Graduate Texts in Physics

Mário J. de Oliveira

Equilibrium
Thermodynamics
Second Edition
Graduate Texts in Physics

Series editors
Kurt H. Becker, Brooklyn, USA
Jean-Marc Di Meglio, Paris, France
Sadri Hassani, Normal, USA
Bill Munro, Kanagawa, Japan
Richard Needs, Cambridge, UK
William T. Rhodes, Boca Raton, USA
Susan Scott, Acton, Australia
H. Eugene Stanley, Boston, USA
Martin Stutzmann, Garching, Germany
Andreas Wipf, Jena, Germany
Graduate Texts in Physics
Graduate Texts in Physics publishes core learning/teaching material for graduate-
and advanced-level undergraduate courses on topics of current and emerging fields
within physics, both pure and applied. These textbooks serve students at the
MS- or PhD-level and their instructors as comprehensive sources of principles,
definitions, derivations, experiments and applications (as relevant) for their mastery
and teaching, respectively. International in scope and relevance, the textbooks
correspond to course syllabi sufficiently to serve as required reading. Their didactic
style, comprehensiveness and coverage of fundamental material also make them
suitable as introductions or references for scientists entering, or requiring timely
knowledge of, a research field.

More information about this series at http://www.springer.com/series/8431

Mário J. de Oliveira

Equilibrium
Thermodynamics

Second Edition

123
Mário J. de Oliveira
Institute of Physics
University of SQao Paulo
SQao Paulo, Brazil

ISSN 1868-4513 ISSN 1868-4521 (electronic)

Graduate Texts in Physics
ISBN 978-3-662-53205-8 ISBN 978-3-662-53207-2 (eBook)
DOI 10.1007/978-3-662-53207-2

Library of Congress Control Number: 2016950892

© Springer-Verlag Berlin Heidelberg 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, express or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer-Verlag GmbH Germany
The registered company address is: Heidelberger Platz 3, 14197 Berlin, Germany
To my beloved wife
Tânia
Preface

If a birth date could be assigned to the science of thermodynamics, it would

certainly be the day of publication of the Clausius article on the first and second
law of thermodynamics, which occurred in 1850. Independently, these two laws
were also laid by Kelvin. Both authors arrived at the two fundamental laws of
thermodynamics relying on the ideas of Carnot on the operation of heat engines and
on the work of Mayer and Joule on the mechanical equivalent of heat. Subsequently,
the thermodynamic theory received contributions from other authors including
Maxwell and Helmholtz. We point out the fundamental contribution of Gibbs whose
formulation of the thermodynamics on the basis of the convexity properties of the
thermodynamic potentials is that we follow in this book starting from Chap. 3.
By its origin, thermodynamics is closely related to the study of heat engines and
thermodynamic processes. However, the science of thermodynamics should also be
understood as the study of thermodynamic properties of substances. In this book,
we adopt the point of view according to which thermodynamics is concerned with
the study of macroscopic properties obtained from macroscopic laws. This does
not mean that the laws of thermodynamics cannot be obtained from microscopic
laws. We just understand that this task belongs to the domain of another discipline,
statistical mechanics.
In the first two chapters of this book, we use the thermodynamic processes
to introduce the first law of thermodynamics, which we call Joule principle,
and the first part of the second law of thermodynamics, which we call Carnot
principle, which allows us to define the absolute temperature and entropy. In
Chap. 3 we present the principle of maximum entropy as introduced by Gibbs,
which corresponds to the second part of the second law of thermodynamics. This
principle leads to the convexity of thermodynamic potentials, studied in Chap. 4
along with the Legendre transformations. Chapter 5 is reserved for the analysis
of the consistency of equations of state and Maxwell relations. The Nernst-Planck
principle, which is the third law of thermodynamics, is presented in Chap. 6.
This book treats not only the fundamental principles of thermodynamics but also
the thermodynamics of phase transitions and critical phenomena observed in various
types of systems. Initially, we study the transitions between ordinary phases, solid,

vii
viii Preface

liquid, and vapor, in pure substances and in binary mixtures. We then examine the
transitions of the order-disorder type in binary alloys and then pass to the study
of phase transitions in magnetic systems, which include the ferromagnetic and
antiferromagnetic materials, and in dielectrics.
The chapter on the Gibbs phase rule and the structure of the phase diagrams of
multicomponent systems is based on the work of Griffiths and Wheeler. The study
of criticality is done in accordance with the view that the systems have universal
behavior in the critical region and can be described by the scaling theory introduced
by Widom. In the study of phase transitions and criticality, we use equations of
state that are introduced ad hoc and whose most notorious example is the van
der Waals equation. Such equations, in spite of being different, describe the same
critical behavior if they have the same symmetries. This universal behavior can be
understood by the Landau theory of phase transitions, also described in this book.
Other topics studied in this book include the thermodynamics of solids, liquid
crystals, thermal radiation, and thermochemistry. We examine the structural phase
transitions and the phase transitions occurring between the mesophases of liquid
crystals. In the study of thermal radiation, we arrive at the Planck distribution by
a thermodynamic reasoning. The last chapter concerns the equilibrium thermody-
namics of chemical reactions.

São Paulo, Brazil Mário J. de Oliveira

Contents

1 Joule Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1
1.2 Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4
1.3 Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7
1.4 Conservation of Energy . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 9
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 11
2 Carnot Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 13
2.1 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 13
2.2 Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 16
2.3 Ideal Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 20
2.4 Cyclic Processes .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 24
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 27
3 Clausius-Gibbs Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
3.1 Thermodynamic Coefficients . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
3.2 Thermodynamic Stability .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 35
3.3 Second Law of Thermodynamics.. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 41
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 46
4 Thermodynamic Potentials .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 49
4.1 Fundamental Relation . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 49
4.2 Extensivity .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 51
4.3 Legendre Transformations . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
4.4 Convexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 62
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 69
5 Thermodynamic Identities. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 73
5.1 Consistency of the Equations of State . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 73
5.2 Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 75
5.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 79
5.4 Gas Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 82
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 87

ix
x Contents

6 Nernst-Planck Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 89
6.1 Nernst Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 89
6.2 Heat Capacity of Solids . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 92
6.3 Planck Postulate .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 100
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 108
7 Phase Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 111
7.1 Pure Substance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 111
7.2 Discontinuous Transition . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 114
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 126
8 Criticality .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 127
8.1 Critical Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 127
8.2 van der Waals Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 129
8.3 Critical Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 139
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 149
9 Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 151
9.1 Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 151
9.2 Fundamental Relation . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 152
9.3 Mixture of Ideal Gases . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 155
9.4 Dilute Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 158
9.5 General Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 163
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 168
10 Binary Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 171
10.1 Fundamental Relation . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 171
10.2 Gibbs Free Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 174
10.3 Phase Transition .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 178
10.4 Completely Miscible Substances . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 184
10.5 Partially Miscible Substances .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 192
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 200
11 Phase Diagrams .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 201
11.1 Gibbs Phase Rule .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 201
11.2 Structure of Phase Diagrams.. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 206
11.3 Griffiths-Landau Theory . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 213
12 Order-Disorder Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 223
12.1 Binary Alloys .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 223
12.2 Bragg-Williams Theory .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 227
12.3 Landau Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 234
13 Magnetic Systems .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 241
13.1 Magnetic Materials .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 241
13.2 Magnetic Thermodynamic Potentials . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 246
13.3 Ideal Paramagnetic Systems . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 251
13.4 Weiss Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 255
Contents xi

13.5 Criticality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 262

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 269
14 Magnetic Ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 273
14.1 Antiferromagnetism .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 273
14.2 Metamagnetic Systems. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 281
14.3 Anisotropy .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 286
14.4 Ferrimagnetism .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 292
15 Dielectrics.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 295
15.1 Dielectric Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 295
15.2 Dielectric Properties.. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 298
15.3 Devonshire Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 304
15.4 Order-Disorder Transition . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 309
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 314
16 Solids . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 315
16.1 Stress and Strain .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 315
16.2 Thermodynamic Coefficients . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 321
16.3 Structural Phase Transition . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 328
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 333
17 Liquid Crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 335
17.1 Mesophases .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 335
17.2 Nematics .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 336
17.3 Smectics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 344
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 347
18 Thermal Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 349
18.1 Electromagnetic Waves . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 349
18.2 Planck Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 354
19 Thermochemistry .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 359
19.1 Extent of Reaction and Affinity . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 359
19.2 Thermodynamic Coefficients . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 366
19.3 Thermochemical Equilibrium.. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 373
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 382

References .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 385

Index . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 395
Chapter 1
Joule Principle

1.1 Introduction

Thermal Phenomena

The thermal phenomena are events that occur around us every day and so are part of
our common experience. We feel the changes in temperature throughout the day
and perceive climate changes caused by variations in atmospheric pressure and
temperature. We notice that the fire is the most striking manifestation of the daily
thermal processes. We learn that it can be used to heat water and make it boil and
discover its enormous potentiality to generate motion and useful work.
In the absence of fire, the scalding water cools spontaneously to room temper-
ature. In this process of spontaneous cooling, we find that the energy dissipates
irreversibly in the form of heat transferred to the environment. However, it is
possible to avoid the dissipation if the heat generated by fire or by other means
is used to obtain useful work. In fact, this is achieved by means of a heat engine
which transforms part of the heat into work that can be used directly or converted
into another form of energy.
We also notice that a body does not cool spontaneously, unless it is placed in
contact with a colder body. On contact, the hotter body cools because it gives off
heat spontaneously to the colder body. However, it is possible to cool a body in the
absence of a colder one. In fact, this can be achieved by the refrigerator, which does
the opposite of the spontaneous process, that is, it removes heat from a colder body
by the consumption of work and gives heat to the warmer body.
The operation principles of heat engines and refrigerators are directly related to
the fundamental laws of thermodynamics, which are laws that govern how the heat
is transformed into work and vice versa. These fundamental laws also govern the
thermal behavior of bodies such as the thermal expansion, the decrease in volume
by compression, the increase in temperature by heat absorption, etc.

© Springer-Verlag Berlin Heidelberg 2017 1

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_1
2 1 Joule Principle

In general, the bodies expand when heated, as happens to a stem of iron or to

mercury inside a thermometer. Some materials, by contrast, contract when heated,
as happens to rubber. We mention also the peculiar behavior of water when heated,
which expands if it is above four degrees Celsius and contracts if it is below this
temperature. The bodies also contract by the increase of the pressure exerted on
them. Some bodies can be compressed more easily than others. In general, the gases
have large variations in volume when compressed, contrary to what happens with
the liquids and solids which have a small decrease in volume.
The amount of heat that the bodies can receive is a fundamental thermodynamic
property. Some bodies can receive a large amount of heat with a small increase
in temperature. The variation in temperature can also be zero, as in a system
comprising two or more thermodynamic phases in coexistence. A trivial example
is the system formed by water in coexistence with ice. If this system receives heat,
a certain amount of ice will melt becoming liquid, but the temperature remains
unchanged.

Viewpoint

We adopt the point of view according to which the science of thermodynamics

concerns the study of the macroscopic properties of the bodies from macroscopic
fundamental laws. This does not mean that the laws of thermodynamics can not
be obtained from the laws governing the microscopic motion of matter. We just
understand that this task is beyond the scope of thermodynamics and belongs to the
field of statistical mechanics.
According to our point of view, the thermodynamic laws and properties do
without reference to the microscopic constitution of matter. However, they acquire
greater understanding if we can resort on the atomic constitution of matter. After
all, the big breakthrough in the study of thermal phenomena, which gave rise to the
science of thermodynamics, was due to the understanding of heat as the agitation of
matter at the microscopic level.

Thermodynamic Equilibrium

The existence of thermodynamic equilibrium states is a fundamental postulate of

thermodynamics. The thermodynamic equilibrium comprises the thermal equilib-
rium and other forms of equilibrium, depending on the type of thermodynamic
system under study. A mechanical system in thermodynamic equilibrium is in
thermal equilibrium and in mechanical equilibrium.
When a hot body is placed in contact with a cold body, a flow of heat is
established from the first to the second up to the moment they reach a situation
in which the flow ceases. From that moment on the two bodies are in thermal
1.1 Introduction 3

equilibrium. The thermal equilibrium can be checked by measuring the temperature

of the bodies. If the two bodies have the same temperatures then they are in thermal
equilibrium. If several bodies are in thermal contact with each other, they will be in
thermal equilibrium if they have the same temperature.
We should also consider the internal thermal equilibrium of a thermodynamic
system. In this case it is useful to understand the system as consisting of several parts
that make up a set of multiple bodies in contact. The internal equilibrium means that
there is no heat flows between the various parts of the system so that they all must
have the same temperature. In summary, all parts of a system in thermal equilibrium
have the same temperature.
If the object of study is a mechanical system, the thermodynamic equilibrium
will also comprise mechanical equilibrium. For example, two gases in a closed
vessel and separated by a movable wall will be in mechanical equilibrium if the
wall is at rest. In this situation, the pressure is the same on both sides of the wall.
Generally speaking, the mechanical equilibrium of a fluid, in the absence of external
forces, requires no pressure gradient inside the fluid. Thus all parts of a system in
mechanical equilibrium have the same pressure, in the absence of external forces.

Thermodynamic Process

The thermodynamic state of a gas can be changed in several ways depending on the
type of wall which delimits the container which encloses the gas. If the walls are
rigid but allow heat exchange, diathermal walls, the state of the gas can be changed
simply by contact of the container, for example, with a warmer body. In this case, the
gas receives heat and the pressure increases while the volume remains unchanged.
If, on the other hand, the walls of the container do not allow the exchange of heat,
adiabatic walls, but are mobile, the state may be modified by movements of the
walls. A compression decreases the volume of the gas increasing its pressure. If the
walls are rigid and adiabatic, then the gas is completely isolated from the outside
and its equilibrium state remains unchanged.
If a system has its thermodynamic state modified, we say that it undergoes
a thermodynamic process. In the study of equilibrium thermodynamics we are
particularly interested in processes whose initial and final states are thermodynamic
equilibrium states. The intermediate states cannot be all equilibrium states because
it is necessary to remove the system from its initial equilibrium state for the process
to take place. This is what happens with the free expansion of a gas. In a container
with two compartments, a gas occupies one of the two compartments while the other
is empty. At a certain moment the wall separating the compartments is removed, the
gas expands freely and after a certain time the equilibrium is restored in a state in
which the gas fills all the container.
We can also imagine, on the other hand, that the intermediate states, although
out of equilibrium, are very close to equilibrium so that the thermodynamic process
can be considered as composed by a succession of equilibrium states, intercalated
4 1 Joule Principle

between the initial and final states. To this end, small perturbations are applied to
the system at regular intervals of time. The interval between two disturbances must
be large enough to enable the system to relax to a new equilibrium state before
it will be disturbed again. If the disturbance is too small, then the equilibrium
state occurring after each disturbance will be very close to the disturbed state so
that the time interval between two disturbance may be reduced. In the limit where
the disturbances are arbitrarily small and infinitely large in number, the successive
equilibrium states form a continuous sequence of equilibrium states that defines a
thermodynamic process which we call quasi-static. If the disturbance is made by the
contact with the external environment, then the system and the external environment
should be close to equilibrium. This implies, for instance, that the temperature of the
external environment and the system must be very close. Notice that this does not
happen with the free expansion of a gas, even if the gas flow from one compartment
to the other is arbitrarily small.
The increase of energy of a system has two contributions. One is the work done
on the system and the other is the heat received. This second contribution can also
be understood as work done on the system, but differs from the former because
it occurs at the microscopic level. However, this distinctiveness between heat and
work is not useful from the thermodynamic point of view. The thermodynamic
distinctiveness between them is made by postulating the existence of adiabatic walls
whose fundamental property is preventing the passage of heat. Later on, we will see
how the use of adiabatic walls will allow us to distinguish between heat and work.

1.2 Work

Representation of Quasi-Static Processes

Consider a quantity of gas confined in a container. Moving the container walls,

we can adjust the volume V of the gas at will. Introducing or removing heat, we
can vary the gas pressure p until it reaches the value we wish. Once the gas is in its
equilibrium state it is characterized by the volume V and the pressure p and therefore
may be represented by a point in pressure versus volume diagram, called Clapeyron
diagram.
Imagine that the gas undergoes a quasi-static process that takes it from an initial
equilibrium state A to a final equilibrium state B. These two states are represented
by two points in the Clapeyron diagram. The intermediate stages of the process are
also states of equilibrium, and therefore have a Clapeyron diagram representation.
Thus, the quasi-static process corresponds to a succession of points in the Clapeyron
diagram, forming a continuous path that connects the points A and B, as can be seen
in Fig. 1.1.
Notice that out of equilibrium processes, that is, such that the gas is not in
equilibrium, do not have representation in the Clapeyron diagram. Suppose, for
1.2 Work 5

Fig. 1.1 Representation of a

quasi-static process
connecting the states A and p
B, in the Clapeyron diagram.
A
The area under the curve
between the points A and B
equals the work done

VA VB V

example, that a gas undergoes a free expansion. Before the expansion, the gas is in
equilibrium and its state is represented by a point in the Clapeyron diagram. After
the expansion and as soon it enters again in equilibrium, the state of the gas will
be represented by another point which is distinct from the first. However, these two
points are not connected by a trajectory because the free expansion, a process that
is out of equilibrium, has no representation in the Clapeyron diagram, although the
initial and final states have it.

Isochoric and Isobaric Processes

The simplest quasi-static process that can be carried out is the one in which the
volume of the gas remains unchanged, called isochoric process. The corresponding
trajectory is an isochoric line, which in the Clapeyron diagram is simply a straight
line parallel to the p-axis. Introducing heat into the gas slowly while the walls of the
container are kept unchanged, we will be facing an isochoric process.
The quasi-static process in which the pressure remains unchanged is called
isobaric process. The corresponding trajectory is an isobaric line or isobar, which
in the Clapeyron diagram is merely a straight line parallel to the V-axis. Let us
imagine that the gas is confined in a cylindrical container having rigid base and side
walls and whose top can move freely. Over the top there is a block and the system
composed by the block, the container, and the gas are in mechanical equilibrium.
Suppose that from a certain instant the gas becomes subjected to a continuous flow
of heat, for example, through the base of the container. Suppose further that the heat
transfer is slow enough so that the system can be considered to be in mechanical
equilibrium at each instant. Due to the continuous flow of heat, the gas expands and
the block rises. Since the block has a fixed weight, the gas pressure remains constant
throughout the expansion and we are faced with an isobaric process.
6 1 Joule Principle

Mechanical Work

Consider the isobaric process just described. If we denote by A the area of the top,
then the force exerted by the gas on the block is F D pA. If the top rises from the
height `1 to the height `2 then the work W done by the gas in the isobaric process
will be W D F.`2  `1 / D pA.`2  `1 /, that is,

W D p.V2  V1 /; (1.1)

where V1 D A`1 is the initial volume and V2 D A`2 is the final volume. The work
done by the gas is represented in the Clapeyron diagram by the area under the line
segment that represents the isobaric process.
Let us imagine next an arbitrary quasi-static process connecting the states A and
B, like the one shown in Fig. 1.1. This process can be accomplished by conveniently
changing the mass of the block which rests on the top of the container as well as the
heat flow. The process is then approximated by a succession of isobaric subprocesses
such that the volume change in each one of them is V. Denoting by pi the pressure
corresponding to the starting point of each subprocess, then the work done is given
by the sum of the works corresponding to each isobaric subprocess,
X
WD pi V: (1.2)
i

In the limit where the number of subprocesses increases without limit, V ! 0 and
the work becomes equal to the integral
Z
WD pdV: (1.3)

The area under the path is equal to the work performed by the gas. When a gas
does work, increasing its volume, W is positive because the pressure of the gas is
positive. When a gas consumes work, reducing its volume, W is negative. Note that
in an isochoric process, the work done by the gas is zero.
As an example of the calculation of work, suppose that a gas undergoes a process
such that its pressure is inversely proportional to the volume, that is,

B
pD (1.4)
V
along the path, where B is constant. The work done by the gas when it expands from
a volume V1 to a volume V2 will be
Z Z
V2 V2
1 V2
WD pdV D B dV D B ln : (1.5)
V1 V1 V V1
1.3 Heat 7

As a second example, consider another process such that the gas pressure varies
according to the equation

A
pD ; (1.6)
V
where A is a constant and  is a numeric constant greater than one. If the gas expands
from an initial volume V1 to a final volume V2 , then the work done by the gas will be
Z V2 Z V2
A  C1  C1
WD pdV D A V  dV D fV  V1 g: (1.7)
V1 V1  C 1 2
 
Using the notation p1 D AV1 and p2 D AV2 , then

1
WD .p1 V1  p2 V2 /: (1.8)
 1

In the International System of Units, the volume is measured in m3 and the

pressure in Pascal (Pa), defined as 1 N/m2 . Another very common unit for pressure,
but not belonging to the International Unit System is the atmosphere (atm), defined
as 101,325 Pa. The unit of work in the International Unit System is the joule (J),
defined as 1 N m.

1.3 Heat

Measure of Heat

By putting a body in contact with a warmer body, there will be heat transfer from
the second to the first. However, in this procedure we will not know how much heat
is transferred from one body to the other. To transfer a determined amount of heat
to a system we proceed as in the following example.
A system comprising a fluid confined in a container with adiabatic walls has in its
interior a propeller shaft which is wounded by an inextensible wire. This wire passes
through a small pulley and is attached at its end to a block of mass m. Initially, the
block is at rest and propeller immobile. Then the block is released and falls slowly
with a constant speed v due to friction of the helix with the fluid. As the speed is
constant the power dissipated by the helix is equal to the work performed by the
force of gravity per unit time mgv, where g is the acceleration of gravity. In a time
interval t the fluid receives an amount of heat Q D mgvt. Such a procedure
in which a measurable amount of heat is introduced into a system we call Joule
procedure.
Another way of performing the Joule procedure is to put within the system a
resistor connected to a battery. If the circuit is closed during a time interval t, the
8 1 Joule Principle

heat received by the system equals Q D .E2 =R/t, where R is the resistance of the
resistor and E is the battery emf.
If a system is brought from a state A to a state B, through an arbitrary quasi-static
process in which it receives heat, the Joule procedure allows the direct determination
of the amount of heat received by the system. If on the other hand, the system
experiences a quasi-static process from a state C to a state D, where it yields heat,
the Joule procedure can not be applied directly because this procedure can only be
used to transfer heat to the system. We note, however, that in the inverse process
the system receives heat and this heat can be measured by the Joule procedure. The
amount of heat given off by the system in direct process CD is the same amount of
heat received in the reverse process DC.
Conventionally, the heat received by the system under consideration is positive
and the heat given off by the system is negative. Denoting by QAB the heat
exchanged in the process AB, then according to this convention QAB > 0. Similarly,
QCD < 0, QDC > 0 and also QCD D QDC .

Calorie

The heat has the same unit as work. In the International System of Units both
quantities are measured in joule (J). Another unit often used for heat, but not
belonging to the International System of Units, is the calorie (cal) defined as the
amount of heat required to raise the temperature of 1 g of water by 1 ı C. This
definition is directly related to the definition of specific heat. However, the specific
heat of water is not constant but varies with temperature, which makes the definition
of calorie inaccurate unless one declares what temperature is taken as reference.
Thus, many calories may be defined such as the thermochemical calorie, equivalent
to 4.184 J; the IT calorie, equivalent to 4.1868 J; the average calorie, equivalent to
4.1900 J; the 15 ı C calorie, equivalent to 4.1858 J, and the 20 ı C calorie, equivalent
to 4.1819 J.
The origin of the calorie as a unit of heat occurred before the development of
thermodynamics, when the heat measurements were not related to the measures
of work. The most common method to measure the heat developed in a process
is to compare it with that necessary to raise the temperature by one unit a certain
quantity of water. Another method for measuring heat is the one used by Lavoisier
and Laplace in which the heat developed by a system is determined by the amount
of ice that the heat can melt. It is worth mentioning that the first method is linked to
the specific heat of water and the second to the latent heat of melting ice.
The first values for calorie are found in the Joule experiments for determining the
mechanical equivalent of heat, that is, the work required to raise the temperature of a
unit a mass of water by one unit of temperature. Joule showed that the heat required
to raise by 1 ı F the temperature of one pound of water is equivalent to the work of
a body of 772 pounds falling from a height of one foot. In the International System
1.4 Conservation of Energy 9

of Units this is equal to 4.16 J. This value obtained by Joule for the mechanical
equivalent of heat, and therefore for the calorie, is very close to the current values of
the calorie. Before Joule the mechanical equivalent of heat was obtained by Mayer.
According to Mayer, the heating of a certain a quantity of water by 1 ı C corresponds
to the falling of the same amount of water from a height of 365 m. This result is
equivalent to 3.58 J.

Quasi-Static Adiabatic Process

Consider a gas confined to a cylindrical vessel with adiabatic walls, so that no heat
can be exchanged between the environment and the system. Then imagine that the
vessel top is raised slowly, leading to a gas expansion. This can be done by assuming
that the top initially contain a certain amount of sand. The sand is then slowly
removed, so that the process is quasi-static. The gas then undergoes a quasi-static
process where no heat is exchanged with the environment, that is, a quasi-static
adiabatic process. The corresponding trajectory is an adiabatic line.

1.4 Conservation of Energy

Internal Energy

To define the internal energy of a gas, we use as a reference one point in the
Clapeyron diagram, labeling it by O D .V0 ; p0 /, and associate arbitrarily to this
point an energy U0 . To determine the energy UA of an arbitrary point A D .VA ; pA /,
we choose a process consisting of an adiabatic line followed by an isochoric process,
as shown in Fig. 1.2. An adiabatic line connects the point O to an intermediate point
B D .VB ; pB / while an isochoric line connects the point B to the point final A. If
we denote by W0B the work done by the gas in the adiabatic process and by QBA the
heat received by the gas in the isochoric process, then the energy UA is defined by

UA D U0  W0B C QBA ; (1.9)

in which the work is determined by the formula (1.3) while the heat is measured
by the Joule procedure. It may happen, however, that QBA is negative, resulting in
the inability to use the Joule procedure. In this case, we use the inverse process and
determine UA by

U0 D UA  WB0 C QAB ; (1.10)

10 1 Joule Principle

Fig. 1.2 The point A is

connected to the reference C
point O in two ways. The first p
consists of an adiabatic line
(OB) followed by an
isochoric line (BA). The
second comprises an
isochoric line (OC) followed A
by an adiabatic line (CA) O

since QAB D QBA > 0. Since any two states can be connected by an adiabatic
line followed by an isochoric line, then the internal energy of gas is defined for any
state.

The Principle of Conservation of Energy

Suppose we choose another path to reach the point A from the point O. For example,
the path OCA, shown in Fig. 1.2, comprising an isochoric line followed by an
adiabatic line. In this case, the work and heat are different from the previous case.
However, the Joule principle, or principle of conservation of energy states that the
energy of an equilibrium state is independent of the process used to reach it. Thus
the energy of the point A is the same regardless of the path used to reach the point
A from the reference point O. We may write then, for any path,

UA D U0 C Q  W; (1.11)

where Q is the heat exchanged and W is the work done by the system along the
chosen path.
Consider now any two points A and B. The energy difference U D UB  UA
between these two points will then be

U D Q  W; (1.12)

where Q is the heat exchanged and W is the work done by the system along any
path between any A and B. Remember that heat exchanged is positive when the
system receives heat and negative when the systems gives off heat; the work actually
performed by the system is positive while the work consumed is negative.
1.4 Conservation of Energy 11

First Law of Thermodynamics

The Joule principle, or principle of conservation of energy, is the first law of

thermodynamics. It was established by several scientists, but mostly by Mayer
and Joule, who assumed that the various forms of work could be converted into
one another and, moreover, that all of them could be dissipated as heat. Although
it seems quite obvious that the work can be transformed into heat, for example,
by friction, one can not conclude that the work dissipated always produce the
same amount of heat. It is necessary to experimentally verify such a law, which
is summarized in the determination of the mechanical equivalent of heat. This
determination was in fact performed by Mayer and exhaustively by Joule by
numerous experiments. By assuming that a certain amount of work always turns
into the same amount of heat, they were adopting the conservation of energy.
From the microscopic point of view, the principle of conservation of energy was
advanced by Helmholtz. Assuming the atomic constitution of matter, Helmholtz
extended the theorem of conservation of mechanical energy to the microscopic
motion of atoms. That is, he assumed that the sum of the kinetic and potential energy
of atoms is constant and is therefore the internal energy of a body. When any form
of work dissipates as heat, this means to say, from the microscopic point of view,
that atoms gain energy. Therefore, according to Helmholtz and other scientists that
have established the kinetic theory of matter, heat is associated to the microscopic
motion of atoms.

Problems

1.1 A gas undergoes a quasi-static process expanding from a state A characterized

by a volume V0 and a pressure p0 to a state B corresponding to a volume V1 . In this
5=3
expansion the pressure varies with volume according p D p0 V0 V 5=3 . Determine
the pressure p1 corresponding to state B. Calculate the work done by the gas when
it expands from state A to the state B. Assuming the expansion to be an adiabatic
line, what is the change in internal energy? Is the energy of the gas increased or
decreased?
1.2 A gas goes through the process described in the previous problem and then
undergoes an isochoric compression until a final state C whose pressure is p0 , the
same pressure of the state A. In this process the gas receives an amount of heat Q.
Suppose now that the gas undergoes an isobaric process from A to C. What is the
amount of heat received in this process?
1.3 For a given gas, the internal energy depends on volume and pressure according
U D .3=2/pV. Determine the work done by the gas when it is expanded from a state
A to a state B, both points A and B belong to the same adiabatic line. Determine the
heat received by the gas when it undergoes an isochoric process from state B to a
12 1 Joule Principle

state C such that C has the same energy as A. Suppose now that the gas undergoes
a process leading from A to C by a quasi-static process in which the energy is
constant. Determine the work and heat during this process. Given: VA , pA and VB ,
pB .
1.4 Suppose that the same gas of the previous problem undergoes a quasi-static
adiabatic expansion starting from a reference state .V0 ; p0 / of the Clapeyron
diagram. Along the adiabatic line the heat exchanged is zero, so that the work W
carried by the gas to a generic point .V; p/ is equal to the energy variation, that is,
Z V
3 3
 pdV D pV  p0 V0 :
V0 2 2

Use this equation to determine the equation of the adiabatic curve which passes
through the reference point. Hint: derive both members of this equation with respect
to V to find a differential equation for p.V/.
1.5 For a given gas the equation of the adiabatic line that passes through a reference
5=3
point .V0 ; p0 / is pV 5=3 D p0 V0 . Furthermore, the heat introduced in a quasi-static
process at constant volume, between the points .V; p1 / and .V; p/ is Qv D .3=2/.p
p1 /V. Determine the internal energy as a function of V and p.
Chapter 2
Carnot Principle

2.1 Temperature

Isothermal Process

When two bodies are placed in thermal contact, the hotter body gives off heat to
the colder body. As long as the temperatures are different, there will be a flow of
heat between them. After a while, the temperature becomes equal and the heat flow
ceases. From that moment the bodies are in thermal equilibrium.
Once a body is in thermal equilibrium its temperature does not change anymore,
unless it is disturbed. To verify possible changes in temperature, we used a tool
comprising a substance having a thermometric property, that is, a property which
varies with temperature. We call this instrument thermoscope and not thermometer,
considering that we did not introduce a temperature scale yet.
Consider a gas confined in a cylindrical container whose base allows the
exchange of heat and whose top can move freely. Over the top we place a certain
amount of sand. Starting from a certain pressure adjusted by the quantity of sand
on the top, we lead the gas through a process along which the temperature remains
constant by monitoring the temperature of the gas by means of the thermoscope.
We place the gas in the initial state and mark the value of the thermometric
property. Then we allow the gas to expand by introducing a small amount of heat.
If the temperature has changed, we vary the pressure by the removal of sand,
until the temperature returns to the same value. This operation is then repeated
many times. If the amount of heat introduced into each stage is small enough, the
procedure approaches a quasi-static process carried out at constant temperature, or
an isothermal process. The corresponding trajectory is an isotherm.
Another way to achieve an isothermal process is as follows. It is an experimental
fact that can be verified by the use of thermoscope that a pure substance, such as pure
water, has its temperature unchanged while it boils at constant pressure. If we place
the cylinder containing the gas in contact with boiling water, the gas temperature

© Springer-Verlag Berlin Heidelberg 2017 13

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_2
14 2 Carnot Principle

is remains constant. Removing sand from the top in small quantities, we will be
conducting an isothermal process.
The above procedures allow a system to go through an isothermal process.
However, we do not know the temperature of the system since we have not
defined a temperature scale. In the following we define a temperature scale, called
absolute temperature T, which allows us to determine the temperature of equilibrium
systems.

Carnot Cycle

Suppose that a system, comprising a fluid contained in vessel, runs a cycle consisting
of two isothermal processes and two adiabatic process. Starting from a state A,
the system undergoes an isothermal expansion to a state B, then an adiabatic
expansion to a state C, after an isothermal compression to a state D and finally
an adiabatic compression back to the initial state A. This cycle, shown in Fig. 2.1, is
referred to as the Carnot cycle. In the isothermal expansion, the system receives
an amount of heat Q1 from a heat reservoir at a temperature T1 and, in the
isothermal compression, gives off an amount of heat Q2 to a heat reservoir at
temperature T2 . The Carnot principle states that the ratio of the work and heat
received by a system which operates according to a Carnot cycle depends only on
the temperature of the reservoirs. Denoting by W the work done by the Carnot
cycle then W=Q1 depends only on T1 and T2 . This principle is universal and is
therefore independent of the substance that comprises the system that undergoes the
cycle.
By the principle of conservation of energy, the work done in a closed cycle equals
the heat received minus the heat given off, W D Q1  jQ2 j, so W=Q1 D 1 
jQ2 j=Q1 . We conclude from the Carnot principle that the ratio of heat given off jQ2 j

Fig. 2.1 Representation of

the Carnot Cycle in the
p A
Clapeyron diagram. AB
represents an isothermal
expansion, BC an adiabatic
expansion, CD an isothermal B
compression and DA an
adiabatic compression

D
C

V
2.1 Temperature 15

and the heat received Q1 depends only on the temperatures of the reservoirs, that
is

jQ2 j
D f .T1 ; T2 /: (2.1)
Q1

If now we use another system which runs a Carnot cycle whose isotherms
correspond to temperatures T2 and T3 then

jQ3 j
D f .T2 ; T3 /; (2.2)
Q02

where Q02 is the heat received and jQ3 j is the heat given off by the system running
this second cycle. We conveniently choose the second cycle so that the heat given off
by the first system is received by the second system so that Q02 D jQ2 j. Multiplying
the two equations (2.1) and (2.2) and taking into account that Q02 D jQ2 j, we get

jQ3 j
D f .T1 ; T2 /f .T2 ; T3 /: (2.3)
Q1

But Q1 and Q3 , appearing on the left side, can be considered respectively as

the heat received and heat given off by a Carnot cycle operating between the
temperatures T1 and T3 . Therefore,

jQ3 j
D f .T1 ; T3 /; (2.4)
Q1

so that

f .T1 ; T2 /f .T2 ; T3 / D f .T1 ; T3 /: (2.5)

A function that satisfies this equation is of the form

.T 0 /
f .T; T 0 / D ; (2.6)
.T/

from which we conclude that

jQ2 j .T2 /
D ; (2.7)
Q1 .T1 /

where .T/ is a function to be defined but which is independent of the substances

which is used to run the cycle.
16 2 Carnot Principle

Next we use the Carnot principle, in the form given by (2.7), to define the
absolute scale of temperature. Following the prescription of Kelvin, we define
absolute temperature T as the quantity such that .T/ / T or, equivalently,

jQ2 j T2
D : (2.8)
Q1 T1

Thus, to construct an absolute temperature scale, it suffices to assign a numerical

value to the temperature of a reference state. Any other temperature is determined
in accordance with (2.8). For the Kelvin scale (K) one uses the triple point of water
as a reference state whose temperature we assign the exact value

T0 D 273:16 K: (2.9)

It is worth mentioning that in the Kelvin scale the temperature of the melting of ice
at a pressure of 1 atm is Tsl D 273:15 K and that the temperature of boiling water at
the same pressure of 1 atm is Tlv D 373:15 K.
The Celsius scale (ı C) of temperature  is defined by

 D T  273:15 ıC; (2.10)

where T is given in kelvins. In the Celsius scale, the temperature of the triple point
of water is 0 D 0:01 ı C, whereas the temperature of the melting ice and boiling
water at the pressure of 1 atm are sl D 0 ı C and lv D 100 ı C, respectively.

2.2 Entropy

Definition of Entropy

The definition of temperature allowed us to determine the temperature of a gas for

any values of p and volume V. Thus, to each point of the Clapeyron diagram is
associated a temperature. The points having the same temperature form an isotherm.
The isotherms are a set of lines that never intersect, as can be seen in Fig. 2.2.
We can also plot on the same Clapeyron diagram, adiabatic lines corresponding
to the same gas. To set up an adiabatic line, we confine the gas in a cylindrical
container with adiabatic walls and slowly raise the top of the cylinder, while
measuring the volume and pressure. Repeating this procedure several times, we get
several adiabatic lines. The adiabatic lines form a set of lines which do not intersect,
as can be seen in Fig. 2.2.
Just as there is a thermodynamic quantity associated with the isotherms, which
is the temperature, already defined, we also associate to the adiabatic lines a
thermodynamic quantity. This quantity should be invariant along an adiabatic line,
2.2 Entropy 17

Fig. 2.2 Set of isotherms and

adiabatic lines in the
p
Clapeyron diagram. The lines T1
AB and CD are isotherms
corresponding to A
temperatures T1 and T2 ,
respectively. The lines AC T2 B
and BD are adiabatic lines
C D

but this is not sufficient to establish its value for each adiabatic line. To set this
magnitude, called entropy, we proceed as follows.
Consider any two adiabatic lines constructed in the manner shown above.
Consider a point A on the first adiabatic line and a point B on the second adiabatic
line such that A and B are in the same isotherm, corresponding to a temperature T1 ,
as shown in Fig. 2.2. If SA is the entropy of the first adiabatic line, then the entropy
SB of the second adiabatic line is defined by

Q1
SB D SA C ; (2.11)
T1

where Q1 is the amount of heat received by the system from A to B along the
isotherm T1 .
Suppose that another pair of points C and D are chosen on the two adiabatic lines,
C over the first adiabatic and D over the second, but belonging to another isotherm,
corresponding to a temperature T2 , as shown in Fig. 2.2. Using the same definition
we also have
Q2
SB D SA C ; (2.12)
T2

where Q2 is the amount of heat received by the system from C to D along the
isotherm T2 . At first sight, expressions (2.11) and (2.12) seem to be inconsistent.
But by the Carnot principle Q1 =T1 D Q2 =T2 so that (2.11) and (2.12) are in fact
consistent and the entropy of the second adiabatic line becomes independent of the
choice of the isotherms.
Thus, taking an adiabatic line as reference for entropy, all other adiabatic lines
will be associated with well defined values of entropy. To each point of the
Clapeyron diagram is associated a certain value of entropy. As the adiabatic lines
connect points with the same entropy value, they are also called isentropic lines. The
quasi-static adiabatic processes may equivalently be called isentropic processes.
18 2 Carnot Principle

Clausius Integral

Next we show that the entropy difference between any two points A and B in the
Clapeyron diagram is given by the integral Clausius
Z
dQ
SB  SA D ; (2.13)
T

which extends over any path connecting the points A and B. This integral should be
understood as follows. A path chosen from A to B is partitioned into a number
of segments. For each segment, we determine the ratio Qi =Ti between the heat
exchanged and the temperature at which the system is found along the segment.
However, as the segment is not necessarily an isotherm, the temperature may vary
along the segment. To avoid ambiguity we set Ti as the temperature of the system at
the beginning of the segment. Thus, we calculate the sum
X Qi
(2.14)
i
Ti

over all segments of the path. The integral in (2.13) should then be understood as
the limit of this sum when the number of segments grows without bound.
Let us demonstrate now that the sum in (2.14) is an approximation to the
difference in entropy SB  SA , while the number of sections is finite, but which
becomes exact in the limit as the number of segments becomes infinite. To this
end, we chose a smooth and monotonic decreasing path. In the extreme points
of the segments, we draw adiabatic curves. We focus next in a segment between
two consecutive adiabatic lines as the one shown in Fig. 2.3. The initial and final
point of the segment are I and F. The curves IG and HF are the two adiabatic
lines and GH is the isotherm built in such a way that the work WIGHF done by the
system along the path IGHF equals the work WIF along the path IF. Thus, the heat
QIGHF received by the system along the path IGHF equals the heat QIF received
along the segment IF. Indeed, from the energy difference between the states F
and I,

UF  UI D QIF  WIF D QIGHF  WIGHF ; (2.15)

calculated along the paths IF and IGHF, we conclude that QIGHF D QIF because
WIF D WIGHF . But QIGHF D QGH , since IG and HF are adiabatic lines, so that

QGH D QIF : (2.16)

2.2 Entropy 19

Fig. 2.3 The curve IF is a

segment of a path in the
SI
Clapeyron diagram. The p
curves IG and HF are the
adiabatic lines corresponding I SF
to the entropies SI and SF ,
respectively. The curve GH is
an isotherm built in such a
way that the works done by
G H
the system along the paths
IGHF and IF are equal
F

From the definition of entropy, we see that the difference SF  SI between the
points F and I is given by
QGH
SF  SI D ; (2.17)
TG
where TG is the temperature corresponding to the isotherm GH. Therefore,
QIF
SF  SI D : (2.18)
TG
Thus the entropy difference SB  SA will be
X X Qi
SB  SA D .SFi  SIi / D IF
; (2.19)
i i
TGi

where the summation is over all segments and the index i indicates that the quantities
refer to the i-th segment. By construction, the expression (2.19) is exact, even for
a finite number of segments. However, it cannot be identified with (2.14) because
TGi and Ti are not equal, although QiIF D Qi . However, in the limit in which the
number of segments increase indefinitely and the size of the segments decreases, TGi
approaches TIi D Ti , and the two summations become equal in this limit and they
are identified as the integral of expression (2.13).

Conservation of Energy in Differential Form

According to (2.18), the heat introduced along the segment IF is related to the
entropy difference between the extreme points by

QIF D TG .SF  SI /: (2.20)

20 2 Carnot Principle

The total heat Q introduced along the whole path is therefore given by
X X
QD QiIF D TGi .SFi  SIi /; (2.21)
i i

where the summation is over all segments of the path. In the limit where the number
of segments growths without bounds, the summation becomes the integral
Z
QD TdS; (2.22)

which extends along the path chosen between A and B. This equation says that the
heat exchanged along a process is identified with the area under the corresponding
trajectory in the T-S diagram.
Using (2.22) and (1.3), then the principle of conservation of energy can be written
in integral form
Z Z
U D TdS  pdV; (2.23)

where U is the internal energy difference between the points A and B, on the
diagram S-V, and the integrals extends over any trajectory that connects these two
points. Equivalently,

dU D TdS  pdV; (2.24)

which is the equation of energy conservation in differential form. From this

equation, we get
   
@U @U
TD and pD ; (2.25)
@S V @V S

where the derivatives should be carried out keeping constant the variable that
appears as an index.

2.3 Ideal Gas

Equation of State

All gases at sufficiently low densities behave similarly, irrespective of the type of
molecules that they are composed. This behavior is called the ideal behavior and the
gas is called an ideal gas. Therefore, an ideal gas must be understood as a shorthand
for a gas in the regime of low densities.
2.3 Ideal Gas 21

If the temperature is kept constant, the pressure of an ideal gas is inversely

proportional to the volume, that is,

A
pD ; (2.26)
V
which is the Boyle law. Experimentally, one finds that the constant A is proportional
to the temperature, which is the Gay-Lussac Law, so that we may write

BT
pD : (2.27)
V
The internal energy U of a gas which follows the law (2.27) is independent of the
volume, depending only on the temperature. This independence on the volume was
demonstrated experimentally by Joule by means of the free expansion of an ideal
gas. In many cases the energy is linear with temperature, so that

U D CT; (2.28)

where we arbitrarily set U D 0 for T D 0. If another behavior of the internal energy

of an ideal gas is not explicitly stated, this means that we will be assuming the
behavior given by (2.28).
To demonstrate that U depends only on the temperature, we use (2.24) in the
form
1 p
dS D dU C dV: (2.29)
T T
Since S is a state function, then dS is an exact differential and, therefore,
 
@ 1 @ p
D : (2.30)
@V T @U T

The right hand side of this equation vanishes because p=T D B=V depends on V but
not on U. Therefore, 1=T does not depend on V but only on U or, in other words, U
depends on T but not on V.
Next, we examine the consequences of assuming that U depends only on
temperature. This is to say that the left side of (2.30) vanishes. Therefore, p=T
depends only on V but not U. The equation of state (2.27) is therefore compatible
with this result although not the most general equation.
22 2 Carnot Principle

Avogadro Hypothesis

Consider two vessels with distinct volumes containing different amounts of the same
gas under the same temperature and pressure. Since the conditions are the same,
then the ratio between the numbers of molecules in the containers must be equal to
the ratio between the volumes. In other words, the number of molecules should be
proportional to the volume. Bearing in mind that, from (2.27), B is proportional to
the volume at T and p fixed, we then conclude that B is proportional to the number
of molecules.
Using the standard for the number of molecules called mole, we may write
B D NR where N is the number of moles and R is the constant of proportionality.
Therefore, for the gas we are considering, the following equation holds

NRT
pD : (2.31)
V
So far R is a constant which in principle should be different for each type of
ideal gas. However, the Avogadro hypothesis implies that the constant R must
be universal, that is, it must be the same for any gas at low densities. In fact,
the Avogadro hypotheses states that equal volumes of distinct ideal gases under
the same temperature and pressure contains the same number of molecules. This
hypothesis implies, for instance, that the ratio NT=pV should be the same for any
ideal gas. But according to (2.31), this ratio is a constant R, which therefore, must
be universal. It is called the universal gas constant and its value is

R D 8:314510 J=mol K: (2.32)

Under normal conditions of temperature and pressure, that is, under the pressure
of 1 atm (101,325 Pa) and temperature of 0 ı C, one mole of gas occupies a volume
equal to 22:41410 dm3 .
An argument similar to that used above leads us to conclude that the constant C
that appears in (2.28) is also proportional to the number of moles, so that

U D NcT: (2.33)

However, the constant c is not universal. For a monoatomic gas c D .3=2/R, for a
diatomic gas c D .5=2/R.
An alternative form of presenting the internal energy of ideal gases that
obey (2.33) is
c
UD pV; (2.34)
R
obtained form (2.31) and (2.33).
2.3 Ideal Gas 23

Adiabatic Curve

To determine the equation of an adiabatic curve, or isentropic line, in the Clapeyron

diagram, we first calculate the differential dU from (2.34),
c
dU D .pdV C Vdp/: (2.35)
R
Substituting (2.35) in (2.29) we obtain

cCR c
dS D pdV C Vdp; (2.36)
RT RT
or
c
dS D . pdV C Vdp/; (2.37)
RT
where the constant  is defined by

cCR
D : (2.38)
c
Along the isentropic line, S is constant so that

 pdV C Vdp D 0: (2.39)

Integrating this differential equation, we get

pV  D const:; (2.40)

that gives the dependence of p on V along the adiabatic line of an ideal gas. The
constant  varies from one gas to another. For a monatomic gas,  D 5=3 because
c D .3=2/R and for a diatomic gas,  D 7=5 because c D .5=2/R.

Entropy

To determine the entropy of an ideal gas corresponding to a point in the Clapeyron

diagram, we use the equation of state (2.31) to eliminate T in (2.37). Then we obtain
the following equation

 1
dS D Nc. dV C dp/: (2.41)
V p
24 2 Carnot Principle

Integrating this equation along a path formed by an isobaric line and by an isochoric
line, starting from a reference point O D .V0 ; p0 /, we find
Z Z
V
1 p
1
S  S0 D Nc dV C Nc dp; (2.42)
V0 V p0 p

or
V p
S D S0 C Nc ln C Nc ln ; (2.43)
V0 p0

where S0 is the entropy of the reference point. But V0 D Nv0 and S0 D Ns0 are
proportional to the number of moles N so that

V p
S D Ns0 C Nc ln C Nc ln ; (2.44)
Nv0 p0

equation that gives the entropy of an ideal gas as a function of volume, pressure and
number of moles.

2.4 Cyclic Processes

Heat Engines

Consider a system composed of a fluid confined in a cylindrical vessel and suppose

that initially it is a particular state. The fluid undergoes a process and returns to its
initial state. Suppose that the cyclic process is such that it performs some positive
work, W > 0. In the Clapeyron diagram, this work corresponds to the area of the
internal region of the cycle. As the fluid is back to the initial state, the change of
internal energy is zero so that the work equals the total heat Q. This total heat is
composed of the heat QR effectively received by gas and the heat QC given way by
the gas, that is, Q D QR C QC . But QR > 0 and QC < 0, so that Q D QR  jQC j.
Therefore,

W D QR  jQC j: (2.45)

Since W > 0, the heat received by the system is greater than the heat given away.
Therefore, in this case, we can say that part of the heat received by the system
becomes work and the other part is transferred to the outside. A heat engine works
in this way, converting heat into useful work.
Because not all the heat effectively received by the system becomes work, it is
important to know what fraction  of the heat received turns into work. This fraction
2.4 Cyclic Processes 25

is called the efficiency of the heat engine and is defined by

W
D ; (2.46)
QR
or

QR  jQC j jQC j
D D1 : (2.47)
QR QR

In general the efficiency is distinct for different cyclic processes and depends on the
substance which runs the cyclic process.
The simplest heat engine is that operating according to a Carnot cycle. As we
have seen, the Carnot cycle consists of four steps: (a) an isothermal expansion, at
a temperature T1 , in which the system receives a heat Q1 > 0; (b) an adiabatic
expansion; (c) an isothermal compression, at a temperature T2 , in which the system
yields a heat Q2 < 0; and (d) an adiabatic compression. The system returns to its
initial state, so that the work done W D Q1 C Q2 D Q1  jQ2 j. The temperature
T1 is greater than T2 . Thus, in this cycle, the system receives an amount of heat Q1
from a heat source at a higher temperature and gives off a smaller amount of heat to
a heat sink at a lower temperature.
According to the Carnot principle, the ratio jQ2 j=Q1 D T2 =T1 is independent of
the substance that undergoes the Carnot cycle. Therefore, the efficiency is

T2 T1  T2
D1 D ; (2.48)
T1 T1

which is independent of the substance and depends only on the temperatures of the
heat source and heat sink.

Otto Cycle

Let us examine a machine that operates with an ideal gas in a cycle called Otto
cycle. It consists of four processes: (a) an isochoric compression from A to B, (b)
an adiabatic expansion from B to C, (c) an isochoric decompression from C to D,
and (d) an adiabatic contraction from D to the starting point A. In the calculations
below we use the equation for the internal energy of an ideal gas in the form given
by (2.34).
In the first process the gas receives an amount of heat QAB . As the gas does not
perform work, then the heat is equal to the change of internal energy, that is,
c c
QAB D UB  UA D p B VB  p A VA : (2.49)
R R
26 2 Carnot Principle

In the second process the gas performs a work WBC . Because there is no heat
exchange, then the work is equal to the decrease in energy, that is,
c c
WBC D UB  UC D p B VB  p C VC : (2.50)
R R
In the third process there is not work done, so that the heat in this process is given
by
c c
QCD D UD  UC D p D VD  p C VC : (2.51)
R R
Finally, in the last process there is no heat exchanged so that the work is
c c
WDA D UD  UA D p D VD  p A VA : (2.52)
R R
The efficiency will be then

WBC C WDA p B VB  p C VC C p D VD  p A VA
D D ; (2.53)
QAB p B VB  p A VA
or
p C VC  p D VD
D1 : (2.54)
p B VB  p A VA

Recalling that VB D VA and VD D VC then

VC p C  p D
D1 : (2.55)
VA p B  p A
 
Now, the points A and D are on the same adiabatic line so that pA VA D pD VD .
 
Likewise, the points B and C are over the same adiabatic line so that pB VB D pC VC .
Recalling that VB D VA and VD D VC and combining these two relations we get
 
pD pC VA
D D : (2.56)
pA pB VC

Using these relations we obtain

  1
VC p C VA
D1 D1 : (2.57)
VA p B VC
2.4 Cyclic Processes 27

Refrigerators

If we run a heat engine in the reverse sense, then we are creating a refrigerator.
For a better understanding, let us examine a Carnot cycle operated in the opposite
direction. We start from a state corresponding to a temperature T1 . First we perform
an adiabatic expansion in which the system cools to a temperature T2 smaller than
T1 . Then we go through an isothermal expansion at the temperature T2 , the end of
which the system will have received a heat Q2 > 0. Then, the system is compressed
adiabatically to the temperature T1 . Finally, it is compressed isothermally at the
temperature T1 to the starting point. In this last stage it gives way a heat Q1 < 0.
The total work W is negative, which means that the system actually consumes work.
As the total energy change is zero, then W D Q1 C Q2 , which can be written as

jWj D jQ1 j  Q2 ; (2.58)

so that the heat given way is greater than the heat received. However, the important
point is that the system takes a certain amount of heat from a body at a lower
temperature and reject a greater amount of heat to a body at a higher temperature.
The purpose of the refrigerator is to remove as much heat as possible for a given
consumption of work. It is therefore convenient to define a performance coefficient
! of the refrigerator by

Q2 Q2
!D D ; (2.59)
jWj jQ1 j  Q2

where Q2 is the heat extract from the colder body and W is the work consumed.
Notice that Q2 can be greater than jWj. Taking into account that for the Carnot cycle
jQ1 j=Q2 D T1 =T2 then the performance coefficient for the cycle is

T2
!D : (2.60)
T1  T2

Paradoxically, a refrigerator may also function as a very efficient heater because

the Q1 rejected to the body at the higher temperature may be very large. It is at least
greater than the work consumed.

Problems

2.1 Determine the work done, the absorbed heat, the energy variation and the
change in entropy of an ideal gas undergoing the following processes. (a) Isothermal
expansion at a temperature T between two states of volumes V1 and V2 . (b) adiabatic
expansion from a state of volume V1 and pressure p1 to a state of volume V2 . (c)
28 2 Carnot Principle

Isobaric expansion at the pressure p between two states of volumes V1 and V2 . (d)
Isochoric compression at the volume V between two states of pressures p1 and p2 .
2.2 An ideal gas goes from an initial state A whose pressure is p0 and volume
is V0 to a final state B whose pressure is 2p0 and volume is 2V0 , through two
different processes comprising: (a) isothermal expansion to a point C followed by an
isochoric increase of pressure, (b) an isothermal compression to a point D followed
by an isobaric expansion. Represent the two processes in the Clapeyron diagram.
Find the volume and the pressure of the point C and the point D. For each process,
calculate for each section the work, the heat absorbed and the change in internal
energy.
2.3 An ideal gas undergoes a Carnot cycle composed by (a) an isothermal expan-
sion (AB) at a temperature T1 , (b) an adiabatic expansion (BC), (c) an isothermal
compression (CD) at a temperature T2 and (d) an adiabatic compression (DA).
Shown that the volumes and the pressures of the four states A, B, C and D, satisfies
the relations VC =VD D VB =VA and pC =pD D pB =pA . Determine the work done and
the heat exchanged in each section of the cycle. Show explicitly that the efficiency
is given by  D 1  T2 =T1 .
2.4 An ideal gas undergoes a cyclic process formed by an isobaric line (AB), an
isochoric line (BC) and an isotherm (CA). Make a sketch of the possible cycles in
the Clapeyron diagram. Determine the efficiency of a heat engine working according
to the cycle such that the temperature TB of state B is greater than the temperature
TA of state A, TB > TA .
2.5 Determine the efficiency of a heat engine that operates with an ideal gas in
accordance with the Brayton-Joule cycle consisting of two adiabatic processes and
two isobaric processes. Make a sketch of the cycle in the Clapeyron diagram.
2.6 Determine the efficiency of a heat engine that works with an ideal gas in accor-
dance with the Diesel cycle compose by (a) an isobaric expansion, (b) an adiabatic
expansion, (c) an isochoric decompression, and (d) an adiabatic compression. Make
a sketch of the cycle in the Clapeyron diagram.
2.7 An ideal gas undergoes the cyclic transformation ABCA composed by an
isochoric process (AB), an adiabatic process (BC) and an isobaric process (CA).
Represent the cycle in the Clapeyron diagram given that the temperature TB of B is
greater than the temperature TA of A, TB > TA . Show that the temperature TC of C
is related to TA and TB by
  1
TC D TB TA :

Determine the heat exchanged, the work done, the variation of the internal energy in
each of the processes as well as the efficiency of a heat engine operating according
to this cycle. Give the answers in terms of the temperatures of the three points.
2.4 Cyclic Processes 29

2.8 An ideal gas undergoes a free expansion from a volume V0 to a volume 2V0 .
Determine the variation of the entropy in this process.
2.9 Consider a straight line in the Clapeyron diagram passing through the point
A and having a slope equal to ˛ < 0. Suppose that an ideal gas undergoes
an expansion along this straight line starting from point A. Determine the heat Q
received by the gas as a function of the gas volume V. Plot Q versus V, indicating
the interval in which the heat increases with volume and the interval in which it
decreases with volume. Do the same with entropy.
2.10 In the Clapeyron diagram, consider the segment of a straight line connecting
two points A and B found in the same adiabatic curve. The volume of A is smaller
than the volume of B, that is, VA < VB . An ideal gas undergoes the cycle formed
by the segment and the adiabatic curve. Initially, the gas expands from A to B along
the straight line. Then, it s compressed adiabatically going back to the point A.
Determine the efficiency of a heat engine operating according to this cycle.
2.11 Show that the efficiency of a heat engine working between two temperatures
T1 and T2 in any cyclic quasi-static process is always smaller than the one operating
according to a Carnot cycle between the same temperatures. The temperatures T1
an T2 must be understood as the maximum and minimum temperatures attained
by the substances undergoing the cycle. Hint: plot the cycle in a T-S diagram and
compare it with the Carnot cycle. Remember that in this diagram the heat exchanged
is identified as the area under the trajectory and that the Carnot cycle is a rectangle.
Chapter 3
Clausius-Gibbs Principle

3.1 Thermodynamic Coefficients

Introduction

A system in thermodynamic equilibrium must be stable. This means to say that small
perturbations do not remove the system from its equilibrium. In a mechanical system
in stable equilibrium described by a potential energy V .x/, the stability implies
that the coefficient d2 V =dx2 is non-negative. Similarly, a gas confined in a vessel
with adiabatic walls, which can be viewed as a mechanical system, the mechanical
stability implies that the coefficient @2 U=@V 2 is nonnegative, where U.V/ is the
internal energy and V the volume of the vessel. This quantity is directly related to
the adiabatic compressibility s , which will be defined later, by @2 U=@V 2 D 1=Vs .
Therefore, the mechanical stability of a gas means that the adiabatic compressibility
is nonnegative.
The overall stability of a thermodynamic systems is related not only to the
adiabatic compressibility, but also to other coefficients such as the isothermal
compressibility and the heat capacities at constant volume and at constant pressure,
as defined below. A system in thermodynamic equilibrium implies that these
coefficients are nonnegative. The positivity of the thermodynamic coefficients,
however, does not necessarily imply global stability of thermodynamic systems.
It implies the local stability, that is, the stability with respect to small disturbances.
The global stability, which implies the local stability, of the thermodynamic systems
is guaranteed by the Clausius-Gibbs principle.

© Springer-Verlag Berlin Heidelberg 2017 31

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_3
32 3 Clausius-Gibbs Principle

Heat Capacity

The heat capacity of a body is defined as the ratio Q=T between the heat received
Q and the corresponding increase in temperature T. More precisely, as the limit of
this ratio when T ! 0. Because the amount of heat is small, then Q D TS, where
S is the increase in entropy. Therefore, the heat capacity is given by TdS=dT. We
however should discriminate the way in which the heat is received by the system. If
done at constant volume, we have the isochoric heat capacity Cv , given by
 
@S
Cv D T : (3.1)
@T V

If done at constant pressure, we have the isobaric heat capacity Cp , given by

 
@S
Cp D T : (3.2)
@T p

Figure 3.1 shows the heat capacity per unit mass, or specific heat, of water for
temperatures between 0 and 100 ı C, at the pressure of 1 atm.
In the first case, being the heat introduced at constant volume, it will be equal to
the variation of the internal energy so that we may write in an equivalent manner
 
@U
Cv D : (3.3)
@T V

4.25
cp (J/gK)

4.20
~

4.15
0 20 40 60 80 100
o
θ ( C)
Fig. 3.1 Isobaric specific heat cQp of water as a function of temperature at the pressure of 1 atm. At
15 ı C, cQp is 4.1858 J/gK. Source: AIP
3.1 Thermodynamic Coefficients 33

In the second case, the heat introduced is equal to the variation of energy added to
the work done by the system. Therefore, in this case
   
@U @V
Cp D Cp : (3.4)
@T p @T p

Using the equations of the ideal gas, we obtain

Cv D Nc; Cp D Nc C NR: (3.5)

Therefore, the constant  defined by (2.38), which appears in (2.40) of the adiabatic
line of an ideal gas, is equal to the ratio between the two heat capacities of the ideal
gas, that is,  D Cp =Cv .
Particularly useful formulas for obtaining entropy from the heat capacities are
as follows. Along an isochoric process, we integrate both members of (3.1), after
dividing by the temperature, to get
Z
Cv
SD dT; (3.6)
.V/ T

where the symbol (V) indicates that the integral must be performed along a path in
which the volume remains constant. Along an isobaric process, we integrate both
member of (3.2), after dividing by the temperature, to get
Z
Cp
SD dT; (3.7)
.p/ T

where the symbol (p) indicates that the integral must be done along a trajectory in
which the pressure is kept constant.

Coefficient of Thermal Expansion

The coefficient of thermal expansion ˛ is related to the ratio between the increase
in the volume and the increase in the temperature in a process at constant pressure.
More precisely, ˛ is defined by
 
1 @V
˛D : (3.8)
V @T p

For the case of an idea gas, it is given by

NR 1
˛D D : (3.9)
pV T
34 3 Clausius-Gibbs Principle

(a) (b)
1.00 1.000

0.99

ρ (g/cm )
ρ (g/cm )

3
3

0.98

0.97

0.96

0.95 0.999
0 20 40 60 80 100 0 5 10 15
o o
θ ( C) θ ( C)

Fig. 3.2 Density of water  as a function of temperature at the pressure of 1 atm. The maximum
value of density 0.999972 g/cm3 occurs when  D 4:0 ı C. Source: LB

The coefficient of thermal expansion can be positive, zero or negative. For

example, at normal pressure, the water has a positive value of ˛ for temperature
above 4 ı C and a negative value of ˛, below 4 ı C. Equivalently, we may say that
above 4 ı C the density of water decreases with temperature and below 4 ı C the
density increases with temperature. At 4 ı C the water has its maximum density as
seen in Fig. 3.2.

Compressibility

The compressibility is related to the ratio between the decrease in the volume and
the increase in the pressure. There are two types of compressibility. The isothermal
compressibility, denoted by T , and the adiabatic compressibility, denoted by s .
The first is defined by
 
1 @V
T D  ; (3.10)
V @p T

whereas the second is defined by

 
1 @V
s D  : (3.11)
V @p S

In the Clapeyron diagram, the isothermal compressibility is related to the slope of

the isotherm whereas the adiabatic compressibility is related to the slope of the
adiabatic curve.
3.2 Thermodynamic Stability 35

Sometimes, it is useful to employ the bulk modulus, defined as the reciprocal of

the compressibility. We have the bulk modulus at constant temperature BT D 1=T
and the adiabatic bulk modulus Bs D 1=s.
Using the equation of state of the ideal gas, V D NRT=p, we get

1
T D ; BT D p: (3.12)
p

Using the equation pV  D const., valid for the ideal gas along an adiabatic line, we
get

1
s D ; Bs D  p: (3.13)
p

3.2 Thermodynamic Stability

Principle of Maximum Entropy

We focus our attention on two distinct states of an isolated system composed of

several bodies, which can be interpreted as the various parts of the same system. In
the first state, the bodies are separated and each one is in thermodynamic equilibrium
at a given temperature and a given pressure. The separation between them is
imposed by the existence of one or more constraints of various types. A rigid wall,
for example, is a mechanical constraint. An adiabatic wall is a thermal constraint.
In this state, the system as a whole is in a state of constrained equilibrium. In the
second state, the constraints are absent, the bodies are in thermal and mechanical
contact with each other forming a system in an unconstrained equilibrium, all at the
same temperature and same pressure.
Next, we imagine a thermodynamic process that takes the system from a
state of constrained equilibrium to a state of unconstrained equilibrium, which is
accomplished simply by the withdrawal of the constraints, by placing the bodies in
thermal and mechanical contact. In this process, the energy and the volume remain
invariant. Under these conditions, the maximum entropy principle states that the
sum of the entropies of the body increases or remains constant. In other words, this
principle states that the entropy of the unconstrained equilibrium state is greater
than or equal to that of the constrained equilibrium, as long as the two states have
the same energy and the same volume.
Now we use the principle of maximum entropy to determine the conditions of
thermodynamic stability of a gas in contact with a thermal reservoir at temperature
T0 and a mechanical reservoir at a pressure p0 . The gas is confined in a vertical
cylindrical vessel having side walls that are rigid and adiabatic and whose base is in
contact with the thermal reservoir. This is an auxiliary system which exchanges
36 3 Clausius-Gibbs Principle

only heat without change in temperature. Because the temperature is constant,

the variation of its entropy SRT and the heat received QRT are related by
SRT D QRT=T0 .
The top of the cylinder is adiabatic and can move freely. Over the top there
is a block which together with the surrounding air, act as a mechanical reservoir,
maintaining a constant pressure on the gas, no matter what the height of the top.
The pressure p0 on the gas is equal to the ratio between the weight of the block
and the area of top plus the pressure of the atmospheric air. Because the pressure
remains constant, the work WRM done by the mechanical reservoir and the variation
VRM of its volume are related by WRM D p0 VRM .
Next, we consider the constrained and unconstrained equilibrium states. In the
first, the gas is in equilibrium but isolated from the two reservoirs. In the second, the
gas is in contact with the two reservoirs forming a composite system in equilibrium.
In the state of constrained equilibrium we denote by T1 and p1 the temperature and
the pressure of the gas. We suppose that in this state its volume is V1 , its energy
is U1 and its entropy is S1 . The variation of the total entropy Stotal , between the
unconstrained and the constrained equilibrium, is given by

QRT
Stotal D S0  S1 C SRT D S0  S1 C ; (3.14)
T0

since the mechanical reservoir does not exchange heat. In addition, the variation of
the total energy Utotal between the two states is

Utotal D U0  U1 C QRT  WRM D U0  U1 C QRT C p0 .V0  V1 /; (3.15)

since WRM D p0 VRM and VRM D .V1  V0 /.

If the composite system is brought from the constrained equilibrium to the
unconstrained equilibrium in such a way that the total energy is constant, then we
have Utotal D 0 and, according to the principle of maximum entropy, Stotal  0.
In other terms, for

U0  U1 C p0 .V0  V1 / C QRT D 0; (3.16)

we should have
QRT
S0  S1 C  0: (3.17)
T0

Replacing QRT , obtained from (3.16), into (3.17), we arrive at the principle o
maximum entropy in the form

1 p0
.S1  S0 /  .U1  U0 /  .V1  V0 /  0; (3.18)
T0 T0
3.2 Thermodynamic Stability 37

(a) (b)
S U
O
S0
O
U0

U0 U S0 S

Fig. 3.3 Illustration of the equivalence between the principle of maximum entropy and the
principle of minimum energy. (a) The curve S.U/ satisfies the inequality S.U/  S0 C .1=T0 /.U 
U0 /, whose right hand side describes the tangent straight line at O. (b) The curve U.S/ satisfies the
inequality U.S/  U0 C T0 .S  S0 /, whose right hand side describes the tangent straight line at O.
The functions S.U/ an U.S/ form a pair of inverse functions

which we call Clausius-Gibbs principle, as illustrated in Fig. 3.3a for the case
V1 D V0 .

Principle of Minimum Energy

Suppose again that the set of various bodies undergoes a thermodynamic process
that leads the set from the state of constrained equilibrium to the state of uncon-
strained equilibrium. However, the process that interests us here is the one in which
the total entropy is kept constant. For the occurrence of this situation, the total
energy can not remain invariant. In fact, according to the principle of minimum
energy, the total energy should decrease or remain the same. In other words, this
principle states that the energy of the state of unconstrained equilibrium is less than
or equal to that of the state of constrained equilibrium, as long as these states have
the same entropy.
We consider again, the gas confined in a cylinder vessel, initially separated from
the thermal and mechanical reservoirs. Then they are placed in contact and reach
equilibrium by a trajectory such that the total variation of entropy is zero. According
to the principle of minimum energy, the total energy should decrease or remain
constant. That is, for Stotal D 0, we should have Utotal  0. In other words, for

QRT
S0  S1 C D 0; (3.19)
T0

we should have

U0  U1 C p0 .V0  V1 / C QRT  0: (3.20)

38 3 Clausius-Gibbs Principle

Replacing QRT , obtained form (3.19), into (3.20), we arrive at the principle of
minimum energy in the form

.U1  U0 / C p0 .V1  V0 /  T0 .S1  S0 /  0; (3.21)

as illustrated in Fig. 3.3b for the case V1 D V0 .

Importantly, the inequalities (3.18) and (3.21) are equivalent, which means to
say that the principles of maximum entropy and minimum energy in the forms
given by these two inequalities are in fact two expressions of one and the same
principle, which we call the principle of Clausius-Gibbs. The equivalence of the
two inequalities for the case where the volume remains constant becomes evident
from Fig. 3.3.

Thermal Stability

We interpret the inequalities (3.18) and (3.21) as the general stability conditions of
a thermodynamic system. To study specific stabilities, we consider special cases of
these inequalities. We imagine that a gas is taken from a reference state with energy
U, entropy S and volume V, and led to another state with energy U 0 , entropy S0 and
volume V 0 , through a particular thermodynamic process. Defining U D U 0  U as
the change in energy, S D S0  S as the change in entropy and V D V 0  V as
the change in the volume, then the inequality (3.21) is written as

U C pV  TS  0; (3.22)

where T and p are the temperature and pressure of the reference state.
Stabilities of various kinds can be examined depending on the type of disturbance
made on the system. First, we analyze the thermal stability, which corresponds to
the case in which the gas exchanges heat without performing work. That is, the gas
suffers an isochoric disturbance. In this case, V D 0, so that the inequality (3.22)
becomes

U  TS  0: (3.23)

Up to terms of second order in S, the variation in energy, along an isochoric line
is calculated from
 
1 @T
U D TS C .S/2 (3.24)
2 @S V
3.2 Thermodynamic Stability 39

where we used the equality .@U=@S/V D T. Replacing this result into the
inequality (3.23) gives
 
@T
.S/2  0; (3.25)
@S V

from which we conclude that

 
@S
Cv D T  0: (3.26)
@T V

That is, the stability by isochoric perturbation implies that the isochoric heat
capacity cannot be negative.

Mechanical Stability

Now we consider the mechanical stability, which corresponds to an adiabatic

perturbation. In this case, S D 0, so that the inequality (3.22) is written as

U C pV  0: (3.27)

If we consider only small variation in the volume, then

 
1 @p
U D pV  .V/2 ; (3.28)
2 @V S

where we used the equality .@U=@V/S D p. Replacing this result into the
inequality (3.27) gives
 
@p
 .V/2  0; (3.29)
@V S

from which we get

 
1 @V
s D   0: (3.30)
V @p S

That is, the stability by adiabatic perturbation implies that the adiabatic compress-
ibility cannot be negative. In other words, the adiabatic expansion (increase of the
volume) of a gas implies the decrease in pressure.
40 3 Clausius-Gibbs Principle

Isothermal Stability

Let us assume now that the disturbance is such that the temperature of the gas
remains constant and equal to T, that is, an isothermal disturbance. In this case, the
variation in the internal energy of the gas is U D QT WT , where QT D TS is the
heat introduced isothermally and WT is the work done by the gas along an isotherm
at temperature T. Considering small changes in the volume, then the pressure varies
according to
 
@p
p0 D p C V: (3.31)
@V T

From this expression we get the isothermal work, given by

 
1 @p
WT D pV C .V/2 : (3.32)
2 @V T

Therefore, for an isothermal perturbation

 
1 @p
U D QT  WT D TS  pV  .V/2 ; (3.33)
2 @V T

which, replaced in (3.22), gives

 
@p
 .V/2  0; (3.34)
@V T

from which we conclude that

 
1 @V
T D   0: (3.35)
V @p T

That is, the stability by an isothermal perturbation implies that the isothermal
compressibility cannot be negative. In other words, the isothermal expansion of a
gas implies a decrease in pressure.

Isobaric Stability

In this case, the disturbance is such that the gas pressure remains the same, which
corresponds to an isobaric disturbance. The variation of internal energy of the gas
will then be U D Qp  Wp , where Wp D pV is the isobaric work done by
3.3 Second Law of Thermodynamics 41

the gas and Qp is the heat introduced at constant pressure. For small values of heat
introduced the temperature varies with entropy according to
 
@T
T0 D T C S; (3.36)
@S p

from which we obtain the result

 
1 @T
Qp D TS C .S/2 : (3.37)
2 @S p

Therefore, for an isobaric perturbation,

 
1 @T
U D Qp  Wp D TS C .S/2  pV; (3.38)
2 @S p

which, replaced in (3.22), gives

 
@T
.S/2  0; (3.39)
@S p

from which follows the result

 
@S
Cp D T  0: (3.40)
@T p

That is, the stability by an isobaric perturbation implies that the isobaric heat
capacity cannot be negative.

3.3 Second Law of Thermodynamics

Statements and Meanings

The second law of thermodynamics is the law that distinguishes thermodynamics

from the other areas of physics. At the same time, it is a fundamental law of nature
that has implications that go beyond the field of physics. Many are their statements
and their meanings. It was introduced independently by Clausius and Kelvin from
the ideas of Carnot on heat engines and cyclic processes. For this reason, the original
statements of this fundamental law refers to the operation of heat engines and cyclic
processes. The statement associated with the name of Kelvin says that no process
is possible whose sole result is the conversion of heat, extracted from a thermal
reservoir, into work. In other words, it is impossible to construct a heat engine that
42 3 Clausius-Gibbs Principle

converts all the heat received into work. The statement associated with the name
of Clausius states that no process is possible whose sole result is the transfer of
heat from a cooler to a hotter body. In other words, it is impossible to build a
refrigerator that operates without the consumption of work. The further development
of thermodynamic theory led to the Clausius definition of entropy as a state function
and the formulation of the second law as a principle of maximum entropy, which is
another form of representation of the second law.
The second law of thermodynamics must be understood as consisting of three
parts. The first one leads to the definition of absolute temperature and entropy.
Combined with the principle of conservation of energy, it allows us to set up
the thermodynamic space and introduce the fundamental relation of systems in
equilibrium. The second part consists of the form postulated by Gibbs for the
principle of maximum entropy. It leads us to the property of convexity of the entropy
and the conditions of stability of thermodynamic systems in equilibrium. These two
parts refer only to equilibrium states and are represented by the Carnot principle and
by the Clausius-Gibbs principle, respectively. These two principles altogether form
the second law of thermodynamics for equilibrium systems.
The third part of the second law of thermodynamics refers to the time evolution
of thermodynamic systems and has to do with the growth of entropy in spontaneous
and irreversible processes. It therefore corresponds to the dynamic aspect of the
principle of maximum entropy. We emphasize, however, that this aspect is not used
explicitly, since we treat only systems in thermodynamic equilibrium. In contrast,
the static aspect of the maximum entropy principle is used explicitly in the form
given by the Clausius-Gibbs principle. When the initial and final states of an
irreversible process are made up of equilibrium states a close relationship exists
between the two aspects, which we analyze below.

Spontaneous Processes

When two bodies at different temperatures are placed in thermal contact, a flow
of heat spontaneously establishes from the hotter to the colder body until the
temperatures become equal and the flow ceases. From that moment on, the bodies
remain in thermal equilibrium. Before being placed in contact, each body separately
is in thermodynamic equilibrium. The moment they are brought into contact, the two
bodies are in an state out of equilibrium that evolves into a state of final equilibrium.
It is important to note that we are considering bodies constituting an isolated system,
which does not exchange heat or work, that is, such that the total energy and volume
remain invariant.
We can imagine the more general case where several bodies, or various parts of
the same system, are separated by the existence of thermal constraints, arising from
adiabatic walls, or mechanical constraints, due to rigid walls. When the constraints
are removed while the system as a whole is isolated, they evolve spontaneously to
3.3 Second Law of Thermodynamics 43

the unrestricted equilibrium. A spontaneous process is therefore a process occurring

in an isolated system, whose initial state is a state of constrained equilibrium and
whose final state is a state of unconstrained equilibrium. The initial and final states
have the same energy and the same volume since the system is isolated.
The principle of maximum entropy leads to conclude that in a spontaneous
process the entropy increases. As an example, consider a body in thermal contact
with a reservoir at a temperature T0 . Suppose that the body finds itself initially at a
temperature T1 and the thermal contact takes place at constant volume. The variation
of the total energy of the reservoir and body is given by

Utotal D U0  U1 C QRT ; (3.41)

where U0 and U1 are the final and initial energies of the body and QRT is the heat
exchanged with the reservoir. The variation of the total entropy of the body and
reservoir is given by

QRT
Stotal D S0  S1 C ; (3.42)
T0

where S0 and S1 are the final and initial entropies of the body and QRT =T0 is the
variation of the entropy of the reservoir. Eliminating QRT from these two equations,
we reach the following relation

1 1
Stotal D S0  S1  .U0  U1 / C Utotal : (3.43)
T0 T0

Along an isochoric line

Z T0
U0  U1 D Cv dT (3.44)
T1

and
Z T0
Cv
S0  S1 D dT; (3.45)
T1 T

where Cv is the isochoric heat capacity of the body. Therefore,

Z T0
1 1 1
Stotal D .  /Cv dT C Utotal : (3.46)
T1 T T0 T0

Considering that the total energy of the system composed by the body and the
reservoir is constant, then Utotal D 0 and therefore the total variation of entropy is
44 3 Clausius-Gibbs Principle

given by
Z T0
1 1
Stotal D .  /Cv dT: (3.47)
T1 T T0

If T1 < T0 , then the expression in parentheses is positive, so that the integral is

positive because, since Cv  0. If T1 > T0 , then the expression in parentheses
is negative. But in this case, the lower limit of the integral is greater than the upper
limit, so that the integral is also positive. If T1 D T0 the integral vanishes. Therefore,
in all cases Stotal  0. Notice that in the second case, T1 > T0 , the energy and the
entropy of the body decrease.

Maximum Work

Suppose now that the body reaches the final state by a process in which the total
energy is not maintained constant. We imagine that the difference in total energy is
transformed into work Wtotal , performed against the outside. Thus, the variation of
the total energy can be nonzero and moreover Wtotal D Utotal . We may ask about
the maximum work that the composite system can perform when the body is taken
from the initial state to the final equilibrium state with the reservoir.
From (3.46), we obtain
Z T0
1 1
Wtotal D Utotal D T0 .  /Cv dT  T0 Stotal : (3.48)
T1 T T0

Taking into account that Stotal  0, then the maximum work is obtained when the
process is such that Stotal D 0. For this case the maximum work is
Z T0
1 1
Wtotal D T0 .  /Cv dT (3.49)
T1 T T0

and is clearly positive.

Microscopic Interpretation of Entropy

Entropy has a microscopic interpretation that is used as the foundation of statistical

mechanics. According to Boltzmann entropy is related to the number of microscopic
states available to the system. The higher the number W of accessible microscopic
states the higher the entropy S. However, entropy does not grow linearly with the
3.3 Second Law of Thermodynamics 45

number of accessible states but grows logarithmically, that is,

S D kB ln W; (3.50)

where kB is a universal proportionality constant called Boltzmann constant.

Let us examine what happens to entropy when an ideal gas expands freely.
Imagine a vessel consisting of two compartments separated by a rigid and adiabatic
wall. An ideal gas at a temperature T occupies one compartment whereas the other
is empty. In a given moment, the separation wall is removed and the gas expands
freely occupying the whole vessel and reaching equilibrium. Because the system is
isolated, there is no variation of energy, so that the final temperature will be the same
as the energy of an ideal gas depends only on the temperature. We suppose that the
volume V of the vessel is ` times greater than the volume V1 of the compartment
where the gas was, that is, V D `V1 . As the gas is ideal, the final pressure p will be
` times smaller than the initial pressure p1 , that is, p D p1 =`. Using (2.44) for the
entropy of an ideal gas, we obtain the following variation of entropy

S D NR ln `: (3.51)

After the expansion, each molecule of the gas may wander in a space ` times
greater which means that it has a number of possibilities, or microscopic states, `
times greater. The gas as a whole has, therefore, a number of microscopic states `n
times greater, where n is the number of gas molecules. The ratio W=W1 between
the final number W and the initial number W1 of microscopic states is, therefore,
W=W1 D `n . Using the Boltzmann formula (3.50), we get for the variation of
entropy the expression

W
S D kB ln W  kB ln W1 D kB ln ; (3.52)
W1
or

S D kB ln `n D nkB ln `; (3.53)

which is identical to the expression (3.51).

The number of molecules divided by the number of moles is equal to the
number of molecules in one mole NA , called Avogadro constant. Comparing
expressions (3.51) and (3.53), we see that the Boltzmann constant kB and the
universal gas constant R are related by

R
kB D : (3.54)
NA
46 3 Clausius-Gibbs Principle

The knowledge of the Avogadro constant allows the determination of the Boltzmann
constant. The value of the Avogadro constant is

NA D 6; 0221367  1023 mol1 (3.55)

and of the Boltzmann constant is

kB D 1:380658  1023 J=K: (3.56)

Problems

3.1 Determine Cv , Cp , T , S and ˛ of an ideal gas.

3.2 The coefficient of linear thermal expansion of a body is defined by ˇ D
.1=L/.@L=@T/p where L is the length of the body along a certain direction. Show
that for an isotropic body ˛ D 3ˇ. If the length of a metallic stem at temperature
T0 is L0 , show that for small values of T D T  T0 the length of the stem varies
according to L D L0 .1 C ˇT/.
3.3 Determine the variation of the entropy of an ideal gas along an isochoric
line and along an isobaric line from the isochoric and isobaric heat capacities,
respectively.
3.4 Find the variation of entropy along an isobaric line between the temperatures
T1 and T2 of a body whose heat capacity at constant pressure varies according to
Cp D A C BT.
3.5 Determine the variation of the entropy of a body when it is placed in thermal
contact with a heat reservoir at temperature T0 . Suppose that initially the body is at
temperature T1 and that the isochoric heat capacity is constant. Find the variation
of the total entropy Stotal and show explicitly that Stotal  0. Determine also the
maximum work that can be obtained when the body passes from the initial state to
the equilibrium state with the reservoir.
3.6 Two identical bodies are found at temperatures T1 and T2 . They are placed
in thermal contact and eventually reach the state of equilibrium with each other.
Determine the equilibrium temperature and the variation of the total entropy Stotal .
Show explicitly that Stotal  0. Find next the maximum work that one can extract
form these bodies and determine, in this case, the final equilibrium temperature.
Which temperature is the biggest? Suppose that the isochoric heat capacities of the
bodies are constant.
3.7 Solve the previous problem considering three identical bodies at temperatures
T1 , T2 and T3 .
3.3 Second Law of Thermodynamics 47

3.8 A body initially at a temperature T1 is placed successively in contact with

a series of n reservoirs numbered from 1 to n. The last reservoir, of number n,
has temperature T0 and the temperature difference between successive reservoirs
is T D .T1  T0 /=n. Determine the entropy variation of the body, of the reservoirs,
and the total entropy variation in the limit n ! 1. Suppose next that the body at
temperature T1 is placed in contact with the last reservoir. Determine in this case
the entropy variation of the body, of the reservoir and the total entropy variation.
Compare the results of both cases. Suppose that the isochoric heat capacity of the
body is constant.
3.9 Use the principle of minimum energy to show that

U11 .S/2 C 2U12 .S/.V/ C U22 .V/2  0;

where U11 , U12 and U22 are the second order derivatives of U with respect to S and
2
V. From this inequality, show that U11  0, U22  0 and U11 U22  U12  0.
3.10 Use the principle of maximum entropy to show that

S11 .U/2 C 2S12 .U/.V/ C S22 .V/2  0:

where S11 , S12 and S22 are the second order derivatives of S with respect to U and V.
2
From this inequality, show that S11  0, S22  0 and S11 S22  S12  0.
Chapter 4
Thermodynamic Potentials

4.1 Fundamental Relation

Energy Representation

The thermodynamic properties of an isotropic fluid, confined in a vessel, involve

several thermodynamic quantities such as pressure, temperature, volume, the inter-
nal energy and entropy. These quantities are not all independent. To determine how
many and which may be treated as independent we start by examining the equation

dU D TdS  pdV: (4.1)

We see that it is natural to imagine as U as dependent on S and V. Furthermore,

if we understand the internal energy U.S; V/ as a function of S and V, then this
equation allows us to obtain also the pressure and temperature as functions of S and
V through the relations

@U @U
TD ; pD : (4.2)
@S @V
The reasonings above were done by considering a given quantity of material
contained in the vessel. However, the amount of substance that composes the fluid
may also be changed. To completely specify the thermodynamic state of the fluid
we should therefore indicate the number of moles N of the substance in the vessel
in addition to S and V. The thermodynamic equilibrium state becomes defined by
the independent variables S, V and N. In an equivalent manner, we may say that
each state of thermodynamic equilibrium of an isotropic fluid composed by a single
pure substance corresponds to a point in the thermodynamic space .S; V; N/. Note
that we are assuming the fluid as composed by a single pure substance. For a fluid

© Springer-Verlag Berlin Heidelberg 2017 49

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_4
50 4 Thermodynamic Potentials

comprising several substances it would be necessary to indicate the number of moles

of each substance.
Since U is also a function of N, it is convenient to define a quantity that is
associated with the variation of energy when one makes an increase in the number of
moles of the pure substance that makes up the fluid. This quantity is called chemical
potential and is defined by

@U
D ; (4.3)
@N
so that we may write the conservation of energy in differential form as

dU D TdS  pdV C dN: (4.4)

The internal energy U.S; V; N/ as a function of the independent variables S,

V and N is called the fundamental relation in the energy representation because
from it we can obtain the thermodynamic properties of the system in consideration.
Deriving the energy with respect to S, V, and N, we obtain equations (4.2) and (4.3),
called equations of state in the energy representation. We assume as a postulate that
the internal energy U and its derivatives T, p and are continuous functions or,
equivalently, that the internal energy is a continuous and differentiable function of
entropy, volume, and number of moles.

