Physical Chemistry For Advanced Level

THERMODYNAMICS
AND
CHEMISTRY
SECOND EDITION
HOWARD DEVOE
Thermodynamics
and Chemistry
Second Edition
Version 6, February 2015
Howard DeVoe
Associate Professor of Chemistry Emeritus
University of Maryland, College Park, Maryland
The first edition of this book was previously published by Pearson Education, Inc. It was
copyright ©2001 by Prentice-Hall, Inc.
The second edition, version 6 is copyright ©2015 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 file 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.
The book was typeset using the LATEX typesetting system and the memoir class. Most of
the figures 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 uni-

versity, you may send a request to hdevoe@umd.edu 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 verification 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 101
5 Thermodynamic Potentials 134
6 The Third Law and Cryogenics 149
7 Pure Substances in Single Phases 163
8 Phase Transitions and Equilibria of Pure Substances 192
9 Mixtures 222
10 Electrolyte Solutions 285
11 Reactions and Other Chemical Processes 302
12 Equilibrium Conditions in Multicomponent Systems 366
13 The Phase Rule and Phase Diagrams 418
14 Galvanic Cells 449
Appendix A Definitions of the SI Base Units 470
4
S HORT C ONTENTS 5
Appendix B Physical Constants 471
Appendix C Symbols for Physical Quantities 472
Appendix D Miscellaneous Abbreviations and Symbols 476
Appendix E Calculus Review 479
Appendix F Mathematical Properties of State Functions 481
Appendix G Forces, Energy, and Work 486
Appendix H Standard Molar Thermodynamic Properties 504
Appendix I Answers to Selected Problems 507
Bibliography 511
Index 520
Thermodynamics and Chemistry, 2nd edition, version 6 © 2015 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 19
1.1 Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.1.1 Amount of substance and amount . . . . . . . . . . . . . . . . . . 21
1.2 Quantity Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.3 Dimensional Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2 Systems and Their Properties 27

2.1 The System, Surroundings, and Boundary . . . . . . . . . . . . . . . . . . 27
2.1.1 Extensive and intensive properties . . . . . . . . . . . . . . . . . . 28
2.2 Phases and Physical States of Matter . . . . . . . . . . . . . . . . . . . . . 30
2.2.1 Physical states of matter . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2 Phase coexistence and phase transitions . . . . . . . . . . . . . . . 31
2.2.3 Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2.4 The equation of state of a fluid . . . . . . . . . . . . . . . . . . . . 33
2.2.5 Virial equations of state for pure gases . . . . . . . . . . . . . . . . 34
2.2.6 Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3 Some Basic Properties and Their Measurement . . . . . . . . . . . . . . . 36
2.3.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3.3 Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.4 Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.5 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 The State of the System . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.4.1 State functions and independent variables . . . . . . . . . . . . . . 45
2.4.2 An example: state functions of a mixture . . . . . . . . . . . . . . 46
2.4.3 More about independent variables . . . . . . . . . . . . . . . . . . 47
6
C ONTENTS 7
2.4.4 Equilibrium states . . . . . . . . . . . . . . . . . . . . . . . . . . 48

2.4.5 Steady states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.5 Processes and Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.6 The Energy of the System . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.6.1 Energy and reference frames . . . . . . . . . . . . . . . . . . . . . 53
2.6.2 Internal energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3 The First Law 56

3.1 Heat, Work, and the First Law . . . . . . . . . . . . . . . . . . . . . . . . 56
3.1.1 The concept of thermodynamic work . . . . . . . . . . . . . . . . 57
3.1.2 Work coefficients and work coordinates . . . . . . . . . . . . . . . 59
3.1.3 Heat and work as path functions . . . . . . . . . . . . . . . . . . . 60
3.1.4 Heat and heating . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1.5 Heat capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.1.6 Thermal energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2 Spontaneous, Reversible, and Irreversible Processes . . . . . . . . . . . . . 62
3.2.1 Reversible processes . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2.2 Irreversible processes . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2.3 Purely mechanical processes . . . . . . . . . . . . . . . . . . . . . 66
3.3 Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.3.1 Heating and cooling . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.3.2 Spontaneous phase transitions . . . . . . . . . . . . . . . . . . . . 68
3.4 Deformation Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.4.1 Gas in a cylinder-and-piston device . . . . . . . . . . . . . . . . . 69
3.4.2 Expansion work of a gas . . . . . . . . . . . . . . . . . . . . . . . 71
3.4.3 Expansion work of an isotropic phase . . . . . . . . . . . . . . . . 73
3.4.4 Generalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.5 Applications of Expansion Work . . . . . . . . . . . . . . . . . . . . . . . 74
3.5.1 The internal energy of an ideal gas . . . . . . . . . . . . . . . . . . 74
3.5.2 Reversible isothermal expansion of an ideal gas . . . . . . . . . . . 75
3.5.3 Reversible adiabatic expansion of an ideal gas . . . . . . . . . . . . 75
3.5.4 Indicator diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.5.5 Spontaneous adiabatic expansion or compression . . . . . . . . . . 78
3.5.6 Free expansion of a gas into a vacuum . . . . . . . . . . . . . . . . 79
3.6 Work in a Gravitational Field . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7 Shaft Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.7.1 Stirring work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.7.2 The Joule paddle wheel . . . . . . . . . . . . . . . . . . . . . . . . 84
3.8 Electrical Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.8.1 Electrical work in a circuit . . . . . . . . . . . . . . . . . . . . . . 86
3.8.2 Electrical heating . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.8.3 Electrical work with a galvanic cell . . . . . . . . . . . . . . . . . 89
3.9 Irreversible Work and Internal Friction . . . . . . . . . . . . . . . . . . . . 91
3.10 Reversible and Irreversible Processes: Generalities . . . . . . . . . . . . . 94
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
C ONTENTS 8
4 The Second Law 101

4.1 Types of Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2 Statements of the Second Law . . . . . . . . . . . . . . . . . . . . . . . . 102
4.3 Concepts Developed with Carnot Engines . . . . . . . . . . . . . . . . . . 105
4.3.1 Carnot engines and Carnot cycles . . . . . . . . . . . . . . . . . . 105
4.3.2 The equivalence of the Clausius and Kelvin–Planck statements . . . 108
4.3.3 The efficiency of a Carnot engine . . . . . . . . . . . . . . . . . . 110
4.3.4 Thermodynamic temperature . . . . . . . . . . . . . . . . . . . . . 113
4.4 Derivation of the Mathematical Statement of the Second Law . . . . . . . 115
4.4.1 The existence of the entropy function . . . . . . . . . . . . . . . . 115
4.4.2 Using reversible processes to define the entropy . . . . . . . . . . . 119
4.4.3 Some properties of the entropy . . . . . . . . . . . . . . . . . . . . 122
4.5 Irreversible Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.5.1 Irreversible adiabatic processes . . . . . . . . . . . . . . . . . . . 123
4.5.2 Irreversible processes in general . . . . . . . . . . . . . . . . . . . 124
4.6 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.6.1 Reversible heating . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.6.2 Reversible expansion of an ideal gas . . . . . . . . . . . . . . . . . 126
4.6.3 Spontaneous changes in an isolated system . . . . . . . . . . . . . 127
4.6.4 Internal heat flow in an isolated system . . . . . . . . . . . . . . . 127
4.6.5 Free expansion of a gas . . . . . . . . . . . . . . . . . . . . . . . . 128
4.6.6 Adiabatic process with work . . . . . . . . . . . . . . . . . . . . . 128
4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.8 The Statistical Interpretation of Entropy . . . . . . . . . . . . . . . . . . . 129
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5 Thermodynamic Potentials 134

5.1 Total Differential of a Dependent Variable . . . . . . . . . . . . . . . . . . 134
5.2 Total Differential of the Internal Energy . . . . . . . . . . . . . . . . . . . 135
5.3 Enthalpy, Helmholtz Energy, and Gibbs Energy . . . . . . . . . . . . . . . 137
5.4 Closed Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.5 Open Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.6 Expressions for Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . 142
5.7 Surface Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.8 Criteria for Spontaneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6 The Third Law and Cryogenics 149

6.1 The Zero of Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.2 Molar Entropies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.2.1 Third-law molar entropies . . . . . . . . . . . . . . . . . . . . . . 151
6.2.2 Molar entropies from spectroscopic measurements . . . . . . . . . 154
6.2.3 Residual entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.3 Cryogenics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.3.1 Joule–Thomson expansion . . . . . . . . . . . . . . . . . . . . . . 156
6.3.2 Magnetization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
C ONTENTS 9
Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7 Pure Substances in Single Phases 163

7.1 Volume Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.2 Internal Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7.3 Thermal Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
7.3.1 The relation between CV;m and Cp;m . . . . . . . . . . . . . . . . . 167
7.3.2 The measurement of heat capacities . . . . . . . . . . . . . . . . . 168
7.3.3 Typical values . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
7.4 Heating at Constant Volume or Pressure . . . . . . . . . . . . . . . . . . . 174
7.5 Partial Derivatives with Respect to T, p, and V . . . . . . . . . . . . . . . 176
7.5.1 Tables of partial derivatives . . . . . . . . . . . . . . . . . . . . . 176
7.5.2 The Joule–Thomson coefficient . . . . . . . . . . . . . . . . . . . 179
7.6 Isothermal Pressure Changes . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.6.1 Ideal gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.6.2 Condensed phases . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.7 Standard States of Pure Substances . . . . . . . . . . . . . . . . . . . . . 181
7.8 Chemical Potential and Fugacity . . . . . . . . . . . . . . . . . . . . . . . 181
7.8.1 Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
7.8.2 Liquids and solids . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.9 Standard Molar Quantities of a Gas . . . . . . . . . . . . . . . . . . . . . 185
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
8 Phase Transitions and Equilibria of Pure Substances 192

8.1 Phase Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8.1.1 Equilibrium conditions . . . . . . . . . . . . . . . . . . . . . . . . 192
8.1.2 Equilibrium in a multiphase system . . . . . . . . . . . . . . . . . 193
8.1.3 Simple derivation of equilibrium conditions . . . . . . . . . . . . . 194
8.1.4 Tall column of gas in a gravitational field . . . . . . . . . . . . . . 195
8.1.5 The pressure in a liquid droplet . . . . . . . . . . . . . . . . . . . 197
8.1.6 The number of independent variables . . . . . . . . . . . . . . . . 198
8.1.7 The Gibbs phase rule for a pure substance . . . . . . . . . . . . . . 199
8.2 Phase Diagrams of Pure Substances . . . . . . . . . . . . . . . . . . . . . 199
8.2.1 Features of phase diagrams . . . . . . . . . . . . . . . . . . . . . . 200
8.2.2 Two-phase equilibrium . . . . . . . . . . . . . . . . . . . . . . . . 203
8.2.3 The critical point . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
8.2.4 The lever rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
8.2.5 Volume properties . . . . . . . . . . . . . . . . . . . . . . . . . . 209
8.3 Phase Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
8.3.1 Molar transition quantities . . . . . . . . . . . . . . . . . . . . . . 211
8.3.2 Calorimetric measurement of transition enthalpies . . . . . . . . . 213
8.3.3 Standard molar transition quantities . . . . . . . . . . . . . . . . . 213
8.4 Coexistence Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
8.4.1 Chemical potential surfaces . . . . . . . . . . . . . . . . . . . . . 214
8.4.2 The Clapeyron equation . . . . . . . . . . . . . . . . . . . . . . . 215
8.4.3 The Clausius–Clapeyron equation . . . . . . . . . . . . . . . . . . 218
C ONTENTS 10
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
9 Mixtures 222
9.1 Composition Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
9.1.1 Species and substances . . . . . . . . . . . . . . . . . . . . . . . . 222
9.1.2 Mixtures in general . . . . . . . . . . . . . . . . . . . . . . . . . . 222
9.1.3 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
9.1.4 Binary solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
9.1.5 The composition of a mixture . . . . . . . . . . . . . . . . . . . . 225
9.2 Partial Molar Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
9.2.1 Partial molar volume . . . . . . . . . . . . . . . . . . . . . . . . . 226
9.2.2 The total differential of the volume in an open system . . . . . . . . 228
9.2.3 Evaluation of partial molar volumes in binary mixtures . . . . . . . 230
9.2.4 General relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
9.2.5 Partial specific quantities . . . . . . . . . . . . . . . . . . . . . . . 234
9.2.6 The chemical potential of a species in a mixture . . . . . . . . . . . 235
9.2.7 Equilibrium conditions in a multiphase, multicomponent system . . 235
9.2.8 Relations involving partial molar quantities . . . . . . . . . . . . . 237
9.3 Gas Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
9.3.1 Partial pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
9.3.2 The ideal gas mixture . . . . . . . . . . . . . . . . . . . . . . . . . 239
9.3.3 Partial molar quantities in an ideal gas mixture . . . . . . . . . . . 239
9.3.4 Real gas mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
9.4 Liquid and Solid Mixtures of Nonelectrolytes . . . . . . . . . . . . . . . . 245
9.4.1 Raoult’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
9.4.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
9.4.3 Partial molar quantities in ideal mixtures . . . . . . . . . . . . . . 248
9.4.4 Henry’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
9.4.5 The ideal-dilute solution . . . . . . . . . . . . . . . . . . . . . . . 252
9.4.6 Solvent behavior in the ideal-dilute solution . . . . . . . . . . . . . 254
9.4.7 Partial molar quantities in an ideal-dilute solution . . . . . . . . . . 255
9.5 Activity Coefficients in Mixtures of Nonelectrolytes . . . . . . . . . . . . 257
9.5.1 Reference states and standard states . . . . . . . . . . . . . . . . . 257
9.5.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
9.5.3 Real mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
9.5.4 Nonideal dilute solutions . . . . . . . . . . . . . . . . . . . . . . . 260
9.6 Evaluation of Activity Coefficients . . . . . . . . . . . . . . . . . . . . . . 261
9.6.1 Activity coefficients from gas fugacities . . . . . . . . . . . . . . . 261
9.6.2 Activity coefficients from the Gibbs–Duhem equation . . . . . . . 264
9.6.3 Activity coefficients from osmotic coefficients . . . . . . . . . . . 265
9.6.4 Fugacity measurements . . . . . . . . . . . . . . . . . . . . . . . . 267
9.7 Activity of an Uncharged Species . . . . . . . . . . . . . . . . . . . . . . 269
9.7.1 Standard states . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
9.7.2 Activities and composition . . . . . . . . . . . . . . . . . . . . . . 271
9.7.3 Pressure factors and pressure . . . . . . . . . . . . . . . . . . . . . 272
9.8 Mixtures in Gravitational and Centrifugal Fields . . . . . . . . . . . . . . 274
C ONTENTS 11
9.8.1 Gas mixture in a gravitational field . . . . . . . . . . . . . . . . . . 274

9.8.2 Liquid solution in a centrifuge cell . . . . . . . . . . . . . . . . . . 276
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
10 Electrolyte Solutions 285

10.1 Single-ion Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
10.2 Solution of a Symmetrical Electrolyte . . . . . . . . . . . . . . . . . . . . 288
10.3 Electrolytes in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.3.1 Solution of a single electrolyte . . . . . . . . . . . . . . . . . . . . 291
10.3.2 Multisolute solution . . . . . . . . . . . . . . . . . . . . . . . . . 292
10.3.3 Incomplete dissociation . . . . . . . . . . . . . . . . . . . . . . . 293
10.4 The Debye–Hückel Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 294
10.5 Derivation of the Debye–Hückel Equation . . . . . . . . . . . . . . . . . . 297
10.6 Mean Ionic Activity Coefficients from Osmotic Coefficients . . . . . . . . 299
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
11 Reactions and Other Chemical Processes 302

11.1 Mixing Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
11.1.1 Mixtures in general . . . . . . . . . . . . . . . . . . . . . . . . . . 303
11.1.2 Ideal mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
11.1.3 Excess quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.1.4 The entropy change to form an ideal gas mixture . . . . . . . . . . 306
11.1.5 Molecular model of a liquid mixture . . . . . . . . . . . . . . . . . 308
11.1.6 Phase separation of a liquid mixture . . . . . . . . . . . . . . . . . 310
11.2 The Advancement and Molar Reaction Quantities . . . . . . . . . . . . . . 312
11.2.1 An example: ammonia synthesis . . . . . . . . . . . . . . . . . . . 313
11.2.2 Molar reaction quantities in general . . . . . . . . . . . . . . . . . 315
11.2.3 Standard molar reaction quantities . . . . . . . . . . . . . . . . . . 318
11.3 Molar Reaction Enthalpy . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
11.3.1 Molar reaction enthalpy and heat . . . . . . . . . . . . . . . . . . . 318
11.3.2 Standard molar enthalpies of reaction and formation . . . . . . . . 319
11.3.3 Molar reaction heat capacity . . . . . . . . . . . . . . . . . . . . . 322
11.3.4 Effect of temperature on reaction enthalpy . . . . . . . . . . . . . . 323
11.4 Enthalpies of Solution and Dilution . . . . . . . . . . . . . . . . . . . . . 324
11.4.1 Molar enthalpy of solution . . . . . . . . . . . . . . . . . . . . . . 324
11.4.2 Enthalpy of dilution . . . . . . . . . . . . . . . . . . . . . . . . . 326
11.4.3 Molar enthalpies of solute formation . . . . . . . . . . . . . . . . . 327
11.4.4 Evaluation of relative partial molar enthalpies . . . . . . . . . . . . 328
11.5 Reaction Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
11.5.1 The constant-pressure reaction calorimeter . . . . . . . . . . . . . 333
11.5.2 The bomb calorimeter . . . . . . . . . . . . . . . . . . . . . . . . 335
11.5.3 Other calorimeters . . . . . . . . . . . . . . . . . . . . . . . . . . 340
11.6 Adiabatic Flame Temperature . . . . . . . . . . . . . . . . . . . . . . . . 341
11.7 Gibbs Energy and Reaction Equilibrium . . . . . . . . . . . . . . . . . . . 342
11.7.1 The molar reaction Gibbs energy . . . . . . . . . . . . . . . . . . . 342
11.7.2 Spontaneity and reaction equilibrium . . . . . . . . . . . . . . . . 342
C ONTENTS 12
11.7.3 General derivation . . . . . . . . . . . . . . . . . . . . . . . . . . 343

11.7.4 Pure phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
11.7.5 Reactions involving mixtures . . . . . . . . . . . . . . . . . . . . . 344
11.7.6 Reaction in an ideal gas mixture . . . . . . . . . . . . . . . . . . . 346
11.8 The Thermodynamic Equilibrium Constant . . . . . . . . . . . . . . . . . 349
11.8.1 Activities and the definition of K . . . . . . . . . . . . . . . . . . 349
11.8.2 Reaction in a gas phase . . . . . . . . . . . . . . . . . . . . . . . . 352
11.8.3 Reaction in solution . . . . . . . . . . . . . . . . . . . . . . . . . 353
11.8.4 Evaluation of K . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
11.9 Effects of Temperature and Pressure on Equilibrium Position . . . . . . . . 356
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
12 Equilibrium Conditions in Multicomponent Systems 366

12.1 Effects of Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
12.1.1 Variation of i =T with temperature . . . . . . . . . . . . . . . . . 366
12.1.2 Variation of ıi =T with temperature . . . . . . . . . . . . . . . . . 367
12.1.3 Variation of ln K with temperature . . . . . . . . . . . . . . . . . . 368
12.2 Solvent Chemical Potentials from Phase Equilibria . . . . . . . . . . . . . 369
12.2.1 Freezing-point measurements . . . . . . . . . . . . . . . . . . . . 370
12.2.2 Osmotic-pressure measurements . . . . . . . . . . . . . . . . . . . 372
12.3 Binary Mixture in Equilibrium with a Pure Phase . . . . . . . . . . . . . . 374
12.4 Colligative Properties of a Dilute Solution . . . . . . . . . . . . . . . . . . 375
12.4.1 Freezing-point depression . . . . . . . . . . . . . . . . . . . . . . 377
12.4.2 Boiling-point elevation . . . . . . . . . . . . . . . . . . . . . . . . 380
12.4.3 Vapor-pressure lowering . . . . . . . . . . . . . . . . . . . . . . . 380
12.4.4 Osmotic pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
12.5 Solid–Liquid Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
12.5.1 Freezing points of ideal binary liquid mixtures . . . . . . . . . . . 383
12.5.2 Solubility of a solid nonelectrolyte . . . . . . . . . . . . . . . . . . 385
12.5.3 Ideal solubility of a solid . . . . . . . . . . . . . . . . . . . . . . . 386
12.5.4 Solid compound of mixture components . . . . . . . . . . . . . . . 386
12.5.5 Solubility of a solid electrolyte . . . . . . . . . . . . . . . . . . . . 389
12.6 Liquid–Liquid Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
12.6.1 Miscibility in binary liquid systems . . . . . . . . . . . . . . . . . 391
12.6.2 Solubility of one liquid in another . . . . . . . . . . . . . . . . . . 391
12.6.3 Solute distribution between two partially-miscible solvents . . . . . 394
12.7 Membrane Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
12.7.1 Osmotic membrane equilibrium . . . . . . . . . . . . . . . . . . . 395
12.7.2 Equilibrium dialysis . . . . . . . . . . . . . . . . . . . . . . . . . 395
12.7.3 Donnan membrane equilibrium . . . . . . . . . . . . . . . . . . . 396
12.8 Liquid–Gas Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
12.8.1 Effect of liquid pressure on gas fugacity . . . . . . . . . . . . . . . 399
12.8.2 Effect of liquid composition on gas fugacities . . . . . . . . . . . . 400
12.8.3 The Duhem–Margules equation . . . . . . . . . . . . . . . . . . . 404
12.8.4 Gas solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
12.8.5 Effect of temperature and pressure on Henry’s law constants . . . . 407
C ONTENTS 13
12.9 Reaction Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408

12.10 Evaluation of Standard Molar Quantities . . . . . . . . . . . . . . . . . . 410
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
13 The Phase Rule and Phase Diagrams 418

13.1 The Gibbs Phase Rule for Multicomponent Systems . . . . . . . . . . . . 418
13.1.1 Degrees of freedom . . . . . . . . . . . . . . . . . . . . . . . . . . 419
13.1.2 Species approach to the phase rule . . . . . . . . . . . . . . . . . . 419
13.1.3 Components approach to the phase rule . . . . . . . . . . . . . . . 421
13.1.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
13.2 Phase Diagrams: Binary Systems . . . . . . . . . . . . . . . . . . . . . . 425
13.2.1 Generalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
13.2.2 Solid–liquid systems . . . . . . . . . . . . . . . . . . . . . . . . . 426
13.2.3 Partially-miscible liquids . . . . . . . . . . . . . . . . . . . . . . . 430
13.2.4 Liquid–gas systems with ideal liquid mixtures . . . . . . . . . . . . 431
13.2.5 Liquid–gas systems with nonideal liquid mixtures . . . . . . . . . . 433
13.2.6 Solid–gas systems . . . . . . . . . . . . . . . . . . . . . . . . . . 436
13.2.7 Systems at high pressure . . . . . . . . . . . . . . . . . . . . . . . 439
13.3 Phase Diagrams: Ternary Systems . . . . . . . . . . . . . . . . . . . . . . 441
13.3.1 Three liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
13.3.2 Two solids and a solvent . . . . . . . . . . . . . . . . . . . . . . . 443
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
14 Galvanic Cells 449

14.1 Cell Diagrams and Cell Reactions . . . . . . . . . . . . . . . . . . . . . . 449
14.1.1 Elements of a galvanic cell . . . . . . . . . . . . . . . . . . . . . . 449
14.1.2 Cell diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
14.1.3 Electrode reactions and the cell reaction . . . . . . . . . . . . . . . 451
14.1.4 Advancement and charge . . . . . . . . . . . . . . . . . . . . . . . 451
14.2 Electric Potentials in the Cell . . . . . . . . . . . . . . . . . . . . . . . . . 452
14.2.1 Cell potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
14.2.2 Measuring the equilibrium cell potential . . . . . . . . . . . . . . . 454
14.2.3 Interfacial potential differences . . . . . . . . . . . . . . . . . . . 455
14.3 Molar Reaction Quantities of the Cell Reaction . . . . . . . . . . . . . . . 457
14.3.1 Relation between r Gcell and Ecell, eq . . . . . . . . . . . . . . . . 458
14.3.2 Relation between r Gcell and r G . . . . . . . . . . . . . . . . . 459
14.3.3 Standard molar reaction quantities . . . . . . . . . . . . . . . . . . 461
14.4 The Nernst Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
14.5 Evaluation of the Standard Cell Potential . . . . . . . . . . . . . . . . . . 464
14.6 Standard Electrode Potentials . . . . . . . . . . . . . . . . . . . . . . . . 464
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Appendix A Definitions of the SI Base Units 470
Appendix B Physical Constants 471
C ONTENTS 14
Appendix C Symbols for Physical Quantities 472
Appendix D Miscellaneous Abbreviations and Symbols 476

D.1 Physical States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
D.2 Subscripts for Chemical Processes . . . . . . . . . . . . . . . . . . . . . . 477
D.3 Superscripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478
Appendix E Calculus Review 479

E.1 Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
E.2 Partial Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
E.3 Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
E.4 Line Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
Appendix F Mathematical Properties of State Functions 481

F.1 Differentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
F.2 Total Differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
F.3 Integration of a Total Differential . . . . . . . . . . . . . . . . . . . . . . 483
F.4 Legendre Transforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484
Appendix G Forces, Energy, and Work 486

G.1 Forces between Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
G.2 The System and Surroundings . . . . . . . . . . . . . . . . . . . . . . . . 490
G.3 System Energy Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
G.4 Macroscopic Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
G.5 The Work Done on the System and Surroundings . . . . . . . . . . . . . . 495
G.6 The Local Frame and Internal Energy . . . . . . . . . . . . . . . . . . . . 495
G.7 Nonrotating Local Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
G.8 Center-of-mass Local Frame . . . . . . . . . . . . . . . . . . . . . . . . . 499
G.9 Rotating Local Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
G.10 Earth-Fixed Reference Frame . . . . . . . . . . . . . . . . . . . . . . . . 503
Appendix H Standard Molar Thermodynamic Properties 504
Appendix I Answers to Selected Problems 507
Bibliography 511
Index 520
B IOGRAPHICAL S KETCHES
Benjamin Thompson, Count of Rumford . . . . . . . . . . . . . . . . . . . . . . . . 63

James Prescott Joule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Sadi Carnot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Rudolf Julius Emmanuel Clausius . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
William Thomson, Lord Kelvin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Max Karl Ernst Ludwig Planck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Josiah Willard Gibbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Walther Hermann Nernst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
William Francis Giauque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Benoit Paul Émile Clapeyron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
William Henry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Gilbert Newton Lewis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Peter Josephus Wilhelmus Debye . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Germain Henri Hess . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
François-Marie Raoult . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Jacobus Henricus van’t Hoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
15
P REFACE TO THE S ECOND E DITION
This second edition of Thermodynamics and Chemistry is a revised and enlarged version
of the first 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 file of this book contains hyperlinks to pages, sections, equations, tables,
figures, 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
thermodynamics—the 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 classi-
cal 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 definitions—the intended meanings of words.
I welcome comments and suggestions for improving this book. My e-mail address ap-
pears below.
Howard DeVoe
hdevoe@umd.edu
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 one-
semester 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 difficult to grasp at first. 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 prop-
erties when appropriate.
In deriving equations, I have strived for rigor, clarity, and a minimum of mathematical complex-
ity. 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 Moffitt, 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
University–Bozeman; David W. Larsen, University of Missouri–St. 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 first 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 increases—the second law of

thermodynamics—holds, 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 Maxwell’s equations—then so much
the worse for Maxwell’s equations. If it is found to be contradicted by
observation—well, 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 one’s 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, 226–232 (1966)
“When I use a word,” Humpty Dumpty said, in rather a scornful tone,

“it means just what I choose it to mean—neither 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—
that’s 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 Système International d’Unités). The Physical Chemistry Division of the Inter-
national Union of Pure and Applied Chemistry, or IUPAC, produces a manual of recom-
mended 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 rec-
ommended 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 sufficient 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 official definitions of the base units are given
in Appendix A.
Table 1.2 lists derived units for some additional physical quantities used in thermody-
namics. The derived units have exact definitions 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 definitions 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 meter a m
mass kilogram kg
time second s
thermodynamic temperature kelvin K
amount of substance mole mol
electric current ampere A
luminous intensity candela cd
a or metre
Table 1.2 SI derived units
Physical quantity Unit Symbol Definition of unit

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

volume litera
L b
1 L D 1 dm3 D 10 3 m3
pressure bar bar 1 bar D 105 Pa
pressure atmosphere atm 1 atm D 101,325 Pa D 1:01325 bar
pressure torr Torr 1 Torr D .1=760/ atm D .101,325/760/ Pa
energy calorie c cal d 1 cal D 4:184 J
a or litre b or l c or thermochemical calorie d or calth
1.1 U NITS 21
Table 1.4 SI prefixes
Fraction Prefix Symbol Multiple Prefix Symbol

1
10 deci d 10 deka da
2
10 centi c 102 hecto h
3
10 milli m 103 kilo k
6
10 micro 106 mega M
9
10 nano n 109 giga G
12
10 pico p 1012 tera T
15
10 femto f 1015 peta P
18
10 atto a 1018 exa E
21
10 zepto z 1021 zetta Z
24
10 yocto y 1024 yotta Y
Any of the symbols for units listed in Tables 1.1–1.3, except kg and ı C, may be preceded
by one of the prefix symbols of Table 1.4 to construct a decimal fraction or multiple of the
unit. (The symbol g may be preceded by a prefix symbol to construct a fraction or multiple
of the gram.) The combination of prefix 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 3g
1 cm D 1 10 2m
1 cm3 D .1 10 2 m/3 D 1 10 6 m3
1.1.1 Amount of substance and amount

The physical quantity formally called amount of substance is a counting quantity for par-
ticles, such as atoms or molecules, or for other chemical entities. The counting unit is
invariably the mole, defined 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 official IUPAC definition. This definition is such that one mole of H2 O
molecules, for example, has a mass of 18:0153 grams (where 18:0153 is the relative molec-
ular mass of H2 O) and contains 6:02214 1023 molecules (where 6:02214 1023 mol 1 is
the Avogadro constant to six significant 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. [118]. 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 dimen-
sionless, 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 ı C to four significant digits in SI base units by the equation
D 9:970 102 kg m 3 (1.2.2)
and in different density units by the equation
D 0:9970 g cm 3 (1.2.3)
We may divide both sides of the last equation by 1 g cm 3 to obtain a new equation
3
=g cm D 0:9970 (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 figures 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
nRT
pD (1.2.5)
V
(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
.5:000 mol/.8:3145 J K 1 mol 1 /.298:15 K/
pD D 3:099 103 J m 3 (1.2.6)
4:000 m3
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 2 (1.2.7)
When we divide both sides of the first relation by 1 J and divide both sides of the second
relation by 1 N m 2 , we obtain the two new relations
2
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:
2
p D .3:099 103 J m 3/ .1 N m=J/ .1 Pa=N m /
D 3:099 103 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 prefixes), 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 definition of the torr in Table 1.3, the calculation becomes
p D .3:099 103 Pa/ .760 Torr=101; 325 Pa/

D 23:24 Torr (1.2.10)
3 Thisis the first equation in this book that, like many others to follow, shows conditions of validity in parenthe-
ses 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 dimen-
sions of a quantity used for a calculation, by checking for dimensional consistency. Here
are some rules that must be satisfied:
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 di-
mensionless
a quantity used as a power is dimensionless
In this book the differential of a function, such as df , refers to an infinitesimal quantity.
If one side of an equation is an infinitesimal 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
the integral T dT has the same dimensions as T 2
R
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 ı / and of p ı in
this expression? The quantity has the same dimensions as nRT (or energy) because the
logarithm is dimensionless. Furthermore, p ı in this expression has dimensions of pressure
in order to make the argument of the logarithm, p=p ı , dimensionless.
Example 3. Find the dimensions of the constants a and b in the van der Waals equation
nRT n2 a
pD
V nb 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, find the
dimensions of the constants a and b and express these dimensions in SI units.
(a) The Dieterici equation:
RT e .an=VRT /
pD
.V =n/ b
(b) The Redlich–Kwong equation:
RT an2
pD
.V =n/ b 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. Fi-
nally, several important concepts needed in later chapters are described: thermodynamic
states and state functions, independent and dependent variables, processes, and internal en-
ergy.
2.1 THE SYSTEM, SURROUNDINGS, AND BOUNDARY
Chemists are interested in systems containing matter—that 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 ma-
terial 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 flask 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 defined by its boundary, may change in time. In short, our choice of
the three-dimensional region that constitutes the system is arbitrary—but 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
influence 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 fixed 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 wall—one 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 field, such as a gravitational field.
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 classified as either extensive
or intensive. We can distinguish these two types of properties by the following considera-
tions.
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.
2 Theenergy in this definition of an isolated system is measured in a local reference frame, as will be explained
in Sec. 2.6.2.
2.1 T HE S YSTEM , S URROUNDINGS , AND B OUNDARY 29
Table 2.1 Symbols and SI units for some com-

mon properties
Symbol Physical quantity SI unit

E energy J
m mass kg
n amount of substance mol
p pressure Pa
T thermodynamic temperature K
V volume m3
U internal energy J
3
density kg m
Sometimes a more restricted definition of an extensive property is used: The property

must be not only additive, but also proportional to the mass or the amount when inten-
sive properties remain constant. According to this definition, 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 fluid), 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 region—it 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 defined 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 ra-
tio is an intensive property. For example concentration, defined as the ratio amount=volume,
is intensive. A mathematical derivative of one such extensive quantity with respect to an-
other is also intensive.
A special case is an extensive quantity divided by the mass, giving an intensive specific
quantity; for example
V 1
Specific volume D D (2.1.1)
m
If the symbol for the extensive quantity is a capital letter, it is customary to use the cor-
responding lower-case letter as the symbol for the specific quantity. Thus the symbol for
specific 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

V
Molar volume D D Vm (2.1.2)
n
V
Molar volume of substance i D D Vm;i (2.1.3)
ni
Molar volume of H2 O D Vm (H2 O) (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 directions—for example, a tall column of gas in
a gravitational field 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 infinite number of uniform phases, each of infinitesimal 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 fluid,
based on the phase’s response to an applied shear stress, and then use additional criteria
to classify a fluid as a liquid, gas, or supercritical fluid. 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 deformation—a 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
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 difficult
for adjacent layers to slide past one another, so that the solid resists deformation.
A fluid responds to shear stress differently, by undergoing continuous relative motion (flow)
of its parts. The flow 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 fixed deformation in a solid and contin-
uous flow in a fluid. 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 fluid. Over a sufficiently long time period, however, detectable flow is likely to occur
in any material under shear stress of any magnitude. Thus, the distinction between solid
and fluid 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 fluid-like behavior
over a longer period. Such materials, that resist deformation by a suddenly-applied shear
stress but undergo flow 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 find 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 ı C.
A phase transition of a pure substance is a change over time in which there is a con-
tinuous transfer of the substance from one phase to another. Eventually one phase can
supercritical
fluid
p b b
B b
C
liquid cp
solid A b b
D
gas
b
tp
Figure 2.2 Pressure–temperature 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 lo-
cus 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 equilib-
rium phase transition. For example, H2 O at 99:61 ı C and 1 bar can undergo an equilibrium
phase transition from liquid to gas (vaporization) or from gas to liquid (condensation). Dur-
ing 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 fluid as either a liquid or a gas. The distinction is important for a
pure substance because the choice determines the treatment of the phase’s standard state
(see Sec. 7.7). To complicate matters, a fluid at high pressure may be a supercritical fluid.
Sometimes a plasma (a highly ionized, electrically conducting medium) is considered a
separate kind of fluid state; it is the state found in the earth’s ionosphere and in stars.
In general, and provided the pressure is not high enough for supercritical phenomena
to exist—usually 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 fluid states of a pure substance in relation to a liquid–gas coexistence
curve and a critical point. If raising the temperature of a fluid at constant pressure causes
a phase transition to a second fluid phase, the original fluid was a liquid and the transition
occurs at the liquid–gas coexistence curve. This curve ends at a critical point, at which
all intensive properties of the coexisting liquid and gas phases become identical. The fluid
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 fluid.
The term vapor is sometimes used for a gas that can be condensed to a liquid by increas-
ing the pressure at constant temperature. By this definition, the vapor state of a substance
exists only at temperatures below the critical temperature.
The designation of a supercritical fluid 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 tempera-
ture or pressure in such a way that the substance changes from what we call a liquid to what
we call a supercritical fluid, 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 fluid 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 fluid mixture (instead of a pure substance) at a high pressure, it
may be difficult to classify the phase as either liquid or gas. The complexity of classification
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 fluid phases to become the less dense phase. In a gravitational field, the two
phases switch positions.
2.2.4 The equation of state of a fluid

Suppose we prepare a uniform fluid phase containing a known amount ni of each constituent
substance i, and adjust the temperature T and pressure p to definite known values. We
expect this phase to have a definite, fixed 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 fluid 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 fluid.
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 fluid has only one constituent (i.e., is a pure substance rather than a mixture), then
at a fixed T and p the volume is proportional to the amount. In this case, the equation of
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 Redlich–Kwong 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 theo-
retical as well as empirical foundation. There are two forms of virial equations for a pure
gas: one a series in powers of 1=Vm :

B C
pVm D RT 1 C C 2 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 coefficients, and the pa-
rameters Bp , Cp , : : : are a set of pressure virial coefficients. Their values depend on the
substance and are functions of temperature. (The first virial coefficient in both power se-
ries is 1, because pVm must approach RT as 1=Vm or p approach zero at constant T .)
Coefficients beyond the third virial coefficient are small and rarely evaluated.
The values of the virial coefficients 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 coef-
ficient 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 find the relation between the virial coefficients 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 2 C (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 2 C
Vm Vm
2 #
1 B
C Cp .RT /2 C 2 C C (2.2.5)
Vm Vm
Then we equate coefficients 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)
C D Bp RTB C Cp .RT /2 D .RT /2 .Bp2 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
RT
Vm CB (2.2.8)
p
(pure gas, p 1 bar)
The compression factor (or compressibility factor) Z of a gas is defined by
def pV pVm
Z D D (2.2.9)
nRT RT
(gas)
When a gas at a particular temperature and pressure satisfies the ideal gas equation, the
value of Z is 1. The virial equations rewritten using Z are
B C
Z D1C C 2 C (2.2.10)
Vm Vm
Z D 1 C Bp p C Cp p 2 C (2.2.11)
These equations show that the second virial coefficient 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 temper-
atures B and Bp are negative, and at higher temperatures they are positive—see 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
significant 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 36
1:2 50
710 K
1:0 0
0:8 50
B /cm3 mol 1
.
Z
0:6 400 K . 100
0:4 150
310 K
0:2 200
0 250
0 200 400 600 200 400 600 800 1000
p /bar T /K
(a) (b)
Figure 2.3 (a) Compression factor of CO2 as a function of pressure at three temper-
atures. At 710 K, the Boyle temperature, the initial slope is zero.
(b) Second virial coefficient 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 equa-
tion of state just as a fluid does, in which p is the pressure of a fluid 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 fiber. 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 ex-
erted on the object by the earth’s gravitational field. The classic balance has a beam and
knife-edge arrangement to compare the gravitational force on the body of interest with the
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 gravita-
tional 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 Sèvres, France.
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 defined 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, defined 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:
1
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 flasks. The National Institute of Standards and Technology in the
United States has established specifications 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 fills 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 defined 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 definition: it was the volume of 1 kilogram of
water at 3:98 ı C, the temperature of maximum density. This definition 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 defined may need a small correction in order to be
consistent with the present definition of the liter.
Table 2.2 Representative measurement methods
Physical Typical Approximate

Method
quantity value uncertainty
Mass analytical balance 100 g 0:1 mg
microbalance 20 mg 0:1 g
Volume pipet, Class A 10 mL 0:02 mL
volumetric flask, Class A 1L 0:3 mL
Density pycnometer, 25-mL capacity 1 g mL 1 2 mg mL 1
magnetic float densimeter 1 g mL 1 0:1 mg mL 1
vibrating-tube densimeter 1 g mL 1 0:01 mg mL 1
Pressure mercury manometer or barometer 760 Torr 0:001 Torr
diaphragm gauge 100 Torr 1 Torr
Temperature constant-volume gas thermometer 10 K 0:001 K
mercury-in-glass thermometer 300 K 0:01 K
platinum resistance thermometer 300 K 0:0001 K
monochromatic optical pyrometer 1300 K 0:03 K
2.3.3 Density
Density, an intensive property, is defined as the ratio of the two extensive properties mass
and volume:
def m
D (2.3.3)
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
M
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 fixed 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. Specifically, it is the normal component of stress exerted
by an isotropic fluid on a surface element.4 The surface can be an interface surface between
the fluid and another phase, or an imaginary dividing plane within the fluid.
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 field are examples of fluids that are not isotropic—that is, that
have different macroscopic properties in different directions.
D
B
S C
(a) (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 filled pycnometer is brought
to the desired temperature in a thermostat bath, dried, and weighed.
(b) Magnetic float 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-filled metal U-tube are clamped to
a stationary block. An oscillating magnetic field 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 col-
umn exactly 1 mm high of a fluid of density equal to exactly 13:5951 g cm 3 (the density
of mercury at 0 ı C) in a place where the acceleration of free fall has its standard value gn
(see Appendix B). One atmosphere is defined as exactly 1:01325 105 Pa (Table 1.3). The
torr is defined by letting one atmosphere equal exactly 760 Torr. One atmosphere is approx-
imately 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 ı . In the past, the value of p ı was usually taken to be 1 atm, but since
1982 the IUPAC has recommended the value p ı D 1 bar.
A variety of manometers and other devices is available to measure the pressure of a
fluid, each type useful in a particular pressure range. Some devices measure the pressure of
the fluid directly. Others measure the differential pressure between the fluid and the atmo-
sphere; the fluid pressure is obtained by combining this measurement with the atmospheric
pressure measured with a barometer.
Within a solid, we cannot define pressure simply as a force per unit area. Macroscopic
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 solid’s exterior surface. Thus, a solid immersed in a uniform isotropic fluid
of pressure p is at pressure p; if the fluid 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 definition. The chosen definition, 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 define a satisfactory scale with any
measuring method that satisfies 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 define 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 find 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 significance 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 first 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 defined by gas thermometry measurements, as described on page 42. The ther-
modynamic temperature scale is defined by the behavior of a theoretical Carnot engine,
as explained in Sec. 4.3.4. These temperature scales correspond to the physical quanti-
ties called ideal-gas temperature and thermodynamic temperature, respectively. Although
the two scales have different definitions, the two temperatures turn out (Sec. 4.3.4) to be
proportional to one another. Their values become identical when the same unit of temper-
ature is used for both. Thus, the kelvin is defined 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 body’s temperature. The
measured temperature is that of the body after thermal equilibrium is achieved.
6 Turner (Ref. [160]) argues that the “zeroth law” is a consequence of the first 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. [142]).
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 dif-
ferent 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 equilib-
rium 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 sufficient reproducibility for high-precision work. The temperature of the ice-water-
air system used to define 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 be-
cause the temperature of coexisting liquid and gas is a sensitive function of the experimental
pressure.
The obsolete centigrade scale was defined to give a value of exactly 0 degrees centi-
grade 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 Cel-
sius scale is the thermodynamic (or ideal-gas) temperature scale shifted by exactly 273:15
kelvins. The temperature unit is the degree Celsius (ı C), identical in size to the kelvin.
Thus, Celsius temperature t is related to thermodynamic temperature T by
t=ı C D T =K 273:15 (2.3.5)
On the Celsius scale, the triple point of water is exactly 0:01 ı C. The ice point is 0 ı C to
within 0:0001 ı C, and the steam point is 99:97 ı 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 defines 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 defined over a very wide temperature range, from 0:65 K up to at least
1358 K. There is a specified procedure for each measurement of T90 , depending on the
range in which T falls: vapor-pressure thermometry (0:65–5:0 K), gas thermometry (3:0–
24:5561 K), platinum-resistance thermometry (13:8033–1234:93 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 meth-
ods, it assigns values of several fixed calibration temperatures. The fixed temperatures are
achieved with the reproducible equilibrium systems listed in Table 2.3 on the next page.
7 Refs. [114] and [136].
Table 2.3 Fixed temperatures of the

International Temperature Scale of 1990
T90 =K Equilibrium system

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

If two different phases of a pure substance coexist at a controlled, constant pressure, the
temperature has a definite fixed value. As shown in Table 2.3, eight of the fixed tempera-
tures 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 tem-
perature and pressure have definite fixed values in a system containing coexisting solid, liq-
uid, and gas phases of a pure substance. The table lists six temperatures fixed 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
definition, is exactly 273:16 K.
Gas thermometry
Only the triple point of water has a defined value on the thermodynamic temperature scale.
How are the values of other fixed 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
H2 O(g)
H2 O(l)
H2 O(s)
H2 O(l)
thermometer bulb
Figure 2.5 Cross-section of a water triple-point cell. The cell has cylindrical sym-
metry. 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 film 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
h
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
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
T2 D T1 lim (2.3.6)
p1 !0 p1 V1
(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 coefficient 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
T2 D lim (2.3.7)
p2 !0 R
(gas)
Constant-volume gas thermometry can be used to evaluate the second virial coefficient
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 Stan-
dards 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 sen-
sitivity (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 in-
stant of time, the system is in some definite 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 define the state at that instant. Whenever the value of any
of these properties changes, the state has changed. If we subsequently find 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.
Table 2.4 Values of state functions of an aqueous su-

crose solution (A = water, B = sucrose)
temperature T D 293:15 K
pressure p D 1:01 bar
amount of water nA D 39:18 mol
amount of sucrose nB D 1:375 mol
volume V D 1000 cm3
mass m D 1176:5 g
3
density D 1:1765 g cm
mole fraction of sucrose xB D 0:03390
osmotic pressure ˘ D 58:2 bar
refractive index, sodium D line nD D 1:400
Various conditions determine what states of a system are physically possible. If a uni-
form 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 define 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 definite, 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 first four properties (T , p, nA , nB ) are ones that we can vary independently, and
their values suffice to define 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 definite value listed—we cannot alter any of the other properties
without changing one or more of the first 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.
We could make other choices of the independent variables for the aqueous sucrose sys-
tem. 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. Sup-
pose the sucrose solution is in a closed system with fixed, 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 vari-
ables, as we can see by the fact that the state is completely defined 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 fixed 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 com-
ponent. The number of components of a system is the minimum number of substances or
mixtures of fixed composition needed to form each phase.10 A system of a single pure sub-
stance 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 num-
ber 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 condi-
tions 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 defines 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.
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 ı C and 6 ı C, so these values would not uniquely
define 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 or parcels,
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., flow), 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 flow in a fluid or a shock wave in
a gas may involve inhomogeneities all the way down to the molecular scale. Macroscopic
variables would not suffice to define 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
indefinitely 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 ı 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
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 fluid (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 influence of an external
field. Here are five examples of such states.
A system with an internal adiabatic partition separating two phases can be in an equi-
librium state that is not in thermal equilibrium. The adiabatic partition allows the two
phases to remain indefinitely 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 372)
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 influenced by an external field is a tall column of gas in a
gravitational field (Sec. 8.1.4). In order for an equilibrium state to be established in
this field, the pressure must decrease continuously with increasing elevation.
Keep in mind that regardless of the presence or absence of internal constraints and
external fields, 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
definition 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 finite 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
T D 400 K T D 300 K
T D 350 K
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 metastable—the 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 infinite.
2.4.5 Steady states

It is important not to confuse an equilibrium state with a steady state, a state that is con-
stant 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 figure 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 definite initial
state and ending with a definite final state. The process is defined 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
p p p
V V V
(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 final 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 defined 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 field. 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 finite 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 final 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 final states of the system, not on the path of the process.
An infinitesimal change of the state function X is written dX. The mathematical op-
eration of summing an infinite number of infinitesimal changes is integration, and the sum
is an integral (see the brief calculus review in Appendix E). The sum of the infinitesimal
2.6 T HE E NERGY OF THE S YSTEM 52
changes of X along a path is a definite integral equal to X :

Z X2
dX D X2 X1 D X (2.5.1)
X1
If dX obeys this relation—that is, if its integral for given limits has the same value regardless
of the path—it 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 final 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 infinites-
imal quantity of a path function. Thus, ¶q and ¶w are infinitesimal quantities of heat and
work. The sum of many infinitesimal 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 infinitesimal 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 definite elevation above sea level. During a climb from the
trailhead to the summit, the hiker’s 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 first law of thermody-
namics (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 ıq and ıw.
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 thermody-
namic system.
In classical mechanics, the energy of a collection of interacting point particles is the
sum of the kinetic energy 12 mv 2 of each particle (where m is the particle’s mass and v is
its velocity), and of various kinds of potential energies. The potential energies are defined
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 addi-
tional contributions to the total energy.
The positions and velocities of particles must be measured in a specified 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 Newton’s laws of motion are obeyed (see Sec. G.1 in Appendix G).
A reference frame whose axes are fixed relative to the earth’s surface is what this book
will call a lab frame. A lab frame for all practical purposes is inertial (Sec. G.10 on
page 503). 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 specified
inertial frame. The system could be located in a weightless environment in outer space, and
the inertial frame could be one that is either fixed or moving at constant velocity relative to
local stars. Usually, however, the system is located in the earth’s gravitational field, and the
appropriate inertial frame is then an earth-fixed 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 function—that 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-fixed lab frame.
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 fixed amount of water contained in a glass beaker. (The glass material of the beaker
is part of the surroundings.) We can define 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 fixed 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
earth’s gravitational field 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 define 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 system’s 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 fixed with respect to the container.
Is it possible to determine a numerical value for the internal energy of a system? The
total energy of a body of mass m when it is at rest is given by the Einstein relation E D mc02 ,
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 mea-
surement methods listed in Table 2.2 on page 38, using the typical values shown. For each of
the five physical quantities listed, which measurement method has the smallest relative uncer-
tainty?
2.3 Table 2.5 lists data obtained from a constant-volume gas thermometer containing samples of
Table 2.5 Helium at a fixed temperature
.1=Vm / =102 mol m 3

.p2 Vm =R/ =K
1.0225 2.7106
1.3202 2.6994
1.5829 2.6898
1.9042 2.6781
2.4572 2.6580
2.8180 2.6447
3.4160 2.6228
3.6016 2.6162
4.1375 2.5965
4.6115 2.5790
5.1717 2.5586
varying amounts of helium maintained at a certain fixed 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 coefficient at that temperature.
Use these data and Eq. 2.2.2 on page 34 to evaluate T2 and the second virial coefficient of he-
lium at temperature T2 . (You can assume the third and higher virial coefficients are negligible.)
2.4 Discuss the proposition that, to a certain degree of approximation, a living organism is a steady-
state system.
2.5 The value of U for the formation of one mole of crystalline potassium iodide from its el-
ements at 25 ı 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 repro-

ducible 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 first 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 first 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 first law, and U D q C w is

the integrated form.
The heat and work appearing in the first law are two different modes of energy transfer.
They can be defined 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 infinitesimal quantity of energy transferred as heat at a surface element of the bound-
ary is written ¶q, and a Rfinite quantity is written q (Sec. 2.5). To obtain the total finite 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 infinitesimal quantity of work is ¶w, and a finite 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 first-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 first 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 con-
duction with good thermal insulation at the boundary, we can eliminate conduction and
convection with a vacuum gap, and we can minimize radiation with highly reflective sur-
faces 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 collec-
tion of interacting particles moving according to the principles of classical mechanics. The
analysis shows how we should evaluate mechanical thermodynamic work. Suppose the dis-
placement 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
¶w D Fxsur dx wD Fxsur dx (3.1.1)
x1
Here Fxsur 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 infinitesimal 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.
The kind of force represented by Fxsur is a short-range contact force. Appendix G shows
that the force exerted by a conservative time-independent external field, such as a gravita-
tional force, should not be included as part of Fxsur . 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).
Newton’s 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
¶w D Fx dx wD Fx dx (3.1.2)
x1
sys
where is the component in the Cx direction of the contact force exerted by the system
Fx
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 system’s surroundings is a change in
the elevation of a weight in the surroundings. The weight must be one that is linked
mechanically to the source of the force Fxsur . Then, provided the local frame is a sta-
tionary lab frame, the work is equal in magnitude and opposite in sign to the change
in the weight’s potential energy: w D mgh where m is the weight’s mass, g is the
acceleration of free fall, and h is the weight’s elevation in the lab frame. This inter-
pretation 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 weight’s 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,
Joule’s 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 99).
In the first-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 find 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 find 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 frame—one
whose origin moves with the system’s center of mass and whose axes do not rotate relative
to the lab frame:5
2
w D wlab 12 m vcm

mgzcm (3.1.4)
4 Sec. G.7. 5 Eq. G.8.12 on page 502.
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 21 mvcm and Ep D mgzcm are the kinetic and potential energies of the system
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 system’s 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 fixed 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 coefficients and work coordinates

If a process has only one kind of work, it can be expressed in the form
Z X2
¶w D Y dX or wD Y dX (3.1.7)
X1
where Y is a generalized force called a work coefficient and X is a generalized displace-

ment called a work coordinate. The work coefficient 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 coefficient is p and the work coordinate is V .
6 Eq. G.7.3 on page 499.
m 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 coefficient
and conjugate work coordinate. In this case the work can be expressed as a sum over the
different kinds labeled by the index i:
X XZ Xi;2
¶w D Yi dXi or wD Yi dXi (3.1.8)
i i Xi;1
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 defined by T1 D 300:0 K, and each has the same final state.
Experiment 1: We surround the system with thermal insulation as shown in the figure 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 final 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 final temperature. Assume the final temperature
is T2 D 300:10 K, an increase of 0:10 kelvins.
7 Thiscalculation 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.
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 final 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 final 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. En-
ergy can now enter the system in the form of heat, and does so because of the temper-
ature gradient at the boundary. By a substitution of heat for mechanical or electrical
work, the system changes to the same final 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 final 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 first 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 flows 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 flow,” 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.
3.2 S PONTANEOUS , R EVERSIBLE , AND I RREVERSIBLE P ROCESSES 62
The notion of heat as an indestructible substance was the essence of the caloric the-
ory. This theory was finally 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 defined as the ratio of an infinitesimal quantity of
heat transferred across the boundary under specified conditions and the resulting infinitesi-
mal temperature change:
def ¶q
heat capacity D (3.1.9)
dT
(closed system)
Since q is a path function, the value of the heat capacity depends on the specified 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. It can be defined as the kinetic
energy of random translational motions of atoms and molecules relative to the local frame,
plus the vibrational and rotational energies of molecules. The thermal energy of a body
or phase depends on its temperature, and increases when the temperature increases. The
thermal energy of a system is a contribution to the internal energy.
It is important to understand that a change of the system’s thermal energy during a
process is not necessarily the same as energy transferred across the system boundary as
heat. The two quantities are equal only if the system is closed and there is no work, volume
change, phase change, or chemical reaction. This is illustrated by the three experiments
described in Sec. 3.1.3: the thermal energy change is the same in each experiment, but only
in experiment 3 is the work negligible and the thermal energy change equal to the heat.
3.2 SPONTANEOUS, REVERSIBLE, AND IRREVERSIBLE PROCESSES
A spontaneous process is a process that can actually occur in a finite 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 define entropy changes in the next chapter,
BIOGRAPHICAL SKETCH
Benjamin Thompson, Count of Rumford (1753–1814)
Being engaged, lately in superintending the bor-
ing 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 exper-
iment,) of the metallic chips separated from it by
the borer. . .
By meditating on the results of all these ex-
periments, we are naturally brought to that great
question which has so often been the subject of
Benjamin Thompson, whose career was re- speculation among philosophers; namely,
markably varied and colorful, collected exper- What is Heat?—Is there any such thing as an
imental evidence of the falseness of the caloric igneous fluid?—Is there any thing that can with
theory—the concept that heat is a material sub- propriety be called caloric?. . .
stance. He was a complex man: energetic, ego- And, in reasoning on this subject, we must
tistical, domineering, and misanthropic. not forget to consider that most remarkable cir-
cumstance, that the source of the heat generated
Thompson was born into a farming fam-
by friction, in these experiments, appeared evi-
ily in Woburn, Massachusetts. He married a
dently to be inexhaustible.
wealthy widow and was admitted into fashion- It is hardly necessary to add, that any thing
able society. At the time of the American Rev- which any insulated body, or system of bod-
olution he was accused of being a loyalist, and ies, can continue to furnish without limitation,
at the age of 23 fled to England, abandoning cannot possibly be a material substance: and it
his wife and daughter. He was an Under Sec- appears to me to be extremely difficult, if not
retary of State in London, returned briefly to quite impossible, to form any distinct idea of any
America as a British cavalry commander, and thing, capable of being excited and communi-
then spent 11 years as a colonel in the Bavar- cated, in the manner the heat was excited and
ian army. In Bavaria, to reward his success in communicated in these experiments, except it be
MOTION.
reorganizing the army and reforming the so-
cial welfare system, he was made a Count of Rumford thought of heat in a solid as har-
the Holy Roman Empire. He chose the name monic vibrations similar to acoustic waves, not
Rumford after the original name of Concord, as random motion or as a form of energy as
New Hampshire, his wife’s home town. later developed by James Joule.
While in Bavaria, Count Rumford carried Rumford also made investigations into bal-
out the cannon-boring experiments for which listics, nutrition, thermometry, light, and fabric
he is best known. The caloric theory held that properties. He invented the Rumford fireplace
heat is a kind of indestructible fluid (“caloric”) and the drip coffee percolator. After living in
that is held in the spaces between the atoms London for fourteen years, he settled in Paris
of a body. Frictional forces were supposed in 1804. The following year, his first wife hav-
to cause a rise in temperature by squeezing ing died in America, he married the widow of
caloric fluid out of a body. Rumford’s experi- the famous French chemist Antoine Lavoisier.
ments involved boring into a horizontally-fixed The marriage was stormy and they soon sepa-
cannon barrel with a blunt steel bit turned by rated.
horse power. He reported the results in 1798:a
a Ref. [147].
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 final 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 film of the events of the process. Each frame of the film is a “snapshot”
picture of the state at one instant. If you run the film 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
sufficiently 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 infinite
slowness.
This book has many equations expressing relations among heat, work, and state func-
tions 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 sponta-
neous 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.)
The sequence of equilibrium states can be approximated, as closely as desired, by
the intermediate states of a real spontaneous process carried out sufficiently 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 suffi-
ciently 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 infinite 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 both—the 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 coefficient Y
in the expression ¶w D Y dX must be finite (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 field, to reverse the direction of the
process.
In the reversible limit, dissipative effects within the system such as internal friction
vanish.
When any infinitesimal step of a reversible process takes place in reverse, the magni-
tudes 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 finite 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 finite rate while the tem-
perature and pressure magically stay perfectly uniform throughout the system. This is an
entirely imaginary process, because there is no temperature or pressure gradient—no phys-
ical “driving force”—that would make the change tend to occur in a particular direction.
This imaginary process is a reversible process—one 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.
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 any-
where. This statement, however, is misleading, because during the period in question
spontaneous changes inevitably occur outside the system. At the very least, the ex-
ternal 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.
(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.
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 film 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 film of the experiment. Then when you run the film 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
difficult 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, spon-
taneous but not irreversible. These are purely mechanical processes involving the motion of
perfectly-elastic macroscopic bodies without friction, temperature gradients, viscous flow,
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 flywheel with
frictionless bearings rotating in a vacuum.
A purely mechanical process proceeding at a finite rate 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 H EAT T RANSFER 67
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 flow, 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 con-
trol. 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 final uniform temper-
ature 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 flow 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 flows 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 dif-
ferent 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 thermally-
insulated jacket while it is in one of these states, the state would change as the temperature
gradient rapidly disappears. Thus, the intermediate states of the spontaneous heating pro-
cess are not equilibrium states, and the rapid heating process is not reversible.
To make the intermediate states more nearly uniform in temperature, with smaller tem-
perature gradients, we can raise the temperature of the bath at a slower rate. The sequence
of states approached in the limit of infinite slowness is indicated in Fig. 3.3(b). In each in-
termediate 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
suffices to define 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.
3.3 H EAT T RANSFER 68
101:2
301:2 101:2
301:2 101:2
301:2
final state final state initial state
101:0
301:0 101:0
301:0 101:0
301:0
100:8
300:8 10 s 100:8
300:8 100:8
300:8
8s 2s
100:6
300:6 100:6
300:6 100:6
300:6
T =K
T =K
T =K
6s 4s
100:4
300:4 100:4
300:4 100:4
300:4
4s 6s
100:2
300:2 100:2
300:2 100:2
300:2 10 s
2s 8s
100:0
300:0 100:0
300:0 100:0
300:0
initial state initial state final state
299:8
99:8 299:8
99:8 299:8
99:8
0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6
r=cm r=cm r=cm
(a) (b) (c)
Figure 3.3 Temperature profiles 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 infinitely slowly.
(c) Bath temperature lowered at the rate of 0:10 K s 1 .
In any real heating process occurring at a finite rate, the sphere’s 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 sponta-
neous heating process—the gradients are in opposite directions. It is physically impossible
for the sequence of states of either process to occur in the reverse chronological order, for
that would have thermal energy flowing in the wrong direction along the temperature gra-
dient. 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 recover-
ability 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
3.4 D EFORMATION W ORK 69
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. 3.5–3.8) 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 Fsur cos ˛ ds (3.4.1)
where Fsur is the magnitude of the contact force exerted by the surroundings on the surface
element, ds is the infinitesimal 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
Eq. 3.1.1 on page 57: ¶w D Fxsur dx.
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 arrange-
ment shown in Fig. 3.4 on the next page. A sample of gas is confined 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 system’s 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 find 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 piston’s velocity when the piston moves.
The friction occurs at the seal between the edge of the piston and the cylinder wall.
We will assume this seal is lubricated, and that Ffric approaches zero as the piston velocity
approaches zero.
8 From Eq. G.6.10 on page 498.
Ffric
Fgas
Fext
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.
Let pb be the average pressure of the gas at the piston—that 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
Newton’s 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 confine 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
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 nonuni-
form, 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 uni-
form are the temperature and pressure. In the limit of infinite slowness, the gas passes
through a continuous sequence of equilibrium states of uniform temperature and pressure.
Let p2 be the pressure in the final state of the infinitely-slow expansion. In this state,
Fext is equal to p2 As . By increasing Fext from this value, we cause spontaneous compres-
sion 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 infinite
slowness. The sequence of equilibrium states, taken in either direction, is a reversible pro-
cess.
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 molecule’s
velocity as it approaches the piston be vx (a positive value). After the collision, the x
component of the molecule’s 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 molecule’s 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:
nM ˝ 2 ˛
2u hvx i C u2

pb D vx (3.4.4)
V
(The angle brackets denote
˝ ˛ averages.) The pressure at the stationary wall of the cylin-
der is p D .nM=V / vx2 . Accordingly, the pressure at the moving piston is given
by9 !
hvx i u2
pb D p 1 2u ˝ 2 ˛ C ˝ 2 ˛ (3.4.5)
vx vx
1=2
˝From
2
˛ kinetic-molecular theory, the averages are given by hvx i D .2RT =M / and
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
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 cylinder-
and-piston device of Fig. 3.4. This kind of deformation work, for both expansion and com-
pression, 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.
9 An equivalent formula is derived in Ref. [9], Eq. 7. A formula that yields similar values of pb appears in Ref.
[14].
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 infinitesimal change dx changes the volume by dV D As dx.
The infinitesimal quantity of work for an infinitesimal volume change is then given by
¶w D pb dV (3.4.7)
(expansion work,
closed system)
and the finite work for a finite volume change, found by integrating from the initial to the
final 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 surround-
ings. During compression (negative dV ), ¶w is positive and the surroundings do work on
the system.
When carrying out dimensional analysis, you will find 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 480). 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 sufficiently slow, we can with
little error assume that the gas has a uniform pressure p throughout, and that the work can
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

Z V2
wD p dV (3.4.10)
V1
(reversible expansion
work, closed system)
Fsur ds
As;
Fsys
(a) (b)
Figure 3.5 Deformation of an isotropic phase (shaded) confined by a wall.

