This fascinating book explains why the materials we can see and

touch behave in the way that they do. In a completely non-technical

style, using only basic arithmetic, the author explains how the
properties of materials result from the way they are composed of
atoms and why it is they have the properties they do: for example,
why copper and rubies are coloured, why metals conduct heat better
than glass does, why magnets attract an iron nail but not a brass pin,
and how superconductors are able to conduct electricity without
resistance.

The book is intended for general readers, and uses mainly words,
pictures and analogies, with only a minimum of very simple
mathematics. Even so, the author explains the basic ideas of
quantum mechanics, the laws .of which, although unfamiliar to most
people, govern the behaviour of all matter in the universe. He
explains how it is possible to understand the basic properties of
matter using these ideas, and translates the technical jargon of
physics into a language that can be understood by anyone with an
interest in science who wants to know why the world around us
behaves in the way that it does.

WHY THINGS ARE THE WAY THEY ARE

B. S. CHANDRASEKHAR

WHY

THINGS

ARE

THE
WAY

THEY

ARE

Cambridge

UNIVERSITY PRESS

PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY

OF CAMBRIDGE

The Pitt Building, Trumpington Street, Cambridge CB2 irp, United

Kingdom CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2ru, United Kingdom 40

West 20th Street, New York, ny 10011—4211, USA

10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1998

This book is in copyright. Subject to statutory exception and to the

provisions of relevant collective licensing agreements, no
reproduction of any part may take place without

the written permission of Cambridge University Press.

First published 1998

Printed in the United Kingdom at the University Press, Cambridge

Typeset in Monotype Columbus 10V2 on i3pt [wv]

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Chandrasekhar, B. S. (Bellur Sivaramiah), 1928—

Why things are the way they are / B.S. Chandrasekhar.

p. cm.

isbn o 521 45039 x (he). - isbn o 521 45660 6 (pbk.)

1. Condensed matter. 2. Quantum theory. I. Title.

QC173.454.C49 1998

53o.4’i-dc2i 96—52937 CIP

isbn o 521 45039 x hardback

isbn o 521 45660 6 paperback

CONTENTS
PREFACE IX
I INTRODUCTION I
II CRYSTALS 8

III PARTICLES AND WAVES 35

IV THE ATOM 63

V STATISTICAL PHYSICS 84

VI THE QUANTUM MECHANICAL CRYSTAL IO3

VII COPPER WIRES AND GLASS RODS I33

VIII SILVER SPOONS AND PLASTIC SPOONS 145

IX GLASS PANES AND ALUMINIUM FOILS 163

X ELECTRIC BULBS AND INSULATED CABLES l88

XI MAGNETS 2O2

XII SUPERCONDUCTORS 2l6

XIII CONCLUSION 235

GLOSSARY 238

INDEX 251
PREFACE
A friend once remarked ruefully to me, “Whenever I hear someone
talk about a well-rounded person, I think of a spherical object with no
distinguishing features on it.’ A closer examination of such a person
does reveal some features; these, however, are likely to be
concentrated in areas that are labelled literature, art, music,
philosophy, history and others generally called the liberal arts or the
humanities. The area called physics, on the other hand, usually
tends to be blank, although physics is as impressive an achievement
of the mind as any of the others. A reason for this state of affairs is
contained in the response I often get from someone who hears that I
am a physicist: ‘Physics was my most difficult subject in school. I
never understood it.’ Instead of giving up, I try to explain to the
person some problem I am working on at the moment, maybe in
superconductivity. I use words, analogies, pictures drawn on a paper
serviette: anything but mathematics that is more advanced than
simple arithmetic. I am then, even if not always, rewarded by the
person with a dawning of interest and a desire to learn more.

Having enjoyed this activity during all these years that I have myself
been learning physics through teaching and research, I decided to
gather and organize these ideas into a book for people (and I believe
there are many) who are curious about the physicist’s picture of the
world, especially the part which forms our immediate surroundings.

There are many books which attempt to explain to a general

readership certain parts of physics: stars and galaxies, quarks and
neutrinos, the nature of space and time, the big bang, and relativity,
to name a few. These involve ideas which have caught the public
imagination and appear often in the science sections of newspapers
and magazines. There is, however, a paucity of books which deal
with that part of physics which is around us all the time and impinges
on our daily lives, namely the physics of condensed matter. This is a
branch of physics that explains the rich variety of properties
displayed by the materials that make up the things that we can see,
touch and use in our daily lives. It is a beautiful synthesis of three of
the cornerstones of the subject, namely quantum mechanics,
statistical physics, and electromagnetism, and is therefore an
excellent means to discovering them. It is also the branch which
continues to have the most applications to the development of
objects used in everyday life: a fact which should please those who
seek relevance in the endeavours of the scientist. This book is the
result of my conviction that no one should be denied some
knowledge of this fascinating field merely because of not being a
physicist.

In writing this book, I have avoided certain features which are found
in some books on popular science for the lay reader. To follow the
line of reasoning in an unfamiliar field requires some effort, and I
have left out things which seem to me irrelevant or distracting. I have
drawn the figures so that they illustrate a point in the text, and
nothing IX
PREFACE
more. They are in effect the sketches on a serviette that I mentioned
earlier. The reader will find no photographs of physicists or of
equipment; it is not clear to me that looking at a picture of Enrico
Fermi helps one to understand what a fermion is, or that gazing upon
a photograph of tubes, cables, electronic instruments and assorted
hardware clarifies the photoelectric effect. I describe the subject as it
is today, and not its history. I name a few of the physicists who made
signal contributions, but do not try to be exhaustive. I do not identify
their nationalities or prizes won: physics is an endeavour which
knows no national boundaries, and is not some kind of Olympics. I
do not list books for further reading, because I think that readers
whose interest is aroused will do much better browsing in libraries
and bookshops and making their own discoveries.

Two rules guided me in the selection of topics for inclusion in the

book: they should show clearly how the basic concepts are applied,
and they should be no more in number than could be included in a
book of manageable size. In short, the book is lean, and intentionally
so.

Several friends in fields like languages, law, history, classics, theatre

and music, as well as physics, read parts of the manuscript and
expressed their opinions, for which I thank them. I wrote the book
while enjoying the hospitality and the environment of the Walther
Meissner Institute of the Bayerische Akademie der Wissenschaften
in Garching, Germany, and for this I am deeply grateful to its director,
Professor Klaus Andres. To my wife Dorothee, who has helped me in
more ways than she knows, I express my eternal gratitude.

B. S. CHANDRASEKHAR

GROBENZELL, GERMANY
INTRODUCTION
When you use a plastic spoon to stir hot coffee, the handle does not
get hot, but if you use a silver spoon, it does. A sheet of glass is
transparent so that you can see right through it, but a much thinner
sheet of aluminium foil is opaque and lets no light through. You can
bend a piece of copper wire back and forth without breaking it, but a
glass rod is brittle and will break if you try to bend it. A magnet will
attract an iron screw, but not a brass screw. A piece of copper is
reddish in colour, while silver is -

well, silvery. The electric current flowing through the tungsten

filament of a light bulb makes it so hot that it gives off light. This
current also flows through the connecting copper wires but does not
make them even warm, and it cannot flow at all through the rubber
insulation that covers the wires. The filament, the wires, and the
insulation behave quite differently from one another as far as the
electric current is concerned.

These examples illustrate the basic point I want to make: we are in

our everyday life surrounded by objects that display a wide variety of
properties. All these objects are made up of tiny things called atoms.
A piece of silver is composed of silver atoms that are all exactly alike
but are different from the atoms of, say, copper. Each atom consists
of a nucleus (plural: nuclei) which is surrounded by a certain number
of electrons. This number for silver is different from the one for
copper, or for any other element. Each nucleus itself is made up of
two kinds of even tinier things called protons and neutrons.

The number of protons equals the number of electrons in the atom,

and this number determines which element it is.

Thus all the different atoms, one for each of the approximately one
hundred elements that exist in the universe, consist of just three
kinds of objects: electrons, protons and neutrons. Yet when they are
put together to give us the things we see around us, the result is a
great range of different properties. One may well wonder how such
complexity can result from putting together atoms composed of just
three simple entities: electrons, protons and neutrons. Indeed, we
shall see later that we need to consider only two entities, namely
electrons and nuclei, to understand such properties of matter. That a
nucleus is itself composed of protons and neutrons will turn out to be
largely irrelevant for this purpose, though it is of crucial importance
when we want to understand specifically nuclear properties such as
radioactivity and nuclear fission.

Consider the silver spoon. It is made up of a very large number of

silver atoms that must be stuck fast to one another, because the
spoon retains its shape and does not collapse into a pile of free
atoms. What is the ‘glue’ that holds the atoms together? Each atom
is a piece of matter and weighs something, however little: it has
mass. There is therefore a force of gravitation between the atoms
which certainly pulls them together, INTRODUCTION

just as a dropped ball is pulled towards the earth by gravitation. Is

this force strong enough to hold the atoms together in the spoon? A
child, or an adult for that matter, playing on the seashore knows how
easy it is to start a mini-avalanche in a sandpile.

The gravitational attraction between the grains of sand, which

certainly is there, is not enough to maintain the shape of the pile
when it is disturbed. One can conclude that the gravitational force
between the atoms will not hold them together in the spoon.

There must be some other force between the atoms that has nothing
to do with their mass, and this force must lead to the rigidity of the
solid. We shall see later that electrons and nuclei have a property
called electric charge, or sometimes simply charge, and this is the
source of the force that we are seeking.

In the past several decades, one of the exciting challenges in

physics has been to understand how the properties of solids can be
related to the way they are made up of atoms. This subject used to
be, and sometimes still is, called solid state physics. A new name
has come into use in recent years: condensed matter physics. The
phrase ‘condensed matter’ means solids as well as liquids, both of
which have atoms rather densely packed together. This similarity
leads to some shared features of the way we can understand the
properties of solids and of liquids. This branch of physics is not only
one of the great intellectual achievements of the twentieth century; it
has also led to practical applications that reach into many aspects of
our everyday life. A famous example is the transistor. This is a
device, made from a small piece of germanium or silicon, which
began to replace valves (also called vacuum tubes) in radio
receivers in the 1950s. A direct descendant of the transistor is
today’s chip, which has found its way into countless household items
and tools of industry, and is at the heart of the modern computer.

So pervasive is the chip now, that one can give the old saying a new
meaning: ‘Chips with everything!’ Some other examples of practical
applications which grew out of the understanding of the physics of
condensed matter are: new magnetic materials used in electric
motors, glass fibres carrying telephone conversations, liquid crystals
and light-emitting diodes which display the numbers in pocket
calculators, superconducting magnets for magnetic imaging systems
which are used in medical diagnosis, high-strength alloys for jet
engines.

I describe in the following chapters the atomic picture of solids which

physics gives us. Some of you may have found that, if you ask a
physicist-friend what she (or he, as the case may be) does, you are
sometimes put off with the reply that you will not understand it if you
do not know advanced mathematics. This attitude is, I think, not
justifiable. It is true that a mastery of the subject in all of its intricate
detail requires some heavy mathematics. I assume that you have no
wish to acquire such mastery if you do not already have it. Rather,
you would like to know the basic features of what physics tells us
about the nature of solids. It is possible to describe these features
using only words and pictures along with some arithmetic, and it is
what I do in this book. After you have read it, you will have a picture
of what a physicist sees in a piece of solid matter, and why it has the
properties it does.

There will be no complicated mathematics in this book. I use only

simple arithmetic, INTRODUCTION

and sometimes symbols to denote numerical quantities, e.g. m to

denote the mass (expressed in grams) of some object, x to denote a
distance (in centimetres). Just as I add, subtract, multiply and divide
with numbers, I do the same with symbols. I use the forms shown
and explained below, for two numbers denoted by xandj: x+y means
‘add xandjy’,

x-y means ‘subtract y from x,

xXy (or xy) means ‘multiply x byy,

x/y (or ^means ‘divide xbyy.

I can go one further step and denote the result of each of the above
operations also by a symbol. For example, I could write the equation

x +y = z.

Here, if x equals 2 andy equals 3, then z equals 5. If x is 23.51

andjyis 12.34, then z is 35.85. If I give any particular values to x and
y, then the value of z is fixed. This single equation stands for every
one of the separate operations of adding all possible pairs of
numbers. The number of such operations is infinite, so that the
equation is a very convenient shorthand indeed.

I give now another example of using such symbols, which illustrates

what you will meet now and then in the book. An important property
of a moving object is its speed.

If the speed is constant, then I get the distance travelled by

multiplying the speed by the time of travel. Using symbols, I write

x = vxt
where x stands for the distance, v for the speed and t for the time of
travel. This equation is read ‘x equals v times f’. If the object
happens to be moving at 106 centimetres per second, then I get the
distance it travels in 2 seconds, say, by multiplying its speed,
106cm/sec, by 2 sec. The equation applies equally well for any
speed and any time of travel. You can see that the equation with its
symbols gives us a convenient shorthand for the general statement,
‘The distance travelled by any object which is moving at a constant
speed is given by multiplying its speed by its time of travel.’

There is one other matter of notation that I have just used, when I
wrote a number as 106, that may not be familiar to all of you. It is
called the exponential (and when used in pocket calculators, the
scientific) notation, and is useful in writing down numbers that are
either very large or very small. For example, the number of atoms in
one gram of a solid is about

1,000,000,000,000,000,000,000.

The width of an atom is about

100,000,000 centimetre.

The actual numbers will depend on what kind of atom it is, since the
atoms of different INTRODUCTION

matter, quantum mechanics should be able to explain their

characteristics as completely as it does those of inanimate things. A
very complicated matter, this, and the subject of much debate
involving not only scientists but also philosophers and theologians.
The question lies rather outside the scope of this book, and so I shall
not go further into it.

In chapter IV, I apply the ideas of chapter III to an atom and find that
quantum mechanics tells us that the atom is not at all what we had
imagined it to be, namely a tiny hard ball. The philosopher
Democritus said a long time ago that if one took a piece of matter
and divided it into smaller and still smaller pieces, one would finally
get to a piece that was so small that it could not be divided any
further, and therefore he called it an atom. One might then think that
an atom of, say, silver would be a tiny speck of silver with a silvery
sheen, quite different from a copper atom which would be reddish in
colour, or a glittering atom of gold. The reality turns out to be quite
different, as we shall see in chapter IV.

A piece of solid weighing a gram has about 1021 atoms, which is a

very large number indeed. There is no way in which we can hope to
describe this piece in terms of what each atom in it is doing.
Consider this: astrophysicists tell us that the big bang, which was the
beginning of our universe, took place 1010 years ago. That is about
3 X1017 seconds. So, even if a computer existed which took only
10”4 seconds to calculate and record the state of each atom, it would
need 1021x 10”4= 1017 seconds to complete the job - about a third
of the life of the universe. In fact, there is no good reason for wanting
such detailed information, other than perhaps the one given by a
climber for wanting to climb Mount Everest: ‘Because it is there.’
Clearly we need some other way of dealing with such large numbers
of atoms. Such a way exists, and you have already seen it applied
to, say, the characteristics of the population of a country.

Suppose I want to talk about the individual ages of the 248 million
inhabitants of the United States. I do not even begin the hopeless
task of listing the age of each person, all 248 million of them. Rather,
I make a list of the numbers of people whose ages lie between
specified limits, say 0—9 years, 10—19 years, and so on. I now
have a compact description of the age distribution of the population,
and have taken the first step into the subject called statistics. When
we use this method in physics, we call it statistical physics. I shall
introduce you to this topic in chapter V. In chapter VI, I replace the
hard-sphere atom of chapter II with the quantum-mechanical atom of
chapter IV, apply the statistical ideas of chapter V, and discover what
happens to the electrons when the atoms find themselves in the
solid. By doing so, I arrive at a picture of what a solid is like. It will
look very different from the orderly stack of spherical atoms which I
had in chapter II. I find that this new picture gives an explanation of
why the solid holds together and does not disintegrate into a pile of
atoms.

I compared earlier the passage through this book to a mountain

ramble. When you reach the end of chapter VI, you will have
reached an altitude with many views. You will be able to enjoy these
views in the succeeding chapters, where the quantum-mechanical
description that has been developed is used to explain the various
properties of solids mentioned earlier. If we both complete the
ramble successfully, you will INTRODUCTION

be able to answer the questions with which this chapter opened, and
many more. You will have a new way of looking at the world around
you, a way which has its own beauty and which adds to the world-
views of art, music, and literature.

You will find, as you read on, questions interspersed in the text. You
can decide whether or not to try to answer each question first before
looking at my answer which follows. In any event, please do not skip
over the questions; they do form an integral part of the whole story.

The last section in each of the following chapters summarizes the

chapter. You could read this section before beginning the chapter to
get a general idea of what it contains, or read it at the end to help
you put together what you have just read in the chapter.
Alternatively, you could read it at the beginning as well as at the end
to enjoy both benefits.

II CRYSTALS

formation. They would then form a single orderly pattern of rows and
columns. If on the other hand I divide them into ten units, each of
one hundred soldiers, and then ask them to fall into parade formation
at ten different points on the parade ground, but do not tell them in
which direction they should face, I would end up with a different
result. It is true that each unit will look exactly like every other unit,
but while all the soldiers in a given unit will be facing in the same
direction, this direction is very likely to be different for the different
units. Something like this is what happens when the liquid is cooled
rapidly, leading to a polycrystalline solid. It is possible to assign an
‘orientation’ to each grain in such a solid, somewhat like calling the
direction in which the soldiers of a given unit are facing as the
‘orientation’ of the unit. If we look at the ordered pattern of atoms in
the whole crystal, we find that the order will be different in different
directions, analogous to what is illustrated in fig. I-1. It shows a tree
nursery, with the seedlings planted in a regular pattern. You can see
that the plants are spaced apart by different distances in different
directions, so that it is possible to distinguish one direction from
another.

Finally we come to the splatt-cooled solid. The atoms in the liquid

have no order in their positions at any given instant of time. The
positions however are changing continuously with time as the atoms
slide past one another. When splatt-cooling takes place, the
temperature drops so fast that the atoms are stopped dead in the
last positions they had in the liquid, which were not in any orderly
arrangement. The result is a glassy solid.

I can also look in another way at what the difference is which causes
the liquid to end up as a crystal or as a glass: namely in terms of
what the atoms are doing as time goes on in the two cases. If I do
the freezing slowly enough, the atoms have enough time to find their
way to the positions in the solid where they would most like to be, so
to speak, and the result is a crystal. If on the other hand I speed up
the process sufficiently, the atoms just do not have enough time to
find their way to their preferred crystalline arrangement. So they
freeze into the disordered arrangement of a glass. It is known from
actual experience that, given a long enough time, the atoms in the
glassy solid will slowly rearrange themselves eventually into an
ordered crystalline arrangement. One possible crystalline form of
ordinary glass is quartz. The glass in some archaeological objects,
known to be many centuries old, is seen to have transformed to
quartz, which is the crystalline form of glass. This process is known
as devitrification. It appears therefore that the solid, given enough
time, always ends up being crystalline.

So now the question arises, why do the atoms prefer to be in an

ordered crystalline arrangement? We shall find the answer in the
next section.
2 The crystalline arrangement of
atoms
I have said that when the liquid is cooled slowly to form a solid, the
atoms end up in their preferred positions, which is the orderly
arrangement of a crystal. There are two features which characterise
this arrangement.

12

THE CRYSTALLINE ARRANGEMENT OF ATOMS

The first of these follows from the fact that the solid is stable: it does
not spontaneously change its shape. This means that the atoms
continue to stay where they are with respect to one another, and do
not move away from these positions: the net force on each atom is
zero. If I take a rod and pull on its two ends, it will get a bit longer; if I
push, it will get shorter. The actual change in size is perhaps invisible
to the naked eye, but it has been measured with special instruments.
In either case, the rod returns to its original shape when I let go. If I
describe this behaviour in terms of what the atoms are doing, I would
say that when the rod is stretched the atoms move a bit further apart,
and a bit closer together when it is compressed. An ordinary coil
spring becomes shorter when compressed, longer when stretched,
and in each case returns to its original length when released. I can
explain the behaviour of the rod by imagining that there are springs
connecting the atoms to one another. There are of course no real
springs in the solid; it is just that the influence of the atoms on one
another due to being made up of charged nuclei and electrons
produces the same effect, and we might talk about virtual springs
connecting the atoms to one another. This property of a solid, of
recovering its shape after being stretched or compressed, is called
its elasticity.
question But what about a copper wire? When I bend it, it remains
bent and does not return to its original shape. This must surely mean
that at least some of the atoms have moved from their original
positions.

answer The reason the wire remains bent is that the grains in the
wire are not perfect crystals, but have small imperfections which
permit some atoms to move about when the wire is bent. We shall
learn about these imperfections in chapter VII, and see how they can
explain the bending of a copper wire. These imperfections have very
little effect on many of the properties of solids that we shall be
Considering.

The second feature of the crystal comes from the fact that all atoms
of silver are identical. More generally, all atoms of any element are
identical, but different from atoms of any other element. Now
suppose that I place myself in the position of a particular atom in the
crystal, and look at the atoms around me. They will be distributed in
their preferred pattern at various distances in various directions
around me. Now I move to another atom in the crystal. Since all the
atoms are identical, the preferred pattern of atoms around this
second atom should be exactly the same as around the first atom.
Whatever factors led to the preferred arrangement around the first
atom should hold equally for the second atom and lead to the same
arrangement. The crystal, on an atomic scale, must look exactly the
same from the viewpoint of every atom in it. The universe around us
seems to have a similar property. When considered on a large scale
that includes enough stars, galaxies and so on, the universe looks
the same from one point in it as from any other point. Astronomers
call this the cosmologicalprinciple. The atoms in a crystal satisfy their
own version of the cosmological principle. There is nothing mystical
about this principle as it applies to solids. It is merely a consequence
of the experimentally established fact that all atoms of the same
element are identical.

II CRYSTALS
So we now have the two general characteristics of a crystal: the
atoms in it are as if connected to one another by virtual springs, and
their positions satisfy the cosmological principle. When we put atoms
together to meet these two requirements, we shall end up with the
ordered atomic arrangement in a crystal. We shall build the crystal in
three stages, for this makes it easy to satisfy the two requirements.
We shall first build a straight line of identical atoms, then assemble a
number of identical lines of atoms to give a plane of atoms. Finally
we shall stack identical planes of atoms above one another to give a
crystal, always making sure at each stage that the two requirements
are met.
3 The one-dimensional crystal
I now introduce an object that is very useful when trying to
understand nature, even though it does not exist in nature. This is a
one-dimensional crystal, consisting of a long row of atoms sitting
equally spaced in a straight line. A straight line has only length, and
zero width and zero thickness. Suppose that I measure distances
along the line from one of its ends. Then to tell you where any point
on the line is, I need to give you only its distance from that end. That
means that just one number, namely this distance, is enough to tell
you exactly where that point is. A ‘point’ in this context is a thing that
has a position in space, but has vanishingly small length, width and
thickness. Since we need only one number to tell us where a
particular point is on a straight line, we say that it is in one-
dimensional space. We assume that the atom is a point with a mass
equal to the mass of the atom, and that the influence of a
neighbouring atom on a given atom is as if a spring connects the
two.

I show in fig. II-5 three stages labelled A, B, and C in building up a

one-dimensional crystal. In stage A, I set down atom 1 and put atom
2 on its right at a distance a. Since 1

-o o o-
531
B
FIGURE II-5. Three successive stages, @ q q ||_

labelled A, B, and C, in building up a r-i r q i

one-dimensional lattice of atoms. The z^

distance between adjacent atoms is a. *s

THE ONE-DIMENSIONAL CRYSTAL

atom 2 now has an atom on its left, so must atom 1 and at the same
distance. So I put down atom 3 to the left of 1 at a distance a from it.
Now atom 1 has an atom on each side, and therefore so must atoms
2 and 3, and this is satisfied by atoms 4 and 5 in stage B. Continued
application of the cosmological principle will lead to stage C and so
on, so that we get a long straight line of equally spaced atoms, with
neighbouring atoms connected by virtual springs of length a and no
net force on each atom.

We can now deduce some features of the virtual springs from the
properties of this one-dimensional crystal. We assume that it has its
undisturbed length, i.e. it can be stretched or compressed, but it will
return to its original length afterwards. This clearly means that the
undisturbed length of each spring is a, the spacing between the
atoms.

Since the virtual springs are just a way of expressing the influence of
one atom on another, and as the atoms are identical, all the springs
must also be identical in their strength. I mean by the strength of a
spring how hard it is to press or to stretch it.
question Will the length of the springs be the same for different kinds
of atoms, say silver and copper?

answer No. The length and the strength of the spring will be
determined by the detailed structure (in terms of nuclei and
electrons) of the two atoms it connects, because this is what
determines how one influences the other.

question The lattice must have two ends somewhere, and the atoms
near there will certainly not satisfy the cosmological principle. Is this
not a problem?

answer The influence of one atom on another weakens rapidly as

their distance apart increases, so that a given atom can feel the
effect of only atoms that are a few times the distance a. from it, and
is blissfully unaware of atoms further away. So it suffices if the
cosmological principle applies for a few neighbours surrounding
each atom. The condition will then hold for all the atoms except a
few at the ends. We shall assume that the total number of atoms is
very large compared to the small number at the ends, so that we can
in practice say that all the atoms satisfy the cosmological principle.

In any crystal, there are some atoms at its boundaries. If the crystal
is one-dimensional, they are at its two ends. They are at the edges
of a two-dimensional crystal, and on the surfaces of a three-
dimensional crystal. We shall assume in everything that follows that
we are looking at the properties in a region of a crystal far enough
away from its boundaries so that effects due to the atoms at those
boundaries can be neglected. Such properties are called bulk
properties, to distinguish them from surface properties which are
determined by the atoms at the boundaries.

The distance a between adjacent atoms is called the lattice spacing

of the crystal. I have introduced here a new term, lattice. When I am
interested only in the relative positions of the atoms, but not in a
specific value of a or what kind of atoms they are, I call it a lattice.
II CRYSTALS
4 Two-dimensional lattices
The one-dimensional lattice is a suitable building-block with which to
assemble a two-dimensional lattice, as it already satisfies the
cosmological principle and has the virtual springs built into it. So I
make myself a large number of identical one-dimensional lattices,
each with spacing a. I label these lattices 1, 2, 3, and so on. I want to
assemble these on a flat tabletop to give me an object that has both
length and breadth, and in such a way that the cosmological
condition continues to be satisfied. I begin by setting down lattice 1
on the plane. Then I place lattices 2 and 3 on either side of lattice 1,
parallel to and at a distance b from it (fig. II-6, shaded portion). I
imagine that corresponding atoms between the lattices are
connected by virtual springs of length b.

Further, I make each of the two angles that I have marked on the
diagram equal to 90

degrees. We shall see later that these angles can also be different
from 90 degrees, and then one would get a different lattice.

question Why is the distance b taken to be different from the

distance a?

All the atoms are identical and so one might expect the virtual
springs to have the same length.

answer Remember that the virtual springs stand for the mutual effect
of the electrons and nuclei of the neighbouring atoms on one
another. The electrons are distributed differently in different
directions around the 6 °

2
 ©

Q® @

P^fl R

90°

3 ‘-

FIGURE H-6. The stages in building

up a two-dimensional lattice,

beginning with the three one-dimensional lattices labelled 1, 2 and

3 (shaded portion).

#a

lis

il

16
TWO-DIMENSIONAL LATTICES

nucleus. This can result in the virtual spring between P and

Q_having a different length and strength from the one between P
and R. It is possible as a special case to have a and b be the same,
as we shall see later.

You can see in the shaded part of fig. II-6 that each atom in lattice 1
has eight near neighbours around it, arranged in exactly the same
pattern. Thus the cosmological principle holds for the atoms in lattice
1, but not for the atoms in lattices 2 and 3 which have only five
nearest neighbours each. However, we know how to solve this
problem, from our experience in building the one-dimensional lattice.
We add pairs of lattices 4 and 5, 6 and 7 (the rest of fig. II-6), and so
on. The arrangement of atoms around each atom is now exactly the
same, and there is no net force on it; we have constructed a two-
dimensional crystal. We can completely specify the crystal by giving
the kind of atom (silver, copper, etc.), the lengths a and b, and the
angle of 90 degrees. We call a and b the lattice parameters. The
numbers, of atoms in each row and of rows, only specify the size of
the lattice, and have nothing to do with how the atoms are arranged.

We could now go on to the next stage, and complete the

construction of a solid crystal by stacking a large number of two-
dimensional crystals on top of one another in such a way that each
atom still feels no force from its neighbours. Before doing that, I
should like us to take a closer look at the two-dimensional crystal
that we have before us. We shall see that such a crystal has some
beautiful and important features which carry over, mutatis mutandis,
into the solid crystal. To do so, we shall be using pictures.

We can make an exact picture of the two-dimensional crystal on a

page, and see those features directly. It is not so easy to do so with
a picture of a solid crystal: that is, unless your imagination can leap
easily from the flat page into the third dimension.

We use the term lattice when we are interested only in the relative
positions of the atoms, but not in their kind. So the lattice is specified
by the lengths a and b being unequal, and the angle of 90 degrees.
What we have just finished constructing is a rectangular lattice. The
whole lattice can be broken up into a lot of rectangles, each with
edges a and b.

Is the rectangular lattice the only one I can have on a plane, just as
there was only one lattice I could construct on a straight line? No,
because I can change the angle and the ratio of a to b, which
determine the lattice. I denote the angle by the Greek letter cc
(alpha). I can now produce four more lattices, to give me a total of
five. Each of them shows order, but in a way which is different from
the others. I list below the five lattices with their names.

1. Simple rectangular lattice, a and b are different, a is 90 degrees.

2. Centred rectangular lattice. I get this by adding an additional atom

at the exact centre of each rectangle formed by four atoms from
adjacent rows of the rectangular lattice.

3. Square lattice, a and b axe. equal, and I use the same symbol a
for both of them.

The angle a is again 90 degrees.

4. Triangular lattice, a and b are again equal, but a is now 60

degrees.

II CRYSTALS

5. Oblique lattice, a and b are unequal, and a is some angle which is

different from 90 degrees.

Figure II-7 shows the positions of the atoms in the basic unit for each
of these lattices.

This unit is called the unit cell of the corresponding lattice. It is the
smallest part of the o
o

,90°

(2)

■o

FIGURE n-7. Unit cells of the five

two-dimensional lattices: (1) simple

rectangular; (2) centred rectangular;

(3) square; (4) triangular; and (5)

oblique.

o-

,90°

(3)

-o
-6

(5)

18

ORDER AND SYMMETRY

lattice which completely specifies what the whole lattice looks like. I
can build up the lattice by assembling the unit cells rather like tiles
on a floor. Each atom at the edge or the corner of a unit cell is then
shared with the immediately adjacent unit cell or cells.

question Why are there no more two-dimensional lattices?

answer These five are the only ones possible which satisfy the
cosmological principle.

What distinguishes each of these lattices is its symmetry, a term that

I shall explain in the next section. If I try to make another lattice that
seems to look different from any of these, it will turn out that it is just
one of these, but with changed values of the lattice parameters and
sometimes the angle. For example, if I add an atom at the centre of
each little square in the square lattice, I again get a square lattice,
but with a different lattice parameter, as you can verify by drawing a
picture.

question What if the added atom is not at the centre of each square?

answer The arrangement of atoms that results does not satisfy the
cosmological principle, because the surroundings of each added
atom are different from those of each atom in the original lattice.

You may amuse yourself by seeing what happens if you add an atom
to the centre of each little triangle in the triangular lattice. By drawing
a picture, you will find that the result is again a triangular lattice, but
with a smaller lattice parameter.
5 Order and symmetry
I have been talking about the order in a crystal or a lattice, implicitly
appealing to your innate sense of what the term means. However,
you might well say that order lies in the eyes of the beholder. To take
an example from the world of hearing rather than of sight, there are
those who find the music of Johann Sebastian Bach highly ordered,
and that of John Cage somewhat less so. But there are others who
detect much order in Cage’s music too. Such differences of
perception contribute to the richness of the world of art, but could
lead to undesirable confusion in the world of physics. We need to
make more precise what we mean by the order in a crystal. We do
so by introducing the idea of symmetry, an idea which plays a central
role in physics.

In our everyday life, we think of some objects as having a certain

symmetry to them, and others not. For example, suppose I look at
two objects on a table, a simple vase and an irregular lump of clay
as in fig. II-8, part (a). I shall get to the vase under the table later. I
say that the vase on the table has symmetry, but the lump has none.
The following is what I mean by this: suppose I look away from the
objects, and you rotate both the vase and the lump without moving
them from their places on the table. The result is shown in fig. II-8,
part (b). When I look at them again, the lump will look different, but
the vase will look exactly the same as before. I assume here that the
vase is absolutely II CRYSTALS

FIGURE n-8. To illustrate the idea of

symmetry. Part (<z) is before, and part

(b) is after, the vases and the lump of

clay have been rotated as described in

the text. The vases have symmetry,

the lump has none, under these

operations.

before

(b)

after

perfect, with no distinguishing marks or patterns on it, and that the

lump is — well, a lump. There is only one axis of rotation for the vase
which will lead to the above result.

If you had rotated it by turning it upside down, it would look different

when I look at it again. I summarize all this in the following
statement: The vase has symmetry for rotation by any angle about
one particular axis, and the lump does not.

20

SYMMETRIES IN A TWO-DIMENSIONAL LATTICE

I can explain the term axis of rotation by an example: the axis of

rotation of the earth is the straight line joining its north and south
poles.

I consider another example: a one-dimensional lattice. Look at a

section of the lattice far from its ends, and imagine that the entire
lattice is moved a distance along its length equal to the lattice
spacing a, so that each point is now in the position where its
neighbour was before. The lattice looks exactly as it was before. I
express this property of the lattice as follows:

A one-dimensional lattice has symmetry for a translation of one

lattice spacing along its length.
I mean by the term translation the moving of a body from here to
there without rotating it.

As a final example, consider the vase under the table, with a pair of
identical handles on opposite sides (fig. II-8). This vase has an axis
of rotational symmetry only for half of a full rotation, namely through
an angle of 180 degrees. A rotation through a further 180 degrees
brings the vase back to its starting position. Thus there are two
positions, each reached from the other by the same angle of rotation,
in which the vase looks exactly the same. I describe this property as
follows: A vase with a pair of identical handles on opposite sides has
a two-fold axis of rotational symmetry.

question Can you see a plane of reflection symmetry in this vase?

answer Yes, a plane passing through the axis and perpendicular to

the plane joining the two handles.

These examples suggest a precise way in which we can talk about

the symmetry of a given object. First, we perform a specific operation
on it. This operation can be a translation, a rotation or a reflection. If
at the end of the operation the object looks exactly the way it did
before, then we say that the object is symmetric under that specific
operation. This is the sense in which we shall talk about the
symmetries of lattices and crystals. When we say that the crystal has
order, we mean that it has certain symmetries.
6 Symmetries in a two-dimensional
lattice
We now look at the symmetries of the lattice, a portion of which is
shown in part (a) of fig. II-9. Imagine identical atoms sitting at each
lattice point. We shall first consider its translational symmetry. I move
the lattice to the right, such that the atom marked A finds itself in the
position where the atom marked B was, and B moves to where C
was, and so on. The crystal now looks as it did before the
translation. We have thus found a translational symmetry of the
lattice, but it is only one of many. For example, if I make a translation
which moves atom A to the position of atom P, then each other atom
moves to the position of another atom which was in the same
relative position to it as P

was to A. This is again a translational symmetry, but different from

the first one, because the size of the displacement and its direction
are different. A third translational II CRYSTALS

FIGURE II-9. A two-dimensional

lattice. Part (a) illustrates its symmetry

under translations which take A to B,

or to P, or to D. How two other

translations look when A goes to D

are shown by lines connecting the

relevant lattice points. Part (b) shows

how the same lattice can be

constructed from different sets of lines

of atoms, as indicated by the dotted

lines and dashed lines. The directions

AB and AC are close-packed, and AD

is not.

(a)

©
©6

(b)

.©■’

©■■’

© ©■’

©■’ ©

©•■’ p’

00
00
©” 0

AB

symmetry is one that takes atom A to the position of atom D. In fact,

any translation that takes a given atom to the position of any other
atom is a symmetry operation for the lattice. We can summarize:

Two-dimensional lattices have translational symmetry for any

translation that takes a lattice point to the position of any other lattice
point.

Closely related to this translational symmetry is a property of a lattice

that gives us a hint why naturally occurring crystals have regular
crystalline shapes. Consider the lattice shown in part (b) of fig. II-9. If
I join the atoms marked A and B by a straight line, and extend it in
both directions, I get a row of atoms that are equally spaced. If 22

II CRYSTALS

FIGURE II-II. Rotational and ® © © © © @ ©

reflection symmetries in a two-dimensional square lattice. Part (a)

shows axes of four-fold symmetry ® © il II ® © ©

through the points A and B, and an

axis of two-fold symmetry through C. ^ # @ @ . © @ ©

An axis of n-fold symmetry means f^

that the lattice can be rotated to n \J%) ~

positions about that axis, and still © @ © # © ® ®

look the same. In part (b) of the A

figure, the broken line through A and

Bis a line of reflection symmetry. The • # ® ® ® @ ©

reflection interchanges atoms C and

D, and all other similar pairs of atoms @ © ® ® @ @ <©

likewise, leaving the lattice looking

the same as before. The dotted line

above A is also a line of reflection

symmetry. © • II ® It © @

©®©®©®©

wAB

©©®®©©©

®•©®©©©

© <§> © ® © ® @

through a further 90 degrees, and at the end of each rotation it still

looks the same as before. A final rotation of 90 degrees brings it
back to the starting position, because a 360-degree-rotation is a full
rotation. I have thus found a rotational axis of four-fold symmetry:
there are four positions reached by successive rotations through 90
degrees where the lattice looks the same. This axis of symmetry
may pass through any lattice point.

question Are there any other axes of rotational symmetry for the
square lattice?

answer You can verify from the upper picture that if the axis passes
24

SYMMETRIES IN A TWO-DIMENSIONAL LATTICE

through the centre of a unit cell, B, it is again an axis of four-fold

symmetry. If it passes through the midpoint between two nearest
neighbouring points, C, then it is an axis of two-fold symmetry.

question What are the rotational symmetries in the triangular lattice

for an axis through a lattice point, and through the centre of a
triangular unit cell?

answer Six-fold and three-fold respectively.

These lattices have one other type of symmetry. Consider the lattice
shown in part (b) of fig. 11-11. Draw a straight line through A and B,
think of it as a mirror, and imagine that the portions of the lattice lying
above and below the mirror are reflected by it. Then, for example,
the reflection of atom C will coincide with atom D, and vice versa.
The reflection of the portion of the lattice that is above the mirror will
look exactly like the portion below, and vice versa. The appearance
of the lattice is unchanged by this reflection. The line through AB is a
line of reflection symmetry.

question Can you find another line of reflection symmetry in the

square lattice?

answer Yes. The straight line through A and C.

question There are six lines of reflection symmetry in this lattice. Can
you find the remaining four?

answer A line parallel to the line joining A and B, and sitting midway
between B and C. Now if you imagine this and the other two lines
each rotated through an angle of 90 degrees, the resulting lines are
also lines of reflection symmetry.

question Which of the five lattices we have been looking at has the
least symmetry?

answer The oblique lattice. It has only translational symmetry and

one axis of two-fold rotational symmetry.

Let us pause and take stock of what we have done. I list the
highlights: 1. The lattice surrounding a lattice site looks exactly the
same as the lattice surrounding any other lattice site. This is the
cosmological principle as applied to a crystal, and is a consequence
of the fact that all the atoms of a given element are identical.

2. The following is what is meant by the symmetry of a lattice: I carry

out a specified operation on the crystal. The operation may be a
translation, or a rotation, or a reflection. If it is a symmetry operation,
then the crystal will look exactly the same after the operation as
before it.

3. The order in a lattice or crystal is the total of its symmetries.

II CRYSTALS

4. Each of the two-dimensional lattices has its own characteristic set

of translation, rotation and reflection symmetries.

I want now to say something about a more general matter. There are
some physicists who, when they have the time, think and even talk
or write about the beauty of physics.
You might recall the physics you were exposed to in your school
days, with its balls rolling down slopes, electric currents in wires
causing nearby compass needles to move, and so on.

You might wonder what exactly is supposed to be beautiful about all

those things.

I think that there is beauty in the way symmetries suddenly appear

when I put atoms together to form a solid. We began with a simple
and very reasonable condition: each atom in the crystal should see
the same pattern of atoms around it. Out of this emerged the variety
of symmetries we see in the crystal. One ordinarily associates
beauty with creations in the arts: a Mozart symphony, a Thyagaraja
song, a Leonardo da Vinci painting, the Taj Mahal. If you can begin
to think of the symmetries of a crystal also as being beautiful, you
will have enlarged and enriched your sense of beauty.

There are other symmetries that exist in nature, in addition to the

ones we have seen in crystals. Consider, for example, a container
filled with helium gas. Each helium atom looks exactly like every
other helium atom. Nothing about the gas would change if I imagined
interchanging the positions of any two atoms in the gas. The gas is
symmetric for such an operation. We shall see later that quantum
mechanics gives this particular symmetry an unexpected new twist,
with far-reaching consequences for our understanding of nature.

The following are two more examples of symmetries in nature. All

objects exist in space and time. Their behaviour is governed by the
laws of physics, such as the law of gravitation and the law of force
between electric charges. We know from experience that these laws
are exactly the same on earth as anywhere else in the universe, and
have been the same for as far back in time as we can determine.
The astronauts’ experiments on the moon obeyed the same laws of
physics as the experiments we do on earth, and the weights which
Galileo dropped from the leaning tower of Pisa obeyed exactly the
same law of gravitation as the pencil I just dropped on the floor. The
laws of physics are the same at all points in space, and at all instants
of time. From the way we have understood symmetry in physics, we
can say:

The laws of physics are symmetric/or displacements in space and in

time.

Having got this far, you might be willing to grant that there is some
sort of beauty in these symmetries, but you might well wonder what
on earth one can do with them.

These last symmetries have profound consequences for all of

physics. For example, the symmetry with respect to time implies that
energy cannot be created or destroyed but can only be changed
from one form to another. To explore these matters fully would take
us far afield from our present concern, which is to understand solids.
So let me give an example of how we might use the symmetry of a
crystal to simplify the task of understanding its properties. We saw
that the neighbourhood of any one atom is exact-26

II CRYSTALS

I end up with a three-dimensional lattice. Figure II-12 shows you

what a small portion of this lattice looks like. I think you can see what
I meant about the difficulty of drawing a picture of a three-
dimensional object on a two-dimensional page. However, if you use
your imagination a bit, you can see that the lattice is made up of a lot
of tiny bricks each with edges a, b, and c. I have marked one of them
at the top left with heavy lines to help you visualize it. This tiny ‘brick’
is the unit cell ofthe lattice, and is shown in fig. 11-13, labelled (1). I
can build up the whole lattice out of these unit cells. The positions of
the atoms surrounding any given atom in the lattice satisfy the
cosmological principle.

There is another way of stacking rectangular lattices so as to satisfy

the fundamental condition. I place the second lattice over the first at
a distance c/2, with each atom lying exactly over the centre of a little
rectangle in the first lattice. I place the third lattice at a FIGURE II-13.
Unit cells for some

three-dimensional lattices: (1) simple

orthorhombic; (2) body-centred

orthorhombic; (3) simple cubic;

(4) body-centred cubic; (5) face-centred cubic (only the three front

faces are shown for clarity); (6) simple

hexagonal. The unit cells (3), (4) and

(5) are shaped like cubes. The lattice

represented by (6) results from a

vertical stacking of the two-dimensional triangular lattice.

(1)

(2)

i,

1—

(3)

(4)
(5)

(6)

THREE-DIMENSIONAL LATTICES

distance cfrom the first, but with each atom now lying exactly over an
atom of the first lattice (there is a similarity to the way we made a
centred rectangular lattice in two dimensions). I show the resulting
unit cell in fig. II-13, labelled (2). This lattice is different from the
previous one, but still satisfies the cosmological principle. More
three-dimensional lattices can be assembled using each of the other
four two-dimensional lattices.

question Is one allowed to take a mixture of several kinds of two-

dimensional lattices to build a three-dimensional lattice?

answer No, because then the cosmological principle will obviously

not be satisfied.

Even with this constraint, you might think that the possibilities are
practically limitless. There are however only fourteen lattices with
different symmetries that I can build satisfying the cosmological
principle. Figure 11-13 shows the unit cells and the names of some
of these lattices. The circles in these figures represent the lattice
points, and the lines joining them are there to guide the eye in seeing
the shapes of the unit cells.

We thus come to the astonishing conclusion that the well-nigh infinite

varieties of crystals that occur in nature or are made in the laboratory
belong to one of just fourteen crystal lattices. I recapitulate the line of
thought that has brought us to this conclusion:

1. All atoms of a given element are identical.

2. As a consequence, the cosmological principle holds for a solid.

3. A lattice is characterized by its symmetries.

4. Two lattices are different if they do not share exactly the same set
of symmetries.

Nature confirms our conclusion: all crystals that occur naturally or

are made in the laboratory belong to one of these fourteen lattice
types.

We talked earlier of a number of symmetries and other

characteristics of two-dimensional lattices. Suitably modified, they
exist in the three-dimensional lattices too, but it is not always easy to
see them in a picture of the lattice on a page. Physicists, who have
the same difficulty in visualizing these lattices, resort to models
assembled from balls and rods. All these properties can be proved
by applying mathematics. I list the results below:

1. There is translational symmetry for displacement from one lattice

point to any other lattice point.

2. A straight line connecting any two lattice points, when extended,

will pass through other equally spaced lattice points, and only those
points.

3. A lattice can be thought of as being made up of identical parallel

planes in more ways than one, each with its own characteristic
spacing within and between planes.

By analogy with the discussion of how a two-dimensional crystal

grows so that it is bounded by straight lines of close-packed atoms,
one can understand why natural crystals are bounded by plane
faces, each a plane of close-packed atoms.

29

II CRYSTALS
FIGURE n-14. Part (a) shows the

graphite structure, which is a

triangular lattice with pairs of atoms

like A and B attached to each lattice

site. Part (b) shows the cubic unit cell

for a crystal of common salt (sodium

chloride).

(a)

chlorine atom
8 Crystal structures
The lattices we have been considering are not real objects, because
there are no atoms in them. They are just sets of points, distributed
in space in a way that satisfies the cosmological principle. It is only
after we associate atoms with the lattice points that we get a physical
object. A crystal of silver has the atoms sitting at the points of a face-
centred cubic lattice, whose unit cell is shown in fig. 11-13, labelled
(5). We say that silver crystallizes in the face-centred cubic structure.
Brass is an alloy in which 65 per cent of the 30

CRYSTAL STRUCTURES

atoms are copper and 3 5 per cent are zinc. A crystal of brass also
has a face-centred cubic structure, in which the copper and zinc
atoms occupy the lattice sites at random.

question In such an alloy there are two kinds of atoms, and so the
cosmological principle cannot be satisfied. What happens to the
lattice spacing?

answer Call the two kinds of atoms A and B. Because an A atom is

different from a B atom, the lengths and also the strengths of the A-
A, A-B and B-B springs will be different from one another. These
differences are usually small. So the crystal of the alloy will have its
atoms displaced, but only slightly, from the exact lattice sites.

Yet another kind of crystal is made up, not of single atoms sitting at
each lattice site, but rather of a group of two or more atoms
associated with each site. Further, the atoms in the group could all
be the same, or they could be different from one another. We could
also have molecules around each lattice site. I give below examples
of these different possibilities.
The element carbon exists in nature with two different crystalline
forms, as graphite and as diamond. Graphite consists of identical
planes, one of which is shown in part (a) of fig. 11-14. The picture
does not look at all like one of the five two-dimensional lattices.

It is in fact a triangular lattice, with pairs of atoms like the ones

marked A and B attached to each lattice point. This arrangement of
atoms is known as the graphite structure.

My next example is common salt, which consists of equal parts of

the elements sodium and chlorine. The sodium atoms form a face-
centred cubic lattice. There is one chlorine atom for each sodium
atom, sitting on its right at a distance a/2, where a is the length of the
cube edge. I show in part (b) of fig. 11-14 the resulting unit cell for a
crystal of common salt. This is called the rocksalt structure, after the
mineral form of common salt. You may have seen large cubic
crystals of rocksalt in museums of science.

Figure II-15(d) shows what common salt looks like through a

microscope. You can see the cubic nature of the grains, arising from
the cubic unit cell.

My last example is a crystal composed of sugar molecules. I show in

fig. ll-15(b) a photograph of grains of sugar made with a microscope.
Each grain is a crystal. The molecule has six atoms of carbon, six of
oxygen, and twelve of hydrogen. It is somewhat more complicated
than the examples we have looked at so far. Nevertheless, the
crystal still forms one of the fourteen lattices, the orthorhombic
lattice, whose unit cell is like a brick with unequal edges. The flat
faces of the crystals are sheets of molecules lying relatively close to
one another.

The crystals in fig. II-15 do not look perfect: they have rounded
edges, and some are broken. This is because of the handling they
have received and their solubility in water.

Note particularly that while their sizes vary, the basic shape is the
same for all crystals of each of the two substances. This is a
beautiful manifestation of the underlying orderly arrangement of the
atoms or molecules in the crystal.

I have so far used symbols like a, b, and so on to denote the

quantities that describe a SUMMARY

seems at first glance to violate the reasoning we have used to

explain the formation of crystals. A little further thought shows that
there is in fact no problem. Remember the virtual springs that we
imagined connecting each atom to its surrounding atoms? The very
rapid cooling leaves the atoms at or very near the positions they
occupied in the liquid, each feeling no net force so that we still have
a solid. However, the springs surrounding each atom do not all have
the same relaxed length that they would have had in the crystal.
Some will find themselves a little compressed, others a little
extended, but altogether in such a way that there is still no net force
on each atom. When we realize that each atom has many springs
connecting it to the surrounding atoms, we can see that there is a
good chance that the atoms can find positions, not ordered, in which
each atom feels no force even though each spring is a bit
compressed or extended. The result is a glassy solid. This picture
also helps us understand why some solids are more easily formed in
the glassy state than others: the exact details of the ‘springs’ play a
crucial role. We recall that a glass, taking a long time that may be
several centuries or more, will change to a crystalline form, which
therefore appears to be the preferred form for the atoms in a solid to
take.
10 Summary
We saw in section 1 that as a liquid is cooled to form a solid, the
result is a single crystal, or a polycrystal, or a glassy solid,
depending on the cooling rate. The atoms sit in an orderly
arrangement in the crystal and polycrystal, but not so in the glassy
solid. The arrangement in which the atoms prefer to be is the
crystalline one.

The cosmological principle as it applies to crystals is introduced in

section 2. It is seen as a consequence just of the fact that all atoms
of a given element are identical to one another. It seems reasonable
therefore that the principle should hold for the preferred atomic
arrangement in the solid.

The simplest solid we can think of is just a line of atoms. If this line is
to satisfy the cosmological principle, the atoms will be equally
spaced along a straight line. This is the one-dimensional crystal,
which is discussed in section 3.

In section 4, the five possible two-dimensional lattices are built up

from a collection of identical one-dimensional lattices.

Section 5 makes precise the idea of order in a crystal by introducing

the concept of symmetry as it applies to crystals. The symmetries of
two-dimensional lattices are discussed in section 6. Several kinds of
symmetry are seen to exist: of translation, of rotation, and of
reflection.

Stacking a number of two-dimensional lattices on top of one another

while maintaining the cosmological principle leads to fourteen
different three-dimensional lattices. Section 7 introduces them and
some of their characteristics.

In section 8, several ways are described in which atoms can be

associated with lattice sites, leading to crystals of materials each
with its characteristic structure.

33

II CRYSTALS

Section 9 deals very briefly with a glassy solid, and points out that
after a long time of possibly centuries, it will have changed to a
crystalline form, which seems to be the preferred form of atoms in a
solid.

This concludes our description of crystals. The crystals we have

assembled have no properties other than their symmetries. Their
physical properties arise only when we put real atoms at the lattice
sites. Before doing that, we need to learn about an isolated atom in
terms of its nucleus and electrons in some detail. We shall then be
able to see what happens to the atom when it finds itself in a crystal.
Finally we shall use the resulting picture of the crystal to explain
various properties of a solid.

I have said that an atom consists of a nucleus with some electrons

around it. To describe it in more detail, I need to introduce the
description of nature called quantum mechanics. This is the subject
of chapter III.

34

Ill PARTICLES AND WAVES

l Some preliminaries

We have, until now, thought of atoms as tiny hard spheres that feel a
force of attraction towards one another. We ignored the fact that an
atom consists of a nucleus surrounded by one or more electrons. If
we stopped at this, we would make little progress towards
understanding the properties of solids, other than that the atoms may
form crystals. We must therefore go on and take a look at the
internal structure of the atom itself, and see how this changes when
the atom is surrounded by other atoms in a crystal. We shall then
find that we need a description that is somewhat removed from our
everyday experience. There is a name for this description, quantum
mechanics-, mechanics, because it deals with the motion of objects,
and quantum, because features of the motion, such as energy, are
quantized, meaning that they can only have any one of a set of
distinct values, and no other value in between. Imagine for example
a car which is so constructed that it can travel only at speeds of 10,
or 50, or 100 km per hour, and not at any other speed. Then we
could say that the speed of the car is quantized. This idea of
quantization is something alien to our usual way of thinking. After all,
a car is able to go at any speed we like up to its maximum speed.
We shall find that there is no such freedom of choice for electrons
and nuclei. In this chapter I shall introduce you to the essentials of
quantum mechanics, and then in the next chapter show you the
description of an atom that it provides. I have summarized the
content of this chapter in section 13 at the end (page 61). You may
want to read it now, so that you have a preview of what lies ahead.
2 Motion
As I write this, I take a moment to look out of the window. I notice
that there is a breeze that is causing the leaves on the trees to
flutter. Birds are flying in the air. I see a stream flowing among the
trees, and there are small waves on its surface that are themselves
also in motion. Although I cannot see them, I know that the
molecules in the air around me are in a perpetual state of motion.
The paper clip on the table before me seems to be sitting there,
doing nothing. If I look at what the atoms in it are doing, using the
right sort of microscope, I find that they are all jiggling a little bit
about their regular crystalline positions. It would appear that the
world around us is always in various kinds of motion. At first glance,
there seems to be a daunting variety of such motions, and of moving
things. However, the essential features of each of these motions can
be 35

Ill PARTICLES AND WAVES

assigned to one or the other of just two types, namely the motion of
particles, or of waves.

Before proceeding, I should make a general observation. I use words

in this book, many of which are taken from everyday usage, to
denote certain concepts. When so used, they have very narrowly
defined meanings, and should not be confused with their everyday
connotations. When I say crystal, I mean a solid with an orderly
arrangement of its atoms as described in chapter II, and not an
expensive piece of glassware. Other examples are energy,
momentum, force, particle, conservation. As Humpty Dumpry said in
Lewis Carroll’s Through the Looking-Glass, ‘When I use a word, it
means just what I choose it to mean, nothing more nor less.’ The
glossary beginning on page 238 lists these words with explanations.
We already have some idea of what is meant by a wave; we have all
seen waves on the surface of a stretch of water, for example. A wave
is spread out in space and moves in some direction. It has many
alternating crests and troughs.

I now explain what I mean by a particle. Here is a definition: A

particle is an object that has a mass and occupies a vanishingly
small volume. As I make the volume smaller and smaller, it finally
becomes a point in space. I can therefore think of a moving particle
as a mass that is concentrated at a point and is moving about. A
particle is thus an idealization of an ordinary moving object. It is in
this sense that I say that electrons and nuclei are particles. An
essential feature of a particle is that it is at some point in space at
each instant of time: it is localized in space and time, in contrast to a
wave, which is spread out in space at every instant of time. Figure
III-1 illustrates the difference between a particle and a wave.

Electrons and nuclei have three other properties in addition to their

mass: they have an electric charge, they behave as if they were
spinning about an axis, and they respond to a bar magnet the way a
magnetic compass needle does, as if they were tiny bar magnets
with a north pole and a south pole. Furthermore, these properties
remain exactly the same at all times and places. There is nothing we
can do which will change the mass, the charge, the spinning, or the
magnetism of an electron or a nucleus. They are what lead ultimately
to our understanding of the properties of matter.

