier-gas experiment, where the 299
/ - = 0 • 239 X CC at 26°C
partial pressure of a volatile component over a solution is measured by 273
'
sweeping away the vapour of the component, under equilibrium condi- I · '• · · · .·· .;. 0·262 cc at 26°C.

l
tions, with an inert carrier gas and weighing the condensate. The activity . . 't. .. ...
of the volatile component in the solution is directly calculable from Thus, the partial pre~sti(.e of-Hg over pure mercury
. \. . . ' .
measurement of the partial pressures of the component over the solution
and over the pure component. The reactions may be written as follows. ! '..;i., .'. = 0·262 X
760
;~ . 102,000
For pure mercury: ,i..·

H&u,ure, I) ~ H~ 111 N1)

'
!': = 19·5 x 10-4 mm Hg.
When passed ove.r thoJme'rcU:ry~thaJlium amalgam, 102 litre of nitrogen
/H~\ contains 0·75 mg mercijry, which occupies 'a' volume of: . :
KP = ( a8J (puro H&)
22,•4 { I '.. . ' .: .: ; . "
= PH,(pµre Hg)
. '5:oo-6 x. CP5·~ ~ 0:082 cc a't, N.T.P.
assuming that mercury vapour behaves as an ideal gas, and as aH, = 1 .
. · 'CC at 26°C.: - ,',,,: 0'.0.89
when referred to pure liquid mercury as the standard state. Over the ,~: ' . ... (

amalgam: Thus the partial pressure ofi{g overthe Hg7Tl amalgam

H~1 i.n amalpm) ~ H&<i in N 1)
.· . : 0·089 ... 7.. 60 .. ·. ·.,...
=·~--.--~.x .. \ ...
.. · 102,000 . _ . ·
K -(~\
P a~ (a:nalpm) = 6;65 x W:-4 mm Hg.
again assuming that the mercury vapour behaves ideally. Therefore, from 4.2.l:
Since the equilibrium constant KP for a reaction in a system is inde- PHa
aH = ---
pendent of the initial composition of the system, it follows that: g P0H11
PHg(amalpm)
4.2.1 6·65 X J0-4
a Hg (amalpm) = PH& (puro H&) 19;5 x.10-4
(Note that this relation is equivalent to the definition of activity for the = 0·34
mercury through 1.53 and 1.54 if the fugacities in the latter are replaced and, from I .74:
by partial pressures.) aH , 0.·34
B - -·-
The volume of the nitrogen was measured at 20°C, whereas the experi- YHa = J,(.
Ha
- 0·316
ment was performed at 26°C. This latter is the temperature for which the
= 1·08.
partial pressure must be calculated. Thus (assuming the nitrogen behaves
ideally): The detailed calculations and corrections of the volume of the N2 to
the temperature of. thf experiment are, unnecessary to calculate tho
Volume (of 100 litre at 20°C) at 26°C = 100 X ~:~ = 102 litre. activity of the m~rcury~~lone. For this value is given by the ratio of tho
. . .: ' " .
. . j~
:... ~·~ ~\'~ . I

·,
'~
6<i THE EQUILIBRIUM CONSTANT THE EQUILIBRIUM CONSTANT 67
weights of mercury carried over by equal volumes of carrier gas from the when the gases are assumed to behave ideally. Substituting the given gas
alloy and pure metal. Thus: composition in equilibrium with the steel:
0·75 · ·
aH, = = 0·34 as calculated previously. KP = O·O 162 = - 0·0064-
- - - -
2.20
or:
4.3 A methane-hydrogen mixture at I atmosphere p essure is allowed to I 0·0064
come to equilibrium with a steel containing 0·60 weight-per cent carbon at ac = 0·0162 · (0·9936)2
925°C. On analysis, the gas is found }O contain 0·64 per cent methane and
99·36 per cent hydrogen by volume. = 0·40
Calculate the activity of carbon in the alloy and the expected analysis i.e. the activity of the carbon in the steel, relative to graphite as the
(by volume) of a carbon monoxide-carbon dioxide mixture which would be standard state, is 0·40.
in equilibrium with the steelat the same temperature and pressure ..
The composition of the carbon monoxide-carbon dioxide atmosphere
When data are available for the equilibrium condition for a reaction in equilibrium with carbon at this activity is obtained from the equilib-
in a given system; it is often possible to determine the equilibrium condi- rium constant for the reaction:
tion for a similar reaction in a related system from knowledge of the
equilibrium constants for the two reactions. This is a typical example of
co2W + C(s. IIJ"&Phite) ~ 2COw. 4.3.2
this principle. The activity of carbonin.the alloy is determined by the The standard free-energy change for this reaction is:
hydrogen-methane gas composition and the equilibriumconstant for the
carburizing reaction. Using this value of the activity; the composition of
"1Go = +40,800-41-7Tcal
a carbon monoxide-carbon dioxide gas mixture can be determined from and hence, from 1.58:
knowledge of the equilibrium constant for the carburizing reaction with
these gases. The determination of activity by partitioning between two
4·575TlogKP = -40,800+41·7T
liquid phases is a further example of this principle. · or, at 925°C (1,198°K):
The standard free-energy change for the reaction: KP= 46·8.
Cc,. vaphtte> + 2Ht<a> ~ CH400 4.3.1 The equilibrium constant for 4.3.2 is:
is given by:
"1Go ••• -21,600+26·2Tcal. K = Ito
P fco,.ac
Hence, from 1.58:
Jrco
4·575TlogKP = +2f.600-26·2T; Pco,.ac
or, at 925°C (1,198°K): assuming ideal behaviour of the gases. Substituting the calculated values
for KP and ac:
KP= 0·0162.
' ' Jrco = 18·72 ·
The equilibrium constant for. reaction 4.3.1 is given by: Pco,
K .,,, _/cH. or, because Pco+Pco, = 1:
p ,2
. JH,.ac rco = 18·72(1-pco)
PcH. or:
""'JrH,.O~
 THE EQUILIBRIUM CONSTANT 69
68 THE EQUILIBRIVM CONSTANT
Now, from 1.58:
which is a quadratic inPco· Solving by the formula method:" LIG0 = -RTln KP
-18·72± v'[(18·72)2+(4x 18·72)] and hence, at 473°K, the equilibrium constant is given by:
Pco = 2 ·
= 0·952 atmosphere KP =
. [ +380 ·
antilog 4·575 x 473
J
and:
Pco, = 0·048 atmosphere. = 1·49(4)•
That is, a gas comprising 95·2 per cent CO and 4·8 per cent CO2 (by But the equilibrium constant for reaction i.4.1 is given by:
volume) at 1 atmosphere pressure is in equilibrium with the steel con-
taining 0·60 weight-per cent carbon at 925°C. KP = I ·49(4) = fll 2·t~J2•
_J __
fNH,
. p~2-pi/2
· 4.4 The stability of ammonia gas depends on pressure as well as on = _!_~ 4.4.2
temperature. Show that this is so by calculating the extent of the dissocia- .· Pr-itt,
tion of ammonia to form hydrogen and nitrogen at 200°C at l ·O and 0·8 if the gases are assumed to behave ideally.
atmosphere total pressure. The two additional r~lations between the partial pressures are obtained
In general, the solution of a problem containing x unknown quantities by consideration bft!ilte'st!'i'cfio~s Ori the system. Thus, as the dissocia-
requires knowledge of x independent simultaneous equations relating tion of ammonia always results in the formation of 3 mole of hydrogen
these unknowns. In the present examples one of these relations can be to I of nitrogen, it follows that: ' ·.i ·
obtained from the equilibrium constant. A second relation may usualy 3pN, =.PH,· 4.4.3
be derived from an imposed restriction of constant pressure or volume.
Thus, in the previous example the values of the two unknowns, Pco and I
In addition, the total 'pressure is fixed at first l ·O and then 0·8 atmo-
I sphere pressure; i.e. fcfr-1 atmosphere pressure:
Pco., were obtained from knowledge of the value of the equilibrium
constant and the imposed restriction that the total pressure (Pco + Pco,)
equalled 1 atmosphere. In the present instance there are three unknowns, I f PN, +pH,+ PNHa = 1. 4.4.4

l
PNH,, p8, and PN,, and three independent simultaneous equations are Substituting first 4.4;'.4 and then 4.4.3 for PNHa and PN, respectively
required. In addition to the two relations used previously, a third into 4.4.2 and rearranging: . ,, . : ·
relation arises from the consideration that the mass is conserved within
the system, i.e. the reaction occurs in a closed system, with no gain or loss
. O· 5~·7 Pii,·+ 1·9~2pi{, ~ 1 ·494 = 0
or:
of mass to the surroundings.j ~
The dissociation reaction is: 1! Hence ;
I : . · :-):46±v'[(~·~)2+(4x2·59)]
4.4.J
NH3W ~ tN2(.a)+ ItH2(.a)
and the standard free-energy change is:
II PH, =1 . . . ;, . .
2: . .

_i-3·46+v'<I·95+ 16:36)
LlGo = 10,400-7·1TlogT-3·79Tcal. - 2
Thus, at 200°C (473°K): _ -3·46+ vc22-31)
.d G° = - 380 cal. - 2
• See appendix 5. -3·46+4·72
t The derivation and manipulation of the simultaneous equations when four or =- 0·63(4)•
more unknowns arc involved are considered in example 4.10.
6
.............................. ~~ '~ -----
I
70 THE EQUILIBRIUM CONSTANT i THE EQUILIBRIUM CONSTANT 71
Substituting in 4.4.3 and 4.4.4:
That is:

I
PN, ""' 0·21(1) Mn0<1,+(;,> ~ Mnc,>+COw .f.5.1

PNH4 = 0· 15(5)• 2Mn0<•> + Cc,> ~ Mll(,> + CO2(&). 4.5.2

Hence at I atmosphere pressure. 84·5 per 'cent (by volume) of the The partial pressures of CO and CO2 at equilibrium with the metal and
ammonia will dissociate. · · · . ·. \ its oxide arc related through a further reaction:
When the total pressure is 0·8 atmosphere: ;
C02+C ~ 2CO. 4.5.3
PN,+PH,+PNH, = 0·8. ; 4.4.5 It is stated that the equilibrium partial pressure of CO2 is negligible at
Substituting in a manner similar · to that adopted previously and temperatures above 900°C when the total pressure of the carbon gases is
rearranging: 1 atmosphere. According to the principle of Le Chatelier, the equilibrium
#.• {1.• - •'_;"-;.
in reaction 4.5.3 will be displaced to the right as the pressure is lowered.
0·571p~.+ l·992pH,-·I · 195 ',"!"' ()\'.
or: Hence in the present example, where the total pressure of the carbon
P1.
+ 3·46p~•...,:i:073 -~· .oi! gases is I0-3 mm Hg, the temperature above whichpc0, is negligible will
Hence: ' '' 1 be significantly less than 900°C. As a first approximation, therefore, it is
- 3.·46 ;!: v'[(3·46)2 + 8· 29~]· reasonable to assume that reaction 4.5.2 occurs to a negligible extent,
' 2· J and that the reduction is adequately represented by 4.5.J. If'the minim.um
<I
\
temperature is determined at which the latter reaction canoccur when
~ 0·52<,, .t Pco = 10-3 mm Hg, then substitution of this temperature and this value
•••• -3·46+4·51_
2
of Pco in the equilibrium constant for 4.5.3 will iridicate whether or not
the assumption that Pco, is negligible is justified.•
and from 4.4.3: PN, .,..; ci'. 17(4),
The standard free-energy change for reaction 4.5.J is obtained from
from 4.4.5: PNM. = 0· 10(3); data for the reactions:
,l : .
Hence, at 0·8 atmosphere pressure, 86·4 per cent;;(by volume) of the MnO<•> ~ Mnc,>+-!Ous>: JG0 = 91,950-17·4Tcal 4.5.4
ammonia will dissociate. ··
Cc,)+!02(&) ~ co(I): JG0 = -26,700-20·95Tcal. 4.5.5
The effect of pressure on the degree of dissociation of ammonia
follows the Le Chatelier principle. Consideration of 4.4.1, in which I mole Adding reactions 4.5.4 and 4.5.5 yields reaction 4.5.l, for which,
of ammonia dissociates to give 2 mole of nitrogen and hydrogen, would therefore:
indicate that the degree of dissociation increases with decrease in total Jc,O = +65,250-38·35Tcal.
pressure. This has been calculatedto be so. The minimum temperature for reaction 4.5.J to proceed will be that at
which this expression equals (-RTinKp) for the reaction, where KP is
4.5 The equilibrium between carbon and its oxides is such that above determined by the conditions prescribed (i.e. pure carbon, manganese
about 900°C the formation of carbon dioxide is negligible at I atmosphere and manganese oxide and carbon monoxide at 10-3 mm Hg pressure),
pressure. With this in mind, calculate the approximate minimum tempera- i.e. when:
turefor the reduction of pure manganese oxide by solid carbon toform pure
65,250-38·35T = -4·57:,Tlog0Mn-/co ,l.5.6
manganese metal in a vacuum of 10-3 mm Hg. If the reduced metal enters aMnO•ac
into solution in iron and the minimum temperature of reduction is thereby
= -4·575Tlogpco
lowered 53°C, calculate the activity of the manganese in solution in the iron.
• Alternatively the simultaneous formation of both CO and CO2 in the reduction
The product of the reaction between the metal oxide (MnO) and solid process can be allowed for in the precise determination of the minimum temperature
carbon is .the' metal . (Mn), carbon monoxide and/or carbon dioxide. for reaction. The more complicated procedure necessary in this case is demonstrated
.', . . ; . j~ in example 4.12. ·

-~,,
 \f:
, T~E. EQUILiBRIUM CONSTANT 73
72 THE EOL'ILIBRIUM CONSTANT
4.6 A gas consisting of 6b) per cent ~2, .39·8 per cent H20 at l atmo-
when all the condensed phases are pure and the carbon monoxide
sphere pressure is in equilibrium with p'l:Jre·y-ironat 910°C. At the same
. behaves ideally. Thus:
temperature the gas composition in ~c;uilibi:ium 'with an iron-nickel alloy
10-3 (containing 0·721 atom fraction of irori)fs 5!'·9per cent H2 and 48·] P'"'
65, 250-38·35T = -4·575Tlog- 760
cent H20. Determine the ~ctivity of irqn in the ~lloy. .
+26·90T
This technique, in which the oxygen potential.over a metal or an alloy
or: 65,250 = 65·25T. is controlled by the gas atmosphere, can be. used to determine the activity
Hence the minimum temperature for the reduction of MnO is 1,000°K of a component of an alloy when that component forms an oxide which
or 727°C. is considerably more stable (i.e. has a more negative free energy of
The partial pressure of carbon dioxide in equilibrium at this tempera- formation) than the oxides of the other elements present in the alloy. At
ture with solid carbon and carbon monoxide at 10-3 mm Hg pressure, is 910°C (1,183°K) the standard free energies of formation of nickel oxide
now obtained from the equilibrium constant for reaction 4.5.3. For th.is (NiO) and ferrous oxide (FeO) are .:..34,800 and -44,350 cal mole+'
reaction: respectively. Hence it is evident that the iron will oxidize in preference to
LlGo = +40,800-41·7Tcal the nickel. It is essential from the point of view of the calculation that the
oxide formed is pure and does not enter into solid solution in the alloy.
-900 cal at 1,000°K
These conditions are well satisfied in the present system and the activity
and:
of the ferrous oxide may be taken as unity.
K = 1·574 = Jrco In equilibrium with pure iron, the reaction is:
P Pco,
assuming ideal behaviour of the gases and the purity of the solid carbon. Fec i+H20<al ~ FeO<•>+H2<ai
1

for which:
Hence:
10-3 l K = aF.o·IH.:.
Pco, = 760 x l ·574 ,} ' P , aFe• f H,O
1'l x 10-12 atm Because both Fe and FeO[are in theirstandard states (i.e, at unit activity)
and the gases may be assJm'ed to be 'ideal, this simplifies to:
= 8·4 x 10-10 mm Hg.
As this value is negligible, relative to Pco, the error in the calculated
;:}( - ·( P~.) :<
} P - PH~O Fe:.pure
equilibrium temperature of 1,000°K will not be significant.
The solution of the manganese in the iron results in a depression of the The corresponding reaerion with the alloy is:
:: I

minimum temperature of reduction by 53°C, i.e. this temperature Fe(Fo-~ii +H20<a> ·~ Fe01i> + Hiw, 4.6.l
becomes 947°K. Equation 4.5.6 becomes:
The activity of the pure fgrroµs oxide is· again unity, but the activity of
65,250-(38·35 x 947) = -4·575 x 9471ogaMa·Pco the iron is now less than Jnfoi' in the alloy. Thus:
.,
where the MnO and the solid carbon are again assumed pure and the
carbon dioxide to behave ideally. Thus: J~

/
',;,··

" · aFe:PH,O ~0-:.Ni

(_:!!_~-·~). '
logaMo =
-28,900
. ~~~ +5·88 However, as the value ot
the' equilibrium constant for the reaction is
unchanged on solution of the iron in the alloy, then:
or:
aMn = 0·159. '(. PH. )· ·: ' : '. ( .~P'tt/ )·
The activity of the manganese in solution in the iron is 0· 159. :, p;;,;,:, Pe; ~ure = \aF."-'Ptt.o Ft>-NI '
................ ._..f!!l'lrn•~-------------------------.....-.-- __..., --,
.:.,

74 THE EQUILIBRIUM CONSTANT

THE EQUILIBRIUM CONSTANT 75
··;.
Rearranging terms: . to give:
Ope=
(pHJ PH,b)Po--~;
. LlG'?a,> = -29,080+3·67Tcal
(pHJ PH,o)Fe,P;'JfO'.
-22,200 cal at 1,873°K (l,600°C).
and substituting the gas compositions given:
From 1.58 the equilibrium constant for reaction 4.7.1 is expressed by:
(0·519/0·481)
Ope = (0·602/0·398) +( 22,200
· ) = I og KP = l og (OMnO, Op0) ·
4·575 X 1,873 aFoO·aMo
= 0·713.,' .. ,
" However, because the two liquid solutions are approximately ideal, the
That is, the activity of the iron in the iron-nickel alloy;Jrelative to pure activities of the reaction components can be replaced by their atom
iron as the standard state, is 0·713. ·· · fractions (equation 1.66). Hence:
It should be noted that the ratio:
I og ( NMnO•
-·--- NF•) - 2 ".)~9
PH, ) ( = 0·602 •••. 1·52) ',.
Np.o,NMn .
( ---- -.-
PH,O Pe. pure\ O· 398 : · .. ;/'
. Substituting the given values of Npe0 and NMno:
which is equal to the equilibrium constant· for reaction 4:6./, can be
NFe 0·70 .1
obtained from a knowledge of the standard free-energy change for the N =
0_30.anti
og2·59
reaction (cf. example 4.5). However, the gas composition in equilibrium Mn
with the alloy can be obtained only by experimental measurement. = 908
and, as:
4.7 At J ,600°C, liquid solutions of MnO in FeO and Mn in Fe are
approximately ideal. Determine. the weight-per cent concentration of therefore:
manganese in iron which is in equilibrium with an oxide melt (slag) con- NFe = 0·9989
taining 0·30 mole fraction of MnO and 0·70 mole fraction of FeO at this NMo = 0·0011.
temperature.
The concentration of manganese in iron is required in terms ofweight-
This example (though a simplified version) is of practical importance per cent and the atom fraction may be converted as follows.
in relation to the refining of steel, and demonstrates how the equilibrium For any binary system the ratio of the weight percentages is equal
relation may be applied to wholly condensed systems in an exactly simply to the product of the ratios of the atomic weights and the atom
similar manner to that used in previous examples involving gaseous fractions of the two components, i.e. for the Mn-Fe system:
phases. In addition, the example illustrates the conversion of concentra-
tion in terms of mole fraction to per cent by weight, an often-necessary wt %Mn M Mn at %Mn
procedure in practical systems. · wt%Fe = Mp0 • at%Fe
The exchange. of iron and manganese between r:rj~tal and slag is where MMn and Mp. are the atomic weights of manganese and iron.
controlled by the reaction: In the present example:
'· ' FeO(l)+Mn(1>· ~ MnO(l)f Fe(l). I 4.7.1 wt %Mn 54·94 0·0011
IOO-wt%Mn = 55·85. 0·9989°
The free-energy change for this reacti~r1:is obtained.by subtracting:
Hence:
Fe(l,+102<&; ~ FeOttJ: .LJG0. = ~55;626+'1oi83Tcal wt%Mn = 0·108.
from: Thus the concentration of manganese in iron in equilibrium with the
Mncl)+102w ~ Mn00,: JOO - . ~-~;7oo-+J4·5T,cal oxide melt is O· l 08 weight-per cent.
' ".
76 THE EQUILIBRIUM CONSTANT THE EQUILIBRIUM CONSTANT 77

4.8 From measurements of the equilibrium between hydrogen and It should be noted that°this result may also be obtained by calculation
hydrogen chloride gas with liquid sodium and solid sodium chloride, it is of L!c,0 at 773°K from ~~73-i and insertion of this value, and that given
calculated that the partial pressure of chlorine gas in equilibrium with the for AH0, in the relationr
two condensed phases is 2·2 x 10-46 atmosphere at 50Q°C.If the standard . :f.1 .:1Go '= LJHO-TLJSO
. '
heat of formation of sodium chloride is assumed to be independent of
temperature and equal to -98·6 kcal mole:", calculate the equilibrium to yield Ll S'', Assuming'~ H0 and A_S0 are independent of temperature
partial pressure of chlorine under the same conditions at 60Q°C. (as does the method of ~},lculatio:i,. shown in detail), L1 0 at 873°K_ may o
now be calculated, from ~vhich 1(8;3.K (and hence Pct, at this temperature)
This example demonstrates the application of the approximate form is found. The temperature 1it1dependencb of Ll!{0 is, of course, inherent
of the van't Hoff relation, 1.61. The known value of the partial pressure in the approximate forni of the van't-Hoff equation, 1.61, employed in
of chlorine is first used to deduce the value of the equilibrium constant ' .
the detailed calculation. · · . . · ." : .· :·. ·..
: .·· :· . .
for the sodium chloride formation reaction at 500°C. That is: j.. ~· . " ,· : : ' ' ;_ ' . . '
4.9 Show that the basis of operation of the iron blast furnace - the
reduction of ferrous oxid~ (Fe6) by a ·co-C02-atmosphe~e _: is a process
which is dependent upon 'marked deviationsfrom equilibrium at tempera-
tures around 600°C.
There are several different ways in which this example can be answered.
I One approach is to determine whether or not the partial pressure of
= pi!,; oxygen in equilibrium with iron and FeO is less than that in equilibrium
with the CO-CO2 atmosphere. This is the necessary condition for the
if sodium and sodium chloride are in their standard pure states and the
reduction reaction to proceed. The equilibrium ratio Pco/Pco, is· fixed,
chlorine behaves as an ideal gas. Thus: :
however, by the equilibrium between these gases and the solid carbon in
K I . the charge. Hence a more simple approach is to determine whether or
P = y(2•2 X 10-46) · not the ratio PcofPco, in equilibrium with graphite at the stated tempera-
ture is greater or Jess than the ratio in equilibrium with iron and ferrous
= 6·74 x 1022 at 773°K. oxide at the same temperature.
As a basis for calculation it will be assumed that the solids, iron, ferrous
Substitution of this value ( = K1) at the temperature T1 in 1.61 yields the
value of the equilibrium constant (Ki) at the temperature 873°K (T2). oxide and carbon, are present in their pure standard states and the total
That is: pressure is 1 atmosphere. The carbon monoxide and carbon dioxide are
produced partly by combustion of the coke in the charge with the oxygen
-98 600( 1 I ) in the air blast and partly by reaction between the coke and the combined
!og(6·74x 1022)-logK2 = 4.5~5 873-773
oxygen in the ore. It will be assumed/ therefore, that the total pressure
or:
of carbon gases (Pco+.Pto,) is approximately 0·5 atmosphere, the
K2 = 4·35 X 1019.
a
remainder of the gas· <~\t totai pressure of 1 atmosphere) being the
Hence as:
diluting nitrogen from the air blast. . .
1
Pei.= K; The free-energy change for the reaction:
therefore: ;' C(•> + C02
~-
w ~ 2COw 4.9.J
1 is given by:
Pei.= (4·35x1019)2 A<}~= +40,800'-.41-7Tcal.
f,;, .
At 600°C (873°K): <,

= 5· 3 x I o....w atmosphere. · \' AG0

,) = +4 , 400 cal '
78 THE EQUILIBRIUM CONSTANT THE EQUILIBRIUM CONSTANT 79
and, from 1.58: Hence:
RTlnKP - -4,400
+390 = -4·575Tlog(Pco)
·= RTln~( . n: )., Pco
.
co,.ac assuming that Fe and FeO are at unit activity and the gases are ideal.
. i' Therefore:
Thus, assuming the carbon is pure graphite and the gasesbehave ideally,
. if:.o . . . , log (Pco) = -0{)9 ;7
log-= -1·104 Pco,
Pco, · and:
i
Pco = 0·799.
['co = 0·079 at 873°K. Pco,
Pco,
However, as: This ratio is larger than that calculated as being in equilibrium with
Pco, = 0·5-Pco graphite and hence, if the latter equilibrium was attained in the blast
therefore:
furnace, on this basis the iron would exist only as the oxide. In practice
P~o - 0·079 (O· 5 - Pco)
the iron oxide is impure and the activity of ferrous oxide is usually very
P~o+0·079Pco-0·040 = 0 much lower than unity. From inspection of equation 4.9.2 it is evident
that a decrease in aFeO would increase proportionately the ratio PcofPco,
a quadratic in Pco- which is solved by the formula method. to give: in equilibrium with the metal and its oxide. For example, if ape0 is 0·5,
· -0-019 ± vrco-019)2+4(o·040)J the equilibrium ratio Pcof Pco, is increased to 1 ·598.
Pco = -----2------ As the activities of the iron and graphite will not be very much less
than unity, it will be realized that the impurity of the solid reactants
= 0· 163 atmosphere. cannot be the basis for an explanation of the practical occurrence of the
Hence: reduction reaction which should not, in theory, take place. The practical
Pco. - Cr337atmosphere realization of the reduction of ferrous oxide in the lower temperature
and: •,i
regions of the blast furnace is, therefore, only possible because of marked
' · Pco -= 0·484. · · deviations from equilibrium conditions. These are mainly in the equilib- .
Pco, rium in reaction 4.9.1. The ratio Pcof Pco, in equilibrium with graphite
The ratio Pcof Pco, in equilibrium with'Fe and FeO under the stated decreases from near unity at about I,000°C to 0·484 at 600°C and to
conditions is now calculated from th~ ·cquilib.riurp: constant for the near zero at about 400°C, but the reaction occurs very slowly in relation
reaction: to the time available as the gases ascend the blast furnace stack. In fact,
Fec,i+COus> 'F·FeOc,>+COw. 4.9.2 experimentally it has been shown that finely divided iron oxide can be
The free-energy .change for this reaction is obui.itied by a.ci'&ing those for converted completely to cementite (Fe3C) by exposure to a CO-CO2 gas
reactions:
' • _?. atmosphere at 600°C.
/

Fe(•l+!-02u> ~ FcO<,): JG°= -62,050+14·95!cal, 4.10 A gas mixture of20 per cent CO, 20 per cent CO2, 10 per cent H2
and:
~:
and 50 per cent N2 (by volume) is fed into a furnace at 900°C. Find the
CO2(¥)~ co~)+!02w: JG0 =·+67,500~20·7~~cal equilibrium composition of the gas if the total pressure in the furnace is
I atmosphere.
to obtain:
A Ci'?o.2) = + 5,450-S:SOT cal This is a problem of obvious practical importance, for mixtures of
=- +390 cal at 873°K. · these gases arc used as controlled atmospheres in commercial heat-

: '-5·
 i
iO THE EQUILIBRIUM CONSTANT l THE EQUILlQRIUM CONSTANT 81

reatment furnaces. The gases will tend to react with each other at 900°C, r
I
In addition, the initial composition of.the: gas mixture is kno~n. As the
amount of any.one ele~nt,is conserved throughout (i,e. only its mode
.quilibrium being finally established for the reaction:
cow+H20w ~ co2W+H2W 4.10.1
I of association is chailgedi, {4rt}ler.equations relating the partial pressures
may be derived. Thus if~1 denotes the number of gram-atoms or moles
Data are available for the formation reactions for each of the com- I of the element or compound i, then, forthe present mixture of gases:
»onents of this reaction. Thus: Ii . . nl =::: nc~ + nc~~. . }
Cr,)+Wi(I> ~ C0,1> : AGO= -26,700-20·95Tcal 4.10.2 ~·, I ' ' • •

C<•>+02w ~ C02W : AGO= -94,200-0·2Tcal 4.10.3

nJ = 2nH,+2nn~o· 4.10.7

Hl(a>+i02W ~ H20w: AGO·-= -58,900+ 13·1Tcal. 4.10.4

•. n}:\ 'S .
h~ ¥2nco,+
. :, .- . ·,
nH;o
where nc, nH and n0 arc constant, independent of the analysis of the gas in
Hence, at 900°C (1,173°K):
the system. Thus these ioay be related to each other from the initial
Ll61'oo.J) ••• -51,300 cal analysis of the gas._ For tl\e analysis: 20 per cent CO, 20 per cent CC>2, _10
.d 61100.J) - - 94,400 cal per cent H2, it is deducetthat.: , ,. : '. ·
AG1'4.I0.4) ••• -43,500cal. nc- = ·2nn .
Reaction 4.10.l is given by 4.10.3-{4.10.2+4.10.4}, and hence: . ncf,= 3nH
..:161'4.J0.1) = -94,400-(-94,800) ; 2nc == 3rio ·.
+400 cal at 900°C. Therefore, substituting tJc ilppropriatc r~lations 4:10:1: ..
·.~ · . · i. . . .: r
Therefore:
-AGO]
Kp(4.10.n ,.,. antilog [4·575T and:

- antilog [ -0·0745]
t
••• 0·842. i As:
P1 = --
. n1
.
p
or p1 cc 111,
n,01at
This equilibrium constant for 4.10.1 is given by:

Kp-
lco,.fH,
lco•IH,O
or, assuming all the gases behave ideally under the prescribed conditions:
I
t!
(where pis the total pressure), these relations may be rewritten in terms
of partial pressures. Thus the relations:
4.10.8
and:
Pco.•PHo. • 0•842. 4.10.5 ;
\ or:
Pco + 2Pco. + PH,o = 6p"•.T 6pa.o
Pco·PH,o 'I
Pco+2Pco, = 6pH,+5pH,o 4.10.9
This is the first relation between the partial pressures of the com-
ponents. As there are four of the latter, it is necessary to find three other are the remaining two equations needed to solve for the four unknowns.
relations between them in order to solve for their values (cf. example 4.4, Of the four equations 4.10.5, 4.10.6, 4.10.8 and 4.10.9, the latter three
arc linear and are best used to substitute for three of the unknowns in
p. 68).
Because the total pressure in the furnace is 1 atmosphere and the gas 4.10.5. · · · f • · . · ' .
mixture contains 50 per cent by volume of inert nitrogen, then: For example, {4.. 10,;6 Cf i·I0.8} gives:
.. ~.-/·,!t. . •· ; . :\ ' '
Pco + Pco. +PH,+ PH.O - O·S (atm). 4.10.6 · ··-;. · ··:-:- 5pH·, +•Spl1,cf= 0·5
··;
\~

: ~I I • •1 '-I'

1\'
i'/,
\'
sz THE EQUILIBRIUM CONSTANT THE EQUILIBRIUM CONSTANT 83
or: a case, the use of determinants= is the simplest way of solving the linear
. PH. - ~l-pH,O 4.10.10 equations. Thus, in the present example, the relations 4.10.6, 4.10.8 and
4.10.9 may be rewritten:
and {4.10.9-4.10.6} gives:
Pco + Pco, + PH. = (0·5- PH10)
Pco.-1pn.-6PH.o - -0·5. ti.JO.I I
Pco + Pco; - 4pH, = 4pH,o
Substituting for PH, from 4.10.10 yields:
4.10.12
Pc'J+2Pco.-6PH. = 5ptt.o
Pco. = 0·2-PH.O from which:
Similarly, substituting 4.10.10 and 4.10.12 in 4.10.6 yields:
(0·5-pH,o)
Pco - 0-2+PH.o· 4.10.13 4pH10 I -4
Relations 4.10.10, 4.10.12 and 4.10.13 may now be substituted in 4.10.5 p I 5pH,O 2 -6
toobtain: co= l
1
(O· 2- PH,o). (0· l -PH.o) ...,
0.842 1 1 -4
(O· 2 + Pn,o). PH,O
or: 2 -6
0· 158pt0 - 0·468.PH.o + 0·02 ;.,, 0
- (O· 5 - PH,o) X 2- I ( - 4pH,o) + 1 (3PH10)
a quadratic in PH,O· Solving by the formula method: - ----- ( +2)-(-2)+ I
PH.o = :o.o43; = 0·2+Pn.o as derived previously (4.10.13).

Resubstituting in 4.10.10, 4.10.12 and 4.10.13 in turn yields: Similarly Pco, and PH, may J-.e expressed in terms of p810.

PH, - 0·057
Pco, = 0· 157 ·
Pco = 0·243
i.e. the equilibrium composition of the gas at 900°C is .,
•..,~.'
50% N2, 4·3% H20, 5·7% H2, 15·7% CO2 and 24·3% CQ (by volume).
·1
This type of calculation is very commonly encountered fo equilibrium
studies. As noted, the determination of the concentrations of n com-
ponents at equilibrium involves the use; 'of the standard free-energy
change (relation 1.58) and (n-1) additionalrelations which are formu-
lated on the basis of the known restrictions inherent in thesystem (e.g. a
fixed total pressure or volume, conservation of mass, etc.), These latter
(n - l) relations are (or can be made) of linear type.· · contact with nickel and its oxide must necessarily be equal to the partial
While the value of n in the present problem is only four, and partial
solution of the (n-1) linear relations is a matter of simple algebraic
manipulations, these become excessively tedious when four, five or six
linear relations must be solved for five, six or seven unknowns. In such
I
!
4.11 It is found experimentally that a gas consisting of 1 ·95 per cent of
CO and 98·05 per cent of CO2 (by volume) is in equilibrium with pure solid
nickel and pure solid nickel oxide at l,S00°C and I atmosphere total
pressure. Calculate the partial pressure of oxygen in equilibrium with pure
nickel and its oxide and the standardfree-energy offormation of the oxide
at that temperature, given only thefollowing standardfree-energy data:
Cc,>+ tOzw ~ co<.s> : ,a c;o - -- 26; 700- 20·95T cal 4.11 .1
Cc•>+Ozw ~ C02ca>: .dG0 = -94,200-0·2Tcal 4.11.2
The solution to this example emphasizes one of the most important
concepts of thermodynamics. That is, if the system as a whole is in
thermodynamic equilibrium, parts of the system cannot be in differing
states of equilibrium.f In this instance, the partial pressure of oxygen in
pressure of oxygen in equilibrium with the mixture of CO and CO2•
• See appendix S.
t This concept has been implicit in previous examples (e.g. 4.S) but appears here
u the CSIC!ltial principle in the solution.