(a) Equal and opposite forces exerted by the surroundings and system at surface ele-
ment (thick curve) of the system boundary.
(b) Change from initial volume (dotted curve) to a smaller volume.
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 specific 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

R device like
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
R is the gas the correct general expression is the one given by Eq. 3.4.8:
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 ofRthe surroundings.
R
In other words, if the integrals Fgas dx and Fext dx have different values, R it is the
first 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 re-
versibly. This is because in the limit of infinite 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
R and is stationary in the
R initial and final states,
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
R to Eq. G.1.5 R on page 488.) In the general irreversible case, however, the
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 fluid or solid phase, and various portions of its boundary undergo displacements in
different directions. Figure 3.5 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
10 For an informative discussion of this topic see Ref. [6]; also comments in Refs. [29], [7], [95], [8], and [121];
also Ref. [92].
3.5 A PPLICATIONS OF E XPANSION W ORK 74
the phase remains uniform. Consider the surface element of the boundary, with area
As; , indicated in the figure by a short thick curve. Because the phase is isotropic, the
sys
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 Fsur exerted by the
sys
surroundings on the system is equal in magnitude to F and is directed in the opposite
direction. The volume change for an infinitesimal 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 find the total reversible expan-
sion 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 displace-
ments of a confining wall, or for a volume change caused by slow temperature changes
at constant pressure. It is valid if the system is an isotropic fluid phase in which other
phases are immersed, provided the fluid phase contacts all parts of the system boundary.
The expression is not necessarily valid for an anisotropic fluid 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 coefficient, p, and the infinitesimal 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 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 definition 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
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 first property, it must also have the second.
On the molecular level, a gas with negligible intermolecular interactions12 fulfills both
of these requirements. Kinetic-molecular theory predicts that a gas containing noninteract-
ing 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
dV V
w D nRT D nRT ln 2 (3.5.1)
V1 V V1
(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 fixed 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 fixed amount of an ideal gas depends only on its temperature (Sec.
3.5.1), an infinitesimal change dT will cause a change dU that depends only on T and dT :
dU D f .T / dT (3.5.2)
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 first 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.
12 This
book uses the terms “intermolecular interactions” and “intermolecular forces” for interactions or forces
between either multi-atom molecules or unbonded atoms.
In a reversible adiabatic expansion with expansion work only, the heat is zero and the
first 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:

nRT
CV dT D dV (3.5.6)
V
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 final states, we obtain
Z T2 Z V2
dT dV
CV D nR (3.5.7)
T1 T V1 V
T2 V2
CV ln D nR ln (3.5.8)
T1 V1
We can rearrange this result into the form
nR=CV
T nR V2 V1
ln 2 D ln D ln (3.5.9)
T1 CV V1 V2
and take the exponential of both sides:
nR=CV
T2 V1
D (3.5.10)
T1 V2
The final temperature is then given as a function of the initial and final volumes by
nR=CV
V1
T2 D T1 (3.5.11)
V2
(reversible adiabatic
expansion, ideal gas)
This relation shows that the temperature decreases during an adiabatic expansion and in-
creases during an adiabatic compression, as expected from expansion work on the internal
energy.
To find 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)
p / bar
4
800
K
600
2 K
400 K
200 K
0
0 0:10 0:20 0:30 0:40
V = dm3
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).
To express the final pressure as a function of the initial and final volumes, we make the
substitutions T1 D p1 V1 =nR and T2 D p2 V2 =nR in Eq. 3.5.11 and obtain
nR=CV
p2 V 2 p V V1
D 1 1 (3.5.13)
nR nR V2
Solving this equation for p2 , we obtain finally

1CnR=CV
V1
p2 D p1 (3.5.14)
V2
The solid curve in Fig. 3.6 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 figure 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 tempera-
tures. 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 flow of heat into the system.
3.5.4 Indicator diagrams

An indicator diagram (or pressure–volume 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 fluid
R that
is essentially uniform. The area under the curve has the same value asR the integral p dV ,
which is the negative of the reversible expansion work given by w D p dV . For example,
Figure 3.7 Indicator with paper-covered roll at left and pressure gauge at right. a
a Ref. [111], page 104.
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.
More generally, an indicator diagram can be a plot of a work coefficient or its negative
as a function of the work coordinate. For example, it could be a plot of the pressure
R pb at
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 instru-

ment 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). 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 Ris less than the
pressure at the stationary wall. Consequently the work given by w D pb dV is less
3.6 W ORK IN A G RAVITATIONAL F IELD 79
gas vacuum gas gas
Figure 3.8 Free expansion into a vacuum.
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
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 equi-
librium 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 WORK IN A GRAVITATIONAL FIELD
Figure 3.9 on the next page depicts a spherical body, such as a glass marble, immersed in
a liquid or gas in the presence of an external gravitational field. 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 body’s elevation above the bottom of the vessel. All displacements are parallel to
the vertical z axis. From Eq. 3.1.1, the work is given by ¶w D Fzsur dz where Fzsur 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 Sec. 3.1.1, this force does not contribute to Fzsur .
Consider first the simple process in Fig. 3.9(a) in which the body falls freely through
the fluid. This process is clearly spontaneous. Here are two choices for the definition of the
system:
The system is the combination of the spherical body and the fluid. The system bound-
ary is where the fluid contacts the air and the vessel walls. Because there is no dis-
3.6 W ORK IN A G RAVITATIONAL F IELD 80
Fstr
Fbuoy Ffric Fbuoy Ffric
mg mg
z z
fluid fluid
(a) (b)
Figure 3.9 Spherical body (shaded) in a gravitational field. The arrows indicate the
directions and magnitudes of contact and gravitational forces exerted on the body.
(a) The body falls freely through a fluid in a vessel.
(b) The body is lowered on a string through the fluid.
placement 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 first law tells us the system’s 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 fluid is in the surroundings. The upward components
of the forces exerted on the body are (1) a gravitational force mg, where m is the
body’s mass and g is the acceleration of free fall; (2) a buoyant force13 Fbuoy D V 0 g,
where is the fluid 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
gravitational force is not included in Fzsur . Therefore the gravitational work is given
by
¶w D Fzsur dz D Fbuoy C Ffric dz

(3.6.1)
and is negative because ˇ dz is negative:
ˇ the body as it falls does work on the fluid.
The positive quantity ˇFbuoy dz ˇ is the work of moving displaced fluid 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 fluid. The moving part of the boundary
is at the point where the string is attached to the body. The force exerted here by the
string is an upward force Fstr , and the gravitational work is given by ¶w D Fzsur dz D
13 The buoyant force is a consequence of the pressure gradient that exists in the fluid in a gravitational field (see
Sec. 8.1.4). We ignore this gradient when we treat the fluid as a uniform phase.
3.7 S HAFT W ORK 81
Fstr dz. According to Newton’s 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/R dz is an infinitesimal change of
the body’s kinetic energy Ek ,14 so that the integral .m dv= dt/ dz is equal to Ek .
The finite 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 con-
trolled from the surroundings. As v approaches zero from either direction, Ffric ap-
proaches zero and the work approaches the reversible limit w D .mg Fbuoy /z.
(If the fluid 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 writ-
ten 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.
The system is the body only. In this case, Fzsur 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 (3.6.5)
For a process that begins and ends in equilibrium states, Ek is zero and the finite
work is w D mgz, unaffected by the velocity v during the process. The expressions
for infinitesimal and finite 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 fluid, 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.
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.
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.15 The sign con-
vention 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 finite 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 find that sys is
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
Z #2 Z #2
wD b d# D sys d# (3.7.1)
#1 #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 finite: sys D mgr where r is the radius of the shaft
at the point where the cord is attached. When ! is finite 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) on the next page shows how the shaft
work given by Eq. 3.7.1 depends on the angular velocity for a fixed value of j#2 #1 j. The
variation of w with ! is due to the frictional forces. System A has finite, reversible shaft
work in the limit of infinite slowness (!!0) given by w D mgr#. The shaft work is
least positive or most negative in the reversible limit.

14 To 1 2
prove this, we write m.dv= dt / dz D m.dz= dt/ dv D mv dv D d 2 mv D dEk .
15 A torque is a moment of tangential force with dimensions of force times distance.
3.7 S HAFT W ORK 83
w w
bc
! bc
!
bc
(a) (b)
Figure 3.11 Shaft work w for a fixed magnitude of shaft rotation # as a function
of the angular velocity ! D d#= dt. The open circles indicate work in the limit of
infinite slowness. (a) System A of Fig. 3.10. (b) System B of Fig. 3.10.
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 finite 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 fixed value of j#2 #1 j depends on ! in the way shown in
Fig. 3.11(b). The work is positive for finite values of ! of either sign, and approaches zero
in the limit of infinite slowness.
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, 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 coefficient
3.7 S HAFT W ORK 84
opening for
thermometer
MIRKO JUNGE / COMMONS.WIKIMEDIA.ORG

water
(a) (b)
Figure 3.12 Joule paddle wheel.

(a) Joule’s 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.
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.
Joule’s apparatus contained the paddle wheel shown in Fig. 3.12. It consisted of eight
sets of metal paddle arms attached to a shaft in a water-filled copper vessel. When the shaft
rotated, the arms moved through openings in four sets of stationary metal vanes fixed 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 flow) in the water
and an increase in the temperature of the entire assembly.
The complete apparatus is depicted in Fig. 3.13 on page 86. 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 first approximation, this work is the negative of the change of the weights’
3.7 S HAFT W ORK 85
BIOGRAPHICAL SKETCH
James Prescott Joule (1818–1889)
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 refined measure-
ment in 1850.a
In a note dated 1885 in his Collected Pa-
pers, Joule wrote:
It was in the year 1843 that I read a paper “On
the Calorific Effects of Magneto-Electricity and
the Mechanical Value of Heat” to the Chemical
Section of the British Association assembled at
James Joule drove the final nails into the cof- Cork. With the exception of some eminent men
fin of the caloric theory by his experimental . . . the subject did not excite much general at-
demonstrations of the mechanical equivalent tention; so that when I brought it forward again
of heat. at the meeting in 1847, the chairman suggested
that, as the business of the section pressed, I
Joule (pronounced like “jewel”) was born
should not read my paper, but confine myself to a
in Salford, near Manchester, England. His fa-
short verbal description of my experiments. This
ther was a prosperous brewery owner; after his I endeavoured to do, and discussion not being
death, James and one of his brothers carried on invited, the communication would have passed
the business until it was sold in 1854. without comment if a young man had not risen
Joule was a sickly child with a minor spinal in the section, and by his intelligent observations
weakness. He was tutored at home, and at the created a lively interest in the new theory. The
age of 16 was a pupil of the atomic theory ad- young man was William Thomson, who had two
vocate John Dalton. years previously passed the University of Cam-
As an adult, Joule was a political conser- bridge with the highest honour, and is now prob-
vative and a member of the Church of Eng- ably the foremost scientific authority of the age.
land. He dressed plainly, was of a somewhat The William Thomson mentioned in Joule’s
nervous disposition, and was a poor speaker. note later became Lord Kelvin. Thomson de-
He was shy and reserved unless with friends, scribed introducing himself to Joule after the
had a strong sense of humor, and loved nature. 1847 meeting, which was in Oxford, as a re-
Joule never attended a university or had a sult of which the two became collaborators and
university appointment, but as an “amateur” life-long friends. Thomson wrote:b
scientist and inventor he published over 100 Joule’s paper at the Oxford meeting made a great
papers (some of them jointly with collabo- sensation. Faraday was there and was much
rators) and received many honors. He in- struck with it, but did not enter fully into the new
vented arc welding and a mercury displace- views. It was many years after that before any of
ment pump. He carried out investigations on the scientific chiefs began to give their adhesion.
electrical heating and, in collaboration with According to a biographer:c
William Thomson, on the cooling accompany- His modesty was always notable. ‘I believe,’ he
ing the expansion of a gas through a porous told his brother on 14 Sept. 1887, ‘I have done
plug (the Joule–Thomson experiment). The two or three little things, but nothing to make a
joule, of course, is now the SI derived unit of fuss about.’ During the later years of his life he
energy. received many distinctions both English and for-
Joule’s best-known experiment was the de- eign.
a Ref. [84]. b Ref. [17]. c Ref. [66].
3.8 E LECTRICAL W ORK 86
F G
D
C
b
H I b
J K
A
10 cm
Figure 3.13 Joule’s apparatus for measuring the mechanical equivalent of heat (re-
drawn from a figure in Ref. [84]).
Key: A—paddle wheel and vessel (see Fig. 3.12); B—wood thermal insulator; C—pin
used to engage paddle wheel shaft to roller; D—roller; E—crank used to wind up the
weights; F, G—strings; H, I—pulley wheels; J, K—weights (round lead disks, viewed
here edge-on).
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
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 99. His
results for the mechanical equivalent of heat, based on 40 such experiments at average
temperatures in the range 13 ı C–16 ı 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 ı C calorie,” the energy needed to raise the temperature
of one gram of water from 14:5 ı C to 15:5 ı C.16
3.8 ELECTRICAL WORK
The electric potential at a point in space is defined as the work needed to reversibly move
an infinitesimal test charge from a position infinitely 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
16 Thethermochemical calorie (cal), often used as an energy unit in the older literature, is defined as 4:184 J.
Thus 1 kcal D 4:184 kJ.
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
charge of a proton). Thus the electrical potential energy of an electron is R e in the right
conductor and L e in the left conductor. The difference in the energies of an electron in
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 infinitesimal
negative charge. During a period of time in which an infinitesimal 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 , be-
cause Qsys is a path function.) This internal energy change is called electrical work. Thus
the general formula for an infinitesimal 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 defined by

def
D R L (3.8.2)
Note that in the expression .R ¶Qsys L ¶Qsys / for the energy difference, the term
R ¶Qsys does not represent the energy transferred across the boundary at the right
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
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 justification 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 define 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
C e
L R
liquid
Figure 3.14 System containing an electrical resistor immersed in a liquid. The

dashed rectangle indicates the system boundary.
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 defining
the system to include both the resistor and the liquid. An external voltage source provides
an electric potential difference across the wires. When electrons flow in the circuit, the
resistor becomes warmer due to the ohmic resistance of the resistor. This phenomenon is
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 (Ohm’s 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 finite current and zero when there is no
current. Figure 3.15 on the next page 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 characteristic
wel
bc
I
Figure 3.15 Work of electrical heating with a fixed magnitude of Qsys as a function
of the electric current I D ¶Qsys = dt. The open circle indicates the limit of infinite
slowness.
b b b b
C C
L cell R e
cell
(a) (b)
Figure 3.16 Galvanic cell and external electrical resistor.

(a) Open circuit with isolated cell in an equilibrium state.
(b) Closed circuit.
of irreversible dissipative work (page 83). Note the resemblance of Fig. 3.15 to Fig. 3.11(b)
on page 83 for dissipative stirring work—they are the same graphs with different labels.
Suppose we redefine 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 poten-
tial 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, 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.
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 defined by
def
Ecell D R L (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
wel
bc
bc
Figure 3.17 Electrical work of a galvanic cell for a fixed magnitude of Qsys as a
function of the electric current I D ¶Qsys = dt . Open circles: reversible limits.
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 figure 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 finite 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. 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.
Thus, unlike the dissipative work of stirring and electrical heating, electrical work with
a galvanic cell has a nonzero reversible limit, as reflected by the difference in the appear-
ance 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 trans-
ferred 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 re-
versible limit is always a state function. Electrical work with a galvanic cell does
3.9 I RREVERSIBLE W ORK AND I NTERNAL F RICTION 91
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 finite 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 limit—that 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 field, 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 flow spontaneously
across the system boundary, or an irreversible chemical reaction takes place spontaneously
in a homogeneous phase. Nor is internal friction necessarily involved when positive work
increases the thermal energy: during an infinitely slow adiabatic compression of a gas,
the temperature and thermal energy increase but internal friction is absent—the process is
reversible.
During a process with irreversible work, energy dissipation can be either partial or com-
plete. Dissipative work, such as the stirring work and electrical heating described in previ-
ous sections, is irreversible work with complete energy dissipation. The final equilibrium
state of an adiabatic process with dissipative work can also be reached by a path in which
positive heat replaces the dissipative work. This is a special case of the minimal work
principle.
As a model for work with partial energy dissipation, consider the gas-filled cylinder-
and-piston device depicted in Fig. 3.18 on the next page. This device has an obvious source
of internal friction in the form of a rod sliding through a bushing. The contact between
the rod and bushing is assumed to be lubricated to allow the piston to move at velocities
infinitesimally close to zero. 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
gas
P
R
B gas
Figure 3.18 Cylinder and piston with internal sliding friction. The dashed rectangle
indicates the system boundary. P—piston; R—internal rod attached to the piston; B—
lubricated bushing fixed inside the cylinder. A fixed amount of an ideal gas fills the
remaining space inside the cylinder.
where x is the piston position and F sys 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:17
F sys D pAs C Ffric

int
(3.9.3)
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
wD p dV .Ffric =As / dV (3.9.4)
V1 V1
The first term on the right is the work ofR expanding or compressing the gas. The second
int
term is the frictional work: wfric D .Ffric =As / dV . The frictional work 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 infinite slowness RFfric 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 finite rate. The fric-
tional 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.
17 Thisequation assumes that the gas pressure is uniform, and that a term for the acceleration of the rod is
negligible.
8 8
bc
bc bc
6 6
.F sys =As / / bar

.F sys =As / / bar
4 4
2 2
bc
bc bc
0 0
0 0:10 0:20 0:30 0 0:10 0:20 0:30
V = dm3 V = dm3
(a) (b)
Figure 3.19 Indicator diagrams for the system of Fig. 3.18.

Solid curves: F sys =As for irreversible adiabatic volume changes at finite rates in the
directions indicated by the arrows.
Dashed curves: F sys =As D p along a reversible adiabat.
Open circles: initial and final equilibrium states.
(a) Adiabatic expansion.
(b) Adiabatic compression.
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
finite 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 flow is present within the system, that would also be a contribution.
Figure 3.19 shows indicator diagrams for adiabatic expansion and compression with
internal friction. The solid curves are for irreversible processes at finite rates, and the dashed
curves are for reversible processes with the same initial states as the irreversible processes.
The areas under the curves confirm 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 final states of the irreversible processes have
greater internal energies and higher temperatures and pressures than the final 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 final 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
3.10 R EVERSIBLE AND I RREVERSIBLE P ROCESSES : G ENERALITIES 94
bc
compression
hdV=dt i
bc
expansion
Figure 3.20 Adiabatic expansion work with internal friction for a fixed magnitude of
V , as a function of the average rate of volume change. The open circles indicate the
reversible limits.
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
difficult for an irreversible process with both work and heat.
Figure 3.20 shows the effect of the rate of change of the volume on the adiabatic work
for a fixed 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 that they have
the same magnitude. The figure 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 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.
Infinitesimal quantities
P of work during a process are calculated from an expression of
the form ¶w D i Yi dXi , where Xi is the work coordinate of kind of work i and
Yi is the conjugate work coefficient.
The work coefficients and work coordinates of reversible work are state functions.
Energy transferred across the boundary by work in a reversible process is fully recov-
ered as work of the opposite sign in the reverse reversible process. It follows from
the first law that heat is also fully recovered in the reverse process.
When work occurs irreversibly at a finite rate, there is partial or complete dissipation
of energy. The dissipation results in a change that could also be accomplished with
positive heat, such as an increase of thermal energy within the system.
Dissipative work is positive irreversible work with complete energy dissipation. The
work coordinate for this type of work is not a state function. Examples are stirring
work (Sec. 3.7.1) and the work of electrical heating (Sec. 3.8.2).
3.10 R EVERSIBLE AND I RREVERSIBLE P ROCESSES : G ENERALITIES 95
If a process is carried out adiabatically and has a reversible limit, 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. The dependence of work on the rate of change
of the work coordinate is shown graphically for examples of dissipative work in Figs.
3.11(b) and 3.15, and for examples of work with partial energy dissipation in 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 lists general formulas for various kinds of work, including those that were
described in detail in Secs. 3.4–3.8.
Table 3.1 Some kinds of work
Kind Formula Definitions

Linear mechanical work ¶w D Fxsur dx Fxsur D x-component of force exerted
by surroundings
dx D displacement in x direction
Shaft work ¶w D b d# b D internal torque at boundary
# D angle of rotation
Expansion work ¶w D pb dV pb D average pressure at moving
boundary
Surface work of a flat surface ¶w D dAs D surface tension, As D surface area
Stretching or compression ¶w D F dl F D stress (positive for tension,
of a rod or spring negative for compression)
l D length
Gravitational work ¶w D mg dh m D mass, h D height
g D acceleration of free fall
Electrical work in a circuit ¶w D ¶Qsys D electric potential difference
D R L
¶Qsys D charge entering system at right
Electric polarization ¶w D E dp E D electric field strength
p D electric dipole moment of system
Magnetization ¶w D B dm B D magnetic flux density
m D magnetic dipole moment of system
18 Ifthe 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 sufficient to determine the amounts in each of the three
possible phases.
P ROBLEMS 96
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 Hooke’s 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 fixed amounts of water and air and an incom-
pressible 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 defined. Esys is the energy change in a lab frame, and
U is the energy change in a specified local frame.
For each of the following definitions 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 through-
out the system differs only infinitesimally from 300:0 K and the pressure on both sides of the
piston differs only infinitesimally from 1:00 bar.
P ROBLEMS 97
p D 3:00 bar porous

V D 0:500 m3 plug
T D 300:0 K piston
gas
Text D 300:0 K
pext D 1:00 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 confined
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 pro-
cesses 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 orienta-
tion, so the elevation of the center of mass of the gas confined by the piston changes as the pis-
ton 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 nitro-
gen 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 confined 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 figure.
P ROBLEMS 98
gas l
Figure 3.23
vacuum weight
ideal
gas h
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 com-
pressed adiabatically and spontaneously. Pressure gradients in the gas, a form of friction,
eventually cause the piston to come to rest at a final position h2 . Find the final 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 sufficiently large, V2 could
be made as close to zero as desired. Actually, however, this is not the case. Find ex-
pressions for V2 and T2 in the limit as m2 approaches infinity, 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 77 shows the path of a reversible adiabatic expansion or
compression of a fixed amount of an ideal gas. Information about the gas is given in the figure
P ROBLEMS 99
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 , find the final temperature to 0:1 K and the work.
p D 3:00 bar p D0
V D 0:500 m3 V D 1:00 m3
T D 300:0 K
gas vacuum
Text D 300:0 K
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 final 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 Joule’s 1850 paper a
and have been converted to SI units.
Properties of the paddle wheel apparatus:

combined mass of the two lead weights . . . . . . . . . . . . . . . . 26:3182 kg
mass of water in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6:04118 kg
mass of water with same heat capacity
as paddle wheel, vessel, and lidb . . . . . . . . . . . . . . . . . . . 0:27478 kg
Measurements during the experiment:
number of times weights were wound up and released . . . 20
change of elevation of weights during each descent . . . . . 1:5898 m
1
final downward velocity of weights during descent . . . . . . 0:0615 m s
initial temperature in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . 288:829 K
final temperature in vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289:148 K
mean air temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289:228 K
a Ref. [84], p. 67, experiment 5.
b Calculated from the masses and specific heat capacities of the materials.
3.10 This problem guides you through a calculation of the mechanical equivalent of heat using data
from one of James Joule’s 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 floor
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 86). The paddle wheel was engaged to the weights through the
P ROBLEMS 100
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 finish of the experiment.
For the purposes of the calculations, define 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
floor. 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
floor.)
(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 repre-
sents 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 floor, 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 oc-
curred 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 jj 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 first 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 questions—a general criterion for spontaneity in a closed system.
4.1 TYPES OF PROCESSES
Any conceivable process is either spontaneous, reversible, or impossible. These three pos-
sibilities were discussed in Sec. 3.2 and are summarized below.
A spontaneous process is a real process that can actually take place in a finite 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 infinite 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. Some-
times 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 pro-
cess 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.
101
CHAPTER 4 THE SECOND LAW
4.2 S TATEMENTS OF THE S ECOND L AW 102
It is true that reversible processes and purely mechanical processes are idealized pro-
cesses that cannot occur in practice, but a spontaneous process can be practically reversible
if carried out sufficiently 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 infinitesimal 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 defined 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 flow 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 first 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 flow from the cool body to the device,
1 During a reversible process, the temperature usually has the same value Tthroughout 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.
warm
warm
q
device
q
cool q
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 flow from the device to the warm body, and the final 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 infinite 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 there-
with, 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 first law. This process is just the reverse of the Joule paddle-
wheel 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. [32], page 117.
3 This hypothetical process is called “perpetual motion of the second kind.”
heat engine b
(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 Kelvin–Planck statement of the second law: It is impossible to construct a heat en-
gine 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 Kelvin–Planck statement assert that certain pro-
cesses, although they do not violate the first law, are nevertheless impossible.
These processes would not be impossible if we could control the trajectories of large
numbers of individual particles. Newton’s laws of motion are invariant to time re-
versal. Suppose we could measure the position and velocity of each molecule of a
macroscopic system in the final 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 hypo-
thetical 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 Kelvin–Planck statement) are
equivalent, in the sense that if one is true, so is the other. An expression is also derived for
the efficiency of a Carnot engine for the purpose of defining thermodynamic temperature.
Section 4.4 combines Carnot cycles and the Kelvin–Planck statement to derive the existence
4 Ref. [133], p. 89.
4.3 C ONCEPTS D EVELOPED WITH C ARNOT E NGINES 105
8 b
A
7
B0
b
6
p=Pa
.
B
5 b
4 b
D
3 b
C0 b
C
2
0:5 1:0 1:5 2:0 2:5
V =m3
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 102,
Sec. 4.6 describes some applications, and Sec. 4.7 is a summary. Finally, Sec. 4.8 briefly
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 Thermodynam-
ics 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 in-
voking thermodynamic cycles. You may avoid the details by skipping Secs. 4.3–4.5.
(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. [104], p. 2.
BIOGRAPHICAL SKETCH
Sadi Carnot (1796–1832)
carried as far as possible.
(3) It should be so arranged that the passage
of the elastic fluid 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 re-
sult of rarefaction [i.e., adiabatic expansion].
Carnot derived these principles from the ab-
stract 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
Sadi Carnot was the eldest son of Lazare zero, whereas today we would say that it is the
Carnot, a famous French anti-royalist politi- net entropy change that is zero.
cian, one of Napoleon’s generals with a great Despite the flaw of assuming that heat is
interest in mathematics. As a boy Sadi was conserved, a view which there is evidence he
shy and sensitive. He studied at the École was beginning to doubt, his conclusion was
Polytechnique, a training school for army en- valid that the efficiency of a reversible cycle
gineers, and became an army officer. operating between two fixed temperatures is
Carnot is renowned for the one book he independent of the working substance. He
wrote: a treatise of 118 pages entitled Reflec- based his reasoning on the impossibility of the
tions on the Motive Power of Fire and on Ma- perpetual motion which would result by com-
chines Fitted to Develop that Power. This was bining the cycle with the reverse of a more ef-
published in 1824, when he was 28 and had ficient cycle. Regarding Carnot’s accomplish-
retired from the army on half pay. ment, William Thomson (later Lord Kelvin)
The book was written in a nontechnical wrote:
style and went virtually unnoticed. Its purpose Nothing in the whole range of Natural Philoso-
was to show how the efficiency of a steam en- phy 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 Carnot’s personal-
that of Britain at the time:a ity as follows:b
Notwithstanding the work of all kinds done by
He was reserved, almost taciturn, with a hatred
steam-engines, notwithstanding the satisfactory
of any form of publicity. . . . his friends all spoke
condition to which they have been brought today,
of his underlying warmth and humanity. Pas-
their theory is very little understood, and the at-
sionately fond of music, he was an excellent vi-
tempts to improve them are still directed almost
olinist who preferred the classical Lully to the
by chance.
“moderns” of the time; he was devoted to litera-
. . . We can easily conceive a multitude of ma-
ture and all the arts.
chines fitted to develop the motive power of heat
through the use of elastic fluids; but in whatever Carnot came down with scarlet fever and,
way we look at it, we should not lose sight of the while convalescing, died—probably of the
following principles: cholera epidemic then raging. He was only 36.
(1) The temperature of the fluid should be Two years later his work was brought to
made as high as possible, in order to obtain a public attention in a paper written by Émile
great fall of caloric, and consequently a large Clapeyron (page 217), who used indicator dia-
production of motive power. grams to explain Carnot’s ideas.
(2) For the same reason the cooling should be
a Ref. [27]. b Ref. [115], page x.
2:5
A B
2:4
p=105 Pa
2:3
2:2
D C
2:1
0 1 2
V =10 4 m3
Figure 4.4 Indicator diagram for a Carnot engine using H2 O as the working sub-
stance. 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 Kelvin–Planck
engine.) Adjustment of the length of path A!B makes the magnitude of w as large or
small as desired—note 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 pro-
cesses. 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 ir-
reversible steps in each cycle: (1) high-pressure steam enters the cylinder from a boiler
Th Th
4 4
1 1
3 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 4.5–4.7 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 Kelvin–Planck statements

We can use the logical tool of reductio ad absurdum to prove the equivalence of the Clausius
and Kelvin–Planck statements of the second law.
Let us assume for the moment that the Clausius statement is incorrect, and that the de-
vice 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 110. We adjust the cycles of the Clausius device and
Carnot engine to transfer equal quantities of heat from and to the cold reservoir. The com-
bination 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 Kelvin–Planck statement is impossible.
Thus, if the Kelvin–Planck statement is correct, it is impossible to operate the Clausius
BIOGRAPHICAL SKETCH
Rudolf Julius Emmanuel Clausius (1822–1888)
therewith, occurring at the same time. Every-
thing we know concerning the interchange of
heat between two bodies of different temperature
confirms this; for heat everywhere manifests a
tendency to equalize existing differences of tem-
perature, and therefore to pass in a contrary di-
rection, i. e. from warmer to colder bodies. With-
out further explanation, therefore, the truth of the
principle will be granted.
In an 1865 paper, he introduced the symbol
U for internal energy, and also coined the word
entropy with symbol S:c
Rudolf Clausius was a German theoretical
We might call S the transformational content of
physicist who was the first to treat thermody-
the body, just as we termed the magnitude U
namics as a rigorous science, based on the ear- its thermal and ergonal content. But as I hold
lier writings of Carnot and Clapeyron. it better to borrow terms for important magni-
He was born in Köslin, Prussia, into a large tudes from the ancient languages, so that they
family. His father was an educator and church may be adopted unchanged in all modern lan-
minister. guages, I propose to call the magnitude S the en-
Clausius was successively a professor at tropy of the body, from the Greek word o , J
universities in Berlin, Zurich, Würzburg, and transformation. I have intentionally formed the
Bonn. In addition to thermodynamics, he did word entropy so as to be as similar as possible
work on electrodynamic theory and the kinetic to the word energy; for the two magnitudes to
be denoted by these words are so nearly allied in
theory of gases.
their physical meanings, that a certain similarity
Max Planck, referring to a time early in his
in designation appears to be desirable.
own career, wrote:a
The 1865 paper concludes as follows, end-
One day, I happened to come across the trea-
tises of Rudolf Clausius, whose lucid style and
ing with Clausius’s often-quoted summations
enlightening clarity of reasoning made an enor- of the first and second laws:d
mous impression on me, and I became deeply ab- If for the entire universe we conceive the same
sorbed in his articles, with an ever increasing en- magnitude to be determined, consistently and
thusiasm. I appreciated especially his exact for- with due regard to all circumstances, which for
mulation of the two Laws of Thermodynamics, a single body I have called entropy, and if at the
and the sharp distinction which he was the first same time we introduce the other and simpler
to establish between them. conception of energy, we may express in the fol-
lowing manner the fundamental laws of the uni-
Clausius based his exposition of the second
verse which correspond to the two fundamental
law on the following principle that he pub-
theorems of the mechanical theory of heat.
lished in 1854:b 1. The energy of the universe is constant.
. . . it appears to me preferable to deduce the gen- 2. The entropy of the universe tends to a max-
eral form of the theorem immediately from the imum.
same principle which I have already employed
Clausius was a patriotic German. During
in my former memoir, in order to demonstrate
the modified theorem of Carnot.
the Franco-Prussian war of 1870–71, he un-
This principle, upon which the whole of the dertook the leadership of an ambulance corps
following development rests, is as follows:— composed of Bonn students, was wounded in
Heat can never pass from a colder to a warmer the leg during the battles, and suffered disabil-
body without some other change, connected ity for the rest of his life.
a Ref. [134], page 16. b Ref. [32], page 117. c Ref. [33], page 357. d Ref. [33], page 365.
Th Th Th Th Th Th
3 4 1 1 4 3
1 1 1 1
3 3 3 3
Tc Tc Tc Tc
„ ƒ‚ … „ ƒ‚ …
(a) (b) (c) (d)
Figure 4.6 (a) A Clausius device combined with the Carnot engine of Fig. 4.5(a).
(b) The resulting impossible Kelvin–Planck engine.
(c) A Kelvin–Planck 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 Kelvin–Planck 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 Kelvin–Planck
statement claims is impossible (a “Kelvin–Planck engine”) by seeing what happens if we
assume this engine is actually possible. We combine a Kelvin–Planck 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 Kelvin–Planck 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 Kelvin–Planck statement must also be correct.
These conclusions complete the proof that the Clausius and Kelvin–Planck 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 efficiency of a Carnot engine

Integrating the first-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 efficiency of a heat engine is defined as the fraction of the heat input qh that is returned
as net work done on the surroundings:
def w
D (4.3.2)
qh
Th Th Th Th Th Th
4 5 1 3 4 1
1 1 1 1
3 4 1 2 3 1
Tc Tc Tc Tc Tc Tc
„ ƒ‚ … „ ƒ‚ …
(a) (b) (c) (d)
Figure 4.7 (a) A Carnot engine of efficiency D 1=4 combined with a Carnot engine
of efficiency D 1=5 run in reverse.
(b) The resulting impossible Clausius device.
(c) A Carnot engine of efficiency D 1=3 combined with the Carnot engine of effi-
ciency D 1=4 run in reverse.
(d) The resulting impossible Clausius device.
By substituting for w from Eq. 4.3.1, we obtain

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

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 efficiency 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 efficiency of the Carnot engine be different from the efficiency 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 tem-
peratures have the same efficiency.
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 efficiency, 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.
Thus, the efficiency 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 efficiency is given by D
1 C qc =qh , the ratio qc =qh must be a unique function of Tc and Th only. To find 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 first law, dU D ¶q C ¶w, and solving for ¶q, we obtain
nRT
¶q D CV dT C dV (4.3.4)
V
(ideal gas, reversible
expansion work only)
Dividing both sides by T gives

¶q C dT dV
D V C nR (4.3.5)
T T T
(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
¶q CV dT V
Z
Path B!C: D C nR ln C D 0 (4.3.6)
T Th T VB
Z Th
¶q CV dT V
Z
Path D!A: D C nR ln A D 0 (4.3.7)
T Tc T VD
Adding these two equations (the integrals shown with limits cancel) gives the relation
VA VC
nR ln D0 (4.3.8)
VB VD
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 qh D nRTh ln.VB =VA / (4.3.10)

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

qc T ln.VD =VC /
D c (4.3.12)
qh Th ln.VB =VA /
By means of Eq. 4.3.9, this ratio becomes

qc Tc
D (4.3.13)
qh Th
(Carnot cycle)
Accordingly, the unique function of Tc and Th we seek that is equal to qc =qh is the ratio
Tc =Th . The efficiency, from Eq. 4.3.3, is then given by
Tc
D1 (4.3.14)
Th
(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 efficiency is less than one as deduced
earlier on page 111. This conclusion is an illustration of the Kelvin–Planck statement of the
second law: A heat engine cannot have an efficiency of unity—that 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 108, 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 efficiency 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 efficiency 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” temper-
ature scale. The physical quantity now called thermodynamic temperature is defined by
the relation
Tc qc
D (4.3.15)
Th qh
(Carnot cycle)
That is, the ratio of the thermodynamic temperatures of two heat reservoirs is equal, by
definition, 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 defined value of the thermodynamic tempera-
ture of one of the heat reservoirs, can establish the thermodynamic temperature of the other
heat reservoir. This defined value is provided by the triple point of H2 O; its thermodynamic
temperature is defined 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
BIOGRAPHICAL SKETCH
William Thomson, Lord Kelvin (1824–1907)
I heard his paper read in the section, and
felt strongly impelled at first to rise and say
that it must be wrong because the true mechan-
ical value of heat given, suppose in warm wa-
ter, 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 clash-
ing with Joule’s ‘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 dis-
William Thomson was born in Belfast, Ireland. covery, and a most important measurement to
His mother died when he was six. In 1832 the bring forward. So instead of rising with my ob-
family moved to Glasgow, Scotland, where his jection to the meeting I waited till it was over,
father had been appointed as the chair of math- and said my say to Joule himself, at the end of
ematics at the University. the meeting. This made my first introduction to
In 1845 Thomson was age 21. He had re- him. After that I had a long talk over the whole
matter at one of the conversaziones of the Asso-
cently graduated from Cambridge University
ciation, and we became fast friends from thence-
with high honors in mathematics, and was in
forward.
Paris for discussions with French physicists
and mathematicians. He had learned about The physicist Charles Everitt described
Carnot’s book from Clapeyron’s 1834 paper Thomson’s personality as follows:c
but could not find a copy—no bookseller in Thomson was the kind of man who created all
Paris had even heard of it. Nevertheless, the in- around him a sense of bustle and excitement.
formation he had was sufficient for him to real- . . . [He] was a man of violent enthusiasms. He
would take up a subject, work at it furiously for
ize that Carnot’s ideas would allow a thermo-
a few weeks, throw out a string of novel ideas,
dynamic temperature scale to be defined, one
and then inexplicably drop everything and pass
that does not depend on any particular gas. on.
The following year, Thomson became the
During his career, Thomson published more
Chair of Natural Philosophy at Glasgow Uni-
than 600 papers and received many honors. In
versity, largely through the influence of his fa-
addition to the theory of heat, his work in-
ther. He remained there until his retirement in
cluded a dynamical theory of electricity and
1899. His concept of a thermodynamic tem-
magnetism, and the invention of the mirror gal-
perature scale, his best-known contribution to
vanometer used as the telegraph receiver for
thermodynamics, was published in 1848.a
the first transatlantic submarine cables.
Thomson published other papers on the the-
Thomson was the chief technical consultant
ory of heat, but his ideas eventually changed
for the initial cable projects. As a result of
as a result of hearing a presentation by James
their success and his involvement in the con-
Joule at a meeting in Oxford in 1847. Joule
struction of a global cable network, he became
was describing his experiments with the con-
extremely wealthy and was knighted in 1866.
version of work to thermal energy by a paddle
In 1892 he became Baron Kelvin of Largs.
wheel. In a letter written to his nephew, J. T.
“Kelvin” is the name of the river that flows past
Bottomley, Thomson wrote:b
Glasgow University, “Largs” is the town where
I made Joule’s acquaintance at the Oxford meet-
he had his home, and kelvin is now the SI unit
ing, and it quickly ripened into a life-long friend-
ship.
of thermodynamic temperature.
a Ref. [89]. b Ref. [17]. c Ref. [51].
4.4 D ERIVATION OF THE M ATHEMATICAL S TATEMENT OF THE S ECOND L AW 115
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 “ex-
perimental cycle.” There are no restrictions on the contents of the experimental system—it
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 temper-
ature and pressure gradients, constraints and external fields 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 Kelvin–Planck statement of the second law. The heat transferred across the bound-
ary of the experimental system in each infinitesimal 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 117). 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 Kelvin–Planck statement to a cycle of the supersystem.6
We assume that we are able to control changes of the work coordinates of the experi-
mental 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 cylinder-
and-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 infinitesimal Carnot cycles with infinitesimal 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. [129], Chap. 4; Ref. [1], Chap. 5; and Ref. [98], p. 53.
BIOGRAPHICAL SKETCH
Max Karl Ernst Ludwig Planck (1858–1947)
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 scientific life that I have but seldom—in
fact, I might say, never—succeeded 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 hap-
pened this time, too. All my sound arguments
fell on deaf ears.
Max Planck, best known for his formulation of Planck became an associate professor of
the quantum theory, had a passionate interest physics at the University of Kiel. In 1889
in thermodynamics in the early part of his ca- he succeeded Kirchhoff as Professor at Berlin
reer. University. By the end of the following year,
He was born in Kiel, Germany, where his at the age of 42, he had worked out his quan-
father was a distinguished law professor. His tum theory to explain the experimental facts of
family had a long tradition of conservatism, blackbody radiation, a formulation that started
idealism, and excellence in scholarship. a revolution in physics. He was awarded the
As a youth, Planck had difficulty deciding 1918 Nobel Prize in Physics “in recognition of
between music and physics as a career, finally the services he rendered to the advancement of
settling on physics. He acquired his interest in Physics by his discovery of energy quanta.”
thermodynamics from studies with Hermann Planck was reserved and only enjoyed so-
von Helmholtz and Gustav Kirchhoff and from cializing with persons of similar rank. He
the writings of Rudolf Clausius. His doctoral was a gifted pianist with perfect pitch, and en-
dissertation at the University of Munich (1879) joyed hiking and climbing well into his old
was on the second law. age. He was known for his fairness, integrity,
In 1897, Planck turned his papers on ther- and moral force.
modynamics into a concise introductory text- He endured many personal tragedies in his
book, Treatise on Thermodynamics. It went later years. His first wife died after 22 years
through at least seven editions and has been of a happy marriage. His elder son was killed
translated into English.a in action during World War I. Both of his twin
Concerning the second law he wrote:b daughters died in childbirth.
Another controversy arose with relation to the Planck openly opposed the Nazi persecu-
question of the analogy between the passage tion of Jews but remained in Germany dur-
of heat from a higher to a lower temperature ing World War II out of a sense of duty.
and the sinking of a weight from a greater to The war brought further tragedy: his house in
a smaller height. I had emphasized the need Berlin was destroyed by bombs, and his sec-
for a sharp distinction between these two pro- ond son was implicated in the failed 1944 at-
cesses. . . However, this theory of mine was con-
tempt to assassinate Hitler and was executed
tradicted by the view universally accepted in
those days, and I just could not make my fellow
by the Gestapo. Planck and his second wife
physicists see it my way. . . escaped the bombings by hiding in the woods
A consequence of this point of view [held by and sleeping in haystacks in the countryside.
others] was that the assumption of irreversibility They were rescued by American troops in
for proving the Second Law of Thermodynamics May, 1945.
a Ref. [133]. b Ref. [134], pages 29–30.
Tres
Tres
¶q 0
Tres
Tb
¶q
Tb Tb
experimental experimental
system 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 infinitesimal
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 system’s 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 infinitesimal Carnot cycle, the
relation Tb =Tres D ¶q= ¶q 0 , or
¶q
¶q 0 D Tres (4.4.1)
Tb
After many infinitesimal Carnot cycles, the experimental cycle is complete, the exper-
imental 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:
¶q
I
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 posi-
tive? If so, the net work would be negative (to make the internal energy change zero) and the
supersystem would have converted heat from a single heat reservoir completely into work,
a process the Kelvin–Planck 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
¶q
I
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 changes—the
path must not include an impossible change, such as the reverse of an irreversible change.
The Clausius inequality H says that if a cyclic path meets this specification, it is impossible
for the cyclic integral .¶q=Tb / to be positive.
If the entire experimental cycle is adiabatic (which is only possible ifH the process is
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 flow of heat into the system, and in which ¶q is either positive or zero in
each path element. The final state of this process is equilibrium state B. If each R B infinitesimal
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
RA a cycle
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 flow 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 finite “one-way” heat (i.e., the
flow 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 flow reversibly into the liquid. It is impossible to
duplicate this change of state by a reversible process without heat—that 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 final state. There is nothing
in the rule that says we can’t 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 infinitesimally close to it that are inaccessible by
a reversible adiabatic process. This is Carathéodory’s principle of adiabatic inaccessibility.7
7 ConstantinCarathéodory in 1909 combined this principle with a mathematical theorem (Carathéodory’s the-
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.
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 final 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 uni-
form 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 adi-
abatic, 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 Carathéodory’s principle
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 final states of a reversible process with one-way heat (i.e.,
each nonzero infinitesimal 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 infinite number of nonintersecting
reversible adiabatic surfaces filling the N-dimensional space. (To visualize this for N D 3,
think of a flexed stack of paper sheets; each sheet represents a different reversible 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 final surface lies on one side of the
initial surface, and if q is negative, the final surface is on the opposite side.
4.4.2 Using reversible processes to define the entropy

The existence of reversible adiabatic surfaces is the justification for defining a new state
function S , the entropy. S is specified to have the same value everywhere on one of these
surfaces, and a different, unique value on each different surface. In other words, the re-
versible adiabatic surfaces are surfaces of constant entropy in the N-dimensional space.
The fact that the surfaces fill this space without intersecting ensures that S is a state func-
tion for equilibrium states, because any point in this space represents an equilibrium state
and also lies on a single reversible adiabatic surface with a definite 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
500
T /K
. 400
300
0:02 0:03 0:04 0:05
V /m3
Figure 4.9 A family of reversible adiabatic curves (two-dimensional reversible adi-

abatic 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 purpose—see 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 figure 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
B
¶q
Z
q 0 D Tres (4.4.4)
A Tb
and it is positive.
Suppose the same experimental system undergoes a second reversible process, not nec-
essarily 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 b
B B
b b
T D T D
!
b b
A A
b b
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 :

D
¶q
Z
00
q D Tres (4.4.5)
C Tb
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 first pro-
cess 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 Kelvin–Planck
statement, neither cycle can have positive netR heat. Therefore q 0 and q 00 must be equal, and
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 defining the entropy change of
a reversible process in a closed system is
¶q
Z
S D (4.4.6)
Tb
(reversible process,
closed system)
This formula satisfies 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
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 infinitesimal path element of a reversible
process, the equation becomes
¶q
dS D (4.4.7)
Tb
(reversible process,
closed system)
It is common to see this equation written in the form dS D ¶qrev =T , where ¶qrev denotes
an infinitesimal 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 infinitesimal change of a state function. The quantity
c=Tb , where c is any nonzero constant, would also be a satisfactory integrating factor;
so the definition of entropy, using cD1, is actually one of an infinite number of possible
choices for assigning values to the reversible adiabatic surfaces.
4.4.3 Some properties of the entropy

It is not difficult 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 infinitesimal 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 con-
strained 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 with-
out 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 defining 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 definition, S for a purely mechanical process (Sec. 3.2.3) is zero.
4.5 I RREVERSIBLE P ROCESSES 123
internal constraints will partition it into practically-uniform regions whose

R entropy is addi-
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 definite value of S.
4.5 IRREVERSIBLE PROCESSES
We know that during a reversible process of a closed system, each infinitesimal R entropy
change dS is equal to ¶q=Tb and the finite 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 infinitesimal irreversible change of
aR closed system, dS is greater than ¶q=Tb , and for an entire process S is greater than
.¶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 par-
ticular initial state A. The final 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 infinitesimal path element of the irreversible adiabatic process
(¶w ¤ 0), we know from experience that this work would be different if the work coor-
dinate or coordinates were changing at a different rate, because energy dissipation from
internal friction would then be different. In the limit of infinite slowness, an adiabatic pro-
cess with initial state A and the same change of work coordinates would become reversible,
and the net work and final internal energy would differ from those of the irreversible pro-
cess. Because the final 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 final states, must be equilib-
rium states. There is therefore a conceptual difficulty 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 infinitesimal path element of the irre-
versible 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 64).
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 124
Tres
Tres
¶q 0
Tres
Tb
¶q
Tb Tb
experimental experimental
system 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 super-
system.
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 infinitesimal R B path element of
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 stateR B back to state A. The Clausius inequality (Eq. 4.4.3)
A
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 final state B is reached by an
irreversible adiabatic process. If the two states are only infinitesimally different, then the
change is infinitesimal. Thus for an infinitesimal 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 fol-
lows. 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 infinitesimal 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 figure. 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 figure.
The supersystem exchanges work, but not heat, with its surroundings. During one
infinitesimal cycle of the Carnot engine, the net entropy change of the Carnot engine is
4.6 A PPLICATIONS 125
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:
¶q
dSss D dS C dSres D dS (4.5.2)
Tb
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 infinitesimal entropy change during the irreversible
change of the experimental system, must be greater than ¶q=Tb :
¶q
dS > (4.5.3)
Tb
(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 final states of the irre-
versible process, we obtain a relation for the finite entropy change corresponding to many
infinitesimal cycles of the Carnot engine:
¶q
Z
S > (4.5.4)
Tb
(irreversible process, closed system)
4.6 APPLICATIONS
The lengthy derivation in Secs. 4.3–4.5 is based on the Kelvin–Planck statement describing
the impossibility of converting completely into work the energy transferred into the sys-
tem 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 102. The mathemati-
cal statement includes an equality, dS D ¶q=Tb , that applies to an infinitesimal reversible
change, and an inequality, dS > ¶q=Tb , that applies to an infinitesimal 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:
¶q
dS (4.6.1)
Tb
( irrev
rev , closed system)
The inequality refers to an irreversible change and the equality to a reversible change, as
indicated by the notation irrev
rev in the conditions of validity. The integrated form of this
relation is
¶q
Z
S (4.6.2)
Tb
( irrev
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 102, 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 definition 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
T2
S D C ln (4.6.4)
T1
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 fluid, is changed reversibly and adiabati-
cally, there is of course no entropy change.
When the volume of an ideal gas is changed reversibly and isothermally, there is expan-
sion 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
V2
S D nR ln (4.6.5)
V1
(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 final 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.
4.6.3 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, uninfluenced by changes in the surroundings—the pro-
cess occurs by itself, of its own accord. The initial state and the intermediate states of the
process must be nonequilibrium states, because by definition 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 infinitesimal 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 assur-
ance that this is a valid concept), we can say that as spontaneous changes occur in the
universe, its entropy continuously increases. Clausius summarized the first 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 flow 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 final equilib-
rium 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 de-
scribe a region that has a continuous temperature gradient, we approximate the region with a
very large number of very small phases or parcels, each having a temperature infinitesimally
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 infinitesimal quantity of
heat transferred during an infinitesimal 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 satisfied,
¶q’“ is zero.
To evaluate the entropy change, we need a reversible path from the initial to the final
state. The Pnet quantity of heat transferred to phase ’ during an infinitesimal time interval
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:
X X ¶q ’ X X ¶q’“
dS D dS ’ D D
’ ’
T’ ’ “¤’
T’
XX ¶q’“ ¶q“’

D C (4.6.7)
’ “>’
T’ T“
There is also the condition of quantitative energy transfer, ¶q“’ D ¶q’“ , which we use
to rewrite Eq. 4.6.7 in the form
XX 1 1

dS D ¶q’“ (4.6.8)
’
T’ T“
“>’
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 finite 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 iso-
lated 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 confine 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 final 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. [99], obtains the same result by a more complicated derivation.
4.7 S UMMARY 129
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 definitions 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 finite 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 infinitesimal quantity ¶q=Tb equals the infinites-
imal change of a state function called the entropy, S. Here ¶q is heat transferred at the
boundary where the temperature is Tb .
In each infinitesimal 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 sum-
marized by the relation dS ¶q=Tb .
Consider two particular equilibrium states 1 and 2 of a closed system. The system can
change
R from state 1 to state 2 by either a reversible process, with
R S equal to the integral
.¶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 130
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 final equi-
R system. The value of S is the same for both paths, but the
librium states of a closed
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 equilib-
rium 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 definite energy. The
microstate is one that is accessible to the system—that is, one whose wave function is com-
patible with the system’s volume and with any other conditions and constraints imposed
on the system. The system, while in the equilibrium state, continually jumps from one ac-
cessible 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 ther-
mal equilibrium with an external constant-temperature heat reservoir. The internal energy
then fluctuates 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 131
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 accom-
panied 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 spread-
ing 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. [100].
12 Ref. [96].
13 Ref. [97].
14 The symbol S for entropy seems originally to have been an arbitrary choice by Clausius; see Ref. [82].
P ROBLEMS 132
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 de-
vice.”
4.2 A system consisting of a fixed 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, find 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 97 and 3.25 on page 99 and described in
R systems, evaluate S for the process that results from opening the
Probs. 3.3 and 3.8. For both
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 133
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 final 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 in-
troduced: enthalpy, Helmholtz energy, and Gibbs energy. These functions, together with
internal energy, are called thermodynamic potentials.1 Some formal mathematical ma-
nipulations 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 defined by a certain
minimum number of state functions, the independent variables. State functions not treated
as independent variables are dependent variables. Infinitesimal changes in any of the inde-
pendent variables will, in general, cause an infinitesimal change in each dependent variable.
A dependent variable is a function of the independent variables. The total differen-
tial of a dependent variable is an expression for the infinitesimal change of the variable in
terms of the infinitesimal 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 indepen-
dent variables and the infinitesimal 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 fixed amount of an ideal gas with pressure given by p D nRT =V ,
the total differential of the pressure can be written
nR nRT
dp D dT dV (5.1.2)
V V2
1 Theterm thermodynamic potential should not be confused with the chemical potential, , to be introduced on
page 136.
134
CHAPTER 5 THERMODYNAMIC POTENTIALS
5.2 T OTAL D IFFERENTIAL OF THE I NTERNAL E NERGY 135
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 first 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 first law in the form
dU D ¶q pb dV C ¶w 0 (5.2.1)
(closed system)
where ¶w 0 is nonexpansion work—that 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 infinitesimal
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 compo-
nent (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 infinite 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
uniform—that is, for a process in which the system stays very close to thermal and mechan-
ical 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 work—work without a re-
versible limit—cannot 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 136
As explained in Appendix F, we may identify the coefficient of each term in an expres-

sion for the total differential of a state function as a partial derivative of the function. We
identify the coefficients on the right side of Eq. 5.2.2 as follows:

@U @U
T D pD (5.2.3)
@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
dU D dS C dV C dn (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 first 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 system contains a mixture of s different substances in a single phase, and the sys-
tem is open so that the amount of each substance can vary independently, there are 2 C s
independent variables and the total differential of U can be written
s
X
dU D T dS p dV C i dni (5.2.6)
iD1 (open system,
P D1, ¶w 0 D0)
The coefficient i is the chemical potential of substance i. We identify it as the partial

derivative .@U=@ni /S;V;nj ¤i .
The term p dV on the right side of Eq. 5.2.6 is the reversible work. However, the
term T dS does not equal the reversible heat as it would if the system were closed. This
is because the entropy change dS is partly due to the entropy
P of the matter transferred
across the boundary. It follows that the remaining term, i i dni (sometimes called
the “chemical work”), should not be interpreted as the energy brought into the system
by the transfer of matter.2
2 Ref. [93].
5.3 E NTHALPY, H ELMHOLTZ E NERGY, AND G IBBS E NERGY 137
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 ENTHALPY, HELMHOLTZ ENERGY, AND GIBBS ENERGY
For the moment we shall confine 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 prod-
ucts 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, defined as follows:
def
Enthalpy H D U C pV (5.3.1)
def
Helmholtz energy A D U TS (5.3.2)
def
Gibbs energy G D U T S C pV D H TS (5.3.3)
These definitions are used whether or not the system has only two independent variables.
The enthalpy, Helmholtz energy, and Gibbs energy are important functions used exten-
sively in thermodynamics. They are state functions (because the quantities used to define
them are state functions) and are extensive (because U , S , and V are extensive). If temper-
ature or pressure are not uniform in the system, we can apply the definitions 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 infinitesimal changes of H , A, and G are obtained by applying the rules
of differentiation to their defining equations:
dH D dU C p dV C V dp (5.3.4)
5.3 E NTHALPY, H ELMHOLTZ E NERGY, AND G IBBS E NERGY 138
BIOGRAPHICAL SKETCH
Josiah Willard Gibbs (1839–1903)
face for H2 O which he sent to Gibbs.
The two papers of 1873 were followed by a
monumental paper in the same journal—in two
parts (1876 and 1878) and over 300 pages in
length!—entitled simply “On the Equilibrium
of Heterogeneous Substances.”d This third pa-
per used an analytical rather than geometri-
cal approach. From the first and second laws
of thermodynamics, it derived the conditions
needed for equilibrium in the general case of
a multiphase, multicomponent system. It in-
Willard Gibbs’s brilliant and rigorous formu- troduced the state functions now known as en-
lation of the theoretical basis of classical ther- thalpy, Helmholtz energy,e Gibbs energy, and
modynamics was essential for further develop- chemical potential. Included in the paper was
ment of the subject. the exposition of the Gibbs phase rule.
Gibbs was born in New Haven, Connecti- The only public comment Gibbs ever made
cut, and lived there all his life. His father was on his thermodynamic papers was in a letter of
a professor in the Yale Divinity School. Gibbs 1881 accepting membership in the American
was Professor of Mathematical Physics at Yale Academy of Arts and Sciences:f
College from 1871 until his death. The leading idea which I followed in my paper
Gibbs never married. In demeanor he was on the Equilibrium of Heterogeneous Substances
serene, kindly, reserved, and self-effacing. A was to develop the rôles of energy and entropy in
biographer wrote:a the theory of thermo-dynamic equilibrium. By
Gibbs’ attitude toward his discoveries is also il- means of these quantities the general condition
luminating as to his character. . . . he made no ef- of equilibrium is easily expressed, and by apply-
fort to “sell” his discoveries (aside from the usual ing this to various cases we are led at once to
distribution of reprints of his papers) or to pop- the special conditions which characterize them.
ularize the results. He was so confident of their We thus obtain the consequences resulting from
rightness and ability to stand on their own feet the fundamental principles of thermo-dynamics
that he was entirely content to let their value and (which are implied in the definitions of energy
importance be “discovered” by others. The fact and entropy) by a process which seems more
that he had made a discovery was to him an ir- simple, and which lends itself more readily to the
relevant matter; the important thing was the truth solution of problems, than the usual method, in
established. which the several parts of a cyclic operation are
explicitly and separately considered. Although
In 1873, when he was 34, the first two of my results were in a large measure such as had
Gibbs’s remarkable papers on theoretical ther- previously been demonstrated by other methods,
modynamics appeared in an obscure journal, yet, as I readily obtained those which were to me
Transactions of the Connecticut Academy.b,c before unknown, I was confirmed in my belief in
These papers explored relations among state the suitableness of the method adopted.
functions using two- and three-dimensional Gibbs had a visit about 1898 from a young
geometrical constructions. Gilbert Lewis. He told Lewis that he was
James Clerk Maxwell promoted Gibbs’s rather lonely at Yale, where few others were
ideas in England, and made a small plaster actively interested in his work.g
model of the three-dimensional S –V –U sur-
a Ref.[171], 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. [171], page
89. g Ref. [131].
5.4 C LOSED S YSTEMS 139
dA D dU T dS S dT (5.3.5)
dG D dU T dS S dT C p dV C V dp (5.3.6)
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 first law expression ¶q p dV C ¶w 0 , to obtain dH D ¶q C V dp C ¶w 0 .
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
R change
R under these conditions is equal to the heat. The integrated form of this
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, ob-
tained from the first 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 find 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. 5.3.4–5.3.6 with
dU D T dS p dV (5.4.1)
to obtain
dH D T dS C V dp (5.4.2)
dA D S dT p dV (5.4.3)
dG D S dT C V dp (5.4.4)
Equations 5.4.1–5.4.4 are sometimes called the Gibbs equations. They are expressions
for the total differentials of the thermodynamic potentials U , H , A, and G in closed sys-
tems of one component in one phase with expansion work only. Each equation shows how
5.4 C LOSED S YSTEMS 140
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 coefficients on the right side of Eqs. 5.4.1–5.4.4, we obtain the fol-
lowing 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

@H
from Eq. 5.4.2: DT (5.4.7)
@S p

@H
DV (5.4.8)
@p S

@A
from Eq. 5.4.3: D S (5.4.9)
@T V

@A
D p (5.4.10)
@V T

@G
from Eq. 5.4.4: D S (5.4.11)
@T p

@G
DV (5.4.12)
@p T
This book now uses for the first 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 (5.4.14)
@y x @x y
The reciprocity relations obtained from the Gibbs equations (Eqs. 5.4.1–5.4.4) are called
Maxwell relations (again valid for a closed system with C D1, P D1, and ¶w 0 D0):

@T @p
from Eq. 5.4.1: D (5.4.15)
@V S @S V

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

@S @p
from Eq. 5.4.3: D (5.4.17)
@V T @T V

@S @V
from Eq. 5.4.4: D (5.4.18)
@p T @T p
5.5 O PEN S YSTEMS 141
5.5 OPEN SYSTEMS
An open system of one substance in one phase, with expansion work only, has three inde-
pendent 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. 5.3.4–5.3.6, we obtain the
following total differentials:
dH D T dS C V dp C dn (5.5.2)
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. 5.4.1–5.4.4) with the addition
of a term dn to allow for a change in the amount of substance.
Identification of the coefficient of the last term on the right side of each of these equa-
tions shows that the chemical potential can be equated to four different partial derivatives:

@U @H @A @G
D D D D (5.5.5)
@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 in-
teresting because it is the rate at which the Gibbs energy increases with the amount of sub-
stance 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 find the total differentials of the thermodynamic
potentials are given by
X
dU D T dS p dV C i dni (5.5.6)
i
X
dH D T dS C V dp C i dni (5.5.7)
i
X
dA D S dT p dV C i dni (5.5.8)
i
X
dG D S dT C V dp C i dni (5.5.9)
i
The independent variables on the right side of each of these equations are the natural vari-
ables of the corresponding thermodynamic potential. Section F.4 shows that all of the infor-
mation contained in an algebraic expression for a state function is preserved in a Legendre
5.6 E XPRESSIONS FOR H EAT C APACITY 142
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. 5.5.6–5.5.9 the fun-
damental 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.3 Problem 5.4 illustrates this useful application of the total
differential of a thermodynamic potential.
In Eqs. 5.5.6–5.5.9, the coefficient i is the chemical potential of species i. The equa-
tions show that i can be equated to four different partial derivatives, similar to the equali-
ties shown in Eq. 5.5.5 for a pure substance:

@U @H @A @G
i D D D D (5.5.10)
@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 defined as the ratio of
an infinitesimal quantity of heat transferred across the boundary under specified conditions
def
and the resulting infinitesimal 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 work—that is, no work at all. The first 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 define 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 ap-
pearing in the equation are state functions whose relations to one another are fixed 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
defined 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 .
3 Ref. [64], p. 86.
5.7 S URFACE W ORK 143
Note that CV is a state function whose value depends on the state of the system—that
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 defined 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 simplified to
dU
CV D (5.6.2)
dT
(closed system, ideal gas)
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 fixed amount of an ideal gas depends only on T (Prob. 5.1), we
can write a relation analogous to Eq. 5.6.2:
dH
Cp D (5.6.4)
dT
(closed system, ideal gas)
5.7 SURFACE WORK
Sometimes we need more than the usual two independent variables to describe an equi-
librium 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 liquid–gas interface behaves somewhat like a stretched membrane. The upper and
lower surfaces of the liquid film in the device depicted in Fig. 5.1 on the next page 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 rod position x. The force also depends on the temperature and pressure.
5.8 C RITERIA FOR S PONTANEITY 144
l F Fext
Figure 5.1 Device to measure the surface tension of a liquid film. The film is
stretched between a bent wire and a sliding rod.
The surface tension or interfacial tension, , is the force exerted by an interfacial sur-
face per unit length. The film shown in Fig. 5.1 has two surfaces, so we have D F=2l
where l is the rod length.
To increase the surface area of the film 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 coefficient 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. Iden-
tifying the coefficient of the last term on the right side as a partial derivative, we find 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
D (5.7.4)
@T p;As @As T;p
It is valid to replace the partial derivative on the left side by .@ =@T /p because is in-
dependent 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 first 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 pro-
cess is irreversible. For example, the volume might be changing at a finite 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
or
¶q T dS (5.8.2)
( irrev
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 first 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)
( irrev
We substitute this relation for dU into the differentials of enthalpy, Helmholtz energy, and
Gibbs energy given by Eqs. 5.3.4–5.3.6 to obtain three more relations:
dH T dS C V dp C ¶w 0 (5.8.4)
( irrev
dA S dT p dV C ¶w 0 (5.8.5)
( irrev
dG S dT C V dp C ¶w 0 (5.8.6)
( irrev
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 expan-

sion work only, the Gibbs energy continuously decreases until the system reaches an
equilibrium state.
Ben-Amotz and Honig4 developed a “rectification” procedure that simplifies the math-
ematical 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 first law expression dU D ¶q p dV C¶w 0 ,
produces
dU D T dS p dV C ¶w 0 T ¶ (5.8.8)
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. 5.3.4–5.3.6, we
obtain equalities equivalent to Eqs. 5.8.4–5.8.6, for example
dG D S dT C V dp C ¶w 0 T ¶ (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.
4 Refs. [11] and [81].
P ROBLEMS 147
PROBLEMS
in Appendix I.
5.1 Show that the enthalpy of a fixed 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 fixed amount of
an ideal gas are functions only of T .
5.3 During the reversible expansion of a fixed 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 sufficient condition for this expression to be an exact differential is that
the reciprocity relation must be satisfied for the independent variables T and V (see Ap-
pendix 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 func-
tion 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 finite 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 confined 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 148
gas
liquid
Figure 5.2
pressure equal to the vapor pressure of the liquid, 2:50 105 Pa, and some of the liquid vapor-
izes. Assume that the process is adiabatic and that T and p remain uniform and constant. The
final 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 =ı C =10 6
J cm 2
15 7:350
20 7:275
25 7:199
30 7:120
35 7:041
a Ref. [163].
5.6 Use the data in Table 5.1 to evaluate .@S=@As /T;p at 25 ı C, which is the rate at which the
entropy changes with the area of the air–water 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,
find 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 (6.0.1)
T!0
(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 finite 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.”
1 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. [104], p. 448.
149
CHAPTER 6 THE THIRD LAW AND CRYOGENICS
6.1 T HE Z ERO OF E NTROPY 150
BIOGRAPHICAL SKETCH
Walther Hermann Nernst (1864–1941)
They had two sons, both killed in World War I,
and three daughters.
Nernst wrote an influential textbook of
physical chemistry, the second in the field, en-
titled Theoretische Chemie vom Standpunkte
der Avogadroschen Regel und der Thermody-
namik. It was first published in 1893 and its
last edition was in 1926.
Nernst began work in 1893 on a novel elec-
tric 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 com-
chemist 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 Nernst’s laboratory, was Chemistry for the year 1920 “in recognition
one of those who developed an enmity to- of his work in thermochemistry.” In his No-
ward him; in later years, Lewis delighted in bel Lecture, describing the heat theorem, he
pointing out what he considered to be errors said:
in Nernst’s writings. The American physicist . . . in all cases chemical affinity and evolution
Robert A. Millikan, who studied with Nernst of heat become identical at low temperatures.
at Göttingen, wrote in a memorial article:a Not, and this is the essential point, in the sense
He was a little fellow with a fish-like mouth and that they intersect at absolute zero, but rather in
other well-marked idiosyncrasies. However, he the sense that they invariably become practically
was in the main popular in the laboratory, despite identical some distance before absolute zero is
the fact that in the academic world he nearly al- reached; in other words the two curves become
ways had a quarrel on with somebody. He lived mutually tangential in the vicinity of absolute
on the second floor of the institute with his wife zero.
and three young children. As we students came If we frame this principle in quite general
to our work in the morning we would not infre- terms, i.e. if we apply it not only to chemical
quently meet him in his hunting suit going out but to all processes, then we have the new heat
for some early morning shooting. . . . His great- theorem which gives rise to a series of very far-
est weakness lay in his intense prejudices and the reaching consequences . . .
personal, rather than the objective, character of
Nernst would have nothing to do with the
some of his judgments. Nazis. When they passed the 1933 law bar-
At Leipzig University, in 1888, he pub- ring Jews from state employment, he refused
lished the Nernst equation, and in 1890 the to fire the Jewish scientists at the Berlin insti-
Nernst distribution law. tute, and instead took the opportunity to retire.
In 1891 he moved to the University of He caused a stir at a meeting by refusing to
Göttingen, where in 1895 he became director stand for the singing of the Horst Wessel Lied.
of the Göttingen Physicochemical Institute. Before his death he ordered that letters he had
In 1892 Nernst married Emma Lohmeyer, received be burned, perhaps to protect his cor-
daughter of a Göttingen medical professor. respondents from the Nazi authorities.b
a Ref. [117]. b Ref. [35].
6.2 M OLAR E NTROPIES 151
According to this principle, every substance (element or compound) in a pure, perfectly-

ordered 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 definition 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
Cp X ntrs H
S D dT C (6.2.1)
0 T Ttrs
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 177) combined with the experimental observation that
the cubic expansion coefficient ˛ approaches zero as T approaches zero kelvins.
The resulting operational equation for the calculation of the molar entropy of the substance
at the temperature and pressure of interest is
T0 Cp;m
S X trs H
Z
0
Sm .T / D D dT C (6.2.2)
n 0 T Ttrs
(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 difficult to make measurements below this temperature. Statistical me-
chanical 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 ap-
propriate 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 con-
stant 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 Cp;m
Z T 00 Z T0 Cp;m
2
dT D a T dT C dT (6.2.3)
0 T 0 T 00 T
where T 00 is the lowest temperature at which Cp;m is measured. The first term on the right
side of Eq. 6.2.3 is
Z T 00
ˇT 00
a T 2 dT D .aT 3 =3/ˇ D a.T 00 /3 =3 0
(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 /
Z T0 Cp;m X trs H
0
Sm .T / D C dT C (6.2.5)
3 T 00 T Ttrs
(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
70 0:5
60
0:4
2 mol 1
1 mol 1
50
0:3
40
=J K
Cp;m =J K
30