Entropy Representation

Inverting the fundamental relation U.S; V; N/ with respect to S, we get the entropy
S.U; V; N/ as a function of the independent variables U, V, and N, which is the
fundamental relation in the entropy representation because from it we can obtain
the thermodynamic properties of the system in consideration. In this representation,
the thermodynamic state is defined by a point in the space .U; V; N/. From (4.4) we
may write

1 p
dS D dU C dV  dN; (4.5)
T T T
from which follow the three equations of state

1 @S p @S @S
D ; D ; D ; (4.6)
T @U T @V T @N
in the entropy representation.
Since S.U/ is the inverse of U.S/ and, being the energy continuous and
differentiable in S, V and N, then the entropy is a continuous and differentiable
function of the energy, volume, and number of moles.
4.2 Extensivity 51

4.2 Extensivity

Extensive Variables

The thermodynamic systems exhibit the property of extensivity. The main con-
sequence of this property is a reduction in number of variables necessary for
the description of a thermodynamic system. For the analysis of the property of
extensivity, we should initially distinguish the variables that are extensible from
those that we call thermodynamic fields, which are not extensive. The volume V, the
internal energy U, the entropy S and the number of moles N are extensive variables.
The pressure p, the temperature T and the chemical potential are thermodynamic
fields. In a system in equilibrium, the thermodynamic fields have the same value in
each part of the system, regardless of the size of each part. The extensive variables
on the other hand take on values that depend on the size of each part.
Let us imagine two copies of the same system such that one of them is an
enlargement of the other in the following sense. The entropies, the volumes, and
the number of moles of the two copies are related by S2 D
S1 , V2 D
V1 and N2 D

N1 , where
is the enlargement factor. The system is described by the fundamental
equation U.S; V; N/ so that the energies are given by U2 D U.S2 ; V2 ; N2 / and
U1 D U.S1 ; V1 ; N1 /. It exhibits the property of extensivity if the energies U1 and
U2 are also related by U2 D
U1 . This means to say that the fundamental relation
obeys the equation

U.S; V; N/ D U.
S;
V;
N/; (4.7)

for any positive values of

. That is, U must be a homogeneous first order function
of S, V, and N.
Deriving both sides of (4.7) with respect to
and then setting
D 1, we get

U D TS  pV C N; (4.8)

which is the Euler equation. From the fundamental relation, the equations of state
are obtained by differentiation. If, however, we do not know the fundamental
relation, but the three equations of state, that is, T, p, and as functions of S, V
and N, then it can be obtained by integration or by replacing into the Euler equation.
In fact, the knowledge of two equations of state plus the property of extensivity
suffices to obtain the fundamental relation.

Molar Quantities

The property of extensivity, given by relation (4.7), enables the description of

a thermodynamic system by a smaller number of independent variables. Such
a description involves quantities called thermodynamic densities, defined as the
52 4 Thermodynamic Potentials

ratio of extensive quantities. Although they are not extensive, the thermodynamic
densities are a class of thermodynamic variables that must be distinguished from the
class of thermodynamic fields. Particularly useful are the thermodynamic densities
obtained by dividing the extensive quantities by the number of moles N, called molar
quantities. Thus, we define the energy per mole, or molar energy, u D U=N, entropy
per mole, or molar entropy, s D S=N and the volume per mole, or molar volume,
v D V=N.
If in the relation (4.7) we set
D 1=N, then the left hand side becomes the molar
energy u and the right hand side becomes a function of two variables only: the molar
entropy s and molar volume v. We may conclude that u.s; v/ is a function s and v
only and that

S V
U.S; V; N/ D Nu. ; /: (4.9)
N N

Therefore, if we know u.s; v/, the fundamental relation can immediately be

recovered by the above formula. That is, u.s; v/ is equivalent to the fundamental
relationship.
From (4.2) and (4.9) we conclude that

@u @u
TD ; pD ; (4.10)
@s @v
which allows to write the differential form for the molar energy,

du D Tds  pdv: (4.11)

The extensivity of S.U; V; N/ allows to conclude also that the thermodynamic

properties can be determined from the molar entropy s.u; v/ as a function of the
molar energy u and the molar volume v. If we know s.u; v/, we can recover the
fundamental relation in the entropy representation through

U V
S.U; V; N/ D Ns. ; /: (4.12)
N N
From this equation and (4.6), we may conclude that

1 @s p @s
D ; D ; (4.13)
T @u T @v
which allows us to write the differential form,

1 p
ds D du C dv; (4.14)
T T
for the molar entropy.
4.2 Extensivity 53

The molar heat capacities cv D Cv =N and cp D Cp =N, or specific heats, as well

as T , s and ˛, may be obtained from the molar quantities
   
@s @s
cv D T ; cp D T ; (4.15)
@T v @T p
     
1 @v 1 @v 1 @v
T D  ; s D  ; ˛D : (4.16)
v @p T v @p s v @T p

Ideal Gas

Let us consider, as an example, an ideal gas whose equations of state pv D RT and

u D cT are written in the form
1 c p R
D ; D : (4.17)
T u T v
Therefore, according to (4.13), we obtain

@s c @s R
D ; D ; (4.18)
@u u @v v
from which we obtain s by the path integral
Z Z
1 p c R
s D s0 C f du C dvg D s0 C f du C dvg: (4.19)
T T u v

Choosing a any path that begins at the reference point .u0 ; v0 /, we get
u v
s D s0 C c ln C R ln : (4.20)
u0 v0

Inverting this relation,

1 v
u D u0 expf Œs  s0  R ln g: (4.21)
c v0

Therefore, the fundamental relations of an ideal gas in the entropy and energy
representations are, respectively

U V
S D Nfs0 C c ln C R ln g (4.22)
Nu0 Nv0
54 4 Thermodynamic Potentials

and
1 S V
U D Nu0 expf Π s0  R ln g: (4.23)
c N Nv0

Performing the derivatives of U with respect to S and V, we recover the equations of

state of the ideal gas, T D U=Nc and p D RU=cV. Indeed, the first gives U D NcT
which, substituted in the second, gives p D RNT=V.

van der Waals Fluid

As a second example, let us consider a van der Waals fluid, that is a fluid that obeys
the van der Waals equation

RT a
pD  (4.24)
v  b v2
and the following equation for the molar internal energy
a
u D cT  : (4.25)
v
From (4.25), we get

1 cv
D ; (4.26)
T uv C a

which can be considered an equation of state in the entropy representation, since the
right hand side depend on u and v. From equation (4.24), we get

p R a 1
D  2 ; (4.27)
T vb v T
or, using equation (4.26),

p R ac
D  ; (4.28)
T v  b v.uv C a/
or
p R c cu
D  C ; (4.29)
T vb v uv C a

which can be considered an equation of state in the entropy representation, since

the right hand side depend on u and v. Integrating these equations of state from a
4.3 Legendre Transformations 55

reference state .u0 ; v0 /, for which the molar entropy is s0 , we have

.uv C a/v0 vb

s D s0 C c ln C R ln : (4.30)
.u0 v0 C a/v v0  b

To obtain the fundamental relation S.U; V; N/, it suffices to do the replacements

s D S=N, u D U=N and v D V=N,

.UV C N 2 a/v0 V  Nb
S D Ns0 C Nc ln C NR ln : (4.31)
.u0 v0 C a/NV .v0  b/N

4.3 Legendre Transformations

Helmholtz Free Energy

The use of the Legendre transformation allows us to get other representations of

the fundamental relation. In other words, it allows the use of other thermody-
namic quantities as independent variables. Let us begin by performing the change
.S; V; N/ ! .T; V; N/. In this new representation the fundamental relation is given
by the Helmholtz free energy F.T; V; N/, defined by the Legendre transformation

F.T/ D minfU.S/  TSg; (4.32)

S

which should be understood as follows. Given a certain value of temperature, we

vary the entropy until we reach the minimum value of the expression inside the
curls, that is, until we find the least difference between U.S/ and TS. Since U.S/ is
differentiable, the minimum occurs when U 0 .S/ D T. Inverting the relation, we get
S.T/ which, replaced in (4.32), gives

F.T/ D U.S/  TS: (4.33)

From (4.33) we obtain F 0 .T/ D S. Therefore the derivatives T.S/ D @U=@S and
S.T/ D @F=@T constitute a pair of inverse functions, as can be seen in Fig. 4.1.
This result provides a way of constructing the Legendre transformation: (a) initially,
we determine the derivative T.S/ of U with respect to S; (b) we take the inverse of
T.S/ to obtain S.T/ which is the derivative of F with respect to T; (c) the integral
of S.T/ gives F.T/ with the opposite sign, except for a constant.
From the relations (4.33) and (4.4), we find

dF D SdT  pdV C dN; (4.34)

56 4 Thermodynamic Potentials

(a) (b)
U F

S T
(c) (d)
T S

S T

Fig. 4.1 Function U.S/ (a) and its Lengendre transform F.T/ (b). Their derivatives (c) T D
@U=@S and (d) S D @F=@T are inverse functions of each other

or equivalently
     
@F @F @F
SD ; pD ; D ; (4.35)
@T VN @V TN @N TV

which are the equations of state.

Enthalpy

The thermodynamic potential called enthalpy H.S; p; N/ is defined by the Legendre

transformation

H. p/ D minfU.V/ C pVg: (4.36)

V

and corresponds to the change of independent variables .S; V; N/ ! .S; p; N/. Since
U.V/ is differentiable, the minimum occurs when U 0 .V/ D p. Inverting p.V/, we
obtain V. p/ which, replaced in the expression between curls, gives

H. p/ D U.V/ C pV: (4.37)

Deriving with respect to p, we have H 0 . p/ D V, so that the derivatives p.V/ D

@U=@V and V. p/ D @H=@p are a pair of inverse functions.
4.3 Legendre Transformations 57

From the relations (4.37) and (4.4), we find

dH D TdS C Vdp C dN; (4.38)

or equivalently
     
@H @H @H
TD ; VD ; D ; (4.39)
@S pN @p SN @N Sp

which are the equations of state.

Gibbs Free Energy

The thermodynamic potential called Gibbs free energy G.T; p; N/ is obtained from
the Helmholtz free energy by the Legendre transformation

G. p/ D minfF.V/ C pVg: (4.40)

V

Using relation (4.34), we get

dG D SdT C Vdp C dN; (4.41)

or equivalently
     
@G @G @G
SD ; VD ; D ; (4.42)
@T pN @p TN @N Tp

which are the equations of state. Alternatively, we may define the Gibbs free energy
from the enthalpy by the Legendre transformation

G.T/ D minfH.S/  TSg; (4.43)

S

which lead us again to relation (4.41) if we use (4.38).

Grand Thermodynamic Potential

Other thermodynamic potentials can be obtained considering transformations that

involve the number of moles. One of them is the grand thermodynamic potential
˚.T; V; /, defined by the Legendre transformation,

˚. / D minfF.N/  Ng; (4.44)

N
58 4 Thermodynamic Potentials

obtained from the Helmholtz free energy. From relation (4.34), we find

d˚ D SdT  pdV  Nd ; (4.45)

which allows to write

     
@˚ @˚ @˚
SD ; pD ; ND : (4.46)
@T V @V T @ TV

Molar Thermodynamic Potentials

The extensivity property of the Helmholtz free energy F.T; V; N/ allows the molar
Helmholtz free energy f .T; v/, which is defined by f D F=N, to be a function of
temperature T and molar volume v only. From f .T; v/, we can recover F.T; V; N/
by

V
F.T; V; N/ D Nf .T; /: (4.47)
N
From this relation, we get

@f @f
sD ; pD ; (4.48)
@T @v
and

df D sdT  pdv: (4.49)

Similarly, from the extensivity of enthalpy, we conclude that the molar enthalpy
h D H=N is a function of the molar entropy s and pressure p only. From h.s; p/ we
recover the enthalpy by

S
H.S; p; N/ D Nh. ; p/: (4.50)
N
From this relation we get

@h @h
TD ; vD ; (4.51)
@s @p

and

dh D Tds C vdp: (4.52)

4.3 Legendre Transformations 59

Due to the extensivity, the Gibbs free energy G.T; p; N/ can be written in the
form

G.T; p; N/ D Ng.T; p/; (4.53)

where g is the molar Gibbs free energy g D G=N and depends only on T and p,
because G is a function of just one extensive variable. From this relation, we get

@g @g
sD ; vD ; (4.54)
@T @p

and

dg D sdT C vdp: (4.55)

In addition, taking into account that D @G=@N, we conclude that D g. That

is, for a system constituted by a pure substance, the chemical potential is identified
with the molar Gibbs free energy. Therefore, from the last equation we obtain the
Gibbs-Duhem equation

d D sdT C vdp: (4.56)

The grand thermodynamic potential ˚ is also a function of just one extensive

variable, so that we may write

˚.T; V; / D V.T; /; (4.57)

where  depends only on the temperature T and chemical potential . The following
relations can be easily obtained:

@ @
sN D  ; N D  ; (4.58)
@T @

where sN D S=V is the entropy per unit volume and N D N=V is the number of moles
per unit volume, or

d D NsdT  d :
N (4.59)

In addition, since p D @˚=@V, then p D , so that

dp D sNdT C d ;
N (4.60)

which is equivalent to the Gibbs-Duhem equation. It suffices to recall that sN D s=v

and N D 1=v.
60 4 Thermodynamic Potentials

Thermodynamic Potentials of an Ideal Gas

From the internal energy of an ideal gas we can determine the Helmholtz free energy
by means of a Legendre transformation. To this end, we replace U D NcT in
expression (4.22) to obtain the entropy

T V
S D Nfs0 C c ln C R ln g (4.61)
T0 Nv0

as a function of T and V, where T0 D u0 =c. Substituting these two results in F D

U  TS, we get
T V
F D TNfc ln C a C R ln g; (4.62)
T0 Nv0
where a D s0  c. The Gibbs free energy is obtained from the transformation G D
F C pV. Using V D NRT=p to eliminate V, we obtain the result
T p
G D NfbT   cT ln C RT ln g; (4.63)
T0 p0
where b D a  R.
The respective molar quantities are given by
T v
s D s0 C c ln C R ln ; (4.64)
T0 v0
T v
f D Tfc ln C a C R ln g; (4.65)
T0 v0
and
T p
g D bT   cT ln C RT ln : (4.66)
T0 p0

Heat, Work and Thermodynamic Potentials

The heat received by a system in an isochoric process equals the variation of the
internal energy, since the work vanishes, that is,

U D Qv : (4.67)

If, on the other hand, we consider an isobaric process, then the variation of energy
will be

U D Qp  pV; (4.68)
4.3 Legendre Transformations 61

because the work done is pV for an isobaric process. But the variation of enthalpy
in an isobaric process is H D U C pV and

H D Qp ; (4.69)

that is, the heat received by the system in an isobaric process equals the variation of
enthalpy.
The work consumed by a system in a quasi-static adiabatic process equals the
variation of the internal energy, since there is no heat exchanged, that is,

U D Wad : (4.70)

If, on the other hand, we consider an isothermal process, then the heat received is
given by QT D TS, so that the variation of internal energy will be

U D TS  WT ; (4.71)

where WT is the work done along an isothermal process. Taking into account that
in an isothermal process the variation of the Helmholtz free energy is F D U 
TS, then

F D WT ; (4.72)

that is, the work consumed by a system along an isothermal process equals the
variation of the Helmholtz free energy.
The isochoric heat capacity Cv is defined as the limit of the ratio Qv =T, when
T ! 0, between the heat introduced at constant volume and the increment in
temperature. Taking into account the equality (4.67), the ratio becomes equal to
U=T along an isochoric line process and therefore

Cv D .@U=@T/V : (4.73)

The isobaric heat capacity Cp is defined as the limit of the ratio Qp =T, when
T ! 0, between the heat introduced at constant pressure and the increment in
temperature. Taking into account the equality (4.69), the ratio becomes equal to
H=T along an isobaric process and therefore

Cp D .@H=@T/p : (4.74)

The molar heat capacities are related with the molar energy and molar enthalpy
by cv D .@u=@T/v and cp D .@h=@T/p .
62 4 Thermodynamic Potentials

4.4 Convexity

Clausius-Gibbs Principle

We saw in the previous chapter that the principle of minimum energy can be
translated by the equation

U.S; V/  U0  T0 .S  S0 /  p0 .V  V0 /; (4.75)

established for a system in which the number of moles is kept constant, where
U0 D U.S0 ; V0 /. To include processes in which the number of moles may vary,
we generalize this inequality to include the number of moles. To this end, we first
divide both sides of (4.75) by the number of moles N to obtain

u.s; v/  u0  T0 .s  s0 /  p0 .v  v0 /; (4.76)

where u0 D u.s0 ; v0 /.
Next, we use the Euler equation in the form u0 D T0 s0  p0 v0 C 0 to eliminate
u0 in expression (4.76) and to obtain the inequality

u.s; v/  T0 s  p0 v C 0 : (4.77)

Multiplying both members of the inequality by the number of moles N, we get

U.S; V; N/  T0 S  p0 V C 0 N; (4.78)

which subtracted from the Euler equation, U0 D T0 S0  p0 V0 C 0 N0 , gives

U  U0  T0 .S  S0 /  p0 .V  V0 / C 0 .N  N0 /; (4.79)

which is the expression of the principle of minimum energy in all three variables.
The entropy obeys the principle of maximum entropy expressed as

1 p0 0
S  S0  .U  U0 / C .V  V0 /  .N  N0 /; (4.80)
T0 T0 T0

which is obtained directly from the inequality (4.79). The inequalities (4.79)
and (4.80) are the expression of the Clausius-Gibbs principle.

Convex and Concave Functions

The principle of minimum energy is equivalent to saying that energy is a convex

function of all extensive variables jointly. Similarly, the principle of maximum
4.4 Convexity 63

(a) (b)
U F

U0 F0

S0 S T0 T

Fig. 4.2 (a) Example of a convex function: the internal energy U.S/ versus entropy S. (b) Example
of a concave function: the Helmholtz free energy F.T/ versus temperature T

entropy is equivalent to saying that entropy is a concave function of all extensive

variables jointly. Furthermore, an important property in respect of convex and
concave functions is that a Legendre transformation converts a convex function
into a concave function and vice versa. That is, both functions that are linked by
a Legendre transformation form a pair of function such that one is convex and the
other is concave. It is therefore appropriate to examine the properties of convex and
concave functions. We analyze initially a convex function of a single variable, the
internal energy U.S/ as a function of the entropy S, and a concave function of a
single variable, the Helmholtz free energy, F.T/ as a function of temperature T. The
variables V and N are taken as constant.
The convexity of a convex or concave function manifest itself if we consider any
tangent to the curve described by the function. All the curve will be entirely on one
side of the tangent line. The convex functions are those curves that lie above any
tangent to the curve whereas the concave curves are those that lie below any tangent
to the curves, as shown in Fig. 4.2.
Consider the tangent to the convex function U.S/ at the point .S0 ; U0 /, where
U0 D U.S0 /. The equation that describes the tangent line is

y.S/ D U.S0 / C U1 .S0 /.S  S0 /; (4.81)

where U1 .S0 / is the slope of the tangent line. Taking into account that U.S/ is
differentiable, then U1 .S0 / may be identified with the derivative of U calculated
at S D S0 . Since U.S/ is a convex function, then U.S/  y.S/, that is,

U.S/  U.S0 / C U1 .S0 /.S  S0 /; (4.82)

valid for any value of S, as can be seen in Fig. 4.2a. This inequality is the same
inequality (4.79) for the case in which the volume and the number of moles remains
constant. It suffices to recall that U1 .S0 / D T0 .
64 4 Thermodynamic Potentials

A convex function U.S/ has two fundamental properties:

(a) The derivative U1 .S/ is monotonic increasing function of S.
(b) If the function U.S/ has the second derivative U11 .S/ at a certain point S D S0 ,
then the convexity condition implies

U11 .S0 /  0: (4.83)

Let us examine now the concave function F.T/. Similarly, we consider a tangent
to the curve described by F.T/ at a point .T0 ; F0 /, where F0 D F.T0 /. The heights
of the points on the tangent line is given by

y.T/ D F.T0 / C F1 .T0 /.T  T0 /; (4.84)

where F1 .T0 / is the slope of the line. At the points where F.T/ is differentiable,
F1 .T0 / is identified with the derivative of F.T/ calculated at T0 . Since F.T/ is
concave then F.T/  y.T/, that is,

F.T/  F.T0 / C F1 .T0 /.T  T0 /; (4.85)

valid for any value of T, as seen in Fig. 4.2b.

The concave function F.T/ has the fundamental properties:
(a) The derivative of F1 .T/ is a monotonic decreasing function of T.
(b) If the function F.T/ has the second derivative F11 .T/ at a certain point T D T0 ,
then the convexity condition implies

F11 .T0 /  0: (4.86)

Convex and Concave Function in Many Variables

Equation (4.82) may be generalized to the case of a function of more than one
variable. Let us consider initially the case of a function of two variable: the internal
energy U.S; V/ as a function of the entropy S and the volume V, only. We are
assuming that the number of moles N remains constant. Taking into account that
U.S; V/ is differentiable, then its convexity is expressed by the inequality

U.S; V/  U.S0 ; V0 / C U1 .S0 ; V0 /.S  S0 / C U2 .S0 ; V0 /.V  V0 /; (4.87)

where U1 .S0 ; V0 / and U2 .S0 ; V0 / are the derivatives of U.S; V/ with respect to S and
V, respectively, calculated at the point .S0 ; V0 /. The right hand site of this inequality
describes a tangent plane to the surface described by U.S; V/ at the point .S0 ; V0 /.
The surface lies entirely above the plane. This inequality is identified with (4.79). It
suffices to recall that U1 .S0 ; V0 / D T0 and U2 .S0 ; V0 / D p0 .
4.4 Convexity 65

Since U.S; V/ is convex in the two variables jointly, it follows that U11 .S0 ; V0 /,
U12 .S0 ; V0 / and U22 .S0 ; V0 / satisfy the properties

U11  0; U22  0 (4.88)

and

.U12 /2  U11 U22  0: (4.89)

Next we analyze the Helmholtz free energy F.T; V/ as a function of temperature

T and volume V, holding the number of moles N fixed. The Helmholtz free energy
is obtained from the internal energy by means of a Legendre transformation. This
makes F.T; V/ a concave function of T while remaining a convex function of V. It
therefore obeys the inequalities

F.T; V/  F.T; V0 / C F2 .T; V0 /.V  V0 / (4.90)

and

F.T; V/  F.T0 ; V/ C F1 .T0 ; V/.T  T0 /; (4.91)

where the coefficients F1 .T; V0 / and F2 .T0 ; V/ of the tangent lines are identified
with the derivatives of F.T; V/ with respect to T and V, respectively, calculated at
the points where F.T; V/ is differentiable. The Helmholtz free energy F.T; V/ is an
example of a concave-convex function.
Since F.T; V/ is concave in T and convex in V, it follows that its second
derivatives F11 .T0 ; V0 / and F22 .T0 ; V0 / satisfies the properties

F11  0; F22  0: (4.92)

The enthalpy is also obtained from the internal energy by a Legendre transfor-
mation. For fixed N, the enthalpy H.S; p/ is a convex function of S and concave in
p and obeys the inequalities

H.S; p/  H.S0 ; p/ C H1 .S0 ; p/.S  S0 / (4.93)

and

H.S; p/  H.S; p0/ C H2 .S; p0 /. p  p0 /; (4.94)

where the coefficients H1 .S0 ; p/ and H2 .S; p0 / of the tangent lines are identified
with the derivatives of H.S; p/ with respect to S and p, respectively, calculated at
the points where H.S; p/ is differentiable. The enthalpy is also an example of a
concave-convex function.
66 4 Thermodynamic Potentials

Since H.S; p/ is convex in S and concave in p, it follows that the second

derivatives H11 .S0 ; p0 / and H22 .S0 ; p0 / satisfies the properties

H11  0; H22  0: (4.95)

We now examine the Gibbs free energy G.T; p/ as a function of temperature

T and pressure p, at constant number of moles N. The Gibbs free energy is
obtained from the internal energy U.S; V/ by means of two successive Legendre
transformations. This makes G.T; P/ concave in both variables T and p, jointly. The
convexity manifests itself by the inequality

G.T; p/  G.T0 ; p0 / C G1 .T0 ; p0 /.T  T0 / C G2 .T0 ; p0 /. p  p0 /; (4.96)

where G1 .T0 ; p0 / and G2 .T0 ; p0 / are the coefficients that defines the slope of the
tangent plane and are identified with the derivatives of G.T; p/ with respect to T and
p, respectively, calculated in the points where G.T; p/ is differentiable.
Since G.T; p/ is concave in two variable jointly, it follows that its second
derivatives G11 .T0 ; p0 /, G12 .T0 ; p0 / and G22 .T0 ; p0 / satisfies the properties

G11  0; G22  0 (4.97)

and

.G12 /2  G11 G22  0: (4.98)

Convex Hull

In some situations, the description of a system in thermodynamic equilibrium is

done by means of a potential defined as the convex hull of a function F.V/, as
the one shown in Fig. 4.3a, which is devoid of the convexity property. The convex
hull Fch .V/ of F.V/ is obtained as follows. We start by determining the Legendre
transformation

G. p/ D minfF.V/ C pVg; (4.99)

V

which is a concave function. Next, we perform the inverse Legendre transformation

Fch .V/ D minfG. p/  Vpg; (4.100)

p

which gives us a convex function, the convex hull of F.V/.

The convex hull Fch .V/ may be obtained more simply by the construction of a
double tangent, as seen in Fig. 4.3a. The double tangent AB gives rise to a horizontal
4.4 Convexity 67

(a) (b)
F p

A
FA
A´ B´
p*
C
B
FB

VA VB V VA VC VB V

Fig. 4.3 (a) Convex hull. (b) Maxwell construction

line segment A0 B0 in the derivative pch .V/ D @Fch =@V of the convex hull, as seen
in Fig. 4.3b. This horizontal segment guarantees that pch .V/ is a monotonic function
although the derivative p.V/ D @F=@V of the original function does not have this
property.
To set up the convex hull Fch .V/ it suffices to determine VA and VB , the values of
V at the tangential points A and B, respectively. Let p be the slope of the double
tangent. According to Fig. 4.3a,

F.VA /  F.VB / D p .VB  VA /: (4.101)

Taking into account that the slopes at the tangential points A and B are equal, then

p.VA / D p ; p.VB / D p : (4.102)

These three equations determine p , VA and VB from F.V/ and its derivative.
Equations (4.101) and (4.102), that determine the localization of the double
tangent AB and therefore the horizontal segment A0 B0 , are valid even if F.V/ has a
singular behavior in the interval between the points VA and VB , as happens when
F.V/ is defined by two unconnected branches. When the function F.V/ is well
behaved, the position of the segment A0 B0 is such that the area above A0 B0 is equal
to the area below A0 B0 , as seen in Fig. 4.3b. The trace of the segment A0 B0 made
according to this result is called Maxwell construction and is equivalent to the
making of the double tangent. To demonstrate this result we proceed as follows.
Integrating p.V/ between VA and VB , we get
Z VB
p.V/dV D F.VA /  F.VB /: (4.103)
VA

Using result (4.101), we see that

Z VB
p.V/dV D p .VB  VA /; (4.104)
VA
68 4 Thermodynamic Potentials

or
Z VB
fp  p.V/gdV D 0: (4.105)
VA

If we denote by C the point where the segment A0 B0 cuts the curve p.V/, then we
can separate the integral in two and obtain
Z VC Z VB
fp  p.V/gdV D fp.V/  p gdV; (4.106)
VA VC

which is the desired result since the right and left hand sides of these equations are,
respectively, equal to the areas above and below the segment that join the two points
A0 and B0 .

Fundamental Properties of the Thermodynamic Potentials

Below we summarize the properties of the thermodynamic potentials, which are

obtained by successive Legendre transformations. We recall initially that the internal
energy is a function of extensive variables only. The thermodynamic fields are
obtained by differentiation of the internal energy. A given thermodynamic field
obtained by differentiation of the internal energy with respect to a certain extensive
variable, constitutes with it a pair of conjugate variables.
At each Legendre transformation, a extensive variable is replaced by its conjugate
thermodynamic field. Each thermodynamic potential is, therefore, a function of a
set of extensive variables and a set of thermodynamic fields, which constitute the
thermodynamic space related to this potential. Notice that just one component of a
pair of conjugate variables makes part of the thermodynamic space. The other pair
component is obtained by differentiation of the thermodynamic potential.
The fundamental properties of a thermodynamic potential associated to a ther-
modynamic space are the followings:
(a) Continuity of the potential.
The potential is a continuous function of all variables of the thermodynamic
space.
(b) Continuity of the thermodynamic field.
The potential is differentiable with respect to the extensive variables of
the thermodynamic space. This property is equivalent to the following.
A thermodynamic field obtained by differentiation of the potential is a
continuous function of all variables of the thermodynamic space.
(c) Extensivity of the potential.
A thermodynamic potential is an extensive function of the extensive variable or,
equivalently, it is homogeneous of the first order in the extensive variables.
4.4 Convexity 69

(d) Convexity of the potential.

A thermodynamic potential is a convex function of the set of extensive variables
and a concave function of the set of thermodynamic fields.
The thermodynamic potentials related to the thermodynamic space containing
just one extensive variable are particularly useful, mainly in the study of phase
transitions. The thermodynamic quantity defined as the ratio between one of such
potentials and the extensive variable depends only on the thermodynamic fields and
is also identified as a thermodynamic field.

Problems

4.1 For each of the fundamental relations below, show that S is extensive. Deter-
mine the equations of state for each case. Find the fundamental relation in the energy
representation.
(a) S D A.UVN/1=3 , (b) S D a.U 2 V 3 N/1=6 ,
(c) S D B.U 3 V/1=4 , (d) S D b.UV/1=2 .
4.2 For each of the fundamental relations below, show that U is extensive.
Determine the equations of state for each case.
(a) U D aS3 =VN, (b) U D AS3 =.V 3 N/1=2 ,
(c) U D b.S4 =V/1=3 , (d) U D BS2 =V.
4.3 For each of the potentials of the previous problem, determine F and H by means
of Legendre transformations. For the potentials of items (a) and (b) determine also
G and ˚.
4.4 Determine the three equations of state for the system that obeys the equation
u D av 1 s2 es=R .
4.5 Find the fundamental equation of a system that obeys the following relations
U D pV and p D BT 2 . Do the same for the system that follows the equations
u D 3pv=2 and u1=2 D BTv 1=3 .
4.6 Determine the molar entropy s.u; v/ of a system whose equations of state are
p D aT 4 and u D 3pv.
4.7 Show that the compressibilities are related with the density N by 
N T D
.@=@p/
N T e 
N s D .@=@p/
N s .
4.8 Consider the equations of state

1 a p c
D C bv; D C f .u/:
T u T v

Determine f .u/ and the fundamental equation knowing that f .0/ D 0.

70 4 Thermodynamic Potentials

4.9 From the equations p D RT=v and cv D c, valid for an ideal gas, obtain the
fundamental relation in the Helmholtz free energy representation.
4.10 Show the following relations between pressure p, density N and chemical
potential of an ideal gas

p D a e =RT ; N D b e =RT ;

where a and b depend only on the temperature. Determine a and b.

4.11 Obtain the fundamental relation of an ideal gas in the grand thermodynamic
potential representation.
4.12 From the van der Waals equation
RT a
pD  ;
v  b v2
and assuming that the molar heat capacity is a constant, cv D c, obtain the
fundamental relation of a van der Waals fluid in the Helmholtz free energy
representation. Show that u D cT  a=v.
4.13 From the fundamental relation of a van der Waals fluid in the entropy
representation, obtain the fundamental relation in the energy representation. From
this relation determine the fundamental relation in the Helmholtz free energy
representation by a Legendre transformation.
4.14 Show that the functions of the problem 4.1 are concave and those of
problem 4.2 are convex.
4.15 Determine the conditions that must be imposed on the parameters a, b and c
so that S D r U a V b N c is extensive and has the convexity property.
4.16 Show that the functions x2 ,  ln x, ex are convex. Determine their Legendre
transformations and show that they are concave functions.
4.17 Determine the convex hull of the functions expfjxj  1g  jxj and .jxj  1/2 .
4.18 Show that the functions

.jxj  1/2 ; jxj  1;
f .x/ D
0; jxj < 1

and

expfjxj  1g  jxj; jxj  1;
f .x/ D
0; jxj < 1

are convex. Determine, for each one of them, the Legendre transform g. p/ D
minx f f .x/  pxg and show that it is concave. Sketch the plots of f .x/, f 0 .x/, g. p/,
and g0 . p/. Show that f 0 .x/ and g0 . p/ are inverse of each other.
4.4 Convexity 71

4.19 Determine the convex hull of the function

f .x/ D x4  ax2 C bx;

where a and b are constants and a > 0. Find the points x1 and x2 corresponding to
the double tangent as functions of a and b. Sketch the plots of f .x/, of the convex
hull and of their derivatives.
Chapter 5
Thermodynamic Identities

5.1 Consistency of the Equations of State

Exact Differential

We have seen that the equations of state are equations deduced from the fundamental
relations by differentiation of the thermodynamic potential. For this reason, there is
an inter-relationship among them which we will reveal. Consider the equations of
state
1 p
D s1 .u; v/ and D s2 .u; v/ (5.1)
T T

in the entropy representation. Since 1=T D @s=@u and p=T D @s=@v, then

@s @s
D s1 .u; v/ and D s2 .u; v/: (5.2)
@u @v

Taking into account that @2 s=@v@u D @2 s=@u@v, we conclude that

@ @
s1 .u; v/ D s2 .u; v/; (5.3)
@v @u

that is, the functions s1 .u; v/ and s2 .u; v/ and, therefore, the equations of state (5.1)
cannot be arbitrary. They must obey the consistency relation given by (5.3). This
condition is equivalent to saying that

1 p
ds D du C dv (5.4)
T T
is an exact differential.

© Springer-Verlag Berlin Heidelberg 2017 73

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_5
74 5 Thermodynamic Identities

As an example, let us imagine that a given system obeys the equations of state,

p D Tw.v/  q.v/ (5.5)

and

u D cT  r.v/; (5.6)

where w.v/, q.v/ and r.v/ are functions of v only and c is a constant. From (5.6),

1 c
D D s1 .u; v/: (5.7)
T u C r.v/

From (5.5) and using (5.7), we get

p c
D w.v/  q.v/ D s2 .u; v/: (5.8)
T u C r.v/

Using now the consistency equation (5.3), we reach the result

q.v/ D r0 .v/; (5.9)

which says that the functions q.v/ and r.v/ cannot be arbitrary. For the van der
Waals fluid, w D R=.v  b/ and r D a=v, so that q D a=v 2 .

Maxwell Relations

Consider the potentials U, F, H and G and their differentials

dU D TdS  pdV; (5.10)

dF D SdT  pdV; (5.11)
dH D TdS C Vdp; (5.12)
dG D SdT C Vdp: (5.13)

Since they are all exact differentials then

   
@T @p
D ; (5.14)
@V S @S V
   
@S @p
D ; (5.15)
@V T @T V
5.2 Identities 75

   
@T @V
D ; (5.16)
@p S @S p
   
@S @V
 D ; (5.17)
@p T @T p

which are the Maxwell relations.

The Maxwell relations are useful, for example, when one wishes to obtain
information about the entropy from the equation of state p.T; V/. The second
Maxwell relation above says that the isothermal variation of the entropy with the
volume is equal to the isochoric increase of pressure with temperature. These
relations are also useful to obtain further thermodynamic identities as we shall see
below.

5.2 Identities

Identities Involving Derivatives

In the manipulation of the thermodynamic derivatives it is necessary to use certain

mathematical identities, which we now demonstrate. Consider five generic variables
X, Y, Z, U, and W, assume that X.Y; Z/ is a function of Y and Z and that Y.U; W/
and Z.U; W/ are functions of U and W. Then by the chain rule,
         
@X @X @Y @X @Z
D C : (5.18)
@U W @Y Z @U W @Z Y @U W

1. For the case U D Y,

       
@X @X @X @Z
D C : (5.19)
@Y W @Y Z @Z Y @Y W

2. For the case W D Z,

     
@X @X @Y
D : (5.20)
@U Z @Y Z @U Z

3. For W D X and U D Y,
     
@X @X @Z
0D C (5.21)
@Y Z @Z Y @Y X
76 5 Thermodynamic Identities

or
   @X 
@Z @Y Z
D   @X  : (5.22)
@Y X @Z Y

4. For U D X and W D Z,
   
@X @Y
D1 (5.23)
@Y Z @X Z

or
 
@Y 1
D  @X  : (5.24)
@X Z @Y Z

Relation Between the Heat Capacities

As an example of the use of these equalities, we see, from (5.22),

   @V 
@p @T p ˛
D   D : (5.25)
@T V @V T
@p T

Taking into account that T  0, this equality says that bodies that expand with
temperature (˛ > 0), get hotter in an isochoric compression (.@p=@T/V > 0). Those
that contract with temperature (˛ < 0), get colder in an isochoric compression
(.@p=@T/V < 0).
Another example can be obtained from (5.19). From this equation,
       
@S @S @S @V
D C ; (5.26)
@T p @T V @V T @T p

or, by Maxwell relation (5.15),

       
@S @S @p @V
D C : (5.27)
@T p @T V @T V @T p

Using the definitions of the thermodynamic coefficients and the result (5.25) for
.@p=@T/V , we get

˛2
Cp D Cv C TV ; (5.28)
T

which gives the relation between the heat capacities.

5.2 Identities 77

This relation, which is one of the best known of thermodynamics, is useful in

determining Cv for solids and liquids because in these cases it is very difficult
to make measurements of quantities defined at constant volume. It is easier to
make measurements of quantities defined at constant pressure, as is the case of the
quantities Cp and ˛. As the second term on the right hand side of (5.28) is never
negative, we conclude that the heat capacity at constant pressure is always greater
than or equal to the heat capacity at constant volume, Cp  Cv .
Another relation particularly useful is obtained from (5.22), namely,
   @V 
@V @S p
D   : (5.29)
@p S
@p
@S V

Using the relation (5.20), we can write the numerator and the denominator in the
forms
       
@V @V @T @V T
D D (5.30)
@S p @T p @S p @T p Cp

and
       
@p @p @T @p T
D D ; (5.31)
@S V @T V @S V @T V Cv

so that
   @V   
@V @T p
Cv @V Cv
D   D ; (5.32)
@p S
@p
Cp @p T Cp
@T V

where we have used again the relation (5.22). From the definitions of the compress-
ibilities, we get

T Cp
D : (5.33)
s Cv

The ratio T =s between the compressibilities is therefore equal to the ration Cp =Cv
between the heat capacities. Replacing (5.28) into (5.33), we find the result

˛2
T D s C TV : (5.34)
Cp

Since the second term is never negative, then the isothermal compressibility is
always greater or equal to the adiabatic compressibility, T  s .
78 5 Thermodynamic Identities

Reduction of Derivatives

A thermodynamic system can be described in various ways, according to the

representation we have chosen. This variety of descriptions leads us to a large
number of equivalent results, which can be revealed by the thermodynamic identities
given above. In general, these identities involve the derivatives of thermodynamic
variables. To achieve identities in a systematic way, we present a scheme to reduce
a given derivative in terms of other previously chosen. In the scheme we present
below, we chose the following quantities: the heat capacity at constant pressure, the
coefficient of thermal expansion and the isothermal compressibility, given by
     
@S 1 @V 1 @V
Cp D T ; ˛D ; T D  ; (5.35)
@T p V @T p V @p T

respectively.
Notice that the expressions derived earlier for the heat capacities and for the
compressibilities are under this scheme. In fact, from (5.28) and (5.34),

˛2 ˛2
Cv D Cp  TV ; s D T  TV : (5.36)
T Cp

The rules for the reduction of derivatives are as follows.

1. Bring the thermodynamic potentials U, F, H and G to the numerator, using
identities (5.22) or (5.24) and, next, eliminate them, using the identities (5.10)–
(5.13).
2. Bring the entropy to the numerator, using identities (5.22) or (5.24) and, then
eliminate them, by the use of the Maxwell relations (5.15) and (5.17) or through
the definitions of the heat capacities Cp or Cv . If these two alternatives are not
appropriate, use the identities
       
@S @S @T Cv @T
D D (5.37)
@p V @T V @p V T @p V

or
       
@S @S @T Cp @T
D D : (5.38)
@V p @T p @V p T @V p

3. The remaining derivatives are of the type .@X=@Y/Z where X, Y and Z are V, p,
and T. Therefore, they can be eliminate by bringing the volume to the numerator
and using the definition of ˛ and T .
5.3 Applications 79

5.3 Applications

Compression

We examine two types of compression: the isothermal and adiabatic. In an isother-

mal compression we are interested in determining the heat given way Q D TS
when the pressure increases by an amount p. The appropriate quantity is then
 
@S
T : (5.39)
@p T

Using the rules above,

   
@S @V
T DT D TV˛: (5.40)
@p T @T p

In an adiabatic compression we are interested in determining the variation of

temperature T when the pressure increases by an amount p. The appropriate
quantity is then
 
@T
: (5.41)
@p S

Using the rules above,

   @V 
  @S
@T @p T @T p TV˛
D  @S  T D D : (5.42)
@p S @T p
Cp Cp

Since Cp  0, this equality allows us to say that the bodies that expand with
temperature (˛ > 0), get hotter in an adiabatic compression (.@T=@p/S > 0). Those
that contract with temperature (˛ < 0), get colder in an adiabatic compression
(.@T=@p/S < 0).

Free Expansion

In a free expansion the internal energy is constant and the volume of the system
increases. If we wish to determine whether there is a variation T in temperature
with the change V in volume it is appropriate to calculate the ratio .T=V/ at
constant energy, that is
 
@T
: (5.43)
@V U
80 5 Thermodynamic Identities

Using the scheme for the reduction of derivatives, we start from the identity (5.22)
to get
   @U 
@T @V T
D   @U  : (5.44)
@V U @T V

Next, we use the chain rule and the identity (5.10) to get
   
@U @S
DT p (5.45)
@V T @V T

and
   
@U @S
DT : (5.46)
@T V @T V

To eliminate the entropy, we use the Maxwell relation (5.15) in the first and the
definition of Cv in the second, obtaining
   
@U @p
DT p (5.47)
@V T @T V

and
 
@U
D Cv : (5.48)
@T V

Finally, we use the identity (5.22) in the first to get

 
@U ˛
DT  p: (5.49)
@V T T

Then,
 
@T 1 ˛
D .p  T /: (5.50)
@V U Cv T

It is worth noting that, for an ideal gas, ˛ D 1=T and T D 1=p, leading to
the results .@U=@V/T D 0 and .@T=@V/U D 0. From these results, we conclude
that the internal energy of an ideal gas does not depend on the volume, but only
on temperature or, equivalently, that there is no variation in temperature in a free
expansion of an ideal gas.
5.3 Applications 81

Joule-Thomson Process

In a Joule-Thomson process, also called throttling process, a gas is forced through

a porous wall so that the enthalpy of the gas remains unchanged. The pressure
difference on both sides of the wall is p and we wish to know the temperature
variation T. For small pressure differences, the variation of temperature is T D
JT p where JT is the Joule-Thomson coefficient
 
@T
JT D : (5.51)
@p H

From the identity (5.22),

 
  @H
@T @p
D  @H  T : (5.52)
@p H @T p

Using the chain rule and the identity (5.12), we get

   
@H @S
DT CV (5.53)
@p T @p T

and
   
@H @S
DT : (5.54)
@T p @T p

Using Maxwell relation (5.17) in the first equation and the definition of Cp in the
second, then
   
@H @V
D T C V D TV˛ C V (5.55)
@p T @T p

and
 
@H
D Cp : (5.56)
@T p

Therefore
 
@T V
D .T˛  1/: (5.57)
@p H Cp
82 5 Thermodynamic Identities

For an ideal gas ˛ D 1=T, so that .@H=@p/T D 0 and .@T=@p/H D 0. From these
results we conclude that the enthalpy of an ideal gas do not depend on pressure, but
only on temperature or, equivalently, that there is no variation in temperature of an
ideal gas under a Joule-Thomson process.

5.4 Gas Properties

Virial Expansion

For sufficiently low densities, the gases follows the ideal behavior pV D NRT which
we write as

p D RT ;
N (5.58)

where N D N=V is the number of moles per unit volume. The plot represented
in Fig. 5.1a shows the pressure of hydrogen as a function of density for several
temperatures. For low densities we see that indeed p is proportional to N and
therefore the hydrogen behaves as an ideal gas in this regime. However, deviations
from this behavior can be seen in this figure, if the density is not sufficiently small.
To examine the deviation from the ideal behavior it is convenient to plot the
compressibility factor Z D p=RT N D pv=RT as a function of , N where v D V=N D
1=N is the molar volume. Figure 5.1b shows such a plot for hydrogen. It is seen
that Z deviates from the value Z D 1, valid for an ideal gas. We may represent the

(a) 50 (b) 1.20

+150 +150
+25 +25
40 -25 1.15
-25
-75 -75
-120
p (MPa)

-120
30 -150 1.10
pv/RT

-150
-175 -175
20 1.05

10 1.00

0 0.95
0 5 10 15 20 0 2 4 6 8 10
_ 3
_ 3
ρ (mol/dm ) ρ (mol/dm )

Fig. 5.1 Isotherms of the hydrogen obtained experimentally by Michels et al. [106] at various
temperatures between 175 and 150 ı C. (a) Pressure p versus number of moles per unit volume
.
N (b) Compressibility factor Z D pv=RT versus N
5.4 Gas Properties 83

behavior of the compressibility factor Z by the expression

pv
D 1 C B2 N C B3 N 2 C : : : ; (5.59)
RT
called virial expansion. The coefficients B2 , B3 , . . . , depend only on temperature.
Equation (5.59) may describe the general behavior of gases, as long as the
expansion contains a sufficient number of terms. Alternatively, it is possible to
describe the general behavior by means of approximate equations of state such as
the van der Waals equation
a
.p C /.v  b/ D RT; (5.60)
v2
the Dieterici equation

p.v  b/ D RTea=RTv ; (5.61)

and the Berthelot equation

a
.p C /.v  b/ D RT: (5.62)
v2 T
These equations of state, unlike the virial expansion, are capable of describing the
liquid phase, in addition to the gas phase, and the transition between them.

Measurement of Compressibility

An experiment to measure the compressibility of gases is performed as follows. We

consider a vessel with adiabatic walls, in the form of a bottle, which encloses the
gas whose compressibility we want to measure. In the neck of the bottle, which has
a cylindrical shape, there is a small object of mass m, that can move without friction.
The weight of the object is such that it is in equilibrium position when the volume
of the gas is V0 and the pressure of the gas is p0 . The object is then displaced from
its equilibrium position and starts to oscillate with a certain angular frequency !.
The equation of motion of the object is given by

d2 x
m D .p  p0 /A; (5.63)
dt2
where A is the cross-sectional area of the bottle neck. The pressure at the equilibrium
point p0 is equal to the pressure due to the weight of the object plus the atmospheric
air pressure patm , that is, p0 D patm C mg=A, where g is the acceleration of gravity.
We are assuming that the equilibrium point occurs at x D 0, so that the volume
depends on x as V D V0 C Ax.
84 5 Thermodynamic Identities

Assuming that the oscillations are adiabatic, the pressure and the volume are
connected by an adiabatic line, so that, for small oscillations, that is, for small
volume variations,
 
p  p0 @p 1
D D ; (5.64)
V  V0 @V S V 0 s

and therefore

d2 x
m D kx; (5.65)
dt2

where the constant k is given by k D A2 =s Vp 0 . But the angular frequency ! of

oscillations is related to the constant k by ! D k=m, from which we get
s
A2
!D : (5.66)
s V 0 m

Therefore, measuring the angular frequency !, knowing the dimensions of the

bottle, that is, the volume V0 and the area A, and the mass m of the object, one
can determine the adiabatic compressibility.
If we know the isothermal compressibility T independently, then this experiment
allows us to determined the ratio between the heat capacities  D Cp =Cv because,
according to (5.33), s D T = . For instance, if the gas can be considered ideal then
T D 1=p0 . p
The speed of sound in a gas is given by vsom D .@p=@/s where  is the gas
density, related to N by  D M N where M is the gas molar mass. Since .@p=@/s D
1=s , the speed of sound is related to the gas compressibility in accordance with
s r
1 
vsom D D : (5.67)
s T

The second equality is obtained from the relation s D T = . If we know the

adiabatic compressibility, or if we know the isothermal compressibility and  , we
can determine the speed of sound. If the gas may be considered as ideal, then
T D M =p
N D M=RT so that
r
 RT
vsom D : (5.68)
M

Considering the atmospheric air as an ideal gas, we can use formula (5.68) to
compute the speed of sound. For normal conditions of pressure and temperature,
that is, pressure of 1 atm and temperature of 0 ı C, the experimental values are
5.4 Gas Properties 85

 D 1:293 kg/m3 and  D 1:403. With these data we find vsom D 331:6 m/s. For
comparison, the value of the speed of sound in atmospheric air measured by direct
means vsom D 331:5 m/s.

Inversion Curve

One way of achieving lower temperatures is through the Joule-Thomson process.

To decrease the temperature of the gas when it passes through the porous wall, it
is necessary that the Joule-Thomson coefficient JT is positive since the pressure
of the gas necessarily decreases. For a given gas, we can draw in the p-T diagram
the region for which JT > 0, which is the region useful for cooling through the
Joule-Thomson process. This region is bounded by the curve JT D 0, called the
inversion curve.
Let us determine the inverse curve for a fluid obeying the van der Waals
equation (5.60). According to the result (5.57), the inversion curve is given by
T D 1=˛, that is, by
 
@T
T Dv : (5.69)
@v p

Determining the derivative @T=@v from (5.60) and substituting into the inversion
curve equation, we arrive at the result
 
RTb b 2
D 1 : (5.70)
2a v

Solving for v and substituting into (5.60), we obtain the inversion curve in the p-T
diagram, given by
r ! r !
a bRT bRT
pD 2 1 3 1 : (5.71)
b 2a 2a

Figure 5.2 shows the inversion curve, according to this equation and the inverse
curve of nitrogen, obtained experimentally. It is seen that the inversion curve
reaches the liquid-vapor coexistence line. We shall see later that the van der Waals
equation in fact provides a liquid-vapor coexistence line and a critical point, shown
in Fig. 5.2a. The critical temperature Tc and critical pressure pc are related to the
parameters a and b by Tc D 8a=27bR and pc D a=27b2.
86 5 Thermodynamic Identities

10 40
(a) (b)

8
30
μJT<0 μJT<0

p (MPa)
6
p/pc

20
4 μJT>0 μJT>0

10
2

0 0
0 2 4 6 8 100 200 300 400 500 600

T/Tc T (K)

Fig. 5.2 (a) Inversion curve for a van der Waals fluid according to (5.71). (b) Inversion curve for
nitrogen according to Roebuck and Osterberg [107]. In both plots, the thick curve represents the
liquid-vapor coexistence and the full circle, the critical point

Molar Heat Capacity

The molar heat capacity of the ideal gases, that is, those that obey the equation of
state pv D RT, depends only on temperature. In fact, from (5.47) it is seen that for
an ideal gas .@u=@v/T D 0 and, therefore, the molar internal energy u is independent
of v. Since the molar enthalpy h D u C pv D u C RT, then h is also independent of
v. Therefore, the molar heat capacities cv D .@u=@T/v and cp D .@h=@T/p , depend
only on T. Moreover, taking into account that h D u C RT then cp D cv C R.
Figure 5.3a shows cp of hydrogen as a function of temperature, for several values of
pressure. Notice that, at high temperatures, cp becomes independent of pressure and
approaches the value .7=2/R.
Figure 5.3b shows cp of several gases that have behavior close to the ideal
behavior. We assume that for these gases cv is a sum of the terms

cv D ctrans C crot C cvib (5.72)

corresponding to the translation motion, rotation and vibration of the molecules.

According to the equipartition of energy ctrans D .3=2/R. The rotational part is
crot D R for diatomic gases and crot D .3=2/R for polyatomic gases. Therefore,
for monoatomic gases ideal gases that have only translational degrees of freedom,
cv D .3=2/R and cp D .5=2/R as seen in Fig. 5.3b for argon.
For diatomic ideal gases, cv D .5=2/R and cp D .7=2/R, as seen in Fig. 5.3b
for nitrogen and approximately for carbon monoxide. Hydrogen has the same
behavior as long as the temperature is not so low. In the low temperature regime,
hydrogen behave as a monoatomic gas. The behavior of cvib with temperature is
more complicated and will be object of study in the next chapter.
5.4 Gas Properties 87

(a) 4.0 (b) 8

CO2
7 H2 O
CO
3.5 6 N2
H2
5
Ar
cp /R

cp /R
3.0 4
100
10 3
1
2.5 0 2

2.0 0
0 200 400 600 0 200 400 600 800

T (K) T (K)

Fig. 5.3 Molar isobaric heat capacity as a function of temperature of (a) hydrogen for the
pressures indicated in atmospheres and (b) several gases at the pressure of 1 atm. Source: TT

Problems

5.1 Show that

   
@u @p
DT  p:
@v T @T v

Use this identity to show that, for a fluid that obeys the equations of state of the type

p.T; v/ D Tw.v/  q.v/ and u.T; v/ D cT  r.v/:

where w.v/, q.v/ and r.v/ depend only on v, the functions q.v/ and r.v/ are related
by r0 .v/ D q.v/. Determine r.v/ for the van der Waals gas for which w D R=.v 
b/ and q.v/ D a=v 2 .
5.2 Reduce the following derivatives to expressions containing ˛, T and Cp .
           
@T @V @T @T @T @V
; ; ; ; ; :
@p V @S p @V U @p H @p U @p S

5.3 Show the following expressions for the heat capacity CVN D T.@S=@T/VN D
.@U=@T/VN at constant volume and number of moles,
   
.@N=@T/2V @U @U .@N=@T/V
CVN D CV  T D  ;
.@N=@ /VT @T V @ VT .@N=@ /VT

where CV D T.@S=@T/V is the heat capacity at constant volume but at constant

chemical potential. These expressions allows the determination of CVN from S, U
and N as functions of T, V and .
88 5 Thermodynamic Identities

5.4 Show that the internal energy of gases that obey the van der Waals, the Dieterici
and the Berthelot equations must depend on volume, besides temperature. Show that
this is also valid for the virial expansion as long as the virial coefficients are not
independent of temperature.
5.5 Determine the second virial coefficient B2 corresponding to the van der Waals,
Dieterici and Berthelot equations of state.
Chapter 6
Nernst-Planck Principle

6.1 Nernst Postulate

Entropy

At sufficiently low temperatures, it is experimentally shown that the thermal

capacity of the solids decreases appreciably and vanishes in the limit of zero
absolute temperature. This behavior however does not necessarily follow from the
laws of thermodynamics seen so far. It is one of the consequences of the Nernst
postulate according to which the entropy of a system does not decrease without
limits, but approaches a constant in the limit of zero absolute temperature, that is,

S.T/ ! S0 when T ! 0: (6.1)

Although the existence of a lower bound for the entropy seems to be quite
evident, we must remember that this does not occur in a system whose heat capacity
is constant, like an ideal gas whose internal energy is linear with temperature.
The entropy of a system with constant heat capacity behaves as ln T and therefore
decreases without bounds as T ! 0. Therefore, the heat capacity can not remain
constant at low enough temperatures if S has a lower bound.
If the temperature of a system is decreased along an isobaric line, the entropy
S.T/ can be determined by
Z T
Cp
S.T/  S0 D dT; (6.2)
0 T

where the integral is performed along and isobaric line. Since S and S0 are finite,
the integral must also be finite. The integral will be finite if the heat capacities at
constant pressure Cp vanish when T ! 0. Similarly, we conclude that the heat
capacity at constant volume vanishes when T ! 0.

© Springer-Verlag Berlin Heidelberg 2017 89

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_6
90 6 Nernst-Planck Principle

The Nernst postulate has a second part: the entropy S0 is the same for any state
of a system at temperature zero. This means to say that the isotherm corresponding
to the absolute zero coincides with an isentropic line. It should be noted that only in
this case an isotherm coincides with an isentropic line. For nonzero temperatures, an
isotherm never coincides with an isentropic line. We also note that S0 is the smallest
value of the entropy of a thermodynamic system because the entropy is a monotonic
increasing function of temperature
Being S0 a constant, the derivative .@S=@p/T vanishes at T D 0. Since by
the Maxwell relation .@S=@p/T D .@V=@T/p , then the coefficient of thermal
expansion ˛ D .1=V/.@V=@T/p also vanishes at the absolute zero of temperature.
Similarly, the derivative .@S=@V/T vanishes when T D 0. Using the Maxwell
relation .@S=@V/T D .@p=@T/v , we conclude that the coefficient ˇ D .@p=@T/v
also vanishes at zero absolute temperature. On the other hand, the compressibilities
T D .1=V/.@V=@p/T and S D .1=V/.@V=@p/S do not necessarily vanish. In
general, these two quantities remain nonzero. We remark that they become identical
at absolute zero, since the isotherm T D 0 coincides with the isentropic line S D S0 .
As an experimental illustration of the Nernst postulate we consider the entropy
of sulfur. This substance can be found in two crystalline forms known as rhombic
and monoclinic. Below T0 D 368:6 K, the stable form is the rhombic. Above this
temperature, the stable form is the monoclinic. At the transition temperature T0 ,
rhombic sulfur consumes q D 402 ˙ 2 J/mol, where q is the latent heat of transition
from rhombic to monoclinic. In this conversion, carried out at constant temperature,
the entropy of sulfur increases by the value q=T0 D 1:091 ˙ 0:005 J/mol K.
From the measurement of heat capacity of rhombic sulfur, it is possible to
determine the variation of molar entropy srhomb  s0 between T D 0 and T D T0
from the formula (6.2). Using the same formula, we can also determine the change in
entropy smon  s0 in the same temperature range for the monoclinic sulfur. Although
the monoclinic form is not stable below T0 , the transformation speed to the rhombic
form is very slow, enough for the experimental determination of its heat capacity.
Therefore, the entropy difference smon  srhomb can be determined by
Z T0
cp
smon  srhomb D dT (6.3)
0 T

where cp D cmon p  crhomb

p is the difference between the molar heat capacities of
the two forms of sulfur. We are assuming according to the Nernst postulate that
S0 is the same for both allotropic forms. Figure 6.1 shows the thermal capacities
and the difference between them as a function of temperature. The integral in (6.3),
performed numerically from the experimental data presented in Fig. 6.1, provides
the value smon  srhomb D 0:90 ˙ 0:17 J/mol K, close to the value obtained above.
Another example is the entropy of tin. This substance has two allotropic forms
known as white tin, which is metallic and has a tetragonal structure, and gray tin,
which is semiconductor and has a cubic structure. The gray tin is stable below 13 ı C
and white tin is stable above this temperature. Under 13 ı C, white tin is metastable
6.1 Nernst Postulate 91

(a) 30 (b) 0.005

Δcp /T (J/mol K )
2
cp (J/mol K)

20

10 S monoclinic
S rhombic

0 0
0 100 200 300 400 0 100 200 300 400

T (K) T (K)

Fig. 6.1 Experimental data of the molar heat capacity of sulfur obtained by Eastman and
McGavock [111]. (a) Molar heat capacity of sulfur in the rhombic and monoclinic form.
(b) Difference cp between the molar heat capacities of sulfur in the rhombic and monoclinic
form. The vertical line indicates the transition temperature from the rhombic to the monoclinic
form

and transforms itself into gray tin, a phenomenon known as tin pest. However, the
transformation is slow enough so that they heat capacity measurements can be made.
From the results of the thermal capacity of the two forms, one obtains the value
8.1 J/mol K for the change in entropy. On the other hand, from the latent heat, which
is 2.28 kJ/mol, we obtain the value 8.0 J/mol K very close to the previous value

Thermodynamic Potentials

Decreasing the temperature of a system at constant volume, the internal energy also
diminishes. In fact, along an isochoric line, the energy is a monotonic increasing
function of temperature because .@U=@T/v D Cv  0. In the limit of absolute zero,
the internal energy approaches its lowest value, which we postulate to be finite. That
is, the internal energy has a lower bound which is reached when T ! 0, that is,

U ! U0 when T ! 0: (6.4)

The corresponding state is called ground state.

Taking into account that the Helmholtz free energy is related to the internal
energy through F D U  TS, then in the limit of absolute zero F coincides with
U, that is,

F ! F 0 D U0 when T ! 0; (6.5)

because S0 is finite.
92 6 Nernst-Planck Principle

Using the relation .@H=@T/p D Cp  0, we see that the enthalpy of a system is

monotonic increasing function of temperature, at constant pressure. In the limit of
absolute zero, the enthalpy approaches its lowest value, which we postulate to be
finite. The enthalpy has therefore a lower bound, which is reached when T ! 0,
that is,

H ! H0 when T ! 0: (6.6)

Taking into account that the Gibbs free energy is related to the enthalpy by G D
H  TS, then in the limit of absolute zero G coincides with H, that is,

G ! G0 D H 0 when T ! 0: (6.7)

We should note that, unlike S0 , the thermodynamic potentials at zero temperature

U0 , F0 , H0 and G0 may depend on other thermodynamic variables.

6.2 Heat Capacity of Solids

Dulong-Petit Law

The heat capacity of the solids is usually measured at constant pressure and not
at constant volume, since the low compressibility of the solid makes it difficult to
maintain the volume constant. However, from cp we get cv using the identity

Tv˛ 2
cp D cv C ; (6.8)
T

provided the coefficient of thermal expansion ˛ and the isothermal compressibility

T are known. In any event, the second term is small but not negligible for solids.
Table 6.1 shows the isobaric heat capacity of several solids at room temperature.
According to the Dulong-Petit law, the molar heat capacity of a solid is constant
and has the same value 3R for any solid. We assume that the Dulong-Petit law is
valid for the isochoric molar heat capacity,

cv D 3R: (6.9)

From Table 6.1 and taking into account that cv is close to cp , we see that
the Dulong-Petit Law is reasonably obeyed at room temperature, but with some
significant deviations. However, the greatest deviation from the Dulong-Petit law is
not apparent in Table 6.1, because it occurs at low temperatures. In this regime the
heat capacity becomes small and vanishes when T ! 0.
6.2 Heat Capacity of Solids 93

Table 6.1 Isobaric heat Qcp M D

capacity of solids at
Solid (J/gK) (g/mol) cp =3R (K)
temperature of 25 ı C and
pressure of 1 kPa. The table Lithium 3:582 6:941 0:997 344
shows the specific heat cQp and Graphite 0:709 12:01 0:341 420
the molar heat capacity cp , in Diamond 0:509 12:01 0:245 2230
addition to the molar mass M Aluminium 0:903 26:98 0:976 428
and the Debye temperature
D . Source: CRC, AIP Silicon 0:705 28:09 0:793 640
Sodium 1:228 22:99 1:132 158
Magnesium 1:023 24:31 0:997 400
Potassium 0:757 39:10 1:187 91
Calcium 0:647 40:08 1:040 230
Titanium 0:523 47:87 1:005 420
Chromium 0:449 52:00 0:936 630
Manganese 0:479 54:94 1:055 410
Iron 0:449 55:84 1:006 467
Nickel 0:444 58:69 1:045 450
Cobalt 0:421 58:93 0:995 445
Copper 0:385 63:55 0:980 343
Zinc 0:388 65:39 1:018 327
Germanium 0:320 72:61 0:931 370
Niobium 0:265 92:91 0:986 275
Silver 0:235 107:9 1:016 225
Cadmium 0:232 112:4 1:043 209
Tin 0:228 118:7 1:087 199
Tellurium 0:202 127:6 1:032 153
Cesium 0:242 132:9 1:291 38
Gadolinium 0:236 157:2 1:485 195
Tungsten 0:132 183:4 0:973 400
Platinum 0:133 195:1 1:037 240
Gold 0:129 197:0 1:019 165
Lead 0:129 207:2 1:068 105
Bismuth 0:122 209:0 1:023 119

From the microscopic point of view, the Dulong-Petit law can be understood if
we consider a harmonic solid, that is, a solid whose interatomic potential energy
contains terms up to second order in the deviations of atoms from their equilibrium
positions. It can be shown that such a system is equivalent to a set of independent
one-dimensional oscillators whose frequencies of oscillation are the frequencies of
the normal modes of vibration of the solid. Appealing to the theorem of equipartition
of energy, valid for high temperatures, we assume that the kinetic and potential
energies of each oscillator contributes each with a term equal to kB T=2 to the
94 6 Nernst-Planck Principle

total energy of the solid, where kB is the Boltzmann constant. Since the number
of oscillators is equal to the number of degrees of freedom 3n, where n is the
total number of atoms of the solid, then the contribution of the total energy of
oscillators is 3nkB T D 3NRT. Taking into account the potential energy of atoms
in its equilibrium position, which we denote by U0 , then the molar energy u of the
solid is

u D u0 C 3RT (6.10)

where u0 D U0 =N depends only on the molar volume v. From this result we get
cv D .@u=@T/v D 3R, which is the Dulong-Petit law.

Debye Theory

At low temperatures, the heat capacity of solids deviates appreciably from the
Dulong-Petit law. According to Debye theory, the heat capacity of the solids
behaves, in this regime, according to the law

cv D aT 3 ; (6.11)

and vanishes as T ! 0, in accordance with the Nernst postulate. The same behavior
applies to cp , as we shall see, which can be seen in Fig. 6.2 where the experimental
data of cp =T for KCl are plotted as a function of T 2 .
The Debye theory also provides an interpolation between this law, valid for low
temperatures, and the law of Dulong-Petit. According to Debye, the isochoric molar
heat capacity of a solid harmonic behaves as

D
cv D 3R C. /; (6.12)
T

Fig. 6.2 Molar heat capacity 6

of KCl at low temperatures
obtained experimentally by
cp /T (mJ/mol K )
2

Keesom and Pearlman [115]

4

0
0 5 10 15 20
2 2
T (K )
6.2 Heat Capacity of Solids 95

Fig. 6.3 Molar heat capacity 1.0

of several solids obtained
experimentally, according to 0.8
Debye [110]. The continuous
line corresponds to the Debye
0.6

c p /3R
function given by (6.13)
diamond
0.4
aluminum
copper
silver
0.2 lead

0
0 0.5 1.0 1.5

T/ΘD

where C.x/ is the Debye function defined by (Fig. 6.3)

Z
3 x
4
C.x/ D 3 d (6.13)
2x 0 cosh  1

and has the following properties: (a) C.x/ D 1  x2 =20 for x << 1 and (b) C.x/ D
4 4 =5x3 for x >> 1. At high temperatures, T >> D , we use the first property
to obtain the Dulong-Petit law. At low temperatures, T << D , we use the second
property to obtain the result

12 4 RT 3
cv D ; (6.14)
5D3

which is the law (6.11) with

12 4 R
aD : (6.15)
5D3

The parameter D , called the Debye temperature, is a characteristic of the solid

and is related to the normal frequencies of vibration of the solid. Usually, these
frequencies are affected by variations of the distances Rij between equilibrium
positions of atoms. This can be understood if we examine the coefficients of the
quadratic terms of the expansion of the potential energy in the deviations of atoms
from their equilibrium positions. These coefficients generally depend on Rij . As the
normal frequencies are obtained from the coefficients of second order, then they
depend on the distances Rij . On the other hand, the distances Rij are proportional
to v 1=3 , from which follows that the normal frequencies and therefore D , depend
on the molar volume v. Only for an ideal harmonic solid, for which the second
order coefficients are independent of the volume, the normal frequencies and D
are independent of the volume.
96 6 Nernst-Planck Principle

The Debye temperature is related to the velocity of propagation of elastic waves

vsom by
 1=3
„vsom 6 2
D D NA : (6.16)
kB v

From the experimental measurement of cv at low temperatures, we obtain the

coefficient a and from this coefficient we may obtain D by (6.15). Experimental
values of D for several solids obtained in this manner are shown in Table 6.1.
The Debye temperature obtained by this method can then be compared with that
obtained from the velocity of elastic waves by the use of formula (6.16).
The Debye theory is valid for solids with harmonic interactions. At low tem-
peratures one expects that the harmonic approximation is reasonable, considering
that the atoms perform small oscillations in this regime, so that the law (6.11) is
in fact obeyed at low temperatures. At high temperatures, anharmonic terms of the
interatomic interactions become important and deviations from the Debye theory, in
this regime, and from the Dulong-Petit law, begin to be observed.
It is important to note also that some solids have very high Debye temperatures
such that the regime of the Dulong-Petit law is not reached at room temperature.
This is the case of diamond whose Debye temperature is 2230 K, well above room
temperature. We also note that some solids melt before the regime of the Dulong-
Petit law is reached.

Thermal Expansion

The thermal expansion ˛ of solids is related, as we will see, to the Grüneisen

coefficient G , defined by

d ln D
G D  : (6.17)
d ln v

To determine ˛, we use the relation ˛=T D .@p=@T/v and the Maxwell relation
.@p=@T/v D .@s=@v/T to write
 
@s
˛ D T : (6.18)
@v T

The entropy is obtained by integrating the expression .@s=@T/v D cv =T. Using

(6.12), we get
Z T
D dT 0
s D s0 C 3R C. / ; (6.19)
0 T0 T0
6.2 Heat Capacity of Solids 97

where s0 is the zero temperature entropy. Then we derive this expression with
respect to v, recalling that D depends on v, to obtain the relation
 
@s D 1 dD d ln D
D 3RC. / D cv ; (6.20)
@v T T D dv dv

where we have taken into account, by the Nernst postulate, that s0 is independent
of v. Using the definition of the Grüneisen coefficient, the following expression
follows
cv
˛D T  G : (6.21)
v
It is worth noting that for an ideal harmonic solid, G D 0 because D is
independent of the volume, so that ˛ D 0 and therefore cp D cv .
Assuming that the compressibility T is weakly dependent on the temperature
and remembering that G is independent of T, we may conclude that the ratio ˛=cv
is independent of the temperature from which follows that ˛ has the same behavior
of cv . Therefore, in the regime where the Dulong-Petit law is valid, ˛ is constant. In
this regime, the relation (6.8) implies that the molar isobaric heat capacity cp has a
correction proportional to temperature. In the regime of low temperatures, ˛ must
have the behavior

˛ D bT 3 : (6.22)

This result together with the relation (6.8) implies that the difference between cp and
cv at low temperatures is of the order T 7 , what justifies to say that cp has the same
behavior as cv in this regime.

Metals

Metals are characterized by electrical conduction, resulting from the mobility of

valence electrons of atoms. These electrons contribute to the heat capacity of solid,
adding to the heat capacity of the lattice described by the Debye theory. Assuming
that the free electrons behave like an ideal gas, we would expect an electronic
contribution equal to 3R= 2 for the molar heat capacity at least for high temperatures,
a value that is half the contribution of the crystal lattice. But, the metals at room
temperature show no discrepancy in the heat capacity, when compared with an
insulator, which could be attributed to the electrons.
According to Sommerfeld, the electronic heat capacity behaves linearly with the
temperature in accordance with

2 T
cv D R ; (6.23)
2 F
98 6 Nernst-Planck Principle

where F is the Fermi temperature given by

 2=3
„2 3 2
F D ne (6.24)
2mkB v

where m is the mass of the electron and ne is the number of electrons per mole.
The result (6.23) is valid at low temperatures, more precisely at temperatures
much smaller than the Fermi temperature. However, given that the Fermi tem-
perature of metals is greater than 104 K, we can use the result (6.23) to room
temperature and concluded that the electronic contribution to the heat is negligible
when compared to the contribution of the crystal lattice.
To observe experimentally the electronic contribution we should consider tem-
peratures much lower than the Debye temperature, in which case the two contribu-
tions have the same order of magnitude. Thus, in this regime, the behavior of the
heat capacity of metals is given by

cv D  T C aT 3 ; (6.25)

where the constant  is

 2R
D : (6.26)
2F

It is convenient thus to present the experimental data of the molar heat capacity in a
plot of cv =T versus T 2 , as that shown in Fig. 6.4 for copper at low temperatures.
To determine the coefficient of thermal expansion of metals, we start from the
molar Helmholtz free energy f of free electrons, given, according to Sommerfeld,
by

3  2 RT 2
f D RF  : (6.27)
5 4F

Fig. 6.4 Molar heat capacity 2.0

of copper at low temperatures
obtained experimentally by
cp /T (mJ/mol K )
2

Corak et al. [109]. The 1.5

straight line fitted to the data
points is described by
c=T D 0:665 C 0:0516T 2 1.0

0.5

0
0 5 10 15 20
2 2
T (K )
6.2 Heat Capacity of Solids 99

The electronic pressure p D @f =@v is obtained from f recalling that F depends on
v. Using the result (6.24) and taking into account that ne is constant, then dF =dv D
.2=3/.F =v/ and we get

2RF  2 RT 2
pD C : (6.28)
5v 6vF

The isothermal compressibility T is obtained from this equation by means of

1=T D v.@p=@v/T . The dominant term in (6.28) gives the result

3v
T D : (6.29)
2RF

Using the identity ˛=T D .@p=@T/v and the result (6.29), we obtain from (6.28)
the following result for the coefficient of thermal expansion of metals

 2T
˛D ; (6.30)
2F2

that is, ˛ depends linearly on temperature. Therefore, the electronic coefficient

of thermal expansion has a behavior similar to the electronic heat capacity and
moreover ˛=cv D 1=RF .
The coefficient of thermal expansion of a metal at low temperatures has also a
contribution corresponding to the lattice, which is proportional to T 3 . Therefore, ˛
has a behavior similar to the heat capacity, namely,

˛ D ıT C bT 3 ; (6.31)

where the first and second terms are the electronic and the lattice contribution,
respectively. Figure 6.5 shows this behavior for copper at low temperatures.

Fig. 6.5 Coefficients of 1.5

thermal expansion ˛ of
cooper a low temperatures
obtained experimentally by
10 α/T (K )
-2

White and Collins [123]. The 1.0

straight line fitted to the data
points is described by
1010 ˛=T D 6:0 C 0:87T 2
8

0.5

0
0 50 100 150
2 2
T (K )
100 6 Nernst-Planck Principle

6.3 Planck Postulate

Entropy

According to the Nernst postulate, the molar entropy s of a system approaches

a finite value s0 in the limit T ! 0 and that, moreover, s0 is independent of
thermodynamic variables, that is, a constant. One could imagine this constant as a
characteristic of the substance and, therefore, different for each substance. However,
Planck postulated that s0 has the same value for any substance in thermodynamic
equilibrium then assigning the value zero to the entropy at absolute zero. Therefore,
the Nernst postulate together with the Planck postulate, which we call the Nernst-
Planck principle or the third law of thermodynamics, tells us that

s.T/ ! 0 when T ! 0; (6.32)

for any substance in thermodynamic equilibrium.

To verify the Planck postulate we must define a reference state whose entropy
sR is known. Next, we determine the entropy difference s between the reference
state and the ground state. Since the entropy of the ground state is zero then we
must have sR D s. In general, the reference state is considered to be a gas at high
temperatures and low densities, so that it can be considered an ideal gas. Let us
first examine how to determine s. Subsequently, we examine the entropy of the
reference state.
Consider for simplicity an isobaric process that takes a system composed by
a pure substance from one state at sufficiently low temperature to the reference
state. In this process the system undergoes the transition from solid to liquid, from
liquid to gas, and possible transitions between different crystalline states. In each
transition, occurring at a well defined temperature, the system gains entropy which is
equal to the latent heat divided by the transition temperature. From the experimental
data of the heat capacity cp and latent heats qi of each transition, the entropy
difference s between the reference state and the ground state is determined by
Z T X qi
cp
s D dT C ; (6.33)
0 T i
Ti

where Ti are the transition temperatures.

Table 6.2 shows the values of the several terms of the entropy of xenon, obtained
from the experimental data of the heat capacity of solid and liquid xenon at the
pressure of 1 atm, and from the fusion and boiling latent heats. Under the pressure
of 1 atm xenon undergoes a solid-liquid transition at the temperature of 161.3 K and
a liquid-gas transition at the temperature of 165.13 K. The melting latent heat is
2.293 kJ/mol and the boiling latent heat is 12.62 kJ/mol. Between 0 and 10 K the
entropy is estimated by using the Debye law assuming D D 55 K. According
6.3 Planck Postulate 101

Table 6.2 Entropy of xenon Interval in T s

according to the experimental
State (K) (J/mol K)
data obtained by Clusius and
Riccoboni [108] at the Solid 0 ! 161.3 62:4
pressure of 1 atm Fusion 161:3 14:2
Liquid 161.3 ! 165.13 3:8
Boiling 165:13 76:4
Correction 0:6
Solid ! vapor 0 ! 165.13 157:4
Sackur-Tetrode 165:13 157:3

to Table 6.2, the difference in entropy between xenon in the vapor state, under
the pressure of 1 atm at the boiling temperature, and the solid xenon, at zero
temperature, is s D 157:4 J/mol K.

Ideal Gases

In the regime of low densities, the average distance between gas molecules is large
enough so that the interaction between them can be neglected. Thus, in this regime,
the gas obeys the equation of state of ideal gas pv D RT and the internal energy
u depends only on temperature. As a result, cv D .@u=@T/v depends only on
temperature and the relation between cp and cv becomes simply cp D cv C R since
cp D .@h=@T/p and h D u C pv D u C RT.
If the interaction between molecules is negligible, it suffices to consider the
motion of molecules in isolation. The movement of a molecule is composed of
translation, rotation and vibration and other movements that we will not consider
here. The translation motion is always independent of the other types of motion. The
rotation and vibration of the molecules are in general interdependent movements.
However, the separation into independent motions constitutes a good approxima-
tion. Thus, we assume that the thermodynamic potential of an ideal gas is equal to
the sum of three terms corresponding to the three types of motions. The isochoric
heat capacity is therefore a sum of three terms

cv D ctrans C crot C cvib : (6.34)

Similarly, the entropy will be the sum of three terms

s D strans C srot C svib : (6.35)

We assume that each type of motion is characterized by a certain energy value

below which the movement is frozen and the corresponding degrees of freedom
suppressed. Thus, we introduce the characteristic energy of each type of motion,
102 6 Nernst-Planck Principle

Table 6.3 Characteristic temperatures of rotation and vibration of molecules of several gases,
in K. The indices d and t indicate the double and triple degeneracies, respectively. Source: LB
Gas r;1 r;2 r;3 v:1 v;2 v;3 v;4
H2 85:3 5995
D2 43:0 4300
O2 2:11 2238
F2 1:27 1283
Cl2 0:36 801
N2 2:92 3352
HCl 15:3 4160
CO 2:82 3081
NO 2:45 2701
CO2 0:57 961d 1924 3379
CS2 0:15 571d 943 2190
HCN 2:16 1024d 3002 4765
N2 O 0:62 847d 1851 3198
H2 S 6:84 10:23 15:05 1777 3753 3861
SO2 3:04 0:47 0:41 755 1647 1918
H2 O 39:4 21:0 13:7 2293 5259 5401
NO2 11:6 0:63 0:60 1079 1901 2324
NH3 14:5 14:5 9:34 1365 2344d 4794 4911d
CH4 7:41 7:41 7:41 1877t 2186d 4190 4343t

which we denote by trans , rot and vib . It is convenient to define the characteristic
temperatures t D trans =kB , r D rot =kB and v D vib =kB . Extending the Nernst-
Planck principle to each type of motion, we conclude that the heat capacity and
the entropy associated to a particular type of movement vanish at temperatures well
below the corresponding characteristic temperature.
In general, the characteristic temperatures are such that

t << r << v : (6.36)

Table 6.3 shows values of the characteristic temperatures of rotation and vibration of
some gases. Therefore, as the temperature of the gas is decreasing, the vibrational,
rotational and translational modes will be successively removed. This can be seen
in Fig. 5.3, where we present the molar heat capacity at constant pressure of several
gases. We must remark however that in general the gases condense long before the
regime where T < t could be reached. Equivalently, we can say that the gases are
usually found in the regime T >> t .
To determine the characteristic energy of translation t , we use the relation p D
h=
between the moment of a molecule and its de Broglie wavelength
, where h is
the Planck constant. The corresponding kinetic energy is p2 =2m D h2 =2m
2 . The
lowest energy, identified as trans , is given by the largest value of
, that we consider
6.3 Planck Postulate 103

N 1=3 , where vN is
to be equal to the distance between gas molecules, that is,
D .v/
2 2=3
the volume per molecule. That is, trans D h =2m.v/
N so that

h2
t D : (6.37)
N 2=3 kB
2m.v/

This temperature can be estimated by assuming an intermolecular distance of the

order of 108 m. For helium, the lightest of gases, we obtain the result t D 2:4 mK.
For other gases the characteristic temperature is even smaller.

Sackur-Tetrode Formula

The translational part of the molar heat capacity of an ideal gas is cv D .3=2/R
because each molecule has three translational degrees of freedom. If the ideal gas
is monoatomic this is the only contribution to the heat capacity. Using the relation
.@s=@T/v D cv =T D 3R=2T and taking into account that .@s=@v/T D .@p=@T/v D
R=v, we reach the following expression for the molar entropy of an ideal gas

3
s D Rf ln T C ln vN C ag; (6.38)
2

where vN D v=NA is the volume per molecule, NA is the Avogadro constant and a is
a constant, independent of T and v.
The constant a must depend only on the characteristic of the gas with respect
to the translational kinetic energy. In this sense, the constant a depends solely on
the mass of the molecules, which is the parameter associated to the translational
kinetic energy. Sackur assumed that a depends on the mass m of the gas molecules
in accordance with
3
aD ln m C b; (6.39)
2
where b is a universal constant. This constant, determined by Tetrode by means of
the semi-classical theory is given by

3 2kB 5
bD ln 2 C ; (6.40)
2 h 2
where h is the Planck constant.
Replacing these results into (6.38) gives the Sackur-Tetrode formula

3 2mkB T 5
s D Rfln vN C ln C g; (6.41)
2 h2 2
104 6 Nernst-Planck Principle

which allows us the determination of the translation absolute entropy of the ideal
gases. Sackur-Tetrode formula can also be written as

3 T 3 5
s D Rf ln C ln  C g; (6.42)
2 t 2 2

recalling that t depends on v.

N
The determination of the entropy through the Sackur-Tetrode formula enables us
to verify the Planck postulate for substances that turn into monoatomic ideal gas at
high temperatures. To this end it suffices to compare s, obtained by Sackur-Tetrode
formula with s obtained from (6.33). As an example, we determine the entropy of
vapor xenon in coexistence with its liquid at the pressure of 1 atm. Formula (6.41)
provides the following value for the absolute molar entropy s D 157:3 J/mol K,
which should be compared with that obtained from (6.33), shown in Table 6.2.
We remark that an ideal gas of diatomic or polyatomic molecules can behave as
a monoatomic gas, that is, as a gas having only translational degrees of freedom if it
can be found in the regime where T << r . However, the diatomic and polyatomic
gases in general are found in the opposite regime, where T >> r . An exception
is hydrogen, which remains in the gaseous state even at temperatures below the
characteristic temperature of rotation r D 85:3 K. In this regime, the molar heat
capacity cv of H2 approaches the value .3=2/R, corresponding to a monoatomic gas,
as shown in Fig. 5.3.