FIGURE ill-l. To illustrate the

■difference between a particle, which is

very localized in space, and a wave,

which is extended over all of space. • * particle

PARTICLES
3 Particles
To begin with, a particle has mass, say M grams. If it is an electron,
then the mass is 9.11Xl028gram. Fora proton, it is 1.67xlO-24gram,
or about 1800 times the mass of the electron. The mass of a particle
is the measure of the amount of matter it contains.

It is the property which gives it its heaviness, its weight, due to

gravity, and also determines how it moves when a force is applied to
it.

As the particle moves about, it is at some point in space at each

instant of time: it is localized in space and time. The particle moves
in some direction with some speed. In our everyday experience with
moving objects, we are often content with knowing the speed, so
many kilometres per hour, for example, and do not explicitly state the
direction of motion. We know that direction can often be important
too: we would not easily get to Cambridge from Oxford or to
Cleveland from New York if we insisted on travelling south all the
time. Similarly, when we try to understand the nature of matter in
terms of its basic constituents (atoms, electrons, etc.), we use a
quantity called velocity denoted by the symbol v (in heavy type)
which tells us not only how fast but also in which direction. When I
say that a particle has a velocity v, I mean that it has a speed of v in
specified units (centimetres per second, for example) in some
specified direction.

Two particles moving with the same speed but in different directions
have different velocities, as also two particles moving in the same
direction but with different speeds.

We call quantities that have both a magnitude (so many centimetres

per second in the case of v) and a direction, vectors, and denote
them in heavy type. When I want to speak only about the magnitude
of the vector, I shall use the same letter but now printed in ordinary
italic type, v in the case of velocity. A quantity that has only a
magnitude and no direction is called a scalar. Thus an interval of
time, t seconds, is a scalar.

In solids we are concerned not only with the motions of the atoms,
but also with how they influence one another. A simple example will
illustrate the point. In air, say, each molecule is whizzing about,
colliding every now and then with another molecule.

After each such collision, the two molecules have velocities that are
different from what they were before. The final velocities will depend
not only on the initial velocities of the two molecules, but also on
their masses. I would certainly notice the difference between a fly
and a cyclist colliding with me as I am walking along a path, even if
they both were travelling at the same velocity. So, in order to
consider collisions (or more generally interactions), between
particles, I need a concept, and a symbol for it, which combines
mass and velocity. I call it linear momentum, and give it the symbol p
(again a vector). Its direction is the direction of the velocity v, and its
magnitude p is given by the mass M multiplied by the speed v. In
short, we can write p = Mv, a vector equation,

and

p = Mv, a scalar equation.

Note that the second equation gives only the magnitude of the vector
p.

37

Ill PARTICLES AND WAVES

There is another kind of motion possible for a body, in which it is

spinning about an axis, but otherwise not going anywhere. The body
as a whole has zero velocity, and therefore zero momentum. All its
atoms however are spinning around in circles about the axis. The
momentum connected with this motion is called angular momentum.
We note that it is essentially different from the linear momentum we
got by multiplying the mass and the velocity. Angular momentum is
also a vector, and its direction is along the axis of rotation. Its
magnitude is fixed by how fast the body is rotating and how its mass
is distributed around its axis of rotation. We recall that electrons and
nuclei behave as if they were spinning about an axis: they have an
angular momentum that is an intrinsic property. In what follows, I
shall simply use ‘momentum’ when the context makes it clear which
kind is meant.

We have one final concept to introduce, that of energy. Here is

another of those words of normal speech, which we use in this book
with a very specific meaning. I start by defining energy, and then give
some examples.

A body has energy either if it is already in a state of motion, or if

even though at rest, it can through appropriate means put itself or
another body into a state of motion.

In the first case, we say that the body has kinetic energy, and in the
second case, potential energy. The kinetic energy is denoted by the
symbol K, and is given by KM

K~ 2

or equivalently

as can be seen from our definition of p as equal to Mv. If a mass of 1

gram is going at a speed of 1 centimetre per second, it is said to
have a kinetic energy of one-half erg. If the mass is travelling at 2
centimetres per second, its kinetic energy is 2 ergs. Note that
doubling the speed quadruples the energy: cyclists and car drivers
should be aware of this. Energy has just a magnitude of so many
ergs, and no direction attached to it, unlike momentum: energy is a
scalar.
question Can you name another scalar quantity?

answer The mass M. It has only a magnitude of so many grams, and

no direction.

I now turn to the second type of energy, potential energy. It is a form

of energy that, under the right conditions, can be converted into
kinetic energy. I illustrate it with a few examples. I place a marble at
the edge of a table. It is not moving, and therefore has no kinetic
energy. I now give it a gentle shove to send it over the edge. It starts
falling, with ever increasing speed, until it hits the floor: it has
acquired kinetic energy. I say that the marble sitting on the table has
potential energy, because when I remove the table’s support the
marble begins to move. Since this property of the marble arises from
gravitation, I call it gravitationalpotential energy. Gravity acting on the
marble causes it to PARTICLES

move when it is no longer held up by the table. You will note that
even when the ball was sitting still on the table, it had the potential to
move; hence the name potential energy. As another example,
consider an electric battery. I connect it to a light bulb with a pair of
wires and a switch. When I close the switch, the battery drives a
stream of electrons — we call it an electric current — through the
bulb causing it to light up. The electrons move because of the
battery. I say that the battery has chemical potential energy.

If I replace the electric battery with a solar battery and shine sunlight
on it, I get the same result. Thus sunlight also has energy that can
be converted to the motion of particles, electrons in this instance.
This energy is commonly called solar energy. A physicist would
probably call it electromagnetic potential energy, because light is
intimately connected with electricity and magnetism, as we shall see
later. Yet another example: if I stretch a rubber band between my
hands and let go, it flies off. The stretched band has elastic potential
energy when it is stretched, and this is converted to kinetic energy
when I let go the band. If it is clear from the context that I am talking
about some particular form of potential energy, then I may simply say
gravitational energy, chemical energy, and so on.
question Can you think of an example that shows that heat is also a
form of energy?

answer A steam locomotive. Energy in the form of heat is converted

to the kinetic energy of motion of the locomotive.

Finally, I give an example that is probably known to everyone: Albert

Einstein’s discovery that a particle of mass AT has a potential energy
by virtue of its mass alone, equal to Me2 where cis the speed of
light, namely 3 X1010 centimetres per second.

The foregoing does not exhaust all possible forms of energy, but is
meant to show that it comes in many different forms. The examples
also illustrate the fact that it is possible to convert the energy from
one form to another: the locomotive converts thermal energy to the
energy of motion. Now it often happens that one and the same
particle can have energy in more than one form at the same time.
For example, a particle sliding on a tabletop has both kinetic energy
and gravitational potential energy. In such cases, we add up all the
separate energies and call it the total energy E. We have E= kinetic
energy + gravitational potential energy + etc.

I summarize now the essential characteristics of a particle: a particle

is localized in space and in time, and it has a mass M, a momentum
p, and a total energy E.

We have so far considered a single particle. The pieces of matter

whose properties we wish to understand consist of many particles
such as atoms and molecules, themselves composed of nuclei and
electrons. If I take a piece of matter, I get its total energy by adding
up the individual energies of each of its constituent particles, and its
total momentum by doing the same with the individual momenta
(plural of momentum).

I have no problem doing this with the energies, because they are
scalars, just numbers. How do I do this with momenta, which are
vectors with both magnitude and 39
Ill PARTICLES AND WAVES

FIGURE Hl-2. Adding two vectors

together: p added to q results in the

vector r. This procedure is applied to

the question about the canoe on the

river. The direction of north (N) is

shown.

direction? We do this as shown in fig. III-2 for vectors p and q, which

for example could be two velocities. I draw an arrow whose length
and direction are the magnitude and direction respectively of p. I
draw a similar arrow for q, with its tail at the head of p. Then the sum
of the two, which I call r, in magnitude and direction is given by the
arrow which connects the tail of p with the head of q. The resultant
velocity is then r.

question I paddle north in a canoe at 4 km/hour (p) across a river

that is flowing east at a speed of 3 km/hour (q). What is my actual
velocity as seen by a person standing on the bank of the river?

answer The canoe is going north in the water at 4 km/hour, and the
water itself is flowing east at 3 km/hour past the person on the river
bank.

Using the above prescription and direct measurement, or the

theorem of Pythagoras and trigonometry, I find my velocity to be 5
km/hour (r) in a nearly northeasterly direction, 37° from north to be
precise, as seen by the person on the river bank. Figure III-2
illustrates this.
40

THE CONSERVATION OF ENERGY AND MOMENTUM

4 The conservation of energy and momentum

The subject of this section is a profound and beautiful basis of all of

physics. These conservation principles are sometimes explicitly, and
often implicitly, contained in much of what follows in the book. I shall
first state what is meant by the conservation of energy and
momentum:

The total energy, the toted linear momentum, and the total angular
momentum of a closed system —are constant at all times: they are
said to be conserved.

The statement summarizes three fundamental principles that govern

all of physics: the conservation of energy, the conservation of linear
momentum, and the conservation of angular momentum. When we
say ‘conservation of energy’ here, we mean something different from
switching lights off when they are not needed. We should of course
be as economical as possible in the use of energy in our daily life.
When we talk here about energy being conserved, it is not a matter
of personal choice but a fact of nature. By ‘closed system’, we mean
some portion of our surroundings that is so fixed up that neither
matter nor energy can enter it or leave it. This is an idealization of
what can be achieved in the real world, where it not easy to exclude
totally a flow of energy and/or matter. Some hot coffee in a closed
thermos flask is an approximation to a closed system. No form of
matter can enter or leave the flask, nor much energy as heat or light.
It is not a perfect closed system: the coffee does cool eventually,
because it slowly loses thermal energy to the outside.

Let us see how the conservation principles apply to the case of a ball
falling off the edge of a table and coming to rest on the floor. At the
start of the fall, the ball has only gravitational potential energy.
During the fall, the potential energy is continually converted to kinetic
energy of motion, in such a way that the total energy remains
constant. Just before the instant when the ball hits the ground, its
initial potential energy has been completely converted to an equal
amount of kinetic energy, so that the total energy is still the same.
After impact, the ball is at rest on the floor, and both its kinetic and
potential energies are zero. Figure III-3 shows the three successive
stages. You may wonder what happened to energy conservation
when the ball came to rest on the floor, since both its kinetic and
potential energies then are zero. Not to worry: the impact makes a
sound and the ball warms up a bit, because the total energy has now
appeared partly as sound and the rest as heat. If you have ever
hammered a nail into the wall and noticed the nail warming up, you
will have seen the conversion of kinetic energy into thermal energy.
But how about the linear momentum? The ball started from rest with
zero momentum, and continually increased its speed as it fell. Its
momentum certainly did not remain constant. Then we note that the
closed system is not just the ball by itself, but the ball plus the earth
towards which it is falling, as seen by an extraterrestrial creature
somewhere out in space. This creature sees the earth move up to
meet the ball with a momentum which at each instant of time is
exactly equal to and opposite to the momentum of the ball, so that
the total momentum at all times is zero: it is conserved.

Ill PARTICLES AND WAVES

FIGURE III-3. From left to right, three

successive stages of a ball falling off a

table. The initial gravitational

potential energy is converted to

kinetic energy and then to the energy

of sound and heat.

question Then should I not include the earth in considering the

energy also?

answer Yes, you should. However, because of the enormous

difference in masses of the two objects, the kinetic energy of the
earth is absolutely negligible compared to that of the ball, even
though their momenta have equal magnitudes. You remember that
the kinetic energy is equal to the square of the momentum divided by
twice the mass. So the kinetic energy of the earth is small compared
to that of the ball in the same proportion as the mass of the ball is
small compared to that of the earth.

As mentioned in chapter II, these conservation principles arise out of

the fundamental symmetries of space and time. To go deeper into
this will lead us too far from our main course, to which we must now
return.
5 Waves
We encounter waves in nature in a great variety of forms: the ripples
which spread out from a stone dropped into a pool of water, sound
waves in air, light which is an electromagnetic wave that travels
through even empty outer space, and great ocean waves, to name a
few. We have a similarly great variety of moving objects in nature,
and in order to distil out the essential aspects of their motion, we
came up with the idea of a particle. We shall now similarly consider
an idealized wave, shown in fig. III-4 at five successive instants of
time. To be specific, imagine it to be a wave on the surface of water.

The wave, unlike a particle, is not localized, but is spread over a

region of space. It consists of alternating crests and troughs, moving
with a certain speed. By including 42

WAVES

the direction in which the wave is travelling, I can talk about the
velocity of the wave.

Although the wave itself is travelling in a certain direction, there is no

general motion of water in that direction. I see this by floating a piece
of cork on the surface of the water. As the successive crests and
troughs go by, the cork moves up and down, but is not carried
forward. The wave does not carry matter along with it, whereas the
motion of a particle is certainly the motion of matter. The maximum
distance that the cork moves from its mean position, as it oscillates
up and down, is called the amplitudeof’the wave and denoted by the
symbol A. At any fixed point along its path, the wave is oscillating up
and down/times per second, and/is called the frequency of the wave.

The distance between adjacent crests is always the same, and equal
to the distance between adjacent troughs. This distance is called the
wavelength, and denoted by the Greek letter X (lambda).

Figure III-4 illustrates some of these points. You can see that this
wave has a wavelength of 1 metre. The broken vertical line near the
left end of the picture intersects the wave at the position of the cork
at successive instants of time. It is on a crest at 0 second, has
moved down to a trough at 2 seconds, and is back again on the
following crest at 4

seconds. It has completed one oscillation in four seconds, and

therefore the frequency is 4 oscillation per second. Each crest is
moving to the right as time elapses, and has moved a distance of 1
metre in 4 seconds. The speed of the wave is therefore \ metre per
second.

question Can you see that, if the speed of the wave is v, then v is
equal to the product of/and A?

answer I look at the cork, and find that crests are passing it each
second. The crests are spaced a distance A. apart. Therefore the
crest, which one second ago was at a distance times X, is now
passing the cork. So the speed of the wave is/times X centimetres
per second. I can write this as v —fX.

If I know the wavelength and frequency of a wave, I just multiply the

two together FIGURE III-4. A wave travelling to

the right, as seen at successive

intervals of one second. The

intersection with each wave of the

broken vertical line near the left

indicates the position of the cork

referred to in the text. You can see

from the figure that there is one

oscillation in four seconds (or a

frequency of one-fourth oscillation

per second), a wavelength of one

metre, and a speed of one-fourth

metre per second.

direction of wave

.oo

distance in metres

43

Ill PARTICLES AND WAVES

to get the speed. This speed depends on the nature of the wave and
the substance in which it travels. The speed of sound waves in air is
331 metres per second, in water 1480 metres per second. The
speed of light waves in free space is 3 X 10s metres per second, in
water 2.24 X 10s metres per second.

The last three attributes of wave motion that we have described,

namely the amplitude, frequency and wavelength, have no
counterparts in particle motion. In addition, I see no obvious way of
talking about the momentum or energy of the wave, as I could with a
particle. I know however that waves also do have energy and
momentum, as is shown by the following effects with sunlight, which
is a form of electromagnetic waves. The energy of sunlight is
converted into the energy of motion of electrons in the current
produced by a solar battery. The momentum of sunlight causes the
tails of comets to stream away from the sun. We shall see later how
quantum mechanics tells us how to associate energy and
momentum with waves.

Since water consists of molecules (each made up of two atoms of

hydrogen and one of oxygen), I could argue that in principle I could
describe water waves in terms of the motion of water molecules: a
task which would need the lifetime of the universe to complete —
perhaps. Then there are electromagnetic waves in the form of light,
which travel through empty space where there are no molecules or
other forms of matter at all.

Light must therefore be a pure wave with nothing in the form of

particles of any sort associated with it. So it would seem that we
must accept particle motion and wave motion as two distinct and
non-overlapping forms in nature.

I now come to yet another difference between particle motion and

wave motion.

When two particles are moving together, I have no difficulty in

thinking of the motion of each particle independently of the other. If
the two collide with each other, they would continue their separate
motions afterwards, still clearly distinguishable from each other. If on
the other hand I get two waves to come together, something very
dramatic happens, as illustrated in fig. III-5.1 take two waves, which I
label (1) and (2), with slightly different wavelengths. I let them
combine so that at each point the new amplitude is the sum of the
amplitudes of the two waves, and get a resulting wave that I have
labelled ‘sum’. This wave has an amplitude that waxes and wanes
FIGURE III-5. The interference of two

waves travelling together. The

separate waves labelled (1) and (2)

add up to give the wave labelled

44

Ill PARTICLES AND WAVES

There is still another difference between particles and waves. If I

place an obstacle in the path of a moving particle, the particle may
either hit the obstacle and change its subsequent motion, or miss it
and continue on its way as if the obstacle did not exist.

On the other hand, the motion of the wave after it passes an

obstacle is quite different from what it would have been without the
obstacle. Put differently, a particle cannot go through an obstacle,
but a wave goes around it and continues although with an altered
form. Waves undergo diffraction by an obstacle, whereas particles
do not.

When I go to an open air concert, I sometimes find a very large

person sitting in front of me, pretty much blocking my view of the
stage. Even though there are no walls or roof to reflect the music
towards me, I can still hear it almost as if the person were not there.
Since light is also a form of wave motion, you may wonder why the
light from the stage is also not diffracted round the person to reach
me, so that I can still see the stage. The answer lies in the difference
in wavelength between the two kinds of waves: in the range of
centimetres to metres for sound, and a few times 10”5

centimetres for light. The size of the person in front of me is

comparable to the wavelength of sound, but is enormous compared
to the wavelength of light. When the size of the obstacle is
comparable to or smaller than the wavelength, we get strong
diffraction effects. If the size of the obstacle is very much greater
than the wavelength, the wave is blocked by the obstacle.

QUESTION If you hold a compact disc so that light from a window is

reflected from its playing surface into your eyes, you will see a
beautiful spectrum of colours. What causes this?
ANSWER There is a regular pattern of dots on the surface, with a
spacing that is comparable to the wavelength of light. So the daylight
is diffracted, and different wavelengths (which correspond to different
colours) are sent in different directions. The result is the spectrum of
colours which you see.

It is this pattern of dots, invisible to the naked eye, which is ‘read’ by

the compact disc player and finally emerges as music.
6 Particles and waves: a comparison
I summarize in table III-1 the characteristics of particles and waves,
so that the differences between them can be seen clearly. This table
shows us that in the everyday world, a particle and a wave are very
distinct entities. Each has certain characteristics that are not
possessed by the other. If I point out to you something that is
moving, you are able to tell me without hesitation whether it is a
particle or a wave. A ball thrown up in the air is a particle, ripples on
water are a wave. You will think that I am somewhat odd if I tell you
that the ball can sometimes behave like a wave, and the ripples like
particles. Well, that in essence is what I am about to tell you!

46

THE ELECTRON: PARTICLE AND WAVE

Table III-1. Properties of particles and waves

particles waves

localized in space and time not localized in space at any time mass,
M cannot imagine mass of a wave

momentum, p wavelength, X

energy, E frequency, /

no interference interference

no diffraction diffraction
7 The electron: particle and wave
We want to understand the properties of matter in terms of the
constituent atoms, which are themselves made up of electrons and
nuclei. To be specific, I talk about electrons here. I describe three
experiments that can tell me whether an electron behaves like a
particle or a wave. The characteristics of electrons that these
experiments will reveal to me are also those of protons and
neutrons, the other fundamental constituents of matter. Similar
experiments are commonly performed in many physics laboratories
all over the world.

In the first experiment, a beam of electrons that are going quite fast
passes into air containing water vapour. Each electron causes
condensation of water along its path and creates a string of tiny
droplets of water marking its trajectory. The gadget, called a cloud
chamber, has been widely used in the study of electrons, protons,
and other basic constituents of matter. A photograph taken with such
a chamber is shown in fig. III-8. The curly track at the top is from an
electron, curly because of the presence of a magnetic field exerting a
force on it. The other tracks are from other particles passing through
the chamber. Clearly the electron is localized, and its momentum
and energy can be measured in various ways. We therefore
conclude that electrons are particles.

In the second experiment, a beam of electrons travelling at various

speeds, and therefore with a spread of momenta and energy,
bounces off the flat surface of a nickel crystal. One might expect that
each electron, being a particle, would scatter offa nickel atom on the
surface and proceed in some arbitrary direction; there should be no
noticeable pattern in the directions of the reflected electrons. What
actually happens is that all the electrons of a given energy come off
in one and the same direction, and this direction is different for
different energies. Just as sunlight is diffracted by the pattern on a
compact disc into a spectrum of different colours (frequencies), the
stream of electrons is diffracted by the regular atomic pattern of the
crystal into a spectrum of different energies.

47

LIGHT: WAVES AND PARTICLES

It appears that in the experiment with the crystal of nickel, the

electrons behave not as particles, but as waves with a frequency that
is somehow related to their energy, and these waves are diffracted
by the regular atomic arrangement in the crystal. Such an
experiment was first done by Clinton ]. Davisson and Lester H.
Germer, and independently and at about the same time by George P.
Thomson and A. Reid.

In the third experiment, the electron waves coming off the nickel
crystal pass through a cloud chamber as in the first experiment. One
would expect them as waves to spread throughout the chamber, but
instead the electrons behave like particles again, making tracks in
the chamber just as in the first experiment. So it seems that,
depending on how they are observed, electrons behave like either
waves or particles. The very same electrons that behaved like waves
when they bounced off the crystal of nickel seem to turn into
particles as they go through the cloud chamber. This situation is
totally alien to everyday experience, and we need a new way of
thinking about it. This way exists, and it is called quantum
mechanics.

The above experiments show what is called the wave-particle duality

of electrons.

I should not leave you with the impression that this is all the
evidence there is for it.

A vast number of other kinds of experiments have been done, and

they all lead to the same conclusion. Experiments with protons,
neutrons, atoms and molecules show that they also exhibit wave-
particle duality and behave, depending on the experiment, like
waves or like particles. There are other primary constituents of
matter whose names you might have sometimes seen in the science
pages of newspapers and magazines, like neutrinos and quarks.
They too show wave-particle duality, and are described by quantum
mechanics. This duality is a fundamental and universal property of
matter. There are more surprises to come, as you will see in the next
section.
8 Light: waves and particles
Light shows interference and diffraction effects. It is a form of
electromagnetic wave motion. These waves come in a tremendous
range of wavelengths and frequencies, of which ordinary light forms
only a very small part. They all travel at the same speed, namely the
speed of light. We call the waves in different ranges of frequencies
by different names. Some examples, in order of increasing
frequency, are radio waves, microwaves, infrared radiation, visible
light, ultraviolet radiation, X-rays, gamma rays. Having seen that
electrons behave sometimes like particles and other times like
waves, we might wonder whether the same is true for light also. Now
the simplest experiment we can do with two particles, two billiard
balls say, is to shoot one at the other so that they collide and move
off in different directions. We find that the total momentum and
energy of the two are the same after the collision as before. A
billiards player can use this result from physics to improve his game.
Now suppose we take a bunch of electrons that are standing still,
and shoot a beam of X-rays at it, and look at 49

Ill PARTICLES AND WAVES

the scattered X-rays. I find that I can describe the result exactly as if
the X-rays consisted of a stream of particles, whose energy and
momentum are related to the frequency and wavelength respectively
of the X-rays, and these particles bounce off the electrons keeping
the total momentum and energy constant. I conclude that these
waves, in this case, behave like particles. These particles of X-rays
are called photons.

Such an experiment was first done by Arthur H. Compton. I note

from its result that wave-particle duality holds for electromagnetic
waves also. As we shall see later, the exact relationships between
energy and frequency, and momentum and wavelength, are the
same for these waves as they are for electrons. Wave-particle duality
is a property of all of nature.
9 Wave-particle duality
Now we face the following question: a tennis ball, say, is made up of
electrons, protons and neutrons, all of which have wavelike
properties. Does that mean that I must then think of the ball also as a
wave that can be diffracted and so on? This should certainly make a
game of tennis rather complicated, with the ball being diffracted
around the racket. Fortunately for tennis, and indeed for all of our
normal activities, we can forget wave-particle duality for everyday
objects. A ball behaves like a particle, ripples on a lake are waves,
with separate and distinctive characteristics that were listed in table
III-1 (page 47). We say that we are looking at the macroscopic world
in such cases. When we wish to understand the properties of matter
in terms of the electrons and nuclei in it, at the so-called microscopic
level, we cannot do without wave-particle duality.

It could be that some of you are having difficulty in accepting this

fundamental duality of nature at the atomic level, a duality that is not
evident in our everyday world.

After all, things around us seem to be clearly distinguishable as

either particles or waves. So how is one to picture an electron as
being both a wave and a particle, exhibiting the characteristics of the
one or the other but not both at the same time, depending on what
one does with it? Well, the following might suggest to us that the
basic idea of duality is not very far-fetched. My friend Brian does
physics, and also plays the piano, but not both at the same time. He
goes into the laboratory and becomes a physicist; he sits down at
the piano and becomes a pianist. In each case he is still the same
person called Brian, displaying one aspect or the other depending on
where he finds himself. Indeed, most people have many aspects to
their personae. We can think of nature as being rather simple in
having only two.
The description of the world that embodies its wave-particle duality
is, as I mentioned before, called quantum mechanics. You may in
your reading elsewhere sometimes come across the term wave
mechanics. It means the same as quantum mechanics, but the term
is falling into disuse. I shall introduce you in the rest of this chapter to
some of the basic results of quantum mechanics.

SOME FORMULAS OF QUANTUM MECHANICS

10 Some formulas of quantum
mechanics
I shall be writing down some very simple formulas in this section.
You can think of them as just an economical way of saying the same
thing in words. I shall express in words the first formula when I get to
it, so that you can see what I mean. You may if you wish do the
same thing for yourself when you encounter later formulas.

We saw that we can characterize a moving particle by giving its

position, momentum (equal to its mass multiplied by its velocity) and
energy, and a wave by giving its amplitude, wavelength and
frequency. As before, we shall use the symbols p for the momentum
vector (and p for the magnitude of the momentum), E for energy, A
for amplitude, A, for wavelength and /for frequency. These symbols
stand for the numerical values of the quantities for any particular
case that we are considering.

Before we proceed, I must introduce you to another way in which to

think about the wavelength of a wave. Instead of saying that the
wavelength is A centimetres, we count the number of crests per
centimetre of the wave (which is equal to 1 A), call it the wave
number and give it the symbol k. Thus k = 1 A per cm, which can
also be written IA cm”1. The symbol cm”1, read ‘reciprocal
centimetre’, stands for ‘per centimetre’. A wavelength of one-quarter
centimetre is the same as a wave number of one divided by one-
quarter, or four, reciprocal centimetres. The wave is also travelling in
some particular direction, and we can include this information also by
introducing the wave vector k, whose size is given by k as defined
above and whose direction is the direction of travel of the wave.

So now we have light and electrons, both of which behave

sometimes like particles with momentum p and energy E, and other
times like waves with wave vector k and frequency f. Is there any
relationship between the particle-like quantities and the wavelike
quantities of either light or electrons? The answer to this question
has been given by innumerable experiments, and reveals a profound
and wonderful simplicity and symmetry of nature that lay hidden from
us till the beginning of the twentieth century. I say symmetry,
because these relationships apply to electrons as well as to light
waves. I first write down the answer, which consists of two simple
formulas, and then explain their meaning:

and

In words: the momentum of an electron (or proton, light wave,…) is

equal to its wave vector multiplied by a particular number that I have
denoted by h, and the total energy is given by the frequency
multiplied by exactly the same number. This is a unique number that
is fixed by nature, and we have no control over it whatsoever. If I
express masses in grams, distances in centimetres, and time in
seconds, then h has the value 6.625 lx 10”27. You might wonder why
I did not just put this number into the formulas, instead of denoting it
by the letter h. The reason is not just laziness, but because the
number would be Ill PARTICLES AND WAVES

different if I chose to measure masses and distances in pounds and

yards respectively, say. Further, its actual numerical value may
change slightly as more and more accurate ways of measuring it are
used. The constant h is called the Planck constant, after the
physicist Max Planck.

I have expressed the energy of the electron in terms of its frequency.

I recall that the kinetic energy K of the electron, namely the energy
due to its motion, can be expressed in terms of its momentum p and
mass m.

and since the momentum is related to the wave number k by p=hk

I put this value of p in the formula for Kznd find that K~ 2m-In words:
the kinetic energy if is equal to the momentum p multiplied by itself
(which I call p-squared) and the result divided by twice the mass m.
Since the momentum itself is equal to the Planck constant h
multiplied by the wave number k, the kinetic energy is got by
multiplying /^-squared by ^-squared and dividing the result by twice
the mass. I think you will agree with me that formulas give us a very
convenient way of expressing what would take a lot of words to say!

Quantities like the Planck constant h, the mass of an electron, a

proton or a neutron, the speed of light in free space (denoted by the
letter c), the magnitude of the electric charge of an electron or a
proton (denoted by the letter e), their intrinsic angular momenta
(plural of momentum) and a measure (called the magnetic moment)
of their behaviour as bar magnets, are all called fundamental
constants. As far as we know, they are the same on earth as
anywhere else in the universe, and they are the same now as they
were at all times past and as we expect them to be in the future. We
do not have much of a clue as to why they are constants, or why
each has the particular value that it has. We must simply accept
them as expressing some profound fundamental properties of the
natural world.
11 Particle-waves: probability
Let us go back now to the electron that behaves sometimes like a
particle and other times like a wave. I talk about the electron here
just to have something specific in mind; the picture applies equally
for any other particle. To think of the electron as a particle does not
present us with any problem; we may imagine it to be like a very tiny
speck of dust moving about. At any given time we can say that it is
there and has that momentum. At the same time it has a wavelike
nature that shows up in the right kind of experiment. This wave has a
frequency f a wavelength X (or wave number k), and an PARTICLE-
WAVES: PROBABILITY

amplitude A. We saw that the frequency and wavelength are

connected with the energy and momentum through the Planck
constant h. We now want to give a meaning to the amplitude. In so
doing, we discover a startling new aspect of nature that is hidden
from us in our everyday experience.

Let me consider an electron that is travelling in a straight line with

some momentum p. Associated with this electron is a wave of
wavelength equal to hip, and an amplitude A. If I multiply the
amplitude A by itself, I get the number A2, which is always positive
or zero. If I now do an experiment to try to find if the electron is in
some particular part of its path, I find that it has a probability of being
there which is proportional to the value of A2 there! The larger the
value of A2, the more likely is it to be there; and if A2 is zero in some
region, the electron will definitely not be found there.

Let us pause a bit and see what exactly is meant by probability in

this context. If a coin is tossed many times, it will come up heads in
half the tosses and tails in the other half. We say then that the
probability that a given toss of the coin gives heads (or tails) is a half.
Of course this does not mean that there is any uncertainty about the
result after the toss has been made: it will definitely be heads, or
definitely tails. The idea of probability as we use it here has meaning
only when applied to the results of a large number of tosses of the
coin. It is in this sense that we talk about the probability associated
with an electron wave. I imagine that I do a very large number of
experiments one after the other, all under identical conditions, to find
where the electron is. I shall then find that the electron is seen in
more experiments in a region where A2 is large than in a region
where it is small. In the limit, if the number of experiments becomes
infinite, then the ratio of the two values of A2 equals the ratio of the
numbers of times the electron is found in the two regions. If the
amplitude is zero in some region, then I shall never find the electron
there.

With this interpretation of the amplitude A of an electron wave, first

proposed by the physicist Max Born, we have left the world of
everyday experience and entered the quantum world. To illustrate
the difference between the two, let me take an example from each
world: a train that I know to be somewhere between two points P and
Q^

and an electron that I also know to be somewhere between those

two points. I know that I can find out exactly where the train is at any
given time, if I want to; probability just does not enter the picture.
Things are quite different with the electron. Since I know definitely
that the electron is not outside the region between P and QJts wave
in this outside region must have zero amplitude. Inside the region the
electron has a wave associated with it, whose precise shape we
shall come to in a moment, and whose amplitude A varies as we go
from P to Q^We cannot even in principle talk about the electron
definitely being in some part of the region between P and Q^but only
of the probability of finding it there, a probability proportional to the
value of A2 there.

Quantum mechanics thus tells us that we can only determine the

probability of the result of an experiment to locate the position of an
electron. This is a profound break with the way we have been used
to thinking about the natural world around us. The statement ‘the
electron is definitely there’ is replaced by ‘there is a probability that I
53

Ill PARTICLES AND WAVES

shall find the electron there if I look for it’. You will note that this
probability refers to an actual effort, a set of experiments, to find out
where the electron is.

Some of you might begin to wonder what this quantum-mechanical

way of looking at things does to our intuitive ideas of the reality of the
world around us. What we were used to thinking of as certainties, we
find we should now think of as probabilities.

Since the advent of quantum mechanics in the early part of the

twentieth century, such questions have engaged many philosophers
of science, and some physicists too, Albert Einstein and Nils Bohr
among them. Most physicists, however, take a somewhat pragmatic
stand on such questions. Their attitude is roughly the following:
quantum mechanics is a description of nature that works correctly
everywhere it is applicable, and leads to predictions that are
subsequently verified; and that is all that one can ask of a theory in
physics. This attitude is in tune with my purpose in writing this book,
namely simply to introduce you to the quantum-mechanical
description of the properties of matter and to resist the temptation to
dwell on philosophical implications.

The quantum-mechanical interpretation of the square of the

amplitude as a probability immediately leads to a dramatic new effect
that is the justification for the name quantum mechanics. I describe
this effect in the next section.

12 Particle-waves: the quantization of energy; the wave function

Suppose I want to find out a bit more about an electron as it is
moving: what its energy is, for example. If all I know about its
location is that it is somewhere, then I would not have an easy time
finding it. It could be anywhere in the universe. Common sense tells
me that I would do better to trap it in a box from which it cannot get
out, so that I know that it is definitely somewhere inside the box,
although I can only talk about the probability that it is in any
particular region in the box. If now I can find out what the wave is
which is associated with this electron, then I can hope to learn
something about its motion. I remind you that there are relations (p.
51) between wave properties like frequency and wavelength on the
one hand, and particle properties like energy and momentum on the
other.

I take a very simple example that leads to a startling new result,

namely the quantization of energy. I mean by this that the energy
can only have any one of a set of discrete values, but not anything
that lies between them. For my example, I imagine that I have
trapped an electron between two parallel walls P and Q_which it
cannot penetrate, and that I have further arranged matters so that it
can only move back and forth along a straight line and cannot stray
from this line (fig. III-9). L is the distance between the walls. I
summarize all this by saying that the electron is trapped inside a
one-dimensional box of length L.

FIGURE III-9. A one-dimensional box of length L extending from P to

Q^An electron is trapped inside the box, and all we know is that it is
somewhere between P and Q^

54

PARTICLE-WAVES: THE QUANTIZATION OF ENERGY; THE

WAVE FUNCTION

FIGURE 111-10. Four of the possible

standing waves which can represent

an electron trapped in a one-dimensional box. The wave labelled

(1) has the longest wavelength, (2)

the next longest, and so on. Note that

no in-between wavelengths are

possible. Each wave is to be imagined

as oscillating between the continuous

and broken lines depicting it.

distance

If I now look to see where the electron is at any given instant of time,
one thing is certain: I shall not find it outside the box. If there is no
electron, there is no wave. So the amplitude of the electron-wave
outside the box is zero, and it remains zero right up to the two walls.

Since the electron is definitely somewhere inside the box, its

amplitude inside cannot be zero. However, its amplitude on the two
walls must be zero, the same as outside the box, because the
amplitude of a wave cannot change abruptly. I now try to draw a
picture of such a wave, and I find that there is more than one
possibility, depending on how many crests (and troughs) of the wave
are fitted inside the length L. Further, I find that there is a maximum
possible wavelength which can be fitted inside the box,
corresponding to just one crest (or trough). Other successive
possibilities with decreasing wavelengths would be one crest and
one trough, two crests and one trough, two crests and two troughs,
…, all fitting inside the box. Figure III-10 illustrates these waves.

You must imagine that each wave is vibrating between the positions
shown by the continuous and broken lines somewhat like the
vibrations of a plucked string in a guitar.
There is an important point to note here: the wavelengths can only
have those values which permit integer numbers (1, 2, 3, …) of
troughs and crests to fit inside the box.

At this point, you may have noticed that the waves I have just
illustrated do not look quite like the waves that we encountered
earlier. Those earlier waves were travelling somewhere with some
velocity, while these waves are trapped inside the box, and are not
going anywhere. Such a trapped wave is called a standing wave, to
distinguish it from the other sort which is called a travelling wave. A
standing wave results from adding together two travelling waves of
the same frequency and wavelength, which are travelling in opposite
directions. This is illustrated in figs. Ill-11 and III-12. The two waves
can be thought of as arising from reflections at the ends P and Q_of
the box in fig. III-9. An everyday example of a standing wave is the
way the strings in a musical instrument like a piano or a guitar
vibrate.

Thus we see that I cannot fit a wave of any arbitrary wavelength

inside the box; it can only have certain wavelengths that are fixed by
the length of the box, and none 55

Ill PARTICLES AND WAVES

FIGURE III-11. A wave travelling to

the right in a one-dimensional box,

and the reflected wave travelling to

the left, as indicated by arrows. Each

wave is shown at the same three

successive instants of time labelled 0,

1, and 2. The wavelength is such that

the two added together result in a

standing wave, as in fig. Ill-12.

FIGURE m-12. What the standing

wave, produced by adding together

the two travelling waves of fig. III-l 1,

looks like at the same three instants

of time 0, 1, and 2

distance

i ..-•‘V——O\

itude

Cu 7 X

CO V

>V

—0

—1

-■_•_■ 2

A :-

distance

that lies in between them. We recall that the wavelength, or

equivalently the wave number k, is related to the energy E of the
electron by the formula r h2k2

h~ 2m ■

So we conclude that the energy of the electron can take only certain
discrete values, and none in between. In such a case, we say that
the energy is quantized. We refer to these values of the energy as
the energy levels of the electron in the box. This result is entirely due
to the wave-particle duality of the electron. If the electron were to be
considered purely as a particle in the box, then it could have any
desired value of the momentum and energy; there would be no
quantization. The smallest energy it could have would be zero, when
the particle is just sitting motionless somewhere in the box.

With the quantization of energy, we have arrived at a fundamental

result of quantum mechanics, which sets it apart from our everyday
understanding of the world around us. The electron cannot sit still; it
is forced to have a minimum energy, corresponding to the longest
possible wavelength, and it can also have only a set of discrete
higher energies. The energy is quantized.

PARTICLE-WAVES: THE QUANTIZATION OF ENERGY; THE

WAVE FUNCTION

We shall now do a simple calculation to get a formula for these

allowed energies, the energy levels. The need to fit a whole number
(1, 2, 3,…) of troughs and crests into the box means that the
wavelengths cannot be made to be anything we wish, but can only
have certain values that are fixed by the length L. The wavelength is
the distance occupied by a crest and its neighbouring trough. Now
look at the longest wave, labelled (1), in fig. Ill-10. There is just one
crest (which with time alternates with a trough because of the
vibrations of the wave), and its length is L, which is also the length of
its trough.
So its wavelength is equal to twice L, or 2L. Its wave number, which
is just the reciprocal of its wavelength, is then 1/(2L). This is the
smallest wave number that the electron-wave can have in the box,
and I shall denote it by the symbol k^.

kl= JL

where the subscript 1 in kt reminds me that I am thinking of the wave

number of the wave labelled (1) in fig. Ill-10.1 now come to the wave
labelled (2). I see from the figure that its wavelength is exactly equal
to L, and therefore its wave number, which I shall call k2, is equal to
l/L, or

For the wave labelled (3), I see from the figure that there are \
wavelengths in the distance L, so that its wavelength is fl, and its
wave number k3 is question What is the wave number kt of the wave
labelled (4)?

answer There are exactly two wavelengths in the length L, so that

the wavelength is 1/2, and so the wave number kt is 2/1, which I can
write as . = ■*,.

I see a pattern beginning to emerge: the wave numbers of the

successive waves which fit inside the length L are ku 2ku 3ku 4klr
…. For the n-th wave, where n is an integer which has any one of the
values 1, 2, 3, 4, …, the wave number, which I shall denote by km
will be n times kt:

kn = nku where k{= yr ■

I have introduced in the foregoing a new symbol, kn. I read it as ‘k

sub n’, to remind me that n is written a little bit below k, i.e. n is a
subscript. If n = 1, then I read it as ‘k sub 1’, and so on. You can see
that with the one formula that I have just written for kn, I do not need
a separate formula for the wave number of each wave. I now have a
useful shorthand when I wish to make a statement involving the
wave number, which applies to all the waves. For example, suppose
I want to find the energy of each of these electron-waves. I assume
that there is no potential energy, so that the total energy is just equal
to the kinetic energy, which is equal to

h2k2

2m

57

Ill PARTICLES AND WAVES

for an electron of mass m and wave number k. Let En be the energy

of the w-th wave with wave number kn. Then we have

h2(kn)2

tn~ 2m ■

We saw above that kn is just n times ku and so we can write h\nkx)2

En~ 2m •

Now, using the value for kx of 1 /(2L), we get

h2n2

E”=8mL2

We get the energy Ex of the wave with the longest wavelength by

putting n equal to 1: h2

£l= 8mL2

and we can now write the energy of the «-th wave as

question We said that the matter waves are also vibrating. What is
the frequency fn of the wave whose energy is £„?
answer Since the energy is given by the Planck constant h multiplied
by the frequency, the frequency fn equals En divided by h.

Let me state in words our result:

The energy of an electron in a one-dimensional box is quantized.

The smallest energy it can have, denoted byElr cannot be zero but
has a finite value that becomes larger as the length of the box
decreases. The other values of energy that it can have, namely its
energy levels, are given by multi-plyingEi by n2, where n is given
one of the values 2,3, 4, …. These are the only possible values of
energy that the electron can have, and they belong to the allowed
quantum states of the electron.

The lowest energy that the particle can have is called its zero-point
energy, and is a direct consequence of its wave nature. If it were
simply a particle, its lowest energy would be zero and it would be
sitting motionless. This however is not allowed because of its wave
nature: instead, it has a zero-point motion which results in its zero-
point energy.

So much for the energy of the electron in a one-dimensional box.

How about its momentum? We saw that the momentum vector p is
given by where k is the wave vector. In the present case, there are
two waves with the same wavelength that are travelling in opposite
directions. The corresponding wave vectors are therefore +k and —k
respectively, and the probable momentum that we would measure is
zero.

I have taken the example of an electron in a one-dimensional box, to

keep things PARTICLE-WAVES: THE QUANTIZATION OF
ENERGY; THE WAVE FUNCTION

simple while bringing out the important consequences of quantum

mechanics. The world around us, however, is three-dimensional: we
shall later, for example, want to talk about electrons in a real crystal.
It turns out that the basic features of the quantum description still
hold. The energies of the electrons can have only discrete values
that are determined by the size and nature of the crystal. Also, the
dependence of energy on the wave number becomes more
complicated than for the simple case that we have considered
above. This is because the electric potential energy of the electrons,
which have an electric charge, must also be taken into account. The
physicist uses a mathematical equation called the Schroedinger
equation, named for Erwin Schroedinger who first wrote it down, in
such cases. If I know the conditions under which the electron is
moving, for example that it is moving back and forth inside a one-
dimensional box of length L, then I can write down the Schroedinger
equation describing its motion, and from it I can calculate the
amplitudes and energies corresponding to the different possible
waves associated with the motion. What I have done above with
words, this method allows me to do with mathematics. The
Schroedinger equation also permits me to calculate the electron’s
energies in more complicated cases, such as when it is in a
hydrogen atom or in a crystal.

We shall not do such calculations in this book, but shall look at some
of the results in later chapters, and see that they are what we would
have expected from the understanding of quantum mechanics that
we are now acquiring here.

I said that the square of the amplitude, A2, at any point is

proportional to the probability of finding the electron there. I show in
fig. Ill-13 what this quantity looks like for the four waves labelled (1)
to (4) in fig. Ill-10. If I were to think of the electron as simply a
particle, I would say that it has the same probability of being
anywhere inside the box. When I use the quantum description of the
electron, I find by looking at the figure that the electron is more likely
to be found in some parts of the box than in others. For example, for
the wave labelled (1), the electron is most likely to be found at the
centre of the box, and progressively less and less likely as I move
away from the centre in either direction. It will definitely not be found
at the two ends of the box or outside it, FIGURE 111-13. The square
of the
amplitude at a point of the wave

describing a particle gives the relative

probability of finding the particle in

the vicinity of the point. This

probability is shown here, labelled

the same way, for each of the waves

which describe an electron in a one-dimensional box (fig. Ill-10).

Note

that the probability is not at all

constant everywhere, but varies

between a maximum and zero.

distance

59

Ill PARTICLES AND WAVES

where A2 is zero. So we have here yet another example of how the

quantum world is totally different from the everyday world. I should
remind you of the meaning of probability here: it refers to the
different outcomes of a very large number of experiments that one
imagines to be done to see where the electron is in the box.

Some of you may have begun to wonder what these seemingly

strange ideas have to do with our usual sense of reality. Let me
assure you that innumerable experiments have shown that quantum
mechanics is the correct description of nature. It provides the
answers to the kind of questions that were posed at the beginning of
chapter I, as you will see in later chapters. We must simply accept
that this is the way nature is at the microscopic level of atoms and
electrons and nuclei, strange as it may seem compared to what our
senses tell us about the macroscopic world around us.

We saw that an electron confined in an enclosure is associated with

a wave that has some amplitude at every point inside the enclosure.
I can label each of these points by using vectors, which you recall
are things that have a magnitude and a direction. I first fix one point,
which I call O (for origin). Then I take the point P which I want to
label, and imagine a straight arrow whose tail is at O and tip at P.
This arrow has a magnitude, namely its length, and a direction, from
O to P. So I can represent the arrow by a vector; I call it r, the
position vector of P with respect to O. By changing the length and/or
the direction of the arrow, which is the same as saying that the
vector r is a variable, I can cover all points in the enclosure. I use the
Greek letter \y (pronounced psi) as the symbol for the amplitude of
the wave, and write the amplitude at the point whose position vector
is r as y(r). This way of writing is to remind me that V|/ is a function of
r, i.e. the value of \|/ can be different at different points in space. I call
\j/(r) the wave function. This single symbol \|/(r) stands for the
amplitude of the wave at every point in the enclosure holding the
electron. I think you can appreciate the usefulness and economy of
this notation. The Schroedinger equation, which I mentioned earlier,
allows me to determine the wave function and the energies of the
electron. I must also keep in mind that the wave function is
oscillating in. time at each point in space, the frequency being equal
to the energy divided by the Planck constant h.

For an electron in a one-dimensional box different waves are

possible, and we labelled them n, where n takes on one of the
values 1, 2, 3, 4, …. Each of these waves can be described by a
wave function. We can write the wave function for a given value of n
as V|/n(x), where x is the distance measured from one end of the
box. So the value of x can vary from zero to L where L is the length
of the box. The corresponding energy level will then be En. The
number n, which first appeared in the formulas for the wave number
and energy (p. 57) and with which I have now labelled the different
wave functions and energy levels, is called a quantum number. As
each value of n corresponds to a particular value of the wave
number kn, I can equally well use the kn as the quantum numbers to
specify the different wave functions. Quantum numbers are a
characteristic feature of quantum mechanics, and we shall see other
examples in later chapters. We see now that the waves shown in fig.
Ill-10 are just pictures of the wave functions for the first four quantum
states of an electron in a one-dimensional box.

6o

SUMMARY

Now some mathematics for those of you who remember

trigonometry. The wave functions for the electron in a one-
dimensional box are = A sin \n L I, with n = 1, 2, 3, 4, …,

as you can verify from fig. Ill-10. If this formula looks opaque to you,
just skip it and carry on!
13 Summary
This chapter contains a number of new ideas and ways of looking at
things. I shall give you now a summary of the principal points in the
chapter, section by section. It may be helpful for you to look up the
relevant sections as you go through this summary.

I look in sections 1 and 2 at the world around us, and note that
everything is in a state of motion. I find that I can put the things that
are moving into one of two classes: they are either particles, or they
are waves. A particle is an idealization of our everyday sense of a
moving object: it has mass and occupies a tiny volume. We think that
electrons and nuclei might be good examples of particles, because
they have a mass and they are certainly very small. We shall find
later that this is not the whole story. The idea of a wave is easy to
grasp: we have all seen waves on the surface of water.

In section 3 I look in more detail at the motion of a particle. Besides

its position and mass, the moving particle has velocity, momentum,
and energy. The energy comes in two forms, kinetic energy which is
due to the motion, and potential energy which requires some special
trick to convert it into kinetic energy and thus to motion. I introduce
the idea of a vector, which is a quantity that has both a magnitude
and a direction, and denote it by a symbol in heavy type. I call a
quantity that has only a magnitude, but no direction associated with
it, a scalar. Momentum and velocity are examples of vectors,
whereas mass and energy are examples of scalars.

Section 4 is a brief diversion from the main course, to introduce you

to the conservation of momentum and energy. These conservation
laws form the fundamental basis for all of physics. They are a
“consequence of the nature of space and time.

I describe the characteristics of waves in section 5. A wave has

amplitude, wavelength, &nd frequency. Unlike a particle which is
localized in space, a wave is spread over space.

It is travelling in some direction with some velocity. A wave is

diffracted around an obstacle in its path, and continues onwards.
Two waves that come together interfere with each other and produce
a new wave.

In section 6, I recapitulate the characteristics of particles and waves,

and present a table that emphasizes the distinction between them. It
appears that in the world we see around us, motion can be
unambiguously assigned either to particles or to waves, with no
overlap between the two.

This picture is dramatically changed when we look at the world on

the scale of an electron (or proton, neutron, etc.). Depending on the
kind of experiment I do on it, an 6i

Ill PARTICLES AND WAVES

electron may move like a particle or like a wave, with the

corresponding features shown in table III-1. This wave-particle
duality is the subject of section 7.

We now come to section 8, and find that nature has not yet finished
with its surprises for us. Light as well as other forms of wave motion
shows particle-like or wavelike properties depending on the specific
way in which they are observed. The particles of light are called
photons. Wave-particle duality thus turns out to be a property of all
nature.

Section 9 contains some general remarks on this duality, and points

out that we can still continue to think of particles and waves as
distinct entities on the scale of the macroscopic world.

The quantitative description of wave-particle duality is called

quantum mechanics. I introduce you in section 10 to the basic
formulas of quantum mechanics relating the wavelike and particle-
like properties of the electron (or proton, neutron, etc.) to each other.
Exactly the same formulas hold between the wavelike and particle-
like properties of light. You may think of this as a manifestation of
some deep symmetry in nature.

I continue the account of quantum mechanics in section 11, and

describe how the amplitude of the wave is related to the probability
of finding the particle there. The amplitude multiplied by itself, which I
call the square of the amplitude, is proportional to this probability.

I illustrate in section 12 this interpretation of the amplitude by looking

at an electron trapped in a one-dimensional box. I find that the
electron can only have any one of a set of values for its energy, and
nothing in between. Furthermore, its lowest possible energy cannot
be zero. This means that the electron cannot be sitting still, but must
necessarily be moving with at least this minimum energy. This is one
more example of the unexpected consequences of wave-particle
duality. I conclude the section by describing what is meant by the
wave function for the electron.

This brings me to the end of my summary. In the chapters to follow I

use quantum mechanics to describe a crystal, and see how I can
then understand its properties. A crystal is composed of a very large
number of atoms. We have so far looked at the quantum mechanics
of an electron in a box, but this is very different from an atom, let
alone a whole crystal. The simplest atom we have is the hydrogen
atom, with just one proton and one electron. I shall show you in the
next chapter what quantum mechanics has to say about the
hydrogen atom. The resulting picture, with some modifications,
applies to other atoms also. We shall then be closer to being able to
describe a crystal in quantum mechanics.

62

IV THE ATOM
l Introduction

I give in this chapter the quantum-mechanical description of an atom.

I look in some detail at the hydrogen atom, which is the simplest
atom we have, with just one proton and one electron. From the
picture that emerges, we shall be able to see how to describe all
other atoms. Before continuing, I need to explain what I mean when I
say that the proton and electron have equal and opposite electric
charges. It is because of these charges that the atom is stable, and
does not spontaneously break up into the two separate particles.
2 Electric charge and potential energy
If I take a ball in my hand and then let go, it drops to the floor. This is
because of the gravitational interaction between the ball and the
earth. The ball had gravitational potential energy when I held it, and
when I release it, this energy is progressively converted to kinetic
energy as it falls. Gravitation makes two bodies move towards each
other, and not away from each other: the gravitational force is always
attractive.

As a result of many experiments done over the years, we find that

there is another kind offeree in nature, somewhat analogous to, but
in one aspect different from, gravitation. I take two idealized
examples, and use them to show the features of this force.

First I consider two electrons that I imagine to be pinned down a

distance r apart. If I now unpin the electrons so that they are free to
move, I find that they fly apart, moving away from each other. There
appears to be a repulsive force between them, which therefore
cannot be due to the gravitation (which attracts them to each other)
between the two masses. Just as the source of gravitation is the
mass of the electron, there must be a source for this new force: I call
it the electric charge, or simply the charge, of the electron.

I should now like to make a small detour, and say a few words that
will give you an insight into some of the ways of physics. Having
been told that an electron has something called charge, you might
be tempted to ask, “Well, what exactly is it? Is it something that
makes the electron glitter, or colours it blue, or something like that?’

The answer is, none of the above. It is just a compact way of saying
that the electron has a property that makes it exert a force, which
has nothing to do with its mass, on another electron. The reasoning
goes as follows:
1. We observe that two electrons repel each other. This is a given
fact of nature.

2. This repulsion cannot be due to the force of gravitation between

the electrons, because then they would be attracted towards each
other.

IV THE ATOM

3. So there must be another property of the electron, besides its

mass, which is responsible for this force. I call this property its
charge.

4. I need to be able to measure the charge — remember, physics is

a quantitative science. So I need a unit and a method of
measurement. I think of the analogy with measuring mass, where I
have a set of standard weights labelled 1 gram, 5

grams and so on, and I find the mass of a bag of potatoes, say, by
comparing the gravitational force on the bag with the force on these
weights. This last step is essentially what a spring balance does. So
with charge: I can make the charge on the electron my unit of
charge. A superb unit it is too, for it is exactly the same for all
electrons, and never changes with time. Then I can measure any
other charge by finding the electric force on it due to an electron. In
practice, such measurements are made by more indirect methods,
but the underlying principle is as I have described it.

To conclude: when you think of electric charge, you must understand

that it is just a name given to that property of the electron which is
responsible for the electric force.

This brings me to the end of my detour.

From the discussion of energy in chapter III, I know that the two
electrons had potential energy when they were pinned a distance r
apart, and this energy is converted to kinetic energy when they are
released. Further experiments show me that the larger the distance r
at which the two electrons are pinned, the smaller the potential
energy that they have.

Now, I repeat all these experiments after replacing the two electrons
by two protons. I find that I get exactly the same results for the
potential and kinetic energies as before. Can I conclude that the
electron and the proton must have exactly the same charge? Not
yet! I do a third set of experiments, this time with one electron and
one proton. Almost everything is as with the first two sets of
experiments, but with one important difference: when set free, the
electron and proton move towards each other and not away from
each other. The force is attractive, and not repulsive as it was in the
two previous experiments. It thus appears that the charge on the
proton is similar to, yet essentially different from, the charge on the
electron. Similar because two protons and two electrons have the
same magnitude of force when they are the same distance apart.
Different because two protons (or two electrons) repel each other,
whereas an electron and a proton attract each other.

I can summarize all this by saying that a proton has a positive

charge + e and an electron has a negative charge —e, where eis
4.80 X10”10 (using the cm, gram, and second as the units of
measurement), and that two like charges (i.e. both positive or both
negative) repel each other and two unlike charges attract each other.
We have also talked earlier about another particle called the neutron.
It has almost exactly the same mass as the proton, but no charge at
all. Neither the electron nor the proton exerts any electric force on
the neutron. I say that the neutron is electrically neutral, it has zero
charge.

