Howard Devoe Thermodynamic and Chemical

THERMODYNAMICS
AND
CHEMISTRY
SECOND EDITION
HOWARD DEVOE
Thermodynamics
and Chemistry
Second Edition
Version 4, March 2012
Howard DeVoe
Associate Professor of Chemistry Emeritus
University of Maryland, College Park, Maryland
The rst edition of this book was previously published by Pearson Education, Inc. It was
copyright 2001 by Prentice-Hall, Inc.
The second edition, version 4 is copyright 2012 by Howard DeVoe.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs

License, whose full text is at
http://creativecommons.org/licenses/by-nc-nd/3.0
You are free to read, store, copy and print the PDF le for personal use. You are not allowed
to alter, transform, or build upon this work, or to sell it or use it for any commercial purpose
whatsoever, without the written consent of the copyright holder.
A
The book was typeset using the LTEX typesetting system and the memoir class. Most of
the gures were produced with PSTricks, a related software program. The fonts are Adobe
Times, MathTime, Helvetica, and Computer Modern Typewriter.
I thank the Department of Chemistry and Biochemistry, University of Maryland, College

Park, Maryland (http://www.chem.umd.edu) for hosting the Web site for this book. The
most recent version can always be found online at
http://www.chem.umd.edu/thermobook
If you are a faculty member of a chemistry or related department of a college or university, you may send a request to [email protected] for a complete Solutions Manual
in PDF format for your personal use. In order to protect the integrity of the solutions,
requests will be subject to verication of your faculty status and your agreement not to
reproduce or transmit the manual in any form.
S HORT C ONTENTS
Biographical Sketches
15
Preface to the Second Edition
16
From the Preface to the First Edition
17
1 Introduction
19
2 Systems and Their Properties
27
3 The First Law
56
4 The Second Law
102
5 Thermodynamic Potentials
135
6 The Third Law and Cryogenics
150
7 Pure Substances in Single Phases
164
8 Phase Transitions and Equilibria of Pure Substances
193
9 Mixtures
223
10 Electrolyte Solutions
286
11 Reactions and Other Chemical Processes
303
12 Equilibrium Conditions in Multicomponent Systems
367
13 The Phase Rule and Phase Diagrams
419
14 Galvanic Cells
450
Appendix A Denitions of the SI Base Units
471
S HORT C ONTENTS
Appendix B Physical Constants
472
Appendix C Symbols for Physical Quantities
473
Appendix D Miscellaneous Abbreviations and Symbols
477
Appendix E Calculus Review
480
Appendix F Mathematical Properties of State Functions
482
Appendix G Forces, Energy, and Work
487
Appendix H Standard Molar Thermodynamic Properties
505
Appendix I
508
Answers to Selected Problems
Bibliography
512
Index
521
Thermodynamics and Chemistry, second edition, version 4 2012 by Howard DeVoe. Latest version: www.chem.umd.edu/thermobook
C ONTENTS
Biographical Sketches
15
Preface to the Second Edition
16
From the Preface to the First Edition
17
1 Introduction
1.1 Units . . . . . . . . . . . . . . . . . . .
1.1.1 Amount of substance and amount
1.2 Quantity Calculus . . . . . . . . . . . .
1.3 Dimensional Analysis . . . . . . . . . .
Problem . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19
19
21
22
24
26
2 Systems and Their Properties

2.1 The System, Surroundings, and Boundary . . . .
2.1.1 Extensive and intensive properties . . . .
2.2 Phases and Physical States of Matter . . . . . . .
2.2.1 Physical states of matter . . . . . . . . .
2.2.2 Phase coexistence and phase transitions .
2.2.3 Fluids . . . . . . . . . . . . . . . . . . .
2.2.4 The equation of state of a uid . . . . . .
2.2.5 Virial equations of state for pure gases . .
2.2.6 Solids . . . . . . . . . . . . . . . . . . .
2.3 Some Basic Properties and Their Measurement .
2.3.1 Mass . . . . . . . . . . . . . . . . . . .
2.3.2 Volume . . . . . . . . . . . . . . . . . .
2.3.3 Density . . . . . . . . . . . . . . . . . .
2.3.4 Pressure . . . . . . . . . . . . . . . . . .
2.3.5 Temperature . . . . . . . . . . . . . . .
2.4 The State of the System . . . . . . . . . . . . .
2.4.1 State functions and independent variables
2.4.2 An example: state functions of a mixture
2.4.3 More about independent variables . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
27
27
28
30
30
31
32
33
34
36
36
36
37
38
38
40
45
45
46
47
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
C ONTENTS
2.4.4 Equilibrium states . . . . .

2.4.5 Steady states . . . . . . . .
2.5 Processes and Paths . . . . . . . . .
2.6 The Energy of the System . . . . .
2.6.1 Energy and reference frames
2.6.2 Internal energy . . . . . . .
Problems . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
48
50
50
52
53
53
55
3 The First Law

3.1 Heat, Work, and the First Law . . . . . . . . . . . . . .
3.1.1 The concept of thermodynamic work . . . . . .
3.1.2 Work coefcients and work coordinates . . . . .
3.1.3 Heat and work as path functions . . . . . . . . .
3.1.4 Heat and heating . . . . . . . . . . . . . . . . .
3.1.5 Heat capacity . . . . . . . . . . . . . . . . . . .
3.1.6 Thermal energy . . . . . . . . . . . . . . . . . .
3.2 Spontaneous, Reversible, and Irreversible Processes . . .
3.2.1 Reversible processes . . . . . . . . . . . . . . .
3.2.2 Irreversible processes . . . . . . . . . . . . . . .
3.2.3 Purely mechanical processes . . . . . . . . . . .
3.3 Heat Transfer . . . . . . . . . . . . . . . . . . . . . . .
3.3.1 Heating and cooling . . . . . . . . . . . . . . .
3.3.2 Spontaneous phase transitions . . . . . . . . . .
3.4 Deformation Work . . . . . . . . . . . . . . . . . . . .
3.4.1 Gas in a cylinder-and-piston device . . . . . . .
3.4.2 Expansion work of a gas . . . . . . . . . . . . .
3.4.3 Expansion work of an isotropic phase . . . . . .
3.4.4 Generalities . . . . . . . . . . . . . . . . . . . .
3.5 Applications of Expansion Work . . . . . . . . . . . . .
3.5.1 The internal energy of an ideal gas . . . . . . . .
3.5.2 Reversible isothermal expansion of an ideal gas .
3.5.3 Reversible adiabatic expansion of an ideal gas . .
3.5.4 Indicator diagrams . . . . . . . . . . . . . . . .
3.5.5 Spontaneous adiabatic expansion or compression
3.5.6 Free expansion of a gas into a vacuum . . . . . .
3.6 Work in a Gravitational Field . . . . . . . . . . . . . . .
3.7 Shaft Work . . . . . . . . . . . . . . . . . . . . . . . .
3.7.1 Stirring work . . . . . . . . . . . . . . . . . . .
3.7.2 The Joule paddle wheel . . . . . . . . . . . . . .
3.8 Electrical Work . . . . . . . . . . . . . . . . . . . . . .
3.8.1 Electrical work in a circuit . . . . . . . . . . . .
3.8.2 Electrical heating . . . . . . . . . . . . . . . . .
3.8.3 Electrical work with a galvanic cell . . . . . . .
3.9 Irreversible Work and Internal Friction . . . . . . . . . .
3.10 Reversible and Irreversible Processes: Generalities . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
56
56
57
59
60
61
62
62
64
64
66
66
67
67
69
69
70
72
73
74
75
75
75
75
77
78
79
80
82
83
84
86
86
88
89
91
95
97
C ONTENTS
4 The Second Law

4.1 Types of Processes . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Statements of the Second Law . . . . . . . . . . . . . . . . . . . . .
4.3 Concepts Developed with Carnot Engines . . . . . . . . . . . . . . .
4.3.1 Carnot engines and Carnot cycles . . . . . . . . . . . . . . .
4.3.2 The equivalence of the Clausius and KelvinPlanck statements
4.3.3 The efciency of a Carnot engine . . . . . . . . . . . . . . .
4.3.4 Thermodynamic temperature . . . . . . . . . . . . . . . . . .
4.4 Derivation of the Mathematical Statement of the Second Law . . . .
4.4.1 The existence of the entropy function . . . . . . . . . . . . .
4.4.2 Using reversible processes to dene the entropy . . . . . . . .
4.4.3 Some properties of the entropy . . . . . . . . . . . . . . . . .
4.5 Irreversible Processes . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.1 Irreversible adiabatic processes . . . . . . . . . . . . . . . .
4.5.2 Irreversible processes in general . . . . . . . . . . . . . . . .
4.6 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.1 Reversible heating . . . . . . . . . . . . . . . . . . . . . . .
4.6.2 Reversible expansion of an ideal gas . . . . . . . . . . . . . .
4.6.3 Spontaneous changes in an isolated system . . . . . . . . . .
4.6.4 Internal heat ow in an isolated system . . . . . . . . . . . .
4.6.5 Free expansion of a gas . . . . . . . . . . . . . . . . . . . . .
4.6.6 Adiabatic process with work . . . . . . . . . . . . . . . . . .
4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.8 The Statistical Interpretation of Entropy . . . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
102
102
103
106
106
109
111
114
116
116
120
123
124
124
125
126
127
127
128
128
129
129
130
130
133
5 Thermodynamic Potentials
5.1 Total Differential of a Dependent Variable . . .
5.2 Total Differential of the Internal Energy . . . .
5.3 Enthalpy, Helmholtz Energy, and Gibbs Energy
5.4 Closed Systems . . . . . . . . . . . . . . . . .
5.5 Open Systems . . . . . . . . . . . . . . . . .
5.6 Expressions for Heat Capacity . . . . . . . . .
5.7 Surface Work . . . . . . . . . . . . . . . . . .
5.8 Criteria for Spontaneity . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
135
135
136
138
140
141
143
144
145
148
.
.
.
.
.
.
.
.
150
150
152
152
155
156
157
157
159
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6 The Third Law and Cryogenics

6.1 The Zero of Entropy . . . . . . . . . . . . . . . . . . . .
6.2 Molar Entropies . . . . . . . . . . . . . . . . . . . . . .
6.2.1 Third-law molar entropies . . . . . . . . . . . . .
6.2.2 Molar entropies from spectroscopic measurements
6.2.3 Residual entropy . . . . . . . . . . . . . . . . . .
6.3 Cryogenics . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1 JouleThomson expansion . . . . . . . . . . . . .
6.3.2 Magnetization . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
C ONTENTS
Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7 Pure Substances in Single Phases
7.1 Volume Properties . . . . . . . . . . . . . . .
7.2 Internal Pressure . . . . . . . . . . . . . . . .
7.3 Thermal Properties . . . . . . . . . . . . . . .
7.3.1 The relation between CV;m and Cp;m . .
7.3.2 The measurement of heat capacities . .
7.3.3 Typical values . . . . . . . . . . . . .
7.4 Heating at Constant Volume or Pressure . . . .
7.5 Partial Derivatives with Respect to T, p, and V
7.5.1 Tables of partial derivatives . . . . . .
7.5.2 The JouleThomson coefcient . . . .
7.6 Isothermal Pressure Changes . . . . . . . . . .
7.6.1 Ideal gases . . . . . . . . . . . . . . .
7.6.2 Condensed phases . . . . . . . . . . .
7.7 Standard States of Pure Substances . . . . . .
7.8 Chemical Potential and Fugacity . . . . . . . .
7.8.1 Gases . . . . . . . . . . . . . . . . . .
7.8.2 Liquids and solids . . . . . . . . . . .
7.9 Standard Molar Quantities of a Gas . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
164
164
166
168
168
169
174
175
177
177
180
181
181
181
182
182
183
186
186
189
8 Phase Transitions and Equilibria of Pure Substances

8.1 Phase Equilibria . . . . . . . . . . . . . . . . . . . . . .
8.1.1 Equilibrium conditions . . . . . . . . . . . . . . .
8.1.2 Equilibrium in a multiphase system . . . . . . . .
8.1.3 Simple derivation of equilibrium conditions . . . .
8.1.4 Tall column of gas in a gravitational eld . . . . .
8.1.5 The pressure in a liquid droplet . . . . . . . . . .
8.1.6 The number of independent variables . . . . . . .
8.1.7 The Gibbs phase rule for a pure substance . . . . .
8.2 Phase Diagrams of Pure Substances . . . . . . . . . . . .
8.2.1 Features of phase diagrams . . . . . . . . . . . . .
8.2.2 Two-phase equilibrium . . . . . . . . . . . . . . .
8.2.3 The critical point . . . . . . . . . . . . . . . . . .
8.2.4 The lever rule . . . . . . . . . . . . . . . . . . . .
8.2.5 Volume properties . . . . . . . . . . . . . . . . .
8.3 Phase Transitions . . . . . . . . . . . . . . . . . . . . . .
8.3.1 Molar transition quantities . . . . . . . . . . . . .
8.3.2 Calorimetric measurement of transition enthalpies
8.3.3 Standard molar transition quantities . . . . . . . .
8.4 Coexistence Curves . . . . . . . . . . . . . . . . . . . . .
8.4.1 Chemical potential surfaces . . . . . . . . . . . .
8.4.2 The Clapeyron equation . . . . . . . . . . . . . .
8.4.3 The ClausiusClapeyron equation . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
193
193
193
194
195
196
198
199
200
200
201
204
206
207
210
212
212
214
214
214
215
216
219
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10
C ONTENTS
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
9 Mixtures
9.1 Composition Variables . . . . . . . . . . . . . . . . . . . . . . . . . .
9.1.1 Species and substances . . . . . . . . . . . . . . . . . . . . . .
9.1.2 Mixtures in general . . . . . . . . . . . . . . . . . . . . . . . .
9.1.3 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.1.4 Binary solutions . . . . . . . . . . . . . . . . . . . . . . . . .
9.1.5 The composition of a mixture . . . . . . . . . . . . . . . . . .
9.2 Partial Molar Quantities . . . . . . . . . . . . . . . . . . . . . . . . .
9.2.1 Partial molar volume . . . . . . . . . . . . . . . . . . . . . . .
9.2.2 The total differential of the volume in an open system . . . . . .
9.2.3 Evaluation of partial molar volumes in binary mixtures . . . . .
9.2.4 General relations . . . . . . . . . . . . . . . . . . . . . . . . .
9.2.5 Partial specic quantities . . . . . . . . . . . . . . . . . . . . .
9.2.6 The chemical potential of a species in a mixture . . . . . . . . .
9.2.7 Equilibrium conditions in a multiphase, multicomponent system
9.2.8 Relations involving partial molar quantities . . . . . . . . . . .
9.3 Gas Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.1 Partial pressure . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.2 The ideal gas mixture . . . . . . . . . . . . . . . . . . . . . . .
9.3.3 Partial molar quantities in an ideal gas mixture . . . . . . . . .
9.3.4 Real gas mixtures . . . . . . . . . . . . . . . . . . . . . . . . .
9.4 Liquid and Solid Mixtures of Nonelectrolytes . . . . . . . . . . . . . .
9.4.1 Raoults law . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.4.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.4.3 Partial molar quantities in ideal mixtures . . . . . . . . . . . .
9.4.4 Henrys law . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.4.5 The ideal-dilute solution . . . . . . . . . . . . . . . . . . . . .
9.4.6 Solvent behavior in the ideal-dilute solution . . . . . . . . . . .
9.4.7 Partial molar quantities in an ideal-dilute solution . . . . . . . .
9.5 Activity Coefcients in Mixtures of Nonelectrolytes . . . . . . . . . .
9.5.1 Reference states and standard states . . . . . . . . . . . . . . .
9.5.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.5.3 Real mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.5.4 Nonideal dilute solutions . . . . . . . . . . . . . . . . . . . . .
9.6 Evaluation of Activity Coefcients . . . . . . . . . . . . . . . . . . . .
9.6.1 Activity coefcients from gas fugacities . . . . . . . . . . . . .
9.6.2 Activity coefcients from the GibbsDuhem equation . . . . .
9.6.3 Activity coefcients from osmotic coefcients . . . . . . . . .
9.6.4 Fugacity measurements . . . . . . . . . . . . . . . . . . . . . .
9.7 Activity of an Uncharged Species . . . . . . . . . . . . . . . . . . . .
9.7.1 Standard states . . . . . . . . . . . . . . . . . . . . . . . . . .
9.7.2 Activities and composition . . . . . . . . . . . . . . . . . . . .
9.7.3 Pressure factors and pressure . . . . . . . . . . . . . . . . . . .
9.8 Mixtures in Gravitational and Centrifugal Fields . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
223
223
223
223
224
225
226
226
227
229
231
233
235
236
236
238
239
240
240
240
243
246
246
248
249
250
253
255
256
258
258
259
259
261
262
262
265
266
268
270
270
272
273
275
11
C ONTENTS
9.8.1 Gas mixture in a gravitational eld . . . . . . . . . . . . . . . . . . 275

9.8.2 Liquid solution in a centrifuge cell . . . . . . . . . . . . . . . . . . 277
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
10 Electrolyte Solutions
10.1 Single-ion Quantities . . . . . . . . . . . . . . . . . . . . .
10.2 Solution of a Symmetrical Electrolyte . . . . . . . . . . . .
10.3 Electrolytes in General . . . . . . . . . . . . . . . . . . . .
10.3.1 Solution of a single electrolyte . . . . . . . . . . . .
10.3.2 Multisolute solution . . . . . . . . . . . . . . . . .
10.3.3 Incomplete dissociation . . . . . . . . . . . . . . .
10.4 The DebyeH ckel Theory . . . . . . . . . . . . . . . . . .
u
10.5 Derivation of the DebyeH ckel Equation . . . . . . . . . .
u
10.6 Mean Ionic Activity Coefcients from Osmotic Coefcients
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11 Reactions and Other Chemical Processes
11.1 Mixing Processes . . . . . . . . . . . . . . . . . . . . . . .
11.1.1 Mixtures in general . . . . . . . . . . . . . . . . . .
11.1.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . .
11.1.3 Excess quantities . . . . . . . . . . . . . . . . . . .
11.1.4 The entropy change to form an ideal gas mixture . .
11.1.5 Molecular model of a liquid mixture . . . . . . . . .
11.1.6 Phase separation of a liquid mixture . . . . . . . . .
11.2 The Advancement and Molar Reaction Quantities . . . . . .
11.2.1 An example: ammonia synthesis . . . . . . . . . . .
11.2.2 Molar reaction quantities in general . . . . . . . . .
11.2.3 Standard molar reaction quantities . . . . . . . . . .
11.3 Molar Reaction Enthalpy . . . . . . . . . . . . . . . . . . .
11.3.1 Molar reaction enthalpy and heat . . . . . . . . . . .
11.3.2 Standard molar enthalpies of reaction and formation
11.3.3 Molar reaction heat capacity . . . . . . . . . . . . .
11.3.4 Effect of temperature on reaction enthalpy . . . . . .
11.4 Enthalpies of Solution and Dilution . . . . . . . . . . . . .
11.4.1 Molar enthalpy of solution . . . . . . . . . . . . . .
11.4.2 Enthalpy of dilution . . . . . . . . . . . . . . . . .
11.4.3 Molar enthalpies of solute formation . . . . . . . . .
11.4.4 Evaluation of relative partial molar enthalpies . . . .
11.5 Reaction Calorimetry . . . . . . . . . . . . . . . . . . . . .
11.5.1 The constant-pressure reaction calorimeter . . . . .
11.5.2 The bomb calorimeter . . . . . . . . . . . . . . . .
11.5.3 Other calorimeters . . . . . . . . . . . . . . . . . .
11.6 Adiabatic Flame Temperature . . . . . . . . . . . . . . . .
11.7 Gibbs Energy and Reaction Equilibrium . . . . . . . . . . .
11.7.1 The molar reaction Gibbs energy . . . . . . . . . . .
11.7.2 Spontaneity and reaction equilibrium . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
286
287
289
292
292
293
294
295
298
300
302
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
303
303
304
304
306
307
309
311
313
314
316
319
319
319
320
323
324
325
325
327
328
329
334
334
336
341
342
343
343
343
12
C ONTENTS
11.7.3 General derivation . . . . . . . . . . . . . . . . . .

11.7.4 Pure phases . . . . . . . . . . . . . . . . . . . . . .
11.7.5 Reactions involving mixtures . . . . . . . . . . . . .
11.7.6 Reaction in an ideal gas mixture . . . . . . . . . . .
11.8 The Thermodynamic Equilibrium Constant . . . . . . . . .
11.8.1 Activities and the denition of K . . . . . . . . . .
11.8.2 Reaction in a gas phase . . . . . . . . . . . . . . . .
11.8.3 Reaction in solution . . . . . . . . . . . . . . . . .
11.8.4 Evaluation of K . . . . . . . . . . . . . . . . . . .
11.9 Effects of Temperature and Pressure on Equilibrium Position
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
12 Equilibrium Conditions in Multicomponent Systems

12.1 Effects of Temperature . . . . . . . . . . . . . . . . . . . . . . . .
12.1.1 Variation of i =T with temperature . . . . . . . . . . . . .
12.1.2 Variation of =T with temperature . . . . . . . . . . . . .
i
12.1.3 Variation of ln K with temperature . . . . . . . . . . . . . .
12.2 Solvent Chemical Potentials from Phase Equilibria . . . . . . . . .
12.2.1 Freezing-point measurements . . . . . . . . . . . . . . . .
12.2.2 Osmotic-pressure measurements . . . . . . . . . . . . . . .
12.3 Binary Mixture in Equilibrium with a Pure Phase . . . . . . . . . .
12.4 Colligative Properties of a Dilute Solution . . . . . . . . . . . . . .
12.4.1 Freezing-point depression . . . . . . . . . . . . . . . . . .
12.4.2 Boiling-point elevation . . . . . . . . . . . . . . . . . . . .
12.4.3 Vapor-pressure lowering . . . . . . . . . . . . . . . . . . .
12.4.4 Osmotic pressure . . . . . . . . . . . . . . . . . . . . . . .
12.5 SolidLiquid Equilibria . . . . . . . . . . . . . . . . . . . . . . .
12.5.1 Freezing points of ideal binary liquid mixtures . . . . . . .
12.5.2 Solubility of a solid nonelectrolyte . . . . . . . . . . . . . .
12.5.3 Ideal solubility of a solid . . . . . . . . . . . . . . . . . . .
12.5.4 Solid compound of mixture components . . . . . . . . . . .
12.5.5 Solubility of a solid electrolyte . . . . . . . . . . . . . . . .
12.6 LiquidLiquid Equilibria . . . . . . . . . . . . . . . . . . . . . . .
12.6.1 Miscibility in binary liquid systems . . . . . . . . . . . . .
12.6.2 Solubility of one liquid in another . . . . . . . . . . . . . .
12.6.3 Solute distribution between two partially-miscible solvents .
12.7 Membrane Equilibria . . . . . . . . . . . . . . . . . . . . . . . . .
12.7.1 Osmotic membrane equilibrium . . . . . . . . . . . . . . .
12.7.2 Equilibrium dialysis . . . . . . . . . . . . . . . . . . . . .
12.7.3 Donnan membrane equilibrium . . . . . . . . . . . . . . .
12.8 LiquidGas Equilibria . . . . . . . . . . . . . . . . . . . . . . . .
12.8.1 Effect of liquid pressure on gas fugacity . . . . . . . . . . .
12.8.2 Effect of liquid composition on gas fugacities . . . . . . . .
12.8.3 The DuhemMargules equation . . . . . . . . . . . . . . .
12.8.4 Gas solubility . . . . . . . . . . . . . . . . . . . . . . . . .
12.8.5 Effect of temperature and pressure on Henrys law constants
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
344
345
345
347
350
350
353
354
355
356
360
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
367
367
367
368
369
370
371
373
375
376
378
381
381
382
384
384
386
387
387
390
392
392
392
395
395
396
396
397
400
400
401
405
406
408
13
C ONTENTS
12.9 Reaction Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409

12.10 Evaluation of Standard Molar Quantities . . . . . . . . . . . . . . . . . . 411
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
13 The Phase Rule and Phase Diagrams
13.1 The Gibbs Phase Rule for Multicomponent Systems . . .
13.1.1 Degrees of freedom . . . . . . . . . . . . . . . . .
13.1.2 Species approach to the phase rule . . . . . . . . .
13.1.3 Components approach to the phase rule . . . . . .
13.1.4 Examples . . . . . . . . . . . . . . . . . . . . . .
13.2 Phase Diagrams: Binary Systems . . . . . . . . . . . . .
13.2.1 Generalities . . . . . . . . . . . . . . . . . . . . .
13.2.2 Solidliquid systems . . . . . . . . . . . . . . . .
13.2.3 Partially-miscible liquids . . . . . . . . . . . . . .
13.2.4 Liquidgas systems with ideal liquid mixtures . . .
13.2.5 Liquidgas systems with nonideal liquid mixtures .
13.2.6 Solidgas systems . . . . . . . . . . . . . . . . .
13.2.7 Systems at high pressure . . . . . . . . . . . . . .
13.3 Phase Diagrams: Ternary Systems . . . . . . . . . . . . .
13.3.1 Three liquids . . . . . . . . . . . . . . . . . . . .
13.3.2 Two solids and a solvent . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
419
419
420
420
422
423
426
426
427
431
432
434
437
440
442
443
444
446
14 Galvanic Cells
14.1 Cell Diagrams and Cell Reactions . . . . . . .
14.1.1 Elements of a galvanic cell . . . . . . .
14.1.2 Cell diagrams . . . . . . . . . . . . . .
14.1.3 Electrode reactions and the cell reaction
14.1.4 Advancement and charge . . . . . . . .
14.2 Electric Potentials in the Cell . . . . . . . . . .
14.2.1 Cell potential . . . . . . . . . . . . . .
14.2.2 Measuring the equilibrium cell potential
14.2.3 Interfacial potential differences . . . .
14.3 Molar Reaction Quantities of the Cell Reaction
14.3.1 Relation between r Gcell and Ecell, eq .
14.3.2 Relation between r Gcell and r G . .
14.3.3 Standard molar reaction quantities . . .
14.4 The Nernst Equation . . . . . . . . . . . . . .
14.5 Evaluation of the Standard Cell Potential . . .
14.6 Standard Electrode Potentials . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
450
450
450
451
452
452
453
454
455
456
458
459
460
462
463
465
465
468
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Appendix A Denitions of the SI Base Units
471
Appendix B Physical Constants
472
14
C ONTENTS
Appendix C Symbols for Physical Quantities
473
Appendix D Miscellaneous Abbreviations and Symbols

477
D.1 Physical States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
D.2 Subscripts for Chemical Processes . . . . . . . . . . . . . . . . . . . . . . 478
D.3 Superscripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Appendix E Calculus Review
E.1 Derivatives . . . . . .
E.2 Partial Derivatives . .
E.3 Integrals . . . . . . . .
E.4 Line Integrals . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
480
480
480
481
481
Appendix F Mathematical Properties of State Functions

F.1 Differentials . . . . . . . . . . . . . . . . . . . .
F.2 Total Differential . . . . . . . . . . . . . . . . . .
F.3 Integration of a Total Differential . . . . . . . . .
F.4 Legendre Transforms . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
482
482
482
484
485
Appendix G Forces, Energy, and Work

G.1 Forces between Particles . . . . . . . . . . . . .
G.2 The System and Surroundings . . . . . . . . . .
G.3 System Energy Change . . . . . . . . . . . . . .
G.4 Macroscopic Work . . . . . . . . . . . . . . . .
G.5 The Work Done on the System and Surroundings
G.6 The Local Frame and Internal Energy . . . . . .
G.7 Nonrotating Local Frame . . . . . . . . . . . . .
G.8 Center-of-mass Local Frame . . . . . . . . . . .
G.9 Rotating Local Frame . . . . . . . . . . . . . .
G.10 Earth-Fixed Reference Frame . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
487
488
491
493
494
496
496
500
500
503
504
.
.
.
.
.
.
.
.
.
.
Appendix H Standard Molar Thermodynamic Properties
505
Appendix I
508
Answers to Selected Problems
Bibliography
512
Index
521
B IOGRAPHICAL S KETCHES
Benjamin Thompson, Count of Rumford

James Prescott Joule . . . . . . . . . .
Sadi Carnot . . . . . . . . . . . . . . .
Rudolf Julius Emmanuel Clausius . . .
William Thomson, Lord Kelvin . . . . .
Max Karl Ernst Ludwig Planck . . . . .
Josiah Willard Gibbs . . . . . . . . . .
Walther Hermann Nernst . . . . . . . .
William Francis Giauque . . . . . . . .
Benoit Paul Emile Clapeyron . . . . . .

William Henry . . . . . . . . . . . . . .
Gilbert Newton Lewis . . . . . . . . . .
Peter Josephus Wilhelmus Debye . . . .
Germain Henri Hess . . . . . . . . . . .
Francois-Marie Raoult . . . . . . . . .
Jacobus Henricus vant Hoff . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
15
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
63
85
107
110
115
117
139
151
160
218
251
271
296
322
380
383
P REFACE TO THE S ECOND E DITION
This second edition of Thermodynamics and Chemistry is a revised and enlarged version
of the rst edition published by Prentice Hall in 2001. The book is designed primarily as a
textbook for a one-semester course for graduate or undergraduate students who have already
been introduced to thermodynamics in a physical chemistry course.
The PDF le of this book contains hyperlinks to pages, sections, equations, tables,
gures, bibliography items, and problems. If you are viewing the PDF on a computer
screen, tablet, or color e-reader, the links are colored in blue.
Scattered through the text are sixteen one-page biographical sketches of some of the
historical giants of thermodynamics. A list is given on the preceding page. The sketches
are not intended to be comprehensive biographies, but rather to illustrate the human side of
thermodynamicsthe struggles and controversies by which the concepts and experimental
methodology of the subject were developed.
The epigraphs on page 18 are intended to suggest the nature and importance of classical thermodynamics. You may wonder about the conversation between Alice and Humpty
Dumpty. Its point, particularly important in the study of thermodynamics, is the need to pay
attention to denitionsthe intended meanings of words.
I welcome comments and suggestions for improving this book. My e-mail address appears below.
Howard DeVoe
[email protected]
16
F ROM THE P REFACE TO THE F IRST

E DITION
Classical thermodynamics, the subject of this book, is concerned with macroscopic aspects of the
interaction of matter with energy in its various forms. This book is designed as a text for a onesemester course for senior undergraduate or graduate students who have already been introduced to
thermodynamics in an undergraduate physical chemistry course.
Anyone who studies and uses thermodynamics knows that a deep understanding of this subject
does not come easily. There are subtleties and interconnections that are difcult to grasp at rst. The
more times one goes through a thermodynamics course (as a student or a teacher), the more insight
one gains. Thus, this text will reinforce and extend the knowledge gained from an earlier exposure
to thermodynamics. To this end, there is fairly intense discussion of some basic topics, such as the
nature of spontaneous and reversible processes, and inclusion of a number of advanced topics, such
as the reduction of bomb calorimetry measurements to standard-state conditions.
This book makes no claim to be an exhaustive treatment of thermodynamics. It concentrates
on derivations of fundamental relations starting with the thermodynamic laws and on applications
of these relations in various areas of interest to chemists. Although classical thermodynamics treats
matter from a purely macroscopic viewpoint, the book discusses connections with molecular properties when appropriate.
In deriving equations, I have strived for rigor, clarity, and a minimum of mathematical complexity. I have attempted to clearly state the conditions under which each theoretical relation is valid
because only by understanding the assumptions and limitations of a derivation can one know when
to use the relation and how to adapt it for special purposes. I have taken care to be consistent in the
use of symbols for physical properties. The choice of symbols follows the current recommendations
of the International Union of Pure and Applied Chemistry (IUPAC) with a few exceptions made to
avoid ambiguity.
I owe much to J. Arthur Campbell, Luke E. Steiner, and William Moftt, gifted teachers who
introduced me to the elegant logic and practical utility of thermodynamics. I am immensely grateful
to my wife Stephanie for her continued encouragement and patience during the period this book
went from concept to reality.
I would also like to acknowledge the help of the following reviewers: James L. Copeland,
Kansas State University; Lee Hansen, Brigham Young University; Reed Howald, Montana State
UniversityBozeman; David W. Larsen, University of MissouriSt. Louis; Mark Ondrias, University
of New Mexico; Philip H. Rieger, Brown University; Leslie Schwartz, St. John Fisher College; Allan
L. Smith, Drexel University; and Paul E. Smith, Kansas State University.
17
A theory is the more impressive the greater the simplicity of its

premises is, the more different kinds of things it relates, and the more
extended is its area of applicability. Therefore the deep impression
which classical thermodynamics made upon me. It is the only physical
theory of universal content concerning which I am convinced that,
within the framework of the applicability of its basic concepts, it will
never be overthrown.
Albert Einstein
Thermodynamics is a discipline that involves a formalization of a large
number of intuitive concepts derived from common experience.
J. G. Kirkwood and I. Oppenheim, Chemical Thermodynamics, 1961
The rst law of thermodynamics is nothing more than the principle of
the conservation of energy applied to phenomena involving the
production or absorption of heat.
Max Planck, Treatise on Thermodynamics, 1922
The law that entropy always increasesthe second law of
thermodynamicsholds, I think, the supreme position among the laws
of Nature. If someone points out to you that your pet theory of the
universe is in disagreement with Maxwells equationsthen so much
the worse for Maxwells equations. If it is found to be contradicted by
observationwell, these experimentalists do bungle things sometimes.
But if your theory is found to be against the second law of
thermodynamics I can give you no hope; there is nothing for it but to
collapse in deepest humiliation.
Sir Arthur Eddington, The Nature of the Physical World, 1928
Thermodynamics is a collection of useful relations between quantities,
every one of which is independently measurable. What do such
relations tell one about ones system, or in other words what do we
learn from thermodynamics about the microscopic explanations of
macroscopic changes? Nothing whatever. What then is the use of
thermodynamics? Thermodynamics is useful precisely because some
quantities are easier to measure than others, and that is all.
M. L. McGlashan, J. Chem. Educ., 43, 226232 (1966)
When I use a word, Humpty Dumpty said, in rather a scornful tone,
it means just what I choose it to meanneither more nor less.
The question is, said Alice,whether you can make words mean
so many different things.
The question is, said Humpty Dumpty, which is to be master
thats all.
Lewis Carroll, Through the Looking-Glass
C HAPTER 1
I NTRODUCTION
Thermodynamics is a quantitative subject. It allows us to derive relations between the

values of numerous physical quantities. Some physical quantities, such as a mole fraction,
are dimensionless; the value of one of these quantities is a pure number. Most quantities,
however, are not dimensionless and their values must include one or more units. This
chapter reviews the SI system of units, which are the preferred units in science applications.
The chapter then discusses some useful mathematical manipulations of physical quantities
using quantity calculus, and certain general aspects of dimensional analysis.
1.1 UNITS
There is international agreement that the units used for physical quantities in science and
technology should be those of the International System of Units, or SI (standing for the
French Syst` me International dUnit s). The Physical Chemistry Division of the Intere
e
national Union of Pure and Applied Chemistry, or IUPAC, produces a manual of recommended symbols and terminology for physical quantities and units based on the SI. The
manual has become known as the Green Book (from the color of its cover) and is referred
to here as the IUPAC Green Book. This book will, with a few exceptions, use symbols recommended in the third edition (2007) of the IUPAC Green Book;1 these symbols are listed
for convenient reference in Appendices C and D.
The SI is built on the seven base units listed in Table 1.1 on the next page. These base
units are independent physical quantities that are sufcient to describe all other physical
quantities. One of the seven quantities, luminous intensity, is not used in this book and is
usually not needed in thermodynamics. The ofcial denitions of the base units are given
in Appendix A.
Table 1.2 lists derived units for some additional physical quantities used in thermodynamics. The derived units have exact denitions in terms of SI base units, as given in the
last column of the table.
The units listed in Table 1.3 are sometimes used in thermodynamics but are not part
of the SI. They do, however, have exact denitions in terms of SI units and so offer no
problems of numerical conversion to or from SI units.
1 Ref.
[36]. The references are listed in the Bibliography at the back of the book.
19
CHAPTER 1 INTRODUCTION
1.1 U NITS
20
Table 1.1 SI base units
Physical quantity
SI unit
Symbol
length
mass
time
thermodynamic temperature
amount of substance
electric current
luminous intensity
meter a
kilogram
second
kelvin
mole
ampere
candela
m
kg
s
K
mol
A
cd
a or
metre
Table 1.2 SI derived units

Physical quantity
Unit
Symbol
Denition of unit
force
pressure
Celsius temperature
energy
power
frequency
electric charge
electric potential
electric resistance
newton
pascal
degree Celsius
joule
watt
hertz
coulomb
volt
ohm
N
Pa
C
J
W
Hz
C
V
1 N D 1 m kg s 2
1 Pa D 1 N m 2 D 1 kg m 1 s 2
t = C D T =K 273:15
1 J D 1 N m D 1 m2 kg s 2
1 W D 1 J s 1 D 1 m2 kg s 3
1 Hz D 1 s 1
1C D 1As
1 V D 1 J C 1 D 1 m2 kg s 3 A 1
1 D 1 V A 1 D 1 m2 kg s 3 A 2
Table 1.3 Non-SI derived units

Physical quantity
volume
pressure
pressure
pressure
energy
a or
litre
Unit
liter
bar
atmosphere
torr
calorie c
b or
c or
Symbol
b
L
bar
atm
Torr
cal d
thermochemical calorie
Denition of unit
1 L D 1 dm3 D 10 3 m3
1 bar D 105 Pa
1 atm D 101,325 Pa D 1:01325 bar
1 Torr D .1=760/ atm D .101,325/760/ Pa
1 cal D 4:184 J
d or
calth
1.1 U NITS
21
Table 1.4 SI prexes
Fraction
Prex
Symbol
Multiple
Prex
Symbol
deci
centi
milli
micro
nano
pico
femto
atto
zepto
yocto
d
c
m
n
p
f
a
z
y
10
102
103
106
109
1012
1015
1018
1021
1024
deka
hecto
kilo
mega
giga
tera
peta
exa
zetta
yotta
da
h
k
M
G
T
P
E
Z
Y
10
10
10
10
10
10
10
10
10
10
2
3
6
9
12
15
18
21
24
Any of the symbols for units listed in Tables 1.11.3, except kg and C, may be preceded
by one of the prex symbols of Table 1.4 to construct a decimal fraction or multiple of the
unit. (The symbol g may be preceded by a prex symbol to construct a fraction or multiple
of the gram.) The combination of prex symbol and unit symbol is taken as a new symbol
that can be raised to a power without using parentheses, as in the following examples:
1 mg D 1 10 3 g
1 cm D 1
1 cm3 D .1
10
2m
10
2 m/3
D1
10
6 m3
1.1.1 Amount of substance and amount

The physical quantity formally called amount of substance is a counting quantity for particles, such as atoms or molecules, or for other chemical entities. The counting unit is
invariably the mole, dened as the amount of substance containing as many particles as the
number of atoms in exactly 12 grams of pure carbon-12 nuclide, 12 C. See Appendix A for
the wording of the ofcial IUPAC denition. This denition is such that one mole of H2 O
molecules, for example, has a mass of 18:0153 grams (where 18:0153 is the relative molecular mass of H2 O) and contains 6:02214 1023 molecules (where 6:02214 1023 mol 1 is
the Avogadro constant to six signicant digits). The same statement can be made for any
other substance if 18:0153 is replaced by the appropriate atomic mass or molecular mass
value.
The symbol for amount of substance is n. It is admittedly awkward to refer to n(H2 O)
as the amount of substance of water. This book simply shortens amount of substance to
amount, a common usage that is condoned by the IUPAC.2 Thus, the amount of water in
the system refers not to the mass or volume of water, but to the number of H2 O molecules
in the system expressed in a counting unit such as the mole.
2 Ref.
[117]. An alternative name suggested for n is chemical amount.
1.2 Q UANTITY C ALCULUS
22
1.2 QUANTITY CALCULUS

This section gives examples of how we may manipulate physical quantities by the rules of
algebra. The method is called quantity calculus, although a better term might be quantity
algebra.
Quantity calculus is based on the concept that a physical quantity, unless it is dimensionless, has a value equal to the product of a numerical value (a pure number) and one or
more units:
physical quantity = numerical value units
(1.2.1)
(If the quantity is dimensionless, it is equal to a pure number without units.) The physical
property may be denoted by a symbol, but the symbol does not imply a particular choice of
units. For instance, this book uses the symbol for density, but can be expressed in any
units having the dimensions of mass divided by volume.
A simple example illustrates the use of quantity calculus. We may express the density
of water at 25 C to four signicant digits in SI base units by the equation
102 kg m
D 9:970
(1.2.2)
and in different density units by the equation

D 0:9970 g cm
We may divide both sides of the last equation by 1 g cm

=g cm
D 0:9970
(1.2.3)
3
to obtain a new equation

(1.2.4)
Now the pure number 0:9970 appearing in this equation is the number of grams in one
cubic centimeter of water, so we may call the ratio =g cm 3 the number of grams per
cubic centimeter. By the same reasoning, =kg m 3 is the number of kilograms per cubic
meter. In general, a physical quantity divided by particular units for the physical quantity is
a pure number representing the number of those units.
Just as it would be incorrect to call the number of grams per cubic centimeter,
because that would refer to a particular choice of units for , the common practice of
calling n the number of moles is also strictly speaking not correct. It is actually the
ratio n=mol that is the number of moles.
In a table, the ratio =g cm 3 makes a convenient heading for a column of density

values because the column can then show pure numbers. Likewise, it is convenient to use
=g cm 3 as the label of a graph axis and to show pure numbers at the grid marks of the
axis. You will see many examples of this usage in the tables and gures of this book.
A major advantage of using SI base units and SI derived units is that they are coherent.
That is, values of a physical quantity expressed in different combinations of these units have
the same numerical value.
1.2 Q UANTITY C ALCULUS
23
For example, suppose we wish to evaluate the pressure of a gas according to the ideal
gas equation3
pD
nRT
V
(1.2.5)
(ideal gas)
In this equation, p, n, T , and V are the symbols for the physical quantities pressure, amount
(amount of substance), thermodynamic temperature, and volume, respectively, and R is the
gas constant.
The calculation of p for 5:000 moles of an ideal gas at a temperature of 298:15 kelvins,
in a volume of 4:000 cubic meters, is
pD
.5:000 mol/.8:3145 J K 1 mol

4:000 m3
1 /.298:15 K/
D 3:099
103 J m
(1.2.6)
The mole and kelvin units cancel, and we are left with units of J m 3 , a combination of
an SI derived unit (the joule) and an SI base unit (the meter). The units J m 3 must have
dimensions of pressure, but are not commonly used to express pressure.
To convert J m 3 to the SI derived unit of pressure, the pascal (Pa), we can use the
following relations from Table 1.2:
1J D 1Nm
1 Pa D 1 N m
(1.2.7)
When we divide both sides of the rst relation by 1 J and divide both sides of the second
relation by 1 N m 2 , we obtain the two new relations
1 D .1 N m=J/
.1 Pa=N m
/D1
(1.2.8)
The ratios in parentheses are conversion factors. When a physical quantity is multiplied
by a conversion factor that, like these, is equal to the pure number 1, the physical quantity
changes its units but not its value. When we multiply Eq. 1.2.6 by both of these conversion
factors, all units cancel except Pa:
p D .3:099
D 3:099
103 J m
3/
.1 N m=J/
.1 Pa=N m
103 Pa
/
(1.2.9)
This example illustrates the fact that to calculate a physical quantity, we can simply
enter into a calculator numerical values expressed in SI units, and the result is the numerical
value of the calculated quantity expressed in SI units. In other words, as long as we use
only SI base units and SI derived units (without prexes), all conversion factors are unity.
Of course we do not have to limit the calculation to SI units. Suppose we wish to
express the calculated pressure in torrs, a non-SI unit. In this case, using a conversion factor
obtained from the denition of the torr in Table 1.3, the calculation becomes
p D .3:099
103 Pa/
D 23:24 Torr
.760 Torr=101; 325 Pa/

(1.2.10)
3 This
is the rst equation in this book that, like many others to follow, shows conditions of validity in parentheses immediately below the equation number at the right. Thus, Eq. 1.2.5 is valid for an ideal gas.
1.3 D IMENSIONAL A NALYSIS
24
1.3 DIMENSIONAL ANALYSIS

Sometimes you can catch an error in the form of an equation or expression, or in the dimensions of a quantity used for a calculation, by checking for dimensional consistency. Here
are some rules that must be satised:
both sides of an equation have the same dimensions
all terms of a sum or difference have the same dimensions
logarithms and exponentials, and arguments of logarithms and exponentials, are dimensionless
a quantity used as a power is dimensionless
In this book the differential of a function, such as df , refers to an innitesimal quantity.
If one side of an equation is an innitesimal quantity, the other side must also be. Thus,
the equation df D a dx C b dy (where ax and by have the same dimensions as f ) makes
mathematical sense, but df D ax C b dy does not.
Derivatives, partial derivatives, and integrals have dimensions that we must take into
account when determining the overall dimensions of an expression that includes them. For
instance:
the derivative dp= dT and the partial derivative .@p=@T /V have the same dimensions
as p=T
the partial second derivative .@2 p=@T 2 /V has the same dimensions as p=T 2
R
the integral T dT has the same dimensions as T 2
Some examples of applying these principles are given here using symbols described in
Sec. 1.2.
Example 1. Since the gas constant R may be expressed in units of J K 1 mol 1 , it has
dimensions of energy divided by thermodynamic temperature and amount. Thus, RT has
dimensions of energy divided by amount, and nRT has dimensions of energy. The products
RT and nRT appear frequently in thermodynamic expressions.
Example 2. What are the dimensions of the quantity nRT ln.p=p / and of p in
this expression? The quantity has the same dimensions as nRT (or energy) because the
logarithm is dimensionless. Furthermore, p in this expression has dimensions of pressure
in order to make the argument of the logarithm, p=p , dimensionless.
Example 3. Find the dimensions of the constants a and b in the van der Waals equation
pD
nRT
V nb
n2 a
V2
Dimensional analysis tells us that, because nb is subtracted from V , nb has dimensions

of volume and therefore b has dimensions of volume/amount. Furthermore, since the right
side of the equation is a difference of two terms, these terms have the same dimensions
as the left side, which is pressure. Therefore, the second term n2 a=V 2 has dimensions of
pressure, and a has dimensions of pressure volume2 amount 2 .
Example 4. Consider an equation of the form
@ ln x
y
D
@T p
R
1.3 D IMENSIONAL A NALYSIS
25
What are the SI units of y? ln x is dimensionless, so the left side of the equation has the
dimensions of 1=T , and its SI units are K 1 . The SI units of the right side are therefore
also K 1 . Since R has the units J K 1 mol 1 , the SI units of y are J K 2 mol 1 .
P ROBLEM
26
PROBLEM
1.1 Consider the following equations for the pressure of a real gas. For each equation, nd the
dimensions of the constants a and b and express these dimensions in SI units.
(a) The Dieterici equation:
pD
RT e .an=VRT /
.V =n/ b
(b) The RedlichKwong equation:

pD
RT
.V =n/ b
an2
T 1=2 V .V C nb/
C HAPTER 2
S YSTEMS AND T HEIR P ROPERTIES
This chapter begins by explaining some basic terminology of thermodynamics. It discusses

macroscopic properties of matter in general and properties distinguishing different physical
states of matter in particular. Virial equations of state of a pure gas are introduced. The
chapter goes on to discuss some basic macroscopic properties and their measurement. Finally, several important concepts needed in later chapters are described: thermodynamic
states and state functions, independent and dependent variables, processes, and internal energy.
2.1 THE SYSTEM, SURROUNDINGS, AND BOUNDARY

Chemists are interested in systems containing matterthat which has mass and occupies
physical space. Classical thermodynamics looks at macroscopic aspects of matter. It deals
with the properties of aggregates of vast numbers of microscopic particles (molecules,
atoms, and ions). The macroscopic viewpoint, in fact, treats matter as a continuous material medium rather than as the collection of discrete microscopic particles we know are
actually present. Although this book is an exposition of classical thermodynamics, at times
it will point out connections between macroscopic properties and molecular structure and
behavior.
A thermodynamic system is any three-dimensional region of physical space on which
we wish to focus our attention. Usually we consider only one system at a time and call it
simply the system. The rest of the physical universe constitutes the surroundings of the
system.
The boundary is the closed three-dimensional surface that encloses the system and
separates it from the surroundings. The boundary may (and usually does) coincide with
real physical surfaces: the interface between two phases, the inner or outer surface of the
wall of a ask or other vessel, and so on. Alternatively, part or all of the boundary may be
an imagined intangible surface in space, unrelated to any physical structure. The size and
shape of the system, as dened by its boundary, may change in time. In short, our choice of
the three-dimensional region that constitutes the system is arbitrarybut it is essential that
we know exactly what this choice is.
We usually think of the system as a part of the physical universe that we are able to
inuence only indirectly through its interaction with the surroundings, and the surroundings
27
CHAPTER 2 SYSTEMS AND THEIR PROPERTIES

2.1 T HE S YSTEM , S URROUNDINGS , AND B OUNDARY
28
as the part of the universe that we are able to directly manipulate with various physical
devices under our control. That is, we (the experimenters) are part of the surroundings, not
the system.
For some purposes we may wish to treat the system as being divided into subsystems,
or to treat the combination of two or more systems as a supersystem.
If over the course of time matter is transferred in either direction across the boundary,
the system is open; otherwise it is closed. If the system is open, matter may pass through a
stationary boundary, or the boundary may move through matter that is xed in space.
If the boundary allows heat transfer between the system and surroundings, the boundary
is diathermal. An adiabatic1 boundary, on the other hand, is a boundary that does not allow
heat transfer. We can, in principle, ensure that the boundary is adiabatic by surrounding the
system with an adiabatic wallone with perfect thermal insulation and a perfect radiation
shield.
An isolated system is one that exchanges no matter, heat, or work with the surroundings,
so that the mass and total energy of the system remain constant over time.2 A closed system
with an adiabatic boundary, constrained to do no work and to have no work done on it, is
an isolated system.
The constraints required to prevent work usually involve forces between the system
and surroundings. In that sense a system may interact with the surroundings even
though it is isolated. For instance, a gas contained within rigid, thermally-insulated
walls is an isolated system; the gas exerts a force on each wall, and the wall exerts an
equal and opposite force on the gas. An isolated system may also experience a constant
external eld, such as a gravitational eld.
The term body usually implies a system, or part of a system, whose mass and chemical
composition are constant over time.
2.1.1 Extensive and intensive properties

A quantitative property of a system describes some macroscopic feature that, although it
may vary with time, has a particular value at any given instant of time.
Table 2.1 on the next page lists the symbols of some of the properties discussed in this
chapter and the SI units in which they may be expressed. A much more complete table is
found in Appendix C.
Most of the properties studied by thermodynamics may be classied as either extensive
or intensive. We can distinguish these two types of properties by the following considerations.
If we imagine the system to be divided by an imaginary surface into two parts, any
property of the system that is the sum of the property for the two parts is an extensive
property. That is, an additive property is extensive. Examples are mass, volume, amount,
energy, and the surface area of a solid.
1 Greek:
impassable.
energy in this denition of an isolated system is measured in a local reference frame, as will be explained
in Sec. 2.6.2.
2 The

2.1 T HE S YSTEM , S URROUNDINGS , AND B OUNDARY
29
Table 2.1 Symbols and SI units for some common properties

Symbol
Physical quantity
SI unit
E
m
n
p
T
V
U
energy
mass
amount of substance
pressure
volume
internal energy
density
J
kg
mol
Pa
K
m3
J
kg m
Sometimes a more restricted denition of an extensive property is used: The property

must be not only additive, but also proportional to the mass or the amount when intensive properties remain constant. According to this denition, mass, volume, amount,
and energy are extensive, but surface area is not.
If we imagine a homogeneous region of space to be divided into two or more parts of

arbitrary size, any property that has the same value in each part and the whole is an intensive
property; for example density, concentration, pressure (in a uid), and temperature. The
value of an intensive property is the same everywhere in a homogeneous region, but may
vary from point to point in a heterogeneous regionit is a local property.
Since classical thermodynamics treats matter as a continuous medium, whereas matter
actually contains discrete microscopic particles, the value of an intensive property at a point
is a statistical average of the behavior of many particles. For instance, the density of a gas at
one point in space is the average mass of a small volume element at that point, large enough
to contain many molecules, divided by the volume of that element.
Some properties are dened as the ratio of two extensive quantities. If both extensive
quantities refer to a homogeneous region of the system or to a small volume element, the ratio is an intensive property. For example concentration, dened as the ratio amount=volume,
is intensive. A mathematical derivative of one such extensive quantity with respect to another is also intensive.
A special case is an extensive quantity divided by the mass, giving an intensive specic
quantity; for example
V
1
D
(2.1.1)
Specic volume D
m
If the symbol for the extensive quantity is a capital letter, it is customary to use the corresponding lower-case letter as the symbol for the specic quantity. Thus the symbol for
specic volume is v.
Another special case encountered frequently in this book is an extensive property for a
pure, homogeneous substance divided by the amount n. The resulting intensive property is
called, in general, a molar quantity or molar property. To symbolize a molar quantity, this
book follows the recommendation of the IUPAC: The symbol of the extensive quantity is
followed by subscript m, and optionally the identity of the substance is indicated either by

2.2 P HASES AND P HYSICAL S TATES OF M ATTER
30
a subscript or a formula in parentheses. Examples are

Molar volume D
V
D Vm
n
Molar volume of substance i D
V
D Vm;i
ni
Molar volume of H2 O D Vm (H2 O)
(2.1.2)
(2.1.3)
(2.1.4)
In the past, especially in the United States, molar quantities were commonly denoted
with an overbar (e.g., V i ).
2.2 PHASES AND PHYSICAL STATES OF MATTER

A phase is a region of the system in which each intensive property (such as temperature and
pressure) has at each instant either the same value throughout (a uniform or homogeneous
phase), or else a value that varies continuously from one point to another. Whenever this
book mentions a phase, it is a uniform phase unless otherwise stated. Two different phases
meet at an interface surface, where intensive properties have a discontinuity or change
over a small distance.
Some intensive properties (e.g., refractive index and polarizability) can have directional
characteristics. A uniform phase may be either isotropic, exhibiting the same values of these
properties in all directions, or anisotropic, as in the case of some solids and liquid crystals.
A vacuum is a uniform phase of zero density.
Suppose we have to deal with a nonuniform region in which intensive properties vary
continuously in space along one or more directionsfor example, a tall column of gas in
a gravitational eld whose density decreases with increasing altitude. There are two ways
we may treat such a nonuniform, continuous region: either as a single nonuniform phase,
or else as an innite number of uniform phases, each of innitesimal size in one or more
dimensions.
2.2.1 Physical states of matter

We are used to labeling phases by physical state, or state of aggregation. It is common
to say that a phase is a solid if it is relatively rigid, a liquid if it is easily deformed and
relatively incompressible, and a gas if it is easily deformed and easily compressed. Since
these descriptions of responses to external forces differ only in degree, they are inadequate
to classify intermediate cases.
A more rigorous approach is to make a primary distinction between a solid and a uid,
based on the phases response to an applied shear stress, and then use additional criteria
to classify a uid as a liquid, gas, or supercritical uid. Shear stress is a tangential force
per unit area that is exerted on matter on one side of an interior plane by the matter on the
other side. We can produce shear stress in a phase by applying tangential forces to parallel
surfaces of the phase as shown in Fig. 2.1 on the next page.
A solid responds to shear stress by undergoing momentary relative motion of its parts,
resulting in deformationa change of shape. If the applied shear stress is constant and
small (not large enough to cause creep or fracture), the solid quickly reaches a certain

31
Figure 2.1 Experimental procedure for producing shear stress in a phase (shaded).
Blocks at the upper and lower surfaces of the phase are pushed in opposite directions,
dragging the adjacent portions of the phase with them.
degree of deformation that depends on the magnitude of the stress and maintains this
deformation without further change as long as the shear stress continues to be applied.
On the microscopic level, deformation requires relative movement of adjacent layers of
particles (atoms, molecules, or ions). The shape of an unstressed solid is determined by
the attractive and repulsive forces between the particles; these forces make it difcult
for adjacent layers to slide past one another, so that the solid resists deformation.
A uid responds to shear stress differently, by undergoing continuous relative motion (ow)
of its parts. The ow continues as long as there is any shear stress, no matter how small,
and stops only when the shear stress is removed.
Thus, a constant applied shear stress causes a xed deformation in a solid and continuous ow in a uid. We say that a phase under constant shear stress is a solid if, after the
initial deformation, we are unable to detect a further change in shape during the period we
observe the phase.
Usually this criterion allows us to unambiguously classify a phase as either a solid or
a uid. Over a sufciently long time period, however, detectable ow is likely to occur
in any material under shear stress of any magnitude. Thus, the distinction between solid
and uid actually depends on the time scale of observation. This fact is obvious when
we observe the behavior of certain materials (such as Silly Putty, or a paste of water and
cornstarch) that exhibit solid-like behavior over a short time period and uid-like behavior
over a longer period. Such materials, that resist deformation by a suddenly-applied shear
stress but undergo ow over a longer time period, are called viscoelastic solids.
2.2.2 Phase coexistence and phase transitions

This section considers some general characteristics of systems containing more than one
phase.
Suppose we bring two uniform phases containing the same constituents into physical
contact at an interface surface. If we nd that the phases have no tendency to change
over time while both have the same temperature and the same pressure, but differ in other
intensive properties such as density and composition, we say that they coexist in equilibrium
with one another. The conditions for such phase coexistence are the subject of later sections
in this book, but they tend to be quite restricted. For instance, the liquid and gas phases of
pure H2 O at a pressure of 1 bar can coexist at only one temperature, 99:61 C.
A phase transition of a pure substance is a change over time in which there is a continuous transfer of the substance from one phase to another. Eventually one phase can

32
supercritical
uid
cp
liquid
solid
gas
tp
Figure 2.2 Pressuretemperature phase diagram of a pure substance (schematic).

Point cp is the critical point, and point tp is the triple point. Each area is labeled
with the physical state that is stable under the pressure-temperature conditions that
fall within the area. A solid curve (coexistence curve) separating two areas is the locus of pressure-temperature conditions that allow the phases of these areas to coexist
at equilibrium. Path ABCD illustrates continuity of states.
completely disappear, and the substance has been completely transferred to the other phase.
If both phases coexist in equilibrium with one another, and the temperature and pressure of
both phases remain equal and constant during the phase transition, the change is an equilibrium phase transition. For example, H2 O at 99:61 C and 1 bar can undergo an equilibrium
phase transition from liquid to gas (vaporization) or from gas to liquid (condensation). During an equilibrium phase transition, there is a transfer of energy between the system and its
surroundings by means of heat or work.
2.2.3 Fluids
It is usual to classify a uid as either a liquid or a gas. The distinction is important for a
pure substance because the choice determines the treatment of the phases standard state
(see Sec. 7.7). To complicate matters, a uid at high pressure may be a supercritical uid.
Sometimes a plasma (a highly ionized, electrically conducting medium) is considered a
separate kind of uid state; it is the state found in the earths ionosphere and in stars.
In general, and provided the pressure is not high enough for supercritical phenomena
to existusually true of pressures below 25 bar except in the case of He or H2 we can
make the distinction between liquid and gas simply on the basis of density. A liquid has a
relatively high density that is insensitive to changes in temperature and pressure. A gas, on
the other hand, has a relatively low density that is sensitive to temperature and pressure and
that approaches zero as pressure is reduced at constant temperature.
This simple distinction between liquids and gases fails at high pressures, where liquid
and gas phases may have similar densities at the same temperature. Figure 2.2 shows how
we can classify stable uid states of a pure substance in relation to a liquidgas coexistence
curve and a critical point. If raising the temperature of a uid at constant pressure causes
a phase transition to a second uid phase, the original uid was a liquid and the transition

33
occurs at the liquidgas coexistence curve. This curve ends at a critical point, at which
all intensive properties of the coexisting liquid and gas phases become identical. The uid
state of a pure substance at a temperature greater than the critical temperature and a pressure
greater than the critical pressure is called a supercritical uid.
The term vapor is sometimes used for a gas that can be condensed to a liquid by increasing the pressure at constant temperature. By this denition, the vapor state of a substance
exists only at temperatures below the critical temperature.
The designation of a supercritical uid state of a substance is used more for convenience
than because of any unique properties compared to a liquid or gas. If we vary the temperature or pressure in such a way that the substance changes from what we call a liquid to what
we call a supercritical uid, we observe only a continuous density change of a single phase,
and no phase transition with two coexisting phases. The same is true for a change from
a supercritical uid to a gas. Thus, by making the changes described by the path ABCD
shown in Fig. 2.2, we can transform a pure substance from a liquid at a certain pressure
to a gas at the same pressure without ever observing an interface between two coexisting
phases! This curious phenomenon is called continuity of states.
Chapter 6 will take up the discussion of further aspects of the physical states of pure
substances.
If we are dealing with a uid mixture (instead of a pure substance) at a high pressure, it
may be difcult to classify the phase as either liquid or gas. The complexity of classication
at high pressure is illustrated by the barotropic effect, observed in some mixtures, in which
a small change of temperature or pressure causes what was initially the more dense of two
coexisting uid phases to become the less dense phase. In a gravitational eld, the two
phases switch positions.
2.2.4 The equation of state of a uid

Suppose we prepare a uniform uid phase containing a known amount ni of each constituent
substance i, and adjust the temperature T and pressure p to denite known values. We
expect this phase to have a denite, xed volume V . If we change any one of the properties
T , p, or ni , there is usually a change in V . The value of V is dependent on the other
properties and cannot be varied independently of them. Thus, for a given substance or
mixture of substances in a uniform uid phase, V is a unique function of T , p, and fnig,
where fnig stands for the set of amounts of all substances in the phase. We may be able
to express this relation in an explicit equation: V D f .T; p; fnig/. This equation (or a
rearranged form) that gives a relation among V , T , p, and fnig, is the equation of state of
the uid.
We may solve the equation of state, implicitly or explicitly, for any one of the quantities
V , T , p, and ni in terms of the other quantities. Thus, of the 3 C s quantities (where s is
the number of substances), only 2 C s are independent.
The ideal gas equation, p D nRT =V (Eq. 1.2.5 on page 23), is an equation of state.
It is found experimentally that the behavior of any gas in the limit of low pressure, as
temperature is held constant, approaches this equation of state. This limiting behavior is
also predicted by kinetic-molecular theory.
If the uid has only one constituent (i.e., is a pure substance rather than a mixture), then
at a xed T and p the volume is proportional to the amount. In this case, the equation of

34
state may be expressed as a relation among T , p, and the molar volume Vm D V =n. The
equation of state for a pure ideal gas may be written p D RT =Vm .
The RedlichKwong equation is a two-parameter equation of state frequently used to
describe, to good accuracy, the behavior of a pure gas at a pressure where the ideal gas
equation fails:
RT
a
pD
(2.2.1)
Vm b Vm .Vm C b/T 1=2
In this equation, a and b are constants that are independent of temperature and depend on
the substance.
The next section describes features of virial equations, an important class of equations
of state for real (nonideal) gases.
2.2.5 Virial equations of state for pure gases

In later chapters of this book there will be occasion to apply thermodynamic derivations to
virial equations of state of a pure gas or gas mixture. These formulas accurately describe
the gas at low and moderate pressures using empirically determined, temperature-dependent
parameters. The equations may be derived from statistical mechanics, so they have a theoretical as well as empirical foundation. There are two forms of virial equations for a pure
gas: one a series in powers of 1=Vm :
C
B
C 2C
pVm D RT 1 C
(2.2.2)
Vm
Vm
and the other a series in powers of p:
pVm D RT 1 C Bp p C Cp p 2 C
(2.2.3)
The parameters B, C , : : : are called the second, third, : : : virial coefcients, and the parameters Bp , Cp , : : : are a set of pressure virial coefcients. Their values depend on the
substance and are functions of temperature. (The rst virial coefcient in both power series is 1, because pVm must approach RT as 1=Vm or p approach zero at constant T .)
Coefcients beyond the third virial coefcient are small and rarely evaluated.
The values of the virial coefcients for a gas at a given temperature can be determined
from the dependence of p on Vm at this temperature. The value of the second virial coefcient B depends on pairwise interactions between the atoms or molecules of the gas, and
in some cases can be calculated to good accuracy from statistical mechanics theory and a
realistic intermolecular potential function.
To nd the relation between the virial coefcients of Eq. 2.2.2 and the parameters Bp ,
Cp , : : : in Eq. 2.2.3, we solve Eq. 2.2.2 for p in terms of Vm
1
B
p D RT
C 2C
(2.2.4)
Vm
Vm
and substitute in the right side of Eq. 2.2.3:
"
1
B
pVm D RT 1 C Bp RT
C 2C
Vm
Vm
#
2
1
B
C Cp .RT /2
C 2C
C
(2.2.5)
Vm
Vm

35
Then we equate coefcients of equal powers of 1=Vm in Eqs. 2.2.2 and 2.2.5 (since both
equations must yield the same value of pVm for any value of 1=Vm ):
B D RTBp
(2.2.6)
2
C D Bp RTB C Cp .RT /2 D .RT /2 .Bp C Cp /
(2.2.7)
In the last equation, we have substituted for B from Eq. 2.2.6.

At pressures up to at least one bar, the terms beyond Bp p in the pressure power series
of Eq. 2.2.3 are negligible; then pVm may be approximated by RT .1 C Bp p/, giving, with
the help of Eq. 2.2.6, the simple approximate equation of state3
Vm
RT
CB
p
(2.2.8)
(pure gas, p 1 bar)
The compression factor (or compressibility factor) Z of a gas is dened by

def
Z D
pV
pVm
D
nRT
RT
(2.2.9)
(gas)
When a gas at a particular temperature and pressure satises the ideal gas equation, the
value of Z is 1. The virial equations rewritten using Z are
Z D1C
C
B
C 2C
Vm
Vm
Z D 1 C Bp p C Cp p 2 C
(2.2.10)
(2.2.11)
These equations show that the second virial coefcient B is the initial slope of the curve of
a plot of Z versus 1=Vm at constant T , and Bp is the initial slope of Z versus p at constant
T.
The way in which Z varies with p at different temperatures is shown for the case of
carbon dioxide in Fig. 2.3(a) on the next page.
A temperature at which the initial slope is zero is called the Boyle temperature, which
for CO2 is 710 K. Both B and Bp must be zero at the Boyle temperature. At lower temperatures B and Bp are negative, and at higher temperatures they are positivesee Fig. 2.3(b).
This kind of temperature dependence is typical for other gases. Experimentally, and also
according to statistical mechanical theory, B and Bp for a gas can be zero only at a single
Boyle temperature.
The fact that at any temperature other than the Boyle temperature B is nonzero is
signicant since it means that in the limit as p approaches zero at constant T and the
gas approaches ideal-gas behavior, the difference between the actual molar volume Vm
and the ideal-gas molar volume RT =p does not approach zero. Instead, Vm RT =p
approaches the nonzero value B (see Eq. 2.2.8). However, the ratio of the actual and
ideal molar volumes, Vm =.RT =p/, approaches unity in this limit.
Virial equations of gas mixtures will be discussed in Sec. 9.3.4.

3 Guggenheim
(Ref. [71]) calls a gas with this equation of state a slightly imperfect gas.

2.3 S OME BASIC P ROPERTIES AND T HEIR M EASUREMENT
/cm mol
36
/bar
(a)
/K
(b)
Figure 2.3 (a) Compression factor of CO2 as a function of pressure at three temperatures. At 710 K, the Boyle temperature, the initial slope is zero.
(b) Second virial coefcient of CO2 as a function of temperature.
2.2.6 Solids
A solid phase responds to a small applied stress by undergoing a small elastic deformation.
When the stress is removed, the solid returns to its initial shape and the properties return to
those of the unstressed solid. Under these conditions of small stress, the solid has an equation of state just as a uid does, in which p is the pressure of a uid surrounding the solid
(the hydrostatic pressure) as explained in Sec. 2.3.4. The stress is an additional independent
variable. For example, the length of a metal spring that is elastically deformed is a unique
function of the temperature, the pressure of the surrounding air, and the stretching force.
If, however, the stress applied to the solid exceeds its elastic limit, the response is plastic
deformation. This deformation persists when the stress is removed, and the unstressed solid
no longer has its original properties. Plastic deformation is a kind of hysteresis, and is
caused by such microscopic behavior as the slipping of crystal planes past one another in a
crystal subjected to shear stress, and conformational rearrangements about single bonds in
a stretched macromolecular ber. Properties of a solid under plastic deformation depend on
its past history and are not unique functions of a set of independent variables; an equation
of state does not exist.
2.3 SOME BASIC PROPERTIES AND THEIR MEASUREMENT

This section macroscopic discusses aspects of the macroscopic properties mass, volume,
density, pressure, and temperature, with examples of how these properties can be measured.
2.3.1 Mass
We may measure the mass of an object with a balance utilizing the downward force exerted on the object by the earths gravitational eld. The classic balance has a beam and
knife-edge arrangement to compare the gravitational force on the body of interest with the

37
gravitational force on a weight of known mass. A modern balance (strictly speaking a

scale) incorporates a strain gauge or comparable device to directly measure the gravitational force on the unknown mass; this type must be calibrated with known masses. For the
most accurate measurements, we must take into account the effect of the buoyancy of the
body and the calibration masses in air. The accuracy of the calibration masses should be
traceable to a national standard kilogram (which in the United States is maintained at the
National Institute of Standards and Technology, formerly the National Bureau of Standards,
in Gaithersburg, Maryland) and ultimately to the international prototype of the kilogram
located at the International Bureau of Weights and Measures in S` vres, France.
e
From the measured mass of a sample of a pure substance, we can calculate the amount
of substance (called simply the amount in this book). The SI base unit for amount is the
mole (Sec. 1.1.1). Chemists are familiar with the fact that, although the mole is a counting
unit, an amount in moles is measured not by counting but by weighing. This is possible
because one mole is dened as the amount of atoms in exactly 12 grams of carbon-12, the
most abundant isotope of carbon (Appendix A). One mole of a substance has a mass of Mr
grams, where Mr is the relative molecular mass (or molecular weight) of the substance, a
dimensionless quantity.
A quantity related to molecular weight is the molar mass of a substance, dened as the
mass divided by the amount:
def m
Molar mass D M D
(2.3.1)
n
(The symbol M for molar mass is an exception to the rule given on page 29 that a subscript
m is used to indicate a molar quantity.) The numerical value of the molar mass expressed
in units of g mol 1 is equal to the relative molecular mass:
M=g mol
D Mr
(2.3.2)
2.3.2 Volume
We commonly measure liquid volumes with precision volumetric glassware such as burets,
pipets, and volumetric asks. The National Institute of Standards and Technology in the
United States has established specications for Class A glassware; two examples are listed
in Table 2.2 on the next page. We may accurately determine the volume of a vessel at one
temperature from the mass of a liquid of known density, such as water, that lls the vessel
at this temperature.
The SI unit of volume is the cubic meter, but chemists commonly express volumes in
units of liters and milliliters. The liter is dened as one cubic decimeter (Table 1.3). One
cubic meter is the same as 103 liters and 106 milliliters. The milliliter is identical to the
cubic centimeter.
Before 1964, the liter had a different denition: it was the volume of 1 kilogram of
water at 3:98 C, the temperature of maximum density. This denition made one liter
equal to 1:000028 dm3 . Thus, a numerical value of volume (or density) reported before
1964 and based on the liter as then dened may need a small correction in order to be
consistent with the present denition of the liter.

38
Table 2.2 Representative measurement methods

Physical
quantity
Mass
Volume
Density
Pressure
Temperature
Method
Typical
value
Approximate
uncertainty
analytical balance
microbalance
pipet, Class A
volumetric ask, Class A
pycnometer, 25-mL capacity
magnetic oat densimeter
vibrating-tube densimeter
mercury manometer or barometer
diaphragm gauge
constant-volume gas thermometer
mercury-in-glass thermometer
platinum resistance thermometer
monochromatic optical pyrometer
100 g
20 mg
10 mL
1L
1 g mL 1
1 g mL 1
1 g mL 1
760 Torr
100 Torr
10 K
300 K
300 K
1300 K
0:1 mg
0:1 g
0:02 mL
0:3 mL
2 mg mL 1
0:1 mg mL 1
0:01 mg mL 1
0:001 Torr
1 Torr
0:001 K
0:01 K
0:0001 K
0:03 K
2.3.3 Density
Density, an intensive property, is dened as the ratio of the two extensive properties mass
and volume:
def m
(2.3.3)
D
V
The molar volume Vm of a homogeneous pure substance is inversely proportional to its
density. From Eqs. 2.1.2, 2.3.1, and 2.3.3, we obtain the relation
Vm D
(2.3.4)
Various methods are available for determining the density of a phase, many of them
based on the measurement of the mass of a xed volume or on a buoyancy technique.
Three examples are shown in Fig. 2.4 on the next page. Similar apparatus may be used for
gases. The density of a solid may be determined from the volume of a nonreacting liquid
(e.g., mercury) displaced by a known mass of the solid, or from the loss of weight due to
buoyancy when the solid is suspended by a thread in a liquid of known density.
2.3.4 Pressure
Pressure is a force per unit area. Specically, it is the normal component of stress exerted
by an isotropic uid on a surface element.4 The surface can be an interface surface between
the uid and another phase, or an imaginary dividing plane within the uid.
Pressure is usually a positive quantity. Because cohesive forces exist in a liquid, it may
be possible to place the liquid under tension and create a negative pressure. For instance,
the pressure is negative at the top of a column of liquid mercury suspended below the closed
4A
liquid crystal and a polar liquid in a electric eld are examples of uids that are not isotropicthat is, that
have different macroscopic properties in different directions.

S
(a)
39
(b)
(c)
Figure 2.4 Three methods for measuring liquid density by comparison with samples
of known density. The liquid is indicated by gray shading.
(a) Glass pycnometer vessel with capillary stopper. The lled pycnometer is brought
to the desired temperature in a thermostat bath, dried, and weighed.
(b) Magnetic oat densimeter. a Buoy B, containing a magnet, is pulled down and
kept in position with solenoid S by means of position detector D and servo control
system C . The solenoid current required depends on the liquid density.
(c) Vibrating-tube densimeter. The ends of a liquid-lled metal U-tube are clamped to
a stationary block. An oscillating magnetic eld at the tip of the tube is used to make it
vibrate in the direction perpendicular to the page. The measured resonance frequency
is a function of the mass of the liquid in the tube.
a Ref.
[69].
end of a capillary tube that has no vapor bubble. Negative pressure in a liquid is an unstable
condition that can result in spontaneous vaporization.
The SI unit of pressure is the pascal. Its symbol is Pa. One pascal is a force of one
newton per square meter (Table 1.2).
Chemists are accustomed to using the non-SI units of millimeters of mercury, torr, and
atmosphere. One millimeter of mercury (symbol mmHg) is the pressure exerted by a column exactly 1 mm high of a uid of density equal to exactly 13:5951 g cm 3 (the density
of mercury at 0 C) in a place where the acceleration of free fall has its standard value gn
(see Appendix B). One atmosphere is dened as exactly 1:01325 105 Pa (Table 1.3). The
torr is dened by letting one atmosphere equal exactly 760 Torr. One atmosphere is approximately 760 mmHg. In other words, the millimeter of mercury and the torr are practically
identical; they differ from one another by less than 2 10 7 Torr.
Another non-SI pressure unit is the bar, equal to exactly 105 Pa. A pressure of one
bar is approximately one percent smaller than one atmosphere. This book often refers to a
standard pressure, p . In the past, the value of p was usually taken to be 1 atm, but since
1982 the IUPAC has recommended the value p D 1 bar.
A variety of manometers and other devices is available to measure the pressure of a
uid, each type useful in a particular pressure range. Some devices measure the pressure of
the uid directly. Others measure the differential pressure between the uid and the atmosphere; the uid pressure is obtained by combining this measurement with the atmospheric
pressure measured with a barometer.
Within a solid, we cannot dene pressure simply as a force per unit area. Macroscopic

40
forces at a point within a solid are described by the nine components of a stress tensor. The
statement that a solid has or is at a certain pressure means that this is the hydrostatic pressure
exerted on the solids exterior surface. Thus, a solid immersed in a uniform isotropic uid
of pressure p is at pressure p; if the uid pressure is constant over time, the solid is at
constant pressure.
2.3.5 Temperature
Temperature scales
Temperature and thermometry are of fundamental importance in thermodynamics. Unlike
the other physical quantities discussed in this chapter, temperature does not have a single
unique denition. The chosen denition, whatever it may be, requires a temperature scale
described by an operational method of measuring temperature values. For the scale to be
useful, the values should increase monotonically with the increase of what we experience
physiologically as the degree of hotness. We can dene a satisfactory scale with any
measuring method that satises this requirement. The values on a particular temperature
scale correspond to a particular physical quantity and a particular temperature unit.
For example, suppose you construct a simple liquid-in-glass thermometer with equally
spaced marks along the stem and number the marks consecutively. To dene a temperature
scale and a temperature unit, you could place the thermometer in thermal contact with a
body whose temperature is to be measured, wait until the indicating liquid reaches a stable
position, and read the meniscus position by linear interpolation between two marks.5
Thermometry is based on the principle that the temperatures of different bodies may
be compared with a thermometer. For example, if you nd by separate measurements with
your thermometer that two bodies give the same reading, you know that within experimental
error both have the same temperature. The signicance of two bodies having the same
temperature (on any scale) is that if they are placed in thermal contact with one another,
they will prove to be in thermal equilibrium with one another as evidenced by the absence
of any changes in their properties. This principle is sometimes called the zeroth law of
thermodynamics, and was rst stated as follows by J. C. Maxwell (1872): Bodies whose
temperatures are equal to that of the same body have themselves equal temperatures.6
Two particular temperature scales are used extensively. The ideal-gas temperature
scale is dened by gas thermometry measurements, as described on page 42. The thermodynamic temperature scale is dened by the behavior of a theoretical Carnot engine,
as explained in Sec. 4.3.4. These temperature scales correspond to the physical quantities called ideal-gas temperature and thermodynamic temperature, respectively. Although
the two scales have different denitions, the two temperatures turn out (Sec. 4.3.4) to be
proportional to one another. Their values become identical when the same unit of temperature is used for both. Thus, the kelvin is dened by specifying that a system containing
the solid, liquid, and gaseous phases of H2 O coexisting at equilibrium with one another
(the triple point of water) has a thermodynamic temperature of exactly 273:16 kelvins. We
5 Of
course, placing the thermometer and body in thermal contact may affect the bodys temperature. The
measured temperature is that of the body after thermal equilibrium is achieved.
6 Turner (Ref. [158]) argues that the zeroth law is a consequence of the rst and second laws and therefore is
not a separate assumption in the axiomatic framework of thermodynamics. The term law for this principle is
also questioned by Redlich (Ref. [140]).

41
set the ideal-gas temperature of this system equal to the same value, 273:16 kelvins. The
temperatures measured on the two scales are then identical.
Formally, the symbol T refers to thermodynamic temperature. Strictly speaking, a different symbol should be used for ideal-gas temperature. Since the two kinds of temperatures
have identical values, this book will use the symbol T for both and refer to both physical
quantities simply as temperature except when it is necessary to make a distinction.
Why is the temperature of the triple point of water taken to be 273:16 kelvins? This
value is chosen arbitrarily to make the steam point approximately one hundred kelvins
greater than the ice point.
The ice point is the temperature at which ice and air-saturated water coexist in equilibrium at a pressure of one atmosphere. The steam point is the temperature at which liquid
and gaseous H2 O coexist in equilibrium at one atmosphere. Neither of these temperatures
has sufcient reproducibility for high-precision work. The temperature of the ice-waterair system used to dene the ice point is affected by air bubbles in the ice and by varying
concentrations of air in the water around each piece of ice. The steam point is uncertain because the temperature of coexisting liquid and gas is a sensitive function of the experimental
pressure.
The obsolete centigrade scale was dened to give a value of exactly 0 degrees centigrade at the ice point and a value of exactly 100 degrees centigrade at the steam point, and
to be a linear function of an ideal-gas temperature scale.
The centigrade scale has been replaced by the Celsius scale, which is based on the
triple point of water rather than on the less reproducible ice point and steam point. The Celsius scale is the thermodynamic (or ideal-gas) temperature scale shifted by exactly 273:15
kelvins. The temperature unit is the degree Celsius ( C), identical in size to the kelvin.
Thus, Celsius temperature t is related to thermodynamic temperature T by
t = C D T =K
273:15
(2.3.5)
On the Celsius scale, the triple point of water is exactly 0:01 C. The ice point is 0 C to
within 0:0001 C, and the steam point is 99:97 C.
The International Temperature Scale of 1990
The International Temperature Scale of 1990 (abbreviated ITS-90) is the most recent scale
devised for practical high-precision temperature measurements.7 This scale denes the
physical quantity called international temperature, with symbol T90 . Each value of T90 is
intended to be very close to the corresponding thermodynamic temperature T .
The ITS-90 is dened over a very wide temperature range, from 0:65 K up to at least
1358 K. There is a specied procedure for each measurement of T90 , depending on the
range in which T falls: vapor-pressure thermometry (0:655:0 K), gas thermometry (3:0
24:5561 K), platinum-resistance thermometry (13:80331234:93 K), or optical pyrometry
(above 1234:93 K). For vapor-pressure thermometry, the ITS-90 provides formulas for T90
in terms of the vapor pressure of the helium isotopes 3 He and 4 He. For the other methods, it assigns values of several xed calibration temperatures. The xed temperatures are
achieved with the reproducible equilibrium systems listed in Table 2.3 on the next page.
7 Refs.
[113] and [134].

42
Table 2.3 Fixed temperatures of the

International Temperature Scale of 1990
T90 =K
13:8033
24:5561
54:3584
83:8058
234:3156
273:16
302:9146
429:7485
505:078
692:677
933:473
1234:93
1337:33
1357:77
Equilibrium system
H2 triple point
Ne triple point
O2 triple point
Ar triple point
Hg triple point
H2 O triple point
Ga melting point at 1 atm
In melting point at 1 atm
Sn melting point at 1 atm
Zn melting point at 1 atm
Al melting point at 1 atm
Ag melting point at 1 atm
Au melting point at 1 atm
Cu melting point at 1 atm
Equilibrium systems for xed temperatures

If two different phases of a pure substance coexist at a controlled, constant pressure, the
temperature has a denite xed value. As shown in Table 2.3, eight of the xed temperatures on the ITS-90 are obtained experimentally with coexisting solid and liquid phases at
a pressure of one atmosphere. These temperatures are normal melting points.
Triple-points of pure substances provide the most reproducible temperatures. Both temperature and pressure have denite xed values in a system containing coexisting solid, liquid, and gas phases of a pure substance. The table lists six temperatures xed on the ITS-90
by such systems.
Figure 2.5 on the next page illustrates a triple-point cell for water whose temperature
is capable of a reproducibility within 10 4 K. When ice, liquid water, and water vapor are
in equilibrium in this cell, the cell is at the triple point of water and the temperature, by
denition, is exactly 273:16 K.
Gas thermometry
Only the triple point of water has a dened value on the thermodynamic temperature scale.
How are the values of other xed temperatures of a scale such as the ITS-90 determined?
The fundamental method is gas thermometry, a method most commonly carried out with
a constant-volume gas thermometer. This device consists of a bulb or vessel containing a
thermometric gas and a means of measuring the pressure of this gas. The thermometric gas
is usually helium, because it has minimal deviations from ideal-gas behavior.
The simple constant-volume gas thermometer depicted in Fig. 2.6 on the next page
uses a mercury manometer to measure the pressure. More sophisticated versions have a
diaphragm pressure transducer between the bulb and the pressure measurement system.
The procedure for determining the value of an unknown temperature involves a pair of
pressure measurements. The gas is brought successively into thermal equilibrium with two

43
H O(g)
H O(l)
H O(s)
H O(l)
thermometer bulb
Figure 2.5 Cross-section of a water triple-point cell. The cell has cylindrical symmetry. Pure water of the same isotopic composition as H2 O in ocean water is distilled
into the cell. The air is pumped out and the cell is sealed. A freezing mixture is placed
in the inner well to cause a thin layer of ice to form next to the inner wall. The freezing
mixture is removed, and some of the ice is allowed to melt to a lm of very pure water
between the ice and inner wall. The thermometer bulb is placed in the inner well as
shown, together with ice water (not shown) for good thermal contact.
Hg
gas
Figure 2.6 Simple version of a constant-volume gas thermometer.
different systems: a reference system of known temperature T1 (such as one of the systems
listed in Table 2.3), and the system whose temperature T2 is to be measured. The pressures
p1 and p2 are measured at these temperatures. In the two equilibrations the amount of gas
is the same and the gas volume is the same except for a small change due to effects of T
and p on the bulb dimensions.
If the gas exactly obeyed the ideal gas equation in both measurements, we would have
nR D p1 V1 =T1 D p2 V2 =T2 or T2 D T1 .p2 V2 =p1 V1 /. Since, however, the gas approaches

44
ideal behavior only in the limit of low pressure, it is necessary to make a series of the paired
measurements, changing the amount of gas in the bulb before each new pair so as to change
the measured pressures in discrete steps. Thus, the operational equation for evaluating the
unknown temperature is
p2 V2
p1 !0 p1 V1
T2 D T1 lim
(2.3.6)
(gas)
(The ratio V2 =V1 differs from unity only because of any change in the bulb volume when T
and p change.) The limiting value of p2 V2 =p1 V1 can be obtained by plotting this quantity
against p1 , 1=Vm , or another appropriate extrapolating function. Note that values of n and
R are not needed.
Another method is possible if the value of the second virial coefcient at the reference
temperature T1 is known. This value can be used in the virial equation (Eq. 2.2.2) together
with the values of T1 and p1 to evaluate the molar volume Vm . Then, assuming Vm is the
same in both equilibrations of a measurement pair, it is possible to evaluate p2 Vm =R at
temperature T2 , and T2 can be found from
p2 V m
p2 !0 R
T2 D lim
(2.3.7)
(gas)
Constant-volume gas thermometry can be used to evaluate the second virial coefcient
of the gas at temperature T2 if the value at T1 is known (Prob. 2.3).
The principles of measurements with a gas thermometer are simple, but in practice great
care is needed to obtain adequate precision and accuracy.8 Corrections or precautions are
required for such sources of error as thermal expansion of the gas bulb, dead volume
between the bulb and pressure measurement system, adsorption of the thermometric gas on
interior surfaces, and desorption of surface contaminants.
The national laboratories of several countries, including the National Institute of Standards and Technology in the United States, maintain stable secondary thermometers (e.g.,
platinum resistance thermometers and thermocouples) that have been calibrated according
to the ITS-90. These secondary thermometers are used as working standards to calibrate
other laboratory and commercial temperature-measuring devices.
Practical thermometers
Liquid-in-glass thermometers use indicating liquids whose volume change with temperature
is much greater than that of the glass. A mercury-in-glass thermometer can be used in the
range 234 K (the freezing point of mercury) to 600 K, and typically can be read to 0:01 K.
A Beckmann thermometer covers a range of only a few kelvins but can be read to 0:001 K.
A resistance thermometer is included in a circuit that measures the electric resistance.
Platinum resistance thermometers are widely used because of their stability and high sensitivity (0:0001 K). Thermistors use metal oxides and can be made very small; they have
greater sensitivity than platinum thermometers but are not as stable over time.
8 In
1899 H. L. Callendar, the British physicist who introduced the platinum resistance thermometer, wrote
(Ref. [26]) It is impossible for those who have never worked with a gas-thermometer to realize the extent of
its shortcomings.

2.4 T HE S TATE OF THE S YSTEM
45
Figure 2.7 A thermopile used to measure the difference in the temperatures of two
bodies.
A thermocouple consists of wires of two dissimilar metals (e.g., constantan alloy and
copper) connected in series at soldered or welded junctions. A many-junction thermocouple
is called a thermopile (Fig. 2.7). When adjacent junctions are placed in thermal contact with
bodies of different temperatures, an electric potential develops that is a function of the two
temperatures.
Finally, two other temperature-measuring devices are the quartz crystal thermometer,
incorporating a quartz crystal whose resonance frequency is temperature dependent, and
optical pyrometers, which are useful above about 1300 K to measure the radiant intensity
of a black body emitter.
2.4 THE STATE OF THE SYSTEM

The thermodynamic state of the system is an important and subtle concept.9 At each instant of time, the system is in some denite state that we may describe with values of the
macroscopic properties we consider to be relevant for our purposes. The values of these
properties at any given instant dene the state at that instant. Whenever the value of any
of these properties changes, the state has changed. If we subsequently nd that each of the
relevant properties has the value it had at a certain previous instant, then the system has
returned to its previous state.
2.4.1 State functions and independent variables

The properties whose values at each instant depend only on the state of the system at that
instant, and not on the past or future history of the system, are called state functions (or
state variables or state parameters). There may be other system properties that we consider
to be irrelevant to the state, such as the shape of the system, and these are not state functions.
9 Do
not confuse the state of the system with the kind of physical state or state of aggregation of a phase
discussed in Sec. 2.2.1. A change of state refers to a change in the state of the system, not necessarily to a phase
transition.

46
Table 2.4 Values of state functions of an aqueous sucrose solution (A = water, B = sucrose)
temperature
pressure
amount of water
amount of sucrose
volume
mass
density
mole fraction of sucrose
osmotic pressure
refractive index, sodium D line
T D 293:15 K
p D 1:01 bar
nA D 39:18 mol
nB D 1:375 mol
V D 1000 cm3
m D 1176:5 g
D 1:1765 g cm
xB D 0:03390
D 58:2 bar
nD D 1:400
Various conditions determine what states of a system are physically possible. If a uniform phase has an equation of state, property values must be consistent with this equation.
The system may have certain built-in or externally-imposed conditions or constraints that
keep some properties from changing with time. For instance, a closed system has constant
mass; a system with a rigid boundary has constant volume. We may know about other
conditions that affect the properties during the time the system is under observation.
We can dene the state of the system with the values of a certain minimum number of
state functions which we treat as the independent variables. Once we have selected a set of
independent variables, consistent with the physical nature of the system and any conditions
or constraints, we can treat all other state functions as dependent variables whose values
depend on the independent variables.
Whenever we adjust the independent variables to particular values, every other state
function is a dependent variable that can have only one denite, reproducible value. For
example, in a single-phase system of a pure substance with T , p, and n as the independent
variables, the volume is determined by an equation of state in terms of T , p, and n; the
mass is equal to nM ; the molar volume is given by Vm D V =n; and the density is given by
D nM=V .
2.4.2 An example: state functions of a mixture

Table 2.4 lists the values of ten state functions of an aqueous sucrose solution in a particular
state. The rst four properties (T , p, nA , nB ) are ones that we can vary independently, and
their values sufce to dene the state for most purposes. Experimental measurements will
convince us that, whenever these four properties have these particular values, each of the
other properties has the one denite value listedwe cannot alter any of the other properties
without changing one or more of the rst four variables. Thus we can take T , p, nA , and
nB as the independent variables, and the six other properties as dependent variables. The
other properties include one (V ) that is determined by an equation of state; three (m, ,
and xB ) that can be calculated from the independent variables and the equation of state;
a solution property ( ) treated by thermodynamics (Sec. 12.4.4); and an optical property
(nD ). In addition to these six independent variables, this system has innumerable others:
energy, isothermal compressibility, heat capacity at constant pressure, and so on.

47
We could make other choices of the independent variables for the aqueous sucrose system. For instance, we could choose the set T , p, V , and xB , or the set p, V , , and xB .
If there are no imposed conditions, the number of independent variables of this system is
always four. (Note that we could not arbitrarily choose just any four variables. For instance,
there are only three independent variables in the set p, V , m, and because of the relation
D m=V .)
If the system has imposed conditions, there will be fewer independent variables. Suppose the sucrose solution is in a closed system with xed, known values of nA and nB ; then
there are only two independent variables and we could describe the state by the values of
just T and p.
2.4.3 More about independent variables

A closed system containing a single substance in a single phase has two independent variables, as we can see by the fact that the state is completely dened by values of T and p or
of T and V .
A closed single-phase system containing a mixture of several nonreacting substances,
or a mixture of reactants and products in reaction equilibrium, also has two independent
variables. Examples are
air, a mixture of gases in xed proportions;
an aqueous ammonia solution containing H2 O, NH3 , NH4 C , HC , OH , and probably
other species, all in rapid continuous equilibrium.
The systems in these two examples contain more than one substance, but only one component. The number of components of a system is the minimum number of substances or
mixtures of xed composition needed to form each phase.10 A system of a single pure substance is a special case of a system of one component. In an open system, the amount of
each component can be used as an independent variable.
Consider a system with more than one uniform phase. In principle, for each phase we
could independently vary the temperature, the pressure, and the amount of each substance
or component. There would then be 2 C C independent variables for each phase, where C
is the number of components in the phase.
There usually are, however, various equilibria and other conditions that reduce the number of independent variables. For instance, each phase may have the same temperature and
the same pressure; equilibrium may exist with respect to chemical reaction and transfer
between phases (Sec. 2.4.4); and the system may be closed. (While these various conditions do not have to be present, the relations among T , p, V , and amounts described by an
equation of state of a phase are always present.) On the other hand, additional independent
variables are required if we consider properties such as the surface area of a liquid to be
relevant.11
We must be careful to choose a set of independent variables that denes the state
without ambiguity. For a closed system of liquid water, the set p and V might be a
10 The
concept of the number of components is discussed in more detail in Chap. 13.
11 The important topic of the number of independent intensive variables is treated by the Gibbs phase rule, which
will be discussed in Sec. 8.1.7 for systems of a single substance and in Sec. 13.1 for systems of more than one
substance.

48
poor choice because the molar volume of water passes through a minimum as T is
varied at constant p. Thus, the values p D 1:000 bar and V D 18:016 cm3 would
describe one mole of water at both 2 C and 6 C, so these values would not uniquely
dene the state. Better choices of independent variables in this case would be either T
and p, or else T and V .
How may we describe the state of a system that has nonuniform regions? In this case we
may imagine the regions to be divided into many small volume elements, each small enough
to be essentially uniform but large enough to contain many molecules. We then describe
the state by specifying values of independent variables for each volume element. If there is
internal macroscopic motion (e.g., ow), then velocity components can be included among
the independent variables. Obviously, the quantity of information needed to describe a
complicated state may be enormous.
We can imagine situations in which classical thermodynamics would be completely
incapable of describing the state. For instance, turbulent ow in a uid or a shock wave in
a gas may involve inhomogeneities all the way down to the molecular scale. Macroscopic
variables would not sufce to dene the states in these cases.
Whatever our choice of independent variables, all we need to know to be sure a system
is in the same state at two different times is that the value of each independent variable is
the same at both times.
2.4.4 Equilibrium states

An equilibrium state is a state that, when present in an isolated system, remains unchanged
indenitely as long as the system remains isolated. (Recall that an isolated system is one
that exchanges no matter or energy with the surroundings.) An equilibrium state of an
isolated system has no natural tendency to change over time. If changes do occur in an
isolated system, they continue until an equilibrium state is reached.
A system in an equilibrium state may have some or all of the following kinds of internal
equilibria:
Thermal equilibrium: each phase has the same temperature.
Mechanical equilibrium: each phase has the same pressure.
Transfer equilibrium: there is equilibrium with respect to the transfer of each species
from one phase to another.
Reaction equilibrium: every possible chemical reaction is at equilibrium.
A homogeneous system has a single phase of uniform temperature and pressure, and so
has thermal and mechanical equilibrium. It is in an equilibrium state if it also has reaction
equilibrium.
A heterogeneous system is in an equilibrium state if each of the four kinds of internal
equilibrium is present.
The meaning of internal equilibrium in the context of an equilibrium state is that no
perceptible change of state occurs during the period we keep the isolated system under
observation. For instance, a system containing a homogeneous mixture of gaseous H2 and
O2 at 25 C and 1 bar is in a state of reaction equilibrium on a time scale of hours or days; but
if a measurable amount of H2 O forms over a longer period, the state is not an equilibrium

49
state on this longer time scale. This consideration of time scale is similar to the one we
apply to the persistence of deformation in distinguishing a solid from a uid (Sec. 2.2.1).
Even if a system is not in internal equilibrium, it can be in an equilibrium state if a
change of state is prevented by an imposed internal constraint or the inuence of an external
eld. Here are ve examples of such states.
A system with an internal adiabatic partition separating two phases can be in an equilibrium state that is not in thermal equilibrium. The adiabatic partition allows the two
phases to remain indenitely at different temperatures. If the partition is rigid, it can
also allow the two phases to have different pressures, so that the equilibrium state
lacks mechanical equilibrium.
An experimental system used to measure osmotic pressure (Fig. 12.2 on page 373)
has a semipermeable membrane separating a liquid solution phase and a pure solvent
phase. The membrane prevents the transfer of solute from the solution to the pure
solvent. In the equilibrium state of this system, the solution has a higher pressure than
the pure solvent; the system is then in neither transfer equilibrium nor mechanical
equilibrium.
In the equilibrium state of a galvanic cell that is not part of a closed electrical circuit
(Sec. 3.8.3), the separation of the reactants and products and the open circuit are
constraints that prevent the cell reaction from coming to reaction equilibrium.
A system containing mixed reactants of a reaction can be in an equilibrium state
without reaction equilibrium if we withhold a catalyst or initiator or introduce an
inhibitor that prevents reaction. In the example above of a mixture of H2 and O2
gases, we could consider the high activation energy barrier for the formation of H2 O
to be an internal constraint. If we remove the constraint by adding a catalyst, the
reaction will proceed explosively.
An example of a system inuenced by an external eld is a tall column of gas in a
gravitational eld (Sec. 8.1.4). In order for an equilibrium state to be established in
this eld, the pressure must decrease continuously with increasing elevation.
Keep in mind that regardless of the presence or absence of internal constraints and
external elds, the essential feature of an equilibrium state is this: if we isolate the system
while it is in this state, the state functions do not change over time.
Three additional comments can be made regarding the properties of equilibrium states.
1. It should be apparent that a system with thermal equilibrium has a single value of
T , and one with mechanical equilibrium has a single value of p, and this allows the
state to be described by a minimal number of independent variables. In contrast, the
denition of a nonequilibrium state with nonuniform intensive properties may require
a very large number of independent variables.
2. Strictly speaking, during a time period in which the system exchanges energy with the
surroundings its state cannot be an equilibrium state. Energy transfer at a nite rate
causes nonuniform temperature and pressure within the system and prevents internal
thermal and mechanical equilibrium. If, however, the rate of energy transfer is small,
then at each instant the state can closely approximate an equilibrium state. This topic
will be discussed in detail in the next chapter.

2.5 P ROCESSES AND PATHS
50
Figure 2.8 Steady state in a metal rod (shaded) with heat conduction. The boxes at
the ends represent heat reservoirs of constant temperature.
3. The concept of an equilibrium state assumes that when the system is in this state and
isolated, no observable change occurs during the period of experimental observation.
If the state does, in fact, undergo a slow change that is too small to be detected during
the experimental observation period texp , the state is metastablethe relaxation
time of the slow change is much greater than texp . There is actually no such thing
as a true equilibrium state, because very slow changes inevitably take place that we
have no way of controlling. One example was mentioned above: the slow formation
of water from its elements in the absence of a catalyst. Atoms of radioactive elements
with long half-lives slowly change to other elements. More generally, all elements are
presumably subject to eventual transmutation to iron-58 or nickel-62, the nuclei with
the greatest binding energy per nucleon. When we use the concept of an equilibrium
state, we are in effect assuming that rapid changes that have come to equilibrium
have relaxation times much shorter than texp and that the relaxation times of all
other changes are innite.
2.4.5 Steady states

It is important not to confuse an equilibrium state with a steady state, a state that is constant during a time period during which the system exchanges matter or energy with the
surroundings.
The heat-conducting metal rod shown in Fig. 2.8 is a system in such a steady state.
Each end of the rod is in thermal contact with a heat reservoir (or thermal reservoir),
which is a body or external system whose temperature remains constant and uniform when
there is heat transfer to or from it.12 The two heat reservoirs in the gure have different
temperatures, causing a temperature gradient to form along the length of the rod and energy
to be transferred by heat from the warmer reservoir to the rod and from the rod to the cooler
reservoir. Although the properties of the steady state of the rod remain constant, the rod is
clearly not in an equilibrium state because the temperature gradient will quickly disappear
when we isolate the rod by removing it from contact with the heat reservoirs.
2.5 PROCESSES AND PATHS

A process is a change in the state of the system over time, starting with a denite initial
state and ending with a denite nal state. The process is dened by a path, which is the
12 A
heat reservoir can be a body that is so large that its temperature changes only imperceptibly during heat
transfer; a thermostat bath whose temperature can be controlled; or an external system of coexisting phases of
a pure substance (e.g., ice and water) at constant pressure.

2.5 P ROCESSES AND PATHS
51
(a)
(b)
(c)
Figure 2.9 Paths of three processes of a closed ideal-gas system with p and V as the
independent variables. (a) Isothermal expansion. (b) Isobaric expansion. (c) Isochoric
pressure reduction.
continuous sequence of consecutive states through which the system passes, including the
initial state, the intermediate states, and the nal state. The process has a direction along
the path. The path could be described by a curve in an N-dimensional space in which each
coordinate axis represents one of the N independent variables.
This book takes the view that a thermodynamic process is dened by what happens
within the system, in the three-dimensional region up to and including the boundary, and by
the forces exerted on the system by the surroundings and any external eld. Conditions and
changes in the surroundings are not part of the process except insofar as they affect these
forces. For example, consider a process in which the system temperature decreases from
300 K to 273 K. We could accomplish this temperature change by placing the system in
thermal contact with either a refrigerated thermostat bath or a mixture of ice and water. The
process is the same in both cases, but the surroundings are different.
Expansion is a process in which the system volume increases; in compression, the
volume decreases.
An isothermal process is one in which the temperature of the system remains uniform
and constant. An isobaric or isopiestic process refers to uniform constant pressure, and an
isochoric process refers to constant volume. Paths for these processes of an ideal gas are
shown in Fig. 2.9.
An adiabatic process is one in which there is no heat transfer across any portion of the
boundary. We may ensure that a process is adiabatic either by using an adiabatic boundary
or, if the boundary is diathermal, by continuously adjusting the external temperature to
eliminate a temperature gradient at the boundary.
Recall that a state function is a property whose value at each instant depends only on
the state of the system at that instant. The nite change of a state function X in a process is
written X. The notation X always has the meaning X2 X1 , where X1 is the value in
the initial state and X2 is the value in the nal state. Therefore, the value of X depends
only on the values of X1 and X2 . The change of a state function during a process depends
only on the initial and nal states of the system, not on the path of the process.
An innitesimal change of the state function X is written dX. The mathematical operation of summing an innite number of innitesimal changes is integration, and the sum
is an integral (see the brief calculus review in Appendix E). The sum of the innitesimal

2.6 T HE E NERGY OF THE S YSTEM
changes of X along a path is a denite integral equal to X :

Z X2
dX D X2 X1 D X
52
(2.5.1)
X1
If dX obeys this relationthat is, if its integral for given limits has the same value regardless
of the pathit is called an exact differential. The differential of a state function is always
an exact differential.
A cyclic process is a process in which the state of the system changes and then returns
H
to the initial state. In this case the integral of dX is written with a cyclic integral sign: dX .
Since a state function X has the same initial and nal values in a cyclic process, X2 is equal
to X1 and the cyclic integral of dX is zero:
I
dX D 0
(2.5.2)
Heat (q) and work (w) are examples of quantities that are not state functions. They are
not even properties of the system; instead they are quantities of energy transferred across
the boundary over a period of time. It would therefore be incorrect to write q or w.
Instead, the values of q and w depend in general on the path and are called path functions.
This book uses the symbol (the letter d with a bar through the stem) for an innitesimal quantity of a path function. Thus, q and w are innitesimal quantities of heat and
work. The sum of many innitesimal quantities of a path function is not the difference of
two values of the path function; instead, the sum is the net quantity:
Z
Z
q D q
w D w
(2.5.3)
The innitesimal quantities q and w, because the values of their integrals depend on the
path, are inexact differentials.13
There is a fundamental difference between a state function (such as temperature or
volume) and a path function (such as heat or work): The value of a state function refers to
one instant of time; the value of a path function refers to an interval of time.
The difference between a state function and a path function in thermodynamics is
analogous to the difference between elevation and trail length in hiking up a mountain.
Suppose a trailhead at the base of the mountain has several trails to the summit. The
hiker at each instant is at a denite elevation above sea level. During a climb from the
trailhead to the summit, the hikers change of elevation is independent of the trail used,
but the trail length from base to summit depends on the trail.
2.6 THE ENERGY OF THE SYSTEM

A large part of classical thermodynamics is concerned with the energy of the system. The
total energy of a system is an extensive property whose value at any one instant cannot be
measured in any practical way, but whose change is the focus of the rst law of thermodynamics (Chap. 3).
13 Chemical
thermodynamicists often write these quantities as dq and dw. Mathematicians, however, frown on
using the same notation for inexact and exact differentials. Other notations sometimes used to indicate that heat
and work are path functions are Dq and Dw, and also q and w.

53
2.6.1 Energy and reference frames

Classical thermodynamics ignores microscopic properties such as the behavior of individual
atoms and molecules. Nevertheless, a consideration of the classical mechanics of particles
will help us to understand the sources of the potential and kinetic energy of a thermodynamic system.
In classical mechanics, the energy of a collection of interacting point particles is the
sum of the kinetic energy 1 mv 2 of each particle (where m is the particles mass and v is
2
its velocity), and of various kinds of potential energies. The potential energies are dened
in such a way that if the particles are isolated from the rest of the universe, as the particles
move and interact with one another the total energy (kinetic plus potential) is constant over
time. This principle of the conservation of energy also holds for real atoms and molecules
whose electronic, vibrational, and rotational energies, absent in point particles, are additional contributions to the total energy.
The positions and velocities of particles must be measured in a specied system of
coordinates called a reference frame. This book will use reference frames with Cartesian
axes. Since the kinetic energy of a particle is a function of velocity, the kinetic energy
depends on the choice of the reference frame. A particularly important kind is an inertial
frame, one in which Newtons laws of motion are obeyed (see Sec. G.1 in Appendix G).
A reference frame whose axes are xed relative to the earths surface is what this book
will call a lab frame. A lab frame for all practical purposes is inertial (Sec. G.10 on
page 504). It is in this kind of stationary frame that the laws of thermodynamics have
been found by experiment to be valid.
The energy E of a thermodynamic system is the sum of the energies of the particles
contained in it and the potential energies of interaction between these particles. Just as for
an individual particle, the energy of the system depends on the reference frame in which it
is measured. The energy of the system may change during a process, but the principle of
the conservation of energy ensures that the sum of the energy of the system, the energy of
the surroundings, and any energy shared by both, all measured in the same reference frame,
remains constant over time.
This book uses the symbol Esys for the energy of the system measured in a specied
inertial frame. The system could be located in a weightless environment in outer space, and
the inertial frame could be one that is either xed or moving at constant velocity relative to
local stars. Usually, however, the system is located in the earths gravitational eld, and the
appropriate inertial frame is then an earth-xed lab frame.
If during a process the system as a whole undergoes motion or rotation relative to the
inertial frame, then Esys depends in part on coordinates that are not properties of the system.
In such situations Esys is not a state function, and we need the concept of internal energy.
2.6.2 Internal energy

The internal energy, U, is the energy of the system measured in a reference frame that
allows U to be a state functionthat is, at each instant the value of U depends only on the
state of the system. This book will call a reference frame with this property a local frame.
A local frame may also be, but is not necessarily, an earth-xed lab frame.

54
Here is a simple illustration of the distinction between the energy Esys of a system
measured in a lab frame and the internal energy U measured in a local frame. Let the system
be a xed amount of water contained in a glass beaker. (The glass material of the beaker
is part of the surroundings.) We can dene the state of this system by two independent
variables: the temperature, T , and pressure, p, of the water. The most convenient local
frame in which to measure U in this case is a frame xed with respect to the beaker.
When the beaker is at rest on the lab bench, the local frame is a lab frame; then the
energies Esys and U are equal and depend only on T and p.
If we place the beaker on a laboratory hot plate and use the hot plate to raise the
temperature of the water, the values of Esys and U increase equally.
Suppose we slide the beaker horizontally along the lab bench while T and p stay
constant. While the system is in motion, its kinetic energy is greater in the lab frame
than in the local frame. Now Esys is greater than when the beaker was at rest, although
the state of the system and the value of U are unchanged.
If we slowly lift the beaker above the bench, the potential energy of the water in the
earths gravitational eld increases, again with no change in T and p. The value of
Esys has increased, but there has been no change in the state of the system or the value
of U .
Section 3.1.1 will show that the relation between changes of the system energy and the
internal energy in this example is Esys D Ek C Ep C U , where Ek and Ep are the
kinetic and potential energies of the system as a whole measured in the lab frame.
Our choice of the local frame used to dene the internal energy U of any particular
system during a given process is to some extent arbitrary. Three possible choices are as
follows.
If the system as a whole does not move or rotate in the laboratory, a lab frame is an
appropriate local frame. Then U is the same as the system energy Esys measured in
the lab frame.
If the systems center of mass moves in the lab frame during the process, we can let
the local frame be a center-of-mass frame whose origin moves with the center of mass
and whose Cartesian axes are parallel to the axes of the lab frame.
If the system consists of the contents of a rigid container that moves or rotates in the
lab, as in the illustration above, it may be convenient to choose a local frame that has
its origin and axes xed with respect to the container.
Is it possible to determine a numerical value for the internal energy of a system? The
2
total energy of a body of mass m when it is at rest is given by the Einstein relation E D mc0 ,
where c0 is the speed of light in vacuum. In principle, then, we could calculate the internal
energy U of a system at rest from its mass, and we could determine U for a process
from the change in mass. In practice, however, an absolute value of U calculated from a
measured mass has too much uncertainty to be of any practical use. For example, the typical
uncertainty of the mass of an object measured with a microbalance, about 0:1 g (Table
2.2), would introduce the enormous uncertainty in energy of about 1010 joules. Only values
of the change U are useful, and these values cannot be calculated from m because the
change in mass during an ordinary chemical process is much too small to be detected.

P ROBLEMS
55
PROBLEMS
2.1 Let X represent the quantity V 2 with dimensions .length/6 . Give a reason that X is or is not
an extensive property. Give a reason that X is or is not an intensive property.
2.2 Calculate the relative uncertainty (the uncertainty divided by the value) for each of the measurement methods listed in Table 2.2 on page 38, using the typical values shown. For each of
the ve physical quantities listed, which measurement method has the smallest relative uncertainty?
2.3 Table 2.5 lists data obtained from a constant-volume gas thermometer containing samples of
Table 2.5 Helium at a xed temperature
.1=Vm / =102 mol m
1.0225
1.3202
1.5829
1.9042
2.4572
2.8180
3.4160
3.6016
4.1375
4.6115
5.1717
.p2 Vm =R/ =K
2.7106
2.6994
2.6898
2.6781
2.6580
2.6447
2.6228
2.6162
2.5965
2.5790
2.5586
varying amounts of helium maintained at a certain xed temperature T2 in the gas bulb.14 The
molar volume Vm of each sample was evaluated from its pressure in the bulb at a reference
temperature of T1 D 7:1992 K, corrected for gas nonideality with the known value of the
second virial coefcient at that temperature.
Use these data and Eq. 2.2.2 on page 34 to evaluate T2 and the second virial coefcient of helium at temperature T2 . (You can assume the third and higher virial coefcients are negligible.)
2.4 Discuss the proposition that, to a certain degree of approximation, a living organism is a steadystate system.
2.5 The value of U for the formation of one mole of crystalline potassium iodide from its elements at 25 C and 1 bar is 327:9 kJ. Calculate m for this process. Comment on the
feasibility of measuring this mass change.
14 Ref.
[13].
C HAPTER 3
T HE F IRST L AW
In science, a law is a statement or mathematical relation that concisely describes reproducible experimental observations. Classical thermodynamics is built on a foundation of
three laws, none of which can be derived from principles that are any more fundamental.
This chapter discusses theoretical aspects of the rst law; gives examples of reversible and
irreversible processes and the heat and work that occur in them; and introduces the extensive
state function heat capacity.
3.1 HEAT, WORK, AND THE FIRST LAW

The box below gives two forms of the rst law of thermodynamics.
In a closed system:
dU D q C w
U D q C w
where U is the internal energy of the system, a state function;
q is heat; and
w is thermodynamic work.
The equation dU D q C w is the differential form of the rst law, and U D q C w is
the integrated form.
The heat and work appearing in the rst law are two different modes of energy transfer.
They can be dened in a general way as follows.
Heat refers to the transfer of energy across the boundary caused by a temperature gradient
at the boundary.
Work refers to the transfer of energy across the boundary caused by the displacement of a
macroscopic portion of the system on which the surroundings exert a force, or because
of other kinds of concerted, directed movement of entities (e.g., electrons) on which an
external force is exerted.
An innitesimal quantity of energy transferred as heat at a surface element of the boundary is written q, and a R
nite quantity is written q (Sec. 2.5). To obtain the total nite heat
for a process from q D q (Eq. 2.5.3), we must integrate over the total boundary surface
and the entire path of the process.
56
CHAPTER 3 THE FIRST LAW

3.1 H EAT, W ORK , AND THE F IRST L AW
57
R
An innitesimal quantity of work is w, and a nite quantity is w D w. To obtain
w for a process, we integrate all kinds of work over the entire path of the process.
Any of these quantities for heat and work is positive if the effect is to increase the
internal energy, and negative if the effect is to decrease it. Thus, positive heat is energy
entering the system, and negative heat is energy leaving the system. Positive work is work
done by the surroundings on the system, and negative work is work done by the system on
the surroundings.
The rst-law equation U D q Cw sets up a balance sheet for the energy of the system,
measured in the local frame, by equating its change during a process to the net quantity of
energy transferred by means of heat and work. Note that the equation applies only to a
closed system. If the system is open, energy can also be brought across the boundary by the
transport of matter.
An important part of the rst law is the idea that heat and work are quantitative energy
transfers. That is, when a certain quantity of energy enters the system in the form of heat,
the same quantity leaves the surroundings. When the surroundings perform work on the
system, the increase in the energy of the system is equal in magnitude to the decrease in
the energy of the surroundings. The principle of conservation of energy is obeyed: the
total energy (the sum of the energies of the system and surroundings) remains constant over
time.1
Heat transfer may occur by conduction, convection, or radiation.2 We can reduce conduction with good thermal insulation at the boundary, we can eliminate conduction and
convection with a vacuum gap, and we can minimize radiation with highly reective surfaces at both sides of the vacuum gap. The only way to completely prevent heat during a
process is to arrange conditions in the surroundings so there is no temperature gradient at
any part of the boundary. Under these conditions the process is adiabatic, and any energy
transfer in a closed system is then solely by means of work.
3.1.1 The concept of thermodynamic work

Appendix G gives a detailed analysis of energy and work based on the behavior of a collection of interacting particles moving according to the principles of classical mechanics. The
analysis shows how we should evaluate mechanical thermodynamic work. Suppose the displacement responsible for the work comes from linear motion of a portion of the boundary
in the Cx or x direction of the local frame. The differential and integrated forms of the
work are then given by3
Z x2
sur
sur
w D Fx dx
wD
Fx dx
(3.1.1)
x1
sur
Here Fx is the component in the Cx direction of the force exerted by the surroundings on
the system at the moving portion of the boundary, and dx is the innitesimal displacement
of the boundary in the local frame. If the displacement is in the same direction as the force,
w is positive; if the displacement is in the opposite direction, w is negative.
1 Strictly speaking, it is the sum of the energies of the system, the surroundings, and any potential energy shared
by both that is constant. The shared potential energy is usually negligible or essentially constant (Sec. G.5).
2 Some thermodynamicists treat radiation as a separate contribution to U , in addition to q and w.
3 These equations are Eq. G.6.11 with a change of notation.

58
sur
The kind of force represented by Fx is a short-range contact force. Appendix G shows
that the force exerted by a conservative time-independent external eld, such as a gravitasur
tional force, should not be included as part of Fx . This is because the work done by this
kind of force causes changes of potential and kinetic energies that are equal and opposite in
sign, with no net effect on the internal energy (see Sec. 3.6).
Newtons third law of action and reaction says that a force exerted by the surroundings
on the system is opposed by a force of equal magnitude exerted in the opposite direction by
the system on the surroundings. Thus the expressions in Eq. 3.1.1 can be replaced by
Z x2
sys
sys
Fx dx
(3.1.2)
w D Fx dx
wD
x1
sys
Fx
where
is the component in the Cx direction of the contact force exerted by the system
on the surroundings at the moving portion of the boundary.
An alternative to using the expressions in Eqs. 3.1.1 or 3.1.2 for evaluating w is to
imagine that the only effect of the work on the systems surroundings is a change in
the elevation of a weight in the surroundings. The weight must be one that is linked
sur
mechanically to the source of the force Fx . Then, provided the local frame is a stationary lab frame, the work is equal in magnitude and opposite in sign to the change
in the weights potential energy: w D mgh where m is the weights mass, g is the
acceleration of free fall, and h is the weights elevation in the lab frame. This interpretation of work can be helpful for seeing whether work occurs and for deciding on
its sign, but of course cannot be used to determine its value if the actual surroundings
include no such weight.
The procedure of evaluating w from the change of an external weights potential
energy requires that this change be the only mechanical effect of the process on the
surroundings, a condition that in practice is met only approximately. For example,
Joules paddle-wheel experiment using two weights linked to the system by strings and
pulleys, described latter in Sec. 3.7.2, required corrections for (1) the kinetic energy
gained by the weights as they sank, (2) friction in the pulley bearings, and (3) elasticity
of the strings (see Prob. 3.10 on page 100).
In the rst-law relation U D qCw, the quantities U , q, and w are all measured in an
arbitrary local frame. We can write an analogous relation for measurements in a stationary
lab frame:
Esys D qlab C wlab
(3.1.3)
Suppose the chosen local frame is not a lab frame, and we nd it more convenient to measure
the heat qlab and the work wlab in a lab frame than to measure q and w in the local frame.
What corrections are needed to nd q and w in this case?
If the Cartesian axes of the local frame do not rotate relative to the lab frame, then the
heat is the same in both frames: q D qlab .4
The expressions for wlab and wlab are the same as those for w and w in Eqs. 3.1.1
and 3.1.2, with dx interpreted as the displacement in the lab frame. There is an especially
simple relation between w and wlab when the local frame is a center-of-mass frameone
whose origin moves with the systems center of mass and whose axes do not rotate relative
to the lab frame:5
2
mgzcm
(3.1.4)
w D wlab 1 m vcm
2
4 Sec.
G.7.
5 Eq.
G.8.12 on page 503.

59
In this equation m is the mass of the system, vcm is the velocity of its center of mass in
the lab frame, g is the acceleration of free fall, and zcm is the height of the center of mass
above an arbitrary zero of elevation in the lab frame. In typical thermodynamic processes
the quantities vcm and zcm change to only a negligible extent, if at all, so that usually to a
good approximation w is equal to wlab .
When the local frame is a center-of-mass frame, we can combine the relations U D
q C w and q D qlab with Eqs. 3.1.3 and 3.1.4 to obtain
Esys D Ek C Ep C U
(3.1.5)
2
where Ek D 1 mvcm and Ep D mgzcm are the kinetic and potential energies of the system
2
as a whole in the lab frame.
A more general relation for w can be written for any local frame that has no rotational
motion and whose origin has negligible acceleration in the lab frame:6
w D wlab
mgzloc
(3.1.6)
Here zloc is the elevation in the lab frame of the origin of the local frame. zloc is usually
small or zero, so again w is approximately equal to wlab . The only kinds of processes
for which we may need to use Eq. 3.1.4 or 3.1.6 to calculate a non-negligible difference
between w and wlab are those in which massive parts of the system undergo substantial
changes in elevation in the lab frame.
Simple relations such as these between q and qlab , and between w and wlab , do not exist
if the local frame has rotational motion relative to a lab frame.
Hereafter in this book, thermodynamic work w will be called simply work. For all
practical purposes you can assume the local frames for most of the processes to be described
are stationary lab frames. The discussion above shows that the values of heat and work
measured in these frames are usually the same, or practically the same, as if they were
measured in a local frame moving with the systems center of mass. A notable exception
is the local frame needed to treat the thermodynamic properties of a liquid solution in a
centrifuge cell. In this case the local frame is xed in the spinning rotor of the centrifuge
and has rotational motion. This special case will be discussed in Sec. 9.8.2.
3.1.2 Work coefcients and work coordinates

If a process has only one kind of work, it can be expressed in the form
w D Y dX
or
wD
X2
Y dX
(3.1.7)
X1
where Y is a generalized force called a work coefcient and X is a generalized displacement called a work coordinate. The work coefcient and work coordinate are conjugate
variables. They are not necessarily actual forces and displacements. For example, we shall
see in Sec. 3.4.2 that reversible expansion work is given by w D p dV ; in this case, the
work coefcient is p and the work coordinate is V .
6 Eq.
G.7.3 on page 500.

60
weight
water
stop
Figure 3.1 System containing an electrical resistor and a paddle wheel immersed in
water. Dashed rectangle: system boundary. Cross-hatched area: removable thermal
insulation.
A process may have more than one kind of work, each with its own work coefcient
and conjugate work coordinate. In this case the work can be expressed as a sum over the
different kinds labeled by the index i:
w D
Yi dXi
or
wD
XZ
i
Xi;2
Xi;1
Yi dXi
(3.1.8)
3.1.3 Heat and work as path functions

Consider the apparatus shown in Fig. 3.1. The system consists of the water together with
the immersed parts: stirring paddles attached to a shaft (a paddle wheel) and an electrical
resistor attached to wires. In equilibrium states of this system, the temperature and pressure
are uniform and the paddle wheel is stationary. The system is open to the atmosphere, so
the pressure is constrained to be constant. We may describe the equilibrium states of this
system by a single independent variable, the temperature T . (The angular position of the
shaft is irrelevant to the state and is not a state function for equilibrium states of this system.)
Here are three experiments with different processes. Each process has the same initial
state dened by T1 D 300:0 K, and each has the same nal state.
Experiment 1: We surround the system with thermal insulation as shown in the gure and
release the external weight, which is linked mechanically to the paddle wheel. The
resulting paddle-wheel rotation causes turbulent churning of the water and an increase in
its temperature. Assume that after the weight hits the stop and the paddle wheel comes
to rest, the nal angular position of the paddle wheel is the same as at the beginning
of the experiment. We can calculate the work done on the system from the difference
between the potential energy lost by the weight and the kinetic energy gained before
it reaches the stop.7 We wait until the water comes to rest and the system comes to
thermal equilibrium, then measure the nal temperature. Assume the nal temperature
is T2 D 300:10 K, an increase of 0:10 kelvins.
7 This
calculation is an example of the procedure mentioned on page 58 in which the change in elevation of an
external weight is used to evaluate work.

61
Experiment 2: We start with the system in the same initial state as in experiment 1, and
again surround it with thermal insulation. This time, instead of releasing the weight we
close the switch to complete an electrical circuit with the resistor and allow the same
quantity of electrical work to be done on the system as the mechanical work done in
experiment 1. We discover the nal temperature (300:10 K) is exactly the same as at the
end of experiment 1. The process and path are different from those in experiment 1, but
the work and the initial and nal states are the same.
Experiment 3: We return the system to its initial state, remove the thermal insulation, and
place the system in thermal contact with a heat reservoir of temperature 300:10 K. Energy can now enter the system in the form of heat, and does so because of the temperature gradient at the boundary. By a substitution of heat for mechanical or electrical
work, the system changes to the same nal state as in experiments 1 and 2.
Although the paths in the three experiments are entirely different, the overall change of
state is the same. In fact, a person who observes only the initial and nal states and has
no knowledge of the intermediate states or the changes in the surroundings will be ignorant
of the path. Did the paddle wheel turn? Did an electric current pass through the resistor?
How much energy was transferred by work and how much by heat? The observer cannot
tell from the change of state, because heat and work are not state functions. The change of
state depends on the sum of heat and work. This sum is the change in the state function U,
as expressed by the integrated form of the rst law, U D q C w.
It follows from this discussion that neither heat nor work are quantities possessed by the
system. A system at a given instant does not have or contain a particular quantity of heat
or a particular quantity of work. Instead, heat and work depend on the path of a process
occurring over a period of time. They are path functions.
3.1.4 Heat and heating

In thermodynamics, the technical meaning of the word heat when used as a noun is energy
transferred across the boundary because of a temperature gradient at the boundary.
In everyday speech the noun heat is often used somewhat differently. Here are three
statements with similar meanings that could be misleading:
Heat is transferred from a laboratory hot plate to a beaker of water.
Heat ows from a warmer body to a cooler body.
To remove heat from a hot body, place it in cold water.
Statements such as these may give the false impression that heat is like a substance that
retains its identity as it moves from one body to another. Actually heat, like work, does not
exist as an entity once a process is completed. Nevertheless, the wording of statements such
as these is embedded in our everyday language, and no harm is done if we interpret them
correctly. This book, for conciseness, often refers to heat transfer and heat ow, instead
of using the technically more correct phrase energy transfer by means of heat.
Another common problem is failure to distinguish between thermodynamic heat and
the process of heating. To heat a system is to cause its temperature to increase. A heated
system is one that has become warmer. This process of heating does not necessarily involve
thermodynamic heat; it can also be carried out with work as illustrated by experiments 1
and 2 of the preceding section.

62
The notion of heat as an indestructible substance was the essence of the caloric theory. This theory was nally disproved by the cannon-boring experiments of Benjamin
Thompson (Count Rumford) in the late eighteenth century, and in a more quantitative
way by the measurement of the mechanical equivalent of heat by James Joule in the
1840s (see Sec. 3.7.2).
3.1.5 Heat capacity

The heat capacity of a closed system is dened as the ratio of an innitesimal quantity of
heat transferred across the boundary under specied conditions and the resulting innitesimal temperature change:
def
heat capacity D
q
dT
(3.1.9)
(closed system)
Since q is a path function, the value of the heat capacity depends on the specied conditions,
usually either constant volume or constant pressure. CV is the heat capacity at constant
volume and Cp is the heat capacity at constant pressure. These are extensive state functions
that will be discussed more fully in Sec. 5.6.
3.1.6 Thermal energy

It is sometimes useful to use the concept of thermal energy, the contribution to the systems
internal energy due to its temperature. Thermal energy has no exact denition, but generally
speaking it includes the kinetic energy of the random translational motions of microscopic
particles (atoms and molecules) as well as the vibrational and rotational energies (if any) of
these particles. When the temperature of a body increases, so do these energies.
If the transfer of energy across the boundary of a closed system is solely by means of
heat, the thermal energy change may be the same as the quantity of energy transferred. This
is not always the case. For example, when there is slow heat transfer into an equilibrium
system of ice and water at constant pressure, the ice partially melts without an increase in
either temperature or thermal energy. It is intermolecular potential energy that has increased
in this case, rather than thermal energy.
When an innitesimal quantity of energy q is transferred as heat into a body or a closed
single-phase system without work, volume change, or chemical reaction, we can assume all
of the transferred energy goes into increasing the thermal energy: dEthermal D q. Under
these conditions of zero work, the rst law becomes dU D q. The heat capacity at constant
volume for such a system, as dened by Eq. 3.1.9, is then CV D dU= dT , and the change
in thermal energy is given by
dEthermal D CV dT
(3.1.10)
(closed single-phase system,
no reaction)
Because thermal energy by denition depends only on T , Eq. 3.1.10 is an expression for its
innitesimal change that is valid for a closed single-phase system without reaction regardless of whether or not the volume is constant.

63
BIOGRAPHICAL SKETCH
Benjamin Thompson, Count of Rumford (17531814)
Benjamin Thompson, whose career was remarkably varied and colorful, collected experimental evidence of the falseness of the caloric
theorythe concept that heat is a material substance. He was a complex man: energetic, egotistical, domineering, and misanthropic.
Thompson was born into a farming family in Woburn, Massachusetts. He married a
wealthy widow and was admitted into fashionable society. At the time of the American Revolution he was accused of being a loyalist, and
at the age of 23 ed to England, abandoning
his wife and daughter. He was an Under Secretary of State in London, returned briey to
America as a British cavalry commander, and
then spent 11 years as a colonel in the Bavarian army. In Bavaria, to reward his success in
reorganizing the army and reforming the social welfare system, he was made a Count of
the Holy Roman Empire. He chose the name
Rumford after the original name of Concord,
New Hampshire, his wifes home town.
While in Bavaria, Count Rumford carried
out the cannon-boring experiments for which
he is best known. The caloric theory held that
heat is a kind of indestructible uid (caloric)
that is held in the spaces between the atoms
of a body. Frictional forces were supposed
to cause a rise in temperature by squeezing
caloric uid out of a body. Rumfords experiments involved boring into a horizontally-xed
cannon barrel with a blunt steel bit turned by
horse power. He reported the results in 1798:a
a Ref.
Being engaged, lately in superintending the boring of cannon, in the workshops of the military
arsenal at Munich, I was struck with the very
considerable degree of heat which a brass gun
acquires, in a short time, in being bored; and
with the still more intense heat (much greater
than that of boiling water, as I found by experiment,) of the metallic chips separated from it by
the borer. . .
By meditating on the results of all these experiments, we are naturally brought to that great
question which has so often been the subject of
speculation among philosophers; namely,
What is Heat?Is there any such thing as an
igneous uid?Is there any thing that can with
propriety be called caloric?. . .
And, in reasoning on this subject, we must
not forget to consider that most remarkable circumstance, that the source of the heat generated
by friction, in these experiments, appeared evidently to be inexhaustible.
It is hardly necessary to add, that any thing
which any insulated body, or system of bodies, can continue to furnish without limitation,
cannot possibly be a material substance: and it
appears to me to be extremely difcult, if not
quite impossible, to form any distinct idea of any
thing, capable of being excited and communicated, in the manner the heat was excited and
communicated in these experiments, except it be
MOTION.
Rumford thought of heat in a solid as harmonic vibrations similar to acoustic waves, not
as random motion or as a form of energy as
later developed by James Joule.
Rumford also made investigations into ballistics, nutrition, thermometry, light, and fabric
properties. He invented the Rumford replace
and the drip coffee percolator. After living in
London for fourteen years, he settled in Paris
in 1804. The following year, his rst wife having died in America, he married the widow of
the famous French chemist Antoine Lavoisier.
The marriage was stormy and they soon separated.
[145].

3.2 S PONTANEOUS , R EVERSIBLE ,
AND I RREVERSIBLE
P ROCESSES
64
3.2 SPONTANEOUS, REVERSIBLE, AND IRREVERSIBLE PROCESSES

A spontaneous process is a process that can actually occur in a nite time period under
the existing conditions. Any change over time in the state of a system that we observe
experimentally is a spontaneous process.
A spontaneous process is sometimes called a natural process, feasible process, possible
process, allowed process, or real process.
3.2.1 Reversible processes

A reversible process is an important concept in thermodynamics. This concept is needed
for the chain of reasoning that will allow us to dene entropy changes in the next chapter,
and will then lead on to the establishment of criteria for spontaneity and for various kinds
of equilibria.
Before reversible processes can be discussed, it is necessary to explain the meaning of
the reverse of a process. If a particular process takes the system from an initial state A
through a continuous sequence of intermediate states to a nal state B, then the reverse of
this process is a change over time from state B to state A with the same intermediate states
occurring in the reverse time sequence. To visualize the reverse of any process, imagine
making a movie lm of the events of the process. Each frame of the lm is a snapshot
picture of the state at one instant. If you run the lm backward through a movie projector,
you see the reverse process: the values of system properties such as p and V appear to
change in reverse chronological order, and each velocity changes sign.
The concept of a reversible process is not easy to describe or to grasp. Perhaps the
most confusing aspect is that a reversible process is not a process that ever actually occurs,
but is only approached as a hypothetical limit. During a reversible process the system
passes through a continuous sequence of equilibrium states. These states are ones that can
be approached, as closely as desired, by the states of a spontaneous process carried out
sufciently slowly. As the spontaneous process is carried out more and more slowly, it
approaches the reversible limit. Thus, a reversible process is an idealized process with a
sequence of equilibrium states that are those of a spontaneous process in the limit of innite
slowness.
This book has many equations expressing relations among heat, work, and state functions during various kinds of reversible processes. What is the use of an equation for a
process that can never actually occur? The point is that the equation can describe a spontaneous process to a high degree of accuracy, if the process is carried out slowly enough for
the intermediate states to depart only slightly from exact equilibrium states. For example,
for many important spontaneous processes we will assume the temperature and pressure are
uniform throughout the system, which strictly speaking is an approximation.
A reversible process of a closed system, as used in this book, has all of the following
characteristics:
It is an imaginary, idealized process in which the system passes through a continuous
sequence of equilibrium states. That is, the state at each instant is one that in an
isolated system would persist with no tendency to change over time. (This kind of
process is sometimes called a quasistatic process.)

AND I RREVERSIBLE
P ROCESSES
65
The sequence of equilibrium states can be approximated, as closely as desired, by

the intermediate states of a real spontaneous process carried out sufciently slowly.
The reverse sequence of equilibrium states can also be approximated, as closely as
desired, by the intermediate states of another spontaneous process carried out sufciently slowly. (This requirement prevents any spontaneous process with hysteresis,
such as plastic deformation or the stretching of a metal wire beyond its elastic limit,
from having a reversible limit.) During the approach to innite slowness, very slow
changes of the type described in item 3 on page 50 must be eliminated, i.e., prevented
with hypothetical constraints.
The spontaneous process of a closed system that has a reversible limit must be a
process with heat, or work, or boththe system cannot be an isolated one. It must be
possible for an experimenter to use conditions in the surroundings to control the rate
at which energy is transferred across the boundary by means of heat and work, and
thus to make the process go as slowly as desired.
If energy is transferred by work during a reversible process, the work coefcient Y
in the expression w D Y dX must be nite (nonzero) in equilibrium states of the
sys
system. For example, if the work is given by w D Fx dx (Eq. 3.1.2), the force
sys
Fx exerted by the system on the surroundings must be present when the system is
in an equilibrium state.
When a spontaneous process with a reversible limit is proceeding slowly enough for
its states to closely approximate those of the reversible process, a small change in
forces exerted on the system by the surroundings or in the external temperature at the
boundary can change the process to one whose states approximate the sequence of
states of the reverse process. In other words, it takes only a small change in external
conditions at the boundary, or in an external eld, to reverse the direction of the
process.
In the reversible limit, dissipative effects within the system such as internal friction
vanish. This is a consequence of the fact that only equilibrium states are present in
this limit.
When any innitesimal step of a reversible process takes place in reverse, the magnitudes of the heat q and work w are unchanged and their signs are reversed. Thus,
energy transferred by heat in one direction across the boundary during a reversible
process is fully recovered as energy transferred by heat in the opposite direction in
the reverse process. Energy transferred by work is recovered in the same way.
We must imagine the reversible process to proceed at a nite rate, otherwise there would
be no change of state over time. The precise rate of the change is not important. Imagine a
gas whose volume, temperature, and pressure are changing at some nite rate while the temperature and pressure magically stay perfectly uniform throughout the system. This is an
entirely imaginary process, because there is no temperature or pressure gradientno physical driving forcethat would make the change tend to occur in a particular direction.
This imaginary process is a reversible processone whose states of uniform temperature
and pressure are approached by the states of a real process as the real process takes place
more and more slowly.

AND I RREVERSIBLE
P ROCESSES
66
It is a good idea, whenever you see the word reversible, to think in the reversible
limit. Thus a reversible process is a process in the reversible limit, reversible work is work
in the reversible limit, and so on.
The reverse of a reversible process is itself a reversible process. As explained above,
the quantities of energy transferred across the boundary by heat and work during a
reversible process are fully recovered when the reversible process is followed by the
reverse process. This characteristic of a reversible process is sometimes described by
the statement that after a reversible change occurs, it is possible to restore both the
system and the local surroundings to their original states with no further changes anywhere. This statement, however, is misleading, because during the period in question
spontaneous changes inevitably occur outside the system. At the very least, the external operations needed to control the rates and directions of energy transfer across
the boundary by heat and work, carried out by a human investigator or by some sort
of automated mechanism, are highly spontaneous in nature and dissipate energy in the
surroundings.
3.2.2 Irreversible processes

An irreversible process is a spontaneous process whose reverse is neither spontaneous nor
reversible. That is, the reverse of an irreversible process can never actually occur and is
impossible. If a movie is made of a spontaneous process, and the time sequence of the events
depicted by the lm when it is run backward could not occur in reality, the spontaneous
process is irreversible.
A good example of a spontaneous, irreversible process is experiment 1 on page 60, in
which the sinking of an external weight immersed in water causes a paddle wheel to rotate
and the temperature of the water to increase. During this experiment mechanical energy is
dissipated into thermal energy. Suppose you insert a thermometer in the water and make a
movie lm of the experiment. Then when you run the lm backward in a projector, you will
see the paddle wheel rotating in the direction that raises the weight, and the water becoming
cooler according to the thermometer. Clearly, this reverse process is impossible in the real
physical world, and the process occurring during the experiment is irreversible. It is not
difcult to understand why it is irreversible when we consider events on the microscopic
level: it is extremely unlikely that the H2 O molecules next to the paddles would happen
to move simultaneously over a period of time in the concerted motion needed to raise the
weight.
3.2.3 Purely mechanical processes

There is a class of spontaneous processes that are also spontaneous in reverse; that is, spontaneous but not irreversible. These are purely mechanical processes involving the motion of
perfectly-elastic macroscopic bodies without friction, temperature gradients, viscous ow,
or other irreversible changes.
A simple example of a purely mechanical process and its reverse is shown in Fig. 3.2.
The ball can move spontaneously in either direction. Another example is a ywheel with
frictionless bearings rotating in a vacuum.

3.3 H EAT T RANSFER
67
(a)
(b)
Figure 3.2 Two purely mechanical processes that are the reverse of one another: a
thrown ball moving through a vacuum (a) to the right; (b) to the left.
A purely mechanical process is not reversible, for its states are not equilibrium states.
Such a process is an idealization, of a different kind than a reversible process, and is of little
interest in chemistry. Later chapters of this book will ignore such processes and will treat
the terms spontaneous and irreversible as synonyms.
3.3 HEAT TRANSFER

This section describes irreversible and reversible heat transfer. Keep in mind that when this
book refers to heat transfer or heat ow, energy is being transferred across the boundary on
account of a temperature gradient at the boundary. The transfer is always in the direction of
decreasing temperature.
We may sometimes wish to treat the temperature as if it is discontinuous at the boundary,
with different values on either side. The transfer of energy is then from the warmer side to
the cooler side. The temperature is not actually discontinuous; instead there is a thin zone
with a temperature gradient.
3.3.1 Heating and cooling

As an illustration of irreversible heat transfer, consider a system that is a solid metal sphere.
This spherical body is immersed in a well-stirred water bath whose temperature we can control. The bath and the metal sphere are initially equilibrated at temperature T1 D 300:0 K,
and we wish to raise the temperature of the sphere by one kelvin to a nal uniform temperature T2 D 301:0 K.
One way to do this is to rapidly increase the external bath temperature to 301:0 K and
keep it at that temperature. The temperature difference across the surface of the immersed
sphere then causes a spontaneous ow of heat through the system boundary into the sphere.
It takes time for all parts of the sphere to reach the higher temperature, so a temporary
internal temperature gradient is established. Thermal energy ows spontaneously from the
higher temperature at the boundary to the lower temperature in the interior. Eventually the
temperature in the sphere becomes uniform and equal to the bath temperature of 301:0 K.
Figure 3.3(a) on the next page graphically depicts temperatures within the sphere at different times during the heating process. Note the temperature gradient in the intermediate
states. Because of the gradient, these states cannot be characterized by a single value of
the temperature. If we were to suddenly isolate the system (the sphere) with a thermallyinsulated jacket while it is in one of these states, the state would change as the temperature

3.3 H EAT T RANSFER
68
101:2
301:2
101:2
301:2
10 s
4s
100:2
300:2
299:8
99:8
2s
3 4
r=cm
(a)
100:6
300:6
100:6
300:6
100:4
300:4
100:2
300:2
299:8
99:8
2s
4s
100:4
300:4
6s
100:2
300:2
100:0
300:0
initial state
100:8
300:8
T =K
6s
100:4
300:4
initial state
101:0
301:0
100:8
300:8
8s
100:6
300:6
100:0
300:0
101:2
301:2
nal state
101:0
301:0
T =K
100:8
300:8
T =K
101:0
301:0
nal state
100:0
300:0
initial state
3 4
r=cm
(b)
10 s
299:8
99:8
8s
nal state
3 4
r=cm
(c)
Figure 3.3 Temperature proles in a copper sphere of radius 5 cm immersed in a

water bath. The temperature at each of the times indicated is plotted as a function of
r, the distance from the center of the sphere. The temperature at distances greater than
5 cm, to the right of the vertical dashed line in each graph, is that of the external water
bath.
(a) Bath temperature raised at the rate of 0:10 K s 1 .
(b) Bath temperature raised innitely slowly.
(c) Bath temperature lowered at the rate of 0:10 K s 1 .
gradient rapidly disappears. Thus, the intermediate states of the spontaneous heating process are not equilibrium states, and the rapid heating process is not reversible.
To make the intermediate states more nearly uniform in temperature, with smaller temperature gradients, we can raise the temperature of the bath at a slower rate. The sequence
of states approached in the limit of innite slowness is indicated in Fig. 3.3(b). In each intermediate state of this limiting sequence, the temperature is perfectly uniform throughout
the sphere and is equal to the external bath temperature. That is, each state has thermal
equilibrium both internally and with respect to the surroundings. A single temperature now
sufces to dene the state at each instant. Each state is an equilibrium state because it would
have no tendency to change if we isolated the system with thermal insulation. This limiting
sequence of states is a reversible heating process.
The reverse of the reversible heating process is a reversible cooling process in which the
temperature is again uniform in each state. The sequence of states of this reverse process is
the limit of the spontaneous cooling process depicted in Fig. 3.3(c) as we decrease the bath
temperature more and more slowly.
In any real heating process occurring at a nite rate, the spheres temperature could not
be perfectly uniform in intermediate states. If we raise the bath temperature very slowly,
however, the temperature in all parts of the sphere will be very close to that of the bath. At
any point in this very slow heating process, it would then take only a small decrease in the
bath temperature to start a cooling process; that is, the practically-reversible heating process
would be reversed.
The important thing to note about the temperature gradients shown in Fig. 3.3(c) for the
spontaneous cooling process is that none resemble the gradients in Fig. 3.3(a) for the spontaneous heating processthe gradients are in opposite directions. It is physically impossible

3.4 D EFORMATION W ORK
69
for the sequence of states of either process to occur in the reverse chronological order, for
that would have thermal energy owing in the wrong direction along the temperature gradient. These considerations show that a spontaneous heat transfer is irreversible. Only in
the reversible limits do the heating and cooling processes have the same intermediate states;
these states have no temperature gradients.
Although the spontaneous heating and cooling processes are irreversible, the energy
transferred into the system during heating can be fully recovered as energy transferred back
to the surroundings during cooling, provided there is no irreversible work. This recoverability of irreversible heat is in distinct contrast to the behavior of irreversible work.
3.3.2 Spontaneous phase transitions

Consider a different kind of system, one consisting of the liquid and solid phases of a pure
substance. At a given pressure, this kind of system can be in transfer equilibrium at only
one temperature: for example, water and ice at 1:01 bar and 273:15 K. Suppose the system
is initially at this pressure and temperature. Heat transfer into the system will then cause
a phase transition from solid to liquid (Sec. 2.2.2). We can carry out the heat transfer by
placing the system in thermal contact with an external water bath at a higher temperature
than the equilibrium temperature, which will cause a temperature gradient in the system and
the melting of an amount of solid proportional to the quantity of energy transferred.
The closer the external temperature is to the equilibrium temperature, the smaller are
the temperature gradients and the closer are the states of the system to equilibrium states.
In the limit as the temperature difference approaches zero, the system passes through a
sequence of equilibrium states in which the temperature is uniform and constant, energy is
transferred into the system by heat, and the substance is transformed from solid to liquid.
This idealized process is an equilibrium phase transition, and it is a reversible process.
3.4 DEFORMATION WORK

This and the four following sections (Secs. 3.53.8) describe some spontaneous, irreversible
processes with various kinds of work and illustrate the concept of a reversible limit for the
processes that have such a limit.
The deformation of a system involves changes in the position, relative to the local frame,
of portions of the system boundary. At a small surface element of the boundary, the work
of deformation is given in general by the expression8
w D F sur cos ds
(3.4.1)
where F sur is the magnitude of the contact force exerted by the surroundings on the surface
element, ds is the innitesimal displacement of the surface element in the local frame, and
is the angle between the directions of the force and the displacement. If the displacement
is entirely parallel to the x axis, the expression becomes equivalent to that already given by
sur
Eq. 3.1.1 on page 57: w D Fx dx.
8 From
Eq. G.6.10 on page 499.