0:2
Cp;m =T
20
0:1
10
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
T =K T =K
70 200
60
150
1 mol 1
50
1 mol 1
40
100
Cp;m =J K
Sm =J K
30
20
50
10
0 0
1 2 3 4 5 6 0 50 100 150 200 250 300
ln.T =K/ T =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 tempera-
ture. (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 Cp;m
Z T DT 0
dT D Cp;m d ln.T =K/ (6.2.6)
T 00 T T DT 00
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 find the dependence of Sm on T . Note how
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 randomly-
mixed 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 perfectly-
ordered 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 final 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 scale—that 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 Sackur–Tetrode equation:
.2M /3=2 .RT /5=2

Sm;trans D R ln C .5=2/R (6.2.7)
ph3 NA4
Here h is the Planck constant and NA is the Avogadro constant. The internal contribu-
tion 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 defined by
X
qint D exp . i =kT / (6.2.9)
i
In Eq. 6.2.9, i is the energy of a molecular quantum state relative to the lowest en-
ergy 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 lev-
els, values of electronic degeneracies, fundamental vibrational frequencies, rotational
constants, and other spectroscopic parameters.
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
Smı =J K 1
mol 1
1 1
Substance calorimetric spectroscopica Sm,0 =J K mol
HCl 186:3 ˙ 0:4b 186:901 0:6 ˙ 0:4
CO 193:4 ˙ 0:4c 197:65 ˙ 0:04 4:3 ˙ 0:4
NO 208:0 ˙ 0:4d 210:758 2:8 ˙ 0:4
N2 O (NNO) 215:3 ˙ 0:4e 219:957 4:7 ˙ 0:4
H2 O 185:4 ˙ 0:2f 188:834 ˙ 0:042 3:4 ˙ 0:2
a Ref. [28]. b Ref. [60]. c Ref. [34]. d Ref. [83]. e Ref. [15]. f Ref. [59].
When the spectroscopic method is used to evaluate Sm with p set equal to the standard
pressure p ı D 1 bar, the value is the standard molar entropy, Smı , 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 Smı 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 Smı 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 meth-
ods. This agreement is typical of most substances, particularly those like HCl whose
molecules are polar and asymmetric with a large energetic advantage of forming perfectly-
ordered 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 Smı 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 find 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 156
(a) ! 0
T ;p
0
V
0
1 V
1
00
T
00
;p
00
!
(b) ! 0
T ;p
0
V
0
2 V
00
2 T
00
;p
00
!
Figure 6.2 Joule–Thomson expansion of a gas through a porous plug. The shaded
area represents a fixed-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 struc-
ture 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 almost-
symmetric 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 ar-
rangement 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 Smı 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 field 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 Joule–Thomson 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 evapora-
tion procedure with a refrigerant fluid is what produces the cooling in an ordinary kitchen
refrigerator.
For further cooling of a fluid, a common procedure is to use a continuous throttling
process in which the fluid is forced to flow through a porous plug, valve, or other con-
striction that causes an abrupt drop in pressure. A slow continuous adiabatic throttling of
a gas is called the Joule–Thomson experiment, or Joule–Kelvin experiment, after the two
scientists who collaborated between 1852 and 1862 to design and analyze this procedure.5
The principle of the Joule–Thomson 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 157
side of a porous plug and at a constant lower pressure p 00 at the right side. Because of the
pressure difference, the gas flows continuously from left to right through the plug. The flow
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 flow 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 fixed 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 p 0 .V20 V10 / p 00 .V200 V100 / (6.3.1)
where the meaning of the volumes V10 , V20 , and so on is indicated in the figure.
The internal energy change U of the gas segment must be equal to w, since q is zero.
Now let us find 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
H D U C .p 0 V20 C p 00 V200 / .p 0 V10 C p 00 V100 / (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 Joule–Thomson 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 Joule–Thomson coefficient
(or Joule–Kelvin coefficient) defined by

def @T
JT D (6.3.3)
@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 Joule–Thomson coefficient to other properties of a gas.
6.3 C RYOGENICS 158
S
bc
A
bc
B bc
Figure 6.3 Adiabatic demagnetization to achieve a low temperature in a paramag-

netic 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 Joule–Thomson expansion to cause further cooling.
The cooling effect of a Joule–Thomson 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 Joule–Thomson coefficient, so that it in turn can be converted by a throttling
process to liquid hydrogen at 20:3 K. Finally, gaseous helium, whose Joule–Thomson coef-
ficient 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 flux density and mmag is the mag-
nitude 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 be-
low 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
sufficiently separated that at 1 K the orientations of the magnetic dipoles are almost com-
pletely 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 tem-
perature dependence of the entropy of a paramagnetic solid in the absence of an applied
6.3 C RYOGENICS 159
BIOGRAPHICAL SKETCH
William Francis Giauque (1895–1982)
graduate students and collaborators, he began
to publish papers on various aspects of the
AIP EMILIO SEGRÈ VISUAL ARCHIVES

third law.b The research included the evalua-
BRITTLE BOOKS COLLECTION

tion of third-law molar entropies and compar-
ison to molar entropies calculated from spec-
troscopic 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 un-
known 17 O and 18 O isotopes of oxygen.
Giauque’s best-known accomplishment is
William Giauque was an American chemist his invention and exploitation of cooling to
who made important contributions to the field very low temperatures by adiabatic demagne-
of cryogenics. He received the 1949 Nobel tization. In 1924, he learned of the unusual
Prize in Chemistry “for his contributions in the properties of gadolinium sulfate octahydrate at
field of chemical thermodynamics, particularly the temperature of liquid helium. In his Nobel
concerning the behaviour of substances at ex- Lecture, he said:c
tremely low temperatures.” I was greatly surprised to find, that the applica-
Giauque was born in Niagara Falls, On- tion of a magnetic field removes a large amount
tario, Canada, but as his father was a citizen of of entropy from this substance, at a temperature
the United States, William was able to adopt so low that it had been thought that there was
American citizenship. His father worked as a practically no entropy left to remove. . . . Those
weighmaster and station agent for the Michi- familiar with thermodynamics will realize that
gan Central Railroad in Michigan. in principle any process involving an entropy
Giauque’s initial career goal after high change may be used to produce either cooling or
heating. Accordingly it occurred to me that adia-
school was electrical engineering. After work-
batic demagnetization could be made the basis of
ing for two years at the Hooker Electrochem-
a method for producing temperatures lower than
ical Company in Niagara Falls, New York, he those obtainable with liquid helium.
left to continue his education with the idea of
becoming a chemical engineer. Hearing of the It wasn’t until 1933 that he was able to build
scientific reputation of G. N. Lewis, the chair the equipment and publish the results from his
of the chemistry department at the University first adiabatic demagnetization experiment, re-
of California at Berkeley (page 270), and mo- porting a temperature of 0:25 K.d
tivated by the free tuition there, he enrolled in- He was married to a botanist and had two
stead in that department.a sons. According to one biography:e
Giauque spent the rest of his life in the Giauque’s students remember pleasant Thanks-
chemistry department at Berkeley, first as an giving dinners at the Giauque home, with Muriel
[his wife] as cook and Frank (as she called him)
undergraduate; then as a graduate student; and
as raconteur, with a keen sense of humor. The
finally, after receiving his Ph.D. in 1922, as stories he most enjoyed telling were those in
a faculty member. Some of his undergradu- which the joke was on him. . . . Giauque’s con-
ate courses were in engineering, which later servatism was legendary. He always appeared at
helped him in the design and construction of the university dressed in an iron-gray tweed suit.
the heavy equipment for producing the high . . . A dominant personality himself, Giauque not
magnetic fields and the liquid hydrogen and only tolerated but respected students who dis-
helium needed in his research. agreed with him, and he was especially pleased
Beginning in 1928, with the help of his when they could prove their point.
a Ref. [132]. b Ref. [154]. c Ref. [61]. d Ref. [58]. e Ref. [132].
6.3 C RYOGENICS 160
magnetic field, and the dashed curve is for the solid in a constant, finite magnetic field. 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 field. 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 sur-
rounding the solid, and the sample is slowly moved away from the magnetic field. 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 158).
To change the independent variables in Eq. 6.3.4 to S, p, and B, we define 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 find the reciprocity relation
@mmag

@T
D (6.3.7)
@B S;p @S p;B
According to Curie’s law of magnetization, the magnetic dipole moment mmag of a

paramagnetic phase at constant magnetic flux 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 find the sign of the entropy change during the isothermal magnetization process
shown as path A in Fig. 6.3 on page 158. In order to use T , p, and B as the independent
def
variables, we define the Legendre transform G 0 D H 0 T S. Its total differential is
dG 0 D S dT C V dp mmag dB (6.3.8)
From this total differential, we obtain the reciprocity relation
@mmag

@S
D (6.3.9)
@B T;p @T p;B
6.3 C RYOGENICS 161
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 de-
magnetization, 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. How-
ever, as is evident from the figure, 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 finite number of stages of adiabatic demagnetization. This
conclusion is called the principle of the unattainability of absolute zero.
P ROBLEM 162
PROBLEM
6.1 Calculate the molar entropy of carbon disulfide at 25:00 ı 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
1 1
T =K Cp;m =J K mol
15:05 6:9
20:15 12:0
29:76 20:8
42:22 29:2
57:52 35:6
75:54 40:0
94:21 45:0
108:93 48:5
131:54 52:6
156:83 56:6
a Ref. [25].
161:11 K, the molar enthalpy of fusion is fus H D 4:39 103 J mol 1 . The molar heat capac-
ity of the liquid in the range 161–300 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 defined as follows:

def 1 @V
cubic expansion coefficient ˛ D (7.1.1)
V @T p

def 1 @V
isothermal compressibility T D (7.1.2)
V @p T
The cubic expansion coefficient is also called the coefficient of thermal expansion and
the expansivity coefficient. Other symbols for the isothermal compressibility are ˇ and
T.
These definitions 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 expres-
sions
1 The cubic expansion coefficient is not always positive. ˛ is negative for liquid water below its temperature
of maximum density, 3:98 ı 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. [110]). The intermetallic compound YbGaGe has been found to have a value of ˛ that is practically zero
in the range 100–300 K (Ref. [148]).
163
CHAPTER 7 PURE SUBSTANCES IN SINGLE PHASES
7.1 VOLUME P ROPERTIES 164
3:0
idea H2
l ga O(g
s )
2:5
2:0
)
ane(l
1
1:5 n-hex
K
.
3
˛=10
1:0 nol(l)
1-propa
0:5 (l)
H 2O
H2 O(s)
0:0
0:5
25 0 25 50 75 100 125 150
t=ı C
Figure 7.1 The cubic expansion coefficient 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 ı C, ˛ is zero at that temperature.
a Based on data in Ref. [49], p. 104; Ref. [70]; and Ref. [167], p. 28.

1 @Vm
˛D (7.1.3)
Vm @T p (pure substance, P D1)

1 @Vm
T D (7.1.4)
Vm @p T (pure substance, P D1)
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 dif-
ferential 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 165
15
l)
ol(
th an
e )
e(l
zen
1
10 n
be
bar
5
T =10
5 H2 O(l)
H2 O(s)
.
0
25 0 25 50 75 100
t=ı C
Figure 7.2 The isothermal compressibility of several substances as a function of tem-

perature at p D 1 bar. (Based on data in Ref. [49]; Ref. [88]; and Ref. [167], p. 28.)
With the substitutions .@V =@T /p D ˛V (from Eq. 7.1.1) and .@V =@p/T D T V (from
Eq. 7.1.2), the expression for the total differential of V becomes
dV D ˛V dT T V dp (7.1.6)
(closed system,
C D1, P D1)
To find 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 ˛
D (7.1.7)
@T V T
(closed system,
C D1, P D1)
7.2 INTERNAL PRESSURE
The partial derivative .@U=@V /T applied to a fluid 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 166
obtain

@U @p
DT p (7.2.1)
@V T @T V (closed system,
fluid phase, C D1)
This equation is sometimes called the “thermodynamic equation of state” of the fluid.
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 (7.2.3)
@V T
(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 defined 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 first part of this definition, expresses the second part. It thus appears that the second
part of the definition is redundant, and that we could define an ideal gas simply as a
gas obeying the ideal gas equation. This argument is valid only if we assume the ideal-
gas 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 can’t
define an ideal gas solely by pV D nRT , where T is the ideal gas temperature.
Here is a simplified interpretation of the significance of the internal pressure. When

the volume of a fluid 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 fluid 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

@U ˛T
D p (7.2.4)
@V T T
(closed system,
fluid 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 167
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
dU D ¶q CV D (7.3.1)
@T V
and for a process at constant pressure we have4

@H
dH D ¶q Cp D (7.3.2)
@T p
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 definitions of
CV and Cp because they are expressed with two independent variables—T and V for CV ,
and T and p for Cp . As mentioned on page 142, additional conditions would have to be
specified to define CV for a more complicated system; the same is true for Cp .
For a closed system of an ideal gas we have5
dU dH
CV D Cp D (7.3.3)
dT dT
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 fixed amount of an ideal gas. Using substitutions from Eq. 7.3.3, we write
dH dU d.H U / d.pV /
Cp CV D D D D nR (7.3.4)
dT dT dT dT
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 (7.3.6)
@T p @T p @T p @T p
Then we write the total differential of U with T and V as independent variables and identify
one of the coefficients 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.
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

@U @V
Cp D CV C Cp (7.3.9)
@V T @T p
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
˛2T V
Cp D CV C (7.3.10)
T
and divide by n to obtain molar quantities:
˛ 2 T Vm
Cp;m D CV;m C (7.3.11)
T
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 flow 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 sufficient 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.
T2 bc
T2 T1 r .t2 t1 /
T1 bc
t
t1 t2
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 88: 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 first 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 first 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 figure 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 figure.
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
dT dT dU 1 ¶wcont
rD D D (7.3.14)
dt dU dt dt
The rate of the continuous work is therefore ¶wcont = dt D r. This rate is constant through-
out 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 find 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
U w C r.t2 t1 /
D D el (7.3.16)
T2 T1 T2 T1
Solving for , we obtain

wel
D (7.3.17)
T2 T1 r.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 inter-
val 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 eval-
uate a quantity defined 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.
T
T1
T2 bc
slope = r2
T1 bc
slope = r1
t
t1 t2
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 elimi-
nated by evacuating the gap. Then heat transfer is by conduction and radiation, and its rate
is given by Newton’s law of cooling
¶q
D k.T Text / (7.3.20)
dt
where k is a constant (the thermal conductance). Heat flows 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 first law and Eq. 7.3.20, we obtain the following relation for the rate at
which the internal energy changes:
dU ¶q ¶w dV ¶wel ¶wcont
D C D k.T Text / p C C (7.3.21)
dt dt dt dt dt dt
For heating at constant volume (dV = dt D 0), this relation becomes
dU ¶wel ¶wcont
D k.T Text / C C (7.3.22)
dt dt dt
(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 flow. Times t1 and t2 are before and after the heater circuit is closed. In any time
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 ¶wel
D k.T T1 / C (7.3.23)
dt dt
(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
U D k .T T1 / dt C wel (7.3.24)
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
wel
D Z t2 (7.3.26)
.T2 T1 / C .k=/ .T T1 / dt
t1
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
dH dU dV ¶wel ¶wcont
D Cp D k.T Text / C C (7.3.27)
dt dt dt dt dt
(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 .T T1 / dt C wel (7.3.28)
t1
(constant p)
At constant p, the energy equivalent is equal to Cp D H=.T2 T1 /, and the final expres-
sion 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 T1 / r2 D .k=/.T2 T1 / (7.3.30)
The solutions are

r1 r2 r1 T2 r2 T1
.k=/ D T1 D (7.3.31)
T2 T1 r1 r2
Finally, k is given by
r1 r2
k D .k=/ D (7.3.32)
T2 T1
When the pressure is constant, this procedure yields the same relations for k=, T1 , and k.
Continuous-flow calorimeters
A flow calorimeter is a third type of calorimeter used to measure the heat capacity of a fluid
phase. The gas or liquid flows 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 flow 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 fluid (where M is the molar
mass). If heat flow is negligible, the molar heat capacity of the substance is given by
w M.¶wel = dt/
Cp;m D D (7.3.33)
nT T .dm= dt/
To correct for the effects of heat flow, T is usually measured over a range of flow rates
and the results extrapolated to infinite flow rate.
7.3.3 Typical values

Figure 7.5 on the next page shows the temperature dependence of Cp;m for several sub-
stances. The discontinuities seen at certain temperatures occur at equilibrium phase transi-
tions. At these temperatures the heat capacity is in effect infinite, since the phase transition
of a pure substance involves finite heat with zero temperature change.
7.4 H EATING AT C ONSTANT VOLUME OR P RESSURE 174
80
H2 O(l)
60 N2 (l)
Cp;m =J K 1 mol 1
. 40 H2 O(g)
N2 (g)
N2 (s)
20
H2 O(s) C(s)
0
0 100 200 300 400 500 600
T =K
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 infinitesimal change of U is given
by
dU D CV dT (7.4.1)
(closed system,
C D1, P D1, constant V )
and the finite change of U between temperatures T1 and T2 is

Z T2
U D CV dT (7.4.2)
T1
(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 finite change of a state function, U , by inte-
grating 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 fluid 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 175
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 final 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 final 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 finite 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 infinitesimal or finite
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 (7.4.3)
and Eq. 7.4.2 becomes
Z T2
Um D CV;m dT (7.4.4)
T1
If, at a fixed volume and over the temperature range T1 to T2 , the value of CV is essen-
tially 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 infinitesimal 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
CV
dS D dT (7.4.6)
T
(closed system,
Integration yields the finite change

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

T2
S D CV ln (7.4.8)
T1
P D1, constant V and CV )
7.5 PARTIAL D ERIVATIVES WITH R ESPECT TO T, p, AND V 176
(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
Z T2
H D Cp dT (7.4.9)
T1
(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
Cp
dS D dT (7.4.11)
T
(closed system,
Integration gives
Z T2 Cp
S D dT (7.4.12)
T1 T
(closed system,
or, with Cp treated as constant,
T2
S D Cp ln (7.4.13)
T1
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
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 General Ideal Partial General Ideal

derivative expression gas derivative expression gas

@p 1 p @A
T pV V
@V T T V V @p T

@V V @A
T V p p
@p T p @V T

@U @G
. ˛T C T p/V 0 V V
@p T @p T

@U ˛T @G 1
p 0 p
@V T T @V T
T
@H @S V
.1 ˛T /V 0 ˛V
@p T @p T T

@H ˛T 1 @S ˛ p
0
@V T T @V T T 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 differen-
tial of any of the eight quantities, and (2) for expressing the finite 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
dS D
dT ˛V dp (7.5.2)
T
Furthermore, the first expression is equivalent to the differential form
Cp
dS D
dT (7.5.3)
T
provided p is constant; we can integrate this equation to obtain the finite 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.
We may derive the general expressions as follows. We are considering differentia-

tion 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
1
.@y=@x/z D (7.5.4)
.@x=@y/z
Table 7.2 Constant pressure: expressions for partial derivatives of state func-
tions with respect to temperature and volume in a closed, single-phase system


@T 1 T @A pV
˛pV S S
@V p ˛V V @T p T

@V V @A S TS
˛V p p
@T p T @V p ˛V V

@U @G
Cp ˛pV CV S S
@T p @T p
Cp

@U CV T @G S TS
p
@V p ˛V V @V p ˛V V
Cp Cp

@H @S
Cp Cp
@T p @T p T T
Cp Cp T Cp Cp

@H @S
@V p ˛V V @V p ˛T V 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


@T T T @A
S S
@p ˛ p @T
V V
@p ˛ p @A T S TS
@T V T T @p V ˛ p

@U @G ˛V pV
CV CV S S
@T V @T V T T
T Cp

@U T CV @G T S TS
˛T V V V
@p V ˛ p @p V ˛ p

@H ˛V @S CV CV
Cp C .1 ˛T / Cp
@T V T @T T T
V
T Cp Cp T T Cp

@H @S CV
C V .1 ˛T / ˛V
@p V ˛ p @p V ˛T 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
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 Joule–Thomson coefficient

The Joule–Thomson coefficient of a gas was defined in Eq. 6.3.3 on page 157 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 @H dT @H
D C (7.5.7)
dp @T p dp @p T
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 .@H=@p/T
D (7.5.9)
@p H .@H=@T /p
The left side of this equation is the Joule–Thomson coefficient. 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
.˛T 1/V .˛T 1/Vm

JT D D (7.5.10)
Cp Cp;m
7 Ref. [22]; Ref. [21], p. 199–241.
7.6 I SOTHERMAL P RESSURE C HANGES 180
Table 7.4 Changes of state functions during an isothermal pressure change in a

closed, single-phase system
State function Approximate expression

change General expression Ideal gas for liquid or solid
Z p2
U . ˛T C T p/V dp 0 ˛T Vp
p1
Z p2
H .1 ˛T /V dp 0 .1 ˛T /Vp
Zp1p2
p2
A T pV dp nRT ln T V .p22 p12 /=2
p1
Zp1p2
p
G V dp nRT ln 2 Vp
p1 p1
Z p2
p2
S ˛V dp nR ln ˛Vp
p1 p1
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.
R expressions by integrating expressions found in Table 7.1. For example, U
We obtain the
is given by .@U=@p/T dp. The results are listed in the second column of Table 7.4.
7.6.1 Ideal gases

Simplifications result when the phase is an ideal gas. In this case, we can make the substi-
tutions 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 165), 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 181
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 ı . 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 recom-
mends the value p ı D 1 bar (exactly 105 Pa). This book uses the latter value unless stated
otherwise. (Note that there is no defined standard temperature.)
A superscript degree symbol (ı) denotes a standard quantity or standard-state condi-
tions. An alternative symbol for this purpose, used extensively outside the U.S., is a super-
script 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 recommenda-
tions 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 ı 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 ı .
The standard state of a pure liquid or solid is the unstressed liquid or solid at pres-
sure p ı 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 definitions (page 141)
def G
D Gm D (7.8.1)
n
(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 definition 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 fixed 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. 61–62.
7.8 C HEMICAL P OTENTIAL AND F UGACITY 182
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)
(pure substance, P D1)
(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 coefficients of the terms on the right side of Eq. 7.8.2 as the partial
derivatives

@
D Sm (7.8.3)
@T p
(pure substance, P D1)
and

@
D Vm (7.8.4)
@p T (pure substance, P D1)
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, ı , 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 ı .
There is no way we can evaluate the absolute value of at a given temperature and
pressure, or of ı at the same temperature,10 but we can measure or calculate the difference
ı . The general procedure is to integrate d D Vm dp (Eq. 7.8.2 with dT set equal to
zero) from the standard state at pressure p ı to the experimental state at pressure p 0 :
Z p0
0 ı
.p / D Vm dp (7.8.5)
pı
(constant T )
7.8.1 Gases
For the standard chemical potential of a gas, this book will usually use the notation ı (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 find 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 p0
RT =p in Eq. 7.8.5: .p 0 / ı D pı .RT =p/ dp D RT ln.p 0 =p ı /. The general relation
for as a function of p, then, is
p
D ı (g) C RT ln (7.8.6)
pı
(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 ı include the molar internal energy
10 At
whose absolute value can only be calculated from the Einstein relation; see Sec. 2.6.2.

C b
B b
Æ (g) A b
0 pÆ f .p0/ p0 p
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 defined by an equation with the same
form as Eq. 7.8.6:
f
D ı (g) C RT ln (7.8.7)
pı
(pure gas)
or
ı (g)

def ı
f D p exp (7.8.8)
RT
(pure gas)
Note that fugacity has the dimensions of pressure. Fugacity is a kind of effective pressure.
Specifically, 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 pressure–volume 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
use Eq. 7.8.5 to write 0 ı (g) D RT ln.f 0 =p ı / and 00 ı (g) D RT ln.f 00 =p ı /, from
which we obtain
f0
0 00 D RT ln 00 (7.8.9)
f
By integrating d D Vm dp from pressure p 00 to pressure p 0 , we obtain

Z 0 Z p0
0 00 D d D Vm dp (7.8.10)
00 p 00
Equating the two expressions for 0 00 and dividing by RT gives

p0
f0 Vm
Z
ln D dp (7.8.11)
f 00 p 00 RT
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 difficult to evaluate. We avoid this difficulty by
subtracting from the preceding equation the identity
p0
p0
Z
dp
ln 00 D (7.8.12)
p p 00 p
which is simply the result of integrating the function 1=p from p 00 to p 0 . The result is
p0
f 0 p 00

Vm 1
Z
ln 00 0 D dp (7.8.13)
f p p 00 RT p
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 0 RT p
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 coefficient of a gas is defined by
def f
D or f D p (7.8.15)
p
(pure gas)
The fugacity coefficient at pressure p 0 is then given by Eq. 7.8.14:

p0
Vm 1
Z
0
ln .p / D dp (7.8.16)
0 RT p
(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
Vm CB (7.8.17)
p
7.9 S TANDARD M OLAR Q UANTITIES OF A G AS 185
Here B is the second virial coefficient, a function of T . With this equation of state, Eq.
7.8.16 becomes
Bp
ln (7.8.18)
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 coefficient from an experimental equation of state or a second virial coefficient.
We can then find 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 coefficient and the dimensionless ratio f =p ı 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
ı C Vm .p pı/ (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 influence 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 0 / D ı (g) C RT ln
pı
p0 f .p 0 /
D ı (g) C RT ln ı C RT ln (7.9.1)
p p0
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

0 ı RT
.p / D (g) C RT ln ı C Vm dp (7.9.2)
p 0 p
(pure gas)
Note that this expression for is not what we would obtain by simply integrating d D
Vm dp from p ı to p 0 , because the real gas is not necessarily in its standard state of ideal-gas
behavior at a pressure of 1 bar.
Table 7.5 Real gases: expressions for differences between molar properties and standard molar
values at the same temperature
General expression Equation of state

Difference at pressure p 0 V D nRT =p C nB
Z p0 " #
@Vm dB
Um ı
Um (g) Vm T dp C RT p 0 Vm pT
0 @T p dT
Z p0 " #
@Vm dB
Hm Hmı (g) Vm T dp p B T
0 @T p dT
Z p0
p0

RT p
Am Aım (g) RT ln ı C Vm dp C RT p 0 Vm RT ln
p 0 p pı
Z p0
p0

RT p
Gm Gmı (g) RT ln ı C Vm dp RT ln C Bp
p 0 p pı
Z p 0 " #
p0

ı @Vm R p dB
Sm Sm (g) R ln ı dp R ln p
p 0 @T p p pı dT
Z p0 2
ı @ Vm d2 B
Cp;m Cp;m (g) T dp pT
0 @T 2 p 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 ı (g) of the gas is the standard molar Gibbs
ı (g). Therefore Eq. 7.9.2 can be rewritten in the form
energy, Gm
Z p0
p0

0 ı RT
Gm .p / D Gm (g) C RT ln ı C Vm dp (7.9.3)
p 0 p
The middle column of Table 7.5 contains an expression for Gm .p 0 / Gm ı (g) taken from this
equation. This expression contains all the information needed to find 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 to ı (g):
ı
ı @ (g)
Sm (g) D (7.9.5)
@T p
By substituting the expression for given by Eq. 7.9.2 into Eq. 7.9.4 and comparing the
result with Eq. 7.9.5, we obtain
0 Z p0 " #
p @V R
Sm .p 0 / D Smı (g) R ln ı m
dp (7.9.6)
p 0 @T p p
The expression for Sm Smı (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 Smı (g).
Now we can use the expressions for Gm and Sm to find 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 Xmı (g) as a function of Gm ı (g) and S ı (g). Substitutions
m
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 Xmı (g) taken.
For example, the expression for Um Umı (g) in the middle column Table 7.5 was derived
as follows. The equation defining 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
Umı (g) D Gm
ı
(g) C T Smı (g) p ı Vmı (g) (7.9.8)
where from the ideal gas law p ı Vmı (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 Umı (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 coefficient (Eq. 7.8.17). Equation 7.9.2 then becomes
p
ı (g) C RT ln C Bp (7.9.9)
pı
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 188
PROBLEMS
in Appendix I.
7.1 Derive the following relations from the definitions 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

RT
Vm D CB
p
(Eq. 2.2.8), where B is the second virial coefficient (a function of T ), find 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 Joule–Thomson coefficient is given by
RT 2 ŒdBp = dT C .dCp = dT /p C
JT D
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 coefficient. James Joule attempted to evaluate
this quantity by measuring the temperature change accompanying the expansion of air into a
vacuum—the “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 coefficient is equal to
p ˛T =T
CV
7.6 p–V –T data for several organic liquids were measured by Gibson and Loeffler.11 The follow-
ing formulas describe the results for aniline.
Molar volume as a function of temperature at p D 1 bar (298–358 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 c D 1:0443 10 4 cm3 K 2 mol 1
b D 0:08852 cm3 K 1 mol 1

d D 1:940 10 7 cm3 K 3 mol 1
Molar volume as a function of pressure at T D 298:15 K (1–1000 bar):
Vm D e f ln.g C p=bar/
where the parameter values are
e D 156:812 cm3 mol 1
f D 8:5834 cm3 mol 1
g D 2006:6
11 Ref. [65].
P ROBLEMS 189
(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 fixed 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 ı C and 500 bar, given that
the value of ˛ is 1:2 10 3 K 1 at 25 ı C and 1 bar. (Use information from Fig. 7.2 on
page 165.)
7.9 Certain equations of state supposed to be applicable to nonpolar liquids and gases are of the
form p D Tf .Vm / a=Vm2 , where f .Vm / is a function of the molar volume only and a is a
constant.
(a) Show that the van der Waals equation of state .p C a=Vm2 /.Vm b/ D RT (where a and
b are constants) is of this form.
(b) Show that any fluid with an equation of state of this form has an internal pressure equal
to a=Vm2 .
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

1 1 2 1
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 K–2000 K is given by
ı c
Cp;m D a C bT C
T2
1 1 3 2 1
a D 44:2 J K mol b D 8:8 10 JK mol cD 8:6 105 J K mol 1
Calculate the enthalpy and entropy changes when one mole of CO2 is heated at 1 bar from
ı
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 190
at the second virial coefficient, B. In the temperature range 300 K–800 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 1
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
ı
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 specific 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 ı C,
she removed a 60:17-g block of aluminum metal from a boiling water bath at 100:00 ı C and
quickly immersed it in the liquid in the container. After the contents of the container had
become thermally equilibrated, she recorded a final temperature of 27:92 ı C. She calculated
the specific 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, find the specific 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 final 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 coefficient, , of this gas as a function of p.
(b) For CO2 (g) at 0:00 ı C, the virial coefficients have the values Bp D 6:67 10 3 bar 1
and Cp D 3:4 10 5 bar 2 . Evaluate the fugacity f at 0:00 ı 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 ı C and
12 pressures.
(a) Evaluate the fugacity coefficient and fugacity of H2 O(g) at 400:00 ı C and 200 bar.
(b) Show that the second virial coefficient B in the virial equation of state, pVm D RT .1 C
B=Vm C C =Vm2 C /, is given by

Vm 1
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 ı C.
P ROBLEMS 191
Table 7.6 Molar volume of H2 O(g) at 400:00 ı C a
p=105 Pa Vm =10 3
m3 mol 1
p=105 Pa Vm =10 3
m3 mol 1
1 55:896 100 0:47575

10 5:5231 120 0:37976
20 2:7237 140 0:31020
40 1:3224 160 0:25699
60 0:85374 180 0:21447
80 0:61817 200 0:17918
a based 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 un-
dergo 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 equi-
librium state.
A general procedure will now be introduced for finding conditions for equilibrium with
given constraints. The procedure is applied to phase equilibria of single-substance, mul-
tiphase 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 five 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
’ 0 P ’
P ’
dS DPdS ’¤’0 dS . (This is valid because entropy is extensive: S D ’S ,
dS D ’ dS .)’
4. Rearrange to obtain an expression for the total differential of the entropy consistent
with the reduced number of independent variables.
192
CHAPTER 8 PHASE TRANSITIONS AND EQUILIBRIA OF PURE SUBSTANCES
8.1 P HASE E QUILIBRIA 193
5. The conditions for an equilibrium state are those that make the infinitesimal entropy
change, dS, equal to zero for all infinitesimal 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 field 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 Pwill be the reference phase. SinceP internal energy is extensive, we can write
’ 0 ’ ’ 0 ’
U D U C ’¤’0 U and dU D dU C ’¤’0 dU . We assume any changes are slow
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 0 0 0 0 0
dU D .T ’ dS ’ p ’ dV ’ C ’ dn’ /
X
C .T ’ dS ’ p ’ dV ’ C ’ dn’ / (8.1.1)
’¤’0
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 (constant internal energy) (8.1.2)

’0
X
dV C dV ’ D 0 (no expansion work) (8.1.3)
’¤’0
0
X
dn’ C dn’ D 0 (closed system) (8.1.4)
’¤’0
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 and dn’ from these re-
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
0 0 0
X X
0 D T ’ dS C .T ’ T ’ / dS ’ .p ’ p ’ / dV ’
’¤’0 ’¤’0
0
X
C .’ ’ / dn’ (8.1.5)
’¤’0
Solving for dS, we obtain

X T ’0 T ’ X p ’0 p ’
dS D ’0
dS ’ ’0
dV ’
0
T 0
T
’¤’ ’¤’
X ’0 ’
C ’0
dn’ (8.1.6)
0
T
’¤’
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 differ-
ent state. Once the isolated system has reached an equilibrium state, an imagined finite
change of any of the independent variables consistent with the constraints (a so-called vir-
tual 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 infinitesimal change of any of the
independent variables of the isolated system.
This requirement is satisfied in the case of the multiphase system only if the coefficient
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 parti-
tions), 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 flow 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.
0
Figure 8.1 Closed system of constant-volume slab-shaped fluid phases stacked in the
vertical direction. The shaded phase is reference phase ’0 .
8.1.4 Tall column of gas in a gravitational field

The earth’s gravitational field is an example of an external force field that acts on a system
placed in it. Usually we ignore its effects on the state of the system. If, however, the
system’s vertical extent is considerable we must take the presence of the field 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 infinite number of uniform phases, each of infinitesimal 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 confining the gas in a rigid adiabatic
container. In order to be able to associate each of the thin slab-shaped phases with a definite
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
figure. 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
X T ’0 T ’ X ’0 ’
’
dS D ’0
dS C ’0
dn’ (8.1.7)
0
T 0
T
’¤’ ’¤’
In the equilibrium state of the isolated system, dS is equal to zero for an infinitesimal change
of any of the independent variables. In this state, therefore, the coefficient 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 earth’s surface
for a reference elevation at which h is zero, and define the standard chemical potential
ı (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:
f .0/
.0/ D ı (g) C RT ln (8.1.8)
pı
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
field 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 81). 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 nM gh, 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:
f .h/
.h/ D ı (g) C RT ln C Mgh (8.1.10)
pı
(pure gas in
gravitational field)
Compare this relation with the equation that defines the fugacity when the effect of a grav-
itational field is negligible: D ı (g) C RT ln.f =p ı / (Eq. 7.8.7 on page 183). 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 “to-
tal chemical potential” or “gravitochemical potential” and reserve the term “chemical
potential” for the function ı (g)CRT ln.f =p ı /. With these definitions, in an equilib-
rium state the “total chemical potential” is the same at all elevations and the “chemical
potential” decreases with increasing elevation.
This book instead defines the chemical potential of a pure substance at any ele-
vation as the molar Gibbs energy at that elevation, as recommended in a 2001 IUPAC
technical report.1 When the chemical potential is defined 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/
ı (g) C RT ln ı C M gh D ı (g) C RT ln ı (8.1.11)
p p
Solving for f .h/ gives
Mgh=RT
f .h/ D f .0/e (8.1.12)
(pure gas at equilibrium
in gravitational field)
1 Ref. [2].
If we treat the gas as ideal, so that the fugacity equals the pressure, this equation becomes
Mgh=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 equi-
librium 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 field. There is, however, a shorter
derivation based on Newton’s second law and not involving the chemical potential. Con-
sider 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 g•h.
If the phase is at rest, the net vertical force is zero: pAs .p C •p/As As g•h D 0, or
•p D g•h. In the limit as the number of phases becomes infinite and •h and •p become
infinitesimal, this becomes
dp D g dh (8.1.14)
(fluid at equilibrium
Equation 8.1.14 is a general relation between changes in elevation and hydrostatic pressure
in any fluid. 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 liquid–gas interface which acts to minimize the ratio of
surface to volume. The interface acts somewhat like the stretched membrane of an inflated
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 infinitesimal change dU of the internal energy is the sum of contributions from the liq-
uid and gas phases and from the surface work dAs , where is the surface tension of the
liquid–gas 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 p l dV l C l dnl
C T g dS g p g dV g C g dng C dAs (8.1.15)
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 significant if r is small,
and decreases as r increases. The limit r!1 represents the flat 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 280).
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 sys-
tems 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 field, 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 sys-
tem are independent? To find 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

significance whether a phase, with particular values of its intensive properties, is con-
tiguous. 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 significant 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 sys-
tem 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 addi-
tional 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 199
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 P (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 free-
dom 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 pressure–volume and
pressure–temperature phase diagrams for a single substance, and Chap. 13 will describe
numerous types of phase diagrams for multicomponent systems.
8.2.1 Features of phase diagrams

Two-dimensional phase diagrams for a single-substance system can be generated as projec-
tions 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 combi-
nation 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 202. In these figures, some areas of the surface
are labeled with a single physical state: solid, liquid, gas, or supercritical fluid. 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 202).
The two-dimensional projections shown in Figs. 8.2(b) and 8.2(c) are pressure–volume
and pressure–temperature 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 pressure–temperature 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 equi-
librium 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 pressure–volume 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 find 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 208.
2 This statement assumes there are no constraints such as internal adiabatic partitions.
scf
10
0
bc
s l g
p=bar
l+g
0
trip g
0
le l
ine
s+
.V
g
=
n
4
/=
0
0
cm
3
m
ol
1
=
K
T
(a)
3
2
5
0
0
0
0
100
scf scf
80 bc bc
60
p=bar
. s l
40 g g
20 l+g
0
0 500 1000 1500 2000 2500 3000 3500 200 250 300 350 400
.V =n/=cm3 mol 1 T =K
(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 fluid). 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) Pressure–volume phase diagram (projection of the surface onto the p–.V =n/
plane).
(c) Pressure–temperature phase diagram (projection of the surface onto the p–T
plane).
a Based on data in Refs. [124] and [3].
10
0
bc
l scf
p=bar
s
g
s+l
l+g g
0
trip
le l
0
ine
.V
s+
=
4
n
0
g
0
/=
cm
3
K
m
=
ol
T
1
3
2
0
0
0
0
Figure 8.3 Three-dimensional p–.V =n/–T surface for CO2 , magnified 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 pressure–volume 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 pressure–temperature phase diagram.
The molar volumes in the three coexisting phases are given by the values of V =n at the
three points on the pressure–volume 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 pressure–temperature 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 liquid–gas 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 solid–liquid–gas 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 superfluidity. It is only at high pressures (10 bar or greater) that solid
helium can exist.
30
25 VIII VII
20
p=kbar
15
VI
10
5 V liquid
II III
I
0
200 250 300 350 400
T =K
Figure 8.4 High-pressure pressure–temperature 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 [130].
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 fixed 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.
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 fixed 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.
1:2
1:0 bc
standard standard bc
melting point boiling point
0:8
s l g
p=bar
. 0:6
0:4
0:2
triple point
0 bc
250 300 350 400

T =K
Figure 8.6 Pressure–temperature phase diagram of H2 O. (Based on data in Ref.

[124].)
The relation between temperature and pressure in a system with two phases in equilib-
rium is shown by the coexistence curve separating the two one-phase areas on the pressure–
temperature diagram (see Fig. 8.6). Consider the liquid–gas 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 ı C) is the boiling point at 1 atm; this temperature is also known
as the steam point. The standard boiling point of water (99:61 ı 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
fluid phase can exist at any volume and pressure (Sec. 2.2.3). The critical point is the point
on a phase diagram corresponding to liquid–gas 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 fluid 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 fluctuations. 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 fluid phase is called a supercritical fluid. Thus, a supercritical fluid of a
pure substance is a fluid that does not undergo a phase transition to a different fluid phase
when we change the pressure at constant temperature or change the temperature at constant
pressure.5
A fluid 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
fluids 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 liquid–gas 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 207. The
observation that the mean density closely approximates a linear function of temperature, as
shown in the figure, 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 fluid 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. [166] and [10].
PHOTOS BY RAY RAKOW

(a) (b) (c) (d)
Figure 8.7 Glass bulb filled 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, approx-
imately equal to, and greater than the critical density. a
(a) Supercritical fluid 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. [152].
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 fixed 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 liquid–gas 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 pressure–volume phase diagram of Fig. 8.9 on page 208, 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
1000 1050 ut
1000
ut
ut ut
ut
ut
950 ut
ut
800 ut
ut
ut
ut
900 ut
ut
850
ut
ut
ut
600 800
= kg m 3
= kg m 3
ut
750
bc ut
bc bc bc bc bc bc bc bc bc bc bc bc bc
bc bc bc
rs
rs
700
rs
400 rs
rs
650 rs
rs
600
rs
rs rs
200 rs rs
rs
rs
550 rs
rs
rs
rs
500 rs
rs
0 450
rs
270 280 290 300 310 315 316 317 318 319
T =K T =K
(a) (b)
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 dia-
mond is at the critical temperature and critical density.
a Based on data in Ref. [116].
b Data of Ref. [127], 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
is the molar volume of the liquid, Vml .
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
V D V l C V g D nl Vml C ng Vm (8.2.1)
n D nl C n g (8.2.2)
The value of V =n at the system point is then given by the equation
g
V nl Vml C ng Vm
D (8.2.3)
n nl C ng
l l+g g
Ll Lg
p b b b
A S B
V =n
Figure 8.9 Tie line (dashed) at constant T and p in the liquid–gas 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

V V g
l
n Vml Dn g
Vm (8.2.4)
n n
g
The quantities Vml V =n and V =n Vm are the lengths Ll and Lg , respectively, defined
in the figure and measured in units of V =n. This gives us the lever rule for liquid–gas
equilibrium:7
ng Ll
nl Ll D ng Lg or D (8.2.5)
nl Lg
(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 a
F D (8.2.6)
b
where a and b are extensive state functions. (In the pressure–volume phase diagram of
Fig. 8.9, these functions are a D V and b D n and the system point position is given
7 Therelation 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.
300
10
00
90 C
Æ
80
0
70
ÆC
C
0
Æ
0
60
250
ÆC
0
50
ÆC
0
ÆC
bc
40
0Æ
200 C
350 Æ C
p=bar
. 150
100
300 Æ C
50
250 Æ C
200 Æ C
0
0 100 200 300 400 500
.V =n/=cm3 mol 1
Figure 8.10 Isotherms for the fluid phases of H2 O. a The open circle indicates the
critical point, the dashed curve is the critical isotherm at 373:95 ı C, and the dotted
curve encloses the two-phase area of the pressure–volume phase diagram. The triple
line lies too close to the bottom of the diagram to be visible on this scale.
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
b“ L’
b ’ L’ D b “ L“ or D (8.2.8)
b’ L“
8.2.5 Volume properties

Figure 8.10 is a pressure–volume phase diagram for H2 O. On the diagram are drawn
isotherms (curves of constant T ). These isotherms define 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 defined by the dotted
curve drawn through the ends of the horizontal segments. The one-phase liquid area lies to
400
390 d c b a
380
ÆCT=
bc
370
360
350
0 50 100 150 200
.V =n/=cm3 mol 1
Figure 8.11 Isobars for the fluid 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 temperature–volume phase diagram.
Solid curves: a, p D 200 bar; b, p D 210 bar; c, p D 230 bar; d, p D 240 bar.
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 tem-
perature 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
1
T D (8.2.9)
Vm slope of isotherm
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 fluid regions, and approach zero at
the critical point. Accordingly, the isothermal compressibility of the gas and the supercrit-
ical fluid is much greater than that of the liquid, approaching infinity at the critical point.
The critical opalescence seen in Fig. 8.7 is caused by local density fluctuations, 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 coefficient ˛ is equal to .1=Vm /.@Vm =@T /p , so we have
1
˛D (8.2.10)
Vm slope of isobar
8.3 P HASE T RANSITIONS 211
The figure shows that the slopes of the isobars are large and positive in the liquid region,
smaller and negative in the gas and supercritical fluid regions, and approach zero at the
critical point. Thus the gas and the supercritical fluid have much larger cubic expansion
coefficients than the liquid. The value of ˛ approaches infinity 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 first 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 pro-
cesses 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 212
g
H D nHm nHml and the molar enthalpy of vaporization is
H g
vap H D D Hm Hml (8.3.1)
n
(pure substance)
In other words, vap H is the enthalpy change per amount vaporized and is also the differ-
ence between the molar enthalpies of the two phases.
A molar property of a phase, being intensive, usually depends on two independent in-
tensive variables such as T and p. Despite the fact that vap H is the difference of the two
g
molar properties Hm and Hml , 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
same chemical potential: “ ’ D 0. Therefore, the molar Gibbs energy of any equilib-
rium phase transition is zero:
trs G D 0 (8.3.3)
(pure substance)
Since the Gibbs energy is defined by G D H T S, in phase ’ we have Gm ’ D G ’ =n’ D

“ “ “
Hm’ ’
T Sm . Similarly, in phase “ we have Gm D Hm T Sm . When we substitute these
“ ’ (Eq. 8.3.2) and set T equal to the transition temperature
expressions in trs G D Gm Gm
Ttrs , we obtain
“ “
trs G D .Hm Hm’ / Ttrs .Sm Sm’ /
D trs H Ttrs trs S (8.3.4)
Then, by setting trs G equal to zero, we find the molar entropy and molar enthalpy of the
equilibrium phase transition are related by
trs H
trs S D (8.3.5)
Ttrs
(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 213
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 first 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 ı , is the enthalpy change when pure
liquid in its standard state at a specified 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 ı ), but
the final state (the gas behaving ideally at pressure p ı ) is hypothetical. The liquid and gas
are not necessarily in equilibrium with one another at pressure p ı and the temperature of
interest, and we cannot evaluate vap H ı from a calorimetric measurement with electrical
work without further corrections. The same difficulty applies to the evaluation of sub H ı .
In contrast, vap H and sub H (without the ı symbol), as well as fus H ı , 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 ı D Xmı (g) Xm (l) and sub X ı D Xmı (g)
Xm (s) are functions only of T . To evaluate vap X ı or sub X ı 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 ı 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 ı . 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 ı or sub X ı .
8.4 COEXISTENCE CURVES
A coexistence curve on a pressure–temperature phase diagram shows the conditions under

which two phases can coexist in equilibrium, as explained in Sec. 8.2.2.
10
gas
1
=kJ mol
liquid
300
0:2
0
K
T=
p =b
ar
1 400
p
T
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 p–T
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
5 0:04
sol C
id bc
4 A bc bc
bc 0:03 A
1 B
3 liq
=kJ mol
uid
p=bar
gas,
. 0:02 solid liquid gas
bc
B
0:00
2
gas,
bar
0:01
1
0
bc
:03
triple point
C bc
bar
0 0
250 260 270 280 290 300 310 250 260 270 280 290 300 310
T =K T =K
(a) (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 negligi-
ble. 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) Pressure–temperature 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 low-
ers 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 solid–gas equilibrium is possible for H2 O.
The liquid phase is not stable at any pressure below the triple-point pressure, as shown by
the pressure–temperature 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 tem-
perature of both phases equally without changing the pressure, the phases will no longer be
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 infinitesimal 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
“
dp Sm Sm’
D “ (8.4.3)
dT Vm Vm’
or
dp S
D trs (8.4.4)
dT trs V
(pure substance)
Equation 8.4.4 is one form of the Clapeyron equation, which contains no approximations.
We find an alternative form by substituting trs S D trs H=Ttrs (Eq. 8.3.5):
dp H
D trs (8.4.5)
dT Ttrs V
(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 solid–gas 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 solid–liquid coexistence curve may be either positive or negative. The absolute
value of fus V is small, causing the solid–liquid coexistence curve to be relatively steep;
see Fig. 8.13(b) for an example.
Most substances expand on melting, making the slope of the solid–liquid coexistence
curve positive. This is true of carbon dioxide, although in Fig. 8.2(c) on page 201 the
curve is so steep that it is difficult 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 203 clearly shows that the coex-
istence 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 liq-
uid, causing the slopes of the other solid–liquid coexistence curves to be positive. The
ice VII–ice 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
BIOGRAPHICAL SKETCH
Benoit Paul Émile Clapeyron (1799–1864)
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’s paper used indicator diagrams
and calculus for a rigorous proof of Carnot’s
conclusion that the efficiency 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.
Clapeyron was a French civil and railroad en- It was not until the appearance of English and
gineer who made important contributions to German translations of this paper that Clapey-
thermodynamic theory. He was born in Paris, ron’s analysis enabled Kelvin to define a ther-
the son of a merchant. modynamic temperature scale and Clausius to
He graduated from the École Polytechnique introduce entropy and write the mathematical
in 1818, four years after Sadi Carnot’s gradu- statement of the second law.
ation from the same school, and then trained Clapeyron’s 1834 paper also derived an ex-
as an engineer at the École de Mines. At this pression for the temperature dependence of the
time, the Russian czar asked the French gov- vapor pressure of a liquid equivalent to what is
ernment for engineers to help create a program now called the Clapeyron equation (Eq. 8.4.5).
of civil and military engineering. Clapeyron The paper used a reversible cycle with vapor-
and his classmate Gabriel Lamé were offered ization at one temperature followed by conden-
this assignment. They lived in Russia for ten sation at an infinitesimally-lower temperature
years, teaching pure and applied mathematics and pressure, and equated the efficiency of this
in Saint Petersburg and jointly publishing en- cycle to that of a gas between the same two
gineering and mathematical papers. In 1831 temperatures. Although the thermodynamic
they returned to France; their liberal political temperature T does not appear as such, it is
views and the oppression of foreigners by the represented by a temperature-dependent func-
new czar, Nicholas I, made it impossible for tion whose relation to the Celsius scale had to
them to remain in Russia. be determined experimentally.b
Back in France, Clapeyron became involved Beginning in 1844, Clapeyron taught the
in the construction of the first French passen- course on steam engines at the École Nationale
ger railroads and in the design of steam loco- des Ponts et Chaussées near Paris, the old-
motives and metal bridges. He was married est French engineering school. In this course,
with a daughter. surprisingly, he seldom mentioned his theory
In a paper published in 1834 in the journal of heat engines based on Carnot’s work.c He
of the École Polytechnique, Clapeyron brought eventually embraced the equivalence of heat
attention to the work of Sadi Carnot (page and work established by Joule’s experiments.d
106), who had died two years before:a At the time of Clapeyron’s death, the rail-
Among studies which have appeared on the the- road entrepreneur Émile Péreire wrote:e
ory of heat I will mention finally a work by S. We were together since 1832. I’ve never done
Carnot, published in 1824, with the title Reflec- important business without consulting him, I’ve
tions on the Motive Power of Fire. The idea never found a judgment more reliable and more
which serves as a basis of his researches seems honest. His modesty was so great and his char-
to me to be both fertile and beyond question; acter so good that I never knew him to have an
his demonstrations are founded on the absur- enemy.
a Ref. [30]. b Ref. [173]. c Ref. [90]. d Ref. [80]. e Ref. [80].
Consider the transition from solid to liquid (fusion). Because of the fact that the cubic
expansion coefficient 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
fus H T2
p2 p1 ln (8.4.7)
fus V T1
or

fus V .p2 p1 /
T2 T1 exp (8.4.8)
fus H
(pure substance)
from which we may estimate the dependence of the melting point on pressure.
8.4.3 The Clausius–Clapeyron 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 g g g
vap V D Vm Vml Vm sub V D Vm Vms Vm (8.4.9)
g
The further approximation that the gas behaves as an ideal gas, Vm RT =p, then changes
Eq. 8.4.5 to
dp ptrs H
(8.4.10)
dT RT 2
(pure substance,
vaporization or sublimation)
Equation 8.4.10 is the Clausius–Clapeyron equation. It gives an approximate expres-

sion for the slope of a liquid–gas or solid–gas 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 fol-
lowed by vaporization, the net change is sublimation. Therefore, the molar transition en-
thalpies 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 liquid–gas
coexistence curve.
We divide both sides of Eq. 8.4.10 by p ı and rearrange to the form
d.p=p ı / H dT
trs 2 (8.4.12)
p=p ı R T
Then, using the mathematical identities d.p=p ı /=.p=p ı / D d ln.p=p ı / and dT =T 2 D

d.1=T /, we can write Eq. 8.4.12 in three alternative forms:
d ln.p=p ı / H
trs 2 (8.4.13)
dT RT
(pure substance,
trs H
d ln.p=p ı / d.1=T / (8.4.14)
R
(pure substance,
d ln.p=p ı / trs H
(8.4.15)
d.1=T / R
(pure substance,
Equation 8.4.15 shows that the curve of a plot of ln.p=p ı / 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 ı appearing in
these equations becomes p=bar. That is, p=p ı 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 ı 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 final 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 220
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, find 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 Smı (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 coefficient 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 Smı (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 Hmı (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 ı 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: ı C.
8.5 Naphthalene has a melting point of 78:2 ı C at 1 bar and 81:7 ı 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 significant figures.
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
b
ln.p=bar/ D a
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 Clausius–Clapeyron 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 ı 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 221
(a) Use the Clapeyron equation to calculate the slope dp= dT of the liquid–gas coexistence
curve at this point.
(b) Repeat the calculation using the Clausius–Clapeyron equation.
(c) Use your results to estimate the standard boiling point of water. (Note: The experimental
value is 99:61 ı 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 ı at this temperature with the help of Table 7.5 and the fol-
lowing data for the second virial coefficient of gaseous H2 O at 372:76 K:
4
BD 4:60 10 m3 mol 1
dB= dT D 3:4 10 6
m3 K 1
mol 1
(c) Would you expect the values of fus H and fus H ı to be equal at the standard freezing
point of water? Explain.
8.9 The standard boiling point of H2 O is 99:61 ı 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
1 1 2 1
a D 75:94 J K mol b D 0:022 J K mol c D 99:61 ı 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 ı C above the standard boiling point, re-
sulting 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 dis-
cusses composition variables and partial molar quantities of mixtures in which no chemical
reaction is occurring. The ideal mixture is defined. Chemical potentials, activity coeffi-
cients, and activities of individual substances in both ideal and nonideal mixtures are dis-
cussed.
Except for the use of fugacities to determine activity coefficients 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 definite 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 defined by
def n def n
xi D P i or yi D P i (9.1.1)
j nj j nj (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.
222
CHAPTER 9 MIXTURES
9.1 C OMPOSITION VARIABLES 223
The mass fraction, or weight fraction, of species i is defined by
def m.i / nM
wi D DP i i (9.1.2)
m j nj Mj (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 defined by
def ni
ci D (9.1.3)
V
(P D1)
1 3.
The symbol M is often used to stand for units of mol L , or mol dm Thus, a concentra-
tion of 0:5 M is 0:5 moles per liter, or 0:5 molar.
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 defined by
def nB
mB D (9.1.4)
m.A/
(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 224
9.1.4 Binary solutions

We may write simplified equations for a binary solution of two substances, solvent A and
solute B. Equations 9.1.1–9.1.4 become
nB
xB D (9.1.5)
nA C nB
(binary solution)
nB MB
wB D (9.1.6)
nA MA C nB MB
(binary solution)
nB nB
cB D D (9.1.7)
V nA MA C nB MB
(binary solution)
nB
mB D (9.1.8)
nA MA
(binary solution)
The right sides of Eqs. 9.1.5–9.1.8 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 :
nB xB
from Eq. 9.1.5 D (9.1.9)
nA 1 xB
(binary solution)
nB MA wB
from Eq. 9.1.6 D (9.1.10)
nA MB .1 wB /
(binary solution)
nB MA cB
from Eq. 9.1.7 D (9.1.11)
nA MB cB
(binary solution)
nB
from Eq. 9.1.8 D MA mB (9.1.12)
nA
(binary solution)
These expressions for nB =nA allow us to find one composition variable as a function of
another. For example, to find molality as a function of concentration, we equate the expres-
sions for nB =nA on the right sides of Eqs. 9.1.11 and 9.1.12 and solve for mB to obtain
cB
mB D (9.1.13)
MB cB
CHAPTER 9 MIXTURES
9.2 PARTIAL M OLAR Q UANTITIES 225
A binary solution becomes more dilute as any of the solute composition variables be-
comes smaller. In the limit of infinite 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 mB (9.1.14)
(binary solution at
infinite dilution)
where a superscript asterisk ( ) denotes a pure phase. We see that, in the limit of infinite
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 ex-
pressed 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 defined 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 fixed 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 fixed composition has fixed 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 sub-
script i identifies a constituent species of the mixture, denotes the partial molar quantity
CHAPTER 9 MIXTURES
of species i defined by

def @X
Xi D (9.2.1)
@ni 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 defined 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 first 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 significance of a partial molar quantity as defined by Eq.
9.2.1, let us first 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 ı 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 ı C with
100:0 cm3 of methanol at 25 ı C, we find the volume of the resulting mixture at 25 ı 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:
VA 100:0 cm3

nA D D D 5:53 mol (9.2.2)
Vm;A 18:07 cm3 mol 1
VB 100:0 cm3

nB D D D 2:45 mol (9.2.3)
Vm;B 40:75 cm3 mol 1
nB 2:45 mol
xB D D D 0:307 (9.2.4)
nA C nB 5:53 mol C 2:45 mol
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. [119], p. 44.
CHAPTER 9 MIXTURES
(a) (b)
Figure 9.1 Addition of pure methanol (substance B) to a water–methanol 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 ı C was 10 , 000.0 cm3 , we find 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 infinitesimal, 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 infinitesimal amount is mixed with a
finite volume of mixture.
The partial molar volume of B is an intensive property that is a function of the compo-
sition 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
by solvent molecules. We denote this limiting value of VB by VB1 , the partial molar volume
of solute B at infinite 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
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 water–ion 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 definite 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
C dn C dn (9.2.5)
@nA T;p;nB A @nB T;p;nA B
(binary mixture)
We know the first two partial derivatives on the right side are given by4

@V @V
D ˛V D T V (9.2.6)
@T p;nA ;nB @p T;nA ;nB
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)
3C 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 fixed
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
A mixture of B
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 vol-
umes 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, thermostat-
ted 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 con-
stituents and the appropriate partial molar volumes for the particular temperature, pressure,
and composition.
For example, given that the partial molar volumes in a water–methanol mixture of com-
position 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 water–methanol mixture described at the beginning of Sec. 9.2.1 as fol-
lows:
V D .17:74 cm3 mol 1 /.5:53 mol/ C .38:76 cm3 mol 1 /.2:45 mol/
D 193:1 cm3 (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 Euler’s theorem on homogeneous functions to V treated
as a function of nA and nB .
CHAPTER 9 MIXTURES
Equation 9.2.12 is the Gibbs–Duhem 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
nB
dVA D dV (9.2.14)
nA B
(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 figure, 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 n B
.V =n/ D D VA xA C VB xB
n
D VA .1 xB / C VB xB D .VB VA /xB C VA (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/ dŒ.VB VA /xB C VA
D
dxB dxB

dVB dVA dV
D VB VA C xB C A
dx dxB dx
B B
dVA dVB
D VB VA C .1 xB / C xB
dxB dxB

dVA dVB
D VB VA C xA C xB (9.2.16)
dxB dxB
The differentials dVA and dVB are related to one another by the Gibbs–Duhem 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
45 0:0 41
40
40
0:5 39 bc VB
mol 1
35 38
3 mol 1
bc
mol 1
1:0 20
37
Vm .mix/=cm3
30
.
.V =n/=cm
19
36
Vi =cm3
0 0:2 0:4 0:6 0:8 1:0
25 bc
1:5 18 bc
20 17
VA
2:0 16
15
15
10 2:5 14
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
xB xB xB
(a) (b) (c)
Figure 9.3 Mixtures of water (A) and methanol (B) at 25 ı 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).
obtain
dVA dVB
xA C xB D 0 (9.2.17)
dxB dxB
and substitute in Eq. 9.2.16 to obtain
d.V =n/
D VB VA (9.2.18)
dxB
Let the partial molar volumes of the constituents of a binary mixture of arbitrary
composition xB0 be VA0 and VB0 . Equation 9.2.15 shows that the value of V =n at the
point on the curve of V =n versus xB where the composition is xB0 is .VB0 VA0 /xB0 C VA0 .
Equation 9.2.18 shows that the tangent to the curve at this point has a slope of VB0 VA0 .
The equation of the line that passes through this point and has this slope, and thus is
the tangent to the curve at this point, is y D .VB0 VA0 /xB C VA0 , where y is the vertical
ordinate on the plot of .V =n/ versus xB . The line has intercepts yDVA0 at xB D0 and
yDVB0 at xB D1.
A variant of the method of intercepts is to plot the molar integral volume of mixing
given by

V (mix) V nA Vm;A nB Vm;B
Vm (mix) D D (9.2.19)
n n
CHAPTER 9 MIXTURES
versus xB , as illustrated in Fig. 9.3(b). V (mix) is the integral volume of mixing—the

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/ xA Vm;A xB Vm;B

D .V =n/ .1 xB /Vm;A xB Vm;B (9.2.20)
We make the substitution .V =n/ D .VB VA /xB C VA from Eq. 9.2.15 and rearrange:

Vm (mix) D VB Vm;B VA Vm;A xB C VA Vm;A (9.2.21)
Differentiation with respect to xB yields

dVm (mix) dVB dVA dV

xB C A

D VB Vm;B VA Vm;A C
dxB dxB dxB dx
B

dVA dVB
D VB Vm;B VA Vm;A C .1 xB / C xB
dxB dx
B

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

dVm (mix)

D VB Vm;B VA Vm;A (9.2.23)
dxB
Equations 9.2.21 and 9.2.23 are analogous to Eqs. 9.2.15 and 9.2.18, with V =n re-

placed 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 find 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 water–methanol 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
Gibbs–Duhem 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 mo-

lecular 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 V for the
m;B
molar volumes of pure A and B. The partial molar volume of a pure substance is the same as
CHAPTER 9 MIXTURES
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, defined by

def @X
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 Gibbs–Duhem equation applied to X can be written in the equivalent forms
X
ni dXi D 0 (9.2.26)
i (constant T and p)
and
X
xi dXi D 0 (9.2.27)
These relations can be applied to a mixture in which each species i is a nonelectrolyte sub-
stance, 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 quanti-
ties 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 mea-
surements 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 226, 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
ion. Consider an aqueous solution of a fully-dissociated electrolyte solute with the

formula MC 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
this ion at infinite dilution is arbitrarily set equal to zero: VH1C D 0.
For example, given the value (at 298:15 K and 1 bar) of the partial molar volume
at infinite dilution of aqueous hydrogen chloride
1
VHCl D 17:82 cm3 mol 1
(9.2.29)
we can find the so-called “conventional” partial molar volume of Cl ion:

1
VCl1 D VHCl VH1C D 17:82 cm3 mol 1
(9.2.30)
1
Going one step further, the measured value VNaCl D 16:61 cm3 mol 1
gives, for NaC
ion, the conventional value
1 1
VNaC D VNaCl VCl1 D .16:61 17:82/ cm3 mol 1
D 1:21 cm3 mol 1
(9.2.31)
9.2.5 Partial specific quantities

A partial specific 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 specific
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 specific quantities and partial specific quantities,
in some applications they have an advantage over molar quantities and partial molar quanti-
ties because they can be evaluated without knowledge of the molar mass. For instance, the
value of a solute’s partial specific 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 specific quantities by replacing amounts by masses, mole
fractions by mass fractions, and partial molar quantities by partial specific
P quantities. Using
volume as an example,
P we can write anPadditivity relation V D i m.i /vi , and Gibbs–
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 specific 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
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, defined by

def @G
i D (9.2.33)
@ni T;p;nj ¤i
(mixture)
If there are work coordinates for nonexpansion work, the partial derivative is taken at con-
stant values of these coordinates.
The chemical potential of a species in a phase plays a crucial role in equilibrium prob-
lems, because it is a measure of the escaping tendency of the species from the phase. Al-
though 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
defined 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 141, often called the Gibbs fundamental equation:
s
X
dG D S dT C V dp C i dni (9.2.34)
i D1 (mixture)
Consider the special case of a mixture containing charged species, for example an aque-
ous 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 inde-
pendently vary the amounts of KC and Cl in the mixture. The chemical potential of the
KC ion is defined 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 C dnC C 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 con-
ditions 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
of species and energy between the phases, and that effects of gravity and other external
force fields 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
0
X
dU D dU ’ C dU ’
’¤’0
’0 ’0 0 0 0 0
X
DT dS p ’ dV ’ C ’i dn’i
i
!
X X
’ ’ ’ ’
C T dS p dV C ’i dn’i (9.2.37)
’¤’0 i
The conditions of isolation are

’0
X
’
dV C dV D0 (no expansion work) (9.2.39)
’¤’0
For each species i:

0 X
dn’i C dn’i D 0 (closed system) (9.2.40)
’¤’0
0 0
We use these relations to substitute for dU , dV ’ , and dn’i in Eq. 9.2.37. After making the
0
further substitution dS ’ D dS ’
P
’¤’0 dS and solving for dS , we obtain
X T ’0 T ’ X p ’0 p ’
dS D ’0
dS ’ ’0
dV ’
0
T 0
T
’¤’ ’¤’
X X ’0 ’
i i
C ’0
dn’i (9.2.41)
0
T
i ’¤’
This equation is like Eq. 8.1.6 on page 194 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 infinitesimal change of any of the independent variables, and the
coefficient of each term on the right side of Eq. 9.2.41 is zero. We find 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 accommo-
dated 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 dn’i 0 equal to
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, multicompo-
nent system without internal partitions, the temperature and pressure are uniform throughout
CHAPTER 9 MIXTURES
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 Gibbs–Duhem equation applied to the chemical po-
tentials can be written in the equivalent forms
X
ni di D 0 (9.2.42)
and
X
xi di D 0 (9.2.43)
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 Gibbs–Duhem equation, without the restriction of con-
stant T and p, is X
S dT V dp C ni di D 0 (9.2.44)
i
This version is derived

P by comparing
P the expression for dG given by Eq. 9.2.34
P with the
differential dGD i i dni C i ni di obtained from the additivity rule GD i i ni .
The Gibbs energy is defined 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 @H @S
D T (9.2.45)
@ni T;p;nj ¤i @ni T;p;nj ¤i @ni T;p;nj ¤i
CHAPTER 9 MIXTURES
9.3 G AS M IXTURES 238
We recognize each partial derivative as a partial molar quantity and rewrite the equation as
i D Hi T Si (9.2.46)
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 i dni (Eq.
9.2.34), we obtain the following reciprocity relations:

@i @S @i @V
D D (9.2.47)
@T p;fnig @ni T;p;nj ¤i @p T;fnig @ni T;p;nj ¤i
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 constant—that 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
D Vi (9.2.49)
@p T;fnig
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

@H X
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

@Hi
Cp;i D (9.2.52)
@T 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
9.3.1 Partial pressure

The partial pressure pi of substance i in a gas mixture is defined 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
Psum of the partial pressures of all substances in a gas mixture is i pi D i yi p D
The
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)
i (gas mixture)
Thus, the sum of the partial pressures equals the pressure of the gas phase. This statement
is known as Dalton’s 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.
P It obeys the ideal gas equation p D nRT =V (where n in a
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
ni RT
pi D (9.3.3)
V
(ideal gas mixture)
Equation 9.3.3 is the ideal gas equation with the partial pressure of a constituent sub-
stance 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 ı , and behaves as an ideal
CHAPTER 9 MIXTURES
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 ıi (g) of gaseous i is the chemical potential of i in
this gas standard state, and is a function of temperature.
To derive an expression for i in an ideal gas mixture relative to ıi (g), we make an
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 ı (g) C RT ln.p=p ı /, 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
pA
A D ıA (g) C RT ln (9.3.4)
pı
In general, for each substance i in an ideal gas mixture, we have the relation
pi
i D ıi (g) C RT ln (9.3.5)
pı
(ideal gas mixture)
where ıi (g) is the chemical potential of i in the gas standard state at the same temperature
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 ı , so that ln.pi =p ı / is zero, the chemical potential is equal to the standard
6 We assume the gas is not adsorbed to a significant extent on the surface of the membrane or in its pores.
CHAPTER 9 MIXTURES
chemical potential. Conceptually, a standard state should be a well-defined state of the

system, which in the case of a gas is the pure ideal gas at pDp ı . 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 definition 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 definition an ideal gas mixture. The nonrigorous nature of the assump-
tion used to obtain Eq. 9.3.5 presents no difficulty if we consider the equation to be the basic
definition.
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:
ı
@i (g) p
Si D R ln ıi
@T p;fnig p
p
D Siı R ln ıi (9.3.6)
p
(ideal gas mixture)
The quantity Siı D Œ@ıi (g)=@T p;fnig is the standard molar entropy of constituent i. It
is the molar entropy of i in its standard state of pure ideal gas at pressure p ı .
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 ıi (g) C T Siı , which is equivalent
to
Hi D Hiı (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
ı
@i (g) @ ln.pi =p ı /

Vi D C RT (9.3.8)
@p T;fnig @p T;fnig
The first partial derivative on the right is zero because ıi (g) is a function only of T . For
the second partial derivative, we write pi =p ı D yi p=p ı . The mole fraction yi is constant
when the amount of each substance is constant, so we have Œ@ ln.yi p=p ı /=@pT;fnig D 1=p.
The partial molar volume is therefore given by
RT
Vi D (9.3.9)
p
(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 Viı D RT =p ı for
an ideal gas.
CHAPTER 9 MIXTURES
From Eqs. 9.2.50, 9.2.52, 9.3.7, and 9.3.9 we obtain the relations
Ui D Uiı (9.3.10)
(ideal gas mixture)
and
ı
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 definition 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 derivationP shows.
From
P Eq. 9.3.9 and the additivity rule, we find the volume is given by V D i ni Vi D
n RT
i i P =p D nRT =p, which is the ideal gas equation. From Eq. 9.3.10 we have
ı
P
U D n U
i i i D n U
i i i , 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 define an ideal gas according to Sec. 3.5.1.
9.3.4 Real gas mixtures

Fugacity
The fugacity f of a pure gas is defined by D ı (g) C RT ln.f =p ı / (Eq. 7.8.7 on
page 183). By analogy with this equation, the fugacity fi of substance i in a real gas
mixture is defined by the relation
i ıi (g)

f def
i D ıi (g) C RT ln iı or fi D p ı exp (9.3.12)
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 pressure–volume 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
fi0
0i 00i D RT ln (9.3.13)
fi00
Integration of di D Vi dp (from Eq. 9.2.49) between these pressures yields

Z p0
0i 00i D Vi dp (9.3.14)
p 00
CHAPTER 9 MIXTURES
When we equate these two expressions for 0i 00i , divide both sides by RT , subtract the
identity
Z p0
p0 dp
ln 00 D (9.3.15)
p p 00 p
and take the ideal-gas behavior limits p 00 !0 and fi00 !yi p 00 D .pi0 =p 0 /p 00 , we obtain
f0 p0
Vi 1
Z
ln i0 D dp (9.3.16)
pi 0 RT p
(gas mixture, constant T )
The fugacity coefficient i of constituent i is defined by

def
fi D i pi (9.3.17)
(gas mixture)
Accordingly, the fugacity coefficient at pressure p 0 is given by

Z p0
0 Vi 1
ln i .p / D dp (9.3.18)
0 RT p
(gas mixture, constant T )
As p 0 approaches zero, the integral in Eqs. 9.3.16 and 9.3.18 approaches zero, fi0 ap-
proaches pi0 , and i .p 0 / approaches unity.
Partial molar quantities

By combining Eqs. 9.3.12 and 9.3.16, we obtain
p0 p 0
RT
Z
0
i .p / D ıi (g) C RT ln ıi C Vi dp (9.3.19)
p 0 p
(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 (9.3.20)
.V =n/ .V =n/2
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 coefficients B; C; : : : depend on
composition as well as temperature.
CHAPTER 9 MIXTURES
Table 9.1 Gas mixture: expressions for differences between partial molar and stan-
dard molar quantities of constituent i
General expression Equation of statea

Difference at pressure p 0 V D nRT =p C nB
Z p0
p0

RT pi
i ıi (g) RT ln ıi C Vi dp RT ln C Bi0 p
p 0 " p pı
Z p0 #
pi0 dB 0

@Vi R p
Si Siı (g) R ln ı dp R ln ıi p i
p 0 @T p p p dT
Z p0 " #
dBi0

@Vi
Hi Hiı (g) Vi T dp p Bi0 T
0 @T p dT
Z p0 " #
dBi0

@Vi
Ui Uiı (g) Vi T dp C RT p 0 Vi pT
0 @T p dT
Z p0 2
ı @ Vi d2 Bi0
Cp;i Cp;i (g) T dp pT
0 @T 2 p dT 2
aB and Bi0 are defined by Eqs. 9.3.24 and 9.3.26
At low to moderate pressures, the simple equation of state

RT
V =n D CB (9.3.21)
p
describes a gas mixture to a sufficiently high degree of accuracy (see Eq. 2.2.8 on page 35).
This is equivalent to a compression factor given by
def pV Bp
Z D D1C (9.3.22)
nRT RT
From statistical mechanical theory, the dependence of the second virial coefficient B of
a binary gas mixture on the mole fraction composition is given by
B D yA2 BAA C 2yA yB BAB C yB2 BBB (9.3.23)

(binary gas mixture)
where BAA and BBB are the second virial coefficients of pure A and B, and BAB is a mixed
second virial coefficient. 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 j (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 coefficient
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
Vi D C Bi0 (9.3.25)
p
CHAPTER 9 MIXTURES
9.4 L IQUID AND S OLID M IXTURES OF N ONELECTROLYTES 245
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
BA0 D BAA C . BAA C 2BAB BBB /yB2 (9.3.27)

BB0 D BBB C . BAA C 2BAB BBB /yA2 (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 coefficient of constituent i and the function Bi0 :
Bi0 p
ln i D (9.3.29)
RT
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 coefficient 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 nonelectrolytes—substances
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 Raoult’s law

In 1888, the French physical chemist François Raoult published his finding 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 Raoult’s 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+C)(g)
pA ; fA

p ;f
A A
p D pA C BC
p pC p D p

A C pC
(A+B)(l) A(l)
xA
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 temper-
ature, while adjusting the partial pressure of C so as to keep p constant. If we find 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 Raoult’s law for partial pressure in this range. This is the same as the origi-
nal Raoult’s 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
Raoult’s law for partial pressure becomes identical to the original law.
A form of Raoult’s 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 Raoult’s law for fugacity.
We can generalize the two forms of Raoult’s law for any constituent i of a liquid mix-
ture:
pi D xi pi (9.4.2)
(Raoult’s law for partial pressure)
fi D xi fi (9.4.3)
(Raoult’s 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 Raoult’s law are equivalent when the gas phases are ideal gas mix-
tures. When it is necessary to make a distinction between the two forms, this book will refer
specifically to Raoult’s law for partial pressure or Raoult’s law for fugacity.
CHAPTER 9 MIXTURES
Raoult’s 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 ıi (g) C RT ln fi =p ı (Eq. 9.3.12). Replacing fi by xi fi
according to Raoult’s law, and rearranging, we obtain
fi

ı
i D i (g) C RT ln ı C RT ln xi (9.4.4)
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 D i C RT ln xi (9.4.5)
Equation 9.4.5 is an expression for the chemical potential in the liquid phase when Raoult’s
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 Raoult’s 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,
Raoult’s 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 defined as a liquid mixture in which, at a given temperature
and pressure, each constituent obeys Raoult’s 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 Raoult’s
law quite closely. In contrast, water and methanol, although miscible in all proportions,
form liquid mixtures that deviate considerably from Raoult’s law. The most commonly
encountered situation for mixtures of organic liquids is that each constituent deviates from
Raoult’s law behavior by having a higher fugacity than predicted by Eq. 9.4.3—a positive
deviation from Raoult’s 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 mix-
ture, 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:
p yp
i D ıi (g) C RT ln ıi D ıi (g) C RT ln i ı (9.4.6)
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
i D ıi (g) C RT ln ı (9.4.7)
p
CHAPTER 9 MIXTURES
By eliminating ıi (g) between these equations and rearranging, we obtain Eq. 9.4.5 with xi
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:
(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 find expressions for partial molar quantities. These expressions find
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 i 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.
CHAPTER 9 MIXTURES
gas mixture fi
liquid mixture xi
Figure 9.6 Equilibrated liquid and gas mixtures. Substance i is present in both
phases.
9.4.4 Henry’s 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 Henry’s law. The proportionality constant kH;i is the Henry’s
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 definition constituent
i obeys Raoult’s 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 non-
electrolyte, Henry’s law behavior occurs in the limit of infinite 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 251. The limiting slope of the
plot of fB versus xB is finite, not zero or infinite. (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 infinite dilution (Eq. 9.1.14), in a very dilute solution the fugacity
CHAPTER 9 MIXTURES
BIOGRAPHICAL SKETCH
William Henry (1774–1836)
the gas to dissolve in the water. His findings
were presented to the Royal Society of London
in 1802 and published the following year:a
The results of a series of at least fifty experi-
ments, on carbonic acid, sulphuretted hydrogen
gas, nitrous oxide, oxygenous and azotic gases,b
with the above apparatus, establish the follow-
ing general law: that, under equal circumstances
of temperature, water takes up, in all cases, the
same volume of condensed gas as of gas un-
der ordinary pressure. But, as the spaces oc-
cupied by every gas are inversely as the com-
William Henry was a British chemist, trained
pressing force, it follows, that water takes up, of
as a physician, who is best known for his for-
gas condensed by one, two, or more additional
mulation of what is now called Henry’s law. atmospheres, a quantity which, ordinarily com-
Henry was born in Manchester, England. pressed, would be equal to twice, thrice, &c. the
His father was an apothecary and industrial volume absorbed under the common pressure of
chemist who established a profitable business the atmosphere.
manufacturing such products as magnesium Henry later confirmed a suggestion made by
carbonate (used as an antacid) and carbonated his close friend John Dalton, that the amount
water. At the age of ten, Henry was severely of a constituent of a gaseous mixture that is ab-
injured by a falling beam and was plagued by sorbed is proportional to its partial pressure.c
pain and ill health for the rest of his life. Henry carried out other important work,
Henry began medical studies at the Univer- chiefly on gases, including the elemental com-
sity of Edinburgh in 1795. He interrupted these positions of hydrogen chloride, ammonia, and
studies to do chemical research, to assist his methane.
father in general medical practice, and to help Because of his poor health and unsuccess-
run the family chemical business. He finally ful surgery on his hands, Henry was unable to
received his diploma of Doctor in Medicine in continue working in the lab after 1824. Twelve
1807. In 1809, in recognition of his research years later, suffering from pain and depression,
papers, he was elected a Fellow of the Royal he committed suicide.
Society. In a biography published the year after
In 1801 the first edition of his influential Henry’s death, his son William Charles Henry
chemistry textbook appeared, originally called wrote:d
An Epitome of Chemistry and in later editions
In the general intercourse of society, Dr. Henry
Elements of Experimental Chemistry. The was distinguished by a polished courtesy, by an
book went through eleven editions over a pe- intuitive propriety, and by a considerate fore-
riod of 28 years. thought and respect for the feelings and opinions
Henry investigated the relation between the of others. . . His comprehensive range of thought
pressure of a gas and the volume of the gas, and knowledge, his proneness to general spec-
measured at that pressure, that was absorbed ulation in contradistinction to detail, his ready
into a given volume of water. He used a simple command of the refinements of language and the
apparatus in which the water and gas were con- liveliness of his feelings and imagination ren-
fined over mercury in a graduated glass vessel, dered him a most instructive and engaging com-
and the contents agitated to allow a portion of panion.
a Ref. [75]. b These gases are respectively CO2 , H2 S, N2 O, O2 , and N2 . c Ref. [76]. d Quoted in Ref.
[156].
CHAPTER 9 MIXTURES
0:4
reference state b
0:4
kH,B
0:3 0:3
.fB =xB /=bar

fB =bar
.
0:2 0:2
0:1 0:1
0 0
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
xB xB
(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 Henry’s 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 Henry’s law constant kH,B .
is proportional to all three of these composition variables. This leads to three versions of
Henry’s law:
mole fraction basis fB D kH,B xB (9.4.16)

(nonelectrolyte solute
at infinite dilution)
concentration basis fB D kc;B cB (9.4.17)
molality basis fB D km;B mB (9.4.18)
In these equations kH,B , kc;B , and km;B are Henry’s law constants defined by

def fB
mole fraction basis kH,B D lim (9.4.19)
xB !0 xB

def fB
concentration basis kc;B D lim (9.4.20)
cB !0 cB

def fB
molality basis km;B D lim (9.4.21)
mB !0 mB
CHAPTER 9 MIXTURES
Note that the Henry’s 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 infinite dilution. The
evaluation of kH,B by this procedure is illustrated in Fig. 9.7(b).
Relations between these Henry’s law constants can be found with the use of Eqs. 9.1.14
and 9.4.16–9.4.18:
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
Henry’s law. On the microscopic level, the requirement is that solute molecules be suffi-
ciently separated to make solute–solute interactions negligible.
Note that an ideal-dilute solution is not necessarily an ideal mixture. Few liquid mix-
tures behave as ideal mixtures, but a solution of any nonelectrolyte solute becomes an ideal-
dilute solution when sufficiently 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
ıB (g) C RT ln.fB =p ı / (Eq. 9.3.12). Substituting fB D kH,B xB (Henry’s law) into this
equation, we obtain
kH,B xB
B D ıB (g) C RT ln
pı
kH,B

D ıB (g) C RT ln ı C RT ln xB (9.4.23)
p
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
x;B , where the x in the subscript reminds us that the reference
state is based on mole fraction. The equation then becomes
B .T; p/ D ref
x;B .T; p/ C RT ln xB (9.4.24)
(ideal-dilute solution
of a nonelectrolyte)
Here the notation emphasizes the fact that B and ref

x;B are functions of T and p.
Equation 9.4.24, derived using fugacity, is valid even if the solute has such low volatil-
ity 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 finite solute fugacity in the gas
phase even if immeasurably small.
It is worthwhile to describe in detail the reference state to which ref
x;B refers. The
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
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 fictitious 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 Henry’s law constant kH,B ; an
example is indicated by the filled circle in Fig. 9.7(a).
By similar steps, combining Henry’s law based on concentration or molality (Eqs.

9.4.17 and 9.4.18) with the relation B D ıB (g) C RT ln.fB =p ı /, we obtain for the solute
chemical potential in the ideal-dilute range the equations
kc;B cB c ı

ı
B D B (g) C RT ln ı
pı c

kc;B c ı
c
D ıB (g) C RT ln ı
C RT ln Bı (9.4.25)
p c
km;B mB mı

B D ıB (g) C RT ln
pı mı
km;B mı

m
D ıB (g) C RT ln ı
C RT ln Bı (9.4.26)
p m
Note how in each equation the argument of a logarithm is multiplied and divided by a
constant, c ı or mı , in order to make the arguments of the resulting logarithms dimension-
less. These constants are called standard compositions with the following values:
standard concentration c ı D 1 mol dm 3 (equal to one mole per liter, or one molar)
standard molality mı 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:
cB
B .T; p/ D ref
c;B .T; p/ C RT ln (9.4.27)
cı
mB
B .T; p/ D ref
m;B .T; p/ C RT ln (9.4.28)
mı
The quantities ref ref

c;B and m;B are the chemical potentials of the solute in hypothetical refer-
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
For consistency with Eqs. 9.4.27 and 9.4.28, we can rewrite Eq. 9.4.24 in the form
xB
B .T; p/ D ref
x;B .T; p/ C RT ln (9.4.29)
xı
with x ı , the standard mole fraction, given by x ı D 1.
9.4.6 Solvent behavior in the ideal-dilute solution

We now use the Gibbs–Duhem equation to investigate the behavior of the solvent in an
ideal-dilute solution of one or more nonelectrolyte solutes. TheP
Gibbs–Duhem equation ap-
i D 0 (Eq. 9.2.43).
plied to chemical potentials at constant T and p can be written i xi dP
We use subscript A for the solvent, rewrite the equation as xA dA C i ¤A xi di D 0,
and rearrange to
1 X
dA D xi di (9.4.30)
xA
i¤A (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
x;i C
RT ln xi and the differential of i at constant T and p is
di D RT d ln xi D RT dxi =xi (9.4.31)
(Here the fact has been used that ref

x;i is a constant at a given T and p.) When we substitute
this expression for di in Eq. 9.4.30, we obtain
RT X
dA D dxi (9.4.32)
xA
i¤A
P
Now since theP sum of all mole fractions is 1, we have the relation i¤A xi D 1 xA whose
differential is i¤A dxi D dxA . Making this substitution in Eq. 9.4.32 gives us
RT
dA D dxA D RT d ln xA (9.4.33)
xA
of nonelectrolytes)
Consider a process in an open system in which we start with a fixed amount of pure
solvent and continuously add the solute or solutes at constant T and p. The solvent mole
fraction decreases from unity to a value xA0 , and the solvent chemical potential changes
from A to 0A . 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:
Z 0A Z xA DxA0
dA D RT d ln xA (9.4.34)

A xA D1
The result is 0A A D RT ln xA0 , or in general
A D A C RT ln xA (9.4.35)
CHAPTER 9 MIXTURES
8 bc
bc
6
law
bc
ry’s
bc
fA =kPa
Hen
bc
bc
4 bc
bc
aw
sl
bc
ult’
bc o
2 Ra
bc
0
0 0:2 0:4 0:6 0:8 1:0
xA
Figure 9.8 Fugacity of ethanol in a gas phase equilibrated with a binary liquid mix-
ture of ethanol (A) and H2 O at 25 ı C and 1 bar. Open circles: experimental measure-
ments. a The dashed lines show Henry’s law behavior and Raoult’s law behavior.
a Ref. [45].
Comparison with Eq. 9.4.5 on page 247 shows that Eq. 9.4.35 is equivalent to Raoult’s law
for fugacity.
Thus, in an ideal-dilute solution of nonelectrolytes each solute obeys Henry’s law and
the solvent obeys Raoult’s law.
An equivalent statement is that a nonelectrolyte constituent of a liquid mixture ap-
proaches Henry’s law behavior as its mole fraction approaches zero, and approaches Raoult’s
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 Raoult’s law and negative deviations from Henry’s 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 Raoult’s 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. 9.4.8–9.4.13 and are collected in the first 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 infinite 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 Henry’s
law, and its chemical potential is given by B D refx;B C RT ln xB (Eq. 9.4.24) where x;B
ref
is a function of T and p, but not of composition. B varies with the composition and goes
to 1 as the solution becomes infinitely dilute (xA !1 and xB !0).
CHAPTER 9 MIXTURES
Table 9.2 Partial molar quantities of solvent and non-

electrolyte solute in an ideal-dilute solution
Solvent Solute
A D A C RT ln xA B D ref
x;B C RT ln xB
ı
D ref
c;B C RT ln.cB =c /
ı
D ref
m;B C RT ln.mB =m /
SA D SA R ln xA ref
SB D Sx;B R ln xB
ref
D Sc;B R ln.cB =c ı /
ref
D Sm;B R ln.mB =mı /
HA D HA HB D HB 1
VA D VA VB D VB1
UA D UA UB D UB1
1
Cp;A D Cp;A Cp;B D Cp;B
For the partial molar entropy of the solute, we use SB D .@B =@T /p;fnig (Eq. 9.2.48)
and obtain !
@ref
x;B
SB D R ln xB (9.4.36)
@T
p
The term .@ref

x;B =@T /p
ref
represents the partial molar entropy Sx;B of B in the fictitious
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 infinite 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
HB D ref ref ref
x;B C T Sx;B D Hx;B (9.4.38)
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
HB D HB1 (9.4.39)
where HB1 is the partial molar enthalpy at infinite dilution. By similar reasoning, using
Eqs. 9.2.49–9.2.52, we find 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 infinite dilution.
The expressions are listed in the second column of Table 9.2.
When the pressure is equal to the standard pressure p ı , the quantities HB1 , VB1 , UB1 ,
1
and Cp;B are the same as the standard values HBı , VBı , UBı , and Cp;B
ı
.
CHAPTER 9 MIXTURES
9.5 ACTIVITY C OEFFICIENTS IN M IXTURES OF N ONELECTROLYTES 257
Table 9.3 Reference states for nonelectrolyte constituents of mixtures. In each refer-
ence state, the temperature and pressure are the same as those of the mixture.
Chemical
Constituent Reference state potential
Substance i in a gas mixture Pure i behaving as an ideal gasa ref

i (g)
Substance i in a liquid or solid Pure i in the same physical state as the i
mixture mixture
Solvent A of a solution Pure A in the same physical state as the A
solution
Solute B, mole fraction basis B at mole fraction 1, behavior extrapo- ref
x;B
lated from infinite dilution on a mole frac-
tion basisa
Solute B, concentration basis B at concentration c ı , behavior extrapo- ref
c;B
lated from infinite dilution on a concen-
tration basisa
Solute B, molality basis B at molality mı , behavior extrapolated ref
m;B
from infinite dilution on a molality basisa
aA hypothetical state.
9.5 ACTIVITY COEFFICIENTS IN MIXTURES OF

NONELECTROLYTES
An activity coefficient 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 behav-
ior 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 i depends only
on the temperature and pressure of the mixture.
If the pressure is the standard pressure p ı , the reference state of species i becomes
its standard state. In the standard state, the chemical potential is the standard chemical
potential ıi , which is a function only of temperature.
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

Since the activity coefficient 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 first 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):
ı pi
id
i (g) D i (g) C RT ln (9.5.1)
pı
The reference state of gaseous substance i is pure i acting as an ideal gas at pressure p. Its
chemical potential is given by
ı p
ref
i (g) D i (g) C RT ln (9.5.2)
pı
Subtracting Eq. 9.5.2 from Eq. 9.5.1, we obtain
pi
id
i (g) ref
i (g) D RT ln (9.5.3)
p
Consider the following expressions for chemical potentials in ideal mixtures and ideal-
dilute solutions of nonelectrolytes. The first equation is a rearrangement of Eq. 9.5.3, and
the others are from earlier sections of this chapter.7
pi
Constituent of an ideal gas mixture id ref
i (g) D i (g) C RT ln (9.5.4)
p

Constituent of an ideal liquid or solid mixture id
Solvent of an ideal-dilute solution id
id
Solute, ideal-dilute solution, mole fraction basis B D ref
x;B C RT ln xB (9.5.7)
cB
Solute, ideal-dilute solution, concentration basis id ref
B D c;B C RT ln ı (9.5.8)
c
m
Solute, ideal-dilute solution, molality basis B D m;B C RT ln Bı
id ref
(9.5.9)
m
Note that the equations for the condensed phases have the general form

id ref composition variable
i D i C RT ln (9.5.10)
standard composition
where ref
i is the chemical potential of component i in an appropriate reference state. (The
standard composition on a mole fraction basis is x ı 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 coefficient. The expression is like one of those for the
ideal case (Eqs. 9.5.4–9.5.9) with the activity coefficient 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 ıi (g) from Eqs. 9.3.12 and
9.5.2, we obtain
fi
i D ref
i (g) C RT ln
p
i pi
D ref
i (g) C RT ln (9.5.11)
p
where fi is the fugacity of constituent i and i is its fugacity coefficient. Here the activity
coefficient is the fugacity coefficient 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 coefficient:

ref composition variable
i D i C RT ln .activity coefficient of i/ (9.5.12)
standard composition
The activity coefficient 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
coefficient is unity and the chemical potential is given by one of the expressions of Eqs.
9.5.4–9.5.9; otherwise, the activity coefficient has the value that gives the actual chemical
potential.
This book will use various symbols for activity coefficients, as indicated in the following
list of expressions for the chemical potentials of nonelectrolytes:

ref pi
Constituent of a gas mixture i D i (g) C RT ln i (9.5.13)
p
Constituent of a liquid or solid mixture i D i C RT ln . i xi / (9.5.14)
Solvent of a solution A D A C RT ln . A xA / (9.5.15)
B D ref

Solute of a solution, mole fraction basis x;B C RT ln x;B xB (9.5.16)
cB
Solute of a solution, concentration basis B D ref
c;B C RT ln c;B ı (9.5.17)
c
m
Solute of a solution, molality basis B D m;B C RT ln m;B Bı
ref
(9.5.18)
m
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 coefficients 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
coefficient 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. 9.5.16–9.5.18, the symbols x;B , c;B , and m;B for activity coefficients of a
nonelectrolyte solute include x, c, or m in the subscript to indicate the choice of the solute
CHAPTER 9 MIXTURES
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 coefficient, on the other hand, depends on the choice
of the solute reference state.
You may find 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 , and x;B xB are effective mole fractions.
c;B cB is an effective concentration.
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 coefficient of each mixture
constituent to approach unity:
Constituent of a gas mixture i ! 1 as p ! 0 (9.5.19)

Constituent of a liquid or solid mixture i !1 as xi ! 1 (9.5.20)
Solvent of a solution A !1 as xA ! 1 (9.5.21)
Solute of a solution, mole fraction basis x;B !1 as xB ! 0 (9.5.22)
Solute of a solution, concentration basis c;B !1 as cB ! 0 (9.5.23)
Solute of a solution, molality basis m;B !1 as mB ! 0 (9.5.24)
9.5.4 Nonideal dilute solutions

How would we expect the activity coefficient 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 interac-
tions 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 261
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 x;B C RT ln xB , can be
ref
described by adding a term proportional to xB : B D x;B C RT ln xB C kx xB , where
kx is a positive or negative constant related to solute-solute interactions.
If we equate this expression for B with the one that defines the activity coeffi-
cient, B D ref
x;B C RT ln. x;B xB / (Eq. 9.5.16), and solve for the activity coefficient,
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 coef-
ficients on a concentration or molality basis. For instance, at low mB the chemical
ı
potential of B should be approximately refm;B C RT ln.mB =m / C km mB , where km is
a constant at a given T and p; then the activity coefficient 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 coefficient in a
dilute solution is a linear function of the composition variable, is borne out experimentally
as illustrated in Fig. 9.10 on page 264. This prediction applies only to a nonelectrolyte
solute; for an electrolyte, the slope of activity coefficient versus molality approaches 1
at low molality (page 290).
9.6 EVALUATION OF ACTIVITY COEFFICIENTS
This section describes several methods by which activity coefficients of nonelectrolyte sub-
stances may be evaluated. Section 9.6.3 describes an osmotic coefficient method that is also
suitable for electrolyte solutes, as will be explained in Sec. 10.6.
9.6.1 Activity coefficients 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 coefficient i in the liquid. The
relation between i and fi will now be derived.
9 This is essentially the result of the McMillan–Mayer solution theory from statistical mechanics.
CHAPTER 9 MIXTURES
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
i C RT ln . i xi / D ref
i (g) C RT ln .fi =p/ (9.6.1)
" #
ref
i (g) i f
i D exp i (9.6.2)
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
i D Ci (9.6.3)
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 260. It says that as xi approaches 1 at constant T and
p, i also approaches 1. We can use this limit to evaluate Ci :
i xi 1
Ci D lim D (9.6.4)
xi !1 fi fi
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:
fi
i D (9.6.5)
xi fi
We can follow the same procedure for a solvent or solute of a liquid solution. We re-
place the left side of Eq. 9.6.1 with an expression from among Eqs. 9.5.15–9.5.18, then
derive an expression analogous to Eq. 9.6.3 for the activity coefficient with a composition-
independent factor, and finally apply the limiting conditions that cause the activity coeffi-
cient to approach unity (Eqs. 9.5.21–9.5.24) and allow us to evaluate the factor. When we
take the limits that cause the solute activity coefficients to approach unity, the ratios fB =xB ,
fB =cB , and fB =mB become Henry’s law constants (Eqs. 9.4.19–9.4.21). The resulting ex-
pressions for activity coefficients as functions of fugacity are listed in Table 9.4 on the next
page.
Examples
Ethanol and water at 25 ı 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 fixed 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 figure, the filled 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 Raoult’s law.
The ratio of these two quantities is the activity coefficient A .
CHAPTER 9 MIXTURES
8 b 4
6 3
fA =kPa bc
A
4 2
bc
fA bc
2 1
xA fA
0 0
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
xA xA
(a) (b)
Figure 9.9 Liquid mixtures of ethanol (A) and H2 O at 25 ı C and 1 bar.

(a) Ethanol fugacity as a function of mixture composition. The dashed line is Raoult’s
law behavior, and the filled circle is the pure-liquid reference state.
(b) Ethanol activity coefficient as a function of mixture composition.
Table 9.4 Activity coefficients as functions of fugac-

ity. For a constituent of a condensed-phase mixture, fi ,
fA , and fB refer to the fugacity in a gas phase equili-
brated with the condensed phase.
Substance Activity coefficient

fi
Substance i in a gas mixture i D
pi
fi
Substance i in a liquid or solid i D
mixture xi fi
fA
Solvent A of a solution A D
xA fA
fB
Solute B, mole fraction basis x;B D
kH,B xB
fB
Solute B, concentration basis c;B D
kc;B cB
fB
Solute B, molality basis m;B D
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 figures
demonstrate how, treating 1-butanol as a solute, we locate the solute reference state by a
CHAPTER 9 MIXTURES
400 8
b
fB =kPa
b
6
fB =kPa
200 4
5 2
0 0
0 0:01 0:5 1 0 0:5 1:0
xB mB =mol kg 1
(a) (b)
1:0 1:0
m;B
x;B
0:5 0:5
0 0
0 0:01 0:02 0 0:5 1:0
xB mB =mol kg 1
(c) (d)
Figure 9.10 Dilute aqueous solutions of 1-butanol (B) at 50:08 ı 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 magnified view. Dashed line:
Henry’s 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: Henry’s law behavior on a molality basis. Filled circle: solute reference
state on this basis.
(c) Activity coefficient on a mole fraction basis as a function of xB .
(d) Activity coefficient on a molality basis as a function of mB .
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 filled circle in
Fig. 9.10(a), fB equals the Henry’s law constant kH,B ; at the reference state in Fig. 9.10(b),
fB equals km;B mı . Note how the activity coefficients plotted in Figs. 9.10(c) and 9.10(d)
approach 1 at infinite 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 coefficients from the Gibbs–Duhem equation

If component B of a binary liquid mixture has low volatility, it is not practical to use its fu-
gacity in a gas phase to evaluate its activity coefficient. If, however, component A is volatile
CHAPTER 9 MIXTURES
enough for fugacity measurements over a range of liquid composition, we can instead use
the Gibbs–Duhem 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 coefficient based on a pure-
liquid reference state: A D fA =.xA fA / (Table 9.4). We wish to use the same fugacity
measurements to determine the activity coefficient of the nonvolatile component, B.
The Gibbs–Duhem equation for a binary liquid mixture in the form given by Eq. 9.2.43
is
xA dA C xB dB D 0 (9.6.6)
where dA and dB are the chemical potential changes accompanying a change of com-
position 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
RT
dA D RT d ln A C RT d ln xA D RT d ln A C dxA (9.6.7)
xA
For component B, we obtain in the same way
RT RT
dB D RT d ln B C dxB D RT d ln B dxA (9.6.8)
xB xB
Substituting these expressions for dA and dB in Eq. 9.6.6 and solving for d ln B, we
obtain the following relation:
xA
d ln B D d ln A (9.6.9)
xB
Integration from xB D 1, where B equals 1 and ln B equals 0, to composition xB0 gives

xB DxB0
xA
Z
0
ln B .xB / D d ln A (9.6.10)
xB D1 xB
(binary mixture,
constant T and p)
Equation 9.6.10 allows us to evaluate the activity coefficient of the nonvolatile component,
B, at any given liquid composition from knowledge of the activity coefficient 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 coefficient 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 infinite. Instead, it is convenient in this case to use the method
described in the next section.
9.6.3 Activity coefficients from osmotic coefficients

It is customary to evaluate the activity coefficient of a nonvolatile solute with a function
m called the osmotic coefficient, or osmotic coefficient on a molality basis. The osmotic
CHAPTER 9 MIXTURES
coefficient of a solution of nonelectrolyte solutes is defined by
def A
m D XA (9.6.11)
RTMA mi (nonelectrolyte solution)
i ¤A
P
The definition of m in Eq. 9.6.11 has the following significance. 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 =2P .1 xA /3 =3 .
Thus, in a very dilute solution we have ln xA .1 xA / D i ¤A xi . In the limit
of infinite dilution, the mole fraction of solute i becomes xi D MA mi (see Eq. 9.1.14).
In the limit of infinite dilution, therefore, we have
X
ln xA D MA mi (9.6.12)
i ¤A (infinite dilution)
and the solvent chemical potential is related to solute molalities by

X
A D A RTMA mi (9.6.13)
i ¤A (infinite dilution)
The deviation of m from unity is a measure of the deviation of A from infinite-

dilution behavior, as we can see by comparing the preceding equation with a rear-
rangement of Eq. 9.6.11:
X
A D A m RTMA mi (9.6.14)
i ¤A
The reason m is called the osmotic coefficient 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 infinite dilution (see Sec. 12.4.4).
Evaluation of m
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 equili-
brated gas phase by A D refA (g) C RT ln.fA =p/ (from Eq. 9.5.11). For the pure solvent,

this relation is A D A (g) C RT ln.fA =p/. Taking the difference between these two
ref
equations, we obtain
f
A A D RT ln A (9.6.15)
fA
which allows us to evaluate m from fugacity measurements.
Osmotic coefficients can also be evaluated from freezing point and osmotic pressure
measurements that will be described in Sec. 12.2.
CHAPTER 9 MIXTURES
Use of m
Suppose we have a solution of a nonelectrolyte solute B whose activity coefficient m;B we
wish to evaluate as a function of mB . For a binary solution, Eq. 9.6.11 becomes
A A
m D (9.6.16)
RTMA mB
(binary nonelectrolyte solution)
Solving for A and taking its differential at constant T and p, we obtain
dA D RTMA d.m mB / D RTMA .m dmB C mB dm / (9.6.17)
From B D ref ı
m;B C RT ln. m;B mB =m / (Eq. 9.5.18), we obtain
m;B mB

dmB
dB D RT d ln D RT d ln m;B C (9.6.18)
mı mB
We substitute these expressions for dA and dB in the Gibbs–Duhem equation in the form
given by Eq. 9.2.26, nA dA C nB dB D 0, make the substitution nA MA D nB =mB , and
rearrange to
1
d ln m;B D dm C m dmB (9.6.19)
mB
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 m0B , with the result
m0B
m 1
Z
0
ln m;B .mB / D m .m0B / 1C dmB (9.6.20)
0 mB
(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 finite 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 coefficient 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 coefficient 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 co-
efficient 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 coefficient of a constituent of a liquid
mixture from its fugacity in a gas phase equilibrated with the mixture. Section 9.6.3 men-
tioned the use of solvent fugacities in gas phases equilibrated with pure solvent and with a
solution, in order to evaluate the osmotic coefficient of the solution.
CHAPTER 9 MIXTURES
0:10 2:5 10
1
0:08 2:0 8
1/=mB =kg mol
1
m;B mB =mol kg
0:06 1:5 6
m;B
0:04 1:0 4
Œ.m
0:02 0:5 2
0 0 0
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
mB =mol kg 1 mB =mol kg 1 mB =mol kg 1
(a) (b) (c)
Figure 9.11 Aqueous sucrose solutions at 25 ı C. a

(a) Integrand of the integral in Eq. 9.6.20 as a function of solution composition.
(b) Solute activity coefficient on a molality basis.
(c) Product of activity coefficient and molality as a function of composition. The
dashed line is the extrapolation of ideal-dilute behavior.
a Based on data in Ref. [145], 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. Dy-
namic 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 deter-
mining 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 269
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 defined by the relation
i ıi

def
ai D exp (9.7.1)
RT
(uncharged species)
or
i D ıi C RT ln ai (9.7.2)
(uncharged species)
where ıi is the standard chemical potential of the species.10 The activity of a species in a
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 definition 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 ele-
vations in a gravitational field, or are at different electric potentials. Section 9.8 10.1 will
describe a modification of the defining equation i D ıi C RT ln ai for a system with
phases of different elevations, and Sec. 10.1 will describe the modification needed for a
charged species.
9.7.1 Standard states

The standard states of different kinds of mixture components have the same definitions 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 ı .
When component i is in its standard state, its chemical potential is the standard chemical
potential ıi . It is important to note from Eq. 9.7.2 that when i equals ıi , the logarithm
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
10 Some chemists define the activity by i D ref

i C RT ln ai . The activity defined this way is not the same as
the activity used in this book unless the phase is at the standard pressure.
CHAPTER 9 MIXTURES
BIOGRAPHICAL SKETCH
Gilbert Newton Lewis (1875–1946)
formation of substances and ions.
UNIVERSITY OF PENNSYLVANIA LIBRARY

In 1912 he became the chair of the chem-
istry department at the University of Cali-
EDGAR FAHS SMITH COLLECTION

fornia 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 mea-
suring the thermodynamic properties of elec-
trolyte solutions, he introduced the concept of
Gilbert Lewis made major contributions to ionic strength (1921).
several fields 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 first to prepare
imental and theoretical thermodynamics, and pure deuterium and D2 O (1933), he formu-
was a chain smoker of vile Manila cigars. lated his generalized definitions 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 No-
theory that introduced for the first time the bel prize, but was never awarded it. Accord-
terms fugacity (1901) and activity (1907). The ing to a history of modern chemistry published
first of these papers was entitled “A New Con- in 2008,c Wilhelm Palmaer, a Swedish electro-
ception of Thermal Pressure and a Theory of chemist, used his position on the Nobel Com-
Solutions” 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 ba-
sary into the nature of the conditions which exist sis of occasional “arithmetic and thermody-
in the interior of any homogeneous phase. It will namic inaccuracy.”d
be the aim of the present paper to study this prob- His career was summarized by his last grad-
lem 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 simplic- Gilbert Lewis once defined physical chemistry
ity and by its ability to explain several important as encompassing “everything that is interesting.”
phenomena which have hitherto received no sat- His own career touched virtually every aspect of
isfactory explanation. science, and in each he left his mark. He is justly
regarded as one of the key scientists in Ameri-
His first faculty position (1905-1912) was at can history. It would be a great omission not to
Boston Tech, now the Massachusetts Institute record the warmth and intellectual curiosity ra-
of Technology, where he continued work in diated by Lewis’ personality. He epitomized the
one of his dissertation subjects: the measure- scientist of unlimited imagination, and the joy of
ment of standard electrode potentials in order working with him was to experience the life of
to determine standard molar Gibbs energies of the mind unhindered by pedestrian concerns.
a Ref. [103]. b Ref. [104]. c Ref. [35], Chap. 7. d Ref. [104], page 6. e Ref. [85].
CHAPTER 9 MIXTURES
components:
Substance i in a gas mixture i D ıi (g) C RT ln ai (g) (9.7.3)
Substance i in a liquid or solid mixture i D ıi C RT ln ai (9.7.4)
Solvent A of a solution A D ıA C RT ln aA (9.7.5)
Solute B, mole fraction basis B D ıx;B C RT ln ax;B (9.7.6)
Solute B, concentration basis B D ıc;B C RT ln ac;B (9.7.7)
Solute B, molality basis B D ım;B C RT ln am;B (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 finding 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
ı ı
potentials, ref
i and i , are equal only if the mixture is at the standard pressure p .
It will be useful to define the following dimensionless quantity:
!
ref ı
def i i
i D exp (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 definition of activity given by
i D ıi C RT ln ai with a rearrangement of Eq. 9.7.9: ref ı
i D i C RT ln i .
ref ı
At a given temperature, the difference i i depends only on the pressure p of the
ı
mixture, and is zero when p is equal to p . Thus i is a function of p with a value of 1
when p is equal to p ı . 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
mB
B D ref
m;B C RT ln m;B ı
m
ı
D m;B C RT ln am;B (9.7.10)
The activity is then given by
ref
m;B ım;B
mB
ln am;B D C ln m;B
RT mı
mB
D ln m;B C ln m;B ı (9.7.11)
m
m
am;B D m;B m;B Bı (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 coefficient, 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.
11 Ref. [105], p. 249.
CHAPTER 9 MIXTURES
Table 9.5 Expressions for activities of nonelectrolytes. For a con-

stituent of a condensed-phase mixture, fi , fA , and fB refer to the
fugacity in a gas phase equilibrated with the condensed phase.
Substance Activity
f
Pure gas a(g) D (g) D
pı
Pure liquid or solid aD
pi f
Substance i in a gas mix- ai (g) D i (g) i D iı
ture p p
fi
Substance i in a liquid or ai D i i xi D i
solid mixture fi
fA
Solvent A of a solution aA D A A xA D A
fA
fB
Solute B, mole fraction ax;B D x;B x;B xB D x;B
basis kH,B
cB fB
Solute B, concentration ac;B D c;B c;B ı D c;B
basis c kc;B c ı
mB fB
Solute B, molality basis am;B D m;B m;B D m;B
mı km;B mı
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 ıi / as functions of pressure to substitute in
the defining equation i D exp Œ .ref
i ıi /=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
on its pressure according to Eq. 7.8.6: D ı (g) C RT ln .p=p ı /. Thus the chemical
ı ı
potential of the reference state of gas component i is ref i (g) D i (g) C RT ln .p=p /,
and i (g) ıi (g) is equal to RT ln .p=p ı /. This gives us the following expression for the
ref
pressure dependence of the pressure factor:

p
i (g) D (9.7.13)
pı
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 composi-
tion is therefore di 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 ı 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
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 di D Vi dp:
Z p0
0
ref
i .p / ıi D Vi dp (9.7.14)
pı
The appropriate partial molar volume Vi is the molar volume Vi or VA of the pure sub-
stance, or the partial molar volume VB1 of solute B at infinite 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 / ıc;B .
0 ı
We can derive a rigorous expression for ref
c;B .p / c;B as follows. Consider an ideal-
dilute solution of solute B at an arbitrary pressure p, with solute chemical potential
ı
given by B D refc;B C RT ln.cB =c / (Table 9.2). From this equation we obtain
!
@ref @ ln.cB =c ı /

@B c;B
D C RT (9.7.15)
@p T;fnig @p @p T;fnig
T
The partial derivative .@B =@p/T;fnig is equal to the partial molar volume VB (Eq.
9.2.49), which in the ideal-dilute solution has its infinite-dilution value VB1 . We
rewrite the second partial derivative on the right side of Eq. 9.7.15 as follows:
@ ln.cB =c ı /

1 @cB 1 @.nB =V /
D D
@p T;fnig cB @p T;fnig nB =V @p T;fnig

@.1=V / 1 @V
DV D
@p T;fnig V @p T;fnig
D T (9.7.16)
Here T is the isothermal compressibility of the solution, which at infinite dilution is

T1 , the isothermal compressibility of the pure solvent. Equation 9.7.15 becomes
!
1
@ref
c;B
VB D C RT T1 (9.7.17)
@p
T
ı 0
Solving for dref
c;B at constant T , and integrating from p to p , we obtain finally
Z p0
0
ref ıc;B D VB1 RT T1 dp

c;B .p / (9.7.18)
pı
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 ı 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 find 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 274
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 Pressure factor at pressure p 0
p0
Substance i in a gas mixture, or the i (g) D
pure gas pı
p0
!
Vi Vi .p 0 p ı /
Z
Substance i in a liquid or solid i D exp dp exp
mixture, or the pure liquid or solid pı RT RT
Z p0 ! 0
VA .p p ı /

VA
Solvent A of a solution A D exp dp exp
p ı RT RT
Z p0 1 ! 1 0
VB .p p ı /

VB
Solute B, mole fraction or molality x;B D m;B D exp dp exp
basis p ı RT RT
"Z 0 #
p
VB1
1 0
VB .p p ı /

1
Solute B, concentration basis c;B D exp T dp exp
pı RT RT
1:49 at 100 bar. For a solution with VB1 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 =mı ,
and so on.
In principle, we can specify any convenient value for the standard pressure p ı . For a
chemist making measurements at high pressures, it would be convenient to specify a
value of p ı within the range of the experimental pressures, for example p ı 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 field, 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 field

Consider a tall column of a gas mixture in an earth-fixed 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 195. We want to find the equilibrium conditions reached spontaneously when the
system is isolated from its surroundings.
CHAPTER 9 MIXTURES
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 ’0
dS C ’0
dn’i (9.8.1)
0
T 0
T
’¤’ i ’¤’
In an equilibrium state, S is at a maximum and dS is zero for an infinitesimal change of

any of the independent variables. This requires the coefficient 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 defined in
this book as the molar Gibbs energy at that elevation (page 196), the chemical potential of
substance i in a mixture at elevation h is the partial molar Gibbs energy at that elevation.
We define the standard potential ıi (g) of component i of the gas mixture as the chem-
ical potential of i under standard state conditions at the reference elevation hD0. At this
elevation, the chemical potential and fugacity are related by
fi .0/
i .0/ D ıi (g) C RT ln (9.8.2)
pı
If we reversibly raise a small sample of mass m of the gas mixture by an infinitesimal
distance dh, without heat and at constant T and V , the fugacity fi remains constant. The
¶w 0 D mg dh. This work contributes to
gravitational work during the elevation process isP
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 T S C pV /
X
D S dT C V dp C i dni C mg dh (9.8.3)
i
From this total differential, we write the reciprocity relation

@i @mg
D (9.8.4)
@h T;p;fnig @ni T;p;nj ¤i ;h
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 di D
Mi g dh. Integrating over a finite elevation change from h D 0 to h D h0 , we obtain
Z h0
i .h0 / i .0/ D Mi g dh D Mi gh0 (9.8.5)
0 0
(fi .h /Dfi .0/ )
The general relation between i , fi , and h that agrees with Eqs. 9.8.2 and 9.8.5 is
fi .h/
i .h/ D ıi (g) C RT ln C Mi gh (9.8.6)
pı
CHAPTER 9 MIXTURES
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
fi .h/ f .0/
ıi (g) C RT ln ı
C Mi gh D ıi (g) C RT ln i ı (9.8.7)
p p
Solving for fi .h/ gives
Mi gh=RT
fi .h/ D fi .0/e (9.8.8)
(gas mixture at equilibrium)
If the gas is an ideal gas mixture, fi is the same as the partial pressure pi :
Mi gh=RT
pi .h/ D pi .0/e (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 197. 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 confined to a cell
embedded in a spinning centrifuge rotor.
The system is the solution. The rotor’s 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 fixed 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 498).
Note that an equilibrium state can only exist relative to the rotating local frame; an
observer fixed 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 system’s 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 body’s velocity in the rotating local frame approaches zero.
The centrifugal and Coriolis forces are apparent or fictitious 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 Newton’s second law of motion to relate the net force on a body and the body’s
acceleration in the rotating frame (see Sec. G.6).
CHAPTER 9 MIXTURES
z
y 0
r
#
x
x 0
(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 fixed 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 negligible—all the measurable variation is along the radial direction.
To find 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 find the dependence of pressure on r as follows. Consider one of the thin slab-
shaped 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 r•r. From
Newton’s second law, the sum of these components is zero at equilibrium:
pAs .p C •p/As C As ! 2 r•r D 0 (9.8.10)
or •p D ! 2 r•r. In the limit as •r and •p are made infinitesimal, this becomes
dp D ! 2 r dr (9.8.11)
CHAPTER 9 MIXTURES
We will assume the density is uniform throughout the solution.13 Then integration of Eq.
9.8.11 yields
p 00 r 00
! 2 h 00 2
Z Z
00 0 2
2 i
p p D dp D ! r dr D r r0 (9.8.12)
p0 r0 2
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 equi-
librium 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 field to be negligible. The reversible work in the local frame needed
to move this small sample an infinitesimal 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 T S C pV /
D S dT C V dp C A dnA C B dnB m! 2 r dr (9.8.14)
We use Eq. 9.8.14 to write the reciprocity relation

@B 2 @m
D ! r (9.8.15)
@r T;p;nA ;nB @nB T;p;nA ;r
Then, using m D nA MA C nB MB , we obtain

@B
D MB ! 2 r (9.8.16)
@r T;p;nA ;nB
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 dB D MB ! 2 r dr.
We integrate from radial position r 0 to position r 00 to obtain
Z r 00
00 0 2
B .r / B .r / D MB ! r dr
r0
h 2 2 i
2
D 1
2 MB ! 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 define the standard chemical potential ıc;B as the solute chemical
13 In
the centrifugal field, this assumption is strictly true only if the solution is incompressible and its density is
independent of composition.
CHAPTER 9 MIXTURES
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 0 / D ıc;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 B and ac;B
at an arbitrary radial position r 00 :
h 2 i
2
B .r 00 / D ıc;B C RT ln ac;B .r 00 / 21 MB ! 2 r 00 r0 (9.8.19)
We found earlier that when the solution is in an equilibrium state, B is independent

of r—that 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:
ac;B .r 00 / MB ! 2 h 00 2 2 i
ln D r r0 (9.8.20)
ac;B .r 0 / 2RT
(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 ı . We

assume the solution is sufficiently dilute for the activity
1coefficient c;B to be approximated
ı
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:
ac;B .r 00 / VB1 .p 00 p 0 / cB .r 00 /
ln D C ln (9.8.21)
ac;B .r 0 / RT cB .r 0 /
We substitute for p 00 p 0 from Eq. 9.8.12. It is also useful to make the substitution VB1 D
MB vB1 , where vB1 is the partial specific volume of the solute at infinite dilution (page 234).
When we equate the two expressions for lnŒac;B .r 00 /=ac;B .r 0 /, we obtain finally
cB .r 00 / MB 1 vB1 ! 2 h 00 2

i
0 2
ln D r r (9.8.22)
cB .r 0 / 2RT
(solution in centrifuge
cell at equilibrium)
This equation shows that if the solution density is less than the effective solute density
1=vB1 , so that vB1 is less than 1, the solute concentration increases with increasing distance
from the axis of rotation in the equilibrium state. If, however, is greater than 1=vB1 , the
concentration decreases with increasing r. The factor 1 vB1 is like a buoyancy factor
for the effect of the centrifugal field 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 ı 2
2 that a plot of ln .cB =c / versus r will be linear,
1
with a slope equal to MB 1 vB ! =2RT . The partial specific volume vB1 is found
from measurements of solution density as a function of solute mass fraction (page 234). By
this means, the molar mass MB of the macromolecule is evaluated.
CHAPTER 9 MIXTURES
P ROBLEMS 280
PROBLEMS
in Appendix I.
9.1 For a binary solution, find 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 ı C. a
xB ŒV (mix)=n=cm3 mol 1

xB ŒV (mix)=n=cm3 mol 1
0 0 0:555 0:005
0:049 0:027 0:597 0:011
0:097 0:050 0:702 0:029
0:146 0:063 0:716 0:035
0:199 0:077 0:751 0:048
0:235 0:073 0:803 0:056
0:284 0:074 0:846 0:058
0:343 0:065 0:897 0:057
0:388 0:053 0:944 0:049
0:448 0:032 1 0
0:491 0:016
a Ref. [158].
(b) The molar volumes of mixing for liquid binary mixtures of 1-hexanol (A) and 1-octene
(B) at 25 ı 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
VB1 of B at infinite 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 coeffi-
cients at this temperature have the values14
BAA D 4:8 10 6 m3 mol 1
BBB D 124:5 10 6 m3 mol 1
6
BAB D 47:5 10 m3 mol 1
14 Refs. [3], [47], and [48].
CHAPTER 9 MIXTURES
P ROBLEMS 281
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 ı C and 1 bar, the Henry’s 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 liquid–gas systems formed by
equilibrating liquid water and air at 25 ı 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 coefficient
BAB is the average of BAA and BBB . Show that when this rule holds, the fugacity coef-
ficient 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 coefficients 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 coefficient 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 find 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. [172]. 16 Ref. [135].
CHAPTER 9 MIXTURES
P ROBLEMS 282
Table 9.8 Activity coefficient of ben-

zene (A) in mixtures of benzene and
1-octanol at 20 ı C. The reference state
is the pure liquid.
xA A xA A
a
0 2:0 0:7631 1:183
0:1334 1:915 0:8474 1:101
0:2381 1:809 0:9174 1:046
0:4131 1:594 0:9782 1:005
0:5805 1:370
a extrapolated
Table 9.9 Liquid and gas compositions in the two-phase sys-

tem of methanol (A) and benzene (B) at 45 ı C a
xA yA p=kPa xA yA p=kPa
0 0 29:894 0:4201 0:5590 60:015
0:0207 0:2794 40:962 0:5420 0:5783 60:416
0:0314 0:3391 44:231 0:6164 0:5908 60:416
0:0431 0:3794 46:832 0:7259 0:6216 59:868
0:0613 0:4306 50:488 0:8171 0:6681 58:321
0:0854 0:4642 53:224 0:9033 0:7525 54:692
0:1811 0:5171 57:454 0:9497 0:8368 51:009
0:3217 0:5450 59:402 1 1 44:608
a Ref. [157].
9.9 Table 9.9 lists measured values of gas-phase composition and total pressure for the binary
two-phase methanol–benzene system at constant temperature and varied liquid-phase compo-
sition. 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 Raoult’s law positive or negative?
(c) Tabulate and plot the activity coefficient 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 Henry’s law constant kH,A of methanol in the benzene environment at 45 ı 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 po-
tential of solute B on composition is given by
mB
B D ref
m;B C RT ln C km mB
mı
CHAPTER 9 MIXTURES
P ROBLEMS 283
where ref
m;B and km are constants at a given T and p. (The derivation of this equation is
sketched in Sec. 9.5.4.) Use the Gibbs–Duhem equation in the form dA D .nB =nA / dB 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 coefficients of aqueous so-
lutions of urea at 25 ı 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
0:050 mB =mı
m D 1:00
1:00 C 0:179 mB =mı
where mı is 1 mol kg 1 . Calculate values of the solvent and solute activity coefficients A
and m;B at various molalities in the range 0–20 mol kg 1 , and plot them versus mB =mı . 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 coef-
ficient 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 ı 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 field.
(a) Derive an expression for ln Œ cB .h/=cB .0/ , where cB .h/ and cB .0/ are the solute concen-
trations at elevations h and 0. Your expression should be a function of h, MB , T , , and
the partial specific volume of the solute at infinite dilution, vB1 . For the dependence of
pressure on elevation, you may use the hydrostatic formula dp D g dh (Eq. 8.1.14 on
page 197) 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
molar mass MB D 10:0 kg mol 1 and partial specific volume vB1 D 0:78 cm3 g 1 . The
solution density is D 1:00 g cm 3 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
authors used the values vB1 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. [150]. 18 Ref. [18].
CHAPTER 9 MIXTURES
P ROBLEMS 284
0:5
0:0
0:5
ln A
1:0
1:5
2:0
2:5
48 49 50 51 52
r 2 /cm2
Figure 9.13 Sedimentation equilibrium of a dilute solution of the FhuA-dodecyl mal-

toside aggregate.
(b) Find the molar mass of the aggregate solute species, and use it to estimate the mass bind-
ing ratio (the mass of bound detergent divided by the mass of protein).
C HAPTER 10
E LECTROLYTE S OLUTIONS
The thermodynamic properties of electrolyte solutions differ in significant 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 infinite dilution is not
finite, but zero. What is the reason for this non-Henry’s law behavior? It must be because
HCl is an electrolyte—it 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 ıC , for instance, denotes the chemical potential of a cation in a standard
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 definitions of chemical potential, activity coefficient, and activity to these different
6
bc
4
pHCl =102 Pa
3
bc
2
bc
1
bc
bc bc bc
bc bc
0 bc bc
0 2 4 6 8 10
mHCl =mol kg 1
Figure 10.1 Partial pressure of HCl in a gas phase equilibrated with aqueous HCl at
25 ı C and 1 bar. Open circles: experimental data from Ref. [5].
285
CHAPTER 10 ELECTROLYTE SOLUTIONS
10.1 S INGLE - ION Q UANTITIES 286
species, but only the activity coefficient 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 defined as the amount of solute formula unit divided
by the mass of solvent.
We first 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 defined as the work needed to reversibly move an infinitesimal test
charge into the phase from a position infinitely 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 infinitesimal amount dnC of the cation is
transferred into a solution phase at constant T and p. The quantity of charge transferred
is ıQ 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 finite electric poten-
tial , the transfer process changes its electrical potential energy by ıQ 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 C .0/ C zC F dnC (10.1.3)
The chemical potential of the cation in a phase of electric potential , defined by the partial
molar Gibbs energy Œ@G./=@nC T;p , is therefore given by
C ./ D C .0/ C zC F (10.1.4)
The corresponding relation for an anion is
./ D .0/ C z F (10.1.5)

1 The Faraday constant (page 452) is the charge per amount of protons.
10.1 S INGLE - ION Q UANTITIES 287
where z is the charge number of the anion ( 1, 2, etc.). For a charged species in general,
we have
i ./ D i .0/ C zi F (10.1.6)
We define 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 mı
with behavior extrapolated from infinite dilution on a molality basis, in a phase of pressure
p D p ı and electric potential D0.
The standard chemical potential ıC or ı of a cation or anion is the chemical potential
of the ion in its standard state. Single-ion activities aC and a in a phase of zero electric
potential are defined by relations having the form of Eq. 9.7.8:
C .0/ D ıC C RT ln aC .0/ D ı C RT ln a (10.1.7)
As explained on page 269, 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 D ıC C RT ln aC C zC F D ı C RT ln a C zi F (10.1.8)
These relations are definitions of single-ion activities in a phase of electric potential .