Rotational Modes

For temperatures well above r , the molar heat capacity of gases composed by
linear diatomic or polyatomic molecules is given by

crot D R; (6.43)

because such molecules have only two rotational degrees of freedom. The rotational
molar entropy has, therefore, the form

1
srot D Rf ln T C a0 g; (6.44)
2

where a0 is a constant. This constant must depend on the characteristic of the gas
with respect to the rotation motion. In this sense, the constant a0 must depend on
the moment of inertia I of the linear molecule, which is the parameter associated to
rotation.
A semi-classical calculation leads us to the following expression

T
srot D Rfln  ln  C 1g; (6.45)
r
6.3 Planck Postulate 105

where r is the characteristic temperature of rotation and is given by

„2
r D : (6.46)
2IkB

The parameter  is related to the symmetries of molecules. For diatomic molecules

consisting of distinct type of atoms  D 1 and for the same type of atoms  D 2.
For asymmetric polyatomic molecules, like N2 O (NNO),  D 1 and for symmetric
polyatomic molecules, like CO2 (OCO),  D 2.
For noncolinear polyatomic molecules, the number of rotational degrees of
freedom is three, so that the rotational contribution to the molar heat capacity is

3
crot D R: (6.47)
2
The corresponding molar entropy is

3 T 1 3
srot D Rf ln  ln  C ln  C g; (6.48)
2 r 2 2

where

r D .r;1 r;2 r;3 /1=3 : (6.49)

The characteristic temperatures r;i are related with the principal moments of inertia
I1 , I2 and I3 of the molecules by

„2
r;i D : (6.50)
2Ii kB

Characteristic temperatures of rotational modes of several molecules are presented

in Table 6.3. The parameter  is related to the symmetries of the molecule. For H2 O
(isosceles triangle),  D 2; for NH3 (pyramid with triangular base),  D 3; and
for CH4 (tetrahedron),  D 12. We remark that the formulas for entropy and the
heat capacity are valid for temperatures well above the characteristic temperatures
of rotation.

Vibrational Modes

The vibrational part of entropy and of the heat capacity can be determined assuming
that the oscillations are harmonic. For a gas of diatomic molecules, the vibrational
molar entropy is given by

v 1
svib D Rf  =T
 ln.1  ev =T /g; (6.51)
T e v 1
106 6 Nernst-Planck Principle

and the molar heat capacity by

v2 ev =T
cvib D R : (6.52)
T 2 .ev =T  1/2

where the characteristic temperature v is related to the frequency  of the sole

mode of vibration of the diatomic molecule by

h
v D : (6.53)
kB

Result (6.52) was obtained originally by Einstein for the heat capacity of solids
assuming that they oscillate with the same frequency.
Polyatomic molecules in general have various modes of vibration, each one
associated to a natural frequency. The vibrational entropy and heat capacity of
polyatomic gases are determined in a manner similar to that of the diatomic gases.
For each normal mode of oscillation of the molecule, we add to the molar entropy a
term equal to (6.51) and, to the molar heat capacity, a term equal to (6.52). The term
corresponding to the normal mode of frequency i must contain the characteristic
temperature

hi
v;i D : (6.54)
kB

Characteristic temperatures of the vibrational modes of various molecules are shown

in Table 6.3.
In the regime where T >> v , the terms of the entropy and of the heat capacity
corresponding to each normal mode are given by

T
svib D Rfln C 1g (6.55)
v

and

cvib D R; (6.56)

respectively. We remark, however, that this regime is never reached because in

general the polyatomic molecules decompose well before this happens.

Residual Entropy of Ice

At the pressure of 1 atm, water freezes at 0 ı C becoming a crystalline solid, that we

call ice. A way of determining the absolute entropy s0 of ice at zero temperature
6.3 Planck Postulate 107

Table 6.4 Entropy of water Interval in T s

from experimental data for
State (K) (J/mol K)
the heat capacity presented in
Fig. 6.6 and from the latent Ice 0 ! 273.15 38:05
heats of fusion and boiling Ice melting 273.15 22:00
Water 273.15 ! 373.15 23:49
Boiling water 373.15 108:96
Ice ! vapor 0 ! 373.15 192:50

Fig. 6.6 Molar isobaric heat 10

capacity cp of water in the
three states at the pressure of 8
1 atm. The dashed lines
indicate the ice-water
transition at 273.15 K and the
cp /R 6
water-vapor transition at
373.15 K. The results for ice 4
were obtained experimentally
by Giauque and Stout [114]
2
ice water vapor
0
0 200 400 600
T (K)

consists in calculating the difference in entropy s between the state at zero

temperature and a reference state, which we choose as the vapor in coexistence
with water at the pressure of 1 atm and temperature of 373.15 K. Table 6.4 shows
the variations of entropy of ice and water, obtained from the experimental data for
the heat capacity, presented in Fig. 6.6.
The integral of cp =T for ice gives 38.05 J/mol K and for the liquid, 23.49 J/mol K.
The melting latent heat of ice is 6.006 kJ/mol, so that the entropy at the ice-
water transition increases by 22.00 J/mol K. The boiling latent heat of water
is 40.660 kJ/mol, so that the entropy at the water-vapor transition increases by
108.96 J/mol K. Therefore, the difference between the molar ice entropy at 0 K and
the vapor at 1 atm is s D 192:50 J/mol K, as long as no other transition takes
places in addition to the ones mentioned.
The molar entropy sR of the reference state, which we assumed to be that of
an ideal gas, is determined from the formulas for the translational, rotational and
vibrational. The translational entropy gives the result strans D 149:33. To determine
the rotational entropy we use the rotational characteristic temperatures of H2 O
presented in Table 6.3. We obtain the following result srot D 46:50 J/mol K. The
vibrational contribution to the entropy is small. From the data of Table 6.3 we
obtain the value svib D 0:12 J/mol K. Therefore, the entropy of water vapor is
sR D 195:95 J/mol K.
108 6 Nernst-Planck Principle

From the values obtained above, we see that ice has a residual entropy s0 D
sR  s D 3:45 J/mol K, in apparent contradiction with the Nernst-Planck principle.
The contradiction disappears if we assume, for example, that at low temperatures ice
finds itself in a metastable state. In fact, in this regime one observes experimentally
another more stable solid phase, called ice XI, whose crystalline structure is distinct
from ordinary ice. The transition between the ordinary ice, called ice I, and ice XI
occurs at 72 K, with a decrease of entropy.
Pauling considers that the hydrogen bonds existing in the ice structure have a
fundamental role in the explanation of the residual entropy. A hydrogen bond binds
atoms of oxygen belonging to different molecules and the hydrogen, actually a
hydrogen ion, may be found next to one or the other oxygen atom, that is, in one of
two possible positions. Since the number of hydrogen bonds is equal to twice the
number n of molecules then the number of microscopic state W D wn would be such
that w D 4. However, this value is overestimated because each oxygen ion must have
only two hydrogen ions in its neighborhood. Taking into account this restriction, we
see that, of the 16 possible arrangements of hydrogen ions around an oxygen ion,
only 6 can be accomplished, what leads to the Pauling result w D 4.6=16/ D 3=2.
Using the statistical interpretation of entropy, S D kB ln W, this approximate
calculation gives the following residual molar entropy s0 =R D ln.3=2/, that is,
s0 D 3:371 J/mol K. A more correct counting of the microscopic arrangements leads
us to the result obtained by Nagle, w D 1:5069, that is, s0 D 3:409 J/mol K, in
accordance with the observed value, s0 D 3:45 ˙ 0:20 J/mol K, taking into account
the experimental errors.
We remark finally that ice XI, in contrast to the ordinary ice, has a crystalline
structure such that the hydrogen ions are ordered resulting in zero entropy at T D 0.

Problems

6.1 At low temperatures, the molar Helmholtz free energy f .T; v/ of a solid is given
by

f D aT 4 v C bv.ln v C c/:

Determine s, p, u, cv , cp , T , s and ˛. Show that these quantities behave according

to the Nernst-Planck principle.
6.2 At low temperatures, the molar heat capacity at constant volume of a solid
behaves as cv D avT 3 . Determine the molar entropy and the quantity  D
.@p=@T/v . Show that the ratio cv = is independent on temperature. Solve the same
problem for the case cv D av 2=3 T.
6.3 Planck Postulate 109

Table 6.5 Entropy of Interval in T s

nitrogen according to the
State (K) (J/mol K)
experimental data obtained by
Giauque and Clayton [113] at Solid II 0 ! 35.61 27:1
the pressure of 1 atm Transio II!I 35:61 6:4
Solid I 35.61 ! 63.14 23:4
Fuso 63:14 11:4
Liquid 63.14 ! 77.32 11:4
Boiling 77:32 72:1
Solid II ! vapor 0 ! 77.32 151:8

6.3 According to the Debye theory, the molar free energy f of a solid is given by

D
f D u0 C 3RT ln.1  eD =T /  RTD. /;
T

where D.x/ is the function defined by

Z
3 x
3
D.x/ D d :
x3 0 e  1

Show that the molar energy u is given by u D u0 C 3RTD.D =T/. Obtain, from this
result, the expression (6.12) for the molar heat capacity cv . Determine the behavior
of f , u and cv at low and high temperatures.
6.4 Demonstrate the Slater relation

1 v @BS 1 1 @ ln BS
G D   D 
6 2BS @v 6 2 @ ln v

from the relation between vsom and the bulk modulus BS .

6.5 Determine the absolute molar entropy s of the nitrogen vapor at the pressure
of 1 atm and boiling temperature Te . Compare it with the variation of the molar
entropy s between the solid nitrogen at T D 0 K and the nitrogen vapor at T D Te ,
obtained by calorimetric measurements shown in Table 6.5. Notice that the solid
nitrogen undergoes a structural transition in addition to the two solid-liquid and
liquid-vapor transitions.
Chapter 7
Phase Transition

7.1 Pure Substance

State of Matter

A substance may occur in different states of aggregation. At high temperatures and

low pressures it occurs in gaseous form. At low temperatures and high pressures, it
occurs in solid form. At intermediate temperatures and pressures, it may be a liquid.
A gas is characterized by its high compressibility and low density. A solid and a
liquid, in contrast, are less compressible, and have moderate or high densities. A
solid, contrary to a liquid, offers resistance to deformation.
At the microscopic level the states of matter are characterized by the form of
aggregation of the molecules that make up the substance. Either as a liquid or a gas
the molecules are very mobile and therefore are spatially disordered. However, the
mobility of the molecules and the average distance between molecules in a gas are
much larger. In a solid, on the other hand, the mobility is greatly reduced and the
movement of the molecules is only local, enabling the regular spatial structures of
molecules, which characterize the crystalline solids. A solid material may have one
or more crystalline structures.
Besides these three states of aggregation of matter, other states may arise in
substances composed of more complex molecules. For example, liquid crystals
consist of long molecules favoring the emergence of a state, called nematic, which
has orientational order but is devoid of spatial order. We must also mention the glass,
traditionally made by the rapid cooling of fused silica. Although a solid material,
glass and other amorphous materials do not have the regular structural found in
crystalline solids. But, the glassy state is not a state of thermodynamic equilibrium
and therefore will not be the subject of our study.

© Springer-Verlag Berlin Heidelberg 2017 111

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_7
112 7 Phase Transition

Thermodynamic Phases

A homogeneous system, that is, completely uniform with regard to specific prop-
erties, constitutes a thermodynamic phase. From the microscopic point of view, the
molecules that comprise a thermodynamic phase are in the same type of aggregation.
The gaseous state, the liquid state and the various crystalline states are possible
thermodynamic phases of a system. A system may also be heterogeneous, that
is, display simultaneously two or more thermodynamic phases. In this case, if the
system is in equilibrium, we are facing a coexistence of phases.
The phenomenon of phase coexistence is particularly useful in determining the
purity of a substance. Suppose that a system consisting of two phases in coexistence
receive heat at constant pressure. In this process, one phase grows at the expenses
of the other which decreases. If the system consists of a pure substance, the
temperature remains unchanged as long as there are phases in coexistence. Thus,
for a given pressure, the transition between two phases of a pure substance occurs
at a well defined temperature.
From the microscopic point of view, a pure substance consists of a single type
of atom (simple substance) or a single type of molecule (composite substance). As
an example, distilled water, formed by molecules of H2 O. Many minerals consist of
pure or nearly pure substances such as diamond and graphite, which are different
crystalline forms of carbon (C), and quartz, consisting almost exclusively of silica
(SiO2 ).

Phase Diagram

The various thermodynamic phases of a pure substance are represented in a

pressure-temperature diagram, called phase diagram. The different phases are rep-
resented by regions. The coexistence of two thermodynamic phases is represented
by a line in the diagram since, for a given pressure, the coexistence of two phases
must occur at a well defined temperature and not in a range of temperatures. Two
lines of coexistence may meet. The meeting point of coexistence lines is a triple
point, that is, the point of coexistence of three phases. The liquid-vapor coexisting
line always ends in a point, called critical point, where the liquid and vapor become
indistinct.
Figure 7.1 shows the phase diagram of pure water (H2 O). At a pressure of 1 atm
(101,325 Pa), ice and water coexist at a temperature of 0 ı C (273:15 K). At the same
pressure, water and vapor coexist at a temperature of 100 ı C (373:15 K). Heating
water, at the pressure of 1 atm, from the room temperature, its temperature will
increase to the point where it starts to boil when reaching 100 ı C. While boiling, its
temperature remains unchanged. The temperature will rise again only when all water
has been evaporated. Similarly, if we remove heat of a certain amount of water at a
pressure of 1 atm, starting from the room temperature, its temperature will decrease
7.1 Pure Substance 113

until reaching 0 ı C when it starts to freeze. While water is being transformed into
ice, the temperature remains unchanged and will decrease only when all water has
been transformed into ice.
The line of coexistence of water and its vapor has a positive slope which means
that at pressures lower than 1 atm boiling occurs at lower temperatures. In places
of high altitudes, where the atmospheric air pressure is lower, the water boils at
lower temperatures. At pressures higher than 1 atm, water boils at temperatures
higher than 100 ı C, which is the working principle of pressure cookers. At still
higher pressures, the line of coexistence of water-vapor eventually ends at a
critical point, where water and steam become indistinct, which happens at the
pressure of 22:06 MPa (217.7 atm) and at temperature of 374:00 ı C (647:15 K),
as shown if Fig. 7.1. The line of coexistence of ice-water has a negative slope, in
contrast to most substances, whose lines of coexistence solid-liquid have positive
slope. Decreasing the pressure, the melting temperature slightly increases. The two
lines of transition, ice-water and water-vapor, meet at the triple point that occurs
at the pressure of 611:7 Pa (0.006037 atm) and at the temperature of 0:01 ı C
(273:16 K).
Figure 7.2 shows the phase diagram of carbon dioxide (CO2 ). The triple point
occurs at a temperature of 56:57 ı C (216:58 K) and at the pressure of 0:518 MPa
(5:11 atm). As the pressure of the triple point is greater than 1 atm, the carbon
dioxide does not exist in the liquid state at a pressure of 1 atm. At this pressure,
the solid carbon dioxide passes directly to the gaseous state at a temperature of
78:45 ı C (194:7 K), a phenomenon known as sublimation. However, it is possible
to obtain liquid carbon dioxide, for example, at the temperature of 20 ı C, increasing
the pressure to 5:73 MPa (56:5 atm). The critical point of CO2 occurs at the
temperature of 30:99 ı C (304:14 K) and at the pressure of 7:375 MPa (72:79
atm).

Fig. 7.1 Phase diagram of 10

2

water. The dotted horizontal

line indicates the pressure of 10
1

1 atm and the vertical dotted

0
line, the temperature of 10
ice water vapor
p (MPa)

20 ı C. The triple point is

-1
represented by a full circle 10
and the critical point by an
-2
empty circle. Source: LB 10

-3
10

-4
10
-100 0 100 200 300 400
o
q ( C)
114 7 Phase Transition

Fig. 7.2 Phase diagram of 10.0

carbon monoxide (CO2 )
showing the gas (G), the L
liquid (L) and solid (S)
phases. The horizontal dotted 1.00

p (MPa)
line indicates the pressure of
1 atm and the vertical dotted G
line, the temperature of
S
20 ı C. The triple point is 0.10
represented by a full circle
and the critical point by an
empty circle. Source: LB
0.01
-120 -80 -40 0 40
o
θ ( C)

7.2 Discontinuous Transition

Densities and Fields

For a pure substance, the fundamental relation in the energy representation is

constituted by the internal energy U.S; V; N/ as function of entropy S, volume
V and number of moles N of the substance. As the internal energy is extensive,
it is possible to describe the thermodynamic properties of a pure substance by
the molar internal energy of u.s; v/ as a function of the molar entropy s and the
molar volume v. The fundamental relation is recovered by means of U.S; V; N/ D
Nu.S=N; V=N/. Performing the Legendre transformations starting from u.s; v/, we
obtain the molar potentials f .T; v/, h.s; p/ and g.T; p/ which describe equivalently
the thermodynamic properties of a pure substance.
The variables such as temperature and pressure are known as thermodynamic
fields while others, such as the molar entropy and molar volume, are known as
thermodynamic density. In a state of thermodynamic equilibrium all parts of a
system will have the same value for each one of the fields. If the system is
homogeneous (one thermodynamic phase), not only the fields but densities have the
same values. In a heterogeneous system (coexistence of one or more thermodynamic
phase) on the other hand, while the fields have the same values for each phase, the
density will have different values.
We state below the fundamental properties of the molar thermodynamic poten-
tials u.s; v/, f .T; v/, h.s; p/ and g.T; p/, with respect to their thermodynamic space.
Let  one of these potentials and ˝ the respective thermodynamic space. Then:
(a)  is a continuous function of all variables of ˝;
(b)  is differentiable with respect to the densities of ˝; or, equivalently, a
thermodynamic field obtained by differentiation of  is a continuous function
of all variable of ˝;
7.2 Discontinuous Transition 115

(c)  is a convex function of the densities of ˝ and a concave function of the fields
of ˝;
(d)  is analytic function inside a one-phase region of ˝.
The following formulas allows us to obtain the equation of state from each one
of the molar thermodynamic potential:

du D Tds  pdv; (7.1)

df D sdT  pdv; (7.2)
dh D Tds C vdp; (7.3)
dg D sdT C vdp; (7.4)

For a pure substance, the molar Gibbs free energy identifies itself with the
chemical potential, that is, g D , so that

d D sdT C vdp: (7.5)

Equivalently

dp D NsdT C d ;
N (7.6)

where N D 1=v is the number of moles per unit volume, and sN D s=v is the entropy
per unit volume.

Volume Change

Suppose that a pure substance in the liquid state is in coexistence with its vapor
whose pressure is kept constant while receiving a certain amount of heat. The heat
introduced transforms part of the liquid into vapor thus increasing the total volume
of the liquid-vapor system. In this process the temperature remains unchanged as
long as the phases are in coexistence. Moreover, considering that each phase is
homogeneous, the densities of the liquid and of the vapor are well defined and
remain invariant. That is, the molar volumes of the liquid vL and the molar volume
of vapor vV remain invariant. The volume change is caused by the transformation
of the liquid, which has smaller molar volume, into vapor, which has greater molar
volume.
If a certain number of mole N of the liquid is transformed into vapor, the variation
in volume will be V D NvG NvL , where vG D 1=NG and vL D 1=NL are the molar
volumes of the vapor and liquid, respectively. Therefore, the variation v D V=N,
corresponding to the transformation of one mole of the substance in the liquid state
into vapor, is equal to the difference between the molar volumes of vapor and liquid,
116 7 Phase Transition

Table 7.1 Liquid-vapor transition at the pressure of 1 atm of several pure substances. The table
presents the boiling temperature Te , the specific boiling latent heat `e , the specific volume of the
liquid vQL and of the vapor vQV . Source: ICT, LB, EG, CRC
Te `e vQL vQV
Substances K J/g cm3 /g dm3 /g
Helium He 4:222 20:7 8:006 0:059
Neon Ne 27:07 84:8 0:831 0:1051
Argon Ar 87:293 161:0 0:716 0:173
Krypton Kr 119:92 108:4 0:413 0:112
Xenon Xe 165:10 96:1 0:327 0:102
Hydrogen H2 20:28 445 14:12 0:7468
Oxygen O2 90:188 213:1 0:871 0:2239
Nitrogen N2 77:35 198:8 1:239 0:2164
Fluorine F2 85:03 174 0:666 0:178
Chlorine Cl2 239:11 288 0:640 0:270
Bromine Br2 331:9 187:5 0:335 0:17
Carb. monox. CO 81:6 215:6 1:267 0:230
Ammonia NH3 239:82 1370 1:466 1:16
Hydr. chlor. HCl 188:05 443:1 0:840 0:48
Water H2 O 373:15 2257 1:04346 1:6731
Methane CH4 111:67 510:8 2:367 0:551
Ethylene C2 H4 169:38 482:3 1:761 0:480
Ethane C2 H6 184:5 488:5 1:830 0:487
Propane C3 H8 231:0 431 1:718 0:413
Butane C4 H10 272:6 386:1 1:663 0:370
Benzene C6 H6 353:24 393:3 1:23 0:36
Methanol CH4 O 337:7 1099 1:332 0:82
Ethanol C2 H6 O 351:44 837 1:358 0:606
Acetone C3 H6 O 329:20 501 1:333 0:5
Diethyl ether C4 H10 O 307:6 358:8 1:436 0:316

that is, v D vG  vL . Table 7.1 presents the specific volumes of liquid and vapor
at the transition, at the pressure of 1 atm, of various pure substances.
Suppose that a certain quantity of gas is compressed isothermally. Figure 7.3c
shows the representation of this process. The volume of the gas decreases and the
pressure increases to the point that the vapor begins to condense. From this point,
the decrease in volume does not change the pressure, known as the vapor pressure,
until the condensation is complete when the pressure then increases again.
The transition from one phase to another along an isotherm is represented by a
horizontal straight line in the p-v diagram because the pressure remains unchanged.
The end points of the segment have coordinates .vL ; p / and .vG ; p /, where p is
the vapor pressure. These two end points represent the thresholds of the liquid-vapor
coexistence and the line segment that joins them is called the tie line.
7.2 Discontinuous Transition 117

(a) (b)
g
f

vA vB v p* p
(c) (d)
v
p

vB
p*

vA

vA vB v p* p

Fig. 7.3 The figures show an isotherm corresponding to a discontinuous phase transition. (a)
Molar Helmholtz free energy f versus molar volume v. (b) Molar Gibbs free energy g versus
pressure p. (c) Pressure p D @f =@v versus v. (d) Molar volume v D @g=@p versus p. Notice that
the plots (a) and (b) form a pair of functions related by a Legendre transformation whereas (c) and
(d) form a pair of inverse functions

Given any point .v; p / on the tie line we can determine the fraction in number
of moles of each phase. The volume occupied by the liquid is VL D vL NL and
by the vapor is VG D vG NG , where NL and NG are the number of moles of the
liquid and vapor, respectively. The fractions in number of mole are rL D NL =N
and rG D NG =N, where N D NL C NG is the total number of moles. Therefore,
V D VL C VG D vL rL N C vG rG N, so that v D V=N is given by

v D vL rL C vG rG ; (7.7)

which is the lever rule. Since rL C rG D 1, we can then obtain from it the fractions
in number of moles
vG  v v  vL
rL D and rG D : (7.8)
vG  vL vG  vL

Therefore, the fraction in number of moles of one phase is proportional do the

distance between the point considered and the end point of the tie line corresponding
to the other phase.
118 7 Phase Transition

Latent Heat

Let us suppose again that a pure substance in the liquid state is in coexistence with
its vapor, whose pressure is kept constant while receiving a certain amount of heat.
The heat introduced transforms part of the liquid into vapor without changing the
temperature. The number of moles of the substance in the liquid state that turns
into vapor is proportional to the heat introduced Le , called boiling latent heat. The
amount of heat `e D Le =N required to evaporate one mole of the liquid is called
molar boiling latent heat. Table 7.1 shows the specific latent heat of several pure
substances corresponding to the liquid-vapor transition at pressure of 1 atm.
Taking into account that each phase is homogeneous, each have a well-defined
molar entropy, which remains invariant if the pressure is kept constant and as long
as there is coexistence. Denoting by sL the molar entropy of the liquid and by sG
that of the vapor, then the increase in entropy corresponding to the transformation
of N moles of liquid into vapor will be S D NsG  NsL . Because this process is
isothermal, then the change in entropy is equal to the heat introduced divided by
the temperature so that Le D T  S, where T  is the temperature of coexistence.
Therefore, the molar latent heat is related to the change in molar entropy s D
sG  sL by `e D T  s.
In the T-s diagram, the phase transition along an isobaric line is also represented
by a tie line between the points .sL ; T  / and .sG ; T  /, which indicate the threshold
of liquid-vapor coexistence, as shown in Fig. 7.4. For a point .s; T  /, the lever rule
also holds

s D sL rL C sG rG : (7.9)

The fractions in number of moles of each phase are given by

sG  s s  sL
rL D e rG D : (7.10)
sG  sL sG  sL

Phase Transition

The phase transition, in the diagram of Helmholtz free energy versus volume, is
represented by an inclined line segment whose slope is equal to p because p D
@f =@v, as shown in Fig. 7.3a. In the range vL  v  vG the free energy f D
.NL fL C NG fG /=N is the weighted average of fL and fG , that is,
vG  v v  vL
f .T; v/ D f .T; vL / C f .T; vG /: (7.11)
vG  vL vG  vL
7.2 Discontinuous Transition 119

(a) (b)
g
h

sA sB s T* T
(c) (d)
T s
s
B

T*
sA

sA sB s T* T

Fig. 7.4 The figures show an isobaric line corresponding to a discontinuous phase transition. (a)
Molar enthalpy h versus molar entropy s. (b) Molar Gibbs free energy g versus temperature T. (c)
Temperature T D @h=@s versus s. (d) Molar entropy s D @g=@T versus T. Notice that the plots
(a) and (b) form a pair of functions related by a Legendre transformation whereas (c) and (d) form
a pair of inverse functions

In this same range, the pressure is

1 1
p .T/ D f .T; vL /  f .T; vG /: (7.12)
vG  vL vG  vL

Inverting the plot of p versus v, we get the plot of v versus p, shown in Fig. 7.3d.
In this case the transition is represented by a jump in the molar volume, equal to
v D vG  vL , at the vapor pressure p D p . Taking into account that v D @g=@p,
the phase transition is represented, in the diagram of Gibbs free energy versus
pressure, by a point whose derivatives at left and at right are distinct, as shown
in Fig. 7.3b.
Similarly, the phase transition, in the diagram of enthalpy versus entropy, is
represented by an inclined line segment because T D @h=@s. In the range sL 
s  sG the enthalpy h D .NL hL C NG hG /=N is the weighted average of hL and hG ,
that is,
sG  s s  sL
h.s; p/ D h.sL ; p/ C h.sG ; p/: (7.13)
sG  sL sG  sL
120 7 Phase Transition

In the same range the temperature is

1 1
T  .p/ D h.sG ; p/  h.sL ; p/: (7.14)
sG  sL sG  sL

The slope is equal to the temperature of coexistence T  at the pressure p , as

shown in Fig. 7.4a. Inverting the plot T versus s, shown in Fig. 7.4c, we get the plot
of s versus T, shown if Fig. 7.4d. The phase transition is represented by a jump in
the molar entropy equal to s D sG  sL D `e =T  that occurs at the temperature
T D T  . Taking into account that s D @g=@T, the phase transition is represented,
in the Gibbs free energy versus temperature, by a point whose derivatives at left and
right are different, as seen in Fig. 7.4b.
The graphs presented in Figs. 7.3 and 7.4 correspond to the isothermal and
isobaric processes, respectively. In the T-p diagram, these processes correspond
to straight lines parallel to the axis T and p, that intersect at the point .T  ; p /.
At this point, where the phase transition occurs, the molar volume and the molar
entropy present a jump. If we imagine other processes corresponding to trajectories
in the phase diagram that intersect the liquid-vapor transition line at the same point
.T  ; p /, the jump in v will be the same because vG and vL depend only on the
point on the coexistence line. Similarly, the jump in s as well as the molar boiling
latent heat `e will be the same because sG , sL and `e depend only on the point on the
coexistence line.
The phase transition just described, where the molar volume and molar entropy
present a jump when one crosses the coexistence line in the T-p diagram is called
discontinuous or first-order phase transition.

Clausius-Clapeyron Equation

Consider two nearby points A0 and B0 in the region corresponding to the liquid phase
in the T-p diagram. The difference in the molar Gibbs free energy g0 between these
two points is given by

g0 D sL T C vL p; (7.15)

since they are close together, where sL and vL are the molar entropy and the molar
volume of the liquid phase and T and p are the difference in temperature and
pressure, respectively, between the two points. Similarly, consider another pair of
nearby points A00 and B00 in the region corresponding to the gas phase such that the
differences in temperature and pressure of these two points are also T and p.
The difference in the molar Gibbs free energy g00 between these two points are

g00 D sG T C vG p; (7.16)

7.2 Discontinuous Transition 121

since the two points are close together, where sG and vG are the molar entropy and
molar volume of the gas phase.
Imagine now two points A and B on the coexistence line such that the difference
in temperature and pressure between these points are also T and p. Suppose next
that the points A0 and A00 approach the point A. As a consequence, B0 and B00 will
approach the point B. Since the Gibbs free energy is a continuous function of T and
p, then g00 D g0 so that

 sG T C vG p D sL T C vL p; (7.17)

or
p sG  sL
D : (7.18)
T vG  vL

Taking the limit in which the points A and B approach each other, then

dp sG  sL
D ; (7.19)
dT vG  vL

which is the Clausius-Clapeyron equation, which can be written in the form

dp `e
D ; (7.20)
dT T.vG  vL /

where `e D T.sG  sL / is the molar boiling latent heat and T is the transition
temperature.
The Clausius-Clapeyron equation relates the slope of the coexistence in the p-
T diagram with the discontinuities of entropy and volume when passing from one
phase to another. For the liquid-vapor transition, sG > sL because the gas phase
occur always at a temperature higher than that of the liquid phase and vG > vL
because the gas always has a molar volume greater than that of the liquid. Therefore
dp=dT > 0 from which we conclude that the liquid-vapor coexistence curve has a
positive slope in the p-T diagram.
The Clausius-Clapeyron equation may also be used to demonstrate that the tie
lines in the volume-entropy diagram are perpendicular to the coexistence line as
illustrated in Fig. 7.5b. Consider a portion of the coexistence line in the pressure-
temperature diagram as shown in Fig. 7.5a. Consider also the volume-entropy
diagram constructed so that the axis s is parallel to the axis T and the axis v is
antiparallel to the axis p. In this diagram we locate the point L=.sL ; vL / and the
point G=.sG ; vG /, corresponding to the liquid and vapor coexisting at the point
O of the p-T diagram. Next we perform a translation of the coexistence line from
the pressure-temperature diagram into the volume-entropy diagram. The slope of
the segment LG is
D .vG  vL /=.sL  sG / and the slope of the coexistence line
is given for dp=dT. But according to the Clausius-Clapeyron equation dp=dT D
122 7 Phase Transition

(a) (b)
p −v

C C
L

O O

T s

Fig. 7.5 (a) Portion of the liquid-vapor coexistence ending on the critical point C, in the pressure-
temperature diagram. (b) Tie lines in the volume entropy diagram. The tie lines are perpendicular to
the line of coexistence obtained by a translation of the (a) diagram. The tie line LG is perpendicular
to the coexistence line at the point O. Notice the change in sign of v in the vertical axis

1=
, which is the condition of perpendicularity. Therefore, the tie line LG is
perpendicular to the coexistence line at the point O.
The Clausius-Clapeyron equation is valid not only for the transition between the
liquid and vapor but for the transition between any two phases. For the solid-liquid
transition,

dp sL  sS `f
D D ; (7.21)
dT vL  vS T.vL  vS /

where `f D T.sL  sS / is the molar melting latent heat and T is the transition
temperature. Table 7.2 presents the specific melting latent heat and the densities of
the solid and liquid, corresponding to the solid-liquid transition, at the pressure of 1
atm, of various pure substances. In general, the liquid phase occurs at a temperature
higher than that of the solid phase, which implies that the molar entropy of the liquid
is greater than the molar entropy of the solid, sL > sS .
For most pure substances, the density of the liquid is smaller than that of the
solid, which means to say that the liquid molar volume is greater than the solid
molar volume, vL > vS . In these case, therefore, dp=dT > 0 and the solid-liquid
coexistence line has a positive slope in the p-T diagram, as can be seen in Fig. 7.2.
An important exception to this general case is the water. As is well known, at the
pressure of 1 atm, the ice floats over the water, which means that the density of ice is
smaller than the density of water, that is, S < L or in an equivalent way vS > vL .
Therefor dp=dT < 0 and the water-ice coexistence line has a negative slope in the
p-T diagram, as seen in Fig. 7.1. However, there are other crystalline forms of ice
that are denser that water, which turns the slope of the coexistence line positive. The
substances whose solid phase is less dense than the liquid phase include, besides
water, bismuth, silicon, germanium and gallium.
7.2 Discontinuous Transition 123

Table 7.2 Solid-liquid transition at the pressure of 1 atm. The table presents the melting
temperature f and Tf , the specific melting latent heat `f and the densities of the solid S and
of the liquid L . Sources: LB, CRC
f Tf `f S L
ı
Substances C K J/g g/cm3 g/cm3
S 115:21 388:36 53:64 1:919 1:819
P 44:15 317:30 21:3 1:80 1:74
Sb 630:63 903:78 163:2 6:59 6:53
Bi 271:40 544:55 54:07 9:71 10:05
Si 1414 1687 1788 2:31 2:57
Ge 938:25 1211:40 508:7 5:31 5:60
Li 180:5 453:6 432 0:521 0:512
Na 97:80 370:95 113 0:951 0:927
K 63:38 336:53 59:6 0:847 0:828
Mg 650 923 349 1:649 1:584
Fe 1538 1811 247:3 7:23 6:98
Co 1495 1768 275 8:19 7:75
Ni 1455 1728 297:8 8:31 7:81
Pt 1768:4 2041:5 113:6 21:06 19:77
Cu 1084:62 1357:77 208:7 8:36 8:02
Ag 961:78 1234:93 104:8 9:75 9:320
Au 1064:18 1337:33 63:72 18:23 17:31
Zn 419:53 692:68 112 7:06 6:57
Hg 38:83 234:32 11:4 14:191 13:69
Al 660:32 933:47 397 2:534 2:375
Ga 29:76 302:91 80:2 5:89 6:08
Sn 231:93 505:08 59:2 7:19 6:99
Pb 327:46 600:61 23:0 11:05 10:66
H2 O 0:00 273:15 333:4 0:9173 0:99984
NaCl 800:7 1073:8 482:1 1:91 1:556
KCl 771 1044 356:0 1:83 1:527
KBr 734 1007 214 2:49 2:127
AgCl 455 728 92 5:26 4:83
AgBr 432 705 48:6 6:03 5:577
NaNO3 307 580 170 2:11 1:90

The Clausius-Clapeyron equation can also describe the solid-vapor coexistence.

For this transition,

dp sG  sS `s
D D ; (7.22)
dT vG  vS T.vG  vS /

where `s D T.sG  sS / is the molar sublimation latent heat and T is the transition
temperature. In sublimation, sG > sS because the vapor always occurs at a
124 7 Phase Transition

Table 7.3 Solid-vapor transition at the pressure of 1 atm. The table presents the sublimation
temperature f and Tf , the specific sublimation latent heat `s and the specific volumes of the solid
vQS and of the vapor vQV . Sources: LB, EG
s Ts `s vQS vQv
ı
Substance C K J/g cm3 /g dm3 /g
Carb. diox. CO2 78:50 194:65 570:7 0:640 0:355
Acetylene C2 H2 83:80 189:35 801:4 1:37 0:578
Arsenic As 615 883 425
Graphite C 3600 3873

temperature higher than that of the solid. Moreover, since vG > vS , then dp=dT > 0
from which we conclude that the sublimation line has a positive slope. At the
pressure of 1 atm, some substance pass directly from solid to vapor. Table 7.3
presents some example, that include the carbon dioxide and graphite.

Triple Point

The triple point corresponds to the coexistence of the three phases solid, liquid
and vapor. Such a point, for a pure substance, occurs at a unique temperature and
pressure since the triple point is the meeting of three lines of coexistence, liquid-
vapor, solid-liquid and solid-vapor. Table 7.4 shows the temperature and pressure
corresponding to the triple point of several pure substances. At the triple point the
latent heat are related by

`s D `f C `e (7.23)

because .sG  sS / D .sG  sL / C .sL  sS /.

At the triple point, the three lines of coexistence meet forming angles always
smaller than 180ı. In an equivalent manner, we may say that the extension of the line
beyond the triple point will be found inside the region between the other two lines
of coexistence. To show this rule we proceed as follows. In a system of coordinates
consisting by the temperature and pressure, we locate the triple point .Tt ; pt / and the
three lines of coexistence, as shown in Fig. 7.6a. In another system of coordinates,
set up in such a way that the molar entropy axis is parallel to the temperature and the
molar volume axis is antiparallel to the pressure, we locate the points S D .vS ; sS /,
L D .vL ; sL / and G D .vG ; sG /, determined by the molar volumes and molar
entropies of the three phases in coexistence. The segments SL, LG and SG constitute
tie lines related to the three transitions. Next we perform a translation of the three
lines of coexistence and of the triple point to the volume-entropy diagram, so that
the triple point is place inside the triangle SLG, as seen in Fig. 7.6b.
7.2 Discontinuous Transition 125

Table 7.4 Triple point of some pure substances. The table presents the temperature of the triple
point Tt and the pressure of the triple point pt . Sources: LB, EG, CRC
Tt pt
Substance K kPa
Neon Ne 24:55 43:3
Argon Ar 83:78 68:7
Krypton Kr 115:95 73:1
Xenon Xe 161:35 81:6
Hydrogen H2 13:947 7:2
Oxygen O2 54:351 0:152
Nitrogen N2 63:148 12:53
Fluorine F2 53:48 0:252
Chlorine Cl2 172:15 1:4
Bromine Br2 265:85 6:1
Iodine I2 386:85 12:21
Carb. monox. CO 68:14 15:35
Carb. diox. CO2 216:58 518:5
Ammonia NH3 195:41 6:08
Hydr. chlor. HCl 158:91 14:0
Water H2 O 273:16 0:61166
Methane CH4 90:68 11:70
Acetylene C2 H2 192:60 128:2
Ethylene C2 H4 103:97 0:12
Ethane C2 H6 89:28 0:0011
Benzene C6 H6 278:69 4:78
Naphthalene C10 H8 353:43 1:00

(a) (b)
B B
p −v
L
S
O A O A

C C G

T s

Fig. 7.6 (a) The three coexistence lines, liquid-vapor OA, solid-liquid OB, and solid-vapor OC,
meet at the triple point O. (b) The sides of the triangle SLG, formed by the tie lines, are
perpendicular do the three coexistence lines
126 7 Phase Transition

Each line of coexistence, starting from inside the triangle SLG intercepts one of
the sides of the triangle. The lines of coexistence solid-liquid OB, liquid-vapor OA
and solid-vapor OC intercept the respective tie lines SL, LG and SG. The intercep-
tions occur at right angles because, as we have seen, the tie line is perpendicular
to the coexistence line. Since the angle formed by two lines of coexistence, which
comprises one of the phases, and the internal angle of the triangle with respect to
this phase are supplementary angles, then, it should be smaller than 180ı .

Problems

7.1 Determine the liquid-vapor transition line of a substance whose molar boiling
latent heat `e is independent of temperature. Assume that the vapor behaves as an
ideal gas and that the vapor molar volume is much greater than the molar liquid
volume. Solve the same problem for the case in which the latent heat depends on
temperature in accordance with `e D a C .b=T/.
7.2 In the neighborhood of the triple point, the liquid-vapor and the solid-vapor
transition lines are described by ln.p=a1 / D b1 =T and ln.p=a2 / D b2 =T,
respectively. Determine the pressure and temperature of the triple point and the
boiling, melting and sublimation latent heats around this point.
7.3 In the neighborhood of the solid-liquid transition the molar Helmholtz free
energies of the solid and liquid are given by

fS D RT ln.v  b1 / C c1 ; fL D RT ln.v  b2 / C c2 ;

where b1 and b2 are constant and c1 and c2 depend only of the temperature. Set up
the molar Gibbs free energy gS and gL of each phase. Determine the pressure p and
the volumes vS and vL of the phases in coexistence.
7.4 At low temperatures, the molar Helmholtz free energy of a van der Waals fluid
can be approximated by two expressions that describe the liquid and vapor phases.
For the liquid
a a
fL D RT ln.v  b/  C 2 .v  b/ C K;
b b
and for the gas

fG D RT ln.v  b/ C K; (7.24)

where K only depends on temperature. Make a Legendre transformation to obtain

the molar Gibbs free energy gL of the liquid and gG of the gas. Determine the line
of coexistence equating the two Gibbs free energy. Determine also vL , vG and the
latent heat ` as functions of temperature.
Chapter 8
Criticality

8.1 Critical Point

Liquid-Vapor Critical Point

Walking along the liquid-vapor coexistence line of a pure substance into direction
of high pressures and high temperatures, the vapor density increases and the liquid
density decreases to the point where the densities of the two phases become equal.
This point which determines the end of the coexistence line and corresponds to the
state in which the two phases become the identical, is called the critical point.
The first experimental observations of liquid-vapor critical point were made in
alcohol, benzene and ether by Cagniard de La Tour. The critical point of these
substances and of others, such as water, which are liquid at room temperature and
at the pressure of 1 atm, can be experimentally achieved in two steps. At first, the
coexistence line is reached by heating the liquid at constant pressure to the boiling
point. Once on the transition line, the substance may be brought to the critical point
by confining the liquid and its vapor in sealed vessel, followed by an increase in
temperature.
Increasing the temperature, the vapor-liquid system will necessarily be found on
the coexistence line, unless one of the two phases disappear, which may happen
before the critical point is reached. However, if the liquid-vapor system is prepared
so that the number of moles divided by the total volume of the vessel equals the
critical density, then the critical point is necessarily reached.
If we use this method to reach the critical point, we see that initially the meniscus
between liquid and vapor is well defined. As we approach the critical point, the
meniscus becomes less defined and disappears at the critical point. The region where
the meniscus was located becomes quite white, meaning that the light scattering is
very intense. This phenomenon, known as critical opalescence, is caused by large
fluctuations in density that occur around the critical point.

© Springer-Verlag Berlin Heidelberg 2017 127

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_8
128 8 Criticality

In the phase diagram, the critical point .Tc ; pc / lies therefore in the terminal point
of the liquid-vapor coexistence line. Near the critical point the coexistence line is
represented by the semi-straight line

p  pc D A.T  Tc /; T < Tc ; (8.1)

where A is a constant strictly positive which is identified as .@p=@T/v calculated at

the critical point. Using the Clausius-Clapeyron equation, we see that the differences
in molar entropy and molar volume in both phases in coexistence are related by

.sG  sL / D A.vG  vL /: (8.2)

Therefore, at the critical point, not only the molar volumes vG and vL become iden-
tical, but the molar entropies sG and sL become identical as well and, consequently,
the boiling latent heat `e vanishes.

Liquefaction

The pure substances that are gases under normal conditions of temperature and
pressure can be liquefied by compression alone, that is, can pass into the liquid state
when subjected isothermally to sufficiently high pressures. This occurs with carbon
dioxide, ammonia, ethane, propane, butane and other gases. These substances have
a critical temperature higher than the room temperature, so that the coexistence line
can be reached by isothermal compression. Carbon dioxide, for example, can be
liquefied at a temperature of 20 ı C under the pressure of 5.73 MPa. Other gases,
on the other hand, such as helium, neon, argon, krypton, xenon, hydrogen, oxygen,
nitrogen, carbon monoxide and methane cannot be liquefy at room temperature, no
matter what the applied pressure. These substances have critical temperatures below
the room temperature. A compression alone at room temperature is insufficient to
meet the liquid-vapor coexisting line.
Figure 8.1 shows the carbon dioxide isotherms at various temperatures above
and below the critical temperature, which occurs at 31.04 ı C. It is seen that, below
the critical temperature, the vapor-liquid coexistence may occur and therefore
liquefaction may happen by compression alone. Along a subcritical isotherm, the
volume decreases by compression and presents a jump at the transition. Along a
supercritical isotherm, the volume increases continuously with pressure, without any
discontinuity. The critical point is reached by compression alone along the critical
isotherm when pressures reaches 72.85 atm. The density of the carbon dioxide at the
critical point, where the liquid and vapor become identical, is 0.468 g/cm 3 . The first
experimental measurements around a critical point were made in carbon dioxide by
Andrews, who proved to be possible to continuously convert steam into liquid and
vice-versa, bypassing the critical point through an appropriate path.
8.2 van der Waals Theory 129

100 75
40,09 32,05
90 34,72 31,52
74
32,05 31,32
80 31,01 31,19
29,93 31,01

p (atm)
p (atm)

28,05
73 30,41
70
25,07 29,93
19,87
60 72
10,82
2,85
50
71
40

30 70
0 2 4 6 8 10 1 2 3 4
~ 3 ~ 3
v (cm /g) v (cm /g)

Fig. 8.1 Isotherms of carbon dioxide (CO2 ) in the diagram pressure p versus specific volume vQ,
for various values of temperature (in ı C), obtained experimentally by Michels et al. [127]

Table 8.1 shows experimental data corresponding to the critical point of several
pure substances. In addition to critical temperature, pressure and density, the table
presents also the compressibility factor Zc D pc vc =RTc determined at the critical
point.

8.2 van der Waals Theory

van der Waals Equation

van der Waals theory provides a quantitative description of the liquid-vapor phase
transition and of the corresponding critical point. Although it does not describe
accurately the behavior of the thermodynamic properties near the critical point it
provides an appropriate description of the phase coexistence and its relation with
the critical point.
The van der Waals equation is given by

RT a
pD  ; (8.3)
v  b v2
where a and b are constants. It describes approximately the behavior of gases and of
the liquid-vapor transition. For convenience we call a system obeying this equation
a van der Waals fluid or van der Waals gas, although it also describe the liquid
phase. From the microscopic point of view, a fluid that undergoes a liquid-vapor
transition must consist of molecules that attract each other over long distances and
repel at short distances. A van der Waals fluid is to be understood as composed of
hard spherical attractive molecules. The repulsion is the result of the rigidity of the
130 8 Criticality

Table 8.1 Critical temperature Tc , critical pressure pc and critical density c of various pure
substances. The last column shows the corresponding value of the compressibility factor Zc D
pc vc =RTc . Source: CRC
Substance Tc pc c Zc
K MPa g/cm3
Helium He 5:1953 0:22746 0:06964 0:303
Neon Ne 44:40 2:760 0:484 0:312
Argon Ar 150:663 4:860 0:531 0:292
Krypton Kr 209:40 5:500 0:919 0:288
Xenon Xe 289:73 5:840 1:110 0:287
Hydrogen H2 32:98 1:293 0:0310 0:306
Oxygen O2 154:581 5:043 0:436 0:288
Nitrogen N2 126:20 3:390 0:313 0:289
Fluorine F2 144:1 5:172 0:58 0:285
Chlorine Cl2 416:9 7:99 0:58 0:284
Bromine Br2 588 10:34 1:26 0:269
Carb. monox. CO 132:91 3:499 0:301 0:294
Carb. diox. CO2 304:14 7:375 0:468 0:274
Ammonia NH3 405:5 11:35 0:237 0:242
Hydrog. chloride HCl 324:7 8:31 0:45 0:249
Water H2 O 647:14 22:06 0:322 0:230
Methane CH4 190:56 4:592 0:1627 0:286
Ethane C2 H6 305:32 4:872 0:207 0:279
Ethylene C2 H4 282:34 5:041 0:214 0:281
Acetylene C2 H2 308:33 6:14 0:230 0:271
Propane C3 H8 369:83 4:248 0:220 0:276
Butane C4 H10 425:12 3:796 0:228 0:274
Benzene C6 H6 562:0 4:89 0:305 0:268
Naphthalene C10 H8 748 4:10 0:315 0:265
Methanol CH4 O 512:5 8:1 0:274 0:222
Ethanol C2 H6 O 514 6:1 0:274 0:241
Diethyl ether C4 H10 O 466 3:6 0:265 0:260
Acetone C3 H6 O 508 4:7 0:278 0:233
Chloroform CHCl3 536 5:47 0:499 0:294

molecules and is related to the parameter b. The attraction is related to the parameter
a, which must be understood as a measure the force of attraction between molecules.
The fundamental relation of a van der Waals fluid, in the Helmholtz free energy
representation, is obtained by integrating p D .@f =@v/T . Using (8.3), we get
a
f D RT ln.v  b/   K; (8.4)
v
8.2 van der Waals Theory 131

where K depends only on temperature. Therefore, the entropy s D .@f =@T/v is

given by

s D R ln.v  b/ C K 0 : (8.5)

To determine K.T/, we assume that the isochoric molar heat capacity cv of the van
der Waals is constant and equal to c. Thus T.@s=@T/v D c, that is, K 00 D c=T, so
that

K 0 D c ln T C c1 (8.6)

and

K D c.T ln T  T/ C c1 T; (8.7)

where c1 is a constant. To determine the molar energy u, we use the relation u D

f C Ts, from which we get
a
u D cT  : (8.8)
v

Maxwell Construction

At high temperatures, the van der Waals isotherms, in the p-v diagram, are
monotonically decreasing and the free energy, given by (8.4) is a convex function. At
low temperatures, however, the isotherms are no longer monotonic and free energy,
accordingly, loses convexity. Therefore, in this temperature regime, we adopt as the
free energy of the van der Waals fluid the convex hull of the function given by (8.4),
which is obtained by construction of a double tangential, as shown in Fig. 4.3a.
The convex hull corresponds to perform a Maxwell construction on the van der
Waals isotherms, which consists in tracing a straight line parallel to the axis of
the volumes, so that the areas between the segment and the isotherm are equal as
shown in Fig. 4.3b. The equivalence between the double tangent construction and
the construction of Maxwell is shown as follows. Let A and B be the two points
where the segment is tangent to the curve f versus v. The coordinates of these points
are .vL ; fG / and .vG ; fG /. Notice that the derivatives .@f =@v/ at the points A and B
are the same which means that the pressures corresponding to these two points are
equal. This common pressure we denote by p and, from the graph Fig. 4.3a, we get

f .vL /  f .vG / D p .vG  vL /: (8.9)

132 8 Criticality

On the other hand

Z vG   Z vG
@f
f .vL /  f .vG / D  dv D p.v/dv; (8.10)
vL @v T vL

and therefore
Z vG
p.v/dv D p .vG  vL /; (8.11)
vL

which is the algebraic expression of the Maxwell construction. For a better

understanding, denote by C the point where the segment cuts the curve in the p-
v diagram. Then the expression above is equivalent to
Z vC Z vG

p .vC  vL /  p.v/dv D p.v/dv  p .vG  vC /: (8.12)
vL vC

The right and left hand sides of this equation are identified as the two areas
mentioned above.
To determine the coexistence line, we use (8.11) or (8.9) and the equation

p D p.vL / D p.vG /: (8.13)

Eliminating vL and vG from these equations, we find p .T/ as a function of T. Using

expressions (8.3) and (8.4), for pressure and free energy, into (8.9) and (8.13), we
get equations whose solutions give, for each isotherm, the values of vL , vG and p ,
that are shown in Fig. 8.2.
We should give a justification for the Maxwell construction since it involves an
integral along a path that passes necessarily through unstable states, which would

(a) (b) 2.0

1.2
1.5
p / pc

p / pc

0.8 1.0

0.5
0.4

0
0 1 2 3 4 0 0.5 1.0 1.5 2.0
v / vc T / Tc

Fig. 8.2 van der Waals fluid. (a) Isotherms in the Clapeyron diagram. (b) Isochoric lines in the
phase diagram pressure versus temperature. The thick line is the line of phase coexistence
8.2 van der Waals Theory 133

be illegitimate. The justification would be the construction of the double tangent

which, as we have seen, leads us to the Maxwell construction but does not involve
unstable states. However, we must remember that, to obtain the free energy from
the equation of state, we have performed an integral that necessarily went through
the same unstable states, and again illegitimate. The unstable states, on the other
hand, can be avoided, as argued in the following if we assume the Griffiths postulate
according to which the free energy must be analytical within a single-phase region.
With this assumption the Maxwell construction is fully justified.
Suppose that we use only the monotonic part of p.T; v/ to determine f .Tv/ by
integration, thus avoiding unstable states. Below the critical temperature, we must
necessarily use two monotonic branches of the same isotherm. The free energy
obtained by integration along an isotherm will also have two branches, both convex.
These two branches connected by a double tangent, constitute with the double
tangent the free energy, which is entirely convex. The two branches, however, are
associated with different integration constants K1 .T/ and K2 .T/, which depend on
T. For f .T; v/ to be analytical within the single phase region it is required that
K1 .T/ and K2 .T/ are analytical in T within the same region. As a consequence, the
analytical continuation of K1 .T/ and K2 .T/, if they exist, should be the same. But in
this case the procedure just presented becomes equivalent to Maxwell construction.
In summary, the placement of the horizontal segment of the isotherm in a
position different from that given by the Maxwell construction would make f .T; v/
a nonalytic function inside a single-phase region.

Critical Point

Increasing the temperature along the coexistence line, the difference v D vG  vL

between the molar volumes decreases and vanishes at a certain temperature that we
call critical temperature Tc . At this temperature, v D 0 and the two phases become
identical. The corresponding point .Tc ; pc / in the phase diagram marks the end of
the coexistence line. Above the critical temperature the van der Waals fluid exhibits
a single phase.
The critical point can be determined by

@p @2 p
D0 and D 0; (8.14)
@v @v 2
because the critical point is both a stationary point and an inflexion point, as seen in
Figs. 8.1 and 8.2. From the van der Waals equation, we get

@p RT 2a
D C 3; (8.15)
@v .v  b/2 v
134 8 Criticality

and

@2 p 2RT 6a
2
D 3
 4; (8.16)
@v .v  b/ v

from which we obtain the critical molar volume

vc D 3b; (8.17)

and the critical temperature

8a
Tc D ; (8.18)
27bR
which substituted in van der Waals equation gives the critical pressure
a
pc D : (8.19)
27b2
Eliminating a and b from these three equations we reach the following relation

pc vc 3
D : (8.20)
RTc 8

The compressibility factor Z of a fluid is defined by Z D pv=RT. Therefore, from

this equation we conclude that at the critical point Zc D 3=8 for a fluid that satisfies
the van der Waals equation. In Table 8.1 we present the values of Zc obtained from
the critical properties of several pure substances. Although the experimental values
for Zc D pc vc =RTc are not equal to 3=8 D 0:375, they are close to each other,
mainly those related to the noble gases.

Expansion Around the Inflexion Point

For temperatures near the critical temperature, the molar volumes of the liquid and
vapor phases, vL and vG , are close to each other and close to the inflexion point v0
of the van der Waals isotherm. Therefore, if we want to determine vL and vG near
the critical temperature, it is reasonable to approximate the van der Waals isotherm
and the free energy by an expansion around v D v0 .
The expansion of p.v/ up to cubic terms gives

p D p0 C A.v  v0 / C B.v  v0 /3 ; (8.21)

8.2 van der Waals Theory 135

where v0 is given by p00 .v0 / D 0, p0 D p.v0 /, A D p0 .v0 / and B D p000 .v0 /=6. The
corresponding free energy f .v/ is given by

A B
f D f0  p0 .v  v0 /  .v  v0 /2  .v  v0 /4 : (8.22)
2 4
Notice that f0 , v0 , p0 , A and B depend only on temperature. The expansions above
are valid for small values of jv  v0 j. For temperatures near Tc , the constants can
be obtained as explicit functions of temperature. The expansion of these constants
around Tc gives

T  Tc
v0 D vc .1 C 2 /; (8.23)
Tc
T  Tc
p0 D pc .1 C 4 /; (8.24)
Tc
pc T  Tc
A D 6 (8.25)
vc Tc

and
3pc
BD : (8.26)
2vc3

To determine the coexistence line p .T/ and the values of vL and vG , we use the
equation that expresses the Maxwell construction

f .vL /  f .vG / D p .vG  vL / (8.27)

and the equations

p D p.vL / D p.vG /: (8.28)

The solution of these equations tell us that p D p0 , that is, the coexistence line (and
its extension) is given by

T  Tc
p D pc .1 C 4 / (8.29)
Tc

and that vL and vG are roots of the equation

A.v  v0 / C B.v  v0 /3 D 0; (8.30)

136 8 Criticality

that are distinct from. Solving this equation, we get

r r
A A
vG D v0 C e vL D v0  ; (8.31)
B B

so that
r
A
vG  vL D 2 : (8.32)
B

If we are very near the critical temperature Tc , then we can use the expres-
sions (8.25) and (8.26) for A and B and write
s
vG  vL Tc  T
D 2vc : (8.33)
2 Tc

Notice that
vG C vL T  Tc
D v0 D vc .1 C 2 /: (8.34)
2 Tc

Equivalently we may write

s
vG  vL pc  p
D vc (8.35)
2 pc

and

vG C vL p  pc
D v0 D vc .1 C /: (8.36)
2 2pc

Compressibility

At temperatures above the critical temperature, the isotherms are strictly monotonic
decreasing so that the isothermal compressibility T D .1=v/.@v=@p/ is positive
(Fig. 8.3). The largest value of T along an isotherm occurs at the inflexion point.
At the critical temperature it diverges because the derivative @p=@v vanishes at this
point. According to (8.21), the pressure varies with the molar volume, along the
critical isotherm, and around the critical point, according to the equation

3pc
p D pc  .v  vc /3 : (8.37)
2vc3
8.2 van der Waals Theory 137

(a) 2.0 (b) 8

1.5 6
κT / κ 0

cv / R
1.0 4

0.5 2

0 0
0 1 2 3 0 0.5 1.0 1.5 2.0

T / Tc T / Tc

Fig. 8.3 van der Waals fluid. (a) Isothermal compressibility T of the fluid along the critical
isochoric (T > Tc ) and of the liquid along the coexistence line (T < Tc ), where 0 D 1=pc .
(b) Isochoric molar heat capacity cv along the critical isochoric line. The jump of cv at T D Tc is
equal to 9R=2

Let us determine next the behavior of T around the critical point, along the
coexistence line and its extension, defined by (8.29), which coincides with the
inflexions points of the isotherms. From (8.21), we get

1 @p
D v D vŒA C 3B.v  v0 /2 : (8.38)
T @v

For temperatures above Tc , the molar volume along the extension of the coexistence
line is given by v D v0 , where v0 depends on T according to (8.23), so that

1 T  Tc
D vc A D 6pc . /; (8.39)
T Tc

that is,
1 Tc
T D . /: (8.40)
6pc T  Tc
If we wish to determine T along the critical isochoric we substitute v D vc
into (8.38). However, the second term inside the square brackets is of higher order
compared to the first and can be neglected. Therefore, the result (8.40) is also valid
for the critical isochoric near the critical point.
For temperatures below Tc , along the coexistence line, we should specify whether
the compressibility is being determined for the liquid (v D vL ) or for the vapor
(v D vG ). In both cases, however, according to (8.31), we have

A Tc  T 2
.v  v0 /2 D D 4. /vc ; (8.41)
B Tc
138 8 Criticality

so that
1 Tc  T
D 2vc A D 12pc. /; (8.42)
T Tc

that is,

1 Tc
T D . /; (8.43)
12pc Tc  T

for both phases. The results (8.40) and (8.43) show that the isothermal compress-
ibility diverges at the critical point according to

T D A˙ jTc  Tj1 ; (8.44)

where the coefficients AC and A , related to the behaviors of the compressibility

above and below the critical temperature, differ by a factor of 2, that is, AC =A D 2.

Molar Heat Capacity

Let us determine the isochoric molar heat capacity along the critical isochoric near
the critical point. Above the critical temperature, there is just a single phase whose
internal energy is given by (8.8). Since cv D .@u=@T/v , then

cv D c for T > Tc : (8.45)

Below the critical temperature, the two phases coexist and along the critical
isochoric the internal energy is given by
a a
u D x.cT  / C .1  x/.cT  /; (8.46)
vL vG

where x is the mole fraction of the liquid given by

vG  vc
xD : (8.47)
vG  vL

Deriving the expression for u with respect to T, we obtain

9
cv D c C R C c .T  Tc / for T < Tc ; (8.48)
2

where c is a positive constant. Therefore, the molar heat capacity cv of the van der
Waals fluid displays, along the critical isochoric, a jump at T D Tc equal to 9R=2.
8.3 Critical Behavior 139

8.3 Critical Behavior

Critical Exponents

Around the critical point, some thermodynamic properties have singular behavior.
The compressibility and the molar heat capacity, for example, grow without bounds
and diverge at the critical point. Figure 8.4 shows the divergence of molar heat
capacity of argon and carbon dioxide at the critical point. It is established that the
singularities are well represented by power laws when the quantities are placed as
functions of the deviations .T  Tc / and . p  pc / of temperature and pressure from
their critical values. Since the critical point marks the end of the coexistence line,
it is natural to examine firstly how the densities of the liquid and vapor approach
each other and become identical at the critical point. We assume that the difference
between them vanishes at the critical point in accordance with the power law

L  G  .Tc  T/ˇ or vG  vL  .Tc  T/ˇ : (8.49)

Another important quantity in the characterization of the critical behavior in

fluids is the isothermal compressibility, which diverges at the critical point. We
assume that along the critical isochoric (v D vc or  D c ) and for T > Tc , it
behaves according to

T  jT  Tc j : (8.50)

Below the critical temperature, T < Tc , we observe that the compressibility along
the coexistence line. In this case, it can be determined either for the liquid or for the
vapor. However, we assume that the compressibility of both phases have the same
behavior around the critical point and that it is given by (8.50).

(a) 250 (b) 300

Ar 250 CO2
200
cv (J/mol K)
cv (J/mol K)

200
150
150
100
100

50 50

0 0
146 148 150 152 154 10 20 30 40 50
o
T (K) θ ( C)

Fig. 8.4 Molar heat capacity cv along the critical isochoric line as a function of temperature, (a)
for the argon, with experimental data obtained by Voronel et al. [128], and (b) for carbon dioxide,
with experimental data obtained by Beck et al. [124]
140 8 Criticality

It is also interesting to determine how the density and pressure approach their
critical values. We assume that along the critical isotherm (T D Tc ) they have the
same behavior

j p  pc j  j  c jı or j p  pc j  jv  vc jı : (8.51)

We consider finally the isochoric molar heat capacity cv . We assume that this
quantity behaves along the critical isochoric according to

cv  jT  Tc j˛ ; (8.52)

both above and below the critical temperature. In general we expect a weak
divergence for the isochoric molar heat capacity. It is possible that this divergence
is of the logarithmic type

cv  ln jT  Tc j: (8.53)

The notation a.x/  b.x/ used to characterize the critical behavior, where both
a and b diverge or both vanish when x ! 0, means that the ratio a=b approaches a
finite constant or, equivalently, that ln a= ln b ! 1 when x ! 0.
If the critical behavior corresponds in fact to a power law, the critical exponent is
defined by the power law. It is convenient, however, to define the critical exponent
more broadly to include other behaviors that are not strict power laws. Thus, we
define the critical exponent , associated with a quantity a.x/, by

ln a.x/
 D lim ; (8.54)
x!0 ln x

or equivalently by a  x . With this definition we see that a logarithm divergence is

associated to an exponent zero.
The exponents ˇ,  , ı and ˛ contained in the power laws (8.49), (8.50), (8.51)
and (8.52), called critical exponents, concisely characterize the critical behavior.
Table 8.2 presents experimental values of the exponents for the critical point related
to the liquid-vapor transition of several pure substances. The results show that there
is a good agreement among the values for each exponent, indicating a universal
character of the critical behavior. The results (8.33), (8.44) and (8.37), obtained from
the van der Waals equation, give us ˇ D 1=2,  D 1 and ı D 3, which are distinct
from the experimental values, which means that the van der Waals equation does
not adequately describe the behavior of the fluid near the critical point. However,
the theory of van der Waals reveals that the critical exponents are independent of
the parameters a and b and in this sense are universal.
As an example of the determination of the exponent ˇ from the experimental
measurements done in carbon dioxide (CO2 ), we show in Fig. 8.5a the graph of L
and G versus temperature. To estimate the exponent ˇ and the critical temperature
from the data of Fig. 8.5, we plot lnŒ.L  G /=c  versus lnŒjTc  Tj=Tc . The value
8.3 Critical Behavior 141

Table 8.2 Critical exponents Substance ˛ ˇ  ı

˛, ˇ,  and ı related to the 3
liquid-vapor critical point of Helium-3 He 0:11 0:36 1:19 4:1
4
several pure substances Helium-4 He 0:13 0:36 1:18
Neon Ne 0:33 1:25
Argon Ar 0:13 0:34 1:21
Krypton Kr 0:36 1:18
Xenon Xe 0:11 0:33 1:23
Hydrogen H2 0:33 1:19
Oxygen O2 0:12 0:35 1:25
Nitrogen N2 0:33 1:23
Carb. diox. CO2 0:11 0:32 1:24
Sulf. hexafl. SF6 0:11 0:32 1:28
Ethylene C2 H4 0:33 1:18 4:4
Ethane C2 H6 0:12 0:34

(a) 1.0 (b) 1

0.8
ρ (g/cm )

0
3

ln(Δρ/ρc)

0.6

0.4
-1

0.2

0 -2
0 10 20 30 40 -10 -8 -6 -4 -2
o
θ ( C) ln|ΔT/Tc|

Fig. 8.5 Experimental data for liquid carbon dioxide in coexistence with its vapor, obtained by
Michels et al. [127]. In (a) we show the density of the liquid L (upper branch), the density of the
vapor G (lower branch) and the mean .L C G /=2 versus temperature. In (b) we show a log-log
plot of =c versus T=Tc , where  D L  G and T D T  Tc . The slope of the straight
line fitted to the data gives the result ˇ D 0:35

used for the critical temperature is the best fitting of a straight line: Tc D 31:06.
According to (8.49) the slope of this straight line is identified as the exponent ˇ.
From the fitting one obtains ˇ D 0:35. From this value one obtains also the critical
pressure pc D 72:85 atm.
The critical density is obtained in an approximate way by using the Cailletet-
Mathias rule. According to this rule, around the critical point, the arithmetic mean
.L C G /=2 is linear with temperature, as seen in Fig. 8.5a The value obtained in
this manner gives c D 0:467 g/cm3 . However, as we will see later this rule is not
valid very near the critical point.
142 8 Criticality

Other thermodynamic quantities can be analyzed as to the behavior around the

critical point. Using the following thermodynamic identity
 2
@p
cp D cv C Tv T (8.55)
@T v

and taking into account that .@p=@T/v is bounded and nonzero, we conclude that the
isobaric molar heat capacity cp diverges as T because the divergence of cv is weak
when compared to the divergence of T . From these result and using the identity
cv
s D T ; (8.56)
cp

we conclude that the adiabatic compressibility s diverges as cv . Therefore, along

the critical isochoric line, the critical behavior of cp is related to the exponent  and
that of s is related to the exponent ˛.
If we recall that the speed of sound in fluids is directly related with the adiabatic
compressibility by

1
vsom D p ; (8.57)
s

then the measurements of the velocity of sound along the critical isochoric line, near
the critical point, provides an experimental method to determine the exponent ˛.

Singular Part

The thermodynamic properties of a pure substance, including the singular behavior

around the critical point, are contained in the fundamental relation. In this sense
it is desirable to establish a form for the thermodynamic potential, valid around
the critical point, from which we power law introduced above could be derived.
The thermodynamic potential more convenient for this purpose is the one for which
all independent variables are thermodynamic field. We therefore choose the molar
Gibbs free energy g.T; p/ which is a natural function of T and p.
To proceed with the analysis of the critical behavior, it is convenient to
separate the molar Gibbs free energy g.T; p/ into one regular part gr .T; p/ and one
singular part gs .T; p/ containing the anomalies related to the critical point and the
discontinuities along the coexistence line. Thus, we write

g.T; p/ D gr .T; p/ C gs .T; p/: (8.58)

According to this separation, any thermodynamic quantity will also have a regular
and a singular part. The singular part of the quantities that are finite at the critical
8.3 Critical Behavior 143

point must be chosen such that it vanishes at this point, which can always be made
by the addition of appropriate terms in gr .T; p/. In this way, we guarantee that the
singular part of any nondiverging quantity will vanish at the critical point.
The regular part of the Gibbs free energy is assumed to have an expansion around
the critical point .Tc ; pc / which, up to linear terms in .TTc / and . ppc /, is given by

gr .T; p/ D gc C vc . p  pc /  sc .T  Tc /; (8.59)

where gc D g.Tc ; pc /, vc and sc are the molar volume and molar entropy at the
critical point.
To explore in a systematic way the consequences of a privileged direction in the
phase diagram, determined by the transition line, it is convenient to make a change
of variables. Near the critical point, the line of coexistence is represented by the
semi straight line defined by (8.1). The axes corresponding tho the new variables
have origin at the critical point and one of them coincides with the coexistence
line (8.1). We adopt the following transformation

" D .T  Tc / C B. p  pc / (8.60)

and

 D A.T  Tc / C . p  pc /; (8.61)

where B is a constant to be defined later. We remember that A is a constant strictly

positive. The line of coexistence (8.1) in the new variables is given by  D 0 and " <
0.
In this new reference system we write the singular part of the molar Gibbs free
energy as

gs .T; p/ D G ."; /; (8.62)

so that

@g
vD D vr C BG1 ."; / C G2 ."; / (8.63)
@p

and

@g
sD D sr  G1 ."; / C AG2 ."; / (8.64)
@T

where G1 D @G =@" and G2 D @G =@ and vr D @gr =@p and sr D @gr =@T are the
regular parts of the volume and entropy.
144 8 Criticality

Deriving the expression (8.63) and (8.64) with respect to the pressure and
temperature, respectively, we get
 
@v
vT D  D B2 G11 ."; /  2BG12 ."; /  G22 ."; /; (8.65)
@p T

and
 
1 @s
cp D D G11 ."; / C 2AG12 ."; /  A2 G22 ."; /; (8.66)
T @T p

where G11 D @2 G =@ 2 , G12 D @2 G =@"@ and G22 D @2 G =@"2 . Near the critical
point we use the relation cv D cp  TvA2 T , obtained from the identity (8.55),
to write
1
cv D .AB C 1/f.AB  1/G11 ."; / C 2AG12 ."; /g: (8.67)
T
In these three expressions above we have omitted the regular part by taking
into account that these quantities diverges at the critical point and are therefore
dominated by the singular part.
Up to now the constant B contained in the transformation of variables (8.60)
and (8.61) has not been specified. Next, we choose this constant in such a way that
the singular part of the Gibbs free energy is symmetric with respect to the transition
line, that is, so that G ."; / is an even function with respect to the variable ,

G ."; / D G ."; /: (8.68)

In addition to this symmetry, the function G ."; / has the usual properties of the
thermodynamic potentials; it is continuous and concave. The derivative of G1 ."; /
is an even function of  and is continuous. The derivative G2 ."; / is an odd function
of  and continuous except over the line of coexistence where it has a discontinuity.
These properties are shown in Fig. 8.6.
Along the coexistence line, the vapor molar volume vG and the liquid molar
volume vL are obtained by taking the limits  ! 0 and  ! 0C , respectively, in
expression (8.63):

vG D vr C BG1 ."; 0/ C G2 ."; 0 / (8.69)

and

vL D vr C BG1 ."; 0/ C G2 ."; 0C /: (8.70)

8.3 Critical Behavior 145

G1
(a) (b)
G
ζ
ζ

ε
ε

(c) (d) −G 22
−G2

ζ
ζ

ε
ε

Fig. 8.6 Singular part of the Gibbs free energy G around the critical point and its derivatives G1 ,
G2 as G22 functions of " and 

The difference between the liquid and vapor molar volumes is, therefore, related
with the discontinuity of G2 ."; / by

vG  vL D G2 ."; 0 /  G2 ."; 0C /: (8.71)

The mean of the liquid and vapor molar volumes is

1
.vG C vL / D vr C BG1 ."; 0/: (8.72)
2
Along " D 0,

v D vr C BG1 .0; / C G2 .0; / (8.73)

From the equality (8.65), we obtain the behavior of the isotherm compressibility
along the coexistence line and its extension

vT D B2 G11 ."; 0/  G22 ."; 0/: (8.74)

146 8 Criticality

Similarly, from the equality (8.67), we obtain the isochoric molar heat capacity
along the coexistence line and its extension

1
cv D .A2 B2  1/G11 ."; 0/; (8.75)
T

where we have taken into account that G12 ."; 0/ D 0.

Widom Scaling Theory

The Widom scaling theory concerns the behavior of the thermodynamic quantities
around the critical point. According to this theory, the value of a certain thermody-
namic quantity in a certain point of he phase diagram can be obtained from another
point by a change is scale. The critical point is taken as the reference point for the
changes of scale. If we consider the following scale transformation

."; / ! .
";
0 /; (8.76)

that connect two points of the phase diagram, then the singular parts of the
Gibbs free energy corresponding to these two points are connected by a scale
transformation, that is,

G .
";
0 / D
00 G ."; /: (8.77)

The parameters
,
0 and
00 are the enlargement factors related to ",  and G ,
respectively. Without losing generality we may write
0 D
b and
00 D
a , so that

G ."; / D
a G .
";
b /; (8.78)

valid for any positive values of

and that expresses the homogeneity of G ."; / with
respect to the variables " and . The exponents a and b are connected to the critical
exponents introduced above, as we will se later.
The homogeneity property of G ."; / implies that its derivatives are also
homogeneous functions. Indeed, deriving (8.78) with respect to " and , we get

G1 ."; / D
aC1 G1 .
";
b / (8.79)

and

G2 ."; / D
aCb G2 .
";
b /: (8.80)
8.3 Critical Behavior 147

Deriving these expressions with respect to " and , respectively, we get in addition

G11 ."; / D
aC2 G11 .
";
b /; (8.81)

and

G22 ."; / D
aC2b G22 .
";
b /: (8.82)

Next we determine the behavior of these quantities along the coexistence line
and its extension. For " < 0 and  D 0, and choosing
D j"j1 , we get

G1 ."; 0/ D j"ja1 G1 .1; 0/; (8.83)

G2 ."; 0˙ / D j"jab G2 .1; 0˙ /; (8.84)
G11 ."; 0/ D j"ja2 G11 .1; 0/; (8.85)
G22 ."; 0/ D j"j a2b
G22 .1; 0/: (8.86)

For " > 0 and  D 0, and choosing

D "1 , we obtain

G11 ."; 0/ D "a2 G11 .1; 0/; (8.87)

G22 ."; 0/ D "a2b G22 .1; 0/: (8.88)

Replacing these results into the expressions (8.71), (8.74) and (8.75), we reach
the results

vG  vL  j"jab ; (8.89)
T  j"ja2b ; (8.90)
cv  j"ja2 : (8.91)

We see therefore that the Widom scaling theory predicts power laws for the
critical behavior of these quantities. The result for the compressibility was obtained
considering that the second term of (8.74) dominates over the first. Comparing these
expressions with the (8.49), (8.50) and (8.52), we find the relations between the
critical exponents and the indices a and b:

˛ D a C 2; (8.92)
ˇ D a  b; (8.93)
 D a C 2b: (8.94)

In doing the comparison, we have taken into account that along  D 0 the variable
" is proportional to T  Tc .
148 8 Criticality

For " D 0 and  > 0, and choosing

D  1=b , we get

G2 .0; / D  .ab/=b G2 .0; 1/: (8.95)

Replacing into (8.73), we reach the result

v  vc   .ab/=b ; (8.96)

where we have taken into account only the dominant term in (8.73) and we have
approximated vr by vc . Comparing with expression (8.51), and taking into account
that along " D 0,  is proportional to p  pc , we get the following relation between
the exponent ı and the indices a and b

1 ab
D : (8.97)
ı b

Notice, however, that the exponent ı is defined by the behavior of v  vc (or   c )

versus p  pc along the critical isotherm and not along " D 0. On the other hand, it
is possible to argue that the critical behavior will be the same for both trajectories.
In fact, it will be the same along any straight line that intercepts the critical point,
that does not coincide with the coexistence line.
The four critical exponents ˛, ˇ,  , and ı are not all independent. We can choose
only two as independent. Eliminating a and b we obtain several relations between
them, that include the Rushbrooke relation

˛ C 2ˇ C  D 2; (8.98)

the Griffiths relation

˛ C ˇ.ı C 1/ D 2 (8.99)

and the Widom relation

 D ˇ.ı  1/: (8.100)

The experimental values shown in Table 8.2 are in accordance with these relations,
within the experimental errors. It is interesting to notice that the values of the
exponents obtained from the van der Waals equation, ˛ D 0, ˇ D 1=2,  D 1
and ı D 3 also satisfy these relations. The jump in the isochoric molar heat capacity
cv , predicted by the de van der Waals theory, must be understood as corresponding
to ˛ D 0.
Replacing the result (8.83) into (8.72), we obtain

1
.vG C vL /  vr  j"j1˛ (8.101)
2
8.3 Critical Behavior 149

where we used the relation a D 2  ˛. This result can be used to determine the
critical exponent ˛ from the measurements of vG and vL . Since the regular part vr
is linear with temperature, we see that the last term is a correction to the Cailletet-
Mathias rule.

Problems

8.1 Determine the critical temperature Tc , the critical pressure pc and the critical
molar volume vc corresponding to the Dieterici and Berthelot equations. Find also
the critical compressibility factor Zc D pc vc =RTc .
8.2 Show that the critical behavior of the latent heat is related to the exponent ˇ.
8.3 Show that the derivative of g.T; p/ along the coexistence line is well defined,
that is, has the same value if we approach it from the liquid side or the vapor
side. Determine its value at the critical point. Along a line perpendicular to the
coexistence line the derivative has a jump. Determine this jump.
8.4 Determine the critical exponent related with the following functions

ln jxj; x1=2 ln jxj; ax C bx1=2 ; ax1 C bx1=2 :

8.5 Show that the Dieterici and Berthelot equations of state provide the same
critical exponent as the van der Waals equation.
8.6 Show that the coefficient of thermal expansion diverges at the critical point with
the same exponent  related to T .
8.7 Demonstrate relation (8.55).
Chapter 9
Mixtures

9.1 Introduction

Mixtures of Pure Substances

The mixtures of pure substances are quite common in nature. Some mixtures are
homogeneous others are heterogeneous. The atmospheric air is a homogeneous
gaseous mixture of nitrogen, oxygen and other gases in smaller proportions. Oil is
a natural mixture of hydrocarbons. The rocks, like granite, are solid heterogeneous
mixtures. The homogeneous mixtures have uniform physical aspect and consists of
a single thermodynamic phase. The heterogeneous mixtures, on the other hand, are
formed by an aggregate of thermodynamic phase in coexistence. Due to gravity,
the different phases of a heterogeneous liquid mixture are arranged one above the
other, separated by menisci and sorted according to density. In a heterogeneous solid
mixture the phases are displayed in a fragmented and intricate arrangement.
The separation of the components of a mixture to obtain pure compounds is a
major problem. Once the separation is made, the pure substances may be used to
obtain new compounds. Distillation is an example of a process of liquid separation
by evaporation and condensation of vapor.
In this chapter and the next, we analyze mixtures of pure substances that do
not lose their identities when they enter the composition of a mixture. That is, the
chemical species in the mixture are the same as the pure substances used to make
the mixture. They do not undergo any change, that is to say there is no chemical
reactions within the mixture. Therefore, the number of moles of each pure substance
remains unchanged before and after the mixture composition.

© Springer-Verlag Berlin Heidelberg 2017 151

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_9
152 9 Mixtures

The chemical species comprising a mixture are generically called constituents

of the mixture. In a molecular mixture, each constituent is represented by a type
of molecule. If there is no chemical reaction, the number of chemical species
corresponding to the number of components of the mixture. In mixtures in which
chemical reactions occur, which will be treated in the last chapter, the number
of components equals the number of chemical species minus the number of
independent chemical reactions.

Solutions

If a mixture displays a single phase, that is, if it is homogeneous, it is also called

a solution. In this case we say that the components are miscible. The component
with the greatest concentration is conventionally called solvent and the others, of
lower concentration, solutes. In a solution, we also say that the component of the
greatest concentration dissolves the ones with lower concentration. If a mixture
has more than one phase, that is, if it is a heterogeneous mixture, then each
phase is a solution. In this case, however, each one of the coexisting solutions is
called saturated solution. It is also possible that a component of the mixture is
completely immiscible, forming a phase different from the others, which is not a
proper solution.
A solution may be gaseous, liquid or solid. The gases are miscible in all
proportions always forming a homogeneous mixture and therefore a single phase.
A heterogeneous mixture can therefore have at most a single gas phase. The liquids
dissolve gases, solids and liquids, sometimes in large quantities. Since the liquids
are not fully miscible it is possible the occurrence of more than one liquid phase
in a heterogeneous mixture. The solids also can dissolve gases, liquids and other
solids, but the limits of miscibility are generally lower than that of liquids. Due to
this reason, the number of solid phases in a heterogeneous mixture can be great.

9.2 Fundamental Relation

Thermodynamic State

An adequate description of the thermodynamic state of a mixture must contain

quantities that determine the amount of each component. We may use for this
purpose the mass, but the number of moles of each component is preferable. Thus,
in the standard representation, the thermodynamic state of a mixture consisting of c
components is defined by the extensive variables: entropy S, volume V, number of
moles N1 of the first component, number of moles N2 of the second component, . . . ,
number of moles of the Nc of the c-th component. The corresponding fundamental
9.2 Fundamental Relation 153

relation is given by the internal energy U.S; V; N1 ; N2 ; : : : ; Nc / whose differential is

dU D TdS  pdV C 1 dN1 C 2 dN2 C : : : C c dNc ; (9.1)

where T is the temperature, p is the pressure, and 1 , 2 , . . . , and c are the

chemical potentials of the components 1, 2, . . . , and c, respectively. Other represen-
tations can also be employed including the Helmholtz representation, the enthalpy
representation and the Gibbs representation. The respective fundamental relations
are given by the Helmholtz free energy F.T; V; N1 ; N2 ; : : : ; Nc /, by the enthalpy
H.S; p; N1 ; N2 ; : : : ; Nc / and by the Gibbs free energy G.T; p; N1 ; N2 ; : : : ; Nc /, which
are obtained one from another by means of Legendre transformations.
The Gibbs representation is particularly useful in the study of mixtures. The
differential of the Gibbs free energy is given by

dG D SdT C Vdp C 1 dN1 C 2 dN2 C : : : C c dNc ; (9.2)

where

@G @G @G
SD ; VD ; i D : (9.3)
@T @p @Ni

Being G an extensive function with respect to the variables Ni , we may write

G D 1 N1 C 2 N2 C : : : C c Nc ; (9.4)

an expression that can be used to recover the fundamental relation G.T; p; N1 ;

N2 ; : : : ; Nc / in case the chemical potentials i .T; p; N1 ; N2 ; : : : ; Nc / are known.
From (9.2) and (9.4) one obtains the Gibbs-Duhem equation

 SdT C Vdp  N1 d 1  N2 d 2  : : :  Nc d c D 0: (9.5)

Molar Quantities

In many applications, we will have the opportunity to use molar quantities such as
the molar Gibbs free energy g, defined by

G
gD ; (9.6)
N
154 9 Mixtures

and the mole fraction xi of the i-th component, defined by

Ni
xi D ; (9.7)
N
where

N D N1 C N2 C : : : C Nc (9.8)

is the total number of moles of the mixture. Notice that the c mole fractions are not
independent because

x1 C x2 C : : : C xc D 1: (9.9)

Equations (9.4) and (9.5) reduce to

g D 1 x1 C 2 x2 : : : : C c xc (9.10)

and

 sdT C vdp  x1 d 1  x2 d 2  : : :  xc d c D 0; (9.11)

where s D S=N is the molar entropy and v D V=N is the molar volume. From (9.10)
and (9.11), we conclude that

dg D sdT C vdp C 1 dx1 C 2 dx2 C : : : C c dxc : (9.12)

Notice, however, that the differentials dxi are not independent but are tied by

dx1 C dx2 C : : : C dxc D 0: (9.13)

Usually we eliminate dx1 in (9.12), using (9.13), which allows us to write

dg D sdT C vdp C . 2  1 /dx2 C : : : C . c  1 /dxc (9.14)

and to conclude that g can be considered as a function of temperature T, pressure p

and mole fractions x2 , x3 , . . . , xc .
In some circumstances it is appropriate to use the activity zi , introduced by Lewis
and defined for each component by

zi D e i =RT : (9.15)
9.3 Mixture of Ideal Gases 155

9.3 Mixture of Ideal Gases

Helmholtz Free Energy

Consider a mixture of c distinct ideal gases in a vessel of volume V at temperature T.