If you take two objects and stick them together, you get the mass of
the resulting object 64

ELECTRIC CHARGE AND POTENTIAL ENERGY

by adding the masses of the two. It is the same with charges; only,
you have to be careful with the plus and minus charges. For
example, the total charge on a hydrogen atom is got by adding the
charge of the proton, +e, to the charge of the electron, -e, resulting in
a total charge of zero. The hydrogen atom is neutral. This electric
neutrality is a property of all atoms in nature. The total negative
charge of the electrons in any atom is exactly balanced by an equal
positive charge of an equal number of protons in the nucleus of the
atom. The nucleus has some neutrons too, but these have no
charge. This neutrality of atoms is why we do not usually see electric
forces in everyday life. There are some occasions, however, when
some electrons are separated from their atoms, and then these
forces appear. This is what happens when you sometimes find your
hair crackling and standing on end when you comb it. The friction
between the comb and the hair causes some electrons to detach
themselves from the hair and go to the comb. So the hair is left with
some ionized atoms, i.e. atoms that have lost some electrons and
are therefore positively charged. Neighbouring hairs, both positively
charged, repel each other, and so the hairs fan out and stand up.
Some of the separated electrons jump back to where they came
from, producing sparks that cause the crackling you hear.

question If an atom of carbon loses an electron, what is the charge

on the resulting ion?

answer Initially the atom had six protons in the nucleus with a charge
+6f (and also some neutrons with no charge) and six electrons with
total charge -6eand was therefore neutral. The ion has five electrons
with charge—5 c and so its net charge is +6e minus 5e, or just +e.

We saw that two electrons held a distance r apart have electric

potential energy, which is progressively converted to their kinetic
energy of motion when they are let go. There is a formula for
calculating the potential energy of two charges qx and q2, say, held
at a distance r apart. If the two charges are an electron and a proton
respectively, then qt is —eand q2 is +e. The energy is given by
multiplying the two charges (remembering the plus and minus
signs!), and dividing the result by r. If I call this energy V, I can write
for short

qiqi

V—

Therefore the potential energy of the electron and the proton held a
distance r apart is —e times +edivided by r, which is equal to —^.
You remember that a negative number multiplied by a positive
number gives a negative number.

question What is the potential energy of two electrons at a distance r

apart?

answer I multiply the two charges together, and divide by r. The

answer is —emultiplied by-eand divided by r, or +-y-Two electrons,
or two protons, have a positive potential energy and repel each IV
THE ATOM

other. An electron and a proton have negative potential energy and

attract each other.

This is an example of the connection, which holds generally,

between the plus or minus sign of the electric potential energy and
the repulsion or attraction.

The electric charge —e is the same for all electrons, everywhere and
always. The same is true for the charge +eof the proton. The charge
is an intrinsic and unchanging property of the particle, just like its
mass. Any electric charge that we might meet in nature, since it must
be made up of electrons and protons, can only have a value that is
an integer multiplied by e, and nothing in between. I can have a
charge of 5 times e, or 6 times e, but not, say, 5.5 times e. I can say
that electric charge is quantized, and the quantum of charge is e. I
should mention here that protons and neutrons are believed to be
made up of particles called quarks,swhich have positive or negative
charges of magnitudes \eox\e. This internal structure of protons and
neutrons has no effect on those properties of matter which are
discussed in this book, and can therefore be ignored.

3 The hydrogen atom: why is it stable?

The hydrogen atom consists of a positively charged proton and a

negatively charged electron, and we have seen that these two
charges, being opposite in sign (plus and minus), will attract each
other. So you might well ask why the electron is not sucked into the
proton so that one ends up with a particle with zero charge,
something like a neutron. We know that hydrogen atoms are quite
stable in nature. Why is this so? The answer lies in the wave nature
of the electron. Consider what would happen if the electron did fall
into the proton and become trapped inside it. We know from many
experiments that the proton is like a tiny sphere with a radius of
about 10~13 centimetre. We recall from the last chapter (page 58)
that when an electron is trapped inside a box, it must have a
minimum kinetic energy that cannot be zero, and which becomes
larger the smaller the box is. Well, if I try to trap the electron in a box
that is as small as the proton, it will have a kinetic energy that is so
great that it would come shooting out of the proton at once! When it
is outside the proton, it still feels the attractive force due to the
proton, and therefore cannot wander off to a great distance away. It
ends up by staying somewhere near the proton all the time, and the
result is the stable hydrogen atom.

Many of you are familiar with a picture of the atom that often appears
in the popular press, and on banners at demonstrations about
nuclear weapons, nuclear power stations, and so on. This picture
looks somewhat like fig. IV-1, with the dark circle in the centre
representing the nucleus, and electrons going around it in elliptical
orbits.

Unfortunately, this picture does not represent reality as we know it,

because it does not take account of the wave nature of the electron.
Quantum mechanics tells us that an electron does not go round in a
fixed orbit like a planet round the sun, but may be found anywhere
near the nucleus with a probability which varies from point to point.
66

THE HYDROGEN ATOM: THE GROUND STATE

FIGURE IV-1. A popular but

misleading representation of an atom,

as a nucleus surrounded by electrons

going around it in various orbits. The

correct picture, as given by quantum

mechanics, is indicated in fig. IV-3.

We have to find out what the wave function of the electron looks like
when it is near the proton. Then we can find out how the probability
varies from point to point and what the energy of the atom is. This
we now proceed to do.

4 The hydrogen atom: the ground state

We saw in chapter III that an electron that is confined to a box has a

lowest energy state, which is the one with just half a wavelength
fitting inside the box. Similarly the hydrogen atom, with the electron
confined to being somewhere near the proton, has a state of lowest
energy. This is a general feature of any quantum-mechanical
system. We call this state of lowest energy the ground state of the
system. The physicist finds the properties of the ground state, and of
the higher energy states, by solving the Schroedinger equation for
the atom. We are not going to do it here. Instead, I shall show you
how we can use the ideas about wave functions and probability that
we learned in chapter III to get a qualitative picture of the ground
state.
Look at fig. IV-2, depicting a hydrogen atom. The dark blob at the
centre represents the proton, and the electron has a probability of
being anywhere in the grey region surrounding it. The region extends
in principle out to infinity, although the probability of finding the
electron rapidly decreases as one moves away from the proton. Any
point P in this region can be specified by the vector r connecting it to
the proton. Each 67

THE HYDROGEN ATOM: THE GROUND STATE

FIGURE IV-3. The relative

probabilities of finding the electron in

any one direction at different

distances from the proton in a

hydrogen atom in the ground state.

The horizontal axis is divided into

ranges lying within 0 to 0.1

angstrom, 0.1 to 0.2 angstrom, ….

For example, the segment labelled 0

is the range from 0 to 0.1 angstrom.

0 0.5 1 1.5 2

distance (in angstroms) from electron to proton

again of the meaning we attach to probability in quantum mechanics.

Suppose I take a hydrogen atom and do many experiments on it to
find how far the electron is from the proton. The heights of the
relevant bars in fig. IV-3 tell me that for every experiment in which
the electron is in some direction 0.6 to 0.7 angstrom away from the
proton, there will be 10 experiments in which it is in that direction
less than 0.1 angstrom away.

So, instead of the popular picture of the electron going around the
proton in an orbit, we now have the quantum-mechanical picture of
the electron smeared out, so to speak, in a probability cloud
extending out to a distance of about an angstrom from the proton.
The energy of the proton with this electronic cloud is the ground
state energy of the atom. Using the Schroedinger equation, one
calculates this energy to be about -13.6 electron-volts. A word is in
order here about why the energy is negative. I have assumed that
when the electron and proton are very far apart, their energy is zero.
The negative sign then means that the energy of the ground state is
lower than the energy of the two particles when very far apart. If now
I want to break up the atom by taking the electron very far away from
the proton, I must feed an energy of 13.6 eV to the atom, because
+13.6 added to -13.6 gives me zero. This is an application of the
principle of conservation of energy.

I have introduced here a new unit with which to measure energy, the
electron-volt.

You are familiar with the fact that the electricity in your home is
supplied at about 230

volts, or at about 115 volts if the home is in north America. You can
think of volts as the pressure difference that drives the electrons to
flow in a current, for example through a light bulb. The electron-volt
is the energy that an electron acquires when it moves under a
pressure difference of one volt. The electron-volt, abbreviated to eV,
and the angstrom which I introduced earlier, are commonly used
when talking about things on the atomic scale, for the same reason
that we talk about distances between cities in miles or kilometres
rather than in inches or centimetres: they fit more comfortably the
scale of the things which are being measured.
69

IV THE ATOM

Let us look at fig. IV-2 again. Every point on the sphere is at exactly
the same distance r from the proton. In the ground state of the atom,
the wave function changes only when r changes, and therefore it
must have the same value at every point on the sphere. Now look at
the curve, labelled C, which represents a circle girdling the sphere:
its equator, so to speak. Suppose I start at some point on it, and go
all the way around the equator till I am back to the same point. Since
every point on the equator is at the same distance from the proton,
the wave function in this state must have the same value at all these
points. I have illustrated this in fig. IV-4. You are looking in this
picture at the proton with the equator around it. To help you get a
sense of depth, I have marked the front and back of the equator.
There is an object like a pin at each of several points on the equator.
The height of the pinhead at each point represents the amplitude of
the wave function, whose square is the probability of finding the
electron there. All these heights are the same, and you should
imagine that the probability is the same between the pins also. Again
to help you with the perspective, I have shown the pinheads in the
front as dark and in the back as clear. What we have here is a wave
of constant amplitude around the proton. It is certainly not at all like
the waves we see on water, or, for that matter, the waves associated
with the electron in a one-dimensional box (fig. III-10). It may help to
think of it as a wave with an extremely long wavelength which is
tending to infinity.

The probability in the ground state depends only on how far the
electron is from the proton, and not on which direction. So we
conclude that we find the same probability if we go along any line of
latitude or of longitude, so long as it is at the same distance r from
the proton.
0o
o
Q

back

proton

front

FIGURE IV-4. A depiction of the wave function of the ground state

for the electron in the hydrogen atom. Imagine a circle of some
radius rat the centre of which sits the proton. The wave function has
the same amplitude at every point on this circle, and this value is
suggested by the height of the pins stuck around on the circle. The
pinheads in the front are shaded dark, to help in the visualisation of
depth in the picture: the usual problem of showing three dimensions
on a sheet of paper.

70
THE HYDROGEN ATOM: EXCITED STATES

proton

60

—■ o-
111
front
FIGURE IV-5. The wave function for an excited state in the hydrogen
atom, going round a circle of longitude with the proton at the centre.
Its amplitude is indicated by the height of the pinhead above the
circle. The pinheads in the front are shaded dark. This looks more
like what one expects from a wave: it has two oscillations in going
round the circle.
5 The hydrogen atom: excited states
We now come to the excited states, which are the states with
energies higher than the ground state energy. We find here that the
probability often depends not only on the distance of the electron
from the proton, but also on its direction. When I work out the
Schroedinger equation, I find that the resulting wave functions (and
consequently the probabilities) cannot take any arbitrary form, but
can only be one of a discrete set, each with its value for the energy:
the energy is quantized. These wave functions, unlike the one for the
ground state, have a recognizable wavelike character.

I show in fig. IV-5 how the wave function looks for one of these
states, if I follow it along a line of longitude. Since the wave must join
smoothly on to itself after going full-circle, its wavelength cannot be
any arbitrary value.

question Why must the wave join smoothly on to itself after going
full-circle?

answer Because the probability of finding the electron near any point
in space has a unique value in a given quantum state, and cannot
change abruptly from point to neighbouring point.

The example I have chosen has exactly two full wavelengths in

going around the circle. A similar condition holds if I go around any
other line of longitude for this wave IV THE ATOM

function. You can see an analogy with the case of an electron in a

one-dimensional box, where also we found that the waves could
have only particular wavelengths because they had to fit exactly
inside the box. It is this feature that is at the root of the quantization
of energy. I can have a wave function with, say, one full wave going
around the line of longitude, or two full waves, but nothing in
between. Each of these two states has its specified energy, which is
one of the discrete energy levels allowed for the atom.
This is what we mean by the quantization of energy.
6 Energy levels and quantum
numbers
We thus find that the hydrogen atom can be described by any one of
a number of wave functions, each of which corresponds to a
particular energy of the atom. Some of these energy levels are
shown pictorially in fig. IV-6. Let us look first at the right-hand half of
this picture. At the bottom is a horizontal line labelled -13.6eV,
representing the energy level of the ground state. The energy
increases (i.e. becomes less negative) as we go up from this line,
and the next line, at —3.4 eV, stands for the energy level of the first
excited state. Yet higher up is the energy level of the second excited
state, at —1.5 eV.

The energy difference between adjacent levels continues to become

smaller and smaller as we go to higher excited states, which I have
not shown in the picture. Finally we reach the level labelled 0, whose
energy is zero and which represents the electron far away and free
from the proton. This is the ionised state of the atom. Any further
increase in energy would produce motion of the electron and proton
relative to each other.

question How much energy must I feed into the atom in the ground
state to put it in its first excited state?

answer Exactly enough to raise its energy from -13.6eV to -3.4eV,

that is to say 13.6 minus 3.4 or 10.2 eV.

question What happens if the atom changes from the second excited
state (-1.5 eV) to the first excited state (-3.4eV)?

answer The atom loses an amount of energy equal to 3.4 minus 1.5
or 1.9 eV. By the conservation of energy, this energy has to appear
somewhere else: it could appear for example as a photon of energy
1.9 eV.
question Can I get the atom in the ground state to absorb a photon
of 10.0 eV?

answer No. If the atom were to absorb an energy of 10.0 eV, then its
final state should have an energy of-3.6 eV. No state with this energy
exists for the atom, which therefore cannot absorb it.

Let us now turn to the left half of fig. IV-6. We see three columns of
numbers, labelled at the top n, I, and E respectively. The numbers
under E (for energy) are the 72

ENERGY LEVELS AND QUANTUM NUMBERS

FIGURE IV-6. On the right are shown

the quantized discrete energy levels of

the hydrogen atom as well as the

continuum of energy levels when it is

ionized. On the left are three columns

of numbers assigned to the discrete

energy levels. The columns labelled n

and / show the quantum numbers

(explained in the text), and the

column labelled E the energy, for

each level.

n
3

0,1,2

0,1

-1.5eV

-3.4eV

y increa

Energ;

-13.6 eV

ionized atom

energy is positive

second excited state

first excited state

stable atom

energy is negative

ground state

energies in eV of the levels, as we have already seen. The numbers

in the columns under n and are called quantum numbers, and are an
essential feature of the quantum nature of the atom. You are to read
these numbers horizontally. I mean by this that, for example, the
second excited state has an energy of-1.5 eV, its n-quantum number
is 3, and its -quantum number can have the values 0, 1, or 2.

We already met a quantum number when we looked at an electron in

a one-dimensional box in chapter III. The number n that appeared in
the formula for the energy on page 58 is a quantum number. When I
work out the wave functions for a given problem, say the hydrogen
atom, I find that quantum numbers appear in these functions. These
numbers are allowed to have only certain values that are determined
by the conditions of the problem, and they set the allowed values of
the energy and 73

IV THE ATOM

other such quantities. In the example of the electron in the one-

dimensional box, the quantum number n was allowed to take only
integer values 1,2,3, … because the electron wave had to fit exactly
inside the box, and this number in turn fixed the energy of the
electron. I shall now say something more about quantum numbers,
using the ground and first excited states as illustrations.

We see from the figure that the ground state has quantum numbers
n equal to 1 and equal to zero. Its wave function contains these two
numbers, and I can use a convenient shorthand to write it as \|(« = 1,
/ = 0). This is not the end of the story. You remember that the
electron has not only a charge, it also acts like a tiny magnet that
can point in two opposite directions in the presence of a magnet, and
each of these orientations is a different state of the electron with its
own wave function. So let me introduce a new quantum number with
the symbol s for this property of the electron, and give it the value +\
for pointing one way and —\ for pointing the other Way. The reason
for choosing these peculiar numbers is that it then simplifies other
formulas which the physicist uses. Then we can write down the
corresponding wave functions as and

The energy is the same, —13.6 eV, for these two wavefunctions, and
is set by the value of n alone. Now if you are getting a bit alarmed by
the complicated symbol for the wave function I have just used, I ask
you to think of it just as a shorthand way of reminding yourself that
the wave function for the electron has the values of the quantum
numbers shown inside the parentheses.

Thus we find that there are two different wave functions for the
ground state with exactly the same energy. You may think this a little
peculiar; it would be quite a coincidence to have two wave functions
that look different and yet have exactly the same energy. Consider
this example: a square that is four centimetres on the side looks
quite different from a rectangle that is sixteen centimetres by one
centimetre, but they have exactly the same area. So it is not difficult
to imagine that there can be two, or even more, things that are
different from one another in certain ways and yet have one property
that is exactly the same for all of them. Physicists have a name for
this occurrence of two or more wave functions having the same
energy. They say that the energy level is degenerate. I have a friend,
not a physicist, who giggles each time I use the term in a physics
context. She says it conjures up for her an image of a person with
bloodshot eyes and ragged hair: a degenerate. This is one more
example of the special meaning in physics that we attach to many
words of common usage.

You recall that for the electron in a box we had a picture of what the
wave function looked like: it looked like a wave. Things get more
complicated with atoms and even more so when we get to crystals,
so that it will not be always possible to picture their wave functions.
So we shall sometimes specify a state of the system by saying what
values the quantum numbers have in that state, and not worry about
what the wave func-74

ENERGY LEVELS AND QUANTUM NUMBERS

tion looks like. You will now see that this is what I did above when I
wrote down the wave functions for the ground state of the atom.

I have so far introduced three quantum numbers:

n= 1,2,3, … ,

/ =0, 1,2, … ,

and s = +1 -i

We have seen that the value of n in the wave function tells me the
energy level to which it belongs: 1 for the ground state, 2 for the first
excited state, and so on. The two possible values of the quantum
number s distinguish between the two opposite orientations of the
electron-magnet. Now what about the quantum number U It contains
information about another property of the atom, namely its angular
momentum. This property, you remember from chapter III, is
connected with rotational motion. Furthermore, the quantum number
/ contains hidden in it another quantum number that I shall call m. I
explain all this by looking now at the first excited state.

This first excited state, as I see from fig. IV-6, has n= 2, while /can be
either 0 or 1

and 5 can be either —\ or +2. The possible values of the quantum

number m are set by what value has. When is zero, m is also zero.
When / has the value 1, m can take any of the three values +1, 0, or
—1.1 show in table IV-1 the different combinations of possible
quantum numbers.

Table IV-1. Quantum numbers for the first excited state n I m s

__

_I

—2

_i

Each row in the table is a set of possible quantum numbers for the
first excited state.

Each set has its wave function, and no two sets are identical. We
see that there are eight such wave functions, each of them having
exactly the same energy of—3.4 eV. This state is even more
degenerate than the ground state, which had only two wave
functions.

You may have begun to see a pattern emerging about how the
quantum numbers n, I and m are interrelated. It is convenient to
express this pattern using algebraic symbols, and this I do now. For
a given value of the quantum number n, lean take any of the integer
values between 0 and n-l:

1=0, 1,2, …,n-l.

75

2
2

+1

+1

0
0

-1

-1

IV THE ATOM

With each value of/, m takes any of the integer values from +/ to —/:
m=+l, +1-1,…,-/+1,-1.

Finally, with each combination of values of/and m, stakes on the

values —\ and +2.

question How many different wave functions are there for the excited
state with n — 3? And for n — 4?

answer When n is equal to 3, can take the values 0, 1, 2. Then I can

write down the possible values ofand m as follows:

/ = 0 and m - 0;

/= 1 and m= +1, 0, or-1;

/= 2andm = +2, +1, 0,-1, or -2.

I thus have a total of 1 + 3 + 5, or nine possible combinations of/and

m.

Each of these can have s equal to either +2 or —\, so that I end up

with a grand total of two times nine or eighteen different wave
functions for the state with n — 3.

When n — 4, can also take the value 3, besides the above values.
You can use the same method as before to work out that there are
14 wave functions with = 3, so that there will be 14+ 18 or 32
different wave functions for n = 4.
I said that the quantum number is related to the atom’s angular
momentum, and we have seen that m is very closely connected with
. If I did an experiment to measure the angular momentum of the
atom along some particular direction, I would find that it can only
have one of the values given by m multiplied by h and divided by
twice 7t, or more compactly expressed,

mh

271

where n is a number approximately equal to 3.14. You may

remember from your geometry in school that n is the number you get
by dividing the circumference of a circle by its diameter, and is
exactly the same for all circles, whatever their size. With this result,
we now meet a new and extremely important consequence of
quantum mechanics, namely that the angular momentum can only
have certain values related to the Planck constant, and cannot take
on any values that lie in between. The angular momentum is
quantized.

The electron not only has an electric charge — e and behaves like a
tiny magnet, but also has an angular momentum as though it were
spinning like a top. When I measure this angular momentum, I find
that it is also quantized: it can only have one of the two values

2tc

where 5 is either +2 or —. In words: the measured angular

momentum of the electron, which is a vector, can point in either of
two opposite directions with the same magni-76

THE ELECTRONIC STRUCTURE OF ATOMS

tude. You can see now why we chose the particular values of plus or
minus one-half for the quantum number s.
7 The electronic structure of atoms
I summarize now the picture we have developed so far. The
hydrogen atom has a set of energy levels. Each level can have more
than one wave function: it is degenerate. We can label each wave
function by the values of its quantum numbers n, I, m, and s. Now
the question is, in which of these states would a hydrogen atom find
itself?

The answer is that the atom will be in its ground state. This is so
because, if it begins in an excited state, any slight disturbance will
cause it to emit a photon (or in some other way lose energy to the
surroundings) and change to the ground state. The energy so lost by
the atom will be exactly equal to the difference in energies of the two
states, because the total energy must be conserved. The ground
state is the stable state of the atom, because it has no state of lower
energy to which it can go.

The next simplest atom must have two electrons with total charge
-2e, and therefore a nucleus with charge +2e, because the atom as
a whole is neutral with zero total charge.

Its energy level picture will be rather different from the one in fig. IV-
6, because now we have to think of the quantum numbers for two
electrons. The energies of the levels will be different from what they
were in the hydrogen atom, because the charge on the nucleus is +2
e instead of+e. What we have now is an atom of helium, which you
know as the gas used to fill balloons on fairgrounds and for
children’s birthday parties. The two electrons go into the ground
state, and their quantum numbers are as shown below: First
electron: n— \,l—Q,s — —\

Second electron: n= 1, /= 0, s= +2

You will note that not all the quantum numbers for the two electrons
are the same. One other point: the nucleus of the helium atom
consists of two protons each with a charge +eand two neutrons with
no charge.

The description of the helium atom which I have given raises two
questions whose answers introduce us to some very important new
ideas. The first question is, since I can balance the charge of—2eof
the two electrons with just two protons each with a charge +e, why
do I want to throw in two neutrons also which, after all, have no
charge? Well, you remember that two like charges repel each other,
so that if I had just two protons in the nucleus, they would fly apart. It
turns out that there is another kind offeree that acts only among
neutrons and protons, and this force attracts them to one another
very strongly. We call this the nuclear force. This attractive nuclear
force, however, is still not strong enough to overcome the repulsive
electric force that is trying to push the protons apart. The two
neutrons increase the attractive nuclear force further, without
affecting the electric force (because they are electrically neutral),
enough to hold the nucleus together. A study of the nuclear force is
at the heart of nuclear physics, which is 77

IV THE ATOM

one of the branches of physics. We shall not dwell further on it, since
what goes on inside the nucleus is not relevant to those properties of
matter that we want to understand in this book.

And now for the second question: When I put the two electrons in the
ground state, why did I not put them both in a state with the same
value for the s quantum number instead of putting one of them in the
s = -2 state and the other in the s = +2 state? The answer lies in a
peculiar property of electrons that I have not mentioned so far. This
is that no two electrons can have all of their quantum numbers
exactly the same. A given state, with certain values for its quantum
numbers, can be occupied by one, and only one, electron. It is as if
an electron, sitting in a given state, excludes any other electron from
getting into the same state. This so-called exclusion principle was
discovered by the physicist Wolfgang Pauli, and is named for him. I
shall say more about it in chapter V. We shall use it now to form a
picture of the wave functions for electrons in a given atom.

To be specific, let us consider an atom of carbon. Its nucleus

consists of six protons and six neutrons. To keep it electrically
neutral, it must have six electrons surrounding the nucleus. We have
seen that the n = 1 level has two states, and, from table IV-1, the n =
2 level has two states with / = 0 and six states with 1=1, with no two
states having all quantum numbers the same. So we start by putting
two electrons in the n = 1 level, and this fills it up. We then put two
electrons in the n = 2, / = 0 level, filling it up. The last two electrons
go into the n = 2,1 = 1 level and the result is an atom of carbon. You
can see that the exclusion principle has been satisfied by this
distribution of electrons.

I hope you are beginning to see now the picture of an atom given by
quantum mechanics. Let me describe it, putting together the points
made in this chapter. All the atoms of each chemical element are
identical. Each atom has a nucleus with a certain number Z of
protons and a number A/of neutrons. I show the values of Z and
JVfor some elements in table IV-2. Surrounding the nucleus are Z
electrons, so that the atom is electrically neutral; the charge +Ze on
the nucleus is exactly cancelled by the charge ~Ze of the Z
electrons. The electrons are to be found in the successive energy
levels, starting at the ground level. According to the Pauli exclusion
principle, no two electrons will have all the quantum numbers the
same. The value of Z, the number of protons in the nucleus (which is
the same as the number of electrons in the atom), identifies the
element uniquely; each element has a different value of Z.

Table IV-2. The number of protons, electrons and neutrons in some

atoms Element Z N

(protons, electrons) (neutrons)

0
2

146

hydrogen

helium

carbon

uranium

92

78

THE ELECTRONIC STRUCTURE OF ATOMS

If you look at fig. IV-3 again, you will be reminded that there is very
little probability of finding the electron at a distance of more than
about an angstrom from the proton in the hydrogen atom. We can
say that the size of that atom is about one angstrom. The electrons
in the other atoms go into successively higher excited states.

In these states, the electron has a somewhat higher probability of

being further away from the nucleus than in the ground state, but
even then not by very much. Thus it turns out that in the atoms of all
the elements, there is very little probability of finding any electron
further from the nucleus than a couple of angstroms. Even though
the number of electrons in atoms of different elements varies
between one and about one hundred, the sizes of the atoms are all a
few angstroms across. The nucleus itself is far, far smaller than this.
Its size depends on how many neutrons and protons it has, and in all
cases is a few times 10~3 angstroms.

I remind you again that no two of these electrons have all of their
quantum numbers the same. You must imagine that there exists a
wave function that describes the probability of finding any one of
these electrons at any specified region around the nucleus, and is
such that the Pauli exclusion principle is obeyed. I am afraid I cannot
draw a picture of this wave function for, say, a carbon atom in the
way I was able to for the much simpler case of a single electron
trapped in a box in chapter III. This is where mathematics comes to
the aid of the physicist. She will say that she does indeed have a
picture of the atom, but it is expressed in mathematical formulas and
not in lines and shadings on a sheet of paper. We shall see as we go
along that we can make much headway even without all that
mathematics.

All the things we see around us are made up of atoms. The only
difference between atoms of different elements is in the number of
electrons around the nucleus and the numbers of protons and
neutrons in the nucleus. This might appear to be too trivial a
difference to account for the amazing variety of properties that things
around us have.

A lump of coal, which is nearly pure carbon, is not likely to be

confused for a lump of silver, or for a diamond which is also pure
carbon. My main task in the rest of this book is to explain how these
properties follow from the way in which the electronic structure of the
atoms is modified when I put them together to form a solid.

The picture of an atom that we have arrived at is not at all what one
might have thought, namely a tiny hard ball a few angstroms across
in size. We can make this dramatically evident if we imagine that we
have somehow magnified an atom so that the nucleus measures a
centimetre across. Then the electrons, numbering at most about a
hundred, will be distributed in a sphere of diameter about one
kilometre. The cloud of electrons around the nucleus is a very thin
cloud indeed; most of the atom looks like empty space. It would
appear that one atom should be able to pass right through a second
atom and keep going. This is not what happens in reality. When I put
atoms together to form a solid, they behave as if they were hard
impenetrable balls stacked together. To take a more immediate
example, I am sitting comfortably on a chair as I write this; my
atoms, and therefore I, do not just simply pass through the atoms in
the chair. Why is this so?

79

IV THE ATOM

The answer lies in the Pauli exclusion principle, which says that one,
and only one, electron can occupy a given state with a given set of
quantum numbers. A second electron must necessarily go to another
state with a different set of quantum numbers. Now let us see what
happens when I try to bring two atoms of carbon, say, together.
Figure IV-7 shows the two atoms some distance apart. You can think
of it as a kind of snapshot. You see in each atom the nucleus at the
centre and six electrons in the surrounding grey space, which
represents the sphere within which each of the electrons is likely to
be found. Please note that this and the following picture are not
drawn to scale. Each electron is in one of the six possible states of
lowest energy. Now suppose I try, as in fig.

IV-8, to bring the two atoms together so that the two grey spheres
overlap. Then in the overlapping region I am forced to have pairs of
electrons, one originally from each atom, trying to occupy the same
state. This, we know, is prohibited by the Pauli exclusion principle,
which allows only one electron to be in each state. The consequence
is that the atoms cannot penetrate into each other, and my chair
supports me and does not let me pass right through it.
I have said that the number of electrons in an atom uniquely
specifies which element it is: one for hydrogen, six for carbon, and
so on. This is also the number of protons in the nucleus. There are
also some neutrons in the nucleus, which have no electric charge.
So their number is not fixed, but can vary a bit without affecting the
electrical FIGURE IV-7. Two carbon atoms A

and B at some distance apart. The

nucleus is at the centre of each atom,

and six electrons surround it. The

grey shading around the nucleus

indicates the space within which the

six electrons are almost certain to be,

according to their wave functions.

Am

FIGURE IV-8. The two carbon atoms • . .

of fig. IV-7 now brought close

enough to each other that their wave * •

functions begin to overlap, thus •

requiring that two electrons, one from »

each atom, have the same wave

function. This is not possible, because *

of the Pauli principle, and so the

atoms cannot penetrate each other. A B

80

SUMMARY

neutrality of the atom or changing it to a different element. I shall

take the simplest example, an atom with one electron which
therefore must have a nucleus with charge +e. I accomplish this by
having a nucleus of just one proton, and this of course is the
hydrogen atom. There also exist two other forms of hydrogen, rare to
be sure, with the nucleus of one consisting of one proton and one
neutron, and the other of one proton and two neutrons. They are
called deuterium, or sometimes heavy hydrogen, and tritium
respectively. Many of you have heard of isotopes. Ordinary
hydrogen, deuterium and tritium constitute three isotopes of the
same element. There are isotopic forms of all the elements. Isotopes
are of interest when one studies the physics of the atomic nucleus,
and they have found applications in certain fields of medicine,
technology, and weaponry. For our purposes in this book, however,
we can ignore their existence, for none of the characteristics of a
solid that we want to understand depends noticeably on its isotopic
composition. Well, almost none; in superconductivity, the isotopic
nature of the material plays a role that gives an important clue to our
understanding of the phenomenon, as we shall see in chapter XII.
8 Summary
The gravitational attraction between the sun and the planets is what
holds the solar system together. While this attraction certainly exists
between the electrons and the nucleus in an atom too, it is far too
weak to hold the atom together. It is a different force, the electric
force, which is at work in the atom. Section 2 introduces you to this
force, and relates it to a property of electrons and protons called
electric charge. This charge is the source of the electric force, just as
mass is the source of the gravitational force. Unlike gravitation which
always pulls two bodies towards each other, the electric force either
pushes them apart (two electrons or two protons) or pulls them
together (an electron and a proton). Because of this, we say that a
charge can be negative or positive, and like charges repel and unlike
charges attract each other.

In section 3 we see an important consequence of the wave aspect of

the electron: it cannot fall into and remain trapped inside the nucleus.
If it did, it would have such an enormous kinetic energy of motion
that it would be shot out of the nucleus right away.

The electron in a one-dimensional box has a state of lowest energy,

with a wave having just one trough (or one crest) fitting exactly inside
the box. We look in section 4

at the analogous lowest energy state, the ground state, of the

hydrogen atom. This is the simplest atom we have, with just one
proton and one electron. We find that the wave function in this state
is such that the electron is most likely to be found very close to the
proton, and less and less likely as we go further away. The likelihood
becomes practically zero at a distance of about an angstrom, or
10~8 centimetres, so that we can think of the atom as measuring
about an angstrom across.
In section 5 we look at the states of the electron with higher
energies, the so-called excited states. Each of these states has a
definite energy, and these energies are 8i

IV THE ATOM

separated by gaps. The electron is allowed to have only these

energies, and nothing in between. This is what we call the
quantization of energy.

We look in section 6 more closely at the energy levels of the atom.

Each of these levels is degenerate, i.e. there is more than one wave
function that corresponds to this energy level. This means that the
electron can have different probability distributions around the
nucleus, all of them having the same energy. We introduce quantum
numbers n. I, m, and s for the wave functions. Each wave function is
labelled by the particular numerical values which these quantum
numbers assume in it. The number n is related to the energy of the
electron in that state, / and m to that part of the angular momentum
of the atom which arises from the way the electron wave is
distributed around the proton, and s to the angular momentum of the
electron by itself. We find that not only the energy but also the
angular momentum is quantized. The allowed values of the angular
momentum are determined by the Planck constant h, and the
numerical values of m and 5. The atom is usually found in its lowest
energy state, namely the ground state. It can be put into an excited
state by giving it an amount of energy that is exactly equal to the
difference in energies between the two states, for example by
making it absorb a photon of this energy.

Using the scheme of energy levels and quantum numbers, we see in

section 7 how the atoms of other elements like helium, carbon and
uranium are built up. The nucleus now has several protons and
neutrons, and the number of protons determines uniquely which
element it is. Surrounding the nucleus is a number of electrons equal
to the number of protons. Each electron has its wave function with a
set of quantum numbers which are different from those of any other
electron. No two electrons can have exactly the same set of values
for the quantum numbers. This is an example of the Pauli exclusion
principle. One can say that the electrons do not seem to be very
friendly characters, each preventing any other electron from entering
its house.

What I have described so far is an atom in isolation from the rest of

the world, not disturbed in any way by all the surrounding atoms
because they are too far away. The (a)

FIGURE IV-9. When two atoms

are far apart as in part (a), the wave

functions and energy levels of each

atom are little influenced by the

presence of the other atom. This is no

longer the case when the atoms are

close together in a solid, as in part (b). ” _

The wave functions and energies in a

solid will thus be different from what

they are in the isolated atom. (Jy)

82

SUMMARY

atoms in a gas are a very close approximation to this state of affairs.

Their average distance apart is much larger than their size. We
remember that it is the electrical force that is the important one in
atoms. Now look at the two atoms in fig. IV-9, part (a). I have shown
them as being very far apart in relation to their size. Then the
electrons and the nucleus of each atom are all at about the same
distance from the other atom. Each atom feels no electrical force due
to the other atom, because the force due to the electrons (negative
charge) will be cancelled by that due to the nucleus (equal positive
charge). So each atom is unaware of the other’s existence, so to
speak. Now look at the two atoms in part (b) of fig. IV-9. Their
distance apart is comparable to their size. You can see qualitatively
that the electrons and the nucleus of each atom must each feel the
influence of the other atom, and we can no longer think of isolated
atoms. The atoms in solids are at distances apart which are like that
of the atoms in part (b): the lattice spacing in crystals is a few
angstroms, and is therefore comparable to the size of atoms. We
can expect that the wave functions and energy levels in solids will be
changed from what they are for the isolated atom. These are the
changes we shall look at in chapter VI. We shall then be able to see
why solids are the way they are.

Before getting there, there is one other matter we have to deal with.
We have seen so far what the quantum-mechanical picture of single
atoms looks like. In an actual solid/there are very large numbers of
atoms packed rather closely together. Very large numbers indeed: a
cube of solid measuring one centimetre along the edge will have
something like 1021 atoms in it. That means that there will also be
comparably large numbers of electrons in the solid. It is obviously a
hopeless task to try to find out what each electron is doing, because
even the lifetime of the universe would not be enough for the task.
We need to invent a new way of dealing with such large numbers of
objects. This way is called statistical physics, and it is the subject of
the next chapter.

STATISTICAL PHYSICS

l Introduction
Any object that we see around us, even though it may be very small
like a grain of sand or a pin, consists of a very large number of
atoms, about 1021. When the atoms are relatively far apart, as in a
gas, the electronic structure, meaning the wave functions, of each
atom is practically unaffected by the presence of the other atoms.
When the atoms come close together in a solid, the wave functions
become different from what they were for the isolated atoms. In
crystals of metallic elements like silver and copper this change is
such that the electrons in the highest energy state of the isolated
atom are no longer bound to stay near the corresponding nucleus,
but are free to wander around in the whole crystal. This freeing of the
outer electrons happens with each atom in the crystal. So we end up
with a crystal in which a large number of electrons, 1021 or more,
are wandering about. These electrons are responsible for many of
the metallic properties of copper or silver: the shiny appearance and
the ability to carry electric current and to conduct heat easily, for
example. We shall see later how this comes about. For the present
we need to find a way to describe the properties and behaviour of
such large numbers of identical entities like electrons or atoms. We
do this with statistics.

The Concise Oxford Dictionary defines the word statistics as

numerical facts systematically collected. It is in this sense that the
word is used when we talk, for example, about weather statistics: the
temperatures, rainfall and so on for various cities that are reported in
the newspapers every day. A typical report might cover some fifty
cities.

What if we want to present the same information for a thousand

places scattered all over the world? Or a million places? We would
then be faced with a hopeless task —quite apart from the question of
what use the resulting compilation would be.

Fortunately, we do not have to undertake a similarly daunting task

when we consider the behaviour of the 1021 or so particles, like
electrons or atoms, that are in a typical piece of solid matter or
volume of gas or liquid. It turns out that for our purpose of
understanding the behaviour of matter we do not need to know what
each particle in this huge assembly is doing, but can be satisfied with
much less information about groups of them. To explain what this
means, I shall first describe a simple example.

Figure V-l shows some information about the age of the populations
of two countries: the U.S.A. in the year 1989 and India in 1990.
Along the horizontal line at the bottom you see the labels 0 to 9, 10
to 19, and so on, which stand for ages up to 9 years, between 10
and 19 years, …. Above each of these labels are two bars, one for
the U.S.A. and the other for India. The length of each bar represents
the percentage of the population that is in the corresponding age
group. The vertical line on the left of the 84
INTRODUCTION
FIGURE v-l. The population

distribution by age in the U.S.A. and

in India in the years 1989 and 1990

respectively. For example, about 14

per cent of people in the U.S.A. and

22 per cent of those in India were

between 10 and 19 years old.

On

On

r-H

On

(N

-i—»

o
On

CO

-4—’

co

On

-t—»

On

in

in

On

^O

4—*

age group (years)

diagram is a scale that enables you to tell how long each of the
rectangular bars is, just as an ordinary scale for measuring lengths
tells you how long something is in centimetres or inches. For
example, the figure shows that about 14 per cent of the population of
the U.S.A. and 22 per cent of Indians are between 10 and 19 years
old.

You will note that this figure, although it tells us something about the
ages of the people, is quite different from a simple, if very long, list of
the age of each person in the two populations. It tells us how the
people are distributed in different age groups. For this reason we say
that fig. V-l shows the distribution of the populations of the U.S.A.

and India by age.

question What are the percentages for the age group from 0 to 9
years in the U.S.A. and India?

answer It is about 15 per cent in the U.S.A, and 25 per cent in India.

There were 248 million people in the U.S.A. and 827 million in India
when these counts were made. You can imagine the impossibility of
trying to make a list showing the age of each person. It suffices for
many purposes just to know what fractions of the population are up
to 9 years old, 10 to 19 years old, and so on. Such information would
be needed in making plans for education, health care, social
security, and other such matters. The picture of the distribution offers
something more: it presents in a visually striking manner something
about the populations that would be invisible in a bare table showing
the age of each person, and not easy to see even in table V-1, which
shows the numbers that were used in drawing the figure. I point out
the following features of the way the percentage of people in a given
age group varies as the age of the group increases:

V STATISTICAL PHYSICS
Table V-1. Distribution by age groups of populations of the U.S.A.
and India age group

(years) 0-9 10-19 20-29 30-39 40^19 50-59 60-69

per cent

(U.S.A.)

per cent

(India)

14.9

24.8

14.0

21.8

16.3

17.7

16.8

12.7

12.2

9.5

8.9

7.0

8.5

4.2
1. For the U.S.A., this percentage slowly increases to a maximum
value at the 30 to 39 year age group, and then begins to decrease.

2. For India, this percentage starts at a relatively high level for the
youngest age group, and then steadily declines as the age
increases. The pattern is strikingly different from that of the U.S.A.

One can make sense of these two features if one thinks about the
relative affluence of the two countries and the consequence for infant
mortality, health care, and nutrition, and also recalls that for some
years in the 1950s, the so-called baby-boom years, the U.S.A. had
an unusually high birth rate. I shall not dwell further on this, for this
book is about physics and not demographics. I do note three
features of this way of presenting the information. The first is that we
know something about the distribution of the population among
different age groups, but nothing about the age of any given
individual. The second is that the pattern of the age distribution
implicitly says something about the population: its affluence, state of
health, and even history. Thirdly, the information in this age
distribution can be used for other purposes such as the planning of
health care, social services, and so on. There are analogies to these
three features in the way we describe a collection of a large number
of particles in physics. This description is called statistical physics.
Before I get to that, I must explain to you what temperature means in
terms of the atomic structure of matter. As you will see soon,
temperature plays a crucial role in statistical physics.
2 Temperature
Boiling water is hot, and ice is cold. I specify how hot or cold
something is by giving its temperature. Water boils in New York at a
temperature of 212 degrees Fahrenheit, and freezes to ice at 32
degrees Fahrenheit. In Rome, on the other hand, water boils at 100

degrees Celsius and freezes at 0 degrees Celsius. This does not

mean that Italian water is somehow different from American water,
but only that people in the two cities use different units with which to
measure temperature. In this book we shall follow the practice of
physicists, and indeed of most of the world, and use Celsius
degrees. We 86

TEMPERATURE

shall write ‘twenty degrees celsius’ as ‘20 °C\ There is yet another
way of expressing temperature, by adding the number 273 to the
degrees Celsius and calling the result as so many kelvin. For
example, 20 °C would be the same as 293 kelvin, which we shall
write as 293 K. You will see in a while that there is some point to
expressing temperatures this way.

We get our idea of something being hot or cold through our sense of
touch. I now want to connect this idea with what the atoms are doing
in a piece of matter. Suppose I take a container filled with helium
gas, say, which is at a temperature of 20 °C. The helium atoms are
not just standing still, but are whizzing about in all directions at all
possible speeds, bumping into one another and with the walls of the
container. When two atoms collide, the speed of each will in general
be different from what it was before. You can see that it is hopeless
to expect to follow the motion of each atom as time goes on.
Fortunately, we do not need to do this. Numerous experiments have
shown that the number of atoms whose speeds lie in any given
range will remain unchanged, so long as the temperature remains
unchanged. The atoms whose speeds change to outside this range
because of collisions are replaced by an equal number of atoms
whose speeds move into this range because of collisions. Now
suppose that I heat the container so that it and the gas inside it are
at a temperature of 300 °C, say.

The atoms will still be moving about at all speeds, but their
distribution in different ranges of speeds will be changed by this
change of temperature.

Figure V-2 shows the distribution of speeds of the helium atoms at

the two temperatures. The picture is based on experiments in which
the speeds of groups of atoms were measured, and on a theory
worked out by the physicists Ludwig Boltzmann and James Clerk
Maxwell. Their predictions agree with the experimental results.

FIGURE V-2. Distribution of atomic

speeds in helium gas at 20 °C and

300 °C. The length of each bar shows

the per cent of atoms whose speeds lie

within the corresponding range

shown along the horizontal axis.

Each range includes the lower value

and ends just below the upper value.

o
00

(N

00

CNJ

(N

00

(N

cn

00
CO

CM

^o

rt

00

OO^Ht-hcSICSCSCOCO

speed, km/second

V STATISTICAL PHYSICS

The horizontal axis carries labels showing speeds within the ranges
0 to 0.4, 0.4 to 0.8, … kilometres per second. I should mention that a
speed of 1 km/sec, which is the same as 2250 miles per hour, is
somewhat more than what we normally meet; the atoms are indeed
speeding about. Above each of these numbers are two bars, one for
each of the two temperatures. The length of a bar, measured on the
vertical scale at the left, gives the percentage of atoms that have
speeds in the corresponding range and at that temperature. For
example, I see from the figure that about 30 per cent of the atoms in
the gas at 20 °C have speeds between 1.2 and 1.6km/s, and this
drops to 20 per cent when the temperature is raised to 300 °C.

question What percentage of atoms have speeds in the range from

0.4 to 0.8 km/s?

answer About 10 per cent at 20 °C and 4 per cent at 300 °C.

question What is the average speed of an atom in the gas at 20 °C?

answer The question is similar to ‘If four apples cost one dollar each,
and six other apples cost two dollars each, what is the average cost
of an apple?’ To answer this, we work out the total cost of the ten
apples as four times one, plus six times two, or 16 dollars, so that
the average cost per apple is 16 divided by 10 or 1.60 dollars. We
can work out the average speed in the same way, substituting atoms
for apples and speed for cost.

Suppose we have 100 atoms in the gas. The actual number will not
matter, because we are working only with percentages. We find from
the figure that about 27 per cent, or 27 of them, have speeds
between 0.8 and 1.2 km/s. We say approximately that each has a
speed of 1.0 km/s, and we multiply 27 by 1.0 to give us the number
27.0. We do the same thing with all the speed ranges, add up the
resulting numbers, and divide by 100, which is the total number of
atoms, to get the average speed per atom. When you work it out,
you will find that the average speed is 1.2 km/s. We could have
expected the answer to be about this, because we see from the
figure that a large fraction of the atoms have speeds around 1.2
km/s. If more than half the apples I bought cost a dollar each, and
there was not too much spread in the prices of the rest, then I would
expect the average cost per apple to be about a dollar.

I repeat that even though the speed of a given atom may be

changing thousands of times per second, the distribution of speeds
is unchanged so long as the temperature of the gas remains
constant. One might wonder whether there is a deeper connection
between the two quantities. There is a slight problem: temperature is
just one number, so many kelvin or degrees Celsius, whereas the
distribution of speeds is many numbers.

So we need to distil some single number out of the distribution that

can be meaningfully related to the temperature.

We start the search for this number by looking at some features of

fig. V-2. While TEMPERATURE AND ENERGY

there are atoms moving at all speeds from practically zero to more
than 4.4 km/s, the percentage reaches a maximum value for speeds
in the range between 1.2 and 1.6 km/s, and drops off at lower as well
as higher speeds. There are some atoms moving at speeds much
greater than 4.4 km/s, but they are such a small percentage of the
total number of atoms that they do not show up on the scale of this
picture.

Now look at the effect of heating up the gas on the distribution of

atomic speeds.

You can see that the distribution, although it has the same qualitative
shape, is quantitatively different. Raising the temperature leaves a
smaller fraction of atoms at the lower speeds and increases the
fraction of atoms at higher speeds. An appreciable percentage of
atoms is now moving at high speeds at which there were practically
no atoms at the lower temperature. The range of speeds at which
the maximum percentage of atoms occurs is now at 1.6 to 2.0 km/s,
while it was at 1.2 to 1.6 km/s at 20 °C. We can calculate the
average speed of an atom for the distribution at 300 °C, and find that
it is 1.7 km/s, distinctly higher than the value of 1.2 km/s at 20 °C. It
would appear that the average speed is a possible candidate for the
single number we are looking for to relate to the temperature. It turns
out that there is an even better candidate, namely the average
energy of an atom in the gas.
3 Temperature and energy
When the container of helium gas was heated from 20 °C to 300 °C,
we see from fig.

V-2 that the only changes were that there were fewer slower atoms
and more faster atoms, and the range of speeds in which the highest
percentage of atoms found themselves was higher. As a
consequence, the average speed increased from 1.2 km/s at 293 K
to 1.7 km/s at 573 K. Notice that I have now changed to kelvin from
degrees Celsius, and for good reason too, as you will soon see.
From the foregoing observations it seems reasonable to conclude
that the temperature must somehow be connected with the state of
motion of the atoms, which was the only thing that changed when
the gas was heated up. So let us try to find out what this connection
might be. To do this, we need to recall some things that we learnt
about the motion of particles in chapter III, section 3 (page 37).

We learnt there that a moving atom with mass M and velocity v has a
momentum p and kinetic energy if which are given by the formulas

p = Mv

and

„ _ Mv2

K — -,

where v is the velocity (with magnitude v) of the atom. We get the

total momentum and total kinetic energy of the gas by adding up for
all the atoms their individual momenta (remember that they are
vectors) and kinetic energies (which are scalars, and always 89
V STATISTICAL PHYSICS

positive). The atoms at any given instant of time will be travelling in

every possible direction inside the container, so that the momentum
of one atom going in a particular direction will be cancelled by the
momentum of some other atom going in the opposite direction with
the same speed. This results in the total momentum of the gas being
zero, and it remains zero, by the same reasoning, when the gas is
heated up. The total momentum of the gas is unchanged when the
temperature changes, and so there can be no direct connection
between the two.

Things are different with the kinetic energy of the gas. We see from
the formula for K that increasing the speed increases K even faster:
doubling the speed quadruples the kinetic energy. So the raising of
the temperature, resulting in a speeding up of the atoms, causes the
total energy (which we shall call E) of the gas to increase. It is
tempting to say that the temperature is somehow related to the total
energy of the gas. But not so fast! If I took two identical containers of
the gas at the same temperature and put them together, the resulting
object would have twice the energy of the separate containers, but
the temperature would be unchanged.

So I cannot relate the total energy of the gas to its temperature.

There are other quantities that are unchanged in the above
experiment of putting two containers together. One of them is the
average speed of an atom. Another is the average kinetic energy of
an atom. I have calculated the latter quantity using the same method
as for .the average speed, and find that the average kinetic energy is
MX 8.5 X109 ergs at 293 K

and

MX 17.0 X 10” ergs at 593 K.

The letter M stands for the mass of the helium atom expressed in
grams. The average speed and the average energy depend only on
the temperature of the container of gas, and not on how many atoms
there are.

The average speed was 1.2 km/s at 293 K and 1.7 km/s at 593 K.
The speed does increase with an increase in temperature, but not
exactly in proportion. A doubling of the temperature causes a much
smaller increase of the average speed. However, when the
temperature is almost doubled from 293 K to 593 K, the average
energy is also doubled. There exists a simple proportionality
between temperature and average energy. You can see now why I
chose to measure the temperature in kelvin. It is only then that this
simple proportionality appears. We have verified this proportionality
for helium gas at two temperatures now. Much work with many
gases at many temperatures has confirmed this proportionality of the
temperature to the average energy of the atoms or molecules in the
gas.

We have thus established a direct connection between the

temperature and the average kinetic energy of the atoms in a gas.
Let us now think a bit about what this means. You would measure
the temperature of the gas by sticking a thermometer in it.

The temperature is a property of the gas as a whole, is the same

everywhere in it, and 90

THE TEMPERATURE OF CRYSTALS

the act of measuring it would seem to have nothing to do with the

fact that the gas is composed of atoms. Now we see that measuring
the temperature is nothing but, as it were, measuring the average
energy of the atoms in the gas. We have found a connection
between a macroscopic property, the temperature, with an atomistic
property, the average energy of atoms. The construction of such
connections between features of the macroscopic world around us
and their microscopic atomic structure, is the task of the field of
statistical physics.
4 The Kelvin scale and the absolute zero

We have seen that the average kinetic energy of an atom in a gas is

proportional to its temperature. We can write this as a formula:

where Kis the kinetic energy, 7” is the temperature in kelvin, and kB

is a constant called the Boltzmann constant, after the physicist
Ludwig Boltzmann. It has the value kB= 1.38 X 10”16 ergs per
degree.

The factor \ in the formula for K is there because of the value chosen
for the constant kB. We see that doubling the temperature doubles
the kinetic energy. If we go in the other direction and halve the
temperature, then the kinetic energy is also halved and the speeds
of the atoms are lower. As we keep reducing the temperature, K gets
smaller and smaller. Is there going to be an end to this? Yes,
because the lowest energy an atom can have is its zero-point
energy, which is a consequence of the wave aspect of the atom
inside a container, just as the particle in a one-dimensional box
(chapter III) has a finite zero-point energy. This must mean, from the
formula above, that the energy of motion (called the thermal energy)
which is to be attributed to temperature is zero at a temperature of
zero K, and the atoms are left with just their zero-point energy. We
shall call this the absolute zero of temperature, 0 K. From the way
we defined the Kelvin temperature scale (sometimes also called the
absolute temperature scale), we see that the absolute zero is at -273
°C. This is the lowest possible temperature that could exist anywhere
at all, corresponding to all the atoms being in their lowest energy
state.
5 The temperature of crystals
In the preceding section, I talked glibly about cooling the gas until it
reaches the absolute zero, where the atoms have no thermal energy.
Well, this does not quite agree with reality.

If I take a container of some gas and cool it to lower and yet lower
temperatures, then at some temperature (called the boiling point) it
will condense into a liquid. On further cooling, it reaches a
temperature (the freezing point) at which it becomes a solid. An
example: I V STATISTICAL PHYSICS

imagine a container at a temperature of 105 °C filled with steam

having the same pressure as the air around us. As the container
cools down, the steam will condense to water at 100 °C, and on
further cooling, the water will freeze to ice at 0 °C. The motions of
atoms in the solid and the liquid will be somewhat different from each
other, and each will certainly be very different from that in the gas.
So what if anything will survive of the simple relation between
temperature and kinetic energy that we found in the gas?

To begin with, it seems reasonable that the temperature in the liquid

and the solid must still be connected with the energy of motion of the
atoms. To see this, let us think of the gas a bit above the boiling
point. Its atoms are moving about at various speeds as we saw,
some of them quite large. Now if the temperature drops a little,
enough to cause the gas to condense into a liquid, the atoms will
have lower speeds and will be bouncing off each other more than
they did in the gas because the atoms are closer together. The
atomic motions will be much more complicated than in the gas, and it
is not possible to present a picture of the distribution of speeds in the
same way as in fig. V-2. We can imagine that there is still some
distribution which is characteristic of the temperature, and that this
distribution shifts to lower speeds as the temperature is lowered.
question Why should not the distribution of speeds in the liquid shift
to higher speeds as the temperature is lowered?

answer Well, if this happens each time we lower the temperature,

the distribution will soon be back up at those speeds where we
should have gas, and not liquid. Of course, the liquid does not
become gas when the temperature is lowered; rather, it will freeze
into a solid.

Now let us look more closely at a solid, a crystal to be specific, for

that is what we are mostly concerned with in this book. You will
remember from chapter II that the atoms in a crystal are in an orderly
arrangement at the lattice points. Suppose the crystal is at the
absolute zero of temperature, so that it has the lowest possible
energy. This would be the case if the atoms are just sitting at the
lattice points and not moving at all, for then the kinetic energy of
each atom would be zero. However, the wave nature of the atoms
does not permit this situation to be realized. Each atom in the crystal
is effectively trapped inside a small finite volume around its lattice
point. The surrounding atoms block it from leaving this volume. So,
just as the particle trapped in a one-dimensional box has a ground-
state energy that is not zero, the atom in a crystal has a non-zero
lowest energy, its zero-point energy.

I should mention here that nature is such that we cannot cool

something down to the absolute zero by any method that has been
used or even can be imagined. The closer we get to 0 K, the more
difficult it becomes to get any closer. The unattainability of the
absolute zero is not a failure of our ingenuity, but rather a
consequence of the wave nature of matter. It is still all right to think
of the crystal at 0 K as an idealization.

Now suppose I raise the temperature of the crystal by adding energy

to it. This energy is taken up by the atoms and produces some sort
of motion in them. If the crystal is to remain a crystal, each atom
cannot wander away from its lattice site. It can only 92
V STATISTICAL PHYSICS

even if the particles are molecules, electrons, ions, and so on. We

can go even further: we saw that light has also particle-like
properties, and we call these particles photons.

