Fundamentals of

and
statistical

thermal physics

F. REIF

Professor of Physics

University of Californiaf Berkeley

International Student Edition

McGRAW-HILL BOOK COMPANY

New York St. Louis San Francisco

London Mexico Panama Sydney Toronto

KOGAKUSHA COMPANY, LTD.

Fundamentals of statistical and thermal physics

INTERNATIONAL STUDENT EDITION

Exclusive rights by Kogakusha Co., Ltd., for manufacture

and export from Japan. This book cannot be re-exported
from the country to which it is consigned by Kogakusha
Co., Ltd., or by McGraw-Hill Book Company or any of
its subsidiaries.
I

Copyright © 1965 by McGraw-Hill, Inc. All rights reserved.

This book, or parts thereof, may not be reproduced in
any form without permission of the publishers. Library
of Congress Catalog Card Number 63-22730.

TOSHO PRINTING CO.. LTD., TOKYO, JAPAN

Preface

this book is devoted to a discussion of some of the basic physical concepts and
methods appropriate for the description of systems involving very many
particles. It is intended, in particular, to present the disciplines of
thermodynamics, statistical mechanics, and kinetic theory from a unified and modern
point of view. Accordingly, the presentation departs from the historical
development in which thermodynamics was the first of these disciplines to
arise as an independent subject. The history of the ideas concerned with heat,
work, and the kinetic theory of matter is interesting and instructive, but it
does not represent the clearest or most illuminating way of developing these
subjects. I have therefore abandoned the historical approach in favor of one
that emphasizes the essential unity of the subject matter and seeks to develop
physical insight by stressing the microscopic content of the theory.
Atoms and molecules are constructs so successfully established in modern
science that a nineteenth-century distrust of them seems both obsolete and
inappropriate. For this reason I havedeliberately chosen to base the entire
discussion on the premise that all macroscopic systems consist ultimately of
atoms obeying the laws of quantum mechanics. A combination of these
microscopic concepts with some statistical postulates then leads readily to
some very general conclusions on a purely macroscopic level of description.

These conclusions are valid irrespective of any particular models that might be
assumed about the nature or interactions of the particles in the systems under
consideration; they possess, therefore, the full generality of the classical laws
of thermodynamics. Indeed, they are more general, since they make clear
that the macroscopic parameters of a system are statistical in nature and
exhibit fluctuations which are calculable and observable under appropriate
conditions. Despite the microscopic point of departure, the book thus
contains much general reasoning on a purely macroscopic level—probably about
as much as a text on classical thermodynamics—but the microscopic content

of the macroscopic arguments remains clear at all stages. Furthermore, if one

is willing to adopt specific microscopic models concerning the particles consti-

vxi
vxxx PREFACE

tuting a system, then it is also apparent how one can calculate macroscopic
quantities on the basis of this microscopic information. Finally, the statistical
concepts used to discuss equilibrium situations constitute an appropriate
preparation for their extension to the discussion of systems which are not
in equilibrium.
This approach has, in my own teaching experience, proved to be no more
difficult than the customary one which begins with classical thermodynamics.
The latter subject, developed along purely macroscopic lines, is conceptually
far from easy. Its reasoning is often delicate and of a type which seems
unnatural to many physics students, and the significance of the fundamental
concept of entropy is very hard to grasp. I have chosen to forego the
subtleties of traditional arguments based on cleverly chosen cycles and to substitute
instead the task of assimilating some elementary statistical ideas. The
following gains are thereby achieved: (a) Instead of spending much time discussing
various arguments based on heat engines, one can introduce the student at an
early stage to statistical methods which are of great and recurring importance
throughout all of physics, (6) The microscopic approach yields much better
physical insight into many phenomena and leads to a ready appreciation of
the meaning of entropy, (c) Much of modern physics is concerned with the
explanation of macroscopic phenomena in terms of microscopic concepts. It
seems useful, therefore, to follow a presentation which stresses at all times the

interrelation between microscopic and macroscopic levels of description. The

traditional teaching of thermodynamics and statistical mechanics as distinct
subjects has often left students with their knowledge compartmentalized and
has also left them ill-prepared to accept newer ideas, such as spin temperature
or negative temperature, as legitimate and natural, (d) Since a unified
presentation is more economical, conceptually as well as in terms of time, it permits
one to discuss more material and some more modern topics.