We postulate that the Helmholtz free energy of the mixture F.T; V; N1 ; N2 ; : : : ; Nc /
is equal to the sum of the Helmholtz free energy of each gas as if each one of them
were alone in the vessel, that is, as if each one were confined in a vessel of volume
V at temperature T. That is,

X
c
V
FD Ni fi .T; /; (9.16)
iD1
Ni

where fi .T; vi / is the molar Helmholtz free energy of the i-th component alone,
given by

T vi
fi .T; vi / D ai T  ci T ln  RT ln ; (9.17)
T0i v0i

where ai , ci , T0i and v0i are characteristic constants of the ideal gas i.
The pressure p D @F=@V is thus

X
c
Ni RT NRT
pD D : (9.18)
iD1
V V

The partial pressure pi , which is the pressure exerted by the i-th ideal gas if it were
alone in the vessel of volume V, is given by pi D @fi =@vi or

RT Ni RT
pi D D : (9.19)
vi V

Therefore,

X
c
pD pi ; (9.20)
iD1

which is Dalton law: the pressure of an ideal gas mixture is the sum of the partial
pressures.
156 9 Mixtures

Gibbs Free Energy

The Gibbs free energy G.T; p; N1 ; N2 ; : : : ; Nc / of an ideal gas mixture is obtained

from the Legendre transformation of the free energy F, given by (9.16). Making the
transformation, we get

X
c
Ni
GD Ni fgi C RT ln g; (9.21)
iD1
N

where gi is the molar Gibbs free energy of the i-th component alone, given by

T p
gi D bi T  .ci C R/T ln C RT ln ; (9.22)
T0i p0i

where bi D ai  R and p0i D RT0i =v0i , and is related with fi .T; vi / by means of a
Legendre transformation.
The entropy S D @G=@T, as a function of T and p, is given by

X
r
Ni
SD Ni fsi  R ln g; (9.23)
iD1
N

where si D @gi =@T is the molar entropy of the i-th component alone, given by

T p
si D ai C ci C .ci C R/ ln  R ln : (9.24)
T0i p0i

The chemical potential i D @G=@Ni of the i-th component is

Ni
i D gi C RT ln : (9.25)
N

It is worth noting that the molar Gibbs free energy g D G=N, the molar entropy
s D S=N and the chemical potential can be written in the forms

X
c
gD xi fgi C RT ln xi g; (9.26)
iD1

X
c
sD xi fsi  R ln xi g (9.27)
iD1

and

i D gi C RT ln xi ; (9.28)
9.3 Mixture of Ideal Gases 157

where xi D Ni =N. Recall that the mole fractions xi are not independent but are
related by (9.9).
The activity zi of the i-th ideal gas is connected to the partial pressure pi by the
linear relation

pi D p0i zi (9.29)

where p0i depends only on temperature. This relation is obtained by noting that for
an ideal gas mixture pi =p D Ni =N D xi and by using (9.22) and (9.28).

Entropy of Mixing

Let us determine what is called the entropy of mixing. Consider a vessel divided
into c compartments separated by diathermal and mobile walls. Each compartment
contains a different type of ideal gas and the i-th compartment contains Ni moles
of ideal gas of type i. As the walls that separate the compartments are diathermal
and mobile, all gases have the same temperature T and the same pressure p. Next
the inner walls are removed and the gases mix. The temperature and pressure of
the ideal gas mixture does not change, but the entropy increases. The entropy of
mixing is the difference between the entropy of the mixed gases and the entropy of
the gases isolated, within their respective compartments, at the same temperature
and pressure.
The total entropy of the isolated gases, each in its compartment, is simply

X
c
SI D Ni si ; (9.30)
iD1

where si is the molar entropy of the i-th component. For an ideal gas mixture, the
entropy is

X
c
Ni
SF D Ni fsi  R ln g; (9.31)
iD1
N

so that

X
c
Ni
S D SF  SI D R Ni ln : (9.32)
iD1
N

As Ni =N  1, then S  0.
158 9 Mixtures

The variation in the Gibbs free energy is

X
c
Ni
G D RT Ni ln D TS; (9.33)
iD1
N

from which one concludes that the change in the enthalpy is zero, since H D
G C TS D 0. We conclude also that the volume change is zero. Indeed, since
V D @G=@p D 0 and being G independent of p, then V D 0. The variation
in the internal energy is also zero, since U D H  pV D 0.
Suppose that the gases are all equal, for example, all of type 1. In this case there
should be no increase in entropy. The initial entropy is

X
c
SI D Ni s1 : (9.34)
iD1

The final entropy, however, is

SF D Ns1 ; (9.35)
P
since the vessel is filled by a single component, that has N moles. As N D i Ni ,
then SF D SI and there is no change in entropy as expected.

9.4 Dilute Solutions

Raoult and Henry Laws

Let us examine here the properties of dilute solutions. We consider a mixture of c

components such that c  1 components have very small concentrations. We adopt
the component 1 as the solvent and the other components as solutes. The sum of the
mole fractions of solutes is denoted by x, that is,

x D x2 C x3 C : : : C xc : (9.36)

For a dilute solution x << 1 so that the mole fraction of the solvent x1 D 1  x is
close to unit.
Denoting by N D N=V the total number of moles of the solution per unit volume
and by Ni D Ni =V the number of moles of the i-th component per unit volume,
then the condition x << 1 implies N2 C N3 C : : : C Nc << .
N That is to say, in a
dilute solution the number of moles per unit volume of the solutes must be small.
Note, however, that the number of moles per unit volume of the solvent need not
necessarily be small, unlike what happens with a mixture of ideal gases, in which
all the components must have low density. On the other hand, a mixture of ideal
gases need not be diluted.
9.4 Dilute Solutions 159

We now examine the behavior of the molar Gibbs free energy g.T; p; x2 ; x3 ;
: : : ; xc / for small values of x. We are considering g to be a function of the mole
fractions of the solutes and not of the solvent, because the mole fractions can not
all be independent. If we use (9.10), we see that the determination of g is reduced
to the determination of the chemical potentials i . Proceeding according to Lewis,
we start with two assumptions concerning the behavior of chemical potentials for
dilute solutions. The first hypothesis states that the chemical potential of the solvent
reduces the chemical potential of pure solvent in the limit x ! 0. The second
hypothesis says that the activities zi D expf i =RTg of the solutes are proportional
to their mole fractions, that is,

zi D i xi ; (9.37)

where the constant i depends on the pressure, temperature and nature of the solvent
and solute, but not of the mole fractions. We call this expression Henry law for dilute
solutions. The chemical potential of the solute behaves, therefore, according to

i D ıi C RT ln xi ; (9.38)

where ıi D RT ln i .
To determine the chemical potential of the solvent, we use the Gibbs-Duhem
equation (9.11). But before, we see that for T and p constant, we have d i D
RTdxi =xi for i D 2; 3; : : : ; c which, replaced in the Gibbs-Duhem equation, gives

.1  x/d 1 C RTdx D 0; (9.39)

because x1 D 1  x and x D x2 C x3 C : : : C xc . Integrating this equation, we get

1 D g1 C RT ln.1  x/; (9.40)

where we take into account that in the limit x ! 0, the chemical potential of the
solvent reduces to the chemical potential of pure solvent, which is identified with
the molar Gibbs free energy of the pure solvent, denoted by g1 . The activity z1 D
expf 1 =RTg of the solvent is given therefore by

z1 D z1 .1  x/; (9.41)

that is, the activity of the solvent of diluted solutions is proportional to the mole
fraction of the solvent x1 D 1  x. Moreover, the constant of proportionality z1 D
expfg1 =RTg is identified with the activity of the pure solvent. We call this expression
Raoult law for dilute solutions.
We remark that the Raoult law concerns the solvent (whose mole fraction is close
to unity), while Henry law concerns the solutes (whose mole fractions approach
zero). If the dilute solution is in coexistence with its vapor, then the equilibrium
condition implies that the activities of the components in the solution should be
160 9 Mixtures

equal to the activities of the components in the vapor phase. If, in addition, the
vapor can be considered an ideal gas mixture, then, according to (9.29), the activities
zi are proportional to the partial pressures pi of the gases. Therefore, from Raoult
and Henry laws, we can write

p1 D p1 .1  x/; (9.42)

for the solvent, and

pi D ki x; (9.43)

for the solutes, where p1 , the vapor pressure of the pure solvent, and ki depend only
on temperature. These relations constitute the original forms of Raoult and Henry
laws, respectively.
Replacing, finally, the expressions for the chemical potentials in (9.10), we obtain
the explicit form for the molar Gibbs free energy of a dilute solution,

X
c X
c
g D g1 .1  x/ C ıi xi C RT.1  x/ ln.1  x/ C RT xi ln xi : (9.44)
iD2 iD2

The chemical potential of the solvent can be written as

1 D g1  RTx; (9.45)

since for small values of x we can approximate ln.1  x/ by x. Importantly, the
chemical potential 1 , both in the form (9.40) or as (9.45), does not depend on
the characteristics of the solutes, but only on the sum of their mole fractions. This
particularity provides the emergency of so-called colligative properties, which are
those properties that depend on the number of moles of the solutes but do not
depend on the type of solutes present in the solution. Typical examples of colligative
properties are:
(a) osmotic pressure;
(b) elevation of the boiling temperature of liquid solutions with non-volatile solutes
(the vapor has a single component: the solvent), and
(c) lowering of the fusion temperature of the solids immiscible in solute (the solid
phase contains a single component: the solvent).

Osmotic Pressure

Consider a system consisting of two compartments separated by a rigid and

diathermal wall so that, at equilibrium, the temperature will be equal, but not the
pressure. On one side there is a dilute solution and on the other there is only the pure
9.4 Dilute Solutions 161

substance number 1, the solvent. The wall is permeable but only to the component
1. The wall being permeable to the solvent, in equilibrium, the chemical potential of
the solvent in solution should be equal to the chemical potential of the pure solvent,
that is,

1 .T; p; x2 ; x3 ; : : : ; xc / D g1 .T; p0 /; (9.46)

where p0 is the pressure of the pure solvent. The osmotic pressure ˘ is defined as
the difference in pressure between the two sides of the wall, that is, ˘ D p  p0 .
If the solution is diluted, then the chemical potential of the solvent is given
by (9.45) which, replaced in (9.46), yields

g1 .T; p/  RTx D g1 .T; p0 /: (9.47)

But for small values of ˘ D p  p0 , we may approximate g1 .T; p/ by

g1 .T; p/ D g1 .T; p0 / C v1 .T; p0 /.p  p0 /; (9.48)

where v1 D @g1 =@p is the molar volume of the pure solvent. Since the solution is
diluted, v1 can be replaced by the molar volume v of the solution. Therefore,

RTx D ˘ v; (9.49)

or

˘ D Nsolute RT; (9.50)

which is the van ’t Hoff equation, where Nsolute D x=v D .N2 C N3 C : : : C Nc /=V is
the total number of moles of solute per unit volume (molarity). For dilute solutions,
the osmotic pressure is thus independent of the nature of the solvent, depending only
on the molarity of the solute.
The experimental measurement of the osmotic pressure can be used for the
determination of molecular weight M of the molecules, especially macromolecules.
For a solution with only one type of solute Nsolute D solute =M, where solute is the
density of the solute (mass per unit volume). Therefore,

RT
M D solute ; (9.51)
˘
so that the experimental measurements of pressure and density provide the molecu-
lar weight of the solute. As the above equation should be valid for dilute solutions
then it is convenient to plot the right side of (9.51) as a function of density for
various values of solute and make the extrapolation solute ! 0.
162 9 Mixtures

Lowering the Melting Point

We will examine here the lowering of the melting temperature of a pure substance
when certain impurities are dissolved in it. Suppose that the impurities and the pure
substance are miscible in the liquid phase but not on the solid phase. Thus we should
examine the coexistence of a liquid phase, which is a dilute solution, and a solid
phase consisting only of the pure solvent. The liquid phase is a dilute solution whose
chemical potential of the solvent is given by (9.45), that is,

1 D gL1 .T; p/  RTx; (9.52)

where gL1 .T; p/ is the molar Gibbs free energy of the pure solvent and x is the sum of
mole fractions of the solutes in the liquid phase. Since we are assuming that the solid
phase is composed only by the solvent, then the chemical potential of the solvent in
the solid phase is independent of x and is equal to the molar Gibbs free energy of
the solid which we denote by gS1 .T; p/. The equilibrium condition is then

gL1 .T; p/  RTx D gS1 .T; p/; (9.53)

so that

RTx D gL1 .T; p/  gS1 .T; p/: (9.54)

Let T0 be the temperature for which gL1 .T0 ; p/ D gS1 .T0 ; p/, which is the melting
temperature of the pure solid. Then for small differences T  T0 we have

gL1 .T; p/  gS1 .T; p/ D ŒsL1 .T0 ; p/  sS1 .T0 ; p/.T  T0 /; (9.55)

or

`f
gL1 .T; p/  gS1 .T; p/ D  .T  T0 /; (9.56)
T0

where `f D T0 .sL1  sS1 / is the molar fusion latent heat of the pure solid. Therefore,

RT02
T0  T D x; (9.57)
`f

which is the van ’t Hoff formula for the lowering of the fusion temperature of the
solvent. The coefficient RT02 =`f depends only on the properties of the solvent and
does not depend on the nature of the solute, but only on its number of moles, and is
known as the cryoscopic constant. Notice that this expression is valid as long as the
liquid solution coexist with the pure solid.
9.5 General Solutions 163

If the solute is composed by a single substance, the measurement of the decrease

in temperature T D T0  T can be used to determine the molar mass of the solute
M2 from the molar mass of the solvent M1 . If the mole fraction x of the solute is
small then x D N2 =N1 . But

N2 m2 M1
D ; (9.58)
N1 m1 M2

where m1 and m2 are the masses of the solvent and solute, respectively. This
equation, combined with the van ’t Hoff formula (9.57), allows the determination
of M2 from M1 .

9.5 General Solutions

Ideal Solutions

So far we have obtained two expressions for the Gibbs free energy. One related to
the mixture of ideal gases, valid in the regime of low densities of all components,
and another related to the dilute solutions, valid in the regime of low mole fractions
of solutes. We now develop expressions of the Gibbs free energy for solutions that
are not necessarily dilute or a mixture of ideal gases.
Suppose that N1 moles of a pure substance 1, N2 moles of a pure substance 2,
. . . and Nc moles of a pure substance c, are used to form a mixture. Initially, all pure
substances which are isolated have the same temperature T and the same pressure p.
The process of formation of the mixture is such that the final mixture has the same
temperature and the same pressure of the isolated substances. In this process, the
change in Gibbs free energy G is given by

X
c
G D G.T; p; N1 ; N2 ; : : : ; Nc /  Ni gi .T; p/; (9.59)
iD1

where G.T; p; N1 ; N2 ; : : : ; Nc / is the Gibbs free energy of the mixture and gi .T; p/
is the molar Gibbs free energy of the i-th pure substance. Similarly, we define the
variation in entropy S, the volume change V, and the enthalpy change H.
In the process of formation of the mixture, there will be certainly an increase in
the entropy and in general one expects a nonzero variation of the enthalpy and of
other thermodynamic properties. In an ideal mixture, or ideal solution, we assume
that the variation in entropy is the same as that occurring in a mixture of ideal gases,
at constant temperature and pressure, that is

X
c
Ni
S D R Ni ln ; (9.60)
iD1
N
164 9 Mixtures

and that the variation in enthalpy vanishes, that is, H D 0. As the mixture of
the substances is an isothermal process, then H D G C TS, so that, for ideal
solutions,

G D TS: (9.61)

A significant consequence is that there is no variation of volume nor variation of

internal energy when pure substances are mixed to form an ideal solution. In fact,
deriving both sides of (9.61) with respect to pressure and taking into account that
S is independent on p, we see that V D @G=@p D 0. Since p is constant then
U D H  pV D 0.
Replacing (9.61) in (9.59) and using (9.60), one obtains the following expression
for the Gibbs free energy of an ideal solution

X
c
Ni
GD Ni fgi C RT ln g; (9.62)
iD1
N

which is the fundamental relation in the Gibbs representation. From it we found the
entropy of the ideal solution S D @G=@T, given by
X
c
Ni
SD Ni fsi  R ln g; (9.63)
iD1
N

where si D @gi =@T is the molar entropy of the i-th pure substance.
Since the enthalpy, the volume and the internal energy do not vary in the process
of formation of an ideal solution then

X
c X
c X
c
HD Ni hi ; VD Ni vi ; UD Ni ui ; (9.64)
iD1 iD1 iD1

where hi D gi CTsi , vi D @gi =@p and ui D hi pvi are the molar enthalpy, the molar
volume and the molar energy of the i-th component in the pure state. The chemical
potential i D @G=@Ni of the i-th component is given by

Ni
i D gi C RT ln : (9.65)
N

The expressions for the molar Gibbs free energy g D G=N and for the molar
entropy s D S=N are given by

X
c
gD xi fgi C RT ln xi g; (9.66)
iD1

X
c
sD xi fsi  R ln xi g; (9.67)
iD1
9.5 General Solutions 165

and for the enthalpy, volume and molar energy by

X
c X
c X
c
hD xi h i ; vD xi vi ; uD xi u i : (9.68)
iD1 iD1 iD1

The i-th chemical potential is given by

i D gi C RT ln xi : (9.69)

Remember that the mole fractions xi are not independent but are linked by (9.9).
From (9.69), we see that the activity zi D e i =RT for the i-th component is related
with the mole fraction by means of

zi D zi xi ; (9.70)

where zi D egi =RT is the activity of the pure component. Therefore, for ideal
solutions, if any component is considered to be a solvent (xi close to unit), Raoult
law is fulfilled. If on the other hand one component is considered to be a solute (xi
close to zero), Henry law is satisfied with i D zi .
In general, we do not expect real solutions to behave as ideal solutions. However,
certain molecular solutions behaves approximately as ideal solutions. In particular
a mixture of ideal gases is an ideal solution. It suffices to compare the equations
for the Gibbs free energies of an ideal gas mixture and of an ideal mixture, given
by (9.21) and (9.62), respectively.

Hildebrand Regular Solutions

A Hildebrand regular solution is defined as one whose change in entropy is the same
as that of an ideal solution. This means that when mixing several pure substances
at the same temperature and pressure, the entropy change S is given by (9.60).
However, unlike what happens in an ideal solution, the change in enthalpy H
in a Hildebrand regular solution does not vanish. Since S is independent of
temperature, it follows that H should also be independent of temperature. Indeed,
deriving H D G C TS with respect to temperature, we get

@ @ @
H D G C S C T S D 0; (9.71)
@T @T @T

because S is independent on T and S D @G=@T.

If we demand that the mixture is formed without change of volume, then H
should be independent of pressure. In fact, deriving H D G C TS with respect
166 9 Mixtures

to pressure,

@ @ @
H D G C T S D V; (9.72)
@p @p @p

since S does not depend on p and V D @G=@p. Therefore, if V D 0, then H

will be independent of pressure. In this case, H D U, since H D U C pV.
Replacing G D H  TS in expression (9.59) and using the notation e GD
H, then the fundamental relation G.T; p; N1 ; N2 ; : : : ; Nc / of a Hildebrand regular
mixture has the form

X
c X
c
Ni
GDe
GC Ni gi C RT Ni ln ; (9.73)
iD1 iD1
N

where eG depends on the numbers of moles of each component, can depend on

pressure, but not on the temperature, as shown in (9.71).
As the function e
G.p; N1 ; N2 ; : : : ; Nc / is extensive, we may write

e
G.p; N1 ; N2 ; : : : ; Nc / D N gQ .p; x1 ; x2 ; : : : ; xc /; (9.74)
P
where xi D Ni =N are the mole fractions and N D i Ni . Therefore the molar Gibbs
free energy g D G=N of a Hildebrand regular solution is given by

X
c
g D gQ C xi fgi C RT ln xi g; (9.75)
iD1

where gQ does not depend on T.

Excess Quantities

In the study of real solutions it is convenient to use as a reference system an ideal

solution with the same composition. Thus, we introduce the thermodynamic excess
quantities, defined as the differences of the thermodynamic quantities from the ideal
forms. The excess Gibbs free energy GE is defined by

X
c
Ni
G D GE C Ni fgi C RT ln g: (9.76)
iD1
N

Clearly, for an ideal solution, GE D 0. Comparing this expression with (9.73), we

see that the Hildebrand regular solution is the one such that GE does not depend on
temperature but only on pressure and number of moles.
9.5 General Solutions 167

Other excess quantities are defined in an analogous way and can be obtained
directly form GE . Thus, the excess entropy is given by SE D @GE =@T and the
excess enthalpy by H E D GE C TSE . The excess chemical potentials Ei D @GE =@Ni
are related to the chemical potentials i by

Ni
i D Ei C gi C RT ln : (9.77)
N

The excess Gibbs free energy GE is related to the variation of the Gibbs free
energy in the formation of a mixture G by

X
c
Ni
G D GE C Ni RT ln : (9.78)
iD1
N

From this relation we conclude that H E D H, that is, the excess enthalpy coincides
with the variation in enthalpy in the formation of a mixture.
Taking into account that GE .T; p; N1 ; N2 ; : : : ; Nc / is an extensive function we
may write

GE .T; p; N1 ; N2 ; : : : ; Nc / D NgE .T; p; x1 ; x2 ; : : : ; xc /: (9.79)

Therefore, the molar Gibbs free energy g D G=N of any solution can be written in
the form

X
c
g D gE C xi fgi C RT ln xi g: (9.80)
iD1

The chemical potentials are given by

i D Ei C gi C RT ln xi : (9.81)

Substituting (9.80) and (9.81) into (9.10), we conclude that

X
c
gE D Ei xi : (9.82)
i

Replacing (9.81) in the Gibbs-Duhem equation in the form given by (9.11), we

obtain the Gibbs-Duhem-Margules equation

X
c
xi d Ei D 0; (9.83)
iD1

valid at T and p constant.

168 9 Mixtures

In the limit of dilute mixtures, the excess quantities must have certain properties
that we will examine. Comparing the expression (9.81) with the expression of the
chemical potential of the solvent, given by (9.40), and with the chemical potentials
of the solutes, given by (9.38), we see that in the limit xi ! 1, the excess chemical
potential Ei vanishes and the others, Ej , approaches ıj  gj . Using these results
in (9.82), we conclude that gE also vanish in this limit.
The knowledge of specific forms of the excess Gibbs free energy gE allows the
determination of the intrinsic properties of the mixtures, as long as we know the
properties of the pure substances. In fact, if gE and gi were known it is possible to
recover g by (9.80). To set up gE we should take into account that gE ! 0 when
xi ! 1 for any component i. The simplest form of gE that meets this property is

1X
gE D wij xi xj ; (9.84)
2 i;j

where wij depends only on T and p, and wij D wji and wii D 0.
We remark that gE can be a singular function of the mole fractions. For example,
in the vicinity of a critical point, we expect that the thermodynamic quantities have
singular behavior. In particular, the excess Gibbs free energy must by a nonanalytic
at the critical point. However, in most cases the systems are not found inside the
critical region and can therefore be described by functions gE that are analytic in xi ,
as is the case of that given by (9.84), which we call regular forms. The mixtures that
are described by such functions are called regular mixtures. Although the regular
forms for gE do not provide a precise description at the critical region, they allows
us, on the other hand, a qualitative description of this region.

Problems

9.1 Show that if G.T; p; N1 ; N2 ; : : : ; Nc / is an extensive function, that is, homoge-

neous function of first order in N1 ; N2 ; : : : ; Nc , then

G D 1 N1 C 2 N2 C : : : C c Nc :

9.2 From H D G C TS, V D @G=@p and U D H  pV, and using the

expressions (9.21) and (9.23) for G and S, show that for an ideal gas mixture the
enthalpy, the volume and the internal energy are given by

X
c X
c X
c
HD Ni hi ; VD Ni vi ; UD Ni ui ;
iD1 iD1 iD1

where hi is the molar enthalpy, vi is the molar volume and ui is the molar energy
of the i-th component, alone. The enthalpy, the volume and the energy of a mixture
9.5 General Solutions 169

of ideal gases are therefore equal to the sum of the enthalpies, the volumes and the
energy of the gases isolated, at the same temperature and pressure of the mixture.
Or, equivalently, the enthalpy, the volume and the internal energy do not change
when ideal gases are mixed at constant temperature and pressure.
9.3 Suppose that a vessel of volume V has r partitions separated by rigid and
diathermal walls containing distinct gases. The i-th partition has volume Vi and
contains Ni moles of a gas of type i. Determine the change in the Helmholtz free
energy and of the entropy after the removal of the walls. Notice that initially the
gases are at the same temperature, since the walls are diathermal, but not necessarily
at the same pressure. Repeat the problem for the case in which all partitions have
the same type of gas.
9.4 Derive a formula similar to the van ’t Hoff one (9.57) for the elevation of the
boiling point of a liquid by the addition of non-volatile solutes. Taking into account
that the vapor contains only the solvent as its sole component, the chemical potential
of the solvent in the gaseous phase is independent of the mole fractions of the
solutes. Show that how it can be used to determine the ratio between the molecular
mass of the solvent and of the solute.
9.5 Show that a regular solution fulfills the Raoult and Henry laws. Determine the
quantity i which appears in the Henry law.
Chapter 10
Binary Mixtures

10.1 Fundamental Relation

Substances that are Binary Mixtures

In this chapter we study mixtures of two components, called binary mixtures which
are quite common in nature, both the homogeneous and heterogeneous. Many
minerals, when not pure substances are homogeneous solid binary mixtures such
as olivine (magnesium silicate and iron silicate) and amalgam (solid solution of
mercury and silver). The manufactured alloys, such as bronze (copper and tin) and
brass (copper and zinc) are also solid binary mixtures. Alcohol is usually found in
association with water, forming a homogeneous binary mixture.
We focus our study on the phase equilibria of binary mixtures. We initially
examine the liquid-vapor coexistence of mixtures of liquids which are miscible in
all proportions. Such mixtures may or may not have an azeotropic point. Next, we
examine the solid-liquid transition of mixtures of substances that are completely
miscible in both the solid phase and liquid phase.
Later, we pass to the study of mixtures of partially miscible liquid. In this
case it is common the occurrence of the coexistence of two liquids of different
compositions. The variation in temperature or pressure may lead the system to a
critical point where both coexisting liquids become identical. We consider, next,
the mixtures of substances partially miscible in the solid phase but fully miscible
liquid phase. The partial solubility provides the appearance of two solid phases
in coexistence. If the pure solid substances have the same crystalline structure,
then the two solid solutions also have the same crystalline structure but different
compositions. In this case, their coexistence may disappear at a critical point at
which the two solid phases become equal.
It is still possible, the occurrence of the melting of the two solid solutions while in
coexistence. In this case we will be facing the equilibrium of three phases, two solid
phases and the liquid phase, and we will have the opportunity to examine systems

© Springer-Verlag Berlin Heidelberg 2017 171

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_10
172 10 Binary Mixtures

with eutectic point. Other types of phase equilibria may occur, but in this chapter,
we limit ourselves to the study of equilibrium types mentioned above.

Thermodynamic State

In the representation of internal energy, the thermodynamic state of a binary mixture

is defined by the extensive variables: entropy S, volume V, number of moles N1
of the first component and number of moles N2 of the second component. The
fundamental relation is given by the corresponding internal energy U.S; V; N1 ; N2 /,
whose differential is

dU D TdS  pdV C 1 dN1 C 2 dN2 ; (10.1)

where T is the temperature, p is the pressure, and 1 and 2 are the chemical
potentials of components 1 and 2, respectively. Other representations can be used in
the study of the binary mixtures, but the most convenient is the Gibbs representation.
The fundamental relation is given by the Gibbs free energy G.T; p; N1 ; N2 /, whose
differential is

dG D SdT C Vdp C 1 dN1 C 2 dN2 ; (10.2)

where

@G @G @G @G
SD ; VD ; 1 D ; 2 D : (10.3)
@T @p @N1 @N2

The activities z1 and z2 of the components 1 and 2 are defined by

z1 D e 1 =RT and z2 D e 2 =RT : (10.4)

Being G extensive we may write

G D 1 N1 C 2 N2 ; (10.5)

which can be uses to recover the fundamental relation G.T; p; N1 ; N2 / in the case the
chemical potentials 1 .T; p; N1 ; N2 / and 2 .T; p; N1 ; N2 / are known. From (10.2)
and (10.5) we get the Gibbs-Duhem equation

 SdT C Vdp  N1 d 1  N2 d 2 D 0: (10.6)

10.1 Fundamental Relation 173

Molar Quantities

We will make frequent use of molar quantities such as the molar Gibbs free energy
g defined by g D G=N and the mole fractions x1 and x2 of the components, defined
by x1 D N1 =N and x2 D N2 =N, where N D N1 C N2 is the total number of moles of
the mixture. Note that the mole fractions are not independent since x1 C x2 D 1.
Dividing both members of (10.5) and (10.6) by N, they reduce to

g D 1 x1 C 2 x2 (10.7)

and

 sdT C vdp  x1 d 1  x2 d 2 D 0; (10.8)

where s D S=N is the molar entropy and v D V=N molar volume. From (10.7)
and (10.8) we conclude that

dg D sdT C vdp C 1 dx1 C 2 dx2 : (10.9)

Notice however that the differentials dxi are not independent but are linked by

dx1 C dx2 D 0: (10.10)

Usually we eliminate dx1 in (10.9), using (10.10), which permits us to write

dg D sdT C vdp C . 2  1 /dx2 : (10.11)

Defining and x by

D 2  1 and x D x2 ; (10.12)

respectively, then, according to (10.11), we have

dg D sdT C vdp C dx; (10.13)

so that the molar Gibbs free energy g.T; p; x/ can be considered a function of
temperature T, pressure p and mole fraction of the second component x. The
equations of state are

@g @g @g
D ; sD ; vD : (10.14)
@x @T @p
174 10 Binary Mixtures

Given g.T; p; x/, the fundamental relation in the Gibbs representation can be
recovered by

G.T; p; N1 ; N2 / D Ng.T; p; x/; (10.15)

where x D N2 =.N1 C N2 /. From this equation we get the following relations,

1 D g  x (10.16)

and

2 D g C .1  x/ ; (10.17)

between the chemical potentials 1 D @G=@N1 and 2 D @G=@N2 of the

components of the mixture and the chemical potential . It suffices to take into
account that @x=@N1 D x=N and that 2 D @G=@N2 .
It is convenient to define the relative activity  of the component 2 as  D
z2 =.z1 C z2 /, which can be written in the form

1
D : (10.18)
1 C e =RT

10.2 Gibbs Free Energy

Dilute Mixtures

In the applications that we will consider below we will use specific forms for the
Gibbs free energy. In the preceding chapter we have seen some forms that are
applied to a mixture of ideal gases and to other systems called ideal and regular
solutions. Next we will write explicitly the various expressions for the case of a
binary mixture. We consider first the Gibbs free energy of a diluted mixture and after
we consider the ideal mixture and regular mixture forms. These expressions should
be understood as approximations to the free energy. Although approximate they
adequately describe it in many cases, the properties of the mixtures, in particular the
phase coexistence.
The molar Gibbs free energy of a dilute mixture was presented in the previous
section and reduces, for the present case of two components, to the following
expression

g D g1 .1  x/ C ı2 x C RT.1  x/ ln.1  x/ C RTx ln x; (10.19)

10.2 Gibbs Free Energy 175

where g1 is the molar Gibbs free energy of the pure solvent, the component 1 of the
mixture, and ı2 depends only on the temperature and pressure. The mole fraction x
is related to the component 2, the solute. The expression above is valid for x << 1.
The chemical potential D @g=@x is given by

D g1 C ı2  RT ln.1  x/ C RT ln x: (10.20)

From and g, we obtain the chemical potential of the solvent 1 D g  x ,

given by

1 D g1 C RT ln.1  x/; (10.21)

which can also be written in the form

1 D g1  RTx (10.22)

taking into account that x << 1. The chemical potential of the solute, 2 D C 1 ,
is given by

2 D ı2 C RT ln x: (10.23)

The activities of the solvent z1 and solute z2 are obtained by exponentiation of both
sides of (10.21) and (10.23) and are therefore given by

z1 D zo1 .1  x/ and z2 D 2 x; (10.24)

where zo1 D expfg1 =RTg is the activity of the pure solvent and 2 D expf ı2 =RTg.
These two expressions are the Raoult and Henry laws, respectively.

Ideal Mixtures

The expression for the molar Gibbs free energy g of an ideal mixture of several
components, seen in the previous chapter, is reduced to the following expression for
the case of two components:

g D xg2 C .1  x/g1 C xRT ln x C .1  x/RT ln.1  x/: (10.25)

where g1 and g2 are the molar Gibbs free energies of the pure substances 1 and 2,
respectively, and x is the mole fraction of component 2.
The molar entropy @g=@T is given by

s D xs2 C .1  x/s1  xR ln x  .1  x/R ln.1  x/; (10.26)

176 10 Binary Mixtures

where s1 D @g1 =@T and s2 D @g2 =@T are the molar entropies of pure substances 1
and 2, respectively. The chemical potential is given by
x
D g2  g1 C RT ln : (10.27)
1x
while the chemical potentials 1 and 2 of each of the components are given by

1 D g1 C RT ln.1  x/ (10.28)

and

2 D g2 C RT ln x: (10.29)

If the i-th pure substance is an ideal gas then the corresponding free energy
gi .T; p/ is known and is given by

T p
gi .T; p/ D bi T  .ci C R/T ln C RT ln ; (10.30)
T0i p0i

where bi , ci , T0i and p0i are constants that characterize the i-th ideal gas. The
corresponding molar entropy si D @gi =@T is

T p
si .T; p/ D ai C ci C .ci C R/ ln  R ln : (10.31)
T0i p0i

Regular Mixtures

The molar Gibbs free energy g.T; p; x/ of a mixture of two components is given by

g D gE C .1  x/g1 C xg2 C RT.1  x/ ln.1  x/ C RTx ln x; (10.32)

where gE .T; p; x/ is the molar excess Gibbs free energy, that depends on temper-
ature, pressure and mole fraction x of component 2. From it we get the chemical
potential D @g=@x, given by
x
D E  g1 C g2 C RT ln ; (10.33)
1x

where E D @gE =@x. The chemical potentials 1 and 2 of the components are
obtained by the relations (10.16) and (10.17) and are

1 D E1 C g1 C RT ln.1  x/ (10.34)
10.2 Gibbs Free Energy 177

and

2 D E2 C g2 C RT ln x; (10.35)

where the excess chemical potential are given by

@gE
E1 D gE  x E D gE  x (10.36)
@x
and

@gE
E2 D gE C .1  x/ E D gE C .1  x/ : (10.37)
@x
In order to describe the intrinsic properties of binary solutions, we must seek
specific forms of gE . A large class of binary mixtures is well described by the excess
Gibbs free energies gE that are analytic functions of x, which we call regular forms.
The mixtures are described by such regular functions are called regular mixtures.
These forms must be constructed so that gE ! 0 when x ! 0 and when x ! 1,
since, taking into account that the function gE describes the deviations of the real
solution from the ideal behavior, it must vanish when the solution is reduced to
a pure substance. We assume, therefore, that gE is the product of x.1  x/ with a
function of x which is finite at x D 0 and x D 1. Under these conditions the simplest
regular form is

gE D wx.1  x/; (10.38)

where w depends only on T and p. Therefore,

g D wx.1  x/ C .1  x/g1 C xg2 C RT.1  x/ ln.1  x/ C RTx ln x: (10.39)

From gE given by (10.38) we obtain the simple regular equations

E1 D wx2 and E2 D w.1  x/2 ; (10.40)

so that the chemical potentials 1 and 2 of each component are

1 D wx2 C g1 C RT ln.1  x/ (10.41)

and

2 D w.1  x/2 C g2 C RT ln x: (10.42)

178 10 Binary Mixtures

A regular form particularly useful is

gE D x.1  x/fa C b.1  2x/ C c.1  2x/2 g; (10.43)

which leads as to the Margules equations:

E1 D x2 f.a C b/ C 2.b C c/.1  2x/ C 3c.1  2x/2 g (10.44)

and

E2 D .1  x/2 f.a  b/ C 2.b  c/.1  2x/ C 3c.1  2x/2 g: (10.45)

They are reduced to the simple regular equations when b D c D 0.

Another regular form is

x.1  x/
gE D ; (10.46)
b.1  x/ C ax

which yields the van Laar equations

ax2
E1 D (10.47)
Œb.1  x/ C ax2

and

b.1  x/2
E2 D : (10.48)
Œb.1  x/ C ax2

These equations are reduced to the simple regular equations when b D a.

10.3 Phase Transition

Two Phase Coexistence

The coexistence of two phases in a binary mixture is described by a linear segment

of g.T; p; x/ as a function of x, as shown in Fig. 10.1. To show this result, let us
consider a process in which the concentration of a mixture, which displays two
coexisting phases, are changed, keeping the temperature and pressure constant.
Under these conditions, one phase grows over the other which decreases, changing
the proportion of each coexisting phase. However, the mole fractions of the
components in each phase remain unchanged.
Denoting by NA and NB the number of moles of each phase, then the fraction in
number of moles of phase A is rA D NA =N and of phase B is rB D NB =N, where
10.3 Phase Transition 179

(a) (b)
g μ

μ*
gA

g
B

0 xA xB x 1 0 xA xB x 1

Fig. 10.1 (a) Molar Gibbs free energy g as a function of the mole fraction x for T and p constant.
(b) Chemical potential D @g=@x as a function of the mole fraction x for T and p constant

N is the total number of moles. Denoting by xA and by xB the mole fractions of

component 2 in phases A and B, respectively, then the number of mole of component
2 in phase A is xA NA and in phase B is xB NB , so that the total number of moles
of component 2 in the mixture is N2 D xA NA C xB NB . But the mole fraction of
component 2 in the mixture is x D N2 =N and therefore

x D xA rA C xB rB ; (10.49)

which is the lever rule. Since rA C rB D 1, we obtain from this rule the relations
x  xB xA  x
rA D and rB D : (10.50)
xA  xB xA  xB

Since xA and xB remain unchanged, the same happens with the molar Gibbs free
energy gA D g.T; p; xA / and gB D g.T; p; xB / of each phase because the temperature
and pressure are constant. The molar Gibbs free energy g of the mixture is then

g D rA gA C rB gB : (10.51)

Substituting (10.50) in (10.51), we get

xB  x  x  xA 
gD g C g ; (10.52)
xB  xA A xB  xA B

and therefore g is linear in x in the range xA  x  xB as gA and gB do not depend
on x but only of T and p.
180 10 Binary Mixtures

In the range xA  x  xB , the chemical potential D @g=@x is constant and is

gB  gA
 D ; (10.53)
xB  xA

from which we get

gA   xA D gB   xB : (10.54)

Thermodynamic Equilibrium

In a system in thermodynamic equilibrium the thermodynamic fields should be

uniform. For a mixture of two components, the condition of thermodynamic
equilibrium means that the temperature T, the pressure p, the chemical potential 1
of component 1 and the chemical potential 2 of component 2 must have the same
value anywhere in the system. In particular, for two phases A and B in coexistence,
the condition of equilibrium is given by

TA D TB ; pA D pB ; 1A D 1B ; 2A D 2B : (10.55)

where TA , pA , 1A and 2A denote the fields in phase A and TB , pB , 1B and 2B

denote the fields in phase B.
Let us imagine that the Gibbs free energy of the thermodynamic phases A and
B are described by the functions gA .T; p; x/ an gB .T; p; x/, respectively. If these
functions are two distinct branches, then, the Gibbs free energy of the mixture
g.T; p; x/ is obtained by the construction of a double tangent between the two
branches, that is, by the construction of a linear segment that connect the two points
of tangency .xA ; gA / and .xB ; gB / and having slope  given by (10.53). Taking into
account that gA D gA .T; p; xA /, that gB D gA .T; p; xB /, and that

 D A .T; p; xA / D B .T; p; xB /; (10.56)

where A D @gA =@x and B D @gB =@x, then, (10.54) reads

gA .T; p; xA /   xA D gB .T; p; xB /   xB : (10.57)

Equations (10.56) and (10.57) are, therefore, the conditions of thermodynamic

equilibrium and can be use to determine xA and xB as functions of T and p.
Using (10.16), it is easy to see that (10.57) is equivalent to equation

1A .T; p; xA / D 1B .T; p; xB /: (10.58)

10.3 Phase Transition 181

Adding (10.56) and (10.57) and using (10.17), we see that the resulting equation is
equivalent to equation

2A .T; p; xA / D 2B .T; p; xB /: (10.59)

These two equations are the conditions of thermodynamic equilibrium stated above
and may equivalently be used in the place of (10.56) and (10.57) to determine xA
and xB .

Condensation and Boiling Lines

As an example of phase transition, we examine the liquid-vapor transition of

the water-methanol mixture whose diagram is shown in Fig. 10.2. The water and
methanol are miscible in all proportions forming a liquid solution. Introducing heat
in a mixture of completely miscible liquid, at constant pressure, the temperature
increases and the mixture passes to the gaseous state. However, unlike what happens
with a pure substance, the transition does not occur in a single temperature, but a
range of temperatures. The transition begins at the boiling temperature and ends at
the condensation temperature. These two temperatures depend on the mole fraction
x of one component, as can be seen in Fig. 10.2a. In this graph, the line that
marks the threshold of boiling is called boiling or bubble line and the line that
marks the threshold of condensation is called condensation or dew line. Similarly
if the temperature is maintained constant, the phase transition occurs between two
pressures, as can be seen in Fig. 10.2b.

(a) 110 (b) 0.4

100 L
G 0.3
co
nd line
90 ing
en boil
p (Mpa)

sa
θ ( C)

tio
nl
ine
o

80 0.2
e
boilin n lin
70 g line
en satio
c ond
0.1
L
60
G
50 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
water x methanol water x methanol

Fig. 10.2 Boiling and condensation lines for water-methanol mixture, where x is the mole fraction
of methanol. (a) Diagram temperature versus mole fraction, at the pressure of 1 atm. (b) Diagram
pressure versus mole fraction, at the temperature of 100 ı C. The horizontal segments are the tie
lines and represent the coexistence between the liquid (L) and gas (G) phases. Source: LB
182 10 Binary Mixtures

To understand the graphs in Fig. 10.2, we recall that the coexistence of two
phases, for T and p fixed, occurs at a well defined value of the chemical potential
 .T; p/. The two phases in coexistence have also well defined mole fractions
xL .T; p/ and xG .T; p/. In a graph of T versus x at constant pressure, xL .T; p/ and
xG .T; p/ constitute the boiling and condensation lines, respectively. The horizontal
line segments that connect two points on these lines are tie lines shown in Fig. 10.2a.
Along the tie line, not just the pressure p and temperature T are constant, but the
chemical potential as well.
If the mixture is in a state corresponding to a point on the tie line, the fraction in
number of moles rL of the liquid phase and the fraction in number of moles rG of
the vapor are given by
x  xG xL  x
rL D and rG D : (10.60)
xL  xG xL  xG

Therefore, the fraction in number of moles of one phase is proportional to the

distance between the point under consideration and the end point of the tie line
corresponding to the other phase.
Heating the solution at constant pressure until a temperature between the boiling
and condensation temperatures, the mixture will be found in a state of coexistence of
liquid and vapor whose compositions are distinct from one another and distinct from
the mixture. One phase is richer in one component than the other. In Fig. 10.2 we
can see that the vapor is rich in methanol, component 2, as the mole fraction xG of
the component 2 in the gas phase is greater than the mole fraction xL of component
2 in the vapor. The method of enrichment of one of the components of a solution by
boiling is called distillation.

Limit of Dilute Mixture

When x D 0 or x D 1, the boiling and condensation lines should meet as shown

in Fig. 10.2 because in both cases the mixture is reduced to a pure substance whose
transition liquid-vapor occurs at a single temperature. The behavior of these lines in
the vicinity of these points is obtained from results valid for dilute solutions because
around them the mixture becomes diluted. Around x D 0, the component 1 plays the
role of solvent and component 2, the solute; around x D 1, the roles are inverted.
For convenience, let us analyze the behavior of the lines around x D 0, in which
case x, the mole fraction of component 2, is very small. Furthermore, let us consider
the general case of a transition between a phase A and phase B. We assume that
phase A occurs at temperatures higher than phase B.
According to Raoult law, in the form given by (10.22), the chemical potential of
the solvent in dilute solutions is given by

A
1 D g1  RTxA ;
A
and B1 D gB1  RTxB ; (10.61)
10.3 Phase Transition 183

where gA B
1 and g1 are the molar Gibbs free energies of the pure solvent in phases
A and B, respectively, and depend only on temperature and pressure. According to
Henry law the activity of the solute of dilute solutions is given by

2 D 2 xA ;
zA zB2 D 2B xB ;
A
and (10.62)

where 2A and 2B depend only on temperature and pressure.

The condition of thermodynamic equilibrium means that the chemical potentials
of the solvent and the solute in each phase are identical. Therefore,

1  RTxA D g1  RTxB
gA B
(10.63)

or
g1
xA  xB D ; (10.64)
RT

where g1 D gA
1  g1 . And
B

2B xB D 2A xA : (10.65)

From these equations we get

g1 2B g1 2A

xA D and xB D ; (10.66)
RT 2B  2A RT 2B  2A

which are the equations describing the lines that determine the thresholds of the
phase transition. Note that if the pure substances are immiscible in one phase, the
results above yield the van’t Hoff formula. Suppose that the miscibility occurs only
at the phase A. Then, 2A D 0, which gives xA D g1 =RT from which we obtain
van’t Hoff formula.
Expanding g1 around the transition pressure p1 .T/ of the pure solvent and
taking into account that g1 D 0 when p D p1 then

xL D a.p  p1 / and xG D b.p  p1 /; (10.67)

where a and b depend only on temperature. Therefore the lines that determine the
thresholds of the transition in the diagram pressure versus mole fraction are linear
near x D 0 and have slopes of the same sign because a=b D 2A =2B > 0, as
seen in Fig. 10.2b. The same behavior holds in the diagram temperature versus mole
fraction, as seen in Fig. 10.2a, what can be verified by the expansion of g1 around
the transition temperature of the pure solvent.
184 10 Binary Mixtures

10.4 Completely Miscible Substances

Phase Coexistence

Here we examine liquid-vapor and solid-liquid transitions in binary mixtures of pure

substances which are miscible in all proportions. When mixed, such substances form
a single liquid solution or a single solid solution. Examples of pairs of completely
miscible liquids are water and methanol, water and ethanol, benzene and ethanol,
acetone and chloroform. Heating a liquid solution, it turns into vapor, which is also
a solution. Many pairs of pure solids also are miscible in all proportions. Examples
of solid solutions are: olivine (magnesium silicate and iron silicate), silicon and
germanium, silver and gold, bismuth and antimony. Such solid solutions become
liquid solutions at temperatures sufficiently high.
Let us suppose that, when isolated, the pure substances 1 and 2 undergo a
transition from phase A to phase B, along the line of coexistence p1 .T/ and p2 .T/,
respectively. The molar Gibbs free energy of the pure substances g1 .T; p/ and
g2 .T; p/ are considered to have two branches each corresponding to the two phases,
that is,

gi .T; p/ D
gAi .T; p/; phase A;
(10.68)
gBi .T; p/; phase B;

for i D 1; 2. The coexistence line pi .T/ for each isolated pure substance is
determined by imposing the equality gA i .T; p/ D gi .T; p/.
B

We imagine that the molar Gibbs free energy of the mixture g.T; p; x/, where x
represents the mole fraction of component 2, is constituted also by two branches
corresponding to the two phases, that is,

gA .T; p; x/; phase A;
g.T; p; x/ D (10.69)
gB .T; p; x/; phase B:

Let us assume that each phase can be approximated by a regular solution so that
the molar Gibbs free energies gA and gB of each phase are of the form (10.32).
The chemical potentials of the components 1 and 2 of the phase A are thus of the
form (10.34) and (10.35), that is,

A
1 .x/ D 1 .x/ C g1 C RT ln.1  x/
EA A
(10.70)

and

2 .x/ D 2 .x/ C g2 C RT ln x:
A EA A
(10.71)
10.4 Completely Miscible Substances 185

Similarly, the chemical potentials of the components 1 and 2 of phase B are also
of the form (10.34) and (10.35), that is,

B1 .x/ D EB
1 .x/ C g1 C RT ln.1  x/
B
(10.72)

and

B2 .x/ D EB
2 .x/ C g2 C RT ln x;
B
(10.73)

where EAi and i are the excess chemical potentials of the components in phases
EB

A and B, respectively.
If there is phase coexistence, the mole fraction of component 2 in phase A will
be denoted by xA and in phase B by xB . According to (10.58) and (10.59), the
coexistence of phases is obtained by imposing the equality of the chemical potentials
A
i .xA / D i .xB / or
B

1 .xA / C g1 C RT ln.1  xA / D 1 .xB / C g1 C RT ln.1  xB /

EA A EB B
(10.74)

and

2 .xA / C g2 C RT ln xA D 2 .xB / C g2 C RT ln xB ;
EA A EB B
(10.75)

which we write in the form

1  xB 1
ln D fg1 C EA
1 .xA /  1 .xB /g
EB
(10.76)
1  xA RT

and
xB 1
ln D fg2 C EA
2 .xA /  2 .xB /g;
EB
(10.77)
xA RT

where g1 D gA 1  g1 and g2 D g2  g2 . These two expressions permit to

B A B

determine xA and xB as functions of pressure and temperature and constitute the

lines that determine the thresholds of the transition between A and B.

Liquid-Vapor Equilibrium

To determine the lines that give the threshold of the transition in the diagram p
versus x or in the diagram T versus x, we must know how the free energies gi .T; p/
of the pure substances depend on T and p. We consider initially the liquid-vapor
transition. We assume that the gas corresponding to the pure substances are ideal
gases so that

i D RT ln p C i ;
gG (10.78)
186 10 Binary Mixtures

where i depends only on temperature and that the liquids corresponding to the pure
substances are described by

gLi D RT ln pi C i ; (10.79)

where pi is the vapor pressure of the i-th pure substance and depends only on
temperature.

From these expressions we have gi D gG i  gi D ln.p=pi / which replaced
L

in (10.76) and (10.77) yields

1  xL p 1 EL
ln D ln   .xL / (10.80)
1  xG p1 RT 1

and
xL p 1 EL
ln D ln   .xL /; (10.81)
xG p2 RT 2

where we have taken into account that the excess chemical potentials of the pure
substances in vapor phase vanish, EGi D 0, because the vapor is a mixture of ideal
gases. These equations describe the condensation and boiling lines in the diagram
pressure versus mole fraction.
They can also be written in the form

1 EL
p.1  xG / D p1 .1  xL / expf .xL /g (10.82)
RT 1
and
1 EL
pxG D p2 xL expf .xL /g: (10.83)
RT 2
Adding these two equations, we get p as a function of xL , which constitutes the
boiling line

1 EL 1
p D p1 .1  xL / expf 1 .xL /g C p2 xL expf EL .xL /g: (10.84)
RT RT 2
The condensation line, p as a function of xG , is obtained in implicit form from the
boiling line and using one of the two equations (10.82) or (10.83). Alternatively,
dividing both (10.82) and (10.83), we get

1 p 1 1
D 1 C 1 .  1/ expf ΠEL .xL /  EL
2 .xL /g; (10.85)
xG p 2 xL RT 1

which determined xG from xL .

10.4 Completely Miscible Substances 187

To determine the condensation and boiling lines in the diagram temperature

versus mole fraction we should explicit the dependence of the vapor pressure pi
of the pure substance with temperature. To this end we use the following equation

p `i 1 1
ln  D .  / (10.86)
pi R T Ti

where Ti is the transition temperature of the i-th pure substance, at pressure p. This
equation can be derived with the help of the Clausius-Clapeyron equation, assuming
that the boiling latent heat `i is a constant and that the volume of the vapor is much
greater than the volume of the liquid. Substituting this result in (10.80) and (10.81),
we get

1  xL `1 1 1 1 EL
ln D .  /  .xL / (10.87)
1  xG R T T1 RT 1

and

xL `2 1 1 1 EL
ln D .  /  .xL /; (10.88)
xG R T T2 RT 2

If the liquid phase is an ideal solution, the excess chemical potentials of the
substances in this phase are also zero, EL
i D 0, so that (10.80) and (10.81) become

1  xL p xL p
D  and D ; (10.89)
1  xG p1 xG p2

and can be solved for xL and xG , with the results

p  p1 p2 p  p1

xL D and xG D : (10.90)
p2  p1 p p2  p1

The boiling and condensation lines are thus

p D p2 xL C p1 .1  xL / (10.91)

and
1 xG 1  xG
D C : (10.92)
p p2 p1

We see that p is linear in x along the boiling line and that 1=p is linear in x along the
condensation line. The two properties are observed experimentally, for example,
in the system heptane-hexane, as can be seen in Fig. 10.3a, which can then be
considered as an ideal mixture.
188 10 Binary Mixtures

(a) 25 (b) 95

20 90 G
L
p (kPa)

T (K)
15 85

10 80
G L

5 75
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
heptane x hexane O2 x N2

Fig. 10.3 (a) Liquid-vapor transition of the system heptane-hexane, at the temperature of 30 ı C.
Diagram p versus mole fraction x of hexane. Source: LB. (b) Liquid-vapor transition of the
system oxygen-nitrogen, at the pressure of 1 atm. Diagram temperature T versus mole fraction
x of nitrogen. The circles represent the experimental data obtained by Dodge and Dunbar [130]

The ideal behavior displayed by the system heptane-hexane is also presented

by other pairs of similar pure substances. However, in general, we expect deviations
from the ideal behavior, that can be relevant such as in azeotropes or small such as in
the oxygen-nitrogen system shown if Fig. 10.3b. These deviations can be described
by means of the regular forms for the excess quantities. Using (10.87) and (10.88),
we can describe the condensation and boiling lines assuming that the liquid solution
is a simple regular solution, that is, such that
2 2
EL
1 D wxL and EL
2 D w.1  xL / : (10.93)

Using the oxygen and nitrogen latent heats, `1 =R D 820 K and `2 =RT D 670 K,
respectively, and the value w=R D 12 K, we get the continuous line shown in
Fig. 10.3b.

Azeotropy

The results obtained so far show boiling and condensation lines that are strictly
monotonic, as seen in Figs. 10.2 and 10.3. However, these lines may have a
maximum or a minimum as seen in Fig. 10.4. At this point, called azeotropic, the
boiling and condensation lines meet so that in the azeotropic point xL D xG D xaz .
A binary mixture, with a composition corresponding to the azeotropic point, passes
from the liquid to the vapor phase, at constant pressure, at a unique temperature
and in this sense it resembles a pure substance. However, an azeotropic mixture
distinguishes from a pure substance because the composition of the azeotropic point
10.4 Completely Miscible Substances 189

(a) 85 (b) 22

20
L
80
G 18

p (kPa)
q ( C)

75
16
o

70 14

12 G
65
L 10

60 8
0 0.2 0,.4 0,.6 0,.8 1 0 0.2 0.4 0.6 0.8 1
benzene x ethanol benzene x methanol

Fig. 10.4 (a) Liquid-vapor transition for the system benzene-ethanol in the diagram temperature
versus mole fraction x of ethanol, at the pressure of 1 atm. The azeotropic point occurs at xaz D
0:45. (b) Liquid-vapor transition for the system benzene-methanol in the diagram pressure versus
mole fraction x of methanol, at the temperature of 20 ı C. The azeotropic point occurs at xaz D 0:52.
Source: LB

depends on pressure. The properties of mixtures related to the azeotropic point are
known as Gibbs-Konovalov laws.
There are many pairs of liquids that displays an azeotropic point including
benzene-ethanol and benzene-methanol, whose diagrams are shown in Fig. 10.4,
chloroform-acetone, water-ethanol, etc. At the pressure of 1 atm, the azeotropic
point of the mixture water-ethanol occurs when the mole fraction of ethanol xaz D
0:894. Decreasing the pressure, the azeotropic point of the system water-ethanol
increases, reaching the value xaz D 0:98 at the pressure of 0.13 atm. An important
consequence of the existence of the azeotropic point is that the distillation process
stops working when one reaches this point. A mixture water-ethanol that is poor in
ethanol may become richer in ethanol by distillation until the concentration reaches
the azeotropic point.
At the azeotropic point the coexistence line p versus , at constant T, reaches
a maximum or a minimum and in both cases dp=d D 0. From the Clausius-
Clapeyron equation

dp xG  xL
D ; (10.94)
d vG  vL

valid at constant temperature, we see that .xG  xL / ! 0 at the azeotropic point, that
is, the length of the tie lines vanishes.
To determine the azeotropic point it suffices to impose xG D xL D xaz in (10.85)
since at this point xG and xL are equal. We get then

p2
EL
1 .xaz /  2 .xaz / D RT ln
EL
: (10.95)
p1
190 10 Binary Mixtures

For the simple regular form E1 D wx2 and muE2 D w.1  x/2 we get the following
result for the azeotropic point

1 RT p1
xaz D  ln : (10.96)
2 2w p2

The azeotropic point exists if the right hand side of this equation is between 0 and
1. The corresponding pressure is given by
q w 1 RT p
paz D p1 p2 expf C .ln 1 /2 g: (10.97)
RT 4 4w p2

Other regular forms can also predict azeotropic points. The continuous lines shown
in Fig. 10.4b were determined from (10.84) and (10.85) and by using the Margules
regular form given by (10.44) and (10.45), with the following parameters: a=RT D
2:04, b=RT D 0:28 and c=RT D 0:30, and T D 293:15 K.

Solid-Liquid Equilibrium

We examine now the solid-liquid equilibrium of substances that are completely

miscible both in the solid and liquid phases or, equivalently, the coexistence of a
solid solution and a liquid solution. Complete miscibility occurs, for example, in
mixtures of substances which in the solid phase has the same crystalline structure.
The solid solutions can be of type substitutional or interstitial. In the first type, the
atoms of one species are replaced by atoms of the other species in the vertices of
the crystalline structure, which remains unchanged. In the second type, the atoms
of the solute are inserted in the interstices between the atoms of the solvent. The
complete solubility can occur only on the substitutional type. Examples of binary
mixtures whose components have the same crystalline structure and are miscible in
all proportions include: silver and gold, gold and copper (above 410 ı C), gold and
nickel (above 810 ı C), gold an platinum (above 1260 ı C), bismuth and antimony,
cadmium and magnesium (above 253 ı C), chromium and molybdenum (above
880 ı C), copper and nickel, copper and platinum (above 816 ı C), germanium and
silicon, selenium and tellurium, magnesium silicate and iron silicate (olivine), and
Al2 O3 and Cr2 O3 .
Raising the temperature of these systems, they melt giving rise to a liquid
solution, which coexist with the solid solution. We denote by xS and xL the mole
fractions of component 2 in the solid and liquid phases, respectively. The mole
fractions xS and xL as functions of temperature or pressure constitute the solid line
(solidus) and the liquid line (liquidus). To determine them, we start from (10.76)
and (10.77), that is,

1  xS 1
ln D fg1 C EL
1 .xL /  1 .xS /g
ES
(10.98)
1  xL RT
10.4 Completely Miscible Substances 191

and
xS 1
ln D fg2 C EL
2 .xL /  2 .xS /g;
ES
(10.99)
xL RT

where g1 D gL1  gS1 and g2 D gL2  gS2 . These two expressions allows the
determination of the solidus and liquidus. The lines can be obtained as long as g1
and g2 are known as functions of temperature and pressure. Let us consider p
constant and use the approximation

`1  `1 
g1 D .T  T/ and g2 D .T  T/; (10.100)
T1 1 T2 2

where T1 and T2 are the melting temperatures and `1 and `2 are the melting latent
heat of the pure substances. Substituting these expression in the previous equations,
we get

1  xS `1 1 1 1
ln D .  /C f EL .xL /  ES
1 .xS /g (10.101)
1  xL R T T1 RT 1

and

xS `2 1 1 1
ln D .  /C f EL .xL /  ES
2 .xS /g; (10.102)
xL R T T2 RT 2

which permit determine the liquidus and solidus in the diagram temperature versus
mole fraction.
As an example of the solid-liquid transition we show in Fig. 10.5 the diagram T
versus x of the system germanium-silicon. Germanium and silicon are completely
miscible both in the liquid and solid phases. Both have the same type of crystalline
structure and their atoms similar sizes and the solid solution is of the type

Fig. 10.5 Solid-liquid 1600

transition for the system
germanium-silicon in the
diagram temperature versus 1400 L
mole fraction x of
θ ( C)

germanium, at the pressure of

o

1 atm. The circles represent 1200

experimental data obtained by
Stöhr and Klemm [132]
1000
S

800
0 0.2 0.4 0.6 0.8 1
Si x Ge
192 10 Binary Mixtures

substitutional. The continuous line over the experimental data of Fig. 10.5 were
obtained by (10.101) and (10.102) using the simple regular forms for the chemical
potentials, that is,
2 2
EL
1 .xL / D wL xL and EL
2 .xL / D wL .1  xL / ; (10.103)

for the liquid solution, and

2 2
1 .xS / D wS xS
ES and ES
2 .xS / D wS .1  xS / ; (10.104)

for the solid solution. The parameters used were wL D 450 K and wL D 800 K, in
addition to the latent heats of the germanium and silicon, given by `1 =R D 1212 K
and `2 =R D 1681 K, respectively.

10.5 Partially Miscible Substances

Coexistence of Two Liquid Phases

When phenol (C6 H6 O) is added to water at room temperature one obtains a

homogeneous solution, that is, a single thermodynamic phase, as long as the mass
of phenol is not above 8 %. Above this value, phenol ceases to dissolve completely
in water causing the appearance of another liquid phase, richer in phenol. Since the
densities of the phases are distinct, the phase with lower density stays above that of
higher density, separated by a meniscus. Adding more phenol, the two phase remain
in coexistence until the mass of phenol in the mixture reaches 72 %. Above this
value, the system returns to have single phase. If the mass of phenol is between 8
and 72 %, the system water-phenol will be, therefore, heterogeneous and composed
by two liquid phases in coexistence. Changing the temperature, the range of phase
coexistence changes, as shown in Fig. 10.6. For any concentration of phenol in the
mixture, the two liquid phases in coexistence have each one a well defined mass
fraction of phenol which, for the room temperature, is 0.08 for the phenol poor
phase and 0.72 for the phenol rich phase.
To characterize the quantity of phenol in the mixture we will use the mole fraction
x of phenol in the place of the mass fraction y, considered as the component 2 of the
mixture. The relation between these two fractions is
y M1 x
D (10.105)
1y M2 1  x

where M1 and M2 are the molar mass of components 1 and 2, respectively. Let
us denote by xA and xB the mole fractions of component 2 in the phenol poor and
phenol rich phases, respectively. Thus, given a mole fraction x of phenol, it is related
to the fractions of phenol in number of moles in the phenol poor phase, rA , and in
10.5 Partially Miscible Substances 193

the phenol rich phase, rB D 1  rA , by the lever rule

x D rA xA C rB xB ; (10.106)

from which we get

xB  x x  xA
rA D and rB D : (10.107)
xB  xA xB  xA

The mole fractions xA and xB of the two solutions in coexistence vary with
temperature, describing the miscibility curve, or solubility, as seen in Fig. 10.6.
Raising the temperature of the system water-phenol, from 20 ı C, xA increases and
xB decreases, so that the compositions of the two solutions become closer and, at a
certain temperature, become equal. Above this temperature, called critical solution
temperature, of consolute temperature, the two solutions become identical and the
system becomes homogeneous. In other words, the two components of the mixture
becomes completely miscible. At the critical temperature, the two mole fractions xA
and xB have the same value xc called critical mole fraction.
When the partial miscibility occurs at temperatures lower than the complete
miscibility, as happens to the system water-phenol, the critical temperature is
called upper consolute temperature. When the opposite occurs, it is called lower
consolute temperature, as is the case of the system methylamine-water which has
complete miscibility below 18.5 ı C and partial miscibility above this temperature.
Some binary mixtures, such as the system water-nicotine, have closed miscibility
curve showing upper and lower consolute temperatures. Between 61 and 208 ı C the
miscibility of the system water-nicotine is partial. Out of this range the miscibility
is total.

Fig. 10.6 Miscibility curve 70

for the system water-phenol,
at the pressure of 1 atm, 60
LA
where y is the mass fraction LB
of phenol. In the region inside
50
θ ( C)

the miscibility curve, the

phenol poor phase (LA ) and
o

the phenol rich phase (LB ) 40

coexist. Above the critical
point, which occurs at 30
y D 0:34 and  D 66 ı C,
water and phenol are
20
completely miscible. Source: 0 0.2 0.4 0.6 0.8
LB water y phenol
194 10 Binary Mixtures

Hildebrand-Heitler Theory

The coexistence of two liquid phases and the critical point in systems of the type
water-phenol can be described qualitatively by the theory of regular solutions. To
this end, we use the expression for the Gibbs free energy g.T; p; x/ given by (10.39),
that is,

g D wx.1  x/ C .1  x/g1 C xg2 C

C RT.1  x/ ln.1  x/ C RTx ln x; (10.108)

where g1 .T; p/ and g2 .T; P/ depend on T and p and w.p/ depends only on pressure.
We suppose that the pure substances do not undergo any transition, so that, unlike
of what we have seen before, the functions g1 and g2 do not describe any transition.
From the expression for g, we determine the chemical potential D @g=@x,
given by
x
D w.1  2x/ C g2  g1 C RT ln : (10.109)
1x
At high temperatures, g is a convex function of x. However at low temperatures,
g ceases to be convex and, in this case, we should consider the convex hull of g,
obtained from a Maxwell construction. Initially, we determine the critical point. To
this end, we calculate the derivatives of with respect to x, given by

@ RT
D 2w C (10.110)
@x x.1  x/

and

@2 2x  1
2
D RT 2 : (10.111)
@x x .1  x/2

The critical point is such that the two derivatives vanish which occur when xc D 1=2
and
w
RTc D : (10.112)
2
The mole fractions xA and xB , related to the two phases, are determined by the
equilibrium conditions

gA  A xA D gB  B xB (10.113)

and

A D B ; (10.114)
10.5 Partially Miscible Substances 195

which are equivalent to a Maxwell construction. In an explicit form

wx2A C RT ln.1  xA / D wx2B C RT ln.1  xB / (10.115)

and
xA xB
 2wxA C RT ln D 2wxB C RT ln ; (10.116)
1  xA 1  xB

equations that determine xA and xB as functions of T. It is easy to see that xB D 1xA

because this condition makes the two equations above equivalent and equivalent to
the equation
xA
w.1  2xA / C RT ln D 0: (10.117)
1  xA

Defining the auxiliary quantity x by x D xB  xA D 1  2xA and, taking into

account that w D 2RTc , we have

T 1 C x
x  ln D 0: (10.118)
2Tc 1  x

This equation can be solved numerically for x, from which we obtain the mole
fractions xB D .1 C x/=2 and xA D .1  x/=2, which describe the miscibility
curve. Around the critical point, x is small, what allows the expansion of the left
hand side of (10.118) in powers of x. Up to cubic terms,

T 1
.1  /x  .x/3 D 0; (10.119)
Tc 3

from which we get

s
Tc  T
x D 3 ; (10.120)
Tc

which gives the behavior of the miscibility curve around the critical point.
It is worth noticing that the quantity .@x=@ /Tp , which we call susceptibility,
diverges at the critical point along x D xc . Replacing x D xc D 1=2 in
expression (10.110), we see that, for T > Tc ,

@x 1
D : (10.121)
@ 4R.T  Tc /
196 10 Binary Mixtures

Using the result (10.120), we get the following expression for the susceptibility,
for T < Tc ,

@x 1
D : (10.122)
@ 8R.Tc  T/

Critical Point

Around the critical point the thermodynamic quantities behave singularly. We

assume that the singularities can be represented by power laws. According to this
assumption, the miscibility gap xB  xA vanishes at the critical point as

xB  xA  jTc  Tjˇ : (10.123)

Experimentally, one finds that the critical exponent ˇ has values that are close to
each other as seen in Table 10.1. The susceptibility .@x=@ /Tp diverges at the critical
point as

@x
 jT  Tc j ; (10.124)
@

Table 10.1 Critical exponents related to the critical point of the liquid-liquid transition of
incomplete miscible substances. The data refer to the pressure of 1 atm
Sistema ˛ ˇ  c (ı C)
Isobutyric acid + water 0:12 0:33 1:24 26.14
Nitroethane + isooctane 0:11 0:32 30.03
Polystyreno + cyclohexane 0:14 0:33 30.50
Nitroethane + 3-methylpentane 0:14 0:34 26.47
Methanol + cyclohexane 0:10 0:33 49.1
Phenol + water 0:33 1:32 65.87
Triethylamina + water 0:11 18.5
Carbon tetrachloride
+ perfluoro methyl cyclohexane 0:33 28.64
Aniline + cyclohexane 0:33 31
Nitrobenzene + heptane 0:33 18
Carbon disulphide + nitromethane 0:32 63.5
Cyclohexane + anhydrous acetic 0:32 52.3
Gallium + mercury 0:34 203.3
Nitroethane + hexane 1:19
10.5 Partially Miscible Substances 197

along x D xc . The value of the exponents ˇ and  obtained from the Hildebrand-
Heitler theory, ˇ D 1=2 and  D 1, are distinct from the experimental values
shown in Table 10.1, what means that this theory does not give a proper quantitative
description if we are very close to the critical point although it gives a qualitative
description.
A quantity that can be directly measured in binary liquid mixture is the molar
heat capacity cpx , at x and p constant, defined by cpx D T.@s=@T/px . Around the
critical point this quantity behaves according to

cpx  jT  Tc j˛ ; (10.125)

along x D xc . Some experimental values of ˛ are presented in Table 10.1. The

isothermal compressibility at x constant and the coefficient of thermal expansion at
x constant also diverge with the same exponent ˛, along x D xc , that is,

1 @v
 . /Tx  jT  Tc j˛ ; (10.126)
v @p
1 @v
. /px  jT  Tc j˛ : (10.127)
v @T

Coexistence of Two Solid Phases

The mixture of solid substances may result in both homogeneous or heterogeneous

mixtures. Crystalline solids with the same structure can be miscible in all propor-
tions, as happens with the system germanium-silicon, but may also give rise to
solid mixtures with partial solubility, as happens to the system gold-platinum below
1260 ı C or with the system silver-copper. If the structures are distinct, the solids
will have partial solubility or will be immiscible. The alloy copper-zinc (brass) is
homogeneous if the mole fraction of zinc is lower than 0.32. Above this value and
until 0.48, the alloy consists of two solutions in coexistence. Copper and tin are
practically immiscible at room temperature, comprising an alloy, the bronze, with
two solid solutions in coexistence.
Below 1260 ı C, the system gold-platinum may present coexistence of two solid
solutions, as seen in Fig. 10.7. One rich in platinum (phase SA ), with mole fraction
xA , and the other rich in gold (phase SB ), with mole fraction xB > xA . The mole
fractions xA and xB as functions of the temperature constitute the solubility curve
(solvus) that delimits the coexistence region from the two solid solutions. If a
fraction rA in number of moles of the solid solution SA is in equilibrium with a
fraction rB in number of moles of the solution SB , then the system is represented by
a point .x; T/ inside the region of coexistence such that

x D rA xA C rB xB : (10.128)
198 10 Binary Mixtures

Fig. 10.7 Diagram 1800

temperature versus mole
fraction of gold for the 1600 L
system gold-platinum. The
mixture presents a liquid

θ ( C)
1400
phase (L) and two solid

o
phases (SA and SB ). The
critical point between the 1200 SA
solid phases occur at
x D 0:39 and  D 1260 ı C. 1000 SB
The circles represent
experimental data obtained by
800
Darling et al. [129] 0 0.2 0.4 0.6 0.8 1
Pt x Au

Equivalently
x  xB xA  x
rA D and rB D : (10.129)
xA  xB xA  xB

Raising the temperature of the system gold-platinum with fixed composition and
equal to x D xc D 0:39, the two solutions in coexistence become identical at the
critical temperature Tc D 1260 ı C, that is, xA  xB ! 0 when T ! Tc . Above this
temperature, gold and platinum become miscible in all proportions.

Eutectic Point

Suppose that the system gold-platinum, with fixed composition, is cooled slowly
from a state corresponding to a point just above the solubility curve until a final state
corresponding to a point below this curve. In the initial state, the system is a single
solid solution. Crossing the solubility curve, the system decomposes into two solid
solutions forming an heterogeneous system. The process of cooling, represented in
the T versus x diagram by a vertical straight line that crosses the solubility line, is
called spinodal decomposition.
If we make the inverse process, that is, we start from a state inside the
coexistence region and introduce heat, at fixed composition, the system will become
homogeneous as we pass the solubility curve. Continuing the process, the solid
solution begins to melt when it reaches the solid line, becoming a liquid solution
when it reaches the liquid line. We see that the system gold-platinum turns into a
single solid solution before it melts. However, there are binary systems for which
the two solid solutions in coexistence melt without becoming a single solution. In
this case we face a coexistence of three phase: the two solid solutions and the liquid
solution.
10.5 Partially Miscible Substances 199

1200

1000 L

SB

θ ( C)
800
SA

o 600

400

200
0 0.2 0.4 0.6 0.8 1
Ag x Cu

Fig. 10.8 Diagram temperature versus mole fraction of copper for the system silver-copper. The
system presents the liquid phase (L) and two solid phases (SA and SB ). The coexistence of the three
phases occurs at  D 780 ı C and the mole fractions of copper are 0.14 in phase SA , 0.40 in the
liquid phase and 0.95 in phase SB . The circle represent experimental data obtained by Murray [131]

The system silver-copper, whose diagram is presented in Fig. 10.8, shows the
coexistence of three phases at the temperature of Tt D 780 ı C, as long as the mole
fraction of copper is between xA D 0:14 and xB D 0:95. Below this temperature the
system displays coexistence between two solid solutions: one rich in silver (phase
SA ) and the other rich in copper (phase SB ). Above this temperature it presents the
coexistence of phase SA and the liquid phase, if x < xE , or the coexistence of phase
SB and the liquid phase, if x > xE , where xE D 0:40. The point .xE ; Tt / is called
eutectic point.
Suppose that heat is introduced in the system silver-copper from a state in which
the two solid phases are in coexistence and such that xA < x < xE , that is,
such that the mole fraction is smaller that the mole fraction of the eutectic point.
The temperature will increase and when reaching Tt the two solid phase begin to
melt. The temperature will remain invariant until the whole phase SB disappear,
remaining only the phase SA and the liquid phase. Just before the beginning of
melting, the fractions rA and rB in number of moles of the phases SA and SB are
given by (10.129). Just after the disappearing of the phase SB , the fractions rA0 and
rL0 in number of moles of phase SA and the liquid phase are given by

x  xE xA  x
rA0 D and rL0 D : (10.130)
xA  xE xA  xE

Notice that rA0 < rA . Similarly, if the initial state is such that xE < x < xB then
phase SA will disappear remaining only phase SB and the liquid phase. If on the
other hand the system is prepared so that it has exactly the composition of the
eutectic point, then both solid phases will disappear simultaneously remaining only
the liquid phase, that is, the system will pass directly to the liquid state, as shown in
Fig. 10.8. In this sense, the heterogeneous alloy with the composition of the eutectic
point behaves as a pure substance, melting entirely at a single temperature.
200 10 Binary Mixtures

Problems

10.1 Show that at the azeotropic point, the derivatives of pressure along the boiling
and condensation line vanish, that is, @p=@xL D 0 and @p=@xG D 0.
10.2 Determine the critical point of the liquid-liquid transition of partially miscible
substances described by the following Gibbs free energy

g D wx.1  x/ C cx2 .1  x/ C .1  x/g1 C xg2 C

CRT.1  x/ ln.1  x/ C RTx ln x;

were c is a constant that depends only on pressure. In this case the critical point
occurs at x D 1=2.
10.3 Show that, if the concentrations xA and xB solve (10.56) and (10.57), then they
also solve the same equations with g.x/ replaced by g.x/ C ax, for any value of a.
The chemical potential changes but not the concentrations.
Chapter 11
Phase Diagrams

11.1 Gibbs Phase Rule

Fields and Densities

We study in this chapter the phase diagrams of multicomponent systems. The

representation that provides the simplest diagrams is that composed only by ther-
modynamic fields. A system of c components is described by c C 2 thermodynamic
fields: the temperature T, the pressure p, and the chemical potentials 1 ; 2 ; : : : ; c
of the c components. However, these fields are not independent but are connected
by the Gibbs-Duhem equation

 SdT C Vdp  N1 d 1  N2 d 2      Nc d c D 0; (11.1)

where S is the entropy, V is the volume, and N1 ; N2 ; : : : ; Nc are the number of moles
of the c components. One of the fields should be chosen as dependent and will work
as a thermodynamic potential. The remaining c C 1 fields, considered independent,
comprises the thermodynamic space of dimension c C 1.
Any of the c C 2 fields can be chosen as the thermodynamic potential. If the
Gibbs-Duhem equation is written in the form

dp D sNdT C N1 d 1 C N2 d 2 C    C Nc d c ; (11.2)

where sN D S=V is entropy per unit volume and Ni D Ni =V is the number of moles
per unit volume, it becomes clear that the pressure can act as a thermodynamic
potential and the fields T; 1 ; 2 ; : : : ; c as independent variables.
As a second example, we show how 1 can be chosen as the thermodynamic
potential. To this end we write the Gibbs-Duhem equation in the form

 sdT C vdp  x1 d 1  x2 d 2      xc d c D 0; (11.3)

© Springer-Verlag Berlin Heidelberg 2017 201

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_11
202 11 Phase Diagrams

where s denotes the molar entropy, v the molar volume, and xi the mole fractions of
the c components, which are related by

x1 C x2 C    C xc D 1: (11.4)

Using (11.4) to eliminate x1 from (11.3), then

d 1 D sdT C vdp  x2 d N 2  x3 d N 3      xc d N c ; (11.5)

where N i D i  1 . The thermodynamic space associated to 1 , which works as a

thermodynamic potential, are composed by the variables T; p; N 2 ; N 3 ; : : : ; N c .
It is worth mentioning that this representation can be obtained by successive
Legendre transformations from the molar Gibbs free energy g.T; p; x2 ; : : : ; xc /,
which we used in the study of mixtures of pure substances and whose differential is
given by

dg D sdT C vdp C N 2 dx2 C N 3 dx3 C    C N c dxc : (11.6)

The molar Gibbs free energy of g in its turn is obtained by Legendre transformation
from the molar energy u.s; v; x2 ; : : : ; xc /, whose differential is given by

du D Tds  pdv C N 2 dx2 C N 3 dx3 C    C N c dxc : (11.7)

Similarly, we can choose other representations involving only thermodynamic

fields. Whatever the choice, the thermodynamic field space of a system of c
components is composed by c C 1 variables.
Let us denote, generally, the c C 1 thermodynamic fields chosen to compose
the thermodynamic space by h0 ; h1 ; h2 ; : : : ; hc and the respective thermodynamic
densities by 0 ; 1 ; 2 ; : : : ; c . These two spaces are called space-h and space-
, respectively. The fields are understood as the components of the vector hE D
.h0 ; h1 ; : : : ; hc /, belonging to space-h, and the densities as components of the vector
E D .0 ; 1 ; : : : ; c /, belonging to space-. Denoting the thermodynamic associated
to the fields by , then
X
d D  i dhi ; (11.8)
i

that is,

@
i D  : (11.9)
@hi

Incidentally, by comparing (11.2) with (11.8), we see that the potential associated
to the independent variables T; 1 ; 2 ; : : : ; c is p and not p. The potential
11.1 Gibbs Phase Rule 203

.h0 ; h1 ; h2 ; : : : ; hc / is linked, by successive Legendre transformation, to the energy

density u.0 ; 1 ; 2 ; : : : ; c / whose differential is
X
du D hi di ; (11.10)
i

that is,
@u
hi D : (11.11)
@i
We remark that u.0 ; 1 ; 2 ; : : : ; c / is a continuous and convex function of the
densities altogether while .h0 ; h1 ; h2 ; : : : ; hc / is a continuous and concave function
of the fields altogether. Although  is a continuous function, its derivatives, the
densities, may not be continuous. And indeed they are not continuous in the points
of space-h corresponding to a phase coexistence. A point of discontinuity in the
densities in this space indicates the occurrence of a phase coexistence.

Manifolds

The Gibbs phase rule concerns the number of phases that may coexist in a
thermodynamic system with c components. The phase rule is a law that can not
be derived from the laws of thermodynamics alone. It should be understood as an
independent law, compatible with the laws of thermodynamics.
In a system composed by a single pure substance (one component) we have seen
that two phases, for example, liquid and vapor coexist along a line in the phase
diagram T versus p. We have also seen that the coexistence of three phases, solid,
liquid and vapor occurs only at a single point of the two-dimensional diagram.
A binary mixture (two components) on the other hand has a three-dimensional
diagram. The usual thermodynamic space includes the temperature T the pressure p
and a third field which we choose as the difference between the chemical potentials
of the two components. In this three-dimensional phase diagram, the coexistence of
two phases occurs on a surface, the coexistence of three phases takes place along a
line and four phases can only coexist in a point of the diagram. These results and
the generalization of them comprises the Gibbs phase rule.
To state the phase rule it is convenient to introduce what is known as the number
of degrees of freedom f for the coexistence of a certain number of phases. If the
coexistence occurs in a single point in space-h, then f D 0; if it occurs along a line,
f D 1; if it occurs on a surface, f D 2, etc. That is, f is identified as the dimension of
the manifold (a point, a line, a surface, etc.) in space-h corresponding to the phase
coexistence. The Gibbs phase rule states that the number of degrees of freedom f
for the occurrence of m thermodynamic phases in a system with c components is
given by

f D c C 2  m: (11.12)
204 11 Phase Diagrams

That is, the dimension of the manifold in space-h corresponding to the coexistence
of m phases is equal to c C 2  m.
The rule may also be expressed in terms of the manifold codimension, defined
as the difference between the dimension of the space in which the manifold is
immersed and the manifold dimension. As the dimension of space-h is c C 1 then a
codimension  of the manifold of phase coexistence is  D .c C 1/  f . Thus, the
Gibbs phase rule is equivalent to say that the codimension of the manifold of the
coexistence of m phases in space-h is m  1, or

 D m  1: (11.13)

Simplices

Suppose that at a given point in space-h, occurs the coexistence of m phases with
densities E .1/ ; E .2/ ; : : : ; E .m/ . In space-, these densities are a set of m distinct points
whose spatial distribution is a simplex of dimension m  1 (Fig. 11.1). A zero-
dimensional simplex is a point. A one-dimensional simplex is a line segment, a
two-dimensional simplex is a triangle, a three-dimensional simplex is a tetrahedron,
etc. A .m  1/-dimensional simplex comprises a set of m generic vertices located
at arbitrary positions of a space with dimension greater than or equal m  1. The
generality of the position means that two vertices of a simplex never coincide, three
vertices are never collinear, four vertices are never at the same plane, and so on.
The proposition that the coexistence of m phases in a system of several
components is represented in space- by a .m  1/-dimensional simplex constitutes
the Griffiths-Wheeler postulate. In the following we show that the Gibbs phase rule
can be derived from this postulate.
Denoting by  .1/ ;  .2/ ; : : : ;  .m/ the thermodynamic potentials that describe each
of the m single phases in the neighborhood of the manifold of the phase coexistence,
then, for each single-phase region, we write the following equation

X
c
. j/
d . j/ D  i dhi : (11.14)
iD0

Fig. 11.1 Simplices: zero-dimensional (point), one-dimensional (line segment), two-dimensional

(triangle) and three-dimensional (tetrahedron)
11.1 Gibbs Phase Rule 205

Subtracting the equations corresponding to j D 2; : : : ; m from that corresponding to

j D 1 and taking into account the equality of the thermodynamic potentials  . j/ at
the points of the manifold of phase coexistence, we reach the Clausius-Clapeyron
equation for a system of c components and m phases in coexistence,

X
c
. j/ .1/
Πi  i dhi D 0; (11.15)
iD0

valid for j D 2; : : : ; m. We may also write the Clausius-Clapeyron equation as

X
c
. j/
i dhi D 0 (11.16)
iD0

. j/
valid for j D 2; : : : ; m, where i are the components of the vector E . j/ defined by

E . j/ D E . j/  E .1/ (11.17)

for j D 2; 3; : : : ; m.
The set of equations (11.16) can be written as a scalar product

E . j/  dhE D 0; (11.18)

valid for j D 2; 3; : : : ; m. This equation describes locally the manifold in space-h

corresponding the coexistence of m phases. Taking into account that the densities
E .1/ ; E .2/ ; : : : ; E .m/ constitute a simplex .m  1/-dimensional, it easy to see that the
vectors E . 2/ ; E .3/ ; : : : ; E .m/ comprises a set of m  1 vectors linearly independent.
On the other hand, each one of the equations given by (11.18) describes a hyperplane
in space-h perpendicular to the vector E . j/ . Therefore, the m  1 (11.18) describe
the intersection of m  1 hyperplanes. Since no hyperplane coincides with another
because the vectors E . j/ are linearly independent, then the intersection of the
hyperplanes comprises a manifold of dimension f D .c C 1/  .m  1/ D c C 2  m,
which is the Gibbs phase rule.
The Gibbs phase rule is applicable to systems of several components that are
devoid of symmetry. It can not be used indiscriminately when symmetry has
an important role. In some cases, however, we can restore the Gibbs rule if
thermodynamic space of symmetric systems is extended properly. For example,
suppose that two or more phases in coexistence are connected by symmetrical
operations. The Gibbs phase rule may be applied provided that the thermodynamic
fields, that completely break the symmetry between the phases, are added to the
space-h.
206 11 Phase Diagrams

11.2 Structure of Phase Diagrams

Ordinary Manifold

The phase diagram in space-h of a system of c components comprises the manifolds

(points, lines, surfaces, etc.) of phase coexistence which are connected to each other
forming a spatial structure. With the exception of zero-dimensional manifold, which
is a point, other manifolds extend continuously in space-h and can reach the bound-
aries of space-h. However, it is possible, and is actually more common, for it to finish
by meeting or by transforming into other manifolds. To examine these possibilities,
we must take into account that each point in space-h that belongs to a manifold of
m coexisting phases is associated with a .m  1/-dimensional simplex in space-.
A simplex corresponding to m coexisting phases is composed by m vertices in
space-. A property relevant to our analysis is that any set of m  1 vertices of this
simplex is also a simplex but of dimension smaller by one unit. Therefore, we can
imagine that a simplex of m vertices has as neighbors m simplices of m  1 vertices.
This implies that in space-h, from a manifold of coexistence of m phases emerge
m manifolds of coexistence of m  1 phases. It is noteworthy that the latter has
dimension one unit smaller.
Based on this property, we can build possible structures of the manifolds of phase
coexistence, which we call ordinary manifolds in opposition to critical manifolds
which we will see later. In a space of dimension c C 1, we start from a zero-
dimensional manifold (a point) which corresponds to the coexistence of cC2 phases.
From this point emerge c C 2 lines of coexistence of c C 1 phases. From each of
these coexistence lines, in turn, spring c C 1 surfaces of coexistence of c phases.
From these surfaces can sprout manifolds of dimension three, and so on.
According to this scheme, we examine the phase structure of a system of one
component. In this case, the space-h is two-dimensional. The zero-dimensional
manifold (a point) corresponds to the coexistence of three phases (triple point).
Three lines of two phase coexistence converge to this point, as shown in Fig. 7.6a.
For a two component system, the space-h is three dimensional. The zero-
dimensional manifold (a point) corresponds to the coexistence of four phases.
To this point, four lines of coexistence of three phases converge, as shown in
Fig. 11.2. To each of these lines, surfaces of two phases coexistence converge.