For a charged species in general, we can write2
i D ıi 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 269.
Of course there is no experimental way to evaluate either C or relative to a refer-
ence 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
C
C D refC C RT ln C ı D ref
C RT ln (10.1.10)
m mı
Here ref
C and
ref are the chemical potentials of the cation and anion in solute reference
states. Each reference state is defined 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 mı , and the ion behaves according to Henry’s law
based on molality. C and are single-ion activity coefficients on a molality basis.
2 Some thermodynamicists call the quantity .ıi C RT ln ai /, which depends only on T , p, and composition,
the chemical potential of ion i , and the quantity .ıi C RT ln ai C zi F / the electrochemical potential with
symbol Q i .
10.2 S OLUTION OF A S YMMETRICAL E LECTROLYTE 288
The single-ion activity coefficients approach unity in the limit of infinite 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 finite solute
molality, the values of C and are the ones that allow Eq. 10.1.10 to give the correct
values of the quantities .C ref C and .
/ ref /. We have no way to actually measure
these quantities experimentally, so we cannot evaluate either C or .
We can define single-ion pressure factors C and as follows:
!
ıC
1
ref VC .p p ı /

def C
C D exp exp (10.1.12)
RT RT
ref ı V 1 .p p ı /

def
D exp exp (10.1.13)
RT RT
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
mC m
aC D C C a D (10.1.14)
mı mı
Then, from Eq. 10.1.9, we have the following relations between the chemical potentials and
molalities of the ions:
C D ıC C RT ln. C C mC =m
ı
/ C zC F (10.1.15)
D ı C RT ln. m =mı / C z F (10.1.16)

Like the values of C and , values of the single-ion quantities aC , a , C, and
cannot be determined by experiment.
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 electrolyte—a 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. In an aqueous solution,
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 defined in the
usual way as the partial molar Gibbs energy

def @G
B D (10.2.1)
@nB T;p;nA
and is a function of T , p, and the solute molality mB . Although B under given conditions
must in principle have a definite 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
differences B ref m;B and B ım;B .
We can write the additivity rule (Eq. 9.2.25) for G as either
G D nA A C nB B (10.2.2)
or
G D nA A C nC C C n (10.2.3)
A comparison of these equations for a symmetrical electrolyte (nB D nC D n ) gives us
the relation
B D C C (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
C C appearing in Eq. 10.2.4 is
C ./ C ./ D C .0/ C .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 ref
m;B D C C
ref
with reference states at D0, set the ion molalities mC and m equal to mB , and obtain
m 2
ref B
B D m;B C RT ln C (10.2.6)
mı
(D2)
The important feature of this relation is the appearance of the second power of mB =mı ,
instead of the first 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 coefficient ˙ , defined for a symmet-
rical electrolyte by
def p
˙ D C (10.2.7)
(D2)
With this definition, Eq. 10.2.6 becomes

m 2
2 B
B D ref
m;B C RT ln . ˙/ (10.2.8)
mı
(D2)
Since it is possible to determine the value of B ref

m;B for a solution of known molality,
˙ is a measurable quantity.
If the electrolyte (e.g., HCl) is sufficiently volatile, its mean ionic activity coefficient 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 coefficients. 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 defined by Eq. 9.7.8 on
page 271:
B D ım;B C RT ln am;B (10.2.9)
Here ım;B is the chemical potential of the solute in its standard state, which is the solute
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
m 2
am;B D m;B . ˙ /2 B
(10.2.10)
mı
(D2)
where m;B is the pressure factor defined by
ım;B
!
def ref
m;B
m;B D exp (10.2.11)
RT
We can use the appropriate expression in Table 9.6 on page 274 to evaluate m;B at an
arbitrary pressure p 0 :
Z p0 1 ! 1 0
VB .p p ı /

0 VB
m;B .p / D exp dp exp (10.2.12)
p ı RT RT
The value of m;B is 1 at the standard pressure, and close to 1 at any reasonably low
pressure (page 274). For this reason it is common to see Eq. 10.2.10 written as am;B D
2 ı 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 infinite dilution, to the square of the HCl molality. In contrast, the activity
of a nonelectrolyte solute is proportional to the first 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 285, and is confirmed 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 =mı /2 . The extension of this line to mB D mı establishes the hypothetical
solute reference state based on molality, indicated by a filled 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 figure is also the solute standard state.)
The solid curve of Fig. 10.2(c) shows how the mean ionic activity coefficient 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 infinite 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
coefficient of a nonelectrolyte solute.
For a symmetrical strong electrolyte, ˙ is the geometric average of the single-ion
activity coefficients 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 291
10 4
1:0 b
1:0
am;B
am;B

0:5 0:5
0 0 0
0 10 4 0 0:5 1:0 0 0:5 1:0
2 =mol2 kg 2
mB m2B =mol2 kg 2 mB =mol kg 1
(a) (b) (c)
Figure 10.2 Aqueous HCl at 25 ı 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 filled circle indicates the solute reference
state.
(c) Mean ionic activity coefficient 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 define
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 mo-
lalities are related to the overall solute molality by
mC D C mB m D mB (10.3.1)
From the additivity rule for the Gibbs energy, we have
G D nA A C nB B
D nA A C C nB C C nB (10.3.2)
giving the relation

B D C C C (10.3.3)
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 ./
D C C .0/ C .0/ C .C zC C z /F (10.3.4)
Electrical neutrality requires that .C zC C z / be zero, giving
B D C C .0/ C .0/ (10.3.5)
By combining Eq. 10.3.5 with Eqs. 10.1.10, 10.3.1, and 10.3.3, we obtain
h mB i
B D ref
B C RT ln CC C
C
(10.3.6)
mı
where ref ref
B D C C C
ref is the chemical potential of the solute in the hypothetical
reference state at D0 in which B is at the standard molality and behaves as at infinite
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 coefficient ˙ is defined in general by

C

˙ D C (10.3.7)
or 1=
C

˙ D C (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
i
mB
h
C
B D ref
B C RT ln C ˙ (10.3.9)
mı
Since B refB is a measurable quantity, so also is ˙ .
The solute activity, defined by B D ım;B C RT ln am;B , is
m
am;B D CC
m;B ˙
B
(10.3.10)
mı
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 ı D ı C ı , we obtain the relation
B D C C C B C C
C
m;B D C (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 D C C C (10.3.12)
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 com-
pletely 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 A C nB B C nC C (10.3.13)
and also, using single-ion quantities, in the form
G D nA A 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 find the following relations must exist between
the chemical potentials of the solute substances and the ion species:
B D .Ba2C / C 2.I / C 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 defined by the relation B D ıB C RT ln am;B (Eq. 10.2.9).
Using this relation together with Eqs. 10.1.7 and 10.1.14, we find that the solute activity is
related to ion molalities by
C m
mC

am;B D m;B ˙ (10.3.16)
mı mı
where the pressure factor m;B is defined in Eq. 10.2.11. The ion molalities in this ex-
pression 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 coefficient 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 pairs—closely associated ions of opposite charge. Furthermore, in so-
lutions 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 signifi-
cance of being the geometric average of the activity coefficients of the actual dissociated
ions, and is called the stoichiometric activity coefficient of the electrolyte.
10.4 T HE D EBYE –H ÜCKEL T HEORY 294
10.4 THE DEBYE–HÜCKEL THEORY
The theory of Peter Debye and Erich Hückel (1923) provides theoretical expressions for
single-ion activity coefficients and mean ionic activity coefficients in electrolyte solutions.
The expressions in one form or another are very useful for extrapolation of quantities that
include mean ionic activity coefficients to low solute molality or infinite 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 ion–ion interactions would be zero,
because each cation–cation or anion–anion repulsion would be balanced by a cation–anion
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 ion–ion interactions, meaning that the single-ion activity
coefficient C becomes less than 1 as the electrolyte molality is increased beyond the ideal-
dilute range. Similarly, also becomes less than 1.
According to the Debye–Hückel theory, the single-ion activity coefficient i of ion i in
a solution of one or more electrolytes is given by
ADH zi2 Im
p
ln i D p (10.4.1)
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, defined by3
def 1
X
Im D 2 mj zj2 (10.4.2)
all ions
ADH and BDH are defined 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 definitions of the quantities ADH and BDH appearing in Eq. 10.4.1 are
def 1=2 3=2
ADH D NA2 e 3 =8 2A

.r 0 RT / (10.4.3)
def 1=2
BDH D NA e 2A .r 0 RT / 1=2

(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. [102]) introduced the term ionic strength, defined by this equation, two years before
the Debye–Hückel theory was published. They found empirically that in dilute solutions, the mean ionic activity
coefficient of a given strong electrolyte is the same in all solutions having the same ionic strength.
BIOGRAPHICAL SKETCH
Peter Josephus Wilhelmus Debye (1884–1966)
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 conduc-
tivity of a dilute solution of a strong electrolyte
is proportional to the square root of the con-
centration. 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 Hückel, to study the literature and find
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. Hückel theory came about.b
Henri Sack, a close associate for 40 years, In a reminiscence of Debye published in
recalled in 1968:a 1972, Erich Hückel wrote:c
He was not only endowed with a most powerful My personal relations with Debye were always
and penetrating intellect and an unmatched abil- completely care-free. Although I was 12 years
ity for presenting his ideas in a most lucid way, younger than he and a complete freshman when
but he also knew the art of living a full life. He I came to Zürich, he always treated me as his
greatly enjoyed his scientific endeavors, he had equal.
a deep love for his family and home life, and he . . . Debye conceived his work—in my
had an eye for the beauties of nature and a taste opinion—as an artist who operates on the ba-
for the pleasure of the out-of-doors as manifested sis of joy in his work and its creations, and who
by his hobbies such as fishing, collecting cacti, was led often by intuition, which was then later
and gardening, mostly in the company of Mrs. on rationally founded in the most plain and clear
Debye. way leaving out everything that was unessential.
Before World War II, Debye held appoint- . . . I never found in Debye any interest in philo-
sophical questions. Debye’s way of life seemed
ments at several universities in The Nether-
to me rather straightforward and uncomplicated.
lands, Switzerland, and Germany. He emi-
He liked a good dinner: when a problem could
grated to America in 1940 and was at Cornell not be solved after a physics lecture, he used to
University in Ithaca, New York until his death. say: “one must enjoy a good evening dinner and
He became an American citizen in 1946. then the inspiration comes by itself”. . . Debye
Debye was responsible for theoretical treat- received an immense number of awards. It did
ments of a variety of subjects, including not seem to matter much to him. When I vis-
molecular dipole moments (for which the de- ited him in Berlin to congratulate him on the
bye is a non-SI unit), X-ray diffraction and Nobel Prize, he interrupted: “Fine that you are
scattering, and light scattering. His theories here.” My congratulations were therefore not
relevant to thermodynamics include the tem- completed.
perature dependence of the heat capacity of Debye was awarded the 1936 Nobel Prize in
crystals at a low temperature (Debye crys- Chemistry “for his contributions to our knowl-
tal theory), adiabatic demagnetization, and the edge of molecular structure through his in-
Debye–Hückel theory of electrolyte solutions. vestigations on dipole moments and on the
In an interview in 1962, Debye said that he diffraction of X-rays and electrons in gases.”
a Ref. [57], page 232. b Ref. [43]. c Translation in Ref. [161], pages 73–74.
1:0
0:9
˙
0:8
0:7
0 0:02 0:04 0:06 0:08 0:10
mB =mol kg 1
Figure 10.3 Mean ionic activity coefficient of aqueous HCl at 25 ı C. Solid curve:
experiment; a dashed curve: Debye–Hückel theory with a D 5 10 10 m; dotted
curve: Debye–Hückel limiting law.
a Ref. [74], Table 11-5-1.
When the solvent is water at 25 ı C, the quantities ADH and BDH have the values
ADH D 1:1744 kg1=2 mol 1=2 (10.4.5)

BDH D 3:285 109 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 ob-
tain the following expression for the logarithm of the mean ionic activity coefficient of an
electrolyte solute: ˇ ˇp
ADH ˇzC z ˇ Im
ln ˙ D p (10.4.7)
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 finite solute molalities, and zero at infinite
dilution, the theory predicts that ˙ is less than 1 at finite solute molalities and approaches
1 at infinite dilution.
Figure 10.3 shows that with the proper choice of the parameter a, the mean ionic activity
coefficient 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
ˇ ˇp
ln ˙ D ADH ˇzC z ˇ Im (10.4.8)
(infinite dilution)
This form is known as the Debye–Hückel 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 ÜCKEL E QUATION 297
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 coefficient 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
by Im D .1=2/ i mi zi2 D .1=2/.C zC 2
C z 2 /mB . With the help of the electroneutrality
P
condition C zC D . z /, this becomes
Im D 21 Œ . z /zC .C zC /z mB

1
D 2Œ . C C /zC z mB
1
ˇ ˇ
D 2
ˇzC z ˇ mB (10.4.9)
We find 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 Debye–Hückel
limiting law (Eq. 10.4.8), but at a low ionic strength the curves begin to deviate significantly
from the linear relations predicted by that law. The full Debye–Hückel equation (Eq. 10.4.7)
fits the experimental curves over a wider range of ionic strength.
10.5 DERIVATION OF THE DEBYE–HÜCKEL EQUATION
Debye and Hückel derived Eq. 10.4.1 using a combination of electrostatic theory, statis-
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 dis-
tribution 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 distribu-
tion, whatever it is, must be spherically symmetric about the central ion; that is, a function
only of the distance r from the center of the ion. The local concentration, cj0 , 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
ions and is symmetric about the central ion, so expressions must be found for cj0 and as
functions of r that are mutually consistent.
10.5 D ERIVATION OF THE D EBYE –H ÜCKEL E QUATION 298
0:1 HCl
0:2
˙
CaCl2
ln
0:3
0:4
0:5
0 0:1 0:2 0:3
1 1=2

Im =mol kg
p
Figure 10.4 Dependence of ln ˙ on Im for aqueous HCl (upper curves) and aque-
ous CaCl2 (lower curves) at 25 ı C. a Solid curves: experimental; dashed curves:
Debye–Hückel equation (a D 5 10 10 m for HCl, a D 4:5 10 10 m for CaCl2 );
dotted lines: Debye–Hückel limiting law.
a Experimental curves from parameter values in Ref. [74], Tables 11-5-1 and 12-1-3a.
Debye and Hückel assumed that cj0 is given by the Boltzmann distribution
cj0 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 con-
stant (k D R=NA ). As r becomes large, approaches zero and cj0 approaches the macro-
scopic concentration cj . As T increases, cj0 at a fixed value of r approaches cj because
of the randomizing effect of thermal energy. Debye and Hückel expanded the exponential
function in powers of 1=T and retained only the first two terms: cj0 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=4r 0 r/e .a r/ =.1 C a/ (10.5.2)
Here is defined by 2 D 2NA2 e 2 IcP=r 0 RT , where Ic is the ionic strength on a concen-
tration basis defined 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 infinite dilution, zi e=4r 0 r, and the
electric potential due to all other ions, 0 . Thus, 0 is equal to zi e=4r 0 r, or
0 D .zi e=4r 0 r/Œe .a r/

=.1 C a/ 1 (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=4r 0 /=.1 C a/. The interaction energy
10.6 M EAN I ONIC ACTIVITY C OEFFICIENTS FROM O SMOTIC C OEFFICIENTS 299
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 final 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 Œ.˛zi e=4r 0 /=.1 C a/ d.˛zi /
˛D0
D .zi2 e 2 =8r 0 /=.1 C a/ (10.5.4)
Since the infinitesimal 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 =8r 0 /=.1 C a/. This quantity is G=ni for the process in
which a solution of fixed composition changes from a hypothetical state lacking ion–ion
interactions to the real state with ion–ion interactions present. G=ni may be equated to
the difference of the chemical potentials of i in the final and initial states. If the chemical
potential without ion–ion interactions is taken to be that for ideal-dilute behavior on a mo-
ı 2 2
lality basis, i D ref
m;i C RT ln.mi =m /, then .zi e NA =8r 0 /=.1 C a/ is equal to
i Œref ı
m;i C RT ln.mi =m / D RT ln m;i . In a dilute solution, ci can with little error be

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 coefficient of a strong electrolyte, or the stoichio-
metric activity coefficient of an electrolyte that does not dissociate completely.
The general procedure described in this section for evaluating ˙ requires knowledge
of the osmotic coefficient 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 coefficient of a binary solution of an electrolyte is defined by
def A A
m D (10.6.1)
RTMA mB
(binary electrolyte solution)
P
That is, for an electrolyte the sum i¤A mi appearing in the definition of m for a nonelec-
trolyte solution (Eq. 9.6.11 on page 266) is replaced by mB , the sum of the ion molalities
assuming complete dissociation. It will now be shown that m defined this way can be used
to evaluate ˙ .
The derivation is like that described in Sec. 9.6.3 for a binary solution of a nonelec-
trolyte. Solving Eq. 10.6.1 for A and taking the differential of A at constant T and p, we
obtain
dA D RTMA .m dmB C mB dm / (10.6.2)
10.6 M EAN I ONIC ACTIVITY C OEFFICIENTS FROM O SMOTIC C OEFFICIENTS 300
From Eq. 10.3.9 on page 292, we obtain

dmB
dB D RT d ln ˙C (10.6.3)
mB
Substitution of these expressions in the Gibbs–Duhem equation nA dA C nB dB D 0,

together with the substitution nA MA D nB =mB , yields
m 1
d ln ˙ D dm C dmB (10.6.4)
mB
Then integration from mB D 0 to any desired molality m0B gives the result
m0B
m 1
Z
0
ln ˙ .mB / D m .m0B / 1C dmB (10.6.5)
0 mB
The right side of this equation is the same expression as derived for ln m;B for a nonelec-
trolyte (Eq. 9.6.20 on page 267).
The integrand of the integral on the right side of Eq. 10.6.5 approaches 1 as mB
approaches zero, making it difficult to evaluate the integral by numerical integration starting
at mB D 0. (This difficulty does not exist when the solute is a nonelectrolyte.) Instead, we
can split the integral into two parts
m0B m00 m0B
m 1 m 1 m 1
Z Z B
Z
dmB D dmB C dmB (10.6.6)
0 mB 0 mB m00
B
mB
where the integration limit m00B is a low molality at which the value of m is available and
at which ˙ can either be measured or estimated from the Debye–Hückel equation.
We next rewrite Eq. 10.6.5 with m0B replaced with m00B :
m00
m 1
Z B
00
ln ˙ .mB / D m .m00B / 1C dmB (10.6.7)
0 mB
By eliminating the integral with an upper limit of m00B from Eqs. 10.6.6 and 10.6.7, we obtain
m0B m0B
m 1 m 1
Z Z
00
dmB D ln ˙ .mB / m .m00B / C1C dmB (10.6.8)
0 mB m00
B
mB
Equation 10.6.5 becomes

m0B
m 1
Z
0
ln ˙ .mB / D m .m0B / m .m00B / C ln 00
˙ .mB / C dmB (10.6.9)
m00
B
mB
The integral on the right side of this equation can easily be evaluated by numerical integra-
tion.
P ROBLEMS 301
PROBLEMS
in Appendix I.
10.1 The mean ionic activity coefficient 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 coefficients 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 coefficients of aqueous NaCl at

298:15 K a
1 1
mB =mol kg m mB =mol kg m
0.1 0.9325 2.0 0.9866
0.2 0.9239 3.0 1.0485
0.3 0.9212 4.0 1.1177
0.5 0.9222 5.0 1.1916
1.0 0.9373 6.0 1.2688
1.5 0.9598
a Ref. [31].
10.2 Rard and Miller5 used published measurements of the freezing points of dilute aqueous so-
lutions of Na2 SO4 to calculate the osmotic coefficients of these solutions at 298:15 K. Use
their values listed in Table 10.2 to evaluate the mean ionic activity coefficient of Na2 SO4 at
298:15 K and a molality of mB D 0:15 mol kg 1 . For the parameter a in the Debye–Hückel
equation (Eq. 10.4.7), use the value a D 3:0 10 10 m.
Table 10.2 Osmotic coefficients of aqueous Na2 SO4

at 298:15 K
1 1
mB =mol kg m mB =mol kg m
0.0126 0.8893 0.0637 0.8111
0.0181 0.8716 0.0730 0.8036
0.0228 0.8607 0.0905 0.7927
0.0272 0.8529 0.0996 0.7887
0.0374 0.8356 0.1188 0.7780
0.0435 0.8294 0.1237 0.7760
0.0542 0.8178 0.1605 0.7616
0.0594 0.8141
4 Ref. [145], Table 9.3. 5 Ref. [141].
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 final 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 final 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.
302
CHAPTER 11 REACTIONS AND OTHER CHEMICAL PROCESSES
11.1 M IXING P ROCESSES 303
mixture of
A B
A and B
Figure 11.1 Initial state (left) and final state (right) of mixing process for liquid sub-
stances 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 final 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 i / (11.1.3)
i
The molar Gibbs energy of mixing is the Gibbs energy of mixing

Pper amount of mix-
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 i / (11.1.4)
i
where xi is the mole fraction of substance i in the final 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 con-
stituents 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
by a superscript “id” as in Gm id (mix). The general definition of an ideal molar mixing
quantity, analogous to Eq. 11.1.5, is

X
Xmid (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)
By combining this relation with Eq. 11.1.4, we find the molar Gibbs energy of mixing to
form an ideal mixture is given by
X
id
Gm (mix) D RT xi ln xi (11.1.8)
i
Since each mole fraction is less than one and the logarithm of a fraction is negative, it
follows that Gmid (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
Smid (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

Hmid (mix) D 0 (11.1.10)
and
Umid (mix) D 0 (11.1.11)
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
Vmid (mix) D 0 (11.1.12)
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
Figure 11.2 on the next page shows how Gm id (mix), TS id (mix), and H id (mix)
m m
depend on the composition of an ideal mixture formed by mixing two pure substances.
Although it is not obvious in the figure, the curves for Gm id (mix) and TS id (mix) have
m
slopes of C1 or 1 at xA D0 and xA D1.
2 Ref.
[113], p. 241.
3 From the fact mentioned on p. 226 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.
TSm
id (mix)
1
quantity / kJ mol 1
Hmid (mix)
0
1
Gm
id (mix)
2
0:0 0:2 0:4 0:6 0:8 1:0
xA
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 defined 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.
An excess molar quantity XmE 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 i
X
Xiid

D xi Xi
i
X X
D xi .Xi Xi / xi .Xiid Xi /
i i
id
D Xm (mix) Xm (mix) (11.1.13)
By substituting expressions for Xmid (mix) from Eqs. 11.1.8–11.1.12 in Eq. 11.1.13, we ob-
tain 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
X
E
Sm D Sm (mix) C R xi ln xi (11.1.15)
i
HmE D Hm (mix) (11.1.16)
UmE D Um (mix) (11.1.17)

VmE 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 coefficients 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
D RT ln i (11.1.20)
@ni
T;p;nj ¤i
To derive Eq. 11.1.20, consider infinitesimal changes in the mixture composition at

constant T and p. From Eq. 11.1.19, we write
X X X
d nGmE D RT

d.ni ln i / D RT ni d ln i C RT .ln i / dni (11.1.21)
i i i
From i D i
C RT ln. i xP
i /, we have di D RT .d ln i C dxi =xi /. Substitution in
the Gibbs–Duhem equation, i xi di D 0, gives
X X
xi d ln i C dxi D 0 (11.1.22)
i i
P P
In Eq. 11.1.22, we set the sum i dxi equal P to zero (because i xi equals 1) and
multiply by the total amount, n, resulting in i ni d ln i D 0. This turns Eq. 11.1.21
into X
d nGmE D RT

.ln i / dni (11.1.23)
i
from which Eq. 11.1.20 follows.
11.1.4 The entropy change to form an ideal gas mixture
P T and p to form an ideal gas mixture, the molar

When pure ideal gases mix at constant
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. De-
spite the fact that the latter process allows the molecules in the two subsystems to inter-
mingle 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 Siı R ln.pi =p ı / (11.1.24)
A B A B
A+B A+B
12 1 2 1 2
(a) (b) (c)
Figure 11.3 Reversible mixing process for ideal gases A and B confined 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 fixed 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 Siı R ln.p=p ı / (11.1.25)
where p is equal to nRT =V .
From Eqs. 11.1.24 Pand 11.1.25, and the fact that the entropy of a mixture is given by the
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 final 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 semiper-
meable 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 final 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 lnŒV2 =V1 (B). Likewise, the reversible
work to expand gas A with piston 2 is the same as if B were absent: nA RT lnŒV2 =V1 (A).
Because the initial and final temperatures and pressures are the same, the mole fractions in
the final 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 fixed temperature.
Let the standard molar internal energy of pure gaseous A be UAı (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 UAı (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 A–A 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 UBı (g) C kBB (11.1.28)
where kBB is for B–B interactions.

We assume that in a liquid mixture of A and B, the numbers of nearest-neighbor mole-
cules of A and B surrounding any given molecule are in proportion to the mole fractions xA
and xB .4 Then the number of A–A interactions is proportional to nA xA , the number of B–B
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.
interactions is proportional to nB xB , and the number of A–B interactions is proportional to

nA xB C nB xA . The internal energy of the liquid mixture is then given by
U (mixt) D nA UAı (g) C nB UBı (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 UB
D nA xA kAA C nB xB kBB C .nA xB C nB xA /kAB nA kAA nB kBB
D nA .xA 1/kAA C nB .xB 1/kBB C .nA xB C nB xA /kAB (11.1.30)
With the identities xA 1 D xB , xB 1 D xA , and nA xB D nB xA D nA nB =n
(where n is the sum nA C nB ), we obtain
nA nB
U (mix) D .2kAB kAA kBB / (11.1.31)
n
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 A–B interaction
equals the average of an A–A interaction and a B–B interaction.
If we write Eq. 11.1.29 in the form
1
U (mixt) D nA UAı (g) C nB UBı (g) C .n2 k C 2nA nB kAB C n2B kBB / (11.1.32)
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
what is called a regular solution.5 The excess molar Gibbs energy of a mixture is Gm E
D
E E E
Um C pVm T Sm . Using the expression of Eq. 11.1.31 with the further assumptions that
VmE and SmE are zero, this model predicts the excess molar Gibbs energy is given by
E U (mix)
Gm D D xA xB .2kAB kAA kBB / (11.1.34)
n
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].
b
1
bc bc
Gm (mix)

bc
b
2
bc
0 xA 1
Figure 11.4 Molar Gibbs energy of mixing as a function of the composition of a

binary liquid mixture with spontaneous phase separation. The inflection points are
indicated by filled 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 Redlich–Kister series given by
E
D xA xB a C b.xA xB / C c.xA xB /2 C

Gm (11.1.35)
where the parameters a; b; c; depend on T and p but not on composition. This function
E E
satisfies a necessary condition for the dependence of Gm on composition: Gm must equal
zero when either xA or xB is zero. 6
E
For many binary liquid systems, the measured dependence of Gm on composition is
reproduced reasonably well by the two-parameter Redlich–Kister series
E
Gm D xA xB Œ a C b.xA xB / (11.1.36)
in which the parameters a and b are adjusted to fit the experimental data. The activity
coefficients in a mixture obeying this equation are found, from Eq. 11.1.20, to be given by
RT ln A D xB2 Œ a C .3 4xB /b RT ln B D xA2 Œ 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 figure is a
plot of Gm (mix) versus xA . It has the form needed to evaluate the quantities .A A /
and .B B / by the variant of the method of intercepts described on page 233. On this
6 Thereason for this condition can be seen by looking at Eq. 11.1.19 on page 306. 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.
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
figure 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 satisfied.
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 separa-
tion, 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
y D y’ C .xA xA’ / (11.1.38)
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“
xA D (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 ’ ’
yD n y C n“ y “ (11.1.40)
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.
“
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 inflection points. The compositions of the two phases are not the compositions at
the inflection 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 Redlich–Kister series (Eq. 11.1.36).
11.2 T HE A DVANCEMENT AND M OLAR R EACTION Q UANTITIES 312
0 1:0
bc bc
bc
3 0:8
3
Gm (mix)=RT 0:1
0:6 2
aA
bc
0:4 1
0:2 2
1 1
0:2
0:3 0
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
xA xA
(a) (b)
Figure 11.5 Binary liquid mixtures at 1 bar. The curves are calculated from the two-
parameter Redlich–Kister 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 com-
position.
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 de-
scribes 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 equa-
tions, or chemical equations, for the conversion of reactants into products. Thus, for the
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 stoichio-
metric 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 specifies neither the initial and final 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 final 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 syn-
thesis 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 five 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)
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 defined
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
five 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 de-
fined by the reaction equation has advanced in the forward direction from specified 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 3 nNH3 D nNH3 ;0 C 2 (11.2.3)
These relations come from the stoichiometry of the reaction as expressed by the stoichio-
metric coefficients 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 find that infinitesimal changes in the amounts
are related to the change of as follows:
dnN2 D d dnH2 D 3 d 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 HN2 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 coefficient 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 (11.2.6)
We identify this coefficient as the partial derivative

@H
r H D (11.2.7)
@ T;p
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.
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 infinitesimal 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 defined 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 C 32 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 coefficients. We can rearrange this equation to
0D aA bB C d D C eE (11.2.9)
In general, the stoichiometric relation for any chemical process is

X
0D i Ai (11.2.10)
i
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 sto-
ichiometric numbers are N2 D 1, H2 D 3, and NH3 D C2. In other words, each
stoichiometric number is the same as the stoichiometric coefficient 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 i (11.2.11)
(closed system)
The infinitesimal change in the amount due to an infinitesimal change in the advancement
is
dni D i 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 (11.2.13)
@T p;fnig @p T;fnig
i
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 coefficient r X is the molar reaction quantity defined by

def
X
r X D i X i (11.2.15)
i
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 finite 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 finite differences be-
tween the final 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:
H (rxn) H.2 / H.1 /

Hm (rxn) D D (11.2.17)
2 1
(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 sys-
tem at constant T and p. The value of r X is in general a function of the independent
variables T , p, and .
!
H
S
0 =mol 1 0 =mol 1
(a) (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 de-
scribed in Sec. 8.3.1. Subscripts for specific 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 con-
stant 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) defined by X= is equal, for any finite 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 figure, 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 318
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 ı . For instance, vap H ı is a
standard molar enthalpy of vaporization (already discussed in Sec. 8.3.3), and r G ı is the
standard molar Gibbs energy of a reaction.
From Eq. 11.2.15, the relation between a standard molar reaction quantity and the stan-
dard molar quantities of the reactants and products at the same temperature is
def
X
r X ı D i Xiı (11.2.18)
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
defined composition and constant pressure p ı .
2. Since the value of a standard molar reaction quantity is independent of , the standard
molar integral and differential quantities are identical (page 317):
Xmı (rxn) D r X ı (11.2.19)
These general concepts will now be applied to some specific chemical processes.
11.3 MOLAR REACTION ENTHALPY
Recall that Hm (rxn) is a molar integral reaction enthalpy

P equal to H (rxn)=, and that
r H is a molar differential reaction enthalpy defined 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
¶q
r H D (11.3.1)
d
(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
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 pro-
cesses 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 ı , is the same as the molar integral reaction en-
thalpy Hm (rxn) for the reaction taking place under standard state conditions (each reactant
and product at unit activity) at constant temperature (page 318).
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 ı , the reaction is exothermic if r H ı is negative and endothermic if r H ı 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 ı , 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 ı ) C 32 H2 (ideal gas, p ı ) C 12 Br2 (l, p ı ) ! CH3 Br(ideal gas, p ı )
The value of f H ı for a given substance depends only on T . By definition, f H ı for the
reference state of an element is zero.
A principle called Hess’s 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
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 ı D 1 bar):
C(s, graphite) C O2 (g) ! CO2 .g/ r H ı D 393:51 kJ mol 1
CO(g) C 21 O2 (g) ! CO2 .g/ r H ı D 282:98 kJ mol 1
(Note that the first reaction, in addition to being the combustion reaction of graphite, is
also the formation reaction of carbon dioxide.) The change resulting from the first reaction
followed by the reverse of the second reaction is the formation reaction of carbon monoxide:
C(s, graphite) C 12 O2 (g) ! CO(g)
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 difficult to prevent the
formation of some CO2 . From Hess’s 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 ı (CO, g, 298:15 K) D . 393:51 C 282:98/ kJ mol 1
D 110:53 kJ mol 1 (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 en-
thalpy r H ı of a reaction using a formula based on Hess’s law. Imagine the reaction to take
place in two steps: First each reactant in its standard state changes to the constituent ele-
ments 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
r H ı D i f H ı .i/ (11.3.3)
i (Hess’s law)
where f H ı .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 Hess’s law the standard molar reaction enthalpy is the sum of the standard mo-
lar 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
coefficient from the reaction equation.
A standard molar enthalpy of formation can be defined for a solute in solution to use in
Eq. 11.3.3. For instance, the formation reaction of aqueous sucrose is
11
12 C(s, graphite) C 11 H2 (g) C 2 O2 (g) ! C12 H22 O11 (aq)
and f H ı 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
BIOGRAPHICAL SKETCH
Germain Henri Hess (1802–1850)
seven editions. He became a full member of
the St. Petersburg Academy Academy in 1834.
Hess published the results of his thermo-
chemical 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 wa-
ter, and another series of measurements of the
heat evolved when the acid in H2 SO4 -water
mixtures is neutralized with aqueous ammo-
nia. The following table from this paper is of
Hess was a Russian chemist and physician historical interest, although it is a bit difficult
whose calorimetric measurements led him to to decipher:
formulate the law of constant heat summation, Acid Heat evolved by Sum
now known as Hess’s law. His given name had ammonia water
several versions: “Germain Henri” in French, P«
H S 595:8 595:8
as shown above; “Hermann Heinrich” in Ger- P2«
H S 518:9 77:8 596:7
man; and “German Iwanowitsch” in Russian. P3«
H S 480:5 116:7 597:2
He was born in Geneva, Switzerland, the P6«
H S 446:2 155:6 601:8
son of an artist. The family moved to Rus- Average 597:9
sia when he was three years old; his father The first 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, « S is SO , and P
H «
S is H SO . Thus
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 sec-
empire) 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 geol-
moiety 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
composition of mineral waters in Russia.a
H2 SO4 .l/ C 2 NH3 .aq/ ! .NH4 /2 SO4 .aq/
Hess was more interested in chemistry than
demonstrates Hess’s law of constant heat sum-
medicine. He briefly studied with the fa-
mation, which he stated in the words:c
mous Swedish chemist Jöns Jakob Berzelius
in Stockholm, and they became life-long The amount of heat evolved during the formation
of a given compound is constant, independent of
friends. Hess then practiced medicine in
whether the compound is formed directly or in-
Irkutsk, Siberia, while continuing his interests
directly in a series of steps.
in mineral chemistry and chemical analysis.
In 1829, after being being elected an adjunct Hess confirmed this law with similar exper-
member of the St. Petersburg Academy of Sci- iments using other acids such as HCl(aq) and
ences, he gave up his medical practice and be- other bases such as NaOH(aq) and CaO(s).
gan teaching chemistry at various institutions After 1848 Hess’s health began to deterio-
of higher learning in St. Petersburg. He wrote rate. He was only 48 when he died. His role
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 influential German
abridgment in 1834 that became the standard physical chemist Wilhelm Ostwald drew atten-
Russian chemistry textbook and went through tion to it about forty years after Hess’s death.
a Refs. [94] and [101]. b Ref. [77]. c Ref. [77]; translation in Ref. [41].
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 infinite-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 ı for substances and nonionic solutes. Aqueous hydrogen ion is the usual reference
ion, to which is assigned the arbitrary value
f H ı (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 12 Cl2 .g/ ! HC .aq/ C Cl .aq/
The standard molar reaction enthalpy at 298:15 K for this reaction is known, from reaction
calorimetry, to have the value r H ı 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 ı , 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 coefficient on a molality basis were 1. From Eq. 11.3.3, we
equate the value of r H ı to the sum
ı ı
1
2 f H (H2 , g)
1
2 f H (Cl2 , g) C f H ı (HC , aq) C f H ı (Cl , aq)
But the first three terms of this sum are zero. Therefore, the value of f H ı (Cl , aq) is
167:08 kJ mol 1 .
Next we can combine this value of f H ı (Cl , aq) with the measured standard molar
enthalpy of formation of aqueous sodium chloride
Na.s/ C 21 Cl2 .g/ ! NaC .aq/ C Cl .aq/
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 ı values of
various ions in aqueous solution.
11.3.3 Molar reaction heat capacity
Preaction enthalpy r H is in general a function of T , p, and . Using the relations

The molar
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
P
@r H @ i i Hi X
D D i Cp;i D r Cp (11.3.5)
@T p; @T 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 ı = dT D r Cpı (11.3.6)
7 This procedure is similar to that described on page 234 for partial molar volumes of ions.
H.1 ; T 00 /
H (rxn, T 00 )
H.1 ; T 0 /
H
H (rxn, T 0 )
H.2 ; T 00 /
H.2 ; T 0 /
T
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 finite 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 find 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 143: Cp D .@H=@T /p; . We integrate dH D Cp dT from T 0 to T 00 at
constant p and , for both the final and initial values of the advancement:
Z T 00
00 0
H.2 ; T / D H.2 ; T / C Cp .2 / dT (11.3.7)
T0
Z T 00
00 0
H.1 ; T / D H.1 ; T / C Cp .1 / dT (11.3.8)
T0
Subtracting Eq. 11.3.8 from Eq. 11.3.7, we obtain
Z T 00
H (rxn, T 00 ) D H (rxn, T 0 ) C Cp dT (11.3.9)
T0
where Cp is the difference between the heat capacities of the system at the final 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 figure, and
the product of Cp and the temperature difference T 00 T 0 equals the change in the value
of H (rxn). The figure 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 324
sol dil
B(s) (A + B)(l) A(l) (A + B)(l)
(a) (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 find 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
Z T 00
00 0
r H.T ; / D r H.T ; / C r Cp .T; / dT (11.3.11)
T0
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:
@H
sol H D (11.4.1)
@sol 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 infinite 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 infinitely 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
sol H 1 D HB1 HB (11.4.3)
Note that because the values of HB1 and HB are independent of the solution composition,
the molar differential and integral enthalpies of solution at infinite dilution are the same.
An integral enthalpy of solution, H(sol), is the enthalpy change for a process in
which a finite amount sol of solute is transferred from a pure solute phase to a specified
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 inte-
gral enthalpy of solution which this book will denote by Hm (sol, mB ), where mB is the
molality of the solution formed:
H(sol)
Hm (sol, mB ) D (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 fixed 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 defined 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:
dH(sol)
sol H D (11.4.5)
dsol
(constant T , p, and nA )
20
40
H.sol/=kJ
. 60
a
80
bc
100
c b
120
0 5 10 15 20
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 satura-
tion 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. [165], page 2-315.
The slope of the curve at sol D0 is sol H 1 , the molar enthalpy of solution at infinite
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 ı , because the standard molar enthalpy of a
solute is the molar enthalpy at pDp ı and infinite 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

P
i i X i becomes
dil H D HA HA (11.4.7)
where HA is the partial molar enthalpy of the solvent in the solution. In the limit of infinite
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 finite amount of solvent from a pure solvent phase to a solution, T and p being the
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 m0B to m00B , this book will use the notation
Hm .dil, m0B !m00B /:
H (dil)
Hm .dil, m0B !m00B / D (11.4.8)
nB
The value of Hm .dil, m0B !m00B / at a given T and p depends only on the initial and final
molalities m0B and m00B .
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 m00B from pure solvent and solute phases. Both paths are at constant T
and p in a closed system.
Path 1: The solution forms directly by dissolution of the solute in the solvent. The en-
thalpy change is nB Hm .sol, m00B /, where the molality of the solution is indicated in
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 m0B (more concentrated than m00B ). The
enthalpy change is nB Hm .sol, m0B /. In a second step of this path, the remaining pure
solvent mixes with the solution to dilute it from m0B to m00B . The enthalpy change of the
second step is nB Hm .dil, m0B !m00B /.
Since both paths have the same initial states and the same final states, both have the same
overall enthalpy change:
nB Hm .sol, m00B / D nB Hm .sol, m0B / C nB Hm .dil, m0B !m00B / (11.4.9)
or
Hm .sol, m00B / D Hm .sol, m0B / C Hm .dil, m0B !m00B / (11.4.10)
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 21 O2 (g) C 12 H2 (g) C 50 H2 O(l) ! NaOH in 50 H2 O
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, m0B !m00B / D f H (B, m00B ) f H (B, m0B ) (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 defined 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 HB /

D nA LA C nB sol H (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 ŒHm (sol, mB ) sol H (11.4.16)
Thus LA depends on the difference between the molar integral and differential enthalpies
of solution.
The relative partial molar enthalpy of a solute is defined by
def
LB D HB HB1 (11.4.17)
The reference state for the solute is the solute at infinite dilution. To relate LB to molar
enthalpies of solution, we write the identity
LB D HB HB1 D .HB HB / .HB1 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 solu-
tion at the molality of interest and at infinite dilution.
For a solution of a given molality, LA and LB can be evaluated from calorimetric mea-
surements 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 231. 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 inter-
cept 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 finite limiting slope at mB D0 whose value for
an electrolyte can be predicted from the Debye–Hückel limiting law, providing a use-
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 fixed amount of solvent. From Eq. 11.4.5, the molar differential enthalpy of solution is
given by sol H D dH(sol)= dsol 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) fitted to the points, and slopes and intercepts can then
be evaluated by numerical methods.
H(sol) D sol Hm (sol, mB ) and take the derivative of the expression with respect to sol :
d Œsol Hm (sol, mB )
sol H D
dsol
dHm (sol, mB )
D Hm (sol, mB ) C sol (11.4.20)
dsol
At constant nA , mB is proportional to sol , so that dsol =sol can be replaced by dmB =mB .
When we combine the resulting expression for sol H with Eq. 11.4.19, we get the follow-
ing expression for the relative partial molar enthalpy of the solute:
dHm (sol, mB )
LB D Hm (sol, mB ) C mB sol H 1 (11.4.21)
dmB
It is convenient to define 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 indepen-
dent of mB , the derivative d˚L = dmB is equal to dHm (sol, mB )= dmB . We can therefore
write Eq. 11.4.21 in the compact form
d˚L
LB D ˚L C mB (11.4.23)
dmB
(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 finite and can be
predicted from the Debye–Hückel limiting law. Accordingly, a satisfactory procedure is to
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 .m00B / ˚L .m0B / D Hm .dil, m0B !m00B / (11.4.25)
We can measure the enthalpy changes for diluting a solution of initial molality m0B to various
p
molalities m00B , plot the values of Hm .dil, m0B !m00B / versus mB , extrapolate the curve to
p
m D0, and shift the origin to the extrapolated intercept, resulting in a plot of ˚L versus
p B
mB .
5:0 1:0 0:5

4:5 0
0:5
1
4:0 0:5
Hm .sol/=kJ mol
1
0 1:0
L =kJ mol
LB =kJ mol
3:5 0:5 1:5
3:0 1:0 2:0
2:5 1:5 2:5
2:0 0 3:0 0 1:0 2:0 3:0 4:0
q
0.5 1.0 1.5 2.0
mB =mol kg 1
mB =mol kg 1
(a) (b)
Figure 11.10 Thermal properties of aqueous NaCl at 25:00 ı 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 from molar enthalpy of formation values in Ref. [165], p. 2-301.
b Based on data in Ref. [126], Table X.
In order to be able to use Eq. 11.4.23, we need to relate the derivative d˚L = dmB to the
p
slope of the curve of ˚L versus mB . We write
p 1 p p
d mB D p dmB dmB D 2 mB d mB (11.4.26)
2 mB
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 d˚L
LB D ˚L C p (11.4.27)
2 d mB
(constant T and p)
The value of ˚L goes to zero at infinite dilution. When the solute is an electrolyte, the
dependence of ˚L on mB in solutions dilute enough for the Debye–Hückel limiting law to
apply is given by
p
˚L D C˚L mB (11.4.28)
(very dilute solution)
For aqueous solutions of a 1:1 electrolyte at 25 ı C, the coefficient C˚L has the value10
C˚L 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.
p p
C˚L is equal to the limiting slope of ˚L versus mB , of Hm (sol, mB ) versus mB ,
and of Hm .dil, m0B !m00B / versus m0B . The value given by Eq. 11.4.29 can be used for
p
extrapolation of measurements at 25 ı C and low molality to infinite dilution.
Equation 11.4.28 can be derived as follows. For simplicity, we assume the pressure
is the standard pressure p ı . At this pressure HB1 is the same as HBı , and Eq. 11.4.17
becomes LB D HB HBı . From Eqs. 12.1.3 and 12.1.6 in the next chapter, we can
write the relations
d.ım;B =T /

@.B =T /
HB D T 2 HBı D T 2 (11.4.30)
@T p;fnig dT
Subtracting the second of these relations from the first, we obtain

@.B ım;B /=T

HB HBı D T 2 (11.4.31)
@T p;fnig
The solute activity on a molality basis, am;B , is defined by B ım;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 =mı / . Accordingly, the relative partial molar enthalpy
of the solute is related to the mean ionic activity coefficient by

@ ln ˙
LB D RT 2 (11.4.32)
@T p;fnig
We assume the solution is sufficiently dilute for the mean ionic activity coefficient
p described by the Debye–Hückel limiting law, Eq. 10.4.8: ln ˙ D
to be adequately
ADH jzC z j Im , where ADH is a temperature-dependent quantity defined on page
294. Then Eq. 11.4.32 becomes

@ADH
LB D RT 2 jzC z j Im
p
(11.4.33)
@T p;fnig
Substitution of the expression given by Eq. 10.4.9 on page 297 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 jzC z j mB
2 @T p;fnig
p
D CLB mB (11.4.34)
The coefficient 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 C˚L represent the limiting slope of ˚L versus mB . In a very dilute solution
p
we have ˚L D C˚L mB , and Eq. 11.4.27 becomes
p p
mB d˚L p mB
LB D ˚L C p D C˚L mB C C˚L (11.4.35)
2 d mB 2
By equating this expression for LB with the one given by Eq. 11.4.34 and solving for
p
C˚L , we obtain C˚L D .2=3/CLB and ˚L D .2=3/CLB mB .
11.5 R EACTION C ALORIMETRY 333
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 sur-
rounded by an outer jacket. The jacket may be of either the adiabatic type or the isothermal-
jacket 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, in-
stead 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 final 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 the advancement changes. For an
exothermic reaction occurring in an adiabatic calorimeter, the heating curve may resemble
that shown in Fig. 7.3 on page 169, and the heating curve in an isothermal-jacket calorimeter
may resemble that shown in Fig. 7.4 on page 171. 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 final 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 eval-
uate 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 170 with wel
set equal to zero), where is the energy equivalent of the calorimeter, r is the slope of the
P.T2 /
H (expt) H.P/
T
R.T1 / P.T1 /
H.rxn, T1 /
Figure 11.11 Enthalpy changes for paths at constant pressure (schematic). R denotes
reactants and P denotes products.
heating curve when no reaction is occurring, and t1 and t2 are the times at temperatures T1
and T2 . For an isothermal-jacket calorimeter, we evaluate H (expt) using Eq. 7.3.28 on
page 172 with wel set equal to zero.
The enthalpy change we wish to find 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 figure have the same initial state and different
final states. The path connecting these two final states is for a change of the temperature
from T1 to T2 with fixed at its final 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 /=

P .T2 T1 / C H (expt)
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, reflecting
the fact that after the reaction takes place in the calorimeter, heat would have to leave the
11 The symbol P refers to the final equilibrium state in which the reaction vessel contains products of the reaction
and any excess reactants.
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.
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). Correc-
tions, usually small, are needed to obtain the standard molar reaction enthalpy r H ı 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 ı ,
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 ı C.) With careful work, using temper-
ature measurements with a resolution of 1 10 4 K or better and detailed corrections, the
precision of c H ı can be of the order of 0:01 percent.
Bomb calorimetry is the principal means by which standard molar enthalpies of com-
bustion of individual elements and of compounds of these elements are evaluated. From
these values, using Hess’s 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 com-
pounds, the standard molar reaction enthalpies of innumerable reactions can be calculated
with Hess’s law (Eq. 11.3.3 on page 320).
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 five 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 final 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 ı .Tref /, is calculated. This
calculation is called reduction to standard states.
5. The standard molar enthalpy of combustion, c H ı .Tref /, is calculated.
These five steps are described below.
ignition
wires stirrer
jacket
calorimeter wall
oxygen inlet
platinum resistance
thermometer air or vacuum
sample holder 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 thick-
walled 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 sub-
stance to be combusted is placed in a metal sample holder. The calculations are simplified
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 confined
in the sample holder by cellulose tape of known combustion properties. If one of the com-
bustion 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 reac-
tion in the bomb vessel. If the reaction is exothermic, the measured temperature rapidly
R.T2 / P.T2 /
U.IBP; T2 /
U.R/ U (expt)
T
R.T1 /

Figure 11.13 Internal energy changes for paths at constant volume in a bomb calo-
rimeter (schematic). R denotes reactants and P denotes products.
increases over the course of several minutes. 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 final time t2 , with final tem-
perature 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 in-
side 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 334, 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 figure. Conceptually, the
overall change of state during this process would be duplicated by a path in which the tem-
perature of the system with the reactants present increases from T1 to T2 ,12 followed by the
isothermal bomb process at temperature T2 . In the figure 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 reac-
tants 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.
the temperature from T1 to T2 . From the first 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 final 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 find 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 mea-
surement 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 mea-
surement. 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 modified version of the Kirchhoff equation (Eq. 11.3.10 on page 323):
U.IBP; Tref / D U.IBP; T2 / C Œ CV (P) CV (R) .Tref T2 / (11.5.8)
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 ı .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 find 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 final 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 ı .cmb; Tref / D U.IBP; Tref / C (Washburn corrections) i r U ı .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
U ı .cmb; Tref /
c U ı .Tref / D (11.5.10)
c
where c is the advancement of the main combustion reaction in the bomb vessel.
Standard molar enthalpy change

The quantity c U ı .Tref / is the molar internal energy change for the main combustion reac-
tion carried out at constant temperature Tref withPeach reactant and product in its standard
state at pressure p ı . 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 ı .Tref / D c U ı .Tref / C p ı i Viı (11.5.11)
i
13 Ref. [170].
Molar volumes of condensed phases are much smaller than those of gases, and to a good
approximation we may write
X g
c H ı .Tref / D c U ı .Tref / C p ı i Viı (g) (11.5.12)
i
where the sum includes only gaseous reactants and products of the main combustion reac-
tion. Since a gas in its standard state is an ideal gas with molar volume equal to RT =p ı ,
the final relation is X g
c H ı .Tref / D c U ı .Tref / C i RTref (11.5.13)
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 final
state of the isothermal bomb process to their standard states.
For example, consider the complete combustion of a solid or liquid compound of car-
bon, 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 pro-
cess, 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 final 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 ı D 1 bar.
We can calculate the amount of each substance in each phase, in both the initial state
and final state of the isothermal bomb process, from the following information: the inter-
nal 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 Henry’s law constants) of O2 and CO2 in the water; and the
stoichiometry of the combustion reaction. Problem 11.7 on page 361 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 briefly
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 341
constitute a univariant subsystem that at constant pressure is at the fixed 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 en-
thalpy, 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 ice–water transition at
0 ı C. The solid–liquid transition of diphenyl ether has a relatively large volume change and
is useful for measurements at 26:9 ı C. Phase-transition calorimeters are especially useful
for slow reactions.
A heat-flow calorimeter is a variation of an isothermal-jacket calorimeter. It uses a ther-
mopile (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 propor-
tional 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 flame calorimeter is a flow system in which oxygen, fluorine, or another gaseous
oxidant reacts with a gaseous fuel. The heat transfer between the flow tube and a heat sink
can be measured with a thermopile, as in a heat-flow calorimeter.
11.6 ADIABATIC FLAME TEMPERATURE
With a few simple approximations, we can estimate the temperature of a flame formed in
a flowing 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 334. 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
Z T2
c H ı .T1 / C Cp .P/ dT D 0 (11.6.1)
T1
where c H ı .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 satisfies Eq. 11.6.1 is the estimated flame temperature. Problem
11.9 presents an application of this calculation. Several factors cause the actual temperature
in a flame to be lower: the process is never completely adiabatic, and in the high temperature
of the flame 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 342
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 definition 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 definition of the molar reaction Gibbs energy or molar Gibbs energy of reaction, r G:
def
X
r G D i i (11.7.1)
i
Equation 11.2.16 shows that this quantity is also given by the partial derivative

@G
r G D (11.7.2)
@ T;p
(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 145). 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 affinity of reaction, defined as the negative of r G. increases spontaneously if the affinity is positive and
decreases spontaneously if the affinity is negative; the system is at equilibrium when the affinity is zero.
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 i i D 0 (11.7.4)
i (reaction equilibrium)
It is important to realize that this condition is independent of whether or not reaction

equilibrium is approached at constant temperature andP pressure. It is a universal criterion
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 236:
0 0 0 0 X 0 0
dU D T ’ dS ’ p ’ dV ’ C ’i dn’i
i
!
X X
C T ’ dS ’ p ’ dV ’ C ’i dn’i (11.7.5)
’¤’0 i
The conditions of isolation are

’0
X
dV C dV ’ D 0 (no expansion work) (11.7.7)
’¤’0
For each species i:

0
X
dn’i C dn’i D i d (closed system) (11.7.8)
’¤’0
00
In Eq. 11.7.8, dn’i 0 should be set equal to zero for a species i 0 that is excluded from phase
’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 0
We use these conditions of isolation to substitute for dU , dV ’ , and dn’i in Eq. 11.7.5,
0
and make the further substitution dS ’ D dS ’
P
’¤’0 dS . Solving for dS, we obtain
X .T ’0 T ’ / X .p ’0 p ’ /
dS D ’0
dS ’ ’0
dV ’
0
T T
’¤’ ’¤’0
0 ’0
X X .’ ’ /
P
i i ’ i i i
C 0 dni 0 d (11.7.9)
0
T’ T’
i ’¤’
The equilibrium condition is that the coefficient 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 containingP that species, is equal to the
0
chemical potential of the species in phase ’ ; and the quantity i i ’i (which is equal to
0
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), in-
volves 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.
G G G

(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 .
! G
bc
0 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 317.)
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 sponta-
neous reaction occurs in a closedPsystem at constant temperature and volume. The system
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 141: X
dA D S dT p dV C i dni (11.7.10)
i
When we make the substitution dni D i d, we obtain an expression for the total differen-
tial of A in a closed system with a chemical reaction:

!
X
dA D S dT p dV C i i d (11.7.11)
i
We identify the coefficient of the last term on the right as a partial derivative:

X @A
i i D (11.7.12)
@ T;V
i
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 an-
other way to express the criterion for spontaneity stated on page 145: 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 equilib-
rium 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 ıi (g) C RT ln.pi =p ı /
(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

ı
G./ D ni i (g) C RT ln ı (11.7.14)
p
i
At D 0, the amounts and partial pressures have their initial values ni;0 and pi;0 :
pi;0
X
ı
G.0/ D ni;0 i (g) C RT ln ı (11.7.15)
p
i
The difference between these two expressions is

X
G./ G.0/ D .ni ni;0 /ıi (g)
i
X pi X pi;0
C RT ni ln RT ni;0 ln (11.7.16)
pı pı
i i
Converting partial pressures to mole fractions with pi D yi p and pi;0 D yi;0 p gives
X X
G./ G.0/ D .ni ni;0 /ıi (g) C RT ni ln yi
i i
X X p
RT ni;0 ln yi;0 C RT .ni ni;0 / ln (11.7.17)
pı
i i
With the substitution ni ni;0 D i (Eq. 11.2.11) in the first and last terms on the right
side of Eq. 11.7.17, the result is
X X
G./ G.0/ D i ıi (g) C RT ni ln yi
i i
!
X X p
RT ni:0 ln yi;0 C RT i ln (11.7.18)
pı
i i
The sum i i ıi (g) in the first term on the right side of Eq. 11.7.18 is r G ı , the
P
standard molar reaction Gibbs energy. Making this substitution gives finally
X X
G./ G.0/ D r G ı C RT ni ln yi RT ni;0 ln yi;0
i i
!
X p
C RT i ln (11.7.19)
pı
i (ideal gas mixture)
There are four terms on the right side of Eq. 11.7.19. The first term is the Gibbs en-
ergy 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 first and last terms are proportional to the advancement and cannot bePthe cause of a
minimum in the curve of the plot of G versus . It is the mixing termP RT i ni ln yi that
is responsible for the observed minimum.16 This term divided by nD i ni is Gm id (mix),
the molar differential Gibbs energy of mixing to form an ideal mixture (see Eq. 11.1.8 on
page 304); the term is also equal to nTSmid (mix) (Eq. 11.1.9), showing that the minimum
is entirely an entropy effect.
Now let us consider specifically 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 ı 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 l’Hospital’s rule (see any calculus textbook).
G G G
bc bc bc
0 1 0 1 0 1
=nA;0
=nA;0 =nA;0
(a) (b) (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 ı < 0. (b) r G ı D 0. (c) r G ı > 0.
The second term on the right side is nGm id (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 .
Since the curve of Gm id (mix) plotted against has a minimum (as shown in Fig. 11.2 on
page 305), G./ G.0/ also has a minimum.

Figure 11.16 illustrates how the position of the minimum, which is the position of re-
action equilibrium, depends on the value of r G ı . The more negative is r G ı , the closer
to the product side of the reaction is the equilibrium position. On the other hand, the more
positive is r G ı , 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 figure 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 filled 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 satisfied 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 349
1:4
D nA,0
1:2
bc
1:0
p=p Æ b
0:8
D0
0:6
0:4
0:2
0 0:5 1:0 1:5 2:0 2:5 3:0
V =V0
Figure 11.17 Dependence of Gibbs energy on volume and pressure, at constant tem-
perature, in a closed system containing an ideal gas mixture of A and B. The reaction is
A ! 2 B with r G ı 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 ı . Filled circle: position
of minimum G for a constant volume of 1:41V0 , where V0 is the initial volume at
pressure p ı .
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 definition of K

Equation 10.1.9 gives the general relation between the chemical potential i and the activity
ai of species i in a phase of electric potential :
i D ıi C RT ln ai C zi F (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 i gives
X X X
r G D i ıi C RT i ln ai C F 0 i zi (11.8.2)
i i i (all ions at D 0 )
The first 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
X
r G ı D i ıi (11.8.3)
i
Since the standard chemical potential ıi of each species i is a function only of T , the value
of r G ı for a given reaction as defined 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 ı C RT i ln ai (11.8.4)
i (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
P are all in the same phase, or in phases of equal electric potential, the
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
i ln ai D ln ai i D ln ai i (11.8.5)
i i i
Q i
The product i ai is called the reaction quotient or activity quotient, Qrxn :
def
Y
Qrxn D ai i (11.8.6)
i
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 coefficient 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.
18 The
ai i is the product
Q Q
symbol stands for a continued product. If, for instance, there are three species, i

.a11 /.a22 /.a33 /.
For instance, for the ammonia synthesis reaction N2 (g) + 3 H2 (g) ! 2 NH3 (g) the reac-
tion quotient is given by
2
aNH 3
Qrxn D 3
(11.8.7)
aN2 aH 2
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 ı C RT ln Qrxn (11.8.8)
The value of Qrxn under equilibrium conditions is the thermodynamic equilibrium con-
stant, K. The general definition of K is
def
Y
K D .ai /eqi (11.8.9)
i
where the subscript eq indicates an equilibrium state. Note that K, like Qrxn , is dimension-
less.
The IUPAC Green Book19 gives K as an alternative symbol for the thermodynamic
equilibrium constant, the appended superscript denoting “standard.” An IUPAC Com-
mission on Thermodynamics20 has furthermore recommended the name “standard
equilibrium constant,” apparently because its value depends on the choice of stan-
dard states. Using this alternative symbol and name could cause confusion, since the
quantity defined 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 im-
portant relation between the standard molar reaction Gibbs energy and the thermodynamic
equilibrium constant:
r G ı D RT ln K (11.8.10)
We can solve this equation for K to obtain the equivalent relation
r G ı

K D exp (11.8.11)
RT
We have seen that the value of r G ı 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 ı is positive and greater than 1 if r G ı is negative. At a fixed
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. [36], p. 58.
20 Ref. [52].
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 348.
Equation 11.8.10 correctly relates r G ı 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 ı , 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 ı (Sec. 9.3.3). In this case,
each activity is given by ai (g) D fi =p ı D i pi =p ı where i is a fugacity coefficient
(Table 9.5). When we substitute this expression into Eq. 11.8.9, we find we can express the
thermodynamic equilibrium constant as the product of three factors:
" #" #
h i

Y Y P
KD .i /eqi .pi /eqi .p ı / i i (11.8.12)
i i (gas mixture)
On the right side of this equation, the first factor is the proper quotient of fugacity coef-
ficients 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 ı needed to make K
dimensionless.
The proper quotient of equilibrium partial pressures is an equilibrium constant on a
pressure basis, Kp :

Y
Kp D .pi /eqi (11.8.13)
i (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 equi-
librium 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 fu-
gacity coefficient 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.
As an example of the difference between K and KpP , consider again the ammonia syn-
thesis 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
!
NH
KD 3
Kp .p ı /2 (11.8.14)
N2 H3 2
eq
where Kp is given by
2
!
pNH 3
Kp D (11.8.15)
pN2 pH3 2
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 infinite dilution, each
solute activity coefficient is unity, and at the standard pressure each pressure factor is unity.
Under these conditions of infinite 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 ı am;B D mB =mı (11.8.16)
In the limit of infinite dilution, the solute composition variables approach values given by
the relations in Eq. 9.1.14 on page 225: 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 find that the activities of
solute B at infinite dilution and pressure p ı are related by
ax;B D VA c ı ac;B D MA mı am;B (11.8.18)

Q
The expression K D i .ai /eqi has a factor .aB /eqB for each solute B that is a reactant
or product. From Eq. 11.8.18, we see that for solutes at infinite dilution at pressure p ı , the
relations between the values of K based on different solute standard states are
Y Y
K(x basis) D .VA c ı /B K(c basis) D .MA mı /B K(m basis) (11.8.19)
B 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.
11.8.4 Evaluation of K
The relation K D exp . r G ı =RT / (Eq. 11.8.11) gives us a way to evaluate the thermo-
dynamic equilibrium constant K of a reaction at a given temperature from the value of the
standard molar reaction Gibbs energy r G ı at that temperature. If we know the value of
r G ı , we can calculate the value of K.
One method is to calculate r G ı from values of the standard molar Gibbs energy of
formation f G ı 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 ı to measurable
P
quantities, wePmake the substitution
P i D Hi T Si (Eq. 9.2.46) in r G D i i i to
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 ı D r H ı Tr S ı (11.8.21)
where the standard molar reaction entropy is given by
X
r S ı D i Siı (11.8.22)
i
If the reaction is the formation reaction of a substance, we have

X
f G ı D f H ı T i Siı (11.8.23)
i
where the sum over i is for the reactants and product of the formation reaction. We can eval-
uate 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 ı for substances and ions. An abbrevi-
ated 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 ı , and then K, from the expression
(analogous to Hess’s law) X
r G ı D i f G ı .i/ (11.8.24)
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 Siı
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 first half of the 20th century.21
For ions in aqueous solution, the values of Smı and f G ı found in Appendix H are based
on the reference values Smı D0 and f G ı D0 for HC (aq) at all temperatures, similar to the
convention for f H ı 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.
22 Note ı in Appendix H for some ions, unlike the values for substances, are negative; this
that the values of Sm
simply means that the standard molar entropies of these ions are less than that of HC (aq).
as reactants or products, these values correctly give r S ı using Eq. 11.8.22, or r G ı 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 ı , f H ı and Smı is
more complicated.
Consider first 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 ı , Smı , and f G ı are zero for the
aqueous HC ion and the fact that f H ı and f G ı are zero for the elements, we can
write the following expressions for standard molar reaction quantities:
r H ı D f H ı .MzC / (11.8.25)
X
r S ı D .zC =2/Smı .H2 / C Smı .MzC / Siı (11.8.26)
elements
r G ı D f G ı .MzC / (11.8.27)
Then, from r G ı D r H ı Tr S ı , we find
f G ı .MzC / D f H ı .MzC /
" #
X
T Smı .MzC / Siı C .zC =2/Smı .H2 / (11.8.28)
elements
For example, the standard molar Gibbs energy of the aqueous mercury(I) ion is found
from
f G ı .Hg2 2C / D f H ı .Hg2 2C / T Smı .Hg2 2C /
2
C 2T Smı .Hg/ 2
T Smı .H2 / (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 find by the same method
f G ı .Xz / D f H ı .Xz /
" #
X
T Smı .Xz / Siı jz =2j Smı .H2 / (11.8.30)
elements
For example, the calculation for the nitrate ion is
f G ı .NO3 / D f H ı .NO3 / T Smı .NO3 /

C 12 T Smı .N2 / C 32 T Smı .O2 / C 21 T Smı .H2 / (11.8.31)
11.9 E FFECTS OF T EMPERATURE AND P RESSURE ON E QUILIBRIUM P OSITION 356
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 specified initial values of these amounts. We
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 345. 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 inde-
pendent variables
dG D S dT C V dp C r G d (11.9.1)
and obtain the reciprocity relations

@r G @S @r G @V
D D (11.9.2)
@T p; @ T;p @p T; @ 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; @p T;
We use these expressions for two of the coefficients 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 coefficient
.@r G=@/T;p is the partial second derivative of G with respect to :
2
@r G @ G
D (11.9.5)
@ T;p @ 2 T;p
We know that at a fixed 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 345). 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 deq , we obtain
2
r H @ G
0D dT C r V dp C d (11.9.6)
T @ 2 T;p eq
(closed system)
which shows how infinitesimal changes in T , p, and eq are related.