: ? V
~4 THE EQUILIBRIUM CONSTANT THE EQUILIBRIUM CONSTANT 85
This principle is frequently employed for the indirect determination of Because the whole system is at equilibrium, the partial pressure of
partial pressures which are too small for direct measurement. Th\1$ the oxygen in equilibrium with the stated values of Pco and Pco. must
vmall value of the partial pressure of oxygen in tho system can be ascer- necessarily be the same as that in equilibrium with pure nickel and its
tained from the equilibrium constant for the reaction: oxide, reaction 4.11.5, at this temperature. Hence the partial pressure of
C0+!02 ~ CO2, oxygen in equ" .brium with pure nickel and its oxide at 1,773°K is
' 6·62 x 10-s atmosphere.
The reaction for which the experimental data are given is: The equilibrium constant for reaction 4.11.5, the formation reaction
Ni<•>+C0200 ~ NiOc•> + COw 4.11.3 for nickel oxide, is: ·
and the equilibrium constant is given by: QNiO
Kp(OIS) ""' aN1. pg;
K aN10,/cx,
,c,JJ.J> - a Nl•./CO.
" I
- :-m
Poo
Pco--- 1

since Ni and NiO are both in their standard pure states and ideal
4.11.4
.
i\"
since Ni and NiO are both in their standard pure states. Hence, inserting
the equilibrium value o~p01: · ·
.Po;

behaviour of the gases is assumed. .· 1

Substituting the experimental values of pcx, andpco. in 4.11.4: Kf <4·!1~ = v(6;o2 X IQ-S)
0-0195 tt., . ',,' ~ . = 122·8)
. .
K,. a.u .J> - 0·9805
Thus, the standard f~~e-ener~ change for the formation of NiO at
-·0-0199. 1,773°K is: ~, . ·'
Reaction 4.11 .3 is given by the sum of the two reactions: -~q> = -4·575TlogKP

N~1>+,0ua> ~ NiO<•> 4.11.5 , = - 17,000 cal.

C0200 ~ COw+,02<,)- 4.11.6 The standard free-en¢rgy' change could also be obtained from the
But 4.11.6 is given by {4.11.1-4.11.2} and hence the free-energy change following considerations, As reaction 4.11.5 is given by the difference
between reactions 4.JJ.{anQ 4.11.6, therefore: ·
for reaction 4.11.6 is:
LIG° - +67,S00-20-7STcal.
or:
Thus:
.•. ' .. · ·. . . · . K,,(4JI.J)
61,SOO 20-7~ RTln K, c,.11 S) RT\n K, o.u .J> •.. R.Tln KP c,.11 .6) RTln K .
log K,c,.11.6) - ---+--·
4·515T 4·575
•••
. .
-
. p~.6)
. T(p (4.11.J)·
and: ,, Kp(4.IIS) "". K ·.. . •
- -3·79 at 1,773°K (t,S00°C) . '. p (4.11.6)

or: · l ·99x 10-2

Thus: Kp\4.11S) = l·( . -
Kp<4Jl.6) - 1·62x lo-4 _Pco,PW.
Pco.
= 122·8.
assuming ideal behaviour of the gases, Substituting the given values of
Hence: Ll~01.:.n = -4·575Tlog 122·8.
Pco and Pco.:
Po. - 6·62 x 10-5 atmosphere. -17,000 cal at 1,773°K.
7

; ;
% THE EQU!LIBRIU!\f CONSTANT THE EQUILIBRIUM CONSTANT 87
4.12 Determine whether or not chromic oxide (Cr203) tan be reduced by more stable than either the metal or the carbide when Po. > 2·95 x 10-21
solid carbon at l,000°C without forming the chromium carbide Cr23C6. atmosphere, whereas when 2·95 x 10-21 >Po.> 3·55 x 10-22 atmos-
The problem may be considered from the point of view of two possible phere, the carbide is most stable, and the metal is stable only when
Po, < 3· 55 >< I 0-22 atmosphere. Obviously, these pressures are all so small
reactions.
l. Reduction 'of chromic oxide: that they are unmeasurable. However, as there is solid carbon in the
system, carbon dioxide and carbon monoxide will be formed, and the
Cri03(•> # 2C~<•)+~2U> 4.12.1 partial pressure of oxygen in the system will be related to the partial
for which: pressures of these . gases through the equilibrium constant for the
AGO - 267,750--62·1T-cal reaction:
orat 1,273°K: 2CO<a> + O:zw ~ 2C02(a), 4.12.3
LIOO - +188,650 cal ·'\ In the presence of solid carbon, the partial pressures of the oxides of
- -RT1np3~ : .. carbon are further related through the equilibrium constant for the
reaction:
assuming the condensed phases are pure and the gas i)deal. ·
2COw ~ CO:zw + G,>· 4.12.4
Hence: ~'
-188 650 ··: il The calculated value of the partial pressure of oxygen at equilibrium
Po.= antilog [ 4·575xl·5x1,273'~! with chromium oxide and chromium metal and that at equilibrium with
' . ·f chromium oxide and chromium carbide will correspond to unique
=antilog(-21·55) ·1 values of Pco and Pco« which may be calculated by the conventional
•• 3·55 x 10-22 atmosphere. · ,: procedure. Thus for reaction 4.12.3:

2. Carburization of chromic oxide: ..dG° = -135,000+41·5Tcal

4.12.2 = -82,100 cal at 1,273°K.
Cr20x,>+H~•> ~ -fser23C«•>+!-O~ Hence:
for which the standard free-energy change may 1?C d~hced from: 82, l 00
.• .I
Pco.
2--2 = an til og ['
Pco-Po. 4·575 x 1,273]
.lfer<•>+C<•> ~ i(:r23C6(,): ..16° - ~16,380,;.;,1·54Tcal
= 1·26 X 101". 4./2.5
and 4./2.J as: But for reaction 4.12.4:
Ac:«4.11.2') - +259.210-62·9Tcal ..de?= +40,800-41·7Tcal
= + 179,0CJO cal at 1).73°K. = -12,300 cal at 1,273°K.
Hence: Hence:
2
·1 [
-179,000 ] [ 12,300 ]
Po. "" anti og 4·575 x 1,273 x 1 ·5 Pco = antilog I,273 x4·575
Pco,

= antilog(-20·47) = 129. 4.12.6

- 2·95 x 10-21 atmosphere.
It is evident, therefore, that the chromium carbide ;Cr23C6 is formed
from the oxide at a higher partial pressure of oxygen (i.e, has a higher
'oxygen potential') than that required for the reduction of the oxide to
the metal. F·oi these calculations indicate that at l,000°C, the oxide is
The partial pressures of CO2, CO and 02 in the system must simul-
taneously satisfy the relations 4.12.5 and 4.12,6. Therefore, by successive
substitution of the two oxygen potentials "or reactions 4.12.1 (the
reduction reaction) and 4.12.2 (the carburization reaction), the condi-
tionswill be found, in terms of the partial pressures of CO2 and CO, under
which these reactions will be at equilibrium.
 !!II THE EQUILIBRIUM CONSTANT
Substituting the value of Po. for reaction 4.12.1 into 4.12.5 yields:
Pco. - 2· 12 X lo--'.
Pco
Combining this ratio with 4.12.6 yields:
Pco = 2·74x 10-2 atm } Eqw'lib num
· fior
Pco, = 5·80 x 10-6 atm reduction of
(Po, ••• 3.55 x io-22 atm) Cr203 to Cr metal.
Similarly, substituting the value of Po. for reaction 4.12.2 into 4.12.5
yields:
Pco. - 6·09 X lo--4.
Pco
Combining this ratio with 4.12.6 yields:
Pco - 7·85 x 10--Zatn:1/:S'}~'.&i~irlum for
Pco, = 4·78 x 10-5 atm •· · carburization of
(Po, = 2.95 x io-21 atm) . Cr203 to Cr23C6.
These conditions for equilibrium between chromium oxide and metal
and chromium oxide and the carbide indicate immediately that the
formation of one or other of these products is a sensitive function of tho
imposed conditions. For at reduced pressures of oxides of carbon greater
than 7·85 x 10-2 atm the chromium oxide is stable in the presence of
solid carbon, whereas in the pressure range 7·85 x 10-2 atm to 2·74 x 10-2
atm (about 60-20 mm Hg) the oxide is reduced directly to the carbide
Cr23C6• At lower pressures than this, the chromium metal is formed.
This example again illustrates one of the most important principles of
· thermodynamic equilibrium, namely, that all parts of a heterogeneous
· system in complete equilibrium must themselves be at equilibrium with
each other. In this example the equilibrium conditions for carbide or
metal formation from chromium oxide and solid carbon are also the
conditions for equilibrium between oxygen and carbon and the oxides
of the latter. By calculation of these conditions, the practical realization
of carbide or metal formation becomes a possibility.
Exercises
4.A Calculate the ratio PcolPco. in equilibrium with pure iron and pure
ferrous oxide at l ,027°C.
EXERCISES 89
4.B · When a gas consisting of ei'ght"parts by volume of hydrogen and
at
two parts by volume of an inert .diluent, a total pressure of I atmos-
phere, is passed over a solid mixture of nickel andnickel.chloride at
400°C, the partial pressures of H2, iJ;Cl and 'the inert gas arc found to be
0·240, 0·560 and 0·200 respectively at the exit to the reaction chamber.
Determine whether or not equilibrium between the gas and the solid is
attained during the passage of the gas through the chamber.
4.c Solid magnesium oxide is mixed intimately with solid silicon and
heated to l,400°C, when the following reaction occurs:
4Mg0c,) + Si<•> .== 2M&a,+ Mg2Si0'4<•)·
Determine the vapour pressure of magnesium in equilibrium with the
solid mixture.
4.o In one process for the separation of zirconium from hafnium, the
gaseous tetra-chlorides of these metals arc allowed to react at 800°C with
a gas containing 67 per cent oxygen and 33 per cent chlorine by volume.
If the standard free-energy changes at this temperature for the oxidation
of the hafnium and zirconium chlorides to give the (di) oxides and
chlorine gas are + 48·85 and - 34· 30 kcal mole=! of the chloride
respectively, determine-the end products of the separation process.
/,
4.E On oxidation of~o'st FO:..Ni alloys at 840°C, pure FcO is formed
which is insoluble in. thb alloy. Calculate the activity of iron in an alloy
which has been equilibrated with a: gasmixture of sis per cent hydrogen
and 42·5 per cent wate~ Vl!,pout at this temperature.
\
t,j>.
4.F At 850°C, the activity of. nickel i~ a solid gold-nickel alloy con-
taining 0·40 NNI is 0-8!2; relative 'to pure nickel as the standard state.
Find the minimum r;tio,°pco'/Pco, in which this alloy can be heated
without oxidation at t~s temperature.
;_
4.o Calculate the s~dard; molar free energy of formation of liquid
KC! from the liquid n{gtal and thlorinc gas at 1,000°K, given that the
0
ratio PHa!PHain equilibrium with the metal and the chlorideis4·0 X 10-13
+
when the total pressure.. :(p~c'.1 ·,·PHJ ..= 1 atm at the stated temperature.
4.H Calculate the ratio 'PH,s!P~.:in ·eqµiiibrium with pure copper and
copper sulphide at 500~,C. If the va)ue·ofthe ratio in equilibrium with a
gold-copper' alloy· (Na;..; 0·85) is 2.·88x'Io-;4 at the same.temperature,
calculate theactivity ~fficient.(y) ofcopper in the alloy. ·
;,. ' • ~- I ··:' . J ' .-.,.. ' . . , ' '
 :it•,'

90 THE EQUILIBRIUM CONSTANT

EXERCISES 91
4.1 Using only the following data,' -calculate the ratio PH;sf PH, in oxide (FeO). When equilibrium is attained the slag is found to contain
equilibrium with. silver and silver sulphide 'at 6,00°C: ' : , O· 376 mole fraction of FeO. Find the activity of the FeO, given that for
4A~,> + Szw ~ 2Ag2Sc,)· : LI ~7l'K
. ' .
= - 30'.570
. i
cal the reaction:

Hua> Szw H2Sw ;· FeOc1,pure)+COw~ Fec,)+C02(&): L!G0 = -12,080+9·868Tcal.

S~98 . 31 ·2 l 54·4 49• 1 cal deg-1 mole-I
LlH~98 0 +31,000 -4,800 cal m'ole-1 4.Q A gas mixture consisting of methane, carbon monoxide, water
. . ~ vapour and hydrogen is circulated continuously over a specimen of pure
Assume that the gases behave ideally and that the heatsof reaction are iron held at 1,050°C. The partial pressure of the water vapour in the
independent of temperature. · '.: mixture is maintained constant at a small value by passing the gas over
•:l a desiccant. When equilibrium is established, the iron contains 0·842
4.J Can pure copper sheet be bright annealed (i.e. without forming a weight-per cent carbon and the sum of the partial pressures of methane
surface oxide layer) at 750.°C in an atmosphere of high~purity nitrogen and carbon monoxide is 0·0081 atm when the total gas pressure is l ·O atm.
gas containing 10-6 volume-per cent ofoxygen? ; · · In a further series of experiments it is found that a gas consisting of
r Pco + PcH, = O·O 190 and the same partial pressure of water vapour is in
4.K Determine the minimum temperature at which calcium carbonate equilibrium with graphite when the total gas pressure is 1 atmosphere.
(CaC03) will dissociate when heated in air containing Q·15 volume-per Find the activity of carbon in the metal sample, relative to graphite as
cent of carbon dioxide. the standard state.

4.L Pure iron can exist in equilibrium with either pure solid wtistite
( FeO) or pure solid magnetite (Fe304), depending upon the temperature.
Determine which of these oxides would be in equilibrium with pure iron
at room temperature and the maximum temperature at which this oxide
is in equilibrium with iron. ·
4.M Repeat the calculation given in the previous exercise, evaluating
the free-energy-temperature relations from standard enthalpy and en-
tropy data, assuming the latter are independent of temperature. Comment
on the difference between the maximum temperature obtained by the
two methods of calculation.
4.N Determine the maximum temperature to which molybdenum can
be heated without formation of the oxide Mo02 in av atmosphere of
69 per cent hydrogen and 31 per cent by volume of steam.
4.o Calcul~te the miri.imuril temperature at which pure solid calcium
orthosilicate .(2Ca0. SiOJ can be reduced by solid carbon to form pure
silicon metal, the partial pressure of carbon monoxide in equilibrium
with the solids being (a) 0·03mni Hg:(b) (}001mzn Hg.
4.R A mixture of ferric oxide (Fe203) and cuprous oxide (CuO) is
heated to 600°C in an atmosphere containing S03 gas with the intention
of converting the copper into a soluble copper sulphate (CuS04) and
leaving the ferric oxide unchanged. Given the following data and using
the appropriate standard free-energy relations of appendix 3, determine
the range of values of Pso, in the atmosphere which will allow this to be
accomplished at this temperature, assuming that the oxides and sul-
phates do not form solid solutions.
Cuc,>+102ea>+S03eal ~ CuS04<•> : LIG0 = -94,375-5·01TlogT
+86·9Tcal
2Fec,>+ l102ea>+3S03ea> ~ Fe2(S04)3<s): LIG0 = -223,370
+ 111 ·04Tcal.
4.s It has been proposed that aluminium could be refined by first
forming AlCl gas which could then be dissociated to give AIC13 gas and
pure aluminium. Deduce the feasibility of this process by calculating the
approximate values of the equilibrium partial pressures of AICI at 1,000
and 2,000°K. if the various equilibria are allowed to establish themselves
at 1 atmosphere total pressure.

4.P A steam-hydrogen gas mixture (i,H,olPHa == 0·462) is equilibrated
. at 1,361 °C with pure solid iron and a molten slag containing ferrous
. . .. . . )~
··.
4.T A stream of nitrogen gas is passed over a series of crucibles in a
furnace, the latter having been designed to give a negative temperature
i
XI
92 THE EQUILIBRIUM CONSTANT
gradient from the gas entry end. The crucibles contain various zinc-
CHAPTER 5 .rw
,;.
.1, .~

aluminium alloys which. after equilibrium has been achieved, are ·,

~:
analysed, their compositions and corresponding temperatures in the
furnace being given in the table. In an independent experiment. the
Solutions; . -. ~..
weights of zinc vapour condensed from equal volumes of nitrogen
passed over pure zinc and the first of the alloys at the same temperature
of l,030°C are 8·20 g and 6·15 g respectively. Show that this alloy ··:,
system conforms to a relation: Most of th~ examples ;presented f~. e;ulier chapters'. have been concerned
RTin YZn = rxNL
with the thermodynamic pr~pe.r!i~:s· of
elements ari4 compounds and
mechanical mixturesofthese. However, a: very profitable.field for the
where y7..., is the activity coefficient of zinc and a. is a constant. Calculate application of thermodynamic concepts isthe study.of liquid andsolid
the value of this constant. solutions. Apart frori) the obvious useof this approach .to+determine
equilibrium conditions in systems containing solutions, thermodynamics
T°C 1,030 860 710 580 530
can sometimes be applied to indicate the physical and chemical factors
0·706 0·699 0·686 0·670 0·661 which determine the behaviour of the various components ofa solution.
One of the simplest ways of expressing such behaviour is through a com-
4.u On heating, sulphur vapour dissociates from its low temperature parison of the thermodynamic properties of the solution with those for a
molecular form S8 to molecules containing six, four, two and one atoms hypothetical 'ideal' solution of the same composition. Alternatively, as
of sulphur. If the equilibrium vapour pressure of sulphur is 2·04 atm at this ideal solution concept is of limited usefulness, the regular solution of
80(>°K, calculate the pressures of the various molecular species at that Hildebrand [l] is often a better basis for describing the thermodynamic
temperature. (A graphical solution of the quartic equation in this behaviour of real solutions. The use of these models involving the rela-
exercise should be used.) tions between activity, activity coefficient and the partial molar quantities
for the components of solutions ·is illustrated in this and successive
4.v · A producer gas containing 2·5 per cent CO2, 28 per cent CO, chapters.
l l per cent H2, 4 per cent H20 by volume, remainder nitrogen, is passed
through a regenerator at l,l00°C. Assuming that the gases behave Ideal and regular solutions
ideally and equilibrium is attained, calculate the composition of the gas 5.1 If a liquid gold-copper alloy containing ·45 atom-per cent of copper
at the regenerator outlet. is assumed to behave ideally at l,OS()S'C'. calculate the hea't absorbed and
the entropy change in}h.e.system when 1 gram of solid copper is dissolved
isothermally at this temperature in a large bath of the alloy of this
• . , :t I • · 0:: ·'·
compositton. ·:
This example conce1rnsthe addition of a solute to a solution which is
assumed to be .ideal. Thus, by definition, the activity of any component
of the solution is equalto its mole fraction, relative to the pure component
(in the same physical f<>r~) as the standard state. Consequent upon this,
the partial !JI. ilar heatiof
;... ~ .
mixing
.
of
. the . component
.
is zero; i.e. from. 1.68:
91 ~_Ji,{'.~ DiM = O . for an ideal solution.
' ,, ·'.
Further, from 1.72: ·'.,
<
S1-S? =\S/'t,;. .;;_RlnN1. for an ideal solution.
. . )t· .": .
 :, \, • ; > , ' : .f '\' "'.;. ~- .!~ ••. ! .

9'-! <: ·· ·.:JT!ONS

As on sol~tfon in the ideal liquid gold-copper alloy heat of mixing
of 1 gram of liquid copper is zero, the only heat involved in the process is
the heat necessary to convert the copperfrom the solid Jo the liquid state
at l,050°C. As the equilibrium melting point of copper is 1,083°C,
negligible error is introduced by assuming that this hea] value is equal to
the latent heat effusion of copper, which is 3,100 cal moJe-1 or 49 cal g-1•
(This is obvious from example 2.5.) The latter quantity will be the heat
· absorbed when I gram of copper is dissolved in the ideal gold-copper
alloy at l ,050°C. .. . 2Cu0>+i02W ~ Cu20<•>
The entropy change in the solution process will be the sum of two
terms; the entropy change on mixing of the copper:
Sc'1i =- -RlnNcu
and the entropy change on fusion of the copper at I,050°C (1,323°K):
S Hr<eu> ,.., Lr(eu)
!(CU) .,. 1,323 - 1,323
where Lrceu> is the latent heat of fusion of copper at the equilibrium
melting point,
Thus:
St[, = -4·575log0·45
- + l ·59 cal deg-1 mole-1
- 0·0250 cal deg-1 g-'1
and:
49
Sr<eu> ""' 1,323
.,;.. 0·0370 cal de~. 1 g-1•
' ' ··..
Thus the entropy change on dissolving :1· gram of solid copper in a 45
atom-per cent copper ideal solution in gold is 0·0620 caldeg+! at I,050°C.
If the copper was, say, at room temperature when added to the liquid
. bath, the enthalpy and entropy changes would, of co~e. be increased
by the change in these· properties on heating solid ci?pper from room
temperature to l,050°C (cf; examples 2.5 arid 3.2). . 2 63.54
IDEAL AND REGULAR SOLUTIOJ'•-: 15
~e solution of copper in gold is ideal at this concentration, the reaction of
the dissolved copper with the oxygen will be identical, as far as the heat
evolved is concerned, with that taking place when oxygen is blown
through pure liquid copper supercooled from the melting point of
1,083°C to l,050°C. This identity follows from the zero heat of mixing
of the copper in the ideal gold-copper alloy.
The oxidation of pure liquid copper can be described by.
and the heat of reaction is given by:
'1Hr = -43,700-3·68T+2·60x I0-3T2-0·20x lO~T-1 cal. 5.2.1
The latter equation is calculated by use of the Kirchhoff equation, 1.20, in the
manner demonstrated in example 2.6. Values of the heat of reaction at 298°K
of solid copper with oxygen to form the solid oxide (-40 kcal mole<), the
latent heat of fusion of copper (3,100cal mole-1) at the meltingpoint (I ,3S6°K)
are required, together with the appropriate heat-capacity-changerelations for
the oxidation of both solid copper:
..:1C• = O·S0+2·2x I0-3T+0·20x IO'T-2caldeg-1molti-1
and liquid copper:
L1C,, = -3·68+S·20x 1Q-JT+0·20x 10sr-2ca1c1eg-1 molc-1•
Equation 5.2.I is strictly applicable only at temperatures above the
i
melting point of pure copper and the calculation of LlH1,323.involves the
assumption that the CP data for Cuc,, are valid down to 1,323°K. How-
ever, as the temperature range of the extrapolation is small, it is reason-
able to assume that the error incurred is not large. Inserting T- 1,323°K
in equation 5.2.1 yields:
-, '
LlH,,323 = -43,830 cal.
Two mole of liquid copper are oxidized in the reaction, so the
heat evolved by blowing oxygen through an ideal gold-copper alloy at
l,050°C is:
-1 f 43,830
43,8 30 ca I ( 2 mole ) o copper = x

~ .. ~ . . . ;
·,. ;• = 345 cal g-1 of copper.
5.2 Calculate the heat evolved per gram of copper oxidized when pure
oxygen is blown.through the bath described bf the prev.io~ example. 5.3 If the liquid gold-copper alloy discussed in the previous examples is
assumed to behave regularly rather than ideally, the activity coefficient of
. When oxygen is blown through the liquid gold1opper alloy at copper in the alloy being equal to 0.41 at l,050°C, calculate the same
; l,()50~C. the copper alone will be oxidized. As it'_is "8&ume4 that the quantity as calculated in example 5.2.

f/:'tJi/:>}:.·:,.. :; .
96 SOLUTIQNS
. ::;. ,.:.- ·.',-·t ·•, ,• THE GIBBS-DUHEM RELATION 97
When it is assumed that t,he liquid . gold-copper alloy is a regular
5.4 From ni;~r~h~nt o/:the.. temperature coefficient of the e.m.f. of
solution, the heat of mixing of copper is not zero (as for an ideal solution)
galvanic cells contai,,ing liquid gold-lead alloy electrodes, the following
but is given, from 1.82, by: Tc/
values of the [JIJf'_tial q,o.{ar entropy ofmixing of lead.Juwe been dttermined
B~ _; RTlnyea. at 830°Cfor vqnoustcompositions.
. ~- · ' ..
(,

Substitution of the values given for T and Yea yields:

ll/1, - -2,350~mole-t.
~~le fracti~~ f:1~~\1 ~JO.:: .~·20 0·30 0·40 0-SO

(As the solution is regular ~s beat of mixir,g, equal to the excess partial Partial molar ~.iitropy'·-.,.-· ---------------
molar free energy of copper, is independent of temperature (sec p. 15).) of mixing 'oflea'.d·, · ·· 8·S0 HX) 3·28 2·20 .1·60
Thus, compared to the previous example, when l mole of copper is
s~ (cal mole-:-~)deg-1)
. . ~.- . :, '·. ~· ' . .·.
oxidized by blowing oxygen through the alloy bath, an additional heat
effect arises from the removal of copper from the solution. This heat Mole
AT
"n
fractjon
.
~--f!~d.
"
,· -
O·~
.
I · 0·70 . 0·80 0-90
effect is the reverse of that occurring on solution of 1 mole of copper in
the alloy, i.e. the reverse of the partial molar heat of mixing of copper at
the temperature of oxidation. Thus a quantity of heat:
Partial molar entropy
of mixing .of lead, , .
St:, (cal mole-I deg-,!) .
11 ·08 . . 0- 71 0·42 . 0- 18

+ 2,350 cal mole-I

or: Calculate the partial ~olar entro/JY~!gold (S A..J in ~.40 atom-per cent Au
+ 37 ca! g-1 of copper ( positive heat-endothermic) alloy at this temperature. · ·· · · ·
~.
must be added to the value of the heat of the oxidation reaction for pure The appropriate Gibbs-Duhem relationis:
copper as calculated in example 5.2. As this heat of oxidation is (-345
cal s'). the tota.l heat evolved by blowing oxygen through the regular i · NPbdSt{,+NAudS1{. • 0
gok'-copper alloy at l,050°C is therefore: l
+ 345-
The Clbbs-Duhem relation
37 ••• 308 cal g-1 of copper. .J
. r.
or, expressed as an indefinite integral: .

\l J d sAu -M
- J NPb S"'
N Aud Pb• , · 5.4.1
In the thermodynamic analysis of solutions, it is very commonly the
situation that data are more readily measurable for one component than
for the other components. Because of this, the Gibbs-Duhem relation,
·t
I
The limits of integration may be chosen as desired. The upper limit will
obviously be that corresponding to the 40 atom-per cent Au alloy. The
lower limit may in theory be chosen as any value, but to be of use, this
derived in terms of the partial molar free energy in /.D, is of great limit must be such as to yield a known value of the entropy of mixing of
importance, for it relates the thermodynamic quantities for the various gold on integration of the left-hand side of 5.4.1. Thus the most con-
components in the solution. It will be recalled that this relation may be venient lower integration limit will correspond to pure gold (NAu - 1).
written in terms of any of the partial molar state variables of the system. For then 5.4.1 becomes:
I

I.
As will be shown, however, its use in the form of 1.43 is rather incon- N.•••.o,,co
I
venient and inaccurate, and modified versions are more c: mmonly used I l'M SM . SM . . N.PbdSM 5.4.2
Au(purc) = Au(N,..-0·40) =.
NAo . Pb
..)Au(N ••• -0·40)-
to obtained values for components other than the one for which experi- I
I
N,..-1
rncnt has yielded data. The following examples illustrate the application I '
as ..)r<M ·o . ".;·
Au (pure) = • ·
of various forms of the Gibbs-Duhem equation to binary solutions. The I The integral in sAv2, can. be-solved graphically as the area under a
extension of these principles to three-component (ternary) solutions is curve of Npb/NAu ver{us St:, between the values of Sf: corresponding to
(
discussed in chapter 8.
I
NAu7"0·40(Npi,/NAui:
. . ',- .
~ i:50) :and':N>.u""
. .' .
l(Np-JN;.u•O). The curve

! J [_:
I
911 SOLUTIONS THE GIBBS-DUHEM RELATION 99
•j, .
'derived from the given data, Fig. 5.4a, is asymptotic i~· both axes and l
where Lr, the latent heat of fusion of gold at Tr (l,063°C), - 3,050 cal
moJe-1 and CPAu(I) - 7·00 cal deg-1 mole-1•
evaluation of the integral will involve quite. serious error, However, the i
shaded area in Fig: 5.4a, which corresponds to the required integration Substituting these values:
limits, yields a value of the partial molar .entropy of mixing of gold
S;.,.'1, = (SAu-Siu) = 3·58 cal deg-1 mole-1 (using the Trapezoidal rule"
with a chord width of .dSft = 0·5). This value will, of course, be too small,
I
I
I
0
SAu(l,830'C) = 11·32+ f
. 1,336
[5·66 3]
-y+ 1·24 X IO- dT+ 3,050
l,336+ f
1,103
1·0
TdT
298 1,336
but is that given by the only reasonable extrapolation of the curve from
the available data. ·
I
) =
1 l ·32+ 9·79 + 2·28-1 ·35
22·04 cal deg=! mole-I.
i Hence SAu• the partial molar entropy of gold in solution in a 40 atom
I per cent gold-lead alloy at 830°C = 22·04+ 3·58 - 25·62caldeg-1 mole=".
l The graphical integration of this form of the Gibb-Duhem relation is

I rather unsatisfactory and, as will be shown presently, it is preferable to

use other forms for which, on integration, it is not necessary to evaluate
the area under an asymptotic curve.

5.5 From vapourpressure measurements, the following values have been

determined for the activity of mercury in liquid mercury-bismuth alloys at
320°C. Calculate the activity of bismuth in a 20 atom-per cent Bi alloy at
this temperature.

~· .•, •·' I
NH, 0·949 0·893 0·851 0·753 0·653 0·537 0·437 0·330 O·'lffT 0·063
0 A'-.w 5
S Pb cal deg ~le-
~;6
··;.7·,., \,w I
aH, : 0·961 0·929 0·908 0·84{) 0·765 0·650 0·542 0·432 0·278 0·092

Fig. 5.4.a:. Graphical solution or·

binary Gibb&:-Duh$ equation. Experimental measurements very commonly yield the activity of a
•· : Solutiou of integra! in equatioi:1•~5.4._2}, / · : · .. component in a solution directly (e.g. vapour pressure measurements,
The molar entropy of pure gold (St,) it 830~C-wiil be given by the equilibrium data). The Gibbs-Duhem relation, 1.43, may readily be
integrated form of 1.26. The required heat capacity data and the standard expressed in terms of the activities of the components by substitution of
molar entropy of gold are: :. , . :; · . the relation: ·
. ' . : ' 1 ... dC1 = RTdlna1
CPAu(a) = 5·66+l ·24 X l(r,3 Taj de$':• ,~.pie-1
Siu, 291 =- 11 · 32 cal. deg-,:1 tnol~-1.'·,. ·.-!., obtained by differentiation of 1.54.
On this substitution, it is obtained that:
However, 'as the 40 atom-per cent gold-lead' alld) fo'r which SAu is
required, is liquid at 830°C, the value of Siu to be caleulated must be that
L Nidlna1 = 0.
for pure liquid gold, supercooled frbni 1',063°C (I,;~36°K) to. 830°C Applied to the binary Hg-Bi system, this becomes:
(1,103°K). Thus, following 1.27: , ., \ .. NHadlnaH1+NBidlnaB1 =0 5.5.l
. :,·· ... · .. , :.'·.·:·,r. ·. '

soAu(l, 830°C). '= soAu 298 + J

1,336
r CPAu~) dT +':I+
·
f(
· 1,10:J:
dT
c.,;AuO>
or, as an indefinite integral:

'· . Zlll,
. .T
.,
Tr
'1,336 ,
T
J dlnas; = - f ~~dlnaHa. 5.S.2
100 SOLUTIONS THE GIBBS-DU HEM RELA TJON 101

The upper integration limit will be NBt - 0·20, while the lower limit is It is apparent, as in the previous example, that the area measured is
chosen so as to yield a known value of an1 on the left-hand side of 5.5.2. too small, and this activity value is too large (the area being a negative
Since by definition, a.91-1 as Nn,_.. l, a lower integration limit of quantity). A better result is obtained if the data are expressed in terms of
N9i = 1 will yield: the activity coefficient instead of the activity, using a further version of
Naf-0-20
NHs the Gibbs-Duhem relation.
lna81(Na,-o-:.io> - - N.s1-dlnaH1 · For a binary system AB, NA+ NB= I or dNA+ dN.s = 0. Multiplying ," ·
!,
Na-I throughout by: i:
or: NA Ne
-=--=!
NA Ne. '
then:

or:

·\

Subtraction of -this rgla~ion from tile binary Gibbs-Duhem relation

.. Id s:
5 ..51 yiet \r .

t: Nai•l
~J\"Adln·yA+Niidlny8 = 0 5.5.3 ;

Shaded 0- ~ d log a+ig • -o-s.u.

~
from the definition of}h.e activity coefficient, y1 = aJN1• As an indefinite
integral, this form of \he Gibbs-Duhern equation yields, for the present
l
!
t
system, ;; . : '· .. ~--- ·
. . .. ~ f N"Na~.dlnyH,
0·20

f-/ln~e.r:=
or, integrating betweeW
r
N~/;,.
, .
r '(where ,YB!= 1). and NBI ::s: 0·20:.

0 O·I 0·2 0·3 <H 6-5 o-e

-log "Hg
07 0-8 0·9 l-o·· ·1,1
\: . ',

. . }. ·; ... ;,'· F~--1

.

ln:Ys1c1~;'.'..o'.~o) -~ ~.::' r
Naa•0-20

i~dlny~.
Bi ":J
Fig. 5.5a. Graphical aolution or binary Oibbl-Duhem equation. Solution or integral ,1 or: t:
in equation (S.S.2). ~. Nm.:.! .
The activity of bismuth in the 20 atom per cent alloy is given, therefore, I I 08Yl31c.N.,-0·20> . ==. + ···s .. N"'dl ogyHs.
-N·- .
as the antilog of the area under a curve of NHJN.s1 versus logaHs between
I
. Bl
Naa-:-0-20· ·
the limits N81 = 1, (NHJNn; = 0), and NB1-0·20, (NHsJNBi = 4). As in
the previous example, the area to be measured is under a curve tending The activity coefficient of bismuth is obtained, therefore, as the antilog
to infinity, as shown in Fig. 5.5a. Bearing the inherent errors in mind, of the area under a curve of NHJN81 versus logyHa between the limits
the shaded portion shown is of area - 0·544 units (using the Trapezoidal . Nei = l (NHJNai = 0) and Nei = 0·20 (NHJNs1 = 4).
rule, L1 logaHa chord width 0·05). This yields a value of: Values of Yl3i are calculated from the table given, by use of the definition
y1 = a;JN1• The area to be integrated is shown shaded in Fig."5.5b. It
QBl(Na,-0·20) ,.. antilog !•456 should be noted that the curve is not asymptotic to the abscissa as
- 0·286. logyH, is finite when N81 ••• I or NH.JN131 == 0. Thus the shaded portion
8
 .-q>.)~.J:c ·.~:~. ,=.

102 SOLUTIONS
THE GIBBS·DUHEM RELATION 103
of Fig, 5.5b is of area +0· 134 units (using the Trapezoidal rule,
Ll logyH11 chord width 0·01), which yields a value of: 5.6 Calculate the activity of bismuth in a 3 atom-per cent bismuth-
mercury alloy at 320°C, using the data given in example 5.5.
'Yl!l(Na,-0.20) = anti1og0·134
As demonstrated, the use of the Gibbs-Duhern relation expressed in
= 1·36. terms of activity coefficient is quite satisfactory when values of the
Hence: activity of the unknown component are required for solutions containing
OBl ~·'YB1:NB1 appreciable amounts of that component. However, when values of the
= U6x0·20 activity are required in dilute solutions of the unknown component,
difficulties are encountered due to the asymptotic nature of the curve (as
- 0·272.
in Fig. 'i.5b) with respect to the ordinate. For example, the required
evaluation of the activity of bismuth in a Bi-Hg solution at NBr - 0·03
would be difficult using equation 5.5.3 as the area to be measured under
the curve would be to a value of NHJN81 = 32·3. Activity values in dilute
solutions are very important, and this difficulty is a severe limitation on
the usefulness of 5.5.3. Fortunately, a modification of this equation has
/ been devised by Wagner [2], which removes the necessity of integrating
to a steeply rising curve.
Division of 5.5.3 by dN8 yields:

dlnyA N dln'YB _
NA. dNB + B. dNB - 0

or, as an indefinite integral:

-~•.r;. f dlnyA = - Jr NA.