The inside of a hot oven is filled not only with hot air but also with
infrared radiation which is after all a form of light and is composed of
photons. It is as if the oven is filled with a gas of photons, all moving
at the speed of light and with an energy distribution that will
correspond to the temperature of the oven. We shall see presently
that this distribution is different from the one for atoms of helium gas.

question We spoke of percentages of the total number of atoms in

different ranges of speeds for the helium gas, leading to the energy
distribution which then connects with the temperature of the gas. The
photons are all travelling at the speed of light, so that there can be
no distribution of their speeds. So what is the quantity whose
distribution is the relevant thing for photons?

answer We recall that the energy £ of a photon is related to its

frequency /through the Planck constant h:

E=hf.

So we expect that the energy distribution for photons is to be

expressed as the number of photons (and correspondingly their total
energy) having frequencies in different ranges.
6 Fermions and bosons
We have so far talked about the atoms, electrons and photons in this
chapter as if they were particles. We have not taken into account
(other than in mentioning their zero-point energy) their particle-wave
duality, which requires that these objects must be described using
quantum mechanics. We shall now do this, and discover a result that
is basic to much of what follows in the book.

I start by recalling some features of a quantum-mechanical

description of a single particle. I shall call it particle A, where A can
be an electron, hydrogen atom, etc. It is associated with a wave that
is described by some wave function \|/(rA,aA) which has some value
at every point rA in space. This notation reminds me that the particle
A is in a particular quantum state labelled by aA. As an example, the
electron in the hydrogen atom is in a state specified by the values of
the quantum numbers n, I, m, and s, and when I say the state 0CA, I
mean these values of the quantum numbers. For the sake of
compactness, I shall later use the single symbol qA to represent both
rA and 0CA. The square of the amplitude that the wave function has
at a given point in space is proportional to the probability of finding
the particle near that point.

Now suppose I take two particles labelled A and B that are the same
kind: they both are electrons, or hydrogen atoms, for example. The
wave function of this system of two particles will certainly be different
from that of one particle. It will depend on their 94

FERMIONS AND BOSONS

positions and quantum numbers, which I shall indicate by the single

symbols qA and qB.
I can then write this wave function as §(qA, qB). Its amplitude varies
as the positions of the particles A and B vary. The square of the
amplitude, <))2, is proportional to the probability of finding the two
particles near the corresponding positions.

question We had some pictures to help us see what some of the

wave functions for a particle in a box and for a hydrogen atom
looked like. Can we also have a picture of what this new wave
function looks like?

answer Alas, not so easily. For one thing, we do not have enough
information about the particles, such as whether they are in a box
and if so what size and shape of box, etc. We do not even know if
both of them are electrons or protons or photons or whatever. If I had
all this needed information, I could in principle calculate the wave
function <|). I would then still have difficulty in making a picture of it,
for its value will depend on the positions of each of the two particles,
and there is no simple way I could show all of this on a flat sheet of
paper.

With such well-nigh total ignorance of the details of the wave

function for two particles, one might think that there is little more we
can say. Before giving up, we note that there is one point we have
not yet made use of: the two particles are identical. If I imagine the
particles to be interchanged so that A becomes B and B becomes A,
the wave function would change from (J)(^A, qB) to ty(qB, qA). But
since the two particles are identical, the system will have the same
properties after the interchange as before. The simplest of the
properties that are related to the wave function is the probability of
finding the two particles near any two chosen points. This probability,
equal to the square of the wave function, must be unchanged by the
interchange of A and B: §(qA, qB) multiplied by itself = §(qB, qA)
multiplied by itself.

There are two ways of satisfying this condition. One is if <t>(?A. 9b)
= <K?b, 1a)>
and the other (remembering that a negative number multiplied by
itself gives a positive number) is if

HiA> 9b) = -<K?b. ?a)-

In words: the wave function for a system of two particles must be

such that it will, when the two particles are interchanged, either
remain exactly the same or at most change its sign (positive or
negative) to the opposite sign.

We shall be interested in systems not with just two particles, but with
a very large number JV, typically 1021, of particles. The wave
function for such a system would have the form §(qlt q2, …, gw),
where each q stands for the position and quantum numbers of one of
the particles. Using the same reasoning as I did for two particles, I
see that interchanging the two particles at q1 and q2 and leaving the
others where they were will give me either the same wave function
or merely change its sign: either 95

V STATISTICAL PHYSICS

or

You will note that in all the foregoing, I have said nothing about the
actual form of the wave function, other than that it describes a
system of Af identical particles. The actual form will depend, for
example, on whether the system is the collection of 92 electrons in
an atom of uranium or the 1021 or so atoms in a container of helium.
In any case, the wave function must be one or the other of the two
kinds described above. What the particles are - electrons, photons,
helium atoms, etc. - determines which of the two kinds is
appropriate.

I introduced in chapter II the idea of symmetry: if after I do something

to an object it looks exactly the same as before, I say that it is
symmetric under what I did to it. In the same spirit I call the wave
function that is unchanged when I interchange two particles a
symmetric wave junction, and the one that changes its sign an
antisymmetric wave function.

The operation I carry out in each case is the interchange of two

particles. Particles that have symmetric wave functions are called
bosons, and those which have antisymmetric wave functions are
called fermions. The names commemorate the physicists
Satyendranath Bose and Enrico Fermi who first clarified these ideas.
Photons and hydrogen molecules are examples of bosons, electrons
and protons of fermions. If I interchange two particles A and B in a
system containing identical particles, then I have for bosons: §s{qA,
qB) = tys(qB, qA),

and for fermions: §a(qA, qB) = -<\>a(qB, qA).

In what I have just written, I have used subscripts s and a to remind

me which wave function is .symmetric and which is antisymmetric.
Also, I have not indicated anything about the other particles, which
remain undisturbed during the interchange of A and B.

So we have two kinds of particles according to quantum mechanics,

fermions and bosons. How do we know which kind a given particle,
say an electron, or a proton, or a photon, is? One way is to measure
the properties of an assembly of such particles. We shall see in the
next section that some of these properties can be strikingly different
depending on whether the particles are fermions or bosons. We
would then find that electrons and protons are fermions, photons and
hydrogen molecules are bosons. This tells us what they are, but not
why. The answer to that question is given when we realize that these
particles are to be described using not only quantum mechanics but
also the Einstein theory of relativity. It turns out then that the spin
quantum number of a given particle can only have one of the values
0, \, \,\,2 and so on. No number lying in between is possible.
Particles with integer spin quantum numbers (0, 1, 2,…) are found to
have symmetric wave functions and therefore are bosons, and the
particles with fractional spin quantum numbers have antisymmetric
wave functions and so are fermions. The electron with spin quantum
number \ is therefore a fermion, and the photon with spin quantum
number 1 is a boson.

96

FERMIONS AND THE PAULI PRINCIPLE

Now whether we have a system of bosons or one of fermions, we

have seen that interchanging two particles leaves the system exactly
as before, in terms of the probability of finding the particles in
specified positions: this after all is what the square of the wave
function’s amplitude gives us. So one might well ask what difference
it makes whether the particles are bosons or fermions. The answer is
that it makes a tremendous difference, as we shall see now.
7 Fermions and the Pauli principle
I go back to the case of two identical fermions A and B, and use the
symbol qA for the position vector and quantum numbers of A, and
similarly qB for B. Interchanging the fermions then changes the sign
of the wave function:

I have assumed so far that the fermions have different sets of

quantum numbers. For example, if I think of the two electrons in a
helium atom, and label them A and B, then their quantum numbers
might be as shown in Table V-2.

Table V-2.

electron

Quantum numbers for the two electrons in a helium atom quantum

nI

10
32
number
m

Now consider the special case where the two sets of quantum
numbers for the two fermions A and B are identical. I denote this
single set by the symbol a. Then the wave function depends on (X
and the position vectors rA and rB of A and B, and can be written as
\|/(rA, rB, a). Interchanging A and B leads to the result \|/(rB, rA, a) =
-\|/(rA,rB, a).

I have labelled the fermions A and B just to keep track of their

interchange. They are nevertheless identical and have exactly the
same quantum numbers, and so interchanging them must leave their
wave function unchanged. Yet because they are fermions, the wave
function must change its sign. The only way this is possible is if the
wave function is equal to zero, for both plus zero and minus zero are
just zero. If the wave function is zero, it means that the probability of
finding the particles is zero: I can never find two fermions in a
situation where they have all their quantum numbers the same. A
given quantum state, defined by its set of quantum numbers, can be
occupied by only one fermion and no more. Fermions are not very
friendly types, it would 97

V STATISTICAL PHYSICS

seem. This is exactly the Pauli principle, which we used in describing

the atom in chapter IV.

The Pauli principle holds only for fermions. For bosons,

interchanging two particles that have the same quantum numbers
does not change the sign of the wave function and so it remains the
same as before, as is to be expected. This means that in an
assembly of bosons, any number of particles can all be in the same
quantum state and share the same quantum numbers. In contrast to
fermions, which do not seem to like one another’s company, bosons
are rather gregarious creatures.

8 The quantum states of particles in a box

I want now to develop a quantum description of a very large number

N of identical particles. If I find out what quantum numbers to use for
this system, then I can see how to assign them to the particles,
depending on whether they are fermions or bosons. I assume that
the particles are confined inside a three-dimensional box, so that the
probability of finding them outside the box is zero. Now you recall
from chapter III (page 55) that the analogous condition for a particle
in a one-dimensional box led to the wave function taking the form of
standing waves with only certain values for the wave number and
nothing in between. These wave numbers kn were the quantum
numbers of that system.

In a three-dimensional box, the waves belonging to a particle are in a

three-dimensional space, and therefore have wave vectors with both
magnitude and direction.
Again, because the waves must have zero amplitude outside the box
and on its walls, we are only allowed standing waves whose wave
vectors kn can only have particular values of the magnitude kn and
point in particular directions. Each allowed \an is then a quantum
number (or strictly speaking, a quantum vector) with its own wave
function, and the energy En of a particle with this wave function is
given by E«~ 2m

where kn is the magnitude of the vector kn. The wave function for
each of these wave vectors will also depend on the spin quantum
number s, and I can write it as \|/(kn, s). If for example the particles
are electrons, then for a given kn there are two wave functions,
vj/(kn,+2) and \y(kn,-2) for the two possible values of the spin
quantum number for the same value of kK. How the particles are
distributed among these wave functions depends on whether they
are fermions or bosons, as we shall see in the next section.
9 Fermion and boson distributions
I shall first consider the case of fermions. I imagine a large number,
like 1021, of fermions in the box, and see how they are distributed
among the different states 98

FERMION AND BOSON DISTRIBUTIONS

\|/(kn, s). I assume to begin with that the box is at the absolute zero
of temperature, 0 K, and then see what happens as I heat the box
and its contents to higher temperatures.

At 0 K, the fermions must have the lowest possible energy: this is

what is meant by the absolute zero of temperature. I show in fig. V-4
a portion of the possible energy FIGURE V-4. The occupation of

successive states of increasing energy

by an assembly of fermions, at

different temperatures labelled in

increasing order from (a) to (d). Each

small rectangle represents a quantum

state which is either empty or has one

particle (Fermi-Dirac statistics). The

energy of the state, allowing for

degeneracy, increases as one goes up

the stack of rectangles. As the

temperature increases, only the top of

the distribution is altered, with some

fermions moving to quantum states of

higher energy and thus leaving empty

states behind.

M.

(a)

(c)

(d)

99

V STATISTICAL PHYSICS

states as boxes stacked in a tower with the lowest energy at the

bottom and the energy increasing as I go up the stack. Each of these
states has a finite number, certainly very much smaller than 102\ of
possible kn vectors and corresponding wave functions \|/(kn, 5).
Since no two fermions can have the same kn and s, we cannot have
all the fermions sitting in the lowest energy state. Rather, they must
occupy progressively higher energy states until all the fermions have
been accommodated. This situation is shown in fig. V-4, stack on the
left labelled (a). Every energy state up to a certain highest energy is
fully occupied and cannot take any more fermions. This highest
energy is called the Fermi energy, and it depends on the number of
particles, their mass, and the size of the box. Even at the absolute
zero of temperature, only one fermion is in each of the states of
lowest energy. All the others are in successively higher energy
states, one in each state, and are whizzing about at various speeds
that can reach quite high values.

question Does this mean that no two fermions can have the same
energy?

answer No. The Pauli principle says that no two fermions can be in
the same quantum state. However, quantum states can have
degeneracy, meaning that more than one such state can have the
same energy. Thus many fermions can have the same energy even
though each is in a different quantum state. In a cubic box, for
example, there are twelve degenerate states all of the same lowest
possible energy.

Now suppose I heat the box and its contents to some temperature
above the absolute zero. This means that I add some energy to the
fermions, so that they move to nearby states of higher energy. Note
that fermions sitting in states well below the Fermi energy cannot go
to states of somewhat higher energy, because these states are
already fully occupied and will not accept any more fermions, thanks
to the Pauli principle.

Only those fermions that are at or near the Fermi energy can find
empty states at higher energies into which they can move. Fermions
have room only at the top. This state of affairs is illustrated in fig. V-
4, stacks labelled (b), (c), and (d) for successively higher
temperatures. A change in the temperature affects only those
fermions with energies near the Fermi energy, which form only a
very small fraction of the total number, and leaves the vast majority
of the fermions in the same states they occupied at the absolute
zero. Electrons are fermions, and many of the properties of solids
that we shall be looking at involve changing the energy of the
electron system. Thus these properties are mainly determined by
just the relatively few electrons with energies near the Fermi energy,
while the rest of them are, so to speak, passive spectators.
The situation is quite different with bosons, as illustrated in fig. V-5.
The part labelled (a) shows the case for 0 K, and (b), (c) and (d) for
successively higher temperatures. At each temperature, the box
labelled (1) represents the state of lowest energy, and (2), (3) and (4)
are three higher energy states. Being bosons, the particles can all
get into the lowest energy state, and this is exactly what happens at
0 K. With increasing 100

SUMMARY

FIGURE v-5. The occupation of states

of increasing energy by an assembly

of bosons (Bose-Einstein statistics), at

different temperatures labelled in

increasing order from (a) to (d). The

same four energy states are shown at

each temperature. They are labelled

(1) to (4) in order of increasing

energy. Note the stark contrast to the

distribution of fermions shown in fig.

V-4, all because the wave function is

symmetric for bosons and

antisymmetric for fermions.

(d)
(c)

(b)

(a)

»•#••••••••®m»»»Bmm®mm

•••••••

••••••••••••••••a**

(4)

(3)

(2)

CD

(4)

(3)

(2)

(1)

(4)

(3)

(2)

(1)
(4)

(3)

(2)

(1)

temperature, more and more of them acquire thermal energy and get
into the higher energy states, as shown in parts (b), (c), and (d) of
the picture. You can see that the resulting distributions are totally
different from those for fermions. It is therefore to be expected that
the properties of boson systems will be quite different from the
properties of fermion systems.
10 Summary
I introduced in section 1 the idea of a statistical distribution as a
convenient way of handling the individual characteristics of the
members of a very large assembly. I illustrated it with the example of
the distribution by age of the population of a country.

Such a distribution contains interesting and useful information about

the population that is not so easily accessible from a list of the ages
of each of the inhabitants.

In section 2, I considered a volume of gas consisting of a very large

number of atoms, something like 1021. These atoms are moving
about at all possible speeds, bouncing off one another and off the
walls of the container. I showed a picture of their distribution by
speed at two temperatures. The distribution changes with a change
in temperature, such that the average speed of an atom increases
when the temperature increases. I introduced a new way of
expressing temperature, in kelvin (K), obtained by adding the
number 273 to the temperature in degrees Celsius (°C). Thus a
temperature of 20 °C is the same as 293 K. The temperature in
kelvin is called the absolute temperature.

I took a closer look, in section 3, at the distributions by speed of

atoms in a gas, and found that the average kinetic energy of an atom
is proportional to the temperature expressed in kelvin: doubling the
temperature of the gas doubles this energy. I thus have IOI

V STATISTICAL PHYSICS

a connection between the average energy, which depends on the

energy of each of the 1021 atoms, and the temperature, which is a
property of the gas as a whole. This connection is through the
Boltzmann constant kB. The making of such connections between
atomistic and macroscopic aspects of matter is the essence of
statistical physics.

I introduced the idea of the absolute zero of temperature in section 4,

as a consequence of the average energy being proportional to the
absolute temperature of the gas. In practice, one can never reach
the absolute zero. Nature is such that the closer one gets to it, the
harder it becomes to get any closer.

I saw in section 5 that the temperature of a crystal is related to the

energy of vibration of its atoms about their average positions. These
thermal vibrations vanish at the absolute zero of temperature,
leaving only the zero-point vibrations of the atoms.

I considered the quantum mechanics of an assembly of identical

particles in section 6. I found that interchanging any two particles in
the assembly must either leave the wave function unchanged (a
symmetric wave function), or must change its sign (plus to minus or
vice versa) and nothing else (an antisymmetric wave function).
Which of the two happens depends on whether the particles are
electrons, or photons, or hydrogen molecules, etc. Particles that are
described by symmetric wave functions are called bosons, and the
other kind are called fermions. For example, photons and hydrogen
molecules are bosons, while electrons and protons are fermions.

I showed in section 7 that because fermions have antisymmetric

wave functions, they obey the Pauli principle: no two fermions in the
assembly can have all their quantum numbers the same. At least
one of the quantum numbers must be difierent for the two fermions.

In section 8 I looked at the quantum mechanics of a particle in a

three-dimensional box. This is an extension of the case of a particle
in a one-dimensional box that I described in chapter III. I found that
the wave vectors and the energies are quantized: they can have only
certain values and nothing in between. These wave vectors, along
with the spin, can therefore be thought of as the quantum numbers.
The quantum states are degenerate: states with different quantum
numbers (i.e. wave vectors and spin) can have the same energy.
I took, in section 9, the case of a large number of particles in the box.
The way these particles occupy the different quantum states is quite
different depending on whether they are fermions or bosons,
because of the Pauli principle. I showed with the help of pictures how
the different states are occupied at different temperatures in the two
cases.

This account of statistical physics, together with the preceding

chapters on crystals, quantum mechanics and atomic structure,
provides the foundation on which we can now build an
understanding of why things are the way they are. As the first step,
we shall look at a crystal that is at some temperature and is
composed of identical atoms each consisting of a nucleus and its
surrounding electrons, to be described using the quantum
mechanics and statistical physics which we have now learned. This
is the subject of the next chapter.

IO2

VI THE QUANTUM MECHANICAL CRYSTAL

l From isolated atom to crystal

A crystal is an orderly arrangement of atoms that are sitting rather

close to each other, as we saw in chapter II. We then looked at the
basic ideas of quantum mechanics in chapter III. We used these
ideas to get a picture of an isolated atom in terms of its nucleus and
especially its electrons in chapter IV. The material things that we deal
with in our daily life consist not of single atoms or electrons, but
rather of very large numbers (like 1021) of particles. We need a
statistical approach to cope with such numbers. We saw in chapter V
that quantum mechanics, when used to give such a description of
assemblies of particles, divides them into two types: fermions and
bosons.
The only property that the crystals of chapter II had was their
symmetry. A real crystal has in addition many other properties like
being a good (or bad) conductor of electric currents, having a colour,
and so on. These properties depend on the particular kind of atoms
that make up the crystal. For example, silver and silicon are both
crystalline, but differ widely in many of their properties. The only
essential difference, other than in the mass, between the atoms of
any two elements is in the number of electrons surrounding the
nucleus. So we can conclude that this difference must play a key
role in determining the specific properties of materials. It is therefore
reasonable to see what happens to the electrons when atoms come
close together to form a crystal.

We proceed to do this using the ideas of both quantum mechanics

(because that is the correct description of nature) and statistical
physics (because of the large number of particles involved, and also
because we shall be interested in effects at different temperatures).

I remind you that the typical size of an atom is a few angstroms (one
angstrom =

10s cm), which is also the typical distance between neighbouring

atoms in a crystal. In an isolated atom that is far from all other
atoms, each electron feels the influence of the electric charge of the
nucleus and of the other electrons in the atom, and nothing else.

The result is the set of energy levels and quantum numbers

described in chapter IV.

When the atoms come together to form a crystal, an electron from a

given atom feels the electric charge not only from that atom, but also
from neighbouring atoms. It seems reasonable that the wave
functions, energy levels and quantum numbers should now be
different from what they were in the isolated atom. We shall find out
in the next section what these differences are.

103
FROM ENERGY LEVELS TO ENERGY BANDS

+47e, is very small compared to the size of the atom. The differently
shaded regions show where electrons with different quantum
numbers n are most likely to be, and the number of these electrons
is shown for each region. Thus there are two electrons with n = 1
very close to the nucleus, and one electron with n = 5 near the
surface of the atom, and the remaining electrons are distributed in
between.

The energy states of the hydrogen atom correspond to the values of

the quantum number n (p. 72). For a given value of n, the different
possible values of the other quantum numbers , m and s gave
different wave functions but the same energy: the energy level was
degenerate. The situation is different in an atom with several
electrons, as happens with silver for example. Here, the energy of a
given electron is determined not only by the nucleus, but also by all
the other electrons. So the wave functions that correspond to a given
value of the quantum number n will have slightly different energies
for the different sets of values of their quantum numbers , m, and s.

figure vi-2. a few of the energy n energy levels

levels for an atom with many

electrons, such as silver. There are, for

example, 18 levels for the quantum

number n — 3 corresponding to the

different possible values of the

quantum numbers /, m and s. -^

A
toO

c3

S-H

105

VI THE QUANTUM MECHANICAL CRYSTAL

I show the resulting picture of the energy levels in fig. VI-2 for the
quantum number n equal to 1, 2, and 3. For each of these values of
n, there is a certain number of levels that are clustered together, and
there are gaps in energy between adjacent clusters. Each level
corresponds to a different wave function with its set of quantum
numbers n, I, m and 5. We know from the Pauli principle that there
can be at most one electron described by each of these wave
functions with the corresponding energy.

An actual piece of solid, weighing a few grams, consists of a very

large number N

(something like 1021) of atoms sitting very close to one another. We

can see what the energy levels for the electrons in such a solid look
like by going through the following imagined experiment. We begin
by putting the N atoms very far apart from one another. Each atom
then feels practically no effect from the presence of the other atoms,
since the electric force decreases as the distance increases. The
energy levels therefore will look practically the same as for the
isolated atom (fig. VI-2). But there will now be JV electrons in each
energy level, one electron belonging to each of the N

atoms. This does not violate the Pauli exclusion principle, because
each of these electrons has a wave function belonging to a particular
atom with vanishing probability of being near any other atom. This is
equivalent to saying that the wave functions of the different atoms do
not overlap.

The atoms are now gradually brought closer and closer together,
until they finally end up in the positions they have in the actual
crystal. The atoms are now crowded close together, and we can no
longer neglect the mutual influence of electrons from nearby atoms
on one another. Because of this, each of the AT electrons that was in
a given energy level when the atoms were far apart will now have a
different energy in the crystal. Each discrete energy level in the
isolated atom now becomes a band of energy levels in the crystal,
lying very close to each other and numbering as many as the
number of atoms in the crystal. This state of affairs is illustrated in
fig. VI-3.1 show the bands as just grey patches in the picture. If you
imagine it hugely magnified, you will see the distinct levels, each
capable of accommodating one electron (the Pauli exclusion
principle). The electrons in the crystal are accommodated in these
energy bands, starting at the lowest energy level in the lowest band,
and going progressively higher in energy. You see that the basic idea
is the same as what we had with the atom, starting with the nucleus
and progressively adding the electrons to the different energy levels.
The difference is that the atomic energy levels arise from a single
isolated atom, whereas the energy bands for the crystal are due to
all the atoms of the crystal taken together.

You notice that the different bands in fig. VI-3 are separated by
energy gaps. It happens in some materials that a pair of adjacent (in
energy) bands overlap, instead of having a gap between them. The
net effect is still that of bands of allowed energies separated by
gaps, as in fig. VI-3. An electron can only have an energy that lies in
one of these bands but not an energy in any of the gaps. An
analogy: it is as if the rules of travel for cars on a highway were such
that they could only move at speeds between, say, zero and 20
km/hour, or between 30 and 40 km/hour, and so on. A speed 106
INSULATORS AND METALS

FIGURE VI-3. The energy levels for

different quantum numbers n in the

isolated atom (left). The further

splitting of the levels for each n (see

fig. VI-2) is not shown. In the crystal

(right), there are bands of energy

levels separated by energy gaps. Each

of these bands has at least as many

energy levels as there are atoms in

the crystal.

energy levels in

isolated atom

energy bands in

crystal

.a

between 20 and 30 km/hour is just not allowed. We would then talk

about bands of allowed speeds and bands of prohibited speeds. The
corresponding rules for electrons in a crystal are those of quantum
mechanics. They say that electrons can only have energies in
certain bands, and not in between: we have allowed and prohibited
bands of energy. This is a fundamental property of electrons in a
crystal, and leads to an explanation of the difference between
insulators which cannot carry an electric current, and metals which
can, as we shall see in the next section.
3 Insulators and metals
We now populate the energy bands with the electrons in the crystal,
beginning with the lowest energy in the lowest band, and going to
progressively higher bands until all the electrons have been
accommodated. The resulting picture is shown in fig. VI-4.

Depending on the total number of electrons per atom, which in turn

depends on which chemical element makes up the crystal, the
highest band that has electrons will either be fully occupied, or only
partially occupied. The two situations are illustrated in the left and
right parts of the figure respectively. When the highest band with
electrons in it is fully occupied, as on the left of the figure, we get an
insulator. When it is only 107

VI THE QUANTUM MECHANICAL CRYSTAL

FIGURE VI-4. Energy bands showing

occupation by electrons, in an

insulator (left) and a metal (right).

In the insulator, the band with all its

states empty (shown at the top), i.e.

without electrons, is separated by an

energy gap from the next lower band,

in which all states are occupied by

electrons. In the metal, the band of

highest energy with electrons in it is

only partly occupied and has empty

states left.

insulator

metal

energy gap J

ID

occupied states

empty states

partially occupied, we get a metal. Electrons in the highest energy

states in a metal can absorb extremely small amounts of energy
because they have suitable empty states to which they can go.
Electrons in an insulator, on the other hand, can only absorb energy
equal at least to the gap energy, and nothing less. Figure VI-5
illustrates the differing responses of an insulator and a metal to small
additions of energy due to a rise in temperature or an applied
voltage, for example.

I have described so far how the electrons are distributed in the

energy bands. One can also think of how the electrons are
distributed in space in the crystal itself. For this I shall take the
example of silver as a typical metal, and common salt and silicon as
typical insulators. In these as in all other solids, only a few electrons
which are farthest from the nucleus are affected significantly; all the
other electrons are nearly in the same 108

INSULATORS AND METALS

FIGURE VI-5. The effect of a small

energy input like an applied voltage

or a rise in temperature on an

insulator (left) and a metal (right).

In the insulator, the electrons are in a

full band, there are no empty states

nearby in energy into which they

could go in response to the energy

input, which consequently has no

effect on the electrons. In the metal,

on the other hand, the electrons

respond by changing energy and

moving to states nearby which were

previously empty (dark circles),

leaving behind empty states (light

circles).

insulator
energy J gap

metal

energyjgap

occupied states

empty states

positions with respect to the nucleus as in the isolated atom. This is

to be expected, for the outermost electrons of a given atom have the
weakest binding to the atom, and will therefore be influenced by
neighbouring atoms most strongly.

I show in fig. VI-6 what a cross-section through a plane of atoms in a

crystal of silver looks like, and for comparison a section through an
isolated silver atom. The picture is not to scale, but shows the
features of interest. The different shaded regions indicate where the
electrons with different quantum numbers n are most likely to be
found. The electrons with quantum numbers from 1 to 4 are little
affected by being in the crystal.

The electrons with n = 5, however, have parted company from their

parent atoms and can be found anywhere in the crystal with equal
probability. Each atom is now an ion with an electric charge +e, and
the crystal is a lattice of these ions immersed in a sea of electrons.

question Why is the charge on the ion equal to +e?

answer The atom has a charge +47e on the nucleus surrounded by

47

electrons each with a charge —e, so that its total charge is zero. The
outermost electron has wandered off in the metal, leaving behind an
ion with a total charge of+47e-46e, or just +e.

109
ELECTRON WAVES IN A CRYSTAL

FIGURE VI-10. What the wave looks

like for a free electron in a region of a

box (broken curve) and for an

electron in the same region of a

crystal that has the same size and

shape (continuous curve). There is a

row of atoms (not shown), one in

each of the segments between the

points labelled 0 and 1, 1 and 2, ….

The wave in the crystal, called a

Bloch wave, has the same shape at

each atom, but its overall amplitude is

like that of a free electron wave.

belonged to a different wave function, which could be associated

with two electrons with spin up and spin down respectively, both
having the same energy.

What do the waves look like for the electrons that were in the outer
shells of the atoms in a crystal? The answer is illustrated in fig. VI-
10. The continuous curve shows a typical wave for such an electron
in a piece of crystalline solid as one goes along a line of atoms in
some region inside it. The broken curve shows a wave for a free
electron in the same region of a box that has the same shape and
size as the solid. The wave for the electron in the crystal has a
pattern of wiggles at each atom, but its overall amplitude changes
from one atom to the next like that for a free electron wave. One
would expect this kind of picture, if one recalls that the wave
amplitude at any point tells us about the probability of finding the
electron in the neighbourhood of that point. The probability of finding
the conduction electron inside the ion is strongly influenced by the
other electrons and the nucleus of the ion itself and leads to those
wiggles. Since the electron can move from atom to atom, it also has
the aspect of a free electron wave. It is rather like some folk dance in
which there are a number of identical groups of dancers at different
places on the floor. These are the electrons in the ions. The outer
electrons are like individual dancers who spend a little time with each
group before moving on to the next group. These waves of the
electrons in a crystal are called Bloch waves, after the physicist Felix
Bloch who first thought of them. Such electrons are referred to as
Bloch electrons, to distinguish them from free electrons in a box.

For the free electron in a box, we saw in chapter III that there is a
relation between the energy E and the wave number k,

b2k2

E=-

2m ■

For the electrons in the solid, I would expect the variation of £ as k

varies to be a little more complicated, because there are all those
ions influencing the electrons. I show in fig. VI-11

how this variation, for fixed direction and changing magnitude of the
vector k, might look for two adjacent energy bands with a gap
between them. As the wave number increases, VI THE QUANTUM
MECHANICAL CRYSTAL
the energy in the lower band increases and the energy in the upper
band decreases. In this figure, each of the little squares, like the one
labelled P, represents a possible state with its specified wave
number and energy. For the state P, I can denote the values of wave
number and energy by £[P] and E[P] respectively, as read offon the
two scales. There can be a maximum of two electrons, with opposite
spins, in each of these states (Pauli exclusion principle). I show a
small number of states in each band to avoid cluttering the picture.
The actual number is comparable to the number of atoms in the
crystal, about 1021.

Figure VI-11 gives us more information about the energy bands than
is shown in fig. VI-3. We see that each band is made up of closely
spaced states, each with its wave function, wave number k, and
energy E. There is no state with energy lying in the gap between the
two bands. If a given electron changes from one state to another, the
FIGURE VI-11. Two neighbouring

energy bands, showing the discrete

quantum states (small squares) of

which P is an example, with its

specific wave number k[P] and energy

£[P], This and the following pictures

of energy bands are qualitative, and

therefore no numbers appear along

the axes.

increasing wave number—

114
ELECTRON WAVES IN A CRYSTAL

figure tells us how much the corresponding changes are in the wave
number and energy of the electron. An electron can go from one
state to another within a band with very small change of energy, but
to go from one band to the other must involve a change of energy
equal at least to the energy gap.

The energy in a given band may either increase or decrease as the

wave number increases. In fig. VI-11,1 have arbitrarily chosen the
lower band to be one way and the upper, the other. Both types occur
among bands in solids, and indeed one can have in the same band
both kinds of dependence of the energy on the wave number.

question With electrons in a box, the energy always increased when

the wave number increased. Why is it sometimes different with
electrons in the crystal?

FIGURE VI-12. Some of the different

possible types of curves showing how

energy varies with wave number for

electrons in a crystal. Such curves are

said to represent dispersion relations

for the electrons, and are called

dispersion curves.

increasing wave number

VI THE QUANTUM MECHANICAL CRYSTAL

answer You see from fig. VI-10 that the electron wave in a crystal is
heavily disturbed near the ions, while in the box it looks quite smooth
everywhere. When the energies for such wave functions are
calculated, it turns out that they sometimes increase, and sometimes
decrease, as the wave number increases.

I show in fig. VI-12 a few of the possible shapes for bands. In this
figure, I have not shown the little squares representing the separate
states in each band as in fig. VI-11, but have joined them as smooth
lines. For an actual crystal, there is one of these graphs of energy £
versus wave number k for each direction that the wave vector points
in the crystal, looking like so many strands of spaghetti in a bowl.
Such curves are called dispersion curves. A collection of all such
pictures for a given solid is called the electronic FIGURE vi-13.
Energy bands in an

insulator (or pure semiconductor) at

the absolute zero of temperature. All

the states in the lower band are

occupied by electrons (filled squares),

and there are no electrons in the states

in the upper band (empty squares).

increasing wave number

n6

ELECTRON WAVES IN A CRYSTAL

band structure, or sometimes simply the band structure, of the solid.

The crystal structure and the particular element or elements making
up the solid determine the actual band structure. It is used to deduce
many of the properties of the solid, and we shall next see an
example of this.

The electrons in the crystal occupy the states in the energy bands,
starting with the lowest energy, until all the electrons have been
accommodated. Each quantum state of given energy takes two
electrons with opposite spin. Then, depending on how many
electrons there are altogether, the last band containing electrons will
be either exactly full, or still have unoccupied states. These two
possibilities are shown in figs. VI-13 and VI-14 respectively.

The situation in fig. VI-13 is that of an insulator at the absolute zero

of temperature.

The lower band is completely full, the upper band is quite empty, and
there is an energy gap between the two. To set up an electric current
means that some electrons have to change to states of slightly
higher energy. They cannot do this since the lower band has no
empty states for the electrons to move into, and the empty states in
the upper band are too far away in energy to be accessible to the
electrons. As a result, applying FIGURE VI-14. Energy bands in

a metal at the absolute zero of

temperature. All the states up to a

maximum wave number (the Fermi

wave number) in the lower band are

occupied by electrons, the remaining

states are empty. The energy of this

highest occupied state is the Fermi

energy.

increasing energ
noon

4—

“X,

♦ occupied

states

° empty states,

upper band

« empty states,

lower band

~ Fermi energy

“‘1™*-

increasing wave number

VI THE QUANTUM MECHANICAL CRYSTAL

an electric voltage to the crystal produces no electric current: we

have an insulator.

Figure VI-14 shows how the bands are occupied in a metal. The
highest band that has electrons still has unoccupied states left. All
the lower bands (which are not shown in the picture) are completely
occupied by electrons, and the next higher band (shown) is empty.
Now if an electric voltage is applied to this crystal, electrons in states
below the one that is marked ‘Fermi wave number’ (more about this
in a moment) have empty states to move to, and so an electric
current results.

The electron waves can be travelling in all directions in the space of

the crystal, and one should properly speak about their wave vectors.
If I look at the lower band in fig.

VI-14 and think about it in terms of wave vectors, then I would say
that all states with wave vectors up to a certain maximum length in
that direction are occupied by electrons, and states with larger wave
vectors are empty. This largest wave vector is called the Fermi wave
vector, after the physicist Enrico Fermi. There is such a vector for
each direction in the crystal, though their lengths may not all be the
same; after all, the crystal looks different in different directions, and
so do the electron waves. The magnitude of this vector is the Fermi
wave number.

Now imagine a point in space from which radiate arrows in all

directions, each being the Fermi vector in its direction. The result will
be a hedgehog-like object, and the surface that encloses it, touching
the tip of each arrow, is the Fermi surface. Figure VI-15 shows what
a cross-section through a Fermi surface and four typical Fermi wave
vectors might look like. All the quantum states whose wave vectors
lie inside the Fermi surface are occupied by electrons, and all those
outside are empty. The energy for each Fermi wave vector is the
same, and is called the Fermi energy.

question Why should the energy for all the Fermi wave vectors be
the same?

FIGURE VI-15. Cross-section of a

Fermi surface, indicated by the closed

contour, at the absolute zero of

temperature. All the states inside the

surface are occupied by electrons, all

the states outside are empty. Each

state is identified by its wave vector

and energy. Four typical Fermi wave

vectors are shown, each going from

the centre to the Fermi surface.

Though they point in different

directions and have different

magnitudes, they represent states

which have the same energy, namely

the Fermi energy.

Fermi surface

empty states

n8

ELECTRON WAVES IN A CRYSTAL

answer Suppose they all had different energies. Then there would be
a shifting of electrons from state to state until they all had the same
energy.

The stable situation is where each electron is in the lowest energy

state possible. This is achieved by filling all the energy states of
different wave vectors with electrons up to the same energy, for then
no electron has an empty state of lower energy to which it can go.

question What does the Fermi surface look like for electrons filling a
box?

answer Since there is no crystalline lattice whose anisotropy would

lead to different properties in different directions, the length of the
Fermi vector must be the same in all directions. So the Fermi surface
will be a sphere.

A metal like silver has tens of electrons per atom. Most of these
electrons in the metal are in completely filled energy bands, and play
no part in carrying a current.

It is only the relatively few electrons in the states of highest energies,

with nearby empty states, that carry the current. Put differently: only
those electrons that are in states very close to the Fermi surface
take part in electrical conduction. All the other electrons are, in a
sense, just passive observers of the scene. This is true not only for
electrical conduction, but also for most other properties of metals, as
we shall see.

So when I think of a metal, it helps not only to think of the regular

crystalline arrangement of the atoms in it, but also of the Fermi
surface associated with the electrons in it. It is in a space in which
lengths are measured in wave numbers: the units are reciprocal
centimetres (cm”1), and not the usual centimetres you find on a tape
measure.

question Can one talk about a Fermi surface for an insulator also?

answer It is true that in an insulator also there is, in each direction in

the crystal, a wave vector of maximum length for the highest
occupied state, as one can see in fig. VI-13.1 can construct a
surface using all these wave vectors in exactly the same way that I
got a Fermi surface for a metal.
The resulting surface, called a Brillouin zone after the physicist Leon
Brillouin, has nothing to do with the electrons at all, but is solely
determined by how the atoms are arranged in the lattice. This comes
about essentially because (a) the electron waves are Bloch waves,
(b) the bands are either fully occupied or empty, and (c) they satisfy
the Pauli exclusion principle. I must let it go at that; to explain it
further will take us far afield. This idea is not useful for an insulator
since we are concerned with its electronic properties, and the
Brillouin zone just represents the crystal lattice and not the electronic
structure. The pictures of energy bands, like those in figs.VI-12, are
on the other hand more meaningful for insulators.

119

VI THE QUANTUM MECHANICAL CRYSTAL

5 Semiconductors
You have probably come across the word semiconductor in
newspapers and elsewhere, and may wonder what it has to do with
the metals and insulators that we have been talking about here. A
semiconductor, as the name implies, lies somewhere between
metals and insulators in its ability to carry an electric current. It is
really an insulator to start with, but something happens along the
way so that it ends up with a few electrons in states that have empty
states very close by in energy, and it is now able to carry an electric
current. What happens to it is that it is at some temperature above
the absolute zero, or is not pure but has some impurities in it, or is
both.

Take the case of temperature first. We saw in chapter V that at any

temperature the thermal energy is distributed among the particles
over a wide range, all the way from very low to very high energies.
So with the insulator at some temperature, there will always be some
electrons whose energies are larger than the gap energy. These
electrons can therefore move into empty states in the upper band
and leave empty states behind in the lower band, as shown in fig. VI-
16. Notice in this figure that there are now empty states close to
states with electrons in both the upper and the lower bands, so that
current can be carried by electrons in both bands. There is a
difference, however: the upper band has a few electrons sitting in an
otherwise empty band, while the lower band has a few empty states
in an otherwise filled band. It turns out that the current in the lower
band is exactly the same as what one would have in a similar band
containing a new kind of particle called a hole, equal in number to
the number of empty states in the original band.

The hole is like an anti-electron: each of its properties (like energy,

wave vector or spin) is the negative of that of the electron that
previously was in that quantum state.
When I work out how a hole moves when I apply a voltage, I find that
it behaves like an electron, except that it has a positive charge +e.
This means that if the electron moves to the right because of the
voltage, the hole moves to the left. Since an electric current is just a
moving stream of charged particles, the upper band gives a current
of electrons, and the lower band gives a current of holes. Because
their charges as well as the voltage-induced motion are opposite, the
two currents are in the same direction and add to each other. The
physical reality is that the current is carried by electrons in both
bands. The concept of a hole is only a way of describing the
properties of the lower band with some empty states. It avoids
having to worry about what the huge number of electrons do, and
instead focuses on the few holes alone.

This picture of a hole applies also in a metal to the Fermi surface

with an empty state left behind by an electron, when it acquires
some energy and moves to a state outside the Fermi surface. The
energy needed to move an electron to another state must at least be
equal to the energy gap in a semiconductor, but can be almost
negligibly small in a metal. It is this difference that essentially
distinguishes a semiconductor from a metal, and ultimately makes
possible the use of semiconductors in transistors, chips, and all the
other wonders of the electronic age.

SEMICONDUCTORS

FIGURE VI-16. An insulator (or pure

semiconductor) at some temperature

above the absolute zero. Some of the

electrons in the lower band have

acquired enough thermal energy to

cross the energy gap and go into

states in the upper band, leaving

empty states behind in the lower

band. Now it is possible to set up an

electric current, which is carried by

electrons in both bands, because there

are empty states next to occupied

states in both. The current in the

lower band can be thought of as

being carried by positively charged

particles called holes.

increasing wave number

An insulator can change to a semiconductor even without its

temperature being raised, when it contains atoms of certain other
elements. We saw how in silicon the four outermost electrons on
each atom form the bonds with neighbouring atoms. There is an
energy gap between the fully occupied band (which is where these
four electrons are) and the next higher band, which is empty of
electrons. Now suppose we melt some silicon, add a pinch of
phosphorus, and then make a crystal out of the mixture. It will be a
crystal of silicon doped with phosphorus, as the people who make
and use such things call it. There are phosphorus atoms in some
places in the crystal where there would have been silicon atoms, but
otherwise the crystal looks the same as one of pure silicon. The
phosphorus atom is only slightly different from the silicon atom: the
charge of the nucleus is 15 e instead of 14e, and there are five
electrons in the outermost VI THE QUANTUM MECHANICAL
CRYSTAL

shell, instead of four as in silicon. Four out of these five electrons

again are used to form the bonds to neighbouring atoms and
completely fill up an energy band, just as in pure silicon. But now we
have an extra electron from each phosphorus atom which must find
a home somewhere. The only place for it to go is the next energy
band, which was empty in pure silicon. This is illustrated in fig. VI-17.
The electrons in states in the upper band have empty states very
close by in energy, and can therefore carry a current If I dope silicon
with a bit of aluminium instead of with phosphorus, the upper band
remains empty but now empty states appear in the lower band. This
is because the aluminium atom has one electron less than the silicon
atom in its outer shell. Some of the bonds therefore have only one
electron instead of two, leaving empty states in the lower band that
behave as holes.

FIGURE VI-17. Energy bands in

phosphorus-doped silicon. All states

in the lower band are occupied by

electrons, just as in pure silicon. The

extra electron from each phosphorus

atom goes into a state in the upper

band, and can contribute to an electric

current.

o
 ■

♦ occupied,

lower band

° empty,

upper band

- occupied,

upper band

”’^

♦♦**

increasing wave number

THERMAL ENERGY OF A CRYSTAL

The two kinds of doped semiconductors described above are said

respectively to be n-type (negative electrons) and p-type (positive
holes). Note that such semiconductors can have electrons (or holes)
even at the absolute zero of temperature. As the temperature is
raised, additional electrons from the lower band gain enough thermal
energy to move to states in the upper band, leaving behind empty
states in the lower band equivalent to holes, as with the pure
semiconductor.
6 Thermal energy of a crystal
We saw in chapter V that changing the temperature of an object
changes the energies of motion of the particles that make up the
object. The total change for all the particles is the change in the
thermal energy of the object produced by the change in temperature.
In a metal, the moving particles are the conduction electrons that are
reasonably free to move around in the crystal, and the ions that
cannot ran away from their lattice sites but can vibrate about their
average positions. In an insulator (or semiconductor), we have the
atoms that again can vibrate about their positions, and we may have
some mobile electrons and holes also.

Consider a metal at some temperature T. (I shall use this symbol for

the temperature when I want to leave its actual value unspecified;
otherwise I shall say that the temperature of something is 300 °C or
800 K or whatever.) The particles in it are the ions and the
conduction electrons. The electrons are fermions, and so their
distribution among their energy states will be as shown in fig. V-4 (p.
99). If an electron is to take on additional energy as thermal energy
because of the temperature, it must have an empty state of energy
that is higher by this amount into which to move. The Fermi energy
for the electrons in a metal is about a hundred times their average
thermal energy. The consequence of this is that only electrons in a
very small range of energies (approximately the average thermal
energy) around the Fermi energy can take on the thermal energy in
addition. Most of the electrons are unaffected by temperature,
because it takes more energy for them to move into empty states
than can be provided by the thermal energy.

Figure VI-18 shows what this state of affairs looks like in terms of the
Fermi surface, which I have taken to be the same as the one in fig.
VI-15. Some electrons that were close to the Fermi surface acquire
thermal energy and move to states of higher energy just outside the
surface. This leaves an equal number of empty states inside the
Fermi surface, which behave like holes. You notice that only a thin
layer of states at the Fermi surface is affected by temperature. It
turns out that our interest is in knowing how much increase in energy
takes place with an increase in temperature, rather than the actual
energy at any given temperature. This means that we can think of
the thermal energy as being used to create electron-hole pairs, an
example of which is shown in fig. VI-18 on the right. An electron in
state A inside the Fermi surface gains thermal energy and changes
to state C outside the surface, leaving an empty state at A. One can
divide this gain in energy into two parts: the energy in going from A
to B (at the Fermi surface) is 123

VI THE QUANTUM MECHANICAL CRYSTAL

empty states

c9

energy o:

Feraii

‘ electron

surface

““X occupied slates /’”” energy uf hole

\ / ‘V

\ / * electron

^■s./’ ° empty state

FIGURE VI-18. The same Fermi surface as in fig. VI-15 but now at
some temperature above the absolute zero (shown on left), and an
electron-hole pair (on right). The thermal energy is very small
compared to the Fermi energy at all temperatures of interest. This
means that only a few electrons very near the Fermi surface have
empty states of slightly higher energy to which they can go after
gaining thermal energy (dark circles). These electrons leave behind
empty states (open circles), equivalent to holes. The distribution of
electrons in all the other states is unaffected by temperature; these
electrons have nowhere to go, because all nearby states are already
occupied by other electrons. The thermal energy is divided between
the electron and the hole as shown on the right and described in the
text.

thought of as the energy of the hole, and the energy in going from B
to C as the energy of the electron. The thermal energy can create
these electron-hole pairs from only a small fraction of the total
number of electrons, namely those with energies close to the Fermi
energy. This means, as we shall see in chapter VII, that only a small
part of the thermal energy of metals is carried by the electrons and
holes. The greater part is in the vibrations of the ions.

With insulators (or semiconductors, pure as well as doped), this is

even more so, for there is an energy gap to the nearest empty states
for the electrons. So even fewer of them make it across the gap with
their additional thermal energies, leaving empty states behind. For all
practical purposes, the whole thermal energy is carried only by the
ions. It is only at very low temperatures, where the total thermal
energy is very small, that a significant fraction of it is in the electrons
and holes.
7 Phonons
We now look more closely at the thermal vibrations of the atoms (or
ions) in the crystal. This is energy of motion, and is therefore
connected with the mass of the ion and not its electric charge. We
want to find a way of describing this motion quantum-124

PHONONS

mechanically. We have seen that quantum mechanics is intimately

connected with wave motion, and so it is reasonable to see if we can
somehow bring waves into the picture of the vibrating atoms in the
crystal. If I send a sound wave of a certain frequency down a crystal,
and imagine that I could look at the individual atoms, I would find that
each atom is vibrating at that frequency. In addition, at any given
instant of time the positions of the atoms would trace out a wave of
the corresponding wavelength. So there seems to be some promise
in trying to connect sound waves with atomic vibrations.

question How do we know that sound can travel through solids?

answer Because if there is a loud enough noise inside a room, I can

hear it from outside even if the doors and windows are shut tight. I
can hear the person in the flat upstairs moving furniture about,
because the sound comes through my ceiling.

In a solid at some temperature, the atoms are vibrating by

themselves even though no sound is fed in from outside. We saw in
chapter II that the influence of atoms on one another in a solid is as if
they were connected by springs. For simplicity consider a short one-
dimensional crystal, with just 11 atoms: very short, indeed. Assume
that the atoms at the two ends are clamped, so that they cannot
vibrate. Figure VI-19 shows five of the waves that could exist in this
lattice, with each wave displaced vertically for clarity. Lines connect
the atoms so that the wave shape is easily seen; you can think of the
lines as representing the imaginary springs between pairs of atoms.
We look first at the lowest wave in the figure, labelled ‘fundamental’.
There is exactly half a wavelength in this length, and so its
wavelength is 2L if the length of the crystal is L. The speed v of a
wave is equal to its wavelength multiplied by its frequency, and so its
frequency^ is given by V

~2L—

FIGURE VI-I9. Some modes of

vibration of a one-dimensional crystal

with 11 atoms, which are shown as

small squares. The lines joining the

atoms are to make it easy to see the

wave shapes. They may be thought

of as indicating the virtual springs

connecting the atoms. Five of the

possible modes are shown. In each

case, the line of atoms is vibrating

between the extreme positions shown

by the continuous and broken lines.

fundamental

positions of atoms

125
VI THE QUANTUM MECHANICAL CRYSTAL

The next wave, labelled ‘second harmonic’, has a wavelength equal

to L. If the speed of sound is assumed to be the same for all
wavelengths, then the frequency^ of this wave is

or

f2 = 2f,

A pattern seems to be emerging for the frequencies of the

successive harmonics. The third harmonic has a wavelength equal
to two-thirds of the length L, and therefore its frequency^ is

3v

fi = v divided by 3I or ^-, which is the same as 3fu

and the fourth harmonic has a frequency ji equal to 4/j. The

harmonics have frequencies given by multiplying the fundamental
frequency fx successively by the integers2, 3, … .

There is a simple experiment you can do on a piano that will

illustrate this property of the harmonics. The predominant tone you
hear when you play the note of middle C

has a frequency of 256 vibrations per second: this is the

fundamental. The successive octaves of this note have frequencies
that are two, three, … times this frequency. Now gently depress the
key an octave above middle C, so that its damper is off the strings
but the note does not sound. Then play middle C and let the key up
at once. You will hear clearly the tone of the octave, which is the
second harmonic of what you just played. This is because when you
played middle C, its strings vibrated not only at the fundamental but
also at the higher harmonic frequencies. The strings an octave
above pick up the sound waves of the second harmonic and begin to
vibrate at the same frequency. You get similar results if instead you
hold down keys which are two, three,…, octaves above the note you
play. The experiment shows that the tones on the piano contain the
fundamental as well as the harmonics.

Now look at the wave labelled ‘top harmonic’ in fig VI-19. The
neighbouring atoms are displaced in opposite directions from where
they would be if there were no wave. The wavelength here is just
twice the interatomic distance: I cannot have a wave that has an
even shorter wavelength. So I conclude that there is a minimum
wavelength for the lattice vibrations and a corresponding maximum
frequency of vibration. Since the wave number is just the reciprocal
of the wavelength, I can equivalently say that there is a maximum
wave number that I shall call &□ and a maximum frequency^) for the
waves. I use the subscript D here to recognise Peter Debye, the
physicist who first thought of the thermal vibrations of atoms in solids
in terms of sound waves.

Each of the vibrations of this lattice has its frequency as well as

wave number, and fig. VI-20 shows how the frequency depends
upon wave number. I have shown in this figure all the possible
modes of vibration, up to the top harmonic. This picture summarises
all the information about the vibrational modes of the crystal, except
how big the vibrations are, namely their amplitudes, and we consider
this aspect next.

126

PHONONS

FIGURE vi-20. The frequencies of the

different harmonics for the example

shown in fig. VI-19. The picture is

qualitative, and no scale is shown for

the frequencies. As the harmonic

number increases (which is the same

as saying the wave number increases),

the frequency of vibration increases.

Since energy is equal to the frequency

multiplied by the Planck constant,

this picture also shows how energy

varies with wave number, and is

therefore the dispersion curve for the

vibrations of the crystal of fig. VI-19.

g-
4567
harmonics
10

question The speed of a wave is equal to its frequency divided by its

wave number. Figure VI-20 shows that this speed is constant for
small wave numbers, but becomes somewhat smaller for the largest
wave numbers. Why is this so?

answer The smallest wave numbers correspond to the longest

wavelengths, and at these wavelengths a single wave extends over
many atoms: something like 107 or so in a crystal about one cm on
the side. So on the scale of the wavelength, the discrete atomic
nature of the crystal is smoothed out, and the speeds are the same.
For the largest wave numbers, there are very few atoms, much less
than ten, in a wavelength. In this case, the discrete atoms must be
explicitly taken into account, and the result is that the speed comes
out to be less than for very long wavelengths.

Let us look again at one of the modes of vibration, say the

fundamental with a frequency^. The larger the amplitude of its
vibration, the greater is its energy. According to quantum mechanics,
this energy cannot take any arbitrary value. It can only take the
values Ei(ri) given by

£„(«) = nhfu

where n has one of the values 0, 1, 2, 3,…, and h is the Planck

constant. I have neglected here the zero-point energy of the
vibrations, because it plays no part in what follows. Thus, as
expected, the energy is quantized. This quantum of energy, hfu is
called a phonon. The energy of this mode is determined by the
number n of phonons in it: the larger its amplitude, the more energy
it has and the greater the number of phonons it has. This applies to
each of the possible modes of vibration in the crystal.

So what exactly is a phonon? A phonon is a quantum of energy of

the atomic vibrations of the crystal, characterised by a wave vector k
and corresponding frequency f and n of these phonons have energy
nhf. I can think of the thermal energy of the lattice as if it were
contained in a gas of phonons, whose quantum states are specified
by 127

VI THE QUANTUM MECHANICAL CRYSTAL

FIGURE vi-21. The numbers of

phonons which lie in different energy

ranges, at three different temperatures.

The picture is qualitative, and no

numerical values for the different

quantities are shown. The number of

phonons is proportional to the height

of the bar, and energies increase from

left to right. As the temperature rises,

the number of phonons in each energy

range increases. This means that the

total energy of all the phonons also

rises with increasing temperature.

li low temperature

CH intermediate

temperature

H high temperature

energy ranges

k andyf Each of these states can be occupied by any number n of

phonons, the actual number giving the total energy of this state. We
recall from chapter V that this is exactly the property of the particles
we called bosons, any number of which can be in the same quantum
state. We conclude that phonons satisfy the statistics of bosons. The
actual number of phonons in each quantum state specified by k
and/is then set by the temperature, and varies as the temperature is
varied. I show this variation for three temperatures in fig. VI-21. This
is a picture of the phonon distribution, which shows several notable
features. The number of phonons, and therefore the energy, in each
energy range increases as the temperature increases. At each
temperature, there is a particular energy range that has the
maximum number of phonons, and the number in higher energy
ranges decreases progressively. If you will now go back to fig. V-2

showing the velocity distribution of atoms in a gas, you will see a

similarity to this picture, but with one important difference. The total
number of atoms in a given quantity of gas does not change with
temperature. You can see from fig. VI-21 that the total number of
phonons, got by adding up the numbers that are in each of the
energy ranges, increases with increasing temperature.

Let me pause now and take stock of what we have done. We began
with a crystal whose atoms were vibrating because of the thermal
energy of the crystal. These vibrations were equivalent to sound
waves, whose energies were quantized as phonons. The phonons
belong to the class of particles called bosons. The state of the crystal
at some temperature can be described, as far as the atomic
vibrations are concerned, by saying that the volume of the crystal is
filled with a gas of particles called phonons. There is a kind of
symmetry here with the quantum description of electrons confined in
a box. In that case, we started with objects that we think of as
particles (electrons), and ended by describing them as waves filling
the whole box. Here, with atomic vibrations in a crystal, we began
with sound waves filling the crystal, and ended up with particles
(phonons) moving about in the space occupied by the crystal,

128

THE QUANTUM MECHANICAL CRYSTAL

question Are there any differences between a gas of electrons and a

gas ofphonons?

answer Yes, many. I tabulate some in table VI-2.