Fig. 11.2 Phase diagram in h2

field space (h0 ; h1 ; h2 ) of a
system with two components.
The surfaces of coexistence
of two phases meet at lines of
coexistence of three phases
(dashed line). These converge
to a point of coexistence of
four phases (quadruple point) h0
h1
11.2 Structure of Phase Diagrams 207

A phase diagram comprises one or more structures of the type shown in Figs. 7.6a
and 11.2. For example, a pure substance can have multiple triple points. The most
notable of these is the point of coexistence solid-liquid-vapor. However, a pure
substance can show various crystal forms that can coexist with each other or together
with the liquid forming several point of three-phase coexistence.
We should also note that cuts performed in a phase diagram in space-h of a
system of c components may represent a possible phase diagram of a system of c1
components. However, this cut should be generic, that is, it should avoid passing
through peculiar points. For example, a cut in the diagram of Fig. 11.2, represented
by h2 D const is a possible diagram of a system of one component, since the cutting
is not done on the quadruple point or tangentially to the line of triple points.

Critical Manifolds

In the study of the possible structures of a phase diagram in space-h, it is convenient

to think that an ordinary manifold of dimension f is described in parametric form
by a set of f parameters. The continuous variation of these parameters allows us to
walk on the manifold. Recall that to each point belonging to the ordinary manifold
of coexistence of m phases corresponds a simplex in the space- whose number of
vertices is equal to m. Therefore, walking along the points of an ordinary manifold,
the number of vertices of the corresponding simplices is invariant. However, the
distances between the vertices may change when walk along the manifold. Thus,
one possible mechanism for the disappearance of a manifold is that for which two
vertices of a simplex collapse into a single vertex, which means that two phases
become identical. The corresponding points in space-h define a manifold that we
call critical manifold, which has dimension equal to that corresponding to three
ordinary phases in coexistence.
One can imagine critical manifolds where more than two phases become equal.
For example, we may suppose that three phases become identical. In this case the
critical manifold has dimension equal to that corresponding to the manifold of five
ordinary phases in coexistence. In general, when i ordinary phases become identical,
the critical manifold has dimension equal to that of the manifold corresponding to
2i  1 ordinary phases in coexistence, which is the Zernike rule.
Let us imagine now a configuration of phases of a system in which there are the
coexistence of m1 ordinary phases, m2 critical phases (two ordinary phases becom-
ing identical), m3 tricritical phases (three ordinary phases becoming identical), etc.
According to the Zernike rule this configuration of phases correspond, as regards to
the Gibbs phase rule, to a number of ordinary phases equal to

m D m1 C 3m2 C 5m3 C    : (11.19)

According to this rule, a system comprising a pure substance, c D 1, may display

only a critical zero-dimensional manifold in which two phases become equal. This
manifold is the critical point, the terminal point of a line of coexistence of two
208 11 Phase Diagrams

phases, which was the object of our study in previous chapters. The phase diagram
in the vicinity of this point are shown in Fig. 7.5, both in space fields and space of
densities.

Critical Line and Critical End Point

The thermodynamic space of fields, corresponding to a binary mixture, consists of

three variables. Usually, they include temperature, pressure, and a third thermody-
namic field that we choose as the difference between the chemical potentials of
the two components. In this space, the coexistence of two phases takes place on a
surface, which can extend to the boundaries of the thermodynamic space or may
end in a line of critical points, which is identified as the edge of the surface of
coexistence, as shown in Fig. 11.3.
The phase diagram of a binary system may also contain lines of coexistence
of three phases, as can be seen in Figs. 11.2 and 11.4. Walking along the line of
coexistence of three phases, it is possible that it terminates at a quadruple point as
seen in Fig. 11.2. It is also possible that at some point two phases become identical
in the presence of third, as illustrated in Fig. 11.4. The corresponding point is called
critical end point because in addition to being the terminal point of a line of triple
points it is the terminal point of a line of critical points. As shown in Fig. 11.4, the
critical line, which is the edge of the surface of coexistence, ends on another surface
of the coexistence of two phases.

Fig. 11.3 Phase diagram in h2

space .h0 ; h1 ; h2 / of a system
of two components. The edge
(thick line) of the surface of
two-phase coexistence is a
critical line

h0
h1

Fig. 11.4 Phase diagram in h2

space .h0 ; h1 ; h2 / of a system
of two components. The
surfaces of two-phase
coexistence meet at the line
(dashed line) of three-phase
coexistence. The end point of
the line of critical points
(thick line) is a critical end h0
h1
point
11.2 Structure of Phase Diagrams 209

(a) (b) h2
h2

h1 h1

h0 h0

(c) (d)
h2
h2

h1
h1

h3 h3

Fig. 11.5 Phase diagram of a ternary system corresponding to cuts in space .h0 ; h1 ; h2 ; h3 / such
that one of the four fields is kept constant. The surfaces correspond to the coexistence of two
phases. The continuous lines are critical lines. The dashed line is a line of triple points. Small filled
circles are critical end points. The full large circle is the tricritical point

Tricritical Point

A tricritical point is the one in which three phases become identical. It should
be distinguished from an ordinary critical in which only two become identical.
According to the Gibbs and Zernike rules three phase can only become identical
in a system that has at least three components. In a three-component system, whose
space-h is four-dimensional, three phases become identical in a zero-dimensional
manifold (a point). As the viewing of a space of more than three dimensions is
impractical, it is expedient to examine projections or cuts of this space. We can also
use the fact that a phase diagram in a given space can be seen as a possible cut of a
phase diagram in a space constituted by an additional field. This additional field can
then be continuously varied to change the phase diagram.
Consider the diagram of Fig. 11.5a, which can be considered a phase diagram of
a binary system and also as a cutting of the four-dimensional space .h0 ; h1 ; h2 ; h3 /
of a ternary system. In this diagram, we are assuming that the variable h3 has a fixed
value. We then vary h3 appropriately so that the critical end point moves toward the
other critical line. The tricritical point will occur when the critical end point touches
the other critical line, as can be seen in Fig. 11.5b. We see that the tricritical point is
the meeting point of three critical lines and is also the terminus of the line of triple
210 11 Phase Diagrams

points. It is also possible to imagine other cuts, like those shown in Fig. 11.5c, d.
Figure 11.5c presents a line of triple points whose ends are critical end points and is
a possible phase diagram of a binary system. Appropriately varying one of the fields,
we can make the two critical end points to approach each other. The tricritical point
occurs when they coalesce, as shown in Fig. 11.5d.

Multicritical Points

In addition to the points mentioned above, we can imagine other points in which
several ordinary phases become identical in the presence or absence of other phases.
However, according to Gibbs and Zernike rules, the observation of such points on a
mixture of pure substances is possible only above a certain number of components.
A bicritical point (two critical phases in coexistence) can only be observed in a
mixture with at least four components. A tetracritical point (four ordinary phases
that become identical), can only be observed in a mixture of at least five components.
It is convenient to use a notation of product type to represent the configuration of
the phases in coexistence. An ordinary phase is represented by the letter A; a critical
phase, in which two ordinary phases become identical, by the letter B; a tricritical
phase, in which three phases become identical, by the letter C, etc. A state composed
of two ordinary phases in coexistence is represented by A2 , a triple point by A3 . A
critical end point by AB, a bicritical point by B2 . In general, a multicritical point is
represented by Am1 Bm2 Cm3 Dm4 : : :. In accordance with the Gibbs and Zernike laws,
the minimum number of components necessary to observe a multicritical point is
given by

c D m1 C 3m2 C 5m3 C 7m4 C     2: (11.20)

Other Spaces

So far, the phase diagrams were constructed in space-h, composed only by

thermodynamic fields. From the experimental point of view, it is sometimes more
convenient to use thermodynamic densities instead of thermodynamic fields as
independent variables. In a mixture of several components it is common the use of
the mole fractions, in addition to temperature and pressure, as control variables. It is
useful, therefore, to know how to construct a diagram in the space of densities from a
diagram in the space of fields. The fundamental rule is the Clausius-Clapeyron equa-
tion (11.16) which tells us that the tie lines should be perpendicular to the lines of
coexistence, provided that the axes of space and  are parallel to the axes of space-h.
Figure 11.6a, c, e present diagrams corresponding to three cuts of the phase
diagram of Fig. 11.4, made at constant pressure. We are identifying the fields h0 ,
h1 and h2 as the temperature T, the activity  and the pressure p, respectively.
11.2 Structure of Phase Diagrams 211

(a) (b) C
C

T
T

A B
A B

(c) (d)
C
C
T
T

A B A
B

(e) (f)
C
C
T
T

A B A B

0 ζ 1 0 x 1

Fig. 11.6 Figures (a), (c) and (e) show the phase diagrams of a binary mixture in space
temperature T versus activity , at constant pressure. They correspond to cuts in the diagram
shown in Fig. 11.4. Figures (b), (d) and (f) are the respective phase diagrams in space T versus
mole fraction x, at constant pressure. The full circle represents an ordinary critical point and an
empty circle represents a critical end point

Figure 11.6a corresponds to a cut made at a pressure just above the pressure of the
critical end point. In this diagram, we see that, above the critical temperature, the
substances become completely miscible, passing to phase C at higher temperatures.
Figure 11.6e corresponds to a cut made at a pressure just below the pressure
of the critical end point. An increase in temperature of the phases A and B
212 11 Phase Diagrams

(a) (b)
h2 ρ2

h1 ρ
1

Fig. 11.7 (a) Phase diagram corresponding to a cut in Fig. 11.5c done at constant h3 and passing
between the two critical end points. (b) Analogous diagram in the space of conjugated densities

ρ3

ρ
ρ 1
2

Fig. 11.8 Phase diagram analogous to that of Fig. 11.5c. The triangles represent the coexistence
of three phases. The two-phase regions and the ordinary critical point are not shown in the figure
but are similar to that shown in Fig. 11.7b. The open circles represent critical end points

in coexistence causes the appearance of phase C when the temperature reaches

the triple point temperature, whereupon the three phases remain in coexisting.
Figure 11.6c corresponds to a cut made at the pressure of the critical end point.
Figure 11.6b, d, f are the respective diagrams in the space temperature versus mole
fraction. Various binary mixtures of partially miscible liquids have diagrams of the
type shown in Fig. 11.6. In this case A and B are liquid phases and C is a gas phase.
We also find diagrams of this type in binary mixtures of partially miscible solids,
such as gold-platinum system whose diagram is shown in Fig. 10.7. In this system,
A and B are solid phases and C is a liquid phase.
Figure 11.7a presents a phase diagram corresponding to a cut in Fig. 11.5c made
between the two critical end points. The diagram analogous in the space of densities
is shown in Fig. 11.7b. We see that the tie lines are perpendicular to the lines of
two-phase coexistence. The triangle represents the coexistence of three phases.
Figure 11.8 shows the analogous of Fig. 11.5c in the space of densities. The
coexistence of three phases, which is a line in the space of fields, is represented
by a stack of triangles in the space of densities.
11.3 Griffiths-Landau Theory 213

11.3 Griffiths-Landau Theory

Landau Function

The Griffiths-Landau theory allows us to determine the topology of phase diagrams

in the vicinity of critical and multicritical points. It is still able to provide
quantitative results, but its importance lies in the simplicity in obtaining the topology
of phase diagrams.
The Griffiths-Landau theory takes as its starting point a thermodynamic potential
f which has a mixed dependency on fields and densities. The potential f is a poly-
nomial function of densities. The polynomial coefficients depend only on the fields.
These two properties are sufficiently generic in such a way that any thermodynamic
potential that is analytic function can be approximated by a potential of Griffiths-
Landau type by means of a Taylor series expansion in the densities. These include,
for example, those relating to the theory of van der Waals and regular solution
theory.
Next we develop the theory for the case in which the thermodynamic potential
f is a function of only one density, which we denote by . The dominant term
of the polynomial is chosen as being convex which means that the highest power
must be even. However, this is not sufficient to guarantee the convexity of f
since a polynomial function do not have in general the property of convexity. The
thermodynamic potential is therefore taken as the convex hull of f , obtained from a
Legendre transformation

 D minf ./g; (11.21)



where

D f ./  h (11.22)

is called Landau function. Being f ./ a polynomial then ./ is also a polynomial
of the same degree. We remark that the Landau function depend on all fields relating
to the potential  and not only on those relating to the potential f . The dependence
on the fields, however, is found only on the coefficients of the polynomial of ./.
Next we introduce a change of variables to make the analysis simpler. Defining
the new density x D   , N we see that .x/ will also be a polynomial and of the
same degree of ./. In addition, the potential  will be given by

 D minf .x/g: (11.23)

x

The choice of N will be done so that one of the coefficients of the polynomial in x,
other than the dominant, vanishes. We point out that, although the coefficients of
the new polynomial are distinct from the older ones, they are functions of the fields
only since N will depend only on the older coefficients and thus only on the fields.
214 11 Phase Diagrams

The description of the topology around the critical and tricritical points, as we
will see below, is obtained from the Landau polynomial

.x/ D a0 C a1 x C a2 x2 C    C an2 xn2 C xn ; (11.24)

where n is the degree of the polynomial, that must be even. The coefficient of the
dominant term was taken as being equal to unity. This expedient is justified by
dividing both sides of (11.21) by the coefficient of the dominant term, which is
strictly positive. The choice of the variable x was done so that the penultimate term
is absent, that is, so that the coefficient of the term xn1 vanishes.
As we said above, the coefficients ai of the terms of the Landau polynomial
depend only on the thermodynamic fields. According to the Griffiths-Landau theory,
this dependence is smooth so that it is always possible to assume that it is linear.
Thus we consider that the coefficients ai are themselves fields, obtained from
the original fields by the linear dependence. This implies that the topology of
the diagram is the same whether we use the original fields or the coefficients ai ,
identified as fields.
Adopting the coefficients ai as fields, it is convenient to determine the density
conjugated to them, which we denote by i and are given by

@
` D  ; (11.25)
@a`

from which we obtain

` D x` ; (11.26)

where x, in this equation, represents an absolute minimum of .x/.

Factorized Representation

It is convenient in our analysis to consider the polynomial Q defined by Q .x/ D

.x/  . This polynomial has the following fundamental properties. If x is an
absolute minimum of the Landau function .x/, then x is identified as the root of
Q .x/ with even multiplicity (double, quadruple, etc.). In fact, according to (11.23),
 D .x / and therefore Q .x / D 0. In addition, according to (11.23), .x/  
and therefore Q .x/  0 for any x. As Q .x/ is analytic then in the neighborhood of
x D x this function behaves as .x  x /k , where k is even.
Based on these results, we can construct the specific Landau functions for each
one of the manifolds of the phase diagram. To this end we use the factorized
representation

Q .x/ D .x  r1 /.x  r2 /    .x  rn /: (11.27)

11.3 Griffiths-Landau Theory 215

(a) ψ (b)
y

Fig. 11.9 (a) Landau function corresponding to the coexistence of an ordinary phase (quadratic
minimum, double root) and a critical phase (quartic minimum, quadruple root). (b) Roots of Q in
the complex plane: small full circle, double root; big full circle, quadruple root; and empty circles,
simple roots

where ri are the roots of the polynomial Q which are chosen according to the
following rules.
(a) To each ordinary phase in coexistence there corresponds a double real root.
(b) To each critical phase (two ordinary phase becoming identical) in coexistence
there corresponds a real quadruple root.
(c) To each tricritical phase (three ordinary phases becoming identical) in coexis-
tence there corresponds a real sixfold root.
(d) In general, to each multicritical phase such that p phases becomes identical there
corresponds a real root of multiplicity 2p.
The remaining roots cannot be real because otherwise Q .x/ would vanish since
x takes on real values. These non real roots must appear in pairs of complex
conjugated numbers because the coefficients of the polynomial are real. The real
roots constitute the absolute minima of Q .x/ and, because they are also minima of
.x/, they are identified as the densities of the phases in coexistence.
As an example of a Landau function, Fig. 11.9 shows a polynomial of degree
eight corresponding to a critical end point (coexistence of an ordinary phase and a
critical phase). The polynomial has one double real root, one quadruple real root
and a pair of complex conjugated roots.
Finally, the absence of the term xn1 in the development of the product (11.27)
means that the sum of the roots vanishes, that is,

r1 C r2 C    C rn D 0: (11.28)

If all roots are equal then they should vanish.

The determination of the equation that describe a certain manifold in the space
of the fields is reduced to the comparison of the development of the product (11.27)
with (11.24). This comparison allow the determination of the relations between
the coefficients ai and the roots ri . Understanding the roots ri as parameters, these
relations constitute the equations that describe the manifold in a parametric way.
216 11 Phase Diagrams

Critical Point

To describe the region of the phase diagram around the critical point it suffices to
consider a polynomial of fourth order in x, Thus, the expression which we use for
the Landau function is

D a 0 C a 1 x C a 2 x2 C x4 : (11.29)

The coefficients of the Landau function define the two-dimensional space .a1 ; a2 /.
To determined the occurrence of two phases in coexistence, the polynomial Q .x/
should have two distinct double roots what allows us to write

Q D .x  r1 /2 .x  r2 /2 : (11.30)

Since the sum of the roots vanishes, then r2 D r1 , so that

Q D .x2  r12 /2 D x4  2r12 x2 C r14 : (11.31)

Therefore,

a2 D 2r12 and a1 D 0; (11.32)

and the conditions for the coexistence of two phases are

a1 D 0 and a2 < 0; (11.33)

which define the line of coexistence. The densities of the two phases en coexistence
are given by

1 1
r1 D p ja2 j1=2 and r2 D  p ja2 j1=2 : (11.34)
2 2

When the two phases become identical, the double roots become identical.
Therefore, a critical point is characterized by the occurrence of quadruple root, that
is, Q D x4 , which gives the result

a1 D 0 and a2 D 0; (11.35)

which are the conditions that determine the critical point. The topology around the
critical point can be seen in the diagram of a1 versus a2 as shown in Fig. 11.10. We
show also the diagram in the variables 1 and a2 , where 1 is the density conjugated
to the field a1 and given by 1 D r1 or 1 D r2 .
11.3 Griffiths-Landau Theory 217

(a) a1 (b) ξ
1

a2 a2

Fig. 11.10 (a) Phase diagram around the critical point in the space of the coefficients a1 and a2 .
The dashed line is a two-phase coexistence line. The full circle is the critical point. (b) Same
diagram in the space density 1 versus a2 , where 1 is the density conjugated to the field a1 . The
region of coexistence of two phases is indicated by tie lines

Tricritical Point

The topology of the region around the tricritical point is obtained from a polynomial
of sixth degree, that is, from the Landau function

D a 0 C a 1 x C a 2 x2 C a 3 x3 C a 4 x4 C x6 : (11.36)

Notice that the term x5 is absent. The coefficients of the Landau function define a
four-dimensional space .a1 ; a2 ; a3 ; a4 /.
To describe the coexistence of three phases (triple point) it is necessary that Q .x/
has three distinct double roots, that is,

Q D .x  r1 /2 .x  r2 /2 .x  r3 /2 ; (11.37)

where r1 , r2 and r3 are distinct and are the densities of each phase. Since the sum of
the roots vanishes then r3 D r1  r2 , what allows to write

Q D .x  r1 /2 .x  r2 /2 .x C r1 C r2 /2 : (11.38)

Developing the product, we get

Q D .x3  px C q/2 D x6  2px4 C 2qx3 C p2 x2  2pqx C q2 (11.39)

where

p D r12 C r22 C r1 r2 and q D .r1 C r2 /r1 r2 ; (11.40)

Comparing expression (11.39) with (11.36), we get the relations

a4 D 2p; a3 D 2q; a2 D p2 ; and a1 D 2pq; (11.41)

218 11 Phase Diagrams

with a4 < 0 because p > 0 for any values of the densities r1 and r2 . Elimination p
and q, we obtain the following relations among the parameters ai

1 2 1
a2 D a ; a1 D a3 a4 and a4 < 0; (11.42)
4 4 2
which define surfaces of coexistence of three phases in a four-dimensional space.
Let us set up now Landau functions corresponding to critical end points. These
points correspond to the occurrence of two coexistence phases, one of them an
ordinary phase and the other a critical phase. Therefore, the critical end points are
characterized by one double root and one quadruple root, that is,

Q D .x  r1 /4 .x  r2 /2 ; (11.43)

where r1 is the density of the critical phase and r2 of the ordinary phase. Since the
sum of the roots vanishes, then r2 D 2r1 yielding the result

Q D .x  r1 /4 .x C 2r1 /2 : (11.44)

Developing the product and comparing it with the expression (11.36), we get the
relations

a4 D 6r12 ; a3 D 4r13 ; a2 D 9r14 ; and a1 D 12r15 : (11.45)

Eliminating r1 we get the following relations among the parameters ai

1 1
a1 D ˙ p ja4 j5=2 ; a3 D ˙ p ja4 j3=2 ;
3 6 3 6

1 2
a2 D a and a4 < 0: (11.46)
4 4
These relations determine lines of critical end points in a four-dimensional space.
The Landau function corresponding to an ordinary critical point is characterized
by the occurrence of quadruple root. Since the critical phase is not in coexistence
with other phases then the other roots must be non real. Therefore

Q D .x  r1 /4 .x  r2 /.x  r3 /; (11.47)

where r1 is the density of the critical phase and r2 and r3 are complex conjugate. As
the sum of the roots vanishes then r2 C r3 D 4r1 what permits us to write Q in the
form

Q D .x  r1 /4 .x2  4r1 x C r2 /; (11.48)

11.3 Griffiths-Landau Theory 219

where r2 D r2 r3 is real and positive. Developing the product and comparing it with
expression (11.36), we get the following relations

a4 D 10r12 C r2 ; a3 D 20r13  4r1 r2 ;

a2 D 15r14 C 6r12 r2 and a1 D 4r15  4r13 r2 : (11.49)

These relation determine in a parametric form a critical surface in a four-

dimensional space. They in fact define three critical surfaces. One of them is
obtained when r1 D 0, which results in

a1 D a2 D a3 D 0 and a4 > 0: (11.50)

The other two correspond to r1 ¤ 0 and are given implicitly by (11.49).

Finally, a tricritical point is described by a Landau function Q D x6 from which
we get

a1 D a2 D a3 D a4 D 0; (11.51)

which determined the tricritical point in a four-dimensional space.

Diagrams

To visualize the topology of the phase diagram around the critical and tricritical
points, we use as the space of thermodynamic fields the space of the coefficients ai .
For the critical point, the space is two-dimensional and is composed by the variables
a1 and a2 . Figure 11.10 shows the critical line and the critical points traced according
to the conditions (11.33) and (11.35), respectively.
For the tricritical point the space is four-dimensional and is compose by the
variables a1 , a2 , a3 and a4 . Due to the impossibility of visualizing such a space we
examine cuts made at constant a3 and at constant a4 . The various manifolds defined
by the conditions obtained above can be seen in Fig. 11.5 considering that a1 , a2 ,
a3 and a4 are, respectively, parallel to h1 , h2 , h3 and h0 . In these three-dimensional
diagrams, the triple points and the critical points are lines. The critical end points
and the tricritical point are isolated points.
Figure 11.5a, b correspond to cuttings made at constant a3 . In the first a3 is
nonzero and in the second a3 D 0. Only in this second diagram we can observe the
tricritical point since this point occurs when a1 D a2 D a3 D a4 D 0. For a3 D 0,
the critical lines are determined explicitly from (11.49). For this it suffices to set
r2 D 5r12 to get

a4 D 5r12 ; a2 D 15r14 and a1 D 16r15 ; (11.52)

220 11 Phase Diagrams

(a) (b)
a2 ξ2

a4 a4

Fig. 11.11 (a) Phase diagram around the tricritical point in the space of the coefficients a2 and a4 ,
corresponding to a1 D a3 D 0. The dashed curve is a three-phase coexistence line. The full circle
is the tricritical point. The thick line is a critical line. The region below the two lines correspond to
points of two-phase coexistence. (b) Same diagram in space 2 versus a4 , where 2 is the density
conjugated to the field a2 . The region of three-phase coexistence is indicated by tie lines

from which we get the equations for both critical lines

3 2 16
a2 D a ; a1 D ˙ p ja4 j5=2 and a4 < 0 (11.53)
5 4 25 5

The other critical line is described by (11.50).

If in the diagram corresponding to a3 D 0 we consider a cutting defined by
a1 D 0, we obtain the two-dimensional diagram in the variables a2 and a4 , as shown
in Fig. 11.11. In this diagram the critical line is given by (11.50) and the line of triple
points by a2 D a24 =4 and a4 < 0, obtained from (11.42). In this figure we show also
the diagram in the variables 2 versus a4 , where 2 , the density conjugated to the
field a2 , is given by 2 D a4 =2 and a4 < 0 or 2 D 0 and a4 < 0.
We can also examine a cut made at constant a4 which are shown in Fig. 11.5c,
d, and correspond respectively to a4 < 0 and a4 D 0. In the diagram of Fig. 11.5c,
the line of triple points finish in two critical end points. The location of these points
are given by (11.46). From them, we see that the distance between these two points
vanishes when a4 ! 0. Equivalently, the line of triple points shrinks and disappears
in this limit, as seen in Fig. 11.5d.
A diagram analogous to that of Fig. 11.5c can be seen if Fig. 11.8. The line of
triple points correspond to a stack of triangles. At the bottom and at the top of the
stack one finds line segments (degenerated triangles) that represent the coexistence
of a critical phase and an ordinary phase (critical end point). When a4 ! 0, the
stack of triangles shrinks to a single point which is the tricritical point. However,
the stack does not shrink uniformly in the three directions. The height of the stack
`3 , which is the distance between the critical end points, decreases as

`3 D ja4 jˇ3 ; ˇ3 D 3=2; (11.54)

11.3 Griffiths-Landau Theory 221

a result that is obtained from (11.46). The length of the base of the triangles `2 is
equal to the length of the tie lines shown in Fig. 11.11b, which behave linearly with
a4 , that is,

`2 D ja4 jˇ2 ; ˇ2 D 1: (11.55)

The height of the triangles `1 is associated to the variable 1 , conjugated to a1 .

Therefore,

`1 D ja4 jˇ1 ; ˇ1 D 1=2: (11.56)

The exponents ˇ1 , ˇ2 and ˇ3 characterize the behavior of the densities around

the tricritical point. According to the Griffiths-Landau theory the values of these
exponents are 1/2, 1 and 3/2, respectively. The experimental results obtained in
ternary and quaternary mixtures are in agreement with these results.
Chapter 12
Order-Disorder Transition

12.1 Binary Alloys

Ordered Alloys

A crystalline solid has as a fundamental property the ordered structure of its atoms.
This structure consists of an array of sites forming a regular three-dimensional
periodic lattice. The ordered structure, strictly speaking, does not mean that the
atoms are located exactly on the sites of the lattice as they are in constant motion
due to thermal agitation. In fact, the average positions of the atoms are the points in
space that should be considered as the sites of the ordered lattice. Only for simplicity
we say that atoms themselves form the crystal lattice.
The possible crystalline structures are classified according to the symmetry
groups. The crystalline solids are ordered according to one or other structure that
defined the thermodynamic state. Many metals are ordered according to the structure
called face-centered cubic (fcc). In this structure the atoms of a simple solid are
located at the vertices and centers of cubes that make up the structure. Among
the metals that are arranged according to the fcc structure are copper, gold, silver,
platinum, nickel, lead and aluminum. Binary alloys of any of these metals form
substitutional disordered alloys with the same structure.
Gold and silver form continuous solid solution for these two metals are miscible
in all proportions. A solution of gold and silver has the fcc spatial structure where
the gold and silver atoms are distributed randomly among the lattice sites, forming
an alloy called disordered alloy. The gold and platinum also form alloys whose
atoms are distributed at random on the sites of a fcc lattice. However, gold and
platinum are not completely miscible. For certain compositions, the gold-platinum
system separates into two phases which are also disordered alloys. These two
phases in coexistence differ in the composition one being gold-rich and the other
platinum-rich as shown in the diagram of Fig. 10.7. Raising the temperature the two

© Springer-Verlag Berlin Heidelberg 2017 223

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_12
224 12 Order-Disorder Transition

disordered alloys become identical at a critical point. Above the critical temperature,
gold and platinum are miscible in all proportions.
The alloy consisting of gold and copper, on the other hand, offers the opportunity
to illustrate an alloy where the atoms are not randomly distributed on the sites of
a fcc lattice but are preferably located at certain sites of the crystal lattice. For
a better understanding, we must consider that a fcc lattice can be understood as
composed of four interpenetrating cubic sublattices. For a mole fraction of copper
around 0.25, the copper atoms are located preferentially in a sublattice and the gold
atoms preferentially in the other three sublattices, thereby forming the ordered alloy
Au3 Cu. Heating this alloy it passes from the ordered state to the disordered state.
The binary alloy consisting of magnesium and cadmium also presents order
of their atoms. Figure 12.1 presents the phase diagram temperature versus mole
fraction of magnesium where one can observe the regions corresponding to the
ordered alloys Cd3 Mg, CdMg and CdMg3 as well as the disordered alloy (Cd,Mg).
The transitions between all these phases are first order including that which occurs
at x D 0:5 and at a temperature of 253 ı C. This transition is congruent, that is, it
occurs with no change in composition.
A binary system particularly interesting is the one consisting of copper and
zinc, and generically named brass. This system has several phases according to the
composition. For a mole fraction of zinc from 0.45 to 0.482 and temperature around
450 ı C, the brass is in a phase called ˇ 0 -brass. At this phase the alloy is in a cubic
ordered structure in which the zinc atoms are preferably at the vertices of the cube
and the copper atoms preferably in the centers of the cubes. Heating the alloy, it
passes to a disordered phase called ˇ-brass, where the atoms of zinc and copper
are distributed indiscriminately at the corners and center of the cubes of the lattice,
comprising a lattice called body centered cubic (bcc). The transition occurs at a

(a) 300 (b) 1

(Cd,Mg)
0.9

200
θ ( C)

0,.
η
o

CdMg 0.7
100
CdMg3
0.6
Cd3Mg

0 0.5
0.2 0.4 0.6 0.8 0 100 200 300
x o
θ ( C)

Fig. 12.1 Phase diagram and order parameter of the binary alloy cadmium-magnesium obtained
experimentally by Franz and Gantois [136]. (a) Diagram temperature  versus mole fraction x of
magnesium. (b) Order parameter  as a function of temperature  along x D 0:5. It displays a
jump at the transition to the ordered phase that occurs at a 253 ı C
12.1 Binary Alloys 225

temperature of 454 ı C for a mole fraction of zinc of 0.448 and at a temperature of

468 ı C for a mole fraction of zinc of 0.482. This transition is a continuous transition.

Order Parameter

Consider an ordered alloy of type brass-ˇ 0 formed by two kinds of atoms A and B,
which are arranged in a cubic lattice in which the atoms are located either at the
vertices or at the centers of the cubes. The vertices and the centers in turn form two
sublattices that are intertwined. Suppose that the fraction x of atoms of type A is
exactly 0.5. The order does not necessarily mean that all the atoms A are found in
one sublattice and all atoms B are found in the other sublattice. This only occurs
when the order is complete, which should occur at low temperatures. The ordering
means that the atoms of one type are found preferably in a sublattice and the atoms
of the other type preferably in the other sublattice. It is convenient therefore to define
the fractions x1 and x2 of atoms A in the sublattice formed by the vertices and centers
of the cubes, respectively. The ordered alloy does not only mean that x1 D 1 and
x2 D 0 or that x1 D 0 and x2 D 1 but that x1 ¤ x2 . The alloy becomes disordered
when x1 D x2 .
Suppose now that the fraction x of atoms of type A is different from 0.5. The
disordered alloy implies that the fraction of atoms of type A is the same in any of
the sublattices and hence x1 D x2 D x. If the alloy is ordered, then x1 ¤ x2 . The
difference

 D x1  x2 (12.1)

is therefore a measure of the degree of ordering of the alloy. The quantity , called
the order parameter, vanishes in the disordered phase and is nonzero in the ordered
phase. Notice that .x1 C x2 /=2 D x and therefore
 
x1 D x C and x2 D x  : (12.2)
2 2
Figures 12.1b and 12.2b show the order parameter of two binary alloys as a
function of temperature. In the cadmium-magnesium alloy the order parameter
decreases and shows a jump at the transition temperature to the disordered phase,
characterizing a discontinuous transition. The ordered and disordered alloys coexist
at the transition. In the copper-zinc alloy on the other hand, the order parameter
decreases continuously and vanishes at the critical temperature, characterizing a
continuous transition.
226 12 Order-Disorder Transition

(a) 500 (b) 1

β β +γ
480 0.8
α+β
θ ( C)

460 0.6
η
o

,
,
440
α +β
,
β β +γ 0.4

420 0.2

400 0
0.44 0.46 0.48 0.50 150 200 250 300 350 400 450 500
x θ ( C)
o

Fig. 12.2 Binary alloy copper-zinc. (a) Experimental phase diagram temperature  versus mole
fraction x of zinc around the order-disorder transition, represented by the thick line, between the
ordered phase ˇ 0 and the disordered phase ˇ. The symbols represent the experimental data. (b)
Order parameter  as a function of temperature  along x D 0:469. The order parameter vanishes
continuously at the critical temperature, c D 467 ı C

Gibbs Free Energy

The thermodynamic space related to the molar Gibbs free energy g of a two
component system consists of the temperature T, the pressure p and the mole
fraction x of the second component. The variables conjugated to these quantities
are respectively the molar entropy s, the molar volume v and chemical potential
which is the difference between the chemical potentials of the components.
However, to treat appropriately the ordered alloys, it is desirable to extend the
thermodynamic space. To this end, let us examine the simplest case of a binary alloy
A-B whose atoms are arranged in two sublattices that we call 1 and 2. A possible
ordered structure is that in which the atoms of type A is preferably located in the
sublattice 1. On the other hand, an equally possible structure is that in which the
A atoms are located preferably in the sublattice 2. Therefore, the system is capable
of displaying two different thermodynamic states, that has to be understood as two
ordered phases in coexistence. We remark that the two ordered phases in coexistence
have the same composition. Therefore, unlike what happens with disordered phases
in coexistence, the ordered phases in coexistence can not be distinguished by the
composition.
To properly describe the coexistence of two ordered phases it is necessary to use
another variable since the composition, described by the mole fraction x of one of
the component, does not distinguish the two ordered phases. The quantity that can
distinguish these phases is the order parameter defined above. Thus we add the order
parameter  to the thermodynamic space which earns an extra dimension. The molar
Gibbs free energy then depends also on  so that

dg D Tds C vdp C dx C  d; (12.3)

12.2 Bragg-Williams Theory 227

where  is the variable conjugate to the order parameter. The chemical potential
 is a thermodynamic field that breaks the symmetry between the sublattices.
An equivalent description is obtained by using the variables x1 and x2 , which are
the fractions of atoms of type A in sublattice 1 and 2, respectively, in place of x and
. In this representation we have

dg D Tds C vdp C 1 dx1 C 2 dx2 ; (12.4)

where 1 and 2 are conjugated to x1 and x2 , respectively. Using (12.2) in (12.4)

we obtain the following relations between the chemical potentials

1
D 1 C 2 and  D . 1  2 /: (12.5)
2
We notice that the properties of the ordered alloy are invariant by the permutation
of the labels of the sublattice. As a consequence, the Gibbs free energy g.T; p; x1 ; x2 /
must be invariant by the permutation of x1 and x2 . Therefore, the phase diagram is
invariant by the permutation of 1 by 2 , so that the coexistence of the two ordered
phases occurs when 1 D 2 , that is, when  D 0.

12.2 Bragg-Williams Theory

Gibbs Free Energy

The orderly arrangement of atoms in an alloy is a direct result of the forces of

attraction and repulsion between atoms. The Bragg-Williams theory considers that
the ordering of an alloy is established due to the greater attractive force between
atoms of different kind than that between atoms of the same type. To understand
the consequences of this idea, let us examine the simple binary alloy in which the
atoms are located on a bcc lattice, composed of two interpenetrating sublattices,
one formed by the vertices of the cube and the other by the centers of the cubes.
Each site belonging to a sublattice has as neighbors only sites belonging to the other
sublattice. Assuming that the dominant interaction is the one that occurs between
atoms located at neighboring sites, then the relevant interaction occurs between
atoms located in different sublattice. But, in accordance with Bragg-Williams, the
energy of interaction between atoms of different kind must be less than that between
atoms of the same type. Therefore, the local configuration of lowest energy is the
one in which one atom of one type are surrounded only by atoms of distinct type,
which are on another sublattice. This favors the segregation of atoms of one type of
atoms in a sublattice and another type in another sublattice, constituting an ordered
alloy.
228 12 Order-Disorder Transition

According to the Bragg-Williams theory the molar energy of a binary alloy has
the following form

u D uAA x1 x2 C uAB .x1 y2 C y1 x2 / C uBBy1 y2 ; (12.6)

where x1 and x2 are the mole fraction of component A of the alloy in sublattice 1
and 2, respectively; and y1 D 1  x1 and y2 D 1  x2 are the mole fraction of
component B of the alloy in sublattice 1 and 2, respectively. The parameters uAA ,
uBB and uAB are related, respectively, to the interaction between two atoms of type
A, two atoms of type B and two atoms of distinct types. Suppose for a moment that
x1 C x2 D 1. If the atoms of the same type are all in the same sublattice (complete
ordering) then x1 D 1 and x2 D 0 or x1 D 0 and x2 D 1. In both cases u D uAB .
If half the atoms of one type is in one sublattice and the other half is in the other
sublattice (disorder) then x1 D x2 D 1=2 and therefore u D .uAA C 2uAB C uBB /=4.
Therefore, the ordering will be favored if uAB < .uAA C uBB /=2, that is, when the
quantity, defined by

w D uAA C uBB  2uAB; (12.7)

is positive, that is, w > 0.

The variation of energy that occurs when an alloy is made up from pure
substances is given by u D u  .u1 C u2 / where u1 and u2 are the molar energies
of the pure substances given by

1 1
u1 D uAA.x1 C x2 / and u2 D uBB .y1 C y2 /: (12.8)
2 2
Therefore, using the results (12.6) and (12.8), we get

1
u D  w.x1 y2 C y1 x2 /; (12.9)
2
where w is given by (12.7).
The second hypothesis of the Bragg-Williams theory concerns the entropy. We
assume that the change in entropy that occurs when we mix two metals A and B
to compose each one of the sublattices of the alloy is the same as that of an ideal
solution. Therefore, the change in molar entropy molar s of an alloy which orders
according to the two sublattices is given by

R
s D  fx1 ln x1 C y1 ln y1 C x2 ln x2 C y2 ln y2 g: (12.10)
2
The variation in the molar Gibbs free energy is g D h  Ts. On the other hand,
the variation in enthalpy is h D u C pv. Under ordinary pressures, the term
pv can be neglected so that g D u  Ts.
12.2 Bragg-Williams Theory 229

The Gibbs free energy g of the alloy is set up from

1 1
gD gA .x1 C x2 / C gB .y1 C y2 / C g; (12.11)
2 2
where the first two terms are the molar Gibbs free energies of the pure substances.
Using the results (12.9) and (12.10) and taking into account that g D u  Ts,
we get

1 1 1
gD gA .x1 C x2 / C gB .y1 C y2 /  w.x1 y2 C y1 x2 /
2 2 2
RT
C fx1 ln x1 C y1 ln y1 C x2 ln x2 C y2 ln y2 g: (12.12)
2
The function g can also be written in the form

b
g D wx1 x2 C .x1 C x2 / C a
2
RT
C fx1 ln x1 C .1  x1 / ln.1  x1 / C x2 ln x2 C .1  x2 / ln.1  x2 /g; (12.13)
2
where b D gA  gB  w and a D gB .
The chemical potential 1 D @g=@x1 and 2 D @g=@x2 are given by

b RT x1
1 D wx2 C C ln (12.14)
2 2 1  x1

and
b RT x2
2 D wx1 C C ln : (12.15)
2 2 1  x2

Two Sublattices

The two symmetric ordered phases occur when the chemical potential  D 1 
2 , conjugated to the order parameter  D x1 x2 , vanishes. In this case 1 D 2 D
=2 and using the relations (12.14) and (12.15), we obtain

 x C 2
D 2w.x  / C b C RT ln  (12.16)
2 1x 2
230 12 Order-Disorder Transition

and

 x  2
D 2w.x C / C b C RT ln : (12.17)
2 1  x C 2

These two equations determine x and  as functions of T and .

Eliminating from these two equation we get

.x C 2 /.1  x C 2 /
2w D RT ln : (12.18)
.x  2 /.1  x  2 /

The coexistence of the ordered phases corresponds to a nonzero solution for . This
solution is monotonically decreasing with temperature and vanishes continuously.
The critical line is defined by the limit  ! 0 of this solution. Dividing both sides
of (12.18) by  and taking this limit we obtain

RT
2w D ; (12.19)
x.1  x/
or

RT D 2wx.1  x/; (12.20)

which is the equation that describes the critical line in the diagram temperature
versus mole fraction, shown in Fig. 12.3. Therefore, the Bragg-Williams theory
predicts a continuous transition occurring along a line of critical points in the
diagram temperature versus mole fraction. Such a line is observed experimentally
in the copper-zinc alloy in the transition between the ordered phase ˇ 0 -brass and the
disordered phase ˇ-brass, as shown in Fig. 12.2a.
Next we determine the behavior of the order parameter near the critical line. For
simplicity we will consider the case of an alloy with the same concentrations of the

Fig. 12.3 Phase diagram of 1.2

the binary alloy A-B in a bcc
lattice according to the 1.0 (A,B)
Bragg-Williams theory. The
transition between the 0.8
ordered and disordered
T/T0

phases is continuous. The 0.6

temperature T0 is defined by
T0 D w=2R 0.4 AB

0.2

0
0 0.2 0.4 0.6 0.8 1
A x B
12.2 Bragg-Williams Theory 231

two components, that is, for x D 1=2. In this case, (12.18) reduces to

1C
w D RT ln ; (12.21)
1

equation that yields  as a function of temperature, which was obtained originally

by Gorsky.
The behavior of  in the neighborhood of the critical point is obtained by the
expansion of the right hand side of (12.21) up to cubic terms in ,

3
.w  2RT/ D 3 : (12.22)
2RT

If T > Tc D w=.2R/ there is only the solution  D 0. For T < Tc the nonzero
solutions are
s
Tc  T
D˙ 3 : (12.23)
Tc

Near the critical point, the order parameter behaves therefore as

  .Tc  T/ˇ : (12.24)

with ˇ D 1=2. It is possible to show from (12.18) that this behavior, with ˇ D 1=2,
is valid not only at x D 1=2 but for any value of x between 0 and 1. That is, the
critical behavior predicted by the Bragg-Williams theory is the same around any
point of the critical line.
Experimentally, one verifies that the exponent ˇ has values distinct from 1/2.
Table 12.1 shows the values of this exponent and other critical exponents for the
order-disorder transition between the ˇ 0 -brass and ˇ-brass.

Four Sublattices

The Gibbs free energy built for two sublattices can be generalized to four sublattices.
In this case, in addition to temperature and pressure, g depends on the mole fractions
x1 , x2 , x3 and x4 of component A in each of four sublattices. The molar Gibbs free

Table 12.1 Critical Liga ˛ ˇ 

exponents related to the
order-disorder transition of CuZn 0:11 0:31 1:25
binary alloys obtained Fe3 Al 0:31
experimentally FeCo 0:30
232 12 Order-Disorder Transition

energy is given by

g D w.x1 x2 C x1 x3 C x1 x4 C x2 x3 C x2 x4 C x3 x4 / C b.x1 C x2 C x3 C x4 / C aC

RT
C fx1 ln x1 C .1  x1 / ln.1  x1 / C x2 ln x2 C .1  x2 / ln.1  x2 /C
4

C x3 ln x3 C .1  x3 / ln.1  x3 / C x4 ln x4 C .1  x4 / ln.1  x4 /g: (12.25)

The relation of the mole fractions of the sublattices with the mole fraction x is given
by

x1 C x2 C x3 C x4
xD : (12.26)
4
From g we get the chemical potentials

RT x1
1 D w.x2 C x3 C x4 / C b C ln ; (12.27)
4 1  x1
RT x2
2 D w.x1 C x3 C x4 / C b C ln ; (12.28)
4 1  x2
RT x3
3 D w.x1 C x2 C x4 / C b C ln ; (12.29)
4 1  x3
RT x4
4 D w.x1 C x2 C x3 / C b C ln : (12.30)
4 1  x4

Let us now consider the case where one of the components preferably lies in
a sublattice and the other component, preferably in the other three sublattices.
Assuming that these three sublattices are 2, 3 and 4, then x2 D x3 D x4 , so that
2 D 3 D 4 . Therefore

RT x1
1 D 3wx2 C b C ln ; (12.31)
4 1  x1
RT x2
2 D w.x1 C 2x2 / C b C ln : (12.32)
4 1  x2

The condition for coexistence is 1 D 2 D =4. On the other hand

x1 C 3x2
xD ; (12.33)
4
which combined with the relation that defines the order parameter

 D x1  x2 ; (12.34)
12.2 Bragg-Williams Theory 233

yields

3 
x1 D x C and x2 D x  : (12.35)
4 4
Therefore

3 x C 3
4
D 4w.3x  / C 4b C RT ln ; (12.36)
4 1  x  3
4

 x  4
D 4w.3x C / C 4b C RT ln : (12.37)
4 1  x C 4

Eliminating from these two equations we get

3
.x C 4
/.1  x C 4 /
4w D RT ln  3
: (12.38)
.x  4 /.1  x  4 /

Let us now consider the case where one of the components is found preferably
and equally in two of the sublattices and the other component preferably and equally
in the other two sublattices. Assuming that the first two sublattices are 1 and 3 and
the other two are 2 and 4, then x1 D x3 and x2 D x4 so that 1 D 3 and 2 D 4 .
Therefore,

RT x1
1 D w.2x2 C x1 / C b C ln ; (12.39)
4 1  x1
RT x2
2 D w.2x1 C x2 / C b C ln : (12.40)
4 1  x2

When there is coexistence, then 1 D 2 D =4. Using relations (12.1)

and (12.2), we get

 x C 2
D 4w.3x  / C 4b C RT ln ; (12.41)
2 1  x  2
 x  2
D 4w.3x C / C 4b C RT ln : (12.42)
2 1  x C 2

Eliminating from these two equations we get

.x C 2 /.1  x C 2 /
4w D RT ln : (12.43)
.x  2 /.1  x  2 /

The solutions of (12.38) and (12.43) allow the construction of the phase diagram
shown in Fig. 12.4. The transitions between the phases are all discontinuous except
234 12 Order-Disorder Transition

Fig. 12.4 Phase diagram of a 2.4

binary alloy A-B in a fcc (A,B)
lattice according to the 2.0
Bragg-Williams theory,
obtained originally by 1.6
Shockley [139]. The

T/T0
transitions between phases 1.2
are discontinuous except that AB
occurring along the x D 0:5 0.8
which is continuous. The A3B AB3
temperature T0 is defined by 0,4
T0 D w=2R
0
0 0.2 0.4 0.6 0.8 1
A x B

that occurring along the x D 0:5 which is a continuous transition. The phase
diagram predicted by the Bragg-Williams theory is distinct from those observed
experimentally for binary alloys that are ordered into four sublattices, as that shown
in Fig. 12.1. Results closer to the experimental data, however, can be obtained from
more elaborate theories.

12.3 Landau Theory

Spontaneous Symmetry Breaking

Landau theory deals with the transition between phases which have different sym-
metries. More specifically between a phase of higher symmetry, usually occurring at
high temperatures, called disordered phase to a lower symmetry one called ordered
phase. The transition to the ordered phase corresponds to a symmetry breaking as
some symmetries existing in the disordered phase will be absent in the ordered
phase. Since the symmetry breaking is not induced by any external field, the
transition is associated with a spontaneous breaking of symmetry.
The ordered phase is actually made up of several distinct ordered phases in
coexistence which have similar structures. The structures are not arbitrary but
are connected to each other by certain symmetry operations which comprises a
group of symmetry, which we denote by G . The symmetry operations of the group
G transform an ordered state into another, thus generating all ordered states in
coexistence from one of them. The disordered phase, on the other hand, is invariant
under the transformations of the group G . According to Landau theory all ordered
states have the same Gibbs free energy which results in the following rule: the Gibbs
free energy is invariant under the symmetry operations of the group G .
To describe the ordered phase we introduce an order parameter: a thermodynamic
variable that takes nonzero values in the ordered phase and vanishes in the
12.3 Landau Theory 235

disordered phase. To address the possibility of the occurrence of distinct ordered

states, we assume that the order parameter has several components which define a
vector space whose vectors we denote by E. Thus, each of the coexisting ordered
states is represented by a point in that space. The positions of representative points
of the ordered states form a geometric figure that is invariant by the symmetry
operations of the group G . The disordered phase is represented by E D 0. The
space vector is constructed so that its dimension is the same as that of the geometric
figure representing the ordered states. The thermodynamic state associated with the
molar Gibbs free energy is thus defined by the components i of the vector E in
addition to the usual variables such as temperature, pressure and mole fractions.
The thermodynamic field hi conjugated to the component i of the order
parameter is given by hi D @g=@i , where g.E / is the molar Gibbs free energy. This
field, when nonzero, breaks the symmetry of the disordered phase and destroys the
coexistence of ordered phases by favoring one of them. In general this field, which
breaks the symmetry, is not accessible experimentally. The order-disorder transition
described by the Landau theory, wherein the symmetry breaking is spontaneous,
corresponds to the case where this field is identically zero. Thus, the molar Gibbs
free energy is such that

@g
D 0: (12.44)
@i

Since g must be a convex function of i then condition (12.44) means that the
ordered and disordered phases correspond to the minima of g with respect to the
order parameter.
The Landau theory for the order-disorder transition assumes that g.E / is a
polynomial in i . In many applications it is enough to assume that the polynomial is
of degree four, that is,

/ D a0 C a2 I2 .E
g.E / C a3 I3 .E
/ C a4 I4 .E
/; (12.45)

where In .E/ is a monomial or a sum of monomials of degree n in i and the

coefficients an are dependent of other thermodynamic variables with the exception
of i . Since g.E
/ is invariant under the symmetry operations of group G , the same
occurs with In .E
/ which for this reason is called invariant of degree n.
The linear term is absent because the invariant of the first degree is nonexistent.
The invariant of second order, on the other hand, always exists and is given by
X
I2 D 2i ; (12.46)
i

ensuring that E D 0 is a minimum of g for a2 > 0. Assuming that a2 varies with

temperature, other minima of g arise leading to the ordered states. We assume that
a2 varies linearly with temperature and vanishes at a temperature T0 . That is, we
236 12 Order-Disorder Transition

assume that a2 D A.T  T0 / with A > 0 so that the disordered phase occurs at high
temperatures.

Continuous Transition

Let us first examine a binary alloy A-B whose atoms can be ordered in a crystalline
structure composed by two sublattices. The ordered alloy is such that the mole
fractions x1 and x2 of component A of the alloy related to sublattices 1 and 2,
respectively, are distinct. An ordered state is such that the atoms of type A are
preferably found in sublattice 1 and is characterized by x1 > x2 . The other ordered
state is the one in which the atoms of type A are preferably found in the sublattice
2. This state is characterized by x2 > x1 . One state can be generated from the other
by a spatial translation which in this case is equivalent to the exchange of atoms
between the sublattices 1 and 2. Therefore, the Gibbs free energy is invariant by the
exchange of x1 by x2 .
Using the order parameter  defined by

 D x1  x2 (12.47)

we see that  > 0 characterizes an ordered state and  < 0 characterizes the other
ordered state. The permutation of x1 by x2 is equivalent therefore to the permutation
of  by  and therefore g./ must be invariant under this permutation, that is,

g./ D g./: (12.48)

According to Landau theory we postulate that g is a polynomial of fourth degree

in . Using the property (12.48) we conclude that the invariants of odd degree are
absent, which allows us to write

g D a0 C a2 2 C a4 4 : (12.49)

Furthermore, still according to Landau theory, the coefficient a2 is positive in the

disordered phase and negative in the ordered phase, implying

a2 D A.T  Tc /; (12.50)

where A > 0. The coefficient a4 is considered to be positive.

The consequences of the result (12.50) are illustrated in Fig. 12.5. For tempera-
tures above Tc , the function g./ is convex and has a single minimum at  D 0. For
temperatures below Tc , the function g./ is no longer convex and has two symmetric
minima. In this case we consider the convex hull of g./. The appearance of two
minima is a manifestation of spontaneous symmetry breaking. It is important to
12.3 Landau Theory 237

(a) (b) g (c)

g g

η η η

Fig. 12.5 Molar Gibbs free energy g versus order parameter  for a binary alloy that is ordered
in two sublattices according to the Landau theory for a temperature above the critical temperature
(a), equal to the critical temperature (b), and below the critical temperature (c)

notice that as the temperature approaches the critical temperature, the minimum
continuously approaches zero.
The coexistence of two ordered phases occurs when the field h D @g=@ vanishes
which is equivalent to saying that the order parameter  corresponds to the minimum
of g. In fact, the convex hull has a horizontal line segment (corresponding to h D 0)
that connects the two minima. From

@g
D 2a2  C 4a4 3 D 0 (12.51)
@

we obtain the two minima

s
ja2 j
D˙ ; (12.52)
2a4

result valid for T < Tc . For T > Tc we get  D 0. Therefore, the order parameter
vanishes continuously and behaves in accordance with

  jT  Tc j1=2 : (12.53)

Discontinuous Transition

Next we study an example where the phase transition is discontinuous, that is, the
order parameter undergoes a jump in the passage from the order to the disorder
phase. We will see that this is a consequence of the existence of an invariant of third
order.
Let us examine a binary alloy A-B that orders in a structure composed by three
sublattices. We denote by x1 , x2 and x3 the mole fractions of component A in each
of the three sublattices. An ordered phase is that in which the atoms of type A are
238 12 Order-Disorder Transition

found preferably in sublattice 1 and indiscriminately in the other two sublattices.

Let us suppose that in this ordered state x1 D a and x2 D x3 D b with a > b. The
other two ordered states are characterized by x2 D a and x3 D x1 D b and by x3 D a
and x1 D x2 D b. The three ordered states are represented by three distinct points in
the space .x1 ; x2 ; x3 / which form an equilateral triangle.
Next we define a new coordinate system such that one axis is perpendicular and
the other two are parallel to the triangle representative of the ordered states. The one
corresponding to the perpendicular axis is defined by

1
xD .x1 C x2 C x3 /; (12.54)
3
which is the mole fraction of component A. The other two are defined by
p
1 3
 D x1  .x2 C x3 / and D .x2  x3 / (12.55)
2 2

and work as components of the vector order parameter .; /, since they vanish when
x1 D x2 D x3 . One of the ordered phase is defined by D 0 and  > 0. The other
are obtained by symmetry operations that leave the equilateral triangle invariant.
Using this symmetry operation, we can determine the invariants up to fourth order,
that are given by

I2 D 2 C 2 ; I3 D 3  3 2 ; I4 D .2 C 2 /2 : (12.56)

For temperatures high enough, g.; / is a convex function with a minimum at  D

D 0. For sufficient low temperatures, g.; / develops three minima that form an
equilateral triangle. In this case the function g.; / ceases to have the convexity
properties and we should consider the convex hull. The three minima correspond
to the three phases in coexistence. One of the minima occurs along D 0. Along
D 0 the molar Gibbs free energy is given by

g D a0 C a2 2 C a3 3 C a4 4 : (12.57)

Figure 12.6 shows the graph of g./ for several values of temperature. We recall that
a2 D A.T  T0 /, where A > 0 and that a4 > 0. For simplicity, the other coefficients
are considered constants.
For low temperatures, the absolute minimum corresponds to a nonzero value of
, characterizing an ordered phase. Raising the temperature, this minimum ceases
to be absolute. The local minimum at  D 0, corresponding to the disordered phase,
becomes then the absolute minimum. The transition, which is discontinuous, occurs
at temperature T1 which is determined by the condition g./ D g.0/, as seen in
Fig. 12.6b. Therefore, at this temperature g./ has the form

g./ D a0 C a4 2 .  b/2 ; (12.58)

12.3 Landau Theory 239

(a) (b) (c)

g g g

η η η

Fig. 12.6 Molar Gibbs free energy g versus the order parameter  for a binary alloy that orders in
three sublattices according to the Landau theory for a temperature above the critical temperature
(a), equal to the critical temperature (b), and below the critical temperature (c). The phase transition
is discontinuous

which, compared to (12.57), gives us a3 D 2ba4 and a2 D b2 a4 . Eliminating b,

we get

a23
a2 D : (12.59)
4a4

Therefore the transition temperature occurs at T1 D T0 C a23 =4Aa4 .

Four Sublattices

Let us study now a binary alloy A-B that orders in four sublattices. We denote by
x1 , x2 , x3 and x4 the mole fraction of component A in each of the four sublattices.
The ordered state in which the atoms A are found preferable in sublattice 1 is
characterized by x1 D a and x2 D x3 D x4 D b with a > b. The other three states are
characterized by x2 D a and x3 D x4 D x1 D b, by x3 D a and x4 D x1 D x2 D b,
and by x4 D a and x1 D x2 D x3 D b. The representative points of the ordered state
make up a tetrahedron in space .x1 ; x2 ; x3 ; x4 /. Next we use a new coordinate system
defined by

1
xD .x1 C x2 C x3 C x4 /; (12.60)
4
which is the mole fraction of component A, and by

 D x1 C x2  x3  x4 ; (12.61)
D x1  x2 C x3  x4 ; (12.62)
 D x1  x2  x3 C x4 : (12.63)
240 12 Order-Disorder Transition

The variables , and  works as components of the order parameter since they
vanish when x1 D x2 D x3 D x4 .
Using the symmetry operation that leaves the tetrahedron invariant we can obtain
the invariants of second, third and fourth order. The invariant of second order is
given by

I2 D 2 C 2 C  2 (12.64)

and that of third order by

I3 D  : (12.65)

There are two invariants of fourth order:

I41 D .2 C 2 C  2 /2 (12.66)

and

I42 D 4 C 4 C  4 : (12.67)

The molar Gibbs free energy is given by

g D a0 C a2 I2 C a3 I3 C a41 I41 C a42 I42 : (12.68)

For temperatures sufficiently small g.; ; / has four minima corresponding to

the four ordered states. One of them occurs along  D D . Along this line the
molar Gibbs free energy is given by

g D a0 C 3a2 2 C a3 3 C .9a41 C 3a42 /4 : (12.69)

The presence of the third order invariant implies a discontinuous phase transition.
Chapter 13
Magnetic Systems

13.1 Magnetic Materials

Introduction

The most evident magnetic materials are iron (Fe), cobalt (Co) and nickel (Ni). At
room temperature they exhibit a natural magnetism or spontaneous magnetization
and therefore are used as permanent magnets. If heated, however, such materials lose
their natural magnetization at a given temperature, called critical temperature Tc or
Curie temperature, and become paramagnetic. The iron loses its natural magnetism
at 770 ı C, cobalt at 1122 ı C and nickel at 358 ı C. Many compounds of the iron
group metals, for example, MnSb, CrTe and CrO2 are also ferromagnetic at room
temperature.
Certain materials, such as gadolinium (Gd) and the compounds CrBr3 , EuO
and EuS, are paramagnetic already at room temperature. To acquire spontaneous
magnetization, it is necessary to lower the temperature. Gadolinium becomes ferro-
magnetic below 20 ı C. The compounds CrBr3 , EuO and EuS become ferromagnetic
at low temperatures.
Magnetite (Fe3 O4 ), mineral known since ancient times by the magnetic prop-
erties as well as other iron oxides with the natural spinel crystal structure, such
as MgFe2 O4 , MnFe2 O4 , NiFe2 O4 , CoFe2 O4 and CuFe2 O4 , also exhibit a natural
magnetization. Although presenting a spontaneous magnetization, these oxides,
called ferrites, are not exactly ferromagnetic but ferrimagnetic. In any event, at suffi-
ciently high temperatures they lose their natural magnetism becoming paramagnetic.
Magnetite loses magnetization at 585 ı C. Another class of ferrimagnetic materials is
that formed by the iron oxides from the crystal structure of the garnet. An important
example of this category is the garnet of iron and yttrium, Y3 Fe5 O12 , whose Curie
temperature is 287 ı C
Other magnetic materials such as manganese (Mn) cease to be paramagnetic but
do not become ferromagnetic. They acquire magnetic states known generically as

© Springer-Verlag Berlin Heidelberg 2017 241

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_13
242 13 Magnetic Systems

antiferromagnetic states, which are characterized by the absence of spontaneous

magnetization. Manganese (Mn) becomes antiferromagnetic below 173 ı C. Like-
wise, the oxides FeO, MnO and Mn2 O3 and the fluorine compounds FeF2 , MnF2
and RbMnF3 are paramagnetic at room temperature, becoming antiferromagnetic at
low temperatures. On the other hand, chromium (Cr) and hematite (Fe2 O3 ) and the
oxides CoO and Cr2 O3 are antiferromagnetic at room temperature becoming param-
agnetic at higher temperature. Chromium becomes paramagnetic above 39 ı C. The
temperature below which the material becomes antiferromagnetic is called Néel
temperature, TN . Although the antiferromagnetic state do not present spontaneous
magnetization it should not be confused with the paramagnetic state.

Microscopic State

From the microscopic viewpoint each of the compounds mentioned above should
be understood as a large collection of microscopic magnetic dipoles which, in the
case of ionic compounds, are located at the magnetic ions. The magnetic ions are
the ions of transition metals, especially those of the iron group of iron, or the rare
earth ions. The interactions between the magnetic dipoles are the cause of several
types of magnetic ordering. These interactions are not properly magnetic, as one
might imagine at first, but originate from the Pauli exclusion principle of quantum
mechanics.
In the ferromagnetic state, the interactions between the dipoles favor the parallel
alignment of them. In the simplest antiferromagnetic state the interactions cause
an antiparallel alignment of dipoles. At high temperatures or if the interactions
between the dipoles are negligible, the thermal agitation causes the dipole moments
of ions to point in arbitrary directions in space giving rise to a magnetically
disordered state, with a zero total dipole moment. This state magnetically disordered
is called paramagnetic state. Thus, when increasing the temperature of the magnetic
substances mentioned above, they undergo a transition from a magnetically ordered
thermodynamic phase, which may be ferromagnetic, antiferromagnetic or other
more complex, to a magnetically disorderly thermodynamic phase, the paramag-
netic phase.

Paramagnetism

The paramagnetic state is characterized from the macroscopic point of view by the
linear response to an applied magnetic field. In absence of the field, a sample of
paramagnetic material does not exhibit magnetization. Applying a field the sample
acquires a magnetization that increases linearly with the field. If, on the other
hand, the field is reduced and vanishes, then the magnetization is also reduced and
vanishes. For small values of the applied field H, the magnetization m of a sample
13.1 Magnetic Materials 243

Table 13.1 Paramagnetic Substances C p

substances and their
respective Curie constants C, Ce(NO3 )3 5H2 O 7:17 2:39
in K A m2 /Tmol, obtained Pr2 (SO4 )3 16:4 3:62
experimentally. The Nd2 (SO4 )3 8H2 O 18:2 3:81
adimensional quantity
p p is Gd2 (SO4 )3 8H2 O 78:0 7:90
defined by p D 3RC= B Tb2 (SO4 )3 8H2 O 118:6 9:74
where B is the molar Bohr
Dy2 (SO4 )3 8H2 O 137:4 10:5
magneton. Source: AIP
Ho2 (SO4 )3 8H2 O 136 10:4
Er2 (SO4 )3 8H2 O 111:8 9:46
Tm2 (SO4 )3 63:3 7:11
Yb2 (SO4 )3 8H2 O 29:2 4:83
CrK(SO4 )2 12H2 O 18:4 3:84
MnF3 30:1 4:91
MnSO4 43:4 5:89
Fe(NH4 )(SO4 )2 12H2 O 43:9 5:92
Co(NO3 )2 6H2 O 25:8 4:54
NiSiF6 6H2 O 13:1 3:24
CuSO4 5H2 O 4:6 1:92

in the paramagnetic state is proportional to the field, that is,

m D 0 H; (13.1)

where 0 , the magnetic susceptibility, is positive. Increasing the field, the behavior
of m with H is no longer linear and at sufficiently high values of the field, the
magnetization saturates reaching a maximum value.
The susceptibility 0 depends on temperature T. For materials which are
paramagnetic at all temperatures, which we call ideal paramagnetism, it behaves
in accordance with the Curie law
C
0 D ; (13.2)
T
where C is a positive constant. Equation (13.1) together with (13.2) define thus from
the thermodynamic point of view, an ideal paramagnet. From the microscopic point
of view, an ideal paramagnetic material corresponds to a system consisting of non
interacting permanent magnetic microscopic dipoles. Examples of such materials
include the paramagnetic salts of the elements of the iron group and rare earths such
as those presented in Table 13.1. In these compounds the magnetic ions are far from
each another so that the interaction between them can be considered negligible.
244 13 Magnetic Systems

(a) 60 (b) 3.5

3.0
(Tmol/Am )
2

2.5

m (Am /mol)
40
2.0

2
1.5 0
1.0070 T
20
1.7775 T
-1

1.0
χ

0.5

0 0
300 400 500 600 0 100 200 300 400 500
o o
θ ( C) θ ( C)

Fig. 13.1 Experimental data for nickel obtained by Weiss and Forrer [148]. (a) Susceptibility at
zero field in the paramagnetic phase. The graph shows the inverse of the susceptibility  as a
function of temperature. The asymptote intercepts the axis of temperatures at 378 ı C, above the
Curie temperature that occurs at 357 ı C. (b) Molar magnetization m as a function of temperature
for various values of the applied magnetic field

For materials that have a paramagnetic phase, but are not ideal paramagnets, the
quantity 0 behaves, at sufficiently high temperatures according to the Curie-Weiss
law
C
0 D : (13.3)
T 
The constant  is positive for materials that undergo the transition to a ferromag-
netic state and is negative for those that undergo transition to an antiferromagnetic
state. The two constants C and  can be determined by fitting an asymptote to the
experimental data of 1=0 versus T, as shown in Fig. 13.1. The asymptote intersects
the axis of temperatures at T D  and the slope provides 1=C. The constant 
should not be confused with the transition temperature.

Ferromagnetism

Suppose that a ferromagnetic substance is subjected to a magnetic field. If the sub-

stance is heated while the field is maintained constant, then the magnetization of the
substance varies continuously with increasing temperature, as shown in Fig. 13.1b.
For any temperature below or above the critical temperature, the substance exhibits
a magnetization due to the presence of the field, as shown in Fig. 13.2. Then assume
that the field is turned off slowly at a given temperature, kept unchanged. If the
temperature exceeds the Curie temperature, T > Tc , the magnetization vanishes
linearly with the field, m D 0 H, characterizing the paramagnetic state. If the
temperature is lower, T < Tc , the magnetization does not disappear and the
13.1 Magnetic Materials 245

1.4 339.36
341.25
343.14
345.02
346.90
1.2 348.78
350.66
352.53
354.40
1.0 356.27
358.14
360.00
m (Am /mol)

0.8 361.84
2

365.36

0.6 369.07

372.77

376.47
0.4 380.15
383.83
387.50
391.16
394.81
0.2

0
0 0.5 1.0 1.5 2.0
H (T)
Fig. 13.2 Magnetization m versus field H for several temperatures below and above the critical
temperature obtained experimentally for nickel by Weiss and Forrer [148]. The temperatures are
given in ı C and the critical temperature occurs at 357:6 ı C. Above the critical temperature, m ! 0
when H ! 0. Below, m approaches the spontaneous magnetization when H ! 0

substance remains showing a magnetization even after the field has been turned
off, which for that reason is called the remanent magnetization. For small values of
the field the magnetization depends on the field according to

m D m C 0 H (13.4)

where the spontaneous magnetization m and the susceptibility 0 depends only on

temperature. The different behaviors, above and below the critical temperature can
be seen in Fig. 13.2, which shows isotherms obtained experimentally for nickel both
above and below the Curie temperature, occurring at 357 ı C.
246 13 Magnetic Systems

Table 13.2 Critical Material Tc M Ms b

temperature Tc , volumetric
(K) (kA/m) (kA/m)
magnetization of saturation,
Ms , and at 20 ı C, M , of Fe 1043 1714 1735 2:22
ferromagnetic materials. The Co 1395 1422 1445 1:71
number b of Bohr magnetons Ni 631 484 509 0:604
is defined by b D = B Gd 289 1980 7:12
where is the molar
magnetization of saturation. MnAs 318 670 3:40
Sources: AIP, LB, CRC MnBi 633 620 3:52
MnSb 587 710 3:53
MnB 578 152 1:92
CrTe 339 247 2:39
CrO2 393 515 2:07
MnP 298 1:20

At zero field, the remanent or spontaneous magnetization, varies with tem-

perature and vanishes continuously at the critical temperature. Above the critical
temperature it is identically zero. The behavior of the spontaneous magnetization
of nickel with the temperature can be seen in Fig. 13.1b. At low temperatures,
the magnetization m reaches its saturation value. Table 13.2 shows the values of
saturation of magnetization and the magnetization at room temperature for several
ferromagnetic materials.
The understanding of the ferromagnetism presents, sometimes, a situation which
seems paradoxical. As mentioned above, a sample of ferromagnetic material such
as iron must have a spontaneous magnetization below the critical temperature.
However, any sample of iron does not display in general the expected permanent
magnetization, which is achieved only after being magnetized, that is, after having
been subjected to a magnetic field. This can be explained assuming that the sample
is composed of several macroscopic magnetic domains, each possessing a magnetic
dipole in a certain direction. The dipoles of different domains point to arbitrary
directions which results in a vanishing total dipole. To restore the natural magnetism
of the sample it is necessary to subject it to a magnetic field, that is, to magnetize it.

13.2 Magnetic Thermodynamic Potentials

Magnetic Work

To study the effects of a magnetic field on a sample of magnetic material we use a

solenoid inside which we place the sample. The solenoid is traversed by an electric
current supplied by a battery, thus producing a magnetic field in its interior. The
13.2 Magnetic Thermodynamic Potentials 247

magnetic induction field B inside the solenoid is associated with the magnetic field
H and the magnetic dipole moment per unit volume or volumetric magnetization,
M of the sample by the relation

B D 0 .H C M /; (13.5)

where 0 is the permeability of vacuum. In the absence of the sample, the magnetic
induction field within the solenoid is given by

Be D 0 He : (13.6)

The local field H and the external field He are produced by the same free currents
that run through the solenoid. However, they may not be identical. For a uniform
external field, we can show that the infinitesimal magnetic work dWmag performed
on the magnetic sample is given by

dWmag D Be  dM; (13.7)

where M is the total magnetic dipole moment of the sample.

To show the result (13.7) we proceed as follows. For simplicity, we consider
a cylindrical solenoid of length ` and cross-sectional area equal to A and that the
sample fills the entire space within the solenoid. We also assume that H and M are
uniform and parallel to the axis of the solenoid and that edge effects are neglected.
The local field H inside the solenoid, which in this case coincides with the external
field, is given by
NI
HD ; (13.8)
`
where N is the number of spiral of the solenoid and I is the electric current through
the solenoid. The magnetic induction field B inside the solenoid varies in time so that
the potential difference V at the ends of the solenoid is equal to the time variation
of the flux of magnetic induction field. Since the total flow is NAB, then
dB
V D NA : (13.9)
dt
The power P transferred to the system solenoid-sample is P D V I. Using (13.8)
and (13.9) we reach the result
dB dB
P D H`A D VH ; (13.10)
dt dt
where V D `A is the volume of the solenoid and is also the volume of the sample.
Let us now determine the power transferred P0 to the solenoid when it is empty.
Using the same arguments, we obtain

dH
P0 D VH 0 : (13.11)
dt
248 13 Magnetic Systems

Therefore, the magnetic work rate dWmag =dt D PP0 performed only on the sample
will be
dWmag dM
D VH 0 ; (13.12)
dt dt

since B  0 H D 0 M . Taking into account that the total magnetic dipole of the
sample is M D VM then

dWmag dM
D 0 H : (13.13)
dt dt
Since in the present case the local field H and the external field He coincide then
Be D 0 He D 0 H and we arrive at the following result for the infinitesimal
magnetic work performed on the magnetic sample:

dWmag D Be dM: (13.14)

In the International System of Units, the magnetic dipole moment M is measured

in ampère square meter (A m2 ) and the magnetic induction fields B and Be are
measured in tesla (T). The magnetic fields H and He as well as the volumetric
magnetization M are measured in ampère per meter (A/m). The magnetic dipole
moment per mole, or molar magnetization, m is given in A m2 /mol.
A standard dipole moment widely used is the Bohr magneton defined as e„=2me ,
where e is the charge of the electron, „ is the Planck constant and me is the mass
of the electron. The Bohr magneton is almost identical to the dipole moment of the
free electron and is 9:2740154  1024 A m2 . Here it is useful to define the molar
Bohr magneton by B D .e„=2me/NA , where NA is the Avogadro constant. In the
International System of Units

B D 5:5849388 A m2 =mol: (13.15)

Thermodynamic Potentials

The infinitesimal change of internal energy dU of a magnetic system is equal to

the sum of heat introduced dQ D TdS and the work done on the magnetic system
dWmag D HdM, that is,

dU D TdS C HdM: (13.16)

To simplify notation we denote by H the external magnetic induction field Be ,

which we call external field or the applied field. Importantly, the total dipole
moment M must be understood as an extensive thermodynamic quantity and H
13.2 Magnetic Thermodynamic Potentials 249

as a thermodynamic field. Thus we understand that internal energy U of a simple

magnetic system is a function of the extensive variables S and M. More precisely
we consider that the internal energy is also a function of the number of moles N
so that U.S; M; N/ is the fundamental relationship of a simple magnetic system in
the representation of internal energy. The entropy S.U; M; N/ is also a fundamental
relationship for which we have
1 H
dS D dU  dM; (13.17)
T T
which we get directly from (13.16).
From U.S; M; N/, we obtain other representations by Legendre transforma-
tions. The magnetic enthalpy H .S; H; N/, the magnetic Helmholtz free energy
F.T; M; N/ and the magnetic Gibbs free energy G.T; H; N/ are defined by the
Legendre transformations

H D min.U  HM/; (13.18)

M

F D min.U  TS/; (13.19)

S

G D min.F  HM/: (13.20)

M

From them we get the results

dH D TdS  MdH; (13.21)

dF D SdT C HdM; (13.22)
dG D SdT  MdH: (13.23)

The thermodynamic potentials have the usual properties of convexity. In particular

U is a convex function of the extensive variables S and M, while G is concave
function of the thermodynamic fields T and H.

Thermodynamic Coefficients

From (13.23), it follows

   
@G @G
SD and MD : (13.24)
@T H @H T

Two thermodynamic coefficients are particularly important: the magnetic isothermal

susceptibility XT , defined by
 
@M
XT D ; (13.25)
@H T
250 13 Magnetic Systems

and the heat capacity at constant field CH , defined by

   
@S @H
CH D T D : (13.26)
@T H @T H

The convexity of G implies the results XT  0 and CH  0. Other thermodynamic

coefficients are: the heat capacity at constant magnetization
   
@S @U
CM D T D ; (13.27)
@T M @T M

and the magnetocaloric coefficient

 
@M
AD : (13.28)
@T H

The convexity of U implies CM  0.

Molar Quantities

The thermodynamic properties of a simple magnetic system can be obtained from

the fundamental relation u.s; m/, where u D U=N is the molar internal energy,
s D S=N is the molar entropy and m D M=N is the molar dipole moment or simply
molar magnetization. For this system, we have

du D Tds C Hdm: (13.29)

The molar entropy s.u; m/ as a function of u and m is also a fundamental relation

and
1 H
ds D du  dm: (13.30)
T T
Other molar quantities can also be defined that include the molar potentials h D
H =N, f D F=N and g D G=N, for which

dh D Tds  mdH; (13.31)

df D sdT C Hdm; (13.32)
dg D sdT  mdH: (13.33)

We also highlight the isothermal molar susceptibility  D XT =N, given by

 
@m
D (13.34)
@H T
13.3 Ideal Paramagnetic Systems 251

and the molar heat capacities cm D CM =N and cH D CH =N at constant

magnetization and field, respectively, given by
   
@s @u
cm D T D (13.35)
@T m @T m

and
   
@s @h
cH D T D (13.36)
@T H @T H

13.3 Ideal Paramagnetic Systems

Equation of State

In analogy with the properties of an ideal gas we postulate that for an ideal
paramagnetic system, the molar energy u.T/ is a function of temperature only.
Considering that 1=T D .@s=@u/ and H=T D .@s=@m/ we see that
   
@ 1 @ H
D : (13.37)
@m T @u T

The left hand side of this equation vanishes identically because, inverting u.T/,
we see that T is a function of u but not of m. As the right hand side must also
be identically zero, we conclude that H=T is independent of u, depending only on
m. We reached then the following result due to Langevin: the magnetization m of
a paramagnetic ideal system depends on temperature T and field H through the
combination H=T.
The result of Langevin leads directly to the Curie law. It suffices to recall that for
a paramagnetic system, the magnetization m is proportional to the field H for small
values of the field, that is, m D 0 H, where 0 depends on T. The dependence 0
with temperature must be such that m is a function of H=T. Therefore, it is necessary
that 0 D C=T so that

H
mDC : (13.38)
T

Equation (13.38) defines the equation of state m.H; T/ of an ideal paramagnetic

system, which is valid in the regime of small values of H. An equation of state valid
for any values of H is the one introduced by Langevin
 
H RT
m D coth  ; (13.39)
RT H
252 13 Magnetic Systems

where is the saturation magnetization and R is the universal gas constant.

The Langevin equation of state, however, is based on a classical description of
microscopic magnetic dipoles. A more appropriate description, which takes into
account the quantum aspects, is that given by the equation of state used by Brillouin.

Brillouin Theory

From the microscopic point of view an ideal paramagnetic solid comprises a set of
non-interacting magnetic dipoles located at the magnetic atoms. As the origin of the
microscopic magnetic dipole moment is the angular momentum of the electrons and
that it is quantized, we must presume that the same occurs with the dipole moment
of magnetic atoms. Assuming that the component of the moment of each dipole in
the direction of the applied field H takes discrete values and equally spaced, then in
accordance with Brillouin, the magnetization m is given by
 
2J C 1 2J C 1 H 1 1 H
mD coth  coth : (13.40)
2J 2J RT 2J 2J RT

The number of values that the component of the magnetic dipole can take is 2J C 1
and J can take one of the following values: 1=2; 1; 3=2; 2; 5=2; : : :. For small values
of the field, the magnetization behaves linearly with the field in accordance with

J C 1 2 H
mD : (13.41)
3J RT
Therefore the molar magnetic susceptibility at zero field behaves according to the
Curie law 0 D C=T where

J C 1 2
CD : (13.42)
3J R
It is usual to present the saturation magnetization in terms of the Bohr
magneton B as follows D b B . It is also customary to present the Curie constant
as follows

p2 2B
CD : (13.43)
3R
Therefore, the saturation magnetization and Curie constant can be equivalently
be given by the dimensionless quantities b and p. If the magnetization of a
paramagnetic substance follows p the Brillouin equation of state then these two
quantities are related by p D b .J C 1/=J.
Table 13.1 shows the experimental data of C obtained by the Curie law
for various paramagnetic substances and the respective values of dimensionless
13.3 Ideal Paramagnetic Systems 253

Fig. 13.3 Magnetization m 7

versus H=T obtained III
experimentally by Henry 6
[145] for three paramagnetic II
5
salts: (I) CrK(SO4 )2 12H2 O,
(II) Fe(NH4 )(SO4 )2 12H2 O

m/mB
4
and (III) Gd2 (SO4 )3 8H2 O
I
3
1.30 K
2 2.00 K
3.00 K
1 4.21 K
Brillouin
0
0 1 2 3 4
-1 -1
HT (TK )

quantity p. Figure 13.3 shows the magnetization of three paramagnetic salts as

functions of H=T. They are consistent with (13.40) corresponding to J D 3=2 for
chromium salt, J D 5=2 for iron salt and J D 7=2 for the gadolinium salt.