70
fric
gas
ext
Figure 3.4 Forces acting on the piston (cross hatched) in a cylinder-and-piston device
containing a gas (shaded). The direction of Ffric shown here is for expansion.
3.4.1 Gas in a cylinder-and-piston device

A useful kind of deformation for the development of thermodynamic theory is a change in
the volume of a gas or liquid.
As a model for the work involved in changing the volume of a gas, consider the arrangement shown in Fig. 3.4. A sample of gas is conned in a cylinder by a piston. The system
is the gas. The piston is not part of the system, but its position given by the variable x
determines the systems volume. Movement of the piston to the right, in the Cx direction,
expands the gas; movement to the left, in the x direction, compresses it.
We will nd it instructive to look in detail at the forces acting on the piston. There are
three kinds: the force Fgas exerted in the Cx direction by the gas; an external force Fext in
the x direction, which we can control in the surroundings; and a frictional force Ffric in
the direction opposite to the pistons velocity when the piston moves. (The friction occurs
at the seal between the edge of the piston and the cylinder wall.)
Let pb be the average pressure of the gas at the pistonthat is, at the moving portion
of the system boundary (the subscript b stands for boundary). Then the force exerted by
the gas on the piston is given by
Fgas D pb As
(3.4.2)
where As is the cross-section area of the cylinder.
The component in the Cx direction of the net force Fnet acting on the piston is given by
Fnet D Fgas
Fext C Ffric
(3.4.3)
Here, Fgas and Fext are taken as positive. Ffric is negative when the piston moves to the
right, positive when the piston moves to the left, and zero when the piston is stationary.
Suppose the system (the gas) initially is in an equilibrium state of uniform temperature
T1 and uniform pressure p1 , and the piston is stationary, so that Ffric is zero. According to
Newtons second law of motion, the net force Fnet is also zero, because otherwise the piston
would be accelerating. Then, from Eqs. 3.4.2 and 3.4.3, the external force needed to keep
the piston from moving is Fext D Fgas D p1 As .
To avoid complications of heat transfer, we conne our attention to a system with an
adiabatic boundary. By reducing Fext from its initial value of p1 As , we cause spontaneous
expansion to begin. As the piston moves to the right, the pressure pb exerted on the left face
of the piston becomes slightly less than the pressure on the stationary cylinder wall. The
molecular explanation of this pressure gradient is that gas molecules moving to the right

71
approach the moving piston at lower velocities relative to the piston than if the piston were
stationary, so that they collide with the piston less frequently and with a smaller transfer of
momentum in each collision. The temperature and pressure within the gas become nonuniform, and we cannot describe intermediate states of this spontaneous process with single
values of T and p. These intermediate states are not equilibrium states.
The more slowly we allow the adiabatic expansion to take place, the more nearly uniform are the temperature and pressure. In the limit of innite slowness, the gas passes
through a continuous sequence of equilibrium states of uniform temperature and pressure.
Let p2 be the pressure in the nal state of the innitely-slow expansion. In this state,
Fext is equal to p2 As . By increasing Fext from this value, we cause spontaneous compression to begin. The gas pressure pb at the piston now becomes slightly greater than at the
stationary cylinder wall, because the piston is moving to the left toward the molecules that
are moving to the right. A different pressure gradient develops than during expansion. The
states approached in the limit as we carry out the compression more and more slowly are
equilibrium states, occurring in the reverse sequence of the states for expansion at innite
slowness. The sequence of equilibrium states, taken in either direction, is a reversible process.
The magnitude of the effect of piston velocity on pb can be estimated with the help of
the kinetic-molecular theory of gases. This theory, of course, is not part of classical
macroscopic thermodynamics.
Consider an elastic collision of a gas molecule of mass m with the left side of the
piston shown in Fig. 3.4. Assume the piston moves at a constant velocity u (positive for
expansion and negative for compression), and let the x component of the molecules
velocity as it approaches the piston be vx (a positive value). After the collision, the x
component of the molecules velocity is vx C 2u, and the momentum that has been
transferred to the piston is 2m.vx u/. With the assumption that the x component
of the molecules velocity remains unchanged between consecutive collisions with the
piston, the time interval between two collisions is 2l=.vx u/ where l is the internal
cylinder length at the time of the earlier collision.
The average force exerted by one molecule on the piston is equal to the momentum
transferred to the piston by a collision, divided by the time interval between collisions.
The total pressure exerted by the gas on the piston is found by summing the average
force over all molecules and dividing by the piston area As D V = l:
pb D
nM 2
vx
V
2u hvx i C u2
(3.4.4)
(The angle brackets denote averages.) The pressure at the stationary wall of the cylin 2
der is p D .nM=V / vx . Accordingly, the pressure at the moving piston is given
by9
!
u2
hvx i
pb D p 1 2u 2 C 2
(3.4.5)
vx
vx
From kinetic-molecular theory, the averages are given by hvx i D .2RT = M /1=2 and
2
vx D RT =M .
Suppose the piston moves at the considerable speed of 10 m s 1 and the gas in the
cylinder is nitrogen (N2 ) at 300 K; then Eq. 3.4.5 predicts the pressure pb exerted on
9 An
equivalent formula is derived in Ref. [9], Eq. 7. A formula that yields similar values of pb appears in Ref.
[14].

72
the piston by the gas is only about 5% lower than the pressure p at the stationary wall
during expansion, and about 5% higher during compression. At low piston speeds the
percentage difference is proportional to the piston speed, so this example shows that
for reasonably-slow speeds the difference is quite small and for practical calculations
can usually be ignored.
3.4.2 Expansion work of a gas

We now consider the work involved in expansion and compression of the gas in the cylinderand-piston device of Fig. 3.4. This kind of deformation work, for both expansion and compression, is called expansion work or pressure-volume work.
Keep in mind that the system is just the gas. The only moving portion of the boundary of
this system is at the inner surface of the piston, and this motion is in the Cx or x direction.
The x component of the force exerted by the system on the surroundings at this portion of
sys
the boundary, Fx , is equal to Fgas . (The other forces shown in Fig. 3.4 are within the
surroundings.) Applying the differential form of Eq. 3.1.2, we have w D Fgas dx which,
with the substitution Fgas D pb As (from Eq. 3.4.2), becomes
w D
pb As dx
(3.4.6)
It will be convenient to change the work coordinate from x to V . The gas volume is
given by V D As x so that an innitesimal change dx changes the volume by dV D As dx.
The innitesimal quantity of work for an innitesimal volume change is then given by
w D
pb dV
(3.4.7)
(expansion work,
closed system)
and the nite work for a nite volume change, found by integrating from the initial to the
nal volume, is
Z V2
wD
pb dV
(3.4.8)
V1
(expansion work,
closed system)
During expansion (positive dV ), w is negative and the system does work on the surroundings. During compression (negative dV ), w is positive and the surroundings do work on
the system.
When carrying out dimensional analysis, you will nd it helpful to remember that the
product of two quantities with dimensions of pressure and volume (such as pb dV ) has
dimensions of energy, and that 1 Pa m3 is equal to 1 J.
The integral on the right side of Eq. 3.4.8 is a line integral (Sec. E.4 on page 481). In
order to evaluate the integral, one must be able to express the integrand pb as a function of
the integration variable V along the path of the expansion or compression process.
If the piston motion during expansion or compression is sufciently slow, we can with
little error assume that the gas has a uniform pressure p throughout, and that the work can

73
be calculated as if the process has reached its reversible limit. Under these conditions, Eq.
3.4.7 becomes
w D
p dV
(3.4.9)
(reversible expansion
work, closed system)
and Eq. 3.4.8 becomes

wD
V2
p dV
V1
(3.4.10)
(reversible expansion
work, closed system)
The appearance of the symbol p in these equations, instead of pb , implies that the equations
apply only to a process in which the system has at each instant a single uniform pressure.
As a general rule, an equation containing the symbol of an intensive property not assigned
to a specic phase is valid only if that property is uniform throughout the system, and this
will not be explicitly indicated as a condition of validity.
Some texts state that expansion work in a horizontal cylinder-and-piston device like
R
that shown in Fig. 3.4 should be calculated from w D
pext dV , where pext is a
pressure in the surroundings that exerts the external force Fext on the piston. However,
if the system is the gas the correct general expression is the one given by Eq. 3.4.8:
R
w D
pb dV . This is because it is the force Fgas D pb As that is exerted by the
system on the surroundings, whereas the force Fext D pext As is exerted by one part of
the surroundings on another part of the surroundings.
R
R
In other words, if the integrals Fgas dx and Fext dx have different values, it is the
R
rst of these two integrals that should be used to evaluate the work: w D
Fgas dx.
Both integrals are equal if the expansion or compression process is carried out reversibly. This is because in the limit of innite slowness the piston has neither friction
(Ffric D0) nor acceleration (Fnet D0), and therefore according to Eq. 3.4.3, Fgas and Fext
are equal throughout the process. Another situation in which the two integrals are
equal is when the piston is frictionless and is stationary in the initial and nal states,
R
R
because then both Ffric and Fnet dx are zero. (The integral Fnet dx can be shown to
be equal to the change in the kinetic energy of the piston, by a derivation similar to
that leading to Eq. G.1.5 on page 489.) In the general irreversible case, however, the
R
R
integrals Fgas dx and Fext dx are not equal.10
3.4.3 Expansion work of an isotropic phase

Expansion work does not require a cylinder-and-piston device. Suppose the system is an
isotropic uid or solid phase, and various portions of its boundary undergo displacements
in different directions. Figure 3.5 on the next page shows an example of compression in
a system of arbitrary shape. The deformation is considered to be carried out slowly, so
that the pressure p of the phase remains uniform. Consider the surface element of the
boundary, with area As; , indicated in the gure by a short thick curve. Because the phase
10 For an informative discussion
of this topic see Ref. [6]; also comments in Refs. [29], [7], [94], [8], and [120];
also Ref. [92].


74
d
sur
s
sys
(a)
(b)
Figure 3.5 Deformation of an isotropic phase (shaded) conned by a wall.

(a) Equal and opposite forces exerted by the surroundings and system at surface element (thick curve) of the system boundary.
(b) Change from initial volume (dotted curve) to a smaller volume.
sys
is isotropic, the force F

D pAs; exerted by the system pressure on the surroundings
is perpendicular to this surface element; that is, there is no shearing force. The force F sur
sys
exerted by the surroundings on the system is equal in magnitude to F and is directed
in the opposite direction. The volume change for an innitesimal displacement ds that
reduces the volume is dV D As; ds , so that the work at this surface element (from Eq.
3.4.1 with D0) is w D p dV .
By summing the work over the entire boundary, we nd the total reversible expansion work is given by the same expression as for a gas in a piston-and-cylinder device:
w D p dV . This expression can be used for deformation caused by reversible displacements of a conning wall, or for a volume change caused by slow temperature changes
at constant pressure. It is valid if the system is an isotropic uid phase in which other
phases are immersed, provided the uid phase contacts all parts of the system boundary.
The expression is not necessarily valid for an anisotropic uid or solid, because the angle
appearing in Eq. 3.4.1 might not be zero.
3.4.4 Generalities
The expression w D p dV for reversible expansion work of an isotropic phase is the
product of a work coefcient, p, and the innitesimal change of a work coordinate, V .
In the reversible limit, in which all states along the path of the process are equilibrium
states, the system has two independent variables, e.g., p and V or T and V . The number of
independent variables is one greater than the number of work coordinates. This will turn out
to be a general rule: The number of independent variables needed to describe equilibrium
states of a closed system is one greater than the number of independent work coordinates
for reversible work.
Another way to state the rule is as follows: The number of independent variables is one
greater than the number of different kinds of reversible work, where each kind i is given by
an expression of the form wi D Yi dXi .

3.5 A PPLICATIONS OF E XPANSION W ORK
75
3.5 APPLICATIONS OF EXPANSION WORK

This book uses expansion work as a general term that includes the work of both expansion
and compression of an isotropic phase.
3.5.1 The internal energy of an ideal gas

The model of an ideal gas is used in many places in the development of thermodynamics.
For examples to follow, the following denition is needed: An ideal gas is a gas
1. whose equation of state is the ideal gas equation, pV D nRT ; and
2. whose internal energy in a closed system is a function only of temperature.11
On the molecular level, a gas with negligible intermolecular interactions fullls both of
these requirements. Kinetic-molecular theory predicts that a gas containing noninteracting
molecules obeys the ideal gas equation. If intermolecular forces (the only forces that depend
on intermolecular distance) are negligible, the internal energy is simply the sum of the
energies of the individual molecules. These energies are independent of volume but depend
on temperature.
The behavior of any real gas approaches ideal-gas behavior when the gas is expanded
isothermally. As the molar volume Vm becomes large and p becomes small, the average
distance between molecules becomes large, and intermolecular forces become negligible.
3.5.2 Reversible isothermal expansion of an ideal gas

During reversible expansion or compression, the temperature and pressure remain uniform.
If we substitute p D nRT =V from the ideal gas equation into Eq. 3.4.10 and treat n and T
as constants, we obtain
Z V2
V
dV
D nRT ln 2
(3.5.1)
w D nRT
V1
V1 V
(reversible isothermal
expansion work, ideal gas)
In these expressions for w the amount n appears as a constant for the process, so it is not
necessary to state as a condition of validity that the system is closed.
3.5.3 Reversible adiabatic expansion of an ideal gas

This section derives temperature-volume and pressure-volume relations when a xed amount
of an ideal gas is expanded or compressed without heat.
First we need a relation between internal energy and temperature. Since the value of
the internal energy of a xed amount of an ideal gas depends only on its temperature (Sec.
3.5.1), an innitesimal change dT will cause a change dU that depends only on T and dT :
dU D f .T / dT
(3.5.2)
11 A
gas with this second property is sometimes called a perfect gas. In Sec. 7.2 it will be shown that if a gas
has the rst property, it must also have the second.

76
where f .T / D dU= dT is a function of T . For a constant-volume process of a closed

system without work, we know from the rst law that dU is equal to q and that q= dT
is equal to CV , the heat capacity at constant volume (Sec. 3.1.5). Thus we can identify the
function f .T / as the heat capacity at constant volume:
dU D CV dT
(3.5.3)
(ideal gas, closed system)
The relation given by Eq. 3.5.3 is valid for any process of a closed system of an ideal gas
of uniform temperature, even if the volume is not constant or if the process is adiabatic,
because it is a general relation between state functions.
By comparing Eq. 3.5.3 with Eq. 3.1.10 on page 62, dEthermal D CV dT , we see that
we can equate a change in the internal energy of an ideal gas to its thermal energy
change: dU D dEthermal . This equality does not apply to a nonideal gas, because the
internal energy of a nonideal gas has contributions from both thermal energy and the
potential energy of intermolecular interactions.
In a reversible adiabatic expansion with expansion work only, the heat is zero and the
rst law becomes
dU D w D p dV
(3.5.4)
We equate these two expressions for dU to obtain
CV dT D
p dV
(3.5.5)
and substitute p D nRT =V from the ideal gas equation:

CV dT D
nRT
dV
V
(3.5.6)
It is convenient to make the approximation that over a small temperature range, CV is

constant. When we divide both sides of the preceding equation by T in order to separate
the variables T and V , and then integrate between the initial and nal states, we obtain
Z
CV
T2
T1
CV ln
V2
dT
D
T
nR
T2
D
T1
nR ln
V1
dV
V
V2
V1
(3.5.7)
(3.5.8)
We can rearrange this result into the form

T
ln 2 D
T1
nR=CV
nR V2
V1
ln
D ln
CV
V1
V2
(3.5.9)
and take the exponential of both sides:

T2
D
T1
V1
V2
nR=CV
(3.5.10)

77
The nal temperature is then given as a function of the initial and nal volumes by
T2 D T1
V1
V2
nR=CV
(3.5.11)
(reversible adiabatic
expansion, ideal gas)
This relation shows that the temperature decreases during an adiabatic expansion and increases during an adiabatic compression, as expected from expansion work on the internal
energy.
To nd the work during the adiabatic volume change, we can use the relation
Z
Z T2
w D U D dU D CV
dT
T1
D CV .T2
T1 /
(3.5.12)
To express the nal pressure as a function of the initial and nal volumes, we make the
substitutions T1 D p1 V1 =nR and T2 D p2 V2 =nR in Eq. 3.5.11 and obtain
p2 V 2
p V
D 1 1
nR
nR
V1
V2
nR=CV
(3.5.13)
Solving this equation for p2 , we obtain nally

p2 D p1
V1
V2
1CnR=CV
(3.5.14)
The solid curve in Fig. 3.6 on the next page shows how the pressure of an ideal gas
varies with volume during a reversible adiabatic expansion or compression. This curve is
an adiabat. The dashed curves in the gure are isotherms showing how pressure changes
with volume at constant temperature according to the equation of state p D nRT =V . In the
direction of increasing V (expansion), the adiabat crosses isotherms of progressively lower
temperatures. This cooling effect, of course, is due to the loss of energy by the gas as it
does work on the surroundings without a compensating ow of heat into the system.
3.5.4 Indicator diagrams

An indicator diagram (or pressurevolume diagram) is usually a plot of p as a function
of V . The curve describes the path of an expansion or compression process of a uid that
R
is essentially uniform. The area under the curve has the same value as the integral p dV ,
R
which is the negative of the reversible expansion work given by w D p dV . For example,
the area under the solid curve of Fig. 3.6 between any two points on the curve is equal to
w for reversible adiabatic expansion or compression. If the direction of the process is to
the right along the path (expansion), the area is positive and the work is negative; but if the
direction is to the left (compression), the area is taken as negative and the work is positive.

78
p / bar
4
800
K
600
K
400 K
200 K
0:10
0:20
V = dm3
0:30
0:40
Figure 3.6 An adiabat (solid curve) and four isotherms (dashed curves) for an ideal
gas (n D 0:0120 mol, CV;m D 1:5R).
More generally, an indicator diagram can be a plot of a work coefcient or its negative
as a function of the work coordinate. For example, it could be a plot of the pressure pb at
R
a moving boundary as a function of V . The area under this curve is equal to pb dV , the
negative of expansion work in general (Eq. 3.4.8).
Historically, an indicator diagram was a diagram drawn by an indicator, an instrument invented by James Watt in the late 1700s to monitor the performance of steam
engines. The steam engine indicator was a simple pressure gauge: a piston moving
in a small secondary cylinder, with the steam pressure of the main cylinder on one
side of the piston and a compressed spring opposing this pressure on the other side.
A pointer attached to the small piston indicated the steam pressure. In later versions,
the pointer was replaced with a pencil moving along a paper-covered roll, which in
turn was mechanically linked to the piston of the main cylinder (see Fig. 3.7 on the
next page). During each cycle of the engine, the pencil moved back and forth along
the length of the roll and the roll rotated in a reciprocating motion, causing the pencil
to trace a closed curve whose area was proportional to the net work performed by one
cycle of the engine.
3.5.5 Spontaneous adiabatic expansion or compression

Section 3.4.1 explained that during a rapid spontaneous expansion of the gas in the cylinder
shown in Fig. 3.4, the pressure pb exerted by the gas at the moving piston R less than the
is
pressure at the stationary wall. Consequently the work given by w D
pb dV is less
negative for a spontaneous adiabatic expansion than for a reversible adiabatic expansion
with the same initial state and the same volume change.
During a rapid spontaneous compression, pb is greater than the pressure at the stationary
wall. The work is positive and greater for a spontaneous adiabatic compression than for a

79
Figure 3.7 Indicator with paper-covered roll at left and pressure gauge at right. a
a Ref.
[110], page 104.
gas
vacuum
gas
gas
Figure 3.8 Free expansion into a vacuum.
reversible adiabatic compression with the same initial state and the same volume change.
These observations are summarized by the statement that, for an adiabatic expansion
or compression with a given change of the work coordinate, starting at a given initial equilibrium state, the work is algebraically smallest (least positive or most negative) in the
reversible limit. That is, in the reversible limit the surroundings do the least possible work
on the system and the system does the maximum possible work on the surroundings. This
behavior will turn out to be true of any adiabatic process of a closed system.
3.5.6 Free expansion of a gas into a vacuum

When we open the stopcock of the apparatus shown in Fig. 3.8, the gas expands from the
vessel at the left into the evacuated vessel at the right. This process is called free expansion.
The system is the gas. The surroundings exert a contact force on the system only at the vessel
walls, where there is no displacement. Thus, there is no work in free expansion: w D 0.
If the free expansion is carried out adiabatically in a thermally-insulated apparatus, there
is neither heat nor work and therefore no change in the internal energy: U D 0. If the gas
is ideal, its internal energy depends only on temperature; thus the adiabatic free expansion
of an ideal gas causes no temperature change.

3.6 W ORK IN A G RAVITATIONAL F IELD
80
str
buoy
buoy
fric
fric
uid
uid
(a)
(b)
Figure 3.9 Spherical body (shaded) in a gravitational eld. The arrows indicate the
directions and magnitudes of contact and gravitational forces exerted on the body.
(a) The body falls freely through a uid in a vessel.
(b) The body is lowered on a string through the uid.
3.6 WORK IN A GRAVITATIONAL FIELD

Figure 3.9 depicts a spherical body, such as a glass marble, immersed in a liquid or gas in
the presence of an external gravitational eld. The vessel is stationary on a lab bench, and
the local reference frame for work is a stationary lab frame. The variable z is the bodys
elevation above the bottom of the vessel. All displacements are parallel to the vertical z axis.
sur
sur
From Eq. 3.1.1, the work is given by w D Fz dz where Fz is the upward component
of the net contact force exerted by the surroundings on the system at the moving portion of
the boundary. There is also a downward gravitational force on the body, but as explained in
sur
Sec. 3.1.1, this force does not contribute to Fz .
Consider rst the simple process in Fig. 3.9(a) in which the body falls freely through
the uid. This process is clearly spontaneous. Here are two choices for the denition of the
system:
The system is the combination of the spherical body and the uid. The system boundary is where the uid contacts the air and the vessel walls. Because there is no displacement of this boundary, no work is being done on or by the system: w D 0.
(We ignore expansion work caused by the small temperature increase.) If the process
is adiabatic, the rst law tells us the systems internal energy remains constant: as
the body loses gravitational potential energy, the system gains an equal quantity of
kinetic and thermal energy.
The system is the body; the uid is in the surroundings. The upward components
of the forces exerted on the body are (1) a gravitational force mg, where m is the
bodys mass and g is the acceleration of free fall; (2) a buoyant force12 Fbuoy D V 0 g,
where is the uid density and V 0 is the volume of the body; and (3) a frictional drag
force Ffric of opposite sign from the velocity v D dz= dt. As mentioned above, the
sur
gravitational force is not included in Fz . Therefore the gravitational work is given
12 The
buoyant force is a consequence of the pressure gradient that exists in the uid in a gravitational eld (see
Sec. 8.1.4). We ignore this gradient when we treat the uid as a uniform phase.

3.6 W ORK IN A G RAVITATIONAL F IELD
81
by
sur
w D Fz dz D Fbuoy C Ffric dz
(3.6.1)
and is negative because dz is negative: the body as it falls does work on the uid.
The positive quantity Fbuoy dz is the work of moving displaced uid upward, and
jFfric dzj is the energy dissipated by friction to thermal energy in the surroundings.
This process has no reversible limit, because the rate of energy transfer cannot be
controlled from the surroundings and cannot be made to approach zero.
Next, consider the arrangement in Fig. 3.9(b) in which the body is suspended by a thin
string. The string is in the surroundings and provides a means for the surroundings to exert
an upward contact force on the system. As before, there are two appropriate choices for the
system:
The system includes both the body and the uid. The moving part of the boundary
is at the point where the string is attached to the body. The force exerted here by the
sur
string is an upward force Fstr , and the gravitational work is given by w D Fz dz D
Fstr dz. According to Newtons second law, the net force on the body equals the
product of its mass and acceleration: . mg C Fbuoy C Ffric C Fstr / D m dv= dt .
Solving this equation for Fstr , we obtain
Fstr D mg
Fbuoy
Ffric C m dv= dt
(3.6.2)
We can therefore express the work in the form

w D Fstr dz D mg
Fbuoy
Ffric C m dv= dt dz
(3.6.3)
This work can be positive or negative, depending on whether the body is being pulled
up or lowered by the string. The quantity .m dv= dt/ dz is an innitesimal change of
R
the bodys kinetic energy Ek ,13 so that the integral .m dv= dt/ dz is equal to Ek .
The nite quantity of work in a process that starts and ends in equilibrium states, so
that Ek is zero, is therefore
Z
Z
w D w D mg Fbuoy z
Ffric dz
(3.6.4)
The work has a reversible limit, because the string allows the velocity v to be controlled from the surroundings. As v approaches zero from either direction, Ffric approaches zero and the work approaches the reversible limit w D .mg Fbuoy /z.
(If the uid is a gas whose density is much smaller than the density of the body,
Fbuoy can be neglected in comparison with mg, and the reversible work can be written w D mgz.) Ffric and dz have opposite signs, so w for a given change of the
work coordinate z is least positive or most negative in the reversible limit.
sur
The system is the body only. In this case, Fz is equal to Fbuoy C Ffric C Fstr which
by substitution from Eq. 3.6.2 is .mg C m dv= dt/. The work is then given by
w D F sur dz D .mg C m dv= dt/ dz

13 To
prove this, we write m.dv= dt / dz D m.dz= dt/ dv D mv dv D d
2
1
2 mv
(3.6.5)
D dEk .

3.7 S HAFT W ORK
82
water
System A
System B
Figure 3.10 Two systems with shaft work. The dashed rectangles indicate the system
boundaries. System A has an internal weight, cord, and pulley wheel in air; system B
has a stirrer immersed in water.
For a process that begins and ends in equilibrium states, Ek is zero and the nite
work is w D mgz, unaffected by the velocity v during the process. The expressions
for innitesimal and nite work in the reversible limit are
w D mg dz
and w D mgz
(3.6.6)
(reversible gravitational
work of a body)
When we compare Eqs. 3.6.3 and 3.6.5, we see that the work when the system is the
body is greater by the quantity Fbuoy C Ffric dz than the work when the system is the
combination of body and uid, just as in the case of the freely-falling body. The difference
in the quantity of work is due to the different choices of the system boundary where contact
forces are exerted by the surroundings.
3.7 SHAFT WORK

Shaft work refers to energy transferred across the boundary by a rotating shaft.
The two systems shown in Fig. 3.10 will be used to illustrate two different kinds of shaft
work. Both systems have a straight cylindrical shaft passing through the system boundary.
Let # be the angle of rotation of the shaft in radians, and ! be the angular velocity d#= dt .
Tangential forces imposed on one of these shafts can create a torque sys at the lower end
within the system, and a torque sur at the upper end in the surroundings.14 The sign convention for a torque is that a positive value corresponds to tangential forces in the rotational
direction in which the shaft turns as # increases.
The condition for ! to be zero, or nite and constant (i.e., no angular acceleration), is
that the algebraic sum of the imposed torques be zero: sys D sur . Under these conditions
of constant !, the torque couple creates rotational shear forces in the circular cross section
of the shaft where it passes through the boundary. These shear forces are described by
an internal torque with the same magnitude as sys and sur . Applying the condition for
zero angular acceleration to just the part of the shaft within the system, we nd that sys is
14 A
torque is a moment of tangential force with dimensions of force times distance.

3.7 S HAFT W ORK
83
(a)
(b)
Figure 3.11 Shaft work w for a xed magnitude of shaft rotation # as a function
of the angular velocity ! D d#= dt. The open circles indicate work in the limit of
innite slowness. (a) System A of Fig. 3.10. (b) System B of Fig. 3.10.
balanced by the internal torque b exerted on this part of the shaft by the part of the shaft in
the surroundings: b D sys . The shaft work is then given by the formula
wD
#2
#1
b d# D
#2
#1
sys d#
(3.7.1)
(shaft work, constant !)
In system A of Fig. 3.10, when ! is zero the torque sys is due to the tension in the cord
from the weight of mass m, and is nite: sys D mgr where r is the radius of the shaft
at the point where the cord is attached. When ! is nite and constant, frictional forces at
the shaft and pulley bearings make sys more negative than mgr if ! is positive, and less
negative than mgr if ! is negative. Figure 3.11(a) shows how the shaft work given by
Eq. 3.7.1 depends on the angular velocity for a xed value of j#2 #1 j. The variation of w
with ! is due to the frictional forces. System A has nite, reversible shaft work in the limit
of innite slowness (!!0) given by w D mgr#. The shaft work is least positive or most
negative in the reversible limit.
In contrast to system A, the shaft work in system B has no reversible limit, as discussed
in the next section.
3.7.1 Stirring work

The shaft work done when a shaft turns a stirrer or paddle to agitate a liquid, as in system
B of Fig. 3.10 on the preceding page, is called stirring work.
In system B, when the angular velocity ! is zero and the water in which the stirrer is
immersed is at rest, the torques sys and b are both zero. When ! is nite and constant, the
water is stirred in a turbulent manner and there is a frictional drag force at the stirrer blades,
as well as frictional forces at the shaft bearings. These forces make the value of sys have
the opposite sign from !, increasing in magnitude the greater is the magnitude of !. As a
result, the stirring work for a xed value of j#2 #1 j depends on ! in the way shown in
Fig. 3.11(b). The work is positive for nite values of ! of either sign, and approaches zero
in the limit of innite slowness.

3.7 S HAFT W ORK
84
Stirring work is an example of dissipative work. Dissipative work is work that is

positive for both positive and negative changes of the work coordinate, that vanishes as the
rate of change of the work coordinate approaches zero, and that therefore cannot be carried
out reversibly. Energy transferred into the system by dissipative work is not recovered as
work done on the surroundings when the work coordinate is reversed. In the case of stirring
work, if the shaft rotates in one direction work is done on the system; if the rotation direction
is reversed, still more work is done on the system. The energy transferred to the system by
stirring work is converted by friction within the system into the random motion of thermal
energy: the energy is completely dissipated.
Because energy transferred to the system by dissipative work is converted to thermal
energy, we could replace this work with an equal quantity of positive heat and produce
the same overall change. The replacement of stirring work with heat was illustrated by
experiment 3 on page 61.
The shaft rotation angle #, which is the work coordinate for stirring work, is a property
of the system but is not a state function, as we can see by the fact that the state of the system
can be exactly the same for # D 0 and # D 2 . The work coordinate and work coefcient
of work with a reversible limit are always state functions, whereas the work coordinate of
any kind of dissipative work is not a state function.
In system B of Fig. 3.10, there is in addition to the stirring work the possibility of
expansion work given by w D p dV . When we take both kinds of work into account,
we must treat this system as having two work coordinates: # for stirring work and V for
expansion work. Only the expansion work can be carried out reversibly. The number of
independent variables in equilibrium states of this system is two, which we could choose as
T and V . Thus, the number of independent variables of the equilibrium states is one greater
than the number of work coordinates for reversible work, in agreement with the general rule
given on page 74.
3.7.2 The Joule paddle wheel

A good example of the quantitative measurement of stirring work is the set of experiments
conducted by James Joule in the 1840s to determine the mechanical equivalent of heat.
In effect, he determined the quantity of dissipative stirring work that could replace the heat
needed for the same temperature increase.
Joules apparatus contained the paddle wheel shown in Fig. 3.12 on page 86. It consisted
of eight sets of metal paddle arms attached to a shaft in a water-lled copper vessel. When
the shaft rotated, the arms moved through openings in four sets of stationary metal vanes
xed inside the vessel, and churned the water. The vanes prevented the water from simply
moving around in a circle. The result was turbulent motion (shearing or viscous ow) in the
water and an increase in the temperature of the entire assembly.
The complete apparatus is depicted in Fig. 3.13 on page 87. In use, two lead weights
sank and caused the paddle wheel to rotate. Joule evaluated the stirring work done on the
system (the vessel, its contents, and the lid) from the change of the vertical position h of the
weights. To a rst approximation, this work is the negative of the change of the weights
potential energy: w D mgh where m is the combined mass of the two weights. Joule
made corrections for the kinetic energy gained by the weights, the friction in the connecting

3.7 S HAFT W ORK
85
BIOGRAPHICAL SKETCH
James Prescott Joule (18181889)

termination of the mechanical equivalent of
heat using a paddle wheel to agitate water (Sec.
3.7.2 and Prob. 3.10). He reported his results in
1845, and published a more rened measurement in 1850.a
In a note dated 1885 in his Collected Papers, Joule wrote:
James Joule drove the nal nails into the cofn of the caloric theory by his experimental
demonstrations of the mechanical equivalent
of heat.
Joule (pronounced like jewel) was born
in Salford, near Manchester, England. His father was a prosperous brewery owner; after his
death, James and one of his brothers carried on
the business until it was sold in 1854.
Joule was a sickly child with a minor spinal
weakness. He was tutored at home, and at the
age of 16 was a pupil of the atomic theory advocate John Dalton.
As an adult, Joule was a political conservative and a member of the Church of England. He dressed plainly, was of a somewhat
nervous disposition, and was a poor speaker.
He was shy and reserved unless with friends,
had a strong sense of humor, and loved nature.
Joule never attended a university or had a
university appointment, but as an amateur
scientist and inventor he published over 100
papers (some of them jointly with collaborators) and received many honors. He invented arc welding and a mercury displacement pump. He carried out investigations on
electrical heating and, in collaboration with
William Thomson, on the cooling accompanying the expansion of a gas through a porous
plug (the JouleThomson experiment). The
joule, of course, is now the SI derived unit of
energy.
Joules best-known experiment was the dea Ref.
[84].
b Ref.
[17].
c Ref.
It was in the year 1843 that I read a paper On

the Caloric Effects of Magneto-Electricity and
the Mechanical Value of Heat to the Chemical
Section of the British Association assembled at
Cork. With the exception of some eminent men
. . . the subject did not excite much general attention; so that when I brought it forward again
at the meeting in 1847, the chairman suggested
that, as the business of the section pressed, I
should not read my paper, but conne myself to a
short verbal description of my experiments. This
I endeavoured to do, and discussion not being
invited, the communication would have passed
without comment if a young man had not risen
in the section, and by his intelligent observations
created a lively interest in the new theory. The
young man was William Thomson, who had two
years previously passed the University of Cambridge with the highest honour, and is now probably the foremost scientic authority of the age.
The William Thomson mentioned in Joules

note later became Lord Kelvin. Thomson described introducing himself to Joule after the
1847 meeting, which was in Oxford, as a result of which the two became collaborators and
life-long friends. Thomson wrote:b
Joules paper at the Oxford meeting made a great
sensation. Faraday was there and was much
struck with it, but did not enter fully into the new
views. It was many years after that before any of
the scientic chiefs began to give their adhesion.
According to a biographer:c
His modesty was always notable. I believe, he
told his brother on 14 Sept. 1887, I have done
two or three little things, but nothing to make a
fuss about. During the later years of his life he
received many distinctions both English and foreign.
[66].

3.8 E LECTRICAL W ORK
86
MIRKO JUNGE / COMMONS.WIKIMEDIA.ORG
opening for
thermometer
(a)
water
(b)
Figure 3.12 Joule paddle wheel.

(a) Joules original paddle wheel on exhibit at the Science Museum, London.
(b) Cross-section elevation of paddle wheel and water in copper vessel. Dark shading:
rotating shaft and paddle arms; light shading: stationary vanes.
strings and pulley bearings, the elasticity of the strings, and the heat gain from the air
surrounding the system.
A typical experiment performed by Joule is described in Prob. 3.10 on page 100. His
results for the mechanical equivalent of heat, based on 40 such experiments at average
temperatures in the range 13 C16 C and expressed as the work needed to increase the
temperature of one gram of water by one kelvin, was 4:165 J. This value is close to the
modern value of 4:1855 J for the 15 C calorie, the energy needed to raise the temperature
of one gram of water from 14:5 C to 15:5 C.15
3.8 ELECTRICAL WORK

The electric potential at a point in space is dened as the work needed to reversibly move
an innitesimal test charge from a position innitely far from other charges to the point of
interest, divided by the value of the test charge. The electrical potential energy of a charge
at this point is the product of and the charge.
3.8.1 Electrical work in a circuit

Electric current is usually conducted in an electrical circuit. Consider a thermodynamic
system that is part of a circuit: in a given time period electrons enter the system through
one wire, and an equal number of electrons leave through a second wire. To simplify the
description, the wires are called the right conductor and the left conductor.
The electric potentials experienced by a electron in the right and left conductors are
R and L , respectively. The electron charge is e, where e is the elementary charge (the
15 The
thermochemical calorie (cal), often used as an energy unit in the older literature, is dened as 4:184 J.
Thus 1 kcal D 4:184 kJ.

87
G
D
C
I
B
K
A
cm
Figure 3.13 Joules apparatus for measuring the mechanical equivalent of heat (redrawn from a gure in Ref. [84]).
Key: Apaddle wheel and vessel (see Fig. 3.12); Bwood thermal insulator; Cpin
used to engage paddle wheel shaft to roller; Droller; Ecrank used to wind up the
weights; F, Gstrings; H, Ipulley wheels; J, Kweights (round lead disks, viewed
here edge-on).
charge of a proton). Thus the electrical potential energy of an electron is

R e in the right
conductor and
e in the left conductor. The difference in the energies of an electron in
L
the two conductors is the difference in the electrical potential energies.
The sum of charges of a small number of electrons can be treated as an innitesimal
negative charge. During a period of time in which an innitesimal charge Qsys enters
the system at the right conductor and an equal charge leaves at the left conductor, the
contribution of the electric current to the internal energy change is the energy difference
. R Qsys
L Qsys / D . R
L / Qsys . (The notation is Qsys instead of dQsys , because Qsys is a path function.) This internal energy change is called electrical work. Thus
the general formula for an innitesimal quantity of electrical work when the system is part
of an electrical circuit is
wel D Qsys
(3.8.1)
(electrical work in a circuit)
where is the electric potential difference dened by
def
(3.8.2)
Note that in the expression . R Qsys

L Qsys / for the energy difference, the term
Qsys does not represent the energy transferred across the boundary at the right
R
conductor, and
L Qsys is not the energy transferred at the left conductor. These
energies cannot be measured individually, because they include not just the electrical
potential energy but also the energy of the rest mass of the electrons. The reason we can
write Eq. 3.8.1 for the electrical work in a circuit is that equal numbers of electrons

88
C
L
e
R
liquid
Figure 3.14 System containing an electrical resistor immersed in a liquid. The

dashed rectangle indicates the system boundary.
enter and leave the system, so that the net energy transferred across the boundary
depends only on the difference of the electric potential energies. Because the number
of electrons in the system remains constant, we can treat the system as if it were closed.
Why should we regard the transfer of energy across the boundary by an electric
current as a kind of work? One justication for doing so is that the energy transfer is
consistent with the interpretation of work discussed on page 58: the only effect on the
surroundings could be a change in the elevation of an external weight. For example, the
weight when it sinks could drive a generator in the surroundings that does electrical
work on the system, and electrical work done by the system could run an external
motor that raises the weight.
What is the meaning of Qsys in the differential Qsys ? We dene Qsys as the total
cumulative charge, positive or negative, that has entered the system at the right conductor
def R
since the beginning of the process: Qsys D Qsys . Qsys is a path function for charge,
and Qsys is its inexact differential, analogous to q and q for heat. Because the charge of
an electron is negative, Qsys is negative when electrons enter at the right conductor and
positive when they leave there.
The electric current I is the rate at which charges pass a point in the circuit: I D
Qsys = dt , where t is time. We take I as negative if electrons enter at the right conductor
and positive if electrons leave there. This relation provides an alternative form of Eq. 3.8.1:
wel D I dt
(3.8.3)
(electrical work in a circuit)
Equations 3.8.1 and 3.8.3 are general equations for electrical work in a system that is
part of a circuit. The electric potential difference which appears in these equations may
have its source in the surroundings, as for electrical heating with a resistor discussed in the
next section, or in the system, as in the case of a galvanic cell (Sec. 3.8.3).
3.8.2 Electrical heating

Figure 3.14 shows an electrical resistor immersed in a liquid. We will begin by dening
the system to include both the resistor and the liquid. An external voltage source provides
an electric potential difference across the wires. When electrons ow in the circuit, the
resistor becomes warmer due to the ohmic resistance of the resistor. This phenomenon is

89
el
Figure 3.15 Work of electrical heating with a xed magnitude of Qsys as a function
of the electric current I D Qsys = dt. The open circle indicates the limit of innite
slowness.
variously called electrical heating, Joule heating, ohmic heating, or resistive heating. The
heating is caused by inelastic collisions of the moving electrons with the stationary atoms of
the resistor, a type of friction. If the resistor becomes warmer than the surrounding liquid,
there will be a transfer of thermal energy from the resistor to the liquid.
The electrical work performed on this system is given by the expressions wel D
Qsys and wel D I dt (Eqs. 3.8.1 and 3.8.3). The portion of the electrical circuit
inside the system has an electric resistance given by Rel D =I (Ohms law). Making the
substitution D IRel in the work expressions gives two new expressions for electrical
work in this system:
wel D IRel Qsys
(3.8.4)
wel D I 2 Rel dt
(3.8.5)
The integrated form of Eq. 3.8.4 when I and Rel are constant is wel D IRel Qsys .
When the source of the electric potential difference is in the surroundings, as it is here, I
and Qsys have the same sign, so wel is positive for nite current and zero when there is
no current. Figure 3.15 shows graphically how the work of electrical heating is positive
for both positive and negative changes of the work coordinate Qsys and vanishes as I , the
rate of change of the work coordinate, approaches zero. These are the characteristics of
irreversible dissipative work (page 84). Note the resemblance of Fig. 3.15 to Fig. 3.11(b)
on page 83 for dissipative stirring workthey are the same graphs with different labels.
Suppose we redene the system to be only the liquid. In this case, electric current passes
through the resistor but not through the system boundary. There is no electrical work, and
we must classify energy transfer between the resistor and the liquid as heat.
3.8.3 Electrical work with a galvanic cell

A galvanic cell is an electrochemical system that, when isolated, exhibits an electric potential difference between the two terminals at the system boundary. The potential difference
has its source at the interfaces between phases within the cell.
Consider the combination of galvanic cell and electrical resistor in Fig. 3.16 on the next
page, and let the system be the cell. When an electric current passes through the cell in
either direction, a cell reaction takes place in one direction or the other.

cell
90
(a)
cell
(b)
Figure 3.16 Galvanic cell and external electrical resistor.

(a) Open circuit with isolated cell in an equilibrium state.
(b) Closed circuit.
In a manner similar to the labeling of the conductors of a circuit, the cell terminals are
called the right terminal and the left terminal. The cell potential Ecell is the electric potential
difference between the terminals, and is dened by
def
Ecell D
(3.8.6)
When the cell is in an isolated zero-current equilibrium state, as in Fig. 3.16(a), the cell
potential is the equilibrium cell potential Ecell, eq . When the cell is part of an electrical
circuit with an electric current passing through the cell, as in Fig. 3.16(b), Ecell is different
from Ecell, eq on account of the internal resistance Rcell of the cell:
Ecell D Ecell, eq C IRcell
(3.8.7)
The sign of the current I is negative when electrons enter the cell at the right terminal, and
positive when electrons leave there.
In the circuit shown in Fig. 3.16(b), the cell does electrical work on the resistor in the
surroundings. The energy for this work comes from the cell reaction. The formula for the
electrical work is given by Eq. 3.8.1 with replaced by Ecell :
wel D Ecell Qsys
(3.8.8)
The gure shows Ecell as positive and Qsys as negative, so for this arrangement wel is
negative.
When current passes through the cell, the work done is irreversible because the internal
resistance causes energy dissipation. We can make this work approach a nite reversible
limit by replacing the external resistor shown in Fig. 3.16(b) with an adjustable voltage
source that we can use to control the cell potential Ecell and the current I . According to
Eq. 3.8.7, Ecell is greater than Ecell, eq when I is positive, and is less than Ecell, eq when I is
negative. This behavior is shown graphically in Fig. 3.17 on the next page. In the limit as
the electric current approaches zero from either direction and the external adjustable voltage
approaches Ecell, eq , the electrical work approaches a reversible limit given by
wel, rev D Ecell, eq Qsys
(3.8.9)
Note that the electrical work is the least positive or most negative in the reversible limit.

3.9 I RREVERSIBLE W ORK AND I NTERNAL F RICTION
91
el
Figure 3.17 Electrical work of a galvanic cell for a xed magnitude of Qsys as a
function of the electric current I D Qsys = dt . Open circles: reversible limits.
Thus, unlike the dissipative work of stirring and electrical heating, electrical work with
a galvanic cell has a nonzero reversible limit, as reected by the difference in the appearance of Fig. 3.17 compared to Figs. 3.11 and 3.15. During irreversible electrical work of
a galvanic cell, there is only partial dissipation of energy within the cell: the energy transferred across the boundary by the work can be partially recovered by returning the work
coordinate Qsys to its initial value.
On page 84 the observation was made that the work coordinate of work with a reversible limit is always a state function. Electrical work with a galvanic cell does
not contradict this statement, because the work coordinate Qsys is proportional to the
extent of the cell reaction, a state function.
The thermodynamics of galvanic cells will be treated in detail in Chap. 14.
3.9 IRREVERSIBLE WORK AND INTERNAL FRICTION

Consider an irreversible adiabatic process of a closed system in which a work coordinate
X changes at a nite rate along the path, starting and ending with equilibrium states. For
a given initial state and a given change X, the work is found to be less positive or more
negative the more slowly is the rate of change of X . The work is least positive or most
negative in the reversible limitthat is, the least work needs to be done on the system, or
the most work can be done by the system on the surroundings. This minimal work principle
has already been illustrated in the sections of this chapter describing expansion work, work
in a gravitational eld, and electrical work with a galvanic cell.
Let the work during an irreversible adiabatic process be wirr , and the reversible adiabatic
work for the same initial state and the same value of X be wrev . wirr is algebraically greater
than wrev , and we can treat the difference wirr wrev as excess work wex that is positive for
an irreversible process and zero for a reversible process.
Conceptually, we can attribute the excess work of an irreversible adiabatic process to
internal friction that dissipates other forms of energy into thermal energy within the system.
Internal friction occurs only during a process with work that is irreversible . Internal friction
is not involved when, for example, a temperature gradient causes heat to ow spontaneously
across the system boundary, or an irreversible chemical reaction takes place spontaneously

gas
92
P
R
gas
Figure 3.18 Cylinder and piston with internal sliding friction. The dashed rectangle
indicates the system boundary. Ppiston; Rinternal rod attached to piston; B
bushing xed inside cylinder. A xed amount of an ideal gas lls the remaining space
inside the cylinder.
in a homogeneous phase. Nor is internal friction necessarily involved when positive work
increases the thermal energy: during an innitely slow adiabatic compression of a gas,
the temperature and thermal energy increase but internal friction is absentthe process is
reversible.
During a process with irreversible work, energy dissipation can be either partial or complete. Dissipative work, such as the stirring work and electrical heating described in previous sections, is irreversible work with complete energy dissipation. The nal equilibrium
state of an adiabatic process with dissipative work can also be reached by a path with positive heat and no work. This is a special case of the minimal work principle.
As a model for work with partial energy dissipation, consider the gas-lled cylinderand-piston device depicted in Fig. 3.18. This device has an obvious source of internal
friction in the form of a rod sliding through a bushing. The system consists of the contents
of the cylinder to the left of the piston, including the gas, the rod, and the bushing; the piston
and cylinder wall are in the surroundings.
From Eq. 3.1.2, the energy transferred as work across the boundary of this system is
Z x2
wD
F sys dx
(3.9.1)
x1
F sys
where x is the piston position and

is the component in the direction of increasing x of
the force exerted by the system on the surroundings at the moving boundary.
For convenience, we let V be the volume of the gas rather than that of the entire system.
The relation between changes of V and x is dV D As dx where As is the cross-section area
of the cylinder. With V replacing x as the work coordinate, Eq. 3.9.1 becomes
Z V2
wD
.F sys =As / dV
(3.9.2)
V1
Equation 3.9.2 shows that a plot of F sys =As as a function of V is an indicator diagram (Sec.
3.5.4), and that the work is equal to the negative of the area under the curve of this plot.
We can write the force F sys as the sum of two contributions:16
int
F sys D pAs C Ffric
(3.9.3)
16 This
equation assumes that the gas pressure is uniform, and that a term for the acceleration of the rod is
negligible.

93
int
Here p is the gas pressure, and Ffric is the force on the rod due to internal friction with sign
opposite to that of the piston velocity dx=dt. Substitution of this expression for F sys in Eq.
3.9.2 gives
Z V2
Z V2
int
p dV
.Ffric =As / dV
(3.9.4)
wD
V1
V1
The rst term on the right is the work of expanding or compressing the gas. The second term
R int
is the frictional work: wfric D
.Ffric =As / dV . wfric is positive or zero, and represents the
energy dissipated within the system by internal sliding friction.
Consider the situation when the piston moves very slowly in one direction or the other.
int
In the limit of innite slowness17RFfric and wfric vanish, and the process is reversible with
expansion work given by w D
p dV .
The situation is different when the piston moves at an appreciable nite rate. The frictional work wfric is then positive. As a result, the irreversible work of expansion is less
negative than the reversible work for the same volume increase, and the irreversible work of
compression is more positive than the reversible work for the same volume decrease. These
effects of piston velocity on the work are consistent with the minimal work principle.
The piston velocity, besides affecting the frictional force on the rod, has an effect
on the force exerted by the gas on the piston as described in Sec. 3.4.1. At large
nite velocities, this latter effect tends to further decrease F sys during expansion and
increase it during compression, and so is an additional contribution to internal friction.
If turbulent ow is present within the system, that would also be a contribution.
Figure 3.19 on the next page shows indicator diagrams for adiabatic expansion and
compression with internal friction. The solid curves are for irreversible processes at nite
rates, and the dashed curves are for reversible processes with the same initial states as the
irreversible processes. The areas under the curves conrm that the work for expansion is less
negative along the irreversible path than along the reversible path, and that for compression
the work is more positive along the irreversible path than along the reversible path.
Because of these differences in work, the nal states of the irreversible processes have
greater internal energies and higher temperatures and pressures than the nal states of the
reversible processes with the same volume change, as can be seen from the positions on
the indicator diagrams of the points for the nal equilibrium states. The overall change
of state during the irreversible expansion or compression is the same for a path in which
the reversible adiabatic volume change is followed by positive heat at constant volume.
Since U must be the same for both paths, the heat has the same value as the excess work
wex D wirr wrev . The excess work and frictional work are not equal, because the thermal
energy released by frictional work increases the gas pressure, making wex less than wfric for
expansion and greater than wfric for compression. There seems to be no general method by
which the energy dissipated by internal friction can be evaluated, and it would be even more
difcult for an irreversible process with both work and heat.
Figure 3.20 on the next page shows the effect of the rate of change of the volume on the
adiabatic work for a xed magnitude of the volume change. Note that the work of expansion
and the work of compression have opposite signs, and that it is only in the reversible limit
17 The
contact between the rod and bushing is assumed to be lubricated to allow the piston to move at velocities
innitesimally close to zero.

6
.F sys =As / / bar
6
.F sys =As / / bar
94
0:10
0:20
V = dm3
0:30
0:10
(a)
0:20
V = dm3
0:30
(b)
Figure 3.19 Indicator diagrams for the system of Fig. 3.18.

Solid curves: F sys =As for irreversible adiabatic volume changes at nite rates in the
directions indicated by the arrows.
Dashed curves: F sys =As D p along a reversible adiabat.
Open circles: initial and nal equilibrium states.
(a) Adiabatic expansion.
(b) Adiabatic compression.
compression
d d
expansion
Figure 3.20 Adiabatic expansion work with internal friction for a xed magnitude of
V , as a function of the average rate of volume change. The open circles indicate the
reversible limits.
that they have the same magnitude. The gure resembles Fig. 3.17 for electrical work of
a galvanic cell with the horizontal axis reversed, and is typical of irreversible work with
partial energy dissipation.

3.10 R EVERSIBLE AND I RREVERSIBLE P ROCESSES : G ENERALITIES
95
3.10 REVERSIBLE AND IRREVERSIBLE PROCESSES:

GENERALITIES
This section summarizes some general characteristics of processes in closed systems. Some
of these statements will be needed to develop aspects of the second law in Chap. 4.
Innitesimal quantities of work during a process are calculated from an expression of
P
the form w D i Yi dXi , where Xi is the work coordinate of kind of work i and
Yi is the conjugate work coefcient.
The work coefcients and work coordinates of reversible work are state functions.
Energy transferred across the boundary by work in a reversible process is fully recovered as work of the opposite sign in the reverse reversible process. It follows from
the rst law that heat is also fully recovered in the reverse process.
When work occurs irreversibly at a nite rate, there is partial or complete dissipation
of energy. The dissipation consists of an increase of thermal energy within the system
due to frictional effects.
Dissipative work is positive work with complete energy dissipation. In this type of
work, the work coefcient vanishes in the limit of innite slowness, and the work
coordinate is not a state function. Examples are stirring work (Sec. 3.7.1) and the
work of electrical heating (Sec. 3.8.2).
If a process is carried out adiabatically, the work for a given initial equilibrium state
and a given change in the work coordinate is least positive or most negative in the
reversible limit. Since dissipative work is positive and vanishes in the limit of innite
slowness, dissipative work can be thought of as a special case of the same principle.
The way in which work depends on the rate of change of the work coordinate is
shown graphically for dissipative work by Figs. 3.11(b) and 3.15, and for work with
partial energy dissipation by Figs. 3.11(a), 3.17, and 3.20.
The number of independent variables needed to describe equilibrium states of a
closed system is one greater than the number of independent work coordinates for
reversible work.18 Thus, we could choose the independent variables to be each of the
work coordinates and in addition either the temperature or the internal energy.19 The
number of independent variables needed to describe a nonequilibrium state is greater
(often much greater) than this.
Table 3.1 on the next page lists general formulas for various kinds of work, including
those that were described in detail in Secs. 3.43.8.
18 If
the system has internal adiabatic partitions that allow different phases to have different temperatures in
equilibrium states, then the number of independent variables is equal to the number of work coordinates plus
the number of independent temperatures.
19 There may be exceptions to this statement in special cases. For example, along the triple line of a pure
substance the values of V and T , or of V and U , are not sufcient to determine the amounts in each of the three
possible phases.

3.10 R EVERSIBLE AND I RREVERSIBLE P ROCESSES : G ENERALITIES
96
Table 3.1 Some kinds of work

Kind
Formula
Denitions
Linear mechanical work
sur
w D Fx dx
Shaft work
w D
Expansion work
w D
Surface work of a at surface

Stretching or compression
of a rod or spring
w D dAs
w D F dl
Gravitational work
w D mg dh
Electrical work in a circuit
w D Qsys
Electric polarization
w D E dp
Magnetization
w D B dm
sur
Fx D x-component of force exerted
by surroundings
dx D displacement in x direction
b D internal torque at boundary
# D angle of rotation
pb D average pressure at moving
boundary
D surface tension, As D surface area
F D stress (positive for tension,
negative for compression)
l D length
m D mass, h D height
g D acceleration of free fall
D electric potential difference
D R
L
Qsys D charge entering system at right
E D electric eld strength
p D electric dipole moment of system
B D magnetic ux density
m D magnetic dipole moment of system
d#
pb dV

P ROBLEMS
97
PROBLEMS
An underlined problem number or problem-part letter indicates that the numerical answer appears
in Appendix I.
3.1 Assume you have a metal spring that obeys Hookes law: F D c.l l0 /, where F is the force
exerted on the spring of length l, l0 is the length of the unstressed spring, and c is the spring
constant. Find an expression for the work done on the spring when you reversibly compress it
from length l0 to a shorter length l 0 .
air
water
Figure 3.21
3.2 The apparatus shown in Fig. 3.21 consists of xed amounts of water and air and an incompressible solid glass sphere (a marble), all enclosed in a rigid vessel resting on a lab bench.
Assume the marble has an adiabatic outer layer so that its temperature cannot change, and that
the walls of the vessel are also adiabatic.
Initially the marble is suspended above the water. When released, it falls through the air into
the water and comes to rest at the bottom of the vessel, causing the water and air (but not
the marble) to become slightly warmer. The process is complete when the system returns to
an equilibrium state. The system energy change during this process depends on the frame of
reference and on how the system is dened. Esys is the energy change in a lab frame, and
U is the energy change in a specied local frame.
For each of the following denitions of the system, give the sign (positive, negative, or zero)
of both Esys and U , and state your reasoning. Take the local frame for each system to be a
center-of-mass frame.
(a) The system is the marble.
(b) The system is the combination of water and air.
(c) The system is the combination of water, air, and marble.
3.3 Figure 3.22 on the next page shows the initial state of an apparatus consisting of an ideal gas
in a bulb, a stopcock, a porous plug, and a cylinder containing a frictionless piston. The walls
are diathermal, and the surroundings are at a constant temperature of 300:0 K and a constant
pressure of 1:00 bar.
When the stopcock is opened, the gas diffuses slowly through the porous plug, and the piston
moves slowly to the right. The process ends when the pressures are equalized and the piston
stops moving. The system is the gas. Assume that during the process the temperature throughout the system differs only innitesimally from 300:0 K and the pressure on both sides of the
piston differs only innitesimally from 1:00 bar.

P ROBLEMS
98
bar
m
K
porous
plug
piston
gas
ext
ext
K
bar
Figure 3.22
(a) Which of these terms correctly describes the process: isothermal, isobaric, isochoric,
reversible, irreversible?
(b) Calculate q and w.
3.4 Consider a horizontal cylinder-and-piston device similar to the one shown in Fig. 3.4 on
page 70. The piston has mass m. The cylinder wall is diathermal and is in thermal contact
with a heat reservoir of temperature Text . The system is an amount n of an ideal gas conned
in the cylinder by the piston.
The initial state of the system is an equilibrium state described by p1 and T D Text . There
is a constant external pressure pext , equal to twice p1 , that supplies a constant external force
on the piston. When the piston is released, it begins to move to the left to compress the gas.
Make the idealized assumptions that (1) the piston moves with negligible friction; and (2) the
gas remains practically uniform (because the piston is massive and its motion is slow) and has
a practically constant temperature T D Text (because temperature equilibration is rapid).
(a) Describe the resulting process.
(b) Describe how you could calculate w and q during the period needed for the piston velocity
to become zero again.
(c) Calculate w and q during this period for 0:500 mol gas at 300 K.
3.5 This problem is designed to test the assertion on page 59 that for typical thermodynamic processes in which the elevation of the center of mass changes, it is usually a good approximation
to set w equal to wlab . The cylinder shown in Fig. 3.23 on the next page has a vertical orientation, so the elevation of the center of mass of the gas conned by the piston changes as the piston slides up or down. The system is the gas. Assume the gas is nitrogen (M D 28:0 g mol 1 )
at 300 K, and initially the vertical length l of the gas column is one meter. Treat the nitrogen as an ideal gas, use a center-of-mass local frame, and take the center of mass to be at the
midpoint of the gas column. Find the difference between the values of w and wlab , expressed
as a percentage of w, when the gas is expanded reversibly and isothermally to twice its initial
volume.
3.6 Figure 3.24 on the next page shows an ideal gas conned by a frictionless piston in a vertical
cylinder. The system is the gas, and the boundary is adiabatic. The downward force on the
piston can be varied by changing the weight on top of it.
(a) Show that when the system is in an equilibrium state, the gas pressure is given by p D
mgh=V where m is the combined mass of the piston and weight, g is the acceleration of
free fall, and h is the elevation of the piston shown in the gure.

P ROBLEMS
99
gas
Figure 3.23
vacuum
weight
ideal
gas
Figure 3.24
(b) Initially the combined mass of the piston and weight is m1 , the piston is at height h1 , and
the system is in an equilibrium state with conditions p1 and V1 . The initial temperature
is T1 D p1 V1 =nR. Suppose that an additional weight is suddenly placed on the piston,
so that m increases from m1 to m2 , causing the piston to sink and the gas to be compressed adiabatically and spontaneously. Pressure gradients in the gas, a form of friction,
eventually cause the piston to come to rest at a nal position h2 . Find the nal volume,
V2 , as a function of p1 , p2 , V1 , and CV . (Assume that the heat capacity of the gas, CV ,
is independent of temperature.) Hint: The potential energy of the surroundings changes
by m2 gh; since the kinetic energy of the piston and weights is zero at the beginning
and end of the process, and the boundary is adiabatic, the internal energy of the gas must
change by m2 gh D m2 gV =As D p2 V .
(c) It might seem that by making the weight placed on the piston sufciently large, V2 could
be made as close to zero as desired. Actually, however, this is not the case. Find expressions for V2 and T2 in the limit as m2 approaches innity, and evaluate V2 =V1 in this
limit if the heat capacity is CV D .3=2/nR (the value for an ideal monatomic gas at room
temperature).
3.7 The solid curve in Fig. 3.6 on page 78 shows the path of a reversible adiabatic expansion or
compression of a xed amount of an ideal gas. Information about the gas is given in the gure

P ROBLEMS
100
caption. For compression along this path, starting at V D 0:3000 dm3 and T D 300:0 K and
ending at V D 0:1000 dm3 , nd the nal temperature to 0:1 K and the work.
bar
m
K
gas
vacuum
ext
Figure 3.25
3.8 Figure 3.25 shows the initial state of an apparatus containing an ideal gas. When the stopcock
is opened, gas passes into the evacuated vessel. The system is the gas. Find q, w, and U
under the following conditions.
(a) The vessels have adiabatic walls.
(b) The vessels have diathermal walls in thermal contact with a water bath maintained at
300: K, and the nal temperature in both vessels is T D 300: K.
3.9 Consider a reversible process in which the shaft of system A in Fig. 3.10 makes one revolution
in the direction of increasing #. Show that the gravitational work of the weight is the same as
the shaft work given by w D mgr#.
Table 3.2 Data for Problem 3.10. The values are from Joules 1850 paper a
and have been converted to SI units.
Properties of the paddle wheel apparatus:
combined mass of the two lead weights . . . . . . . . . . . . . . . .
mass of water in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
mass of water with same heat capacity
as paddle wheel, vessel, and lidb . . . . . . . . . . . . . . . . . . .
Measurements during the experiment:
number of times weights were wound up and released . . .
change of elevation of weights during each descent . . . . .
nal downward velocity of weights during descent . . . . . .
initial temperature in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . .
nal temperature in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . .
mean air temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
a Ref.
26:3182 kg
6:04118 kg
0:27478 kg
20
1:5898 m
0:0615 m s
288:829 K
289:148 K
289:228 K
[84], p. 67, experiment 5.

from the masses and specic heat capacities of the materials.
b Calculated
3.10 This problem guides you through a calculation of the mechanical equivalent of heat using data
from one of James Joules experiments with a paddle wheel apparatus (see Sec. 3.7.2). The
experimental data are collected in Table 3.2.
In each of his experiments, Joule allowed the weights of the apparatus to sink to the oor
twenty times from a height of about 1:6 m, using a crank to raise the weights before each
descent (see Fig. 3.13 on page 87). The paddle wheel was engaged to the weights through the

P ROBLEMS
101
roller and strings only while the weights descended. Each descent took about 26 seconds, and
the entire experiment lasted 35 minutes. Joule measured the water temperature with a sensitive
mercury-in-glass thermometer at both the start and nish of the experiment.
For the purposes of the calculations, dene the system to be the combination of the vessel, its
contents (including the paddle wheel and water), and its lid. All energies are measured in a
lab frame. Ignore the small quantity of expansion work occurring in the experiment. It helps
conceptually to think of the cellar room in which Joule set up his apparatus as being effectively
isolated from the rest of the universe; then the only surroundings you need to consider for the
calculations are the part of the room outside the system.
(a) Calculate the change of the gravitational potential energy Ep of the lead weights during
each of the descents. For the acceleration of free fall at Manchester, England (where
Joule carried out the experiment) use the value g D 9:813 m s 2 . This energy change
represents a decrease in the energy of the surroundings, and would be equal in magnitude
and opposite in sign to the stirring work done on the system if there were no other changes
in the surroundings.
(b) Calculate the kinetic energy Ek of the descending weights just before they reached the
oor. This represents an increase in the energy of the surroundings. (This energy was
dissipated into thermal energy in the surroundings when the weights came to rest on the
oor.)
(c) Joule found that during each descent of the weights, friction in the strings and pulleys
decreased the quantity of work performed on the system by 2:87 J. This quantity represents an increase in the thermal energy of the surroundings. Joule also considered the
slight stretching of the strings while the weights were suspended from them: when the
weights came to rest on the oor, the tension was relieved and the potential energy of the
strings changed by 1:15 J. Find the total change in the energy of the surroundings during
the entire experiment from all the effects described to this point. Keep in mind that the
weights descended 20 times during the experiment.
(d) Data in Table 3.2 show that change of the temperature of the system during the experiment
was
T D .289:148 288:829/ K D C0:319 K
The paddle wheel vessel had no thermal insulation, and the air temperature was slighter
warmer, so during the experiment there was a transfer of some heat into the system. From
a correction procedure described by Joule, the temperature change that would have occurred if the vessel had been insulated is estimated to be C0:317 K.
Use this information together with your results from part (c) to evaluate the work needed
to increase the temperature of one gram of water by one kelvin. This is the mechanical
equivalent of heat at the average temperature of the system during the experiment. (As
mentioned on p. 86, Joule obtained the value 4:165 J based on all 40 of his experiments.)
3.11 Refer to the apparatus depicted in Fig. 3.1 on page 60. Suppose the mass of the external weight
is m D 1:50 kg, the resistance of the electrical resistor is Rel D 5:50 k , and the acceleration
of free fall is g D 9:81 m s 2 . For how long a period of time will the external cell need to
operate, providing an electric potential difference j j D 1:30 V, to cause the same change in
the state of the system as the change when the weight sinks 20:0 cm without electrical work?
Assume both processes occur adiabatically.
C HAPTER 4
T HE S ECOND L AW
The second law of thermodynamics concerns entropy and the spontaneity of processes. This
chapter discusses theoretical aspects and practical applications.
We have seen that the rst law allows us to set up a balance sheet for energy changes
during a process, but says nothing about why some processes occur spontaneously and
others are impossible. The laws of physics explain some spontaneous changes. For instance,
unbalanced forces on a body cause acceleration, and a temperature gradient at a diathermal
boundary causes heat transfer. But how can we predict whether a phase change, a transfer of
solute from one solution phase to another, or a chemical reaction will occur spontaneously
under the existing conditions? The second law provides the principle we need to answer
these and other questionsa general criterion for spontaneity in a closed system.
4.1 TYPES OF PROCESSES

Any conceivable process is either spontaneous, reversible, or impossible. These three possibilities were discussed in Sec. 3.2 and are summarized below.
A spontaneous process is a real process that can actually take place in a nite time
period.
A reversible process is an imaginary, idealized process in which the system passes
through a continuous sequence of equilibrium states. This sequence of states can be
approached by a spontaneous process in the limit of innite slowness, and so also can
the reverse sequence of states.
An impossible process is a change that cannot occur under the existing conditions,
even in a limiting sense. It is also known as an unnatural or disallowed process. Sometimes it is useful to describe a hypothetical impossible process that we can imagine
but that does not occur in reality. The second law of thermodynamics will presently
be introduced with two such impossible processes.
The spontaneous processes relevant to chemistry are irreversible. An irreversible process is a spontaneous process whose reverse is an impossible process.
There is also the special category, of little interest to chemists, of purely mechanical
processes. A purely mechanical process is a spontaneous process whose reverse is also
spontaneous.
102
CHAPTER 4 THE SECOND LAW

4.2 S TATEMENTS OF THE S ECOND L AW
103
It is true that reversible processes and purely mechanical processes are idealized processes that cannot occur in practice, but a spontaneous process can be practically reversible
if carried out sufciently slowly, or practically purely mechanical if friction and temperature
gradients are negligible. In that sense, they are not impossible processes. This book will
reserve the term impossible for a process that cannot be approached by any spontaneous
process, no matter how slowly or how carefully it is carried out.
4.2 STATEMENTS OF THE SECOND LAW

A description of the mathematical statement of the second law is given in the box below.
dS D q=Tb for a reversible change of a closed system;
dS > q=Tb for an irreversible change of a closed system;
where S is an extensive state function, the entropy, and
q is an innitesimal quantity of energy transferred
by heat at a portion of the boundary where the
thermodynamic temperature is Tb .
The box includes three distinct parts. First, there is the assertion that a property called
entropy, S, is an extensive state function. Second, there is an equation for calculating
the entropy change of a closed system during a reversible change of state: dS is equal to
q=Tb .1 Third, there is a criterion for spontaneity: dS is greater than q=Tb during an
irreversible change of state. The temperature Tb is a thermodynamic temperature, which
will be dened in Sec. 4.3.4.
Each of the three parts is an essential component of the second law, but is somewhat
abstract. What fundamental principle, based on experimental observation, may we take
as the starting point to obtain them? Two principles are available, one associated with
Clausius and the other with Kelvin and Planck. Both principles are equivalent statements of
the second law. Each asserts that a certain kind of process is impossible, in agreement with
common experience.
Consider the process depicted in Fig. 4.1(a) on the next page. The system is isolated,
and consists of a cool body in thermal contact with a warm body. During the process, energy
is transferred by means of heat from the cool to the warm body, causing the temperature of
the cool body to decrease and that of the warm body to increase. Of course, this process
is impossible; we never observe heat ow from a cooler to a warmer body. (In contrast,
the reverse process, heat transfer from the warmer to the cooler body, is spontaneous and
irreversible.) Note that this impossible process does not violate the rst law, because energy
is conserved.
Suppose we attempt to bring about the same changes in the two bodies by interposing a
device of some sort between them, as depicted in Fig. 4.1(b). Here is how we would like the
device to operate in the isolated system: Heat should ow from the cool body to the device,
1 During a reversible process, the temperature usually has the same value T
throughout the system, in which case

we can simply write dS D q=T . The equation dS D q=Tb allows for the possibility that in an equilibrium
state the system has phases of different temperatures separated by internal adiabatic partitions.

104
warm
warm
cool
device
cool
(a)
(b)
Figure 4.1 Two impossible processes in isolated systems.

(a) Heat transfer from a cool to a warm body.
(b) The same, with a device that operates in a cycle.
an equal quantity of heat should ow from the device to the warm body, and the nal state
of the device should be the same as its initial state. In other words, we want the device to
transfer energy quantitatively by means of heat from the cool body to the warm body while
operating in a cycle. If the device could do this, there would be no limit to the quantity of
energy that could be transferred by heat, because after each cycle the device would be ready
to repeat the process. But experience shows that it is impossible to build such a device! The
proposed process of Fig. 4.1(b) is impossible even in the limit of innite slowness.
The general principle was expressed by Rudolph Clausius2 in the words: Heat can
never pass from a colder to a warmer body without some other change, connected therewith, occurring at the same time. For use in the derivation to follow, the statement can be
reworded as follows.
The Clausius statement of the second law: It is impossible to construct a device whose
only effect, when it operates in a cycle, is heat transfer from a body to the device and
the transfer by heat of an equal quantity of energy from the device to a warmer body.
Next consider the impossible process shown in Fig. 4.2(a) on the next page. A Joule
paddle wheel rotates in a container of water as a weight rises. As the weight gains potential
energy, the water loses thermal energy and its temperature decreases. Energy is conserved,
so there is no violation of the rst law. This process is just the reverse of the Joule paddlewheel experiment (Sec. 3.7.2) and its impossibility was discussed on page 66.
We might again attempt to use some sort of device operating in a cycle to accomplish
the same overall process, as in Fig. 4.2(b). A closed system that operates in a cycle and
does net work on the surroundings is called a heat engine. The heat engine shown in Fig.
4.2(b) is a special one. During one cycle, a quantity of energy is transferred by heat from
a heat reservoir to the engine, and the engine performs an equal quantity of work on a
weight, causing it to rise. At the end of the cycle, the engine has returned to its initial
state. This would be a very desirable engine, because it could convert thermal energy into
an equal quantity of useful mechanical work with no other effect on the surroundings.3 The
2 Ref.
3 This
[32], page 117.

hypothetical process is called perpetual motion of the second kind.

105
heat engine
(a)
(b)
Figure 4.2 Two more impossible processes.

(a) A weight rises as a liquid becomes cooler.
(b) The same, with a heat engine.
engine could power a ship; it would use the ocean as a heat reservoir and require no fuel.
Unfortunately, it is impossible to construct such a heat engine!
The principle was expressed by William Thomson (Lord Kelvin) in 1852 as follows:
It is impossible by means of inanimate material agency to derive mechanical effect from
any portion of matter by cooling it below the temperature of the coldest of the surrounding
objects. Max Planck4 gave this statement: It is impossible to construct an engine which
will work in a complete cycle, and produce no effect except the raising of a weight and the
cooling of a heat-reservoir. For the purposes of this chapter, the principle can be reworded
as follows.
The KelvinPlanck statement of the second law: It is impossible to construct a heat engine whose only effect, when it operates in a cycle, is heat transfer from a heat reservoir
to the engine and the performance of an equal quantity of work on the surroundings.
Both the Clausius statement and the KelvinPlanck statement assert that certain processes, although they do not violate the rst law, are nevertheless impossible.
These processes would not be impossible if we could control the trajectories of large
numbers of individual particles. Newtons laws of motion are invariant to time reversal. Suppose we could measure the position and velocity of each molecule of a
macroscopic system in the nal state of an irreversible process. Then, if we could
somehow arrange at one instant to place each molecule in the same position with its
velocity reversed, and if the molecules behaved classically, they would retrace their
trajectories in reverse and we would observe the reverse impossible process.
The plan of the remaining sections of this chapter is as follows. In Sec. 4.3, a hypothetical device called a Carnot engine is introduced and used to prove that the two physical
statements of the second law (the Clausius statement and the KelvinPlanck statement) are
equivalent, in the sense that if one is true, so is the other. An expression is also derived for
the efciency of a Carnot engine for the purpose of dening thermodynamic temperature.
Section 4.4 combines Carnot cycles and the KelvinPlanck statement to derive the existence
4 Ref.
[131], p. 89.

4.3 C ONCEPTS D EVELOPED WITH C ARNOT E NGINES
106
Pa
Figure 4.3 Indicator diagram for a Carnot engine using an ideal gas as the working
substance. In this example, Th D 400 K, Tc D 300 K, D 1=4, CV;m D .3=2/R,
n D 2:41 mmol. The processes of paths A!B and C!D are isothermal; those of
paths B!C, B0 !C0 , and D!A are adiabatic. The cycle A!B!C!D!A has net
work w D 1:0 J; the cycle A!B0 !C0 !D!A has net work w D 0:5 J.
and properties of the state function called entropy. Section 4.5 uses irreversible processes
to complete the derivation of the mathematical statements given in the box on page 103,
Sec. 4.6 describes some applications, and Sec. 4.7 is a summary. Finally, Sec. 4.8 briey
describes a microscopic, statistical interpretation of entropy.
Carnot engines and Carnot cycles are admittedly outside the normal experience of
chemists, and using them to derive the mathematical statement of the second law may
seem arcane. G. N. Lewis and M. Randall, in their classic 1923 book Thermodynamics and the Free Energy of Chemical Substances,5 complained of the presentation of
cyclical processes limping about eccentric and not quite completed cycles. There
seems, however, to be no way to carry out a rigorous general derivation without invoking thermodynamic cycles. You may avoid the details by skipping Secs. 4.34.5.
(Incidently, the cycles described in these sections are complete!)
4.3 CONCEPTS DEVELOPED WITH CARNOT ENGINES

4.3.1 Carnot engines and Carnot cycles
A heat engine, as mentioned in Sec. 4.2, is a closed system that converts heat to work
and operates in a cycle. A Carnot engine is a particular kind of heat engine, one that
performs Carnot cycles with a working substance. A Carnot cycle has four reversible
steps, alternating isothermal and adiabatic; see the examples in Figs. 4.3 and 4.4 in which
the working substances are an ideal gas and H2 O, respectively.
5 Ref.
[103], p. 2.

107
BIOGRAPHICAL SKETCH
Sadi Carnot (17961832)

carried as far as possible.
(3) It should be so arranged that the passage
of the elastic uid from the highest to the lowest
temperature should be due to increase of volume;
that is, it should be so arranged that the cooling
of the gas should occur spontaneously as the result of rarefaction [i.e., adiabatic expansion].
Carnot derived these principles from the abstract reversible cycle now called the Carnot
cycle. He assumed the validity of the caloric
theory (heat as an indestructible substance),
which requires that the net heat in the cycle be
zero, whereas today we would say that it is the
net entropy change that is zero.
Despite the aw of assuming that heat is
conserved, a view which there is evidence he
was beginning to doubt, his conclusion was
valid that the efciency of a reversible cycle
operating between two xed temperatures is
independent of the working substance. He
based his reasoning on the impossibility of the
perpetual motion which would result by combining the cycle with the reverse of a more efcient cycle. Regarding Carnots accomplishment, William Thomson (later Lord Kelvin)
wrote:
Sadi Carnot was the eldest son of Lazare

Carnot, a famous French anti-royalist politician, one of Napoleons generals with a great
interest in mathematics. As a boy Sadi was
shy and sensitive. He studied at the Ecole

Polytechnique, a training school for army engineers, and became an army ofcer.
Carnot is renowned for the one book he
wrote: a treatise of 118 pages entitled Reections on the Motive Power of Fire and on Machines Fitted to Develop that Power. This was
published in 1824, when he was 28 and had
retired from the army on half pay.
The book was written in a nontechnical
style and went virtually unnoticed. Its purpose
Nothing in the whole range of Natural Philosowas to show how the efciency of a steam enphy is more remarkable than the establishment
gine could be improved, a very practical matter
of general laws by such a process of reasoning.
since French power technology lagged behind
A biographer described Carnots personalthat of Britain at the time:a
ity as follows:b
Notwithstanding the work of all kinds done by
steam-engines, notwithstanding the satisfactory
condition to which they have been brought today,
their theory is very little understood, and the attempts to improve them are still directed almost
by chance.
. . . We can easily conceive a multitude of machines tted to develop the motive power of heat
through the use of elastic uids; but in whatever
way we look at it, we should not lose sight of the
following principles:
(1) The temperature of the uid should be
made as high as possible, in order to obtain a
great fall of caloric, and consequently a large
production of motive power.
(2) For the same reason the cooling should be
a Ref.
[27].
b Ref.
He was reserved, almost taciturn, with a hatred

of any form of publicity. . . . his friends all spoke
of his underlying warmth and humanity. Passionately fond of music, he was an excellent violinist who preferred the classical Lully to the
moderns of the time; he was devoted to literature and all the arts.
Carnot came down with scarlet fever and,

while convalescing, diedprobably of the
cholera epidemic then raging. He was only 36.
Two years later his work was brought to
public attention in a paper written by Emile

Clapeyron (page 218), who used indicator diagrams to explain Carnots ideas.
[114], page x.

108
Pa
Figure 4.4 Indicator diagram for a Carnot engine using H2 O as the working substance. In this example, Th D 400 K, Tc D 396 K, D 1=100, w D 1:0 J. In
state A, the system consists of one mole of H2 O(l). The processes (all carried out
reversibly) are: A!B, vaporization of 2:54 mmol H2 O at 400 K; B!C, adiabatic
expansion, causing vaporization of an additional 7:68 mmol; C!D, condensation of
2:50 mmol at 396 K; D!A, adiabatic compression returning the system to the initial
state.
The steps of a Carnot cycle are as follows. In this description, the system is the working
substance.
Path A!B: A quantity of heat qh is transferred reversibly and isothermally from a heat
reservoir (the hot reservoir) at temperature Th to the system, also at temperature Th .
qh is positive because energy is transferred into the system.
Path B!C: The system undergoes a reversible adiabatic change that does work on the
surroundings and reduces the system temperature to Tc .
Path C!D: A quantity of heat qc is transferred reversibly and isothermally from the system
to a second heat reservoir (the cold reservoir) at temperature Tc . qc is negative.
Path D!A: The system undergoes a reversible adiabatic change in which work is done on
the system, the temperature returns to Th , and the system returns to its initial state to
complete the cycle.
In one cycle, a quantity of heat is transferred from the hot reservoir to the system, a
portion of this energy is transferred as heat to the cold reservoir, and the remainder of the
energy is the negative net work w done on the surroundings. (It is the heat transfer to
the cold reservoir that keeps the Carnot engine from being an impossible KelvinPlanck
engine.) Adjustment of the length of path A!B makes the magnitude of w as large or
small as desirednote the two cycles with different values of w described in the caption of
Fig. 4.3.
The Carnot engine is an idealized heat engine because its paths are reversible processes. It does not resemble the design of any practical steam engine. In a typical
working steam engine, such as those once used for motive power in train locomotives
and steamships, the cylinder contains an open system that undergoes the following irreversible steps in each cycle: (1) high-pressure steam enters the cylinder from a boiler

Th
109
Th
4
1
1
3
Tc
Tc
(a)
(b)
Figure 4.5 (a) One cycle of a Carnot engine that does work on the surroundings.
(b) The same system run in reverse as a Carnot heat pump.
Figures 4.54.7 use the following symbols: A square box represents a system (a
Carnot engine or Carnot heat pump). Vertical arrows indicate heat and horizontal
arrows indicate work; each arrow shows the direction of energy transfer into or out of
the system. The number next to each arrow is an absolute value of q/J or w/J in the
cycle. For example, (a) shows 4 joules of heat transferred to the system from the hot
reservoir, 3 joules of heat transferred from the system to the cold reservoir, and 1 joule
of work done by the system on the surroundings.
and pushes the piston from the closed end toward the open end of the cylinder; (2) the
supply valve closes and the steam expands in the cylinder until its pressure decreases
to atmospheric pressure; (3) an exhaust valve opens to release the steam either to the
atmosphere or to a condenser; (4) the piston returns to its initial position, driven either
by an external force or by suction created by steam condensation.
The energy transfers involved in one cycle of a Carnot engine are shown schematically
in Fig. 4.5(a). When the cycle is reversed, as shown in Fig. 4.5(b), the device is called a
Carnot heat pump. In each cycle of a Carnot heat pump, qh is negative and qc is positive.
Since each step of a Carnot engine or Carnot heat pump is a reversible process, neither
device is an impossible device.
4.3.2 The equivalence of the Clausius and KelvinPlanck statements

We can use the logical tool of reductio ad absurdum to prove the equivalence of the Clausius
and KelvinPlanck statements of the second law.
Let us assume for the moment that the Clausius statement is incorrect, and that the device the Clausius statement claims is impossible (a Clausius device) is actually possible.
If the Clausius device is possible, then we can combine one of these devices with a Carnot
engine as shown in Fig. 4.6(a) on page 111. We adjust the cycles of the Clausius device and
Carnot engine to transfer equal quantities of heat from and to the cold reservoir. The combination of the Clausius device and Carnot engine is a system. When the Clausius device
and Carnot engine each performs one cycle, the system has performed one cycle as shown
in Fig. 4.6(b). There has been a transfer of heat into the system and the performance of an
equal quantity of work on the surroundings, with no other net change. This system is a heat
engine that according to the KelvinPlanck statement is impossible.
Thus, if the KelvinPlanck statement is correct, it is impossible to operate the Clausius

110
BIOGRAPHICAL SKETCH
Rudolf Julius Emmanuel Clausius (18221888)

therewith, occurring at the same time. Everything we know concerning the interchange of
heat between two bodies of different temperature
conrms this; for heat everywhere manifests a
tendency to equalize existing differences of temperature, and therefore to pass in a contrary direction, i. e. from warmer to colder bodies. Without further explanation, therefore, the truth of the
principle will be granted.
Rudolf Clausius was a German theoretical

physicist who was the rst to treat thermodynamics as a rigorous science, based on the earlier writings of Carnot and Clapeyron.
He was born in K slin, Prussia, into a large
o
family. His father was an educator and church
minister.
Clausius was successively a professor at
universities in Berlin, Zurich, W rzburg, and
u
Bonn. In addition to thermodynamics, he did
work on electrodynamic theory and the kinetic
theory of gases.
Max Planck, referring to a time early in his
own career, wrote:a
One day, I happened to come across the treatises of Rudolf Clausius, whose lucid style and
enlightening clarity of reasoning made an enormous impression on me, and I became deeply absorbed in his articles, with an ever increasing enthusiasm. I appreciated especially his exact formulation of the two Laws of Thermodynamics,
and the sharp distinction which he was the rst
to establish between them.
Clausius based his exposition of the second

law on the following principle that he published in 1854:b
. . . it appears to me preferable to deduce the general form of the theorem immediately from the
same principle which I have already employed
in my former memoir, in order to demonstrate
the modied theorem of Carnot.
This principle, upon which the whole of the
following development rests, is as follows:
Heat can never pass from a colder to a warmer
body without some other change, connected
a Ref.
[132], page 16.
b Ref.
[32], page 117.
c Ref.
In an 1865 paper, he introduced the symbol

U for internal energy, and also coined the word
entropy with symbol S:c
We might call S the transformational content of
the body, just as we termed the magnitude U
its thermal and ergonal content. But as I hold
it better to borrow terms for important magnitudes from the ancient languages, so that they
may be adopted unchanged in all modern languages, I propose to call the magnitude S the entropy of the body, from the Greek word o ,
J
transformation. I have intentionally formed the
word entropy so as to be as similar as possible
to the word energy; for the two magnitudes to
be denoted by these words are so nearly allied in
their physical meanings, that a certain similarity
in designation appears to be desirable.
The 1865 paper concludes as follows, ending with Clausiuss often-quoted summations
of the rst and second laws:d
If for the entire universe we conceive the same
magnitude to be determined, consistently and
with due regard to all circumstances, which for
a single body I have called entropy, and if at the
same time we introduce the other and simpler
conception of energy, we may express in the following manner the fundamental laws of the universe which correspond to the two fundamental
theorems of the mechanical theory of heat.
1. The energy of the universe is constant.
2. The entropy of the universe tends to a maximum.
Clausius was a patriotic German. During

the Franco-Prussian war of 187071, he undertook the leadership of an ambulance corps
composed of Bonn students, was wounded in
the leg during the battles, and suffered disability for the rest of his life.
[33], page 357.
d Ref.
[33], page 365.

Th
Th
Th
Th
1
1
1
3
Tc
(a)
Th
Tc
Th
1
1
111
Tc
(b)
(c)
Tc
(d)
Figure 4.6 (a) A Clausius device combined with the Carnot engine of Fig. 4.5(a).
(b) The resulting impossible KelvinPlanck engine.
(c) A KelvinPlanck engine combined with the Carnot heat pump of Fig. 4.5(b).
(d) The resulting impossible Clausius device.
device as shown, and our provisional assumption that the Clausius statement is incorrect
must be wrong. In conclusion, if the KelvinPlanck statement is correct, then the Clausius
statement must also be correct.
We can apply a similar line of reasoning to the heat engine that the KelvinPlanck
statement claims is impossible (a KelvinPlanck engine) by seeing what happens if we
assume this engine is actually possible. We combine a KelvinPlanck engine with a Carnot
heat pump, and make the work performed on the Carnot heat pump in one cycle equal to
the work performed by the KelvinPlanck engine in one cycle, as shown in Fig. 4.6(c). One
cycle of the combined system, shown in Fig. 4.6(d), shows the system to be a device that
the Clausius statement says is impossible. We conclude that if the Clausius statement is
correct, then the KelvinPlanck statement must also be correct.
These conclusions complete the proof that the Clausius and KelvinPlanck statements
are equivalent: the truth of one implies the truth of the other. We may take either statement
as the fundamental physical principle of the second law, and use it as the starting point for
deriving the mathematical statement of the second law. The derivation will be taken up in
Sec. 4.4.
4.3.3 The efciency of a Carnot engine

Integrating the rst-law equation dU D q C w over one cycle of a Carnot engine, we
obtain
0 D qh C qc C w
(4.3.1)
(one cycle of a Carnot engine)
The efciency of a heat engine is dened as the fraction of the heat input qh that is returned
as net work done on the surroundings:
def
w
qh
(4.3.2)

Th
Th
5
1
Th
Tc
(b)
Tc
Th
4
1
Tc
(a)
Th
Tc
Th
112
Tc
(c)
Tc
(d)
Figure 4.7 (a) A Carnot engine of efciency D 1=4 combined with a Carnot engine
of efciency D 1=5 run in reverse.
(b) The resulting impossible Clausius device.
(c) A Carnot engine of efciency D 1=3 combined with the Carnot engine of efciency D 1=4 run in reverse.
(d) The resulting impossible Clausius device.
By substituting for w from Eq. 4.3.1, we obtain

D1C
qc
qh
(4.3.3)
(Carnot engine)
Because qc is negative, qh is positive, and qc is smaller in magnitude than qh , the efciency

is less than one. The example shown in Fig. 4.5(a) is a Carnot engine with D 1=4.
We will be able to reach an important conclusion regarding efciency by considering a
Carnot engine operating between the temperatures Th and Tc , combined with a Carnot heat
pump operating between the same two temperatures. The combination is a supersystem, and
one cycle of the engine and heat pump is one cycle of the supersystem. We adjust the cycles
of the engine and heat pump to produce zero net work for one cycle of the supersystem.
Could the efciency of the Carnot engine be different from the efciency the heat pump
would have when run in reverse as a Carnot engine? If so, either the supersystem is an
impossible Clausius device as shown in Fig. 4.7(b), or the supersystem operated in reverse
(with the engine and heat pump switching roles) is an impossible Clausius device as shown
in Fig. 4.7(d). We conclude that all Carnot engines operating between the same two temperatures have the same efciency.
This is a good place to pause and think about the meaning of this statement in light of
the fact that the steps of a Carnot engine, being reversible changes, cannot take place in
a real system (Sec. 3.2). How can an engine operate that is not real? The statement is
an example of a common kind of thermodynamic shorthand. To express the same idea
more accurately, one could say that all heat engines (real systems) operating between
the same two temperatures have the same limiting efciency, where the limit is the
reversible limit approached as the steps of the cycle are carried out more and more
slowly. You should interpret any statement involving a reversible process in a similar
fashion: a reversible process is an idealized limiting process that can be approached
but never quite reached by a real system.

113
Thus, the efciency of a Carnot engine must depend only on the values of Tc and Th
and not on the properties of the working substance. Since the efciency is given by D
1 C qc =qh , the ratio qc =qh must be a unique function of Tc and Th only. To nd this function
for temperatures on the ideal-gas temperature scale, it is simplest to choose as the working
substance an ideal gas.
An ideal gas has the equation of state pV D nRT . Its internal energy change in a
closed system is given by dU D CV dT (Eq. 3.5.3), where CV (a function only of T ) is
the heat capacity at constant volume. Reversible expansion work is given by w D p dV ,
which for an ideal gas becomes w D .nRT =V / dV . Substituting these expressions for
dU and w in the rst law, dU D q C w, and solving for q, we obtain
q D CV dT C
nRT
dV
V
(4.3.4)
(ideal gas, reversible
expansion work only)
Dividing both sides by T gives

q
C dT
dV
D V
C nR
T
T
T
(4.3.5)
(ideal gas, reversible
expansion work only)
In the two adiabatic steps of the Carnot cycle, q is zero. We obtain a relation among the
volumes of the four labeled states shown in Fig. 4.3 by integrating Eq. 4.3.5 over these steps
and setting the integrals equal to zero:
Z Tc
Z
CV dT
V
q
D
C nR ln C D 0
(4.3.6)
Path B!C:
T
T
VB
Th
Z
Z Th
q
CV dT
V
Path D!A:
D
C nR ln A D 0
(4.3.7)
T
T
VD
Tc
Adding these two equations (the integrals shown with limits cancel) gives the relation
nR ln
VA VC
D0
VB VD
(4.3.8)
which we can rearrange to

ln.VB =VA / D
ln.VD =VC /
(4.3.9)
(ideal gas, Carnot cycle)
We obtain expressions for the heat in the two isothermal steps by integrating Eq. 4.3.4 with
dT set equal to 0.
Path A!B W
Path C!D W
qh D nRTh ln.VB =VA /

qc D nRTc ln.VD =VC /
(4.3.10)
(4.3.11)
The ratio of qc and qh obtained from these expressions is

qc
T
D c
qh
Th
ln.VD =VC /
ln.VB =VA /
(4.3.12)

114
By means of Eq. 4.3.9, this ratio becomes

qc
D
qh
Tc
Th
(4.3.13)
(Carnot cycle)
Accordingly, the unique function of Tc and Th we seek that is equal to qc =qh is the ratio
Tc =Th . The efciency, from Eq. 4.3.3, is then given by
D1
Tc
Th
(4.3.14)
(Carnot engine)
In Eqs. 4.3.13 and 4.3.14, Tc and Th are temperatures on the ideal-gas scale. As we have
seen, these equations must be valid for any working substance; it is not necessary to specify
as a condition of validity that the system is an ideal gas.
The ratio Tc =Th is positive but less than one, so the efciency is less than one as deduced
earlier on page 112. This conclusion is an illustration of the KelvinPlanck statement of the
second law: A heat engine cannot have an efciency of unitythat is, it cannot in one
cycle convert all of the energy transferred by heat from a single heat reservoir into work.
The example shown in Fig. 4.5 on page 109, with D 1=4, must have Tc =Th D 3=4 (e.g.,
Tc D 300 K and Th D 400 K).
Keep in mind that a Carnot engine operates reversibly between two heat reservoirs. The
expression of Eq. 4.3.14 gives the efciency of this kind of idealized heat engine only. If
any part of the cycle is carried out irreversibly, dissipation of mechanical energy will cause
the efciency to be lower than the theoretical value given by Eq. 4.3.14.
4.3.4 Thermodynamic temperature

The negative ratio qc =qh for a Carnot cycle depends only on the temperatures of the two heat
reservoirs. Kelvin (1848) proposed that this ratio be used to establish an absolute temperature scale. The physical quantity now called thermodynamic temperature is dened by
the relation
Tc
D
Th
qc
qh
(4.3.15)
(Carnot cycle)
That is, the ratio of the thermodynamic temperatures of two heat reservoirs is equal, by
denition, to the ratio of the absolute quantities of heat transferred in the isothermal steps
of a Carnot cycle operating between these two temperatures. In principle, a measurement of
qc =qh during a Carnot cycle, combined with a dened value of the thermodynamic temperature of one of the heat reservoirs, can establish the thermodynamic temperature of the other
heat reservoir. This dened value is provided by the triple point of H2 O; its thermodynamic
temperature is dened as exactly 273:16 kelvins (page 40).
Just as measurements with a gas thermometer in the limit of zero pressure establish
the ideal-gas temperature scale (Sec. 2.3.5), the behavior of a heat engine in the reversible
limit establishes the thermodynamic temperature scale. Note, however, that a reversible
Carnot engine used as a thermometer to measure thermodynamic temperature is only a

115
BIOGRAPHICAL SKETCH
William Thomson, Lord Kelvin (18241907)
William Thomson was born in Belfast, Ireland.

His mother died when he was six. In 1832 the
family moved to Glasgow, Scotland, where his
father had been appointed as the chair of mathematics at the University.
In 1845 Thomson was age 21. He had recently graduated from Cambridge University
with high honors in mathematics, and was in
Paris for discussions with French physicists
and mathematicians. He had learned about
Carnots book from Clapeyrons 1834 paper
but could not nd a copyno bookseller in
Paris had even heard of it. Nevertheless, the information he had was sufcient for him to realize that Carnots ideas would allow a thermodynamic temperature scale to be dened, one
that does not depend on any particular gas.
The following year, Thomson became the
Chair of Natural Philosophy at Glasgow University, largely through the inuence of his father. He remained there until his retirement in
1899. His concept of a thermodynamic temperature scale, his best-known contribution to
thermodynamics, was published in 1848.a
Thomson published other papers on the theory of heat, but his ideas eventually changed
as a result of hearing a presentation by James
Joule at a meeting in Oxford in 1847. Joule
was describing his experiments with the conversion of work to thermal energy by a paddle
wheel. In a letter written to his nephew, J. T.
Bottomley, Thomson wrote:b
I made Joules acquaintance at the Oxford meeting, and it quickly ripened into a life-long friendship.
a Ref.
[89].
b Ref.
[17].
c Ref.
I heard his paper read in the section, and

felt strongly impelled at rst to rise and say
that it must be wrong because the true mechanical value of heat given, suppose in warm water, must, for small differences of temperature,
be proportional to the square of its quantity. I
knew from Carnot that this must be true (and it is
true; only now I call it motivity, to avoid clashing with Joules mechanical value.) But as I
listened on and on, I saw that (though Carnot
had vitally important truth, not to be abandoned)
Joule had certainly a great truth and a great discovery, and a most important measurement to
bring forward. So instead of rising with my objection to the meeting I waited till it was over,
and said my say to Joule himself, at the end of
the meeting. This made my rst introduction to
him. After that I had a long talk over the whole
matter at one of the conversaziones of the Association, and we became fast friends from thenceforward.
The physicist Charles Everitt described

Thomsons personality as follows:c
Thomson was the kind of man who created all
around him a sense of bustle and excitement.
. . . [He] was a man of violent enthusiasms. He
would take up a subject, work at it furiously for
a few weeks, throw out a string of novel ideas,
and then inexplicably drop everything and pass
on.
During his career, Thomson published more

than 600 papers and received many honors. In
addition to the theory of heat, his work included a dynamical theory of electricity and
magnetism, and the invention of the mirror galvanometer used as the telegraph receiver for
the rst transatlantic submarine cables.
Thomson was the chief technical consultant
for the initial cable projects. As a result of
their success and his involvement in the construction of a global cable network, he became
extremely wealthy and was knighted in 1866.
In 1892 he became Baron Kelvin of Largs.
Kelvin is the name of the river that ows past
Glasgow University, Largs is the town where
he had his home, and kelvin is now the SI unit
of thermodynamic temperature.
[51].

4.4 D ERIVATION OF THE M ATHEMATICAL S TATEMENT OF THE S ECOND L AW
116
theoretical concept and not a practical instrument, since a completely-reversible process

cannot occur in practice.
It is now possible to justify the statement in Sec. 2.3.5 that the ideal-gas temperature
scale is proportional to the thermodynamic temperature scale. Both Eq. 4.3.13 and Eq.
4.3.15 equate the ratio Tc =Th to qc =qh ; but whereas Tc and Th refer in Eq. 4.3.13 to the
ideal-gas temperatures of the heat reservoirs, in Eq. 4.3.15 they refer to the thermodynamic
temperatures. This means that the ratio of the ideal-gas temperatures of two bodies is equal
to the ratio of the thermodynamic temperatures of the same bodies, and therefore the two
scales are proportional to one another. The proportionality factor is arbitrary, but must be
unity if the same unit (e.g., kelvins) is used in both scales. Thus, as stated on page 41, the
two scales expressed in kelvins are identical.
4.4 DERIVATION OF THE MATHEMATICAL STATEMENT OF THE

SECOND LAW
4.4.1 The existence of the entropy function
This section derives the existence and properties of the state function called entropy.
Consider an arbitrary cyclic process of a closed system. To avoid confusion, this system
will be the experimental system and the process will be the experimental process or experimental cycle. There are no restrictions on the contents of the experimental systemit
may have any degree of complexity whatsoever. The experimental process may involve
more than one kind of work, phase changes and reactions may occur, there may be temperature and pressure gradients, constraints and external elds may be present, and so on. All
parts of the process must be either irreversible or reversible, but not impossible.
We imagine that the experimental cycle is carried out in a special way that allows us to
apply the KelvinPlanck statement of the second law. The heat transferred across the boundary of the experimental system in each innitesimal path element of the cycle is exchanged
with a hypothetical Carnot engine. The combination of the experimental system and the
Carnot engine is a closed supersystem (see Fig. 4.8 on page 118). In the surroundings of
the supersystem is a heat reservoir of arbitrary constant temperature Tres . By allowing the
supersystem to exchange heat with only this single heat reservoir, we will be able to apply
the KelvinPlanck statement to a cycle of the supersystem.6
We assume that we are able to control changes of the work coordinates of the experimental system from the surroundings of the supersystem. We are also able to control the
Carnot engine from these surroundings, for example by moving the piston of a cylinderand-piston device containing the working substance. Thus the energy transferred by work
across the boundary of the experimental system, and the work required to operate the Carnot
engine, is exchanged with the surroundings of the supersystem.
During each stage of the experimental process with nonzero heat, we allow the Carnot
engine to undergo many innitesimal Carnot cycles with innitesimal quantities of heat and
work. In one of the isothermal steps of each Carnot cycle, the Carnot engine is in thermal
contact with the heat reservoir, as depicted in Fig. 4.8(a). In this step the Carnot engine
has the same temperature as the heat reservoir, and reversibly exchanges heat q 0 with it.
6 This
procedure is similar to ones described in Ref. [1], Chap. 5, and Ref. [97], p. 53.

117
BIOGRAPHICAL SKETCH
Max Karl Ernst Ludwig Planck (18581947)

was declared to be unessential; furthermore, the
existence of an absolute zero of temperature was
disputed, on the ground that for temperature, just
as for height, only differences can be measured.
It is one of the most painful experiences of my
entire scientic life that I have but seldomin
fact, I might say, neversucceeded in gaining
universal recognition for a new result, the truth
of which I could demonstrate by a conclusive,
albeit only theoretical proof. This is what happened this time, too. All my sound arguments
fell on deaf ears.
Max Planck, best known for his formulation of

the quantum theory, had a passionate interest
in thermodynamics in the early part of his career.
He was born in Kiel, Germany, where his
father was a distinguished law professor. His
family had a long tradition of conservatism,
idealism, and excellence in scholarship.
As a youth, Planck had difculty deciding
between music and physics as a career, nally
settling on physics. He acquired his interest in
thermodynamics from studies with Hermann
von Helmholtz and Gustav Kirchhoff and from
the writings of Rudolf Clausius. His doctoral
dissertation at the University of Munich (1879)
was on the second law.
In 1897, Planck turned his papers on thermodynamics into a concise introductory textbook, Treatise on Thermodynamics. It went
through at least seven editions and has been
translated into English.a
Concerning the second law he wrote:b
Another controversy arose with relation to the
question of the analogy between the passage
of heat from a higher to a lower temperature
and the sinking of a weight from a greater to
a smaller height. I had emphasized the need
for a sharp distinction between these two processes. . . However, this theory of mine was contradicted by the view universally accepted in
those days, and I just could not make my fellow
physicists see it my way. . .
A consequence of this point of view [held by
others] was that the assumption of irreversibility
for proving the Second Law of Thermodynamics
a Ref.
[131].
b Ref.
Planck became an associate professor of

physics at the University of Kiel. In 1889
he succeeded Kirchhoff as Professor at Berlin
University. By the end of the following year,
at the age of 42, he had worked out his quantum theory to explain the experimental facts of
blackbody radiation, a formulation that started
a revolution in physics. He was awarded the
1918 Nobel Prize in Physics in recognition of
the services he rendered to the advancement of
Physics by his discovery of energy quanta.
Planck was reserved and only enjoyed socializing with persons of similar rank. He
was a gifted pianist with perfect pitch, and enjoyed hiking and climbing well into his old
age. He was known for his fairness, integrity,
and moral force.
He endured many personal tragedies in his
later years. His rst wife died after 22 years
of a happy marriage. His elder son was killed
in action during World War I. Both of his twin
daughters died in childbirth.
Planck openly opposed the Nazi persecution of Jews but remained in Germany during World War II out of a sense of duty.
The war brought further tragedy: his house in
Berlin was destroyed by bombs, and his second son was implicated in the failed 1944 attempt to assassinate Hitler and was executed
by the Gestapo. Planck and his second wife
escaped the bombings by hiding in the woods
and sleeping in haystacks in the countryside.
They were rescued by American troops in
May, 1945.
[132], pages 2930.

Tres
118
Tres
q 0
Tres
Tb
Tb
experimental
system
q
Tb
experimental
system
(a)
(b)
Figure 4.8 Experimental system, Carnot engine (represented by a small square box),
and heat reservoir. The dashed lines indicate the boundary of the supersystem.
(a) Reversible heat transfer between heat reservoir and Carnot engine.
(b) Heat transfer between Carnot engine and experimental system. The innitesimal
quantities q 0 and q are positive for transfer in the directions indicated by the arrows.
The sign convention is that q 0 is positive if heat is transferred in the direction of the arrow,
from the heat reservoir to the Carnot engine.
In the other isothermal step of the Carnot cycle, the Carnot engine is in thermal contact
with the experimental system at a portion of the systems boundary. as depicted in Fig.
4.8(b). The Carnot engine now has the same temperature, Tb , as the experimental system at
this part of the boundary, and exchanges heat with it. The heat q is positive if the transfer
is into the experimental system.
The relation between temperatures and heats in the isothermal steps of a Carnot cycle
is given by Eq. 4.3.15. From this relation we obtain, for one innitesimal Carnot cycle, the
relation Tb =Tres D q= q 0 , or
q
q 0 D Tres
(4.4.1)
Tb
After many innitesimal Carnot cycles, the experimental cycle is complete, the experimental system has returned to its initial state, and the Carnot engine has returned to its
initial state in thermal contact with the heat reservoir. Integration of Eq. 4.4.1 around the
experimental cycle gives the net heat entering the supersystem during the process:
I
q
0
q D Tres
(4.4.2)
Tb
The integration here is over each path element of the experimental process and over each
surface element of the boundary of the experimental system.
H
Keep in mind that the value of the cyclic integral q=Tb depends only on the path of
the experimental cycle, that this process can be reversible or irreversible, and that Tres is a
positive constant.
In this experimental cycle, could the net heat q 0 transferred to the supersystem be positive? If so, the net work would be negative (to make the internal energy change zero) and the

119
supersystem would have converted heat from a single heat reservoir completely into work,
a process the KelvinPlanck statement of the second law says is impossible. Therefore it is
impossible for q 0 to be positive, and from Eq. 4.4.2 we obtain the relation
I
q
0
(4.4.3)
Tb
(cyclic process of a closed system)
This relation is known as the Clausius inequality. It is valid only if the integration is taken
around a cyclic path in a direction with nothing but reversible and irreversible changesthe
path must not include an impossible change, such as the reverse of an irreversible change.
The Clausius inequality says that if a cyclic path meets this specication, it is impossible
H
for the cyclic integral .q=Tb / to be positive.
If the entire experimental cycle is adiabatic (which is only possible if the process is
H
reversible), the Carnot engine is not needed and Eq. 4.4.3 can be replaced by .q=Tb / D 0.
Next let us investigate a reversible nonadiabatic process of the closed experimental
system. Starting with a particular equilibrium state A, we carry out a reversible process in
which there is a net ow of heat into the system, and in which q is either positive or zero in
each path element. The nal state of this process is equilibrium state B. If each innitesimal
RB
quantity of heat q is positive or zero during the process, then the integral A .q=Tb / must
be positive. In this case the Clausius inequality tells us that if the system completes a cycle
RA
by returning from state B back to state A by a different path, the integral B .q=Tb / for
this second path must be negative. Therefore the change B!A cannot be carried out by any
adiabatic process.
Any reversible process can be carried out in reverse. Thus, by reversing the reversible
nonadiabatic process, it is possible to change the state from B to A by a reversible process
with a net ow of heat out of the system and with q either negative or zero in each element
of the reverse path. In contrast, the absence of an adiabatic path from B to A means that it
is impossible to carry out the change A!B by a reversible adiabatic process.
The general rule, then, is that whenever equilibrium state A of a closed system can be
changed to equilibrium state B by a reversible process with nite one-way heat (i.e., the
ow of heat is either entirely into the system or else entirely out of it), it is impossible for the
system to change from either of these states to the other by a reversible adiabatic process.
A simple example will relate this rule to experience. We can increase the temperature
of a liquid by allowing heat to ow reversibly into the liquid. It is impossible to
duplicate this change of state by a reversible process without heatthat is, by using
some kind of reversible work. The reason is that reversible work involves the change
of a work coordinate that brings the system to a different nal state. There is nothing
in the rule that says we cant increase the temperature irreversibly without heat, as we
can for instance with stirring work.
States A and B can be arbitrarily close. We conclude that every equilibrium state of a
closed system has other equilibrium states innitesimally close to it that are inaccessible by
a reversible adiabatic process. This is Carath odorys principle of adiabatic inaccessibility.7
e
7 Constantin
Carath odory in 1909 combined this principle with a mathematical theorem (Carath odorys thee
e
orem) to deduce the existence of the entropy function. The derivation outlined here avoids the complexities of
that mathematical treatment and leads to the same results.

120
Next let us consider the reversible adiabatic processes that are possible. To carry out
a reversible adiabatic process, starting at an initial equilibrium state, we use an adiabatic
boundary and slowly vary one or more of the work coordinates. A certain nal temperature
will result. It is helpful in visualizing this process to think of an N-dimensional space
in which each axis represents one of the N independent variables needed to describe an
equilibrium state. A point in this space represents an equilibrium state, and the path of a
reversible process can be represented as a curve in this space.
A suitable set of independent variables for equilibrium states of a closed system of uniform temperature consists of the temperature T and each of the work coordinates (Sec.
3.10). We can vary the work coordinates independently while keeping the boundary adiabatic, so the paths for possible reversible adiabatic processes can connect any arbitrary
combinations of work coordinate values.
There is, however, the additional dimension of temperature in the N-dimensional space.
Do the paths for possible reversible adiabatic processes, starting from a common initial
point, lie in a volume in the N-dimensional space? Or do they fall on a surface described
by T as a function of the work coordinates? If the paths lie in a volume, then every point
in a volume element surrounding the initial point must be accessible from the initial point
by a reversible adiabatic path. This accessibility is precisely what Carath odorys principle
e
of adiabatic inaccessibility denies. Therefore, the paths for all possible reversible adiabatic
processes with a common initial state must lie on a unique surface. This is an .N 1/dimensional hypersurface in the N-dimensional space, or a curve if N is 2. One of these
surfaces or curves will be referred to as a reversible adiabatic surface.
Now consider the initial and nal states of a reversible process with one-way heat (i.e.,
each nonzero innitesimal quantity of heat q has the same sign). Since we have seen that
it is impossible for there to be a reversible adiabatic path between these states, the points for
these states must lie on different reversible adiabatic surfaces that do not intersect anywhere
in the N-dimensional space. Consequently, there is an innite number of nonintersecting
reversible adiabatic surfaces lling the N-dimensional space. (To visualize this for N D 3,
think of a exed stack of printer paper; each sheet represents a different adiabatic surface
in three-dimensional space.) A reversible, nonadiabatic process with one-way heat is represented by a path beginning at a point on one reversible adiabatic surface and ending at a
point on a different surface. If q is positive, the nal surface lies on one side of the initial
surface, and if q is negative, the nal surface is on the opposite side.
4.4.2 Using reversible processes to dene the entropy

The existence of reversible adiabatic surfaces is the justication for dening a new state
function S , the entropy. S is specied to have the same value everywhere on one of these
surfaces, and a different, unique value on each different surface. In other words, the reversible adiabatic surfaces are surfaces of constant entropy in the N-dimensional space.
The fact that the surfaces ll this space without intersecting ensures that S is a state function for equilibrium states, because any point in this space represents an equilibrium state
and also lies on a single reversible adiabatic surface with a denite value of S.
We know the entropy function must exist, because the reversible adiabatic surfaces exist.
For instance, Fig. 4.9 on the next page shows a family of these surfaces for a closed system
of a pure substance in a single phase. In this system, N is equal to 2, and the surfaces

121
T /K
500
400
300
0:02
0:03
0:04
0:05
V /m3
Figure 4.9 A family of reversible adiabatic curves (two-dimensional reversible adiabatic surfaces) for an ideal gas with V and T as independent variables. A reversible
adiabatic process moves the state of the system along a curve, whereas a reversible
process with positive heat moves the state from one curve to another above and to the
right. The curves are calculated for n D 1 mol and CV;m D .3=2/R. Adjacent curves
differ in entropy by 1 J K 1 .
are two-dimensional curves. Each curve is a contour of constant S. At this stage in the
derivation, our assignment of values of S to the different curves is entirely arbitrary.
How can we assign a unique value of S to each reversible adiabatic surface? We can
order the values by letting a reversible process with positive one-way heat, which moves the
point for the state to a new surface, correspond to an increase in the value of S. Negative
one-way heat will then correspond to decreasing S. We can assign an arbitrary value to the
entropy on one particular reversible adiabatic surface. (The third law of thermodynamics
is used for this purposesee Sec. 6.1.) Then all that is needed to assign a value of S to
each equilibrium state is a formula for evaluating the difference in the entropies of any two
surfaces.
Consider a reversible process with positive one-way heat that changes the system from
state A to state B. The path for this process must move the system from a reversible adiabatic
surface of a certain entropy to a different surface of greater entropy. An example is the path
A!B in Fig. 4.10(a) on the next page. (The adiabatic surfaces in this gure are actually
two-dimensional curves.) As before, we combine the experimental system with a Carnot
engine to form a supersystem that exchanges heat with a single heat reservoir of constant
temperature Tres . The net heat entering the supersystem, found by integrating Eq. 4.4.1, is
q 0 D Tres
B
A
q
Tb
(4.4.4)
and it is positive.
Suppose the same experimental system undergoes a second reversible process, not necessarily with one-way heat, along a different path connecting the same pair of reversible
adiabatic surfaces. This could be path C!D in Fig. 4.10(a). The net heat entering the

B
D
122
A
C
C
!V
!V
(a)
(b)
Figure 4.10 Reversible paths in V T space. The thin curves are reversible adiabatic
surfaces.
(a) Two paths connecting the same pair of reversible adiabatic surfaces.
(b) A cyclic path.
supersystem during this second process is q 00 :

00
q D Tres
D
C
q
Tb
(4.4.5)
We can then devise a cycle of the supersystem in which the experimental system undergoes
the reversible path A!B!D!C!A, as shown in Fig. 4.10(b). Step A!B is the rst process described above, step D!C is the reverse of the second process described above, and
steps B!D and C!A are reversible and adiabatic. The net heat entering the supersystem
in the cycle is q 0 q 00 . In the reverse cycle the net heat is q 00 q 0 . In both of these cycles the
heat is exchanged with a single heat reservoir; therefore, according to the KelvinPlanck
statement, neither cycle can have positive net heat. Therefore q 0 and q 00 must be equal, and
R
Eqs. 4.4.4 and 4.4.5 then show the integral .q=Tb / has the same value when evaluated
along either of the reversible paths from the lower to the higher entropy surface.
Note that since the second path (C!D) does not necessarily have one-way heat, it
can take the experimental system through any sequence of intermediate entropy values,
provided it starts at the lower entropy surface and ends at the higher. Furthermore, since the
pathRis reversible, it can be carried out in reverse resulting in reversal of the signs of S
and .q=Tb /.
It should now be apparent that a satisfactory formula for dening the entropy change of
a reversible process in a closed system is
Z
q
S D
(4.4.6)
Tb
(reversible process,
closed system)
This formula satises the necessary requirements: it makes the value of S positive if the
process has positive one-way heat, negative if the process has negative one-way heat, and

123
zero if the process is adiabatic. It gives the same value of S for any reversible change
between the same two reversible adiabatic surfaces, and it makes the sum of the S values
of several consecutive reversible processes equal to S for the overall process.
In Eq. 4.4.6, S is the entropy change when the system changes from one arbitrary
equilibrium state to another. If the change is an innitesimal path element of a reversible
process, the equation becomes
dS D
q
Tb
(4.4.7)
(reversible process,
closed system)
It is common to see this equation written in the form dS D qrev =T , where qrev denotes
an innitesimal quantity of heat in a reversible process.
In Eq. 4.4.7, the quantity 1=Tb is called an integrating factor for q, a factor that makes
the product .1=Tb / q be the innitesimal change of a state function. The quantity
c=Tb , where c is any nonzero constant, would also be a satisfactory integrating factor;
so the denition of entropy, using cD1, is actually one of an innite number of possible
choices for assigning values to the reversible adiabatic surfaces.
4.4.3 Some properties of the entropy

It is not difcult to show that the entropy of a closed system in an equilibrium state is an
extensive property. Suppose a system of uniform temperature T is divided into two closed
subsystems A and B. When a reversible innitesimal change occurs, the entropy changes of
the subsystems are dSA D qA =T and dSB D qB =T and of the system dS D q=T . But
q is the sum of qA and qB , which gives dS D dSA C dSB . Thus, the entropy changes
are additive, so that entropy must be extensive: S =SA +SB .8
How can we evaluate the entropy of a particular equilibrium state of the system? We
must assign an arbitrary value to one state and then evaluate theR entropy change along a
reversible path from this state to the state of interest using S D .q=Tb /.
We may need to evaluate the entropy of a nonequilibrium state. To do this, we imagine
imposing hypothetical internal constraints that change the nonequilibrium state to a constrained equilibrium state with the same internal structure. Some examples of such internal
constraints were given in Sec. 2.4.4, and include rigid adiabatic partitions between phases of
different temperature and pressure, semipermeable membranes to prevent transfer of certain
species between adjacent phases, and inhibitors to prevent chemical reactions.
We assume that we can, in principle, impose or remove such constraints reversibly without heat, so there is no entropy change. If the nonequilibrium state includes macroscopic
internal motion, the imposition of internal constraints involves negative reversible work to
bring moving regions of the system to rest.9 If the system is nonuniform over its extent, the
8 The
argument is not quite complete, because we have not shown that when each subsystem has an entropy of
zero, so does the entire system. The zero of entropy will be discussed in Sec. 6.1.
9 This concept amounts to dening the entropy of a state with macroscopic internal motion to be the same as
the entropy of a state with the same internal structure but without the motion, i.e., the same state frozen in time.
By this denition, S for a purely mechanical process (Sec. 3.2.3) is zero.

4.5 I RREVERSIBLE P ROCESSES
124
internal constraints will partition it into practically-uniform regions whose entropy is addiR
tive. The entropy of the nonequilibrium state is then found from S D .q=Tb / using a
reversible path that changes the system from an equilibrium state of known entropy to the
constrained equilibrium state with the same entropy as the state of interest. This procedure
allows every possible state (at least conceptually) to have a denite value of S.
4.5 IRREVERSIBLE PROCESSES

We know that during a reversible process of a closed system, each innitesimal entropy
R
change dS is equal to q=Tb and the nite change S is equal to the integral .q=Tb /
but what can we say about dS and S for an irreversible process?
The derivation of this section will show that for an innitesimal irreversible change of
a closed system, dS is greater than q=Tb , and for an entire process S is greater than
R
.q=Tb /. That is, the equalities that apply to a reversible process are replaced, for an
irreversible process, by inequalities.
The derivation begins with irreversible processes that are adiabatic, and is then extended
to irreversible processes in general.
4.5.1 Irreversible adiabatic processes

Consider an arbitrary irreversible adiabatic process of a closed system starting with a particular initial state A. The nal state B depends on the path of this process. We wish to
investigate the sign of the entropy change SA!B . Our reasoning will depend on whether
or not there is work during the process.
If there is work along any innitesimal path element of the irreversible adiabatic process
(w 0), we know from experience that this work would be different if the work coordinate or coordinates were changing at a different rate, because energy dissipation from
internal friction would then be different. In the limit of innite slowness, an adiabatic process with initial state A and the same change of work coordinates would become reversible,
and the net work and nal internal energy would differ from those of the irreversible process. Because the nal state of the reversible adiabatic process is different from B, there is
no reversible adiabatic path with work between states A and B.
All states of a reversible process, including the initial and nal states, must be equilibrium states. There is therefore a conceptual difculty in considering reversible paths
between two states if either of these states are nonequilibrium states. In such a case
we will assume that the state has been replaced by a constrained equilibrium state of
the same entropy as described in Sec. 4.4.3.
If, on the other hand, there is no work along any innitesimal path element of the irreversible adiabatic process (wD0), the process is taking place at constant internal energy U
in an isolated system. A reversible limit cannot be reached without heat or work (page 65).
Thus any reversible adiabatic change from state A would require work, causing a change of
U and preventing the system from reaching state B by any reversible adiabatic path.
So regardless of whether or not an irreversible adiabatic process A!B involves work,
there is no reversible adiabatic path between A and B. The only reversible paths between

4.5 I RREVERSIBLE P ROCESSES
Tres
125
Tres
q 0
Tres
Tb
experimental
system
Tb
q
Tb
experimental
system
Figure 4.11 Supersystem including the experimental system, a Carnot engine (square
box), and a heat reservoir. The dashed rectangle indicates the boundary of the supersystem.
these states must be nonadiabatic. It follows that the entropy change SA!B , given by the
value of q=Tb integrated over a reversible path from A to B, cannot be zero.
Next we ask whether SA!B could be negative. In each innitesimal path element of
RB
the irreversible adiabatic process A!B, q is zero and the integral A .q=Tb / along the
path of this process is zero. Suppose the system completes a cycle by returning along a
different, reversible path from state B back to state A. The Clausius inequality (Eq. 4.4.3)
RA
tells us that in this case the integral B .q=Tb / along the reversible path cannot be positive.
But this integral for the reversible path is equal to SA!B , so SA!B cannot be negative.
We conclude that because the entropy change of the irreversible adiabatic process A!B
cannot be zero, and it cannot be negative, it must be positive.
In this derivation, the initial state A is arbitrary and the nal state B is reached by an
irreversible adiabatic process. If the two states are only innitesimally different, then the
change is innitesimal. Thus for an innitesimal change that is irreversible and adiabatic,
dS must be positive.
4.5.2 Irreversible processes in general

To treat an irreversible process of a closed system that is nonadiabatic, we proceed as follows. As in Sec. 4.4.1, we use a Carnot engine for heat transfer across the boundary of
the experimental system. We move the boundary of the supersystem of Fig. 4.8 so that the
supersystem now includes the experimental system, the Carnot engine, and a heat reservoir
of constant temperature Tres , as depicted in Fig. 4.11.
During an irreversible change of the experimental system, the Carnot engine undergoes
many innitesimal cycles. During each cycle, the Carnot engine exchanges heat q 0 at
temperature Tres with the heat reservoir and heat q at temperature Tb with the experimental
system, as indicated in the gure. We use the sign convention that q 0 is positive if heat is
transferred to the Carnot engine, and q is positive if heat is transferred to the experimental
system, in the directions of the arrows in the gure.
The supersystem exchanges work, but not heat, with its surroundings. During one
innitesimal cycle of the Carnot engine, the net entropy change of the Carnot engine is

4.6 A PPLICATIONS
126
zero, the entropy change of the experimental system is dS, the heat transferred between the
Carnot engine and the experimental system is q, and the heat transferred between the heat
reservoir and the Carnot engine is given by q 0 D Tres q=Tb (Eq. 4.4.1). The heat transfer
between the heat reservoir and Carnot engine is reversible, so the entropy change of the heat
reservoir is
q 0
q
dSres D
D
(4.5.1)
Tres
Tb
The entropy change of the supersystem is the sum of the entropy changes of its parts:
dSss D dS C dSres D dS
q
Tb
(4.5.2)
The process within the supersystem is adiabatic and includes an irreversible change within
the experimental system, so according to the conclusions of Sec. 4.5.1, dSss is positive.
Equation 4.5.2 then shows that dS, the innitesimal entropy change during the irreversible
change of the experimental system, must be greater than q=Tb :
dS >
q
Tb
(4.5.3)
(irreversible change, closed system)
This relation includes the case of an irreversible adiabatic change, because it shows that if
q is zero, dS is greater than zero.
By integrating both sides of Eq. 4.5.3 between the initial and nal states of the irreversible process, we obtain a relation for the nite entropy change corresponding to many
innitesimal cycles of the Carnot engine:
Z
q
(4.5.4)
S >
Tb
(irreversible process, closed system)
4.6 APPLICATIONS
The lengthy derivation in Secs. 4.34.5 is based on the KelvinPlanck statement describing
the impossibility of converting completely into work the energy transferred into the system by heat from a single heat reservoir. The derivation has now given us all parts of the
mathematical statement of the second law shown in the box on page 103. The mathematical statement includes an equality, dS D q=Tb , that applies to an innitesimal reversible
change, and an inequality, dS > q=Tb , that applies to an innitesimal irreversible change.
It is convenient to combine the equality and inequality in a single relation that is a general
mathematical statement of the second law:
dS
q
Tb
(4.6.1)
( irrev ,
rev
closed system)
The inequality refers to an irreversible change and the equality to a reversible change, as
indicated by the notation irrev in the conditions of validity. The integrated form of this
rev

4.6 A PPLICATIONS
127
relation is
S
q
Tb
(4.6.2)
( irrev ,
rev
closed system)
During a reversible process, the states are equilibrium states and the temperature is
usually uniform throughout the system. The only exception is if the system happens to have
internal adiabatic partitions that allow phases of different temperatures in an equilibrium
state. As mentioned in the footnote on page 103, when the process is reversible and the
temperature is uniform, we can replace dS D q=Tb by dS D q=T .
The rest of Sec. 4.6 will apply Eqs. 4.6.1 and 4.6.2 to various reversible and irreversible
processes.
4.6.1 Reversible heating

The denition of the heat capacity C of a closed system is given by Eq. 3.1.9 on page 62:
def
C D q= dT . For reversible heating or cooling of a homogeneous phase, q is equal to

T dS and we can write
Z T2
C
S D
dT
(4.6.3)
T1 T
where C should be replaced by CV if the volume is constant, or by Cp if the pressure is
constant (Sec. 3.1.5). If the heat capacity has a constant value over the temperature range
from T1 to T2 , the equation becomes
S D C ln
T2
T1
(4.6.4)
Heating increases the entropy, and cooling decreases it.
4.6.2 Reversible expansion of an ideal gas

When the volume of an ideal gas, or of any other uid, is changed reversibly and adiabatically, there is of course no entropy change.
When the volume of an ideal gas is changed reversibly and isothermally, there is expansion work given by w D nRT ln.V2 =V1 / (Eq. 3.5.1). Since the internal energy of an ideal
gas is constant at constant temperature, there must be heat of equal magnitude and opposite
sign: q D nRT ln.V2 =V1 /. The entropy change is therefore
S D nR ln
V2
V1
(4.6.5)
(reversible isothermal volume
change of an ideal gas)
Isothermal expansion increases the entropy, and isothermal compression decreases it.
Since the change of a state function depends only on the initial and nal states, Eq. 4.6.5
gives a valid expression for S of an ideal gas under the less stringent condition T2 D T1 ; it
is not necessary for the intermediate states to be equilibrium states of the same temperature.

4.6 A PPLICATIONS
128
4.6.3 Spontaneous changes in an isolated system

An isolated system is one that exchanges no matter or energy with its surroundings. Any
change of state of an isolated system that actually occurs is spontaneous, and arises solely
from conditions within the system, uninuenced by changes in the surroundingsthe process occurs by itself, of its own accord. The initial state and the intermediate states of the
process must be nonequilibrium states, because by denition an equilibrium state would not
change over time in the isolated system.
Unless the spontaneous change is purely mechanical, it is irreversible. According to
the second law, during an innitesimal change that is irreversible and adiabatic, the entropy
increases. For the isolated system, we can therefore write
dS > 0
(4.6.6)
(irreversible change, isolated system)
In later chapters, the inequality of Eq. 4.6.6 will turn out to be one of the most useful for
deriving conditions for spontaneity and equilibrium in chemical systems: The entropy of an
isolated system continuously increases during a spontaneous, irreversible process until it
reaches a maximum value at equilibrium.
If we treat the universe as an isolated system (although cosmology provides no assurance that this is a valid concept), we can say that as spontaneous changes occur in the
universe, its entropy continuously increases. Clausius summarized the rst and second laws
in a famous statement: Die Energie der Welt ist constant; die Entropie der Welt strebt einem
Maximum zu (the energy of the universe is constant; the entropy of the universe strives
toward a maximum).
4.6.4 Internal heat ow in an isolated system

Suppose the system is a solid body whose temperature initially is nonuniform. Provided
there are no internal adiabatic partitions, the initial state is a nonequilibrium state lacking
internal thermal equilibrium. If the system is surrounded by thermal insulation, and volume
changes are negligible, this is an isolated system. There will be a spontaneous, irreversible
internal redistribution of thermal energy that eventually brings the system to a nal equilibrium state of uniform temperature.
In order to be able to specify internal temperatures at any instant, we treat the system
as an assembly of phases, each having a uniform temperature that can vary with time. To
describe a region that has a continuous temperature gradient, we approximate the region
with a very large number of very small phases, each having a temperature innitesimally
different from its neighbors.
We use Greek letters to label the phases. The temperature of phase at any given
instant is T . We can treat each phase as a subsystem with a boundary across which there
can be energy transfer in the form of heat. Let q represent an innitesimal quantity of
heat transferred during an innitesimal interval of time to phase from phase . The heat
transfer, if any, is to the cooler from the warmer phase. If phases and are in thermal
contact and T is less than T , then q is positive; if the phases are in thermal contact
and T is greater than T , q is negative; and if neither of these conditions is satised,
q is zero.

4.6 A PPLICATIONS
129
To evaluate the entropy change, we need a reversible path from the initial to the nal
state. The P quantity of heat transferred to phase during an innitesimal time interval
net
is q D q . The entropy change of phase is the same as it would be for the
reversible transfer of this heat from a heat reservoir of temperature T : dS D q =T .
The entropy change of the entire system along the reversible path is found by summing over
all phases:
dS D
D
dS D
XX
>
X q
X X q
T

q
q
C
T
T
There is also the condition of quantitative energy transfer, q D

to rewrite Eq. 4.6.7 in the form
XX 1
1
dS D
q
T T
(4.6.7)
q , which we use
(4.6.8)
>
Consider an individual term of the sum on the right side of Eq. 4.6.8 that has a nonzero
value of q due to nite heat transfer between phases and . If T is less than T , then
both q and .1=T 1=T / are positive. If, on the other hand, T is greater than T ,
both q and .1=T 1=T / are negative. Thus each term of the sum is either zero or
positive, and as long as phases of different temperature are present, dS is positive.
This derivation shows that during a spontaneous thermal equilibration process in an isolated system, starting with any initial distribution of the internal temperatures, the entropy
continuously increases until the system reaches a state of thermal equilibrium with a single
uniform temperature throughout.10 The result agrees with Eq. 4.6.6.
4.6.5 Free expansion of a gas

Consider the free expansion of a gas shown in Fig. 3.8 on page 79. The system is the gas.
Assume that the vessel walls are rigid and adiabatic, so that the system is isolated. When the
stopcock between the two vessels is opened, the gas expands irreversibly into the vacuum
without heat or work and at constant internal energy. To carry out the same change of state
reversibly, we conne the gas at its initial volume and temperature in a cylinder-and-piston
device and use the piston to expand the gas adiabatically with negative work. Positive
heat is then needed to return the internal energy reversibly to its initial value. Because the
reversible path has positive heat, the entropy change is positive.
This is an example of an irreversible process in an isolated system for which a reversible
path between the initial and nal states has both heat and work.
4.6.6 Adiabatic process with work

In general (page 95), an adiabatic process with a given initial equilibrium state and a given
change of a work coordinate has the least positive or most negative work in the reversible
10 Leff,
in Ref. [98], obtains the same result by a more complicated derivation.

4.7 S UMMARY
130
limit. Consider an irreversible adiabatic process with work wirr . The same change of state
can be accomplished reversibly by the following two steps: (1) a reversible adiabatic change
of the work coordinate with work wrev , followed by (2) reversible transfer of heat qrev with
no further change of the work coordinate. Since wrev is algebraically less than wirr , qrev
must be positive in order to make U the same in the irreversible and reversible paths.
The positive heat increases the entropy along the reversible path, and consequently the
irreversible adiabatic process has a positive entropy change. This conclusion agrees with
the second-law inequality of Eq. 4.6.1.
4.7 SUMMARY
Some of the important terms and denitions discussed in this chapter are as follows.
Any conceivable process is either spontaneous, reversible, or impossible.
A reversible process proceeds by a continuous sequence of equilibrium states.
A spontaneous process is one that proceeds naturally at a nite rate.
An irreversible process is a spontaneous process whose reverse is impossible.
A purely mechanical process is an idealized process without temperature gradients,
and without friction or other dissipative effects, that is spontaneous in either direction.
This kind of process will be ignored in the remaining chapters of this book.
Except for a purely mechanical process, the terms spontaneous and irreversible are
equivalent.
The derivation of the mathematical statement of the second law shows that during a
reversible process of a closed system, the innitesimal quantity q=Tb equals the innitesimal change of a state function called the entropy, S. Here q is heat transferred at the
boundary where the temperature is Tb .
In each innitesimal path element of a process of a closed system, dS is equal to q=Tb
if the process is reversible, and is greater than q=Tb if the process is irreversible, as summarized by the relation dS q=Tb .
Consider two particular equilibrium states 1 and 2 of a closed system. The system can
change from state 1 to state 2 by either a reversible process, with S equal to the integral
R
R
.q=Tb /, or an irreversible process, with S greater than .q=Tb /. It is important to
keep in mind the point made by Fig. 4.12 on the next page: because S is a state function, it
is the value of the integral that is different in the two cases, and not the value of S.
The second law establishes no general relation between entropy changes and heat in an
open system, or for an impossible process. The entropy of an open system may increase or
decrease depending on whether matter enters or leaves. It is possible to imagine different
impossible processes in which dS is less than, equal to, and greater than q=Tb .
4.8 THE STATISTICAL INTERPRETATION OF ENTROPY

Because entropy is such an important state function, it is natural to seek a description of its
meaning on the microscopic level.
Entropy is sometimes said to be a measure of disorder. According to this idea, the
entropy increases whenever a closed system becomes more disordered on a microscopic

4.8 T HE S TATISTICAL I NTERPRETATION OF E NTROPY
131
R reversible
q=Tb D S
state 1
state 2
R irreversible
q=Tb < S
Figure 4.12 Reversible and irreversible paths between the same initial and nal equilibrium states of a closed system. The value of S is the same for both paths, but the
R
values of the integral .q=Tb / are different.
scale. This description of entropy as a measure of disorder is highly misleading. It does not
explain why entropy is increased by reversible heating at constant volume or pressure, or
why it increases during the reversible isothermal expansion of an ideal gas. Nor does it seem
to agree with the freezing of a supercooled liquid or the formation of crystalline solute in a
supersaturated solution; these processes can take place spontaneously in an isolated system,
yet are accompanied by an apparent decrease of disorder.
Thus we should not interpret entropy as a measure of disorder. We must look elsewhere
for a satisfactory microscopic interpretation of entropy.
A rigorous interpretation is provided by the discipline of statistical mechanics, which
derives a precise expression for entropy based on the behavior of macroscopic amounts of
microscopic particles. Suppose we focus our attention on a particular macroscopic equilibrium state. Over a period of time, while the system is in this equilibrium state, the system
at each instant is in a microstate, or stationary quantum state, with a denite energy. The
microstate is one that is accessible to the systemthat is, one whose wave function is compatible with the systems volume and with any other conditions and constraints imposed
on the system. The system, while in the equilibrium state, continually jumps from one accessible microstate to another, and the macroscopic state functions described by classical
thermodynamics are time averages of these microstates.
The fundamental assumption of statistical mechanics is that accessible microstates of
equal energy are equally probable, so that the system while in an equilibrium state spends an
equal fraction of its time in each such microstate. The statistical entropy of the equilibrium
state then turns out to be given by the equation
Sstat D k ln W C C
(4.8.1)
where k is the Boltzmann constant k D R=NA , W is the number of accessible microstates,

and C is a constant.
In the case of an equilibrium state of a perfectly-isolated system of constant internal
energy U , the accessible microstates are the ones that are compatible with the constraints
and whose energies all have the same value, equal to the value of U .
It is more realistic to treat an equilibrium state with the assumption the system is in thermal equilibrium with an external constant-temperature heat reservoir. The internal energy
then uctuates over time with extremely small deviations from the average value U , and the
accessible microstates are the ones with energies close to this average value. In the language

4.8 T HE S TATISTICAL I NTERPRETATION OF E NTROPY
132
of statistical mechanics, the results for an isolated system are derived with a microcanonical
ensemble, and for a system of constant temperature with a canonical ensemble.
A change Sstat of the statistical entropy function given by Eq. 4.8.1 is the same as the
change S of the macroscopic second-law entropy, because the derivation of Eq. 4.8.1 is
based on the macroscopic relation dSstat D q=T D .dU w/=T with dU and w given
by statistical theory. If the integration constant C is set equal to zero, Sstat becomes the
third-law entropy S to be described in Chap. 6.
Equation 4.8.1 shows that a reversible process in which entropy increases is accompanied by an increase in the number of accessible microstates of equal, or nearly equal,
internal energies. This interpretation of entropy increase has been described as the spreading and sharing of energy11 and as the dispersal of energy.12 It has even been proposed
that entropy should be thought of as a spreading function with its symbol S suggesting
spreading.13,14
11 Ref.
[99].
[95].
13 Ref. [96].
14 The symbol S for entropy seems originally to have been an arbitrary choice by Clausius; see Ref. [82].
12 Ref.

P ROBLEMS
133
PROBLEMS
in Appendix I.
4.1 Explain why an electric refrigerator, which transfers energy by means of heat from the cold
food storage compartment to the warmer air in the room, is not an impossible Clausius device.
4.2 A system consisting of a xed amount of an ideal gas is maintained in thermal equilibrium
with a heat reservoir at temperature T . The system is subjected to the following isothermal
cycle:
1. The gas, initially in an equilibrium state with volume V0 , is allowed to expand into a
vacuum and reach a new equilibrium state of volume V 0 .
2. The gas is reversibly compressed from V 0 to V0 .
H
H
For this cycle, nd expressions or values for w, q=T , and dS.
4.3 In an irreversible isothermal process of a closed system:
(a) Is it possible for S to be negative?
(b) Is it possible for S to be less than q=T ?
4.4 Suppose you have two blocks of copper, each of heat capacity CV D 200:0 J K 1 . Initially one
block has a uniform temperature of 300:00 K and the other 310:00 K. Calculate the entropy
change that occurs when you place the two blocks in thermal contact with one another and
surround them with perfect thermal insulation. Is the sign of S consistent with the second
law? (Assume the process occurs at constant volume.)
4.5 Refer to the apparatus shown in Figs. 3.22 on page 98 and 3.25 on page 100 and described in
Probs. 3.3 and 3.8. For both systems, evaluate S for the process that results from opening the
R
stopcock. Also evaluate q=Text for both processes (for the apparatus in Fig. 3.25, assume
the vessels have adiabatic walls). Are your results consistent with the mathematical statement
of the second law?
water
Figure 4.13
4.6 Figure 4.13 shows the walls of a rigid thermally-insulated box (cross hatching). The system
is the contents of this box. In the box is a paddle wheel immersed in a container of water,
connected by a cord and pulley to a weight of mass m. The weight rests on a stop located
a distance h above the bottom of the box. Assume the heat capacity of the system, CV , is

P ROBLEMS
134
independent of temperature. Initially the system is in an equilibrium state at temperature T1 .

When the stop is removed, the weight irreversibly sinks to the bottom of the box, causing the
paddle wheel to rotate in the water. Eventually the system reaches a nal equilibrium state
with thermal equilibrium. Describe a reversible process with the same entropy change as this
irreversible process, and derive a formula for S in terms of m, h, CV , and T1 .
C HAPTER 5
T HERMODYNAMIC P OTENTIALS
This chapter begins with a discussion of mathematical properties of the total differential
of a dependent variable. Three extensive state functions with dimensions of energy are introduced: enthalpy, Helmholtz energy, and Gibbs energy. These functions, together with
internal energy, are called thermodynamic potentials.1 Some formal mathematical manipulations of the four thermodynamic potentials are described that lead to expressions for
heat capacities, surface work, and criteria for spontaneity in closed systems.
5.1 TOTAL DIFFERENTIAL OF A DEPENDENT VARIABLE

Recall from Sec. 2.4.1 that the state of the system at each instant is dened by a certain
minimum number of state functions, the independent variables. State functions not treated
as independent variables are dependent variables. Innitesimal changes in any of the independent variables will, in general, cause an innitesimal change in each dependent variable.
A dependent variable is a function of the independent variables. The total differential of a dependent variable is an expression for the innitesimal change of the variable in
terms of the innitesimal changes of the independent variables. As explained in Sec. F.2
of Appendix F, the expression can be written as a sum of terms, one for each independent
variable. Each term is the product of a partial derivative with respect to one of the independent variables and the innitesimal change of that independent variable. For example, if the
system has two independent variables, and we take these to be T and V , the expression for
the total differential of the pressure is

@p
@p
dp D
dT C
dV
(5.1.1)
@T V
@V T
Thus, in the case of a xed amount of an ideal gas with pressure given by p D nRT =V ,
the total differential of the pressure can be written
dp D
nR
dT
V
1 The
nRT
dV
V2
(5.1.2)
term thermodynamic potential should not be confused with the chemical potential, , to be introduced on
page 137.
135
CHAPTER 5 THERMODYNAMIC POTENTIALS

5.2 T OTAL D IFFERENTIAL OF THE I NTERNAL E NERGY
136
5.2 TOTAL DIFFERENTIAL OF THE INTERNAL ENERGY

For a closed system undergoing processes in which the only kind of work is expansion work,
the rst law becomes dU D q C w D q pb dV . Since it will often be useful to make
a distinction between expansion work and other kinds of work, this book will sometimes
write the rst law in the form
dU D q
pb dV C w 0
(5.2.1)
(closed system)
where w 0 is nonexpansion workthat is, any thermodynamic work that is not expansion
work.
Consider a closed system of one chemical component (e.g., a pure substance) in a single
homogeneous phase. The only kind of work is expansion work, with V as the work variable.
This kind of system has two independent variables (Sec. 2.4.3). During a reversible process
in this system, the heat is q D T dS, the work is w D p dV , and an innitesimal
internal energy change is given by
dU D T dS
p dV
(5.2.2)
(closed system, C D1,
P D1, w 0 D0)
In the conditions of validity shown next to this equation, C D1 means there is one component (C is the number of components) and P D1 means there is one phase (P is the number
of phases).
The appearance of the intensive variables T and p in Eq. 5.2.2 implies, of course, that
the temperature and pressure are uniform throughout the system during the process. If they
were not uniform, the phase would not be homogeneous and there would be more than two
independent variables. The temperature and pressure are strictly uniform only if the process
is reversible; it is not necessary to include reversible as one of the conditions of validity.
A real process approaches a reversible process in the limit of innite slowness. For all
practical purposes, therefore, we may apply Eq. 5.2.2 to a process obeying the conditions
of validity and taking place so slowly that the temperature and pressure remain essentially
uniformthat is, for a process in which the system stays very close to thermal and mechanical equilibrium.
Because the system under consideration has two independent variables, Eq. 5.2.2 is an
expression for the total differential of U with S and V as the independent variables. In
general, an expression for the differential dX of a state function X is a total differential if
1. it is a valid expression for dX, consistent with the physical nature of the system and
any conditions and constraints;
2. it is a sum with the same number of terms as the number of independent variables;
3. each term of the sum is a function of state functions multiplied by the differential of
one of the independent variables.
Note that the work coordinate of any kind of dissipative workwork without a reversible limitcannot appear in the expression for a total differential, because it is not a
state function (Sec. 3.10).