The basic plan of the book is the following: The first chapter is designed to
introduce some basic probability concepts. Statistical ideas are then applied
to systems of particles in equilibrium so as to develop the basic notions of
statistical mechanics and to derive therefrom the purely macroscopic
general statements of thermodynamics. The macroscopic aspects of the
theory are then discussed and illustrated at some length; the same is then done
for the microscopic aspects of the theory. Some more complicated equilibrium
situations, such as phase transformations and quantum gases, are taken up
next. At this point the text turns to a discussion of nonequilibrium situations
and treats transport theory in dilute gases at varying levels of sophistication.
Finally, the last chapter deals with some general questions involving
irreversible processes and fluctuations. Several appendices contain mostly various
useful mathematical results.
The book is intended chiefly as a text for an introductory course in
statistical and thermal physics for college juniors or seniors. The mimeographed
notes on which it is based have been used in this way for more than two years
by myself and several of my colleagues in teaching such a course tat the
University of California in Berkeley. No prior knowledge of heat or thermo-
PREFACE IX

dynamics presupposed; the necessary prerequisites are only the equivalents

is
of a in introductory physics and of an elementary course in atomic
course

physics. The latter course is merely supposed to have given the student
sufficient background in modern physics (a) to know that quantum mechanics
describes systems in terms of quantum states and wave functions, (6) to have
encountered the energy levels of a simple harmonic oscillator and to have seen
the quantum description of a free particle in a box, and (c) to have heard of the
Heisenberg uncertainty and Pauli exclusion principles. These are essentially
all the quantum ideas that are needed.
The material included here is more than can be covered in a one-semester
undergraduate course. This was done purposely (a) to include a discussion of
those basic ideas likely to be most helpful in facilitating the student's later
access to advanced works, (b) to allow students with some curiosity to
more

read beyond the minimum on a given topic, (c) to give the instructor some
possibility of selecting between alternate topics, and (d) to anticipate current
revisions of the introductory physics course curriculum which should make
upper-division students in the near future much more sophisticated and better
prepared to handle advanced material than they are now. In actual practice
I have successfully covered the first 12 chapters (omitting Chapter 10 and most
starred sections) in a one-semester course. Chapter 1 contains a discussion of
probability concepts more extensive than is needed for the understanding of
subsequent chapters. In addition, the chapters are arranged in such a way
that it is readily possible, after the first eight chapters, to omit some chapters
in favor of others without encountering difficulties with prerequisites.
The book should also be suitable for use in an introductory graduate
course if one includes the starred sections and the last three chapters, which

contain somewhat more advanced material. Indeed, with students who have
studied classical thermodynamics but have had no significant exposure to the
ideas of statistical mechanics in their undergraduate career, one cannot hope to
cover in a one-semester graduate course appreciably more subject matter than

is treated here. One of my colleagues has thus used the material in our
Berkeley graduate course on statistical mechanics (a course which is, as yet,
mostly populated by students with this kind of preparation).
Throughout the book I have tried to keep the approach well-motivated and
to strive for simplicity of presentation. It has not been my aim to pursue rigor
in the formal mathematical sense. I have, however, attempted to keep
the basic physical ideas in the forefront and to discuss them with care. In the
process the book has become longer than it might have otherwise, for I have
not hesitated to increase the ratio of words to formulas, to give illustrative

examples, or to present several ways of looking at a question whenever I felt

that it would enhance understanding. My aim has been to stress physical
insight and important methods of reasoning, and I advise most earnestly that
the student stress these aspects of the subject instead of trying to memorize
various formulas meaningless in themselves. To avoid losing the reader in
irrelevant details, I have often refrained from presenting the most general case
of a problem and have sought instead to treat relatively simple cases by power-
X PREFACE

ful and easily generalizable methods. The book is not meant to be

encyclopaedic;merely intended to provide a basic skeleton of some fundamental
it is
ideas most likely to be useful to the student in his future work. Needless to
say, some choices had to be made. For example, I thought it important to
introduce the Boltzmann equation, but resisted the temptation of discussing
applications of the Onsager relations to various irreversible phenomena such
as thermoelectric effects.
It is helpful if a reader can distinguish material of secondary importance
from that which is essential to the main thread of the argument. Two devices
have been used to indicate subject matter of subsidiary importance: (a)
Sections markedby a star (asterisk) contain material which is more advanced or
more detailed; they can be omitted (and probably should be omitted in a first
reading) without incurring a handicap in proceeding to subsequent sections.
(b) Many remarks, examples, and elaborations are interspersed throughout the
text and are set off on a gray background. Conversely, black marginal
pointers have been used to emphasize important results and to facilitate
reference to them.
The book contains about 230 problems, which should be regarded as an

essential part of the text. It is indispensable that the student solve an

appreciable fraction of these problems if he is to gain a meaningful understanding of
the subject matter and not merely a casual hearsay acquaintance with it.
I am indebted to several of my colleagues for many valuable criticisms
and suggestions. In particular, I should like to thank Prof. Eyvind H.
Wichmann, who read an older version of the entire manuscript with meticulous
care, Prof. Owen Chamberlain, Prof. John J. Hopfield, Dr. Allan N. Kaufman,
and Dr. John M. Worlock. Needless to say, none of these people should be
blamed for the flaws of the final product.
Acknowledgements are also due to Mr. Roger F. Knacke for providing
the answers to the problems. Finally, I am particularly grateful to my
secretary, Miss Beverly West, without whose devotion and uncanny ability
to transform pages of utterly illegible handwriting into a perfectly typed
technical manuscript this book could never have been written.
It has been said that "an author never finishes a book, he merely abandons
it." I have come to appreciate vividly the truth of this statement and dread
to see the day when, looking at the manuscript in print, I am sure to realize
that so many things could have been done better and explained more clearly.
If I abandon the book nevertheless, it is in the modest hope that it may be
useful to others despite its shortcomings.