Case J D 1=2

Let us consider here the simplest case of an ideal paramagnetic system, which is the
one corresponding to J D 1=2. In this case the expression (13.40) is reduced to

H
m D tanh : (13.44)
RT
For small fields the magnetization behaves linearly with the field according to m D
2 H=RT and the susceptibility at zero field is in agreement with the Curie Law
0 D C=T with C D 2 =R. From the magnetization we get by differentiation the
molar susceptibility  D .@m=@H/T ,
 
2 H 2
D sech ; (13.45)
RT RT

and the molar magnetocaloric coefficient ˛ D .@m=@T/H ,

 
2 H H 2
˛D sech : (13.46)
RT 2 RT

Integrating m D .@g=@H/T in H, we get the molar Gibbs free energy

H
g D g0  RT ln.2 cosh /; (13.47)
RT
254 13 Magnetic Systems

where g0 depends on temperature only. The derivative of this expression with respect
to temperature gives the molar entropy

H H H
s D s0 C R ln.2 cosh / tanh ; (13.48)
RT RT RT

where s0 D dg0 =dT. To determine s0 we notice on one hand that at zero field,
s D s0 C R ln 2. On the other hand, at zero field the entropy must be the same for
any temperature and its value for one mole of dipoles, according to the microscopic
interpretation of entropy, must be equal to R ln 2 since each dipole moment takes
two values. From this result we conclude that s0 D 0 and that g0 is a constant which
we choose as being zero.
From the above results for g and s we note that the molar enthalpy h D g C Ts
is h D Hm which replaced in u D h C Hm gives u D 0. This result leads us
to conclude that the molar heat capacity at constant magnetization cm D .@u=@T/m
vanishes identically. The molar heat capacity at constant field cH D .@h=@T/m ,
however, is nonzero and is given by
 2
H H
cH D R sech : (13.49)
RT RT

Inverting expression (13.44), we get H as a function of m,

RT C m
HD ln ; (13.50)
2  m

from which we reach the molar Helmholtz free energy f by integration of H D

.@f =@m/T . Performing the integration, we reach the result
 
Cm Cm m m
f D RT ln C ln : (13.51)
2 2 2 2

where the constant of integration was fitted so that f calculated at m D 0 is equal to

g calculated at H D 0 since these two quantities are related by g D f  Hm.
The molar entropy is obtained from s D .@f =@T/m , which gives
 
Cm Cm m m
s D R ln C ln : (13.52)
2 2 2 2

From these two last results we see that f D Ts from which follows u D 0, as seen
before.
13.4 Weiss Theory 255

13.4 Weiss Theory

Spontaneous Magnetization

To explain the emergence of spontaneous magnetization in ferromagnetic materials,

Weiss imagined that each microscopic dipole magnetic material is subjected to the
action of the field produced by neighboring dipoles. Weiss assumed that this field
should be proportional to the magnetization m so that the total field on a dipole
would equal H C
m, where the parameter
> 0. Thus, even in the absence of
applied field H, the dipole could be subjected to a local field, called by Weiss the
molecular field, which would orient the dipole.
If in the equation of state of an ideal paramagnetic system we replace H by H C

m, we reach the equation of state of a system that undergoes a ferromagnetic-

paramagnetic transition. Using (13.44), we get

m D tanh. .H C
m//: (13.53)
RT
For zero applied field, H D 0, this equation becomes

m D tanh. m/: (13.54)

RT
To see that this equation has a nonzero solution, corresponding to a spontaneous
magnetization, let us resort to a graphic method. Using an auxiliary variable x D

m=RT, equation (13.54) can be written in the form

RT
x D tanh x: (13.55)

2
The function tanh x has the aspect shown in Fig. 13.4a. On the graph we draw
the straight line .RT=
2 /x. One solution is x D 0 and exists for any value of
temperature. It is the sole solution if the slope of the straight line is greater or equal
to the slope of tanh x at x D 0, that is, as long as RT=
2  1. This condition is
equivalent to T  Tc , where Tc is given by

2
Tc D : (13.56)
R
For T < Tc , there are two nontrivial solutions corresponding to the interception
of the straight line with the curve, as seen in Fig. 13.4a. The positive solutions are
presented in Fig. 13.4b for each value of temperature.
The value of the constant
can be determine from (13.56) as long as the
saturation magnetization is known. It is more interesting however to determine
the value of the Weiss molecular field HW D
D RTc = . To estimate HW we
assume that is of the order of magnitude of the Bohr magneton B . Therefore
256 13 Magnetic Systems

(a) 1.0 (b) 1

0.8 0.8

0.6 0.6

m/μ
m/μ

0.4 0.4

0.2 0.2

0 0
0 0.5 1.0 1.5 2.0 0 0.2 0.4 0.6 0.8 1.0 1.2
x T/Tc

Fig. 13.4 (a) Graphic solution of (13.55). The curve represents the function tanh x. The straight
lines correspond to the cases in which the temperature is greater, equal to or smaller than the critical
temperature Tc . (b) Magnetization corresponding to the graphic solution of (a) as a function of
temperature

HW  RTc = B and using the value of the critical temperature of iron, we get the
estimate HW  103 T.
Let us compare now HW with the field Hd created by a magnetic dipole of the
order of the Bohr magneton at a point situated at a distance a corresponding to the
spacing between magnetic atoms in a solid. Such a field is Hd D 107 2B =NA2 a3 .
Using the value for a equal to 2  1010 m, we obtain the estimate Hd  101 T.
Thus we see that Hd is much smaller than HW , which means that in a ferromagnetic
material the interaction between the magnetic dipoles cannot be of magnetic origin,
as we should expect. We conclude that ferromagnetism is due to a nonmagnetic
interaction between atoms. In fact, this interaction has quantum origin and is based
on the Pauli exclusion principle.

Free Energy

Inverting (13.53), we get

RT C m
H D 
m C ln ; (13.57)
2  m

which must be understood as an equation of state in the representation of the

Helmholtz free energy f .T; m/. The graph of H versus m, for several temperatures is
presented in Fig. 13.5. The spontaneous magnetization m obeys (13.57) for H D 0,
that is,

RT C m

m D ln ; (13.58)
2  m
13.4 Weiss Theory 257

(a) H (b) H (c) H

m m m

Fig. 13.5 External field H versus magnetization for a ferromagnetic system according to the Weiss
theory for a temperature above the critical temperature (a), equal to the critical temperature (b),
and below the critical temperature (c)

(a) (b) (c)

f f f

m m m

Fig. 13.6 Molar magnetic Helmholtz free energy f versus magnetization for a ferromagnetic
system according the Curie theory. For a temperature above the critical temperature (a), equal
to the critical temperature (b), and below the critical temperature (c)

which is equivalent to (13.54). The condition that gives the critical temperature Tc is
obtained when the nonzero solution m ! 0. Dividing both sides by m and taking
the limit, we get
D RTc = 2 , which is equivalent to (13.56).
Using H D .@f =@m/T and integrating (13.57), we get
 
m2 Cm Cm m m
f D 
C RT ln C ln : (13.59)
2 2 2 2 2
where we have chosen the integration constant in such a way that f .T; m/ reduces
to the expression (13.51) when
D 0.
Figure 13.6 shows f as a function of magnetization m for several temperatures.
For temperatures below the critical temperature, given by (13.56), f .m/ ceases to
have the convexity properties and in these cases we need to consider the convex
hull. The convex hull is obtained by the construction of the double tangent which
in the present case corresponds to a horizontal line segment. The double tangent
connects the two minima of f that occur at m D ˙m , where m is the positive root
of H.m / D 0. Thus we distinguish three regimes:
(a) When T > Tc , corresponding to Figs. 13.5a and 13.6a, (13.57) and (13.59) are
valid for any values of m. Since m vanishes linearly with H when H ! 0 then
we are facing a paramagnetic state.
258 13 Magnetic Systems

(b) When T D Tc , corresponding to Figs. 13.5b and 13.6b, (13.57) and (13.59) are
also valid for any values of m. In addition m vanishes with H when H ! 0, but
not linearly. We call this state critical.
(c) When T < Tc , corresponding to Figs. 13.5c and 13.6c, (13.57) and (13.59) are
valid only for jmj  m . For jmj < m , f is constant and H D 0. In this case,
when H ! 0 from positive values m ! m > 0. Therefore, we are facing a
ferromagnetic state and m is identified with the spontaneous magnetization.
The diagram H versus temperature T displays a coexistence line which occurs
along H D 0 and for T < Tc . On the line two thermodynamic phases coexist: one
corresponding to the magnetization m D m and the other corresponding to the
magnetization m D m .
The entropy s D .@f =@T/m is obtained from (13.59) and is given by
 
Cm Cm m m
s D R ln C ln : (13.60)
2 2 2 2

Replacing s in u D f C Ts and using (13.59), we obtain

m2
u D 
: (13.61)
2

Susceptibility and Heat Capacity

The susceptibility  is obtained from its inverse, given by 1= D .@H=@m/T .

Deriving the expression (13.57) with respect to m, we get

1 RT
D 
C 2 : (13.62)
  m2

Along H D 0 and for T > Tc , the magnetization vanishes yielding the result

1 RT
D 
C 2 : (13.63)


Using the result (13.56), Tc D

2 =R, we get

2
D ; (13.64)
R.T  Tc /

which is the Curie-Weiss law (13.3) with  D Tc and C D 2 =R.

Below the critical temperature and at zero field, the two phases are in coexistence
and we should consider a susceptibility of each one of them. However, both phase
have the same susceptibility. The susceptibility as a function of temperature is
13.4 Weiss Theory 259

(a) 20 (b) 2.0

15 1.5

cH/R
λχ

10 1.0

5 0.5

0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
T/Tc T/Tc

Fig. 13.7 Behavior of a ferromagnetic system at zero field according to the Weiss theory. (a)
Molar susceptibility  versus temperature T. (b) Molar heat capacity cH versus temperature

obtained by the replacement of the nonzero solution m of (13.58) in (13.62). The

result is shown in Fig. 13.7a.
To determine the molar heat capacity at constant field cH D .@h=@T/H , we need
the molar enthalpy h D u  Hm, which is

m2
h D 
 Hm: (13.65)
2
At zero field and for T > Tc , the magnetization vanishes, so that h D 0 and cH D
0. At zero field and for T < Tc , we should take into account that the phases are
found in coexistence. However, both phase have the same value of h, given by h D

.m /2 =2, so that

d
cH D  .m /2 : (13.66)
2 dT
Figure 13.7b shows the graph of cH versus T.
Next we determine the molar heat capacity at constant magnetization, cm D
.@u=@T/m , along H D 0. For T  Tc the magnetization vanishes so that the
result (13.61) gives u D 0, from which we conclude that cm D 0. For T < Tc ,
we should consider the molar heat capacity of each one of the coexisting phases.
However, they have the same molar energy given by u D 
.m /2 =2, which
becomes equal to the molar enthalpy h. Therefore, cm D cH along H D 0.
The equality between the molar heat capacities along H D 0 can be obtained
directly from the following relation between them
 2
@H
cH D cm C T : (13.67)
@T m
260 13 Magnetic Systems

It suffices to take into account that .@H=@T/m D 0 along H D 0. This behavior is

distinct from that related to fluids for which cp and cv are different along the critical
density.

Critical Point

Near the critical point the magnetization is small and therefore it can be determined
by an expansion of the right hand side of (13.57) in powers of m. Up to cubic terms
in m, we get

T
3
HD.  1/
m C m: (13.68)
Tc 3 2

The nonzero root of H.m/ D 0 is given by

T
.m /2 D 3 2 .1  /; (13.69)
Tc

from which we get the positive root

s
T
m D 3.1  /: (13.70)
Tc

To determine the behavior of  just below the critical temperature, we use

result (13.69) in (13.62) to obtain

2
D : (13.71)
2R.Tc  T/

To determined the behavior of cH just below the critical temperature, we replace

result (13.69) in (13.66) to conclude that cH D 3R=2 what means that cH has a jump
in Tc as seen in Fig. 13.7b. To obtain the linear behavior seen in the same figure we
should use the expression of .m /2 up to quadratic terms in .Tc  T/. Using such an
expression we reach the result

3 12 T
cH D R C R.  1/: (13.72)
2 5 Tc

Anisotropy

The results obtained so far are valid for isotropic magnetic systems and, therefore,
such that the spontaneous magnetization can appear in any direction. The magnetic
13.4 Weiss Theory 261

system, however, can be anisotropic so that the spontaneous magnetization may

appear only in certain crystalline directions. In systems with uniaxial anisotropy, for
example, it appears along the anisotropic axis. If the magnetic field is applied along
this preferential direction, the free energy (13.59) and other results obtained above
remain valid. But, if the field is applies in another direction, the free energy is no
longer appropriated. In these cases, we adopt the following molar Helmholtz free
energy

 2

f D .mx C m2y /  m2z C

2 2
 
Cm Cm m m
C RT ln C ln ; (13.73)
2 2 2 2

where mx , my and mz are the Cartesian components of the magnetization and m D

.m2x C m2y C m2z /1=2 . This free energy describes a ferromagnetic system with uniaxial
anisotropy, whose preferential axis extends along the z direction. The parameter 
is a measure of the anisotropy with respect to the preferential axis z. The value
 D 1 corresponds to the isotropic case and  D 0 to the extreme anisotropy.
We consider  < 1, so that the spontaneous magnetization will occur along the
preferential direction z.
In the present case the following relation is valid

df D sdT C Hx dmx C Hy dmy C Hz dmz ; (13.74)

where Hx , Hy and Hz are the components of the external field. From f we found the
equations of state

RT mx Cm
Hx D 
 mx C ln. /; (13.75)
2 m m
RT my Cm
Hy D 
 my C ln. / (13.76)
2 m m
RT mz Cm
Hz D 
mz C ln. / (13.77)
2 m m

If the external field is applies along the preferential direction z, that is, if Hx D
Hy D 0, then mx D my D 0 and equation (13.77) reduces to equation (13.57).
The results obtained previously become valid for this case and, in particular, the
susceptibility obtained previously becomes what we call parallel susceptibility.
Next we examine the case where the field is applied in a direction perpendicular
to the preferential axis z, that is, such that Hy D Hz D 0 and Hx ¤ 0. We wish to
determine the perpendicular susceptibility ? D @mx =@Hx at zero field. To this end
we examine (13.75) for small values of Hx . In this case, mx will also be small and
will vanish when Hx ! 0 even in the ferromagnetic phase since the spontaneous
262 13 Magnetic Systems

magnetization emerges only in direction z. For small values of m (13.75) reduces to

RT mx C m
Hx D 
 mx C ln. /; (13.78)
2 m  m

in which we set mz D m .
Above the critical temperature m D 0 and we get

T
Hx D 
 mx C
mx ; (13.79)
Tc

where we used the result (13.56). Below the critical temperature, we use the
result (13.58) to obtain

Hx D 
 mx C
mx ; (13.80)

From these equations we reach the expressions for the perpendicular susceptibility
at zero field
Tc
? D ; (13.81)

.T   Tc /

valid above the critical temperature and

1
? D ; (13.82)

.1   /

valid below the critical temperature. Therefore, the perpendicular susceptibility is

finite and constant below the critical temperature.

13.5 Criticality

Critical Exponents

In the neighborhood of the critical point, the thermodynamic quantities have a

singular behavior. The molar susceptibility  has a divergent behavior and the spon-
taneous magnetization m vanishes T ! Tc , at zero field. The results of the previous
section, obtained from the Weiss theory, describe the ferromagnetic-paramagnetic
transition and in particular predict the occurrence of these singularities. The results
predicted by the Weiss theory however show deviations from the experimental
results as seen in Fig. 13.3a where 1 deviates from the linear behavior predicted
by (13.64). The experimental data show that only at high temperature this linear
behavior occurs, what constitutes the Curie-Weiss law.
13.5 Criticality 263

Table 13.3 Critical Ferromagnets ˛ ˇ  ı Tc

exponents and critical
temperature, in K, of Fe 0:12 0:39 1:34 4:2 1043
ferromagnetic materials Ni 0:10 0:38 1:34 4:6 631
obtained experimentally Co 0:42 1:23 1395
CrBr3 0:37 1:22 4:3 32.7
EuO 0:04 0:37 1:40 4:5 69.4
EuS 0:13 0:36 1:39 16.4
Cu(NH4 )2 Cl4 2H2 O 1:37 0.701
CuK2 Cl4 2H2 O 1:36 0.877
Cu(NH4 )2 Br4 2H2 O 0:38 1:33 4:3 1.83
CuRb2 Br4 2H2 O 0:37 1.87
CuCs2 Cl4 2H2 O 1:33

To describe appropriately the behavior near the critical point we assume that the
spontaneous magnetization vanishes according to

m  .Tc  T/ˇ ; (13.83)

the susceptibility at zero field, H D 0, diverges according to

  jT  Tc j ; (13.84)

and along the critical isotherm, T D Tc , the magnetization and the field are related
by

H  mı ; (13.85)

where ˇ,  and ı are certain critical exponents. The results (13.70), (13.64), (13.71)
and (13.68), obtained from the Weiss theory, predicts the following values for the
critical exponents: ˇ D 1=2,  D 1 and ı D 3. The experimental data on the other
hand give distinct values for these exponents, as can be seen in Table 13.3 for several
magnetic materials.
Experimentally, one also observes that the molar heat capacity cH at zero field
has singular behavior. This quantity may diverge according to

cH  jT  Tc j˛ ; (13.86)

where ˛ is positive. It is possible that cH does not diverge although has a singular
behavior. In this case, to characterize appropriately the singularity we assume the
following behavior

cc  cH  jT  Tc j˛ ; (13.87)
264 13 Magnetic Systems

(a) 42 (b) 2

40
cp (J/Kmol)

ln|dcp/dT|
38 0

36

34 -2

32

30 -4
580 600 620 640 660 680 -2 0 2 4

T (K) ln|T-Tc|

Fig. 13.8 (a) Molar isobaric heat capacity cp of nickel obtained experimentally by Connelly et al.
[142], as a function of temperature T. (b) Log-log plot of dcp =dT versus T. The slope of the straight
line fitted to the experimental data gives 0:90 which yields an exponent ˛ D 0:10

with ˛ negative, where cc is the finite value of cH at the critical point. Figure 13.8
shows the molar heat capacity of nickel whose exponent ˛ is negative. To determine
the exponent, it is convenient to use the quantity dcH =dT to eliminate the constant
cc . According to (13.87) this quantity behaves as jT  Tc j1˛ .
A singularity characterized by an exponent ˛ D 0 must be described by a more
specific form. It can be a divergence of the logarithm type

cH  ln jT  Tc j; (13.88)

but can also corresponds to a jump in cH , as predicted by the Weiss theory and
shown in Fig. 13.7b.
The critical exponents ˛, ˇ,  and ı are not independent but keep certain relations
such as the Rushbrooke relation,

˛ C 2ˇ C  D 2; (13.89)

the Griffiths relation,

˛ C ˇ.ı C 1/ D 2; (13.90)

and the Widom relation,

 D ˇ.ı  1/: (13.91)

The relations between the exponents can be demonstrated from the Widom scaling
theory, as we will see below.
13.5 Criticality 265

Singular Part

To analyze the critical behavior, it is convenient to separate the molar Gibbs free
energy g.T; H/ in a regular part gr .T; H/ and a singular part gs .T; H/ that contains
the singularities associated to the critical point as well as the discontinuities along
the line of coexistence. Thus, we write

g.T; H/ D gr .T; H/ C gs .T; H/: (13.92)

Any quantity obtained from g will also have a regular and a singular part. The
singular part of the nondiverging quantities is chosen so that it vanishes at the critical
point, what always can be done by the addition of appropriate terms in gr .
The regular part of the molar Gibbs free energy has an expansion around the
critical point whose first terms are
cc
gr D gc  sc .T  Tc /  .T  Tc /2 ; (13.93)
2Tc

where gc , sc and cc are constant. The last term is absent if the molar heat capacity is
divergent.
It is convenient to use variables related to the critical point which occurs at H D 0
and T D Tc . We define thus " D T  Tc and write

gs .T; H/ D G ."; H/: (13.94)

Notice that gs .T; H/ is an even function of H and the same happens to G ."; H/, that
is,

G ."; H/ D G ."; H/: (13.95)

The magnetization m and the molar entropy s are given by

@g
mD D G2 ."; H/ (13.96)
@H
and

@g cc
sD D sc C "  G1 ."; H/; (13.97)
@T Tc

where G1 D @G =@" and G2 D @G =@H.

Deriving m and s with respect to H and to T, respectively, we get
 
@m
D D G22 ."; H/ (13.98)
@H T
266 13 Magnetic Systems

and
 
1 @s cc
cH D D  G11 ."; H/; (13.99)
T @T H Tc

where G11 D @2 G =@"2 and G22 D @2 G =@H 2 .

Along the coexistence line, H D 0 and " < 0, the spontaneous magnetization m
is obtained by taking the limit H ! 0 in (13.96),

m D G2 ."; 0C /: (13.100)

Along " D 0

m D G2 .0; H/: (13.101)

Along H D 0, the susceptibility  is given by

 D G22 ."; 0/ (13.102)

and the molar heat capacity at constant field is given by

cH D cc  Tc G11 ."; 0/: (13.103)

Scaling Theory

According to the Widom scaling theory, the singular part of the Gibbs free energy
around the critical point is a generalized homogeneous function, that is,

G ."; H/ D
a G .
";
b H/; (13.104)

valid for any positive values of

. The exponent a and b are related to the critical
exponents introduced above, as we will see below.
The homogeneity of G implies the homogeneity of its derivatives. Indeed,
deriving (13.104) with respect to " and to H, we get

G1 ."; H/ D
aC1 G1 .
";
b H/ (13.105)

and

G2 ."; H/ D
aCb G2 .
";
b H/: (13.106)
13.5 Criticality 267

Deriving these expressions with respect to " and to H, respectively, we obtain in

addition

G11 ."; H/ D
aC2 G11 .
";
b H/; (13.107)

and

G22 ."; H/ D
aC2b G22 .
";
b H/: (13.108)

Next we determine the behavior of these quantities along H D 0. For " < 0, we
choose
D j"j1 to obtain

G1 ."; 0/ D j"ja1 G1 .1; 0/; (13.109)

G2 ."; 0C / D j"jab G2 .1; 0C /; (13.110)
G11 ."; 0/ D j"j a2
G11 .1; 0/; (13.111)
G22 ."; 0/ D j"ja2b G22 .1; 0/: (13.112)

For " > 0 and H D 0, we choose

D "1 to obtain

G11 ."; 0/ D "a2 G11 .1; 0/; (13.113)

G22 ."; 0/ D "a2b G22 .1; 0/: (13.114)

Replacing these results into the expressions (13.100), (13.102) and (13.103), we
get

m  j"jab ; (13.115)
  j"ja2b ; (13.116)
cH  j"ja2 ; (13.117)

or

cc  cH  j"ja2 ; (13.118)

if cH is finite. We see therefore that the Widom scaling theory predicts power laws
for the critical behavior of these quantities.
Comparing these expressions with (13.83), (13.84), (13.86) and (13.87) we find
the following relations between the critical exponents and the indices a and b:

˛ D a C 2; (13.119)
ˇ D a  b; (13.120)
 D a C 2b: (13.121)
268 13 Magnetic Systems

For " D 0 and H > 0, we choose

D H 1=b to obtain

G2 .0; H/ D H .ab/=b G2 .0; 1/; (13.122)

which, replaced in (13.101), gives

m  H .ab/=b : (13.123)

Comparing with expression (13.85), we get the following relation between the
exponent ı and the indices a and b:

1 ab
D : (13.124)
ı b

The four exponents ˛, ˇ,  and ı are not independent. As they are related to the
indices a and b, then only two of them can be chosen as independent. Eliminating
a and b, we find several relations that include the Rushbrooke relation (13.89), the
Griffiths relation (13.90) and the Widom relation (13.91).

Universal Function

The homogeneity of the singular part of the free energy and its derivatives implies,
as we have seen, certain relations between the critical exponents. It also give rise
to certain geometric relations between the thermodynamic quantities, which we
examine now. Let us consider initially the magnetization m D G2 ."; H/. If, in
the regime " > 0, we choose
D "1 and substitute it in (13.106) we get the result

G2 ."; H/ D "ˇ G2 .1; H" /: (13.125)

where  D ˇ C  . That is

m D "ˇ .H" /; (13.126)

where .x/ D G .C1; x/ is a function of one variable. Similarly we can obtain the
relation valid in the regime " < 0.
The result above tell us that the isotherms m versus H must coincide if we
perform a change of scale such that the axis-m is multiplied by j"jˇ and the axis-H
by j"j . The experimental data of magnetization m as a function of the field H
and of the temperature T can then be analyzed in the following way. In a diagram
x-y we place y D mj"jˇ versus x D Hj"j . The data will collapse in a curve
y D .x/ which for that reason is called universal function. Figure 13.9 shows such
collapse obtained from the experimental data presented in Fig. 13.2 for the isotherms
of nickel.
13.5 Criticality 269

Fig. 13.9 Graph of mj"jˇ 10

versus Hj"j for nickel
obtained from the data of 8
Fig. 13.2. The upper and T < Tc
lower branches correspond,
6

−β
respectively, to the points

m|ε|
below and above the critical
temperature 4
T > Tc
2

0
0 0.2 0.4 0.6 0.8 1.0
−Δ
H|ε|

Similar results for the susceptibility  and for the molar heat capacity cH can be
obtained from (13.107) and (13.108). For " > 0 we find

 D " .H" / (13.127)

and

cH D "˛ f .H" /; (13.128)

or

cc  cH D "˛ f .H" /; (13.129)

in case cH is finite, where .x/ D G22 .1; x/ and f .x/ D G11 .1; x/ are universal
functions. Analogous results can be derived for " < 0.

Problems

13.1 Obtain the Maxwell relations for a magnetic system:

       
@T @H @T @M
D ; D ;
@M S @S M @H S @S H
       
@S @H @S @M
D ; D :
@M T @T M @H T @T H

13.2 Magnetocaloric effect. Show that .@T=@H/S D TA=CH , where A is the

magnetocaloric effect, defined by A D .@M=@T/H . Since CH  0, materials
270 13 Magnetic Systems

whose magnetization decreases with temperature at constant field (A < 0) has the
temperature decreased with the decrease in the field in a adiabatic process.
13.3 Show that

TA2
CH D CM C :
XT

Verify this result for a paramagnetic system.

13.4 Shown that
 2
@H
CH D CM C T XT :
@T M

13.5 Equation (13.40), for the magnetization of an ideal paramagnetic, can be

written as
H
m D B. /;
RT

where B.x/ is the Brillouin function, defined by

2J C 1 2J C 1 1 x
B.x/ D coth. x/  coth. /:
2J 2J 2J 2J

Show that (a) it is a monotonic increasing function and such that B.x/ ! 1 when
x ! 1; (b) for small values of the argument, it behaves according to B.x/ D
.J C 1/x=3J; (c) for J D 1=2 it reduces to tanh.x/; and (d) for J ! 1 it reduces to
the Langevin equation L.x/, defined by

1
L.x/ D coth x  :
x

13.6 Show that the critical temperature of a ferromagnetic system described by the
equation of state

m D B. .H C
m//
RT
is given by

.J C 1/
2
Tc D :
3JR
Determine the behavior of the spontaneous magnetization, susceptibility and heat
capacity around the critical point.
13.5 Criticality 271

13.7 Near the critical point, the molar Helmholtz free energy of a ferromagnetic
system is described by

1 1
f D f0 C a.T  Tc /m2 C b m4 ;
2 4
where a and b are constant and f0 depends only on temperature. Determine
the spontaneous magnetization and susceptibility at zero field as functions of
temperature. Determine the critical exponents ˇ,  and ı.
13.8 Show that the spontaneous magnetization, given by (13.58), has the following
expansion
 2    
m T 12 T 2
D3 1  1 :
Tc 5 Tc

Use this expansion to determine the behavior of the molar heat capacity just
below Tc .
13.9 Show that

m D H 1=ı ˚."H 1= /:

13.10 Show that (13.68) yields the following form for the universal function

1
y D .x/ D ˙x C x3 :
3
where the signs C and  are valid for temperatures above and below the critical
temperature, respectively. The variables x and y are defined by x D j"j1=2 m= and
y D j"j3=2 =.
/ where " D .T  Tc /=Tc .
Chapter 14
Magnetic Ordering

14.1 Antiferromagnetism

Antiferromagnetic Materials

The antiferromagnetic crystals undergo a transition to the paramagnetic state at a

temperature called the Néel temperature (Table 14.1). Above it, in the paramagnetic
state, the dipoles point in random directions due to the predominance of thermal
agitation, producing a zero total magnetization. Below the Néel temperature in
antiferromagnetic state, the total magnetization is also zero, but for another reason.
In the antiferromagnetic state, the lattice formed by magnetic atoms is divided into
two or more equivalent sublattices, each presenting a spontaneous magnetization in
a certain direction. The interaction between the dipoles causes a magnetic ordering
such that the vector sum of the spontaneous magnetization of the sublattice vanishes,
which defines the antiferromagnetic state.
The simplest antiferromagnetic structure consists of a magnetic lattice that
is divided into only two equivalent and interpenetrating sublattices. In this case
the interaction between neighboring magnetic dipoles, which belong to different
sublattices, favors an antiparallel alignment of them, resulting in the ordering of
the magnetizations of the sublattices in opposite directions. The simple antiferro-
magnetism is represented by the fluorides FeF2 and MnF2 , whose magnetic ions
form a body-centered tetragonal lattice, and the compounds KMnF3 , KNiF3 and
RbMnF3 , which have the structure of perovskite and whose magnetic ions form
a simple cubic lattice. Other compounds have more complex antiferromagnetic
structure, consisting of various magnetic sublattices as is the case of the compounds
MnO, FeO, CoO, NiO, whose magnetic lattices are face-centered cubic. In these
compounds, the magnetization of the sublattices are all collinear.
A magnetic structure in which the magnetization of sublattices are not collinear
is represented by the compound NiF2 . The magnetic ions of this compound form
a body-centered tetragonal lattice that is divided into two equivalent sublattices.

© Springer-Verlag Berlin Heidelberg 2017 273

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_14
274 14 Magnetic Ordering

Table 14.1 Néel temperature Material TN Type

TN , in K, and type of ordering
of substances that become FeF2 90 Simple
antiferromagnetic just below MnF2 75 Simple
the the Néel temperature. KMnF3 95 Simple
Some have weak KNiF3 275 Simple
ferromagnetism (wf). Source: RbMnF3 82 Simple
AIP
MnO 120 Complex
FeO 198 Complex
CoO 291 Complex
NiO 530 Complex
NiF2 83 Simple, wf
Cr2 O3 318 Alternating
˛-Fe2 O3 948 Alternating, wf
MnCO3 32 Alternating, wf
CoCO3 20 Alternating, wf
CrSb 723 Metamagnet
MnTe 323 Metamagnet
CrCl2 20 Metamagnet
FeCl2 24 Metamagnet
CoCl2 25 Metamagnet
MnO2 84 Helical
FeCl3 15 Helical
MnAu2 363 Helical
Cr 312
Mn 95

The magnetization of the sublattices are not exactly opposite, but form an angle
giving rise to a nonzero total magnetization. As the deviation from the ideal
antiferromagnet state is small this compound exhibits a weak ferromagnetism.
The oxide Cr2 O3 , the hematite ˛-Fe2 O3 and the compounds MnCO3 and CoCO3
have a structure in which the magnetic dipoles alternate along a particular crystal
direction. The last three are also examples of antiferromagnetic substances that have
weak ferromagnetism. The metamagnets CrSb, MnTe, CrCl2 , FeCl2 and CoCl2 are
constituted by a succession of ferromagnetic layers of alternating signs.
More complex magnetic structures include the periodic in which the magnetiza-
tion varies, in magnitude or direction, along a crystal axis. The compounds MnO2 ,
FeCl3 and MnAu are examples of substances that exhibit antiferromagnetic periodic
helical structures.
Among the metals, chromium (Cr) is found in the antiferromagnetic state at
room temperature, becoming a paramagnetic 39 ı C. Manganese (Mn), which is
paramagnetic at room temperature, becomes antiferromagnetic below 95 K.
14.1 Antiferromagnetism 275

Order Parameter and Free Energy

For a suitable description of the antiferromagnetic state, as well as other types of

ordering, it is necessary to use an expanded thermodynamic space relatively to
that of a ferromagnetic. The most natural way is to use a space composed by the
magnetization of each magnetic sublattice and the temperature. To this end, let us
examine the simplest case of a system that is ordered according to two equivalent
and interpenetrating magnetic sublattices. Suppose further that the magnetizations
of the sublattices occur along a certain crystal direction. Denoting by m1 and m2
the magnetizations corresponding to the sublattices 1 and 2, respectively, then the
molar Helmholtz free energy f .T; m1 ; m2 / is a function of m1 and m2 in addition to
temperature. Denoting by H1 and H2 the fields conjugated to m1 and m2 , then

df D sdT C H1 dm1 C H2 dm2 : (14.1)

The antiferromagnetic state means that the magnetization of the sublattices are
different. They need not necessarily have the same magnitude. This happens only in
restricted antiferromagnetic state, which occurs in the absence of external field, for
which m1 D m2 and the total magnetization

1
mD .m1 C m2 / (14.2)
2
vanishes. In the presence of an external field it is possible the occurrence of a generic
antiferromagnetic state for which m ¤ 0.
Suppose that in an antiferromagnetic phase the values of the magnetization of
the sublattices are m1 D a and m2 D b with a ¤ b. Another antiferromagnetic
phase equally possible is that for which m1 D b and m2 D a, as the sublattices
are equivalent. Therefore, an antiferromagnetic system displays two different
thermodynamic states that should be understood as two thermodynamic phases in
coexistence. We note that these two thermodynamic phases have the same total
magnetization m D .a C b/=2 and therefore can not be distinguished by m. They
are distinguished by the order parameter  defined by

1
D .m1  m2 /; (14.3)
2
that takes a positive value in one phase and a negative value in the other. The order
parameter  makes also the distinction of the antiferromagnetic phase, ordered,
from the paramagnetic phase, disordered. In the paramagnetic phase the sublattices
cannot have different magnetizations, so that m1 D m2 , that is,  D 0.
An alternative and more interesting description is the one in which the thermo-
dynamic space is formed by the total magnetization m and the order parameter  in
276 14 Magnetic Ordering

addition to the temperature. From (14.2) and (14.3) we get

m1 D m C  and m2 D m  ; (14.4)

which replaced in (14.1) gives

df D sdT C Hdm C H d; (14.5)

where

H D H1 C H2 and H D H1  H2 : (14.6)

the quantity H is the variable conjugated to m and thus is identified with the external
field. The quantity H is the field conjugated to , which we call staggered field.
Experimentally, this field is identically zero.
In certain applications, it is convenient to use the molar Gibbs free energy
g.T; H; /, obtained from f .T; m; / by the Legendre transformation

g.T; H; / D minf f .T; m; /  Hmg (14.7)

m

for which

dg D sdT  mdH C H d: (14.8)

In the study of phase coexistence is particularly useful the potential .T; H; H /

obtained by the Legendre transformation

.T; H; H / D minf f .T; m; /  Hm  H g: (14.9)

m;

As  is a function only of thermodynamic fields then, at the coexistence of phases

it must have the same value in each phase.

Néel Theory

The Néel theory corresponds to the application of molecular field theory of Weiss to
antiferromagnetic systems. According to the Néel theory the magnetization of each
of the sublattices of a simple antiferromagnetic system is given by

m1 D tanh. .H 
m2 // (14.10)
RT
14.1 Antiferromagnetism 277

and

m2 D tanh. .H 
m1 //; (14.11)
RT

where 
m2 and 
m1 are the molecular fields of Weiss acting on the dipole
moments corresponding to the sublattices 1 and 2, respectively, because the
magnetic dipoles in a sublattice interact only with those belonging to the other
sublattice. That is, the molecular field acting on the dipoles of the sublattice 1 is due
to the dipoles of sublattice 2 and vice versa. The negative sign induces an antiparallel
dipole alignment between different sublattices.
Equations (14.10) and (14.11) introduced by Néel can be derived from the
following molar Helmholtz free energy
 

RT C m1 C m1  m1  m1
f D m1 m2 C ln C ln C
2 2 2 2 2 2
 
RT C m2 C m2  m2  m2
C ln C ln : (14.12)
2 2 2 2 2

Indeed, from H1 D @f =@m1 and H2 D @f =@m2 , we obtain

1 RT C m1
H1 D .
m2 C ln / (14.13)
2 2  m1

and
1 RT C m2
H2 D .
m1 C ln /: (14.14)
2 2  m2

Taking into account that H1 D H2 D H=2, because H D 0, the results (14.10)

and (14.11) follows directly from these equations.
To determine the magnetizations of the sublattices, we should solve (14.10)
and (14.11) for m1 and m2 . When H D 0, it is easy to see that the solution is such
that m1 D m2 and that the order parameter  D m1 D m2 obeys the equation

 D tanh. /; (14.15)

RT
which is equivalent to the equation

RT C 

 D ln ; (14.16)
2  

obtained from (14.13) and (14.14).

278 14 Magnetic Ordering

The solution of these equations is such that for temperatures above the critical
temperature TN , given by

2
TN D ; (14.17)
R

we have  D 0 and m1 D m2 D 0. Below this temperature,  ¤ 0 and therefore

m1 D m2 ¤ 0, that define the antiferromagnetic state.

Susceptibility

To determine the susceptibility  D @m=@H it is convenient to calculate first 1 D

@m1 =@H and 2 D @m2 =@H and then determine  D .1 C 2 /=2. Deriving both
members of (14.10) and (14.11) with respect to H, we obtain two equations which
can be solved for 1 and 2 . From the expressions for 1 and 2 we reach the
following result for  at zero field

1 RT
D
C 2 ; (14.18)
  2

valid both in the paramagnetic and antiferromagnetic phases.

Above the Néel temperature,  D 0, so that

1 RT
D
C 2: (14.19)


Using (14.17), we get the result

TN
D ; (14.20)

.T C TN /

which is the Curie-Weiss law for antiferromagnetic systems.

Below the Néel temperature, we should replace the order parameter , nonzero
solution of (14.15), into (14.18). The result is shown in Fig. 14.1. Just below the
Néel temperature,
s
TN  T
D 3 ; (14.21)
TN

yielding the result

T
D : (14.22)
2
TN
14.1 Antiferromagnetism 279

Fig. 14.1 Susceptibility of 0.8

an antiferromagnetic system
according to the Néel theory.
The lower and upper curves 0.6
are respectively the parallel
and perpendicular

λχ
susceptibilities. The middle 0.4
curve corresponds to the
weighted average (14.54)
0.2

0
0 1 2 3
T/TN

valid for temperatures near TN . According to these results, the susceptibility is finite
at the critical temperature and has the value  D 1=2
, as seen in Fig. 14.1.

Criticality

Near the Néel temperature, the order parameter behaves according to

  .TN  T/ˇ : (14.23)

As we see from (14.21), the value predicted by the Néel theory for the exponent ˇ
is 1/2. Experimentally, however, the observed values are different and are shown in
Table 14.2 for various antiferromagnetic compounds.
Other exponents can be defined for the phase transition between paramagnetic
and antiferromagnetic phase. The exponent ˛ is related to the molar heat capacity
cH , which behaves according to

cH  jTN  Tj˛ ; (14.24)

with ˛ > 0 for the case where cH diverges. If cH is finite at the critical temperature,
with the value c0 , then the behavior is given by

c0  cH  jTN  Tj˛ ; (14.25)

where, in this expression, ˛ < 0. Table 14.2 shows the value of this exponent for
various antiferromagnetic compounds. Note that for some compounds ˛ is positive
and for others ˛ is negative.
The critical exponent  , also shown in Table 14.2, governs the critical behavior
of the susceptibility  associated to the order parameter  and defined by
280 14 Magnetic Ordering

Table 14.2 Critical Compound ˛ ˇ  TN

exponents and Néel
temperature, in K, of RbMnF3 0:14 0:32 1:37 83:1
antiferromagnetic MnF2 0:34 1:27 67:3
compounds, obtained FeF2 C0:14 0:33 1:25 78:3
experimentally DyPO4 0:31 3:39
DyAlO3 0:31 3:52

0.3 0.28

χ/C (K )
χ/C (K )

0.2 0.24

-1
-1

0.1 b 0.20
p
c

0 0.16
0 1 2 3 4 1.2 1.4 1.6 1.8 2.0 2.2 2.4

T (K) T (K)

Fig. 14.2 Susceptibility of the antiferromagnetic compound MnCl2 4H2 O obtained experimen-
tally by Lasher et al. [153]. The curves c and b correspond to the susceptibility of a single crystal
determined along the directions parallel and perpendicular to the preferential axis, respectively.
The curve p represents the susceptibility of the powder. According to Fisher the critical temperature
occurs at the inflection point immediately below the maximum of the susceptibility, as shown in
detail for the parallel susceptibility. C is the Curie constant

 D @=@H . It is possible to show from the Néel theory that the exponent  is
equal to 1. The experimental values are distinct from those predicted by the Néel
theory, as seen in Table 14.2.
The susceptibility  of antiferromagnetic systems must be distinguished from
the susceptibility  D @m=@H, related to the magnetization m. The susceptibility
 is finite, as predicted by the Néel theory. However, the Néel temperature does
not correspond to a maximum of . According to Fisher, the Néel temperature
corresponds to the inflection point in the susceptibility , which occurs just below
the maximum of , as seen in Fig. 14.2. Yet according to Fisher, near the critical
point,  is related to the molar heat capacity as follows

@
.T/  cH ; (14.26)
@T
from which we conclude that the singular part of  is governed by the exponent
1  ˛.
14.2 Metamagnetic Systems 281

14.2 Metamagnetic Systems

Free Energy

The behavior of antiferromagnetic systems in the presence of an external field is

very different from that observed in ferromagnetic systems. In the latter, an external
field as small as it is destroys the phase transition. In antiferromagnetic systems,
by contrast, the transition is still present, unless the field is high enough. The
different behaviors are due to the role of H with respect to the order parameter.
In ferromagnetic systems, H is the field conjugate to the order parameter. In
antiferromagnetic systems that role is reserved to the staggered field H and not
the external field H.
An antiferromagnetic system subject to an external field can display a phase
diagram containing various types of magnetic ordering. In order to determine the
phase diagram in the space H versus T we start from the following molar Helmholtz
free energy

12
11 2
f D m1 m2  .m1 C m22 /C
2 4
 
RT C m1 C m1  m1  m1
C ln C ln C
2 2 2 2 2
 
RT C m2 C m2  m2  m2
C ln C ln : (14.27)
2 2 2 2 2

where
12 > 0 and
11  0 are parameters and m1 and m2 are the magnetizations of
the sublattices. The first term is the interaction between magnetic dipoles located in
different sublattices and favors the antiparallel alignment of them. The second is the
interaction between magnetic dipoles located on the same sublattice and favors the
parallel alignment of them. Considering that the sublattices are equivalent, the free
energy is invariant by permutation of m1 and m2 .
Several systems are described by the antiferromagnetic free energy (14.27).
When
11 D 0, we recover the expression (14.12), which describes a simple
antiferromagnetic system, consisting of two interpenetrating sublattices. When

11 > 0, the free energy (14.27) describes metamagnetic systems that are anti-
ferromagnetic systems consisting of magnetic layers. The odd layers constitute one
sublattice and the even layers the other sublattice. The interaction between layers
is antiferromagnetic. Within the layer the interaction is ferromagnetic. Examples of
metamagnetic systems are shown in Table 14.1.
It is convenient here to use the variables m and  which are, respectively, the total
magnetization of the sample and the order parameter, related to m1 and m2 by (14.2)
282 14 Magnetic Ordering

and (14.3). Using (14.4) we may write

1 2
0 2
f D m   C
2 2
 
RT CmC CmC m m
C ln C ln C
2 2 2 2 2
 
RT Cm Cm mC mC
C ln C ln : (14.28)
2 2 2 2 2

where
0 D
12 C
11 and
1 D
12 
11 .
To determine the thermodynamic phases we consider the thermodynamic poten-
tial .T; H; H / defined by (14.9). For H D 0,

.T; H; 0/ D minf f .T; m; /  Hmg; (14.29)

m;

that is, the quantities m and  correspond to the minima of the expression between
curls. Deriving this expression with respect to m and  and equating both results to
zero, we get the following conditions for a minimum

RT . C m/2  2
H D
1 m C ln (14.30)
4 .  m/2  2

and

RT . C /2  m2

0  D ln : (14.31)
4 .  /2  m2

These two equations of state implicitly provide m and  as functions of T and H. The
solutions are of two types: (a)  D 0, corresponding to the paramagnetic state, (b)
 ¤ 0, corresponding to the antiferromagnetic state. This type of solution appears in
pairs, since the equations of state are invariant under the change of sign of . Thus
the antiferromagnetic region is a region of coexistence of two antiferromagnetic
phases, defined by  and , nonzero.

Tricritical Point

The two kinds of solution divide the phase diagram into two regions, as shown in
Fig. 14.3a for the case
11 D 2
12 . The line separating the two regions corresponds
to a continuous phase transition or critical line for low field. For high fields, it
corresponds to a discontinuous phase transition in which three phases coexist:
the paramagnetic and two antiferromagnetic. The point separating the continuous
14.2 Metamagnetic Systems 283

(a) 1.2 (b) 1.2

1.0 1.0

0.8 P 0.8
H/H0

m/μ
0.6 0.6

0.4 AF 0.4

0.2 0.2

0 0
0 0.5 1.0 1.5 0 0.5 1.0 1.5
T/TN T/TN

Fig. 14.3 Phase diagram of a metamagnetic system described by the free energy (14.27) for
11 D
2
12 . (a) Diagram field H versus temperature T, where H0 D
12 . The transition line separates
the paramagnetic phase (P) and the two antiferromagnetic phases (AF) in coexistence. The solid
curve represents a critical line and the dotted line the coexistence of three phases: the paramagnetic
and two antiferromagnetic. The full circle represents the tricritical point. (b) Magnetization m
versus temperature T. Along the continuous curves the field is constant. The thick lines represent
the coexistence curve and the critical line

and discontinuous transitions is the tricritical point where the three phases become
identical.
The critical line occurs when the nonzero solution  ! 0. Dividing both sides
of (14.31) by  and taking the limit  ! 0, we get

RT

0 D (14.32)
2  m2

This equation together with the equation

RT C m
H D
1 m C ln (14.33)
2  m

obtained from (14.30) by setting  D 0, give the critical line in the diagram H versus
T in parametric form. In particular, for H D 0 and therefore m D 0, we get the Néel
temperature

RTN

0 D : (14.34)
2

To determine the line of coexistence of three phases we must look to the

thermodynamic potential  given by (14.29), which is a function of thermodynamic
fields only and therefore, at the coexistence of phases, has the same value in each
phase. Denoting by mp and ma the magnetizations corresponding to the solutions
paramagnetic ( D 0) and antiferromagnetic ( ¤ 0) of (14.30) and (14.31), then
284 14 Magnetic Ordering

the condition for coexistence is given by

f .T; ma ; /  Hma D f .T; mp ; 0/  Hmp : (14.35)

Thus, we determine the coexistence line in the diagram H versus T, shown in

Fig. 14.3a. In particular, one can show that at the tricritical point

m2
12
D : (14.36)
2 3
11

Therefore, this condition together with the equations of the critical line (14.32)
and (14.33), determine the tricritical point. Since jmj  , then the condition for
the existence of tricritical point is
12  3
11 . If this condition is not satisfied, the
transition line is reduced to a critical line only.
The magnetization m as a function of temperature for constant H is shown in
Fig. 14.3b. Defining the variable D mp  ma , then it is possible to show that along
the coexistence of three phases, behaves near the tricritical point according to

 jT  Tt j; (14.37)

where Tt the temperature of the tricritical point. The variable gives the length
of the tie lines which, according to (14.37), varies linearly with distance from the
tricritical point, as shown in Fig. 14.3b. Experimental results for a metamagnetic
system are shown in Fig. 14.4.

(a) 0.4 (b) 700

600
0.3
500
(kA/m)

P
H (T)

400
0.2
AF 300

200
0.1
100

0 0
1.0 1.5 2.0 2.5 3.0 1.0 1.5 2.0 2.5 3.0

T (K) T (K)

Fig. 14.4 Phase diagram of the metamagnetic compound Dy3 Al5 O12 obtained experimentally by
Landau et al. [152]. (a) Diagram field H versus temperature T. The transition line separates the
paramagnetic phase (P) and the antiferromagnetic phase (AF). The continuous curve represents a
critical line and a dotted line represents a coexistence line. The full circle represents the tricritical
point. (b) Volumetric magnetization M versus temperature T. Along the continuous curves the
field is constant. The thick lines represent the coexistence curve and the critical line
14.2 Metamagnetic Systems 285

Landau Theory

Landau theory allows the determination of the topology of the phase diagrams.
An important ingredient of the Landau theory is related to the symmetries of the
system which are reflected in the symmetries of the free energy. In antiferromagnetic
systems, the equivalence of sublattices implies that the free energy is invariant by
permutation of the magnetization of sublattices or, equivalently, invariant by the
change of sign of the order parameter .
According to Landau theory, we postulate for the molar free energy g.T; H; /,
defined by (14.7), the following form

g D a0 C a2 2 C a4 4 C a6 6 ; (14.38)

which is an even function of  due to the symmetry g.T; H; / D g.T; H; /. The
parameters ai are linear functions of T and H and a6 is strictly positive. As the
staggered field H D @g=@ vanishes then according to (14.9), the order parameter
 corresponds to an absolute minimum of the polynomial (14.38).
The number of coexisting phases equals the number of absolute minima of g.
To determine the conditions for existence of a certain number of minima, it is
convenient to write the free energy in the form

g D g0 C a6 ./; (14.39)

where g0 is the value of the absolute minimum and ./ is a polynomial of sixth
degree in . Thus, the minima of g become the double roots of ./.
Since is a pair polynomial of sixth-degree, then we can have at most three
phases in coexistence. Initially, we determine the conditions for the coexistence of
three phases. For this to occur it is necessary that

D 2 .  0 /2 . C 0 /2 D 2 .2  20 /2 ; (14.40)

from which we conclude that

a4 a2
D 220 and D 40 ; (14.41)
a6 a6
or

a24
a2 D and a4 < 0: (14.42)
4a6

The three phases in coexistence are:  D 0, the paramagnetic phase, and  D ˙0 ,
the two antiferromagnetic phases.
286 14 Magnetic Ordering

A critical line (two identical phases) is described by a polynomial having a

quadruple root, that is, D 4 .2 C a2 /, from which we conclude that

a2 D 0 and a4 > 0: (14.43)

Finally, the tricritical point (three identical phases) corresponds to a sixfold root so
that D 6 , from which follows a4 D a2 D 0. The phase diagram in space a2
versus a4 is shown in Fig. 11.11a.
To determine the magnetization m D @g=@H, we assume that the coefficients
ai depend linearly on T and H. Thus up to quadratic order in  we obtain m D
c0  c2 2 , where c0 and c2 are constant. On the three-phase coexistence line, the
magnetization of the paramagnetic phase mp D c0 and that of the antiferromagnetic
phase ma D c0  c2 20 . Therefore, the variable D mp  ma is related with 0 by
D c2 20 . But according to (14.41), 0 behaves as

0  ja4 j1=2 ; (14.44)

which yields the result

 ja4 j; (14.45)

that is, approaches linearly the tricritical point, as shown in Fig. 11.11b, which
justify result (14.37).

14.3 Anisotropy

Perpendicular Susceptibility

The calculations developed in the previous sections are appropriate for a system
with uniaxial anisotropy, that is, such that the magnetizations of the sublattices
are parallel to the preferred direction determined by the crystalline anisotropy and
applied field along the preferred direction. For a more general treatment that takes
into account both applied fields in other directions and the possible emergence of
magnetization in other directions, we start with the following molar Helmholtz free
energy

f D .m1x m2x C m1y m2y / C m1z m2z C

2 2
 
RT C m1 C m1  m1  m1
C ln C ln C
2 2 2 2 2
 
RT C m2 C m2  m2  m2
C ln C ln : (14.46)
2 2 2 2 2
14.3 Anisotropy 287

where m1 D .m21x C m21y C m21z /1=2 and m2 D .m22x C m22y C m22z /1=2 . The components
of the total magnetization are given by mx D .m1x C m2x /=2, my D .m1y C m2y /=2
and mz D .m1z C m2z /=2. The parameter  measures the anisotropy with respect to
the preferred axis z. The value  D 1 corresponds to the isotropic case and  D 0
corresponds to the maximum anisotropy. We consider here the anisotropic case,
0   < 1, so that the spontaneous magnetization of the sublattices will emerge
along the preferential axis z.
If the external field is applied along the z direction, then m1x D m1y D m2x D
m2y D 0 and the expression (14.46) is reduced to (14.12). The results obtained
previously for the susceptibility correspond in the present case to the parallel
susceptibility k D @mz =@Hz . To determine the perpendicular susceptibility ? D
@mx =@Hx we consider a field applied in the x direction. In this case m1x D m2x D mx
and, in addition, m1y D m2y D 0 and m21z D m22z D 2 . The free energy (14.46)
becomes then

 2
2
f D m  C
2 x 2
 
RT C m2 C m2  m2  m2
C ln C ln : (14.47)
2 2 2 2 2

where m D .2 C m2x /1=2 , so that the field Hx D @f =@mx is given by

mx RT C m
Hx D
 mx C ln : (14.48)
m 2  m

To determine ? at zero field, it suffices to use the above expression for small
values of Hx .The magnetization mx will also be small and will vanish when Hx ! 0.
For small values of hx , we get

mx RT C 
Hx D
 mx C ln : (14.49)
 2  

Above the Néel temperature,  D 0 and we obtain

RT T
Hx D
 mx C mx D
 mx C
mx ; (14.50)
2 TN

where we used the result (14.17). Below the Néel temperature, we use the
result (14.16) to obtain

Hx D
 mx C
mx ; (14.51)
288 14 Magnetic Ordering

From these two equation, we reach the results

TN
? D ; (14.52)

.T C  TN /

valid above the Néel temperature and

1
? D ; (14.53)

.1 C  /

valid below the Néel temperature.

The results for k and ? at zero field are shown in Fig. 14.1 for  D 1=2. Below
the Néel temperature, ? is constant whereas k increases monotonically from
the zero value at T D 0. Above TN both susceptibilities decrease monotonically.
For samples of antiferromagnetic substances in the form of powder the measured
susceptibility  corresponds to a weighted mean of the two susceptibilities. For
uniaxial anisotropies with a preferred axis the following relation is valid

1 2
D k C ? ; (14.54)
3 3
which is shown in Fig. 14.1. In Fig. 14.2 we shown experimental data of the
susceptibility of the compound MnCl2 4H2 O both in the single crystal and powder
forms

Phase Diagram

Let us determine next the possible thermodynamic phases predicted by the free
energy (14.46) for the field H applied along the preferential axis. In addition to the
paramagnetic and antiferromagnetic phases, a system described by (14.46) exhibits
a new magnetic phase, as seen in the phase diagram of Fig. 14.5. This phase, called
spin flop, is such that the magnetization of the sublattices point to directions distinct
that of the preferential axis although the applied field is parallel to the preferential
direction. The phase transitions from the antiferromagnetic state to the paramagnetic
state and from the spin flop state to the paramagnetic state are continuous. However,
the transition from the antiferromagnetic state to the spin flop state is discontinuous.
The phase diagram has two critical lines and one coexistence line, as seen in
Fig. 14.5 and Fig. 14.6.
To determined the phase diagram, it is more convenient to consider the thermo-
dynamic potential , which is function only of the thermodynamic fields, given by

 D minf f  Hmz g; (14.55)

14.3 Anisotropy 289

2.0

1.5

SF P

H/H0
1.0

0.5
AF

0
0 0.2 0.4 0.6 0.8 1.0 1.2

T/TN

Fig. 14.5 Phase diagram of a magnetic system described by the free energy (14.46) for  D 0:75.
The system displays three phases: paramagnetic (P), antiferromagnetic (AF) and spin flop (SF).
The thick line between the phases AF and P and between SF and P are critical lines. The dashed
line between the phases AF and SF is a coexistence line. The three lines meet at the bicritical point

15
4.50

10 P SF
SF
H (T)
H (T)

4.45 P

5
4.40

AF AF
0 4.35
0 1 2 3 4 5 6 3.87 3.92 3.97 4.02

T (K) T (K)

Fig. 14.6 Phase diagram of the antiferromagnetic salt NiCl2 6H2 O obtained experimentally by
Oliveira et al. [155]. The system presents three phases: paramagnetic (P), antiferromagnetic (AF)
and spin flop (SF). The lines between the phases AF and P and between SF and P are critical lines.
The line between the phases AF and SF is a coexistence line. The three lines meet at the bicritical
point

where the minimization is done on the Cartesian components of the sublattice

magnetizations. With the help of the potential  we can calculate the coexistence
line since  must have the same value on the two phases in coexistence.
In the antiferromagnetic and paramagnetic phases, the magnetizations are par-
allel to the preferential direction, so that mix D miy D 0 and m1z D mz C  and
290 14 Magnetic Ordering

m2z D mz  . Therefore, for these phases,

2
f D .m  2 /C
2 z
 
RT C mz C  C mz C   mz    mz  
C ln C ln C
2 2 2 2 2
 
RT C mz   C mz    mz C   mz C 
C ln C ln : (14.56)
2 2 2 2 2

The minimization leads to the equations

RT C mz C  RT C mz  
H D
mz C ln C ln (14.57)
4  mz   4  mz C 

and
RT C mz C  RT C mz  

 D ln  ln : (14.58)
4  mz   4  mz C 

The critical line that separates the antiferromagnetic and paramagnetic is

obtained by taking the limit  ! 0. The first equation gives

RT C mz
H D
mz C : (14.59)
2  mz

Dividing both sides of the second equation by  and taking the limit  ! 0, we get

RT

D (14.60)
2  m2z

These two equations give parametrically the critical line shown in Fig. 14.5.
The spin flop phase is such that the magnetizations have the same magnitude
but have distinct direction from each other and from the preferential direction. The
components are given parametrically by m1z D m2z D m cos , m1x D m2x D
m sin  and m1y D m2y D 0. Using this parametrization we get

2
f D m .cos2    sin2 /C
2
 
Cm Cm m m
C RT ln C ln (14.61)
2 2 2 2

and

 D minff  Hm cos g: (14.62)

m;
14.3 Anisotropy 291

The minimization in m and  lead us to the equations

RT C m

m D ln (14.63)
2  m

and

H D
m.1 C  / cos : (14.64)

The critical line of transition from the spin flop and the paramagnetic phases is
obtained when  D 0 which gives

H D
m.1 C  /: (14.65)

This equation together with (14.63) gives parametrically the critical line shown in
Fig. 14.5.
To obtain the coexistence line of the antiferromagnetic and the spin flop we
compare the potential  obtained on the two phases. The result is shown in Fig. 14.5
for  D 0:75. The meeting point of the three lines determine the bicritical point.

Bicritical Point

The bicritical point is located at the end of the coexistence line of the antiferro-
magnetic and spin flop phases. The antiferromagnetic phase is described by the
order parameter  D .m1z  m2z /=2, which vanishes both in the paramagnetic
and the spin flop phases. The spin flop phase can also be defined by an order
parameter !, given by ! D .m1x  m2x /=2, which vanishes both in the paramagnetic
and antiferromagnetic phases. Thus the adequate description of the bicritical point
requires the concurrence of two order parameters.
According to the Landau theory, we postulate for the molar Gibbs free energy
g.T; H; ; !/ the following form

g D a00 C a20 2 C a02 ! 2 C a40 4 C a22 2 ! 2 C a04 ! 4 ; (14.66)

which is an even function both in  and !. The parameters aij are linear functions of
T and H and a40 and a04 are strictly positive. The conjugate fields to  and ! must
vanish so that, according to (14.9), the order parameters  and ! correspond to the
minima of (14.66).
The possible phase diagrams are shown in Fig. 14.7. For a222 < 4a40 a04 the
diagram presents three phases: paramagnetic and other two which we identify as
the antiferromagnetic and the spin flop. The transition between the last two are
discontinuous. The other two lines are continuous transition and they meet at the
bicritical point. For a222 > 4a40 a04 , on the other hand, there is a fourth phase,
292 14 Magnetic Ordering

(a) (b)
a02 a02
P P
A A

a20 a20

B C B

Fig. 14.7 Phase diagram around the bicritical and tetracritical points in the space a20 versus a02 .
The dashed line is a coexistence line. The thick lines are critical lines. The phases are such that
 D ! D 0 in P phase,  ¤ 0 and ! D 0 in A phase,  D 0 and ! ¤ 0 in B phase,  ¤ 0 and
! ¤ 0 in C phase

characterized by  ¤ 0 and ! ¤ 0. In this case the transitions are continuous.

The four lines meet at a point called tetracritical point.

14.4 Ferrimagnetism

Ferrimagnetic Materials

The ferrimagnetic materials also order in sublattices that present each of them spon-
taneous magnetization in distinct directions. However, unlike the antiferromagnetic
materials, the sublattices are not equivalent resulting in a non vanishing sum of
the magnetizations. The ferrimagnetic materials present therefore, a spontaneous
magnetization. This is the case of the magnetite, Fe3 O4 , mineral known since
ancient times by its natural magnetism, which disappears at the temperature of
585 ı C.
Magnetite is the most notorious example of a class of ferrimagnetic materials
known as ferrites which have chemical formula MFe2 O4 where M is a divalent metal
and has the same crystalline structure of spinel. The properties of some ferrites are
presented in Table 14.3. In these compounds, the metallic ions, are located in two
types of sites, denoted by A and B, that constitute the two nonequivalent sublattices.
The site of type A is surrounded by four oxygen atoms located at the vertices of a
tetrahedron whereas site B is surrounded by six oxygen atoms located at the vertices
of an octahedron. The number of atoms of type B is equal to twice the number of
atoms of type A.
Another class of ferrimagnetic materials is that constituted by the compounds
with the same crystalline structure of the garnet described by the chemical formula
M3 Fe5 O12 , where M is a trivalent metal and Fe is the trivalent ferric ion. The
properties of some ferrimagnetic garnets are presented in Table 14.3 and in
14.4 Ferrimagnetism 293

Table 14.3 Critical temperature Tc , of ferrimagnetic materials. The table shows also the magne-
tization M (magnetic dipole moment per unit volume) at 20 ı C and b D = B of ferrites; and
compensating temperature Tcomp of the iron garnet. Source: AIP
Ferrite Tc (T) M (kA/m) b Garnet Tcomp (K) Tc (K)
Fe3 O4 858 477 4:2 Y3 Fe5 O12 560
CoFe2 O4 793 422 3:7 Yb3 Fe5 O12 0 548
NiFe2 O4 858 257 2:4 Er3 Fe5 O12 83 556
CuFe2 O4 728 135 1:3 Ho3 Fe5 O12 137 567
MnFe2 O4 568 390 4:5 Dy3 Fe5 O12 226 563
MgFe2 O4 713 115 1:8 Tb3 Fe5 O12 246 568
Li1=2 Fe5=2 O4 943 310 2:6 Gd3 Fe5 O12 286 564

(a) (b) 0.6

16
Gd 0.5
Tb
Dy
12 Ho 0.4
m/μB

Er
m/μ

Tm 0.3
8 Yb
Lu
Y 0.2
4
0.1

0 0
0 200 400 600 0 0.2 0.4 0.6 0.8 1.0 1.2

T (K) T/Tc

Fig. 14.8 (a) Molar spontaneous magnetization m versus temperature of iron garnets M3 Fe5 O12 ,
where M is a rare-earth metal or yttrium, obtained experimentally by Pauthenet [156]. (b) Molar
spontaneous magnetization m versus temperature according to the Néel theory for p D 0:25,
22 D
0 and
11 D 2
12

Fig. 14.8a. The garnet of iron and yttrium, Y3 Fe5 O12 , is particularly interesting
because the only magnetic ions are the ferric ions. The total magnetization is the
result of the sum of opposite magnetizations corresponding to the two sublattices
with distinct number of magnetic atoms. This compound loses the spontaneous
magnetization at the temperature of 287 ı C. At room temperature, the magnetization
is equal to 130 kA/m and the saturation magnetization is equal to 200 kA/m. In
addition b D = B D 5:0.

Néel Theory

Let us consider the simplest case of a ferrimagnetic system in which the magnetic
lattice is divided into two sublattice, in general not equivalent. The total molar
magnetization is the weighted mean of the magnetization of sublattices 1 and 2,

m D pm1 C qm2 ; (14.67)

294 14 Magnetic Ordering

where p and q D 1  p are the mole fractions of the magnetic atoms that are in
sublattices 1 and 2, respectively. We start from the following molar Helmholtz free
energy

11 2
22 2
f D pq
12 m1 m2  p
m1  q mC
2 2 2
 
C m1 C m1  m1  m1
CpRT ln C ln C (14.68)
2 2 2 2
 
C m2 C m2  m2  m2
CqRT ln C ln ; (14.69)
2 2 2 2

where
12 ,
11 and
22 are positive. The corresponding molar Gibbs free energy is
given by

g D minf f  H. pm1 C qm2 /g: (14.70)

The magnetizations m1 and m2 are obtained as the minima of the expression

between curls. The minimization leads us to the following equations

RT C m1
H D q
12 m2 
11 m1 C ln ; (14.71)
2  m1
RT C m2
H D p
12 m1 
22 m2 C ln ; (14.72)
2  m2

which can be written in the form

m1 D tanh. .H  q
12 m2 C
11 m1 // (14.73)
RT
and

m2 D tanh. .H  p
12 m1 C
22 m2 //: (14.74)
RT
At zero field H D 0, in addition to the trivial solution m1 D m2 D 0 these equations
give nontrivial solutions such that m1 ¤ 0 and m2 ¤ 0. Figure 14.8b presents curves
of spontaneous magnetization m versus temperature for the case p D 0:25,
22 D 0
and
11 D 2
12 . It is worth mentioning that, in this case, the total magnetization m
vanishes at a temperature, called compensating temperature Tcomp , smaller than the
critical temperature Tc , as seen in Fig. 14.8b. However, this point does not represent
any phase transition. Experimentally, this point is indeed observed, as shown in
Fig. 14.8a. Table 14.3 shows the compensating temperature as well as the critical
temperature of several ferrimagnetic garnets.
Chapter 15
Dielectrics

15.1 Dielectric Materials

Ordinary Dielectrics

Dielectric materials are insulators characterized by the response to an electric field.

An ordinary dielectric, such as water, placed in the presence of an electric field
acquires an electric polarization that disappears when the field is removed. For small
values of the applied electric field E, the polarization (electric dipole moment per
unit volume) of an ordinary dielectric is proportional to the field, that is,

P D E;
N (15.1)

where N is the constant of proportionality. If an ordinary dielectric is placed between

the plates of a capacitor, the capacitance increases by a factor . This factor, which
is the dielectric constant is related to N by means of

N D 0 .  1/; (15.2)

where 0 is the permittivity of vacuum. The dielectric constant  is a characteristics

of the dielectric substance and in general depends on the temperature. The dielectric
constant of water at room temperature is equal to 80.
The electric polarization of a dielectric substance is due to the molecular
microscopic dipoles of the molecules. Such molecular dipoles can be permanent,
as happens with polar molecules, or can be induced by the applied field. In a
dielectric composed of non-polar molecules, the polarization is only due to the effect
of electrical induction, and therefore should disappear when the field vanish. In a
dielectric composed by molecules with permanent dipole moment, the polarization
also vanishes in the absence of field, but this is due to thermal agitation, which
causes a disordering of the electric dipoles. The dipoles point to arbitrary directions

© Springer-Verlag Berlin Heidelberg 2017 295

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_15
296 15 Dielectrics

in space resulting in a null vector sum of dipole moments. An applied field causes
the partial orientation of the dipoles, giving rise to a polarization.

Ferroelectrics

Ferroelectrics are materials that have a polarization even in the absence of an applied
electric field. This spontaneous polarization generally decreases with increasing
temperature and disappears above a temperature, called transition temperature T0 ,
as shown in Fig. 15.1. Above the transition temperature the material becomes
paraelectric. In this state the polarization behaves according to (15.1), that is, the
material becomes an ordinary dielectric: the polarization arises if we apply an
electric field and disappears if we turn off the field. Rochelle salt or potassium
sodium tartrate tetrahydrate (NaKC4 H4 O6 4H2 O), the substance in which ferroelec-
tricity has been observed for the first time, is ferroelectric at room temperature, but
loses polarization spontaneously if heated above 24 ı C. Above this temperature, it
becomes paraelectric. Interestingly, the Rochelle salt also ceases to be ferroelectric
below 18 ı C. Other ferromagnetic materials are listed on Table 15.1.
The ferroelectric crystals can basically be classified into two groups. In the first
group are those consisting of permanent dipoles. The transition to the ferroelectric
state occurs through the ordering of the permanent dipoles and so the transition
is of the order-disorder type. In the second group are those consisting of induced
dipoles. The transition to the ferroelectric is due to the relative displacement of two
sublattices, one composed by anions and the other by cations.
The crystals of the first group are represented by potassium dihydrogen phos-
phate (KH2 PO4 ) or KDP. In this substance phosphate groups are connected by
hydrogen bonds. Each hydrogen ion can be located near one or the other of two

(a) 0.04 (b) 0.3

0.03
0.2
P (C/m )
P (C/m )

2
2

0.02

TGS 0.1
0.01 TGBe

0 0
0 20 40 60 80 20 40 60 80 100 120
o o
θ ( C) θ ( C)

Fig. 15.1 (a) Spontaneous polarization of triglycine sulfate (TGS) and triglycine fluoberyllate
(TGFB) obtained experimentally by Hoshino et al. [161]. (b) Spontaneous polarization of barium
titanate obtained experimentally by Merz [163]
15.1 Dielectric Materials 297

Table 15.1 Transition Ferroelectrics T0

temperature T0 , in K, to the BaTiO3 380
paraelectric phase of
ferroelectric and KNbO3 691
antiferroelectric crystalline PbTiO3 763
substances. Some substances LiNbO3 1483
(d) decompose before losing LiTaO3 938
the ferroelectricity. Source: SbSI 295
LB
HCl 98
HBr 90
NaNO2 437
KNO3 398
KH2 PO4 123
RbH2 PO4 147
CsH2 PO4 152
KH2 AsO4 97
RbH2 AsO4 110
CsH2 AsO4 143
KD2 PO4 213
(NH4 )2 SO4 224
(NH4 )2 BeF4 176
CH3 NH3 Al(SO4 )2 12H2 O 177
C(NH2 )3 Al(SO4 )2 6H2 O d
Ca2 B6 O11 5H2 O 266
SC(NH2 )2 169
N(CH3 )4 HgCl3 d
(NH2 CH2 COOH)3 H2 SO4 321
(NH2 CH2 COOH)3 H2 SeO4 295
(NH2 CH2 COOH)3 H2 BeF4 343
NaKC4 H4 O6 4H2 O 297
Antiferroelectrics T0
PbZrO3 503
NaNbO3 638
NH4 H2 PO4 148
NH4 H2 AsO4 216
ND4 D2 PO4 242

phosphate groups, which determines the orientation of the dipole attached to it. The
ordering of the dipoles, which characterizes the ferroelectric state, is a consequence
of the ordering of hydrogen ions. At high temperatures, thermal agitation causes
the disordering of hydrogen ions and therefore of the dipoles, which characterizes
the paraelectric state. To observe the ferroelectric state in KDP it is necessary to
reduce its temperature below 150 ı C. Other representatives of this group are the
triglycine sulphate ((NH2 CH2 COOH)3 H2 SO4 ) or TGS and triglycine fluoberyllate
298 15 Dielectrics

((NH2 CH2 COOH)3 H2 BeF4 ) or TGFB whose spontaneous polarizations are shown
in Fig. 15.1a as functions of temperature.
The ferroelectrics of the second group are represented by the barium titanate
(BaTiO3 ). In this substance there are two sublattices, one of them consisting of
negative ions and the other of positive ions. The ferroelectric state arises by displace-
ment of an entire sublattice in relation to the other. At room temperature BaTiO3 is
ferroelectric and has a tetragonal structure. If heated, it undergoes a transition to
the paraelectric phase, at a temperature of 107 ıC, as shown in Fig. 15.1b, passing
to have a cubic structure. If cooled, however, it passes through two ferroelectric
transitions. One at the temperature of 0 ı C, acquiring the orthorhombic structure,
and another at 90 ı C acquiring the rhombohedral structure. All three transitions
are discontinuous.
The ferroelectric transitions represent a special subclass of the class of structural
transitions which are those in which the crystalline structure undergoes a change.
As an example of a structural transition, we mention the compound SrTiO3 which
undergoes such a transition at a temperature of 110 K. Above this temperature,
the compound has the structure of perovskite. Below that, the compound displays
a structure that is distorted in relation to the original structure. This distortion,
however, does not entail the emergence of a spontaneous polarization in SrTiO3 .
If the distortion is accompanied by the emergence of the spontaneous polarization,
then it is also a ferroelectric transition, as happens with the compound BaTiO3 . Is
also worth mentioning that some crystals have other types of electrical ordering,
with zero total polarization. Such states are called antiferroelectrics as is the case,
for example, of PbZrO3 crystal.

15.2 Dielectric Properties

Electric Work

To study the effects of the electric field on a dielectric we use a capacitor within
which the dielectric sample is placed. The capacitor is maintained at a potential
difference which generates an electric field within the capacitor. The electric dipole
moment per unit volume or volumetric polarization P and the electric field E are
related to the electric displacement D by

D D 0 E C P; (15.3)

where 0 is the permittivity of vacuum. For a uniform electric field, we can show
that the infinitesimal work performed on the dielectric dWele is given by

dWele D E  dP (15.4)

where P is the total electric dipole of the sample.

15.2 Dielectric Properties 299

The demonstration of this result is as follows. Consider a capacitor of parallel

plates of area A and distance between plates equal to ` within which is the dielectric.
The capacitor is maintained at a potential difference equal to V . We assume that
the electric field E and the electric displacement D inside the capacitor as well as
the volumetric polarization P are perpendicular to the plates of the capacitor and
neglect border effects. The work dW required to carry a free charge dqliv against the
potential is given by

dW D V dqliv : (15.5)

But V D E` and the free charges are related to the electric displacement D by
qliv D DA so that

dW D E`AdD D EVdD; (15.6)

where V D A` is the volume of the capacitor. On the other hand, if the capacitor
was empty the work would be

dW0 D 0 EVdE: (15.7)

Therefore, the electric energy transferred to the electric material inside the capacitor
will be dWele D dW  dW0 or

dWele D EVdP; (15.8)

or yet

dWele D EdP; (15.9)

where P D VP is the total electric dipole moment of the sample.

In the International System of Units, the electric dipole moment P is measured
in coulomb meter (C m) and the electric field E in volts per meter (V/m). The
electric displacement D and the volumetric polarization P are measured in C/m2 .
The electric dipole moment per mole, or molar polarization, is given in C m/mol.

Thermodynamic Potentials and Coefficients

The infinitesimal change in internal energy of a dielectric equals the sum of the heat
introduced dQ D TdS and the work performed on the electrical system dWele D
EdP, that is,

dU D TdS C EdP: (15.10)

300 15 Dielectrics

We remark that the total dipole moment P is an extensive quantity. Thus, we can
understand that the internal energy U of a sample of dielectric is a function of
the extensive variables S and P. Taking into account the number of moles N of
the sample, we consider U.S; P; N/ to be also a function of N, what represents
the fundamental relation of a dielectric in the energy representation. The entropy
S.U; P; N/ is also a fundamental relation for which

1 E
dS D dU  dP: (15.11)
T T

From U.S; P; N/ we can get other representations by means of Legendre

transformations. The enthalpy H.S; E; N/, the Helmholtz free energy F.T; P; N/ and
the Gibbs free energy G.T; E; N/ are defined by the Legendre transformations

H D min.U  EP/; (15.12)

P

F D min.U  TS/; (15.13)

P

G D min.F  EP/: (15.14)

P

From them, we obtain the results

dF D SdT C EdP; (15.15)

dH D TdS  PdE; (15.16)
dG D SdT  PdE: (15.17)

From (15.17), we see that

   
@G @G
SD ; PD : (15.18)
@T E @E T

From these quantities we can determine certain thermodynamic coefficients includ-

ing the isothermal electric susceptibility and the pyroelectric coefficient A, defined
by
   
@P @P
XT D ; AD ; (15.19)
@E T @T E

and the heat capacities at constant field and at constant polarization, defined by
   
@S @S
CE D T : CP D T ; (15.20)
@T E @T P
15.2 Dielectric Properties 301

The property of convexity of the thermodynamic potentials implies that the suscep-
tibility and the heat capacities are nonnegative quantities, XT  0, CE  0 and
CP  0.
The variation of the electrical polarization of a ferroelectric material caused by
the change in temperature is called pyroelectric effect. A measure of this effect is
given by the pyroelectric coefficient defined above.

Molar Quantities

The thermodynamic properties of a dielectric can be derived from the fundamental

relation u.s; p/, where u D U=N is the molar internal energy, s D S=N is the molar
entropy and p D P=N is the molar dipole moment or simply polarization. For a
dielectric

du D Tds C Edp: (15.21)

Inverting the relation u.s; p/ we get the fundamental relation in the entropy
representation s.u; p/. We can also define other molar quantities. Among them we
find the molar thermodynamic potentials: h D H=N, f D F=N and g D G=N for
which

dh D Tds  pdE; (15.22)

df D sdT C Edp; (15.23)
dg D sdT  pdE: (15.24)

The electric susceptibility and the molar heat capacities are given by
     
@p @s @s
D ; cE D T ; cP D T : (15.25)
@E T @T E @T P

Debye Law

Let us first examine the properties of an ideal dielectric, understood as a system

consisting of noninteracting microscopic dipoles or such that the interactions are
negligible. Examples of such systems are gases at low pressures. In this state, the
molecules are sufficiently far from each other so that the electric field acting on a
molecule, due to the other, can be neglected when compared with the applied field.
302 15 Dielectrics

The molar polarization p of an ideal dielectric is composed of two parts: an induced,

p1 , and another, p2, permanent. The induced part is independent of temperature and
simply proportional to the applied electric field

p1 D ˛E; (15.26)

where ˛ is the molar polarizability. The permanent part depends on the temperature.
By analogy with ideal paramagnetism we postulate the following form due to
Langevin
 
p0 E RT
p2 D p0 coth  ; (15.27)
RT p0 E

where p0 is the permanent molar polarization. Therefore, the molar polarization

p D p1 C p2 of an ideal dielectric is given by
 
p0 E RT
p D ˛E C p0 coth  : (15.28)
RT p0 E

For small values of the electric field, the expression in parentheses behaves as
p0 E=3RT, so that the polarization varies with the field in accordance with

p D 0 E; (15.29)

where

p20
0 D ˛ C ; (15.30)
3RT
which is the Debye law. The second term, which is analogous to the Curie law,
corresponds to the contribution of permanent dipoles. The first term is due to
induced dipoles that may also be present in the polar molecules.
In an ordinary dielectric, as we have seen, the relation between the volumetric
polarization (dipole moment per unit volume) P and the field is P D E. N Given
that the dipole moment per unit volume P and dipole moment per mole p are
connected by p D vP, where v is the molar volume, then 0 D v , N from which
follows the relation between 0 and the dielectric constant 

0 D v0 .  1/: (15.31)

Thus we see that 0 can be experimentally determined from the measurement of

dielectric constant and the molar volume.
15.2 Dielectric Properties 303

Fig. 15.2 Dielectric constant 0.010

 of several gases obtained H 2O
experimentally by Zahn SO2

[167], and Zahn and Miles 0.008 NH3

HCl
[167] and of the water vapor H 2S
obtained experimentally by

(κ-1)v*
HBr
Sanger et al. [164]. The graph 0.006 CO2
shows the variation of CO
.  1/v  with the inverse of N2
0.004
temperature T, where
v  D v=v0 and v0 is the
molar volume of an ideal gas 0.002
at the pressure of 1 atm and at
0 ıC
0
0 0.001 0.002 0.003 0.004 0.005
-1
1/T (K )

Replacing (15.31) into (15.30), Debye law can be written in the form

1 p2
.  1/v D .˛ C 0 /: (15.32)
0 3RT

Figure 15.2 shows .  1/v versus 1=T for various gases. By Debye law (15.32),
the slope of the straight line fitted to the experimental data gives the permanent
polarization p0 .
The dense gases as well as the liquids and solids deviate from ideal behavior
and Debye law ceases to be valid. In these states, due to the proximity between
molecules, one can not neglect the mutual action between the dipoles. A dipole does
not feel only the external field but also the field generated by the other molecule
dipoles. An approximate form for the dielectric properties of these systems is that
given by the Clausius-Mossotti relation, which is obtained by replacing the field
E in (15.29) by E C P=30 , where the second term is the Lorentz field. With the
Lorentz correction, Debye law becomes

1 1 p2
vD .˛ C 0 /: (15.33)
C2 30 3RT

For nonpolar molecules, p0 D 0, and this equation is reduced to the Clausius-

Mossotti relation
 1 ˛
vD : (15.34)
C2 30
304 15 Dielectrics

15.3 Devonshire Theory

Discontinuous Transition

To explain the transition between paraelectric and the ferroelectric phases occurring
in barium titanate, Devonshire postulated the following expression for the molar
Helmholtz free energy

f D f0 C b.T  T1 /p2  cp4 C dp6 ; (15.35)

where the constants b, c, d and T1 are positive and f0 depends only on temperature.
The negative coefficient of p4 guarantees that the transition is discontinuous as we
will see below. The coefficient of p2 can be either positive or negative according to
whether T > T1 or T < T1 .
The electric field E D .@f =@p/T is given by

E D 2b.T  T1 /p  4cp3 C 6dp5 ; (15.36)

and the susceptibility  D .@p=@E/T is given by

1
D 2b.T  T1 /  12cp2 C 30dp4: (15.37)


Therefore, in the paraelectric phase, for which p D 0, the susceptibility is given by

1
D : (15.38)
2b.T  T1 /

Expression (15.35) is not always convex in p. In the cases where it is not

convex, the free energy of the system should be understood as the convex hull of
expression (15.35) obtained by the minimization of f  Ep.
At zero field, E D 0, the thermodynamic properties are therefore obtained by
the minimization of expression (15.35). That is, the polarization corresponds to the
absolute minimum of (15.35).
In the paraelectric phase p D 0 is a absolute minimum of f . In the ferroelectric
phase, the absolute minimum occurs at p D p , where p is the spontaneous
polarization, which is the solution of (15.36) for E D 0, that is,

b.T  T1 /  2cp2 C 3dp4 D 0; (15.39)

whose solution is
s r
c c 3bd
p D C 1 .T  T1 /: (15.40)
3d 3d c2
15.3 Devonshire Theory 305

which gives the variation of p with temperature.

The phase transition occurs when the Helmholtz free energy of the ferromagnetic
phase f . p / becomes equal to that of the paraelectric phase f .0/, that is, when
f . p / D f .0/, which leads us to equation

b.T0  T1 /  cp2 C dp4 D 0: (15.41)

This equation together with equation (15.39) determine the transition temperature,
which we denote by T0 , and the value of p at the transition, which we denote by
p0 , that is,

c2
T0 D T1 C ; (15.42)
4bd
r
c
p0 D : (15.43)
2d

Eliminating T1 from expression (15.40), we can write the dependence of p with

temperature in the following form
s r
2 1 12bd
p D p0 C 1 .T  T0 /; (15.44)
3 3 c2

valid for T  T0 . Above T0 , the polarization vanishes so that the polarization has a
jump at T D T0 equal to p0 , characterizing a discontinuous transition.
If the system described by (15.35) is subject to an electric field, the phase
transition does not disappear but is displaced to another value of temperature as
long as the field does not exceed a certain value Ec . For fields above Ec the transition
disappears. Thus the phase diagram E versus T has a discontinuous transition line
that starts at the point T D T0 and E D 0 and ends at the critical point T D Tc and
E D Ec , as seen in Fig. 15.3
The properties described above are obtained by the minimization of . p/ D
f .T; p/  Ep, given by

. p/ D f0 C b.T  T1 /p2  cp4 C dp6  Ep: (15.45)

Along the transition line two phase coexist: a paraelectric phase characterized by
p D p1 and another characterized by p D p2 . Therefore, these two values of p must
be the absolute minima of . p/ what results in the following form

.p/ D 0 C d. p  p1 /2 . p  p2 /2 Œ. p C r1 /2 C r22 ; (15.46)

306 15 Dielectrics

(a) (b)
E p

0
T0 Tc T E

Fig. 15.3 (a) Phase diagram electric field E versus temperature T of a system described by the free
energy (15.35). The thick lines are discontinuous transitions that meet at the triple point T D T0
and E D 0. The full circles are critical points. (b) Isotherms of polarization p versus E. Three
isotherms are shown: one below the temperature of the triple point T0 , another between T0 and the
critical temperature Tc and another one above Tc

where r1 and r2 are nonzero. Developing the product and comparing with (15.45),
we get the relations

r1 D p1 C p2 ; r22 D p1 p2 ; (15.47)
c D d.2r12  3r22 /; b.T  T1 / D d.r14 C 3r24  r12 r22 / (15.48)

and

E D 2dr13 r22 : (15.49)

These equations give implicitly the coexistence line between the paraelectric and
ferroelectric phases. Near the transition point T D T0 at zero field E D 0, we get
the following behavior

1
b.T  T0 / D jEj: (15.50)
p0

The terminal point of the line define the critical point, which is obtained form
equations above when p2 D p1 , that is, when r2 D r12 =4, which gives the results

3c2  c 5=2
Tc D T1 C and Ec D 16d ; (15.51)
5bd 5d
or yet

7c2
Tc D T0 C : (15.52)
20bd
15.3 Devonshire Theory 307

Continuous Transition

The majority of ferroelectric substances pass to the paraelectric phase through a

first order transition as happens, for example, with the barium titanate. Some fer-
roelectrics, however, undergoes a continuous transition. According the Devonshire
theory, the free energy of such systems is give by

f D f0 C b.T  Tc /p2 C cp4 ; (15.53)

where the constants b and c are positive and f0 depends only on temperature.
The electric field E is given by

E D 2b.T  Tc /p C 4cp3 ; (15.54)

from which we get the inverse of susceptibility

1
D 2b.T  Tc / C 12cp2: (15.55)


At zero field, p D 0 for T  0 and in this case the susceptibility is

1
D : (15.56)
2b.T  Tc /

For T < Tc , a spontaneous polarization emerges, given by

r
b
pD .Tc  T/; (15.57)
2c

which vanishes continuously when T ! Tc , characterizing a continuous phase

transition. In this case the susceptibility is

1
D : (15.58)
4b.Tc  T/

The molar pyroelectric coefficient ˛ D A=N D .@p=@T/E at zero field has therefore
the following behavior
s
b
˛D ; (15.59)
8c.Tc  T/

diverging when T ! Tc .
308 15 Dielectrics

Ferroelectric Transitions

In addition to the transition between the paraelectric and ferroelectric phases which
occurs at 107 ı C, the barium titanate undergoes two more transitions that occur
at 0 ı C and at 90 ı C. These transitions are discontinuous and occur among
ferroelectric phases. In the paraelectric phase the barium titanate has cubic crystal
structure. In the ferroelectric phases the structures are tetragonal, orthorhombic and
rhombic.
The three discontinuous phase transitions occurring in BaTiO3 can be obtained
from the following free energy

f D f0 C b.T  T1 /. p2x C p2y C p2z /  c. p4x C p4y C p4z /

C a. p2x p2y C p2x p2z C p2y p2z / C d. p6x C p6y C p6z /; (15.60)

where f0 depends only on temperature and the coefficients a, b, c and d are positive.
The paraelectric phase is characterized by the absence of spontaneous polarization,
that is, px D py D pz D 0. The three ferroelectric phases are characterized by

px D py D 0 and pz ¤ 0; (15.61)
px D 0 and py D pz ¤ 0; (15.62)
px D py D pz ¤ 0: (15.63)

If we set px D py D 0 and pz D p, the free energy is reduced to the expression

f D f0 C b.T  T1 /p2  cp4 C dp6 ; (15.64)

which we have used previously to describe the ferroelectric to paraelectric transition.