5.2 T OTAL D IFFERENTIAL OF THE I NTERNAL E NERGY
137
As explained in Appendix F, we may identify the coefcient of each term in an expression for the total differential of a state function as a partial derivative of the function. We
identify the coefcients on the right side of Eq. 5.2.2 as follows:
@U
@U
pD
(5.2.3)
T D
@S V
@V S
Now let us consider some of the ways a system might have more than two independent
variables. Suppose the system has one phase and one substance, with expansion work
only, and is open so that the amount n of the substance can vary. Such a system has three
independent variables. Let us write the formal expression for the total differential of U with
S, V , and n as the three independent variables:
@U
@U
@U
dS C
dV C
dn
dU D
(5.2.4)
@S V;n
@V S;n
@n S;V
(pure substance,
P D1, w 0 D0)
We have seen above that if the system is closed, the partial derivatives are .@U=@S/V D T
and .@U=@V /S D p. Since both of these partial derivatives are for a closed system in
which n is constant, they are the same as the rst two partial derivatives on the right side of
Eq. 5.2.4.
The quantity given by the third partial derivative, .@U=@n/S;V , is represented by the
symbol (mu). This quantity is an intensive state function called the chemical potential.
With these substitutions, Eq. 5.2.4 becomes
dU D T dS
p dV C
dn
(5.2.5)
(pure substance,
P D1, w 0 D0)
and this is a valid expression for the total differential of U under the given conditions.
If a single-phase system contains a mixture of s different substances, and the amount of
each substance can be varied independently, there are 2 C s independent variables and the
total differential of U can be written
dU D T dS
p dV C
s
X
dni
iD1
(5.2.6)
(P D1, w 0 D0)
The coefcient i is the chemical potential of substance i. We identify it as the partial

derivative .@U=@ni /S;V;nj i .
Suppose that in addition to expansion work, other kinds of reversible work are possible.
Each work coordinate adds an additional independent variable. Thus, for a closed system
of one component in one phase, with reversible nonexpansion work given by w 0 D Y dX ,
the total differential of U becomes
dU D T dS
p dV C Y dX
(5.2.7)
(closed system,
C D1, P D1)

5.3 E NTHALPY, H ELMHOLTZ E NERGY, AND G IBBS E NERGY
138
5.3 ENTHALPY, HELMHOLTZ ENERGY, AND GIBBS ENERGY

For the moment we shall conne our attention to closed systems with one component in one
phase. The total differential of the internal energy in such a system is given by Eq. 5.2.2:
dU D T dS p dV . The independent variables in this equation, S and V , are called the
natural variables of U .
In the laboratory, entropy and volume may not be the most convenient variables to
measure and control. Entropy is especially inconvenient, as its value cannot be measured
directly. The way to change the independent variables is to make Legendre transforms, as
explained in Sec. F.4 in Appendix F.
A Legendre transform of a dependent variable is made by subtracting one or more products of conjugate variables. In the total differential dU D T dS p dV , T and S are
conjugates (that is, they comprise a conjugate pair), and p and V are conjugates. Thus
the products that can be subtracted from U are either T S or pV , or both. Three Legendre
transforms of the internal energy are possible, dened as follows:
def
Enthalpy
H D U C pV
(5.3.1)
Helmholtz energy
A D U
def
TS
(5.3.2)
def
G D U
T S C pV D H
Gibbs energy
TS
(5.3.3)
These denitions are used whether or not the system has only two independent variables.
The enthalpy, Helmholtz energy, and Gibbs energy are important functions used extensively in thermodynamics. They are state functions (because the quantities used to dene
them are state functions) and are extensive (because U , S , and V are extensive). If temperature or pressure are not uniform in the system, we can apply the denitions to constituent
phases, or to subsystems small enough to be essentially uniform, and sum over the phases
or subsystems.
Alternative names for the Helmholtz energy are Helmholtz function, Helmholtz free
energy, and work function. Alternative names for the Gibbs energy are Gibbs function
and Gibbs free energy. Both the Helmholtz energy and Gibbs energy have been called
simply free energy, and the symbol F has been used for both. The nomenclature in
this book follows the recommendations of the IUPAC Green Book (Ref. [36]).
Expressions for innitesimal changes of H , A, and G are obtained by applying the rules
of differentiation to their dening equations:
dH D dU C p dV C V dp
(5.3.4)
dA D dU
S dT
(5.3.5)
S dT C p dV C V dp
(5.3.6)
dG D dU
T dS
T dS
These expressions for dH , dA, and dG are general expressions for any system or phase
with uniform T and p. They are not total differentials of H , A, and G, as the variables in
the differentials in each expression are not independent.
A useful property of the enthalpy in a closed system can be found by replacing dU in
Eq. 5.3.4 by the rst law expression q p dV C w 0 , to obtain dH D q C V dp C w 0 .

5.3 E NTHALPY, H ELMHOLTZ E NERGY, AND G IBBS E NERGY
139
BIOGRAPHICAL SKETCH
Josiah Willard Gibbs (18391903)

monumental paper in the same journalin two
parts (1876 and 1878) and over 300 pages in
length!entitled simply On the Equilibrium
of Heterogeneous Substances.d This third paper used an analytical rather than geometrical approach. From the rst and second laws
of thermodynamics, it derived the conditions
needed for equilibrium in the general case of
a multiphase, multicomponent system. It introduced the state functions now known as enthalpy, Helmholtz energy,e Gibbs energy, and
Willard Gibbss brilliant and rigorous formu- chemical potential. Included in the paper was
lation of the theoretical basis of classical ther- the exposition of the Gibbs phase rule.
modynamics was essential for further developThe only public comment Gibbs ever made
ment of the subject.
on his thermodynamic papers was in a letter of
Gibbs was born in New Haven, Connecti- 1881 accepting membership in the American
cut, and lived there all his life. His father was Academy of Arts and Sciences:f
a professor in the Yale Divinity School. Gibbs
The leading idea which I followed in my paper
was Professor of Mathematical Physics at Yale
on the Equilibrium of Heterogeneous Substances
College from 1871 until his death.
was to develop the r les of energy and entropy in
o
Gibbs never married. In demeanor he was
the theory of thermo-dynamic equilibrium. By
means of these quantities the general condition
serene, kindly, reserved, and self-effacing. A
of equilibrium is easily expressed, and by applybiographer wrote:a
Gibbs attitude toward his discoveries is also illuminating as to his character. . . . he made no effort to sell his discoveries (aside from the usual
distribution of reprints of his papers) or to popularize the results. He was so condent of their
rightness and ability to stand on their own feet
that he was entirely content to let their value and
importance be discovered by others. The fact
that he had made a discovery was to him an irrelevant matter; the important thing was the truth
established.
In 1873, when he was 34, the rst two of

Gibbss remarkable papers on theoretical thermodynamics appeared in an obscure journal,
Transactions of the Connecticut Academy.b,c
These papers explored relations among state
functions using two- and three-dimensional
geometrical constructions.
James Clerk Maxwell promoted Gibbss
ideas in England, and made a small plaster
model of the three-dimensional S V U surface for H2 O which he sent to Gibbs.
The two papers of 1873 were followed by a
ing this to various cases we are led at once to

the special conditions which characterize them.
We thus obtain the consequences resulting from
the fundamental principles of thermo-dynamics
(which are implied in the denitions of energy
and entropy) by a process which seems more
simple, and which lends itself more readily to the
solution of problems, than the usual method, in
which the several parts of a cyclic operation are
explicitly and separately considered. Although
my results were in a large measure such as had
previously been demonstrated by other methods,
yet, as I readily obtained those which were to me
before unknown, I was conrmed in my belief in
the suitableness of the method adopted.
While Gibbs is best known for his papers on

classical thermodynamics, he also published
works on vector analysis and statistical mechanics.
Gibbs had a visit about 1898 from a young
Gilbert Lewis. He told Lewis that he was
rather lonely at Yale, where few others were
actively interested in his work.g
a Ref.
[169], page 83. b Ref. [62]. c Ref. [63]. d Ref. [64]. e Hermann von Helmholtz, a German
physiologist and physicist, introduced the term free energy for this quantity in 1882. f Ref. [169], page
89. g Ref. [129].

5.4 C LOSED S YSTEMS
140
Thus, in a process at constant pressure (dp D 0) with expansion work only (w 0 D0), we
have
dH D q
(5.3.7)
(closed system, constant p,
w 0 D0)
The enthalpy change under these conditions is equal to the heat. The integrated form of this
R
R
relation is dH D q, or
H D q
(5.3.8)
(closed system, constant p,
w 0 D0)
Equation 5.3.7 is analogous to the following relation involving the internal energy, obtained from the rst law:
dU D q
(5.3.9)
(closed system, constant V ,
w 0 D0)
That is, in a process at constant volume with expansion work only, the internal energy
change is equal to the heat.
5.4 CLOSED SYSTEMS

In order to nd expressions for the total differentials of H , A, and G in a closed system
with one component in one phase, we must replace dU in Eqs. 5.3.45.3.6 with
dU D T dS
p dV
(5.4.1)
dH D T dS C V dp
(5.4.2)
p dV
(5.4.3)
dG D
S dT C V dp
(5.4.4)
to obtain
dA D
S dT
Equations 5.4.15.4.4 are sometimes called the Gibbs equations. They are expressions
for the total differentials of the thermodynamic potentials U , H , A, and G in closed systems of one component in one phase with expansion work only. Each equation shows how
the dependent variable on the left side varies as a function of changes in two independent
variables (the natural variables of the dependent variable) on the right side.
By identifying the coefcients on the right side of Eqs. 5.4.15.4.4, we obtain the following relations (which again are valid for a closed system of one component in one phase
with expansion work only):
@U
from Eq. 5.4.1:
DT
(5.4.5)
@S V
@U
D p
(5.4.6)
@V S

5.5 O PEN S YSTEMS
from Eq. 5.4.2:
from Eq. 5.4.3:
from Eq. 5.4.4:
141
@H
@S
@H
@p
S
@A
@T V
@A
@V T
@G
@T p
@G
@p T
DT
(5.4.7)
DV
(5.4.8)
(5.4.9)
(5.4.10)
(5.4.11)
DV
(5.4.12)
This book now uses for the rst time an extremely useful mathematical tool called the
reciprocity relation of a total differential (Sec. F.2). Suppose the independent variables are
x and y and the total differential of a dependent state function f is given by
df D a dx C b dy
(5.4.13)
where a and b are functions of x and y. Then the reciprocity relation is

@a
@b
D
@y x
@x y
(5.4.14)
The reciprocity relations obtained from the Gibbs equations (Eqs. 5.4.15.4.4) are called
Maxwell relations (again valid for a closed system with C D1, P D1, and w 0 D0):
@p
@T
D
from Eq. 5.4.1:
(5.4.15)
@V S
@S V

@T
@V
from Eq. 5.4.2:
D
(5.4.16)
@p S
@S p

@S
@p
(5.4.17)
from Eq. 5.4.3:
D
@V
@T V
T
@S
@V
from Eq. 5.4.4:
D
(5.4.18)
@p T
@T p
5.5 OPEN SYSTEMS

An open system of one substance in one phase, with expansion work only, has three independent variables. The total differential of U is given by Eq. 5.2.5:
dU D T dS
p dV C
dn
(5.5.1)
In this open system the natural variables of U are S, V , and n. Substituting this expression
for dU into the expressions for dH , dA, and dG given by Eqs. 5.3.45.3.6, we obtain the
following total differentials:
dH D T dS C V dp C
dn
(5.5.2)

5.5 O PEN S YSTEMS
142
dA D
S dT
p dV C
dn
(5.5.3)
dG D
S dT C V dp C
dn
(5.5.4)
Note that these are the same as the four Gibbs equations (Eqs. 5.4.15.4.4) with the addition
of a term dn to allow for a change in the amount of substance.
Identication of the coefcient of the last term on the right side of each of these equations shows that the chemical potential can be equated to four different partial derivatives:
@U
@H
@A
@G
D
D
D
(5.5.5)
D
@n S;V
@n S;p
@n T;V
@n T;p
All four of these partial derivatives must have the same value for a given state of the system;
the value, of course, depends on what that state is.
The last partial derivative on the right side of Eq. 5.5.5, .@G=@n/T;p , is especially interesting because it is the rate at which the Gibbs energy increases with the amount of substance added to a system whose intensive properties remain constant. Thus, is revealed
to be equal to Gm , the molar Gibbs energy of the substance.
Suppose the system contains several substances or species in a single phase (a mixture)
whose amounts can be varied independently. We again assume the only work is expansion
work. Then, making use of Eq. 5.2.6, we nd the total differentials of the thermodynamic
potentials are given by
X
dU D T dS p dV C
(5.5.6)
i dni
i
dH D T dS C V dp C
dA D
S dT
p dV C
dG D
S dT C V dp C
dni
(5.5.7)
dni
(5.5.8)
dni
(5.5.9)
X
i
X
i
The independent variables on the right side of each of these equations are the natural variables of the corresponding thermodynamic potential. Section F.4 shows that all of the information contained in an algebraic expression for a state function is preserved in a Legendre
transform of the function. What this means for the thermodynamic potentials is that an
expression for any one of them, as a function of its natural variables, can be converted to
an expression for each of the other thermodynamic potentials as a function of its natural
variables.
Willard Gibbs, after whom the Gibbs energy is named, called Eqs. 5.5.65.5.9 the fundamental equations of thermodynamics, because from any single one of them not only the
other thermodynamic potentials but also all thermal, mechanical, and chemical properties
of the system can be deduced.2 Problem 5.4 illustrates this useful application of the total
differential of a thermodynamic potential.
2 Ref.
[64], p. 86.

5.6 E XPRESSIONS FOR H EAT C APACITY
143
In Eqs. 5.5.65.5.9, the coefcient i is the chemical potential of species i. The equations show that i can be equated to four different partial derivatives, similar to the equalities shown in Eq. 5.5.5 for a pure substance:
@H
@A
@G
@U
D
D
D
D
(5.5.10)
i
@ni S;V;nj i
@ni S;p;nj i
@ni T;V;nj i
@ni T;p;nj i
The partial derivative .@G=@ni /T;P;nj i is called the partial molar Gibbs energy of species
i, another name for the chemical potential as will be discussed in Sec. 9.2.6.
5.6 EXPRESSIONS FOR HEAT CAPACITY

As explained in Sec. 3.1.5, the heat capacity of a closed system is dened as the ratio of
an innitesimal quantity of heat transferred across the boundary under specied conditions
def
and the resulting innitesimal temperature change: heat capacity D q= dT . The heat
capacities of isochoric (constant volume) and isobaric (constant pressure) processes are of
particular interest.
The heat capacity at constant volume, CV , is the ratio q= dT for a process in a closed
constant-volume system with no nonexpansion workthat is, no work at all. The rst law
shows that under these conditions the internal energy change equals the heat: dU D q
(Eq. 5.3.9). We can replace q by dU and write CV as a partial derivative:
@U
CV D
(5.6.1)
@T V
(closed system)
If the closed system has more than two independent variables, additional conditions
are needed to dene CV unambiguously. For instance, if the system is a gas mixture
in which reaction can occur, we might specify that the system remains in reaction
equilibrium as T changes at constant V .
Equation 5.6.1 does not require the condition w 0 D0, because all quantities appearing in the equation are state functions whose relations to one another are xed by
the nature of the system and not by the path. Thus, if heat transfer into the system
at constant V causes U to increase at a certain rate with respect to T , and this rate is
dened as CV , the performance of electrical work on the system at constant V will
cause the same rate of increase of U with respect to T and can equally well be used to
evaluate CV .
Note that CV is a state function whose value depends on the state of the systemthat
is, on T , V , and any additional independent variables. CV is an extensive property: the
combination of two identical phases has twice the value of CV that one of the phases has
by itself.
For a phase containing a pure substance, the molar heat capacity at constant volume
def
is dened by CV;m D CV =n. CV;m is an intensive property.
If the system is an ideal gas, its internal energy depends only on T , regardless of whether
V is constant, and Eq. 5.6.1 can be simplied to
CV D
dU
dT
(5.6.2)
(closed system, ideal gas)

5.7 S URFACE W ORK
144
Fext
Figure 5.1 Device to measure the surface tension of a liquid lm. The lm is
stretched between a bent wire and a sliding rod.
Thus the internal energy change of an ideal gas is given by dU D CV dT , as mentioned

earlier in Sec. 3.5.3.
The heat capacity at constant pressure, Cp , is the ratio q= dT for a process in a
closed system with a constant, uniform pressure and with expansion work only. Under
these conditions, the heat q is equal to the enthalpy change dH (Eq. 5.3.7), and we obtain
a relation analogous to Eq. 5.6.1:
@H
Cp D
(5.6.3)
@T p
(closed system)
Cp is an extensive state function. For a phase containing a pure substance, the molar heat
capacity at constant pressure is Cp;m D Cp =n, an intensive property.
Since the enthalpy of a xed amount of an ideal gas depends only on T (Prob. 5.1), we
can write a relation analogous to Eq. 5.6.2:
Cp D
dH
dT
(5.6.4)
(closed system, ideal gas)
5.7 SURFACE WORK

Sometimes we need more than the usual two independent variables to describe an equilibrium state of a closed system of one substance in one phase. This is the case when,
in addition to expansion work, another kind of work is possible. The total differential of
U is then given by dU D T dS p dV C Y dX (Eq. 5.2.7), where Y dX represents the
nonexpansion work w 0 .
A good example of this situation is surface work in a system in which surface area is
relevant to the description of the state.
A liquidgas interface behaves somewhat like a stretched membrane. The upper and
lower surfaces of the liquid lm in the device depicted in Fig. 5.1 exert a force F on the
sliding rod, tending to pull it in the direction that reduces the surface area. We can measure
the force by determining the opposing force Fext needed to prevent the rod from moving.
This force is found to be proportional to the length of the rod and independent of the surface
area. The force also depends on the temperature and pressure.
The surface tension or interfacial tension, , is the force exerted by an interfacial surface per unit length. The lm shown in Fig. 5.1 has two surfaces, so we have D F=2l
where l is the rod length.

5.8 C RITERIA FOR S PONTANEITY
145
To increase the surface area of the lm by a practically-reversible process, we slowly

pull the rod to the right in the Cx direction. The system is the liquid. The x component of
sys
the force exerted by the system on the surroundings at the moving boundary, Fx , is equal
sys
to F (F is positive and Fx is negative). The displacement of the rod results in surface
sys
work given by Eq. 3.1.2: w 0 D Fx dx D 2 l dx. The increase in surface area, dAs ,
is 2l dx, so the surface work is w 0 D dAs where is the work coefcient and As is the
work coordinate. Equation 5.2.7 becomes
dU D T dS
p dV C dAs
(5.7.1)
Substitution into Eq. 5.3.6 gives

dG D
S dT C V dp C dAs
(5.7.2)
which is the total differential of G with T , p, and As as the independent variables. Identifying the coefcient of the last term on the right side as a partial derivative, we nd the
following expression for the surface tension:
@G
D
(5.7.3)
@As T;p
That is, the surface tension is not only a force per unit length, but also a Gibbs energy per
unit area.
From Eq. 5.7.2, we obtain the reciprocity relation

@
@S
(5.7.4)
D
@T p;As
@As T;p
It is valid to replace the partial derivative on the left side by .@ =@T /p because is independent of As . Thus, the variation of surface tension with temperature tells us how the
entropy of the liquid varies with surface area.
5.8 CRITERIA FOR SPONTANEITY

In this section we combine the rst and second laws in order to derive some general relations
for changes during a reversible or irreversible process of a closed system. The temperature
and pressure will be assumed to be practically uniform during the process, even if the process is irreversible. For example, the volume might be changing at a nite rate but very
slowly, or there might be a spontaneous homogeneous reaction in a mixture of uniform
temperature and pressure.
The second law states that dS is equal to q=T if the process is reversible, and is greater
than q=T if the process is irreversible:
dS
q=T
(5.8.1)
( irrev , closed system)
rev
or
q T dS
(5.8.2)
( irrev ,
rev
closed system)

146
The inequalities in these relations refer to an irreversible process and the equalities to a
reversible process, as indicated by the notation irrev .
rev
When we substitute q from Eq. 5.8.2 into the rst law in the form dU D q p dV C
w 0 , where w 0 is nonexpansion work, we obtain the relation
dU T dS
p dV C w 0
(5.8.3)
rev
We substitute this relation for dU into the differentials of enthalpy, Helmholtz energy, and
Gibbs energy given by Eqs. 5.3.45.3.6 to obtain three more relations:
dH T dS C V dp C w 0
p dV C w 0
dA
S dT
dG
S dT C V dp C w 0
(5.8.4)
rev
(5.8.5)
( irrev ,
rev
closed system)
(5.8.6)
( irrev ,
rev
closed system)
The last two of these relations provide valuable criteria for spontaneity under common
laboratory conditions. Equation 5.8.5 shows that during a spontaneous irreversible change
at constant temperature and volume, dA is less than w 0 . If the only work is expansion
work (i.e., w 0 is zero), the Helmholtz energy decreases during a spontaneous process at
constant T and V and has its minimum value when the system reaches an equilibrium state.
Equation 5.8.6 is especially useful. From it, we can conclude the following:
Reversible nonexpansion work at constant T and p is equal to the Gibbs energy
change. For example, if the system is a galvanic cell operated in the reversible limit
(Sec. 3.8.3) at constant T and p, the electrical work is given by wel, rev D dG. There
is an application of this relation in Sec. 14.3.1.
During a spontaneous process at constant T and p in a closed system with expansion work only, the Gibbs energy continuously decreases until the system reaches an
equilibrium state.
Ben-Amotz and Honig3 developed a rectication procedure that simplies the mathematical manipulation of inequalities. Following this procedure, we can write
dS D q=T C
(5.8.7)
where is an excess entropy function that is positive for an irreversible change and
zero for a reversible change (
0). Solving for q gives the expression q D
T dS T that, when substituted in the rst law expression dU D q p dV Cw 0 ,
produces
dU D T dS p dV C w 0 T
(5.8.8)
3 Refs.
[11] and [81].

147
The equality of this equation is equivalent to the combined equality and inequality of
Eq. 5.8.3. Then by substitution of this expression for dU into Eqs. 5.3.45.3.6, we
obtain equalities equivalent to Eqs. 5.8.45.8.6, for example
dG D
S dT C V dp C w 0
(5.8.9)
Equation 5.8.9 tells us that during a process at constant T and p, with expansion work
only (w 0 D0), dG has the same sign as T : negative for an irreversible change
and zero for a reversible change.

P ROBLEMS
148
PROBLEMS
in Appendix I.
5.1 Show that the enthalpy of a xed amount of an ideal gas depends only on the temperature.
5.2 From concepts in this chapter, show that the heat capacities CV and Cp of a xed amount of
an ideal gas are functions only of T .
5.3 During the reversible expansion of a xed amount of an ideal gas, each increment of heat is
given by the expression q D CV dT C .nRT =V / dV (Eq. 4.3.4).
(a) A necessary and sufcient condition for this expression to be an exact differential is that
the reciprocity relation must be satised for the independent variables T and V (see Appendix F). Apply this test to show that the expression is not an exact differential, and that
heat therefore is not a state function.
(b) By the same method, show that the entropy increment during the reversible expansion,
given by the expression dS D q=T , is an exact differential, so that entropy is a state
function.
5.4 This problem illustrates how an expression for one of the thermodynamic potentials as a function of its natural variables contains the information needed to obtain expressions for the other
thermodynamic potentials and many other state functions.
From statistical mechanical theory, a simple model for a hypothetical hard-sphere liquid
(spherical molecules of nite size without attractive intermolecular forces) gives the following
expression for the Helmholtz energy with its natural variables T , V , and n as the independent
variables:
V
A D nRT ln cT 3=2
b
nRT C na
n
Here a, b, and c are constants. Derive expressions for the following state functions of this
hypothetical liquid as functions of T , V , and n.
(a) The entropy, S
(b) The pressure, p
(c) The chemical potential,
(d) The internal energy, U
(e) The enthalpy, H
(f) The Gibbs energy, G
(g) The heat capacity at constant volume, CV
(h) The heat capacity at constant pressure, Cp (hint: use the expression for p to solve for V
as a function of T , p, and n; then use H D U C pV )
5.5 Figure 5.2 on the next page depicts a hypothetical liquid in equilibrium with its vapor. The
liquid and gas are conned in a cylinder by a piston. An electrical resistor is immersed in the
liquid. The system is the contents of the cylinder to the left of the piston (the liquid, gas, and
resistor). The initial state of the system is described by
V1 D 0:2200 m3
T1 D 300:0 K
p1 D 2:50
105 Pa
A constant current I D 0:5000 A is passed for 1600 s through the resistor, which has electric
resistance Rel D 50:00 . The piston moves slowly to the right against a constant external

P ROBLEMS
149
gas
liquid
Figure 5.2
pressure equal to the vapor pressure of the liquid, 2:50 105 Pa, and some of the liquid vaporizes. Assume that the process is adiabatic and that T and p remain uniform and constant. The
nal state is described by
V2 D 0:2400 m3
T2 D 300:0 K
p2 D 2:50
105 Pa
(a) Calculate q, w, U , and H .

(b) Is the process reversible? Explain.
(c) Devise a reversible process that accomplishes the same change of state, and use it to
calculate S.
(d) Compare q for the reversible process with H . Does your result agree with Eq. 5.3.8?
Table 5.1 Surface tension
of water at 1 bar a
t = C
=10
15
20
25
30
35
a Ref.
J cm
7:350
7:275
7:199
7:120
7:041
[161].
5.6 Use the data in Table 5.1 to evaluate .@S=@As /T;p at 25 C, which is the rate at which the
entropy changes with the area of the airwater interface at this temperature.
5.7 When an ordinary rubber band is hung from a clamp and stretched with constant downward
force F by a weight attached to the bottom end, gentle heating is observed to cause the rubber
band to contract in length. To keep the length l of the rubber band constant during heating,
F must be increased. The stretching work is given by w 0 D F dl. From this information,
nd the sign of the partial derivative .@T =@l/S;p ; then predict whether stretching of the rubber
band will cause a heating or a cooling effect.
(Hint: make a Legendre transform of U whose total differential has the independent variables
needed for the partial derivative, and write a reciprocity relation.)
You can check your prediction experimentally by touching a rubber band to the side of your
face before and after you rapidly stretch it.
C HAPTER 6
T HE T HIRD L AW AND C RYOGENICS
The third law of thermodynamics concerns the entropy of perfectly-ordered crystals at zero
kelvins.
When a chemical reaction or phase transition is studied at low temperatures, and all
substances are pure crystals presumed to be perfectly ordered, the entropy change is found
to approach zero as the temperature approaches zero kelvins:
lim S D 0
T!0
(6.0.1)
(pure, perfectly-ordered crystals)
Equation 6.0.1 is the mathematical statement of the Nernst heat theorem1 or third law of
thermodynamics. It is true in general only if each reactant and product is a pure crystal
with identical unit cells arranged in perfect spatial order.
6.1 THE ZERO OF ENTROPY

There is no theoretical relation between the entropies of different chemical elements. We
can arbitrarily choose the entropy of every pure crystalline element to be zero at zero
kelvins. Then the experimental observation expressed by Eq. 6.0.1 requires that the entropy
of every pure crystalline compound also be zero at zero kelvins, in order that the entropy
change for the formation of a compound from its elements will be zero at this temperature.
A classic statement of the third law principle appears in the 1923 book Thermodynamics
and the Free Energy of Chemical Substances by G. N. Lewis and M. Randall:2
If the entropy of each element in some crystalline state be taken as zero at the
absolute zero of temperature: every substance has a nite positive entropy, but
at the absolute zero of temperature the entropy may become zero, and does so
become in the case of perfect crystalline substances.
Nernst preferred to avoid the use of the entropy function and to use in its place the partial
derivative .@A=@T /V (Eq. 5.4.9). The original 1906 version of his heat theorem was in the form
limT!0 .@A=@T /V D0 (Ref. [39]).
2 Ref. [103], p. 448.
150
CHAPTER 6 THE THIRD LAW AND CRYOGENICS

6.1 T HE Z ERO OF E NTROPY
151
BIOGRAPHICAL SKETCH
Walther Hermann Nernst (18641941)

They had two sons, both killed in World War I,
and three daughters.
Nernst wrote an inuential textbook of
physical chemistry, the second in the eld, entitled Theoretische Chemie vom Standpunkte
der Avogadroschen Regel und der Thermodynamik. It was rst published in 1893 and its
last edition was in 1926.
Nernst began work in 1893 on a novel electric incandescent lamp based on solid-state
electrolytes. His sale of the patent in 1898
Walther Nernst was a German physical made him wealthy, but the lamp was not comchemist best known for his heat theorem, also mercially successful.
known as the third law of thermodynamics.
In 1905 Nernst was appointed director of
Nernst was born in Briesen, West Prussia the Berlin Physicochemical Institute; at the
(now Poland). His father was a district judge. end of that year he reported the discovery of
From all accounts, Nernst was not an easy his heat theorem.
person to get along with. Gilbert Lewis, who
Nernst was awarded the Nobel Prize in
spent a semester in Nernsts laboratory, was Chemistry for the year 1920 in recognition
one of those who developed an enmity to- of his work in thermochemistry. In his Noward him; in later years, Lewis delighted in bel Lecture, describing the heat theorem, he
pointing out what he considered to be errors said:
in Nernsts writings. The American physicist
. . . in all cases chemical afnity and evolution
Robert A. Millikan, who studied with Nernst
of heat become identical at low temperatures.
at G ttingen, wrote in a memorial article:a
o
Not, and this is the essential point, in the sense
He was a little fellow with a sh-like mouth and
other well-marked idiosyncrasies. However, he
was in the main popular in the laboratory, despite
the fact that in the academic world he nearly always had a quarrel on with somebody. He lived
on the second oor of the institute with his wife
and three young children. As we students came
to our work in the morning we would not infrequently meet him in his hunting suit going out
for some early morning shooting. . . . His greatest weakness lay in his intense prejudices and the
personal, rather than the objective, character of
some of his judgments.
that they intersect at absolute zero, but rather in

the sense that they invariably become practically
identical some distance before absolute zero is
reached; in other words the two curves become
mutually tangential in the vicinity of absolute
zero.
If we frame this principle in quite general
terms, i.e. if we apply it not only to chemical
but to all processes, then we have the new heat
theorem which gives rise to a series of very farreaching consequences . . .
Nernst would have nothing to do with the

Nazis. When they passed the 1933 law barAt Leipzig University, in 1888, he pub- ring Jews from state employment, he refused
lished the Nernst equation, and in 1890 the to re the Jewish scientists at the Berlin institute, and instead took the opportunity to retire.
Nernst distribution law.
In 1891 he moved to the University of He caused a stir at a meeting by refusing to
G ttingen, where in 1895 he became director stand for the singing of the Horst Wessel Lied.
o
Before his death he ordered that letters he had
of the G ttingen Physicochemical Institute.
o
In 1892 Nernst married Emma Lohmeyer, received be burned, perhaps to protect his cordaughter of a G ttingen medical professor. respondents from the Nazi authorities.b
o
a Ref.
[116].
b Ref.
[35].

6.2 M OLAR E NTROPIES
152
According to this principle, every substance (element or compound) in a pure, perfectlyordered crystal at 0 K, at any pressure,3 has a molar entropy of zero:
Sm (0 K) D 0
(6.1.1)
(pure, perfectly-ordered crystal)
This convention establishes a scale of absolute entropies at temperatures above zero kelvins
called third-law entropies, as explained in the next section.
6.2 MOLAR ENTROPIES

With the convention that the entropy of a pure, perfectly-ordered crystalline solid at zero
kelvins is zero, we can establish the third-law value of the molar entropy of a pure substance
at any temperature and pressure. Absolute values of Sm are what are usually tabulated for
calculational use.
6.2.1 Third-law molar entropies

Suppose we wish to evaluate the entropy of an amount n of a pure substance at a certain
temperature T 0 and a certain pressure. The same substance, in a perfectly-ordered crystal at
zero kelvins and the same pressure, has an entropy of zero. The entropy at the temperature
R T0
and pressure of interest, then, is the entropy change S D 0 q=T of a reversible heating
process at constant pressure that converts the perfectly-ordered crystal at zero kelvins to the
state of interest.
Consider a reversible isobaric heating process of a pure substance while it exists in a
single phase. The denition of heat capacity as q= dT (Eq. 3.1.9) allows us to substitute
Cp dT for q, where Cp is the heat capacity of the phase at constant pressure.
If the substance in the state of interest is a liquid or gas, or a crystal of a different
form than the perfectly-ordered crystal present at zero kelvins, the heating process will
include one or more equilibrium phase transitions under conditions where two phases are
in equilibrium at the same temperature and pressure (Sec. 2.2.2). For example, a reversible
heating process at a pressure above the triple point that transforms the crystal at 0 K to a gas
may involve transitions from one crystal form to another, and also melting and vaporization
transitions.
Each such reversible phase transition requires positive heat qtrs . Because the pressure
is constant, the heat is equal to the enthalpy change (Eq. 5.3.8). The ratio qtrs =n is called
the molar heat or molar enthalpy of the transition, trs H (see Sec. 8.3.1). Because the
phase transition is reversible, the entropy change during the transition is given by trs S D
qtrs =nTtrs where Ttrs is the transition temperature.
With these considerations, we can write the following expression for the entropy change
of the entire heating process:
Z T0
X ntrs H
Cp
dT C
(6.2.1)
S D
T
Ttrs
0
3 The
entropy becomes independent of pressure as T approaches zero kelvins. This behavior can be deduced
from the relation .@S=@p/T D V (Table 7.1 on page 178) combined with the experimental observation that
the cubic expansion coefcient approaches zero as T approaches zero kelvins.

153
The resulting operational equation for the calculation of the molar entropy of the substance
at the temperature and pressure of interest is
S
D
Sm .T / D
n
0
T0
X trs H
Cp;m
dT C
T
Ttrs
(6.2.2)
(pure substance,
constant p)
where Cp;m D Cp =n is the molar heat capacity at constant pressure. The summation is over
each equilibrium phase transition occurring during the heating process.
Since Cp;m is positive at all temperatures above zero kelvins, and trs H is positive
for all transitions occurring during a reversible heating process, the molar entropy of a
substance is positive at all temperatures above zero kelvins.
The heat capacity and transition enthalpy data required to evaluate Sm .T 0 / using Eq.
6.2.2 come from calorimetry. The calorimeter can be cooled to about 10 K with liquid
hydrogen, but it is difcult to make measurements below this temperature. Statistical mechanical theory may be used to approximate the part of the integral in Eq. 6.2.2 between
zero kelvins and the lowest temperature at which a value of Cp;m can be measured. The appropriate formula for nonmagnetic nonmetals comes from the Debye theory for the lattice
vibration of a monatomic crystal. This theory predicts that at low temperatures (from 0 K to
about 30 K), the molar heat capacity at constant volume is proportional to T 3 : CV;m D aT 3 ,
where a is a constant. For a solid, the molar heat capacities at constant volume and at constant pressure are practically equal. Thus for the integral on the right side of Eq. 6.2.2 we
can, to a good approximation, write
Z
T0
0
Cp;m
dT D a
T
T 00
T dT C
T0
T 00
Cp;m
dT
T
(6.2.3)
where T 00 is the lowest temperature at which Cp;m is measured. The rst term on the right
side of Eq. 6.2.3 is
Z T 00
T 00
a
T 2 dT D .aT 3 =3/ D a.T 00 /3 =3
(6.2.4)
0
But a.T 00 /3 is the value of Cp;m at T 00 , so Eq. 6.2.2 becomes

Cp;m .T 00 /
C
Sm .T / D
3
0
T0
T 00
X trs H
Cp;m
dT C
T
Ttrs
(6.2.5)
(pure substance,
constant p)
In the case of a metal, statistical mechanical theory predicts an electronic contribution

to the molar heat capacity, proportional to T at low temperature, that should be added
to the Debye T 3 term: Cp;m D aT 3 C bT . The error in using Eq. 6.2.5, which ignores
the electronic term, is usually negligible if the heat capacity measurements are made
down to about 10 K.
We may evaluate the integral on the right side of Eq. 6.2.5 by numerical integration.
We need the area under the curve of Cp;m =T plotted as a function of T between some low

J K mol

m J K mol
154
J K mol

m J K mol
ln K
Figure 6.1 Properties of hydrogen chloride (HCl): the dependence of Cp;m , Cp;m =T ,
and Sm on temperature at a pressure of 1 bar. The discontinuities are at a
solid!solid phase transition, the melting temperature, and the vaporization temperature. (Condensed-phase data from Ref. [60]; gas-phase data from Ref. [28], p. 762.)
temperature, T 00 , and the temperature T 0 at which the molar entropy is to be evaluated.

Since the integral may be written in the form
Z
T0
T 00
Cp;m
dT D
T
T DT 0
T DT 00
Cp;m d ln.T =K/
(6.2.6)
we may also evaluate the integral from the area under a curve of Cp;m plotted as a function
of ln.T =K/.
The procedure of evaluating the entropy from the heat capacity is illustrated for the
case of hydrogen chloride in Fig. 6.1. The areas under the curves of Cp;m =T versus T ,
and of Cp;m versus ln.T =K/, in a given temperature range are numerically identical (Eq.
6.2.6). Either curve may be used in Eq. 6.2.2 to nd the dependence of Sm on T . Note how

155
the molar entropy increases continuously with increasing T and has a discontinuity at each
phase transition.
As explained in Sec. 6.1, by convention the zero of entropy of any substance refers to
the pure, perfectly-ordered crystal at zero kelvins. In practice, experimental entropy
values depart from this convention in two respects. First, an element is usually a
mixture of two or more isotopes, so that the substance is not isotopically pure. Second,
if any of the nuclei have spins, weak interactions between the nuclear spins in the
crystal would cause the spin orientations to become ordered at a very low temperature.
Above 1 K, however, the orientation of the nuclear spins become essentially random,
and this change of orientation is not included in the Debye T 3 formula.
The neglect of these two effects results in a practical entropy scale, or conventional
entropy scale, on which the crystal that is assigned an entropy of zero has randomlymixed isotopes and randomly-oriented nuclear spins, but is pure and ordered in other
respects. This is the scale that is used for published values of absolute third-law
molar entropies. The shift of the zero away from a completely-pure and perfectlyordered crystal introduces no inaccuracies into the calculated value of S for any
process occurring above 1 K, because the shift is the same in the initial and nal states.
That is, isotopes remain randomly mixed and nuclear spins remain randomly oriented.
6.2.2 Molar entropies from spectroscopic measurements

Statistical mechanical theory applied to spectroscopic measurements provides an accurate
means of evaluating the molar entropy of a pure ideal gas from experimental molecular
properties. This is often the preferred method of evaluating Sm for a gas. The zero of
entropy is the same as the practical entropy scalethat is, isotope mixing and nuclear spin
interactions are ignored. Intermolecular interactions are also ignored, which is why the
results apply only to an ideal gas.
The statistical mechanics formula writes the molar entropy as the sum of a translational
contribution and an internal contribution: Sm D Sm;trans C Sm;int . The translational
contribution is given by the SackurTetrode equation:
Sm;trans D R ln
.2 M /3=2 .RT /5=2

C .5=2/R
4
ph3 NA
(6.2.7)
Here h is the Planck constant and NA is the Avogadro constant. The internal contribution is given by
Sm;int D R ln qint C RT .d ln qint = dT /
(6.2.8)
where qint is the molecular partition function dened by
X
qint D
exp . i =kT /
(6.2.9)
In Eq. 6.2.9, i is the energy of a molecular quantum state relative to the lowest energy level, k is the Boltzmann constant, and the sum is over the quantum states of one
molecule with appropriate averaging for natural isotopic abundance. The experimental
data needed to evaluate qint consist of the energies of low-lying electronic energy levels, values of electronic degeneracies, fundamental vibrational frequencies, rotational
constants, and other spectroscopic parameters.

156
Table 6.1 Standard molar entropies of several substances (ideal gases

at T D 298:15 K and p D 1 bar) and molar residual entropies
Sm =J K
mol
Substance
calorimetric
spectroscopica
Sm,0 =J K
HCl
CO
NO
N2 O (NNO)
H2 O
186:3 0:4b
193:4 0:4c
208:0 0:4d
215:3 0:4e
185:4 0:2f
186:901
197:65 0:04
210:758
219:957
188:834 0:042
0:6 0:4
4:3 0:4
2:8 0:4
4:7 0:4
3:4 0:2
a Ref.
[28].
b Ref.
[60].
c Ref.
[34].
d Ref.
[83].
e Ref.
[15].
mol
f Ref.
[59].
When the spectroscopic method is used to evaluate Sm with p set equal to the standard
pressure p D 1 bar, the value is the standard molar entropy, Sm , of the substance in the
gas phase. This value is useful for thermodynamic calculations even if the substance is not
an ideal gas at the standard pressure, as will be discussed in Sec. 7.9.
6.2.3 Residual entropy

Ideally, the molar entropy values obtained by the calorimetric (third-law) method for a gas
should agree closely with the values calculated from spectroscopic data. Table 6.1 shows
that for some substances this agreement is not present. The table lists values of Sm for
ideal gases at 298:15 K evaluated by both the calorimetric and spectroscopic methods. The
quantity Sm,0 in the last column is the difference between the two Sm values, and is called
the molar residual entropy.
In the case of HCl, the experimental value of the residual entropy is comparable to its
uncertainty, indicating good agreement between the calorimetric and spectroscopic methods. This agreement is typical of most substances, particularly those like HCl whose
molecules are polar and asymmetric with a large energetic advantage of forming perfectlyordered crystals.
The other substances listed in Table 6.1 have residual entropies that are greater than
zero within the uncertainty of the data. What is the meaning of this discrepancy between
the calorimetric and spectroscopic results? We can assume that the true values of Sm at
298:15 K are the spectroscopic values, because their calculation assumes the solid has only
one microstate at 0 K, with an entropy of zero, and takes into account all of the possible
accessible microstates of the ideal gas. The calorimetric values, on the other hand, are
based on Eq. 6.2.2 which assumes the solid becomes a perfectly-ordered crystal as the
temperature approaches 0 K.4
The conventional explanation of a nonzero residual entropy is the presence of random
rotational orientations of molecules in the solid at the lowest temperature at which the heat
4 The
calorimetric values in Table 6.1 were calculated as follows. Measurements of heat capacities and heats
of transition were used in Eq. 6.2.2 to nd the third-law value of Sm for the vapor at the boiling point of the
substance at p D 1 atm. This calculated value for the gas was corrected to that for the ideal gas at p D 1 bar
and adjusted to T D 298:15 K with spectroscopic data.

6.3 C RYOGENICS
(a)
(b)
157
Figure 6.2 JouleThomson expansion of a gas through a porous plug. The shaded
area represents a xed-amount sample of the gas (a) at time t1 ; (b) at a later time t2 .
capacity can be measured, so that the crystals are not perfectly ordered. The random structure is established as the crystals form from the liquid, and becomes frozen into the crystals
as the temperature is lowered below the freezing point. This tends to happen with almostsymmetric molecules with small dipole moments which in the crystal can have random
rotational orientations of practically equal energy. In the case of solid H2 O it is the arrangement of intermolecular hydrogen bonds that is random. Crystal imperfections such
as dislocations can also contribute to the residual entropy. If such crystal imperfection is
present at the lowest experimental temperature, the calorimetric value of Sm for the gas at
298:15 K is the molar entropy increase for the change at 1 bar from the imperfectly-ordered
solid at 0 K to the ideal gas at 298:15 K, and the residual entropy Sm,0 is the molar entropy
of this imperfectly-ordered solid.
6.3 CRYOGENICS
The eld of cryogenics involves the production of very low temperatures, and the study of
the behavior of matter at these temperatures. These low temperatures are needed to evaluate
third-law entropies using calorimetric measurements. There are some additional interesting
thermodynamic applications.
6.3.1 JouleThomson expansion

A gas can be cooled by expanding it adiabatically with a piston (Sec. 3.5.3), and a liquid
can be cooled by pumping on its vapor to cause evaporation (vaporization). An evaporation procedure with a refrigerant uid is what produces the cooling in an ordinary kitchen
refrigerator.
For further cooling of a uid, a common procedure is to use a continuous throttling
process in which the uid is forced to ow through a porous plug, valve, or other constriction that causes an abrupt drop in pressure. A slow continuous adiabatic throttling of
a gas is called the JouleThomson experiment, or JouleKelvin experiment, after the two
scientists who collaborated between 1852 and 1862 to design and analyze this procedure.5
The principle of the JouleThomson experiment is shown in Fig. 6.2. A tube with
thermally insulated walls contains a gas maintained at a constant pressure p 0 at the left
5 William
Thomson later became Lord Kelvin.

6.3 C RYOGENICS
158
side of a porous plug and at a constant lower pressure p 00 at the right side. Because of the
pressure difference, the gas ows continuously from left to right through the plug. The ow
is slow, and the pressure is essentially uniform throughout the portion of the tube at each
side of the plug, but has a large gradient within the pores of the plug.
After the gas has been allowed to ow for a period of time, a steady state develops in
the tube. In this steady state, the gas is assumed to have a uniform temperature T 0 at the left
side of the plug and a uniform temperature T 00 (not necessarily equal to T 0 ) at the right side
of the plug.
Consider the segment of gas whose position at times t1 and t2 is indicated by shading
in Fig. 6.2. This segment contains a xed amount of gas and expands as it moves through
the porous plug from higher to lower pressure. We can treat this gas segment as a closed
system. During the interval between times t1 and t2 , the system passes through a sequence
of different states, none of which is an equilibrium state since the process is irreversible. The
energy transferred across the boundary by heat is zero, because the tube wall is insulated
and there is no temperature gradient at either end of the gas segment. We calculate the
energy transferred by work at each end of the gas segment from w D pb As dx, where
pb is the pressure (either p 0 or p 00 ) at the moving boundary, As is the cross-section area of
the tube, and x is the distance along the tube. The result is
wD
0
p 0 .V2
0
V1 /
00
p 00 .V2
00
V1 /
(6.3.1)
0
0
where the meaning of the volumes V1 , V2 , and so on is indicated in the gure.
The internal energy change U of the gas segment must be equal to w, since q is zero.
Now let us nd the enthalpy change H . At each instant, a portion of the gas segment is in
the pores of the plug, but this portion contributes an unchanging contribution to both U and
H because of the steady state. The rest of the gas segment is in the portions on either side
of the plug, with enthalpies U 0 C p 0 V 0 at the left and U 00 C p 00 V 00 at the right. The overall
enthalpy change of the gas segment must be
0
00
H D U C .p 0 V2 C p 00 V2 /
0
00
.p 0 V1 C p 00 V1 /
(6.3.2)
which, when combined with the expression of Eq. 6.3.1 for w D U , shows that H
is zero. In other words, the gas segment has the same enthalpy before and after it passes
through the plug: the throttling process is isenthalpic.
The temperatures T 0 and T 00 can be measured directly. When values of T 00 versus p 00
are plotted for a series of JouleThomson experiments having the same values of T 0 and
p 0 and different values of p 00 , the curve drawn through the points is a curve of constant
enthalpy. The slope at any point on this curve is equal to the JouleThomson coefcient
(or JouleKelvin coefcient) dened by

@T
def
(6.3.3)
JT D
@p H
For an ideal gas, JT is zero because the enthalpy of an ideal gas depends only on T (Prob.
5.1); T cannot change if H is constant. For a nonideal gas, JT is a function of T and
p and the kind of gas.6 For most gases, at low to moderate pressures and at temperatures
6 See
Sec. 7.5.2 for the relation of the JouleThomson coefcient to other properties of a gas.

6.3 C RYOGENICS
159
Figure 6.3 Adiabatic demagnetization to achieve a low temperature in a paramagnetic solid.
not much greater than room temperature, JK is positive. Under these conditions, a Joule
Thomson expansion to a lower pressure has a cooling effect, because T will decrease as p
decreases at constant H . Hydrogen and helium, however, have negative values of JK at
room temperature and must be cooled by other means to about 200 K and 40 K, respectively,
in order for a JouleThomson expansion to cause further cooling.
The cooling effect of a JouleThomson expansion is often used to cool a gas down to its
condensation temperature. This procedure can be carried out continuously by pumping the
gas through the throttle and recirculating the cooler gas on the low-pressure side through a
heat exchanger to help cool the gas on the high-pressure side. Starting at room temperature,
gaseous nitrogen can be condensed by this means to liquid nitrogen at 77:4 K. The liquid
nitrogen can then be used as a cooling bath for gaseous hydrogen. At 77:4 K, hydrogen
has a positive JouleThomson coefcient, so that it in turn can be converted by a throttling
process to liquid hydrogen at 20:3 K. Finally, gaseous helium, whose JouleThomson coefcient is positive at 20:3 K, can be converted to liquid helium at 4:2 K. Further cooling of
the liquid helium to about 1 K can be carried out by pumping to cause rapid evaporation.
6.3.2 Magnetization
The work of magnetization of an isotropic paramagnetic phase can be written w 0 D
B dmmag , where B is the magnitude of the magnetic ux density and mmag is the magnitude of the magnetic dipole moment of the phase. The total differential of the internal
energy of a closed isotropic phase with magnetization is given by
dU D T dS
p dV C B dmmag
(6.3.4)
with S , V , and mmag as the independent variables.

The technique of adiabatic demagnetization can be used to obtain temperatures below 1 K. This method, suggested by Peter Debye in 1926 and independently by William
Giauque in 1927, requires a paramagnetic solid in which ions with unpaired electrons are
sufciently separated that at 1 K the orientations of the magnetic dipoles are almost completely random. Gadolinium(III) sulfate octahydrate, Gd2 .SO4 /3 8H2 O, is commonly used.
Figure 6.3 illustrates the principle of the technique. The solid curve shows the temperature dependence of the entropy of a paramagnetic solid in the absence of an applied

6.3 C RYOGENICS
160
BIOGRAPHICAL SKETCH
`
AIP EMILIO SEGRE VISUAL ARCHIVES
BRITTLE BOOKS COLLECTION
William Francis Giauque (18951982)
William Giauque was an American chemist

who made important contributions to the eld
of cryogenics. He received the 1949 Nobel
Prize in Chemistry for his contributions in the
eld of chemical thermodynamics, particularly
concerning the behaviour of substances at extremely low temperatures.
Giauque was born in Niagara Falls, Ontario, Canada, but as his father was a citizen of
the United States, William was able to adopt
American citizenship. His father worked as a
weighmaster and station agent for the Michigan Central Railroad in Michigan.
Giauques initial career goal after high
school was electrical engineering. After working for two years at the Hooker Electrochemical Company in Niagara Falls, New York, he
left to continue his education with the idea of
becoming a chemical engineer. Hearing of the
scientic reputation of G. N. Lewis, the chair
of the chemistry department at the University
of California at Berkeley (page 271), and motivated by the free tuition there, he enrolled instead in that department.a
Giauque spent the rest of his life in the
chemistry department at Berkeley, rst as an
undergraduate; then as a graduate student; and
nally, after receiving his Ph.D. in 1922, as
a faculty member. Some of his undergraduate courses were in engineering, which later
helped him in the design and construction of
the heavy equipment for producing the high
magnetic elds and the liquid hydrogen and
helium needed in his research.
Beginning in 1928, with the help of his
a Ref.
[130].
b Ref.
[152].
c Ref.
[61].
d Ref.
[58].
graduate students and collaborators, he began

to publish papers on various aspects of the
third law.b The research included the evaluation of third-law molar entropies and comparison to molar entropies calculated from spectroscopic data, and the study of the residual
entropy of crystals. Faint unexplained lines
in the absorption spectrum of gaseous oxygen
led him to the discovery of the previously unknown 17 O and 18 O isotopes of oxygen.
Giauques best-known accomplishment is
his invention and exploitation of cooling to
very low temperatures by adiabatic demagnetization. In 1924, he learned of the unusual
properties of gadolinium sulfate octahydrate at
the temperature of liquid helium. In his Nobel
Lecture, he said:c
I was greatly surprised to nd, that the application of a magnetic eld removes a large amount
of entropy from this substance, at a temperature
so low that it had been thought that there was
practically no entropy left to remove. . . . Those
familiar with thermodynamics will realize that
in principle any process involving an entropy
change may be used to produce either cooling or
heating. Accordingly it occurred to me that adiabatic demagnetization could be made the basis of
a method for producing temperatures lower than
those obtainable with liquid helium.
It wasnt until 1933 that he was able to build

the equipment and publish the results from his
rst adiabatic demagnetization experiment, reporting a temperature of 0:25 K.d
He was married to a botanist and had two
sons. According to one biography:e
Giauques students remember pleasant Thanksgiving dinners at the Giauque home, with Muriel
[his wife] as cook and Frank (as she called him)
as raconteur, with a keen sense of humor. The
stories he most enjoyed telling were those in
which the joke was on him. . . . Giauques conservatism was legendary. He always appeared at
the university dressed in an iron-gray tweed suit.
. . . A dominant personality himself, Giauque not
only tolerated but respected students who disagreed with him, and he was especially pleased
when they could prove their point.
e Ref.
[130].

6.3 C RYOGENICS
161
magnetic eld, and the dashed curve is for the solid in a constant, nite magnetic eld. The
temperature range shown is from 0 K to approximately 1 K. At 0 K, the magnetic dipoles
are perfectly ordered. The increase of S shown by the solid curve between 0 K and 1 K is
due almost entirely to increasing disorder in the orientations of the magnetic dipoles as heat
enters the system.
Path A represents the process that occurs when the paramagnetic solid, surrounded by
gaseous helium in thermal contact with liquid helium that has been cooled to about 1 K, is
slowly moved into a strong magnetic eld. The process is isothermal magnetization, which
partially orients the magnetic dipoles and reduces the entropy. During this process there is
heat transfer to the liquid helium, which partially boils away. In path B, the thermal contact
between the solid and the liquid helium has been broken by pumping away the gas surrounding the solid, and the sample is slowly moved away from the magnetic eld. This step
is a reversible adiabatic demagnetization. Because the process is reversible and adiabatic,
the entropy change is zero, which brings the state of the solid to a lower temperature as
shown.
The sign of .@T =@B/S;p is of interest because it tells us the sign of the temperature
change during a reversible adiabatic demagnetization (path B of Fig. 6.3 on page 159).
To change the independent variables in Eq. 6.3.4 to S, p, and B, we dene the Legendre
transform
def
H 0 D U C pV Bmmag
(6.3.5)
(H 0 is sometimes called the magnetic enthalpy.) From Eqs. 6.3.4 and 6.3.5 we obtain the
total differential
dH 0 D T dS C V dp mmag dB
(6.3.6)
From it we nd the reciprocity relation

@T
D
@B S;p
@mmag
@S
(6.3.7)
p;B
According to Curies law of magnetization, the magnetic dipole moment mmag of a

paramagnetic phase at constant magnetic ux density B is proportional to 1=T . This law
applies when B is small, but even if B is not small mmag decreases with increasing T . To
increase the temperature of a phase at constant B, we allow heat to enter the system, and S
then increases. Thus, .@mmag =@S/p;B is negative and, according to Eq. 6.3.7, .@T =@B/S;p
must be positive. Adiabatic demagnetization is a constant-entropy process in which B
decreases, and therefore the temperature also decreases.
We can nd the sign of the entropy change during the isothermal magnetization process
shown as path A in Fig. 6.3 on page 159. In order to use T , p, and B as the independent
def
variables, we dene the Legendre transform G 0 D H 0 T S. Its total differential is
dG 0 D
S dT C V dp
mmag dB
From this total differential, we obtain the reciprocity relation

@mmag
@S
D
@B T;p
@T
p;B
(6.3.8)
(6.3.9)

6.3 C RYOGENICS
162
Since mmag at constant B decreases with increasing T , as explained above, we see that the
entropy change during isothermal magnetization is negative.
By repeatedly carrying out a procedure of isothermal magnetization and adiabatic demagnetization, starting each stage at the temperature produced by the previous stage, it has
been possible to attain a temperature as low as 0:0015 K. The temperature can be reduced
still further, down to 16 microkelvins, by using adiabatic nuclear demagnetization. However, as is evident from the gure, if in accordance with the third law both of the entropy
curves come together at the absolute zero of the kelvin scale, then it is not possible to attain
a temperature of zero kelvins in a nite number of stages of adiabatic demagnetization. This
conclusion is called the principle of the unattainability of absolute zero.

P ROBLEM
163
PROBLEM
6.1 Calculate the molar entropy of carbon disulde at 25:00 C and 1 bar from the heat capacity
data for the solid in Table 6.2 and the following data for p D 1 bar. At the melting point,
Table 6.2 Molar heat capacity
of CS2 (s) at p D 1 bar a
T =K
15:05
20:15
29:76
42:22
57:52
75:54
94:21
108:93
131:54
156:83
a Ref.
Cp;m =J K
mol
6:9
12:0
20:8
29:2
35:6
40:0
45:0
48:5
52:6
56:6
[25].
161:11 K, the molar enthalpy of fusion is fus H D 4:39 103 J mol 1 . The molar heat capacity of the liquid in the range 161300 K is described by Cp;m D a C bT , where the constants
have the values a D 74:6 J K 1 mol 1 and b D 0:0034 J K 2 mol 1 .
C HAPTER 7
P URE S UBSTANCES IN S INGLE P HASES
This chapter applies concepts introduced in earlier chapters to the simplest kind of system,
one consisting of a pure substance or a single component in a single phase. The system has
three independent variables if it is open, and two if it is closed. Relations among various
properties of a single phase are derived, including temperature, pressure, and volume. The
important concepts of standard states and chemical potential are introduced.
7.1 VOLUME PROPERTIES

Two volume properties of a closed system are dened as follows:
@V
def 1
cubic expansion coefcient
D
V @T p
1 @V
def
isothermal compressibility T D
V @p T
(7.1.1)
(7.1.2)
The cubic expansion coefcient is also called the coefcient of thermal expansion and
the expansivity coefcient. Other symbols for the isothermal compressibility are and
T.
These denitions show that is the fractional volume increase per unit temperature
increase at constant pressure, and T is the fractional volume decrease per unit pressure
increase at constant temperature. Both quantities are intensive properties. Most substances
have positive values of ,1 and all substances have positive values of T , because a pressure
increase at constant temperature requires a volume decrease.
If an amount n of a substance is in a single phase, we can divide the numerator and
denominator of the right sides of Eqs. 7.1.1 and 7.1.2 by n to obtain the alternative expressions
1 The
cubic expansion coefcient is not always positive. is negative for liquid water below its temperature
of maximum density, 3:98 C. The crystalline ceramics zirconium tungstate (ZrW2 O8 ) and hafnium tungstate
(HfW2 O8 ) have the remarkable behavior of contracting uniformly and continuously in all three dimensions
when they are heated from 0:3 K to about 1050 K; is negative throughout this very wide temperature range
(Ref. [109]). The intermetallic compound YbGaGe has been found to have a value of that is practically zero
in the range 100300 K (Ref. [146]).
164
CHAPTER 7 PURE SUBSTANCES IN SINGLE PHASES

7.1 VOLUME P ROPERTIES
165
3:0
H2
idea
l ga
s
O(g
)
2:5
2:0
)
ane(l
n-hex
1:5
=10
1:0
nol(l)
1-propa
0:5
H 2O
(l)
H2 O(s)
0:0
0:5
25
25
50
75
100
125
150
t= C
Figure 7.1 The cubic expansion coefcient of several substances and an ideal gas as
functions of temperature at p D 1 bar. a Note that because liquid water has a density
maximum at 4 C, is zero at that temperature.
a Based
on data in Ref. [49], p. 104; Ref. [70]; and Ref. [165], p. 28.
1
D
Vm
T D
1
Vm
@Vm
@T
@Vm
@p
(7.1.3)
(pure substance, P D1)
(7.1.4)
where Vm is the molar volume. P in the conditions of validity is the number of phases. Note
that only intensive properties appear in Eqs. 7.1.3 and 7.1.4; the amount of the substance
is irrelevant. Figures 7.1 and 7.2 show the temperature variation of and T for several
substances.
If we choose T and p as the independent variables of the closed system, the total differential of V is given by
@V
@V
dV D
dT C
dp
(7.1.5)
@T p
@p T

7.2 I NTERNAL P RESSURE
166
l(l)
no
tha
e
)
e(l
zen
n
be
10
T =10
bar
15
H2 O(l)
H2 O(s)
.
0
25
25
50
75
100
t= C
Figure 7.2 The isothermal compressibility of several substances as a function of temperature at p D 1 bar. (Based on data in Ref. [49]; Ref. [88]; and Ref. [165], p. 28.)
With the substitutions .@V =@T /p D V (from Eq. 7.1.1) and .@V =@p/T D
Eq. 7.1.2), the expression for the total differential of V becomes
dV D V dT
T V dp
T V (from
(7.1.6)
(closed system,
C D1, P D1)
To nd how p varies with T in a closed system kept at constant volume, we set dV equal
to zero in Eq. 7.1.6: 0 D V dT T V dp, or dp= dT D =T . Since dp= dT under
the condition of constant volume is the partial derivative .@p=@T /V , we have the general
relation

@p
(7.1.7)
D
@T V
T
(closed system,
C D1, P D1)
7.2 INTERNAL PRESSURE

The partial derivative .@U=@V /T applied to a uid phase in a closed system is called the
internal pressure. (Note that U and pV have dimensions of energy; therefore, U=V has
dimensions of pressure.)
To relate the internal pressure to other properties, we divide Eq. 5.2.2 by dV : dU= dV D
T .dS= dV / p. Then we impose a condition of constant T : .@U=@V /T D T .@S=@V /T p.
When we make a substitution for .@S=@V /T from the Maxwell relation of Eq. 5.4.17, we

7.2 I NTERNAL P RESSURE
obtain
@U
@V
DT
@p
@T
167
p
V
(7.2.1)
(closed system,
uid phase, C D1)
This equation is sometimes called the thermodynamic equation of state of the uid.
For an ideal-gas phase, we can write p D nRT =V and then

@p
nR
p
D
D
(7.2.2)
@T V
V
T
Making this substitution in Eq. 7.2.1 gives us
@U
D0
@V T
(7.2.3)
(closed system of an ideal gas)
showing that the internal pressure of an ideal gas is zero.

In Sec. 3.5.1, an ideal gas was dened as a gas (1) that obeys the ideal gas equation, and
(2) for which U in a closed system depends only on T . Equation 7.2.3, derived from
the rst part of this denition, expresses the second part. It thus appears that the second
part of the denition is redundant, and that we could dene an ideal gas simply as a
gas obeying the ideal gas equation. This argument is valid only if we assume the idealgas temperature is the same as the thermodynamic temperature (Secs. 2.3.5 and 4.3.4)
since this assumption is required to derive Eq. 7.2.3. Without this assumption, we cant
dene an ideal gas solely by pV D nRT , where T is the ideal gas temperature.
Here is a simplied interpretation of the signicance of the internal pressure. When

the volume of a uid increases, the average distance between molecules increases and the
potential energy due to intermolecular forces changes. If attractive forces dominate, as they
usually do unless the uid is highly compressed, expansion causes the potential energy to
increase. The internal energy is the sum of the potential energy and thermal energy. The
internal pressure, .@U=@V /T , is the rate at which the internal energy changes with volume
at constant temperature. At constant temperature, the thermal energy is constant so that the
internal pressure is the rate at which just the potential energy changes with volume. Thus,
the internal pressure is a measure of the strength of the intermolecular forces and is positive
if attractive forces dominate.2 In an ideal gas, intermolecular forces are absent and therefore
the internal pressure of an ideal gas is zero.
With the substitution .@p=@T /V D =T (Eq. 7.1.7), Eq. 7.2.1 becomes
T
@U
D
p
(7.2.4)
@V T
T
(closed system,
uid phase, C D1)
The internal pressure of a liquid at p D 1 bar is typically much larger than 1 bar (see Prob.
7.6). Equation 7.2.4 shows that, in this situation, the internal pressure is approximately
equal to T =T .
2 These
attractive intermolecular forces are the cohesive forces that can allow a negative pressure to exist in a
liquid; see page 38.

7.3 T HERMAL P ROPERTIES
168
7.3 THERMAL PROPERTIES

For convenience in derivations to follow, expressions from Chap. 5 are repeated here that
apply to processes in a closed system in the absence of nonexpansion work (i.e., w 0 D0).
For a process at constant volume we have3
@U
(7.3.1)
dU D q
CV D
@T V
and for a process at constant pressure we have4
dH D q
Cp D
@H
@T
(7.3.2)
A closed system of one component in a single phase has only two independent variables.
In such a system, the partial derivatives above are complete and unambiguous denitions of
CV and Cp because they are expressed with two independent variablesT and V for CV ,
and T and p for Cp . As mentioned on page 143, additional conditions would have to be
specied to dene CV for a more complicated system; the same is true for Cp .
For a closed system of an ideal gas we have5
CV D
dU
dT
Cp D
dH
dT
(7.3.3)
7.3.1 The relation between CV;m and Cp;m

The value of Cp;m for a substance is greater than CV;m . The derivation is simple in the case
of a xed amount of an ideal gas. Using substitutions from Eq. 7.3.3, we write
Cp
CV D
dU
d.H U /
d.pV /
D
D
D nR
dT
dT
dT
dH
dT
(7.3.4)
Division by n to obtain molar quantities and rearrangement then gives

Cp;m D CV;m C R
(7.3.5)
(ideal gas, pure substance)
For any phase in general, we proceed as follows. First we write
@H
@.U C pV /
@U
@V
Cp D
D
D
Cp
@T p
@T
@T p
@T p
p
(7.3.6)
Then we write the total differential of U with T and V as independent variables and identify
one of the coefcients as CV :
@U
@U
@U
dU D
dT C
dV D CV dT C
dV
(7.3.7)
@T V
@V T
@V T
3 Eqs.
5.3.9 and 5.6.1.
4 Eqs.
5.3.7 and 5.6.3.
5 Eqs.
5.6.2 and 5.6.4.

169
When we divide both sides of the preceding equation by dT and impose a condition of
constant p, we obtain
@U
@U
@V
D CV C
(7.3.8)
@T p
@V T @T p
Substitution of this expression for .@U=@T /p in the equation for Cp yields
@V
@U
Cp D CV C
Cp
@V T
@T p
(7.3.9)
Finally we set the partial derivative .@U=@V /T (the internal pressure) equal to .T =T / p
(Eq. 7.2.4) and .@V =@T /p equal to V to obtain
Cp D CV C
2T V
T
(7.3.10)
and divide by n to obtain molar quantities:

Cp;m D CV;m C
2 T Vm
T
(7.3.11)
Since the quantity 2 T Vm =T must be positive, Cp;m is greater than CV;m .
7.3.2 The measurement of heat capacities

The most accurate method of evaluating the heat capacity of a phase is by measuring the
temperature change resulting from heating with electrical work. The procedure in general
is called calorimetry, and the apparatus containing the phase of interest and the electric
heater is a calorimeter. The principles of three commonly-used types of calorimeters with
electrical heating are described below.
Adiabatic calorimeters
An adiabatic calorimeter is designed to have negligible heat ow to or from its surroundings.
The calorimeter contains the phase of interest, kept at either constant volume or constant
pressure, and also an electric heater and a temperature-measuring device such as a platinum
resistance thermometer, thermistor, or quartz crystal oscillator. The contents may be stirred
to ensure temperature uniformity.
To minimize conduction and convection, the calorimeter usually is surrounded by a
jacket separated by an air gap or an evacuated space. The outer surface of the calorimeter
and inner surface of the jacket may be polished to minimize radiation emission from these
surfaces. These measures, however, are not sufcient to ensure a completely adiabatic
boundary, because energy can be transferred by heat along the mounting hardware and
through the electrical leads. Therefore, the temperature of the jacket, or of an outer metal
shield, is adjusted throughout the course of the experiment so as to be as close as possible
to the varying temperature of the calorimeter. This goal is most easily achieved when the
temperature change is slow.
To make a heat capacity measurement, a constant electric current is passed through the
heater circuit for a known period of time. The system is the calorimeter and its contents.

170
Figure 7.3 Typical heating curve of an adiabatic calorimeter.
The electrical work wel performed on the system by the heater circuit is calculated from the
integrated form of Eq. 3.8.5 on page 89: wel D I 2 Rel t , where I is the electric current, Rel
is the electric resistance, and t is the time interval. We assume the boundary is adiabatic
and write the rst law in the form
dU D
p dV C wel C wcont
(7.3.12)
where p dV is expansion work and wcont is any continuous mechanical work from stirring
(the subscript cont stands for continuous). If electrical work is done on the system by a
thermometer using an external electrical circuit, such as a platinum resistance thermometer,
this work is included in wcont .
Consider rst an adiabatic calorimeter in which the heating process is carried out at
constant volume. There is no expansion work, and Eq. 7.3.12 becomes
dU D wel C wcont
(7.3.13)
(constant V )
An example of a measured heating curve (temperature T as a function of time t) is shown in

Fig. 7.3. We select two points on the heating curve, indicated in the gure by open circles.
Time t1 is at or shortly before the instant the heater circuit is closed and electrical heating
begins, and time t2 is after the heater circuit has been opened and the slope of the curve has
become essentially constant.
In the time periods before t1 and after t2 , the temperature may exhibit a slow rate of
increase due to the continuous work wcont from stirring and temperature measurement. If
this work is performed at a constant rate throughout the course of the experiment, the slope
is constant and the same in both time periods as shown in the gure.
The relation between the slope and the rate of work is given by a quantity called the
energy equivalent, . The energy equivalent is the heat capacity of the calorimeter under
the conditions of an experiment. The heat capacity of a constant-volume calorimeter is
given by D .@U=@T /V (Eq. 5.6.1). Thus, at times before t1 or after t2 , when wel is zero
and dU equals wcont , the slope r of the heating curve is given by
rD
dT dU
1 wcont
dT
D
D
dt
dU dt
dt
(7.3.14)

171
The rate of the continuous work is therefore wcont = dt D r. This rate is constant throughout the experiment. In the time interval from t1 to t2 , the total quantity of continuous work
is wcont D r.t2 t1 /, where r is the slope of the heating curve measured outside this time
interval.
To nd the energy equivalent, we integrate Eq. 7.3.13 between the two points on the
curve:
U D wel C wcont D wel C r.t2
t1 /
(7.3.15)
(constant V )
Then the average heat capacity between temperatures T1 and T2 is

D
U
w C r.t2 t1 /
D el
T2 T1
T2 T1
(7.3.16)
wel
T1 r.t2
(7.3.17)
Solving for , we obtain

D
T2
t1 /
The value of the denominator on the right side is indicated by the vertical line in Fig. 7.3. It
is the temperature change that would have been observed if the same quantity of electrical
work had been performed without the continuous work.
Next, consider the heating process in a calorimeter at constant pressure. In this case the
enthalpy change is given by dH D dU C p dV which, with substitution from Eq. 7.3.12,
becomes
dH D wel C wcont
(7.3.18)
(constant p)
We follow the same procedure as for the constant-volume calorimeter, using Eq. 7.3.18 in
place of Eq. 7.3.13 and equating the energy equivalent to .@H=@T /p , the heat capacity of
the calorimeter at constant pressure (Eq. 5.6.3). We obtain the relation
H D wel C wcont D wel C r.t2
t1 /
(7.3.19)
(constant p)
in place of Eq. 7.3.15 and end up again with the expression of Eq. 7.3.17 for .
The value of calculated from Eq. 7.3.17 is an average value for the temperature interval from T1 to T2 , and we can identify this value with the heat capacity at the temperature
of the midpoint of the interval. By taking the difference of values of measured with and
without the phase of interest present in the calorimeter, we obtain CV or Cp for the phase
alone.
It may seem paradoxical that we can use an adiabatic process, one without heat, to evaluate a quantity dened by heat (heat capacity D q= dT ). The explanation is that energy
transferred into the adiabatic calorimeter as electrical work, and dissipated completely to
thermal energy, substitutes for the heat that would be needed for the same change of state
without electrical work.

172
slope =
slope =
Figure 7.4 Typical heating curve of an isothermal-jacket calorimeter.
Isothermal-jacket calorimeters
A second common type of calorimeter is similar in construction to an adiabatic calorimeter,
except that the surrounding jacket is maintained at constant temperature. It is sometimes
called an isoperibol calorimeter. A correction is made for heat transfer resulting from the
difference in temperature across the gap separating the jacket from the outer surface of the
calorimeter. It is important in making this correction that the outer surface have a uniform
temperature without hot spots.
Assume the outer surface of the calorimeter has a uniform temperature T that varies
with time, the jacket temperature has a constant value Text , and convection has been eliminated by evacuating the gap. Then heat transfer is by conduction and radiation, and its rate
is given by Newtons law of cooling
q
D
dt
k.T
Text /
(7.3.20)
where k is a constant (the thermal conductance). Heat ows from a warmer to a cooler
body, so q= dt is positive if T is less than Text and negative if T is greater than Text .
The possible kinds of work are the same as for the adiabatic calorimeter: expansion
work p dV , intermittent work wel done by the heater circuit, and continuous work wcont .
By combining the rst law and Eq. 7.3.20, we obtain the following relation for the rate at
which the internal energy changes:
dU
q
w
D
C
D
dt
dt
dt
k.T
Text /
dV
wel
wcont
C
C
dt
dt
dt
(7.3.21)
For heating at constant volume (dV = dt D 0), this relation becomes

dU
D
dt
k.T
Text / C
wel
wcont
C
dt
dt
(7.3.22)
(constant V )
An example of a heating curve is shown in Fig. 7.4. In contrast to the curve of Fig.
7.3, the slopes are different before and after the heating interval due to the changed rate
of heat ow. Times t1 and t2 are before and after the heater circuit is closed. In any time

173
interval before time t1 or after time t2 , the system behaves as if it is approaching a steady
state of constant temperature T1 (called the convergence temperature), which it would
eventually reach if the experiment were continued without closing the heater circuit. T1 is
greater than Text because of the energy transferred to the system by stirring and electrical
temperature measurement. By setting dU= dt and wel = dt equal to zero and T equal to
T1 in Eq. 7.3.22, we obtain wcont = dt D k.T1 Text /. We assume wcont = dt is constant.
Substituting this expression into Eq. 7.3.22 gives us a general expression for the rate at
which U changes in terms of the unknown quantities k and T1 :
dU
D
dt
k.T
T1 / C
wel
dt
(7.3.23)
(constant V )
This relation is valid throughout the experiment, not only while the heater circuit is closed.
If we multiply by dt and integrate from t1 to t2 , we obtain the internal energy change in the
time interval from t1 to t2 :
Z t2
(7.3.24)
U D k
.T T1 / dt C wel
t1
(constant V )
All the intermittent work wel is performed in this time interval.

The derivation of Eq. 7.3.24 is a general one. The equation can be applied also
to a isothermal-jacket calorimeter in which a reaction is occurring. Section 11.5.2
will mention the use of this equation for an internal energy correction of a reaction
calorimeter with an isothermal jacket.
The average value of the energy equivalent in the temperature range T1 to T2 is

Z t2
.k= /
.T T1 / dt C wel
U
t1
D
D
(7.3.25)
T2 T1
T2 T1
Solving for , we obtain
D
.T2
wel
Z
T1 / C .k= /
(7.3.26)
t2
.T
t1
T1 / dt
The value of wel is known from wel D I 2 Rel t , where t is the time interval during which
the heater circuit is closed. The integral can be evaluated numerically once T1 is known.
For heating at constant pressure, dH is equal to dU C p dV , and we can write
dU
dV
dH
D
Cp
D
dt
dt
dt
k.T
Text / C
wel
wcont
C
dt
dt
(7.3.27)
(constant p)
which is analogous to Eq. 7.3.22. By the procedure described above for the case of constant
V , we obtain
Z t2
H D k
(7.3.28)
.T T1 / dt C wel
t1
(constant p)

174
At constant p, the energy equivalent is equal to Cp D H=.T2 T1 /, and the nal expression for is the same as that given by Eq. 7.3.26.
To obtain values of k= and T1 for use in Eq. 7.3.26, we need the slopes of the heating
curve in time intervals (rating periods) just before t1 and just after t2 . Consider the case of
constant volume. In these intervals, wel = dt is zero and dU= dt equals k.T T1 / (from
Eq. 7.3.23). The heat capacity at constant volume is CV D dU= dT . The slope r in general
is then given by
dT
dT dU
rD
D
D .k= /.T T1 /
(7.3.29)
dt
dU dt
Applying this relation to the points at times t1 and t2 , we have the following simultaneous
equations in the unknowns k= and T1 :
r1 D
.k= /.T1
The solutions are

.k= / D
r1
T2
T1 /
r2 D
r2
T1
T1 D
Finally, k is given by
k D .k= / D
r1
T2
.k= /.T2
r1 T2
r1
r2
T1
r2 T1
r2
T1 /
(7.3.30)
(7.3.31)
(7.3.32)
When the pressure is constant, this procedure yields the same relations for k= , T1 , and k.
Continuous-ow calorimeters
A ow calorimeter is a third type of calorimeter used to measure the heat capacity of a uid
phase. The gas or liquid ows through a tube at a known constant rate past an electrical
heater of known constant power input. After a steady state has been achieved in the tube,
the temperature increase T at the heater is measured.
If wel = dt is the rate at which electrical work is performed (the electric power) and
dm= dt is the mass ow rate, then in time interval t a quantity w D .wel = dt/t of
work is performed on an amount n D .dm= dt /t =M of the uid (where M is the molar
mass). If heat ow is negligible, the molar heat capacity of the substance is given by
Cp;m D
w
M.wel = dt/
D
nT
T .dm= dt/
(7.3.33)
To correct for the effects of heat ow, T is usually measured over a range of ow rates
and the results extrapolated to innite ow rate.
7.3.3 Typical values

Figure 7.5 on the next page shows the temperature dependence of Cp;m for several substances. The discontinuities seen at certain temperatures occur at equilibrium phase transitions. At these temperatures the heat capacity is in effect innite, since the phase transition
of a pure substance involves nite heat with zero temperature change.

7.4 H EATING AT C ONSTANT VOLUME OR P RESSURE
175
H O(l)

m J K mol
N (l)
H O(g)
N (g)
N (s)
H O(s)
C(s)
Figure 7.5 Temperature dependence of molar heat capacity at constant pressure (p D

1 bar) of H2 O, N2 , and C(graphite).
7.4 HEATING AT CONSTANT VOLUME OR PRESSURE

Consider the process of changing the temperature of a phase at constant volume.6 The rate
of change of internal energy with T under these conditions is the heat capacity at constant
volume: CV D .@U=@T /V (Eq. 7.3.1). Accordingly, an innitesimal change of U is given
by
dU D CV dT
(7.4.1)
(closed system,
C D1, P D1, constant V )
and the nite change of U between temperatures T1 and T2 is

U D
T2
T1
CV dT
(7.4.2)
(closed system,
Three comments, relevant to these and other equations in this chapter, are in order:
1. Equation 7.4.2 allows us to calculate the nite change of a state function, U , by integrating CV over T . The equation was derived under the condition that V is constant
during the process, and the use of the integration variable T implies that the system
has a single, uniform temperature at each instant during the process. The integrand
6 Keeping
the volume exactly constant while increasing the temperature is not as simple as it may sound. Most
solids expand when heated, unless we arrange to increase the external pressure at the same time. If we use solid
walls to contain a uid phase, the container volume will change with temperature. For practical purposes, these
volume changes are usually negligible.

7.4 H EATING AT C ONSTANT VOLUME OR P RESSURE
176
CV may depend on both V and T , and we should integrate with V held constant and
CV treated as a function only of T .
2. Suppose we want to evaluate U for a process in which the volume is the same in the
initial and nal states (V2 D V1 ) but is different in some intermediate states, and the
temperature is not uniform in some of the intermediate states. We know the change
of a state function depends only on the initial and nal states, so we can still use Eq.
7.4.2 to evaluate U for this process. We integrate with V held constant, although V
was not constant during the actual process.
In general: A nite change X of a state function, evaluated under the condition
that another state function Y is constant, is the same as X under the less stringent
condition Y2 D Y1 . (Another application of this principle was mentioned in Sec.
4.6.2.)
3. For a pure substance, we may convert an expression for an innitesimal or nite
change of an extensive property to an expression for the change of the corresponding
molar property by dividing by n. For instance, Eq. 7.4.1 becomes
dUm D CV;m dT
and Eq. 7.4.2 becomes
Um D
(7.4.3)
T2
T1
CV;m dT
(7.4.4)
If, at a xed volume and over the temperature range T1 to T2 , the value of CV is essentially constant (i.e., independent of T ), Eq. 7.4.2 becomes
U D CV .T2
T1 /
(7.4.5)
P D1, constant V and CV )
An innitesimal entropy change during a reversible process in a closed system is given

according to the second law by dS D q=T . At constant volume, q is equal to dU which
in turn equals CV dT . Therefore, the entropy change is
dS D
CV
dT
T
(7.4.6)
(closed system,
Integration yields the nite change

S D
T2
T1
CV
dT
T
(7.4.7)
(closed system,
If CV is treated as constant, Eq. 7.4.7 becomes

S D CV ln
T2
T1
(7.4.8)
P D1, constant V and CV )

7.5 PARTIAL D ERIVATIVES WITH R ESPECT TO T, p, AND V
177
(More general versions of the two preceding equations have already been given in Sec.
4.6.1.)
Since CV is positive, we see from Eqs. 7.4.2 and 7.4.7 that heating a phase at constant
volume causes both U and S to increase.
We may derive relations for a temperature change at constant pressure by the same
methods. From Cp D .@H=@T /p (Eq. 7.3.2), we obtain
H D
T2
T1
Cp dT
(7.4.9)
(closed system,
C D1, P D1, constant p)
If Cp is treated as constant, Eq. 7.4.9 becomes

H D Cp .T2
T1 /
(7.4.10)
P D1, constant p and Cp )
From dS D q=T and Eq. 7.3.2 we obtain for the entropy change at constant pressure
dS D
Cp
dT
T
(7.4.11)
(closed system,
Integration gives
S D
T2
T1
Cp
dT
T
(7.4.12)
(closed system,
or, with Cp treated as constant,

S D Cp ln
T2
T1
(7.4.13)
P D1, constant p and Cp )
Cp is positive, so heating a phase at constant pressure causes H and S to increase.

The Gibbs energy changes according to .@G=@T /p D S (Eq. 5.4.11), so heating at
constant pressure causes G to decrease.
7.5 PARTIAL DERIVATIVES WITH RESPECT TO T, p, AND V

7.5.1 Tables of partial derivatives
The tables in this section collect useful expressions for partial derivatives of the eight state
functions T , p, V , U , H , A, G, and S in a closed, single-phase system. Each derivative is
taken with respect to one of the three easily-controlled variables T , p, or V while another

178
Table 7.1 Constant temperature: expressions for partial derivatives of state

functions with respect to pressure and volume in a closed, single-phase system
Partial
derivative
@p
@V T
@V
@p T
@U
@p T
@U
@V T
@H
@p T
@H
@V T
General
expression
Ideal
gas
1
T V
p
V
V
p
T V
. T C T p/V
T
T
.1
T /V
T 1
T
Partial
derivative

@A
@p T
@A
@V T
@G
@p T
@G
@V
T
@S
@p T
@S
@V T
General
Ideal
expression gas
T pV
p
V
p
V
1
T
V
T
p
T
of these variables is held constant. We have already seen some of these expressions, and the
derivations of the others are indicated below.
We can use these partial derivatives (1) for writing an expression for the total differential of any of the eight quantities, and (2) for expressing the nite change in one of these
quantities as an integral under conditions of constant T , p, or V . For instance, given the
expressions

Cp
@S
@S
D
and
D V
(7.5.1)
@T p
T
@p T
we may write the total differential of S , taking T and p as the independent variables, as
Cp
dT V dp
T
Furthermore, the rst expression is equivalent to the differential form
dS D
(7.5.2)
Cp
dT
(7.5.3)
T
provided p is constant; we can integrate this equation to obtain the nite change S under
isobaric conditions as shown in Eq. 7.4.12.
Both general expressions and expressions valid for an ideal gas are given in Tables 7.1,
7.2, and 7.3.
dS D
We may derive the general expressions as follows. We are considering differentiation with respect only to T , p, and V . Expressions for .@V =@T /p , .@V =@p/T , and
.@p=@T /V come from Eqs. 7.1.1, 7.1.2, and 7.1.7 and are shown as functions of
and T . The reciprocal of each of these three expressions provides the expression for
another partial derivative from the general relation
.@y=@x/z D
1
.@x=@y/z
(7.5.4)

179
Table 7.2 Constant pressure: expressions for partial derivatives of state functions with respect to temperature and volume in a closed, single-phase system
Partial
derivative
@T
@V p
@V
@T p
@U
@T p
@U
@V p
@H
@T p
@H
@V p
General
expression
Ideal
gas
1
V
T
V
V
T
V
Cp
pV
Cp
V
Partial
General
derivative expression

@A
pV S
@T p
@A
S
p
@V p
V
@G
S
@T p
@G
S
@V p
V

Cp
@S
@T p
T
Cp
@S
@V p
T V
CV
CV T
V
Cp
Cp
Cp
V
Cp T
V
Ideal
gas
pV
T
p
S
TS
V
S
TS
V
Cp
T
Cp
V
Table 7.3 Constant volume: expressions for partial derivatives of state functions with
respect to temperature and pressure in a closed, single-phase system
Partial
derivative

@T
@p
V
@p
@T V
@U
@T V
@U
@p V
@H
@T V
@H
@p V
General
expression
Ideal
gas
T
p
p
T
CV
CV
T Cp
Cp C
T CV
p
T V
V
.1
T
T Cp
C V .1
T /
T /
Cp
Cp T
p
Partial
General
derivative expression

@A
S
@T
V
@A
T S
@p V
@G
V
S
@T V
T
@G
T S
V
@p V

@S
CV
@T
T
V
T Cp
@S
V
@p V
T
Ideal
gas
S
TS
p
pV
S
T
TS
V
p
CV
T
CV
p
This procedure gives us expressions for the six partial derivatives of T , p, and V .
The remaining expressions are for partial derivatives of U , H , A, G, and S. We
obtain the expression for .@U=@T /V from Eq. 7.3.1, for .@U=@V /T from Eq. 7.2.4,
for .@H=@T /p from Eq. 7.3.2, for .@A=@T /V from Eq. 5.4.9, for .@A=@V /T from Eq.
5.4.10, for .@G=@p/T from Eq. 5.4.12, for .@G=@T /p from Eq. 5.4.11, for .@S=@T /V
from Eq. 7.4.6, for .@S=@T /p from Eq. 7.4.11, and for .@S=@p/T from Eq. 5.4.18.
We can transform each of these partial derivatives, and others derived in later
steps, to two other partial derivatives with the same variable held constant and the
variable of differentiation changed. The transformation involves multiplying by an

180
appropriate partial derivative of T , p, or V . For instance, from the partial derivative

.@U=@V /T D .T =T / p, we obtain
@U
@U
@V
T
D
D
p . T V / D . T C T p/ V (7.5.5)
@p T
@V T @p T
T
The remaining partial derivatives can be found by differentiating U D H pV , H D
U C pV , A D U T S , and G D H T S and making appropriate substitutions.
Whenever a partial derivative appears in a derived expression, it is replaced with an
expression derived in an earlier step. The expressions derived by these steps constitute
the full set shown in Tables 7.1, 7.2, and 7.3.
Bridgman7 devised a simple method to obtain expressions for these and many
other partial derivatives from a relatively small set of formulas.
7.5.2 The JouleThomson coefcient

The JouleThomson coefcient of a gas was dened in Eq. 6.3.3 on page 158 by JT D
.@T =@p/H . It can be evaluated with measurements of T and p during adiabatic throttling
processes as described in Sec. 6.3.1.
To relate JT to other properties of the gas, we write the total differential of the enthalpy
of a closed, single-phase system in the form
@H
@H
dH D
dT C
dp
(7.5.6)
@T p
@p T
and divide both sides by dp:
dH
D
dp
@H
@T
dT
C
dp
@H
@p
(7.5.7)
Next we impose a condition of constant H ; the ratio dT = dp becomes a partial derivative:
@H
@T
@H
0D
C
(7.5.8)
@T p @p H
@p T
Rearrangement gives
@T
@p
.@H=@p/T
.@H=@T /p
(7.5.9)
The left side of this equation is the JouleThomson coefcient. An expression for the partial
derivative .@H=@p/T is given in Table 7.1, and the partial derivative .@H=@T /p is the heat
capacity at constant pressure (Eq. 5.6.3). These substitutions give us the desired relation
JT
7 Ref.
.T
1/V
Cp
.T 1/Vm
Cp;m
(7.5.10)
[22]; Ref. [21], p. 199241.

7.6 I SOTHERMAL P RESSURE C HANGES
181
Table 7.4 Changes of state functions during an isothermal pressure change in a

closed, single-phase system
State function
change
U
H
A
G
S
General expression
Z p2
. T C T p/V dp
p1
Z p2
.1 T /V dp
Zp1 2
p
T pV dp
Zp1 2
p
V dp
p1
Z p2
V dp
p1
Ideal gas
0
0
p2
p1
p
nRT ln 2
p1
p
nR ln 2
p1
nRT ln
Approximate expression
for liquid or solid
T Vp
.1
T /Vp
2
T V .p2
2
p1 /=2
Vp
Vp
7.6 ISOTHERMAL PRESSURE CHANGES

In various applications, we will need expressions for the effect of changing the pressure at
constant temperature on the internal energy, enthalpy, entropy, and Gibbs energy of a phase.
We obtain the expressions by integrating expressions found in Table 7.1. For example, U
R
is given by .@U=@p/T dp. The results are listed in the second column of Table 7.4.
7.6.1 Ideal gases

Simplications result when the phase is an ideal gas. In this case, we can make the substitutions V D nRT =p, D 1=T , and T D 1=p, resulting in the expressions in the third
column of Table 7.4.
The expressions in the third column of Table 7.4 may be summarized by the statement
that, when an ideal gas expands isothermally, the internal energy and enthalpy stay constant,
the entropy increases, and the Helmholtz energy and Gibbs energy decrease.
7.6.2 Condensed phases

Solids, and liquids under conditions of temperature and pressure not close to the critical
point, are much less compressible than gases. Typically the isothermal compressibility,
T , of a liquid or solid at room temperature and atmospheric pressure is no greater than
1 10 4 bar 1 (see Fig. 7.2 on page 166), whereas an ideal gas under these conditions has
T D 1=p D 1 bar 1 . Consequently, it is frequently valid to treat V for a liquid or solid
as essentially constant during a pressure change at constant temperature. Because T is
small, the product Tp for a liquid or solid is usually much smaller than the product T .
Furthermore, T for liquids and solids does not change rapidly with p as it does for gases,
and neither does .
With the approximations that V , , and T are constant during an isothermal pressure
change, and that Tp is negligible compared with T , we obtain the expressions in the last

7.7 S TANDARD S TATES OF P URE S UBSTANCES
182
column of Table 7.4.
7.7 STANDARD STATES OF PURE SUBSTANCES

It is often useful to refer to a reference pressure, the standard pressure, denoted p . The
standard pressure has an arbitrary but constant value in any given application. Until 1982,
chemists used a standard pressure of 1 atm (1:01325 105 Pa). The IUPAC now recommends the value p D 1 bar (exactly 105 Pa). This book uses the latter value unless stated
otherwise. (Note that there is no dened standard temperature.)
A superscript degree symbol () denotes a standard quantity or standard-state conditions. An alternative symbol for this purpose, used extensively outside the U.S., is a superscript Plimsoll mark ( ).8
A standard state of a pure substance is a particular reference state appropriate for the
kind of phase and is described by intensive variables. This book follows the recommendations of the IUPAC Green Book9 for various standard states.
The standard state of a pure gas is the hypothetical state in which the gas is at pressure
p and the temperature of interest, and the gas behaves as an ideal gas. The molar
volume of a gas at 1 bar may have a measurable deviation from the molar volume
predicted by the ideal gas equation due to intermolecular forces. We must imagine
the standard state in this case to consist of the gas with the intermolecular forces
magically turned off and the molar volume adjusted to the ideal-gas value RT =p .
The standard state of a pure liquid or solid is the unstressed liquid or solid at pressure p and the temperature of interest. If the liquid or solid is stable under these
conditions, this is a real (not hypothetical) state.
Section 9.7 will introduce additional standard states for constituents of mixtures.
7.8 CHEMICAL POTENTIAL AND FUGACITY

The chemical potential, , of a pure substance has as one of its denitions (page 142)
def
D Gm D
G
n
(7.8.1)
(pure substance)
That is, is equal to the molar Gibbs energy of the substance at a given temperature and
pressure. (Section 9.2.6 will introduce a more general denition of chemical potential that
applies also to a constituent of a mixture.) The chemical potential is an intensive state
function.
The total differential of the Gibbs energy of a xed amount of a pure substance in a
single phase, with T and p as independent variables, is dG D S dT C V dp (Eq. 5.4.4).
8 The Plimsoll mark is named after the British merchant Samuel Plimsoll, at whose instigation Parliament passed
an act in 1875 requiring the symbol to be placed on the hulls of cargo ships to indicate the maximum depth for
safe loading.
9 Ref. [36], p. 6162.

7.8 C HEMICAL P OTENTIAL AND F UGACITY
183
Dividing both sides of this equation by n gives the total differential of the chemical potential
with these same independent variables:
d D
Sm dT C Vm dp
(7.8.2)
(Since all quantities in this equation are intensive, it is not necessary to specify a closed
system; the amount of the substance in the system is irrelevant.)
We identify the coefcients of the terms on the right side of Eq. 7.8.2 as the partial
derivatives

@
(7.8.3)
D Sm
@T p
and
@
@p
D Vm
(7.8.4)
Since Vm is positive, Eq. 7.8.4 shows that the chemical potential increases with increasing
pressure in an isothermal process.
The standard chemical potential, , of a pure substance in a given phase and at a
given temperature is the chemical potential of the substance when it is in the standard state
of the phase at this temperature and the standard pressure p .
There is no way we can evaluate the absolute value of at a given temperature and
pressure, or of at the same temperature,10 but we can measure or calculate the difference
. The general procedure is to integrate d D V dp (Eq. 7.8.2 with dT set equal to
m
zero) from the standard state at pressure p to the experimental state at pressure p 0 :
Z p0
0
Vm dp
(7.8.5)
.p /
D
p
(constant T )
7.8.1 Gases
For the standard chemical potential of a gas, this book will usually use the notation (g)
to emphasize the choice of a gas standard state.
An ideal gas is in its standard state at a given temperature when its pressure is the
standard pressure. We nd the relation of the chemical potential of an ideal gas to its
pressure and its standard chemical potential at the same temperature by setting Vm equal to
R 0
D p .RT =p/ dp D RT ln.p 0 =p /. The general relation
RT =p in Eq. 7.8.5: .p 0 /
p
for as a function of p, then, is
D
(g) C RT ln
p
p
(7.8.6)
(pure ideal gas, constant T )
This function is shown as the dashed curve in Fig. 7.6 on the next page.
least not to any useful degree of precision. The values of and include the molar internal energy
whose absolute value can only be calculated from the Einstein relation; see Sec. 2.6.2.
10 At

184
(g)
Figure 7.6 Chemical potential as a function of pressure at constant temperature, for

a real gas (solid curve) and the same gas behaving ideally (dashed curve). Point A is
the gas standard state. Point B is a state of the real gas at pressure p 0 . The fugacity
f .p 0 / of the real gas at pressure p 0 is equal to the pressure of the ideal gas having the
same chemical potential as the real gas (point C).
If a gas is not an ideal gas, its standard state is a hypothetical state. The fugacity, f , of
a real gas (a gas that is not necessarily an ideal gas) is dened by an equation with the same
form as Eq. 7.8.6:
D
(g) C RT ln
f
p
(7.8.7)
(pure gas)
or
def
f D p exp
(g)
RT
(7.8.8)
(pure gas)
Note that fugacity has the dimensions of pressure. Fugacity is a kind of effective pressure.
Specically, it is the pressure that the hypothetical ideal gas (the gas with intermolecular
forces turned off) would need to have in order for its chemical potential at the given
temperature to be the same as the chemical potential of the real gas (see point C in Fig. 7.6).
If the gas is an ideal gas, its fugacity is equal to its pressure.
To evaluate the fugacity of a real gas at a given T and p, we must relate the chemical
potential to the pressurevolume behavior. Let 0 be the chemical potential and f 0 be
the fugacity at the pressure p 0 of interest; let 00 be the chemical potential and f 00 be the
fugacity of the same gas at some low pressure p 00 (all at the same temperature). Then we
(g) D RT ln.f 0 =p / and 00
(g) D RT ln.f 00 =p /, from
use Eq. 7.8.5 to write 0
which we obtain
f0
0
00
D RT ln 00
(7.8.9)
f

185
By integrating d D Vm dp from pressure p 00 to pressure p 0 , we obtain

0
Equating the two expressions for
00
00
ln
d D
00
p0
p 00
Vm dp
(7.8.10)
and dividing by RT gives
f0
D
f 00
p0
p 00
Vm
dp
RT
(7.8.11)
In principle, we could use the integral on the right side of Eq. 7.8.11 to evaluate f 0 by
choosing the lower integration limit p 00 to be such a low pressure that the gas behaves as
an ideal gas and replacing f 00 by p 00 . However, because the integrand Vm =RT becomes
very large at low pressure, the integral is difcult to evaluate. We avoid this difculty by
subtracting from the preceding equation the identity
p0
ln 00 D
p
p0
p 00
dp
p
(7.8.12)
which is simply the result of integrating the function 1=p from p 00 to p 0 . The result is
f 0 p 00
ln 00 0 D
f p
p0
p 00
Vm
RT
1
dp
p
(7.8.13)
Now we take the limit of both sides of Eq. 7.8.13 as p 00 approaches zero. In this limit, the gas
at pressure p 00 approaches ideal-gas behavior, f 00 approaches p 00 , and the ratio f 0 p 00 =f 00 p 0
approaches f 0 =p 0 :
Z p0
f0
Vm
1
ln 0 D
dp
(7.8.14)
p
RT
p
0
The integrand .Vm =RT 1=p/ of this integral approaches zero at low pressure, making it
feasible to evaluate the integral from experimental data.
The fugacity coefcient of a gas is dened by
def
f
p
or
f D p
(7.8.15)
(pure gas)
The fugacity coefcient at pressure p 0 is then given by Eq. 7.8.14:

0
ln .p / D
p0
0
Vm
RT
1
dp
p
(7.8.16)
(pure gas, constant T )
The isothermal behavior of real gases at low to moderate pressures (up to at least 1 bar)
is usually adequately described by a two-term equation of state of the form given in Eq.
2.2.8:
RT
CB
(7.8.17)
Vm
p

7.9 S TANDARD M OLAR Q UANTITIES OF A G AS
186
Here B is the second virial coefcient, a function of T . With this equation of state, Eq.
7.8.16 becomes
Bp
(7.8.18)
ln
RT
For a real gas at temperature T and pressure p, Eq. 7.8.16 or 7.8.18 allows us to evaluate
the fugacity coefcient from an experimental equation of state or a second virial coefcient.
We can then nd the fugacity from f D p.
As we will see in Sec. 9.7, the dimensionless ratio D f =p is an example of an
activity coefcient and the dimensionless ratio f =p is an example of an activity.
7.8.2 Liquids and solids

The dependence of the chemical potential on pressure at constant temperature is given by
Eq. 7.8.5. With an approximation of zero compressibility, this becomes
C Vm .p
p/
(7.8.19)
(pure liquid or solid,
constant T )
7.9 STANDARD MOLAR QUANTITIES OF A GAS

A standard molar quantity of a substance is the molar quantity in the standard state at the
temperature of interest. We have seen (Sec. 7.7) that the standard state of a pure liquid or
solid is a real state, so any standard molar quantity of a pure liquid or solid is simply the
molar quantity evaluated at the standard pressure and the temperature of interest.
The standard state of a gas, however, is a hypothetical state in which the gas behaves
ideally at the standard pressure without inuence of intermolecular forces. The properties
of the gas in this standard state are those of an ideal gas. We would like to be able to relate
molar properties of the real gas at a given temperature and pressure to the molar properties
in the standard state at the same temperature.
We begin by using Eq. 7.8.7 to write an expression for the chemical potential of the real
gas at pressure p 0 :
f .p 0 /
p
p0
f .p 0 /
D (g) C RT ln C RT ln
p
p0
.p 0 / D
(g) C RT ln
(7.9.1)
We then substitute from Eq. 7.8.14 to obtain a relation between the chemical potential, the
standard chemical potential, and measurable properties, all at the same temperature:
Z p0
p0
RT
0
.p / D (g) C RT ln C
dp
Vm
(7.9.2)
p
p
0
(pure gas)
Note that this expression for is not what we would obtain by simply integrating d D
Vm dp from p to p 0 , because the real gas is not necessarily in its standard state of ideal-gas
behavior at a pressure of 1 bar.

187
Table 7.5 Real gases: expressions for differences between molar properties and standard molar
values at the same temperature
General expression
at pressure p 0
#
Z p0 "
@Vm
Um (g)
Vm T
dp C RT p 0 Vm
@T p
0
#
Z p0 "
@Vm
Vm T
dp
Hm (g)
@T p
0
Z p0
p0
RT
A (g)
RT ln C
Vm
dp C RT p 0 Vm
m
p
p
0
Z p0
RT
p0
Vm
dp
Gm (g)
RT ln C
p
p
0
#
Z p 0 "
@Vm
p0
R
Sm (g)
R ln
dp
p
@T p p
0
Z p0 2
@ Vm
T
dp
Cp;m (g)
@T 2 p
0
Difference
Um
Hm
Am
Gm
Sm
Cp;m
Equation of state
V D nRT =p C nB
pT
dB
dT
p B
dB
T
dT
RT ln
p
p
RT ln
p
C Bp
p
R ln
pT
p
p
dB
dT
d2 B
dT 2
Recall that the chemical potential of a pure substance is also its molar Gibbs energy
Gm D G=n. The standard chemical potential (g) of the gas is the standard molar Gibbs
energy, Gm (g). Therefore Eq. 7.9.2 can be rewritten in the form
Z p0
p0
RT
0
dp
(7.9.3)
Gm .p / D Gm (g) C RT ln C
Vm
p
p
0
The middle column of Table 7.5 contains an expression for Gm .p 0 / Gm (g) taken from this
equation. This expression contains all the information needed to nd a relation between any
other molar property and its standard molar value in terms of measurable properties. The
way this can be done is as follows.
The relation between the chemical potential of a pure substance and its molar entropy
is given by Eq. 7.8.3:

@
Sm D
(7.9.4)
@T p
The standard molar entropy of the gas is found from Eq. 7.9.4 by changing

@ (g)
Sm (g) D
@T
p
By substituting the expression for
result with Eq. 7.9.5, we obtain
Sm .p 0 / D Sm (g)
to
(g):
(7.9.5)
given by Eq. 7.9.2 into Eq. 7.9.4 and comparing the

p0
R ln
p
p0
0
"
@Vm
@T
#
R
dp
p
(7.9.6)

188
The expression for Sm Sm (g) in the middle column of Table 7.5 comes from this equation.
The equation, together with a value of Sm for a real gas obtained by the calorimetric method
described in Sec. 6.2.1, can be used to evaluate Sm (g).

Now we can use the expressions for Gm and Sm to nd expressions for molar quantities
such as Hm and Cp;m relative to the respective standard molar quantities. The general
procedure for a molar quantity Xm is to write an expression for Xm as a function of Gm and
Sm and an analogous expression for Xm (g) as a function of Gm (g) and Sm (g). Substitutions
for Gm and Sm from Eqs. 7.9.3 and 7.9.6 are then made in the expression for Xm , and the
difference Xm Xm (g) taken.
For example, the expression for Um Um (g) in the middle column Table 7.5 was derived
as follows. The equation dening the Gibbs energy, G D U T S C pV , was divided by
the amount n and rearranged to
Um D Gm C T Sm
pVm
(7.9.7)
The standard-state version of this relation is
Um (g) D Gm (g) C T Sm (g)
p Vm (g)
(7.9.8)
where from the ideal gas law p Vm (g) can be replaced by RT . Substitutions from Eqs. 7.9.3
and 7.9.6 were made in Eq. 7.9.7 and the expression for Um (g) in Eq. 7.9.8 was subtracted,
resulting in the expression in the table.
For a real gas at low to moderate pressures, we can approximate Vm by .RT =p/ C B
where B is the second virial coefcient (Eq. 7.8.17). Equation 7.9.2 then becomes
(g) C RT ln
p
C Bp
p
(7.9.9)
The expressions in the last column of Table 7.5 use this equation of state. We can see what
the expressions look like if the gas is ideal simply by setting B equal to zero. They show that
when the pressure of an ideal gas increases at constant temperature, Gm and Am increase,
Sm decreases, and Um , Hm , and Cp;m are unaffected.

P ROBLEMS
189
PROBLEMS
in Appendix I.
7.1 Derive the following relations from the denitions of , T , and :

1 @
1 @
D
T D
@T p
@p T
7.2 Use equations in this chapter to derive the following expressions for an ideal gas:
D 1=T
T D 1=p
7.3 For a gas with the simple equation of state

Vm D
RT
CB
p
(Eq. 2.2.8), where B is the second virial coefcient (a function of T ), nd expressions for ,
T , and .@Um =@V /T in terms of dB= dT and other state functions.
7.4 Show that when the virial equation pVm D RT .1 C Bp p C Cp p 2 C / (Eq. 2.2.3) adequately
represents the equation of state of a real gas, the JouleThomson coefcient is given by
JT
RT 2 dBp = dT C .dCp = dT /p C
Cp;m
Note that the limiting value at low pressure, RT 2 .dBp = dT /=Cp;m , is not necessarily equal to
zero even though the equation of state approaches that of an ideal gas in this limit.
7.5 The quantity .@T =@V /U is called the Joule coefcient. James Joule attempted to evaluate
this quantity by measuring the temperature change accompanying the expansion of air into a
vacuumthe Joule experiment. Write an expression for the total differential of U with T
and V as independent variables, and by a procedure similar to that used in Sec. 7.5.2 show that
the Joule coefcient is equal to
p T =T
CV
7.6 pV T data for several organic liquids were measured by Gibson and Loefer.11 The following formulas describe the results for aniline.
Molar volume as a function of temperature at p D 1 bar (298358 K):
Vm D a C bT C cT 2 C d T 3
where the parameters have the values
a D 69:287 cm3 mol 1
b D 0:08852 cm3 K 1 mol
c D 1:0443 10 4 cm3 K 2 mol

d D 1:940 10 7 cm3 K 3 mol 1
Molar volume as a function of pressure at T D 298:15 K (11000 bar):

Vm D e
f ln.g C p=bar/
where the parameter values are

e D 156:812 cm3 mol
11 Ref.
f D 8:5834 cm3 mol
g D 2006:6
[65].

P ROBLEMS
190
(a) Use these formulas to evaluate , T , .@p=@T /V , and .@U=@V /T (the internal pressure)
for aniline at T D 298:15 K and p D 1:000 bar.
(b) Estimate the pressure increase if the temperature of a xed amount of aniline is increased
by 0:10 K at constant volume.
7.7 (a) From the total differential of H with T and p as independent variables, derive the relation
.@Cp;m =@p/T D T .@2 Vm =@T 2 /p .
(b) Evaluate .@Cp;m =@p/T for liquid aniline at 300:0 K and 1 bar using data in Prob. 7.6.
7.8 (a) From the total differential of V with T and p as independent variables, derive the relation
.@=@p/T D .@T =@T /p .
(b) Use this relation to estimate the value of for benzene at 25 C and 500 bar, given that
the value of is 1:2 10 3 K 1 at 25 C and 1 bar. (Use information from Fig. 7.2 on
page 166.)
7.9 Certain equations of state supposed to be applicable to nonpolar liquids and gases are of the
2
form p D Tf .Vm / a=Vm , where f .Vm / is a function of the molar volume only and a is a
constant.
2
(a) Show that the van der Waals equation of state .p C a=Vm /.Vm
b are constants) is of this form.
b/ D RT (where a and
(b) Show that any uid with an equation of state of this form has an internal pressure equal
2
to a=Vm .
7.10 Suppose that the molar heat capacity at constant pressure of a substance has a temperature
dependence given by Cp;m D a C bT C cT 2 , where a, b, and c are constants. Consider the
heating of an amount n of the substance from T1 to T2 at constant pressure. Find expressions
for H and S for this process in terms of a, b, c, n, T1 , and T2 .
7.11 At p D 1 atm, the molar heat capacity at constant pressure of aluminum is given by
Cp;m D a C bT
where the constants have the values
a D 20:67 J K
mol
b D 0:01238 J K
mol
Calculate the quantity of electrical work needed to heat 2:000 mol of aluminum from 300:00 K
to 400:00 K at 1 atm in an adiabatic enclosure.
7.12 The temperature dependence of the standard molar heat capacity of gaseous carbon dioxide in
the temperature range 298 K2000 K is given by
Cp;m D a C bT C
c
T2

a D 44:2 J K
mol
b D 8:8
10
JK
mol
cD
8:6
105 J K mol
Calculate the enthalpy and entropy changes when one mole of CO2 is heated at 1 bar from
300:00 K to 800:00 K. You can assume that at this pressure Cp;m is practically equal to Cp;m .
7.13 This problem concerns gaseous carbon dioxide. At 400 K, the relation between p and Vm at
pressures up to at least 100 bar is given to good accuracy by a virial equation of state truncated

P ROBLEMS
191
at the second virial coefcient, B. In the temperature range 300 K800 K the dependence of B
on temperature is given by
B D a0 C b 0 T C c 0 T 2 C d 0 T 3
a0 D
521 cm3 mol
b 0 D 2:08 cm3 K 1 mol 1

c 0 D 2:89 10 3 cm3 K 2 mol 1
d 0 D 1:397 10 6 cm3 K 3 mol 1
(a) From information in Prob. 7.12, calculate the standard molar heat capacity at constant
pressure, Cp;m , at T D 400:0 K.

(b) Estimate the value of Cp;m under the conditions T D 400:0 K and p D 100:0 bar.
7.14 A chemist, needing to determine the specic heat capacity of a certain liquid but not having an
electrically heated calorimeter at her disposal, used the following simple procedure known as
drop calorimetry. She placed 500:0 g of the liquid in a thermally insulated container equipped
with a lid and a thermometer. After recording the initial temperature of the liquid, 24:80 C,
she removed a 60:17-g block of aluminum metal from a boiling water bath at 100:00 C and
quickly immersed it in the liquid in the container. After the contents of the container had
become thermally equilibrated, she recorded a nal temperature of 27:92 C. She calculated
the specic heat capacity Cp =m of the liquid from these data, making use of the molar mass of
aluminum (M D 26:9815 g mol 1 ) and the formula for the molar heat capacity of aluminum
given in Prob. 7.11.
(a) From these data, nd the specic heat capacity of the liquid under the assumption that its
value does not vary with temperature. Hint: Treat the temperature equilibration process
as adiabatic and isobaric (H D 0), and equate H to the sum of the enthalpy changes
in the two phases.
(b) Show that the value obtained in part (a) is actually an average value of Cp =m over the
temperature range between the initial and nal temperatures of the liquid given by
Z T2
.Cp =m/ dT
T1
T2
T1
7.15 Suppose a gas has the virial equation of state pVm D RT .1 C Bp p C Cp p 2 /, where Bp and
Cp depend only on T , and higher powers of p can be ignored.
(a) Derive an expression for the fugacity coefcient, , of this gas as a function of p.
(b) For CO2 (g) at 0:00 C, the virial coefcients have the values Bp D 6:67 10 3 bar
and Cp D 3:4 10 5 bar 2 . Evaluate the fugacity f at 0:00 C and p D 20:0 bar.
7.16 Table 7.6 on the next page lists values of the molar volume of gaseous H2 O at 400:00 C and
12 pressures.
(a) Evaluate the fugacity coefcient and fugacity of H2 O(g) at 400:00 C and 200 bar.
(b) Show that the second virial coefcient B in the virial equation of state, pVm D RT .1 C
2
B=Vm C C =Vm C /, is given by
1
Vm
B D RT lim
p!0 RT
p
where the limit is taken at constant T . Then evaluate B for H2 O(g) at 400:00 C.

P ROBLEMS
192
Table 7.6 Molar volume of H2 O(g) at 400:00 C a

p=105 Pa
1
10
20
40
60
80
a based
Vm =10
m3 mol
55:896
5:5231
2:7237
1:3224
0:85374
0:61817
p=105 Pa
100
120
140
160
180
200
Vm =10
m3 mol
0:47575
0:37976
0:31020
0:25699
0:21447
0:17918
on data in Ref. [70]
C HAPTER 8
P HASE T RANSITIONS AND E QUILIBRIA OF

P URE S UBSTANCES
A system of two or more phases of a single substance, in the absence of internal constraints,
is in an equilibrium state when each phase has the same temperature, the same pressure, and
the same chemical potential. This chapter describes the derivation and consequences of this
simple principle, the general appearance of phase diagrams of single-substance systems,
and quantitative aspects of the equilibrium phase transitions of these systems.
8.1 PHASE EQUILIBRIA

8.1.1 Equilibrium conditions
If the state of an isolated system is an equilibrium state, this state does not change over
time (Sec. 2.4.4). We expect an isolated system that is not in an equilibrium state to undergo a spontaneous, irreversible process and eventually to reach an equilibrium state. Just
how rapidly this process occurs is a matter of kinetics, not thermodynamics. During this
irreversible adiabatic process, the entropy increases until it reaches a maximum in the equilibrium state.
A general procedure will now be introduced for nding conditions for equilibrium with
given constraints. The procedure is applied to phase equilibria of single-substance, multiphase systems in the next section, to transfer equilibria in multicomponent, multiphase
systems in Sec. 9.2.7, and to reaction equilibria in Sec. 11.7.3.
The procedure has ve steps:
1. Write an expression for the total differential of the internal energy U consistent with
any constraints and with the number of independent variables of the system.
2. Impose conditions of isolation for the system, including dU D 0, thereby reducing
the number of independent variables.
3. Designate a particular phase, 0 , as a reference phase and make the substitution
P
P
0
dS D dS
0 dS . (This is valid because entropy is extensive: S D
S ,
P
dS D dS .)
4. Rearrange to obtain an expression for the total differential of the entropy consistent
with the reduced number of independent variables.
193
CHAPTER 8 PHASE TRANSITIONS AND EQUILIBRIA OF PURE SUBSTANCES

8.1 P HASE E QUILIBRIA
194
5. The conditions for an equilibrium state are those that make the innitesimal entropy
change, dS, equal to zero for all innitesimal changes of the independent variables
of the isolated system.
8.1.2 Equilibrium in a multiphase system

In this section we consider a system of a single substance in two or more uniform phases
with distinctly different intensive properties. For instance, one phase might be a liquid and
another a gas. We assume the phases are not separated by internal partitions, so that there is
no constraint preventing the transfer of matter and energy among the phases. (A tall column
of gas in a gravitational eld is a different kind of system in which intensive properties of
an equilibrium state vary continuously with elevation; this case will be discussed in Sec.
8.1.4.)
Phase 0 P be the reference phase. Since internal energy is extensive, we can write
will
P
0 C
0 C
U DU
0 dU . We assume any changes are slow
0 U and dU D dU
enough to allow each phase to be practically uniform at all times. Treating each phase as an
open subsystem with expansion work only, we use the relation dU D T dS p dV C dn
(Eq. 5.2.5) to replace each dU term:
0
dU D .T dS p dV C dn /
X
C
.T dS p dV C dn /
(8.1.1)
This is an expression for the total differential of U when there are no constraints.
We isolate the system by enclosing it in a rigid, stationary adiabatic container. The
constraints needed to isolate the system, then, are given by the relations
dU D 0
dV
(8.1.2)
dV D 0
(no expansion work)
(8.1.3)
dn D 0
(closed system)
(8.1.4)
dn C
(constant internal energy)
Each of these relations is an independent restriction that reduces the number of independent
0
0
variables by one. When we substitute expressions for dU , dV , P dn from these reand
0
lations into Eq. 8.1.1, make the further substitution dS D dS

0 dS , and collect
term with the same differentials on the right side, we obtain
X
X
0
0
0
0 D T dS C
.T T / dS
.p p / dV
C
/ dn
(8.1.5)

195
Solving for dS, we obtain

dS D
X T 0 T
dS
0
T
0
X
0
T 0
X p 0 p
dV
0
T
0
dn
(8.1.6)
This is an expression for the total differential of S in the isolated system.

In an isolated system, an equilibrium state cannot change spontaneously to a different state. Once the isolated system has reached an equilibrium state, an imagined nite
change of any of the independent variables consistent with the constraints (a so-called virtual displacement) corresponds to an impossible process with an entropy decrease. Thus,
the equilibrium state has the maximum entropy that is possible for the isolated system. In
order for S to be a maximum, dS must be zero for an innitesimal change of any of the
independent variables of the isolated system.
This requirement is satised in the case of the multiphase system only if the coefcient
of each term in the sums on the right side of Eq. 8.1.6 is zero. Therefore, in an equilibrium
0
state the temperature of each phase is equal to the temperature T of the reference phase,
0
the pressure of each phase is equal to p , and the chemical potential in each phase is equal
0
to . That is, at equilibrium the temperature, pressure, and chemical potential are uniform
throughout the system. These are, respectively, the conditions described in Sec. 2.4.4 of
thermal equilibrium, mechanical equilibrium, and transfer equilibrium. These conditions
must hold in order for a multiphase system of a pure substance without internal partitions to
be in an equilibrium state, regardless of the process by which the system attains that state.
8.1.3 Simple derivation of equilibrium conditions

Here is a simpler, less formal derivation of the three equilibrium conditions in a multiphase
system of a single substance.
It is intuitively obvious that, unless there are special constraints (such as internal partitions), an equilibrium state must have thermal and mechanical equilibrium. A temperature
difference between two phases would cause a spontaneous transfer of heat from the warmer
to the cooler phase; a pressure difference would cause spontaneous ow of matter.
When some of the substance is transferred from one phase to another under conditions
of constant T and p, the intensive properties of each phase remains the same including the
chemical potential. The chemical potential of a pure phase is the Gibbs energy per amount
of substance in the phase. We know that in a closed system of constant T and p with
expansion work only, the total Gibbs energy decreases during a spontaneous process and
is constant during a reversible process (Eq. 5.8.6). The Gibbs energy will decrease only
if there is a transfer of substance from a phase of higher chemical potential to a phase of
lower chemical potential, and this will be a spontaneous change. No spontaneous transfer
is possible if both phases have the same chemical potential, so this is a condition for an
equilibrium state.

196
Figure 8.1 Closed system of constant-volume slab-shaped uid phases stacked in the
vertical direction. The shaded phase is reference phase 0 .
8.1.4 Tall column of gas in a gravitational eld

The earths gravitational eld is an example of an external force eld that acts on a system
placed in it. Usually we ignore its effects on the state of the system. If, however, the
systems vertical extent is considerable we must take the presence of the eld into account
to explain, for example, why gas pressure varies with elevation in an equilibrium state.
A tall column of gas whose intensive properties are a function of elevation may be
treated as an innite number of uniform phases, each of innitesimal vertical height. We
can approximate this system with a vertical stack of many slab-shaped gas phases, each
thin enough to be practically uniform in its intensive properties, as depicted in Fig. 8.1.
The system can be isolated from the surroundings by conning the gas in a rigid adiabatic
container. In order to be able to associate each of the thin slab-shaped phases with a denite
constant elevation, we specify that the volume of each phase is constant so that in the rigid
container the vertical thickness of a phase cannot change.
We can use the phase of lowest elevation as the reference phase 0 , as indicated in the
gure. We repeat the derivation of Sec. 8.1.2 with one change: for each phase the volume
change dV is set equal to zero. Then the second sum on the right side of Eq. 8.1.6, with
terms proportional to dV , drops out and we are left with
dS D
X
X T 0 T
dS C
0
T
0
0
dn
(8.1.7)
In the equilibrium state of the isolated system, dS is equal to zero for an innitesimal change
of any of the independent variables. In this state, therefore, the coefcient of each term in
the sums on the right side of Eq. 8.1.7 must be zero. We conclude that in an equilibrium
state of a tall column of a pure gas, the temperature and chemical potential are uniform
throughout. The equation, however, gives us no information about pressure.
We will use this result to derive an expression for the dependence of the fugacity f on
elevation in an equilibrium state. We pick an arbitrary position such as the earths surface
for a reference elevation at which h is zero, and dene the standard chemical potential
(g) as the chemical potential of the gas under standard state conditions at this reference
elevation. At hD0, the chemical potential and fugacity are related by Eq. 7.8.7 which we
write in the following form, indicating the elevation in parentheses:
.0/ D
(g) C RT ln
f .0/
p
(8.1.8)

197
Imagine a small sample of gas of mass m that is initially at elevation hD0. The vertical
extent of this sample should be small enough for the variation of the gravitational force
eld within the sample to be negligible. The gravitational work needed to raise the gas to
an arbitrary elevation h is w 0 D mgh (page 82). We assume this process is carried out
reversibly at constant volume and without heat, so that there is no change in T , p, V , S, or
f . The internal energy U of the gas must increase by mgh D nMgh, where M is the molar
mass. Then, because the Gibbs energy G depends on U according to G D U T S C pV ,
G must also increase by nM gh.
The chemical potential is the molar Gibbs energy G=n. During the elevation process,
f remains the same and increases by Mgh:
.h/ D .0/ C Mgh
(8.1.9)
(f .h/Df .0/ )
From Eqs. 8.1.8 and 8.1.9, we can deduce the following general relation between chemical
potential, fugacity, and elevation:
.h/ D
(g) C RT ln
f .h/
C Mgh
p
(8.1.10)
(pure gas in
gravitational eld)
Compare this relation with the equation that denes the fugacity when the effect of a gravitational eld is negligible:
D (g) C RT ln.f =p / (Eq. 7.8.7 on page 184). The
additional term Mgh is needed when the vertical extent of the gas is considerable.
Some thermodynamicists call the expression on the right side of Eq. 8.1.10 the total chemical potential or gravitochemical potential and reserve the term chemical
potential for the function (g)CRT ln.f =p /. With these denitions, in an equilibrium state the total chemical potential is the same at all elevations and the chemical
potential decreases with increasing elevation.
This book instead denes the chemical potential of a pure substance at any elevation as the molar Gibbs energy at that elevation, as recommended in a 2001 IUPAC
technical report.1 When the chemical potential is dened in this way, it has the same
value at all elevations in an equilibrium state.
We know that in the equilibrium state of the gas column, the chemical potential .h/
has the same value at each elevation h. Equation 8.1.10 shows that in order for this to be
possible, the fugacity must decrease with increasing elevation. By equating expressions
from Eq. 8.1.10 for .h/ at an arbitrary elevation h, and for .0/ at the reference elevation,
we obtain
f .h/
f .0/
(g) C RT ln C Mgh D (g) C RT ln

(8.1.11)
p
p
Solving for f .h/ gives
f .h/ D f .0/e
1 Ref.
Mgh=RT
(8.1.12)
(pure gas at equilibrium
in gravitational eld)
[2].

198
If we treat the gas as ideal, so that the fugacity equals the pressure, this equation becomes
M gh=RT
p.h/ D p.0/e
(8.1.13)
(pure ideal gas at equilibrium
Equation 8.1.13 is the barometric formula for a pure ideal gas. It shows that in the equilibrium state of a tall column of an ideal gas, the pressure decreases exponentially with
increasing elevation.
This derivation of the barometric formula has introduced a method that will be used in
Sec. 9.8.1 for dealing with mixtures in a gravitational eld. There is, however, a shorter
derivation based on Newtons second law and not involving the chemical potential. Consider one of the thin slab-shaped phases of Fig. 8.1. Let the density of the phase be , the
area of each horizontal face be As , and the thickness of the slab be h. The mass of the phase
is then m D As h. The pressure difference between the top and bottom of the phase is p.
Three vertical forces act on the phase: an upward force pAs at its lower face, a downward
force .p C p/As at its upper face, and a downward gravitational force mg D
As gh.
If the phase is at rest, the net vertical force is zero: pAs .p C p/As
As gh D 0, or
p D
gh. In the limit as the number of phases becomes innite and h and p become
innitesimal, this becomes
dp D
g dh
(8.1.14)
(uid at equilibrium
Equation 8.1.14 is a general relation between changes in elevation and hydrostatic pressure
in any uid. To apply it to an ideal gas, we replace the density by D nM=V D M=Vm D
Mp=RT and rearrange to dp=p D .gM=RT / dh. Treating g and T as constants, we
integrate from hD0 to an arbitrary elevation h and obtain the same result as Eq. 8.1.13.
8.1.5 The pressure in a liquid droplet

The equilibrium shape of a small liquid droplet surrounded by vapor of the same substance,
when the effects of gravity and other external forces are negligible, is spherical. This is the
result of the surface tension of the liquidgas interface which acts to minimize the ratio of
surface to volume. The interface acts somewhat like the stretched membrane of an inated
balloon, resulting in a greater pressure inside the droplet than the pressure of the vapor in
equilibrium with it.
We can derive the pressure difference by considering a closed system containing a
spherical liquid droplet and surrounding vapor. We treat both phases as open subsystems.
An innitesimal change dU of the internal energy is the sum of contributions from the liquid and gas phases and from the surface work dAs , where is the surface tension of the
liquidgas interface and As is the surface area of the droplet (Sec. 5.7):
dU D dU l C dU g C dAs
D T l dS l
C T g dS g
p l dV l C
dnl
p g dV g C
dng C dAs
(8.1.15)

199
Note that Eq. 8.1.15 is not an expression for the total differential of U , because V l and
As are not independent variables. A derivation by a procedure similar to the one used in
Sec. 8.1.2 shows that at equilibrium the liquid and gas have equal temperatures and equal
chemical potentials, and the pressure in the droplet is greater than the gas pressure by an
amount that depends on r:
2
pl D pg C
(8.1.16)
r
Equation 8.1.16 is the Laplace equation. The pressure difference is signicant if r is small,
and decreases as r increases. The limit r!1 represents the at surface of bulk liquid with
p l equal to p g .
The derivation of Eq. 8.1.16 is left as an exercise (Prob. 8.1). The Laplace equation
is valid also for a liquid droplet in which the liquid and the surrounding gas may both be
mixtures (Prob. 9.3 on page 281).
The Laplace equation can also be applied to the pressure in a gas bubble surrounded
by liquid. In this case the liquid and gas phases switch roles, and the equation becomes
p g D p l C 2 =r.
8.1.6 The number of independent variables

From this point on in this book, unless stated otherwise, the discussions of multiphase systems will implicitly assume the existence of thermal, mechanical, and transfer equilibrium.
Equations will not explicitly show these equilibria as a condition of validity.
In the rest of this chapter, we shall assume the state of each phase can be described by
the usual variables: temperature, pressure, and amount. That is, variables such as elevation
in a gravitational eld, interface surface area, and extent of stretching of a solid, are not
relevant.
How many of the usual variables of an open multiphase one-substance equilibrium system are independent? To nd out, we go through the following argument. In the absence of
any kind of equilibrium, we could treat phase as having the three independent variables
T , p , and n , and likewise for every other phase. A system of P phases without thermal,
mechanical, or transfer equilibrium would then have 3P independent variables.
We must decide how to count the number of phases. It is usually of no thermodynamic
signicance whether a phase, with particular values of its intensive properties, is contiguous. For instance, splitting a crystal into several pieces is not usually considered to
change the number of phases or the state of the system, provided the increased surface
area makes no signicant contribution to properties such as internal energy. Thus, the
number of phases P refers to the number of different kinds of phases.
Each independent relation resulting from equilibrium imposes a restriction on the system and reduces the number of independent variables by one. A two-phase system with
thermal equilibrium has the single relation T D T . For a three-phase system, there are
two such relations that are independent, for instance T D T and T D T . (The additional relation T D T is not independent since we may deduce it from the other two.) In
general, thermal equilibrium gives P 1 independent relations among temperatures.

8.2 P HASE D IAGRAMS OF P URE S UBSTANCES
200
By the same reasoning, mechanical equilibrium involves P 1 independent relations

among pressures, and transfer equilibrium involves P
1 independent relations among
chemical potentials.
The total number of independent relations for equilibrium is 3.P 1/, which we subtract
from 3P (the number of independent variables in the absence of equilibrium) to obtain
the number of independent variables in the equilibrium system: 3P
3.P
1/ D 3.
Thus, an open single-substance system with any number of phases has at equilibrium three
independent variables. For example, in equilibrium states of a two-phase system we may
vary T , n , and n independently, in which case p is a dependent variable; for a given value
of T , the value of p is the one that allows both phases to have the same chemical potential.
8.1.7 The Gibbs phase rule for a pure substance

The complete description of the state of a system must include the value of an extensive
variable of each phase (e.g., the volume, mass, or amount) in order to specify how much of
the phase is present. For an equilibrium system of P phases with a total of 3 independent
variables, we may choose the remaining 3 P variables to be intensive. The number of these
intensive independent variables is called the number of degrees of freedom or variance,
F , of the system:
F D3
(8.1.17)
(pure substance)
The application of the phase rule to multicomponent systems will be taken up in Sec.
13.1. Equation 8.1.17 is a special case, for C D 1, of the more general Gibbs phase
rule F D C P C 2.
We may interpret the variance F in either of two ways:

F is the number of intensive variables needed to describe an equilibrium state, in
addition to the amount of each phase;
F is the maximum number of intensive properties that we may vary independently
while the phases remain in equilibrium.
A system with two degrees of freedom is called bivariant, one with one degree of freedom is univariant, and one with no degrees of freedom is invariant. For a system of a
pure substance, these three cases correspond to one, two, and three phases respectively.
For instance, a system of liquid and gaseous H2 O (and no other substances) is univariant
(F D 3 P D 3 2 D 1); we are able to independently vary only one intensive property,
such as T , while the liquid and gas remain in equilibrium.
8.2 PHASE DIAGRAMS OF PURE SUBSTANCES

A phase diagram is a two-dimensional map showing which phase or phases are able to exist
in an equilibrium state under given conditions. This chapter describes pressurevolume and
pressuretemperature phase diagrams for a single substance, and Chap. 13 will describe
numerous types of phase diagrams for multicomponent systems.

201
8.2.1 Features of phase diagrams

Two-dimensional phase diagrams for a single-substance system can be generated as projections of a three-dimensional surface in a coordinate system with Cartesian axes p, V =n, and
T . A point on the three-dimensional surface corresponds to a physically-realizable combination of values, for an equilibrium state of the system containing a total amount n of the
substance, of the variables p, V =n, and T .
The concepts needed to interpret single-substance phase diagrams will be illustrated
with carbon dioxide.
Three-dimensional surfaces for carbon dioxide are shown at two different scales in
Fig. 8.2 on the next page and in Fig. 8.3 on page 203. In these gures, some areas of
the surface are labeled with a single physical state: solid, liquid, gas, or supercritical uid.
A point in one of these areas corresponds to an equilibrium state of the system containing
a single phase of the labeled physical state. The shape of the surface in this one-phase area
gives the equation of state of the phase (i.e., the dependence of one of the variables on the
other two). A point in an area labeled with two physical states corresponds to two coexisting phases. The triple line is the locus of points for all possible equilibrium systems of
three coexisting phases, which in this case are solid, liquid, and gas. A point on the triple
line can also correspond to just one or two phases (see the discussion on page 203).
The two-dimensional projections shown in Figs. 8.2(b) and 8.2(c) are pressurevolume
and pressuretemperature phase diagrams. Because all phases of a multiphase equilibrium
system have the same temperature and pressure,2 the projection of each two-phase area
onto the pressuretemperature diagram is a curve, called a coexistence curve or phase
boundary, and the projection of the triple line is a point, called a triple point.
How may we use a phase diagram? The two axes represent values of two independent
variables, such as p and V =n or p and T . For given values of these variables, we place a
point on the diagram at the intersection of the corresponding coordinates; this is the system
point. Then depending on whether the system point falls in an area or on a coexistence
curve, the diagram tells us the number and kinds of phases that can be present in the equilibrium system.
If the system point falls within an area labeled with the physical state of a single phase,
only that one kind of phase can be present in the equilibrium system. A system containing a
pure substance in a single phase is bivariant (F D 3 1 D 2), so we may vary two intensive
properties independently. That is, the system point may move independently along two
coordinates (p and V =n, or p and T ) and still remain in the one-phase area of the phase
diagram. When V and n refer to a single phase, the variable V =n is the molar volume Vm
in the phase.
If the system point falls in an area of the pressurevolume phase diagram labeled with
symbols for two phases, these two phases coexist in equilibrium. The phases have the same
pressure and different molar volumes. To nd the molar volumes of the individual phases,
we draw a horizontal line of constant pressure, called a tie line, through the system point
and extending from one edge of the area to the other. The horizontal position of each end
of the tie line, where it terminates at the boundary with a one-phase area, gives the molar
volume in that phase in the two-phase system. For an example of a tie line, see Fig. 8.9 on
page 209.
2 This
statement assumes there are no constraints such as internal adiabatic partitions.

202
scf
bar
l+g
trip
le l
ine
s+
cm
ol
m
(a)
scf
bar
scf
s
g
l+g
cm mol
(b)
(c)
Figure 8.2 Relations among p, V =n, and T for carbon dioxide. a Areas are labeled
with the stable phase or phases (scf stands for supercritical uid). The open circle
indicates the critical point.
(a) Three-dimensional p.V =n/T surface. The dashed curve is the critical isotherm
at T D 304:21 K, and the dotted curve is a portion of the critical isobar at p D
73:8 bar.
(b) Pressurevolume phase diagram (projection of the surface onto the p.V =n/
plane).
(c) Pressuretemperature phase diagram (projection of the surface onto the pT
plane).
a Based
on data in Refs. [123] and [3].

203
bar
scf
l
s
s+l
l+g
cm
s+
ol
m
trip
le l
ine
Figure 8.3 Three-dimensional p.V =n/T surface for CO2 , magnied along the
V =n axis compared to Fig. 8.2. The open circle is the critical point, the dashed curve
is the critical isotherm, and the dotted curve is a portion of the critical isobar.
The triple line on the pressurevolume diagram represents the range of values of V =n
in which three phases (solid, liquid, and gas) can coexist at equilibrium.3 A three-phase
one-component system is invariant (F D 3 3 D 0); there is only one temperature (the
triple-point temperature Ttp ) and one pressure (the triple-point pressure ptp ) at which the
three phases can coexist. The values of Ttp and ptp are unique to each substance, and
are shown by the position of the triple point on the pressuretemperature phase diagram.
The molar volumes in the three coexisting phases are given by the values of V =n at the
three points on the pressurevolume diagram where the triple line touches a one-phase
area. These points are at the two ends and an intermediate position of the triple line. If the
system point is at either end of the triple line, only the one phase of corresponding molar
volume at temperature Ttp and pressure ptp can be present. When the system point is on the
triple line anywhere between the two ends, either two or three phases can be present. If the
system point is at the position on the triple line corresponding to the phase of intermediate
molar volume, there might be only that one phase present.
At high pressures, a substance may have additional triple points for two solid phases and
the liquid, or for three solid phases. This is illustrated by the pressuretemperature phase
diagram of H2 O in Fig. 8.4 on the next page, which extends to pressures up to 30 kbar. (On
this scale, the liquidgas coexistence curve lies too close to the horizontal axis to be visible.)
The diagram shows seven different solid phases of H2 O differing in crystal structure and
3 Helium
is the only substance lacking a solidliquidgas triple line. When a system containing the coexisting
liquid and gas of 4 He is cooled to 2:17 K, a triple point is reached in which the third phase is a liquid called
He-II, which has the unique property of superuidity. It is only at high pressures (10 bar or greater) that solid
helium can exist.

204
30
25
VIII
VII
p=kbar
20
15
VI
10
5
0
200
liquid
V
II
III
250
300
T =K
350
400
Figure 8.4 High-pressure pressuretemperature phase diagram of H2 O. a The roman

numerals designate seven forms of ice.
a Based
on data in Refs. [49], Table 3.5, and [128].
designated ice I, ice II, and so on. Ice I is the ordinary form of ice, stable below 2 bar. On
the diagram are four triple points for two solids and the liquid and three triple points for
three solids. Each triple point is invariant. Note how H2 O can exist as solid ice VI or ice
VII above its standard melting point of 273 K if the pressure is high enough (hot ice).
8.2.2 Two-phase equilibrium

A system containing two phases of a pure substance in equilibrium is univariant. Both
phases have the same values of T and of p, but these values are not independent because
of the requirement that the phases have equal chemical potentials. We may vary only one
intensive variable of a pure substance (such as T or p) independently while two phases
coexist in equilibrium.
At a given temperature, the pressure at which solid and gas or liquid and gas are in
equilibrium is called the vapor pressure or saturation vapor pressure of the solid or
liquid.4 The vapor pressure of a solid is sometimes called the sublimation pressure. We
may measure the vapor pressure of a liquid at a xed temperature with a simple device
called an isoteniscope (Fig. 8.5 on the next page).
At a given pressure, the melting point or freezing point is the temperature at which
solid and liquid are in equilibrium, the boiling point or saturation temperature is the
temperature at which liquid and gas are in equilibrium, and the sublimation temperature
or sublimation point is the temperature at which solid and gas are in equilibrium.
4 In a system of more than one substance, vapor pressure can refer to the partial pressure of a substance in a gas
mixture equilibrated with a solid or liquid of that substance. The effect of total pressure on vapor pressure will
be discussed in Sec. 12.8.1. This book refers to the saturation vapor pressure of a liquid when it is necessary to
indicate that it is the pure liquid and pure gas phases that are in equilibrium at the same pressure.

205
bath
Figure 8.5 An isoteniscope. The liquid to be investigated is placed in the vessel and
U-tube, as indicated by shading, and maintained at a xed temperature in the bath. The
pressure in the side tube is reduced until the liquid boils gently and its vapor sweeps
out the air. The pressure is adjusted until the liquid level is the same in both limbs of
the U-tube; the vapor pressure of the liquid is then equal to the pressure in the side
tube, which can be measured with a manometer.
standard
melting point
standard
boiling point
bar
triple point
Figure 8.6 Pressuretemperature phase diagram of H2 O. (Based on data in Ref.

[123].)
The relation between temperature and pressure in a system with two phases in equilibrium is shown by the coexistence curve separating the two one-phase areas on the pressure
temperature diagram (see Fig. 8.6). Consider the liquidgas curve. If we think of T as the
independent variable, the curve is a vapor-pressure curve showing how the vapor pressure
of the liquid varies with temperature. If, however, p is the independent variable, then the
curve is a boiling-point curve showing the dependence of the boiling point on pressure.
The normal melting point or boiling point refers to a pressure of one atmosphere, and
the standard melting point or boiling point refers to the standard pressure. Thus, the normal
boiling point of water (99:97 C) is the boiling point at 1 atm; this temperature is also known

206
as the steam point. The standard boiling point of water (99:61 C) is the boiling point at the
slightly lower pressure of 1 bar.
Coexistence curves will be discussed further in Sec. 8.4.
8.2.3 The critical point

Every substance has a certain temperature, the critical temperature, above which only one
uid phase can exist at any volume and pressure (Sec. 2.2.3). The critical point is the point
on a phase diagram corresponding to liquidgas coexistence at the critical temperature, and
the critical pressure is the pressure at this point.
To observe the critical point of a substance experimentally, we can evacuate a glass
vessel, introduce an amount of the substance such that V =n is approximately equal to the
molar volume at the critical point, seal the vessel, and raise the temperature above the
critical temperature. The vessel now contains a single uid phase. When the substance is
slowly cooled to a temperature slightly above the critical temperature, it exhibits a cloudy
appearance, a phenomenon called critical opalescence (Fig. 8.7 on the next page). The
opalescence is the scattering of light caused by large local density uctuations. At the
critical temperature, a meniscus forms between liquid and gas phases of practically the
same density. With further cooling, the density of the liquid increases and the density of the
gas decreases.
At temperatures above the critical temperature and pressures above the critical pressure,
the one existing uid phase is called a supercritical uid. Thus, a supercritical uid of a
pure substance is a uid that does not undergo a phase transition to a different uid phase
when we change the pressure at constant temperature or change the temperature at constant
pressure.5
A uid in the supercritical region can have a density comparable to that of the liquid,
and can be more compressible than the liquid. Under supercritical conditions, a substance
is often an excellent solvent for solids and liquids. By varying the pressure or temperature,
the solvating power can be changed; by reducing the pressure isothermally, the substance
can be easily removed as a gas from dissolved solutes. These properties make supercritical
uids useful for chromatography and solvent extraction.
The critical temperature of a substance can be measured quite accurately by observing
the appearance or disappearance of a liquidgas meniscus, and the critical pressure can be
measured at this temperature with a high-pressure manometer. To evaluate the density at
the critical point, it is best to extrapolate the mean density of the coexisting liquid and gas
phases, . l C g /=2, to the critical temperature as illustrated in Fig. 8.8 on page 208. The
observation that the mean density closely approximates a linear function of temperature, as
shown in the gure, is known as the law of rectilinear diameters, or the law of Cailletet
and Matthias. This law is an approximation, as can be seen by the small deviation of the
mean density of SF6 from a linear relation very close to the critical point in Fig. 8.8(b). This
failure of the law of rectilinear diameters is predicted by recent theoretical treatments.6
5 If,
however, we increase p at constant T , the supercritical uid will change to a solid. In the phase diagram
of H2 O, the coexistence curve for ice VII and liquid shown in Fig. 8.4 extends to a higher temperature than the
critical temperature of 647 K. Thus, supercritical water can be converted to ice VII by isothermal compression.
6 Refs. [164] and [10].
207
PHOTOS BY RAY RAKOW

(a)
(b)
(c)
(d)
Figure 8.7 Glass bulb lled with CO2 at a value of V =n close to the critical value,
viewed at four different temperatures. The three balls have densities less than, approximately equal to, and greater than the critical density. a
(a) Supercritical uid at a temperature above the critical temperature.
(b) Intense opalescence just above the critical temperature.
(c) Meniscus formation slightly below the critical temperature; liquid and gas of nearly
the same density.
(d) Temperature well below the critical temperature; liquid and gas of greatly different
densities.
a Ref.
[150].
8.2.4 The lever rule

Consider a single-substance system whose system point is in a two-phase area of a pressure
volume phase diagram. How can we determine the amounts in the two phases?
As an example, let the system contain a xed amount n of a pure substance divided into
liquid and gas phases, at a temperature and pressure at which these phases can coexist in
equilibrium. When heat is transferred into the system at this T and p, some of the liquid
vaporizes by a liquidgas phase transition and V increases; withdrawal of heat at this T
and p causes gas to condense and V to decrease. The molar volumes and other intensive
properties of the individual liquid and gas phases remain constant during these changes at
constant T and p. On the pressurevolume phase diagram of Fig. 8.9 on page 209, the
volume changes correspond to movement of the system point to the right or left along the
tie line AB.
When enough heat is transferred into the system to vaporize all of the liquid at the given

208
kg m
kg m
(b)
(a)
Figure 8.8 Densities of coexisting gas and liquid phases close to the critical point
as functions of temperature for (a) CO2 ; a (b) SF6 . b Experimental gas densities are
shown by open squares and experimental liquid densities by open triangles. The mean
density at each experimental temperature is shown by an open circle. The open diamond is at the critical temperature and critical density.
a Based
b Data
on data in Ref. [115].

of Ref. [126], Table VII.
T and p, the system point moves to point B at the right end of the tie line. V =n at this point
g
must be the same as the molar volume of the gas, Vm . We can see this because the system
point could have moved from within the one-phase gas area to this position on the boundary
without undergoing a phase transition.
When, on the other hand, enough heat is transferred out of the system to condense all
of the gas, the system point moves to point A at the left end of the tie line. V =n at this point
l
is the molar volume of the liquid, Vm .
When the system point is at position S on the tie line, both liquid and gas are present.
Their amounts must be such that the total volume is the sum of the volumes of the individual
phases, and the total amount is the sum of the amounts in the two phases:
g
l
V D V l C V g D nl Vm C ng Vm
(8.2.1)
n D nl C ng
(8.2.2)
The value of V =n at the system point is then given by the equation

g
l
V
nl Vm C ng Vm
D
n
nl C ng
(8.2.3)

l+g
l
Ll
209
Lg
V =n
Figure 8.9 Tie line (dashed) at constant T and p in the liquidgas area of a pressure
volume phase diagram. Points A and B are at the ends of the tie line, and point S is a
system point on the tie line. Ll and Lg are the lengths AS and SB, respectively.
which can be rearranged to
l
n Vm
l
V
n
Dn
V
n
g
Vm
(8.2.4)
l
The quantities Vm V =n and V =n Vm are the lengths Ll and Lg , respectively, dened
in the gure and measured in units of V =n. This gives us the lever rule for liquidgas
equilibrium:7
nl Ll D ng Lg
or
ng
Ll
D g
nl
L
(8.2.5)
(coexisting liquid and gas
phases of a pure substance)
In Fig. 8.9 the system point S is positioned on the tie line two thirds of the way from
the left end, making length Ll twice as long as Lg . The lever rule then gives the ratio of
amounts: ng =nl D Ll =Lg D 2. One-third of the total amount is liquid and two-thirds is
gas.
We cannot apply the lever rule to a point on the triple line, because we need more than
the value of V =n to determine the relative amounts present in three phases.
We can derive a more general form of the lever rule that will be needed in Chap. 13 for
phase diagrams of multicomponent systems. This general form can be applied to any
two-phase area of a two-dimensional phase diagram in which a tie-line construction is
valid, with the position of the system point along the tie line given by the variable
def
F D
a
b
(8.2.6)
where a and b are extensive state functions. (In the pressurevolume phase diagram of
Fig. 8.9, these functions are a D V and b D n and the system point position is given
7 The
relation is called the lever rule by analogy to a stationary mechanical lever, each end of which has the
same value of the product of applied force and distance from the fulcrum.