NB dln'YB
dNB .dNu.
... On integration by parts, this yields:

NB
f dlnyA = - NA!nya+ J lnyp.
»i:> 5.6.1

To yield the activity of bismuth in the 3 atom per cent bismuth alloy,
O·U 0·1 0-18
relation 5.6.1 becomes:

Fig. 5.5b. Graphical solution of binary Gibbs-Duhern equation. - [NHa

logyBi(Na,-0·03)- - N-:-logyH.11
B1
]Na,-0·03
Na,-1
+
J
Naa-0-03
logyH,
-N2 dNa,. 5.6.2
Bl
This value is, of course, smaller than that obtained in the first part of Na,-1
this example, but is closer to the true value of the activity of bismuth. The integral in this equation is given by the area under a
curve of
The only errors in the value of a81 = 0·272 arise from curve drawing and JogyHJ.VJ; versus NH, which, as shown in Fig. 5.6a, is not asymptotic
the use of the Trapezoidal' rule, whereas previously a large error arose to either of the coordinates. The shaded portion between the limits
.from the impracticability of area measurement under an asymptotic Nn; = 0·03 (NH1 = 0·97) and Nn;-= l (NH1 = 0) is of area 0·539 units
(measured using the Trapezoidal rule, LlNa,chord width 0·05). The

'L
 ·:~, 'i

04 SOLUTIONS
1

i. · THE GIBliS-DUHEM,RELATION ,05

it her term in 5.6.2 requires the value of logy81 at NB1 = 0·03, which an equation vaiid for a specified temperature range, which is us~lly,
.alue is read off from the curve of Fig. 5.6a. Thus: though not always, the range in which the.experimental data have been
obtained, In such cases, the Gibbs_:Duhem relation may be integrated
analytically instead of graphically. This is illustrated in the following
example.
>r:
NH, )
(NBl log '>'Ha Na,-0-03 - 0·064 26
2·•
ind:
n
( ~"'logy.Ha)
Bl N •• -t
•• o. 2·0
0,

1·8
lence: 1·6
)og'YBl(Nat-0-03) = -0-064+0·539 tfl·•
•••• 0·475 1
~ 1·2
.r : ro
0-8
'YBl(Na,-0-03) ,. 2·99,
C>6,
fence:
Ou; (N.,-0-()J) = 2•99 X 0·03
= 0·0897 as required.

It will be noted that this value is not particularly sensitive to the exact :lf. .'!
.
-osition of the curve of Fig. 5.6a in the region 0·9 < NHa < l, where the Fig. !5.6a. Graphical solutjon of binary Gibbe-Duhem CQUJllion. Solution
. of the intcara} in equation (5.6.2).
tatter of the experimental points necessitates a rather arbitrary fixing S.7 From e.m.f. measurerilenrs,·the activity coefficient ·of zinc in liquid
.f the position of the curve.
cadmium-zinc alloys .at 435fC
1, has
. been shown to conform to the relation: .
Using this analysis, but integrating 5.6.J up to NB;= 0·20 only, yields:
· In Yz.f= 0·87 Nt.i-0·30Nbi.
j· •
log i'Bi (Na,-0·20) - -0· 163 +0•282
Calculate the activity of caPmium in i:i 30 atom-per cent Cd alloy at this
- 0·119 temperature. · · t · ~"'f •
,r:
The appropriate form of~the Gibbs-Duhem relation is:
YBl(Na,-0·20) a: 1·315.
fence: N2n d (µ r:Z.:+Nc.i d ln Ye.;= 0
or, as a definite integral: f.'': ·
/hich compares well with the value of auicN •• -o-20) = 0·272, previously
brained. · ln.Yc.s
,· . J
\ , w ••-0-1

'7" '.""' · ·
N2n
-·-dlnyz..
..
5.7.J
. .· NCd ••
If the experimental data are adequate, it is sometimes possible to ,:
N._.o
x press the thermodynamic behaviour of a solution in terms of empirical From the relation given.].
q uations relating the various thermodynamic parameters with compo-
lnyz,}'=
\r
o:s1N&:..;o,30Nbi.
..
uion. Most commonly, the activity coefficient of a component of a then: ' . :
olution is expressed as a function of composition and temperature by dlny,..A =]1·74NCd.:..0,90Nt.t]d.NCd.

o I\
 EXCESS THERMODYNAMIC QUANTITIES 107
106 SOLUTIONS
Hence, for a binary solution described by a relation for component A of
Substituting in 5.7..J and replacing N7.a by (1-Nc.i), cancelling where
this type, there is no need to use the Gibbs-Duhem equation to obtain
possible and expressing the integration 1,imi.ts in terms of. the component
Cd gives: · ' . (C1 · data for component B.
As noted in chapter : . p. 15, it may be; shown that for a binary regular
Noo-0-3
lnyCd - - J (1·74-2·64NCd+0-90N&)~NCd 5.7.2 solution, defined by 1.80, the functions
Noa-I In y ") ( mRTInNi'YB)
\
which rruty be integrated.to yield:
a.A
( =RT N~ and (XB

\
lnya"" 0-310 are equal and constant and that the quantity RTlny1 is independent of
or: temperature. Thus activity coefficient data for regular solutions are
'Yos - 1 ·363. expressible by relations of the form of 1.84 and the observations just
Thus: made will be applicable to such solutions. However, it should be empha-
aCd<Nc.-o-J) - 1·363 x0-3 sized again that the existence of such a functional relation for activity
- 0-409. coefficient data for a solution does not; of itself, imply that the solution
is regular.
This method is usually the simplest way of analytically integrating the The following example demonstrates the ease with which data for the
Gibbs-Duhem equation. That is, the integral is expressed in terms of the unknown component may be deduced when the known activity coeffi-
unknown component (Cd in this example) and the integration limits are cient data conforms to this functional relation.
changed accordingly; for then the denominator in the' integrand ( e.g.
N01 in 5.7.1) may be eliminated by cancellation, as in: 5.7.2. The alterna- 5.8 From extrapolated data for the solid alloys, the activity cot!jficient
tive, of expressing the integral in terms of the known compor ent, leads of aluminium in liquid Al-Zn alloys, relative to liquid aluminium as the
to an expression with ( 1- N1cnown compooenJ in the denominator of the standard state, is given by:
integrand which, in general, then has to be integrated by parts. RTlnyAl(O = l,750(1-NAt)2•
It is commonly found that activity coefficient data for a solution are
expressed by a relation of the form: Calculate the activity of zinc at 550°C (i.e. just above the liquidus) in the
equi-atomic solution.
RTln y A - «Ni
From the relation given, it is apparent that it is unnecessary to use the
where oc is a constant. Such a relation leads to a similar expression for the
Gibbs-Duhem relation. Thus:
activity coefficient of component B of the binary solution. For, from this
relation: RTlny"' = RT!nyz.n = 1 750
RTdlnyA ••• -2oc(l-N,.)dNA Nbi NL ' .
Hence:
which on substitution in the integrated form of 5.5.3: 'YZn(Naa-0·$) Cl 1•308
and: Oz,, (Naa-0·$) "'" 0·654.
ln')'B = -
. f NA
-dln')'A
NB
Excess thermodynamic quantities
yields:
1 n 'YB
•• 2oc
RT
f ·
NA(l-N,.)dN ·
NB • " ::
5.9 From e.m.f. measurements, the partial molar free energy of mixing
of silver in a liquid gold-silver alloy containing 80 atom-per cent Ag is

_.__._
2a
RT
Ni
2
;
·
calculated to be - 700 cal mole=! at l,085°C, relative to liquid silver as the
standard state. Assuming that gold-silver alloys behave regularly at this
temperature, calculate the excess integral molar free energy of the solution
or: of this composition
RTln YB - ··ocNi,.
\~
•• .•••••'!'~-
.
 f;._.·
~--
'. ··~i.·! ' ·. . .
!08 SOLUTIONS , EXCE S5 THERMODYNAMIC
0 QUANTITIES 109
,i· .

In addition to involving the use of excess quantities, this example Substituting the given valiies: · ·
'urther illustrates the considerable simplification of thermodynamic
: .'\/ ../' '·',\:·.(0·80)2
.alculations in systems which adhere to regular solution behaviour. In (),A~·=:---97 X -:
oarticular, such behaviour enables quantities to be calculated from much
'• r. .
. · · ·. 0·20 . ' '·. r ••

nore limited experimental data than is possible for a non-regular (non- = . .;:, 1,552 cal mole-I.
to
deal) solution. The single quantity a,""' -
700 cal mole-1 in a binary From 1.42 applied
0

ef~ss free eti~igies in the Au-Ag system, the

.olution N Ai= 0·80, would in general be inadequate as a basis for the
excess integral molar f~ee ~nergy of Ole ~oi.utionis given by:
.alculation of any integral quantity, since the latter requires thermo-
lynamic data for both components of a binary system. However, with . GE = N GE + N· Aa GE
. A11 A,u .Aa
he assumption of regular solution behaviour, this single quantity for one or:
:omponent may be used to calculate data for the other component. For, GfN,..-o-a/= 0·2x(-'i,S52)+0:8(-97)
rorn 1.75, the excess partial molar free energy of a component i of a = - 388 cal mole-\ of solution.
.olution is given by:
5.10 At 705°K, the available experimental data for integral heats and
entropies of mixing of the non-ideal, non-regular liquid systems Cd-Pb,
From the definition _of quantities of mixing, this is equivalent to: Cd-Sn and Cd-Zn (designated 1-2), may be described by empirical
equations of theform:
Cf= (Gt'+G?)-(G{"+Gf)1c1ea1.
Hence: HM= AN1N2+BN2IogN2
Cf= G{"-C{"W•I SM= -C(N1logN1+N2logN~
= (:f' - RTin N1 from 1.67. where the valuesfor the constants A, B and C for the three alloy systems
This relation is completely general and applicable to a component of any are listed below. Compare the deviations of these systems from ideal
behaviour in terms of the values of the excess partial molar free energy of
solution.
cadmium In the equi-atomic solutions.
From the data given for the Au-Ag system at l,085°C (1,358°K):
System A B C
GI11(N .•• -0·8) = -700-RTinN,..
Cd-Pb + 1,430 -1,660 · ·+4·8
-= -700+603
·Cd-Sn· -1,290
'. +880 +5·3
= -97 cal mole+'. ca-ze>. ,, 1 600 . -830 +4·6
However, if a system is assumed to behave regularly, the function «. t:7' 1 '- • ':.
·,
defined in J .84 as: This example demonstrates how experimental thermodynamic data,
RTlny1 expressed in this iIJstanceibfmeans 'of ~~pirical functions of composi-
«1-= (l -NJ2 tion, may be used to derije .the quantities most commonly employed to
describe the behaviour ot:inon-ideal; non-regular solutions. '
may be assumed to be a constant for the components of the solution (cf. Using 1.63 applied to iptcgral, quantities of mixing:
previous example). Thus, for a regular Au-Az alloy: /· ··. ·' ,'
.:,GM= HM··TSM
't{"
RTln YAu RTln y Af
N,...
2 -= 2
NAU
the integral frec energy o( nuxirig of the alloy systems at 705°K may be
deduced, as a function i;,(1composition,·from the relations given for. the
or, from 1.79:
enthalpy and entropy. Th;a~:;, '•,:
Gfu C~
Nfs,.. Ni~· G"' • AN1N2+&05CN1lb~N'1+(B+705C)N2logN2, S.10.I
r;. .
,.
 I
I
I

110 SOLUTIONS
i INTERRELATION OF QUANTITIES 111

The partial molar free energy of mixing (Gi") for comP9nent I can be The magnitudes of these positive excess partial molar free energies show
obtained from 'this, using 1.47 applied to free energies of mixing: immediately the extent of the positive deviations of the solutions from
' .
:dGM
G{'I""' GM+(l-N1) dN •
.
. :

l
1·
'!
·i 5.10.2 I
I
ideal behaviour. Thus the Cd-Sn system is more ideal than the Cd-Zn,
which in tum deviates less from ideal behaviour than does the 'Cd-Pb
system.
Differentiating 5.10.1 with respect to N1: As would be expected from 1.79, the activity coefficients have values
which also indicate the respective order of the deviations of these systems
from ideality. However, because of the logarithmic relation of 1.79, the
relative magnitudes of the deviations from ideality, as indicated by the
. 5.10.3 activity coefficients and the excess partial molar free energies of cadmium,
Substituting 5.10.1 and 5.10.3 into 5.10.1 yieldsr are markedly different. Although the activity coefficient is the quantity
employed as a measure of the deviation from ideality in, for example,
Gt'=. ANi+705ClogN1 -0·434BN2• equilibrium calculations, the excess partial molar free energy is the more
Thus for N1 = N2 = 0·5, useful criterion when considering the causes for the non-ideal behaviour
of a system.
Gt!i = 0·25A-212·IC-0·217B
The interrelation of quantities In solution thermodynamics
- 300 cal moJe-1 for Cd-Pb system }
Several of the previous examples in this chapter have illustrated the use
- 625 cal mole-1 for Cd-Sn system at N01 = 0·5. of the various quantities whica describe the thermodynamic behaviour
- 395 cal moJe-1 for Cd-Zn system of ideal, regular and general solutions. However, when teaching thermo-
dynamics, it has been the experience of the authors that students com-
From 1.54: monly confuse the various quantities involved in problems of this type.
Cf" = RTin a1 For example, some students find difficulty in distinguishing between
and hence: excess and relative molar quantities. With this in mind, it has been found
acd = 0·806 for Cd-Pb l r-- . . . useful for the student to consider problems of an algebraic. nature
involving these quantities. By this means, he must understand and

·. . =
= 0·640 for Cd-Sn ~ at Ned
J
= 0·5
/
. ~
interrelate the quantities without the 'distracting' requirement of slide-
0·7~5 for Cd-Zn I . • I rule manipulation. The last two examples in this chapter are of this
and:

·~:~~·;: ~=:: '..}·~t;Cd. . '::o! algebraic type, illustrating the relations between the functions involved in

,< ,,I , . •...•.•

.
i the description of the thermodynamics of solutions.
YCd: -·0·5....
. \: I 5.11 Describe how to calculate the value of the integral molar entropy of
1·51 for Cd-Zn . ;. · . '
a non-ideal, non-regular solution AB (containing nA mole of A and nit mole
However, from 1.79:

Hence:
I -
,
G !S -"' RT1n..,.r,• , ' . ·.'· .)
·
'L ' ·' ·
:,,

1·,
.
· ·.- /
,· ·t
of B), assuming that thefollowing data have been determined al a constant
temperature T:
(a) The partial molar enthalpy of component A over a range of composi-
G& = + 667 cal moJe-1 for Cd-Pb }· . -·-.· ).
tion from pure B up to and including the composition of the solution.
(b) The activity of component A over the same range of composition.
+ 346 cal mole-' for Cd-Sn .· at lVc, "'; 0·5. (c) The standard molar free energies of components A and B and the
+ 578 cal moie-1 for Cd-Zn '
..,~
1. standard molar enthalpy of component 'B.

I 9'(
:.

I
t 12 SOLUTIONS
( ¥ 1 'Il!ffERRELATION OF QUANTITIES 113
If these data were not available, hut it could be assumed that the solution ,-
t Similarly, from 5./ J,j and the known value of Hg in conjunction with
behaved regularly, what single state variable of the pure components would •
the Gibbs-Duhem relatiori applied topanial molar 'enthalpies:
he needed to deduce the integral molar entropy?
It should be noted that the integral molar entropy and not the integral
molar entropy of'rnixing of the solution is required. The activity data for
I
[
. ,· ... '
· f
!-':. . . . .
J
. '0
_ f.B_-1!~,= - .
·
'

.
x.-
N;.
.
,NB dll;.
component A yield values of the partial molar free energy of mixing, ! 0 •'

which must then be combined with the available standard free energy of the partial molar enthai°py of B atthe composition X A of the solution
A to obtain the partial molar free energy of component A as a function may be calculated; Th~ i~tegral molar enthalpy H of the solution is
of composition in the solution. These data can then be used to calculate derived from 1.42 appliJd't9 enthalpies: ·. . .
corresponding data for component B and hence the integral molar free
energy of the solution. Similar manipulation of the partial molar
fJ'f~: N/ll~+ Ne s;
enthalpy data yields the integral molar enthalpy. The integral molar The i:eqajred quant:ti:. the ·int~gra.l molar, entropy of the solution, is
entropy can then be obtained by use of the definitional relation, 1.31. then obtained from: :.·: · · · ·
In addition to the known quantities, ~. Gg and Hg, the available data . • .. ·H..;.,Q
may be expressed by: S' ... :.T .•
---
RA = fi(N,.) 5.JJ.l
If tho solution is regular (and n~ri.~·ofthe previous data is avaiijtble),
a;. =flN,J 5.11.2 then, from 1.80 for a regular solution AB, the partial molar entropies of
mixing are given by: "'. · ·· · ' ··' · ·· ·
these relations being known over the range of composition O ~ NA ~ X;.,
where: . S~ = S;.-S! = -RlriN;. }
nA M o . 5.lJ,,1
and: SB ;:: Se-Se = -RlnNB
X;.-= n;.+ne'
From 1.54: where Si and sg are the standard molar entropies of the pure com-
G;.-61 = RTlna;., ponents A and Bat the temperature T. Hence if these latter quantities
are known, the partial molar entropies of A and B in the solution may be
and, from the known value of~ and 5.11.2, the partial molar free energy calculated from 5.11.4. Using these values the required quantity, the
of A may be calculated. That is: integral molar entropy of the solution, may be calculated from 1.42
GA = fiN ,.). 5.11.3
applied to entropies:

However, from the Gibbs-Duhem relation 1.43:

NAdG;.+N8d0» - 0 5.12 The integral molar entropy of a non-ideal, non-regular liquid
or: solution A-B is known as a completefunction of composition (0 ~ NA ~ 1)
x,.
G8-Gi = -J.
0
N;.dG;.
Ne
at a temperature T, together with the free '!nergy_of the pure component A
and the activity of A in a. solution containing X A atom fraction of that
component. Show how to calculate the partial molar enthalpy of A and the
excess partial molar ent;P,PY of Bin the solution XA at the temperature T.
and by insertion of the differential of 5.11.3 and the value of Gg, G8 may
be calculated at the composition X;. of the solution. The integral molar The available data m~1 be represented by:
free energy G of the solution is then derived using 1.42:
~ <:> S .:= /j(NiJ 5.12.1
G - N;.G;.+NsC., with 61 and a;. i.nthe sJlution X;..
I .
o/'.·
 ''ii. l:• 'j,,,;,, l'f ·!"-:&-..;;,. '..'.;,·· ;- l.' I. :1.~ '1:;~ •. , .t:i.;, >~':,. ·, 'l

114 SOLUTIONS EXERCISES ns

The first step is to calculate the partial ~~i~/entrJ~ies of the com-
ponents in the solution X A• From 1.46 and 1.47,' partial.molar quantities
may be found for any solution (either: graphically or analytically) if the
integral quantity is known as a function of composition. Thus for the
partial molar entropies:
[ SA ••• s+NA0 a:JJN.• -x .•
Exercises
5.A Liquid tin-bismuth alloys are approximately ideal at l,000°C, at
which temperature the vapour pressures of the pure components are
~- l ·66 x I0-4 and I·910 mm Hg respectively. If the latent heats of'vaporiza-
tion of pure tin and bismuth are taken to be 68·20 and 44·53 kcal mole-1
respectively at this temperature, calculate the latent heat of vaporization
of an alloy containing 3 atom-per cent bismuth at the same temperature.

[ SB= S-(1-N,.)-
· · asj 5.B An inert gas containing O· 3 per cent by volume of sulphur vapour
(S2) is pumped at 2 atmosphere pressure through a liquid iron-man-
oN
-x .• N ••
ganese alloy containing 30 atom-per cent manganese at l,600°C. Under
by means of which SA and SB in the present example may be calculated these conditions, the sulphur reacts only with the manganese, with a heat
analytically. of reaction of -63· 1 kcal mole"! of MnS formed. If it is assumed that
From 1.54:' Fe-Mn alloys behave ideally at this temperature, calculate the heat
CA= ~ +RTinaA evolved in the bath per thousand litres of gas blown through. What is the
and hence, from the available values of ~ and aA (in the solution of heat effect if the alloy is assumed to behave regularly, rather than ideally;
composition X ,J, the partial molar free energy of c< rnponent A in the the activity coefficient of manganese being taken as l · 14?
solution X A may be calculated. 5.c The latent heat of vaporization of manganese is 53·7 kcal mole=!
Then, from 1}1 applied to partial molar quantities: at the boiling point, 2,095°C. Assuming that this value does not change
nA = cA+rsA significantly with temperature, determine the vapour pressure of
manganese at 1,600°C.
the partial molar entbalpy.of'component Ain the solution XAis found, as Assuming that manganese behaves ideally when dissolved in a liquid
required. · .
steel, determine whether or not a steel containing l ·O weight-per cent
The excess partial molar entropy of component };3 in the solution XA manganese would lose any of this element when exposed to a vacuum
is given, following 1.75, by: · · ·
of O· l mm Hg pressure at l,600°C.
Sfi = SB-'-.)~
5.o An addition of 40 g of pure chromium: s added to a bath of I kg of
= SB-(St<+Sii°)id•! 5.12.2 pure iron at I ,600°C. If the densities of Cr and Fe are 6· l and 6·9 g cc-1
where St'ldear is the partial molar entropy of mixing of component B in respectively at this temperature, and liquid ferrochrome is assumed to
a hypothetical ideal solution of composition XA. : . · ;; behave ideally, what is the density of the resulting alloy?
But from 1.72: .\
s:flda.l = -RinNB: 5.E From measurements of the e.m.f. and its temperature variation for
galvanic cells with liquid alu.ninium anodes and liquid aluminium-
-Rln(l-X>.) bismuth alloy cathodes, the following smoothed values of the partial
and this quantity may be calculated directly. In addition, S~, the molar molar enthalpy of mixing of aluminium have been determined at 900°C
entropy of pure component B is given by 5.12.1 when NA= 0 (i.e. at for alloys containing up to 26 atom-per cent aluminium:
pure B) and may also be calculated. ,
Hence, with the value of SB deduced previously, all.the quantities on Atom-
pcrcentAl I 1 2 4 6 8 10 12
·. the right-hand side of 5.12.2 are known and the value of the excess
.· . partial molar entropy of component B in the solution XA may be R M
kcal
Al• mole? I 5·76
I
S·S3 5·23 5·00 · 4·80 4·63 4-47
. ' .calculated. t
?(!_
lll'i SO!..UTIONS
EXERCISES 117
Atom- / 14 16 18 20 22 24 26 S.1 From e.m.f. measurements, the. partial molar entropy of mixing of
per cent Al
1 zinc in liquid Zn-Ga alloys (relative to liquid zinc as the standard state)
R~. '4·30 4·14 3-98 Ht 3·65 3·48 3·32 has been determined at 490°C over the complete composition range.
kcal molc=!
From these data, given in the table, calculate the integral molar entropy
Calculate the partial molar heat of mixing of bismuth in the 80 atom-per of mixing of the solution containing 70 atom-per cent gallium.
cent Bi alloy at this temperature.
0·2 0·3 0·4 O·S 0·6 0·7 0·8 0-9
5.P From measurements of the equilibrium distribution of aluminium Stf,, cal
between liquid iron and silver, the activity coefficient of aluminium in I
mcle=! deg·1 I 0·32 0·645 l ·00 l ·42 • 1-88' 2·42 3·09 4·00 5·50
Fe-Al alloys (0 < N Al< 0·25) is given at l,600°C by: ·,\ . '

logyAI = 2·60NA1-1·51. 5.K From vapour P'r~ssure measurements at 324°C, the following
values of the activity c~efficient of mercury in mercury-tin amalgams
Calculate the activity of iron in the 80 atom-per cent Fe alloy at this may be calculated: ·
temperature.
NH.·~ 0·3. 0·4 0·5
5.o The activity coefficient of sodium in liquid Hg-Na alloys at 25°C
conforms to the relation: YHs ~. I ·58 MO 1·44 1 ·38

IogyN. = -12·81+15·6INNa+7·53Ni. Assuming thatthis syst~in is regular, calculate the integral free energy of
referred to pure liquid sodium as the standard state. Calculate the mixing of the 30 atom-per cent Hg solution.
activity of mercury in a 90 atom-per cent Hg alloy.
, 5.L If liquid Fe-Al alldys at l,6Q0°C conform to the relation:
5.H From e.m.f. measurements, the partial molar free energy of mixing Ci1 = -/I2,900+22,250NAi, cal mole-1,
!t
of magnesium in liquid Mg-Pb alloys at 560°C has been calculated at
I ' > • •

various compositions, as shown in the table. and the vapour pressurtj(in mm Hg) of pure molten iron is given by:
'
.':

. ;. 20 150 . . ·
NPb I 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 logppe =;= .-~r7[.1·27 logT+ 13·98
I
C::1 •. kcal mole"! i -0·8 -l·S3 -2·75 -3·75 -4·52 -5·50 -6·70 -8·10
calculate the vapour pressure onr~~· Qve/an allo~ containing 80 atom-
Calculate the excess integral molar free energy of the equi-atomic per cent Fe at 1 600°C. f ·• · . . '''. · · · . , _ ·
, ' ':. -.. , '
solution.
- 5.M Liquid brasses conform to the.relation:
5.1 The volume changes on alloying liquid zinc and liquid tin to form RTln YZn = --:- 5,po()N6,
liquid alloys at 420°C are given in the table. From these data, calculate
the partial molar volume of mixing of tin in an alloy containing 40 atom- and the vapour pressureof pure copper is given by:
per cent zinc at the same temperature'. 17,520
logp(mm Hg) = - -T- l ·21 log T + I 3·21.
NZA ! o-IO 0·20 (}30 0·40 O·SO
Calculate the vapour pressure of copper over a 60:40 brass at I,200°C.
V,.,, cc mole:" 1 +0·0539 0·0964 0-1274 0·1542 0·1763
5.N Liquid copper-silver alloys are found to be regular at l,150°C, at
N'UI I
I
0·60 0·70 (}80 (}90
which temperature the partial molar heat of mixing of copper is as given,
v.v. cc mole=! I 0·1888 (}1779 (} 1441 0-0890 for various compositions, in· the table. Derive an expression for the
9 . . ' '
118 SOLUTIONS EXERCISES 119
variation of the activity coefficient of silver with composition at this 5.s At 700°C, the integral heats of mixing of the liquid Cu-Cd system
temperature. arc found to conform to the relation:

(1·6 0·7 HM= 200NcuN01, cal moJe-1,

0·2 0·3 0·4 0·5
I
the partial molar entropy of mixing of cadmium being 1 · 142 cal mole-1
Rt;.. cal mole"! ; 40 145 325 560 825 1,160 1,590
deg-1 in the equi-atomic solution and the free energy of pure liquid
5.o The thermodynamic behaviour of ct brasses at .298°K may be cadmium being - 15· 3 kcal mole-1 at this temperature. Calculate the
described by the relations: activity coefficient and the partial molar free energy of cadmium in the
equi-atomic solution.
fl .ff, = - 5,350Ntu..,..2,~30N ce- cal mole;1
S.J; = -3·71ogNz,,, cal molet ' deg-1., '
5.T The system AB behaves regularly, the activity of component A in a
solution containing X A mole fraction of A being (aA.)1 at a temperature
Derive a relation between the composition andthe.excess integral molar T1• Show how the integral heat of formation of the solution at a tempera-
free energy of the solution. · ture T2 ( # T1) may be calculated.
5.P At 930°C, Fe-Ni austenites arc found to beha*-. regularly, the 5.u At a temperature T, a volatile component B forms an extensive
integral enthalpy of mixing being given by: · .i' terminal solid solution with component A. Measurements of the vapour
HM= - l,300NF.NN;, cal mole:". 1· pressure of pure Band over solutions of various compositions are readily
. ' made. Indicate how the non-regular behaviour of these solutions may be
Calculate the excess partial molar free energy of iron in)he 30 atom-per
deduced from such measurements, and how to calculate the composition
cent Ni solution. } dependence of the excess integral free energy of the solution AB at the
:~
temperature T.
s.o Impure liquid lead, containing copper in solutiQn, is refined by
adding lead sulphide to the bath. The reaction which occurs may be
References
represented by:
1. J. H. HILDEBRAND, J. Am. Chem. Soc., 51, 1929, p. 66.
PbScl) + 2Cll(n ~ Cu2S(I) ~ ~bc1),
the standard free-energy change being -17·5 k'cal at 1.5.:50°C. Assuming
I
- I-
I
2. c. w AGNER, Thermodynamicsof Alloys, Addison Wesley, 19S2.

that the sulphides are immiscible and that at low concentrations, copper
forms a regular solution in lead, described by: ·

'
Bt'., = 2,700N°fi.o, cal moler',
'

calculate the extent to which the lead can be purified at 1,550°C.

5.R When more than about 2 weight-per cent of nickel is dissolved in

iron, a continuous series of solid· solutions is formed at 840°C. The
dissolved iron is more readily oxidized than the nickel, and the FeO
formed is insoluble in the alloy. An atmosphere of 57·5 per cent H2 and
42·5 per cent H20 by volume is in equilibrium at I atmosphere pressure
with an Fe-Ni alloy containing 85·4 atom-per cent Fe at this temperature.
What is the value of the excess partial molar free energy of iron in the
~·lloy? : '
\\ '· ·. . .
CHAPTER 6 DILUTE SOLUTIONS 'ANP ALTERNATIVE STANDARD STiTES Ill

l 6.1 The melting poinr.~f pure titanium _isJ ,660°C 'andthe h~at of fusion
l
is 4,500 cal moie-1• The activity coefficient of titanium in iron at infinite
i dilution, relative to pure solid titaniu. 1 as th/standard state (i.e. y°} is
Dilute Solutions and I
iI
0·011 at l,623°C. The atomic weights , . titanium and iron are respectively
47·90 and 55·85. Calculate the change in the free energy when the standard
Aiternative Standard States • I
I state is transferred from pure solid titanium to the infinitely dilute, weight-
per cent solution of titanium in iron at 1,623°C.
i The free-energy change at 1,623°C (1,900°K) accompanying the
' change of standard state from pure liquid titanium to an infinitely dilute,
weight-per cent solution of titanium in liquid iron is obtained by use of
equation 1.97:
As indicated in chapter l, when a solute is present in a solution at
6.1.1
relatively low concentration it is sometimes more convenient to refer the
activity of the solute to an infinitely dilute standard state, in preference LI G0 = C? X, dilute) - G?puro)
to the pure substance standard state. For example, in many solutions of
_ J O ( MF• )
interest, the pure solute exists in a different physical state to that in which - RT ny . Mr1· i,00
it appears in the solution. In such cases, it is obviously more physically
realistic to refer the solute to an infinitely dilute solution rather than to where MF. and Mr1 are the atomic weights of iron and titanium. Substi-
the pure solute as the standard state. From the use of such alternative. tuting the given valuesrin this relation:
. ;, ... '
standard states, the determination of the activity of a solute in a dilute,
LIG0 ~ 'ts15 x f9b01 0·011 ( 55·SS )
multicomponent solution becomes possible by combination of the ) ' og 47·90 X J()()
t .
appropriate interaction parameters measured on simple binary and = + 33,830 cal mole-1.
ternary alloys containing the solute.
The equilibrium constant for a reaction can be expressed in terms of This free-energy change applies to the dilution of pure liquid titanium.
activities referred to a number of standard states, and can then be However, pure titaniuin is solid at l,623°C and hence the free-energy
I
manipulated in'a manner similar to those demonstrated in the examples I change accompanying t-he fusion of titanium must be taken into account.
I
of chapter 4. The only requirement for such manipulation is that the I The heat of fusion is}given as 4·,soo cal mole-! at the melting point
I
standard free-energy change for the reaction, with which the equilibrium i
)
(1,933°K), from which;ih~-~ntropy offusion is obtained by use of I.JOO:
constant is related through 1.58, is also expressed in terms of the ;~ Lr .
respective standard states chosen for each of the components of \he ~,.=.Tr.·.'..
reaction. The most common method of calculation of the free-energy
change accompanying a change of standard state of a solute is demon-
strated in example 6.1. The combination of such free-energy changes
I
j
I
i· . 4, soo ',
'-i-'·,,
t
V\{

7._ l,933: '

with that for a reaction in which all the components are referred to I =;= 2·33 cal c;!eg-l mo!e-1.
their pure states, enabling the equilibrium constant for the reaction to be Assuming Sr and Lr are i~d..:Jn.d.;n.t of temperature, the free-energy
determined when one or more solutes are referred to an alternative change for the reaction: ·· · · · ·
standard state, is illustrated in example 6.4. The usual method for the
determination of interaction coefficients from experimental data is
illustrated in examples 6.3 and 6.5 and the utilization of interaction
Il is given by:
6.1.2

coefficients for the solution of an equilibrium problem in a multi- :' Gr= +4,500-2·33Tcal
component system is demonstrated in example 6.4.
l
i . + 70 cal at l ,623°C (1,900°K).
!22 DILUTE SOLUTIONS AND ALTERNATIVE STANDARD STATES DlLUTE SOLUTIONS AND ALTERNATIVE STANDARD STATES 123

The free-energy change accompanying the overall change in standard From the given data, the activity of silicon at I atom-per cent silicon,
state, that is: relative to pure silicon as the standard state, is:
T~,.pure) .,a Ti(~.:Y.. ~!lute In Pe)
Gs; (pure) = Ns1 • Yii
is then obtained by adding 6.1.J and .6.1.2;·
= 0·01 X 0•0014
L1 c;o = - 33,830 + 70
= 0·000014.
= -33,760 cal mcle=! at 1,623°C:,
' . y The activity of silicon at this concentration, relative to the infinitely
~~
6.2 From experimental measurements of the equilibriaibetween H2-H20 dilute, weight-per cent solution of silicon in iron as the standard state,
gas mixtures; pure solid silica and silicon diss~!ved in liq}id iron, the free- is given, therefore, by 1.98. That is: , •. , ' ( I'--
energy change accompanying the transfer of standa c{ state from pure '\. <. !.
1

liquid silicon to the infinitely dilute, weight-per cent so/.'il.rion' of silicon in

iron, that is: ·,;
Si(pure, l) ~ Si(wt X, diluto in Fe ')_
Substituting the calculated values of Os; (pur~), yi1 and the atomic weights:
has been evaluated as: asi (dilute) = O· 55.