Table VI-2. Some properties of electrons and phonons

property

mass

electric charge

statistics

total number in

crystal

electrons

yes

yes
fermion

fixed by number and

kind of atoms

phonons

no

no

boson

varies with

temperature
8 The quantum mechanical crystal
We have come a long way from the crystal lattice of chapter II to the
quantum mechanical crystal of this chapter. The only property that
we gave the crystal lattice was its translational symmetry, from which
emerged other symmetries like those of rotation and reflection. We
then put atoms at the lattice points, let the crystal be at some
temperature, and used the results of the quantum mechanics of
atoms (chapters III and IV) and statistical physics (chapter V) to
describe the resulting situation. We find that the atomic nuclei sit at
the lattice points. The wave functions of the few outermost electrons
are significantly altered from what they were in the isolated atom,
and the remaining electrons are nearly unaffected by being in the
crystal. It is the outer electrons that contribute to those properties of
crystals that are of interest to us. Their energies lie in bands that are
separated by energy gaps. Within each band is a characteristic
relation between the electron’s energy and the wave number of the
electron wave. The electrons are fermions, and so a given quantum
state can hold at most two electrons with spins up and down
respectively. At the absolute zero of temperature the electrons are in
successive energy states in the bands, starting with the state of
lowest energy. The highest energy band with any electrons at all in it
will either have all its states occupied (an insulator), or will have
some fraction of the states empty (a metal). As the temperature of
the crystal is raised, a few electrons which are in the states of
highest energies acquire additional thermal energy, cross the energy
gap and go into empty states in the upper band if it is an insulator or
semiconductor. If it is a metal, electrons move to nearby empty
states in the same energy band. Empty states, equivalent to holes,
are left behind in either case. Most of the electrons in both cases do
not change their quantum states when the temperature of the crystal
is changed.

The effect of temperature on the atoms in the crystal is to set all of

them vibrating about their lattice positions. These vibrations are
equivalent to the fundamental and harmonics of sound waves filling
the crystal. The application of quantum mechanics to these waves
leads to the quantization of their energies, and these quanta are
called 129

VI THE QUANTUM MECHANICAL CRYSTAL

phonons. A detailed study of phonons shows that, just as for

electrons, phonon energies also lie in bands with gaps in between.
One has a relation between the energy (which is the frequency
multiplied by the Planck constant h) of a phonon and its wave
number, again analogous to electrons.

An ordinary crystal will thus present one of two appearances,

depending on how I look at it. With my eyes aided by a suitable
magnifying device, I would see the atoms sitting in a lattice and
jiggling about their lattice positions. I have shown a portion of such a
crystal in fig. VI-9. I could look at this picture and admire its
symmetry, but I would be hard put to conclude anything about its
various physical properties —electrical, optical, magnetic, etc.

Now if I visualise the same crystal incorporating the quantum nature

of matter, then it will be like the left half of fig. VI-22. There is the
volume of the crystal, but inside it are not the atoms, but rather a gas
of electrons and holes, mixed up with a gas of phonons, all colliding
with one another. I also see curves showing how the energy
depends on wave vector for electrons and holes as well as for
phonons, as shown in the right half of fig. VI-22. These are the
dispersion curves for electrons and holes, and for phonons.

Further, if it is a metal, I also see the Fermi surface. It is this view of

the crystal that forms the starting point for understanding why it has
its particular physical properties.

In going to the quantum-mechanical view of the crystal, I seem to

have lost touch with the symmetry of the lattice. Figure VI-22 seems
to contain no indication of the translational symmetry of the lattice. It
only seems so: you recall that for the phonons in the one-
dimensional crystal the largest wave number corresponded to a
wavelength equal to two lattice spacings. The detailed working out of
the quantum mechanics of the electrons in the lattice leads to a
similar limit on the largest possible wave number for the electrons
and holes. The essential aspect of translational symmetry is the
lattice spacing, which determines how the atoms are spaced out.
Thus the existence of such a maximum wave number (or a set of
maximum wave vectors for the three-dimensional FIGURE VI-22. A
crystal of silicon as

seen by an eye that gets the quantum

mechanical picture. The eye sees a

box filled with electrons, holes and

phonons moving about. They show

wave-particle duality and have

dispersion relations for the variation

of their energy £ with wave number k.

If the crystal is a metal, it also has a

Fermi surface in addition to these

features.

AA

HI 1

AA1

o
AA ^

AA <

AA

AA r

w f\f\

AA t

AA |

AA

A/

AA

AA

1 f\f\
o

AA

AA

AA

electron AA phonon

Ohole

phonons

electrons, holes

Fermi surface

(only for metal)

130

SUMMARY

crystal) is related to the translational symmetry of the lattice. In

addition, the shape of the Fermi surface for a metal contains
implicitly the symmetry of the corresponding crystal lattice. Thus the
quantum mechanical view of the crystal not only gives important
information about the electrons and about the effect of temperature
on the crystal, but also contains in it the crystal’s symmetry.

question In the view of quantum mechanics, how does the crystal of

germanium differ from that of silver?

answer Both crystals have electrons and phonons. The difference

will be in their dispersion curves, and in the distribution of electrons
among their quantum states. This difference explains why their
properties are not the same.
9 Summary
Section 1 outlines the task undertaken by us in this chapter: to arrive
at a quantum mechanical picture of a crystal in the same sense in
which we got such a picture of an isolated atom in chapter IV.

We imagine in section 2 a process of assembling a crystal starting

with isolated atoms that are far apart from one another, and
gradually bringing them close to one another. When the atoms are
far apart, the energy levels for the electrons are the same as for the
isolated atom, except that each level is highly degenerate, because it
can accommodate one electron from each of the 1021 or so atoms
that will eventually make up the crystal. In the crystal, on the other
hand, this degeneracy is removed, and each atomic energy level
becomes a band of closely spaced levels. Each electron now has a
slightly different energy, the reason being the influence of electrons
and nuclei from neighbouring atoms. There are gaps between
adjacent bands, meaning that there are no states for electrons with
energy values lying inside these gaps. These effects are most
pronounced for the outermost electrons of the atom because they
are closest to, and therefore feel most strongly the influence of,
neighbouring atoms in the crystal.

The electrons in the crystal occupy these band states beginning at

the lowest energy and progressing upwards until all the electrons in
the crystal are housed. If at this stage the highest occupied band is
exactly full and the next one is separated by an energy gap and
empty, we have an insulating crystal. If on the other hand the highest
band is only partially occupied, we have a metal. This is explained in
section 3, which then goes on to describe some examples of the
probability distribution of electrons in actual crystals, which
complements the description in terms of energy bands.

The energy states of the electrons have corresponding wave

functions, just as with particles in a box and with isolated atoms. We
need not worry about how these waves look. Considering that the
crystal is a more complex object than the other two, we would expect
the wave functions to be more complicated, and this indeed is so.
Each wave will still have its wave vector k, and represent a state of
energy E. We can make a picture for each band of how E varies with
k from the bottom of the band to the top.

VI THE QUANTUM MECHANICAL CRYSTAL

We use such pictures, called dispersion curves, in section 4 to see

again how metals and insulators differ from each other. Using the
wave vector of the electron and the fact that electrons are fermions,
we are led to the picture of a Fermi surface for metals.

In section 5,1 introduce the semiconductor as a special form of

insulator. We use the pictures of dispersion curves to see how
temperature and impurities produce semiconducting behaviour in an
insulator. The name itself is suggestive: the ability to carry an electric
current for a semiconductor is somewhere between that for a metal
(very good) and an insulator (very bad). We find that an otherwise
filled band that has a quantum state with no electron in it behaves
like a particle, named a hole. Its characteristics like energy, wave
vector, charge and spin are the opposite of the missing electron: an
anti-electron, so to speak.

I take a first look at what temperature does to the crystal in section 6.

Temperature is related to the energy of motion of the electrons, holes
and atoms. Raising the temperature increases this energy. The
energy of motion is determined by the mass of the moving object,
and not its electric charge. In a crystal we have for each atom a few
of the outermost electrons that may be wandering around in the
crystal. The ions that are left cannot get away from their lattice
positions, but can only vibrate with increasing amplitude (which
means increasing energy of motion) as the temperature of the crystal
is increased. This vibrational energy depends on the mass and not
on the electric charge of the vibrating ion. The mass of the ion is
practically the same as that of the atom. I can therefore talk about
vibrating atoms, even though they may actually be ions. The total of
the vibrational energy of all the atoms, plus the additional energy
(usually small in comparison to the other) which the electrons and
holes pick up due to temperature, is the thermal energy of the
crystal.

I give a quantum description in section 7 of the thermal vibrations of

the atoms in the crystal. I find that these vibrations are equivalent to
sound waves in the crystal, and their energies are quantized
following the rules of quantum mechanics. These quanta of energy
are called phonons, and they follow the statistics of bosons. They,
like electrons, have dispersion curves that show how their energy
varies with wave vector.

Finally in section 8,1 bring together the different elements of the

quantum picture of a crystal that have been developed in this
chapter: electrons and holes and phonons, dispersion curves of
energy versus wave vector (or wave number), Fermi surface for
metals, energy gaps for semiconductors.

We have thus arrived at a quantum-mechanical description of a solid

based on its crystalline atomic structure. We are ready now to
understand specific properties of solids in terms of this description.
We start with a group which gives the word ‘solid’

its meaning as defined by the Concise Oxford Dictonary: of stable

shape, having some rigidity. A piece of solid has some shape, and it
resists - at least up to a point - attempts to change it. These
properties are the easiest to observe experimentally: they need
nothing more than the experimenter’s eyes and hands, and not the
complicated equipment needed with the other properties we consider
later. These mechanical properties, as they are called, are the
subject of chapter VII.

132
VII COPPER WIRES AND GLASS RODS

l Elasticity, plasticity and brittleness

We have looked in the preceding chapters at the atomic structure of

solids, described in terms of quantum mechanics. We shall now see
how this description explains why solids have the various properties
that they exhibit. We begin by looking in this chapter at a group of
properties that defines what a solid is as distinguished from a liquid
or a gas. The atoms in a solid stay where they are, and do not slide
past one another as in a liquid, or fly about in the enclosing volume
as in a gas. A solid holds its size and shape, and resists attempts to
change them. The ideal solid is a perfect crystal, whose atoms are
bound to one another by the electrical forces among their electrons
and nuclei. It is as if there were springs connecting each atom to
neighbouring atoms. Any stretching or compression produced by an
applied force is taken up by the springs, and the solid returns to its
original state when the force is removed. The response of a real
solid to such a force, however, is more complicated, and I illustrate it
with the following experiments.

I take a copper wire and a glass rod, both of the same size. If I grasp
the ends of the copper wire and bend it just a little and let go, it
springs back to its original shape. If I bend it more than a certain
amount and then let go, it stays bent. The same experiment with the
glass rod ends differently. The rod will bend a bit initially and recover
when I let go, but the attempt to bend it further will cause it to snap
into two or more pieces. I find yet another result if I try the
experiment with a wire of the kind of steel used to make springs. I
can bend this wire a fair amount, but it will spring back to its original
shape when I let go. I go back to the permanently bent copper wire,
straighten it out and again bend it and so on, repeating the cycle a
number of times. I find that it gets harder to change the shape of the
wire with each successive cycle: the copper wire begins to resemble
more and more the steel spring. Finally, I take the springy copper
wire from this last experiment and heat it up to 500 °C, say, and hold
it there for some time. This kind of heat treatment is called
annealing. I test the wire after it has cooled down, and find that it has
lost its springiness.

All these materials are made up of atoms, and they are held together
by electric forces derived from the charges of the electrons and
nuclei. Nevertheless, these experiments show that they respond in
different ways to a mechanical force that tries to change their shape.
These different types of behaviour can be understood when one
looks closely at how the atoms are arranged in real solids, as distinct
from the idealised perfect lattices that we have assumed so far.
Before I get to that, I summarize the behaviour of a solid when
subjected to forces as mentioned above.

The experiments all have something in common: I apply a

mechanical force to a 133

VII COPPER WIRES AND GLASS RODS

sample, and note the resulting change in its shape. For a small
enough force, the change is proportional to the force, and is
reversible, i.e. the change is always the same for a given strength of
force. In this regime the material is said to show elasticity. This
prevails up to some maximum force, beyond which the deformation
of the material does not disappear if the force is taken away. At this
point the material has reached its elastic limit, and its size and shape
will have changed by a few parts in ten thousand. Beyond the elastic
limit, the material will either break (brittkness), or remain
permanently deformed (plastic deformation). I can summarize the
task now as an attempt to understand the elastic, plastic and brittle
behaviour of matter in terms of its atomic structure.
2 The springs
The elastic behaviour of solids is as though there were springs
connecting pairs of neighbouring atoms. The electric force between
the electrons and the nuclei in a solid is the only mechanism we
have to produce the effect of these springs. We saw in chapter VI
three different ways that the outer electrons of an atom arranged
themselves in solids. They were the following:

1. As free conduction electrons in a metal like silver.

2. An electron leaves each A atom and joins a B atom in an AB

compound, thus creating a positive A ion and a negative B ion. The
resulting solid is an ionic crystal, and sodium chloride (common salt)
is an example.

3. Two electrons, one from each atom of a neighbouring pair, come

together to bind the atoms together, as in silicon.

I take first an example of a metal, which I assume for simplicity to

have one free electron per atom. This electron is a Bloch wave which
is spread throughout the whole solid. However, the solid is
electrically neutral, and therefore on the average there is one
electron in the space around each ion. I show this space bounded by
broken lines around four of the ions in the lattice in fig. VII-1, part (a).
The polygons formed by the broken lines are all identical, because of
the translational symmetry of the lattice. We therefore need to find
the energy of just one ion and its associated electron; the same
result will hold for all other atoms in the crystal. The ion and electron
form a kind of atom, but the electron is free to move from one ion to
the next. The physicists Eugene Wigner and Frederick Seitz worked
out the quantum mechanics of such an atom. They found that the
electron stayed closer to the nucleus than it did in the isolated atom,
and therefore had a lower energy. This means that the metal is
stable, and does not disintegrate into a pile of free atoms. The
difference between this energy and the higher energy of the isolated
atom has to be supplied from outside if one is to tear an atom away
from the solid.

question Why does the electron have a lower energy when it stays
closer to the nucleus?

134

THE SPRINGS

figure VII-1. Part (a): Two-dimensional lattice of atoms in a metal

with one free electron per atom. An

identical space around each atom, like

the four that are shown, is imagined

to be enclosed by broken lines. One

free electron is to be found in each of

these spaces, which are called

Wigner-Seitz cells. Part (b): The bond

between two neighbouring silicon

atoms in the crystal. An electron

leaves each atom and is mostly to be

found in the space between the two

ions which are left behind.

(a)

(b)

+e

-2e

+e

answer The potential energy of two charges is equal to their product

divided by their distance apart (cf. p. 65). Because the electron and
the nucleus are oppositely charged, the energy is negative, and
becomes more negative, i.e. lower, when the electron is closer to the
nucleus.

In an ionic crystal like sodium chloride, the electric force of attraction

between the positive chlorine ions and the negative sodium ions
holds the crystal together. This effect is illustrated in figs. VI-7 and
VI-8. Common salt dissolves easily in water because of the electrical
properties of the latter. They cause the attraction between the
sodium and chlorine ions to decrease by a factor of 80, and make it
too weak to hold them together.

In silicon, whose crystal structure is shown in fig. VI-9, there is a

bond between each pair of ions consisting of one electron from each
atom. This is illustrated in fig. VII-1, part (b). The charge distribution
of+eon each of the ions and —2ein the electron cloud VII COPPER
WIRES AND GLASS RODS

between the ions causes electric forces which are as if a spring

connected the two ions.

If the normal interatomic distance is changed by applying a force, the

charge distribution changes in such a way as to oppose this force.

The three kinds of bonding between atoms in solids which I have

described have these things in common: each is based on the
quantum mechanical description of the atom, is derived from the
force between electric charges, and involves a balance between
attractive and repulsive forces between neighbouring atoms when
they are at the observed lattice spacing apart.

question Can you clarify the third point above?

answer Since there are both positive and negative charges in the
atoms, there must be both attractive and repulsive forces at play.
Each atom is sitting still in the crystal, apart from its zero-point and
thermal vibrations.

There is therefore no net force on it. This means that the attractive
and repulsive forces must cancel each other out. Stretching the
crystal changes the distribution of charges in such a way that the
attractive force dominates, whereas compressing the crystal causes
the repulsive force to be the larger.
3 Dislocations
I start with a perfect crystal of some element. I take it to be two-
dimensional, so that I can draw pictures that are easier to follow than
if it were three-dimensional. I show in part (a) of fig. VII-2 such a
crystal. The springs between neighbouring atoms in the FIGURE VII-
2. A crystal is depicted by

rows of atoms interconnected by

springs which stand for the forces

which hold the atoms together. Part

(a) represents the crystal with no force

applied. In part (b), stretching forces

(shown by the two sets of arrows) are

applied to the ends of the crystal,

causing the springs in the direction of

the forces to stretch and those at right

angles to shrink, producing a

corresponding deformation of the

crystal. When the forces are removed,

the crystal returns to the state shown

in part (a) if the elongation was

within the elastic limit.

(a)

(b)

136

DISLOCATIONS

horizontal and vertical directions look different, to suggest that the

crystal is anisotropic. In part (b) the crystal is stretched in the
horizontal direction as indicated by the arrows. The horizontal
springs are then elongated, the other springs are compressed, and
the crystal is correspondingly deformed. When the force is removed,
the springs as well as the crystal return to the state shown in part (a).
This is the elastic response of the crystal. The same stretching force
acting in the vertical direction would produce a different elongation of
the crystal: the crystal shows anisotropy in its elastic properties,
because the springs in the two directions are different.

We now come to the plastic region, where a strong enough force

produces a permanent deformation of the solid. The underlying
rearrangement of atoms is pretty complicated, and so I consider a
very simple deformation, illustrated in fig. VII-3. I FIGURE VII-3.
Atoms near the edge

(shown as a vertical line on the left) of

a two-dimensional crystal. The part

labelled (a) is before, and (b) is after, a

small deformation. To go from (a) to

(b), all the atoms below the horizontal

broken line must be moved as a unit

to the right by one lattice spacing.

(a)

00

o
o

o
o

(b)

ooooooo

ooooooo

oo o”o 6 6

oooooo

oo^ooo

OO-OGO

J37

VII COPPER WIRES AND GLASS RODS

show in part (a) the atoms in a small portion of a perfect two-

dimensional crystal near its edge. Now suppose I want to deform it,
so that it ends up looking like part (b) of the figure. One can think of
a large deformation as made up of a large number of such small
deformations. To produce one such deformation, I would have to
move all the atoms below the horizontal broken line simultaneously
one lattice spacing to the right. This would require the simultaneous
breaking of all the bonds between atoms on either side of the broken
line in part (a), and then reuniting the two pieces as in part (b). The
initial breaking of bonds requires a very large amount of energy.

question How do I know that this operation would require a large

amount of energy?

FIGURE vn-4. Part (a) shows a crystal

with a dislocation in the centre.

The atoms in its immediate

neighbourhood are only slightly

displaced from where they would be

in the perfect crystal, and the atoms

further away are practically unaffected.

The broken line marks the line of

atoms which ends at the dislocation.

In part (b), the atoms marked darker

grey have moved slightly to the right,

and the dislocation has moved to the

left.

(50

./

n
 \

v.-‘ ^ ; Kj \

;i

) ‘■ )

> r\ r-\

) i. j (a) {J ( j ; ; ( j

CD

/■■ “\ .— “n

l :■•

.’■’”- ■””- ”’-■

i■;I

138

DISLOCATIONS

answer Think of trying to pull apart a piece of solid into two pieces;
this also requires a similar breaking of the bonds between all the
atoms on either side of the fracture.

So a perfect crystal cannot be easily deformed in the manner shown

above.
However, real crystals are not perfect crystals. As they are being
formed, not every atom settles down in its legal lattice site. Mistakes
are bound to occur, and I show one such in part (a) of fig. VII-4.
There is a small empty region in the middle, whose volume is about
what would normally be occupied by one atom, surrounded by atoms
that are only slightly displaced from the exact lattice positions. A very
small force, enough to break just one bond, the one between the
atoms marked A and B, will cause the row of FIGURE VII-5. The
same crystal as in

fig. VII-3, but now the dislocation has

moved one more step to the left in

part (a), and finally disappeared at the

surface in part (b), leaving behind a

deformed crystal. The same atoms are

marked darker grey as are in fig. VII-4.

o o ooooo

:■ * o o o o

o©ooo

o•ooo

oo•ooo

o o•oo o

(b)

ooooooo

oosoooo
oo©ooo

jOOOOO

OGOOOO

oo oooo

139

FRACTURE

FIGURE vil-7. An alloy atom C

which has trapped a dislocation. A

movement of the dislocation like that

in fig. VII-3 now requires a

movement of the atom C also, and

this needs considerably more force

than when

momo©o©

o #C0 o ©

©o•ooo

to a high enough temperature, the increased amplitude of thermal

vibrations of the atoms makes it possible for dislocations to break
loose from the tangled network and move about. Those that reach
the surface of the wire will disappear. The net result is that the work-
hardening disappears, and the wire can again be bent easily.
A pure metal can be hardened by another method also, namely by
alloying. Pure copper is soft, but an alloy of copper with some zinc is
hard enough to be made into pots and pans: it is called brass.
Similarly, pure iron deforms too easily to make it suitable as a
construction material, but alloying it with a small amount of carbon
produces a kind of steel that is widely used in the construction of
buildings, bridges and so on.

The basic reason for such hardening is that the atoms of the alloying
element along with some of the atoms of the host element take up
such positions that they act as traps for dislocations and inhibit them
from moving, as illustrated in fig. VII-7. This effect of alloying is
known as solution hardening or precipitate hardening, depending on
whether the alloying atoms are distributed uniformly or lie in small
clusters. By a judicious combination of the hardening methods, one
can produce for example the steels that stand large amounts of
elastic deformation and are used to make springs.
4 Fracture
There are materials like glass, which deform elastically a little bit, but
suffer fracture upon further deformation. We know that the atoms in
glass are not in regularly ordered positions as in a crystal. It is
therefore not possible to imagine in glass a dislocation such as is
illustrated for a crystal in fig. VII-4. It would seem that the only way
for glass to deform 141

VII COPPER WIRES AND GLASS RODS

plastically is for large numbers of interatomic bonds all to be broken

at the same time and then reformed soon after: a highly unlikely
process that also requires a large amount of energy. How then does
a glass rod fracture when bent beyond a certain point?

I start by reminding you of an experience which most of you have

had: trying to tear open the plastic sheet in which things like audio-
cassettes, new books, nuts, and innumerable other things are
packed these days. One cannot tear the plastic sheet at any arbitrary
point along its smooth edge. Rather, one must first look for the small
nick in the edge that has been helpfully put there by the
manufacturer, and start there. I illustrate the process in fig. VII-8 and
explain it in the caption.

(a)

FIGURE VII-8. The tearing of a plastic

sheet. The magnification is supposed

to be such that the molecules show up

as wiggles on the sheet, (d) The force,

applied to the edge of the sheet at the

places shown by the arrows, is

divided among many molecular

bonds in parallel, and the resulting

force on each bond is too small to

break it. (b) If the edge of the sheet

has a sharp notch, the force is

concentrated at the tip of the notch

and acts on only a few bonds there.

The force on each bond there can

then become large enough to break it,

causing the notch to become deeper,

and the process continues.

142

SUMMARY

A similar mechanism is at work when a glass rod is fractured by

bending it.

Although its surface may look smooth to the eye, there are always
present some microscopically small notches on the surface. When
the rod is bent, the concentration of force at one of these notches
reaches a level at which it begins to grow, and the rod fractures
there.

question How can I get a glass rod to break at a particular point

along its length by bending it?

answer By making a scratch at that point with a glass-cutting knife,

holding the rod on both sides of the scratch and then bending the
rod.

The force is then mostly concentrated at this scratch, and the rod will
break there. This method is used by glassblowers to break glass
rods and tubes to a desired length.
5 Summary
A given solid responds to a mechanical force applied to it in different
ways: it deforms slightly and reversibly, or permanently, or it breaks. I
describe these different behaviours, called elastic, plastic, or brittle,
in section 1.

In section 2 I show how the springy binding between atoms occurs in

metals like silver, ionic crystals like salt, and semiconductors like
silicon.

I introduce in section 3 the idea of a dislocation, which is a small

deviation from the perfect arrangement of the atoms in a crystal.
Dislocations are formed naturally during the freezing of the solid from
the melt, and are also created when the solid is being deformed. I
show how the motion of dislocations leads to plastic deformation.
Such motion is inhibited by alloying: the alloying atoms trap the
dislocations and prevent them from moving.

I consider the brittle behaviour of glass in section 4. It is caused by

the presence of very small scratches on the surface. The applied
force becomes concentrated at these scratches, and finally becomes
strong enough at one of them to break the molecular bonds locally,
leading to brittle fracture.

I have used pictures of a two-dimensional solid to illustrate several

points in this chapter. The same pictures, suitably modified, hold for
three-dimensional solids too.

For example, a dislocation in two dimensions causes a departure

from perfect order over an area surrounding one atomic site. A
dislocation in a three-dimensional solid causes a similar departure
over a volume surrounding a line of atomic sites.

Dislocations are a departure from perfect atomic order in a crystal,

and are essential for an understanding of the mechanical properties
of solids. However, the vast majority of atoms even in a deformed
piece of matter are still in their ordered crystalline positions.
Mechanical deformation involves bodily movement of atoms from
one position to another, but there is no such movement in the case
of the other properties that we 143

VII COPPER WIRES AND GLASS RODS

shall be considering. These properties arise from two features: the

perfect crystalline arrangement of the atoms and the resulting
electronic band structures and phonons.

The dislocations cause only a small perturbation of these features,

and do not affect the essential aspects of such properties, like the
one that we consider next.

Temperature is a very pervasive thing in everyday life. The sense of

something being warm or cold is acquired very early in life, certainly
before understanding something about electric charge, say. So it
seems reasonable to take up some thermal properties first, before
going on to things electrical.

144

VIII

SILVER SPOONS AND PLASTIC SPOONS

l Some preliminaries

When I stir hot coffee with a silver spoon, I find that the handle
becomes warm very quickly. If on the other hand I use a plastic
spoon, its handle stays cool. I conclude that silver conducts heat
better than plastic. I find from experience that I can divide materials
into two classes: those that conduct heat well, and others that do
not. We know that metals like silver, copper or aluminium are good
conductors of heat and also of electric currents, while materials like
glass or plastics are poor conductors of both. So we may expect that
there is something common to the ways in which materials conduct
electricity and heat, and we shall see later that this is the case.

In order to understand clearly what we mean by the conduction of

heat, consider a simple example illustrated in fig. VIII-1. A bar of
some material, silver or glass for example, connects two chambers
of which one is hot and the other is cold. If I measure the
temperature of the bar along its length, I find that it is hot at one end
and becomes gradually cooler as I go along its length. What I have
here is an idealised version of the spoon in the cup of coffee. Energy
in the form of heat is being continually conducted from the hot
chamber along the bar to the cold chamber.

question How do we know that energy is actually moving in this

manner?

answer If I were to replace the cold chamber with a block of ice, I

would find that the ice gradually melts because of the thermal energy
it is receiving from the bar.

FIGURE VIII-1. Thermal energy is

conducted in the bar of material from

the hot end to the cold end.

hot

(low of heat —^

bar of material

cold

145
VIII SILVER SPOONS AND PLASTIC SPOONS

I now do another experiment, where I connect two blocks of ice with

the bar after first cooling it to the temperature of ice. I conclude in
this case that no heat flows from one to the other through the bar,
because I see that the blocks do not melt. I have assumed in all this
that no heat flows between the outside and the blocks or the bar.

These two experiments lead to an important conclusion: Thermal

energy flows along a bar only when its two ends are at different
temperatures. It is the difference of temperature which drives the
flow of heat along the bar.

Now I do a third experiment. I take a metal spoon, and while holding

it at one end between my fingers, I place the other end in a cup of
hot tea. That end of the spoon immediately gets hot, and the heat
travels slowly down the spoon, as I can tell by touching it at various
points at successive times. It takes seconds, perhaps even minutes,
for the changes of temperature down the spoon to become evident. I
can make a pic-FIGURE VIII-2. The temperature

along the length of a metal spoon at

four successive instants of time,

labelled 1, 2, 3, and 4 respectively,

after one end has been dipped in hot

tea. Curve 1 shows the situation at

that instant: that end is hot, but the

rest of the spoon has not yet received

the message, so to speak. Curves 2, 3,

and 4, at successively later times,

show how heat travels slowly down

the spoon.

1’ ■■

\\

\ ‘■

\\

\■

:\

\\

\
 \

\ \ 4\

\\

\\

\%

\\

distance along spoon

146

SPECIFIC HEAT

ture, as in fig. VIII-2, of how the temperature along the spoon varies
at different instants of time. The curves labelled 1 to 4 show the
temperatures at the initial moment of heating and at three later
successive times respectively. The sharp rise of temperature which
occurred at one end travels slowly down the spoon.

I summarize what we have learnt from these three experiments

about the flow of heat down a bar of material: a difference of
temperature is necessary in order to produce heat flow, the flow is
from the hot end to the cool end, and is rather slow. We would like
now to explain these features using the picture of a solid that we
developed in chapter VI. To do so, we need two new ideas: a
measure of how much thermal energy must be fed into the material
in order to change its temperature, and a way of taking account of
the collisions which the electrons and phonons suffer as they go
whizzing about in the solid. I explain these ideas in the next two
sections.
2 Specific heat
When I supply a certain amount of energy in the form of heat to a
piece of material, it becomes hotter. If I repeat the experiment with
different pieces, all the same mass but made of different materials,
then I find that the rise in temperature is different for each material,
even though the amount of thermal energy was the same. To put it
conversely, it takes different amounts of thermal energy to produce
the same rise in temperature in different materials. I illustrate this in
table VIII-1 showing the amounts of thermal energy (in joules)
needed to heat one gram of different materials from 20 °C

to 21 °C. This quantity is called the specific heat per gram of the
material. Since we know from other experiments the number of
atoms in one gram of each material, we can also calculate the
specific heat per atom, and this is shown in the second column of the
table. The table shows that the specific heat per gram is different for
different materials, whereas the specific heat per atom has about the
same value for the first three, and is lower for diamond. I shall return
to this point presently, but first we look at what happens to the
electrons and phonons in a solid when some thermal energy is
added to it.

Table VIII-1. Specific heats at room temperature of some materials

material

copper

silver

lead

diamond

water
specific heat per gram

(joule per °C)

0.39

0.24

0.13

0.51

4.2

specific heat per atom

(10-23 joule per °C)

4.1

4.3

4.48

1.05

12.6 (per H2O molecule)

147

VIII SILVER SPOONS AND PLASTIC SPOONS

I start with the thermal energy that is in the phonons, representing

the vibrations of the ions (or atoms) in the crystal. I assume that the
crystal is at 20 °C (which is the same as 293 kelvin). I refer to this as
room temperature, and denote it by the symbol T. You recall that
there is a maximum phonon frequency^) (p. 126). If this frequency is
such that the energy k^Tis much larger than the phonon energy hfo,
then we find from the statistical physics of phonons that the average
thermal energy per atom is just 3k%T. The symbol kB is the
Boltzmann constant, a constant of nature with the value 1.38 X
10~23 joule per degree. At a temperature which is one degree
higher, the thermal energy per atom is 3k^(T+ 1). So the extra
energy per atom needed to raise the temperature by one degree,
which we have called the specific heat per atom, is the difference
between these two energies, or just 3£B, which is 4.14 X 10~23
joule per degree.

Thus the specific heat per atom should have this value for all
materials. The table above shows that this is approximately true for
copper, silver and lead, but not for diamond.

The reason for this deviation for diamond is that itsj^, is much larger
than thej^ for the other three materials, and one must go to much
higher temperatures to get kBTto be greater than hfo. When this
condition is not met, the statistical physics of phonons tells us that
the specific heat is smaller, and in fact heads towards zero as the
temperature approaches the absolute zero. All these predictions are
borne out by experiments.

Looking at the table again, we see that the specific heat per atom for
the three metals is in fact slightly larger than the value of 3kB which
comes from the phonons. This excess is partly due to the presence
of conduction electrons in the metals, which can receive thermal
energy and move to neighbouring quantum states that were
previously unoccupied, and thus add to the specific heat. We now
estimate this electronic contribution.

question The electronic contribution to the specific heat is absent or

at best negligible in insulators. Why?

answer The electrons in insulators have no unoccupied quantum

states near enough in energy to which they could move by absorbing
thermal energy.
Let us consider a metal that has N atoms per cc. Their thermal
vibrations, as we saw, give a phonon specific heat which I shall call
CL:

CL = 3A*b.

The subscript L reminds me that this comes from the /attice. Now
suppose that each atom sets free one electron; these electrons then
form a Fermi surface. At OK, all quantum states are occupied by
electrons up to a maximum energy called the Fermi energy £F, which
is about a thousand times the thermal energy per atom at room
temperature. At a temperature T, the additional energy which an
electron can take on as thermal energy is about k^T, which is a very
small fraction of the Fermi energy. As shown in fig. VI-18 (p. 124),
this means that only those electrons whose energies are less than
the Fermi energy by about ks T can move to empty states of higher
energy leaving empty states, equivalent to holes, behind. The
number of such electrons and 148

MEAN FREE PATH

holes is therefore a fraction (kBT/E^) of the total number N. Each of

these has on the average an energy kBT, and so their total thermal
energy is equal to their number, N(kBT/EF), multiplied by kBT, or

M:B2T2/£F.

If now I warm the crystal by one degree from Tto T+ 1, the energy of
the electrons and holes increases to

NkB2(T+l)2/EF.

The difference between these two energies is the energy per cc that
must be fed to the electrons and holes in the crystal in order to raise
the temperature by one degree, and is therefore the electronic
specific heat, which I denote by the symbol Cg (the subscript E
because this comes from the electrons and the associated holes):
CE = NkB2(T+ 1)2/£F minus NkB2T2/EF,

which is very nearly the same as

I have used a little bit of algebra to get to the last step, and it is easy
to verify it by putting in some numbers. We took Tto be 293 degrees
and so we must subtract 293

multiplied by itself from 294 multiplied by itself; the result is 587,

which is very close to twice 293.

So we now know how much energy must be given to the phonons,

and how much to the electrons and holes, in order to raise the
temperature of the crystal by one degree.

The specific heat of the crystal as a whole is got by adding these

two. We can find out their relative magnitudes by forming the fraction
C^/C^-.

CE/CL = 2NkB2 T/EF divided by 3N&B

which is equal to 2kBTdivided by 3£F. Since the Fermi energy is a

thousand times the thermal energy, this means that the electronic
specific heat is about a thousandth of that due to the phonons. We
conclude from this that at room temperature the electronic specific
heat is a very small fraction of the total specific heat. Most of the
energy needed to warm the crystal goes to the phonons, and only a
very small part goes to the conduction electrons, in the case of a
metal. In an insulator or semiconductor, there are no free electrons
(or they are at best negligible in number), and practically all the
energy goes to the phonons.
3 Mean free path
I go now back to the bar of material illustrated in fig. VIII-1. Energy in
the form of heat is flowing steadily from the hot end to the cold. I
want to describe this in terms of the electrons, holes and phonons.
They are the entities which contain the thermal energy 149

VIII SILVER SPOONS AND PLASTIC SPOONS

hot end cold end

FIGURE VIH-3. Phonons in a bar along which heat is flowing.

Although phonons of all frequencies are present, only those of a
single frequency are shown. Each arrow represents a phonon
travelling in the direction of the arrow. The density of phonons, i.e.
the number in a given volume, decreases gradually from the hot end
to the cold end. This is what eventually leads to heat flow from hot to
cold.

and therefore they must be responsible for its transport down the
bar.

Consider a bar of insulator first, in which there are only phonons to

carry the thermal energy, for there are no free electrons in it. We
shall see later that what we find out about the phonons will apply to
the electrons and holes in a metal also, with suitable modifications.
Figure VIII-3 shows the bar. The arrows represent phonons of one of
the many frequencies present, and remind us that they are speeding
about in all directions with the velocity of sound in the material. The
density, i.e. the number of phonons per unit volume, gradually
decreases as one proceeds from the hot end to the cold end. Exactly
the same holds for phonons of all frequencies in the bar, and so the
results we get for the phonons of a single frequency will apply to all
phonons.
question Why is the density of phonons higher at higher
temperatures?

answer The higher the temperature, the more thermal energy there
is in a given section of the bar. This energy is contained in the
phonons, and so there must be more phonons in this section. You
can see this in fig. VI-21

(p. 128).

Consider a phonon towards the hot end of the bar in the region
marked A. It will travel a certain distance before it collides with
another phonon. Suppose that this collision takes place in the region
marked B. The collision changes the phonons to other phonons
whose total energy is equal to that of the colliding phonons because
the total energy is conserved. Now there is more energy, and so
there are more phonons, in the region of the collision than is
prescribed by the temperature there. These surplus phonons
continue on until the same thing happens to them further down the
bar. There is at the same time a flow of thermal energy in the
opposite direction at B, due to phonons arriving there from a region
like C which lies towards the cooler end of the bar. The net flow of
energy is then the difference between these two.

The distance travelled by a phonon before it suffers a collision is not

the same for all phonons. To simplify matters, we assume that there
is an average distance, and that all phonons travel this distance
between one collision and the next. I call this distance the mean free
path, and denote it by . This means that on the average, a phonon
carries its energy a distance down the bar before it has a collision
and gives up the energy to 150

THERMAL CONDUCTIVITY OF INSULATORS

other phonons, and they in turn continue the process. Experiments

have been done to measure /, and I show some results in table Vlll-
2a.

Table VIII-2d. Mean free paths of phonons at room temperature

material mean free path in angstroms

quartz 100

silicon 430

common salt 70

The electrons and holes in a metal also suffer collisions, with

phonons and with alloying atoms in the case of alloys. They
therefore have a mean free path too, and some values are listed in
table VIII-2&.

Table VIII-26. Mean free paths of electrons and holes at room

temperature metal mean free path in angstroms

copper 850

silver 1140

iron 440

The distance between neighbouring atoms in a solid is a few

angstroms (one centimetre is equal to one hundred million
angstroms). So you can see that all these particles go a fair distance
compared to interatomic distances before they undergo collisions. If
they were particles in the old-fashioned sense, one might expect
their mean free path to be of the order of the interatomic distance. It
is the quantum behaviour of the particles which is responsible for the
long mean free paths.
4 Thermal conductivity of insulators
I now take a closer look at how thermal energy is carried down the
bar by phonons, and use fig. VIII-4 for this purpose. Let the heat flow
be from left to right, which means heat flow —^

FIGURE VIII-4. A, B, and C are

imaginary planes each of area Sand

spaced a mean free path /apart,

cutting the bar along which heat is

flowing. They are at temperatures T,,

T, and T2 respectively, with T, the

highest and T2 the lowest. “J” iT”

VIII SILVER SPOONS AND PLASTIC SPOONS

that the temperature decreases as I go in this direction along the bar.

Imagine three cross-sections A, B, and C of the bar which have the
same area of S square centimetres, at a distance apart, where is the
mean free path in centimetres. Suppose that the temperatures at A,
B, and C are Tlt T, and T2 respectively. Because of the direction of
the heat flow, we know that Tt is higher than T, and Tis higher than
T2. The phonons at A have an energy E(Tt) per cc, the thermal
energy at temperature 7. On the average, half of them are going to
the right and reach B before they suffer collisions. How much energy
per second flowing to the right do these phonons deliver at B? The
problem is similar to the following: I stand at the side of a road on
which pass cars, all at the same speed of 100 km per hour and so
spaced apart that there are 10 cars per km of the road.

How many cars do I see going past in one hour? The answer is 10
cars/km multiplied by 100 km/hour, or 1000 cars per hour. By a
similar reasoning, the thermal energy flowing to the right per second
at the surface B is given by 2 £(7^) times the area S times the speed
v of the phonons.

Note that the speed of the phonons is the same as the speed of
sound, which in most solids is a few thousand metres per second.
There are also phonons at B which are travelling to the left. These
phonons arrived there from C where they had a thermal energy
E(T2) per cubic cm, and they bring a thermal energy per second
flowing to the left equal to

2 E(T2) times the area S times the speed v of the phonons.

So the net thermal energy flowing to the right per second, which I
denote by Ql> is equal to the difference of the above two quantities:

2l = net heat flow per second across B = \[E(Ti) - E(T2)]Sv.

I now do a little bit of arithmetic to get the above formula for <2l into
a form which shows me how to relate Ql t0 the specific heat and the
mean free path, which are properties of the phonons in the material.
I multiply gL by two numbers which are each equal to one, so that
the result is stil

<2l = Ql times j*_j* times 2/

Now notice that QL contains in it the factor [E(Tt) - E(T2)] which is

just the difference in the thermal energies per cc between the
temperatures Tx and T2. This difference divided by the difference in
temperature is the specific heat of the material, namely the energy
needed to raise the temperature of one cc of the material by one
degree. I denote this as before by CL:
The temperature of the bar drops from Ti and T2 over a distance 21.
The quantity —’.. 2 (which I denote by R) is the rate of change, or
the gradient, of temperature 152

THERMAL CONDUCTIVITY OF INSULATORS

along the bar. I can then write the amount of thermal energy going
down the bar per second as

QL = vCLlSR.

I have assumed so far that all the phonons are travelling either to the
left or to the right along the bar. In fact, of course, they are travelling
in all directions. When I take this also into account, the formula for
Q^ is modified to <2L = jvClISR,

which I can write as

QL = KtSR.

I have used the symbol K^ to stand for \vC\l. The quantity K^ is

purely a property of the material; it depends on the specific heat CL
due to the phonons, their speed v, and their mean free path . The
subscript L reminds us that the heat is being carried by the attice
vibrations, i.e. the phonons. K^ is called the lattice thermal
conductivity of the material. The larger it is, the better is the
conduction of heat in the material.

This picture of how heat is carried down a bar by phonons allows us

to understand the features of heat conduction mentioned in section
1: heat flow down a bar requires a difference in temperature between
the two ends of the bar, and takes place rather slowly. The formula
given above for Ql shows that when R is zero, i.e. when there is no
temperature difference, the heat flow is also zero. The rather slow
speed with which heat travels down the spoon when one end is
dipped in hot coffee is because of the collisions which the phonons
suffer after travelling the distance of a mean free path which is very
small in comparison to the length of a spoon. If a rod which is a
single crystal is cooled to a temperature very near the absolute zero,
then the mean free path for phonons becomes very large, and a
pulse of heat fed in at one end travels to the other end at a speed
comparable to the speed of sound in the material. One says then
that the thermal energy is carried by second sound.

Now consider a bar of material which is in the shape of a cube, one

cm on the side, as illustrated in fig. VIII-5. By connecting the cube to
suitable warm plates and so on, one can arrange that the faces A
and B are held at temperatures of 21 °C

and 20 °C respectively. The bar has a cross-section of one square

cm, and the temperature gradient is one degree per cm. From the
formula for QL given above, we can see that the heat flowing per
second from the face A to the face B is just KL. Thus the thermal
conductivity as we have defined it above is the thermal energy
flowing per second between two opposite faces of a cube one cm on
the side, when the temperature difference between the faces is one
degree. I show in table VIII-3 this quantity in joules per second for
various materials in the form of such a cube.

VIII SILVER SPOONS AND PLASTIC SPOONS

FIGURE VIII-5. The cube of material

used to define its thermal

conductivity. Its edges are one

centimetre long, and the pair of

opposite faces A and B are at 21 °C

and 20 °C respectively.
 \

o
21 C
A
\

O
20 C
B
\

Table VIII-3. Flow of heat per second for the cube shown in fig. VIII-5

flow of heat in joules per second

material of cube

quartz

glass

silicon

common salt

silver

copper

stainless steel

brass

typical plastic

flow 0
0.07

0.01

1.5

0.064

4.2

4.0

0.14

0.8

0.1

A joule is the unit with which we measure amounts of energy. A joule

of thermal energy added to one millilitre (which is one cc) of water
raises its temperature by almost one quarter of a degree Celsius.

The numbers in table VIII-3 show some interesting regularities.

Metals like copper and brass conduct heat better than insulators like
glass and plastic, and a pure metal like copper is a better conductor
of heat than an alloy like brass, which is a mixture of copper and
zinc. The principal reason why metals conduct more heat than
crystalline THERMAL CONDUCTIVITY OF METALS

insulators under the same conditions is that, in addition to the

phonons which are common to both kinds of materials, the metals
have free electrons coming from the outermost shells of the atoms.
These electrons can move about the material and make an
additional contribution to the flow of thermal energy. We look at this
electronic part of the thermal conduction in metals in the next
section. We shall then be able to see why alloys do not conduct heat
so well as pure metals.
5 Thermal conductivity of metals
The thermal energy in insulators and semiconductors is almost all
contained in the phonons. In a metal on the other hand, the thermal
energy is contained not only in the phonons, but also in the electron-
hole pairs which appear near the Fermi surface because of
temperature (fig. VI-18, p. 124). The flow of heat in a metal is
therefore due to the phonons and also the electrons and holes.

The way that thermal energy is conducted by the electrons and holes
is exactly the same as the way by phonons. I get the same formula
for the thermal conductivity as in the case of the phonons, except
that I must use in it the corresponding quantities for the electrons
and holes: the specific heat CE, the Fermi velocity vp, and the
electron’s mean free path E. Calling this electronic thermal
conductivity CE, I get So the thermal energy carried per second by
the electrons and holes is 2E = KESR,

and the total thermal energy is Q, with

Q = Qe + Ql—

Notice that in each case the thermal conductivity is determined by

the product of the specific heat, the speed and the mean free path of
the particles in question.

We now estimate the relative magnitudes of the electron and the

phonon contributions to the conduction of heat. We consider the
mean free paths first. An electron or a hole collides with a phonon
long before it meets another electron or hole. The most likely
collisions for phonons also are with other phonons. The mean free
paths for all of them therefore tend to be about the same, as can be
seen in tables VTII-2<z and VIII-2&. Although the electronic specific
heat is smaller than the lattice specific heat, the Fermi velocity is
much larger than the speed of sound in the material. The result is
that the heat conduction due to electrons and holes is comparable to
or even larger than that due to phonons.

In an alloy, we have more than one kind of atom in the lattice: copper
and zinc atoms in brass, for example. This means that we have lost
the translational symmetry of a pure metal in which every lattice
point is occupied by the same kind of atom. This departure from
perfection acts as a disturbance to the particle (electron, hole or VIII
SILVER SPOONS AND PLASTIC SPOONS

phonon) in its quantum state. It responds to the disturbance by

moving to a different quantum state, which is just a different way of
saying that the particle is scattered. This additional scattering means
that the thermal conductivity is decreased by alloying.

You will notice from table VIII-3 (p. 154) that quartz is a much better
conductor of heat than glass, even though they both consist of
molecules of silicon dioxide. So the mean free path for phonons in
glass must be smaller than in quartz. The difference between the two
materials is that quartz is a crystal and the molecules are arranged in
it with translational symmetry, whereas in glass there is no such
symmetry. So we come to the general conclusion that loss of
translational symmetry, whether caused by alloying or by conversion
into a glass, causes the mean free paths of phonons and electrons
to decrease. As a result of this the thermal conductivity becomes
less.
6 Thermal expansion
When a metal cap on a glass jar has been screwed on so tightly that
it cannot be unscrewed by hand, an old kitchen trick is to heat the
cap, for example by running hot water over it. This works in some
cases, and the cap comes off easily. This could be because a bit of
whatever was in the jar had gone hard in the space between cap and
jar, and the heating simply softened it. But the method can also work
when everything is clean. The reason then is that the heating causes
the cap to expand in size so that it is no longer a tight fit on the jar.
Such a change in size of a material when its temperature is changed
is called its thermal expansion. In practically all situations that we are
likely to encounter, it is extremely small and would not be seen by
the unaided eye.

Table VIII-4 shows the thermal expansion of some common

materials.

Table VIII-4. Thermal expansion at room temperature of some

materials (The number in the second column gives the increase in
length of a rod of the material one cm in length when its temperature
is raised from0oCtol00cC)

material increase in length in cm

copper 17 x 10”4

iron 12 X 10”4

aluminium 24xlO~4

stainless steel lOxlO”4

glass 5xlO”4
You notice that different materials expand by different amounts for
the same rise in temperature, and that glass expands less than the
listed metals. One sees now why a stuck bottle cap can be freed by
heating it, even if the glass gets a bit warm.

156

THERMAL EXPANSION

FIGURE viii-6. The upper part shows

the expansion of a warmed balloon,

and the lower part that of a warmed

solid. The molecules of the gas (the

phonons, as also the free electrons if a

metal) inside the balloon (in the solid)

exert a pressure on the balloon wall

(on the surface of the solid). The

pressure goes up if the temperature

goes up, and the balloon (the solid)

expands a bit as a result.

cool

gas-filled balloon

warm
phonons

electrons

phonons

electrons

cool

warm

a solid

What causes a solid to expand when it is heated? I take a simple

example of something which shows similar behaviour, namely a
rubber balloon filled with gas, shown in the upper part of fig. VIII-6.
The pressure of the gas which is trying to expand the balloon is
exactly balanced by the elastic inward pressure of the balloon which
is trying to shrink it. It is this balance which determines how big the
balloon will be for a given pressure of the gas inside it. When I warm
up the balloon, the pressure of the gas increases, and this is
counteracted by the balloon expanding and therefore increasing its
inward pressure so as to continue to balance the gas pressure. The
amount of gas in the balloon remains the same through all this.

Now imagine, as shown in the lower part of fig. VIII-6, the balloon
replaced by the surface of the solid, and the gas molecules by the
phonons in the solid, and the free electrons also if the solid is a
metal. These phonons and electrons exert a pressure on the
surface, and the pressure increases with increasing temperature.
This results in the thermal expansion of the solid.

So far, so good: but how are we to understand the pressure in terms

of the molecules in a gas, or the phonons and electrons in a solid? I
take the case of a gas first. Figure VIII-7 shows a portion of the
balloon wall and the paths of some of the many molecules VIII
SILVER SPOONS AND PLASTIC SPOONS
FIGURE VIH-7. Molecules of a gas

bouncing off the containing wall

(of a balloon, for example) transfer

momentum to the wall. The amount

of momentum so transferred per

second is proportional to the pressure

on the wall.

paths of molecules^

bouncing off the wall \

a small portion

of balloon wall

that are hitting it and bouncing back all the time. As each molecule
bounces back, it causes a recoil of the wall directed to the right in the
figure. What is at work here is the conservation of momentum. The
momentum of the molecule changes as a result of the collision with
the wall. This change must be compensated by a change of
momentum of the wall, such that the total momentum remains
unaltered.

The cumulative effect of such recoils due to all the molecules hitting
the wall is that the gas pressure on that side of the wall tends to
move it to the right. At the same time, the wall suffers similar
collisions from molecules on its other side too, and these give it an
impulse which tends to move it to the left. The impulse, and therefore
the gas pressure, increases when the speeds of the molecules
increase, which is what happens when the temperature of the gas
rises. Thus we see why the balloon expands when the gas inside it
gets warmer.
question Where does the elasticity of the balloon come into all this?

answer The excess pressure of the warm gas inside the balloon over
that of the cool gas outside is balanced by an effective pressure of
the balloon acting inwards, resulting from the fact that it has
increased in size. This is a bit like a stretched rubber band trying to
return to its original size.

158

SOME PRACTICAL EXAMPLES

The free electrons in a metal, which are largely responsible for its
thermal conductivity, are moving about inside the metal, but cannot
get out through its surface. The attraction between the negatively
charged electrons and the positively charged ions in the metal holds
the electrons trapped inside the metal. The effect of this attraction is
as if there were an impermeable but elastic sheath all over the
surface of a piece of metal. The electrons are confined inside this
sheath, with which they are constantly colliding and recoiling, in
analogy with the gas molecules and the balloon wall. There is a
resulting pressure due to the electrons on the surface of the metal.
The average speed of the electrons increases when the temperature
rises, thus causing this pressure to increase and the metal to
expand. We call this the electronic contribution to the thermal
expansion.

Now how about the phonons? The electrons have mass, and they
are moving about at various speeds. It is easy to imagine that they
transfer impulse to the wall when they collide with it. The picture is a
little different with phonons. They represent the quanta of energy of
the vibrating atoms, and there is no obvious way in which we can
associate a mass with a phonon. A phonon carries no momentum in
the sense in which a molecule in a gas does. You can see this when
you recall that the phonon is the quantum aspect of a sound wave in
the solid. For each atom in the sound wave with a given speed in
some direction at any given instant, there is another atom with the
same speed but in the opposite direction, so that the total
momentum of the wave will be zero. We cannot talk about a phonon
colliding with the surface of the solid and transferring momentum to
it.

There is a different way of looking at things which gives meaning to

the idea of the pressure due to phonons. To begin with, the
frequency, and thus the energy, of a phonon decreases when the
volume of the crystal increases. If there is nothing to hold the crystal
together, this would mean that even when the temperature is
constant, the crystal would keep expanding because its energy due
to the phonons would then keep getting lower. One can think of this
as a pressure inside the crystal caused by phonons which tends to
expand it. The crystal of course does nothing of the sort, because
there are internal forces due to the charged nuclei and electrons
which hold the crystal together. If now the crystal is heated up a bit,
the distribution of phonons in the different energy levels changes,
and consequently the phonon pressure also changes. This results in
a change in volume of the crystal.

This way of relating thermal expansion to the change in quantum

energy levels produced by a change in volume is an example of how
quantum mechanics and statistical mechanics are combined. It can
also be used to get the contribution of the conduction electrons in a
metal to its thermal expansion.
7 Some practical examples
We have looked at three properties of materials which are related to
temperature: specific heat, thermal conductivity and thermal
expansion. I give some examples of everyday situations where these
properties come into play.

159

VIII SILVER SPOONS AND PLASTIC SPOONS

While eating a meal, you may have on occasion put a potato in your
mouth which was much hotter than you expected. You may then
have noticed how even a sip of water is enough to cool off the
potato. This is because the specific heat of water is much greater
than that of the potato. Even though the amount of water is small, it
undergoes only a modest rise in temperature as it absorbs the
thermal energy coming out of the potato as it cools. The specific heat
of water is one of the highest for any material we know, as is evident
in table VIII-1. This is why, for example, if a pan of water and a
garden tool are left outside in sunlight, the tool may become too hot
to touch whereas the water may be barely warm. It is this ability of
water to absorb large amounts of thermal energy with only modest
changes in its temperature which is an important factor in
determining the earth’s climate.

In solid materials, thermal energy is carried by the process of

conduction which I have described here: energy moves, matter stays
where it is. In liquids and gases, on the other hand, the process is
predominantly convection: matter itself moves, carrying energy with
it. There is still a third way in which thermal energy moves, namely
as radiation, and I shall say more about it in the next chapter. I
consider here a common situation in which thermal conduction plays
a central role: household heating in the winter months, or cooling in
the summer months.

Figure VIII-8 is a bare-bones impression of a house in winter. The

outside temperature is 5 °C and the inside is heated to hold a
temperature of 22 °C. Since the inside temperature is constant, there
can be no energy being absorbed there. Thus all of the energy being
put out by the heating system finds its way to the outside world by
conduction through the walls, roof, and other boundaries of the
house. Once the inside temperature has been brought up to its
steady level, all the energy supplied by the heating system is
warming up the world outside the house. This is of course the reason
for using very good thermal insulation between the inside and
outside of the house; one wants to reduce the thermal energy which
is being conducted to the outside. We saw earlier (p. 152) that the
rate of energy flow is proportional to the difference in temperature
which produces this flow.

question What is the saving in energy if the house temperature is set

at 18 °C instead of at 22 °C in the example above?

answer The energy flow from inside to outside is proportional to the

difference of temperature between the two. When the setting is
changed from 22 °C to 18 °C, this difference changes from 17
degrees to 13

degrees. Thus the heating power now needed will be 13/17, which is
72

per cent, of what was needed before.