F. REIF
 Contents

Preface vii

Introduction to statistical methods

RANDOM WALK AND BINOMIAL DISTRIBUTION

1 1 Elementary statistical concepts and examples Qj

1 2 The simple random walk problem in one dimension
1 3 General discussion of mean values Qj)
1 4 Calculation of mean values for the random walk problem uS^
1 5
¦

Probability distribution for large N 17

1 6 Gaussian probability distributions 21

GENERAL DISCUSSION OF THE RANDOM WALK

1 7 Probability distributions involving several variables 26

1 8 Comments on continuous probability distributions 27
1 9 General calculation of mean values for the random walk 32
*1 10 Calculation of the probability distribution 36
*1 11 Probability distribution for large N 37

Statistical description of systems of particles

STATISTICAL FORMULATION OF THE MECHANICAL PROBLEM

2 1 Specification of the state of a system 48

22 Statistical ensemble 62

jfj^? Basic postulates 63

2 4 Probability calculations 60
2 5 Behavior of the density of states 61
 INTERACTION BETWEEN MACROSCOPIC SYSTEMS

2-6 Thermal interaction 66

2 7
-
Mechanical interaction 68
2 8
¦
General interaction 73
2 9
¦

Quasi~static processes 74
2 10 Quasi-static work done by pressure 76
2-11 Exact and "inexact" differentials 78

O Statistical thermodynamics 87

IRREVERSIBILITY AND THE ATTAINMENT OF EQUILIBRIUM

3 1
¦

Equilibrium conditions and constraints 87

3 2
-
Reversible and irreversible processes 91

THERMAL INTERACTION BETWEEN MACROSCOPIC SYSTEMS

3 3
¦
Distribution of energy between systems in equilibrium 94
3 4
¦
The approach to thermal equilibrium 100
3 5 Temperature 102
3-6 Heat reservoirs 106
3-7 Sharpness of the probability distribution 108

GENERAL INTERACTION BETWEEN MACROSCOPIC SYSTEMS

3 8
¦

Dependence of the density of states on the external parameters 112

3-9 Equilibrium between interacting systems 114
3 10 Properties of the entropy 117

SUMMARY OF FUNDAMENTAL RESULTS

3 11 Thermodynamic laws and basic statistical relations 122

3 12 Statistical calculation of thermodynamic quantities 124

4 Macroscopic parameters and their measurement 128

4 1 Work and internal energy 128

4-2 Heat 131
4-3 Absolute temperature 138
4 4 Heat capacity and specific heat 189
4-5 Entropy 14M
4 6 Consequences of the absolute definition of entropy 146
4-7 Extensive and intensive parameters 148

D Simple applications of macroscopic thermodynamics 152

PROPERTIES OF IDEAL GASES

5 1 Equation of state and internal energy 168

5 2 Specific heats 166
 5-3 Adiabatic expansion or compression 168
5-4 Entropy 160

GENERAL RELATIONS FOR A HOMOGENEOUS SUBSTANCE

5 5 Derivation of general relations 161

5 6
¦

Summary of MaxweU relations and thermodynamic functions 164

5-7 Specific heats 166
5-8 Entropy and internal energy 171

FREE EXPANSION AND THROTTLING PROCESSES

5-9 Freeexpansion of a gas 176

5 10 ThrotMng (or Joule-Thomson) process 178

HEAT ENGINES AND REFRIGERATORS

5-11 Heat engines 184

5 12 Refrigerators 190

Basic methods and results of statistical mechanics 201

ENSEMBLES REPRESENTATIVE OF SITUATIONS OF PHYSICAL INTEREST

6 1 Isolated system 201

6 2
¦

System in contact with a heat reservoir 202

6-3 Simple applications of the canonical distribution 206
6-4 System with specified mean energy 211
6 5 Calculation of mean values in a canonical ensemble
-

212
6-6 Connection with thermodynamics 214

APPROXIMATION METHODS

6-7 Ensembles used as approximations 219

*6-8 Mathematical approximation methods 221

GENERALIZATIONS AND ALTERNATIVE APPROACHES

*6 9 Grand canonical and other ensembles 226

*6-10 Alternative derivation of the canonical distribution 229

Simple applications of statistical mechanics 237

GENERAL METHOD OF APPROACH

7 1 ¦

Partition functions and their properties 2$7

IDEAL MONATOMIC GAS

7 2 Calculation of thermodynamic quantities 239

7 3 ¦

Gibbs paradox 243

7 4 ¦

Validity of the classical approximation 246