L1G0 = -28,500-5·8Tcal moJe-1., This is the activity of silicon in the I atom-per cent silicon alloy on
the infinitely dilute, weight-per cent standard state scale. Thus, from the
At l ,600°C the activity coefficient of silicon in iron, relative to pure silicon definitional relation 1.92 the activity coefficient, relative to this scale, is
as the standard state, is 0·0014 at 1 atom-per cent sili1pri. Calculate the obtained by dividing the activity by the weight-per cent concentration
activity coefficient of silicon, relative to the weight-pet cent standard state,
at this concentration.
of silicon. However, the concentration or
silicon in the alloy is given as
the atom fraction, so this must be converted into weight-per cent. At
The first step is to calculate the activity of silicon referred to the dilute Ns; = O·O!:
standard state, making use of I .97 and 1.93. wt%Si
At l,600°C (1,873°K) the free-energy change accompanying the 28·09
0·01
transfer of standard state from pure liquid silicon to the dilute, weitht- wt %Si (I 00- wt %Si)
per cent solution is: --28-·09 + --5-5-.3-5--

L1G0 = -28,500-(5·8 x 1,873) where 28·09 and 55·85 are the atomic weights of silicon and iron
respectively. Thus:
= -39,3f;O cal mole-1•
weight-per cent silicon = 0·50.
The activity coefficient of silicon at infinite dilution (relative to pure
silicon as thestandard state) can be obtained, therefore, by substituting The activity coefficient of silicon, relative to the in.finitelydilute, weight-
this value and the atomic weights for silicon (28·09) and iron (55·85) per cent standard state, is then given by:
into 1.97. That is: ~ _ as; (dilute) _ 0·55
MFe·YSi 0 ) Si - -;,;t %ST - 0·50
LlGo = RT!n(Ms;~OO
or: = l·l.
. 0 ).
55·85. Ys;
._.39,360 ·= 4·575x I,873!og(28·09xf00 . 6.3 A pure iron-manganese alloy containing 0·10 weight-per cent of
. \ ' ..
Hence: manganese is brought to equilibrium at 1,335°C with a gas mixture of
hydrogen sulphide and hydrogen in the ratio PH.sfPH, - 5·70 x lCr". On
1::4 DILUTE SOLUTIONS AND ALTERNATIVE STANDARD STATES DILUTE SOLUTIONs}AND ALTERNATIVE STANDARD STATES 1~
analysis, the equilibrated alloy specimen isfound to contain 0·0032 weight- The interaction parc.\fneter;e~.fois
;;-· : -.·,.·
,., . . defined by 1.102
.
as:
per cent of sulphur. If the activity coefficient of sulphur (relative to the
infinitely dilute, weight-per cent standard stC'te) is unity in dilute binary
. Mn ·. 0. log f [;in
' es . =; --·-o --. •
Fc:-S alloys, calculate the interaction parameter for the effect of manganese .. o.wt%Mn
on the activity coefficient of sulphur at 1,335°C. Making the reasonabl~' a~s~~ptfoi, }hat ·logfsM" is a linear function of
the manganese content up to at. least the concentration present in the
For the reaction: ternary alloy.ithen: i · ' ·
6.3.1 ·
es
Mo
= Jogfr~
-. --·-
wt%Mn·
the standard free-energy change in the relevant temperature range is
given by: 0·0301
.JG'° •• -9,870+9·39Tcal -0-70
where the standard state for sulphur dissolved in iron is the infinitely r- . = :;- 0•043, , ,•• / <,:')
() ,11' ,,.A L ·1 / (- ,AJ'' ~ ' ;
- ' I/ /·J. /·
dilute, weight-per cent solution.
6.4 Calculate the residual oxygen cOJJlent of liquid iron containing
From 1.58, the equilibrium constant for reaction 6.3.1 is given by:
0· 10 weight-per cent silicon in equilibrium with solid silica at I,600°C.
9,870 9·39
logK., = + 4·575T- 4·575
The maximum solubility of oxygen in pure liquid iron is 0·23 weight-
per cent at l ,600°C, the solubility being lowered when silicon is dissolved
which at I ,335°C (1,608°K): in the iron. Because the silicon and oxygen are present as dilute solutions
= -0·718 in the liquid iron, it is preferable to solve this problem in terms of the
or: infinitely dilute, weight-per cent standard state. The following interaction
K.,1,608 = 0·191. parameters, valid at l,600°C, for the system Fe-Si-0 are available:
Assuming that the gases behave ideally, the equilibrium constant for e~i = +0·32 e8 = -0·20
this reaction can be written as: ·
e~ = ;J
-0·24 /~ = -0·14
K = PH,S The free energy of formation of pure solid silica from pure liquid
P PH, x (wt %S x fs) silicon and oxygen gas, in the. temperature range 1,700 to 1,986°K is
given by: · \ ' · ' · ·
where fs is the activity coefficient of sulphur in the ternary Fe-Mn-S
alloy. Inserting the given values for PH.sf PH, and wt %S, in the ternary Si(I) +02Cir ~ S.,.i02(1) .: .1G0 . =: '-226,500+47·50Tcal. 6.4.1
.,.
alloy:
The reaction required tb yield the answer is the formation of pure solid
fs = 0·933.
silica from silicon and ibxygeil
' \".
dissolved in liquid iron. It is necessary,
from 1.101 the activity of sulphur in the ternary alloy is equal to the therefore, to know thd/Jree energies .of solution of both oxygen and
product of ft the activity coefficient of sulphur in binary Fe-S melts silicon in iron. If the sitindard state selected for consideration of these
containing the same concentration of sulphur as the ternary alloy, and solutes is the infinitely ct'i!ute, weight-per cent solution, the free-energies
f ;:1 ", the effect of manganese on this activity coefficient. That is: of solution are equivalent to those for the transfer of standard state from
fs = fi X rr-n. the pure form, stable a( l,600°G,·to the infinitely dilute· solution. These
are given by: .. /,
But it is stated that n =- I,therefore: :r ,,, .. , .. -:· :·o ·.
· 02<,) ~ 20cf. ;~ Fei : Ll G = ;- 55,800- l·46T cal 6.4.2
fs = f~" - 0·93,;. !
Si(!) ~ Sic~ in ~~l : LI G0 = .:. 28,500- 6· l T cal. 6.4.3

9/\
·,.
 . .
126 DILUTE i'iOLUTIONS AND ALTERNATIVE S·TANDARD STATE~ DILUTE SOLUTIONS AND ALTERNATIVE STANDARD STATES !2'1
~-
Subtracting 6.4.2 and 6.4.3 from 6.4.1 yields: ·, Equation 6.4.5 can be written as:
/
Sicx1nFc)+20cXinPo) ~ Si02(,): LlCo = -142,200-,/55·06Tcal 6.4.4 (fs;- %Si).(f0. %0)2 = 2·74x 10-s
or:
or, at 1,873°K (I,600°C): logf5;+2logf0+1og %Si+2log %0 = log(2·74x 10-5) •
/ ..dG?.873 - -39,100 cal. Substituting the relations for log(5; and logf0 at 0· 1 weight-per cent
The equilibrium constant for 6.4.4 is given by: silicon, this becomes:
0·032-0·24 %0-0·40%0-0·028-1·0+2log %() = -4·562
or:
But, from 1.58: 2log %0-0·64%0 = -3·558.
This relation can be solved by plotting the expression:
4·575T
+ . 39,100 (2log %0-0·64%0) versus %0
4·575 x 1,873 at 1,873°K. and determining the value of %0 at -3·558 on the graph. The value so
Hence: obtained is 0·017. Hence 0·017 weight-per cent oxygen is in equilibrium
Kpt,873 = 36,560. with 0· I weight-per cent silicon, both dissolved in otherwise pure liquid
The liquid iron solution is stated to be in equilibrium with pure solid iron at l,600°C.
silica. This is the standard state for silica (i.e. asio, •• 1) in reaction 6.4.4 6.5 The following data have been obtained at 1,600°( for tht! carbon and
nnd, therefore: "''·' oxygen contents of liquid iron in equilibrium with carbon monoxide at
l atmosphere pressure:
{~ = as;,a2o = = 2·74x IO-~. 6.4.5
36}60 Carbon wt% 0·05 0·10 0·20 0·40 0·60 0·80
To obtain the equilibrium oxygen content of the liquid iron from this
relation, the activity coefficients for silicon and oxygen in dilute Fe-Si-0.
melts containing O· !O weight-per cent of silicon must be known. Thus,
II Oxygen wt% 0·0482 0·0245 0·0127 0·0066 0·0047 0·0038
In pure liquid I·e-C alloys at 1,600°C the activity coefficient of carbon is
using 1./01 and 1.105: I given by the relation:

Jog(s; =: logf~i+Jog{£ log~= [%C.e~] = 0·30%C.

=
,.
%Si.e:t+ %0.e§]
and, inserting the values given for the· interaction parameters, at O· l
Il Assuming that in pure liquid Fe-0 alloys the parameter
mine the values of the interaction parameters e2 and e8.
e8 is zero, deter·

weight-per cent silicon: · · The data given apply for the reaction:

log/s; = (0·1 x 0·32)+(%0 x -:-0·24) c(in Fe)+ o(i~ Fe) ~ COw

= 0·032-0·24%0. for which the equilibrium constant is:
Similarly;
logfo = %0.e8+ %Si.·e~
= (%0 X -0·20)+(0·! X -0(14) Pco
-0·20 %0 _: 0·014.. . . ·{ = wt%C.wt%0.fc.fo
.r;
·'·
 F '!.'
128 DILUTE SOLUTIONS AND ALTERN.A.TIVE STANI)ARD STATES .DILUTE SO_LUTIO·('{S,AND ALTERNATIVE STANDARD STATES 129
assuming that CO behaves as an ideal gas. The numerical value of KP is ~ • ,1' ' ,.·

I Within the concentration range to which Henry's law applies to

not given in the example but, because Pco is given as l atmosphere, then t carbon and oxygen irt liquid iron, the terms logfg and logf2 are linear
functions of the coricentrations of .oxygen and carbon respectively.
Kp.fc-fo= %Cl %_=K' 6.5.1 That is:
wt O • wt 0
lt>gff = %O.e2 6:5.5
where K' is a composition-dependent term, values of which can be t.
(:, ..... , 1·og rt: %·c·
o = o • eo c
calculated from the table given. Taking logarithms and rearranging . . ;,t ·.:\ .. ' ' I' ~
6.5.6
terms, 6.5.1 becomes: where e8 and eg arethe interaction parameters which are to be deter-
log K' - log KP = log fc + log fo. 6.5.2 mined. Further, at hi'.gh dilution, these parameters are related by 1.107:

By definition of the infinitely dilute, weight-per cent standard state, 6.5.7

re-
1o?,(c tends to zero (i.e. I) as weight-per cent carbon tends to zero.
Also, logf O ( = %0. e8 + %C. eg) will tend to zero at O per cent carbon where 16 and 12 are tbe approximat iatomic weights of oxygen and car-
if e8 = 0. Hence extrapolation of a plot of logK' versus weight-per cent bon respectively. Substituting 6.5.5i 6.5.6an·d 6.5.7 into 6.5.4, and solving
carbon to O per cent carbon yields the value of log KP. Further, at any . for e2 yields: · · · · ·. ·
chosen carbon content (from 6.5.2), the slope of the graph will be equal
to: 16%'c· · o ...: · 0·.424%C
o ... ec+fi o.·ec-
0
%0 - v o v-- 6.5.8
logfc+logfo
-----
wt%C But, from the given data, O·Ol27 weight-per cent of oxygen is in equilib-
at that particular carbon content. rium with 0·20 weight-per cent of carbon at l ,600°C. Thus:
Plotting the given data in this manner, logK' is found to bea linear 0·0127eg+0·267eg = -0·0848
function of the carbon content, \)'le slope of this line being -0·124. or:
He11ce: e2 = -0·30
logfc+logf0 = -0·124%C. 6.5.3 and, from 6.5.7:
These activity coefficients, fc and fo in the ternary alloys, can be eg = -0·40.
expressed as the products of the respective activity coefficients in the The values of eg and eS so calculated are approximately independent
Fe-C or Fe-0 binary alloys and the respective interaction coefficients of composition in the range 0·2--0·8 weight-per cent carbon, but become
(cf. equation 1.101). That is: significantly larger when the data given for the two lower carbon con-

or:

Similarly:
,F.C (ternary)
. -- Ire
C X fOC

logfc(terriaryJ= Iogfg+Iogfg.

log fo (ternary) = log f8 + log rg.

I
~:
[,
;:
centrations are substituted in 6.5.8. This is due primarily to the initial
assumption that e8 = 0. In fact, the experimental value for this parameter
is -0·20. At very low carbon concentrations the oxygen content of
liquid iron increases markedly and hence the value of log f8 (assumed
zero in 6.5.4) becomes increasingly significant. For example, using the
; data given, logf8 is calculated to be -0·00964 at 0·05 per cent carbon,
Substitution of these equalities in 6.5.3 yields: ~tf this value decr~asing .~o - 0·00254at 0·2 per cent carbon and to 0·00076
~
logf~+logfg+Iogf8+logf5 = -0·124%C ~ at 0·8 per cent cart;,pn. Thus the effect of equating logf8 to zero is
f, negligible only at rrioderate and high carbon contents.
or, inserting the given values of log ff and logfg ( = %0. e8 = 0) and
rearranging: ! A further probable source of error is the use of equation 6.5.7 to
deduce the value·of e~. There are as·'yet relatively few experimental data
log~+logf5'"" -0·424%C. 6.5.4
tI 'available with which the validity of this relation can be tested and, whilst
'.1 .;'_ •

~v
~
~
. /
" ,.•
,.
}
., ,' . ·.
 i\'!Mi~•••••••••••••••••••••
DO D!LL'.T!C: SOLUTIONS A~D ALTE.RNATIVE STANDARD STATES EXERCISES 131
. ~
the equality is thermodynamically justified at infinit~ dilution, it is 0·068. Calculate the free-energy change accompanying the transfer of
possible that the derived value of
1 '! 1' •
eg
is not accurate for more concentrated
~ . '. , I \j ". .
standard state from pure solid vanadium to the infinitely dilute, weight-
(although stLI dilute) solutions. , '.; , , per cent solution of vanadium in liquid iron.
It should be noted that when the actual value of log'fg is taken into
account, the intercept of the logK' versus weight-per cent carbon plot 6.E At 1,100°K the activity of component B in a face-centred-cubic
on the zero carbon axis is equal to logKP+ logf8co-/,caroon)· To obtain a solid solution AB is given as a function of N8, the atom fraction ofB, by:
solution by the method described here, it is then necessary to plot !ogK'
a8 = 0·24N8. (0 < N8 < 0·025)
versus (wt %C. f8), the slope of which is equal to:
where a8 is referred to pure B as the standard state. The standard free-
logf6+logns+1og r~ energy change accompanying the oxidation of pure B is - 74,000 cal
-- wt%C -·
mole "! at 1,100°K. On heating, pure B undergoes an allotropic change
The procedure for solution is then similar to that shown. from a body-centred-cubic to a face-centred-cubic structure at 1,180°K,
accompanied by a latent heat of transformation of 220 cal mole-1•
Calculate the equilibrium constant at 1,100°K for the reaction:
Exercises
B(2 atom-per cent in A)+ 02(g) ~ B02 (puresolid)
(,.A Pure hydrogen is <:_i.r.culated at ! atmosphere pressure in a closed
system over pure zirconium heated to 800°C and then over nickel, heated referred to a standard state of the infinitely dilute atom fraction solution
to the same temperature, to remove dissolved oxygen from the nickel. of Bin A.
The solubility of oxygen in nickel in equilibrium with the oxide l'-iO is
O·O 19 weight-per cent at this temperature. If oxygen dissolved in nickel 6.F The alloy system AB obeys Henry's law up to 3 weight-per cent of
conforms to Henry's law at all concentrations below the solubility limit, component B, the slope of the Henry's law line being 0· 17 at 1,000°K,
.,/
'\; /
calculate the· oxygen concentration remaining in the nickel when with respect to pure B as the standard state. On oxidation of a molten
equilibrium is attained .. alloy containing 1 ·5 weight-per cent Bat 1,000°K, the solid oxide film on
the surface of the melt is found to be pure B02• The standard free-energy
· 6.B In one refining process for lead, copper-is removed from the.lea I by change for the oxidation of pure Bis -60·0 kcal moJe-1 at 1,000°K and
reaction with lead sulphide (PbS) to· form copper sulphide (Cu2S). The pure B melts at 1,025°K with a latent heat of fusion of 2·6 kcal mole=".
sulphides are solid at the reaction temperature o(400~Cand the maxi- If the atomic weights of A and Bare 97 and 73 respectively, calculate the
mum solubility of copper in liquid lead is restricted! being only 0·40 standard free-energy change for the oxidation of B in the 1 ·5 weight-per
atom-per cent at the reaction temperature. Assuming: that, within the cent alloy at 1,000°K with respect to the infinitely dilute, weight-per cent
composition range of the homogeneous liquid solutiorr, copper in liquid standard state of B in A.
lead conforms to Henry's law, find the atom-per cent/concentration of
copper in equilibrium with liquid lead, PbS and .Cu2S at the reaction 6.G A hydrogen-steam atmosphere containing 16·67 per cent by
temperature. volume of steam is equilibrated with pure liquid iron at various tempera-
tures. On analysis of quenched samples, the iron is found to contain
6.c If liquid iron-manganese solutions can be regarded as ideal, 0·0472, 0·0597 and 0·0752 weight-per cent oxygen when equilibrated
determine the change in free energy .accornpanying the transfer of respectively at J ,550, 1,600 and 1,650°C. The activity coefficient of
standard state from pure liquid manganese to the infinitely dilute, weight- oxygen, relative to the infinitely dilute, weight-per cent standard state,
per cent solution of manganese in liquid iron. ' varies with temperature as follows:
Temperature, °C 1,550 1,600 1,650
6.D At l ,600°C, the activity coefficient of vanadium at· infinite dilution
in liquid iron, relative to pure solid vanadium as the ~tandard state, is logfo -0·23(%0] -0·20(%0} -0·18(%0]
<-·~i·· t· ,, ·· ... /--:·· .

Vl
 ..
02 OIL'l)TE SOLUTIONS AND ALTERNATIVE STANDARD STATES J EXER~ISE~1"
Express as a function of temperature the free-energy change accompany- relative.to the infinitely dilute, weighj-pe]'cent standard state. Determine
ing the transfer of oxygen gas into an infinitely dilute, weight-per cent the oxygen content of. liquid iron. held in a solid magnesia crucible if
so.ution in. t:q•.::c iron (i.e. the free energy change for the reaction: equilibrium is. established at I.,6()()°C under a vacuum of (a) 0·001
atmosphere pressure, (b) 0·01 mm Hg pressure. Use the data for the
!02W ~ OcwiXlnPo)). transfer· in 'standard stat~ given inexercise 6.1: ·

I
·. ;·
}~- ' . . . ' ,' . -· .. . . ,,.: .
6.H The following values have been determined for the activity of 6.L At one stage during'rhe refining of a. bath of molten steel/the metal
carbon at 925°C in solid iron-carbon (austenitic) alloys, relative to composition was found to be: .· · · · · · · ·. · ·
graphite as the standard state for carbon: I .. · Element C s p Mn
0·005 0·010 0·015 0·020 0·025 0·030 1· r ,.,../ 9-:,~~f ~ht-pe~tf_e,n~ . 0·06 0·028 0·025 0·13
~1'
I
0·064 0·132 0·205 0·280 0·361 0·445 At .•.the time ii'~ili'\ of the bath the temperature of the latter was
l,570°C and the activity bf the P205 in the slag was 9·6 X 10-19, relative
Determine the activity of carbon at 0·5 weight-per cent concentration, to pure liquid 'p205 as \he standard state .. Assuming that the· partial

l
relative to the infinitely dilute, weight-per cent standard state.
pressure of carbon monoxide formed by oxidation of the carbon was
I atmosphere and given the following data, determine whether or not
6.1 An iron-I weight-per cent vanadium alloy is equilibrated at the carbon and phosphorus oxidation re .•ctions were in equilibrium with
I .623~Cwith a hydrogen-steam mixture containing 5 per cent by volume respect to the oxygen content of the metal bath.
of steam. On analysis the metal is found to contain 0·033 weight-per cent
oxygen. Assuming that oxygen dissolved in iron obeys Henry's law, find
the effect of vanadium on the activity coefficient of oxygen. The free-
energy change accompanvins the change of standard state from oxygen
2P(,; in Fe)+ 50(% in Fe) .== P20s(I) : .:1G0 = -164;174+ 138·63Tcal
gas to the infinitely dilute, weight-per cent solution of oxygen in iron is where the standard state for C, P and O in iron is the infinitely dilute,
given by: weight-per cent soiution;: ,
. ···:·. Solute X
.;02,1) ~ o(wtXlnPo): JG0 = -27,790-0·79Tcal. .· . QL ;. . '~:. P:·> .Mn
0·2l 0·045 0·047 -0·002
6.J The ratio PH,s!PH, in equilibrium with iron containing 0·04 weight-
6•11, 0·041.' :.·· 0 -0·012
per cent sulphur and 1 ·2 weight-per cent carbon isl ·40 x 10-4 at l,600°C. i! .·

In dilute Fe-S alloys the interaction parameter e~ is ~0-028, relative to

the infinitely dilute, weight-per cent standard state of sulphur in iron.
Find the effect of carbon on the activity coefficient of sulphur (i.e. e~) in
the ternary alloy. The free-energy change accompanying the transfer of
standard state from sulphur gas to the infinitely dilute, weight-per cent
solution in iron is given by: ;>;

~
!S2w ~ ScwtY.inFe): L1G0 = -31,520+'5·27Tcal.

6.K The solubility of oxygen in pure liquid iron is 0·230 weight-per cent
at I ,600~C and, within this solubility range, the activity coefficient of the
dissolved oxygen is given by:
logfo - -0·20(%0]
10
 CHAPTER 7
Evaluation of Experim·ental'Data/ il
. .
. .
. :·; ,' ' .\.; ·~· ~J
~·
)'
'~-
A variety of experimental methods are used to obtain data from which
the thermodynamic properties of a system, or a component or phase in
a system, can be determined. Several detailed reviews these experi-of
mental methods have been published (1, 2, 3]..and bompr.ehensive a dG = -SW
assessment of the accuracies and limitations of the vari~us techniques
has been given (3]. These aspects of experimental therrriodynarnics are
not, therefore, considered here. However, f).a:viig obtained the experi-
mental data, the problem remains as to howthese can best be used to
obtain the required thermodynamic properties.
Many of the worked examples presented in the earlier chapters are, in
reality, calculations starting with experimental data. For example, the
heat capacity/temperature relations which form the basis of most of the ·
problems presented in chapters 2 and 3 have been evaluated from·
calorimetric experiments. The use of experimental data to evaluate the
equilibrium constant for a reaction, which in turn is used to obtain other
thermodynamic quantities, is demonstrated in chapters 4, S and 6.
Evaluations of the thermodynamic properties of a component of a
solution, from experimental measurements of properties of another
component, are described in chapter 5.
The thermodyn .rnic relations upon which these evaluations are based
are all presented in outline in chapter 1, and the examples can be solved
by application of the appropriate basic relations. However, there are two.
additional techniques which are commonly in use for obtaining thermo- ·
'dynamic information from experimental data, the treatments involved.·
being somewhat less obvious from the relations given in chapter 1. These
tcchniq ues are: . ·. . . .. .
1. Measurement of equilibrium e.m.f.s of galvanic (electrochemical) ·
cells. .. . · .. ·_· . , , · ' . .. . ·
2. Calculations using data directly derivablefrom .phase equilibrium
diagrams. · · · · ·. · · '
This chapter contains a brief review of thl thc~6dynamic· principles '.
involved in these methods and. worked exampfos ;tq;: illustrate their ·.
-,~:;;~;Jii!t{ ·., , · · ·• , . · .;,:'. :·:
E.M.F.~ OF GALVANIC CELLS 135
Calculation of thermodynamic quantities from measured e.m.f.s of
galvanic cells
Measurement of the e.m.f. generated when a reaction occurs in a galvanic
cell is, in theory, one of the most direct and simple methods for obtaining
free energy and other thermodynamic data for the reaction and the .
components in the reaction. From the combined First and Second Laws
of Thermodynamics, the change in the free energy accompanying a
reaction which is made to proceed reversibly at constant temperature
and pressure is given, following equation 1.33, by:
where Wis the maximum work, other than mechanical work, which the
system is capable of performing. Integration at constant temperature and
pressure yields, for a finite reversible process:
LlG = -W. 7.a
This work may be available in electrical form if the system comprises
a suitable galvanic cell. In this case the maximum work (W) available,
following a given amount of reaction, is equivalent to the product of the
potential drop across the cell and the quantity of electricity which passes.
Thus, for a reaction involving 1 mole of a substance,
W = zFE 7.b
where E (volt) is the potential drop across the cell, z is the number of
electrons transferred, per molecule of substance, Fis the charge on !gram
electron (96,494 coulomb g-equiv-1) and 1 gram electron is the number
of electrons corresponding to 1 gram mole of univalent ions ( = Avogad-
ro's number).
The magnitude of the external work obtained from the cell is dependent
upon the value of the e.m.f. generated by the cell, this latter being a
maximum when the cell reaction can be regarded as taking place
reversibly. Such a condition is realized when an external potential source
is connected in opposition to, and adjusted to balance exactly, the cell
e.m.f. such that no current flows through the external circuit. Under
these conditions the work term (W) in equations 7.a and 7.b can be
equated and the reversible free-energy change in the reaction is given by:
LlG = -zFE. 7.c
Changes in other thermodynamic quantities during a reaction can also
be expressed in terms of the e.m.f. generated when the reaction takes
place reversibly in a cell. Thus, applying equation 1.33 to the changes of
extensive properties in a reaction:
d.dG = .d VdP-ASdT
 r· J ', :·

136 EVALUATION OF EXPERIMENTAL DATA ,.. '.; :•}:M.F.s OF GALVANIC CELLS 137
or, at constant pressure: Activation·overpotential·(TJ:J.is directly connected with the activa-

Ct7) p == -.dS.
tion energy
.
of the ra'.te controlling
'I ,,. .·.
reaction
..
final state in the eleS!r'ode metalsurface or gas.
in the transfer of ions to the
.

But differentiation of equation 7.c with respect to temperature yields: The above descripti~ri ot)rr~v'ersible·electrode processes is brief, since
the subject is outside. the scope pf thermodynamics and has been
( o.dG)
oT P = -zF oT
(oE) P '
described in detail elsewheje (4]i However, such effects are of the utmost
importance in considering practical problems associated with galvanic
Hence: cells, such as electrolysis, corrosjon, etc .. To summarize, when obtaining .
.dS = zF(~~t. 7.d useful work from a s~p~taneou~ iJp reaction, the cell e.m.f. is less than
the reversible value and conversely, ·.vhen driving the cell by an external
Similarly, substitution of 7.c and 7.d in the definitional relation 1.31 energy source, an e.rn.f, greater than the reversible value is required.
(applied to th'! property changes in a reaction) gives: In the examples which follow it'is assumed that the ions participating in
the reaction exhibit only 'one valency and the recorded e.m.f, is the true
7.e potential.. under reversible
·,· .-
conditionsfor
. . . . .
the reaction
. .,. .
being considered.

In practice the cell e.m.f. is measured potentiometrically as the value
of the external e.m.f. necessary to balance the cell. When the reaction
involves I mole of a single species of ion, z is then equal to the valency
exhibited by that ion in the solution. Clearly, if z is to be unique, the ions
participating in the reaction must exhibit only one valency in the par-
ticular reaction under consideration.
It is important to appreciate that the cell e.m.f., £, occurring in the
above equations is that measured reversibly; i.e. under conditions in
which no energy is being introduced or extracted from the cell. In
practice, however, the majority of electrode processes involve the
passage of a net current across the electrode/solution interface; the cell
reaction thus proceeds at a finite speed, is therefore irreversible and
cannot be treated directly by thermodynamic methods. The e.m.f. (ei)
of a cell electrode operating at a current density i differs from the re-
versible value (e.), This difference, TJt, is called the overpotential (or·
overvoltage). That is:
T/i = e1-e,.
Y/; does not arise froi:n a single cause, but is made up of three separate
components:
Ohmic overpotential (f)o) (which is zero when no current flows)
arises from the ohmic resistance of the solution and electrode.
Concentration overpotential ('Y]c) results from the concentration
gradient in the electrolyte immediately surrounding the electrode
which is necessary to enable the ions to diffuse to or from the electrode
at a rate equal to their discharge rate.
I to metallic silver: · ;~ • .
i t '. ',:.
7.1 The e.miof the iatv~~jc·c~l(
'Pbc,J I PbCl2c,> jAgClc,>·IA&,>
where all components iare present ~s pure solids in contact with HCI
electrolyte, is OA90 volt at 25°C and, at that temperature, the temperature
coefficient of tire e.m.f. is -- l ·84 x 10-• volt deg-1• The standard entropies
(i.e at 25°C)for Pb, Ag.and AgCl are 15·5, 10·2 and23·0 cal deg t! mole=!
respectively. Calculate: tire free-energy and entropy changes for· the cell
reaction and the standard entropyof PbC]i at 25°C.
The first step is to determine the cell reaction. In all problems concerned
with galvanic cells it is necessary to appreciate the reactions that occur
when work is being withdrawn from, or injected into, the cell. Reference
to 7.c shows that a positive cell e.m.f. is associated with a negative value
of L!G (i.e. a spontaneous reaction) and it is. necessary, therefore, to
specify the direction of spontaneity. For this ·purpose, the convention
adopted is to write the cell such that the spontaneous reaction (associated
with a positive e.rn.f.) involves the passageof positive ions through the
cell from left to 'right.rand ~f negative' ions from right to left. Returning
now to the example. jinder consideration," the spontaneous reaction
(obtained by shortcirg,lfiti!)gthe electrodes) involves the dissolution of
the lead electrode ; .,. . ·' ...
Pbc,> -> Pb+++ 2e-
where e represents an ~lectron. Sim~Itaneously silver ions are converted
. ·
t; . 2Ag,t + 2e- -., 2A&,>·

Iv~
!
}·
~;.: ~-.

} . . :< ..
,
 ~~ . ~ ;§';~:.- ----
··.ti
,.l.
I.JS EVALUATION OF EXPERIMENTAL DATA
', . . E.M.F., OF GALVANIC CELLS 139
The net spontaneous cell reaction is therefore:
Substituting the given values:
Pb(a)+2AAa)- PbC12(a}+2A&i>· ,·::.'
,: •. :- ' 'jl
,: i s~ .. = -8·5-(2x 10·2)+15·5-(2x23·0)
This means that the passage of 2 gram electron charge; 2,f'araday)
through the cell.causes the solution of I gram atom· of lead.and the
(i'.e. I
l
= 32·6 cal deg-1 mole-I.

simultaneous precipitation of 2 gram

atom of silver. J' J
7.2 A galvanic cell is set up with electrodes of solid aluminium and solid
The components of the reaction· are all present in their respective '
I aluminium-zinc alloy, and an electrolyte of a fused AlClrNaCl mixture.
standard states and hence the free-energy change accompanying the When the mole fraction of aluminium in the alloy electrode is 0·38, the
reaction, given by equation 7.c, is equal to the standard free-energy e.m.f, of the cell is 7·43 millivolt at 380°C and the temperature coefficient
change for the reaction at the tem~rature ofthe measurements. That is: ! of the e.m.f. is 2·9 x IO-~ volt deg=", Calculate the activity, and the partial
l;, molar free energy and enthalpy of mixing, of aluminium in the alloy at
.100"" -zFE0 'i 380°C.
where E0 is the e.m.f. of the reaction under the standard conditions. The cell can be written as:
If .ac
0 is to be expressed in calories and Eo·is measured in volts, the
Al(,. pura)lA!Cl3 - NaCl IA!(,, In AJ..'b!).
units of the Faraday must be converted from coulomb g-equiv=! to cal
volt-1 g-equiv-'. Thus, since 1 cal=4·184joule: · and the cell reaction is:
96 494 . Al<•. pure> -+ Alc1, 1n AJ-'Zll>
F = 4_;84 = 23,066 cal volt-1 g-:equiv:1•
At the pure aluminium anode, 1 .?,ram atom of aluminium ions (AI+++)
In the example, • the maximum e.m.f. when the ·.cell is operating is formed by the passage of 3 Faraday of electricity. Simultaneously
reversibly (i:e. the balancing external· c.mi) · is 0·490 volt at 25°C. 1 gram atom of aluminium ions is discharged at the alloy cathode and
Therefore: . . .. . . . · . . '. dissolves in the alloy.
:L1G°m ~· -(2x23,~xP·490). ). The measured e.m.f. of the cell at equilibrium is therefore related to
the transfer of 1 gram atom of aluminium from pure aluminium into the
·-= -2.i,600.cal. · · alloy. It follows from the definition of quantities of mixing (p. 11) th.at
'·.-. .:_
the e.m.f. data will yield relative partial molar properties of aluminium
The standard entropy change for the reaction is .obtained by use of
in the alloy.
equation 7.d applied to standard changes: . .
. : •· . , ·o , ·. , In order to demonstrate the method for solving this example, it is
..:1so - zF(oE. ).::' •.. : . .'':'··, ·,. necessary to derive the dependence of the free-energy change for the cell
. . er ,. . , , . l
1 •
reaction on the activity of the aluminium in the alloy.
In the general case, for a reaction between components X and Y to
The temperature coefficient of the e.m.f., (oE.0/~f),. .is given as .t,. produce P and Q:
- I ·g4 X 1Q-"4 volt deg-! at 25°C. Therefore: , · ~i . ,. bX+cY .=
mP+nQ,
.
LJSO ""2x23,066x(--1·84x·lo-4)
. . .
. . . .

'
,-,~
' .,.
.
.,,
·i
...
the relation between the free-energy change and the activity of the
components is obtained by combining 1.55 and 1.56:
. .· \
The entropy change for the reaction ·is. the differ6nce ·;~tween the .dG-.dG0 =
·
Rnn(~°?).
tfx,ay
7.2.1
entropies of the products and the reactants: ,
.{
• This type of cell, in which the electrochemical differencebetween thee~ la
As0 ••• s~+2st~.si~2Sto·).
. . ~, .
·. . . ; . (, the concentration of the transported component, is called a concffltratkm cell.
. ·r
.' ,.
I

V 11 I·
140 EVALt)ATION OF EXPERIMENTAL DATA
E.M.F.,, OF ·GALVANIC CELLS 141

As in example 7.1, when all the components are present in their The. valency (z) of aluminium is 3; F is equal to 23,066 cal volt-1
standard states in a galvanic cell, the e.m.f (E°} of the cell is related to g-equiv-1 and E is given as 0·00743 volt at 380°C (653°K). Hence,
the standard free-energy change for the reaction by the expression: substituting in 7.2.5, at 0·38 mole fraction of aluminium in: the alloy:

AOO • -zFE6•
Similarly when one or more of the components are not in their standard
~-,~Ai :=_:...;3x23,066-x
1 · . 4:575 653
.f .. ·
o'.00743
'
x
.
·

states, the measured e.m.f. (E) is related to the free-energy change by: ·;· ~ ::::~:'t.12
7.2.2 or:
AG ••• -zFE. , OAt. '!=, 0•673.
Substituting these equalities in 7.2.1 yields: The parti~I; ~~i~r f~e~:::en~rgy 6,f ~xing of aluminium in the alloy is
E-Eo - - RT.ui(a';·°?).
zF ttx.ay
7.2.3 obtained by use of eq{1ati9n 7.2.6:. . . · ·.
. . i/ .
IM
'
.·:

It will be noted that the activity term has the same form as the equilib-
~-.... ··" .
G.1.-,.=,_.-zF~
. ·..
l : =·. ~ (3 x· 13 066 X 0·()()743)
. .