Another example of the use of our knowledge about thermal

conductivity is found in the kitchen: stainless steel cooking pots with
copper bottoms. You can see from table VIII-3 (p. 154) that copper is
thirty times as good a conductor of heat as stainless steel.

The copper bottom spreads the heat from the stove uniformly and
therefore helps to prevent hot spots with their deleterious effect on
what is in the pot.

160

SUMMARY

FIGURE VIII-8. Impression of a

heated house in winter: 5 °C outside

and 22 °C inside. All the thermal

energy supplied by the heating

system, once the inside temperature

has stabilised, flows through the floor,

walls and roof of the house to heat up

the outside world.

5°C

outside

22°C

inside

The thermal expansion of materials between the highest and lowest

temperatures that we encounter around us is quite small. For
example, a rod of iron one metre long would increase in length by
about one millimetre between minus 50 °C and plus 50 °C.

Although this effect is small enough to be negligible for many

purposes, it has to be explicitly taken into account in the design of
many outdoor structures: railway tracks, roads, bridges, and
buildings, for example.
8 Summary
I apply in this chapter the quantum mechanical picture of a solid to
the thermal properties of a solid. One of these is the conduction of
heat: section 1 considers some examples and deduces some
general features of how thermal energy flows down a rod.

A difference of temperature must exist between the ends of the rod,

and the flow is from the hot end to the cool end and is rather slow.

161

VIII SILVER SPOONS AND PLASTIC SPOONS

The specific heat of a solid is one of its thermal properties which is

not only of intrinsic interest, but also important for the thermal
conduction. It is a measure of how much energy must be fed into a
solid in order to raise its temperature by a given amount. Section 2
describes the contribution to the specific heat from phonons (which
exists for all solids) and from free electrons (which applies to metals
only).

The carriers of the energy in thermal conduction are phonons,

electrons, and holes.

Their motion in the solid is not unhindered: they undergo collisions

which can change their energy and momentum. We can think about
the average distance such a particle travels between collisions, and
call it the mean free path. This is the subject of section 3.

The ideas from sections 2 and 3 are brought together in section 4 to

give a picture of how thermal energy is conducted down an insulator
by phonons. We see how the specific heat and the mean free path
are involved in thermal conduction.
Section 5 takes up thermal conduction in metals. Here the
contribution from free electrons is added to that from phonons. This
explains why in general metals conduct heat better than insulators.
Since the atomic arrangement in alloys is less ordered than in pure -
metals, the mean free paths for electrons as well as for phonons are
smaller in alloys. This is the principal reason why, for example,
stainless steel is a poor conductor of heat compared to copper.

Section 6 deals with thermal expansion, which is the change in size

of a piece of matter when its temperature changes. This arises from
the fact that the quantum states of the electrons and phonons are
modified when the volume of the piece changes.

Some examples from everyday experience are mentioned in section

7 to illustrate the thermal effects discussed in this chapter.

This brings us to the end of our consideration of heat and matter. In

the next chapter, I shall describe to you how light interacts with
matter. We go on to where there will be more light and less heat.

162

IX GLASS PANES AND ALUMINIUM FOILS

l Some preliminaries

A sheet of glass is transparent to light, whereas a much thinner

sheet of aluminium foil is opaque: materials can be divided into those
which let light go through, and others which block it. Examples of the
first kind are glass, clear plastics, crystals like sugar and common
salt. Metals on the other hand are opaque to light. We know that all
these materials are made up of atoms. We saw in chapter IV that the
atom itself is mostly empty space, because its nucleus and electrons
occupy a tiny fraction of the volume of the atom. We might think
therefore that light, which can travel freely through empty space,
should be able to go right through all materials, but this clearly is not
so. The transparency or opacity of materials results from what
happens to the quantum energy states of the electrons when the
atoms come together to form a solid.

Colour is another property of materials which must have something

to do with how light interacts with them. Metallic copper for example
is red and gold is yellow, in daylight. Sapphire is blue and ruby is red.
The explanation again lies in the electronic structure of such solids,
together with the quantum nature of light, namely that it consists of
photons with energy equal to the frequency multiplied by the Planck
constant.

The kind of wave motion, of which visible light is an example,

extends over a tremendous range of frequencies. Radio waves,
microwaves, and X-rays are the same kind of waves as light,
differing only in frequency. These waves collectively are known as
electromagnetic waves, for reasons which will become clear later.
We shall find out what an electromagnetic wave is, and then look at
some examples of how these waves interact with matter. I shall
consider here mainly visible light, because this is the kind of
electromagnetic wave that we see every day. The picture that I shall
develop applies to all the other waves too.

2 What is light?

In chapter VII, I described heat as a form of energy which a piece of

material has by virtue of its temperature. This thermal energy was
seen to be the energy of motion of the atoms and electrons in the
material. We cannot therefore imagine this kind of energy in the
absence of matter.

There is another example of energy, namely the energy we receive

from the sun in the form of light. Sunlight reaches us after travelling a
path from the sun which is mostly empty space, and can go through
materials like window glass but is stopped by 163
IV GLASS PANES AND ALUMINIUM FOILS

other materials like aluminium foil. We have seen in chapter III that
light has the property of duality: it behaves like waves showing
interference and diffraction, and like particles (which we call photons)
in the effect discovered by Arthur H. Compton when a beam of X-
rays is scattered by electrons. The momentum p and the energy £ of
a photon are related to the wave vector k and the frequency/of the
corresponding light wave by the rules of quantum mechanics:

p = bk

and

E=hf

where h is the Planck constant. The waves which correspond to

visible light have wavelengths between about 8 X 10~5 cm for red
light and 4 X 10~5 cm for violet light.

The other colours have wavelengths between these two values. The
frequency is given by the speed of light, 3 X 1010 cm/sec, divided by
the wavelength.

In a wave on the surface of water, it is the water itself which is

moving up and down.

In a sound wave travelling through air, the pressure at each point

oscillates up and down at the frequency of the sound wave. In both
these waves, there is a material, water or air, which carries the wave
and oscillates in a known manner. But what about a light wave? It
can travel through empty space, and therefore does not need a
material medium which does the oscillating. It goes through a
vacuum and also through certain, but not all, materials. In order to
see why glass is transparent to light, whereas aluminium foil is not,
we need to look more closely at exactly what it is that is oscillating in
a beam of light. Before doing that, I would like to tell you something
about the wavelength and frequency of these waves.
Visible light has a wavelength, depending on its colour, between 4 X
10‘5cm and 8 X 105 cm. There also are waves with wavelengths
outside this range, not seen by our eyes, but otherwise exactly like
light waves. Such waves exist in nature, or can be produced with
suitable instruments. Their wavelengths have been observed to
range from 1010 cm all the way down to 10~14 cm. There can also
exist such waves with wavelengths beyond these limits.They all
travel through vacuum with the speed of visible light, 3 X 1010cm per
second. These waves, all having the same basic character and
spanning this vast range of wavelengths and frequencies, are
collectively known as electromagnetic radiation.

Figure IX-1 shows the spectrum of electromagnetic radiation. It is

divided into different regions whose approximate boundaries are
shown by broken lines. The waves in each region have their own
name: gamma rays, ultraviolet light and so on. The ways of
producing the waves are different for the different regions. For
example, gamma rays are generated by radioactive nuclei, while
microwaves are produced using special vacuum tubes. Frequencies
and wavelengths belonging together are connected horizontally in
the figure: gamma rays of frequency 1020 hertz (point A) have a
wavelength of 3 X 10”10 cm (point B), for example. Since the speed
is equal to the frequency multiplied by the wavelength, these gamma
rays travel at 3 x 10locm per second, which is 164

ELECTRIC AND MAGNETIC FIELDS

FIGURE IX-1. The different regions of

the electromagnetic spectrum. Note

the scales showing frequency and

wavelength: each step denotes a

change by a factor of 10. Such scales,

called logarithmic scales, are used

when the numbers vary over a

tremendous range of values, as is the

case here. 1 hertz = 1 oscillation per

IO24

IO22

1020

10

18

£ 10

16

I10‘4

lio12

10

in

10

106

104

102

1
gamma rays

-A

X-rays

ultraviolet light

ImsIBIeillgEt”!

infrared light

microwaves

television, FM and

AM radio

long wavelength

waves

10

,-14

10

10

10

,-10

-6

10
-2 M
IO4
IO6

IO8

10

10

also the speed of visible light. You can verify from the figure that this
is the speed of all the waves in the entire spectrum.
3 Electric and magnetic fields
In this section, I describe what it is that is waving in an
electromagnetic wave. In order to do this, I must first explain what is
meant by the term field as used in physics. This is 165

IX GLASS PANES AND ALUMINIUM FOILS

FIGURE IX-2. The arrows suggest the

motion of different parts of the air

inside a heated oven. The temperature

of the air and the direction and

magnitude of its speed vary from

point to point. One speaks of

temperature and velocity fields

for the air in the oven.

■••■■7

/\r

*p

/
a term which we use here with a special meaning which is different
from what it has in everyday speech. I begin with a simple example:
a kitchen oven which has been turned on. I show the oven in fig. IX-2
as a hollow box. The air inside the oven is steadily getting hotter. At
some instant of time, which I denote by the symbol t, I measure the
temperature T at every point inside the oven. I can specify any point
P by noting its position vector r from one corner of the oven, as
shown in the figure. Changing the direction and/or magnitude of r will
let me place P anywhere in the oven. What I have done, at the
chosen instant of time, is to make a snapshot picture, so to speak, of
the temperature everywhere inside the oven. I represent this picture
by the symbol T(r,t).

I choose this form for the symbol to be reminded of the following: 1.

It stands for the temperature T.

2. The temperature depends on the position r and time t.

This symbol stands for the temperature at every point in the oven at
all different times, and is therefore a very convenient shorthand for a
lot of information. In words, I say that there is a temperature field
inside the oven, which depends on position r and time t, and is
denoted by T(r,t).

question Imagine that the air inside the oven is not still, but moving
about as suggested by the curved arrows in fig. IX-2. Can you think
of a field associated with this motion?

166

ELECTRIC AND MAGNETIC FIELDS

answer Yes. At each point P the air is moving with a certain speed in
some direction, both of which may change with time and with the
position of P. I can therefore think of a velocity field which I can
denote by v(r, t) for the air in the oven.
The temperature field is a quantity which has a magnitude (so many
degrees Celsius, for example), but no direction. We call such a
quantity a scalar. The temperature inside the oven forms a scalar
field. The velocity of the regions of air inside the oven forms a vector
field. What is common to both fields is that, over a given volume of
space and span of time, a certain quantity which can be a scalar
(e.g. temperature) or a vector (e.g.

velocity) is specified at every point.

The two fields I have described, temperature and velocity, are

properties of the medium, namely air, which fills the space of
interest, namely the oven. I now look at another example which will
present a new and very important aspect of the field. Consider an
electric charge, say an electron, which is fixed at some point A in
space. If I now place a unit positive electric charge at a point B
anywhere in the space surrounding A, it will feel a force because of
the electron’s charge. Figure IX-3 shows how this force decreases
as the distance between A and B increases. The force is a vector,
since it has both a magnitude and a direction, and it can be
measured at all points in the space surrounding the electron A. Thus
this force satisfies the conditions for a vector field. I give this force a
name and a symbol: the electric field due to an electron, with the
symbol E(r, t). The vector r here is the vector from A to B, and
specifies the point in space at which I measure the electric field at
time 1.1 shall sometimes denote the electric field more simply with
just the symbol E, keeping in mind that it may vary with time and
from point to point.

I can write down a formula for the force which the unit positive
charge feels due to the charge of the electron. I just have to multiply
the charges together and divide by the square of the distance apart:

E = 1 X e/r2 - e/r2 in magnitude, pointing towards the electron.

FIGURE IX-3. The force on a unit

positive electric charge placed at B

due to an electron with charge —e

placed at A, as the distance between

the two is varied; the force decreases

as the distance increases.

<■!

S■i

due

oq

in

■ -i

! JltlllllTllTlTllTOlirinTirnrii

increasing distance from B to A

167

ELECTRIC AND MAGNETIC FIELDS

resemble those of a large bar magnet lying along its axis of rotation.
I illustrate in part (a) of fig. IX-4 the influence of a bar magnet on
compass needles placed in various positions in its neighbourhood.
At each position, there is a definite direction in which the needle
points. To turn the needle away from this direction requires a twisting
torque which varies from point to point, becoming smaller the further
it is from the magnet. Thus the influence of the bar magnet upon the
compass needle has two aspects: a direction (the direction of the
needle) and a magnitude (proportional to the torque needed to
disturb the needle). Further, it is an influence which pervades the
space surrounding the magnet. I can include all of this in the
statement that a bar magnet creates a vector field in the space
around it, called a magnetic field and denoted by the symbol B. As
with the electric field E, the magnetic field B may in general vary with
time and position. The direction of the magnetic field vector
determines the direction of the compass needle, and its magnitude
tells us how strongly the needle sticks to this direction.

A bar magnet is not the only source for a magnetic field. An electric
current flowing in a wire will also produce a magnetic field in the
surrounding space, as illustrated in part (b) of fig. IX-4. It shows a
short section of the current-carrying wire, and the orientation of
compass needles placed at different points around the wire. The
orientation shows the direction of the magnetic field, whose strength
decreases with increasing distance from the wire.

I summarize now what I have said so far in this section. A field is a

quantity which may be a scalar or a vector, whose value is specified
everywhere in some volume of space and at different times. An
electric field is produced by electric charges, and can be sensed by
the force it exerts on an electric charge. A magnetic field is produced
by a bar magnet or by an electric current, and makes itself known by
its effect on a compass needle. Both these fields are vector fields.

Let us now imagine some volume of space which is empty of matter

but has both an electric and a magnetic field everywhere in it. To
begin with, let us assume that these fields do not change with time,
although they may be different at different points in the volume. They
are produced by some distribution of charges, currents, and magnets
all lying outside the volume. We need to know nothing about them
other than that they exist: our attention is focused on the field
distribution inside the volume.

Now suppose something is done to the charges which are

responsible for the electric field so that the field begins to change
with time. We find a new phenomenon: the changing electric field
produces a magnetic field which has nothing to do with the magnets
and current-carrying wires outside the volume. Its connection is only
with the changing electric field. A similar thing happens when the
roles of the two fields are interchanged: a changing magnetic field
produces an electric field. You can see some sort of symmetry
between electric and magnetic fields here. A change with time of
either field produces the other. The exact rules which govern this
behaviour were written down in mathematical form by James Clerk
Maxwell. The symmetry between the two fields was shown by Albert
Einstein to follow from the theory of relativity.

169

GENERATION OF LIGHT

it. The wave consists of electric and magnetic fields, and to sense
them I must place something suitable at P. To see what the electric
field is doing, I imagine an electron at P. It senses the oscillating
electric field of the wave moving past it. It feels an increasing force
which reaches a maximum when the crest of the wave (point C in the
figure) reaches it. The force then begins to decrease to zero and
reaches a maximum in the opposite direction when the point D of the
wave reaches P, and so on. Thus the electron feels an oscillating
force at a frequency f, the frequency of the wave. The electron also
has a magnetic moment, and so it senses the magnetic field too.

Consequently, it also feels a force tending to align it along the

magnetic field, oscillating at a frequency/
You see from fig. IX-5 that the oscillating electric field remains in a
fixed plane, So also does the magnetic field, but in a plane which is
perpendicular to that of the electric field. This is so for all
electromagnetic waves. If I look at the bundle of waves which make
up a beam of light coming from the sun or from a lamp, I find that
these two planes, while still perpendicular to each other, are
differently oriented for the different waves. Such light is called
unpolarized light. By passing this light through certain materials
generally known as polarizers, it is possible to filter out all waves
except those having the electric field in a single specified plane. The
resulting light is called polarized light. A material which is often used
in sunglasses contains layers of certain long molecules packed
parallel to one another. Their electronic structure is such that they
absorb waves whose electric fields are parallel to their length, and let
through those whose fields are perpendicular to their length. The
light reaching us from the sky is sunlight which has been scattered
by the molecules in the atmosphere, and is partially polarized as a
consequence. You can verify this by looking at the sky through
polarizing sunglasses, and noting the change in brightness as you
rotate the glasses one way and the other.

We have achieved a very powerful simplification by thinking of an

electromagnetic wave in terms of its electric and magnetic fields. The
wave of course needs some source to produce it: a light bulb if it is in
the visible spectrum, a radioactive nucleus if it is a gamma ray, and
so on. We do not need to worry about the complicated processes
going on in the source which lead to the production of the wave in
order to understand what happens when the wave meets a solid.
Nevertheless, I describe briefly in the next section two ways in which
electromagnetic waves can be generated from matter: this is where
the wave leaves the source.
4 Generation of light
You have seen neon tubes which give off red light and are used in
advertising signs.

Such a tube contains neon gas, and is so constructed that when an

electric voltage is applied between its ends, a current of electrons
passes through the gas. These electrons collide with the neon atoms
and transfer some of their kinetic energy to the atoms, 171

IX GLASS PANES AND ALUMINIUM FOILS

which therefore change to one of their higher quantum energy

states. The atoms subsequently lose this extra energy by emitting a
photon and return to their previous state, and the process begins all
over again. Thus the photons can only have those energies (and the
corresponding frequencies) which are the differences between the
higher energy states and the initial state of the atoms. In the case of
neon, most of the photons come from a state whose energy is higher
than the initial state by an amount which is the energy of photons of
the colour red. With other gases, other colours may dominate the
light that comes out. A characteristic of such light is that it consists of
waves of only certain frequencies, which are fixed by the quantum
energy levels of the atoms.

I can also produce light from a body if I make it hot enough:

examples are the light from an incandescent electric bulb in which a
filament of tungsten is made white-hot by passing an electric current
through it, or from the sun which stays hot because of processes
going on in it involving nuclei. When I pass such light through a glass
prism, which bends light of different frequencies by different
amounts, I get a continuous spread of colours from violet to red. By
using suitable instruments, I find that the light also contains
frequencies which the human eye cannot see, such as infrared
(frequencies below the red) and ultraviolet (frequencies above the
violet). In fact, the entire electromagnetic spectrum is present in such
light, though most of the photons in sunlight are in the range of
visible light, and there are fewer photons the further away one goes
from this range.

How does this continuous range of frequencies come about? We

know that the thermal energy of a body is just the energy of motion
of its constituent particles, which are atoms, nuclei, electrons and so
forth. The energies of individual particles vary from zero to very large
values, and the distribution of the particles among the different
energy ranges depends on the temperature of the body. We saw all
this in chapter VI. It is this very complicated thermal motion of the
electric charges, namely electrons and nuclei, which produces the
continuous spectrum of electromagnetic radiation, which is therefore
called thermal radiation. We know that if the temperature of the body
is T kelvin, then the average energy of motion per particle is about
k& T, where k^ is the Boltzmann constant. I shall refer to this quantity
as the thermal energy per particle.

If I express the photon energy as so many times the thermal energy,

and look at the distribution of the photons in the thermal radiation
among different energy ranges, J

get the result shown in fig. IX-6. The shape of this distribution results
from the fact that photons are bosons, and is the same whatever the
temperature of the body: 3000 K

for the filament in an electric bulb, 6000 K for the surface of the sun,
and so on. You can see that the largest number of photons are at
energies which are about three times the thermal energy. As the
frequency of the photon is proportional to its energy, this means that,
for example, the largest numbers of photons in sunlight are at
frequencies which are twice those in the light from an electric bulb.
When I put the actual numbers in, I find that most photons from the
sun are in the red part of the spectrum with a wavelength of about 8
X10”5 cm, while they are in the infrared (wavelengths around 1.6
X10”4 cm) in an electric lamp.
172

ABSORPTION OF LIGHT

FIGURE ix-6. The energy distribution

in thermal radiation, showing the

relative numbers of photons in

different energy ranges at any

temperature T. The horizontal scale

shows the ratio of photon energy to

the thermal energy k%T. While

photons of all frequencies are

produced whatever the temperature of

the body, most of them are in a range

of energies which are about 3&BT.

This range is in the visible and

infrared for light from the sun, whose

surface temperature is about 6000 K.

5 10
photon energy/thermal energy

question As you read this, you and your surroundings are also giving
off a continuous spectrum of electromagnetic radiation. What
frequencies do most of these photons have relative to those from an
electric lamp?

answer The temperature around you is about 20 °C which is the

same as 293 K, which in turn is about one-tenth of the temperature
of the filament in the electric bulb. So the radiation has most of its
photons at frequencies which are one-tenth of those from the bulb.
One can call this the far infrared part of the spectrum. It is this
radiation which is picked up by devices used for night vision.

One can see from fig. IX-6 that the lower the temperature, the lower
the range of photon energies at which the largest numbers of
photons are found. The universe is filled with electromagnetic
radiation which corresponds to a temperature of about 3 K; this
means that the largest number of photons in this radiation have
wavelengths in the vicinity of a millimetre. This radiation is a residue
of the big bang which was the start of our universe 1O10 years ago.
5 Absorption of light
All solids consist of particles, electrons and nuclei, which have two
properties which make them feel the influence of the electric and
magnetic fields in an electromagnetic wave: they are electrically
charged, and they are permanent magnets. In addition, quantum
mechanics tells us that the energies of the electrons are to be
described in terms of energy bands, and the energy of the
electromagnetic wave is quantized in terms of photons. Using this
information, we now seek answers to questions like why glass lets
light through and aluminium foil does not, and why copper is red and
sapphire is blue.

173

IX GLASS PANES AND ALUMINIUM FOILS

In this section, I concentrate on the effects the electric field in a light

wave has on the atoms which make up a piece of matter. The
electric field acts on the electric charges, namely the electrons and
nuclei which make up the atoms. The field acting on a given atom
will push the nucleus (positive charge) in one direction and the
surrounding cloud of electrons (negative charge) in the opposite
direction. This relative displacement of the two stops when the force
due to the electric field is balanced by the attractive force between
the nucleus and the electron cloud. I shall come back to this effect
later.

The light also behaves like photons, quanta of energy. A photon can
be absorbed by an electron, whose energy is then increased by the
energy of the absorbed photon. For this absorption to be possible,
there must be a quantum state of this higher energy into which the
electron can move. If there is no such state, then the photon cannot
be absorbed and will therefore continue on its way. Now we begin to
see why some materials are transparent and some are opaque.

Take glass as an example of a transparent material. It is an insulator,

and we know qualitatively what its electronic structure in terms of
energy bands must be. I show the two energy bands which are of
interest for the present purpose in fig. IX-7. The lower energy band is
full; every quantum state in it is occupied by an electron. Above it
and separated by an energy gap is an empty band with no electrons.
Lengths measured vertically on this figure represent energies. An
incoming photon with energy represented by the line AB, which is
less than the energy gap, cannot be absorbed by any of the
electrons in the filled band, because the resulting final state will
either be in the filled band and therefore already occupied, or must
be in the energy gap where there are no quantum states for the
electron at all. So such a photon would pass through the material.

Now consider photons of energy larger than the gap energy, CD for
example. An electron can absorb such a photon and go to a
previously empty state in the upper band.

We see now why glass is transparent to visible light. The band

structure of glass is such that the photons of visible light do not have
enough energy to make an electron cross the energy gap, and so
they pass through without being absorbed. But photons of ultraviolet
light have higher frequencies and therefore higher energies, and
these are enough to raise an electron from the full band to the empty
band, so that such light would be absorbed by the glass. In the
figure, AB would be a typical photon of visible light and CD of
ultraviolet light. The energy gap in quartz is larger than in ordinary
glass, and not only photons of visible light, but even those of
ultraviolet light, have insufficient energy to raise an electron to a
state across th’e gap. This is the reason that ultraviolet lamps are
made of quartz and not of ordinary glass.

I summarize what we have seen so far: light can be absorbed by the

electrons in a material if they can move to quantum states whose
energies are higher than the initial energy by an amount equal to the
energy of the photon. If there are no such states available in the
band structure of a given material, the photons go through the
material: it is transparent to light of that colour. This picture also tells
us why a material can be transparent to waves in one part of the
electromagnetic spectrum and opaque in another part.

174

ABSORPTION OF LIGHT

FIGURE ix-7. The electronic energy

bands in a transparent material like

glass. Photons with energies like AB,

that are less than the energy gap

between the occupied states and the

empty states, cannot be absorbed by

the electrons because they have no

final states available of the right

energy. Such photons will pass

through the material, while those

with higher energies, like CD, are

absorbed.

photon erergy
empty

energy gap

full

Now et us look at the case of light falling on the surface of a metal.

We know that the metal conducts electricity well, because the outer
electrons of each atom are free to wander all through the metal.
Figure IX-8 shows a beam of light reflected from the surface of the
metal. The free electrons closest to the surface feel a force due to
the oscil-latmg electric field in the light wave, and so they start
oscillating at the same frequency as the light. One of these
electrons, labelled 1, is shown in the figure. This oscillating layer of
electrons is taking energy from the light wave, which therefore is
weakened and produces smaller oscillations in the next layer of
electrons (labelled 2) and so on IX GLASS PANES AND
ALUMINIUM FOILS

FIGURE IX-8. Absorption and

reflection of light for a metal. The

oscillating electric field of the light

wave sets the conduction electrons

(four shown with two labelled 1 and

2) in oscillation. The amplitudes of

the oscillations are suggested by the

arrows. The electrons thus absorb

some of the photons’ energy and re—

radiate the rest.

surface of metal

©1

<• ©2 >

<■■•• © •■•>

<• © >

inside the metal

until at some depth below the surface the light wave and the electron
oscillations have both died out. This depth in most metals is very
small: a few tens of angstroms (recall that one angstrom is 10”8 cm).
So we can see now why even a very thin sheet of metal lets no light
through.

You might wonder where the light which is reflected from the surface
of the metal comes from, since it looks as if all the light falling on the
surface got absorbed in setting up the oscillations of the electrons.
You recall that electromagnetic waves are generated by oscillating
electric charges. The surface layers of electrons are oscillating at the
same frequency as the incoming light. These oscillations produce
electromagnetic waves of the same frequency, and these waves are
the reflected light.

What I have described so far seems all right for a metal like
aluminium, which reflects daylight without changing the colour of the
reflected light. But what about metals like copper and gold, which
look red and yellow respectively in ordinary daylight which has all the
colours of the visible spectrum? The answer lies in a special feature
of the electronic band structure of these metals, which I illustrate in
fig. IX-9. The conduction band, labelled band 1 in the figure, has all
its quantum states occupied by electrons up to the Fermi energy,
and the states of higher energy are unoccupied. This part of the
band 176

ABSORPTION OF LIGHT

FIGURE ix-9. The electronic energy

bands in copper and gold. A narrow

energy band, labelled band 2,

overlaps the occupied states in the

conduction band labelled 1. The large

number of electrons in band 2 absorb

photons of energies equal to or

greater than CD, thus removing these

colours from the reflected light. A

photon with an energy like AB

cannot be absorbed by these

electrons. The result is the colour that

we see of copper and gold.

empty states

B
A

occupied stales

&- Fermi energy

Z8l?r.:.:.-:..;.i:

Hi

structure is the same as that for aluminium, and should therefore

produce the same behaviour with respect to light. There is however
a new feature in this figure: a second band (labelled band 2) spread
over a narrow range of energies which are lower than the Fermi
energy and separated from it by a gap in energy. Every quantum
state in this band is occupied by an electron, and there are a lot of
them, about ten times the number of electrons in the conduction
band. Photons with energy of the amount indicated by the line AB in
the figure, where B can lie anywhere below the Fermi energy, cannot
be absorbed by electrons in band 2, because all the states with this
additional energy are already occupied by electrons. These photons
can however set the conduction electrons in band 1 into oscillation
and eventually lead to reflected light just as with aluminium.

On the other hand, photons with energy equal to or greater than the
amount corresponding to CD can be swallowed up by electrons in
band 2 which then move to previously empty states above the Fermi
energy. The reflected light will therefore be missing all these photons
and will be different from the light which hit the surface of the metal.

The critical photon energy corresponding to CD in fig. IX-9

corresponds to yellow for copper metal, and to green for gold metal.
This means that all the colours in daylight from yellow to blue are
absorbed in copper, and only red is reflected: hence the red colour of
the metal. By similar reasoning, you can see why gold has a yellow
colour.
question The electron which went from the state C to the state D by
absorbing the energy of a photon can later go back to C emitting an
identical photon. So why should there be any change in the colour of
the reflected light?

177

IX GLASS PANES AND ALUMINIUM FOILS

answer Yes, the electron will go back to C; but the photons emitted
in this process come out in all directions, and so the number in the
direction of the reflected beam will be much smaller than that in the
incoming beam of light.

question Brass is an alloy made by dissolving some zinc in copper.

The colour of brass is yellow instead of red. Why?

answer The addition of zinc to copper changes the energy band

structure such that the critical energy CD becomes larger than it was
in pure copper, changing the colour corresponding to the energy CD
from yellow (as in copper) to green. Thus only the reddish-yellow
part of the spectrum is reflected, and the rest is absorbed, by brass.

question How does one explain the red colour of ruby or the blue of
sapphire, which are insulators and therefore have no free electrons?

answer Both ruby and sapphire are a crystalline form of a compound

containing two atoms of aluminium for every three atoms of oxygen.
In its pure form this compound is called corundum, and is a clear
colourless solid. Ruby is a form of this compound containing about
one percent of chromium atoms replacing the same number of
aluminium atoms, and sapphire has iron atoms instead of chromium
atoms replacing the aluminium atoms. To be specific, I shall consider
ruby. Each chromium atom is surrounded by aluminium and oxygen
atoms, whose influence changes the energy levels of the chromium
atom from what they were in the isolated atom. These modified
levels are so distributed that most photons with wavelengths shorter
than that for red light are absorbed. So when ruby is viewed in
ordinary light, only the red comes through. In sapphire, the iron
atoms play the same role, except that now the energy levels are
different from what they were for chromium, and all portions of the
spectrum other than the blue are absorbed by the crystal.

It is the interaction of the photons with the electrons in the material

which determines what happens to the light falling on glasses,
metals, and minerals. The details depend on how the quantum
energy levels of the electrons are arranged in a given material. In a
material which is made up of molecules, there are other energy
levels due to the internal motions of each molecule. A photon of the
right energy would be absorbed by a molecule which thereby moves
to one of these higher energy states. To summarize:

If a photon of a certain energy, or equivalently an electromagnetic

wave of a certain frequency, is absorbed by a given material, then
one can conclude that the absorbed energy has put the material in a
new quantum state whose energy is higher by exactly that amount.
There can be no absorption of the photon if such a quantum state
with the right energy is not available.

178

ABSORPTION OF LIGHT

The most widely occuring molecule in the world is the molecule of

water, consisting of two atoms of hydrogen joined to one atom of
oxygen. The atoms vibrate about their average positions, and the
molecule as a whole rotates about an axis. These motions lead to a
spectrum of quantized energy levels which are in addition to the
electronic energy levels of each atom in the molecule. You can
imagine that all this leads to a complex set of energy levels capable
of absorbing photons of a wide range of frequencies. There is one
remarkable exception which makes life as we know it on this planet
possible. In that multitude of energy levels, there are very few in that
narrow range which would cause photons of visible light to be
absorbed. Figure IX-10 shows the thickness of water needed to
absorb 99 per cent of the electromagnetic waves of different
frequencies passing through it. Over most of the range of
frequencies from microwaves to X-rays, only a fraction of a
centimetre of water is enough to absorb almost all the radiation. For
the narrow range of frequencies of visible light, however, the
absorption becomes very weak: it takes tens of metres of water to
absorb 99 per cent of such light. When water is in the form of vapour,
as in the atmosphere, there are fewer molecules per metre of
thickness, and the absorption is therefore even less. If this narrow
window of transparency of water to visible sunlight did not exist, such
light would not have reached the surface of the earth, and no life as
we know it could have evolved. Given that it is solar energy which
sustains life on earth, one might say that a quantum-mechanical
peculiarity of the energy levels in water permitted that particular form
of life to evolve on earth which utilizes the photons of visible light.
One might speculate on the form that life might have taken if the
window of transparency lay in some other part of the solar spectrum:
such is the stuff of science fiction.

The microwave oven found in many kitchens is an application of the

absorption of electromagnetic waves by water and fat. The oven
contains a generator which produces standing microwaves at a
frequency of about 2 X 109 hertz in the chamber where the food is
placed. These photons are absorbed mostly by the molecules of
water and fat FIGURE IX-10. The thickness of a

sheet of water which will absorb 99

per cent of electromagnetic waves ,

passing through it, at different

frequencies. The inset shows the

kinds of rotational and vibrational

motions (indicated by arrows) of a

water molecule, whose quantum

energy levels are responsible for this

absorption.

.2 0.00001

a.

On

On

0.001

iooo—

increasing frequency

179

IX GLASS PANES AND ALUMINIUM FOILS

that are present in all foodstuffs, thus raising their energy and
temperature. The plastic or glass of which the container is made
does not have those energy levels which would permit the
absorption of these photons, and therefore remains relatively cool.
The conduction electrons in a metal, on the other hand, can absorb
these photons and thus heat the metal, as anyone who has
mistakenly left a metal spoon in the oven will have noticed.

question Why must the food be placed on a rotating platform in a

microwave oven, but not so in an ordinary oven?
answer Since there are standing waves in the microwave oven, with
a wavelength comparable to the size of the oven, the strength of the
electromagnetic field varies from zero to its maximum value at
different points inside the oven. Without rotation, the food would be
heated non-uniformly. The heating in an ordinary oven is due to
infrared radiation of much shorter wavelength, and the problem of
non-uniform heating does not arise.
6 Refraction of light
We have seen in the previous section how the existence of quantum
energy levels in a solid leads to the absorption of electromagnetic
waves as they pass through it. Such absorption means that the
radiation progressively loses energy as it moves through the solid.
The radiation has some other characteristics too: it has a
wavelength, a frequency and a speed. Fixing any two of these
quantities automatically fixes the third, because frequency multiplied
by wavelength equals speed of wave. So let us see what happens to
the frequency and the wavelength when radiation passes from empty
space into a transparent material like glass, where the absorption is
negligible.

The frequency of the radiation is determined initially by the source:

the sun, or an electric light, for example. When this radiation passes
through the material, the oscillating electromagnetic field acts in
opposite directions on the charged electrons and nucleus of each
atom and pulls them apart a bit. The atom in this condition is said to
be in a state of polarization, or is polarized. Because the field which
produces the polarization is oscillating in magnitude as well as
direction, so will the polarization itself, and at the same frequency
too. This oscillating polarization produces radiation at the same
frequency, and so the process continues: a complicated interaction
between the electromagnetic field of the radiation and the charged
electrons and nuclei in the material. In any event, the conclusion is
that the frequency is unchanged in the material.

How about the wavelength of the radiation when it is in the material?

Figure IX-11

shows an electromagnetic wave travelling from air into a crystal. For

the present purpose, air can be treated as empty space, because the
wavelength is practically the same in both, and we denote it by X(0),
as indicated in the left half of the figure. If the 180
REFRACTION OF LIGHT

FIGURE IX-11. A light wave of

wavelength A.(0) in air enters a transparent

crystal, in which its wavelength is

reduced to A,(l). The atoms, shown as

black dots like A and B, are polarized

to different extents by the electric

field of the light wave, which is itself

modified by the polarization. The

amount of polarization of each atom

is suggested by the distance and

direction of the hollow square from

the black dot representing the atom.

The result is a shortening of the

wavelength in the crystal.

wave had gone on through the crystal without being affected by it,
then it would be like the broken line in the right half of the figure, with
the same wavelength X(0). But in reality the wave is affected by the
crystal, because it produces displacements of the electrons and
nuclei which make up the atoms. The magnitude and direction of this
atomic polarization varies from atom to atom, and I have indicated
this by a small square against each atomic position at a distance
proportional to the polarization. For example, the polarization of atom
A is about equal in magnitude but opposite in direction to that of
atom B. The electric field is tied to the polarization at each atom, and
so we have an oscillating electric field with the same wavelength as
the oscillating polarization, and this is then the light wave in the
crystal. Note that the light wave in the crystal is associated with the
movement of things with mass, namely electrons and nuclei. This is
in contrast to the same light wave in free space, where there is no
mass to be moved. It is this difference which leads to the wavelength
of light being shorter in matter than in free space. Since speed is
equal to frequency multiplied by wavelength, the speed of light in any
material medium is less than its speed in free space.

The fact that the speed of light is less in a solid than it is in free
space causes a beam of light to change its direction as it goes from
one to the other. This effect, known as the refraction of light, is what
makes possible a host of optical equipment: eyeglasses, cameras,
telescopes, and microscopes, to name a few. I illustrate how
refraction comes about in fig. IX-12. A beam of light comes from the
left and enters a transparent solid.

If there were no solid, the wave would continue in the same

direction, as shown by the broken lines. The motion of the wave is
perpendicular to the wavefront, as you can see for example with
waves on water. Now consider the wavefront AB in the figure. The
part of the wave at B has just reached the surface of the solid, while
the part A still has the distance AC to go before it reaches the
surface. During the time it takes for this to happen, the part of the
wave at B will have travelled to D, and the distance BD will be less
than the distance AC because of the lower speed in the solid. The
wavefront in the solid will be the line CD, and the wave will be
travelling at right angles to this line.

181

IX GLASS PANES AND ALUMINIUM FOILS

FIGURE IX-12. Refraction of a beam

of light on entering a crystal from

empty space. The light takes the same

time to get from A to C as it does

from B to D, because its speed in the

crystal is less than its speed in empty

space.

empty space

crystal

This last condition can be satisfied only if the beam in the solid is in a
specific direction, and none other, relative to the beam in empty
space.

question In terms of the lengths in fig. IX-12, what is the speed of

light in the solid, v, in terms of the speed in empty space c?

answer Light travels the distance AC in free space in the same time
that it travels the distance BD in the solid. Therefore the ratio of vto
cis equal to the ratio of the length BD to the length AC.

question Is it possible to see from fig. IX-12 why a pool of water

appears shallower than it actually is when viewed from above?

answer Yes. Imagine the solid replaced by water, and the arrows
showing the direction of the beam of light to be reversed. So now the
light is coming from the bottom of the pool, say, and its direction is
changed as shown when it emerges from the water. A person
looking into the water will therefore see the bottom in the direction
shown by the broken lines.
The bottom will thus appear to be nearer to the surface of the water
than it actually is. In fact, the actual bending of the beam of light as if
emerges from water to air is such that a pool which is two metres
deep will seem to be only a metre and a half deep: inexperienced
swimmers, beware!
7 Dispersion of light
The speed of an electromagnetic wave in empty space, whatever its
frequency, is the same,“and we denote it by the symbol c, with a
value of 3 X 1010 cm per second. When the wave travels through a
material medium, not only is its speed reduced, but the amount of
reduction depends on both the frequency and the medium. This
effect, 182

DISPERSION OF LIGHT

figure ix-13. How a ray of sunlight

is refracted and reflected in a raindrop

to produce a rainbow. The slight

differences in the speed of light in

water for different colours is an

essential factor in this phenomenon.

sunlight

violet

rainbow ^ red

called dispersion, is greater the more densely packed are the atoms
in the medium. For example, the dispersion of light in air is negligibly
small, so that light of any colour has practically the same speed, c, in
air. In water or in glass, on the other hand, there is a significant
decrease in speed which is different in the two substances and for
different colours. The amount of bending of light as it enters (or
leaves) water or glass from air therefore depends on the colour. This
is the origin of the rainbow, and the spectrum of colours produced
when sunlight passes through a glass prism, to name two examples
of dispersion. I illustrate these effects in figs. IX-13 and IX-14.

Figure IX-13 shows the path of a ray of sunlight as it enters a

raindrop, is reflected once and emerges into the air again. The parts
of the ray corresponding to different colours are bent differently
during this passage, and emerge in slightly different directions as
shown. If I stop at this point, I would have no rainbow, for the
following reason.

Suppose the ray of sunlight entering the raindrop at a particular spot

is sending red light in my direction. A ray entering at another spot
may send blue light in my direction, and so on, so that in the end my
eye receives all the colours and therefore sees just white light.

But this does not happen. The rules which govern reflection and
refraction of light combine with the geometrical properties of the
sphere (which is the shape of the raindrop) to make the light coming
out in a particular direction (which changes slightly with colour) to be
more bright than in all other directions. The result is the rainbow.

The fact that the speed in water of blue light is lower than that of red
light is what ultimately causes blue to be on the inside and red on the
outside of the rainbow. Only a part of the light emerges from the
raindrop after the first reflection. The rest is reflected once more at
the raindrop’s surface, and a part of this emerges and gives rise to
the second rainbow which can sometimes be seen. The order of
colours in this rainbow is the reverse of that in the first. This fact of
nature has been overlooked by some artists famous for their
landscape paintings. The Neue Pinakothek in Munich for example
has two landscapes with double rainbows, painted by Jan Wildens
and by Josef Anton Koch, where the colours are not in the order that
nature says they should be.

183
IX GLASS PANES AND ALUMINIUM FOILS

FIGURE IX-14. How a glass prism

produces the spectrum of sunlight

through two successive refractions.

sunlight

spectrum

Figure IX-14 shows how a triangular glass prism breaks up sunlight

into its constituent colours. This is the basic effect which is
responsible for the sparkling colours of cut diamonds and the beads
in a glass chandelier, for example. It is used in the construction of
one type of an instrument called a spectroscope, which shows the
spectrum of frequencies in light which passes through it. We have
seen that each type of atom -

oxygen, carbon, etc. — has its own characteristic set of energy

levels which is different from that of all other types. When the atom
absorbs some energy (for example when the material is heated
sufficiently), it goes to a higher energy level and subsequently
returns to a state of lower energy. The difference in energy E is given
off by the atom as a photon, which is light of a certain frequency
/related to this energy through the Planck constant k.

E=hf.

Thus the light coming out has a spectrum of colours with certain
frequencies which are characteristic of the atoms producing it.
Knowing the spectrum, one can deduce which type of atom it comes
from. The rainbow is the spectrum of the thermal radiation from the
sun produced by a spectroscope that occurs in nature, namely rain.
8 Solar batteries and LEDs
A solar battery is a semiconductor device which converts some of
the energy it receives from sunlight into electrical energy, just as an
ordinary battery converts chemical energy into electrical energy. The
LED (acronym for /ight-emitting rfiode) is a semiconductor device
which does the reverse: it uses electrical energy to produce light. A
solar battery is a practical application of what we have learnt about
the energy bands in semiconductors.

Figure IX-15 shows how a solar battery works. In part (a), an n-type
and a p-type semiconductor are close together but not in contact.
The n-type has electrons, and the p-type has holes, moving about. In
part (b), the two semiconductors are brought into contact to form
what is called a p-n junction. The electrons and holes at the junction
184

SOLAR BATTERIES AND LEDS

FIGURE IX-15. The working of a

solar battery, (a) Isolated p-and n-type semiconductors, (b) A p-n

junction. The electrons and holes

neutralise each other at the junction,

leaving a layer of positive ions and

one of negative ions facing each other

and forming, as it were, the positive

and negative terminals of a battery, (c)

Sunlight falling on the junction

creates additional electrons and holes

which are driven as an electric current

through a gadget connected to the

battery with wires.

n-type

n-type

(a)

p-type

(b)

n-type

sunlight

G+i-R © e

©+©e

e e Jbs © © ©

©I!

©© ©I!-©©©©

e©e | e g

holes, current.
gadget

p-type (c)

electrons

© electron

© hole

- negative ion

+ positive ion

move across and neutralize each other, leaving positive ions on one
side and negative ions on the other.

question Why does the process not continue, thus neutralizing all the
moving charges on both sides?

answer An electron from the left which tries to move through the
junction meets the phalanx of negative ions on the right and is
repelled; it cannot get through. The positive holes on the right are
similarly stopped by the positive ions on the left. Remember that like
charges repel one another.

185

IX GLASS PANES AND ALUMINIUM FOILS

The situation in part (b) is what obtains in a solar battery. When

sunlight falls on this device, as in part (c) of the figure, some of the
photons give electrons in the valence band enough energy to move
them to the conduction band, leaving empty states in the valence
band. The electrons in the p-type and the holes in the n-type feel a
force due to the ions at the junction shown in part (b), which makes
them flow in the directions shown through a gadget (which could be
a light bulb or motor or whatever) connected with wires to the device.
The direction of the electric current itself is taken to be opposite to
the direction of electron flow and in the direction of the hole flow, and
is therefore as shown in the figure.

question What happens to the electrons in the n-type and the holes
in the p-type which are left behind?

answer: They find their way to the region of the junction, where they
neutralize each other.
9 Summary
This chapter deals with the nature of light, and how it interacts with
material bodies.

In section 2, we are reminded that light manifests itself as either

waves or particles, depending on the mode of observation: it has the
property of duality, just like electrons, protons, neutrons, etc. Visible
light is just a small part of electromagnetic radiation, whose
wavelengths can range from 10Iocm to 10”14cm and even beyond
these limits. All these waves travel at the same speed through empty
space of about 3 x 1010cm per second, denoted by the symbol c.
This speed is one of the constants of nature.

I introduce the idea of a field in section 3 with the examples of the

temperature and the velocity of circulation of air at different points in
an oven. I then describe electric and magnetic fields, and show how
light waves (as well as all other electromagnetic waves) consist of
these fields. In what follows, I consider mostly visible light, but the
basic ideas apply with suitable modifications to all electromagnetic
radiation.

I describe in section 4 two ways in which light can be generated. In

one, the quantum state of an atom or molecule changes to one of
lower energy, and the difference in energy is carried away by a
photon. Since the energy levels are discrete, the resulting photons
can have only certain frequencies and nothing in between. An
example is the light from a neon tube. The other way to get light is
when the energy of motion of the constituent atoms, electrons, etc.
due to the temperature of the material is transformed into photons.
The energies of these motions are quantized, but lie so close
together that the resulting photons have a practically continuous
distribution of frequencies, examples of which are the light from the
sun or from an incandescent electric lamp.
Some materials are transparent to light, and others are opaque. I
show in section 5

186

SUMMARY

how this difference depends on the relationship between the photon

energies and the quantum energy levels in the material. This
relationship also explains the colours of materials like copper, gold,
and minerals.

The speed of light depends on the substance through which it is

passing, and is highest in empty space. A consequence of this is
refraction, which is the change in direction of a light beam in passing
from one substance to another. How this comes about is described
in section 6.

In section 7,1 find that the reduction in the speed of light is different
for different wavelengths and substances. This is called the
dispersion of light. This is the origin of the rainbow, and the principle
is used in an instrument called a spectroscope which analyses light
from a source into its constituent wavelengths.

I describe in section 8 a practical application of the interaction of light

with matter, namely a solar battery. We see how the nature of the
quantum energy levels in a semiconductor p-n junction helps to
convert the energy of the photons in sunlight into electrical energy.

This concludes my description of electromagnetic waves and their

interaction with materials. It is also possible to have electric and
magnetic fields which are not part of an electromagnetic wave: an
electric battery, for example, is the source of an electric field, and a
magnet produces a magnetic field. Such fields also produce effects
in a piece of matter, because it has electrons and nuclei which not
only have electric charge but are also like permanent magnets. I
shall look at some of these effects in the next two chapters.

187

ELECTRIC BULBS AND INSULATED

CABLES

l Preliminaries

Electricity is one of the most widely used forms of energy: in the

home, in industry, for transportation and for communication, to name
just a few of its applications. It is generated at one place and
transported to another which may be hundreds of kilometres away. It
is carried in cables made of copper or aluminium strung all over the
landscape. Electricity inside a home is at about 230 volts in most
places and at 115

volts in North America. It is distributed by copper wires which are

covered with some sort of insulation. The rate at which electrical
energy is used in appliances is measured in watts: a 100-watt bulb, a
2000-watt furnace, and so on. I show all this, stripped down to the
barest essentials, in fig. X-1. The generator produces electricity,
which is carried through cables and wiring and finally arrives at the
appliance where it is used.

The figure also suggests an analogy to a hydraulic system. Table X-1

sets forth this analogy: similar items are set side by side in the two
columns.

In the hydraulic (electric) system, the pump (generator) raises the

pressure (voltage) of water (electrons), consequently raising its
(their) energy. The stream of water (electrons) flows down the pipe
(cable) to the turbine (toaster), where the energy is converted to
kinetic (thermal) energy. The water (electrons) returns (return) to
where it all started but with lower energy, and the whole cycle
continues. Notice that there is no loss of the working substance in
either case. It just keeps going round and round the circuit. There is
one difference between the two systems, which is not important for
the analogy I have drawn. The electric current is oscillating back and
forth 50 times (60 times in North America) per second, whereas the
water keeps going in the same direction all the time.

FIGURE x-l. A schematic picture

of an electricity supply. Only the

generator and the household

appliance are shown; practically

everything in between is in the open

gap. Underneath are named the

corresponding parts of a hydraulic

system which uses water under

pressure to drive a turbine. The

analogy is explained in the text.

transformers

switches

meters

etc.

appliance
\ household wiring
230 V
or
115 V,
30 W
water pump
pipes

pipes

turbine

188

PRELIMINARIES

Table X-1. Analogy between electric and hydraulic systems hydraulic

system electric system

working substance: water working substance: electrons

water pump electric generator

space inside pipe which carries metal (copper, aluminium) wire the
water which carries the electricity

water meters electricity meters

pressure voltage

appliance, e.g. turbine, which appliance, e.g. toaster which converts

potential energy of converts electrical energy water into kinetic
energy of into thermal energy

turbine
water current (amount of water electric current (amount of per sec)
electric charge per sec)

I shall say a few words about volts and amperes, which are
respectively the units for measuring voltage and electric current. If
the voltage of a battery is two volts, for example, it means that the
difference in potential energy of an electron in moving from one
terminal of the battery to the other is the charge multiplied by the
voltage, or 2e, where eis the magnitude of the electron’s charge. It is
this energy which gets converted to other forms of energy when the
electron goes through the appliance. The electric field and the
voltage are closely related. In a battery, for example, there is a
voltage between the terminals, and there is an electric field
everywhere in the space between the terminals.

The electric current is the amount of charge that is circulating per

second past any part of the circuit, and we have just seen that it is
the same everywhere in the circuit: in the wires connected to the wall
socket as well as through the toaster, for example. It is measured in
amperes. One ampere is a current of about 6 x 1018 electrons per
second.

In this chapter, we look at what happens when an electric field (or

equivalently a voltage) is applied to a solid, some of whose electrons
are free to move about under the influence of the field and thus
produce an electric current. This means that we shall be concerned
mostly with metals and semiconductors, since insulators do not have
mobile electrons. We shall find out how the electronic band structure
of metals and semiconductors leads to their characteristic electrical
properties, and why it is more difficult to pass an electric current
through an alloy than through a pure metal.
2 Electrical resistance
I consider a thin wire of tungsten connected to a battery. The electric
field between the terminals of the battery exerts a force on the
electrons in the metal, causing them continually to pick up speed in
the direction of the positive terminal. If there were nothing 189

X ELECTRIC BULBS AND INSULATED CABLES

to prevent it, the speed would continue to increase, and so would the
current. But with a long and thin enough wire, the current quickly
reaches a constant value and stays there. I find that the size /of this
current is proportional to the voltage V. So the voltage divided by the
current is a constant:

?=*

The quantity R is called the electrical resistance (or often simply

resistance) of the wire, and is in ohms if Fis in volts and I in
amperes.

The resistance of a wire depends on its size and shape. A

meaningful comparison of the resistances of different materials is
therefore possible only if they all were wires of identical size and
shape. Scientists have agreed on what this shall be: a cube of the
material with an edge of one cm. Its resistance measured between a
pair of opposite faces is called the resistivity of the material, and is
expressed in ohm-cm.

question An electric bulb carries a current of half an ampere at 220

volts.

What is its resistance?

answer The voltage is equal to the resistance multiplied by the
current.

Therefore the resistance is the voltage divided by the current, or 220

divided by a half, or 440 ohms.

We now want to understand how the quantum description of a metal

and a semiconductor explains the features of their electrical
resistance, and why an insulator does not carry a current when a
voltage is applied to it. I describe in section 3 the essential facts, and
go on to their explanation in the succeeding sections.
3 The observed properties
The first property to note is that the resistivity is different for different
materials. Table X-2 shows the resistivities of some metals, a
semiconductor, and an insulator, all measured at 20 °C. You can see
how greatly the three types differ from one another in their Table X-2.
Resistivities in ohm-cm of various materials at 20 °C

material resistivity in ohm-cm

copper (metal) 1.7 X10”6

silver (metal) 1.6 xlO6

lead (metal) 22 X 10”6

germanium 1

(semiconductor) (value is very sensitive to impurity content) glass

(insulator) 1012

(depends on type, but still very large)

190

THE OBSERVED PROPERTIES

resistivities. There is no other property of a solid which covers such a

vast range of values, differing at the extremes by a factor 1018 (or
even more if low temperatures and superconductors are included).

The effect of a change of temperature on the resistivity of metals and

semiconductors is shown in fig. X-2. The resistivity of a metal drops
with decreasing temperature, appearing to head towards zero at the
absolute zero of temperature. However, this decrease levels off at a
low temperature of a few kelvin. In a semiconductor like germanium,
the effect of lowering the temperature is the opposite: the resistivity
increases.

There are, in addition, many metals and alloys whose resistivity,

instead of levelling off, drops abruptly to zero at some temperature
which is different for different materials.

These are called superconductors, and are the subject of chapter

XII.

figure x-2. The variation of

electrical resistivity with temperature

for a pure metal, an alloy, and a pure

semiconductor. Note the logarithmic

scale for resistivity in the larger

picture which compares a metal and

a semiconductor. The inset compares

the resistivities of a pure metal and an

alloy based on the same metal, and

the resistivity scale here is the usual

linear one. The resistivity of the alloy

is greater than that of the metal by a

constant amount at all temperatures.

106i
a

4=1

co

CO

0>

10

-6

10

rl2

semiconductor

metal

100

temperature

200.

metal

temperature in kelvin 300

191
X ELECTRIC BULBS AND INSULATED CABLES

I look next at the effect of adding small amounts of a second element

to a metal and a semiconductor. Very roughly, one can say that
adding one per cent of a second type of atom to a pure metal
increases its resistivity at all temperatures by a constant amount of
about one microhm-cm, which is 106 ohm-cm. The inset in fig. X-2
shows the typical behaviour of a metallic alloy. The effect of an
added second element on a semiconductor is more dramatic and
complicated. Its resistivity can drop by a factor of as much as a
million, and the drop can be different at different temperatures.
4 Scattering of electrons
Suppose that I think of the electrons as particles moving about in a
metal wire and bouncing off the ions in the lattice. When I apply a
voltage to the wire, the electrons accelerate down the wire,
increasing their energy and producing a current. They would still be
colliding with the ions and thus losing some energy, and eventually
some steady current should result. I might think that this is the way
electrical resistance appears. It is like driving a car through a forest
without roads. Progress would be somewhat slow because of
collisions with the trees.

Nothing about this picture should change if I cool the wire to a lower
temperature, and I would therefore expect the resistance to be
unchanged. But, as fig. X-2 shows, this is not what happens: the
resistance drops dramatically as the temperature drops.

The resistance of a pure metal at a few kelvin may be a million times

smaller than that at 300 kelvin. It is quite clear that there is
something wrong with the picture of particle-like electrons bouncing
off ions to give rise to electrical resistance. What is wrong is that I
have not used quantum mechanics to describe the electrons. I have
not taken account of their wavelike nature. A better analogy than the
car in the forest would be a person trying to call out to someone else
in the forest; the trees do not greatly impede the sound waves.

We saw in chapter VI that electrons in crystalline solids are

represented by Bloch waves. As shown in fig. VI-10, it is a wave,
with some additional wiggles at the position of each ion. It is the
appearance of these wiggles, and not scattering, that is the result of
the interaction between the charge of the electron and charges of the
ions.

Once we take Bloch waves to represent the electrons, we can forget

about the ions. The Bloch waves can be thought of as being in
empty space, and there is nothing to hinder the motion of the
electrons they represent. Then what else could it be which scatters
the electrons, thus preventing a runaway current when a voltage is
applied?

I said that we could forget about the ions. Well, not quite: it is only
the static ions that we can forget. There are still the vibrations of the
ions about their mean positions.

It is the kinetic energy of these vibrations which constitutes the

thermal energy of the lattice and gives the solid its temperature. We
saw in chapter VI that the quantization of these vibrations gives us
phonons. So the Bloch waves are not in empty space, but in a space
filled with these particles called phonons, all moving about at the
speed of sound 192

SCATTERING OF ELECTRONS

in the solid. It is the scattering of the electrons when they collide with
the phonons which limits the electric current, and leads to a
resistance.