Setting px D 0 and py D pz D p we get

f D f0 C 2b.T  T1 /p2  .2c  a/p4 C 2dp6 : (15.65)

Finally, setting px D py D pz D p we obtain

f D f0 C 3b.T  T1 /p2  3.c  a/p4 C 3dp6 : (15.66)

At zero field, the spontaneous polarization are obtained by the minimization of

the above expressions and the transition temperatures by comparison of the free
energies corresponding to the several thermodynamic phases.
15.4 Order-Disorder Transition 309

15.4 Order-Disorder Transition

Hydrogen Bonds

In ferroelectric materials such that the hydrogen bonds have a relevant role, the fer-
roelectric state is described by the ordering of the hydrogen ions and consequently
by the ordering of the permanent dipoles which result in a spontaneous polarization.
In the paraelectric phase the dipoles are disordered so that the polarization vanishes.
Among ferroelectrics whose mechanism of transition is of the type order-disorder
we find the potassium dihydrogen phosphate dihydrate (KH2 PO4 ) or KDP and its
isomorphs and the triglycine sulfate, (NH2 CH2 COOH)3 H2 SO4 , and its isomorphs.
In analogy with ferromagnetic systems we postulate the following free energy
for such systems
 
p2 p0 C p p0 C p p0  p p0  p
f D ˛ C RT ln C ln ; (15.67)
2 2p0 2p0 2p0 2p0

where ˛ is a positive constant. The electric field E D @f =@p is given by

RT p0 C p
E D ˛p C ln : (15.68)
2p0 p0  p

At zero field, we get

RT p0 C p
˛p D ln : (15.69)
2p0 p0  p

or equivalently
p0 ˛
p D p0 tanh. p/: (15.70)
RT
From this equation, we conclude that above the critical temperature Tc , given by

p20 ˛
Tc D ; (15.71)
R
the polarization vanishes and the system finds itself in the paraelectric phase. Below
this temperature, p ¤ 0 and the system exhibits a spontaneous polarization. Near
the critical temperature, p behaves as
s
T
p D p0 3.1  /: (15.72)
Tc
310 15 Dielectrics

Isotopic Effect

An important effect observed in phase transitions of compounds of the KDP type is

that the critical temperature becomes approximately twice greater when hydrogen is
replaced by deuterium, called isotopic effect. This means that the quantum tunneling
of the hydrogen between two equilibrium positions located at the hydrogen bonds
has a relevant role. When deuterium, which has a larger mass, replaces hydrogen,
the tunneling becomes less frequent causing a greater stability of the ferroelectric
state.
According to de Gennes the tunneling corresponding to the model introduced by
Blind can be written by means of a transverse field, that is, by means of a field that
acts perpendicularly to the direction of the electric dipoles. This field give rise to
a polarization in the transversal direction, which we denote by q. Thus in analogy
with a ferromagnetic system in the presence of perpendicular field we postulate the
following free energy
 
˛ p0 C r p0 C r p0  r p0  r
f D  p2 C RT ln C ln ; (15.73)
2 2p0 2p0 2p0 2p0

where r is given by
p
rD q2 C p2 : (15.74)

In fact, the free energy must be understood as the convex hull of the right hand side
of (15.73) which is equivalent to minimize the expression f  Ep   q where  is
the transverse field.
The transverse field  D @f =@q is given by

RTq p0 C r
 D ln ; (15.75)
2p0 r p0  r

and the electric field E D @f =@p, by

RTp p0 C r
E D ˛p C ln : (15.76)
2p0 r p0  r

At zero field, the solution of this last equation is p D 0 and corresponds to the
paraelectric phase. In this case

RT p0 C q p0 
 D ln or q D p0 tanh : (15.77)
2p0 p0  q RT
15.4 Order-Disorder Transition 311

The spontaneous polarization, which characterizes the ferroelectric phase is the

nonzero solution of (15.76) for E D 0, that is, solution of equation

RT p0 C r
˛D ln : (15.78)
2p0 r p0  r

Comparing with (15.75), we see that the transverse field  is related with q through


qD ; (15.79)
˛
which replaced in the previous equation gives
p p p 
0
 2 C ˛ 2 p2 D ˛p0 tanh  2 C ˛ 2 p2 : (15.80)
RT
The spontaneous polarization vanishes continuously characterizing a continuous
transition between the ferroelectric and paraelectric phases. Taking the limit p ! 0
in (15.80) we get the critical temperature Tc , given by

p0 
 D ˛p0 tanh. /: (15.81)
RTc

Near the critical point the spontaneous polarization behaves as

p  .Tc  T/1=2 : (15.82)

It is worth noticing that in the limit T ! 0, we get from (15.80) the following
expression for the polarization
r
2
pD p20  ; (15.83)
˛2

that is, the saturation polarization depends on the ratio  =˛ between the transverse
field  and the interaction ˛ between dipoles.

Slater-Takagi Theory

As we have seen, the ferroelectric state of KDP is related with the ordering of the
protons located at the hydrogen bonds that connect the phosphate groups. Each
phosphate group is represented by a tetrahedron which is characterized according to
the configurations of the protons in its four vertices. To each one of the 16 possible
configurations of protons in the vertices of the tetrahedron, the Slater-Takagi theory
associates a certain dipole moment and a certain energy as follows. Among the six
312 15 Dielectrics

configurations that have two protons, four have moment equal to zero and energy 0 ,
one has moment equal to C2 and energy zero and one has moment equal to 2
and energy zero. Among the eight configurations that have one or three protons, four
have moment equal to C and energy 1 and four have moment equal to  and
energy 1 . The other configurations are forbidden.
The molar Gibbs free energy g.T; E/ is determined by

g D minf f  Epg; (15.84)

p

where the molar Helmholtz free energy f and the molar polarization p depend on
the mole fractions of each type of configurations of the tetrahedrons. According to
Ishibashi, the molar Helmholtz free energy f that is obtained from the Slater-Takagi
theory is given by

f D u  Ts; (15.85)

where u is the molar energy, given by

u D 40 c0 C 41 .dC C d /; (15.86)

and s is the molar entropy, given by

s D Rf2.cC C 2c0 C 3dC C d / ln.cC C 2c0 C 3dC C d /

C2.c C 2c0 C dC C 3d / ln.c C 2c0 C dC C 3d /

 cC ln cC  c ln c  4c0 ln c0  4dC ln dC  4d ln d g; (15.87)

where c0 , cC , c , dC and d are the mole fractions of each type of configuration of

the tetrahedrons and are connected by the relation

4c0 C cC C c C 4dC C 4d D 1: (15.88)

The polarization p is given by

p D Œ.cC  c / C 2.dC  d /p0 ; (15.89)

where p0 is the saturation polarization.

At zero field, which interest us here, g is obtained by the minimization of the
expression (15.85) for f , subjected to the restriction (15.88). The minimization leads
to the following results

1=2
c0 D cC c1=2
 0 ; (15.90)
15.4 Order-Disorder Transition 313

3=4
dC D cC c1=4
 1 ; (15.91)
1=4
d D c3=4
 cC 1 ; (15.92)

where 0 D e0 =RT and 1 D e1 =RT and

.cC C 2c0 C d /d D .c C 2c0 C dC /dC : (15.93)

These relations together with (15.88) determine the five mole fractions as functions
of temperature.
One solution is such that cC D c and corresponds to the paraelectric phase
since in this case d D dC and p D 0. The corresponding free energy is

1
g D RT ln .1 C 20 C 41 /: (15.94)
2

The solution such that cC ¤ c corresponds to the ferroelectric phase. It leads

us to the following value for the polarization
p
.1  20 /2  .21 /2
pD p0 ; (15.95)
1  20

and to the following free energy

 p 
g D RT ln 1 C 1 1  . p=p0 /2 : (15.96)

The transition from the ferroelectric phase to the paraelectric phase occurs
when the expressions (15.94) and (15.96) for the Gibbs free energy become equal.
The equality between them leads to the following expression for the transition
temperature T0

2e0 =RT0 C 2e1 =RT0 D 1: (15.97)

The original Slater theory corresponds to the case in which only the six configu-
rations with two protons are allowed. The other are forbidden what is equivalent
to saying that 1 ! 1 (1 D 0). In this case the ferroelectric phase has constant
polarization p D p0 for any temperature T < T0 and the transition temperature is
given by RT0 =0 D 1= ln 2. When 1 is finite, which corresponds to the Slater-Takagi
theory, the polarization decreases with temperature and vanishes continuously at the
transition temperature T0 .
It is worth mentioning that the Slater theory (1 D 0), for the case in which
0 D 0 (0 D 1), is reduced to the theory introduced by Pauling to explain the
residual entropy of ordinary ice. In this case, the six allowed configurations of the
tetrahedron are associated to the same energy and the system exhibits only the
314 15 Dielectrics

paraelectric phase. From (15.94) we see that g D RT ln.3=2/, from which we
conclude that s D @g=@T D R ln.3=2/ which is the result due to Pauling for the
residual entropy of the ordinary ice.

Problems

15.1 Electrocaloric effect. Show that .@T=@E/S D TA=CE , where A is the

pyroelectric effect defined by A D .@P=@T/E . Since CE  0, materials whose
polarization decreases with temperatures at constant field (A < 0) have the
temperature decreased when the field is decreased adiabatically.
15.2 Show that the molar free energy g.T; E/ of a dielectric that obeys the equation
of state (15.28) is given by
 
1 2 p0 E p0 E
g D  ˛E  RT ln sinh  ln :
2 RT RT

Determine the molar entropy s and show that the molar energy u D g C Ts C pE is
given by

1 2
uD ˛E C u0 ;
2
where u0 depends only on temperature.
15.3 At zero temperature the free energy (15.73) is reduced to
˛
f D  p2 ; p2 C q2  p20 :
2
Show that the minimization of f  Ep   q results in the polarization given
by (15.83).
15.4 Find the transition temperatures among the three ferromagnetic phases and
the paramagnetic phase of a system described by the free energy given by the
expression (15.60). Determine also the spontaneous polarization of the three
ferroelectric phases.
Chapter 16
Solids

16.1 Stress and Strain

Anisotropy

A solid substance is characterized by having a spatial structure consisting of a three-

dimensional lattice at whose vertices the atoms are located. Because the atoms are
in continuous motion due to thermal agitation, the vertices of the lattice are defined,
more properly, as the equilibrium positions around which the atoms vibrate. This
structure gives to the solid the rigidity that opposes not only the compression and
stretch but also shear. In other words, offering resistance to the volume and shape
changes. The spatial structures can be ordered as those of the crystalline solid or
disordered such as those of the amorphous solids and glasses.
Crystal structures are classified according to the symmetries they exhibit. The set
of symmetry operations of a crystalline structure comprises a symmetry group. The
symmetry groups are brought together in classes and the classes in crystal systems.
There are seven crystal systems: triclinic, monoclinic, orthorhombic, tetragonal,
trigonal, hexagonal and cubic. The structures of the cubic system can be simple
cubic (sc), body-centered cubic (bcc) and face-centered cubic (fcc). In a simple
solid, composed of one kind of atom, that orders according to the sc structure,
the atoms are located only at the vertices of the cubes forming the ordered spatial
structure. In a simple solid with ccc structure the atoms are at the vertices and at the
centers of the cubes and in a simple solid with fcc structure, they lie at the vertices
and centers of the sides of the cubes.
Due to the spatially ordered structure, crystalline solids, in contrast to gases and
liquids, are anisotropic. Elasticity may depend on the direction along which the solid
is pulled or compressed. The thermal expansion on a direction may differ from that
on a different direction. It should be noted that variations in properties arising from
the anisotropy are observed in single crystals. Polycrystals, and the amorphous and
glassy solids, are from the macroscopic point of view, examples of isotropic solids.

© Springer-Verlag Berlin Heidelberg 2017 315

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_16
316 16 Solids

This does not mean that they behave, as to the elastic properties, like liquids and
gases, which are also isotropic. Solids, isotropic or not, are distinguished from fluids
by resisting shear.

Stress Tensor

External forces acting on the surface of a solid can induce the appearance of
stress and shear inside the solid. In a fluid external forces only induce hydrostatic
compressions and possibly internal motions. The hydrostatic compressions are
forces perpendicular to the surface of any internal region and have the same
magnitude per unit area at any point on the surface. In a solid, in general, the forces
on an internal region are not perpendicular to the surface and do not have the same
magnitude per unit area. This complexity of the forces requires a more detailed
description of the mechanical equilibrium than that given simply by hydrostatic
compression. The internal forces in a solid are described by means of a quantity
similar to the pressure, called stress. Unlike pressure, which is a scalar, stress is a
tensor quantity and is defined as follows.
Consider a sample of a solid in the form of a rectangular parallelepiped of edges
`1 , `2 and `3 , with edges parallel to a Cartesian coordinate system, as shown in
Fig. 16.1a. We denote by Ef1 , Ef2 and Ef3 the forces acting on the faces A, B and C,
respectively. The forces acting on opposite sides are, respectively, Ef1 , Ef2 and Ef3
so that the resultant of the forces on the sample is zero. The decomposition of these
Cartesian vectors

Ef1 D f11 xO C f12 yO C f13 zO; (16.1)

Ef2 D f21 xO C f22 yO C f23 zO; (16.2)
Ef3 D f31 xO C f32 yO C f33 zO; (16.3)

z z
(a) R’
R (b)

C
O Q’
O Q
P
A B y x y
x P’

Fig. 16.1 Sample of a homogeneous solid in the form of a parallelepiped. (a) Before the
deformation the length and direction of the edges are described by the vectors OPD `1 xO , OQD `2 yO
and ORD `3 zO. (b) After the deformation they are described by OP’D È1 , OQ’D È2 and OR’D È3
16.1 Stress and Strain 317

defines the Cartesian components fij of these vectors. Since the sample is in
mechanical equilibrium torques on it should vanish. This requirement applied to
each of the three Cartesian axes results in the following condition

`i fij D `j fji : (16.4)

In order to describe efforts which are independent of sample size, what can be
done by considering homogeneous samples, we introduce forces per unit area

fij
ij D ; (16.5)
Ai

where Ai are the area of the three faces of the parallelepiped, given by A1 D `2 `3 ,
A2 D `3 `1 and A3 D `1 `2 . They have the unit of pressure and comprise the elements
of a tensor  called stress tensor. The diagonal elements describe the stretches and
are positive in a proper stretch and negative in a compression. The off-diagonal
elements describe the shear stresses. Due to property (16.4) we see that

ij D ji ; (16.6)

which means that the stress tensor is symmetrical.

The stress tensor can be decomposed into a tensor of zero trace and a tensor that
represents a hydrostatic compression. Indeed, from the identity

ij D .ij C pıij /  pıij ; (16.7)

where p D .11 C 22 C 33 /=3, we see that the tensor represented by the first
term has zero trace. The second term constitutes the components of a tensor that
has only diagonal elements and are all equal. This second tensor, defined by pıij ,
represents a hydrostatic compression at a pressure p since the forces Efi that describe
a hydrostatic compression are perpendicular to the faces of the parallelepiped of
Fig. 16.1a and are proportional to the areas of the faces, that is, fij D pAi ıij .

Strain Tensor

The description of the mechanical state of a sample of a solid can not be made
only by its volume as happens to a fluid in thermodynamic equilibrium. Due to
anisotropy, the deformations resulting from forces of the same intensity may have
different magnitudes and may depend on the direction of the forces. In other words,
the deformations are not isotropic. To properly describe the mechanical state of a
solid we introduce a quantity called strain tensor, defined as follows.
Consider again the sample of a homogeneous solid in the form of a rectangular
parallelepiped with edges `1 , `2 , and `3 , as shown in Fig. 16.1a. Then consider a
318 16 Solids

deformation of this body as shown in Fig. 16.1b. The lengths and directions of edges
of the parallelepiped are described by the vectors È1 , È2 and È3 , whose Cartesian
decomposition

È1 D `11 xO C `12 yO C `13 zO; (16.8)

È2 D `21 xO C `22 yO C `23 zO; (16.9)
È3 D `31 xO C `32 yO C `33 zO; (16.10)

define the Cartesian components `ij of these vectors.

In order to describe the deformation of the sample regardless of size we introduce
the relative displacements ij defined by

`ij D `i .ıij C ij /: (16.11)

They are dimensionless and comprise the elements of a tensor called strain tensor.
The diagonal elements describe stretch deformations and the off-diagonal elements
describe shear deformations. Given that the sample is homogeneous, the strain
tensor does not depend on the size of the sample used to determine it. The volume
V0 of the region before the deformation is

V 0 D `1 `2 `3 : (16.12)

After deformation the volume V is

V D j È1  È2  È3 j: (16.13)

The mechanical state of a solid body in equilibrium is defined not only by the
strain tensor but also by a reference state from which we define this tensor. We
adopt as the reference state a rectangular parallelepiped of edges `1 , `2 and `3 like
that shown in Fig. 16.1a. Any other state, like the one shown in Fig. 16.1b, is defined
by the vectors È1 , È2 and È3 , which are related to the strain tensor and the reference
state by means of (16.11).
The components of the strain tensor describe deformations of any magnitude.
However, elastic deformations of a solid body are small in general which allows us
to restrict ourselves to the study of the case in which the components of the strain
tensor is small compared with the unit, that is, ij << 1. In this regime of small
strain the sample volume, given by (16.13) reduces to

V D V0 .1 C 11 C 22 C 33 /; (16.14)

expression valid up to linear terms in ij and depends only on the diagonal

components of the strain tensor.
16.1 Stress and Strain 319

An equivalent description of the mechanical state of a solid in equilibrium can be

made through the variable V given by (16.14) together with the variables ij defined
by

1
ij D ij  ıij . 11 C 22 C 33 /; (16.15)
3
which constitute the components of the tensor . This tensor has zero trace and for
this reason only two of its diagonal components can be chosen as independent. In a
process in which the volume remains constant, it describes the change in shape.

Work of Elastic Forces

To determine the work of elastic forces we look at the forces Ef1 , Ef2 and Ef3 acting on
the faces A, B and C, whose Cartesian components are defined by (16.1)–(16.3).
The forces acting on opposite faces are considered to be equal in magnitude but in
opposite directions. These forces are related to the stress tensor in accordance with

fij D Ai ij ; (16.16)

where Ai are the areas of the projections of the faces of the oblique parallelepiped
of Fig. 16.1b on the three planes of the Cartesian system. As we are in the regime
of small deformations the areas of the sides A, B and C can be approximated by
A1 D `2 `3 , A2 D `3 `1 and A3 D `1 `2 .
To determine the work of forces we consider the displacements of the faces.
Since the forces acting on opposite faces are antiparallel, we assume that the
displacements of the faces are antiparallel. Denoting by d Èi =2 the infinitesimal
displacement of a face, on which acts the force Efi , then the infinitesimal displacement
of the opposite face will be d Èi =2 and will be subject to a force Efi . The
infinitesimal work of the forces acting on the six faces of the sample is then
X X
dW D Efi  d Èi D fij d`ij : (16.17)
i ij

According to the definition of the strain tensor, d`ij D `i d ij . Using this result and
the relation (16.16) between the forces and the stress tensor, we obtain the following
result for the infinitesimal work
X
dW D V0 ij d ij : (16.18)
ij

where V0 D `1 `2 `3 is the volume of the sample in the reference state. The volume
V in state ij , in the regime of small deformations, is given by (16.14) from which
320 16 Solids

we conclude that

dV D V0 .d 11 C d 22 C d 33 /: (16.19)

When the stress tensor  corresponds to a hydrostatic pressure, that is, ij D pıij ,
we see that the work is reduced to dW D pdV, which is the form used in gases
and liquids.
Given that the strain tensor is symmetric, there are only six independent
components of the tensor. The expression for the work will have as a consequence
only six independent terms. Thus it is appropriate to introduce the stress components
i , defined by 1 D 11 , 2 D 22 , 3 D 33 , 4 D 23 , 5 D 13 and 6 D 12 , and
the strain components i , defined by 1 D 11 , 2 D 22 , 3 D 33 , 4 D 23 C 32 ,
5 D 13 C 31 and 6 D 12 C 21 . The infinitesimal work reads then

6
X
dW D V0 k d k ; (16.20)
kD1

and the infinitesimal volume reads

3
X
dV D V0 d k : (16.21)
kD1

Thermodynamic Potentials

The variation of the internal energy U of a sample is the sum of the heat introduced
and the work of the forces acting on the sample, given above. In differential form

6
X
dU D TdS C V0 k d k : (16.22)
kD1

The internal energy U.S; / is thus a function of entropy S and the strain components
k . From U we get other thermodynamic potentials such as the Helmholtz free
energy F D U  TS, whose differential is

6
X
dF D SdT C V0 k d k ; (16.23)
kD1

from which we get the relations

 
1 @F
i D ; (16.24)
V0 @ i T
16.2 Thermodynamic Coefficients 321

where the derivative is carried out by keeping constant the other strain components
and the temperature. The Helmholtz free energy F.T; / is a function of temperature
T and the strain components k .
The Gibbs free energy G D F C pV is obtained by a Legendre transformation
from F. Using the description of the mechanical state defined by the tensor  and
the volume V, we get the following differential form
X
dG D SdT C Vdp C V0 i di : (16.25)
i

We must remember however that 1 , 2 and 3 are not independent because 1 C

2 C 3 D 0.
In addition to F and G other thermodynamic potentials, obtained by Legendre
transformations, can be used to describe the elastic properties a solid. A particularly
useful potential is the one defined by ˚ D F  V0 .1 1 C : : : C 6 6 / and whose
differential form is
6
X
d˚ D SdT  V0 k dk ; (16.26)
kD1

from which follows

 
1 @˚
i D  ; (16.27)
V0 @i T

where the derivative is performed keeping constant the other stress components and
the temperature.

16.2 Thermodynamic Coefficients

Elastic Coefficients

Several thermodynamic coefficients are used to characterize the thermodynamic

properties of solids. These coefficients are similar to those used in the description of
gases and liquids such as the compressibility, the coefficient of thermal expansion
and heat capacity. The coefficients that describe the elasticity relate the variations
of stress and variations in strain in the form of quotients. The isothermal stiffness
coefficient is defined by
 
@i
cij D ; (16.28)
@ j T
322 16 Solids

where the derivative is performed by keeping constant the temperature and the other
strain components. The property of convexity of the potential F with respect the set
of variables i implies that matrix with coefficients cij has nonnegative eigenvalues.
particularly the convexity of F implies cii  0.
The isothermal compliance coefficient is defined by
 
@ i
ij D ; (16.29)
@j T

where the derivative is performed keeping constant the temperature and the other
stress components. The property of convexity of the potential ˚ with respect to the
set of variables i implies that the matrix with the coefficients ij has nonnegative
eigenvalues. Particularly the convexity of ˚ implies ii  0.
The coefficient ij is not generally the inverse of cij . However, they are closely
linked. Considering in one hand that the stress components i are functions of strain
components j and on the other hand that latter are functions of the former then the
following identities can be achieved

X6    
@i @ j
D ıik : (16.30)
jD1
@ j T @k T

From this identity is follows immediately the result

6
X
cij jk D ıik ; (16.31)
jD1

which says the matrix with elements ij is the inverse of the matrix with elements
cjk .
The coefficients cij are not all independent. Using relations analogous to Maxwell
relations
   
@i @j
D ; (16.32)
@ j T @ i T

it follows cij D cji . Similarly we may conclude that ij D ji . Therefore, from
the 36 coefficients cij , or ij , only 21 are independent. The number of independent
coefficients can be further reduced if we take into account the symmetries of the
crystalline structure.
The triclinic system does not impose any restriction and has therefore 21
independent coefficients. However, in this case it is possible to orient the sample
such that three coefficients vanish reducing the number of independent coefficients
to 18. In the monoclinic system, the number of independent coefficients is 13 but
the orientation of the sample can reduce them to 12. In the orthorhombic system
the number is nine. Some classes of the trigonal and tetragonal systems have 7
16.2 Thermodynamic Coefficients 323

independent coefficients but they may be reduced to 6 by the orientation of the

sample. Other classes of these two systems have six independent coefficients. The
hexagonal system has 5 and cubic system has only 3.
In the cubic system we take as independent c11 , c12 and c44 . The other are given
by

c33 D c22 D c11 ; c23 D c13 D c12 ; c66 D c55 D c44 : (16.33)

The remaining coefficients vanish. These results are summarized in the matrix
0 1
c11 c12 c12 0 0 0
B C
B c12 c11 c12 0 0 0 C
B C
B c12 c12 c11 0 0 0 C
B C: (16.34)
B 0 0 0 c44 0 0 C
B C
@ 0 0 0 0 c44 0 A
0 0 0 0 0 c44

The convexity of F implies that the matrix (16.34) has nonnegative eigenvalues
which result in the following restrictions

c11  c12  0; c11 C 2c12  0; c44  0: (16.35)

From the firs two we get c11  0.

Relations similar to those of the stiffness coefficients cij related to the cubic
system are valid for the compliance coefficients

33 D 22 D 11 ; 23 D 13 D 12 ; 66 D 55 D 44 : (16.36)

The other coefficients ij vanish. The convexity of the potential ˚ implies the
conditions

11  12  0; 11 C 212  0; 44  0; (16.37)

and 11  0, which follows from the first two.

Compressibility

A solid body immersed in a liquid undergoes compression which are perpendicular

to its surface and of the same intensity anywhere in the surface. Under these
conditions the solid body undergoes a hydrostatic compression represented by the
324 16 Solids

pressure p. The compressibility  is defined as the ratio of the relative decrease of

the volume and the increase in hydrostatic pressure
 
1 @V
D : (16.38)
V0 @p T

This expression defined the isothermal compressibility. The adiabatic compressibil-

ity is defined similarly.
The compressibility of solids is related with the elasticity coefficients. Indeed,
using the result .@V=@ k / D V0 , k D 1; 2; 3, which follows from (16.21), we see
that
3 
X  X 3    
@ i @ i @j
D D : (16.39)
iD1
@p T i;jD1
@j T @p

Since 1 D 2 D 3 D p for a hydrostatic compression and using the definition

of the coefficient ij , we get the relation

3
X
D ij : (16.40)
i;jD1

The bulk modulus B is defined as the inverse of the compressibility, B D 1=.

For a solid with cubic symmetry

 D 3.11 C 212 /: (16.41)

Using the relation between ij and cij we get B D 1= in terms of cij for the cubic
system

1
BD .c11 C 2c12 /: (16.42)
3

Thermal Expansion and Heat Capacity

The coefficients of thermal expansion are defined by

 
@ i
˛i D ; (16.43)
@T

where the derivative is carried out by keeping all the six stress components j fixed.
Recalling that the stress tensor is symmetric, we can make a rotation of the Cartesian
axes so as to transform the stress tensor in diagonal form. From this result we
16.2 Thermodynamic Coefficients 325

conclude that only three coefficients of expansion are independent. This happens
with crystalline solids of the systems triclinic, monoclinic and orthorhombic. Solids
of the tetragonal, trigonal and hexagonal crystalline systems have greater symmetry
and are described by two independent coefficients of expansion. The crystalline
solids of the cubic system, the highest symmetry, are described by only one
coefficient of expansion and accordingly behave like isotropic solids. Therefore, for
a solid cubic symmetry, ˛1 D ˛2 D ˛3 . An immediate consequence of this result is
that thermal expansion of polycrystalline samples of cubic structure do not produce
tensions between monocrystalline grains because each expands isotropically. In
contrast, the thermal expansion of polycrystalline samples of noncubic structure
causes internal tensions because the monocrystalline grains expand in distinct forms
along a same direction.
In analogy with fluid, for which one defines heat capacity at constant volume
and at constant pressure, here we define the heat capacity for which all the strain
components are kept constant, given by
 
@S
C D T ; (16.44)
@T

and the heat capacity for which all stress components i are kept constant, given by
 
@S
C D T : (16.45)
@T 

Hooke Law

Consider a sample of a crystalline solid in a nondeformed state at a temperature T0 ,

that is, such that the stresses and compressions are absent. This nondeformed state
is described by the vanishing of the strain components, which we consider to be the
reference state. Next we consider a deformation of the sample which passes to a be
in a state defined by the strain components i and temperature T.
Assuming the strain to be small, that is, such that i << 1, we expand the free
energy f D F=V0 in power of the strain components up to quadratic terms to obtain
the result
X 1X X
f D f0  .T  T0 / ai i C cii i2 C cij i j ; (16.46)
i
2 i i<j

where f0 depends only on T. The coefficients ai and cij are determined at the
temperature T0 .
326 16 Solids

From this expansion for f and using (16.24) we get the stress components,

6
X
i D ai .T  T0 / C cij j : (16.47)
jD1

If the process is isothermal, T D T0 , this equation is the expression of the Hooke

law for crystalline solids.
Taking into account that the matrix ij is the inverse of the matrix cij , we reach
from (16.47) the result

6
X
i D ˛i .T  T0 / C ij j ; (16.48)
jD1

where
6
X
˛i D ai ij ; (16.49)
jD1

is the coefficient of thermal expansion. If the process is isothermal equation (16.48)

is again the expression of the Hooke law.
Suppose next a thermal expansion carried out in the absence of tensions and
compressions. This means a process such that i D 0. In this case,

i D ˛i .T  T0 /: (16.50)

For solids with cubic symmetry, the deformation is isotropic, ˛3 D ˛2 D ˛1 and

3 D 2 D 1 .

Isotropic Solids

We have seen that the compliance coefficients that describe a crystalline solid
are not all independent. Each crystal system imposes constraints arising from the
symmetries that reduce the number of the independent coefficients. The lowest
number is three and corresponds to the crystal system with highest symmetry which
is the cubic system. The isotropic solids impose an additional restriction reducing
the number to only two. Many solid behave as isotropic with respect to the elastic
properties. Examples of isotropic solids include the glassy and amorphous solids
and somehow also the polycrystalline solids.
The relation among the stiffness coefficients are the same as that valid for solids
of the cubic system, given by (16.33), with the additional restriction c11 D c12 C2c44
which is equivalent to the parametrization c11 D
C 2 , c12 D
and c44 D ,
16.2 Thermodynamic Coefficients 327

where
and are the Lamé coefficients. The matrix of the coefficients is
0 1

C 2

0 0 0
B

C 2
C
B 0 0 0 C
B C
B


C 2 0 0 0 C
B C: (16.51)
B 0 0 0 0 0 C
B C
@ 0 0 0 0 0 A
0 0 0 0 0

The coefficient D c44 is identified with the ratio between the increase in the
shear stress and the shear strain, determined by keeping constant the other strain
components. The coefficient
D c12 is the ratio between the increase in the stretch
stress and the deformation perpendicular to it, keeping constant the other strain
components. Since the eigenvalues of the matrix (16.51) should be nonnegative,
the Lamé coefficients obey the conditions

 0; 3
C 2  0: (16.52)

Similarly the compliance coefficients of an isotropic solid have the same relations
valid for the cubic system, given by (16.36), with the additional restriction 11 D
12 C 44 =2, which is equivalent to the parametrization

1  2
11 D ; 12 D  ; 44 D .1 C /; (16.53)
E E E
where E is the Young modulus and  is the Poisson ratio. The Young modulus is the
ratio between the increase in the stretch stress and the deformation along the stretch
carried out keeping constant the other stress components. That is,
 
1 @1
ED D ; (16.54)
11 @ 1 T

where the derivative is performed at constant i , i ¤ 1, and should not be confused

with c11 . The Poisson ratio is the ratio between the transverse and longitudinal
deformation when the sample is pulled longitudinally keeping constant the pressure
and other stress components, that is,
 
12 @ 2
D D ; (16.55)
11 @ 1 T

where the derivative is carried at constant i , i ¤ 1. The relations between these two
quantities and the Lamé coefficients are given by

.3
C 2 /

ED ; D ; (16.56)

C 2.
C /
328 16 Solids

which are obtained recalling that the matrices cij and ij are the inverses of each
other.
The conditions of stability (16.51) lead us to the following restriction on E and 

1
E  0; 1    : (16.57)
2
Although the Poisson ratio  can be negative there is no known substance with such
a property.
Using relation (16.41) between the compressibility and the compliance coeffi-
cients and the parametrization (16.53) we see that the compressibility is given by

3
D .1  2/: (16.58)
E

The bulk modulus B D 1= is given by

2
B D
C ; (16.59)
3
expression obtained by using (16.56).

16.3 Structural Phase Transition

Polymorphism

Solids of the same chemical composition but with different crystal structures are
called structural polymorphic forms. Iron, for example, can be found in three
polymorphic forms, known as ˛,  and ı. Phase ˛, or ferrite, has a bcc structure
is the stable phase up to 910 ı C. From this temperature up to 1400 ı C, the stable
phase is the phase  , or austenite, which has a fcc structure. Above 1400 ı C to the
melting point, which occurs at 1540 ı C, the stable phase is the phase ı, which is also
bcc. Silica (SiO2 ) can also be found in various polymorphic forms, among which we
cite quartz, tridymite and cristobalite. The diamond and graphite are polymorphic
forms of carbon. It should be noted that many polymorphic forms are metastable,
some with decay time so great that may be considered stable, as diamond. At
high pressures however diamond becomes stable and graphite metastable. Some
polymorphic forms on the other hand are always metastable, as happens to the phase
of iron-carbon known as martensite, obtained by rapid cooling of austenite.
When subjected to temperature and pressure variations, a solid can undergo
a structural transformation, passing from a polymorphic form to another. The
change in structural form is identified with a phase transition, but should be
distinguished from that occurring between liquid and vapor. These two phases are
quantitatively different, because they differ by the density, but are qualitatively
16.3 Structural Phase Transition 329

similar in structure, because the two are isotropic. This similarity allows for example
the vapor to change continuously into liquid, that is, without undergoing a phase
transition, if it is conducted by an appropriate trajectory in the phase diagram
that bypasses the liquid-vapor critical point. Two phases with different structures
on the other hand are qualitatively distinct because they differ by the symmetries
they display. The passage from one structure to another occurs always through
a phase transition, which is characterized by the occurrence of singularities in
thermodynamic properties.
The structural transitions are governed by different microscopic mechanisms.
The transition of order-disorder type as that which occurs in the copper-zinc alloy
causes a change in crystal structure. The ferroelectric transition of certain substances
such as KH2 PO4 or NaNO2 are also of order-disorder type and are accompanied by
a crystallographic change. A structural transition of another kind is that caused by
coordinated movement of atoms of a sublattice of the crystal structure and called
displacive transition. This type of transition can involve two or more sublattices
moving in different directions as happens to SrTiO3 . If a sublattice is composed
of ions, its displacement can produce an electric polarization giving rise to a
ferroelectric phase, as with BaTiO3 .
Another type of structural transition is that caused by the elastic distortion,
such as with TeO2 . This compound undergoes a continuous transition by pressure
increase from a tetragonal structure to an orthorhombic structure. The latter, con-
sidered distorted with respect to the former, is called ferroelastic. The undistorted
structure is called paraelastic. The class of compounds which undergo this same
type of structural transition includes the compound V3 Si, which undergoes a
discontinuous transition from a cubic to a tetragonal structure below 17 K, and the
compound LaNbO4 , that undergoes a continuous transition from a tetragonal to a
monoclinic structure below 495 ı C.

Ferroelastic Transition

We wish to analyze here the ferroelastic-paraelastic transition occurring in a

crystalline solid when the temperature and pressure are varied. To this end we
examine a sample at a given temperature T and subject to a hydrostatic compression
p. The only mechanical stress on the sample is the hydrostatic compression that is to
say the stretch components of the stress are equal and the shear stress components
are zero. Here it is convenient to use the Gibbs potential because these conditions
are naturally satisfied when the tensor  takes the values corresponding to minimum
of G./ subject to the constraint 1 C 2 C 3 D 0.
The order parameter of the ferroelastic transition is defined as the tensor . In
the paraelastic phase the tensor  vanishes and in the ferroelastic phase one of its
components becomes nonzero. According to Landau theory, which we adopt here,
the function g./ D G=V0 has a polynomial form and should be invariant under the
symmetry operations corresponding to the paraelastic phase. Since  has trace zero
330 16 Solids

we use the following parameterization for diagonal components:

p p
1 3 1 3
1 D  x C y; 2 D  x  y; 3 D x: (16.60)
2 2 2 2
Initially we consider a transition such that the paraelastic phase has cubic
structure and the ferroelastic phase has tetragonal structure, described by 4 D
5 D 6 D 0. In this case it suffices to consider g as a function of x and y only.
A polynomial of fourth degree with cubic invariance is given by

g D g0 C A1 .x2 C y2 / C B1 x.x2  3y2 / C C1 .x2 C y2 /2 ; (16.61)

where g0 , A1 , B1 and C1 depend only on T and p. The free energy g should be

understood as the convex hull of the expression on the right hand side of this
equation. Using the parametrization x D r cos  and y D r sin , r > 0,

g D g0 C A1 r2 C B1 r3 cos 3 C C1 r4 : (16.62)

The minimum of g related to the ferroelastic phase occurs when cos 3 D 1, for
B > 0, and cos 3 D 1, for B < 0. In both cases,

g D g0 C A1 r2  jB1 jr3 C C1 r4 ; (16.63)

where r is given by

2A1  3jB1 jr C 4C1 r2 D 0: (16.64)

The three solutions corresponding to the three possible values of  are equivalent,
that is, they have the same value of g and describe a tetragonal structural. The
absolute minimum of g occurs at r D 0 for A1 > A10 D B21 =4C1 and at a value
r ¤ 0 given by (16.64) for A1 < A10 . The transition occurs when A1 D A10 and
corresponds to a discontinuous transition, that is, the order parameter r has a jump
at A1 D A10 .
Next we consider the case in which the paraelastic phase has cubic structure and
the ferroelastic phase has trigonal structure, described by 1 D 2 D 3 that is by
x D y D 0. A polynomial of the fourth degree with cubic invariance in the variables
4 , 5 and 6 is given by

g D g0 C A2 .24 C 25 C 26 / C B2 4 5 6 C C2 .44 C 45 C 46 /: (16.65)

The minimum of g corresponding to the ferroelastic phase occurs when j4 j D

j5 j D j6 j D z with 4 5 6 < 0 if B2 > 0 and vice-versa. In both case g is

g D g0 C 3A2 z2  jB2 jz3 C 3C2 z4 ; (16.66)

16.3 Structural Phase Transition 331

and z is given by

2A2 z  jB2 jz2 C 4C2 z3 D 0: (16.67)

There are four ferroelastic solutions and they have the same value of g. The absolute
minimum of g occurs at z D 0 for A1 > A10 D B21 =36C1 and at a value z ¤ 0 given
by (16.67) for A1 < A10 . The transition occurs at A1 D A10 and corresponds to a
discontinuous transition, that is, the order parameter z has a jump at A1 D A10 .

Devonshire Theory

In the structural transitions of displacive type as those occurring in the compounds

SrTiO3 and BaTiO3 , the order parameter is not identified with the strain tensor .
In displacive transitions, the order parameter is identified with the vector qE that
defines the equilibrium position of a particular atom relative to the equilibrium
position it possessed in the structure of higher symmetry. When the displacement
qE becomes nonzero it induces the appearance of distortions in the crystalline solid,
characterized by the strain tensor . We therefore assume the existence of a coupling
between the order parameter qE and the strain tensor . If the displacement qE gives
rise to a polarization then the phase of lower symmetry is also a ferroelectric phase
with a spontaneous polarization proportional to qE.
To describe the structural phase transitions we assume that g is a function not
only of T, p and , but also of qE . The values taken by  and qE in each phase are those
that minimize g. According to Landau g must be invariant under the operations
of symmetry of the phase of highest symmetry, which we assume to be the cubic
structure. We assume that the potential g has three terms:

g D gq C ge C ga : (16.68)

The first depends only on qE and in accordance with Devonshire we suppose that gq is
a polynomial of the sixth degree in qi . Since gq should be invariant by the symmetry
operation qE $ Eq, the polynomial contains only monomials of even degree. Other
symmetry operations include the permutations among the Cartesian components q1 ,
q2 and q3 of the vector qE. The polynomial g contains only one monomial of the
second degree and two monomials of the fourth order. As to monomials of the sixth
order, we consider just one of them, although there are others. In accordance with
these considerations the thermodynamic potential gq reads

gq D g0 C b.T  T1 /.q21 C q22 C q23 /  c.q41 C q42 C q43 /C

C a.q21 q22 C q21 q23 C q22 q23 / C d.q61 C q62 C q63 /; (16.69)
332 16 Solids

where the coefficients a, b, c and d do not depend on temperature. The second term
ge depends only on the strain components i of zero trace. We assume that it is given
only by the invariants of the second order in ,

ge D A.21 C 22 C 23 / C B.24 C 25 C 26 /: (16.70)

The third term describes the coupling between the strain and the order parameter,
which we assume to be linear in  and quadratic in qE ,

ga D k.1 q21 C 2 q22 C 3 q23 /  h.4 q2 q3 C 5 q1 q3 C 6 q1 q2 /: (16.71)

Minimizing g with respect to i , under the restriction 1 C 2 C 3 D 0, we get

k h
1 D .2q2  q22  q23 /; 4 D q2 q3 ; (16.72)
6A 1 2B
and similar equations for 2 , 3 , 5 and 6 . The equations obtained by minimizing
g with respect to the variables qi together with (16.72) lead us to the following
solutions:
(a) qi D 0 and i D 0, that describes the cubic phase;
(b) q1 D q2 D 0, q3 ¤ 0, 1 D 2 ¤ 3 , 4 D 5 D 6 D 0, that describes the
tetragonal phase;
(c) q1 D 0 and q2 D q3 ¤ 0, 2 D 3 ¤ 1 4 ¤ 0, 5 D 6 D 0, that describe the
orthorhombic phase; and
(d) q1 D q2 D q3 ¤ 0, 1 D 2 D 3 , 3 D 4 D 6 , that describes the trigonal
phase.
Let us examine in more detail the transition from cubic to tetragonal phase.
To this end it suffices to take into account the description of the tetragonal phase
presented in item (b) and consider g as as function of q3 and 3 . From the expression
for gq , ge and ga , we get
3
g D g0 C b.T  T1 /q23  cq43 C dq63 C A23  k3 q23 : (16.73)
2
The minimization leads us to 3 D kq23 =3A and

b.T  T1 /q3  2cq33 C 3dq53  k3 q3 D 0; (16.74)

or

b.T  T1 /q3  2c0 q33 C 3dq53 D 0; (16.75)

where

k2
c0 D c C : (16.76)
6A
16.3 Structural Phase Transition 333

Equation (16.75) can be understood as coming from the minimization of an

effective potential given by

gef D g0 C b.T  T1 /q23  c0 q43 C dq63 : (16.77)

As seen in the previous chapter, a free energy of this type predicts a discontinuous
transition that occurs at a temperature T0 given by

.c0 /2
T0 D T1 C ; (16.78)
4bd
p
with a jump in the order parameter equal to q0 D c0 =2d.

Problems

16.1 Show that a crystalline solid belonging to a system of cubic symmetry has
only three independent stiffness coefficients and that they are given by (16.34).
16.2 Obtain the compliance coefficients ij of a solid with cubic structure from the
stiffness coefficients cii .
16.3 Show that an isotropic solid has only two independent stiffness coefficients
and that they are given by (16.51).
16.4 Show that the Young modulus and the Poisson ration are related with the Lamé
coefficients by the expression (16.56).
16.5 Use the Hooke law to determined the deformations of a cylindric isotropic
solid subjected only to stretch stress applies at its ends. Suppose next that the
lateral surface of the cylinder is fixed, that is, that there are no lateral deformations.
Determine in this case the deformation along the axis of the cylinder and the lateral
compressions. In both cases determine also the variations in the volume of the
cylinder. Present the results in terms of the Young modulus E and the Poisson ratio .
16.6 The Gibbs free energy density

g D g0 C ay2 C bx2 C cxy2 C dy4 ;

where x and y are related to the components of the traceless tensor  by (16.60),
describes the transition from the paraelastic phase with tetragonal structure to
the ferroelastic phase with orthorhombic structure. Show that the transition is
continuous an determine x and y as functions of the parameter a. Assume that the
coefficients b, c and d are positive and that the coefficient a is a linear function of T
and p, taking positive or negative values.
334 16 Solids

16.7 The Gibbs free energy density

g D g0 C a26 C b6 y C cy2 C d46 ;

where y is related to the components of the traceless tensor  by (16.60), describes

the transition from the paraelastic phase with tetragonal structure to the ferroelastic
phase with monoclinic structure. Show that the transition is continuous an determine
x and 6 as functions of the parameter a. Assume that the coefficients b, c and d are
positive and that the coefficient a is a linear function of T and p, taking positive or
negative values.
Chapter 17
Liquid Crystals

17.1 Mesophases

Thermotropics

Liquid crystals are substances which exhibit thermodynamic phases intermediate

between crystalline solid and ordinary liquid, called mesophases, which are distinct
from both ordinary liquid and crystalline solid. They have the property of fluidity,
that characterizes the liquids, and at the same time exhibit the phenomenon of
birefringence, as happens to crystalline solids. Microscopically, the liquid crystals
are distinguished by displaying only one of two types of ordering that characterize
the molecular crystalline solids. In a molecular crystal, the positions of the centers
of the molecules form a spatial ordered structure and in addition the molecules are
ordered as to the orientation. In an ordinary liquid, both positional and orientational
orders are absent. In a liquid crystal molecules have orientational order but
positional order is absent. In the opposite situation to liquid crystals, the molecules
of a plastic crystal exhibit positional order but orientational order is absent.
The shape of the molecules has a fundamental role in the formation the
orientational order. The packing of rod-shaped molecules, for example, induces the
formation of a structure with orientational order, called nematic, in which the axes
of the molecules become parallel. In fact, the axes of the molecules are not exactly
parallel but fluctuate around a given direction due to thermal agitation.
Several types of mesophases can be found between the crystalline phase and
the ordinary liquid phase, which we call isotropic phase. The simplest mesophase
is called nematic and occurs in the compound p-azoxyanisole (PAA) which is a
solid at room temperature. The transition from solid to the nematic phase occurs
at 118 ı C and from that to the isotropic phase at 135 ı C. A nematic liquid crystal
stable around the room temperature is the compound p-methoxybenzylidene-p-n-
butylaniline (MBBA). This compound passes from the solid to the nematic phase
at 21 ı C and from this phase to the isotropic liquid at 45 ı C. One of the first

© Springer-Verlag Berlin Heidelberg 2017 335

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_17
336 17 Liquid Crystals

liquid crystal studied, the cholesteryl benzoate passes from the solid phase to the
mesophase chiral nematic, known also as cholesteric phase, at 145 ı C and from this
phase to the isotropic phase at 178 ı C.
A liquid crystal can display not one but several distinct mesophases. Increasing
the temperature, the compound cholesteryl myristate passes successively through
the phases solid, smectic-A, chiral nematic and isotropic. The transitions
occur at 71 ı C, at 79 ı C and at 85 ı C, respectively. The compound called 4-n-
pentylbenzenethio-40-n-decyloxybenzoate has three mesophases between the solid
and isotropic phases, that occur in the sequence: smectic-C, smectic-A and nematic.
The transition temperatures from one phase to another are, respectively, 60, 63, 80
and 86 ı C.
The liquid crystals mentioned above are known as thermotropic because the
simplest way to induce the transitions is by varying the temperature. It is important
to note that thermotropic liquid crystals are composed by pure substances. In
another class of liquid crystals, made by mixing two or more pure substances,
the mesophases can arise not only by changing the temperature but especially by
the variation of the concentration of the components of the mixture. Those liquid
crystals are known as lyotropic.

Lyotropics

In general, the lyotropics are formed by long molecules, called amphiphilic, whose
skeleton is a carbon chain and whose ends have opposite behavior in the presence
of water molecules. One end attracts them while the other repels them. When mixed
with water, this behavior favors the development of various types of molecular
structures. The most common among them is the spherical structure called micelle
whose surface is composed by hydrophilic ends of the molecules and whose
central part consists of the hydrophobic ends. The amphiphilic compounds are
also known as surfactants. Two major groups of amphiphiles are saponaceous and
phospholipids. A typical example of the first group is the sodium laurate and of the
second group is dipalmitoylphosphatidylcholine. If the concentration of amphiphilic
water is small, a solution is formed. From a certain concentration, there is the
formation of spherical structures such as micelles or, depending on the type of
amphiphilic, of vesicles. If the concentration increase even more other structures
are formed such as lamellar structure, the cubic structure and hexagonal structure.

17.2 Nematics

Order Parameter

The orientational ordering of a sample of a liquid crystal which is in the isotropic

phase can be achieved submitting the sample to non-uniform external forces.
Denoting by Fi the Cartesian components of an external force F E that acts at a
17.2 Nematics 337

point Er with Cartesian components xi , the non-uniformity means that the derivative
ij D @Fi =@xj is nonzero. These forces are irrotational and have zero divergence
which is equivalent to saying that ij have the properties
X
ji D ij ; ii D 0; (17.1)
i

and can be understood as the components of a traceless symmetric tensor.

We assume that the sample consists of a set of n constituent units and that
the action of external forces occurs through the dipole moments and quadrupole
moment of the constituent units. A further assumption is that the external forces act
on certain points of each constituent unit and that the density of such points is given
by the function .Er/. From , we define the Cartesian components pi of the dipole
moment vector,
Z
pi D xi  d3 r; (17.2)

and the Cartesian components qij of the quadrupole moment tensor,

Z
qij D .xi xj  ıij r2 / d3 r; (17.3)

of a given constituent unit.

The energy of interaction of the external forces with the dipoles is
X
Ep D  Fi Pi ; (17.4)
i

where Pi D npi is the total dipole moment, and the interaction energy with the
quadrupoles is
X
Eq D  ij Qij ; (17.5)
ij

where Qij D nqij is the total quadrupole.

Next we assume that the dipole moments are absent. The infinitesimal work of the
external forces due to an infinitesimal variation of the moments of the quadrupole is
X
dW D  ij dQij ; (17.6)
ij

so that the infinitesimal variation of energy U of the sample is given by

X
dU D TdS C ij dQij : (17.7)
ij
338 17 Liquid Crystals

An example of external forces F E having the properties mentioned above are those
generated by an electric field. In this case pi and qij are identified as the electric
dipoles and electric quadrupoles.
The tensorial quantity Q of components Qij define the order parameter of the
nematic phase. In the absence of external forces, Qij D 0 in the isotropic phase.
In the nematic phase one of the components of Q becomes nonzero even in the
absence of external forces. From the definition of qij and taking into account that
Qij D nqij , we see that the tensor Q is symmetric and traceless. It is convenient to
use a reference system whose axes are parallel to the singular axes of the nematic
liquid crystal. In this reference system the tensor Q is diagonal and is given by
0 1
Q1 0 0
Q D @ 0 Q2 0 A ; (17.8)
0 0 Q3

where Q1 , Q2 and Q3 are the principal components and also the eigenvalues of the
tensor Q. Since the trace is invariant under any rotation of the Cartesian axes then
the trace of Q vanishes.

Q1 C Q2 C Q3 D 0: (17.9)

When the three principal components are distinct from each other the tensor Q
describes a biaxial nematic phase. When two of them are equal but distinct from the
third, the tensor Q describes a uniaxial nematic phase which can be of rod or disc
type. When all three are equal, and therefore zero, we are faced with the isotropic
phase.

Landau-de Gennes Theory

The theory of Landau-de Gennes describes the transition between the nematic phase
and the isotropic phase. According to this theory, the molar Gibbs free energy must
be invariable by the symmetry operation relating to the isotropic phase. This means
that g.Q1 ; Q2 ; Q3 / must be invariant by any permutation of variables Qi . The theory
of Landau-de Gennes assumes that the free energy is a linear combination of the
invariants In of order n. An invariant of order n contains only powers of Qi of order n.
The invariant of first order is identified with the trace of Q and therefore vanishes
identically. The invariant of second and third order are given by

I2 D Tr Q2 ; I3 D Tr Q3 : (17.10)
17.2 Nematics 339

The other invariants are I4 D I22 , I5 D I2 I3 , I6 D I32 and I60 D I23 . Notice that there
are two invariants of sixth order. Up to sixth order, the free energy g reads

g D g0 C a2 I2 C a3 I3 C a4 I22 C a5 I2 I3 C a6 I32 C a06 I23 : (17.11)

Notice that the free energy should be understood as the convex hull of the expression
in the right hand side of (17.11). Taking into account that the thermodynamic field
conjugate to the variables Qi should vanish, then the thermodynamic states are
identified as the minima of this expression.
Initially we examine the transition from the isotropic phase to the uniaxial
nematic phase which is described by the tensor Q such that two components are
equal. Assuming that the two components are Q1 and Q2 we can use the following
parameterization Q1 D Q2 D q=2 and Q3 D q. In this case

3 2 3 3
I2 D q ; I3 D q ; (17.12)
2 4
and using invariants up to fourth order, the function g becomes

3 3 9
g D g0 C a2 q2 C a3 q3 C a4 q4 : (17.13)
2 4 4
We assume that a4 is positive and that a2 is linear with temperature and pressure
and takes positive or negative values. The coefficient a3 is considered to be negative
but the solution for the positive case is obtained from the case a3 < 0 by the
transformation of q into q.
For values of the parameter a2 large enough the absolute minimum of g occurs at
q D 0 which determines the isotropic phase. Decreasing a2 , we obtain the nematic
phase for which q ¤ 0. The value of q for the nematic phase corresponds to the
nonzero root of @g=@q D 0, that is,

3
a2 q C a3 q2 C 3a4 q3 D 0: (17.14)
4
From the two nonzero solutions the one corresponding to the lower free energy is
given by
0 s 1
1 @ ja3 j a23 a2 a4 A
qD C  : (17.15)
a4 8 64 3

The transition from the nematic phase to the isotropic phase is discontinuous
due to the presence of a third order invariant in the free energy g. To determine the
transition temperature which corresponds to the coexistence of phases isotropic and
340 17 Liquid Crystals

nematic we impose the following form for the free energy

9
g D g0 C a4 q2 .q  q0 /2 ; (17.16)
4
which describes the coexistence between the isotropic phase, q D 0, and the nematic
phase, q D q0 . Comparing with (17.13) we see that
s
a2 2a2
a2 D 3 ; q0 D : (17.17)
24a4 3a4

Thus we can draw the following conclusions. For a2 > a0 D a23 =24a4, the order
parameter vanishes, and the phase is isotropic. For a2 < a0 the phase is nematic
with q given by (17.15) with a jump equal to q0 at the transition point a2 D a0 .

Landau Point

Next we will analyze the possibility of the occurrence of a biaxial nematic phase
characterized by distinct values of Q1 , Q2 and Q3 . We assume a free energy of the
form (17.11) with a5 D 0 and a06 D 0,

g D g0 C a2 I2 C a3 I3 C a4 I22 C a6 I32 ; (17.18)

and such that the coefficients a2 and a3 have linear dependence with temperature and
pressure taking positive or negative values and that a4 and a6 are strictly positive.
Taking into account that Q is a traceless tensor, it is convenient to use the
following parametrization

1 1
Q1 D  .q  /; Q2 D  .q C /; Q3 D q; (17.19)
2 2
from which we obtain the following results

1 2 3
I2 D .3q C 2 /; I3 D q.q2  2 /: (17.20)
2 4

The minima of g are obtained from @g=@q D 0 and @g=@ D 0.

To determined the transition from the isotropic to a uniaxial we impose the
condition  D 0. In this case I2 D 3q2 =2 and I3 D 3q3 =4 which replaced in (17.18)
results in the following form for the free energy

3 3 9 9
g D g0 C a2 q2 C a3 q3 C a4 q4 C a6 q6 : (17.21)
2 4 4 16
17.2 Nematics 341

The nematic-isotropic phase transition, which is first order, is determined by

comparing (17.21) with a sixth degree polynomial that has two minima with the
same value of g, one at a D 0 (isotropic phase) and the other at q D q0 (uniaxial
nematic phase), which describes the coexistence of the two phases. This comparison
gives the equation for the nematic-isotropic transition line and the jump in the order
parameter along the coexistence line,
s
2a2 2a2
a2 D 3 ; q0 D ; (17.22)
3a6 3a4

valid for small values of a2 . The coexistence line is shown in Fig. 17.1a. Notice that
q0 vanishes at the point a3 D 0 and a2 D 0.
As long as a3 is nonzero, the isotropic phase is contiguous to a uniaxial nematic
phase as seen in Fig. 17.1a and the transition is discontinuous. When a3 D 0,
the transition ceases to be discontinuous, and becomes a continuous transition that
occurs at a2 D 0. The point a2 D 0, a3 D 0 is a special point of the phase diagram
and is called Landau point.
When  becomes nonzero there is the emergence of a biaxial phase. The minima
of g such that  ¤ 0 and q ¤ 0 are determined by
a2 a3
I2 D  ; I3 D  ; (17.23)
2a4 2a6

which is obtained by the minimization of the free energy (17.18). Recall that I2
and I3 are related to q and  by (17.20) so that (17.23) gives implicitly q and  as
functions of a2 and a3 . The transition line from the biaxial nematic to the uniaxial
nematic phases is obtained by taking the limit  ! 0 because this transition is

(a) a2 (b) a2

I I

a3 a3
N+ N− N+ N−

Fig. 17.1 Phase diagram for the transition between the isotropic (I), uniaxial nematic (NC and
N ) and biaxial (B) phases according to the Landau-de Gennes theory. The dotted and continuous
lines describe the discontinuous and continuous transition, respectively. (a) Diagram corresponding
to a5 D 0. (b) Diagram corresponding to a value of a5 such that a25 > 4a4 a6
342 17 Liquid Crystals

continuous, what can be understood noticing that g, given by (17.18), contains only
even powers of the order parameter  of the biaxial phase and a4 > 0. Taking the
limit  ! 0 in (17.23) and eliminating q, we get

a2 D cja3 j2=3 ; (17.24)

where c D .12a34 =a26 /1=3 , that describes the critical line corresponding to the biaxial
uniaxial transition as shown in Fig. 17.1a.
If the coefficient a5 is nonzero and such that a25 < 4a4 a6 , the phase diagram has
the topology shown in Fig. 17.1a with the critical lines approaching the Landau point
with a certain slope. If a25 > 4a4 a6 the biaxial phase no longer exists. In this case
the phase diagram exhibits only the two uniaxial phase in addition to the isotropic
phase. The two uniaxial phases become contiguous and the transition between them
is discontinuous. The line of coexistence between the two nematic phase meet the
nematic-isotropic coexistence line at the point a3 D a2 D 0 as seen in Fig. 17.1b.
Notice that in both cases the order parameters q and  vanish continuously at the
Landau point.

Coupling to the Electric and Magnetic Fields

Let us examine here the response of a uniaxial nematic liquid crystal to a static
electric field. We will restrict ourselves to nematics which behave as an ordinary
dielectric, that is, nematics such that the dipole moment of the sample is proportional
to the applied field. In an isotropic liquid, the constant of proportionality, which is
the electric susceptibility at zero field, is independent of direction. In a nematic it
depends on the direction of the applied field. In other terms, the susceptibility has an
anisotropy. If the electric field E is applied along the direction corresponding to the
axis of the rod-shaped molecules then the electric dipole moment per unit volume
P is

P D k E; (17.25)

where k is the parallel susceptibility. If the field is perpendicular then

P D ? E; (17.26)

where ? is the perpendicular susceptibility. Denoting by n the unit vector parallel

to the direction of the axis of the rod-shaped molecules then both formulas above
can be written as

P D ? E C .n  E/n; (17.27)

17.2 Nematics 343

where the difference

 D k  ? (17.28)

is a measure of the anisotropy, which can be positive or negative.

The electric contribution gele to the free energy per unit volume will be

? 2 
gele D  E  .n  E/2 ; (17.29)
2 2

which is obtained by integration of the relation P D @gele =@E. The first term is
independent of orientation. The second term instead depends on the orientation of
nematic liquid crystal sample. If  > 0 then it favors a parallel alignment with the
applied field. If  < 0, it favors an alignment perpendicular to the applied field. In
any case, the electric field becomes a method for the alignment of a nematic sample
in a given direction.
We may also consider the response to a magnetic field B. Similarly the
magnetization of a sample subject to a magnetic field is given by

M D m
? B C  .n  B/n;
m
(17.30)

where

m D m
k  ?
m
(17.31)

is the difference between m k

and m? , the magnetic susceptibilities parallel and
perpendicular to the director n, respectively.
The magnetic contribution gmag to the free energy per unit volume is

m
? 2 m
gmag D  B  .n  B/2 ; (17.32)
2 2

which is obtained by integration of the relation M D @gmag =@B.

The order parameter of the nematic phase can be experimentally obtained by
means of experimental measurements of the anisotropic part  of the electric
susceptibility or the anisotropic part m of the magnetic susceptibility. Among
other forms of experimental measurement of the order parameter we mention the
measurement of birefringence,

n D nk  n? ; (17.33)

the difference between nk and n? , the index of refraction parallel and perpendicular
to the director, respectively.
344 17 Liquid Crystals

17.3 Smectics

Order Parameter

In addition to the orientational order, smectic liquid crystals have positional order
along one direction. We imagine the smectic liquid crystals as succession of layers
inside which the molecules are in the nematic state described by the director n. The
direction perpendicular to the layers defines the axis z. If the director n is parallel to
the axis z, the liquid crystal is called smectic -A. If the director n form an angle with
the axis z, then the liquid crystal is called smectic -C.
Next let us examine the phase transitions occurring in a smectic -A liquid crystal.
To this end we need to define the order parameter that describes this phase. Denoting
the local density of the liquid crystal by .z/, we see that .z/ is a periodic function
of z with a period equal to L, the distance between two successive layers. Taking
into account only the first terms of the expansion of .z/ in Fourier series, we write

2
 D 0 C  cos. z C '/; (17.34)
L
where  gives the measure of change in density when one passes from one to the
next layer and, therefore, vanishes when the liquid crystal undergoes a transition to
the isotropic phase or to the nematic phase.
It is important to note that two states that differ only by a translation along the
z direction are described by same value of  but different values of phase '. Thus,
a smectic -A state is described by the variables  and ' which can be combined to
define the complex variable

D ei' ; (17.35)

which determined the order parameter of the smetic -A phase. It vanishes both in
the isotropic and nematic phases. Equation (17.34) can therefore be written as

 D 0 C Ref ei2z=L g: (17.36)

Phase Transition

To describe the phase transitions in a liquid crystal that can be found in isotropic,
nematic and smectic -A, we adopt the following free energy

g D g 0 C a 2 q 2 C a 3 q 3 C a 4 q 4 C b 2  2 C b 4  4  c4 q 2  2 ; (17.37)
17.3 Smectics 345

where the coefficients depend on the temperature and possibly on other thermody-
namic fields. We consider that the coefficients a4 , b4 are strictly positive and that
c24 < 4a4 b4 to ensure the convexity of g with respect to the set of variables q and
. The possible thermodynamic phases correspond to the various types of absolute
minima of g, obtained when the coefficients are varied. The isotropic phase occurs
when q D 0 and  D 0, the nematic phase, when q ¤ 0 and  D 0, and the
smectic -A phase, when q ¤ 0 and  ¤ 0. To exclude minima corresponding to
q D 0 and  ¤ 0, we assume b2 > 0.
Let us start by determining the transition between the nematic and isotropic
phases. In this case,  D 0 in both phases so that the free energy (17.37) is
reduced to

g D g0 C a2 q2 C a3 q3 C a4 q4 : (17.38)

The order parameter of the nematic phase q is given by the equation

2a2 C 3a3 q C 4a4 q2 D 0; (17.39)

obtained from @g=@q D 0. The free energy (17.38) is similar to that given by
(17.13), which describes the nematic isotropic transition. Therefore, using the same
reasoning used in that case, we conclude that the transition is first order and occurs
when

a23
a2 D ; (17.40)
4a4

shown as a horizontal line in the phase diagram of Fig. 17.2.

Fig. 17.2 Phase diagram of a

liquid crystal which can be
found in the isotropic (I),
a2 I
nematic (N) and smectic -A
phases, according to the free
energy given by (17.37). The
dotted and continuous lines
describe the discontinuous
and continuous transitions, N c
4
respectively. The filled circle
represents a tricritic point and
the empty circle represents a A
triple point
346 17 Liquid Crystals

To find the other transitions, we determine the equations that give the minima
of g. They are

2a2 q C 3a3 q2 C 4a4 q3  2c4 q 2 D 0; (17.41)

3 2
2b2  C 4b4   2c4 q  D 0: (17.42)

The nematic phase is described by  D 0 and q given by the root of (17.39).

The smectic -A phase corresponds to the nonzero solutions, q ¤ 0 and  ¤ 0,
determined by equations

.2a2 b4 C b2 c4 / C 3a3 b4 q C .4a4 b4  c24 /q2 D 0; (17.43)

1
q2 D .b2 C 2b4  2 /: (17.44)
c4

From the solutions of the equations that determine q and  for each phase, we
determine the respective free energies. When the transition between two phases
is discontinuous, it is determined by comparing the free energies. From the
solutions of (17.43) and (17.44) and using a numerical procedure to determine
the lowest free energy, we construct the phase diagram in the variables a2 and
c4 , shown in Fig. 17.2. The diagram shows lines of coexistence between nematic
and smectic -A phases, nematic and isotropic and smectic -A and isotropic. The
three coexistence lines meet at a triple point. The diagram also gives a criti-
cal line between the smectic -A and the nematic phases that ends at a tricritic
point.
To determine the critical line and the tricritical point we replace (17.44)
into (17.43) and expand the result in powers of . The critical line occurs when
 ! 0, that is, when the constant term vanishes, which gives us the result
s
b2 b2
2a2 C 3a3 C 4a4 D 0; (17.45)
c4 c4

valid when c4 > 0. The tricritical point occurs when the constant term and the
coefficient of  2 vanish what occurs when
s
b2
3a3 b4 C 2.4a4b4  c24 / D 0: (17.46)
c4
17.3 Smectics 347

Problems

17.1 From the free energy g given by (17.13), which describes the nematic isotropic
transition, determined the density  D 1=v, v D @g=@p, as a function of temperature
and shown that  has a jump at the transition point. To this end assume that the
coefficient a2 is linear with pressure and temperature and that g0 D RT ln pCbpCa0 ,
where b and a0 are constant.
17.2 A liquid crystalline elastomer is a substance that combines the properties of
anisotropy of liquid crystals and the elastic properties of materials such as rubber.
Its thermodynamic properties can be described by the free energy

1
g D g0 C a2 q2 C a3 q3 C a4 q4  bq C K 2
2

where q is the order parameter of the nematic phase and is the strain along a certain
direction, a2 , a3 , a4 and b are parameters, and K is the Young’s modulus. The stress
 is given by  D @g=@ . Determine the stress  as a function of the strain and
shown that it has a horizontal segment.
Chapter 18
Thermal Radiation

18.1 Electromagnetic Waves

Radiation

When we are outdoors on a sunny day we feel the effects of heat on our skin even
if the environment is at a low temperature. Likewise, when we are near the fire, we
feel more heat on the surface of our body that is exposed to fire than on the opposite
surface. These effects are caused by radiative heat transfer from the emitting source,
the sun or fire, to your body. This radiative transfer is distinguished from other
forms of heat transfer because it does not require a material medium to propagate.
In material media heat transfer occurs by thermal conduction or convection. The
thermal conduction is carried out in the solids by vibration of the crystal lattice and
in electrical conductors also by free electrons. In liquids and gases it occurs mainly
by the collision between molecules. The heat transfer by convection is that occurring
by the macroscopic motion of matter.
The radiative transfer does not require the presence of a material medium and is
the main form of heat transfer when the bodies are separated. It occurs by means of
electromagnetic waves that carry energy and can be absorbed and emitted by matter.
The photons, that are the elementary particles resulting from the quantization of the
electromagnetic field, are continually emitted and absorbed by bodies by means of
several processes: transitions between electronic, vibrational and rotational levels in
atoms and molecules, ionization or combination of free electron with ions, change
of the kinetic energy of free electrons, excitations of the crystal lattice in solids.
These processes give rise to a continuous spectrum of emission, or absorption,
superimposed by spectral lines resulting from transitions between discrete energy
levels.
The radiative transfer is the mechanism by which two bodies, physically isolated
from each other, come into equilibrium. If they are at different temperatures, the
hotter emits more radiation than it absorbs and the colder absorbs more than it

© Springer-Verlag Berlin Heidelberg 2017 349

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_18
350 18 Thermal Radiation

emits. In thermal equilibrium, each of the bodies absorbs and emits equal amounts of
radiation. A bar of iron heated to redness cools when removed from the heat source
and reaches room temperature mainly by radiative energy loss to the environment.
The same occurs with a filament of an electric lamp, which is practically materially
isolated from the environment due to the vacuum inside the lamp, when one
interrupts the electric power supply.
Electromagnetic waves propagate in vacuum at the speed of light, c D 2:9979 
108 m/s. The monochromatic waves are characterized by the frequency , or
equivalently by the wavelength
since these two quantities are related by c D
.
Electromagnetic waves have distinct names according to its wavelength. Visible
light corresponds to a very small range of wavelength between 0:38  106 and
0:76  106 m. Wavelengths above the visible up to 103 m correspond to the
infrared radiation. Above the infrared are radio waves. Below the visible down
to 109 m, correspond to the ultraviolet radiation. The X-rays have wavelengths
between 108 and 1012 m and the gamma rays, below 1010 m. Thermal radiation
comprises the ultraviolet and infrared radiation in addition to the visible light.
Electromagnetic waves carry not only energy but also momentum. There is a
close relationship between the momentum p of an electromagnetic wave and its
energy E ,
E
pD ; (18.1)
c
a relation due to Maxwell. The same relation is valid for the momentum p of a
photon and its energy E . Considering that waves have momentum it is natural to
imagine that they exert pressure on bodies. Consider a cubic cavity of edge L and the
total energy U of the electromagnetic radiation within the cavity. As electromagnetic
waves move in three independent spatial directions we can assume that 1=3 of them
move in a particular direction, say along the x direction, parallel to one edge of the
cubic cavity. The momentum transferred to one of the walls perpendicular to the x
direction during a time interval 2L=c is .1=3/2U=c. The force on the wall is the ratio
between these two quantities and thus equal to U=3L. Therefore, the pressure p on
the wall, which is the ratio between the force and the area of the face of the cubic
cavity, is p D .U=3L/=L2 D U=3V, where V D L3 is the cavity volume, that is,
u
pD ; (18.2)
3
which is the relation between the radiation pressure p and the energy per unit volume
u D U=V of the radiation, relation due to Maxwell.

Kirchhoff Law

To characterize the spectrum of electromagnetic radiation of a body at a given

temperature, we measure the energy emitted in each one of the frequencies of
18.1 Electromagnetic Waves 351

electromagnetic wave. More precisely we measure the energy per unit time and
per unit area of the emitter  emitted by the emitter within the frequency range
between  and  C . The monochromatic emissive power , as a function of
, is different from one substance to another. However, there is a hull that can be
considered as the emissive power common to all bodies and therefore independent of
the emitter substance, which we call universal spectrum or black body spectrum. The
idea of black body was introduced by Kirchhoff who established the universality of
the radiation spectrum. According to Kirchhoff, the power spectral density of a black
body is independent of the particular characteristics of the emitter, depending only
on the frequency of the radiation and the temperature.
Certain bodies can be seen approximately as black bodies as the sun or as the
tungsten filament of a lit electric bulb. The black body radiation can be obtained by a
cavity whose walls are at a certain temperature T. The emitted radiation is observed
through a small hole made in the cavity. Considering that the photons almost do not
interact, the thermal equilibrium of radiation contained in a cavity occurs through
the mechanism of emission and absorption of photons by the walls.

Stefan-Boltzmann Law

The radiative heat flux emitted by a surface is characterized by emissive power ˚

defined as the radiative energy emitted by a surface by unit time and per unit area. It
is convenient to define the emissive spectral power ./ with respect to a frequency
. This quantity is defined such that the product ./ is the radiative energy
emitted by a surface per unit time and per unit area, by the radiation with frequencies
between  and  C . The two quantities are related by
Z 1
˚D ./d: (18.3)
0

For the analysis of radiation within a cavity, it is convenient to introduce the

spectral density of energy ./ defined such that du D  is the energy per unit
volume of the radiation with frequencies between  and  C d. The density of
energy, that is, the energy per unit volume u of electromagnetic radiation is related
to  by means of
Z 1
uD ./d: (18.4)
0

Since the radiation that emerges from the cavity through the hole has velocity c,
then the emissive powers and the energy densities are related by  D c=4 and
˚ D uc=4. The factor 1=4 must be introduced by the following reason. Consider a
point within the cavity situated near the cavity opening and define the direction z as
one that arises at this point and goes out of the cavity through the hole. Since the
352 18 Thermal Radiation

radiation in the cavity is isotropic, then the fraction of the radiation passing through
the vicinity of that point and emerging through the hole should be proportional
to the component of the propagation direction kO of the electromagnetic wave in
the direction z > 0. Using spherical coordinates and taking into account that the
component of kO in the z direction is cos , this fraction is determined by
R 2 R =2
0 0 .cos / sin dd  1
R 2 R  D D : (18.5)
4 4
0 0 sin dd

The Kirchhoff universal law of radiation states that the spectral emissive power of
a blackbody depends only on the temperature T and on the frequency . Therefore,
the power spectral density .T; / only depends on T and  and the energy density
u.T/ depends only on T. As the radiation pressure p is related to u by p D u=3 then
the pressure radiation p.T/ also depends only T. Stefan has shown experimentally
that the emissive power ˚ is proportional to the fourth power of the absolute
temperature,

˚ D T 4 : (18.6)

The constant of proportionality constant is called Stefan-Boltzmann constant and is

 D 5:6697  108 W m2 K4 : (18.7)

Relation (18.6), called Stefan-Boltzmann law, was demonstrated by Boltzmann

from the expression obtained by Maxwell for the radiation pressure, p D u=3. This
demonstration will be presented below.
Using relations u D 4˚=c and p D u=3 we see that

4 4
uD T ; (18.8)
c
4 4
pD T : (18.9)
3c
Thus, the density of energy and the pressure are also proportional to the fourth power
of the temperature.

Free Energy

Consider a cavity of volume V whose walls are at a temperature T. Thermody-

namically, radiation within the cavity is characterized by only two independent
thermodynamic quantities and not by three as in a system with one chemical
component. In this sense, we can say that the radiation contained in a cavity
18.1 Electromagnetic Waves 353

is actually a thermodynamic system with zero component. We can compare the

radiation in a cavity with the saturated vapor of a one component system for which
the vapor pressure only depends on the temperature. Similarly to what happens with
the liquid-vapor coexistence in which a volume variation does not cause change in
the pressure if the temperature is kept constant, so does with radiation. By varying
the volume of the cavity, a certain amount of radiation will be absorbed or emitted
by the walls but the pressure remains unchanged.
In the representation of energy, U.S; V/ depends only on entropy S and volume
V. The differential form of energy is given by

dU D TdS  pdV: (18.10)

The extensivity of the energy allows us to reduce the fundamental relation into the
form U.S; V/ D Vu.s/ where u D U=V is the energy density and s D S=V is
the entropy density and the system can be described simply by u.s/. The relation
T D du=ds is also valid. The Euler equation,

U D TS  pV; (18.11)

implies u D Ts  p.
Using the Helmholtz free energy representation F D U  TS we see that F.T; V/
depends only on T and V. In addition,

dF D SdT  pdV: (18.12)

The extensivity allows us to write F.T; V/ D Vf .T/ where f D F=V is the free
energy density. Notice that f .T/ depends only on the temperature. From the Euler
equation

F D pV; (18.13)

and therefore p D f from which follows that the pressure depends only on
temperature. In addition, f D u  Ts and s D df =dT and the entropy density
depends only on temperature.
We have seen that the radiation pressure is related with energy density by p D
u=3. Since f D p then f D u=3. Replacing this last result in f D u  Ts, we
obtain the result 4f D Ts. Deriving both sides of this equation with respect to
temperature and taking into account that s D df =dT, we obtain the following
differential equation for the entropy

ds
3s D T : (18.14)
dT
Integrating this equation, we get the entropy as a function of temperature

s D AT 3 ; (18.15)
354 18 Thermal Radiation

result obtained assuming that the entropy vanishes at the absolute zero of
temperature.
Since s D df =dT and p D f then s D dp=dT. Integrating this equation we get
the radiation pressure

1 4
pD AT ; (18.16)
4
where we assumed that the pressure vanishes at T D 0. The energy density is u D 3p
so that
3 4
uD AT ; (18.17)
4
a result consistent with the Stefan-Boltzmann law. Comparing with (18.9), we see
that the constant A is related to the Stefan-Boltzmann constant by

16
AD : (18.18)
3c
When we vary the volume of a cavity at a constant temperature the pressure
remains invariant. However, a certain amount of radiation will be absorbed or
emitted by the walls. Consequently, a certain amount of heat will be exchanged
with the cavity walls. To determine the amount of heat, we calculate the enthalpy
H D U CpV because the variation is enthalpy equals the heat exchanged at constant
pressure. Using relation p D U=3V, we get H D 4pV and therefore the amount of
heat exchanged Q is equal to Q D H D 4pV. The work done W by the radiation
in this process is equal to the decrease in the free energy because the temperature
remains invariant. Since F D pV, then W D F D pV. The variation of
energy is U D 3pV.

18.2 Planck Law

Wien Displacement Law

According to Kirchhoff the spectral density of energy .; T/ of a black body

depends only on the frequency  of the radiation and temperature. Wien showed
that the spectral density must satisfy the following scaling law

.; T/ D  3 F . /; (18.19)
T
known as Wien displacement law. The Stefan-Boltzmann law is a direct conse-
quence of this law. Indeed, integrating the spectral density of energy over all
18.2 Planck Law 355

frequencies, we get the total density of energy

Z 1 Z 1

u.T/ D .; T/d D  3 F . /d: (18.20)
0 0 T

Doing the following change of integration variable, x D =T, we get the result
Z 1
4
u.T/ D T x3 F .x/dx: (18.21)
0

But the integral does not depend on T and we may conclude that u.T/ is proportional
to the fourth power of temperature.
The spectral density of energy .; T/ has a maximum at a certain value of  that
we denote by max . To get max , we derive .; T/ with respect to  to obtain the
following equation

 3 
3 2 F . / C F 0 . / D 0; (18.22)
T T T
or yet

3F .x/ C xF 0 .x/ D 0; (18.23)

where x D =T. Suppose that x0 is the solution of the equation above. We conclude
that

max D x0 T; (18.24)

that is, raising the temperature of a black body the frequency value for which the
spectral density of energy is maximum moves to the region of high frequencies,
known as Wien displacement law.

Planck Radiation Law

The radiation in a cavity can be viewed as a collection of harmonic oscillators

vibrating with different frequencies. Each oscillator corresponds to a normal mode
of vibration of the electromagnetic waves. Considering a cubic cavity of volume
V D L3 , the normal modes correspond to waves propagating parallel to the sides of
the cube. The wavelength
must be a divisor of the length L of the edge of the cube,
that is,
D L=n, where n is an integer. Since
and the frequency  are related by
 D c=
it follows that the normal modes have frequencies n D nc=L. Next we
determine how many vibration modes exist or equivalently how many oscillators
exist with frequencies between  and  C . Considering just one direction, this
356 18 Thermal Radiation

number equals 2.L=c/, where the factor 2 is due to the two independent modes
of propagation of the electromagnetic waves, which are transversal. Considering the
three directions this number becomes 8.L=c/3  2  D 8.V=c3 / 2 .
The spectral density of energy .; T/ is therefore

8 2
D E; (18.25)
c3

where E.; T/ is the mean energy of an oscillator. To determined E.; T/ we proceed

as Planck.
The formula introduced by Wien for the spectral density is

 D A 3 ea=T ; (18.26)

valid for high frequencies, where A and a are two constants. Comparing
with (18.25), we see that it corresponds to the following expression for E,

E D bea=T ; (18.27)

where b is another constant, related to A by A D 8b=c3 . Writing 1=T as a function

of E and using the relation 1=T D @S=@E we conclude that

@S 1 E
D  ln ; (18.28)
@E a b
from which we get

@2 S 1
2
D : (18.29)
@E aE
In the regime of low frequencies (18.27) ceases to be valid and should be replaced
by another expression. Following Planck we assume, in this regime, the following
expression for the energy of an oscillator

E D kT; (18.30)

from which we obtain @S=@E D 1=T D k=E and therefore

@2 S k
2
D  2: (18.31)
@E E
Proceeding according to Planck, we assume the following interpolation formula
between (18.29) and (18.31),

@2 S k
D ; (18.32)
@E2 E.h C E/
18.2 Planck Law 357

Fig. 18.1 Spectral density of 12

energy for the background
cosmic radiation. The graph

Wm )
shows  D c2 =4 as a

2
function of   D =c. The 8
circles represent

-16
observational data at the

ρ* (10
microwave band, collected by
Fixsen et al. [175]. The 4
continuous line represent the
function

 D 2hc2  3 =.ehc =kT  1/
0
corresponding to 0 5 10 15 20 25
T D 2:725 K -1
ν* (cm )

which we consider to be valid for any frequency. For large values of E, it reduces
to (18.31). For small values of E, it reduces to (18.29), if a D h=k. Integrating, we
get

1 @S k h C E
D D ln ; (18.33)
T @E h E
or, isolating E,

h
ED ; (18.34)
eh=kT  1
from which we get

8h 3 1
D ; (18.35)
c3 eh=kT  1
which is the spectral density of energy due to Planck for the black body radiation.
This formula contains two universal constants. The Boltzmann constant k and the
Planck constant h.
We remark that (18.30) should be also considered as the mean energy of and
oscillator at high temperature what justifies the identification of the constant k as the
Boltzmann constant. As is well known the average energy of a harmonic oscillator
is equal to kT only in the classical regime. This result ceases to be valid at low
temperatures and must be replaced by the result of Planck, given by (18.34) valid in
the quantum context.
The background cosmic radiation, discovered by Penzias and Wilson, occurs
at the microwave band and has a black body spectrum corresponding to the
temperature of 2.725 K, as can be seen in Fig. 18.1.
Chapter 19
Thermochemistry

19.1 Extent of Reaction and Affinity

Chemical Reactions

A chemical reaction is a process in which one or more substances, the reactants,

are transformed into another or other substances, the products. A flame is a
striking manifestation of combustion, which consists of the intense combination
of a substance with oxygen. The reaction with oxygen can be mild such as rust,
spontaneously occurring in iron exposed to atmospheric air. The decomposition of
a chemical compound into chemical elements, as in the electrolysis of water, or
the formation of a chemical compound from its constituents, as in the synthesis
of ammonia from nitrogen and hydrogen, are examples of chemical reactions
performed in the laboratory. A vessel with several chemical species inside of which
chemical reactions take place, is a thermochemical system. Chemical reactions can
be endothermic when heat is absorbed by the thermochemical system, or exothermic
when heat is released.
Chemical reactions are represented by chemical equations constituted by the
chemical formulas of the reactants and products and the respective stoichiometric
coefficients. The formation of water from hydrogen and oxygen is represented by
the chemical equation

2H2 C O2 ! 2H2 O; (19.1)

which means that two moles of hydrogen react with one mole of oxygen to
produce two moles of water. A chemical equation shows the quantitative relations
between reactants and products but do not necessarily represent the molecular
process among the reactants or the reaction mechanism. At the molecular level
the process may include intermediate reactions and involve other chemical species.
The combustion of hydrogen, represented by (19.1) is actually composed by chain

© Springer-Verlag Berlin Heidelberg 2017 359

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2_19
360 19 Thermochemistry

reactions comprising other chemical species in addition to H2 and O2 . Though

relevant to the development of the reaction, the quantities of other chemical species
remain invariant and are not represented in the chemical equation, that shows only
the species whose quantities change, decreasing for the reactants and increasing for
products.
The reversed reaction of that given by (19.1) corresponds to the decomposition
of water, represented by

2H2 O ! 2H2 C O2 ; (19.2)

which occurs by the process of electrolysis, when an electric current passes through
a dilute salt solution. Electrolysis, briefly represented by chemical equation (19.2),
consists of several intermediate reactions including those which occur at the
cathode, which produces hydrogen, and at the anode, where oxygen is produced.
The occurrence of the equilibrium of a reaction requires that its reverse also
occurs as is the case, for example, of the formation of ammonia from nitrogen and
hydrogen and its reverse, the decomposition of ammonia, both represented by

N2 C 3H2 • 2NH3 : (19.3)

In this equation, conventionally, the chemical species on the left are called reactants
and those on the right are known as products. The competition between the reactions
leads the system to equilibrium. If the reaction rate to the right is greater than the rate
of reaction to the left, there is an increase of the products which causes an increase
in the rate of reaction to the left approaching it to the rate of reaction to the right.
Eventually, the rates become equal and the chemical equilibrium is established. In
equilibrium, the quantities of products and reactants remain invariant. Later we
will analyze in detail the conditions of thermochemical equilibrium and will see
that in this regime the concentrations are interrelated in a manner consistent with
the chemical reactions. We will also examine the behavior of the thermochemical
equilibrium with temperature and pressure.