·:·t '. •' · ... ·. '. •,' ~ : .

rium constant for the reaction. However, in a galvanic cell reaction, the
concentrations and hence the activities of the components of the two
·. f · = -514 cal mole- 1•

electrodes differ from the equilibrium values; otherwise AG (and hence From equation 7:e ippliecfto thepresent.conditione ;
E) would be equal to zero. In practice the cell reaction is prevented from · }'·j,·>': '.: . :.·.·.· es ]
reaching equilibrium by applying a back e.m.f. equal to E. Provided this .·· · · ill',.._. :f=
(· .·.: :·
~zF[ ..;-r(ar)
·. '.
. . .. : .P
condition is achieved and no chemical reaction occurs in the cell, the
activities of the components, at the concentrations at which they are and (oE/oT)p is givenas 2-9 x' i0"-5 volt deg:..1• Therefore:
present in the cell, can be determined from measurement of E and £0, /Jfi =' ~r{i3,Q66.[(7,1J>< I0-3)-653(2·9 .x 10-5)] ·
using equation 7.2.3. In the general case it is not necessary to measure
E°. as the value can be calculated from knowledge of the standard free- = . + 800·;.\'. cal ' mole>
. ''·1 .· ..
. .. ... · · · , -

energy change for the reaction. Experimental data from ·e.m,f.ce.lls can be combined, for example,
In the present example the reaction involves the transfer of aluminium with heat capacity daia .to derive free-energy changes aµd; equilibrium
from the pure aluminium electrode (i.e. aA1 - I) to the solid solution. constants for reactions in a -similar manner to that demonstrated in
Hence equation 7.2.3 becomes: earlier chapters. In each case; the first step is usually to convert the e.m.f.
data into one or othe~ of the thermodynamic functions in the manner
E-Eo = _RT_ln(aAl(alloy)). 7_2.4 outlined in the preceding examples. However, the USC of galvanic cells
zF aAl (pure>
is not confined to the evaluation of thermodynamic quantities.' As well
In the specific case of both electrodes being pure aluminium, the last as being used for electricity storage, etc., galvanic cells are widely applied
term in 7.2.4 is zero, since aA1 - 1 for both electrodes. This cell would in the extraction and refining of metals. The two following examples
generate no e.m.f. and hence, as E = 0, E0 is also zero. Therefore: illustrate typical applications of thermodynamic principles to derive
RT operating data for such cells.
E = - z.F .lnaAl(alloy) 7·2·5'
7.3 in the electrolytic refining of copper, a galvanic cell is formed in
-(CiA1~dl1),.. _ ~J 7.2.6 .;.,hich one electrode is impure copper a~d the other ls pure copper.
Assuming that th, activity of the copper is unity at both electrodes, how
from 1.54, where G!{ is the relative partial molar free energy of many kilowatt-hours of electricity are consumed per lb of copper refined'
aluminium in the alloy.
if the cell is operated at 0·2 volt and the efficiency is 85 i,,er cent 'I

·-:

.r /.~··
 142 EVALUATION OF EXPERIMENTAL DATA
E.M F.1 OF GALVANIC CELLS 143
Because the activity of the copper is unity at both electrodes, the
Assuming that the condensed phases are pure and the CO2 is formed at
change in the free energy accompanying the transfer of copper from one
1 atmosphere pressure, determine the theoretical minimum voltage
electrode to the other is zero. It follows. therefore, from the equation:
required to effect the decomposition if the cell is operated at 950°C.
J:1(,0 .., -zFE0 .
The standard free-energy change for the reduction reaction can be
that the e.m.f. generated by such a cell is also zero. Hence, theoretically, obtained by adding the two reactions:
. any imposed Positive e.m.f. :will ca1,1SC transfer of copper from anode to
,f)._ iCc,>+t02ea> ~ tC02ea>: AG0 = -70,650-0·lSTcal,
cathode and O· 2 volt should be adequate to maintain the transfer reaction.
The reactions occurring in the cell arc, at the anode! )f, !Al203c,>~ Alo>+t02ea>: J(rO = 202,880+ 1·88TlogT-46·11TcaJ.
~.} ,-+ Cu+++2e- to obtain:
and, at the cathode:
!Al20x,i+iCc,> ~ Alcn+tCO.zw: AGO=- 132,230+1·88TlogT
Cu+++2e--+, ~l)'. ·. ·. ..
-46·26Tcal.
Th us th~ transfer: of 1 h:iote. of eop~ fro~· one .electt~e to the other
requires the passage of 2 gram electron (i.e. 2 Faraday) of electricity At 950°C (1,223°K) the standard free-energy change for this reaction is
+ 82,800 cal. The positive value signifies that, in the absence of an
through the cell. The gra:fu-moloctilar ~eight of coppetis 63·54 and 1 lb
external driving force, the reaction as written tends to go from right to
is equal to 453·6 gram. 'Therefore, the transfer of 1 lb of copper across
0

the cen requires the passage of: ' .. . . . ' ' . ' left and alumina cannot be reduced spontaneously by carbon at this
temperature. In the con.-nercial process, the driving force required to
· 453·6 . . . · :.. . •· :'. . ',' > . make the reaction go from left to right is provided by the imposed e.m.f.,
.: 63_54 x2 = 14·272 Faraday.~£: el~~i;r:: •, 1, , •, .••,A .,
the magnitude of which must be at least sufficient to provide the free
energy equal to the positive value of the standard free-e iergy change for
The answe~ is required in kilowatt~h~urs, ~ tlie:'ilitl& of the Faraday the reaction at the operating temperature.
must be expressed in ampere-hours. Now: . '·
The relation between the free-energy change for a reaction and the
I Faraday = 96,494 coulomb g-equiv-1 . ,
-:.· ,·.,
. : t\.'

e.m.f. generated when this reaction takes place in an electrolytic cell is

• • •: 'A.
given by equation Z,c:
'7 96,494 x 2·78 x·10··'.4.a~~e-hour'g-equiv'"'1;
Therefore: '.,;,' JG0 = -zFE0•
,. ''. • , ._"J, ·,

14·272 Faraday - 382'.6a.m~re-hour·g-eqwv-1 When all the components arc in their standard states the calculated free-
. .·, . . ., . . . ,, ~ ·,. . ' '
energy change for the reaction under consideration is for the formation
and, if the cell operates at 0·2 volt at 85 per· cent e.fficie4cy,the electrical
of I mole of aluminium from the ox.ide and, as 3 gram electron arc
energy consumed per lb of copper refinedi\ _.· < ._- :~-.'· ·1 ·. transferred in the electrolytic cell per mole of aluminium formed (i.e,
382·6 X 0•2 }(X) . · .• · '· · .· .. ';' · · z = 3 for aluminium), the e.m.f. of the cell is:
· x- = 0·090 kilowatt-hour. z
1,000 85 · · i .•. ~·· ;~
82,800
Eo =
3 X 23,066
7.4 In the commerical production of aluminium, alumina 'is decomposed .1-
hy electrolysis in the presence of carbon, according to the reaction:" .. -1·20 volt
(where F= 23,066 cal volt-1 g-cquiv-1).
!.f(,. JAl203<.,+~c•> ~ A1ro+iC0200• I•
The negative sign of the standard e.m.f. signifies that if a cell is con-
<>::' · .• Although this is the ~ual way in which the reduction of alumina is expressed, in .·
structed such that the reaction, as written, can occur, aluminium ions at

v!·:
\;'. . . ._. . . practice carbon monoxide .is formed at the high temperatures at the carbon electrode, .
the liquid aluminium electrode will combine continuously with oxygen
:>t(:fa'.t~t~~
;,,i( .; '{and this .is then oxidized on its path to atmosphere to give the overall reaction con-
~re. . ·.. . ' ' .
~
to form alumina. However, an applied e.m.f. of 1·20 volt will prevent

\I .,
 ,,,,·.
.,

144 EVALUATION OF EXPERIMENTAL DATA THERMODYNAMIC. DATA FROM PHASE 'DIAGRAM~ 14!!

this reaction occurring, the direction of the reaction being reversed (to
form liquid aluminium from the: alumina) by an applied e.m.f. of greater
than this value. In commercial practice, between 6·5 and 7·5 volt is
applied to a cell which usually operates at a current density of about
700 amp ft-2 of electrode, the efficiency being 75-90 per cent.
Derivation of tbermodyn,amic dat11 f',rom comtitntional (phase) diagrams
It is possible to derive certain thermodynamic properties for the com-
ponents of a binary system from a knowledge of the location of the
liquidus and solidus lines on the constitution~ diagram. (Conversely
the phase diagram can be constructed from knowledge of certain thermo-
dynamic properties of the system.) The basic principle which is used
arises from the equality of the partial molar free energy of any one
component in each of two or more coexisting phases at equilibrium.
Thus at constant temperature the partial molar free energies of a com-
ponent A in a solid solution of solidus composition and in a liquid
solution of liquidus composition are equal, i.e.
C,. (1, at aolldua) = CA(!, at l!Q11ldus)• 7./
If experimental measurements .of the required thermodynamic
property of the melt are not available, it is often acceptable to calculate
values of this property at the liquidus from data for the solid, using the
above principle. Then, assuming regular solution behaviour, these
Jiquidus values may be extrapolated to any temperature required in
the melt.
The accuracy of the calculated liquidus values of the property is
dependent on the approximations necessary in the treatment adopted,
and on the accuracy of both the thermodynamic data used for the pure
component or compound and the location of the liquid us and solidus
lines on the constitutional diagram. In certain cases, where the latter
have been determined very precisely, the property values calculated for
the liquid may be more reliable than the available experimental measure-
ments. More frequently the calculated values are subject to an un-
certainty of at least ± 10 per cent, in addition to any error incurred in
The standard state for a .cornponent of a solution is conventionally
selected as the pure substance iri the state which is stable at the tempera-
ture under consideration. Thus, at temperatures below the melting point
of the pure component,' the standard state to which, for example, the
activity of the component in solution is referred is usually chosen as the ·
pure solid substance. In this case the, activity pf a solid component A at
the solidus would be given by equation 1:54 in the form:
. CA (s, at solidus) - G'i.c,) = RT In aA(•. at solldlll) 7.g
and the activity of A at the liquidµs would be given by:.
CA. (1, at liquidu•>-:-G'i.c,) = RTln:a,..:,i:~I liq11Jd~)• . 7.h
When G'i. is defined a'.s above.Iin both cases) as the free energy of pure
solid A, it follows fromfquation 7./th,at:
.···-:. A1~,
«:at iolidus)
. '. .: Q
:, '·/ :_ ..,.,
.,;.;. •',,: .
aA (I, at Uq11ldua)•
If the solid which separates from themelt is practically pure A (i.e. the
solubility of B in:A is negligible), then ·ii'~<•, at solldlll) can be taken as unity.
It follows then that the a.c~ivityof the liquid at the liquidus, referred to
A
the same pure solid st!ndar~ state, is· ~,1so unity. . .
Asan alternative to tht conventionjustdescribed, it is sometimes more
convenient to refer the-~tivitie~.of:a component in a molt~n solution to
the pure liquid as the stjndard state at temperatures below the melting
point of the pure compcinerit. "This is particularly desirable when it is
required to extrapolate ,thes~ acdvity values to temperatures above the
melting point of the pur~coinp9rt¢t;il;where the pure liquid is, of course,·
the normally accepted st'an'da:td stat~.. . : ·• ' ; . . .. . .
At temperatures below th,e m~ltjng point the pure liquid is less stable
(i.e. has a less negative fr~e energy) than the pure solid. Thus expressing
1.54 in terms 'of the pure, suptrcool~d liquid as the standard state:
· ' . Q ~ (pur~ ,i;;~)...:
~(I(~ _:RT!n aA (p11re solid{ . 7.J
Hence the activity of th~ solid, relative to' the pure, supercooled liquid
standard State; is Jess tn"an unity for 'ail temperatures below tbc melting
point. •. ~ ·· · · · . ' :, · ,· ·· . ·. , · . · ·
Dividing 7.j throughout by T and differentiating with respect to Tat

I
the temperature extrapolation for the liquid when the actual solution constant pressure, it is obtained that:
deviates markedly from regular solution behaviour. Hence it must be
realized that the values of thermodynamic properties calculated in this
way for temperatures above the liquidus may be only very approximate. (a1naAJ'r't1;d>) _ ~[a(~)
oT
. ;
_ a(~>)] ·.
.· oT p · p .
However, even a rough approx·mation on these lines is more useful than . 0
the alternative assumption that the solution behaves ideally (i.e. con- . .,;, !!M>RT
"." ' 2H A(s) 7.k
. Jorms to Raoult's Jaw). · ~'

I V(\
 146 1':VALUATION OF EXPERIMENTAL DA.TA THERMODYNAMIC DATA FROM PHASE DIAGRAMS J.47

from the Gibbs-Helmholtz relation 1.34. The quantity HA(•) is the molar Because the solid solubility of tin in zinc is very restricted, the activity
enthalpy of pure solid A. and (H0A(l)- HA{a)) is equal to the enthalpy of zinc in liquid alloys along the zinc Iiquidus can be equated to the
change for the transformation from pure solid to pure liquid (JH1r) at activities of pure solid zinc at corresponding temperatures, relative to the

J »« _ ·
the temperature T. Therefore, at constant pressure: pure liquid iinc standard state. The first step, therefore, is to calculate .,,

. , . ln·~A~~~Hd) :=
which can be solved to obtain ~A(pureaolid),:With·r~pectto the pure,
supercooled liquid standard state, if .1Htr.i; known as a function of
temperature. This latter can be obtained from use• of the Kirchhoff
equation 1.18. Alternatively, ~Ht~ may pc a.ssum~.to independent of
temperature over a small temperature· interval; beini;: equal to Lr, the
latent heat of fusion of pure A at the rn~l~in,g po{n.t., 't~e~e methods are
illustrated inthe worked examples which follow." •. ·; i' '
In alloy systems where the solubility, of the.solute B;'is negligible, the
liquid at the liquid us is in equilibrium with the pure iiulid at the same
7./
the activities of zinc along the zinc liquidus, using equation 7.m•
The heat capacities of liquid and solid zinc are given as:
CP:zao, = 7·50 cal deg-1 mole-!
CP:za<o> = 5· 35 + 2·4 x I 0-3T cal deg-1 mole-1•
Because of the inaccuracies involved in the experimental measurements,
be the known CP data for liquid zinc may be adequately represented by a
temperature.independent constant. If the variation of C,, with tempera-
ture in the liquid state was known with greater precision, it is probable
that the empirical equation would contain one or more temperature-
j; dependent terms. However, with the presently available data, there is no

I
significant decrease in the accuracy of the calculated activity values if
temperature, and it follows that the activities of thei two phases are C,,Zn<o> is also represented as a temperature-independent constant. For this
equal when both activities are measured ~ith the same pure r~~qo purpose, a mean value of C,,:zac., (6·84 cal deg-1 mole-:') for the tempera-
liquid standard state. In this case, equation 7./ can be :written as.;
·
lnaA(l,.,atUquidus)
. =.::. f-. iJRri·~!·.
tt; .... · 'j .
,: J, 7.m
i
t
ture range of interest (160-420°C) will be used.
Thus, for the fusion reaction, Zn(,).= Zn(!):

... r" ;. ! .:1 C,, = Cpz,,ru - Cpz,,<o>

When the solid solubility of the second component i~ appreciable, the
activity of pure solid A, relative
. to the
. j,-qr~:'liquid.
. . . . . . ~fahdard
. . state, is
obtained for the required temperature in the nrai:me.~. previously out-
lined. However, in order to derive the activities at. t/,le solidus (to be
,)
I
f
= 0·66 cal deg-1 mole-",
Combination of the Kirchhoff relation in indefinit- integral form
(equation 1.20), i.e.:

equated to the liquid activities at the liquidusjit is necessary to know the

variation with composition of the activity of A in the solid solution with equation 7.m yields:
range. This is sometimes available from experimental data. In other
cases some limiting assumption, such as conformity to Raoult's or
Henry's law, or regular solution behaviour of the solid solution, is made
Hence:
RlnaZn\l,atliquidus)= I ;2[.:1no+ I 0·66.dT].dT.
in order to estimate the required data. :
.dHo
R In aZn (I, at Uquidus) = -T +0·66 In T + I 7.5.l
7.5. The zinc-tin system is a simple binary eutectic, the eutectic point
t·.. occurrini at.0·85. atom fraction of tin and 198~C. The solid solubility of tin where .dH0 and Lare integration constants which may be evaluated from
f,.. . in zinc is small; being less than O· I weight-per cent at the eutectic tempera- a knowledge of either lnaZn <• •• 1 liquldus) at two temperatures or
j},.:i_:". turi>The t~mpetaiures corresponding to various compositions on the zinc
~{°' .I '·i•~: ,)/, '. ·. '::' ,.. . ,.' '•·. " :::·:~, •\'.,'°., - ' ·\ • , '
an;
InaZn<•. a1 liquidu•J and at one temperature. The latter information is
"&?h(;::Il~q'uJd,,;ts,,reai!;:fr9_11J.}1~ phas~,aragrizm given by. H ansen [ 5 ], are given in presently available, for at 420°C (693°K), the melting point of pure zinc:
;tJ\fttI)t,hffifs(iwo colimfils .of Table J: l'Assuming that; the liquid alloys behave as QZn(I) = 1

ttif
",:iar. sob1tions;. calculate ihe .activity of zinc as a/unction of composition .
1