Now we are ready to see how eq is affected by changes in T or p. Solving Eq. 11.9.6
for deq gives
r H
dT r V dp
deq D T (11.9.7)
@2 G

(closed system)
@ 2 T;p
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
T (closed system)
@ 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 Châtelier’s principle: When a change is made to a closed system at equilib-
rium, the equilibrium shifts in the direction that tends to oppose the change. Here are two
examples.
1. Suppose r H is negative—the 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 con-
stant 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 Châtelier’s principle. Consider the solution
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 326. The equilibrium position (saturation) with one kilogram of water is at
sol 15 mol, indicated in the figure 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 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 Châtelier’s principle.23
Another kind of change for which Le Châtelier’s principle gives an incorrect prediction
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 require-
ment that Qrxn must equal K or, equivalently, that r G must be zero.
23 Ref. [20].
P ROBLEMS 359
PROBLEMS
in Appendix I.
11.1 Use values of f H ı and f G ı 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 54 O2 .g/ C 21 H2 O.l/ ! HC .aq/ C NO3 .aq/
11.2 In 1982, the International Union of Pure and Applied Chemistry recommended that the value
of the standard pressure p ı 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 Hmı , Smı , and Gmı 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 approxima-
tions.
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) r H ı D 55:82 kJ mol 1
Na(s) C H2 O(l) ! NaC (aq/ C OH (aq) C 12 H2 (g) r H ı D 184:52 kJ mol 1

NaOH(s) ! NaOH(aq) sol H 1 D 44:75 kJ mol 1
NaOH in 5 H2 O ! NaOH in 1 H2 O Hm (dil) D 4:93 kJ mol 1
NaOH(s) f H ı D 425:61 kJ mol 1
Using only these values, calculate:

(a) f H ı 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 360
Table 11.1 Data for Problem 11.5 a

1 1
Substance f H /kJ mol M /g mol
H2 O(l) 285:830 18:0153
Na2 S2 O3 5H2 O.s/ 2607:93 248:1828
Na2 S2 O3 in 50 H2 O 1135:914
Na2 S2 O3 in 100 H2 O 1133:822
Na2 S2 O3 in 200 H2 O 1132:236
Na2 S2 O3 in 300 H2 O 1131:780
a Ref. [165], 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 in-

tegral enthalpies of dilution of aqueous HCl
(m0B D 3:337 mol kg 1 ) at 25 ı C.a
m00B =mol kg 1
Hm .dil, m0B !m00B /=kJ mol 1
0:295 2:883
0:225 2:999
0:199 3:041
0:147 3:143
0:113 3:217
0:0716 3:325
0:0544 3:381
0:0497 3:412
0:0368 3:466
0:0179 3:574
0:0128 3:621
a Ref. [155].
11.6 Use the experimental data in Table 11.2 to evaluate LA and LB at 25 ı C for an aqueous HCl
solution of molality mB D 0:0900 mol kg 1 .
P ROBLEMS 361
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 coefficient . . . . . . . . . . . . . ˛ 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 coefficient . . . . . . . . . . . . . ˛ D 2:59 10 4 K 1
standard molar energy of vaporization . . . vap U ı D 41:53 kJ mol 1
Second virial coefficients, 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
Henry’s 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:
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VB1 D 31 cm3 mol 1
CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VC1 D 33 cm3 mol 1
Standard molar energies of solution (solvent = H2 O):
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . sol U ı D 9:7 kJ mol 1
CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . sol U ı 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 eval-
uate 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
19
C6 H14 .l/ C 2 2
O .g/ ! 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 final states of the isothermal bomb
process. The temperature is the reference temperature of 298:15 K.
P ROBLEMS 362
(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 find 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 find the amount of O2 present, just
the change.) Also find the final 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, find 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 coefficient
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 find 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 find 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 244 to calculate the second virial coefficient. Then solve the equation of
state of Eq. 9.3.21 on page 244 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 final gas phases, in part (g) you
use gas equations of state to evaluate the fugacity coefficients of the H2 O (as well as of
the O2 and CO2 ), and then in part (h) you find the amounts of H2 O in the initial and final
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 185 to es-
timate 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 liquid–gas system is
discussed in Sec. 12.8.1; use Eq. 12.8.3 on page 400 to find 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 coefficients 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 185 to calculate O2 . The other calculations require Eq. 9.3.29 on page 245, 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 363
Use the fugacity coefficient and partial pressure of O2 to evaluate its fugacity in states 1
and 2; likewise, find 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 coefficient of a constituent of a gas mixture,
you can calculate the partial pressure with Eq. 9.3.17 on page 243, then the mole fraction
with yi D pi =p, and finally the amount with ni D yi n. Use this method to find 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 Henry’s 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 408 to find 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 final 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 pure-
liquid 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 ı . Using the
value of vap U ı for H2 O given in Table 11.3, calculate U for the vaporization of liquid
H2 O at pressure p ı 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 ı .
(k) Next, consider the dissolved O2 and CO2 , for which gas standard states are used. Assume
that the solutions are sufficiently dilute to have infinite-dilution behavior; then the partial
molar internal energy of either solute in the solution at the standard pressure p ı 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 ı 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 ı 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 ı , 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 ı to p1 and the aqueous solution of state 2 is
decompressed from p2 to p ı . Use an approximate expression from Table 7.4, and treat
the cubic expansion coefficient of the aqueous solutions as being the same as that of pure
water.
(m) The final 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 Umı ) and ending at pressure p1 to form
P ROBLEMS 364
the gas phase present in state 1. Use the approximate expression for Um Umı (g) in Table
7.5 to calculate U D U.p1 / nUmı (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 P of the internal energy of the gas mixture is
given, to a good approximation, by U D i Uiı (g) npT dB= dT , where B is the second
virial coefficient of the gas mixture; this expression is the analogy for a gas mixture of the
approximate expression for Um Umı (g) in Table 7.5. Calculate the value of dB= dT for
the mixture of O2 and CO2 in state 2 (you need P Eq. 9.3.23 on page 244 and the values of
dBij = dT in Table 11.3) and evaluate U D i ni Uiı (g) U.p2 / for the gas expansion.
(n) Add the internal energy changes you calculated in parts (j)–(m) to find 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 n-
hexane 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 flowing 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 coefficients 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 flame 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 ı 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
H2 .g/ C 12 O2 .g/ ! H2 O.l/
P ROBLEMS 365
(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 find 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 find 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 G•D, based on a standard pressure
p ı 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 tempera-
ture 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 ideal-
gas behavior.
(a) For values of the advancement ranging from 0 to 1 mol, at an interval of 0:1 mol or
less, calculate Œ 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./ G.0/ as a function of , and draw a smooth curve through the
points.
(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
“ Q
’i D i for transfer equilibrium, the other based on i i i D 0 or K D i ai i for
P
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 fixed composition maintained at con-
stant 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 fixed amount of each species is given by1

@ .i =T / 1 @i i
D (12.1.1)
@T p;fnig T @T p;fnig T 2
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 be-
comes
@ .i =T / Si i T Si C i
D D (12.1.2)
@T p;fnig T T2 T2
The further substitution i D Hi T Si (Eq. 9.2.46) gives finally

@ .i =T / Hi
D (12.1.3)
@T p;fnig T2
1 Thisrelation 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 .
366
CHAPTER 12 EQUILIBRIUM CONDITIONS IN MULTICOMPONENT SYSTEMS
12.1 E FFECTS OF T EMPERATURE 367
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 p T2
This is the Gibbs–Helmholtz equation.
12.1.2 Variation of ıi =T with temperature

If we make the substitution i D ıi C RT ln ai in Eq. 12.1.3 and rearrange, we obtain
d.ıi =T /

Hi @ ln ai
D R (12.1.5)
dT T2 @T p;fnig
Because ıi =T is a function only of T , its derivative with respect to T is itself a function
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.ıi =T /= dT at a given
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 Hiı (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 coefficient i for the
ideal gas is unity, and pi =p D yi is constant at constant fnig, so that the partial derivative
Œ@ 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 ı , and a mixture composition in the limit given by Eqs. 9.5.20–
9.5.24 in which the activity coefficient is unity. Hi is then the standard molar enthalpy Hiı ,
and the activity is given by an expression in Table 9.5 with the pressure factor and activity
coefficient set equal to 1: ai Dxi , aA DxA , ax;B DxB , ac;B DcB =c ı , or am;B DmB =mı .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, Œ@ ln ai (g)=@T p;fnig is zero and Eq. 12.1.5 becomes
d.ıi =T / Hiı
D (12.1.6)
dT T2
(standard state not based
on concentration)
Equation 12.1.6, as the conditions of validity indicate, does not apply to a solute stan-
dard 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 consid-
ering the following behavior of ln.cB =c ı /:
@ ln.cB =c ı /

1 @cB 1 @.nB =V /
D D
@T p;fnig cB @T p;fnig nB =V @T p;fnig

@.1=V / 1 @V
DV D
@T p;fnig V @T 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 368
Here ˛ is the cubic expansion coefficient of the solution (Eq. 7.1.1). If the activity coeffi-
cient is to be unity, the solution must be an ideal-dilute solution, and ˛ is then ˛A , the cubic
expansion coefficient of the pure solvent. Eq. 12.1.5 for a nonelectrolyte becomes
d.ıc;B =T / HBı
D C R˛A (12.1.8)
dT T2
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 TP. By equat-
ing two expressions for the standard molar reaction Gibbs energy, r G ı D i i ıi and
r G ı D RT ln K (Eqs. 11.8.3 and 11.8.10), we obtain
1 X
ln K D i ıi (12.1.9)
RT
i
The rate at which ln K varies with T is then given by

d ln K 1 X d.ıi =T /
D i (12.1.10)
dT R dT
i
Combining Eq. 12.1.10 with Eqs. 12.1.6 or 12.1.8, and recognizing that i i Hiı is the
P
standard molar reaction enthalpy r H ı , we obtain the final expression for the temperature
dependence of ln K:
d ln K Hı X
D r 2 ˛A i (12.1.11)
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 Hı
D r 2 (12.1.12)
dT RT
(no solute standard states
based on concentration)
We can rearrange Eq. 12.1.12 to

d ln K
r H ı D RT 2 (12.1.13)
dT
We can convert this expression for r H ı to an equivalent form by using the mathematical
identity d.1=T / D .1=T 2 / dT :
d ln K
r H ı D R (12.1.14)
d.1=T /
12.2 S OLVENT C HEMICAL P OTENTIALS FROM P HASE E QUILIBRIA 369
Table 12.1 Comparison of the Clausius–Clapeyron and van’t Hoff equations for vaporiza-
tion of a liquid.
Clausius–Clapeyron equation van’t Hoff equation

d ln.p=p ı / d ln K
vap H R vap H ı D R
d.1=T / d.1=T /
Derivation assumes Vm (g) Vm (l) and An exact relation.
ideal-gas behavior.
vap H is the difference of the molar en- vap H ı is the difference of the molar en-
thalpies of the real gas and the liquid at the thalpies of the ideal gas and the liquid at
saturation vapor pressure of the liquid. pressure p ı .
p is the saturation vapor pressure of the liq- K is equal to a(g)=a(l) D .f =p ı /= (l),
uid. and is only approximately equal to p=p ı .
Equations 12.1.13 and 12.1.14 are two forms of the van’t 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 ı =R
at the corresponding temperature T .
A simple way to derive the equation for this last procedure is to substitute r G ı D
r H ı Tr S ı in r G ı D RT ln K and rearrange to
r H ı 1 Sı

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 ı and

r S ı 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 ı =R, and the tangent to a point on the
curve has its intercept at 1=T D0 equal to r S ı =R.
When we apply Eq. 12.1.14 to the vaporization process A(l)!A(g) of pure A, it resem-
bles the Clausius–Clapeyron equation for the same process (Eq. 8.4.15 on page 219). 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 coefficient m;B of a nonelectrolyte
solute of a binary solution if we know the variation of the osmotic coefficient of the solution
from infinite dilution to the molality of interest. A similar procedure for the mean ionic
activity coefficient of an electrolyte solute was described in Sec. 10.6.
The physical measurements needed to find the osmotic coefficient m of a binary solu-
tion 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 solvent’s chemical potential.
To calculate m from A A , we use Eq. 9.6.16 on page 267 for a nonelectrolyte
solute, or Eq. 10.6.1 on page 299 for an electrolyte solute. Both equations are represented
by
A A
m D (12.2.1)
RTMA mB
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 coefficient3 by a suitable integration from infinite dilution
to the molality of interest.
Sections 12.2.1 and 12.2.2 will describe freezing-point and osmotic-pressure measure-
ments, 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 first appears is pure A. For
example, for a dilute aqueous solution the solid would be ice. The temperature at which
solid A first 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
determine the value of A .l; T 0 / A .sln; T 0 /, where A .l; T 0 / 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 figure shows
A =T for the solvent in the pure solid phase, in the pure liquid phase, and in the fixed-
composition 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 366, are included
in the figure. The desired value of A .l; T 0 / A .sln; T 0 / is the product of T 0 and the
difference of the values of A =T at points e and a. To find this difference, we integrate the
3A measurement of A A also gives us the solvent activity coefficient, based on the pure-solvent reference
state, through the relation A D A C RT ln. A xA / (Eq. 9.5.15).
so
slope D
H (s)
T2
A
lid
b
c
b
d
H (l)
b
liq
slope D
uid
T2
A
A =T
e
so
H
lut
b
slope D
A (sln)
ion
a
b
T2
Tf Tf T0
T
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 over T along the path abcde:
A .l; T 0 / Tf Tf

A .sln; T 0 / HA (sln) HA (sln)
Z Z
D dT dT
T0 T0 T0 T2 Tf T2
Z Tf HA (s)
Z T0 HA (l)
dT dT (12.2.2)
Tf T2 Tf T2
By combining integrals that have the same range of integration, we turn Eq. 12.2.2 into
A .l; T 0 / Tf HA (sln) HA (s)

A .sln; T 0 /
Z
D dT
T0 T0 Tf T2
Z T0 HA (sln) HA (l)
C dT (12.2.3)
Tf T2
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 first 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 tem-
perature Tf . It is usually sufficient 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).

A (l) (A + B)(sln)
p0 p 00
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 first integral and rearranging, we obtain finally

0 0 0
1 1
A .l; T / A .sln; T / D T sol,A H.Tf / Tf sol,A Cp
Tf Tf
0
T T
dil H
Z
C T 0 sol,A Cp ln f C T 0 dT (12.2.5)
Tf Tf T2
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 schemat-
ically 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 imper-
meable to the solute.
The presence of the membrane makes this system different from the multiphase, mul-
ticomponent system of Sec. 9.2.7, used there to derive conditions for transfer equilibrium.
By a modification 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
figure must be fixed in stationary positions. This keeps the volume of each phase constant:
dV ’ D dV “ D 0. Equation 9.2.41 on page 236, expressing the total differential of the
entropy in an isolated multiphase, multicomponent system, becomes
“
T“ T’ A ’A
dS D dS ’ C dn’A (12.2.6)
T“ T“
“
In an equilibrium state, the coefficients .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.
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 flow 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 defined 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 flow 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 solution’s temperature, pressure, and composition. Strictly speaking, ˘ in an equilib-
rium 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 temper-
ature 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 0 C ˘ / D A .p 0 / (12.2.7)
(equilibrium state)
We can use this relation to derive an expression for A .p 0 / A .p 0 / 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 dA D VA dp and integrating at constant temperature and composition from p 0
to p 0 C ˘ , we obtain
Z p 0 C˘
0 0
A .p C ˘ / A .p / D VA dp (12.2.9)
p0
4 Greek for push.
12.3 B INARY M IXTURE IN E QUILIBRIUM WITH A P URE P HASE 374
Substitution from Eq. 12.2.7 changes this to

Z p 0 C˘
A .p 0 / A .p 0 / D VA dp (12.2.10)
p0
(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
A .p 0 / A .p 0 / D VA ˘ (12.2.11)
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 find general relations among changes of
temperature, pressure, and mixture composition in the two-phase equilibrium system that
can be applied to specific 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 dT C dp C dxA (12.3.1)
@T p;xA @p T;xA @xA T;p
With substitutions from Eqs. 9.2.49 and 12.1.3, this becomes

HA VA @.A =T /
d.A =T / D dT C dp C dxA (12.3.2)
T2 T @xA T;p
Next we write the total differential of A =T for pure solid or gaseous A. The indepen-
dent variables are T and p; the expression is like Eq. 12.3.2 with the last term missing:
HA VA
d.A =T / D dT C dp (12.3.3)
T2 T
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 satisfied. 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)
T2 T @xA T;p
12.4 C OLLIGATIVE P ROPERTIES OF A D ILUTE S OLUTION 375
which we can rewrite as

sol,A H sol,A V

@.A =T /
dT dp D dxA (12.3.5)
T2 T @xA T;p (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:
T2

@T @.A =T /
D (12.3.6)
@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 find 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, com-
pared 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 373).
Note that all four properties are defined 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 fixed pressure, pure liquid water is in equilibrium
with ice at the freezing point and with steam at the boiling point. These are the temperatures
12 H2
O(s
)
ut
bc
10
Tf
8 H
1
2O
so (l)
A =kJ mol
lu
tio
6 n
4
bc
ut
H 2O
Tb
(g)
0
240 280 320 360 400
T=K
Figure 12.3 Freezing-point depression and boiling-point elevation of an aqueous so-
lution. 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 figure. 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 12.4.1–12.4.4 will derive theoretical relations between each of the four col-
ligative properties and solute composition variables in the limit of infinite dilution. The
expressions show that the colligative properties of a dilute binary solution depend on prop-
erties 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 coefficient of a
solute, they are useful for estimating the solute molar mass. For example, from a measure-
ment 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 approxi-
mate 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 .
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 375 gives the general dependence of temperature on the com-
position 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 infinite 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 infinite
dilution, we use the fact that the solvent activity coefficient A approaches 1 in this limit.
Then the solvent chemical potential is given by the Raoult’s law relation
(solution at infinite 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 de-
scribed in Sec. 9.4.6 for an ideal-dilute binary solution of a nonelectrolyte. We must calcu-
late xA from the amounts of all species present at infinite dilution. In the limit of infinite
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 nB
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 infinite dilution, then, Eq. 12.4.1 becomes

R.Tf /2

@Tf
lim D (12.4.5)
xA !1 @xA p fus,A H
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 nB
5 At the freezing point of the mixture, the reference state is an unstable supercooled liquid.
In the limit of infinite 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 225, we obtain the relations
dxA D d.1 xA / D d.nB =nA /

D VA dcB
D MA dmB (12.4.7)
(binary solution at
infinite dilution)
which transform Eq. 12.4.5 into the following:6
VA R.Tf /2

@Tf
lim D
cB !0 @cB p fus,A H
MA R.Tf /2

@Tf
lim D (12.4.8)
mB !0 @mB p fus,A H
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
Tf D mB (12.4.9)
fus,A H
The molal freezing-point depression constant or cryoscopic constant, Kf , is defined

for a binary solution by
def Tf
Kf D lim (12.4.10)
mB !0 mB
and, from Eq. 12.4.9, has a value given by
MA R.Tf /2
Kf D (12.4.11)
fus,A H
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 depres-
sion of the freezing point, the mean ionic activity coefficient deviates greatly from unity and
the relation is not accurate.
6A small dependence of VA on T has been ignored.
BIOGRAPHICAL SKETCH
François-Marie Raoult (1830–1901)
Starting in 1886 Raoult began publishing
his measurements of the vapor pressures of di-
lute 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 up-
per 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 flask connected to a reflux 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, respec-
A
in northern France into a family of modest tively, both at the same temperature, and N
means—his father was an official in the local is one-hundred times the solute mole fraction
customs service. He supported himself with x . This relation is equivalent to the Raoult’s
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 doctor’s degree from the University of wrote:d
Paris in 1863.
With a view to ascertain whether this remark-
Raoult began teaching at the University of able law is general, I dissolved in ether com-
Grenoble in 1867, and three years later was ap- pounds 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
whom died before him. one another; and I measured the vapor pressures
His strength was in experimental measure- of the solutions obtained. In every case I found
0
ments rather than theory. He constructed most . . . that the ratio ff Nf is very nearly 0.01.
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 sol-
versal 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 confirmed the theory of so-
for his doctoral thesis, and later at Grenoble, lutions being developed by J. H. van’t Hoff.
was on the thermochemistry of galvanic cells. Raoult’s work brought him many honors.
His first measurements of freezing-point de- He was most proud of being named Comman-
pressions 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 Raoult’s per-
of a substance from the freezing point of its sonality as follows:f
dilute solution, and gave specific examples of Though modest and retiring, Raoult’s devotion
this procedure. He was the first to show exper- to 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 ad-
formula unit (Eq. 12.4.12). vancement of science.
a Ref. [16]. b Ref. [139]. c Ref. [140]. d Ref. [139]. e Ref. [140]. f Ref. [137].
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
lim D (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 defined
for a binary solution by
def Tb
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
MA R.Tb /2
Kb D (12.4.15)
vap,A H
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 mea-
surement 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 375 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
infinite 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

@p @p MA RT
lim D lim D (12.4.17)
cB !0 @cB T vap,A V mB !0 @mB T 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 MA 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.
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 infinite dilu-
tion are
p VA pA cB and p MA pA mB (12.4.19)
We see that the lowering in this limit depends on the kind of solvent and the solution com-
position, but not on the kind of solute.
12.4.4 Osmotic pressure

The osmotic pressure ˘ is an intensive property of a solution and was defined 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
˘D A (12.4.20)
VA
In the limit of infinite dilution, A A approaches RT ln xA (Eq. 12.4.2) and VA be-
comes the molar volume VA of the pure solvent. In this limit, Eq. 12.4.20 becomes
RT ln xA
˘D (12.4.21)
VA
from which we obtain the equation

@˘ RT
lim D (12.4.22)
xA !1 @xA T;p VA
The relations in Eq. 12.4.7 transform Eq. 12.4.22 into

@˘
lim D RT (12.4.23)
cB !0 @cB T;p

@˘ RTMA
lim D D A RT (12.4.24)
mB !0 @mB T;p VA
Equations 12.4.23 and 12.4.24 show that the osmotic pressure becomes independent of
the kind of solute as the solution approaches infinite dilution. The integrated forms of these
equations are
˘ D cB RT (12.4.25)
RTMA
˘D mB D A RT mB (12.4.26)
VA
Equation 12.4.25 is van’t Hoff’s equation
P for osmotic pressure.
P If there is more than one
solute species, cB can be replaced by i¤A ci and mB by i¤A mi in these expressions.
In Sec. 9.6.3, it was stated that ˘=mB is equal to the product of m and
P the limiting
value of ˘=mB at infinite dilution, where m D .A A /=RTMA i ¤A mi is the
osmotic coefficient. This relation follows directly from Eqs. 12.2.11 and 12.4.26.
BIOGRAPHICAL SKETCH
Jacobus Henricus van’t Hoff (1852–1911)
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, van’t Hoff wrote:b
. . . the relation found permits of an important ex-
tension of the law of Avogadro, which now finds
application also to all [nonelectrolyte] solutions,
van’t Hoff was a Dutch chemist who gained if only osmotic pressure is considered instead of
elastic pressure. At equal osmotic pressure and
fame from epoch-making publications in sev-
equal temperature, equal volumes of the most
eral fields of theoretical chemistry. He was an
widely different solutions contain an equal num-
introvert who valued nature, philosophy, po- ber of molecules, and, indeed, the same number
etry, and the power of imagination.a which, at the same pressure and temperature, is
van’t Hoff was born in Rotterdam, the son contained in an equal volume of a gas. . . . the
of a physician, and studied at Delft, Leyden, existence of the so-called normal molecular low-
Bonn, and Paris before obtaining his doctor’s ering of the freezing-point and diminution of the
degree at Utrecht in 1874. For 18 years he was vapor-pressure were not discovered until Raoult
Professor of Chemistry, Mineralogy, and Geol- employed the organic compounds. These sub-
ogy at the University of Amsterdam. In 1896, stances, almost without exception, behave nor-
mainly to escape burdensome lecture duties, mally. It may, then, have appeared daring to give
Avogadro’s law for solutions such a prominent
he accepted an appointment as Professor of
place, and I should not have done so had not Ar-
Physical Chemistry at the University of Berlin,
rhenius pointed out to me, by letter, the probabil-
where he remained for the rest of his life. ity that salts and analogous substances, when in
In the same year he received his doctor’s desolution, break down into ions.
gree, he published the first explanation of opti-
In 1901 van’t Hoff was awarded the first
cal activity based on the stereoisomerism of an
Nobel Prize in Chemistry “in recognition of
asymmetric carbon atom. Similar ideas were
the extraordinary services he has rendered by
published independently two months later by
the discovery of the laws of chemical dynam-
the French chemist Joseph Le Bel.
ics and osmotic pressure in solutions.”
In 1884 he entered the field of physical
van’t Hoff was married with two daughters
chemistry with his book Études de Dynamique
and two sons. He died of tuberculosis at age
Chimique, a systematic study of theories of re-
fifty-eight. In a memorial article, Frederick
action kinetics and chemical equilibrium. The
Donnanc wrote:d
book introduced, in the form of Eq. 12.1.12,
his expression for the temperature dependence The present writer is one of those whose priv-
ilege it is to have worked under van’t Hoff.
of an equilibrium constant.
. . . Every day endeared van’t Hoff to the small
van’t Hoff next used thermodynamic cycles band of workers in his laboratory. His joy in his
to reason that solutions of equal osmotic pres- work, the simple and unaffected friendliness of
sure should have the same values of freezing- his nature, and the marvellous power of his mind
point depression and of vapor-pressure lower- affected us most deeply. All who worked with
ing. Then, from Raoult’s extensive measure- van’t Hoff quickly learned to love and respect
ments of these values, he found that the os- him, and we were no exception to the rule.
a Ref.[123]. b Ref. [162]. 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 383
12.5 SOLID–LIQUID 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 tempera-
ture) 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 infinite 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 mix-
ture, when the mixture is in equilibrium with pure solid A, is given by Eq. 12.3.6:
T2

@T @.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 Œ@.A =T /=@xA T;p D R=xA . With these
substitutions, Eq. 12.5.1 becomes
2
@Tf;A RTf;A

D (12.5.2)
@xA p xA sol,A H
(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 benzene–toluene and
benzene–cyclohexane. 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
figure. The agreement is good over a wide range of compositions for benzene–toluene mix-
tures (open circles), which are known to closely approximate ideal liquid mixtures. The
agreement for benzene–cyclohexane mixtures (open triangles), which are not ideal liquid
mixtures, is confined 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
integrate as follows from an arbitrary mixture composition xA0 at the freezing point Tf;A
0
to

pure liquid A at its freezing point Tf;A :

Tf;A 1
R
Z Z
dT dxA
2
D (12.5.3)
0
Tf;A T fus,A H xA0 xA
280 bc
bc
270
bc ut
Tf;A =K
ut
260 bc
ut
bc
250 ut
bc
ut
240
0 0:2 0:4 0:6 0:8 1:0
xB
Figure 12.4 Dependence on composition of the freezing point of binary liquid mix-
tures with benzene as component A.a Solid curve: calculated for an ideal liquid mix-
ture (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. [122].
5
2
Tf;A =Tf
1
0:5
0
0 xB 1
Figure 12.5 Freezing-point curves of ideal binary liquid mixtures. The solid is com-
ponent 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 1 1
ln xA D (12.5.4)
R Tf;A Tf;A
(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.
12.5.2 Solubility of a solid nonelectrolyte

Suppose we find that a solution containing solute B at a particular combination of temper-
ature, 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 frac-
tion, concentration, or molality of B in the saturated solution. We can also define 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
B (s) K
xB D (12.5.5)
x;B x;B
If we measure the solubility at the standard pressure, the pressure factors B (s) and x;B
are unity and the solubility is given by
K
xB D (12.5.6)
x;B (solubility of solid B, pDp ı )
If the pressure is not exactly equal to p ı , 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 find the standard molar enthalpy of solution of B from the temperature depen-
dence of the solubility. Combining Eqs. 12.1.12 and 12.5.6, we obtain
d ln. x;B xB /
sol,B H ı D RT 2 (12.5.7)
dT
(pDp ı )
The solubility may be small enough for us to be able to set the solute activity coefficient
equal to 1, in which case Eq. 12.5.7 becomes
d ln xB
sol,B H ı D RT 2 (12.5.8)
dT
(pDp ı , x;B D1)
If the solubility xB increases with increasing temperature, sol,B H ı must be positive

and the solution process is endothermic. A decrease of solubility with increasing tem-
perature implies an exothermic solution process. These statements refer to a solid of low
solubility; see page 357 for a discussion of the general relation between the temperature
dependence of solubility and the sign of the molar differential enthalpy of solution at satu-
ration.
8 Inthis 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.
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 =mı . If we use a solute
standard state based on concentration, the expressions become slightly more complicated.
The solubility in this case is given by
ı
B (s) Kc
cB D (12.5.9)
c;B c;B
From Eq. 12.1.11, we obtain, for a nonelectrolyte solid of low solubility, the relation
ı
ı 2 d ln.cB =c /
sol,B H D RT C ˛A (12.5.10)
dT
(pDp ı , 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 benzene–toluene mixtures shown by the open circles in
Fig. 12.4 on page 384, 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 fixed proportion. This kind of solid is called
a solid compound, or stoichiometric addition compound. Examples are salt hydrates (salts
with fixed numbers of waters of hydration in the formula unit) and certain metal alloys.
1:0
0:8
Tf;B D 380 K
0:6
xB
0:4
Tf;B D 400 K
0:2
0
300 350 400
T =K
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 fixed. 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
P Tf of the liquid. At equilibrium, the molar reaction Gibbs energy defined by r G D
point
i i i is zero:
aA bB 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 dŒ(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 find expressions for the total differentials of these quantities at
constant p with the help of Eq. 12.1.3 on page 366:

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
Hm (s)
dŒ(s)=T D dT (12.5.16)
T2
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 Gibbs–Duhem
equation xA dA C xB dB D 0 (Eq. 9.2.27 on page 233), which applies to changes at
constant T and p. We rearrange the Gibbs–Duhem equation to dB D .xA =xB / dA 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 (12.5.19)
dxA sol H xA xB @xA T;p
which can also be written in the slightly rearranged form

dTf bTf a xA @A
D (12.5.20)
dxA sol H b 1 xA @xA T;p
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 Raoult’s 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 a b
D (12.5.21)
dxA sol H xA xB
900
bc
850
800
Tf = K
750
700
650
0 0:2 0:4 0:6 0:8 1:0
Mg mole fraction Zn Zn
Figure 12.7 Solid curve: freezing-point curve of a liquid melt of Zn and Mg that
solidifies 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:
Z T 00 Z xA00 Z xB00 !
f dT
f R a b
2
D dxA C dxB (12.5.22)
Tf0 Tf
fus H xA0 xA xB0 xB
(The second integral on the right side comes from changing dxA to dxB .) The result of
the integration is
xA00 xB00

1 1 R
D 00 C a ln 0 C b ln 0 (12.5.23)
Tf0 Tf fus H xA xB
(ideal liquid mixture in
equilibrium with solid
compound, sol H Dfus H )
Let Tf0 be the freezing point of a liquid mixture of composition xA0 and xB0 D 1 xA0 , and
let Tf00 be the melting point of the solid compound of composition xA00 D a=.aCb/ and xB00 D
b=.a C b/. Figure 12.7 shows an example of a molten metal mixture that solidifies to an
alloy of fixed 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 solu-
tion containing the solvated ions:
MC X (s) • C MzC (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
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 sat-
urated 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 293, we have am;B D m;B ˙ .mC =mı /C .m =mı / . This
zC z
expression is valid whether or not the ions M and X 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
C m
mC

m;B
Ks D ˙ (12.5.25)
B mı mı
where D C C is the total number of ions per formula unit. ˙ is the mean ionic
activity coefficient 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 first 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

Ks D r ˙ CC B
(12.5.26)
mı
(no common ion)
If the ionic strength of the saturated salt solution is sufficiently low (i.e., the solubility is
sufficiently low), it may be practical to evaluate the solubility product with Eq. 12.5.26 and
an estimate of ˙ from the Debye–Hückel limiting law (see Prob. 12.19). The most accurate
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 sparingly-
soluble salt MC 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 391
z
0 Y 0 at molality m . The common ion in this example is the cation M C . The
salt MC C
expression for the solubility product is now
0
Ks D r ˙ .C mB C C mC /C . mB / =.mı / (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
coefficient ˙ 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 find the temperature dependence of Ks by applying Eq. 12.1.12:
d ln Ks sol,B H ı
D (12.5.28)
dT RT 2
At the standard pressure, sol,B H ı is the same as the molar enthalpy of solution at infinite
dilution, sol,B H 1 .
12.6 LIQUID–LIQUID 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 mer-
cury, when equilibrated with water, has some H2 O dissolved in it, and some mercury dis-
solves 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 fixed; the compositions depend only on the identity of the substances and the
temperature and pressure.
Figure 13.5 on page 431 shows a phase diagram for a typical binary liquid mixture that
spontaneously separates into two phases when the temperature is lowered. The thermody-
namic 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 Raoult’s 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.
We can define 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 fixed 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
KD “ D “ “ “
(12.6.1)
aB B B Bx
The solubility of B is then given by
“ “ “
B B xB
xB’ D ’ ’ K (12.6.2)
x;B x;B
The values of the pressure factors and activity coefficients 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
d ln xB’ d ln K sol,B H ı
D (12.6.3)
dT dT RT 2
where sol,B H ı is the molar enthalpy change for the transfer at pressure p ı of pure liquid
solute to the solution at infinite 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 mini-
mum at about 12 ı C. Equation 12.6.3 allows us to deduce from this behavior that sol,B H ı
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.
bc
12:9
13:0 bc
ln xB
bc
13:1
bc
bc
13:2 bc
bc
bc bc
bc bc
13:3
0 10 20 30 40 50
t=ı C
Figure 12.8 Aqueous solubility of liquid n-butylbenzene as a function of temperature

(Ref. [125]).
Pure-liquid reference state

“
The condition for transfer equilibrium of component B is ’B D B . If we use a pure-liquid
reference state for B in both phases, this condition becomes
“ “
B C RT ln. ’ ’
B xB / D B C RT ln. B xB / (12.6.4)
This results in the following relation between the compositions and activity coefficients:
’ ’ “ “
B xB D 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 non-
polar liquid solute that has very low solubility in water, so that the aqueous solution is
essentially at infinite dilution. Let the nonpolar solute be component B, and let the aqueous
phase that is equilibrated with liquid B be phase ’. The activity coefficient B’ is then a
limiting activity coefficient or activity coefficient at infinite 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 Henry’s 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 ’ ’ x ’ (Table 9.4).
D fB =kH,B B
“ “
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
“ “
B xB fB
xB’ D ’ ’ (12.6.6)
x;B kH,B
12.7 M EMBRANE E QUILIBRIA 394
“
If we assume as before that the activity coefficients 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
’ ’ ’ ’
ax;C x;C x;C xC
KD “
D “ “ “
(12.6.7)
ax;C x;C x;C xC
We define K 0 as the ratio of the mole fractions of C in the two phases at equilibrium:
“ “
0 def xC’ x;C x;C
K D “
D ’ ’ K (12.6.8)
xC x;C x;C
At a fixed 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’C =mC and cC’ =cC also approach constant values at a given T and p. The
ratio of concentrations is called the partition coefficient or distribution coefficient.
In the limit of infinite 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 443.
12.7 MEMBRANE EQUILIBRIA
A semipermeable membrane used to separate two liquid phases can, in principle, be per-
meable to certain species and impermeable to others. A membrane, however, may not be
perfect in this respect over a long time period (see page 373). 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 two-
phase 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
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 372).
Consider a system with transfer equilibrium of the solvent across a membrane sepa-
rating phases ’ and “. The phases have equal solvent chemical potentials but different
pressures:
“
A .p “ / D ’A .p ’ / (12.7.1)
The dependence of A on pressure in a phase of fixed 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 “ / D A .p ’ / C VA .p “ p’/ (12.7.2)
“
By equating the two expressions for A .p “ / and rearranging, we obtain the following ex-
pression for the pressure difference needed to achieve transfer equilibrium:
“
’A .p ’ / A .p ’ /
p“ p’ D “
(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 374, 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 !
’
V
p “ p ’ D ˘ “ .p ’ / A
“
˘ ’ .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 macro-
molecule 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. Measure-
ments 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 coeffi-
cients, allow the amount of ligand bound per macromolecule to be calculated.
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 ’B D B , or
“
.ım;B /’ C RT ln am;B
’
D .ım;B /“ C RT ln am;B (12.7.5)
Solute B has the same standard state in the two phases, so that .ım;B /’ and .ım;B /“ are
’ “
equal. The activities am;B and am;B are therefore equal at equilibrium. Using the expression
for solute activity from Eq. 10.3.16, which is valid for a multisolute solution, we find that at
transfer equilibrium the following relation must exist between the molalities of the salt ions
in the two phases:
C
’ ’ ’ C ’ “ “ “
m“

m;B ˙ m C m D m;B ˙ m C (12.7.6)
To find an expression for the Donnan potential, we can equate the single-ion chemical
“
potentials of the salt cation: ’C . ’ / D C . “ /. When we use the expression of Eq.
10.1.15 for C ./, we obtain
“ “ “
’ “ RT C C mC
D ln ’ ’ ’ (12.7.7)
zC F C C mC (Donnan potential)
The condition needed for an osmotic membrane equilibrium related to the solvent can
be written
“
A .p “ / ’A .p ’ / D 0 (12.7.8)
The chemical potential of the solvent is A D ıA C RT ln aA D ıA C RT ln. A A xA /.
From Table 9.6, we have to a good approximation the expression RT ln A D VA .p p ı /.
“ “
A xA

RT ln ’ ’ C VA p “ p’ D 0 (12.7.9)
A xA
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
RT xA’
p“ p’ ln (12.7.10)
VA xA
“
P
In the limit of infinite dilution, ln xA can be replaced by MA i ¤A mi (Eq. 9.6.12 on
page 266), giving the relation
M RT X “ X “
p“ p’ A mi m’i D A
RT mi m’i (12.7.11)
VA
i ¤A i¤A
Example
As a specific 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
an 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 confined 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 figure, 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 coefficients are unity. Then for the present
example, with C D D 1, the equation becomes
“
m’C m’ mC m“ (12.7.12)
phase phase
phase 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’C D m’ 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’C is less than mC . These equilibrium molalities are depicted in Fig. 12.9(b).
The chemical potential of a cation, its activity, and the electric potential of the phase are
related by Eq. 10.1.9 on page 287: C D ıC C RT ln aC C zC F . In order for MC to
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 coefficients approximated by 1 and
with zC for this example set equal to 1:
“
RT m
’ “
ln C (12.7.15)
F m’C
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 399
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 sufficient to cause the pressure difference.
It should be clear that the existence of a Donnan membrane equilibrium introduces
complications that would make it difficult 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 LIQUID–GAS 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 Œ@i (l)=@pT;fnig D Vi (l) and i (g) D ıi (g) C
RT ln.fi =p ı / (Eqs. 9.2.49 and 9.3.12), we obtain
d ln.fi =p ı / V (l)
D i (12.8.1)
dp RT
(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 com-
position, 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.
The integral in the Poynting factor is simplified if we make the approximation that Vi (l)
is independent of pressure. Then we obtain the approximate relation

Vi (l).p2 p1 /
fi .p2 / fi .p1 / exp (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 mo-
lar 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
fixed. 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 246. A binary liquid mixture of two volatile com-
ponents, 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:
fA fB
A (l) D ıA (g) C RT ln B (l) D ıB (g) C RT ln (12.8.4)
pı pı
Suppose we make an infinitesimal change in the liquid composition at constant T and
p. This causes infinitesimal changes in the chemical potentials and fugacities:

dfA dfB
dA (l) D RT dB (l) D RT (12.8.5)
fA fB
By inserting these expressions in the Gibbs–Duhem equation xA dA D xB dB (Eq.
9.2.43), we obtain
xA xB
df D df (12.8.6)
fA A fB B
(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 Henry’s 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 (12.8.8)
dxB xB
With the substitution dxB D dxA and rearrangement, Eq. 12.8.8 becomes
xB
df D dxA (12.8.9)
fB B
Combined with Eq. 12.8.6, this is .xA =fA / dfA D dxA , which we can rearrange and inte-
grate as follows within the ideal-dilute region:
Z fA0 Z xA0 fA0
dfA dxA
D ln D ln xA0 (12.8.10)
fA fA 1 xA fA
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 Raoult’s law for fugacity applied to component
A.
If component B obeys Henry’s law at all compositions, then the Henry’s law constant
kH,B is equal to fB and B obeys Raoult’s 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 Henry’s law,
the solvent must obey Raoult’s law. (A similar result was derived in Sec. 9.4.6 for a
solution with any number of solutes.)
fA =kPa and fB =kPa

bc
fB bc
4
2 fA
0
0 0:2 0:4 0:6 0:8 1:0
xB
Figure 12.10 Fugacities in a gas phase equilibrated with a binary liquid mixture of
H2 O (A) and ethanol (B) at 25 ı C and 1 bar. a The dashed lines show Raoult’s law
behavior. The dotted lines illustrate the inequality .dfB = dxB / < .fB =xB /.
2. If one component of a binary liquid mixture obeys Raoult’s law at all compositions,
so also must the other component. This is the definition of an ideal binary liquid
mixture (Sec. 9.4.2).
Suppose we have a nonideal binary liquid mixture in which component B exhibits posi-
tive deviations from Raoult’s law. An example of this behavior for the water–ethanol 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 figure), except that the two slopes become equal at
xB D1:
dfB f
B (12.8.12)
dxB xB
As we can see from the figure, this relation must apply to any component whose fugacity
curve exhibits a positive deviation from Raoult’s law and has only one inflection 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
dfB f
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 (12.8.14)
fB C DxB
11 This procedure is similar to the rectification procedure described on page 146.
40
30
fA =kPa and fB =kPa

20 fB
10
fA
0
0 0:2 0:4 0:6 0:8 1:0
xB
Figure 12.11 Fugacities in a gas phase equilibrated with a binary liquid mixture of
chloroform (A) and ethanol (B) at 35 ı C (Ref. [149]).
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 dfA dx
df dxA or A (12.8.16)
fA A fA xA
We can integrate both sides of the second relation as follows:12

Z fA0 Z xA0 fA0
dfA dxA
ln ln xA0 fA xA fA (12.8.17)
fA fA 1 xA fA
Thus, if the curve of fugacity versus mole fraction for one component of a binary liquid
mixture exhibits only positive deviations from Raoult’s law, with only one inflection point,
so also must the curve of the other component. In the water–ethanol system shown in
Fig. 12.10, both curves have positive deviations from Raoult’s law, and both have a single
inflection point.
By the same method, we find that if the fugacity curve of one component has only
negative deviations from Raoult’s law with a single inflection 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 Raoult’s 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.
possible because both fugacity curves have two inflection points instead of the usual one.
Other types of unusual nonideal behavior are possible.13
12.8.3 The Duhem–Margules equation

When we divide both sides of Eq. 12.8.6 by dxA , we obtain the Duhem–Margules equa-
tion:
xA dfA xB dfB
D (12.8.18)
fA dxA fB dxA
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 Duhem–Margules equation then becomes
xA dpA xB dpB
D (12.8.19)
pA dxA pB dxA
with ideal gas, constant T and p)
Solving Eq. 12.8.19 for dpB = dxA , we obtain
dpB xA pB dpA
D (12.8.20)
dxA xB pA dxA
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 liquid–gas 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 xA =xB
D 1 (12.8.21)
dxA yA =yB
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 liquid–gas 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. [112].
Thus compared to the liquid, the gas phase is richer in the component whose addi-
tion to the liquid at constant temperature causes the total pressure to increase. This
statement is a version of Konowaloff’s rule.
In some binary liquid–gas 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 equi-
librium constant is K D aB (sln)=aB (g).
The activity of B in the gas phase is given by aB (g) D fB =p ı . If the solute is a nonelec-
trolyte 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
x;B x;B xB
KD (12.8.22)
fB =p ı
and the solubility, expressed as the equilibrium mole fraction of solute in the solution, is
given by
KfB =p ı
xB D (12.8.23)
x;B x;B (nonelectrolyte solute in
equilibrium with gas)
At a fixed 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 coefficient 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 ı , we can with negligible error set the pres-
sure factor x;B equal to 1. When the gas solubility is low and the solution contains no
other solutes, the activity coefficient x;B is close to 1. If furthermore we assume ideal gas
behavior, then Eq. 12.8.23 becomes
pB
xB D K (12.8.24)
pı
(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 285.
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 fixed
partial pressure of the gas:
sol,B H ı

@ ln xB d ln K
D D (12.8.25)
@T pB dT RT 2
At the standard pressure, sol,B H ı is the same as the molar enthalpy of solution at infinite
dilution.
Since the dissolution of a gas in a liquid is invariably an exothermic process, sol,B H ı
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 tem-
perature 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 ı

@ ln xB
D (12.8.26)
@.1=T / pB R
We can use this form to evaluate sol,B H ı 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 Raoult’s 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
xB D B (12.8.27)
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 liquid–vapor 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 fixed value of pB .
1. The ideal solubility, expressed as a mole fraction, is independent of the kind of sol-
vent.
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.
following solubilities of gaseous Cl2 in several dissimilar solvents at 0 ı 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 Henry’s 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 ı
The Henry’s law constant kH,B is related to fB and xB by
fB
kH,B D (12.8.29)
x;B xB
(see Table 9.4), and is therefore related to K as follows:

ı
x;B p
kH,B D (12.8.30)
K
(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 ı 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 quan-
tity kH,B =p ı on temperature:
d ln.kH,B =p ı / d ln K sol,B H ı
D D (12.8.31)
dT dT RT 2
(pDp ı )
d ln.kH,B =p ı / d ln K sol,B H ı
D D (12.8.32)
d.1=T / d.1=T / R
(pDp ı )
These expressions can be used with little error at any pressure that is not much greater than
p ı , say up to at least 2 bar, because under these conditions x;B does not differ appreciably
from unity (page 274).
To find 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:
0 ı Z p0 1 !
0 x;B .p / p pı VB
kH,B .p / D D exp dp (12.8.33)
K K p ı RT
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 408
Table 12.2 Expressions for activities (from Table 9.5 and Eqs.
10.1.14 and 10.3.16)
Species Activity
f
Pure gas a(g) D
pı
Pure liquid or solid aD
fi
Substance i in a gas mix- ai (g) D
ture pı
Substance i in a liquid or ai D i i xi
solid mixture
Solvent A of a solution aA D A A xA
Nonelectrolyte solute B, ax;B D x;B x;B xB
mole fraction basis
cB
Nonelectrolyte solute B, ac;B D c;B c;B ı
concentration basis c
mB
Nonelectrolyte solute B, am;B D m;B m;B
molality basis mı
m C m
C
Electrolyte solute B am;B D m;B ˙
mı mı
mC m
Ion in solution aC D C C ı a D
m mı
An approximate version of this relation, found by treating VB1 as independent of pressure,

is 1
VB .p2 p1 /
kH,B .p2 / kH,B .p1 / exp (12.8.35)
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 definition of the thermodynamic equilibrium constant of a reaction or other chemical

process is given by Eq. 11.8.9:

Y
KD .ai /eqi (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 409
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:
aC a2 C
2m
Cm
2 =.mı /3
KD D r (12.9.2)
aCaCO3 aCO2 aH2 O fCO2 =p ı H2 O xH2 O
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
2
C
r D (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 constant—one using more
than one kind of standard state. From the definition of the mean ionic activity coefficient
(Eq. 10.3.7), we can replace the product C 2 by ˙ 3 , where
˙ is the mean ionic activity
coefficient of aqueous Ca(HCO3 )2 :
3 2 ı 3
˙ mC m =.m /
KD r (12.9.4)
fCO2 =p ı H2 O xH2 O
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
am;B
KD (12.9.5)
aCaCO3 aCO2 aH2 O
We then obtain Eq. 12.9.4 by replacing am;B with the expression in Table 12.2 for an elec-
trolyte 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
2
C mC m ˙ mC m
Ka D r ı
D r ı
(12.9.6)
m,HA mHA m m,HA mHA m
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
15 Theproduct C 2 in the numerator of Eq. 12.9.3 is the pressure factor m;B for the solute Ca(HCO3 )2 (see
Eq. 10.3.11 on page 292).
12.10 E VALUATION OF S TANDARD M OLAR Q UANTITIES 410
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 =mı
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 ı D r H ı Tr S ı (12.10.1)
and r S ı can be calculated from the standard molar entropies of the reactants and products
using Eq. 11.8.22: X
r S ı D i Siı (12.10.2)
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 ı for a reaction (Sec. 11.5). Calorimetric measurements of heat capacity and phase-
transition enthalpies can be used to obtain the value of Siı for a solid or liquid (Sec. 6.2.1).
For a gas, spectroscopic measurements can be used to evaluate Siı (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 ı and r H ı through the relations r G ı D RT ln K and
r H ı 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 ı , r G ı , and r S ı to be
evaluated for such a reaction.
An efficient 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 di-
rectly. The relations between standard molar reaction and formation quantities (Sec. 11.3.2)
are X X
r H ı D i f H ı .i/ r G ı D i f G ı .i/ (12.10.3)
i i
and for ions the conventions used are
f H ı (HC , aq) D 0 f G ı (HC , aq) D 0 Smı (HC , aq) D 0 (12.10.4)
Appendix H gives an abbreviated set of values of f H ı , Smı , and f G ı at 298:15 K.
12.10 E VALUATION OF S TANDARD M OLAR Q UANTITIES 411
For examples of the evaluation of standard molar reaction quantities and standard mo-
lar formation quantities from measurements made by various experimental techniques, see
Probs. 12.18–12.20, 14.3, and 14.4.
P ROBLEMS 412
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 =ı C p=Torr t=ı C p=Torr

842:3 343:0 904:3 879:0
852:9 398:6 906:5 875:0
854:5 404:1 937:0 1350
868:9 510:9 937:0 1340
a Ref. [153].
(a) What is the approximate relation between p and K?

(b) Plot these data in the form ln K versus 1=T , or fit 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 tf tb fus H vap H
1
18:0153 g mol 0:00 ı C 99:61 ı C 6:010 kJ mol 1
40:668 kJ mol 1
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 ı C. Estimate the molar
mass of this protein.
12.6 Figure 12.8 on page 393 shows a curve fitted to experimental points for the aqueous solubility
of n-butylbenzene. The curve has the equation ln xB D a.t=ı 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 frac-
tion, and calculate sol,B H ı and sol,B S ı at 5:00 ı C, 12:13 ı C (the temperature of minimum
solubility), and 25:00 ı C.
P ROBLEMS 413
12.7 Consider a hypothetical system in which two aqueous solutions are separated by a semiper-
meable 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 fill 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 ı 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 mem-
brane 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 ı C is 0:997 g cm 3 .)
12.8 The derivation of Prob. 9.3 on page 280 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 ı 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 equili-
brated with this water droplet, and the fugacity in air equilibrated at the same temperature
and pressure with a pool of liquid water having a flat surface. Liquid water at 25 ı 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,
find the relation between sol G ı (solute standard state based on mole fraction) and the Henry’s
law constant kH,B .
12.10 Crovetto16 reviewed the published data for the solubility of gaseous CO2 in water, and fitted
the Henry’s law constant kH,B to a function of temperature. Her recommended values of kH,B at
five temperatures are 1233 bar at 15:00 ı C, 1433 bar at 20:00 ı C, 1648 bar at 25:00 ı C, 1874 bar
at 30:00 ı C, and 2111 bar at 35 ı 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 ı C expressed
both as a mole fraction and as a molality.
(b) Find the standard molar enthalpy of solution at 25:00 ı 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 first ionization constant of
carbonic acid.) Combine the kH,B data with data in Appendix H to evaluate K and r H ı
for the ionization reaction at 25:00 ı 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 414
12.11 The solubility of gaseous O2 at a partial pressure of 1:01 bar and a temperature of 310:2 K, ex-
pressed 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. Calcu-
late the activity coefficient 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 es-
timate the percent change in the value of the Henry’s 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 nonideal-
ity 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 Second virial coefficients:

5 6
xB D 2:02142 10 BAA D 1152 10 m3 mol 1
g 3 6
.nB =V g / D 35:9957 mol m BBB D 16:2 10 m3 mol 1
pA D 3167:13 Pa BAB D 27:0 10 6

m3 mol 1
VA D 18:069 10 6
m3 mol 1
VB1 D 31:10 10 6
m3 mol 1
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 Henry’s
law constant of oxygen to six significant figures. Your own calculations should also be carried
out to six significant figures. For the gas constant, use the value R D 8:31447 J K 1 mol 1 .
The method you will use to evaluate the Henry’s law constant kH,B D fB =xB at the experimen-
tal temperature and pressure is as follows. The value of xB is known, and you need to find 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 coefficients. 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 A , the fugacity
coefficient of A in the gas mixture.
(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. [106]. 18 Ref. [143].
P ROBLEMS 415
(d) Use your most recently calculated values of p, A , and fA to calculate an improved value
of yB .
(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 Henry’s 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
Henry’s 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 ı / 1=.T =K/ ln .kH,B =p ı /

0:00363029 10:0569 0:00329870 10:6738
0:00359531 10:1361 0:00319326 10:8141
0:00352175 10:2895 0:00314307 10:8673
0:00347041 10:3883 0:00309444 10:9142
0:00341111 10:4951 0:00304739 10:9564
0:00335390 10:5906
of ln .kH,B =p ı / at eleven temperatures in the range 275 K–328 K and at pressures close to 1 bar.
Use these data to evaluate sol,B H ı and sol,B Cpı at T D 298:15 K. This can be done by a
graphical method. Better precision will be obtained by making a least-squares fit of the data to
the three-term polynomial
ln .kH,B =p ı / D a C b.1=T / C c.1=T /2
and using the values of the coefficients a, b, and c for the evaluations.
12.15 Liquid water and liquid benzene have very small mutual solubilities. Equilibria in the binary
water–benzene 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 coefficient.
At a fixed 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 AxB
xB
Here kH,B is the Henry’s law constant and A is a constant related to deviations from Henry’s
law. At 30 ı C, the measured values were kH,B D 385:5 bar and A D 2:24 104 bar.
19 Ref. [144]. 20 Ref. [159].
P ROBLEMS 416
(a) Treat benzene (B) as the solute and find its activity coefficient on a mole fraction basis,
ı 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 ı C is fB D
0:1576 bar. Estimate the mole fraction solubility of liquid benzene in water at this tem-
perature.
(c) The calculation of x;B in part (a) treated the benzene as a single solute species with
deviations from infinite-dilution behavior. Tucker et al suggested a dimerization model to
explain the observed negative deviations from Henry’s 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 define the mole fractions
def nB n
xB D B
nA C nB nA
def nM n def nD n
xM D M xD D D
nA C n M C nD nA nA C n M C n D 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 coefficients
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
xB xM
KD 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, find 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 find 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 Debye–Hückel limiting law (Eq. 10.4.8) to calculate ˙ . You can do this cal-
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 Debye–Hückel limiting law for ˙ .
Compare with the value in part (a).
12.18 Use the following experimental information to evaluate the standard molar enthalpy of for-
mation and the standard molar entropy of the aqueous chloride ion at 298:15 K, based on the
P ROBLEMS 417
conventions f H ı .HC , aq/ D 0 and Smı .HC , aq/ D 0 (Secs. 11.3.2 and 11.8.4). (Your calcu-
lated 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.)
For the reaction 21 H2 .g/ C 12 Cl2 .g/ ! HCl.g/, the standard molar enthalpy of reaction
at 298:15 K measured in a flow calorimeter21 is r H ı D 92:312 kJ mol 1 .
The standard molar entropy of gaseous HCl at 298:15 K calculated from spectroscopic
data is Smı D 186:902 J K 1 mol 1 .
From five calorimetric runs,22 the average experimental value of the standard molar en-
thalpy of solution of gaseous HCl at 298:15 K is sol,B H ı 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 five 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 Debye–Hückel limiting law, calculate the solubility product Ks
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 ı .Cl , aq/ D 131:22 kJ mol 1 and
f G ı .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
1
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 final solutions to infinite dilution,
the standard molar reaction enthalpy at 298:15 K was found to be r H ı D 43:042 kJ mol 1 .
The same workers used solution calorimetry to obtain the molar enthalpy of solution at infinite
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 ı .Cl , aq/ D 167:08 kJ mol 1 and
f H ı .AgCl, s/ D 127:01 kJ mol 1 from Appendix H. (These values come from cal-
culations 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. [146]. 22 Ref. [72]. 23 Ref. [138]. 24 Ref. [67]. 25 Ref. [164].
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 the phase rule may be written either in the form
F D2CC P (13.1.1)
or
F D2Cs r P (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 fixed-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.
418
CHAPTER 13 THE PHASE RULE AND PHASE DIAGRAMS
13.1 T HE G IBBS P HASE RULE FOR M ULTICOMPONENT S YSTEMS 419
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 definite 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 semiper-
meable or impermeable membranes. Furthermore, every conceivable reaction among the
species is either at reaction equilibrium or else is frozen at a fixed 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 fix 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 finite (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 vari-
ables 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 definite 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
system remains in thermal and mechanical equilibrium, but not necessarily in transfer equi-
librium, 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 vari-
ables.
The equilibrium system is, however, in transfer equilibrium, which requires each species
“ ”
to have the same chemical potential in each phase: i D ’i , i D ’i , and so on. There
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 perfectly-
ordered 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 in-
tensive 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 aQreaction equilibrium, the relation is r G D i i i D 0, or the equivalent
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 infinite variety of possible choices of the independent variables (both exten-
sive and intensive) for the equilibrium system, but the total number of independent variables
is fixed at 2 C s r. Keeping intensive properties fixed, we can always vary how much of
each phase is present (e.g., its volume, mass, or amount) without destroying the equilib-
rium. 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 423. 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
1 Relations such as
P P
pi D p for a gas phase or i xi D 1 for a phase in general have already been accounted
i
for in the derivation by the specification of p and the amount of each species.
of freedom, our expression for F based on species is
F D2Cs r P (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 fixed 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 fixed-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 tem-
perature and pressure of the whole system, and for each phase we can independently vary
the mole fraction of all Pbut one of the substances (the value of the omitted mole fraction
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 in-
dependent. 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 , i D ’i , etc.) and
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 Œ2 C P .C 1/ C.P 1/ D 2 C C P (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 componentsPby a reaction, there is an additional mole fraction variable
and also the additional relation i i i D 0 for the reaction; again the changes cancel.
We may need to remove a component from a phase to achieve the final composition.
Note that it is not necessary to consider additional relations for electroneutrality or
initial conditions; they are implicit in the definitions 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 r (13.1.5)
13.1.4 Examples
The five 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 r P
D2C1 0 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, 2H2 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 r 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.
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 r P D2 (13.1.8)
If we wish to calculate F by the components approach, we must decide on the mini-

mum 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 significance of there being two degrees of freedom when the reaction equi-
libria 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.
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 2Cl 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 r 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 definite values.
Now suppose we impose an initial condition by preparing the system with water and
dry air of a fixed 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
.nlN2 C nN2 /
g Da (13.1.13)
.nlO2 C nO2 /
2 Arbitrary,that is, within the limits that would allow the two phases to coexist.
3 The fact that the compositions of both phases depend on the relative amounts of the phases is illustrated in
Prob. 9.5.
13.2 P HASE D IAGRAMS : B INARY S YSTEMS 425
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 fixed composition, as
before, but consider the solubilities of N2 and O2 in water to be negligible. Then nlN2 and
g g
nlO2 are zero and Eq. 13.1.13 becomes nN2 =nO2 D a, or yN2 D ayO2 , which is a relation
between intensive variables. In this case, r is 1 and the phase rule becomes
F D2Cs r 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 fixed 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 five species, one relation (for reaction equilib-
rium), and three phases. The phase rule gives
F D2Cs r P D3 (13.1.15)
It is more difficult 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
diagram, we usually draw a two-dimensional phase diagram that is either a temperature–

composition diagram at a fixed pressure or a pressure–composition diagram at a fixed tem-
perature. The position of the system point on one of these diagrams then corresponds to a
definite 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 compo-
sitions 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 com-
position 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 fixed, so
there are two other independent intensive variables. For example, on a temperature–
composition phase diagram, the pressure is fixed and the temperature and composi-
tion 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 temperature–composition phase diagram) or
constant pressure (on a pressure–composition 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 fixed 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 hori-
zontal 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 Solid–liquid systems

Figure 13.1 on the next page is a temperature–composition phase diagram at a fixed pres-
sure. 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 com-
prise 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
Tf;A
b
b
A(s) + liquid
Tf;B
liquid b
c
b b b
e
T
d f liquid
+ B(s)
g
Te b b
a
b
h
A(s) + B(s)
0 0:2 0:4 0:6 0:8 1:0

A zB B
Figure 13.1 Temperature–composition 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 temper-
ature 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) b–h. The cool-
ing rate depends on the temperature gradient at the system boundary and the system’s 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 e–f. The compositions of the two phases
are given by the values of zB at the ends of the tie line: xBs D 0 for the solid and xBl D 0:50
for the liquid. From the general lever rule (Eq. 8.2.8 on page 209), the ratio of the amounts
in these phases is
nl zB xBs 0:40 0
s
D l
D D 4:0 (13.2.1)
n xB zB 0:50 0:40
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.
260 1500
240 liquid liquid

s + l
l + s
220 s
s
T= K
T= K
CHCl3 (s) + l
1000
200 l + CCl4 (s)
s + s
180
CHCl3 (s) + CCl4 (s)
160 500
0 0:2 0:4 0:6 0:8 1:0 0 20 40 60 80 100
CHCl3 mass fraction CCl4 CCl4 Ag mass percent Cu Cu
(a) (b)
Figure 13.2 Temperature–composition phase diagrams with single eutectics.