210
bar
C
C
cm mol
Figure 8.10 Isotherms for the uid phases of H2 O. a The open circle indicates the
critical point, the dashed curve is the critical isotherm at 373:95 C, and the dotted
curve encloses the two-phase area of the pressurevolume phase diagram. The triple
line lies too close to the bottom of the diagram to be visible on this scale.
a Based
by F D V =n.) We repeat the steps of the derivation above, labeling the two phases by
superscripts and instead of l and g. The relation corresponding to Eq. 8.2.4 is
b .F
F / D b .F
F /
(8.2.7)
If L and L are lengths measured along the tie line from the system point to the
ends of the tie line at single phases and , respectively, Eq. 8.2.7 is equivalent to the
general lever rule
L
b
D
b L D b L
or
(8.2.8)
b
L
8.2.5 Volume properties

Figure 8.10 is a pressurevolume phase diagram for H2 O. On the diagram are drawn
isotherms (curves of constant T ). These isotherms dene the shape of the three-dimensional
p.V =n/T surface. The area containing the horizontal isotherm segments is the two-phase
area for coexisting liquid and gas phases. The boundary of this area is dened by the dotted
curve drawn through the ends of the horizontal segments. The one-phase liquid area lies to

211
cm mol
Figure 8.11 Isobars for the uid phases of H2 O. a The open circle indicates the
critical point, the dashed curve is the critical isobar at 220:64 bar, and the dotted curve
encloses the two-phase area of the temperaturevolume phase diagram.
Solid curves: a, p D 200 bar; b, p D 210 bar; c, p D 230 bar; d, p D 240 bar.
a Based
the left of this curve, the one-phase gas area lies to the right, and the critical point lies at the
top.
The diagram contains the information needed to evaluate the molar volume at any temperature and pressure in the one-phase region and the derivatives of the molar volume with
respect to temperature and pressure. At a system point in the one-phase region, the slope
of the isotherm passing through the point is the partial derivative .@p=@Vm /T . Since the
isothermal compressibility is given by T D .1=Vm /.@Vm =@p/T , we have
T D
Vm
1
slope of isotherm
(8.2.9)
We see from Fig. 8.10 that the slopes of the isotherms are large and negative in the liquid
region, smaller and negative in the gas and supercritical uid regions, and approach zero at
the critical point. Accordingly, the isothermal compressibility of the gas and the supercritical uid is much greater than that of the liquid, approaching innity at the critical point.
The critical opalescence seen in Fig. 8.7 is caused by local density uctuations, which are
large when T is large.
Figure 8.11 shows isobars for H2 O instead of isotherms. At a system point in the
one-phase region, the slope of the isobar passing through the point is the partial derivative
.@T =@Vm /p . The cubic expansion coefcient is equal to .1=Vm /.@Vm =@T /p , so we have
D
Vm
1
slope of isobar
(8.2.10)

8.3 P HASE T RANSITIONS
212
The gure shows that the slopes of the isobars are large and positive in the liquid region,
smaller and negative in the gas and supercritical uid regions, and approach zero at the
critical point. Thus the gas and the supercritical uid have much larger cubic expansion
coefcients than the liquid. The value of approaches innity at the critical point, meaning
that in the critical region the density distribution is greatly affected by temperature gradients.
This may account for the low position of the middle ball in Fig. 8.7(b).
8.3 PHASE TRANSITIONS

Recall (Sec. 2.2.2) that an equilibrium phase transition of a pure substance is a process in
which some or all of the substance is transferred from one coexisting phase to another at
constant temperature and pressure.
8.3.1 Molar transition quantities

The quantity vap H is the molar enthalpy change for the reversible process in which liquid
changes to gas at a temperature and pressure at which the two phases coexist at equilibrium.
This quantity is called the molar enthalpy of vaporization.8 Since the pressure is constant
during the process, vap H is equal to the heat per amount of vaporization (Eq. 5.3.8).
Hence, vap H is also called the molar heat of vaporization.
The rst edition of this book used the notation vap Hm , with subscript m, in order to
make it clear that it refers to a molar enthalpy of vaporization. The most recent edition
of the IUPAC Green Book9 recommends that p be interpreted as an operator symbol:
def
p D @=@ p , where p is the abbreviation for a process at constant T and p (in

this case vap) and p is its advancement. Thus vap H is the same as .@H=@ vap /T;p
where vap is the amount of liquid changed to gas.
Here is a list of symbols for the molar enthalpy changes of various equilibrium phase
transitions:
vap H molar enthalpy of vaporization (liquid!gas)
sub H molar enthalpy of sublimation (solid!gas)
fus H molar enthalpy of fusion (solid!liquid)
trs H
molar enthalpy of a transition between any two phases in general
Molar enthalpies of vaporization, sublimation, and fusion are positive. The reverse processes of condensation (gas!liquid), condensation or deposition (gas!solid), and freezing
(liquid!solid) have negative enthalpy changes.
The subscripts in the list above are also used for other molar transition quantities. Thus,
there is the molar entropy of vaporization vap S, the molar internal energy of sublimation
sub U , and so on.
A molar transition quantity of a pure substance is the change of an extensive property
divided by the amount transferred between the phases. For example, when an amount n
in a liquid phase is allowed to vaporize to gas at constant T and p, the enthalpy change is
8 Because
vap H is an enthalpy change per amount of vaporization, it would be more accurate to call it the
molar enthalpy change of vaporization.
9 Ref. [36], p. 58.

8.3 P HASE T RANSITIONS
g
H D nHm
213
l
nHm and the molar enthalpy of vaporization is
H
g
D Hm
n
vap H D
l
Hm
(8.3.1)
(pure substance)
In other words, vap H is the enthalpy change per amount vaporized and is also the difference between the molar enthalpies of the two phases.
A molar property of a phase, being intensive, usually depends on two independent intensive variables such as T and p. Despite the fact that vap H is the difference of the two
g
l
molar properties Hm and Hm , its value depends on only one intensive variable, because the
two phases are in transfer equilibrium and the system is univariant. Thus, we may treat
vap H as a function of T only. The same is true of any other molar transition quantity.
The molar Gibbs energy of an equilibrium phase transition, trs G, is a special case. For
the phase transition !, we may write an equation analogous to Eq. 8.3.1 and equate the
molar Gibbs energy in each phase to a chemical potential (see Eq. 7.8.1):
trs G D Gm
Gm D
(8.3.2)
(pure substance)
But the transition is between two phases at equilibrium, requiring both phases to have the
D 0. Therefore, the molar Gibbs energy of any equilibsame chemical potential:
rium phase transition is zero:
trs G D 0
(8.3.3)
(pure substance)
Since the Gibbs energy is dened by G D H T S, in phase we have Gm D G =n D
. Similarly, in phase we have G D H

T Sm
T Sm . When we substitute these
m
m
(Eq. 8.3.2) and set T equal to the transition temperature

expressions in trs G D Gm Gm
Ttrs , we obtain
Hm
trs G D .Hm
D trs H
Hm /
Ttrs .Sm
Sm /
Ttrs trs S
(8.3.4)
Then, by setting trs G equal to zero, we nd the molar entropy and molar enthalpy of the
equilibrium phase transition are related by
trs S D
trs H
Ttrs
(8.3.5)
(pure substance)
where trs S and trs H are evaluated at the transition temperature Ttrs .
We may obtain Eq. 8.3.5 directly from the second law. With the phases in equilibrium,
the transition process is reversible. The second law gives S D q=Ttrs D H=Ttrs .
Dividing by the amount transferred between the phases gives Eq. 8.3.5.

8.4 C OEXISTENCE C URVES
214
8.3.2 Calorimetric measurement of transition enthalpies

The most precise measurement of the molar enthalpy of an equilibrium phase transition uses
electrical work. A known quantity of electrical work is performed on a system containing
coexisting phases, in a constant-pressure adiabatic calorimeter, and the resulting amount of
substance transferred between the phases is measured. The rst law shows that the electrical
work I 2 Rel t equals the heat that would be needed to cause the same change of state. This
heat, at constant p, is the enthalpy change of the process.
The method is similar to that used to measure the heat capacity of a phase at constant
pressure (Sec. 7.3.2), except that now the temperature remains constant and there is no need
to make a correction for the heat capacity of the calorimeter.
8.3.3 Standard molar transition quantities

The standard molar enthalpy of vaporization, vap H , is the enthalpy change when pure
liquid in its standard state at a specied temperature changes to gas in its standard state at
the same temperature, divided by the amount changed.
Note that the initial state of this process is a real one (the pure liquid at pressure p ), but
the nal state (the gas behaving ideally at pressure p ) is hypothetical. The liquid and gas
are not necessarily in equilibrium with one another at pressure p and the temperature of
interest, and we cannot evaluate vap H from a calorimetric measurement with electrical
work without further corrections. The same difculty applies to the evaluation of sub H .
In contrast, vap H and sub H (without the symbol), as well as fus H , all refer to
reversible transitions between two real phases coexisting in equilibrium.
Let X represent one of the thermodynamic potentials or the entropy of a phase. The
standard molar transition quantities vap X D Xm (g) Xm (l) and sub X D Xm (g)
Xm (s) are functions only of T . To evaluate vap X or sub X at a given temperature, we
must calculate the change of Xm for a path that connects the standard state of the liquid or
solid with that of the gas. The simplest choice of path is one of constant temperature T with
the following steps:
1. Isothermal change of the pressure of the liquid or solid, starting with the standard
state at pressure p and ending with the pressure equal to the vapor pressure pvap of
the condensed phase at temperature T . The value of Xm in this step can be obtained
from an expression in the second column of Table 7.4, or from an approximation in
the last column of the table.
2. Reversible vaporization or sublimation to form the real gas at T and pvap . The change
of Xm in this step is either vap X or sub X , which can be evaluated experimentally.
3. Isothermal change of the real gas at pressure pvap to the hypothetical ideal gas at
pressure p . Table 7.5 has the relevant formulas relating molar quantities of a real
gas to the corresponding standard molar quantities.
The sum of Xm for these three steps is the desired quantity vap X or sub X .
8.4 COEXISTENCE CURVES

A coexistence curve on a pressuretemperature phase diagram shows the conditions under
which two phases can coexist in equilibrium, as explained in Sec. 8.2.2.

215
kJ mol
gas
liquid
b
ar
Figure 8.12 Top: chemical potential surfaces of the liquid and gas phases of H2 O;
the two phases are at equilibrium along the intersection (heavy curve). (The vertical
scale for has an arbitrary zero.) Bottom: projection of the intersection onto the pT
plane, generating the coexistence curve. (Based on data in Ref. [70].)
8.4.1 Chemical potential surfaces

We may treat the chemical potential of a pure substance in a single phase as a function
of the independent variables T and p, and represent the function by a three-dimensional
surface. Since the condition for equilibrium between two phases of a pure substance is that
both phases have the same T , p, and , equilibrium in a two-phase system can exist only
along the intersection of the surfaces of the two phases as illustrated in Fig. 8.12.
The shape of the surface for each phase is determined by the partial derivatives of the
chemical potential with respect to temperature and pressure as given by Eqs. 7.8.3 and 7.8.4:

@
@
D Sm
D Vm
(8.4.1)
@T p
@p T
Let us explore how varies with T at constant p for the different physical states of a
substance. The stable phase at each temperature is the one of lowest , since transfer of a
substance from a higher to a lower at constant T and p is spontaneous.
From the relation .@ =@T /p D Sm , we see that at constant p the slope of versus
T is negative since molar entropy is always positive. Furthermore, the magnitude of the
slope increases on going from solid to liquid and from liquid to gas, because the molar

0:04
C
A
0:03
liq
uid
B
0
gas,
bar
0:02
260
270
280 290
T =K
300
(a)
liquid
solid
gas
0:01
bar
:03
1
0
250
p=bar
sol
id
3
0:00
gas,
=kJ mol
216
triple point
310
0
250
260
270
280 290
T =K
300
310
(b)
Figure 8.13 Phase stability of H2 O. a

(a) Chemical potentials of different physical states as functions of temperature. (The
scale for has an arbitrary zero.) Chemical potentials of the gas are shown at 0:03 bar
and 0:003 bar. The effect of pressure on the curves for the solid and liquid is negligible. At p D 0:03 bar, solid and liquid coexist at T D 273:16 K (point A) and liquid
and gas coexist at T D 297:23 K (point B). At p D 0:003 bar, solid and gas coexist at
T D 264:77 K (point C).
(b) Pressuretemperature phase diagram with points corresponding to those in (a).
a Based
on data in Refs. [70] and [87].
entropies of sublimation and vaporization are positive. This difference in slope is illustrated
by the curves for H2 O in Fig. 8.13(a). The triple-point pressure of H2 O is 0:0062 bar. At
a pressure of 0:03 bar, greater than the triple-point pressure, the curves for solid and liquid
intersect at a melting point (point A) and the curves for liquid and gas intersect at a boiling
point (point B).
From .@ =@p/T D Vm , we see that a pressure reduction at constant temperature lowers the chemical potential of a phase. The result of a pressure reduction from 0:03 bar to
0:003 bar (below the triple-point pressure of H2 O) is a downward shift of each of the curves
of Fig. 8.13(a) by a distance proportional to the molar volume of the phase. The shifts of
the solid and liquid curves are too small to see ( is only 0:002 kJ mol 1 ). Because
the gas has a large molar volume, the gas curve shifts substantially to a position where it
intersects with the solid curve at a sublimation point (point C). At 0:003 bar, or any other
pressure below the triple-point pressure, only a solidgas equilibrium is possible for H2 O.
The liquid phase is not stable at any pressure below the triple-point pressure, as shown by
the pressuretemperature phase diagram of H2 O in Fig. 8.13(b).
8.4.2 The Clapeyron equation

If we start with two coexisting phases, and , of a pure substance and change the temperature of both phases equally without changing the pressure, the phases will no longer be

217
in equilibrium, because their chemical potentials change unequally. In order for the phases
to remain in equilibrium during the temperature change dT of both phases, there must be
a certain simultaneous change dp in the pressure of both phases. The changes dT and dp
must be such that the chemical potentials of both phases change equally so as to remain
equal to one another: d D d .
The innitesimal change of in a phase is given by d D Sm dT CVm dp (Eq. 7.8.2).
Thus, the two phases remain in equilibrium if dT and dp satisfy the relation
Sm dT C Vm dp D
Sm dT C Vm dp
(8.4.2)
which we rearrange to
Sm
dp
D
dT
Vm
Sm
Vm
(8.4.3)
or
dp
S
D trs
dT
trs V
(8.4.4)
(pure substance)
Equation 8.4.4 is one form of the Clapeyron equation, which contains no approximations.
We nd an alternative form by substituting trs S D trs H=Ttrs (Eq. 8.3.5):
dp
H
D trs
dT
Ttrs V
(8.4.5)
(pure substance)
Equations 8.4.4 and 8.4.5 give the slope of the coexistence curve, dp= dT , as a function
of quantities that can be measured. For the sublimation and vaporization processes, both
trs H and trs V are positive. Therefore, according to Eq. 8.4.5, the solidgas and liquid
gas coexistence curves have positive slopes. For the fusion process, however, fus H is
positive, but fus V may be positive or negative depending on the substance, so that the
slope of the solidliquid coexistence curve may be either positive or negative. The absolute
value of fus V is small, causing the solidliquid coexistence curve to be relatively steep;
see Fig. 8.13(b) for an example.
Most substances expand on melting, making the slope of the solidliquid coexistence
curve positive. This is true of carbon dioxide, although in Fig. 8.2(c) on page 202 the
curve is so steep that it is difcult to see the slope is positive. Exceptions at ordinary
pressures, substances that contract on melting, are H2 O, rubidium nitrate, and the
elements antimony, bismuth, and gallium.
The phase diagram for H2 O in Fig. 8.4 on page 204 clearly shows that the coexistence curve for ice I and liquid has a negative slope due to ordinary ice being less
dense than liquid water. The high-pressure forms of ice are more dense than the liquid, causing the slopes of the other solidliquid coexistence curves to be positive. The
ice VIIice VIII coexistence curve is vertical, because these two forms of ice have
identical crystal structures, except for the orientations of the H2 O molecule; therefore,
within experimental uncertainty, the two forms have equal molar volumes.
We may rearrange Eq. 8.4.5 to give the variation of p with T along the coexistence
curve:
H dT
dp D trs
(8.4.6)
trs V T

218
BIOGRAPHICAL SKETCH
Benoit Paul Emile Clapeyron (17991864)

dity of the possibility of creating motive power
[i.e., work] or heat out of nothing. . . . This new
method of demonstration seems to me worthy of
the attention of theoreticians; it seems to me to
be free of all objection . . .
Clapeyron was a French civil and railroad engineer who made important contributions to
thermodynamic theory. He was born in Paris,
the son of a merchant.
He graduated from the Ecole Polytechnique

in 1818, four years after Sadi Carnots graduation from the same school, and then trained
as an engineer at the Ecole de Mines. At this

time, the Russian czar asked the French government for engineers to help create a program
of civil and military engineering. Clapeyron
and his classmate Gabriel Lam were offered
e
this assignment. They lived in Russia for ten
years, teaching pure and applied mathematics
in Saint Petersburg and jointly publishing engineering and mathematical papers. In 1831
they returned to France; their liberal political
views and the oppression of foreigners by the
new czar, Nicholas I, made it impossible for
them to remain in Russia.
Back in France, Clapeyron became involved
in the construction of the rst French passenger railroads and in the design of steam locomotives and metal bridges. He was married
with a daughter.
In a paper published in 1834 in the journal
of the Ecole Polytechnique, Clapeyron brought

attention to the work of Sadi Carnot (page
107), who had died two years before:a
Among studies which have appeared on the theory of heat I will mention nally a work by S.
Carnot, published in 1824, with the title Reections on the Motive Power of Fire. The idea
which serves as a basis of his researches seems
to me to be both fertile and beyond question;
his demonstrations are founded on the absura Ref.
[30].
b Ref.
[171].
c Ref.
[90].
d Ref.
[80].
Clapeyrons paper used indicator diagrams

and calculus for a rigorous proof of Carnots
conclusion that the efciency of a reversible
heat engine depends only on the temperatures
of the hot and cold heat reservoirs. However,
it retained the erroneous caloric theory of heat.
It was not until the appearance of English and
German translations of this paper that Clapeyrons analysis enabled Kelvin to dene a thermodynamic temperature scale and Clausius to
introduce entropy and write the mathematical
statement of the second law.
Clapeyrons 1834 paper also derived an expression for the temperature dependence of the
vapor pressure of a liquid equivalent to what is
now called the Clapeyron equation (Eq. 8.4.5).
The paper used a reversible cycle with vaporization at one temperature followed by condensation at an innitesimally-lower temperature
and pressure, and equated the efciency of this
cycle to that of a gas between the same two
temperatures. Although the thermodynamic
temperature T does not appear as such, it is
represented by a temperature-dependent function whose relation to the Celsius scale had to
be determined experimentally.b
Beginning in 1844, Clapeyron taught the
course on steam engines at the Ecole Nationale

des Ponts et Chauss es near Paris, the olde
est French engineering school. In this course,
surprisingly, he seldom mentioned his theory
of heat engines based on Carnots work.c He
eventually embraced the equivalence of heat
and work established by Joules experiments.d
At the time of Clapeyrons death, the rail
road entrepreneur Emile P reire wrote:e
e
We were together since 1832. Ive never done
important business without consulting him, Ive
never found a judgment more reliable and more
honest. His modesty was so great and his character so good that I never knew him to have an
enemy.
e Ref.
[80].

219
Consider the transition from solid to liquid (fusion). Because of the fact that the cubic
expansion coefcient and isothermal compressibility of a condensed phase are relatively
small, fus V is approximately constant for small changes of T and p. If fus H is also
practically constant, integration of Eq. 8.4.6 yields the relation
p2
fus H T2
ln
fus V
T1
p1
(8.4.7)
fus V .p2 p1 /
fus H
(8.4.8)
or
T2
T1 exp
(pure substance)
from which we may estimate the dependence of the melting point on pressure.
8.4.3 The ClausiusClapeyron equation

When the gas phase of a substance coexists in equilibrium with the liquid or solid phase,
and provided T and p are not close to the critical point, the molar volume of the gas is
much greater than that of the condensed phase. Thus, we may write for the processes of
vaporization and sublimation
g
vap V D Vm
l
Vm
Vm
sub V D Vm
s
Vm
g
The further approximation that the gas behaves as an ideal gas, Vm

Eq. 8.4.5 to
dp
dT
ptrs H
RT 2
Vm
(8.4.9)
RT =p, then changes
(8.4.10)
(pure substance,
vaporization or sublimation)
Equation 8.4.10 is the ClausiusClapeyron equation. It gives an approximate expression for the slope of a liquidgas or solidgas coexistence curve. The expression is not valid
for coexisting solid and liquid phases, or for coexisting liquid and gas phases close to the
critical point.
At the temperature and pressure of the triple point, it is possible to carry out all three
equilibrium phase transitions of fusion, vaporization, and sublimation. When fusion is followed by vaporization, the net change is sublimation. Therefore, the molar transition enthalpies at the triple point are related by
fus H C vap H D sub H
(8.4.11)
Since all three of these transition enthalpies are positive, it follows that sub H is greater
than vap H at the triple point. Therefore, according to Eq. 8.4.10, the slope of the solid
gas coexistence curve at the triple point is slightly greater than the slope of the liquidgas
coexistence curve.
We divide both sides of Eq. 8.4.10 by p and rearrange to the form
d.p=p /
p=p
trs H
R
dT
T2
(8.4.12)

220
Then, using the mathematical identities d.p=p /=.p=p / D d ln.p=p / and dT =T 2 D

d.1=T /, we can write Eq. 8.4.12 in three alternative forms:
d ln.p=p /
dT
d ln.p=p /
d ln.p=p /
d.1=T /
trs H
RT 2
trs H
d.1=T /
R
trs H
R
(8.4.13)
(pure substance,
(8.4.14)
(pure substance,
(8.4.15)
(pure substance,
Equation 8.4.15 shows that the curve of a plot of ln.p=p / versus 1=T (where p is the
vapor pressure of a pure liquid or solid) has a slope at each temperature equal, usually to
a high degree of accuracy, to vap H=R or sub H=R at that temperature. This kind of
plot provides an alternative to calorimetry for evaluating molar enthalpies of vaporization
and sublimation.
If we use the recommended standard pressure of 1 bar, the ratio p=p appearing in
these equations becomes p=bar. That is, p=p is simply the numerical value of p
when p is expressed in bars. For the purpose of using Eq. 8.4.15 to evaluate trs H , we
can replace p by any convenient value. Thus, the curves of plots of ln.p=bar/ versus
1=T , ln.p=Pa/ versus 1=T , and ln.p=Torr/ versus 1=T using the same temperature
and pressure data all have the same slope (but different intercepts) and yield the same
value of trs H .
If we assume vap H or sub H is essentially constant in a temperature range, we may

integrate Eq. 8.4.14 from an initial to a nal state along the coexistence curve to obtain
p2
trs H 1
1
ln
(8.4.16)
p1
R
T2 T1
(pure substance,
Equation 8.4.16 allows us to estimate any one of the quantities p1 , p2 , T1 , T2 , or trs H ,

given values of the other four.

P ROBLEMS
221
PROBLEMS
in Appendix I.
8.1 Consider the system described in Sec. 8.1.5 containing a spherical liquid droplet of radius r
surrounded by pure vapor. Starting with Eq. 8.1.15, nd an expression for the total differential
of U . Then impose conditions of isolation and show that the equilibrium conditions are T g D
T l , g D l , and p l D p g C 2 =r, where is the surface tension.
8.2 This problem concerns diethyl ether at T D 298:15 K. At this temperature, the standard molar
entropy of the gas calculated from spectroscopic data is Sm (g) D 342:2 J K 1 mol 1 . The
saturation vapor pressure of the liquid at this temperature is 0:6691 bar, and the molar enthalpy
of vaporization is vap H D 27:10 kJ mol 1 . The second virial coefcient of the gas at this
temperature has the value B D 1:227 10 3 m3 mol 1 , and its variation with temperature is
given by dB= dT D 1:50 10 5 m3 K 1 mol 1 .
(a) Use these data to calculate the standard molar entropy of liquid diethyl ether at 298:15 K.
A small pressure change has a negligible effect on the molar entropy of a liquid, so that it
is a good approximation to equate Sm (l) to Sm (l) at the saturation vapor pressure.

(b) Calculate the standard molar entropy of vaporization and the standard molar enthalpy of
vaporization of diethyl ether at 298:15 K. It is a good approximation to equate Hm (l) to

Hm (l) at the saturation vapor pressure.
8.3 Explain why the chemical potential surfaces shown in Fig. 8.12 are concave downward; that
is, why .@ =@T /p becomes more negative with increasing T and .@ =@p/T becomes less
positive with increasing p.
8.4 Potassium has a standard boiling point of 773 C and a molar enthalpy of vaporization vap H D
84:9 kJ mol 1 . Estimate the saturation vapor pressure of liquid potassium at 400: C.
8.5 Naphthalene has a melting point of 78:2 C at 1 bar and 81:7 C at 100 bar. The molar volume
change on melting is fus V D 0:019 cm3 mol 1 . Calculate the molar enthalpy of fusion to
two signicant gures.
8.6 The dependence of the vapor pressure of a liquid on temperature, over a limited temperature
range, is often represented by the Antoine equation, log10 .p=Torr/ D A B=.t C C /, where
t is the Celsius temperature and A, B, and C are constants determined by experiment. A
variation of this equation, using a natural logarithm and the thermodynamic temperature, is
ln.p=bar/ D a
b
T Cc
The vapor pressure of liquid benzene at temperatures close to 298 K is adequately represented
by the preceding equation with the following values of the constants:
a D 9:25092
b D 2771:233 K
cD
53:262 K
(a) Find the standard boiling point of benzene.

(b) Use the ClausiusClapeyron equation to evaluate the molar enthalpy of vaporization of
benzene at 298:15 K.
8.7 At a pressure of one atmosphere, water and steam are in equilibrium at 99:97 C (the normal
boiling point of water). At this pressure and temperature, the water density is 0:958 g cm 3 , the
steam density is 5:98 10 4 g cm 3 , and the molar enthalpy of vaporization is 40:66 kJ mol 1 .

P ROBLEMS
222
(a) Use the Clapeyron equation to calculate the slope dp= dT of the liquidgas coexistence
curve at this point.
(b) Repeat the calculation using the ClausiusClapeyron equation.
(c) Use your results to estimate the standard boiling point of water. (Note: The experimental
value is 99:61 C.)
8.8 At the standard pressure of 1 bar, liquid and gaseous H2 O coexist in equilibrium at 372:76 K,
the standard boiling point of water.
(a) Do you expect the standard molar enthalpy of vaporization to have the same value as the
molar enthalpy of vaporization at this temperature? Explain.
(b) The molar enthalpy of vaporization at 372:76 K has the value vap H D 40:67 kJ mol 1 .
Estimate the value of vap H at this temperature with the help of Table 7.5 and the following data for the second virial coefcient of gaseous H2 O at 372:76 K:
BD
4:60
10
m3 mol
dB= dT D 3:4
10
m3 K
mol
(c) Would you expect the values of fus H and fus H to be equal at the standard freezing
point of water? Explain.
8.9 The standard boiling point of H2 O is 99:61 C. The molar enthalpy of vaporization at this
temperature is vap H D 40:67 kJ mol 1 . The molar heat capacity of the liquid at temperatures
close to this value is given by
Cp;m D a C b.t c/
where t is the Celsius temperature and the constants have the values
a D 75:94 J K
mol
b D 0:022 J K
mol
c D 99:61 C
Suppose 100:00 mol of liquid H2 O is placed in a container maintained at a constant pressure

of 1 bar, and is carefully heated to a temperature 5:00 C above the standard boiling point, resulting in an unstable phase of superheated water. If the container is enclosed with an adiabatic
boundary and the system subsequently changes spontaneously to an equilibrium state, what
amount of water will vaporize? (Hint: The temperature will drop to the standard boiling point,
and the enthalpy change will be zero.)
C HAPTER 9
M IXTURES
A homogeneous mixture is a phase containing more than one substance. This chapter discusses composition variables and partial molar quantities of mixtures in which no chemical
reaction is occurring. The ideal mixture is dened. Chemical potentials, activity coefcients, and activities of individual substances in both ideal and nonideal mixtures are discussed.
Except for the use of fugacities to determine activity coefcients in condensed phases,
a discussion of phase equilibria involving mixtures will be postponed to Chap. 13.
9.1 COMPOSITION VARIABLES

A composition variable is an intensive property that indicates the relative amount of a
particular species or substance in a phase.
9.1.1 Species and substances

We sometimes need to make a distinction between a species and a substance. A species is
any entity of denite elemental composition and charge and can be described by a chemical
formula, such as H2 O, H3 OC , NaCl, or NaC . A substance is a species that can be prepared
in a pure state (e.g., N2 and NaCl). Since we cannot prepare a macroscopic amount of a
single kind of ion by itself, a charged species such as H3 OC or NaC is not a substance.
Chap. 10 will discuss the special features of mixtures containing charged species.
9.1.2 Mixtures in general

The mole fraction of species i is dened by
n
def
xi D P i
j nj
or
n
def
yi D P i
j nj
(9.1.1)
(P D1)
where ni is the amount of species i and the sum is taken over all species in the mixture.
The symbol xi is used for a mixture in general, and yi is used when the mixture is a gas.
223
CHAPTER 9 MIXTURES
9.1 C OMPOSITION VARIABLES
224
The mass fraction, or weight fraction, of species i is dened by

def
wi D
m.i/
nM
DP i i
m
j nj Mj
(9.1.2)
(P D1)
where m.i/ is the mass of species i and m is the total mass.

The concentration, or molarity, of species i in a mixture is dened by
def
ci D
ni
V
The symbol M is often used to stand for units of mol L

tion of 0:5 M is 0:5 moles per liter, or 0:5 molar.
(9.1.3)
(P D1)
1
, or mol dm
3.
Thus, a concentra-
Concentration is sometimes called amount concentration or molar concentration to

avoid confusion with number concentration (the number of particles per unit volume).
An alternative notation for cA is [A].
A binary mixture is a mixture of two substances.
9.1.3 Solutions
A solution, strictly speaking, is a mixture in which one substance, the solvent, is treated in a
special way. Each of the other species comprising the mixture is then a solute. The solvent
is denoted by A and the solute species by B, C, and so on.1 Although in principle a solution
can be a gas mixture, in this section we will consider only liquid and solid solutions.
We can prepare a solution of varying composition by gradually mixing one or more
solutes with the solvent so as to continuously increase the solute mole fractions. During
this mixing process, the physical state (liquid or solid) of the solution remains the same as
that of the pure solvent. When the sum of the solute mole fractions is small compared to xA
(i.e., xA is close to unity), the solution is called dilute. As the solute mole fractions increase,
we say the solution becomes more concentrated.
Mole fraction, mass fraction, and concentration can be used as composition variables
for both solvent and solute, just as they are for mixtures in general. A fourth composition
variable, molality, is often used for a solute. The molality of solute species B is dened by
def
mB D
nB
m.A/
(9.1.4)
(solution)
where m.A/ D nA MA is the mass of solvent. The symbol m is sometimes used to stand
for units of mol kg 1 , although this should be discouraged because m is also the symbol
for meter. For example, a solute molality of 0:6 m is 0:6 moles of solute per kilogram of
solvent, or 0:6 molal.
1 Some
chemists denote the solvent by subscript 1 and use 2, 3, and so on for solutes.
CHAPTER 9 MIXTURES
9.1 C OMPOSITION VARIABLES
225
9.1.4 Binary solutions

We may write simplied equations for a binary solution of two substances, solvent A and
solute B. Equations 9.1.19.1.4 become
nB
nA C nB
(binary solution)
n B MB
n A MA C n B MB
(binary solution)
nB
nB
D
V
n A MA C n B MB
(binary solution)
xB D
wB D
cB D
mB D
(9.1.5)
nB
nA MA
(9.1.6)
(9.1.7)
(9.1.8)
(binary solution)
The right sides of Eqs. 9.1.59.1.8 express the solute composition variables in terms of the
amounts and molar masses of the solvent and solute and the density of the solution.
To be able to relate the values of these composition variables to one another, we solve
each equation for nB and divide by nA to obtain an expression for the mole ratio nB =nA :
from Eq. 9.1.5
from Eq. 9.1.6
nB
xB
D
nA
1 xB
nB
MA wB
D
nA
MB .1 wB /
from Eq. 9.1.7
nB
D
nA
from Eq. 9.1.8
nB
D MA m B
nA
MA c B
MB c B
(9.1.9)
(binary solution)
(9.1.10)
(binary solution)
(9.1.11)
(binary solution)
(9.1.12)
(binary solution)
These expressions for nB =nA allow us to nd one composition variable as a function of

another. For example, to nd molality as a function of concentration, we equate the expressions for nB =nA on the right sides of Eqs. 9.1.11 and 9.1.12 and solve for mB to obtain
mB D
cB
MB cB
(9.1.13)
CHAPTER 9 MIXTURES
9.2 PARTIAL M OLAR Q UANTITIES
226
A binary solution becomes more dilute as any of the solute composition variables becomes smaller. In the limit of innite dilution, the expressions for nB =nA become:
nB
D xB
nA
M
D A wB
MB
MA
D
cB D Vm;A cB
A
D MA m B
(9.1.14)
(binary solution at
innite dilution)
where a superscript asterisk ( ) denotes a pure phase. We see that, in the limit of innite
dilution, the composition variables xB , wB , cB , and mB are proportional to one another.
These expressions are also valid for solute B in a multisolute solution in which each solute
is very dilute; that is, in the limit xA !1.
The rule of thumb that the molarity and molality values of a dilute aqueous solution
are approximately equal is explained by the relation MA cB = A D MA mB (from Eq.
9.1.14), or cB = A D mB , and the fact that the density A of water is approximately
1 kg L 1 . Hence, if the solvent is water and the solution is dilute, the numerical value
of cB expressed in mol L 1 is approximately equal to the numerical value of mB expressed in mol kg 1 .
9.1.5 The composition of a mixture

We can describe the composition of a phase with the amounts of each species, or with any
of the composition variables dened earlier: mole fraction, mass fraction, concentration, or
molality. If we use mole fractions or mass fractions to describe the composition, we need
the values for all but one of the species, since the sum of all fractions is unity.
Other composition variables are sometimes used, such as volume fraction, mole ratio,
and mole percent. To describe the composition of a gas mixture, partial pressures can be
used (Sec. 9.3.1).
When the composition of a mixture is said to be xed or constant during changes of
temperature, pressure, or volume, this means there is no change in the relative amounts or
masses of the various species. A mixture of xed composition has xed values of mole
fractions, mass fractions, and molalities, but not necessarily of concentrations and partial
pressures. Concentrations will change if the volume changes, and partial pressures in a gas
mixture will change if the pressure changes.
9.2 PARTIAL MOLAR QUANTITIES

The symbol Xi , where X is an extensive property of a homogeneous mixture and the subscript i identies a constituent species of the mixture, denotes the partial molar quantity
CHAPTER 9 MIXTURES
227
of species i dened by
def
Xi D
@X
@ni
(9.2.1)
T;p;nj i
(mixture)
This is the rate at which property X changes with the amount of species i added to the
mixture as the temperature, the pressure, and the amounts of all other species are kept
constant. A partial molar quantity is an intensive state function. Its value depends on the
temperature, pressure, and composition of the mixture.
Keep in mind that as a practical matter, a macroscopic amount of a charged species (i.e.,
an ion) cannot be added by itself to a phase because of the huge electric charge that would
result. Thus if species i is charged, Xi as dened by Eq. 9.2.1 is a theoretical concept whose
value cannot be determined experimentally.
An older notation for a partial molar quantity uses an overbar: X i . The notation Xi0
was suggested in the rst edition of the IUPAC Green Book,2 but is not mentioned in
later editions.
9.2.1 Partial molar volume

In order to gain insight into the signicance of a partial molar quantity as dened by Eq.
9.2.1, let us rst apply the concept to the volume of an open single-phase system. Volume
has the advantage for our example of being an extensive property that is easily visualized.
Let the system be a binary mixture of water (substance A) and methanol (substance B), two
liquids that mix in all proportions. The partial molar volume of the methanol, then, is the
rate at which the system volume changes with the amount of methanol added to the mixture
at constant temperature and pressure: VB D .@V =@nB /T;p;nA .
At 25 C and 1 bar, the molar volume of pure water is Vm;A D 18:07 cm3 mol 1 and that
of pure methanol is Vm;B D 40:75 cm3 mol 1 . If we mix 100:0 cm3 of water at 25 C with
100:0 cm3 of methanol at 25 C, we nd the volume of the resulting mixture at 25 C is not
the sum of the separate volumes, 200:0 cm3 , but rather the slightly smaller value 193:1 cm3 .
The difference is due to new intermolecular interactions in the mixture compared to the pure
liquids.
Let us calculate the mole fraction composition of this mixture:
nA D
VA
100:0 cm3
D
Vm;A
18:07 cm3 mol
D 5:53 mol
VB
100:0 cm3
D
D 2:45 mol
Vm;B
40:75 cm3 mol 1
2:45 mol
nB
D
D 0:307
xB D
nA C nB
5:53 mol C 2:45 mol
nB D
(9.2.2)
(9.2.3)
(9.2.4)
Now suppose we prepare a large volume of a mixture of this composition (xB D 0:307)
and add an additional 40:75 cm3 (one mole) of pure methanol, as shown in Fig. 9.1(a). If
2 Ref.
[118], p. 44.
CHAPTER 9 MIXTURES
228
(a)
(b)
Figure 9.1 Addition of pure methanol (substance B) to a watermethanol mixture at

constant T and p.
(a) 40:75 cm3 (one mole) of methanol is placed in a narrow tube above a much greater
volume of a mixture (shaded) of composition xB D 0:307. The dashed line indicates
the level of the upper meniscus.
(b) After the two liquid phases have mixed by diffusion, the volume of the mixture has
increased by only 38:8 cm3 .
the initial volume of the mixture at 25 C was 10 , 000.0 cm3 , we nd the volume of the
new mixture at the same temperature is 10 , 038.8 cm3 , an increase of 38.8 cm3 see Fig.
9.1(b). The amount of methanol added is not innitesimal, but it is small enough compared
to the amount of initial mixture to cause very little change in the mixture composition: xB
increases by only 0:5%. Treating the mixture as an open system, we see that the addition of
one mole of methanol to the system at constant T , p, and nA causes the system volume to
increase by 38:8 cm3 . To a good approximation, then, the partial molar volume of methanol
in the mixture, VB D .@V =@nB /T;p;nA , is given by V =nB D 38:8 cm3 mol 1 .
The volume of the mixture to which we add the methanol does not matter as long as
it is large. We would have observed practically the same volume increase, 38:8 cm3 , if we
had mixed one mole of pure methanol with 100 , 000.0 cm3 of the mixture instead of only
10 , 000.0 cm3 .
Thus, we may interpret the partial molar volume of B as the volume change per amount
of B added at constant T and p when B is mixed with such a large volume of mixture
that the composition is not appreciably affected. We may also interpret the partial molar
volume as the volume change per amount when an innitesimal amount is mixed with a
nite volume of mixture.
The partial molar volume of B is an intensive property that is a function of the composition of the mixture, as well as of T and p. The limiting value of VB as xB approaches 1
(pure B) is Vm;B , the molar volume of pure B. We can see this by writing V D nB Vm;B for
pure B, giving us VB .xB D1/ D .@nB Vm;B =@nB /T;p;nA D Vm;B .
If the mixture is a binary mixture of A and B, and xB is small, we may treat the mixture
as a dilute solution of solvent A and solute B. As xB approaches 0 in this solution, VB
approaches a certain limiting value that is the volume increase per amount of B mixed with
a large amount of pure A. In the resulting mixture, each solute molecule is surrounded only
1
by solvent molecules. We denote this limiting value of VB by VB , the partial molar volume
of solute B at innite dilution.
It is possible for a partial molar volume to be negative. Magnesium sulfate, in aqueous
solutions of molality less than 0:07 mol kg 1 , has a negative partial molar volume.
CHAPTER 9 MIXTURES
229
Physically, this means that when a small amount of crystalline MgSO4 dissolves at
constant temperature in water, the liquid phase contracts. This unusual behavior is due
to strong attractive waterion interactions.
9.2.2 The total differential of the volume in an open system

Consider an open single-phase system consisting of a mixture of nonreacting substances.
How many independent variables does this system have?
We can prepare the mixture with various amounts of each substance, and we are able
to adjust the temperature and pressure to whatever values we wish (within certain limits
that prevent the formation of a second phase). Each choice of temperature, pressure, and
amounts results in a denite value of every other property, such as volume, density, and
mole fraction composition. Thus, an open single-phase system of C substances has 2 C C
independent variables.3
For a binary mixture (C D 2), the number of independent variables is four. We may
choose these variables to be T , p, nA , and nB , and write the total differential of V in the
general form
@V
@V
dV D
dT C
dp
@T p;nA ;nB
@p T;nA ;nB
@V
@V
(9.2.5)
C
dn C
dn
@nA T;p;nB A
@nB T;p;nA B
(binary mixture)
We know the rst two partial derivatives on the right side are given by4
@V
@V
D V
D T V
@T p;nA ;nB
@p T;nA ;nB
(9.2.6)
We identify the last two partial derivatives on the right side of Eq. 9.2.5 as the partial molar
volumes VA and VB . Thus, we may write the total differential of V for this open system in
the compact form
dV D V dT
T V dp C VA dnA C VB dnB
(9.2.7)
(binary mixture)
If we compare this equation with the total differential of V for a one-component closed
system, dV D V dT T V dp (Eq. 7.1.6), we see that an additional term is required for
each constituent of the mixture to allow the system to be open and the composition to vary.
When T and p are held constant, Eq. 9.2.7 becomes
dV D VA dnA C VB dnB
(9.2.8)
(binary mixture,
constant T and p)
3 C in this kind of system is actually the number of components. The number of components is usually the
same as the number of substances, but is less if certain constraints exist, such as reaction equilibrium or a xed
mixture composition. The general meaning of C will be discussed in Sec. 13.1.
4 See Eqs. 7.1.1 and 7.1.2, which are for closed systems.
CHAPTER 9 MIXTURES
230
mixture of
A and B
Figure 9.2 Mixing of water (A) and methanol (B) in a 2:1 ratio of volumes to form a
mixture of increasing volume and constant composition. The system is the mixture.
We obtain an important relation between the mixture volume and the partial molar volumes by imagining the following process. Suppose we continuously pour pure water and
pure methanol at constant but not necessarily equal volume rates into a stirred, thermostatted container to form a mixture of increasing volume and constant composition, as shown
schematically in Fig. 9.2. If this mixture remains at constant T and p as it is formed, none of
its intensive properties change during the process, and the partial molar volumes VA and VB
remain constant. Under these conditions, we can integrate Eq. 9.2.8 to obtain the additivity
rule for volume:5
V D VA nA C VB nB
(9.2.9)
(binary mixture)
This equation allows us to calculate the mixture volume from the amounts of the constituents and the appropriate partial molar volumes for the particular temperature, pressure,
and composition.
For example, given that the partial molar volumes in a watermethanol mixture of composition xB D 0:307 are VA D 17:74 cm3 mol 1 and VB D 38:76 cm3 mol 1 , we calculate
the volume of the watermethanol mixture described at the beginning of Sec. 9.2.1 as follows:
V D .17:74 cm3 mol
D 193:1 cm3
1 /.5:53 mol/
C .38:76 cm3 mol
1 /.2:45 mol/
(9.2.10)
We can differentiate Eq. 9.2.9 to obtain a general expression for dV under conditions of
constant T and p:
dV D VA dnA C VB dnB C nA dVA C nB dVB
(9.2.11)
But this expression for dV is consistent with Eq. 9.2.8 only if the sum of the last two terms
on the right is zero:
nA dVA C nB dVB D 0
(9.2.12)
(binary mixture,
constant T and p)
5 The
equation is an example of the result of applying Eulers theorem on homogeneous functions to V treated
as a function of nA and nB .
CHAPTER 9 MIXTURES
231
Equation 9.2.12 is the GibbsDuhem equation for a binary mixture, applied to partial
molar volumes. (Section 9.2.4 will give a general version of this equation.) Dividing both
sides of the equation by nA C nB gives the equivalent form
xA dVA C xB dVB D 0
(9.2.13)
(binary mixture,
constant T and p)
Equation 9.2.12 shows that changes in the values of VA and VB are related when the
composition changes at constant T and p. If we rearrange the equation to the form
dVA D
nB
dV
nA B
(9.2.14)
(binary mixture,
constant T and p)
we see that a composition change that increases VB (so that dVB is positive) must make VA
decrease.
9.2.3 Evaluation of partial molar volumes in binary mixtures

The partial molar volumes VA and VB in a binary mixture can be evaluated by the method
of intercepts. To use this method, we plot experimental values of the quantity V =n (where
n is nA C nB ) versus the mole fraction xB . V =n is called the mean molar volume.
See Fig. 9.3(a) on the next page for an example. In this gure, the tangent to the
curve drawn at the point on the curve at the composition of interest (the composition used
as an illustration in Sec. 9.2.1) intercepts the vertical line where xB equals 0 at V =n D
VA D 17:7 cm3 mol 1 , and intercepts the vertical line where xB equals 1 at V =n D VB D
38:8 cm3 mol 1 .
To derive this property of a tangent line for the plot of V =n versus xB , we use Eq. 9.2.9
to write
V A nA C V B nB
D VA xA C VB xB
n
D VA .1 xB / C VB xB D .VB VA /xB C VA
.V =n/ D
(9.2.15)
When we differentiate this expression for V =n with respect to xB , keeping in mind that
VA and VB are functions of xB , we obtain
d.V =n/
d.VB
D
dxB
D VB
D VB
D VB
VA /xB C VA
dxB
dV
dVB dVA
VA C
xB C A
dx
dxB
dx
B
B
dVB
dVA
.1 xB / C
xB
VA C
dxB
dxB
dVA
dVB
VA C
xA C
xB
dxB
dxB
(9.2.16)
The differentials dVA and dVB are related to one another by the GibbsDuhem equation
(Eq. 9.2.13): xA dVA C xB dVB D 0. We divide both sides of this equation by dxB to
CHAPTER 9 MIXTURES
232
m mix
cm mol
cm mol
cm mol
(a)
(b)
(c)
Figure 9.3 Mixtures of water (A) and methanol (B) at 25 C and 1 bar. a
(a) Mean molar volume as a function of xB . The dashed line is the tangent to the curve
at xB D 0:307.
(b) Molar volume of mixing as a function of xB . The dashed line is the tangent to the
curve at xB D 0:307.
(c) Partial molar volumes as functions of xB . The points at xB D 0:307 (open circles)
are obtained from the intercepts of the dashed line in either (a) or (b).
a Based
obtain
dVA
dxB
xA C
dVB
dxB
xB D 0
(9.2.17)
VA
(9.2.18)
and substitute in Eq. 9.2.16 to obtain

d.V =n/
D VB
dxB
Let the partial molar volumes of the constituents of a binary mixture of arbitrary
0
0
0
composition xB be VA and VB . Equation 9.2.15 shows that the value of V =n at the
0
0
0
0
0
point on the curve of V =n versus xB where the composition is xB is .VB VA /xB C VA .
0
0
Equation 9.2.18 shows that the tangent to the curve at this point has a slope of VB VA .
The equation of the line that passes through this point and has this slope, and thus is
0
0
0
the tangent to the curve at this point, is y D .VB VA /xB C VA , where y is the vertical
0
ordinate on the plot of .V =n/ versus xB . The line has intercepts yDVA at xB D0 and
0
yDVB at xB D1.
A variant of the method of intercepts is to plot the molar integral volume of mixing
given by
V nA Vm;A nB Vm;B
V (mix)
Vm (mix) D
D
(9.2.19)
n
n
CHAPTER 9 MIXTURES
233
versus xB , as illustrated in Fig. 9.3(b). V (mix) is the integral volume of mixingthe

volume change at constant T and p when solvent and solute are mixed to form a mixture of
volume V and total amount n (see Sec. 11.1.1). The tangent to the curve at the composition
of interest has intercepts VA Vm;A at xB D0 and VB Vm;B at xB D1.
To see this, we write
Vm (mix) D .V =n/
D .V =n/
We make the substitution .V =n/ D .VB
Vm (mix) D
VB
Vm;B
VA
xA Vm;A xB Vm;B
.1 xB /Vm;A xB Vm;B
(9.2.20)
VA /xB C VA from Eq. 9.2.15 and rearrange:

Vm;A
xB C V A
Vm;A
(9.2.21)
Differentiation with respect to xB yields

dVm (mix)
D VB
dxB
Vm;B
VA
Vm;A
D VB
Vm;B
VA
Vm;A
D VB
Vm;B
VA
Vm;A
dVB dVA
dV
C
xB C A
dxB dxB
dx
B
dVA
dVB
C
.1 xB / C
xB
dxB
dx
B
dVA
dVB
C
xA C
xB
dxB
dxB
(9.2.22)
With a substitution from Eq. 9.2.17, this becomes

dVm (mix)
D VB
dxB
Vm;B
VA
Vm;A
(9.2.23)
Equations 9.2.21 and 9.2.23 are analogous to Eqs. 9.2.15 and 9.2.18, with V =n replaced by Vm (mix), VA by .VA Vm;A /, and VB by .VB Vm;B /. Using the same
reasoning as for a plot of V =n versus xB , we nd the intercepts of the tangent to a
point on the curve of Vm (mix) versus xB are at VA Vm;A and VB Vm;B .
Figure 9.3 shows smoothed experimental data for watermethanol mixtures plotted in
both kinds of graphs, and the resulting partial molar volumes as functions of composition.
Note in Fig. 9.3(c) how the VA curve mirrors the VB curve as xB varies, as predicted by the
GibbsDuhem equation. The minimum in VB at xB 0:09 is mirrored by a maximum in VA
in agreement with Eq. 9.2.14; the maximum is much attenuated because nB =nA is much
less than unity.
Macroscopic measurements are unable to provide unambiguous information about molecular structure. Nevertheless, it is interesting to speculate on the implications of the
minimum observed for the partial molar volume of methanol. One interpretation is that
in a mostly aqueous environment, there is association of methanol molecules, perhaps
involving the formation of dimers.
9.2.4 General relations

The discussion above of partial molar volumes used the notation Vm;A and Vm;B for the
molar volumes of pure A and B. The partial molar volume of a pure substance is the same as
CHAPTER 9 MIXTURES
234
the molar volume, so we can simplify the notation by using VA and VB instead. Hereafter,
this book will denote molar quantities of pure substances by such symbols as VA , HB , and
Si .
The relations derived above for the volume of a binary mixture may be generalized for
any extensive property X of a mixture of any number of constituents. The partial molar
quantity of species i, dened by
@X
def
Xi D
(9.2.24)
@ni T;p;nj i
is an intensive property that depends on T , p, and the composition of the mixture. The
additivity rule for property X is
X
XD
ni X i
(9.2.25)
i
(mixture)
and the GibbsDuhem equation applied to X can be written in the equivalent forms
X
ni dXi D 0
(9.2.26)
(constant T and p)
and
X
i
xi dXi D 0
(9.2.27)
(constant T and p)
These relations can be applied to a mixture in which each species i is a nonelectrolyte substance, an electrolyte substance that is dissociated into ions, or an individual ionic species.
In Eq. 9.2.27, the mole fraction xi must be based on the different species considered to
be present in the mixture. For example, an aqueous solution of NaCl could be treated as
a mixture of components A=H2 O and B=NaCl, with xB equal to nB =.nA C nB /; or the
constituents could be taken as H2 O, NaC , and Cl , in which case the mole fraction of NaC
would be xC D nC =.nA C nC C n /.
A general method to evaluate the partial molar quantities XA and XB in a binary mixture
is based on the variant of the method of intercepts described in Sec. 9.2.3. The molar mixing
quantity X(mix)=n is plotted versus xB , where X(mix) is .X nA XA nB XB /. On this
plot, the tangent to the curve at the composition of interest has intercepts equal to XA XA
at xB D0 and XB XB at xB D1.
We can obtain experimental values of such partial molar quantities of an uncharged
species as Vi , Cp;i , and Si . It is not possible, however, to evaluate the partial molar quantities Ui , Hi , Ai , and Gi because these quantities involve the internal energy brought into the
system by the species, and we cannot evaluate the absolute value of internal energy (Sec.
2.6.2). For example, while we can evaluate the difference Hi Hi from calorimetric measurements of enthalpies of mixing, we cannot evaluate the partial molar enthalpy Hi itself.
We can, however, include such quantities as Hi in useful theoretical relations.
As mentioned on page 227, a partial molar quantity of a charged species is something
else we cannot evaluate. It is possible, however, to obtain values relative to a reference
CHAPTER 9 MIXTURES
235
ion. Consider an aqueous solution of a fully-dissociated electrolyte solute with the

formula M C X , where C and
are the numbers of cations and anions per solute
formula unit. The partial molar volume VB of the solute, which can be determined
experimentally, is related to the (unmeasurable) partial molar volumes VC and V of
the constituent ions by
VB D C VC C V
(9.2.28)
For aqueous solutions, the usual reference ion is HC , and the partial molar volume of
1
this ion at innite dilution is arbitrarily set equal to zero: VHC D 0.
For example, given the value (at 298:15 K and 1 bar) of the partial molar volume
at innite dilution of aqueous hydrogen chloride
1
VHCl D 17:82 cm3 mol
(9.2.29)
we can nd the so-called conventional partial molar volume of Cl ion:

1
1
VCl D VHCl
1
VHC D 17:82 cm3 mol
1
Going one step further, the measured value VNaCl D 16:61 cm3 mol
ion, the conventional value
1
1
VNaC D VNaCl
1
VCl D .16:61
17:82/ cm3 mol
(9.2.30)
1
gives, for NaC
1:21 cm3 mol
(9.2.31)
9.2.5 Partial specic quantities

A partial specic quantity of a substance is the partial molar quantity divided by the molar
mass, and has dimensions of volume divided by mass. For example, the partial specic
volume vB of solute B in a binary solution is given by
VB
@V
vB D
D
(9.2.32)
MB
@m.B/ T;p;m.A/
where m.A/ and m.B/ are the masses of solvent and solute.
Although this book makes little use of specic quantities and partial specic quantities,
in some applications they have an advantage over molar quantities and partial molar quantities because they can be evaluated without knowledge of the molar mass. For instance, the
value of a solutes partial specic volume is used to determine its molar mass by the method
of sedimentation equilibrium (Sec. 9.8.2).
The general relations in Sec. 9.2.4 involving partial molar quantities may be turned
into relations involving partial specic quantities by replacing amounts by masses, mole
fractions by mass fractions, and partial molar quantities by partial specic quantities. Using
P
volume as an example, we can write an additivity relation V D i m.i/vi , and Gibbs
P
P
Duhem relations i m.i / dvi D 0 and i wi dvi D 0. For a binary mixture of A and B,
we can plot the specic volume v versus the mass fraction wB ; then the tangent to the curve
at a given composition has intercepts equal to vA at wB D0 and vB at wB D1. A variant of
this plot is v wA vA wB vB versus wB ; the intercepts are then equal to vA vA and
vB vB .
CHAPTER 9 MIXTURES
236
9.2.6 The chemical potential of a species in a mixture

Just as the molar Gibbs energy of a pure substance is called the chemical potential and given
the special symbol , the partial molar Gibbs energy Gi of species i in a mixture is called
the chemical potential of species i, dened by
@G
def
(9.2.33)
i D
@ni T;p;nj i
(mixture)
If there are work coordinates for nonexpansion work, the partial derivative is taken at constant values of these coordinates.
The chemical potential of a species in a phase plays a crucial role in equilibrium problems, because it is a measure of the escaping tendency of the species from the phase. Although we cannot determine the absolute value of i for a given state of the system, we
are usually able to evaluate the difference between the value in this state and the value in a
dened reference state.
In an open single-phase system containing a mixture of s different nonreacting species,
we may in principle independently vary T , p, and the amount of each species. This is a
total of 2Cs independent variables. The total differential of the Gibbs energy of this system
is given by Eq. 5.5.9 on page 142, often called the Gibbs fundamental equation:
dG D
S dT C V dp C
s
X
dni
(9.2.34)
(mixture)
i D1
Consider the special case of a mixture containing charged species, for example an aqueous solution of the electrolyte KCl. We could consider the constituents to be either the
substances H2 O and KCl, or else H2 O and the species KC and Cl . Any mixture we can
prepare in the laboratory must remain electrically neutral, or virtually so. Thus, while we
are able to independently vary the amounts of H2 O and KCl, we cannot in practice independently vary the amounts of KC and Cl in the mixture. The chemical potential of the
KC ion is dened as the rate at which the Gibbs energy changes with the amount of KC
added at constant T and p while the amount of Cl is kept constant. This is a hypothetical
process in which the net charge of the mixture increases. The chemical potential of a ion is
therefore a valid but purely theoretical concept. Let A stand for H2 O, B for KCl, C for KC ,
and for Cl . Then it is theoretically valid to write the total differential of G for the KCl
solution either as
dG D S dT C V dp C A dnA C B dnB
(9.2.35)
or as
dG D
S dT C V dp C
A dnA
C dnC
dn
(9.2.36)
9.2.7 Equilibrium conditions in a multiphase, multicomponent system

This section extends the derivation described in Sec. 8.1.2, which was for equilibrium conditions in a multiphase system containing a single substance, to a more general kind of
system: one with two or more homogeneous phases containing mixtures of nonreacting
species. The derivation assumes there are no internal partitions that could prevent transfer
CHAPTER 9 MIXTURES
237
of species and energy between the phases, and that effects of gravity and other external
force elds are negligible.
The system consists of a reference phase, 0 , and other phases labeled by 0 . Species
are labeled by subscript i. Following the procedure of Sec. 8.1.1, we write for the total
differential of the internal energy
X
0
dU D dU C
dU
0
DT
dS
p dV C
T dS
0
i
dn
i
p dV C
i dni
(9.2.37)
The conditions of isolation are

dU D 0
dV

dV
For each species i:

X
0
dn C
dn D 0
i
i
(no expansion work)
(9.2.39)
(closed system)
D0
(9.2.38)
(9.2.40)
We use these relations to substitute for dU , dV , and dn in Eq. 9.2.37. After making the
i
P
0
and solving for dS , we obtain
further substitution dS D dS
0 dS
dS D
X T 0 T
dS
0
T
0
XX
i
0
i
T 0
X p 0 p
dV
0
T
0
dn
i
(9.2.41)
This equation is like Eq. 8.1.6 on page 195 with provision for more than one species.
In the equilibrium state of the isolated system, S has the maximum possible value, dS
is equal to zero for an innitesimal change of any of the independent variables, and the
coefcient of each term on the right side of Eq. 9.2.41 is zero. We nd that in this state each
phase has the same temperature and the same pressure, and for each species the chemical
potential is the same in each phase.
Suppose the system contains a species i 0 that is effectively excluded from a particular
phase, 00 . For instance, sucrose molecules dissolved in an aqueous phase are not accommodated in the crystal structure of an ice phase, and a nonpolar substance may be essentially
00
insoluble in an aqueous phase. We can treat this kind of situation by setting dn0 equal to
i
zero. Consequently there is no equilibrium condition involving the chemical potential of
this species in phase 00 .
To summarize these conclusions: In an equilibrium state of a multiphase, multicomponent system without internal partitions, the temperature and pressure are uniform throughout
CHAPTER 9 MIXTURES
238
the system, and each species has a uniform chemical potential except in phases where it is
excluded.
This statement regarding the uniform chemical potential of a species applies to both a
substance and an ion, as the following argument explains. The derivation in this section
begins with Eq. 9.2.37, an expression for the total differential of U . Because it is a total
differential, the expression requires the amount ni of each species i in each phase to be
an independent variable. Suppose one of the phases is the aqueous solution of KCl used
as an example at the end of the preceding section. In principle (but not in practice),
the amounts of the species H2 O, KC , and Cl can be varied independently, so that it
is valid to include these three species in the sums over i in Eq. 9.2.37. The derivation
then leads to the conclusion that KC has the same chemical potential in phases that
are in transfer equilibrium with respect to KC , and likewise for Cl . This kind of
situation arises when we consider a Donnan membrane equilibrium (Sec. 12.7.3) in
which transfer equilibrium of ions exists between solutions of electrolytes separated
by a semipermeable membrane.
9.2.8 Relations involving partial molar quantities

Here we derive several useful relations involving partial molar quantities in a single-phase
system that is a mixture. The independent variables are T , p, and the amount ni of each
constituent species i.
From Eqs. 9.2.26 and 9.2.27, the GibbsDuhem equation applied to the chemical potentials can be written in the equivalent forms
X
(9.2.42)
ni d i D 0
(constant T and p)
and
X
xi d
D0
(9.2.43)
(constant T and p)
These equations show that the chemical potentials of different species cannot be varied
independently at constant T and p.
A more general version of the GibbsDuhem equation, without the restriction of constant T and p, is
X
S dT V dp C
ni d i D 0
(9.2.44)
i
This version is derived by comparing the expression for dG given by Eq. 9.2.34 with the
P
P
P
differential dGD i i dni C i ni d i obtained from the additivity rule GD i i ni .
The Gibbs energy is dened by G D H T S. Taking the partial derivatives of both
sides of this equation with respect to ni at constant T , p, and nj i gives us
@G
@ni
T;p;nj i
@H
@ni
T;p;nj i
@S
@ni
(9.2.45)
T;p;nj i
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
239
We recognize each partial derivative as a partial molar quantity and rewrite the equation as
i
D Hi
T Si
This is analogous to the relation D G=n D Hm T Sm for a pure substance.

P
From the total differential of the Gibbs energy, dG D S dT C V dp C i
9.2.34), we obtain the following reciprocity relations:
@ i
@S
@ i
@V
D
D
@T p;fnig
@ni T;p;nj i
@p T;fnig
@ni T;p;nj i
(9.2.46)
dni (Eq.
(9.2.47)
The symbol fnig stands for the set of amounts of all species, and subscript fnig on a partial
derivative means the amount of each species is constantthat is, the derivative is taken at
constant composition of a closed system. Again we recognize partial derivatives as partial
molar quantities and rewrite these relations as follows:
@ i
D Si
(9.2.48)
@T p;fnig
@ i
@p
T;fnig
D Vi
(9.2.49)
These equations are the equivalent for a mixture of the relations .@ =@T /p D Sm and
.@ =@p/T D Vm for a pure phase (Eqs. 7.8.3 and 7.8.4).
Taking the partial derivatives of both sides of U D H pV with respect to ni at constant
T , p, and nj i gives
Ui D Hi pVi
(9.2.50)
Finally, we can obtain a formula for Cp;i , the partial molar heat capacity at constant
pressure of species i, by writing the total differential of H in the form
X @H
@H
@H
dH D
dT C
dp C
dn
@T p;fnig
@p T;fnig
@ni T;p;nj i i
i
X
@H
D Cp dT C
dp C
Hi dni
(9.2.51)
@p T;fnig
i
from which we have the reciprocity relation .@Cp =@ni /T;p;nj i D .@Hi =@T /p;fnig , or
Cp;i D
@Hi
@T
(9.2.52)
p;fnig
9.3 GAS MIXTURES

The gas mixtures described in this chapter are assumed to be mixtures of nonreacting
gaseous substances.
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
240
9.3.1 Partial pressure

The partial pressure pi of substance i in a gas mixture is dened as the product of its mole
fraction in the gas phase and the pressure of the phase:
def
pi D yi p
(9.3.1)
(gas mixture)
P
P
The sum of the partial pressures of all substances in a gas mixture is i pi D i yi p D
P
p i yi . Since the sum of the mole fractions of all substances in a mixture is 1, this sum
becomes
X
pi D p
(9.3.2)
(gas mixture)
Thus, the sum of the partial pressures equals the pressure of the gas phase. This statement
is known as Daltons Law. It is valid for any gas mixture, regardless of whether or not the
gas obeys the ideal gas equation.
9.3.2 The ideal gas mixture

As discussed in Sec. 3.5.1, an ideal gas (whether pure or a mixture) is a gas with negligible
intermolecular interactions. It obeys the ideal gas equation p D nRT =V (where n in a
P
mixture is the sum i ni ) and its internal energy in a closed system is a function only of
temperature. The partial pressure of substance i in an ideal gas mixture is pi D yi p D
yi nRT =V ; but yi n equals ni , giving
pi D
ni RT
V
(9.3.3)
(ideal gas mixture)
Equation 9.3.3 is the ideal gas equation with the partial pressure of a constituent substance replacing the total pressure, and the amount of the substance replacing the total
amount. The equation shows that the partial pressure of a substance in an ideal gas mixture
is the pressure the substance by itself, with all others removed from the system, would have
at the same T and V as the mixture. Note that this statement is only true for an ideal gas
mixture. The partial pressure of a substance in a real gas mixture is in general different
from the pressure of the pure substance at the same T and V , because the intermolecular
interactions are different.
9.3.3 Partial molar quantities in an ideal gas mixture

We need to relate the chemical potential of a constituent of a gas mixture to its partial
pressure. We cannot measure the absolute value of a chemical potential, but we can evaluate
its value relative to the chemical potential in a particular reference state called the standard
state.
The standard state of substance i in a gas mixture is the same as the standard state of
the pure gas described in Sec. 7.7: It is the hypothetical state in which pure gaseous i has
the same temperature as the mixture, is at the standard pressure p , and behaves as an ideal
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
241
A(g)
(A + B)(g)
p D p0
pA D p 0
p D pA C pB
Figure 9.4 System with two gas phases, pure A and a mixture of A and B, separated
by a semipermeable membrane through which only A can pass. Both phases are ideal
gases at the same temperature.
gas. The standard chemical potential (g) of gaseous i is the chemical potential of i in
i
this gas standard state, and is a function of temperature.
To derive an expression for i in an ideal gas mixture relative to (g), we make an
i
assumption based on the following argument. Suppose we place pure A, an ideal gas, in
a rigid box at pressure p 0 . We then slide a rigid membrane into the box so as to divide
the box into two compartments. The membrane is permeable to A; that is, molecules of
A pass freely through its pores. There is no reason to expect the membrane to affect the
pressures on either side,6 which remain equal to p 0 . Finally, without changing the volume
of either compartment, we add a second gaseous substance, B, to one side of the membrane
to form an ideal gas mixture, as shown in Fig. 9.4. The membrane is impermeable to B, so
the molecules of B stay in one compartment and cause a pressure increase there. Since the
mixture is an ideal gas, the molecules of A and B do not interact, and the addition of gas B
causes no change in the amounts of A on either side of the membrane. Thus, the pressure
of A in the pure phase and the partial pressure of A in the mixture are both equal to p 0 .
Our assumption, then, is that the partial pressure pA of gas A in an ideal gas mixture in
equilibrium with pure ideal gas A is equal to the pressure of the pure gas.
Because the system shown in Fig. 9.4 is in an equilibrium state, gas A must have the
same chemical potential in both phases. This is true even though the phases have different
pressures (see Sec. 9.2.7). Since the chemical potential of the pure ideal gas is given by
D (g) C RT ln.p=p /, and we assume that pA in the mixture is equal to p in the pure
gas, the chemical potential of A in the mixture is given by
A
A (g)
C RT ln
pA
p
(9.3.4)
In general, for each substance i in an ideal gas mixture, we have the relation
i
i (g)
C RT ln
pi
p
(9.3.5)
(ideal gas mixture)
where (g) is the chemical potential of i in the gas standard state at the same temperature
i
as the mixture.
Equation 9.3.5 shows that if the partial pressure of a constituent of an ideal gas mixture
is equal to p , so that ln.pi =p / is zero, the chemical potential is equal to the standard
6 We
assume the gas is not adsorbed to a signicant extent on the surface of the membrane or in its pores.
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
242
chemical potential. Conceptually, a standard state should be a well-dened state of the

system, which in the case of a gas is the pure ideal gas at pDp . Thus, although a
constituent of an ideal gas mixture with a partial pressure of 1 bar is not in its standard
state, it has the same chemical potential as in its standard state.
Equation 9.3.5 will be taken as the thermodynamic denition of an ideal gas mixture.
Any gas mixture in which each constituent i obeys this relation between i and pi at all
compositions is by denition an ideal gas mixture. The nonrigorous nature of the assumption used to obtain Eq. 9.3.5 presents no difculty if we consider the equation to be the basic
denition.
By substituting the expression for i into .@ i =@T /p;fnig D Si (Eq. 9.2.48), we
obtain an expression for the partial molar entropy of substance i in an ideal gas mixture:

@ i (g)
pi
Si D
R ln
@T
p
p;fnig
pi
D Si R ln
(9.3.6)
p
(ideal gas mixture)
The quantity Si D @ (g)=@T p;fnig is the standard molar entropy of constituent i. It

i
is the molar entropy of i in its standard state of pure ideal gas at pressure p .
Substitution of the expression for i from Eq. 9.3.5 and the expression for Si from Eq.
9.3.6 into Hi D i C T Si (from Eq. 9.2.46) yields Hi D (g) C T Si , which is equivalent
i
to
Hi D Hi
(9.3.7)
(ideal gas mixture)
This tells us that the partial molar enthalpy of a constituent of an ideal gas mixture at a given
temperature is independent of the partial pressure or mixture composition; it is a function
only of T .
From .@ i =@p/T;fnig D Vi (Eq. 9.2.49), the partial molar volume of i in an ideal gas
mixture is given by

@ i (g)
@ ln.pi =p /
Vi D
C RT
(9.3.8)
@p
@p
T;fnig
T;fnig
The rst partial derivative on the right is zero because (g) is a function only of T . For
i
the second partial derivative, we write pi =p D yi p=p . The mole fraction yi is constant
when the amount of each substance is constant, so we have @ ln.yi p=p /=@pT;fnig D 1=p.
The partial molar volume is therefore given by
Vi D
RT
p
(9.3.9)
(ideal gas mixture)
which is what we would expect simply from the ideal gas equation. The partial molar
volume is not necessarily equal to the standard molar volume, which is Vi D RT =p for
an ideal gas.
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
243
From Eqs. 9.2.50, 9.2.52, 9.3.7, and 9.3.9 we obtain the relations
Ui D Ui
(9.3.10)
(ideal gas mixture)
and
Cp;i D Cp;i
(9.3.11)
(ideal gas mixture)
Thus, in an ideal gas mixture the partial molar internal energy and the partial molar heat
capacity at constant pressure, like the partial molar enthalpy, are functions only of T .
The denition of an ideal gas mixture given by Eq. 9.3.5 is consistent with the criteria
for an ideal gas listed at the beginning of Sec. 3.5.1, as the following derivation shows.
P
From Eq. 9.3.9 and the additivity rule, we nd the volume is given by V D i ni Vi D
P
i ni RT =p D nRT =p, which is the ideal gas equation. From Eq. 9.3.10 we have
P
P
U D
i ni Ui D
i ni Ui , showing that U is a function only of T in a closed
system. These properties apply to any gas mixture obeying Eq. 9.3.5, and they are the
properties that dene an ideal gas according to Sec. 3.5.1.
9.3.4 Real gas mixtures

Fugacity
The fugacity f of a pure gas is dened by
D (g) C RT ln.f =p / (Eq. 7.8.7 on
page 184). By analogy with this equation, the fugacity fi of substance i in a real gas
mixture is dened by the relation
f
def
i
i (g)
D (g) C RT ln i or fi D p exp
(9.3.12)
i
i
p
RT
(gas mixture)
Just as the fugacity of a pure gas is a kind of effective pressure, the fugacity of a constituent
of a gas mixture is a kind of effective partial pressure. That is, fi is the partial pressure
substance i would have in an ideal gas mixture that is at the same temperature as the real
gas mixture and in which the chemical potential of i is the same as in the real gas mixture.
To derive a relation allowing us to evaluate fi from the pressurevolume properties
of the gaseous mixture, we follow the steps described for a pure gas in Sec. 7.8.1. The
temperature and composition are constant. From Eq. 9.3.12, the difference between the
chemical potentials of substance i in the mixture at pressures p 0 and p 00 is
0
i
Integration of d
00
i
D RT ln
fi0
fi00
(9.3.13)
D Vi dp (from Eq. 9.2.49) between these pressures yields

0
i
00
i
p0
p 00
Vi dp
(9.3.14)
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
244
00
When we equate these two expressions for 0
i
i , divide both sides by RT , subtract the
identity
Z p0
p0
dp
ln 00 D
(9.3.15)
p
p 00 p
0
and take the ideal-gas behavior limits p 00 !0 and fi00 !yi p 00 D .pi =p 0 /p 00 , we obtain
f0
ln i0 D
pi
The fugacity coefcient
p0
1
dp
p
Vi
RT
(9.3.16)
(gas mixture, constant T )
of constituent i is dened by
def
fi D
i pi
(9.3.17)
(gas mixture)
Accordingly, the fugacity coefcient at pressure p 0 is given by
Z p0
Vi
1
0
ln i .p / D
dp
RT
p
0
(9.3.18)
(gas mixture, constant T )
As p 0 approaches zero, the integral in Eqs. 9.3.16 and 9.3.18 approaches zero, fi0 ap0
proaches pi , and i .p 0 / approaches unity.
Partial molar quantities
By combining Eqs. 9.3.12 and 9.3.16, we obtain
i .p
/D
i (g)
0
pi
C RT ln C
p
p 0
Z
0
Vi
RT
p
dp
(9.3.19)
(gas mixture,
constant T )
which is the analogue for a gas mixture of Eq. 7.9.2 for a pure gas. Section 7.9 describes
the procedure needed to obtain formulas for various molar quantities of a pure gas from
Eq. 7.9.2. By following a similar procedure with Eq. 9.3.19, we obtain the formulas for
differences between partial molar and standard molar quantities of a constituent of a gas
mixture shown in the second column of Table 9.1 on the next page. These formulas are
obtained with the help of Eqs. 9.2.46, 9.2.48, 9.2.50, and 9.2.52.
Equation of state
The equation of state of a real gas mixture can be written as the virial equation
B
C
pV =n D RT 1 C
C
C
.V =n/ .V =n/2
(9.3.20)
This equation is the same as Eq. 2.2.2 for a pure gas, except that the molar volume Vm is
replaced by the mean molar volume V =n, and the virial coefcients B; C; : : : depend on
composition as well as temperature.
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES
245
Table 9.1 Gas mixture: expressions for differences between partial molar and standard molar quantities of constituent i
Difference
i
i (g)
Si
Si (g)
Hi
Hi (g)
Ui
Ui (g)
Cp;i
aB
Cp;i (g)
General expression
at pressure p 0
Z p0
pi0
RT
RT ln C
Vi
dp
p
p
0 "
#
Z p0
pi0
@Vi
R
dp
R ln
p
@T p p
0
#
Z p0 "
@Vi
Vi T
dp
@T p
0
#
Z p0 "
@Vi
Vi T
dp C RT p 0 Vi
@T p
0
Z p0 2
@ Vi
T
dp
@T 2 p
0
Equation of statea
V D nRT =p C nB
pi
C Bi0 p
p
dB 0
pi
R ln p i
p
dT
dBi0
p Bi0 T
dT
RT ln
pT
dBi0
dT
pT
d2 Bi0
dT 2
and Bi0 are dened by Eqs. 9.3.24 and 9.3.26
At low to moderate pressures, the simple equation of state

V =n D
RT
CB
p
(9.3.21)
describes a gas mixture to a sufciently high degree of accuracy (see Eq. 2.2.8 on page 35).
This is equivalent to a compression factor given by
def
Z D
pV
Bp
D1C
nRT
RT
(9.3.22)
From statistical mechanical theory, the dependence of the second virial coefcient B of
a binary gas mixture on the mole fraction composition is given by
2
2
B D yA BAA C 2yA yB BAB C yB BBB
(9.3.23)
(binary gas mixture)
where BAA and BBB are the second virial coefcients of pure A and B, and BAB is a mixed
second virial coefcient. BAA , BBB , and BAB are functions of T only. For a gas mixture
with any number of constituents, the composition dependence of B is given by
XX
BD
yi yj Bij
(9.3.24)
i
(gas mixture, Bij DBj i )
Here Bij is the second virial of i if i and j are the same, or a mixed second virial coefcient
if i and j are different.
If a gas mixture obeys the equation of state of Eq. 9.3.21, the partial molar volume of
constituent i is given by
RT
C Bi0
(9.3.25)
Vi D
p
CHAPTER 9 MIXTURES
9.4 L IQUID AND S OLID M IXTURES OF N ONELECTROLYTES
246
where the quantity Bi0 , in order to be consistent with Vi D .@V =@ni /T;p;nj i , is found to
be given by
X
Bi0 D 2
yj Bij B
(9.3.26)
j
For the constituents of a binary mixture of A and B, Eq. 9.3.26 becomes

0
BA D BAA C . BAA C 2BAB
2
BBB /yB
0
BB D BBB C . BAA C 2BAB
2
BBB /yA
(9.3.27)
(9.3.28)
When we substitute the expression of Eq. 9.3.25 for Vi in Eq. 9.3.18, we obtain a relation
between the fugacity coefcient of constituent i and the function Bi0 :
ln
Bi0 p
RT
(9.3.29)
The third column of Table 9.1 gives formulas for various partial molar quantities of
constituent i in terms of Bi0 and its temperature derivative. The formulas are the same as the
approximate formulas in the third column of Table 7.5 for molar quantities of a pure gas,
with Bi0 replacing the second virial coefcient B.
9.4 LIQUID AND SOLID MIXTURES OF NONELECTROLYTES

Homogeneous liquid and solid mixtures are condensed phases of variable composition.
Most of the discussion of condensed-phase mixtures in this section focuses on liquids.
The same principles, however, apply to homogeneous solid mixtures, often called solid
solutions. These solid mixtures include most metal alloys, many gemstones, and doped
semiconductors.
The relations derived in this section apply to mixtures of nonelectrolytessubstances
that do not dissociate into charged species. Solutions of electrolytes behave quite differently
in many ways, and will be discussed in the next chapter.
9.4.1 Raoults law

In 1888, the French physical chemist Francois Raoult published his nding that when a
dilute liquid solution of a volatile solvent and a nonelectrolyte solute is equilibrated with a
gas phase, the partial pressure pA of the solvent in the gas phase is proportional to the mole
fraction xA of the solvent in the solution:
pA D xA pA
(9.4.1)
Here pA is the saturation vapor pressure of the pure solvent (the pressure at which the pure
liquid and pure gas phases are in equilibrium).
In order to place Raoults law in a rigorous thermodynamic framework, consider the
two systems depicted in Fig. 9.5 on the next page. The liquid phase of system 1 is a binary
CHAPTER 9 MIXTURES
(A+B+C)(g)
A
(A+C)(g)

A
B
247
(A+B)(l)
A(l)
system 1
system 2
Figure 9.5 Two systems with equilibrated liquid and gas phases.
solution of solvent A and solute B, whereas the liquid in system 2 is the pure solvent. In
system 1, the partial pressure pA in the equilibrated gas phase depends on the temperature
and the solution composition. In system 2, pA depends on the temperature. Both pA and
pA have a mild dependence on the total pressure p, which can be varied with an inert gas
constituent C of negligible solubility in the liquid.
Suppose that we vary the composition of the solution in system 1 at constant temperature, while adjusting the partial pressure of C so as to keep p constant. If we nd that
the partial pressure of the solvent over a range of composition is given by pA D xA pA ,
where pA is the partial pressure of A in system 2 at the same T and p, we will say that the
solvent obeys Raoults law for partial pressure in this range. This is the same as the original Raoults law, except that pA is now the vapor pressure of pure liquid A at the pressure
p of the liquid mixture. Section 12.8.1 will show that unless p is much greater than pA ,
pA is practically the same as the saturation vapor pressure of pure liquid A, in which case
Raoults law for partial pressure becomes identical to the original law.
A form of Raoults law with fugacities in place of partial pressures is often more useful:
fA D xA fA , where fA is the fugacity of A in the gas phase of system 2 at the same T and
p as the solution. If this relation is found to hold over a given composition range, we will
say the solvent in this range obeys Raoults law for fugacity.
We can generalize the two forms of Raoults law for any constituent i of a liquid mixture:
pi D xi pi
fi D xi fi
(9.4.2)
(Raoults law for partial pressure)
(9.4.3)
(Raoults law for fugacity)
Here xi is the mole fraction of i in the liquid mixture, and pi and fi are the partial pressure
and fugacity in a gas phase equilibrated with pure liquid i at the same T and p as the liquid
mixture. Both pA and fi are functions of T and p.
These two forms of Raoults law are equivalent when the gas phases are ideal gas mixtures. When it is necessary to make a distinction between the two forms, this book will refer
specically to Raoults law for partial pressure or Raoults law for fugacity.
CHAPTER 9 MIXTURES
248
Raoults law for fugacity can be recast in terms of chemical potential. Section 9.2.7
showed that if substance i has transfer equilibrium between a liquid and a gas phase, its
chemical potential i is the same in both equilibrated phases. The chemical potential in
the gas phase is given by i D (g) C RT ln fi =p (Eq. 9.3.12). Replacing fi by xi fi
i
according to Raoults law, and rearranging, we obtain
fi
(9.4.4)
i D
i (g) C RT ln C RT ln xi
p
The expression in brackets is independent of the mixture composition. We replace this
expression by a quantity i , a function of T and p, and write
i
C RT ln xi
(9.4.5)
Equation 9.4.5 is an expression for the chemical potential in the liquid phase when Raoults
law for fugacity is obeyed. By setting xi equal to 1, we see that i represents the chemical
potential of pure liquid i at the temperature and pressure of the mixture. Because Eq. 9.4.5
is valid for any constituent whose fugacity obeys Eq. 9.4.3, it is equivalent to Raoults law
for fugacity for that constituent.
9.4.2 Ideal mixtures

Depending on the temperature, pressure, and identity of the constituents of a liquid mixture,
Raoults law for fugacity may hold for constituent i at all liquid compositions, or over only
a limited composition range when xi is close to unity.
An ideal liquid mixture is dened as a liquid mixture in which, at a given temperature
and pressure, each constituent obeys Raoults law for fugacity (Eq. 9.4.3 or 9.4.5) over the
entire range of composition. Equation 9.4.3 applies only to a volatile constituent, whereas
Eq. 9.4.5 applies regardless of whether the constituent is volatile.
Few liquid mixtures are found to approximate the behavior of an ideal liquid mixture.
In order to do so, the constituents must have similar molecular size and structure, and the
pure liquids must be miscible in all proportions. Benzene and toluene, for instance, satisfy
these requirements, and liquid mixtures of benzene and toluene are found to obey Raoults
law quite closely. In contrast, water and methanol, although miscible in all proportions,
form liquid mixtures that deviate considerably from Raoults law. The most commonly
encountered situation for mixtures of organic liquids is that each constituent deviates from
Raoults law behavior by having a higher fugacity than predicted by Eq. 9.4.3a positive
deviation from Raoults law.
Similar statements apply to ideal solid mixtures. In addition, a relation with the same
form as Eq. 9.4.5 describes the chemical potential of each constituent of an ideal gas mixture, as the following derivation shows. In an ideal gas mixture at a given T and p, the
chemical potential of substance i is given by Eq. 9.3.5:
pi
yi p
(9.4.6)
i D i (g) C RT ln D i (g) C RT ln
p
p
Here yi is the mole fraction of i in the gas. The chemical potential of the pure ideal gas
(yi D1) is
p
(9.4.7)
i D i (g) C RT ln
p
CHAPTER 9 MIXTURES
249
By eliminating (g) between these equations and rearranging, we obtain Eq. 9.4.5 with xi
i
replaced by yi .
Thus, an ideal mixture, whether solid, liquid, or gas, is a mixture in which the chemical
potential of each constituent at a given T and p is a linear function of the logarithm of the
mole fraction:
i
C RT ln xi
(9.4.8)
(ideal mixture)
9.4.3 Partial molar quantities in ideal mixtures

With the help of Eq. 9.4.8 for the chemical potential of a constituent of an ideal mixture,
we will now be able to nd expressions for partial molar quantities. These expressions nd
their greatest use for ideal liquid and solid mixtures.
For the partial molar entropy of substance i, we have Si D .@ i =@T /p;fnig (from Eq.
9.2.48) or, for the ideal mixture,
@ i
Si D
R ln xi D Si
R ln xi
(9.4.9)
@T p
(ideal mixture)
Since ln xi is negative in a mixture, the partial molar entropy of a constituent of an ideal

mixture is greater than the molar entropy of the pure substance at the same T and p.
For the partial molar enthalpy, we have Hi D i C T Si (from Eq. 9.2.46). Using the
expressions for i and Si gives us
Hi D
C T Si D Hi
(9.4.10)
(ideal mixture)
Thus, Hi in an ideal mixture is independent of the mixture composition and is equal to the
molar enthalpy of pure i at the same T and p as the mixture. In the case of an ideal gas
mixture, Hi is also independent of p, because the molar enthalpy of an ideal gas depends
only on T .
The partial molar volume is given by Vi D .@ i =@p/T;fnig (Eq. 9.2.49), so we have
@ i
Vi D
D Vi
(9.4.11)
@p T
(ideal mixture)
Finally, from Eqs. 9.2.50 and 9.2.52 and the expressions above for Hi and Vi , we obtain
Ui D Hi
pVi D Ui
(9.4.12)
(ideal mixture)
and
Cp;i D .@Hi =@T /p;fnig D Cp;i
(9.4.13)
(ideal mixture)
Note that in an ideal mixture held at constant T and p, the partial molar quantities Hi , Vi ,
Ui , and Cp;i do not vary with the composition. This is basically because of the absence of
intermolecular forces in an ideal mixture.
CHAPTER 9 MIXTURES
gas mixture
fi
liquid mixture
250
xi
Figure 9.6 Equilibrated liquid and gas mixtures. Substance i is present in both
phases.
9.4.4 Henrys law

Consider the system shown in Fig. 9.6, in which a liquid mixture is equilibrated with a gas
phase. Transfer equilibrium exists for substance i, a constituent of both phases. Substance
i is assumed to have the same molecular form in both phases, and is not, for instance, an
electrolyte. We can vary the mole fraction xi in the liquid and evaluate the fugacity fi in
the gas phase.
Suppose we allow xi to approach zero at constant T and p while the relative amounts
of the other liquid constituents remain constant. It is found experimentally that the fugacity
fi becomes proportional to xi :
fi ! kH;i xi
as
xi ! 0
(9.4.14)
(constant T and p)
This behavior is called Henrys law. The proportionality constant kH;i is the Henrys
law constant of substance i. The value of kH;i depends on the temperature and the total
pressure, and also on the relative amounts of the constituents other than i in the liquid
mixture.
If the liquid phase happens to be an ideal liquid mixture, then by denition constituent
i obeys Raoults law for fugacity at all values of xi . In that case, kH;i is equal to fi , the
fugacity when the gas phase is equilibrated with pure liquid i at the same temperature and
pressure as the liquid mixture.
If we treat the liquid mixture as a binary solution in which solute B is a volatile nonelectrolyte, Henrys law behavior occurs in the limit of innite dilution:
fB ! kH,B xB
as xB ! 0
(9.4.15)
(constant T and p)
An example of this behavior is shown in Fig. 9.7(a) on page 252. The limiting slope of the
plot of fB versus xB is nite, not zero or innite. (The fugacity of a volatile electrolyte, such
as HCl dissolved in water, displays a much different behavior, as will be shown in Chap.
10.)
Equation 9.4.15 can be applied to a solution of more than one solute if the combination
of constituents other than B is treated as the solvent, and the relative amounts of these
constituents remain constant as xB is varied.
Since the mole fraction, concentration, and molality of a solute become proportional to
one another in the limit of innite dilution (Eq. 9.1.14), in a very dilute solution the fugacity
CHAPTER 9 MIXTURES
251
BIOGRAPHICAL SKETCH
William Henry (17741836)

the gas to dissolve in the water. His ndings
were presented to the Royal Society of London
in 1802 and published the following year:a
William Henry was a British chemist, trained

as a physician, who is best known for his formulation of what is now called Henrys law.
Henry was born in Manchester, England.
His father was an apothecary and industrial
chemist who established a protable business
manufacturing such products as magnesium
carbonate (used as an antacid) and carbonated
water. At the age of ten, Henry was severely
injured by a falling beam and was plagued by
pain and ill health for the rest of his life.
Henry began medical studies at the University of Edinburgh in 1795. He interrupted these
studies to do chemical research, to assist his
father in general medical practice, and to help
run the family chemical business. He nally
received his diploma of Doctor in Medicine in
1807. In 1809, in recognition of his research
papers, he was elected a Fellow of the Royal
Society.
In 1801 the rst edition of his inuential
chemistry textbook appeared, originally called
An Epitome of Chemistry and in later editions
Elements of Experimental Chemistry. The
book went through eleven editions over a period of 28 years.
Henry investigated the relation between the
pressure of a gas and the volume of the gas,
measured at that pressure, that was absorbed
into a given volume of water. He used a simple
apparatus in which the water and gas were conned over mercury in a graduated glass vessel,
and the contents agitated to allow a portion of
a Ref.
[75].
b These
The results of a series of at least fty experiments, on carbonic acid, sulphuretted hydrogen
gas, nitrous oxide, oxygenous and azotic gases,b
with the above apparatus, establish the following general law: that, under equal circumstances
of temperature, water takes up, in all cases, the
same volume of condensed gas as of gas under ordinary pressure. But, as the spaces occupied by every gas are inversely as the compressing force, it follows, that water takes up, of
gas condensed by one, two, or more additional
atmospheres, a quantity which, ordinarily compressed, would be equal to twice, thrice, &c. the
volume absorbed under the common pressure of
the atmosphere.
Henry later conrmed a suggestion made by

his close friend John Dalton, that the amount
of a constituent of a gaseous mixture that is absorbed is proportional to its partial pressure.c
Henry carried out other important work,
chiey on gases, including the elemental compositions of hydrogen chloride, ammonia, and
methane.
Because of his poor health and unsuccessful surgery on his hands, Henry was unable to
continue working in the lab after 1824. Twelve
years later, suffering from pain and depression,
he committed suicide.
In a biography published the year after
Henrys death, his son William Charles Henry
wrote:d
In the general intercourse of society, Dr. Henry
was distinguished by a polished courtesy, by an
intuitive propriety, and by a considerate forethought and respect for the feelings and opinions
of others. . . His comprehensive range of thought
and knowledge, his proneness to general speculation in contradistinction to detail, his ready
command of the renements of language and the
liveliness of his feelings and imagination rendered him a most instructive and engaging companion.
gases are respectively CO2 , H2 S, N2 O, O2 , and N2 .
c Ref.
[76].
d Quoted
in Ref.
[154].
CHAPTER 9 MIXTURES
reference state
H,B
bar
bar
B B
252
(a)
(b)
Figure 9.7 Liquid solutions of 2,3-dimethylbutane (B) in cyclooctane at 298:15 K

and 1 bar. a
(a) Fugacity of B in an equilibrated gas phase as a function of solution composition.
The dashed line, tangent to the curve at xB D 0, is Henrys law behavior, and its slope
is kH,B .
(b) Fugacity divided by mole fraction as a function of composition; the limiting value
at xB D 0 is the Henrys law constant kH,B .
a Based
is proportional to all three of these composition variables. This leads to three versions of
Henrys law:
mole fraction basis
concentration basis
molality basis
fB D kH,B xB
fB D kc;B cB
fB D km;B mB
(9.4.16)
(nonelectrolyte solute
at innite dilution)
(9.4.17)
at innite dilution)
(9.4.18)
at innite dilution)
In these equations kH,B , kc;B , and km;B are Henrys law constants dened by

fB
def
mole fraction basis kH,B D lim
xB !0 xB

fB
def
concentration basis kc;B D lim
cB !0 cB
fB
def
molality basis
km;B D lim
mB !0 mB
(9.4.19)
(9.4.20)
(9.4.21)
CHAPTER 9 MIXTURES
253
Note that the Henrys law constants are not dimensionless, and are functions of T and p.
To evaluate one of these constants, we can plot fB divided by the appropriate composition
variable as a function of the composition variable and extrapolate to innite dilution. The
evaluation of kH,B by this procedure is illustrated in Fig. 9.7(b).
Relations between these Henrys law constants can be found with the use of Eqs. 9.1.14
and 9.4.169.4.18:
kc;B D VA kH,B
km;B D MA kH,B
(9.4.22)
9.4.5 The ideal-dilute solution

An ideal-dilute solution is a real solution that is dilute enough for each solute to obey
Henrys law. On the microscopic level, the requirement is that solute molecules be sufciently separated to make solutesolute interactions negligible.
Note that an ideal-dilute solution is not necessarily an ideal mixture. Few liquid mixtures behave as ideal mixtures, but a solution of any nonelectrolyte solute becomes an idealdilute solution when sufciently dilute.
Within the composition range that a solution effectively behaves as an ideal-dilute
solution, then, the fugacity of solute B in a gas phase equilibrated with the solution is
proportional to its mole fraction xB in the solution. The chemical potential of B in the
gas phase, which is equal to that of B in the liquid, is related to the fugacity by B D
B (g) C RT ln.fB =p / (Eq. 9.3.12). Substituting fB D kH,B xB (Henrys law) into this
equation, we obtain
B
kH,B xB
p
kH,B
D
C RT ln xB
B (g) C RT ln
p
B (g)
C RT ln
(9.4.23)
where the composition variable xB is segregated in the last term on the right side.
The expression in brackets in Eq. 9.4.23 is a function of T and p, but not of xB , and
represents the chemical potential of B in a hypothetical solute reference state. This chemical
potential will be denoted by ref , where the x in the subscript reminds us that the reference
x;B
state is based on mole fraction. The equation then becomes
B .T; p/
ref
x;B .T; p/
Here the notation emphasizes the fact that
and
C RT ln xB
ref
x;B
(9.4.24)
(ideal-dilute solution
of a nonelectrolyte)
are functions of T and p.
Equation 9.4.24, derived using fugacity, is valid even if the solute has such low volatility that its fugacity in an equilibrated gas phase is too low to measure. In principle, no
solute is completely nonvolatile, and there is always a nite solute fugacity in the gas
phase even if immeasurably small.
It is worthwhile to describe in detail the reference state to which ref refers. The
x;B
general concept is also applicable to other solute reference states and solute standard
states to be encountered presently. Imagine a hypothetical solution with the same
constituents as the real solution. This hypothetical solution has the magical property
CHAPTER 9 MIXTURES
254
that it continues to exhibit the ideal-dilute behavior described by Eq. 9.4.24, even when
xB increases beyond the ideal-dilute range of the real solution. The reference state is
the state of this hypothetical solution at xB D1. It is a ctitious state in which the mole
fraction of B is unity and B behaves as in an ideal-dilute solution, and is sometimes
called the ideal-dilute solution of unit solute mole fraction.
By setting xB equal to unity in Eq. 9.4.24, so that ln xB is zero, we see that ref
x;B
is the chemical potential of B in the reference state. In a gas phase equilibrated with
the hypothetical solution, the solute fugacity fB increases as a linear function of xB all
the way to xB D1, unlike the behavior of the real solution (unless it happens to be an
ideal mixture). In the reference state, fB is equal to the Henrys law constant kH,B ; an
example is indicated by the lled circle in Fig. 9.7(a).
By similar steps, combining Henrys law based on concentration or molality (Eqs.

9.4.17 and 9.4.18) with the relation B D (g) C RT ln.fB =p /, we obtain for the solute
B
chemical potential in the ideal-dilute range the equations
kc;B cB c
B D B (g) C RT ln
p
c
kc;B c
c
D
C RT ln B
(9.4.25)
B (g) C RT ln
p
c
km;B mB m
D (g) C RT ln
B
B
p
m
km;B m
m
D
(g) C RT ln
C RT ln B
(9.4.26)
B
p
m
Note how in each equation the argument of a logarithm is multiplied and divided by a
constant, c or m , in order to make the arguments of the resulting logarithms dimensionless. These constants are called standard compositions with the following values:
standard concentration c D 1 mol dm 3 (equal to one mole per liter, or one molar)
standard molality m D 1 mol kg 1 (equal to one molal)

Again in each of these equations, we replace the expression in brackets, which depends
on T and p but not on composition, with the chemical potential of a solute reference state:
B .T; p/
B .T; p/
ref
c;B .T; p/
ref
m;B .T; p/
C RT ln
cB
c
C RT ln
mB
m
(9.4.27)
(9.4.28)
The quantities ref and ref are the chemical potentials of the solute in hypothetical referm;B
c;B
ence states that are solutions of standard concentration and standard molality, respectively,
in which B behaves as in an ideal-dilute solution. Section 9.7.1 will show that when the
pressure is the standard pressure, these reference states are solute standard states.
CHAPTER 9 MIXTURES
255
For consistency with Eqs. 9.4.27 and 9.4.28, we can rewrite Eq. 9.4.24 in the form
B .T; p/
ref
x;B .T; p/
C RT ln
xB
x
(9.4.29)
with x , the standard mole fraction, given by x D 1.
9.4.6 Solvent behavior in the ideal-dilute solution

We now use the GibbsDuhem equation to investigate the behavior of the solvent in an
ideal-dilute solution of one or more nonelectrolyte solutes. TheP
GibbsDuhem equation applied to chemical potentials at constant T and p can be written i xi d PD 0 (Eq. 9.2.43).
i
We use subscript A for the solvent, rewrite the equation as xA d A C i A xi d i D 0,
and rearrange to
d
1 X
xi d
xA
iA
(9.4.30)
(constant T and p)
This equation shows how changes in the solute chemical potentials, due to a composition
change at constant T and p, affect the chemical potential of the solvent.
In an ideal-dilute solution, the chemical potential of each solute is given by i D ref C
x;i
RT ln xi and the differential of i at constant T and p is
d
D RT d ln xi D RT dxi =xi
(9.4.31)
(Here the fact has been used that ref is a constant at a given T and p.) When we substitute
x;i
this expression for d i in Eq. 9.4.30, we obtain
d
RT X
dxi
xA
(9.4.32)
iA
P
Now since theP of all mole fractions is 1, we have the relation iA xi D 1 xA whose
sum
differential is iA dxi D dxA . Making this substitution in Eq. 9.4.32 gives us
d
RT
dxA D RT d ln xA
xA
(9.4.33)
of nonelectrolytes)
Consider a process in an open system in which we start with a xed amount of pure
solvent and continuously add the solute or solutes at constant T and p. The solvent mole
0
fraction decreases from unity to a value xA , and the solvent chemical potential changes
0
from A to A . We assume the solution formed in this process is in the ideal-dilute solution
range, and integrate Eq. 9.4.33 over the path of the process:
0
Z 0A
Z xA DxA
d A D RT
d ln xA
(9.4.34)
xA D1
The result is
0
A
0
D RT ln xA , or in general
A
C RT ln xA
(9.4.35)
CHAPTER 9 MIXTURES
kPa
Hen
rys
law
256
law
s
ult
o
Ra
Figure 9.8 Fugacity of ethanol in a gas phase equilibrated with a binary liquid mixture of ethanol (A) and H2 O at 25 C and 1 bar. Open circles: experimental measurements. a The dashed lines show Henrys law behavior and Raoults law behavior.
a Ref.
[45].
Comparison with Eq. 9.4.5 on page 248 shows that Eq. 9.4.35 is equivalent to Raoults law
for fugacity.
Thus, in an ideal-dilute solution of nonelectrolytes each solute obeys Henrys law and
the solvent obeys Raoults law.
An equivalent statement is that a nonelectrolyte constituent of a liquid mixture approaches Henrys law behavior as its mole fraction approaches zero, and approaches Raoults
law behavior as its mole fraction approaches unity. This is illustrated in Fig. 9.8, which
shows the behavior of ethanol in ethanol-water mixtures. The ethanol exhibits positive
deviations from Raoults law and negative deviations from Henrys law.
9.4.7 Partial molar quantities in an ideal-dilute solution

Consider the solvent, A, of a solution that is dilute enough to be in the ideal-dilute range.
In this range, the solvent fugacity obeys Raoults law, and the partial molar quantities of the
solvent are the same as those in an ideal mixture. Formulas for these quantities were given
in Eqs. 9.4.89.4.13 and are collected in the rst column of Table 9.2 on the next page.
The formulas show that the chemical potential and partial molar entropy of the solvent, at
constant T and p, vary with the solution composition and, in the limit of innite dilution
(xA !1), approach the values for the pure solvent. The partial molar enthalpy, volume,
internal energy, and heat capacity, on the other hand, are independent of composition in the
ideal-dilute region and are equal to the corresponding molar quantities for the pure solvent.
Next consider a solute, B, of a binary ideal-dilute solution. The solute obeys Henrys
law, and its chemical potential is given by B D ref C RT ln xB (Eq. 9.4.24) where ref
x;B
x;B
is a function of T and p, but not of composition. B varies with the composition and goes
to 1 as the solution becomes innitely dilute (xA !1 and xB !0).
CHAPTER 9 MIXTURES
257
Table 9.2 Partial molar quantities of solvent and nonelectrolyte solute in an ideal-dilute solution
Solvent
A
Solute
A
SA D SA
C RT ln xA
R ln xA
HA D HA
VA D VA
UA D UA
Cp;A D Cp;A
D ref C RT ln xB
x;B
D ref C RT ln.cB =c /
c;B
D ref C RT ln.mB =m /
m;B
ref
SB D Sx;B R ln xB
ref
D Sc;B R ln.cB =c /
ref
D Sm;B R ln.mB =m /
1
HB D HB
1
VB D VB
1
UB D UB
1
Cp;B D Cp;B
B
For the partial molar entropy of the solute, we use SB D .@

and obtain
!
@ ref
x;B
SB D
R ln xB
@T
B =@T /p;fnig
(Eq. 9.2.48)
(9.4.36)
ref
x;B =@T /p
ref
The term .@
represents the partial molar entropy Sx;B of B in the ctitious
reference state of unit solute mole fraction. Thus, we can write Eq. 9.4.36 in the form
ref
SB D Sx;B
R ln xB
(9.4.37)
This equation shows that the partial molar entropy varies with composition and goes to
C1 in the limit of innite dilution. From the expressions of Eqs. 9.4.27 and 9.4.28, we can
derive similar expressions for SB in terms of the solute reference states on a concentration
or molality basis.
The relation HB D B C T SB (from Eq. 9.2.46), combined with Eqs. 9.4.24 and 9.4.37,
yields
ref
ref
HB D ref C T Sx;B D Hx;B
(9.4.38)
x;B
showing that at constant T and p, the partial molar enthalpy of the solute is constant
throughout the ideal-dilute solution range. Therefore, we can write
1
HB D HB
(9.4.39)
1
where HB is the partial molar enthalpy at innite dilution. By similar reasoning, using
Eqs. 9.2.499.2.52, we nd that the partial molar volume, internal energy, and heat capacity
of the solute are constant in the ideal-dilute range and equal to the values at innite dilution.
The expressions are listed in the second column of Table 9.2.
1
1
1
When the pressure is equal to the standard pressure p , the quantities HB , VB , UB ,
1
and Cp;B are the same as the standard values HB , VB , UB , and Cp;B .
CHAPTER 9 MIXTURES
9.5 ACTIVITY C OEFFICIENTS IN M IXTURES OF N ONELECTROLYTES
258
Table 9.3 Reference states for nonelectrolyte constituents of mixtures. In each reference state, the temperature and pressure are the same as those of the mixture.
Constituent
Reference state
Substance i in a gas mixture

Substance i in a liquid or solid
mixture
Solvent A of a solution
Pure i behaving as an ideal gasa

Pure i in the same physical state as the
mixture
Pure A in the same physical state as the
solution
B at mole fraction 1, behavior extrapolated from innite dilution on a mole fraction basisa
B at concentration c , behavior extrapolated from innite dilution on a concentration basisa
B at molality m , behavior extrapolated
from innite dilution on a molality basisa
Solute B, mole fraction basis
Solute B, concentration basis
Solute B, molality basis

aA
Chemical
potential
ref
i (g)
i
A
ref
x;B
ref
c;B
ref
m;B
hypothetical state.
9.5 ACTIVITY COEFFICIENTS IN MIXTURES OF

NONELECTROLYTES
An activity coefcient of a species is a kind of adjustment factor that relates the actual
behavior to ideal behavior at the same temperature and pressure. The ideal behavior is
based on a reference state for the species.
We begin by describing reference states for nonelectrolytes. The thermodynamic behavior of an electrolyte solution is more complicated than that of a mixture of nonelectrolytes,
and will be discussed in the next chapter.
9.5.1 Reference states and standard states

A reference state of a constituent of a mixture has the same temperature and pressure as the
mixture. When species i is in its reference state, its chemical potential ref depends only
i
on the temperature and pressure of the mixture.
If the pressure is the standard pressure p , the reference state of species i becomes
its standard state. In the standard state, the chemical potential is the standard chemical
potential , which is a function only of temperature.
i
Reference states are useful for derivations involving processes taking place at constant
T and p when the pressure is not necessarily the standard pressure.
Table 9.3 describes the reference states of nonelectrolytes used in this book, and lists
symbols for chemical potentials of substances in these states. The symbols for solutes
include x, c, or m in the subscript to indicate the basis of the reference state.
CHAPTER 9 MIXTURES
259

Since the activity coefcient of a species relates its actual behavior to its ideal behavior at
the same T and p, let us begin by examining behavior in ideal mixtures.
Consider rst an ideal gas mixture at pressure p. The chemical potential of substance i
in this ideal gas mixture is given by Eq. 9.3.5 (the superscript id stands for ideal):
id
i (g)
i (g)
C RT ln
pi
p
(9.5.1)
The reference state of gaseous substance i is pure i acting as an ideal gas at pressure p. Its
chemical potential is given by
ref
i (g)
i (g)
C RT ln
p
p
(9.5.2)
D RT ln
pi
p
(9.5.3)
Subtracting Eq. 9.5.2 from Eq. 9.5.1, we obtain

id
i (g)
ref
i (g)
Consider the following expressions for chemical potentials in ideal mixtures and idealdilute solutions of nonelectrolytes. The rst equation is a rearrangement of Eq. 9.5.3, and
the others are from earlier sections of this chapter.7
Constituent of an ideal gas mixture
Constituent of an ideal liquid or solid mixture
Solvent of an ideal-dilute solution
Solute, ideal-dilute solution, mole fraction basis
id
i (g)
id
i D
id
A D
id
B D
Solute, ideal-dilute solution, concentration basis
id
B
Solute, ideal-dilute solution, molality basis
id
B
D
i
ref
i (g)
C RT ln
C RT ln xi
A C RT ln xA
ref
x;B C RT ln xB
c
D ref C RT ln B
c;B
c
m
D ref C RT ln B
m;B
m
Note that the equations for the condensed phases have the general form
composition variable
id
ref
i D i C RT ln
standard composition
pi
p
(9.5.4)
(9.5.5)
(9.5.6)
(9.5.7)
(9.5.8)
(9.5.9)
(9.5.10)
where ref is the chemical potential of component i in an appropriate reference state. (The
i
standard composition on a mole fraction basis is x D1.)
9.5.3 Real mixtures

If a mixture is not ideal, we can write an expression for the chemical potential of each
component that includes an activity coefcient. The expression is like one of those for the
ideal case (Eqs. 9.5.49.5.9) with the activity coefcient multiplying the quantity within the
logarithm.
7 In
order of occurrence, Eqs. 9.4.8, 9.4.35, 9.4.24, 9.4.27, and 9.4.28.
CHAPTER 9 MIXTURES
Consider constituent i of a gas mixture. If we eliminate

9.5.2, we obtain
D
ref
i (g)
C RT ln
ref
i (g)
C RT ln
i (g)
260
from Eqs. 9.3.12 and
fi
p
i pi
(9.5.11)
where fi is the fugacity of constituent i and i is its fugacity coefcient. Here the activity
coefcient is the fugacity coefcient i .
For components of a condensed-phase mixture, we write expressions for the chemical
potential having a form similar to that in Eq. 9.5.10, with the composition variable now
multiplied by an activity coefcient:
composition variable
ref
(9.5.12)
i D i C RT ln .activity coefcient of i/
standard composition
The activity coefcient of a species is a dimensionless quantity whose value depends
on the temperature, the pressure, the mixture composition, and the choice of the reference
state for the species. Under conditions in which the mixture behaves ideally, the activity
coefcient is unity and the chemical potential is given by one of the expressions of Eqs.
9.5.49.5.9; otherwise, the activity coefcient has the value that gives the actual chemical
potential.
This book will use various symbols for activity coefcients, as indicated in the following
list of expressions for the chemical potentials of nonelectrolytes:
pi
ref
Constituent of a gas mixture
(9.5.13)
i D i (g) C RT ln
i
p
(9.5.14)
Constituent of a liquid or solid mixture
i D i C RT ln . i xi /
Solvent of a solution
Solute of a solution, mole fraction basis
Solute of a solution, concentration basis
Solute of a solution, molality basis
A C RT ln . A xA /
ref
x;B C RT ln x;B xB
c
D ref C RT ln c;B B
c;B
c
m
D ref C RT ln m;B B
m;B
m
(9.5.15)
(9.5.16)
(9.5.17)
(9.5.18)
Equation 9.5.14 refers to a component of a liquid or solid mixture of substances that

mix in all proportions. Equation 9.5.15 refers to the solvent of a solution. The reference
states of these components are the pure liquid or solid at the temperature and pressure of
the mixture. For the activity coefcients of these components, this book uses the symbols
i and A .
The IUPAC Green Book (Ref. [36], p. 59) recommends the symbol fi for the activity
coefcient of component i when the reference state is the pure liquid or solid. This
book instead uses symbols such as i and A , in order to avoid confusion with the
symbol usually used for fugacity, fi .
In Eqs. 9.5.169.5.18, the symbols x;B , c;B , and m;B for activity coefcients of a
nonelectrolyte solute include x, c, or m in the subscript to indicate the choice of the solute
CHAPTER 9 MIXTURES
261
reference state. Although three different expressions for B are shown, for a given solution
composition they must all represent the same value of B , equal to the rate at which the
Gibbs energy increases with the amount of substance B added to the solution at constant T
and p. The value of a solute activity coefcient, on the other hand, depends on the choice
of the solute reference state.
You may nd it helpful to interpret products appearing on the right sides of Eqs. 9.5.13
9.5.18 as follows.
i pi is an effective partial pressure.
i xi ,
A xA ,
c;B cB
is an effective concentration.
and
x;B xB
are effective mole fractions.
m;B mB
is an effective molality.
In other words, the value of one of these products is the value of a partial pressure or
composition variable that would give the same chemical potential in an ideal mixture as
the actual chemical potential in the real mixture. These effective composition variables
are an alternative way to express the escaping tendency of a substance from a phase; they
are related exponentially to the chemical potential, which is also a measure of escaping
tendency.
A change in pressure or composition that causes a mixture to approach the behavior of
an ideal mixture or ideal-dilute solution must cause the activity coefcient of each mixture
constituent to approach unity:
Constituent of a gas mixture
Constituent of a liquid or solid mixture
Solvent of a solution
Solute of a solution, mole fraction basis
x;B
Solute of a solution, concentration basis
c;B
Solute of a solution, molality basis
m;B
!1
as p ! 0
(9.5.19)
!1
as xA ! 1
(9.5.21)
as cB ! 0
(9.5.23)
!1
!1
!1
!1
as xi ! 1
(9.5.20)
as xB ! 0
(9.5.22)
as mB ! 0
(9.5.24)
9.5.4 Nonideal dilute solutions

How would we expect the activity coefcient of a nonelectrolyte solute to behave in a
dilute solution as the solute mole fraction increases beyond the range of ideal-dilute solution
behavior?
The following argument is based on molecular properties at constant T and p.
We focus our attention on a single solute molecule. This molecule has interactions with nearby solute molecules. Each interaction depends on the intermolecular
distance and causes a change in the internal energy compared to the interaction of the
solute molecule with solvent at the same distance.8 The number of solute molecules
in a volume element at a given distance from the solute molecule we are focusing
on is proportional to the local solute concentration. If the solution is dilute and the
8 In
Sec. 11.1.5, it will be shown that roughly speaking the internal energy change is negative if the average
of the attractive forces between two solute molecules and two solvent molecules is greater than the attractive
force between a solute molecule and a solvent molecule at the same distance, and is positive for the opposite
situation.
CHAPTER 9 MIXTURES
9.6 E VALUATION OF ACTIVITY C OEFFICIENTS
262
interactions weak, we expect the local solute concentration to be proportional to the

macroscopic solute mole fraction. Thus, the partial molar quantities UB and VB of the
solute should be approximately linear functions of xB in a dilute solution at constant
T and p.
From Eqs. 9.2.46 and 9.2.50, the solute chemical potential is given by B D UB C
pVB T SB . In the dilute solution, we assume UB and VB are linear functions of xB
as explained above. We also assume the dependence of SB on xB is approximately the
same as in an ideal mixture; this is a prediction from statistical mechanics for a mixture
in which all molecules have similar sizes and shapes. Thus we expect the deviation
of the chemical potential from ideal-dilute behavior, B D ref C RT ln xB , can be
x;B
described by adding a term proportional to xB : B D ref C RT ln xB C kx xB , where
x;B
kx is a positive or negative constant related to solute-solute interactions.
If we equate this expression for B with the one that denes the activity coefcient, B D ref C RT ln. x;B xB / (Eq. 9.5.16), and solve for the activity coefcient,
x;B
we obtain the relation9 x;B D exp .kx xB =RT /. An expansion of the exponential in
powers of xB converts this to
x;B
D 1 C .kx =RT /xB C
(9.5.25)
Thus we predict that at constant T and p, x;B is a linear function of xB at low xB .

An ideal-dilute solution, then, is one in which xB is much smaller than RT =kx so that
x;B is approximately 1. An ideal mixture requires the interaction constant kx to be
zero.
By similar reasoning, we reach analogous conclusions for solute activity coefcients on a concentration or molality basis. For instance, at low mB the chemical
potential of B should be approximately ref C RT ln.mB =m / C km mB , where km is
m;B
a constant at a given T and p; then the activity coefcient at low mB is given by
m;B
D exp .km mB =RT / D 1 C .km =RT /mB C
(9.5.26)
The prediction from the theoretical argument above, that a solute activity coefcient in a
dilute solution is a linear function of the composition variable, is borne out experimentally
as illustrated in Fig. 9.10 on page 265. This prediction applies only to a nonelectrolyte
solute; for an electrolyte, the slope of activity coefcient versus molality approaches 1
at low molality (page 291).
9.6 EVALUATION OF ACTIVITY COEFFICIENTS

This section describes several methods by which activity coefcients of nonelectrolyte substances may be evaluated. Section 9.6.3 describes an osmotic coefcient method that is also
suitable for electrolyte solutes, as will be explained in Sec. 10.6.
9.6.1 Activity coefcients from gas fugacities

Suppose we equilibrate a liquid mixture with a gas phase. If component i of the liquid
mixture is a volatile nonelectrolyte, and we are able to evaluate its fugacity fi in the gas
phase, we have a convenient way to evaluate the activity coefcient i in the liquid. The
relation between i and fi will now be derived.
9 This
is essentially the result of the McMillanMayer solution theory from statistical mechanics.
CHAPTER 9 MIXTURES
263
When component i is in transfer equilibrium between two phases, its chemical potential
is the same in both phases. Equating expressions for i in the liquid mixture and the
equilibrated gas phase (from Eqs. 9.5.14 and 9.5.11, respectively), and then solving for i ,
we have
ref
(9.6.1)
i C RT ln . i xi / D i (g) C RT ln .fi =p/
#
"
ref
fi
i (g)
i
(9.6.2)
D exp
i
RT
xi p
On the right side of Eq. 9.6.2, only fi and xi depend on the liquid composition. We can
therefore write
fi
(9.6.3)
i D Ci
xi
where Ci is a factor whose value depends on T and p, but not on the liquid composition.
Solving Eq. 9.6.3 for Ci gives Ci D i xi =fi .
Now consider Eq. 9.5.20 on page 261. It says that as xi approaches 1 at constant T and
p, i also approaches 1. We can use this limit to evaluate Ci :
Ci D lim
xi !1
i xi
fi
1
fi
(9.6.4)
Here fi is the fugacity of i in a gas phase equilibrated with pure liquid i at the temperature
and pressure of the mixture. Then substitution of this value of Ci (which is independent of
xi ) in Eq. 9.6.3 gives us an expression for i at any liquid composition:
i
fi
xi f i
(9.6.5)
We can follow the same procedure for a solvent or solute of a liquid solution. We replace the left side of Eq. 9.6.1 with an expression from among Eqs. 9.5.159.5.18, then
derive an expression analogous to Eq. 9.6.3 for the activity coefcient with a compositionindependent factor, and nally apply the limiting conditions that cause the activity coefcient to approach unity (Eqs. 9.5.219.5.24) and allow us to evaluate the factor. When we
take the limits that cause the solute activity coefcients to approach unity, the ratios fB =xB ,
fB =cB , and fB =mB become Henrys law constants (Eqs. 9.4.199.4.21). The resulting expressions for activity coefcients as functions of fugacity are listed in Table 9.4 on the next
page.
Examples
Ethanol and water at 25 C mix in all proportions, so we can treat the liquid phase as a liquid
mixture rather than a solution. A plot of ethanol fugacity versus mole fraction at xed T
and p, shown earlier in Fig. 9.8, is repeated in Fig. 9.9(a) on the next page. Ethanol
is component A. In the gure, the lled circle is the pure-liquid reference state at xA D1
where fA is equal to fA . The open circles at xA D 0:4 indicate fA , the actual fugacity in
a gas phase equilibrated with a liquid mixture of this composition, and xA fA , the fugacity
the ethanol would have if the mixture were ideal and component A obeyed Raoults law.
The ratio of these two quantities is the activity coefcient A .
CHAPTER 9 MIXTURES
264
3
A
6
fA =kPa
fA
2
0

xA f A
0:2
0:4
0:6
0:8
1:0
0:2
0:4
0:6
xA
(a)
0:8
1:0
xA
(b)
Figure 9.9 Liquid mixtures of ethanol (A) and H2 O at 25 C and 1 bar.

(a) Ethanol fugacity as a function of mixture composition. The dashed line is Raoults
law behavior, and the lled circle is the pure-liquid reference state.
(b) Ethanol activity coefcient as a function of mixture composition.
Table 9.4 Activity coefcients as functions of fugacity. For a constituent of a condensed-phase mixture, fi ,
fA , and fB refer to the fugacity in a gas phase equilibrated with the condensed phase.
Substance
Activity coefcient
fi
pi

mixture
fi
xi f i
fA
xA f A
x;B
fB
kH,B xB
c;B
fB
kc;B cB
m;B
fB
km;B mB
Figure 9.9(b) shows how A varies with composition. The open circle is at xA D 0:4
and A D fA =.xA fA /. Note how A approaches 1 as xA approaches 1, as it must according
to Eq. 9.5.20.
Water and 1-butanol are two liquids that do not mix in all proportions; that is, 1-butanol
has limited solubility in water. Figures 9.10(a) and 9.10(b) on the next page show the
fugacity of 1-butanol plotted as functions of both mole fraction and molality. The gures
demonstrate how, treating 1-butanol as a solute, we locate the solute reference state by a
CHAPTER 9 MIXTURES
265
kPa
kPa
B mol kg
(a)
(b)

B
(c)
mol kg
(d)
Figure 9.10 Dilute aqueous solutions of 1-butanol (B) at 50:08 C and 1 bar. a
(a) fB in an equilibrated gas phase as a function of xB , measured up to the solubility
limit at xB D 0:015. The dilute region is shown in a magnied view. Dashed line:
Henrys law behavior on a mole fraction basis. Filled circle: solute reference state
based on mole fraction.
(b) fB as a function of mB , measured up to the solubility limit at mB D 0:85 mol kg 1 .
Dashed line: Henrys law behavior on a molality basis. Filled circle: solute reference
state on this basis.
(c) Activity coefcient on a mole fraction basis as a function of xB .
(d) Activity coefcient on a molality basis as a function of mB .
a Based
linear extrapolation of the fugacity to the standard composition. The fugacity fB is quite
different in the two reference states. At the reference state indicated by a lled circle in
Fig. 9.10(a), fB equals the Henrys law constant kH,B ; at the reference state in Fig. 9.10(b),
fB equals km;B m . Note how the activity coefcients plotted in Figs. 9.10(c) and 9.10(d)
approach 1 at innite dilution, in agreement with Eqs. 9.5.22 and 9.5.24, and how they
vary as a linear function of xB or mB in the dilute solution as predicted by the theoretical
argument of Sec. 9.5.4.
9.6.2 Activity coefcients from the GibbsDuhem equation

If component B of a binary liquid mixture has low volatility, it is not practical to use its fugacity in a gas phase to evaluate its activity coefcient. If, however, component A is volatile
CHAPTER 9 MIXTURES
266
enough for fugacity measurements over a range of liquid composition, we can instead use
the GibbsDuhem equation for this purpose.
Consider a binary mixture of two liquids that mix in all proportions. We assume that
only component A is appreciably volatile. By measuring the fugacity of A in a gas phase
equilibrated with the binary mixture, we can evaluate its activity coefcient based on a pureliquid reference state: A D fA =.xA fA / (Table 9.4). We wish to use the same fugacity
measurements to determine the activity coefcient of the nonvolatile component, B.
The GibbsDuhem equation for a binary liquid mixture in the form given by Eq. 9.2.43
is
xA d A C xB d B D 0
(9.6.6)
where d A and d B are the chemical potential changes accompanying a change of composition at constant T and p. Taking the differential at constant T and p of A D A C
RT ln. A xA / (Eq. 9.5.14), we obtain
d
D RT d ln
C RT d ln xA D RT d ln
RT
dxA
xA
(9.6.7)
For component B, we obtain in the same way

d
D RT d ln
RT
dxB D RT d ln
xB
Substituting these expressions for d

obtain the following relation:
and d
d ln
DD
xA
d ln
xB
Integration from xB D 1, where

ln
0
xB DxB
xB D1
RT
dxA
xB
(9.6.8)
in Eq. 9.6.6 and solving for d ln
equals 1 and ln
0
B .xB /
B,
we
(9.6.9)
0
equals 0, to composition xB gives
xA
d ln
xB
(9.6.10)
(binary mixture,
constant T and p)
Equation 9.6.10 allows us to evaluate the activity coefcient of the nonvolatile component,
B, at any given liquid composition from knowledge of the activity coefcient of the volatile
component A as a function of composition.
Next consider a binary liquid mixture in which component B is neither volatile nor able
to mix in all proportions with A. In this case, it is appropriate to treat B as a solute and
to base its activity coefcient on a solute reference state. We could obtain an expression
for ln x;B similar to Eq. 9.6.10, but the integration would have to start at xB D0 where the
integrand xA =xB would be innite. Instead, it is convenient in this case to use the method
described in the next section.
9.6.3 Activity coefcients from osmotic coefcients

It is customary to evaluate the activity coefcient of a nonvolatile solute with a function
m called the osmotic coefcient, or osmotic coefcient on a molality basis. The osmotic
CHAPTER 9 MIXTURES
267
coefcient of a solution of nonelectrolyte solutes is dened by

def
m
RTMA
A
X
(9.6.11)
mi
(nonelectrolyte solution)
i A
P
The denition of m in Eq. 9.6.11 has the following signicance. The sum i A mi is
the total molality of all solute species. In an ideal-dilute solution, the solvent chemical
potential is A D A C RT ln xA . The expansion of the function ln xA in powers of
.1 xA / gives the power series ln xA D .1 xA / .1 xA /2 =2 .1 xA /3 =3
.
P
Thus, in a very dilute solution we have ln xA
.1 xA / D
xi . In the limit
i A
of innite dilution, the mole fraction of solute i becomes xi D MA mi (see Eq. 9.1.14).
In the limit of innite dilution, therefore, we have
X
ln xA D MA
mi
(9.6.12)
(innite dilution)
i A
and the solvent chemical potential is related to solute molalities by
X
RTMA
mi
A D A
i A
(9.6.13)
(innite dilution)
The deviation of m from unity is a measure of the deviation of A from innitedilution behavior, as we can see by comparing the preceding equation with a rearrangement of Eq. 9.6.11:
X
mi
(9.6.14)
A D A
m RTMA
i A
The reason m is called the osmotic coefcient has to do with its relation to the osmotic
pressure of the solution: The ratio =mB is equal to the product of m and the
limiting value of =mB at innite dilution (see Sec. 12.4.4).
Evaluation of
Any method that measures A

A , the lowering of the solvent chemical potential caused
by the presence of a solute or solutes, allows us to evaluate m through Eq. 9.6.11.
The chemical potential of the solvent in a solution is related to the fugacity in an equilibrated gas phase by A D ref (g) C RT ln.fA =p/ (from Eq. 9.5.11). For the pure solvent,
A
this relation is A D ref (g) C RT ln.fA =p/. Taking the difference between these two
A
equations, we obtain
fA
(9.6.15)
A D RT ln
A
fA
which allows us to evaluate m from fugacity measurements.
Osmotic coefcients can also be evaluated from freezing point and osmotic pressure
measurements that will be described in Sec. 12.2.
CHAPTER 9 MIXTURES
Use of
268
Suppose we have a solution of a nonelectrolyte solute B whose activity coefcient

wish to evaluate as a function of mB . For a binary solution, Eq. 9.6.11 becomes
m
Solving for
d
From
m;B
we
(9.6.16)
RTMA mB
(binary nonelectrolyte solution)
and taking its differential at constant T and p, we obtain
ref
m;B
RTMA d.
C RT ln.
d
m mB /
m;B mB =m /
D RT d ln
A
B
m;B
m dmB
C mB d
m/
(9.6.17)
(Eq. 9.5.18), we obtain
m;B mB
m
We substitute these expressions for d

given by Eq. 9.2.26, nA d A C nB d
rearrange to
d ln
RTMA .
D RT
d ln
m;B C
dmB
mB
(9.6.18)
and d B in the GibbsDuhem equation in the form

D 0, make the substitution nA MA D nB =mB , and
Dd
mB
dmB
(9.6.19)
We integrate both sides of this equation for a composition change at constant T and p from
mB D 0 (where ln xB is 0 and m is 1) to any desired molality m0 , with the result
B
ln
0
m;B .mB /
0
m .mB /
1C
m0
B
0
mB
dmB
(9.6.20)
(binary
nonelectrolyte solution)
When the solute is a nonelectrolyte, the integrand . m 1/=mB is found to be a slowly

varying function of mB and to approach a nite value as mB approaches zero.
Once m has been measured as a function of molality from zero up to the molality
of interest, Eq. 9.6.20 can be used to evaluate the solute activity coefcient m;B at that
molality.
Figure 9.11(a) on the next page shows the function . m 1/=mB for aqueous sucrose
solutions over a wide range of molality. The dependence of the solute activity coefcient on
molality, generated from Eq. 9.6.20, is shown in Fig. 9.11(b). Figure 9.11(c) is a plot of the
effective sucrose molality m;B mB as a function of composition. Note how the activity coefcient becomes greater than unity beyond the ideal-dilute region, and how in consequence
the effective molality m;B mB becomes considerably greater than the actual molality mB .
9.6.4 Fugacity measurements

Section 9.6.1 described the evaluation of the activity coefcient of a constituent of a liquid
mixture from its fugacity in a gas phase equilibrated with the mixture. Section 9.6.3 mentioned the use of solvent fugacities in gas phases equilibrated with pure solvent and with a
solution, in order to evaluate the osmotic coefcient of the solution.
CHAPTER 9 MIXTURES
269
10
0:08
2:0
1:5
0:06
0:04
1:0
0:02
0:5
m;B
m;B mB =mol kg
1/=mB =kg mol
2:5
.m
0:10
1 2 3 4
mB =mol kg 1
6
4
2
(a)
1 2 3 4
mB =mol kg 1
(b)
1 2 3 4
mB =mol kg 1
(c)
Figure 9.11 Aqueous sucrose solutions at 25 C. a

(a) Integrand of the integral in Eq. 9.6.20 as a function of solution composition.
(b) Solute activity coefcient on a molality basis.
(c) Product of activity coefcient and molality as a function of composition. The
dashed line is the extrapolation of ideal-dilute behavior.
a Based
on data in Ref. [143], Appendix 8.6.
Various experimental methods are available for measuring a partial pressure in a gas
phase equilibrated with a liquid mixture. A correction for gas nonideality, such as that
given by Eq. 9.3.16, can be used to convert the partial pressure to fugacity.
If the solute of a solution is nonvolatile, we may pump out the air above the solution and
use a manometer to measure the pressure, which is the partial pressure of the solvent. Dynamic methods involve passing a stream of inert gas through a liquid mixture and analyzing
the gas mixture to evaluate the partial pressures of volatile components. For instance, we
could pass dry air successively through an aqueous solution and a desiccant and measure
the weight gained by the desiccant.
The isopiestic vapor pressure technique is one of the most useful methods for determining the fugacity of H2 O in a gas phase equilibrated with an aqueous solution. This is
a comparative method using a binary solution of the solute of interest, B, and a nonvolatile
reference solute of known properties. Some commonly used reference solutes for which
data are available are sucrose, NaCl, and CaCl2 .
In this method, solute B can be either a nonelectrolyte or electrolyte. Dishes, each
containing water and an accurately weighed sample of one of the solutes, are placed in wells
drilled in a block made of metal for good thermal equilibration. The assembly is placed in
a gas-tight chamber, the air is evacuated, and the apparatus is gently rocked in a thermostat
for a period of up to several days, or even weeks. During this period, H2 O is transferred
among the dishes through the vapor space until the chemical potential of the water becomes
the same in each solution. The solutions are then said to be isopiestic. Finally, the dishes
CHAPTER 9 MIXTURES
9.7 ACTIVITY OF AN U NCHARGED S PECIES
270
are removed from the apparatus and weighed to establish the molality of each solution. The
H2 O fugacity is known as a function of the molality of the reference solute, and is the same
as the H2 O fugacity in equilibrium with the solution of solute B at its measured molality.
The isopiestic vapor pressure method can also be used for nonaqueous solutions.
9.7 ACTIVITY OF AN UNCHARGED SPECIES

The activity ai of uncharged species i (i.e., a substance) is dened by the relation
def
i
i
ai D exp
(9.7.1)
RT
(uncharged species)
or
i
C RT ln ai
(9.7.2)
(uncharged species)
where is the standard chemical potential of the species.10 The activity of a species in a
i
given phase is a dimensionless quantity whose value depends on the choice of the standard
state and on the intensive properties of the phase: temperature, pressure, and composition.
The quantity ai is sometimes called the relative activity of i, because it depends on the
chemical potential relative to a standard chemical potential. An important application of the
activity concept is the denition of equilibrium constants (Sec. 11.8.1).
For convenience in later applications, we specify that the value of ai is the same in
phases that have the same temperature, pressure, and composition but are at different elevations in a gravitational eld, or are at different electric potentials. Section 9.8 10.1 will
describe a modication of the dening equation i D C RT ln ai for a system with
i
phases of different elevations, and Sec. 10.1 will describe the modication needed for a
charged species.
9.7.1 Standard states

The standard states of different kinds of mixture components have the same denitions as
those for reference states (Table 9.3), with the additional stipulation in each case that the
pressure is equal to the standard pressure p .
When component i is in its standard state, its chemical potential is the standard chemical
potential . It is important to note from Eq. 9.7.2 that when i equals , the logarithm
i
i
of ai is zero and the activity in the standard state is therefore unity.
The following equations in the form of Eq. 9.7.2 show the notation used in this book
for the standard chemical potentials and activities of various kinds of uncharged mixture
chemists dene the activity by i D ref C RT ln ai . The activity dened this way is not the same as
i
the activity used in this book unless the phase is at the standard pressure.
10 Some
CHAPTER 9 MIXTURES
271
BIOGRAPHICAL SKETCH
Gilbert Newton Lewis (18751946)

EDGAR FAHS SMITH COLLECTION
UNIVERSITY OF PENNSYLVANIA LIBRARY
formation of substances and ions.

In 1912 he became the chair of the chemistry department at the University of California at Berkeley, which he turned into a
renowned center of chemical research and
teaching. In 1916 he published his theory of
the shared electron-pair chemical bond (Lewis
structures), a concept he had been thinking
about since at least 1902. In the course of measuring the thermodynamic properties of electrolyte solutions, he introduced the concept of
Gilbert Lewis made major contributions to ionic strength (1921).
several elds of physical chemistry. He was
In 1923, at age 48, he consolidated his
born in Weymouth, Massachusetts. His father knowledge of thermodynamics in the great
was a lawyer and banker.
classic Thermodynamics and the Free Energy
Lewis was reserved, even shy in front of a of Chemical Substancesb with Merle Randall
large audience. He was also ambitious, had as coauthor. After that his interests changed
great personal charm, excelled at both exper- to other subjects. He was the rst to prepare
imental and theoretical thermodynamics, and pure deuterium and D2 O (1933), he formuwas a chain smoker of vile Manila cigars.
lated his generalized denitions of acids and
Lewis considered himself to be a disciple bases (Lewis acids and bases, 1938), and at
of Willard Gibbs. After completing his Ph.D the time of his death he was doing research on
dissertation at Harvard University in 1899, he photochemical processes.
published several papers of thermodynamic
Lewis was nominated 35 times for the Notheory that introduced for the rst time the bel prize, but was never awarded it. Accordterms fugacity (1901) and activity (1907). The ing to a history of modern chemistry published
rst of these papers was entitled A New Con- in 2008,c Wilhelm Palmaer, a Swedish electroception of Thermal Pressure and a Theory of chemist, used his position on the Nobel ComSolutions and began:a
mittee for Chemistry to block the award to
For an understanding of all kinds of physico- Lewis. Palmaer was a close friend of Walther
chemical equilibrium a further insight is neces- Nernst, whom Lewis had criticized on the basary into the nature of the conditions which exist sis of occasional arithmetic and thermodyin the interior of any homogeneous phase. It will namic inaccuracy.d
be the aim of the present paper to study this probHis career was summarized by his last gradlem in the light of a new theory, which, although uate student, Michael Kasha, as follows:e
opposed to some ideas which are now accepted
as correct, yet recommends itself by its simplicity and by its ability to explain several important
phenomena which have hitherto received no satisfactory explanation.
His rst faculty position (1905-1912) was at

Boston Tech, now the Massachusetts Institute
of Technology, where he continued work in
one of his dissertation subjects: the measurement of standard electrode potentials in order
to determine standard molar Gibbs energies of
Gilbert Lewis once dened physical chemistry

as encompassing everything that is interesting.
His own career touched virtually every aspect of
science, and in each he left his mark. He is justly
regarded as one of the key scientists in American history. It would be a great omission not to
record the warmth and intellectual curiosity radiated by Lewis personality. He epitomized the
scientist of unlimited imagination, and the joy of
working with him was to experience the life of
the mind unhindered by pedestrian concerns.
a Ref.
[103], page 6.
[102].
b Ref.
[103].
c Ref.
[35], Chap. 7.
d Ref.
e Ref.
[85].
CHAPTER 9 MIXTURES
272
components:
Substance i in a liquid or solid mixture
D
D
D
i (g) C RT ln ai (g)
i C RT ln ai
A C RT ln aA
x;B C RT ln ax;B
c;B C RT ln ac;B
m;B C RT ln am;B
(9.7.3)
(9.7.4)
(9.7.5)
(9.7.6)
(9.7.7)
(9.7.8)
9.7.2 Activities and composition

We need to be able to relate the activity of component i to the mixture composition. We
can do this by nding the relation between the chemical potential of component i in its
reference state and in its standard state, both at the same temperature. These two chemical
potentials, ref and , are equal only if the mixture is at the standard pressure p .
i
i
It will be useful to dene the following dimensionless quantity:
!
def
D exp
ref
i
(9.7.9)
RT
The symbol i for this quantity was introduced by Pitzer and Brewer.11 They called it
the activity in a reference state. To see why, compare the denition of activity given by
ref
i D i C RT ln ai with a rearrangement of Eq. 9.7.9: i D i C RT ln i .

ref
At a given temperature, the difference i

i depends only on the pressure p of the
. Thus
mixture, and is zero when p is equal to p
i is a function of p with a value of 1
when p is equal to p . This book will call i the pressure factor of species i .
To understand how activity is related to composition, let us take as an example the
activity am;B of solute B based on molality. From Eqs. 9.5.18 and 9.7.8, we have
m
D ref C RT ln m;B B
B
m;B
m
D m;B C RT ln am;B
(9.7.10)
The activity is then given by
m;B
ref
m;B
ln am;B D
RT
D ln
m;B
C ln
mB
m
mB
m;B
m
C ln
m;B
(9.7.11)
mB
(9.7.12)
m
The activity of a constituent of a condensed-phase mixture is in general equal to the product
of the pressure factor, the activity coefcient, and the composition variable divided by the
standard composition.
Table 9.5 on the next page gives explicit expressions for the activities of various kinds
of nonelectrolyte substances.
am;B D
11 Ref.
m;B m;B
[104], p. 249.
CHAPTER 9 MIXTURES
273
Table 9.5 Expressions for activities of nonelectrolytes. For a constituent of a condensed-phase mixture, fi , fA , and fB refer to the
fugacity in a gas phase equilibrated with the condensed phase.
Substance
Activity
Pure gas
a(g) D
Pure liquid or solid
aD
ai (g) D
i (g)
Substance i in a liquid or
solid mixture
ai D
aA D
Solute B, mole fraction

basis
ax;B D
x;B
x;B
Solute B, concentration
basis
ac;B D
c;B
cB
c;B
c
am;B D
(g)
pi
f
D i
p
p
xi D
A
f
p
xA D
m;B
fi
fi
A
fA
fA
xB D
m;B
x;B
c;B
mB
D
m
fB
kH,B
fB
kc;B c
m;B
fB
km;B m
9.7.3 Pressure factors and pressure

At a given temperature, the pressure factor i of component i of a mixture is a function
only of pressure. To derive the pressure dependence of i for various kinds of mixture
components, we need expressions for . ref

i
i / as functions of pressure to substitute in
ref
the dening equation i D exp . i

i /=RT .
For component i of a gas mixture, the reference state is pure gas i at the pressure of
the mixture, behaving as an ideal gas. The chemical potential of a pure ideal gas depends
D (g) C RT ln .p=p /. Thus the chemical
on its pressure according to Eq. 7.8.6:
potential of the reference state of gas component i is ref (g) D (g) C RT ln .p=p /,
i
i
and ref (g) (g) is equal to RT ln .p=p /. This gives us the following expression for the
i
i
pressure dependence of the pressure factor:
i (g)
p
p
(9.7.13)
For a mixture in a condensed phase, we will make use of .@ i =@p/T;fnig D Vi (Eq.