How does the electron know that there is a phonon? A phonon is a

quantum of a sound wave, which is a wave of alternating
compression and expansion of the lattice of ions. The conduction
electrons keep in step with these oscillations of the ionic lattice, but
not perfectly. The result is that the compressed (expanded) parts of
the sound wave have a net positive (negative) charge. The sound
wave is thus accompanied by a superposed wave of charge
oscillating between positive and negative values. This charge wave
influences the conduction electrons and causes them to change from
one quantum state to another. This is the mechanism of scattering of
electrons by phonons.

Electrons are fermions, and this means that each quantum state with
a specified k vector can have no more than two electrons with
opposite spins. I described in chapter VI (p. 118) how this leads to
the picture of a Fermi surface inside which lie the wave vectors of
these electrons. Remembering that the momentum is equal to the
wave vector multiplied by the Planck constant, and also to the
velocity multiplied by the mass, I can think of the Fermi surface as
giving me a picture of the velocities of the electrons. Figure X-3

shows what happens to a Fermi surface, which I have assumed to

be spherical so as not FIGURE x-3. The effect of an electric

current on the Fermi surface. The

parts A and B together represent the

Fermi surface when there is no

electric current. In the presence of a

current, the Fermi surface is bodily

displaced by the vector v and is now

represented by the parts B and C

taken together.

states occupied without

current, empty with current

states occupied with or

without current

states occupied with current,

empty without current

X ELECTRIC BULBS AND INSULATED CABLES

to complicate the picture, when a current flows. What follows applies

to any other shape of Fermi surface also. The parts labelled A and B
in fig. X-3 taken together form a sphere with its centre at the point O.
The velocity vectors for all the electrons lie within this sphere, which
shows the state of affairs when there is no current flowing in the
metal.

Although the electrons are moving about in all possible directions

with velocities from zero up to a maximum, there is no net current.
For each electron moving in a given direction with a given speed,
there is another one moving in exactly the opposite direction with the
same speed, and the two add up to give zero current.

Now I apply a voltage and thus exert a force on each electron,

causing it to get an added velocity in the direction of the force. It is
scattered after a certain time which I denote by the Greek letter X
(tau), at which instant each electron has had its velocity increased by
an amount v. The electron whose initial velocity is shown by the
vector labelled 1 in the figure, for example, now has the velocity
vector labelled 2. At this instant the electron is scattered by a phonon
and changes its wave vector to the one labelled 3, say. If you
imagine this process occurring with each of the velocity vectors, you
can see that the net effect is to displace the whole sphere to the right
by the vector v, so that it is now comprised of the portions labelled B
and C. But it is not a static picture: application of a voltage moves
the sphere a certain amount and no further, because electrons are
constantly disappearing, because of scattering, from quantum states
at the leading edge in region C and reappearing in states at the
trailing edge in region B.

The current is the charge passing a given point per second, and is
therefore proportional to v, which is called the drift velocity. The time
X which elapses before the electron is scattered is called the
scattering time. A longer scattering time means a larger drift velocity
and therefore a larger current too. Furthermore, the current is also
proportional to the number per unit volume (called the density) of
electrons which are there to carry it. So we should look to the
scattering time and the density of electrons for an explanation of the
different behaviours of the electrical resistance of metals, alloys and
semiconductors shown in fig. X-2.

We take the case of a pure metal first. When its temperature goes
up, so does the amount of thermal energy in it. This energy consists
of the vibrations of its atoms about their average positions, and this
motion is quantized into phonons. The number of phonons increases
with an increase in temperature, and so the frequency of collisions
between electrons and phonons also increases. This means that the
scattering time (which is the time between collisions) and therefore
the conductivity decrease as the temperature rises. In other words,
the electrical resistance should decrease with decreasing
temperature in a metal, and this is what happens. The electron
density may vary from metal to metal, but does not change with
temperature in a given metal. We thus have a basic explanation of
the different resistivities of different metals, and how they change
with temperature.

In an alloy, we have more than one kind of atom in the lattice: copper
and zinc atoms in brass, for example. We saw in chapter VIII (p.
155) that this leads to an additional scattering of the electrons, due
to the loss of translational symmetry. This in turn 194

SCATTERING OF ELECTRONS

causes an electrical resistance which is added to that due to the

scattering of electrons by phonons. Furthermore, this resistance will
not change with temperature, because the two types of atoms do not
change their positions. This is the behaviour of the resistance of an
alloy that is actually observed, as shown in fig. X-2.
Finally we come to the electrical resistance of a pure semiconductor
like silicon or germanium. There are two striking differences from the
resistance of metals: it is very large, and gets even larger as the
temperature drops. We again look to the two quantities, the
scattering time and the number of electrons per unit volume, for an
explanation. The scattering time is mainly determined by the number
of phonons, which in turn is determined by the temperature of the
material. We would thus not expect spectacular differences in
scattering times between metals and semiconductors.

The situation however is quite different with the number of electrons

available to carry the current.

A semiconductor is essentially an insulator: the energy levels for the

electrons are distributed as shown in fig. X-4. There is a band of
levels, called the valence band. Above this band is a range of
energy, called the energy gap: there are no quantum states for
electrons in this gap. Above the gap there is again a band of
possible energies, called the conduction band. At very low
temperatures, almost every quantum state in the valence band is
occupied by an electron, and the conduction band has very few
electrons. An applied voltage produces negligible current, and thus
we have an insulator.

When the temperature is raised, some electrons in the valence band

acquire enough additional energy in the form of thermal energy to
put them into previously unoccupied states in the conduction band.
This additional energy must be equal at least to the energy gap,
because there are no states for the electrons in the energy gap itself.
An equal number of empty states is left behind in the valence band,
which becomes equivalent to a band of holes. The number of
electrons and holes increases with increasing temperature. This
means that a given applied voltage will produce more current the
higher the temperature, because there are more electrons and holes
to carry it.

Therefore the resistivity of a semiconductor should decrease as the

temperature goes up. This is what happens, as can be seen in fig. X-
2.

How about the scattering time, which also influences the resistivity of
the semiconductor? The scattering is still by phonons, and is
stronger at higher temperatures just as with metals. But its effect in
changing the resistivity is weak in comparison with the effect of the
change in number of electrons and holes, and the net result is a drop
in resistivity with increasing temperature, as is actually observed.

The addition of small amounts of other elements to a semiconductor

- phosphorus or aluminium to silicon, for example — will result in
electrons in the conduction band or empty states in the valence band
(equivalent to a band of holes) being present even at the lowest
temperatures, as explained in chapter VI (page 121). There will also
be thermal excitation of electrons and holes as in a pure
semiconductor, and so the variation of resistivity of such doped
semiconductors will be somewhere between that of metals and pure
semiconductors.

X ELECTRIC BULBS AND INSULATED CABLES

FIGURE X-4. The valence and COld llOt

conduction bands of a semiconductor

at low and high temperatures. More

electrons in the conduction band and

empty states in the valence band

(equivalent to a band of holes) are

present when the semiconductor is

hot than when it is cold.

3

energy gap .B energy gap

conduction band

valence band

• occupied state

O empty state

An insulator, such as the material which covers the copper wire used
in households for carrying electric currents, has essentially the same
band structure as a semiconductor, but with one important qualitative
difference. The energy gap between the filled band and empty band
above it is so large, that at ordinary temperatures there are very few
electrons with enough thermal energy to raise them from the filled
band to the empty band. This means that negligible current is
produced when a voltage is applied: we have an insulator.

196

SEMICONDUCTORS AT WORK

We thus see that the main features of the electrical resistivity of

metals, semiconductors and insulators can be understood in terms of
the quantum energy levels of electrons and the quanta of thermal
energy, namely the phonons. As noted earlier, the resistivities of
different materials have the largest range of values of any of their
physical properties. Nevertheless, the same basic quantum
mechanical description applied to all these materials gives us an
explanation of this tremendous variability.

Quantum mechanics is not something which becomes important only

when one is thinking of single electrons and atoms. It is essential
also for the understanding of effects on an everyday scale, such as
the flow of electric current through a wire.
5 Semiconductors at work
I suppose that everyone has heard of a transistor, though not all may
be quite clear what it is. A small portable radio is often called a
transistor, but this is not the meaning which was given to the term
when it was invented. I shall use it here with its original meaning, to
refer to an object made of a semiconductor so as to have certain
desired electrical properties. All the electronic gadgets that surround
us at home, at work and everywhere else contain such objects,
ranging in number from a few to many millions.

Examples are radios, television sets, video recorders, pocket

calculators, automobiles, washing machines, computers, and on and
on: one could make a very long list.

In order to give an example of what a transistor does, I consider a

radio. It receives from the broadcasting station an electromagnetic
wave which produces a very feeble varying electric current in the
aerial. The radio converts this weak signal into a much larger current
that flows through the loudspeaker and reproduces the broadcast.

Transistors are the key element in this amplification of the signal. It is

basically this function, of converting an incoming signal into a
different outgoing signal, which is performed by the transistor.

Transistors come in many different forms depending on what they

are intended for.

I take the example of a type of transistor which the professionals call

a MOSFET, an acronym for metal-oxide-jemiconductor-/leld-
effect-/Tansistor. There are a million or more of these or similar
transistors on a piece of silicon a centimetre or so across, at the
heart of a computer. This is an example of the ubiquitous chip.

Figure X-5 is a sketch of a MOSFET. A section of p-type silicon is

sandwiched between two n-type sections called source and drain
respectively. A thin film of silicon oxide is sandwiched between the p-
type silicon and a metal film which is called the gate. A voltage
between the source and the drain should in general cause a current
of electrons to flow from one to the other, since the source, being n-
type, can supply electrons. The size of the current depends on how
many electrons are available in the material between the source and
the drain to carry it. This number can be changed by varying a
voltage which is applied to the gate. Variations of this voltage (which
could be the incoming signal in a radio, for example) are reproduced
as variations of the 197

X ELECTRIC BULBS AND INSULATED CABLES

current between the source and the drain, and there we have the
basic functioning of a transistor. There is a similarity to a rubber hose
through which water is flowing at a certain rate because of a
pressure difference between its ends. This flow rate can be varied by
pinching the hose at some point so that its cross-section changes
there.

Other types of semiconductor devices are also in use, designed to fit

particular applications. Fundamental to the working of all of them are
two characteristics which are peculiar to semiconductors. The first is
the special arrangement of the quantum energy levels into a valence
band and a conduction band separated by an energy gap. The
second is the ability to control the number and distribution in space
and in energy of the electrons and holes by introducing junctions, by
alloying, and by applying voltages.
6 Electric currents at work
The use of electricity is pervasive in the way of life today. In addition
to the ever-increasing role of semiconductor devices, we use electric
currents to produce motion, heat, light, and magnetism, to mention
just a few of the applications. I describe now the basic effects of
electric currents which make such applications possible.

The possibility of deriving motion from an electric current, as for

example in an electric motor, depends on the fact that a wire carrying
a current feels a force on it when it is placed in a magnetic field. I
illustrate this in fig. X-6, which shows a current-carrying wire in the
magnetic field between the north and south poles of a suitably
shaped magnet. The wire feels a push in a direction which is
perpendicular to the directions of the current and the magnetic field.
Reversing the direction of either the current or the field reverses the
direction of the push. By suitably arranging magnets and coils of wire
and the way the current is fed to the coils, one ends up with a motor.

When an electric heater is switched on, current flows through it, and
heat is given off. Suppose the current is amperes, and the voltage is
V volts. A current ofamperes means that / coulombs of charge, in the
form of electrons, are flowing through the heater per second. The
potential energy of the electrons is just the charge multiplied by
oxide

FIGURE X-5. A MOSFET. Changing

the voltage on the gate changes the

number of electrons available to carry

a current in the part marked p, and

this causes the current from the source

to the drain to be changed

accordingly.

source

drain

198

ELECTRIC CURRENTS AT WORK

the voltage. Since the voltage along the heater drops by F volts, the
potential energy of the electrons leaving the heater per second is 7x
V joules less than that of the electrons entering the heater. This
energy which is lost by the electrons in their passage through the
heater appears as thermal energy, at a rate of IX Fjoules per second,
or /x Fwatts.

Now how exactly is the energy of the electrons converted to thermal

energy of the heater coils? We saw that when a current is flowing in
a resistance such as a heater coil, electrons are being constantly
speeded up and then scattered back to states of lower energy, and
the difference in energy appears as additional phonons. More
phonons in FIGURE x-6. Current-carrying wire

in a magnetic field. The force is

directed into the paper for the

directions of current and magnetic

field shown. Reversing either of these

directions reverses the direction of the

force also.
current—

199

X ELECTRIC BULBS AND INSULATED CABLES

the wire means a higher temperature.

In an electric heater, the conditions are such that the energy

produced as phonons is enough to heat the coils to a dull red heat.
In an incandescent electric bulb, on the other hand, the current
flowing through the filament causes it to become white hot, thus
giving ofTlight in addition to heat. Only a fraction of the electric
energy supplied to a bulb appears as visible light; the rest just heats
up the bulb and the outside. Fluorescent lights operate on a different
principle. There, the kinetic energy of the electrons is absorbed by
certain atoms which are therefore excited to higher energy states.
The atoms then return to a lower energy state giving off the
difference in energy as a photon of light. Such lights convert
electrical energy to photon energy more efficiently than do electric
bulbs.

We have seen that an electric current flowing in a wire produces a

magnetic field around it, as can be seen by its effect on a compass
needle. This property of a current is used to produce magnets
consisting of a cylindrical coil of wire wrapped around an iron core
and carrying a current. I shall discuss the magnetic properties of iron
and some other materials in the next chapter.
7 Summary
The chapter is concerned with how an electric current is carried by
metals and semiconductors, and why insulators cannot carry a
current. I make an analogy in section 1

between the flow of electrons and the flow of water. I introduce the
units used in measuring electric currents.

I describe in section 2 the property of a material called its electrical

resistance, which is a measure of the ability to carry a current when
a voltage is applied; the greater the resistance, the smaller the
current for a given voltage. The resistance depends on the size and
shape of the material through which the current is flowing. We define
the resistivity of the material as the resistance of a one-cm cube of
the material between a pair of opposite faces. One can then
meaningfully compare the resistivities of different materials.

I note in section 3 that the resistivities of different materials vary over

a tremendous range, the highest resistivities being more than 1018
times the lowest. The resistivity changes with temperature, and the
variation is qualitatively different between metals and
semiconductors. Alloying a metal increases its resistivity, and
alloying a semiconductor changes its resistivity in a complicated
manner.

In section 4,1 describe the electrical current using the quantum

mechanical picture of a solid. I find that the resistance in a pure
metal or a semiconductor is produced by the scattering of electrons
by phonons. The difference in the magnitude, and variation with
temperature, of resistance in metals and in semiconductors is a
consequence of their different energy band structures. I get a picture
of how alloying changes the resistance.

2OO
SUMMARY

I already described a solar battery, which is one practical application

of the quantum mechanical picture of a semiconductor, in chapter IX.
I give in section 5 a description of another example, namely a
transistor. This is a basic unit in devices called integrated circuits (or
chips) used in computers and many other things in everyday use.

I describe briefly in section 6 some practical applications of electric

currents: to produce heat, motion, and magnetic fields. Electric
currents occupy a central place in our daily lives.

We have so far looked at the thermal, optical and electrical

properties of matter. We have seen how they follow from the atomic
structure of matter as described by quantum mechanics. The two
properties of electrons and nuclei which play a role in these
properties are their mass and electric charge. We have not taken
any notice of a third property that electrons, as well as most nuclei,
have, namely that they are magnetic. They behave like permanent
magnets which can be influenced by a magnetic field, just as their
charge causes them to be affected by an electric field. We shall look
in chapter XI at some consequences of this property.

201

XI

MAGNETS

l What is a magnet?

Magnets are a part of many everyday objects, although one is not

always aware of their presence. You might be using one to hold your
shopping list on the side of the refrigerator in the kitchen. The motor
in the refrigerator contains a magnet, and its door is held shut by a
magnet. Most household electronic equipment (radios, television
sets and the like) contain magnets. Electric motors work because of
the magnets in them. Magnets occupy a not insignificant place in the
things used in our daily life.

I illustrate the basic properties of a magnet in fig. XI-1. When a

magnet is mounted so that it is free to rotate, one end always turns
towards the north: I have shaded this end in the diagram and
labelled it N. I shall refer to this as the north pole and the other as
the south pole of the magnet. The figure is divided into four parts
labelled (d) to (d), and each illustrates a property of a magnet:

(a) When I bring two magnets together so that the north pole of one
approaches the FIGURE XI-1. Basic features of a

magnet. The north pole is marked N

and shaded to identify it. (a) Unlike

poles attract each other, (b) Like poles

repel each other, (c) Either pole

attracts a rod of iron, which then itself

becomes a magnet capable of

attracting a second bar of iron, which

then itself becomes a magnet, and so

on. The same happens if the rod is of

nickel or cobalt or certain other

materials, (d) The vast majority of

materials, whether metals,

semiconductors or insulators, feel

practically no force from a magnet.

(a)

attract IN!

magnet

(P)

repel

(c)

JED attract1
1 attractc
iron
iron

(d)

nothing C

glass

202

ELEMENTARY MAGNETS

south pole of the other, the magnets attract each other: there is an
attractive force between unlike poles.

(b) If two poles of the same kind approach each other, there is a
force of repulsion between them.

(c) A rod of iron approaching either pole of a magnet is attracted

towards it.

Furthermore, the iron itself behaves like a magnet so long as it is

close to the original magnet, and will attract another piece of iron.
This induced magnetism of the iron vanishes when it is moved far
enough away from the magnet. The magnet itself carries its
magnetism all the time, which justifies calling it a permanent magnet
to distinguish it from the iron rod. I get the same induced magnetism
if I substitute cobalt, nickel, or certain alloys of these elements, for
iron. Such materials are said to be ferromagnetic, and the
phenomenon itself is ferromagnetism. If I heat the iron rod, I find that
its ferromagnetism becomes weaker as the rod gets hotter, and
finally vanishes when the rod is at 770 °C; the magnet exerts no
force on the rod when it is above this temperature. The same effect
is seen with other ferromagnetic materials, but at other
temperatures.

(d) If the rod is of almost any other material (copper, glass, plastic,
etc.), it feels a negligible force from the magnet. Such materials are
not ferromagnetic.

We take the properties (a) and (b) as given by nature, rather like the
forces between electric charges. In fact, it is tempting to push the
analogy and to imagine that there are separate north and south
poles, which could be called magnetic monopoles. The analogy here
is with electrons and positrons which can be called electric
monopoles. A magnet would then be a concentration of north poles
at one end and south poles at the other end of the bar. However,
many experimenters have tried without success to see if such
magnetic monopoles actually exist. Our experience so far is that any
magnet, whether of the kitchen variety or an electron or proton or
neutron, always comes with both a north pole and a south pole. The
two are inseparable: breaking a magnet in two just gives two smaller
magnets, and not separated north and south poles. So it seems that
an atomic description of magnetic materials must be based on the
magnetism of the constituent particles as an immutable fact of
nature, like for example the charge of the electron.
2 Elementary magnets
All matter is made up of electrons and nuclei, and just as with the
other properties, we want to understand the magnetic properties in
terms of these fundamental constituents of matter. I have said earlier
that an electron, and most nuclei, behave as if they were spinning
tops as well as permanent magnets. I go into this now in a little more
detail.

I begin with the electron. It is a magnet, but a very weak one indeed,
compared to an ordinary permanent magnet. I would need
something like 1021 of these electron-magnets all pointing the same
way and crowded into a volume equal to that of an ordinary bar
magnet, in order to produce the same magnetic field as the magnet.

203

XI MAGNETS

question Why must the electron-magnets all point the same way for
this effect to happen?

answer Otherwise, the magnetic fields produced at any point by the

individual electrons would be pointing in all directions, and the net
field at that point will be practically zero.

I measure the strength of a magnet using a unit called a Bohr

magneton, named for the physicist Niels Bohr. The value of the Bohr
magneton is fixed by the charge and mass of the electron and the
Planck constant. It is therefore also one of the fundamental
constants of nature. The electron-magnet is 1.001159652 Bohr
magneton in strength, and also has a specific direction, namely from
its south pole to its north pole. So I can represent the magnetism of
the electron (or any other magnet for that matter) by a vector with
magnitude and direction, and give it a name: magnetic moment.

A brief digression: you may have wondered at the extraordinary

precision (one part in a thousand million) of the value quoted above
for the electron-magnet’s strength, its magnetic moment. It is the
value found not only experimentally, but is also the value calculated
by the combined theories of quantum mechanics, relativity and
electromagnetism. It is a tribute to the extraordinary accuracy of
some experimental work in physics, and a convincing proof that the
physics used to calculate the value is correct.

The electron magnetic moments would tend to line up just like

compass needles if I were to put them in a magnetic field. This
tendency is opposed by their thermal energy which wants to make
them point in all directions. The result at ordinary temperatures is
that only a small fraction of the moments line up with the field. If the
field is removed, even these moments become disordered. In a
ferromagnet, the electron-magnets are in zero field.

Nevertheless, they do line up and point in the same direction. This is

the distinctive mark of a ferromagnetic substance and we shall see
how it comes about later in this chapter.

The electron is also like a spinning top: it has a permanent angular

momentum which is one of the constants of nature. The physicist
Paul Dirac combined quantum mechanics and Einstein’s relativity
theory for the electron, and found that angular momentum and
magnetic moment are like the two faces of a coin: one cannot have
the one without the other. The quantization of angular momentum
measured in a specified direction implies that the magnetic moment
in that direction is also quantized. For the rest of this chapter, I shall
talk only about magnetic moments, for we are concerned here only
with magnetic properties.

The proton and the neutron are also magnets, but much weaker than
the electron. It would take a couple of thousand of them to equal the
magnetic moment of an electron.
An atomic nucleus is made up of protons and neutrons, and so in
general would also have a magnetic moment which is much smaller
than a Bohr magneton. We shall see later that even these small
moments are important in some cases.

question Is it possible for some nuclei to have zero magnetic

moment?

answer Yes. Suppose the numbers and the orientations of the

proton-and neutron-magnets in some nucleus are such that the
magnetic effects are 204

FERROMAGNETS

cancelled pairwise. This will happen, for example, if all the protons
and neutrons are separately paired up with the spins in each pair
pointing in opposite directions, thus: t J-. Then the total magnetic
moment of the nucleus will be zero.

I said that about 1021 electrons crowded together with all the
moments pointing the same way would have the strength of an
ordinary magnet, say a compass needle. This number is in fact
comparable to the number of electrons in a piece of solid of that size.
This spontaneous lining up of the electron moments takes place only
in a few materials, among them iron, nickel and cobalt. Furthermore,
we do not have to worry about the magnetism of the nuclei in this
case, since it is thousands of times weaker than that of the electrons.
How does this lining up of the electron moments take place? The
answer is in the next section.
3 Ferromagnets
In an atom, most or all of the electron moments cancel out one
another in pairs, with the two moments in each pair pointing in
opposite directions. Then there may remain a few electrons whose
moments have not got so cancelled out. Atoms with such unpaired
electrons are potential candidates for forming ferromagnets. We
know that there are forces of attraction and repulsion between
magnets. So it is tempting to think that somehow these magnetic
forces between the unpaired electrons on neighbouring atoms may
lead to all the moments lining up, giving a ferromagnet. But when
these forces are worked out in detail, they turn out to be too weak to
give rise to the known ferromagnets like iron and nickel. We must
look elsewhere for the origin of this property.

What we have ignored so far is the wave aspect of the electrons,

and now we include this explicitly, as first done by the physicist
Werner Heisenberg. Consider two neighbouring atoms in the solid,
and suppose that each has one unpaired electron. Then quantum
mechanics tells us that there is a wave function for the two electrons
which gives the probability of finding each electron in any particular
region in the two atoms. The quantum states (and the wave
functions) for the two cases, of the two electron moments pointing in
the same direction (parallel moments) or in opposite directions
(antiparallel moments), are different. This leads to a difference in the
potential energy due to the electric force between the two electrons
in the two cases. What prevails in nature is the quantum state of
lower energy. For iron and the other materials which are
ferromagnetic, this lower energy obtains for parallel moments, while
for all other materials it is the one with antiparallel moments. It is
noteworthy that while ferromagnetism requires the magnetism of the
electrons, it is not the magnetic force between them which leads to
ferromagnetism, but rather the electric force and the wavelike nature
of the electrons.
Now one might ask, if the moments in iron all like to line up in the
same direction, then why is every piece of iron not already a
magnet? Why does one have to bring up a normal magnet to the iron
in order to make it behave like one? How is it that a perma-205

XI MAGNETS

nent magnet comes already equipped with its magnetism, unlike the
piece of iron? I shall answer these questions next.

The atoms in an iron crystal sit in a body-centred cubic structure,

whose unit cell is shown in fig. 11-13, part (4) (p. 28). The distance
between neighbouring atoms is different in different directions, and
one expects that the interaction between electrons on neighbouring
atoms will also depend upon the direction. All this taken together
leads to the result that the magnetic moments will point more easily
along certain directions in the crystal, called the easy directions, than
along others. In a lump of iron there are many very small crystals
called grains which are all stuck together. These crystals point in all
possible directions, and in each of them the moments point in one of
the easy directions. So each grain is a magnet, but the different little
grain-magnets point in different directions, and therefore the lump of
iron itself has a total magnetism which is practically zero. I illustrate
this in fig. XI-2, showing an enlarged view of the grains, each a
magnet pointing a different way.

One might think that one would get around this problem by using a
piece of iron which is entirely one crystal. There being no grains in it,
the piece should have the electron moments all pointing in one easy
direction, and therefore become a magnet. But this does not happen.
The easy direction for the moments to point in iron is along an edge
of the unit cube in the crystal: but there are six such directions in a
cube, as you can see in fig. XI-3, part (a). The crystal gets divided
into regions, and the moments in each of these regions will point in
one of the easy directions, but they will be different in the different
regions. Each of these regions is called a magnetic domain. I show
in fig. XI-3, part (b), a possible arrangement of domains in a crystal
shaped like a rectangular bar. The final result is the same as with the
lump of iron: the single crystal bar will not show strong magnetism,
because the effects of the different domains cancel one another out.

FIGURE XI-2. A microscopic view of

a lump of iron showing its grains,

each with its spins pointing in the

direction of the arrow. These

directions being different in the

different grains, the lump as a whole

shows no magnetism.

206

XI MAGNETS

magnetic particle prefers to have just one domain rather than

several, because then its energy is lower. A permanent magnet
contains such single-domain ferromagnetic particles. These particles
are suspended in a substance which itself need not be
ferromagnetic. During the manufacturing process, there is a stage
when the particles are free to rotate. This is when a magnetic field is
turned on, causing the particles to rotate so that the magnetism in
each particle is pointing in the direction of the field. As the process
continues, the substance surrounding the particles hardens and they
become frozen in their new orientations. The result is a permanent
magnet, a microscopic view of which is shown in fig. XI-4. Another
process starts with the particles alone, without the surrounding
substance. They are aligned in a magnetic field as before, and then
heated to a high enough temperature so that they stick to one
another and are no longer free to move. This way of consolidating
particles is called sintering.

And now to the bar of iron which becomes a magnet when it is

brought near a permanent magnet. What happens here is that, if the
magnetic field of the permanent magnet is strong enough, the
magnetism in each of the domains of the iron will rotate to an easy
direction which is closest to the direction of that field. The bar then
behaves like a magnet. When the permanent magnet is removed,
the domains return to their random orientations, and the magnetism
of the bar as a whole is lost. I am reminded of a classroom in which,
before the teacher comes, the pupils are clustered in small groups
with each group talking about a different topic. The teacher enters,
and all eyes and attention turn to her, and there is only one topic,
hers. When she leaves, the initial condition is, at least usually,
restored.

One can also use the magnetic field produced by an electric current
flowing through a spool of copper wire wrapped around a rod of iron
to make it magnetic, and then one has an electromagnet. This is the
kind of magnet used in many applications where magnetic fields are
needed.

Finally, I consider how the magnetism of iron and other

ferromagnetic materials changes with temperature. The
ferromagnetism becomes weaker as the temperature of *

**«

®$*$$$

FIGURE XI-4. A microscopic view of <S9 /$>

the magnetic particles in a permanent <$> w ^) S)

magnet. Each particle is a single ^ ^

domain, and the magnetism, shown ^* w $

by an arrow, points in the same ($? <$)

direction for each domain.

208

FERROMAGNETS

Table XI-1. Curie temperatures (in kelvin)

material

Curie temperature (K)

iron

nickel

cobalt

gadolinium

dysprosium

1043

631

1403

289

105

the material is raised, and finally vanishes at a well-defined

temperature which is called the Curie temperature after the physicist
Pierre Curie. Table XI-1 shows the Curie temperatures of several
ferromagnetic substances. Note that the temperatures are in kelvin,
so that 0 °C is 273 K. The Curie temperature varies from well below
room temperature to very high temperatures, depending on the
material.

Figure XI-5 shows how the strength of the magnetism varies as the
temperature changes from near the absolute zero to the Curie
temperature. It has the highest value at the lowest temperature,
gradually drops as the temperature is raised, and becomes zero at
the Curie temperature.

question Can this effect be because the electron gradually loses its
magnetic moment as the temperature goes up?

answer No. The electron’s magnetic moment, like its mass or

charge, is an intrinsic property which does not change with
temperature.

FIGURE XI-5. Variation of magnetic

strength with temperature for a

ferromagnetic substance.

zero

temperature

Curie

209

XI MAGNETS

The reason for the decrease in the magnetic strength is that with
rising temperature more and more of the electron moments begin to
point in the opposite direction, until at the Curie temperature there
are as many moments pointing one way as there are the opposite
way, and the net magnetic strength then becomes zero. This loss of
ferromagnetism on going through the Curie temperature is an
example of what is called a phase transformation. One gets the
ferromagnetic phase below the Curie temperature, and it transforms
into a non-magnetic phase on being heated above that temperature.

There are other phase transformations in nature, and trying to

understand how they come about has been, and continues to be,
one of the challenging problems in physics.

I shall introduce you to this topic in the next section.

4 Phase transformations
I describe a few examples of phase transformations, from which we
shall be able to see some essential common features. There are two
examples from our daily experience: the melting of ice to water at 0
°C, and the boiling of water to steam at 100 °C. In each case, the
change occurs at a sharply defined temperature. The atomic
arrangements are totally different in the phases below and above the
transformation temperature. Consider ice: even if the temperature is
just a little below 0 °C, it is all ice, and just a little above, all water.
The change does not take place one molecule at a time, so to speak,
but all at once. Ice is a crystal: there is order in the arrangement of
the molecules. This order suddenly disappears when ice melts to
water. I can talk about order in a ferromagnet also, measured by the
net fraction of moments pointing in one direction. This order does not
disappear abruptly at the Curie temperature, but decreases smoothly
and finally vanishes at that temperature. I list some common features
of these transformations: 1. There is a sharply defined
transformation temperature, which I denote by To.

2. There is more order in the phase below To than in the one above
7^,. The actual type of order can be different in different cases:
crystalline order in melting, the ordered orientation of electron
magnetic moments in ferromagnetism.

3. The change in this order in going through 7^ can be smooth, as in

a ferromagnet, or jumpy, as in ice.

4. The change affects all the atoms (or electrons or molecules, as

the case may be). Ice does not melt one molecule at a time.

The physical properties of the two phases are dramatically different,

and yet all it takes to go from one to the other is ever so small a
change in temperature. So you can see that there must be subtle
features of the statistical physics of the constituent particles (the
molecules in ice, the electrons and atoms in a ferromagnet) which
must be carefully considered in understanding such phase
transformations.

I shall devote chapter XII to a famous example of a phase

transformation, namely superconductivity. It is one of the best
understood transformations, and provides us a 210

NUCLEAR MAGNETIC RESONANCE

splendid opportunity to put together the different strands of the

quantum mechanical description of solids which we have learnt so
far.
5 Nuclear magnetic resonance
It was all right to neglect the magnetism of the atomic nuclei in
considering ferromagnetism, because the magnetism of the
electrons, which is the source of ferromagnetism, is thousands of
times stronger. However, there is another difference between the two
which gives nuclear magnetism some features absent in electron
magnetism. Magnetically considered, all electrons are identical, but
all nuclei are not.

Different nuclei have different magnetic moments and spins, as table

XI-2 shows.

Table XI-2. The magnetic moment of various nuclei

nucleus magnetic moment spin (S)

(isotope number) (in nuclear magnetons)

hydrogen(l) +2.792743 i

carbon(13) +0.702381 i

oxygen(17) -1.89370 \

phosphorus(31) +1.13162 \

copper(63) +2.22664 \

The isotope number shown in parenthesis is the total number of

protons and neutrons in the nucleus. The nuclear magneton is a unit
which is smaller than the Bohr magneton by a factor of about 5.4 X
10”4. The magnetic moments differ in magnitude for different nuclei.
They can be positive or negative, which indicates whether the
nuclear magnetic moment points in the same direction as the
angular momentum vector or in the opposite direction.
The angular momentum of a nucleus is specified by the spin
quantum number S, whose values are also shown in the table. When
I put the nucleus in a magnetic field, its energy changes owing to the
action of the field on the magnetic moment. Because of quantum
mechanics, we are not surprised to find that this change is
quantized. The energy of the nucleus can only take one of a certain
number of discrete values, and nothing in between. Even the number
of these allowed quantum energy levels is not arbitrary: it is equal to
2S + 1, where S is the spin quantum number. This is because there
are only 2S +1 specific directions in which the magnetic moment can
point with respect to the field. We have here one more example of
how the quantum nature of matter leads to results which are quite
different from our everyday experience. It is as if a compass needle,
instead of pointing north, could only point, say, either northeast or
southeast, and in no other direction.

There are thus two levels in hydrogen (2S + 1 = 2 if S = 2), six in

oxygen, and so 211

XI MAGNETS

on. You see here one consequence of the intimate connection

between magnetic moment and angular momentum. For a given
nucleus, the energy difference is the same between each pair of
adjacent levels, and is directly proportional to the strength of the
magnetic field. This energy difference also depends on the magnetic
moment of the nucleus, and is therefore different for different
materials, even when the magnetic field is the same. I illustrate these
features in the energy level diagrams in figs. XI-6 and XI-7.

Suppose I put some hydrogen in a magnetic field, so that each

nucleus has two energy levels differing by an amount of energy E. I
now shine photons of frequency/

on the hydrogen. Each of these photons has energy hf, where h is

the Planck constant.
A nucleus which is in its lower energy level can absorb a photon and
move to its higher energy level only if the photon energy is exactly
equal to the difference in energy between the two levels:

E=hf.

If this condition is not satisfied, the photon cannot be absorbed and

will just move right through the material. The absorption, as the
frequency of the photons is varied, is shown in part (a) of fig. XI-8.
There will be no absorption except at the frequency/ we speak of an
absorption peak at that frequency.

Now suppose I take some material which contains several kinds of

atoms, and place it in a magnetic field. Because of the different
nuclear magnetic moments, each set of field

FIGURE XI-6. Effect of a magnetic

field on a nucleus of spin 5. Its

magnetic moment can only point in

one of two directions labelled 1 and

2, but not in any other direction, as

shown in part (a) of the figure. Part (b)

shows the quantum energy levels for

these two orientations. The difference

in energy between these levels

increases from E to 2£ when the field

is doubled.

(a)
zero field

some field

S= 111

“Z

field doubled

212

NUCLEAR MAGNETIC RESONANCE

FIGURE XI-7. The energy levels in the

same magnetic field of the nuclei of

hydrogen (which is just a proton),

carbon, and oxygen. The spin

quantum numbers for these nuclei are

respectively 1,2, and |. Note that the

energy differences between levels are

different for different nuclei.

oxygen

hydrogen
carbon

FIGURE XI-8. Nuclear magnetic

resonance. Part (a) illustrates the sharp

rise in absorption of energy from an

electromagnetic wave at a certain

frequency, by nuclei placed in a given

magnetic field. Part (b) shows the case

of two different kinds of nuclei,

hydrogen and carbon, say, where the

frequency is constant and the

magnetic field is varied. Each peak in

the absorption arises from one of the

two types of nucleus.

/ frequency

-4—»

(b)

magnetic field
213

XI MAGNETS

identical nuclei will set up its own distinctive group of energy levels,
different from that of any other nucleus. If I now shine photons of a
single frequency on the material and increase the magnetic field
gradually, I get a strong absorption from each type of nucleus at that
value of the field for which its energy spacing is equal to the photon
energy. I show in fig. XI-8, part (b), what the absorption might look
like for a mixture of two different types of atoms at fixed photon
frequency and changing magnetic field.

In practice it is much easier to keep the photon frequency constant

and to vary the magnetic field, and so this is what is usually done.

This absorption of photons by nuclei placed in a magnetic field is

called nuclear magnetic resonance. It is at the heart of magnetic
resonance imaging (MRI), a technique used in medical diagnostics.
Since each distinct type of nucleus carries its own signature in the
form of a resonance signal, it is possible to make a picture of the
inside of the human body point by point, as it were, without the
unwelcome necessity of invasive procedures.

The electron of course also has a magnetic moment, and so one

would expect to see similar resonance effects due to electrons. They
are indeed observed, and the phenomenon is called electron spin
resonance. Both types of resonance find many applications in the
study of materials, living as well as inanimate.
6 Weak magnetism
I have considered so far only ferromagnetic materials; but these form
only a very small minority of all the materials we have. All the others
have electrons too: the electrons in shells around the nuclei, and in
addition electrons which are free to move about in the case of
metals. The electrons in shells have in general an angular
momentum, and therefore a magnetic moment too, which is in
addition to the intrinsic magnetic moment of the isolated electron. A
free electron in a metal, when it sees a magnetic field, feels a force
which is always perpendicular to both its direction of motion and the
direction of the magnetic field. The net result is that the electron finds
itself going round in a circle or a helix. This in turn is just like a loop
of wire carrying an electric current, and therefore it produces its own
magnetic field. An electron going round in a circle or helix thus has a
magnetic moment which is in addition to the intrinsic magnetic
moment arising from its spin.

When one of these materials is brought into a magnetic field, I get a

spectrum of energy levels because of the effect of the field on the
various types of magnetic moments I have described. These levels
are occupied by the electrons in accordance with the Pauli principle.
Each level corresponds to a different magnetic moment.

When they are all added up, most of them cancel one another out,
and what is finally left is a very weak net magnetic moment, much
smaller than what we find in a ferromagnet. All materials which are
not ferromagnetic show this kind of weak magnetism.

214

SUMMARY
7 Summary
I consider in this chapter the consequences of the magnetic
properties of electrons and nuclei. The one which is closest to
everyday experience is ferromagnetism, as for example in iron. I
describe in section 1 the basic properties of ferromagnets.

Section 2 deals with the magnetic properties of electrons, and

introduces the magnetic moment as a measure of the strength of a
magnet. The magnetic moment is intimately connected to the spin
(or equivalently the angular momentum) of the electron.

I take up ferromagnetism in section 3.1 find that the electrons’

magnetic moments line up in these materials not, as one might have
expected, because of the magnetic force between them. Rather, it is
via the electric force between the charges combined with the nature
of their wave function: a purely quantum-mechanical effect. I find the
reason why the strength of a magnet varies with temperature. I
explain why an ordinary piece of iron is not a permanent magnet.

In section 4 I describe what is meant by the term phase

transformation. Examples are the melting of ice to form water, and
the change of iron from a weak magnetic material to a ferromagnet
as it is cooled through its Curie point.

I go in section 5 to the magnetism associated with nuclei. I explain

the phenomenon of nuclear magnetic resonance, which is at the
heart of the magnetic imaging systems used in medical diagnostic
work.

Section 6 deals with materials which do not show ferromagnetism,

but which nevertheless become weakly magnetic when placed in a
magnetic field. This magnetism, which vanishes when the field is
removed, is also caused by the quantum energy levels of the
electrons which are produced by the effect of the magnetic field on
their motion and their magnetic moments.
I introduced the idea of a phase transformation in section 4. I take up
in the next chapter a spectacular example of such a transformation,
namely the total disappearance of electrical resistance when certain
materials are cooled to low temperatures. This is the phenomenon
called superconductivity.

215

XII SUPERCONDUCTORS

l What is a superconductor?

I have noticed that physicists are rarely given to excesses of

exuberance in talking about their subject, even among themselves.
So when they use a prefix like super for a material, it must be
something quite extraordinary. I shall first tell you what the
phenomenon is, and then explain why it is extraordinary. Suppose I
take a wire of the metal lead which has a resistance of one ohm at
room temperature, about 290 K, and measure its electrical
resistance at different temperatures as I cool it down. I show the
result in fig. XII-1. The resistance drops smoothly as the temperature
drops. Near 7 K, the resistance drops abruptly by a factor of more
than 1020 below its value at room temperature, and remains so at all
lower temperatures. Put differently, the wire now conducts electric
currents at least 1020 times better than it did before. In fact, we can
only set an upper limit to this number because of the limits on the
sensitivity of the measuring instruments. For all we know, the wire in
this condition may be a perfect conductor, which means that a
current can pass through it with zero voltage applied to it. Well,
super is a bit less than perfect, and so we are content to say that
lead has become superconducting, and to call the phenomenon itself
superconductivity.

If I now warm up the lead wire from the superconducting state, I find
that its resistance jumps back to its original value at the same
temperature at which the resistance had disappeared when cooling
the wire. I call this temperature the superconducting transition
temperature, and use the symbol 7^ (read T-nought) to denote it.

When the wire is above this temperature and therefore not

superconducting, I say that it is in the normal state. The change from
one state to the other is very abrupt: a change of temperature of less
than a thousandth of a degree is enough to cause the transition.

We begin to see why superconductivity is so extraordinary. There is

the enormous change in the electrical resistivity, and then there is
the abruptness (meaning the narrowness of the temperature range)
with which it occurs. We saw in chapter IX

how electrical resistance arises. It involves interactions between the

conduction electrons, phonons, and alloying atoms: a complicated
interplay of huge numbers of particles (more than 1020 electrons
alone, not to mention the others). This cacophony is abruptly and
totally stilled just by cooling the wire a thousandth of a degree
through the transition temperature. Something happens which
affects the interplay of all these particles, and which dramatically
changes their behaviour. Here is an analogy: I hear a din emerging
from a classroom, and it suddenly ceases when a bell rings.
Subsequent research shows me that the reason for the silence was
the 216

WHAT IS A SUPERCONDUCTOR?

FIGURE XII-1. Superconductivity of

lead metal. Note that the scales are

logarithmic, and not linear. This is a

convenient way to show the variation

of quantities over a large range of

values. Thus here the temperature

varies by a factor of a few hundred,

and the resistance by a factor of 1020.

The resistance drops precipitously to

zero at a temperature of about 7 K.

io-lof—

10
-15 _
10
-20
11
resistance (ohms)

10

temperature (K)
100 200
entrance of the teacher. So the task is to find out what plays the role
of the teacher in superconductivity.

We are accustomed to seeing small changes in properties with small

changes in temperature: the thermal expansion of solids, for
example. This implies that the fundamental structure of the material,
meaning its basic atomic and electronic structure, also undergoes
correspondingly small changes. The situation is quite different with
superconductivity. The abrupt disappearance of electrical resistance
implies that some new feature must have appeared in the
superconducting state which just did not exist in the normal state.
This is an example of a phase transformation. Ferromagnetism and
the melting of solids are other examples of such transformations.
They are all characterized by a sudden change in the group (as
distinct from individual) behaviour of all the relevant particles in the
material: the conduction electrons in a superconductor, the electron
magnetic moments in a ferromagnet, the atoms in a solid undergoing
melting. Because of the large numbers of particles involved and the
complicated interactions among 217

XII SUPERCONDUCTORS

them, it is a formidable challenge to explain them in detail. The effort

to do so has been most successful with superconductivity, and I shall
describe it later in this chapter.

Superconductivity is, perhaps surprisingly, anything but a rare

occurrence in materials. In addition to lead, there are twenty-seven
other naturally occurring elements which become superconducting,
each with its own characteristic transition temperature. There are
hundreds of alloys and compounds which are also superconducting.
Superconductivity is a far more widespread phenomenon than
ferromagnetism, which appears in only a few materials.

Table XII-1. Some superconducting transition temperatures

superconductor transition temperature (K)

tungsten 0.012

tin 3.7

lead 7.2

an alloy of niobium and titanium 11

a compound of niobium and tin 18

a compound of yttrium, barium, 90

copper and oxygen

I list in table XII-1 the transition temperatures of a few

superconductors. I have selected them to bring out two points: the
variety of materials which can become superconducting, and the
great range of transition temperatures. The two extreme
temperatures in the table differ by a factor of almost eight thousand. I
should mention here as an aside that in physics, a comparison of
two temperatures is often more meaningful in terms of their ratio
than their difference. This is because in statistical physics, which is
at the heart of the idea of temperature, the ratios of the energies of
quantum levels to the thermal energy (which is basically what
temperature is) are more significant than the temperature by itself.
The percentage change in this ratio is the same in going from, say,
one degree to ten degrees, as from a hundred degrees to a
thousand degrees.

question Even though superconductivity is much more prevalent

than ferromagnetism, why is it that I see so many uses of magnets
around me and none of superconductors?
answer Well, at least part of the answer lies in a comparison of
tables XI-1 and XII-1. We have materials which are ferromagnets at
ordinary temperatures (around 300 K, say) but none which are
superconducting at these temperatures. Any use of superconductors
would require refrigeration with liquid air (90 K) or liquid helium (4 K)
using special equipment, and this rules out the everyday use of
superconductors on a scale anywhere near that of magnets. The
search continues for materials which are superconducting at room
temperature, so far without success.

218

SOME OTHER PROPERTIES OF SUPERCONDUCTORS

2 Some other properties of
superconductors
The disappearance of electrical resistance is the most spectacular
property which distinguishes a superconductor from a normal metal.
A matter of terminology: I use the term normal in this chapter as
synonymous with non-superconducting. Thus lead is normal above
7.2 K, and superconducting below that temperature.

Now imagine that I have the wire of lead sitting at a temperature of

around 4 K, which is the same as minus 269 °C. I get to such low
temperatures by using liquified helium gas. The wire has an electric
current flowing through it but with no voltage, because it is
superconducting. Can I do anything to the wire to disturb its
superconductivity? Shaking it or shining light on it does nothing.
Knowing that electric current is carried by electrons, and that a
magnetic field influences their motion, I let a magnet approach the
wire. At first nothing happens; the current continues to flow
undiminished. But when the strength of the field at the wire has
reached some apparently critical level, the current in the wire
suddenly drops and a voltage appears: the wire has lost its
superconductivity and now shows its normal electrical resistance. I
try a few more tests, and find that at each temperature, there is a
critical strength of magnetic field, below which the lead wire is
superconducting and above which it is normal. I show in fig. XII-2
how this critical strength varies with temperature, beginning with very
small values near 7^ and levelling off at a maximum of about 800
gauss at the lowest temperatures. The gauss is the unit in which the
strength of a magnetic field is measured. The strength of the earth’s
field is about half a gauss. Different superconductors show the same
qualitative variation of critical field as the temperature is varied, but
have different values for this maximum critical field, for which I use
the symbol HM. Table XII-2 lists HM for several superconductors.
One sees that superconductors fall into two groups in terms
of//M.There are those whose critical fields are a few hundred gauss
or less, and then those with fields of hundreds of thousands of figure
xil-2. The smallest magnetic

field (in gauss) which will destroy

superconductivity in lead at different

temperatures. Fields less than this do

not affect the superconductor.

00

magnetic

600—

400-

2000-

^-

^™

^~

rr
0 12 3 4 5 6
temperature (kelvin)

219

XII SUPERCONDUCTORS

Table XII-2. Maximum field.H^ for superconductivity in some

materials superconductor

maximum critical field,

in gauss

tungsten

tin

lead

an alloy of niobium and titanium

a compound of niobium and tin

a compound of yttrium, barium,

copper and oxygen

1
300

800

120,000

300,000

estimated to be more than

1,000,000

FIGURE xn-3. (a) The field of a

magnet goes right through a sheet of

normal metal and acts on an iron nail

on the other side, (b) The magnetic

field is stopped by the

superconducting sheet and does not appear

on the other side; the nail lies quietly

on the tabletop. (c) the magnet is

brought closer to the sheet, until the

field it sees is big enough to make it

normal. The iron nail now feels the

magnetic field and is attracted

towards the sheet.

N
 \

magnet S

superconductor –,,

(b)

tabletop

\\

\“

(c)

magnet

superconductor turned normat

iron—

tabletop

22O

SOME OTHER PROPERTIES OF SUPERCONDUCTORS

gauss. I shall refer to them as the low-field (LF) and the high-field
(HF) superconductors respectively. The reason for the different
behaviours lies in the details of how the superconductor reacts to a
magnetic field, as we shall see now.

A magnetic field goes right through a material which is not

superconducting: a magnet will attract an iron nail even if I interpose
a sheet of normal metal in between, as shown in fig. XII-3, part (a). If
I replace the normal metal by a superconducting sheet, as in part (b)
of fig. XII-3, the magnetic field is stopped by it: the iron nail is not
attracted by the magnet. If now I bring the magnet gradually closer to
the superconducting sheet, I find that at a certain point the magnetic
field apparently punches through the superconductor, because now
the nail is attracted towards the magnet, as in part (c) of fig. XII-3.

question Can you identify one essential detail which is missing in fig.
XII-3?

answer Depending on what it is made of, the superconductor must

be sitting in a bath of liquid air or liquid helium, because we have no
materials which are superconducting at room temperature.

Further experimenting shows that the magnetic field punches

through the superconductor if it is the LF type when its strength is
equal to H^, which is also the field at which the material loses its
superconductivity. I illustrate this effect in fig. XII-4. In part (a) of the
figure, the field is less than H^, and cannot break into the
superconductor (shown as an egg-shaped object in the picture). This
is known as the Meissner effect, for Walther Meissner who in
collaboration with R. Ochsenfeld discovered it. In part (b), the field is
larger than HM, and becomes the same inside and outside the
object, which is now normal. It is as if the superconductor tries to
keep the field out because that is how it can remain
superconducting, and the excluded field exerts a sort of pressure to
get into the material. This pressure increases as the field increases,
until finally the superconductor can no longer stand it. The field then
punches its way in and kills the superconductivity.

On the other hand, if the object is an HF superconductor, its

behaviour is more cunning, so to speak, than that of the LF type. I
illustrate it in fig. XII-5. In small fields, part (a), the HF type excludes
the field just like the LF type. But at some medium field which is
much less than //M for that material, it makes a compromise and lets
a part of the field in, as shown in part (b) of the figure. The material
then becomes an intimate mixture of superconducting and normal
regions, and is said to be in the mixed state. As the field increases, it
admits a larger and larger fraction of the field, and the normal
fraction of the volume grows at the expense of the superconducting
fraction. Finally at Hm the field inside and outside the material
become equal, as shown in part (c), and the material becomes
normal throughout its volume.

The two kinds of magnetic behaviour described above go under the

names of type I and type II behaviour respectively. All HF
superconductors show type II behaviour, and they are the ones
which are mostly used in practical applications.

XII SUPERCONDUCTORS

FIGURE XII-4. A type I

superconductor, shown as an egg-shaped object, (a) keeps a small

field

out, and (b) lets a large field in. The

changeover is at the field HM, at

which superconductivity ends. The

state of the sample is indicated: S =

superconducting, N = normal.

(b)

large field

In a normal metal, I need to apply a voltage in order to send an

electric current through it. In a superconductor, there is no resistance
and a current can flow even without a voltage. If I start a current
flowing in a superconducting ring, for example, it will continue for
ever without diminishing, because there is no resistance. I call it a
persistent current. This current going round the loop produces a
magnetic field: it has a magnetic moment. This, like everything else
in nature, is described by quantum mechanics. The magnetic
moment is quantized. It can take only certain discrete values, and
nothing in between. Measurements of this quantization give an
unexpected result: the particles in the persistent current are not
electrons acting singly and separately, but rather pairs of electrons
acting as a unit. This is quite different from the normal state of the
metal, in which the properties are those expected from electrons
behaving singly. It is as if I made experiments with oxygen gas, and
found that the 222

SOME OTHER PROPERTIES OF SUPERCONDUCTORS

FIGURE XII-5. (a) A type II

superconductor, shown as an egg-shaped object, keeps a small

enough

field out. (b) Intermediate fields leak

partially into the superconductor, (c)

When the field HM for the particular

material is reached, the field enters

completely into the object and its

superconductivity ends. The state of

the sample is indicated: S =

superconducting, M = mixed, N =

normal.
A

small field

(fl)

medium field

(P)

results made sense only if the gas consisted not of single oxygen
atoms, but rather of pairs of atoms bound together.

Some of the properties in the superconducting state of a material are

the same as in the normal state, and others are quite different. The
density and crystal structure do not change. The thermal conductivity
and the specific heat, on the other hand, are quite different in the two
states. The thermal conductivity does exactly the opposite of what
we might expect from our understanding of normal metals. We saw
there that a metal which is a good electrical conductor is also a good
conductor of heat: copper for example. We might therefore expect
that a superconductor would be also a super thermal conductor. Not
so at all; the thermal conductivity of a metal in the superconducting
state is smaller than what it is in the normal state. In the metal lead,
for example, the thermal conductivity near 0 K in the
superconducting state is about a 223

XII SUPERCONDUCTORS

million times smaller than in its normal state. The metal, although
electrically a superconductor, becomes thermally an insulator. Since
heat in metals is conducted primarily by electrons, it would appear
that though the electrons in a superconductor carry an electric
current with no resistance at all, they are strangely reluctant to carry
a heat current. The specific heat in the superconducting state also is
less than in the normal state.
3 Some clues
The explanation of how superconductivity comes about is one of the
most impressive achievements of physics. It represents a beautiful
bringing together of several aspects of the physics of solids that I
have described earlier in this book: the quantum mechanics of
electrons in a solid, phonons, the nature of electrical conduction,
magnetic effects.

Let me list some basic facts about superconductivity, and indicate

how they may suggest where to look, or for that matter where not to
look, for an explanation of the phenomenon. This is not the historical
sequence; I wish rather to give you the picture as it is today.

1. In contrast to ferromagnetism, superconductivity is a rather

common occurrence: there are hundreds of known superconductors.
They come in many different crystal structures and chemical
compositions. The only thing they have in common, apart from their
superconductivity, is that they are metallic conductors above their
transition temperatures. So we must seek an explanation in some
features which are quite general in such conductors, and do not
depend specially on crystal structure and chemical composition.

2. The transition temperatures are mostly well below 100 K. We

recall that temperature is a measure of the thermal energy of the
system. The average kinetic energy of the conduction electrons in a
metal, expressed as a temperature, is around 100,000 K.

Thus the energies involved in superconductivity are a minute fraction

of the total electronic energies. The needle-in-a-haystack image
springs to one’s mind.

3. The average energies of phonons are not too far from the
energies corresponding to the transition temperatures. One might
therefore wonder whether phonons should be somehow brought into
the picture. This thought is reinforced by the fact that the transition
temperatures of different isotopes of the same substance differ
slightly, and we know that the isotopes differ slightly in their phonons,
but not in their electronic structure.

4. The fact that electron pairs are the units which carry the
supercurrent suggests that we should look for something which will
be the glue, as it were, binding a pair of electrons together.

5. We recall that heat is conducted in metals predominantly by the

conduction electrons and holes which have been raised to slightly
higher energy states near the Fermi energy because of their thermal
energy, as illustrated in fig. XII-6. The poor thermal conduction in a
superconductor suggests that fewer electrons get to these states
224

SOME CLUES

than in a normal metal. This could happen is if there is a gap

between the Fermi energy and the energy of the lowest available
state for electrons of thermal energy, as shown in fig. XII-7. If such a
gap opened up when the metal becomes superconducting, then
there would be fewer electrons to carry heat in this state than there
are in the normal state, and the superconductor would be a poor
thermal conductor and also have a lower specific heat, as is
observed.

question With reference to the third point above, why should the
isotopes differ slightly in their phonons?

answer The isotopes have different nuclear, and hence atomic,

masses.

The phonons are the quanta of the vibrational energies of the atoms
in the solid, and these energies are different for different atomic
masses.
figure xn-6. Relative numbers of

electrons in different states of energy

above the Fermi energy in a normal

metal at some temperature. The level

labelled 0 is nearest the Fermi energy,

the others are at progressively higher

energies.