AH1r = Lr= 1,740 cal mole-1

~~~t~t {Js{gf. · I
 \\: ' ' •'"' .
14R EVALUATION OF EXPERIMENTAL DATA
. . . . T~ER..~O~Yf~~IC::P1,~A_:l!~?M PHA~E IDI,AG~AMS '14!>
where Lr is the latent heat of fusion of pure zinc. Hence: : Extrapolation ;of th~se ,data ·to-.703°K. requires t,he a.ssu.tilption of
./JH693 - 1,740""' iJHo+ 0·66.dTf regular ~oluti?n-beha~~>ur for theijquid alJoys. For~Jrc>m·~~$c8Su,mp-
tion it follows that the func_tii;>ri,Rfl~
y1 is Independenfoftcmpcratu,reat
from which: constant composition (~ee e:ica:mple 5.8). Thus equa~ing.the V(ihi~'oft.J:Lis
iJHo -= 1,280. function for zinc at (he liquidus temperat4re and. the: 'tetn~ratu,re
Substituting this value in 7.5.1 yields: required (430°C or 7031K).: ,· . .

1,280 T(liquidu,/ ln ,,z,; (I, at liquid~) = 703 In YZa (l, 7QJ)

RIn Oz,, (1, at liquldlll) =- - -;y- + 0·66 ln T +I. or:
:'. · ·. . 7<11ciolda1).
But: log;,2nci;,,o3> = logy:z.ic1,at11quld111)· 703
ln O:z.i (I, at liqu.ldua) ••• 0 at 693°K. . ' .
Therefore, substituting T-. 693°K yields: This yields the activity coefficient <;>f zinc in liquid alloys at 7()3°K at t,he
compositions for which the activity coefficient.of zinc is known 'at the
I= -2·464 appropriate liquidus temperature, as in Table 7.1.
and: The extrapolated values of y;i;,(I) at 7Q3°K, obtained by use of this
logozno.a1 uquidua> -
1 ( --;y-+0·661n
1,280 T-2·464 ) · equation and the data in column 4 ofTable,-7J; are given in column S.
4.575 Column 6 gives the corresponding values of the activities of zinc in
This equation may now be used to obtain values for the activity of liquid alloys at 703°K. ·
zinc in liquid alloys along the zinc liquidus line. The values so obtained Kubaschewski and Evans list values of D.JJ. and. S.Jf.. for various
are given in the third column of Table 7.1 and the corresponding values compositions of Zn-Sn alloys at793°K. Thevalues were obtained from a
of the activity coefficient are given in the fourth column. f critical assessment of experimental measurements (mainly e.m.f.) of the
! thermodynamic :propeij/es. qf these solutions; The activities of zinc in
Table 7.1 iI the liquid alloys at 703~k, calculated from these data (through 1.63, tne
definition of C.J:f,. arid 1(54) aie given in the last column of Table 7.1.
I Comparison of the ex~rimental and calculated values shows that the
Liquidus
Calculated
Experi- l differences ar~ )!lsignifi.carit for :rinc··cc,ncentratio~.:s betw~n 0·9 and

I
Tern- Com- mental
position 0·6 atom fraction, but !·he. error increases and becomes quite large at
pcrature D2nm 'YZn<ll Y:z.nu> aZn(I) Ozn<H
OK Nzr, (liquidus) (liquidus) 703°K 703°K 703°K low zinc concentrations'>.The values of S-1{. given·by Kubaschewski and
Evans are some 40 per ce·nt la'~gerthanthe ideal values (i.e. S-l,; is posi:
693
673
661
649
636
1·0
0·95
0·90
0·80
0·70
1
0·967
0·941
0·921
0·895
' 1
l ·018
1·046
l·ISO
1-279
1
1·017
1·043
1·140
1·250
1
0·966
0·939
0·912
0·875
0·9S3
0·914
0·874
I
I
I

)
tive), so the actual soltfzi.6n;,do riot
conform' to regular behaviour as
assumed in the calculati9hs_;Jriview of'this, the close agreement between
the calculated and meat'ured activities of zinc at zinc concentrations as
low as 0·6 atom fraction is remarkable.
623 0·60 0·867 1·446 1·388 6·832 0·830
0·50 0·846 l ·691 1·580 0·790 0·757 •f: 7.6 Pure cadmium mil~~ii/s2{°C,j;,d./!J°efreezin,gpoilit is lowered2·4°C
610
0·40 0·799 1·997 1-781 0·712 0·665 by the addition of 0·51 ~iom~p.er ce~i
oi bismuth. The s_o/id solubility of

I
586
551 0·30 0·730 2·437 2·010 0·603 0·536 bismuth in cadmium is hegligibly srfiall: Calculate the heat of fusion of
504 0·20 0·634 3·170 2·285 0·457 0·395 cadmium at the meltingfloirit/321 °C; ·
471 0·15 O·S66 3·775 2·438 0·36S
433• 0·10· 0·488 4·88 2·657 0·266 0·210 .The method addj,te.d\~··s~lve this problem depends on.the fact that
a solvent con,forms to ~aol!lt1s 1a·~-(Le. exhibits idcal,behaviour) as the
• Extrapolated. ' atom fraction. of the solvent-tends to)1nity. Thus, at very low coneentra,
11 (; ' . ·• ' ' ' . . i ...
·~·

I
 '
i
t!SO EVALUATION OF. EXPERIMENTAL DATA- THERMODYNAMIC DATA FROM PHASE DIAGRAMS 151
tions of solute, the activity of tho solvent is .equal tQ its atom fraction. t· an exponential relation between composition and temperature, applic-
Because the solid solubility is negligib1e, the concentration of ~dmium
may be. related to the heat of fusion· through_ equation 7.in which, in
differential form, for the present system may bewritten as:
l able when the solvent obeys Raoult's law. This 'ideal' liquidus can be
calculated for any solute concentration and the temperature difference
between the 'ideal' and measured liquidi can then be related to the
dlnNCd<~at1~u1d~>··. ~H.,.· j excess free energy of the solution at the composition where tho tempera-
. . dT ... -~?r:-:'·i_f 7.6.1
~
ture difference is measured.

where NCd(l,•t liquid •••> is the composition oftheliquid~s at temperature T. I

It is reasonable to assume that .dlfir will.not change appreciably with 7. 7 Thefree-energy change accompanying the fusion ofpure iron is given
temperature over the small range considered here,: lhus, at the melting
point (TM) of pure cadmium, an; will equal the Iatfuit heat of fusion of
I as a function of temperature by the relation:

cadmium (L,). Integration of 7.6.1 fromT(--S91·6~it) fo TM (- 594°K) l LIG0 = 3,130-0·3T1nT+0·518T.

yields: . ' ... ·_·_ ·•· ...• ·--~'. . · . ·
.. Lr( 1 1)_'..
InNCd(ta1n - ~- TM ~T
t'
;;'. 7.6.2
At l,477°C, a 12 atom-per cent liquid solution of copper in Iron I.J In
equilibrium with solid (S) iron containing 5 atom-per cent of copper.
(since at TM, 1), or: · · · · · \: Ca/cu/are the activity of iron in the liquid iron-copper solution, relative to
Ned(!)=

, . Lr '( .1' . l) '~.jk . I

Io,gNCd(ta1n.,. 4·57.5 TM :T.
Substitution of the given values for NCd~atn, TanirM.yieids:•
Lr~ '1~470 cal mole-1.
The value fo~ the heat offusio~ given by Kubaschewski ~nd Evans is
1,530 ( ± 40) cal mole ", which is in good agreement with the calculated
value. The accuracy of the computed value depends critically on the
accuracy of location of the liquidus line at small concentrations of
solute. Reasonable values of Lr are obtained only when the freezing
point depression is known with an accuracy of better than ± O· l °C. In
the calculation it is assumed that the solvent obeys Raoult's law up to
the solute concentration for which the froezitig point depression is
known. The error arising from this assumption will, in general, decrease
as this concentration of solute decreases. · to 0·95, relative to pure solid iron as the standard state.
From knowledge of the heat of fusion of a solvent, the reverse com-
putation can be carried out, and the initial freezing-point depression for
this solvent can be calculated as a function of the concentration for any
solute over the concentration .range in which the solvent conforms· to
ideal bcha~io~~'r'.- Thus-the 'ideal iiquidus~ror'a solvent with latent heat
of fusion Lr (olr'ows from 7;6:2: · · ·
. .· . . " .. ·'.'... . ·.=... exp_·-[Li(.1
. N{l.deaIJlquidw)
· · 1)·](
J( TM_-:7''f·:)
I pure liquid irf)n as the standard state.
I Since solid iron dissolves a significant amount of copper, the activity
I
I
of the iron in the iron-copper solid solution will be less than the activity
of pure solid iron at the same temperature, both activities being measured
relative to the same standard state. As no information is given concerning
the variation of the activity of iron with composition of the solid solution,
it will be assumed as an approximation that iron conforms to Raoult's
law up to 0·05 atom fraction of copper. Thus in the terminal solid
solution, the activity of iron, referred to the pure solid metal as unity, is
assumed equal to its atom fraction. On this basis, the activity of iron at
the solidus at l,477°C, is equal to 0·95, relative to pure solid iron as the
standard state. Since iiquid and solid of the respective compositions are
in equilibrium at this temperature, it follows that the activity of iron in
the 12 atom-per cent Cu liquid solution (i.e, at the liquidus) is also equal
The problem requires that the activity in the melt is referred to the
pure liquid standard state. The necessary change of standard state can
be effected by first calculating the activity ofpure solid iron at l,477°C,
relative to pure (supercooled) liquid iron at the same temperature.
From 7./:
_ JAH1,
RT2 · dT
lnaPe(pureaolid) - 7.7.1

where the activity of tlu: pure solid iron is measured with respect to the
• The use or 4-figure log tables is ~ in'~ calculati;n,
. . ,. ,' ·, .supercooled pw~ li9u;~ .as the standard state. The data given for the

I\ l
 152 EVALUATION OF EXPERIMENTAL DATA
EXERCISES 153
fusion reaction are in terms of the free-energy change. However, insertion
The assumption W~S,:il)~de, at the. ~iinning of this example Ulat iron
in 7.7.1 of 1.35, the Gibbs-Helmholtz relation applied to a reaction: behaves ideally in the S,qlid solution at concentrations up to O·OS atom
Aei'?r _
T
-·J r2 · iJHi dT fraction of copper. This ·a"s&\llllption is, of course, justified· at Whute
of
dilution copper sine.( as\tated in chapter 1, the activity oft.he solvent
yields: tends towards Rao1,1itiiti.,.~h~vi9uras its atom fraction tends to 1.Ul,ity.
However, liquid fron-e,<:Qp~r: .. solutions. show pronounced positive
deviations from ideality,.ifl apditiort to which it is usuallyfound that the
magnitude of such qe~J#iq"ii~' is_.'.grc:at'cr in the solid Ulan in the liquid
(Alternatively, this relation may be deduced from 1.54 applied to the state. Hence, on 'this ~)isis.j(appeatS' very doubtful that the i,ron soijd
activity of solid iron, relative to pure liquid as the standard state.) solution can be regarded ;is !,eh,v~g idca)ly up to i.he solidu.s composi-
Substituting the relation given for .dGO: tion considered here: Ir{'g~_rier11l, ff may be anticipated that uncertainties
in calculated' liquidus activHics may tie·very large when,.itis necessary to
I . [3,130-0·3Tln T +0·518T]
n apo (puro aolld) ••• - RT • make· the 'sornetimes ' douqtfo!" 11,ssuriiption' of ideal '.~1:µlviour of the
solvent in a terminal sJJid' sohition ·containing apptc,¢iable solute. This
Hence, at T= l,477°C (1,750°K): error does not arise, of course, if experimental data a.re'availa.\)le.relating
the activity of the solvent withthe c6mpo~itiordn the'solid ·so:lut:jon at
Opo (J>uroaolld) - 0·967, the temperature of interest. · · · · · · · ·
This is the activity of pure solid iron, relative to the pure supercooled
liquid at 1,750°K as the standard state. From this value and that pre- Exercises
viously calculated for the activity of iron in the 12 atom-per cent Cu 7.A The e.m.f. of the cell:
liquid solution (relative to pure solid iron as the standard state), the
required value of the activity of iron in this solution (relative to pure A8(~)jAgCl(s)IPt, Cl2(t atm)
liquid iron as the standard state) may be calculated. This is most easily is I· I 3 volt at 25°C, and the temperature coefficient of the e.m.f. is
appreciated if the activities are expressed in terms of the fuaacity (f) of
iron in its various states. Thus:
j ..
l -6·0 x .10-4 volt dea-1 at this temri.ratYttl, Cillo1.JJ•t• th•
entropy, enthalpy and free energy of AgCl at 298°K.
•~d..-d
'
ape (liquidus) = Ji-fFoPo(puro
.
(liquidus)
solid)
=
O
. 95 , as calculated !-'I. 7.B The standard oxidation· potential ~fthe Pt, d12(&), CI- electrode is
(Relative to Fe, - I ·358 volt at 25°C. At the same temperature the standard oxidation
pure solid) potential of the Pb,,>, PbC12(,), CJ- electrodeis +0·263 volt. Determine
the partial pressure of chlorine gas in equilibrium with Pb and PbC12
Ope (pure solid)
=
. Ji~UfO IOlfd}
IJ Id)
IOIC 0•967, as calculated, j. (i.e. the dissociation pressure of PbCl2) a; this temperature,
Po(puro qu
(Relative to Fe,
pure liquid)
'· 7.c Find the ~in.itnu~ porenti~l which must be1applied in order to
produce pure liquid ti if from pure, liquid SnC}z in an electrolytic cell
: -··.. (·- • ' i·._ • ·:· ' • •
Therefore: operated at 500°C. · •.:.,. · .. ,, ·

OFo (liquldus) = Ji.f Po (llquldus) - O ' 95 X O ·967 7.o An electrolytic qll is set UP.· in -which one electrode is liquid
Fe (pure liquid) cadmium and the .other, a liquid cadmium-lead alloy. When the atom
(Relative to Fe,
pure liquid)
fraction of cadmiumin. the alloy is 0·45; the e.m.f. is 0·0125 volt at
480°C. Calculate the adiivity coefficient of cadmium in the alloy at this
= 0·92. temperature; . ··, t ·. · ';:_. ' ·
. -: >::: ....
·.-·:ft:·)·"
/) ':··.···:, ·-.-'
;•:
,_ . ~ ,. . ...
-., ,.
 .;,..,. .· {
... -~ ' EXERCISES
J!l-4 E_VALUATION OF EXPERIMENTAL DATA )_ ·
I-
7.P. The e.m.f. of' the cell: · t,,,
:Cd(pure,f>lLiC1.-..KC1--01Cl2IQic1i/~,
is 0·0225 volt at ~°C and the temperature coeftfcient ~f'the e.m.f. is
6·7 x 10-5 volt deg-1 when the atom fraction of cadmium in the alloy is
0·269. Find the partial molar enthalpy of mixing, th~exces~ partial molar
free energy and the excess partial molar entropy of cadmium in the alloy.
7 .'F An electrolytic cell is set up with an Al-Bi alloy (NAt - 0· 167) as
one etecucce and pure aluntlniui:n as the other electrode. The cell e.m.f.
is O·O 1304 volt· at 900°C and the temperature coefficient of the e.m.f. is
6·26 X lQ-5 volt dek-1, Determine the activity. the activity coefficient,
and the partial molar entropy, enthalpy and free energy of mixing of
aluminium in the alloy. By calculating the enthalpy of pure aluminium
at 900"C, find the partial molar enthalpy of aluminium in the alloy at
this temperature.
7 .o The e.m.f. of p.n electrolytic cell comprising a liquid Pb-Zn alloy
(Nz,, = 0·094) as one electrode and pure liquid zinc as the other electrode
is 0·0105 volt at 600°C. Assuming zinc vapour behaves ideally, determine
the vapour pressure of zinc over the alloy. The free-energy change for the
vaporization of zinc is given by:
,1(]0 = 30,300+5~74TlogT-43·2_5T, ca] mole=".
7.H The e.m.f, of an. electrolytic .cell eomprising a zinc-tin alloy
(N7_,, = 0•50) as one electrode and pure zinc as ¢e other electrode is
10·8 x 10-3. volt at 479°C. The e.m.f. of a similar cell comprising a
zinc-tin (Nz,, =0·50) and a zinc-tin (N~-0·173) alloy electrodes is
26·9 x 10-3 volt at the same temperature -. Find the ', attivity coefficient of
zinc in the (Nz.,.;., 0· 173) alloy. · · · · · 7.N The freezing point of silver is lowered by 3·5°C by the addition o ·
7 .1 The standard electrode potential or the A.si..{g+electrode (i.e. the
e.m.f. of the cell comprising pure silver, and hydr~g~n (PH, - 1 atmos-
phere) electrodes) is -0·799 volt at 25°C.. Thi standard electrode
potential of the SnjSn+-+: electrode is +0·140 volt atthe same tempera-
ture. Determine the maximum e.m.f. generated between pure silver and
pure tinelectrodes when they arc imniersed in a·Jµitable electrolyte at
25°C. What would be the maximum value ofthe e~.f. if the activity of
tin in the tin electrode was. 0·5? '
. . . . . .
j
~ . .
. .
7.J At 25°C the activity of mercury in a thalliumimalgam is 0·95 when
NH,= 0·96 and the activity decreases to 0·50 at NHa •• 0·66. Find the
'(
155
maximum e.m.f, which could be developed between electrodes of these
two amalgams when immersed in a suitable electrolyte containing Hg+'
ions at 25°C.
7.K The standard oxidation potentials at 25°C for the Fe, Fe++ and Cu,
Cu+r electrodes arc +0·440 and -0·340 voltrespectively, Show that
when solid iron is placed in a normal solution of copper sulphate at
25°C the copper is almost completely precipitated from solution.
. (Assume that the solid iron and copper are pure metals.) ·
7.L The following e.m.f. measurements at 900°C have been obtained
with electrolytic cells comprising pure liquid aluminium as one electrode
and liquid aluminium-lead alloys as the other electrode:
E, millivolt
N Al (In Al-Pb)
0·0017 100·8
0·0067 56·2
0·0084 48·9
0·0131 35·5
0·0165 29·4
0·0404 6·15
Find the activity of aluminium in the (NA1 -0-0404) alloy at this tern·
perature relative to the infinitely dilute, atom fraction standard state foi
aluminium.
7.M Assuming that the latent heat of fusion of nickel is independent o:
temperature, draw a graph of the 'ideal liquidus' for dilute alloy:
containing nickel as the solvent.
0·5 atom-per cent barium. If the solid solubility of barium in silver i
negligible, calculate the latent heat of fusion of silvtr at the meltin,
point, 960·5°C. .
7.o A cadmium-zinc alloy containing 7·5 atom-per cent of zinc start
to solidify on cooling at 300°C. Calculate the approximate compositio:
of the first solid which separates from the melt. Assume that LrcC4> i
independent of temperature and that cadmium conforms to idea
behaviour in both liquid and solid solutions.
7.P In dilute solution the logarithm of the atom fraction solubility o
nickel in solid aluminium is approximately a linear function of th,

. '/',:'
I\~-
1S6 EVALUATION OF EXPERIMENTAL DA'L', EXERCISES 1S7
reciprocal of the temperature in degrees absolute. The solubility varies Find the activity of copper in the alloy of the solidus composition,
from Ns; = 0·0022 at 500°C to NN1 •••• 0·02,2 at 650°C. Estimate the relative to pure solid.copper as the standard state, u~ing'the free-energy
partial molar heat of mixing (i.e. the heat of solution) of nickel i,n these data for the solidification ·oC:coppergiven in' Uie previous exercise.
solutions.

7.Q The cobalt-lead system shows extensive liquid immiscibility, the

{
,,
+
'.
:!,
7. o . A silicon~~itvir: ~i~r
·~ptai~ing)6·9 weight-per cent silver lies on
the silicon-rich Iiquidus:at 1,289°C ~d;at this composition, the activity
liquidus temperature rising rapidly from the eutectic point, close.to the 'i~·. of silicon (relative to pur~ liquid sµ~9<?D as standard stai:c) is 0-72 at
melting point of lead (327°C), as shown by the following data for "·:•: l,4Q0°C. The:solid solubility of.silvetip· silicon is neg),igiblc.
liquid us compositions: Assuming that In asi\ varies Hn~tly with tcmpctaturo . between the
liquidus and ·1;400°C,t1deteri\lin~· the···activity and t,hc partial molar
Liquidus, °C 913 969 1,076 entropy and enthaipy, it'~ilico,n inJh{liquid alloy at 1,789°C. As an
Nc0 at liquidus 0·022 0·029 0·055 approximation,' the ~i~:r~nce in the heat capacities of solid a;nd liquid

Assuming that the latent heat of fusion of cobalt is independent of

silicon can betaker; as\'. :>:~ .': '.
:·: .·' .· I l
temperarure, estimate the heat of solution ol soljd coba,lt in the al)oys. :
~C,.~·-.-0:5.~dq-
( ,. - . . .. . . . . ·.
mole-.

7.v The following v~\~~S:b'.i~~.~n\ictennin.cdfor the liquid~ and

T»; A motor-car radiator is filled with an 'anti-freeze' mixture con-
solidus compositions ai' various temper~tures in the system Fe-Pd:
sisting of water and ethylene glycol (HOCH2CH20H, molecular
weight= 62·07), the freezing point of which is -15°C. Calculate the ~empcrature
G . ·. .
} i \Uq~id~s .• . Solidus ,
weight-per cent of ethylene glycol required: 0
.fatorri~pcr ~rit Pd · atom-~r cent Pd
(a) a_ssuming that the latent heat of fusion of ice is independent of 1,535: ,, :;" ·, 0,. .•· .. :_6: ,
temperature, 1.,s11 : 2,os···. · · .. 1,05 ..
(h) using the heat-capacity data for water and ice, which can be taken 1,497 . '; 4·36 . . · 1,20 .
as ! 8·0 and 9·0 cal deg-1 mo!e-1 respectively. · l,480 f .
6:50, . · . . 3·55:: ·
Calculate the latent heat.of.fusion of pure iron.
7.s At I ,050°C a solid gold-copper alloy containing 6 atom-per cent of , ·r· I
gold is in equilibrium with a liquid alloy of the composition 0·90 Neu,
References
O· l O N Au· Assuming that the copper in the solid alloy conforms to
Raoulr's law, determine the activity of copper in the liquid alloy at the I. c. WA9Nl!R, Thermod}namics of Alloys, Addison Wctley, 19S2. .·.·•
Jiquidus temperature. 2. F. D. ROSSINI, Experimental -Thermochemistry, Intertci~ Now York,
l~~ ·. . . ..
The free-energy change accompanying the solidification of pure copper 3. 0. KUBASCHl!WSK.l and E. LL. EVANS, Metalliqg/cal Tnermoclumi.siry,
can be represented as a function of temperature (cf. example ,J.3) by: Pergamon, 1958. · · · ·
4. (;, KORTUM and J. o'M, BOCKRIS, Textbook of ElettrtX'lttmi:try, Elsevier
~Gr= - l,650+2·09TlnT-0·75 x l0-3T2-12·85Tcal mole+'. PublishingCo., London, 1951.
s. M. liANSEN, Constitution of Binary Alloys; MtGraw-ron, 19!18.
7.T At 1,000°C a copper-zinc alloy containing 0·160 atom fraction of
zinc lies on the solidus and that containing 0·206 atom fraction lies on
the liquidus. The activity coefficient of zinc in liquid Cu-Zn alloys,
relative to pure liquid zinc as .the standard state, can be represented by:
; ,· ·'
-~-]r~ •
RT!n 'Y:zn - -4,600Ntu• ,t··

,:
'
I
 , .. •
. :~/
CHAPTER 8 ·-~·-
GRAPHICAL (ISO-ACTIVITY) METHOD 1S9
I
Integrating 8.1.3 between the composition limits X and x (assuming

The Ternary Gibbs~Duhern Equation· I; that data for component A are available and data for B are required)
then:
I
I
I 8.1.4
.)
Graphical (iso-aetlvlty enrre slope) method
8.1. In. recent years, thermodynamic data. on ternary systems have I where X is a composition on n8/nc == constant for which GB is known
become more readily available, and as wi.th binary systems, values for and xis the composition for which GB is desired. The integration may be
one component oftep yield to expcrimontal determination more readily made, therefore, if the experimental data for ()A are adequate to enable
than values· for other components. It is very desirable to be able to the partial derivative (onAf on»)o •.. 110 to be evaluated, as a function of GA,
calculate these unknown values from the measured '.quantities. This will
be possible if the Gibbs-Duhem relation 1.43, applied to a three-com-
r along the integration path n»/nc,.. constant between the limits C,.1,., and
CAo•··
ponent system, can be expressed in a practically usable form. Various This partial derivative (on,Jon9)0,.,..., is the direction, expressed in
treatments of this problem have been made, being either graphical, terms of composition, of the tangent to a curve along which CA is
analytical or
a combination of
both. However, all t~ese methods may be constant (i.e, the. tangent to an "iso-activity" curve) at a point which
regarded as originatiag from 1.43 applied to at~~
~ .. I .
system A-B-C, i.e.
. ..
must also be on the integration path given by n»f n - constant. Thus, on
the isothermal triangle, tangents to iso-activity curves at the points where
·. 'TAdGA+nndG»+ncd(.ic ~ 0L '8.1.1
the latter cut the integration path are extended to intersect the A-B side
where nA, nB and nc'ar:e the number·~fm~l~ otJmponents, A, B, C of the triangle. These points of intersection, expressed in terms of the
respectively. Thu$ complete differentiation of.1.41:/ · · · mole ratio of component A to component B, yield values of (iJnAfiJn9)
G' - nACA+,,.c~+ncd,i ,; corresponding to the values of GA for the iso-activity curves.• The area
applied .to the same ternary system, and subtract;ion·bf 8.1.1 yields:
. dG' = CAdJZA. +l'B~B+'.<1~&'6:
.. .
. . 8.1.2
'
f.
Under the curve of (onAfeJn9) versus (JA, between the limftl
CA,., will then yield a value for the integral of 8.1.4.
This technique may, of course, be applied equally readily to either
0Ac•> &ftd

~ activity or activity coefficient data. For by substitution of:

where G' is the free energy of the whole. system ~~kming nA+ n8 + nc
mole. Equation 8.! 2 is an expression.for tlie C9mpl~te differential of the
free energy of the system at constant temperature :itnd pressure. Thus,
! or:
from the fundamental laws of partial ·d.i£rcfrentiai1ori• and with the
restriction of nc = constant (or dnc\= 0): · ·:: , ) ...
. ' '. . - ,. II
I
dC1 = RTd1ny1N1
into 8.1.3, relations similar to 8.1.4 may be obtained with loga1 or logy1
(oG~\ = ·-(anA), . :J 8.1.3 replacing Ci. The available data are then plotted in the form of constant
\ oG-;}"•,0,.,"".·. {
ll<J ·011J!
log a,\ or 16gyA curves in the ternary diagram, and the procedure used to·
which may be regarded as a further alternative. foim of 8.1.1. For a obtain the solution of the appropriate integral is identical to that just ·
single-phase field in an isothermal section teini?1 system, 8.1.. 3 may or'a described to evaluate 8.1.4.
be integrated along a path of constantfru and:nc which corresponds to a The following example illustrates the use of the 'tangent intercept
path of constant mole ratio n»/nc ( =N»/NckSuch ·a: path is a straight method' [IJ for solving the Gibbs-Duhem relation for a ternary system.
line passing through: the comer. N1,. = I on the ternary chart.
• In physical terms, this point of intersection of the tangent Jives tho composition
• Sec appendix 5. or a binary A-B alloy which could be added (in infinitesimal amount) to tho ternary
alloy on the integration path without changing the value of a,. in the latter.
160 THE TERNARY GJBBS-DUHEM EQUATION
Liquid aluminium-bismuth-lead alloys exhibit a miscibility gap, the Bi-Pb
rich liquid containing 5 atom-per cent Al in Pb-rich melts and up to 35
atom-per cent Al in Bi-rich melts at 900°C. The Bi-Pb-rich liquid has been
extensively investigated at this temperature [2] using an e.m.]. technique
with AJ+++ as tire current-carrying ion. Ternary alloy compositions were
selected along various lines of constant mole ratio N8JNPb in the ternary
isotherm. The e.m.f,s generated in cells containing such alloy anodes with
; GRAPHICAL (ISO-ACTiVITY) METHOD 161
The' unknowri' compo~ition x will pc N~1 =- 0·60, NPb - 0·20, NAl~ 0·20,
and thus-the lirie of the integration path is NnJNPb""' 3, as shown on the
bismuth corner of the ternary coordinate chart, Fig. 8.1 b. The only point
along this path at which the activity of qi~muth is known is at the inter-
section of this path with the binary Pb,.aBi:sideof the triangle. That is,
the composition X will be given by (Nsi = 0·75, NPb - 0·25, N Al= 0).

cf
+ "
t.

0 (}()2 0-QA 006 008 0-10 0-12 0:14

NAI

Fig. R. la. E.M.F.'s of composition cells in Al-Bi-Pb system at 900°C. (2).

pure liquid aluminium cathodes are given in Fig. 8.1 a as a function of the
atom fraction of aluminium. If for binary Bi-Pb alloys:
RTlny8; = -1,179N~ O·l
·· ·
' 0·2
{ ·. ·
\ .·0·3·: , .
N Cqmpds,tion X
, -Pb-... .
calculate tire activity of bismuth in an alloy of composition:
Fig. 8.1 b. Grllphjcal ~olutio~of t~~ary Gibb!--Duhe~ equation'. (Tarlgient.lntercept
(Ne1 = 0·60,Npb = 0·20,NAt = 0·20). . . . : k method(!)). . . .... ·. . . . . . ' ' :

As activity data for aluminium in the Al-Bi-Pb system are available

;s
The inte·g~ationof 8.1.5 in~de;·iherefore,~lqn8tbeheavy ~ritinti~us
line shown in Fig. 8.lb.\ . . . . .· .
and the activity of bismuth is required, equation 8./.4, expressed in
• It is required that th¢ available data be expressed in terms of the
terms of activity, will be of the form:
compositions for which values oflogaAl are constant. For the concentra-
tion cells used to obtainthe data of Fig. 8. la, the activity of aluminium
r Joga9;,., = log Oei,x, in the alloy anode will be related to the reversible e.m.f. E generated (in
l volts), by the relation: ·
lo11a.,,.. ] RTlnaA1 = --3FE.
On Al .
( ---
J \ one)I01041, ""• d logaA1 ~'!.' - !'f..~.' - eon11an1
. 8.1.5 a"'
where 1 is the activity of aluminium in the alloy with respect to pure
lo1aAJuU n,.b Nr• aluminium as the standard state, Fis the Faraday, equal to 23,066 cal

'I
 162 THE TERNARY GIBBS-DUHEM EQUATION
ALTERNATIVE GRAPHICAL (ISO-COMPOSITION) METHOD 16.3
volt-equiv-t, and the valence of the aluminium ion AI+++ is taken as 3.
Measurement of this area (using the trapezoidal rule" with chord
At 900°C (1,173°K), logaAl • -12·89£, and the compositions for
width 6,. log aAt= 0·05) yields a value of the integral in 8./.5 ofO·l 11. As
which kigaA1 is constant will be those for which the e.m.f.s of the cells for binary Pb-Bi alloys:
are constant. Thus, from Fig. 8. ta, at values of constant e.m.f., a series
of composition values are obtained, expressed in terms of NAt for various RTln;,81 = -l,179Nib
values of the mole ratio N»JNPb. These, with the CQrresponding values of then:
IogaAt, are listed in Table 8.1. !so-activity curves f6r the various constant logaa;,,.,(NB1 = 0·75,NPb = 0·25) • -0·139.
values of lqgqAI may DOW
' ' :.~ . . be,
: plotted on.the ternajy coordinate
' ; '' . ' ~.
Chart Of Hence:
logos;,.,= -0·139-0·lll
Table 8.1
.. I
. ~- -0·250
and, at the required composition:

as;(Na1 = 0·6,NAt = 0·2,Npb = 0·2) - 0·563.

60 0·773 0-033 · I 0·024 . . 0,020 '; ''fo,b!4 ,,,r,

50 0·644 0·045 0·032 , 0·026, .. ')-1 0·019 .0·034 ·
40 0·515 . .· '• ' . . 0·045 .<,
. t:•
0·062 0·044 0'034 ~. 0·026
30 0·387 0·088 0·064 0·064
20 0·049 ,· :Y ,:0·035 · 0·094
0·258 0·127 0·093 0·074 •-, ... Jio-052 0·141
10 0·!29 0·195 0·140 . 0· 110 <io·078
5 0·064 0·220
! ·667
0·290 0·180 · O· 138 lo.097 0·305 Shoded 0tea,
0·02! - 0·218 0· 161 ,. :· ;-0· I l l 0·408
': t .
. . ~ . .
109 °AJ • • 1•20
Fig. 8.1 b and tangents· drawn to each ·of these curves' at their points of
intersection with the integration path N8JNPb = 3> The intersection of • ~Ill
these tangents with the Bi-.AJ side of the tetri~ ch~(sl:i.o~n arrowed in
l-ig, 8. J b) yield values which, when converted intoth'~ratio of the number
of mole of aluminium to · bismuth, correspond. .to the quantity
( cnAJonBi)loea.,. ,,, •. These are also. listed. in
Table. 8.1 and are plotted
against - IogaA1 in: Fig. 8.lc. It will be noted that the resulting curve is
asymptotic to both axes (cf. the curve of example 5.5).
The area to be integrated is between · logaA1 at X (Ns; = 0·75, -log ••.•.•
Npb = 0·25. N At= 0) for which onAJon81 ••• 0 and logaA1 at x (N81 = 0·6,. Fig. 8.lc. Graphical solution of ternary Gibbs-Duhem equation. Solution or the
.Vr6 = 0·2, NA1 = 0·2). This latte'tvalueisfound to belogaA1 = -0·016 by integral in equation (8.J.5).
plotting values of this quantity versus the values of NA1 along the integra-
Alternative graphical {i~omposition curve slope) method
l ion path and extrapolating to yield the value corresponding toNA1 =;: 0·2.
lhus the area to be integrated is as shown shaded in Fig. 8.Ic where it is 8.2 An alternative treatment [3) of the solution of the Gibbs-Duhem
' '
seen that an error will be.involved, due to the necessity of area measure- equation applied· to ternary systems is based on a further relation
ment under a curve going to infinity. With the present data, this error is between the quantities in equation 8.1.2, arising from the fundamental
f'air!y smal], though in general, it will be significant.'}
• See appendix S.
164 THE Tl!;RNARY GIBBS-DUHEM EQUATION ALTERNATIVE GR..\.PHicA, {~s9-co_MPos1r10N> M.ETHOD 165
' . . ~ .. . .. . .'. . ·. ' ' . ' ' .

laws of partial differentiation. For, it may readily be shown that, with at which· data for component · B, arc desired and any. composition for

l· · · · ·'
the restriction ttc = constant (or dnc = 0) applied to 8.1.2, then: which data for the latter component are known. Stich ·an 'integration
must, of course, be along a path of constant: y,. passing Jbri:H,1gh· the
8.2.1 'unknown' com[ positi°f ..Tha.i·isx: · · .·

which is an alternative expression of the Gibbs-Duhem relation with the

restriction nc = constant applied. In this form, the relation is incon-
f
CB,., ==. qB,x>
'

·
·
+J ~ (soa/)
X
· 2
x dx
1-co111tu,t
8.2.3

venient to apply in practice and can be improved by substituting the

variables: t where X is a composition at which GB is known, this composition and
the whole integration being· on a path corresponding toy...; constant.

tt:
Such paths are shown as the shorter-dashed lines in Uie ternary coordinate
chart of Fig. 8.2a. ·
nc [ Ne ] ,. Al

y = nB+nc - N~+Nc 'I'""

where N,.,, NB and Ne are the atom or mole fractions of components
:·.r , 1"''
1' I 11\ '"'
l·r
I I I 1,1

/ , . \ . \
I I I I I
A, Band C respectively. As: 0 80j O 1.., I . I,...\.,_\.
r-;IMI ~I ,o1¥:,
?'&' 01 "?10\.
ro~\ = -nc. '•• ~I1...,1.,1 11'1 01\1\

r
;,,i. ~\ ;r>\
\On°J,,.,,,,nc (nA +nc)2

I
, · I , I ·I I I
I I : I" I I
then:
0·60 :·,·,·'· :.I·
. ·,:- ,t::. r .. \ I\
r. onA=7
-ncax]
""'""
·
I ,. I
; I
I :1 'I'
I
I
\ I
I
.I
I
I
\
\
Similarly: ,· ,I I· 1· I I

-nc2-oy
[onB = -
l,
1/' ,/,I
i I
?. , f
·,:.'·
·I

i
· I\

\ ·.
\\
.~r
o·'> ~ -,:.
1f
y ,Jn.-,IIC

Substituting these into 8.2.1 and noting that the constancy subscripts
. .,'( '·. · , · /. f '
·I
I :,•
,7;,
f

'.,· .
· I
I . •.
I :
t

, /ff'
,""' I
.,· '.
I

,
/
,

.
. /

I
r, ,,,,-/'"\\
I
I
l.o·.l!-
).....,1('.,
o-60\
can be replaced by x and y: . / :{: .. )/./ I _.........--i , .

0·20,
- -:- ,,,.
I

/)..,,f)~,;,~· '
! ,· ..i ~<-.
"

::r---1·., ,,r'r' ·

·:--,i,.....-.
t"r/

_.Jr/,t·~-
.·
\

I
I

\
.rr7

-' .,..,._ -
- - _,,_, d" ,,"·. ,. ;,----~- - ~'
then:
~:::.:--'- -<>-1.f ·;-t..:.:.--c.;-.-i----,l,,,_
;,:,
.·· t' - . r: -.-: . Y

(ac~) (acoyA) . =------P-- ,, . . . -c-

' .. . •• ·-"' ----
\ ~-o-9
= 1~ 8.2.2 .1 ' . u • •\\ \

OX y X :x B, b·8C)L.:
.
K·
Compo,11t1on x
0·60-.. 0·40 0·2
ompo,i'tion·x :· : · ·
If, for purposes of example, it is assumed that data for component A are . . .:........,_:.. fila;
available. the relation 8.2.2 may be integrated between the composition
Fig. 8.2a. Ternary co-ordinate chait sh~~irig details of solution of equation (8.2.5).
. . t' . . ' .,
• These variables x and y are particularly convenient to use, for they are always
finite, having values between zero and one, and become constant when the restrictions
on the right- and the left-hand sides (respectively)of 8.2.I are applied, as in deriving The integrated ternab, dibbs~D,uherrt relation _expressed'in 8;1.3 is
8.1.2.
not particularly convenient to' from a practical standpoint, an,d it is \lse
12 ·'.
,,,
166 THE TERNARY GIBBS-DUHEM
. . . EQUATION
.
\ ALTERNATIVE (?RAPHICAL (ISO-COMPOSITION) METHOD 167
preferable to.express 8.2.J in terms of activities rather than partial molar composition x, and the integration in 8.2.5 must be made along this line
rec energies. The s~bstittitions: · · ·· · · · · from a composition X for which a81 is known. From the data for the
cA·=.ci1+RTlnaA·· froml.54. binary Pb-Bi system, aB1 may be calculated at NB1 • 0·75 or y ••• 0·25,
and the integration is made between this composition on the Pb-Bi
.md the equivalent for component B are made.int~ 8.J.l and yield a binary side of the isotherm and the composition x (i,e, along the heavy
relation corresponding to 8.2.3: · · · ,. ' line of Fig -. 8.2a).
r
I JogaB,.,
. =
. :.fx T(;log~A)•: ·:·,1'.·.-'il ..•
logaBcx>+. ":"1 -0-.- .. dx \ .. 8.2.4
The available data for the ternary Al-Bi-Pb alloys are for composi-
tions along lines of constant mole ratio N8JNPb, which correspond to
I . J X y ,· JC·, ' · lines of constant y as shown in Table 8.2. These lines arc those shown in
1.. X . . _y-c;omtant