9.2.49). The relation between changes of i and p at constant temperature and composition is therefore d i D Vi dp. Recall (Sec. 9.1.5) that constant composition means that
the mole fraction or molality of each component, but not necessarily the concentration, is
constant.
Consider a process in which the system initially consists of a phase with component i
in its standard state. We change the pressure isothermally from p to the pressure p 0 of
the mixture of interest. For a pure-liquid, pure-solid, or solvent reference state, or a solute
CHAPTER 9 MIXTURES
274
reference state based on mole fraction or molality, this process brings the system to the
reference state of component i at pressure p 0 . The change of i in this case is given by
integration of d i D Vi dp:
ref 0
i .p /
p0
Vi dp
(9.7.14)
The appropriate partial molar volume Vi is the molar volume Vi or VA of the pure sub1
stance, or the partial molar volume VB of solute B at innite dilution.
Suppose we want to use a reference state for solute B based on concentration. Because
the isothermal pressure change involves a small change of volume, cB changes slightly
during the process, so that the right side of Eq. 9.7.14 is not quite the correct expression for
ref
0
c;B .p /
c;B .
We can derive a rigorous expression for ref .p 0 /

c;B
c;B as follows. Consider an idealdilute solution of solute B at an arbitrary pressure p, with solute chemical potential
given by B D ref C RT ln.cB =c / (Table 9.2). From this equation we obtain
c;B
@ B
@p
T;fnig
ref
c;B
@p
C RT
@ ln.cB =c /
@p
(9.7.15)
T;fnig
The partial derivative .@ B =@p/T;fnig is equal to the partial molar volume VB (Eq.
1
9.2.49), which in the ideal-dilute solution has its innite-dilution value VB . We
rewrite the second partial derivative on the right side of Eq. 9.7.15 as follows:
1 @cB
1
@ ln.cB =c /
@.nB =V /
D
D
@p
cB @p T;fnig
nB =V
@p
T;fnig
T;fnig
@.1=V /
1 @V
DV
D
@p
V @p T;fnig
T;fnig
D T
(9.7.16)
Here T is the isothermal compressibility of the solution, which at innite dilution is

1
T , the isothermal compressibility of the pure solvent. Equation 9.7.15 becomes
!
@ ref
c;B
1
1
VB D
C RT T
(9.7.17)
@p
T
Solving for d
ref
c;B
at constant T , and integrating from p to p 0 , we obtain nally

ref
0
c;B .p /
c;B
p0
1
VB
1
RT T dp
(9.7.18)
We are now able to write explicit formulas for i for each kind of mixture component.
They are collected in Table 9.6 on the next page.
Considering a constituent of a condensed-phase mixture, by how much is the pressure
factor likely to differ from unity? If we use the values p D 1 bar and T D 300 K, and
assume the molar volume of pure i is Vi D 100 cm3 mol 1 at all pressures, we nd that
i is 0:996 in the limit of zero pressure, unity at 1 bar, 1:004 at 2 bar, 1:04 at 10 bar, and
CHAPTER 9 MIXTURES
9.8 M IXTURES IN G RAVITATIONAL AND C ENTRIFUGAL F IELDS
275
Table 9.6 Expressions for the dependence of pressure factors of nonelectrolytes on pressure. The
approximate expressions assume the phase is incompressible, or the solute partial molar volume is
independent of pressure.
Substance
Substance i in a gas mixture, or the
pure gas
mixture, or the pure liquid or solid
Solute B, mole fraction or molality
basis
Pressure factor at pressure p 0

p0
p
Z
D exp
i (g)
p0
Vi .p 0 p /
RT
p
!
Z p0
VA
V .p 0 p /
dp
exp A
A D exp
RT
p RT
1 0
Z p0 1 !
VB
V .p p /
dp
exp B
x;B D m;B D exp
RT
p RT
#
"Z 0
1 0
p
1
VB
V .p p /
1
T dp
exp B
D exp
c;B
RT
RT
p
i
Vi
dp
RT
exp
1
1:49 at 100 bar. For a solution with VB D 100 cm3 mol 1 , we obtain the same values as
these for x;B , m;B , and c;B . These values demonstrate that it is only at high pressures
that the pressure factor differs appreciably from unity. For this reason, it is common to see
expressions for activity in which this factor is omitted: ai D i xi , am;B D m;B mB =m ,
and so on.
In principle, we can specify any convenient value for the standard pressure p . For a
chemist making measurements at high pressures, it would be convenient to specify a
value of p within the range of the experimental pressures, for example p D 1 kbar,
in order that the value of each pressure factor be close to unity.
9.8 MIXTURES IN GRAVITATIONAL AND CENTRIFUGAL FIELDS

A tall column of a gas mixture in a gravitational eld, and a liquid solution in the cell
of a spinning centrifuge rotor, are systems with equilibrium states that are nonuniform in
pressure and composition. This section derives the ways in which pressure and composition
vary spatially within these kinds of systems at equilibrium.
9.8.1 Gas mixture in a gravitational eld

Consider a tall column of a gas mixture in an earth-xed lab frame. Our treatment will
parallel that for a tall column of a pure gas in Sec. 8.1.4. We imagine the gas to be divided
into many thin slab-shaped phases at different elevations in a rigid container, as in Fig. 8.1
on page 196. We want to nd the equilibrium conditions reached spontaneously when the
system is isolated from its surroundings.
CHAPTER 9 MIXTURES
276
The derivation is the same as that in Sec. 9.2.7, with the additional constraint that for
each phase , dV is zero in order that each phase stays at a constant elevation. The result
is the relation
X T 0 T
X X 0
i
i
dS D
dS C
dn
(9.8.1)
0
0
i
T
T
0
0
i
In an equilibrium state, S is at a maximum and dS is zero for an innitesimal change of

any of the independent variables. This requires the coefcient of each term in the sums on
the right side of Eq. 9.8.1 to be zero. The equation therefore tells that at equilibrium the
temperature and the chemical potential of each constituent are uniform throughout the gas
mixture. The equation says nothing about the pressure.
Just as the chemical potential of a pure substance at a given elevation is dened in
this book as the molar Gibbs energy at that elevation (page 197), the chemical potential of
substance i in a mixture at elevation h is the partial molar Gibbs energy at that elevation.
We dene the standard potential (g) of component i of the gas mixture as the chemi
ical potential of i under standard state conditions at the reference elevation hD0. At this
elevation, the chemical potential and fugacity are related by
i .0/
i (g)
C RT ln
fi .0/
p
(9.8.2)
If we reversibly raise a small sample of mass m of the gas mixture by an innitesimal

distance dh, without heat and at constant T and V , the fugacity fi remains constant. The
gravitational work during the elevation process isP 0 D mg dh. This work contributes to
w
the internal energy change: dU D T dS p dV C i i dni Cmg dh. The total differential
of the Gibbs energy of the sample is
dG D d.U
D
T S C pV /
S dT C V dp C
X
i
dni C mg dh
(9.8.3)
From this total differential, we write the reciprocity relation
@ i
@mg
D
@h T;p;fnig
@ni T;p;nj i ;h
(9.8.4)
P
With the substitution m D i ni Mi in the partial derivative on the right side, the partial
derivative becomes Mi g. At constant T , p, and composition, therefore, we have d i D
Mi g dh. Integrating over a nite elevation change from h D 0 to h D h0 , we obtain
0
i .h /
The general relation between

i .h/
i .0/ D
i,
h0
0
Mi g dh D Mi gh0
(9.8.5)
0
(fi .h /Dfi .0/ )
fi , and h that agrees with Eqs. 9.8.2 and 9.8.5 is
i (g)
C RT ln
fi .h/
C Mi gh
p
(9.8.6)
CHAPTER 9 MIXTURES
277
In the equilibrium state of the tall column of gas, i .h/ is equal to i .0/. Equation 9.8.6
shows that this is only possible if fi decreases as h increases. Equating the expressions
given by this equation for i .h/ and i .0/, we have
i (g)
C RT ln
fi .h/
C Mi gh D
p
i (g)
C RT ln
fi .0/
p
(9.8.7)
Solving for fi .h/ gives

fi .h/ D fi .0/e
Mi gh=RT
(9.8.8)
(gas mixture at equilibrium)
If the gas is an ideal gas mixture, fi is the same as the partial pressure pi :
pi .h/ D pi .0/e
Mi gh=RT
(9.8.9)
(ideal gas mixture at equilibrium)
Equation 9.8.9 shows that each constituent of an ideal gas mixture individually obeys the
barometric formula given by Eq. 8.1.13 on page 198. P
The pressure at elevation h is found from p.h/ D i pi .h/. If the constituents have
different molar masses, the mole fraction composition changes with elevation. For example,
in a binary ideal gas mixture the mole fraction of the constituent with the greater molar mass
decreases with increasing elevation, and the mole fraction of the other constituent increases.
9.8.2 Liquid solution in a centrifuge cell

This section derives equilibrium conditions of a dilute binary solution conned to a cell
embedded in a spinning centrifuge rotor.
The system is the solution. The rotors angle of rotation with respect to a lab frame
is not relevant to the state of the system, so we use a local reference frame xed in the
rotor as shown in Fig. 9.12(a) on the next page. The values of heat, work, and energy
changes measured in this rotating frame are different from those in a lab frame (Sec. G.9
in Appendix G). Nevertheless, the laws of thermodynamics and the relations derived from
them are obeyed in the local frame when we measure the heat, work, and state functions in
this frame (page 499).
Note that an equilibrium state can only exist relative to the rotating local frame; an
observer xed in this frame would see no change in the state of the isolated solution over
time. While the rotor rotates, however, there is no equilibrium state relative to the lab frame,
because the systems position in the frame constantly changes.
We assume the centrifuge rotor rotates about the vertical z axis at a constant angular
velocity !. As shown in Fig. 9.12(a), the elevation of a point within the local frame is given
by z and the radial distance from the axis of rotation is given by r.
In the rotating local frame, a body of mass m has exerted on it a centrifugal force
F centr D m! 2 r directed horizontally in the outward Cr radial direction (Sec. G.9).12 The
12 There
is also a Coriolis force that vanishes as the bodys velocity in the rotating local frame approaches zero.
The centrifugal and Coriolis forces are apparent or ctitious forces, in the sense that they are caused by the
acceleration of the rotating frame rather than by interactions between particles. When we treat these forces as if
they are real forces, we can use Newtons second law of motion to relate the net force on a body and the bodys
acceleration in the rotating frame (see Sec. G.6).
CHAPTER 9 MIXTURES
278
(a)
(b)
Figure 9.12 (a) Sample cell of a centrifuge rotor (schematic), with Cartesian axes x,
y, z of a stationary lab frame and axes x 0 , y 0 , z of a local frame xed in the spinning
rotor. (The rotor is not shown.) The axis of rotation is along the z axis. The angular
velocity of the rotor is ! D d#= dt. The sample cell (heavy lines) is stationary in the
local frame.
(b) Thin slab-shaped volume elements in the sample cell.
gravitational force in this frame, directed in the downward z direction, is the same as the
gravitational force in a lab frame. Because the height of a typical centrifuge cell is usually
no greater than about one centimeter, in an equilibrium state the variation of pressure and
composition between the top and bottom of the cell at any given distance from the axis of
rotation is completely negligibleall the measurable variation is along the radial direction.
To nd conditions for equilibrium, we imagine the solution to be divided into many thin
curved volume elements at different distances from the axis of rotation as depicted in Fig.
9.12(b). We treat each volume element as a uniform phase held at constant volume so that
it is at a constant distance from the axis of rotation. The derivation is the same as the one
used in the preceding section to obtain Eq. 9.8.1, and leads to the same conclusion: in an
equilibrium state the temperature and the chemical potential of each substance (solvent and
solute) are uniform throughout the solution.
We nd the dependence of pressure on r as follows. Consider one of the thin slabshaped volume elements of Fig. 9.12(b). The volume element is located at radial position
r and its faces are perpendicular to the direction of increasing r. The thickness of the
volume element is r, the surface area of each face is As , and the mass of the solution in
the volume element is m D As r. Expressed as components in the direction of increasing
r of the forces exerted on the volume element, the force at the inner face is pAs , the force
at the outer face is .p C p/As , and the centrifugal force is m! 2 r D As ! 2 rr. From
Newtons second law, the sum of these components is zero at equilibrium:
pAs
.p C p/As C As ! 2 rr D 0
(9.8.10)
or p D ! 2 rr. In the limit as r and p are made innitesimal, this becomes

dp D ! 2 r dr
(9.8.11)
CHAPTER 9 MIXTURES
is uniform throughout the solution.13 Then integration of Eq.
We will assume the density

9.8.11 yields
p
00
p D
p 00
p0
279
dp D !
! 2 h 00
r
2
r 00
r0
r dr D
r0
(9.8.12)
where the superscripts 0 and 00 denote positions at two different values of r in the cell. The
pressure is seen to increase with increasing distance from the axis of rotation.
Next we investigate the dependence of the solute concentration cB on r in the equilibrium state of the binary solution. Consider a small sample of the solution of mass m.
Assume the extent of this sample in the radial direction is small enough for the variation of
the centrifugal force eld to be negligible. The reversible work in the local frame needed
to move this small sample an innitesimal distance dr at constant z, T , and p, using an
external force F centr that opposes the centrifugal force, is
w 0 D F sur dr D . F centr / dr D
m! 2 r dr
(9.8.13)
This work is a contribution to the change dU of the internal energy. The Gibbs energy of
the small sample in the local frame is a function of the independent variables T , p, nA , nB ,
and r, and its total differential is
dG D d.U
D
T S C pV /
S dT C V dp C
A dnA
B dnB
m! 2 r dr
We use Eq. 9.8.14 to write the reciprocity relation
@ B
@m
2
D ! r
@r T;p;nA ;nB
@nB T;p;nA ;r
Then, using m D nA MA C nB MB , we obtain
@ B
D
@r T;p;nA ;nB
(9.8.14)
(9.8.15)
MB ! 2 r
(9.8.16)
Thus at constant T , p, and composition, which are the conditions that allow the activity
ac;B to remain constant, B for the sample varies with r according to d B D MB ! 2 r dr.
We integrate from radial position r 0 to position r 00 to obtain
00
B .r /
B .r / D
MB !
2
1
2 MB !
r 00
r dr
r0
r 00
r0
(9.8.17)
(ac;B .r 00 /Dac;B .r 0 / )
Let us take r 0 as a reference position, such as the end of the centrifuge cell farthest from
the axis of rotation. We dene the standard chemical potential as the solute chemical
c;B
13 In
the centrifugal eld, this assumption is strictly true only if the solution is incompressible and its density is
independent of composition.
CHAPTER 9 MIXTURES
280
potential under standard state conditions on a concentration basis at this position. The solute
chemical potential and activity at this position are related by
B .r
/D
c;B
C RT ln ac;B .r 0 /
(9.8.18)
From Eqs. 9.8.17 and 9.8.18, we obtain the following general relation between
at an arbitrary radial position r 00 :
i
h
2
2
.r 00 / D C RT ln ac;B .r 00 / 1 MB ! 2 r 00
r0
B
c;B
2
and ac;B
(9.8.19)
We found earlier that when the solution is in an equilibrium state, B is independent

of rthat is, B .r 00 / is equal to B .r 0 / for any value of r 00 . When we equate expressions
given by Eq. 9.8.19 for B .r 00 / and B .r 0 / and rearrange, we obtain the following relation
between the activities at the two radial positions:
ln
ac;B .r 00 /
M ! 2 h 00
D B
r
ac;B .r 0 /
2RT
r0
(9.8.20)
(solution in centrifuge
cell at equilibrium)
The solute activity is related to the concentration cB by ac;B D c;B c;B cB =c . We

assume the solution is sufciently dilute for the activity coefcient c;B to be approximated
1
by 1. The pressure factor is given by c;B
exp VB .p p /=RT (Table 9.6). These
relations give us another expression for the logarithm of the ratio of activities:
ln
ac;B .r 00 /
V 1 .p 00 p 0 /
c .r 00 /
D B
C ln B 0
ac;B .r 0 /
RT
cB .r /
(9.8.21)
1
We substitute for p 00 p 0 from Eq. 9.8.12. It is also useful to make the substitution VB D
1
1
MB vB , where vB is the partial specic volume of the solute at innite dilution (page 235).
When we equate the two expressions for lnac;B .r 00 /=ac;B .r 0 /, we obtain nally
ln
1
MB 1 vB
cB .r 00 /
D
cB .r 0 /
2RT
!2 h
r 00
r0
(9.8.22)
(solution in centrifuge
cell at equilibrium)
This equation shows that if the solution density is less than the effective solute density
1
1
1=vB , so that vB is less than 1, the solute concentration increases with increasing distance
1
from the axis of rotation in the equilibrium state. If, however, is greater than 1=vB , the
1
concentration decreases with increasing r. The factor 1 vB
is like a buoyancy factor
for the effect of the centrifugal eld on the solute.
Equation 9.8.22 is needed for sedimentation equilibrium, a method of determining the
molar mass of a macromolecule. A dilute solution of the macromolecule is placed in the cell
of an analytical ultracentrifuge, and the angular velocity is selected to produce a measurable
solute concentration gradient at equilibrium. The solute concentration is measured optically
as a function of r. The equation predicts that a plot of ln .cB =c / versus r 2 will be linear,
1
1
with a slope equal to MB 1 vB
! 2 =2RT . The partial specic volume vB is found
from measurements of solution density as a function of solute mass fraction (page 235). By
this means, the molar mass MB of the macromolecule is evaluated.
CHAPTER 9 MIXTURES
P ROBLEMS
281
PROBLEMS
in Appendix I.
9.1 For a binary solution, nd expressions for the mole fractions xB and xA as functions of the
solute molality mB .
9.2 Consider a binary mixture of two liquids, A and B. The molar volume of mixing, V (mix)=n,
is given by Eq. 9.2.19.
(a) Find a formula for calculating the value of V (mix)=n of a binary mixture from values
of xA , xB , MA , MB , , A , and B .
Table 9.7 Molar volumes of mixing of binary mixtures of 1-hexanol (A)
and 1-octene (B) at 25 C. a
xB
V (mix)=n=cm3 mol
0
0:049
0:097
0:146
0:199
0:235
0:284
0:343
0:388
0:448
0:491
a Ref.
0
0:027
0:050
0:063
0:077
0:073
0:074
0:065
0:053
0:032
0:016
V (mix)=n=cm3 mol
xB
0:555
0:597
0:702
0:716
0:751
0:803
0:846
0:897
0:944
1
0:005
0:011
0:029
0:035
0:048
0:056
0:058
0:057
0:049
0
[156].
(b) The molar volumes of mixing for liquid binary mixtures of 1-hexanol (A) and 1-octene
(B) at 25 C have been calculated from their measured densities. The data are in Table
9.7. The molar volumes of the pure constituents are VA D 125:31 cm3 mol 1 and VB D
157:85 cm3 mol 1 . Use the method of intercepts to estimate the partial molar volumes of
both constituents in an equimolar mixture (xA D xB D 0:5), and the partial molar volume
1
VB of B at innite dilution.
9.3 Extend the derivation of Prob. 8.1, concerning a liquid droplet of radius r suspended in a gas,
to the case in which the liquid and gas are both mixtures. Show that the equilibrium conditions
g
are T g D T l , i D li (for each species i that can equilibrate between the two phases), and
l
g
p D p C2 =r, where is the surface tension. (As in Prob. 8.1, the last relation is the Laplace
equation.)
9.4 Consider a gaseous mixture of 4:0000 10 2 mol of N2 (A) and 4:0000 10 2 mol of CO2
(B) in a volume of 1:0000 10 3 m3 at a temperature of 298:15 K. The second virial coefcients at this temperature have the values14
BAA D
BBB D
BAB D
14 Refs.
4:8 10 6 m3 mol 1
124:5 10 6 m3 mol
47:5
10
m3 mol
[3], [47], and [48].
CHAPTER 9 MIXTURES
P ROBLEMS
282
Compare the pressure of the real gas mixture with that predicted by the ideal gas equation. See
Eqs. 9.3.20 and 9.3.23.
9.5 At 25 C and 1 bar, the Henrys law constants of nitrogen and oxygen dissolved in water are
kH,N2 D 8:64 104 bar and kH,O2 D 4:41 104 bar.15 The vapor pressure of water at this
temperature and pressure is pH2 O D 0:032 bar. Assume that dry air contains only N2 and
O2 at mole fractions yN2 D 0:788 and yO2 D 0:212. Consider liquidgas systems formed by
equilibrating liquid water and air at 25 C and 1:000 bar, and assume that the gas phase behaves
as an ideal gas mixture.
Hint: The sum of the partial pressures of N2 and O2 must be .1:000 0:032/ bar D 0:968 bar.
If the volume of one of the phases is much larger than that of the other, then almost all of the
N2 and O2 will be in the predominant phase and the ratio of their amounts in this phase must
be practically the same as in dry air.
Determine the mole fractions of N2 and O2 in both phases in the following limiting cases:
(a) A large volume of air is equilibrated with just enough water to leave a small drop of liquid.
(b) A large volume of water is equilibrated with just enough air to leave a small bubble of
gas.
9.6 Derive the expression for
m;B
given in Table 9.4, starting with Eq. 9.5.18.
9.7 Consider a nonideal binary gas mixture with the simple equation of state V D nRT =p C nB
(Eq. 9.3.21).
(a) The rule of Lewis and Randall states that the value of the mixed second virial coefcient
BAB is the average of BAA and BBB . Show that when this rule holds, the fugacity coefcient of A in a binary gas mixture of any composition is given by ln A D BAA p=RT .
By comparing this expression with Eq. 7.8.18 for a pure gas, express the fugacity of A in
the mixture as a function of the fugacity of pure A at the same temperature and pressure
as the mixture.
(b) The rule of Lewis and Randall is not accurately obeyed when constituents A and B are
chemically dissimilar. For example, at 298:15 K, the second virial coefcients of H2 O
(A) and N2 (B) are BAA D 1158 cm3 mol 1 and BBB D 5 cm3 mol 1 , respectively,
whereas the mixed second virial coefcient is BAB D 40 cm3 mol 1 .
When liquid water is equilibrated with nitrogen at 298:15 K and 1 bar, the partial pressure
of H2 O in the gas phase is pA D 0:03185 bar. Use the given values of BAA , BBB , and
BAB to calculate the fugacity of the gaseous H2 O in this binary mixture. Compare this
fugacity with the fugacity calculated with the value of BAB predicted by the rule of Lewis
and Randall.
9.8 Benzene and 1-octanol are two liquids that mix in all proportions. Benzene has a measurable
vapor pressure, whereas 1-octanol is practically nonvolatile. The data in Table 9.8 on the next
page were obtained by Platford16 using the isopiestic vapor pressure method.
(a) Use numerical integration to evaluate the integral on the right side of Eq. 9.6.10 at each
of the values of xA listed in the table, and thus nd B at these compositions.
(b) Draw two curves on the same graph showing the effective mole fractions A xA and B xB
as functions of xA . Are the deviations from ideal-mixture behavior positive or negative?
15 Ref.
[170].
16 Ref.
[133].
CHAPTER 9 MIXTURES
P ROBLEMS
283
Table 9.8 Activity coefcient of benzene (A) in mixtures of benzene and
1-octanol at 20 C. The reference state
is the pure liquid.
xA
0
0:1334
0:2381
0:4131
0:5805
2:0
1:915
1:809
1:594
1:370
xA
0:7631
0:8474
0:9174
0:9782
1:183
1:101
1:046
1:005
a extrapolated
Table 9.9 Liquid and gas compositions in the two-phase system of methanol (A) and benzene (B) at 45 C a
xA
yA
p=kPa
xA
yA
p=kPa
0
0:0207
0:0314
0:0431
0:0613
0:0854
0:1811
0:3217
0
0:2794
0:3391
0:3794
0:4306
0:4642
0:5171
0:5450
29:894
40:962
44:231
46:832
50:488
53:224
57:454
59:402
0:4201
0:5420
0:6164
0:7259
0:8171
0:9033
0:9497
1
0:5590
0:5783
0:5908
0:6216
0:6681
0:7525
0:8368
1
60:015
60:416
60:416
59:868
58:321
54:692
51:009
44:608
a Ref.
[155].
9.9 Table 9.9 lists measured values of gas-phase composition and total pressure for the binary
two-phase methanolbenzene system at constant temperature and varied liquid-phase composition. xA is the mole fraction of methanol in the liquid mixture, and yA is the mole fraction of
methanol in the equilibrated gas phase.
(a) For each of the 16 different liquid-phase compositions, tabulate the partial pressures of A
and B in the equilibrated gas phase.
(b) Plot pA and pB versus xA on the same graph. Notice that the behavior of the mixture is far
from that of an ideal mixture. Are the deviations from Raoults law positive or negative?
(c) Tabulate and plot the activity coefcient B of the benzene as a function of xA using a
pure-liquid reference state. Assume that the fugacity fB is equal to pB , and ignore the
effects of variable pressure.
(d) Estimate the Henrys law constant kH,A of methanol in the benzene environment at 45 C
by the graphical method suggested in Fig. 9.7(b). Again assume that fA and pA are equal,
and ignore the effects of variable pressure.
9.10 Consider a dilute binary nonelectrolyte solution in which the dependence of the chemical potential of solute B on composition is given by
B
ref
m;B
C RT ln
mB
C km mB
m
CHAPTER 9 MIXTURES
P ROBLEMS
284
where ref and km are constants at a given T and p. (The derivation of this equation is
m;B
sketched in Sec. 9.5.4.) Use the GibbsDuhem equation in the form d A D .nB =nA / d B to
obtain an expression for A
A as a function of mB in this solution.
9.11 By means of the isopiestic vapor pressure technique, the osmotic coefcients of aqueous solutions of urea at 25 C have been measured at molalities up to the saturation limit of about
20 mol kg 1 .17 The experimental values are closely approximated by the function
m
D 1:00
0:050 mB =m
1:00 C 0:179 mB =m
where m is 1 mol kg 1 . Calculate values of the solvent and solute activity coefcients A
and m;B at various molalities in the range 020 mol kg 1 , and plot them versus mB =m . Use
enough points to be able to see the shapes of the curves. What are the limiting slopes of these
curves as mB approaches zero?
9.12 Use Eq. 9.2.49 to derive an expression for the rate at which the logarithm of the activity coefcient of component i of a liquid mixture changes with pressure at constant temperature and
composition: .@ ln i =@p/T;fnig D
9.13 Assume that at sea level the atmosphere has a pressure of 1:00 bar and a composition given by
yN2 D 0:788 and yO2 D 0:212. Find the partial pressures and mole fractions of N2 and O2 ,
and the total pressure, at an altitude of 10:0 km, making the (drastic) approximation that the
atmosphere is an ideal gas mixture in an equilibrium state at 0 C. For g use the value of the
standard acceleration of free fall listed in Appendix B.
9.14 Consider a tall column of a dilute binary liquid solution at equilibrium in a gravitational eld.
(a) Derive an expression for ln cB .h/=cB .0/ , where cB .h/ and cB .0/ are the solute concentrations at elevations h and 0. Your expression should be a function of h, MB , T , , and
1
the partial specic volume of the solute at innite dilution, vB . For the dependence of
pressure on elevation, you may use the hydrostatic formula dp D g dh (Eq. 8.1.14 on
page 198) and assume the solution density is the same at all elevations. Hint: use the
derivation leading to Eq. 9.8.22 as a guide.
(b) Suppose you have a tall vessel containing a dilute solution of a macromolecule solute of
1
molar mass MB D 10:0 kg mol 1 and partial specic volume vB D 0:78 cm3 g 1 . The
3
solution density is D 1:00 g cm and the temperature is T D 300 K. Find the height h
from the bottom of the vessel at which, in the equilibrium state, the concentration cB has
decreased to 99 percent of the concentration at the bottom.
9.15 FhuA is a protein found in the outer membrane of the Escherichia coli bacterium. From the
known amino acid sequence, its molar mass is calculated to be 78:804 kg mol 1 . In aqueous
solution, molecules of the detergent dodecyl maltoside bind to a FhuA molecule to form an
aggregate that behaves as a single solute species. Figure 9.13 on the next page shows data
collected in a sedimentation equilibrium experiment with a dilute solution of the aggregate.18
In the graph, A is the absorbance measured at a wavelength of 280 nm (a property that is a
linear function of the aggregate concentration) and r is the radial distance from the axis of
rotation of the centrifuge rotor. The experimental points fall very close to the straight line
shown in the graph. The sedimentation conditions were ! D 838 s 1 and T D 293 K. The
1
authors used the values vB D 0:776 cm3 g 1 and D 1:004 g cm 3 .
(a) The values of r at which the absorbance was measured range from 6:95 cm to 7:20 cm.
Find the difference of pressure in the solution between these two positions.
17 Ref.
[148].
18 Ref.
[18].
CHAPTER 9 MIXTURES
P ROBLEMS
285
0:5
0:0
ln A
0:5
1:0
1:5
2:0
2:5
48
49
50
r 2 /cm2
51
52
Figure 9.13 Sedimentation equilibrium of a dilute solution of the FhuA-dodecyl maltoside aggregate.
(b) Find the molar mass of the aggregate solute species, and use it to estimate the mass binding ratio (the mass of bound detergent divided by the mass of protein).
C HAPTER 10
E LECTROLYTE S OLUTIONS
HCl Pa
The thermodynamic properties of electrolyte solutions differ in signicant ways from the
properties of mixtures of nonelectrolytes.
Here is an example. Pure HCl (hydrogen chloride) is a gas that is very soluble in water.
A plot of the partial pressure of gaseous HCl in equilibrium with aqueous HCl, as a function
of the solution molality (Fig. 10.1), shows that the limiting slope at innite dilution is not
nite, but zero. What is the reason for this non-Henrys law behavior? It must be because
HCl is an electrolyteit dissociates (ionizes) in the aqueous environment.
It is customary to use a molality basis for the reference and standard states of electrolyte
solutes. This is the only basis used in this chapter, even when not explicitly indicated for
ions. The symbol , for instance, denotes the chemical potential of a cation in a standard
C
state based on molality.
In dealing with an electrolyte solute, we can refer to the solute (a substance) as a whole
and to the individual charged ions that result from dissociation. We can apply the same
general denitions of chemical potential, activity coefcient, and activity to these different
HCl mol kg
Figure 10.1 Partial pressure of HCl in a gas phase equilibrated with aqueous HCl at
25 C and 1 bar. Open circles: experimental data from Ref. [5].
286
CHAPTER 10 ELECTROLYTE SOLUTIONS

10.1 S INGLE - ION Q UANTITIES
287
species, but only the activity coefcient and activity of the solute as a whole can be evaluated
experimentally.
10.1 SINGLE-ION QUANTITIES

Consider a solution of an electrolyte solute that dissociates completely into a cation species
and an anion species. Subscripts C and will be used to denote the cation and anion,
respectively. The solute molality mB is dened as the amount of solute formula unit divided
by the mass of solvent.
We rst need to investigate the relation between the chemical potential of an ion species
and the electric potential of the solution phase.
The electric potential in the interior of a phase is called the inner electric potential, or
Galvani potential. It is dened as the work needed to reversibly move an innitesimal test
charge into the phase from a position innitely far from other charges, divided by the value
of the test charge. The electrical potential energy of a charge in the phase is the product of
and the charge.
Consider a hypothetical process in which an innitesimal amount dnC of the cation is
transferred into a solution phase at constant T and p. The quantity of charge transferred
is Q D zC F dnC , where zC is the charge number (C1, C2, etc.) of the cation, and F
is the Faraday constant.1 If the phase is at zero electric potential, the process causes no
change in its electrical potential energy. However, if the phase has a nite electric potential , the transfer process changes its electrical potential energy by Q D zC F dnC .
Consequently, the internal energy change depends on according to
dU. / D dU.0/ C zC F dnC
(10.1.1)
where the electric potential is indicated in parentheses. The change in the Gibbs energy of
the phase is given by dG D d.U T S C pV /, where T , S, p, and V are unaffected by the
value of . The dependence of dG on is therefore
dG. / D dG.0/ C zC F dnC
(10.1.2)
P
The Gibbs fundamental equation for an open system, dG D S dT CV dpC i i dni
(Eq. 9.2.34), assumes the electric potential is zero. From this equation and Eq. 10.1.2, the
Gibbs energy change during the transfer process at constant T and p is found to depend on
according to
dG. / D
dnC
(10.1.3)
C .0/ C zC F
The chemical potential of the cation in a phase of electric potential , dened by the partial
molar Gibbs energy @G. /=@nC T;p , is therefore given by
C.
/D
C .0/
C zC F
(10.1.4)
.0/ C z F
(10.1.5)
The corresponding relation for an anion is

. /D
1 The
Faraday constant (page 453) is the charge per amount of protons.

10.1 S INGLE - ION Q UANTITIES
288
where z is the charge number of the anion ( 1, 2, etc.). For a charged species in general,
we have
(10.1.6)
i . / D i .0/ C zi F
We dene the standard state of an ion on a molality basis in the same way as for a
nonelectrolyte solute, with the additional stipulation that the ion is in a phase of zero electric
potential. Thus, the standard state is a hypothetical state in which the ion is at molality m
with behavior extrapolated from innite dilution on a molality basis, in a phase of pressure
p D p and electric potential D0.
The standard chemical potential or of a cation or anion is the chemical potential
C
of the ion in its standard state. Single-ion activities aC and a in a phase of zero electric
potential are dened by relations having the form of Eq. 9.7.8:
C .0/
C RT ln aC
.0/ D
C RT ln a
(10.1.7)
As explained on page 270, aC and a should depend on the temperature, pressure, and
composition of the phase, and not on the value of .
From Eqs. 10.1.4, 10.1.5, and 10.1.7, the relations between the chemical potential of a
cation or anion, its activity, and the electric potential of its phase, are found to be
C
C RT ln aC C zC F
C RT ln a C zi F
(10.1.8)
These relations are denitions of single-ion activities in a phase of electric potential .

For a charged species in general, we can write2
i
C RT ln ai C zi F
(10.1.9)
Note that we can also apply this equation to an uncharged species, because the charge
number zi is then zero and Eq. 10.1.9 becomes the same as Eq. 9.7.2 on page 270.
Of course there is no experimental way to evaluate either C or
relative to a reference state or standard state, because it is impossible to add cations or anions by themselves
to a solution. We can nevertheless write some theoretical relations involving C and
.
For a given temperature and pressure, we can write the dependence of the chemical
potentials of the ions on their molalities in the same form as that given by Eq. 9.5.18 for a
nonelectrolyte solute:
m
m
D ref C RT ln C C
D ref C RT ln
(10.1.10)
C
C
m
m
Here ref and ref are the chemical potentials of the cation and anion in solute reference
C
states. Each reference state is dened as a hypothetical solution with the same temperature,
pressure, and electric potential as the solution under consideration; in this solution, the
molality of the ion has the standard value m , and the ion behaves according to Henrys law
based on molality. C and
are single-ion activity coefcients on a molality basis.
thermodynamicists call the quantity . C RT ln ai /, which depends only on T , p, and composition,
i
the chemical potential of ion i , and the quantity . C RT ln ai C zi F / the electrochemical potential with
i
symbol Q i .
2 Some

10.2 S OLUTION OF A S YMMETRICAL E LECTROLYTE
289
The single-ion activity coefcients approach unity in the limit of innite dilution:
C
! 1 and
! 1 as
mB ! 0
(10.1.11)
(constant T , p, and )
In other words, we assume that in an extremely dilute electrolyte solution each individual
ion behaves like a nonelectrolyte solute species in an ideal-dilute solution. At a nite solute
molality, the values of C and
are the ones that allow Eq. 10.1.10 to give the correct
ref
ref /. We have no way to actually measure
values of the quantities . C
/ and .
C
these quantities experimentally, so we cannot evaluate either C or .
We can dene single-ion pressure factors C and
as follows:
!
1
ref
V .p p /
def
C
C
D exp
(10.1.12)
exp C
C
RT
RT
def
D exp
ref
RT
exp
V 1 .p p /
RT
(10.1.13)
The approximations in these equations are like those in Table 9.6 for nonelectrolyte solutes;
they are based on the assumption that the partial molar volumes VC and V are independent
of pressure.
From Eqs. 10.1.7, 10.1.10, 10.1.12, and 10.1.13, the single-ion activities are related to
the solution composition by
aC D
C C
mC
m
m
m
a D
(10.1.14)
Then, from Eq. 10.1.9, we have the following relations between the chemical potentials and
molalities of the ions:
D
C RT ln.
C RT ln.
C C mC =m
/ C zC F
(10.1.15)
m =m / C z F
(10.1.16)
Like the values of C and , values of the single-ion quantities aC , a ,

cannot be determined by experiment.
C,
and
10.2 SOLUTION OF A SYMMETRICAL ELECTROLYTE

Let us consider properties of an electrolyte solute as a whole. The simplest case is that of a
binary solution in which the solute is a symmetrical strong electrolytea substance whose
formula unit has one cation and one anion that dissociate completely. This condition will be
indicated by D 2, where is the number of ions per formula unit. The solute with equal
to 2 might be a 1:1 salt such as NaCl, a 2:2 salt such as MgSO4 , or a strong monoprotic
acid such as HCl.
In this binary solution, the chemical potential of the solute as a whole is dened in the
usual way as the partial molar Gibbs energy
@G
def
(10.2.1)
B D
@nB T;p;nA

290
and is a function of T , p, and the solute molality mB . Although B under given conditions
must in principle have a denite value, we are not able to actually evaluate it because we
have no way to measure precisely the energy brought into the system by the solute. This
energy contributes to the internal energy and thus to G. We can, however, evaluate the
ref
differences B
m;B or B
m;B .
We can write the additivity rule (Eq. 9.2.25) for G as either
G D nA
C nB
(10.2.2)
Cn
(10.2.3)
or
G D nA
C nC
A comparison of these equations for a symmetrical electrolyte (nB D nC D n ) gives us

the relation
D
(10.2.4)
( D2)
We see that the solute chemical potential in this case is the sum of the single-ion chemical
potentials.
The solution is a phase of electric potential . From Eqs. 10.1.4 and 10.1.5, the sum
appearing in Eq. 10.2.4 is
CC
C.
/C
. /D
C .0/
.0/ C .zC C z /F
(10.2.5)
For the symmetrical electrolyte, the sum .zC C z / is zero, so that B is equal to C .0/ C
.0/. We substitute the expressions of Eq. 10.1.10, use the relation ref D ref C ref
m;B
C
with reference states at D0, set the ion molalities mC and m equal to mB , and obtain
mB 2
ref
D m;B C RT ln C
(10.2.6)
B
m
( D2)
The important feature of this relation is the appearance of the second power of mB =m ,
instead of the rst power as in the case of a nonelectrolyte. Also note that B does not
depend on , unlike C and
.
Although we cannot evaluate C or
individually, we can evaluate the product C .
This product is the square of the mean ionic activity coefcient , dened for a symmetrical electrolyte by
def
(10.2.7)
( D2)
With this denition, Eq. 10.2.6 becomes

B
ref
m;B
C RT ln .
Since it is possible to determine the value of

is a measurable quantity.
2
/
ref
m;B
mB 2
m
(10.2.8)
( D2)
for a solution of known molality,

291
If the electrolyte (e.g., HCl) is sufciently volatile, its mean ionic activity coefcient in
a solution can be evaluated from partial pressure measurements of an equilibrated gas
phase. Section 10.6 will describe a general method by which can be found from
osmotic coefcients. Section 14.5 describes how, in favorable cases, it is possible to
evaluate from the equilibrium cell potential of a galvanic cell.
The activity am;B of a solute substance on a molality basis is dened by Eq. 9.7.8 on
page 272:
(10.2.9)
B D m;B C RT ln am;B
Here is the chemical potential of the solute in its standard state, which is the solute
m;B
reference state at the standard pressure. By equating the expressions for B given by Eqs.
10.2.8 and 10.2.9 and solving for the activity, we obtain
mB 2
am;B D m;B . /2
(10.2.10)
m
( D2)
where
m;B
is the pressure factor dened by

def
m;B
D exp
ref
m;B
m;B
RT
We can use the appropriate expression in Table 9.6 on page 275 to evaluate
arbitrary pressure p 0 :
1 0
Z p0 1 !
VB
V .p p /
0
dp
exp B
m;B .p / D exp
RT
p RT
(10.2.11)
m;B
at an
(10.2.12)
The value of m;B is 1 at the standard pressure, and close to 1 at any reasonably low
pressure (page 275). For this reason it is common to see Eq. 10.2.10 written as am;B D
2
2
.mB =m / , with m;B omitted.
Equation 10.2.10 predicts that the activity of HCl in aqueous solutions is proportional,
in the limit of innite dilution, to the square of the HCl molality. In contrast, the activity
of a nonelectrolyte solute is proportional to the rst power of the molality in this limit.
This predicted behavior of aqueous HCl is consistent with the data plotted in Fig. 10.1
on page 286, and is conrmed by the data for dilute HCl solutions shown in Fig. 10.2(a).
The dashed line in Fig. 10.2(a) is the extrapolation of the ideal-dilute behavior given by
am;B D .mB =m /2 . The extension of this line to mB D m establishes the hypothetical
solute reference state based on molality, indicated by a lled circle in Fig. 10.2(b). (Since
the data are for solutions at the standard pressure of 1 bar, the solute reference state shown
in the gure is also the solute standard state.)
The solid curve of Fig. 10.2(c) shows how the mean ionic activity coefcient of HCl
varies with molality in approximately the same range of molalities as the data shown in
Fig. 10.2(b). In the limit of innite dilution, approaches unity. The slope of the curve
approaches 1 in this limit, quite unlike the behavior described in Sec. 9.5.4 for the activity
coefcient of a nonelectrolyte solute.
For a symmetrical strong electrolyte, is the geometric average of the single-ion
activity coefcients C and . We have no way of evaluating C or
individually, even
if we know the value of . For instance, we cannot assume that C and
are equal.

10.3 E LECTROLYTES IN G ENERAL
292
mol kg
B
mol kg
B
(a)
(b)
B mol kg
(c)
Figure 10.2 Aqueous HCl at 25 C and 1 bar. a

(a) HCl activity on a molality basis as a function of molality squared. The dashed line
is the extrapolation of the ideal-dilute behavior.
(b) Same as (a) at a greatly reduced scale; the lled circle indicates the solute reference
state.
(c) Mean ionic activity coefcient of HCl as a function of molality.
a Curves
based on experimental parameter values in Ref. [74], Table 11-5-1.
10.3 ELECTROLYTES IN GENERAL

The formula unit of a nonsymmetrical electrolyte solute has more than two ions. General
formulas for the solute as a whole are more complicated than those for the symmetrical case
treated in the preceding section, but are derived by the same reasoning.
Again we assume the solute dissociates completely into its constituent ions. We dene
the following symbols:
C D the number of cations per solute formula unit
D the number of anions per solute formula unit
D the sum C C
For example, if the solute formula is Al2 (SO4 )3 , the values are C D2, D3, and D5.
10.3.1 Solution of a single electrolyte

In a solution of a single electrolyte solute that is not necessarily symmetrical, the ion molalities are related to the overall solute molality by
mC D
C mB
m D
mB
(10.3.1)
nB
(10.3.2)
From the additivity rule for the Gibbs energy, we have

G D nA
D nA
C nB
AC
C nB
giving the relation

B
(10.3.3)

293
in place of Eq. 10.2.4. The cations and anions are in the same phase of electric potential .
We use Eqs. 10.1.4 and 10.1.5 to obtain
C
C.
/C
. /
C .0/
Electrical neutrality requires that .

B
C zC
.0/ C .
C zC
z /F
(10.3.4)
z / be zero, giving
C .0/
.0/
By combining Eq. 10.3.5 with Eqs. 10.1.10, 10.3.1, and 10.3.3, we obtain
h
mB i
C
C
ref
B D B C RT ln
C
C
m
(10.3.5)
(10.3.6)
ref is the chemical potential of the solute in the hypothetical

where ref D C ref C
B
C
reference state at D0 in which B is at the standard molality and behaves as at innite
dilution. Equation 10.3.6 is the generalization of Eq. 10.2.6. It shows that although C and
depend on , B does not.
The mean ionic activity coefcient is dened in general by
D CC
(10.3.7)
or
1=
(10.3.8)
Thus is a geometric average of C and

weighted by the numbers of the cations and
anions in the solute formula unit. With a substitution from Eq. 10.3.7, Eq. 10.3.6 becomes
h
mB i
C
D ref C RT ln
(10.3.9)
B
B
C
m
ref
Since B
B is a measurable quantity, so also is .
The solute activity, dened by B D C RT ln am;B , is
m;B
am;B D
m;B
mB
m
(10.3.10)
where m;B is the pressure factor that we can evaluate with Eq. 10.2.12. Equation 10.3.10 is
the generalization of Eq. 10.2.10. From Eqs. 10.1.12, 10.1.13, and 10.2.11 and the relations
ref
ref
ref and D
, we obtain the relation
C CC
B D C C C
B
m;B
(10.3.11)
10.3.2 Multisolute solution

Equation 10.3.3 relates the chemical potential of electrolyte B in a binary solution to the
single-ion chemical potentials of its constituent ions:
B
(10.3.12)

294
This relation is valid for each individual solute substance in a multisolute solution, even
when two or more of the electrolyte solutes have an ion species in common.
As an illustration of this principle, consider a solution prepared by dissolving amounts
nB of BaI2 and nC of CsI in an amount nA of H2 O. Assume the dissolved salts are completely dissociated into ions, with the I ion common to both. The additivity rule for the
Gibbs energy of this solution can be written in the form
G D nA
C nB
C nC
(10.3.13)
and also, using single-ion quantities, in the form

G D nA
C nB .Ba2C / C 2nB .I / C nC .CsC / C nC .I /
(10.3.14)
Comparing Eqs. 10.3.13 and 10.3.14, we nd the following relations must exist between
the chemical potentials of the solute substances and the ion species:
B
D .Ba2C / C 2 .I /
D .CsC / C .I /
(10.3.15)
These relations agree with Eq. 10.3.12. Note that .I /, the chemical potential of the ion
common to both salts, appears in both relations.
The solute activity am;B is dened by the relation B D C RT ln am;B (Eq. 10.2.9).
B
Using this relation together with Eqs. 10.1.7 and 10.1.14, we nd that the solute activity is
related to ion molalities by
mC C m
am;B D m;B
(10.3.16)
m
m
where the pressure factor m;B is dened in Eq. 10.2.11. The ion molalities in this expression refer to the constituent ions of solute B, which in a multisolute solution are not
necessarily present in the same stoichiometric ratio as in the solute substance.
For instance, suppose we apply Eq. 10.3.16 to the solution of BaI2 and CsI used above
as an illustration of a multisolute solution, letting am;B be the activity of solute substance
BaI2 . The quantities mC and m in the equation are then the molalities of the Ba2C and I
ions, and is the mean ionic activity coefcient of the dissolved BaI2 . Note that in this
solution the Ba2C and I ions are not present in the 1:2 ratio found in BaI2 , because I is a
constituent of both solutes.
10.3.3 Incomplete dissociation

In the preceding sections of this chapter, the electrolyte solute or solutes have been assumed
to be completely dissociated into their constituent ions at all molalities. Some solutions,
however, contain ion pairsclosely associated ions of opposite charge. Furthermore, in solutions of some electrolytes (often called weak electrolytes), an equilibrium is established
between ions and electrically-neutral molecules. In these kinds of solutions, the relations
between solute molality and ion molalities given by Eq. 10.3.1 are no longer valid. When
dissociation is not complete, the expression for B given by Eq. 10.3.9 can still be used.
However, the quantity appearing in the expression no longer has the physical signicance of being the geometric average of the activity coefcients of the actual dissociated
ions, and is called the stoichiometric activity coefcient of the electrolyte.
10.4 T HE D EBYE H UCKEL T HEORY
295
10.4 THE DEBYEHUCKEL THEORY

The theory of Peter Debye and Erich H ckel (1923) provides theoretical expressions for
u
single-ion activity coefcients and mean ionic activity coefcients in electrolyte solutions.
The expressions in one form or another are very useful for extrapolation of quantities that
include mean ionic activity coefcients to low solute molality or innite dilution.
The only interactions the theory considers are the electrostatic interactions between
ions. These interactions are much stronger than those between uncharged molecules, and
they die off more slowly with distance. If the positions of ions in an electrolyte solution
were completely random, the net effect of electrostatic ionion interactions would be zero,
because each cationcation or anionanion repulsion would be balanced by a cationanion
attraction. The positions are not random, however: each cation has a surplus of anions in
its immediate environment, and each anion has a surplus of neighboring cations. Each ion
therefore has a net attractive interaction with the surrounding ion atmosphere. The result
for a cation species at low electrolyte molality is a decrease of C compared to the cation
at same molality in the absence of ionion interactions, meaning that the single-ion activity
coefcient C becomes less than 1 as the electrolyte molality is increased beyond the idealdilute range. Similarly,
also becomes less than 1.
According to the DebyeH ckel theory, the single-ion activity coefcient i of ion i in
u
a solution of one or more electrolytes is given by
p
ADH zi2 Im
(10.4.1)
ln i D
p
1 C BDH a Im
where
zi D the charge number of ion i (C1, 2, etc.);
Im D the ionic strength of the solution on a molality basis, dened by3

X
def 1
2
Im D
2
mj zj
(10.4.2)
all ions
ADH and BDH are dened functions of the kind of solvent and the temperature;
a is an adjustable parameter, equal to the mean effective distance of closest approach of
other ions in the solution to one of the i ions.
The denitions of the quantities ADH and BDH appearing in Eq. 10.4.1 are
def
2
ADH D NA e 3 =8
def
BDH D NA e 2
2
1=2
1=2
A
r 0 RT /
r 0 RT /
1=2
3=2
(10.4.3)
(10.4.4)
where NA is the Avogadro constant, e is the elementary charge (the charge of a proton),
A and r are the density and relative permittivity (dielectric constant) of the solvent,
and 0 is the electric constant (or permittivity of vacuum).
3
Lewis and Randall (Ref. [101]) introduced the term ionic strength, dened by this equation, two years before
the DebyeH ckel theory was published. They found empirically that in dilute solutions, the mean ionic activity
u
coefcient of a given strong electrolyte is the same in all solutions having the same ionic strength.
296
BIOGRAPHICAL SKETCH
Peter Josephus Wilhelmus Debye (18841966)

actually had not been interested in electrolytes
at all. He had been at a colloquium at which a
new theory of electrolytes had been described
that was supposed to explain why the conductivity of a dilute solution of a strong electrolyte
is proportional to the square root of the concentration. Debye, on hearing this description,
objected that the theory neglected the effects
of Brownian motion. The discussion became
heated, and some of those present told Debye
you will have to do something about it. What
Peter Debye made major contributions to vari- Debye did about it was to ask his assistant,
ous areas of chemistry and physics.
Erich H ckel, to study the literature and nd
u
He was born in Maastricht, The Nether- out what they were missing. That, according
lands, where his father was foreman in a ma- to Debye in the interview, is how the Debye
chine workshop.
H ckel theory came about.b
u
Henri Sack, a close associate for 40 years,
In a reminiscence of Debye published in
recalled in 1968:a
1972, Erich H ckel wrote:c
u
He was not only endowed with a most powerful
and penetrating intellect and an unmatched ability for presenting his ideas in a most lucid way,
but he also knew the art of living a full life. He
greatly enjoyed his scientic endeavors, he had
a deep love for his family and home life, and he
had an eye for the beauties of nature and a taste
for the pleasure of the out-of-doors as manifested
by his hobbies such as shing, collecting cacti,
and gardening, mostly in the company of Mrs.
Debye.
Before World War II, Debye held appointments at several universities in The Netherlands, Switzerland, and Germany. He emigrated to America in 1940 and was at Cornell
University in Ithaca, New York until his death.
He became an American citizen in 1946.
Debye was responsible for theoretical treatments of a variety of subjects, including
molecular dipole moments (for which the debye is a non-SI unit), X-ray diffraction and
scattering, and light scattering. His theories
relevant to thermodynamics include the temperature dependence of the heat capacity of
crystals at a low temperature (Debye crystal theory), adiabatic demagnetization, and the
DebyeH ckel theory of electrolyte solutions.
u
In an interview in 1962, Debye said that he
a Ref.
[57], page 232.
b Ref.
[43].
c Translation
My personal relations with Debye were always

completely care-free. Although I was 12 years
younger than he and a complete freshman when
I came to Z rich, he always treated me as his
u
equal.
. . . Debye conceived his workin my
opinionas an artist who operates on the basis of joy in his work and its creations, and who
was led often by intuition, which was then later
on rationally founded in the most plain and clear
way leaving out everything that was unessential.
. . . I never found in Debye any interest in philosophical questions. Debyes way of life seemed
to me rather straightforward and uncomplicated.
He liked a good dinner: when a problem could
not be solved after a physics lecture, he used to
say: one must enjoy a good evening dinner and
then the inspiration comes by itself. . . Debye
received an immense number of awards. It did
not seem to matter much to him. When I visited him in Berlin to congratulate him on the
Nobel Prize, he interrupted: Fine that you are
here. My congratulations were therefore not
completed.
Debye was awarded the 1936 Nobel Prize in

Chemistry for his contributions to our knowledge of molecular structure through his investigations on dipole moments and on the
diffraction of X-rays and electrons in gases.
in Ref. [159], pages 7374.
297
1:0
0:9
0:8
0:7
0:02
0:04
0:06
mB =mol kg 1
0:08
0:10
Figure 10.3 Mean ionic activity coefcient of aqueous HCl at 25 C. Solid curve:
experiment; a dashed curve: DebyeH ckel theory with a D 5 10 10 m; dotted
u
curve: DebyeH ckel limiting law.
u
a Ref.
[74], Table 11-5-1.
When the solvent is water at 25 C, the quantities ADH and BDH have the values
ADH D 1:1744 kg1=2 mol
BDH D 3:285
1=2
(10.4.5)
109 m 1 kg1=2 mol 1=2
(10.4.6)
From Eqs. 10.3.8 and 10.4.1 and the electroneutrality condition C zC D z , we obtain the following expression for the logarithm of the mean ionic activity coefcient of an
electrolyte solute:
p
ADH zC z Im
(10.4.7)
ln D
p
1 C BDH a Im
In this equation, zC and z are the charge numbers of the cation and anion of the solute.
Since the right side of Eq. 10.4.7 is negative at nite solute molalities, and zero at innite
dilution, the theory predicts that is less than 1 at nite solute molalities and approaches
1 at innite dilution.
Figure 10.3 shows that with the proper choice of the parameter a, the mean ionic activity
coefcient of HCl calculated from Eq. 10.4.7 (dashed curve) agrees closely with experiment
(solid curve) at low molalities.
As the molalities of all solutes become small, Eq. 10.4.7 becomes
p
(10.4.8)
ln D ADH zC z Im
(innite dilution)
This form is known as the DebyeHuckel limiting law. Note that the limiting law contains
no adjustable parameters. The dotted curve in Fig. 10.3 shows that the limiting law agrees
with experiment only at quite low molality.
10.5 D ERIVATION OF THE D EBYE H UCKEL E QUATION
298
The ionic strength Im is calculated from Eq. 10.4.2 with the molalities of all ions in the
solution, not just the molality of the ion or solute whose activity coefcient we are interested
in. This is because, as explained above, the departure of C and
from the ideal-dilute
value of 1 is caused by the interaction of each ion with the ion atmosphere resulting from
all other ions in the solution.
In a binary solution of a single electrolyte solute, assumed to be completely dissociated,
the relation between the ionic strength and the solute molality depends on (the number of
ions per solute formula unit) and the charge numbers zC and z . The ionic strength is given
P
2
by Im D .1=2/ i mi zi2 D .1=2/. C zC C z 2 /mB . With the help of the electroneutrality
condition C zC D . z /, this becomes
Im D 1 .
2
D
1
2
1
2
z /zC
C zC /z
mB
. C C /zC z mB
zC z mB
(10.4.9)
We nd the following relations hold between Im and mB in the binary solution, depending
on the stoichiometry of the solute formula unit:
For a 1:1 electrolyte, e.g., NaCl or HCl: Im D mB
For a 1:2 or 2:1 electrolyte, e.g., Na2 SO4 or CaCl2 : Im D 3mB

For a 2:2 electrolyte, e.g., MgSO4 : Im D 4mB
For a 1:3 or 3:1 electrolyte, e.g., AlCl3 : Im D 6mB
For a 3:2 or 2:3 electrolyte, e.g., Al2 (SO4 )3 : Im D 15mB
p
Figure 10.4 on the next page shows ln as a function of Im for aqueous HCl and
CaCl2 . The experimental curves have the limiting slopes predicted by the DebyeH ckel
u
limiting law (Eq. 10.4.8), but at a low ionic strength the curves begin to deviate signicantly
from the linear relations predicted by that law. The full DebyeH ckel equation (Eq. 10.4.7)
u
ts the experimental curves over a wider range of ionic strength.
10.5 DERIVATION OF THE DEBYEHUCKEL EQUATION

Debye and H ckel derived Eq. 10.4.1 using a combination of electrostatic theory, statisu
tical mechanical theory, and thermodynamics. This section gives a brief outline of their
derivation.
The derivation starts by focusing on an individual ion of species i as it moves through
the solution; call it the central ion. Around this central ion, the time-average spatial distribution of any ion species j is not random, on account of the interaction of these ions of
species j with the central ion. (Species i and j may be the same or different.) The distribution, whatever it is, must be spherically symmetric about the central ion; that is, a function
0
only of the distance r from the center of the ion. The local concentration, cj , of the ions of
species j at a given value of r depends on the ion charge zj e and the electric potential at
that position. The time-average electric potential in turn depends on the distribution of all
0
ions and is symmetric about the central ion, so expressions must be found for cj and as
functions of r that are mutually consistent.
10.5 D ERIVATION OF THE D EBYE H UCKEL E QUATION
299
0
0:1
HCl
ln
0:2
CaCl2
0:3
0:4
0:5
0:1
0:2
Im =mol kg
0:3

1 1=2
p
Figure 10.4 Dependence of ln on Im for aqueous HCl (upper curves) and aqueous CaCl2 (lower curves) at 25 C. a Solid curves: experimental; dashed curves:
DebyeH ckel equation (a D 5 10 10 m for HCl, a D 4:5 10 10 m for CaCl2 );
u
dotted lines: DebyeH ckel limiting law.
u
a Experimental
curves from parameter values in Ref. [74], Tables 11-5-1 and 12-1-3a.
0
Debye and H ckel assumed that cj is given by the Boltzmann distribution
u
0
cj D cj e
zj e =kT
(10.5.1)
where zj e is the electrostatic energy of an ion of species j , and k is the Boltzmann con0
stant (k D R=NA ). As r becomes large, approaches zero and cj approaches the macro0
scopic concentration cj . As T increases, cj at a xed value of r approaches cj because
of the randomizing effect of thermal energy. Debye and H ckel expanded the exponential
u
0
function in powers of 1=T and retained only the rst two terms: cj
cj .1 zj e =kT /.
The distribution of each ion species is assumed to follow this relation. The electric potential
function consistent with this distribution and with the electroneutrality of the solution as a
whole is
D .zi e=4 r 0 r/e .a r/ =.1 C a/
(10.5.2)
2
=
Here is dened by 2 D 2NA e 2 IcP r 0 RT , where Ic is the ionic strength on a concentration basis dened by Ic D .1=2/ i ci zi2 .
The electric potential at a point is assumed to be a sum of two contributions: the
electric potential the central ion would cause at innite dilution, zi e=4 r 0 r, and the
electric potential due to all other ions, 0 . Thus, 0 is equal to
zi e=4 r 0 r, or
0
D .zi e=4
r 0 r/e
.a r/
=.1 C a/
(10.5.3)
This expression for 0 is valid for distances from the center of the central ion down to a, the
distance of closest approach of other ions. At smaller values of r, 0 is constant and equal
to the value at r D a, which is 0 .a/ D .zi e=4 r 0 /=.1 C a/. The interaction energy

10.6 M EAN I ONIC ACTIVITY C OEFFICIENTS FROM O SMOTIC C OEFFICIENTS
300
between the central ion and the surrounding ions (the ion atmosphere) is the product of the
central ion charge and 0 .a/.
The last step of the derivation is the calculation of the work of a hypothetical reversible
process in which the surrounding ions stay in their nal distribution, and the charge of the
central ion gradually increases from zero to its actual value zi e. Let zi e be the charge at
each stage of the process, where is a fractional advancement that changes from 0 to 1.
Then the work w 0 due to the interaction of the central ion with its ion atmosphere is 0 .a/
integrated over the charge:
Z D1
0
w D
.zi e=4 r 0 /=.1 C a/ d.zi /
D
D0
.zi2 e 2 =8
r 0 /=.1
C a/
(10.5.4)
Since the innitesimal Gibbs energy change in a reversible process is given by dG D

S dT C V dp C w 0 (Eq. 5.8.6), this reversible nonexpansion work at constant T and
p is equal to the Gibbs energy change. The Gibbs energy change per amount of species
i is w 0 NA D .zi2 e 2 NA =8 r 0 /=.1 C a/. This quantity is G=ni for the process in
which a solution of xed composition changes from a hypothetical state lacking ionion
interactions to the real state with ionion interactions present. G=ni may be equated to
the difference of the chemical potentials of i in the nal and initial states. If the chemical
potential without ionion interactions is taken to be that for ideal-dilute behavior on a molality basis, i D ref C RT ln.mi =m /, then .zi2 e 2 NA =8 r 0 /=.1 C a/ is equal to
m;i
ref C RT ln.mi =m / D RT ln m;i . In a dilute solution, ci can with little error be
i
m;i
set equal to A mi , and Ic to A Im . Equation 10.4.1 follows.
10.6 MEAN IONIC ACTIVITY COEFFICIENTS FROM OSMOTIC

COEFFICIENTS
Recall that is the mean ionic activity coefcient of a strong electrolyte, or the stoichiometric activity coefcient of an electrolyte that does not dissociate completely.
The general procedure described in this section for evaluating requires knowledge
of the osmotic coefcient m as a function of molality. m is commonly evaluated by
the isopiestic method (Sec. 9.6.4) or from measurements of freezing-point depression (Sec.
12.2).
The osmotic coefcient of a binary solution of an electrolyte is dened by
def
m
RTMA mB
(10.6.1)
(binary electrolyte solution)
That is, for an electrolyte the sum iA mi appearing in the denition of m for a nonelectrolyte solution (Eq. 9.6.11 on page 267) is replaced by mB , the sum of the ion molalities
assuming complete dissociation. It will now be shown that m dened this way can be used
to evaluate .
The derivation is like that described in Sec. 9.6.3 for a binary solution of a nonelectrolyte. Solving Eq. 10.6.1 for A and taking the differential of A at constant T and p, we
obtain
d A D RTMA . m dmB C mB d m /
(10.6.2)

10.6 M EAN I ONIC ACTIVITY C OEFFICIENTS FROM O SMOTIC C OEFFICIENTS
301
From Eq. 10.3.9 on page 293, we obtain

d
D RT
d ln
dmB
C
mB
(10.6.3)
Substitution of these expressions in the GibbsDuhem equation nA d

together with the substitution nA MA D nB =mB , yields
d ln
Dd
mB
C nB d
dmB
D 0,
(10.6.4)
Then integration from mB D 0 to any desired molality m0 gives the result

B
ln
0
.mB /
0
m .mB /
1C
m0
B
mB
dmB
(10.6.5)
The right side of this equation is the same expression as derived for ln m;B for a nonelectrolyte (Eq. 9.6.20 on page 268).
The integrand of the integral on the right side of Eq. 10.6.5 approaches 1 as mB
approaches zero, making it difcult to evaluate the integral by numerical integration starting
at mB D 0. (This difculty does not exist when the solute is a nonelectrolyte.) Instead, we
can split the integral into two parts
Z
m0
B
mB
dmB D
m00
B
mB
dmB C
m0
B
mB
m00
B
dmB
(10.6.6)
where the integration limit m00 is a low molality at which the value of m is available and
B
at which can either be measured or estimated from the DebyeH ckel equation.
u
We next rewrite Eq. 10.6.5 with m0 replaced with m00 :
B
B
ln
00
.mB /
00
m .mB /
1C
m00
B
mB
dmB
(10.6.7)
By eliminating the integral with an upper limit of m00 from Eqs. 10.6.6 and 10.6.7, we obtain
B
Z
m0
B
0
mB
dmB D ln
00
.mB /
00
m .mB /
C1C
m0
B
m00
B
mB
dmB
(10.6.8)
dmB
(10.6.9)
Equation 10.6.5 becomes

ln
0
.mB /
0
m .mB /
00
m .mB /
C ln
00
.mB /
m0
B
m00
B
mB
The integral on the right side of this equation can easily be evaluated by numerical integration.

P ROBLEMS
302
PROBLEMS
in Appendix I.
10.1 The mean ionic activity coefcient of NaCl in a 0.100 molal aqueous solution at 298:15 K
has been evaluated with measurements of equilibrium cell potentials,4 with the result ln D
0:2505. Use this value in Eq. 10.6.9, together with the values of osmotic coefcients in
Table 10.1, to evaluate at each of the molalities shown in the table; then plot as a
function of mB .
Table 10.1 Osmotic coefcients of aqueous NaCl at
298:15 K a
mB =mol kg
0.1
0.2
0.3
0.5
1.0
1.5
a Ref.
0.9325
0.9239
0.9212
0.9222
0.9373
0.9598
mB =mol kg
2.0
3.0
4.0
5.0
6.0
0.9866
1.0485
1.1177
1.1916
1.2688
[31].
10.2 Rard and Miller5 used published measurements of the freezing points of dilute aqueous solutions of Na2 SO4 to calculate the osmotic coefcients of these solutions at 298:15 K. Use
their values listed in Table 10.2 to evaluate the mean ionic activity coefcient of Na2 SO4 at
298:15 K and a molality of mB D 0:15 mol kg 1 . For the parameter a in the DebyeH ckel
u
equation (Eq. 10.4.7), use the value a D 3:0 10 10 m.
Table 10.2 Osmotic coefcients of aqueous Na2 SO4
at 298:15 K
mB =mol kg
0.0126
0.0181
0.0228
0.0272
0.0374
0.0435
0.0542
0.0594
4 Ref.
[143], Table 9.3.
5 Ref.
0.8893
0.8716
0.8607
0.8529
0.8356
0.8294
0.8178
0.8141
mB =mol kg
0.0637
0.0730
0.0905
0.0996
0.1188
0.1237
0.1605
0.8111
0.8036
0.7927
0.7887
0.7780
0.7760
0.7616
[139].
C HAPTER 11
R EACTIONS AND OTHER C HEMICAL

P ROCESSES
This chapter discusses the thermodynamics of mixing processes and processes described
by reaction equations (chemical equations). It introduces the important concepts of molar
mixing and reaction quantities, advancement, and the thermodynamic equilibrium constant.
The focus is on chemical processes that take place in closed systems at constant pressure,
with no work other than expansion work. Under these conditions, the enthalpy change is
equal to the heat (Eq. 5.3.7). The processes either take place at constant temperature, or
have initial and nal states of the same temperature.
Most of the processes to be described involve mixtures and have intermediate states that
are nonequilibrium states. At constant temperature and pressure, these processes proceed
spontaneously with decreasing Gibbs energy (Sec. 5.8).1 When the rates of change are slow
enough for thermal and mechanical equilibrium to be maintained, the spontaneity is due
to lack of transfer equilibrium or reaction equilibrium. An equilibrium phase transition of
a pure substance, however, is a special case: it is a reversible process of constant Gibbs
energy (Sec. 8.3).
11.1 MIXING PROCESSES

A mixing process is a process in which a mixture is formed from pure substances. In the
initial state the system has two or more separate phases, each containing a different pure
substance at the same temperature and pressure. The nal state is a single-phase mixture at
this temperature and pressure.
The process is illustrated schematically in Fig. 11.1 on the next page. When the partition
is withdrawn, the two pure liquids mix spontaneously at constant pressure to form a single
homogeneous phase. If necessary, heat transfer is used to return the phase to the initial
temperature.
1 Processes
in which G decreases are sometimes called exergonic.
303
CHAPTER 11 REACTIONS AND OTHER CHEMICAL PROCESSES

11.1 M IXING P ROCESSES
304
mixture of
A and B
Figure 11.1 Initial state (left) and nal state (right) of mixing process for liquid substances A and B.
11.1.1 Mixtures in general

First let us consider changes in the Gibbs energy G. Since this is an extensive property, G
in the initial state 1 is the sum of G for each pure phase:
X
G1 D
ni i
(11.1.1)
i
Here i is the chemical potential (i.e., the molar Gibbs energy) of pure substance i at the
initial temperature and pressure. For the nal state 2, we use the additivity rule for a mixture
X
G2 D
ni i
(11.1.2)
i
where i is the chemical potential of i in the mixture at the same temperature and pressure
as the initial state. The overall change of G, the Gibbs energy of mixing, is then
X
G(mix) D G2 G1 D
ni . i
(11.1.3)
i/
i
The molar Gibbs energy of mixing is the Gibbs energy of mixing per amount of mixP
ture formed; that is, Gm (mix) D G(mix)=n, where n is the sum i ni . Dividing both
sides of Eq. 11.1.3 by n, we obtain
X
Gm (mix) D
xi . i
(11.1.4)
i/
i
where xi is the mole fraction of substance i in the nal mixture.

Following the same procedure for an extensive state function X , we derive the following
general relation for its molar mixing quantity:
X
Xm (mix) D
xi .Xi Xi /
(11.1.5)
i

When the mixture formed is an ideal mixture (gas, liquid, or solid), and the pure constituents have the same physical state as the mixture, the expressions for various molar
mixing quantities are particularly simple. An ideal molar mixing quantity will be indicated

305
id
by a superscript id as in Gm (mix). The general denition of an ideal molar mixing
quantity, analogous to Eq. 11.1.5, is
X
id
Xm (mix) D
xi .Xiid Xi /
(11.1.6)
i
The chemical potential of constituent i of an ideal mixture is related to the mole fraction
xi by the relation (Eq. 9.4.8)
(11.1.7)
i D i C RT ln xi
By combining this relation with Eq. 11.1.4, we nd the molar Gibbs energy of mixing to
form an ideal mixture is given by
X
id
xi ln xi
(11.1.8)
Gm (mix) D RT
i
Since each mole fraction is less than one and the logarithm of a fraction is negative, it
id
follows that Gm (mix) is negative for every composition of the mixture.
We obtain expressions for other molar mixing quantities by substituting formulas for
partial molar quantities of constituents of an ideal mixture derived in Sec. 9.4.3 into Eq.
11.1.5. From Si D Si
R ln xi (Eq. 9.4.9), we obtain
X
id
Sm (mix) D R
xi ln xi
(11.1.9)
i
This quantity is positive.

Although the molar entropy of mixing to form an ideal mixture is positive, this is not
true for some nonideal mixtures. McGlashan2 cites the negative value Sm (mix) D
8:8 J K 1 mol 1 for an equimolar mixture of diethylamine and water at 322 K.
From Hi D Hi (Eq. 9.4.10) and Ui D Ui (Eq. 9.4.12), we have

id
Hm (mix) D 0
(11.1.10)
id
Um (mix) D 0
(11.1.11)
id
Vm (mix) D 0
(11.1.12)
and
Thus, the mixing of liquids that form an ideal mixture is an athermal process, one in which
no heat transfer is needed to keep the temperature constant.
From Vi D Vi (Eq. 9.4.11), we get
showing that the ideal molar volume of mixing is zero. Thus an ideal mixture has the same
volume as the sum of the volumes of the pure components at the same T and p.3
id
id
id
Figure 11.2 on the next page shows how Gm (mix), TSm (mix), and Hm (mix)
depend on the composition of an ideal mixture formed by mixing two pure substances.
id
id
Although it is not obvious in the gure, the curves for Gm (mix) and TSm (mix) have
slopes of C1 or 1 at xA D0 and xA D1.
2 Ref.
[112], p. 241.
the fact mentioned on p. 227 that the volume of a mixture of water and methanol is different from the
sum of the volumes of the pure liquids, we can deduce that this mixture is nonideal, despite the fact that water
and methanol mix in all proportions.
3 From

306
id
m (mix)
quantity / kJ mol
id
m (mix)
id
m (mix)
Figure 11.2 Molar mixing quantities for a binary ideal mixture at 298:15 K.
11.1.3 Excess quantities

An excess quantity X E of a mixture is dened as the difference between the value of the
extensive property X of the real mixture and X id , the value for a hypothetical ideal mixture
at the same temperature, pressure, and composition.
E
An excess molar quantity Xm is the excess quantity divided by n, the total amount of
all constituents of the mixture. Examining the dependence of excess molar quantities on
composition is a convenient way to characterize deviations from ideal-mixture behavior.
Excess molar quantities are related to molar mixing quantities as follows:
!
X
X
E
id
id
Xm D .X X /=n D
ni X i
ni Xi =n
i
D
D
xi Xi
Xiid
xi .Xi
Xi /
D Xm (mix)
xi .Xiid
Xi /
i
id
Xm (mix)
(11.1.13)
id
By substituting expressions for Xm (mix) from Eqs. 11.1.811.1.12 in Eq. 11.1.13, we obtain the following expressions for the excess molar Gibbs energy, entropy, enthalpy, internal
energy, and volume:
X
E
Gm D Gm (mix) RT
xi ln xi
(11.1.14)
i
Sm D Sm (mix) C R
E
Hm D Hm (mix)
xi ln xi
(11.1.15)
(11.1.16)

E
Um D Um (mix)
307
(11.1.17)
E
Vm D Vm (mix)
(11.1.18)
By substitution from Eqs. 9.5.14 and 11.1.4 in Eq. 11.1.14, we can relate the excess molar
Gibbs energy to the activity coefcients of the mixture constituents based on pure-liquid
reference states:
X
E
Gm D RT
xi ln i
(11.1.19)
i
It is also possible to derive the useful relation

"
#
E
@ nGm
@ni
T;p;nj i
D RT ln
(11.1.20)
To derive Eq. 11.1.20, consider innitesimal changes in the mixture composition at

constant T and p. From Eq. 11.1.19, we write
X
X
X
E
d nGm D RT
d.ni ln i / D RT
ni d ln i C RT
.ln i / dni (11.1.21)
i
From i D i C RT ln. i xP we have d i D RT .d ln i C dxi =xi /. Substitution in

i /,
the GibbsDuhem equation, i xi d i D 0, gives
X
X
xi d ln i C
dxi D 0
(11.1.22)
i
P
In Eq. 11.1.22, we set the sum i dxi equal to zero (because i xi equals 1) and
P
multiply by the total amount, n, resulting in i ni d ln i D 0. This turns Eq. 11.1.21
into
X
E
.ln i / dni
(11.1.23)
d nGm D RT
i
from which Eq. 11.1.20 follows.
11.1.4 The entropy change to form an ideal gas mixture

When pure ideal gases mix at constant T and p to form an ideal gas mixture, the molar
P
id
entropy change Sm (mix) D R i yi ln yi (Eq. 11.1.9) is positive.
Consider a pure ideal-gas phase. Entropy is an extensive property, so if we divide this
phase into two subsystems with an internal partition, the total entropy remains unchanged.
The reverse process, the removal of the partition, must also have zero entropy change. Despite the fact that the latter process allows the molecules in the two subsystems to intermingle without a change in T or p, it cannot be considered mixing because the entropy
does not increase. The essential point is that the same substance is present in both of the
subsystems, so there is no macroscopic change of state when the partition is removed.
From these considerations, one might conclude that the fundamental reason the entropy
increases when pure ideal gases mix is that different substances become intermingled. This
conclusion would be mistaken, as we will now see.
The partial molar entropy of constituent i of an ideal gas mixture is related to its partial
pressure pi by Eq. 9.3.6:
Si D Si R ln.pi =p /
(11.1.24)

308
B
A+B
A+B
12
(a)
(b)
(c)
Figure 11.3 Reversible mixing process for ideal gases A and B conned in a cylinder.
Piston 1 is permeable to A but not B; piston 2 is permeable to B but not A.
(a) Gases A and B are in separate phases at the same temperature and pressure.
(b) The pistons move apart at constant temperature with negative reversible work,
creating an ideal gas mixture of A and B in continuous transfer equilibrium with the
pure gases.
(c) The two gases are fully mixed at the initial temperature and pressure.
But pi is equal to ni RT =V (Eq. 9.3.3). Therefore, if a xed amount of i is in a container

at a given temperature, Si depends only on the volume of the container and is unaffected by
the presence of the other constituents of the ideal gas mixture.
When Eq. 11.1.24 is applied to a pure ideal gas, it gives an expression for the molar
entropy
Si D Si R ln.p=p /
(11.1.25)
where p is equal to nRT =V .
From Eqs. 11.1.24 and 11.1.25, and the fact that the entropy of a mixture is given by the
P
additivity rule S D i ni Si , we conclude that the entropy of an ideal gas mixture equals
the sum of the entropies of the unmixed pure ideal gases, each pure gas having the same
temperature and occupying the same volume as in the mixture.
We can now understand why the entropy change is positive when ideal gases mix at
constant T and p: Each substance occupies a greater volume in the nal state than initially.
Exactly the same entropy increase would result if the volume of each of the pure ideal gases
were increased isothermally without mixing.
The reversible mixing process depicted in Fig. 11.3 illustrates this principle. The initial
state shown in Fig. 11.3(a) consists of volume V1 (A) of pure ideal gas A and volume V1 (B)
of pure ideal gas B, both at the same T and p. The hypothetical semipermeable pistons are
moved apart reversibly and isothermally to create an ideal gas mixture, as shown in Fig.
11.3(b). According to an argument in Sec. 9.3.3, transfer equilibrium across the semipermeable pistons requires partial pressure pA in the mixture to equal the pressure of the pure
A at the left, and partial pressure pB in the mixture to equal the pressure of the pure B at
the right. Thus in intermediate states of the process, gas A exerts no net force on piston 1,
and gas B exerts no net force on piston 2.
In the nal state shown in Fig. 11.3(c), the gases are fully mixed in a phase of volume
V2 DV1 (A)CV1 (B). The movement of piston 1 has expanded gas B with the same reversible
work as if gas A were absent, equal to nB RT lnV2 =V1 (B). Likewise, the reversible
work to expand gas A with piston 2 is the same as if B were absent: nA RT lnV2 =V1 (A).
Because the initial and nal temperatures and pressures are the same, the mole fractions in

309
the nal mixture are yA D V1 (A)=V2 and yB D V1 (B)=V2 . The total work of the reversible
mixing process is therefore w D nA RT ln yA C nB RT ln yB , the heat needed to keep the
internal energy constant is q D w, and the entropy change is
S D q=T D
nA R ln yA
nB R ln yB
(11.1.26)
It should be clear that isothermal expansion of both pure gases from their initial volumes
to volume V2 without mixing would result in the same total work and the same entropy
change.
When we divide Eq. 11.1.26 by n D nA C nB , we obtain the expression for the molar
entropy of mixing given by Eq. 11.1.9 with xi replaced by yi for a gas.
11.1.5 Molecular model of a liquid mixture

We have seen that when two pure liquids mix to form an ideal liquid mixture at the same T
and p, the total volume and internal energy do not change. A simple molecular model of
a binary liquid mixture will elucidate the energetic molecular properties that are consistent
with this macroscopic behavior. The model assumes the excess molar entropy, but not
necessarily the excess molar internal energy, is zero. The model is of the type sometimes
called the quasicrystalline lattice model, and the mixture it describes is sometimes called
a simple mixture. Of course, a molecular model like this is outside the realm of classical
thermodynamics.
The model is for substances A and B in gas and liquid phases at a xed temperature.
Let the standard molar internal energy of pure gaseous A be UA (g). This is the molar
energy in the absence of intermolecular interactions, and its value depends only on the
molecular constitution and the temperature. The molar internal energy of pure liquid A is
lower because of the attractive intermolecular forces in the liquid phase. We assume the
energy difference is equal to a sum of pairwise nearest-neighbor interactions in the liquid.
Thus, the molar internal energy of pure liquid A is given by
UA D UA (g) C kAA
(11.1.27)
where kAA (approximately the negative of the molar internal energy of vaporization) is the
interaction energy per amount of A due to AA interactions when each molecule of A is
surrounded only by other molecules of A.
Similarly, the molar internal energy of pure liquid B is given by
UB D UB (g) C kBB
(11.1.28)
where kBB is for BB interactions.

We assume that in a liquid mixture of A and B, the numbers of nearest-neighbor molecules of A and B surrounding any given molecule are in proportion to the mole fractions xA
and xB .4 Then the number of AA interactions is proportional to nA xA , the number of BB
4 This assumption requires the molecules of A and B to have similar sizes and shapes and to be randomly mixed
in the mixture. Statistical mechanics theory shows that the molecular sizes must be approximately equal if the
excess molar entropy is to be zero.

310
interactions is proportional to nB xB , and the number of AB interactions is proportional to

nA xB C nB xA . The internal energy of the liquid mixture is then given by
U (mixt) D nA UA (g) C nB UB (g)
C nA xA kAA C nB xB kBB C .nA xB C nB xA /kAB
(11.1.29)
where kAB is the interaction energy per amount of A when each molecule of A is surrounded
only by molecules of B, or the interaction energy per amount of B when each molecule of
B is surrounded only by molecules of A.
The internal energy change for mixing amounts nA of liquid A and nB of liquid B is
now
U (mix) D U (mixt)
nA UA
nB U B
D nA xA kAA C nB xB kBB C .nA xB C nB xA /kAB

D nA .xA
1/kAA C nB .xB
1/kBB C .nA xB C nB xA /kAB
With the identities xA 1 D xB , xB

(where n is the sum nA C nB ), we obtain
U (mix) D
nA kAA
1 D
nA nB
.2kAB
n
nB kBB
(11.1.30)
xA , and nA xB D nB xA D nA nB =n
kAA
kBB /
(11.1.31)
If the internal energy change to form a mixture of any composition is to be zero, as it is for
an ideal mixture, the quantity .2kAB kAA kBB / must be zero, which means kAB must
equal .kAA C kBB /=2. Thus, one requirement for an ideal mixture is that an AB interaction
equals the average of an AA interaction and a BB interaction.
If we write Eq. 11.1.29 in the form
U (mixt) D nA UA (g) C nB UB (g) C
1
.n2 k C 2nA nB kAB C n2 kBB / (11.1.32)
B
nA C nB A AA
we can differentiate with respect to nB at constant nA to evaluate the partial molar internal
energy of B. The result can be rearranged to the simple form
UB D UB C .2kAB
kAA
kBB / .1
xB /2
(11.1.33)
where UB is given by Eq. 11.1.28. Equation 11.1.33 predicts that the value of UB decreases
with increasing xB if kAB is less negative than the average of kAA and kBB , increases for the
opposite situation, and is equal to UB in an ideal liquid mixture.
When the excess molar volume and entropy are set equal to zero, the model describes
E
what is called a regular solution.5 The excess molar Gibbs energy of a mixture is Gm D
E
E
E
Um C pVm T Sm . Using the expression of Eq. 11.1.31 with the further assumptions that
E
E
Vm and Sm are zero, this model predicts the excess molar Gibbs energy is given by
E
Gm D
U (mix)
D xA xB .2kAB
n
kAA
kBB /
(11.1.34)
This is a symmetric function of xA and xB . It predicts, for example, that coexisting liquid
layers in a binary system (Sec. 11.1.6) have the same value of xA in one phase as the value
of xB in the other.
5 Ref.
[79].

311
m (mix)
Figure 11.4 Molar Gibbs energy of mixing as a function of the composition of a

binary liquid mixture with spontaneous phase separation. The inection points are
indicated by lled circles.
Molar excess Gibbs energies of real liquid mixtures are often found to be unsymmetric
functions. To represent them, a more general function is needed. A commonly used function
for a binary mixture is the RedlichKister series given by
E
Gm D xA xB a C b.xA
xB / C c.xA
x B /2 C
(11.1.35)
where the parameters a; b; c;

depend on T and p but not on composition. This function
E
E
satises a necessary condition for the dependence of Gm on composition: Gm must equal
6
zero when either xA or xB is zero.
E
For many binary liquid systems, the measured dependence of Gm on composition is
reproduced reasonably well by the two-parameter RedlichKister series
E
Gm D xA xB a C b.xA
xB /
(11.1.36)
in which the parameters a and b are adjusted to t the experimental data. The activity
coefcients in a mixture obeying this equation are found, from Eq. 11.1.20, to be given by
RT ln
2
D xB a C .3
4xB /b
RT ln
2
D xA a C .4xA
3/b
(11.1.37)
11.1.6 Phase separation of a liquid mixture

A binary liquid mixture in a system maintained at constant T and p can spontaneously
separate into two liquid layers if any part of the curve of a plot of Gm (mix) versus xA is
concave downward. To understand this phenomenon, consider Fig. 11.4. This gure is a
plot of Gm (mix) versus xA . It has the form needed to evaluate the quantities . A
A/
and . B
/ by the variant of the method of intercepts described on page 234. On this
B
6 The
reason for this condition can be seen by looking at Eq. 11.1.19 on page 307. For a binary mixture, this
E
equation becomes Gm D RT .xA ln A C xB ln B /. When xA is zero, B is 1 and ln B is zero. When xB is
E
zero, A is 1 and ln A is zero. Thus Gm must be zero in both cases.

312
plot, the tangent to the curve at any given composition has intercepts equal to . B
B / at
xA D0 and . A
A / at xA D1.
In order for two binary liquid phases to be in transfer equilibrium, A must be the
same in both phases and B must also be the same in both phases. The dashed line in the
gure is a common tangent to the curve at the points labeled and . These two points
are the only ones having a common tangent, and what makes the common tangent possible
is the downward concavity (negative curvature) of a portion of the curve between these
points. Because the tangents at these points have the same intercepts, phases and of
compositions xA and xA can be in equilibrium with one another: the necessary conditions
A D A and B D B are satised.

Now consider point 1 on the curve. A phase of this composition is unstable. It will
spontaneously separate into the two phases of compositions xA and xA , because the Gibbs
energy per total amount then decreases to the extent indicated by the vertical arrow from
point 1 to point 2. We know that a process in which G decreases at constant T and p in a
closed system, with expansion work only, is a spontaneous process (Sec. 5.8).
To show that the arrow in Fig. 11.4 represents the change in G=n for phase separation, we let y represent the vertical ordinate and write the equation of the dashed line
through points and (y as a function of xA ):
!
y y
.xA xA /
(11.1.38)
y D y C
xA xA
In the system both before and after phase separation occurs, xA is the mole fraction of
component A in the system as a whole. When phases and are present, containing
amounts n and n , xA is given by the expression
xA n C xA n
(11.1.39)
n C n
By substituting this expression for xA in Eq. 11.1.38, after some rearrangement and
using n C n D n, we obtain
1
(11.1.40)
yD
n y C n y
n
which equates y for a point on the dashed line to the Gibbs energy change for mixing
pure components to form an amount n of phase and an amount n of phase ,
divided by the total amount n. Thus, the difference between the values of y at points
1 and 2 is the decrease in G=n when a single phase separates into two equilibrated
phases.
xA D
Any mixture with a value of xA between xA and xA is unstable with respect to separation
into two phases of compositions xA and xA . Phase separation occurs only if the curve of
the plot of Gm (mix) versus xA is concave downward, which requires the curve to have at
least two inection points. The compositions of the two phases are not the compositions at
the inection points, nor in the case of the curve shown in Fig. 11.4 are these compositions
the same as those of the two local minima.
By varying the values of parameters in an expression for the excess molar Gibbs energy,
we can model the onset of phase separation caused by a temperature change. Figure 11.5
shows the results of using the two-parameter RedlichKister series (Eq. 11.1.36).

11.2 T HE A DVANCEMENT AND M OLAR R EACTION Q UANTITIES
m (mix)
313
(a)
(b)
Figure 11.5 Binary liquid mixtures at 1 bar. The curves are calculated from the twoparameter RedlichKister series using the following parameter values.
Curve 1: a D b D 0 (ideal liquid mixture).
Curve 2: a=RT D 1:8, b=RT D 0:36.
Curve 3: a=RT D 2:4, b=RT D 0:48.
(a) Molar Gibbs energy of mixing as a function of composition.
(b) Activity of component A (using a pure-liquid standard state) as a function of composition.
E
If the properties of the mixture are such that Gm is positive at each mixture composition
(except at the extremes xA D0 and xA D1 where it must be zero), and no portion of the
curve of Gm (mix) versus xA is concave downward, there can be no phase separation and
the activity aA increases monotonically with xA . This case is illustrated by curve 2 in Figs.
11.5(a) and 11.5(b).
If a portion of the Gm (mix)xA curve is concave downward, the condition needed
for phase separation, then a maximum appears in the curve of aA versus xA . This case
is illustrated by curve 3, and the compositions of the coexisting phases are indicated by
open circles. The difference of the compositions at the two circles is a miscibility gap. The
portion of curve 3 between these compositions in Fig. 11.5(b) is dashed to indicate it describes unstable, nonequilibrium states. Although the two coexisting phases have different
compositions, the activity aA is the same in both phases, as indicated in Fig. 11.5(b) by the
horizontal dashed line. This is because component A has the same standard state and the
same chemical potential in both phases.
Coexisting liquid phases will be discussed further in Secs. 12.6 and 13.2.3.
11.2 THE ADVANCEMENT AND MOLAR REACTION QUANTITIES

Many of the processes of interest to chemists can be described by balanced reaction equations, or chemical equations, for the conversion of reactants into products. Thus, for the

314
vaporization of water we write

H2 O.l/ ! H2 O.g/
For the dissolution of sodium chloride in water, we write
NaCl.s/ ! NaC .aq/ C Cl .aq/
For the Haber synthesis of ammonia, the reaction equation can be written
N2 .g/ C 3 H2 .g/ ! 2 NH3 .g/
The essential feature of a reaction equation is that equal amounts of each element and
equal net charges appear on both sides; the equation is said to be balanced. Thus, matter
and charge are conserved during the process, and the process can take place in a closed
system. The species to the left of a single arrow are called reactants, the species to the right
are called products, and the arrow indicates the forward direction of the process.
A reaction equation is sometimes written with right and left arrows
N2 .g/ C 3 H2 .g/ 2 NH3 .g/
to indicate that the process is at reaction equilibrium. It can also be written as a stoichiometric equation with an equal sign:
N2 .g/ C 3 H2 .g/ D 2 NH3 .g/
A reaction equation shows stoichiometric relations among the reactants and products. It
is important to keep in mind that it species neither the initial and nal states of a chemical
process, nor the change in the amount of a reactant or product during the process. For
example, the reaction equation N2 + 3 H2 ! 2 NH3 does not imply that the system initially
contains only N2 and H2 , or that only NH3 is present in the nal state; and it does not mean
that the process consists of the conversion of exactly one mole of N2 and three moles of H2
to two moles of NH3 (although this is a possibility). Instead, the reaction equation tells us
that a change in the amount of N2 is accompanied by three times this change in the amount
of H2 and by twice this change, with the opposite sign, in the amount of NH3 .
11.2.1 An example: ammonia synthesis

It is convenient to indicate the progress of a chemical process with a variable called the
advancement. The reaction equation N2 + 3 H2 ! 2 NH3 for the synthesis of ammonia synthesis will serve to illustrate this concept. Let the system be a gaseous mixture of N2 , H2 ,
and NH3 .
If the system is open and the intensive properties remain uniform throughout the gas
mixture, there are ve independent variables. We can choose them to be T , p, and the
amounts of the three substances. We can write the total differential of the enthalpy, for
instance, as
@H
@H
dH D
dT C
dp
@T p;fnig
@p T;fnig
C HN2 dnN2 C HH2 dnH2 C HNH3 dnNH3
(11.2.1)

315
The notation fnig stands for the set of amounts of all substances in the mixture, and the
quantities HN2 , HH2 , and HNH3 are partial molar enthalpies. For example, HN2 is dened
by
!
@H
HN2 D
(11.2.2)
@nN2
T;p;nH2 ;nNH3
If the system is closed, the amounts of the three substances can still change because of
the reaction N2 + 3 H2 ! 2 NH3 , and the number of independent variables is reduced from
ve to three. We can choose them to be T , p, and a variable called advancement.
The advancement (or extent of reaction), , is the amount by which the reaction dened by the reaction equation has advanced in the forward direction from specied initial
conditions. The quantity has dimensions of amount of substance, the usual unit being the
mole.
Let the initial amounts be nN2 ;0 , nH2 ;0 , and nNH3 ;0 . Then at any stage of the reaction
process in the closed system, the amounts are given by
nN2 D nN2 ;0
nH2 D nH2 ;0
nNH3 D nNH3 ;0 C 2
(11.2.3)
These relations come from the stoichiometry of the reaction as expressed by the stoichiometric coefcients in the reaction equation. The second relation, for example, expresses
the fact that when one mole of reaction has occurred ( D 1 mol), the amount of H2 in the
closed system has decreased by three moles.
Taking the differentials of Eqs. 11.2.3, we nd that innitesimal changes in the amounts
are related to the change of as follows:
dnN2 D
dnH2 D
3d
dnNH3 D 2 d
(11.2.4)
These relations show that in a closed system, the changes in the various amounts are not
independent. Substitution in Eq. 11.2.1 of the expressions for dnN2 , dnH2 , and dnNH3 gives
@H
@H
dH D
dT C
dp
@T p;
@p T;
C
HN 2
3HH2 C 2HNH3 d
(11.2.5)
(closed system)
(The subscript fnig on the partial derivatives has been replaced by to indicate the same
thing: that the derivative is taken with the amount of each species held constant.)
Equation 11.2.5 gives an expression for the total differential of the enthalpy with T , p,
and as the independent variables. The coefcient of d in this equation is called the molar
reaction enthalpy, or molar enthalpy of reaction, r H :
r H D
HN2
3HH2 C 2HNH3
We identify this coefcient as the partial derivative
@H
r H D
@ T;p
(11.2.6)
(11.2.7)
That is, the molar reaction enthalpy is the rate at which the enthalpy changes with the
advancement as the reaction proceeds in the forward direction at constant T and p.

316
The partial molar enthalpy of a species is the enthalpy change per amount of the species
added to an open system. To see why the particular combination of partial molar
enthalpies on the right side of Eq. 11.2.6 is the rate at which enthalpy changes with
advancement in the closed system, we can imagine the following process at constant
T and p: An innitesimal amount dn of N2 is removed from an open system, three
times this amount of H2 is removed from the same system, and twice this amount of
NH3 is added to the system. The total enthalpy change in the open system is dH D
. HN2 3HH2 C 2HNH3 / dn. The net change in the state of the system is equivalent to
an advancement d D dn in a closed system, so dH= d in the closed system is equal
to . HN2 3HH2 C 2HNH3 / in agreement with Eqs. 11.2.6 and 11.2.7.
Note that because the advancement is dened by how we write the reaction equation,
the value of r H also depends on the reaction equation. For instance, if we change the
reaction equation for ammonia synthesis from N2 + 3 H2 ! 2 NH3 to
1
2 N2
3
C 2 H2 ! NH3
then the value of r H is halved.
11.2.2 Molar reaction quantities in general

Now let us generalize the relations of the preceding section for any chemical process in a
closed system. Suppose the stoichiometric equation has the form
aA C bB D d D C eE
(11.2.8)
where A and B are reactant species, D and E are product species, and a, b, d , and e are the
corresponding stoichiometric coefcients. We can rearrange this equation to
0D
aA
bB C d D C eE
In general, the stoichiometric relation for any chemical process is

X
0D
i Ai
(11.2.9)
(11.2.10)
where i is the stoichiometric number of species Ai , a dimensionless quantity taken as

negative for a reactant and positive for a product. In the ammonia synthesis example of
the previous section, the stoichiometric relation is 0 D N2 3H2 C 2NH3 and the stoichiometric numbers are N2 D 1, H2 D 3, and NH3 D C2. In other words, each
stoichiometric number is the same as the stoichiometric coefcient in the reaction equation,
except that the sign is negative for a reactant.
The amount of reactant or product species i present in the closed system at any instant
depends on the advancement at that instant, and is given by
ni D ni;0 C
(11.2.11)
(closed system)

317
The innitesimal change in the amount due to an innitesimal change in the advancement
is
dni D
(11.2.12)
(closed system)
In an open system, the total differential of extensive property X is
X
@X
@X
dX D
dT C
dp C
Xi dni
@T p;fnig
@p T;fnig
(11.2.13)
where Xi is a partial molar quantity. We restrict the system to a closed one with T , p, and
as the independent variables. Then, with the substitution dni D i d from Eq. 11.2.12,
the total differential of X becomes
@X
@X
dX D
dT C
dp C r X d
(11.2.14)
@T p;
@p T;
(closed system)
where the coefcient r X is the molar reaction quantity dened by

def
r X D
i Xi
(11.2.15)
Equation 11.2.14 allows us to identify the molar reaction quantity as a partial derivative:
@X
r X D
(11.2.16)
@ T;p
(closed system)
It is important to observe the distinction between the notations X, the nite change
of X during a process, and r X , a differential quantity that is a property of the system in
a given state. The fact that both notations use the symbol can be confusing. Equation
11.2.16 shows that we can think of r as an operator.
In dealing with the change of an extensive property X as changes, we must distinguish
between molar integral and molar differential reaction quantities.
X= is a molar integral reaction quantity, the ratio of two nite differences between the nal and initial states of a process. These states are assumed to have
the same temperature and the same pressure. This book will use a notation such
as Hm (rxn) for a molar integral reaction enthalpy:
Hm (rxn) D
H (rxn)
H. 2 /
D
H. 1 /
1
(11.2.17)
(T2 DT1 ; p2 Dp1 )
r X is a molar differential reaction quantity. Equation 11.2.16 shows that r X is the

rate at which the extensive property X changes with the advancement in a closed system at constant T and p. The value of r X is in general a function of the independent
variables T , p, and .
318

mol
(a)
mol
(b)
Figure 11.6 Enthalpy and entropy as functions of advancement at constant T and p.

The curves are for a reaction A!2B with positive r H taking place in an ideal gas
mixture with initial amounts nA;0 D 1 mol and nB;0 D 0.
The notation for a molar differential reaction quantity such as r H includes a subscript
following the symbol to indicate the kind of chemical process. The subscript r denotes
a reaction or process in general. The meanings of vap, sub, fus, and trs were described in Sec. 8.3.1. Subscripts for specic kinds of reactions and processes are listed in
Sec. D.2 of Appendix D and are illustrated in sections to follow.
For certain kinds of processes, it may happen that a partial molar quantity Xi remains
constant for each species i as the process advances at constant T and p. If Xi remains
constant for each i, then according to Eq. 11.2.15 the value of r X must also remain constant as the process advances. Since r X is the rate at which X changes with , in such a
situation X is a linear function of . This means that the molar integral reaction quantity
Xm (rxn) dened by X= is equal, for any nite change of , to r X.
An example is the partial molar enthalpy Hi of a constituent of an ideal gas mixture,
an ideal condensed-phase mixture, or an ideal-dilute solution. In these ideal mixtures, Hi
is independent of composition at constant T and p (Secs. 9.3.3, 9.4.3, and 9.4.7). When
a reaction takes place at constant T and p in one of these mixtures, the molar differential
reaction enthalpy r H is constant during the process, H is a linear function of , and r H
and Hm (rxn) are equal. Figure 11.6(a) illustrates this linear dependence for a reaction in
an ideal gas mixture.
In contrast, Fig. 11.6(b) shows the nonlinearity of the entropy as a function of during
the same reaction. The nonlinearity is a consequence of the dependence of the partial molar
entropy Si on the mixture composition (Eq. 11.1.24). In the gure, the slope of the curve
at each value of equals r S at that point; its value changes as the reaction advances and
the composition of the reaction mixture changes. Consequently, the molar integral reaction
entropy Sm (rxn) D S(rxn)= approaches the value of r S only in the limit as
approaches zero.

11.3 M OLAR R EACTION E NTHALPY
319
11.2.3 Standard molar reaction quantities

If a chemical process takes place at constant temperature while each reactant and product
remains in its standard state of unit activity, the molar reaction quantity r X is called the
standard molar reaction quantity and is denoted by r X . For instance, vap H is a
standard molar enthalpy of vaporization (already discussed in Sec. 8.3.3), and r G is the
standard molar Gibbs energy of a reaction.
From Eq. 11.2.15, the relation between a standard molar reaction quantity and the standard molar quantities of the reactants and products at the same temperature is
X
def
r X D
(11.2.18)
i Xi
i
Two comments are in order.

1. Whereas a molar reaction quantity is usually a function of T , p, and , a standard
molar reaction quantity is a function only of T . This is evident because standard-state
conditions imply that each reactant and product is in a separate phase of constant
dened composition and constant pressure p .
2. Since the value of a standard molar reaction quantity is independent of , the standard
molar integral and differential quantities are identical (page 318):
Xm (rxn) D r X
(11.2.19)
These general concepts will now be applied to some specic chemical processes.
11.3 MOLAR REACTION ENTHALPY

Recall that Hm (rxn) is a molar integral reaction enthalpy equal to H (rxn)= , and that
P
r H is a molar differential reaction enthalpy dened by i i Hi and equal to .@H=@ /T;p .
11.3.1 Molar reaction enthalpy and heat

During a process in a closed system at constant pressure with expansion work only, the
enthalpy change equals the energy transferred across the boundary in the form of heat:
dH D q (Eq. 5.3.7). Thus for the molar reaction enthalpy r H D .@H=@ /T;p , which
refers to a process not just at constant pressure but also at constant temperature, we can
write
r H D
q
d
(11.3.1)
(constant T and p, w 0 D0)
Note that when there is nonexpansion work (w 0 ), such as electrical work, the enthalpy
change is not equal to the heat. For example, if we compare a reaction taking place in a
galvanic cell with the same reaction in a reaction vessel, the heats at constant T and p for
a given change of are different, and may even have opposite signs. The value of r H is
the same in both systems, but the ratio of heat to advancement, q= d , is different.
An exothermic reaction is one for which r H is negative, and an endothermic reaction
is one for which r H is positive. Thus in a reaction at constant temperature and pressure

320
with expansion work only, heat is transferred out of the system during an exothermic process
and into the system during an endothermic process. If the process takes place at constant
pressure in a system with thermally-insulated walls, the temperature increases during an
exothermic process and decreases during an endothermic process.
These comments apply not just to chemical reactions, but to the other chemical processes at constant temperature and pressure discussed in this chapter.
11.3.2 Standard molar enthalpies of reaction and formation

A standard molar reaction enthalpy, r H , is the same as the molar integral reaction enthalpy Hm (rxn) for the reaction taking place under standard state conditions (each reactant
and product at unit activity) at constant temperature (page 319).
At constant temperature, partial molar enthalpies depend only mildly on pressure. It is
therefore usually safe to assume that unless the experimental pressure is much greater than
p , the reaction is exothermic if r H is negative and endothermic if r H is positive.
The formation reaction of a substance is the reaction in which the substance, at a given
temperature and in a given physical state, is formed from the constituent elements in their
reference states at the same temperature. The reference state of an element is usually chosen
to be the standard state of the element in the allotropic form and physical state that is stable
at the given temperature and the standard pressure. For instance, at 298:15 K and 1 bar the
stable allotrope of carbon is crystalline graphite rather than diamond.
Phosphorus is an exception to the rule regarding reference states of elements. Although
red phosphorus is the stable allotrope at 298:15 K, it is not well characterized. Instead, the
reference state is white phosphorus (crystalline P4 ) at 1 bar.
At 298:15 K, the reference states of the elements are the following:
For H2 , N2 , O2 , F2 , Cl2 , and the noble gases, the reference state is the ideal gas at
1 bar.
For Br2 and Hg, the reference state is the liquid at 1 bar.
For P, as mentioned above, the reference state is crystalline white phosphorus at 1 bar.
For all other elements, the reference state is the stable crystalline allotrope at 1 bar.
The standard molar enthalpy of formation (or standard molar heat of formation),
f H , of a substance is the enthalpy change per amount of substance produced in the
formation reaction of the substance in its standard state. Thus, the standard molar enthalpy
of formation of gaseous methyl bromide at 298:15 K is the molar reaction enthalpy of the
reaction
C(s, graphite, p ) C 3 H2 (ideal gas, p ) C 1 Br2 (l, p ) ! CH3 Br(ideal gas, p )
2
2
The value of f H for a given substance depends only on T . By denition, f H for the
reference state of an element is zero.
A principle called Hesss law can be used to calculate the standard molar enthalpy of
formation of a substance at a given temperature from standard molar reaction enthalpies at
the same temperature, and to calculate a standard molar reaction enthalpy from tabulated
values of standard molar enthalpies of formation. The principle is an application of the
fact that enthalpy is a state function. Therefore, H for a given change of the state of the

321
system is independent of the path and is equal to the sum of H values for any sequence
of changes whose net result is the given change. (We may apply the same principle to a
change of any state function.)
For example, the following combustion reactions can be carried out experimentally in a
bomb calorimeter (Sec. 11.5.2), yielding the values shown below of standard molar reaction
enthalpies (at T D 298:15 K, p D p D 1 bar):
r H D
C(s, graphite) C O2 (g) ! CO2 .g/
r H D
CO(g) C 1 O2 (g) ! CO2 .g/

2
393:51 kJ mol
282:98 kJ mol
(Note that the rst reaction, in addition to being the combustion reaction of graphite, is
also the formation reaction of carbon dioxide.) The change resulting from the rst reaction
followed by the reverse of the second reaction is the formation reaction of carbon monoxide:
C(s, graphite) C 1 O2 (g) ! CO(g)
2
It would not be practical to measure the molar enthalpy of this last reaction by allowing
graphite to react with oxygen in a calorimeter, because it would be difcult to prevent the
formation of some CO2 . From Hesss law, the standard molar enthalpy of formation of CO
is the sum of the standard molar enthalpies of the reactions that have the formation reaction
as the net result:
f H (CO, g, 298:15 K) D . 393:51 C 282:98/ kJ mol
D
110:53 kJ mol
(11.3.2)
This value is one of the many standard molar enthalpies of formation to be found in
compilations of thermodynamic properties of individual substances, such as the table in
Appendix H. We may use the tabulated values to evaluate the standard molar reaction enthalpy r H of a reaction using a formula based on Hesss law. Imagine the reaction to take
place in two steps: First each reactant in its standard state changes to the constituent elements in their reference states (the reverse of a formation reaction), and then these elements
form the products in their standard states. The resulting formula is
X
(11.3.3)
r H D
i f H .i/
(Hesss law)
where f H .i/ is the standard molar enthalpy of formation of substance i. Recall that the
stoichiometric number i of each reactant is negative and that of each product is positive, so
according to Hesss law the standard molar reaction enthalpy is the sum of the standard molar enthalpies of formation of the products minus the sum of the standard molar enthalpies
of formation of the reactants. Each term is multiplied by the appropriate stoichiometric
coefcient from the reaction equation.
A standard molar enthalpy of formation can be dened for a solute in solution to use in
Eq. 11.3.3. For instance, the formation reaction of aqueous sucrose is
12 C(s, graphite) C 11 H2 (g) C
11
2 O2 (g)
! C12 H22 O11 (aq)
and f H for C12 H22 O11 (aq) is the enthalpy change per amount of sucrose formed when
the reactants and product are in their standard states. Note that this formation reaction does

322
BIOGRAPHICAL SKETCH
Germain Henri Hess (18021850)

seven editions. He became a full member of
the St. Petersburg Academy Academy in 1834.
Hess published the results of his thermochemical research between 1839 and 1842.
His 1840 paperb describes his measurements
of the heat evolved when pure sulfuric acid,
H2 SO4 , is mixed with various amounts of water, and another series of measurements of the
heat evolved when the acid in H2 SO4 -water
mixtures is neutralized with aqueous ammonia. The following table from this paper is of
Hess was a Russian chemist and physician historical interest, although it is a bit difcult
whose calorimetric measurements led him to to decipher:
formulate the law of constant heat summation,
Acid
Heat evolved by
Sum
ammonia water
now known as Hesss law. His given name had
PS
H
595:8
595:8
several versions: Germain Henri in French,
P2S
as shown above; Hermann Heinrich in GerH
518:9
77:8 596:7
P3S
man; and German Iwanowitsch in Russian.
H
480:5 116:7 597:2
He was born in Geneva, Switzerland, the
P6S
H
446:2 155:6 601:8
son of an artist. The family moved to RusAverage 597:9
sia when he was three years old; his father The rst column gives the relative amounts of
had found work there as a tutor on an es- acid and water in the notation of Berzelius: H
P
tate. Hess studied medicine at the University is H O, is SO , and H is H SO . Thus
PS
S
2
3
2
4
of Tartu in Estonia (then part of the Russian P 3
H S, for example, is H2 SO4 C2H2 O. The secempire) and received his doctor of medicine
ond and third columns show the heat evolved
degree in 1825. In addition to his medical stud( q) in units of calories per gram of the SO3
ies, Hess took courses in chemistry and geolmoiety of the H2 SO4 . The near-equality of the
ogy and wrote a doctoral dissertation on the
sums in the last column for the overall reaction
a
composition of mineral waters in Russia.
H2 SO4 .l/ C 2 NH3 .aq/ ! .NH4 /2 SO4 .aq/
Hess was more interested in chemistry than
demonstrates Hesss law of constant heat summedicine. He briey studied with the famation, which he stated in the words:c
mous Swedish chemist J ns Jakob Berzelius
o
The amount of heat evolved during the formation
in Stockholm, and they became life-long
of a given compound is constant, independent of
friends. Hess then practiced medicine in
whether the compound is formed directly or inIrkutsk, Siberia, while continuing his interests
directly in a series of steps.
in mineral chemistry and chemical analysis.
Hess conrmed this law with similar experIn 1829, after being being elected an adjunct
iments using other acids such as HCl(aq) and
member of the St. Petersburg Academy of Sciences, he gave up his medical practice and be- other bases such as NaOH(aq) and CaO(s).
After 1848 Hesss health began to deteriogan teaching chemistry at various institutions
rate. He was only 48 when he died. His role
of higher learning in St. Petersburg. He wrote
a two-volume textbook, Fundamentals of Pure as the real founder of thermochemistry was
Chemistry, in 1831, followed by a one-volume largely forgotten until the inuential German
abridgment in 1834 that became the standard physical chemist Wilhelm Ostwald drew attenRussian chemistry textbook and went through tion to it about forty years after Hesss death.
a Refs.
[93] and [100].
b Ref.
[77].
c Ref.
[77]; translation in Ref. [41].

323
not include the formation of the solvent H2 O from H2 and O2 . Instead, the solute once
formed combines with the amount of pure liquid water needed to form the solution. If the
aqueous solute is formed in its standard state, the amount of water needed is very large so
as to have the solute exhibit innite-dilution behavior.
There is no ordinary reaction that would produce an individual ion in solution from its
element or elements without producing other species as well. We can, however, prepare
a consistent set of standard molar enthalpies of formation of ions by assigning a value to
a single reference ion.7 We can use these values for ions in Eq. 11.3.3 just like values of
f H for substances and nonionic solutes. Aqueous hydrogen ion is the usual reference
ion, to which is assigned the arbitrary value
f H (HC , aq) D 0
(at all temperatures)
(11.3.4)
To see how we can use this reference value, consider the reaction for the formation of
aqueous HCl (hydrochloric acid):
1
2 H2 .g/
C 1 Cl2 .g/ ! HC .aq/ C Cl .aq/

2
The standard molar reaction enthalpy at 298:15 K for this reaction is known, from reaction
calorimetry, to have the value r H D 167:08 kJ mol 1 . The standard states of the
gaseous H2 and Cl2 are, of course, the pure gases acting ideally at pressure p , and the
standard state of each of the aqueous ions is the ion at the standard molality and standard
pressure, acting as if its activity coefcient on a molality basis were 1. From Eq. 11.3.3, we
equate the value of r H to the sum
1
2 f H (H2 , g)
1
2 f H (Cl2 , g)
C f H (HC , aq) C f H (Cl , aq)
But the rst three terms of this sum are zero. Therefore, the value of f H (Cl , aq) is
167:08 kJ mol 1 .
Next we can combine this value of f H (Cl , aq) with the measured standard molar
enthalpy of formation of aqueous sodium chloride
Na.s/ C 1 Cl2 .g/ ! NaC .aq/ C Cl .aq/
2
to evaluate the standard molar enthalpy of formation of aqueous sodium ion. By continuing
this procedure with other reactions, we can build up a consistent set of f H values of
various ions in aqueous solution.
11.3.3 Molar reaction heat capacity

The molar reaction enthalpy r H is in general a function of T , p, and . Using the relations
P
r H D i i Hi (from Eq. 11.2.15) and Cp;i D .@Hi =@T /p; (Eq. 9.2.52), we can write
X
@r H
@ i i Hi
D
D
(11.3.5)
i Cp;i D r Cp
@T
@T
p;
p;
i
where r Cp is the molar reaction heat capacity at constant pressure, equal to the rate at
which the heat capacity Cp changes with at constant T and p.
Under standard state conditions, Eq. 11.3.5 becomes
dr H = dT D r Cp
7 This
(11.3.6)
procedure is similar to that described on page 235 for partial molar volumes of ions.

324
(rxn, )
(rxn, )
Figure 11.7 Dependence of reaction enthalpy on temperature at constant pressure.
11.3.4 Effect of temperature on reaction enthalpy

Consider a reaction occurring with a certain nite change of the advancement in a closed
system at temperature T 0 and at constant pressure. The reaction is characterized by a change
of the advancement from 1 to 2 , and the integral reaction enthalpy at this temperature is
denoted H (rxn, T 0 ). We wish to nd an expression for the reaction enthalpy H (rxn, T 00 )
for the same values of 1 and 2 at the same pressure but at a different temperature, T 00 .
The heat capacity of the system at constant pressure is related to the enthalpy by Eq. 5.6.3
on page 144: Cp D .@H=@T /p; . We integrate dH D Cp dT from T 0 to T 00 at constant p
and , for both the nal and initial values of the advancement:
00
00
H. 2 ; T / D H. 2 ; T / C
H. 1 ; T / D H. 1 ; T / C
T 00
T0
Cp . 2 / dT
(11.3.7)
T 00
T0
Cp . 1 / dT
(11.3.8)
(11.3.9)
Subtracting Eq. 11.3.8 from Eq. 11.3.7, we obtain

H (rxn, T 00 ) D H (rxn, T 0 ) C
T 00
T0
Cp dT
where Cp is the difference between the heat capacities of the system at the nal and initial
values of , a function of T : Cp D Cp . 2 / Cp . 1 /. Equation 11.3.9 is the Kirchhoff
equation.
When Cp is essentially constant in the temperature range from T 0 to T 00 , the Kirchhoff
equation becomes
H (rxn, T 00 ) D H (rxn, T 0 ) C Cp .T 00
T 0/
(11.3.10)
Figure 11.7 illustrates the principle of the Kirchhoff equation as expressed by Eq.
11.3.10. Cp equals the difference in the slopes of the two dashed lines in the gure, and
the product of Cp and the temperature difference T 00 T 0 equals the change in the value
of H (rxn). The gure illustrates an exothermic reaction with negative Cp , resulting in
a more negative value of H (rxn) at the higher temperature.

11.4 E NTHALPIES OF S OLUTION AND D ILUTION
sol
B(s)
325
dil
(A + B)(l)
A(l)
(a)
(A + B)(l)
(b)
Figure 11.8 Two related processes in closed systems. A: solvent; B: solute. The
dashed rectangles represent the system boundaries.
(a) Solution process.
(b) Dilution process.
We can also nd the effect of temperature on the molar differential reaction enthalpy
r H . From Eq. 11.3.5, we have .@r H=@T /p; D r Cp . Integration from temperature
T 0 to temperature T 00 yields the relation
00
r H.T ; / D r H.T ; / C
T 00
T0
r Cp .T; / dT
(11.3.11)
This relation is analogous to Eq. 11.3.9, using molar differential reaction quantities in place
of integral reaction quantities.
11.4 ENTHALPIES OF SOLUTION AND DILUTION

The processes of solution (dissolution) and dilution are related. The IUPAC Green Book8
recommends the abbreviations sol and dil for these processes.
During a solution process, a solute is transferred from a pure solute phase (solid, liquid,
or gas) to a solvent or solution phase. During a dilution process, solvent is transferred
from a pure solvent phase to a solution phase. We may specify the advancement of these
two kinds of processes by sol and dil , respectively. Note that both processes take place
in closed systems that (at least initially) have two phases. The total amounts of solvent
and solute in the systems do not change, but the amounts in pure phases diminish as the
processes advance and sol or dil increases (Fig. 11.8).
The equations in this section are about enthalpies of solution and dilution, but you
can replace H by any other extensive state function to obtain relations for its solution and
dilution properties.
11.4.1 Molar enthalpy of solution

First let us consider a solution process in which solute is transferred from a pure solute
phase to a solution. The molar differential enthalpy of solution, sol H , is the rate of
change of H with the advancement sol at constant T and p, where sol is the amount of
8 Ref.
[36], Sec. 2.11.1.

solute transferred:
sol H D
@H
@ sol
326
(11.4.1)
T;p;nA
The value of sol H at a given T and p depends only on the solution molality and not on
the amount of solution.
P When we write the solution reaction as B ! B(sln), the general relation r X D
i i Xi (Eq. 11.2.15) becomes
sol H D HB
HB
(11.4.2)
where HB is the partial molar enthalpy of the solute in the solution and HB is the molar
enthalpy of the pure solute at the same T and p.
The molar enthalpy of solution at innite dilution, sol H 1 , is the rate of change
of H with sol when the solute is transferred to a solution with the thermal properties of
an innitely dilute solution. We can think of sol H 1 as the enthalpy change per amount
of solute transferred to a very large volume of pure solvent. According to Eq. 11.4.2, this
quantity is given by
1
sol H 1 D HB
HB
(11.4.3)
1
Note that because the values of HB and HB are independent of the solution composition,
the molar differential and integral enthalpies of solution at innite dilution are the same.
An integral enthalpy of solution, H(sol), is the enthalpy change for a process in
which a nite amount sol of solute is transferred from a pure solute phase to a specied
amount of pure solvent to form a homogeneous solution phase with the same temperature
and pressure as the initial state. Division by the amount transferred gives the molar integral enthalpy of solution which this book will denote by Hm (sol, mB ), where mB is the
molality of the solution formed:
Hm (sol, mB ) D
H(sol)
(11.4.4)
sol
An integral enthalpy of solution can be evaluated by carrying out the solution process in
a constant-pressure reaction calorimeter, as will be described in Sec. 11.5.1. Experimental
values of H(sol) as a function of sol can be collected by measuring enthalpy changes
during a series of successive additions of the solute to a xed amount of solvent, resulting
in a solution whose molality increases in stages. The enthalpy changes are cumulative, so
the value of H(sol) after each addition is the sum of the enthalpy changes for this and the
previous additions.
The relations between H(sol) and the molar integral and differential enthalpies of
solution are illustrated in Fig. 11.9 on the next page with data for the solution of crystalline
sodium acetate in water. The curve shows H(sol) as a function of sol , with sol dened as
the amount of solute dissolved in one kilogram of water. Thus at any point along the curve,
the molality is mB D sol =.1 kg/ and the ratio H(sol)= sol is the molar integral enthalpy
of solution Hm (sol, mB ) for the solution process that produces solution of this molality.
The slope of the curve is the molar differential enthalpy of solution:
sol H D
dH(sol)
d sol
(11.4.5)
(constant T , p, and nA )

327
sol kJ
sol
mol
Figure 11.9 Enthalpy change for the dissolution of NaCH3 CO2 (s) in one kilogram
of water in a closed system at 298:15 K and 1 bar, as a function of the amount sol of
dissolved solute. a The open circle at sol D15 mol indicates the approximate saturation limit; data to the right of this point come from supersaturated solutions. At the
composition mB D15 mol kg 1 , the value of Hm (sol, mB ) is the slope of line a and
the value of sol H is the slope of line b. The value of sol H 1 is the slope of line c.
a Data
from Ref. [163], page 2-315.
The slope of the curve at sol D0 is sol H 1 , the molar enthalpy of solution at innite
dilution. If the measurements are made at the standard pressure, sol H 1 is the same as
the standard molar enthalpy of solution, sol H , because the standard molar enthalpy of a
solute is the molar enthalpy at pDp and innite dilution.
11.4.2 Enthalpy of dilution

Next let us consider a dilution process in which solvent is transferred from a pure solvent
phase to a solution phase. The molar differential enthalpy of dilution is the rate of change
of H with the advancement dil at constant T and p of the dilution process, where dil is the
amount of solvent transferred:
@H
dil H D
(11.4.6)
@ dil T;p;nB
For the dilution reaction A ! A(sln), the general relation r X D
dil H D HA
HA
i i Xi
becomes
(11.4.7)
where HA is the partial molar enthalpy of the solvent in the solution. In the limit of innite
dilution, HA must approach the molar enthalpy of pure solvent, HA ; then Eq. 11.4.7 shows
that dil H approaches zero in this limit.
An integral enthalpy of dilution, H (dil), refers to the enthalpy change for transfer
of a nite amount of solvent from a pure solvent phase to a solution, T and p being the

328
same before and after the process. The molar integral enthalpy of dilution is the ratio of
H (dil) and the amount of solute in the solution. For a dilution process at constant solute
amount nB in which the molality changes from m0 to m00 , this book will use the notation
B
B
Hm .dil, m0 !m00 /:
B
B
H (dil)
Hm .dil, m0 !m00 / D
(11.4.8)
B
B
nB
The value of Hm .dil, m0 !m00 / at a given T and p depends only on the initial and nal
B
B
molalities m0 and m00 .
B
B
There is a simple relation between molar integral enthalpies of solution and dilution,
as the following derivation demonstrates. Consider the following two ways of preparing a
solution of molality m00 from pure solvent and solute phases. Both paths are at constant T
B
and p in a closed system.
Path 1: The solution forms directly by dissolution of the solute in the solvent. The enthalpy change is nB Hm .sol, m00 /, where the molality of the solution is indicated in
B
parentheses.
Path 2: Starting with the unmixed solvent and solute, the solute dissolves in a portion of
the solvent to form a solution of composition m0 (more concentrated than m00 ). The
B
B
enthalpy change is nB Hm .sol, m0 /. In a second step of this path, the remaining pure
B
solvent mixes with the solution to dilute it from m0 to m00 . The enthalpy change of the
B
B
second step is nB Hm .dil, m0 !m00 /.
B
B
Since both paths have the same initial states and the same nal states, both have the same
overall enthalpy change:
nB Hm .sol, m00 / D nB Hm .sol, m0 / C nB Hm .dil, m0 !m00 /
B
B
B
B
(11.4.9)
Hm .sol, m00 / D Hm .sol, m0 / C Hm .dil, m0 !m00 /

B
B
B
B
(11.4.10)
or
Equation 11.4.10 is the desired relation. It shows how a measurement of the molar integral
enthalpy change for a solution process that produces solution of a certain molality can
be combined with dilution measurements in order to calculate molar integral enthalpies of
solution for more dilute solutions. Experimentally, it is sometimes more convenient to carry
out the dilution process than the solution process, especially when the pure solute is a gas
or solid.
11.4.3 Molar enthalpies of solute formation

Molar integral enthalpies of solution and dilution are conveniently expressed in terms of
molar enthalpies of formation. The molar enthalpy of formation of a solute in solution is
the enthalpy change per amount of solute for a process at constant T and p in which the
solute, in a solution of a given molality, is formed from its constituent elements in their
reference states. The molar enthalpy of formation of solute B in solution of molality mB
will be denoted by f H (B, mB ).
As explained in Sec. 11.3.2, the formation reaction of a solute in solution does not
include the formation of the solvent from its elements. For example, the formation reaction
for NaOH in an aqueous solution that has 50 moles of water for each mole of NaOH is
Na(s) C 1 O2 (g) C 1 H2 (g) C 50 H2 O(l) ! NaOH in 50 H2 O
2
2

329
Consider a solution process at constant T and p in which an amount nB of pure solute

(solid, liquid, or gas) is mixed with an amount nA of pure solvent, resulting in solution of
molality mB . We may equate the enthalpy change of this process to the sum of the enthalpy
changes for the following two hypothetical steps:
1. An amount nB of the pure solute decomposes to the constituent elements in their
reference states. This is the reverse of the formation reaction of the pure solute.
2. The solution is formed from these elements and an amount nA of the solvent.
The total enthalpy change is then H(sol) D nB f H (B ) C nB f H (B, mB ). Dividing
by nB , we obtain the molar integral enthalpy of solution:
Hm (sol, mB ) D f H (B, mB )
f H (B )
(11.4.11)
By combining Eqs. 11.4.10 and 11.4.11, we obtain the following expression for a molar
integral enthalpy of dilution in terms of molar enthalpies of formation:
Hm .dil, m0 !m00 / D f H (B, m00 )
B
B
B
f H (B, m0 )
B
(11.4.12)
From tabulated values of molar enthalpies of formation, we can calculate molar integral
enthalpies of solution with Eq. 11.4.11 and molar integral enthalpies of dilution with Eq.
11.4.12. Conversely, calorimetric measurements of these molar integral enthalpies can be
combined with the value of f H (B ) to establish the values of molar enthalpies of solute
formation in solutions of various molalities.
11.4.4 Evaluation of relative partial molar enthalpies

Although it is not possible to determine absolute values of partial molar enthalpies, we can
evaluate HA and HB relative to appropriate solvent and solute reference states.
The relative partial molar enthalpy of the solvent is dened by
def
LA D HA
HA
(11.4.13)
This is the partial molar enthalpy of the solvent in a solution of given composition relative
to pure solvent at the same temperature and pressure.
LA can be related to molar differential and integral enthalpies of solution as follows.
The enthalpy change to form a solution from amounts nA and nB of pure solvent and solute
is given, from the additivity rule, by H(sol) D .nA HA C nB HB / .nA HA C nB HB /. We
rearrange and make substitutions from Eqs. 11.4.2 and 11.4.13:
H(sol) D nA .HA
HA / C nB .HB
D nA LA C nB sol H
HB /
(11.4.14)
H(sol) is also given, from Eq. 11.4.4, by

H(sol) D nB Hm (sol, mB )
(11.4.15)
Equating both expressions for H(sol), solving for LA , and replacing nB =nA by MA mB ,
we obtain
LA D MA mB Hm (sol, mB ) sol H
(11.4.16)

330
Thus LA depends on the difference between the molar integral and differential enthalpies
of solution.
The relative partial molar enthalpy of a solute is dened by
def
LB D HB
1
HB
(11.4.17)
The reference state for the solute is the solute at innite dilution. To relate LB to molar
enthalpies of solution, we write the identity
LB D HB
1
HB D .HB
HB /
1
.HB
HB /
(11.4.18)
From Eqs. 11.4.2 and 11.4.3, this becomes

LB D sol H
sol H 1
(11.4.19)
We see that LB is equal to the difference between the molar differential enthalpies of solution at the molality of interest and at innite dilution.
For a solution of a given molality, LA and LB can be evaluated from calorimetric measurements of H(sol) by various methods. Three general methods are as follows.9
LA and LB can be evaluated by the variant of the method of intercepts described on
page 232. The molar integral enthalpy of mixing, Hm (mix) D H(sol)=.nA CnB /,
is plotted versus xB . The tangent to the curve at a given value of xB has intercepts
LA at xB D0 and HB HB D sol H at xB D1, where the values of LA and sol H
are for the solution of composition xB . The tangent to the curve at xB D0 has intercept sol H 1 at xB D1. LB is equal to the difference of these values of sol H and
sol H 1 (Eq. 11.4.19).
Values of H(sol) for a constant amount of solvent can be plotted as a function of
sol , as in Fig. 11.9. The slope of the tangent to the curve at any point on the curve
is equal to sol H for the molality mB at that point, and the initial slope at sol D0 is
equal to sol H 1 . LB at molality mB is equal to the difference of these two values,
and LA can be calculated from Eq. 11.4.16.
A third method for the evaluation of LA and LB is especially useful for solutions
of an electrolyte solute. This method takes advantage of the fact that a plot of
p
p
Hm (sol, mB ) versus mB has a nite limiting slope at mB D0 whose value for
an electrolyte can be predicted from the DebyeH ckel limiting law, providing a useu
ful guide for the extrapolation of Hm (sol, mB ) to its limiting value sol H 1 . The
remainder of this section describes this third method.
The third method assumes we measure the integral enthalpy of solution H(sol) for
varying amounts sol of solute transferred at constant T and p from a pure solute phase to
a xed amount of solvent. From Eq. 11.4.5, the molar differential enthalpy of solution is
given by sol H D dH(sol)= d sol when nA is held constant. We make the substitution
9 The
descriptions refer to graphical plots with smoothed curves drawn through experimental points. A plot can
be replaced by an algebraic function (e.g., a power series) tted to the points, and slopes and intercepts can then
be evaluated by numerical methods.

H(sol) D
sol Hm (sol, mB )
331
and take the derivative of the expression with respect to
sol H D
sol :
sol Hm (sol, mB )
sol
D Hm (sol, mB ) C
sol
dHm (sol, mB )
d sol
(11.4.20)
At constant nA , mB is proportional to sol , so that d sol = sol can be replaced by dmB =mB .
When we combine the resulting expression for sol H with Eq. 11.4.19, we get the following expression for the relative partial molar enthalpy of the solute:
LB D Hm (sol, mB ) C mB
dHm (sol, mB )
dmB
sol H 1
(11.4.21)
It is convenient to dene the quantity

def
L D Hm (sol, mB )
sol H 1
(11.4.22)
known as the relative apparent molar enthalpy of the solute. Because sol H 1 is independent of mB , the derivative dL = dmB is equal to dHm (sol, mB )= dmB . We can therefore
write Eq. 11.4.21 in the compact form
LB D L C mB
dL
dmB
(11.4.23)
(constant T and p)
Equation 11.4.23 allows us to evaluate LB at any molality from the dependence of L on

mB , with L obtained from experimental molar integral enthalpies of solution according to
Eq. 11.4.22.
Once L and LB have been evaluated for a given molality, it is a simple matter to
calculate LA at that molality. By combining Eqs. 11.4.16 and 11.4.22, we obtain the relation
LA D MA mB .L
LB /
(11.4.24)
For an electrolyte solute, a plot of Hm (sol, mB ) versus mB has a limiting slope of C1

p
at mB D0, whereas the limiting slope of Hm (sol, mB ) versus mB is nite and can be
predicted from the DebyeH ckel limiting law. Accordingly, a satisfactory procedure is to
u
p
plot Hm (sol, mB ) versus mB , perform a linear extrapolation of the experimental points
p
to mB D0, and then shift the origin to the extrapolated intercept. The result is a plot of L
p
versus mB . An example for aqueous NaCl solutions is shown in Fig. 11.10(a) on the next
page.
We can also evaluate L from experimental enthalpies of dilution. From Eqs. 11.4.10
and 11.4.22, we obtain the relation
L .m00 /
B
L .m0 / D Hm .dil, m0 !m00 /

B
B
B
(11.4.25)
We can measure the enthalpy changes for diluting a solution of initial molality m0 to various
B
p
molalities m00 , plot the values of Hm .dil, m0 !m00 / versus mB , extrapolate the curve to
B
B
B
p
mB D0, and shift the origin to the extrapolated intercept, resulting in a plot of L versus
p
mB .

0.5
1.0
1.5
kJ mol
kJ mol
m sol
kJ mol
2.0
B mol kg
332

B mol kg
(b)
(a)
Figure 11.10 Thermal properties of aqueous NaCl at 25:00 C.

(a) Left axis: molar integral enthalpy of solution to produce solution of molality mB .a
The dashed line has a slope equal to the theoretical limiting value of the slope of the
curve. Right axis: relative apparent molar enthalpy of the solute.
(b) Relative partial molar enthalpy of the solute as a function of molality.b
a Calculated
b Based
from molar enthalpy of formation values in Ref. [163], p. 2-301.

on data in Ref. [125], Table X.
In order to be able to use Eq. 11.4.23, we need to relate the derivative dL = dmB to the
p
slope of the curve of L versus mB . We write
1
p
d mB D p
dmB
2 mB
p
p
dmB D 2 mB d mB
(11.4.26)
Substituting this expression for dmB into Eq. 11.4.23, we obtain the following operational
p
equation for evaluating LB from the plot of L versus mB :
p
mB dL
LB D L C
(11.4.27)
p
2 d mB
(constant T and p)
The value of L goes to zero at innite dilution. When the solute is an electrolyte, the
dependence of L on mB in solutions dilute enough for the DebyeH ckel limiting law to
u
apply is given by
p
L D CL mB
(11.4.28)
(very dilute solution)
For aqueous solutions of a 1:1 electrolyte at 25 C, the coefcient CL has the value10
CL D 1:988
103 J kg1=2 mol
3=2
(11.4.29)
10 The fact that C
L is positive means, according to Eq. 11.4.25, that dilution of a very dilute electrolyte solution
is an exothermic process.

333
p
p
CL is equal to the limiting slope of L versus mB , of Hm (sol, mB ) versus mB ,
p 0
and of Hm .dil, m0 !m00 / versus mB . The value given by Eq. 11.4.29 can be used for
B
B
extrapolation of measurements at 25 C and low molality to innite dilution.
Equation 11.4.28 can be derived as follows. For simplicity, we assume the pressure
1
is the standard pressure p . At this pressure HB is the same as HB , and Eq. 11.4.17
becomes LB D HB HB . From Eqs. 12.1.3 and 12.1.6 in the next chapter, we can
write the relations
d. =T /
@. B =T /
m;B
HB D T 2
HB D T 2
(11.4.30)
@T
dT
p;fnig
Subtracting the second of these relations from the rst, we obtain
@. B
m;B /=T
HB HB D T 2
@T
p;fnig
(11.4.31)
The solute activity on a molality basis, am;B , is dened by B

m;B D RT ln am;B .
The activity of an electrolyte solute at the standard pressure, from Eq. 10.3.10, is given
by am;B D . CC
/ .mB =m / . Accordingly, the relative partial molar enthalpy
of the solute is related to the mean ionic activity coefcient by
@ ln
LB D RT 2
(11.4.32)
@T
p;fnig
We assume the solution is sufciently dilute for the mean ionic activity coefcient
to be adequately described by the DebyeH ckel limiting law, Eq. 10.4.8: ln D
u
p
ADH jzC z j Im , where ADH is a temperature-dependent quantity dened on page
295. Then Eq. 11.4.32 becomes
p
@ADH
(11.4.33)
LB D RT 2 jzC z j Im
@T p;fnig
Substitution of the expression given by Eq. 10.4.9 on page 298 for Im in a solution of
a single completely-dissociated electrolyte converts Eq. 11.4.33 to
#
"
RT 2 @ A ADH
3=2 p
LB D p
mB
jzC z j
@T
2
p;fnig
p
D CLB mB
(11.4.34)
The coefcient CLB (the quantity in brackets) depends on T , the kind of solvent, and
the ion charges and number of ions per solute formula unit, but not on the solute
molality.
p
Let CL represent the limiting slope of L versus mB . In a very dilute solution
p
we have L D CL mB , and Eq. 11.4.27 becomes
p
p
mB dL
mB
p
LB D L C
CL
D CL mB C
p
2 d mB
2
(11.4.35)
By equating this expression for LB with the one given by Eq. 11.4.34 and solving for
p
CL , we obtain CL D .2=3/CLB and L D .2=3/CLB mB .

11.5 R EACTION C ALORIMETRY
334
11.5 REACTION CALORIMETRY

Reaction calorimetry is used to evaluate the molar integral reaction enthalpy Hm (rxn) of a
reaction or other chemical process at constant temperature and pressure. The measurement
actually made, however, is a temperature change.
Sections 11.5.1 and 11.5.2 will describe two common types of calorimeters designed for
reactions taking place at either constant pressure or constant volume. The constant-pressure
type is usually called a reaction calorimeter, and the constant-volume type is known as a
bomb calorimeter or combustion calorimeter.
In either type of calorimeter, the chemical process takes place in a reaction vessel surrounded by an outer jacket. The jacket may be of either the adiabatic type or the isothermaljacket type described in Sec. 7.3.2 in connection with heat capacity measurements. A
temperature-measuring device is immersed either in the vessel or in a phase in thermal
contact with it. The measured temperature change is caused by the chemical process, instead of by electrical work as in the determination of heat capacity. One important way in
which these calorimeters differ from ones used for heat capacity measurements is that work
is kept deliberately small, in order to minimize changes of internal energy and enthalpy
during the experimental process.
11.5.1 The constant-pressure reaction calorimeter

The contents of a constant-pressure calorimeter are usually open to the atmosphere, so this
type of calorimeter is unsuitable for processes involving gases. It is, however, a convenient
apparatus in which to study a liquid-phase chemical reaction, the dissolution of a solid or
liquid solute in a liquid solvent, or the dilution of a solution with solvent.
The process is initiated in the calorimeter by allowing the reactants to come into contact.
The temperature in the reaction vessel is measured over a period of time starting before the
process initiation and ending after the advancement has reached a nal value with no further
change.
The heating or cooling curve (temperature as a function of time) is observed over a
period of time that includes the period during which changes. For an exothermic reaction
occurring in an adiabatic calorimeter, the heating curve may resemble that shown in Fig. 7.3
on page 170, and the heating curve in an isothermal-jacket calorimeter may resemble that
shown in Fig. 7.4 on page 172. Two points are designated on the heating or cooling curve:
one at temperature T1 , before the reaction is initiated, and the other at T2 , after has reached
its nal value. These points are indicated by open circles in Figs. 7.3 and 7.4.
Figure 11.11 on the next page depicts three paths at constant pressure. The enthalpy
change of the experimental process, in which reactants at temperature T1 change to products
at temperature T2 , is denoted H (expt).
The value of H (expt) at constant pressure would be zero if the process were perfectly
adiabatic and the only work were expansion work, but this is rarely the case. There may be
unavoidable work from stirring and from electrical temperature measurement. We can evaluate H (expt) by one of the methods described in Sec. 7.3.2. For an adiabatic calorimeter,
the appropriate expression is H (expt) D r.t2 t1 / (Eq. 7.3.19 on page 171 with wel
set equal to zero), where is the energy equivalent of the calorimeter, r is the slope of the
heating curve when no reaction is occurring, and t1 and t2 are the times at temperatures T1

335
P
(expt)
rxn,
Figure 11.11 Enthalpy changes for paths at constant pressure (schematic). R denotes
reactants and P denotes products.
and T2 . For an isothermal-jacket calorimeter, we evaluate H (expt) using Eq. 7.3.28 on

page 173 with wel set equal to zero.
The enthalpy change we wish to nd is the reaction enthalpy H.rxn, T1 /, which is
the change for the same advancement of the reaction at constant temperature T1 . The paths
labeled H (expt) and H.rxn, T1 / in the gure have the same initial state and different
nal states. The path connecting these two nal states is for a change of the temperature
from T1 to T2 with xed at its nal value; the enthalpy change for this path is denoted
H.P/.11 The value of H.P/ can be calculated from
H.P/ D
P .T2
T1 /
(11.5.1)
where P is the energy equivalent (the average heat capacity of the calorimeter) when the
calorimeter contains the products. To measure P , we can carry out a second experiment
involving work with an electric heater included in the calorimeter, similar to the methods
described in Sec. 7.3.2.
Since the difference of enthalpy between two states is independent of the path, we can
write H (expt) D H.rxn, T1 / C P .T2 T1 /, or
H.rxn, T1 / D
P .T2
T1 / C H (expt)
(11.5.2)
The molar integral reaction enthalpy at temperature T1 is the reaction enthalpy divided
by , the advancement during the experimental process:
Hm (rxn) D H.rxn, T1 /=
T1 / C H (expt)
P .T2
D
(11.5.3)
(constant-pressure
calorimeter)
Note that H (expt) is small, so that Hm (rxn) is approximately equal to P .T2 T1 /= .

If T2 is greater than T1 (the process is exothermic), then Hm (rxn) is negative, reecting
the fact that after the reaction takes place in the calorimeter, heat would have to leave the
system in order for the temperature to return to its initial value. If T2 is less than T1 (the
process is endothermic), Hm (rxn) is positive.
11 The symbol P refers to the nal equilibrium state in which the reaction vessel contains products of the reaction
and any excess reactants.


336
Most reactions cause a change in the composition of one or more phases, in which case
Hm (rxn) is not the same as the molar differential reaction enthalpy, r H D .@H=@ /T;p ,
unless the phase or phases can be treated as ideal mixtures (see Sec. 11.2.2). Corrections, usually small, are needed to obtain the standard molar reaction enthalpy r H from
Hm (rxn).
11.5.2 The bomb calorimeter

A bomb calorimeter typically is used to carry out the complete combustion of a solid or
liquid substance in the presence of excess oxygen. The combustion reaction is initiated with
electrical ignition. In addition to the main combustion reaction, there may be unavoidable
side reactions, such as the formation of nitrogen oxides if N2 is not purged from the gas
phase. Sometimes auxiliary reactions are deliberately carried out to complete or moderate
the main reaction.
From the measured heating curve and known properties of the calorimeter, reactants,
and products, it is possible to evaluate the standard molar enthalpy of combustion, c H ,
of the substance of interest at a particular temperature called the reference temperature, Tref .
(Tref is often chosen to be 298:15 K, which is 25:00 C.) With careful work, using temperature measurements with a resolution of 1 10 4 K or better and detailed corrections, the
precision of c H can be of the order of 0:01 percent.
Bomb calorimetry is the principal means by which standard molar enthalpies of combustion of individual elements and of compounds of these elements are evaluated. From
these values, using Hesss law, we can calculate the standard molar enthalpies of formation
of the compounds as described in Sec. 11.3.2. From the formation values of only a few compounds, the standard molar reaction enthalpies of innumerable reactions can be calculated
with Hesss law (Eq. 11.3.3 on page 321).
Because of their importance, the experimental procedure and the analysis of the data
it provides will now be described in some detail. A comprehensive problem (Prob. 11.7)
based on this material is included at the end of the chapter.
There are ve main steps in the procedure of evaluating a standard molar enthalpy of
combustion:
1. The combustion reaction, and any side reactions and auxiliary reactions, are carried
out in the calorimeter, and the course of the resulting temperature change is observed.
2. The experimental data are used to determine the value of U.IBP; T2 /, the internal
energy change of the isothermal bomb process at the nal temperature of the reaction.
The isothermal bomb process is the idealized process that would have occurred if
the reaction or reactions had taken place in the calorimeter at constant temperature.
3. The internal energy change of the isothermal bomb process is corrected to yield
U.IBP; Tref /, the value at the reference temperature of interest.
4. The standard molar internal energy of combustion, c U .Tref /, is calculated. This
calculation is called reduction to standard states.
5. The standard molar enthalpy of combustion, c H .Tref /, is calculated.
These ve steps are described below.

ignition
wires
337
stirrer
jacket
calorimeter wall
oxygen inlet
platinum resistance
thermometer
sample holder
air or vacuum
water
bomb vessel
Figure 11.12 Section view of a bomb calorimeter.
Experimental
The common form of combustion bomb calorimeter shown in Fig. 11.12 consists of a thickwalled cylindrical metal vessel to contain the reactants of the combustion reaction. It is
called a bomb because it is designed to withstand high pressure. The bomb can be sealed
with a gas-tight screw cap. During the reaction, the sealed bomb vessel is immersed in
water in the calorimeter, which is surrounded by a jacket. Conceptually, we take the system
to be everything inside the jacket, including the calorimeter walls, water, bomb vessel, and
contents of the bomb vessel.
To prepare the calorimeter for a combustion experiment, a weighed sample of the substance to be combusted is placed in a metal sample holder. The calculations are simplied
if we can assume all of the sample is initially in a single phase. Thus, a volatile liquid is
usually encapsulated in a bulb of thin glass (which shatters during the ignition) or conned
in the sample holder by cellulose tape of known combustion properties. If one of the combustion products is H2 O, a small known mass of liquid water is placed in the bottom of the
bomb vessel to saturate the gas space of the bomb vessel with H2 O. The sample holder and
ignition wires are lowered into the bomb vessel, the cap is screwed on, and oxygen gas is
admitted through a valve in the cap to a total pressure of about 30 bar.
To complete the setup, the sealed bomb vessel is immersed in a known mass of water
in the calorimeter. A precision thermometer and a stirrer are also immersed in the water.
With the stirrer turned on, the temperature is monitored until it is found to change at a slow,
practically-constant rate. This drift is due to heat transfer through the jacket, mechanical
stirring work, and the electrical work needed to measure the temperature. A particular time
is chosen as the initial time t1 . The measured temperature at this time is T1 , assumed to be
practically uniform throughout the system.
At or soon after time t1 , the ignition circuit is closed to initiate the combustion reaction in the bomb vessel. If the reaction is exothermic, the measured temperature rapidly

IBP
338
(expt)
Figure 11.13 Internal energy changes for paths at constant volume in a bomb calorimeter (schematic). R denotes reactants and P denotes products.
increases over the course of several minutes as chemical energy is converted to thermal energy. For a while the temperature in the system is far from uniform, as energy is transferred
by heat through the walls of the bomb vessel walls to the water outside.
When the measured temperature is again observed to change at a slow and practically
constant rate, the reaction is assumed to be complete and the temperature is assumed once
more to be uniform. A second time is now designated as the nal time t2 , with nal temperature T2 . For best accuracy, conditions are arranged so that T2 is close to the desired
reference temperature Tref .
Because the jacket is not gas tight, the pressure of the water outside the bomb vessel
stays constant at the pressure of the atmosphere. Inside the bomb vessel, the changes in
temperature and composition take place at essentially constant volume, so the pressure inside the vessel is not constant. The volume change of the entire system during the process
is negligible.
The isothermal bomb process
The relations derived here parallel those of Sec. 11.5.1 for a constant-pressure calorimeter.
The three paths depicted in Fig. 11.13 are similar to those in Fig. 11.11 on page 335, except
that instead of being at constant pressure they are at constant volume. We shall assume the
combustion reaction is exothermic, with T2 being greater than T1 .
The internal energy change of the experimental process that actually occurs in the
calorimeter between times t1 and t2 is denoted U (expt) in the gure. Conceptually, the
overall change of state during this process would be duplicated by a path in which the temperature of the system with the reactants present increases from T1 to T2 ,12 followed by the
isothermal bomb process at temperature T2 . In the gure these paths are labeled with the
internal energy changes U.R/ and U.IBP; T2 /, and we can write
U (expt) D U.R/ C U.IBP; T2 /
(11.5.4)
To evaluate U.R/, we can use the energy equivalent R of the calorimeter with reactants present in the bomb vessel. R is the average heat capacity of the system between T1
and T2 that is, the ratio q=.T2 T1 /, where q is the heat that would be needed to change
12 When
one investigates a combustion reaction, the path in which temperature changes without reaction is best
taken with reactants rather than products present because the reactants are more easily characterized.