Thermodynamic State

To understand how one defines the thermodynamic state of a thermochemical

system in equilibrium, we examine a specific example. A vessel contains a gas
mixture of nitrogen, hydrogen and ammonia, within which the chemical reaction
represented by (19.3) takes place. The state of thermodynamic equilibrium in the
representation of energy U is defined by the entropy S, the volume V and the number
of moles N1 of nitrogen, N2 of hydrogen and N3 of ammonia. In differential form
the internal energy U is given by

dU D TdS  pdV C 1 dN1 C 2 dN2 C 3 dN3 ; (19.4)

19.1 Extent of Reaction and Affinity 361

where the temperature T, the pressure p and the chemical potentials i of the three
species are given by
@U @U @U
TD ; pD ; i D : (19.5)
@S @V @Ni
This description, however, is only suitable in the absence of reaction, in which case
N1 , N2 and N3 may in fact be considered independent variables. In a closed system
and in the absence of reaction, they remain invariant. If the system is closed but there
are chemical reactions, they can vary. The variations, however, can not be arbitrary
but must be in accordance to the rules imposed by the chemical reaction. In any
event, it is possible to find combinations of N1 , N2 and N3 that, in a closed system,
can be invariant and therefore be considered independent. Since there is only one
reaction and three chemical species, we should find two independent variables, as
shown below.
Equation (19.3) means that one mole of nitrogen reacts with three moles of
hydrogen to produce two moles of ammonia. If, due to chemical reaction, the
amount of nitrogen is decreased by one mole then the amount of hydrogen will
decrease by three moles which means that

X2 D N2  3N1 (19.6)

remains invariant under the chemical reaction. The amount of ammonia, in turn, will
increase by two moles and

X3 D N3 C 2N1 (19.7)

will also remain invariant. Next we use the two variables X2 and X3 to describe the
thermodynamic state in the place of N1 , N2 and N3 . However, they are not sufficient
and must be supplemented by one more variable, which we choose to be D N1 ,
called the extent of reaction. With this expedient, the relation between the number
of moles and the new variables becomes

N1 D  ; N2 D X2  3 ; N3 D X3 C 2 ; (19.8)

which replaced into the differential form of energy (19.4) gives

dU D TdS  pdV  Ad C 2 dX2 C 3 dX3 ; (19.9)

where

A D 1 C 3 2  2 3 ; (19.10)

is called affinity, a quantity introduced by De Donder and defined by

@U
AD ; (19.11)
@

keeping S, V, X2 and X3 constant.

362 19 Thermochemistry

In this new description, the variables X2 and X3 do not have their values changed
by the chemical reaction. In other words, they can be regarded as independent
variables. This property however is not shared by the variable that, in equilibrium,
has its value dictated by the chemical reaction. In other terms, it should be
understood as a dependent variable. The system should therefore be understood as
a system of two components and not three and the state is now defined only by
the variables X2 and X3 , in addition to S and V. This means that, in equilibrium, U
should be independent of so that the derivative .@U=@ /S;V;X2 ;X3 D A must vanish,
which leads us to the result

1 C 3 2  2 3 D 0; (19.12)

which is the condition for thermochemical equilibrium for the reaction represented
by (19.3). At equilibrium, the chemical potentials are not independent but are related
by (19.12) which is again the manifestation that, in equilibrium, the system in fact
should be considered a two-component system rather than three.

Single Reaction

Here we extend the results above for a generic reaction among q chemical species.
At the same time we highlight the role of two conjugate quantities, the extent of
reaction and affinity, which are relevant for the description of the thermodynamics
of chemical reactions.
The definition of the extent of reaction for a generic reaction is done as follows.
Denote by Bi the chemical formula for each one of the q chemical species of a
reactive system and suppose the occurrence of the following reaction among them
X X
i0 Bi • i0 Bi ; (19.13)
i .reag:/ i .prod:/

where i0 > 0 are the stoichiometric coefficients. It is convenient to write the
chemical equation above in the equivalent form

X
q
i Bi • 0; (19.14)
iD1

where i D i0 < 0 for the reactants, i D i0 > 0 for the products and i D 0 for
the chemical species that do not participate in the reaction.
The extent of reaction is defined by D N1 =1 and the complementary vari-
ables Xi are defined such that they are invariant under the chemical reaction (19.14),
i
Xi D Ni  N1 ; i D 2; : : : ; q: (19.15)
1
19.1 Extent of Reaction and Affinity 363

The inverse relation is given by

N1 D 1 ; (19.16)
Ni D Xi C i ; i D 2; : : : ; q: (19.17)

The differential form for the internal energy

X
q
dU D TdS  pdV C i dNi ; (19.18)
iD1

becomes, in terms of the new variables, the following form

X
q
dU D TdS  pdV  Ad C i dXi ; (19.19)
iD2

where A is the affinity, given by

X
q
AD i i ; (19.20)
iD1

and, according to (19.19), defined by the derivative

@U
AD ; (19.21)
@

keeping constant S, V and Xi . In thermochemical equilibrium we should have A D 0

which gives the condition for equilibrium

X
q
i i D 0; (19.22)
iD1

which is the generalization of the condition (19.12).

Multiple Reactions

Now consider the occurrence of r chemical reactions among the q chemical species,
represented by

X
q
ij Bi • 0; j D 1; 2; : : : ; r: (19.23)
iD1
364 19 Thermochemistry

To each chemical reaction corresponds an extent of reaction j defined as follows.

For each reaction j we choose one chemical species that participate in this reaction
and attach to it the label j. The only requirement is that they are all distinct. The
extents of reaction j are defined by the equation

X
r
Ni D ij j ; i D 1; 2; : : : ; r; (19.24)
jD1

and the complementary variables Xi , by the equation

X
r
Ni D Xi C ij j ; i D r C 1; : : : ; q: (19.25)
jD1

In order that j actually describe the chemical reactions it is necessary that the
variables Xi are invariant under each of the chemical reactions. From (19.24),

X
r
j D aji Ni ; j D 1; 2; : : : ; r; (19.26)
iD1

where the coefficients aij are understood as the elements of the matrix that is the
inverse of the matrix whose elements are the stoichiometric coefficients ij and are
related by

X
r
aji ik D ıjk : (19.27)
iD1

Let us suppose that the k-th reaction causes a change in Ni equal to Ni D ik .
From (19.26) and from relation (19.27), we conclude that the change in j will be
 j D ıjk . Substituting these results in

X
r
Ni D Xi C ij  j ; i D r C 1; : : : ; q: (19.28)
jD1

which is obtained from (19.25), we conclude that Xi D 0, a result that tell us that
Xi is indeed invariant under the k-th reaction.
Replacing (19.26) into (19.25) and solving for Xi , we also get

X
r X
r
Xi D Ni  ij ajk Nk ; i D r C 1; : : : ; q; (19.29)
jD1 kD1

which together with (19.26) comprises the inverse transformation of the transforma-
tion defined by (19.24) and (19.25).
19.1 Extent of Reaction and Affinity 365

Substituting the results (19.24) and (19.25) in (19.18), we obtain

X
r X
q
dU D TdS  pdV  Aj d j C i dXi ; (19.30)
jD1 iDrC1

where the affinities Aj are given by

X
q
Aj D  i ij ; (19.31)
iD1

and, according to (19.30), defined by the derivative

@U
Aj D  ; (19.32)
@ j

keeping constant S, V, Xi and the other extents of reaction. In equilibrium Aj D 0

for each reaction and therefore the following condition must be fulfilled

X
q
i ij D 0; (19.33)
iD1

which is the condition for thermochemical equilibrium for the case of multiple
reactions.
The equilibrium condition (19.33) tells us that the chemical potentials i are not
independent. The number of independent chemical potential equals the number of
chemical species q minus the number of chemical reactions r, that is, c D q  r. A
system of r chemical reactions and q chemical species must therefore be understood
as a mixture of c D qr components. The chemical reactions (19.23) or equivalently
(19.33), which connect the chemical potentials, should be linearly independent,
which means that none can be obtained from a linear combination of the others.

Closed System and Thermodynamics Potentials

A closed system allows the exchange of heat and work but not the exchange
of particles with the external environment. The only variations of the number of
moles Ni are those dictated by the chemical reactions. As the quantities Xi are not
affected by these variations, they remain constant. For a closed system the following
differential form is valid for the internal energy U

X
r
dU D TdS  pdV  Aj d j : (19.34)
jD1
366 19 Thermochemistry

A thermodynamic system can be described in an equivalent manner by means of

other thermodynamic potentials obtained from the internal energy U by Legendre
transformations. From U, we introduce the Helmholtz free energy F D U  TS, the
enthalpy H D U C pV and the Gibbs free energy G D H  TS. The differential
forms of these potentials differ from that related to the internal energy by the first
two terms of (19.30) or (19.34). For closed systems with only one chemical reaction,
described by the extension reaction , the differential forms of the internal energy
and other thermodynamic potentials are given by

dU D TdS  pdV  Ad ; (19.35)

dF D SdT  pdV  Ad ; (19.36)
dH D TdS C Vdp  Ad ; (19.37)
dG D SdT C Vdp  Ad : (19.38)

The affinity A is given by the following equivalent forms

       
@U @F @H @G
AD D D D : (19.39)
@ SV @ TV @ Sp @ Tp

It is worth emphasizing that the equilibrium properties are those obtained by taking
the limit A ! 0.

19.2 Thermodynamic Coefficients

Heat of Reaction

We examine the thermochemical processes in a closed system with only one

chemical reaction, described by the extent of reaction , for which the differential
forms (19.35)–(19.38) are valid. There are two basic thermodynamic processes as
to how heat is transferred to a closed system. In one, the volume is kept constant,
dV D 0. In this case there is no work done and heat introduced TdS becomes equal
to the change of internal energy, dU D TdS, because A D 0. In the other case, the
pressure p is kept constant. In this case, the system does work equal to pdV and
the heat introduced becomes equal to the change in enthalpy. Indeed, taking into
account that A D 0 and p is constant, then TdS D dU C pdV D d.U C pV/ D dH.
The characterization of a chemical reaction as to the heat developed is made by
means of a quantity known as heat of reaction, defined as the ratio of heat exchanged
and variation in the extension reaction during a process at constant temperature. We
19.2 Thermodynamic Coefficients 367

define two types of heat of reaction. One is the isochoric heat of reaction, where the
volume is kept constant in addition to the temperature and defined by
   
@S @U
QV D T D : (19.40)
@ TV @ TV

The second equality is valid only when A D 0. The other is the isobaric heat
of reaction where the pressure is kept constant in addition to the temperature and
defined by
   
@S @H
Qp D T D : (19.41)
@ Tp @ Tp

The second equality is valid only when A D 0. Both heats of reaction are measured
in units of energy per mole. When the heat of reaction is negative, the system
releases heat to the environment and the reaction is exothermic. When the heat of
reaction is positive, the system absorbs heat from the environment and the reaction
is endothermic.
In a thermochemical process, in addition to the heat exchanged work may be
developed. At constant temperature the work pdV performed by the system is equal
to the decrease in Helmholtz free energy F D U  TS because, taking into account
that A D 0 and that T is constant, pdV D dU C TdS D d.U  TS/ D dF.
At constant pressure, the determination of the work is reduced to determining the
increase in volume.
To characterize a chemical reaction with respect to changes in volume, we
introduce the volume of reaction defined as the ratio between the volume change
at constant pressure and variation of the extent of reaction. There are two types of
volume of reaction depending on which quantity, in addition to pressure, is kept
constant: entropy or pressure. They are defined by
   
@V @V
VT D ; VS D : (19.42)
@ Tp @ Sp

The derivatives that define the heats of reaction and the volumes of reaction obey
the Maxwell relations
       
@S @A @S @A
D ; D ; (19.43)
@ TV @T V @ Tp @T p
       
@V @A @V @A
D ; D : (19.44)
@ Tp @p T @ Sp @p S
368 19 Thermochemistry

Heat Capacity and Compressibility

The heat capacities at constant volume Cv and at constant pressure Cp are defined
in the usual form,
   
@S @U
Cv D T D ; (19.45)
@T VA @T VA
   
@S @H
Cp D T D ; (19.46)
@T pA @T pA

where A is kept constant. Again, the second equality in each equation is valid only
when A D 0.
Other thermodynamic coefficients that can be used to characterize the thermo-
chemical processes are the isothermal and adiabatic compressibilities
   
1 @V 1 @V
T D  ; S D  ; (19.47)
V @p TA V @p SA

and the coefficient of thermal expansion

   
1 @V 1 @S
˛D D : (19.48)
V @T pA V @p TA

In the three cases, the affinity A remains constant.

Deriving the heat of reaction Qp with respect to temperature and using the
definition of isobaric heat capacity, we obtain the Kirchhoff law

@Qp @Cp
D ; (19.49)
@T @

whose integrated form,

Z T1
@Cp
Qp .T1 /  Qp .T0 / D dT; (19.50)
T0 @

allows the determination of the heat of reaction at a certain temperature when the
heat of reaction is known at another temperature.
Deriving the volume of reaction VT with respect to pressure and using the
definition of isothermal compressibility, we get

@VT @
D  .VT /: (19.51)
@p @
19.2 Thermodynamic Coefficients 369

Thermochemical Coefficient

Here we examine the properties resulting from the thermochemical stability. We

consider a closed system within which a single chemical reaction takes place which
is described by the extension reaction . We analyze the variations of this quantity
as a result of changes in other thermodynamic variables such as temperature and
pressure. We remark that a positive change,  > 0, means a displacement of the
reaction to the right, that is, there is an increase of the products and a decrease of
the reactants. A negative variation,  < 0, means a shift of the reaction to the left,
that is, there is an increase of the reactants and a decrease of the products.
If a thermochemical system is moved from equilibrium by a disturbance in the
extension reaction, the thermodynamic stability requires that the affinity resulting
from this changes restores the equilibrium. That is, if  > 0 then A < 0 and if
 < 0 then A > 0. In other terms, @A=@  0 and the quantity ˝, defined by
 
@A
˝D ; (19.52)
@ Tp

and which we call thermochemical coefficient, holds the property

˝  0; (19.53)

which is the expression of the thermochemical stability.

We may equally rely on the convexity of the thermodynamic potentials to get the
stability properties of thermochemical systems. We assume that the energy U is a
convex function of all variables S, V and . As a result of Legendre transformations,
the other potentials are convex functions only of the set of extensive variable. The
Helmholtz free energy F is convex function of V and , the enthalpy H of S and ,
and the Gibbs free energy of alone. The convexity of any one of these potentials
with respect to the extent of reaction leads to the result (19.53). For example, the
convexity of the Gibbs free energy means that locally we have @2 G=@ 2  0. Taking
into account that the affinity A D .@G=@ /Tp then ˝ D @A=@  0.
Other thermochemical coefficients are defined by using other thermodynamic
potentials. For example, we define
 
@A
˝V D  ; (19.54)
@ TV

which is equal to ˝V D @2 F=@ 2 . Since F is a convex function of then ˝V  0.

370 19 Thermochemistry

Thermochemical Identities I

The most direct consequence of the stability conditions of a thermodynamic system

has to do with the way a given thermodynamic variable changes as a result of
changes in its conjugate. The introduction of heat, that is, of the entropy never leads
to lowering the temperature, which means that the thermal capacity is positive or
zero. A compression, that is, an increase in pressure never leads to an increase in
volume, which means that the compressibility is positive or zero. As noted above,
an increase of the extension reaction never leads to an increase in the affinity, which
means that the thermochemical coefficient is positive or zero.
There are however other indirect consequences of stability, involving more than
two thermodynamic quantities. The bodies that expand with increasing temperature,
get warmer in compression. Those which contracts with increasing temperature, get
cooler in a compression. This result can be shown by using the thermodynamic
identity
 
@p .@V=@T/p ˛
D D : (19.55)
@T V .@V=@p/T T

Since T  0 then @p=@T and ˛ have the same sign. Therefore, the bodies that
expand with increasing temperature at constant pressure (˛ > 0), get warmer
with increasing pressure (@p=@T > 0). The bodies that contracts with increasing
temperature at constant pressure (˛ < 0), get cooler with increasing pressure
(@p=@T < 0). Using the identity
 
@p .@S=@T/p .@S=@T/p Cp
D D D ; (19.56)
@T S .@S=@p/T .@V=@T/p TV˛

we get the same rule for an adiabatic compression.

Next we present the extension of these results, due to De Donder, for situations
where and A are involved. The variation of the extent of reaction as a result of a
variation in temperature at constant pressure is characterized by the derivative
 
@ .@A=@T/p .@S=@ /Tp
D D ; (19.57)
@T pA .@A=@ /Tp .@A=@ /Tp

where we made use of a Maxwell relation. Therefore, from the definitions of Qp and
˝, we reach the following result
 
@ Qp
D : (19.58)
@T pA T˝

As the denominator is positive, since ˝  0, then @ =@T and Qp have the same
sign. If the reaction is endothermic (Qp > 0) then an increase in the temperature at
19.2 Thermodynamic Coefficients 371

constant pressure shifts the reaction to the right, toward the products (@ =@T > 0).
If the reaction is exothermic (Qp < 0) then an increase in temperature at constant
pressure shifts the reaction to the left, toward reactants (@ =@T < 0). In both cases,
the equilibrium is shifted in the direction where heat is absorbed, a rule established
by van ’t Hoff and by Le Chatelier. The same result can be achieved if the volume is
maintained constant in the place of pressure. To this end it suffices to use the identity
 
@ QV
D ; (19.59)
@T VA T˝V

which is obtained in same manner.

Similarly, the variation in the extent of reaction as a result of changes in pressure
made at constant temperature is characterized by the derivative
 
@ .@A=@p/T .@V=@ /Tp
D D ; (19.60)
@p TA .@A=@ /Tp .@A=@ /Tp

where we have used a Maxwell relation. Using the definitions of VT and ˝, we get
the result
 
@ VT
D : (19.61)
@p T;A ˝

Since the denominator is positive then .@ =@p/ and VT have opposite signs. In
reactions that are accompanied by a decrease in the volume at constant temperature
and pressure (VT < 0), an increase in pressure shifts the reaction to the right, toward
the products (@ =@p > 0). Those that are accompanied by an increase in volume
at constant temperature and pressure (VT > 0), an increase in pressure shifts the
reaction to the left, toward the reactants (@ =@p < 0). In both cases, the equilibrium
is shifted in the direction in which the volume decreases, a rule established by Le
Chatelier. The same result is obtained if one considers the variation in the extent of
reaction as a result of an adiabatic change in pressure.

Thermochemical Identities II

We examine now the identities analogous to that existing between the heat capacity
at constant pressure and at constant volume,

TV˛
Cp D CV C : (19.62)
T

Suppose that we could insert or remove heat from a thermochemical system while
preventing the reactions to occur. In this situation the extent of reaction would
372 19 Thermochemistry

remain constant while heat is exchanged. The thermal capacity of the system in this
situation should be defined by
   
@S @H
Cp D T D ; (19.63)
@T p @T p

where the derivative is performed keeping constant instead of A constant. This

heat capacity is distinct from Cp as defined by (19.46) where A is kept constant. We
show that in fact Cp  Cp . To this end we start from the identity
       
@S @S @S @
D C : (19.64)
@T pA @T p @ Tp @T pA

Using the results above, we find

Q2p
Cp D Cp C : (19.65)
T˝
Since ˝  0, then Cp  Cp , that is, the heat capacity of a reactive system would be
greater if the reaction did not occur. This result tells us in addition that an increase in
the temperature induces an endothermic reaction such that the heat input is greater
than the heat required in the absence the reaction.
Analogously one can imagine the change in volume in the situation where is
taken as constant. This situation is characterized by the compressibility at constant
in the place of A constant,
 
1 @V
T D  : (19.66)
V @p T

Using the results above, we get

       
@V @V @V @
D C ; (19.67)
@p TA @p T @ Tp @p TA

VT2
T D T C : (19.68)
V˝
Since ˝  0, then T  T , that is, the compressibility of a reaction system is
greater than the compressibility in the absence of reaction. This result also says that
an increase in the volume causes a decrease in pressure smaller than that which
would be obtained in the absence of the reaction.
The two results above can be summarized as follows. The variation of an
extensive quantity X due to a variation in its conjugate x is greater than the variation
that would result in the absence of reaction, .@X=@x/A > .@X=@x/ , which is the
statement of the Le Chatelier-Braun principle due to Ehrenfest.
19.3 Thermochemical Equilibrium 373

19.3 Thermochemical Equilibrium

Spontaneous Processes

Imagine an isolated system consisting of a vessel within which several chemical

species take place. The species are inert and the system is in equilibrium in a
given state. At a certain moment they cease to be inert and begin to react. We can
imagine, for example, that the reactants are separated and that at a particular instant
the walls that separate them are removed. From that moment the system evolves
spontaneously to a new equilibrium state whose entropy is greater than that of the
initial state. The entropy change obeys the inequality S  0, which can therefore
be used as a criterion for occurrence of spontaneous reactions in isolated systems.
In systems that are not isolated, spontaneous processes may occur with decreas-
ing entropy. In these situations we must use other criteria. If the temperature T and
volume V are kept constant, the system evolves to a state in which the Helmholtz
free energy is smaller. The change in the Helmholtz free energy obeys the inequality

F D U  TS  0; (19.69)

which is the criterion for the occurrence of spontaneous reactions systems main-
tained at T and V constant. If the temperature T and pressure p are constant, the
system evolves to a state in which the Gibbs free energy is smaller. The change in
the Gibbs free energy obeys the inequality

G D H  TS  0; (19.70)

which is the criterion for the occurrence of spontaneous reactions in systems

maintained at T and p constant.
The exothermicity of a chemical reaction, characterized by H < 0, is not
sufficient to ensure spontaneity in processes at T and p constant, unless S > 0 or
the temperature is sufficiently low. In this regime, according to the Nernst principle,
S becomes very small, and vanishes when T ! 0 so that H approaches G.
The criterion H < 0 was recognized by Berthelot and Thomsen, but it alone is
valid only at sufficiently low temperatures. Endothermic reactions, characterized by
H > 0, may indeed occur spontaneously at high temperatures, regime in which
the second portion can prevail over the first
The obedience to the above criteria do not guarantee that the reaction will
actually occur because the criteria do not include information on reaction rates.
According to Table 19.1, the formation of water from hydrogen and oxygen is very
favorable but the reaction rate is very low frustrating the reactive process.
374 19 Thermochemistry

Table 19.1 Standard enthalpy h, standard entropy s and standard Gibbs free energy g of
some reactions at the pressure of 1 atm and temperature of 25 ı C. Source: CRC
Reactants Products h (kJ/mol) s (J/mol) g (kJ/K mol)
2H2 + O2 2H2 O(l) 571:6 326:8 474:2
N2 + 3H2 2NH3 91:8 198:3 32:8
2CO + O2 2CO2 566:0 173:1 514:4
2NO + O2 2NO2 116:2 146:6 72:6
2N2 O 2N2 + O2 163:2 148:5 207:4
CH4 + 2O2 CO2 + 2H2 O(l) 889:5 243:3 818:1
C2 H6 O(l) + 3O2 2CO2 + 3H2 O(l) 1368 138:9 1326

Standard Enthalpy of Reaction

In the study of thermochemical processes it is essential to determine the variation of

thermodynamic potentials in cases in which the reactants and products are separated
in their pure states. More precisely, in the initial state there are only reactants,
which are separated in their pure states, at a certain temperature and pressure, and
whose quantities in moles are equal to the stoichiometric coefficients of reaction.
In the final stage there are only products, which are separated in their pure states,
at the same temperature and pressure of the reactants, and whose quantities in
moles are equal to the stoichiometric coefficients of reaction. As the reactants
and products are separated, the initial thermodynamic potential is the sum of the
potential of the individual reactants and the final potential is the sum of potential of
the products individually. Taking as reference the reaction (19.13), the variations of
thermodynamic potentials, u, f , h and g, are given by
X X
u D i u i ; f D i fi ; (19.71)
i i
X X
h D i h i ; g D i g i ; (19.72)
i i

and are called standard energy of reaction, standard Helmholtz free energy of
reaction, standard enthalpy of reaction and standard Gibbs free energy of reaction.
The quantities ui , fi , hi and gi are respectively the molar quantities of each of
the chemical species alone. Note that the four expressions above are in fact
differences between the thermodynamic potentials of the products and reactants of
the reaction (19.13) because, according to the convention used the stoichiometric
coefficients are positive for products and negative for the reactants.
It is worth mentioning that the standard enthalpy of reaction as well as other
standard potentials have the additivity property. If we consider that a certain
chemical reaction, as the reaction (19.14), is equivalent to several intermediate
reactions, which we consider to be those given by (19.23), that is, such that
19.3 Thermochemical Equilibrium 375

P
i D j ij then the standard enthalpy of reaction h is equal to the sum of the
standard enthalpies h j of the intermediate reactions
X
h D h j ; (19.73)
j

which is the Hess law. The demonstration is as follows

X XX XX X
h j D ij hi D ij hi D i hi D h: (19.74)
j j i i j i

The standard entropy of reaction and the standard volume of reaction are given
by
X X
s D i si ; v D i vi ; (19.75)
i i

where si and vi are the molar entropy and the molar volume, respectively of the
chemical species alone. ForP ideal gases vi D v D RT=p is independent
P of the gas
and in this case v D v i i and the work done is pv D RT i i . Therefore,
if the sum of the stoichiometric coefficients of the products is greater than those of
the reactants, the volume will increase and there will be work done. The standard
potentials of reaction are related by

h D u C pv; (19.76)
f D u  Ts; (19.77)
g D h  Ts: (19.78)

Table 19.1 shows the values of h, s and g for some reactions. The values for
the reverse reactions are obtained by changing the signs of these values.
The standard enthalpy of reaction corresponding to the synthesis of one mole of
a substance from its constituents is called enthalpy of formation. By convention,
constituents are simple chemical substances stable at the pressure of 1 atm and
temperature of 25 ı C. The enthalpy of formation water is the standard enthalpy of
the reaction H2 +(1/2)O2 ! H2 O. Note that hydrogen and oxygen are gases at the
standard state while the water is liquid at the standard state. The heat of combustion
is defined as the standard enthalpy of reaction of one mole of a substance with
oxygen having as products CO2 and H2 O.
376 19 Thermochemistry

Equilibrium Constant

We have seen that in thermochemical equilibrium the chemical potentials of the

various chemical species are not independent but are related in accordance with
(19.22),

X
q
i i D 0; (19.79)
iD1

valid for a system within which a single chemical reaction occurs, represented
by (19.14). An equivalent way of describing the thermochemical equilibrium is
through the activity of each chemical species defined as follows. For each chemical
species, we introduce a reference chemical potential ıi which depends only on T
and p. We define the relative activity of ai , or simply activity by
ı
ai D e. i  i /=RT ; (19.80)

and therefore

i D ıi C RT ln ai : (19.81)

Replacing (19.81) into (19.79), we get the condition for equilibrium in terms of
activities
X
i f ıi C RT ln ai g D 0; (19.82)
i

which we write in the form


a11 a22 : : : aqq D K; (19.83)

where K, called equilibrium constant, is given by

X
RT ln K D  i ıi ; (19.84)
i

and depends only on T and p. Since the sum at right is the standard Gibbs free
energy we reach the following relation

RT ln K D g; (19.85)

which can be written as

K D eg=RT : (19.86)
19.3 Thermochemical Equilibrium 377

Guldberg-Waage Law

The equilibrium equation (19.83) is as general as the condition (19.79), being valid
for any system. To use any of these two conditions for specific cases, we must
know the dependence of the potential chemicals or activities with T, p and Ni . Next
we consider the case of dilute solutions and of ideal solutions, which include the
mixtures of ideal gases. For these systems the reference chemical potential can be
chosen such that the relative activity ai of a chemical species coincides with its mole
fraction xi , that is,

a i D xi ; (19.87)

or equivalently

i D ıi C RT ln xi : (19.88)
P
The mole fractions are defined by xi D Ni =N, where N D i Ni . The condition
for equilibrium in terms of the mole fraction reads therefore

x11 x22 : : : xqq D K: (19.89)

Recall that K depends only on temperature T and pressure p. Note that the
expression at left is actually a ratio because some stoichiometric coefficients are
negative and others positive. For the reaction (19.3), 1 D 1, 2 D 3 and
3 D 2 so that the left side equals x23 =x1 x32 . We remark that an increase in the
equilibrium constant means an increase of the products at the expense of reactants:
the reaction shifts to the right or toward the products. A decrease in the constant
means a decrease of the products and an increase of the reactants: the reaction shifts
to the left or towards the reactants.
The chemical potential i of a component of an ideal gas mixture is given by

i D gi C RT ln xi ; (19.90)

where gi is the molar Gibbs free energy of component i alone and is given by

RT
gi D ˛i  RT ln : (19.91)
p

The first term ˛i only depends on temperature and is given by ˛i D bi T ci T ln RT,
where bi and ci are constants specific to each gas. We can therefore identify gi with
the reference chemical potential ıi and write the equilibrium constant (19.84) as

RT
ln K D ln Kc C  ln ; (19.92)
p
378 19 Thermochemistry

P
where  D i i and Kc is defined by
X
RT ln Kc D  i ˛i ; (19.93)
i

and depends only on temperature. We see that the equilibrium constant K corre-
sponding to a mixture of ideal gases varies with pressure according to
 
RT
K D Kc : (19.94)
p

Notice that Kc depends only on temperature and therefore an increased in pressure

may increase the equilibrium constant as long as  < 0 which occurs when the sum
of the stoichiometric coefficients of the products is less than that of the reactants. In
other words, when the reaction occurs with a decrease in number of molecules.
The equilibrium condition can be written in terms of concentrations, defined by
i D Ni =V. Since xi D Ni =N then xi D i V=N D i RT=p. Replacing this result
(19.89) and using (19.94) we arrive at the equilibrium equation in terms of the
concentrations,

11 22 : : : qq D Kc ; (19.95)

valid for a mixture of ideal gases. Notice that the equilibrium constant Kc depends
only on temperature. Equation (19.95) is known as the law of mass action or
Guldberg-Waage law of chemical equilibrium.
It is also possible to write the condition for equilibrium in terms of the partial
pressures. Using relation pi D i RT between the partial pressure pi of each gas and
the concentration i , we get the relation

p11 p22 : : : pqq D Kp ; (19.96)

valid for a mixture of ideal gases where Kp D Kc .RT/ and depends only on
temperature.

Van ’t Hoff Equation

The behavior of the equilibrium constant with temperature can be obtained from
standard enthalpy through the van ’t Hoff equation, obtained as follows. From
(19.85),

g
R ln K D  : (19.97)
T
19.3 Thermochemical Equilibrium 379

Taking the derivative with respect to temperature

@ ln K s g
R D C 2; (19.98)
@T T T

where we used the relation s D @g=@T. Using in addition the relation h D

g C Ts, we arrive at the van ’t Hoff equation
 
@ ln K h
D ; (19.99)
@T p RT 2

which relates the isobaric variations of the equilibrium constant with temperature
and the standard enthalpy h. If the reaction is endothermic (h > 0) then an
increase in temperature at constant pressure shifts the reaction to the right, toward
the products (@K=@T > 0). If the reaction is exothermic (h < 0) then an increase
in temperature at constant pressure shifts the reaction to the left, toward the reactants
(@K=@T < 0). In both cases the equilibrium is displaced in the direction in which
heat is absorbed, a rule introduced originally by van ’t Hoff.
Similarly, we obtain the variations of the equilibrium constant with pressure,
at constant temperature. Deriving (19.85) with respect to pressure, and taking into
account that @g=@p D v we reach the result
 
@ ln K v
D : (19.100)
@p T RT

In reactions that are accompanied by a decrease in the volume at constant tempera-

ture and pressure (v < 0), an increase in pressure shifts the reaction to the right,
towards the products (@K=@p > 0). Those that are accompanied by an increase in
volume at constant temperature and pressure (v > 0), an increase in pressure
shifts the reaction to the left, towards the reactants (@K=@p < 0). In both cases, the
equilibrium is shifted in the direction where the volume decreases.

Ideal Gas Mixture

Here we determine the Gibbs free energy of a ideal gas mixture enclosed in a closed
vessel at a given temperature and pressure, given by
X
GD Ni .gi C RT ln xi /; (19.101)
i

P
where xi D Ni =N, N D i Ni and gi is the molar Gibbs free energy of the i-
th gas alone and depends only on of T and p. We consider a single chemical
reaction, represented by (19.14) and described by the extent of reaction . The
380 19 Thermochemistry

number of moles depends on according to (19.16) and (19.17). Here we examine

the situation in which at the minimum value of , which we consider to be zero,
the vessel contains only reactants, whose numbers of moles we denote by NiR . In
addition the number of moles is equal to the stoichiometric coefficients, that is,
NiR D .ji j  i /=2. At the maximum value of , which we consider to be the unity,
there are only products whose number of moles we denote by NiP . In addition the
number of moles is equal to the stoichiometric coefficient, that is, NiP D .ji jCi /=2.
From these conditions we find the relation between Ni and , given by

Ni D NiR .1  / C NiP : (19.102)

From these relation we find

X
Ni gi D gR .1  / C gP ; (19.103)
i

P P
where gR D i NiR gi is the free energy of the reactants alone and gP D i NiP gi is
the free energy of the products alone. The mole fraction is given by

NiR .1  / C NiP
xi D ; (19.104)
NR .1  / C NP
P P
where NR D i NiR and NP D i NiP . The equilibrium is determined by @G=@ D 0
and is given by
X
g C RT i ln xi D 0; (19.105)
i

where g D gP  gR , which can be written in the form

Y
xi i D eg=RT D K: (19.106)

Let us apply the above results for the synthesis of ammonia from its constituents
whose reaction is given by
1
=2 N2 C 3=2 H2 • NH3 : (19.107)

In this case 1 D 1=2, 2 D 3=2 and 3 D 1 so that NR D 2 and Np D 1 and

therefore

1 3.1  /
x1 D ; x2 D ; x3 D ; (19.108)
2.2  / 2.2  / 2
19.3 Thermochemical Equilibrium 381

which replaced in the equilibrium equation

x3
1=2 3=2
D K; (19.109)
x1 x2

leads us to the equation that determines

4 .2  /
p D K; (19.110)
3 3.1  /2

whose solution is
2
D1 p p : (19.111)
3 3K C 4

Table 19.1 gives the value g D 16:4 kJ/mol for the reaction (19.107), at the
pressure of 1 atm and temperature of 25 ı C, from which we can calculate K by
means of (19.106) with the following result K D 747 which gives D 0:968.
This value for the production of ammonia is large but the rate of reaction at room
temperature is very small. At high temperatures the rate increases and the process
becomes feasible as long as is appreciable. At the pressure of 1 atm and the
temperature of 500 ı C we use the result of problem 19.2 to obtain the following
value for the equilibrium constant K D 0:0034 from which we get D 0:0022,
which is very small. Increasing the pressure it is possible to increase . Indeed,
according to (19.94) the equilibrium constant is proportional to p . In the present
case  D 1 C 2 C 3 D 1 and thus K is proportional to the pressure p.
Therefore, for a pressure of 100 atm, and temperature of 500 ı C, K D 0:34 which
gives D 0:16, a reasonable value. The use of high pressures ant high temperatures,
together with the use of a catalyst, comprises the Haber method for the production
of ammonia.

Consistency of the Equilibrium Constants

We have seen that a system consisting of various chemical reactions, represented

by (19.23), the equilibrium conditions must take into account each one of the
reactions. These conditions are given by (19.33), valid for each one of the chemical
equations. From them we get the equilibrium conditions for each equation in terms
of the activities, mole fractions, and concentrations. The forms of these conditions
are analogous to those valid for a single chemical equation. For example, the
equilibrium condition in terms of the activities with respect to the j-th chemical
equation is given by
  
a1j1 a2j2 : : : aqjq D Kj ; (19.112)
382 19 Thermochemistry

where Kj is the equilibrium constant associated to the j-th chemical equation.

A chemical reaction such as that represented by the chemical equation (19.14)
may be composed by a set of intermediate reactions. We assume that they form a
set of r reactions represented by the chemical equations (19.23). As the intermediate
reactions should result in the overall reaction then the stoichiometric coefficients are
related by

X
r
i D ji : (19.113)
jD1

In the thermodynamic equilibrium, the chemical equilibrium equations must be

valid for the overall reaction and for each of the intermediate reactions.
The consistency of the equilibrium equations must be such that the equilibrium
equations of intermediate reactions should result in the equilibrium equation of the
overall reaction. In fact, using the result (19.113) we see that (19.33), which describe
the thermochemical equilibrium of the intermediate reactions in terms of chemical
potentials, imply (19.22), which represent the thermochemical equilibrium of the
overall reaction in terms of the chemical potentials. The same consistency exists for
the equilibrium equation in terms of activities, and also in terms of mole fractions
and concentrations. For each intermediate reaction the condition (19.112) is valid.
Multiplying (19.112) for j D 1; 2; : : : ; r and using the result (19.113), the product
results in the equilibrium equation (19.89) with K given by

K D K1 K2 : : : Kr ; (19.114)

or equivalently by

ln K D ln K1 C ln K2 C : : : C ln Kr : (19.115)

The validity of the Guldberg-Waage law for each intermediate reaction results
in the validity of the law for the overall reaction or any reaction that can be
considered as a combination of intermediate reactions. The equilibrium constants
of the intermediate reactions are related to the equilibrium constant of the overall
reaction by (19.114) or equivalently (19.115).

Problems

19.1 The derivative of any function  that depends on the number of moles Ni , with
respect to the extent of reaction can be carried out as follows

@ X @
D i ;
@ i
@Ni
19.3 Thermochemical Equilibrium 383

because @Ni =@ D i . Use this result to show that for a ideal mixture the following
relations are valid

@H X
Qp D D i hi D h;
@ i

@V X @Cp X
VT D D i vi D v; D i ci ;
@ i
@ i

where ci are the molar heat capacities of the chemical species alone.
19.2 Use the results of the previous problem to show that
P the Kirchhoff law for an
ideal mixture is given by @h=@T D c, where c D i i ci . Supposing that c is
constant, h D h0 C c.T  T0 /. Use this result to obtain the following expression
for the equilibrium constant

K c T 1 1 1
ln D ln  .h0  cT0 /.  /:
K0 R T0 R T T0

Use this expression to determine K at 500 ı C for th reaction (19.107) from the data
at 25 ı C. The heat capacities of N2 , H2 and NH3 are 29.1, 28.8 and 35.1 J/mol K at
25 ı C.
19.3 Show that the thermochemical coefficient ˝ D .@A=@ /Tp can be obtained
explicitly for systems described by the chemical potential (19.88) and is
 2
RT X i
˝D   xi ;
N i xi
P
where  is defined by  D i i , which shows clearly that ˝  0.
19.4 Determine as a function of the equilibrium constant for the reaction

CO C 1=2 O2 • CO2 ;

where the reactants and products are considered to be ideal gases. Calculate at
1 atm and 25 ı C.
19.5 Show that the following van ’t Hoff equation is valid for an ideal gas mixture
 
@ ln Kc u
D ;
@T p RT 2

where
P Kc is the equilibrium constant in terms of the concentrations and u D
i i ui , where ui is the molar energy of the i-th component of the mixture.
References

General

1. M. Bailyn, A Survey of Thermodynamics (American Institute of Physics, New York, 1994)

2. H.B. Callen, Thermodynamics (Wiley, New York, 1960); Thermodynamics and an Introduc-
tion to Thermostatistics, 2nd edn. (Wiley, New York, 1985)
3. S. Carnot, Réflexions sur la Puissance Motrice du Feu et sur les Machines propes à
Developper cette Puissance (Bachelier, Paris, 1824)
4. É. Clapeyron, Mémoire sur la puissance motrice de la chaleur. J. de l’Ecole Royale
Polytechnique 14, 153 (1834)
5. R. Clausius, Über die bewegende Kraft der Wärme und die Gesetze welche sich daraus für
die Wärmelehre selbst ableiten lassen. Ann. der Phys. und Chem. 79, 368, 500 (1850)
6. R. Clausius, Über verschiedene für die Anwendung bequeme Formen der Hauptgleichungen
der mechanischen Wärmetheorie”. Ann. der Phys. und Chem. 125, 353 (1865)
7. P. Duhem, Thermodynamique et Chimie (Hermann, Paris, 1902); 2de édition (1910)
8. P. Duhem, Traité d’Énergétique ou de Thermodynamique Générale, 2 vols. (Gauthier-Villars,
Paris, 1911)
9. P.S. Epstein, Textbook of Thermodynamics (Wiley, New York, 1937)
10. E. Fermi, Thermodynamics (Prentice Hall, New York, 1937); (Dover, New York, 1956)
11. J.W. Gibbs, A method of geometrical representation of the thermodynamic properties of
substances by means of surfaces. Trans. Connecticut Acad. 2, 382 (1873)
12. J.W. Gibbs, On the equilibrium of heterogeneous substances. Trans. Connecticut Acad. 3, 108
(1876); 3, 343 (1878)
13. E.A. Guggenheim, Thermodynamics (North Holland, Amsterdam, 1949); 2nd edn. (1950);
3rd edn. (1957); 4th edn. (1959); 5th edn. (1967); 6th edn. (1977); 7th edn. (1985)
14. J.P. Joule, On the calorific effects of magneto-electricity, and on the mechanical value of heat.
Philos. Mag. 23, 263, 347, 435 (1843)
15. J.P. Joule, On the mechanical equivalent of heat. Philos. Trans. R. Soc. 140, 61 (1850)
16. C. Kittel, Thermal Physics (Wiley, New York, 1969)
17. D. Kondepudi, I. Prigogine, Modern Thermodynamics (Wiley, New York, 1998)
18. R. Kubo, Thermodynamics (North-Holland, Amsterdam, 1966)
19. L.D. Landau, E.M. Lifshitz, Statistical Physics (Clarendon Press, Oxford, 1938); (Pergamon,
Oxford, 1958); 2nd edn. (1969)
20. J.C. Maxwell, Theory of Heat (Longmans, London, 1871); 5th edn. (1877); 9th edn. (1888)

© Springer-Verlag Berlin Heidelberg 2017 385

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2
386 References

21. J.R. Mayer, Bemerkungen über die Kräfte der unbelebten Natur. Ann. der Chem. und Pharm.
42, 233 (1842)
22. W. Nernst, Ueber die Berechnung chemischer Gleichgewichte aus thermischen Messungen.
Kgl. Ges. d. Wiss. Gött. 1906, 1–40 (1906)
23. W. Nernst, Die theoretischen und experimentellen Grundlagen des neuen Wärmesatzes
(Knapp, Halle, 1918); 2te Auflage (1924)
24. A.B. Pippard, The Elements of Classical Thermodynamics (Cambridge University Press,
London, 1957)
25. M. Planck, Vorlesungen über Thermodynamik (Veit, Leipzig, 1897); 2te Auflage (1905); 3te
Auflage (1911); Walter de Gruyter, Berlin, 7te Auflage (1922); 9te Auflage (1930)
26. F. Reif, Fundamentals of Statistical and Thermal Physics (McGraw-Hill, New York, 1965)
27. D. Ruelle, Thermodynamic Formalism (Addison-Wesley, Reading, 1978)
28. Yu.B. Rumer, M.Sh. Ryvkin, Thermodynamics, Statistical Physics, and Kinetics (Mir,
Moscow, 1980)
29. F.W. Sears, Thermodynamics, the Kinetic Theory of Gases, and Statistical Mechanics (Addi-
son Wesley, Reading, 1950); 2nd edn. (1953); F.W. Sears, G.L. Salinger, Thermodynamics,
Kinetic Theory, and Statistical Thermodynamics, 3rd edn. (Addison-Wesley, Reading, 1975)
30. A. Sommerfeld, Thermodynamik und Statistik (Dieterich’sche Verlagsbuchhandlung, Wies-
baden, 1952)
31. R.E. Sonntag, G.J. Van Wylen, Introduction to Thermodynamics: Classical and Statistical
(Wiley, New York, 1971)
32. W. Thomson (Lord Kelvin), On an absolute thermodynamic scale, founded on Carnot’s theory
of the motive power of heat, and calculated from Regnault’s observations. Philos. Mag. 33,
313 (1848)
33. W. Thomson (Lord Kelvin), On the dynamical theory of heat, with numerical results from Mr.
Joule’s equivalent of a thermal unit, and M. Regnault’s observations on steam. Philos. Mag.
4, 8, 105, 168 (1852)
34. L. Tisza, Generalized Thermodynamics (MIT Press, Cambridge, 1966)
35. J.D. van der Waals, Over de Continuiteit van den Gas- en Vloeistoftoestand (Sijthoff, Leiden,
1873)
36. H.L.F. von Helmholtz, Über die Erhaltung der Kraft (Reimer, Berlin, 1847)
37. M.W. Zemansky, Heat and Thermodynamics (McGraw-Hill, New York, 1937); 2nd edn.
(1943); 3rd edn. (1951); 4th edn. (1957); 5th edn. (1968); M.W. Zemansky, R.H. Dittman,
Heat and Thermodynamics, An Intermediate Textbook, 6th edn. (McGraw-Hill, New York,
1981); 7th edn. (1996)

Liquids, Mixtures and Thermochemistry

38. P.W. Atkins, Physical Chemistry (Oxford University Press, Oxford, 1978); 2nd edn. (1982);
3rd edn. (1986); 4th edn. (1990); 5th edn. (1994); 6th edn. (1998)
39. T. De Donder, L’Affinité (Lamertin, Bruxelles, 1923)
40. A. Findlay, The Phase Rule and its Application (Longmans, London, 1904); 5th edn. (1923);
9th edn. (Dover, New York, 1951)
41. E.A. Guggenheim, Mixtures (Clarendon Press, Oxford, 1952)
42. W. Heitler, Zwei Beiträge zur Theorie konzentrierter Lösungen. Ann. Physik 80, 629 (1926)
43. J.H. Hildebrand, Solubility. XII. Regular solutions. J. Am. Chem. Soc. 51, 66 (1929)
44. J.H. Hildebrand, Solubility of Nonelectrolytes, 2nd edn. (Reinhold, New York, 1936); J.H.
Hildebrand, R.L. Scott, Solubility of Nonelectrolytes, 3rd edn. (Reinhold, New York, 1950)
45. J.G. Kirkwood, I. Oppenheim, Chemical Thermodynamics (McGraw-Hill, New York, 1961)
46. A.N. Krestóvnikov, V.N. Vigdoróvich, Termodinámica Química (Editorial Mir, Moscú, 1980)
47. G.N. Lewis, M. Randall, Thermodynamics and the Free Energy of Chemical Substances
(McGraw-Hill, New York, 1923); 2nd edn., 1961, revised by K.S. Pitzer and L. Brewer
References 387

48. W.J. Moore, Physical Chemistry (Prentice-Hall, New York, 1950); 2nd edn. (1955); 3rd edn.
(Prentice-Hall, Englewood Cliffs, NJ, 1962); 4th edn. (1972)
49. L.K. Nash, Elements of Chemical Thermodynamics (Addison-Wesley, Reading, 1962)
50. I. Prigogine, R. Defay, Thermodynamique chimique (Dunot, Paris, 1944); (Desoer, Liège,
1950)
51. J.S. Rowlinson, Liquids and Liquid Mixtures (Butterworths, London, 1959); J.S. Rowlinson,
F.L. Swinton, Liquids and Liquid Mixtures, 3rd edn. (Butterworth Scientific, London, 1982)
52. J. Waser, Basic Chemical Thermodynamics (Benjamin, New York, 1966)
53. E.N. Yeremin, Fundamentals of Chemical Thermodynamics (Mir, Moscow, 1981)

Solids and Alloys

54. N.W. Ashcroft, N.D. Mermin, Solid State Physics (Holt, Rinehart and Winston, New York,
1976)
55. W.L. Bragg, E.J. Williams, The effect of thermal agitation on atomic arrangement in alloys.
Proc. R. Soc. A 145, 699 (1934)
56. A.D. Bruce, R.A. Cowley, Structural Phase Transitions (Taylor and Francis, London, 1981)
57. P. Gordon, Principles of Phase Diagrams in Materials Systems (McGraw-Hill, New York,
1968)
58. W. Gorsky, Röntgenographische Untersuchung von Umwandlungen in der CuAu. Z. Physik
50, 64 (1928)
59. C. Kittel, Introduction to Solid State Physics (Wiley, New York, 1953); 4th edn. (1971); 7th
edn. (1996)
60. M.A. Krivoglaz, A.A. Smirnov, The Theory of Order-Disorder in Alloys (Elsevier, New York,
1965)
61. R. Kubo, T. Nagamiya, Solid State Physics (MacGraw-Hill, New York, 1969)
62. L.D. Landau, Zh. Éksp. Teor. Fiz. 7 19, 627 (1937)
63. L.D. Landau, E.M. Lifshitz, Theory of Elasticity (Pergamon, Oxford, 1959); 2nd edn. (1970)
64. T. Muto, Y. Takagi, The theory of order-disorder transitions in alloys. Solid State 1, 193 (1955)
65. D.C. Wallace, Thermodynamics of Crystals (Wiley, New York, 1972)

Magnetic and Ferromagnetic Materials

66. M.F. Collins, Magnetic Critical Scattering (Oxford University Press, New York, 1989)
67. P. Curie, Propriétés magnétiques des corps a diverses températures. Ann. de Chim. et de Phys.
5, 289 (1895)
68. L.J. de Jongh, A.R. Miedema, Experiments on simple magnetic model systems. Adv. Phys.
23, 1 (1974)
69. D. Jiles, Introduction to Magnetism and Magnetic Materials (Chapman and Hall, London,
1991)
70. L. Landau, Eine mögliche erklärung der feldabhängigkeit der suszeptibilität bei niedrigen
temperaturen. Phys. Z. Sowjetunion 4, 675 (1933)
71. R.A. McCurrie, Ferromagnetic Materials (Academic, London, 1994)
72. L. Néel, Propriétés magnétiques des ferrites; ferrimagnétisme et antiferromagnétisme. Ann.
Phys. 3, 137 (1948)
73. J.S. Smart, Effective Field Theories of Magnetism (Saunders, Philadelphia, 1966)
74. J.H. Van Vleck, The Theory of Electric and Magnetic Susceptibilities (Oxford University
Press, London, 1932)
388 References

75. P. Weiss, L’hypothèse du champ moléculaire et la propriété ferromagétique. J. Phys. 6, 661

(1907)
76. P. Weiss, G. Foëx, Le Magnétisme (Librairie Armand Colin, Paris, 1926)

Dielectrics and Ferroelectrics

77. J.C. Burfoot, Ferroelectrics (D. van Nostrand, London, 1967)

78. P. Debye, Einige Resultate einer kinetischen Theorie der Isolatoren. Phys. Z. 13, 97 (1912)
79. P. Debye, Polare Molekeln (Hirzel, Leipzig, 1929)
80. A.F. Devonshire, Theory of barium titanate. Part I. Philos. Mag. 40, 1040 (1949); Theory of
barium titanate. Part II. Philos. Mag. 42, 1065 (1951)
81. E. Fatuzzo, W.J. Merz, Ferroeletricity (North-Holland, Amsterdam, 1967)
82. W. Känzig, Ferroelectrics and Antiferroelectrics (Academic, New York, 1957)
83. M.E. Lines, A.M. Glass, Principles and Applications of Ferroelectrics and Related Materials
(Clarendon Press, Oxford, 1979)
84. J.C. Slater, Theory of the transition in KH2 PO4 . J. Chem. Phys. 9, 16–33 (1941)

Liquid Crystals

85. P.J. Collings, Liquid Crystals (Princeton University Press, Princeton, 1990)
86. P.G. de Gennes, The Physics of Liquid Crystals (Clarendon Press, Oxford, 1974); P.G. de
Gennes, J. Prost, The Physics of Liquid Crystals, 2nd edn. (Clarendon Press, Oxford, 1993)
87. A.M. Figueiredo Neto, S.R.A. Salinas, The Physics of Lyotropic Liquid Crystals (Oxford
University Press, Oxford, 2005)
88. E.B. Priestley, P.J. Wojtowicz, P. Sheng (eds.), Introduction to Liquid Crystals (Plenum, New
York, 1974)

Thermal Radiation

89. M. Planck, Ueber eine Verbesserung der Wien’schen Spectralgleichung. Verh. Dtsch. Phys.
Ges. 2, 202 (1900)
90. M. Planck, Vorlesungen über Theorie der Wärmestrahlung (Barth, Leipzig, 1906); 2te auflage
(1913)

Phase Transitions and Critical Phenomena

91. C. Domb, The Critical Point (Taylor and Francis, London, 1996)
92. M.E. Fisher, The theory of equilibrium critical phenomena. Rep. Prog. Phys. 30, 615 (1967)
93. R.B. Griffiths, Thermodynamic functions for fluids and ferromagnets near the critical point.
Phys. Rev. 158, 176 (1967)
94. R.B. Griffiths, Thermodynamic model for tricritical points in ternary and quaternary fluid
mixtures. J. Chem. Phys. 60, 195 (1974)
References 389

95. R.B. Griffiths, Phase diagrams and higher-order critical points. Phys. Rev. B 12, 345 (1975)
96. R.B. Griffiths, J.C. Wheeler, Critical points in multicomponent systems. Phys. Rev. A 2, 1047
(1970)
97. R.B. Griffiths, J.C. Wheeler, The Thermodynamics of Phase Transitions (1976)
98. P. Heller, Experimental investigations of critical phenomena. Rep. Prog. Phys. 30, 731 (1967)
99. H.E. Stanley, Introduction to Phase Transitions and Critical Phenomena (Oxford University
Press, New York, 1971)
100. J.-C. Tolédano, P. Tolédano, The Landau Theory of Phase Transitions (World Scientific,
Singapore, 1987)
101. B. Widom, Equation of state in the neighborhood of the critical point. J. Chem. Phys. 43,
3898 (1965)
102. J. Zernike, General considerations concerning the number of virtual phases. Rec. Trav. Chim.
68, 585 (1949)

Collectanea

103. J. Kestin (ed.), The Second Law (Dowden, Hutchinson and Ross, Stroudsburg, Pennsylvania,
1976)
104. R.B. Lindsay (ed.), Energy: Historical Development of the Concept (Dowden, Hutchinson
and Ross, Stroudsburg, Pennsylvania, 1975)
105. W.F. Magie, A Source Book in Physics (McGraw-Hill, New York, 1935)

Chapter 5

106. A. Michels, W. de Graaff, T. Wassenaar, J.M.H. Levelt, P. Louwerse, Physica 25, 25 (1959)
107. J.R. Roebuck, H. Osterberg, Phys. Rev. 48, 450 (1935)

Chapter 6

108. K. Clusius, L. Riccoboni, Z. Phys. Chem. B 38, 81 (1937)

109. W.S. Corak, M.P. Garfunkel, C.B. Satterthwaite, A. Wexler, Phys. Rev. 98, 1699 (1955)
110. P. Debye, Ann. Phys. 39, 789 (1912)
111. E.D. Eastman, W.C. McGavock, J. Am. Chem. Soc. 59, 145 (1937)
112. A. Einstein, Ann. Phys. 22, 180 (1907)
113. W.F. Giauque, J.O. Clayton, J. Am. Chem. Soc. 55, 4875 (1933)
114. W.F. Giauque, J.W. Stout, J. Am. Chem. Soc. 58, 1144 (1936)
115. P.H. Keesom, N. Pearlman, Phys. Rev. 91, 1354 (1953)
116. J.F. Nagle, J. Math. Phys. 7, 1484 (1966)
117. W. Nernst, Kgl. Ges. d. Wiss. Gött. 1906, 1–40 (1906)
118. L. Pauling, J. Am. Chem. Soc. 57, 2680 (1935)
119. M. Planck, Vorlesungen über Thermodynamik (Veit, Leipzig, 1911); 3te Auflage, p. 266
120. O. Sackur, Ann. Phys. 36, 958 (1911)
121. A. Sommerfeld, Z. Phys. 47, 1 (1928)
122. H. Tetrode, Ann. Phys. 38, 434 (1912)
123. G.K. White, J.G. Collins, J. Low Temp. Phys. 7, 43 (1972)
390 References

Chapter 8

124. L. Beck, G. Ernst, J. Gürtner, J. Chem. Thermodyn. 34, 277 (2002)

125. Table 8.2. R.D. Goodwin, L.A. Weber, J. Res. NBS A 73, 1 (1969); M. Barmatz, Phys. Rev.
Lett. 24, 651 (1970); H.D. Bale, B.C. Dobbs, J.S. Lin, P.W. Schmidt, Phys. Rev. Lett. 25,
1556 (1970); L.A. Weber, Phys. Rev. A 2, 2379 (1970); G.R. Brown, H. Meyer, Phys. Rev.
A 6, 364 (1972); A. Tominaga, Y. Narahara, Phys. Lett. A 41, 353 (1972); C.E. Chase, G.O.
Zimmerman, J. Low Temp. Phys. 11, 551 (1973); H.A. Kierstead, Phys. Rev. A 7, 242 (1973);
A.V. Voronel, V.G. Gorbunova, V.A. Smirnov, N.G. Shmakov, V.V. Shchekochikhina, Sov.
Phys. JETP 36, 505 (1973); J.M.H. Levelt-Sengers, Physica 73, 73 (1974); M. Barmatz,
P.C. Hohenberg, A. Kornblit, Phys. Rev. B 12, 1947 (1975); R. Hocken, M.R. Moldover,
Phys. Rev. Lett. 37, 29 (1976); A.V. Voronel, in Phase Transitions and Critical Phenomena,
vol. 5b, p. 343, ed. by C. Domb, M.S. Green (Academic, New York, 1976); D.R. Douslin,
R.H. Harrison, J. Chem. Thermodyn. 8, 301 (1976); D. Balzarini, M. Burton, Can. J. Phys.
57, 1516 (1979); M.W. Pestak, M.H.W. Chan, Phys. Rev. B 30, 274 (1984); J.R. de Bruyn,
D.A. Balzarini, Phys. Rev. A 36, 5677 (1987); J.R. de Bruyn, D.A. Balzarini, Phys. Rev. B
39, 9243 (1989); L. Beck, G. Ernst, J. Gürtner, J. Chem. Thermodyn. 34, 277 (2002)
126. R.B. Griffiths, Phys. Rev. 158, 176 (1967)
127. A. Michels, B. Blaisse, C. Michels, Proc. R. Soc. A 160, 358 (1937)
128. A.V. Voronel, V.G. Gorbunova, V.A. Smirnov, N.G. Shmakov, V.V. Shchekochikhina, Sov.
Phys. JETP 36, 505 (1973)

Chapter 10

129. A.S. Darling, R.A. Mintern, J.C. Chaston, J. Inst. Met. 81, 125 (1952–1953)
130. B.F. Dodge, A.K. Dunbar, J. Am. Chem. Soc. 49, 591 (1927)
131. J.L. Murray, Metall. Trans. A 15, 261 (1984)
132. H. Stöhr, W. Klemm, Z. Anorg. Allgem. Chem. 241, 305 (1939)
133. Table 10.1. D.R. Thompson, O.K. Rice. J. Am. Chem. Soc. 86, 3547 (1964); B. Chu, F.J.
Schoenes, W.P. Kao, J. Am. Chem. Soc. 90, 3042 (1968); A.M. Wims, D. McIntyre, F. Hynne,
J. Chem. Phys. 50, 616 (1969); P.N. Pusey, W.I. Goldburg, Phys. Rev. A 3, 766 (1971); H.K.
Schurmann, R.D. Parks, Phys. Rev. Lett. 26, 367 (1971); D. Balzarini, Can. J. Phys. 52, 499
(1974); E.S.R. Gopal et al., Phys. Rev. Lett. 32, 284 (1974); S.C. Greer, R. Hocken, J. Chem.
Phys. 63, 5067 (1975); J.I. Lataille, T.S. Venkataraman, L.M. Narducci, Phys. Lett. A 53, 359
(1975); S.C. Greer, Phys. Rev. A 14, 1770 (1976); D.T. Jacobs, et al., Chem. Phys. 20, 219
(1977); D. Beysens, A. Bourgon, Phys. Rev. A 19, 2407 (1979); M.A. Anisimov et al., Sov.
Phys. JETP 49, 844 (1979); D. Beysens, J. Chem. Phys. 71, 2557 (1979); A. Sivaraman et
al., Ber. Bunsen Phys. Chem. 84, 196 (1980); J. Shelton, D. Balzarini, Can. J. Phys. 59, 934
(1981); M. Nakata, T. Dobashi, N. Kuwahara, M. Kaneko, B. Chu, Phys. Rev. A 18, 2683
(1987)

Chapter 12
134. W.L. Bragg, E.J. Williams, Proc. R. Soc. A 145, 699 (1934)
135. Table 12.1. O.W. Dietrich, J. Als-Nielsen, Phys. Rev. 153, 711 (1967); J. Als-Nielsen, O.W.
Dietrich, Phys. Rev. 153, 717 (1967); J. Ashman, P. Handler, Phys. Rev. Lett. 23, 642 (1969);
M.B. Salamon, F.L. Lederman, Phys. Rev. B 10, 4492 (1974); M.F. Collins, Magnetic Critical
Scattering (Oxford University Press, Oxford, 1989)
References 391

136. C. Franz, M. Gantois, J. Appl. Crystallogr. 4, 387 (1971)

137. W. Gorsky, Z. Phys. 50, 64 (1928)
138. L.D. Landau, Zh. Éksp. Teor. Fiz. 7 19, 627 (1937)
139. W. Shockley, J. Chem. Phys. 6, 130 (1938)
140. Figure 12.2. C. Sykes, H. Wilkinson, J. Inst. Met. 61, 223 (1937); O. Rathmann, J. Als-
Nielsen, Phys. Rev. B 9, 3921 (1974); P.K. Kumar, L. Muldawer, Phys. Rev. B 14, 1972
(1976); M. Hansen, Constitution of Binary Alloys, 2nd edn. (McGraw-Hill, New York, 1958)

Chapter 13

141. L. Brillouin, J. Phys. 8, 74 (1927)

142. D.L. Connelly, J.S. Loomis, D.E. Mapother, Phys. Rev. B 3, 924 (1971)
143. P. Curie, Ann. Chim. Phys. 5, 289 (1895)
144. Table 13.3. P. Heller, Rep. Prog. Phys. 30, 731 (1967); W. Rocker, R. Kohlhass, Z.
Naturforsch. A 22, 291(1967); J.S. Kouvel, J.B. Comly, J. Appl. Phys. 20, 1237 (1968); J.T.
Ho, J.D. Litster, Phys. Rev. B 2, 4523 (1970); Phys. Rev. Lett. 22, 603 (1969); N. Menyuk,
K. Dwight, T.B. Reed, Phys. Rev. B 3, 1689 (1971); G. Ahlers, A. Kornblit, Phys. Rev. B 12,
1938 (1975); A. Kornblit, G. Ahlers, E. Buehler, Phys. Rev. B 17, 282 (1978); C.J. Glinka,
V.J. Minikiewicz, PRB 16, 4084 (1977); M.F. Collins, Magnetic Critical Scattering (Oxford
University Press, New York, 1989); T. Tanaka, K. Miyatani, J. Appl. Phys. 82, 5658 (1997)
145. W.E. Henry, Phys. Rev. 88, 559 (1952)
146. P. Langevin, Ann. Chim. Phys. 5, 70 (1905); J. Phys. 4, 678 (1905)
147. P. Weiss, J. Phys. 6, 661 (1907)
148. P. Weiss, R. Forrer, Ann. Phys. 5, 153 (1926)

Chapter 14

149. M.E. Fisher, Physica 26, 618 (1960); Philos. Mag. 7, 1731 (1962)
150. Table 14.2. A. Kornblit, G. Ahlers, Phys. Rev. B 8, 5163 (1973); L.J. de Jongh, A.R. Miedema,
Adv. Phys. 23, 1 (1974); G. Ahlers, A. Kornblit, Phys. Rev. B 12, 1938 (1975); M. Barmatz,
P.C. Hohenberg, A.Kornblit, Phys. Rev. B 12, 1947 (1975); M.F. Collins, Magnetic Critical
Scattering (Oxford University Press, New York, 1989)
151. L. Landau, Phys. Z. Sowjetunion 4, 675 (1933)
152. D.P. Landau, B.E. Keen, B. Schneider, W.P. Wolf, Phys. Rev. B 3, 2310 (1970)
153. M.A. Lasher, J. van den Broek, C.J. Gorter, Physica 24, 1061, 1076 (1958)
154. L. Néel, Ann. Phys. 3, 137 (1948)
155. N.F. Oliveira Jr., A.P. Filho, S.R.A. Salinas, in AIP Conference Proceedings, vol. 29 (1975),
p. 463
156. R. Pauthenet, Ann. Phys. 3, 424 (1958)

Chapter 15

157. R. Blinc, J. Phys. Chem. Solids 13, 204 (1960)

158. P.G. de Gennes, Solid State Commun. 1, 132 (1963)
159. P. Debye, Phys. Z. 13, 97 (1912)
392 References

160. A.F. Devonshire, Philos. Mag. 40, 1040 (1949); Philos. Mag. 42, 1065 (1951)
161. S. Hoshino, T. Mitsui, F. Jona, R. Pepinsky, Phys. Rev. 107, 1255 (1957)
162. Y. Ishibashi, J. Phys. Soc. Jpn. 56, 2089 (1987)
163. W.J. Merz, Phys. Rev. 91, 513 (1953)
164. R. Sänger, O. Steiger, Gächter, Helv. Phys. Acta 5, 200 (1932)
165. J.C. Slater, J. Chem. Phys. 9, 16–33 (1941)
166. Y. Takagi, J. Phys. Soc. Jpn. 3, 271, 273 (1948)
167. C.T. Zahn, Phys. Rev. 24, 400 (1924); Phys. Rev. 27, 455 (1926); C.T. Zahn, J.B. Miles, Jr.,
Phys. Rev. 32, 497 (1928)

Chapter 16

168. A.D. Bruce, R.A. Cowley, Structural Phase Transitions (Taylor and Francis, London, 1981)
169. A.F. Devonshire, Philos. Mag. 40, 1040 (1949); Philos. Mag. 40, 1065 (1949)
170. L.D. Landau, E.M. Lifshitz, Theory of Elasticity (Pergamon Press, Oxford, 1959); 2nd edn.
(1970)
171. P. Toledano, M.M. Fejer, B.A. Auld, Phys. Rev. B 27, 5717 (1983)

Chapter 17

172. P.J. Collings, Liquid Crystals (Princeton University Press, Princeton, 1990)
173. P.G. de Gennes, J. Prost, The Physics of Liquid Crystals, 2nd edn. (Clarendon Press, Oxford,
1993)
174. I. Lelides, G. Durand, Phys. Rev. Lett. 73, 672 (1994)

Chapter 18

175. D.J. Fixsen, E.S. Cheng, J.M. Gales, J.C. Mather, R.A. Shafer, E.L. Wright, Astrophys. J.
473, 576 (1996)
176. M.F. Modest, Radiate Heat Transfer (McGraw-Hill, New York, 1992); 2nd edn. (Academic,
Amsterdam, 2003)
177. M. Planck, Verh. Dtsch. Phys. Ges. 2, 202 (1900)

Chapter 19

178. J. Heer, The principle of Le Chatelier and Braun. J. Chem. Educ. 34(8), 375–380 (1957)
179. L.K. Nash, Elements of Chemical Thermodynamics (Addison-Wesley, Reading, 1962)
180. J. Waser, Basic Chemical Thermodynamics (Benjamin, New York, 1966)
References 393

Tables e Handbooks

[EG] Encyclopedie de Gaz, L’Air Liquide, Division Scientifique (Elsevier, Amsterdam, 1976)
[AIP] D.E. Gray (coordinating editor), American Institute of Physics Handbook, 3rd edn.
(McGraw-Hill, New York, 1972)
[TT] J. Hilsenrath (ed.), Tables of Thermodynamic and Transport Properties of Air, Argon, Carbon
Dioxide, Carbon Monoxide, Hydrogen, Nitrogen, and Steam (Pergamon, New York, 1960)
[LB] Landolt-Börnstein, Zahlenwerte und Funktionen aus Physik, Chemie, Astronomie, Geophysik
und Technik, 6te Auflage (Springer, Berlin, 1950–1980); Neue Series (1961–1985)
[CRC] D.R. Lide (editor-in-chief), CRC Handbook of Chemistry and Physics, 78th edn. (CRC
Press, Boca Raton, FL, 1997–1998)
[ICT] E.W. Washburn (editor-in-chief), International Critical Tables (McGraw-Hill, New York,
1926–1930)
Index

A Chemical equation, 359–360, 362–363

Absolute temperature, 16 Chemical potential, 51, 59, 115, 153, 156,
Absolute zero, 89–92, 100 159, 164, 167, 172, 174–177,
Activity, 154, 157, 159, 165, 174–175, 183, 179, 180, 184, 194, 201–202,
376 226, 229, 232, 361, 362, 365,
Adiabatic line, 9 376, 377
Adiabatic process, 9 Chemical reaction, 359–366
Affinity, 361, 363, 365–366 Clapeyron diagram, 4
Ammonia, 360, 380, 381 Clausius integral, 18–19
Anisotropy Clausius-Mossotti relation, 303
magnetic, 260, 286–292 Coefficient
in solids, 315–316 compliance, 322
Antiferromagnetism, 273–280 elastic, 321–323
order parameter, 275–276 Grüneisen, 96
Argon, 139 Lamé, 326–328
Avogadro hypothesis, 22 magnetocaloric, 250, 369–370
Azeotropy, 188–190 pyroelectric, 300, 301, 314
stiffness, 321, 326
thermal expansion, 33–34, 78, 92, 96–99,
B 324–325, 368
Binary alloys, 223–227 thermochemical, 369
ordered, 223–225 Colligative properties, 160
order parameter, 225–226, 230 boiling point elevation, 169
Black body, 351 melting point depression, 162–163
Bohr magneton, 248 osmotic pressure, 160–161
Boiling, 116–117 Compressibility, 31, 34–35, 40, 77, 83–85, 92,
Brass, 225 99, 136–138, 142, 323–324, 368
Bulk modulus, 35, 324, 328 Compressibility factor, 129, 130, 134
Compression, 79
Conservation of energy
C in differential form, 9–11, 19–20
Calorie, 8–9 Constant
Carbon dioxide, 113–114, 128–129, Avogadro, 45
139–141 Boltzmann, 45–46
Celsius scale of temperature, 16 universal gas, 22, 45

© Springer-Verlag Berlin Heidelberg 2017 395

M.J. de Oliveira, Equilibrium Thermodynamics, Graduate Texts in Physics,
DOI 10.1007/978-3-662-53207-2
396 Index

Convex and concave functions Dieterici, 83, 88

multiple variables, 64–66 Euler, 51
single variable, 62–64 Gibbs-Duhem, 59, 172, 201
Convex hull, 66–68, 131 Gibbs-Duhem-Margules, 167
Convexity, 62–69 Margules, 178
Cooper, 99 van der Waals, 54–55, 70, 74, 83, 86, 87,
Critical behavior, 139–149 129–131
antiferromagnetic, 279, 280 van Laar, 178
ferromagnetic, 262, 263, 270 van.t’ Hoff (osmotic pressure), 161
liquid-vapor, 140–141 van.t’ Hoff (thermochemistry), 378–379
Critical exponent, 139–142, 262–264 Equation of state, 50, 54, 56, 57, 251–252
antiferromagnetic, 279–280 Equilibrium constant, 376–379, 381–382
ferromagnetic, 262–263 Exact differential, 73–74
liquid-liquid, 196 Extensivity, 51–55, 68
liquid-vapor, 140–141 Extent of reaction, 359–366
order-disorder in alloys, 231
Criticality, 127–149, 262–269, 279–280
Critical line, 208–209 F
Critical opalescence, 127 Ferrimagnetism, 292–294
Critical pressure, 130 Ferromagnetism, 244–246
Crystalline state, 112, 315 First law of thermodynamics, 11
Cycle Free expansion, 3, 79–80
Brayton-Joule, 28 Function
Carnot, 14–16, 25, 29 Brillouin, 252–253, 270
Diesel, 28 Landau, 213–214
Otto, 25–26 Langevin, 251–252, 270
Fundamental relation, 49–50, 152–154,
171–174
D
Dielectric, 295–314
Dielectric constant, 295, 302, 303 G
Distillation, 151 Gases, 82–87, 111
Gibbs free energy, 57, 66, 92, 156–157,
172–178, 194, 226–227, 231, 234,
E 237–240, 249, 253, 291, 294,
Effect 312–313, 321, 329–333, 338, 366,
electrocaloric, 314 374, 379
magnetocaloric, 269–270 Gibbs-Konovalov laws, 189
pyroelectric, 300 Grand thermodynamic potential, 57–58
Elastomer, 347
Electric field, 295–298, 342
Electromagnetic waves, 349–354 H
Enthalpy, 56–57, 65, 92, 158, 165, 300, 366, Heat, 7–9, 14–19, 24–25, 32, 35–36, 40, 41,
374 248, 350, 366–367
and heat, 60–61 Heat capacity, 31–33, 39, 41, 61, 76–77,
Entropy, 16–20, 35–37, 50–54, 62, 89–91, 84, 249–250, 300, 320–321, 368,
100–101, 228 371–372
Boltzmann formula, 45 Heat engine, 1, 24–25
microscopic interpretation, 44–46 efficiency, 25
mixture, 157–158 Heat of reaction, 366–367
and spontaneous process, 42–44 Helmholtz free energy, 55–56, 63, 65, 91, 153,
Equation 249, 254, 256, 257, 275, 277, 281,
Berthelot, 83, 88 300–301, 304, 305, 312, 320–321,
chemical equilibrium, 339–340, 365 353, 366–367, 373
Clausius-Clapeyron, 120–124, 205 and work, 60–61
Index 397

Hydrogen, 82, 86, 87 Raoult, 158–160, 175, 182

Hydrogen bonds, 108, 309 Stefan-Boltzmann, 351–352
Wien, 354
Legendre transformation, 55–61, 68
I and convexity, 63
Ice, 106–108, 113 Lewis assumptions, 159
melting temperature, 16, 113 Line
residual entropy, 106–108 boiling, 181–182
Ideal gas, 20–24, 53–54, 86–87, 101–106 buble, 181
adiabatic curve, 23 coexistence, 120–121, 128–129, 133, 142
coefficient of thermal expansion, 31–32 condensation, 181
compressibility, 34–35 dew, 181
entropy, 23–24 liquid, 190
equation of state, 20–21 solid, 190
heat capacity, 32, 86 tie, 116, 117, 122, 182
thermodynamic potentials, 60 Liquefaction, 128
Identity between derivatives, 75 Liquid, 111
Internal energy, 9–10, 34–36, 49–53, 62–64, Liquid crystal, 335–346
91, 172, 248, 299, 320, 337, 353, Lyotropics, 336
360, 365
Inversion curve, 85–86
Isentropic line, 17 M
Isobaric line, 5 Magnetic field, 247, 342–343
Isochoric line, 5 Magnetic ordering, 241–242, 273–294
Isotherm, 14, 17 Magnetization, 241–242, 245–246, 252, 255,
Isotopic effect, 310–311 265, 268, 275–276, 281, 283, 284,
286–287, 292
spontaneous, 241, 246, 255–256, 260–263,
J 266, 287, 292, 293
Joule procedure, 7, 9 volumetric, 246–248
Manifold, 203–204
critical, 207–208
K ordinary, 206–207
Kelvin scale of temperature, 16 Materials
antiferroelectric, 279–280
antiferromagnetic, 273–274, 290
L dielectric, 295–298
Latent heat, 118–119, 123–124 ferrimagnetic, 292–293
Law ferroelectric, 296–298
Boyle, 31 ferromagnetic, 241, 246
Curie, 243, 252, 302 magnetic, 241–246
Curie-Weiss, 244 Maxwell construction, 67, 131–133
Dalton, 155 Maxwell relations, 74–75, 269, 322, 367
Debye, 301–303 Mechanical equivalent of heat, 8
Dulong-Petit, 92 Melting, 123
Gay-Lussac, 20 Mesophases, 335–336
Guldberg-Waage, 377–378 Metals, 97–99
Henry, 158–160, 175, 183 Metamagnets, 281–286
Hess, 375 Miscibility, 152, 171, 193
Hooke, 325–326 Miscibility curve, 193
Kirchoff (termochemistry), 368 Mixture, 151–168
Kirchoff (thermal radiation), 350–351 binary, 171–199
mass action, 378 dilute, 174–175, 182–183
Planck, 354–357 entropy, 157–158
398 Index

heterogeneous, 152 first order, 114–126, 237–239

homogeneous, 152 liquid-liquid, 192–196
ideal, 175–176 liquid-vapor, 112–122, 127–128, 141, 181,
ideal gases, 155–158, 379–381 185–188
regular, 176–178 magnetic systems, 281–286
Molar heat capacity, 53, 61, 86–87, 91–93, nematic, 338–340, 344–346
97–98, 102, 104–105, 138, 142, 146, nematic-smectic, 344–346
197, 251, 254, 259, 263–264, 266, order-disorder, 223–240, 296–298,
279, 301 309–314, 329
Molar quantities, 51–53, 153–154, 173–174, second order, 236–237, 307
250–251, 301 solid-liquid, 112–113, 122, 190–192
Mole, 22 solid-solid, 197–198
Molecular field, 255, 276–277 solid-vapor, 113, 123–124
Mole fraction, 154–155, 178 structural, 298, 328–333
Moment of electric dipole, 295, 298 Point
Moment of magnetic dipole, 252 bicritical, 291–292
critical, 127–129, 133–134, 139, 196–197,
207–208, 216–217, 231, 265, 280
N ferromagnetic, 261
Nematics, 336–343 liquid-vapor, 127–128
coupling to electric field, 342–343 critical end point, 208–209, 212
coupling to magnetic field, 342–343 eutectic, 198–199
order parameter, 336–338 Landau, 340–342
Nickel, 241, 244, 245, 246, 264, 269 multicritical, 210
Nitrogen, 85, 86, 109 quadruple, 208
Number of moles, 22, 49 tetracritical, 292
tricritical, 209–210, 217–219, 282–284
triple, 112–113, 124–126
O Poisson ratio, 327
Order parameter, 225–226, 275–276, 329, Polarization, 295
336–338, 343 spontaneous, 296, 298, 304, 307–311
Osmotic pressure, 161 volumetric, 295, 298
Polymorphism, 328–329
Postulate
P Griffiths-Wheeler, 204
Paramagnetism, 242–244 Nernst, 89–92
ideal, 251–254 Planck, 100–108
Phase coexistence, 112, 115, 178–180, Potassium chloride, 93
184–185, 192–193, 197–198 Power law, 139
Phase diagram, 112–114, 201–221 Pressure, 3–4
Phase equilibria partial, 155
liquid-liquid, 192–193 Principle
liquid-vapor, 185–186 Carnot, 13–27, 42
solid-liquid, 190–192 statement, 14
solid-solid, 197–198 Clausius-Gibbs, 31–46, 62
Phase transition, 111–126, 178–183 expression, 38, 62
antiferromagnetic, 241–242, 273–279 conservation of energy, 9–11, 14
binary mixtures, 178–183 Joule, 1–11
continuous, 236–237, 307 statement, 10
discontinuous, 114–126, 237–239 Le Chatelier-Braun, 372
displacive, 296, 329 maximum entropy, 35–37, 42, 47
ferroelastic, 329–331 minimum energy, 37–38, 47
ferroelectric, 296–298, 304, 306–308 Nernst-Planck, 89–109
ferromagnetic, 241–242 statement, 100
Index 399

Process Solution, 152, 158–163

cyclic, 24–27 dilute, 158–163, 168
irreversible, 42 ideal, 163–165
isentropic, 17 regular, 168
isobaric, 6 regular of Hildebrand, 165
isochoric, 6 saturated, 152
isothermal, 13–14 solid, 184, 190
Joule-Thomson, 81–82 Solvent, 152
out of equilibrium, 4 Specific heat, 32, 93
quasi-static, 4 Spontaneous symmetry breaking, 234–236
quasi-static adiabatic, 9 Standard enthalpy of reaction, 374–375
spontaneous, 42–44, 373–374 States of matter, 111
thermochemical, 366, 374 Strain, 315–321
thermodynamic, 3–4 Stress, 315–321
Pure substance, 49–50, 111–114, 151–152 Sublimation, 113, 123, 124
compound, 112 Substances
simple, 112 completely miscible, 184–192
heterogeneous, 171
homogeneous, 171
R paramagnetic, 243, 253
Reduction of derivatives, 78 partially miscible, 192–199
Refrigerator, 1, 27 Sulfur, 91–92
performance coefficient, 27 Susceptibility
Relation between critical exponents electric, 300, 301, 342
Griffiths, 148, 264 magnetic, 249, 257–259, 280–281, 344
Rushbrooke, 148, 264 parallel, 287
Widom, 148, 264 perpendicular, 262, 287
Reservoir System(s)
mechanical, 35–37 closed, 361, 365, 369
thermal, 35–36 ideal paramagnetic, 251–254
Rule isolated, 373
Cailletet-Mathias, 141, 149 magnetic, 241–271
expression, 203 metamagnetic, 281–286
Gibbs phase rule, 201–205
Le Chatelier, 371
lever, 117, 118, 179, 193 T
van’t Hoff, 371, 378 Temperature
Zernike, 207 absolute, 14, 18
boiling, 116
characteristic, 102, 105
S compensation, 293, 294
Sackur-Tetrode formula, 103–104 consolute, 193
Second law of thermodynamics, 41–46 critical, 130, 196, 246, 263, 293
Clausius statement, 41–42 Curie, 241, 243
Kelvin statement, 41–42 Debye, 92, 94
Simplex, 204–205 Fermi, 98
Smectics, 344–346 melting, 123
order parameter, 344 Néel, 274, 280
Solids, 93–94, 96, 111, 223, 315–334 sublimation, 124
heat capacity, 92–99 transition, 297
isotropic, 326–328 triple point, 125
thermal expansion, 96–97 Tenso
Solubility. See Miscibility strain, 317–319
Solute, 152, 162 stress, 316–317
400 Index

Theory Third law of thermodynamics, 100

Bragg-Williams, 227–234 Tin, 90
Brillouin, 252–254
Debye, 9–96
Devonshire, 304–308, 331–333 U
Griffiths-Landau, 213–221 Universal function, 268–269
Hildebrand-Heitler, 194–196
Landau, 234–240, 285–286
V
Landau-de Gennes, 338–340
van der Waals fluid, 54, 129, 136
Néel, 276–278, 293–294
van der Waals gas. See van der Waals fluid
Slater-Takagi, 311–314
Vapor pressure, 115
van der Waals, 129–138
Variables
Weiss, 255–262
extensive, 51
Widom, 146–149, 266–268
intensive (see Thermodynamic field)
Thermal phenomena, 1
Velocity of sound, 84, 142
Thermal radiation, 349–357
Virial expansion, 82
Thermochemical equilibrium,
373–383
Thermochemical identity, 370–371 W
Thermochemistry, 359–383 Walls
Thermodynamic coefficient, 31–35, adiabatic, 3
249–250 diathermal, 3
Thermodynamic density, 52, 114–115 Water, 1
Thermodynamic equilibrium, 2–3, 180–181 boiling, 16, 112
Thermodynamic field, 51, 114–115, 201–203 critical point, 113
Thermodynamic identity, 73–77 density, 34
Thermodynamic phase, 112 entropy, 105
Thermodynamic potential, 49–69, 91, phase diagram, 113
320, 365 specific heat, 33, 108
electric, 299 triple point, 16, 113
fundamental properties, 68, 114–115 Work, 4–7, 24–26, 36, 40, 44, 60, 246–248,
magnetic, 230–234, 246–251 298, 319, 366
molar, 58–59
Thermodynamic space, 49, 202, 210
Thermodynamic stability, 33, 35–41 X
Thermodynamic state, 3, 49, 50, 152–153, Xenon, 94, 100
360–362
Thermometer, 13
Thermoscope, 13 Y
Thermotropics, 335–336 Young modulus, 327