Fig. 8.2a. The quantity (ologaA.Joy) is required for the condition:
where Xis the composition at which~ is kno~Il. \(: l;:·· ·
The solution of 8.2.4 for a ternary alloy of.a: pa.rticJ~r composition x NJ>'o
x = N N - constant.
Pb+ Al
i" obtained through the following procedure: The·dat.~:ifor component A
a re assembled j n the form Of Values o·f log aA for constant values Of the Lines for which (x = constant) pass through the Bi - 1 comer of the
composition variable x. This may entail interpolatio~·~t~een available · AI:-Bi-Pb ternary isotherm, those for increments of 0·10 between
experimental values, though in an original study ofa syitem;.the.necessity x = 0·9 and x = 0·5 being shown as long-dashed lines in Fig. 8.2a. It will
tor the data to be in thi/; form would. bekept in The.assembled mintl. be noted that the line x = 0·5 passes through the 'unknown' com-
data are plotted in the form of log aAvers~ the composition variable y position x.
for the constant values of x. The slopes oftangenfa;to t~ iesul_tingfamily In accordance with the outline given earlier, values are first required
of curves, at the particular constant value of y cotr~spondi,ng to the for logaA1 at constant values of the composition variable x. With the
integration path, yield values of'the quantity (c11ogoAJay)" for the various presently available data, these values oflogaA.1will be obtained for alloy
values of x, These, multiplied by:(y2/x2), are then plotted against this compositions in the ternary chart of Fig. 8.2a at the intersections (shown
composition variable x,
the area under
. the resulting graph between .the circled) of the short-dashed lines y = constant (for which data arc
composition X and the 'unknown' composition x yielding the integral available) with the long-dashed lines x = constant (for which the data
in 8.2.4. The application of this method to the data of example 8:1 are required). The points of intersection may be deduced graphically
will illustrate the necessary procedure. from an enlarged version of Fig. 8.2a, or, more accurately, may be
The ternary coordinate chart of Fig. 8.2a shows the composition x of calculated individually from the three simultaneous equations which
the alloy for which the activity of bismuth is required. Equation 8.2.4 describe the constancies of x and y and the condition that the sum of the
applied to the present problem: . mole fractions is unity. For example, the composition of the intersection
of the line y = 0·333 and the line x = 0·7 is calculated as NA.1 = 0·1249,
r log' . = log OBicx, + Jx· y2
"'z
(iHog aAl)·. J ,·
a-·y -
dx ) 8.2.5
NR; = 0·5834, Npb = 0·2917 (to four-figure accuracy) from the equations:

l .
Oni,.,
X
X
·
JC
.
··•
.f-constant ---NPb
Npb+Na1
= 0·333 or N8;-2N'Pb - 0
has to be solved along the path:
NPb
N . . --
Npb+NA.I
= 0·7 ' or NA.1-0·428NPb - 0
y = N .PbN = constant
Pb+. Bi. . . . I NA1+N8;+NPb = 0.
which must pass through the' unkn~wn' composition x'. Paths of constant
In practice, only the value of NA.I for each point ofintersectionis required
y arc straight-lines passing through the corner Al= .i of the Al-Bi-Pb
and values calculated in the foregoing manner are listed in Table 8.2. (p.
ternary isotherm as shewn by short-dashed liries in Fig. 8.2a. The heavy ·
170)These compositions (i.e, values of NA.I) are then used to deduce values
continuous line of Fig; 8.2a is that for y =;: 0;25, which passesthrough the .
. ' . ;:. ' .·. '
I
I
of the c.m.f.s of cells containing such alloys from the appropriate

I
! I\
I
q
 ALTERNATIVE G~i.(.~J;[ICAL (ISO-COMPOSITION) METHOD 169
16$ THE TERNARY GIBBS•DUHEM EQUATION ·. .' '!>- ': ·: . . . ·.. '
The relation bctw~ri:
:
the reversible
". ~· -. . . . .· ..
,·
e.m.f generated (in volts) and
. ·, '
activity of aluminium iuthe'all'Ciy··anode:
' \ •• •,.: .',:; ; ' . I' '

:,·
·
..:'-.'RTlna
. . Al
~- --3FE
·,. ' . ','.· 1.'.::~-- .. :

o,sl .:.. .:. .'(ir:.:::•.\ ~

0·7
. 1~·~ :t. ,.o
~y /$

. • 0·104
0·6

0·5
\ )( X
"'':"
0
I
0
""

0·3

0·2

0·1

0 O·l 0·2 O·A <>5 O·

Compooition,y

Fig. 8.2b. Plots of (-log aAJ). versus the composition variable y. 08

y = constant curve of Fig. 8. la. The e.m.f.s so deduced are also given in
the table."
• It will be noted that the available data for the Al-Bi-Pb system are not readily
amenable to solution by this method, for only a small number of the calculated points
of inte~tion of the x • constant and y- constant lines are inside the composition
field for which data are available. Hence the scarcity of points in Fig. 8.2c. However, .
this must not be regarded as a shortcoming of the method (compared, for instance,
to the tangent-intercept method) but rather that the experimental data were not
obtained with the use of this method in mind (sec example 8.3).
...
Fig. 8.2c. Graphical solution of..temary Gibbs-Duhcm equation. Solution
' :of integral in equation (8.2.5). ·
·~
is again applicable, and' values ofloga1,.; may be calculated directly from
the e.m.f. values of th~\abte·: These are'also listed. The data are plotted
as loga"j versus y for ioris~nt values of x as in Fig. 8.2b, the slopes of
these curves at y = 0·2~:yi~ldingvalues of (tlloga,Jtly)x for the various
values of x. These lattir·q~antities are also listed in Table 8.2, as is the
' '
 ',,
170 THE TERNARY GIBBS-DUHEM EQUATI6N ANALYTICAL (DATA CURVE-FITTING) METHOD 171

} value of y2/x2 for the various values of x. The product of this and the
- ~ :::! ::: ~ ·i ~ 8 partial differential is then plotted against x as in Fig. 8.2c. The scarcity
"t.._ j~ ~ 9 ·$
O
~
O
) ~
0
:(l
6
of the derived points docs uot merit curve-fitting and a straight line has
, ::.., o ,, · been drawn in this figure.
' The integral in 8.2.S is evaluated as the area under this graph between
~1:1 ~
0 ""
00 N°'
00
I oo
00 o
00 x - l ·O and x - 0·5, along the integration path of Fig. 8.2a. Tho. value
g> ~
-I ~
11
-..
~
...-. . 1·
.!-. . · ·-1 I
::::
~
1 of the integral is given by this area as:

-- "'
n:, . -
'-'.
+
'-'
+ .
'-' .
+ :
'-'
+
J 2(~1 )
[ x-o-, ]

' I I ,, 1· ~" , ? ~ c:_?o~aA, "dx = -0·104.

~ ~ $ ~. ~ ..td~} _._, y-0·2$
6 I 6 I 0~ , i
.
~
0
i i ~
0
i i ~
0
i I· ;
0
, , J_:,-RJ,, ·. _ if_.(
.}••..
Thus, from 8.2.5:

logae,,., ••• logas,uo-0·104

":"'
.. I °' r--
~.,.,'<t
0 8s
°'
· M
:::
N
N
~ ''""""""
g.
~ <;"!Jl!J:-'1.··.
:.~i-fl·
. :·~;4.f.\ ·':f~
which, using the value of logaB1,,., calculated in 8.1. (p. 163), gives:
o l_
o
~/ .
O O O
. .
O
. i , .•. I
u
i
O
I ,
O
. I I ··:-.A·.,~..
. ,;,;II· c,,. logaB,•<•> = -0·139-0·104
I · I ··:·.1:~l•• t;.
1 _
= 1·757
or:
I 99':'l
.,., - .,.,
~ ~ 0
- 8~
~ 'R ~ ~ 8 ~ as,,., = 0·572
0
' ! 6"' ":'".9
r--
":'"II.
"'
':"II·
....,
":'II
0 0 0 0 0 0 0
~ ~ which value compares very well with that calculated in example 8.1.

ee r-.
"' 00
°' o0\
\C ~
. .o, - vi
• t'"-,, -
°'...,
-
°' t"""'i
>!
:r; ·:,;::\1· .,,,. j
<'}~· - ··i:•
Analytical (data curve-fitting) method
i,. ,~ . 8
0
$ 0$ ~
0
~ 09 ~. 0~ 0?. ~--·
0
~§s
O
~ 1 , :i'~J ,:}, 8.3 With all graphical techniques, errors arise from the difficulties
·· ~ ~ '=t·r
~ ~
0
· ):;h- };.
:t:, ·?·
_t ••
involved
•
in drawing tangents and measuring areas under curves. In the
:/;t'. ~;~. previous two examples these errors would appear to be small, but this
J .,.. "'· : . . _ · } VI N :. r- ..., _ ::it~/; is not ?ecess,arilytrue in al! cases. Hence it is ~esirable to be able t~ apply
. N O:2 00
~
0
~ ~· K; ~- ~ ::; ~ :!: ~ . ~- 'R ;;;; :\/,. }: analytically the ternary Gibbs-Duhern relation to a set of experimental
r ity of t hi s app 1·ication
0 0...,.
I c;>
..,. t:I
o o
O6 o
·1
6 0 o· · . o.o...,
1
ooO
·· · J:'
- 0 -
o o o·
:
o ~ 0
o· 6
1
.· .•,...
.,,.. "t
· ·

/·
d t Th
a a. . bili
e practica .
w1
'II d
epen d on the d ata
available, for if the latter yield readily to expression in the form of
.,., .. empirical equations relating to composition, the analytical solution of
..., 1 '!'. ::fJ~ the ternary Gibbs-Duhem relation is straightforward. On the other
.i ~ J ·~ < · · -~·· .i ~ < ~ < .::;~j'\ hand, if considerable time must be spent in trying to fit composition
: llio 1'-" ~ < >. < >.l- . ~< ~.-;,>~ ~ ~ >. ~ ~~ ~ ))_~.sf·;. relations to the experimental data, the use of a graphical technique is
I ~w- ~~- · "-l - - -fl{. I: probably preferable.
I -~ g J },;~:; : · It is obvious that the probability of finding relations to describe the
M
00
·' "3 :~ ~ C'!). · r:- ·.:
0
~
O
vi
6
\~ijf experimental data for a newly investigated system will be improved if the
~ 8. o · ·0 ,'.j
11 • ' 11 11
:}~is~'f:•; . latter are expressed as functions which, in other systems, have been found
E II II be .. I I d . . S h f . . h .
:;
0
>< • >< >< . \_' >< >< /?,~
·';i.l .·
to. . sirnp y re ate t?
_con:ipos1t1on. uc a ~nct1on rs t ~ quan~ty «,
;! I I U ~ . ·t\l( previously noted as arising in the Wagner solution for the binary Gibbs-

l'
~~f
"'~-,
"\?J'
-~'!..""4
·~-~-~·
q
.
(
• l•. \J
172 THE TERNARY GIBBS-DUHEM EQUATION ANALYTICAL (DAT~ CURVE-lflTTING)
' .'. .
METHOD . .175.
. ., . ,_. 'I
Duhern relation (see chapter 1, p. 15 and equation 5.6.1, p. 103). But. from the fundamental laws of par:tial differentiation: , i ~1.i"''".~,.
The use of functions of this type is found to facilitate greatly ~ analytical
solution of the Gibbs-Duhem relation applied to ternary s~ [4].
As in the previous examples, the general relation 8.1.J is \be starting
point for the analytical method. The composition variable: where Mis a function of x and y. Hence 8.3.7 reduces to:
y=--
nc
nA+nc
8.3.1 ac ")
( oNB' ;
Jti,
rP
· (ada)- ·- _ · ._
.;..'$u)2 . oy - •Ila
(oGc)
oNu y •
0
'
8.3.8
is used, where nA, (ns) and nc are the number of moles of components . f. . .', ,;,:. '· · ..

A,.(B) and C respectively. Thus:

Multiplying 8.3:8 by y ~.;id adding.'to 8.3.6 yields:

nA • (nA +nc)(l-y) 8.3.2 o(j

( oNs
A) . '. r)_v ; : ·(oG~)· -
:, • ; ('1-:-N-.i· , a;.: Na""."
( NB ) (oCin)
1.-NB ~NB
8.3 9
'
and: 7•
8.3.3 . . . ' )t~·-:·.:~::'; ; ... ,·. .. . .
nc - (nA +nc)y. Alternatively; multiplyipit S.'3,8_ _ by_:p ~-y) and .subtracting from 8.3.6
Also as:
- •. , ns
yields: . . . r . . ·_ .. · '.' ' '
HB - ---
nA+nB+nc
_ 8.3.10
where Ns is the mole fraction of component B, then:
Relations 8,3.9 and f3;10 ate,:ustful forms of th~ Gibbs:-Dubem
ns =- NB(n,.+nc)
-----·
(1-NB)
8.3.4 equation, for they may :i>e i~tegrated. to yield OA,or ~~ if 0.is' known
such that its variation with the composition variables y and N3 can be
Dividing 8././, the general Gibbs-Duhem relation for a ternary system evaluated. · :_· · · · · _-- · ·· · · ·· · ·
A-B-C, by dy, substituting 8.3.2, 8.3.3 and 8.3.4 and then dividing By expressing 8.3.9 arid 8.3.10 in terms of activity coefficients or excess
throughout by (n,. +nc) yields: partial molar free energies (see equation 1.79), tae practicability of
solving the integrals by analytical means is imprfed· Thus _8.3.9 jn
(l -y)(acA) +(~) (0GB) +y(oCc)oy
oy Na I-NB oy Na Na
_ o. 8.3.5 terms of excess partial molar free energies is: ·

Similarly, dividing 8.1.1 by dNB, substituting 8.3.2, 8.3.3 and 8.3.4 and 8.3.11
dividing by (nA +nc) yields:
which may be integrated between NB - 0 (i,e, the binary system A-C)
(1-y) ( oC
--
aNB
A) + ( --NB ) (cCB)
y
---·
1-NB oNB r
(iJCc)
+y --
oNB
••• . 0.
y
8.3.6 and Nu = NB, along the pseudo-binary path y - constant in the ternary
isotherm. Integrating by parts: · · ·

l[(I ~!.r(t((t -t)')]

Differentiating 8.3.5 with respect to NB and 8.3.6 with respect toy, and
subtracting, yields:
'"•> - Gf '"•-~ +
f (ac
L.aN;,.) y-
(acoy,.) N•. ( oNu
oy ) c J1 [(oCu) (acB)
+(1-Nu)2 aj oNB
(oNs)
aj i,
JN.- y0
( Gf

. c-) .
dN•

. ,';: . NB ii . . .. :,
aGc) ( oy_ \ (iJCc)] . 8·3·7
+ [( oy "; oN;/0 oN~ r • O. 0-
. ·_ -.,,. ----
; ' ·,,
. •. .·'.
(1-Ns)
t~ •' . . '. ~ . ,,:-con.taa:,
• 8 •.,.12

,_.
•'.
 -~.
174 THE TERNARY GIBBS-DUHEM EQUATION
ANALYTICAL (DATA CURVE-FITTING) METHOD 115
If a function o: is defined such that:
However, as an approximation, the quantity Cf <.-+O• the excess partial
. RTlny1 molar free energy of component A in infinitely dilute solution in B, may be
«t .,. (1-NJ2. 8.3.13 taken to be equal to the same quantity in the binary system A-B. That is:

cf (N-1) • [Cf (Na-+t)lNo-o or y-0·

- Cf
(I-NJ2 from 1.79, Integrating 8.J.J 1 by parts between the limits N11 •• 0 and N• • 1, with the,
condition y = 0, yields:
then 8.3.12 becomes:

(
lvt(Na) -
/"E CEA(Na-0)+0
. I ·r
Na

oy d NB.
o:B-Y cJcxBJ
. GE
A<N-•> = - ( f
0
t
NB aCf
(1-NJ·aNa.dNs l y-0

-«BNB~l--NB)) • 8.3.U
.. , .'. y- comtant
Similarly, for component C, from 83.10:. .L
• (i (1 ~!B)2 dNa) 8.J.11

{c~,~~ - c~,~;-~+ n~.-(1-yt~~]dN,

. . . . 0 '·. . ' . -~
where the assumption is made that:
0 y-0

cg
-- ..• O as N8 -. 1.
- o:~ NB(l ~ NB)J' .· ,; 8.3.15
I-NB

. ·, .:· Y- eonatant Inscrting8.J.17into8.3.16 and substituting the cxfunctionsasdcfincd by8J.JJ:

Thus, if the quantity 0: can be expressed as a rUhction of composition, f ,.,a

),_,+ tf [•,-
t }
8.3.14 or 8.3.15 may be solved analytically to.yici~ Cf or cg respectively
at any composition NB along the pscudo-binarfpath, y = constant, As
Cf,.._,, = [ •• dN,
{ y ::•] dN 8- N,(1-N,.) •• ' , 8.J.18

expressions of the form lnyJ(l ....:Ni)2 arc generally found to be well

Similarly for component C:
behaved functions of composition (c.g. lnyJ(I _;;N1)2.is independent of
composition in a regular solution); the cxpressioh· of cx1 as a function of
composition is usually an easy matter.': · ··. { ;

l
Ol, •• ,, - u o,,dN,L. +{!'r•,+(1-y)·~;]dNa-Na(I-N,.)•• J,
If data for the binary system A-C ·a.r:c not available, somewhat less accurate
solution of 8.3.11 may be obtained by integrating between. the composition
· :.},~,· ':' 8.J.19
·\)f
~,~~;.
N11-+I and the composition 'desired. Thus: 1 · ·
M before, 8.3.18 or 8.3.19 may be solved analytically to yield G!, or ~ (~
Na . ,,
·,:11 spectively), and their associated functions, at any composition along the

I
.
Cf <N"> = GfcN-o+
I
cg
J [o-Naf a ',
y·a;;(ti ~t>
. : .. '
CF.
2)] dNa ;;:
f.:~f
pseudo-binary path, y = constant, if the quantity ex is known as a function of
composition.

i'.ili
.. - NBCg
. ) .. .
. ''.~),
,.
.
·~ f.~·
The data for the Al-Bi-Pb system shown in Fig. 8.la will again be
used to derive the activity of bismuth in the alloy (NAt - 0·20, Na1..; 0-60 •
1-NB 8.3.16 ,,:~~.{.
.i;j;~l NP'b ••• 0·20). These data arc amenable to the analytical treatment just
y•con1tant
:-.~~f outlined, as the «At function is found to be related very sim,plyto compo-
'.::t:i
:~~
'./:~
;:;;
176 THE TERNARY GIBBS-DUHEM EQUATION ANALYTICAL (DATA CURVEaFITTING) METRO·» 17'7
sition. In fact, this analytical method was used by the original workers [2]
to assess their data. 8
N
(O· l > il > 08-0)
To treat this problem by the method just outlined, the data of Fig. 8.ta 6 .( "lL·£- 8t·8 - y
must be transformed into the function O:AJ as defined by 8.3.]3.
As: (O· l > .< > L99-0)

ti:
RTlnaAI - -3FE ,<6H-9?·8 • Y
therefore:
•..• . 00
-3FE-RTlnNA1 ••••.. V"l . <<>· t > .< > L99-o>
O:AI = (l-NALJ2 - "'~it
.;.. ~ ,< IH-0?·8 - t/
where E is the e.m.f. of the concentration cell (in volts) containing the
· . }~{
· ,,s,i~
•:ii'J>· (O· l > .< > L99-0)
alloy of the composition given by NA1 and the mole ratio NBJNPb, the ,< ~9·£-tl·8 • 'Cf
latter corresponding to a constant value of the composition variable:
(O·I >A>~-0)
y- . ,< tL-£- Il·8 • 'Cf
NB1+NYb
:;: ..

Thus the data of Fig. 8.ta may readily be made to yield the quantity o: as ~.
t'-·
00
•.•• (O·l >.A> £££-0)
functions of N Al along the various pseudo-binary lines y = constant ~ ,< !L·E-10-8• t/
,.,.,•':!.·
·:~
· (which correspond to the va'r:ious mole ratios of Fig. 8.la). On plotting
the derived values of o: as a function of N Al for each value of y, it is found
that a linear relation adequately expresses this function. These relations
are shown in Table 8.3, with the ranges of N Al over which they apply.
-?i1
~~
~~!
-
N
'If'

In the present problem, data are available for the binary Bi-Pb system. \!'.
Hence equation 8.3.15 can be used to obtain the desired value for
the activity of bismuth in the alloy (N Al= 0·20, NB1 = 0·60, NPb 0·20).
As this alloy lies on the path y = 0·75, 8.3.15 must be applied with y
=
constant at this value. Thus:
~

I[
N,.,
~~_!]
-ti:
( c~i(N,.i) = C[l(N,.,-o)+ CXA1-Cl-y)

-o:A1NA1<1-NA0)
dNAI

8.3.20
I
f'l
,,-0·7S

The solution of this relation requires that the quantity cxA1 be expressed
as a function of the composition variable y for constant values of NA1 .11~
between NAl= 0 and NAl= 0·20. To this end, values of cxA1 at various
values of N Al (at 0·025 intervals) are calculated from the o:Aif NAl equa- al/ I
~'-:. I
:.·:·1
•....
~I- "s
N
tions of the table. These values of cxA1 are shown also in Table 8.3~ On 0 M 0 0
$
plotting these o:A1 values versus y, for the various constant values of N Ah . :\it ,.Q
I"' o
I<. 6
N
6
M
6
V)
6 6
00
6
essentially linear relations are found to apply. The equations for the . ::::}/r- r"'tt I =-,

;~1 q l.
··!?.:·
 1711 THE TERNARY CIBBS-DUHEM EQUATION
ANALYTICAL (DATA CURVE-FITTING) METHOD 17'
latter are also shown in the table, with the limiting values of y between Thus:
which they are valid. From these equations, the integrand of 8.3.20:
CCX At .
«AJ+(I-y) oy (• A)
- -73•6.
may be calculated for the various values of NAt· As a consequence of tho Hence:
linear rxAify relations, the values of the integrand are independent of y, i'Bi {N,.,-0,20} = 0•969
y-0·1'
and are listed in the bottom row of Table 8.3./They are equal to the or:
values of cxA1 at y = 1. . . .-.-.
On plotting the-values of'the integrand versus•N~1,a linear relation is OBI {N,.1-0·~ = 0·581,
Na,-0·60
found to apply, being adequately·cxpressed by: f Nn-0·20

. A - 4·01 +3·80NAl• { This value compares well with those previously calculated, and is
probably equally accurate, It will be appreciated that the method j~ ,

r
Substituting this relation into 8'.3.20: . . demonstrated is not much less tedious than the graphical techniques for

(GS,'"·" - cg, '"t~ + {4·01 + J·8N.04N~,t,(l

. ' . J,; ·.
- N.,) ···L~ lf;
. ·.;~sg. .1t
working out a single unknown value and could be much more so if the
« function is not simply related to composition. However, in such cases
as presently considered, the analytical technique is well suited for the
calculation of thermodynamic data for the unknown components over
This equation may now be' used to yield- On; at any value of NAJ
( ~ 0·20) along the path y = 0·75; the value of rx,.,; in the last term being
that corresponding to the value of NAJ Withy== Thus, atNAI = 0·20: r.n;s.
?t:f~- the whole range of composition investigated. For once the composition
relations for et have been derived, they will yield the required data for

~-~ any alloy within the composition range to which they are applicable
. J, '

' ·,},
~ .::~!t
{ £.
~U1(N .• ,-0·20>
= E
Cnt(N.u-~+.
,
J( . · (4. , • 0.13
l . < .
8NAJJ dNAr ~'."-·'
much more easily than a graphical technique, where almost the whole
operation must be repeated to gain each value.

or:
C Elli ( y-0·1'}
\N .• ,-O·W
.,.CEBi {
~N~(I-NA0(8·48L3·72y)}

y-0·75}
N,.,-0
·

·
.

+ [4·01NA1+ 1·9N,ti-5·69NA1(I -NAtJ],v...,-0,20

1

.
y-0,15

{i
.~·~
• The value of Gf,( ,._ o-"} is insignificantly different from the value ot
N1r,,1
NA1 •0·lO
Cf,{ ,._-oo-,.,,.I Thi., is because, in the present example, the integration path CY• 0-75)
corresponds quite closelyto the line in the isothermal triangle along which the activity
coeflicient of bismuth is constant. In the general case, the change in the value of the
excess partial molar free energy along the integration path will be significant, and
could be of similar magnitude to that for the binary composition.
Refereeces
I. It. SCHUHMANN, JR., Acta Met., 3, 1955, p. 219.
However: 2. T. c. WILDP.R and}. P. ELLIOTT, J. Electrochem. Soc., 111, No. 3, 1964,
p. 352.
Cf;; { N .•y-0·75}
,-o
= RT!n i'Bi in Pb-Bi binary at Npb = 0·25. 3. N. A. OOKCEN.J. Phys. Chem., 64, 1960, p. 401.
4. c. w AGNER, Thermodynamics of Alloys, Addison Wesley, 1952
The data for the Pb-Bi binary system are adequately represented by:
RT1nyn1 = -l,179Nl,b
and hence:
RTlnY111 = -73·6 at NPb - 0·25.
 APPENDIX l ANSWERS TO EXEllCISES m
'/J
Jt The relation:

Answers to Exercises r,

is used to yield p2 at 8'S0°C(1,123°K) as it is known that p1 -1 atm at

the boiling point 1,18Q~K~ ·
2.1 Approximate pressure required e 0·3 mm Hg.
2.K Correction = 0·0'119deg mrrr-'. .
2.A Specific heat of ice •• 0·40 cal deg-I g-1• The exerciseis concerned with the elevation of the boiling temperature
2.s Weight of steam required - 6,100 lb. of nitrogen by the pressure over atmospheric necessary to operate tho
2.c Heat required= 15·6 kcal. flow-meter. The values of integrated heat capacity will need correction
2.o Temperature of water- 45°C. by this temperature increment. The pressurevariationof the temperature
2.e Time for isothermal solidification - 71 ·4 min of a transformation is given by the Clapeyron equation
Cooling rate after solidification - 4·55°C m1n-1•
2.F Maximum depth of groove - 0·0188 nun er Tt:JV
From deceleration rate, the deceleration force (F) is calculated from <lP - t:1ii"
F= m.a as 2·97 x 106 dyne. Thus work done on ice for each cm of
2.L Heat of reaction ;= - I IO· l kcal.
movement is 2·97 x 106 erg, or 0·0710 cal. From data for ice, volume just
2.l'd Standard beat of'forrnation ofFe203 • (- )196·3 kcal mole-1.
melted in each cm with this heat is 9·42 x 10-• cc.
2.N Standard heat ~[ formation · of Al,4C3 = ( - )45·30 ± 15• !S kcal
2.G Instantaneous acceleration of aircraft-« 0·325 g. (Acceleration due mole-1• ~· · -
to gravity.)
2.o Heat of formation. of FeO at l,390°C- -55,200 cal mole-1.
Volume of incoming air at 400 m.p.h. is 35,764 litre sec-I. The theo-
2.P Heat of forrnatiorr of CaO at l,000°C = -:-153;300cal mole-1,
retical thrust of ramjet will be given by this volume of air at 273°K and
Calcium undergoes ah allotropic change at 440°C and melts at 8S0°C
0·8 atm pressure becoming CO2, H20 and N2 at TM °Kand X atm
The heats of these transf.<?rmationsand the relevant heat capacity data in
pressure, assuming ideal gas behaviour of combustion products. Rate of
appendix .3 enable the operations illustrated in example 2.6 to be applied
heat supply in burning fuel is,27,400 kcal sec-1 which can be equated to to the oxidation reacti'* for' calcium. . /
T11. 2.Q Optimum ratio P~ofPH,= 1·24. .
f L Cp
273
CProdw:u>. dT Heat of reaction for- decomposition of ferrous oxide at 1,273°K
(calculated, or from example 2.6) is 63,156 cal mole-I. Heat ofreaction
for oxidation of carbqp'· monoxide at 1,273°K (calculated, or from
if adiabatic conditions are assumed. Thus TM calculated as 2,410°K.
example 7..7) is -67,42ttalinoie:-i ·(ofCO). Heat of reaction of oxida-
From fuel data and volume of incoming air, volume of exhaust gases at
273°K and 0·8 atm is 38,280 litre sec-1, which is at 7·06 atm at 2,410°K.
is
tion of hydrogen at I ,27tK calculated as .2. 59,6401cal mole-I (of H,.).
Heat loss from charge i~ 79~ .cal mole'-1 of FeO. Thus proportion (X)
Thus maximum theoretical thrust= 6·48 kg cm-2• Hence thrust on air- of H2 in gas for processto be just autogenous is given by: ·
craft with 5 per cent efficiency= 648 kg.
2.H · Difference in molar heat capacities= 0· 106 cal deg-I inole-I. •. : . LX)
. 59,64-0Xs+.67,423(1 ..
·= (63,156+
. 793).
In the relation CP- C., = a:2 VT/{3, V is the atomic volume, which is
given by the atomic weight divided by the density. The units of CP-C., 2.R Minimum volumeof air= 2,713 cu ft.
thus calculated will be cc-atm deg"! mo!e-1, which is converted to Heat ot reactions for~burning of CO and H2 at' 973°K calculated as
ca! deg-1 mole=! by the factor in appendix 2. -67,800 and -59, l 70 c~l (per mole of these gases) respectively. Exhaust
2.1 Vapour pressure at 850°C = 0·555 atm, (420 mm Hg). gas composition calculajed from producer gas analysis and required air
13
182 APPENDIX l
ANSWERS TO EXERCISES 183
(calculate on basis of25 mole of CO). Additional aif (say N mole) must be of appendix 3. Applying Hess's law at the melting point of aluminium
fed in to keep adiabatic temperature
. of exhaust
.
gases
~ to 1,873°K. Thus to enthalpy and free-energy changes in the oxidation and fusion reactions
1,873 . . . (the free-energy change in the latter being zero) gives the basis for the
J [l:
~J
cp(Produrn)+~ c,(~1=:rJdr
'
~ (67,~~x 25)+(59,l70x 14)._
repeat calculation for the oxidation of liquid aluminium.
3.o Calculated standard heat of formation of NH3 - -10·5 kcal
mole+',
Total air required (for 100 mole of producer gasburnt) is N+air pro- From the heat capacity data of appendix 3, the free energy of the
viding oxygen for combustion, · . · '.;,
formation reaction for ammonia may be calculated as a function of
2.s Maximum temperature attainables, 3,3009 d: temperature, the calculation involving the two integration constants
The assumption, is made that the Cp data'for·thci/eaction products are (.1li0 and I) which may be deduced by substitution of the two known
valid up. to this temperature. Part of the heat of reaction at l,600°C is used values of the free energy. The heat of formation is then given by the
to heat the oxygen and diluting nitrogen in air fr6nt2rc to this tempera- integrated Kirchhoff equation with the constant i!JH0 inserted. The
ture. The remainder heats the reaction products, \· calculated value of .1 H 298 is some 0·5 kcal mole-! lower than the generally
2.T Approximate weight of steel ••• 87 Ib. · ·. -1
accepted value. This discrepancy arises primarily from the uncertainty in
From heat given out by charge, the apparent h~at capacity of charge the free-energy values given.
plus crucible is calculated as 92·5 kcaldeg-'. The itltegrated heat capacity 3.E Graph of the variation of the entropy of Zn, 25 +920°C.
of iron 25-+2,40i}C_is calculated as 26·34 kcal_Toie-1• Thus mass of 3.F The free-energy _change, Fey.:= Fee., is given by:
steel given by: . ··· · ; ,
92·5x200
-26. mole.
i
:~
AGr = -4,290-2·34Tln T-0·63 x I0-3T2+20·85T(T < 1,033°1'.).
• 34--,,-
. .' . .~.,1/
The variation of free energy with temperature (T > 1,033°K) is
3.A Enthalpy change at l,000°K-:+32·5 kcal. < calculated using the Gibbs-Helmholtz relation with the relevant enthalpy
Entropy change at 1,000°K - 34·23 cal deg-1• and heat capacity data from appendix 3, the integration constants being
deduced by substitution of the . known . values of AH and AG (- 0) for
Free-energy change at 1,000°K cs -1 ·7 kcal. the transformation. The calculation is repeated after application of
3.R Free-energy change at 1,000°K - -1 ·8 kcal. Hess's law at the Curie temperature, 1,033°K, to yield the above relation.
This value is calculated using the Gibbs-Helmholtz relation and It should be noted that this latter deviates markedly from the relation
differs from that calculated in. 3.A due to the rounding off of the values calculable for pure iron from experimental data for they/« equilibrium
of the free energy and enthalpy. The uncertainty in free-energy values is in iron-nickel alloys. These deviations indicate the doubtful validity of
usually much greater than this difference, being as large as ± 10 kcal but the present extrapolation of the CP data for they phase to low tempera-
more commonly about ± 2-3 kcal. · tures (cf. example 3.5).
3.G (a) Theoretical transformation temperature y +«. - 1,16S°K.
3.c LJ.G7 = -404,370-6·76TlnT+2·17x I0-3T2+3·80x 105T-1 (b) Theoretical transformation temperature y-+«. - 1,075 ± 3°K.
+ 123·IT(T < 932°K). The calculations of the values are made by the methods shown in .
example 3.5. It is essential that five-figure log tables arc used for the latter
..1G7 = -408,I00-2·64TlnT-0·79x IQ-3T2+3·80x IQ5T-' calculation, as a slide rule is not nearly accurate enough because of the
+ 101 ·8T (T > 932°K).· relatively slow change of free energy with temperature. This gives rise to
values of .1G less than 200 cal mole-1.as the difference between terms of
The diqbs-Helmholtz equation is used. to deduce the free-energy- magnitude of tens of kilocalories.
temperature relation for the oxidation of solid aluminium, applying the · 3.H • Entropy of ordering' of Cu3Au • 2·191 cal deg-1 mole-1•
appropriate standard heat of formation, entropy_~d heat capacity data
. . . . , ' .. ~ . ' The data are plotted as Cp/T versus T (Tin °K) and the area between
~~ .
~·
 184 APPENDIX 1 ,4,.NSWERS TO EXERCISES l8S
· the estimated background curve and the data curve is measured using the f~ ' . : ' ' .-· ..
4.o End products.of separation process arc Zr02, HfC14•
trapezoidal rule formula as elaborated in appendix S (2).
The standard fretf energy changes for the reactions:
3.I Entropy of fusion of hexagonal thallium at 297°C = 1 •99 cal deg-1 ,
mole-1•· , l ~r9~ .+ ~i<i> ~: Zr02,,>+ 2Cl2w .
The hypothetical melting point of Tic••> is calculated (using five-figure
.= Hf02c.a)+2Cl2(J)
log tables) by the method shown in example 3.5. By application of Hess's
law to the heats of the transformation Tlc•• >.=Tl<ft> and the fusion
T:,pi ~ Tl(n at this temperature, the heat, and hence the entropy, of
l
I
: l-J(Gf;(j)+d;~;
·.:.- . ' . '. . .
. . .
are given at 800°C ~nd J;:e used: to calculate the equilibrium constants.
From these values ahd the known gas composition, the partial pressures
.

J,
fusion of Tl(a> is deduced. l. ate
of'the tetra fluorides calculated and the reaction end products deduced.
4.E Activity of iron = 0·827. · ,. . ·
4.A Pcof Pco, = 2·26 at l,027°C. The standard free-energy change for the reaction:
The standard free-energy change for the reaction:
:;Fee,>+H20~i ·~.
Fee,>+CO2<,> ~ FeO<•> + CO,v ,:. . ; . ,: . .
F_eO<i> +'Hu.>:
,. :' . .. ' .. . '
is calculated at 1, 113°K from the data in appendix 3 for the formation
is calculated through Hess's law, from the free-energy changes for the reactions for FeO a;6d H20. The calculated equilibrium constant, with
reactions: the known equilibrium gas composition, yields the activity of iron in the
c<•> + t02ea> ~ co<a> alloy, as the activity of FeO will be unity.
4.F Minimum value of ratio Pcof Pco, = 3·50 X I0-3•
c,,> + 02ea> .= C02ea> From the data for the reactions for formation of Ni 0, CO and CO2 i11
Fe(s)+tOiw ~ FeO(s) appendix 3, the standard free-energy change for the reaction:
which are available in appendix 3. Thus from equation 1.58 the equilib-
rium constant is deduced and, assuming the solid components are pure
Nie,)+ C02(a) .= NiOc,)+ COw .

(a = I), the ratio of the gas partial pressures is calculated. is calculated, yielding the equilibrium constant and thus the gas ratio
4.n Equilibrium is not attained in the reaction chamber. required when the known activityofnickel in the alloy is substituted
The standard free-energy change for the reaction: with aN;o = 1. .
4.G Standard free energy of formation of K.<;l at l,000°K- -80·9
NiCl2c,l + H2(,> .= Nie,>+2HC1<,> kcal mole:". .·. · . I ·
available in appendix 3, is used to calculate the equilibrium constant and From the ratio Pttcif Ptt, given, the ~ta11dardfree-energy change for the
thus the ratio of the gas partial pressures, assuming the solid components reaction.."
arc pure. This ratio is compared with that given by the gas analysis. KClc1>+tHi<a> ~ ~1)+HClea>
4.c Vapour pressure of magnesium= 0.40 mm Hg.
· may be calculated for l,000.°K-, Frum the data of appendix 3 the same
The standard free-energy change for the reaction given is calculated
quantity for the formation 'reaction for HC! is calculated at the same
from the data of appendix 3 for the reactions: temperature. By application of'.Hess's law to the two reactions, the
Sils> + 02<1> ~ Si02c,> required free energy of formation of
KC! from Kand Cl at 1,000°K is
obtained. ·,. ·.
MJ?<a>+!02ea> .= MgO<s> · 4:H Rati~ Ptt,s!Pttfin equilibrium with pure Cu and Cu2S = l ·66 x 10-•.
2Mg0(s) + Si02(s) .= 2Mg0. Si02(1) Activity coeffi6/ent of copper it;t alloy (Ycu) = 0·895. . .
from the data fo;the formation reactions for Cu2S and H2S given in
yielding the equilibrium constant for the reaction through equation 1.58.
appendix 3, the starid~r:d free-energy change for the reaction: .
The solid components arc assumed to be pure and thus the pressure of
magnesium is calculated. ,'H2S'1r+ 2Cuc,> ~ Cu2Sc,> + H2W
l . '. .,·.·.
 '{i
.)

1116 APPENDIX J
' ANSWERS TO EXERCISES 187
is calcul~ted at 500°C, from which, as the solid ~6:iriio~ents are pure, the in appendix 3, the equilibrium constant is obtained as a function of
ratio Pn,slP11, over pure copper ls deduced, The:~ctivity of copper in temperature. The minimum dissociation temperature under the condi-
the alloy is given by this value divided by the equil{~rium ratio over the . tions specified is given by the solution of this relation with:
alloy. The activity coefficient follows from equatiori:1.74.
4.1 Ratio Ptt,s!Ptt, in equilibrium with Ag and Ag~S at 600°C = 0·150. K = Pco, = 0·0015
From the data given, the standard fr~-energy cha~gc for·the reaction:
the solid components being assumed to be pure.
2H2(a) + S2<a> ~ 2H2Sea> 1
A 4.L The oxide in equilibrium with pure iron at room temperature is
magnetite, Fe304• Maximum temperature of this equilibrium
is calculated at 298°K. Thus the equilibri~~.S~nstait: ~ = 567°C.
. ·~· .
2 •;:_;
The standard free-energy changes for the formation reactions of solid
KA == _:_PH,S, :·.·.
'. . ·p2Ha• p·Sa FeO and Fe304 are given in appendix 3. The 'low temperature' oxide is
stable up to the temperature at which both oxides and iron are in equilib-
is deduced. By application of the van't Hoff equation 1.59 (and assuming rium. Thus the free-energy change for the reaction:
;j H independent of temperature), the value of KA at 873°K may be
calculated. In addition, the data given for the reaction: 2Fe0<•> .= tFeJ')4<•>+tFecs>
/ 4A8(1) + S2c,) ~- 2Ag2S(,) is given by subtraction of those for the two formation reactions. This
B
free-energy change is equated to zero to yield the three-phase equilibrium
yields the equilibrium constant: temperature. At lower temperatures 4'JGO becomes negative and hence
the reaction goes from left-to-right to form Fe304 •.
ai~.s - , ···-1
Ks - -·-- 4.M Maximum temperature of Fe304-Fe equilibrium= 771 °C.
tf,.,.ps, Ps, The free-energy equations used in the previous example were derived
if the solids are assumed pure. Subtraction of equations A and B gives partly on the basis of the known equilibrium temperature at which the
the equilibrium between silver, hydrogen sulphide, hydrogen and silver two oxides and the metal coexist. Thus they are consistent with this
sulphide, the equilibrium constant being: ;; temperature. The calculation in this exercise is, of course, subject to
)
errors in the assumption of temperature independent enthalpy and
entropy values in addition to the errors in the values themselves.
4.N Maximum temperature without formation of Mo02 = 920°C.