339
the temperature from T1 to T2 . From the rst law, with expansion work assumed negligible,
the internal energy change equals this heat, giving us the relation
U.R/ D
R .T2
T1 /
(11.5.5)
The initial and nal states of the path are assumed to be equilibrium states, and there may
be some transfer of reactants or H2 O from one phase to another within the bomb vessel
during the heating process.
The value of R is obtained in a separate calibration experiment. The calibration is
usually carried out with the combustion of a reference substance, such as benzoic acid,
whose internal energy of combustion under controlled conditions is precisely known from
standardization based on electrical work. If the bomb vessel is immersed in the same mass
of water in both experiments and other conditions are similar, the difference in the values
of R in the two experiments is equal to the known difference in the heat capacities of the
initial contents (reactants, water, etc.) of the bomb vessel in the two experiments.
The internal energy change we wish to nd is U.IBP; T2 /, that of the isothermal bomb
process in which reactants change to products at temperature T2 , accompanied perhaps by
some further transfer of substances between phases. From Eqs. 11.5.4 and 11.5.5, we obtain
U.IBP; T2 / D
.T2
T1 / C U (expt)
(11.5.6)
The value of U (expt) is small. To evaluate it, we must look in detail at the possible
sources of energy transfer between the system and the surroundings during the experimental
process. These sources are
1. electrical work wign done on the system by the ignition circuit;
2. heat transfer, minimized but not eliminated by the jacket;
3. mechanical stirring work done on the system;
4. electrical work done on the system by an electrical thermometer.
The ignition work occurs during only a short time interval at the beginning of the process,
and its value is known. The effects of heat transfer, stirring work, and temperature measurement continue throughout the course of the experiment. With these considerations, Eq.
11.5.6 becomes
U.IBP; T2 / D
.T2
T1 / C wign C U 0 (expt)
(11.5.7)
where U 0 (expt) is the internal energy change due to heat, stirring, and temperature measurement. U 0 (expt) can be evaluated from the energy equivalent and the observed rates
of temperature change at times t1 and t2 ; the relevant relations for an isothermal jacket are
Eq. 7.3.24 (with wel set equal to zero) and Eq. 7.3.32.
Correction to the reference temperature
The value of U.IBP; T2 / evaluated from Eq. 11.5.7 is the internal energy change of the
isothermal bomb process at temperature T2 . We need to correct this value to the desired
reference temperature Tref . If T2 and Tref are close in value, the correction is small and can
be calculated with a modied version of the Kirchhoff equation (Eq. 11.3.10 on page 324):
U.IBP; Tref / D U.IBP; T2 / C CV (P)
CV (R) .Tref
T2 /
(11.5.8)

340
Here CV (P) and CV (R) are the heat capacities at constant volume of the contents of the
bomb vessel with products and reactants, respectively, present.
Reduction to standard states
We want to obtain the value of c U .Tref /, the molar internal energy change for the main
combustion reaction at the reference temperature under standard-state conditions. Once we
have this value, it is an easy matter to nd the molar enthalpy change under standard-state
conditions, our ultimate goal.
Consider a hypothetical process with the following three isothermal steps carried out at
the reference temperature Tref :
1. Each substance initially present in the bomb vessel changes from its standard state to
the state it actually has at the start of the isothermal bomb process.
2. The isothermal bomb process takes place, including the main combustion reaction
and any side reactions and auxiliary reactions.
3. Each substance present in the nal state of the isothermal bomb process changes to
its standard state.
The net change is a decrease in the amount of each reactant in its standard state and an
increase in the amount of each product in its standard state. The internal energy change of
step 2 is U.IBP; Tref /, whose value is found from Eq. 11.5.8. The internal energy changes
of steps 1 and 3 are called Washburn corrections.13
Thus, we calculate the standard internal energy change of the main combustion reaction
at temperature Tref from
X
U .cmb; Tref / D U.IBP; Tref / C (Washburn corrections)
i r U .i/ (11.5.9)
i
where the sum over i is for side reactions and auxiliary reactions if present. Finally, we
calculate the standard molar internal energy of combustion from
c U .Tref / D
where
U .cmb; Tref /
c
(11.5.10)
is the advancement of the main combustion reaction in the bomb vessel.
Standard molar enthalpy change

The quantity c U .Tref / is the molar internal energy change for the main combustion reaction carried out at constant temperature Tref with each reactant and product in its standard
P
state at pressure p . From the relations c H D i i Hi (Eq. 11.2.15) and Hi D Ui C pVi
(from Eq. 9.2.50), we get
X
c H .Tref / D c U .Tref / C p
(11.5.11)
i Vi
i
13 Ref.
[168].

341
Molar volumes of condensed phases are much smaller than those of gases, and to a good
approximation we may write
X g
c H .Tref / D c U .Tref / C p
(11.5.12)
i Vi (g)
i
where the sum includes only gaseous reactants and products of the main combustion reaction. Since a gas in its standard state is an ideal gas with molar volume equal to RT =p ,
the nal relation is
X g
c H .Tref / D c U .Tref / C
(11.5.13)
i RTref
i
Washburn corrections
The Washburn corrections needed in Eq. 11.5.9 are internal energy changes for certain
hypothetical physical processes occurring at the reference temperature Tref involving the
substances present in the bomb vessel. In these processes, substances change from their
standard states to the initial state of the isothermal bomb process, or change from the nal
state of the isothermal bomb process to their standard states.
For example, consider the complete combustion of a solid or liquid compound of carbon, hydrogen, and oxygen in which the combustion products are CO2 and H2 O and there
are no side reactions or auxiliary reactions. In the initial state of the isothermal bomb process, the bomb vessel contains the pure reactant, liquid water with O2 dissolved in it, and
a gaseous mixture of O2 and H2 O, all at a high pressure p1 . In the nal state, the bomb
vessel contains liquid water with O2 and CO2 dissolved in it and a gaseous mixture of O2 ,
H2 O, and CO2 , all at pressure p2 . In addition, the bomb vessel contains internal parts of
constant mass such as the sample holder and ignition wires.
In making Washburn corrections, we must use a single standard state for each substance
in order for Eq. 11.5.9 to correctly give the standard internal energy of combustion. In
the present example we choose the following standard states: pure solid or liquid for the
reactant compound, pure liquid for the H2 O, and pure ideal gases for the O2 and CO2 , each
at pressure p D 1 bar.
We can calculate the amount of each substance in each phase, in both the initial state
and nal state of the isothermal bomb process, from the following information: the internal volume of the bomb vessel; the mass of solid or liquid reactant initially placed in the
vessel; the initial amount of H2 O; the initial O2 pressure; the water vapor pressure; the
solubilities (estimated from Henrys law constants) of O2 and CO2 in the water; and the
stoichiometry of the combustion reaction. Problem 11.7 on page 362 guides you through
these calculations.
11.5.3 Other calorimeters

Experimenters have used great ingenuity in designing calorimeters to measure reaction
enthalpies and to improve their precision. In addition to the constant-pressure reaction
calorimeter and bomb calorimeter described above, three additional types will be briey
mentioned.
A phase-change calorimeter has two coexisting phases of a pure substance in thermal
contact with the reaction vessel and an adiabatic outer jacket. The two coexisting phases

11.6 A DIABATIC F LAME T EMPERATURE
342
constitute a univariant subsystem that at constant pressure is at the xed temperature of

the equilibrium phase transition. The thermal energy released or absorbed by the reaction,
instead of changing the temperature, is transferred isothermally to or from the coexisting
phases and can be measured by the volume change of the phase transition. A reaction enthalpy, of course, can only be measured by this method at the temperature of the equilibrium
phase transition. The well-known Bunsen ice calorimeter uses the icewater transition at
0 C. The solidliquid transition of diphenyl ether has a relatively large volume change and
is useful for measurements at 26:9 C. Phase-transition calorimeters are especially useful
for slow reactions.
A heat-ow calorimeter is a variation of an isothermal-jacket calorimeter. It uses a thermopile (Fig. 2.7) to continuously measure the temperature difference between the reaction
vessel and an outer jacket acting as a constant-temperature heat sink. The heat transfer takes
place mostly through the thermocouple wires, and to a high degree of accuracy is proportional to the temperature difference integrated over time. This is the best method for an
extremely slow reaction, and it can also be used for rapid reactions.
A ame calorimeter is a ow system in which oxygen, uorine, or another gaseous
oxidant reacts with a gaseous fuel. The heat transfer between the ow tube and a heat sink
can be measured with a thermopile, as in a heat-ow calorimeter.
11.6 ADIABATIC FLAME TEMPERATURE

With a few simple approximations, we can estimate the temperature of a ame formed in
a owing gas mixture of oxygen or air and a fuel. We treat a moving segment of the gas
mixture as a closed system in which the temperature increases as combustion takes place.
We assume that the reaction occurs at a constant pressure equal to the standard pressure,
and that the process is adiabatic and the gas is an ideal-gas mixture.
The principle of the calculation is similar to that used for a constant-pressure calorimeter
as explained by the paths shown in Fig. 11.11 on page 335. When the combustion reaction
in the segment of gas reaches reaction equilibrium, the advancement has changed by
and the temperature has increased from T1 to T2 . Because the reaction is assumed to be
adiabatic at constant pressure, H (expt) is zero. Therefore, the sum of H.rxn; T1 / and
H.P/ is zero, and we can write
c H .T1 / C
T2
T1
Cp .P/ dT D 0
(11.6.1)
where c H .T1 / is the standard molar enthalpy of combustion at the initial temperature,
and Cp .P/ is the heat capacity at constant pressure of the product mixture.
The value of T2 that satises Eq. 11.6.1 is the estimated ame temperature. Problem
11.9 presents an application of this calculation. Several factors cause the actual temperature
in a ame to be lower: the process is never completely adiabatic, and in the high temperature
of the ame there may be product dissociation and other reactions in addition to the main
combustion reaction.

11.7 G IBBS E NERGY AND R EACTION E QUILIBRIUM
343
11.7 GIBBS ENERGY AND REACTION EQUILIBRIUM

This section begins by examining the way in which the Gibbs energy changes as a chemical
process advances in a closed system at constant T and p with expansion work only. A
universal criterion for reaction equilibrium is derived involving the molar reaction Gibbs
energy.
11.7.1 The molar reaction Gibbs energy

Applying the general denition of a molar differential reaction quantity (Eq. 11.2.15) to the
Gibbs energy of a closed system with T , p, and as the independent variables, we obtain
the denition of the molar reaction Gibbs energy or molar Gibbs energy of reaction, r G:
def
r G D
(11.7.1)
Equation 11.2.16 shows that this quantity is also given by the partial derivative
@G
r G D
@ T;p
(11.7.2)
(closed system)
The total differential of G is then

dG D
S dT C V dp C r G d
(11.7.3)
(closed system)
11.7.2 Spontaneity and reaction equilibrium

In Sec. 5.8, we found that the spontaneous direction of a process taking place in a closed
system at constant T and p, with expansion work only, is the direction of decreasing G. In
the case of a chemical process occurring at constant T and p, r G is the rate at which G
changes with . Thus if r G is positive, spontaneously decreases; if r G is negative,
spontaneously increases. During a spontaneous process d and r G have opposite signs.14
Note how the equality of Eq. 11.7.3 agrees with the inequality dG < S dT C V dp,
a criterion of spontaneity in a closed system with expansion work only (Eq. 5.8.6 on
page 146). When d and r G have opposite signs, r G d is negative and dG D
. S dT C V dp C r G d / is less than . S dT C V dp/.
If the system is closed and contains at least one phase that is a mixture, a state of reaction
equilibrium can be approached spontaneously at constant T and p in either direction of the
reaction; that is, by both positive and negative changes of . In this equilibrium state,
therefore, G has its minimum value for the given T and p. Since G is a smooth function
14 Sometimes
reaction spontaneity at constant T and p is ascribed to the driving force of a quantity called
the afnity of reaction, dened as the negative of r G. increases spontaneously if the afnity is positive and
decreases spontaneously if the afnity is negative; the system is at equilibrium when the afnity is zero.

344
of , its rate of change with respect to is zero in the equilibrium state. The condition for
reaction equilibrium, then, is that r G must be zero:
X
r G D
(11.7.4)
i i D0
(reaction equilibrium)
It is important to realize that this condition is independent of whether or not reaction

equilibrium is approached at constant temperature and pressure. It is a universal criterion
P
of reaction equilibrium. The value of r G is equal to i i i and depends on the state of
the system. If the state is such that r G is positive, the direction of spontaneous change
is one that, under the existing constraints, allows r G to decrease. If r G is negative,
the spontaneous change increases the value of r G. When the system reaches reaction
equilibrium, whatever the path of the spontaneous process, the value of r G becomes zero.
11.7.3 General derivation

We can obtain the condition of reaction equilibrium given by Eq. 11.7.4 in a more general
and rigorous way by an extension of the derivation of Sec. 9.2.7, which was for equilibrium
conditions in a multiphase, multicomponent system.
Consider a system with a reference phase, 0 , and optionally other phases labeled by
0 . Each phase contains one or more species labeled by subscript i, and some or all of
the species are the reactants and products of a reaction.
The total differential of the internal energy is given by Eq. 9.2.37 on page 237:
X 0
0
0
0
0
0
dU D T dS p dV C
i dni
i
T dS
p dV C
i dni
(11.7.5)
The conditions of isolation are

dU D 0
dV
X
0

(no expansion work)
dV D 0
For each species i:

X
0
dn C
dn D
i
i
(11.7.6)
(11.7.7)
(closed system)
(11.7.8)
00
In Eq. 11.7.8, dn0 should be set equal to zero for a species i 0 that is excluded from phase
i
00 , and i 00 should be set equal to zero for a species i 00 that is not a reactant or product of
the reaction.

0
345
0
We use these conditions of isolation to substitute for dU , dV , and dn in Eq. 11.7.5,

i
P
0
and make the further substitution dS D dS

0 dS . Solving for dS, we obtain
X .T 0 T /
dS D
dS
T 0
0
XX .
i
0
i
X .p 0 p /
dV
T 0
0
P
0
i i i
i/
dni
d
0
T
(11.7.9)
The equilibrium condition is that the coefcient multiplying each differential on the right
side of Eq. 11.7.9 must be zero. We conclude that at equilibrium the temperature of each
phase is equal to that of phase 0 ; the pressure of each phase is equal to that of phase 0 ;
the chemical potential of each species, in each phase containing that species, is equal to the
P
0
chemical potential of the species in phase 0 ; and the quantity i i (which is equal to
i
r G) is zero.
In short, in an equilibrium state each phase has the same temperature and the same
pressure, each species has the same chemical potential in the phases in which it is present,
and the molar reaction Gibbs energy of each phase is zero.
11.7.4 Pure phases

Consider a chemical process in which each reactant and product is in a separate pure phase.
For example, the decomposition of calcium carbonate, CaCO3 (s) ! CaO(s) + CO2 (g), involves three pure phases if no other gas is allowed to mix with the CO2 .
As this kind of reaction advances at constant T and p, the chemical potential of each
substance remains constant, and r G is therefore constant. The value of r G for this
reaction depends only on T and p. If r G is negative, the reaction proceeds spontaneously
to the right until one of the reactants is exhausted; the reaction is said to go to completion.
If r G is positive, the reaction proceeds spontaneously to the left until one of the products
is exhausted.15 The reactants and products can remain in equilibrium only if T and p are
such that r G is zero. These three cases are illustrated in Fig. 11.14 on the next page.
Note the similarity of this behavior to that of an equilibrium phase transition of a pure
substance. Only one phase of a pure substance is present at equilibrium unless trs G
is zero. A phase transition is a special case of a chemical process.
11.7.5 Reactions involving mixtures

If any of the reactants or products of a chemical process taking place in a closed system is
a constituent of a mixture, a plot of G versus (at constant T and p) turns out to exhibit a
minimum with a slope of zero; see the example in Fig. 11.15 on the next page. At constant
T and p, changes spontaneously in the direction of decreasing G until the minimum is
reached, at which point r G (the slope of the curve) is zero and the system is in a state of
reaction equilibrium.
15 Keep
in mind that whether a species is called a reactant or a product depends, not on whether its amount
decreases or increases during a reaction process, but rather on which side of the reaction equation it appears.

346
(a)
(b)
(c)
Figure 11.14 Gibbs energy versus advancement at constant T and p in systems of

pure phases. G is a linear function of with slope equal to r G.
(a) r G is negative; spontaneously increases.
(b) r G is positive; spontaneously decreases.
(c) r G is zero; the system is in reaction equilibrium at all values of .
eq
Figure 11.15 Gibbs energy as a function of advancement at constant T and p in a

closed system containing a mixture. The open circle is at the minimum value of G.
(The reaction is the same as in Fig. 11.6 on page 318.)
P
The condition of reaction equilibrium given by r GD0 or i i i D0 is a general one
that is valid whether or not the reaction proceeds at constant T and p. Suppose a spontaneous reaction occurs in a closed system at constant temperature and volume. The system
P
is at reaction equilibrium when i i i becomes equal to zero. To relate this condition to
the change of a thermodynamic potential, we take the expression for the total differential of
the Helmholtz energy of an open system, with expansion work only, given by Eq. 5.5.8 on
page 142:
X
dA D S dT p dV C
(11.7.10)
i dni
i
When we make the substitution dni D
d , we obtain an expression for the total differen-

347
tial of A in a closed system with a chemical reaction:

dA D
S dT
p dV C
(11.7.11)
We identify the coefcient of the last term on the right as a partial derivative:

X
@A
i i D
@ T;V
(11.7.12)
This equation shows that as the reaction proceeds spontaneously at constant T and V , it
reaches reaction equilibrium at the point where .@A=@ /T;V is zero. This is simply another way to express the criterion for spontaneity stated on page 146: If the only work is
expansion work, the Helmholtz energy of a closed system decreases during a spontaneous
process at constant T and V and has its minimum value when the system attains an equilibrium state.
11.7.6 Reaction in an ideal gas mixture

Let us look in detail at the source of the minimum in G for the case of a reaction occurring in
an ideal gas mixture in a closed system at constant T and p. During this process the system
has only one independent variable, which it is convenient to choose as the advancement .
The additivity rule (Eq. 9.2.25) for the Gibbs energy is
X
GD
ni i
(11.7.13)
i
where both ni and i depend on . Thus, G is a complicated function of .

For the chemical potential of each substance, we write i D (g) C RT ln.pi =p /
i
(Eq. 9.3.5), where pi is the partial pressure of i in the mixture. Substitution in Eq. 11.7.13
gives, for the Gibbs energy at any value of ,
X
pi
G. / D
ni
(11.7.14)
i (g) C RT ln
p
i
At
D 0, the amounts and partial pressures have their initial values ni;0 and pi;0 :
X
pi;0
G.0/ D
ni;0 (g) C RT ln
(11.7.15)
i
p
i
The difference between these two expressions is

X
G. / G.0/ D
.ni ni;0 / (g)
i
i
C RT
X
i
ni ln
pi
p
RT
X
i
ni;0 ln
pi;0
p
(11.7.16)

348
Converting partial pressures to mole fractions with pi D yi p and pi;0 D yi;0 p gives
G. /
G.0/ D
X
.ni
i
RT
X
i
C RT
ni;0 ln yi;0 C RT
With the substitution ni ni;0 D

side of Eq. 11.7.17, the result is
X
G. / G.0/ D
i
i (g)
X
i
ni ln yi
.ni
ni;0 / ln
p
p
(11.7.17)
(Eq. 11.2.11) in the rst and last terms on the right
RT
i (g)
ni;0 /
C RT
ni ln yi
ni:0 ln yi;0 C RT
ln
p
p
(11.7.18)
P
The sum i i (g) in the rst term on the right side of Eq. 11.7.18 is r G , the
i
standard molar reaction Gibbs energy. Making this substitution gives nally
X
X
G. / G.0/ D r G C RT
ni ln yi RT
ni;0 ln yi;0
i
C RT
X
i
!
i
ln
p
p
(11.7.19)
(ideal gas mixture)
There are four terms on the right side of Eq. 11.7.19. The rst term is the Gibbs energy change for the reaction of pure reactants to form pure products under standard-state
conditions, the second is a mixing term, the third term is constant, and the last term is an
adjustment of G from the standard pressure to the pressure of the gas mixture. Note that
the rst and last terms are proportional to the advancement and cannot be the cause of a
P
minimum in the curve of the plot of G versus . It is the mixing termP
RT i ni ln yi that
id
is responsible for the observed minimum.16 This term divided by nD i ni is Gm (mix),
the molar differential Gibbs energy of mixing to form an ideal mixture (see Eq. 11.1.8 on
id
page 305); the term is also equal to nTSm (mix) (Eq. 11.1.9), showing that the minimum
is entirely an entropy effect.
Now let us consider specically the simple reaction
A.g/ ! B.g/
in an ideal gas mixture, for which A is 1 and B is C1. Let the initial state be one of pure
A: nB;0 D0. The initial mole fractions are then yA;0 D1 and yB;0 D0. In this reaction, the
total amount n D nA C nB is constant. Substituting these values in Eq. 11.7.19 gives17
G. /
G.0/ D r G C nRT .yA ln yA C yB ln yB /
(11.7.20)
16 This
term also causes the slope of the curve of G. / G.0/ versus to be 1 and C1 at the left and right
extremes of the curve.
17 Note that although ln y approaches 1 as y approaches zero, the product y ln y approaches zero in
A
A
A
A
this limit. This behavior can be proved with lHospitals rule (see any calculus textbook).

=nA;0
(a)
349
1
=nA;0
(b)
1
=nA;0
(c)
Figure 11.16 Gibbs energy as a function of the advancement of the reaction A ! B

in an ideal gas mixture at constant T and p. The initial amount of B is zero. The
equilibrium positions are indicated by open circles.
(a) r G < 0. (b) r G D 0. (c) r G > 0.
id
The second term on the right side is nGm (mix), the Gibbs energy of mixing pure ideal
gases A and B at constant T and p to form an ideal gas mixture of composition yA and yB .
id
Since the curve of Gm (mix) plotted against has a minimum (as shown in Fig. 11.2 on
page 306), G. / G.0/ also has a minimum.
Figure 11.16 illustrates how the position of the minimum, which is the position of reaction equilibrium, depends on the value of r G . The more negative is r G , the closer
to the product side of the reaction is the equilibrium position. On the other hand, the more
positive is r G , the smaller is the value of at equilibrium. These statements apply to any
reaction in a homogeneous mixture.
As the reaction A ! B proceeds, there is no change in the total number of molecules,
and therefore in an ideal gas mixture at constant temperature and volume there is no pressure
change. The point of reaction equilibrium is at the minimum of G when both V and p are
constant.
The situation is different when the number of molecules changes during the reaction.
Consider the reaction A ! 2 B in an ideal gas mixture. As this reaction proceeds to the
right at constant T , the volume increases if the pressure is held constant and the pressure
increases if the volume is held constant. Figure 11.17 on the next page shows how G
depends on both p and V for this reaction. Movement along the horizontal dashed line in
the gure corresponds to reaction at constant T and p. The minimum of G along this line is
at the volume indicated by the open circle. At this volume, G has an even lower minimum
at the pressure indicated by the lled circle, where the vertical dashed line is tangent to one
of the contours of constant G. The condition needed for reaction equilibrium, however, is
that r G must be zero. This condition is satised along the vertical dashed line only at the
position of the open circle.
This example demonstrates that for a reaction occurring at constant temperature and

11.8 T HE T HERMODYNAMIC E QUILIBRIUM C ONSTANT
350
A,
Figure 11.17 Dependence of Gibbs energy on volume and pressure, at constant temperature, in a closed system containing an ideal gas mixture of A and B. The reaction is
A ! 2 B with r G D0. Solid curves: contours of constant G plotted at an interval of
0:5nA;0 RT . Dashed curve: states of reaction equilibrium (r G D 0). Dotted curves:
limits of possible values of the advancement. Open circle: position of minimum G
(and an equilibrium state) at the constant pressure p D 1:02p . Filled circle: position
of minimum G for a constant volume of 1:41V0 , where V0 is the initial volume at
pressure p .
volume in which the pressure changes, the point of reaction equilibrium is not the point of
minimum G. Instead, the point of reaction equilibrium in this case is at the minimum of the
Helmholtz energy A (Sec. 11.7.5).
11.8 THE THERMODYNAMIC EQUILIBRIUM CONSTANT

11.8.1 Activities and the denition of K
Equation 10.1.9 gives the general relation between the chemical potential
ai of species i in a phase of electric potential :
i
C RT ln ai C zi F
and the activity
(11.8.1)
The electric potential affects i only if the charge number zi is nonzero, i.e., only if species
i is an ion.
Consider a reaction in which any reactants and products that are ions are in a single
phase of electric potential 0 , or in several phases of equal electric potential 0 . Under these

P
conditions, substitution of the expression above for i in r G D i i
X
X
X
r G D
C RT
ln ai C F 0
i i
i
i zi
i
351
i
gives
(11.8.2)
(all ions at D 0 )
The rst term on the right side of Eq. 11.8.2 is the standard molar reaction Gibbs energy,
or standard molar Gibbs energy of reaction:
def
r G D
(11.8.3)
Since the standard chemical potential of each species i is a function only of T , the value
i
of r G for a given reaction as dened by the reaction equation depends only on T and on
the choice of a standard state for each reactant and product. P
The last term on the right side of Eq. 11.8.2 is the sum i i zi . Because charge is
conserved during the advancement of a reaction in a closed system, this sum is zero.
X
r G D r G C RT
(11.8.4)
i ln ai
(all ions at same )
This relation enables us to say that for a reaction at a given temperature in which any charged
reactants or products are all in the same phase, or in phases of equal electric potential, the
P
value of r G and i i i depends only on the activities of the reactants and products and
is independent of what the electric potentials of any of the phases might happen to be.
Unless a reaction involving ions is carried out in a galvanic cell, the ions are usually
present in a single phase, and this will not be shown as a condition of validity in the rest of
this chapter. The special case of a reaction in a galvanic cell will be discussed in Sec. 14.3.
We may use properties of logarithms to write the sum on the right side of Eq. 11.8.4 as
follows:18
X
X
Y
ln ai D
ln ai i D ln
ai i
(11.8.5)
i
i
The product
ai i is called the reaction quotient or activity quotient, Qrxn :

def
Qrxn D
ai i
(11.8.6)
Qrxn consists of a factor for each reactant and product. Each factor is the activity raised to
the power of the stoichiometric number i . Since the value of i is positive for a product and
negative for a reactant, Qrxn is a quotient in which the activities of the products appear in the
numerator and those of the reactants appear in the denominator, with each activity raised to
a power equal to the corresponding stoichiometric coefcient in the reaction equation. Such
a quotient, with quantities raised to these powers, is called a proper quotient. The reaction
quotient is a proper quotient of activities.
Q
Q
symbol stands for a continued product. If, for instance, there are three species, i ai i is the product
.a11 /.a22 /.a33 /.
18 The

352
For instance, for the ammonia synthesis reaction N2 (g) + 3 H2 (g) ! 2 NH3 (g) the reaction quotient is given by
2
aNH3
(11.8.7)
Qrxn D
3
aN2 aH2
Qrxn is a dimensionless quantity. It is a function of T , p, and the mixture composition, so
its value changes as the reaction advances.
The expression for the molar reaction Gibbs energy given by Eq. 11.8.4 can now be
written
r G D r G C RT ln Qrxn
(11.8.8)
The value of Qrxn under equilibrium conditions is the thermodynamic equilibrium constant, K. The general denition of K is
def
K D
i
.ai /eq
(11.8.9)
where the subscript eq indicates an equilibrium state. Note that K, like Qrxn , is dimensionless.
The IUPAC Green Book19 gives K as an alternative symbol for the thermodynamic
equilibrium constant, the appended superscript denoting standard. An IUPAC Commission on Thermodynamics20 has furthermore recommended the name standard
equilibrium constant, apparently because its value depends on the choice of standard states. Using this alternative symbol and name could cause confusion, since the
quantity dened by Eq. 11.8.9 does not refer to reactants and products in their standard
states but rather to reactants and products in an equilibrium state.
Substituting the equilibrium conditions r GD0 and Qrxn DK in Eq. 11.8.8 gives an important relation between the standard molar reaction Gibbs energy and the thermodynamic
equilibrium constant:
r G D RT ln K
(11.8.10)
We can solve this equation for K to obtain the equivalent relation
r G
K D exp
RT
(11.8.11)
We have seen that the value of r G depends only on T and the choice of the standard
states of the reactants and products. This being so, Eq. 11.8.11 shows that the value of K for
a given reaction depends only on T and the choice of standard states. No other condition,
neither pressure nor composition, can affect the value of K. We also see from Eq. 11.8.11
that K is less than 1 if r G is positive and greater than 1 if r G is negative. At a xed
temperature, reaction equilibrium is attained only if and only if the value of Qrxn becomes
equal to the value of K at that temperature.
The thermodynamic equilibrium constant K is the proper quotient of the activities of
species in reaction equilibrium. At typical temperatures and pressures, an activity cannot be
19 Ref.
20 Ref.
[36], p. 58.
[52].

353
many orders of magnitude greater than 1. For instance, a partial pressure cannot be greater
than the total pressure, so at a pressure of 10 bar the activity of a gaseous constituent cannot
be greater than about 10. The molarity of a solute is rarely much greater than 10 mol dm 3 ,
corresponding to an activity (on a concentration basis) of about 10. Activities can, however,
be extremely small.
These considerations lead us to the conclusion that in an equilibrium state of a reaction
with a very large value of K, the activity of at least one of the reactants must be very small.
That is, if K is very large then the reaction goes practically to completion and at equilibrium
a limiting reactant is essentially entirely exhausted. The opposite case, a reaction with a very
small value of K, must have at equilibrium one or more products with very small activities.
These two cases are the two extremes of the trends shown in Fig. 11.16 on page 349.
Equation 11.8.10 correctly relates r G and K only if they are both calculated with the
same standard states. For instance, if we base the standard state of a particular solute species
on molality in calculating r G , the activity of that species appearing in the expression for
K (Eq. 11.8.9) must also be based on molality.
11.8.2 Reaction in a gas phase

If a reaction takes place in a gaseous mixture, the standard state of each reactant and product
is the pure gas behaving ideally at the standard pressure p (Sec. 9.3.3). In this case,
each activity is given by ai (g) D fi =p D i pi =p where i is a fugacity coefcient
(Table 9.5). When we substitute this expression into Eq. 11.8.9, we nd we can express the
thermodynamic equilibrium constant as the product of three factors:
"
#"
#
h
i
P
Y
Y
i
i
(11.8.12)
KD
. i /eq
.pi /eq .p / i i
i
(gas mixture)
On the right side of this equation, the rst factor is the proper quotient of fugacity coefcients in the mixture at reaction equilibrium, the second factor is the proper quotient of
partial pressures in this mixture, and the third factor is the power of p needed to make K
dimensionless.
The proper quotient of equilibrium partial pressures is an equilibrium constant on a
pressure basis, Kp :
Kp D
Y
i
i
.pi /eq
(11.8.13)
(gas mixture)
P
Note that Kp is dimensionless only if i i is equal to zero.
The value of Kp can vary at constant temperature, so Kp is not a thermodynamic equilibrium constant. For instance, consider what happens when we take an ideal gas mixture
at reaction equilibrium and compress it isothermally. As the gas pressure increases, the fugacity coefcient of each constituent changes from its low pressure value of 1 and the gas
mixture becomes nonideal. In order for the mixture to remain in reaction equilibrium, and
the product of factors on the right side of Eq. 11.8.12 to remain constant, there must be a
change in the value of Kp . In other words, the reaction equilibrium shifts as we increase p
at constant T , an effect that will be considered in more detail in Sec. 11.9.

354
As an example of the difference between K and KpP

, consider again the ammonia synthesis N2 .g/ C 3 H2 .g/ ! 2 NH3 .g/ in which the sum i i equals 2. For this reaction,
the expression for the thermodynamic equilibrium constant is
!
2
KD
NH3
3
N2 H2
Kp .p /2
eq
where Kp is given by
Kp D
(11.8.14)
2
pNH3
3
pN2 pH2
(11.8.15)
eq
11.8.3 Reaction in solution

If any of the reactants or products are solutes in a solution, the value of K depends on the
choice of the solute standard state.
For a given reaction at a given temperature, we can derive relations between values of
K that are based on different solute standard states. In the limit of innite dilution, each
solute activity coefcient is unity, and at the standard pressure each pressure factor is unity.
Under these conditions of innite dilution and standard pressure, the activities of solute B
on a mole fraction, concentration, and molality basis are therefore
ax;B D xB
ac;B D cB =c
am;B D mB =m
(11.8.16)
In the limit of innite dilution, the solute composition variables approach values given by
the relations in Eq. 9.1.14 on page 226: xB D VA cB D MA mB , Combining these with
ax;B D xB from Eq. 11.8.16, we write
ax;B D VA cB D MA mB
(11.8.17)
Then, using the relations for ac;B and am;B in Eq. 11.8.16, we nd that the activities of
solute B at innite dilution and pressure p are related by
ax;B D VA c ac;B D MA m am;B
(11.8.18)
Q
i
B
The expression K D i .ai /eq has a factor .aB /eq for each solute B that is a reactant
or product. From Eq. 11.8.18, we see that for solutes at innite dilution at pressure p , the
relations between the values of K based on different solute standard states are
Y
Y
K(x basis) D
.VA c / B K(c basis) D
.MA m / B K(m basis)
(11.8.19)
B
For a given reaction at a given temperature, and with a given choice of solute standard state,
the value of K is not affected by pressure or dilution. The relations of Eq. 11.8.19 are
therefore valid under all conditions.

355
11.8.4 Evaluation of K
The relation K D exp . r G =RT / (Eq. 11.8.11) gives us a way to evaluate the thermodynamic equilibrium constant K of a reaction at a given temperature from the value of the
standard molar reaction Gibbs energy r G at that temperature. If we know the value of
r G , we can calculate the value of K.
One method is to calculate r G from values of the standard molar Gibbs energy of
formation f G of each reactant and product. These values are the standard molar reaction
Gibbs energies for the formation reactions of the substances. To relate f G to measurable
P
quantities, we make the substitution i D Hi T Si (Eq. 9.2.46) in r G D i i i to
P
P
give r G D i i Hi T i i Si , or
r G D r H
Tr S
(11.8.20)
When we apply this equation to a reaction with each reactant and product in its standard
state, it becomes
r G D r H Tr S
(11.8.21)
where the standard molar reaction entropy is given by
X
r S D
i Si
(11.8.22)
If the reaction is the formation reaction of a substance, we have

X
f G D f H T
i Si
(11.8.23)
where the sum over i is for the reactants and product of the formation reaction. We can evaluate the standard molar Gibbs energy of formation of a substance, then, from its standard
molar enthalpy of formation and the standard molar entropies of the reactants and product.
Extensive tables are available of values of f G for substances and ions. An abbreviated version at the single temperature 298:15 K is given in Appendix H. For a reaction of
interest, the tabulated values enable us to evaluate r G , and then K, from the expression
(analogous to Hesss law)
X
r G D
(11.8.24)
i f G .i/
i
The sum over i is for the reactants and products of the reaction of interest.
Recall that the standard molar enthalpies of formation needed in Eq. 11.8.23 can be
evaluated by calorimetric methods (Sec. 11.3.2). The absolute molar entropy values Si
come from heat capacity data or statistical mechanical theory by methods discussed in Sec.
6.2. Thus, it is entirely feasible to use nothing but calorimetry to evaluate an equilibrium
constant, a goal sought by thermodynamicists during the rst half of the 20th century.21
For ions in aqueous solution, the values of Sm and f G found in Appendix H are based
D0 and G D0 for HC (aq) at all temperatures, similar to the
on the reference values Sm
f
convention for f H values discussed in Sec. 11.3.2.22 For a reaction with aqueous ions
21 Another method, for a reaction that can be carried out reversibly in a galvanic cell, is described in Sec. 14.3.3.
that the values of Sm in Appendix H for some ions, unlike the values for substances, are negative; this
simply means that the standard molar entropies of these ions are less than that of HC (aq).
22 Note

11.9 E FFECTS OF T EMPERATURE AND P RESSURE ON E QUILIBRIUM P OSITION
356
as reactants or products, these values correctly give r S using Eq. 11.8.22, or r G using
Eq. 11.8.24.
The relation of Eq. 11.8.23 does not apply to an ion, because we cannot write a for
mation reaction for a single ion. Instead, the relation between f G , f H and Sm is
more complicated.
Consider rst a hypothetical reaction in which hydrogen ions and one or more
elements form H2 and a cation MzC with charge number zC :
zC HC (aq) C elements ! .zC =2/H2 (g) C MzC (aq)
For this reaction, using the convention that f H , Sm , and f G are zero for the
C
aqueous H ion and the fact that f H and f G are zero for the elements, we can
write the following expressions for standard molar reaction quantities:
r H D f H .MzC /
r S D .zC =2/Sm .H2 / C Sm .MzC /
(11.8.25)
Si
(11.8.26)
elements
r G D f G .MzC /
Then, from r G D r H
Tr S , we nd
f G .MzC / D f H .MzC /
"
T
(11.8.27)
Sm .MzC /
Si
elements
.zC =2/Sm .H2 /
(11.8.28)
For example, the standard molar Gibbs energy of the aqueous mercury(I) ion is found
from
f G .Hg2 2C / D f H .Hg2 2C /
C 2T Sm .Hg/
T Sm .Hg2 2C /
2
T Sm .H2 /
2
(11.8.29)
For an anion Xz with negative charge number z , using the hypothetical reaction
jz =2j H2 (g) C elements ! jz j HC (aq) C Xz (aq)
we nd by the same method
f G .Xz / D f H .Xz /
"
T
Sm .Xz
Si
elements
jz
=2j Sm .H2 /
(11.8.30)
For example, the calculation for the nitrate ion is

f G .NO3 / D f H .NO3 /
T Sm .NO3 /
C 1 T Sm .N2 / C 3 T Sm .O2 / C 2 T Sm .H2 /

2
2
(11.8.31)
11.9 EFFECTS OF TEMPERATURE AND PRESSURE ON

EQUILIBRIUM POSITION
The advancement of a chemical reaction in a closed system describes the changes in the
amounts of the reactants and products from specied initial values of these amounts. We

357
have seen that if the system is maintained at constant temperature and pressure, changes
spontaneously in the direction that decreases the Gibbs energy. The change continues until
the system reaches a state of reaction equilibrium at the minimum of G. The value of
the advancement in this equilibrium state will be denoted eq , as shown in Fig. 11.15 on
page 346. The value of eq depends in general on the values of T and p. Thus when we
change the temperature or pressure of a closed system that is at equilibrium, eq usually
changes also and the reaction spontaneously shifts to a new equilibrium position.
To investigate this effect, we write the total differential of G with T , p, and as independent variables
dG D S dT C V dp C r G d
(11.9.1)
and obtain the reciprocity relations
@S
@r G
D
@T
@ T;p
p;
@r G
@p
T;
@V
@
(11.9.2)
T;p
We recognize the partial derivative on the right side of each of these relations as a molar
differential reaction quantity:
@r G
@r G
D r S
D r V
(11.9.3)
@T
@p T;
p;
We use these expressions for two of the coefcients in an expression for the total differential
of r G:
@r G
dr G D r S dT C r V dp C
d
(11.9.4)
@
T;p
(closed system)
Since r G is the partial derivative of G with respect to at constant T and p, the coefcient
.@r G=@ /T;p is the partial second derivative of G with respect to :
@r G
@
T;p
@2 G
@ 2
(11.9.5)
T;p
We know that at a xed T and p, a plot of G versus has a slope at each point equal to r G
and a minimum at the position of reaction equilibrium where is eq . At the minimum of
the plotted curve, the slope r G is zero and the second derivative is positive (see Fig. 11.15
on page 346). By setting r G equal to zero in the general relation r G D r H Tr S ,
we obtain the equation r S D r H=T which is valid only at reaction equilibrium where
equals eq . Making this substitution in Eq. 11.9.4, and setting dr G equal to zero and d
equal to d eq , we obtain
0D
r H
dT C r V dp C
T
which shows how innitesimal changes in T , p, and
@2 G
@ 2
eq
eq
T;p
(11.9.6)
(closed system)
are related.

Now we are ready to see how

for d eq gives
eq
eq
358
is affected by changes in T or p. Solving Eq. 11.9.6
r H
dT r V dp
D T
@2 G
@ 2 T;p
(11.9.7)
(closed system)
The right side of Eq. 11.9.7 is the expression for the total differential of in a closed
system at reaction equilibrium, with T and p as the independent variables. Thus, at constant
pressure the equilibrium shifts with temperature according to
@ eq
H
D r
(11.9.8)
@T p
@2 G
(closed system)
T
@ 2 T;p
and at constant temperature the equilibrium shifts with pressure according to
@ eq
V
D r
(11.9.9)
@p T
@2 G
(closed system)
@ 2 T;p
Because the partial second derivative .@2 G=@ 2 /T;p is positive, Eqs. 11.9.8 and 11.9.9
show that .@ eq =@T /p and r H have the same sign, whereas .@ eq =@p/T and r V have
opposite signs.
These statements express the application to temperature and pressure changes of what
is known as Le Ch teliers principle: When a change is made to a closed system at equiliba
rium, the equilibrium shifts in the direction that tends to oppose the change. Here are two
examples.
1. Suppose r H is negativethe reaction is exothermic. Since .@ eq =@T /p has the
same sign as r H , an increase in temperature causes eq to decrease: the equilibrium
shifts to the left. This is the shift that would reduce the temperature if the reaction
were adiabatic.
2. If r V is positive, the volume increases as the reaction proceeds to the right at constant T and p. .@ eq =@p/T has the opposite sign, so if we increase the pressure
isothermally by reducing the volume, the equilibrium shifts to the left. This is the
shift that would reduce the pressure if the reaction occurred at constant T and V .
It is easy to misuse or to be misled by Le Ch teliers principle. Consider the solution
a
process B (s) ! B(sln) for which .@ eq =@T /p , the rate of change of solubility with T , has
the same sign as the molar differential enthalpy of solution sol H at saturation. The sign
of sol H at saturation may be different from the sign of the molar integral enthalpy of
solution, Hm(sol). This is the situation for the dissolution of sodium acetate shown in
Fig. 11.9 on page 327. The equilibrium position (saturation) with one kilogram of water
is at sol
15 mol, indicated in the gure by an open circle. At this position, sol H is
positive and Hm(sol) is negative. So, despite the fact that the dissolution of 15 moles

359
of sodium acetate in one kilogram of water to form a saturated solution is an exothermic

process, the solubility of sodium acetate actually increases with increasing temperature,
contrary to what one might predict from Le Ch teliers principle.23
a
Another kind of change for which Le Ch teliers principle gives an incorrect prediction
a
is the addition of an inert gas to a gas mixture of constant volume. Adding the inert gas
at constant V increases the pressure, but has little effect on the equilibrium position of a
gas-phase reaction regardless of the value of r V . This is because the inert gas affects the
activities of the reactants and products only slightly, and not at all if the gas mixture is ideal,
so there is little or no effect on the value of Qrxn . (Note that the dependence of eq on p
expressed by Eq. 11.9.9 does not apply to an open system.)
The rigorous criterion for the equilibrium position of a reaction is always the requirement that Qrxn must equal K or, equivalently, that r G must be zero.
23 Ref.
[20].

P ROBLEMS
360
PROBLEMS
in Appendix I.
11.1 Use values of f H and f G in Appendix H to evaluate the standard molar reaction enthalpy
and the thermodynamic equilibrium constant at 298:15 K for the oxidation of nitrogen to form
aqueous nitric acid:
1
N .g/
2 2
C 5 O2 .g/ C 1 H2 O.l/ ! HC .aq/ C NO3 .aq/

4
2
11.2 In 1982, the International Union of Pure and Applied Chemistry recommended that the value
of the standard pressure p be changed from 1 atm to 1 bar. This change affects the values of
some standard molar quantities of a substance calculated from experimental data.
(a) Find the changes in Hm , Sm , and Gm for a gaseous substance when the standard pressure
is changed isothermally from 1:01325 bar (1 atm) to exactly 1 bar. (Such a small pressure
change has an entirely negligible effect on these quantities for a substance in a condensed
phase.)
(b) What are the values of the corrections that need to be made to the standard molar enthalpy
of formation, the standard molar entropy of formation, and the standard molar Gibbs
energy of formation of N2 O4 (g) at 298:15 K when the standard pressure is changed from
1:01325 bar to 1 bar?
11.3 From data for mercury listed in Appendix H, calculate the saturation vapor pressure of liquid
mercury at both 298:15 K and 273:15 K. You may need to make some reasonable approximations.
11.4 Given the following experimental values at T D 298:15 K, p D 1 bar:
HC (aq) C OH (aq) ! H2 O(l)
1
Na(s) C H2 O(l) ! NaC (aq/ C OH (aq) C 2 H2 (g)
NaOH(s) ! NaOH(aq)
NaOH in 5 H2 O ! NaOH in 1 H2 O
NaOH(s)
r H D
55:82 kJ mol
r H D 184:52 kJ mol 1
sol H 1 D 44:75 kJ mol 1
Hm (dil) D 4:93 kJ mol 1
f H D 425:61 kJ mol 1
Using only these values, calculate:

(a) f H for NaC (aq), NaOH(aq), and OH (aq);
(b) f H for NaOH in 5 H2 O;
(c) Hm(sol) for the dissolution of 1 mol NaOH(s) in 5 mol H2 O.

P ROBLEMS
361
Table 11.1 Data for Problem 11.5 a

f H /kJ mol
Substance
285:830
2607:93
1135:914
1133:822
1132:236
1131:780
H2 O(l)
Na2 S2 O3 5H2 O.s/
Na2 S2 O3 in 50 H2 O
Na2 S2 O3 in 100 H2 O
Na2 S2 O3 in 200 H2 O
Na2 S2 O3 in 300 H2 O
a Ref.
M /g mol
18:0153
248:1828
[163], pages 2-307 and 2-308.
11.5 Table 11.1 lists data for water, crystalline sodium thiosulfate pentahydrate, and several sodium
thiosulfate solutions. Find H to the nearest 0:01 kJ for the dissolution of 5:00 g of crystalline
Na2 S2 O3 5H2 O in 50:0 g of water at 298:15 K and 1 bar.
Table 11.2 Data for Problem 11.6. Molar integral enthalpies of dilution of aqueous HCl
(m0 D 3:337 mol kg 1 ) at 25 C.a
B
m00 =mol kg
B
0:295
0:225
0:199
0:147
0:113
0:0716
0:0544
0:0497
0:0368
0:0179
0:0128
a Ref.
Hm .dil, m0 !m00 /=kJ mol

B
B
2:883
2:999
3:041
3:143
3:217
3:325
3:381
3:412
3:466
3:574
3:621
[153].
11.6 Use the experimental data in Table 11.2 to evaluate LA and LB at 25 C for an aqueous HCl
solution of molality mB D 0:0900 mol kg 1 .

P ROBLEMS
362
Table 11.3 Data for Problem 11.7. The values of intensive properties are
for a temperature of 298:15 K and a pressure of 30 bar unless otherwise
stated. Subscripts: A = H2 O, B = O2 , C = CO2 .
Properties of the bomb vessel:
internal volume . . . . . . . . . . . . . . . . . . . . . . . 350:0 cm3
mass of n-hexane placed in bomb . . . . . . . 0:6741 g
mass of water placed in bomb . . . . . . . . . . 1:0016 g
Properties of liquid n-hexane:
molar mass . . . . . . . . . . . . . . . . . . . . . . . . . . . M D 86:177 g mol 1
density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D 0:6548 g cm 3
cubic expansion coefcient . . . . . . . . . . . . . D 1:378 10 3 K 1
Properties of liquid H2 O:
molar mass . . . . . . . . . . . . . . . . . . . . . . . . . . . M D 18:0153 g mol 1
density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D 0:9970 g cm 3
cubic expansion coefcient . . . . . . . . . . . . . D 2:59 10 4 K 1
standard molar energy of vaporization . . . vap U D 41:53 kJ mol 1
Second virial coefcients, 298:15 K:
BAA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1158 cm3 mol 1
BBB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 cm3 mol 1
dBBB = dT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0:21 cm3 K 1 mol 1
BCC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 cm3 mol 1
dBCC = dT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0:97 cm3 K 1 mol 1
BAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 cm3 mol 1
BAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 cm3 mol 1
BBC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43:7 cm3 mol 1
dBBC = dT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0:4 cm3 K 1 mol 1
Henrys law constants at 1 bar (solvent = H2 O):
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . km;B D 796 bar kg mol 1
CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . km;C D 29:7 bar kg mol 1
Partial molar volumes of solutes in water:
1
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VB D 31 cm3 mol 1
1
CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VC D 33 cm3 mol 1
Standard molar energies of solution (solvent = H2 O):
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . sol U D 9:7 kJ mol 1
CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . sol U D 17:3 kJ mol 1
11.7 This 16-part problem illustrates the use of experimental data from bomb calorimetry and other
sources, combined with thermodynamic relations derived in this and earlier chapters, to evaluate the standard molar combustion enthalpy of a liquid hydrocarbon. The substance under
investigation is n-hexane, and the combustion reaction in the bomb vessel is
C6 H14 .l/ C
19
O .g/
2 2
! 6 CO2 .g/ C 7 H2 O.l/
Assume that the sample is placed in a glass ampoule that shatters at ignition. Data needed for
this problem are collected in Table 11.3.
States 1 and 2 referred to in this problem are the initial and nal states of the isothermal bomb
process. The temperature is the reference temperature of 298:15 K.

P ROBLEMS
363
(a) Parts (a)(c) consist of simple calculations of some quantities needed in later parts of the
problem. Begin by using the masses of C6 H14 and H2 O placed in the bomb vessel, and
their molar masses, to calculate the amounts (moles) of C6 H14 and H2 O present initially
in the bomb vessel. Then use the stoichiometry of the combustion reaction to nd the
amount of O2 consumed and the amounts of H2 O and CO2 present in state 2. (There is
not enough information at this stage to allow you to nd the amount of O2 present, just
the change.) Also nd the nal mass of H2 O. Assume that oxygen is present in excess
and the combustion reaction goes to completion.
(b) From the molar masses and the densities of liquid C6 H14 and H2 O, calculate their molar
volumes.
(c) From the amounts present initially in the bomb vessel and the internal volume, nd the
volumes of liquid C6 H14 , liquid H2 O, and gas in state 1 and the volumes of liquid H2 O
and gas in state 2. For this calculation, you can neglect the small change in the volume of
liquid H2 O due to its vaporization.
(d) When the bomb vessel is charged with oxygen and before the inlet valve is closed, the
pressure at 298:15 K measured on an external gauge is found to be p1 D 30:00 bar. To a
good approximation, the gas phase of state 1 has the equation of state of pure O2 (since
the vapor pressure of water is only 0:1 % of 30:00 bar). Assume that this equation of state
is given by Vm D RT =p C BBB (Eq. 2.2.8), where BBB is the second virial coefcient
of O2 listed in Table 11.3. Solve for the amount of O2 in the gas phase of state 1. The
gas phase of state 2 is a mixture of O2 and CO2 , again with a negligible partial pressure
of H2 O. Assume that only small fractions of the total amounts of O2 and CO2 dissolve
in the liquid water, and nd the amount of O2 in the gas phase of state 2 and the mole
fractions of O2 and CO2 in this phase.
(e) You now have the information needed to nd the pressure in state 2, which cannot be measured directly. For the mixture of O2 and CO2 in the gas phase of state 2, use Eq. 9.3.23
on page 245 to calculate the second virial coefcient. Then solve the equation of state of
Eq. 9.3.21 on page 245 for the pressure. Also calculate the partial pressures of the O2 and
CO2 in the gas mixture.
(f) Although the amounts of H2 O in the gas phases of states 1 and 2 are small, you need to
know their values in order to take the energy of vaporization into account. In this part,
you calculate the fugacities of the H2 O in the initial and nal gas phases, in part (g) you
use gas equations of state to evaluate the fugacity coefcients of the H2 O (as well as of
the O2 and CO2 ), and then in part (h) you nd the amounts of H2 O in the initial and nal
gas phases.
The pressure at which the pure liquid and gas phases of H2 O are in equilibrium at 298:15 K
(the saturation vapor pressure of water) is 0:03169 bar. Use Eq. 7.8.18 on page 186 to estimate the fugacity of H2 O(g) in equilibrium with pure liquid water at this temperature
and pressure. The effect of pressure on fugacity in a one-component liquidgas system is
discussed in Sec. 12.8.1; use Eq. 12.8.3 on page 401 to nd the fugacity of H2 O in gas
phases equilibrated with liquid water at the pressures of states 1 and 2 of the isothermal
bomb process. (The mole fraction of O2 dissolved in the liquid water is so small that you
can ignore its effect on the chemical potential of the water.)
(g) Calculate the fugacity coefcients of H2 O and O2 in the gas phase of state 1 and of H2 O,
O2 , and CO2 in the gas phase of state 2.
For state 1, in which the gas phase is practically-pure O2 , you can use Eq. 7.8.18 on
page 186 to calculate O2 . The other calculations require Eq. 9.3.29 on page 246, with
the value of Bi0 found from the formulas of Eq. 9.3.26 or Eqs. 9.3.27 and 9.3.28 (yA is so
small that you can set it equal to zero in these formulas).

P ROBLEMS
364
Use the fugacity coefcient and partial pressure of O2 to evaluate its fugacity in states 1
and 2; likewise, nd the fugacity of CO2 in state 2. [You calculated the fugacity of the
H2 O in part (f).]
(h) From the values of the fugacity and fugacity coefcient of a constituent of a gas mixture,
you can calculate the partial pressure with Eq. 9.3.17 on page 244, then the mole fraction
with yi D pi =p, and nally the amount with ni D yi n. Use this method to nd the
amounts of H2 O in the gas phases of states 1 and 2, and also calculate the amounts of
H2 O in the liquid phases of both states.
(i) Next, consider the O2 dissolved in the water of state 1 and the O2 and CO2 dissolved in
the water of state 2. Treat the solutions of these gases as ideal dilute with the molality of
solute i given by mi D fi =km;i (Eq. 9.4.21). The values of the Henrys law constants of
these gases listed in Table 11.3 are for the standard pressure of 1 bar. Use Eq. 12.8.35 on
page 409 to nd the appropriate values of km;i at the pressures of states 1 and 2, and use
these values to calculate the amounts of the dissolved gases in both states.
(j) At this point in the calculations, you know the values of all properties needed to describe
the initial and nal states of the isothermal bomb process. You are now able to evaluate
the various Washburn corrections. These corrections are the internal energy changes, at
the reference temperature of 298:15 K, of processes that connect the standard states of
substances with either state 1 or state 2 of the isothermal bomb process.
First, consider the gaseous H2 O. The Washburn corrections should be based on a pureliquid standard state for the H2 O. Section 7.9 shows that the molar internal energy of a
pure gas under ideal-gas conditions (low pressure) is the same as the molar internal energy
of the gas in its standard state at the same temperature. Thus, the molar internal energy
change when a substance in its pure-liquid standard state changes isothermally to an ideal
gas is equal to the standard molar internal energy of vaporization, vap U . Using the
value of vap U for H2 O given in Table 11.3, calculate U for the vaporization of liquid
H2 O at pressure p to ideal gas in the amount present in the gas phase of state 1. Also
calculate U for the condensation of ideal gaseous H2 O in the amount present in the gas
phase of state 2 to liquid at pressure p .
(k) Next, consider the dissolved O2 and CO2 , for which gas standard states are used. Assume
that the solutions are sufciently dilute to have innite-dilution behavior; then the partial
molar internal energy of either solute in the solution at the standard pressure p D 1 bar
is equal to the standard partial molar internal energy based on a solute standard state (Sec.
9.7.1). Values of sol U are listed in Table 11.3. Find U for the dissolution of O2 from
its gas standard state to ideal-dilute solution at pressure p in the amount present in the
aqueous phase of state 1. Find U for the desolution (transfer from solution to gas phase)
of O2 and of CO2 from ideal-dilute solution at pressure p , in the amounts present in the
aqueous phase of state 2, to their gas standard states.
(l) Calculate the internal energy changes when the liquid phases of state 1 (n-hexane and
aqueous solution) are compressed from p to p1 and the aqueous solution of state 2 is
decompressed from p2 to p . Use an approximate expression from Table 7.4, and treat
the cubic expansion coefcient of the aqueous solutions as being the same as that of pure
water.
(m) The nal Washburn corrections are internal energy changes of the gas phases of states 1
and 2. H2 O has such low mole fractions in these phases that you can ignore H2 O in these
calculations; that is, treat the gas phase of state 1 as pure O2 and the gas phase of state 2
as a binary mixture of O2 and CO2 .
One of the internal energy changes is for the compression of gaseous O2 , starting at a
pressure low enough for ideal-gas behavior (Um D Um ) and ending at pressure p1 to form

P ROBLEMS
365
the gas phase present in state 1. Use the approximate expression for Um Um (g) in Table
7.5 to calculate U D U.p1 / nUm (g); a value of dB= dT for pure O2 is listed in Table
11.3.
The other internal energy change is for a process in which the gas phase of state 2 at
pressure p2 is expanded until the pressure is low enough for the gas to behave ideally,
and the mixture is then separated into ideal-gas phases of pure O2 and CO2 . The molar
internal energies of the separated low-pressure O2 and CO2 gases are the same as the
standard molar internal energies of these gases. The internal energy of unmixing ideal
gases is zero (Eq. 11.1.11). The dependence of the internal energy of the gas mixture is
P
given, to a good approximation, by U D i Ui (g) npT dB= dT , where B is the second
virial coefcient of the gas mixture; this expression is the analogy for a gas mixture of the
approximate expression for Um Um (g) in Table 7.5. Calculate the value of dB= dT for
the mixture of O2 and CO2 in state 2 (you need Eq. 9.3.23 on page 245 and the values of
P
dBij = dT in Table 11.3) and evaluate U D i ni Ui (g) U.p2 / for the gas expansion.
(n) Add the internal energy changes you calculated in parts (j)(m) to nd the total internal
energy change of the Washburn corrections. Note that most of the corrections occur in
pairs of opposite sign and almost completely cancel one another. Which contributions are
the greatest in magnitude?
(o) The internal energy change of the isothermal bomb process in the bomb vessel, corrected
to the reference temperature of 298:15 K, is found to be U.IBP; Tref / D 32:504 kJ.
Assume there are no side reactions or auxiliary reactions. From Eqs. 11.5.9 and 11.5.10,
calculate the standard molar internal energy of combustion of n-hexane at 298:15 K.
(p) From Eq. 11.5.13, calculate the standard molar enthalpy of combustion of n-hexane at
298:15 K.
11.8 By combining the results of Prob. 11.7(p) with the values of standard molar enthalpies of
formation from Appendix H, calculate the standard molar enthalpy of formation of liquid nhexane at 298:15 K.
11.9 Consider the combustion of methane:
CH4 .g/ C 2 O2 .g/ ! CO2 .g/ C 2 H2 O.g/
Suppose the reaction occurs in a owing gas mixture of methane and air. Assume that the
pressure is constant at 1 bar, the reactant mixture is at a temperature of 298:15 K and has
stoichiometric proportions of methane and oxygen, and the reaction goes to completion with
no dissociation. For the quantity of gaseous product mixture containing 1 mol CO2 , 2 mol H2 O,
and the nitrogen and other substances remaining from the air, you may use the approximate
formula Cp .P/ D a C bT , where the coefcients have the values a D 297:0 J K 1 and b D
8:520 10 2 J K 2 . Solve Eq. 11.6.1 for T2 to estimate the ame temperature to the nearest
kelvin.
11.10 The standard molar Gibbs energy of formation of crystalline mercury(II) oxide at 600:00 K has
the value f G D 26:386 kJ mol 1 . Estimate the partial pressure of O2 in equilibrium with
HgO at this temperature: 2 HgO(s) 2 Hg(l) C O2 (g).
11.11 The combustion of hydrogen is a reaction that is known to go to completion.
(a) Use data in Appendix H to evaluate the thermodynamic equilibrium constant at 298:15 K
for the reaction
1
H2 .g/ C 2 O2 .g/ ! H2 O.l/

P ROBLEMS
366
(b) Assume that the reaction is at equilibrium at 298:15 K in a system in which the partial
pressure of O2 is 1:0 bar. Assume ideal-gas behavior and nd the equilibrium partial
pressure of H2 and the number of H2 molecules in 1:0 m3 of the gas phase.
(c) In the preceding part, you calculated a very small value (a fraction) for the number of H2
molecules in 1:0 m3 . Statistically, this fraction can be interpreted as the fraction of a given
length of time during which one molecule is present in the system. Take the age of the
universe as 1:0 1010 years and nd the total length of time in seconds, during the age of
the universe, that a H2 molecule is present in the equilibrium system. (This hypothetical
value is a dramatic demonstration of the statement that the limiting reactant is essentially
entirely exhausted during a reaction with a large value of K.)
11.12 Let G represent carbon in the form of graphite and D represent the diamond crystal form. At
298:15 K, the thermodynamic equilibrium constant for GD, based on a standard pressure
p D 1 bar, has the value K D 0:31. The molar volumes of the two crystal forms at this
temperature are Vm .G/ D 5:3 10 6 m3 mol 1 and Vm .D/ D 3:4 10 6 m3 mol 1 .
(a) Write an expression for the reaction quotient Qrxn as a function of pressure. Use the
approximate expression of the pressure factor given in Table 9.6.
(b) Use the value of K to estimate the pressure at which the D and G crystal forms are in
equilibrium with one another at 298:15 K. (This is the lowest pressure at which graphite
could in principle be converted to diamond at this temperature.)
11.13 Consider the dissociation reaction N2 O4 .g/ ! 2 NO2 .g/ taking place at a constant temperature of 298:15 K and a constant pressure of 0:0500 bar. Initially (at D 0) the system contains
1:000 mol of N2 O4 and no NO2 . Other needed data are found in Appendix H. Assume idealgas behavior.
(a) For values of the advancement ranging from 0 to 1 mol, at an interval of 0:1 mol or
less, calculate G. / G.0/ to the nearest 0:01 kJ. A computer spreadsheet would be a
convenient way to make the calculations.
(b) Plot your values of G. /
points.
G.0/ as a function of , and draw a smooth curve through the
(c) On your curve, indicate the estimated position of eq . Calculate the activities of N2 O4 and
NO2 for this value of , use them to estimate the thermodynamic equilibrium constant K,
and compare your result with the value of K calculated from Eq. 11.8.11.
C HAPTER 12
E QUILIBRIUM C ONDITIONS IN
M ULTICOMPONENT S YSTEMS
This chapter applies equilibrium theory to a variety of chemical systems of more than one
component. Two different approaches will be used as appropriate: one based on the relation
P
Q i
i i i D 0 or K D
i ai for
i D
i for transfer equilibrium, the other based on
reaction equilibrium.
12.1 EFFECTS OF TEMPERATURE

For some of the derivations in this chapter, we will need an expression for the rate at which
the ratio i =T varies with temperature in a phase of xed composition maintained at constant pressure. This expression leads, among other things, to an important relation between
the temperature dependence of an equilibrium constant and the standard molar reaction
enthalpy.
12.1.1 Variation of
i =T
with temperature
In a phase containing species i, either pure or in a mixture, the partial derivative of i =T

with respect to T at constant p and a xed amount of each species is given by1
@ . i =T /
1 @ i
i
D
(12.1.1)
@T
T @T p;fnig T 2
p;fnig
This equality comes from a purely mathematical operation; no thermodynamics is involved.
The partial derivative .@ i =@T /p;fnig is equal to Si (Eq. 9.2.48), so that Eq. 12.1.1 becomes
@ . i =T /
Si
T Si C i
i
D
D
(12.1.2)
@T
T
T2
T2
p;fnig
The further substitution
D Hi
T Si (Eq. 9.2.46) gives nally
@.
i =T /
@T
p;fnig
1 This
Hi
T2
(12.1.3)
relation is obtained from the formula d.uv/= dx D u.dv= dx/ C v.du= dx/ (Appendix E), where u is
1=T , v is i , and x is T .
367
CHAPTER 12 EQUILIBRIUM CONDITIONS IN MULTICOMPONENT SYSTEMS

12.1 E FFECTS OF T EMPERATURE
368
For a pure substance in a closed system, Eq. 12.1.3 when multiplied by the amount n
becomes
@ .G=T /
H
D
(12.1.4)
@T
T2
p
This is the GibbsHelmholtz equation.
12.1.2 Variation of
i =T
If we make the substitution

d.
with temperature
i D i
i =T /
C RT ln ai in Eq. 12.1.3 and rearrange, we obtain
Hi
@ ln ai
D
R
(12.1.5)
T2
@T p;fnig
dT
Because =T is a function only of T , its derivative with respect to T is itself a function

i
only of T . We can therefore use any convenient combination of pressure and composition
in the expression on the right side of Eq. 12.1.5 in order to evaluate d. =T /= dT at a given
i
temperature.
If species i is a constituent of a gas mixture, we take a constant pressure of the gas that
is low enough for the gas to behave ideally. Under these conditions Hi is the standard molar
enthalpy Hi (Eq. 9.3.7). In the expression for activity, ai (g) D i (g) i pi =p (Table 9.5),
the pressure factor i (g) is constant when p is constant, the fugacity coefcient i for the
ideal gas is unity, and pi =p D yi is constant at constant fnig, so that the partial derivative
@ ln ai (g)=@T p;fnig is zero.
For component i of a condensed-phase mixture, we take a constant pressure equal to
the standard pressure p , and a mixture composition in the limit given by Eqs. 9.5.20
9.5.24 in which the activity coefcient is unity. Hi is then the standard molar enthalpy Hi ,
and the activity is given by an expression in Table 9.5 with the pressure factor and activity
coefcient set equal to 1: ai Dxi , aA DxA , ax;B DxB , ac;B DcB =c , or am;B DmB =m .2 With
the exception of ac;B , these activities are constant as T changes at constant p and fnig.
Thus for a gas-phase species, or a species with a standard state based on mole fraction
or molality, @ ln ai (g)=@T p;fnig is zero and Eq. 12.1.5 becomes
d.
i =T /
dT
Hi
T2
(12.1.6)
(standard state not based
on concentration)
Equation 12.1.6, as the conditions of validity indicate, does not apply to a solute standard state based on concentration, except as an approximation. The reason is the volume
change that accompanies an isobaric temperature change. We can treat this case by considering the following behavior of ln.cB =c /:
@ ln.cB =c /
1 @cB
1
@.nB =V /
D
D
@T
cB @T p;fnig
nB =V
@T
p;fnig
p;fnig
@.1=V /
1 @V
DV
D
@T
V @T p;fnig
p;fnig
D
(12.1.7)
2 If
solute B is an electrolyte, am;B is given instead by Eq. 10.3.10; like am;B for a nonelectrolyte, it is constant
as T changes at constant p and fnig.

12.1 E FFECTS OF T EMPERATURE
369
Here is the cubic expansion coefcient of the solution (Eq. 7.1.1). If the activity coefcient is to be unity, the solution must be an ideal-dilute solution, and is then A , the cubic
expansion coefcient of the pure solvent. Eq. 12.1.5 for a nonelectrolyte becomes
d.
c;B =T /
dT
HB
C RA
T2
(12.1.8)
12.1.3 Variation of lnK with temperature

The thermodynamic equilibrium constant K, for a given reaction equation and a given
choice of reactant and product standard states, is a function of T and only of T . By equatP
ing two expressions for the standard molar reaction Gibbs energy, r G D i i and
i
r G D RT ln K (Eqs. 11.8.3 and 11.8.10), we obtain
1 X
RT
ln K D
(12.1.9)
The rate at which ln K varies with T is then given by

d ln K
D
dT
1 X
R
d.
i
i =T /
dT
(12.1.10)
P
Combining Eq. 12.1.10 with Eqs. 12.1.6 or 12.1.8, and recognizing that i i Hi is the
standard molar reaction enthalpy r H , we obtain the nal expression for the temperature
dependence of ln K:
X
d ln K
H
D r 2
A
(12.1.11)
i
dT
RT
solutes,
conc. basis
The sum on the right side includes only solute species whose standard states are based on
concentration. The expression is simpler if all solute standard states are based on mole
fraction or molality:
d ln K
H
D r 2
dT
RT
(12.1.12)
(no solute standard states
based on concentration)
We can rearrange Eq. 12.1.12 to

r H D RT 2
d ln K
dT
(12.1.13)
We can convert this expression for r H to an equivalent form by using the mathematical
identity d.1=T / D .1=T 2 / dT :
r H D
d ln K
d.1=T /
(12.1.14)

12.2 S OLVENT C HEMICAL P OTENTIALS FROM P HASE E QUILIBRIA
370
Table 12.1 Comparison of the ClausiusClapeyron and vant Hoff equations for vaporization of a liquid.
ClausiusClapeyron equation
d ln.p=p /
d.1=T /
Derivation assumes Vm (g)
ideal-gas behavior.
vap H
vant Hoff equation

d ln K
d.1=T /
An exact relation.
vap H D
Vm (l) and
vap H is the difference of the molar enthalpies of the real gas and the liquid at the
saturation vapor pressure of the liquid.
vap H is the difference of the molar enthalpies of the ideal gas and the liquid at
pressure p .
p is the saturation vapor pressure of the liquid.
K is equal to a(g)=a(l) D .f =p /= (l),

and is only approximately equal to p=p .
Equations 12.1.13 and 12.1.14 are two forms of the vant Hoff equation. They allow us
to evaluate the standard molar reaction enthalpy of a reaction by a noncalorimetric method
from the temperature dependence of ln K. For example, we can plot ln K versus 1=T ; then
according to Eq. 12.1.14, the slope of the curve at any value of 1=T is equal to r H =R
at the corresponding temperature T .
A simple way to derive the equation for this last procedure is to substitute r G D
r H Tr S in r G D RT ln K and rearrange to

r H 1
S
ln K D
C r
(12.1.15)
R
T
R
Suppose we plot ln K versus 1=T . In a small temperature interval in which r H and
r S are practically constant, the curve will appear linear. According to Eq. 12.1.15,
the curve in this interval has a slope of r H =R, and the tangent to a point on the
curve has its intercept at 1=T D0 equal to r S =R.
When we apply Eq. 12.1.14 to the vaporization process A(l)!A(g) of pure A, it resembles the ClausiusClapeyron equation for the same process (Eq. 8.4.15 on page 220). These
equations are not exactly equivalent, however, as the comparison in Table 12.1 shows.
12.2 SOLVENT CHEMICAL POTENTIALS FROM PHASE EQUILIBRIA

Section 9.6.3 explained how we can evaluate the activity coefcient m;B of a nonelectrolyte
solute of a binary solution if we know the variation of the osmotic coefcient of the solution
from innite dilution to the molality of interest. A similar procedure for the mean ionic
activity coefcient of an electrolyte solute was described in Sec. 10.6.
The physical measurements needed to nd the osmotic coefcient m of a binary solution must be directed to the calculation of the quantity A
A , the difference between the
chemical potentials of the pure solvent and the solvent in the solution at the temperature and
pressure of interest. This difference is positive, because the presence of the solute reduces
the solvents chemical potential.

371
To calculate m from A
A , we use Eq. 9.6.16 on page 268 for a nonelectrolyte
solute, or Eq. 10.6.1 on page 300 for an electrolyte solute. Both equations are represented
by
m
RTMA mB
(12.2.1)
where for a nonelectrolyte is 1 and for an electrolyte is the number of ions per formula
unit.
The sequence of steps, then, is (1) the determination of A
A over a range of molality
at constant T and p, (2) the conversion of these values to m using Eq. 12.2.1, and (3) the
evaluation of the solute activity coefcient3 by a suitable integration from innite dilution
to the molality of interest.
Sections 12.2.1 and 12.2.2 will describe freezing-point and osmotic-pressure measurements, two much-used methods for evaluating A
A in a binary solution at a given T and
p. The isopiestic vapor-pressure method was described in Sec. 9.6.4. The freezing-point
and isopiestic vapor-pressure methods are often used for electrolyte solutions, and osmotic
pressure is especially useful for solutions of macromolecules.
12.2.1 Freezing-point measurements

This section explains how we can evaluate A
A for a solution of a given composition
at a given T and p from the freezing point of the solution combined with additional data
obtained from calorimetric measurements.
Consider a binary solution of solvent A and solute B. We assume that when this solution
is cooled at constant pressure and composition, the solid that rst appears is pure A. For
example, for a dilute aqueous solution the solid would be ice. The temperature at which
solid A rst appears is Tf , the freezing point of the solution. This temperature is lower
than the freezing point Tf of the pure solvent, a consequence of the lowering of A by the
presence of the solute. Both Tf and Tf can be measured experimentally.
Let T 0 be a temperature of interest that is equal to or greater than Tf . We wish to
0
0
determine the value of A .l; T 0 /
A .sln; T /, where A .l; T / refers to pure liquid solvent
and A .sln; T 0 / refers to the solution.
Figure 12.1 on the next page explains the principle of the procedure. The gure shows
A =T for the solvent in the pure solid phase, in the pure liquid phase, and in the xedcomposition solution, plotted as functions of T at constant p. Since A is the same in the
solution and solid phases at temperature Tf , and is the same in the pure liquid and solid
phases at temperature Tf , the curves intersect at these temperatures as shown.
Formulas for the slopes of the three curves, from Eq. 12.1.3 on page 367, are included
0
0
in the gure. The desired value of A .l; T 0 /
A .sln; T / is the product of T and the
difference of the values of A =T at points e and a. To nd this difference, we integrate the
3A
measurement of A
state, through the relation
A
A
also gives us the solvent activity coefcient, based on the pure-solvent reference
D A C RT ln. A xA / (Eq. 9.5.15).

372
A
(s)
so
lid
slope
c
d
slope
ion
lut
so
A(l)
uid
liq
A (sln)
slope
Figure 12.1 Integration path abcde at constant pressure for determining A

A
at temperature T 0 from the freezing point Tf of a solution (schematic). The dashed
extensions of the curves represent unstable states.
slope d.
A =T /= dT
A .l; T
T0
0/
over T along the path abcde:

A .sln; T
T0
0/
Z
Z
Tf
T0
Tf
Tf
HA (sln)
dT
T2
HA (s)
dT
T2
Tf
Tf
T0
Tf
HA (sln)
dT
T2
HA (l)
dT
T2
(12.2.2)
By combining integrals that have the same range of integration, we turn Eq. 12.2.2 into
A .l; T
T0
0/
A .sln; T
T0
0/
Tf
HA (sln) HA (s)
dT
T2
Tf
T0
Tf
HA (sln) HA (l)
dT
T2
(12.2.3)
For convenience of notation, this book will use sol,A H to denote the molar enthalpy
difference HA (sln) HA (s). sol,A H is the molar differential enthalpy of solution of solid
A in the solution at constant T and p. The rst integral on the right side of Eq. 12.2.3
requires knowledge of sol,A H over a temperature range, but the only temperature at which
it is practical to measure this quantity calorimetrically is at the equilibrium transition temperature Tf . It is usually sufcient to assume sol,A H is a linear function of T :
sol,A H.T / D sol,A H.Tf / C sol,A Cp .T
Tf /
(12.2.4)
The molar differential heat capacity of solution sol,A Cp D Cp;A (sln) Cp;A (s) is treated
as a constant that can be determined from calorimetric measurements.
The quantity HA (sln) HA (l) in the second integral on the right side of Eq. 12.2.3 is the
molar differential enthalpy of dilution of the solvent in the solution, dil H (see Eq. 11.4.7).

373
(A + B)(sln)

A (l)
Figure 12.2 Apparatus to measure osmotic pressure (schematic). The dashed line
represents a membrane permeable only to the solvent A. The cross-hatched rectangles
represent moveable pistons.
This quantity can be measured calorimetrically at any temperature higher than Tf . Making
this substitution in Eq. 12.2.3 together with that of Eq. 12.2.4, carrying out the integration
of the rst integral and rearranging, we obtain nally
1
1
0
0
0
Tf sol,A Cp
A .sln; T / D T sol,A H.Tf /
A .l; T /
Tf Tf
Z T0
T
dil H
C T 0 sol,A Cp ln f C T 0
dT
(12.2.5)
Tf
T2
Tf
12.2.2 Osmotic-pressure measurements

A second method for evaluating A
A uses the solution property called osmotic pressure.
A simple apparatus to measure the osmotic pressure of a binary solution is shown schematically in Fig. 12.2. The system consists of two liquid phases separated by a semipermeable
membrane. Phase is pure solvent and phase is a solution with the same solvent at the
same temperature. The semipermeable membrane is permeable to the solvent and impermeable to the solute.
The presence of the membrane makes this system different from the multiphase, multicomponent system of Sec. 9.2.7, used there to derive conditions for transfer equilibrium.
By a modication of that procedure, we can derive the conditions of equilibrium for the
present system. We take phase as the reference phase because it includes both solvent and
solute. In order to prevent expansion work in the isolated system, both pistons shown in the
gure must be xed in stationary positions. This keeps the volume of each phase constant:
dV D dV D 0. Equation 9.2.41 on page 237, expressing the total differential of the
entropy in an isolated multiphase, multicomponent system, becomes
dS D
T
T
dS C
dn
A
(12.2.6)
In an equilibrium state, the coefcients .T T /=T and . A

A /=T must be zero.
Therefore, in an equilibrium state the temperature is the same in both phases and the solvent
has the same chemical potential in both phases. The presence of the membrane, however,
allows the pressures of the two phases to be unequal in the equilibrium state.

374
Suppose we start with both phases shown in Fig. 12.2 at the same temperature and
pressure. Under these conditions, the value of A is less in the solution than in the pure
liquid, and a spontaneous ow of solvent will occur through the membrane from the pure
solvent to the solution. This phenomenon is called osmosis.4 If we move the right-hand
piston down slightly in order to increase the pressure p 00 of the solution in phase , A
increases in this phase. The osmotic pressure of the solution, , is dened as the additional
pressure the solution must have, compared to the pressure p 0 of the pure solvent at the same
temperature, to establish an equilibrium state with no ow of solvent in either direction
through the membrane: p 00 D p 0 C .
In practice, the membrane may not be completely impermeable to a solute. All that is
required for the establishment of an equilibrium state with different pressures on either
side of the membrane is that solvent transfer equilibrium be established on a short time
scale compared to the period of observation, and that the amount of solute transferred
during this period be negligible.
The osmotic pressure is an intensive property of a solution whose value depends on

the solutions temperature, pressure, and composition. Strictly speaking, in an equilibrium state of the system shown in Fig. 12.2 refers to the osmotic pressure of the solution
at pressure p 0 , the pressure of the pure solvent. In other words, the osmotic pressure of a
solution at temperature T and pressure p 0 is the additional pressure that would have to be
exerted on the solution to establish transfer equilibrium with pure solvent that has temperature T and pressure p 0 . A solution has the property called osmotic pressure regardless of
whether this additional pressure is actually present, just as a solution has a freezing point
even when its actual temperature is different from the freezing point.
Because in an equilibrium state the solvent chemical potential must be the same on both
sides of the semipermeable membrane, there is a relation between chemical potentials and
osmotic pressure given by
A .p
00
/D
A .p
C / D
A .p
(12.2.7)
(equilibrium state)
0
We can use this relation to derive an expression for A .p 0 /
A .p / as a function of .
The dependence of A on pressure is given according to Eq. 9.2.49 by
@ A
D VA
(12.2.8)
@p T;fnig
where VA is the partial molar volume of the solvent in the solution. Rewriting this equation
in the form d A D VA dp and integrating at constant temperature and composition from p 0
to p 0 C , we obtain
A .p
4 Greek
C /
A .p
/D
p 0 C
p0
VA dp
(12.2.9)
for push.

12.3 B INARY M IXTURE IN E QUILIBRIUM WITH A P URE P HASE
375
Substitution from Eq. 12.2.7 changes this to

0
A .p /
0
A .p / D
p 0 C
p0
VA dp
(12.2.10)
(constant T )
which is the desired expression for A A at a single temperature and pressure. To evaluate
the integral, we need an experimental value of the osmotic pressure of the solution. If
we assume VA is constant in the pressure range from p 0 to p 0 C , Eq. 12.2.10 becomes
simply
0
0
(12.2.11)
A .p / D VA
A .p /
12.3 BINARY MIXTURE IN EQUILIBRIUM WITH A PURE PHASE

This section considers a binary liquid mixture of components A and B in equilibrium with
either pure solid A or pure gaseous A. The aim is to nd general relations among changes of
temperature, pressure, and mixture composition in the two-phase equilibrium system that
can be applied to specic situations in later sections.
In this section, A is the chemical potential of component A in the mixture and A is
for the pure solid or gaseous phase. We begin by writing the total differential of A =T
with T , p, and xA as the independent variables. These quantities refer to the binary liquid
mixture, and we have not yet imposed a condition of equilibrium with another phase. The
general expression for the total differential is
@. A =T /
@. A =T /
@. A =T /
d. A =T / D
dxA (12.3.1)
dT C
dp C
@T
@p
@xA
p;xA
T;xA
T;p
With substitutions from Eqs. 9.2.49 and 12.1.3, this becomes
HA
@. A =T /
V
d. A =T / D
dT C A dp C
T2
T
@xA
dxA
(12.3.2)
T;p
Next we write the total differential of A =T for pure solid or gaseous A. The independent variables are T and p; the expression is like Eq. 12.3.2 with the last term missing:
d.
A =T /
HA
V
dT C A dp
T2
T
(12.3.3)
When the two phases are in transfer equilibrium, A and A are equal. If changes occur
in T , p, or xA while the phases remain in equilibrium, the condition d. A =T / D d. A =T /
must be satised. Equating the expressions on the right sides of Eqs. 12.3.2 and 12.3.3 and
combining terms, we obtain the equation
HA HA
VA VA
@. A =T /
dT
dp D
dxA
(12.3.4)
2
T
T
@xA
T;p

12.4 C OLLIGATIVE P ROPERTIES OF A D ILUTE S OLUTION
376
which we can rewrite as

sol,A H
dT
T2
sol,A V
@. A =T /
dp D
T
@xA
dxA
T;p
(12.3.5)
(phases in
equilibrium)
Here sol,A H is the molar differential enthalpy of solution of solid or gaseous A in the
liquid mixture, and sol,A V is the molar differential volume of solution. Equation 12.3.5 is
a relation between changes in the variables T , p, and xA , only two of which are independent
in the equilibrium system.
Suppose we set dp equal to zero in Eq. 12.3.5 and solve for dT = dxA . This gives us the
rate at which T changes with xA at constant p:
@T
@. A =T /
T2
(12.3.6)
D
@xA p
sol,A H
@xA
T;p
(phases in
equilibrium)
We can also set dT equal to zero in Eq. 12.3.5 and nd the rate at which p changes with xA
at constant T :
@p
T
@. A =T /
D
(12.3.7)
@xA T
sol,A V
@xA
T;p
(phases in
equilibrium)
Equations 12.3.6 and 12.3.7 will be needed in Secs. 12.4 and 12.5.
12.4 COLLIGATIVE PROPERTIES OF A DILUTE SOLUTION

The colligative properties of a solution are usually considered to be:
1. Freezing-point depression: the decrease in the freezing point of the solution, compared to pure solvent at the same pressure.
2. Boiling-point elevation: the increase in the boiling point of a solution containing
nonvolatile solutes, compared to pure solvent at the same pressure.
3. Vapor-pressure lowering: the decrease in the vapor pressure of a solution containing
nonvolatile solutes, compared to the vapor pressure of the pure solvent at the same
temperature.
4. Osmotic pressure: the increase in the pressure of the solution that places the solvent
in transfer equilibrium with pure solvent at the same temperature and pressure as the
original solution (page 374).
Note that all four properties are dened by an equilibrium between the liquid solution and a
solid, liquid, or gas phase of the pure solvent. The properties called colligative (Latin: tied
together) have in common a dependence on the concentration of solute particles that affects
the solvent chemical potential.
Figure 12.3 on the next page illustrates the freezing-point depression and boiling-point
elevation of an aqueous solution. At a xed pressure, pure liquid water is in equilibrium
with ice at the freezing point and with steam at the boiling point. These are the temperatures

377
H
O(s
)
kJ mol
H
O
so
(l)
lu
tio
n
(g)
H O
Figure 12.3 Freezing-point depression and boiling-point elevation of an aqueous solution. Solid curves: dependence on temperature of the chemical potential of H2 O (A)
in pure phases and in an aqueous solution at 1 bar. Dashed curves: unstable states. The
A values have an arbitrary zero. The solution curve is calculated for an ideal-dilute
solution of composition xA D 0:9.
at which H2 O has the same chemical potential in both phases at this pressure. At these
temperatures, the chemical potential curves for the phases intersect, as indicated by open
circles in the gure. The presence of dissolved solute in the solution causes a lowering of
the H2 O chemical potential compared to pure water at the same temperature. Consequently,
the curve for the chemical potential of H2 O in the solution intersects the curve for ice at a
lower temperature, and the curve for steam at a higher temperature, as indicated by open
triangles. The freezing point is depressed by Tf , and the boiling point (if the solute is
nonvolatile) is elevated by Tb .
Sections 12.4.112.4.4 will derive theoretical relations between each of the four colligative properties and solute composition variables in the limit of innite dilution. The
expressions show that the colligative properties of a dilute binary solution depend on properties of the solvent, are proportional to the solute concentration and molality, but do not
depend on the kind of solute.
Although these expressions provide no information about the activity coefcient of a
solute, they are useful for estimating the solute molar mass. For example, from a measurement of any of the colligative properties of a dilute solution and the appropriate theoretical
relation, we can obtain an approximate value of the solute molality mB . (It is only approximate because, for a measurement of reasonable precision, the solution cannot be extremely
dilute.) If we prepare the solution with a known amount nA of solvent and a known mass of
solute, we can calculate the amount of solute from nB D nA MA mB ; then the solute molar
mass is the solute mass divided by nB .

378
12.4.1 Freezing-point depression

As in Sec. 12.2.1, we assume the solid that forms when a dilute solution is cooled to its
freezing point is pure component A.
Equation 12.3.6 on page 376 gives the general dependence of temperature on the composition of a binary liquid mixture of A and B that is in equilibrium with pure solid A.
We treat the mixture as a solution. The solvent is component A, the solute is B, and the
temperature is the freezing point Tf :
Tf2
@Tf
@. A =T /
D
(12.4.1)
@xA p
sol,A H
@xA
T;p
Consider the expression on the right side of this equation in the limit of innite dilution.
In this limit, Tf becomes Tf , the freezing point of the pure solvent, and sol,A H becomes
fus,A H , the molar enthalpy of fusion of the pure solvent.
To deal with the partial derivative on the right side of Eq. 12.4.1 in the limit of innite
dilution, we use the fact that the solvent activity coefcient A approaches 1 in this limit.
Then the solvent chemical potential is given by the Raoults law relation
A
C RT ln xA
(12.4.2)
(solution at innite dilution)
where A is the chemical potential of A in a pure-liquid reference state at the same T and
p as the mixture.5
If the solute is an electrolyte, Eq. 12.4.2 can be derived by the same procedure as described in Sec. 9.4.6 for an ideal-dilute binary solution of a nonelectrolyte. We must calculate xA from the amounts of all species present at innite dilution. In the limit of innite
dilution, any electrolyte solute is completely dissociated to its constituent ions: ion pairs
and weak electrolytes are completely dissociated in this limit. Thus, for a binary solution
of electrolyte B with ions per formula unit, we should calculate xA from
nA
xA D
(12.4.3)
nA C n B
where nB is the amount of solute formula unit. (If the solute is a nonelectrolyte, we simply
set equal to 1 in this equation.)
From Eq. 12.4.2, we can write
@. A =T /
! R as xA ! 1
(12.4.4)
@xA
T;p
In the limit of innite dilution, then, Eq. 12.4.1 becomes
R.Tf /2
@Tf
lim
D
xA !1 @xA p
fus,A H
(12.4.5)
It is customary to relate freezing-point depression to the solute concentration cB or

molality mB . From Eq. 12.4.3, we obtain
nB
1 xA D
(12.4.6)
nA C n B
5 At
the freezing point of the mixture, the reference state is an unstable supercooled liquid.

379
In the limit of innite dilution, when nB is much smaller than nA , 1 xA approaches the
value nB =nA . Then, using expressions in Eq. 9.1.14 on page 226, we obtain the relations
dxA D
xA / D
d.1
d.nB =nA /
VA dcB
MA dmB
(12.4.7)
(binary solution at
innite dilution)
which transform Eq. 12.4.5 into the following:6

lim
cB !0
lim
mB !0
@Tf
@cB
@Tf
@mB
VA R.Tf /2
fus,A H
MA R.Tf /2
fus,A H
(12.4.8)
We can apply these equations to a nonelectrolyte solute by setting equal to 1.

As cB or mB approaches zero, Tf approaches Tf . The freezing-point depression (a
negative quantity) is Tf D Tf Tf . In the range of molalities of a dilute solution in which
.@Tf =@mB /p is given by the expression on the right side of Eq. 12.4.8, we can write
MA R.Tf /2
mB
fus,A H
Tf D
(12.4.9)
The molal freezing-point depression constant or cryoscopic constant, Kf , is dened

for a binary solution by
Tf
def
Kf D
lim
(12.4.10)
mB !0 mB
and, from Eq. 12.4.9, has a value given by
Kf D
MA R.Tf /2
fus,A H
(12.4.11)
The value of Kf calculated from this formula depends only on the kind of solvent and the
pressure. For H2 O at 1 bar, the calculated value is Kb D 1:860 K kg mol 1 (Prob. 12.4).
In the dilute binary solution, we have the relation
Tf D
Kf mB
(12.4.12)
(dilute binary solution)
This relation is useful for estimating the molality of a dilute nonelectrolyte solution ( D1)
from a measurement of the freezing point. The relation is of little utility for an electrolyte
solute, because at any electrolyte molality that is high enough to give a measurable depression of the freezing point, the mean ionic activity coefcient deviates greatly from unity and
the relation is not accurate.
6A
small dependence of VA on T has been ignored.

380
BIOGRAPHICAL SKETCH
Francois-Marie Raoult (18301901)
Starting in 1886 Raoult began publishing

his measurements of the vapor pressures of dilute solutions of nonvolatile solutes. He used
two methods: (1) For a highly-volatile solvent
such as diethyl ether, the solution sample was
introduced above a mercury column at the upper closed end of a vertical barometer tube,
and the pressure determined from the height of
the column.b (2) The solution was placed in a
heated ask connected to a reux condenser,
and the pressure was reduced at the desired
Raoult was a French physical chemist best temperature until boiling was observed.c
known for his painstaking measurements of the
His results for diethyl ether as the solvent
freezing points and vapor pressures of dilute led him to propose the relation f f 0 D 0:01,
fN
solutions.
where f and f 0 are the vapor pressures pA and
Raoult was born in Fournes-en-Weppes p of the pure solvent and the solution, respecA
in northern France into a family of modest tively, both at the same temperature, and N
meanshis father was an ofcial in the local is one-hundred times the solute mole fraction
customs service. He supported himself with x . This relation is equivalent to the Raoults
B
various teaching posts until he was able to at- law equation p D x p (Eq. 9.4.1). He
A
A A
tain his doctors degree from the University of wrote:d
Paris in 1863.
With a view to ascertain whether this remarkRaoult began teaching at the University of
able law is general, I dissolved in ether comGrenoble in 1867, and three years later was appounds taken from the different chemical groups,
pointed Professor and chair of chemistry. He
and chosen from those whose boiling points are
remained in Grenoble for the rest of his life.
the highest; the compounds having molecular
He was married and had three children, two of
weights which are very widely different from
one another; and I measured the vapor pressures
whom died before him.
of the solutions obtained. In every case I found
His strength was in experimental measuref0
. . . that the ratio ff N is very nearly 0.01.
ments rather than theory. He constructed most
of his own apparatus; some of it was displayed His measurements with dilute solutions of
in Paris at the Centennial Museum of the Uni- nonelectrolyte solutes in various other solversal Exposition of 1900.a
vents, including benzene, ethanol, and water,
In all Raoult published more than 100 pa- gave the same results.e He was pleased that
pers based on his measurements. The research his measurements conrmed the theory of sofor his doctoral thesis, and later at Grenoble, lutions being developed by J. H. vant Hoff.
was on the thermochemistry of galvanic cells.
Raoults work brought him many honors.
His rst measurements of freezing-point de- He was most proud of being named Commanpressions appeared in 1878. He pointed out der of the French Legion of Honor in 1900.
the advantages of determining the molar mass
Sir William Ramsey described Raoults perof a substance from the freezing point of its sonality as follows:f
dilute solution, and gave specic examples of
Though modest and retiring, Raoults devotion
this procedure. He was the rst to show experto his work, dignity of character and sweetness
imentally that the freezing-point depression of
of temper gained him many friends. He was not
a dilute aqueous solution of an electrolyte is
an ambitious man, but was content to work on,
proportional to the number of ions per solute
happy if his discoveries contributed to the advancement of science.
formula unit (Eq. 12.4.12).
a Ref.
[16].
b Ref.
[137].
c Ref.
[138].
d Ref.
[137].
e Ref.
[138].
f Ref.
[135].

381
12.4.2 Boiling-point elevation

We can apply Eq. 12.3.6 to the boiling point Tb of a dilute binary solution. The pure phase of
A in equilibrium with the solution is now a gas instead of a solid.7 Following the procedure
of Sec. 12.4.1, we obtain
MA R.Tb /2
@Tb
D
lim
(12.4.13)
mB !0 @mB p
vap,A H
where vap,A H is the molar enthalpy of vaporization of pure solvent at its boiling point Tb .
The molal boiling-point elevation constant or ebullioscopic constant, Kb , is dened
for a binary solution by
Tb
def
Kb D lim
(12.4.14)
mB !0 mB
where Tb D Tb
Tb is the boiling-point elevation. Accordingly, Kb has a value given by

Kb D
MA R.Tb /2
vap,A H
(12.4.15)
For the boiling point of a dilute solution, the analogy of Eq. 12.4.12 is
Tb D Kb mB
(12.4.16)
Since Kf has a larger value than Kb (because fus,A H is smaller than vap,A H ), the measurement of freezing-point depression is more useful than that of boiling-point elevation for
estimating the molality of a dilute solution.
12.4.3 Vapor-pressure lowering

In a binary two-phase system in which a solution of volatile solvent A and nonvolatile solute
B is in equilibrium with gaseous A, the vapor pressure of the solution is equal to the system
pressure p.
Equation 12.3.7 on page 376 gives the general dependence of p on xA for a binary
liquid mixture in equilibrium with pure gaseous A. In this equation, sol,A V is the molar
differential volume change for the dissolution of the gas in the solution. In the limit of
innite dilution, sol,A V becomes vap,A V , the molar volume change for the vaporization
of pure solvent. We also apply the limiting expressions of Eqs. 12.4.4 and 12.4.7. The result
is
VA RT
MA RT
@p
@p
D
lim
D
(12.4.17)
lim
mB !0 @mB T
cB !0 @cB T
vap,A V
vap,A V
If we neglect the molar volume of the liquid solvent compared to that of the gas, and
assume the gas is ideal, then we can replace vap,A V in the expressions above by VA (g) D
RT =pA and obtain
@p
@p
lim
VA pA
lim
M A pA
(12.4.18)
cB !0 @cB T
mB !0 @mB T
7 We
must assume the solute is nonvolatile or has negligible partial pressure in the gas phase.

382
where pA is the vapor pressure of the pure solvent at the temperature of the solution.
Thus, approximate expressions for vapor-pressure lowering in the limit of innite dilution are
p
VA pA cB
and
p
M A pA mB
(12.4.19)
We see that the lowering in this limit depends on the kind of solvent and the solution composition, but not on the kind of solute.
12.4.4 Osmotic pressure

The osmotic pressure is an intensive property of a solution and was dened in Sec. 12.2.2.
In a dilute solution of low , the approximation used to derive Eq. 12.2.11 (that the partial
molar volume VA of the solvent is constant in the pressure range from p to p C ) becomes
valid, and we can write
A
(12.4.20)
VA
In the limit of innite dilution, A
A approaches RT ln xA (Eq. 12.4.2) and VA becomes the molar volume VA of the pure solvent. In this limit, Eq. 12.4.20 becomes
D
from which we obtain the equation
lim
xA !1
@
@xA
RT ln xA
VA
T;p
(12.4.21)
RT
VA
The relations in Eq. 12.4.7 transform Eq. 12.4.22 into
@
D RT
lim
cB !0 @cB T;p
@
RTMA
lim
D
D
mB !0 @mB T;p
VA
(12.4.22)
(12.4.23)
A RT
(12.4.24)
Equations 12.4.23 and 12.4.24 show that the osmotic pressure becomes independent of
the kind of solute as the solution approaches innite dilution. The integrated forms of these
equations are
D cB RT
RTMA
mB D
VA
(12.4.25)
A RT
mB
(12.4.26)
Equation 12.4.25 is vant Hoffs equation for osmotic pressure. If there is more than one
P
P
solute species, cB can be replaced by iA ci and mB by iA mi in these expressions.
In Sec. 9.6.3, it was stated that =mB is equal to the product of m and the limiting
P
value of =mB at innite dilution, where m D . A
A /=RTMA
i A mi is the
osmotic coefcient. This relation follows directly from Eqs. 12.2.11 and 12.4.26.

383
BIOGRAPHICAL SKETCH
Jacobus Henricus vant Hoff (18521911)

motic pressure of a dilute solution is described
by D icB RT (Eq. 12.4.25 with replaced
by i ). The Swedish chemist Svante Arrhenius
later interpreted i as the number of particles
per solute formula unit, helping to validate his
theory of ionic dissociation of electrolytes.
In a celebrated 1887 summary of his theory
of osmotic pressure, vant Hoff wrote:b
vant Hoff was a Dutch chemist who gained

fame from epoch-making publications in several elds of theoretical chemistry. He was an
introvert who valued nature, philosophy, poetry, and the power of imagination.a
vant Hoff was born in Rotterdam, the son
of a physician, and studied at Delft, Leyden,
Bonn, and Paris before obtaining his doctors
degree at Utrecht in 1874. For 18 years he was
Professor of Chemistry, Mineralogy, and Geology at the University of Amsterdam. In 1896,
mainly to escape burdensome lecture duties,
he accepted an appointment as Professor of
Physical Chemistry at the University of Berlin,
where he remained for the rest of his life.
In the same year he received his doctors degree, he published the rst explanation of optical activity based on the stereoisomerism of an
asymmetric carbon atom. Similar ideas were
published independently two months later by
the French chemist Joseph Le Bel.
In 1884 he entered the eld of physical
chemistry with his book Etudes de Dynamique

Chimique, a systematic study of theories of reaction kinetics and chemical equilibrium. The
book introduced, in the form of Eq. 12.1.12,
his expression for the temperature dependence
of an equilibrium constant.
vant Hoff next used thermodynamic cycles
to reason that solutions of equal osmotic pressure should have the same values of freezingpoint depression and of vapor-pressure lowering. Then, from Raoults extensive measurements of these values, he found that the os-
. . . the relation found permits of an important extension of the law of Avogadro, which now nds
application also to all [nonelectrolyte] solutions,
if only osmotic pressure is considered instead of
elastic pressure. At equal osmotic pressure and
equal temperature, equal volumes of the most
widely different solutions contain an equal number of molecules, and, indeed, the same number
which, at the same pressure and temperature, is
contained in an equal volume of a gas. . . . the
existence of the so-called normal molecular lowering of the freezing-point and diminution of the
vapor-pressure were not discovered until Raoult
employed the organic compounds. These substances, almost without exception, behave normally. It may, then, have appeared daring to give
Avogadros law for solutions such a prominent
place, and I should not have done so had not Arrhenius pointed out to me, by letter, the probability that salts and analogous substances, when in
solution, break down into ions.
In 1901 vant Hoff was awarded the rst

Nobel Prize in Chemistry in recognition of
the extraordinary services he has rendered by
the discovery of the laws of chemical dynamics and osmotic pressure in solutions.
vant Hoff was married with two daughters
and two sons. He died of tuberculosis at age
fty-eight. In a memorial article, Frederick
Donnanc wrote:d
The present writer is one of those whose privilege it is to have worked under vant Hoff.
. . . Every day endeared vant Hoff to the small
band of workers in his laboratory. His joy in his
work, the simple and unaffected friendliness of
his nature, and the marvellous power of his mind
affected us most deeply. All who worked with
vant Hoff quickly learned to love and respect
him, and we were no exception to the rule.
a Ref.
[122]. b Ref. [160]. c Donnan was an Irish physical chemist after whom the Donnan membrane
equilibrium and Donnan potential (Sec. 12.7.3) are named. d Ref. [46].

12.5 S OLID L IQUID E QUILIBRIA
384
12.5 SOLIDLIQUID EQUILIBRIA

A freezing-point curve (freezing point as a function of liquid composition) and a solubility
curve (composition of a solution in equilibrium with a pure solid as a function of temperature) are different ways of describing the same physical situation. Thus, strange as it may
sound, the composition xA of an aqueous solution at the freezing point is the mole fraction
solubility of ice in the solution.
12.5.1 Freezing points of ideal binary liquid mixtures

Section 12.2.1 described the use of freezing-point measurements to determine the solvent
chemical potential in a solution of arbitrary composition relative to the chemical potential
of the pure solvent. The way in which freezing point varies with solution composition in
the limit of innite dilution was derived in Sec. 12.4.1. Now let us consider the freezing
behavior over the entire composition range of an ideal liquid mixture.
The general relation between temperature and the composition of a binary liquid mixture, when the mixture is in equilibrium with pure solid A, is given by Eq. 12.3.6:
@T
T2
@. A =T /
D
(12.5.1)
@xA p
sol,A H
@xA
T;p
We can replace T by Tf;A to indicate this is the temperature at which the mixture freezes to
form solid A. From the expression for the chemical potential of component A in an ideal
liquid mixture, A D A C RT ln xA , we have @. A =T /=@xA T;p D R=xA . With these
substitutions, Eq. 12.5.1 becomes
@Tf;A
@xA
2
RTf;A
xA sol,A H
(12.5.2)
(ideal liquid mixture)
Figure 12.4 on the next page compares the freezing behavior of benzene predicted by
this equation with experimental freezing-point data for mixtures of benzenetoluene and
benzenecyclohexane. Any constituent that forms an ideal liquid mixture with benzene
should give freezing points for the formation of solid benzene that fall on the curve in this
gure. The agreement is good over a wide range of compositions for benzenetoluene mixtures (open circles), which are known to closely approximate ideal liquid mixtures. The
agreement for benzenecyclohexane mixtures (open triangles), which are not ideal liquid
mixtures, is conned to the ideal-dilute region.
If we make the approximation that sol,A H is constant over the entire range of mixture
composition, we can replace it by fus,A H , the molar enthalpy of fusion of pure solid A at
its melting point. This approximation allows us to separate the variables in Eq. 12.5.2 and
0
0
integrate as follows from an arbitrary mixture composition xA at the freezing point Tf;A to
pure liquid A at its freezing point Tf;A :
Z
Tf;A
0
Tf;A
dT
R
D
2
T
fus,A H
1
0
xA
dxA
xA
(12.5.3)

385
280
Tf;A =K
270
260
250
240
0:2
0:4
0:6
0:8
1:0
xB
Figure 12.4 Dependence on composition of the freezing point of binary liquid mixtures with benzene as component A.a Solid curve: calculated for an ideal liquid mixture (Eq. 12.5.2), taking the temperature variation of sol,A H into account. Open
circles: B = toluene. Open triangles: B = cyclohexane.
a Experimental
data from Ref. [121].
f A f
Figure 12.5 Freezing-point curves of ideal binary liquid mixtures. The solid is component A. Each curve is calculated from Eq. 12.5.4 and is labeled with the value of
fus,A H=RTf;A .
The result, after some rearrangement, is

fus,A H
ln xA D
R
1
Tf;A
1
Tf;A
(12.5.4)
(ideal liquid mixture,
sol,A H Dfus,A H )
This equation was used to generate the curves shown in Fig. 12.5. Although the shape of the
freezing-point curve (Tf;A versus xB ) shown in Fig. 12.4 is concave downward, Fig. 12.5
shows this is not always the case. When fus,A H=RTf;A is less than 2, the freezing-point
curve at low xB is concave upward.

386
12.5.2 Solubility of a solid nonelectrolyte

Suppose we nd that a solution containing solute B at a particular combination of temperature, pressure, and composition can exist in transfer equilibrium with pure solid B at the
same temperature and pressure. This solution is said to be saturated with respect to the
solid. We can express the solubility of the solid in the solvent by the value of the mole fraction, concentration, or molality of B in the saturated solution. We can also dene solubility
as the maximum value of the solute mole fraction, concentration, or molality that can exist
in the solution without the possibility of spontaneous precipitation.
This section considers the solubility of a solid nonelectrolyte. For the solution process
B(s) ! B(sln), the general expression for the thermodynamic equilibrium constant is K D
aB (sln)=aB (s).8 The activity of the pure solid is aB (s) D B (s). Let us use a solute standard
state based on mole fraction; then the solute activity is aB (sln) D x;B x;B xB . From these
relations, the solubility expressed as a mole fraction is
xB D
B (s) K
(12.5.5)
x;B x;B
If we measure the solubility at the standard pressure, the pressure factors

are unity and the solubility is given by
xB D
B (s) and
x;B
(12.5.6)
x;B
(solubility of solid B, pDp )
If the pressure is not exactly equal to p , but is not very much greater, the values of the
pressure factors are close to unity and Eq. 12.5.6 is a good approximation.
We can nd the standard molar enthalpy of solution of B from the temperature dependence of the solubility. Combining Eqs. 12.1.12 and 12.5.6, we obtain
sol,B H D RT 2
d ln.
x;B xB /
(12.5.7)
dT
(pDp )
The solubility may be small enough for us to be able to set the solute activity coefcient
equal to 1, in which case Eq. 12.5.7 becomes
sol,B H D RT 2
d ln xB
dT
(12.5.8)
(pDp ,
x;B D1)
If the solubility xB increases with increasing temperature, sol,B H must be positive

and the solution process is endothermic. A decrease of solubility with increasing temperature implies an exothermic solution process. These statements refer to a solid of low
solubility; see page 358 for a discussion of the general relation between the temperature
dependence of solubility and the sign of the molar differential enthalpy of solution at saturation.
8 In
this and other expressions for equilibrium constants in this chapter, activities will be assumed to be for
equilibrium states, although not indicated by the eq subscripts used in Chap. 11.

387
For a solute standard state based on molality, we can derive equations like Eqs. 12.5.7
and 12.5.8 with x;B replaced by m;B and xB replaced by mB =m . If we use a solute
standard state based on concentration, the expressions become slightly more complicated.
The solubility in this case is given by
cB D
B (s) Kc
(12.5.9)
c;B c;B
From Eq. 12.1.11, we obtain, for a nonelectrolyte solid of low solubility, the relation
2 d ln.cB =c /
sol,B H D RT
C A
(12.5.10)
dT
(pDp ,
c;B D1)
12.5.3 Ideal solubility of a solid

The ideal solubility of a solid at a given temperature and pressure is the solubility calculated
on the assumptions that (1) the liquid is an ideal liquid mixture, and (2) the molar differential
enthalpy of solution equals the molar enthalpy of fusion of the solid (sol,B H Dfus,B H ).
These were the assumptions used to derive Eq. 12.5.4 for the freezing-point curve of an
ideal liquid mixture. In Eq. 12.5.4, we exchange the constituent labels A and B so that the
solid phase is now component B:
!
fus,B H
1
1
ln xB D
(12.5.11)
R
Tf;B T
(ideal solubility of solid B)
Here Tf;B is the melting point of solid B.

According to Eq. 12.5.11, the ideal solubility of a solid is independent of the kind of
solvent and increases with increasing temperature. For solids with similar molar enthalpies
of fusion, the ideal solubility is less at a given temperature the higher is the melting point.
This behavior is shown in Fig. 12.6 on the next page. In order for the experimental solubility
of a solid to agree even approximately with the ideal value, the solvent and solute must be
chemically similar, and the temperature must be close to the melting point of the solid so
that sol,B H is close in value to fus,B H .
From the freezing behavior of benzenetoluene mixtures shown by the open circles in
Fig. 12.4 on page 385, we can see that solid benzene has close to ideal solubility in
liquid toluene at temperatures not lower than about 20 K below the melting point of
benzene.
12.5.4 Solid compound of mixture components

Binary liquid mixtures are known in which the solid that appears when the mixture is cooled
is a compound containing both components in a xed proportion. This kind of solid is called
a solid compound, or stoichiometric addition compound. Examples are salt hydrates (salts
with xed numbers of waters of hydration in the formula unit) and certain metal alloys.

388
fB
fB
Figure 12.6 Ideal solubility of solid B as a function of T . The curves are calculated
for two solids having the same molar enthalpy of fusion (fus,B H D 20 kJ mol 1 ) and
the values of Tf;B indicated.
The composition of the liquid mixture in this kind of system is variable, whereas the
composition of the solid compound is xed. Suppose the components are A and B, present
in the liquid mixture at mole fractions xA and xB , and the solid compound has the formula
Aa Bb . We assume that in the liquid phase the compound is completely dissociated with
respect to the components; that is, that no molecules of formula Aa Bb exist in the liquid.
The reaction equation for the freezing process is
aA(mixt) C bB(mixt) ! Aa Bb (s)
When equilibrium exists between the liquid and solid phases, the temperature is the freezing
point
P Tf of the liquid. At equilibrium, the molar reaction Gibbs energy dened by r G D
i i i is zero:
a A b B C (s) D 0
(12.5.12)
Here A and B refer to chemical potentials in the liquid mixture, and (s) refers to the
solid compound.
How does the freezing point of the liquid mixture vary with composition? We divide
both sides of Eq. 12.5.12 by T and take differentials:
a d.
A =T /
b d.
B =T /
C d (s)=T D 0
(12.5.13)
(phase equilibrium)
The pressure is constant. Then A =T and B =T are functions of T and xA , and (s)=T
is a function only of T . We nd expressions for the total differentials of these quantities at
constant p with the help of Eq. 12.1.3 on page 367:
HA
1 @ A
d. A =T / D
dT C
dx
(12.5.14)
T2
T @xA T;p A
HB
1 @ B
d. B =T / D
dT C
dx
(12.5.15)
T2
T @xA T;p A

d (s)=T D
Hm (s)
dT
T2
389
(12.5.16)
When we substitute these expressions in Eq. 12.5.13 and solve for dT = dxA , setting T equal
to Tf , we obtain
"
#
dTf
Tf
@ A
@ B
D
a
Cb
(12.5.17)
dxA
aHA C bHB Hm (s)
@xA T;p
@xA T;p
The quantity aHA C bHB Hm (s) in the denominator on the right side of Eq. 12.5.17
is sol H , the molar differential enthalpy of solution of the solid compound in the liquid
mixture. The two partial derivatives on the right side are related through the GibbsDuhem
equation xA d A C xB d B D 0 (Eq. 9.2.27 on page 234), which applies to changes at
constant T and p. We rearrange the GibbsDuhem equation to d B D .xA =xB / d A and
divide by dxA :
@ B
xA @ A
D
(12.5.18)
@xA T;p
xB @xA T;p
Making this substitution in Eq. 12.5.17, we obtain the equation
dTf
x T
a
b
@ A
D A f
dxA
sol H xA xB
@xA T;p
which can also be written in the slightly rearranged form
dTf
bTf
a
xA
@ A
D
dxA
sol H b 1 xA
@xA T;p
(12.5.19)
(12.5.20)
Suppose we heat a sample of the solid compound to its melting point to form a liquid
mixture of the same composition as the solid. The molar enthalpy change of the fusion
process is the molar enthalpy of fusion of the solid compound, fus H , a positive quantity.
When the liquid has the same composition as the solid, the dissolution and fusion processes
are identical; under these conditions, sol H is equal to fus H and is positive.
Equation 12.5.20 shows that the slope of the freezing-point curve, Tf versus xA , is zero
when xA =.1 xA / is equal to a=b, or xA D a=.aCb/; that is, when the liquid and solid have
the same composition. Because .@ A =@xA /T;p is positive, and sol H at this composition
is also positive, we see from the equation that the slope decreases as xA increases. Thus,
the freezing-point curve has a maximum at the mixture composition that is the same as
the composition of the solid compound. This conclusion applies when both components of
the liquid mixture are nonelectrolytes, and also when one component is an electrolyte that
dissociates into ions.
Now let us assume the liquid mixture is an ideal liquid mixture of nonelectrolytes in
which A obeys Raoults law for fugacity, A D A C RT ln xA . The partial derivative
.@ A =@xA /T;p then equals RT =xA , and Eq. 12.5.19 becomes
RTf2
dTf
D
dxA
sol H
a
xA
b
xB
(12.5.21)

390
f K
Mg
mole fraction Zn
Zn
Figure 12.7 Solid curve: freezing-point curve of a liquid melt of Zn and Mg that
solidies to the solid compound Zn2 Mg. a The curve maximum (open circle) is at
00
the compound composition xZn D 2=3 and the solid compound melting point Tf00 D
861 K. Dashed curve: calculated using Eq. 12.5.23 with fus H D 15:8 kJ mol 1 .
a Ref.
[50], p. 603.
By making the approximations that sol H is independent of T and xA , and is equal to

fus H , we can separate the variables and integrate as follows:
!
00
00
Z T 00
Z xA
Z xB
f dT
R
a
b
f
D
dxA C
dxB
(12.5.22)
0
0
0
fus H
Tf2
Tf
xA xA
xB xB
(The second integral on the right side comes from changing dxA to
the integration is
00
00
xA
xB
1
1
R
D 00 C
a ln 0 C b ln 0
Tf0
Tf
fus H
xA
xB
dxB .) The result of
(12.5.23)
(ideal liquid mixture in

equilibrium with solid
compound, sol H Dfus H )
0
0
0
Let Tf0 be the freezing point of a liquid mixture of composition xA and xB D 1 xA , and
00
00
00
let Tf be the melting point of the solid compound of composition xA D a=.aCb/ and xB D
b=.a C b/. Figure 12.7 shows an example of a molten metal mixture that solidies to an
alloy of xed composition. The freezing-point curve of this system is closely approximated
by Eq. 12.5.23.
12.5.5 Solubility of a solid electrolyte

Consider an equilibrium between a crystalline salt (or other kind of ionic solid) and a solution containing the solvated ions:
M
(s)
CM
zC
(aq) C
Xz (aq)
Here C and
are the numbers of cations and anions in the formula unit of the salt, and
zC and z are the charge numbers of these ions. The solution in equilibrium with the

391
solid salt is a saturated solution. The thermodynamic equilibrium constant for this kind of
equilibrium is called a solubility product, Ks .
We can readily derive a relation between Ks and the molalities of the ions in the saturated solution by treating the dissolved salt as a single solute substance, B. We write the
equilibrium in the form B (s) B(sln), and write the expression for the solubility product
as a proper quotient of activities:
am;B
Ks D
(12.5.24)
aB
From Eq. 10.3.16 on page 294, we have am;B D m;B .mC =m / C .m =m / . This
expression is valid whether or not the ions MzC and Xz are present in solution in the same
ratio as in the solid salt. When we replace am;B with this expression, and replace aB with
B (Table 9.5), we obtain
mC C m
m;B
Ks D
(12.5.25)
m
m
B
where D C C
is the total number of ions per formula unit. is the mean ionic
activity coefcient of the dissolved salt in the saturated solution, and the molalities mC and
m refer to the ions MzC and Xz in this solution.
The rst factor on the right side of Eq. 12.5.25, the proper quotient of pressure factors
for the reaction B (s)!B(sln), will be denoted r (the subscript r stands for reaction).
The value of r is exactly 1 if the system is at the standard pressure, and is otherwise
approximately 1 unless the pressure is very high.
If the aqueous solution is produced by allowing the salt to dissolve in pure water, or
in a solution of a second solute containing no ions in common with the salt, then the ion
molalities in the saturated solution are mC D C mB and m D
mB where mB is the
solubility of the salt expressed as a molality. Under these conditions, Eq. 12.5.25 becomes9
m
B
C
(12.5.26)
Ks D r C
m
(no common ion)
If the ionic strength of the saturated salt solution is sufciently low (i.e., the solubility is
sufciently low), it may be practical to evaluate the solubility product with Eq. 12.5.26 and
an estimate of from the DebyeH ckel limiting law (see Prob. 12.19). The most accurate
u
method of measuring a solubility product, however, is through the standard cell potential of
an appropriate galvanic cell (Sec. 14.3.3).
Since Ks is a thermodynamic equilibrium constant that depends only on T , and r
depends only on T and p, Eq. 12.5.26 shows that any change in the solution composition at
constant T and p that decreases must increase the solubility. For example, the solubility
of a sparingly-soluble salt increases when a second salt, lacking a common ion, is dissolved
in the solution; this is a salting-in effect.
Equation 12.5.25 is a general equation that applies even if the solution saturated with
one salt contains a second salt with a common ion. For instance, consider the sparinglysoluble salt M C X in transfer equilibrium with a solution containing the more soluble
9 We
could also have obtained this equation by using the expression of Eq. 10.3.10 for am;B .

12.6 L IQUID L IQUID E QUILIBRIA
392
0
salt M C Y 0 at molality mC . The common ion in this example is the cation MzC . The
expression for the solubility product is now
Ks D
r . C mB
0
C mC /
mB /
=.m /
(12.5.27)
(common cation)
where mB again is the solubility of the sparingly-soluble salt, and mC is the molality of the
second salt. Ks and r are constant if T and p do not change, so any increase in mC at
constant T and p must cause a decrease in the solubility mB . This is called the common ion
effect.
From the measured solubility of a salt in pure solvent, or in an electrolyte solution
with a common cation, and a known value of Ks , we can evaluate the mean ionic activity
coefcient through Eq. 12.5.26 or 12.5.27. This procedure has the disadvantage of being
limited to the value of mB existing in the saturated solution.
We nd the temperature dependence of Ks by applying Eq. 12.1.12:
sol,B H
d ln Ks
D
(12.5.28)
dT
RT 2
At the standard pressure, sol,B H is the same as the molar enthalpy of solution at innite
dilution, sol,B H 1 .
12.6 LIQUIDLIQUID EQUILIBRIA

12.6.1 Miscibility in binary liquid systems
When two different pure liquids are unable to mix in all proportions, they are said to be
partially miscible. When these liquids are placed in contact with one another and allowed
to come to thermal, mechanical, and transfer equilibrium, the result is two coexisting liquid
mixtures of different compositions.
Liquids are never actually completely immiscible. To take an extreme case, liquid mercury, when equilibrated with water, has some H2 O dissolved in it, and some mercury dissolves in the water, although the amounts may be too small to measure.
The Gibbs phase rule for a multicomponent system to be described in Sec. 13.1 shows
that a two-component, two-phase system at equilibrium has only two independent intensive
variables. Thus at a given temperature and pressure, the mole fraction compositions of both
phases are xed; the compositions depend only on the identity of the substances and the
temperature and pressure.
Figure 13.5 on page 432 shows a phase diagram for a typical binary liquid mixture that
spontaneously separates into two phases when the temperature is lowered. The thermodynamic conditions for phase separation of this kind were discussed in Sec. 11.1.6. The phase
separation is usually the result of positive deviations from Raoults law. Typically, when
phase separation occurs, one of the substances is polar and the other nonpolar.
12.6.2 Solubility of one liquid in another

Suppose substances A and B are both liquids when pure. In discussing the solubility of
liquid B in liquid A, we can treat B as either a solute or as a constituent of a liquid mixture.
The difference lies in the choice of the standard state or reference state of B.

393
We can dene the solubility of B in A as the maximum amount of B that can dissolve
without phase separation in a given amount of A at the given temperature and pressure.
Treating B as a solute, we can express its solubility as the mole fraction of B in the phase
at the point of phase separation. The addition of any more B to the system will result in
two coexisting liquid phases of xed composition, one of which will have mole fraction xB
equal to its solubility.10
Consider a system with two coexisting liquid phases and containing components A
and B. Let be the A-rich phase and be the B-rich phase. For example, A could be water
and B could be benzene, a hydrophobic substance. Phase would then be an aqueous phase
polluted with a low concentration of dissolved benzene, and phase would be wet benzene.
xB would be the solubility of the benzene in water, expressed as a mole fraction.

Below, relations are derived for this kind of system using both choices of standard state
or reference state.
Solute standard state
Assume that the two components have low mutual solubilities, so that B has a low mole
fraction in phase and a mole fraction close to 1 in phase . It is then appropriate to treat
B as a solute in phase and as a constituent of a liquid mixture in phase . The value of xB

is the solubility of liquid B in liquid A.
The equilibrium when two liquid phases are present is B() B(), and the expression
for the thermodynamic equilibrium constant, with the solute standard state based on mole
fraction, is
ax;B
x;B x;B xB
(12.6.1)
KD D

aB
xB
B B
The solubility of B is then given by
xB D

B B xB
K
x;B x;B
(12.6.2)
The values of the pressure factors and activity coefcients are all close to 1, so that the
solubility of B in A is given by xB
K. The temperature dependence of the solubility is
given by
sol,B H
d ln xB
d ln K
D
(12.6.3)
dT
dT
RT 2
where sol,B H is the molar enthalpy change for the transfer at pressure p of pure liquid
solute to the solution at innite dilution.
H2 O and n-butylbenzene are two liquids with very small mutual solubilities. Figure 12.8 on the next page shows that the solubility of n-butylbenzene in water exhibits
a minimum at about 12 C. Equation 12.6.3 allows us to deduce from this behavior that
sol,B H is negative below this temperature, and positive above.
10 Experimentally,
the solubility of B in A can be determined from the cloud point, the point during titration of
A with B at which persistent turbidity is observed.

394
12:9
ln xB
13:0
13:1
13:2
13:3
10
20
30
40
50
t= C
Figure 12.8 Aqueous solubility of liquid n-butylbenzene as a function of temperature

(Ref. [124]).
Pure-liquid reference state

The condition for transfer equilibrium of component B is D
B
reference state for B in both phases, this condition becomes
B
C RT ln.

B xB /
C RT ln.
B.
If we use a pure-liquid

B xB /
(12.6.4)
This results in the following relation between the compositions and activity coefcients:

B xB

B xB
(12.6.5)
As before, we assume the two components have low mutual solubilities, so that the B
rich phase is almost pure liquid B. Then xB is only slightly less than 1, B is close to 1, and
Eq. 12.6.5 becomes xB

1= B . Since xB is much less than 1, B must be much greater
than 1.
In environmental chemistry it is common to use a pure-liquid reference state for a nonpolar liquid solute that has very low solubility in water, so that the aqueous solution is
essentially at innite dilution. Let the nonpolar solute be component B, and let the aqueous
phase that is equilibrated with liquid B be phase . The activity coefcient B is then a
limiting activity coefcient or activity coefcient at innite dilution. As explained above,
the aqueous solubility of B in this case is given by xB 1= B , and B is much greater than
1.
We can also relate the solubility of B to its Henrys law constant kH,B . Suppose the two
liquid phases are equilibrated not only with one another but also with a gas phase. Since
B is equilibrated between phase and the gas, we have x;B D fB =kH,B xB (Table 9.4).
From the equilibration of B between phase and the gas, we also have B D fB =xB fB .
By eliminating the fugacity fB from these relations, we obtain the general relation
xB D

B xB fB
x;B kH,B
(12.6.6)

12.7 M EMBRANE E QUILIBRIA
395
If we assume as before that the activity coefcients and xB are close to 1, and that the gas
phase behaves ideally, the solubility of B is given by xB pB =kH,B , where pB is the vapor
pressure of the pure solute.
12.6.3 Solute distribution between two partially-miscible solvents

Consider a two-component system of two equilibrated liquid phases, and . If we add a
small quantity of a third component, C, it will distribute itself between the two phases. It is
appropriate to treat C as a solute in both phases. The thermodynamic equilibrium constant
for the equilibrium C./ C./, with solute standard states based on mole fraction, is
KD
ax;C
ax;C
x;C x;C xC
x;C x;C xC
(12.6.7)
We dene K 0 as the ratio of the mole fractions of C in the two phases at equilibrium:
0 def
K D
xC
xC
x;C x;C
K
x;C x;C
(12.6.8)
At a xed T and p, the pressure factors and equilibrium constant are constants. If xC is
low enough in both phases for x;C and x;C to be close to unity, K 0 becomes a constant for
the given T and p. The constancy of K 0 over a range of dilute composition is the Nernst
distribution law.
Since solute molality and concentration are proportional to mole fraction in dilute solu

tions, the ratios m =mC and cC =cC also approach constant values at a given T and p. The
C
ratio of concentrations is called the partition coefcient or distribution coefcient.
In the limit of innite dilution of C, the two phases have the compositions that exist
when only components A and B are present. As C is added and xC and xC increase beyond
the region of dilute solution behavior, the ratios xB =xA and xB =xA may change. Continued
addition of C may increase the mutual solubilities of A and B, resulting, when enough C
has been added, in a single liquid phase containing all three components. It is easier to
understand this behavior with the help of a ternary phase diagram such as Fig. 13.17 on
page 444.
12.7 MEMBRANE EQUILIBRIA

A semipermeable membrane used to separate two liquid phases can, in principle, be permeable to certain species and impermeable to others. A membrane, however, may not be
perfect in this respect over a long time period (see page 374). We will assume that during
the period of observation, those species to which the membrane is supposed to be permeable
quickly achieve transfer equilibrium, and only negligible amounts of the other species are
transferred across the membrane.
Section 12.2.2 sketched a derivation of the conditions needed for equilibrium in a twophase system in which a membrane permeable only to solvent separates a solution from
pure solvent. We can generalize the results for any system with two liquid phases separated
by a semipermeable membrane: in an equilibrium state, both phases must have the same

396
temperature, and any species to which the membrane is permeable must have the same
chemical potential in both phases. The two phases, however, need not and usually do not
have the same pressure.
12.7.1 Osmotic membrane equilibrium

An equilibrium state in a system with two solutions of the same solvent and different solute
compositions, separated by a membrane permeable only to the solvent, is called an osmotic
membrane equilibrium. We have already seen this kind of equilibrium in an apparatus
that measures osmotic pressure (Fig. 12.2 on page 373).
Consider a system with transfer equilibrium of the solvent across a membrane separating phases and . The phases have equal solvent chemical potentials but different
pressures:
(12.7.1)
A .p / D A .p /
The dependence of A on pressure in a phase of xed temperature and composition is given
by .@ A =@p/T;fnig D VA (from Eq. 9.2.49), where VA is the partial molar volume of A in
the phase. If we apply this relation to the solution of phase , treat the partial molar volume
VA as independent of pressure, and integrate at constant temperature and composition from
the pressure of phase to that of phase , we obtain
A .p /
A .p /
C VA .p
p/
(12.7.2)
By equating the two expressions for A .p / and rearranging, we obtain the following expression for the pressure difference needed to achieve transfer equilibrium:
p
p D
A .p /
A .p /
(12.7.3)
VA
The pressure difference can be related to the osmotic pressures of the two phases. From
Eq. 12.2.11 on page 375, the solvent chemical potential in a solution phase can be written
A .p/ D A .p/ VA .p/. Using this to substitute for A .p / and A .p / in Eq. 12.7.3,
we obtain
!
VA
p p D .p /
.p /
(12.7.4)
VA
12.7.2 Equilibrium dialysis

Equilibrium dialysis is a useful technique for studying the binding of a small uncharged
solute species (a ligand) to a macromolecule. The macromolecule solution is placed on
one side of a membrane through which it cannot pass, with a solution without the macromolecule on the other side, and the ligand is allowed to come to transfer equilibrium across
the membrane. If the same solute standard state is used for the ligand in both solutions, at
equilibrium the unbound ligand must have the same activity in both solutions. Measurements of the total ligand molality in the macromolecule solution and the ligand molality in
the other solution, combined with estimated values of the unbound ligand activity coefcients, allow the amount of ligand bound per macromolecule to be calculated.

397
12.7.3 Donnan membrane equilibrium

If one of the solutions in a two-phase membrane equilibrium contains certain charged solute
species that are unable to pass through the membrane, whereas other ions can pass through,
the situation is more complicated than the osmotic membrane equilibrium described in Sec.
12.7.1. Usually if the membrane is impermeable to one kind of ion, an ion species to which
it is permeable achieves transfer equilibrium across the membrane only when the phases
have different pressures and different electric potentials. The equilibrium state in this case
is a Donnan membrane equilibrium, and the resulting electric potential difference across
the membrane is called the Donnan potential. This phenomenon is related to the membrane
potentials that are important in the functioning of nerve and muscle cells (although the cells
of a living organism are not, of course, in equilibrium states).
A Donnan potential can be measured electrically, with some uncertainty due to unknown
liquid junction potentials, by connecting silver-silver chloride electrodes (described in Sec.
14.1) to both phases through salt bridges.
General expressions
Consider solution phases and separated by a semipermeable membrane. Both phases
contain a dissolved salt, designated solute B, that has C cations and
anions in each
formula unit. The membrane is permeable to these ions. Phase also contains a protein or
other polyelectrolyte with a net positive or negative charge, together with counterions of the
opposite charge that are the same species as the cation or anion of the salt. The presence of
the counterions in phase prevents the cation and anion of the salt from being present in
stoichiometric amounts in this phase. The membrane is impermeable to the polyelectrolyte,
perhaps because the membrane pores are too small to allow the polyelectrolyte to pass
through.
The condition for transfer equilibrium of solute B is D B , or

B
.
m;B /
C RT ln am;B D .
m;B /
Solute B has the same standard state in the two phases, so that .
The activities am;B
C RT ln am;B
m;B /
(12.7.5)
and .
m;B /
are
and am;B
equal.
are therefore equal at equilibrium. Using the expression
for solute activity from Eq. 10.3.16, which is valid for a multisolute solution, we nd that at
transfer equilibrium the following relation must exist between the molalities of the salt ions
in the two phases:
m C m
mC
m
(12.7.6)
D m;B
m;B
C
To nd an expression for the Donnan potential, we can equate the single-ion chemical
potentials of the salt cation: . / D C . /. When we use the expression of Eq.

C
10.1.15 for C . /, we obtain
RT
D
ln
zC F
C C mC

C C mC
(12.7.7)
(Donnan potential)

398
The condition needed for an osmotic membrane equilibrium related to the solvent can
be written
(12.7.8)
A .p / D 0
A .p /
The chemical potential of the solvent is A D C RT ln aA D C RT ln. A
A
A
From Table 9.6, we have to a good approximation the expression RT ln A D VA .p
RT ln

A xA

A xA
C VA p
p D 0
A xA /.
p /.
(12.7.9)
We can use this equation to estimate the pressure difference needed to maintain an equilib
rium state. For dilute solutions, with A and A set equal to 1, the equation becomes
p
x
RT
ln A
VA
xA
(12.7.10)
In the limit of innite dilution, ln xA can be replaced by MA

page 267), giving the relation
X
MA RT X
mi m D A RT
mi
p p
i
VA
i A
i A mi
m
i
(Eq. 9.6.12 on
(12.7.11)
iA
Example
As a specic example of a Donnan membrane equilibrium, consider a system in which an
aqueous solution of a polyelectrolyte with a net negative charge, together with a counterion
MC and a salt MX of the counterion, is equilibrated with an aqueous solution of the salt
across a semipermeable membrane. The membrane is permeable to the H2 O solvent and
to the ions MC and X , but is impermeable to the polyelectrolyte. The species in phase
are H2 O, MC , and X ; those in phase are H2 O, MC , X , and the polyelectrolyte. In
a equilibrium state, the two phases have the same temperature but different compositions,
electric potentials, and pressures.
Because the polyelectrolyte in this example has a negative charge, the system has more
MC ions than X ions. Figure 12.9(a) on the next page is a schematic representation of
an initial state of this kind of system. Phase is shown as a solution conned to a closed
dialysis bag immersed in phase . The number of cations and anions shown in each phase
indicate the relative amounts of these ions.
For simplicity, let us assume the two phases have equal masses of water, so that the
molality of an ion is proportional to its amount by the same ratio in both phases. It is clear
that in the initial state shown in the gure, the chemical potentials of both MC and X are
greater in phase (greater amounts) than in phase , and this is a nonequilibrium state. A
certain quantity of salt MX will therefore pass spontaneously through the membrane from
phase to phase until equilibrium is attained.
The equilibrium ion molalities must agree with Eq. 12.7.6. We make the approximation
that the pressure factors and mean ionic activity coefcients are unity. Then for the present
example, with C D
D 1, the equation becomes
m m
C
mC m
(12.7.12)

phase
phase
phase
399
phase
(a)
(b)
Figure 12.9 Process for attainment of a Donnan membrane equilibrium (schematic).

The dashed ellipse represents a semipermeable membrane.
(a) Initial nonequilibrium state.
(b) Final equilibrium state.
There is furthermore an electroneutrality condition for each phase:
m D m
C
mC D m C jzP jmP
(12.7.13)
Here zP is the negative charge of the polyelectrolyte, and mP is its molality. Substitution of
these expressions into Eq. 12.7.12 gives the relation
2
m
m C jzP jmP m
(12.7.14)
This shows that in the equilibrium state, m is greater than m . Then Eq. 12.7.12 shows
that m is less than mC . These equilibrium molalities are depicted in Fig. 12.9(b).
C
The chemical potential of a cation, its activity, and the electric potential of the phase are
related by Eq. 10.1.9 on page 288: C D C RT ln aC C zC F . In order for MC to
C
have the same chemical potential in both phases, despite its lower activity in phase , the
electric potential of phase must be greater than that of phase . Thus the Donnan potential
in the present example is positive. Its value can be estimated from Eq. 12.7.7 with
the values of the single-ion pressure factors and activity coefcients approximated by 1 and
with zC for this example set equal to 1:
mC
RT
ln
F
mC
(12.7.15)
The existence of a Donnan potential in the equilibrium state is the result of a very small
departure of the phases on both sides of the membrane from exact electroneutrality. In
the example, phase has a minute net positive charge and phase has a net negative
charge of equal magnitude. The amount of MC ion transferred across the membrane
to achieve equilibrium is slightly greater than the amount of X ion transferred; the
difference between these two amounts is far too small to be measured chemically. At
equilibrium, the excess charge on each side of the membrane is distributed over the
boundary surface of the solution phase on that side, and is not part of the bulk phase
composition.

12.8 L IQUID G AS E QUILIBRIA
400
The pressure difference p p at equilibrium can be estimated with Eq. 12.7.11, and
for the present example is found to be positive. Without this pressure difference, the solution
in phase would move spontaneously through the membrane into phase until phase
completely disappears. With phase open to the atmosphere, as in Fig. 12.9, the volume
of phase must be constrained in order to allow its pressure to differ from atmospheric
pressure. If the volume of phase remains practically constant, the transfer of a minute
quantity of solvent across the membrane is sufcient to cause the pressure difference.
It should be clear that the existence of a Donnan membrane equilibrium introduces
complications that would make it difcult to use a measured pressure difference to estimate
the molar mass of the polyelectrolyte by the method of Sec. 12.4, or to study the binding of
a charged ligand by equilibrium dialysis.
12.8 LIQUIDGAS EQUILIBRIA

This section describes multicomponent systems in which a liquid phase is equilibrated with
a gas phase.
12.8.1 Effect of liquid pressure on gas fugacity

If we vary the pressure of a liquid mixture at constant temperature and composition, there is
a small effect on the fugacity of each volatile component in an equilibrated gas phase. One
way to vary the pressure at essentially constant liquid composition is to change the partial
pressure of a component of the gas phase that has negligible solubility in the liquid.
At transfer equilibrium, component i has the same chemical potential in both phases:
i (l) D i (g). Combining the relations @ i (l)=@pT;fnig D Vi (l) and i (g) D i (g) C
/ (Eqs. 9.2.49 and 9.3.12), we obtain
RT ln.fi =p
d ln.fi =p /
V (l)
D i
dp
RT
(12.8.1)
(equilibrated liquid and
gas mixtures, constant T
and liquid composition)
Equation 12.8.1 shows that an increase in pressure, at constant temperature and liquid composition, causes an increase in the fugacity of each component in the gas phase.
Integration of Eq. 12.8.1 between pressures p1 and p2 yields
"Z
#
p2
Vi (l)
fi .p2 / D fi .p1 / exp
dp
(12.8.2)
p1 RT

The exponential on the right side is called the Poynting factor.

401
The integral in the Poynting factor is simplied if we make the approximation that Vi (l)
is independent of pressure. Then we obtain the approximate relation
V (l).p2 p1 /
fi .p2 / fi .p1 / exp i
(12.8.3)
RT

The effect of pressure on fugacity is usually small, and can often be neglected. For
typical values of the partial molar volume Vi (l), the exponential factor is close to unity
unless jp2 p1 j is very large. For instance, for Vi (l)D100 cm3 mol 1 and T D300 K,
we obtain a value for the ratio fi .p2 /=fi .p1 / of 1:004 if p2 p1 is 1 bar, 1:04 if p2 p1
is 10 bar, and 1:5 if p2 p1 is 100 bar. Thus, unless the pressure change is large, we
can to a good approximation neglect the effect of total pressure on fugacity. This
statement applies only to the fugacity of a substance in a gas phase that is equilibrated
with a liquid phase of constant composition containing the same substance. If the
liquid phase is absent, the fugacity of i in a gas phase of constant composition is of
course approximately proportional to the total gas pressure.
We can apply Eqs. 12.8.2 and 12.8.3 to pure liquid A, in which case Vi (l) is the molar volume VA (l). Suppose we have pure liquid A in equilibrium with pure gaseous A at
a certain temperature. This is a one-component, two-phase equilibrium system with one
degree of freedom (Sec. 8.1.7), so that at the given temperature the value of the pressure is
xed. This pressure is the saturation vapor pressure of pure liquid A at this temperature.
We can make the pressure p greater than the saturation vapor pressure by adding a second
substance to the gas phase that is essentially insoluble in the liquid, without changing the
temperature or volume. The fugacity fA is greater at this higher pressure than it was at the
saturation vapor pressure. The vapor pressure pA , which is approximately equal to fA , has
now become greater than the saturation vapor pressure. It is, however, safe to say that the
difference is negligible unless the difference between p and pA is much greater than 1 bar.
As an application of these relations, consider the effect of the size of a liquid droplet on
the equilibrium vapor pressure. The calculation of Prob. 12.8(b) shows that the fugacity of
H2 O in a gas phase equilibrated with liquid water in a small droplet is slightly greater than
when the liquid is in a bulk phase. The smaller the radius of the droplet, the greater is the
fugacity and the vapor pressure.
12.8.2 Effect of liquid composition on gas fugacities

Consider system 1 in Fig. 9.5 on page 247. A binary liquid mixture of two volatile components, A and B, is equilibrated with a gas mixture containing A, B, and a third gaseous
component C of negligible solubility used to control the total pressure. In order for A and
B to be in transfer equilibrium, their chemical potentials must be the same in both phases:
A (l)
A (g)
C RT ln
fA
p
B (l)
B (g)
C RT ln
fB
p
(12.8.4)
Suppose we make an innitesimal change in the liquid composition at constant T and

402
p. This causes innitesimal changes in the chemical potentials and fugacities:

d
A (l)
D RT
dfA
fA
D RT
B (l)
dfB
fB
By inserting these expressions in the GibbsDuhem equation xA d

9.2.43), we obtain
xA
df D
fA A
xB
df
fB B
(12.8.5)
A
xB d
(Eq.
(12.8.6)
(binary liquid mixture equilibrated
with gas, constant T and p)
This equation is a relation between changes in gas-phase fugacities caused by a change in

the liquid-phase composition. It shows that a composition change at constant T and p that
increases the fugacity of A in the equilibrated gas phase must decrease the fugacity of B.
Now let us treat the liquid mixture as a binary solution with component B as the solute.
In the ideal-dilute region, at constant T and p, the solute obeys Henrys law for fugacity:
fB D kH,B xB
(12.8.7)
For composition changes in the ideal-dilute region, we can write

dfB
f
D kH,B D B
dxB
xB
With the substitution dxB D
(12.8.8)
dxA and rearrangement, Eq. 12.8.8 becomes

xB
df D dxA
fB B
(12.8.9)
Combined with Eq. 12.8.6, this is .xA =fA / dfA D dxA , which we can rearrange and integrate as follows within the ideal-dilute region:
Z
0
fA
fA
dfA
D
fA
0
xA
dxA
xA
ln
0
fA
0
D ln xA
fA
(12.8.10)
The result is
fA D xA fA
(12.8.11)
(ideal-dilute binary solution)
Here fA is the fugacity of A in a gas phase equilibrated with pure liquid A at the same T
and p as the mixture. Equation 12.8.11 is Raoults law for fugacity applied to component
A.
If component B obeys Henrys law at all compositions, then the Henrys law constant
kH,B is equal to fB and B obeys Raoults law, fB D xB fB , over the entire range of xB .
We can draw two conclusions:
1. In the ideal-dilute region of a binary solution, where the solute obeys Henrys law,
the solvent must obey Raoults law. (A similar result was derived in Sec. 9.4.6 for a
solution with any number of solutes.)

kPa
403
kPa and
Figure 12.10 Fugacities in a gas phase equilibrated with a binary liquid mixture of
H2 O (A) and ethanol (B) at 25 C and 1 bar. a The dashed lines show Raoults law
behavior. The dotted lines illustrate the inequality .dfB = dxB / < .fB =xB /.
a Based
2. If one component of a binary liquid mixture obeys Raoults law at all compositions,
so also must the other component. This is the denition of an ideal binary liquid
mixture (Sec. 9.4.2).
Suppose we have a nonideal binary liquid mixture in which component B exhibits positive deviations from Raoults law. An example of this behavior for the waterethanol system
is shown in Fig. 12.10. At each point on the curve of fB versus xB , the slope dfB = dxB is
less than the slope fB =xB of a line drawn from the origin to the point (as illustrated by
the open circles and dotted lines in the gure), except that the two slopes become equal at
xB D1:
dfB
f
B
(12.8.12)
dxB
xB
As we can see from the gure, this relation must apply to any component whose fugacity
curve exhibits a positive deviation from Raoults law and has only one inection point.
Algebraic operations on an inequality must be carried out with care: multiplying both
sides by a quantity that can be negative may change the inequality to one with the wrong
sign. In order to simplify manipulation of the inequality of Eq. 12.8.12, it helps to convert
it to the following equality:11
f
dfB
CD D B
(12.8.13)
dxB
xB
Here D represents the difference between fB =xB and dfB = dxB ; its value is a function
of xB and is, according to Eq. 12.8.12, either positive or zero. We make the substitution
dxB D dxA and rearrange to
xB dfB
D dxA
fB C DxB
11 This
(12.8.14)
procedure is similar to the rectication procedure described on page 146.