FIGURE XII-7. Relative numbers of

electrons in different states of energy

above the Fermi energy in a

superconductor at some temperature.

The lowest allowed energy for an

electron is now in the level labelled 5.

The lower energy levels are

prohibited: we have an energy gap.

By comparing with fig. xn-6, one can

see that there are fewer thermally

excited electrons, and therefore a

lower thermal conductivity and lower

specific heat in the superconducting

state than in the normal state.

tele

o-

<u

“1.

fl

<u

relati

l_

10

energy levels

15

20

15

20
energy levels

225

XII SUPERCONDUCTORS
4 Cooper pairs and BCS
The physicist Leon Cooper showed how a certain kind of interaction
between electrons can lead to the formation of electron pairs
composed of opposite momenta and spin, which have therefore
come to be called Cooper pairs. A Cooper pair is a peculiar object. It
is not two electrons somehow bound to each other so that they stay
close together and move about as a unit: an electronic molecule
consisting of just two electrons, so to speak, just as an oxygen
molecule consists of two closely bound oxygen atoms moving about.
In fact, it is not possible for two electrons by themselves to be bound
together: the electric force of repulsion between their charges would
make them fly apart. One needs a mediator between them to hold
them together in spite of this repulsion. An example of such a
mediator is the nucleus of the helium atom, which holds two
electrons bound close to each other. This is because each electron
is attracted to the positive charge of the nucleus, and this attraction
more than overcomes the repulsion between the electrons. I can talk
about an electron pair in the helium atom bound to each other, and it
takes a finite energy to break up the pair: this would be the energy
needed to remove one electron and leave behind a helium ion.

The mediator that holds the two electrons of a Cooper pair together
is a phonon, which is the quantum of energy of the thermal vibrations
of the atoms in the solid. I illustrate how this comes about in fig. XII-
8. Two electrons A and B approach each other travelling in opposite
directions with the Fermi velocity. Electron A emits a phonon and
moves off in a different direction. Electron B absorbs the phonon and
changes its direction of motion by the same amount as A, so that the
two electrons are still moving in opposite directions. The pair
remains intact, bound together, so to speak, by the phonon. This is a
Cooper pair. This process happens to all pairs of electrons moving
with opposite velocities all over the Fermi surface, and the result is
the superconducting state. Remembering that momentum is mass
times velocity, we can say that a Cooper pair is two electrons of
opposite momenta (and therefore of zero net momentum) bound
together by the emission and absorption of a phonon. The two
electrons also have opposite spins: this is established through
experiments and predicted by the theory.

question How does an electron emit or absorb a phonon?

answer The negative charge of the electron pulls towards it the

positively charged ions surrounding it, causing a local increase of
pressure. This is also what a sound wave would do, and so one can
speak about the emission of a phonon by the electron. A half-cycle
of the sound wave later, what was a compression has become an
expansion of the lattice. If now a second electron comes along, it
pulls (through the electric force) the surrounding ions back to their
normal positions, and this corresponds to the absorption of the
phonon.

We now have a picture of how Cooper pairs are formed through the
mediation of phonons. But there is more to it than just the formation
of pairs: it turns out that the 226

COOPER PAIRS AND BCS

FIGURE XII-8. Formation of a Cooper

pair. Electrons A and B approach each

other travelling in opposite directions.

Electron A emits a phonon which is

absorbed by B, and they both change

their directions of motion which are

still opposite to each other.

electron A

energy of the electrons in such a pair is less than the energy that the
electrons would have if they were moving independently. When
electron A creates and emits a phonon, what is happening is that it is
pulling towards itself the positive ions in its neighbourhood because
of the electric force between its negative charge and their positive
charges: you remember that a phonon represents a compressional
wave in the solid. The potential energy of the electron is therefore
lower than it would be if the ions were in their undisturbed positions.
Electron B, which arrives there half a cycle of the phonon later, sees
these ions being further from it than would otherwise be the case,
and therefore its energy is raised. However, this increase is smaller
than the decrease in energy of electron A.

The net energy of the Cooper pair is thus lower than that of the two
electrons when they are moving independently. The difference
between the two is the binding energy of the Cooper pair, and must
be supplied to it in order to break up the pair.

I have now described how one Cooper pair is formed. All the
electrons in the vicinity of the Fermi surface are paired up in this
manner at the absolute zero of 227

XII SUPERCONDUCTORS

temperature in a superconductor. You will notice that the formation of

a pair involves the two electrons going from an initial pair of states to
a final pair through the emission and absorption of a phonon, and
this is only possible if the final pair of states is not already occupied
by electrons.

question Why is it necessary for the final states to be empty?

answer Because electrons are fermions, and no more than one

electron can be in any given state. If there is already an electron in
that state, then a second electron cannot enter it.

The answer above means that Cooper pair formation depends on

what all the other electrons are doing: there is strong correlation
among the Cooper pairs, they are not independent entities. I give a
rough analogy: the electrons in a normal metal are like the molecules
in a gas, they are all moving about independently. The electrons in a
superconductor are like the molecules in a crystal. These molecules
are all strongly correlated with one another, so that only the crystal
can move as a whole, and not the single molecules independently of
one another. Just as we talk about the molecules in a gas
condensing to form a crystal, we have a sort of condensation of the
electrons in the superconductor into a new state which is the
superconducting state. The analogy is only rough: the molecules in a
crystal condense into an ordered state with respect to their positions,
and this is not the case with the electrons in a superconductor. They
condense into an ordered state with respect to their momenta. But
there is still something left in the analogy. The atoms in the gas have
momentum because they are moving about, whereas the atoms in
the crystal are sitting still (apart from their thermal vibrations) with
zero momentum. Similarly, the electrons in a normal metal are all
moving about, whereas each Cooper pair in the superconductor has
zero momentum and is effectively sitting still. The crystal can be
moved only as a whole, and each molecule in it has exactly the
same motion as the crystal itself. One of the molecules can have a
different motion only if it is pulled out of the crystal, and this requires
a finite amount of energy. Similarly, all the Cooper pairs in the
condensed state move as a whole, and to pull single electrons out of
it requires a finite amount of energy, namely the binding energy of a
pair.

Figure XII-9 suggests what the essential difference is between the

normal and superconducting states of a metal. You must imagine as
you look at the figure that the individual electrons in all cases are
speeding about, but the Cooper pairs are stationary (with zero
current) or have a common velocity (with current). In the normal
metal, I need in principle to keep track of all the electrons to
understand its behaviour. In the superconductor, however, I need to
know the behaviour of just one Cooper pair: all the others will be
doing exactly the same thing.

We can now see why an electric current in a superconductor flows

without resistance.

The current is carried by the condensed Cooper pairs which are all
moving at the same velocity. Electrical resistance is caused by the
scattering of electrons moving independently of one another. Such
electrons can only be pulled out of the Cooper pair condensate when
at least a finite minimum amount of energy is supplied to it. This 228

COOPER PAIRS AND BCS

FIGURE xil-9. In a normal metal, part

(a), electrons are moving about

independently in all possible

directions. In a superconductor, parts

(b) and (c), Cooper pairs are formed

by electrons with opposite velocities

and spins. The total momentum of a

pair is zero, so that its effective

velocity is also zero, when there is no

electric current, as in part (b). They all

move with the same velocity in the

presence of a current, as in part (c).

The short arrows indicate directions

of motion of individual electrons, and

the long arrows show the direction of

motion of each Cooper pair when

there is a current.

normal metal

superconductor, zero current

superconductor with current

energy would, for example, be available in the kinetic energy of

motion of the pairs when the electric current is large enough. The
energy could then go into breaking up the pairs, and normal
resistance would be restored. This provides a mechanism whereby
the superconductor can carry only a specific maximum current,
called its critical current, before it loses its superconductivity and
becomes a normal metal with resistance.

The picture of a superconductor which I have described above was

proposed by the physicists John Bardeen, Leon Cooper and Robert
Schrieffer, and is known in the world of physics as the BCS theory.
They, and others following them, showed that by working out its
consequences in detail, one could explain all the observed
properties of superconductors. The BCS theory is one of the great
achievements of the quantum mechanical description of nature. It
was originally developed to show how the large 229

XII SUPERCONDUCTORS
number of electrons in a metal can influence one another through
the mediation of phonons and thereby form the superconducting
state. But its significance is far broader, because it showed a way of
looking in general at fermions which are interacting with one another.
It has been used, for example, to explain properties of atomic nuclei,
where the fermions are protons and neutrons, and of certain
astronomical objects called neutron stars, composed of very densely
packed neutrons.
5 Josephson oscillations
The physicist Brian Josephson used the BCS theory to predict an
effect one should see in superconductors. The effect was indeed
found experimentally, and is named after him. Two superconducting
pieces A and B are brought together with a thin insulating layer C
interposed between them as shown schematically in fig. XII-10, thus
forming a junction. A battery is connected to A and B, so that there is
a voltage of F volts between FIGURE XII-10. The Josephson effect.

Two superconductors A and B are

separated by a thin insulator C. A

voltage of V volts is applied between

A and B. An oscillating supercurrent

begins to flow between A and B

through C, and this produces an

emission of photons (electromagnetic

waves) of frequency equal to 2eV

divided by the Planck constant h.

photon 2eVlh

230

JOSEPHSON OSCILLATIONS
them. The layer C is thin enough that some current, even though
small, can flow through it. If the rest of the electrical circuitry (not
shown in the figure) is suitably set up, then photons (electromagnetic
waves) of frequency/= leVlh, where e is the electronic charge and h
is the Planck constant, are emitted by the junction. Given the
voltage, the frequency /depends only on two universal constants,
and not on which particular superconductor is used: a remarkable
result indeed. The effect follows purely from the existence of Cooper
pairs (hence the factor 2e in the formula for the frequency) and the
rules of quantum mechanics (and therefore the factor h).

You can understand how the Josephson effect comes about by

recalling some basic results of quantum mechanics. The
superconducting state consists of Cooper pairs which all have zero
momentum (i.e. no current flowing) or the same non-zero momentum
(i.e. some finite supercurrent flowing). This state can be described by
a wave function. I do not need to know, for my present purpose, the
exact form of this wave function. But I know that there is some part
of it which tells me what the energy and velocity (or equivalently the
frequency and wavelength) are for each Cooper pair.

Both of these latter quantities stand for oscillations of the wave

function, the frequency in time and the wavelength in space. I can
combine both of these oscillating properties of a wave function in a
single quantity called the phase, which is simply an angle, as shown
in fig. XII-11. The phase has the following properties, which are a
consequence of its being a part of the wave function:

1. Its frequency of oscillation /“is directly related to the energy E by

the quantum mechanical formula

E=hf.

2. If the phase is different at two points in the material, then there

must be a flow of matter between the two points. If this difference is
constant with time, then the flow is also constant. But if the phase
difference oscillates in time, so does the flow.
FIGURE XII-11. The phase for a wave

function. At the top is one complete

cycle of a wave, as it varies with time

(relating to its frequency) or at different

distances (relating to its wavelength).

Now imagine an arrow which is

rotating so that it completes one

revolution during each cycle of the

wave. The direction of the arrow

represents the phase of the wave, and

five examples are shown for five points

on the wave. The magnitude of the

phase is given by the angle which the

arrow makes with some fixed direction,

as shown at the bottom of the figure.

wave

time

phase

angle

90
distance
180 270
231
XII SUPERCONDUCTORS

Now we look at what these properties mean for the junction shown in
fig. XII-10.

There is a voltage difference of F volts between the superconductors

A and B on either side of the insulating layer. The potential energy of
a charge is equal to the charge multiplied by the voltage it sees. The
voltages seen by the Cooper pairs in A and B

differ by Fvolts. Therefore the energy of a Cooper pair, with charge

2e, differs in the two superconductors by an amount 2eV. This
means that the phase difference between the two is oscillating at a
frequency/given by 2 eV divided by h: 2eV

f= h ■

This oscillating phase difference means that there is an oscillating

current at the same frequency flowing between the superconductors
A and B. Such an oscillating electric current, as we know, is a
generator for photons of the same frequency, and this is exactly what
is seen to happen. This is the Josephson effect.

One aspect of the Josephson effect is especially striking. A basic

feature of quantum mechanics is that the wave function has a
frequency which is determined by the energy of the system. This
frequency does not, in the many cases I have talked about in this
book, make a direct and explicit appearance. But in the Josephson
effect, it makes a very palpable appearance, through the frequency
of electromagnetic waves (photons) which have been detected and
measured in the laboratory. Any lingering doubt one might have
about the reality of wave functions should be dispelled by the
existence of the Josephson effect.
6 Superconductors at work
The most important practical use of superconductors so far has been
to make magnets consisting of spools of superconducting wire
carrying an electric current. The wire is made of an alloy of the
metals niobium and titanium, and it is superconducting below about
UK. The magnet itself must be cooled in liquid helium to about 4 K,
but the whole equipment is so constructed that the magnetic field
can be made available in a volume at room temperature. Such
magnets are used for magnetic resonance imaging (MRI) for medical
diagnostics, and for experiments in physics and chemistry which
need magnetic fields. The MRI technique uses the nuclear magnetic
resonance of various nuclei in the human body to give a picture of
the body’s interior.

Before superconducting magnets became practical and available,

magnets consisted of spools of copper wire wound around an iron
core. Copper is a good conductor of electricity but still has some
electrical resistance. Therefore the passage of a current heated up
the spool: electrical energy was being converted into thermal energy.
This had two undesirable consequences, namely the need for
significant amounts of electrical power (power = rate of energy use),
and a limitation on the strength of the fields that could be produced.
The need for electrical power to produce a steady magnetic field is
especially disappointing, because it should in fact take no energy at
all to 232

SUMMARY

maintain such a field: a permanent magnet does so with no energy

being fed into it. All the energy supplied to a copper-spool magnet
gets converted to heat, and the heat has to be somehow carried
away in order to prevent a possible meltdown of the magnet.
In a superconducting magnet, the wire has zero resistance and
therefore the electrical current produces no thermal energy at all.
Because of the very high critical magnetic fields for the alloys used
to make this wire, superconducting magnets produce much higher
fields than are possible with the more usual copper-spool magnets.

Magnets are by far the most succesful practical application of

superconductors to date. There have been proposals, and some
experiments, to use such materials in other areas of technology.
Examples are superconducting cables to distribute electric power
from generating stations, and Josephson junctions used in electronic
devices. These efforts continue, but the results till now have not
approached the level of superconducting magnets in their availability,
reliability and usefulness.
7 Summary
This chapter is concerned with superconductivity, a phase transition
which occurs in a large number of metallic elements, alloys and
compounds. The basic effect is described in section 1: the total
disappearance of electrical resistance below a certain critical
temperature.

Section 2 looks at some other properties of superconductors. The

superconductivity is destroyed in a magnetic field larger than a
certain critical value which is different for different superconductors.
They appear to fall into two groups, one with relatively low critical
fields and the other with rather high critical fields. The difference lies
in the way in which the magnetic field penetrates into the
superconductor.

Some clues towards understanding how superconductivity comes

about are presented in section 3. The phenomenon is very general, it
somehow involves phonons, the electrons appear to act as bound
pairs, a finite amount of energy is needed to break up such a pair:
these were some of the clues based on experiments done on
superconductors.

Section 4 describes how Leon Cooper showed that electrons

interacting with one another can form pairs. With this result, John
Bardeen, Leon Cooper and Robert Schrieffer used the quantum
mechanics of metals with electrons interacting via phonons to
explain the phenomenon of superconductivity in all its details.

The Josephson effect is the subject of section 5. Brian Josephson

showed that a voltage applied between two superconductors
separated by a thin insulating layer produces electromagnetic
radiation. One might think that quantum mechanics shows its special
features primarily when it is applied to matter on an atomic scale.
The Josephson effect is a beautiful example of the universal validity
of quantum mechanics: the effect is purely the result of the quantum
behaviour of pieces of matter one can see with the naked eye and
hold in one’s hand.

233

XII SUPERCONDUCTORS

Section 6 takes up practical applications for superconductors. By far

the most successful such application has been in magnets for
producing high fields. Some other possible applications, which are
not as far along as magnets, are mentioned.

The explanation of superconductivity given by the BCS theory is one

of the most impressive accomplishments of the quantum mechanical
description of matter as developed in the earlier chapters of this
book. The theory in its basic form, as I have outlined it here, involves
electrons, phonons, and their mutual interaction. Refinements and
extensions of the theory have been made to take account of, among
other things, the differences in band structure of different metals,
effects due to the presence of magnetic impurities, and effects of
electromagnetic radiation. The results of experiments are in every
instance well explained by the theory. With the explanation of how
superconductivity comes about, we have attained one of the
pinnacles of physics.

234

XIII CONCLUSION

I compared our journey through this book to a mountain ramble. The

last chapter, about superconductivity, is like a high peak in the
landscape of condensed matter physics. I should like to look back at
the terrain we have covered to get there. What I have in mind here is
not a summary of the preceding chapters, but rather a review of the
key elements necessary for our understanding of the properties of
condensed matter. These elements belong to three fundamental
parts of physics: quantum mechanics, statistical physics, and
electromagnetism. We refer to them as theories, but in a very
precise sense. They are supported by vast numbers of experiments,
and they predict correctly the results of new experiments. The word
theory as used here has a significance different from what it has in
everyday language as in ‘I have a theory about why it always rains
on Saturdays.’

Quantum mechanics is the language in physics that we use to

describe the world. It incorporates the fact that everything around us
is sometimes particle-like, and sometimes wavelike. This basic
duality of nature leads to a kind of graininess: just as matter comes
in discrete units like electrons and protons and not fractions thereof,
quantities like energy and momentum are quantized and cannot take
on any arbitrary value we wish. The wave aspect of nature is an
expression of a fundamental indeterminism. A simple example of this
is an electron contained in a box. If I do a large number of
experiments to find out where exactly it is, I find that a pattern
emerges. There is a definite probability of finding it in a given region
inside the box, and this probability can vary between different
regions. What I cannot do is to predict where I shall find the electron
the next time I look for it, after I have found out where it is now;
herein lies the indeterminism of nature. What is not indeterminate is
the probability of finding the electron in some specific region, and
this probability is predicted correctly by quantum mechanics. Even in
considering just one particle, we need to use the idea of probability
which is derived from statistics.

We also need statistics in the usual sense of a way of handling large

numbers of objects, in order to apply quantum mechanics to matter
involving, for example, 1021

electrons. We then get a remarkable result, that all particles can be

classified into two types, named bosons and fermions, with strikingly
different properties. There is another basic result of statistical
physics which I merely stated, namely the distribution of particles in
different energy ranges at a given temperature. If I fix the total
energy and the total number of particles, there are many different
ways in which the particles can be distributed in the different energy
levels. There is one particular distribution which occurs much more
often than any other, and this is the one which actually occurs in
nature. The following example illustrates this point. If I toss ten coins
repeatedly, I 235

CONCLUSION

find different numbers of heads and tails each time. There will
however be many more times with five heads and five tails than any
other possible combination. The larger the number of coins, the
more likely will be the result with equal numbers of heads and tails.
With the 1021 or so particles that we are concerned with in solids,
the most probable distribution becomes a practical certainty.

The third and final element in our description of matter around us is

the theory of electromagnetism. Nature exhibits four kinds of forces,
labelled electromagnetic, nuclear, gravitational, and weak. The
electromagnetic force is responsible for the properties that we have
considered in this book. The nuclear and weak forces determine the
properties of atomic nuclei and their constituents, and the
gravitational force becomes dominant at the other end of the scale,
with astronomical objects. Quantum mechanics and statistical
physics play a fundamental role in all these cases.

Physics as it is today gives a completely satisfactory account of the

properties of condensed matter. In the other two areas, namely the
nuclear particles and the cosmos, however, there are some
observations which cannot be accommodated within the present
structure of physics. This is not to say that the physics has failed, but
rather that something needs to be added to it so that it can
encompass the nuclear and gravitational forces too. A bicycle
without a lamp on which I can ride about quite happily during
daytime is not a failure because I cannot ride it at night; I just need to
mount a lamp on it.

Physics explains why things are the way they are, in the realm of
condensed matter.

This however does not mean the end of condensed matter physics
as an active area of study and research; indeed, far from it. It is like
cooking; perhaps not too far-fetched an analogy, for an
experimenter’s laboratory is somewhat like a kitchen. One often
encounters physicists exchanging, or being secretive about, their
recipes, and it is not something to eat that they have in mind. The
analogy: the basic principles and techniques of cooking are well
established, but this by no means spells the end of the art (or
science) of cooking. Quite the contrary: there is no end in sight of
possible variations on the well-known or the invention of the new,
though the basic knowledge and techniques remain the same. So it
is in condensed matter physics too. The basic principles, as I have
described them in the preceding chapters, are well established, and
they give us a good explanation of material properties. But what is
not now, and may never be, possible is to predict exactly all the
properties of a given material. For example, you might give me a
new material, tell me its composition and crystal structure, and ask
me if it is going to be superconducting. My answer will be ‘Perhaps;
but then again, perhaps not.’ The reason for such indecision is that
one has a very large number of particles interacting with one
another, and one can make a theory of its properties only after
making approximations. The problem otherwise will be too
complicated to solve, even though we know how to formulate it. It
could be that a particular approximation causes something to be
ignored which was precisely what would have made your new
material become a superconductor. One of the skills a physicist
needs in this field, as she approximates, is to know what to keep and
how 236
CONCLUSION

much of it, and what to throw away — rather like a cook in his
kitchen. A related question is, can we design a material so that it has
specified properties, for example a superconducting transition at 300
K? The answer at best can only be ‘Not yet.’

The great triumph of condensed matter physics has been its

explanation of the properties of things in general: elastic, optical,
electrical, magnetic, superconducting and so on. Its task, and its
challenge, in the future will be to explain quantitatively and
completely the properties of the ever more complicated materials
that are being made, and to design new materials which have
specified properties. All this still is within the world of inanimate
things. Living things are also composed of the same kinds of atoms
that make up the inanimate part of the world. Can one then apply the
same principles of physics to obtain an understanding of, say,
memory, which is comparable to what we already have of
superconductivity? It is not clear whether a description that is based
purely on physics can be given of all aspects of the living world.
There is already progress, however, in analysing some basic
biological materials and processes at an atomic level. This field of
biophysics, along with condensed matter physics, will continue to be
lively areas of study and research. The basic physics is already
there. What will be needed as before is artistry in approximating, and
the wisdom to know when to stop.

237

GLOSSARY

absolute zero A temperature of—273.16 degrees on the Celsius

scale at which all thermal motion ceases.

alloy A metallic solid containing atoms of more than one element.

Common examples are brass which contains copper and zinc, and
stainless steel which is composed of iron, nickel and chromium.

ampere The unit for measuring electric current: a flow of 6 X1018

electrons per second is one ampere.

angstrom A unit of length such that there are a hundred million

angstroms in one centimetre; useful in considering lengths on an
atomic scale.

angular momentum A quantity which is for a rotating body what

linear momentum is for a moving particle. It is equal to the rotational
speed multiplied by a quantity which depends on how the mass of
the body is distributed around the rotational axis.

annealing Heating a solid, usually an alloy, to an elevated

temperature and holding it there for some time. This treatment can
change some of the mechanical and other properties of the solid.

antinode The parts of a standing wave where the amplitude of

vibration is a maximum.

antisymmetric wave function A wave function which only goes from

positive to negative, or vice versa, without changing its magnitude,
when two of the particles described by it are interchanged.

atom The smallest entity, a few angstroms in size, which

characterises a chemical element. It has a positively charged
nucleus and a number of negatively charged electrons, which
number

determines the element of which it is an atom. The atom itself is

electrically neutral, i.e. its total charge is zero.

axis of rotation An imaginary line in a rotating body, around which

each part of the body is moving in its own circle. Such an axis for the
rotating earth is the line joining the north and south poles.
band A group of very closely spaced quantum energy levels for
electrons in a solid which are derived from a single energy level of
the isolated atom.

band structure The dispersion curves for the electrons in a solid.

biophysics The study of biological materials and processes using the

principles and techniques of physics.

238

GLOSSARY

Bloch electron

Bohr magneton

Boltzmann constant

boson

Brillouin zone

brittleness

bulk properties

c.g.s. units

cc

charge

chip

condensed matter physics

conduction band

conduction electrons

conservation of energy

conservation of momentum

An electron in a crystal whose properties take account of the atomic

structure of the crystal.

A unit, and a fundamental constant, for measuring the strength of the

magnetic behaviour of electrons; the magnetic strength of a free
electron is very close to one Bohr magneton.

One of the fundamental constants of nature, denoted by &b,

particularly important in statistical physics; provides for example the
connection between the energy of vibration of the atoms in a solid on
the one hand, and the temperature of the solid on the other.

The name for particles which obey Bose-Einstein statistics; the wave
function for a system of bosons is symmetric, i.e. it is unchanged
when two particles are interchanged. Examples are helium atoms,
photons and phonons.

A polyhedron in wave vector space, within which are contained all

the possible wave vectors for the electrons in a crystal.

The tendency of a body to break rather than continue bending when

pushed beyond a certain point.

The properties of a piece of matter taken as a whole, to be

distinguished from the properties of its surface.

A system of measuring physical quantities in which the units of

length, mass and time are centimetre, gram and second

respectively.
Abbreviation of ‘cubic centimetre’; also called a millilitre.

A property of matter which leads to the electromagnetic force, just as

mass is the property which leads to the gravitational force.

A small thin piece of matter, usually made of silicon, on which

transistors and other devices are deposited, all on a microscopic
scale, to form an electrical circuit.

The physics of solids and liquids, where the atoms are so close to
one another that they cannot be treated as independent particles.

A band of quantum energy levels in a solid, only partially occupied

by electrons. These electrons are capable of carrying an electric
current, hence the name.

The electrons in a conduction band.

A fundamental law of nature: the total energy of a system isolated

from the outside is always the same, and the only change possible is
from one form to another.

A fundamental law of nature: the total linear momentum and the total
angular momentum of a system, isolated from the outside, each
remains the same for all time.

239

GLOSSARY

continuous spectrum

Cooper pair

cosmological principle

coulomb
critical current

crystal

crystal structure

Curie temperature

current

current density

degeneracy

density

diffraction

dimensions of space

diode

dislocation

dispersion

Light has such a spectrum if all frequencies within some range are
present in it; sunlight has a continuous spectrum.

The entity found in a superconductor, consisting of two electrons with

opposite momenta and spins which can only move as a whole and
not independently of each other.

A term borrowed from astronomy; means the arrangement of atoms

around an atom in a crystal is exactly the same as around any other
atom.

The unit for measuring electric charge, equal to the total charge of
about 6 X1018 electrons.
The maximum electric current that a superconductor can carry and
still remain a superconductor with zero electrical resistance.

A solid whose atoms or molecules are in positions having

translational symmetry.

The result of placing specific atoms or molecules at the sites of a

crystal lattice.

The maximum temperature at which ferromagnetic behaviour is

shown; the ferromagnetism disappears above this temperature.

The term often used when electric current is meant.

The current flowing through per square centimetre of the cross-

section of a wire.

The situation in which more than one wave function

corresponds to the same quantum energy level.

The density of something is the amount of that thing per unit volume,
usually taken to be one cc. Thus the density of electrons is the
number of electrons per cc. One also says the density of silver to
mean the mass of silver per cc of volume. The context makes clear
which sense is meant.

The bending of light or other waves as they pass past an obstacle.

A straight line is one-dimensional: one number specifies every point

on it, namely its distance from some fixed point on the line. A flat
surface is two-dimensional, because a point on it is fixed by its
distances from two perpendicular lines on the surface. Similarly a
point in the space around us requires three numbers, its distances
from three intersecting planes

perpendicular to one another.

A device that converts an alternating electric current which changes
its direction of flow periodically to direct current, which always flows
in the same direction.

A type of imperfection in a crystal which permits plastic deformation.

The fact that the speed of light in a substance is different for different
frequencies.

240

GLOSSARY

dispersion curve

doped semiconductor

elastic limit

elasticity

electric charge

electric field

electric neutrality

electromagnet

electromagnetic radiation

electron

electron-volt

element
elementary particle physics

energy

energy level

equation

erg

A pictorial representation of how the energy (frequency) of a particle

(wave) varies with momentum (wave vector). The term is used for
particles like electrons and photons as well as for waves like Bloch
electrons and light.

A semiconductor containing a small controlled amount of another

element, called the dopant, in order to set its content of electrons or
holes.

The maximum deformation which an object can tolerate and still

show elastic behaviour; if deformed beyond this limit, the object
either fractures or shows plastic behaviour.

The small change in size and shape of an object produced by

applying a force to it disappears when the force is removed, leaving
the object in exactly the same condition as before.

Same as charge.

A region of space is said to be permeated by an electric field if an

electric charge, an electron for example, feels a force when placed
anywhere in it.

The state of an object in which the total positive and negative

charges are exactly equal in magnitude.

A magnet consisting of a suitably shaped piece of iron or other

ferromagnetic material, surrounded by a coil of wire through which
an electric current flows. Sometimes, and particularly when the wire
is a superconductor, the ferromagnetic material is omitted.

Wave motion of coupled electric and magnetic fields, with a speed in

empty space of c= 3 X 1010cm/sec.

One of the fundamental particles in nature, characterized by its

mass, negative electric charge, magnetic moment and spin.

A unit of energy, equal to the energy gained by an electron in moving

across a voltage difference of one volt.

A substance composed of only one kind of atom, silver for example.

The physics of the basic interactions among the fundamental

particles: electrons, protons, photons, quarks, and so on.

An object is said to have energy either because it is moving, or

because by some suitable process motion can be produced in it or in
another object.

A definite amount of energy that an object can have according to the

rules of quantum mechanics.

A shorthand way, using symbols, to say that something is equal to

something else.

The name of the unit in which energy is measured. An object having

a mass of two grams and travelling at a speed of one cm per second
has a kinetic energy of one erg.

241

GLOSSARY

excited state
exponential notation

Fermi energy

Fermi momentum

Fermi surface

An object, an atom for example, is said to be in an excited state if its

energy is greater than the lowest possible value it can have.

A compact way, explained on p. 3, of writing very small or very large

numbers which are often encountered in physics.

The highest energy which conduction electrons in a metal can have

at the absolute zero of temperature.

The largest momentum which conduction electrons in a metal can

have at the absolute zero of temperature.

The surface in wave vector space, formed by connecting the tips of

all possible Fermi wave vectors. All quantum states for wave vectors
inside this surface are occupied by electrons, and all states outside
are unoccupied, at the absolute zero of temperature.

A wave vector that corresponds to the quantum state of maximum

energy, i.e. the Fermi energy, occupied by an electron in a metal.
States of higher energy are unoccupied.

An elementary particle (examples: electron, proton, neutron) to

which the Pauli exclusion principle applies.

A material in which the magnetic moments of one or more electrons

from each atom line up on their own in the same direction even when
no magnetic field is present. Examples are iron, nickel and cobalt.

A phase transformation characterized by the abrupt appearance of

permanent magnetism below a certain temperature.
A scalar or vector quantity whose value depends in general on
position and time. Examples are the temperature in the earth’s
atmosphere (a scalar field), and the electric and magnetic fields in a
light wave (vector fields).

A force applied to an object changes its velocity (in magnitude and/or

in direction).

A statement that something is equal to something else, usually

expressed in terms of algebraic symbols like x, y and so on.

The abrupt breaking into pieces of a material when an attempt is

made to deform it.

The number of oscillations per second, often expressed as so many

hertz (one hertz = one oscillation per second).

A quantity is said to be a function of another quantity x if the value

ofdepends on the value taken by x. This dependence is expressed
by writing the function asf(x). One can also have a function
depending on several quantities x,y, … and written fundamental
constants Certain unchanging properties that are found in nature:
Fermi wave vector

fermion

ferromagnetic material

ferromagnetism

field

force

formula

fracture

frequency
function

242

GLOSSARY

gauss

glass fibres

glassy solid

grain

gravitation

ground state

hertz

high-field (HF) superconductor

hole

infinity

insulator

interaction

interference

Josephson oscillations

joule

kinetic energy
lattice

lattice parameters

examples are the mass and electric charge of the electron, the
speed of light in free space, the Planck constant, the Boltzmann
constant, the Bohr magneton.

The unit in which the strength of a magnetic field is measured.

The earth’s magnetic field is about half a gauss.

Thin strands drawn from glass.

A solid in which the atoms or molecules are in disordered positions,

without translational symmetry, as also is the case in ordinary glass.

Most solids are made up of tiny crystals called grains which are
stuck together.

A basic force between two bodies, derived from their masses.

The quantum state of lowest energy of a system.

One oscillation per second.

A superconducting alloy or compound which retains its

superconductivity up to magnetic fields which are hundreds or

thousands of times larger than those which most

superconducting elements like tin or lead can stand before they lose
their superconductivity.

A fictitious particle, a kind of anti-electron, whose motion replicates

the behaviour of a band in which one quantum state is empty and all
the others are occupied by electrons.
Suppose x stands for a quantity which continues to grow larger and
larger without end. Then one says that x tends to infinity.

A solid which at ordinary temperatures permits no electric current to

flow through it when a voltage is applied to it.

The influence which one body exerts on another. The origin of ■the
interaction could be their masses, in which case the interaction is
gravitational. If it is their electric charges and magnetic moments,
then the interaction is called

electromagnetic.

The overlapping of two waves to produce a new resulting wave.

The oscillating electric current produced between two

superconductors which are just barely in contact with each other

when a steady voltage is applied between them; also called the
Josephson effect.

A unit for measuring energy, equal to 107 ergs.

The energy possessed by a body because it is moving.

An orderly arrangement of points in space with translational

symmetry.

The lengths and angles which are necessary to specify a crystal

lattice completely.

243

GLOSSARY

lattice site
lattice spacing

line spectrum

linear momentum

low-field (LF) superconductor

macroscopic

magnetic domain

magnetic field

magnetic moment

magnetic monopole

mass

matter

mean free path

metal

microscopic

mode of vibration

molecule

momentum

MOSFET

MRI

A point in a lattice.
The distance between two adjacent points in a lattice.

Light with a line spectrum has only certain frequencies in it, and not
others. Light from a neon tube is an example.

The mass of a moving object multiplied by its velocity is equal to its

linear momentum.

A superconductor, usually an element (as distinct from an alloy or

compound of more than one element), whose maximum

critical magnetic field is less than about a thousand gauss in

strength.

Term used to denote the scale of everyday things around us.

A portion of a ferromagnetic material within which all the electron

magnetic moments are pointing in the same direction.

A region of space is said to be permeated by a magnetic field if a

magnet, the one associated with an electron for example, feels a
torque when placed anywhere in it.

A measure of the strength of a magnet.

An isolated north (or south) pole of a magnet, existing by itself.

It has not so far been found in nature or produced in the laboratory.

The property of an object that leads to the gravitational force, and to

the fact that a force is necessary in order to change the velocity of
the object.

Stuff which has mass; the difference between matter and energy has
become a bit blurred since Einstein’s E = mc2.

The average distance travelled by an electron or phonon between

two successive collisions.
A solid in which there are electrons, in number comparable to the
number of atoms, free to wander about at all temperatures down to
the absolute zero.

On an atomic scale.

A sound wave of a given frequency and corresponding wave vector

that occurs in a crystal; the thermal vibrations of the atoms are
equivalent to the sum of many such modes:

Two or more atoms bound to one another and treated as a single

entity.

A quantity which combines information about the mass and the

motion of a body: linear momentum for translational motion, and
angular momentum for rotational motion.

Acronym for metal-oxide-semiconductor-field-effect-transistor.

Acronym for magnetic resonance imaging, a technique which uses

nuclear magnetic resonance for medical diagnostics.

244

GLOSSARY

n-type semiconductor

neutrino

neutron

node

normal state

nuclear fission
nuclear magnetic resonance

nuclear magneton

nucleus

ohm

operation

order

p-n junction

p-type semiconductor

particle

Pauli exclusion principle

persistent current

A semiconductor which is so doped that the majority of current

carriers are (negative) electrons.

An elementary particle with no mass or electric charge, but which still

has energy and momentum. This state of affairs is allowed by the
theory of relativity for particles travelling at the speed of light.

An elementary particle with no charge, and a mass

approximately the same as that of the proton.

The part of a standing wave where the amplitude of vibration is zero.

The state of a superconductor which has lost its

superconductivity because it is above the transition temperature, or

because it is in a magnetic field larger than the critical field.
The splitting of an atomic nucleus into two or more fragments.

The increased absorption of photons, if they have certain

frequencies, by the nucleus of an atom placed in a magnetic field.

A unit, and a fundamental constant, used to express the magnetic

moment of neutrons, protons and nuclei; it is about 5.4 X10”4

times the Bohr magneton.

What is left of an atom if all its electrons are stripped off. The
nucleus is made up of protons and neutrons, contains most of the
atom’s mass, and is about 10~I3cm across.

A unit in which electrical resistance is expressed; one volt applied

across a resistance of one ohm causes a current of one ampere to
flow through it.

The act of doing something to an object, such as rotating it about an

axis through an angle.

A term to denote the symmetries of a crystal.

Two different types of semiconductors intimately bonded together

across an interface. One is p-type and the other is n-type, meaning
that the electric current is predominantly carried by holes and by
electrons respectively.

A semiconductor which is so doped that the majority of current

carriers are (positive) holes.

An idealization of a material object, as something with mass but a

vanishingly small size.

Certain elementary particles (examples: electrons, protons,

neutrons) are such that a quantum state cannot be occupied by
more than one particle. Such particles are called fermions.
An electric current that starts going around in a superconducting ring
will continue indefinitely without diminution, because it encounters no
resistance to its flow.

245

GLOSSARY

phase of wave function

phase transformation

phonon

photon

Planck constant

plastic deformation

polarization

polarized light

polarizer

polycrystal

position vector

positron

potential energy

precipitate hardening

probability
A wave function oscillates in time as well as in space. The oscillating
amplitude at any instant and point can be related to the angle swept
by a body rotating at the same rate. This angle is the phase of the
wave function.

An abrupt change in the structure of a body which occurs at some

characteristic temperature (or pressure) and is accompanied by
abrupt changes in some of its properties. The melting of ice to form
water is an example.

The particle aspect of the quantum description of a solid’s thermal

energy, whose wave aspect is a sound wave in the solid.

The particle aspect of the quantum description of

electromagnetic radiation, whose wave aspect is an

electromagnetic wave, e.g. visible light.

A fundamental constant, denoted by h, which when multiplied by the

frequency or the wave vector gives the energy or the momentum
respectively.

A change in shape and size of an object produced by an applied

force, which persists when the force is removed.

When applied to electromagnetic waves, this means a beam of light,

for example, in which the electric field points in the same direction
everywhere. When applied to an atom, it means the slight movement
of the electrons and the nucleus relative to each other, produced by
an electric field.

A beam of light in which the electric field points in the same direction
everywhere.

Ordinary light passing through a polarizer comes out polarized.

This is the usual form of most solids, consisting of tiny crystals called
grains which are held together by the forces between atoms from
adjacent grains..

Any point in space can be identified by a vector, called its position

vector, drawn to it from some fixed point.

A particle which is exactly like the electron in all respects except one:
its electric charge is positive and equal in magnitude to the charge of
the electron.

The energy that an object has even when it is not moving.

Examples are the gravitational energy of a body, and the energy of a

charge in an electric field.

The process of making a metal harder to deform permanently, by

annealing so as to produce small clusters of the added atoms
distributed through the volume of the metal.

The probability of a given result of an experiment, which could give

more than one result, is the fraction of a large number of 246

GLOSSARY

proton

quantization

quantum mechanics

quantum number

quantum state

quark

radiation
radioactivity

reflection symmetry

refraction

resistance

rotation symmetry

scalar

scattering time

Schroedinger equation

scientific notation

such experiments in which that particular result appears. Thus the

probability of heads when I toss a coin is one-half.

A particle with a positive charge of the same magnitude as the

charge of the electron, and a mass which is about 1840 times the
electron’s mass.

Quantities like energy and momentum for an object can take only
certain discrete values, and nothing in between. This is called the
quantization of energy, momentum, etc.

A description of nature which includes both the particle-and wavelike

aspects of matter, and explains its observed properties.

A number which labels each of the discrete quantum states of

energy, momentum, spin, etc.

Any one of the possible stable states in which an object can find
itself, with an energy, a momentum, etc which are specific to that
state.
The name of the particles which make up protons and neutrons.

A proton and a neutron each consists of three quarks.

A term used here interchangeably with electromagnetic waves.

A process whereby a nucleus spontaneously ejects one or more

particles; an electron, a helium nucleus, a positron, and a photon are
among the possibile ejected particles.

An object has a plane of reflection symmetry if it looks the same

when every part of it is imagined to be reflected by the plane acting
as a mirror. In two-and one-dimensional objects, it would be a line
and a point of reflection symmetry respectively.

The change in direction of an electromagnetic wave, light for

example, in going from one medium to another, caused by a
difference in speed in the two media.

A term often used when electrical resistance is meant; the property

of a conductor whereby passage of an electrical current through it is
associated with a voltage across it.

An object has an axis of rotation symmetry if it looks the same before

and after being rotated about this axis through a specified angle.

A quantity which can be specified just by giving its magnitude: the

mass of an object, for example.

The average time between two successive collisions that particles

like electrons and phonons in a solid suffer.

A fundamental mathematical expression of quantum mechanics.

It is used to calculate the wave function, the energy levels and other
quantities of interest.

The practice of expressing very large and very small numbers as

numbers between 1 and 10 multiplied by a power of 10.
247

GLOSSARY

second sound

semiconductor

sintering

solar battery

solid state physics

solution hardening

specific heat

spectroscope

standing wave

state

statistical physics

statistics

superconducting transition

temperature

superconductivity

superconductor

surface properties
The passage of a temperature pulse along a piece of matter with a
speed somewhat less than the speed of sound; can occur when the
mean free path of the phonons is very long.

A solid with a modest energy gap between the highest fully occupied
band (the valence band) and the lowest empty band (the conduction
band), so that the effective numbers of electrons and holes can be
sensitively controlled through temperature and doping.

The process of consolidating fine powder which has been pressed

together, by heating it to a high enough temperature without melting
it.

A device made from a semiconductor, which converts sunlight into

electricity.

The physics of matter which is in the form of solids; also known as

condensed matter physics.

The process of making a metal harder to deform permanently, by

alloying it with other elements.

The amount of thermal energy needed to raise the temperature of

one gram of a substance by one degree Celsius. It is a characteristic
property of a given substance.

An instrument which enables one to sort out and measure the

different frequencies in a beam of light.

Produced by the combination of two waves of the same

wavelength and frequency travelling in opposite directions in the

same space. The standing wave has nodes and antinodes which are
fixed in space.

A term which is sometimes used, to mean quantum state.

The physics of large numbers of particles acting as a whole, the

numbers being so large that only the methods of statistics can be
meaningfully applied.

A way of handling the features of very large numbers of objects,

where one looks not at specific individuals but rather at groups.

The temperature up to which superconductivity exists in a given

material. This temperature is a characteristic of the material and is
different for different superconductors.

A phase transformation characterized by the abrupt and total loss of

electrical resistance below a certain temperature.

An electrically conducting material whose electrical resistivity

abruptly vanishes when it is cooled below a certain

temperature.

Those properties of an object which are governed primarily by one

layer, or at most a few layers, of atoms at its surface.

248

GLOSSARY

symmetric wave function

symmetry

thermal conduction

thermal conductivity

thermal energy

thermal expansion

thermal radiation
transistor

translation

translation symmetry

travelling wave

unit cell

unpolarized light

valence band

variable

vector

velocity

volt

watt

A wave function which is unchanged when two of the particles

described by it are interchanged.

An object is said to show symmetry under a given operation if it

appears unchanged after undergoing that operation.

The flow of energy in the form of heat from the hot part to the cold
part of a piece of matter.

The amount of thermal energy flowing per second from one face to
the opposite face of a one-cm cube, when the temperature
difference between the two is a degree Celsius. It is a characteristic
property of a given substance.

The energy possessed by an object by virtue of its

temperature.

The change in size and shape of an object produced by a change in

its temperature.

The radiation given off by a body because of its temperature.

A device made of a semiconductor, often silicon, which is a basic

building block in all electronic equipment.

The operation of moving something from here to there without

rotating it.

This is the symmetry an object has if it looks the same after being
moved from here to there through a specified distance without
rotation.

The usual wave, in which one sees the crests and troughs actually
going somewhere.

A small portion of a crystal lattice, consisting of just enough lattice

points to show the symmetry.

A beam of light in which there is no single direction in which the

electric field points.

The band of highest energy levels for electrons in a solid in which all
the states are occupied by electrons. The next higher band is either
empty or only partially occupied by electrons, and is called the
conduction band.

An algebraic quantity denoted by a symbol like x, to which different

values can be assigned.

A quantity specified by its magnitude as well as its direction.

Velocity and momentum are examples.

The speed of a moving object as well as the direction of its motion
are covered by this term. The velocity, having magnitude as well as
direction, is a vector.

A unit for measuring the potential energy of an electric charge.

The unit for measuring power, which is the rate at which energy is
used or produced; equal to one joule per second.

249

GLOSSARY

wave Something which oscillates, i.e. repeats its shape, at regular

intervals in space as well as in time.

wave function A mathematical function that depends on position and

time, and embodies the properties of the wave which is associated
with a particle or a group of particles.

wave number A quantity obtained by dividing 1 by the wavelength

(i.e. taking the reciprocal of the wavelength). It is the same as the
number of waves per cm.

wave vector A vector whose magnitude is the wave number and

whose direction is the direction in which the wave is travelling.

wave-particle duality The basic aspect of the world around us which

leads to quantum mechanics as its proper description: one and the
same entity shows wave-or particle-like behaviour depending on the
way in which it is observed.

The distance between two adjacent crests, or two adjacent troughs,

in a wave.

A polyhedron surrounding an atom in a crystal, of such a shape that

every point within it is nearer to this atom than to any neighbouring
atom.

work-hardening The increased resistance to plastic deformation

which can be produced in a material by a prior process of repeated

deformation.

zero-point energy A fundamental consequence of the quantum

nature of matter: the lowest possible energy of a system cannot be
zero, but must have a non-zero value which is called its zero-point
energy.

wavelength

Wigner-Seitz cell

250

INDEX

absolute zero, 91

absorption of light, 173

angular momentum, 38

atom, 75

conservation, 41

electron, 76, 204

nucleus, 211

quantized, 76

annealing, 133
atoms

chlorine, no

electronic structure, 77

impenetrability, 79—80

isotopes, 81

misleading representation, 66

Pauli exclusion principle, 78

silver, 104-106,109

size, 79

sodium, no

thermal vibrations in a crystal, 93

see also hydrogen atom

band structure, 117

copper and gold, 176—177

glass, 174

semiconductors, 195

see also energy bands

Bloch electron, 113

Bohr magneton, 204

Boltzmann constant, 91
boson, 96

distribution, 100

spin, 96

symmetric wave function, 96

Brillouin zone, 119

brittleness, 134

bulk properties, 15

charge, see electric charge

chip, 2,197

close-packed direction, 23

colour of things, 176-178

condensed matter physics, 2

conduction band, 195

conduction electrons, 112

light absorption, 175-176

specific heat, 148-149

thermal conductivity, 155

thermal energy, 123

thermal expansion, 159

conservation principles
energy, 41

momentum, 41

Cooper pair, 226-229

cosmological principle, 13

one-dimensional lattice, 15

three-dimensional lattice, 29

two-dimensional lattice, 17

critical current, 229

crystal

electron waves, 112—113

one-dimensional, 14—15

quantum-mechanical picture, 130

shape, 23, 29

silver, 8

spacial probability distribution of

electrons, 108—112

three-dimensional, 27—29

two-dimensional, 16—19

crystal structures

diamond, 112
face-centred cubic, 30

graphite, 31

rocksalt (also common salt, sodium

chloride), 31

silicon, 112

Curie temperature, 209

Debye wave number and frequency, 126

degenerate energy level, 74

diffraction, 46

light and electrons, 48

dislocations, 136-141

dispersion curve

electrons, 116

phonons, 126,132

quantum mechanical crystal, 130

see also energy bands

dispersion of light, 182-184

prism, 184

rainbow, 183

distribution
atomic speeds in helium gas, 87

bosons, 100

fermions, 98

phonons, 128

populations by age, 84

thermal radiation, 172—173

domains, see magnetic domains

elastic limit, 134

elasticity, 13, 134

electric charge, 63

electron, 64

potential energy, 65

proton, 64

source of electric field, 167

electric field, 167

effect on atom, 174,180

effect on free electrons, 175,189

in electromagnetic wave, 171

electrical resistance

definition, 190
effect of alloying, 192

effect of temperature, 191

insulators, 196

metals, 194

resistivity, 190

semiconductors, 195

table of values, 190

electromagnet, 208

electromagnetic radiation, 164-165

see also thermal radiation

electron

electric charge, 63, 64

magnetic moment, 203-204

mass, 37

particle and wave, 47-49

spin, 36, 76

spin resonance, 214

energy

and frequency, 51

definition, 38
kinetic, 38

potential, 38, 65

251

INDEX

energy (cont.)

quantization, see quantization of energy

total, 39

zero-point, 58, 92

energy bands

crystal, 106,114—117

insulators, 107-108,117

metals, 107-108,118

semiconductors, 120—123

energy levels

hydrogen atom, 72-77

many-electron atom, 77-79,105

nucleus in magnetic field, 212-213

particle in one-dimensional box, 57-58

excited state, 71
exponential notation, 3-4

Fermi energy, 100

compared to thermal energy, 123

Fermi surface, 118-119

Cooper pairs, 226, 227

effect of current, 193-194

effect of temperature, 123—124

Fermi wave vector, 118

fermion, 96

antisymmetric wave function, 96

distribution, 98

Pauli principle, 97

spin, 96

ferromagnets, 205—210

Curie temperature, 209

domains, 206—208

effect of temperature, 208-209

Heisenberg model, 205

see also magnets

field, 165-167
electric, 167-168

electromagnetic wave, 170-171

magnetic, 168-169

scalar, 167

temperature, 166

vector, 167

velocity, 167

fracture, 141-143

fundamental constants, 52

glassy solid, 4, 32

fracture, 141-143

preparation, 10,12

grains

as single magnetic domains, 206—207

in polycrystalline solid, 10

of salt and sugar, 31

ofsand, 2

gravitation, 2

ground state, 67

hydrogen atom, 67-70

high-field (HF) superconductor, 221

hole, 120,123

hydrogen atom

energy levels, 72

excited states, 71

ground state, 67-70

quantum numbersv72-77

stability, 66

impenetrability of matjer, 79-80

insulators, 107

absorption of light, 174

electrical conduction, 196 ■

energy bands, 117

thermal conductivity, 151—154

interference, 45

isotopes, 81

Josephson oscillations, 230-232

lattice, 15

parameters, 17

thermal conductivity, 153

spacing, 15

specific heat, 148

light

absorption, 173-180

colours of things, 176-178

fat, 179

glass, 174

metals, 175-176

water, 179—180

dispersion, 182-184

generation, 171—173

refraction, 180-182

linear momentum, 37

conservation, 41

wave vector, 51

low-field (LF) superconductor, 221

magnetic domains, 206-208

magnetic field, 168

effect on nucleus, 211

effect on superconductor, 219

in electric motor, 198

in electromagnetic wave, 170

magnetic moment

current loop, 222

electron, 204

nuclei, 211

proton and neutron, 204

magnetic monopole, 203

magnets

basic properties, 202—203

elementary (electron, nuclei), 203—205

permanent, 207—208

weak magnetism, 214

see also ferromagnets

mean free path

electrons and holes, 151

phonons, 150

tables of values, 151

see also scattering

mixed state, 221

modes of vibration in crystal, 125—127

MOSFET, 197

motion, 35

MRI, 214

n-type semiconductor, 123

in a MOSFET, 197

in solar battery, 184

normal state, 216, 219, 223

nuclear magnetic resonance, 211-214

nuclear magneton, 211

p-n junction, 184

p-type semiconductor, 123

in a MOSFET, 197

in solar battery, 184

particles

comparison to waves, 46

definition, 36

duality, 49, 50

energy, 38

light, 49
momentum, 37

properties, 37-40

the electron, 47

Pauli exclusion principle

antisymmetric wave function, 97

electrons in atoms, 78

persistent current, 222

phase of wave function, 231

phase transformations, 210

phonon, 124—128

Cooper pair, 226—227

distribution, 128

252

INDEX

energy, 127

mean free path, 151

scattering of electrons. 192-193

specific heat, 148

thermal conduction, 151—153

thermal expansion, 159

photon, 50

. distribution in energy, 94

energy, 51

interaction with matter, 163-187

momentum, 51

see also light

Planck constant, 52

plastic deformation, 134, 137-141

polarization

atoms, 180

light, 171

polycrystalline solid, 10

precipitate hardening, 141

probability

particle-waves, 52—54, 59—60

statistical physics, 235

quantization of angular momentum

electron, 76

nuclei, 211—214
quantization of energy

electron in box, 54—58

hydrogen atom, 72—74

phonons, 127

photons, 51

quantum mechanics, 49

some formulas, 51

quantum number, 60, 72—77

rainbow, 183

reflection symmetry, 25

refraction of light, 180—182

resistance, see electrical resistance

rotational symmetry, 23-24

scalar, 37

scalar field, 167

scattering

of electrons by phonons, 192—193

scattering time, 194-195

see also mean free path

Schroedinger equation, 59
scientific notation, 3-4

second sound, 153

semiconductor

conduction band, 195

doped, 121—122

energy bands, 120—123

n-type, 123

p-type, 123

valence band, 195

sintering, 208

solar battery, 184-186

solid state physics, 2

solution hardening, 141

specific heat

definition, 147

free electrons, 148-149

lattice (same as phqrions), 148

superconductor, 225

table of values, 147

spectroscope, 184
standing waves

electrons in box, 55, 98

microwaves in oven, 180

spin

electron, 36, 76

nucleus, 36, 211

springs between atoms, 134—136

metal, 134

silicon, 135

sodium chloride, 135

see also virtual springs

statistics, 6, 84, 235

see also distribution

superconductors, 216—234

BCS theory, 226—230

Cooper pairs, 226—227

critical current, 229

critical magnetic field, 2r9~220

Josephson oscillations, 230—232

magnets, 232
Meissner effect, 221

normal state, 216, 219, 223

persistent current, 222

phase transformation, 217

some clues, 224-225

thermal conductivity, 223, 224

transition temperature, 2r6, 218

type I (LF), 221

type II (HF), 221

zero resistance, 216, 228

surface properties, 15

symmetry, 19

reflection, 25

rotational, 23-24

space and time, 26

translational, 21-22, 29

vase and lump, 19-20

temperature, 86

absolute zero, 91

atomic speeds in helium gas, 87—89

crystals, 91-94

energy, 89

thermal conductivity

alloys, 155

insulators, 151—153

metals, 155

superconductors, 223, 224

table of values, 154

thermal energy, 9t

atomic vibrations in crystals, 124—128

electrons in metals, 123—124

thermal expansion, 156—159

table of values, 156

thermal radiation, 172—173

see also electromagnetic radiation

transistor, 2,197

transition temperature, 216, 218

translation, 21

translational symmetry, 21—22, 29

unit cell
three-dimensional lattice, 28

two-dimensional lattice, 18

unpolarized light, 171

valence band, t95

vector, 37

addition, 40

vector field, r67

velocity, 37

virtual springs, 13

see also springs between atoms

waves, 36, 42

amplitude, 43, 53

comparison to particles, 46

diffraction, 46

duality, 49, 50

electromagnetic, 170-171

see also electromagnetic radiation,

thermal radiation

frequency, 43

interference, 45
matter, 58

sound, 125

standing, 55

travelling, 55

253

&yes,(cont) degeneracy, 74 light, 49

velocity, 43 ferromagnet, 205 weak magnetism, 214

wave number, 51 hydrogen atom, 70-72 Wigner-Seitz cell, 135

wave vector, 51 identical particles, 95 work-hardening, 140

wavelength, 43 particle in box, 60, 61, 98

‘ave function phase, 231 zero-point energy

antisymmetric, 96 symmetric, 96 crystal, 92

Bloch electrons, 113 wave-particle duality, 50 particle in a box, 58,

91

Cooper pair, 231 electrons, 47