(a) Two pure solids and a liquid mixture. a
(b) Two solid solutions and a liquid mixture.
a Ref. [168], 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 first 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 finally 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.
Temperature–composition phase diagrams such as this are often mapped out experimen-
tally 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 temper-
ature 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 temperature–composition phase diagrams with single eutectic
points. The left-hand diagram is for the binary system of chloroform and carbon tetrachlo-
ride, 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 silver–copper 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
50
40 liquid
30
l+B(s)
t=Æ C
A(s)+1 AB(s)+1
20 l+AB(s)
10
A(s)+AB(s) AB(s)+B(s)
0
0 0:2 0:4 0:6 0:8 1:0
A zB B
Figure 13.3 Temperature–composition phase diagram for the binary system of ˛-

naphthylamine (A) and phenol (B) at 1 bar (Ref. [128]).
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 figure 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 ı 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 388). 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 cusp—like 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 NaCl2H2 O is 61.9% NaCl by mass. It decomposes at
0 ı 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 ı 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 ı C, where the phases are ice, the solution at its eutectic point, and the
solid hydrate; and 109 ı 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
120
H2 O(g) C NaCl(s)
100
H2 O(g)
C sln
80
60
sln sln C NaCl(s)
t=Æ C
40
20
H2 O(s) peritectic
C sln
point
0
sln C NaCl 2H2 O(s)
NaCl 2H2 O(s)
20 C NaCl(s)
H2 O(s) C NaCl 2H2 O(s)
40
0 10 20 30 40 50 60 70 80 90 100
H2 O mass percent NaCl NaCl
Figure 13.4 Temperature–composition phase diagram for the binary system of H2 O

and NaCl at 1 bar. (Data from Refs. [37] and [167].)
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 fixed 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 tempera-
ture 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 205) is observed in the
vicinity of this point, caused by large local composition fluctuations. 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 compo-
nent B (carbon disulfide) in a cylindrical vessel and adjust the temperature to 200 K. The
230
bc
220
D1
b
P d
210
T =K

l l
200 b b
c
b
b a
190 P D2
180
0 0:2 0:4 0:6 0:8 1:0
C 3 H6 O2 zB CS2
Figure 13.5 Temperature–composition phase diagram for the binary system of

methyl acetate (A) and carbon disulfide (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 find the two liquid layers have compositions xB’ D 0:20 and xB D 0:92. Since carbon
disulfide is the more dense of the two pure liquids, the bottom layer is phase “, the layer
that is richer in carbon disulfide. According to the lever rule, the ratio of the amounts in the
two phases is given by
n“ z xB’ 0:40 0:20

’
D B“ D D 0:38 (13.2.2)
n xB zB 0:92 0:40
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 disulfide 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 fixed positions, neither phase changes its composition, but the
amount of phase “ increases at the expense of phase ’. The liquid–liquid 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 a–d. 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 liquid–liquid 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 Liquid–gas systems with ideal liquid mixtures

Toluene and benzene form liquid mixtures that are practically ideal and closely obey Raoult’s
law for partial pressure. For the binary system of these components, we can use the vapor
0:8 390
liquidus vaporus
0:6 liquid 380 gas
p=bar g+l g+l
T =K
0:4 370
gas vaporus
0:2 360 liquid
liquidus
0 350
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
benzene zA toluene benzene zA toluene
(a) (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 pres-
sures of the pure liquids.
(a) Pressure–composition diagram at T D 340 K.
(b) Temperature–composition diagram at p D 1 bar.
pressures of the pure liquids to generate the liquidus and vaporus curves of the pressure–
composition and temperature–composition 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
Raoult’s law for partial pressure (Eq. 9.4.2):
pA D xA pA pB D .1 xA /pB (13.2.3)
Strictly speaking, Raoult’s law applies to a liquid–gas system maintained at a constant pres-
sure 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 liquid–gas 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 xA /pB
D pB C .pA 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
function of xA . The mole fraction composition of the gas in the two-phase system is given
by
p xA pA
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 fixed composition. We can calculate the liquid composition by rearranging Eq.
13.2.4:
p pB
xA D (13.2.6)
pA pB
The gas composition is then given by
pA x p
yA D D A A
p p
p pB pA

D (13.2.7)
pA pB p
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 pressure–composition or temperature–composition 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 –p–zA
phase diagram, as shown in Fig. 13.7 on the next page. In this figure, the liquidus surface is
in view at the front and the vaporus surface is hidden behind it.
13.2.5 Liquid–gas systems with nonideal liquid mixtures

Most binary liquid mixtures do not behave ideally. The most common situation is positive
deviations from Raoult’s law.5 Some mixtures, however, have specific A–B 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 Raoult’s law. If the deviations
from Raoult’s 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 435 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 Inthe molecular model of Sec. 11.1.5, positive deviations correspond to a less negative value of kAB than the
average of kAA and kBB .
p=bar
42
0
0
ben0
=K
T
zen
e zA
1
34
tol
0
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 temper-
atures 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 cross-
sections 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 Raoult’s 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 pressure–composition 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 composi-
tions to boil in an equilibrium still at a fixed temperature or pressure. When the liquid
and gas phases have become equilibrated, samples of each are withdrawn for analy-
sis. 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 composi-

tion not corresponding to one of the pure components, which is the case for the methanol–
6 Thisbehavior is consistent with the statement in Sec. 12.8.2 that if one constituent of a binary liquid mixture
exhibits positive deviations from Raoult’s law, with only one inflection point in the curve of fugacity versus
mole fraction, the other constituent also has positive deviations from Raoult’s law.
70 70
liquid
60 p 60 bc
50 50 l+g g+l
40 40
p=kPa
p=kPa
30 30
gas
20 20
pA pB
10 10
0 0
0 0:2 0:4 0:6 0:8 1:0 0 0:2 0:4 0:6 0:8 1:0
xA zA
(a) (b)
Figure 13.8 Binary system of methanol (A) and benzene at 45 ı C. a

(a) Partial pressures and total pressure in the gas phase equilibrated with liquid mix-
tures. The dashed lines indicate Raoult’s law behavior.
(b) Pressure–composition phase diagram at 45 ı C. Open circle: azeotropic point at
zA D 0:59 and p D 60:5 kPa.
a Ref. [157].
benzene system, then the liquid and gas phases are mixtures of identical compositions at
this pressure. This behavior was deduced on page 405 at the end of Sec. 12.8.3. On the
pressure–composition 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 Raoult’s law can have an isothermal liquidus curve with a mini-
mum 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 un-
changed). 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 pressure–composition phase
diagram, then they exhibit a minimum on a temperature–composition phase diagram. This
relation is explained for the methanol–benzene 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
figure is the same as the pressure–composition diagram of Fig. 13.8(b), and the hatched
cross-section at the top of the figure is a temperature–composition 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
69
:5
p=kPa
25
:0
70
ben0
zen
C
t= Æ
e zA
me 1
45
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. [157].
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 temperature–composition phase diagrams of binary systems. If the system does not
form an azeotrope (zeotropic behavior), the equilibrated gas phase is richer in one compo-
nent 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 Solid–gas systems

As an example of a two-component system with equilibrated solid and gas phases, con-
sider the components CuSO4 and H2 O, denoted A and B respectively. In the pressure–
composition phase diagram shown in Fig. 13.12 on page 438, 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
g
g g
T
l
l
l
0 zxAA 1 0 xzAA 1 0 zxA

A 1
(a) (b) (c)
Figure 13.10 Temperature–composition phase diagrams of binary systems exhibit-

ing (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.
g
g g
T
T
lα
lα lβ lα
lβ
lβ
0 zxA
A 1 0 zxA
A 1 0 zxA
A 1
(a) (b) (c)
Figure 13.11 Temperature–composition phase diagrams of binary systems with

partially-miscible liquids exhibiting (a) the ability to be separated into pure compo-
nents 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 1
2 CuSO4 3H2 O.s/ • 2 CuSO4 H2 O.s/ C H2 O.g/
1 1
2 CuSO4 5H2 O.s/ • 2 CuSO4 3H2 O.s/ C H2 O.g/
The equilibria are written above with coefficients that make the coefficient 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.
AB5 (s) sln

3:2
AB3 (s) + AB5 (s)

+ sln sln + B(g)
3:09
3:0 AB5 (s)
+ B(g)
1:05
1:0
p=10 2 bar A(s) + AB(s) AB(s) + AB3 (s) AB3 (s) + B(g)
0:8
0:70
0:6
AB(s) + B(g)
0:4
0:005
0:003
A(s) + B(g)
0
0 1
CuSO4 zB H2 O
Figure 13.12 Pressure–composition phase diagram for the binary system of CuSO4
(A) and H2 O (B) at 25 ı C. a
a Ref. [107]; Ref. [169], 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 three-
phase 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 equi-
librium constant is almost exactly equal to pd =p ı , where pd is the dissociation pressure of
the higher hydrate in the reaction. Thus, a hydrate cannot exist in equilibrium with water va-
por 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 dis-
sociation equilibria is established, the partial pressure pH2 O of H2 O is equal (approx-
imately) 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 phe-
nomenon is called efflorescence (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,
b b
bc
bc
bc
90
80 bc
70 bc
60
p=bar
50 bc
a
40 b
54
30
0:
5
20
50
0
10
45
0
0
0
40
=K
0
eth
T
ane
35
zA
0
1
30
hep
0
tan
e
Figure 13.13 Pressure–temperature–composition behavior in the binary heptane–
ethane system. a The open circles are critical points; the dashed curve is the critical
curve. The dashed line a–b 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 fluid).
If the two-component equilibrium system contains only two phases, it is bivariant cor-
responding 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 fluid (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 figure shows sections of a three-dimensional phase diagram
at five temperatures. Each section is a pressure–composition 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,
40
bc
0
30
0
p=bar 20
0
bc
10
0
29
bc
5
0
29
0
0
=K
He
28
T
5
zA
1
28
0
Xe
Figure 13.14 Pressure–temperature–composition 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 a–b. The system point moves from
the area for a gas phase into the two-phase gas–liquid 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 xenon–helium 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 gas–gas equilibrium, or gas–gas
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 fluids, but are instead high-pressure
fluids 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 441
A A
triangle height
zC
zB
b b
zA
B C B C
(a) (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 two-
dimensional 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 in-
stead 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 oppo-
site 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 figure, 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
filled circle in the figure represents the overall composition zA D 0:20, zB D 0:30, and
b
D E
P
a
B 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 de-
scribed 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.
ethanol
P D1
bc b
c
P D2 b
b
b
a
benzene water
Figure 13.17 Ternary phase diagram for ethanol, benzene, and water at 30 ı 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 figure 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 disap-
pears, 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.
H2 O
sln
sln + NaCl(s) + KCl(s)
NaCl KCl
Figure 13.18 Ternary phase diagram for NaCl, KCl, and water at 25 ı C and 1 bar. a
The dashed lines are tie lines in the two-phase areas.
a Data from Ref. [168], p. 314.
A system of three components and three phases has two degrees of freedom; at fixed
values of T and p, each phase must have a fixed composition. The fixed 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. Al-
though 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 445
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, find 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 equilib-
rium NH4 Cl.s/ • NH3 .g/ C HCl.g/ exists.
(a) Suppose you prepare the system by placing solid NH4 Cl in an evacuated flask 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 equi-
librium constant. Assume that the gas phase is an ideal gas mixture and use the approxi-
mate expression K D pNH3 pHCl =.p ı /2 .
13.4 Consider the lime-kiln process CaCO3 .s/ ! CaO.s/ C CO2 .g/. Find the number of inten-
sive 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 fixed 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 430. Suppose the system contains 36:0 g
(2:00 mol) H2 O and 58:4 g (1:00 mol) NaCl at 25 ı C and 1 bar.
P ROBLEMS 446
(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 ı C.
(c) Describe the changes that occur if the temperature is raised from 25 ı C to 120 ı C at con-
stant pressure.
(d) Describe the system after 200 g H2 O is added at 25 ı C.
Table 13.1 Aqueous solubilities of

sodium sulfate decahydrate and anhy-
drous sodium sulfate a
Na2 SO4 10H2 O Na2 SO4

t =ı C xB t =ı C xB
10 0:011 40 0:058
15 0:016 50 0:056
20 0:024
25 0:034
30 0:048
a Ref. [55], p. 179–180.
13.8 Use the following information to draw a temperature–composition phase diagram for the binary
system of H2 O (A) and Na2 SO4 (B) at p D 1 bar, confining t to the range 20 to 50 ı C and
zB to the range 0–0:2. The solid decahydrate, Na2 SO4 10H2 O, is stable below 32:4 ı 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 ı 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 ı 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 70:0ı C and zA D 0:100
2. t D 50:0ı C and zA D 0:275
13.10 Figure 13.19 on the next page is a temperature–composition phase diagram for the binary
system of water (A) and phenol (B) at 1 bar. These liquids are partially miscible below 67 ı 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 ı 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 447
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 =ı C xA t=ı C xA t =ı C
0.000 0.0 0.119 35.0 0.286 56.0
0.020 10:0 0.143 37.0 0.289 55.0
0.032 20:5 0.157 36.0 0.293 60.0
0.037 27:5 0.173 33.0 0.301 69.0
0.045 40:0 0.183 30.0 0.318 72.5
0.052 55:0 0.195 27.4 0.333 73.5
0.053 41:0 0.213 32.0 0.343 72.5
0.056 27:0 0.222 32.5 0.358 70.0
0.076 0.0 0.232 30.0 0.369 66.0
0.083 10.0 0.238 35.0 0.369 80.0
0.093 20.0 0.259 50.0 0.373 100.0
0.106 30.0 0.277 55.0
a Data from Ref. [55], page 193.
70
60
50
P D2 P D1
t=Æ C
40
30
20
0 0:1 0:2 0:3 0:4
zB
Figure 13.19 Temperature–composition phase diagram for the binary system of wa-
ter (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 ı 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 ı C and that of n-butane is 0:2 ı C. Table 13.3
on the next page lists vapor pressure data for the pure liquids. Assume that the liquid mixtures
obey Raoult’s law.
(a) Calculate the compositions, xA , of the liquid mixtures with boiling points of 10:0 ı C,
20:0 ı C, and 30:0 ı C at a pressure of 1 bar.
(b) Calculate the compositions, yA , of the equilibrium vapor at these three temperatures.
P ROBLEMS 448
Table 13.3 Saturation vapor pressures

of propane (A) and n-butane (B)
t=ı C pA =bar pB =bar

10:0 3:360 0:678
20:0 2:380 0:441
30:0 1:633 0:275
(c) Plot the temperature–composition 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 ı 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 ı C a
0 0 29:89 0:5504 0:3692 35:32
0:0472 0:1467 33:66 0:6198 0:3951 34:58
0:0980 0:2066 35:21 0:7096 0:4378 33:02
0:2047 0:2663 36:27 0:8073 0:5107 30:28
0:2960 0:2953 36:45 0:9120 0:6658 25:24
0:3862 0:3211 36:29 0:9655 0:8252 21:30
0:4753 0:3463 35:93 1:0000 1:0000 18:14
a Ref. [24].
13.12 Use the data in Table 13.4 to draw a pressure–composition phase diagram for the 2-propanol–
benzene system at 45 ı C. Label the axes and each area.
Table 13.5 Liquid and gas compositions in the two-

phase system of acetone (A) and chloroform at 35:2 ı C
a
0 0 39:08 0:634 0:727 36:29
0:083 0:046 37:34 0:703 0:806 38:09
0:200 0:143 34:92 0:815 0:896 40:97
0:337 0:317 33:22 0:877 0:936 42:62
0:413 0:437 33:12 0:941 0:972 44:32
0:486 0:534 33:70 1:000 1:000 45:93
0:577 0:662 35:09
a Ref. [167], p. 286.
13.13 Use the data in Table 13.5 to draw a pressure–composition phase diagram for the acetone–
chloroform system at 35:2 ı 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 equi-
librium or electrochemical equilibrium, however, reaction equilibrium is not necessarily
present—that 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 coefficients 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 solu-
tion, 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
449
CHAPTER 14 GALVANIC CELLS
14.1 C ELL D IAGRAMS AND C ELL R EACTIONS 450
Cu Cu
H2 (g)
AgCl
b b
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 distin-
guish the left and right electrodes. In this way, we establish a left–right 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 silver–silver 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 figure, 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 silver–silver 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 451
The same cell can be described by a slightly different cell diagram that omits the copper
terminals seen in the figure 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 cancella-
tion of the electrons.
For instance, we can write the electrode reactions of the cell of Fig. 14.1 as follows.
oxidation at left: H2 .g/ ! 2 HC .aq/ C 2 e

reduction at right: 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 composi-
tion 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 zinc–copper 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 defined 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 452
e
b b b b
L C R C
Zn Cu Zn Cu
ZnSO4 (aq) CuSO4 (aq) ZnSO4 (aq) CuSO4 (aq)
(a) (b)
Figure 14.2 Zinc–copper galvanic cell with porous barrier (heavy dashed line) sepa-
rating 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 je 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 je j, the advance-
ments of these reactions and of the cell reaction are all described by the same advancement
variable . For an infinitesimal 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 defined 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 five significant figures 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
defined as the change in the electrical potential energy of an infinitesimal test charge when
it is brought to this point from a position infinitely far from other charges, divided by the
charge.
We are concerned with the electric potential within a phase—the inner electric potential,
or Galvani potential. We can measure the difference between the values of this electric po-
tential 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 metal–metal 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.
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 defined by Eq. 3.8.6:
def
Ecell D R L (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 junc-
tion is inevitable.
In a cell without a liquid junction, the reactants of the cell reaction can react di-
rectly 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 zinc–copper 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 flow 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
flow 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.
C C
b b
G
L R L R
C C
galvanic galvanic
cell 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 sponta-
neously. 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 poten-
tial. 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 cell’s 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 high-
impedance digital voltmeter (a voltmeter that draws negligible current) instead of with a
b R
RT
RE
Ecell, eq
RT
b R
I
Ecell
I RE
b L b L
LT LT
LE
LE
(a) (b)
Figure 14.4 Galvani potential profile 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 finite 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 poten-
tial 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 profile 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 profile 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 briefly 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.
Metal–metal contacts
An electric potential difference at an interface between two metals is called a contact poten-
tial. 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
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 con-
tact 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 ’e of free elec-
trons in metal phase ’ on the electric potential ’ of the phase is given by the relation of
Eq. 10.1.6 on page 287, with the charge number zi set equal to 1:
’e ./ D ’e .0/ F ’ (14.2.2)
Here ’e .0/ is the electron chemical potential in a phase with the same intensive properties
as phase ’ but at zero electric potential. ’e .0/ depends only on the temperature and the
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 ’e and e to be equal. We equate ’e and e ,
“
substitute from Eq. 14.2.2 to obtain ’e .0/ F ’ D e .0/ F “ , and rearrange to
“
“ ’ e .0/ ’e .0/
D (14.2.3)
F
(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
potential “ ’ depends only on the temperature and the compositions of the two metals.3
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.
Metal–electrolyte 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 P the cell, the electrode
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 457
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 zinc–copper 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 elec-
trode 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 satisfied:
.Cu/ .Cu2C / 2e .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
specific 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 direc-
tion 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 elec-
tric 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 filled 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 defined by
def
X
r Gcell D i i (14.3.1)
i
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 i i C ze .LE/ D 0 (14.3.2)
i
X
At the right electrode j j ze .RE/ D 0 (14.3.3)
j
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
i i C j j C zŒ e .LE/ e .RE/ D 0 (14.3.4)
i j
The first 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
i i C j j D r Gcell (14.3.5)
i j
Substituting from Eq. 14.3.5 into Eq. 14.3.4 and solving for r Gcell , we obtain
r Gcell D zŒ 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 zŒ e .LT/ e .RT/ (14.3.7)
Making substitutions from Eq. 14.2.2 for e .LT/ and e .RT/, and recognizing that e .0/
is the same in both terminals because they have the same composition, we obtain
r Gcell D zF .R L /
D zFEcell, eq (14.3.8)
We can see from Eq. 14.3.1 that the value of r Gcell has nothing to do with the composi-
tion 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 145,
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-defined chemical composition.
The activity of each reactant and product of the cell reaction therefore has a definite value
in this state.
Now imagine a reaction vessel that has the same temperature and pressure as the gal-
vanic 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 453.
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
P Gibbs energy r G of the direct reaction. Both r Gcell
reaction is equal to the molar reaction
and r G are defined 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 first 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
P vessel during the direct reaction.
When all ions are in the same phase, the value of i i i is independent of the electric
potentials of any of the phases (see the comment following Eq. 11.8.4 on page 350), 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 i i .i / C j j .j / (14.3.10)
i j
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 i i .0/ C j j .0/ C i zi F ’ C j zj F “ (14.3.11)
i j i j
The sum of the first 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 beP the same if the
Pions were in a single phase of any electric potential. Consequently
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 P of the electrode reactions
P at the left
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)
where Ej D “ ’ is the liquid junction potential.

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 :
r G
Ecell, eq D (14.3.13)
zF
(cell without liquid junction)
r G
Ecell, eq D C Ej (14.3.14)
zF
(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 con-
ditions. 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.
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 reac-
tion 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
ı
the standard cell potential of the cell reaction, Ecell, eq . An experimental procedure for
ı
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 ı . From Eq. 14.3.13, or Eq. 14.3.14 with Ej assumed equal to zero, we have
r G ı D ı
zFEcell, eq (14.3.15)
r G ı 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 ı is equal to RT ln K
(Eq. 11.8.10), we can write
zF ı
ln K D E (14.3.16)
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.
ı
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 368:
d ln K
r H ı D RT 2
dT
ı
!
dEcell, eq ı
D zF T Ecell, eq (14.3.17)
dT
14.4 T HE N ERNST E QUATION 462
Finally, by combining Eqs. 14.3.15 and 14.3.17 with r G ı D r H ı Tr S ı , we obtain

an expression for the standard molar reaction entropy:
ı
ı
dEcell, eq
r S D zF (14.3.18)
dT
Because G, H , and S are state functions, the thermodynamic equilibrium constant

ı ı
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 318). During
a reversible cell reaction, dS must equal ¶q=T , and ¶q= d is therefore equal to Tr S ı
during a cell reaction taking place reversibly under standard state conditions at constant T
and p.
14.4 THE NERNST EQUATION

ı
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, ı
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 ı from Eqs. 14.3.13 and Eq. 14.3.15 into r G D r G ı C RT ln Qrxn
(Eq. 11.8.8 on page 351) and solve for Ecell, eq :
ı RT
Ecell, eq D Ecell, eq ln Qrxn (14.4.1)
zF
(no liquid junction, or Ej D0)
Equation 14.4.1 is the Nernst equation for the cell reaction. Q Here Qrxn is the reaction

quotient for the cell reaction defined 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
ı
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 ter-
minals with an external wire, the value of Qrxn becomes equal to the thermodynamic equi-
ı
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 463
ı
The term .RT =zF / ln K is equal to Ecell, eq (Eq. 14.3.16), so Ecell, eq becomes zero—the
cell is “dead” and is incapable of performing electrical work on the surroundings.
At T D298:15 K (25:00 ı C), the value of RT =F is 0:02569 V, and we can write the
Nernst equation in the compact form
ı 0:02569 V
Ecell, eq D Ecell, eq ln Qrxn (14.4.2)
z
(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
2 2 2
aC a aAg
Qrxn D 2
(14.4.3)
aH2 aAgCl
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
ı m =mı , and the hydrogen activity is aH2 D fH2 =p ı . The ion
C mC =m and a D
molalities mC and m are equal to the HCl molality mB . The expression for Qrxn becomes
2 2
C .mB =mı /4 4
˙ .mB =m /
ı 4
Qrxn D D (14.4.4)
fH2 =p ı fH2 =p ı
and the Nernst equation for this cell is

4 ı 4
ı RT ˙ .mB =m /
Ecell, eq D Ecell, eq ln
2F fH2 =p ı
ı 2RT 2RT m RT fH
D Ecell, eq ln ˙ ln Bı C ln ı2 (14.4.5)
F F m 2F p
By measuring Ecell, eq for a cell with known values of mB and fH2 , and with a derived value
ı
of Ecell, eq , we can use this equation to find the mean ionic activity coefficient ˙ of the HCl
solute. This is how the experimental curve for aqueous HCl in Fig. 10.3 on page 296 was
obtained.
We can always multiply each of the stoichiometric coefficients 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 coefficients 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
ı RT .mB =mı /2
Ecell, eq D Ecell, eq ln ˙ (14.4.6)
F .fH2 =p ı /1=2
14.5 E VALUATION OF THE S TANDARD C ELL P OTENTIAL 464
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 ı 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

ı
As we have seen, the value of the standard cell potential Ecell, eq of a cell reaction has useful
ı
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 infinite dilution
where ionic activity coefficients 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
ı 2RT 2RT m RT fH
Ecell, eq D Ecell, eq C ln ˙ C ln Bı ln ı2 (14.5.1)
F F m 2F p
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 coefficient ˙ of the electrolyte. We cannot know the exact
ı
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–
Hückel formula of Eq. 10.4.7 on page 296 is a theoretical expression for ln ˙ that more
closely approximates the actual value the lower is the ionic strength. Accordingly, we define
the quantity
p
A mB fH

0 2RT 2RT m RT
Ecell D Ecell, eq C p C ln Bı ln ı2 (14.5.2)
F 1 C Ba mB F m 2F p
The expression in parentheses is the Debye–Hückel formula for ln ˙ with Im replaced by

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 con-
tributions to .2RT =F / ln ˙ not accounted for by the Debye–Hückel formula. Since these
contributions approach zero in the limit of infinite dilution, the extrapolation of measured
0 ı
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
ı
extrapolated value indicated by the filled 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
ı
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 465
0.2225
bc
bc
b
bc bc
bc bc
bc
0 =V
bc bc bc
0.2220
Ecell
bc bc bc
bc bc
bc
0.2215
0 0.01 0.02 0.03
mB =mol kg 1
0
Figure 14.5 Ecell (defined 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 fit to a linear relation.
a Data from Ref. [73] with f set equal to pH2 and the parameter a set equal to 4:3 10 10 m.
H2
ı
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 electrode—not one we can actually construct in the
laboratory.
A standard electrode potential E ı is defined 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 silver–silver chloride electrode at the right. The
standard electrode potential of the silver–silver 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 poten-
tial of zero, the standard electrode potential of the hydrogen electrode is zero at all temper-
atures. 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 dif-
ferent 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 ELı of electrode L.
14.6 S TANDARD E LECTRODE P OTENTIALS 466
Cell 3 has the hydrogen electrode at the left and electrode R at the right; its standard
cell potential is the standard electrode potential ERı of electrode R.
ı
We wish to calculate the standard cell potential Ecell, eq of cell 1 from the standard electrode
ı ı
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 find 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 infinitesimal 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 ı 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 ı (reaction 1) C r G ı (reaction 2) D r G ı (reaction 3) (14.6.2)

ı
Making the substitution r G ı D zFEcell, eq (Eq. 14.3.15), with the same value of z for
ı ı
each reaction, gives us Ecell, eq C EL D ERı , or
ı ı
Ecell, eq D ER ELı (14.6.3)
ı ı ı
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 appar-
ent 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 difficult to work with experimentally.
Using Eq. 14.6.3 to calculate standard cell potentials from standard electrode potentials
ı
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 ı for
ten electrodes other than E ı D0 for the hydrogen electrode. From these ten values of E ı ,
ı
values of Ecell, eq can be calculated for 35 other cells without hydrogen electrodes.
P ROBLEMS 467
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.
ı ı
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

ı
yielded, at 298:15 K and using a standard pressure of 1 bar, the values Ecell, eq D 0:22217 V
ı 4 1
and dEcell, eq = dT D 6:462 10 V K . (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 ı , r S ı , and r H ı 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 ı .HC , aq/ D 0. If this value is
combined with the value of r H ı 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 ı .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 spectro-
scopic and calorimetric data:
Smı .H2 , g/ D 130:68 J K 1 mol 1 Smı .Cl2 , g/ D 223:08 J K 1 mol 1
Smı .Cl , aq/ D 56:60 J K 1 mol 1 Smı .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-
ı
tial was found to be Ecell, eq D 1:13579 V, and its temperature derivative was found to be
ı
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 468
(b) Use the data to evaluate the standard molar enthalpy of formation and the standard mo-
lar 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
1
Pt H2 (g, f D1 bar) HCl(aq, 0:500 mol kg ) Cl2 (g, f D1 bar) Pt
ı
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 coefficient 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.
ı
(b) At 298:15 K, the standard cell potential of this cell is Ecell, eq D 0:2680 V. Using the value
ı
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 equi-
librium 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 poten-

tial as a function of HBr molality mB .
1
mB / (mol kg ) Ecell, eq / V
0.0004042 0.47381
0.0008444 0.43636
0.0008680 0.43499
0.0013554 0.41243
0.001464 0.40864
0.001850 0.39667
0.002396 0.38383
0.003719 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 469
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 ı , r G ı ,
and r S ı 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.
ı
(e) Evaluate dEcell, eq = dT , the temperature coefficient of the standard cell potential.
A PPENDIX A
D EFINITIONS OF THE SI BASE U NITS
The official definitions 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 hyperfine levels of the ground state of the caesium 133 atom.
This definition 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 thermody-
namic temperature of the triple point of water. This definition refers to water having
the isotopic composition defined 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 specified and may be atoms, molecules, ions,
electrons, other particles, or specified groups of such particles. In this definition, 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 conduc-
tors of infinite 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 monochro-
matic 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.

2 Thisbook 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 Sèvres near Paris, France.
470
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 2010 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 29(27) 1023 mol 1
19
elementary charge e 1.602 176 565(35) 10 C
Faraday constant F 9.648 533 65(21) 104 C mol 1
1 1
gas constant a R 8.314 4621(75) J K mol
7 2
magnetic constant b 0 4 10 NA (exact)
12
electric constant c 0 8.854 187 817: : : 10 C2 J 1
m 1
(exact)
speed of light in vacuum c0 2.997 924 58 108 m s 1
(exact)
2
standard acceleration of free fall d gn 9.806 65 m s (exact)
a or molar gas constant

b or permeability of vacuum
c or permittivity of vacuum
d or standard acceleration of gravity
1 Online at http://physics.nist.gov/cuu/Constants/index.html; IUPAC Green Book (Ref.

[36]), p. 111.
471
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. [120]) except
for quantities followed by an asterisk ( ).
Roman letters
A Helmholtz energy J
As surface area m2
a activity (dimensionless)
B second virial coefficient m3 mol 1
C number of components (dimensionless)
Cp heat capacity at constant pressure JK 1
CV heat capacity at constant volume JK 1
c concentration mol m 3
E energy J
electrode potential V
E electric field strength Vm 1
Ecell cell potential V
Ej liquid junction potential V
Esys system energy in a lab frame J
F force N
number of degrees of freedom (dimensionless)
f fugacity Pa
g acceleration of free fall ms 2
G Gibbs energy J
h height, elevation m
H enthalpy J
H magnetic field strength Am 1
I electric current A
Im ionic strength, molality basis mol kg 1
Ic ionic strength, concentration basis mol m 3
472
A PPENDIX C S YMBOLS FOR P HYSICAL Q UANTITIES 473
(continued from previous page)
K thermodynamic equilibrium constant (dimensionless)

Ka acid dissociation constant (dimensionless)
P
Kp equilibrium constant, pressure basis Pa
Ks solubility product (dimensionless)
kH;i Henry’s law constant of species i,
mole fraction basis Pa
kc;i Henry’s law constant of species i,
concentration basis Pa m3 mol 1
km;i Henry’s law constant of species i,

molality basis Pa kg mol 1
l length, distance m
L relative partial molar enthalpy J mol 1
M molar mass kg mol 1
M magnetization Am 1
Mr relative molecular mass (molecular weight) (dimensionless)
m mass kg
mi molality of species i mol kg 1
N number of entities (molecules, atoms, ions,
formula units, etc.) (dimensionless)
n amount of substance mol
P number of phases (dimensionless)
p pressure Pa
partial pressure Pa
P dielectric polarization Cm 2
Q electric charge C
Qsys charge entering system at right conductor C
Qrxn reaction quotient (dimensionless)
q heat J
Rel electric resistance
S entropy JK 1
s solubility mol m 3
number of species (dimensionless)
T thermodynamic temperature K
t time s
Celsius temperature ıC
U internal energy J
V volume m3
v specific volume m3 kg 1
velocity, speed ms 1
w work J
mass fraction (weight fraction) (dimensionless)
wel electrical work J

w0 nonexpansion work J
x mole fraction in a phase (dimensionless)
Cartesian space coordinate m
y mole fraction in gas phase (dimensionless)
Z compression factor (compressibility factor) (dimensionless)
z mole fraction in multiphase system (dimensionless)
charge number of an ion (dimensionless)
electron number of cell reaction (dimensionless)
Greek letters
alpha
˛ degree of reaction, dissociation, etc. (dimensionless)
cubic expansion coefficient K 1
gamma
surface tension Nm 1, Jm 2
i activity coefficient of species i,

pure liquid or solid standard state (dimensionless)
m;i activity coefficient of species i,
molality basis (dimensionless)
c;i activity coefficient of species i,
concentration basis (dimensionless)
x;i activity coefficient of species i,
mole fraction basis (dimensionless)
˙ mean ionic activity coefficient (dimensionless)
pressure factor (activity of a reference state) (dimensionless)
epsilon
efficiency of a heat engine (dimensionless)
energy equivalent of a calorimeter JK 1
theta
# angle of rotation (dimensionless)
kappa
reciprocal radius of ionic atmosphere m 1
T isothermal compressibility Pa 1
mu
chemical potential J mol 1
JT Joule–Thomson coefficient K Pa 1
nu
number of ions per formula unit (dimensionless)
stoichiometric number (dimensionless)

C number of cations per formula unit (dimensionless)
number of anions per formula unit (dimensionless)
xi
advancement (extent of reaction) mol
pi
˘ osmotic pressure Pa
rho
density kg m 3
tau
torque J
phi
fugacity coefficient (dimensionless)
electric potential V
electric potential difference V
m osmotic coefficient, molality basis (dimensionless)
˚L relative apparent molar enthalpy of solute J mol 1
omega
! angular velocity s 1
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 paren-
theses 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 gas or vapor
l liquid
f fluid (gas or liquid)
s solid
cd condensed phase (liquid or solid)
cr crystalline
mixt mixture
sln solution
aq aqueous solution
aq; 1 aqueous solution at infinite dilution
476
A PPENDIX D M ISCELLANEOUS A BBREVIATIONS AND S YMBOLS 477
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. 59–60).
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 vaporization, evaporation (l ! g)

sub sublimation (s ! g)
fus melting, fusion (s ! l)
trs transition between two phases
mix mixing of fluids
sol solution of a solute in solvent
dil dilution of a solution
ads adsorption
dpl displacement
imm immersion
r reaction in general
at atomization
c combustion reaction
f formation reaction
A PPENDIX D M ISCELLANEOUS A BBREVIATIONS AND S YMBOLS 478
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).
ı standard
pure substance
0 Legendre transform of a thermodynamic potential
1 infinite dilution
id ideal
int integral
E excess quantity
ref 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
d.u=v/ 1 du dv
D v u
dx v2 dx dx
d ln.ax/ 1
D
dx x
d.e ax /
D ae ax
dx
df .u/ df .u/ du
D
dx du 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 vari-
ables
are being used as the independent variables. For example, if the partial derivative is
@f
then f is being treated as a function of y, a, and b.
@y a;b
479
A PPENDIX E C ALCULUS R EVIEW 480
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
x0
P
is the sum i fi xi in the limit as each xi approaches zero and the number of terms in
the sum approaches infinity. 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
x 0
Z x 00 ˇ 00 ˇ
dx ˇx ˇ
D ln ˇˇ 0 ˇˇ
x0 x x
Z x 00
1 00 aC1
x a dx D .x 0 /aC1

.x / (a is a constant other than 1)
x0 aC1
Z x 00
1 ˇˇ ax 00 C b ˇˇ
ˇ ˇ
dx
D ln ˇ 0 (a is a constant)
x 0 ax C b a ax C b ˇ
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 00 2
.x 0 /2

x dx D .x /
x0 2
Z x 00
dx 1 1
2
D
x0 x x 00 x 0
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 coordi-
nates 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 infinitesimal change of f . Since the value

of a state function by definition depends only on the state of the system, integrating df
between an initial state 1 and a final 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 inde-

pendent 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 vari-
ables. 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 in-
dependent 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
481
A PPENDIX F M ATHEMATICAL P ROPERTIES OF S TATE F UNCTIONS 482

@f
slope in f -x plane D
@x y
@f
slope in f -y plane D
f @y x
o 9
@f
dy >
>
) @y x

@f @f
=
df D dx C dy
@f @x @y x
dx> y
>
@x y
;
y
dy
dx
Figure F.1
form
df D a dx C b dy C c dz (F.2.2)
where we can identify the coefficients as

@f @f @f
aD bD cD (F.2.3)
@x y;z @y x;z @z x;y
These coefficients are themselves, in general, functions of the independent variables and
may be differentiated to give mixed second partial derivatives; for example:
@2f @2f

@a @b
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 sufficient condition for df to be an exact differential is
that the reciprocity relation is satisfied for each pair of independent variables.
F.3 INTEGRATION OF A TOTAL DIFFERENTIAL
If the coefficients 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 coefficients 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 inte-
gration.
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

Z x0 Z y0 Z z0
0 2
f D .2x 0/ dx C Œ.x / C 0 dy C .y 0 9z 2 / dz C C
0 0 0
D 0 C x 2 y C .yz 9z 3 =3/ C C
D x 2 y C yz 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 coefficients 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
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 0 0 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 indepen-
dent 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 first example of a Legendre transform, we define 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/ a dx x da
D 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 coefficients 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 (F.4.6)
@y a;z @a y;z
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 cz (F.4.7)
whose total differential is

df2 D df b dy y db c dz z dc
D a dx y db z dc (F.4.8)
Here b has replaced y and c has replaced z as independent variables. Again, we can identify
the coefficients as partial derivatives and write new reciprocity relations.
If we have an algebraic expression for a state function as a function of independent vari-
ables, 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 find 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 find 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 system—for instance, quantum properties of atoms are ignored—the 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 New-
ton’s 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 define the kinetic energy of the collection of particles as the sum for all particles
of 12 mv 2 (where m is mass and v is velocity). We define the potential energy as the sum
over pairwise particle–particle 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 definitions and Newton’s 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 ther-
modynamic system and the remaining particles to constitute the surroundings. The system
particles may interact with an external force field, such as a gravitational field, 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 field or fields. 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 first law of
thermodynamics.
A similar derivation, using a slightly different notation, is given in Ref. [44].
486
APPENDIX G FORCES, ENERGY, AND WORK
G.1 F ORCES BETWEEN PARTICLES 487
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.
Newton’s laws of motion are obeyed only in an inertial reference frame. A reference
frame that is fixed or moving at a constant velocity relative to local stars is practically an
inertial reference frame. To a good approximation, a reference frame fixed relative to the
earth’s 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 Newton’s laws.
It will be assumed that the Cartesian components of all vector quantities appearing in
Sections G.1–G.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 Newton’s second law of motion, the net force Fi acting on particle i is equal
to the product of its mass mi and its acceleration:
dvi
F i D mi (G.1.2)
dt
Here vi is the particle’s 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 defined by the integral3 Z
Wi D Fi dri (G.1.3)
where ri is the position vector of the particle—a 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 definition 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.
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 particle’s displace-
ment is to be integrated overRtime during the time interval. The integral can be written
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 three-
dimensional 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
D 12 mi vi2

(G.1.4)
where vi is the magnitude of the velocity.
The quantity 12 mi vi2 is called the kinetic energy of particle i. This kinetic energy de-
pends only on the magnitude of the velocity (i.e., on the speed) and not on the particle’s
position.
The total work Wtot done by all forcesP acting on all particles during the time interval is
4
the sum of Wi for all particles: Wtot D i Wi . Equation G.1.4 then gives us
!
X X
1 2 1 2

Wtot D 2 mi vi D 2 mi vi (G.1.5)
i 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 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 Newton’s 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
Wtot D Fij dri D
.Fij eij / drij (G.1.8)
i j ¤i i j >i
4 The work Wtot defined here is not the same as the thermodynamic work appearing in the first law of thermo-
dynamics.
5 Ref. [151], p. 95.
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, Newton’s
law of universal gravitation gives the magnitude of a gravitational force as Gmi mj =rij2 ,
where G is the gravitational constant. Coulomb’s law gives the magnitude of an electrical
force between stationary charged particles as Qi Qj =.40 rij2 /, 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 Newton’s 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 Newton’s 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 define a potential function, ˚ij , for this force that will be a contribution to
the potentialR energy. To seeR how ˚ij is related to Fij , we look at Eq. G.1.7. The left-most
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 defined 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 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
1 2
2 mi vi C @ ˚ij A D 0 (G.1.11)
i i j >i
This equation shows that the quantity

X XX
1 2
Etot D 2 mi vi C ˚ij (G.1.12)
i i j >i
is constant over time as the particles move in response to the forces acting on them. The
first 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 particle–particle potential functions; this term is called
G.2 T HE S YSTEM AND S URROUNDINGS 490
bc bc
bc bc
bc b bc b
bc bc
bc b bc b
bc bc
˚ij bc
b ˚ij bc
b
bc b bc b
b b
˚kl b b
˚i k b
b
b
˚i k 00 ˚i k 0 b
b
b
b b b b
b u
b u
b b b u u u
b b b u u u
b u
b b b u u u
b b u u
b b u u
b u
(a) (b)
Figure G.1 Assignment of particles to groups, and some representative particle–

particle potential functions (schematic). The closed dashed curve represents the sys-
tem boundary.
(a) The open circles represent particles in the system, and the filled circles are particles
in the surroundings.
(b) The filled triangles are particles in the surroundings that are the source of a con-
servative force field 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 significance of Eq. G.1.11 is that the total energy Etot defined 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 particle–particle
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 X XX
1 2 1 2
Etot D 2 m i vi C ˚ ij C ˚ ik C 2 m k vk C ˚kl (G.2.1)
i i j >i i k k k l>k
A portion of the surroundings may create a time-independent conservative force field

(an “external” field) for a particle in the system. In order for such a field to be present,
its contribution to the force exerted on the particle and to the particle’s potential energy
must depend only on the particle’s position in the lab frame. The usual gravitational and
electrostatic fields are of this type.
In order to clarify the properties of a conservative external field, the index k 0 will be
used for those particles in the surroundings that are not the source of an external field, and
k 00 for those that are, as indicated
P in Fig. G.1(b). Then the force exerted on system particle
field
i due to the field is Fi D k 00 Fik 00 . If this were the only force acting on particle i ,
the change in its kinetic energy during a time interval would be Fifield dri (Eq. G.1.4).
R
G.2 T HE S YSTEM AND S URROUNDINGS 491
For conservation of energy, the potential energy change in the time interval should have the
same magnitude and the opposite sign:
Z
field
˚i D Fifield dri (G.2.2)
Only if the integral Fifield dri has the same value for all paths between the initial and
R
final positions of the particle does a conservative force field exist; otherwise the concept of
a potential energy ˚ifield is not valid.
Taking a gravitational field as an example of aR conservative external field, we replace
grav grav grav grav
Fifield and ˚ifield by Fi and ˚i : ˚i D Fi dri . The gravitational force on
particle i is, from Newton’s first law, the product of the particle mass and its acceleration
grav
gez in the gravitational field: 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
earth’s radius, so that each system particle experiences essentially the same constant value
of g.) Thus we can define the gravitational potential energy of particle i, which is a function
grav
only of the particle’s 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 ˚ik 0 C ˚ifield (G.2.4)
i k i k0 i
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
1 2
Etot D 4 2 mi vi C ˚ij C ˚ifield 5
i i j >i i
" # " #
XX X XX
1 2
C ˚ik 0 C 2 mk vk C ˚kl (G.2.5)
i k0 k k l>k
The terms on the right side of this equation are shown grouped with brackets into three
quantities. The first 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
1 2
Esys D m v
2 i i C ˚ ij C ˚ifield (G.2.6)
i i j >i 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 field.
G.3 S YSTEM E NERGY C HANGE 492
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 ˚ik 0 C Esurr (G.2.7)
i k0
P P
The quantity i k 0 ˚ik 0 represents potential energy shared by both the system and sur-
roundings on account of forces acting across the system boundary, other than gravita-
tional
P Pforces or forces from other external fields. The forces responsible for the quantity
i k 0 ˚ik 0 are generally significant 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 Fifield C Fisur (G.3.1)
j ¤i
where Fij isP the force exerted on particle i by particle j , both particles being in the system,
and Fisur D k 0 Fik 0 is the net force exerted on particle i by the particles in the surround-
ings that are not the source of an external field. 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 Fifield dri C Fisur dri (G.3.2)
i i j ¤i i i
We can replace the first three sums in this equation with new expressions. Using Eq.
G.1.4, we have !
XZ X
1 2
Fi dri D 2 mi vi (G.3.3)
i i
From Eqs. G.1.8 and G.1.9 we obtain

0 1
XXZ XX
Fij dri D @ ˚ij A (G.3.4)
i j ¤i i j >i
where the sums are over the system particles. From Eq. G.2.2 we can write
!
XZ X
Fifield dri D ˚ifield (G.3.5)
i i
G.4 M ACROSCOPIC W ORK 493
Combining Eqs. G.3.2–G.3.5 and rearranging, we obtain

0 1
XZ X XX X
1 2
Fisur dri D @ 2 mi vi C ˚ij C ˚ifield A (G.3.6)
i i i j >i i
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 field. Thus Esys is equal to
the total work done on the system by the surroundings, other than work done by an external
field such as a gravitational field.
It might seem strange that work done by an external field is not included in Esys . The
reason it is not included is that ˚ifield was defined 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 field. 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 fluid 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 define 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 494
rs ri
b
ri
R
bc
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 fixed in system boundary at
segment ; filled circle: particle i.
to a point fixed 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 Fsur be the totalPcontact force exerted by the surroundings on the system particles in
segment : Fsur 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 i
XZ XXZ
D Fsur dR C •i Fisur dri (G.4.1)
i
We recognize the integral Fsur dR as the macroscopic work at surface element ,
R
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 Fsur dR (G.4.2)

Heat, qlab , can be defined 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 (G.4.3)
with wlab given by the expression in Eq. G.4.2 and qlab given by
XXZ
qlab D •i Fisur dri (G.4.4)
i
G.5 T HE W ORK D ONE ON THE S YSTEM AND S URROUNDINGS 495
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 redefinition 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 fluid contacts a moving piston or paddle.
Suppose, however, that the system is stationary and an interacting part of the surround-
ings 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 con-
stant? 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 flow 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 Pthan
P from a time-independent
external field, and is represented in Eq. G.2.7 by the sum i k 0 ˚ik 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 system’s 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 define 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
particle i (see Fig. G.3). The velocity of the particle in the local frame is vi0 D dri0 = dt.
We continue to treat the earth-fixed 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. 47–56.
7 The ambiguity can be removed by redefining 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 system’s extensive properties are essentially unaffected. With this redefined boundary, the energy
transfer across the boundary is entirely by means of heat.
G.6 T HE L OCAL F RAME AND I NTERNAL E NERGY 496
rs ri
b
R0
z0
ri0 ri
ut y0
R
z
x0
Rloc
bc 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 fixed in system
boundary at segment ; filled 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 finite acceleration
and is noninertial. In this case the motion of particle i in the local frame does not obey
Newton’s second law as it does in an inertial frame. We can, however, define an effective
net force Fieff whose relation to the particle’s acceleration in the local frame has the same
form as Newton’s 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 particle’s motion obey Newton’s second law even though the frame is not inertial.
ThePnet force on particle i from interactions with other particles is given by Eq. G.3.1:
Fi D j ¤i Fij C Fifield 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
fictitious force introduced to make it possible to write Eq. G.6.1 which resembles Newton’s
second law. The motion of particle i in an inertial frame is given by mi dvi = dt D Fi ,
whereas the motion in the local frame is given by mi dvi0 = 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 fix
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 object’s 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 fictitious centrifugal force having
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.1–G.4 can be applied to an arbi-
trary local frame. To carry out the derivations we replace Fi by Fieff , ri by ri0 , and vi by vi0 ,
and use the local frame to measure the Cartesian components of all vectors. We need two
new potential energy functions for the local frame, defined by the relations
Z
0 def
˚i field D Fifield dri0 (G.6.3)
Z
def
˚iaccel D Fiaccel dri0 (G.6.4)
0
Both ˚i field 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.1–G.4 will not be repeated here, but the reader
can verify the following results. The total energy
P Pof the system0 and surroundings measured
0
in the local frame is given by Etot D U C i k 0 ˚ik 0 C Esurr where the index k 0 is for
particles in the surroundings that are not the source of an external field for the system. The
energy of the system (the internal energy) is given by
X XX X 0 X
1 0 2
U D 2 mi .vi / C ˚ij C ˚i field C ˚iaccel (G.6.5)
i i j >i i i
where the indices i and j are for system particles. The energy of the surroundings measured
in the local frame is
X XX X
0 1 0 2
Esurr D m
2 k k.v / C ˚ kl C ˚kaccel (G.6.6)
k k l>i k
where k and l are indices for particles in the surroundings. The value of Etot0 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 (G.6.7)
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.6.8)
i
XZ
wD Fsur dR0 (G.6.9)

In these equations R0 is a vector from the origin of the local frame to a point fixed in the
system boundary at segment , and ri is a vector from this point to particle i (see Fig.
G.3).
We expect that an observer in the local frame will find 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-fixed 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 first 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 defined 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:
The equation w D Fsur dR0 has been derived for a closed system.
P R
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 Fsur exerted by the surroundings
and the displacement dR0 in the local frame.
We could equally well calculate w from the force exerted by the system on the sur-
sys
roundings. According to Newton’s third law, the force F exerted by segment has
sys
the same magnitude as F and the opposite direction: F D Fsur .
sur
During a process, a point fixed in the system boundary at segment is either station-
ary 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 infinitesimal quantity Fsur dR0 in the form Fsur cos ˛ ds , where Fsur is the
magnitude of the force, ds is an infinitesimal 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 Fsur cos ˛ ds ¶w D Fsur cos ˛ ds (G.6.10)

If only one portion of the boundary moves in the local frame, and this portion has
linear motion parallel to the x 0 axis, we can replace Fsur dR0 by Fxsur 0
0 dx , where
sur 0
Fx 0 is the x component of the force exerted by the surroundings on the moving
boundary and dx 0 is an infinitesimal displacement of the boundary. In this case we
can write the following integrated and differential forms of the work:
Z
0 0
w D Fxsur 0 dx ¶w D Fxsur
0 dx (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.6.9–G.6.11, we do not include forces from an
external field such as a gravitational field, or fictitious forces Fiaccel if present.
G.7 N ONROTATING L OCAL F RAME 499
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 infinitesimal 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 R0 has the same Cartesian components in the lab frame as in the nonrotating
local frame, so we can write R R0 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 R0 / equal to Rloc , we obtain the relation
!
XZ Z X
sur 0 sur
w wlab D F d.R R / D F dRloc (G.7.2)

The sum Fsur is the net contact force exerted on the system by the surroundings.
P
For example, suppose the system is a fluid in a gravitational field. Let the system boundary
be at the inner walls of the container, and let the local frame be fixed 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 fluid on the container wall is equal in
magnitude and opposite in direction to the contact force exerted by the surroundings on the
fluid. 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 mea-
sured 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 500
ri0 b
ut
ri
Rcm
bc
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; filled 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 defined
by P
i mi r i def
Rcm D (G.8.1)
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):
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
vi D vcm C vi0 (G.8.3)
where the vector vcm gives theP velocity of the P
center of mass in the lab
Pframe. Substitution
from Eq. G.8.2 into the sum Pi mi ri gives i mi ri D mRcm C i mi ri0 , and a rear-
rangementP of Eq.0 G.8.1 gives i mi ri D mRcm . Comparing these two equalities, we see
the sum i mi ri must be zero. Therefore the first and second derivatives of i mi ri0 with
P
respect to time must also be zero:
X X dvi0
mi vi0 D 0 mi D0 (G.8.4)
dt
i i
From Eqs. G.1.2, G.6.1, G.6.2, and G.8.3 we obtain
d.vi0 vi / dvcm
Fiaccel D mi D mi (G.8.5)
dt dt
Equation G.8.5 is valid only for a nonrotating cm frame.
G.8 C ENTER - OF - MASS L OCAL F RAME 501
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
1 0 2 1 2
U Esys D 2 mi .vi / 2 mi vi
i i
! !
0
X X X
field
C ˚i ˚ifield C ˚iaccel (G.8.6)
i i i
We will find new expressions for the three terms on the right side of this equation.
The first 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
1 2 1 0 0
m v
2 i i D 2 mi .vcm C vi / .vcm C vi /
i i
!
2
X 2 X
D 1
2 mvcm C 1
2 mi vi0 C vcm mi vi0 (G.8.7)
i i
2
The quantity 21 mvcm is the bulk kinetic energy of the system in the lab frame—that is, the
translational energy of a Pbody having the same mass as the system and moving with the
center of mass. The sum i mi vi0 is zero (Eq. G.8.4). Therefore the first term on the right
side of Eq. G.8.6 is
" #
X X
1 0 2 1 2 2
D 21 mvcm

2 mi .vi / 2 mi vi (G.8.8)
i 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
˚i field
˚ifield
D Fifield d.ri0 ri /
i i i
Z !
X
D Fifield dRcm (G.8.9)
i
grav
Suppose the only external field is gravitational: Fifield 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 field ˚ifield D mi gez dRcm
i i i
Z Z
D mg ez dRcm D mg dzcm
D mgzcm (G.8.10)
G.9 ROTATING L OCAL F RAME 502
ri
b
ei
zi
y 0
# y
x
x 0
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 filled 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 system’s bulk gravitational potential energy in the lab frame—the change
in the potential energy of a body of mass m undergoing the same change in elevation as the
system’s 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
0
˚iaccel
D Fi accel
dri D mi cm dri0
dt
i i i
! !
dri0
Z X Z X
0
D mi dv
cm D mi vi dvcm (G.8.11)
dt
i i
The sum i mi vi0 in the integrand of the last integral on the right side is zero (Eq. G.8.4)
P
so the integral is also zero.
2
With these substitutions, Eq. G.8.6 becomes U Esys D 21 m vcm

mgzcm .
Since U Esys is equal to w wlab when the local frame is nonrotating (Eq. G.7.1), we
have
2
w wlab D 12 m vcm

mgzcm (G.8.12)
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 503
x 0 axis of the local frame. There is a gravitational force in the z direction; this force is
responsible for the only external field, 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)
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
Ficentr D mi ! 2 ri ei (G.9.2)
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
to the rotating frame. This force has magnitude 2mi !vi0 and is directed perpendicular to
both vi0 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 fixed relative to the earth’s surface is an inertial frame. This is not exactly true, be-
cause 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 Newton’s second law to be obeyed exactly in the lab frame.9
The earth’s movement around the sun makes only a minor contribution to these correc-
tion 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 significant correction the centrifugal force on the particle from
the earth’s spin about its axis. This force is directed perpendicular to the earth’s axis and has
magnitude mi ! 2 ri , where ! is the earth’s angular velocity, mi is the particle’s mass, and
ri is the radial distance of the particle from the earth’s axis. The correction can be treated
as a small modification 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
earth’s surface. Since Fi depends only on position, we can treat gravity as a conservative
force field in the earth-fixed lab frame.
8 The derivation, using a different notation, can be found in Ref. [108], Chap. 10.
9 Ref. [68], Sec. 4–9.
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 ı C) and the standard pres-
sure p ı 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 Sci-
ence and Technology (CODATA) to establish a set of recommended, internally consistent
values of thermodynamic properties. The values of f H ı and Smı shown with uncertainties
are values recommended by CODATA.1
f H ı Smı f G ı
Species
kJ mol 1 J K 1 mol 1 kJ mol 1
Inorganic substances
Ag(cr) 0 42.55 ˙ 0:20 0
AgCl(cr) 127.01 ˙ 0:05 96.25 ˙ 0:20 109.77
C(cr, graphite) 0 5.74 ˙ 0:10 0
CO(g) 110.53 ˙ 0:17 197.660 ˙ 0:004 137.17
CO2 (g) 393.51 ˙ 0:13 213.785 ˙ 0:010 394.41
Ca(cr) 0 41.59 ˙ 0:40 0
CaCO3 (cr, calcite) 1206.9 92.9 1128.8
CaO(cr) 634.92 ˙ 0:90 38.1 ˙ 0:4 603.31
Cl2 (g) 0 223.081 ˙ 0:010 0
F2 (g) 0 202.791 ˙ 0:005 0
H2 (g) 0 130.680 ˙ 0:003 0
HCl(g) 92.31 ˙ 0:10 186.902 ˙ 0:005 95.30
HF(g) 273.30 ˙ 0:70 173.779 ˙ 0:003 275.40
HI(g) 26.50 ˙ 0:10 206.590 ˙ 0:004 1.70
H2 O(l) 285.830 ˙ 0:040 69.95 ˙ 0:03 237.16
H2 O(g) 241.826 ˙ 0:040 188.835 ˙ 0:010 228.58
H2 S(g) 20.6 ˙ 0:5 205.81 ˙ 0:05 33.44
1 Ref. [38]; also available online at http://www.codata.org.
504
A PPENDIX H S TANDARD M OLAR T HERMODYNAMIC P ROPERTIES 505

f H ı Smı f G ı
Species
Hg(l) 0 75.90 ˙ 0:12 0

Hg(g) 61.38 ˙ 0:04 174.971 ˙ 0:005 31.84
HgO(cr, red) 90.79 ˙ 0:12 70.25 ˙ 0:30 58.54
Hg2 Cl2 (cr) 265.37 ˙ 0:40 191.6 ˙ 0:8 210.72
I2 (cr) 0 116.14 ˙ 0:30 0
K(cr) 0 64.68 ˙ 0:20 0
KI(cr) 327.90 106.37 323.03
KOH(cr) 424.72 78.90 378.93
N2 (g) 0 191.609 ˙ 0:004 0
NH3 (g) 45.94 ˙ 0:35 192.77 ˙ 0:05 16.41
NO2 (g) 33.10 240.04 51.22
N2 O4 (g) 9.08 304.38 97.72
Na(cr) 0 51.30 ˙ 0:20 0
NaCl(cr) 411.12 72.11 384.02
O2 (g) 0 205.152 ˙ 0:005 0
O3 (g) 142.67 238.92 163.14
P(cr, white) 0 41.09 ˙ 0:25 0
S(cr, rhombic) 0 32.054 ˙ 0:050 0
SO2 (g) 296.81 ˙ 0:20 248.223 ˙ 0:050 300.09
Si(cr) 0 18.81 ˙ 0:08 0
SiF4 (g) 1615.0 ˙ 0:8 282.76 ˙ 0:50 1572.8
SiO2 (cr, ˛-quartz) 910.7 ˙ 1:0 41.46 ˙ 0:20 856.3
Zn(cr) 0 41.63 ˙ 0:15 0
ZnO(cr) 350.46 ˙ 0:27 43.65 ˙ 0:40 320.48
Organic compounds
CH4 (g) 74.87 186.25 50.77
CH3 OH(l) 238.9 127.2 166.6
CH3 CH2 OH(l) 277.0 159.9 173.8
C2 H2 (g) 226.73 200.93 209.21
C2 H4 (g) 52.47 219.32 68.43
C2 H6 (g) 83.85 229.6 32.00
C3 H8 (g) 104.7 270.31 24.3
C6 H6 (l, benzene) 49.04 173.26 124.54
Ionic solutes
AgC (aq) 105.79 ˙ 0:08 73.45 ˙ 0:40 77.10
CO3 2 (aq) 675.23 ˙ 0:25 50.0 ˙ 1:0 527.90
Ca2C (aq) 543.0 ˙ 1:0 56.2 ˙ 1:0 552.8
Cl (aq) 167.08 ˙ 0:10 56.60 ˙ 0:20 131.22
F (aq) 335.35 ˙ 0:65 13.8 ˙ 0:8 281.52
A PPENDIX H S TANDARD M OLAR T HERMODYNAMIC P ROPERTIES 506

f H ı Smı f G ı
Species
HC (aq) 0 0 0
HCO3 (aq) 689.93 ˙ 2:0 98.4 ˙ 0:5 586.90
HS (aq) 16.3 ˙ 1:5 67 ˙ 5 12.2
HSO4 (aq) 886.9 ˙ 1:0 131.7 ˙ 3:0 755.4
Hg2 2C (aq) 166.87 ˙ 0:50 65.74 ˙ 0:80 153.57
I (aq) 56.78 ˙ 0:05 106.45 ˙ 0:30 51.72
KC (aq) 252.14 ˙ 0:08 101.20 ˙ 0:20 282.52
NH4 C (aq) 133.26 ˙ 0:25 111.17 ˙ 0:40 79.40
NO3 (aq) 206.85 ˙ 0:40 146.70 ˙ 0:40 110.84
NaC (aq) 240.34 ˙ 0:06 58.45 ˙ 0:15 261.90
OH (aq) 230.015 ˙ 0:040 10.90 ˙ 0:20 157.24
S2 (aq) 33.1 14.6 86.0
SO4 2 (aq) 909.34 ˙ 0:40 18.50 ˙ 0:40 744.00
A PPENDIX I
A NSWERS TO S ELECTED P ROBLEMS
3.3 (b) q D w D 1:00 105 J 7.12 H D 2:27 104 J, S D 43:6 J K 1
3.4 (c) w D 1:99 103 J, q D 1:99 103 J. ı

7.13 (a) Cp;m D 42:3 J K 1
mol 1
3.5 0:0079% 1 1
(b) Cp;m 52:0 J K mol
3.6 (c) V2 ! nRV1 =.CV C nR/, T2 ! 1. 1 1
7.14 (a) 2:56 J K g
For CV D .3=2/nR, V2 =V1 ! 0:4.
7.15 (b) f D 17:4 bar
3.11 9:58 103 s (2 hr 40 min)
7.16 (a) D 0:739, f D 148 bar
1
4.4 S D 0:054 J K
5
(b) B D 7:28 10 m3 mol 1
1
R D 549 J K for 1both processes;
4.5 S
¶q=Text D 333 J K and 0. 8.2 (a) Smı (l) D 253:6 J K 1
mol 1

3=2 V
(b) vap S ı D 88:6 J K 1 mol 1 ,
5.4 (a) S D nR ln cT b C vap H ı D 2:748 104 J mol 1
n
5 3
nR 8.4 4:5 10 bar
2
1
8.5 19 J mol
5.5 (a) q D 0, w D 1:50 104 J,
U D 1:50 104 J, 8.6 (a) 352:82 K
H D 2:00 104 J
(b) 3:4154 104 J mol 1
1
(c) S D 66:7 J K
8.7 (a) 3:62 103 Pa K 1
1 1
6.1 Sm 151:6 J K mol
(b) 3:56 103 Pa K 1
4 1
7.6 (a) ˛ D 8:519 10 K
t D 4:671 10 5 bar 1 (c) 99:60 ı C
1
.@p=@T /V D 18:24 bar K 8.8 (b) vap H ı D 4:084 104 J mol 1
.@U=@V /T D 5437 bar
8.9 0:93 mol
(b) p 1:8 bar
9.2 (b) VA .xB D 0:5/ 125:13 cm3 mol 1
7.7 (b) .@Cp;m =@p/T VB .xB D 0:5/ 158:01 cm3 mol 1
8 1 1 1
D 4:210 10 JK Pa mol VB1 157:15 cm3 mol 1
4 1
7.8 (b) 8 10 K
9.4 real gas: p D 1:9743 bar
7.11 5:001 10 J3 ideal gas: p D 1:9832 bar
507
A PPENDIX I A NSWERS TO S ELECTED P ROBLEMS 508
6
9.5 (a) xN2 D 8:83 10 11.7 (a) State 1:
xO2 D 4:65 10 6 nC6 H14 D 7:822 10 3 mol
yN2 D 0:763 nH2 O D 0:05560 mol
yO2 D 0:205 amount of O2 consumed: 0:07431 mol
State 2:
6
(b) xN2 D 9:85 10 nH2 O D 0:11035 mol
6
xO2 D 2:65 10 nCO2 D 0:04693 mol
yN2 D 0:851 mass of H2 O=1:9880 g
yO2 D 0:117
(b) Vm (C6 H14 ) D 131:61 cm3 mol 1
9.7 (b) fA D 0:03167 bar, fA D 0:03040 bar
Vm (H2 O) D 18:070 cm3 mol 1
9.8 (a) In the mixture of composition
xA D 0:9782, the activity coefficient is (c) State 1: V (C6 H14 ) D 1:029 cm3
B 11:5. V (H2 O) D 1:005 cm3
V g D 348:0 cm3
9.9 (d) kH,A 680 kPa State 2:
9.11 Values for mB =mı D 20: A D 1:026, V (H2 O) D 1:994 cm3
m;B D 0:526; the limiting
V g D 348:0 cm3
slopes are d A = d.mB =mı / D 0,
d m;B = d.mB =mı / D 0:09 (d) State 1:
nO2 D 0:429 mol
9.13 pN2 D 0:235 bar State 2:
yN2 D 0:815 nO2 D 0:355 mol
pO2 D 0:0532 bar yO2 D 0:883
yO2 D 0:185 yCO2 D 0:117
p D 0:288 bar
9.14 (b) h D 1:2 m (e) State 2:
p2 D 27:9 bar
9.15 (a) p.7:20 cm/ p.6:95 cm/ D 1:2 bar pO2 D 24:6 bar
pCO2 D 3:26 bar
(b) MB D 187 kg mol 1
mass binding ratio D 1:37
(f) fH2 O .0:03169 bar/ D 0:03164 bar
10.2 ˙ D 0:392 State 1: fH2 O D 0:03234 bar
State 2: fH2 O D 0:03229 bar
11.1 r H ı D 63:94 kJ mol 1
K D 4:41 10 2 (g) State 1:

ı
11.2 (b) f H : no change H2 O D 0:925
f S ı : subtract 0:219 J K 1
mol 1 O2 D 0:981
f G ı : add 65 J mol 1 fO2 D 29:4 bar
State 2:
6
11.3 p.298:15 K/ D 2:6 10 bar H2 O D 0:896
7
p.273:15 K/ D 2:7 10 bar O2 D 0:983
11.4 (a) 240:34 kJ mol 1
, 470:36 kJ mol 1
, CO2 D 0:910
230:02 kJ mol 1 fO2 D 24:2 bar
fCO2 D 2:97 bar
1
(b) 465:43 kJ mol
1 (h) State 1:
(c) 39:82 kJ mol g
nH2 O D 5:00 10 4 mol
11.5 H D 0:92 kJ nlH2 O D 0:05510 mol
g 4
11.6 LA D 0:405 J mol 1 State 2: nH2 O D 5:19 10 mol
LB D 0:810 kJ mol 1 nlH2 O D 0:10983 mol
(i) State 1: 12.6 sol,B H ı =kJ mol 1 D 3:06; 0; 6:35

km;O2 D 825 bar kg mol 1 sol,B S ı =J K 1 mol 1
nO2 D 3:57 10 5 mol D 121:0; 110:2; 88:4
State 2: 12.7 (a) m’C D m’ D 1:20 10 3 mol kg 1
km;O2 D 823 bar kg mol 1
1 m“C D 1:80 10 3 mol kg 1
km;CO2 D 30:8 bar kg mol
m“ D 0:80 10 3 mol kg 1
nO2 D 5:85 10 5 mol
mP D 2:00 10 6 mol kg 1
nCO2 D 1:92 10 4 mol
12.8 (a) p l D 2:44 bar
(j) H2 O vaporization: U D C20:8 J
H2 O condensation: U D 21:6 J (b) f .2:44 bar/ f .1:00 bar/
D 3:4 10 5 bar
(k) O2 dissolution: U D 0:35 J
O2 desolution: U D 0:57 J 12.10 (a) xB D 1:8 10 7
CO2 desolution: U D 3:32 J mB D 1:0 10 5 mol kg 1
(b) sol,B H ı D 1:99 104 J mol 1

(l) C6 H14 (l) compression:
U D 1:226 J (c) K D 4:4 10 7
solution compression: r H ı D 9:3 kJ mol 1
U D 0:225 J
12.13 (a) p D 92399:6 Pa, yB D 0:965724
solution decompression:
U D 0:414 J (b) A D 0:995801
(m) O2 compression: U D 81 J (c) fA D 3164:47 Pa
gas mixture: dB= dT D
(d) yB D 0:965608
0:26 10 6 m3 K 1 mol 1
gas mixture expansion: U D 87 J (e) Z D 0:999319
(n) U D 8 J (f) p D 92347:7 Pa
(o) c U ı D 4154:4 kJ mol 1 (g) kH,B D 4:40890 109 Pa

12.15 (a) D 0:9826
(p) c H ı D 4163:1 kJ mol 1 x;B
4
(b) xB D 4:19 10
11.8 f H ı D 198:8 kJ mol 1
6
12.16 K D 1:2 10
11.9 T2 D 2272 K
12.17 (a) ˛ D 0:129
5
11.10 p(O2 ) D 2:55 10 bar mC D 1:29 10 3
mol kg 1
11.11 (a) K D 3:5 1041 (b) ˛ D 0:140

(b) pH2 D 2:8 10 42
bar 12.18 f H ı .Cl , aq/ D 167:15 kJ mol 1
NH2 D 6:9 10 17 Smı .Cl , aq/ D 56:46 J K 1 mol 1

10
(c) t D 22 s 12.19 (a) Ks D 1:783 10
12.20 (a) r H ı D 65:769 kJ mol 1
11.12 (b) p 1:5 104 bar
(b) f H ı (AgC , aq) D 105:84 kJ mol 1
11.13 (c) K D 0:15
13.1 (a) F D 4
12.1 (b) T D 1168 K
r H ı D 1:64 105 J mol 1 (b) F D 3
12.4 Kf D 1:860 K kg mol 1 (c) F D 2

Kb D 0:5118 K kg mol 1 13.10 (a) xB (top) D 0:02, xB (bottom) D 0:31
4 1
12.5 MB 5:6 10 g mol (b) nA D 2:1 mol, nB D 1:0 mol
14.3 (a) r G ı D 21:436 kJ mol 1 14.7 (b) f G ı D 210:72 kJ mol 1
r S ı D 62:35 J K 1 mol 1
18
r H ı D 40:03 kJ mol 1 (c) Ks D 1:4 10
(b) f H ı .AgCl, s/ D 127:05 kJ mol 1
14.8 E ı D 0:071 V
(c) Smı .AgCl, s/ D 96:16 J K 1 mol 1 ı
f S ı .AgCl, s/ D 57:93 J K 1 mol 1 14.9 (c) Ecell, eq D 1:36 V
f G ı .AgCl, s/ D 109:78 kJ mol 1
(d) In the cell:
14.4 (b) f H ı .AgCl, s/ D 126:81 kJ mol 1 ¶q= d D 2:27 kJ mol 1
f G ı .AgCl, s/ D 109:59 kJ mol 1 In a reaction vessel:
1
14.5 Ks D 1:76 10 10 ¶q= d D 259:67 kJ mol
ı 5 1
14.6 (b) ˙ D 0:756 (e) dEcell, eq = dT D 3:9 10 VK
B IBLIOGRAPHY
[1] C. J. Adkins, Equilibrium Thermodynamics, 3rd edition. Cambridge University Press,