404
kPa and
kPa
Figure 12.11 Fugacities in a gas phase equilibrated with a binary liquid mixture of
chloroform (A) and ethanol (B) at 35 C (Ref. [147]).
When D is zero, this equation becomes xB dfB =fB D dxA . When D is positive, the
left side of the equation is less than xB dfB =fB and is equal to dxA , so that dxA is less
than xB dfB =fB . Since D cannot be negative, Eq. 12.8.14 is equivalent to the following
relation:
xB
df
dxA
(12.8.15)
fB B
A substitution from Eq. 12.8.6 gives us
xA
df
fA A
dxA
or
dfA
fA
dxA
xA
(12.8.16)
We can integrate both sides of the second relation as follows:12

Z
0
fA
fA
dfA
fA
0
xA
dxA
xA
ln
0
fA
fA
0
ln xA
fA
xA f A
(12.8.17)
Thus, if the curve of fugacity versus mole fraction for one component of a binary liquid
mixture exhibits only positive deviations from Raoults law, with only one inection point,
so also must the curve of the other component. In the waterethanol system shown in
Fig. 12.10, both curves have positive deviations from Raoults law, and both have a single
inection point.
By the same method, we nd that if the fugacity curve of one component has only
negative deviations from Raoults law with a single inection point, the same is true of the
other component.
Figure 12.11 illustrates the case of a binary mixture in which component B has only
positive deviations from Raoults law, whereas component A has both positive and negative
deviations (fA is slightly less than xA fA for xB less than 0.3). This unusual behavior is
12 The
equalities are the same as Eqs. 12.8.10 and 12.8.11, with the difference that here xA is not restricted to
the ideal-dilute region.

405
possible because both fugacity curves have two inection points instead of the usual one.
Other types of unusual nonideal behavior are possible.13
12.8.3 The DuhemMargules equation

When we divide both sides of Eq. 12.8.6 by dxA , we obtain the DuhemMargules equation:
xA dfA
D
fA dxA
xB dfB
fB dxA
(12.8.18)
with gas, constant T and p)
If we assume the gas mixture is ideal, the fugacities are the same as the partial pressures,
and the DuhemMargules equation then becomes
xA dpA
D
pA dxA
xB dpB
pB dxA
(12.8.19)
with ideal gas, constant T and p)
Solving Eq. 12.8.19 for dpB = dxA , we obtain

dpB
D
dxA
xA pB dpA
xB pA dxA
(12.8.20)
To a good approximation, by assuming an ideal gas mixture and neglecting the effect
of total pressure on fugacity, we can apply Eq. 12.8.20 to a liquidgas system in which the
total pressure is not constant, but instead is the sum of pA and pB . Under these conditions,
we obtain the following expression for the rate at which the total pressure changes with the
liquid composition at constant T :
dp
d.pA C pB /
dpA xA pB dpA
dpA
xA =xB
D
D
D
1
dxA
dxA
dxA xB pA dxA
dxA
pA =pB
dpA
D
1
dxA
xA =xB
yA =yB
(12.8.21)
Here yA and yB are the mole fractions of A and B in the gas phase given by yA D pA =p
and yB D pB =p.
We can use Eq. 12.8.21 to make several predictions for a binary liquidgas system at
constant T .
If the ratio yA =yB is greater than xA =xB (meaning that the mole fraction of A is
greater in the gas than in the liquid), then .xA =xB /=.yA =yB / is less than 1 and
dp= dxA must have the same sign as dpA = dxA , which is positive.
Conversely, if yA =yB is less than xA =xB (i.e., the mole fraction of B is greater in the
gas than in the liquid), then dp= dxA must be negative.
13 Ref.
[111].

406
Thus compared to the liquid, the gas phase is richer in the component whose addition to the liquid at constant temperature causes the total pressure to increase. This
statement is a version of Konowaloffs rule.
In some binary liquidgas systems, the total pressure at constant temperature exhibits
a maximum or minimum at a particular liquid composition. At this composition, dp= dxA
is zero but dpA = dxA is positive. From Eq. 12.8.21, we see that at this composition xA =xB
must equal yA =yB , meaning that the liquid and gas phases have identical mole fraction
compositions. The liquid with this composition is called an azeotrope. The behavior of
systems with azeotropes will be discussed in Sec. 13.2.5.
12.8.4 Gas solubility

For the solution process B(g) ! B(sln), the general expression for the thermodynamic equilibrium constant is K D aB (sln)=aB (g).
The activity of B in the gas phase is given by aB (g) D fB =p . If the solute is a nonelectrolyte and we choose a standard state based on mole fraction, the activity in the solution is
aB (sln) D x;B x;B xB . The equilibrium constant is then given by
KD
x;B x;B xB
fB =p
(12.8.22)
and the solubility, expressed as the equilibrium mole fraction of solute in the solution, is
given by
xB D
KfB =p
x;B x;B
(12.8.23)
(nonelectrolyte solute in
equilibrium with gas)
At a xed T and p, the values of K and x;B are constant. Therefore any change in the
solution composition that increases the value of the activity coefcient x;B will decrease
the solubility for the same gas fugacity. This solubility decrease is often what happens
when a salt is dissolved in an aqueous solution, and is known as the salting-out effect (Prob.
12.11).
Unless the pressure is much greater than p , we can with negligible error set the pressure factor x;B equal to 1. When the gas solubility is low and the solution contains no
other solutes, the activity coefcient x;B is close to 1. If furthermore we assume ideal gas
behavior, then Eq. 12.8.23 becomes
xB D K
pB
p
(12.8.24)
(nonelectrolyte solute in equilibrium
with ideal gas, x;B D1, x;B D1)
The solubility is predicted to be proportional to the partial pressure. The solubility of a gas
that dissociates into ions in solution has a quite different dependence on partial pressure.
An example is the solubility of gaseous HCl in water to form an electrolyte solution, shown
in Fig. 10.1 on page 286.

407
If the actual conditions are close to those assumed for Eq. 12.8.24, we can use Eq.
12.1.13 to derive an expression for the temperature dependence of the solubility for a xed
partial pressure of the gas:
sol,B H
@ ln xB
d ln K
D
(12.8.25)
D
@T pB
dT
RT 2
At the standard pressure, sol,B H is the same as the molar enthalpy of solution at innite
dilution.
Since the dissolution of a gas in a liquid is invariably an exothermic process, sol,B H
is negative, and Eq. 12.8.25 predicts the solubility decreases with increasing temperature.
Note the similarity of Eq. 12.8.25 and the expressions derived previously for the temperature dependence of the solubilities of solids (Eq. 12.5.8) and liquids (Eq. 12.6.3). When
we substitute the mathematical identity dT D T 2 d.1=T /, Eq. 12.8.25 becomes
sol,B H
@ ln xB
D
(12.8.26)
@.1=T / pB
R
We can use this form to evaluate sol,B H from a plot of ln xB versus 1=T .
The ideal solubility of a gas is the solubility calculated on the assumption that the
dissolved gas obeys Raoults law for partial pressure: pB D xB pB . The ideal solubility,
expressed as a mole fraction, is then given as a function of partial pressure by
p
(12.8.27)
xB D B
pB
(ideal solubility of a gas)
Here pB is the vapor pressure of pure liquid solute at the same temperature and total pressure
as the solution. If the pressure is too low for pure B to exist as a liquid at this temperature,
we can with little error replace pB with the saturation vapor pressure of liquid B at the same
temperature, because the effect of total pressure on the vapor pressure of a liquid is usually
negligible (Sec. 12.8.1). If the temperature is above the critical temperature of pure B, we
can estimate a hypothetical vapor pressure by extrapolating the liquidvapor coexistence
curve beyond the critical point.
We can use Eq. 12.8.27 to make several predictions regarding the ideal solubility of a
gas at a xed value of pB .
1. The ideal solubility, expressed as a mole fraction, is independent of the kind of solvent.
2. The solubility expressed as a concentration, cB , is lower the greater is the molar
volume of the solvent. This is because at constant xB , cB decreases as the solution
volume increases.
3. The more volatile is the pure liquid solute at a particular temperature (i.e., the greater
is pB ), the lower is the solubility.
4. The solubility decreases with increasing temperature, since pB increases.
Of course, these predictions apply only to solutions that behave approximately as ideal
liquid mixtures, but even for many nonideal mixtures the predictions are found to have
good agreement with experiment.
As an example of the general validity of prediction 1, Hildebrand and Scott14 list the
14 Ref.
[78], Chap. XV.

408
following solubilities of gaseous Cl2 in several dissimilar solvents at 0 C and a partial

pressure of 1:01 bar: xB D 0:270 in heptane, xB D 0:288 in SiCl4 , and xB D 0:298 in
CCl4 . These values are similar to one another and close to the ideal value pB =pB D
0:273.
12.8.5 Effect of temperature and pressure on Henrys law constants

Consider the solution process B(g) ! B(soln) for a nonelectrolyte solute B. The expression
for the thermodynamic equilibrium constant, with a solute standard state based on mole
fraction, is
a (sln)
x;B x;B xB
KD B
D
(12.8.28)
aB (g)
fB =p
The Henrys law constant kH,B is related to fB and xB by
kH,B D
fB
x;B xB
(12.8.29)
(see Table 9.4), and is therefore related to K as follows:

kH,B D
x;B p
(12.8.30)
(nonelectrolyte solute)
The pressure factor x;B is a function of T and p, and K is a function only of T . The value
of kH,B therefore depends on both T and p.
At the standard pressure p D 1 bar, the value of x;B is unity, and Eqs. 12.1.13 and
12.1.14 then give the following expressions for the dependence of the dimensionless quantity kH,B =p on temperature:
d ln.kH,B =p /
D
dT
d ln K
D
dT
sol,B H
RT 2
(12.8.31)
d ln.kH,B =p /
D
d.1=T /
sol,B H
d ln K
D
d.1=T /
R
(12.8.32)
(pDp )
(pDp )
These expressions can be used with little error at any pressure that is not much greater than
p , say up to at least 2 bar, because under these conditions x;B does not differ appreciably
from unity (page 275).
To nd the dependence of kH,B on pressure, we substitute x;B in Eq. 12.8.30 with the
expression for x;B at pressure p 0 found in Table 9.6:
Z p0 1 !
0
VB
p
x;B .p / p
0
kH,B .p / D
D
exp
dp
(12.8.33)
RT
K
K
p
We can use Eq. 12.8.33 to compare the values of kH,B at the same temperature and two
different pressures, p1 and p2 :
Z p2 1 !
VB
kH,B .p2 / D kH,B .p1 / exp
dp
(12.8.34)
p1 RT

12.9 R EACTION E QUILIBRIA
409
Table 12.2 Expressions for activities (from Table 9.5 and Eqs.
10.1.14 and 10.3.16).
Species
Activity
Pure gas
a(g) D
Pure liquid or solid
aD
ai (g) D
Substance i in a liquid or
solid mixture
ai D
aA D
Nonelectrolyte solute B,
mole fraction basis
ax;B D
x;B
x;B
concentration basis
ac;B D
c;B
cB
c;B
c
molality basis
am;B D
m;B
m;B
Electrolyte solute B
am;B D
m;B
Ion in solution
aC D
f
p
fi
p
i
xi
A
xA
C C
mC
m
xB
mB
m
mC
m
a D
m
m
m
m
1
An approximate version of this relation, found by treating VB as independent of pressure,
is
1
V .p2 p1 /
(12.8.35)
kH,B .p2 / kH,B .p1 / exp B
RT
Unless jp2 p1 j is much greater than 1 bar, the effect of pressure on kH,B is small; see
Prob. 12.12 for an example.
12.9 REACTION EQUILIBRIA

The denition of the thermodynamic equilibrium constant of a reaction or other chemical
process is given by Eq. 11.8.9:
Y
i
KD
.ai /eq
(12.9.1)
i
The activity ai of each reactant or product species is based on an appropriate standard state.
We can replace each activity on the right side of Eq. 12.9.1 by an expression in Table 12.2.
For example, consider the following heterogeneous equilibrium that is important in the
formation of limestone caverns:
CaCO3 .cr; calcite/ C CO2 .g/ C H2 O.sln/ Ca2C .aq/ C 2HCO3 .aq/

12.9 R EACTION E QUILIBRIA
410
If we treat H2 O as a solvent and Ca2C and HCO3 as the solute species, then we write the
thermodynamic equilibrium constant as follows:
KD
aC a2
D
aCaCO3 aCO2 aH2 O
2m
fCO2
Cm
=p
2 =.m /3
H2 O xH2 O
(12.9.2)
The subscripts C and refer to the Ca2C and HCO3 ions, and all quantities are for the
system at reaction equilibrium. r is the proper quotient of pressure factors, given for this
reaction by15
r
(12.9.3)
CaCO3 H2 O
Unless the pressure is very high, we can with little error set the value of r equal to unity.
Equation 12.9.2 is an example of a mixed equilibrium constantone using more
than one kind of standard state. From the denition of the mean ionic activity coefcient
3
(Eq. 10.3.7), we can replace the product C 2 by , where is the mean ionic activity
coefcient of aqueous Ca(HCO3 )2 :
KD
3
2
3
mC m =.m /
fCO2 =p H2 O xH2 O
(12.9.4)
Instead of treating the aqueous Ca2C and HCO3 ions as solute species, we can regard the
dissolved Ca(HCO3 )2 electrolyte as the solute and write
KD
am;B
aCaCO3 aCO2 aH2 O
(12.9.5)
We then obtain Eq. 12.9.4 by replacing am;B with the expression in Table 12.2 for an electrolyte solute.
The value of K depends only on T , and the value of r depends only on T and p.
Suppose we dissolve some NaCl in the aqueous phase while maintaining the system at
constant T and p. The increase in the ionic strength will alter and necessarily cause a
compensating change in the solute molarity in order for the system to remain in reaction
equilibrium.
An example of a different kind of reaction equilibrium is the dissociation (ionization)
of a weak monoprotic acid such as acetic acid
HA.aq/ HC .aq/ C A .aq/
for which the thermodynamic equilibrium constant (the acid dissociation constant) is
Ka D
mC m
D
m,HA mHA m
C
2
mC m
m,HA mHA m
(12.9.6)
Suppose the solution is prepared from water and the acid, and HC from the dissociation of
H2 O is negligible compared to HC from the acid dissociation. We may then write mC D
product C 2 in the numerator of Eq. 12.9.3 is the pressure factor
Eq. 10.3.11 on page 293).
15 The
m;B
for the solute Ca(HCO3 )2 (see

12.10 E VALUATION OF S TANDARD M OLAR Q UANTITIES
411
m D mB , where is the degree of dissociation and mB is the overall molality of the

acid. The molality of the undissociated acid is mHA D .1 /mB , and the dissociation
constant can be written
2 2 m =m
B
Ka D r
(12.9.7)
/
m,HA .1
From this equation, we see that a change in the ionic strength that decreases when T , p,
and mB are held constant must increase the degree of dissociation (Prob. 12.17).
12.10 EVALUATION OF STANDARD MOLAR QUANTITIES

Some of the most useful experimentally-derived data for thermodynamic calculations are
values of standard molar reaction enthalpies, standard molar reaction Gibbs energies, and
standard molar reaction entropies. The values of these quantities for a given reaction are
related, as we know (Eq. 11.8.21), by
r G D r H
Tr S
(12.10.1)
and r S can be calculated from the standard molar entropies of the reactants and products
using Eq. 11.8.22:
X
r S D
(12.10.2)
i Si
i
The standard molar quantities appearing in Eqs. 12.10.1 and 12.10.2 can be evaluated
through a variety of experimental techniques. Reaction calorimetry can be used to evaluate
r H for a reaction (Sec. 11.5). Calorimetric measurements of heat capacity and phasetransition enthalpies can be used to obtain the value of Si for a solid or liquid (Sec. 6.2.1).
For a gas, spectroscopic measurements can be used to evaluate Si (Sec. 6.2.2). Evaluation
of a thermodynamic equilibrium constant and its temperature derivative, for any of the
kinds of equilibria discussed in this chapter (vapor pressure, solubility, chemical reaction,
etc.), can provide values of r G and r H through the relations r G D RT ln K and
r H D R d ln K= d.1=T /.
In addition to these methods, measurements of cell potentials are useful for a reaction
that can be carried out reversibly in a galvanic cell. Section 14.3.3 will describe how the
standard cell potential and its temperature derivative allow r H , r G , and r S to be
evaluated for such a reaction.
An efcient way of tabulating the results of experimental measurements is in the form
of standard molar enthalpies and Gibbs energies of formation. These values can be used to
generate the values of standard molar reaction quantities for reactions not investigated directly. The relations between standard molar reaction and formation quantities (Sec. 11.3.2)
are
X
X
r H D
r G D
(12.10.3)
i f H .i/
i f G .i/
i
and for ions the conventions used are

f H (HC , aq) D 0
f G (HC , aq) D 0
Sm (HC , aq) D 0
(12.10.4)
Appendix H gives an abbreviated set of values of f H , Sm , and f G at 298:15 K.


12.10 E VALUATION OF S TANDARD M OLAR Q UANTITIES
412
For examples of the evaluation of standard molar reaction quantities and standard molar formation quantities from measurements made by various experimental techniques, see
Probs. 12.1812.20, 14.3, and 14.4.

P ROBLEMS
413
PROBLEMS
in Appendix I.
12.1 Consider the heterogeneous equilibrium CaCO3 .s/ CaO.s/ C CO2 .g/. Table 12.3 lists
pressures measured over a range of temperatures for this system.
Table 12.3 Pressure of an equilibrium system
containing CaCO3 (s), CaO(s), and CO2 (g) a
t= C
p=Torr
t= C
p=Torr
842:3
852:9
854:5
868:9
343:0
398:6
404:1
510:9
904:3
906:5
937:0
937:0
879:0
875:0
1350
1340
a Ref.
[151].
(a) What is the approximate relation between p and K?

(b) Plot these data in the form ln K versus 1=T , or t ln K to a linear function of 1=T . Then,
evaluate the temperature at which the partial pressure of the CO2 is 1 bar, and the standard
molar reaction enthalpy at this temperature.
12.2 For a homogeneous reaction in which the reactants and products are solutes in a solution,
write a rigorous relation between the standard molar reaction enthalpy and the temperature
dependence of the thermodynamic equilibrium constant, with solute standard states based on
concentration.
12.3 Derive an expression for the standard molar reaction entropy of a reaction that can be used to
calculate its value from the thermodynamic equilibrium constant and its temperature derivative.
Assume that no solute standard states are based on concentration.
Table 12.4 Properties of H2 O at 1 bar
M
18:0153 g mol
tf
1
tb
0:00 C
99:61 C
fus H
6:010 kJ mol
vap H
1
40:668 kJ mol
12.4 Use the data in Table 12.4 to evaluate the molal freezing-point depression constant and the
molal boiling-point elevation constant for H2 O at a pressure of 1 bar.
12.5 An aqueous solution of the protein bovine serum albumin, containing 2:00 10 2 g of protein
per cubic centimeter, has an osmotic pressure of 8:1 10 3 bar at 0 C. Estimate the molar
mass of this protein.
12.6 Figure 12.8 on page 394 shows a curve tted to experimental points for the aqueous solubility
of n-butylbenzene. The curve has the equation ln xB D a.t = C b/2 C c, where the constants
have the values a D 3:34 10 4 , b D 12:13, and c D 13:25. Assume that the saturated
solution behaves as an ideal-dilute solution, use a solute standard state based on mole fraction, and calculate sol,B H and sol,B S at 5:00 C, 12:13 C (the temperature of minimum
solubility), and 25:00 C.

P ROBLEMS
414
12.7 Consider a hypothetical system in which two aqueous solutions are separated by a semipermeable membrane. Solution is prepared by dissolving 1:00 10 5 mol KCl in 10:0 g water.
Solution is prepared from 1:00 10 5 mol KCl and 1:00 10 6 mol of the potassium salt of
a polyelectrolyte dissolved in 10:0 g water. All of solution is used to ll a dialysis bag, which
is then sealed and placed in solution .
Each polyelectrolyte ion has a charge of 10. The membrane of the dialysis bag is permeable
to the water molecules and to the KC and Cl ions, but not to the polyelectrolyte. The system
comes to equilibrium at 25:00 C. Assume that the volume of the dialysis bag remains constant.
Also make the drastic approximation that both solutions behave as ideal-dilute solutions.
(a) Find the equilibrium molality of each solute species in the two solution phases.
(b) Describe the amounts and directions of any macroscopic transfers of ions across the membrane that are required to establish the equilibrium state.
(c) Estimate the Donnan potential,
(d) Estimate the pressure difference across the membrane at equilibrium. (The density of
liquid H2 O at 25:00 C is 0:997 g cm 3 .)
12.8 The derivation of Prob. 9.3 on page 281 shows that the pressure in a liquid droplet of radius r is
greater than the pressure of the surrounding equilibrated gas phase by a quantity 2 =r, where
is the surface tension.
(a) Consider a droplet of water of radius 1:00 10 6 m at 25 C suspended in air of the same
temperature. The surface tension of water at this temperature is 0:07199 J m 2 . Find the
pressure in the droplet if the pressure of the surrounding air is 1:00 bar.
(b) Calculate the difference between the fugacity of H2 O in the air of pressure 1:00 bar equilibrated with this water droplet, and the fugacity in air equilibrated at the same temperature
and pressure with a pool of liquid water having a at surface. Liquid water at 25 C and
1 bar has a vapor pressure of 0:032 bar and a molar volume of 1:807 10 5 m3 mol 1 .
12.9 For a solution process in which species B is transferred from a gas phase to a liquid solution,
nd the relation between sol G (solute standard state based on mole fraction) and the Henrys
law constant kH,B .
12.10 Crovetto16 reviewed the published data for the solubility of gaseous CO2 in water, and tted
the Henrys law constant kH,B to a function of temperature. Her recommended values of kH,B at
ve temperatures are 1233 bar at 15:00 C, 1433 bar at 20:00 C, 1648 bar at 25:00 C, 1874 bar
at 30:00 C, and 2111 bar at 35 C.
(a) The partial pressure of CO2 in the atmosphere is typically about 3 10 4 bar. Assume
a fugacity of 3:0 10 4 bar, and calculate the aqueous solubility at 25:00 C expressed
both as a mole fraction and as a molality.
(b) Find the standard molar enthalpy of solution at 25:00 C.
(c) Dissolved carbon dioxide exists mostly in the form of CO2 molecules, but a small fraction
exists as H2 CO3 molecules, and there is also some ionization:
CO2 .aq/ C H2 O.l/ ! HC .aq/ C HCO3 .aq/
(The equilibrium constant of this reaction is often called the rst ionization constant of
carbonic acid.) Combine the kH,B data with data in Appendix H to evaluate K and r H
for the ionization reaction at 25:00 C. Use solute standard states based on molality, which
are also the solute standard states used for the values in Appendix H.
16 Ref.
[40].

P ROBLEMS
415
12.11 The solubility of gaseous O2 at a partial pressure of 1:01 bar and a temperature of 310:2 K, expressed as a concentration, is 1:07 10 3 mol dm 3 in pure water and 4:68 10 4 mol dm 3
in a 3:0 M aqueous solution of KCl.17 This solubility decrease is the salting-out effect. Calculate the activity coefcient c;B of O2 in the KCl solution.
12.12 At 298:15 K, the partial molar volume of CO2 (aq) is 33 cm3 mol 1 . Use Eq. 12.8.35 to estimate the percent change in the value of the Henrys law constant kH,B for aqueous CO2 at
298:15 K when the total pressure is changed from 1:00 bar to 10:00 bar.
12.13 Rettich et al18 made high-precision measurements of the solubility of gaseous oxygen (O2 ) in
water. Each measurement was made by equilibrating water and oxygen in a closed vessel for
a period of up to two days, at a temperature controlled within 0:003 K. The oxygen was
extracted from samples of known volume of the equilibrated liquid and gas phases, and the
amount of O2 in each sample was determined from p-V -T measurements taking gas nonideality into account. It was then possible to evaluate the mole fraction xB of O2 in the liquid phase
g
and the ratio .nB =V g / for the O2 in the gas phase.
Table 12.5 Data for Problem 12.13 (A = H2 O, B = O2 )
T D 298:152 K
xB D 2:02142
Second virial coefcients:

5
10
.nB =V g / D 35:9957 mol m

pA D 3167:13 Pa
VA D 18:069
1
VB D 31:10
BAA D
BBB D
10
m3 mol
10
m3 mol
BAB D
1152
10
m3 mol
16:2
10
m3 mol
27:0
10
m3 mol
Table 12.5 gives values of physical quantities at T D 298:152 K needed for this problem. The
g
values of xB and .nB =V g / were obtained by Rettich et al from samples of liquid and gas phases
equilibrated at temperature T , as explained above. pA is the saturation vapor pressure of pure
liquid water at this temperature.
Your calculations will be similar to those used by Rettich et al to obtain values of the Henrys
law constant of oxygen to six signicant gures. Your own calculations should also be carried
out to six signicant gures. For the gas constant, use the value R D 8:31447 J K 1 mol 1 .
The method you will use to evaluate the Henrys law constant kH,B D fB =xB at the experimental temperature and pressure is as follows. The value of xB is known, and you need to nd the
fugacity fB of the O2 in the gas phase. fB can be calculated from B and pB . These in turn
can be calculated from the pressure p, the mole fraction yB of O2 in the gas phase, and known
values of second virial coefcients. You will calculate p and yB by an iterative procedure.
Assume the gas has the virial equation of state .V g =ng / D .RT =p/ C B (Eq. 9.3.21) and use
relevant relations in Sec. 9.3.4.
(a) For the equilibrated liquid-gas system, calculate initial approximate values of p and yB
g
by assuming that pA is equal to pA and pB is equal to .nB =V g /RT .
(b) Use your approximate values of p and yB from part (a) to calculate
coefcient of A in the gas mixture.
A,
the fugacity
(c) Evaluate the fugacity fA of the H2 O in the gas phase. Assume p, yB , and A have the
values you calculated in parts (a) and (b). Hint: start with the value of the saturation vapor
pressure of pure water.
17 Ref.
[105].
18 Ref.
[141].

P ROBLEMS
(d) Use your most recently calculated values of p,
of yB .
A,
416
and fA to calculate an improved value
(e) Use your current values of p and yB to evaluate the compression factor Z of the gas
mixture, taking nonideality into account.
g
(f) Derive a general expression for p as a function of .nB =V g /, T , yB , and Z. Use this
expression to calculate an improved value of p.
(g) Finally, use the improved values of p and yB to evaluate the Henrys law constant kH,B at
the experimental T and p.
12.14 The method described in Prob. 12.13 has been used to obtain high-precision values of the
Henrys law constant, kH,B , for gaseous methane dissolved in water.19 Table 12.6 lists values
Table 12.6 Data for Prob. 12.14
1=.T =K/
ln .kH,B =p /
1=.T =K/
ln .kH,B =p /
0:00363029
0:00359531
0:00352175
0:00347041
0:00341111
0:00335390
10:0569
10:1361
10:2895
10:3883
10:4951
10:5906
0:00329870
0:00319326
0:00314307
0:00309444
0:00304739
10:6738
10:8141
10:8673
10:9142
10:9564
of ln .kH,B =p / at eleven temperatures in the range 275 K328 K and at pressures close to 1 bar.
Use these data to evaluate sol,B H and sol,B Cp at T D 298:15 K. This can be done by a
graphical method. Better precision will be obtained by making a least-squares t of the data to
the three-term polynomial
ln .kH,B =p / D a C b.1=T / C c.1=T /2
and using the values of the coefcients a, b, and c for the evaluations.
12.15 Liquid water and liquid benzene have very small mutual solubilities. Equilibria in the binary
waterbenzene system were investigated by Tucker, Lane, and Christian20 as follows. A known
amount of distilled water was admitted to an evacuated, thermostatted vessel. Part of the water
vaporized to form a vapor phase. Small, precisely measured volumes of liquid benzene were
then added incrementally from the sample loop of a liquid-chromatography valve. The benzene
distributed itself between the liquid and gaseous phases in the vessel. After each addition, the
pressure was read with a precision pressure gauge. From the known amounts of water and
benzene and the total pressure, the liquid composition and the partial pressure of the benzene
were calculated. The fugacity of the benzene in the vapor phase was calculated from its partial
pressure and the second virial coefcient.
At a xed temperature, for mole fractions xB of benzene in the liquid phase up to about
3 10 4 (less than the solubility of benzene in water), the fugacity of the benzene in the
equilibrated gas phase was found to have the following dependence on xB :
fB
D kH,B
xB
AxB
Here kH,B is the Henrys law constant and A is a constant related to deviations from Henrys
law. At 30 C, the measured values were kH,B D 385:5 bar and A D 2:24 104 bar.
19 Ref.
[142].
20 Ref.
[157].

P ROBLEMS
417
(a) Treat benzene (B) as the solute and nd its activity coefcient on a mole fraction basis,
4
x;B , at 30 C in the solution of composition xB D 3:00 10 .
(b) The fugacity of benzene vapor in equilibrium with pure liquid benzene at 30 C is fB D
0:1576 bar. Estimate the mole fraction solubility of liquid benzene in water at this temperature.
(c) The calculation of x;B in part (a) treated the benzene as a single solute species with
deviations from innite-dilution behavior. Tucker et al suggested a dimerization model to
explain the observed negative deviations from Henrys law. (Classical thermodynamics, of
course, cannot prove such a molecular interpretation of observed macroscopic behavior.)
The model assumes that there are two solute species, a monomer (M) and a dimer (D), in
reaction equilibrium: 2M D. Let nB be the total amount of C6 H6 present in solution,
and dene the mole fractions
nB
nB
def
xB D
nA C nB
nA
def
xM D
nM
nA C nM C nD
nM
nA
def
xD D
nD
nA C nM C nD
nD
nA
where the approximations are for dilute solution. In the model, the individual monomer
and dimer particles behave as solutes in an ideal-dilute solution, with activity coefcients
of unity. The monomer is in transfer equilibrium with the gas phase: xM D fB =kH,B . The
equilibrium constant expression (using a mole fraction basis for the solute standard states
2
and setting pressure factors equal to 1) is K D xD =xM . From the relation nB D nM C2nD ,
and because the solution is very dilute, the expression becomes
KD
xB xM
2
2xM
Make individual calculations of K from the values of fB measured at xB D 1:00 10 4 ,

xB D 2:00 10 4 , and xB D 3:00 10 4 . Extrapolate the calculated values of K to
xB D0 in order to eliminate nonideal effects such as higher aggregates. Finally, nd the
fraction of the benzene molecules present in the dimer form at xB D 3:00 10 4 if this
model is correct.
12.16 Use data in Appendix H to evaluate the thermodynamic equilibrium constant at 298:15 K for
the limestone reaction
CaCO3 .cr; calcite/ C CO2 .g/ C H2 O.l/ ! Ca2C .aq/ C 2HCO3 .aq/
12.17 For the dissociation equilibrium of formic acid, HCO2 H.aq/ HC .aq/ C HCO2 .aq/, the
acid dissociation constant at 298:15 K has the value Ka D 1:77 10 4 .
(a) Use Eq. 12.9.7 to nd the degree of dissociation and the hydrogen ion molality in a
0.01000 molal formic acid solution. You can safely set r and m,HA equal to 1, and
use the DebyeH ckel limiting law (Eq. 10.4.8) to calculate . You can do this calu
culation by iteration: Start with an initial estimate of the ionic strength (in this case 0),
calculate and , and repeat these steps until the value of no longer changes.
(b) Estimate the degree of dissociation of formic acid in a solution that is 0.01000 molal in
both formic acid and sodium nitrate, again using the DebyeH ckel limiting law for .
u
Compare with the value in part (a).
12.18 Use the following experimental information to evaluate the standard molar enthalpy of formation and the standard molar entropy of the aqueous chloride ion at 298:15 K, based on the

P ROBLEMS
418
conventions f H .HC , aq/ D 0 and Sm .HC , aq/ D 0 (Secs. 11.3.2 and 11.8.4). (Your calculated values will be close to, but not exactly the same as, those listed in Appendix H, which are
based on the same data combined with data of other workers.)
1
1
For the reaction 2 H2 .g/ C 2 Cl2 .g/ ! HCl.g/, the standard molar enthalpy of reaction
at 298:15 K measured in a ow calorimeter21 is r H D 92:312 kJ mol 1 .
The standard molar entropy of gaseous HCl at 298:15 K calculated from spectroscopic
data is Sm D 186:902 J K 1 mol 1 .

From ve calorimetric runs,22 the average experimental value of the standard molar enthalpy of solution of gaseous HCl at 298:15 K is sol,B H D 74:84 kJ mol 1 .
From vapor pressure measurements of concentrated aqueous HCl solutions,23 the value
of the ratio fB =am;B for gaseous HCl in equilibrium with aqueous HCl at 298:15 K is
5:032 10 7 bar.
12.19 The solubility of crystalline AgCl in ultrapure water has been determined from the electrical
conductivity of the saturated solution.24 The average of ve measurements at 298:15 K is sB D
1:337 10 5 mol dm 3 . The density of water at this temperature is A D 0:9970 kg dm 3 .
(a) From these data and the DebyeH ckel limiting law, calculate the solubility product Ks
u
of AgCl at 298:15 K.
(b) Evaluate the standard molar Gibbs energy of formation of aqueous AgC ion at 298:15 K,
using the results of part (a) and the values f G .Cl , aq/ D 131:22 kJ mol 1 and
f G .AgCl, s/ D 109:77 kJ mol 1 from Appendix H.
12.20 The following reaction was carried out in an adiabatic solution calorimeter by Wagman and
Kilday:25
AgNO3 (s) C KCl(aq, mB D 0:101 mol kg
) ! AgCl(s) C KNO3 (aq)
The reaction can be assumed to go to completion, and the amount of KCl was in slight excess,
so the amount of AgCl formed was equal to the initial amount of AgNO3 . After correction
for the enthalpies of diluting the solutes in the initial and nal solutions to innite dilution,
the standard molar reaction enthalpy at 298:15 K was found to be r H D 43:042 kJ mol 1 .
The same workers used solution calorimetry to obtain the molar enthalpy of solution at innite
dilution of crystalline AgNO3 at 298:15 K: sol,B H 1 D 22:727 kJ mol 1 .
(a) Show that the difference of these two values is the standard molar reaction enthalpy for
the precipitation reaction
AgC .aq/ C Cl .aq/ ! AgCl.s/
and evaluate this quantity.
(b) Evaluate the standard molar enthalpy of formation of aqueous AgC ion at 298:15 K,
using the results of part (a) and the values f H .Cl , aq/ D 167:08 kJ mol 1 and
f H .AgCl, s/ D 127:01 kJ mol 1 from Appendix H. (These values come from calculations similar to those in Probs. 12.18 and 14.4.) The calculated value will be close to,
but not exactly the same as, the value listed in Appendix H, which is based on the same
data combined with data of other workers.
21 Ref.
[144].
22 Ref.
[72].
23 Ref.
[136].
24 Ref.
[67].
25 Ref.
[162].
C HAPTER 13
T HE P HASE RULE AND P HASE D IAGRAMS
We encountered the Gibbs phase rule and phase diagrams in Chap. 8 in connection with
single-substance systems. The present chapter derives the full version of the Gibbs phase
rule for multicomponent systems. It then discusses phase diagrams for some representative
types of multicomponent systems, and shows how they are related to the phase rule and to
equilibrium concepts developed in Chaps. 11 and 12.
13.1 THE GIBBS PHASE RULE FOR MULTICOMPONENT SYSTEMS

In Sec. 8.1.7, the Gibbs phase rule for a pure substance was written F D 3 P . We now
consider a system of more than one substance and more than one phase in an equilibrium
state. The phase rule assumes the system is at thermal and mechanical equilibrium. We
shall assume furthermore that in addition to the temperature and pressure, the only other
state functions needed to describe the state are the amounts of the species in each phase;
this means for instance that surface effects are ignored.
The derivations to follow will show that phase rule may be written either in the form
F D2CC
(13.1.1)
or
F D2Cs
(13.1.2)
where the symbols have the following meanings:

F = the number of degrees of freedom (or variance)
= the maximum number of intensive variables that can be varied independently while
the system remains in an equilibrium state;
C = the number of components
= the minimum number of substances (or xed-composition mixtures of substances) that
could be used to prepare each phase individually;
P = the number of different phases;
s = the number of different species;
r = the number of independent relations among intensive variables of individual phases
other than relations needed for thermal, mechanical, and transfer equilibrium.
419
CHAPTER 13 THE PHASE RULE AND PHASE DIAGRAMS

13.1 T HE G IBBS P HASE RULE FOR M ULTICOMPONENT S YSTEMS
420
If we subdivide a phase, that does not change the number of phases P . That is, we treat
noncontiguous regions of the system that have identical intensive properties as parts of the
same phase.
13.1.1 Degrees of freedom

Consider a system in an equilibrium state. In this state, the system has one or more phases;
each phase contains one or more species; and intensive properties such as T , p, and the
mole fraction of a species in a phase have denite values. Starting with the system in this
state, we can make changes that place the system in a new equilibrium state having the
same kinds of phases and the same species, but different values of some of the intensive
properties. The number of different independent intensive variables that we may change in
this way is the number of degrees of freedom or variance, F , of the system.
Clearly, the system remains in equilibrium if we change the amount of a phase without
changing its temperature, pressure, or composition. This, however, is the change of an
extensive variable and is not counted as a degree of freedom.
The phase rule, in the form to be derived, applies to a system that continues to have
complete thermal, mechanical, and transfer equilibrium as intensive variables change. This
means different phases are not separated by adiabatic or rigid partitions, or by semipermeable or impermeable membranes. Furthermore, every conceivable reaction among the
species is either at reaction equilibrium or else is frozen at a xed advancement during the
time period we observe the system.
The number of degrees of freedom is the maximum number of intensive properties
of the equilibrium system we may independently vary, or x at arbitrary values, without
causing a change in the number and kinds of phases and species. We cannot, of course,
change one of these properties to just any value whatever. We are able to vary the value
only within a certain nite (sometimes quite narrow) range before a phase disappears or a
new one appears.
The number of degrees of freedom is also the number of independent intensive variables needed to specify the equilibrium state in all necessary completeness, aside from the
amount of each phase. In other words, when we specify values of F different independent
intensive variables, then the values of all other intensive variables of the equilibrium state
have denite values determined by the physical nature of the system.
Just as for a one-component system, we can use the terms bivariant, univariant, and
invariant depending on the value of F (Sec. 8.1.7).
13.1.2 Species approach to the phase rule

This section derives an expression for the number of degrees of freedom, F , based on
species. Section 13.1.3 derives an expression based on components. Both approaches yield
equivalent versions of the phase rule.
Recall that a species is an entity, uncharged or charged, distinguished from other species
by its chemical formula (Sec. 9.1.1). Thus, CO2 and CO3 2 are different species, but
CO2 (aq) and CO2 (g) is the same species in different phases.
Consider an equilibrium system of P phases, each of which contains the same set of
species. Let the number of different species be s. If we could make changes while the

421
system remains in thermal and mechanical equilibrium, but not necessarily in transfer equilibrium, we could independently vary the temperature and pressure of the system as a whole
and the amount of each species in each phase; there would then be 2CP s independent variables.
The equilibrium system is, however, in transfer equilibrium, which requires each species
to have the same chemical potential in each phase: i D , i D , and so on. There
i
i
are P 1 independent relations like this for each species, and a total of s.P 1/ independent
relations for all species. Each such independent relation introduces a constraint and reduces
the number of independent variables by one. Accordingly, taking transfer equilibrium into
account, the number of independent variables is 2 C P s s.P 1/ D 2 C s.
We obtain the same result if a species present in one phase is totally excluded from
another. For example, solvent molecules of a solution are not found in a pure perfectlyordered crystal of the solute, undissociated molecules of a volatile strong acid such as HCl
can exist in a gas phase but not in aqueous solution, and ions of an electrolyte solute are
usually not found in a gas phase. For each such species absent from a phase, there is one
fewer amount variable and also one fewer relation for transfer equilibrium; on balance, the
number of independent variables is still 2 C s.
Next, we consider the possibility that further independent relations exist among intensive variables in addition to the relations needed for thermal, mechanical, and transfer
equilibrium.1 If there are r of these additional relations, the total number of independent
variables is reduced to 2 C s r. These relations may come from
1. reaction equilibria,
2. the requirement of electroneutrality in a phase containing ions, and
3. initial conditions determined by the way the system is prepared.
P
In the case of a reaction equilibrium, the relation is r G D i i i D 0, or the equivalent
Q
relation K D i .ai / i for the thermodynamic equilibrium constant. Thus, r is the sum of
the number of independent reaction equilibria, the number of phases containing ions, and
the number of independent initial conditions. Several examples will be given in Sec. 13.1.4.
There is an innite variety of possible choices of the independent variables (both extensive and intensive) for the equilibrium system, but the total number of independent variables
is xed at 2 C s r. Keeping intensive properties xed, we can always vary how much of
each phase is present (e.g., its volume, mass, or amount) without destroying the equilibrium. Thus, at least P of the independent variables, one for each phase, must be extensive.
It follows that the maximum number of independent intensive variables is the difference
.2 C s r/ P .
It may be that initial conditions establish relations among the amounts of phases, as
will be illustrated in example 2 on page 424. If present, these are relations among
extensive variables that are not counted in r. Each such independent relation decreases
the total number of independent variables without changing the number of independent
intensive variables calculated from .2 C s r/ P .
Since the maximum number of independent intensive variables is the number of degrees
P
pi D p for a gas phase or i xi D 1 for a phase in general have already been accounted
for in the derivation by the specication of p and the amount of each species.
1 Relations such as

422
of freedom, our expression for F based on species is

F D2Cs
(13.1.3)
13.1.3 Components approach to the phase rule

The derivation of the phase rule in this section uses the concept of components. The number
of components, C , is the minimum number of substances or mixtures of xed composition
from which we could in principle prepare each individual phase of an equilibrium state of
the system, using methods that may be hypothetical. These methods include the addition or
removal of one or more of the substances or xed-composition mixtures, and the conversion
of some of the substances into others by means of a reaction that is at equilibrium in the
actual system.
It is not always easy to decide on the number of components of an equilibrium system.
The number of components may be less than the number of substances present, on account
of the existence of reaction equilibria that produce some substances from others. When we
use a reaction to prepare a phase, nothing must remain unused. For instance, consider a
system consisting of solid phases of CaCO3 and CaO and a gas phase of CO2 . Assume the
reaction CaCO3 (s) ! CaO.s/ C CO2 .g/ is at equilibrium. We could prepare the CaCO3
phase from CaO and CO2 by the reverse of this reaction, but we can only prepare the CaO
and CO2 phases from the individual substances. We could not use CaCO3 to prepare either
the CaO phase or the CO2 phase, because CO2 or CaO would be left over. Thus this system
has three substances but only two components, namely CaO and CO2 .
In deriving the phase rule by the components approach, it is convenient to consider
only intensive variables. Suppose we have a system of P phases in which each substance
present is a component (i.e., there are no reactions) and each of the C components is present
in each phase. If we make changes to the system while it remains in thermal and mechanical
equilibrium, but not necessarily in transfer equilibrium, we can independently vary the temperature and pressure of the whole system, and for each phase we can independently vary
the mole fraction of all P one of the substances (the value of the omitted mole fraction
but
comes from the relation i xi D 1). This is a total of 2 C P .C 1/ independent intensive
variables.
When there also exist transfer and reaction equilibria, not all of these variables are independent. Each substance in the system is either a component, or else can be formed from
components by a reaction that is in reaction equilibrium in the system. Transfer equilibria
establish P 1 independent relations for each component ( i D , i D , etc.) and

i
i
a total of C.P 1/ relations for all components. Since these are relations among chemical
potentials, which are intensive properties, each relation reduces the number of independent
intensive variables by one. The resulting number of independent intensive variables is
F D 2 C P .C
1/
C.P
1/ D 2 C C
(13.1.4)
If the equilibrium system lacks a particular component in one phase, there is one fewer
mole fraction variable and one fewer relation for transfer equilibrium. These changes cancel
in the calculation of F , which is still equal to 2 C C P . If a phase contains a substance
that is formed from components by a reaction, there is an additional mole fraction variable
P
and also the additional relation i i i D 0 for the reaction; again the changes cancel.

423
We may need to remove a component from a phase to achieve the nal composition.
Note that it is not necessary to consider additional relations for electroneutrality or
initial conditions; they are implicit in the denitions of the components. For instance,
since each component is a substance of zero electric charge, the electrical neutrality of
the phase is assured.
We conclude that, regardless of the kind of system, the expression for F based on
components is given by F D 2 C C P . By comparing this expression and F D 2 C s
r P , we see that the number of components is related to the number of species by
C Ds
(13.1.5)
13.1.4 Examples
The ve examples below illustrate various aspects of using the phase rule.
Example 1: liquid water
For a single phase of pure water, P equals 1. If we treat the water as the single species
H2 O, s is 1 and r is 0. The phase rule then predicts two degrees of freedom:
F D2Cs
D2C1
1D2
(13.1.6)
Since F is the number of intensive variables that can be varied independently, we could for
instance vary T and p independently, or T and , or any other pair of independent intensive
variables.
Next let us take into account the proton transfer equilibrium
2 H2 O.l/ H3 OC .aq/ C OH .aq/
and consider the system to contain the three species H2 O, H3 OC , and OH . Then for the
species approach to the phase rule, we have s D 3. We can write two independent relations:
1. for reaction equilibrium, 2 H2 O C H3 OC C OH D 0;
2. for electroneutrality, mH3 OC D mOH .
Thus, we have two relations involving intensive variables only. Now s is 3, r is 2, P is 1,
and the number of degrees of freedom is given by
F D2Cs
P D2
(13.1.7)
which is the same value of F as before.

If we consider water to contain additional cation species (e.g., H5 O2 C ), each such
species would add 1 to s and 1 to r, but F would remain equal to 2. Thus, no matter
how complicated are the equilibria that actually exist in liquid water, the number of degrees
of freedom remains 2.
Applying the components approach to water is simple. All species that may exist in
pure water are formed, in whatever proportions actually exist, from the single substance
H2 O. Thus, there is only one component: C D 1. The component version of the phase
rule, F D 2 C C P , gives the same result as the species version: F D 2.

424
Example 2: carbon, oxygen, and carbon oxides

Consider a system containing solid carbon (graphite) and a gaseous mixture of O2 , CO, and
CO2 . There are four species and two phases. If reaction equilibrium is absent, as might be
the case at low temperature in the absence of a catalyst, we have r D 0 and C D s r D 4.
The four components are the four substances. The phase rule tells us the system has four
degrees of freedom. We could, for instance, arbitrarily vary T , p, yO2 , and yCO .
Now suppose we raise the temperature or introduce an appropriate catalyst to allow the
following reaction equilibria to exist:
1. 2 C.s/ C O2 .g/ 2 CO.g/
2. C.s/ C O2 .g/ CO2 .g/
These equilibria introduce two new independent relations among chemical potentials and
among activities. We could also consider the equilibrium 2 CO.g/ C O2 .g/ 2 CO2 .g/,
but it does not contribute an additional independent relation because it depends on the other
two equilibria: the reaction equation is obtained by subtracting the reaction equation for
equilibrium 1 from twice the reaction equation for equilibrium 2. By the species approach,
we have s D 4, r D 2, and P D 2; the number of degrees of freedom from these values is
F D2Cs
P D2
(13.1.8)
If we wish to calculate F by the components approach, we must decide on the minimum number of substances we could use to prepare each phase separately. (This does not
refer to how we actually prepare the two-phase system, but to a hypothetical preparation of
each phase with any of the compositions that can actually exist in the equilibrium system.)
Assume equilibria 1 and 2 are present. We prepare the solid phase with carbon, and we
can prepare any possible equilibrium composition of the gas phase from carbon and O2 by
using the reactions of both equilibria. Thus, there are two components (C and O2 ) giving
the same result of two degrees of freedom.
What is the signicance of there being two degrees of freedom when the reaction equilibria are present? There are two ways of viewing the situation:
1. We can arbitrarily vary the two intensive variables T and p. When we do, the mole
fractions of the three substances in the gas phase change in a way determined by
equilibria 1 and 2.
2. If we specify arbitrary values of T and p, each of the mole fractions has only one
possible value that will allow the two phases and four substances to be in equilibrium.
Now to introduce an additional complexity: Suppose we prepare the system by placing
a certain amount of O2 and twice this amount of carbon in an evacuated container, and wait
for the reactions to come to equilibrium. This method of preparation imposes an initial
condition on the system, and we must decide whether the number of degrees of freedom is
affected. Equating the total amount of carbon atoms to the total amount of oxygen atoms in
the equilibrated system gives the relation
nC C nCO C nCO2 D 2nO2 C nCO C 2nCO2
or
nC D 2nO2 C nCO2
(13.1.9)
Either equation is a relation among extensive variables of the two phases. From them, we
are unable to obtain any relation among intensive variables of the phases. Therefore, this
particular initial condition does not change the value of r, and F remains equal to 2.

425
Example 3: a solid salt and saturated aqueous solution

In this example, the equilibrium system consists of crystalline PbCl2 and an aqueous phase
containing the species H2 O, Pb2C (aq), and Cl (aq).
Applying the components approach to this system is straightforward. The solid phase
is prepared from PbCl2 and the aqueous phase could be prepared by dissolving solid PbCl2
in H2 O. Thus, there are two components and two phases:
F D2CC
P D2
(13.1.10)
For the species approach, we note that there are four species (PbCl2 , Pb2C , Cl , and
H2 O) and two independent relations among intensive variables:
1. equilibrium for the dissolution process,
PbCl2 C Pb2C C 2 Cl D 0;
2. electroneutrality of the aqueous phase, 2mPb2C D mCl .

We have s D 4, r D 2, and P D 2, giving the same result as the components approach:
F D2Cs
P D2
(13.1.11)
Example 4: liquid water and water-saturated air

For simplicity, let air be a gaseous mixture of N2 and O2 . The equilibrium system in this
example has two phases: liquid water saturated with the dissolved constituents of air, and
air saturated with gaseous H2 O.
If there is no special relation among the total amounts of N2 and O2 , there are three
components and the phase rule gives
F D2CC
P D3
(13.1.12)
Since there are three degrees of freedom, we could, for instance, specify arbitrary values2
of T , p, and yN2 ; then the values of other intensive variables such as the mole fractions
yH2 O and xN2 would have denite values.
Now suppose we impose an initial condition by preparing the system with water and
dry air of a xed composition. The mole ratio of N2 and O2 in the aqueous solution is not
necessarily the same as in the equilibrated gas phase; consequently, the air does not behave
like a single substance. The number of components is still three: H2 O, N2 , and O2 are
all required to prepare each phase individually, just as when there was no initial condition,
giving F D 3 as before.3
We can reach the same conclusion with the species approach. The initial condition can
be expressed by an equation such as
g
.nl 2 C nN2 /
N
g
.nl 2 C nO2 /
O
Da
(13.1.13)
2 Arbitrary,
that is, within the limits that would allow the two phases to coexist.
fact that the compositions of both phases depend on the relative amounts of the phases is illustrated in
Prob. 9.5.
3 The

13.2 P HASE D IAGRAMS : B INARY S YSTEMS
426
where a is a constant equal to the mole ratio of N2 and O2 in the dry air. This equation
cannot be changed to a relation between intensive variables such as xN2 and xO2 , so that r
is zero and there are still three degrees of freedom.
Finally, let us assume that we prepare the system with dry air of xed composition, as
before, but consider the solubilities of N2 and O2 in water to be negligible. Then nl 2 and
N
g
g
nl 2 are zero and Eq. 13.1.13 becomes nN2 =nO2 D a, or yN2 D ayO2 , which is a relation
O
between intensive variables. In this case, r is 1 and the phase rule becomes
F D2Cs
P D2
(13.1.14)
The reduction in the value of F from 3 to 2 is a consequence of our inability to detect any
dissolved N2 or O2 . According to the components approach, we may prepare the liquid
phase with H2 O and the gas phase with H2 O and air of xed composition that behaves as a
single substance; thus, there are only two components.
Example 5: equilibrium between two solid phases and a gas phase
Consider the following reaction equilibrium:
3 CuO.s/ C 2 NH3 .g/ 3 Cu.s/ C 3 H2 O.g/ C N2 .g/
According to the species approach, there are ve species, one relation (for reaction equilibrium), and three phases. The phase rule gives
F D2Cs
P D3
(13.1.15)
It is more difcult to apply the components approach to this example. As components,

we might choose CuO and Cu (from which we could prepare the solid phases) and also NH3
and H2 O. Then to obtain the N2 needed to prepare the gas phase, we could use CuO and
NH3 as reactants in the reaction 3 CuO C 2 NH3 ! 3 Cu C 3 H2 O C N2 and remove the
products Cu and H2 O. In the components approach, we are allowed to remove substances
from the system provided they are counted as components.
13.2 PHASE DIAGRAMS: BINARY SYSTEMS

As explained in Sec. 8.2, a phase diagram is a kind of two-dimensional map that shows
which phase or phases are stable under a given set of conditions. This section discusses
some common kinds of binary systems, and Sec. 13.3 will describe some interesting ternary
systems.
13.2.1 Generalities
A binary system has two components; C equals 2, and the number of degrees of freedom is
F D 4 P . There must be at least one phase, so the maximum possible value of F is 3.
Since F cannot be negative, the equilibrium system can have no more than four phases.
We can independently vary the temperature, pressure, and composition of the system as
a whole. Instead of using these variables as the coordinates of a three-dimensional phase

427
diagram, we usually draw a two-dimensional phase diagram that is either a temperature

composition diagram at a xed pressure or a pressurecomposition diagram at a xed temperature. The position of the system point on one of these diagrams then corresponds to a
denite temperature, pressure, and overall composition. The composition variable usually
varies along the horizontal axis and can be the mole fraction, mass fraction, or mass percent
of one of the components, as will presently be illustrated by various examples.
The way in which we interpret a two-dimensional phase diagram to obtain the compositions of individual phases depends on the number of phases present in the system.
If the system point falls within a one-phase area of the phase diagram, the composition variable is the composition of that single phase. There are three degrees
of freedom. On the phase diagram, the value of either T or p has been xed, so
there are two other independent intensive variables. For example, on a temperature
composition phase diagram, the pressure is xed and the temperature and composition can be changed independently within the boundaries of the one-phase area of the
diagram.
If the system point is in a two-phase area of the phase diagram, we draw a horizontal
tie line of constant temperature (on a temperaturecomposition phase diagram) or
constant pressure (on a pressurecomposition phase diagram). The lever rule applies.
The position of the point at each end of the tie line, at the boundary of the two-phase
area, gives the value of the composition variable of one of the phases and also the
physical state of this phase: either the state of an adjacent one-phase area, or the
state of a phase of xed composition when the boundary is a vertical line. Thus, a
boundary that separates a two-phase area for phases and from a one-phase area
for phase is a curve that describes the composition of phase as a function of T or
p when it is in equilibrium with phase . The curve is called a solidus, liquidus, or
vaporus depending on whether phase is a solid, liquid, or gas.
A binary system with three phases has only one degree of freedom and cannot be
represented by an area on a two-dimensional phase diagram. Instead, there is a horizontal boundary line between areas, with a special point along the line at the junction
of several areas. The compositions of the three phases are given by the positions of
this point and the points at the two ends of the line. The position of the system point
on this line does not uniquely specify the relative amounts in the three phases.
The examples that follow show some of the simpler kinds of phase diagrams known for
binary systems.
13.2.2 Solidliquid systems

Figure 13.1 on the next page is a temperaturecomposition phase diagram at a xed pressure. The composition variable zB is the mole fraction of component B in the system as a
whole. The phases shown are a binary liquid mixture of A and B, pure solid A, and pure
solid B.
The one-phase liquid area is bounded by two curves, which we can think of either as
freezing-point curves for the liquid or as solubility curves for the solids. These curves comprise the liquidus. As the mole fraction of either component in the liquid phase decreases
from unity, the freezing point decreases. The curves meet at point a, which is a eutectic

428
fA
b
A(s) +
liquid
liquid
c
d
fB
liquid
+ B(s)
a
h
A(s) + B(s)
Figure 13.1 Temperaturecomposition phase diagram for a binary system exhibiting

a eutectic point.
point. At this point, both solid A and solid B can coexist in equilibrium with a binary liquid
mixture. The composition at this point is the eutectic composition, and the temperature here
(denoted Te ) is the eutectic temperature. Te is the lowest temperature for the given pressure
at which the liquid phase is stable.4
Suppose we combine 0:60 mol A and 0:40 mol B (zB D 0:40) and adjust the temperature so as to put the system point at b. This point is in the one-phase liquid area, so the
equilibrium system at this temperature has a single liquid phase. If we now place the system
in thermal contact with a cold reservoir, heat is transferred out of the system and the system
point moves down along the isopleth (path of constant overall composition) bh. The cooling rate depends on the temperature gradient at the system boundary and the systems heat
capacity.
At point c on the isopleth, the system point reaches the boundary of the one-phase area
and is about to enter the two-phase area labeled A(s) + liquid. At this point in the cooling
process, the liquid is saturated with respect to solid A, and solid A is about to freeze out
from the liquid. There is an abrupt decrease (break) in the cooling rate at this point, because
the freezing process involves an extra enthalpy decrease.
At the still lower temperature at point d, the system point is within the two-phase solid
liquid area. The tie line through this point is line ef. The compositions of the two phases
s
l
are given by the values of zB at the ends of the tie line: xB D 0 for the solid and xB D 0:50
for the liquid. From the general lever rule (Eq. 8.2.8 on page 210), the ratio of the amounts
in these phases is
s
z
xB
nl
0:40 0
D B
D
D 4:0
(13.2.1)
l
s
n
0:50 0:40
xB zB
Since the total amount is ns C nl D 1:00 mol, the amounts of the two phases must be
ns D 0:20 mol and nl D 0:80 mol.
4 Eutectic
comes from the Greek for easy melting.

429
liquid
liquid
CHCl (s) + l
l + CCl (s)
s + l
l + s
s + s
CHCl (s) + CCl (s)
CHCl
mass fraction CCl
(a)
CCl
Ag
mass percent Cu
Cu
(b)
Figure 13.2 Temperaturecomposition phase diagrams with single eutectics.

(a) Two pure solids and a liquid mixture. a
(b) Two solid solutions and a liquid mixture.
a Ref.
[166], p. 98.
When the system point reaches the eutectic temperature at point g, cooling halts until
all of the liquid freezes. Solid B freezes out as well as solid A. During this eutectic halt,
there are at rst three phases: liquid with the eutectic composition, solid A, and solid B.
As heat continues to be withdrawn from the system, the amount of liquid decreases and the
amounts of the solids increase until nally only 0:60 mol of solid A and 0:40 mol of solid B
are present. The temperature then begins to decrease again and the system point enters the
two-phase area for solid A and solid B; tie lines in this area extend from zB D0 to zB D1.
Temperaturecomposition phase diagrams such as this are often mapped out experimentally by observing the cooling curve (temperature as a function of time) along isopleths of
various compositions. This procedure is thermal analysis. A break in the slope of a cooling
curve at a particular temperature indicates the system point has moved from a one-phase
liquid area to a two-phase area of liquid and solid. A temperature halt indicates the temperature is either the freezing point of the liquid to form a solid of the same composition, or
else a eutectic temperature.
Figure 13.2 shows two temperaturecomposition phase diagrams with single eutectic
points. The left-hand diagram is for the binary system of chloroform and carbon tetrachloride, two liquids that form nearly ideal mixtures. The solid phases are pure crystals, as in
Fig. 13.1. The right-hand diagram is for the silvercopper system and involves solid phases
that are solid solutions (substitutional alloys of variable composition). The area labeled s
is a solid solution that is mostly silver, and s is a solid solution that is mostly copper. Tie
lines in the two-phase areas do not end at a vertical line for a pure solid component as they
do in the system shown in the left-hand diagram. The three phases that can coexist at the
eutectic temperature of 1,052 K are the melt of the eutectic composition and the two solid
solutions.
Section 12.5.4 discussed the possibility of the appearance of a solid compound when a

430
liquid
l+B(s)
A(s)+1
AB(s)+1
l+AB(s)
A(s)+AB(s)
AB(s)+B(s)
Figure 13.3 Temperaturecomposition phase diagram for the binary system of naphthylamine (A) and phenol (B) at 1 bar (Ref. [127]).
binary liquid mixture is cooled. An example of this behavior is shown in Fig. 13.3, in which
the solid compound contains equal amounts of the two components -naphthylamine and
phenol. The possible solid phases are pure A, pure B, and the solid compound AB. Only one
or two of these solids can be present simultaneously in an equilibrium state. The vertical line
in the gure at zB D 0:5 represents the solid compound. The temperature at the upper end
of this line is the melting point of the solid compound, 29 C. The solid melts congruently
to give a liquid of the same composition. A melting process with this behavior is called a
dystectic reaction. The cooling curve for liquid of this composition would display a halt at
the melting point.
The phase diagram in Fig. 13.3 has two eutectic points. It resembles two simple phase
diagrams like Fig. 13.1 placed side by side. There is one important difference: the slope of
the freezing-point curve (liquidus curve) is nonzero at the composition of a pure component,
but is zero at the composition of a solid compound that is completely dissociated in the
liquid (as derived theoretically on page 389). Thus, the curve in Fig. 13.3 has a relative
maximum at the composition of the solid compound (zB D 0:5) and is rounded there,
instead of having a cusplike a Romanesque arch rather than a Gothic arch.
An example of a solid compound that does not melt congruently is shown in Fig. 13.4
on the next page. The solid hydrate NaCl 2H2 O is 61.9% NaCl by mass. It decomposes at
0 C to form an aqueous solution of composition 26.3% NaCl by mass and a solid phase of
anhydrous NaCl. These three phases can coexist at equilibrium at 0 C. A phase transition
like this, in which a solid compound changes into a liquid and a different solid, is called
incongruent or peritectic melting, and the point on the phase diagram at this temperature at
the composition of the liquid is a peritectic point.
Figure 13.4 shows there are two other temperatures at which three phases can be present
simultaneously: 21 C, where the phases are ice, the solution at its eutectic point, and the
solid hydrate; and 109 C, where the phases are gaseous H2 O, a solution of composition
28.3% NaCl by mass, and solid NaCl. Note that both segments of the right-hand boundary
of the one-phase solution area have positive slopes, meaning that the solubilities of the solid

H O(g)
431
NaCl(s)
H O(g)
sln
sln
sln
NaCl(s)
peritectic
point
H O(s)
sln
sln
H O(s)
H O
NaCl H O(s)
NaCl H O(s)
mass percent NaCl
NaCl H O(s)
NaCl(s)
NaCl
Figure 13.4 Temperaturecomposition phase diagram for the binary system of H2 O

and NaCl at 1 bar. (Data from Refs. [37] and [165].)
hydrate and the anhydrous salt both increase with increasing temperature.
13.2.3 Partially-miscible liquids

When two liquids that are partially miscible are combined in certain proportions, phase
separation occurs (Sec. 11.1.6). Two liquid phases in equilibrium with one another are
called conjugate phases. Obviously the two phases must have different compositions or
they would be identical; the difference is called a miscibility gap. A binary system with
two phases has two degrees of freedom, so that at a given temperature and pressure each
conjugate phase has a xed composition.
The typical dependence of a miscibility gap on temperature is shown in Fig. 13.5 on
the next page. The miscibility gap (the difference in compositions at the left and right
boundaries of the two-phase area) decreases as the temperature increases until at the upper
consolute temperature, also called the upper critical solution temperature, the gap vanishes.
The point at the maximum of the boundary curve of the two-phase area, where the temperature is the upper consolute temperature, is the consolute point or critical point. At this point,
the two liquid phases become identical, just as the liquid and gas phases become identical
at the critical point of a pure substance. Critical opalescence (page 206) is observed in the
vicinity of this point, caused by large local composition uctuations. At temperatures at
and above the critical point, the system is a single binary liquid mixture.
Suppose we combine 6:0 mol of component A (methyl acetate) and 4:0 mol of component B (carbon disulde) in a cylindrical vessel and adjust the temperature to 200 K. The

432
C H O
CS
Figure 13.5 Temperaturecomposition phase diagram for the binary system of

methyl acetate (A) and carbon disulde (B) at 1 bar. a All phases are liquids. The
open circle indicates the critical point.
a Data
from Ref. [54].
overall mole fraction of B is zB D 0:40. The system point is at point a in the two-phase
region. From the positions of points b and c at the ends of the tie line through point a,
we nd the two liquid layers have compositions xB D 0:20 and xB D 0:92. Since carbon
disulde is the more dense of the two pure liquids, the bottom layer is phase , the layer
that is richer in carbon disulde. According to the lever rule, the ratio of the amounts in the
two phases is given by
z
n
D B
n
xB
xB
zB
0:40
0:92
0:20
D 0:38
0:40
(13.2.2)
Combining this value with n C n D 10:0 mol gives us n D 7:2 mol and n D 2:8 mol.
If we gradually add more carbon disulde to the vessel while gently stirring and keeping
the temperature constant, the system point moves to the right along the tie line. Since the
ends of this tie line have xed positions, neither phase changes its composition, but the
amount of phase increases at the expense of phase . The liquidliquid interface moves
up in the vessel toward the top of the liquid column until, at overall composition zB D 0:92
(point c), there is only one liquid phase.
Now suppose the system point is back at point a and we raise the temperature while
keeping the overall composition constant at zB D 0:40. The system point moves up the
isopleth ad. The phase diagram shows that the ratio .zB xB /=.xB zB / decreases during
this change. As a result, the amount of phase increases, the amount of phase decreases,
and the liquidliquid interface moves down toward the bottom of the vessel until at 217 K
(point d) there again is only one liquid phase.
13.2.4 Liquidgas systems with ideal liquid mixtures

Toluene and benzene form liquid mixtures that are practically ideal and closely obey Raoults
law for partial pressure. For the binary system of these components, we can use the


liquidus
liquid
gas
vaporus
vaporus
gas
g+l
bar
433
g+l
liquid
liquidus
benzene
toluene
benzene
(a)
toluene
(b)
Figure 13.6 Phase diagrams for the binary system of toluene (A) and benzene (B).
The curves are calculated from Eqs. 13.2.6 and 13.2.7 and the saturation vapor pressures of the pure liquids.
(a) Pressurecomposition diagram at T D 340 K.
(b) Temperaturecomposition diagram at p D 1 bar.
vapor pressures of the pure liquids to generate the liquidus and vaporus curves of the
pressurecomposition and temperaturecomposition phase diagram. The results are shown
in Fig. 13.6. The composition variable zA is the overall mole fraction of component A
(toluene).
The equations needed to generate the curves can be derived as follows. Consider a
binary liquid mixture of components A and B and mole fraction composition xA that obeys
Raoults law for partial pressure (Eq. 9.4.2):
pA D xA pA
pB D .1
xA /pB
(13.2.3)
Strictly speaking, Raoults law applies to a liquidgas system maintained at a constant pressure by means of a third gaseous component, and pA and pB are the vapor pressures of
the pure liquid components at this pressure and the temperature of the system. However,
when a liquid phase is equilibrated with a gas phase, the partial pressure of a constituent of
the liquid is practically independent of the total pressure (Sec. 12.8.1), so that it is a good
approximation to apply the equations to a binary liquidgas system and treat pA and pB as
functions only of T .
When the binary system contains a liquid phase and a gas phase in equilibrium, the
pressure is the sum of pA and pB , which from Eq. 13.2.3 is given by
p D xA pA C .1
D pB C .pA
xA /pB
pB /xA
(13.2.4)
(C D2, ideal liquid mixture)
where xA is the mole fraction of A in the liquid phase. Equation 13.2.4 shows that in the
two-phase system, p has a value between pA and pB , and that if T is constant, p is a linear

434
function of xA . The mole fraction composition of the gas in the two-phase system is given
by
xA pA
p
yA D A D
(13.2.5)
p
pB C .pA pB /xA
A binary two-phase system has two degrees of freedom. At a given T and p, each phase
must have a xed composition. We can calculate the liquid composition by rearranging Eq.
13.2.4:
xA D
p pB
pA pB
(13.2.6)
The gas composition is then given by

x p
pA
D A A
p
p
p pB pA
D
pA pB p
yA D
(13.2.7)
If we know pA and pB as functions of T , we can use Eqs. 13.2.6 and 13.2.7 to calculate the
compositions for any combination of T and p at which the liquid and gas phases can coexist,
and thus construct a pressurecomposition or temperaturecomposition phase diagram.
In Fig. 13.6(a), the liquidus curve shows the relation between p and xA for equilibrated
liquid and gas phases at constant T , and the vaporus curve shows the relation between p
and yA under these conditions. We see that p is a linear function of xA but not of yA .
In a similar fashion, the liquidus curve in Fig. 13.6(b) shows the relation between T and
xA , and the vaporus curve shows the relation between T and yA , for equilibrated liquid and
gas phases at constant p. Neither curve is linear.
A liquidus curve is also called a bubble-point curve or a boiling-point curve. Other
names for a vaporus curve are dew-point curve and condensation curve. These curves are
actually cross-sections of liquidus and vaporus surfaces in a three-dimensional T pzA
phase diagram, as shown in Fig. 13.7 on the next page. In this gure, the liquidus surface is
in view at the front and the vaporus surface is hidden behind it.
13.2.5 Liquidgas systems with nonideal liquid mixtures

Most binary liquid mixtures do not behave ideally. The most common situation is positive
deviations from Raoults law.5 Some mixtures, however, have specic AB interactions,
such as solvation or molecular association, that prevent random mixing of the molecules
of A and B, and the result is then negative deviations from Raoults law. If the deviations
from Raoults law, either positive or negative, are large enough, the constant-temperature
liquidus curve exhibits a maximum or minimum and azeotropic behavior results.
Figure 13.8 on page 436 shows the azeotropic behavior of the binary methanol-benzene
system at constant temperature. In Fig. 13.8(a), the experimental partial pressures in a
5 In
the molecular model of Sec. 11.1.5, positive deviations correspond to a less negative value of kAB than the
average of kAA and kBB .

435
bar
ben
zen
e
tol
uen
e
Figure 13.7 Liquidus and vaporus surfaces for the binary system of toluene (A) and
benzene. Cross-sections through the two-phase region are drawn at constant temperatures of 340 K and 370 K and at constant pressures of 1 bar and 2 bar. Two of the
cross-sections intersect at a tie line at T D 370 K and p D 1 bar, and the other crosssections are hatched in the direction of the tie lines.
gas phase equilibrated with the nonideal liquid mixture are plotted as a function of the
liquid composition. The partial pressures of both components exhibit positive deviations
from Raoults law,6 and the total pressure (equal to the sum of the partial pressures) has a
maximum value greater than the vapor pressure of either pure component. The curve of p
versus xA becomes the liquidus curve of the pressurecomposition phase diagram shown in
Fig. 13.8(b). Points on the vaporus curve are calculated from p D pA =yA .
In practice, the data needed to generate the liquidus and vaporus curves of a nonideal
binary system are usually obtained by allowing liquid mixtures of various compositions to boil in an equilibrium still at a xed temperature or pressure. When the liquid
and gas phases have become equilibrated, samples of each are withdrawn for analysis. The partial pressures shown in Fig. 13.8(a) were calculated from the experimental
gas-phase compositions with the relations pA D yA p and pB D p pA .
If the constant-temperature liquidus curve has a maximum pressure at a liquid composition not corresponding to one of the pure components, which is the case for the methanol
6 This
behavior is consistent with the statement in Sec. 12.8.2 that if one constituent of a binary liquid mixture
exhibits positive deviations from Raoults law, with only one inection point in the curve of fugacity versus
mole fraction, the other constituent also has positive deviations from Raoults law.

liquid
kPa
kPa
436
(a)
g+l
gas
l+g
(b)
Figure 13.8 Binary system of methanol (A) and benzene at 45 C. a

(a) Partial pressures and total pressure in the gas phase equilibrated with liquid mixtures. The dashed lines indicate Raoults law behavior.
(b) Pressurecomposition phase diagram at 45 C. Open circle: azeotropic point at
zA D 0:59 and p D 60:5 kPa.
a Ref.
[155].
benzene system, then the liquid and gas phases are mixtures of identical compositions at
this pressure. This behavior was deduced on page 406 at the end of Sec. 12.8.3. On the
pressurecomposition phase diagram, the liquidus and vaporus curves both have maxima
at this pressure, and the two curves coincide at an azeotropic point. A binary system with
negative deviations from Raoults law can have an isothermal liquidus curve with a minimum pressure at a particular mixture composition, in which case the liquidus and vaporus
curves coincide at an azeotropic point at this minimum. The general phenomenon in which
equilibrated liquid and gas mixtures have identical compositions is called azeotropy, and
the liquid with this composition is an azeotropic mixture or azeotrope (Greek: boils unchanged). An azeotropic mixture vaporizes as if it were a pure substance, undergoing an
equilibrium phase transition to a gas of the same composition.
If the liquidus and vaporus curves exhibit a maximum on a pressurecomposition phase
diagram, then they exhibit a minimum on a temperaturecomposition phase diagram. This
relation is explained for the methanolbenzene system by the three-dimensional liquidus
and vaporus surfaces drawn in Fig. 13.9 on the next page. In this diagram, the vaporus
surface is hidden behind the liquidus surface. The hatched cross-section at the front of the
gure is the same as the pressurecomposition diagram of Fig. 13.8(b), and the hatched
cross-section at the top of the gure is a temperaturecomposition phase diagram in which
the system exhibits a minimum-boiling azeotrope.
A binary system containing an azeotropic mixture in equilibrium with its vapor has two
species, two phases, and one relation among intensive variables: xA D yA . The number
of degrees of freedom is then F D 2 C s r P D 2 C 2 1 2 D 1; the system is

kPa
437
ben
zen
e
me
tha
n
ol
Figure 13.9 Liquidus and vaporus surfaces for the binary system of methanol (A)
and benzene. a Cross-sections are hatched in the direction of the tie lines. The dashed
curve is the azeotrope vapor-pressure curve.
a Ref.
[155].
univariant. At a given temperature, the azeotrope can exist at only one pressure and have
only one composition. As T changes, so do p and zA along an azeotrope vapor-pressure
curve as illustrated by the dashed curve in Fig. 13.9.
Figure 13.10 on the next page summarizes the general appearance of some relatively
simple temperaturecomposition phase diagrams of binary systems. If the system does not
form an azeotrope (zeotropic behavior), the equilibrated gas phase is richer in one component than the liquid phase at all liquid compositions, and the liquid mixture can be separated
into its two components by fractional distillation. The gas in equilibrium with an azeotropic
mixture, however, is not enriched in either component. Fractional distillation of a system
with an azeotrope leads to separation into one pure component and the azeotropic mixture.
More complicated behavior is shown in the phase diagrams of Fig. 13.11. These are
binary systems with partially-miscible liquids in which the boiling point is reached before
an upper consolute temperature can be observed.
13.2.6 Solidgas systems

As an example of a two-component system with equilibrated solid and gas phases, consider the components CuSO4 and H2 O, denoted A and B respectively. In the pressure
composition phase diagram shown in Fig. 13.12 on page 439, the composition variable zB
is as usual the mole fraction of component B in the system as a whole.
The anhydrous salt and its hydrates (solid compounds) form the series of solids CuSO4 ,
CuSO4 H2 O, CuSO4 3H2 O, and CuSO4 5H2 O. In the phase diagram these formulas are

438
g
g
g
l
l
l
xA
A
xA
A
(a)
xA
A
(b)
(c)
Figure 13.10 Temperaturecomposition phase diagrams of binary systems exhibiting (a) no azeotropy, (b) a minimum-boiling azeotrope, and (c) a maximum-boiling
azeotrope. Only the one-phase areas are labeled; two-phase areas are hatched in the
direction of the tie lines.
l
l
l
l
xA
A
(a)
xA
A
(b)
xA
A
(c)
Figure 13.11 Temperaturecomposition phase diagrams of binary systems with

partially-miscible liquids exhibiting (a) the ability to be separated into pure components by fractional distillation, (b) a minimum-boiling azeotrope, and (c) boiling at a
lower temperature than the boiling point of either pure component. Only the one-phase
areas are labeled; two-phase areas are hatched in the direction of the tie lines.
abbreviated A, AB, AB3 , and AB5 . The following dissociation equilibria (dehydration
equilibria) are possible:
CuSO4 H2 O.s/ CuSO4 .s/ C H2 O.g/
1
2 CuSO4
1
2 CuSO4
3H2 O.s/
5H2 O.s/
1
2 CuSO4
1
2 CuSO4
H2 O.s/ C H2 O.g/
3H2 O.s/ C H2 O.g/
The equilibria are written above with coefcients that make the coefcient of H2 O(g) unity.
When one of these equilibria is established in the system, there are two components and
three phases; the phase rule then tells us the system is univariant and the pressure has only
one possible value at a given temperature. This pressure is called the dissociation pressure
of the higher hydrate.

AB (s) + AB (s)
bar
439
sln
sln + B(g)
AB (s)
+ sln
AB (s)
+ B(g)
AB(s) + AB (s)
A(s) + AB(s)
AB (s) + B(g)
AB(s) + B(g)
CuSO
A(s) + B(g)
H O
Figure 13.12 Pressurecomposition phase diagram for the binary system of CuSO4
(A) and H2 O (B) at 25 C. a
a Ref.
[106]; Ref. [167], p. 263.
The dissociation pressures of the three hydrates are indicated by horizontal lines in Fig.
13.12. For instance, the dissociation pressure of CuSO4 5H2 O is 1:05 10 2 bar. At the
pressure of each horizontal line, the equilibrium system can have one, two, or three phases,
with compositions given by the intersections of the line with vertical lines. A fourth threephase equilibrium is shown at p D 3:09 10 2 bar; this is the equilibrium between solid
CuSO4 5H2 O, the saturated aqueous solution of this hydrate, and water vapor.
Consider the thermodynamic equilibrium constant of one of the dissociation reactions.
At the low pressures shown in the phase diagram, the activities of the solids are practically
unity and the fugacity of the water vapor is practically the same as the pressure, so the equilibrium constant is almost exactly equal to pd =p , where pd is the dissociation pressure of
the higher hydrate in the reaction. Thus, a hydrate cannot exist in equilibrium with water vapor at a pressure below the dissociation pressure of the hydrate because dissociation would
be spontaneous under these conditions. Conversely, the salt formed by the dissociation of
a hydrate cannot exist in equilibrium with water vapor at a pressure above the dissociation
pressure because hydration would be spontaneous.
If the system contains dry air as an additional gaseous component and one of the dissociation equilibria is established, the partial pressure pH2 O of H2 O is equal (approximately) to the dissociation pressure pd of the higher hydrate. The prior statements
regarding dissociation and hydration now depend on the value of pH2 O . If a hydrate
is placed in air in which pH2 O is less than pd , dehydration is spontaneous; this phenomenon is called eforescence (Latin: blossoming). If pH2 O is greater than the vapor
pressure of the saturated solution of the highest hydrate that can form in the system,

440
hep
tan
e
eth
ane
bar
Figure 13.13 Pressuretemperaturecomposition behavior in the binary heptane

ethane system. a The open circles are critical points; the dashed curve is the critical
curve. The dashed line ab illustrates retrograde condensation at 450 K.
a Ref.
[86].
the anhydrous salt and any of its hydrates will spontaneously absorb water and form
the saturated solution; this is deliquescence (Latin: becoming uid).
If the two-component equilibrium system contains only two phases, it is bivariant corresponding to one of the areas in Fig. 13.12. Here both the temperature and the pressure can
be varied. In the case of areas labeled with two solid phases, the pressure has to be applied
to the solids by a uid (other than H2 O) that is not considered part of the system.
13.2.7 Systems at high pressure

Binary phase diagrams begin to look different when the pressure is greater than the critical
pressure of either of the pure components. Various types of behavior have been observed
in this region. One common type, that found in the binary system of heptane and ethane,
is shown in Fig. 13.13. This gure shows sections of a three-dimensional phase diagram
at ve temperatures. Each section is a pressurecomposition phase diagram at constant T .
The two-phase areas are hatched in the direction of the tie lines. At the left end of each
tie line (at low zA ) is a vaporus curve, and at the right end is a liquidus curve. The vapor
pressure curve of pure ethane (zA D0) ends at the critical point of ethane at 305:4 K; between
this point and the critical point of heptane at 540:5 K, there is a continuous critical curve,

441
bar
He
Xe
Figure 13.14 Pressuretemperaturecomposition behavior in the binary xenon

helium system. a The open circles are critical points; the dashed curve is the critical
curve.
a Ref.
[42].
which is the locus of critical points at which gas and liquid mixtures become identical in
composition and density.
Consider what happens when the system point is at point a in Fig. 13.13 and the pressure
is then increased by isothermal compression along line ab. The system point moves from
the area for a gas phase into the two-phase gasliquid area and then out into the gas-phase
area again. This curious phenomenon, condensation followed by vaporization, is called
retrograde condensation.
Under some conditions, an isobaric increase of T can result in vaporization followed by
condensation; this is retrograde vaporization.
A different type of high-pressure behavior, that found in the xenonhelium system, is
shown in Fig. 13.14. Here, the critical curve begins at the critical point of the less volatile
component (xenon) and continues to higher temperatures and pressures than the critical
temperature and pressure of either pure component. The two-phase region at pressures
above this critical curve is sometimes said to represent gasgas equilibrium, or gasgas
immiscibility, because we would not usually consider a liquid to exist beyond the critical
points of the pure components. Of course, the coexisting phases in this two-phase region
are not gases in the ordinary sense of being tenuous uids, but are instead high-pressure
uids of liquid-like densities. If we want to call both phases gases, then we have to say
that pure gaseous substances at high pressure do not necessarily mix spontaneously in all
proportions as they do at ordinary pressures.

13.3 P HASE D IAGRAMS : T ERNARY S YSTEMS
442
A
triangle height
(a)
C
(b)
Figure 13.15 Representing the composition of a ternary system by a point in an

equilateral triangle.
If the pressure of a system is increased isothermally, eventually solid phases will appear;
these are not shown in Figs. 13.13 and Fig. 13.14.
13.3 PHASE DIAGRAMS: TERNARY SYSTEMS

A ternary system is one with three components. We can independently vary the temperature,
the pressure, and two independent composition variables for the system as a whole. A twodimensional phase diagram for a ternary system is usually drawn for conditions of constant
T and p.
Although we could draw a two-dimensional phase diagram with Cartesian coordinates
to express the mole fractions of two of the components, there are advantages in using instead the triangular coordinates shown in Fig. 13.15. Each vertex of the equilateral triangle
represents one of the pure components A, B, or C. A point on the side of the triangle opposite a vertex represents a binary system of the other two components, and a point within the
triangle represents a ternary system with all three components.
To determine the mole fraction zA of component A in the system as a whole represented
by a point within the triangle, we measure the distance to the point from the side of the
triangle that is opposite the vertex for pure A, then express this distance as a fraction of the
height of the triangle. We follow the same procedure to determine zB and zC . The concept
is shown in Fig. 13.15(a).
As an aid for the conversion between the position of a point and the overall composition,
we can draw equally-spaced lines within the triangle parallel to the sides as shown in Fig.
13.15(b). One of these lines, being at a constant distance from one side of the triangle,
represents a constant mole fraction of one component. In the gure, the lines divide the
distance from each side to the opposite vertex into ten equal parts; thus, adjacent parallel
lines represent a difference of 0:1 in the mole fraction of a component, starting with 0 at
the side of the triangle and ending with 1 at the vertex. Using the lines, we see that the
lled circle in the gure represents the overall composition zA D 0:20, zB D 0:30, and

443

D
B
E
C
Figure 13.16 Proof that the sum of the lengths a, b, and c is equal to the height h of
the large equilateral triangle ABC. ADE and FDP are two smaller equilateral triangles.
The height of triangle ADE is equal to h a. The height of triangle FDP is equal to the
height of triangle ADE minus length b, and is also equal to length c: h a b D c.
Therefore, a C b C c D h.
zC D 0:50.
The sum of zA , zB , and zC must be 1. The method of representing composition with a
point in an equilateral triangle works because the sum of the lines drawn from the point to
the three sides, perpendicular to the sides, equals the height of the triangle. The proof is
shown in Fig. 13.16.
Two useful properties of this way of representing a ternary composition are as follows:
1. Points on a line parallel to a side of the triangle represent systems in which one of the
mole fractions remains constant.
2. Points on a line passing through a vertex represent systems in which the ratio of two
of the mole fractions remains constant.
13.3.1 Three liquids

Figure 13.17 on the next page is the ternary phase diagram of a system of ethanol, benzene,
and water at a temperature and pressure at which the phases are liquids. When the system
point is in the area labeled P D1, there is a single liquid phase whose composition is described by the position of the point. The one-phase area extends to the side of the triangle
representing binary mixtures of ethanol and benzene, and to the side representing binary
mixtures of ethanol and water. In other words, ethanol and benzene mix in all proportions,
and so also do ethanol and water.
When the overall composition is such that the system point falls in the area labeled
P D2, two liquid phases are present. The compositions of these phases are given by the
positions of the ends of a tie line through the system point. Four representative tie lines
are included in the diagram, and these must be determined experimentally. The relative
amounts of the two phases can be determined from the lever rule.7 In the limit of zero mole
7 The
lever rule works, according to the general derivation in Sec. 8.2.4, because the ratio nA =n, which is equal
to zA , varies linearly with the position of the system point along a tie line on the triangular phase diagram.

444
ethanol
b
a
benzene
water
Figure 13.17 Ternary phase diagram for ethanol, benzene, and water at 30 C and
1 bar. a The dashed lines are tie lines; the open circle indicates the plait point.
a Ref.
[19].
fraction of ethanol, the tie line falls along the horizontal base of the triangle and displays a
miscibility gap for the binary system of benzene and water. (The conjugate phases are very
nearly pure benzene and pure water).
The plait point shown as an open circle in the gure is also called a critical solution
point. As the system point approaches the plait point from within the two-phase area, the
length of the tie line through the system point approaches zero, the miscibility gap disappears, and the compositions of the two conjugate liquid phases become identical.
Suppose we have the binary system of benzene and water represented by point a. Two
liquid phases are present: one is wet benzene and the other is water containing a very small
mole fraction of benzene. If we gradually stir ethanol into this system, the system point
moves along the dotted line from point a toward the vertex for pure ethanol, but can never
quite reach the vertex. At point b, there are still two phases, and we can consider the ethanol
to have distributed itself between two partially-miscible solvents, benzene and water (Sec.
12.6.3). From the position of point b relative to the ends of the tie line passing through
point b, we see that the mole fraction of ethanol is greater in the water-rich phase. As we
continue to add ethanol, the amount of the water-rich phase increases and the amount of the
benzene-rich phase decreases, until at point c the benzene-rich phase completely disappears.
The added ethanol has increased the mutual solubilities of benzene and water and resulted
in a single liquid phase.
13.3.2 Two solids and a solvent

The phase diagram in Fig. 13.18 on the next page is for a ternary system of water and two
salts with an ion in common. There is a one-phase area for solution, labeled sln; a pair
of two-phase areas in which the phases are a single solid salt and the saturated solution;
and a triangular three-phase area. The upper vertex of the three-phase area, the eutonic
point, represents the composition of solution saturated with respect to both salts. Some
representative tie lines are drawn in the two-phase areas.

445
H2 O
sln
sln + NaCl(s) + KCl(s)

NaCl
KCl
Figure 13.18 Ternary phase diagram for NaCl, KCl, and water at 25 C and 1 bar. a
The dashed lines are tie lines in the two-phase areas.
a Data
from Ref. [166], p. 314.
A system of three components and three phases has two degrees of freedom; at xed
values of T and p, each phase must have a xed composition. The xed compositions
of the phases that are present when the system point falls in the three-phase area are the
compositions at the three vertices of the inner triangle: solid NaCl, solid KCl, and solution
of the eutonic composition xNaCl D 0:20 and xKCl D 0:11.
From the position of the curved boundary that separates the one-phase solution area
from the two-phase area for solution and solid KCl, we can see that adding NaCl to the
saturated solution of KCl decreases the mole fraction of KCl in the saturated solution. Although it is not obvious in the phase diagram, adding KCl to a saturated solution of NaCl
decreases the mole fraction of NaCl. These decreases in solubility when a common ion is
added are examples of the common ion effect mentioned in Sec. 12.5.5.

P ROBLEMS
446
PROBLEMS
in Appendix I.
13.1 Consider a single-phase system that is a gaseous mixture of N2 , H2 , and NH3 . For each
of the following cases, nd the number of degrees of freedom and give an example of the
independent intensive variables that could be used to specify the equilibrium state, apart from
the total amount of gas.
(a) There is no reaction.
(b) The reaction N2 .g/ C 3 H2 .g/ ! 2 NH3 .g/ is at equilibrium.
(c) The reaction is at equilibrium and the system is prepared from NH3 only.
13.2 How many components has a mixture of water and deuterium oxide in which the equilibrium
H2 O C D2 O 2 HDO exists?
13.3 Consider a system containing only NH4 Cl(s), NH3 (g), and HCl(g). Assume that the equilibrium NH4 Cl.s/ NH3 .g/ C HCl.g/ exists.
(a) Suppose you prepare the system by placing solid NH4 Cl in an evacuated ask and heating
to 400 K. Use the phase rule to decide whether you can vary the pressure while both
phases remain in equilibrium at 400 K.
(b) According to the phase rule, if the system is not prepared as described in part (a) could
you vary the pressure while both phases remain in equilibrium at 400 K? Explain.
(c) Rationalize your conclusions for these two cases on the basis of the thermodynamic equilibrium constant. Assume that the gas phase is an ideal gas mixture and use the approximate expression K D pNH3 pHCl =.p /2 .
13.4 Consider the lime-kiln process CaCO3 .s/ ! CaO.s/ C CO2 .g/. Find the number of intensive variables that can be varied independently in the equilibrium system under the following
conditions:
(a) The system is prepared by placing calcium carbonate, calcium oxide, and carbon dioxide
in a container.
(b) The system is prepared from calcium carbonate only.
(c) The temperature is xed at 1000 K.
13.5 What are the values of C and F in systems consisting of solid AgCl in equilibrium with an
aqueous phase containing H2 O, AgC (aq), Cl (aq), NaC (aq), and NO3 (aq) prepared in the
following ways? Give examples of intensive variables that could be varied independently.
(a) The system is prepared by equilibrating excess solid AgCl with an aqueous solution of
NaNO3 .
(b) The system is prepared by mixing aqueous solutions of AgNO3 and NaCl in arbitrary
proportions; some solid AgCl forms by precipitation.
13.6 How many degrees of freedom has a system consisting of solid NaCl in equilibrium with
an aqueous phase containing H2 O, NaC (aq), Cl (aq), HC (aq), and OH (aq)? Would it be
possible to independently vary T , p, and mOH ? If so, explain how you could do this.
13.7 Consult the phase diagram shown in Fig. 13.4 on page 431. Suppose the system contains 36:0 g
(2:00 mol) H2 O and 58:4 g (1:00 mol) NaCl at 25 C and 1 bar.

P ROBLEMS
447
(a) Describe the phases present in the equilibrium system and their masses.
(b) Describe the changes that occur at constant pressure if the system is placed in thermal
contact with a heat reservoir at 30 C.
(c) Describe the changes that occur if the temperature is raised from 25 C to 120 C at constant pressure.
(d) Describe the system after 200 g H2 O is added at 25 C.
Table 13.1 Aqueous solubilities of
sodium sulfate decahydrate and anhydrous sodium sulfate a
Na2 SO4 10H2 O
t= C
10
15
20
25
30
a Ref.
Na2 SO4
xB
t = C
0:011
0:016
0:024
0:034
0:048
40
50
xB
0:058
0:056
[55], p. 179180.
13.8 Use the following information to draw a temperaturecomposition phase diagram for the binary
system of H2 O (A) and Na2 SO4 (B) at p D 1 bar, conning t to the range 20 to 50 C and
zB to the range 00:2. The solid decahydrate, Na2 SO4 10H2 O, is stable below 32:4 C. The
anhydrous salt, Na2 SO4 , is stable above this temperature. There is a peritectic point for these
two solids and the solution at xB D 0:059 and t D 32:4 C. There is a eutectic point for
ice, Na2 SO4 10H2 O, and the solution at xB D 0:006 and t D 1:3 C. Table 13.1 gives the
temperature dependence of the solubilities of the ionic solids.
13.9 Iron(III) chloride forms various solid hydrates, all of which melt congruently. Table 13.2 on
the next page lists the temperatures t of aqueous solutions of various compositions that are
saturated with respect to a solid phase.
(a) Use these data to construct a t zB phase diagram for the binary system of FeCl3 (A) and
H2 O (B). Identify the formula and melting point of each hydrate. Hint: derive a formula
for the mole ratio nB =nA as a function of xA in a binary mixture.
(b) For the following conditions, determine the phase or phases present at equilibrium and
the composition of each.
1. t D 70:0 C and zA D 0:100
2. t D 50:0 C and zA D 0:275
13.10 Figure 13.19 on the next page is a temperaturecomposition phase diagram for the binary
system of water (A) and phenol (B) at 1 bar. These liquids are partially miscible below 67 C.
Phenol is more dense than water, so the layer with the higher mole fraction of phenol is the
bottom layer. Suppose you place 4:0 mol of H2 O and 1:0 mol of phenol in a beaker at 30 C
and gently stir to allow the layers to equilibrate.
(a) What are the compositions of the equilibrated top and bottom layers?
(b) Find the amount of each component in the bottom layer.

P ROBLEMS
448
Table 13.2 Data for Problem 13.9. Temperatures of

saturated solutions of aqueous iron(III) chloride at p D
1 bar (A = FeCl3 , B = H2 O) a
xA
t= C
xA
t= C
xA
t = C
0.000
0.020
0.032
0.037
0.045
0.052
0.053
0.056
0.076
0.083
0.093
0.106
0.0
10:0
20:5
27:5
40:0
55:0
41:0
27:0
0.0
10.0
20.0
30.0
0.119
0.143
0.157
0.173
0.183
0.195
0.213
0.222
0.232
0.238
0.259
0.277
35.0
37.0
36.0
33.0
30.0
27.4
32.0
32.5
30.0
35.0
50.0
55.0
0.286
0.289
0.293
0.301
0.318
0.333
0.343
0.358
0.369
0.369
0.373
56.0
55.0
60.0
69.0
72.5
73.5
72.5
70.0
66.0
80.0
100.0
a Data
from Ref. [55], page 193.
Figure 13.19 Temperaturecomposition phase diagram for the binary system of water (A) and phenol (B) at 1 bar. a Only liquid phases are present.
a Ref.
[55], p. 95.
(c) As you gradually stir more phenol into the beaker, maintaining the temperature at 30 C,
what changes occur in the volumes and compositions of the two layers? Assuming that
one layer eventually disappears, what additional amount of phenol is needed to cause this
to happen?
13.11 The standard boiling point of propane is 41:8 C and that of n-butane is 0:2 C. Table 13.3
on the next page lists vapor pressure data for the pure liquids. Assume that the liquid mixtures
obey Raoults law.
(a) Calculate the compositions, xA , of the liquid mixtures with boiling points of
20:0 C, and 30:0 C at a pressure of 1 bar.
10:0 C,
(b) Calculate the compositions, yA , of the equilibrium vapor at these three temperatures.

P ROBLEMS
449
Table 13.3 Saturation vapor pressures

of propane (A) and n-butane (B)
t= C
pA =bar
pB =bar
10:0
20:0
30:0
3:360
2:380
1:633
0:678
0:441
0:275
(c) Plot the temperaturecomposition phase diagram at p D 1 bar using these data, and label
the areas appropriately.
(d) Suppose a system containing 10:0 mol propane and 10:0 mol n-butane is brought to a
pressure of 1 bar and a temperature of 25 C. From your phase diagram, estimate the
compositions and amounts of both phases.
Table 13.4 Liquid and gas compositions in the two-phase
system of 2-propanol (A) and benzene at 45 C a
xA
yA
p=kPa
xA
yA
p=kPa
0
0:0472
0:0980
0:2047
0:2960
0:3862
0:4753
0
0:1467
0:2066
0:2663
0:2953
0:3211
0:3463
29:89
33:66
35:21
36:27
36:45
36:29
35:93
0:5504
0:6198
0:7096
0:8073
0:9120
0:9655
1:0000
0:3692
0:3951
0:4378
0:5107
0:6658
0:8252
1:0000
35:32
34:58
33:02
30:28
25:24
21:30
18:14
a Ref.
[24].
13.12 Use the data in Table 13.4 to draw a pressurecomposition phase diagram for the 2-propanol
benzene system at 45 C. Label the axes and each area.
Table 13.5 Liquid and gas compositions in the twophase system of acetone (A) and chloroform at 35:2 C
a
xA
yA
p=kPa
xA
yA
p=kPa
0
0:083
0:200
0:337
0:413
0:486
0:577
0
0:046
0:143
0:317
0:437
0:534
0:662
39:08
37:34
34:92
33:22
33:12
33:70
35:09
0:634
0:703
0:815
0:877
0:941
1:000
0:727
0:806
0:896
0:936
0:972
1:000
36:29
38:09
40:97
42:62
44:32
45:93
a Ref.
[165], p. 286.
13.13 Use the data in Table 13.5 to draw a pressurecomposition phase diagram for the acetone
chloroform system at 35:2 C. Label the axes and each area.
C HAPTER 14
G ALVANIC C ELLS
An electrochemical cell is a system in which passage of an electric current through an

electrical circuit is linked to an internal cell reaction. A galvanic cell, or voltaic cell, is
an electrochemical cell that, when isolated, has an electric potential difference between its
terminals; the cell is said to be a seat of electromotive force.
The cell reaction in a galvanic cell differs in a fundamental way from the same reaction
(i.e., one with the same reaction equation) taking place in a reaction vessel that is not part of
an electrical circuit. In the reaction vessel, the reactants and products are in the same phase
or in phases in contact with one another, and the reaction advances in the spontaneous
direction until reaction equilibrium is reached. This reaction is the direct reaction.
The galvanic cell, in contrast, is arranged with the reactants physically separated from
one another so that the cell reaction can advance only when an electric current passes
through the cell. If there is no current, the cell reaction is constrained from taking place.
When the electrical circuit is open and the cell is isolated from its surroundings, a state of
thermal, mechanical, and transfer equilibrium is rapidly reached. In this state of cell equilibrium or electrochemical equilibrium, however, reaction equilibrium is not necessarily
presentthat is, if the reactants and products were moved to a reaction vessel at the same
activities, there might be spontaneous advancement of the reaction.
As will be shown, measurements of the cell potential of a galvanic cell are capable of
yielding precise values of molar reaction quantities of the cell reaction and thermodynamic
equilibrium constants, and of mean ionic activity coefcients in electrolyte solutions.
14.1 CELL DIAGRAMS AND CELL REACTIONS

14.1.1 Elements of a galvanic cell
We will treat a galvanic cell as a system. The cell has two metal wires called terminals that
pass through the system boundary. Within the cell are phases that can conduct an electric
current and are collectively called electrical conductors. Each terminal is attached to an
electron conductor that is usually a metal, but might also be graphite or a semiconductor.
Each electron conductor is in contact with an ionic conductor, usually an electrolyte solution, through which ions but not electrons can move. Both of the electron conductors can
be in contact with the same ionic conductor; or they can be in contact with separate ionic
conductors, in which case the ionic conductors contact one another at a liquid junction. The
450
CHAPTER 14 GALVANIC CELLS

14.1 C ELL D IAGRAMS AND C ELL R EACTIONS
Cu
451
Cu
H (g)
AgCl
Pt
Ag
HCl(aq)
Figure 14.1 A galvanic cell without liquid junction.
general arrangement of the physical elements of a galvanic cell is therefore

terminal electron conductor ionic conductor(s) electron conductor terminal
Both terminals must be the same metal (usually copper) in order for it to be possible to
measure the electric potential difference between them.
The combination of an electron conductor and the ionic conductor in contact with it is
called an electrode,1 or half-cell. To describe a galvanic cell, it is conventional to distinguish the left and right electrodes. In this way, we establish a leftright association with the
reactants and products of the reactions at the electrodes.
14.1.2 Cell diagrams

Consider the galvanic cell depicted in Fig. 14.1. This cell has a hydrogen electrode at the
left and a silversilver chloride electrode at the right. The hydrogen electrode is a strip of
platinum in contact with hydrogen gas and with aqueous hydrochloric acid, which is the
ionic conductor. In the type of hydrogen electrode shown in the gure, hydrogen gas is
introduced through a side tube into a closed-end glass jacket that surrounds the platinum
strip and is immersed in the hydrochloric acid; the gas bubbles out through holes near the
bottom of the tube. The silversilver chloride electrode is a silver strip or wire that dips into
the hydrochloric acid and is coated with solid silver chloride.
The cell in Fig. 14.1 is compactly described by the following cell diagram:
Cu Pt H2 .g/ HC .aq/; Cl .aq/ AgCl.s/ Ag Cu
A cell diagram indicates which electrode is at the left and which is at the right, and shows
the reactants and products of the two electrode reactions. A single vertical bar represents a
phase boundary. Commas are used to separate different species in the same phase.
1 The
term electrode is sometimes used to refer to just the electron conductor.

14.1 C ELL D IAGRAMS AND C ELL R EACTIONS
452
The same cell can be described by a slightly different cell diagram that omits the copper
terminals seen in the gure and shows the electrolyte solute instead of the ion species:
Pt H2 .g/ HCl.aq/ AgCl.s/ Ag
The reason it is not necessary to include the terminals is that the property whose value
we seek, the zero-current cell potential, is the same regardless of the metal used for the
terminals.
14.1.3 Electrode reactions and the cell reaction

A cell diagram, with its designation of the left and right electrodes, allows us to write
reaction equations for the cell. These equations are written according to the convention that
electrons enter at the right terminal and leave at the left terminal.
At each electrode there is an electrode reaction, or half-reaction, one for reduction at
the right electrode and the other for oxidation at the left electrode. The reaction equations
for the electrode reactions include electrons as either a reactant (at the right terminal) or a
product (at the left terminal). The cell reaction describes the overall chemical change; its
reaction equation is the sum of the equations for the two electrode reactants with cancellation of the electrons.
For instance, we can write the electrode reactions of the cell of Fig. 14.1 as follows.
oxidation at left:
reduction at right:
H2 .g/ ! 2 HC .aq/ C 2 e
2 AgCl.s/ C 2 e
! 2 Ag.s/ C 2 Cl .aq/
As written here, the stoichiometric numbers of the electrons have the same absolute value
(2) in both reaction equations. This allows the electrons to cancel when we add the electrode
reactions to form the cell reaction:
H2 .g/ C 2 AgCl.s/ ! 2 HC .aq/ C 2 Cl .aq/ C 2 Ag.s/
The cell of Fig. 14.1 has a single electrolyte phase with essentially the same composition at both electrodes, and is an example of a cell without liquid junction or cell without
transference. As an example of a cell with transference, consider the cell diagram
Zn Zn2C .aq/ Cu2C .aq/ Cu
This is the zinccopper cell depicted in Fig. 14.2 on the next page, sometimes called a
Daniell cell. The two electrolyte phases are separated by a liquid junction represented in
the cell diagram by the dashed vertical bar. If the liquid junction potential can be assumed
to be negligible, the liquid junction is instead represented by a pair of dashed vertical bars:
14.1.4 Advancement and charge

The electron number or charge number, z, of the cell reaction is dened as the amount
of electrons entering at the right terminal per unit advancement of the cell reaction. z is a

14.2 E LECTRIC P OTENTIALS IN THE C ELL
453
e
Zn
Cu
Zn
Cu
ZnSO (aq)
CuSO (aq)
ZnSO (aq)
CuSO (aq)
(a)
(b)
Figure 14.2 Zinccopper galvanic cell with porous barrier (heavy dashed line) separating two electrolyte solutions. The dashed rectangle indicates the system boundary.
(a) Open circuit with isolated system in equilibrium state.
(b) Closed circuit.
positive dimensionless quantity equal to j e j, where e is the stoichiometric number of the

electrons in either of the electrode reactions whose sum is the cell reaction.
Because both electrode reactions are written with the same value of j e j, the advancements of these reactions and of the cell reaction are all described by the same advancement
variable . For an innitesimal change d , an amount of electrons equal to z d enters the
system at the right terminal, an equal amount of electrons leaves at the left terminal, and
there is no buildup of charge in any of the internal phases.
The Faraday constant F is a physical constant dened as the charge per amount of
protons, and is equal to the product of the elementary charge (the charge of a proton) and the
Avogadro constant: F D eNA . Its value to ve signicant gures is F D 96; 485 C mol 1 .
The charge per amount of electrons is F . Thus, the charge entering the right terminal
during advancement d is
Qsys D zF d
(14.1.1)
14.2 ELECTRIC POTENTIALS IN THE CELL

As explained at the beginning of Sec. 3.8, the electric potential at a point in space is
dened as the change in the electrical potential energy of an innitesimal test charge when
it is brought to this point from a position innitely far from other charges, divided by the
charge.
We are concerned with the electric potential within a phasethe inner electric potential,
or Galvani potential. We can measure the difference between the values of this electric potential in the two terminals of a galvanic cell, provided the terminals have the same chemical
composition. If the terminals were of different metals, at least one of them would have an
unknown metalmetal contact potential in its connection to the external measuring circuit.
Since we will be applying the concept of electric potential to macroscopic phases, the
value of the Galvani potential at a point in a phase should be interpreted as the average value
in a small volume element at this point that is large enough to contain many molecules.

454
14.2.1 Cell potential

The cell potential of a galvanic cell is the electric potential difference between terminals of
the same metal, and is dened by Eq. 3.8.6:
def
Ecell D
(14.2.1)
The subscripts R and L refer to the right and left terminals. The equilibrium cell potential,
Ecell, eq , is the cell potential measured under conditions of zero current when the cell is
assumed to be in an equilibrium state.2
Over a relatively long period of time, the state of an isolated galvanic cell is found to
change. Nevertheless, the assumption of an equilibrium state is valid if the changes
are very slow compared to the period during which we measure Ecell .
The long-term changes can be of two types. If there is a liquid junction between
electrolyte solutions of different composition, slow diffusion of ions through the junction is inevitable.
In a cell without a liquid junction, the reactants of the cell reaction can react directly without the passage of an electric current. For instance, in the cell of Fig. 14.1
the electrolyte solution is saturated with respect to gaseous H2 and solid AgCl, and
therefore contains very small concentrations of dissolved H2 molecules and AgC ions.
The direct reaction H2 C 2 AgC ! 2 HC C 2 Ag occurs irreversibly and continuously
in the solution, but is slow on account of the low concentrations.
It is entirely arbitrary whether we show a particular electrode at the left or the right
of the cell diagram, although often there is a preference to place the electrode attached to
the positive terminal at the right. If we exchange the positions of the two electrodes in the
diagram, then we must reverse the reaction equations for the electrode reactions and the cell
reaction.
For example, it is found that the zinccopper cell of Fig. 14.2, with typical electrolyte
molalities, has its positive terminal at the copper electrode. When we write the cell diagram
as
then Ecell and Ecell, eq are positive. If we connect the two terminals by an external resistor
as depicted in Fig. 14.2(b), electrons will ow from the left terminal through the external
resistor and wires to the right terminal, and the cell reaction
Zn C Cu2C .aq/ ! Zn2C .aq/ C Cu
will occur spontaneously in the forward direction.
If, however, we draw the cell diagram the other way around:
Cu Cu2C .aq/ Zn2C .aq/ Zn
then the positive terminal is at the left, Ecell and Ecell, eq are negative, and electrons will
ow through an external resistor from the right terminal to the left terminal. Since the cell
2 The
equilibrium cell potential used to be called the electromotive force, or emf. These names are deprecated
by the IUPAC Green Book (Ref. [36], p. 71) because a potential difference is not a force.

455
G
L
galvanic
cell
galvanic
cell
(a)
(b)
Figure 14.3 Potentiometer to measure the zero-current cell potential of a galvanic

cell.
(a) Galvanic cell with zero current.
(b) Galvanic cell included in potentiometer circuit; G is a galvanometer.
reaction should show reduction at the right electrode and oxidation at the left, we must now
write it as
Cu C Zn2C .aq/ ! Cu2C .aq/ C Zn
even though the arrow is not in the direction of the reaction that actually occurs spontaneously. In other words, the cell reaction is written according to the cell diagram, not
according to the direction of the spontaneous change.
14.2.2 Measuring the equilibrium cell potential

Figure 14.3 shows how we can use a potentiometer to determine the equilibrium cell potential. Consider Fig. 14.3(a). Outside the galvanic cell is an external circuit with a battery that
allows an electric current to pass through a slidewire resistor. The cells negative terminal is
connected to the negative terminal of the battery. Since the cell is not part of this circuit, no
current passes through the cell, and R
L is the zero-current cell potential Ecell, eq . The
left end of the slidewire is at the same electric potential as the left terminal of the cell.
In the setup shown in Fig. 14.3(a), the electric potential within the slidewire is a linear
function of the distance from the left end. At some position along the slidewire, the electric
potential is equal to R . We can determine this position by connecting the right terminal
of the cell to a slidewire contact as shown in Fig. 14.3(b). When we place the contact
at this particular position along the slidewire, there is no electric potential gradient in the
connecting wire, and the galvanometer indicates a condition of zero current in the wire. It
is a straightforward procedure to evaluate R
L from the zero-current position of the
contact; this value is still equal to Ecell, eq . When we keep the slidewire contact in this
position, no current passes through the cell; but if we displace the contact from this position
in either direction along the slidewire, current will pass in one direction or the other through
the cell.
In practice, it is more convenient to measure the zero-current cell potential with a highimpedance digital voltmeter (a voltmeter that draws negligible current) instead of with a

456
RT
RE
cell, eq
RT
I
RE
I
L
LT
LE
cell
L
LT
LE
(a)
(b)
Figure 14.4 Galvani potential prole across a galvanic cell (schematic). LT and RT
are the left and right terminals, LE and RE are the left and right electron conductors,
and I is an ionic conductor such as an electrolyte solution.
(a) Cell with zero current.
(b) The same cell with nite current.
potentiometer circuit.
14.2.3 Interfacial potential differences

What is the source of an open-circuit, zero-current cell potential? When no electric current
passes through the cell, the electric potential must be uniform within each bulk phase that
is an electrical conductor, because otherwise there would be a spontaneous movement of
charged particles (electrons or ions) through the phase. Electric potential differences in a
cell without current therefore exist only at phase boundaries. The equilibrium cell potential is the cumulative result of these potential differences at interfaces between different
conducting phases within the cell.
An interfacial potential difference appears as a vertical step in a prole of the Galvani
potential, as shown schematically in Fig. 14.4(a). The zero-current cell potential, Ecell, eq ,
is the algebraic sum of the interfacial potential differences within the cell.
When an external resistor is connected to the terminals to form a circuit, current passes
through the cell and the cell performs electrical work on the surroundings. Figure 14.4(b)
shows what happens to the potential prole in this case: the interfacial potential differences
are still present within the cell, and the internal resistance of the electrical conductors causes
Ecell to be reduced in magnitude compared to Ecell, eq .
We shall next look briey at the origin and consequences of potential differences at
interfaces between (1) two different metals, (2) a metal and an electrolyte solution, and
(3) two different electrolyte solutions. Keep in mind that these potential differences are
theoretical concepts whose values cannot be measured experimentally.
Metalmetal contacts
An electric potential difference at an interface between two metals is called a contact potential. When two different metals are placed in contact, the local densities of the free (mobile)
electrons change so as to form an electrical double layer with an excess positive charge on

457
one side of the interface and an excess negative charge of equal magnitude on the other side.
The electrical double layer creates the contact potential.
To understand why a stable equilibrium state of two metals in contact includes a contact potential, we can consider the chemical potential of the free electrons. The concept of
chemical potential (i.e., partial molar Gibbs energy) applies to the free electrons in a metal
just as it does to other species. The dependence of the chemical potential of free elece
trons in metal phase on the electric potential of the phase is given by the relation of
Eq. 10.1.6 on page 288, with the charge number zi set equal to 1:
e.
/D
e .0/
(14.2.2)
Here .0/ is the electron chemical potential in a phase with the same intensive properties
e
as phase but at zero electric potential. .0/ depends only on the temperature and the
e
composition of phase . (The dependence on pressure is so small for a solid that we will
ignore it.)
Consider two or more electron conductors that are so arranged that electrons can freely
transfer among them. There is the usual condition for transfer equilibrium in these phases:
the chemical potential (in this case e ) is the same in each phase. Thus, electron transfer
equilibrium between phases and requires and e to be equal. We equate and e ,

e
e
substitute from Eq. 14.2.2 to obtain .0/ F D e .0/ F , and rearrange to

e
e .0/
.0/
e
(14.2.3)
(phases in electron
transfer equilibrium)
The quantities on the right side of Eq. 14.2.3 are functions only of the temperature and the
compositions of phases and . If the phases have the same temperature and composition
and are in electron transfer equilibrium, and are equal.
For an equilibrium state of metals and in contact, Eq. 14.2.3 shows that the contact
depends only on the temperature and the compositions of the two metals.3
potential
Equation 14.2.3 explains why a galvanic cell must have at least one electrical conductor
that is not an electron conductor. If electrons were free to pass from one terminal
through the system to the other terminal of the same temperature and composition,
then in a zero-current equilibrium state e would be the same in both terminals. In
that case there would be no potential difference between the terminals, and the system
would not be a galvanic cell.
Metalelectrolyte interfaces
An electrode reaction of a galvanic cell takes place at the interface between a metal electron
conductor and an electrolyte solution. In an equilibrium state of the cell, the electrode
P
reaction is at equilibrium. The condition for this equilibrium is i i i D 0, where the
sum is over the reactants and products of the electrode reaction, including the electrons.
The chemical potentials of the ions and electrons in the electrode reaction are functions of
3 The
temperature dependence of a contact potential between two different metals is the basis of the operation
of a thermocouple or thermopile to measure temperature (Sec. 2.3.5).

14.3 M OLAR R EACTION Q UANTITIES OF THE C ELL R EACTION
458
the electric potentials of their phases. Consequently, in order for the sum to be zero, the
metal and solution must in general have different electric potentials.
For example, consider the zinccopper cell of Fig. 14.2. The electrode reaction of the
copper electrode at the right is
Cu2C .aq/ C 2 e .Cu/ ! Cu
where the metal phase of the electrons is indicated in parentheses. In order for this electrode reaction to be at equilibrium, the interfacial potential difference between the copper
conductor and the solution containing Cu2C ions must be such that the following condition
is satised:
.Cu/
.Cu2C / 2 e .Cu/ D 0
(14.2.4)
The interfacial potential difference can arise from a combination of charge separation
across the interface, orientation of polar molecules on the solution side of the interface, and
specic adsorption of ions. The thickness of the zones in which properties differ from those
in the bulk phases is probably no greater than 10 11 m on the metal side and 10 7 m on the
solution side.
Liquid junctions
Some galvanic cells contain two electrolyte solutions with different compositions. These
solutions must be separated by a porous barrier or some other kind of junction in order to
prevent rapid mixing. At this liquid junction in the zero-current cell, there is in general
a liquid junction potential caused by diffusion of ions between the two bulk electrolyte
phases.
To understand this phenomenon, imagine the situation that would exist at the junction if
both solution phases had the same electric potential. An ion species with different chemical
potentials in the two solutions would spontaneously diffuse across the junction in the direction of lower chemical potential. Different ions would diffuse at different rates, resulting in
a net charge transfer across the junction and an electric potential difference. It is this electric potential difference in the equilibrium state of the cell that prevents further net charge
transfer under zero-current conditions.
The liquid junction may consist of a bridging solution in a salt bridge. A commonly
used kind of salt bridge is a glass tube lled with gel made from agar and concentrated
aqueous KCl or KNO3 ; this type of liquid junction is believed to reduce the liquid junction
potential to several millivolts or less.
14.3 MOLAR REACTION QUANTITIES OF THE CELL REACTION

This book will denote the molar reaction Gibbs energy of a cell reaction by r Gcell . This
notation distinguishes it from the molar reaction Gibbs energy r G of the direct reaction,
which may have a different value because in the cell the chemical potential of an ionic
species is affected by the electric potential of its phase. r Gcell is dened by
def
r Gcell D
(14.3.1)

459
where the sum is over the reactants and products of the cell reaction. r Gcell is also equal
to the partial derivative .@Gcell =@ /T;p , where is the advancement of the cell reaction.
14.3.1 Relation between r Gcell and Ecell, eq

When a galvanic cell is in a zero-current equilibrium state, both electrode reactions are at
reaction equilibrium. In the electrode reaction at the left electrode, electrons are a product
with stoichiometric number equal to z. At the right electrode, electrons are a reactant with
stoichiometric number equal to z. We can write the conditions for electrode reaction
equilibria as follows:
X
At the left electrode
(14.3.2)
i i C z e .LE/ D 0
i
At the right electrode
e .RE/
D0
(14.3.3)
In these equations, the sum over i is for the chemical species (excluding electrons) of the
electrode reaction at the left electrode, and the sum over j is for the chemical species of the
electrode reaction at the right electrode. e .LE/ is the chemical potential of electrons in the
electron conductor of the left electrode, and e .RE/ is the chemical potential of electrons
in the electron conductor of the right electrode.
Adding Eqs. 14.3.2 and 14.3.3, we obtain
X
X
(14.3.4)
i i C
j j C z e .LE/
e .RE/ D 0
i
The rst two terms on the left side of Eq. 14.3.4 are sums over all the reactants and products
of the cell reaction. From Eq. 14.3.1, we recognize the sum of these terms as the molar
reaction Gibbs energy of the cell reaction:
X
X
(14.3.5)
i i C
j j D r Gcell
i
Substituting from Eq. 14.3.5 into Eq. 14.3.4 and solving for r Gcell , we obtain
r Gcell D
e .LE/
e .RE/
(14.3.6)
In a zero-current equilibrium state, there is electron transfer equilibrium between the left
electron conductor and the left terminal, and between the right electron conductor and the
right terminal: e .LE/ D e .LT/ and e .RE/ D e .RT/, where e .LT/ and e .RT/ are
the chemical potentials of electrons in the left terminal and right terminal, respectively.
Thus we can rewrite Eq. 14.3.6 as
r Gcell D
e .LT/
e .RT/
(14.3.7)
Making substitutions from Eq. 14.2.2 for e .LT/ and e .RT/, and recognizing that
is the same in both terminals because they have the same composition, we obtain
r Gcell D
D
zF .
zFEcell, eq
e .0/
L/
(14.3.8)

460
We can see from Eq. 14.3.1 that the value of r Gcell has nothing to do with the composition of the terminals. The relations of Eq. 14.3.8 were derived for a cell with both terminals
made of the same metal. We can make the following deductions for such a cell:
1. Neither the potential difference R
L nor the equilibrium cell potential Ecell, eq
depend on the kind of metal used for the terminals.
2. If we interpose a metal conductor of any composition between the electron conductor
and the terminal of one of the electrodes, e will have the same value in all three
conductors and there will be no effect on the value of Ecell, eq .
Equation 14.3.8 can be derived by a different route. According to Eq. 5.8.6 on page 146,
reversible electrical work at constant T and p is equal to the Gibbs energy change:
wel, rev D dGcell . Making the substitution wel, rev D Ecell, eq Qsys (from Eq. 3.8.8),
with Qsys set equal to zF d (Eq. 14.1.1), followed by division by d , gives
zFEcell, eq D .@Gcell =@ /T;p , or r Gcell D zFEcell, eq .
Strictly speaking, this derivation applies only to a cell without a liquid junction.
In a cell with a liquid junction, the electric current is carried across the junction by
different ions depending on the direction of the current, and the cell is therefore not
reversible.
14.3.2 Relation between r Gcell and r G

Suppose we have a galvanic cell in a particular zero-current equilibrium state. Each phase
of the cell has the same temperature and pressure and a well-dened chemical composition.
The activity of each reactant and product of the cell reaction therefore has a denite value
in this state.
Now imagine a reaction vessel that has the same temperature and pressure as the galvanic cell, and contains the same reactants and products at the same activities as in the cell.
This reaction vessel, unlike the cell, is not part of an electrical circuit. In it, the reactants
and products are in direct contact with one another, so there is no constraint preventing a
spontaneous direct reaction. For example, the reaction vessel corresponding to the zinc
copper cell of Fig. 14.2 would have zinc and copper strips in contact with a solution of both
ZnSO4 and CuSO4 . Another example is the slow direct reaction in a cell without liquid
junction described on page 454.
Let the reaction equation of the direct reaction be written with the same stoichiometric
numbers i as in the reaction equation for the cell reaction. The direct reaction in the
reaction vessel is described by this equation or its reverse, depending on which direction is
spontaneous for the given activities.
The question now arises whether the molar reaction Gibbs energy r Gcell of the cell
reaction is equal to the molar reaction Gibbs energy r G of the direct reaction. Both r Gcell
P
and r G are dened by the sum i i i . Both reactions have the same values of i , but
the values of i for charged species are in general different in the two systems because the
electric potentials are different.
Consider rst a cell without a liquid junction. This kind of cell has a single electrolyte
solution, and all of the reactant and product ions of the cell reaction are in this solution
phase. The same solution phase is present in the reaction vessel during the direct reaction.
P
When all ions are in the same phase, the value of i i i is independent of the electric

461
potentials of any of the phases (see the comment following Eq. 11.8.4 on page 351), so that
the molar reaction Gibbs energies are the same for the cell reaction and the direct reaction:
r Gcell D r G
(14.3.9)
(no liquid junction)
Next, consider a cell with two electrolyte solutions separated by a liquid junction. For
the molar reaction Gibbs energy of the cell reaction, we write
X
X
r Gcell D
(14.3.10)
i i. i/ C
j j. j/
i
The sums here include all of the reactants and products appearing in the cell reaction, those
with index i being at the left electrode and those with index j at the right electrode. Let the
solution at the left electrode be phase and the solution at the right electrode be phase .
Then making the substitution i . / D i .0/ C zi F (Eq. 10.1.6) gives us
X
X
X
X
r Gcell D
C
(14.3.11)
i i .0/ C
j j .0/ C
i zi F
j zj F
i
The sum of the rst two terms on the right side of Eq. 14.3.11 is the molar reaction
Gibbs energy of a reaction in which the reactants and products are in phases of zero electric
potential. According to the comment following Eq. 11.8.4, the molar reaction Gibbs energy
would be the same if the ions were in a single phase of any electric potential. Consequently
P
P
the sum i i i .0/C j j j .0/ is equal to r G for the direct reaction.
The conservation of charge during advancement of the electrode reactions at the left
P
P
electrode and the right electrode is expressed by i i zi z D 0 and j j zj C z D 0,
respectively. Equation 14.3.11 becomes
r Gcell D r G
zFEj
(14.3.12)
(cell with liquid junction)
is the liquid junction potential.

where Ej D
Finally, in Eqs. 14.3.9 and 14.3.12 we replace r Gcell by zFEcell, eq (Eq. 14.3.8) and
solve for Ecell, eq :
Ecell, eq D
Ecell, eq D
r G
zF
r G
C Ej
zF
(14.3.13)
(cell without liquid junction)
(14.3.14)
(cell with liquid junction)
Ecell, eq can be measured with great precision. If a reaction can be carried out in a galvanic
cell without liquid junction, Eq. 14.3.13 provides a way to evaluate r G under given conditions. If the reaction can only be carried out in a cell with a liquid junction, Eq. 14.3.14
can be used for this purpose provided that the liquid junction potential Ej can be assumed
to be negligible or can be estimated from theory.

462
Note that the cell has reaction equilibrium only if r G is zero. The cell has thermal,
mechanical, and transfer equilibrium when the electric current is zero and the cell potential
is the zero-current cell potential Ecell, eq . Equations 14.3.13 and 14.3.14 show that in order
for the cell to also have reaction equilibrium, Ecell, eq must equal the liquid junction potential
if there is a liquid junction, or be zero otherwise. These are the conditions of an exhausted,
dead cell that can no longer do electrical work.
14.3.3 Standard molar reaction quantities

Consider a hypothetical galvanic cell in which each reactant and product of the cell reaction is in its standard state at unit activity, and in which a liquid junction if present has
a negligible liquid junction potential. The equilibrium cell potential of this cell is called
the standard cell potential of the cell reaction, Ecell, eq . An experimental procedure for
evaluating Ecell, eq will be described in Sec. 14.5.

In this hypothetical cell, r Gcell is equal to the standard molar reaction Gibbs energy
r G . From Eq. 14.3.13, or Eq. 14.3.14 with Ej assumed equal to zero, we have
r G D
zFEcell, eq
(14.3.15)
r G is the molar reaction Gibbs energy when each reactant and product is at unit activity
and, if it is an ion, is in a phase of zero electric potential. Since r G is equal to RT ln K
(Eq. 11.8.10), we can write
zF
E
(14.3.16)
ln K D
RT cell, eq
Equation 14.3.16 allows us to evaluate the thermodynamic equilibrium constant K of
the cell reaction by a noncalorimetric method. Consider for example the cell
Ag AgC .aq/ Cl .aq/ AgCl.s/ Ag
in which the pair of dashed vertical bars indicates a liquid junction of negligible liquid
junction potential. The electrode reactions are
Ag(s) ! AgC (aq) C e
AgCl(s) C e
! Ag(s) C Cl (aq)
and the cell reaction is

AgCl.s/ ! AgC .aq/ C Cl .aq/
The equilibrium constant of this reaction is the solubility product Ks of silver chloride (Sec.
12.5.5). At 298:15 K, the standard cell potential is found to be Ecell, eq D 0:5770 V. We

can use this value in Eq. 14.3.16 to evaluate Ks at 298:15 K (see Prob. 14.5).
Equation 14.3.16 also allows us to evaluate the standard molar reaction enthalpy by
substitution in Eq. 12.1.13 on page 369:
r H D RT 2
D zF
d ln K
dT
dEcell, eq
T
dT
Ecell, eq
(14.3.17)

14.4 T HE N ERNST E QUATION
463
Finally, by combining Eqs. 14.3.15 and 14.3.17 with r G D r H

an expression for the standard molar reaction entropy:
r S D zF
dEcell, eq
dT
Tr S , we obtain
(14.3.18)
Because G, H , and S are state functions, the thermodynamic equilibrium constant
and the molar reaction quantities evaluated from Ecell, eq and dEcell, eq = dT are the same
quantities as those for the reaction when it takes place in a reaction vessel instead of in a
galvanic cell. However, the heats at constant T and p are not the same (page 319). During
a reversible cell reaction, dS must equal q=T , and q= d is therefore equal to Tr S
during a cell reaction taking place reversibly under standard state conditions at constant T
and p.
14.4 THE NERNST EQUATION
The standard cell potential Ecell, eq of a cell reaction is the equilibrium cell potential of
the hypothetical galvanic cell in which each reactant and product of the cell reaction is
in its standard state and there is no liquid junction potential. The value of Ecell, eq for a
given cell reaction with given choices of standard states is a function only of temperature.
The measured equilibrium cell potential Ecell, eq of an actual cell, however, depends on the
activities of the reactants and products as well as on temperature and the liquid junction
potential, if present.
To derive a relation between Ecell, eq and activities for a cell without liquid junction, or
with a liquid junction of negligible liquid junction potential, we substitute expressions for
r G and for r G from Eqs. 14.3.13 and Eq. 14.3.15 into r G D r G C RT ln Qrxn
(Eq. 11.8.8 on page 352) and solve for Ecell, eq :
Ecell, eq D Ecell, eq
RT
ln Qrxn
zF
(14.4.1)
(no liquid junction, or Ej D0)
Equation 14.4.1 is the Nernst equation for the cell reaction. Here Qrxn is the reaction
Q
quotient for the cell reaction dened by Eq. 11.8.6: Qrxn D i ai i .
The rest of this section will assume that the cell reaction takes place in a cell without
liquid junction, or in one in which Ej is negligible.
If each reactant and product of the cell reaction is in its standard state, then each activity
is unity and ln Qrxn is zero. We can see from the Nernst equation that the equilibrium cell
potential Ecell, eq in this case has its standard value Ecell, eq , as expected. A decrease in
product activities or an increase in reactant activities decreases the value of ln Qrxn and
increases Ecell, eq , as we would expect since Ecell, eq should be greater when the forward cell
reaction has a greater tendency for spontaneity.
If the cell reaction comes to reaction equilibrium, as it will if we short-circuit the cell terminals with an external wire, the value of Qrxn becomes equal to the thermodynamic equi
librium constant K, and the Nernst equation becomes Ecell, eq D Ecell, eq .RT =zF / ln K.

14.4 T HE N ERNST E QUATION
464
The term .RT =zF / ln K is equal to Ecell, eq (Eq. 14.3.16), so Ecell, eq becomes zerothe
cell is dead and is incapable of performing electrical work on the surroundings.
At T D298:15 K (25:00 C), the value of RT =F is 0:02569 V, and we can write the
Nernst equation in the compact form
0:02569 V
ln Qrxn
z
(14.4.2)
(T D298:15 K)
As an illustration of an application of the Nernst equation, consider the reaction equation

H2 .g/ C 2 AgCl.s/ ! 2 HC .aq/ C 2 Cl .aq/ C 2 Ag.s/
This reaction takes place in a cell without liquid junction (Fig. 14.1), and the electrolyte
solution can be aqueous HCl. The expression for the reaction quotient is
Qrxn D
2
2
aC a2 aAg
2
aH2 aAgCl
(14.4.3)
We may usually with negligible error approximate the pressure factors of the solids and
solutes by unity. The activities of the solids are then 1, the solute activities are aC D
m =m , and the hydrogen activity is aH2 D fH2 =p . The ion

C mC =m and a D
molalities mC and m are equal to the HCl molality mB . The expression for Qrxn becomes
Qrxn D
.mB =m /4
D
fH2 =p
2 2
C
4
4
.mB =m /
fH2 =p
(14.4.4)
and the Nernst equation for this cell is

Ecell, eq D
Ecell, eq
D Ecell, eq
RT
ln
2F
2RT
ln
F
4
4
.mB =m /
fH2 =p
fH
2RT
m
RT
ln B C
ln 2
F
m
2F
p
(14.4.5)
By measuring Ecell, eq for a cell with known values of mB and fH2 , and with a derived value
of Ecell, eq , we can use this equation to nd the mean ionic activity coefcient of the HCl
solute. This is how the experimental curve for aqueous HCl in Fig. 10.3 on page 297 was
obtained.
We can always multiply each of the stoichiometric coefcients of a reaction equation
by the same positive constant without changing the meaning of the reaction. How does
this affect the Nernst equation for the reaction equation above? Suppose we decide to
multiply the stoichiometric coefcients by one-half:
1
H .g/
2 2
C AgCl.s/ ! HC .aq/ C Cl .aq/ C Ag(s)
With this changed reaction equation, the value of z is changed from 2 to 1 and the
Nernst equation becomes
2
.mB =m /2
RT
ln
F
.fH2 =p /1=2
(14.4.6)

14.5 E VALUATION OF THE S TANDARD C ELL P OTENTIAL
465
which yields the same value of Ecell, eq for given cell conditions as Eq. 14.4.5. This
value must of course be unchanged, because physically the cell is the same no matter
how we write its cell reaction, and measurable physical quantities such as Ecell, eq are
unaffected. However, molar reaction quantities such as r G and r G do depend on
how we write the cell reaction, because they are changes per extent of reaction.
14.5 EVALUATION OF THE STANDARD CELL POTENTIAL
As we have seen, the value of the standard cell potential Ecell, eq of a cell reaction has useful
thermodynamic applications. The value of Ecell, eq for a given cell reaction depends only on
temperature. To evaluate it, we can extrapolate an appropriate function to innite dilution
where ionic activity coefcients are unity.
To see how this procedure works, consider again the cell reaction H2 .g/C2 AgCl.s/ !
2 HC .aq/ C 2 Cl .aq/ C 2 Ag.s/. The cell potential depends on the molality mB of the HCl
solute according to Eq. 14.4.5. We can rearrange the equation to
Ecell, eq D Ecell, eq C
2RT
ln
F
2RT
m
ln B
F
m
fH
RT
ln 2
2F
p
(14.5.1)
For given conditions of the cell, we can measure all quantities on the right side of Eq. 14.5.1
except the mean ionic activity coefcient of the electrolyte. We cannot know the exact
value of ln for any given molality until we have evaluated Ecell, eq . We do know that
as mB approaches zero, approaches unity and ln must approach zero. The Debye
H ckel formula of Eq. 10.4.7 on page 297 is a theoretical expression for ln that more
u
closely approximates the actual value the lower is the ionic strength. Accordingly, we dene
the quantity
p
fH
A mB
2RT
2RT
m
RT
0
Ecell D Ecell, eq C
C
ln B
ln 2
(14.5.2)
p
F
1 C Ba mB
F
m
2F
p
The expression in parentheses is the DebyeH ckel formula for ln with Im replaced by
u
mB . The constants A and B have known values at any temperature (Sec. 10.4), and a is an
ion-size parameter for which we can choose a reasonable value. At a given temperature, we
0
can evaluate Ecell experimentally as a function of mB .
The expression on the right side of Eq. 14.5.1 differs from that of Eq. 14.5.2 by contributions to .2RT =F / ln not accounted for by the DebyeH ckel formula. Since these
u
contributions approach zero in the limit of innite dilution, the extrapolation of measured
0
values of Ecell to mB D0 yields the value of Ecell, eq .

Figure 14.5 on the next page shows this extrapolation using data from the literature. The
extrapolated value indicated by the lled circle is Ecell, eq D 0:2222 V, and the uncertainty
is on the order of only 0:1 mV.
14.6 STANDARD ELECTRODE POTENTIALS

Section 14.5 explained how, by measuring the equilibrium cell potential of a galvanic cell
at different electrolyte molalities, we can evaluate the standard cell potential Ecell, eq of the
cell reaction. It is not necessary to carry out this involved experimental procedure for each

14.6 S TANDARD E LECTRODE P OTENTIALS
466
cell
0.2225
0.2220
0.2215
0.01
0.02
B mol kg
0.03
0
Figure 14.5 Ecell (dened by Eq. 14.5.2) as a function of HCl molality for the cell of
Fig. 14.1 at 298:15 K. a The dashed line is a least-squares t to a linear relation.
a Data from Ref. [73] with f
H2
set equal to pH2 and the parameter a set equal to 4:3
10
10 m.
individual cell reaction of interest. Instead, we can calculate Ecell, eq from standard electrode
potentials.
By convention, standard electrode potentials use a standard hydrogen electrode as a
reference electrode. A standard hydrogen electrode is a hydrogen electrode, such as the
electrode shown at the left in Fig. 14.1, in which the species H2 (g) and HC (aq) are in their
standard states. Since these are hypothetical gas and solute standard states, the standard
hydrogen electrode is a hypothetical electrodenot one we can actually construct in the
laboratory.
A standard electrode potential E is dened as the standard cell potential of a cell
with a hydrogen electrode at the left and the electrode of interest at the right. For example,
the cell in Fig. 14.1 with cell diagram

has a hydrogen electrode at the left and a silversilver chloride electrode at the right. The
standard electrode potential of the silversilver chloride electrode, therefore, is equal to the
standard cell potential of this cell.
Since a cell with hydrogen electrodes at both the left and right has a standard cell potential of zero, the standard electrode potential of the hydrogen electrode is zero at all temperatures. The standard electrode potential of any other electrode is nonzero and is a function
only of temperature.
Consider the following three cells constructed from various combinations of three different electrodes: a hydrogen electrode, and two electrodes denoted L and R.
Cell 1 has electrode L at the left and electrode R at the right.
Cell 2 has the hydrogen electrode at the left and electrode L at the right; its standard
cell potential is the standard electrode potential EL of electrode L.


14.6 S TANDARD E LECTRODE P OTENTIALS
467
Cell 3 has the hydrogen electrode at the left and electrode R at the right; its standard
cell potential is the standard electrode potential ER of electrode R.
We wish to calculate the standard cell potential Ecell, eq of cell 1 from the standard electrode
potentials EL and ER .
If we write the cell reactions of cells 1 and 2 using the same value of the electron number
z for both, we nd that their sum is the cell reaction for cell 3 with the same value of z. Call
these reactions 1, 2, and 3, respectively:
(reaction 1) C (reaction 2) D (reaction 3)
(14.6.1)
The relation of Eq. 14.6.1 shows that an innitesimal advancement d of reaction 1

combined with an equal advancement of reaction 2 causes the same changes in amounts as
the advancement d of reaction 3. Because r G for each reaction is the rate at which G
changes with at constant T when the reactants and products are in their standard states,
the following relation applies when the reactions take place at the same temperature:
r G (reaction 1) C r G (reaction 2) D r G (reaction 3)
(14.6.2)
Making the substitution r G D zFEcell, eq (Eq. 14.3.15), with the same value of z for
each reaction, gives us Ecell, eq C EL D ER , or
Ecell, eq D ER
EL
(14.6.3)
where Ecell, eq , ER , and EL all refer to cell 1.

Equation 14.6.3 is a general relation applicable to any galvanic cell. It should be apparent that we can use the relation to calculate the standard electrode potential of an electrode
from the standard electrode potential of a different electrode and the standard cell potential
of a cell that contains both electrodes. Neither electrode has to be a hydrogen electrode,
which is difcult to work with experimentally.
Using Eq. 14.6.3 to calculate standard cell potentials from standard electrode potentials
saves a lot of experimental work. For example, measurement of Ecell, eq for ten different
cells, only one of which needs to include a hydrogen electrode, provides values of E for
ten electrodes other than E D0 for the hydrogen electrode. From these ten values of E ,
values of Ecell, eq can be calculated for 35 other cells without hydrogen electrodes.

P ROBLEMS
468
PROBLEMS
in Appendix I.
14.1 The state of a galvanic cell without liquid junction, when its temperature and pressure are
uniform, can be fully described by values of the variables T , p, and . Find an expression
for dG during a reversible advancement of the cell reaction, and use it to derive the relation
r Gcell D zFEcell, eq (Eq. 14.3.8). (Hint: Eq. 3.8.8.)
14.2 Before 1982 the standard pressure was usually taken as 1 atm. For the cell shown in Fig.
14.1, what correction is needed, for a value of Ecell, eq obtained at 25 C and using the older
convention, to change the value to one corresponding to a standard pressure of 1 bar? Equation
14.3.15 can be used for this calculation.
14.3 Careful measurements4 of the equilibrium cell potential of the cell
yielded, at 298:15 K and using a standard pressure of 1 bar, the values Ecell, eq D 0:22217 V
and dEcell, eq = dT D 6:462 10 4 V K 1 . (The requested calculated values are close to, but
not exactly the same as, the values listed in Appendix H, which are based on the same data
combined with data of other workers.)
(a) Evaluate r G , r S , and r H at 298:15 K for the reaction

1
H .g/
2 2
C AgCl.s/ ! HC .aq/ C Cl .aq/ C Ag.s/
(b) Problem 12.18 showed how the standard molar enthalpy of formation of the aqueous
chloride ion may be evaluated based on the convention f H .HC , aq/ D 0. If this value is
combined with the value of r H obtained in part (a) of the present problem, the standard
molar enthalpy of formation of crystalline silver chloride can be evaluated. Carry out
this calculation for T D 298:15 K using the value f H .Cl , aq/ D 167:08 kJ mol 1
(Appendix H).
(c) By a similar procedure, evaluate the standard molar entropy, the standard molar entropy of
formation, and the standard molar Gibbs energy of formation of crystalline silver chloride
at 298:15 K. You need the following standard molar entropies evaluated from spectroscopic and calorimetric data:
Sm .H2 , g/ D 130:68 J K 1 mol 1
Sm .Cl , aq/ D 56:60 J K 1 mol 1
Sm .Cl2 , g/ D 223:08 J K 1 mol
Sm .Ag, s/ D 42:55 J K 1 mol 1
14.4 The standard cell potential of the cell

Ag AgCl.s/ HCl.aq/ Cl2 .g/ Pt
has been determined over a range of temperature.5 At T D298:15 K, the standard cell poten
tial was found to be Ecell, eq D 1:13579 V, and its temperature derivative was found to be
dEcell, eq = dT D 5:9863 10 4 V K 1 .
(a) Write the cell reaction for this cell.
4 Ref.
[4].
5 Ref.
[53].

P ROBLEMS
469
(b) Use the data to evaluate the standard molar enthalpy of formation and the standard molar Gibbs energy of formation of crystalline silver chloride at 298:15 K. (Note that this
calculation provides values of quantities also calculated in Prob. 14.3 using independent
data.)
14.5 Use data in Sec. 14.3.3 to evaluate the solubility product of silver chloride at 298:15 K.
14.6 The equilibrium cell potential of the galvanic cell
Pt H2 (g, f D1 bar) HCl(aq, 0:500 mol kg
) Cl2 (g, f D1 bar) Pt
is found to be Ecell, eq D 1:410 V at 298:15 K. The standard cell potential is Ecell, eq D 1:360 V.
(a) Write the cell reaction and calculate its thermodynamic equilibrium constant at 298:15 K.
(b) Use the cell measurement to calculate the mean ionic activity coefcient of aqueous HCl
at 298:15 K and a molality of 0:500 mol kg 1 .
14.7 Consider the following galvanic cell, which combines a hydrogen electrode and a calomel
electrode:
Pt H2 .g/ HCl.aq/ Hg2 Cl2 .s/ Hg.l/ Pt
(a) Write the cell reaction.
(b) At 298:15 K, the standard cell potential of this cell is Ecell, eq D 0:2680 V. Using the value
of f G for the aqueous chloride ion in Appendix H, calculate the standard molar Gibbs
energy of formation of crystalline mercury(I) chloride (calomel) at 298:15 K.
(c) Calculate the solubility product of mercury(I) chloride at 298:15 K. The dissolution equilibrium is Hg2 Cl2 .s/ Hg2 2C .aq/ C 2 Cl .aq/. Take values for the standard molar
Gibbs energies of formation of the aqueous ions from Appendix H.
Table 14.1 Equilibrium cell potential as a function of HBr molality mB .
mB / (mol kg
0.0004042
0.0008444
0.0008680
0.0013554
0.001464
0.001850
0.002396
0.003719
Ecell, eq / V
0.47381
0.43636
0.43499
0.41243
0.40864
0.39667
0.38383
0.36173
14.8 Table 14.1 lists equilibrium cell potentials obtained with the following cell at 298:15 K:6
Pt H2 (g, 1:01 bar) HBr(aq, mB ) AgBr(s) Ag
Use these data to evaluate the standard electrode potential of the silver-silver bromide electrode
at this temperature to the nearest millivolt. (Since the electrolyte solutions are quite dilute, you
p
may ignore the term Ba mB in Eq. 14.5.2.)
6 Ref.
[91].