Hence the gas ratio ii1 this equilibrium is t;alculatedJassuming the solid The free-energy changes for the formation reactions for molybdenum
components are pure... ; oxide and steam in appendix 3 are combined to yield the value for the
4.J The copper cannot.be bright annealed.under these.conditions. required equilibrium. The procedure is identical to 4.K.
The data for the equilibrium: · .. · '-· · · · · 4.o (a) Minimum reduction temperature with 0·3 mm Hg pressure of.
co= I ,363°C.
Cu20c,> .= Cll(,> + 102C1> , ' ..
,, (h) Minimum reduction temperature with 0·001 mm Hg pressure
of co = I ,092°C.
in appendix 3 enables the equilibrium value pf p~~ ·~o :be· calculated as 4.P Activity of FeO at 1,361 °C = 0·473.
2·7 x 10-10atm. Thus J0-6 vol-per cent Qi in nitrogen (10-8 atm) will
oxidize copper at 750°C. · . • · ; .. :, . From the data given and the standard frec-energy change for the
reaction:
4.K Minimum dissociation temperature pf CaCO~J 579°C.
From the standard free-energy change fo.r-the.i:e~r,kion,:
C02(1)+H2c,) .= co(a)+H20(a)
CaC03c,> .= C~Oc,>~COi<a> ;j .
•. . . i, .
calculable from the data of appendix 3, the equilibrium constant for the
reaction between the steam-hydrogen mixture and the iron-iron oxide
.~
. ~
'
·~ ·..
 188 APPENDIX I
. ANSWERS TO EX.ERCISES lit
slag is obtained. Thus'with the known gas mixture, the activity of FeO
is calculated. is calculated, which, as~.umingthe ~6lid components are pure, yields the
4.Q Activity of carbon in the metal sample e 0·43. value of Pso, atequili~riuin_. SJmilarly,.the data for the reaction:
The standard free-energy change for the equilibrium: . . ~l CuO+S03 .=.CuS04
co<a>+3Hw .= CH,4(a)+H20w A ~+•·;;.-,~ is deduced and yi~i'~s (he :~~tue of ~~o.:in this equilibrium. The range of
is calculated from the data in appendix 3. Assuming the partial pressure Pso, for sulphating ofthe 9opper only is between these two values.
of water vapour is small, then from the graphite equilibrium data: 4.s · Approximate eqtiffibriu~ partial 'pressures of AICl at 1,000 and
=
. 2,000°K S ~ -IO{~_ari~f latrn respectively. . ·
PH, = I- (PcH, + Pco) = 0·981. Combination of thef{ree-<friergy.data for the formation reactions for
Further, from the data of appendix 3 for the reaction: AICl and AIC13.inapp~nd~iyields the data for the equilibrium:
C<•>+2H2<a> ~ CH4<1> B ~Alc;).·+AIClJ<i> .= 3A1Clw
applied to the same graphite equilibrium, the equilibrium constant: frorri which the valuesitt~e\atidpi,~JPAICI~ are calculated for the two
temperatures. · ·. j' '. ':, •... ·;: · ' · .·' . · . ·' · ,
K8 = Pct• ( • 0·0069) It cannot be iminedialc!yiassunied that (PA1¢1 + PAIClt) - 1, for chlorine
PH,
gas will be present in the. system, However, consideration 9f either of the
is calculated, from which PcH, ••• 0·0066, and hence Pco = 0·0124. Thus
from: formatlon reaction equilibr'i~ shows that Pei. will be yery small at both
1,000 and 2.000~K. andJhesum of the.chloride paitjalpressures may be
KA - PcH,·P~,o(- 5·13 x IO-') taken as I atmosphere}The individual values of thes~ partial pressures
Pco·PH, · are calculated by combfnation of this fact With the gas ~atio vaiues. The
the value of Ptt,o = 8·9 x I0-5 is calculated, a constant for both experi- solution of the resulta-ht cubic equation-In PAtci involves 'use of the
ments (and, as originally assumed, a small value). Therefore, for the gas formula method given fn appendix s. ' ' '
equilibrium with the iron, The operation of a refining process by the method of this exercise is
' '

Pck, + Pco = 0·0081 therefore feasible by cycling the gas mixture between two chambers, one
containing impure aluminium at about 750°C, the other at about l,700°C
and hence:
PH,O = 8·9 x 10-5 yielding pure aluminium if'none of the impurities ha~ volatile chlorides;
4.T Value of constante ss I, 750.
PH, - 0·9909.
From the weights of zinc condensed from nitrogen passed over pure
Thus from the value of KA, the ratio: ' zinc and the alloy (Nzn = 0·706), the activity of .zinc in the latter is
calculated. In the equilibrium experiment, the activities of zinc in all the
Pett, = 0·561
Pco alloys will be the same, and hence this value applies to all the alloys: that
and therefore PcH, = 0·00291. Hence: is, the composition/temperature data of'the table represent an iso-activity
curve for this system. By plotting the temperature versus: .
- Pk. (l-Nzn)2
Oc = KaPcH,
4·575logyzn'
4.R Range of values of Pso. = 6 x 10-4 to 0·3 atm.
From the data given and that for the equilibria between iron, its the linear relation demonstrates the conformity or the d~ta with the
various oxides and oxygen in appendix 3, the standard frec-energy relation given, the slope' 'of the line giving the constant IX.
change for the reaction: " 4.u Pressures of molecular species ~2• S4, S6, S8, are 0· 166, 0· 168, 0·822
and 0·892 atrn·.respeciivelyat-800°K. The pressure of monatomic
Fe203<•> + 3S0.)(a> .= Fe2(S04)J<•> sulphur is ncgligi~),Y small.
 ·1

190 APPENDIX 1
ANSWERS TO EXERCISES Ul
From the standard free-energy ~hangc'data. t-r
appendix 3, the ratios
of example 5.4, involving the free energies of mixing, or alternatively, the
of the pressures of the various species are ~!~\dated. After the appr<r
method of example 5.5, involving activity coefficients. The latter are
priate substitution of these into the relation expressing the sum of the given by:
pressures (,.;, 2·04 atm), ·the. resulting quartic equation in S2 is solved .. GM
graphically. . logyM1 = logaM1-logNM, =
4.5;;T-logN~.
4.v Gas composition at regenerator outlet: 3·4 per cent CO2, 27·1 per
cent CO, 3· l per cent H20, 11 •9 per cent H2, S4•S per cent N2• Either method yields the same value of G-t{, for the equi-atomic solution.
From the data of appendix 3, the equilibrium constant for the reaction: Values of Gjb and Gt1 follow, for:

H20w+COw ¢ C02(a)+ H2(a)

Cf = Gi -Cjdeal = G{"-G{" Ideal
and:
is calculated at l,373·K. Consideration of the conservation of the Gt' ide11t = RTln Ni.
number of atonis of each element in the final gas analysis, in relation to The integral excess molar free energy follows from equation 1.42.
the initial analysis, leads to expression of thi~,cquilibrium constant in 5.r Partial molar volume of mixing of tin -0·06 ± O·OOS cc mole-1•
terms of one component only. Thus, from the initial analysis, the number This quantity is obtained by application of equation 1.47 to the .
of atoms of carbon, hydrogen and oxygen are .';30·5, 30 and 37 per 100 volumes of mixing, as in Fig. la.
mole of gas respectively. Then, if the finalgas mixture contains X mole 5.1 Integral molar entropy of mixing of solution• 1 ·39 cal mole-1
of CO2, the nurpber of mole of CO, H20 and H2 arc deduced to be deg-1•
(30·5--,X), (6·5-X) and (8·5+X) respectively. Substitution of these S.K Integral free energy of mixing of solution= -554 cal moJe-1•
quantitiesjn·the
: ...
·, '
exprcssionfor
. . .
the equilibrium
. . ~constant gives X. As the solution is regular, logyJ(l -Ni)2 = constant, independent of
the component. Thus the constant is determined from the data given for
5.A Latent heat, of vaporization Of alloy ...; 44·~ kcal mole-t.
Hg, then Ysn is calculated at the required composition. Values of Gtf,
As the alloys are ideal, the vapour pressure o( a component i is given and Gt:, follow through equation I .54.
by p1 = N;pp. The latent heat ofthe alloy will.~ the weighted mean (in S.L Vapour pressure of iron =0·083 mm Hg.
terms of the partial 'pressures) of the latent heats of the components. The data for Gf, is used to calculate GJ0 in the alloy and hence YFe
5.B Heat evolved by 1,000 litre of gas blowrj through ideal Fe-Mn from equation /.79. The value of aFe then gives the vapour pressure of
alloy= 4,920 cal.
iron over the alloy from the value over pure iron.
Heat evolved by J.,()()() litre of .gas blown \hrough regular Fe-Mn 5.M Vapour pressure of copper= 1·41 x I0-3 mm Hg.
· alloy = 4,960 cal. · · ? .
· ·. · . ·. · ..
The relation for Yzn enables Ycu to be calculated directly (p. 107).
The answers are given by the methods elaborated in examples 5.2 and 5.N logy Ac = 0·491Ntu•
5.3. ,.
The regular behaviour allows use of equation J.82.
5.c Vapour pressure of manganese ~t 1,600~<;:J= 0·0491 atm. S.o Excess integral molar free energy given by:
Manganese will be lost from the stee].·expos,edto a vacuum of O· I
mm Hg. · ' · . i: ... • GE= 3,339NculnNcu-5,350Ncu(l-Nc.J
5.D Density of resulting alloy= 6·86 g ci:~1 •.. ) . + 509(1 -Neu) In (I -Neu)
5.t Partial molar heat of mixing.of bismuth e, io1 cal mole-1• The data given yield Cf,, which, through the Gibbs-Duhem relation,
The method of example 5.4 is.applied to the data given.· gives C~~ as a function of composition. The expressions for GI.a and ~u
5.F Activity of iron = 0·695. · , ' follow from:
The method of example 5.7 is applied tothe di+,ta given. Gf = Ct'-C('1idcat.
5.G Activity of mercury= 0·874. · · ·: . · . 'J · · ..
5.P Excess partial molar free energy = - I I 7 cal mole+',
5.H Excess integral molar free energy of solution= -1,746 cal moJe-1,
From HM data given, Bf'e calculated through equation 1.46. As the
The value of Gt{, is calculated, from the data given, by either the method solution is regular, equation 1.83 yields Gt•.
192 APPENDIX 1
ANSWtRS TO EXERCISES 193
5.Q Extent of lead purification is Neu• 4·26 x 10-2. 6.e Concentration 'of copper in equilibrium with Pb, PbS and
The activities of the liquid metals are calculated in terms of Neu from Cu2S = 99·76 atom-per cent;
the data given, and substituted in the ~prossion for the reaction equilib- From data for Cu2S and PbS formation reactions in appendix 3, the
rium constant. equilibrium constant for the reaction: ·
5.R Excess partial molar free energy of iron= -68 cal. . '
The activity coefficient of iron is calculated from the equilibrium ' 2Cu(,)+ PbS(I) ~ PbS,i) + Cu2Sc,)
constant for the reaction: is calculated. as. 4'.0 )( ,10-8.=· '!Pb/~~;' Henry's law appl,ies for As .
Fec,>+H20C1>,... Fc0c,>+H2CI>. Cu(Ncu < 0·004) the~: Pb: is i4.eal .over this composition range (p. 17) •
The value of Gfe follows from equation J .79. Further, as ac~.""': l ;\~hen N~u =i"0;004 (as activity of Cu at solubility
5.s Activity coefficient of cadmium - l · l S6. llmit e activit{o(p·J.fe ioH&ct:ipper) then i'iu = 2S0. Substituting in
Partial molar free energy of cadmiwn - -16·36 kcal mole+'. equilibrium constant ~quation, then: .
••• 'J :·.:.: ·:-.,- •• :'·: ./:

b X,. RT 1n<a..J1 : , { (l....:.N).. . .

M.
5 .T H T, •
IS given y 1- x,..
1 x,. . .... ·. \: (250N Cul ~}·.O.x 10-s,
. .. ,, .. · .... Cu.·.
From data, y,. at T1 calculated and, as the solution is regular:
RTlny· .
a q~adratic'in Neu, sJi~ed by the f6rm;ula method. . . .
( _ N} = constant, independent of component. 6.c Change in free e~:e~gy ;for tf~nsfer of standard state • 9• 11 T cal.
1
For the transfer _cf i}~:~~ardstate: ·
Thus i'A and i'B may be calculated at 12 and and n~
deduced from nr
equation 1.82 for regular solution. The quantity HM, independent of Mn(lf ": ¥0.(l;wt X diluic); .
temperature, follows. · AGoi: ..>~-; .. ,: .: , . . (P:
. , . f;,; Yi Mn,!, iN!Xdilute)- (~n, I, pure)
5.u Indication of non-regular solution behaviour.
Method of obtaining GE from data given:
.. " f : '- .· . : . · .. o:558s : ,
The vapour pressure measurements will yield as = /1(N8) from which 'y=/~·57~Tkig 54·94· •.
the non-conformity of YB = /2(N8) with relation: 6.i> · Change in.free ~ri.~r~y for'.t~an$ferof standard. state ·
In YB = «{l -N8)2 . . }~ - 7"s,s~:,11.circa1.
indicates the non-regular behaviour of the solutions.
For the transfer of siandard·s.tate:.:·
From i'B = fi(NB), i',. = /J(N8) may be calculated through the Gibbs-
Duhem relation. Thus Gf. and Gf are deduced from equation 1.79 and Vu>~ Vu.~Y.di:ui~> •.
GE obtained as a function of composition.
;_ .. ' · . . .• -. · - 0·5585
Jc;? = 4·5?5 X l",873 X log;,0+4·5757:'log 50.95
6.A Oxygen concentration in nickel at equilibrium ••• 8·2 x IQ-31 ,! • •
weight-per cent. = 10,000-8·97T.
From the data for the formation reactions for Zr02 and NiO, the :::
equilibrium partial pressures of oxygen over the metals and oxides are But vanadium is so!idli:t l,600°C, melting at 1,860°C with a latent heat
calculated as l ·3 x 10-43 and 3·0 x I0-14 respectively. Thus at any of 4·5 kcal mole=! (appendix 3). Thus:
pressure of oxygen less than that at equilibrium: AG,..; 4,S00-2·11T.
Po. = Oo(laNI) ••• %0 Hence:
Po. (u.tn.) ao (In NI auacn.) O·O 19 JG0 .. = <i'?v.1, wt Xdilute)7"' qc'(V, ,; pure)

as oxygen dissolved in nickel conforms to Henry's law. ••• A<;?+AG,.

 --1

194 APPENDIX 1
ANSWERS TO EXERCISES 195
6.E Equilibrium constant K, =·2·4 x 1014• Hence:
The free-energy change for the transfer of standard state: .dGo = -31,700+ 14·64Tcal.
B(F.C.C. pure) 'F ~.c.c. dilute, 2 atom~ In A) The quantity required is obtained by subtraction of this from the
is given by:
expression for the free energy of formation of steam from hydrogen and
.d G0 = RTln 2y0 oxygen gas.
where y0, the Henrian activity coefficient relative to pure B as standard 6.H Activity of carbon relative to wt% standard state - 0·57.
state, is given as 0·24. Further, for the transformation: The logarithm of the activity coefficient of carbon (logyc) is plotted
B(B.C.C, PW'•) ~ B(F.C.C. pureii
versus Ne and extrapolated to Ne= 0 to yield ri.
Thus, from equation
1.94:
· .dG, • 220-/:~J;T. r
Yi = 1,243 = OC(pure) • wt%
Oc(dltute) Ne
C.

The standard free-energy c.bangc for the. reacti<>n given· is obtained by

As at wt% C = 0·5, Ne ••• 0·0256, .substitution of the appropriate value
application of Hess's law to these ~ction.s and that for tbe oxidation of of OC(pure) from the table yields OC(diluto) as required.
pure B. The equilibrium constant follows. -: ...• i ', . . .·-·
6.1 Interaction coefficient fi = 0·57. Interaction parametere~ - -0·24.
6.F Standard free-energy change;,,; ....;.•47,1·52 cat inole~1:
From the free-energy change for the standard state transfer given, and
The method is sfmilar to the previous exairipie;Jhe free-energy change that for the steam formation reaction, the equilibrium constant for the
for the transfer in standard state:;.·. . · · · · 3. reaction:
',· . . ; . ._ .. : . ;:-'>~i :
B(pure: I) ~ Bcd1i~,.; •. , -in~; tn'·A) · ; H2cai + Ocwt X !11 Fe) ? H20w
being given by: ( may be calculated. Thus:
i100 - RTlri_1·5;.yt1 !~{,
. ·- . -·. :l a0 = PH,o = 0·0188 .
and for the fusion: . ,, ·•. -~:
PH •. K, .
B(plll'e, aolid) •~- B~ e, liquidH Since this dissolved oxygen obeys Henry's law, then in the binary
. . . . ~
· '· · ··. 2·600 Fe-0 alloy, wt% 0 = 0·0188 at a0 = 0·0188. Hence:
.d 61 =. ~ 2•~
.
":"" I ''0J5
. .
r. V 0•0188
Combination of these with the oxidation reactioh for pure B yields the
fo = -----
0·033
standard free-energy change required.v, ,\.;·.' . i · · and:
6.G Free-en~tgy change ..dG0 = -:~~.2C>0;1:54t:ca,L_ V Jogfti
The data given are for the reaction: : .." : . i, eo= --·
. ,; . ~,--
wt%V
6.1 Interaction parameter e~ = 0· 1 I.
for which:
Combination of the standard state transfer given and the formation
K . PH10
---z: PH,O reaction for H2S yields the reaction:
' PH,.ao PH,•/o. %0
H2c,>+ScwtX!oF•> ~ H2Sw
This equilibrium constant can ~- calculate~ for the three temperatures .
and found to conform to the relation: · for which the equilibrium constant:
· 6,930
logK, - ~-,.3·20. K = PH1S
' PH,,fs, %S
\. ~
 . .' :. :, ~ •.
196 APPENDIX J. f . ANSWERS TO EXERCISES 191
is calculated as 2·58 x I0-3 at 1,873°K. Thus fs is calculated by insertion 7.D · Activity coeffici~nt of cadmium in the alloy - l ·Sl. · . . .
of the data given, and the required quantity ei obtained by application is
. The potential of tlie composition cell used to calculate the partial
of equation 1.105; molar free energy ofjmixing of cadmium in the alloy by the method
logfs = %S,rs+ %C.ei. shown in example 7 .2> .The activity of cadmium follows from equation
6.K (a) Oxygen content = 0·003 weight-per cent. 1.54. :::. .
(b) Oxygen content= 0·024 weight-per cent. 7.B Partial molar enthalpy of mixing of cadmium-1,486 cal mole-1,
Excess partial molar free energy ofcadm,h,im-1,092 cal mole-1•
From the data for the standard state transfer reaction and the formation
Excess partial molar- entropy of cadmium =0·4 cal deg-1 mote-1 •
. reaction for MgO, the equilibrium constant for the reaction:
The values of Btr,, and St{,, are calculated from the e.m.f. data of the
M8<a)+Ocwt,t.lDPe)?MgO(,) composition cell using equations 7.e and 7.d respectively. Calculation of
is calculated. Thus as: C~ in the alloy yields ac.1, .The activity coefficient of cadmium enables
1 Ced to be calculated through equation !;79. The value of S& is given by
K,- PMa· ~o.wt %- o
the difference between Sc": and StJ for an ideal solution (- -RlnNc,s).
7.P Activity of aluminium= 0·678.
and fo is given, the wt% 0 can be calculated for the two values of PM,· Activity coefficient of aluminium .••. 4·07.
6.L Carbon and phosphorus are not in equilibrium. Partial molar. ciliropy of mixing· of aluminium - 4· 33 cal deg-1
The values of a0 in equilibrium with phosphorus and carbon are I _( =:·.-
mo e : . .··; . . . , .
calculated and compared. Thus in equilibrium with carbon: Partial molar enthalpy of mixing of aluminium = 4,176 cal mole='.
Partial mol~r trei~n~tgy\;r'fuixing of aluminium - 902 cal mole-1•
K = Pco ••• Pco
' ,._ "~ - "c.wt%C.aoec, Partial molar enthalpy of aluminium = 12,649 cal mole-1.
;-: f . ' ., . ·•_.. ,,
7.o Vapour pressure,tof·zinc·over alloy= 10·64 mm Hg.

· . · :· r· , . · ·. at . ·.
However, by insertion of the composition and interaction parameter
data given into equation 1.105, fc is calculated and thus aoec, is obtained. / ·
,/ For the vaporizatiotof pure zinc 873°K:
o
A similar calculation yields aoeP>· As this latter value is much larger than ( f/
• .1(/0 +.'Z,;270~1- '.-RTln[~]·
aO<C>• the phosphorus and carbon cannot be in equilibrium. , . ~ '. : . ·:< : , :· . . OZn(I) .
However, as Ozn(I> ~ itf?( pure zinc, then Pin is qatculable. From the
7.A Standard entropy of AgCl • 23·0 cal deg-I mole-1. ·or
e.m.f. data, the activifa, zinc in the· alloy is calculated and thus the
Standard enthalpy of AgCl-= -30,300 cal mole-I. partial pressure of zin~ over the alioy is given by: . ·
Standard free energy of AgCl ••• -37,150 cal mole-I.
._ : if}(:;:}p~)= p~az,;. · .
The values are calculated by the method of example 7.1 using the 7.H Activity coefficient of zincin' (/fir,=· 0; 173) alloy - 1 ·80.
standard entropies of the components in appendix 3. For the first com,pJ~i{i~'riiceil:..Gl:n--G'~ - ~zFE1• For the second
7.B Partial pressure of chlorine w l-d x 10-ss. composition cell: G~" ,;.,Cit~-~ ·':""zFE11: As
free energy is a state variable,
The standard free-energy change for the cell reaction: these values' are·additiyfand so:;:..:. i . ; . . . .
Pb<•>+ Cl2w ? PbC!2<•> lf .',-,o. .
G zn-Pzn.,":'· {GZn
MJII ': . . . F(EI £")..
.·;::: -z + ... · '
is calculated from the potential of the cell, the latter being given by the' is
H~nce a~n(N~n = 0· 17j) calcul~t~d.. .
difference between the oxygen potentials of the two electrodes. From 7.I E.m.f. generated ~y K~~urc)is'~(pure) cell= 0·869 volt.
this, the equilibrium constant yields the required partial pressure. E.m.f. generated ~iA~i,~rc>/Sn<,u~r> cell'= 0•851 volt.:
7.c Minimum potential at 500°C- 0·626 volt. · As the hydrogen electrode at pH.= I atm is the arbitrary zero potential,
From the standard free-energy change for the formation reaction for · the cell reaction: '.! · · · · · · ·
SnCl 2, the potential of the cell is calculated as -0·626 volt. ;,_!Sn+Ag+ ~ tSn+++Ag
.·,
14
 198 APPENDIX I
ANSWERS TO EXERCISES 199
will generate a voltage:
7 .P Partial molar heat of mixing of nickel = - 21,760 cal molc-1•
E~ !Eso..;.E~ For a solute in the composition range where Henry's law holds,
This will be 'the standard potential (.E") for the ~II when all the com- equation 7.k becomes (in the present context):
ponents are ii:i their standard states.· When a,.;· in the Sn electrode is O· S,
the relation 7.2.4 is applicable. Thus: · ; a1n NNI [RN,-H~ul Rlfi
. .'
·. '.·,'RT·. 1·:'
.i ar " - RT2 • RT2 •
~oa1101> - eo __
F tn·:m·; , Hence, when the log of the atom fraction is a reciprocal functioq of the
. Z ·,'. Dlii \
absolute temperature:
7.1 Maximum e.m.f. generated - 8·26 millivQlt.·i
7.K Value of logap •• ~/aeu •• •, 26·9: Thqs as a~•• i: ~ 1, then acu++ must I [NNi<92l>] Rlfi ( 1 1 )
be very small, and 'the concentration DlUSt be:Cbfl1Spondingly low. og NNi(773) • 4•575 923-773 •
By consideration of the two half cells, (bo'th\written as reduction
reactions): · · · · · _-; ; ·., ~. · 7.Q Heat of solution of solid cobalt= 11,950 cal mole-1•
The heat of solution of liquid cobalt is calculated, as in 7.P, from
Fe+++2e ~ Fe : Ep.·-:·:~·.--
· , ~ R1t;
zf/n
·r a
0:.:.
.]
equation 7.lc, using the liquidus temperatures and compositions as given
in the table. The heat of solution of solid cobalt is given by this value
Cu++J.2e ~Cu: E~· •• -~;·_ RJ1n[. 0cu·J_ subtracted from the heat of fusion of pure cobalt, assumed to apply
' ' .. •· ', : z.·;.; acu++ at these temperatures.
•. ~-· .
7.R (a) Weight-per cent of ethylene glycol - 36·5.
and as Ocu = Ope = 1 and EFo = Eeu a_t equilibrium; then:'
(b) Weight-per cent of ethylene glycol - 34·2.
;F
o Eo .} i'.
1 og [-aFe+•] .,. - (.E· Pe- Cu •.'.
..
As no activity or liquid us/solidus data are given, it must be assumed
' ~~ . ' ~T...:..:, .;.;~-.,{':· :,j . t~at the mixture is an ideal solution .. Thus equatio~J,:~6J i.~...~PRlia~,
7.L Activity of aluminium in (NA1 ~ 0·0404) alloy, relative to infinitely with Lr assumed (a) constant and (b) calculated as a function or tem-
dilute atom fraction standardstate» 0·0278.:\ · perature using the Kirchhoff equation 1.20.
From the data given, the quantity l6gyAI cari be calculated and plotted 7.s Activity of copper at liquidus temperature - 0·91.
versus NAl to yield :y~ by.extrapolation to NAl"" 0. Then from equation The answer is obtained by the method of example 7.7.
l.88: 7.T Activity of copper in alloy of solidus composition - 0•797•
. QAl(pure) From the activity coefficient data, the activity of (liquid) copper in an
QAl (W'. dilute) ""' '11 alloy ofliquidus composition at 1,273°K may be calculated as 0·735 with
7 .M Graph of' ideal liquid us ' of nickel. · .respect to pure liquid copper as the standard state. This is equal to the
The ideal liquidus for any solvent is given by equation 7.6.2 where the activity of solid copper at the solidus with respect to the same standard
latent heat of fusion of the solvent (in this case nickel) is assumed state. From the free-energy data, the activity of pure solid copper with
independent of temperature. · · respect to a standard state of pure liquid copper may also be calculated
7.N Latent heat of fusion of Ag-2,620 cal mole=". for the same temperature. The required activity of copper at the solidus,
The data are applied to equation 7.6.2 with silver as the solvent. with respect to the pure solid copper standard state, is given by the ratio
7 .o . First solid separating at 300°C contains 3 atom-per cent Zn. of these two quantities.
Equation 7.6.2 is applied to the present system showing solid solu- 7.u Activity of silicon in liquid alloy at 1,289°C - 0·75.
bility. Thus: · · Partial molar entropy of silicon in same alloy-1-17 cal deg-1
N.] · .. · . Lr '( I ' i\ mole-1•
log [::9!Q1 ·= -·-· ~-- .
. . · NCd(•> · 4·57~ .Tr · T J Partial molar enthalpy of silicon in same aUoy - 940 cal mole-1•
7. v Latent heat off usion of iron = 3,600 cal molc-1•
 .. . },,.,
APPENDIX 2 APPENDIX 3.. . :,.:

'·
Table of Constants and Thermochemicar 'Data .
Useful Conversion Factors
Heats of foniladoa, ttamforinadoa apd fmloa, and st,alard •ffflBIM
o Solid, o Liquid, nba~: ... ' ., . . ...
Constants
Avogadro's number N - 6·02 x 1023 mole-t Heat of:. EnJropy of Transformation,<
Boltzmann's constant k - · 1 • 38 x I0-23 joule deg-1 Element formation fonnation ·· orfusion
Electronic charge
Oas constant
e - l ·60 x 10-19 coulomb
R •• Nk ••• 8·314 joule deg-1 mole-1
or
compound
-AH29•
kcal mole"'1
S291
cal deg'"'• molc-'1
.Reaction· . ©
C L1orL,
kcal mo1e-1
= l ·987 cal deg-1 mote-1 (Ag)
Faraday's constant F =- Ne - 96,494 coulomb g-cquiv-1 0 10·2
(AgCl) 30·3 23·0
••• 23,066 cal volt-1 g-equiv=! (Al) 0 6·77 S-+:L 659 2·.S
Gravitational constant g = 980·7 cm sec-2 · <Al203) 400·0 12-2
(Au) .. 0 . 11·32 . S.-+L . 1,063 3·0.S
Conversion Factors <C> 0 I ·361
[CH4) 17-89 44·S
1 atmosphere = 1,033·2 g cm-2 [CO] 26·40 47·3
= 14·7 lb in-2 [CO2] 94·0S SH
I calorie =- 4· 184 joule <Ca) 0 9·9S s•.• s,, 440 0-24
""'4·184x 107 erg S11-+L 8S0 2·1
(CaO) IS1·S . 9·S
= 0·00397 B.Th. U
(Cd) 0 12'3 S-+L 321 1-53
= 41·293 cc-atm [0] 0 sn .. I
I coulomb = 2·78 x IO-" ampere-hr <Co> .0 7'18 S-+L 1,495 3·7.S
1 electron-volt = l ·60 x 10-19 joule (Cr) 0 5-68 .S-+L 1,8S0 4·6
I electron-volt mclecule=! =- 23·05 kcal mole-1 (Cr203) 270·0 ,,! · 19,4'
1 gram-molecular volume = 22·4 litre at N.T.P. (Cu) · .9· A 0 71'}7 S-+L 1,083 3-1
(Cu20) 40·0 ··22:45
I cubic foot = 28·3 litre (Fe) 0 ,::.:' 6·49 s•.• s, 760 0·66
1 pound = 453·6 g mag, RQD• (~)
1 mile = l ·609 kilometre 'f.. . · mag.
60 m.p.h. = 88 ft sec-1 7~
,'r s•.• s,. 910 0-22
loge (In) = 2·303log1o non-(
mag. a,)
Rloge Un) == 4·S75log10 S,.-+S, 1,401 0-21
Se-+L 1,S39 3·7
(FcO) 63·2 .,:,. 14·0S S-+L 1,378 7·4
(Fe304) 266·9 -36·2
[H2] 0 . , 31·2
[HCI] 22·0
' ' {-i{~~~:~f ' ~-

,!, '
'·
 202 APPENDIX 3 '; . 't-~
,,·, t THERMOCHEMICAL DATA 203
( > Solid, () Liquid, [] Gu.
'. Element or Temperature
Heat of Entropy or.
. Trwfinmtion I compound a b >< 101 ex 10-s . ran,c, OK
Element formation - formation _,
or fusion
·I·I
r ;

or -iJH29a - Sita ., Li or Li (Ca)11 . 1-SO

compound oc 7·74 2·$ 713-1,123
kcal mole-1 cal deg-1 mo)e-1 Reacti?,D . kcal moJe-1 (Ca)
(CaO)
7·4 - - 1,123-1,220
".I'. i
I
11·86 1·08 -1·66 298-1,177
(H20) S-.L - 0 1·436 i (Cr) 5-84 2·36 -0·88 298-2.123
(HzO) 68·32· 16·75 1 (Cr) 9·40 - - 2,123-
[H20]
(Mn)
51·80
0
4H
7·6 s.-.s.
. ·s.s,,-..s.;_
,) . -s;;
1
'(i
120
1,100
0·48
0·55
(Cr201)
(Cu) .
(Cu)
28·53
5-41
7-50
2·20
1·50
-
-3·74
-
-
3».1,800
298-1,356
1,356-1,600
1,136 - 0·43 (Cu20) 14·90 5·10 - 298-1,200
(Na) 0 12·3 s-L' 9_7-8 0-63 [ (Fo),_ 4·18 5·92 - 273-1,033
(NaO) · - 98·6 17·4 (Fe)-. non-ma1.(0J 9·0 - - 1,033-1,183
(Na20) J0()-7 _ 17-0 • (Fe),. 1·84 4·66 -.
- - 1,183-1,674
(Ni) 0 7·12 S-+L l,4S5 4·22 (Fe)., 10·5 - 1,674-1,812
(02]
(Pb)
0 49-02 (Fe) cl l
(FeO)
10·0
11.-66
- - 1,812-1,873
l5·S S-+-L 327 r-is 2·00 -0-67 298-1,651
(PbO) 5~-4 16·2 (FeO) 16·30 - - 1,651-1,800
(Si> 0 4·S S-+-L 1,420 12·1 (Fe304)a 21·88 48·2 - 298-900
(Ti) 0 7·3 s.-..s.
S11-+-L
882
1,660
0·83 (Fe104)11
[H2]
48·0
6·52
-
0·78
-
0-12
900-1,800
4·!1 298-3,000
(Tl) 0 1S·4 I
s.-s. 234 0-09 [H20] 7·17 2·56 0-08 298-2,SOO
S11-+L 304 1·03 (Mn).
(Mn>11
. 5·16 3·81 - 298-993
(V) 0 7-0 S-+L 1,860 4·5 8·33 0·66 - 993-1,373
(Zn) 0 9·9S s-r. 419·5 J,74 <Mn),. 10·70 - - 1,373-1,410
(Zn02) 99-5 25·9 (Mn)., 11-30 - - 1,41~1,Sl7
[N2J
[NH3]
6·66
7·11
1·02
6·00
-
-0-37
298-2,SOO
298-1,800
[02] 7·16 1·00 -0-40 298-3,000
Heat capacides
(Pb)
(Ti).
5·63
5·28
2-33
2·4
-
-
298-600
298-1,lSS
C,= a+ bT+.cr-2, cal d~-1 moJe-1 - -,
-:' ~
(Ti)11 6·91 - - 1,ISS-1933
(Ti) 8·00 - - 1,933-
Element or ~: - Temperature i
(11). 5·26 3·40 - 298-SOS
compound -a bx 103 . . CX 1()-S . range, °K
<11>11 7-30 - - 50S-S11
I (Tl) 7-SO - - S11-800
<Al) 4·94 2·96 ·
..
. ..;.., 298-932 .
(Zn)
(Zn)
S·3S
1-SO
2·40
- -
-
298-693
693-1.200
(Al)
(Al203)
7·0()
27·38 3·08 .•.
....;
.:,-·s-~
932-1,273
298-1;800
[ZnJ 4·97 - - 298-1,200
(Au) S·66 .. 1·24 .• -:, -'
298-1,336 -
(Au) 7·00 - . . ,. \:,t
. ·,·.- ' 1,336-1;600
(C) 4·10 1-02 -2·20 298-2,300
[CO] 6·79 . _ 0·98. - · -nn 298-2,500 l
[CO2] lO·SS 2·16 -:-.2·9.4 298-2,SOO
(Ca)a 5·31 3·33 ;a{ 273-713
·.;~· .
. :~
'{
.
. ,i
:' . L., £/
204 APPENDIX 3

o u

~~- .,..
·'-+°
~ r--: .
I I I I I I I M\0\0 0
M
,:..I
·"P· ~. <i> ""
I I I I I I 1 · I I -,. I'- I I 1· .•. IM' I I..-""
I I I I, I + I .I I

-e <
 .'
APPENDIX 4
APPENDIX 5

Atomic Weights of Stable Elements·

. ' ' . . . Mathematical Operations

Actinium Ac 227 There arc various mathematical operations used in this book with which
Aluminium Neodymium Nd 144·27
Al 26-98 Neon Ne some students may not be familiar, or which they find not immediately
Antimony Sp 121-76 20·18
Argon Nickel Ni 58-71 obvious in the context of the examples and exercises. To clarify their use
A 39•94
Arsenic As 74·91
Niobium Nb 92·91 and for the 1ake or completeness, these operations arc here elaborated as
Barium Nitrogen N 14·01 they are used in the book.
Ba · 137·36 Osmium Os
Beryllium Be 9·01 190·2
Bismuth Oxygen 0 16·000
Bi 209·00 Palladium Pd
Boron B 10·82 Phosphorus P
106·7 l. Some fundamental Jaws or partial dlfl'erentlatloa
Bromine Br 79·92 30·98
Cadmium Platinum Pt 195·09 If the quantity z = /(x, y, u ... ) then:
Calcium §1 112·41
40-08
Polonium
Potassium ·
Po
K
210
-oz) ,dx+ ("z)
Carbon 39·10
Cerium.
Caesium
C
Ce
12·01
140•13
Praseodymium
Protactinium:
Pr
Pa
140·92
231
dz= ( ox ,...... oy
au ,,,,,,du.....•
- ,dy+ (oz)
11,11...
Cs 132'.91 .Radium Ra
Chlorine a. 226-05
Chromium er·
35·46 ·
52·01
Radon
Rhenium ;
Rn
Re
222 where (oz/ox),. 11 ••• is the first partial derivative of z with respect to x,
Cobalt Co'. 58·94 186·22 while y, u ... remain constant. This relation is widely used for expression
Copper Rhodium· '" Rh 102·91
Cu 63·S4 Rubidium ; Rb of the incremental changes in the thermodynamic state variables of a
Dysprosium ·Dy·: . 162·51 85·48
Erbium Rl!thcnium :: Ru 101-1 system, which are, of course, functions of the other state variables (e.g.
Er' 167·27 ·Samarium · Sm
Europium Eu 152·0 150·35 chapter I, page 7, derivation of 1.43).
Sc.a.odium . · Sc 44·96
Fluorine F 19·00 Selenium Sc Further if:
Gadolinium Gd 157·26 78·96
Gallium Silicon Si 28·09 z = fi.(x,y)
Ga 69·72 Silver Ag and:
Germanium Ge 72·60: 107-88
Gold Sodium Na 22·99
Au 197·0 Strontiµui Sr v = f2(x,y)
Hafnium Hf 178·58 . 87·63 then, of course:
Sulphur ·, S
Helium
Holmium
He 4·00 Tantalum - , Ta
32·07
180·95 Z = J;(fl,X)
Ho 164·94 Tcll~um_. .,. : ;·; Tc
Hydrogen H 1·008 127·61
Terbium · ·. ,, Th 158·93
Indium In 114·82 Thallium . ·};. · TI z = /.(t1,y)
Iodine 204·39 and the identities:
I 126·91 Thorium · · ·:f _ Th • 232·05
Iridium Ir 192·2
Iron Fe SS·8S Tltulium , ". Tm 168·94
Krypton Tin '.\ Sn 118·70 (:;),, = (;;),,·(!;),,
Kr 83-8 · Titanium · · - ; t; Ti 47·90
Lanthanum La 138·92 Tungsten · · w.
Lead Pb 183-86
Lithium
207·21 Uranium - :~ :u 238·07
Li 6·94 Vanadium ::, V .. (!;),, = (~),,·(~),,
Lutetium Lu 50·95
174·99 Xenon 1 Xe 131·30
Magnesium Mg 24·32 Ytte:b.iwn. ~ .Y~, arc valid, In addition, as:
Manganese Mo 173·04 -
54·94 Yttrium . ·Y··. 88·92
Mercury Hg 200·61
Molybdenum Mo 9S·95
Zinc
Zirconium.
.~
·
'zr{
U·
65·38
91·22 dz= (!;),, dx+(:;t dy,
 ··~
20'1 APP.E.NDIX s
MATHEMATICAL OPERATIONS 209
then dividing throughout by dy yields:
of n ! terms, the sign of any given term being plus or minus as the number
dz (
dy -
az) dx (az)
ax ., dy + ay 1t
of inversions from the natural order" of the numbers 1, 2, 3, ... , n in the
sequence i, j, k, ... , I is even or odd respectively.
For example, the third-order determinant:
which will be applicable for all values of dz/dy and hence when dz/dy = O.
Thus:
811 812 813
· (az)
ay 1t - -
( ax
az) , . (ax)
ay : 821 822 823
831 832 833
Such identities between the partial differentials of the state variables are is given by:
repeatedly applied in the development of the relations used in thermo-
dynamics (e.g. in examples 8.1, 8.2 and 8.3).
I I I
a11 a22 a23 -a12 a21 a23 +a13 a21 •221
832 833 831 833
I I 831 832
2. Trapezoidal rule for measurement or areas llllder graphs = a11(a22 an -a23 a32)-a12(a21 a33-a23 ll31) +a13(a21 a32-a22a31).
If Yo, Yi, Yi, •.• , Ya arc the lengths of a series of equally spaced ordinates Further, the solutions of a series of linear equations:
to a curve and h is their distance apart, the area enclosed by the curve
between the ordinates Yo and Yn is given by:

A - h[!<Yo+YJ+Y1+.Y2+ ... +yn_iJ. arc given by:

This is the trapezoidal rule which, in effect, subdivides the area under a If

curve into a series of adjacent trapezia, of constant width, the latter being
chosen such that the curve approximates to a straight line over the length where Ci, C2, etc., are the determinant D when the elements of the first,
that forms the end of any trapezium. The sum of the areas of these second, etc., column are replaced by Ci, c2, etc., respectively.
trapezia is then given by the formula or, equivalently, by the sum of the For example, the solutions to the relations:
mean heights of the trapezia multiplied by their width.
2x1 +x2-2X3 = -6
3. Application of determinants to the solution or simultaneous linear
X1+X2+X3 = 2
I
equations
A determinant of order n: -X1 -2x2+3X3 = 12
are given by:
a11 a12
a21 a22
a1,,
a2,,
1-6 -21
! I~ -2
I 11
3 8
.
D= X1 I
= i
- ..
2 ···--.:::-2 = s= l,
'
·'
, _; -2
I I
31
I

a,,1 a.a ••• a,., • The number of inversions in a term is the total number o( pairs of numben in the
is defined as the sum: sequence which arc inverted from their natural order. For example, for the fifth-order
term (43521), the inversions arc 43, 42, 41, 32, 31, $2, $1 and 21 -that is, ei&ht Inver,
sions which, being an even number, would make this term pc,sitive. On the other hand,
the third-order term (321) will be negative, as the number of inveniona ia three.
 / .;

210 MATHEMATICAL OPERATIONS MATHEMATICAL OPERATIONS 211

2 -6 -2 lf:
1 2 1
b2 + O 1 · · . 1 d · h th · ·
4 27 > , one so ution rs rca an t e o ers unagmary,
al
-1 12 31 ·-- -16
= I- 2
X2

. I
1
1
-2 8 - -2,
!:,2 + al = 0, all solutions are real and at least two arc equal,
4 27
-I -2 3
b2 3 ·
+ ;1 < 0, all solutions are real and unequal.
2 I -6 4
I 1 2 The solutions of the quartic equation:
-1 -2
X3 = ! 12, .• 3. x4+ax2+bx+c = 0
2 I -2
1 l 1 may be obtained by a formula method, but, as the latter is tedious, a
-1 -2 3 simpler method is the graphical one which, of course, may be used to
deduce the solutions for any equation. Thus, by plotting the expression
on the left-hand side of the equation as a function of x, the intersections
4. Solution of qu~tic, cubic and quartic algebraic equaifons of the curve with the x-axis gives the solutions of the equation. By
increasing the scale of the graph in the region of a solution, the latter may
The solutions of the quadratic:
be obtained with any desired accuracy.
ax2+bx+c - 0
arc given by:
-b± y'(b2-4ac)
X = . .
2a
If:
b2 - 4ac > 0, the solutions arc real and unequal,

b2-4ac = 0, the solutions arc real and equal,

b2 -4ac <.0, the solutions are imaginary.

The solutions of the cubic:

x3+ax+b - 0 .
y'

are given by:

Xi""" A+B

where:
(
i2 = - .![-b
l, A = :/ T + J(b24 + 27
al)] • B - ;/ _/(-b
T- J(b2 al)J .
4 + 27

\l\