Cambridge, 1983.
[2] Robert A. Alberty, “Use of Legendre Transforms in Chemical Thermodynamics (IUPAC
Technical Report).” Pure Appl. Chem., 73, 1349–1380 (2001).
[3] S. Angus, B. Armstrong, and K. M. de Reuck, International Thermodynamic Tables of the
Fluid State. Vol. 3. Carbon Dioxide. Pergamon Press, Oxford, 1976.
[4] Roger G. Bates and Vincent E. Bower, “Standard Potential of the Silver–Silver-Chloride
Electrode from 0ı to 95ı C and the Thermodynamic Properties of Dilute Hydrochloric Acid
Solutions.” J. Res. Natl. Bur. Stand. (U.S.), 53, 283–290 (1954).
[5] 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, 1991–2001 (1919).
[6] Robert P. Bauman, “Work of Compressing an Ideal Gas.” J. Chem. Educ., 41, 102–104
(1964).
[7] Robert P. Bauman, “Maximum Work Revisited (letter).” J. Chem. Educ., 41, 675 (1964).
[8] Robert P. Bauman, “Maximum Work Revisited (letter).” J. Chem. Educ., 41, 676–677
(1964).
[9] Robert P. Bauman and Howard L. Cockerham III, “Pressure of an Ideal Gas on a Moving
Piston.” Am. J. Phys., 37, 675–679 (1969).
[10] 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 321–367. Royal Society of Chemistry,
Cambridge, 2010.
[11] Dor Ben-Amotz and J. M. Honig, “Rectification of Thermodynamic Inequalities.” J. Chem.
Phys., 118, 5932–5936 (2003).
[12] 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, 791–804 (1980).
[13] K. H. Berry, “NPL-75: A Low Temperature Gas Thermometry Scale from 2.6 K to 27.1 K.”
Metrologia, 15, 89–115 (1979).
[14] Gary L. Bertrand, “Thermodynamic Calculation of Work for Some Irreversible Processes.”
J. Chem. Educ., 82, 874–877 (2005).
[15] 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, 991–997
(1935).
[16] M. Boirac, “Nécrologie: François Raoult, Discours de M. Boirac, Recteur.” Ann.
l’Université de Grenoble, 13, i–v (1901).
[17] J. T. Bottomley, “Scientific Worthies. XX.—James Prescott Joule.” Nature, 26, 617–620
511
B IBLIOGRAPHY 512
(1882).
[18] Pascale Boulanger, Marc le Maire, Mélanie Bonhivers, Sandrine Dubois, Michel Desmadril,
and Lucienne Letellier, “Purification and Structural and Functional Characterization of
FhuA, a Transporter of the Escherichia coli Outer Membrane.” Biochemistry, 14,
14216–14224 (1996).
[19] Vincenzo Brandani, Angelo Chianese, and Maurizio Rossi, “Ternary Liquid-Liquid
Equilibrium Data for the Water-Ethanol-Benzene System.” J. Chem. Eng. Data, 30, 27–29
(1985).
[20] L. K. Brice, “Le Chatelier’s Principle: The Effect of Temperature on the Solubility of Solids
in Liquids.” J. Chem. Educ., 60, 387–389 (1983).
[21] P. W. Bridgman, The Thermodynamics of Electrical Phenomena in Metals and a Condensed
Collection of Thermodynamic Formulas. Dover, New York, 1961.
[22] P. W. Bridgman, “A Complete Collection of Thermodynamic Formulas.” Phys. Rev., 3,
273–281 (1914).
[23] P. W. Bridgman, The Nature of Thermodynamics. Harvard University Press, Cambridge,
Massachusetts, 1941.
[24] I. Brown, W. Fock, and F. Smith, “Thermodynamic Properties of Alcohol Solutions.” Aust. J.
Chem., 9, 364–372 (1956).
[25] Oliver L. I. Brown and George G. Manov, “The Heat Capacity of Carbon Disulfide from 15
to 300ı K. The Entropy and Heat of Fusion of Carbon Disulfide.” J. Am. Chem. Soc., 59,
500–502 (1937).
[26] H. L. Callendar, “On a Practical Thermometric Standard.” Phil. Mag., 48, 519–547 (1899).
[27] Sadi Carnot, “Reflections on the Motive Power of Fire, and on Machines Fitted to Develop
that Power.” In E. Mendoza, editor, Reflections on the Motive Power of Fire: Sadi Carnot
and other Papers on the Second Law of Thermodynamics by É. Clapeyron and R. Clausius,
pages 1–59. Dover, Mineola, New York, 1988. Translation by R. H. Thurston of Réflexions
sur la Puissance Motrice du Feu et sur les Machines Propres à Développer cette Puissance,
Bachelier, Paris, 1824.
[28] 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.
[29] John P. Chesick, “Maximum Work Revisited (letter).” J. Chem. Educ., 41, 674–675 (1964).
[30] É. Clapeyron, “Memoir on the Motive Power of Heat.” In E. Mendoza, editor, Reflections on
the Motive Power of Fire: Sadi Carnot and other Papers on the Second Law of
Thermodynamics by É. Clapeyron and R. Clausius, pages 71–105. Dover, Mineola, New
York, 1988. Translation by E. Mendoza of “Mémoire sur la Puissance Motrice de la
Chaleur,” J. de l’École Polytechnique, 14, 153–190 (1834).
[31] E. Colin W. Clarke and David N. Glew, “Evaluation of Thermodynamic Functions for
Aqueous Sodium Chloride from Equilibrium and Calorimetric Measurements below 154ı C.”
J. Phys. Chem. Ref. Data, 14, 489–610 (1985).
[32] R. Clausius, “Fourth Memoir. On a Modified 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
11–135. John Van Voorst, London, 1867. Translation of “Ueber eine Veränderte Form des
Zweiten Hauptsatzes der Mechanischen Wärmetheorie,” Ann. Phys. Chem. (Leipzig), 93,
481–506 (1854).
[33] 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
327–365. John Van Voorst, London, 1867. Translation of “Ueber Verschiedene für die
Anwendung Bequeme Formen der Hauptgleichungen der Mechanischen Wärmetheorie,”
Ann. Phys. Chem. (Leipzig), 125, 353–400 (1865).
B IBLIOGRAPHY 513
[34] 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 5000ı K from
Spectroscopic Data.” J. Am. Chem. Soc., 54, 2610–2626 (1932).
[35] Patrick Coffey, Cathedrals of Science: The Personalities and Rivalries That Made Modern
Chemistry. Oxford University Press, New York, 2008.
[36] E. Richard Cohen, Tomislav Cvitaš, Jeremy G. Frey, Bertil Holmström, Kozo Kuchitsu,
Roberto Marquardt, Ian Mills, Franco Pavese, Martin Quack, Jürgen Stohner, Herbert L.
Strauss, Michio Takami, and Anders J Thor, Quantities, Units and Symbols in Physical
Chemistry, 3rd edition. RSC Publishing, Cambridge, 2007.
[37] 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.
[38] J. D. Cox, D. D. Wagman, and V. A. Medvedev, CODATA Key Values for Thermodynamics.
Hemisphere Publishing Corp., New York, 1989.
[39] William H. Cropper, “Walther Nernst and the Last Law.” J. Chem. Educ., 64, 3–8 (1987).
[40] 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, 575–589 (1991).
[41] Thomas W. Davis, “A Common Misunderstanding of Hess’s Law.” J. Chem. Educ., 28,
584–585 (1951).
[42] J. de Swann Arons and G. A. M. Diepen, “Gas-Gas Equilibria.” J. Chem. Phys., 44,
2322–2330 (1966).
[43] 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.
[44] Howard DeVoe, “Particle Model for Work, Heat, and the Energy of a Thermodynamic
System.” J. Chem. Educ., 84, 504–512 (2007).
[45] Hilyard John Eglinton Dobson, “The Partial Pressures of Aqueous Ethyl Alcohol.” J. Chem.
Soc., 127, 2866–2873 (1925).
[46] Frederick G. Donnan, “Prof. J. H. van’t Hoff.” Nature, 86, 84–86 (1911).
[47] J. H. Dymond and E. B. Smith, The Virial Coefficients of Pure Gases and Mixtures, A
Critical Compilation. Clarendon Press, Oxford, 1980.
[48] A. E. Edwards and W. E. Roseveare, “The Second Virial Coefficients of Gaseous Mixtures.”
J. Am. Chem. Soc., 64, 2816–2819 (1942).
[49] D. Eisenberg and W. Kauzmann, The Structure and Properties of Water. Oxford University
Press, New York, 1969.
[50] Rodney P. Elliot, Constitution of Binary Alloys, First Supplement. McGraw-Hill, New York,
1965.
[51] C. W. F. Everitt, “Maxwell’s Scientific Papers.” Applied Optics, 6, 639–646 (1967).
[52] M. B. Ewing, T. H. Lilley, G. M. Olofsson, M. T. Rätzsch, and G. Somsen, “Standard
Quantities in Chemical Thermodynamics. Fugacities, Activities, and Equilibrium Constants
for Pure and Mixed Phases (IUPAC Recommendations 1994).” Pure Appl. Chem., 66,
533–552 (1994).
[53] 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, 340–343
(1967).
[54] P. Ferloni and G. Spinolo, Int. DATA Ser., Sel. Data Mixtures, Ser. A, 70 (1974).
[55] Alexander Findlay and A. N. Campbell, The Phase Rule and Its Applications, 8th edition.
Dover, New York, 1945.
[56] Kai Fischer and Jurgen Gmehling, “P-x and 1 Data for the Different Binary Butanol-Water
Systems at 50 ı C.” J. Chem. Eng. Data, 39, 309–315 (1994).
[57] George Fleck, “Peter Debye.” In Laylin K. James, editor, Nobel Laureates in Chemistry,
1901–1992, pages 228–235. Chemical Heritage Foundation, 1993.
B IBLIOGRAPHY 514
[58] W. F. Giauque and D. P. MacDougall, “Attainment of Temperatures Below 1ı Absolute by

Demagnetization of Gd2 .SO4 /3 8H2 O.” Phys. Rev., 43, 768 (1933).
[59] 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 273ı K.” J. Am. Chem. Soc., 58,
1144–1150 (1936).
[60] W. F. Giauque and R. Wiebe, “The Entropy of Hydrogen Chloride. Heat Capacity from
16ı K. to Boiling Point. Heat of Vaporization. Vapor Pressures of Solid and Liquid.” J. Am.
Chem. Soc., 50, 101–122 (1928).
[61] 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.
[62] J. Willard Gibbs, “Graphical Methods in the Thermodynamics of Fluids.” In Henry Andrews
Bumstead and Ralph Gibbs Van Name, editors, The Scientific Papers of J. Willard Gibbs,
Ph.D., LL.D., Vol. I, Thermodynamics, pages 1–32. Ox Bow Press, Woodbridge,
Connecticut, 1993. Reprint of book originally published by Longmans, Green in 1906.
[63] 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 Scientific Papers of J. Willard Gibbs, Ph.D., LL.D., Vol.
I, Thermodynamics, pages 33–54. Ox Bow Press, Woodbridge, Connecticut, 1993. Reprint
of book originally published by Longmans, Green in 1906.
[64] J. Willard Gibbs, “On the Equilibrium of Heterogeneous Substances.” In Henry Andrews
Bumstead and Ralph Gibbs Van Name, editors, The Scientific Papers of J. Willard Gibbs,
Ph.D., LL.D., Vol. I, Thermodynamics, pages 55–353. Ox Bow Press, Woodbridge,
Connecticut, 1993. Reprint of book originally published by Longmans, Green in 1906.
[65] R. E. Gibson and O. H. Loeffler, “Pressure-Volume-Temperature Relations in Solutions. II.
The Energy-Volume Coefficients of Aniline, Nitrobenzene, Bromobenzene and
Chlorobenzene.” J. Am. Chem. Soc., 61, 2515–2522 (1939).
[66] R. T. Glazebrook, “Joule, James Prescott.” In Sidney Lee, editor, Dictionary of National
Biography, Vol. 30, pages 208–214. Macmillan, New York, 1892.
[67] J. A. Gledhill and G. McP. Malan, “The Solubilities of Sparingly Soluble Salts in Water.”
Trans. Faraday Soc., 48, 258–262 (1952).
[68] Herbert Goldstein, Classical Mechanics. Addison-Wesley, Cambridge, Massachusetts, 1950.
[69] Sandra C. Greer, M. R. Moldover, and R. Hocken, “A Differential Transformer as a Position
Detector in a Magnetic Densimeter.” Rev. Sci. Instrum., 45, 1462–1463 (1974).
[70] Ulrich Grigull, Johannes Straub, and Peter Schiebener, Steam Tables in SI-Units.
Springer-Verlag, Berlin, 1990.
[71] E. A. Guggenheim, Thermodynamics: An Advanced Treatment for Chemists and Physicists,
8th edition. Elsevier Science, 1985.
[72] Stuart R. Gunn and Leroy G. Green, “Heat of Solution of Hydrogen Chloride.” J. Chem.
Eng. Data, 8, 180 (1963).
[73] Herbert S. Harned and Russell W. Ehlers, “The Dissociation Constant of Acetic Acid from 0
to 35ı Centigrade.” J. Am. Chem. Soc., 54, 1350–1357 (1932).
[74] Herbert S. Harned and Benton B. Owen, The Physical Chemistry of Electrolytic Solutions,
3rd edition. Reinhold, New York, 1958.
[75] William Henry, “Experiments on the Quantity of Gases Absorbed by Water, at Different
Temperatures, and under Different Pressures.” Philos. Trans. R. Soc. London, 93, 29–42
(1803).
[76] William Henry, “Appendix to Mr. William Henry’s Paper, on the Quantity of Gases
Absorbed by Water, at Different Temperatures, and under Different Pressures.” Philos.
Trans. R. Soc. London, 93, 274–276 (1803).
[77] H. Hess, “Thermochemische Untersuchungen.” Ann. Phys. Chem. (Leipzig), 50, 385–404
(1840).
B IBLIOGRAPHY 515
[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 École Polytechnique: Livre du Centenaire, 1794–1894,
Vol. I, pages 194–198. Gauthier-Villars et fils, 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, 132–137 (2006).
[82] Irmgard K. Howard, “S Is for Entropy. U is for Energy. What Was Clausius Thinking?” J.
Chem. Educ., 78, 505–508 (2001).
[83] H. L. Johnston and W. F. Giauque, “The Heat Capacity of Nitric Oxide from 14ı 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, 3194–3214 (1929).
[84] James Prescott Joule, “On the Mechanical Equivalent of Heat.” Philos. Trans. R. Soc.
London, 140, 61–82 (1850).
[85] Michael Kasha, “The Triplet State. An Example of G. N. Lewis’ Research Style.” J. Chem.
Educ., 61, 204–215 (1984).
[86] W. B. Kay, “Liquid-Vapor Phase Equilibria Relations in the Ethane–n-Heptane System.” Ind.
Eng. Chem., 30, 459–465 (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
0ı to 150ı C: Correlations and Tables for Atmospheric Pressure and Saturation Reviewed
and Expressed on 1968 Temperature Scale.” J. Chem. Eng. Data, 20, 97–105 (1975).
[89] Lord Kelvin, “On an Absolute Thermometric Scale Founded on Carnot’s Theory of the
Motive Power of Heat, and Calculated from Regnault’s Observations.” Philos. Mag., 33,
313–317 (1848). Online at http://zapatopi.net/kelvin/papers/.
[90] Milton Kerker, “Sadi Carnot and the Steam Engine Engineers.” Isis, 51, 257–270 (1960).
[91] Albert S. Keston, “A Silver–Silver Bromide Electrode Suitable for Measurements in Very
Dilute Solutions.” J. Am. Chem. Soc., 57, 1671–1673 (1935).
[92] Daniel Kivelson and Irwin Oppenheim, “Work in Irreversible Expansions.” J. Chem. Educ.,
43, 233–235 (1966).
[93] Jan T. Knuiman, Peter A. Barneveld, and Nicolaas A. M. Besseling, “On the Relation
between the Fundamental Equation of Thermodynamics and the Energy Balance Equation in
the Context of Closed and Open Systems.” J. Chem. Educ., 89, 968–972 (2012).
[94] Angela Köhler-Krützfeld and Klaus Christmann. “Hermann Heinrich Hess,” online at
http://www.chemiedidaktik.uni-oldenburg.de/download/Hermann_Heinrich_
Hess.pdf (accessed July 2011).
[95] Richard J. Kokes, “Maximum Work Revisited (letter).” J. Chem. Educ., 41, 675–676 (1964).
[96] Frank L. Lambert, “Entropy Is Simple, Qualitatively.” J. Chem. Educ., 79, 1241–1246
(2002).
[97] 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, 94–98 (2009).
[98] Peter T. Landsberg, Thermodynamics and Statistical Mechanics. Dover Publications, Inc.,
New York, 1990.
[99] Harvey S. Leff, “Multisystem Temperature Equilibration and the Second Law.” Am. J. Phys.,
45, 252–254 (1977).
[100] Harvey S. Leff, “Thermodynamic Entropy: The Spreading and Sharing of Energy.” Am. J.
Phys., 64, 1261–1271 (1996).
[101] Henry M. Leicester, “Germain Henri Hess and the Foundation of Thermochemistry.” J.
B IBLIOGRAPHY 516
Chem. Educ., 28, 581–583 (1951).

[102] Gilbert N. Lewis and Merle Randall, “The Activity Coefficient of Strong Electrolytes.” J.
Am. Chem. Soc., 1112–1154 (1921).
[103] Gilbert Newton Lewis, “A New Conception of Thermal Pressure and a Theory of Solutions.”
Proc. Am. Acad. Arts and Sciences, 36, 145–168 (1900).
[104] Gilbert Newton Lewis and Merle Randall, Thermodynamics and the Free Energy of
Chemical Substances. McGraw-Hill, New York, 1923.
[105] Gilbert Newton Lewis and Merle Randall, Thermodynamics, 2nd edition. McGraw-Hill,
New York, 1961. Revised by Kenneth S. Pitzer and Leo Brewer.
[106] David R. Lide (editor), CRC Handbook of Chemistry and Physics, 73rd edition. CRC Press,
Boca Raton, Florida, 1992.
[107] Thomas S. Logan, “Textbook Errors: Guest Column. XVI: The Vapor Pressure of Hydrated
Cupric Sulfate.” J. Chem. Educ., 35, 148–149 (1958).
[108] Jerry B. Marion and Stephen T. Thornton, Classical Dynamics of Particles and Systems, 4th
edition. Harcourt Brace, New York, 1995.
[109] K. N. Marsh, Int. DATA Ser., Sel. Data Mixtures, Ser. A, 2, 73 (1974).
[110] 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, 90–92 (1996).
[111] James Clerk Maxwell, Theory of Heat, 9th edition. Longmans, Green and Co., London,
1888.
[112] M. L. McGlashan, “Deviations from Raoult’s Law.” J. Chem. Educ., 40, 516–518 (1963).
[113] M. L. McGlashan, Chemical Thermodynamics. Academic Press, London, 1979.
[114] M. L. McGlashan, “The International Temperature Scale of 1990 (ITS-90).” J. Chem.
Thermodyn., 22, 653–663 (1990).
[115] E. Mendoza, Reflections on the Motive Power of Fire: Sadi Carnot and other Papers on the
Second Law of Thermodynamics. Dover, Mineola, New York, 1988.
[116] 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, 358–375
(1937).
[117] Robert A. Millikan, “Walther Nernst, a Great Physicist, Passes.” Scientific Monthly, 54,
84–86 (1942). Online at http://www.nernst.de/nernst_millikan1942.htm.
[118] I. M. Mills, “The Choice of Names and Symbols for Quantities in Chemistry.” J. Chem.
Educ., 66, 887–889 (1989).
[119] Ian Mills, Tomislav Cvitas̆, Klaus Homann, Nikola Kallay, and Kozo Kuchitsu, Quantities,
Units and Symbols in Physical Chemistry. Blackwell, Oxford, 1988.
[120] Ian Mills, Tomislav Cvitas̆, Klaus Homann, Nikola Kallay, and Kozo Kuchitsu, Quantities,
Units and Symbols in Physical Chemistry, 2nd edition. Blackwell, Oxford, 1993.
[121] Karol J. Mysels, “Maximum Work Revisited (letter).” J. Chem. Educ., 41, 677 (1964).
[122] R. Negishi, “The Solubility of Solid Benzene in Several Non-Polar Liquids.” Rev. Phys.
Chem. Japan, 15, 98–116 (1941).
[123] Nobelprize.org, Jacobus H. van’t Hoff - Biography. 1901. Online at http:
//nobelprize.org/nobel_prizes/chemistry/laureates/1901/hoff-bio.html.
[124] National Institute of Standards and Technology, NIST Chemistry WebBook. Online at
http://webbook.nist.gov/chemistry/.
[125] John W. Owens, Stanley P. Wasik, and Howard DeVoe, “Aqueous Solubilities and Enthalpies
of Solution of n-Alkylbenzenes.” J. Chem. Eng. Data, 31, 47–51 (1986).
[126] Vivian B. Parker, Thermal Properties of Aqueous Uni-Univalent Electrolytes. National
Standard Data Reference Series. National Bureau of Standards (U.S.), 1965.
[127] 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, 599–614 (1987).
[128] J. C. Philip, “Freezing-Point Curves for some Binary Mixtures of Organic Substances,
B IBLIOGRAPHY 517
Chiefly Phenols and Amines.” J. Chem. Soc., 83, 814–834 (1903).

[129] A. B. Pippard, Elements of Classical Thermodynamics for Advanced Students of Physics.
Cambridge University Press, Cambridge, 1966.
[130] Carl W. F. T. Pistorius, Martha C. Pistorius, J. P. Blakey, and L. J. Admiraal, “Melting Curve
of Ice VII to 200 kbar.” J. Chem. Phys., 38, 600–602 (1963).
[131] Kenneth S. Pitzer, “Gilbert N. Lewis and the Thermodynamics of Strong Electrolytes.” J.
Chem. Educ., 61, 104–107 (1984).
[132] Kenneth S. Pitzer and David A. Shirley, “William Francis Giauque.” In Biographical
Memoirs, Vol. 69. National Academies Press, Washington, 1996.
[133] Max Planck, Treatise on Thermodynamics, 3rd edition. Dover, Mineola, New York, 1990.
Translation by Alexander Ogg of Vorlesungen über Thermodynamik, de Gruyter, Berlin,
1922.
[134] Max Planck, “A Scientific Autobiography.” In Scientific Autobiography and Other Papers.
Williams and Norgate, London, 1950. Translation by Frank Gaynor of Wissenschaftliche
Selbstbiographie, Barth, Leipzig, 1948.
[135] R. F. Platford, “Oil-Water Partition Coefficients from Solvent Activities.” J. Solution Chem.,
5, 645–651 (1976).
[136] H. Preston-Thomas, “The International Temperature Scale of 1990 (ITS-90).” Metrologia,
27, 3–10 (1990).
[137] William Ramsey, “Prof. François Marie Raoult.” Nature, 64, 17–18 (1901).
[138] Merle Randall and Leona Esther Young, “The Calomel and Silver Chloride Electrodes in
Acid and Neutral Solutions. The Activity Coefficient of Aqueous Hydrochloric Acid and the
Single Potential of the Deci-Molal Calomel Electrode.” J. Am. Chem. Soc., 50, 989–1004
(1928).
[139] F. M. Raoult, “On the Vapor-Pressure of Ethereal Solutions.” In Harry C. Jones, editor, The
Modern Theory of Solution: Memoirs by Pfeffer, van’t Hoff, Arrhenius, and Raoult, pages
95–121. Harper and Brothers, New York, 1899. Translation by Harry C. Jones of “Sur les
Tensions de Vapeur des Dissolutions Faites dans l’Éther,” Ann. Chim. Phys., 6th Ser., 15,
375–407 (1888).
[140] F.-M. Raoult, “Sur les Tensions de Vapeur des Dissolutions.” Ann. Chim. Phys., 6th Ser., 20,
297–371 (1890).
[141] Joseph A. Rard and Donald G. Miller, “Isopiestic Determination of the Osmotic Coefficients
of Aqueous Na2 SO4 , MgSO4 , and Na2 SO4 –MgSO4 at 25ı C.” J. Chem. Eng. Data, 26,
33–38 (1981).
[142] Otto Redlich, “The So-Called Zeroth Law of Thermodynamics.” J. Chem. Educ., 47,
740–741 (1970).
[143] T. R. Rettich, Rubin Battino, and Emmerich Wilhelm, “Solubility of Gases in Liquids. 22.
High-Precision Determination of Henry’s Law Constants of Oxygen in Liquid Water from
T D 274 K to T D 328 K.” J. Chem. Thermodyn., 32, 1145–1156 (2000).
[144] Timothy R. Rettich, Y. Paul Handa, Rubin Battino, and Emmerich Wilhelm, “Solubility of
Gases in Liquids. 13. High-Precision Determination of Henry’s Constants for Methane and
Ethane in Liquid Water at 275 to 328 K.” J. Phys. Chem., 85, 3230–3237 (1981).
[145] R. A. Robinson and R. H. Stokes, Electrolyte Solutions, 2nd edition. Butterworths, London,
1959.
[146] Frederick D. Rossini, “Heat of Formation of Hydrogen Chloride and some Related
Thermodynamic Data.” J. Res. Natl. Bur. Stand. (U.S.), 9, 679–702 (1932).
[147] Benjamin Count of Rumford, “An Inquiry Concerning the Source of the Heat Which is
Excited by Friction.” Philos. Trans. R. Soc. London, 88, 80–102 (1798).
[148] James R. Salvador, Fu Guo, Tim Hogan, and Mercouri G. Kanatzidis, “Zero Thermal
Expansion in YbGaGe Due to an Electronic Valence Transition.” Nature, 425, 702–705
(2003).
[149] George Scatchard and C. L. Raymond, “Vapor–Liquid Equilibrium. II. Chloroform–Ethanol
B IBLIOGRAPHY 518
Mixtures at 35, 45 and 55ı .” J. Am. Chem. Soc., 60, 1278–1287 (1938).
[150] 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
Acid, Sucrose, Urea and Glycerol at 25ı .” J. Am. Chem. Soc., 60, 3061–3070 (1938).
[151] Francis Weston Sears and Mark W. Zemansky, University Physics, 4th edition.
Addison-Wesley, Reading, Massachusetts, 1970.
[152] Jan V. Sengers and Anneke Levelt Sengers, “The Critical Region.” Chem. Eng. News, June
10, 104–118 (1968).
[153] F. Hastings Smyth and Leason H. Adams, “The system, Calcium Oxide–Carbon Dioxide.” J.
Am. Chem. Soc., 45, 1167–1184 (1923).
[154] Anthony N. Stranges, “William Francis Giauque: An Adventure in Low-Temperature
Research.” J. Chem. Educ., 67, 187–193 (1990).
[155] Julian M. Sturtevant, “The Heat of Dilution of Aqueous Hydrochloric Acid at 25ı .” J. Am.
Chem. Soc., 62, 3265–3266 (1940).
[156] Craig Thornber, Thomas Henry, FRS and his son, William Henry, MD, FRS, GS. Online at
http://www.thornber.net/cheshire/ideasmen/henry.html.
[157] 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,
63–67 (1994).
[158] 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 25ı C.” J. Solution Chem., 31, 455–466
(2002).
[159] 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, 1–20 (1981).
[160] Louis A. Turner, “Zeroth Law of Thermodynamics.” Am. J. Phys., 29, 71–76 (1961).
[161] 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.
[162] J. H. van’t Hoff, “The Rôle of Osmotic Pressure in the Analogy between Solutions and
Gases.” In Harry C. Jones, editor, The Modern Theory of Solution: Memoirs by Pfeffer, van’t
Hoff, Arrhenius, and Raoult, pages 13–42. Harper and Brothers, New York, 1899.
Translation by Harry C. Jones of “Die Rolle des Osmotischen Druckes in der Analogie
zwischen Lösungen and Gasen,” Z. Phys. Chem. (Leipzig), 1, 481–508 (1887).
[163] N. B. Vargaftik, B. N. Volkov, and L. D. Voljak, “International Tables of the Surface Tension
of Water.” J. Phys. Chem. Ref. Data, 12, 817–820 (1983).
[164] 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, 569–579 (1973).
[165] 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).
[166] Jingtao Wang and Mikhail A. Anisimov, “Nature of Vapor-Liquid Asymmetry in Fluid
Criticality.” Phys. Rev. E, 75, 051107 (2007).
[167] 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.
[170] Edward W. Washburn, “Standard States for Bomb Calorimetry.” J. Res. Natl. Bur. Stand.
B IBLIOGRAPHY 519
(U.S.), 10, 525–558 (1933).

[171] Lynde Phelps Wheeler, Josiah Willard Gibbs: The History of a Great Mind, revised edition.
Yale University Press, New Haven, Connecticut, 1952.
[172] Emmerich Wilhelm, Rubin Battino, and Robert J. Wilcock, “Low-Pressure Solubility of
Gases in Liquid Water.” Chem. Rev., 77, 219–262 (1977).
[173] Jaime Wisniak, “Benoit Paul Emile Clapeyron: A Short Bibliographical Sketch.” Chem.
Educator, 5, 83–87 (2000).
I NDEX
A boldface page number refers to a definition. 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, 161 Activity quotient, 350

Acid dissociation constant, 409 Additivity rule, 229, 233, 237, 242, 289, 291,
Activity, 269, 408 293, 303, 307, 328, 346
of an electrolyte solute, 292 Adiabat, 77
of a gas, 185, 272 Adiabatic
of an ion, 288 boundary, 28
of a mixture constituent, 272 calorimeter, 168–170, 333
of a pure liquid or solid, 272 demagnetization, 158, 160
relative, 269 flame temperature, 341
of a solute, 272 process, 51, 57, 95, 128
of a solvent, 272 Advancement, 314
of a symmetrical electrolyte, 290 Affinity of reaction, 342n
Activity coefficient, 257, 258 Amount, 21, 37
approach to unity, 260 Amount of substance, 21, 37, 470
of a gas, 185 Ampere, 470
from the Gibbs–Duhem equation, Anisotropic phase, 30, 74
264–265 Antoine equation, 220p
of an ion, 287 Athermal process, 304
from the Debye–Hückel theory, 294 Avogadro constant, 21, 471
mean ionic Azeotrope, 405, 435
from the Debye–Hückel theory, 296 minimum-boiling, 435
of an electrolyte solute, 292 vapor-pressure curve, 436
from the Nernst equation, 463 Azeotropic behavior, 433
from the osmotic coefficient, 299 Azeotropy, 435
from solubility measurement, 391
of a symmetrical electrolyte, 289 Bar, 39
from the osmotic coefficient, 265–267 Barometric formula, 197, 276
of a solute Barotropic effect, 33
in dilute solution, 260 Base units, 19, 470
from gas fugacity, 261–264 Binary mixture, 223
of a solvent, 370n in equilibrium with a pure phase, 374
from gas fugacity, 261–264 Binary solution, 224
stoichiometric, 293 Bivariant system, 199, 419
Body, 28
520
I NDEX 521
Boiling point, 203 without liquid junction, 451

curve, 204, 433 without transference, 451
elevation in a solution, 375, 380 Cell potential, 89, 453
Bomb calorimeter, 320, 333, 335–340 equilibrium, 89
Bomb calorimetry, 361p Celsius scale, 41
Boundary, 27 Center of mass, 500
adiabatic, 28 Center-of-mass frame, 54, 58, 499–502
diathermal, 28 Centigrade scale, 41
Boyle temperature, 35 Centrifugal force, 503
Brewer, Leo, 271 Centrifuge, 274
Bridgman, Percy, 179, 495 cell, 276–279
Bubble-point curve, 433 Charge
electric, 452
Cailletet and Matthias, law of, 205 number, 294
Caloric theory, 62 Chemical amount, 21n
Calorie, 86n Chemical equation, 312
Calorimeter, 168 Chemical potential, 136, 142
adiabatic, 168–170, 333 of an electrolyte solute, 292
bomb, 320, 333, 335–340 of electrons, 456
Bunsen ice, 341 as a function of T and p, 214
combustion, 333 of a liquid or solid, 185
constant-pressure, 170, 213 of a pure substance, 181
constant-volume, 169 of a solvent
continuous-flow, 173 from the freezing point, 370–372
flame, 341 from the osmotic coefficient, 369
heat-flow, 341 from osmotic pressure, 372–374
isoperibol, 171, 333, 334, 341 of a species in a mixture, 235
isothermal-jacket, 171–173, 333, 334, standard, 257, 269
341 of a gas, 182
phase-change, 340 of a gas constituent, 240
reaction, 333–335 of an ion, 287
Calorimetry of a pure substance, 182
bomb, 335, 361p of a symmetrical electrolyte, 289
drop, 190p total, 196
to evaluate an equilibrium constant, 354 Chemical process, 302
to measure heat capacities, 168 subscript for, 477
to measure transition enthalpies, 213 Chemical work, 136
reaction, 322, 333–341, 410 Circuit
Candela, 470 electrical, 86–88
Carathéodory’s principle of adiabatic heater, 168, 169, 171, 172
inaccessibility, 118 ignition, 336, 338
Carnot Clapeyron equation, 216
cycle, 105, 105–108 Clapeyron, Émile, 217b
engine, 105, 105–108 Clausius
heat pump, 108 inequality, 118
Carnot, Sadi, 106b statement of the second law, 103
Cell Clausius, Rudolf, 102, 103, 109b, 127, 131
diagram, 450 Clausius–Clapeyron equation, 218, 369
electrochemical, 449 CODATA, 504
galvanic, 449 Coefficient of thermal expansion, 163
reaction, 451 Coexistence curve, 200, 213
with transference, 451 liquid–gas, 33
I NDEX 522
Colligative property, 375 equation

to estimate solute molar mass, 376 for a mean ionic activity coefficient,
Common ion effect, 391, 444 296, 297, 464
Component, 421 for a single-ion activity coefficient,
Components, number of, 47, 228n 294
Composition variable, 222, 225 limiting law, 296, 330, 332, 390
relations at infinite dilution, 225 theory, 294–299
Compressibility factor, 35 Deformation, 30
Compression, 51 elastic, 36
Compression factor, 35 plastic, 36
Concentration, 29, 223 work, 69–74
standard, 253 Degree of dissociation, 410
Condensation curve, 433 Degrees of freedom, 199, 419
Conditions of validity, 23n Deliquescence, 439
Congruent melting, 429 Density, 29
Conjugate measurement of, 38
pair, 137 Dependent variable, 46
phases Derivative, 479
in a binary system, 430 formulas, 479
in a ternary system, 443 Dew-point curve, 433
variables, 59, 137 Dialysis, equilibrium, 395
Consolute point, 430 Diathermal boundary, 28
Constants, physical, values of, 471 Dieterici equation, 26p
Contact Differential, 24, 481
force, 493 exact, 52, 481
potential, 455, 456 inexact, 52
Continuity of states, 33 total, 134, 481
Convergence temperature, 172 of the internal energy, 135–137
Conversion factor, 23 Dilution process, 324
Coriolis force, 276n, 503 Dimensional analysis, 24–25
Coulomb’s law, 489 Disorder, 129
Critical Dissipation of energy, 65, 66, 80, 83, 90–92,
curve, 439 94, 95, 113, 123, 129, 170
opalescence, 205 Dissipative work, 83, 89, 91, 94, 135
point Dissociation pressure of a hydrate, 437
of partially-miscible liquids, 430 Distribution coefficient, 394
of a pure substance, 33, 205 Donnan
pressure, 205 membrane equilibrium, 396, 396–399
temperature, 205 potential, 396
Cryogenics, 156–161 Duhem–Margules equation, 404
Cryoscopic constant, 378
Cubic expansion coefficient, 163, 210, 218 Ebullioscopic constant, 380
of an ideal gas, 188p Efficiency
negative values of, 163n of a Carnot engine, 111–113
Curie’s law of magnetization, 160 of a heat engine, 110
Current, electric, 86, 87, 168, 169, 449, 470 Efflorescence, 438
Cyclic process, 52 Einstein energy relation, 54, 182n
Elastic deformation, 36
Dalton’s law, 239 Electric
Debye crystal theory, 152 charge, 452
Debye, Peter, 158, 294, 295b, 297, 298 current, 86, 87, 168, 169, 449, 470
Debye–Hückel potential, 45, 286, 297, 298, 452
I NDEX 523
inner, 286, 452 of an ion, 322

potential difference, 87, 88, 396, 453 of a solute, 320
power, 173 of mixing to form an ideal mixture, 304
resistance, 44, 88, 169 molar reaction, 314
Electrical molar, effect of temperature on,
circuit, 86–88 323–324
conductor of a galvanic cell, 449 partial molar, 248
force, 489 in an ideal gas mixture, 241
heating, 88–89, 168, 169, 173 relative, of a solute, 329
neutrality, 235, 422 relative, of the solvent, 328
resistor, 60, 88, 90 of a solute in an ideal-dilute solution,
work, 61, 86–91, 95, 168–170, 173, 256
213, 338, 455 reaction
Electrochemical standard molar, 319, 320, 368, 369,
cell, 449 410
potential, 287n standard molar of a cell reaction, 461
Electrode, 450 relative apparent, of a solute, 330
hydrogen, 450 of solution
standard, 465 at infinite dilution, 325
potential, standard, 465 integral, 325
reaction, 451 molar differential, 324, 325, 385
Electrolyte molar integral, 325, 328, 357
solution, 285–300 of vaporization
symmetrical, 288–290 molar, 211
Electromotive force, 453n standard molar, 213
Electron Entropy, 102, 119, 129
chemical potential, 456 change
conductor of a galvanic cell, 449 at constant pressure, 176
number, 451 at constant volume, 175
Emf, 453n an extensive property, 122
Endothermic reaction, 318 as a measure of disorder, 129
Energy, 53, 488–495 of mixing
dissipation of, 65, 66, 80, 83, 90–92, 94, to form an ideal mixture, 304, 306
95, 113, 123, 129, 170 negative value, 304
Gibbs, 137 molar, 151–156
Helmholtz, 137 from calorimetry, 152
internal, 53, 53–54 of a nonequilibrium state, 122
kinetic, 488 partial molar, 248
potential, 490 of a solute in an ideal-dilute solution,
of the system, 52–54 256
thermal, 62, 495 reaction
Energy equivalent, 169, 170, 172, 173, 333, standard molar, 410, 412p
334, 337, 338 standard molar of a cell reaction, 462
Enthalpy, 137 residual, 155–156
change at constant pressure, 176 scale
of combustion, standard molar, 335, 340 conventional, 154
of dilution practical, 154
integral, 326 standard molar, 155, 410
molar differential, 326 of a gas, 186, 241
molar integral, 327 third-law, 151
of formation of a solute, molar, 327 zero of, 149–151
of formation, standard molar, 319 Equation
I NDEX 524
chemical, 312 function, 146

reaction, 312 quantity, 305
stoichiometric, 313 work, 91
Equation of state, 46, 47 Exergonic process, 302n
of a fluid, 33 Exothermic reaction, 318
of a gas at low pressure, 35, 244, 281p, Expansion, 51
362p free, 79
of an ideal gas, 33 reversible, of an ideal gas, 126
thermodynamic, 166 work, 71, 71–79, 95
virial, 34 reversible, 77
Equilibrium Expansivity coefficient, 163
dialysis, 395 Extensive property, 28
gas–gas, 440 Extent of reaction, 314
liquid–gas, 399–408 External field, 28, 49, 58, 195
liquid–liquid, 391–394
mechanical, 48 Faraday constant, 286, 452, 471
phase transition, 69, 344 Field
position, effect of T and p on, 356–358 external, 28, 49, 58, 195
reaction, 48, 343, 408–410 gravitational, 28, 36, 49, 54, 195, 274
solid–liquid, 383–391 magnetic, 160
thermal, 48 First law of thermodynamics, 56, 135
transfer, 48 Fluid, 31, 32–33
Equilibrium cell potential, 89, 453 supercritical, 32, 33, 205, 210, 211
Equilibrium conditions Flux density, magnetic, 158
for a gas mixture in a gravitational field, Force, 486–495
274–276 apparent, 276n, 496
in a gravitational field, 195 centrifugal, 503
in a multiphase multicomponent contact, 493
system, 235–237 Coriolis, 276n, 503
in a multiphase one-component system, effective, 496
193–194 electrical, 489
for reaction, 343–349 fictitious, 276n, 496, 498
for a solution in a centrifuge cell, frictional, 69
276–279 gravitational, 37, 197, 489, 503
Equilibrium constant Formation reaction, 319
mixed, 409 Frame
on a pressure basis, 352 center-of-mass, 54, 58, 499–502
thermodynamic, 351 lab, 53, 58, 59, 79, 274, 276, 487,
of a cell reaction, 461 490–493, 495, 496, 499, 500
temperature dependence, 368 local, 53, 57, 58, 69, 79, 276, 278, 495,
Equilibrium state, 48, 48–50 497, 498
Euler reciprocity relation, 482 nonrotating, 499–502
Eutectic rotating, 276, 496, 502–503
composition, 427 reference, 28, 53
halt, 428 earth-fixed, 503
point, 426, 429 inertial, 53, 486, 487, 490
temperature, 427, 428 Free expansion, 79, 128
Eutonic Freezing point, 203
composition, 444 curve, 383
point, 443 for a binary solid–liquid system, 426
Exact differential, 52, 481 of an ideal binary mixture, 384
Excess for solid compound formation, 388
I NDEX 525
depression in a solution, 375–378 Gibbs, Willard, 138b

to evaluate solvent chemical potential, Gibbs–Duhem equation, 230, 232, 233, 254,
370–372 264–265, 267, 300, 306, 388, 401
of an ideal binary mixture, 383–384 Gibbs–Helmholtz equation, 367
Friction Gravitational
internal, 91–94 field, 28, 36, 49, 54, 195, 274
sliding, 495 force, 37, 197, 489, 503
Frictional force, 69 work, 79–81, 95
Fugacity Gravitochemical potential, 196
effect of liquid composition on, Green Book, see IUPAC Green Book
400–404
effect of liquid pressure on, 399 Hückel, Erich, 294, 297, 298
of a gas, 183 Heat, 56, 89, 494
of a gas mixture constituent, 242 flow in an isolated system, 127–128
Fugacity coefficient reservoir, 50, 61, 103, 104, 107, 113,
of a gas, 184 115, 117, 118, 120, 121, 124, 125,
of a gas mixture constituent, 243, 245 128, 130
Fundamental equation, Gibbs, 235 technical meaning of, 61
Fundamental equations, Gibbs, 142, 286 transfer, 67–69
Heat capacity, 62, 142–143
Galvani potential, 286, 452 at constant pressure, 62, 143
Galvanic cell, 89, 449 molar, 143
in an equilibrium state, 49 partial molar, 248
Gas, 32 at constant volume, 62, 142
ideal, 74 of an ideal gas, 75
mixture, 239, 241 molar, 143
perfect, 74n at constant volume and constant
solubility, 405–407 pressure, relation between, 167
thermometry, 42–44 measurement of, by calorimetry, 168
Gas constant, 471 molar reaction, 322
Gas–gas immiscibility, 440 Heat engine, 103, 105
Giauque, William, 158, 159b Heater circuit, 168, 169, 171, 172
Gibbs Heating
equations, 139, 141 at constant volume or pressure, 174–176
fundamental equation, 235 curve, of a calorimeter, 169, 171, 173
fundamental equations, 142, 286 electrical, 88–89, 168, 169, 173
phase rule reversible, 126
for a multicomponent system, Helium, 202n
418–425 Helmholtz energy, 137
for a pure substance, 199 Henry’s law, 249
Gibbs energy, 137 not obeyed by electrolyte solute, 285
of formation, standard molar, 354 Henry’s law constant, 249
of an ion, 354 effect of pressure on, 407
of mixing, 303 effect of temperature on, 407
to form an ideal mixture, 304 evaluation of, 252
molar, 303 Henry’s law constants, relations between
molar, 141, 181 different, 252
molar reaction, 342 Henry, William, 250b
of a cell reaction, 457–461 Hess’s law, 319, 320, 335
reaction, standard molar, 350, 410 Hess, Germain, 321b
total differential of, for a mixture, 235 Homogeneous phase, 30
Gibbs, Josiah Willard, 142 Hydrogen electrode, 450
I NDEX 526
standard, 465 effect on reaction equilibrium, 409

Irreversible process, 66, 101, 123–125,
Ice point, 41 127–129
Ice, high pressure forms, 203 Isobaric process, 51
Ideal gas, 74 Isochoric process, 51
equation, 23, 33, 74 Isolated system, 28, 48
internal pressure, 166 spontaneous changes in, 127
mixture, 239, 241 Isoperibol calorimeter, 171, 333, 334, 341
in a gravitational field, 274–276 Isopiestic
and Raoult’s law, 247 process, 51
Ideal mixture, 248, 258 solution, 268
mixing process, 303 vapor pressure technique, 268
and Raoult’s law, 247 Isopleth, 427
Ideal solubility Isoteniscope, 203
of a gas, 406 Isotherm, 77, 209
of a solid, 386 Isothermal
Ideal-dilute solution, 252 bomb process, 335, 337–338, 340
partial molar quantities in, 255–256 compressibility, 163, 210, 218
solvent behavior in, 254–255, 401 of an ideal gas, 188p
Ideal-gas temperature, 40, 115 of a liquid or solid, 180
Ignition circuit, 336, 338 magnetization, 160
Impossible process, 66 pressure changes, 180–181
Independent variables, 46, 192 of a condensed phase, 180
of an equilibrium state, 119 of an ideal gas, 180
number of, 198, 420 process, 51
Indicator diagram, 77 Isotropic phase, 30
Inertial reference frame, 53, 487, 490 IUPAC, 19
Inexact differential, 52 IUPAC Green Book, 19, 137, 181, 211,
Inner electric potential, 452 259n, 324, 351, 453n, 470
Integral, 480
formulas, 480 Joule
line, 72, 480 coefficient, 188p
Integral enthalpy of dilution, 326 experiment, 188p
Integral enthalpy of solution, 325 paddle wheel, 84–86, 103
Integrand, 480 Joule, James Prescott, 58, 62, 84, 85b, 86,
Integrating factor, 122 99p, 188p
Intensive property, 29 Joule–Kelvin
Interface surface, 30 coefficient, 157
Internal experiment, 156
friction, 91–94 Joule–Thomson
pressure, 165, 165–166, 168 coefficient, 157, 179
of an ideal gas, 166 experiment, 156
resistance, 455
Internal energy, 53, 53–54 Kelvin (unit), 40, 470
of an ideal gas, 74 Kelvin, Baron of Largs, 102, 104, 113, 114b
of mixing to form an ideal mixture, 304 Kelvin–Planck statement of the second law,
partial molar, 248 104, 118
International System of Units, see SI Kilogram, 470
International temperature scale, 41 Kinetic energy, 488
Invariant system, 199, 419 Kirchhoff equation, 323, 338
Ionic conductor of a galvanic cell, 449 Konowaloff’s rule, 405
Ionic strength, 294, 296, 298
I NDEX 527
Lab frame, 53, 58, 59, 79, 274, 276, 487, Membrane, semipermeable, 49, 372
490–493, 495, 496, 499, 500 Metastable state, 50
Laplace equation, 198, 280p Meter, 470
Law Method of intercepts, 230, 231
of Cailletet and Matthias, 205 Metre, 470
of rectilinear diameters, 205 Microstate, 130
scientific, 56 Milliter, 37
Le Châtelier’s principle, 357, 358 Minimal work principle, 91
Legendre transform, 137, 141, 160, 484, Miscibility gap
484–485 in a binary system, 312, 430
Lever rule, 208, 426 in a ternary system, 443
for a binary phase diagram, 427 Mixing process, 302
general form, 208 Mixture
for one substance in two phases, 206 binary, 223
for partially-miscible liquids, 431 of fixed composition, 225
for a ternary system, 442 gas, in a gravitational field, 274–276
Lewis and Randall rule, 281p ideal, 258
Lewis, Gilbert Newton, 105, 149, 270b, 294 and chemical potential, 248
Line integral, 72, 480 and Raoult’s law, 247
Liquid, 32 simple, 308
Liquid junction, 449, 451, 457 Molal boiling-point elevation constant, 380
potential, 451, 457 Molal freezing-point depression constant,
Liquidus curve, 429 378
for a binary system, 426 Molality, 223
for a binary liquid–gas system, 432–435 standard, 253
for a binary solid–liquid system, 426 Molar
at high pressure, 439 differential reaction quantity, 316
Liter, 37 excess quantity, 305
Local frame, 53, 57, 58, 69, 79, 276, 278, integral reaction quantity, 316
495, 497, 498 mass, 37
nonrotating, 499–502 from a colligative property, 376
rotating, 496 from sedimentation equilibrium, 279
quantity, 29, 175
Magnetic reaction quantity, 316
enthalpy, 160 standard, 318
field, 160 Mole, 21, 37, 470
flux density, 158 Mole fraction, 222
Magnetization, isothermal, 160 standard, 254
Mass, 470 Mole ratio, 224
meaurement of, 36–37 Molecular mass, relative, 37
Mass fraction, 223 Molecular weight, 37
Maxwell relations, 140
Maxwell, James Clerk, 40 Natural variables, 137, 140, 141
McMillan–Mayer theory, 261n Nernst
Mean ionic activity coefficient, see Activity distribution law, 394
coefficient, mean ionic equation, 462
Mean molar volume, 230 heat theorem, 149
Melting point, 203 Nernst, Walther, 149n, 150b
normal, 42 Neutrality, electrical, 235, 422
Membrane equilibrium Newton’s law of cooling, 171
Donnan, 396, 396–399 Newton’s law of universal gravitation, 489
osmotic, 395 Newton’s second law of motion, 70, 197, 487
I NDEX 528
Newton’s third law of action and reaction, in an ideal gas mixture, 239
488, 489, 498 Partition coefficient, 394
Nonexpansion work, 135, 145 Pascal (unit), 39
Normal Path, 50
boiling point, 204 Path function, 52, 61
melting point, 42, 204 Peritectic point, 429
Perpetual motion of the second kind, 103n
Optical pyrometer, 45 Phase, 30
Osmosis, 373 anisotropic, 30, 74
Osmotic coefficient, 265, 381 boundary, 200
evaluation, 269 coexistence, 31
of a mean ionic activity coefficient, homogeneous, 30
299 isotropic, 30
Osmotic membrane equilibrium, 395 rule, see Gibbs phase rule
Osmotic pressure, 49, 373, 375, 381, 395 separation of a liquid mixture, 310–312,
to evaluate solvent chemical potential, 391, 430
372–374 transition, 31
van’t Hoff’s equation for, 381 equilibrium, 32, 69, 151, 344
Phase diagram
Paddle wheel, 60 for a binary liquid–gas system, 432
Joule, 84–86, 103 for a binary liquid–liquid system, 430
Partial for a binary solid–gas system, 436
specific quantity, 234 for a binary solid–liquid system, 426
specific volume, 234 for a binary system, 425
Partial derivative, 479 at high pressure, 439
expressions at constant T , p, and V , of a pure substance, 199
176–179 for a ternary system, 441
Partial molar Physical constants, values of, 471
enthalpy, 248 Physical quantities, symbols for, 472–475
in an ideal gas mixture, 241 Physical state, 30
relative, of a solute, 329 symbols for, 476
relative, of the solvent, 328 Pitzer, Kenneth, 271
of a solute in an ideal-dilute solution, Plait point, 443
256 Planck, Max, 102, 104, 116b
entropy, 248 Plasma, 32
in an ideal gas mixture, 241 Plastic deformation, 36
of a solute in an ideal-dilute solution, Plimsoll mark, 181
256 Potential
Gibbs energy, 142 chemical, 136
heat capacity at constant pressure, 248 electric, 452
internal energy, 248 energy, 490
quantity, 225 function, 489
of a gas mixture constituent, 244 standard cell, 461, 462
in general, 233 evaluation of, 464
general relations, 237–238 thermodynamic, 134
in an ideal mixture, 248 Potentiometer to measure equilibrium cell
in an ideal-dilute solution, 255–256 potential, 454
volume, 226–228, 248 Poynting factor, 399
in an ideal gas mixture, 241 Prefixes, 21
interpretation, 227 Pressure
negative value of, 227 changes, isothermal, 180–181
Partial pressure, 239 of a condensed phase, 180
I NDEX 529
of an ideal gas, 180 Quasicrystalline lattice model, 308

dissociation, of a hydrate, 437 Quasistatic process, 64
internal, 165, 165–166, 168
in a liquid droplet, 197 Randall, Merle, 105, 149, 294
measurement of, 38–40 Raoult’s law
negative, 166n deviations from, 402–404, 433
partial, 239 for fugacity, 246, 247
standard, 39, 181, 274, 359p, 467p in a binary liquid mixture, 402
sublimation, 203 in an ideal-dilute solution, 255
vapor, see Vapor pressure for partial pressure, 246, 246
Pressure factor, 271, 272–274 in a binary system, 432
of an electrolyte solute, 292 Raoult, François, 245
of an ion, 288 Raoult, François-Marie, 379b
of a symmetrical electrolyte, 290 Reactant, 313
Pressure–volume diagram, 77 Reaction
Process, 50 between pure phases, 344
adiabatic, 51, 57, 95, 128 cell, 451
chemical, 302 endothermic, 318
subscript for, 477 equation, 312
compression, 51 equilibrium, 408–410
cyclic, 52 exothermic, 318
dilution, 324 in a gas phase, 352–353
expansion, 51 in an ideal gas mixture, 346–349
impossible, 66, 101, 102–104 in a mixture, 344–346
irreversible, 66, 101, 123–125, 127–129 quotient, 350, 462
isenthalpic, 157 in solution, 353
isobaric, 51 Reaction quantity
isochoric, 51 molar, 316
isopiestic, 51 molar differential, 316
isothermal, 51 molar integral, 316
mechanical, 129 Reciprocity relation, 140, 147p, 482
mixing, 302 Rectilinear diameters, law of, 205
purely mechanical, 66 Redlich–Kister series, 310
quasistatic, 64 Redlich–Kwong equation, 26p, 34
reverse of a, 64 Reduction to standard states, 335
reversible, 62, 62–66, 94–95, 101, 102, Reference frame, 28, 53
129 earth-fixed, 503
solution, 324 inertial, 486
spontaneous, 62, 64–66, 101, 129, 342 Reference state, 257
throttling, 156 of an element, 319
Product, 313 of an ion, 287
Proper quotient, 350 of a mixture constituent, 259
Property of a solute, 252–253, 259
extensive, 28 of a solvent, 259
intensive, 29 Regular solution, 309
molar, 175 Relative activity, 269
Relative apparent molar enthalpy of a solute,
Quantity 330
molar, 29 Relative molecular mass, 37
specific, 29 Relative partial molar enthalpy
Quantity calculus, 22 of a solute, 329
Quartz crystal thermometer, 45 of the solvent, 328
I NDEX 530
Residual entropy, 155–156 Solidus curve for a binary system, 426

Resistance Solubility
electric, 44, 88, 169 curve, 383
internal, 455 for a binary solid–liquid system, 426
Resistor, electrical, 60, 88, 90 of a gas, 405–407
Retrograde ideal, 406
condensation, 440 of a liquid, relation to Henry’s law
vaporization, 440 constant, 393
Reverse of a process, 64 of a solid, 385
Reversible of a solid electrolyte, 389–391
adiabatic expansion of an ideal gas, 75 of a solid nonelectrolyte, 385–386
adiabatic surface, 119 Solubility product, 390, 461
expansion and compression, 71 temperature dependence, 391
expansion of an ideal gas, 126 Solute, 223
expansion work, 77 reference state, 252–253, 259
heating, 126 Solution, 223
isothermal expansion of an ideal gas, 75 binary, 224
phase transition, 69 in a centrifuge cell, 276–279
process, 62, 62–66, 94–95, 101, 102, ideal-dilute, 252
129 multisolute electrolyte, 292
work, 94 process, 324
Rotating local frame, 276, 502–503 regular, 309
Rubber, thermodynamics of, 148p saturated, 385
Rumford, Count, 62, 63b solid, 245, 428
Solvent, 223
Salt bridge, 457 activity coefficient of, 370n
Salting-out effect on gas solubility, 405, 414p behavior in an ideal-dilute solution,
Saturated solution, 385 254–255, 401
Saturation Species, 222, 419
temperature, 203 Specific
vapor pressure, 203 quantity, 29
Second, 470 volume, 29
Second law of thermodynamics Spontaneous process, 62, 64–66, 101, 129,
Clausius statement, 103 342
equivalence of Clausius and Standard
Kelvin–Planck statements, boiling point, 204
108–110 cell potential, 461, 462
Kelvin–Planck statement, 104, 118, 121 evaluation, 464
mathematical statement, 102, 125 chemical potential, see Chemical
derivation, 115–123 potential, standard
Sedimentation equilibrium, 279 composition, 253
Shaft work, 81, 81–86 concentration, 253
Shear stress, 30 electrode potential, 465
SI, 19 hydrogen electrode, 465
base units, 20, 470 melting point, 204
derived units, 20 molality, 253
prefixes, 21 mole fraction, 254
Simple mixture, 308 pressure, 39, 181, 274, 359p, 467p
Solid, 30, 36 Standard molar
viscoelastic, 31 properties, values of, 504–506
Solid compound, 386, 428, 436 quantity, 185
of mixture components, 386–389 evaluation of, 410–411
I NDEX 531
of a gas, 185–187 Surface tension, 144

reaction quantity, 318 Surface work, 143–144
Standard state Surroundings, 27
of a gas, 181 Symbols for physical quantities, 472–475
of a gas mixture constituent, 239 System, 27
of an ion, 287 closed, 28
of a mixture component, 269 isolated, 28, 48
of a pure liquid or solid, 181 open, 28, 227, 228, 235
of a pure substance, 181 state of, 45, 45–50
State System point, 200
of aggregation, 30, 476 Systéme International d’Unités, see SI
equilibrium, 48, 48–50
metastable, 50 Temperature
physical, 30, 476 Boyle, 35
standard, see Standard state convergence, 172
steady, 50 critical, 205
of a system, 45, 45–50 equilibrium systems for fixed values, 42
State function, 45, 45–47 ideal-gas, 40, 115, 166
change of, 51 international scale, 41
Statistical mechanics, 34, 130, 131 measurement of, 40–45
Boyle temperature, 35 scales, 40–41
Debye crystal theory, 152 thermodynamic, 40, 113, 113–115, 166,
Debye–Hückel theory, 297 470
ideal mixture, 261 upper consolute, 430
McMillan–Mayer theory, 261n upper critical solution, 430
mixture theory, 308 Thermal
molar entropy of a gas, 154 analysis, 428
molar heat capacity of a metal, 152 conductance, 171
second virial coefficient, 244 energy, 62, 495
virial equations, 34 reservoir, 50
Steady state, 50 Thermocouple, 45, 456n
Steam engine, 107 Thermodynamic
Steam point, 41, 205 equation of state, 166
Stirring work, 83, 83–84 equilibrium constant, 351
Stoichiometric potential, 134
activity coefficient, 293 temperature, 40, 113, 113–115, 166,
addition compound, see Solid 470
compound Thermometer
coefficient, 315 Beckmann, 44
equation, 313 constant-volume gas, 42
number, 315, 320, 350 liquid-in-glass, 40, 44
Sublimation optical pyrometer, 45
point, 203 quartz crystal, 45
pressure, 203 resistance, 44
temperature, 203 thermocouple, 45
Subscripts for chemical processes, 477 Thermometry, gas, 42–44
Substance, 222 Thermopile, 45, 341, 456n
Subsystem, 28 Third-law entropy, 151
Supercritical fluid, 32, 33, 205, 210, 211 Third law of thermodynamics, 120, 149
Superscripts, 478 Thompson, Benjamin, 62, 63b
Supersystem, 28, 115, 117, 120, 121, 124, Thomson, William, 104, 114b
125 Throttling process, 156
I NDEX 532
Tie line, 200 equation

on a binary phase diagram, 426 for a gas mixture, 243
on a ternary phase diagram, 442 for a pure gas, 34
Torque, 82 Virtual displacement, 194
Total differential, 134, 481 Viscoelastic solid, 31
of the Gibbs energy of a mixture, 235 Volume
of the Gibbs energy of a pure substance, mean molar, 230
181 meaurement of, 37
of the internal energy, 135–137 of mixing to form an ideal mixture, 304
of the volume, 228 molar, 30, 38
Triple partial molar, 248
line, 95n, 200 in an ideal gas mixture, 241
point, 200 specific, 29
cell, 42 total differential in an open system, 228
of H2 O, 40
Washburn corrections, 339, 340, 363p
Ultracentrifuge, 276–279 Work, 56, 57–59, 487–495
Units, 22 chemical, 136
Non-SI, 20 coefficient, 59
SI, 20 coordinate, 59, 119
SI derived, 20 deformation, 69–74
Univariant system, 199, 419 dissipative, 83, 89, 91, 94, 135
Upper consolute temperature, 430 of electric polarization, 95
Upper critical solution temperature, 430 electrical, 61, 86–91, 95, 168–170, 173,
213, 338, 455
van’t Hoff, Jacobus, 382b excess, 91
van’t Hoff equation, 369 expansion, 71, 71–79, 95
van’t Hoff’s equation for osmotic pressure, gravitational, 79–81, 95
381 of magnetization, 95, 158
Vapor, 33 mechanical, 95
Vapor pressure, 203 nonexpansion, 135, 145
curve, 204 reversible, 94
of a liquid droplet, 400 reversible expansion, 77
lowering in a solution, 375, 380–381 shaft, 81, 81–86, 95
saturation, 203 stirring, 83, 83–84
Vaporization, 369 stretching, 95
molar enthalpy of, 211 surface, 95, 143–144
molar heat of, 211 Working substance, 105, 113
Vaporus curve
for a binary system, 426 Zeotropic behavior, 436
for a binary liquid–gas system, 432, Zeroth law of thermodynamics, 40
434, 435
at high pressure, 439
Variables
conjugate, 137
dependent, 46
independent, 46
natural, 137, 140, 141
number of independent, 47
Variance, 199, 419
Virial
coefficient, 34

Physical Chemistry For Advanced Level

Uploaded by

Copyright:

Available Formats

Physical Chemistry For Advanced Level

Uploaded by

Document Information

Original Title

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

Physical Chemistry For Advanced Level

Uploaded by

Copyright:

Available Formats

THERMODYNAMICS

The second edition, version 6 is copyright ©2015 by Howard DeVoe.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs

I thank the Department of Chemistry and Biochemistry, University of Maryland, College

If you are a faculty member of a chemistry or related department of a college or uni-

Preface to the Second Edition 16

From the Preface to the First Edition 17

2 Systems and Their Properties 27

3 The First Law 56

4 The Second Law 101

5 Thermodynamic Potentials 134

6 The Third Law and Cryogenics 149

7 Pure Substances in Single Phases 163

8 Phase Transitions and Equilibria of Pure Substances 192

10 Electrolyte Solutions 285

11 Reactions and Other Chemical Processes 302

12 Equilibrium Conditions in Multicomponent Systems 366

13 The Phase Rule and Phase Diagrams 418

14 Galvanic Cells 449

Appendix A Definitions of the SI Base Units 470

Appendix B Physical Constants 471

Appendix C Symbols for Physical Quantities 472

Appendix D Miscellaneous Abbreviations and Symbols 476

Appendix E Calculus Review 479

Appendix F Mathematical Properties of State Functions 481

Appendix G Forces, Energy, and Work 486

Appendix H Standard Molar Thermodynamic Properties 504

Appendix I Answers to Selected Problems 507

Preface to the Second Edition 16

From the Preface to the First Edition 17

2 Systems and Their Properties 27

2.4.4 Equilibrium states . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3 The First Law 56

4 The Second Law 101

5 Thermodynamic Potentials 134

6 The Third Law and Cryogenics 149

7 Pure Substances in Single Phases 163

8 Phase Transitions and Equilibria of Pure Substances 192

9.8.1 Gas mixture in a gravitational field . . . . . . . . . . . . . . . . . . 274

10 Electrolyte Solutions 285

11 Reactions and Other Chemical Processes 302

11.7.3 General derivation . . . . . . . . . . . . . . . . . . . . . . . . . . 343

12 Equilibrium Conditions in Multicomponent Systems 366

12.9 Reaction Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408

13 The Phase Rule and Phase Diagrams 418

14 Galvanic Cells 449

Appendix A Definitions of the SI Base Units 470

Appendix B Physical Constants 471

Appendix C Symbols for Physical Quantities 472

Appendix D Miscellaneous Abbreviations and Symbols 476

Appendix E Calculus Review 479

Appendix F Mathematical Properties of State Functions 481

Appendix G Forces, Energy, and Work 486

Appendix H Standard Molar Thermodynamic Properties 504

Appendix I Answers to Selected Problems 507

Benjamin Thompson, Count of Rumford . . . . . . . . . . . . . . . . . . . . . . . . 63

Thermodynamics is a discipline that involves a formalization of a large

The first law of thermodynamics is nothing more than the principle of

The law that entropy always increases—the second law of

D 9:970 102 kg m 3 (1.2.2)

changes of X along a path is a definite integral equal to X :

The equation dU D ¶q C ¶w is the differential form of the first law, and U D q C w is

Esys D Ek C Ep C U (3.1.5)

w D wlab mgzloc (3.1.6)