P ROBLEMS
470
14.9 The cell diagram of a mercury cell can be written

Zn.s/ ZnO.s/ NaOH.aq/ HgO.s/ Hg.l/
(a) Write the electrode reactions and cell reaction with electron number z D 2.
(b) Use data in Appendix H to calculate the standard molar reaction quantities r H , r G ,
and r S for the cell reaction at 298:15 K.
(c) Calculate the standard cell potential of the mercury cell at 298:15 K to the nearest 0:01 V.
(d) Evaluate the ratio of heat to advancement, q= d , at a constant temperature of 298:15 K
and a constant pressure of 1 bar, for the cell reaction taking place in two different ways:
reversibly in the cell, and spontaneously in a reaction vessel that is not part of an electrical
circuit.
(e) Evaluate dEcell, eq = dT , the temperature coefcient of the standard cell potential.
A PPENDIX A
D EFINITIONS OF THE SI BASE U NITS

The ofcial denitions of the base units given in the IUPAC Green Book1 are as follows.
The metre2 is the length of path traveled by light in vacuum during a time interval of
1/299 792 458 of a second.
The kilogram is the unit of mass; it is equal to the mass of the international prototype of
the kilogram.3
The second is the duration of 9 192 631 770 periods of the radiation corresponding to the
transition between the two hyperne levels of the ground state of the caesium 133 atom.
This denition refers to a caesium atom at rest at a temperature of 0 K.
The kelvin, unit of thermodynamic temperature, is the fraction 1/273:16 of the thermodynamic temperature of the triple point of water. This denition refers to water having
the isotopic composition dened exactly by the following amount-of-substance ratios:
0.000 155 76 mole of 2 H per mole of 1 H, 0.000 379 9 mole of 17 O per mole of 16 O, and
0.002 005 2 mole of 18 O per mole of 16 O.
The mole is the amount of substance of a system which contains as many elementary entities
as there are atoms in 0:012 kilogram of carbon 12; its symbol is mol. When the mole
is used, the elementary entities must be specied and may be atoms, molecules, ions,
electrons, other particles, or specied groups of such particles. In this denition, it
is understood that unbound atoms of carbon 12, at rest and in their ground state, are
referred to.
The ampere is that constant current which, if maintained in two straight parallel conductors of innite length, of negligible circular cross-section, and placed 1 metre apart in
vacuum, would produce between these conductors a force equal to 2 10 7 newton per
metre of length.
The candela is the luminous intensity, in a given direction, of a source that emits monochromatic radiation of frequency 540 1012 hertz and that has a radiant intensity in that
direction of 1/683 watt per steradian.
1 Ref.
[36], Sec. 3.3.

book uses the alternative spelling meter.
3 The international prototype is a platinum-iridium cylinder stored in a vault of the International Bureau of
Weights and Measures in S` vres near Paris, France.
e
2 This
471
A PPENDIX B
P HYSICAL C ONSTANTS
The following table lists values of fundamental physical constants used in thermodynamic
calculations.1 Except for those marked exact, they are the 2006 CODATA (Committee on
Data for Science and Technology) recommended values. The number in parentheses at the
end of a value is the standard deviation uncertainty in the right-most digits of the value.
Physical quantity
Symbol
Value in SI units
Avogadro constant
NA
6.022 141 79(30)
elementary charge
1.602 176 487(40)
Faraday constant
9.648 533 99(24)
gas constant a
8.314 472(15) J K
magnetic constant b
electric constant c
0
0
10
NA
8.854 187 817: : :
speed of light in vacuum
c0
gn
9.806 65 m s
19
10
104 C mol
1
mol
(exact)
10
108 m s
2.997 924 58
standard acceleration of free fall d
1023 mol
12
C2 J
(exact)
(exact)
(exact)
a or
molar gas constant

permeability of vacuum
c or permittivity of vacuum
d or standard acceleration of gravity
b or
1 Online
at http://physics.nist.gov/cuu/Constants/index.html; IUPAC Green Book (Ref.

[36]), p. 111.
472
A PPENDIX C
S YMBOLS FOR P HYSICAL Q UANTITIES
This appendix lists the symbols for most of the variable physical quantities used in this
book. The symbols are those recommended in the IUPAC Green Book (Ref. [119]) except
for quantities followed by an asterisk ( ).
Symbol
Physical quantity
SI unit
Roman letters
A
Helmholtz energy
As
surface area
a
activity
B
second virial coefcient
C
number of components
Cp
heat capacity at constant pressure
CV
heat capacity at constant volume
c
concentration
E
energy
electrode potential
E
electric eld strength
Ecell
cell potential
Ej
liquid junction potential
Esys
system energy in a lab frame
F
force
number of degrees of freedom
f
fugacity
g
acceleration of free fall
G
Gibbs energy
h
height, elevation
H
enthalpy
H
magnetic eld strength
I
electric current
Im
ionic strength, molality basis
Ic
ionic strength, concentration basis
473
J
m2
(dimensionless)
m3 mol 1
(dimensionless)
JK 1
JK 1
mol m 3
J
V
Vm 1
V
V
J
N
(dimensionless)
Pa
ms 2
J
m
J
Am 1
A
mol kg 1
mol m 3
474
A PPENDIX C S YMBOLS FOR P HYSICAL Q UANTITIES
(continued from previous page)
Symbol
K
Ka
Kp
Ks
kH;i
kc;i
km;i
l
L
M
M
Mr
m
mi
N
n
P
p
P
Q
Qsys
Qrxn
q
Rel
S
s
T
t
U
V
v
w
Physical quantity
SI unit
thermodynamic equilibrium constant

acid dissociation constant
equilibrium constant, pressure basis
solubility product
Henrys law constant of species i,
mole fraction basis
concentration basis
molality basis
length, distance
relative partial molar enthalpy
molar mass
magnetization
relative molecular mass (molecular weight)
mass
molality of species i
number of entities (molecules, atoms, ions,
formula units, etc.)
amount of substance
number of phases
pressure
partial pressure
dielectric polarization
electric charge
charge entering system at right conductor
reaction quotient
heat
electric resistance
entropy
solubility
number of species
time
Celsius temperature
internal energy
volume
specic volume
velocity, speed
work
mass fraction (weight fraction)
(dimensionless)
(dimensionless)
P
Pa
(dimensionless)
Pa
Pa m3 mol
Pa kg mol 1
m
J mol 1
kg mol 1
Am 1
(dimensionless)
kg
mol kg 1
(dimensionless)
mol
(dimensionless)
Pa
Pa
Cm 2
C
C
(dimensionless)
J
JK 1
mol m 3
(dimensionless)
K
s
C
J
m3
m3 kg 1
ms 1
J
(dimensionless)
475
Symbol
wel
w0
x
y
Z
z
Physical quantity
SI unit
electrical work
nonexpansion work
mole fraction in a phase
Cartesian space coordinate
mole fraction in gas phase
compression factor (compressibility factor)
mole fraction in multiphase system
charge number of an ion
electron number of cell reaction
J
J
(dimensionless)
m
(dimensionless)
m
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
m
Greek letters
alpha
degree of reaction, dissociation, etc.

cubic expansion coefcient
gamma
surface tension
activity coefcient of species i,
i
pure liquid or solid standard state
m;i
molality basis
c;i
concentration basis
x;i
mole fraction basis
mean ionic activity coefcient
pressure factor (activity of a reference state)

epsilon
efciency of a heat engine
energy equivalent of a calorimeter
theta
#
angle of rotation
kappa
reciprocal radius of ionic atmosphere

T
isothermal compressibility
mu
chemical potential
JouleThomson coefcient
JT
nu
number of ions per formula unit
(dimensionless)
K 1
Nm
1,
Jm
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
JK 1
(dimensionless)
m
Pa
1
1
J mol
K Pa
1
1
(dimensionless)
476
Symbol
SI unit
stoichiometric number
number of cations per formula unit
number of anions per formula unit
(dimensionless)
(dimensionless)
(dimensionless)
advancement (extent of reaction)
mol
osmotic pressure
Pa
density
kg m
torque
Physical quantity
fugacity coefcient
electric potential
electric potential difference
osmotic coefcient, molality basis
relative apparent molar enthalpy of solute
(dimensionless)
V
V
(dimensionless)
J mol 1
angular velocity
xi
pi
rho
tau
phi
L
omega
!
A PPENDIX D
M ISCELLANEOUS A BBREVIATIONS AND

S YMBOLS
D.1 PHYSICAL STATES

These abbreviations for physical states (states of aggregation) may be appended in parentheses to chemical formulas or used as superscripts to symbols for physical quantities. All
but mixt are listed in the IUPAC Green Book (Ref. [36], p. 54).
g
l
f
s
cd
cr
mixt
sln
aq
aq; 1
gas or vapor
liquid
uid (gas or liquid)
solid
condensed phase (liquid or solid)
crystalline
mixture
solution
aqueous solution
aqueous solution at innite dilution
477
A PPENDIX D M ISCELLANEOUS A BBREVIATIONS AND S YMBOLS
478
D.2 SUBSCRIPTS FOR CHEMICAL PROCESSES

These abbreviations are used as subscripts to the symbol. They are listed in the IUPAC
Green Book (Ref. [36], p. 5960).
The combination p , where p is any one of the abbreviations below, can be interpreted
def
as an operator: p D @=@ p where p is the advancement of the given process at constant

temperature and pressure. For example, c H D .@H=@ c /T;p is the molar differential
enthalpy of combustion.
vap
sub
fus
trs
mix
sol
dil
ads
dpl
imm
r
at
c
f
vaporization, evaporation (l ! g)
sublimation (s ! g)
melting, fusion (s ! l)
transition between two phases
mixing of uids
solution of a solute in solvent
dilution of a solution
adsorption
displacement
immersion
reaction in general
atomization
combustion reaction
formation reaction
A PPENDIX D M ISCELLANEOUS A BBREVIATIONS AND S YMBOLS
479
D.3 SUPERSCRIPTS
These abbreviations and symbols are used as superscripts to symbols for physical quantities.
All but 0 , int, and ref are listed as recommended superscripts in the IUPAC Green Book (Ref.
[36], p. 60).
1
id
int
E
ref
standard
pure substance
Legendre transform of a thermodynamic potential
innite dilution
ideal
integral
excess quantity
reference state
A PPENDIX E
C ALCULUS R EVIEW
E.1 DERIVATIVES
Let f be a function of the variable x, and let f be the change in f when x changes by
x. Then the derivative df = dx is the ratio f =x in the limit as x approaches zero.
The derivative df = dx can also be described as the rate at which f changes with x, and as
the slope of a curve of f plotted as a function of x.
The following is a short list of formulas likely to be needed. In these formulas, u and v
are arbitrary functions of x, and a is a constant.
d.ua /
du
D aua 1
dx
dx
d.uv/
dv
du
Du
Cv
dx
dx dx
1
du
d.u=v/
D
v
dx
v2
dx
d ln.ax/
1
D
dx
x
d.e ax /
D ae ax
dx
df .u/
df .u/ du
D
dx
du
dx
dv
dx
E.2 PARTIAL DERIVATIVES

If f is a function of the independent variables x, y, and z, the partial derivative .@f =@x/y;z
is the derivative df = dx with y and z held constant. It is important in thermodynamics
to indicate the variables that are held constant, as .@f =@x/y;z is not necessarily equal to
.@f =@x/a;b where a and b are variables different from y and z.
The variables shown at the bottom of a partial derivative should tell you which variables being used as the independent variables. For example, if the partial derivative is
are
@f
then f is being treated as a function of y, a, and b.
@y a;b
480
481
A PPENDIX E C ALCULUS R EVIEW
E.3 INTEGRALS
Let f be a function of the variable x. Imagine the range of x between the limits x 0 and x 00
to be divided into many small increments of size xi .i D 1; 2; : : :/. Let fi be the value of
f when x is in the middle of the range of the ith increment. Then the integral
Z x 00
f dx
P
x0
is the sum i fi xi in the limit as each xi approaches zero and the number of terms in
the sum approaches innity. The integral is also the area under a curve of f plotted as a
function of x, measured from x D x 0 to x D x 00 . The function f is the integrand, which
is integrated over the integration variable x.
This book uses the following integrals:
Z x 00
dx D x 00 x 0
0
x
00
Z x 00
x
dx
D ln 0
x
x
x0
Z x 00
1
x a dx D
.x 00 /aC1 .x 0 /aC1
(a is a constant other than 1)
0
aC1
x
Z x 00
1 ax 00 C b
dx
D ln 0
(a is a constant)
a
ax C b
x 0 ax C b
Here are examples of the use of the expression for the third integral with a set equal to 1
and to 2:
Z x 00
1
x dx D
.x 00 /2 .x 0 /2
2
x0
Z x 00
dx
1
1
D
2
x 00 x 0
x0 x
E.4 LINE INTEGRALS

A line integral is an integral with an implicit single integration variable that constraints the
integration to a path.
R
The most frequently-seen line integral in this book, p dV , will serve as an example.
The integral can be evaluated in three different ways:
1. The integrand p can be expressed as a function of the integration variable V , so that
there is only one variable. For example, if p equals c=V where c is a constant, the
R
RV
line integral is given by p dV D c V 2 .1=V / dV D c ln.V2 =V1 /.
1
2. If p and V can be written as functions of another variable, such as time, that coordinates their values so that they follow the desired path, this new variable becomes the
integration variable.
R
3. The desired path can be drawn as a curve on a plot of p versus V ; then p dV is
equal in value to the area under the curve.
A PPENDIX F
M ATHEMATICAL P ROPERTIES OF S TATE

F UNCTIONS
A state function is a property of a thermodynamic system whose value at any given instant
depends only on the state of the system at that instant (Sec. 2.4).
F.1 DIFFERENTIALS
The differential df of a state function f is an innitesimal change of f . Since the value
of a state function by denition depends only on the state of the system, integrating df
between an initial state 1 and a nal state 2 yields the change in f , and this change is
independent of the path:
Z f2
df D f2 f1 D f
(F.1.1)
f1
A differential with this property is called an exact differential. The differential of a state
function is always exact.
F.2 TOTAL DIFFERENTIAL

A state function f treated as a dependent variable is a function of a certain number of independent variables that are also state functions. The total differential of f is df expressed
in terms of the differentials of the independent variables and has the form

@f
@f
@f
df D
dx C
dy C
dz C : : :
(F.2.1)
@x
@y
@z
There are as many terms in the expression on the right side as there are independent variables. Each partial derivative in the expression has all independent variables held constant
except the variable shown in the denominator.
Figure F.1 on the next page interprets this expression for a function f of the two independent variables x and y. The shaded plane represents a small element of the surface
f D f .x; y/.
Consider a system with three independent variables. If we choose these independent
variables to be x, y, and z, the total differential of the dependent state function f takes the
482
A PPENDIX F M ATHEMATICAL P ROPERTIES OF S TATE F UNCTIONS
483

@f
slope in f -x plane D
@x y

@f
slope in f -y plane D
@y x
9
o
@f
>

dy >
=
@f
@f
) @y
x
dx C
dy
df D
@x y
@y x
>
@f
;
dx>
@x y
dx
dy
Figure F.1
form
df D a dx C b dy C c dz
where we can identify the coefcients as

@f
@f
aD
bD
@x y;z
@y x;z
cD
(F.2.2)
@f
@z
(F.2.3)
x;y
These coefcients are themselves, in general, functions of the independent variables and
may be differentiated to give mixed second partial derivatives; for example:

@a
@2f
@b
@2f
D
D
(F.2.4)
@y x;z
@y@x
@x y;z
@x@y
The second partial derivative @2f =@y@x, for instance, is the partial derivative with respect
to y of the partial derivative of f with respect to x. It is a theorem of calculus that if a
function f is single valued and has continuous derivatives, the order of differentiation in a
mixed derivative is immaterial. Therefore the mixed derivatives @2f =@y@x and @2f =@x@y,
evaluated for the system in any given state, are equal:

@a
@b
D
(F.2.5)
@y x;z
@x y;z
The general relation that applies to a function of any number of independent variables is

@X
@Y
D
(F.2.6)
@y
@x
where x and y are any two of the independent variables, X is @f =@x, Y is @f =@y, and
each partial derivative has all independent variables held constant except the variable shown
in the denominator. This general relation is the Euler reciprocity relation, or reciprocity
relation for short. A necessary and sufcient condition for df to be an exact differential is
that the reciprocity relation is satised for each pair of independent variables.
484
F.3 INTEGRATION OF A TOTAL DIFFERENTIAL

If the coefcients of the total differential of a dependent variable are known as functions
of the independent variables, the expression for the total differential may be integrated to
obtain an expression for the dependent variable as a function of the independent variables.
For example, suppose the total differential of the state function f .x; y; z/ is given by
Eq. F.2.2 and the coefcients are known functions a.x; y; z/, b.x; y; z/, and c.x; y; z/.
Because f is a state function, its change between f .0; 0; 0/ and f .x 0 ; y 0 ; z 0 / is independent
of the integration path taken between these two states. A convenient path would be one with
the following three segments:
R x0
1. integration from .0; 0; 0/ to .x 0 ; 0; 0/: 0 a.x; 0; 0/ dx
R y0
2. integration from .x 0 ; 0; 0/ to .x 0 ; y 0 ; 0/: 0 b.x 0 ; y; 0/ dy
R z0
3. integration from .x 0 ; y 0 ; 0/ to .x 0 ; y 0 ; z 0 /: 0 c.x 0 ; y 0 ; z/ dz
The expression for f .x; y; z/ is then the sum of the three integrals and a constant of integration.
Here is an example of this procedure applied to the total differential
df D .2xy/ dx C .x 2 C z/ dy C .y
9z 2 / dz
(F.3.1)
An expression for the function f in this example is given by the sum

f D
x0
0
.2x 0/ dx C
D 0 C x 2 y C .yz
D x 2 y C yz
y0
0
0 2
.x / C 0 dy C
z0
.y 0
9z 2 / dz C C
9z 3 =3/ C C
3z 3 C C
(F.3.2)
where primes are omitted on the second and third lines because the expressions are supposed
to apply to any values of x, y, and z. C is an integration constant. You can verify that the
third line of Eq. F.3.2 gives the correct expression for f by taking partial derivatives with
respect to x, y, and z and comparing with Eq. F.3.1.
In chemical thermodynamics, there is not likely to be occasion to perform this kind
of integration. The fact that it can be done, however, shows that if we stick to one set of
independent variables, the expression for the total differential of an independent variable
contains the same information as the independent variable itself.
A different kind of integration can be used to express a dependent extensive property
in terms of independent extensive properties. An extensive property of a thermodynamic
system is one that is additive, and an intensive property is one that is not additive and has
the same value everywhere in a homogeneous region (Sec. 2.1.1). Suppose we have a state
function f that is an extensive property with the total differential
df D a dx C b dy C c dz C : : :
(F.3.3)
where the independent variables x; y; z; : : : are extensive and the coefcients a; b; c; : : :

are intensive. If the independent variables include those needed to describe an open system
(for example, the amounts of the substances), then it is possible to integrate both sides of
485
the equation from a lower limit of zero for each of the extensive functions while holding the
intensive functions constant:
Z f0
Z x0
Z y0
Z z0
df D a
dx C b
dy C c
dz C : : :
(F.3.4)
0
f 0 D ax 0 C by 0 C cz 0 C : : :
(F.3.5)
Note that a term of the form c du where u is intensive becomes zero when integrated with
intensive functions held constant, because du is this case is zero.
F.4 LEGENDRE TRANSFORMS

A Legendre transform of a state function is a linear change of one or more of the independent variables made by subtracting products of conjugate variables.
To understand how this works, consider a state function f whose total differential is
given by
df D a dx C b dy C c dz
(F.4.1)
In the expression on the right side, x, y, and z are being treated as the independent variables.
The pairs a and x, b and y, and c and z are conjugate pairs. That is, a and x are conjugates,
b and y are conjugates, and c and z are conjugates.
For the rst example of a Legendre transform, we dene a new state function f1 by
subtracting the product of the conjugate variables a and x:
def
f1 D f
ax
(F.4.2)
The function f1 is a Legendre transform of f . We take the differential of Eq. F.4.2

df1 D df
a dx
x da
(F.4.3)
and substitute for df from Eq. F.4.1:

df1 D .a dx C b dy C c dz/
D
a dx
x da
x da C b dy C c dz
(F.4.4)
Equation F.4.4 gives the total differential of f1 with a, y, and z as the independent variables.
The functions x and a have switched places as independent variables. What we did in order
to let a replace x as an independent variable was to subtract from f the product of the
conjugate variables a and x.
Because the right side of Eq. F.4.4 is an expression for the total differential of the state
function f1 , we can use the expression to identify the coefcients as partial derivatives of
f1 with respect to the new set of independent variables:
@f1
@f1
@f1
xD
bD
cD
(F.4.5)
@a y;z
@y a;z
@z a;y
We can also use Eq. F.4.4 to write new reciprocity relations, such as

@x
@b
D
@y a;z
@a y;z
(F.4.6)
486
We can make other Legendre transforms of f by subtracting one or more products of

conjugate variables. A second example of a Legendre transform is
def
f2 D f
by
b dy
y db
cz
(F.4.7)
whose total differential is

df2 D df
D a dx
y db
z dc
c dz
z dc
(F.4.8)
Here b has replaced y and c has replaced z as independent variables. Again, we can identify
the coefcients as partial derivatives and write new reciprocity relations.
If we have an algebraic expression for a state function as a function of independent variables, then a Legendre transform preserves all the information contained in that expression.
To illustrate this, we can use the state function f and its Legendre transform f2 described
above. Suppose we have an expression for f .x; y; z/this is f expressed as a function of
the independent variables x, y, and z. Then by taking partial derivatives of this expression,
we can nd according to Eq. F.2.3 expressions for the functions a.x; y; z/, b.x; y; z/, and
c.x; y; z/.
Now we perform the Legendre transform of Eq. F.4.7: f2 D f by cz with total
differential df2 D a dx y db z dc (Eq. F.4.8). The independent variables have been
changed from x, y, and z to x, b, and c.
We want to nd an expression for f2 as a function of these new variables, using the
information available from the original function f .x; y; z/. To do this, we eliminate z
from the known functions b.x; y; z/ and c.x; y; z/ and solve for y as a function of x, b,
and c. We also eliminate y from b.x; y; z/ and c.x; y; z/ and solve for z as a function
of x, b, and c. This gives us expressions for y.x; b; c/ and z.x; b; c/ which we substitute
into the expression for f .x; y; z/, turning it into the function f .x; b; c/. Finally, we use
the functions of the new variables to obtain an expression for f2 .x; b; c/ D f .x; b; c/
by.x; b; c/ cz.x; b; c/.
The original expression for f .x; y; z/ and the new expression for f2 .x; b; c/ contain
the same information. We could take the expression for f2 .x; b; c/ and, by following the
same procedure with the Legendre transform f D f2 C by C cz, retrieve the expression
for f .x; y; z/. Thus no information is lost during a Legendre transform.
A PPENDIX G
F ORCES , E NERGY,
AND
W ORK
The aim of this appendix is to describe a simple model that will help to clarify the meaning
of energy and mechanical work in macroscopic systems. The appendix applies fundamental
principles of classical mechanics to a collection of material particles representing a closed
system and its surroundings. Although classical mechanics cannot duplicate all features of
a chemical systemfor instance, quantum properties of atoms are ignoredthe behavior
of the particles and their interactions will show us how to evaluate the thermodynamic work
in a real system.
In broad outline the derivation is as follows. An inertial reference frame in which Newtons laws of motion are valid is used to describe the positions and velocities of the particles.
The particles are assumed to exert central forces on one another, such that between any two
particles the force is a function only of the interparticle distance and is directed along the
line between the particles.
We dene the kinetic energy of the collection of particles as the sum for all particles
of 1 mv 2 (where m is mass and v is velocity). We dene the potential energy as the sum
2
over pairwise particleparticle interactions of potential functions that depend only on the
interparticle distances. The total energy is the sum of the kinetic and potential energies.
With these denitions and Newtons laws, a series of mathematical operations leads to the
principle of the conservation of energy: the total energy remains constant over time.
Continuing the derivation, we consider one group of particles to represent a closed thermodynamic system and the remaining particles to constitute the surroundings. The system
particles may interact with an external force eld, such as a gravitational eld, created by
some of the particles in the surroundings. The energy of the system is considered to be the
sum of the kinetic energy of the system particles, the potential energy of pairwise particle
particle interactions within the system, and the potential energy of the system particles in
any external eld or elds. The change in the system energy during a time interval is then
found to be given by a certain sum of integrals which, in the transition to a macroscopic
model, becomes the sum of heat and thermodynamic work in accord with the rst law of
thermodynamics.
A similar derivation, using a slightly different notation, is given in Ref. [44].
487
APPENDIX G FORCES, ENERGY, AND WORK

G.1 F ORCES BETWEEN PARTICLES
488
G.1 FORCES BETWEEN PARTICLES

A material particle is a body that has mass and is so small that it behaves as a point,
without rotational energy or internal structure. We assume each particle has a constant mass,
ignoring relativistic effects that are important only when the particle moves at a speed close
to the speed of light.
Consider a collection of an arbitrary number of material particles that have interactions
only among themselves and with no other particles. Later we will consider some of the
particles to constitute a thermodynamic system and the others to be the surroundings.
Newtons laws of motion are obeyed only in an inertial reference frame. A reference
frame that is xed or moving at a constant velocity relative to local stars is practically an
inertial reference frame. To a good approximation, a reference frame xed relative to the
earths surface is also an inertial system (the necessary corrections are discussed in Sec.
G.10). This reference frame will be called simply the lab frame, and treated as an inertial
frame in order that we may apply Newtons laws.
It will be assumed that the Cartesian components of all vector quantities appearing in
Sections G.1G.4 are measured in an inertial lab frame.
Classical mechanics is based on the idea that one material particle acts on another by
means of a force that is independent of the reference frame. Let the vector Fij denote the
force exerted on particle i by particle j .1 The net force Fi acting on particle i is the vector
sum of the individual forces exerted on it by the other particles:2
X
Fi D
Fij
(G.1.1)
j i
(The term in which j equals i has to be omitted because a particle does not act on itself.)
According to Newtons second law of motion, the net force Fi acting on particle i is equal
to the product of its mass mi and its acceleration:
F i D mi
dvi
dt
(G.1.2)
Here vi is the particles velocity in the lab frame and t is time.

A nonzero net force causes particle i to accelerate and its velocity and position to
change. The work done by the net force acting on the particle in a given time interval
is dened by the integral3
Z
Wi D
Fi dri
(G.1.3)
where ri is the position vector of the particlea vector from the origin of the lab frame to
the position of the particle.
1 This and the next two footnotes are included for readers who are not familiar with vector notation.
The quantity
Fij is printed in boldface to indicate it is a vector having both magnitude and direction.
2 The rule for adding vectors, as in the summation shown here, is that the sum is a vector whose component
along each axis of a Cartesian coordinate system is the sum of the components along that axis of the vectors
being added. For example, the vector C D A C B has components Cx D Ax C Bx , Cy D Ay C By , and
Cz D Az C Bz .
3 The dot between the vectors in the integrand indicates a scalar product or dot product, which is a nonvector
quantity. The general denition of the scalar product of two vectors, A and B, is A B D AB cos where A
and B are the magnitudes of the two vectors and is the angle between their positive directions.

489
The integral on the right side of Eq. G.1.3 is an example of a line integral. It indicates
that the scalar product of the net force acting on the particle and the particles displacement is to be integrated over time during the time interval. The integral can be written
R
without vectors in the form Fi cos .ds= dt / dt where Fi is the magnitude of the net
force, ds= dt is the magnitude of the velocity of the particle along its path in threedimensional space, and is the angle between the force and velocity vectors. The
three quantities Fi , cos , and ds= dt are all functions of time, t, and the integration is
carried out with time as the integration variable.
By substituting the expression for Fi (Eq. G.1.2) in Eq. G.1.3, we obtain

Z
Z
Z
Z
dvi
dri
Wi D mi
dri D mi
dvi D mi vi dvi D mi vi dvi
dt
dt
2
D 1 mi vi
2
(G.1.4)
where vi is the magnitude of the velocity.

2
The quantity 1 mi vi is called the kinetic energy of particle i. This kinetic energy de2
pends only on the magnitude of the velocity (i.e., on the speed) and not on the particles
position.
The total work Wtot done by all forces acting on all particles during the time interval is
P
the sum of Wi for all particles: Wtot D i Wi .4 Equation G.1.4 then gives us
!
X
X
2
2
1
1
Wtot D
2 mi vi D
(G.1.5)
2 mi vi
i
Equation G.1.5 shows that the total work during a time interval is equal to the change in
the total kinetic energy in this interval. This result is called the work-energy principle by
physicists.5
From Eqs. G.1.1 and G.1.3 we obtain a second expression for Wtot :
X
XZ X
XXZ
Wtot D
Wi D
Fij dri D
Fij dri
(G.1.6)
i
i j i
j i
The double sum in the right-most expression can be written as a sum over pairs of particles,
the term for the pair i and j being
Z
Z
Z
Z
Fij dri C Fj i drj D Fij dri
Fij drj
Z
Z
D Fij d.ri rj / D .Fij eij / drij
(G.1.7)
Here we have used the relations Fj i D Fij (from Newtons third law) and .ri rj / D
eij rij , where eij is a unit vector pointing from j to i and rij is the distance between the
particles. Equation G.1.6 becomes
XXZ
XXZ
(G.1.8)
.Fij eij / drij
Wtot D
Fij dri D
i j i
j >i
4 The
work Wtot dened here is not the same as the thermodynamic work appearing in the rst law of thermodynamics.
5 Ref. [149], p. 95.

490
Next we look in detail at the force that particle j exerts on particle i . This force depends
on the nature of the two particles and on the distance between them. For instance, Newtons
2
law of universal gravitation gives the magnitude of a gravitational force as Gmi mj =rij ,
where G is the gravitational constant. Coulombs law gives the magnitude of an electrical
2
force between stationary charged particles as Qi Qj =.4 0 rij /, where Qi and Qj are the
charges and 0 is the electric constant (or permittivity of vacuum). These two kinds of forces
are central forces that obey Newtons third law of action and reaction, namely, that the forces
exerted by two particles on one another are equal in magnitude and opposite in direction and
are directed along the line joining the two particles. (In contrast, the electromagnetic force
between charged particles in relative motion does not obey Newtons third law.)
We will assume the force Fij exerted on particle i by particle j has a magnitude that
depends only on the interparticle distance rij and is directed along the line between i and
j , as is true of gravitational and electrostatic forces and on intermolecular forces in general.
Then we can dene a potential function, ij , for this force that will be a contribution to
the potential energy. To see how ij is related to Fij , we look at Eq. G.1.7. The left-most
R
R
expression, Fij dri C Fj i drj , is the change in the kinetic energies of particles i and
j during a time interval (see Eq. G.1.4). If these were the only particles, their total energy
should be constant for conservation of energy; thus ij should have the same magnitude
and the opposite sign of the kinetic energy change:
Z
ij D
.Fij eij / drij
(G.1.9)
The value of ij at any interparticle distance rij is fully dened by Eq. G.1.9 and the choice
of an arbitrary zero. The quantity .Fij eij / is simply the component of the force along the
line between the particles, and is negative for an attractive force (one in which Fij points
from i to j ) and positive for a repulsive force. If the force is attractive, the value of ij
increases with increasing rij ; if the force is repulsive, ij decreases with increasing rij .
Since ij is a function only of rij , it is independent of the choice of reference frame.
Equations G.1.8 and G.1.9 can be combined to give
0
1
XX
XX
Wtot D
ij D @
ij A
(G.1.10)
i
j >i
j >i
By equating the expressions for Wtot given by Eqs. G.1.5 and G.1.10 and rearranging, we
obtain
0
1
!
X
XX
2
1
C @
ij A D 0
(G.1.11)
2 mi vi
i
j >i
This equation shows that the quantity

X
XX
2
1
Etot D
ij
2 mi vi C
i
(G.1.12)
j >i
is constant over time as the particles move in response to the forces acting on them. The
rst term on the right side of Eq. G.1.12 is the total kinetic energy of the particles. The
second term is the pairwise sum of particleparticle potential functions; this term is called

G.2 T HE S YSTEM AND S URROUNDINGS
491
ij
i k
ij
kl
i k 00 i k 0
(a)
(b)
Figure G.1 Assignment of particles to groups, and some representative particle

particle potential functions (schematic). The closed dashed curve represents the system boundary.
(a) The open circles represent particles in the system, and the lled circles are particles
in the surroundings.
(b) The lled triangles are particles in the surroundings that are the source of a conservative force eld for particles in the system.
the potential energy of the particles. Note that the kinetic energy depends only on particle
speeds and the potential energy depends only on particle positions.
The signicance of Eq. G.1.11 is that the total energy Etot dened by Eq. G.1.12 is
conserved. This will be true provided the reference frame used for kinetic energy is inertial
and the only forces acting on the particles are those responsible for the particleparticle
potential functions.
G.2 THE SYSTEM AND SURROUNDINGS

Now we are ready to assign the particles to two groups: particles in the system and those in
the surroundings. This section will use the following convention: indices i and j refer to
particles in the system; indices k and l refer to particles in the surroundings. This division
of particles is illustrated schematically in Fig. G.1(a). With this change in notation, Eq.
G.1.12 becomes
X
XX
XX
XX
X
2
2
1
1
Etot D
mi vi C
ij C
ik C
mk vk C
kl (G.2.1)
2
2
i
j >i
k l>k
A portion of the surroundings may create a time-independent conservative force eld

(an external eld) for a particle in the system. In order for such a eld to be present,
its contribution to the force exerted on the particle and to the particles potential energy
must depend only on the particles position in the lab frame. The usual gravitational and
electrostatic elds are of this type.
In order to clarify the properties of a conservative external eld, the index k 0 will be
used for those particles in the surroundings that are not the source of an external eld, and
k 00 for those that are, as indicated in Fig. G.1(b). Then the force exerted on system particle
P
i due to the eld is Field D k 00 Fik 00 . If this were the only force acting on particle i ,
R
the change in its kinetic energy during a time interval would be Field dri (Eq. G.1.4).

G.2 T HE S YSTEM AND S URROUNDINGS
492
For conservation of energy, the potential energy change in the time interval should have the
same magnitude and the opposite sign:
Z
eld
i D
Field dri
(G.2.2)
R
Only if the integral Field dri has the same value for all paths between the initial and
nal positions of the particle does a conservative force eld exist; otherwise the concept of
a potential energy ield is not valid.
Taking a gravitational eld as an example of a conservative external eld, we replace
R grav
grav
grav
grav
Field and ield by Fi
and i : i D
Fi
dri . The gravitational force on
particle i is, from Newtons rst law, the product of the particle mass and its acceleration
grav
gez in the gravitational eld: Fi
D mi gez where g is the acceleration of free fall
and ez is a unit vector in the vertical (upward) z direction. The change in the gravitational
potential energy given by Eq. G.2.2 is
Z
grav
i D mi g ez dri D mi g zi
(G.2.3)
(The range of elevations of the system particles is assumed to be small compared with the
earths radius, so that each system particle experiences essentially the same constant value
of g.) Thus we can dene the gravitational potential energy of particle i, which is a function
grav
only of the particles vertical position in the lab frame, by i D mi gzi C Ci where Ci is
an arbitrary constant.
Returning to Eq. G.2.1 for the total energy, we can now write the third term on the right
side in the form
XX
XX
X
ik D
i k 0 C
ield
(G.2.4)
i
k0
To divide the expression for the total energy into meaningful parts, we substitute Eq. G.2.4
in Eq. G.2.1 and rearrange in the form
2
3
X
XX
X
2
1
Etot D 4
ij C
ield 5
2 mi vi C
i
"
XX
k0
j >i
ik 0 C
"
2
1
2 mk vk
XX
kl
(G.2.5)
k l>k
The terms on the right side of this equation are shown grouped with brackets into three
quantities. The rst quantity depends only on the speeds and positions of the particles in
the system, and thus represents the energy of the system:
X
XX
X
2
1
Esys D
mi vi C
ij C
ield
(G.2.6)
2
i
j >i
The three terms in this expression for Esys are, respectively, the kinetic energy of the system
particles relative to the lab frame, the potential energy of interaction among the system
particles, and the total potential energy of the system in the external eld.

G.3 S YSTEM E NERGY C HANGE
493
The last bracketed quantity on the right side of Eq. G.2.5 depends only on the speeds
and positions of all the particles in the surroundings, so that this quantity is the energy of
the surroundings, Esurr . Thus, an abbreviated form of Eq. G.2.5 is
XX
Etot D Esys C
i k 0 C Esurr
(G.2.7)
k0
P P
The quantity i k 0 ik 0 represents potential energy shared by both the system and surroundings on account of forces acting across the system boundary, other than gravitational forces or forces from other external elds. The forces responsible for the quantity
P P
i
k 0 ik 0 are generally signicant only between particles in the immediate vicinity of
the system boundary, and will presently turn out to be the crucial forces for evaluating
thermodynamic work.
G.3 SYSTEM ENERGY CHANGE

This section derives an important relation between the change Esys of the energy of the
system measured in a lab frame, and the forces exerted by the surroundings on the system
particles. The indices i and j will refer to only the particles in the system.
We write the net force on particle i in the form
X
Fi D
Fij C Field C Fisur
(G.3.1)
j i
where Fij isP force exerted on particle i by particle j , both particles being in the system,
the
and Fisur D k 0 Fik 0 is the net force exerted on particle i by the particles in the surroundings that are not the source of an external eld. During a given period of time, the work
done by forces acting on only the system particles is
XZ
XXZ
XZ
XZ
Fi dri D
Fij dri C
Field dri C
Fisur dri (G.3.2)
i
i j i
We can replace the rst three sums in this equation with new expressions. Using Eq.
G.1.4, we have
!
XZ
X
2
1
Fi dri D
(G.3.3)
2 mi vi
i
From Eqs. G.1.8 and G.1.9 we obtain

XXZ
i j i
Fij dri D
0
1
XX
@
ij A
i
where the sums are over the system particles. From Eq. G.2.2 we can write
!
XZ
X
Field dri D
ield
i
(G.3.4)
j >i
(G.3.5)

G.4 M ACROSCOPIC W ORK
494
Combining Eqs. G.3.2G.3.5 and rearranging, we obtain

0
1
XZ
X
XX
X
2
1
Fisur dri D @
ij C
ield A
2 mi vi C
i
j >i
(G.3.6)
Comparison of the expression on the right side of this equation with Eq. G.2.6 shows that
the expression is the same as the change of Esys :
XZ
Esys D
Fisur dri
(G.3.7)
i
Recall that the vector Fisur is the force exerted on particle i, in the system, by the particles in
the surroundings other than those responsible for an external eld. Thus Esys is equal to
the total work done on the system by the surroundings, other than work done by an external
eld such as a gravitational eld.
It might seem strange that work done by an external eld is not included in Esys . The
reason it is not included is that ield was dened to be a potential energy belonging
only to the system, and is thus irrelevant to energy transfer from or to the surroundings.
As a simple example of how this works, consider a system consisting of a solid
body in a gravitational eld. If the only force exerted on the body is the downward
gravitational force, then the body is in free fall but Esys in the lab frame is zero; the
loss of gravitational potential energy as the body falls is equal to the gain of kinetic
energy. On the other hand, work done on the system by an external force that opposes
the gravitational force is included in Esys . For example, if the body is pulled upwards
at a constant speed with a string, its potential energy increases while its kinetic energy
remains constant, and Esys increases.
G.4 MACROSCOPIC WORK

In thermodynamics we are interested in the quantity of work done on macroscopic parts of
the system during a process, rather than the work done on individual particles. Macroscopic
work is the energy transferred across the system boundary due to concerted motion of many
particles on which the surroundings exert a force. Macroscopic mechanical work occurs
when there is displacement of a macroscopic portion of the system on which a short-range
contact force acts across the system boundary. This force could be, for instance, the pressure
of an external uid at a surface element of the boundary multiplied by the area of the surface
element, or it could be the tension in a cord at the point where the cord passes through the
boundary.
The symbol wlab will refer to macroscopic work measured with displacements in the
lab frame.
At any given instant, only the system particles that are close to the boundary will have
nonnegligible contact forces exerted on them. We can dene an interaction layer, a thin
shell-like layer within the system and next to the system boundary that contains all the
system particles with appreciable contact forces. We imagine the interaction layer to be
divided into volume elements, or segments, each of which either moves as a whole during
the process or else is stationary. Let R be a position vector from the origin of the lab frame

G.4 M ACROSCOPIC W ORK
495
ri

ri
R
Figure G.2 Position vectors within the system. Segment of the interaction layer
lies within the heavy curve (representing the system boundary) and the dashed lines.
Open circle: origin of lab frame; open square: point xed in system boundary at
segment ; lled circle: particle i.
to a point xed in the boundary at segment , and let ri be a vector from this point to
particle i (Fig. G.2). Then the position vector for particle i can be written ri D R C ri .
Let F sur be the total contact force exerted by the surroundings on the system particles in
P
segment : F sur D i i Fisur , where i is equal to 1 when particle i is in segment and
is zero otherwise.
The change in the system energy during a process is, from Eq. G.3.7,
XZ
XXZ
sur
Esys D
Fi
dri D
i Fisur d .R C ri /
i
XZ
sur
dR C
XXZ
i Fisur dri
(G.4.1)
R
We recognize the integral F sur dR as the macroscopic work at surface element ,
because it is the integrated scalar product of the force exerted by the surroundings and the
displacement. The total macroscopic work during the process is then given by
XZ
wlab D
F sur dR
(G.4.2)
Heat, qlab , can be dened as energy transfer to or from the system that is not accounted
for by macroscopic work. This transfer occurs by means of chaotic motions and collisions
of individual particles at the boundary. With this understanding, Eq. G.4.1 becomes
Esys D qlab C wlab
with wlab given by the expression in Eq. G.4.2 and qlab given by
XXZ
qlab D
i Fisur dri
(G.4.3)
(G.4.4)

G.5 T HE W ORK D ONE ON THE S YSTEM AND S URROUNDINGS
496
G.5 THE WORK DONE ON THE SYSTEM AND SURROUNDINGS

An additional comment can be made about the transfer of energy between the system and
the surroundings. We may use Eq. G.4.2, with appropriate redenition of the quantities on
the right side, to evaluate the work done on the surroundings. This work may be equal in
magnitude and opposite in sign to the work wlab done on the system. A necessary condition
for this equality is that the interacting parts of the system and surroundings have equal
displacements; that is, that there be continuity of motion at the system boundary. We expect
there to be continuity of motion when a uid contacts a moving piston or paddle.
Suppose, however, that the system is stationary and an interacting part of the surroundings moves. Then according to Eq. G.4.2, wlab is zero, whereas the work done on or by that
part of the surroundings is not zero. How can this be, considering that Etot remains constant? One possibility, discussed by Bridgman,6 is sliding friction at the boundary: energy
lost by the surroundings in the form of work is gained by the system and surroundings in
the form of thermal energy. Since the effect on the system is the same as a ow of heat
from the surroundings, the division of energy transfer into heat and work can be ambiguous
when there is sliding friction at the boundary.7
Another way work can have different magnitudes for system and surroundings is a
change in potential energy shared by the system and surroundings. This shared energy
is associated with forces acting across the boundary, other than from a time-independent
P P
external eld, and is represented in Eq. G.2.7 by the sum i k 0 i k 0 . In the usual types
of processes this sum is either practically constant, or else each term falls off so rapidly
with distance that the sum is negligible. Since Etot is constant, during such processes the
quantity Esys C Esurr remains essentially constant.
G.6 THE LOCAL FRAME AND INTERNAL ENERGY

As explained in Sec. 2.6.2, a lab frame may not be an appropriate reference frame in which
to measure changes in the systems energy. This is the case when the system as a whole
moves or rotates in the lab frame, so that Esys depends in part on external coordinates that
are not state functions. In this case it may be possible to dene a local frame moving with
the system in which the energy of the system is a state function, the internal energy U .
As before, ri is the position vector of particle i in a lab frame. A prime notation will
be used for quantities measured in the local frame. Thus the position of particle i relative
to the local frame is given by vector ri0 , which points from the origin of the local frame to
0
particle i (see Fig. G.3). The velocity of the particle in the local frame is vi D dri0 = dt.
We continue to treat the earth-xed lab frame as an inertial frame, although this is not
strictly true (Sec. G.10). If the origin of the local frame moves at constant velocity in the lab
frame, with Cartesian axes that do not rotate with respect to those of the lab frame, then the
6 Ref.
[23], p. 4756.
ambiguity can be removed by redening the system boundary so that a thin stationary layer next to
the sliding interface, on the side that was originally part of the system, is considered to be included in the
surroundings instead of the system. The layer removed from the system by this change can be so thin that the
values of the systems extensive properties are essentially unaffected. With this redened boundary, the energy
transfer across the boundary is entirely by means of heat.
7 The

G.6 T HE L OCAL F RAME AND I NTERNAL E NERGY
497
ri
0
R
z0
ri0
y0
x0
ri
R
Rloc
y
Figure G.3 Vectors in the lab and local reference frames. Open circle: origin of
lab frame; open triangle: origin of local frame; open square: point xed in system
boundary at segment ; lled circle: particle i. The thin lines are the Cartesian axes
of the reference frames.
local frame is also inertial but U is not equal to Esys and the change U during a process
is not necessarily equal to Esys .
If the origin of the local frame moves with nonconstant velocity in the lab frame, or if the
local frame rotates with respect to the lab frame, then the local frame has nite acceleration
and is noninertial. In this case the motion of particle i in the local frame does not obey
Newtons second law as it does in an inertial frame. We can, however, dene an effective
net force Fieff whose relation to the particles acceleration in the local frame has the same
form as Newtons second law:
dv0
Fieff D mi i
(G.6.1)
dt
To an observer who is stationary in the local frame, the effective force will appear to make
the particles motion obey Newtons second law even though the frame is not inertial.
The net force on particle i from interactions with other particles is given by Eq. G.3.1:
P
Fi D j i Fij C Field C Fisur . The effective force can be written
Fieff D Fi C Fiaccel
(G.6.2)
where Fiaccel is the contribution due to acceleration. Fiaccel is not a true force in the sense of
resulting from the interaction of particle i with other particles. Instead, it is an apparent or
ctitious force introduced to make it possible to write Eq. G.6.1 which resembles Newtons
second law. The motion of particle i in an inertial frame is given by mi dvi = dt D Fi ,
0
whereas the motion in the local frame is given by mi dvi = dt D Fi C Fiaccel .
A simple example may make these statements clear. Consider a small unattached object
suspended in the weightless environment of an orbiting space station. Assume the object
is neither moving nor spinning relative to the station. Let the object be the system, and x
the local frame in the space station. The local frame rotates with respect to local stars as the
station orbits around the earth; the local frame is therefore noninertial. The only true force
exerted on the object is a gravitational force directed toward the earth. This force explains
the objects acceleration relative to local stars. The fact that the object has no acceleration
in the local frame can be explained by the presence of a ctitious centrifugal force having

498
the same magnitude as the gravitational force but directed in the opposite direction, so that
the effective force on the object as a whole is zero.
The reasoning used to derive the equations in Secs. G.1G.4 can be applied to an arbi0
trary local frame. To carry out the derivations we replace Fi by Fieff , ri by ri0 , and vi by vi ,
and use the local frame to measure the Cartesian components of all vectors. We need two
new potential energy functions for the local frame, dened by the relations
Z
0
def
i eld D
Field dri0
(G.6.3)
def
iaccel D
Fiaccel dri0
(G.6.4)
Both i eld and accel must be time-independent functions of the position of particle i in
the local frame in order to be valid potential functions. (If the local frame is inertial, Fiaccel
and iaccel are zero.)
The detailed line of reasoning in Secs. G.1G.4 will not be repeated here, but the reader
can verify the following results. The total energy of the system and surroundings measured
P P
0
0
in the local frame is given by Etot D U C i k 0 i k 0 C Esurr where the index k 0 is for
particles in the surroundings that are not the source of an external eld for the system. The
energy of the system (the internal energy) is given by
X
XX
X 0
X
0 2
1
U D
ij C
i eld C
iaccel
(G.6.5)
2 mi .vi / C
i
j >i
where the indices i and j are for system particles. The energy of the surroundings measured
in the local frame is
XX
X
X
0
0
accel
1
mk .vk /2 C
Esurr D
kl C
k
(G.6.6)
2
k
l>i
0
where k and l are indices for particles in the surroundings. The value of Etot is found to
be constant over time, meaning that energy is conserved in the local frame. The internal
energy change during a process is the sum of the heat q measured in the local frame and the
macroscopic work w in this frame:
U D q C w
The expressions for q and w, analogous to Eqs. G.4.4 and G.4.2, are found to be
XXZ
qD
i Fisur dri
(G.6.7)
(G.6.8)
wD
XZ
F sur dR 0
(G.6.9)
In these equations R 0 is a vector from the origin of the local frame to a point xed in the
system boundary at segment , and ri is a vector from this point to particle i (see Fig.
G.3).

499
We expect that an observer in the local frame will nd the laws of thermodynamics
are obeyed. For instance, the Clausius statement of the second law (Sec. 4.2) is as valid
in a manned orbiting space laboratory as it is in an earth-xed laboratory: nothing the
investigator can do will allow energy to be transferred by heat from a colder to a warmer
body through a device operating in a cycle. Equation G.6.7 is a statement of the rst law
of thermodynamics (box on page 56) in the local frame. Accordingly, we may assume that
the thermodynamic derivations and relations treated in the body of this book are valid in
any local frame, whether or not it is inertial, when U and w are dened by Eqs. G.6.5 and
G.6.9.
In the body of the book, w is called the thermodynamic work, or simply the work. Note
the following features brought out by the derivation of the expression for w:
P R sur
The equation w D
F
dR 0 has been derived for a closed system.
The equation shows how we can evaluate the thermodynamic work w done on the
system. For each moving surface element of the system boundary at segment of the
interaction layer, we need to know the contact force F sur exerted by the surroundings
and the displacement dR 0 in the local frame.
We could equally well calculate w from the force exerted by the system on the sursys
roundings. According to Newtons third law, the force F exerted by segment has
sur and the opposite direction: F sys D F sur .
the same magnitude as F
During a process, a point xed in the system boundary at segment is either stationary or traverses a path in three-dimensional space. At each intermediate stage of the
process, let s be the length of the path that began in the initial state. We can write
the innitesimal quantity F sur dR 0 in the form F sur cos ds , where F sur is the
magnitude of the force, ds is an innitesimal change of the path length, and is
the angle between the directions of the force and the displacement. We then obtain
the following integrated and differential forms of the work:
XZ
X
wD
F sur cos ds
w D
F sur cos ds
(G.6.10)
If only one portion of the boundary moves in the local frame, and this portion has
sur
linear motion parallel to the x 0 axis, we can replace F sur dR 0 by Fx 0 dx 0 , where
sur
0 component of the force exerted by the surroundings on the moving
Fx 0 is the x
boundary and dx 0 is an innitesimal displacement of the boundary. In this case we
can write the following integrated and differential forms of the work:
Z
sur
sur
w D Fx 0 dx 0
w D Fx 0 dx 0
(G.6.11)
The work w does not include work done internally by one part of the system on
another part.
In the calculation of work with Eqs. G.6.9G.6.11, we do not include forces from an
external eld such as a gravitational eld, or ctitious forces Fiaccel if present.

G.7 N ONROTATING L OCAL F RAME
500
G.7 NONROTATING LOCAL FRAME

Consider the case of a nonrotating local frame whose origin moves in the lab frame but
whose Cartesian axes x 0 , y 0 , z 0 remain parallel to the axes x, y, z of the lab frame. In this
case the Cartesian components of Fisur for particle i are the same in both frames, and so
also are the Cartesian components of the innitesimal vector displacement dri . According
to Eqs. G.4.4 and G.6.8, then, for an arbitrary process the value of the heat q in the local
frame is the same as the value of the heat qlab in the lab frame.
From Eqs. G.4.3 and G.6.7 with qlab set equal to q, we obtain the useful relation
U
Esys D w
wlab
(G.7.1)
This equation is not valid if the local frame has rotational motion with respect to the lab
frame.
The vector R 0 has the same Cartesian components in the lab frame as in the nonrotating
local frame, so we can write R
R 0 D Rloc where Rloc is the position vector in the lab
frame of the origin of the local frame (see Fig. G.3). From Eqs. G.4.2 and G.6.9, setting
.R
R 0 / equal to Rloc , we obtain the relation
!
Z X
XZ
sur
0
sur
w wlab D
F
d.R
R /D
F
dRloc
(G.7.2)
P
The sum
F sur is the net contact force exerted on the system by the surroundings.
For example, suppose the system is a uid in a gravitational eld. Let the system boundary
be at the inner walls of the container, and let the local frame be xed with respect to the
container and have negligible acceleration in the lab frame. At each surface element of the
boundary, the force exerted by the pressure of the uid on the container wall is equal in
magnitude and opposite in direction to the contact force exerted by the surroundings on the
uid. The horizontal components of the contact forces on opposite sides of the container
cancel, but the vertical components do not cancel because of the hydrostatic pressure. The
net contact force is mgez , where m is the system mass and ez is a unit vector in the vertical
Cz direction. For this example, Eq. G.7.2 becomes
w
wlab D
mgzloc
(G.7.3)
where zloc is the elevation in the lab frame of the origin of the local frame.
G.8 CENTER-OF-MASS LOCAL FRAME

If we use a center-of-mass frame (cm frame) for the local frame, the internal energy change
during a process is related in a particularly simple way to the system energy change measured in a lab frame. A cm frame has its origin at the center of mass of the system and its
Cartesian axes parallel to the Cartesian axes of a lab frame. This is a special case of the
nonrotating local frame discussed in Sec. G.7. Since the center of mass may accelerate in
the lab frame, a cm frame is not necessarily inertial.
The indices i and j in this section refer only to the particles in the system.

G.8 C ENTER - OF - MASS L OCAL F RAME
501
ri0
ri
Rcm
Figure G.4 Position vectors in a lab frame and a center-of-mass frame. Open circle:
origin of lab frame; open triangle: center of mass; lled circle: particle i . The thin
lines represent the Cartesian axes of the two frames.
The center of mass of the system is a point whose position in the lab frame is dened
by
i mi r i
(G.8.1)
Rcm D
m
P
where m is the system mass: m D i mi . The position vector of particle i in the lab frame
is equal to the sum of the vector Rcm from the origin of the lab frame to the center of mass
and the vector ri0 from the center of mass to the particle (see Fig. G.4):
def
ri D Rcm C ri0
(G.8.2)
We can use Eqs. G.8.1 and G.8.2 to derive several relations that will be needed presently.
Because the Cartesian axes of the lab frame and cm frame are parallel to one another (that
is, the cm frame does not rotate), we can add vectors or form scalar products using the
vector components measured in either frame. The time derivative of Eq. G.8.2 is dri = dt D
dRcm = dt C dri0 = dt , or
0
vi D vcm C vi
(G.8.3)
where the vector vcm gives theP
velocity of the P
center of mass in the lab frame. Substitution
P
from Eq. G.8.2 into the sum Pi mi ri gives i mi ri D mRcm C i mi ri0 , and a rearrangement of Eq. G.8.1 gives i mi ri D mRcm . Comparing these two equalities, we see
P
P
the sum i mi ri0 must be zero. Therefore the rst and second derivatives of i mi ri0 with
respect to time must also be zero:
X
i
0
m i vi D 0
mi
0
dvi
D0
dt
(G.8.4)
From Eqs. G.1.2, G.6.1, G.6.2, and G.8.3 we obtain

Fiaccel D mi
0
d.vi vi /
D
dt
mi
dvcm
dt
(G.8.5)
Equation G.8.5 is valid only for a nonrotating cm frame.

G.8 C ENTER - OF - MASS L OCAL F RAME
502
The difference between the energy changes of the system in the cm frame and the lab
frame during a process is given, from Eqs. G.2.6 and G.6.5, by
"
#
X
X
0 2
2
1
1
U Esys D
2 mi .vi /
2 mi vi
i
eld
ield
iaccel
(G.8.6)
We will nd new expressions for the three terms on the right side of this equation.
The rst term is the difference between the total kinetic energy changes measured in
the cm frame and lab frame. We can derive an important relation, well known in classical
mechanics, for the kinetic energy in the lab frame:
X
X
2
0
0
1
1
mi vi D
2
2 mi .vcm C vi / .vcm C vi /
i
2
1
2 mvcm
1
2 mi
0 2
vi
C vcm
0
m i vi
(G.8.7)
2
The quantity 1 mvcm is the bulk kinetic energy of the system in the lab framethat is, the
2
translational energy of a P having the same mass as the system and moving with the
body
0
center of mass. The sum i mi vi is zero (Eq. G.8.4). Therefore the rst term on the right
side of Eq. G.8.6 is
"
#
X
X
0 2
2
2
1
1
1
D 2 mvcm
(G.8.8)
2 mi .vi /
2 mi vi
i
Only by using a nonrotating local frame moving with the center of mass is it possible to
derive such a simple relation among these kinetic energy quantities.
The second term on the right side of Eq. G.8.6, with the help of Eqs. G.2.2, G.6.3, and
G.8.2 becomes
!
X 0
X
XZ
eld
eld
i
i
D
Field d.ri0 ri /
i
Field
dRcm
(G.8.9)
i
grav
Suppose the only external eld is gravitational: Field D Fi

D mi gez where ez is a
unit vector in the vertical (upward) Cz direction. In this case we obtain
!
Z X !
X 0
X
i eld
ield D
mi gez dRcm
i
ez dRcm D
mg
mgzcm
mg
dzcm
(G.8.10)

G.9 ROTATING L OCAL F RAME
503
Figure G.5 Relation between the Cartesian axes x, y, z of a lab frame and the axes
x 0 , y 0 , z of a rotating local frame. The lled circle represents particle i .
where zcm is the elevation of the center of mass in the lab frame. The quantity mgzcm is
the change in the systems bulk gravitational potential energy in the lab framethe change
in the potential energy of a body of mass m undergoing the same change in elevation as the
systems center of mass.
The third term on the right side of Eq. G.8.6 can be shown to be zero when the local
frame is a cm frame. The derivation uses Eqs. G.6.4 and G.8.5 and is as follows:
!
X
XZ
XZ
dv
accel
accel
0
i
D
Fi
dri D
mi cm dri0
dt
i
i
i
!
!
Z X
Z X
dri0
0
dvcm D
m i vi
dvcm
(G.8.11)
D
mi
dt
i
0
The sum i mi vi in the integrand of the last integral on the right side is zero (Eq. G.8.4)
so the integral is also zero.
2
With these substitutions, Eq. G.8.6 becomes U Esys D 1 m vcm
mgzcm .
2
Since U Esys is equal to w wlab when the local frame is nonrotating (Eq. G.7.1), we
have
2
w wlab D 1 m vcm
mgzcm
(G.8.12)
2
G.9 ROTATING LOCAL FRAME

A rotating local frame is the most convenient to use in treating the thermodynamics of
a system with rotational motion in a lab frame. A good example of such a system is a
solution in a sample cell of a spinning ultracentrifuge (Sec. 9.8.2).
We will make several simplifying assumptions. The rotating local frame has the same
origin and the same z axis as the lab frame, as shown in Fig. G.5. The z axis is vertical
and is the axis of rotation for the local frame. The local frame rotates with constant angular
velocity ! D d#= dt, where # is the angle between the x axis of the lab frame and the

G.10 E ARTH -F IXED R EFERENCE F RAME
504
x 0 axis of the local frame. There is a gravitational force in the z direction; this force is
responsible for the only external eld, whose potential energy change in the local frame
0 grav
during a process is i
D mi g zi (Eq. G.2.3).
The contribution to the effective force acting on particle i due to acceleration when !
is constant can be shown to be given by8
Fiaccel D Ficentr C FiCor
(G.9.1)
Ficentr D mi ! 2 ri ei
(G.9.2)
where Ficentr is the so-called centrifugal force and FiCor is called the Coriolis force.
The centrifugal force acting on particle i is given by
Here ri is the radial distance of the particle from the axis of rotation, and ei is a unit vector
pointing from the particle in the direction away from the axis of rotation (see Fig. G.5). The
direction of ei in the local frame changes as the particle moves in this frame.
The Coriolis force acting on particle i arises only when the particle is moving relative
0
0
to the rotating frame, has magnitude 2mi !vi , and is directed perpendicular to both vi and
the axis of rotation.
In a rotating local frame, the work during a process is not the same as that measured in
a lab frame. The heats q and qlab are not equal to one another as they are when the local
frame is nonrotating, nor can general expressions using macroscopic quantities be written
for U Esys and w wlab .
G.10 EARTH-FIXED REFERENCE FRAME

In the preceding sections of Appendix G, we assumed that a lab frame whose coordinate
axes are xed relative to the earths surface is an inertial frame. This is not exactly true, because the earth spins about its axis and circles the sun. Small correction terms, a centrifugal
force and a Coriolis force, are needed to obtain the effective net force acting on particle i
that allows Newtons second law to be obeyed exactly in the lab frame.9
The earths movement around the sun makes only a minor contribution to these correction terms. The Coriolis force, which occurs only if the particle is moving in the lab frame,
is usually so small that it can be neglected.
This leaves as the only signicant correction the centrifugal force on the particle from
the earths spin about its axis. This force is directed perpendicular to the earths axis and has
magnitude mi ! 2 ri , where ! is the earths angular velocity, mi is the particles mass, and
ri is the radial distance of the particle from the earths axis. The correction can be treated
as a small modication of the gravitational force acting on the particle that is at most, at the
equator, only about 0.3% of the actual gravitational force. Not only is the correction small,
but it is completely taken into account in the lab frame when we calculate the effective
grav
gravitational force from Fi
D mi gez , where g is the acceleration of free fall and ez
is a unit vector in the Cz (upward) direction. The value of g is an experimental quantity
that includes the effect of Ficentr , and thus depends on latitude as well as elevation above the
grav
earths surface. Since Fi depends only on position, we can treat gravity as a conservative
force eld in the earth-xed lab frame.
8 The
9 Ref.
derivation, using a different notation, can be found in Ref. [107], Chap. 10.
[68], Sec. 49.
A PPENDIX H
S TANDARD M OLAR T HERMODYNAMIC

P ROPERTIES
The values in this table are for a temperature of 298:15 K (25:00 C) and the standard pressure p D 1 bar. Solute standard states are based on molality. A crystalline solid is denoted
by cr.
Most of the values in this table come from a project of the Committee on Data for Science and Technology (CODATA) to establish a set of recommended, internally consistent
values of thermodynamic properties. The values of f H and Sm shown with uncertainties

are values recommended by CODATA.1
Species
Inorganic substances
Ag(cr)
AgCl(cr)
C(cr, graphite)
CO(g)
CO2 (g)
Ca(cr)
CaCO3 (cr, calcite)
CaO(cr)
Cl2 (g)
F2 (g)
H2 (g)
HCl(g)
HF(g)
HI(g)
H2 O(l)
H2 O(g)
H2 S(g)
1 Ref.
f H
kJ mol 1
0
127.01 0:05
0
110.53 0:17
393.51 0:13
0
1206.9
634.92 0:90
0
0
0
92.31 0:10
273.30 0:70
26.50 0:10
285.830 0:040
241.826 0:040
20.6 0:5
Sm
J K 1 mol
42.55 0:20
96.25 0:20
5.74 0:10
197.660 0:004
213.785 0:010
41.59 0:40
92.9
38.1 0:4
223.081 0:010
202.791 0:005
130.680 0:003
186.902 0:005
173.779 0:003
206.590 0:004
69.95 0:03
188.835 0:010
205.81 0:05
[38]; also available online at http://www.codata.org.
505
f G
kJ mol 1
0
109.77
0
137.17
394.41
0
1128.8
603.31
0
0
0
95.30
275.40
1.70
237.16
228.58
33.44
506
A PPENDIX H S TANDARD M OLAR T HERMODYNAMIC P ROPERTIES
Species
Hg(l)
Hg(g)
HgO(cr, red)
Hg2 Cl2 (cr)
I2 (cr)
K(cr)
KI(cr)
KOH(cr)
N2 (g)
NH3 (g)
NO2 (g)
N2 O4 (g)
Na(cr)
NaCl(cr)
O2 (g)
O3 (g)
P(cr, white)
S(cr, rhombic)
SO2 (g)
Si(cr)
SiF4 (g)
SiO2 (cr, -quartz)
Zn(cr)
ZnO(cr)
Organic compounds
CH4 (g)
CH3 OH(l)
CH3 CH2 OH(l)
C2 H2 (g)
C2 H4 (g)
C2 H6 (g)
C3 H8 (g)
C6 H6 (l, benzene)
Ionic solutes
AgC (aq)
CO3 2 (aq)
Ca2C (aq)
Cl (aq)
F (aq)
f H
kJ mol 1
Sm
J K 1 mol
f G
kJ mol 1
0
61.38 0:04
90.79 0:12
265.37 0:40
0
0
327.90
424.72
0
45.94 0:35
33.10
9.08
0
411.12
0
142.67
0
0
296.81 0:20
0
1615.0 0:8
910.7 1:0
75.90 0:12
174.971 0:005
70.25 0:30
191.6 0:8
116.14 0:30
64.68 0:20
106.37
78.90
191.609 0:004
192.77 0:05
240.04
304.38
51.30 0:20
72.11
205.152 0:005
238.92
41.09 0:25
32.054 0:050
248.223 0:050
18.81 0:08
282.76 0:50
41.46 0:20
0
31.84
58.54
210.72
0
0
323.03
378.93
0
16.41
51.22
97.72
0
384.02
0
163.14
0
0
300.09
0
1572.8
856.3
350.46 0:27
43.65 0:40
320.48
74.87
238.9
277.0
226.73
52.47
83.85
104.7
49.04
105.79 0:08
675.23 0:25
543.0 1:0
167.08 0:10
335.35 0:65
41.63 0:15
186.25
127.2
159.9
200.93
219.32
229.6
270.31
173.26
73.45 0:40
50.0 1:0
56.2 1:0
56.60 0:20
13.8 0:8
50.77
166.6
173.8
209.21
68.43
32.00
24.3
124.54
77.10
527.90
552.8
131.22
281.52
507
A PPENDIX H S TANDARD M OLAR T HERMODYNAMIC P ROPERTIES
Species
HC (aq)
HCO3 (aq)
HS (aq)
HSO4 (aq)
Hg2 2C (aq)
I (aq)
KC (aq)
NH4 C (aq)
NO3 (aq)
NaC (aq)
OH (aq)
S2 (aq)
SO4 2 (aq)
f H
kJ mol 1
0
689.93 2:0
16.3 1:5
886.9 1:0
166.87 0:50
56.78 0:05
252.14 0:08
133.26 0:25
206.85 0:40
240.34 0:06
230.015 0:040
33.1
909.34 0:40
Sm
J K 1 mol
0
98.4 0:5
67 5
131.7 3:0
65.74 0:80
106.45 0:30
101.20 0:20
111.17 0:40
146.70 0:40
58.45 0:15
10.90 0:20
14.6
18.50 0:40
f G
kJ mol 1
0
586.90
12.2
755.4
153.57
51.72
282.52
79.40
110.84
261.90
157.24
86.0
744.00
A PPENDIX I
A NSWERS TO S ELECTED P ROBLEMS
3.3 (b) q D
105 J
w D 1:00
3.4 (c) w D 1:99
103 J, q D
1:99
103 J.
7.13 (a) Cp;m D 42:3 J K
3.5 0:0079%
(b) Cp;m
3.6 (c) V2 ! nRV1 =.CV C nR/, T2 ! 1.

For CV D .3=2/nR, V2 =V1 ! 0:4.
3.11 9:58
103 s (2 hr 40 min)
6.1 Sm
151:6 J K
7.8 (b) 8
7.11 5:001
10
7:28
10
8.5 19 J mol
10
mol
8.6 (a) 352:82 K

104 J mol
(b) 3:4154
103 Pa K
103 Pa K
(b) 3:56
bar
8.7 (a) 3:62
JK
m3 mol
8.4 4:5
(c) 99:60 C
1
8.8 (b) vap H D 4:084
Pa
104 J mol
8.9 0:93 mol
1:8 bar
7.7 (b) .@Cp;m =@p/T

D 4:210 10
mol
(b) vap S D 88:6 J K 1 mol 1 ,

vap H D 2:748 104 J mol
7.6 (a) D 8:519 10 K

t D 4:671 10 5 bar 1
.@p=@T /V D 18:24 bar K
.@U=@V /T D 5437 bar
(b) p
8.2 (a) Sm (l) D 253:6 J K
mol
D 0:739, f D 148 bar
(b) B D
5.5 (a) q D 0, w D 1:50 104 J,

U D 1:50 104 J,
H D 2:00 104 J
1
mol
7.15 (b) f D 17:4 bar
4.5 R D 549 J K 1 for both processes;

S
q=Text D 333 J K 1 and 0.
3=2 V
5.4 (a) S D nR ln cT
b
C
n

5
nR
2
(c) S D 66:7 J K
52:0 J K
7.14 (a) 2:56 J K
7.16 (a)
4.4 S D 0:054 J K
104 J, S D 43:6 J K
7.12 H D 2:27
mol
9.2 (b) VA .xB D 0:5/ 125:13 cm3 mol

VB .xB D 0:5/ 158:01 cm3 mol
1
VB
157:15 cm3 mol 1
9.4 real gas: p D 1:9743 bar

ideal gas: p D 1:9832 bar
10 J
508
1
1
509
A PPENDIX I A NSWERS TO S ELECTED P ROBLEMS
9.5 (a) xN2

xO2
yN2
yO2
D 8:83 10
D 4:65 10
D 0:763
D 0:205
(b) xN2
xO2
yN2
yO2
D 9:85 10
D 2:65 10
D 0:851
D 0:117
11.7 (a) State 1:

nC6 H14 D 7:822 10 3 mol
nH2 O D 0:05560 mol
amount of O2 consumed: 0:07431 mol
State 2:
nH2 O D 0:11035 mol
nCO2 D 0:04693 mol
mass of H2 O=1:9880 g
9.7 (b) fA D 0:03167 bar, fA D 0:03040 bar

9.8 (a) In the mixture of composition
xA D 0:9782, the activity coefcient is
11:5.
B
9.9 (d) kH,A
680 kPa
9.11 Values for mB =m D 20: A D 1:026,

m;B D 0:526; the limiting
slopes are d A = d.mB =m / D 0,
d m;B = d.mB =m / D 0:09
p.6:95 cm/ D 1:2 bar
(b) MB D 187 kg mol 1

mass binding ratio D 1:37
(f) fH2 O .0:03169 bar/ D 0:03164 bar

State 1: fH2 O D 0:03234 bar
State 2: fH2 O D 0:03229 bar
D 0:392
11.1 r H D 63:94 kJ mol

K D 4:41 10 2
11.2 (b) f H : no change

f S : subtract 0:219 J K
f G : add 65 J mol 1
11.3 p.298:15 K/ D 2:6 10
p.273:15 K/ D 2:7 10
1
240:34 kJ mol
230:02 kJ mol
(b)
465:43 kJ mol
(c)
39:82 kJ mol
11.4 (a)
6
7
mol
bar
bar
, 470:36 kJ mol
11.5 H D 0:92 kJ
11.6 LA D 0:405 J mol 1
LB D 0:810 kJ mol 1
(c) State 1: V (C6 H14 ) D 1:029 cm3

V (H2 O) D 1:005 cm3
V g D 348:0 cm3
State 2:
V (H2 O) D 1:994 cm3
V g D 348:0 cm3
(e) State 2:
p2 D 27:9 bar
pO2 D 24:6 bar
pCO2 D 3:26 bar
9.14 (b) h D 1:2 m
10.2
(d) State 1:
nO2 D 0:429 mol
State 2:
nO2 D 0:355 mol
yO2 D 0:883
yCO2 D 0:117
9.13 pN2 D 0:235 bar

yN2 D 0:815
pO2 D 0:0532 bar
yO2 D 0:185
p D 0:288 bar
9.15 (a) p.7:20 cm/
(b) Vm (C6 H14 ) D 131:61 cm3 mol

Vm (H2 O) D 18:070 cm3 mol 1
(g) State 1:
H2 O D 0:925
O2 D 0:981
fO2 D 29:4 bar
State 2:
H2 O D 0:896
O2 D 0:983
CO2 D 0:910
fO2 D 24:2 bar
fCO2 D 2:97 bar
(h) State 1:
g
nH2 O D 5:00 10 4 mol
nlH2 O D 0:05510 mol
g
State 2: nH2 O D 5:19 10
nlH2 O D 0:10983 mol
mol
510
(i) State 1:
km;O2 D 825 bar kg mol 1
nO2 D 3:57 10 5 mol
State 2:
km;O2 D 823 bar kg mol 1
km;CO2 D 30:8 bar kg mol
nO2 D 5:85 10 5 mol
nCO2 D 1:92 10 4 mol
12.6 sol,B H =kJ mol 1 D 3:06; 0; 6:35

sol,B S =J K 1 mol 1
D 121:0; 110:2; 88:4
1
(j) H2 O vaporization: U D C20:8 J

H2 O condensation: U D 21:6 J
(k) O2 dissolution: U D 0:35 J
O2 desolution: U D 0:57 J
CO2 desolution: U D 3:32 J
(l) C6 H14 (l) compression:
U D 1:226 J
solution compression:
U D 0:225 J
solution decompression:
U D 0:414 J
12.7 (a) m D m D 1:20 10 3 mol kg

C
m D 1:80 10 3 mol kg 1
C
m D 0:80 10 3 mol kg 1
mP D 2:00 10 6 mol kg 1
12.8 (a) p l D 2:44 bar

(b) f .2:44 bar/ f .1:00 bar/
D 3:4 10 5 bar
12.10 (a) xB D 1:8 10 7

mB D 1:0 10 5 mol kg
(b) sol,B H D
104 J mol
1:99
(c) K D 4:4 10 7
r H D 9:3 kJ mol
12.13 (a) p D 92399:6 Pa, yB D 0:965724

(b)
D 0:995801
(c) fA D 3164:47 Pa
(m) O2 compression: U D 81 J
gas mixture: dB= dT D
0:26 10 6 m3 K 1 mol 1
gas mixture expansion: U D 87 J
(d) yB D 0:965608
(n) U D 8 J
(f) p D 92347:7 Pa
(o) c U D
4154:4 kJ mol
(p) c H D
4163:1 kJ mol
11.8 f H D
198:8 kJ mol
10
11.11 (a) K D 3:5
(b) pH2 D 2:8 10

NH2 D 6:9 10
bar
42
10
10
mol kg
12.18 f H .Cl , aq/ D 167:15 kJ mol
Sm .Cl , aq/ D 56:46 J K 1 mol 1
bar
17
12.19 (a) Ks D 1:783

12.20 (a) r H D
104 bar
10
10
65:769 kJ mol
(b) f H (AgC , aq) D 105:84 kJ mol
11.13 (c) K D 0:15

12.1 (b) T D 1168 K
r H D 1:64
13.1 (a) F D 4
105 J mol
12.4 Kf D 1:860 K kg mol 1

Kb D 0:5118 K kg mol 1
5:6
D 0:9826
(b) D 0:140
(c) t D 22 s
12.5 MB
x;B
12.17 (a) D 0:129

mC D 1:29 10
1041
1:5
12.15 (a)
12.16 K D 1:2
11.10 p(O2 ) D 2:55
109 Pa
(g) kH,B D 4:40890

(b) xB D 4:19
11.9 T2 D 2272 K
11.12 (b) p
(e) Z D 0:999319
10 g mol
(b) F D 3
(c) F D 2
13.10 (a) xB (top) D 0:02, xB (bottom) D 0:31
(b) nA D 2:1 mol, nB D 1:0 mol
511
14.3 (a) r G D 21:436 kJ mol 1

r S D 62:35 J K 1 mol
r H D 40:03 kJ mol 1
(b) f H .AgCl, s/ D
(c)
14.7 (b) f G D
127:05 kJ mol
(c) Ks D 1:4
1
Sm .AgCl, s/ D 96:16 J K 1 mol 1

f S .AgCl, s/ D 57:93 J K 1 mol 1
f G .AgCl, s/ D 109:78 kJ mol 1
14.4 (b) f H .AgCl, s/ D 126:81 kJ mol 1

f G .AgCl, s/ D 109:59 kJ mol 1
14.5 Ks D 1:76
14.6 (b)
10
D 0:756
10
210:72 kJ mol
10
18
14.8 E D 0:071 V
14.9 (c) Ecell, eq D 1:36 V
(d) In the cell:

q= d D 2:27 kJ mol 1
In a reaction vessel:
q= d D 259:67 kJ mol
(e) dEcell, eq = dT D 3:9
10
VK
B IBLIOGRAPHY
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
C. J. Adkins, Equilibrium Thermodynamics, 3rd edition. Cambridge University Press,

Cambridge, 1983.
Robert A. Alberty, Use of Legendre Transforms in Chemical Thermodynamics (IUPAC
Technical Report). Pure Appl. Chem., 73, 13491380 (2001).
S. Angus, B. Armstrong, and K. M. de Reuck, International Thermodynamic Tables of the
Fluid State. Vol. 3. Carbon Dioxide. Pergamon Press, Oxford, 1976.
Roger G. Bates and Vincent E. Bower, Standard Potential of the SilverSilver-Chloride
Electrode from 0 to 95 C and the Thermodynamic Properties of Dilute Hydrochloric Acid
Solutions. J. Res. Natl. Bur. Stand. (U.S.), 53, 283290 (1954).
Stuart J. Bates and H. Darwin Kirschman, The Vapor Pressures and Free Energies of the
Hydrogen Halides in Aqueous Solution; the Free Energy of Formation of Hydrogen
Chloride. J. Am. Chem. Soc., 41, 19912001 (1919).
Robert P. Bauman, Work of Compressing an Ideal Gas. J. Chem. Educ., 41, 102104
(1964).
Robert P. Bauman, Maximum Work Revisited (letter). J. Chem. Educ., 41, 675 (1964).
Robert P. Bauman, Maximum Work Revisited (letter). J. Chem. Educ., 41, 676677
(1964).
Robert P. Bauman and Howard L. Cockerham III, Pressure of an Ideal Gas on a Moving
Piston. Am. J. Phys., 37, 675679 (1969).
Hassan Behnejad, Jan V. Sengers, and Mikhail A. Anisimov, Thermodynamic Behaviour of
Fluids near Critical Points. In A. R. H. Goodwin, J. V. Sengers, and C. J. Peters, editors,
Applied Thermodynamics of Fluids, pages 321367. Royal Society of Chemistry,
Cambridge, 2010.
Dor Ben-Amotz and J. M. Honig, Rectication of Thermodynamic Inequalities. J. Chem.
Phys., 118, 59325936 (2003).
George C. Benson and Osamu Kiyohara, Thermodynamics of Aqueous Mixtures of
Nonelectrolytes. I. Excess Volumes of Water - n-Alcohol Mixtures at Several Temperatures.
J. Solution Chem., 9, 791804 (1980).
K. H. Berry, NPL-75: A Low Temperature Gas Thermometry Scale from 2.6 K to 27.1 K.
Metrologia, 15, 89115 (1979).
Gary L. Bertrand, Thermodynamic Calculation of Work for Some Irreversible Processes.
J. Chem. Educ., 82, 874877 (2005).
R. W. Blue and W. F. Giauque, The Heat Capacity and Vapor Pressure of Solid and Liquid
Nitrous Oxide. The Entropy from its Band Spectrum. J. Am. Chem. Soc., 57, 991997
(1935).
M. Boirac, N crologie: Francois Raoult, Discours de M. Boirac, Recteur. Ann.
e
lUniversit de Grenoble, 13, iv (1901).

e
J. T. Bottomley, Scientic Worthies. XX.James Prescott Joule. Nature, 26, 617620
512
B IBLIOGRAPHY
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
513
(1882).
Pascale Boulanger, Marc le Maire, M lanie Bonhivers, Sandrine Dubois, Michel Desmadril,
e
and Lucienne Letellier, Purication and Structural and Functional Characterization of
FhuA, a Transporter of the Escherichia coli Outer Membrane. Biochemistry, 14,
1421614224 (1996).
Vincenzo Brandani, Angelo Chianese, and Maurizio Rossi, Ternary Liquid-Liquid
Equilibrium Data for the Water-Ethanol-Benzene System. J. Chem. Eng. Data, 30, 2729
(1985).
L. K. Brice, Le Chateliers Principle: The Effect of Temperature on the Solubility of Solids
in Liquids. J. Chem. Educ., 60, 387389 (1983).
P. W. Bridgman, The Thermodynamics of Electrical Phenomena in Metals and a Condensed
Collection of Thermodynamic Formulas. Dover, New York, 1961.
P. W. Bridgman, A Complete Collection of Thermodynamic Formulas. Phys. Rev., 3,
273281 (1914).
P. W. Bridgman, The Nature of Thermodynamics. Harvard University Press, Cambridge,
Massachusetts, 1941.
I. Brown, W. Fock, and F. Smith, Thermodynamic Properties of Alcohol Solutions. Aust. J.
Chem., 9, 364372 (1956).
Oliver L. I. Brown and George G. Manov, The Heat Capacity of Carbon Disulde from 15
to 300 K. The Entropy and Heat of Fusion of Carbon Disulde. J. Am. Chem. Soc., 59,
500502 (1937).
H. L. Callendar, On a Practical Thermometric Standard. Phil. Mag., 48, 519547 (1899).
Sadi Carnot, Reections on the Motive Power of Fire, and on Machines Fitted to Develop
that Power. In E. Mendoza, editor, Reections on the Motive Power of Fire: Sadi Carnot
and other Papers on the Second Law of Thermodynamics by E. Clapeyron and R. Clausius,
pages 159. Dover, Mineola, New York, 1988. Translation by R. H. Thurston of R exions
e
sur la Puissance Motrice du Feu et sur les Machines Propres a D velopper cette Puissance,
` e
Bachelier, Paris, 1824.
Malcolm W. Chase Jr., NIST-JANAF Thermochemical Tables, Fourth edition. American
Institute of Physics, Woodbury, New York, 1998. Appeared as J. Phys. Chem. Ref. Data,
Monograph 9.
John P. Chesick, Maximum Work Revisited (letter). J. Chem. Educ., 41, 674675 (1964).
E. Clapeyron, Memoir on the Motive Power of Heat. In E. Mendoza, editor, Reections on

the Motive Power of Fire: Sadi Carnot and other Papers on the Second Law of
Thermodynamics by E. Clapeyron and R. Clausius, pages 71105. Dover, Mineola, New

York, 1988. Translation by E. Mendoza of M moire sur la Puissance Motrice de la
e
Chaleur, J. de lEcole Polytechnique, 14, 153190 (1834).

E. Colin W. Clarke and David N. Glew, Evaluation of Thermodynamic Functions for
Aqueous Sodium Chloride from Equilibrium and Calorimetric Measurements below 154 C.
J. Phys. Chem. Ref. Data, 14, 489610 (1985).
R. Clausius, Fourth Memoir. On a Modied Form of the Second Fundamental Theorem in
the Mechanical Theory of Heat. In T. Archer Hirst, editor, The Mechanical Theory of Heat,
with its Applications to the Steam-Engine and to the Physical Properties of Bodies, pages
11135. John Van Voorst, London, 1867. Translation of Ueber eine Ver nderte Form des
a
Zweiten Hauptsatzes der Mechanischen W rmetheorie, Ann. Phys. Chem. (Leipzig), 93,
a
481506 (1854).
R. Clausius, Ninth Memoir. On Several Convenient Forms of the Fundamental Equations of
the Mechanical Theory of Heat. In T. Archer Hirst, editor, The Mechanical Theory of Heat,
with its Applications to the Steam-Engine and to the Physical Properties of Bodies, pages
327365. John Van Voorst, London, 1867. Translation of Ueber Verschiedene f r die
u
Anwendung Bequeme Formen der Hauptgleichungen der Mechanischen W rmetheorie,
a
Ann. Phys. Chem. (Leipzig), 125, 353400 (1865).
B IBLIOGRAPHY
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
514
J. O. Clayton and W. F. Giauque, The Heat Capacity and Entropy of Carbon Monoxide.
Heat of Vaporization. Vapor Pressures of Solid and Liquid. Free Energy to 5000 K from
Spectroscopic Data. J. Am. Chem. Soc., 54, 26102626 (1932).
Patrick Coffey, Cathedrals of Science: The Personalities and Rivalries That Made Modern
Chemistry. Oxford University Press, New York, 2008.
E. Richard Cohen, Tomislav Cvita, Jeremy G. Frey, Bertil Holmstr m, Kozo Kuchitsu,
s
o
Roberto Marquardt, Ian Mills, Franco Pavese, Martin Quack, J rgen Stohner, Herbert L.
u
Strauss, Michio Takami, and Anders J Thor, Quantities, Units and Symbols in Physical
Chemistry, 3rd edition. RSC Publishing, Cambridge, 2007.
Roger Cohen-Adad and John W. Lorimer, Alkali Metal and Ammonium Chlorides in Water
and Heavy Water (Binary Systems). Solubility Data Series, Vol. 47. Pergamon Press,
Oxford, 1991.
J. D. Cox, D. D. Wagman, and V. A. Medvedev, CODATA Key Values for Thermodynamics.
Hemisphere Publishing Corp., New York, 1989.
William H. Cropper, Walther Nernst and the Last Law. J. Chem. Educ., 64, 38 (1987).
Rosa Crovetto, Evaluation of Solubility Data of the System CO2 -H2 O from 273 K to the
Critical Point of Water. J. Phys. Chem. Ref. Data, 20, 575589 (1991).
Thomas W. Davis, A Common Misunderstanding of Hesss Law. J. Chem. Educ., 28,
584585 (1951).
J. de Swann Arons and G. A. M. Diepen, Gas-Gas Equilibria. J. Chem. Phys., 44,
23222330 (1966).
Peter Debye. Interviewed by C. Uhlenbeck on 3 May, 1962, Niels Bohr Library & Archives,
American Institute of Physics, College Park, MD USA, online at
http://www.aip.org/history/ohilist/4568_2.html.
Howard DeVoe, Particle Model for Work, Heat, and the Energy of a Thermodynamic
System. J. Chem. Educ., 84, 504512 (2007).
Hilyard John Eglinton Dobson, The Partial Pressures of Aqueous Ethyl Alcohol. J. Chem.
Soc., 127, 28662873 (1925).
Frederick G. Donnan, Prof. J. H. vant Hoff. Nature, 86, 8486 (1911).
J. H. Dymond and E. B. Smith, The Virial Coefcients of Pure Gases and Mixtures, A
Critical Compilation. Clarendon Press, Oxford, 1980.
A. E. Edwards and W. E. Roseveare, The Second Virial Coefcients of Gaseous Mixtures.
J. Am. Chem. Soc., 64, 28162819 (1942).
D. Eisenberg and W. Kauzmann, The Structure and Properties of Water. Oxford University
Press, New York, 1969.
Rodney P. Elliot, Constitution of Binary Alloys, First Supplement. McGraw-Hill, New York,
1965.
C. W. F. Everitt, Maxwells Scientic Papers. Applied Optics, 6, 639646 (1967).
M. B. Ewing, T. H. Lilley, G. M. Olofsson, M. T. R tzsch, and G. Somsen, Standard
a
Quantities in Chemical Thermodynamics. Fugacities, Activities, and Equilibrium Constants
for Pure and Mixed Phases (IUPAC Recommendations 1994). Pure Appl. Chem., 66,
533552 (1994).
G. Faita, P. Longhi, and T. Mussini, Standard Potentials of the Cl2 /Cl Electrode at Various
Temperatures with Related Thermodynamic Functions. J. Electrochem. Soc., 114, 340343
(1967).
P. Ferloni and G. Spinolo, Int. DATA Ser., Sel. Data Mixtures, Ser. A, 70 (1974).
Alexander Findlay and A. N. Campbell, The Phase Rule and Its Applications, 8th edition.
Dover, New York, 1945.
Kai Fischer and Jurgen Gmehling, P-x and 1 Data for the Different Binary Butanol-Water
Systems at 50 C. J. Chem. Eng. Data, 39, 309315 (1994).
George Fleck, Peter Debye. In Laylin K. James, editor, Nobel Laureates in Chemistry,
19011992, pages 228235. Chemical Heritage Foundation, 1993.
B IBLIOGRAPHY
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
515
W. F. Giauque and D. P. MacDougall, Attainment of Temperatures Below 1 Absolute by

Demagnetization of Gd2 .SO4 /3 8H2 O. Phys. Rev., 43, 768 (1933).
W. F. Giauque and J. W. Stout, The Entropy of Water and the Third Law of
Thermodynamics. The Heat Capacity of Ice from 15 to 273 K. J. Am. Chem. Soc., 58,
11441150 (1936).
W. F. Giauque and R. Wiebe, The Entropy of Hydrogen Chloride. Heat Capacity from
16 K. to Boiling Point. Heat of Vaporization. Vapor Pressures of Solid and Liquid. J. Am.
Chem. Soc., 50, 101122 (1928).
William F. Giauque. Some Consequences of Low Temperature Research in Chemical
Thermodynamics: Nobel Lecture, December 12, 1949. Online at http://nobelprize.
org/nobel_prizes/chemistry/laureates/1949/giauque-lecture.pdf.
J. Willard Gibbs, Graphical Methods in the Thermodynamics of Fluids. In Henry Andrews
Bumstead and Ralph Gibbs Van Name, editors, The Scientic Papers of J. Willard Gibbs,
Ph.D., LL.D., Vol. I, Thermodynamics, pages 132. Ox Bow Press, Woodbridge,
Connecticut, 1993. Reprint of book originally published by Longmans, Green in 1906.
J. Willard Gibbs, A Method of Geometrical Representation of the Thermodynamic
Properties of Substances by Means of Surfaces. In Henry Andrews Bumstead and
Ralph Gibbs Van Name, editors, The Scientic Papers of J. Willard Gibbs, Ph.D., LL.D., Vol.
I, Thermodynamics, pages 3354. Ox Bow Press, Woodbridge, Connecticut, 1993. Reprint
of book originally published by Longmans, Green in 1906.
J. Willard Gibbs, On the Equilibrium of Heterogeneous Substances. In Henry Andrews
Bumstead and Ralph Gibbs Van Name, editors, The Scientic Papers of J. Willard Gibbs,
Ph.D., LL.D., Vol. I, Thermodynamics, pages 55353. Ox Bow Press, Woodbridge,
Connecticut, 1993. Reprint of book originally published by Longmans, Green in 1906.
R. E. Gibson and O. H. Loefer, Pressure-Volume-Temperature Relations in Solutions. II.
The Energy-Volume Coefcients of Aniline, Nitrobenzene, Bromobenzene and
Chlorobenzene. J. Am. Chem. Soc., 61, 25152522 (1939).
R. T. Glazebrook, Joule, James Prescott. In Sidney Lee, editor, Dictionary of National
Biography, Vol. 30, pages 208214. Macmillan, New York, 1892.
J. A. Gledhill and G. McP. Malan, The Solubilities of Sparingly Soluble Salts in Water.
Trans. Faraday Soc., 48, 258262 (1952).
Herbert Goldstein, Classical Mechanics. Addison-Wesley, Cambridge, Massachusetts, 1950.
Sandra C. Greer, M. R. Moldover, and R. Hocken, A Differential Transformer as a Position
Detector in a Magnetic Densimeter. Rev. Sci. Instrum., 45, 14621463 (1974).
Ulrich Grigull, Johannes Straub, and Peter Schiebener, Steam Tables in SI-Units.
Springer-Verlag, Berlin, 1990.
E. A. Guggenheim, Thermodynamics: An Advanced Treatment for Chemists and Physicists,
8th edition. Elsevier Science, 1985.
Stuart R. Gunn and Leroy G. Green, Heat of Solution of Hydrogen Chloride. J. Chem.
Eng. Data, 8, 180 (1963).
Herbert S. Harned and Russell W. Ehlers, The Dissociation Constant of Acetic Acid from 0
to 35 Centigrade. J. Am. Chem. Soc., 54, 13501357 (1932).
Herbert S. Harned and Benton B. Owen, The Physical Chemistry of Electrolytic Solutions,
3rd edition. Reinhold, New York, 1958.
William Henry, Experiments on the Quantity of Gases Absorbed by Water, at Different
Temperatures, and under Different Pressures. Philos. Trans. R. Soc. London, 93, 2942
(1803).
William Henry, Appendix to Mr. William Henrys Paper, on the Quantity of Gases
Absorbed by Water, at Different Temperatures, and under Different Pressures. Philos.
Trans. R. Soc. London, 93, 274276 (1803).
H. Hess, Thermochemische Untersuchungen. Ann. Phys. Chem. (Leipzig), 50, 385404
(1840).
B IBLIOGRAPHY
516
[78]
Joel H. Hildebrand and Robert L. Scott, The Solubility of Nonelectrolytes, 3rd edition.
Dover, New York, 1964. Corrected republication of the 3rd (1950) edition.
[79] Joel H. Hildebrand and Robert L. Scott, Regular Solutions. Prentice-Hall, Englewood Cliffs,
New Jersey, 1962.
[80] Joseph Hirsch, Clapeyron. In Ecole Polytechnique: Livre du Centenaire, 17941894,

Vol. I, pages 194198. Gauthier-Villars et ls, Paris, 1895. Online at
http://www.annales.org/archives/x/clapeyron.html.
[81] J. M. Honig and Dor Ben-Amotz, The Analysis of Spontaneous Processes Using
Equilibrium Thermodynamics. J. Chem. Educ., 83, 132137 (2006).
[82] Irmgard K. Howard, S Is for Entropy. U is for Energy. What Was Clausius Thinking? J.
Chem. Educ., 78, 505508 (2001).
[83] H. L. Johnston and W. F. Giauque, The Heat Capacity of Nitric Oxide from 14 K. to the
Boiling Point and the Heat of Vaporization. Vapor Pressures of Solid and Liquid Phases. The
Entropy from Spectroscopic Data. J. Am. Chem. Soc., 51, 31943214 (1929).
[84] James Prescott Joule, On the Mechanical Equivalent of Heat. Philos. Trans. R. Soc.
London, 140, 6182 (1850).
[85] Michael Kasha, The Triplet State. An Example of G. N. Lewis Research Style. J. Chem.
Educ., 61, 204215 (1984).
[86] W. B. Kay, Liquid-Vapor Phase Equilibria Relations in the Ethanen-Heptane System. Ind.
Eng. Chem., 30, 459465 (1938).
[87] Joseph H. Keenan, Frederick G. Keyes, Philip G. Hill, and Joan G. Moore, Steam Tables.
Thermodynamic Properties of Water Including Vapor, Liquid, and Solid Phases. Krieger
Publishing Company, Malabar, Florida, 1992.
[88] George S. Kell, Density, Thermal Expansivity, and Compressibility of Liquid Water from
0 to 150 C: Correlations and Tables for Atmospheric Pressure and Saturation Reviewed
and Expressed on 1968 Temperature Scale. J. Chem. Eng. Data, 20, 97105 (1975).
[89] Lord Kelvin, On an Absolute Thermometric Scale Founded on Carnots Theory of the
Motive Power of Heat, and Calculated from Regnaults Observations. Philos. Mag., 33,
313317 (1848). Online at http://zapatopi.net/kelvin/papers/.
[90] Milton Kerker, Sadi Carnot and the Steam Engine Engineers. Isis, 51, 257270 (1960).
[91] Albert S. Keston, A SilverSilver Bromide Electrode Suitable for Measurements in Very
Dilute Solutions. J. Am. Chem. Soc., 57, 16711673 (1935).
[92] Daniel Kivelson and Irwin Oppenheim, Work in Irreversible Expansions. J. Chem. Educ.,
43, 233235 (1966).
[93] Angela K hler-Kr tzfeld and Klaus Christmann. Hermann Heinrich Hess, online at
o
u
http://www.chemiedidaktik.uni-oldenburg.de/download/Hermann_Heinrich_
Hess.pdf (accessed July 2011).
[94] Richard J. Kokes, Maximum Work Revisited (letter). J. Chem. Educ., 41, 675676 (1964).
[95] Frank L. Lambert, Entropy Is Simple, Qualitatively. J. Chem. Educ., 79, 12411246
(2002).
[96] Frank L. Lambert and Harvey S. Leff, The Correlation of Standard Entropy with Enthalpy
Supplied from 0 to 298.15 K. J. Chem. Educ., 86, 9498 (2009).
[97] Peter T. Landsberg, Thermodynamics and Statistical Mechanics. Dover Publications, Inc.,
New York, 1990.
[98] Harvey S. Leff, Multisystem Temperature Equilibration and the Second Law. Am. J. Phys.,
45, 252254 (1977).
[99] Harvey S. Leff, Thermodynamic Entropy: The Spreading and Sharing of Energy. Am. J.
Phys., 64, 12611271 (1996).
[100] Henry M. Leicester, Germain Henri Hess and the Foundation of Thermochemistry. J.
Chem. Educ., 28, 581583 (1951).
[101] Gilbert N. Lewis and Merle Randall, The Activity Coefcient of Strong Electrolytes. J.
Am. Chem. Soc., 11121154 (1921).
B IBLIOGRAPHY
517
[102] Gilbert Newton Lewis, A New Conception of Thermal Pressure and a Theory of Solutions.
Proc. Am. Acad. Arts and Sciences, 36, 145168 (1900).
[103] Gilbert Newton Lewis and Merle Randall, Thermodynamics and the Free Energy of
Chemical Substances. McGraw-Hill, New York, 1923.
[104] Gilbert Newton Lewis and Merle Randall, Thermodynamics, 2nd edition. McGraw-Hill,
New York, 1961. Revised by Kenneth S. Pitzer and Leo Brewer.
[105] David R. Lide (editor), CRC Handbook of Chemistry and Physics, 73rd edition. CRC Press,
Boca Raton, Florida, 1992.
[106] Thomas S. Logan, Textbook Errors: Guest Column. XVI: The Vapor Pressure of Hydrated
Cupric Sulfate. J. Chem. Educ., 35, 148149 (1958).
[107] Jerry B. Marion and Stephen T. Thornton, Classical Dynamics of Particles and Systems, 4th
edition. Harcourt Brace, New York, 1995.
[108] K. N. Marsh, Int. DATA Ser., Sel. Data Mixtures, Ser. A, 2, 73 (1974).
[109] T. A. Mary, J. S. O. Evans, T. Vogt, and A. W. Sleight, Negative Thermal Expansion from
0.3 to 1050 Kelvin in ZrW2 O8 . Science, 272, 9092 (1996).
[110] James Clerk Maxwell, Theory of Heat, 9th edition. Longmans, Green and Co., London,
1888.
[111] M. L. McGlashan, Deviations from Raoults Law. J. Chem. Educ., 40, 516518 (1963).
[112] M. L. McGlashan, Chemical Thermodynamics. Academic Press, London, 1979.
[113] M. L. McGlashan, The International Temperature Scale of 1990 (ITS-90). J. Chem.
Thermodyn., 22, 653663 (1990).
[114] E. Mendoza, Reections on the Motive Power of Fire: Sadi Carnot and other Papers on the
Second Law of Thermodynamics. Dover, Mineola, New York, 1988.
[115] A. Michels, B. Blaisse, and C. Michels, The Isotherms of CO2 in the Neighbourhood of the
Critical Point and Round the Coexistence Line. Proc. R. Soc. London, Ser. A, 160, 358375
(1937).
[116] Robert A. Millikan, Walther Nernst, a Great Physicist, Passes. Scientic Monthly, 54,
8486 (1942). Online at http://www.nernst.de/nernst_millikan1942.htm.
[117] I. M. Mills, The Choice of Names and Symbols for Quantities in Chemistry. J. Chem.
Educ., 66, 887889 (1989).
[118] Ian Mills, Tomislav Cvita, Klaus Homann, Nikola Kallay, and Kozo Kuchitsu, Quantities,
s
Units and Symbols in Physical Chemistry. Blackwell, Oxford, 1988.
[119] Ian Mills, Tomislav Cvita, Klaus Homann, Nikola Kallay, and Kozo Kuchitsu, Quantities,
s
Units and Symbols in Physical Chemistry, 2nd edition. Blackwell, Oxford, 1993. Online at
http://www.iupac.org/publications/books/gbook/index.html.
[120] Karol J. Mysels, Maximum Work Revisited (letter). J. Chem. Educ., 41, 677 (1964).
[121] R. Negishi, The Solubility of Solid Benzene in Several Non-Polar Liquids. Rev. Phys.
Chem. Japan, 15, 98116 (1941).
[122] Nobelprize.org, Jacobus H. vant Hoff - Biography. 1901. Online at http:
//nobelprize.org/nobel_prizes/chemistry/laureates/1901/hoff-bio.html.
[123] National Institute of Standards and Technology, NIST Chemistry WebBook. Online at
http://webbook.nist.gov/chemistry/.
[124] John W. Owens, Stanley P. Wasik, and Howard DeVoe, Aqueous Solubilities and Enthalpies
of Solution of n-Alkylbenzenes. J. Chem. Eng. Data, 31, 4751 (1986).
[125] Vivian B. Parker, Thermal Properties of Aqueous Uni-Univalent Electrolytes. National
Standard Data Reference Series. National Bureau of Standards (U.S.), 1965.
[126] M. W. Pestak, Raymond E. Goldstein, M. H. W. Chan, J. R. de Bruyn, D. A. Balzarini, and
N. W. Ashcroft, Three-body Interactions, Scaling Variables, and Singular Diameters in the
Coexistence Curves of Fluids. Phys. Rev. B, 36, 599614 (1987).
[127] J. C. Philip, Freezing-Point Curves for some Binary Mixtures of Organic Substances,
Chiey Phenols and Amines. J. Chem. Soc., 83, 814834 (1903).
[128] Carl W. F. T. Pistorius, Martha C. Pistorius, J. P. Blakey, and L. J. Admiraal, Melting Curve
B IBLIOGRAPHY
518
of Ice VII to 200 kbar. J. Chem. Phys., 38, 600602 (1963).

[129] Kenneth S. Pitzer, Gilbert N. Lewis and the Thermodynamics of Strong Electrolytes. J.
Chem. Educ., 61, 104107 (1984).
[130] Kenneth S. Pitzer and David A. Shirley, William Francis Giauque. In Biographical
Memoirs, Vol. 69. National Academies Press, Washington, 1996. Online at
http://www.nasonline.org/site/PageServer?pagename=MEMOIRS_A.
[131] Max Planck, Treatise on Thermodynamics, 3rd edition. Dover, Mineola, New York, 1990.
Translation by Alexander Ogg of Vorlesungen uber Thermodynamik, de Gruyter, Berlin,
1922.
[132] Max Planck, A Scientic Autobiography. In Scientic Autobiography and Other Papers.
Williams and Norgate, London, 1950. Translation by Frank Gaynor of Wissenschaftliche
Selbstbiographie, Barth, Leipzig, 1948.
[133] R. F. Platford, Oil-Water Partition Coefcients from Solvent Activities. J. Solution Chem.,
5, 645651 (1976).
[134] H. Preston-Thomas, The International Temperature Scale of 1990 (ITS-90). Metrologia,
27, 310 (1990).
[135] William Ramsey, Prof. Francois Marie Raoult. Nature, 64, 1718 (1901).
[136] Merle Randall and Leona Esther Young, The Calomel and Silver Chloride Electrodes in
Acid and Neutral Solutions. The Activity Coefcient of Aqueous Hydrochloric Acid and the
Single Potential of the Deci-Molal Calomel Electrode. J. Am. Chem. Soc., 50, 9891004
(1928).
[137] F. M. Raoult, On the Vapor-Pressure of Ethereal Solutions. In Harry C. Jones, editor, The
Modern Theory of Solution: Memoirs by Pfeffer, vant Hoff, Arrhenius, and Raoult, pages
95121. Harper and Brothers, New York, 1899. Translation by Harry C. Jones of Sur les
Tensions de Vapeur des Dissolutions Faites dans lEther, Ann. Chim. Phys., 6th Ser., 15,
375407 (1888).
[138] F.-M. Raoult, Sur les Tensions de Vapeur des Dissolutions. Ann. Chim. Phys., 6th Ser., 20,
297371 (1890).
[139] Joseph A. Rard and Donald G. Miller, Isopiestic Determination of the Osmotic Coefcients
of Aqueous Na2 SO4 , MgSO4 , and Na2 SO4 MgSO4 at 25 C. J. Chem. Eng. Data, 26,
3338 (1981).
[140] Otto Redlich, The So-Called Zeroth Law of Thermodynamics. J. Chem. Educ., 47,
740741 (1970).
[141] T. R. Rettich, Rubin Battino, and Emmerich Wilhelm, Solubility of Gases in Liquids. 22.
High-Precision Determination of Henrys Law Constants of Oxygen in Liquid Water from
T D 274 K to T D 328 K. J. Chem. Thermodyn., 32, 11451156 (2000).
[142] Timothy R. Rettich, Y. Paul Handa, Rubin Battino, and Emmerich Wilhelm, Solubility of
Gases in Liquids. 13. High-Precision Determination of Henrys Constants for Methane and
Ethane in Liquid Water at 275 to 328 K. J. Phys. Chem., 85, 32303237 (1981).
[143] R. A. Robinson and R. H. Stokes, Electrolyte Solutions, 2nd edition. Butterworths, London,
1959.
[144] Frederick D. Rossini, Heat of Formation of Hydrogen Chloride and some Related
Thermodynamic Data. J. Res. Natl. Bur. Stand. (U.S.), 9, 679702 (1932).
[145] Benjamin Count of Rumford, An Inquiry Concerning the Source of the Heat Which is
Excited by Friction. Philos. Trans. R. Soc. London, 88, 80102 (1798).
[146] James R. Salvador, Fu Guo, Tim Hogan, and Mercouri G. Kanatzidis, Zero Thermal
Expansion in YbGaGe Due to an Electronic Valence Transition. Nature, 425, 702705
(2003).
[147] George Scatchard and C. L. Raymond, VaporLiquid Equilibrium. II. ChloroformEthanol
Mixtures at 35, 45 and 55 . J. Am. Chem. Soc., 60, 12781287 (1938).
[148] George Scatchard, W. J. Hamer, and S. E. Wood, Isotonic Solutions. I. The Chemical
Potential of Water in Aqueous Solutions of Sodium Chloride, Potassium Chloride, Sulfuric
B IBLIOGRAPHY
[149]
[150]
[151]
[152]
[153]
[154]
[155]
[156]
[157]
[158]
[159]
[160]
[161]
[162]
[163]
[164]
[165]
[166]
[167]
[168]
[169]
519
Acid, Sucrose, Urea and Glycerol at 25 . J. Am. Chem. Soc., 60, 30613070 (1938).
Francis Weston Sears and Mark W. Zemansky, University Physics, 4th edition.
Addison-Wesley, Reading, Massachusetts, 1970.
Jan V. Sengers and Anneke Levelt Sengers, The Critical Region. Chem. Eng. News, June
10, 104118 (1968).
F. Hastings Smyth and Leason H. Adams, The system, Calcium OxideCarbon Dioxide. J.
Am. Chem. Soc., 45, 11671184 (1923).
Anthony N. Stranges, William Francis Giauque: An Adventure in Low-Temperature
Research. J. Chem. Educ., 67, 187193 (1990).
Julian M. Sturtevant, The Heat of Dilution of Aqueous Hydrochloric Acid at 25 . J. Am.
Chem. Soc., 62, 32653266 (1940).
Craig Thornber, Thomas Henry, FRS and his son, William Henry, MD, FRS, GS. Online at
http://www.thornber.net/cheshire/ideasmen/henry.html.
Hossein Toghiani, Rebecca K. Toghiani, and Dabir S. Viswanath, Vapor-Liquid Equilibria
for the Methanol-Benzene and Methanol-Thiophene Systems. J. Chem. Eng. Data, 39,
6367 (1994).
Andrzej Treszczanowicz, Teresa Treszczanowicz, Teresa Kasprzycka-Guttman, and
Tomasz S. Pawlowski, Excess Molar Volumes for (Binary Mixtures of 1-Alkanol and
1-Alkene). I. The System 1-Alkanol and 1-Octene at 25 C. J. Solution Chem., 31, 455466
(2002).
Edwin E. Tucker, Edwin H. Lane, and Sherril D. Christian, Vapor Pressure Studies of
Hydrophobic Interactions. Formation of Benzene-Benzene and Cyclohexane-Cyclohexane
Dimers in Dilute Aqueous Solution. J. Solution Chem., 10, 120 (1981).
Louis A. Turner, Zeroth Law of Thermodynamics. Am. J. Phys., 29, 7176 (1961).
Gijs van Ginkel, Prof. Peter J. W. Debye (1884-1966) in 1935-1945: An Investigation of
Historical Sources, revised edition. RIPCN, The Netherlands, 2006. Online at
http://www.theochem.ru.nl/~pwormer/HistoricalsourcesDebye1935-1945.pdf.
J. H. vant Hoff, The R le of Osmotic Pressure in the Analogy between Solutions and
o
Gases. In Harry C. Jones, editor, The Modern Theory of Solution: Memoirs by Pfeffer, vant
Hoff, Arrhenius, and Raoult, pages 1342. Harper and Brothers, New York, 1899.
Translation by Harry C. Jones of Die Rolle des Osmotischen Druckes in der Analogie
zwischen L sungen and Gasen, Z. Phys. Chem. (Leipzig), 1, 481508 (1887).
o
N. B. Vargaftik, B. N. Volkov, and L. D. Voljak, International Tables of the Surface Tension
of Water. J. Phys. Chem. Ref. Data, 12, 817820 (1983).
Donald D. Wagman and Marthada V. Kilday, Enthalpies of Precipitation of Silver Halides;
Entropy of the Aqueous Silver Ion. J. Res. Natl. Bur. Stand. (U.S.), 77A, 569579 (1973).
Donald D. Wagman, William H. Evans, Vivian B. Parker, Richard H. Schumm, Iva Halow,
Sylvia M. Bailey, Kenneth L. Churney, and Ralph L. Nuttall, The NBS Tables of Chemical
Thermodynamic Properties. Selected Values for Inorganic and C1 and C2 Organic
Substances in SI Units. J. Phys. Chem. Ref. Data, 11, Supplement No. 2 (1982).
Jingtao Wang and Mikhail A. Anisimov, Nature of Vapor-Liquid Asymmetry in Fluid
Criticality. Phys. Rev. E, 75, 051107 (2007).
E. W. Washburn, International Critical Tables of Numerical Data, Physics, Chemistry and
Technology, Vol. III. McGraw-Hill, New York, 1928.
Technology, Vol. IV. McGraw-Hill, New York, 1928.
Technology, Vol. VII. McGraw-Hill, New York, 1930.
Edward W. Washburn, Standard States for Bomb Calorimetry. J. Res. Natl. Bur. Stand.
(U.S.), 10, 525558 (1933).
Lynde Phelps Wheeler, Josiah Willard Gibbs: The History of a Great Mind, revised edition.
Yale University Press, New Haven, Connecticut, 1952.
B IBLIOGRAPHY
520
[170] Emmerich Wilhelm, Rubin Battino, and Robert J. Wilcock, Low-Pressure Solubility of
Gases in Liquid Water. Chem. Rev., 77, 219262 (1977).
[171] Jaime Wisniak, Benoit Paul Emile Clapeyron: A Short Bibliographical Sketch. Chem.
Educator, 5, 8387 (2000).
I NDEX
A boldface page number refers to a denition. A page number followed by b is for a

biographical sketch, one followed by n is for a footnote, and one followed by p is for a
problem.
Absolute zero, unattainability of, 162
Acid dissociation constant, 410
Activity, 270, 409
of an electrolyte solute, 293
of a gas, 186, 273
of an ion, 289
of a mixture constituent, 273
of a pure liquid or solid, 273
relative, 270
of a solute, 273
of a solvent, 273
of a symmetrical electrolyte, 291
Activity coefcient, 258, 259
approach to unity, 261
of a gas, 186
from the GibbsDuhem equation,
265266
of an ion, 288
from the DebyeH ckel theory, 295
u
mean ionic
from the DebyeH ckel theory, 297
u
from the Nernst equation, 464
from the osmotic coefcient, 300
from solubility measurement, 392
of a solute
in dilute solution, 261
from gas fugacity, 262265
of a solvent, 371n
from gas fugacity, 262265
stoichiometric, 294
Activity quotient, 351

Additivity rule, 230, 234, 238, 243, 290, 292,
294, 304, 308, 329, 347
Adiabat, 77
Adiabatic
boundary, 28
calorimeter, 169171, 334
demagnetization, 159, 161
ame temperature, 342
process, 51, 57, 95, 129
Advancement, 315
Afnity of reaction, 343n
Amount, 21, 37
Amount of substance, 21, 37, 471
Ampere, 471
Anisotropic phase, 30, 74
Antoine equation, 221p
Athermal process, 305
Avogadro constant, 21, 472
Azeotrope, 406, 436
minimum-boiling, 436
vapor-pressure curve, 437
Azeotropic behavior, 434
Azeotropy, 436
Bar, 39
Barometric formula, 198, 277
Barotropic effect, 33
Base units, 19, 471
Binary mixture, 224
in equilibrium with a pure phase, 375
Binary solution, 225
Bivariant system, 200, 420
Body, 28
521
I NDEX
Boiling point, 204

curve, 205, 434
elevation in a solution, 376, 381
Bomb calorimeter, 321, 334, 336341
Bomb calorimetry, 362p
Boundary, 27
adiabatic, 28
diathermal, 28
Boyle temperature, 35
Brewer, Leo, 272
Bridgman, Percy, 180, 496
Bubble-point curve, 434
Cailletet and Matthias, law of, 206
Caloric theory, 62
Calorie, 86n
Calorimeter, 169
adiabatic, 169171, 334
bomb, 321, 334, 336341
Bunsen ice, 342
combustion, 334
constant-pressure, 171, 214
constant-volume, 170
continuous-ow, 174
ame, 342
heat-ow, 342
isoperibol, 172, 334, 335, 342
isothermal-jacket, 172174, 334, 335,
342
phase-change, 341
reaction, 334336
Calorimetry
bomb, 336, 362p
drop, 191p
to evaluate an equilibrium constant, 355
to measure heat capacities, 169
to measure transition enthalpies, 214
reaction, 323, 334342, 411
Candela, 471
Carath odorys principle of adiabatic
e
inaccessibility, 119
Carnot
cycle, 106, 106109
engine, 106, 106109
heat pump, 109
Carnot, Sadi, 107b
Cell
diagram, 451
electrochemical, 450
galvanic, 450
reaction, 452
with transference, 452
522
without liquid junction, 452

without transference, 452
Cell potential, 90, 454
equilibrium, 90
Celsius scale, 41
Center of mass, 501
Center-of-mass frame, 54, 58, 500503
Centigrade scale, 41
Centrifugal force, 504
Centrifuge, 275
cell, 277280
Charge
electric, 453
number, 295
Chemical amount, 21n
Chemical equation, 313
Chemical potential, 137, 143
of electrons, 457
as a function of T and p, 215
of a liquid or solid, 186
of a pure substance, 182
of a solvent
from the freezing point, 371373
from osmotic pressure, 373375
of a species in a mixture, 236
standard, 258, 270
of a gas, 183
of a gas constituent, 241
of an ion, 288
total, 197
Chemical process, 303
subscript for, 478
Circuit
electrical, 8689
heater, 169, 170, 172, 173
ignition, 337, 339
Clapeyron equation, 217
Clapeyron, Emile, 218b

Clausius
inequality, 119
statement of the second law, 104
Clausius, Rudolf, 103, 104, 110b, 128, 132
ClausiusClapeyron equation, 219, 370
CODATA, 505
Coefcient of thermal expansion, 164
Coexistence curve, 201, 214
liquidgas, 33
Colligative property, 376
523
I NDEX
to estimate solute molar mass, 377

Common ion effect, 392, 445
Component, 422
Components, number of, 47, 229n
Composition variable, 223, 226
relations at innite dilution, 226
Compressibility factor, 35
Compression, 51
Compression factor, 35
Concentration, 29, 224
standard, 254
Condensation curve, 434
Conditions of validity, 23n
Congruent melting, 430
Conjugate
pair, 138
phases
in a binary system, 431
in a ternary system, 444
variables, 59, 138
Consolute point, 431
Constants, physical, values of, 472
Contact
force, 494
potential, 456, 457
Continuity of states, 33
Convergence temperature, 173
Conversion factor, 23
Coriolis force, 277n, 504
Coulombs law, 490
Critical
curve, 440
opalescence, 206
point
of partially-miscible liquids, 431
of a pure substance, 33, 206
pressure, 206
temperature, 206
Cryogenics, 157162
Cryoscopic constant, 379
Cubic expansion coefcient, 164, 211, 219
of an ideal gas, 189p
negative values of, 164n
Curies law of magnetization, 161
Current, electric, 86, 88, 169, 170, 450, 471
Cyclic process, 52
Daltons law, 240
Debye crystal theory, 153
Debye, Peter, 159, 295, 296b, 298, 299
DebyeH ckel
u
equation
for a mean ionic activity coefcient,

297, 298, 465
for a single-ion activity coefcient,
295
limiting law, 297, 331, 333, 391
theory, 295300
Deformation, 30
elastic, 36
plastic, 36
work, 6974
Degree of dissociation, 411
Degrees of freedom, 200, 420
Deliquescence, 440
Density, 29
measurement of, 38
Dependent variable, 46
Derivative, 480
formulas, 480
Dew-point curve, 434
Dialysis, equilibrium, 396
Diathermal boundary, 28
Dieterici equation, 26p
Differential, 24, 482
exact, 52, 482
inexact, 52
total, 135, 482
of the internal energy, 136137
Dilution process, 325
Dimensional analysis, 2425
Disorder, 130
Dissipation of energy, 66, 81, 84, 9095,
114, 124, 130, 171
Dissipative work, 84, 89, 92, 95, 136
Dissociation pressure of a hydrate, 438
Distribution coefcient, 395
Donnan
membrane equilibrium, 397, 397400
potential, 397
DuhemMargules equation, 405
Ebullioscopic constant, 381
Efciency
of a Carnot engine, 112114
of a heat engine, 111
Eforescence, 439
Einstein energy relation, 54, 183n
Elastic deformation, 36
Electric
charge, 453
current, 86, 88, 169, 170, 450, 471
potential, 45, 287, 298, 299, 453
inner, 287, 453
I NDEX
potential difference, 87, 88, 397, 454

power, 174
resistance, 44, 89, 170
Electrical
circuit, 8689
conductor of a galvanic cell, 450
force, 490
heating, 8889, 169, 170, 174
neutrality, 236, 423
resistor, 60, 8890
work, 61, 8691, 96, 169171, 174,
214, 339, 456
Electrochemical
cell, 450
potential, 288n
Electrode, 451
hydrogen, 451
standard, 466
potential, standard, 466
reaction, 452
Electrolyte
solution, 286301
symmetrical, 289291
Electromotive force, 454n
Electron
chemical potential, 457
conductor of a galvanic cell, 450
number, 452
Emf, 454n
Endothermic reaction, 319
Energy, 53, 489496
dissipation of, 66, 81, 84, 9095, 114,
124, 130, 171
Gibbs, 138
Helmholtz, 138
internal, 53, 5354
kinetic, 489
potential, 491
of the system, 5254
thermal, 62, 496
Energy equivalent, 170, 171, 173, 174, 334,
335, 338, 339
Enthalpy, 138
change at constant pressure, 177
of combustion, standard molar, 336, 341
of dilution
integral, 327
molar differential, 327
molar integral, 328
of formation of a solute, molar, 328
of formation, standard molar, 320
of an ion, 323
524
of a solute, 321
of mixing to form an ideal mixture, 305
molar reaction, 315
molar, effect of temperature on,
324325
partial molar, 249
in an ideal gas mixture, 242
relative, of a solute, 330
relative, of the solvent, 329
of a solute in an ideal-dilute solution,
257
reaction
standard molar, 320, 321, 369, 370,
411
standard molar of a cell reaction, 462
relative apparent, of a solute, 331
of solution
at innite dilution, 326
integral, 326
molar differential, 325, 326, 386
molar integral, 326, 329, 358
of vaporization
molar, 212
standard molar, 214
Entropy, 103, 120, 130
change
at constant pressure, 177
at constant volume, 176
an extensive property, 123
as a measure of disorder, 130
of mixing
to form an ideal mixture, 305, 307
negative value, 305
molar, 152157
from calorimetry, 153
of a nonequilibrium state, 123
partial molar, 249
257
reaction
standard molar, 411, 413p
standard molar of a cell reaction, 463
residual, 156157
scale
conventional, 155
practical, 155
standard molar, 156, 411
of a gas, 187, 242
third-law, 152
zero of, 150152
Equation
chemical, 313
525
I NDEX
reaction, 313
stoichiometric, 314
Equation of state, 46, 47
of a uid, 33
of a gas at low pressure, 35, 245, 282p,
363p
of an ideal gas, 33
thermodynamic, 167
virial, 34
Equilibrium
dialysis, 396
gasgas, 441
liquidgas, 400409
liquidliquid, 392395
mechanical, 48
phase transition, 69, 345
position, effect of T and p on, 356359
reaction, 48, 344, 409411
solidliquid, 384392
thermal, 48
transfer, 48
Equilibrium cell potential, 90, 454
Equilibrium conditions
for a gas mixture in a gravitational eld,
275277
in a gravitational eld, 196
in a multiphase multicomponent
system, 236238
in a multiphase one-component system,
194195
for reaction, 344350
for a solution in a centrifuge cell,
277280
Equilibrium constant
mixed, 410
on a pressure basis, 353
thermodynamic, 352
of a cell reaction, 462
temperature dependence, 369
Equilibrium state, 48, 4850
Euler reciprocity relation, 483
Eutectic
composition, 428
halt, 429
point, 427, 430
temperature, 428, 429
Eutonic
composition, 445
point, 444
Exact differential, 52, 482
Excess
function, 146
quantity, 306
work, 91
Exergonic process, 303n
Exothermic reaction, 319
Expansion, 51
free, 79
reversible, of an ideal gas, 127
work, 72, 7279, 96
reversible, 77
Expansivity coefcient, 164
Extensive property, 28
Extent of reaction, 315
External eld, 28, 49, 58, 196
Faraday constant, 287, 453, 472
Field
external, 28, 49, 58, 196
gravitational, 28, 36, 49, 54, 196, 275
magnetic, 161
First law of thermodynamics, 56, 136
Fluid, 31, 3233
supercritical, 32, 33, 206, 211, 212
Flux density, magnetic, 159
Force, 487496
apparent, 277n, 497
centrifugal, 504
contact, 494
Coriolis, 277n, 504
effective, 497
electrical, 490
ctitious, 277n, 497, 499
frictional, 70
gravitational, 37, 198, 490, 504
Formation reaction, 320
Frame
center-of-mass, 54, 58, 500503
lab, 53, 58, 59, 80, 275, 277, 488,
491494, 496, 497, 500, 501
local, 53, 57, 58, 69, 80, 277, 279, 496,
498, 499
nonrotating, 500503
rotating, 277, 497, 503504
reference, 28, 53
earth-xed, 504
inertial, 53, 487, 488, 491
Free expansion, 79, 129
Freezing point, 204
curve, 384
for a binary solidliquid system, 427
of an ideal binary mixture, 385
for solid compound formation, 389
depression in a solution, 376379
526
I NDEX
to evaluate solvent chemical potential,

371373
of an ideal binary mixture, 384385
Friction
internal, 9194
sliding, 496
Frictional force, 70
Fugacity
effect of liquid composition on,
401405
effect of liquid pressure on, 400
of a gas, 184
of a gas mixture constituent, 243
Fugacity coefcient
of a gas, 185
of a gas mixture constituent, 244, 246
Fundamental equations, Gibbs, 142, 236, 287
Galvani potential, 287, 453
Galvanic cell, 89, 450
in an equilibrium state, 49
Gas, 32
ideal, 75
mixture, 240, 242
perfect, 75n
solubility, 406408
thermometry, 4244
Gas constant, 472
Gasgas immiscibility, 441
Giauque, William, 159, 160b
Gibbs
equations, 140, 142
fundamental equations, 142, 236, 287
phase rule
for a multicomponent system,
419426
for a pure substance, 200
Gibbs energy, 138
of formation, standard molar, 355
of an ion, 355
of mixing, 304
to form an ideal mixture, 305
molar, 304
molar, 142, 182
molar reaction, 343
of a cell reaction, 458462
reaction, standard molar, 351, 411
total differential of, for a mixture, 236
Gibbs, Josiah Willard, 142
Gibbs, Willard, 139b
GibbsDuhem equation, 231, 233, 234, 255,
265266, 268, 301, 307, 389, 402
GibbsHelmholtz equation, 368

Gravitational
eld, 28, 36, 49, 54, 196, 275
force, 37, 198, 490, 504
work, 8082, 96
Gravitochemical potential, 197
Green Book, see IUPAC Green Book
H ckel, Erich, 295, 298, 299
u
Heat, 56, 89, 495
ow in an isolated system, 128129
reservoir, 50, 61, 104, 105, 108, 114,
116, 118, 119, 121, 122, 125, 126,
129, 131
technical meaning of, 61
transfer, 6769
Heat capacity, 62, 143144
at constant pressure, 62, 144
molar, 144
partial molar, 249
at constant volume, 62, 143
of an ideal gas, 76
molar, 143
at constant volume and constant
pressure, relation between, 168
measurement of, by calorimetry, 169
molar reaction, 323
Heat engine, 104, 106
Heater circuit, 169, 170, 172, 173
Heating
at constant volume or pressure, 175177
curve, of a calorimeter, 170, 172, 174
electrical, 8889, 169, 170, 174
reversible, 127
Helium, 203n
Helmholtz energy, 138
Henrys law, 250
not obeyed by electrolyte solute, 286
Henrys law constant, 250
effect of pressure on, 408
effect of temperature on, 408
evaluation of, 253
Henrys law constants, relations between
different, 253
Henry, William, 251b
Hesss law, 320, 321, 336
Hess, Germain, 322b
Homogeneous phase, 30
Hydrogen electrode, 451
standard, 466
Ice point, 41
I NDEX
Ice, high pressure forms, 204

Ideal gas, 75
equation, 23, 33, 75
internal pressure, 167
mixture, 240, 242
in a gravitational eld, 275277
and Raoults law, 248
Ideal mixture, 249, 259
mixing process, 304
Ideal solubility
of a gas, 407
of a solid, 387
Ideal-dilute solution, 253
partial molar quantities in, 256257
solvent behavior in, 255256, 402
Ideal-gas temperature, 40, 116
Ignition circuit, 337, 339
Impossible process, 66
Independent variables, 46, 193
of an equilibrium state, 120
number of, 199, 421
Indicator diagram, 77
Inertial reference frame, 53, 488, 491
Inexact differential, 52
Inner electric potential, 453
Integral, 481
formulas, 481
line, 72, 481
Integral enthalpy of dilution, 327
Integral enthalpy of solution, 326
Integrand, 481
Integrating factor, 123
Intensive property, 29
Interface surface, 30
Internal
friction, 9194
pressure, 166, 166167, 169
of an ideal gas, 167
resistance, 456
Internal energy, 53, 5354
of an ideal gas, 75
partial molar, 249
International System of Units, see SI
International temperature scale, 41
Invariant system, 200, 420
Ionic conductor of a galvanic cell, 450
Ionic strength, 295, 297, 299
effect on reaction equilibrium, 410
Irreversible process, 66, 102, 124126,
128130
527
Isobaric process, 51
Isochoric process, 51
Isolated system, 28, 48
spontaneous changes in, 128
Isoperibol calorimeter, 172, 334, 335, 342
Isopiestic
process, 51
solution, 269
vapor pressure technique, 269
Isopleth, 428
Isoteniscope, 204
Isotherm, 77, 210
Isothermal
bomb process, 336, 338339, 341
compressibility, 164, 211, 219
of an ideal gas, 189p
of a liquid or solid, 181
magnetization, 161
pressure changes, 181182
of a condensed phase, 181
process, 51
Isotropic phase, 30
IUPAC, 19
IUPAC Green Book, 19, 138, 182, 212,
260n, 325, 352, 454n, 471
Joule
coefcient, 189p
experiment, 189p
paddle wheel, 8486, 104
Joule, James Prescott, 58, 62, 84, 85b, 86,
100p, 189p
JouleKelvin
coefcient, 158
experiment, 157
JouleThomson
coefcient, 158, 180
experiment, 157
Kelvin (unit), 40, 471
Kelvin, Baron of Largs, 103, 105, 114, 115b
KelvinPlanck statement of the second law,
105, 119
Kilogram, 471
Kinetic energy, 489
Kirchhoff equation, 324, 339
Konowaloffs rule, 406
Lab frame, 53, 58, 59, 80, 275, 277, 488,
491494, 496, 497, 500, 501
Laplace equation, 199, 281p
I NDEX
Law
of Cailletet and Matthias, 206
of rectilinear diameters, 206
scientic, 56
Le Ch teliers principle, 358, 359
a
Legendre transform, 138, 142, 161, 485,
485486
Lever rule, 209, 427
for a binary phase diagram, 428
general form, 209
for one substance in two phases, 207
for partially-miscible liquids, 432
for a ternary system, 443
Lewis and Randall rule, 282p
Lewis, Gilbert Newton, 106, 150, 271b, 295
Line integral, 72, 481
Liquid, 32
Liquid junction, 450, 452, 458
potential, 452, 458
Liquidus curve, 430
for a binary system, 427
for a binary liquidgas system, 433436
at high pressure, 440
Liter, 37
Local frame, 53, 57, 58, 69, 80, 277, 279,
496, 498, 499
nonrotating, 500503
rotating, 497
Magnetic
enthalpy, 161
eld, 161
ux density, 159
Magnetization, isothermal, 161
Mass, 471
meaurement of, 3637
Mass fraction, 224
Maxwell relations, 141
Maxwell, James Clerk, 40
McMillanMayer theory, 262n
Mean ionic activity coefcient, see Activity
coefcient, mean ionic
Mean molar volume, 231
Melting point, 204
normal, 42
Membrane equilibrium
Donnan, 397, 397400
osmotic, 396
Membrane, semipermeable, 49, 373
Metastable state, 50
Meter, 471
528
Method of intercepts, 231, 232

Metre, 471
Microstate, 131
Milliter, 37
Minimal work principle, 91
Miscibility gap
in a binary system, 313, 431
in a ternary system, 444
Mixing process, 303
Mixture
binary, 224
of xed composition, 226
gas, in a gravitational eld, 275277
ideal, 259
and chemical potential, 249
simple, 309
Molal boiling-point elevation constant, 381
Molal freezing-point depression constant,
379
Molality, 224
standard, 254
Molar
differential reaction quantity, 317
excess quantity, 306
integral reaction quantity, 317
mass, 37
from a colligative property, 377
from sedimentation equilibrium, 280
quantity, 29, 176
reaction quantity, 317
standard, 319
Mole, 21, 37, 471
Mole fraction, 223
standard, 255
Mole ratio, 225
Molecular mass, relative, 37
Molecular weight, 37
Natural variables, 138, 140142
Nernst
distribution law, 395
equation, 463
heat theorem, 150
Nernst, Walther, 150n, 151b
Neutrality, electrical, 236, 423
Newtons law of cooling, 172
Newtons law of universal gravitation, 490
Newtons second law of motion, 70, 198, 488
Newtons third law of action and reaction,
489, 490, 499
Nonexpansion work, 136, 146
I NDEX
Normal
boiling point, 205
melting point, 42, 205
Optical pyrometer, 45
Osmosis, 374
Osmotic coefcient, 266, 382
evaluation, 270
of a mean ionic activity coefcient,
300
Osmotic membrane equilibrium, 396
Osmotic pressure, 49, 374, 376, 382, 396
to evaluate solvent chemical potential,
373375
vant Hoffs equation for, 382
Paddle wheel, 60
Joule, 8486, 104
Partial
specic quantity, 235
specic volume, 235
Partial derivative, 480
expressions at constant T , p, and V ,
177180
Partial molar
enthalpy, 249
relative, of a solute, 330
relative, of the solvent, 329
257
entropy, 249
257
Gibbs energy, 143
heat capacity at constant pressure, 249
internal energy, 249
quantity, 226
in general, 234
general relations, 238239
in an ideal mixture, 249
in an ideal-dilute solution, 256257
volume, 227229, 249
interpretation, 228
negative value of, 228
Partial pressure, 240
Partition coefcient, 395
Pascal (unit), 39
529
Path, 50
Path function, 52, 61
Peritectic point, 430
Perpetual motion of the second kind, 104n
Phase, 30
anisotropic, 30, 74
boundary, 201
coexistence, 31
homogeneous, 30
isotropic, 30
rule, see Gibbs phase rule
separation of a liquid mixture, 311313,
392, 431
transition, 31
equilibrium, 32, 69, 152, 345
Phase diagram
for a binary liquidgas system, 433
for a binary liquidliquid system, 431
for a binary solidgas system, 437
for a ternary system, 442
Physical constants, values of, 472
Physical quantities, symbols for, 473476
Physical state, 30
symbols for, 477
Pitzer, Kenneth, 272
Plait point, 444
Planck, Max, 103, 105, 117b
Plasma, 32
Plastic deformation, 36
Plimsoll mark, 182
Potential
chemical, 137
electric, 453
energy, 491
function, 490
standard cell, 462, 463
evaluation of, 465
thermodynamic, 135
Potentiometer to measure equilibrium cell
potential, 455
Poynting factor, 400
Prexes, 21
Pressure
changes, isothermal, 181182
of a condensed phase, 181
dissociation, of a hydrate, 438
internal, 166, 166167, 169
I NDEX
in a liquid droplet, 198

measurement of, 3840
negative, 167n
partial, 240
standard, 39, 182, 275, 360p, 468p
sublimation, 204
vapor, see Vapor pressure
Pressure factor, 272, 273275
of an ion, 289
Pressurevolume diagram, 77
Process, 50
adiabatic, 51, 57, 95, 129
chemical, 303
subscript for, 478
compression, 51
cyclic, 52
dilution, 325
expansion, 51
impossible, 66, 102, 103105
irreversible, 66, 102, 124126, 128130
isenthalpic, 158
isobaric, 51
isochoric, 51
isopiestic, 51
isothermal, 51
mechanical, 130
mixing, 303
purely mechanical, 66
quasistatic, 64
reverse of a, 64
reversible, 64, 6466, 95, 102, 103, 130
solution, 325
spontaneous, 64, 6466, 102, 130, 343
throttling, 157
Product, 314
Proper quotient, 351
Property
extensive, 28
intensive, 29
molar, 176
Quantity
molar, 29
specic, 29
Quantity calculus, 22
Quartz crystal thermometer, 45
Quasicrystalline lattice model, 309
Quasistatic process, 64
Randall, Merle, 106, 150, 295
530
Raoults law
deviations from, 403405, 434
for fugacity, 247, 248
in a binary liquid mixture, 403
in an ideal-dilute solution, 256
for partial pressure, 247, 247
in a binary system, 433
Raoult, Francois, 246
Raoult, Francois-Marie, 380b
Reactant, 314
Reaction
between pure phases, 345
cell, 452
endothermic, 319
equation, 313
equilibrium, 409411
exothermic, 319
in a gas phase, 353354
in a mixture, 345347
quotient, 351, 463
in solution, 354
Reaction quantity
molar, 317
molar differential, 317
molar integral, 317
Reciprocity relation, 141, 148p, 483
Rectilinear diameters, law of, 206
RedlichKister series, 311
RedlichKwong equation, 26p, 34
Reduction to standard states, 336
Reference frame, 28, 53
earth-xed, 504
inertial, 487
Reference state, 258
of an element, 320
of an ion, 288
of a mixture constituent, 260
of a solute, 253254, 260
of a solvent, 260
Regular solution, 310
Relative activity, 270
Relative apparent molar enthalpy of a solute,
331
Relative molecular mass, 37
Relative partial molar enthalpy
of a solute, 330
of the solvent, 329
Residual entropy, 156157
Resistance
electric, 44, 89, 170
internal, 456
I NDEX
Resistor, electrical, 60, 8890

Retrograde
condensation, 441
vaporization, 441
Reverse of a process, 64
Reversible
adiabatic expansion of an ideal gas, 75
adiabatic surface, 120
expansion and compression, 71
expansion of an ideal gas, 127
expansion work, 77
heating, 127
isothermal expansion of an ideal gas, 75
phase transition, 69
process, 64, 6466, 95, 102, 103, 130
work, 95
Rotating local frame, 277, 503504
Rubber, thermodynamics of, 149p
Rumford, Count, 62, 63b
Salt bridge, 458
Salting-out effect on gas solubility, 406, 415p
Saturated solution, 386
Saturation
temperature, 204
vapor pressure, 204
Second, 471
Second law of thermodynamics
Clausius statement, 104
equivalence of Clausius and
KelvinPlanck statements,
109111
KelvinPlanck statement, 105, 119, 122
mathematical statement, 103, 126
derivation, 116124
Sedimentation equilibrium, 280
Shaft work, 82, 8286
Shear stress, 30
SI, 19
base units, 20, 471
derived units, 20
prexes, 21
Simple mixture, 309
Solid, 30, 36
viscoelastic, 31
Solid compound, 387, 429, 437
of mixture components, 387390
Solidus curve for a binary system, 427
Solubility
curve, 384
of a gas, 406408
531
ideal, 407
of a liquid, relation to Henrys law
constant, 394
of a solid, 386
of a solid electrolyte, 390392
of a solid nonelectrolyte, 386387
Solubility product, 391, 462
temperature dependence, 392
Solute, 224
reference state, 253254, 260
Solution, 224
binary, 225
in a centrifuge cell, 277280
ideal-dilute, 253
multisolute electrolyte, 293
process, 325
regular, 310
saturated, 386
solid, 246, 429
Solvent, 224
activity coefcient of, 371n
behavior in an ideal-dilute solution,
255256, 402
Species, 223, 420
Specic
quantity, 29
volume, 29
Spontaneous process, 64, 6466, 102, 130,
343
Standard
boiling point, 205
cell potential, 462, 463
evaluation, 465
chemical potential, see Chemical
potential, standard
composition, 254
concentration, 254
electrode potential, 466
hydrogen electrode, 466
melting point, 205
molality, 254
mole fraction, 255
pressure, 39, 182, 275, 360p, 468p
Standard molar
properties, values of, 505507
quantity, 186
evaluation of, 411412
of a gas, 186188
reaction quantity, 319
Standard state
of a gas, 182
532
I NDEX
of an ion, 288
of a mixture component, 270
of a pure liquid or solid, 182
State
of aggregation, 30, 477
equilibrium, 48, 4850
metastable, 50
physical, 30, 477
standard, see Standard state
steady, 50
of a system, 45, 4550
State function, 45, 4547
change of, 51
Statistical mechanics, 34, 131, 132
Boyle temperature, 35
Debye crystal theory, 153
DebyeH ckel theory, 298
u
ideal mixture, 262
McMillanMayer theory, 262n
mixture theory, 309
molar entropy of a gas, 155
molar heat capacity of a metal, 153
second virial coefcient, 245
virial equations, 34
Steady state, 50
Steam engine, 108
Steam point, 41, 206
Stirring work, 83, 8384
Stoichiometric
activity coefcient, 294
addition compound, see Solid
compound
coefcient, 316
equation, 314
number, 316, 321, 351
Sublimation
point, 204
pressure, 204
temperature, 204
Subscripts for chemical processes, 478
Substance, 223
Subsystem, 28
Supercritical uid, 32, 33, 206, 211, 212
Superscripts, 479
Supersystem, 28, 116, 118, 121, 122, 125,
126
Surface tension, 144
Surface work, 144145
Surroundings, 27
Symbols for physical quantities, 473476
System, 27
closed, 28
isolated, 28, 48
open, 28, 228, 229, 236
state of, 45, 4550
System point, 201
Syst me International dUnit s, see SI
e
e
Temperature
Boyle, 35
convergence, 173
critical, 206
equilibrium systems for xed values, 42
ideal-gas, 40, 116, 167
international scale, 41
measurement of, 4045
scales, 4041
thermodynamic, 40, 114, 114116, 167,
471
upper consolute, 431
upper critical solution, 431
Thermal
analysis, 429
conductance, 172
energy, 62, 496
reservoir, 50
Thermocouple, 45, 457n
Thermodynamic
equation of state, 167
equilibrium constant, 352
potential, 135
temperature, 40, 114, 114116, 167,
471
Thermometer
Beckmann, 44
constant-volume gas, 42
liquid-in-glass, 40, 44
optical pyrometer, 45
quartz crystal, 45
resistance, 44
thermocouple, 45
Thermometry, gas, 4244
Thermopile, 45, 342, 457n
Third-law entropy, 152
Third law of thermodynamics, 121, 150
Thompson, Benjamin, 62, 63b
Thomson, William, 105, 115b
Throttling process, 157
Tie line, 201
on a binary phase diagram, 427
on a ternary phase diagram, 443
Torque, 82
Total differential, 135, 482
533
I NDEX
of the Gibbs energy of a mixture, 236

of the Gibbs energy of a pure substance,
182
of the internal energy, 136137
of the volume, 229
Triple
line, 95n, 201
point, 201
cell, 42
of H2 O, 40
Ultracentrifuge, 277280
Units, 22
Non-SI, 20
SI, 20
SI derived, 20
Univariant system, 200, 420
Upper consolute temperature, 431
Upper critical solution temperature, 431
vant Hoff, Jacobus, 383b
vant Hoff equation, 370
vant Hoffs equation for osmotic pressure,
382
Vapor, 33
Vapor pressure, 204
curve, 205
of a liquid droplet, 401
lowering in a solution, 376, 381382
saturation, 204
Vaporization, 370
molar enthalpy of, 212
molar heat of, 212
Vaporus curve
for a binary liquidgas system, 433,
435, 436
Variables
conjugate, 138
dependent, 46
independent, 46
natural, 138, 140142
number of independent, 47
Variance, 200, 420
Virial
coefcient, 34
equation
for a gas mixture, 244
for a pure gas, 34
Virtual displacement, 195
Viscoelastic solid, 31
Volume
mean molar, 231
meaurement of, 37
molar, 30, 38
partial molar, 249
specic, 29
total differential in an open system, 229
Washburn corrections, 340, 341, 364p
Work, 56, 5759, 488496
coefcient, 59
coordinate, 59, 120
deformation, 6974
dissipative, 84, 89, 92, 95, 136
of electric polarization, 96
electrical, 61, 8691, 96, 169171, 174,
214, 339, 456
excess, 91
expansion, 72, 7279, 96
gravitational, 8082, 96
of magnetization, 96, 159
mechanical, 96
nonexpansion, 136, 146
reversible, 95
reversible expansion, 77
shaft, 82, 8286, 96
stirring, 83, 8384
stretching, 96
surface, 96, 144145
Working substance, 106, 114
Zeotropic behavior, 437
Zeroth law of thermodynamics, 40

Howard Devoe Thermodynamic and Chemical

Uploaded by

Copyright:

Available Formats

Howard Devoe Thermodynamic and Chemical

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Howard Devoe Thermodynamic and Chemical

Uploaded by

Copyright:

Available Formats

THERMODYNAMICS

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs

I thank the Department of Chemistry and Biochemistry, University of Maryland, College

Preface to the Second Edition

From the Preface to the First Edition

2 Systems and Their Properties

3 The First Law

4 The Second Law

6 The Third Law and Cryogenics

7 Pure Substances in Single Phases

8 Phase Transitions and Equilibria of Pure Substances

11 Reactions and Other Chemical Processes

12 Equilibrium Conditions in Multicomponent Systems

13 The Phase Rule and Phase Diagrams

Appendix A Denitions of the SI Base Units

Appendix B Physical Constants

Appendix C Symbols for Physical Quantities

Appendix D Miscellaneous Abbreviations and Symbols

Appendix E Calculus Review

Appendix F Mathematical Properties of State Functions

Appendix G Forces, Energy, and Work

Appendix H Standard Molar Thermodynamic Properties

Answers to Selected Problems

Preface to the Second Edition

From the Preface to the First Edition

2 Systems and Their Properties

2.4.4 Equilibrium states . . . . .

3 The First Law

4 The Second Law

6 The Third Law and Cryogenics

8 Phase Transitions and Equilibria of Pure Substances

9.8.1 Gas mixture in a gravitational eld . . . . . . . . . . . . . . . . . . 275

11.7.3 General derivation . . . . . . . . . . . . . . . . . .

12 Equilibrium Conditions in Multicomponent Systems

12.9 Reaction Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409

Appendix A Denitions of the SI Base Units

Appendix B Physical Constants

Appendix C Symbols for Physical Quantities

Appendix D Miscellaneous Abbreviations and Symbols

Appendix F Mathematical Properties of State Functions

Appendix G Forces, Energy, and Work

Appendix H Standard Molar Thermodynamic Properties

Answers to Selected Problems

Benjamin Thompson, Count of Rumford

Benoit Paul Emile Clapeyron . . . . . .

Jacobus Henricus vant Hoff . . . . . .

P REFACE TO THE S ECOND E DITION

F ROM THE P REFACE TO THE F IRST

A theory is the more impressive the greater the simplicity of its

Thermodynamics is a quantitative subject. It allows us to derive relations between the

Table 1.2 SI derived units

Table 1.3 Non-SI derived units

1.1.1 Amount of substance and amount

[117]. An alternative name suggested for n is chemical amount.

1.2 QUANTITY CALCULUS

and in different density units by the equation

We may divide both sides of the last equation by 1 g cm

to obtain a new equation

In a table, the ratio =g cm 3 makes a convenient heading for a column of density

.5:000 mol/.8:3145 J K 1 mol